Rat supraoptic magnocellular neurones show distinct large conductance, Ca²⁺-activated K⁺ channel subtypes in cell bodies versus nerve endings

Abstract

1. Large conductance, Ca²⁺-activated K⁺ (BK) channels were identified in freshly dissociated rat supraoptic neurones using patch clamp techniques.

2. The single channel conductance of cell body BK channels, recorded from inside-out patches in symmetric 145 mM K⁺, was 246.1 pS, compared with 213 pS in nerve ending BK channels (P < 0.01).

3. At low open probability (P_o), the reciprocal of the slope in the ln(NP_o)-voltage relationship (N, number of available channels in the patch) for cell body and nerve ending channels were similar: 11 vs. 14 mVper e-fold change in NP_o, respectively.

4. At 40 mV, the [Ca²⁺]_i producing half-maximal activation was 273 nM, as opposed to > 1.53 μM for the neurohypophysial channel, indicating the higher Ca²⁺ sensitivity of the cell body isochannel.

5. Cell body BK channels showed fast kinetics (open time constant, 8.5 ms; fast closed time constant, 1.6 and slow closed time constant, 12.7 ms), identifying them as ‘type I’ isochannels, as opposed to the slow gating (type II) of neurohypophysial BK channels.

6. Cell body BK activity was reduced by 10 nM charybdotoxin (NP_o, 37 % of control), or 10 nM iberiotoxin (NP_o, 5 % of control), whereas neurohypophysial BK channels are insensitive to charybdotoxin at concentrations as high as 360 nM.

7. Whilst blockade of nerve ending BK channels markedly slowed the repolarization of evoked single spikes, blockade of cell body channels was without effect on repolarization of evoked single spikes.

8. Ethanol reversibly increased neurohypophysial BK channel activity (EC₅₀, 22 mM; maximal effect, 100 mM). In contrast, ethanol (up to 100 mM) failed to increase cell body BK channel activity.

9. In conclusion, we have characterized BK channels in supraoptic neuronal cell bodies, and demonstrated that they display different electrophysiological and pharmacological properties from their counterparts in the nerve endings.

Rat magnocellular neurones with cell bodies in the supraoptic and paraventricular nuclei of the hypothalamus send their axonal nerve endings into the neurohypophysis, where they release oxytocin or vasopressin into the bloodstream during highly specific firing patterns. Peptide release from the neurohypophysis is closely controlled by an interplay between the duration and frequency of the action potential burst, and the silence that separates the bursts generated by these neurones (Cazalis et al. 1985). Thus, understanding of peptide release from the hypothalamic-neurohypophysial system requires characterization of the conductances that shape action potentials in both the cell bodies and their axonal endings. The physical separation of cell bodies and nerve endings in this system allows a unique opportunity to elucidate the differences in neuronal excitability, and underlying conductances in different cellular domains.

Large conductance Ca²⁺-activated K⁺ (BK) channels are ubiquitous in neurones. They typically exhibit high permeability and selectivity for K⁺, and activation upon increases in [Ca²⁺]_i and/or depolarization (for reviews see: McManus, 1991; Kaczorowski et al. 1996), which make them effective ‘negative feedback’ regulators of depolarization and Ca²⁺ entry, and subsequent processes, such as neurotransmitter release (Robitaille & Charlton, 1992). They contribute to repolarization of individual action potentials, and/or hyperpolarizations following trains of spikes (Crest & Gola, 1993). Mutations in the Drosophila slowpoke gene, which encodes BK channels (Kaczorowski et al. 1996), cause abnormally broad action potentials (Elkins et al. 1986).

Ca²⁺ entry produces opposite effects on action potential shape in neurohypophysial nerve endings vs. supraoptic cell bodies. Exposure of nerve endings to extracellular solutions containing either zero Ca²⁺, Ca²⁺-channel blockers, or the tetraethylammonium ion, slows repolarization and broadens individual spikes (Wang & Lemos, 1995), which suggests the participation of Ca²⁺-sensitive voltage-gated K⁺ conductances in spike repolarization. Indeed, BK channels have been characterized at both macroscopic and single channel levels in these nerve endings (Wang et al. 1992). In contrast, exposure of cell bodies to zero Ca²⁺ or Ca²⁺-channel blockers speeds repolarization and narrows the spike (Widmer et al. 1998), consistent with a reduced Ca²⁺ influx, but little, if any, contribution of BK channel activity to spike repolarization. Furthermore, iberiotoxin, a specific blocker of BK channels, fails to modify the shape of the cell body spike (Widmer et al. 1998). Using a slice preparation, it has recently been shown that the slow after-hyperpolarization in supraoptic neurones is blocked by charybdotoxin, but not by iberiotoxin, these data being interpreted as a blockade of an intermediate conductance, Ca²⁺-activated K⁺ channel (Greffrath et al. 1998). Together, these results are consistent with either the absence of BK channels in the cell bodies, or the existence of a BK channel phenotype with properties markedly different from those of the nerve ending channel.

Differential distribution of BK channel types has been reported among different hair cells in the cochlea using the RT-PCR technique (Navaratnam et al. 1997; Rosenblatt et al. 1997). BK channels from vascular smooth muscle or brain, when incorporated into lipid bilayers, exhibit a variety of phenotypes (isochannels) (Reinhardt et al. 1989; Toro et al. 1991; Sansom & Stockand, 1994). However, the topology and function of distinct BK isochannels regionally segregated in the same neurone have not been reported. In the present paper, we describe for the first time, biophysical and pharmacological characteristics of a BK channel subtype in supraoptic cell bodies, allowing a comparison with the BK channel of the corresponding nerve endings in the neurohypophysis. We also investigate the contribution of each BK channel to single spike repolarization within their particular neuronal domain. Moreover, nerve ending BK channels have been shown to be targets of alcohol action (Dopico et al. 1996), possibly contributing to drug-induced reduction of neuropeptide release from the neurohypophysis. The physical separation of cell bodies and nerve endings in the hypothalamo-neurohypophysial system allow us to examine the regionalization of responses of an ion channel class within a neurone to ethanol. A preliminary set of data has appeared in an abstract form (Dopico et al. 1998b).

METHODS

Isolation of supraoptic neurones

One pair of male CD rats (Charles River, Boston, MA, USA), aged 29–44 days, was used for each experiment. Animals were decapitated according to local and national guidelines, and the supraoptic nuclei were dissected out and prepared as previously described (Dayanithi et al. 1996; Widmer et al. 1998). The neurones selected for recording had cell bodies with a diameter larger than 14 μm and often retained two to three short processes (Dayanithi et al. 1996; Widmer et al. 1998). Using immunocytochemical labelling, it has been shown that > 96 % of these large neurones stain positive for either oxytocin or vasopressin and, thus, are identified as supraoptic neurones (Oliet & Bourque, 1992). Immediately after being isolated, neurones were allowed to attach to the bottom of a dish coated with poly-L-lysine hydrobromide (10 mg ml⁻¹). Before recording, supraoptic neurones were rinsed for 10 min with a solution containing (mM): 100 sodium gluconate, 45 potassium gluconate, 5 EGTA, 4.11 CaCl₂, 1 MgCl₂, 15 Hepes, 10 glucose, pH 7.35.

Isolation of neurohypophysial nerve endings

Male rats were rendered unconscious with CO₂, then decapitated. The neurohypophysis was removed and homogenized, and the homogenate was then dropped into a poly-L-lysine-coated 35 mm dish, as previously described (Wang et al. 1992). Isolated neurohypophysial endings, characterized by lack of nuclei, spherical shape, size (3–8 μm in diameter), and golden colour under Hoffman-modulation optics, could be easily identified using an inverted microscope (Nikon, Japan). Nerve endings were then perfused with normal Locke's saline (mM): 145 NaCl, 5 KCl, 2.2 CaCl₂, 2 MgCl₂, 10 Hepes, 15 glucose, pH 7.3.

Single channel recordings from the cell bodies/nerve endings of supraoptic neurones

Unitary current recordings were obtained from excised, inside-out (I/O) and outside-out (O/O) membrane patches using standard patch clamp techniques. Unless otherwise stated, I/O recordings were obtained in ‘symmetrical conditions’, that is, the same solution bathed both the extracellular and the cytosolic sides of the patch membrane. In experiments to study the Ca²⁺ sensitivity of BK channels, the intracellular surface of I/O patches was exposed to solutions in which the final free [Ca²⁺] ranged from 0.01–10 μM. Bath solutions in which the final free [Ca²⁺] was < 1 μM, contained (mM): 145 potassium gluconate, 5 EGTA, 1 MgCl₂, 15 Hepes, pH 7.35, with varying amounts of CaCl₂. Bath solutions in which the final free [Ca²⁺] was ≥ 1 μM, contained (mM): 145 potassium gluconate, 5–6.5 EGTA, 1 MgCl₂, 15 Hepes, pH 7.35, with varying amounts of HEDTA and CaCl₂. A similar solution was used as extracellular solution, the free [Ca²⁺] of which was usually 3–10 μM. For O/O patches, the ‘electrode’ solution contained (mM): 145 potassium gluconate, 5 EGTA, 3 MgCl₂, 15 Hepes, ∼0.001 free Ca²⁺, pH 7.35. The bath solution contained (mM): 145 potassium gluconate, 5 EGTA, 1 MgCl₂, 15 Hepes, ∼0.001 free Ca²⁺, pH 7.35. In all cases, the free [Ca²⁺] was adjusted to the desired nominal value by adding CaCl₂ to the solutions. The calculation of the final [Ca²⁺]_free was computed using a custom-written computer program, based upon calculations by Fabiato (1988). In all solutions, the pH was adjusted by addition of the hydroxide of the bulk cation.

Electrodes were made as previously described (Dopico et al. 1996) to give resistances of 5–10 MΩ when filled with electrode solutions (for compositions, see above). An Ag-AgCl pellet connected through an agar bridge was used as bath electrode. After excision from the cell body/nerve ending, the membrane patch was exposed to a stream of experimental solution flowing from a micropipette (‘sewer pipette’; 1 mm diameter, WPI Inc., Sarasota, FL, USA). Whereas solutions containing charybdotoxin, iberiotoxin, apamin, or tetrandrine were applied to the extracellular surface of O/O patches, solutions containing varying amounts of ethanol or Ca²⁺ were applied to the intracellular surface of I/O patches. Superfusion with dextrose isosmotically replacing ethanol was used as control for ethanol superfusion. All the blockers were diluted from stocks to final concentration in bath solution for O/O patches (for composition, see above). Ethanol was diluted in bath solution for I/O patches. Both ethanol and the different blockers used were freshly diluted in the corresponding solution immediately before each experiment.

Single channel currents were recorded using a patch clamp amplifier (EPC7, List Electronics, Darmstadt, Germany) at a bandwidth of 3 kHz, and low-pass filtered at 1 kHz using an eight-pole Bessel filter (model 902LPF, Frequency Devices, Haverhill, MA, USA). Data were acquired and stored using an A/D converter (model TL-1, Axon Instruments, Inc., Foster City, CA, USA) and an IBM-compatible computer. Data acquisition and analysis were performed using pCLAMP software, version 6.0.2 (Axon Instruments, Inc.). Data were sampled at a bandwidth of 5 kHz. The product of the total number of functional channels present in the patch membrane (N) and the probability that a particular channel is open under steady-state conditions (P_o) was used as an index of channel activity. NP_o values were calculated from all-points amplitude histograms, as described in Dopico et al. (1998a). In multichannel patches of unknown N, knowing NP_o, and the number of openings (X_o) during a period of observation (T), allowed calculation of the mean open time (t_o) from the relationship:

(N)P_o data as a function of voltage were fitted with a Boltzmann function of the type:

where K is the logarithmic potential sensitivity and V_0.5 the potential at which P_o is half-maximal. When the (N)P_o-voltage relationship is fitted by a Boltzmann curve, a plot of lnNP_o (or lnP_o) as a function of voltage is linear at low values of P_o (Dopico et al. 1998a). In this plot, the reciprocal of the slope is the potential needed to produce an e-fold change in (N)P_o, which is routinely used as a measure of the voltage dependence of gating of BK channels (Reinhardt et al. 1989; Toro et al. 1990; Dopico et al. 1996, 1998a). Then, the effective valency (z) was calculated from: 1/slope = RT/zF, where R, T and F have their usual meanings (Toro et al. 1990).

The relationship between [Ca²⁺]_i and the channel steady-state P_o is usually a sigmoid function that can be fitted according to a Hill equation of the type:

where P_o is the steady-state open probability, P_o,max is the maximum open probability, n_H is the slope factor (or Hill coefficient), and the nth root of K gives an estimate of the midpoint of the activation curve and the sensitivity of the channel to [Ca²⁺]_i. The slope factor provides a minimum estimate of the number of Ca²⁺ ions involved in maximally activating the channel (McManus, 1991). Thus, the required [Ca²⁺]_i to achieve the midpoint of activation (P_0.5) at 0 mV, usually designated [Ca²⁺]_0.5, is a widely used parameter to measure the [Ca²⁺] sensitivity of BK channels.

In single channel patches, durations of open and closed times were measured with half-amplitude threshold analysis. A maximum-likelihood minimization routine was used to fit curves to the distribution of open and closed times. Determination of the minimum number of terms for adequate fit was established using a standard F statistic table (significance level, P < 0.01). The slope of the unitary current amplitude (i)-voltage relationship, was used as the unitary conductance (γ). Values for i were obtained from the Gaussian fit of all-points amplitude histograms by measuring the distance between the modes corresponding to the closed state and the first opening level. For all experiments, voltages given correspond to the potential at the intracellular side of the membrane.

Single spike recording in the cell bodies of supraoptic neurones

Single evoked spikes were recorded using the ‘whole-cell’ configuration of the patch clamp technique. This configuration allowed us to control the free [Ca²⁺]_i by using highly buffered Ca²⁺ solutions, which were prepared as described above (see Single channel recordings in the cell bodies of supraoptic neurones). Action potentials were recorded using electrode solutions that contained (mM): 10 NaCl, 130 KCl, 3 MgCl₂, 1.5 EGTA, 2 ATP, 10 Hepes, pH 7.2, with varying amounts of free Ca²⁺ (0.34 or 0.54 μM); or (mM): 10 NaCl, 130 KCl, 3 MgCl₂, 1.5 EGTA, 6 HEDTA, 2 ATP, 10 Hepes, pH 7.2, with free Ca²⁺ ∼1 μM. As indicated above for voltage clamp experiments, a program based upon calculations by Fabiato (1988) was used to calculate nominal free [Ca²⁺] in the solutions. The bath solution contained (mM): 130 NaCl, 5 KCl, 1 CaCl₂, 2 MgCl₂, 10 Hepes, 10 glucose, pH 7.4. The osmolality was routinely measured with an osmometer (μ-Osmette, Precision Systems, Natick, MA, USA), and adjusted to 295 ± 5 mosmol kg⁻¹. Iberiotoxin was diluted from stock to final concentration (10 nM) in bath solution. Electrodes were made as previously described (Widmer et al. 1998), having a tip resistance of 6 MΩ when filled with electrode solution. The whole-cell configuration was obtained by applying a brief pulse of suction to the back of the patch pipette. Action potentials were recorded using an Axopatch 200B amplifier (Axon Instruments, Inc.) and low-pass filtered at a bandwidth of 5 kHz using an eight-pole Bessel filter (model 902LPF, Frequency Devices). Single spikes were evoked as described in Widmer et al. (1998). Briefly, 2.5 ms current pulses were injected, to bring membrane voltage to slightly suprathreshold values and ultimately trigger the spike. For each neurone and condition (in the presence and absence of iberiotoxin, at a given [Ca²⁺]_i), six to eight evoked spikes were selected, which arose from resting potential values within 5 mV of each other. Thus, each neurone served as its own control. The membrane potential and the injected current were acquired on two separated channels, at a sampling rate of 20 kHz per channel. Data acquisition and analysis were performed using pCLAMP software, version 6.0.2 (Axon Instruments, Inc.). The analysis of action potential characteristics has been described in detail elsewhere (Widmer et al. 1998).

Single spike recording in isolated neurosecretory nerve endings

Single evoked spikes were recorded using the ‘perforated patch’ configuration of the patch clamp technique. Glass pipettes were made as previously described (Wang et al. 1992). To achieve the perforated patch configuration, amphotericin-B (240–300 μg ml⁻¹; Rae et al. 1991) was included in the electrode solution, which contained (mM): 130 potassium glutamate, 20 KCl, 2 CaCl₂, Hepes 10, glucose 5, pH 7.2. The bath solution was similar to that described above (see Single spike recording in the cell bodies of supraoptic neurones). Tetrandrine was freshly diluted from stock solution to a final concentration of 0.5 μM in the bath immediately before each experiment. After obtaining the perforated patch configuration, the nerve ending was exposed to a stream of experimental solution flowing from a ‘sewer pipette’ (1 mm diameter; WPI Inc.). In general, isolated nerve endings with an access resistance of 3–5 MΩ were used for electrophysiological recordings. To evoke single spikes, a stimulating current with a duration of 5 ms and an amplitude range of 20–30 pA was applied under current clamp conditions, using a patch clamp amplifier (EPC7, List Electronics). Spikes recorded in the presence and absence of tetrandrine were recorded in the same nerve ending. Data were low-pass filtered at 3 kHz, sampled at 5 kHz, and stored on high density diskettes. Data acquisition and analysis were performed using pCLAMP software, version 5.5.1 (Axon Instruments, Inc.).

Chemicals

Charybdotoxin, iberiotoxin, and apamin were purchased from Alomone Labs (Jerusalem, Israel). Deionized (100 % purity) ethanol was purchased from American Bioanalytical (Natick, MA, USA). Tetrandrine was obtained from Jinhua Pharmaceuticals Inc. (Zheijiang Province, China). All other chemicals were purchased from Sigma Chemical Co. Charybdotoxin, iberiotoxin, and apamin were maintained as stock solutions in water at −80°C. Tetrandrine was dissolved in acidified water (+ HCl, pH 3), and then titrated with NaOH to pH 6 to obtain a 10 mM stock solution.

All experiments were carried out at room temperature (20°C). Data are expressed as means ± standard error of the mean (s.e.m.).

RESULTS

Permeability and selectivity for K⁺

Single-channel recordings, obtained in symmetrical potassium gluconate, from excised patches of rat supraoptic cell bodies revealed a variety of conductances. The subject of this study is the channel with the largest unitary current amplitude observed (Fig. 1A). This channel type was observed in 32 of 68 (47 %) active (i.e. where channel openings were observed) I/O patches, and 31 of 50 (62 %) active O/O patches, yielding an overall frequency of observing channels of 53 % (63 of 118 active patches; each patch excised from a different cell). Patches containing a single functional channel were very infrequent, patches mostly contained either no channels or multiple channels (usually N > 3).

Figure 1. BK channels in the cell bodies of supraoptic magnocellular neurones.

A, single BK channel recordings from an I/O patch of a supraoptic neurone cell body in symmetrical 145 mM potassium gluconate, with a nominal free [Ca²⁺]_i of 560 nM. Outward and inward currents are indicated by upward and downward deflections, respectively. Arrows indicate the baseline. For composition of solutions see Methods. B, unitary current-voltage (i-V) relationships for BK channels in symmetrical 145 mM K⁺ (•) and 145 mM K⁺_∘/45 mM K⁺_i (Na⁺ substituting for K⁺; ▾) from I/O patches. Individual points were obtained from all-points amplitude histograms. Plots have been fitted using first-order regressions. Each data point (n) and bar indicates the mean + s.e.m. of 3–16 determinations; each determination was performed in a different patch, and each patch excised from a different cell. The unitary conductance, obtained from the slope of the plot in symmetrical [K⁺] was 246.1 ± 3.4 pS (n = 12; r = 0.99), with a reversal potential of −1.07 mV. The reversal potential shifted 28.3 mV in a positive direction (r = 0.99) with 45 mM K⁺_i, a value very close to the 29.5 mV predicted by the Nernst equation for a channel highly selective for K⁺ over Na⁺.

Figure 1B shows that the channel under study has a unitary conductance, obtained from the slope of the i-V plot in symmetrical 145 mM potassium gluconate, of 246.1 ± 3.4 pS (n = 12; r = 0.99), with a reversal potential near 0 mV (−1.07 mV), and displays no open channel rectification in the range of potentials tested (± 60 mV). Figure 1B also shows that the reversal potential shifted 28.3 mV in a positive direction upon changing to 100 mM Na⁺/45 mM K⁺ on the intracellular side, a value very close to the theoretical shift of 29.5 mV predicted by the Nernst equation for a channel highly selective for K⁺ over Na⁺. Therefore, the channel under study exhibits the combination of a large unitary conductance and high selectivity for K⁺ over Na⁺, characteristic of BK channels (for reviews see McManus, 1991; Kaczorowski et al. 1996). The single channel conductance of BK channels, recorded under identical conditions, from the isolated secretory nerve endings of these neurones is 213 ± 15 pS (n = 3; r = 0.99). This value is in agreement with that obtained in our previous study of the nerve ending channel, which reported a unitary conductance of 218.3 ± 5.3 pS in symmetric 145 mM potassium gluconate (Dopico et al. 1996), and indicates that unitary conductance values for cell body and terminal BK channels, although close, are significantly different (P < 0.01, Student's unpaired t test).

Voltage and Ca²⁺ activation

A key feature of BK channels is that steady-state activity is increased upon increases in [Ca²⁺]_i and/or membrane depolarization (McManus, 1991). Figure 2A shows that for a fixed [Ca²⁺]_i, the NP_o (or P_o, when N was known)-voltage relationship of the channel under study could be well fitted with a Boltzmann relationship. The reciprocal of the slope in the ln(NP_o)-V relationship (at low P_o) was 10.96 ± 1.39 mV per e-fold change in NP_o (n = 6), which yields a z of 2.31 ± 0.29 (see Methods). These values are within the range of values previously reported for both cloned BK channels encoded by slo genes (Dopico et al. 1998a), and native BK channels (Toro et al. 1990). The P_o-V relationship of the nerve ending BK channel also followed a Boltzmann function: at low P_o, the slope factor was 13.92 ± 2.31 mV per e-fold change in P_o (n = 3), which yields a z of 2.76 ± 0.49. These values are identical to those previously reported for the voltage-sensitive type of BK channel in the isolated neurohypophysial endings (10–15 mV per e-fold change in NP_o, with z = 1.68–2.64; Wang et al. 1992; Dopico et al. 1996). Therefore, the channels under study and BK channels in neurohypophysial secretory endings show an identical voltage dependence of channel gating. Figure 2A also shows that the P_o-V curve is shifted along the voltage axis to more negative potentials as [Ca²⁺]_i is increased, reflected as a linear decrease in channel V_0.5 with increasing [Ca²⁺]_i plotted on a logarithmic scale (Fig. 2B). Thus, the cell body channel is characterized by both voltage and Ca²⁺ dependence of channel gating, which, together with its large unitary conductance and high selectivity for K⁺ over Na⁺, identify this conductance as a BK channel (McManus, 1991; Kaczorowski et al. 1996).

Figure 2. Voltage and Ca²⁺ dependence of gating of BK channels from supraoptic neurone cell bodies, evaluated in I/O patches.

A, plots of P_o as a function of voltage at different [Ca²⁺]_i: 100 (•), 316 (▾), and 560 (▪) nM, are fitted with a Boltzmann relationship. At low (N)P_o values, the limiting slope is ≈11 mV per e-fold change in (N)P_o. The P_o-V curve is shifted along the voltage axis to more negative potentials as [Ca²⁺]_i is increased. B, this parallel shift is reflected in V_0.5 reaching more negative values as [Ca²⁺]_i increases. V_0.5 values, obtained from the Boltzmann fits displayed in A, are now plotted as a function of [Ca²⁺]_i on a logarithmic scale. Data points are fitted by linear regression (r = 0.99).

Fitting channel P_o values obtained at 0 mV as a function of [Ca²⁺]_i to a Hill equation (see Methods), yields a slope factor of 2.38 ± 0.47 (n = 3) and a [Ca²⁺]_0.5 of 470 nM. The Hill coefficient obtained indicates that at least three Ca²⁺ ions are involved in maximal activation of the channel, which is similar to previous data from both neuronal and non-neuronal BK channels (McManus, 1991). However, the [Ca²⁺]_0.5 value of the channel under study is well below values reported for neuronal BK channels (for a review, see McManus, 1991). The [Ca²⁺]_0.5 value for the BK channel in the nerve endings of these neurones was 1.53 μM, calculated from P_o values obtained at 40 mV (Wang et al. 1992). When we fitted the P_o-[Ca²⁺]_i relationship at this potential, we obtained a [Ca²⁺]_0.5 of 273 nM, which is markedly smaller than the value from the neurohypophysial channel, indicating a significantly higher Ca²⁺ sensitivity for the cell body BK channel. We confirmed the differential Ca²⁺ sensitivity of cell body and terminal BK channels by recording channel activity in the nerve terminals at 0.27, 1.53 and 10 μM free Ca²⁺₁. At any of these three different free [Ca²⁺]_i, the P_o-V relationships could be well fitted with a Boltzmann relationship, with P_o/P_o,max values of 0.004 (n = 2), 0.090 ± 0.037 (n = 4), and 0.642 (n = 2), respectively. In contrast, for the cell body BK channel, we found that at 1 μM Ca²⁺₁, the P_o-V relationship could no longer be well fitted with a Boltzmann function, but followed a bell-shaped curve: P_o values increased from 0.04 at −60 mV to 0.56 at 30 mV (where P_o/P_o,max = 1), and decreased at more positive potentials (50 mV, 0.46; 60 mV, 0.43; 70 mV, 0.39). The decrease in steady-state P_o at potentials more positive than 40 mV and [Ca²⁺]_i higher than 1 μM was related to the higher frequency of appearance of long-closed events, possibly reflecting the channel sojourning into an inactivated state(s), as reported for BK channels from cultured rat muscle at [Ca²⁺]_i ≥ 10 μM (Rothberg et al. 1996)

Kinetics

The mean open time of the supraoptic cell body BK channel, calculated from multichannel patches as described in Methods, is 11.37 ± 1.06 ms (n = 5), recorded at +40 mV and 316 nM free Ca²⁺₁. This brief t_o was confirmed by studying the channel dwell time distribution. The distribution of open times could be well-fitted with a single component (Fig. 3A), which is consistent with results obtained from both the neurohypophysial BK channel (Wang et al. 1992) and BK channel types I and II characterized in planar lipid bilayers (Reinhardt et al. 1989). The open time constant (τ_o) for the channel under study was 8.53 ms, consistent with the t_o values obtained from multichannel patches. The distribution of closed times showed a more complex pattern, best fitted by two exponentials, with time constants of 1.57 (τ_c,fast) and 12.67 ms (τ_c,slow) (Fig. 3B). These data, in particular the brief channel open times, indicate that the BK channel in supraoptic cell bodies displays fast gating kinetics, which is characteristic of ‘type I’ BK channels (Reinhardt et al. 1989). At fixed [Ca²⁺]_i and voltage, we never observed a modal ‘switch’ of the cell body channel from the fast gating type to the slow gating class (type II; Reinhardt et al. 1989), characteristic of BK channels in the neurohypophysial secretory endings (Wang et al. 1992). Since type I BK channels are characterized by both short dwell (in particular, open) times, and block by nanomolar concentrations of charybdotoxin, a pharmacological analysis was used to confirm the type I identity of the supraoptic cell body BK channel.

Figure 3. Dwell time distribution of BK channels from supraoptic neurones obtained from a single channel patch.

The voltage was set to +40 mV and the free [Ca²⁺]_i was 316 nM. A, the open times distribution could be well fitted with a single component function, with τ_O = 8.53 ms. B, the closed time distribution was well fitted with a function of two components, with τ_c,fast = 1.57 ms (88.7 %), and τ_c,slow = 12.67 ms (11.3 %). These values indicate that BK channels in the cell bodies of supraoptic neurones belong to the ‘fast kinetics’ subtype (type I) of BK channels (Reinhardt et al. 1989). Durations of open or closed times were measured with half-amplitude threshold analysis. Curves were fitted using a maximum-likelihood minimization routine. The dotted lines represent the individual fitted components, while the continuous line is the composite fit. A statistical comparison of different fitting models was done using a standard F table (P < 0.01). The duration of each particular component is given in milliseconds, and the relative contribution of each particular component to the total fit is given as a percentage in parentheses.

Responses to Ca²⁺-activated K⁺ channel blockers

Figure 4 summarizes the response of the channel to a variety of Ca²⁺-activated K⁺ channel blockers. Apamin characteristically blocks small conductance, Ca²⁺-activated K⁺ channels at low nanomolar concentrations (Garcia et al. 1991). In rat supraoptic neurones, this peptide has been reported to block both a Ca²⁺-activated K⁺ current and the slow Ca²⁺-dependent after-hyperpolarization that follows trains of action potentials evoked by depolarizing current pulses (Kirkpatrick & Bourque, 1996). Figure 4A shows that the steady-state activity of the channel under study was unmodified by 10 nM apamin. In contrast, the activity of a small conductance, Ca²⁺-activated K⁺ channel (∼85.9 pS in symmetric 145 mM potassium gluconate), recorded in the same patches in which apamin was ineffective on BK channel activity, was almost totally abolished by this concentration of the peptide (n = 2). These data suggest that the cell body BK channel does not play a major role in the slow after-hyperpolarization that follows trains of action potentials in these neurones.

Figure 4. Effect of different blockers on BK channel unitary currents in supraoptic cell bodies.

A, effects on channel NP_o following acute exposure (< 1 min) of O/O patches to each blocker (Ap, apamin; Tt, tetrandrine; ChTX, charybdotoxin; IbTX, iberiotoxin). Values are shown as percentage of control. B, single channel recordings obtained from different O/O patches, each exposed to a different blocker, in symmetrical 145 mM K⁺; free [Ca²⁺]_i, 0.32 μM. Arrows indicate the baseline. Whereas ChTX characteristically decreases BK channel NP_o by introducing long silent periods (top traces; V, −20 mV), tetrandrine at maximal concentrations for BK channel blockade reduces NP_o by introducing a ‘flickery’ block (bottom traces; V, 40 mV). This results in a characteristic decrease in unitary current amplitude (8.38 ± 0.45 vs. 7.20 ± 0.59 pA, n = 3; P < 0.01) due to poor resolution (bandwidth, 1 kHz) of individual open-to-blocked transitions (τ_O = 0.38 ms, in the presence of 3 μM tetrandrine; Wang & Lemos, 1992).

Most BK channels described to date are blocked by nanomolar concentrations of charybdotoxin (Garcia et al. 1991). Exceptions include the type II BK channel from rat brain vesicles reconstituted into lipid bilayers (Reinhardt et al. 1989), and BK channels from the neurosecretory nerve endings of supraoptic neurones (Wang et al. 1992). Figure 4A shows that supraoptic cell body BK channels were blocked by 100 nM charybdotoxin. The block by charybdotoxin was evident at concentrations as low as 10 nM (NP_o = 37 % of control; n = 2), whereas BK channels in the nerve endings of these neurones are insensitive to charybdotoxin at concentrations as high as 360 nM (Wang et al. 1992). Iberiotoxin, which is about tenfold more potent and also more selective on BK channels than charybdotoxin (García et al. 1991), also decreased the NP_o of the cell body channel (Fig. 4A). Both charybdotoxin and iberiotoxin decreased BK channel activity in the cell body by introducing long quiescent interburst periods without modifying intraburst kinetics, a characteristic feature of these peptide blockers (Fig. 4B).

BK channels from neurohypophysial nerve endings are typically blocked by the alkaloid tetrandrine (Wang & Lemos, 1992). Figure 4A shows that BK channels in the cell bodies of supraoptic neurones are also blocked by tetrandrine. Tetrandrine (10 μM) decreased channel activity by introducing a ‘flickery’ block (Fig. 4B), characteristic of BK channel inhibition by maximal concentrations of this alkaloid (τ_o = 0.38 ms, in the presence of 3 μM tetrandrine; Wang & Lemos, 1992). This tetrandrine-induced channel ‘flickering’ resulted, due to the limits of temporal resolution of our recording system, in a decrease in apparent unitary current amplitude (8.38 ± 0.45 vs. 7.20 ± 0.59 pA at 40 mV; P < 0.01).

Differential role of cell body and nerve ending BK channels in single spike repolarization

BK channel activity has been shown to control neuronal excitability by contributing to action potential repolarization (Crest & Gola, 1993). In light of the differential response of BK channels from cell bodies and nerve endings to blockers, we explored the effect of these blockers on evoked single spikes in cell bodies and nerve endings, to determine the contribution of each channel subtype to action potential repolarization within its particular neuronal domain. Exposure of isolated secretory nerve endings to 0.5 μM tetrandrine slowed the repolarization phase of evoked single spikes, recorded in the perforated-patch configuration (Fig. 5A). The concentration of tetrandrine chosen is close to the IC₅₀ for the inhibitory action of the drug on BK single channel activity in these nerve endings (0.21 μM; Wang & Lemos, 1992). The widening of the action potential by 0.5 μM tetrandrine is best explained by a drug-induced inhibition of the neurohypophysial BK channel, since tetrandrine at these concentrations has been reported not to affect other conductances present in these nerve endings (Wang & Lemos, 1992).

Figure 5. Differential effect of selective BK channel blockers on evoked single spike repolarization in nerve endings vs. cell bodies of supraoptic neurones.

A, exposure to 0.5 μM tetrandrine, which selectively blocks BK channel steady-state activity (Wang & Lemos, 1992), slows the repolarization of evoked single spikes recorded from isolated neurohypophysial endings in the perforated-patch configuration. To evoke single spikes, a stimulating current with a duration of 5 ms and amplitude ranging from 20–30 pA was applied under current clamp conditions; resting potential, −80 mV. Continuous line, control; dotted line, 0.5 μM tetrandrine. Both spikes were recorded in the same nerve ending. B, exposure to 10 nM iberiotoxin, which markedly decreased cell body BK channel steady-state activity (Fig. 4A) fails to shape evoked single spikes recorded from supraoptic cell bodies in the whole-cell configuration. The nominal free [Ca²⁺]_i was set to 540 nM. Spikes were elicited by 2.5 ms current pulses under current clamp conditions, which evoked the depolarization preceding the notch in the rising phase of the spike; resting potential, −70 mV. Continuous line, control; dotted line, 10 nM iberiotoxin. Both spikes were recorded in the same neurone.

In contrast, exposure of supraoptic cell bodies to an effective BK channel blocker failed to slow the repolarization phase of evoked single spikes (Fig. 5B). In this case, we used 10 nM iberiotoxin, which totally abolished BK single channel activity in supraoptic cell bodies (Fig. 4A). The whole-cell configuration was used in these experiments to allow control of the steady-state activity of BK channels, by exposing the cytosolic side of the membrane, where Ca²⁺ sensors seem to be located (McManus, 1991), to highly buffered free Ca²⁺ solutions. We selected a free [Ca²⁺]_i (i.e. 540 nM) at which most BK channels are active within the voltage range spanned by the action potential (Fig. 2A). Iberiotoxin was without effect not only on the repolarization phase of the action potential (half-width, 1.52 ± 0.22 and 1.51 ± 0.24 ms, in the presence and absence of the blocker, respectively; n = 3), but also on all other measured characteristics of evoked single spikes in the cell bodies (threshold for activation, amplitude, and rise time), in all three cells examined (data not shown). We also evaluated iberiotoxin action on spike repolarization using both a higher (1 μM) and a lower (340 nM) free [Ca²⁺]_i. As shown in Fig. 2, these concentrations markedly modify the steady-state activity-voltage relationship of the cell body channel. As found with 540 nM Ca²⁺₁, 10 nM iberiotoxin repeatedly failed to modify the threshold for activation, amplitude, rise time, or half-width of evoked single spikes in the cell bodies, at either 340 nM (3 of 3 neurones) or 1 μM (2 of 2 neurones) Ca²⁺₁. The differential action of selective BK channel blockers on action potential repolarization in the nerve ending versus the cell body cannot be explained by the disruption of the intracellular milieu and consequent loss of some critical modulatory factor in the whole-cell configuration, since identical results were obtained when 10 nM iberiotoxin was applied to cell bodies, and spikes were recorded in the absence and presence of the blocker in the perforated-patch configuration (Widmer et al. 1998). In summary, these results indicate that regionally segregated supraoptic BK channels with different biophysical and pharmacological properties, differ in their apparent contribution to rapid spike repolarization.

Response to ethanol

We have previously demonstrated that ethanol, at circulating concentrations found after moderate drinking, reversibly increases BK channel activity in excised, I/O or O/O membrane patches from rat neurohypophysial endings (Dopico et al. 1996), possibly contributing to the decreased vasopressin release and plasma levels following alcohol ingestion. In contrast, acute application of ethanol to the cytosolic surface of I/O patches from the cell bodies of supraoptic neurones repeatedly failed to increase BK channel NP_o, under conditions similar to those previously used in isolated neurohypophysial endings (Fig. 6). This insensitivity to ethanol was observed at concentrations as high as 100 mM, at which the drug reaches its maximal effect on the activity of the BK channel in neurohypophysial endings (EC₅₀, 22 mM; Dopico et al. 1996). The characteristics of supraoptic cell body and nerve ending BK channel subtypes are summarized in Table 1.

Figure 6. Concentration-response curves for the effect of ethanol on BK channel activity in the cell bodies (•) vs. the nerve endings (^) of magnocellular neurones.

Results are expressed as the ratio of NP_o values obtained in the presence and absence of ethanol, determined in the same patch. Each point of the graph is the mean ± s.e.m. of 4–20 determinations; each determination was obtained in a different patch. In all cases, the patch membrane potential was set to 20–40 mV. Values from BK channels in the terminals were obtained from Dopico et al. 1996. The maximal effect of ethanol on these channels was obtained at 50–100 mM, with an EC₅₀ of 22 mM.

Table 1.

Comparison of the properties of BK isochannels from cell bodies versus nerve endings of rat supraoptic neurones

	Cell bodies	Nerve endings
Unitary conductance (pS)	246.1 ± 3.4	213 ± 15 (P < 0.01)
Gating
Voltage dependence	2.31 ± 0.29	2.76 ± 0.49
Ca2+ dependence (μm)	0.272	>1.53
Kinetic type	Fast (I)	Slow (II)
Pharmacology
Blockade by ChTX (nm)	Yes (observed at 10)	No (up to 360)
Blockade by TD (μm)	Yes (evaluated at 10)	Yes (0.1–10)
EtOH activation (mm)	No (up to 100)	Yes (EC50, 22; Emax, 100*)
Role in AP repolarization	No	Yes

Unitary conductance was measured in symmetric 145 mm K⁺. Voltage dependence is given by z obtained from the slope of lnNP_o vs. V at low P_o. Calcium dependence is given by [Ca²⁺]_0.5 at 40 mV. Each value is the mean ± s.e.m. of 3–20 determinations; each determination was obtained in a different cell body/nerve ending. AP, action potential; ChTX, charybdotoxin; EtOH, ethanol; TD, tetrandrine.

^*From Dopico et al. 1996; BK channel characterization at the single channel and macroscopic levels in neurohypophysial nerve endings is described in detail in Wang et al. 1992.

DISCUSSION

Cell body and nerve ending channels are distinct functional types

We have identified, according to both electrophysiological and pharmacological criteria, BK channels in rat supraoptic neuronal cell bodies, consistent with previous findings of slo expression, measured by [³H]iberiotoxin-D19 binding, in the hypothalamus (Knaus et al. 1996). Thus, the differential actions of both Ca²⁺ entry blockade and BK channel blockers on cell bodies versus secretory endings (Wang & Lemos, 1995; Widmer et al. 1998), result not from the absence of BK channels in the cell bodies, but most likely from the different properties of BK channel subtypes in cell bodies vs. nerve endings. Although BK channels in neurohypophysial endings and cell bodies are similar with respect to K⁺/Na⁺ selectivity, gating charge, and blockade by tetrandrine, they differ markedly in Ca²⁺- and ethanol sensitivity, blockade by charybdotoxin, and kinetics (Wang et al. 1992; Wang & Lemos, 1992; Dopico et al. 1996). In addition, the unitary conductance of cell body BK channels is slightly larger than that of neurohypophysial ending channels. In spite of the well-known ‘wanderlust’ kinetics of BK channels (Silberberg et al. 1996), we never observed a gating ‘switch’ from type I (characteristic of cell bodies) to type II (characteristic of neurohypophysial endings) kinetics. Together, these data indicate that BK channels in cell bodies and nerve endings constitute different functional types or ‘isochannels’. We did not observe type II isochannels in any of the 118 supraoptic cell bodies in which BK channel activity was recorded. Whether type I BK isochannels are differentially distributed between subsets of supraoptic cell bodies (i.e. vasopressinergic versus oxytocinergic), remains to be established.

The existence of isochannels has been previously reported following reconstitution of BK channels from rat brain (Reinhart et al. 1989) or bovine mesenteric vascular smooth muscle (Sansom & Stockand, 1994) into lipid bilayers. The phenotypes observed differed primarily in gating or Ca²⁺ sensitivity. In addition, Toro et al. (1991) showed that pig coronary artery BK isochannels reconstituted into lipid bilayers differed not only in Ca²⁺ sensitivity, but also in unitary conductance, as found in the present study. Electrophysiological studies of BK channel isoforms in situ (Solaro et al. 1995) or other channel isoforms regionally segregated in intact neurones (Fisher & Bourque, 1996) are very scarce. The distinct electrophysiological properties of BK isochannels from supraoptic neuronal cell bodies versus nerve endings demonstrated here might reflect an adaptation to the regional neuronal microenvironment of each isochannel, as discussed below.

Ca²⁺ sensitivity

One of the most striking differences between cell body and nerve ending BK channels is their Ca²⁺ sensitivity, which may correspond to regional differences in [Ca²⁺]_i. Single channel recordings show that BK currents in the neurohypophysial endings have a Ca²⁺ threshold greater than 1 μM, with an EC₅₀ for Ca²⁺ of 1.53 μM (Wang et al. 1992). Fura-2 measurements in these nerve endings show a bulk [Ca²⁺]_i of 0.05 μM during resting conditions (Stuenkel, 1990), > 1 μM during electrical stimulation (Brethes et al. 1987), and > 1.5 μM during depolarization elicited under voltage-clamp conditions (Lindau et al. 1992). Furthermore, studies on squid presynaptic endings at giant synapses demonstrate that during neurotransmitter release, Ca²⁺ in the vicinity of the membrane is much higher than the bulk [Ca²⁺]_i measured by fura-2 (Adler et al. 1991). In addition, BK channels from Helix neurones are reported to be concentrated at presynaptic nerve terminals, colocalized with Ca²⁺ channels in the membrane in ‘Ca²⁺ domains’, where [Ca²⁺]_i may reach as high as 100 μM (Gola & Crest, 1993). A similar clustering of Ca²⁺ and BK channels at presynaptic active zones was found in hair cells (Issa & Hudspeth, 1994). In permeabilized neurohypophysial endings, half-maximal peptide release occurs at ∼1.7 μM Ca²⁺₁ (Cazalis et al. 1987). The micromolar concentrations of Ca²⁺₁ required for activation of neurohypophysial ending BK channels at depolarized potentials, reported both previously (Wang et al. 1992) and in the present work, fit perfectly within the range of [Ca²⁺]_i values in the studies noted above. Thus, the Ca²⁺ sensitivity of the neurohypophysial ending channel is well suited to the high [Ca²⁺]_i present in the secretory endings during stimulation and neuropeptide release. The activation of these BK isochannels by a rise in [Ca²⁺]_i to micromolar levels would result in K⁺ efflux and, consequently, faster repolarization, which would limit Ca²⁺ influx through voltage-gated Ca²⁺ channels, and thus, neurosecretion. In summary, the Ca²⁺ sensitivity of the neurohypophysial ending channel is consistent with a putative role for BK channels in providing negative feedback during depolarization-secretion coupling in these nerve endings.

In contrast, BK channels in the cell bodies have a Ca²⁺ threshold well-below 1 μM, with a [Ca²⁺]_0.5 of ∼0.273 μM at +40 mV. Fura-2 measurements show a bulk [Ca²⁺]_i of 0.04–0.05 μM during resting conditions, and 0.05–0.1 μM during spontaneous oscillations in these cell bodies (Lambert et al. 1994). Since ‘hot spots’, or Ca²⁺ domains have been reported only near release sites, it can be postulated that the differential Ca²⁺ sensitivity of BK channels from cell bodies and nerve endings in the same neurone is appropriate for the [Ca²⁺]_i in the cellular compartment and microenvironment of each isochannel. Although there are numerous electrophysiological studies on the role of Ca²⁺ in the shaping of electrical activity in supraoptic neurones (for a review, see Bourque et al. 1998), simultaneous measurements of [Ca²⁺]_i and electrical recording are rare. Our data indicate that individual spike shaping involves BK channels in the nerve ending, but not in the cell body. A delayed recovery of evoked single spikes in neurohypophysial endings by exposure to a BK channel blocker is consistent with previous results obtained in axonal nerve endings at the lizard neuromuscular junction, which showed that blockade of BK channels broadens action potentials (Morita & Barrett, 1990). In the cell body, the role of BK channels may involve the patterning of spikes, difficult to explore in the dissociated cell preparation used. In particular, given the high Ca²⁺ sensitivity of this BK isochannel, its activity may be involved in the regulation of membrane potential and action potential threshold. Interestingly, it has recently been shown in a slice preparation that up to 200 nM iberiotoxin fails to block the slow after-hyperpolarization, the fast after-hyperpolarization, and the depolarizing after-potential that may follow train of spikes in supraoptic neurones (Greffrath et al. 1998). Our single channel data clearly demonstrate that lower concentrations of this peptide are sufficient to almost totally suppress cell body BK channel activity in supraoptic neurones (Fig. 4). Thus, the failure of iberiotoxin to block either the repolarization phase of the single spike or the after-potential components cannot be attributed to the existence of a BK phenotype insensitive to the toxin, but rather to the lack of major involvement of this isochannel in the generation of these waveforms. Since the slow after-hyperpolarization and the depolarizing after-potentials were blocked by charybdotoxin, but not by iberiotoxin, Greffrath et al. (1998) postulate the contribution of intermediate conductance, Ca²⁺-activated K⁺ channel activity to these waveform components.

Differential ethanol sensitivity of BK isochannels

An increase in BK channel activity upon acute exposure to relevant concentrations of ethanol has been documented in a variety of systems, including isolated neurohypophysial terminals (Dopico et al. 1996), GH3 pituitary clonal cells (Jakab et al. 1997), cloned BK channels expressed in Xenopus oocytes (Dopico et al. 1998b), and skeletal muscle BK channels reconstituted into planar lipid bilayers (Chu et al. 1998). In addition to the difference in Ca²⁺ sensitivity of supraoptic cell body and nerve terminal BK isochannels, their difference in ethanol sensitivity was striking. Although differential actions of ethanol have been reported for voltage-gated calcium channels as a function of brain region (Leslie et al. 1983), for potassium channels among different identified neurones in invertebrates (Treistman & Wilson, 1987), for GABA_A receptors in hippocampal CA1 cells, dependent upon the presynaptic stimulus site (Weiner et al. 1997), and for PC12 cell voltage-gated calcium channels, according to whether cells are differentiated or not (Mullikin-Kilpatrick & Treistman, 1995), the present finding is the first report of such a sharply defined difference in the ethanol sensitivity of a channel class between specialized regions of neurones. Thus, this result forces us to consider that rather than describing a channel class within a neuronal subtype as sensitive or insensitive to ethanol, it may be necessary to describe the sensitivity of particular isochannel of that class, considering, in addition, the cell domain in which they reside.

The differences between BK isochannels are probably due to structural diversity in the channel complex itself and/or its immediate proteolipid environment

Since BK channel activity in both nerve endings and cell bodies was recorded in cell-free patches, > 10 min after patch excision, in the absence of nucleotides or other regenerating systems, it is very unlikely that the regional differences in the biophysical and pharmacological properties of these BK isochannels can be explained by differential modulatory action(s) of freely diffusible second messengers or BK-associated kinases or G-proteins. Rather, these differences are probably attributable to structural diversity in the channel complex itself and/or its immediate proteolipid environment. Native BK channels from mammalian tissues consist of two structurally distinct subunits, termed α and β. The β subunit is a modulatory subunit that, when coexpressed with the pore-forming α subunit in Xenopus oocytes, markedly increases the Ca²⁺ sensitivity of the channel, evidenced by a parallel shift to the left of the NP_o-V relationship (Kaczorowski et al. 1996). Although differential efficacy of coupling between α and β subunits, and/or different levels of expression of the β subunit might explain the shift in the P_o-V relationship between the cell body and the nerve ending channels, other data suggest that at least part of the difference between these isochannels results from structural differences in the α subunit itself: (1) the channels differ in unitary conductance, consistent with structural differences in the pore of the channel, located in the α subunit (Wei et al. 1994); (2) the response of these isochannels to acute application of 25–100 mM ethanol is drastically different, and this drug most probably modulates BK channel function by a direct interaction with the α subunit (Dopico et al. 1998b). In addition, the presence of different α subunits might contribute to the different t_o of these isochannels, since this electrophysiological variable is partially determined by the ‘core’ domain of the α subunit (Wei, et al. 1994). The marked differences in Ca²⁺ sensitivity between these isochannels, could also reflect differences between α subunits, since Ca²⁺-sensing sites have been located in the ‘tail’ domain of this subunit (Wei et al. 1994; Schreiber & Salkoff, 1997). However, the molecular underpinnings of the regional heterogeneity of BK channel function remain to be determined, and differences in either post-translational processing, and/or lipid microenvironment of the channel protein remain possible explanations.

Targeting of ion channels in neurones

The appropriate targeting of ion channels to specific cellular localizations in the neurone is essential for efficient electrical signalling. During the last decade, many light and electron microscopic immunocytochemical studies have demonstrated that differential localization of both ligand- and voltage-gated ion channels to specific compartments of the cell is a widespread phenomenon in central neurones (for review see Sheng & Wyszynski, 1997). Outside the CNS, it was recently found that several BK channel isoforms, generated by alternative splicing of a single gene (cSlo), are differentially distributed along avian cochlea (Navaratnam et al. 1997; Rosenblatt et al. 1997). Heterologously expressed chicken cSlo clones showed differences in both voltage and Ca²⁺ sensitivity, which, together with their differential spatial distribution, may help to determine the avian tonotopic map (Rosenblatt et al. 1997). Here, we have demonstrated that BK channels in supraoptic cell bodies display different electrophysiological and pharmacological properties from their counterparts in neurohypophysial endings. Their distinct electrophysiological properties, in particular their differential Ca²⁺ sensitivity, probably reflect the characteristics of the regional microenvironment in which each BK isochannel subserves a particular neuronal function. Their differential response to ethanol forces a re-examination of the concept of a ‘cellular’ target of ethanol, preventing the extrapolation of results obtained in one particular neuronal microdomain to another.