DUAL EFFECT OF Zn 2 ON MULTIPLE TYPES OF VOLTAGE- DEPENDENT Ca 2 CURRENTS IN RAT PALAEOCORTICAL NEURONS

Transcription

1 Neuroscience 117 (2003) DUAL EFFECT OF Zn 2 ON MULTIPLE TYPES OF VOLTAGE- DEPENDENT Ca 2 CURRENTS IN RAT PALAEOCORTICAL NEURONS J. MAGISTRETTI,* L. CASTELLI, V. TAGLIETTI AND F. TANZI Dipartimento di Scienze Fisiologiche-Farmacologiche Cellulari-Molecolari, Sezione di Fisiologia Generale e Biofisica Cellulare, Università degli Studi di Pavia, Via Forlanini 6, Pavia, Italy were used as depolarizing stimuli. We conclude that Zn 2 exerts a dual action on multiple types of voltage-gated Ca 2 channels, causing a blocking effect and altering the speed at which channels are delivered to conducting states, with mechanism(s) that could be distinct IBRO. Published by Elsevier Science Ltd. All rights reserved. Abstract The effects of Zn 2 were evaluated on high-voltage-activated Ca 2 currents expressed by pyramidal neurons acutely dissociated from rat piriform cortex. Whole-cell, patch-clamp experiments were carried out using Ba 2 (5 mm) as the charge carrier. Zn 2 blocked total high-voltage-activated Ba 2 currents with an IC 50 of approximately 21 M. In addition, after application of non-saturating Zn 2 concentrations, residual currents activated with substantially slower kinetics than control Ba 2 currents. Both of the above-mentioned effects of Zn 2 were also observed in high-voltageactivated currents recorded in the presence of nearly-physiological concentrations of extracellular Ca 2 (1 and 2 mm) rather than Ba 2. Under the latter conditions, 30 M Zn 2 inhibited high-voltage-activated currents somewhat less than observed in extracellular Ba 2 ( 47% and 41%, respectively, vs. 59%), but slowed Ca 2 -current activation to very similar degrees. All of the pharmacological components in which Ba 2 currents could be dissected (L-, N-, P/Q-, and R-type) were inhibited by Zn 2, the percentage of current blocked by 30 MZn 2 ranging from 34 to 57%. Moreover, the activation kinetics of all pharmacological Ba 2 current components were slowed by Zn 2. Hence, the lower activation speed observed in residual Ba 2 currents after Zn 2 block is due to a true slowing of macroscopic Ca 2 -current activation kinetics and not to the preferential inhibition of a fast-activating current component. The inhibitory effect of Zn 2 on Ba 2 current amplitude was voltage-independent over the whole voltage range explored ( 60 to 30 mv), hence the Zn 2 - dependent decrease of Ba 2 current activation speed is not the consequence of a voltage- and time-dependent relief from block. Zn 2 also caused a slight, but significant, reduction of Ba 2 current deactivation speed upon repolarization, which is further evidence against a depolarization-dependent unblocking mechanism. Finally, the slowing effect of Zn 2 on Ca 2 -channel activation kinetics was found to result in a significant, extra reduction of Ba 2 current amplitude when action-potential-like waveforms, rather than step pulses, *Corresponding author. Tel: ; ; fax: address: jmlab1@unipv.it (J. Magistretti). Abbreviations: -AgaTx IVA, -agatoxin IVA; -CTx GVIA, -conotoxin GVIA; -CTx MVIIC, -conotoxin MVIIC; APWs, action-potentiallike waveforms; EGTA, ethylenebis(oxyethylenenitrilo)tetraacetic acid; HEPES, N-[2-hydroxyethyl]piperazine-N'-[2-ethanesulphonic acid]; HVA, high-voltage activated; I Ba, barium current; I Ca, calcium current; NMDA, N-methyl-D-aspartate; PC, piriform cortex; R s, series resistance; VDCC, voltage-dependent calcium currents /03$ IBRO. Published by Elsevier Science Ltd. All rights reserved. doi: /s (02) Key words: calcium channels, zinc, block, activation kinetics, channel gating, patch clamp. The biological importance of zinc ions (Zn 2 ) is well known (Vallee and Falchuk, 1993). Besides being a constitutive part of several metal-dependent enzymes, in the CNS Zn 2 is contained, at high concentrations and in free form ( chelatable Zn 2 ), in the synaptic vesicles of many synapses (Frederickson, 1989; Cuajungco and Lees, 1997), especially at excitatory glutamatergic synapses. The location of this chelatable Zn 2 is highly specific and differential in different regions and structures of the CNS (see Cuajungco and Lees, 1997). For instance, strong reactivity for chelatable, synaptic Zn 2 is present in the synaptic terminations sent by hippocampal mossy fibers to CA3 pyramidal neurons (Haug, 1967; Frederickson et al., 1983), as well as in specific laminar and regional subdivisions of allocortical structures such as the piriform, entorhinal, and perirhinal cortices (Haug, 1976; Schwerdtfeger et al., 1985; Slomianka, 1992). Major interest as to the possible physiological roles of vesicular Zn 2 has been risen by the finding that Zn 2 can modulate synaptic excitatory and inhibitory receptors (Smart et al., 1994), especially glutamatergic N-methyl-D-aspartate (NMDA) receptors (Peters et al., 1987; Westbrook and Mayer, 1987). Importantly, it has also been demonstrated that synaptic Zn 2 can be exocytotically released as a consequence of synaptic activity, and can reach remarkably high concentrations in the extracellular space during periods of intense activity (Assaf and Chung, 1984; Howell et al., 1984; Aniksztejn et al., 1987). Zn 2 levels of 300 mol/l have been estimated to be reached in the extracellular space secondarily to epileptiform (or ictal-like) discharges (Assaf and Chung, 1984). More recently, it has been shown that Zn 2 released at the hippocampal mossy fiber-ca3 pyramidal neuron synapse acts as a physiological modulator on postsynaptic NMDA receptors by exerting a significant inhibitory action on the corresponding synaptic currents (Vogt et al., 2000). Such inhibitory action appeared to be both tonic (being present under basal conditions, namely in the absence of electrical stimulation of the afferent pathway), and inducible upon mossy-fiber discharge.

2 250 J. Magistretti et al. / Neuroscience 117 (2003) Therefore, Zn 2, more than other group VIIA-IIB divalent metal cations, appears as a potentially important modulatory factor of neuronal electroresponsiveness. It has been shown that Zn 2, like many other divalent metal cations, can influence several neuronal membrane conductances, including K (Bardoni and Belluzzi, 1994; Easaw et al., 1999), Na (Gilly and Armstrong, 1982; White et al., 1993) and Ca 2 conductances. The blocking effects of cations such as Mn 2,Ni 2, and Cd 2 on specific voltagedependent Ca 2 channels are well known and characterized (see Carbone and Swandulla, 1989). Whereas in the case of the latter ions such effects are purely pharmacological and have been exploited as experimental tools, it is possible that the interactions of Zn 2 with voltage-dependent membrane conductances represent a physiological, endogenous mechanism of modulatory control. Inhibitory effects of Zn 2 on high-threshold voltage-dependent Ca 2 currents (VDCC) have been reported in a few studies on peripheral (Buesselberg et al., 1994) and central (Nam and Hockberger, 1992; Easaw et al., 1999; Kerchner et al., 2000) mammalian neurons. However, the fine properties and mechanisms of the block exerted by Zn 2 on the same currents, as well as the possible existence of additional effects of Zn 2 on Ca 2 -channel gating, have never been investigated. Moreover, multiple pharmacological and molecular subtypes of voltage-gated Ca 2 channels are known to be expressed by central neurons, and each of them could, or have been shown to, subserve different functional roles in various neuronal populations (see Tsien et al., 1995; Catterall, 1998). No data on the effects of Zn 2 on specific Ca 2 -current and -channel subtypes are currently available. The piriform cortex (PC) is a paleocortical structure that represents the primary olfactory area of mammalian brain. PC is known to contain high densities of vesicular Zn 2 (Haug, 1976; Schwerdtfeger et al., 1985), especially at those laminar levels in which the bulk of the synaptic terminals directed to layer-ii pyramidal neurons reach their dendritic targets, namely layers Ib and III. The present study was undertaken to investigate the effects of Zn 2 on the high-voltage-activated (HVA) Ca 2 currents expressed by pyramidal neurons of rat PC. Cell preparation EXPERIMENTAL PROCEDURES Young (P12 P22) Wistar rats of either sex were decapitated, according to a procedure approved by the University of Pavia Ethical Committee and compliant with the national laws on animal research. The brain was quickly extracted under hypothermic conditions, the two hemispheres were separated, and each was cut with a McIlwain tissue chopper into 350- m thick slices, the plane of which was normal to the main axis of the lateral olfactory tract. Layer II of anterior piriform cortex was carefully dissected from each slice under microscopic control. During this operation the slices were submerged in an ice-cold solution (dissection buffer) composed of (in mmol/l): 115 NaCl, 3 KCl, 3 MgCl 2, 0.2 CaCl 2, 20 piperazine-n,n'-bis(2-ethanesulphonic acid) 1.5 Na (PIPES-NA), and 25 D-glucose (ph 7.4 with NaOH, bubbled with pure O 2 ). The tissue fragments obtained were then transferred into a 20-ml stirring flask filled with dissection buffer added with 1 mg/ml pronase (protease type XIV, Sigma, St. Louis, MO, USA) and continuously bubbled with O 2. The flask was submerged in a thermostated bath at C and kept gently stirring for a time variable as a function of the animal s age (4 min for P12 rats up to 7 min for P18 P22 rats). The enzymatic reaction was stopped by removing the solution and by rinsing the tissue with a solution (calcium-free buffer) containing (in mmol/l): NaCl, 3 KCl, 3 MgCl 2, 20 PIPES-Na 3 ethylene glycol-bis ( -aminoethyl ether) N,N,N,N -tetraacetic acid (EGTA), 25 D-glucose, and 2 mg/ml bovine serum albumin (Sigma fraction V) (ph 7.4 with NaOH). The tissue fragments were then placed in a holding chamber filled with continuously oxygenated dissection buffer, and there kept at room temperature for at least 1 h before further dissociation. When needed, tissue fragments were removed from the chamber, resuspended in 2 ml of calcium-free buffer and triturated with a few passages through Pasteur pipettes of progressively decreasing tip diameter. After sedimentation of the undissociated tissue, the supernatant was transferred into the recording chamber, on a concanavaline A (Sigma, type V)-coated, 16-mm diameter round coverslip. The dissociated cells were allowed to settle down for 15 min before starting the recordings. Patch-clamp recordings The recording chamber was mounted on the stage of an Axiovert 100 microscope (Zeiss, Oberkochen, FRG). The cells were observed at 400 magnification. After cell seeding, the chamber was perfused with an oxygenated extracellular solution suitable for isolating Ba 2 currents conducted through Ca 2 channels, containing (in mmol/l): 88 choline-cl, 40 tetraethylammonium-cl (TEA-Cl), 3 KCl, 2 MgCl 2, 5 BaCl 2, 3 CsCl, 10 HEPES, 5 4- aminopyridine, and 25 D-glucose (ph 7.4 with HCl). In some experiments, BaCl 2 was used at lower concentrations (1 or 3 mm) or substituted with CaCl 2 (1, 2, or 5 mm). When BaCl 2 or CaCl 2 were used at less than 5 mmol/l, the osmolarity was corrected with D-glucose. Perfusion rate was about 0.5 ml/min. Patch pipettes were fabricated from thick-wall borosilicate glass capillaries (CEI GC ; Harvard Apparatus, Edenbridge, UK) by means of a Sutter P-87 horizontal puller (Sutter Instruments, Novato, CA, USA). The pipette solution contained (in mmol/l): 78 Cs methanesulphonate, 40 TEA-Cl, 10 HEPES, 10 EGTA, 20 phosphocreatine di-tris salt, 2 adenosine 5'-triphosphate-Mg, 0.2 guanosine 5'-triphosphate-Na, and 20 U/ml creatinephosphokinase (ph adjusted to 7.2 with CsOH). The patch pipettes had a resistance of M when filled with the above solution. Tight seals ( 10 G ) and the whole-cell configuration were obtained according to the standard technique (Hamill et al., 1981). Voltage-clamp recordings of Ba 2 or Ca 2 currents were performed at room temperature (22 C) by means of an EPC7 patch-clamp amplifier (List Electronics, Darmstadt, FRG). After the establishment of the whole-cell condition, voltage-dependent currents were allowed to stabilize for a few minutes before starting data acquisition. Series resistance (R s ) was evaluated on line by canceling the whole-cell capacitive transients evoked by 5-mV voltage square pulses with the amplifier s compensation section, and reading out the corresponding values. R s averaged M (n 145), was always compensated by 50 70%, and was continually monitored during the experiment. Recordings in which R s levels varied with time by more than 2 M were discarded. Voltage protocols were commanded and current signals were acquired with a Pentium personal computer interfaced to an Axon TL-1 interface, using the Clampex program of the pclamp software package (Axon Instruments). In all recordings the general holding potential was 70 mv. Every test voltage protocol was preceded by a 2-s conditioning prepulse at 60 mv. Current signals were filtered at 5 khz, digitized at 100 khz (tail protocols) or 50 khz (other protocols), and on-line leak subtracted via a P/4 protocol.

3 J. Magistretti et al. / Neuroscience 117 (2003) Drug application ZnCl 2 (Sigma) and organic Ca 2 -channel blockers were applied through a local-perfusion system consisting of a multibarrel pipette (diameter at the tip 150 m), each barrel of which was connected to a separate perfusion channel. Upon opening of each channel, controlled by operating on remote-commanded electrovalves (Sirai, Milano, Italy), the drug-containing solution flowed by gravity, thus forming a laminar-flux cone. The tip of the perfusion pipette was positioned in close proximity of the recording site, so that the recorded cell was fully drenched by the laminar-flux cone. The same local-perfusion system was also employed in most experiments in which the effects of Zn 2 were tested in the presence of extracellular Ca 2 or low concentrations of Ba 2. In these cases, the solution used for the general perfusion of cells and for starting the recording contained 5 mm Ba 2, then the control and Zn 2 -added solutions containing Ca 2 or low concentrations of Ba 2 were delivered locally. Concentrated stock solutions of the drugs were prepared, divided in small aliquots, and stored at 20 C. Nifedipine (Sigma) was dissolved in dimethylsulfoxide at 10 mmol/l; -conotoxin GVIA ( -CTx GVIA), -conotoxin MVIIC ( -CTx MVIIC), and -agatoxin IVA ( -AgaTx IVA) (all from Bachem, Bubendorf, Switzerland) were dissolved in pure water at 1, 1, and 0.1 mmol/l, respectively. The aliquots were then diluted to the final concentrations in the recording solution described in the previous paragraph. -CTx MVIIC and -AgaTx IVA aliquots were dissolved in the presence of lysozyme (Sigma; 1 mg/ml), in order to minimize aspecific binding to recipient walls. The solvents and additional substances used for each drug (dimethylsulfoxide and lysozyme) were also added, in the same amounts, to the control solution and the other drug-containing solutions used in the same experiment. Due to the drug s light sensitivity, nifedipine was prepared and stored in the dark, and the perfusion channel containing nifedipine was light shielded. Data analysis Current traces were analyzed by means of the Clampfit program of pclamp Voltage-clamp conditions were evaluated off-line by analyzing the kinetics of capacitive current transients elicited by 10-mV voltage square pulses. Exponential fittings of capacitivetransient decay returned an average time constant of s (n 24). This value coincides with the average time constant of the voltage clamp operating inside the cell in the absence of R s compensation. Activation of R s compensation to the percentages normally achieved (50 70%: see above) allowed for even higher (by 2 3 times) clamp speed. Ba 2 and Ca 2 currents (I Ba s and I Ca s, respectively) were normally re-filtered off-line at 1 2 khz, unless preservation of high-frequency signal components was required (i.e. when tail currents were analyzed). Current amplitude was always measured at the peak of each trace, unless otherwise specified. Data fittings with exponential functions, I A i exp( t/ i ) C, were carried out using Clampfit. Dose-response plots were fitted with a Langmuir adsorption isotherm function, in the form: I Ba (C) I Ba (Zn 2 )]/I Ba (C) [Zn 2 n /(IC 50 n [Zn 2 ] n ), (1) where I Ba (C) and I Ba (Zn 2 ) are the current amplitudes in control and under Zn 2, respectively, IC 50 is the Zn 2 concentration corresponding to 50% I Ba block, and n is the Hill coefficient. Langmuir fittings were performed using Origin 6.0 (MicroCal Software, Northampton, MA, USA). Average values were expressed as mean S.E.M. Statistical significance was evaluated by means of the two-tail Student s t-test for paired or, when not explicitly stated, unpaired data. Differences were considered statistically significant for P Fig. 1. Zn 2 reduces both the amplitude and activation speed of HVA I Ba s in rat piriform-cortex neurons. Voltage-clamp protocol (A) and the resulting HVA Ba 2 currents recorded in a representative neuron in control conditions, in the presence of 30 MZn 2, and after was-out of Zn 2 (B). (C) The current fraction blocked by Zn 2, obtained by subtraction. Calibration bars: 100 pa, 10 ms. (D) Detail of the activation phases of control and Zn 2 -resistant currents, normalized to their peak amplitudes and shown over an expanded time scale (calibration bar: 5 ms). Note that the traces are depicted as separate points to highlight the high temporal resolution of the acquisition. The inset shows normalized Zn 2 -resistant and Zn 2 -sensitive currents superimposed (same time span as in the main panel). RESULTS Voltage-dependent I Ba s and I Ca s were recorded in 145 pyramidal neurons from rat PC layer II. Currents were routinely elicited by delivering 50-ms depolarizing steps starting from 2-s prepulses at 60 mv, which allowed for recording of HVA currents in isolation (see Magistretti and de Curtis, 1998). Unless otherwise explicitly stated, 5 mm Ba 2 was routinely used as the charge carrier. Dual effect of Zn 2 on HVA currents Fig. 1B, shows typical currents recorded in response to test pulses at 0 mv in a representative neuron. The application of 30 M Zn 2 reduced the peak amplitude of total I Ba sby 60% on average. The inhibitory effect of Zn 2 was readily reversible upon washout (Fig. 1B). In addition,

4 252 J. Magistretti et al. / Neuroscience 117 (2003) reached) not far from one ( 0.8). Hence, Zn 2 is a potent blocker of total HVA I Ba s under the experimental conditions here adopted. In order to quantify the effects of Zn 2 on residualcurrent activation kinetics, exponential fittings of I Ba onset were performed at various test potentials. I Ba activation phase could be consistently best fitted by biexponential functions (Fig. 3A2). To obtain a synthetic description of I Ba activation kinetics, fitting parameters were then used to derive the following quantity: act act1 A 1 /(A 1 A 2 ) act2 A 2 /(A 1 A 2 ), (2) Fig. 2. Concentration-dependence of I Ba block by Zn 2. (A) I Ba s recorded in a representative neuron in response to 50-ms depolarizing test pulses at 0 mv in control conditions and during application of increasing concentrations (6 200 M) of Zn 2. Calibration bars: 200 pa, 10 ms. (B) Average concentration-dependence plot of Zn 2 -induced I Ba block. Fractional block was measured considering peak values of control currents and residual currents under Zn 2. Note the logarithmic scale of the x axis. Data points were best fitted with a Langmuir isotherm function (continuous line), which returned the fitting parameters specified inside the panel. n 41 (30 M), 4 (0.6 and 2000 M), or 13 (all the other data points). we observed that the activation phase of residual I Ba s recorded in the presence of the same concentration of Zn 2 was consistently slower than in control currents. This is better illustrated in Fig. 1D, where total and residual currents are shown, over an expanded time scale, normalized to their maximal amplitudes. As a consequence, Zn 2 -sensitive currents, obtained by subtraction (Fig. 1C), appeared to be substantially faster-activating than both control and Zn 2 -resistant currents (Fig. 1D). Similarly to the blocking effect of Zn 2, the activation-speed decrease observed in unblocked currents was also reverted by Zn 2 washout (Fig. 1B). The two above-illustrated effects of Zn 2 on HVA currents were than characterized in more detail. Increasing levels of Zn 2 (6 200 M) were delivered to analyze the concentration dependence of Zn 2 -induced I Ba block (Fig. 2A). Average results are illustrated in Fig. 2B. The plot of I Ba fractional inhibition [measured at the peak of the current-voltage relationship (I V) and considering peak I Ba amplitudes] as a function of Zn 2 concentration could be properly fitted with a Langmuir isotherm (see the Methods, Eq. 1), which returned an IC 50 of 21.4 M and an n coefficient (related to the number of Zn 2 ions that must bind to any Ca 2 channel for the blocking effect to be which represents a lumped, or global, activation time constant (see Magistretti et al., 2001). Fig. 3 illustrates how Zn 2 reduced the current activation speed over the whole voltage range of I Ba activation. act was consistently increased by Zn 2 (Fig. 3B1). This effect was voltage dependent, since the percent act increase in Zn 2 averaged 180% close to I Ba activation threshold ( 30 mv), and reached a steady value of 50% at the most positive test potentials used ( 10 to 30 mv) (Fig. 3B2). The average effects of Zn 2 on the individual fitting parameters ( act1, act2, and relative amplitude coefficients) are illustrated in Fig. 3B3 5. Modifications at all of these parameters charge contributed to the global slowing of residual I Ba activation kinetics carried out by Zn 2 and synthetically described by act. Dependence of Zn 2 effects on the identity and concentration of the permeant ion We then verified whether the two above-described effects of Zn 2 on HVA I Ba s (blocking action and slowing of residual-current activation kinetics) could also be observed in a more physiological situation, namely in the presence of Ca 2, rather than Ba 2, as the permeant ion. When Ca 2 was used at the same concentration as that employed for Ba 2 in the experiments illustrated so far (5 mm), 30 M Zn 2 reduced the amplitude of total HVA currents by % on average (n 4; Fig. 4A3 and B). This percentage of block is about twofold lower than that measured using 5 mm extracellular Ba 2 ( %, n 41; P 10 4 ). Since a difference was noticed in the blocking potency of Zn 2 using Ca 2 vs. Ba 2 as the charge carrier, we hypothesized that Zn 2 produces its blocking effect by binding to a site normally occupied by the permeant ion, thus competing with it. In this case, the competition phenomenon would be influenced by the identity and concentration of the permeant ion. To verify this hypothesis, further experiments were performed using 2 mm or 1 mm extracellular Ca 2 (Fig. 4A1 and A2). In 2 and 1 mm Ca 2, 30 MZn 2 decreased I Ca amplitude by % (n 8) and % (n 5), respectively. These percentages of inhibitions are significantly higher than those observed in 5 mm Ca 2 (P 0.05 and 0.001, respectively). Assuming that Zn 2 competes with Ca 2 for a single binding site within Ca 2 channels to produce its blocking effect, the

5 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 3. Analysis of the effects of Zn 2 on I Ba activation kinetics. (A) Detail of the activation phases of I Ba s recorded in a representative neuron in response to step depolarizations at various voltage levels (see the voltage protocol in A1), before and during application of 30 M Zn 2. All currents are Cd 2 -subtracted, and have been normalized to their peak amplitudes. The slower-activating currents are those recorded in the presence of Zn 2. Enhanced lines are biexponential best fittings to current activation phases. The values of the lumped activation time constant ( act ; see text for details) are also shown. Calibration bar: 5 ms. (B) Average plots of the voltage dependence of kinetic parameters returned by biexponential fittings of I Ba activation phase: lumped activation time constant, act (B1), individual fast and slow activation time constants, act1 and act2 (B3 and B4, respectively), relative amplitude coefficient of the slow exponential component (B5). Empty symbols, continuous lines: control; filled symbols, dotted lines: 30 M Zn 2. ***, P 0.001; **, P 0.01; *, P 0.05 (t test for paired data). Panel B2 shows the average values of percent act increase caused by Zn 2 at the different test potentials. n 22 throughout.

6 254 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 4. Effects of Zn 2 on Ca 2 -current amplitude and activation speed in the presence of different concentrations of Ca 2 or Ba 2 as the charge carrier. (A) Current traces recorded at the peak of the I V relationship in 1,2,or5mMextracellular Ca 2 (A1 3, respectively) or 1 mm extracellular Ba 2 (A4) in four different representative neurons, either in control conditions or in the presence of 30 M Zn 2. The upper traces of each sub-panel are original currents, the lower traces are the same currents normalized to their peak amplitudes and shown over an expanded time scale. Calibration bars for original currents: 100 pa (A1), 83 pa (A2), 115 pa (A3), 54 pa (A4); 10 ms. x-axis scale for normalized traces: 5 ms. (B) Bar diagram of average, percent current-amplitude reduction (open bars) and percent act increase (hatched bars) induced by 30 M Zn 2 in the presence of various concentrations of Ca 2 or Ba 2 as the charge carrier. The numbers of observations are specified over each bar pair. (C) Plot of the quantity (1 fi)/fi (where fi is the fractional reduction of Ca 2 -current amplitude induced by 30 MZn 2 ) as a function of extracellular Ca 2 (diamonds) or Ba 2 (triangle) concentration (see the text for details). The straight lines are linear best fittings to data points obtained applying Eq. 3. The fitting parameters are K DZn M, K DCa M (plot of Ca 2 data); K DZn M, K DBa M (plot of Ba 2 data). dependence on Ca 2 concentration of I Ca -amplitude fractional inhibition (fi) caused by a given concentration of Zn 2 would be predicted to be: fi [Zn 2 ]/{[Zn 2 ] (K DZn /K DCa ) [Ca 2 ] K DZn }, where K DZn and K Dca are the dissociation constants for Zn 2 and Ca 2, respectively. The above equation yields: (1 fi)/fi {(K DZn /K DCa ) [Ca 2 ] K DZn }/[Zn 2 ], (3)

7 J. Magistretti et al. / Neuroscience 117 (2003) hence a linear dependence on [Ca 2 ] of the ratio (1 fi)/fi. Therefore, this quantity was calculated in each cell and average values were plotted as a function of extracellular [Ca 2 ] (Fig. 4C, diamonds). This plot showed a clearly linear behavior, and linear regression returned values of 25.2 M and 2.83 mm for K DZn and K DCa, respectively. Low concentrations of Ba 2 were then used to test the dependence of Zn 2 -induced I Ba inhibition on Ba 2 concentration. In 3 mm Ba 2 and1mmba 2 (Fig. 4A4), 30 M Zn 2 reduced I Ba amplitude by % (n 9) and % (n 5), respectively. The percent inhibition observed in 1 mm Ba 2 was significantly higher than that found in 5 mm Ba 2 (P 0.05). The plot of the quantity (1 fi)/fi as a function of the permeant ion s concentration was also constructed for the data obtained in extracellular Ba 2 (Fig. 4C, triangles). The linear fitting obtained applying Eq. 3 returned values of 10.5 M and 5.36 mm for K DZn and K DBa, respectively. To further confirm that the potency of Zn 2 in blocking the Ca 2 channels present in our experimental model depends on the identity and concentration of the permeant ion, dose-response experiments similar to those illustrated in Fig. 2 were carried out in 1 mm Ca 2,5mMCa 2, and 1mMBa 2 (n 3 in all cases). Langmuir fittings of the corresponding plots returned IC 50 values of 34.1 M, M, and 6.6 M, respectively. Finally, the effects of Zn 2 in slowing Ca 2 -current activation kinetics were also analyzed at the various concentrations of extracellular Ca 2 and Ba 2 (Fig. 4A1 4, lower traces). act was measured at the peak of the I-V relationship in all cases. (The voltage corresponding to the peak of the I V relationship and in general the position of the I V relationship over the voltage axis was influenced by the identity and concentration of the charge carrier used. This is expected because Ca 2 and Ba 2 are know to exert surface-charge shielding effects in a concentration-dependent manner, Ca 2 being more effective than Ba 2 in this action. In 1, 2, and 5 mm extracellular Ca 2 the peak of the I V relationship was observed between 10 and 0 mv, at 0 mv, and at 10 mv, respectively; in 1, 3, and 5 mm extracellular Ba 2, the peak was at 10 mv, between 10 and 0 mv, and at 0 mv, respectively.) act was significantly increased in all conditions considered (Fig. 4B, hatched bars). For instance, in 1 mm and 5 mm Ca 2, 30 M Zn 2 increased act by % and %, respectively. The differences in act percent increase found in the various conditions were not statistically significant (P 0.15 in all cases, considering the measures obtained in either 5 mm Ba 2 or 1mMCa 2 as the reference). Hence, the modifications induced by Zn 2 on HVA currents recorded using 5 mm Ba 2 as the charge carrier were qualitatively (and, considering the slowing of residual-current activation kinetics, also quantitatively) similar to those observed using Ca 2 at nearly-physiological concentrations (1 and 2 mm). The experiments illustrated from this point on were performed using 5 mm Ba 2 as the charge carrier, which augmented current amplitude by 2.5 and 4.0 times as compared with 2 mm and 1 mm Ca 2, respectively, thus substantially increasing the signal-to-noise ratio. Moreover, using extracellular Ba 2 instead of Ca 2 considerably improved the stability of recordings. Zn 2 affects multiple pharmacological types of HVA I Ba s Neuronal whole-cell Ca 2 currents consist of multiple pharmacological components, each of which is underlain by specific Ca 2 -channel subtypes. To isolate the single pharmacological Ca 2 -current components expressed by the neurons under study and determine the effects of Zn 2 on each of them, selective L-, N-, and P/Q-type Ca 2 - channel blockers were used, namely nifedipine, -CTx GVIA, and -AgaTx IVA, respectively. Since the blocking actions of -CTx GVIA and -AgaTx IVA on their specific targets are known to be largely irreversible or incompletely reversible (Bean, 1989; Mintz et al., 1992), whereas the here-reported effects of Zn 2 were completely and quickly reversible (see above), the protocol illustrated in Fig. 5A for the case of -AgaTx IVA was employed. 20 M Zn 2 was first applied (which induced a 50% inhibition of total I Ba s), then one of the above organic blockers was delivered in association with Zn 2 (which resulted in a further, partial block), then Zn 2 was washed out in the presence of the organic blocker (which caused a partial recovery). The traces marked 1 and 4 in Fig. 5A were used to isolate, by subtraction, the total -AgaTx IVA-sensitive (or P/Q-type) component, the traces marked 2 and 3 were used to derive, again by subtraction, the residual P/Q-type current in the presence of Zn 2 block. Fig. 5B1 3 shows the total and Zn 2 -resistant P/Q-, N-, and L-type current fractions obtained by applying the above procedure in three representative cells. The R-type current component was isolated by simultaneously applying 10 M nifedipine, 1 M -CTx GVIA, and the N- and P/Q-channel blocker, -conotoxin MVIIC (1 M), then 20 M Zn 2 was delivered in addition to this organic-blocker mix (Fig. 5B4). These experiments showed that all of the above pharmacological Ca 2 -current components are sensitive to Zn 2 block. 20 M Zn 2 inhibited L-type currents by % on average (n 6), N-type currents by % (n 5), P/Q-type currents by % (n 4), and R-type currents by % (n 4). In addition, the activation kinetics of the single pharmacological Ca 2 -current components were analyzed both before and after Zn 2 inhibition. Fig. 6A D shows L-, N-, P/Q-, and R-type currents recorded in four other representative neurons in the absence and in the presence of 20 M Zn 2. Control and Zn 2 -resistant currents are shown normalized to their peak amplitudes, superimposed, and expanded in the time scale to highlight the current activation phase. It clearly appears that all of the single current components displayed slower activation speed after Zn 2 inhibition. Double exponential fittings of current activation phases were carried out, and act values were derived (see above). In all cases, act was significantly increased by Zn 2,by 52 to 63% (Table 1). The above data clearly demonstrate that both effects

8 256 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 5. Zn 2 inhibits multiple pharmacological types of Ca 2 currents. (A) Effects of sequential application of 20 M Zn 2 and -agatoxin IVA ( -AgaTx IVA, 100 nm) on I Ba amplitude in a representative neuron. (A1) illustrates the time course of the effects of blocker delivery on I Ba peak amplitude. The empty horizontal bars correspond to the periods of Zn 2 application, the filled bar indicates the time of -AgaTx IVA application. Currents were evoked by 50-ms depolarizing square pulses at 0 mv delivered every 7 s. (A2) and (A3) show original current traces from the same experiment, recorded in the absence (A2) or presence (A3) of 20 M Zn 2 either without or with application of -AgaTx IVA. Calibration bars: 200 pa, 10 ms. (B) Effects of 20 M Zn 2 on the amplitude of P/Q- (B1), N- (B2), L- (B3), and R- (B4)-type currents recorded in four representative neurons. The single pharmacological current components (except R currents) were isolated by subtraction on the basis of recordings carried out as illustrated in (A). Traces in (B1) were derived by subtraction from the same currents shown in (A2) and (A3). The traces marked by asterisks are currents obtained in the presence of Zn 2. Calibration bars: 80 pa (B1), 60 pa (B2), 130 pa (B3), 30 pa (B4); 10 ms. exerted by Zn 2 on total I Ba s (i.e. block and slowing of activation kinetics in residual currents) can be also found in each one of the pharmacological components that sum up to form the same currents. Hence, the lower activation speed found in residual currents after Zn 2 block is unlikely to be caused by the preferential inhibition of a fastactivating current component that would unveil sloweractivating component(s). Rather, it must be the consequence of some molecular event(s) mediated by Zn 2 resulting in modifications of the macroscopic activation process of multiple Ca 2 -current components. Zn 2 -induced slowing of I Ba activation kinetics: analysis of the mechanism of action As discussed in more detail elsewhere for the specific case of Ni 2 (Magistretti et al., 2001), to explain the slowing

9 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 6. Zn 2 slows the activation kinetics of multiple pharmacological Ca 2 -current types. (A D) P/Q-, N-, L-, and R-type current fractions isolated in four representative neurons, in the absence and in the presence of 20 MZn 2. Test potential was 0 mv. Currents have been normalized to their peak amplitudes and are shown over an expanded time scale to highlight the effects of Zn 2 on activation kinetics. Calibration bar: 5 ms. effect exerted by an extracellular blocking cation on the activation phase of voltage-dependent Ca 2 currents, two orders of mechanisms can, in principle, be considered: 1) the block exerted by the cation on Ca 2 channels is intrinsically voltage-dependent: upon depolarization, the block is partly relieved with relatively slow kinetics, which would account for an apparent decrease of the activation speed of residual currents; 2) the block has no intrinsic voltage dependence: the cation decreases the speed at which Ca 2 channels reach a conducting state, with a mechanism that could be independent of the blocking action in itself. Under the hypothesis that the first mechanism applies to the here studied case of Zn 2, the steady-state blocking effect of Zn 2 on I Ba s would be expected to be progressively smaller with increasingly positive voltage levels of the pulses used to elicit I Ba s themselves. To test this possibility, the voltage-dependence of Zn 2 block was analyzed with the experiments illustrated in Fig. 7. Currents recorded at test potentials ranging from 50 to 30 mv both in the absence and in the presence of 30 MZn 2 (Fig. 7A1 and 2) were subtracted of the currents recorded after application of 200 M Ca 2 (Fig. 7A3), thus abolishing possible contaminations from residual outward currents that could be occasionally revealed at voltages positive to 10 mv. The I Ba s thus obtained (Fig. 7A, right Table 1. Average act values measured in P/Q-, N-, L-, and R-type current fractions (test potential 0 mv), both in the absence and in the presence of 20 M Zn 2. Statistical significance was determined on the basis of the Student s t-test for paired data Current type act (ms) n Level of significance Control Zn 2 (20 M) P/Q type P N type P 0.05 L type P 0.01 R type P 0.05 traces) were used to construct I V plots. Average I V relationships derived from 13 neurons in control conditions and after Zn 2 block are shown in Fig. 7B1. When normalized for maximal amplitude, the two I V plots appeared almost exactly overlapping (Fig. 7B2). Moreover, the fractional block induced by Zn 2 was basically constant at the different test potential explored (Fig. 7B2, inset). Hence, Zn 2 block of total I Ba s is not voltage dependent over the examined voltage range of I Ba activation. Average I V relationships were also constructed for the single pharmacological components in which the total I Ba s under study can be dissected. Normalized I V plots (Fig. 8) clearly showed that no significant modifications in voltage dependence were induced in the presence of Zn 2 inhibition in either L-type, N-type, P/Q-type, or R-type current fractions. To conclusively demonstrate that the slower activation kinetics observed in residual currents after Zn 2 block is not due to a mechanism of slow relief from an intrinsically voltage-dependent block upon depolarization, it is necessary to exclude that such an unblocking process is complete and saturating between 60 mv (the usual pre-step level) and 30 mv (the first voltage level at which sizable I Ba s start to be recorded) (see Magistretti et al., 2001, for more details on this issue). Under the latter hypothesis, no Zn 2 -dependent slowing of residual-current activation kinetics should be observed when delivering depolarizing test pulses from 30 mv rather than from the usual prestep level of 60 mv. To test this possibility, we performed experiments in which I Ba s were evoked, both in control conditions and during application of 30 M Zn 2, by test pulses at 0 mv delivered either from the usual level of 60 mv, or after a conditioning prepulse at 30 mv (Fig. 9, top traces). The duration of this prepulse was 25 ms, namely four times the average act observed under Zn 2 at the same voltage level. The normalized currents of Fig. 9 (bottom traces) clearly show that no detectable reduction

10 258 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 7. Zn 2 does not modify the voltage dependence of activation of total I Ba s. (A) I Ba s recorded in a representative neuron in response to step depolarizations at various voltage levels (see the voltage protocol in the bottom-right part of the panel), before (top traces) and during (middle traces) application of 30 MZn 2. The families of traces marked 1, 2, and 3 are original traces recorded in control conditions, under Zn 2, and in the presence of 200 M Cd 2, respectively. In the right half of the panel, Cd 2 -subtracted traces are shown. Calibration bars: 300 pa, 10 ms. (B) Average current-voltage (I V) relationships for control currents (empty symbols, continuous lines) and residual currents recorded during application of 30 M Zn 2 (filled symbols, dotted lines) (n 13). In each cell, I Vs were normalized for the maximal current value observed in control conditions, then I-Vs were averaged among cells. (B1) shows the original plots obtained in this way, (B2) the same plots, both normalized for the maximal value of each. The inset in (B2) is a plot of average I Ba -amplitude fractional reduction caused by 30 M Zn 2 as a function of test potential. The data points were derived from the mean values shown in (B1). The linear regression of the plot (straight line) returned no significant correlation between the two variables (slope coefficient mv 1, P 0.669). of the slowing action exerted by Zn 2 on I Ba activation kinetics was caused by the depolarizing prepulse at 30 mv. On average, the percent act increase in residual currents during Zn 2 application was % in control conditions, and % after the conditioning prepulse at 30 mv (n 10). These results provide a conclusive demonstration that the decrease in I Ba activation speed carried out by Zn 2 is not due to a time-dependent process of relief from an intrinsically voltage-dependent block. Effects of Zn 2 on I Ba deactivation kinetics We further examined whether Zn 2, besides decreasing I Ba activation speed, also affects I Ba deactivation kinetics upon repolarization. The voltage protocol routinely applied for this purpose is illustrated in Fig. 10A1. Fifteen-ms test pulses delivered to elicit I Ba s were followed by step repolarizations at voltage levels variable from 20 to 60 mv. In these experiments, the 2-s conditioning prepulse and the test pulse were set at 50 mv and 10 mv, respectively, so as to reduce the current size and thereby optimize the clamp control of tail currents. Fig. 10A2 shows currents recorded in a representative neuron in response to such voltage protocol, both in the absence and presence of 30 MZn 2. Currents are shown normalized to the peak amplitude of the tail currents resulting from step repolarization. It can be seen that tail-current decay was slowed somewhat during Zn 2 application. Such slowing effect was readily reversible upon Zn 2 washout (not shown). The decay phase of tail currents consistently followed a biexponential time course (Fig. 10A2, enhanced lines).

11 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 8. Zn 2 does not modify the voltage dependence of activation of single pharmacological I Ba components. (A D) Average I V plots derived for P/Q-, N-, L-, and R-type current fractions in the absence (open symbols, continuous lines) and in the presence (filled symbols, dotted lines) of 30 M Zn 2. I-V plots were constructed as explained in Fig. 7 legend, and normalized for the maximal value found in each. n 3 in all cases. Zn 2 application increased the lumped deactivation time constant, deact (calculated in the same way as the lumped activation time constant, act : see above, Eq. 2) by 10-19%, depending on the voltage level (Fig. 10B2). This slight increase reached a level of statistical significance (P 0.05, t test for paired data) between 40 and 60 mv (Fig. 10B1). The decrease of I Ba deactivation speed observed during Zn 2 application is further evidence that Zn 2 -induced changes of I Ba kinetics are not the consequence of an intrinsic voltage dependence of Zn 2 blocking action. Indeed, if a process of slow relief from block was acting Fig. 9. Slowing of residual I Ba s during Zn 2 application is not affected by depolarizing prepulses at 30 mv. The figure illustrates the experimental protocol used for testing the effect of depolarizing pre-steps on Zn 2 -induced slowing of I Ba activation kinetics. The voltage-clamp protocol applied is shown in the upper part of the figure. The current traces recorded in a representative neuron, both in the absence and in the presence (slower-activating currents) of 30 M Zn 2, are shown in the lower part of the figure. Currents have been normalized to their peak values. Calibration bar: 5 ms. during depolarizing test pulses as a consequence of depolarization itself, one would expect that, upon repolarization, the tendency of Zn 2 to re-block open channels would create an additional escape path from conducting to nonconducting states, thus accelerating, rather than slowing, the apparent deactivation kinetics (for details, see Magistretti et al., 2001). In this case, Zn 2 would produce, if anything, a decrease, and not an increase, of deact values derived from tail-current decay, which is in contrast with the above-illustrated data. Effects of Zn 2 on I Ba s elicited by action-potentiallike waveforms Because Zn 2 decreased the activation speed of residual I Ba s, the inhibitory action of Zn 2 on I Ba amplitude was greater when the early phases of I Ba activation, rather than peak currents, were considered. As a consequence, it can be expected that Zn 2 -dependent inhibition of peak I Ba amplitude is potentiated when these currents are elicited by short, phasic depolarizations, such as action potentials, instead of the step pulses employed experimentally. Moreover, since the effects of Zn 2 on I Ba deactivation speed upon repolarization were found to be relatively minor (see above), it can also be expected that a similar potentiation of Zn 2 -dependent inhibition may be observed when the current integral over time (which corresponds to the amount of charge transferred), rather than current amplitude, is considered. To test these hypotheses, we performed experiments in which I Ba s were evoked with actionpotential-like waveforms (APWs) of variable duration (Fig.

12 260 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 10. Zn 2 slows repolarization-induced tail currents. (A) I Ba s recorded in a representative neuron in response to a fixed depolarizing step at 10 mv, followed by repolarizing steps at variable voltage levels (see the voltage protocol in A1), before and during application of 30 M Zn 2. The slower-activating and -deactivating currents are those recorded in the presence of Zn 2. Currents have been normalized to the peak amplitudes of tail currents (I tail s) elicited upon step repolarization. Enhanced lines are biexponential best fittings to I tail deactivation phases. The values of the lumped deactivation time constant ( deact ; see text for details) are also specified. Calibration bar: 5 ms. The inset is a highlight, over an expanded time scale, of I tail s recorded at 60 mv before and during application of 30 M Zn 2 (calibration bar: 1 ms). (B) Average plots of the voltage dependence of the lumped deactivation time constant, deact (B1; empty symbols, continuous line: control; filled symbols, dotted line: 30 MZn 2 ), and of the percent deact increase during Zn 2 application (B2). n 12 throughout. 11A2) in addition to the usual depolarizing step pulses (Fig. 11A1). The currents thus recorded both in the absence and in the presence of 30 M Zn 2 in a representative neuron are shown in Fig. 11B (all of the currents used for this kind of analysis were Cd 2 -subtracted). Step pulses returned a level of Zn 2 -dependent percent inhibition of 60% (horizontal, dashed line). The peak amplitude of currents recorded in response to an APW of 100 mv in amplitude and 1 ms in half width (similar to the action potentials produced by PC superficial pyramidal neurons: see Tseng and Haberly, 1989) was decreased by Zn 2 to a higher degree (Fig. 11B, left panel). The same was observed when current integral was considered. When the APW half width was increased to 2 and 4 ms, the extra inhibition of current peak amplitude and current integral progressively decreased to reach values similar to that observed when using step depolarizing test pulses (Fig. 11B, middle and right panels). Average data for current peak amplitude and current integral are illustrated in Fig. 11D. In both cases, APWs of 1 and 2 ms in half width were accompanied by significant degrees of Zn 2 -dependent additional inhibition equal to 8 and 3%, respectively. In the case of APWs of 4 ms in half width, no significant differences in percent inhibition were observed as compared with currents evoked by step pulses. These findings are consistent with an extra Zn 2 -dependent inhibitory action being exerted on I Ba s elicited by short, phasic depolarizations, as a consequence of the Zn 2 -induced reduction of I Ba activation speed. As expected, the effects of this slowing of I Ba activation kinetics appeared to be overcome as the duration of the stimulus applied was made progressively longer. DISCUSSION The present study shows that zinc ions can exert a relatively potent blocking action on multiple subtypes of HVA Ca 2 currents in mammalian paleocortical neurons. Although most group VIIA-IIB metal divalent cations (including Mn 2,Co 2,Ni 2, and Cd 2 ) are known to interact with voltage-gated Ca 2 channels in a number of cell systems, thus exerting, on the corresponding currents, more or less potent inhibitory effects that have been characterized in detail (Lansman et al., 1986; Swandulla and

13 J. Magistretti et al. / Neuroscience 117 (2003) Fig. 11. Enhancement of Zn 2 inhibitory action in I Ba s elicited by short, action-potential-like phasic depolarizing stimuli. (A) Step (A1) and action-potential-like (A2) depolarizing test pulses delivered to evoke total I Ba s. The y-axis labels are mv, every x-axis division is 1 ms. Note the different half-widths of action-potential-like waveforms (APWs). (B) I Ba s recorded in a representative neuron in response to the voltage-clamp protocols illustrated in panel A, both in control conditions and in the presence of 30 MZn 2 (all currents are Cd 2 -subtracted). In panel B2, the continuous lines are current traces, the dotted lines are the corresponding integrals over time (which represent the amount of charge transferred). Note that in both panel B1 and B2, currents and integrals were normalized to the peak values observed in control conditions. Actual control-current peak amplitudes were pa (B1), pa (B2, left), pa (B2, middle), pa (B2, right); actual maximal values of control-current integrals were fc (B2, left), fc (B2, middle), fc (B2, right). The horizontal, dashed line marks the relative level of I Ba -amplitude inhibition observed in 30 MZn 2 using the step protocol. (C) The same currents depicted in the left part of panel B2 (evoked with an APW of 1 ms in half width) are shown normalized to their peak values, to highlight the scarce effect of Zn 2 on the kinetics of current deactivation phase. The dotted-line trace is the current recorded in 30 M Zn 2. Calibration bar: 1 ms. (D) Average plots of Zn 2 -dependent percent inhibition of I Ba peak amplitude (D1) and current integral (D2) as a function of the half width of APWs used as depolarizing stimuli (n 9). The horizontal, continuous line in each sub-panel correspond to the average level of Zn 2 -dependent percent inhibition of I Ba amplitude as measured with step protocols (test potential 0 mv). Statistical significance was calculated considering the percent inhibition measured with step protocols as the reference. **, P 0.01; *, P 0.05 (t test for paired data). Armstrong, 1989; Carbone and Swandulla, 1989; Winegar et al., 1991), the data so far available as to the effects of Zn 2 on VDCC in central neurons were scarce. This is surprising, especially when considering that free Zn 2 is present at a number of central (especially glutamatergic) synapses where it is stored in synaptic vesicles, can be synaptically released, and is likely to act as a physiological, endogenous modulator of specific postsynaptic ion-channel targets, including NMDA receptors (see the Introduction section). In the experimental conditions routinely employed in this study, which included 5 mm extracellular Ba 2 as the charge carrier, Zn 2 blocked total HVA Ca 2 currents in PC pyramidal neurons with an IC 50 of 20 M. The blocking potency of Zn 2 observed using Ca 2, instead of Ba 2, at the concentration closest to the physiological extracellular levels (1 mm), was not very different (IC M). In the studies in which inhibitory effects of Zn 2 on HVA currents of mammalian peripheral or central neurons have been described, considerably diverse blocking potencies were found. The IC 50 values reported are 7 M in neurons from the rat diagonal band of Broca (in 2 mm Ba 2 as the charge carrier; Easaw et al., 1999), 210 M in mouse cultured neocortical neurons (in 2 mm Ca 2 ; Kerchner et al., 2000), 69 M in rat cultured dorsal-root-ganglion neurons (in 10 mm Ba 2 ; Buesselberg et al., 1994), and 300 M in rat cerebellar Purkinje cells (in 5 mm Ca 2 ; Nam and Hockberger, 1992). One obvious, possible reason of such diversity in Zn 2 potency consists in the different neuronal systems studied. However, our results strongly suggest that an additional source of heterogeneity may be given by the different recording conditions adopted. We showed that higher degrees of Zn 2 -dependent inhibition can be observed with lower concentrations of the permeant ion (either Ca 2 or Ba 2 ). Our findings are consistent with the existence of a competition between Zn 2 and the permeant ion for a binding site the occupancy of which by Zn 2 results in Ca 2 -channel block. An obvious possibility