Overlapping selectivity of neurotoxin and dihydropyridine calcium channel blockers in cerebellar granule neurones

Neuropharmacology 39 (2000) 1740–1755 www.elsevier.com/locate/neuropharm

Overlapping selectivity of neurotoxin and dihydropyridine calcium channel blockers in cerebellar granule neurones J. Russell Burley *, Annette C. Dolphin Department of Pharmacology, Medawar Building, University College London, Gower Street, London WC1E 6BT, UK Accepted 1 December 1999

Abstract Calcium (Ca2+) currents have been studied extensively in cerebellar granule neurones, but much of the whole-cell pharmacology is inconsistent. Ca2+ channel currents were recorded from granule neurones to investigate whether the commonly used Ca2+ channel blockers show overlapping selectivity. Using combinations of toxin channel blockers, 45% of the total current was shown to be carried by Ca2+ channels susceptible to block by the combined, or cumulative application of, ω-agatoxin IVA, ω-conotoxin GVIA and ω-conotoxin MVIIC, thus representing P/Q- and N-type channel currents. However, sequential application of these toxins showed that substantial overlap occurred in the proportions of current sensitive to individual toxins. Application of the 1,4-dihydropyridine nicardipine at 1 µM, a concentration reported to be selective for L-type channels, blocked 16% of the total current, without reducing the current sensitive to the toxins used. However, greater concentrations of nicardipine (⬎10 µM) blocked a proportion of the total current that could not be accounted for by L-type channels alone. These results demonstrate that a pharmacological approach based on the L, N, P/Q, and R classification does not adequately describe the Ca2+ channel subtypes found in cerebellar granule neurones due to substantial cross-selectivity to the drugs and toxins used.  2000 Elsevier Science Ltd. All rights reserved. Keywords: Calcium channel; Rat cerebellar granule neurone; Nicardipine; ω-Agatoxin IVA; ω-Conotoxin GVIA; ω-Conotoxin MVIIC

1. Introduction In initial pharmacological attempts to characterise Ca2+ channel currents in cerebellar granule neurones De Waard et al. (1991) showed that ω-conotoxin GVIA (ωCTx GVIA) blocked 63% of the mean current. Nicardipine in this study reduced currents by 78% leading to the conclusion that more than half of the channels in cerebellar granule cells had a mixed pharmacology. Pearson et al. (1993) also demonstrated currents that were sensitive to both dihydropyridine (DHP) and ω-CTx GVIA, although accounting for less of the total current. These more recent studies also used ω-agatoxin IVA (ω-Aga IVA) to demonstrate further diversity and showed that 40% of the current was blocked by ω-Aga IVA but this

* Corresponding author. Tel.: +44-020-7419-3285; fax: +44-0207813-2808. E-mail address: [email protected] (J.R. Burley).

overlapped completely with current sensitive to ω-CTx GVIA and the DHP antagonist (⫺)-202-791. Interestingly, Randall and Tsien (1995) showed that 46% of the whole-cell current was sensitive to ω-Aga IVA but identified two distinct components, “P” constituting 11% of the mean total current and 50% blocked by 1–3 nM and “Q” making up 35% and 50% blocked by 90 nM both showing no overlap with nimodipine and ω-CTx GVIA, but both being completely blocked by ω-conotoxin MVIIC (ω-CTx MVIIC). In both these studies a component of current, being 20–30% of the total, was resistant to block by organic compounds but sensitive to Cd 2+ and Ni2+, designated “R”-type current. A greater diversity than initially proposed has also been shown by studies of single Ca2+ channels in cerebellar granule neurones. Pietrobon’s group have shown two distinct DHP-sensitive L-type channels with different inactivation properties (Forti and Pietrobon, 1993). In addition, P- and R-type channels have been shown to be functionally diverse. Three distinct channels were

0028-3908/00/$ - see front matter  2000 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 0 8 ( 9 9 ) 0 0 2 6 6 - X

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

identified, each having a characteristic conductance and activation threshold, but similar inactivation properties (Forti et al., 1994). Later studies (Tottene et al., 1996) showed that one of these channels could be classified as a novel P-type channel, being insensitive to ω-CTx GVIA and nimodipine, but blocked by ω-Aga IVA and ω-CTx MVIIC. The two remaining channels were both insensitive to all Ca2+ channel blockers and displayed voltage-dependent properties similar to low-voltage activated (LVA) channels. Similarities were apparent to the R-type current described by Randall and Tsien (1995) in that Ni2+ could block the residual current in wholecell recordings with a biphasic concentration dependence. Taken together, data from whole-cell pharmacology and single channels show that in many cases the Ca2+ channels in cerebellar granule neurones fall outside the established classification. Here we show that Ca2+ current components in granule neurones exhibit overlapping sensitivity to the neurotoxin channel blockers currently in use, thus highlighting the need for a cautious approach to the pharmacological definition of Ca2+ channels. The initial pharmacological classification of currents into L, N and P, and the later addition of Q- and R-types, represents an oversimplification.

2. Methods 2.1. Cell culture Cerebellar granule neurones were dispersed from the cerebella of P6 Sprague-Dawley rats. Dissociated cells (5×105) were plated onto 13-mm-diameter glass coverslips coated in poly-l-lysine (15 µg/ml), in modified minimum essential medium containing 10% foetal calf serum, 2.5% chick embryo extract, 39 mM glucose, 2 mM glutamine, 23 mM KCl, penicillin (50 IU/ml) and streptomycin (50 µg/ml). After 48 h, 5-fluoro-2⬘-deoxyuridine (80 µM) was added to the medium to reduce proliferation of non-neuronal cells, and after 72 h, horse serum replaced foetal calf serum. Culture medium was renewed every 5 days. Media and serum additives were obtained from Gibco BRL. Whole-cell patch-clamp recordings were carried out on cells at 7–14 days in vitro (DIV). Pharmacological data were pooled from recordings made during this time-window. In some cases data were pooled from serial cultures.

1741

mOsM (sucrose). During recording, cells were bathed in a solution containing (mM): TEA-acetate 70, NMDG 70, KOH 3, Mg-acetate 0.6, glucose 4, barium or calcium acetate 10, HEPES 10, TTX 0.001, pH 7.4 (acetic acid), 320 mOsM (sucrose). 2.3. Electrophysiological measurements All experiments were carried out at room temperature (20–22°C) using an Axopatch 1D patch-clamp amplifier (Axon Instruments) interfaced via a TL-1 to a personal computer. Data were captured online at a digitization rate of 5 kHz following electronic compensation of capacity transients and subtraction of linear leak using a P/6 subtraction protocol (in which six hyperpolarising pulses of one-sixth the amplitude of the test pulse were summed and added to the test pulse). Data acquisition and analysis were carried out using pCLAMP6 software (Axon Instruments). To minimise space-clamp problems, isolated cells with short processes were selected. Cells with a series resistance greater than 25 M⍀ were discarded and series resistance was generally compensated to 80%. Typically, voltage-error for uncompensated series resistance was no greater than 1 mV. The junction potential measured for these solutions was ⫺12 mV and all figures show corrected values. To prevent voltage offset due to changing activity of Cl⫺ encountered when a composite Ag/AgCl electrode is immersed directly in the bath, the reference electrode consisted of a short agar bridge bathed in an isolated well of 1 M KCl into which an Ag/AgCl pellet was fixed. 2.4. Statistical analysis Data are shown as the mean±s.e.m of n observations and Student’s t-test was used to examine statistical significance. 2.5. Drug supply and use All fine chemicals were obtained from Sigma. ω-Aga IVA (Bachem Bioscience Inc.), ω-CTx GVIA (Peninsula Laboratories) and ω-CTx MVIIC (Peptide Institute) were stored as 100 µM stock solutions in water. Nicardipine (Sigma) was stored in the dark as a 10 mM stock in ethanol; experiments involving nicardipine were carried out in dim light. All drugs were applied by pressure ejection from a 5 µm perfusion pipette placed within 50 µm of the cell being studied.

2.2. Recording solutions

3. Results

Patch pipettes were fabricated from borosilicate glass capillaries (Plowden and Thompson) with a resistance of 5–8 M⍀ and filled with a solution containing (in mM): 100 HEPES, 30 EGTA, 0.57 CaCl2, 2.25 MgCl2, 3.68 ATP(K2), 0.1 GTP(Na2), pH 7.2 (CsOH), 320

3.1. General properties of Ca2+ channel currents in cerebellar granule neurones Ca2+ channel currents were not present in granule cells on the day of dissociation (Fig. 1A). Currents appeared

1742

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 1. (A) Development of Ca2+ channel currents in cerebellar granule neurones over the first 14 DIV. The open circles show changes in Cm and the filled circles measure the peak inward current at different developmental stages. Each point represents the mean±s.e.m. for a group of four cells. (B) Current–voltage relationship for a group of 310 cells. The data were fit with a Boltzmann equation of the form: I⫽

g(V−Vrev) 1+exp[(V−V50(act))/k]

where g is the conductance in nS, Vrev is the null potential in mV, V50(act) is the voltage for half-activation (mV) and k is the slope factor (mV). The global averages of individual fit values for this group were: g, 4.6±0.06; Vrev, +45±0.2; V50(act), ⫺14±0.3; k, 8±0.2. Error bars are smaller than the symbols. (C) Family of Ba2+ currents recorded from a holding potential of ⫺92 mV and stepping to ⫺12, +8, +28 and +48 mV for 100 ms. (D) Comparison of current inactivation recorded in 10 mM Ca2+ and 10 mM Ba2+. Currents from 10 cells were normalised and averaged. Error bars are shown at several points. (E) Example recording of a longer duration (10 s) Ba2+ current showing inactivation, the current was fit with a second order exponential (see text). (F) Complete block of current by 1 mM Cd 2+. Traces shown were recorded before and during application of Cd 2+.

and began to increase in amplitude over the first 7 DIV, reaching a stable plateau in their second week of culture. Over the initial week a small, coincident increase in whole-cell capacitance was evident. The mean Ba2+ current–voltage relationship for all cells is shown in Fig. 1B. The current was entirely highvoltage activated (HVA), significant inward current not being activated until the voltage-step reached about ⫺40 mV. Maximal inward current was evoked at ⫺2 mV and

there was no evidence of more than one distinguishable peak in either the global average or individual current records. Activation was rapid and became faster as the depolarising step was increased (Fig. 1C). Inactivation of peak Ba2+ currents was not evident over 100 ms steps, although recordings made in the presence of Ca2+ inactivated significantly over this duration. Fig. 1D compares the kinetics of normalised currents recorded in either 10 mM Ba2+ or 10 mM Ca2+. Generally, Ca2+ currents were

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

smaller in amplitude, activated more slowly, and inactivated with a faster time-course, and had tail currents that deactivated more slowly. For currents evoked at ⫺2 mV carried by Ca2+, decay of the current could be described by a single-order exponential with τ(Ca)=57±9 ms. Although there was little or no inactivation of Ba2+ currents at this test potential over 100 ms, 10 s steps could be fit to a second-order exponential. Of the total current inactivated, 16% had a τ(Ba fast) of 645±56 ms and 65% a τ(Ba slow) of 7.4±1 s (n=7, Fig. 1E). A completely noninactivating component remained that constituted 18% of the total current. Ca2+ channel currents could be completely blocked by Cd 2+. Application of 1 mM Cd 2+ to cells caused a completely reversible block of 90±6% (n=3, Fig. 1F). The rapid run-down of Ca2+ currents and their small amplitude favoured the use of Ba2+ in the rest of this study. 3.2. Effect of w-Aga IVA on Ca2+ channel currents A range of concentrations of ω-Aga IVA was used to determine the maximal blocking concentration of toxin for subsequent studies and to examine the features of the inhibition. The effect of ω-Aga IVA on Ca2+ channel currents is demonstrated in Fig. 2. Block by ω-Aga IVA was extremely potent, the EC50 for inhibition was 60 pM. Maximal inhibition was seen at a concentration of 1 nM, at which currents evoked at ⫺2 mV were blocked by just over 20%. Although concentrations of up to 500 nM were applied to cells, there was not a significantly greater block at these concentrations compared to the inhibition at 1 nM. The onset of block was rapid, developing in 10–20 s, and usually irreversible within the course of an experiment. The mean current–voltage relationship is shown for cells in the absence and presence of 100 nM ω-Aga IVA (Fig. 2B). Following application of ω-Aga IVA, peak current was reduced by 25±5% and the midpoint of activation was shifted 5 mV to a slightly but significantly more hyperpolarised potential (⫺14±2 mV to ⫺19±3 mV; p=0.04; n=5). This shift and the observation that no current was blocked at potentials below ⫺12 mV suggest that channels with a higher activation threshold than the ω-Aga IVA-resistant current are blocked by this toxin. An example of the inhibition is shown in Fig. 2C. 3.3. Block of Ca2+ channels in cerebellar granule cells by w-CTx MVIIC The concentration–inhibition relationship for different concentrations of ω-CTx MVIIC showed that maximal inhibition occurred at 5 µM (Fig. 3A). At this concentration, whole-cell Ca2+ channel currents were reduced by 20±2%. Inhibition by ω-CTx MVIIC was rapid, with maximum block occurring in 10–20 s, and showed par-

1743

tial recovery. From a logistic fit to the data the EC50 was found to be 640 nM. Fig. 3B shows the average current–voltage relationship for 25 cells across a number of cell cultures before and during application of 5 µM ω-CTx MVIIC. ω-CTx MVIIC reduced the inward current with no effect on the voltage-dependence of activation. The inhibition observed in a single cell is depicted in Fig. 3C. 3.4. Effect of w-CTx GVIA on Ca2+ channel currents Previous studies have shown that 1 µM ω-CTx GVIA is a saturating concentration for its binding sites in neuronal preparations (Boland et al., 1991; Kasai et al., 1987; Plummer et al., 1989; Regan et al., 1991), so this concentration was used throughout. Fig. 4A demonstrates the effect of 1 µM ω-CTx GVIA on currents evoked by test pulses between ⫺82 and +58 mV. The toxin caused an inhibition of current at all voltages with no significant change in the voltage-dependence of activation. In a group of 18 cells from three different experiments the average inhibition of the peak Ca2+ channel current was 34±3%. Current examples (Fig. 4A inset) show inhibition of the peak current after application of ω-CTx GVIA in a single granule cell. ωCTx GVIA caused no discernible change in the kinetics of the current and the ω-CTx GVIA-sensitive current had similar kinetics to the control trace. The inhibition of currents by ω-CTx GVIA was rapid and completely irreversible even with long periods after removal of the drug. The irreversible nature of the block can be seen in Fig. 4B, which shows cumulative inhibition by ω-CTx GVIA and ω-CTx MVIIC. In this experiment application of 1 µM ω-CTx GVIA to a group of five cells caused 29±5% inhibition of the whole-cell current. After ceasing application of ω-CTx GVIA no recovery from inhibition was seen. After block by ω-CTx GVIA, 5 µM ωCTx MVIIC caused a rapid yet reversible block (reversible in four out of five cells) of 14±3% (as a percentage of the total current before block with ω-CTx GVIA). ω-CTx MVIIC at the same concentration produced an inhibition of 28±2% of the current when applied to naive cells. Prior application of ω-CTx GVIA appears from the data shown in Fig. 4B to reduce the proportion of current sensitive to ω-CTx MVIIC, which is in agreement with studies in other systems where ωCTx MVIIC and ω-CTx GVIA have been shown to compete for binding sites (Hillyard et al., 1992). Statistical examination of the data showed that the application of ω-CTx GVIA caused a significant reduction in the ωCTx MVIIC-sensitive fraction of current (p=0.0002; unpaired t-test). As the application of ω-CTx MVIIC produced a smaller block of current than the sum of individual uses of ω-CTx GVIA (34%) and ω-Aga IVA (25%, Figs. 2

1744

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 2. Effect of ω-Aga IVA on whole-cell Ca2+ channel currents in cerebellar granule neurones. (A) Concentration–response relationship showing potent block of Ca2+ channel currents by ω-Aga IVA. The data were described by a logistic fit (broken line). The EC50 was estimated to be 60 pM and h was found to be very close to 1. Each point represents the mean response from 4–7 cells. (B) Average current–voltage data for a group of five cells before and during challenge with 100 nM ω-Aga IVA. The data were fit using a Boltzmann equation. Averaged values for each parameter before and after addition of ω-Aga IVA were: g (nS), 5.2±0.7 and 3.2±0.2 (p=0.04; n=5); Erev (mV), +47±3 and +47±3; V50(act) (mV), ⫺14±2 and ⫺19±3 (p=0.04; n=5); k (mV), 8±1 and 7±1. (C) Example current recordings before and during application of 100 nM ω-Aga IVA. The trace labelled difference shows the ω-Aga IVA-sensitive current.

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

1745

Fig. 3. The effect of ω-CTx MVIIC on Ca2+ channel currents in cerebellar granule neurones. (A) Concentration–response curve for the effect of increasing concentrations of ω-CTx MVIIC on peak currents. The broken line shows a logistic fit to the data where the EC50=640 nM and h=3. Each point represents the mean response from 4–10 cells. (B) Current–voltage relationship for currents recorded before and during application of ω-CTx MVIIC. The data are the mean±s.e.m. of 25 cells pooled from several experiments and were fit with a Boltzmann equation. Averaged values for each parameter before and after addition of ω-CTx MVIIC were: g (nS), 4.4±0.3 and 2.9±0.3 (p=0.000002; n=25); Erev (mV), +48±2 and +51±1; V50(act) (mV), ⫺14±0.3 and ⫺15±1; k (mV), 8±1 and 7±0.3. (C) Example current recordings evoked by stepping to a command voltage of ⫺2 mV in a cell before and during application of ω-CTx MVIIC. The difference current was calculated by subtracting the ω-CTx MVIICresistant current from the control current.

1746

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 4. Effect of ω-CTx GVIA on Ca2+ channel currents in granule cells and the co-application of ω-CTx MVIIC. (A) Current–voltage relation showing inhibition by 1 µM ω-CTx GVIA in a group of nine cells. The individual data were fit with Boltzmann equations. Averaged values for each parameter before and after addition of ω-CTx were: g (nS), 3.4±0.4 and 1.9±0.3 (p⬍0.0001; n=9); Erev (mV), +48±2 and +51±1; V50(act) (mV), ⫺11±2 and ⫺13±2; k (mV), 8±0.7 and 8±0.5. The inset shows currents recorded in a single cell before and during application. The difference current illustrates the typical kinetics of the ω-CTx GVIA-sensitive current. (B) Time-course of application of 1 µM ω-CTx GVIA and 5 µM ωCTx MVIIC in a single neurone. Bars show periods of drug delivery. The example current traces show currents at the times indicated by the arrows.

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

and 3), ω-CTx MVIIC (28%) does not appear to target exactly the same populations of channels. The following section explores this discrepancy. 3.5. w-CTx MVIIC incompletely blocks w-CTx GVIAand w-Aga IVA-sensitive currents The broader selectivity of ω-CTx MVIIC for Ca2+ channel subtypes present on neurones is further investigated in Fig. 5. This figure depicts a determination of the proportion of ω-CTx GVIA-sensitive current and ωAga IVA-sensitive current remaining in cells after the application of a maximally blocking dose of ω-CTx MVIIC. In Fig. 5A, 5 µM ω-CTx MVIIC produced an average inhibition of 32±2% (n=5) which is in line with the inhibition shown previously here. In the continued presence of ω-CTx MVIIC, co-application of 1 µM ωCTx GVIA caused a further inhibition of Ca2+ channel currents which was calculated as a percentage of the total current before challenge with ω-CTx MVIIC to be 10±2%. ω-CTx MVIIC significantly reduced the ω-CTx GVIA-sensitive fraction of whole-cell current (p⬍0.0001; unpaired t-test; n=5, compared to application of ω-CTx GVIA without pre-treatment with ωCTx MVIIC; n=9). Fig. 5B shows a similar experiment in which cells were challenged with 5 µM ω-CTx MVIIC and in its continued presence were subsequently exposed to 100 nM ω-Aga IVA. Here, ω-CTx MVIIC produced a 30±3% inhibition of peak current (n=8). ω-Aga IVA further reduced the current by 16±4% in the presence of ω-CTx MVIIC. Comparing this to inhibition by ω-Aga IVA alone (Fig. 2B, 25±5%; n=5), the prior application of ω-CTx MVIIC reduced the ω-Aga IVA-sensitive component of current, although this did not reach statistical significance when inhibition was measured at a single potential. These experiments suggest that in these cells ω-CTx MVIIC blocks a greater proportion of the ω-CTx GVIAsensitive current than the ω-Aga IVA-sensitive current, but in neither case is there complete overlap. The fact that ω-CTx MVIIC causes a smaller block of total current than ω-CTx GVIA and ω-Aga IVA shows that ωCTx MVIIC does not target all ω-CTx GVIA-sensitive channels. In addition, substantial inhibition by ω-Aga IVA remained after incubation with ω-CTx MVIIC implying that this blocker does not target all P/Q-type channels.

1747

rent was blocked. In the continued presence of ω-CTx GVIA, application of 100 nM ω-Aga IVA only reduced the currents by 11±3%. This was not significantly different from the extent of inhibition (22±4%) by ω-Aga IVA applied alone to a group of 10 cells (p=0.11; unpaired t-test). Fig. 6B illustrates the converse situation. ω-Aga IVA (100 nM) was first applied to a group of five cells and after a steady-state inhibition had been reached, 1 µM ω-CTx GVIA was applied in conjunction with ω-Aga IVA. ω-Aga IVA reduced the evoked currents by an average of 17±2%. Prior application of ω-Aga IVA reduced the ω-CTx GVIA-sensitive current to 18±3% which was a very significant reduction from the ω-CTx GVIA-sensitive component in a group of 18 cells (34±3%) where ω-CTx GVIA was applied alone (p=0.0073; unpaired t-test). Thus it appears that ω-Aga IVA can significantly block channels that are also sensitive to ω-CTx GVIA. 3.7. No evidence of a Q-type current component in cerebellar granule neurones Other studies have demonstrated a component of current in cerebellar granule neurones (Randall and Tsien, 1995) and other neurones (Wheeler et al., 1994) that was less sensitive to ω-Aga IVA than the P-type current but blocked by ω-CTx MVIIC. However, in the present study the dose–response curve to concentrations of ωAga IVA did not show a biphasic relationship. To establish whether a component of “Q-type” current remained after block of N- and P-type currents, a low dose of ωAga IVA (1 nM as used in previous studies) was applied to cells together with 1 µM ω-CTx GVIA, and then subsequently challenged with ω-CTx MVIIC. Fig. 6C shows an example time-course of peak Ca2+ channel currents measured in a group of six cells first exposed to 1 nM ω-Aga IVA and 1 µM ω-CTx GVIA and then to a combination of 5 µM ω-CTx MVIIC, 1 nM ω-Aga IVA and 1 µM ω-CTx GVIA. The combined effects of ω-Aga IVA and ω-CTx GVIA decreased the current 42±3%. The subsequent addition of ω-CTx MVIIC did not produce a further significant decrease in the elicited current (3±3%, p=0.16; paired t-test; n=6). Thus it would appear that under the conditions used here a Q-type current was not present. 3.8. Block of Ca2+ channels in cerebellar granule neurones by the DHP antagonist nicardipine

3.6. w-Aga IVA blocks w-CTx GVIA-sensitive currents The overlap between ω-CTx GVIA- and ω-Aga IVAsensitive currents was further studied in the experiments shown in Fig. 6. These toxins were applied to groups of cells in combination. In a group of four cells that were exposed to 1 µM ω-CTx GVIA, 30±5% of the peak cur-

Fig. 7A demonstrates the concentration-dependence of the inhibition by nicardipine of the Ca2+ channel current in cerebellar granule neurones. Block of current was only seen at relatively high concentrations, and the EC50 for block was 4.4 µM. Maximal block was achieved by the application of 100 µM nicardipine and produced a

1748

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 5. Effect of the co-application of ω-CTx MVIIC and ω-CTx GVIA or ω-Aga IVA. (A) Average current–voltage curves for five cells recorded in the absence of toxin, in the presence of 5 µM ω-CTx MVIIC and in the presence of 5 µM ω-CTx MVIIC and 1 µM ω-CTx GVIA. Example current recordings in each condition are shown to the right of the figure. (B) Current–voltage data showing the effects of 5 µM ω-CTx MVIIC and 100 nM ω-Aga IVA on peak currents in granule neurones. Each point is the average of eight cells. Representative traces from a single cell are shown on the right.

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

1749

Fig. 6. Effects of the co-application of ω-Aga IVA and ω-CTx GVIA on peak Ca2+ channel currents in cerebellar granule neurones. (A) Timecourse of application of 1 µM ω-CTx GVIA and 100 nM ω-Aga IVA. After reaching a steady-state inhibition with ω-CTx GVIA the drug perfusion was switched to a solution containing ω-CTx GVIA and ω-Aga IVA. Bars show periods of drug delivery. Representative current traces are shown taken from the time-points indicated. (B) Similar experiment to (A) where steady-state inhibition with 100 nM ω-Aga IVA was obtained before application of both toxins. (C) 5 µM ω-CTx MVIIC produced no further block in the presence of 1 nM ω-Aga IVA and 5 µM ω-CTx GVIA. The time-course shown was representative of a group of six cells. 1 nM ω-Aga IVA and 5 µM ω-CTx GVIA were applied to cells until a steady-state inhibition had been reached. In the continued presence of ω-Aga IVA and ω-CTx GVIA the cells were challenged with 5 µM ω-CTx MVIIC. Example current recordings were taken from the times shown.

1750

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 7. Dose-dependence and effect on voltage-dependence of currents exposed to nicardipine. (A) Concentration–response curve for the effect of nicardipine. Percentage inhibition was plotted against drug concentration. Each point represents the mean of 4–14 observations. The dotted line is a logistic fit to the data with EC50=4.4 µM and h=1. (B) Current–voltage relationship for cells before and during application of 1 µM nicardipine. Points are the mean±s.e.m. of six separate applications (error bars are only shown in one direction for clarity). The individual data were fit with Boltzmann equations. Averaged values for each parameter before and after addition of nicardipine were: g (nS), 4.0±0.4 and 3.3±0.3 (p=0.03; n=6); Erev (mV), +48±3 and +49±5; V50(act) (mV), ⫺22±1 and ⫺24±3 (p=0.03; n=6); k (mV), 6±0.7 and 6±0.5. (C) Example recordings are shown to illustrate the reduction in current amplitude and kinetics of the nicardipine-sensitive current.

greater than 60% decrease in peak current amplitude. Complete recovery of current was usually seen after removal of the drug, and maximum inhibition occurred in around 60 s, developing more slowly than block with the Ca2+ channel toxins. The holding potential in these

experiments was ⫺92 mV. The voltage-dependence of DHP activity is well described in these cells (Marchetti et al., 1995) and other systems (Hamilton et al., 1987), and the EC50 would be expected to be reduced at more depolarised holding potentials. Fig. 7B shows the cur-

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

rent–voltage data for block with 1 µM nicardipine. At this concentration, the peak current was reduced by 16±7% (n=6). A small but statistically significant shift in the voltage-dependence was also observed on analysis of Boltzmann fits to the curves. V50(act) changed from ⫺22±1 mV in control to ⫺24±2 mV after application of nicardipine (p=0.03). Fig. 7C shows example traces from a cell before and during application of 1 µM nicardipine. 3.9. Nicardipine at 1–10 mM does not reduce the toxin-sensitive current component To determine how a low concentration of nicardipine, reported previously to be selective for L-type Ca2+ channels, blocked cerebellar Ca2+ channel currents in the presence of non-L-type blockers, studies were carried out in which 1 µM nicardipine was applied in the presence of one or more peptide toxins. Fig. 8A shows pooled current–voltage data and an example of current recordings from a single cell where 1 µM nicardipine was applied to cells following the application of 5 µM

1751

ω-CTx MVIIC. The average inhibition of peak current by nicardipine was 17±4% (n=9; calculated as the percentage of control current before application of ω-CTx MVIIC) and was not significantly different from the inhibition seen in Fig. 7 indicating that the 1 µM nicardipine-sensitive current was completely additive with the ω-CTx MVIIC-sensitive current. In the experiment illustrated in Fig. 7B a combination of 100 nM ω-Aga IVA, 1 µM ω-CTx GVIA and 5 µM ω-CTx MVIIC was applied to cells simultaneously. A decrease in the amplitude of Ca2+ channel current of 38±4% (n=7) was seen. In the presence of these toxins 1 µM nicardipine further reduced peak current by 17±3% as a percentage of the control amplitude. Under these conditions the 1 µM nicardipine-sensitive current was equivalent to the nicardipine-sensitive current recorded when applied to a group of six cells alone. These data suggest that 1 µM nicardipine reduced a component of the whole-cell current, presumably L-type current, that was not susceptible to the effects of any of the toxins used here.

Fig. 8. Inhibition by nicardipine in the presence of toxins. (A) 1 µM nicardipine was applied to a group of cells in the presence of 5 µM ω-CTx MVIIC. The current–voltage relationships show reduction of currents in these conditions. The inset shows recordings taken from the peak of the curve in control, in the presence of ω-CTx MVIIC and in the presence of nicardipine and ω-CTx MVIIC. (B) Time-course showing inhibition of 1 µM nicardipine following block with 100 nM ω-Aga IVA, 1 µM ω-CTx GVIA and 5 µM ω-CTx MVIIC. Representative currents are taken from the time-points indicated. (C) Change in the current–voltage relationship for a group of six cells exposed to 10 µM nicardipine. Currents recorded in each condition are shown in the inset. (D) Time-course of inhibition by 10 µM nicardipine and the subsequent addition of 5 µM ωCTx MVIIC. The inset shows currents recorded in control conditions and at each of the time-points indicated.

1752

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

3.10. Concentrations of nicardipine above 10 mM show less selectivity for L-type currents Fig. 8C shows the effect of the application of 10 µM nicardipine: the average reduction in amplitude of currents measured at their peak was 50±7% (n=6). From current–voltage relationships an extremely significant hyperpolarising shift in the voltage for half-activation was seen (⫺12.7±3 mV to ⫺19.6±3 mV, p=0.0008). The substantial inhibition brought about by applying 10 µM nicardipine to the cells begs the question as to whether it is selective at these concentrations for L-type Ca2+ channels. To test this hypothesis 10 µM nicardipine was applied to a group of cells. After a steady-state of inhibition had been reached, 5 µM ω-CTx MVIIC was applied without removal of the nicardipine. In a group of six cells, 10 µM nicardipine produced 59±7% inhibition. In the presence of 10 µM nicardipine, Ca2+ channel currents in the same cells were further reduced by 5 µM ω-CTx MVIIC by 21±4% as a percentage of the total current before challenge with nicardipine (Fig. 8D). If nicardipine was causing non-specific blockade of nonL-type Ca2+ channels at concentrations above 1 µM, the ω-CTx MVIIC-sensitive component of the total wholecell current might be expected to be reduced. In this case 10 µM nicardipine did not affect the ω-CTx MVIICsensitive contribution of the total current. It would appear that the populations of Ca2+ channels that are sensitive to up to 10 µM nicardipine do not overlap with those sensitive to ω-CTx MVIIC. It is clear from Fig. 7 that greater concentrations than 10 µM cause a reduction of whole-cell current to an extent that cannot be solely accounted for by L-type channels. It was shown that the non-L-type current blocked by the combined effect of the toxins used amounted to 45% of the total current, which suggests that nicardipine must block part of this current at concentrations in excess of 10 µM. 4. Discussion The early classification of HVA Ca2+ channels into L-, N- and P-type channels was largely based on the unique sensitivity of L-type channels to DHP antagonists (Fox et al., 1987a,b; Nowycky et al., 1985a,b), of Ntype channels to ω-CTx GVIA (McCleskey et al., 1987), and of P-type channels to ω-Aga IVA (Mintz et al., 1992b). However, part of the Ca2+ currents in some tissues are insensitive to these agents (Bossu et al., 1994; Mintz et al., 1992a) and, along with evidence from molecular cloning techniques, there has been shown to be a much greater diversity of Ca2+ channel types in the brain than was previously recognised. Correlating the structurally diverse Ca2+ channels identified by molecular cloning with those expressed in situ has recently become a highly debated issue.

Unfortunately, molecularly homogenous Ca2+ channels can show very different properties from one study to another, and in different expression systems. For instance, recent work has shown that ω-CTx GVIA can block an inactivating and a non-inactivating HVA Ca2+ current (Aosaki and Kasai, 1989; Artalejo et al., 1992; Mynlieff and Beam, 1992; Plummer and Hess, 1991; Plummer et al., 1989) demonstrating that inactivation and biophysical properties in general may not be reliable factors for relating results from one study to another. Pharmacological criteria would therefore seem to be more appropriate for defining channel types but problems are also inherent in this approach. The tools presently available do not have complete selectivity and show varied effects in different tissues (De Waard et al., 1991; Scott et al., 1991). It should also be noted that in the case of a toxin interacting with a Ca2+ channel protein, complete block of conductance is often not seen. It is often impossible to say in such situations whether the partial inhibition is the result of a second Ca2+ channel subtype being present that is resistant to the toxin or whether the residual current results from partial toxin block. Analysis of the voltage-dependence and biophysical properties of the current may not help to decipher this problem since the binding of a drug or toxin to an ion channel may change the basic parameters of conductance. Where combinations of toxins are used it should also be borne in mind that binding of one may affect the binding of the second toxin as reported for Ca2+ channels in undifferentiated neuroblastoma cells (Reeve et al., 1994). The results presented here demonstrate some of the problems inherent in dissecting out the individual Ca2+ current components from cells containing multiple channel subtypes. Overlapping selectivity of toxins to different classes of neuronal Ca2+ channel hinder clear conclusions being made. The assumption that ω-CTx GVIAsensitive currents are equivalent to N-type currents described elsewhere, and similarly for ω-Aga IVA-sensitive currents, may lead to erroneous definition of channel species. Fig. 9A summarises the relative effects of the toxins in inhibiting Ca2+ channel currents in cerebellar granule neurones. With regard to the proportions of total current carried by identified channel types, the data show that 40–45% of the Ca2+ channel current in these cells was carried through N- and P-type channels. It is impossible to establish exact proportions of both N- and P-types on the basis of these data alone, since a component of ωCTx GVIA-sensitive current was blocked by ω-Aga IVA, and a component of ω-Aga IVA-sensitive current was blocked by ω-CTx GVIA. In addition, ω-CTx MVIIC was shown to block a smaller proportion of total current than the combined application of ω-Aga IVA and ω-CTx GVIA demonstrating that not all N- and P-type currents are sensitive to this toxin. One possible expla-

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

Fig. 9. (A) Summary of the effects of ω-CTx MVIIC, ω-CTx GVIA and ω-Aga IVA applied to granule cells individually or in combination. Bars represent the mean inhibition±s.e.m. where n=5–25. (B) Summary of the effects of nicardipine applied to granule cells individually or in combination with various toxins. Bars represent the mean inhibition±s.e.m. where n=5–9. The lowest division represents the drug(s) applied first in every case.

nation of these results is that a novel subset of channels carrying approximately 10–15% of the total current exists with mixed pharmacology, being sensitive to both ω-Aga IVA and ω-CTx GVIA, yet insensitive to ω-CTx MVIIC. In this scenario, current components contributing to the total whole-cell current would be 20% ω-CTx GVIA-sensitive N-type, 10% ω-Aga IVA-sensitive Ptype, and 10–15% accounted for by channels with mixed pharmacology, but insensitive to ω-CTx MVIIC. An overlapping pharmacology is not without support. Cells of the sympathetic ganglion have been shown to have currents sensitive to ω-Aga IVA and ω-CTx GVIA

1753

(Mintz et al., 1992a). Also, a low-affinity binding site for ω-Aga IVA has been shown to be present on N-type and possibly other Ca2+ channel species, introducing the idea that this toxin is not as selective as was initially believed (Mintz and Sidach, 1998). However, in the present study all block by ω-Aga IVA was of high affinity. In agreement with some studies (Pearson et al., 1995), but not others (Randall and Tsien, 1995), under the conditions used here no evidence was obtained for the presence of a Q-type current based on the sensitivity of currents to ω-Aga IVA, as currents were maximally blocked (25% of the total whole-cell current) by a concentration of 1 nM, and the ω-Aga IVA and ω-CTx MVIIC-sensitive currents were non-inactivating. Other studies that demonstrated a Q-type component in cerebellar granule neurones (Randall and Tsien, 1995) showed a subdivision of the current sensitive to ω-Aga IVA. They showed that only 10% of the total current was blocked by 1–3 nM ω-Aga IVA, but a further 35% of total current was sensitive to 90 nM ω-Aga IVA. Similar concentrations of ω-CTx MVIIC as used here were found to block both of these ω-Aga IVA-sensitive currents. This difference may result from differences in culture or recording conditions. In the latter case cells were cultured in a non-depolarising medium containing 5 mM K +, which may lead to a different pattern of expression of Ca2+ channels. Recent work has shown that P- and Qtype Ca2+ channels result from the expression of alternate splice forms of the α1A subunit gene which may explain the differences in biophysical and pharmacological properties of ω-Aga IVA-sensitive channels seen under different conditions (Bourinet et al., 1999). Interestingly, Pearson et al. (1995) showed a similar proportion of current blocked by 100 nM ω-Aga IVA (40%) and the toxin appeared to be blocking just one population of channels with an EC50 of 2.7 nM. Another important factor worth considering is the age-dependent change in Ca2+ channel expression. In our studies significant changes in the proportion of Ca2+ current components over the 7–14 DIV range were not apparent. The study of Harrold et al. (1997) showed that Ca2+ channels contributing to K +-evoked Ca2+-entry into the soma change over time in culture. Granule neurones maintained for longer than 8 DIV showed a decrease in the DHP-sensitive component of the plateau Ca2+ elevation and a corresponding increase in the 200 nM ω-Aga IVA sensitive component up to 28 DIV. It is possible that a component of the Ca2+ current with sensitivity to high concentrations of ω-Aga IVA develops at a later stage in granule cell cultures. Between different studies in cerebellar granule neurones, the proportion of current reported to be sensitive to ω-CTx GVIA is largely consistent. Pearson et al. (1993, 1995) reported that 25% of the current was blocked by 1 µM ω-CTx GVIA and Randall and Tsien

1754

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

(1995) reported 20% block at this concentration. Regarding the proportion of L-type/DHP-sensitive currents, estimates vary. Nimodipine (1 µM) in Randall’s study blocked just 15% of the total current, whereas Pearson showed that (⫺)-202-791 blocked 35% of total current at 1 µM. In De Waard’s study (1991) nicardipine at 200 µM maximally blocked 78% of current and showed an EC50 of 0.5 µM. Thus, a high degree of variation is shown in response to DHP compounds and their selectivity to L-type channels is questionable. Here, in agreement with De Waard et al. (1991), a large component of the Ca2+ channel current in cerebellar granule neurones was found to be sensitive to the DHP antagonist nicardipine. When concentrations in the range 1–10 µM were applied to cells, the current sensitive to nicardipine was completely additive with the components of the whole-cell current sensitive to ω-CTx MVIIC or a combination of ω-CTx MVIIC, ω-Aga IVA and ω-CTx GVIA. This suggests that nicardipine blocks a population of channels at these concentrations that is clearly distinct from those blocked by ω-CTx MVIIC, ω-Aga IVA and ω-CTx GVIA. At higher concentrations nicardipine inhibited currents in these cells by a proportion which could not be solely attributed to non-N/non-P-type channels. At 100 µM, currents were blocked by 64±5% following application of nicardipine. In previous studies in granule neurones a similarly large reduction in Ca2+ channel currents has been observed after exposure to nicardipine (De Waard et al., 1991). This study also examined the effect of ω-CTx GVIA and concluded that a population of channels was present with a mixed pharmacology being sensitive to nicardipine and ω-CTx GVIA. More recently, again fitting in with the results here, Diochot et al. (1995) showed that DHPs and other Ca2+ antagonists have a substantial inhibitory effect on neuronal non-L-type Ca2+ channels. Nicardipine in the latter study blocked HVA currents in DRG neurones completely at a concentration of 200 µM. It seems reasonable to conclude from the data here that nicardipine at high concentrations has non-selective blocking effects on various Ca2+ channel currents including ω-Aga IVA-sensitive and ω-CTx GVIA-sensitive currents. However, at lower concentrations up to 10 µM it appears not to have a significant interaction with channels carrying these toxin-sensitive currents. A proportion of the nicardipine-sensitive current must be carried by L-type channels of either the C or D class since characteristic effects of both DHP agonists (not shown) and antagonists have been observed and both these Ca2+ channel subunits are known to be expressed at high levels in the granule cells of the cerebellum (Hell et al., 1993). Some of the inhibition may also be accounted for by the effect of nicardipine against the α1E Ca2+ channel which has been shown to be sensitive to nicardipine (Stephens et al., 1997) and is also present in the cerebellum (Yokoyama et al., 1995). The α1E channel was first

put forward as the molecular counterpart of the R-type current component identified in cerebellar granule neurones (Randall and Tsien, 1995) but recently it has been shown that a toxin that blocks α1E channels (SNX-482) has no effect on R-type current in these cells (Newcomb et al., 1998). It is also possible that nicardipine blocks novel and as yet unidentified Ca2+ channels present in cerebellar granule cells. The effect of nicardipine in association with other channel antagonists is summarised graphically in Fig. 9B. At a concentration usually referred to as being selective for L-type Ca2+ channels (1 µM), nicardipine blocked just less than 20% of the total whole-cell current. At this concentration, and up to 10 µM, the nicardipine-sensitive current did not overlap with currents carried by toxin-sensitive Ca2+ channels. Higher concentrations (⬎10 µM) of nicardipine, however, blocked more of the total current than could be solely attributed to L-type channels. In this study we have shown that it is difficult to distinguish distinct populations of Ca2+ channels in cells that are known to express several different classes. The Ca2+ channel blockers may indeed be rather less selective than is generally presumed and their use in the definition of native channel species should be approached with careful consideration.

Acknowledgements This work was supported by the Wellcome Trust.

References Aosaki, T., Kasai, H., 1989. Characterization of two kinds of highvoltage-activated Ca-channel currents in chick sensory neurons. Differential sensitivity to dihydropyridines and omega-conotoxin GVIA. Pfluegers Archives 414, 150–156. Artalejo, C.R., Perlman, R.L., Fox, A.P., 1992. Omega-conotoxin GVIA blocks a Ca2+ current in bovine chromaffin cells that is not of the “classic” N type. Neuron 8, 85–95. Boland, L.M., Allen, A.C., Dingledine, R., 1991. Inhibition by bradykinin of voltage-activated barium current in a rat dorsal root ganglion cell line: role of protein kinase C. Journal of Neuroscience 11, 1140–1149. Bossu, J.L., De Waard, M., Fagni, L., Tanzi, F., Feltz, A., 1994. Characteristics of calcium channels responsible for voltage-activated calcium entry in rat cerebellar granule cells. European Journal of Neuroscience 6, 335–344. Bourinet, E., Soong, T.W., Sutton, K., Slaymaker, S., Mathews, E., Monteil, A. et al., 1999. Splicing of a1A subunit gene generates phenotypic variants of P- and Q-type calcium channels. Nature Neuroscience 2, 407–415. De Waard, M., Feltz, A., Bossu, J.L., 1991. Properties of a high-threshold voltage-activated calcium current in rat cerebellar granule cells. European Journal of Neuroscience 3, 771–777. Diochot, S., Richard, S., Baldy-Moulinier, M., Nargeot, J., Valmier, J., 1995. Dihydropyridines, phenylalkylamines and benzothiazepines

J.R. Burley, A.C. Dolphin / Neuropharmacology 39 (2000) 1740–1755

block N-, P/Q- and R-type calcium currents. Pfluegers Archives 431, 10–19. Forti, L., Pietrobon, D., 1993. Functional diversity of L-type calcium channels in rat cerebellar neurons. Neuron 10, 437–450. Forti, L., Tottene, A., Moretti, A., Pietrobon, D., 1994. Three novel types of voltage-dependent calcium channels in rat cerebellar neurons. Journal of Neuroscience 14, 5243–5256. Fox, A.P., Nowycky, M.C., Tsien, R.W., 1987a. Kinetic and pharmacological properties distinguishing three types of calcium currents in chick sensory neurones. Journal of Physiology 394, 149–172. Fox, A.P., Nowycky, M.C., Tsien, R.W., 1987b. Single-channel recordings of three types of calcium channels in chick sensory neurones. Journal of Physiology 394, 173–200. Hamilton, S.L., Yatani, A., Brush, K., Schwartz, A., Brown, A.M., 1987. A comparison between the binding and electrophysiological effects of dihydropyridines on cardiac membranes. Molecular Pharmacology 31, 221–231. Harrold, J., Ritchie, J., Nicholls, D., Smith, W., Bowman, D., Pocock, J., 1997. The development of Ca2+ channel responses and their coupling to exocytosis in cultured cerebellar granule cells. Neuroscience 77, 683–694. Hell, J.W., Westenbroek, R.E., Warner, C., Ahlijanian, M.K., Prystay, W., Gilbert, M.M. et al., 1993. Identification and differential subcellular localization of the neuronal class C and class D L-type calcium channel alpha 1 subunits. Journal of Cell Biology 123, 949–962. Hillyard, D.R., Monje, V.D., Mintz, I.M., Bean, B.P., Nadasdi, L., Ramachandran, J. et al., 1992. A new Conus peptide ligand for mammalian presynaptic Ca2+ channels. Neuron 9, 69–77. Kasai, H., Aosaki, T., Fukuda, J., 1987. Presynaptic Ca-antagonist omega-conotoxin irreversibly blocks N-type Ca-channels in chick sensory neurons. Neuroscience Research 4, 228–235. McCleskey, E.W., Fox, A.P., Feldman, D.H., Cruz, L.J., Olivera, B.M., Tsien, R.W., Yoshikami, D., 1987. Omega-conotoxin: direct and persistent blockade of specific types of calcium channels in neurons but not muscle. Proceedings of the National Academy of Sciences of the United States of America 84, 4327–4331. Marchetti, C., Amico, C., Usai, C., 1995. Functional characterization of the effect of nimodipine on the calcium current in rat cerebellar granule cells. Journal of Neurophysiology 73, 1169–1180. Mintz, I.M., Sidach, S., 1998. Low-affinity block of the N-type and related Ca channels by omega-Aga-IVA. Society for Neuroscience 24, 407.1. Mintz, I.M., Adams, M.E., Bean, B.P., 1992a. P-type calcium channels in rat central and peripheral neurons. Neuron 9, 85–95. Mintz, I.M., Venema, V.J., Swiderek, K.M., Lee, T.D., Bean, B.P., Adams, M.E., 1992b. P-type calcium channels blocked by the spider toxin omega-Aga-IVA. Nature 355, 827–829. Mynlieff, M., Beam, K.G., 1992. Characterization of voltage-dependent calcium currents in mouse motoneurons. Journal of Neurophysiology 68, 85–92.

1755

Newcomb, R., Szoke, B., Palma, A., Wang, G., Chen, X.H., Hopkins, W. et al., 1998. Selective peptide antagonist of the class E calcium channel from the venom of the tarantula Hysterocrates gigas. Biochemistry 37, 15353–15362. Nowycky, M.C., Fox, A.P., Tsien, R.W., 1985a. Long-opening mode of gating of neuronal calcium channels and its promotion by the dihydropyridine calcium agonist Bay K 8644. Proceedings of the National Academy of Sciences of the United States of America 82, 2178–2182. Nowycky, M.C., Fox, A.P., Tsien, R.W., 1985b. Three types of neuronal calcium channel with different calcium agonist sensitivity. Nature 316, 440–443. Pearson, H.A., Sutton, K.G., Scott, R.H., Dolphin, A.C., 1993. Ca2+ currents in cerebellar granule neurones: role of internal Mg2+ in altering characteristics and antagonist effects. Neuropharmacology 32, 1171–1183. Pearson, H.A., Sutton, K.G., Scott, R.H., Dolphin, A.C., 1995. Characterization of Ca2+ channel currents in cultured rat cerebellar granule neurones. Journal of Physiology 482, 493–509. Plummer, M.R., Hess, P., 1991. Reversible uncoupling of inactivation in N-type calcium channels. Nature 351, 657–659. Plummer, M.R., Logothetis, D.E., Hess, P., 1989. Elementary properties and pharmacological sensitivities of calcium channels in mammalian peripheral neurons. Neuron 2, 1453–1463. Randall, A., Tsien, R.W., 1995. Pharmacological dissection of multiple types of Ca2+ channel currents in rat cerebellar granule neurons. Journal of Neuroscience 15, 2995–3012. Reeve, H.L., Vaughan, P.F., Peers, C., 1994. Calcium channel currents in undifferentiated human neuroblastoma (SH-SY5Y) cells: actions and possible interactions of dihydropyridines and omega-conotoxin. European Journal of Neuroscience 6, 943–952. Regan, L.J., Sah, D.W., Bean, B.P., 1991. Ca2+ channels in rat central and peripheral neurons: high- threshold current resistant to dihydropyridine blockers and omega-conotoxin. Neuron 6, 269–280. Scott, R.H., Pearson, H.A., Dolphin, A.C., 1991. Aspects of vertebrate neuronal voltage-activated calcium currents and their regulation. Progress in Neurobiology 36, 485–520[Review]. Stephens, G.J., Page, K.M., Burley, J.R., Berrow, N.S., Dolphin, A.C., 1997. Functional expression of rat brain cloned α1E calcium channels in COS-7 cells. Pfluegers Archives 433, 523–532. Tottene, A., Moretti, A., Pietrobon, D., 1996. Functional diversity of P-type and R-type calcium channels in rat cerebellar neurones. Journal of Neuroscience 16, 6353–6363. Wheeler, D.B., Randall, A., Tsien, R.W., 1994. Roles of N-type and Q-type Ca2+ channels in supporting hippocampal synaptic transmission. Science 264, 107–111[see comments]. Yokoyama, C.T., Westenbroek, R.E., Hell, J.W., Soong, T.W., Snutch, T.P., Catterall, W.A., 1995. Biochemical properties and subcellular distribution of the neuronal class E calcium channel alpha 1 subunit. Journal of Neuroscience 15, 6419–6432.