Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Plasmalemmal voltage-gated calcium channels (VGCCs) initiate intracellular calcium transients and control many aspects of neuronal processes, including the generation of calcium-dependent action potentials, neurotransmitter release, regulation of neuronal death, synapse formation and elimination, phenotypic differentiation, and gene expression (Ghosh and Greenberg, 1995; Gu and Spitzer, 1997). These channels are activated by membrane depolarization, leading to a transmembrane calcium influx and a transient increase in cytoplasmic free calcium concentration. It has become clear, however, that these voltage-dependent calcium transients also are generated by mechanisms involving the release of calcium from intracellular stores in the sarco-endoplasmic reticulum. In neurons, at least two types of calcium stores have been identified in the sarco-endoplasmic reticulum, based on calcium-release channels: caffeine-ryanodine-sensitive and inositol-1,3,4-triphosphate-sensitive calcium channels. The relationships among voltage-activated calcium influx, release of calcium from cytoplasmic stores, and intracellular calcium transients are complex and are only just beginning to be understood.

Five VGCC subtypes (T, L, N, P/Q, and R) were initially defined according to electrophysiological and pharmacological characteristics (Birnbaumer et al., 1994) and, more recently, these definitions have been extended to take into account amino acid sequences (Perez-Reyes and Schneider, 1994; Perez-Reyes et al., 1998). Similarly, at least three intracellular calcium channel subtypes have been characterized at the molecular level for both the ryanodine and inositol-1,3,4-triphosphate receptors. Such diversity raises the possibility that calcium channels may combine in different ways in a single neuron, thereby controlling specific pathways for calcium signaling, which may in turn have different functions. Therefore, the pharmacological isolation of channels with specific agents is essential if we are to elucidate the roles of the various channel families in the regulation of neuronal calcium signaling.

The chemical 2,5-di(tert-butyl)-1,4-benzohydroquinone (tBHQ) belongs to a large family of synthetic phenolic antioxidants commonly used as food preservatives. In addition to this beneficial effect, these compounds are potentially carcinogenic and, paradoxically, play a significant role in oxidative stress (van Esch, 1986; Yu et al., 1997). Independent of its ability to modify the thiol groups of proteins (Moore et al., 1987), tBHQ also has been shown to inhibit the calcium ATPase of the sarco-endoplasmic reticulum (SERCA), which is involved in the refilling of [Ca²⁺]_i stores (Kostyuk and Verkhratsky, 1994). This inhibition of SERCA activity has been used in experiments designed to detect [Ca²⁺]_i stores in neuronal preparations under various experimental conditions and to evaluate their relative contributions. One study reported that tBHQ inhibited L-type calcium currents in neuroendocrine cells (Nelson et al., 1994), but no data are available concerning the effects of this compound on neuronal VGCCs. We show herein that, at concentrations used to block SERCA, tBHQ also inhibited N- and P/Q-type VGCCs in peripheral and central neurons without affecting neuronal T-, L-, and R-type VGCCs. This effect seemed to involve a direct action on VGCCs, and may generate new strategies for the development of new classes of "calcium antagonists."

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cell Cultures. Spinal motoneurons from embryonic day 15 (E15) Sprague-Dawley rat embryos were purified by a two-step metrizamide-panning method, as previously described (Camu et al., 1993). Briefly, ventral spinal cords were treated with trypsin, dissociated, and centrifuged on 6.5% metrizamide cushions. The yield of large cells, which were not dense enough to pass through the metrizamide, was increased by immunopanning on Petri dishes coated with an antibody directed against the p75 neurotrophin receptor (low-affinity nerve growth factor receptor), specifically expressed by motoneurons at this stage (Yan and Johnson, 1988).

Electrophysiological Recordings. Ca²⁺ current (ICa) from dorsal root ganglions (DRGs) was recorded 2 h after dissociation. For motoneurons, ICa was recorded after 1 or 2 days in culture. ICa in hippocampal pyramidal cell-like neurons was recorded after 3 days in culture. ICa in NG108-15 cells was recorded at 4 or 5 days of differentiation. Whole-cell recordings were made at 20-22°C under conditions optimized so as to ensure the isolation of ICa from other voltage-activated currents. The bathing solution contained 115 mM tetraethylammonium chloride, 5 mM BaCl₂ (10 mM for NG108-15 cells), 10 mM HEPES, 10 mM glucose, and 1 µM tetrodotoxin with the pH adjusted to 7.35 with CsOH. Recording pipettes were filled with the following solution: 135 mM CsCl, 20 mM 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA), 10 mM HEPES, 3 mM Mg-ATP, 1 mM Mg-GTP, and 10 mM glucose, pH 7.35 (adjusted with CsOH). The pipette resistance was 4 M. Whole-cell currents were recorded with a Bio-Logic RK300 patch-clamp amplifier (Claix, France). After seal formation and membrane disruption, series resistance was estimated from the capacitative transient evoked by a +10-mV test pulse. Membrane capacity was calculated as Cm = /Rs, with  being the time constant of capacitative transient and Rs the series resistance. Because current amplitudes were always 1 nA, voltage errors due to uncompensated series resistance were negligible (<4 mV). All experimental parameters were controlled with a computer equipped with a DigiData 1200 analog interface (Axon Instruments, Foster City, CA). Data acquisition and analysis were performed with the pClamp software (version 6.03; Axon Instruments). Current signals were sampled at 5 kHz and filtered at 3 kHz. They were then digitized and stored.

Dye Loading and Measurements of [Ca²⁺]_i. Dye loading and measurement of [Ca²⁺]_i was performed as previously described (Dayanithi et al., 1996). Briefly, the culture dishes were washed and loaded by incubation with 2.5 µM Fura-2-AM and 0.05% w/v Pluronic F-127 (Molecular Probes, Eugene, OR) in Locke buffer (140 mM NaCl, 1.2 mM MgSO₄, 1.8 mM CaCl₂, 10 mM glucose, 5 mM KH₂PO₄, and 10 mM HEPES-NaOH, pH 7.25) at 34°C for 40 min. Loaded cells were washed with Locke buffer and fluorescence measurements were performed at room temperature. [Ca²⁺]_i in single cells was measured with the digital imaging microfluorimetry system (Axon Instruments) based on an inverted microscope equipped with epifluorescence optics (Nikon, Champigny-sur-Marne, France). Interference filters of 340/10 nm and 380/10 nm were alternately mounted on the filter wheel and the excitation light beam was deflected through an oil-immersion objective (40× 0.75 numerical aperture; Nikon). Fluorescence measurements were converted to [Ca²⁺]_i with a standard equation (Grynkiewicz et al., 1985) for ratiometric dyes [Ca²⁺]_i = K_d × (S_f2/S_b2) × (R  R_min)/(R_max  R) where K_d is the dissociation constant (0.145 µM for Fura-2; Molecular Probes); S_f2/S_b2 is the ratio of the signals obtained at 380 nm in the nominal absence and presence of saturating concentrations of Ca²⁺, respectively; R_min and R_max are the fluorescence ratios determined in the nominal absence and presence of saturating concentrations of Ca²⁺; and R is the fluorescence ratio measured. For 11 motoneurons, S_f2/S_b2 was 2.7 ± 0.1, R_min was 0.45± 0.01, and R_max was 2.6 ± 0.1.

Drugs. Nitrendipine (Bayer AG, Wuppertal, Germany), tBHQ, cyclopiazonic acid (CPA), and thapsigargin (Sigma, Saint Quentin Fallavier, France) were dissolved in DMSO to make concentrated stock solutions (10, 100, and 1 mM, respectively) and stored at 20°C. Controls showed that the solvent had no effect on ICa at the maximal final dilutions used herein (<0.05%). -Conotoxin GVIA (GVIA; Sigma) and -agatoxin IVA (AgaIVA; Peptide International, Louisville, KY, and Pfizer, New York, NY) were dissolved in double distilled water at 0.1 and 1 mg/ml, respectively, to give stock solutions. Stock solutions of NiCl₂ and CdCl₂ were prepared in double distilled water at 10 mM. Test solutions were prepared daily with aliquots from frozen stocks to obtain the working concentrations. Results are expressed as mean ± S.D. Means were compared with Student's t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

tBHQ and Thapsigargin Do Not Reduce [Ca²⁺]_i to the Same Extent in Cultured Motoneurons. The basal [Ca²⁺]_i of rat embryonic motoneurons was 36 ± 2 nM and application of 100 mM K⁺ induced an increase in [Ca²⁺]_i due to the opening of calcium channels, with the concentration reaching 475 ± 63 nM (n = 43; Fig. 1). To evaluate the spontaneous run-down of the response, a second depolarization with K⁺ was applied 10 min after the first. The run-down after the second application of K⁺ amounted to a 20 ± 2% decrease in the peak [Ca²⁺]_i (n = 30; Fig. 1A). In another series of experiments, cells were incubated for 10 min with 250 nM thapsigargin or 10 µM CPA, the maximal inhibitory doses known to completely inhibit SERCA (Thastrup et al., 1990). Thapsigargin decreased peak [Ca²⁺]_i by 51 ± 2% (n = 25) and CPA by 50 ± 2% (n = 17). These values were significantly different from those obtained for run-down (P < .001). After correction for run-down, thapsigargin was found to have caused a 31 ± 2% (n = 25) decrease in maximal amplitude of the intracellular calcium transient and CPA at 30 ± 3% (n = 17; Fig. 1B). This suggests that intracellular stores and calcium influx contributed 30 and 70%, respectively, of the intracellular calcium transient induced by depolarizing stimuli. Application of 10 µM tBHQ, which gives maximal inhibition of SERCA (Moore et al., 1987), resulted in a 60 ± 3% decrease in the intracellular calcium transient (n = 15; P < .001 relative to run-down). Corrected for run-down, tBHQ gave 40 ± 3% inhibition (n = 15; Fig. 1C), and therefore had a significantly stronger effect than thapsigargin (P < .05).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Effects of SERCA pump inhibitors on K+-induced [Ca2+]i responses in embryonic rat motoneurons. A, effects of two successive applications of 100 mM K+ on the [Ca2+]i responses are illustrated. The cell was allowed to recover for 10 min between each depolarization. B and C, effects of 10-min incubation of SERCA pumps inhibitors thapsigargin and tBHQ, respectively, on K+-induced [Ca2+]i responses.

tBHQ Reduced ICa in Cultured Motoneurons. It has been reported that tBHQ inhibits L-type ICa in non-neuronal preparations. We therefore used the patch-clamp technique to evaluate and to compare the effects of thapsigargin and tBHQ on the voltage-activated ICa of embryonic motoneurons. ICa were recorded with a 300-ms depolarization to 0 or +10 mV from a 100-mV holding potential (HP) to ensure maximal amplitude. At this voltage, ICa was composed of a sustained and a transient component. The sustained component was quantified by measuring the absolute amplitude of ICa at 250 ms, whereas the transient component was calculated as the difference between peak inward current and sustained current. After breaking the membrane patch, repetitive cell stimulation every 10 s initially induced an increase in ICa amplitude, ICa run-up, as is usually reported in cardiac as well other neuronal cells (Tiaho et al., 1993; Diochot et al., 1995). Once ICa had reached a steady amplitude, on four cells tested, the application of thapsigargin (250 nM) had no effect (Fig. 2A). In seven other cells, the application of 10 µM tBHQ induced a decrease in ICa amplitude corresponding to 24 ± 8% inhibition for the transient component and 13 ± 4% inhibition for the sustained component (n = 7; Fig. 2B). The onset of ICa inhibition was fast and complete within 10 s. Dose-response curves for tBHQ were constructed from the transient and sustained components of macroscopic ICa. tBHQ inhibited both the transient and sustained components of ICa with similar half-maximal inhibitory concentrations, IC₅₀ = 35 µM, and the maximum inhibition was 75 to 80% of macroscopic ICa (Fig. 2C).

View larger version (13K): [in this window] [in a new window]   Fig. 2.   Effects of SERCA pump inhibitors on the ICa amplitude of embryonic rat motoneurons. A, superfusion of 250 nM thapsigargin had no effects on ICa amplitude. B, at 10 µM, tBHQ induced a substantial decrease in ICa amplitude. C, dose-response curves for tBHQ inhibition of transient () and sustained () components of macroscopic ICa. Current recorded from a HP of 100 mV to a depolarization to +10 mV, every 10 s.

tBHQ Selectively Inhibited N- and P-Types of ICa But Did Not Affect L- and R-Types of ICa in Motoneurons. In a previous study, we showed that the macroscopic ICa of E15 rat motoneurons consists of four high voltage-activated ICa, the N, P, L, and R-type ICa, and we characterized these currents pharmacologically, kinetically, and functionally (Scamps et al., 1998; Table 1). A low voltage-activated, T-type current was observed in 10% of the cells studied with a negligible amplitude as previously reported (Magnelli et al., 1998) for a similar preparation after 7 days in culture. Because 75% of macroscopic ICa were sensitive to tBHQ, we investigated whether tBHQ had some specificity with respect to calcium channel types. After application of a supramaximal concentration of tBHQ (300 µM), we evaluated the effects of 3 µM GVIA, 1 µM nitrendipine, and 50 nM AgaIVA as specific antagonists of the N-, L-, and P type channels, respectively (Fig. 3A). In the presence of 300 µM tBHQ, GVIA and AgaIVA had no effect (n = 4), but nitrendipine induced a 27 ± 3% decrease relative to the sustained ICa amplitude, an effect similar to that obtained under control conditions (n = 7; Fig. 3B). These results suggest that tBHQ totally inhibits the N- and P-type channels without affecting L- and R-type channels.

View larger version (24K): [in this window] [in a new window]   Fig. 3.   Specificity of tBHQ inhibition of Ca2+ channel type in embryonic rat motoneurons. A, macroscopic high voltage-activated ICa of E15 rat embryonic motoneuron is composed of the P-, N-, L-, and R-type Ca2+ channels as shown by the use of 50 nM AgaIVA, 3 µM GVIA, and 1 µM nitrendipine, respectively. Nickel/cadmium (Ni/Cd; 100 µM) was added to evaluate the true zero current (A1, time course of ICa inhibition; A2, corresponding current traces). B, application of a supramaximal dose of tBHQ (300 µM) decreased ICa amplitude and prevented the effects of the specific inhibitors of the P- and N-type channels (B1, time course effect of tBHQ and subsequent application of the specific Ca2+ channels inhibitors; B2, corresponding current traces). Same protocol as in Fig. 2.

N-, P-, and Q-Type ICa But Not T-, L-, and R-Type ICa Is Inhibited by tBHQ in Hippocampal and DRG Neurons. We checked whether the observed effects were tissue-specific with two other preparations, originating from the central nervous system (hippocampus) and from the peripheral nervous system (DRGs). After 3 days in culture, the macroscopic ICa of the pyramidal cells from E17 hippocampus consisted of Q-, L-, and R-type channels because these neurons did not respond to the application of 3 µM GVIA (n = 6) and were sensitive only to a high concentration (250-500 nM) of AgaIVA (45 ± 6% inhibition; n = 5) and to 1 µM nitrendipine (35 ± 3% inhibition; n = 6; Fig. 4A₁). The R-type current accounted for ~10 to 20% of the sustained macroscopic current. A T-type current was observed in 40% of the cells studied and its amplitude was no >50 pA. This preparation was therefore of value for studying specifically the effects of tBHQ on the Q-type ICa. Superfusion with 300 µM tBHQ decreases sustained ICa amplitude by 61 ± 10% (n = 5). Under these conditions, AgaIVA had no effect, nitrendipine induced a decrease in current amplitude, and an R-type current was detected in the presence of these agents (Fig. 4A₂).

View larger version (25K): [in this window] [in a new window]   Fig. 4.   Specificity of tBHQ inhibition of Ca2+ channel type in embryonic hippocampal pyramidal neurons and DRG neurons. High voltage-activated ICa of E17 embryonic pyramidal neurons is composed of the Q-, L-, and R-type Ca2+ channels (A1), as determined by the use of AgaIVA and nitrendipine, respectively. Application of 300 µM tBHQ decreased ICa amplitude and prevented the effects of AgaIVA, but not of nitrendipine (A2). High voltage-activated ICa of E13 embryonic DRG neurons is composed of the P-, Q-, N-, L-, and R-type Ca2+ channels (B1), as determined by the use of AgaIVA [50 nM (P type) + 500 nM (Q type; data not shown)], GVIA, and nitrendipine, respectively. Application of 300 µM tBHQ decreased ICa amplitude and partially prevented the effects of GVIA while totally blocking the effects of AgaIVA (B2). For other tissues, the effects of nitrendipine were not affected despite a low contribution of the L-type channel to the macroscopic current.

View larger version (16K): [in this window] [in a new window]   Fig. 5.   Sensitivity of T-type ICa and of P/Q- and N-type ICa to tBHQ in a DRG neuron. Kinetics of tBHQ effects. A, effects of tBHQ on the T-type ICa and on the P/Q- and N-type ICa. A1, to record the P/Q-type ICa, the cell was incubated in the presence of 250 nM nitrendipine and 3 µM GVIA. A2, to record the N-type ICa, the cell was incubated in the presence of 250 nM nitrendipine and 250 nM AgaIVA. As shown below current traces, a two-pulse protocol was applied to record T-type ICa at a 40-mV depolarization and the drug-selected, high voltage-activated ICa at a 0-mV depolarization, 1-s interpulse every 10 s, 100-mV HP. Concentrations of 30 and 100 µM tBHQ were cumulatively applied to the cell. B, kinetics of tBHQ inhibition and reversibility. Application of supramaximal dose of tBHQ induced a rapid inhibition of macroscopic ICa recorded from a DRG neuron (no inhibitors). Washout of the drug allows for complete recovery of inhibition.

tBHQ Inhibits Currents Generated by Injection of cDNAs Encoding 1A, 1B, 1C, and 1E into Oocytes. To determine more accurately the site of action of tBHQ, we injected the -subunits encoded by the A, B, C, and E genes into oocytes. tBHQ at 30 µM inhibited the 1A, 1B, 1C, and 1E subunit currents, suggesting that this compound acts specifically on the pore-forming subunit of the calcium channel (Fig. 6A). At 30 µM, tBHQ inhibits barium currents recorded from oocytes expressing the 1A, 1B, 1C, or 1E calcium channel subunits by 58 ± 14 (n = 8), 64 ± 10 (n = 4), 30 ± 10 (n = 6), and 67 ± 13% (n = 6), respectively, when coexpressed with the 2- and 1 auxiliary subunits. Therefore the order of potency of tBHQ was 1E > 1B > 1A > 1C. As with the native proteins, the kinetics of tBHQ inhibition was rapid (Fig. 6B) and reversible. The insensitivity of the L- and R-type channels in neurons is in contradiction with the results obtained after expression of the 1C and specifically of the 1E subunits. One possible explanation of this discrepancy is the expression of specific combinations of auxiliary subunits with the 1E subunits in DRG, hippocampal neurons, and motoneurons. We have therefore tested the effects of coexpression of the 2, 3, and 4 subunits with the 1E and 1C subunits on the inhibition by tBHQ. As seen in Table 2, block of the 1E or 1C subunits by tBHQ was statistically (P > .05) identical for all subunit combinations tested.

View larger version (17K): [in this window] [in a new window]   Fig. 6.   tBHQ blocks neuronal VGCCs expressed in Xenopus oocytes. A, typical current traces showing steady-state inhibition of 1E, 1C, 1B, and 1A calcium channels by 30 µM tBHQ. Each 1 Ca2+ channel subunit was expressed with the subunits cDNA 2- and 1 (1E, 1B, and 1C) or 2 (1A). Currents were recorded during a typical depolarization at +10 mV from an HP of 80 mV in a Cl-free, 10 mM Ba2+-containing solution. B, time course of the Ba2+ current inhibition induced by perfusion of tBHQ. tBHQ (30 µM) was applied by superfusion to oocytes injected with 1E, 1C, 1B, or 1A. Peak currents were recorded during a typical depolarization from 80 to +10 mV applied every 15 s, and subsequently normalized according to the current recorded just before tBHQ application. Note the lower sensitivity of the 1C combination.

View this table: [in this window] [in a new window]   TABLE 2 Percentage of inhibition of tBHQ on Ba2+ currents recorded from oocytes expressing 1C or 1E calcium channel subunit with the 2- and 1, 2, 3, or 4 auxiliary subunits

tBHQ Inhibits L-Type ICa Recorded in Differentiated Mouse Neuroblastoma and Rat Glioma Hybrid Cell Line. In an attempt to elucidate whether L-type ICa generated by the 1D subunit gene was sensitive to tBHQ, as suggested by Nelson et al. (1994) in neuroendocrine cells, we used a cell line that expresses the 1D subunit gene on neuronal differentiation, the mouse neuroblastoma and rat glioma hybrid cell line NG108-15 (Kamp et al., 1995). Cell depolarization from a 40-mV HP to 0 mV elicited a sustained inward current that was sensitive to 1 µM nitrendipine in 50% of the cell population (four of eight cells). The nitrendipine-sensitive current amounted to 29 ± 10% of the macroscopic ICa (n = 4; Fig. 7A). In another series of experiments, application of 300 µM tBHQ induced an 84 ± 5% decrease in macroscopic ICa, and addition of 1 µM nitrendipine to tBHQ did not induce a further decrease (n = 10; Fig. 7B).

View larger version (22K): [in this window] [in a new window]   Fig. 7.   tBHQ inhibits the L-type ICa in NG108-15 cells. A, NG108-15 cells possess a nitrendipine-sensitive ICa. Currents were recorded during a depolarization from a 40-mV HP to 0 mV, every 10 s. B, at 300 µM, tBHQ induced a decrease in macroscopic ICa. Application of nitrendipine did not induce a further decrease.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The main finding of this study is that tBHQ, a well characterized inhibitor of SERCA, reversibly blocks neuronal ICa at a range of concentrations overlapping that used to inhibit mobilization of [Ca²⁺]_i. This VGCC block by tBHQ is not typical of other SERCA pump antagonists such as thapsigargin and CPA, and is specific with respect to ICa subtypes. tBHQ specifically inhibited N-, P-, and Q-type VGCCs without affecting the T, L, or R types. This effect appears to be a general feature of the action of tBHQ in the nervous system because tBHQ-induced VGCC inhibition was effective in neurons from both the peripheral (sensory neurons) and central (hippocampal and motor neurons) nervous systems. The effects of tBHQ may be due to direct, specific binding sites on VGCCs as suggested by VGCC 1-subunit expression studies. The specificity of this compound should be taken into account when assessing the physiological roles of endoplasmic reticulum calcium stores in neuronal preparations. It opens up the possibility of developing novel selective VGCC antagonists or broad-specificity intracellular calcium transient inhibitors.

For motoneurons, the IC₅₀ for tBHQ effects was 35 µM and the saturating concentration was in the 100 µM range. The maximal inhibitory effect of tBHQ amounted to 75% of the macroscopic HVA ICa, which suggests some specificity with respect to ICa types. Testing the selective blocking by tBHQ of N-, P-, and Q-type channels and not of T-, L-, and R-type VGCCs requires that the different components of the ICa be clearly distinguished. As previously shown, N-type GVIA-sensitive and P/Q-type AgaIVA-sensitive VGCCs are distinct, nonoverlapping ICa components that differ from the dihydropyridine-sensitive L-type and GVIA-, AgaIVA-, dihydropyridine-insensitive R-type current components, both in the DRG and motoneurons (Diochot et al., 1995; Scamps et al., 1998). Based on this pharmacological dissection of the components of macroscopic ICa (GVIA, AgaIVA, and nitrendipine for the N, P/Q, and L types, respectively) and on the differential expression of channel types in the preparations (N, L, P, and R types in motoneurons; L, Q, and R types in hippocampal neurons; and T, N, L, P, Q, and R types in DRG neurons), we demonstrated that a supramaximal dose of tBHQ specifically inhibited N-, P-, and Q-type ICa without affecting other VGCC subtypes. In addition, the N- and P/Q-type ICa displayed similar sensitivity over the range of tBHQ inhibition. However, it should be noted that in DRG neurons, the GVIA induced a further 10% decrease in ICa amplitude in the presence of a supramaximal inhibitory dose of tBHQ, an effect not found in motoneurons (Figs. 3 and 4). Two hypotheses may be put forward to explain these pharmacological differences. First, this effect may be related to a nonspecific effect of GVIA (at this concentration) toward non-N-type ICa in DRG neurons (i.e., T-, L-, and R-type ICa). It has been consistently reported that the L- and T-type ICa were partially inhibited by GVIA in various cell types, including DRG neurons (Kasai et al., 1987; McCleskey et al., 1987). Second, pharmacological differences between N-type ICa recorded in mouse DRG and rat motoneurons also may exist as a result of either interspecies variations in the amino acid sequence of the pore-forming 1B subunit of the N-type channels or differential expression of specific splice variants of the 1B subunit in DRG and motoneurons (Lu and Dunlap, 1999). The low voltage-activated T-type ICa, which has been shown to be clearly different from the R-type current and to be more frequently found in DRG neurons than in motoneurons and hippocampal neurons (Hilaire et al., 1997), also was insensitive to tBHQ.

Interestingly, we observed that tBHQ did not inhibit the neuronal L-type ICa, whereas it inhibited non-neuronal L-type ICa (i.e., in neuroendocrine cells; Nelson et al., 1994), probably in smooth muscle cells (Philippe et al., 1995) and L-type ICa in differentiated NG108-15 cells (present study). Moreover, the poor sensitivity of the neuronal 1C subunit to tBHQ in the oocyte expression system appears to be consistent with the inefficacy of this drug in the various neuronal preparations tested. The difference between neuronal and non-neuronal cells may be due to L-type ICa diversity. It has been reported that at least four different genes encode the pore-forming 1-subunits responsible for the L-type current: 1S mainly present in skeletal muscle (Tanabe et al., 1987); 1C found in cardiac muscle, smooth muscle (Lory et al., 1991), and neurons (Hell et al., 1993); 1D predominantly found in neurosecretory cells (Chin et al., 1992) but also in neurons, cardiac myocytes, and differentiated NG108-15 cells (Hell et al., 1993; Kamp et al., 1995; Wyatt et al., 1997); and a recently described retina-specific 1F (Strom et al., 1998). Selectivity with respect to the L-type ICa was reported for the calcium channel antagonists (Hockerman et al., 1997; Striessnig et al., 1998). It also has been reported that different splice variants of the 1C gene confer tissue-specific dihydropyridine sensitivity on the channel (Welling et al., 1997) and that alternative splicing of the 1A gene results in channels with different pharmacological properties (Bourinet et al., 1999). Therefore, the differential expression of 1 isoforms (we show that the auxiliary 1, 2, 3, and 4 do not account for the weak 1C inhibition by tBHQ) may confer a specific pharmacological profile, accounting for differences in the effects of tBHQ on different L types of ICa. Overall, these data support the notion that in situ neuronal and non-neuronal L-type VGCCs differ in their interactions with tBHQ.

The 1E subunit appeared clearly to be more sensitive to tBHQ inhibition than the native R-type ICa. We have no experimental data to explain this difference in sensitivity, which is not due to the composition of the auxiliary -subunits. Splice variants could presumably modulate the effects of tBHQ on the 1E subunit, but we cannot rule out the possibility that the R-type ICa that we record in our preparations does not belong to the gene family coding for the 1E subunit.

These data suggest that tBHQ acts directly on the channel. The effect of tBHQ on endogenous N-, P-, and Q-type currents was not an indirect result of depleting intracellular pools because this effect was not shared with thapsigargin and our whole-cell recording conditions do not permit changes of [Ca²⁺]_i. The onset of ICa inhibition by tBHQ was rapid. Its effects were reproduced at the level of the expressed proteins, as seen by injecting the 1A and 1B genes into oocytes. So the site of action of tBHQ on the N-, P-, and Q-type currents was probably the pore-forming protein. However we did not determine whether the inhibitory effects of tBHQ were related to its ability to modify the thiol groups of the pore-forming proteins. This is relevant given that this is the first report showing that tBHQ blocks neuronal ICa at concentrations in the range at which it is active both against the SERCA and in the control of cell redox state.

These data demonstrate for the first time that micromolar concentrations of tBHQ inhibit voltage-gated ICa in neurons, in addition to SERCA. This newly identified site of action could lead to misinterpretation or overestimation of data relating to variation in [Ca²⁺]_i movement and may therefore explain some of the conflicting results in the literature (for review, see Taylor and Broad, 1998). It also opens up the possibility of developing new calcium channel antagonists with more potent and selective activity against either VGCCs or SERCA. Finally, calcium overload is deleterious to cells in pathological conditions such as epilepsy and ischemia. So, the design of mixed calcium channel antagonists that inhibit both calcium influx and [Ca²⁺]_i release may be an important strategy in the treatment of damage to the central nervous system.