The Journal of Neurophysiology Vol. 80 No. 5 November 1998, pp. 2316-2322
Copyright ©1998 by the American Physiological Society

Facilitation of Miniature GABAergic Currents by Ruthenium Red in Neonatal Rat Hippocampal Neurons

Marina Sciancalepore, Nataa Savi, János Györi, and Enrico Cherubini

Neuroscience Programme and Istituto Nazionale Fisica della Materia Unit, International School for Advanced Studies, 34014 Trieste, Italy

	ABSTRACT

Abstract Introduction Methods Results Discussion References

Sciancalepore, Marina, Nataa Savi, János Györi, and Enrico Cherubini. Facilitation of miniature GABAergic currents by ruthenium red in neonatal rat hippocampal neurons. J. Neurophysiol. 80: 2316-2322, 1998. The whole cell configuration of the patch-clamp technique was used to study the modulation -aminobutyric acid (GABA)-mediated postsynaptic currents by ruthenium red in CA3 hippocampal neurons in slices obtained from postnatal (P) days P6-P10 old rats. In the presence of kynurenic acid (1 mM), ruthenium red (100 µM) completely blocked stimulus-elicited GABA-mediated postsynaptic currents and reduced by 50% the amplitude of the spontaneous ones. Ruthenium red (100 µM) increased the frequency but not the amplitude of miniature GABAergic currents recorded in the presence of tetrodotoxin (1 µM) and kynurenic acid (1 mM), an effect that was prevented by heparin (100 µM). Ruthenium red did not modify the kinetics of miniature postsynaptic currents and the currents induced by exogenous application of GABA (10 µM) in the presence of tetrodotoxin, suggesting that its action was presynaptic in origin. The effects of ruthenium red on quantal GABA release was independent of external calcium. In a nominally Ca²⁺-free solution the potentiating effect induced by this polyvalent cation on miniature postsynaptic currents was still present. Intracellular calcium stores were not involved in ruthenium red action, because this polyvalent cation was able to facilitate miniature currents also in the presence of thapsigargin (10-20 µM). These results indicate that ruthenium red has a dual action on GABA release from GABAergic interneurons: it reduces the amplitude of spontaneous events and increases the frequency of miniature currents. The former effect is calcium-dependent, whereas the latter is calcium independent.

	INTRODUCTION

Abstract Introduction Methods Results Discussion References

Neuronal communication takes place at synaptic terminals by means of neurotransmitter release. Neurotransmitters are stored in synaptic vesicles and released in packets or "quanta," each of them containing thousands of transmitter molecules. This phenomenon is triggered by a rise in [Ca²⁺]_i (Katz and Miledi 1967). Calcium enters into the cell through voltage-gated calcium channels, clustered in specialized area of presynaptic plasma membrane, called active zones, where synaptic vesicles are docked (Simon and Llinás 1985). Transmitter release can be also modulated by substances that affect internal calcium stores (Kostyuk and Verkhratsky 1994) or by factors that influence the release machinery downstream from calcium influx through voltage-dependent calcium channels and independently of intracellular calcium stores, probably acting on proteins that are directly involved in exocytosis. One of these factors is the polyvalent cation ruthenium red (RR) (Trudeau et al. 1996a,b). In early studies on the neuromuscular junction, it has been suggested that the increased rate of spontaneous miniature end-plate potentials by RR is due to a block of mitochondrial calcium uptake (Alnaes and Rahamimoff 1975). To exert this action, RR should enter into the cell. However, this compound seems to be impermeable to intact cell membranes (Korte and Rosenbluth 1982; Tani and Ametani 1971). Although a recent study advances the hypothesis that RR might bind to an external site of the presynaptic membranes (Trudeau et al. 1996a), the mechanism of action of this drug is still unknown. Here we report that, in neonatal rat hippocampal slices, RR blocks the evoked while increases the quantal release of -aminobutyric acid (GABA) from presynaptic nerve terminals in a way that is independent of external calcium.

	METHODS

Abstract Introduction Methods Results Discussion References

Slice preparation

Transverse hippocampal slices were prepared from hippocampi obtained from 6- to 10-day-old Wistar rats according to the methods described by Edwards et al. (1989). Rats were decapitated under anesthesia (5% urethan ip), and their brains were rapidly removed and placed in ice-cold artificial cerebrospinal fluid (ACSF) of the following composition (in mM): 126 NaCl, 3.5 KCl, 2 CaCl₂, 1.2 NaH₂PO₄, 1.3 MgCl₂, 25 NaHCO₃, and 11 glucose, gassed with 95% O₂-5% CO₂ (pH 7.3). After bisecting the brain, the tissue was immersed in low temperature (2-4°C), oxygenated ACSF solution. Slices (250 µm thick) were cut with a vibrating microslicer (Vibracut, FTB, Weinheim, Germany) and incubated at 32°C for 1 h. Individual slices were transferred to a recording chamber where they were continuously superfused at room temperature (22-24°C) with oxygenated ACSF at a rate of 3 ml/min. In some experiments, slices were superfused with a nominally Ca²⁺-free solution in which Ca²⁺ was replaced by Mg²⁺.

Drugs

Drugs were dissolved in ACSF and applied by superfusion in the bath. In some experiments, GABA (10 µM) was dissolved in ACSF and applied by pressure (usually 15-20 PSI for 10 s) from a fine pipette (4-5 M) using a Picospritzer II (General Valve, Fairfield, NJ). The tip of the pipette was placed ~100 µm away from the recording cell. Tetrodotoxin (TTX), RR, Na₂-ATP, and heparin were purchased from Sigma (St. Louis, MO), and thapsigargin from Alomone labs (Israel). Thapsigargin was dissolved in dimethylsulfoxide (DMSO, 0.001%). Control experiments were performed using DMSO alone; at the concentration used, it did not affect the frequency or amplitude of miniature postsynaptic currents (mPSCs).

Patch-clamp recordings

Spontaneous or evoked GABA-mediated synaptic currents were routinely recorded from CA3 pyramidal neurons, held at 70 mV, in the presence of kynurenic acid (1 mM) to block ionotropic glutamatergic currents, using the whole cell configuration of the patch-clamp technique. Patch pipettes had a resistance of 2-4 M when filled with an intracellular solution containing (in mM) 100 CsCl, 10 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), 1 CaCl₂, 1 MgCl₂, and 3-tetracesium bis(o-aminophenoxy) ethane-N,N,N',N'-tetraacetic acid (BAPTA, 10 mM). Na₂ATP (2 mM) was routinely added to minimize rundown of spontaneous GABAergic currents (Chen et al. 1990). Miniature GABAergic currents were routinely recorded in the presence of TTX (1 µM) to block voltage-gated sodium channels and kynurenic acid (1 mM) to block ionotropic glutamate receptors. Nominally Ca²⁺-free solutions were obtained, substituting Ca²⁺ with Mg²⁺. Membrane currents were recorded with a standard patch-clamp amplifier (EPC-7, List Medical Instruments) after optimizing capacitance and series resistance compensation. Typical values for series resistance ranged between 10 and 15 M; recordings were rejected if series resistance values changed more than 15% during the experiment. The membrane input resistance was estimated in voltage-clamp recordings by analyzing current transients during 10 mV, 100 ms long voltage pulses. GABAergic currents were evoked in the presence of kynurenic acid at the frequency of 0.05 Hz (4- to 10-V intensity, 40 µs duration) by a stimulating electrode (patch pipette filled with standard ACSF), positioned 100-200 µm from the recorded cell.

Data acquisition and analysis

Continuous data of spontaneous currents were stored on a magnetic tape and transferred to a computer after digitization with a A/D converter (Digidata 1200). Currents were sampled every 0.2 ms and filtered at 2 kHz. Data acquisition was done using pClamp software from Axon Instruments. The amplitude and frequency of mPSCs were analyzed with the use of a peak detector program with an adjustable threshold, set at 6-10 pA and kept constant for a given experiment. Decay time constants of mPSCs were calculated by performing a least-squares fitting of experimental records with a single exponential function. Rise times were measured as the time required for the current to reach from 10 to 90% of its peak amplitude. Unless otherwise stated, data are expressed as means ± SD. Amplitudes and interevent intervals of spontaneous synaptic currents in control and in test condition were compared with the use of the Kolmogorov-Smirnov test with the criterion of P < 0.05.

Due to the large variability observed from cell to cell, the values of amplitude and frequency of the currents are expressed as I_t/I_c and F_t/F_c, where I_t and F_t are the amplitude and frequency in test condition and I_c and F_c are the amplitude and frequency in control condition, respectively. Mean ratios were calculated from individual test/control ratios in each cell.

	RESULTS

Abstract Introduction Methods Results Discussion References

Experiments were performed from 50 CA3 pyramidal cells. Dialysis of the neuron with Cs⁺ usually resulted in stabilization of the cell membrane potential close to 0 mV. In control conditions, membrane input resistance, measured 1 min after breaking into the whole cell configuration was 255 ± 23 (SD) M. On average, neurons could be recorded in the whole cell configuration for at least 1 h without deterioration.

RR blocked evoked GABA_A-mediated postsynaptic currents

Focal stimulation of GABAergic fibers in the vicinity of the patched cell from a holding potential of 70 mV, delivered in the presence of kynurenic acid (1 mM), evoked a postsynaptic current (mean amplitude 315 ± 81 pA, n = 6) that was GABA_A-mediated because it was reversibly blocked by bicuculline (10 µM). Bath application of RR (100 µM) completely blocked the evoked GABAergic postsynaptic currents without any apparent change in cell input resistance or holding current (n = 6, see Fig. 1). The effect was rapid in onset (2-3 min) and difficult to reverse (at least after 30-40 min wash). The effect of RR on the evoked GABA-mediated postsynaptic currents was associated with 50% decrease in the amplitude of action potential-dependent spontaneous currents.

View larger version (8K): [in this window] [in a new window]   FIG. 1. Ruthenium red (RR) blocks stimulus-elicited GABAergic currents. Average of 10 consecutive stimulus-elicited -aminobutyric acid (GABA)-mediated postsynaptic currents before and after application of RR. In this and the following figures, the holding potential was 70 mV.

RR increased the frequency but not the amplitude of miniature GABAergic currents

In our standard recording conditions with a high internal Cl solution (E_Cl near 0) and in the presence of kynurenic acid (1 mM) and TTX (1 µM) in the bath solution, inwardly directed, spontaneous synaptic currents of variable amplitude (from 14 to 54 pA, on average 28 ± 9 pA, n = 50) and frequency (from 0.09 to 2.4 Hz, on average 0.5 ± 0.4 Hz, n = 50) were observed from hippocampal pyramidal cells, voltage clamped at 70 mV. These currents were reversibly blocked by bicuculline (10 µM), and therefore they reflected the input of GABAergic interneurons to principal cells.

Bath application of RR (100 µM) markedly increased the frequency but not the amplitude of GABA-mediated mPSCs in all cells tested (n = 12). At lower concentrations (50 µM), the effect of RR was often absent or less consistent, and therefore in the majority of the experiments, a concentration of 100 µM RR was used. The effect was rapid in onset (1.5-2 min), reached a steady-state level within 5 min, and was irreversible (at least after 30-40 min wash, Fig. 2). No changes in membrane input resistance or holding current were observed. On average, in 12 cells a 4 ± 2-fold increase in frequency of GABA-mediated mPSCs was obtained (P < 0.001). In contrast, no significant (P > 0.09) changes in amplitude of miniature GABAergic currents were detected during RR application (Fig. 2). An increase in the baseline noise was observed after RR application. Because this effect was blocked by bicuculline, it was thought to be due to the increase in frequency of small-amplitude mPSCs difficult to measure with our detection system. On average, the amplitude of GABA-mediated PSCs was 32 ± 14 pA in control solution and 30 ± 12 pA in the presence of RR (amplitude ratio of RR over control was 0.9 ± 0.1, see Fig. 5). A representative example of cumulative frequency and amplitude distribution of GABAergic mPSCs in the absence or presence of RR is shown in Fig. 2B. Kinetics parameters of miniature GABAergic currents were not altered by RR application. On average, in 10 neurons, rise time values were 2.1 ± 0.1 ms and 2. ± 0.2 ms before and after RR application, respectively. Decay time constants were 30 ± 0.6 ms and 31 ± 0.6 ms before and after RR, respectively (Fig. 2C).

View larger version (36K): [in this window] [in a new window]   FIG. 2. RR enhances the frequency of GABA-mediated miniature postsynaptic currents. A: samples of continuous recordings of miniature GABAergic currents before (control) and during superfusion of RR. B: cumulative distributions of interevent intervals and amplitude of miniature GABAergic events (shown in A) before and after application of RR. The sampling time was the same for both experimental conditions (480 s). C: average of 40 GABA-mediated miniature postsynaptic currents (mPSCs) recorded in 1 representative cell before and during application of RR. The synaptic currents were normalized and aligned with their peaks. The continuous line superimposed on the decay time course of the averaged trace represents a single exponential (decay time constant = 30 ms in control and 32 ms in RR; rise time = 1.9 ms in control and 2 ms in RR).

View larger version (38K): [in this window] [in a new window]   FIG. 5. Pooled data of the effects of RR in control conditions (RR, n = 12), in nominally Ca2+-free solution (0 Ca2+, n = 4) and in thapsigargin (th, n = 5) on the frequency and amplitude of mPSCs. Each column represents the mean ratio (test over control) of event frequency and amplitude. Bars represent SD. *P < 0.05.

The effect of RR on the frequency but not on the amplitude or kinetics of spontaneous miniature GABA-mediated postsynaptic currents suggests a presynaptic site of action. To further investigate this point, the effects of RR on the currents evoked by direct application of GABA to the postsynaptic cells were tested. Pressure application of GABA (10 µM for 10 s), from a holding potential of 70 mV, induced inward currents that were not affected by RR (mean peak amplitudes were 1.4 ± 0.4 nA and 1.3 ± 0.4 nA in the absence and presence of RR, respectively, n = 4, not shown).

Effect of RR on the frequency of miniature GABAergic currents was independent of external calcium

To see whether external calcium was necessary for the potentiating effect of RR on miniature GABA-mediated synaptic currents, the experiments were performed on neurons superfused with a nominally Ca²⁺-free solution. In Ca²⁺-free solutions, no significant changes in amplitude and frequency of mPSCs were found comparing with the control condition. The mean amplitude of GABA-mediated postsynaptic currents varied from 32 ± 4 pA (in the presence of external Ca²⁺) to 28 ± 4 pA in Ca²⁺-free solution, whereas the mean frequency changed from 0.9 ± 0.9 Hz to 0.7 ± 0.6 Hz (n = 7). Bath application of RR, in the absence of external Ca²⁺ induced a significant (P < 0.001) increase in the frequency of synaptic events without change in amplitude distribution (mean frequency and amplitude ratios were 4.6 ± 2.3 and 1 ± 0.2, respectively, n = 4, Figs. 3 and 5).

View larger version (36K): [in this window] [in a new window]   FIG. 3. Facilitatory effect of RR on miniature synaptic currents was independent of external calcium. A: samples of continuous recordings of miniature GABAergic currents in control, in Ca2+-free solution and in Ca2+-free plus RR. B: cumulative distributions of interevent intervals and amplitude of miniature GABAergic events recorded in Ca2+-free solutions (shown in A), before and during application of RR. The sampling time was the same in both experimental conditions (360 s).

Effect of RR on the frequency of miniature GABAergic currents was independent of calcium released from internal stores

It has been demonstrated that RR can affect ryanodine-sensitive Ca²⁺ release channels (Tsien and Tsien 1990) and block mitochondria calcium uniporter (Moore 1971). To test at least for one of these possibilities, RR was superfused in the presence of thapsigargin, a blocker of Ca²⁺-ATPase, known to deplete Ca²⁺ from intracellular stores. To avoid contamination with calcium-induced calcium release mechanisms, RR and thapsigargin were applied in a Ca²⁺-free medium. Thapsigargin (10-20 µM) was tested in five cells, where it did not change frequency and amplitude of miniature events (mean frequency and amplitude ratio were 0.9 ± 0.3 and 0.9 ± 0.1, respectively). After 5 min of incubation in thapsigargin, RR increased the frequency, but not the amplitude of miniature GABAergic events (mean frequency and amplitude ratios were 3.3 ± 2.6 and 1 ± 0.1, respectively, n = 5, Figs. 4 and 5). These results suggest that the potentiation of miniature events by RR is a calcium-independent process.

View larger version (44K): [in this window] [in a new window]   FIG. 4. Effect of RR was independent of calcium released from internal stores. A: samples of continuous recordings of miniature GABAergic currents recorded in thapsigargin and thapsigargin plus RR. B: cumulative distributions of interevent intervals and amplitude of miniature GABAergic events (shown in A). Sampling time was the same for both experimental conditions (360 s).

Heparin antagonizes the action of RR on miniature GABA-mediated currents

In agreement with a previous study on miniature glutamatergic currents in cultured hippocampal neurons (Trudeau et al. 1996a), also the potentiating effect of RR on GABA-mediated mPSCs was antagonized by heparin (100 µM). Heparin, applied immediately after washing RR, was able to block the potentiating effect of RR. In three cells, the mean frequency of miniature currents varied from 0.7 ± 0.2 Hz in control condition to 2.2 ± 0.3 Hz in RR and 0.7 ± 0.1 Hz after heparin superfusion. These data suggest that heparin and RR interact at the external site of the presynaptic terminal (see also Trudeau et al. 1996a).

	DISCUSSION

Abstract Introduction Methods Results Discussion References

The main conclusions that can be drawn from the present experiments are that in neonatal hippocampal neurons, RR 1) blocks stimulus-elicited GABA-mediated postsynaptic currents, 2) depresses the amplitude of action potential-dependent spontaneous events, and 3) facilitates the quantal release of GABA in a calcium-independent way.

RR blocked evoked GABA_A-mediated postsynaptic currents

As previously observed in the neuromuscular junction (Alnaes and Rahamimoff 1975), RR was found to be able to completely block the stimulus-elicited GABA-mediated postsynaptic currents also in central neurons. This effect probably reflects the inhibitory action of RR on voltage-dependent calcium currents (Gomis et al. 1994; Trudeau et al. 1996a). Further evidence in favor of a calcium-dependent effect is provided by the observation that on the same cells in which evoked GABAergic currents were blocked by RR, a clear reduction in the amplitude of spontaneous currents was found after RR application. The larger events were lost, and the smaller currents, supposed to be quantal in origin, increased in their number. This indicates that RR has a dual effect on synaptic currents: it depresses calcium-dependent release while it facilitates the calcium-independent one. The lack of changes in frequency of action potential-dependent spontaneous events with RR can be explained by the fact that a decrease in frequency of the calcium-dependent component was counterbalanced by a concomitant increase in frequency of calcium-independent events.

Potentiating effect of RR on miniature GABAergic currents was presynaptic in origin

This work extends previous observations on the effects of RR on glutamate and GABA release from hippocampal neurons in culture (Trudeau et al. 1996a,b). In comparison to the culture conditions, in brain slices, due to the reduction of the extracellular space and differences in diffusion properties, higher concentrations of RR have to be used to get similar responses (see Haas et al. 1993). As in previous studies, the effect of RR observed in the present work is presynaptic in origin, as suggested by the following points: 1) RR induced only an increase in frequency and not in amplitude of miniature events; 2) the effects of RR on miniature currents were not associated with any change in membrane input conductance or holding current; 3) the kinetics of spontaneous GABAergic responses was unaffected by RR; and 4) RR did not modify the responses to exogenously applied GABA in the presence of TTX.

Effect of RR on miniature events was calcium independent

As suggested also by Trudeau et al. (1996a,b), also in our case the effects of RR were calcium independent. Thus the increase in frequency of miniature GABAergic currents persisted when RR was applied in a nominally Ca²⁺-free medium. In these conditions the observed increased efficacy of RR might be attributed to a competition for binding sites between RR and Ca²⁺ on presynaptic membranes (Trudeau et al. 1996a).

The experiments with thapsigargin allow us to exclude the involvement of intracellular calcium stores in RR action because the increase in frequency of miniature events still occurred when Ca²⁺-ATPases were blocked. However, in cells loaded with fluorescent calcium-sensitive dyes, intracellular application (Llano et al. 1994; Markram et al. 1995) or extracellular superfusion (Trudeau et al. 1996a; unpublished observations) of RR failed to induce any change in resting calcium either at somatic (Llano et al. 1994; Trudeau et al. 1996a) or dendritic (Markram et al. 1995) level.

Recently, on the basis of the observation that RR-evoked transmitter release was blocked by tetanus toxin, it has been proposed that its action occurs at a late step in the vesicle cycle, downstream of vesicle docking and calcium influx, probably interfering with vesicle proteins such as synaptobrevin (Trudeau et al. 1996b). Our observation that in CNS RR depresses calcium-dependent and facilitates calcium-independent release makes this polyvalent cation an interesting assay to test secretory processes.

	ACKNOWLEDGEMENTS

  This work was partially supported by a grant from Consiglio Nazionale delle Ricerche (CNR), from Ministero Universita' Ricerca Scientifica e Tecnologica (MURST), and Human Capital and Mobility Program network from the European Union.

	FOOTNOTES

  Address for reprint requests: M. Sciancalepore, Biophysics Sector, International School for Advanced Studies, Via Beirut 2-4, 34014 Trieste, Italy.

  Received 26 March 1998; accepted in final form 16 July 1998.

	REFERENCES

Abstract Introduction Methods Results Discussion References

• ALNAES, E., RAHAMIMOFF, R. On the role of mitochondria in transmitter release from motor nerve terminals. J. Physiol. (Lond.) 248: 285-306, 1975.[Abstract/Free Full Text]
• CHEN, Q. X., STELZER, A., KAY, A. R., WONG, R.K.S. GABA_A receptor function is regulated by phosphorylation in acutely dissociated guinea-pig hippocampal neurons. J. Physiol. (Lond.) 420: 207-221, 1990.[Abstract/Free Full Text]
• EDWARDS, F. A., KONNERTH, A., SAKMANN, B., TAKAHASHI, T. A thin slice preparation for patch-clamp recordings from neurons of the mammalian central nervous system. Pflügers Arch. 414: 600-612, 1989.[Medline]
• GOMIS, A., GUTIERREZ, L. M., SALA, F., VINIEGRA, S., REIG, J. A. Ruthenium red inhibits selectively chromaffin cell calcium channels. Biochem. Pharmacol. 47: 225-231, 1994.[Medline]
• HAAS, H. L., HÄRTTER, S., HERMES, B., HIEMKE, C. Bath perfusion of slices with drugs. In: Practical Electrophysiological Methods: A Guide for In Vitro Studies in Vertebrate Neurobiology, edited by H. Kettenmann, and R. Grantyn New York: Wiley-Liss, 1993, p. 129-131.
• KATZ, B., MILEDI, R. The timing of calcium action during neuromuscular transmission. J. Physiol. (Lond.) 189: 535-544, 1967.[Abstract/Free Full Text]
• KORTE, G. E., ROSENBLUTH, J. Anionic sites on the surface of frog ependymal astrocytes and mouse ependymal cells. Anat. Rec. 204: 95-100, 1982.[Medline]
• KOSTYUK, P., VERKHRATSKY, A. Calcium stores in neurons and glia. Neuroscience 63: 381-404, 1994.[Medline]
• LLANO, I., DIPOLO, R., MARTY, A. Calcium-induced calcium release in cerebellar Purkinje cells. Neuron 12: 663-673, 1994.[Medline]
• MARKRAM, H., HELM, P. J., SAKMANN, B. Dendritic calcium transients evoked by single back-propagating action potentials in rat neocortical pyramidal neurons. J. Physiol. (Lond.) 485: 1-20, 1995.[Medline]
• MOORE, C. L. Specific inhibition of mitochondrial Ca⁺⁺ transport by ruthenium red. Biochem. Biophys. Res. Commun. 42: 298-305, 1971.[Medline]
• SIMON, S. M., LLINÁS, R. R. Compartmentalization of the submembrane calcium activity during calcium influx and its significance in transmitter release. Biophys. J. 48: 485-498, 1985.[Abstract/Free Full Text]
• TANI, E., AMETANI, T. Extracellular distribution of ruthenium red-positive substances in cerebral cortex. J. Ultrastruct. Rec. 34: 1-14, 1971.
• TRUDEAU, L., DOYLE, R. T., EMERY, D. G., HAYDON, P. G. Calcium-independent activation of the secretory apparatus by ruthenium red in hippocampal neurons: a new tool to assess modulation of presynaptic function. J. Neurosci. 16: 46-54, 1996a.[Abstract/Free Full Text]
• TRUDEAU, L. E., DOYLE, R. T., EMERY, D. G., HAYDON, P. G. Direct modulation of the secretory machinery underlies PKA-dependent synaptic facilitation in hippocampal neurons. Neuron 17: 789-797, 1996b.[Medline]
• TSIEN, R. W., TSIEN, R. Y. Calcium channels, stores, and oscillations. Annu. Rev. Cell Biol. 6: 715-760, 1990.

0022-3077/98 $5.00 Copyright ©1998 The American Physiological Society