The role of Ca²⁺ stores in the muscarinic inhibition of the K⁺ current I_K(SO) in neonatal rat cerebellar granule cells

Abstract

1. Cerebellar granule neurons (CGNs) possess a standing outward potassium current (I_K(SO)) which shares many similarities with current through the two-pore domain potassium channel TASK-1 and which is inhibited following activation of muscarinic acetylcholine receptors.

2. The action of muscarine on I_K(SO) was unaffected by the M₂ receptor antagonist methoctramine (100 nm) but was blocked by the M₃ antagonist zamifenacin, which, at a concentration of 100 nm, shifted the muscarine concentration-response curve to the right by around 50-fold.

3. Surprisingly, M₃ receptor activation rarely produced a detectable increase in [Ca²⁺]_i unless preceded by depolarization of the cells with 25 mm K⁺. Experiments with thapsigargin and ionomycin suggested that the endoplasmic reticulum Ca²⁺ stores in CGNs were depleted at rest. In contrast, cerebellar glial cells in the same fields of cells possessed substantial endoplasmic reticulum Ca²⁺ stores at rest.

4. Pretreatment of the cells with BAPTA AM, thapsigargin or the phospholipase C (PLC) inhibitor U-73122 all blocked the muscarine-induced Ca²⁺ signal but had little or no effect on muscarinic inhibition of I_K(SO). Raising [Ca²⁺]_i directly with ionomycin caused a small but significant inhibition of I_K(SO).

5. It is concluded that muscarine acts on M₃ muscarinic acetylcholine receptors both to inhibit I_K(SO) and to mobilize Ca²⁺ from intracellular stores in CGNs. While the mobilization of Ca²⁺ occurs through activation of PLC, this does not seem to be the primary mechanism underlying muscarinic inhibition of I_K(SO).

Cerebellar granule neurons (CGNs) possess a standing outward potassium current, I_K(SO), which has a major role regulating the excitability of these cells (Watkins & Mathie, 1996a). I_K(SO) shares many of the properties of the current through the two-pore domain potassium channel TASK-1 (Duprat et al. 1997; Leonoudakis et al. 1998; Reyes et al. 1998; Millar et al. 2000; see also North, 2000). Of particular interest is the observation that both of these currents are inhibited following activation of muscarinic acetylcholine receptors (Millar et al. 2000). Indeed, TASK-1 current may be inhibited following activation of several different seven-transmembrane domain (7-TM) receptors, which couple to the G protein G_q (Talley et al. 2000), although it is not affected by activators of protein kinase C (Duprat et al. 1997; Leonoudakis et al. 1998). For I_K(SO) in CGNs, the muscarinic inhibition is mediated through a pertussis toxin-insensitive G protein (Watkins & Mathie, 1996a) but there is no direct evidence as to the subtype of muscarinic receptor involved. It is known that CGNs express both M₂ and M₃ muscarinic receptors but have no detectable levels of M₁ or M₄ muscarinic receptor expression (Fukamauchi et al. 1991).

Since I_K(SO) is a non-inactivating current that is inhibited following muscarinic receptor activation, it can be considered to have a number of pharmacological similarities to the M current, a K⁺ current found in many mammalian neurons (Brown, 1988; Marrion, 1997). However, unlike I_K(SO), the M current is a voltage-gated current thought to be formed from two members of the six-transmembrane domain (6-TM), KCNQ channel family, KCNQ2 and KCNQ3 (Wang et al. 1998; Selyanko et al. 2000; Shapiro et al. 2000). There is a significant amount of literature concerning the link between muscarinic receptor activation and M current inhibition (see Marrion, 1997); however, despite intense effort, no clear consensus exists as to the pathway involved. The same pathway may underlie voltage-independent inhibition of voltage-gated Ca²⁺ channels in mammalian neurons (see Hille, 1994).

Much evidence for the mechanism of modulation of M current by muscarinic receptor activation originally pointed to a role for the activated G_qα-subunit (Caulfield et al. 1994; Haley et al. 1998) in stimulating the enzyme phospholipase C (PLC) (Hille, 1994). This leads to a mobilization of both diacylglycerol and IP₃ and a subsequent rise in [Ca²⁺]_i. However, for each step in this pathway after activation of G_qα, available evidence seems equivocal (see Marrion, 1997). For example, it seems clear that Ca²⁺ can inhibit M current directly (Selyanko & Brown, 1996) but this inhibition may not underlie the mechanism by which muscarine acts to inhibit the current physiologically (Marrion, 1997; Cruzblanca et al. 1998; Del Rio et al. 1999; Shapiro et al. 2000). On the other hand, bradykinin, acting on B₂ receptors, is known to inhibit M current through mobilization of intracellular Ca²⁺ (Cruzblanca et al. 1998).

For I_K(SO), we have found that removing extracellular Ca²⁺ reduced the inhibitory action of muscarine on I_K(SO) (Watkins & Mathie 1996a). This observation suggested a potential role for Ca²⁺ in the modulatory pathway. The aim of this study was to determine the subtype of muscarinic receptor underlying inhibition of I_K(SO) in cerebellar granule cells. We then sought to clarify the role of Ca²⁺ in the modulatory pathway; in particular, we asked whether a rise in [Ca²⁺]_i, following mobilization from intracellular stores, is a key step in muscarinic inhibition of I_K(SO). In cerebellar granule cells it is often difficult to obtain a detectable rise in [Ca²⁺]_i following muscarinic receptor activation unless the Ca²⁺ stores are precharged by depolarizing the cells (see Irving et al. 1992; Masgrau et al. 2000). Therefore a further aim of this study was to investigate whether this is due to the Ca²⁺ stores being functionally depleted at rest.

Preliminary accounts of some of these observations have been published previously (Watkins et al. 1996; Boyd & Mathie, 2000).

METHODS

Tissue culture

Granule neurons were isolated using previously described methods (Huston et al. 1993; Watkins & Mathie, 1996a) from the cerebella of 6- to 8-day-old Sprague-Dawley rats of either sex which had been killed by decapitation. Following dispersion, cells were plated onto 13 mm glass coverslips coated with poly-ornithine or poly-l-lysine and allowed to adhere. They were then covered with a minimum essential medium (MEM) comprising Earle’s balanced salt solution supplemented with 10% fetal calf serum, 50 i.u. ml⁻¹ chick embryo extract, 39 mm glucose and 2 mm glucamine, and maintained in 5% CO₂ at 37°C. The medium also contained 23 mm potassium, which has been shown to enhance the viability of cerebellar granule neurons in culture.

Electrophysiological recordings

Electrophysiological recordings were made from CGNs maintained in culture for up to 14 days in a physiological saline solution containing (mm): NaCl 120, KCl 2.5, CaCl₂ 0.5, MgCl₂ 2, glucose 5 and Hepes 10, adjusted to pH 7.4. The neurons were whole-cell voltage clamped using amphotericin B-permeablized patches to minimize disruption of the intracellular composition (Rae et al. 1991). The pipette solution contained (mm): KCl 125, MgCl₂ 5, Hepes 5 and BAPTA 0.1, with 240 μg ml⁻¹ amphotericin B, adjusted to pH 7.4. External solutions were applied by bath perfusion at a rate of 4–5 ml min⁻¹ and complete exchange of the bath solution occurred within 20–40 s. Solutions were applied at room temperature (20–23°C).

Neurons were voltage clamped using an Axopatch-1D amplifier (Axon Instruments, USA) and low-pass filtered at 5 kHz before sampling and capture on-line to a PC with a Digidata 1200 interface (Axon Instruments). Data acquisition and analysis were carried out using pCLAMP software (Axon Instruments), Excel (Microsoft, USA) and Origin (Microcal Inc., USA). Voltage protocols are described in the figure legends.

Imaging intracellular calcium changes

Ca²⁺ measurements were made with fura-2 loaded into the cells as its AM ester for 20 min at 37°C. Measurements were made using dual excitation (340 and 380 nm) with emission above 420 nm. All experiments were conducted at room temperature (20–23°C). For fluorescence imaging, light was collected via a cooled CCD camera (Hamamatsu, Japan) and analysed using software supplied by Kinetic Imaging (Liverpool, UK). To improve the signal-to-noise ratio, the collection times were adjusted to increase the total number of photons accumulated at the lowest intensity of emission. Thus the collection period for 340 nm excitation was three times that for 380 nm excitation (the collection period was generally 600 ms compared to 200 ms). The F₃₄₀/F₃₈₀ fluorescence ratio values in the figures are not corrected for these different collection times. For determination of the local [Ca²⁺] from the ratio images, specific areas of interest were chosen. The average fluorescence ratio value for the designated areas was then calculated and plotted as a function of time (see Prothero et al. 1998; Jones et al. 2000).

Calibration of the fluorescence ratios in terms of [Ca²⁺] was carried out in situ in several experiments using the ionophore ionomycin in the presence of CCCP (carbonyl cyanide m-chlorophenyl hydrazone) and thapsigargin (see Jones et al. 2000). The average values from these calibrations were used to quantify intracellular [Ca²⁺] where this was considered appropriate, using the following equation (Grynkiewicz et al. 1985):

where R is the measured ratio of interest, R_min and R_max are the ratios recorded with zero and high extracellular calcium, respectively, S_f2 and S_b2 are the signals at 380 nm in zero and high calcium, respectively, and K_d is the apparent dissociation constant for fura-2 (224 nm).

We have previously found that the resting calcium concentration was around 20 nm (Jones et al. 2000), close to previously published values for these cells (e.g. Chen et al. 1999).

Throughout the text values are given as means ±s.e.m. with n as the number of cells examined. Statistical analysis was performed using Student’s t test with paired comparisons if relevant. Probabilities are given for two-tailed tests. Results were considered significantly different when P < 0.05.

All chemicals were obtained from Sigma (UK) except fetal calf serum, MEM, penicillin and streptomycin (Gibco, UK), chick embryo extract (ICN Pharmaceuticals, UK), thapsigargin and methoctramine (RBI), fura-2 AM (Molecular Probes) and BAPTA AM (Calbiochem). Zamifenacin was a generous gift from Pfizer.

RESULTS

Inhibition of I_K(SO) by M₃ muscarinic receptors

The fundamental properties of I_K(SO) and its inhibition by muscarine are illustrated in Fig. 1A and B. I_K(SO) was seen as a non-inactivating outward current when cells were held at −20 mV, which was then reduced in amplitude when the potential was stepped to −60 mV, closer to the potassium equilibrium potential (see also Watkins & Mathie 1996a; Millar et al. 2000). Routinely, I_K(SO) was measured as the standing outward current at −20 mV immediately before the step to −60 mV, i.e. after the cell had been held continuously at −20 mV for 4.2 s. The maximum inhibition obtained with 10 μm muscarine in control cells was found to be 67 ± 2% (n = 74). A concentration-response relationship obtained for concentrations of muscarine between 0.1 and 100 μm gave a control EC₅₀ for muscarine of 0.25 μm in this series of experiments (Fig. 1C and D), comparable with what we have found previously (Watkins & Mathie, 1996a). To determine the subtype of muscarinic receptor underlying the inhibition of I_K(SO), two antagonists were used. These were methoctramine, a potent antagonist at M₂ receptors (Hulme et al. 1990) and zamifenacin, which is most effective as an M₃ muscarinic receptor antagonist (Wallis, 1995). Following pretreatment with 100 nm methoctramine the concentration-response curve for muscarine was virtually unaltered (see Fig. 1C) with an EC₅₀ for muscarine of 0.44 μm and no significant difference in the amount of inhibition seen at any muscarine concentration. For example, with 1 μm muscarine, in control cells, there was 48 ± 2% (n = 16) inhibition of I_K(SO), whereas in the presence of 100 nm methoctramine, 1 μm muscarine produced 49 ± 4% inhibition (n = 7, P = 0.81, unpaired t test; see also Fig. 1A). In contrast, 1 μm muscarine produced no detectable inhibition of I_K(SO) in the presence of 100 nm zamifenacin (Fig. 1B). The concentration-response relationship for muscarine was shifted to the right by around 50-fold with an EC₅₀ of 11.8 μm for muscarine in the presence of 100 nm zamifenacin (Fig. 1D). The results with these two blocking drugs strongly suggest that M₃ receptors underlie the muscarinic inhibition of I_K(SO) (see also Discussion).

Figure 1. Muscarine inhibits I_K(SO) through an action on M₃ muscarinic receptors.

A and B, in perforated-patch recordings, I_K(SO) is seen as a non-inactivating current at −20 mV and is instantaneously reduced in amplitude when cells are stepped to −60 mV for 800 ms, once every 5 s. The current is inhibited by muscarine (1 μm) and this inhibition is unaffected by methoctramine (100 nm; see A) but blocked by zamifenacin (100 nm; see B). Concentration-response relationships for inhibition of I_K(SO) by muscarine were obtained in control conditions and in the presence of methoctramine (C) and zamifenacin (D).

M₃ muscarinic receptor-mediated Ca²⁺ signalling is enhanced by prior depolarization of CGNs

Muscarinic receptor activation is known to cause a rise in [Ca²⁺]_i in CGNs through stimulation of M₃ receptors (Whitham et al. 1991). However, as has been seen previously (e.g. Irving et al. 1992; Masgrau et al. 2000), we have found it quite difficult to obtain detectable rises in [Ca²⁺]_i in these cells under normal conditions. Calcium responses to muscarine can be obtained much more reliably, however, following prior exposure of the cells to a depolarizing, high K⁺, extracellular solution. High K⁺ itself caused a large rise in [Ca²⁺]_i in CGNs but little change in cerebellar glial cells (see Fig. 2). In one series of experiments in normal solution, muscarine (10 μm) caused a detectable transient increase in [Ca²⁺]_i in 8 out of 87 cells (9%); in the remaining 79 cells no response was observed. Prior exposure to 25 mm K⁺, however (see Fig. 3), increased the number of responsive cells to 51% (33 out of 65), and in those cells muscarine induced a peak rise in [Ca²⁺]_i with a F₃₄₀/F₃₈₀ ratio of 0.88 ± 0.03 from a resting level of 0.67 ± 0.03, which corresponds to a peak increase in [Ca²⁺]_i of around 380 nm.

Figure 2. Depolarization of CGNs with 25 mm K⁺ increases [Ca²⁺]_i.

A, phase-contrast image of cerebellar granule cells and glial cells in culture. CGNs are identified as small circular cells (cells 1–5) with bright halos. A glial cell (cell 6) is also identified. B, plot of fluorometric ratio against time for the cells highlighted in A before, during and after bath application of K⁺ (25 mm). An increase in the F₃₄₀/F₃₈₀ ratio reflects an increase in [Ca²⁺]_i. Note, in contrast to the CGNs, the lack of an increase in the glial cell. C-E, individual images taken at rest (C), during K⁺ application (D) and after wash (E). Dark colours (black, blue) indicate low [Ca²⁺], while progressively warmer colours (green, yellow, red) indicate higher [Ca²⁺]. Scale bar (A), 20 μm.

Figure 3. Muscarine increases [Ca²⁺]_i in CGNs following prior depolarization of the cells with 25 mm K⁺.

Following prior depolarization with 25 mm K⁺, 33 out of 65 cells (51%) responded to muscarine (10 μm). F₃₄₀/F₃₈₀ ratio values were obtained as shown for Fig. 2.

M₃ muscarinic receptor-mediated mobilization of intracellular Ca²⁺ and the inhibition of I_K(SO)

To determine whether there is a link between M₃ muscarinic receptor-mediated rises in [Ca²⁺]_i and inhibition of I_K(SO), experiments were carried out to disrupt muscarinic receptor-mediated mobilization of Ca²⁺. The effects of these manipulations on M₃ muscarinic receptor-mediated inhibition of I_K(SO) were then measured.

BAPTA AM

The Ca²⁺ signal evoked following activation of M₃ muscarinic receptors could be completely inhibited by pretreating the cells with the Ca²⁺ chelator BAPTA in its membrane-permeable form, BAPTA AM (see Watkins et al. 1996). It was of interest to determine whether pretreatment with BAPTA AM would also affect M₃ muscarinic receptor coupling to I_K(SO). However, we, and others, have shown previously that BAPTA AM treatment can reversibly and directly inhibit certain potassium currents including I_K(SO) (Watkins & Mathie 1996b; Urbano & Buno, 1998). Exactly the same experimental protocol that in our previous study produced block of muscarinic receptor-mediated Ca²⁺ mobilization was used in these experiments on I_K(SO). This comprised a 10 min incubation in 25 μm BAPTA AM followed by a 5 min wash. Following this treatment, I_K(SO) amplitude recovered to only 80 ± 5% (n = 11) of its magnitude before exposure to BAPTA AM. Notwithstanding this effect of BAPTA AM, muscarine inhibition of I_K(SO) was, itself, significantly reduced following treatment with BAPTA AM, with a control inhibition by muscarine of 74 ± 3% (n = 11) but only 53 ± 5% (n = 11, P < 0.005, paired t test) inhibition in the same cells following BAPTA AM treatment. Thus treatment with BAPTA AM completely abolished the rise in [Ca²⁺]_i produced by 10 μm muscarine and significantly, but only partially, reduced the inhibition of I_K(SO) by muscarine.

Thapsigargin

CGNs were exposed to the SERCA (sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase) inhibitor thapsigargin (1 μm) to deplete ER Ca²⁺ stores, in Ca²⁺-containing external solution (Fig. 4A). This induced a surprisingly small rise in [Ca²⁺]_i in CGNs compared with that seen in cerebellar glial cells in the same fields of cells. In CGNs, [Ca²⁺]_i increased from a resting fluorescence ratio of 0.58 ± 0.02 (n = 51) to a plateau level of 0.63 ± 0.02 following treatment with thapsigargin. In the glial cells in the same fields, there was a much more substantial, sustained [Ca²⁺]_i rise from a resting fluorescence ratio of 0.73 ± 0.09 (n = 7) to 1.12 ± 0.13. Nevertheless, in CGNs, subsequent application of muscarine, following depletion of the Ca²⁺ stores with thapsigargin, produced no detectable rise in [Ca²⁺]_i in any cells with or without prior depolarization with high K⁺. In contrast, pretreatment with thapsigargin had no effect on the inhibition of I_K(SO) by muscarine (10 μm; Fig. 4B). If anything, inhibition was more marked following thapsigargin pretreatment, with 83 ± 4% (n = 6) inhibition obtained.

Figure 4. Thapsigargin has little direct effect on [Ca²⁺]_i in CGNs (in contrast to glial cells) and does not block muscarinic inhibition of I_K(SO).

A, the effect of thapsigargin (1 μm) on [Ca²⁺]_i in CGNs (▪, n = 51) and glia (^, n = 7). Thapsigargin was applied by continuous bath perfusion at the arrow. Note the small rise in [Ca²⁺]_i in CGNs compared to glia. B, muscarine (10 μm) inhibits I_K(SO) following pretreatment of the cells with thapsigargin (1 μm for 15 min). I_K(SO) was measured as the standing outward current at −30 mV once every 5 s.

U-73122

The muscarinic receptor-induced rise in [Ca²⁺]_i was abolished by pretreating cells with U-73122 (10 μm), an inhibitor of the enzyme PLC (see Fig. 5A). As with BAPTA AM pretreatment, muscarinic inhibition of I_K(SO) was significantly reduced, but not blocked, by 10 μm U-73122. In five cells, inhibition by 10 μm muscarine was 75 ± 4% following the first muscarine application but only 63 ± 4% in the same five cells following 5 min incubation in 10 μm U-73122 (Fig. 5B). This is a small but significant reduction in the response to muscarine (P = 0.003, paired t test). In contrast, in eight control cells the second application of muscarine following 5 min perfusion in control external solution produced an inhibition that was not significantly different from the first application in the same cells.

Figure 5. U-73122 blocks muscarinic mobilization of Ca²⁺ in CGNs but does not block muscarinic inhibition of I_K(SO).

A, the effect of muscarine (1 μm) on [Ca²⁺]_i in control cells (▪ SR, n = 20) and cells pretreated with U-73122 (10 μm; ^, n = 20). B, muscarine (10 μm) inhibits I_K(SO) following treatment of the cells with U-73122 (10 μm). I_K(SO) was measured as the standing outward current at −20 mV once every 5 s.

Taken together, these three sets of experiments suggest that M₃ receptor stimulation of PLC and subsequent Ca²⁺ mobilization may account for around a quarter, at best, of the inhibition of I_K(SO) following muscarinic receptor activation and that some other as yet unidentified modulatory pathway must underlie the bulk of the inhibition.

Store depletion and Ca²⁺ entry pathways

Both the small number of CGNs giving a detectable rise in [Ca²⁺]_i in response to muscarine (without pre-depolarization of the cells) and the small changes in [Ca²⁺]_i evoked by thapsigargin are surprising observations; however, similar results have been obtained by others (e.g. Irving et al. 1992; Toescu, 1998) for these neurons. One explanation may be that, at rest, the Ca²⁺ stores in CGNs are depleted and that they are subsequently transiently replenished following pre-depolarization of the neurons with high K⁺. This contrasts with what is often seen in other mammalian neurons and glial cells (e.g. cortical neurons and glia; Prothero et al. 1998, 2000) where substantial [Ca²⁺]_i rises following activation of G_q-coupled metabotropic receptors can be observed without the requirement for any loading of intracellular stores. Related to this, there does not seem to be a significant stimulation of Ca²⁺ entry following store depletion in CGNs compared to what has been observed in many other cell types. This is illustrated in Fig. 6, where cells were treated with thapsigargin in the absence of extracellular Ca²⁺, then Ca²⁺ was subsequently restored to the extracellular solution. In the glial cells, there was a large rise in [Ca²⁺]_i of around 670 nm (which was inhibited by La³⁺ and Zn²⁺ ions, data not shown) when Ca²⁺ was added back to the external recording solution. This is indicative of the presence of a store-operated Ca²⁺ entry pathway (see e.g. Hoth & Penner, 1993; Prothero et al. 1998). In contrast, a much smaller rise in [Ca²⁺]_i of around 40 nm was seen in the CGNs when they were treated in the same manner.

Figure 6. Restoring extracellular Ca²⁺ to thapsigargin-treated cells induces a large rise in [Ca²⁺]_i in glia and a smaller rise in CGNs.

The effect of thapsigargin (1 μm) pretreatment (in Ca²⁺-free external solution) on the [Ca²⁺]_i in CGNs (▪, n = 44) and glia (^, n = 5). Ca²⁺ (2 mm) was restored to the extracellular solution by continuous bath perfusion at the arrow and removed again as indicated. Note the small rise in [Ca²⁺] in CGNs compared to glia.

Status of Ca²⁺ stores in CGNs

We examined the status of the CGN Ca²⁺ stores further using the Ca²⁺ ionophore ionomycin. Ionomycin (5 μm), as expected, caused a large sustained rise in [Ca²⁺]_i in the presence of external Ca²⁺ (2 mm) and a smaller but significant transient rise in the absence of external Ca²⁺ (see Fig. 7), the latter reflecting release of Ca²⁺ from intracellular stores. Prior stimulation with high K⁺ increased the response to ionomycin in Ca²⁺-free external solution (also containing 100 μm BAPTA) in CGNs. In one series of experiments in Ca²⁺-free solution the peak percentage increase in the F₃₄₀/F₃₈₀ ratio in response to ionomycin was to 122 ± 4% (n = 51) of the resting level in control conditions but to 138 ± 3% (n = 92) of the resting level following pre-stimulation with high K⁺ in the presence of extracellular Ca²⁺.

Figure 7. Thapsigargin has different effects on the ionomycin response in CGNs and glia.

A, the effect of ionomycin (5 μm) in Ca²⁺-free external solution on [Ca²⁺]_i in control CGNs (^, n = 46) and cells pretreated with thapsigargin (1 μm;, n = 50). Ratio values are normalized to the average resting value obtained before addition of ionomycin in each condition (F/F_O). B, the same experiment as in A for glial cells seen in the same fields of cells for control cells (^, n = 5) and cells pretreated with thapsigargin (▪, n = 8). Note the difference in scale and the much larger control responses in the glia.

It is possible to determine the contribution of thapsigargin-sensitive stores to the ionomycin-evoked response in Ca²⁺-free solutions by comparing the ionomycin responses obtained under normal conditions with those obtained in cells pretreated with thapsigargin. In such experiments, thapsigargin pretreatment had no effect on the ionomycin-induced rise in [Ca²⁺]_i in CGNs in Ca²⁺-free external solution (see Fig. 7A). In one series of paired experiments on CGNs (see Fig. 7), the ionomycin-induced rise in the F₃₄₀/F₃₈₀ ratio was to 133 ± 3% (n = 46) of the resting level in control CGNs and to 137 ± 4% of the resting level (n = 50) in thapsigargin-treated cells (1 μm for 15 min). In contrast, the ionomycin-induced response seen in control glial cells was much larger than that seen in CGNs and thapsigargin pretreatment greatly reduced this response (Fig. 7B). In glial cells in the same fields of cells, the rise in the F₃₄₀/F₃₈₀ ratio was to 289 ± 108% (n = 5) of the resting level in control cells but to only 137 ± 9% (n = 8) of the resting level in thapsigargin-treated cells. These results provide strong evidence to support the hypothesis that in CGNs (in contrast to cerebellar glial cells) the Ca²⁺ stores are depleted at rest.

Direct inhibition of I_K(SO) by Ca²⁺

In view of the observations above that the Ca²⁺ stores of CGNs are depleted at rest, it is possible that muscarine-induced Ca²⁺ mobilization may make a small contribution to the inhibition of I_K(SO) simply because the stores do not contain much Ca²⁺ under resting conditions. We examined this further by considering more directly the contribution of Ca²⁺ to inhibition of I_K(SO) by treating cells with ionomycin in the presence of extracellular Ca²⁺ (Fig. 8). Ionomycin treatment led to a reversible 14 ± 4% (n = 7) inhibition of I_K(SO)– about one-fifth the magnitude of the response seen to muscarine in the same cells (Fig. 8). This suggests that Ca²⁺ has only a small inhibitory action on I_K(SO) and that it is not the depleted state of the Ca²⁺ stores, per se, which determines the comparatively small contribution of this pathway to the inhibition of I_K(SO) by muscarine.

Figure 8. Ionomycin treatment causes a small inhibition of I_K(SO).

Ionomycin (5 μm) (in the presence of extracellular Ca²⁺) inhibits I_K(SO) but to a much smaller degree than the inhibition evoked by muscarine (10 μm). I_K(SO) was measured as the standing outward current at −20 mV once every 5 s. Representative current traces are shown in the inset for control currents and currents in the presence of either ionomycin or muscarine.

DISCUSSION

M₃ muscarinic receptors inhibit I_K(SO)

We have shown previously that muscarine inhibits I_K(SO) in cerebellar granule cells; however, the subtype of muscarinic receptor involved has not been identified (Watkins & Mathie, 1996a). CGNs possess mRNA for both M₂ and M₃ muscarinic receptors but not M₁ and M₄ (Fukamauchi et al. 1991) so we have used two pharmacological agents, methoctramine, which is a relatively selective antagonist at M₂ receptors, and zamifenacin, a relatively selective antagonist at M₃ receptors, to help distinguish, functionally, which receptor is important.

Zamifenacin has a pA₂ value (−log of the concentration producing a 2-fold rightward shift of the concentration-response curve) of either 9.2 or 7.9 for endogenous M₃ receptors depending on their tissue location, and a pA₂ value of 7.14 for M₂ receptors (Wallis, 1995). Methoctramine has a pA₂ value of 7.9 for M₂ receptors and 6.2 for M₃ receptors (Hulme et al. 1990). In this study, we used each compound at a concentration of 100 nm. Thus, if the endogenous receptors underlying the modulation were M₂, we can calculate, from the published pA₂ values, that 100 nm methoctramine would be expected to shift the muscarine concentration-response curve to the right by about 20-fold while 100 nm zamifenacin would shift the same concentration-response curve to the right by only around 2-fold. If, on the other hand, the receptors were of the M₃ subtype, we can calculate an expected shift of the muscarine concentration-response curve to the right by, at most, about 1.3-fold by methoctramine while zamifenacin should cause a shift to the right by at least 20-fold. The results we have obtained show virtually no effect of 100 nm methoctramine on the muscarine concentration-response curve while 100 nm zamifenacin shifts it to the right by about 50-fold. These results strongly suggest that M₃ muscarinic receptors transduce the modulation of I_K(SO) by muscarine.

Release of Ca²⁺ from intracellular stores

The most well-established second messenger pathway that is stimulated following activation of M₃ muscarinic receptors is the G_qα-PLC pathway. This G protein (G_q) is pertussis toxin insensitive and in agreement with this we have found that muscarinic inhibition of I_K(SO) is also pertussis toxin insensitive (Watkins & Mathie, 1996a).

In cerebellar granule neurons it has been shown that activation of M₃ muscarinic receptors stimulates the G_qα-PLC pathway to mobilize intracellular Ca²⁺ (see e.g. Whitham et al. 1991; Simpson et al. 1996). However, we have found that in control conditions, very few cells respond to muscarine with a detectable rise in [Ca²⁺]_i. This has been observed previously, both in cerebellar granule cells (e.g. Irving et al. 1992; Masgrau et al. 2000) and in other mammalian neurons (e.g. Irving & Collingridge, 1998). The response to muscarine is considerably larger and more frequently observed, however, if the neurons are pretreated with a high K⁺ external solution (25 mm K⁺) which will depolarize neurons such as CGNs to a potential of around −40 to −45 mV (e.g Marchetti et al. 1995; Millar et al. 2000).

The most likely explanation for this enhanced response to muscarine is that the Ca²⁺ stores are depleted under resting conditions and need, somehow, to be charged in order to allow a significant muscarine-mediated mobilization of Ca²⁺ from the stores. The content of intracellular Ca²⁺ stores can be estimated from the ionomycin-induced Ca²⁺ response in Ca²⁺-free external solution. Our data, which show that the stores in CGNs are enhanced by prior depolarization and are relatively unaffected by pretreatment with the SERCA blocker thapsigargin, provide direct evidence to support this idea. In hippocampal neurons, Ca²⁺ stores are similarly depleted at rest and it has been shown that these stores can be charged by Ca²⁺ entering the cell through L-type voltage-gated Ca²⁺ channels (Irving & Collingridge, 1998). The small response in CGNs to thapsigargin together with the low apparent Ca²⁺ content of the stores suggests that these cells possess little in the way of a capacitative Ca²⁺ entry pathway (e.g. Putney, 1986; Parekh & Penner, 1997). In our further experiments in cells pretreated with thapsigargin, the small elevation in [Ca²⁺]_i seen when Ca²⁺ was added back to the external solution would also suggest that there is comparatively little Ca²⁺ entry through store-operated Ca²⁺ channels. In contrast, cerebellar glial cells present in the same fields of cells showed a substantial initial rise in [Ca²⁺]_i in response to thapsigargin, a large sustained rise in [Ca²⁺]_i when Ca²⁺ was added back to the external solution bathing cells pretreated with thapsigargin and a large thapsigargin-dependent response to ionomycin in Ca²⁺-free external solution. These three sets of data suggest that, in contrast to CGNs and hippocampal neurons (Irving & Collingridge, 1998), cerebellar glial cells possess functional Ca²⁺ stores and a Ca²⁺-entry pathway activated on store depletion. A similar Ca²⁺ entry pathway has been seen in glial cells in many other regions of the mammalian CNS (see Verkhratsky et al. 1998) and in certain mammalian neurons (e.g. Prothero et al. 1998, 2000).

In some mammalian neurons it has been proposed that intracellular Ca²⁺ stores may amplify the Ca²⁺ entry signal following membrane depolarization through Ca²⁺-induced Ca²⁺ release from the stores (e.g. Llano et al. 1994). Therefore, the fact that the Ca²⁺ stores in CGNs are depleted at rest makes their functional role in these cells less obvious. Depleted stores would allow the ER to buffer Ca²⁺ during periods of synaptic activity (Toescu, 1998). However, more pertinently, the increased response to G protein-coupled receptors such as M₃ muscarinic receptors following depolarization of the membrane may allow these receptors to act, physiologically, as co-incidence detectors responding only when the G protein-coupled receptors are activated at around the same time as depolarization following synaptic activity (Irving & Collingridge, 1998). Similarly, in this way, the degree of accumulation of Ca²⁺ into ER stores may be determined by recent synaptic activity. After a period of intensive activity the Ca²⁺ concentration in the stores may be sufficient to allow transmission of a regenerative Ca²⁺ signal to different spatial locations in the neuron (Berridge, 1998).

M₃ muscarinic receptor-mediated inhibition of I_K(SO): the role of Ca²⁺ release from stores

In terms of the regulation of I_K(SO) following M₃ muscarinic receptor activation, the fact that the Ca²⁺ stores are depleted at rest might suggest that this is unlikely to be the mechanism by which the receptors inhibit the current. However, it may be that small changes in [Ca²⁺]_i, below the levels of detection using whole-cell Ca²⁺ imaging, are sufficient to inhibit the current. In this study, three sets of experiments suggest that this pathway plays, at best, only a minor role in the inhibition of I_K(SO) following muscarinic receptor activation. Treating cells with the Ca²⁺ chelator BAPTA AM, with thapsigargin or with the PLC inhibitor U-73122 led to complete inhibition of muscarine-induced changes in [Ca²⁺]_i but had either no effect or only a small effect on muscarinic receptor-mediated inhibition of I_K(SO). Both BAPTA AM treatment and U-73122 treatment significantly reduced the degree of inhibition of I_K(SO) by muscarine but the majority of the inhibition remained. For M₁ muscarinic receptor-mediated inhibition of the M current in rat sympathetic neurons, it is clear that although these receptors stimulate the G_q-PLC pathway, this is not the mechanism by which they inhibit M current (Cruzblanca et al. 1998; Del Rio et al. 1999). On the other hand this is exactly the mechanism whereby bradykinin acting on B₂ receptors inhibits M current in the same cells (Cruzblanca et al. 1998). Therefore it is not unreasonable to suggest that M₃ muscarinic receptors in CGNs inhibit I_K(SO) in part through Ca²⁺ mobilization following G_q activation of PLC, just as bradykinin inhibits M current. However the primary mechanism of inhibition may be through an as yet unidentified second messenger pathway as would seem to be the case for muscarinic inhibition of M current.

It does not appear that the G_q-PLC pathway would necessarily contribute a greater proportion of the muscarinic modulation of I_K(SO) if the ER Ca²⁺ stores were not depleted at rest, since raising [Ca²⁺]_i directly with the Ca²⁺ ionophore ionomycin only led to around 14% inhibition of the current. Interestingly, a direct modulatory action of Ca²⁺ on M current in sympathetic neurons has been shown (Selyanko & Brown, 1996), although the [Ca²⁺]_i rises induced by bradykinin in these cells do not appear to be large enough to suggest that it is Ca²⁺ itself that mediates the inhibition of M current. Rather, some subsequent Ca²⁺-dependent intermediatory process has been proposed (Cruzblanca et al. 1998). Similarly, our original observation that removing extracellular Ca²⁺ increased the amplitude of I_K(SO) and reduced the degree of muscarinic inhibition of the current (Watkins & Mathie, 1996a) suggests that Ca²⁺ has an important role in the modulation of I_K(SO) even if it does not appear to act as the primary second messenger in the modulatory pathway.

While a rise in [Ca²⁺]_i does not seem to be the primary mechanism whereby muscarine inhibits I_K(SO), the rise in [Ca²⁺]_i evoked by ionomycin does inhibit the current. A number of other K⁺ channels are subject to modulation following changes in [Ca²⁺]_i. For example, we have shown that the delayed-rectifier current in CGNs is inhibited following AMPA receptor-mediated [Ca²⁺]_i rises (Jones et al. 2000; see also Muller et al. 1992). Similarly, certain 6-TM domain K⁺ channels, such as Kv1.2, can be inhibited following rises in [Ca²⁺]_i, either directly, or through stimulation of protein kinases such as the tyrosine kinase PYK2 (e.g. Lev et al. 1995). Furthermore, the recently discovered Kv channel-interacting proteins (ChiPs) both bind Ca²⁺ ions and regulate the activity of A-type 6-TM K⁺ channels (An et al. 2000).

In conclusion, we have shown that muscarine acts on M₃ muscarinic acetylcholine receptors to inhibit I_K(SO) in CGNs. These receptors can also mobilize Ca²⁺ from intracellular stores in CGNs through activation of PLC but the Ca²⁺ stores in these cells are depleted at rest. Furthermore, the inhibition of I_K(SO) by M₃ receptor activation does not seem to be mediated primarily through activation of PLC and mobilization of intracellular Ca²⁺.