Zn²⁺ current is mediated by voltage-gated Ca²⁺ channels and enhanced by extracellular acidity in mouse cortical neurones

Abstract

1. Mammalian neuronal voltage-gated Ca²⁺ channels have been implicated as potential mediators of membrane permeability to Zn²⁺. We tested directly whether voltage-gated Ca²⁺ channels can flux Zn²⁺ in whole-cell voltage-clamp recordings from cultured murine cortical neurones.

2. In the presence of extracellular Zn²⁺ and no Na⁺, K⁺, or other divalent cations, a small, non-inactivating, voltage-gated inward current was observed exhibiting a current-voltage relationship characteristic of high-voltage activated (HVA) Ca²⁺ channels. Inward current was detectable at Zn²⁺ levels as low as 50 μm, and both the amplitude and voltage sensitivity of the current depended upon Zn²⁺ concentration. This Zn²⁺ current was sensitive to blockade by Gd³⁺ and nimodipine and, to a lesser extent, by ω-conotoxin GVIA.

3. Zn²⁺ could permeate Ca²⁺ channels in the presence of Ca²⁺ and other physiological cations. Inward currents recorded with 2 mm Ca²⁺ were attenuated by Zn²⁺ (IC₅₀ = 210 μm), and currents recorded with Zn²⁺ were unaffected by up to equimolar Ca²⁺ concentrations. Furthermore, the Zn²⁺-selective fluorescent dye Newport Green revealed a depolarisation-activated, nimodipine-sensitive Zn²⁺ influx into cortical neurones that were bathed in a physiological extracellular solution plus 300 μm ZnCl₂.

4. Surprisingly, while lowering extracellular pH suppressed HVA Ca²⁺ currents, Zn²⁺ current amplitude was affected oppositely, varying inversely with pH with an apparent pK of 7·4. The acidity-induced enhancement of Zn²⁺ current was associated with a positive shift in reversal potential but no change in the kinetics or voltage sensitivity of channel activation.

5. These results provide evidence that L- and N-type voltage-gated Ca²⁺ channels can mediate Zn²⁺ entry into cortical neurones and that this entry may be enhanced by extracellular acidity.

Zinc is concentrated in the synaptic vesicles of central excitatory nerve terminals (Perez-Clausell & Danscher, 1985) and is released upon synaptic activation (Assaf & Chung, 1984; Howell et al. 1984). The total concentration of Zn²⁺ in the hippocampus and cortical grey matter has been estimated at 220–300 μm (Frederickson, 1989), and peak concentrations of synaptically released Zn²⁺ may reach hundreds of micromolar in hippocampal slices (Assaf & Chung, 1984; Vogt et al. 2000) – concentrations more than sufficient for Zn²⁺ to exhibit profound modulatory effects on a variety of agonist- and voltage-gated mammalian ion channels, including attenuation of current through NMDA and GABA receptor-gated channels, potentiation of AMPA, glycine, and ATP receptor-mediated currents, modulation of the transient outward K⁺ current I_A, and blockade of voltage-gated Na⁺ and Ca²⁺ channels (Harrison & Gibbons, 1994; Smart et al. 1994).

Excessive exposure to extracellular Zn²⁺ can also be toxic to neurones (Yokoyama et al. 1986), inducing apoptosis at lower concentrations (Manev et al. 1997; Lobner et al. 2000) and necrosis at higher concentrations (Choi et al. 1988; Lobner et al. 2000). Zn²⁺ neurotoxicity has been implicated in the pathogenesis of selective neuronal death following transient global cerebral ischaemia and prolonged seizures (Choi & Koh, 1998). A critical step in Zn²⁺-induced neuronal death appears to be its permeation across the plasma membrane (Tønder et al. 1990; Weiss et al. 1993; Koh et al. 1996). Raising extracellular Ca²⁺ reduces Zn²⁺ neurotoxicity (Weiss et al. 1993; Koh & Choi, 1994), suggesting that routes implicated in membrane permeability to Ca²⁺ may also mediate Zn²⁺ flux (Choi & Koh, 1998). Putative routes of neuronal Zn²⁺ entry include Ca²⁺-permeable AMPA (Yin & Weiss, 1995; Sensi et al. 1997) and NMDA (Christine & Choi, 1990; Koh & Choi, 1994) receptor-gated channels, as well as transporter-mediated exchange with intracellular Na⁺ (Sensi et al. 1997). An especially prominent route of toxic Zn²⁺ entry may be voltage-gated Ca²⁺ channels, as neuronal depolarisation enhanced both Zn²⁺ influx and resultant neuronal death, and Ca²⁺ channel antagonists attenuated both indices (Weiss et al. 1993; Freund & Reddig, 1994; Manev et al. 1997).

Zn²⁺ is a potent antagonist of voltage-gated Ca²⁺ channels in invertebrates (Hagiwara & Takahashi, 1967) and mammals (Winegar & Lansman, 1990; Büsselberg et al. 1992). There is also clear evidence that Zn²⁺ can permeate voltage-gated Ca²⁺ channels in invertebrates: in the absence of Ca²⁺, Zn²⁺ can support full Ca²⁺ channel-mediated action potentials in insect muscle membranes (Fukuda & Kawa, 1977) and snail neurones (Kawa, 1979). In addition, nanoampere-sized whole-cell Zn²⁺ currents through verapamil- and Co²⁺-sensitive channels have been characterised in Helix neurones (Oyama et al. 1982). However, the evidence for Zn²⁺ permeation of voltage-gated Ca²⁺ channels in mammals has so far been only indirect. Voltage-gated Ca²⁺ channel antagonists, including dihydropyridine compounds and ω-conotoxin GVIA in particular, attenuated depolarisation-induced intracellular Zn²⁺ accumulation as measured by transcription of a reporter gene coupled to a Zn²⁺-activated metallothionein promoter in pituitary tumour cells (Atar et al. 1995), fluorescence of a Zn²⁺-sensitive dye loaded into living cortical neurones (Sensi et al. 1997), and cortical neuronal uptake of ⁶⁵Zn²⁺ (H. Ying and D. W. Choi, unpublished observations), suggesting that Zn²⁺ may permeate L- and N-type HVA Ca²⁺ channels. Although two reports exist of small (< 50 pA) inward whole-cell currents recorded in the presence of extracellular Zn²⁺ in bovine chromaffin cells (Vega et al. 1994) and rat ventricular myocytes (Atar et al. 1995), neither study directly implicated Zn²⁺ as the charge carrier or provided any characterisation of the currents observed. In this study, we sought to test whether cortical neuronal voltage-gated Ca²⁺ channels flux Zn²⁺. We showed that Zn²⁺ could indeed permeate these channels and, in addition, that it could do so in the ionic conditions normally encountered in the brain. Furthermore, we have demonstrated that extracellular acidity enhanced the permeation of Zn²⁺ relative to Ca²⁺ through voltage-gated Ca²⁺ channels. An abstract based on these results has appeared (Kerchner et al. 1998).

METHODS

Murine cortical cell cultures

Mixed cortical cultures, containing both neurones and glia, were prepared from embryonic day 15 mice as described previously (Rose et al. 1993). Embryos were harvested acutely from pregnant female mice killed under halothane anaesthesia by cervical dislocation. Embryos were decapitated, and neocortices were dissected, dissociated, and plated in Eagle's minimal essential medium (MEM, Earle's salts) supplemented with 20 mm glucose (final concentration 25 mm), 2 mm glutamine, 5 % fetal bovine serum, and 5 % horse serum at a density of 0·4 hemispheres per dish onto 35 mm culture dishes (Becton Dickinson and Company, Plymouth, UK), which were prepared by incubation overnight with 1 mg ml⁻¹ poly-D-lysine (Sigma Chemical Co.) and 20 μg ml⁻¹ laminin (Collaborative Biomedical Products, Bedford, MA, USA) and subsequent rinsing with water. The medium was changed after 9 days in vitro (DIV 9) to MEM containing 25 mm glucose and 2 mm glutamine. Cultures were maintained in a 37°C, humidified incubator in a 5 % CO₂ atmosphere. All experiments were performed between DIV 14 and 17. Protocols for the handling of experimental animals were approved by the Animal Studies Committee at Washington University School of Medicine, in accordance with local and national ethical guidelines.

Electrophysiology

Whole-cell voltage-clamp recordings were performed on cultured murine cortical neurones in 35 mm dishes on the stage of an inverted microscope (Nikon, Japan) using an Axopatch 200A amplifier (Axon Instruments Inc.). Patch electrodes were pulled from borosilicate glass capillary tubes (World Precision Instruments) using a Flaming-Brown micropipette puller (Sutter Instrument Co.) and fire polished to a final tip resistance of 7–12 MΩ when filled with the internal pipette solution (Table 1). Capacitance neutralisation and series resistance compensation were routinely applied, and input resistance was typically ∼400–500 MΩ in the standard external solution (Table 1). Current was digitally sampled at 20 kHz and filtered by a 5 kHz, 4-pole low-pass Bessel filter. Current traces were displayed and stored on an IBM PC-compatible computer (Comtrade Express) using the A/D converter interface Digidata 1200 and the data acquisition/analysis software package pCLAMP 6.0 (Axon Instruments Inc.). Leak subtraction was performed by pCLAMP by delivering two negative-going voltage pulses, each half the height of the test pulse, from a holding potential of −90 mV, before delivery of each depolarising test pulse from −70 mV. All experiments were performed at room temperature (21–22°C). Data are reported as means ± standard error of the mean (s.e.m.).

Table 1.

Composition of electrophysiological recording solutions (concentrations in mm except as indicated)

Solute	Internal (all expts)	External
		Expts at constant pH	Expts at varying pH
NMDG	—	122	120
ZnCl2 or CaCl2	—	As indicated	As indicated
TEA-Cl	130	20	20
Hepes	10	5	—
Hepes or Pipes	—	—	10
Glucose	—	10	10
MgCl2	3	—	—
BAPTA	2.5	—	—
Na2-ATP	5	—	—
Li-GTP	0.3	—	—
TTX	—	200 nm	200 nm
HCl	—	To pH 7.4	To indicated pH
TEA-OH	To pH 7.4	—	—

Abbreviations: NMDG, N-methyl-d-glucamine; TEA-Cl, tetraethylammonium chloride; Hepes, N-(2-hydroxyethyl) piperazine-N,N′-bis(2-ethanesulfonic acid); Pipes, piperazine-N,N′-bis (2-ethanesulfonic acid); BAPTA, 1,2-bis (2-aminophenoxy) -ethane-NN,N′,N′-tetraacetic acid; TTX, tetrodotoxin.

Solution preparation and delivery

The contents of the standard external bath and internal pipette solutions are itemised in Table 1. To aid in the formation of a tight electrical seal between the glass pipette and the neuronal membrane, cells were initially bathed in standard external solution supplemented with 2 mm MgCl₂. During recordings, a neurone was bathed in a control solution with no divalent cations, one supplemented with CaCl₂ or ZnCl₂, or one supplemented both with ZnCl₂ and a pharmacological agent. In each case, solutions were delivered directly to the tested neurone by continuous perfusion with a gravity-fed delivery pipette of approximately 300 μm internal diameter. Solution changes occurred at a cell in less than 100 ms.

Current-voltage curve fits

In Figs 4 and 9, current-voltage relations were fitted with curves to allow determination and comparison of the voltage sensitivity of channel activation in different conditions. At some voltage V:

where I is the magnitude of current measured in relative units, P_o is the proportion of channels open at V, V_rev is the reversal potential of the current, and g is proportional to membrane conductance when channels are fully activated. P_o is described by the Boltzmann relation:

where V_h is the voltage of half-maximal channel activation, and m is a slope factor. These equations apply whether I represents the contribution of a single ion (e.g. Zn²⁺) flowing into a cell or the net current resulting from multiple conductances (e.g. Zn²⁺ flowing into and Na⁺ flowing out of a cell); in the latter case, g would represent the sum of conductances, and V_rev would represent the net reversal potential of the current, an average of the V_rev values for each ion weighted by the relative contribution of each ion's conductance to the total membrane conductance. Curves were fitted by least-squares regression.

Figure 4.

Zn²⁺ concentration affected both maximal current amplitude and voltage sensitivity

A, increasing [Zn²⁺]_o from 50 μm (○-○) to 2 mm (□ – – □) resulted in both an increased maximum steady-state current amplitude (P < 0·05, Student's paired t test comparing I–V curve peaks in 50 μm and 2 mm Zn²⁺), and a rightward shift of the I–V relationship. Five cells were tested, each at both Zn²⁺ concentrations to maximise comparability of the data, and I–V curves for each cell were normalised to the peak steady-state inward current in 2 mm Zn²⁺. As each cell was first tested in 50 μm Zn²⁺ and then in 2 mm Zn²⁺, some current run-down is the likely explanation for why the current at 0 mV in 2 mm Zn²⁺ was not as much larger than the current in 50 μm Zn²⁺ as would be predicted from the data in Fig. 3. Points represent means ± s.e.m. Curves fitted to the points are described by the equation:

(see Methods). In 2 mm Zn²⁺, V_h = 2·1 mV, m = 9·4, V_rev = 38 mV, and g = 0·044; in 50 μm Zn²⁺, V_h = −22 mV, m = 5·3, V_rev = 14 mV, and g = 0·025. B, increasing [Zn²⁺]_o from 50 μm (continuous line) to 2 mm (dashed line) resulted in a shift of the voltage sensitivity of channel activation to more depolarised potentials. Boltzmann curves, described by the equation I_relative = {1 + exp[(V_h− V) m⁻¹]}⁻¹, are plotted using values for V_h and m derived from curve fits to the I–V data (A) to ease comparison of those two variables between conditions.

Figure 9.

Extracellular acidity enhanced Zn²⁺ current amplitude with no effect on voltage sensitivity

A, the I–V relationships of currents recorded in the presence of 500 μm Zn²⁺ at pH_o 7·4 (○-○) and pH_o 6·4 (□ – – □) were plotted. Twelve cells were tested, each at both pH_o values, and I–V curves for each cell were normalised to the peak inward steady-state current at pH_o 7·4. Points represent means ± s.e.m., and curves were fitted as in Fig. 4. At pH_o 7·4, V_h = −13 mV, m = 9·2, V_rev = 29 mV, and g = 0·038; at pH_o 6·4, V_h = −11 mV, m = 9·3, V_rev = 59 mV, and g = 0·037. B, decreasing pH_o from 7·4 (continuous line) to 6·4 (dashed line) resulted in little to no shift of the voltage sensitivity of channel activation. Curves were plotted as in Fig. 4.

Fluorescence detection of intracellular Zn²⁺

The intracellular Zn²⁺ concentration ([Zn²⁺]_i) was assessed using Newport Green as previously described (Canzoniero et al. 1999). Cell cultures were first washed with a Hepes-buffered physiological saline solution containing (mm): NaCl 120, KCl 5·4, MgCl₂ 0·8, Hepes 20, glucose 15, CaCl₂ 1·8, NaOH 10 (pH 7·4). Cells were then loaded with 5 μm Newport Green diacetate (excitation wavelength (λ), 485 nm; emission λ, 530 nm; Molecular Probes) in the presence of 0·02 % Pluronic F-127 for 30 min at room temperature. Neurones were washed and incubated for an additional 30 min in the Hepes-buffered saline. Before each experiment, cultures were washed twice with the same solution.

All experiments were performed at room temperature under constant perfusion (2 ml min⁻¹) on the stage of a Nikon Diaphot inverted microscope equipped with a 75 W xenon lamp and a Nikon × 40, 1·3 NA epifluorescence oil immersion objective lens. Background fluorescence was subtracted at the beginning of each experiment. Images were acquired with a CCD camera (Quantex, CA, USA) and digitised using Metafluor 4.0 software (Universal Imaging, West Chester, PA, USA).

[Zn²⁺]_i was calculated using in situ calibration at the end of each experiment. For each field of neurones, maximum (F_max) and minimum (F_min) fluorescence values were determined by exposure to 1 mm Zn²⁺plus 50 μm sodium pyrithione or to the heavy metal chelator N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine (TPEN, 100 μm) in a Zn²⁺-, Mg²⁺- and Ca²⁺-free medium, respectively. An assumed K_d of 1 μm (Haugland, 1996) was applied to the formula described by Grynkiewicz et al. (1985), which gives that for a fluorescence value F:

The experiments were performed twice with similar results.

Atomic zinc analysis

Samples (prepared according to Table 3) were transferred to a quartz beaker and treated with concentrated sulphuric acid. A sample was then repeatedly charred on a hot plate at 250°C and treated with nitric acid until the sample no longer charred, at which point the sample was resuspended in doubly deionised water. A determination of the zinc content of these samples was performed with an inductively-coupled plasma atomic emission spectrometer, the Thermo Jarrell Ash IRIS Advantage Duo-View System (Thermo Jarrell Ash Corp., Franklin, MA, USA).

Table 3.

Atomic emission spectroscopic analysis of zinc recovered from experimental recording solutions in various conditions

Nominal [Zn2+](μm)	pH	Exposure to cultures	Zinc recovered (p.p.m.)
500	7.4	−	28.8
500	6.4	−	29.5
0	7.4	−	< 0.5
0	6.4	−	< 0.5
500	7.4	+	28.0
500	6.4	+	29.1
0	7.4	+	<0.5
0	6.4	+	< 0.5

Extracellular recording solutions (Table 1) were prepared at the indicated pH and [Zn²⁺] and stored in the same polypropylene containers used to store solutions during electrophysiological experiments. Some solutions were washed through cortical neuronal cultures before storage.

Chemicals and pharmacological agents

All chemicals were purchased from Sigma Chemical Co. except ω-conotoxin GVIA which was from Bachem Biosciences Inc. (King of Prussia, PA, USA).

RESULTS

A voltage-gated current is observed in the presence of Zn²⁺

When extracellular Ca²⁺ is lowered below micromolar levels, voltage-gated Ca²⁺ channels become highly permeable to other divalent and monovalent ions including Na⁺ and K⁺ (Kostyuk et al. 1983; Almers & McCleskey, 1984). To isolate any Zn²⁺ permeation through voltage-gated Ca²⁺ channels, voltage-clamp recordings of cultured cortical neurones were made using an internal solution containing TEA to block voltage-gated K⁺ channels and an external solution nominally devoid of any divalent cations, Na⁺ and K⁺ (Table 1) to eliminate contributions of these ions to inward whole-cell currents. Under these conditions, no membrane current was observed during a voltage step from −70 to +10 mV, whereas in the presence of 2 mm extracellular Zn²⁺, a sustained inward current appeared (Fig. 1A) with an average steady-state amplitude of −19 ± 1 pA (n = 33) and a current-voltage (I–V) relationship consistent with mediation by voltage-gated Ca²⁺ channels (Fig. 1B). No significant voltage-dependent steady-state inactivation was observed in inward currents recorded in the presence of Zn²⁺ after prepulses of 1 s duration, although currents carried by Ca²⁺ showed substantial inactivation in the same conditions (Fig. 2). To compare the activation kinetics of inward currents recorded in the presence of either Zn²⁺ or Ca²⁺, the 10–90 % rise times of currents elicited by voltage steps to 0 mV were measured, revealing a slower activation phase when Zn²⁺ rather than Ca²⁺ was used as the putative charge carrier (Table 2).

Figure 1.

Zn²⁺ produced an inward current in cultured cortical neurones

A, an inward current was elicited in the presence but not the absence (Control) of 2 mm Zn²⁺ by a voltage step to +10 mV. Illustrated are superimposed traces from a representative cell (left; data filtered at 0·5 kHz for display) and the averaged current from 33 cells in 2 mm Zn²⁺ (right; no additional filter). Dotted line indicates the baseline zero current level for these leak-subtracted traces (see Methods). B, the I–V relationship of steady-state current recorded during voltage steps from −70 mV in the presence of 2 mm Zn²⁺ is shown as the mean ± s.e.m. of 5 cells. Sample traces over a range of voltages (mV) are illustrated for a representative cell (inset; current traces here and in subsequent figures were filtered at 1 or 2 kHz to optimise legibility). ‘Steady-state current’ in these and the other experiments reported here refers to an average of the sustained component of current, minus baseline, during a voltage step; for instance, the average current recorded during the last 100 ms of a 200 ms pulse. A reversal potential appears in this I–V relationship, and an outward current was evident at positive potentials. This outward current was probably not carried by Zn²⁺ that had accumulated within a cell during experiments, as recorded cells were perfused with BAPTA (2 mm; K_d for Zn²⁺ = 1–10 nm; Aballay et al. 1995). More likely, intracellular cations such as Na⁺ and Li⁺ (Table 1), as well as some residual K⁺, may have accounted for the outward conductance, by analogy with the known effect of intracellular monovalent cations to influence the experimentally observed reversal potential of voltage-gated Ca²⁺ currents (Hille, 1992).

Figure 2.

Zn²⁺ current exhibited no voltage-dependent steady-state inactivation

A, after 1 s prepulses to the indicated potentials, the amplitude of steady-state inward currents elicited by immediately subsequent voltage steps to 0 mV in the presence of 500 μm Zn²⁺ (○; n = 6) or 2 mm Ca²⁺ (□; n = 3) are plotted relative to values obtained after a prepulse to −100 mV. In each cell tested, families of voltage pulses were delivered twice – once in an ascending order (−100 to +50 mV), and once in a descending order, to help compensate for any non-voltage-dependent current run-down. * Significant difference between relative current amplitudes recorded in Zn²⁺versus Ca²⁺ after the indicated prepulse potentials (two-way ANOVA with Bonferroni's t test). Points represent means ± s.e.m.B, traces obtained by voltage steps from −70 to 0 mV after a 1 s prepulse to −100 mV (left) or +5 mV (right) are illustrated for the cell for which the greatest degree of inactivation was observed in the presence of 500 μm Zn²⁺. The dotted line indicates the zero current level. The sustained component (latter 100 ms) of the inward currents were compared to generate the graph in A.

Table 2.

Rise times of currents carried by Zn²⁺ or Ca²⁺

Charge carrier	pH	10–90%(ms) rise time	n
Ca2+ (1 mm)	7.4	5.5 ± 0.5	7
Zn2+ (0.5 mm)	7.4	12 ± 2	12
Zn2+ (0.5 mm)	6.4	14 ± 2	12

Rise times were calculated from the rising phase of currents evoked by voltage steps from −70 to 0 mV. Rise times differ significantly between Zn²⁺ and Ca²⁺ currents at pH 7.4 (P < 0.05, Student's t test) but not between Zn²⁺ currents at pH 7.4 and 6.4.

Zn²⁺ concentration affects current amplitude and voltage sensitivity

To test the hypothesis that Zn²⁺ was the charge carrier for the inward current described above, the effects of varying the extracellular Zn²⁺ concentration ([Zn²⁺]_o) were explored. As [Zn²⁺]_o was increased through a range from 200 μm to 1 mm, the steady-state inward current amplitudes elicited by voltage steps from −70 to 0 mV increased (Fig. 3). Increasing [Zn²⁺]_o to 2 mm did not increase inward current amplitude any further.

Figure 3.

Inward current amplitude at 0 mV depended upon extracellular Zn²⁺ concentration

Increasing [Zn²⁺]_o increased the amplitude of steady-state currents elicited by voltage steps to 0 mV. Points represent means ± s.e.m. of 4–7 cells and are plotted relative to currents elicited in the presence of 2 mm Zn²⁺.

To determine whether this positive correlation between [Zn²⁺]_o and steady-state current amplitude at 0 mV was due to an increase in maximum steady-state current amplitude or a shift in the I–V relationship, I–V curves were constructed in the presence of 50 μm and 2 mm extracellular Zn²⁺ (Fig. 4A). The maximum steady-state current amplitude recorded in the presence of 2 mm Zn²⁺ was larger than in 50 μm Zn²⁺. In addition, the I–V curve peak in 2 mm Zn²⁺ occurred at a more depolarised membrane potential than in 50 μm Zn²⁺. When curves were fitted to the I–V data, the voltage dependence of channel activation could be approximated (see Methods). Raising the Zn²⁺ concentration caused the channel activation-voltage relationship to shift rightwards with some reduction in slope; V_h was approximately −22 mV in 50 μm Zn²⁺ and +2·1 mV in 2 mm Zn²⁺ (Fig. 4B).

Zn²⁺ current is blocked by voltage-gated Ca²⁺ channel antagonists

To investigate whether voltage-gated Ca²⁺ channels mediated the inward current recorded in the presence of 2 mm Zn²⁺ and to determine which subtypes of these channels were involved, we tested voltage-gated Ca²⁺ channel antagonists for their effects on the current (Fig. 5A). Gd³⁺ (10 μm), a non-subtype-selective Ca²⁺ channel antagonist (Lansman, 1990; Canzoniero et al. 1993), suppressed steady-state current amplitude by 86 ± 5 % (n = 8). Nimodipine (1 μm), an L-type Ca²⁺ channel antagonist, attenuated a larger portion of the inward current (81 ± 10 %; n = 6) than ω-conotoxin GVIA (1 μm; 33 ± 4 %; n = 6), an N-type Ca²⁺ channel antagonist. In contrast, MK-801 (1 μm), an NMDA receptor-gated channel blocker, had no effect on the current amplitude (0·0 ± 8 % compared to controls; n = 3). Further experiments with nimodipine demonstrated that its inhibition of a Zn²⁺ current was reversible (Fig. 5B) and occurred at all voltage levels (Fig. 5C).

Figure 5.

Zn²⁺ current was sensitive to blockade by non-specific and specific blockers of voltage-gated Ca²⁺ channels

A, steady-state Zn²⁺ current was inhibited by 10 μm Gd³⁺ (Gd), 1 μm nimodipine (Nimo), and 1 μm ω-conotoxin GVIA (Cono), but not by 1 μm MK-801. Zn²⁺ current was evoked by a voltage step to +10 mV in the presence of 2 mm Zn²⁺. The level of steady-state current recorded in the presence of an antagonist was normalised to current amplitude measured with Zn²⁺ alone and plotted as the mean + s.e.m. of 3–8 cells per condition. Cells were perfused with drugs or control solution for 30 s between voltage step trials. * Significant difference between normalised steady-state current amplitudes measured in the presence and absence of an agent; † Significant difference compared to currents measured after treatment with nimodipine or Gd³⁺ (P < 0·05, one-way ANOVA with Bonferroni's t test). B, current traces from a representative cell demonstrate reversible inhibition of Zn²⁺ current by 1 μm nimodipine; the cell was treated as in A. The dotted line represents the zero current level. C, the I–V relationship of steady-state currents recorded in the presence of 2 mm Zn²⁺ in the absence (○) or presence (□) of 1 μm nimodipine is plotted. I–V curves from 3 cells were normalised, each to its peak, and averaged. Points represent means ± s.e.m.

Voltage-gated Ca²⁺ channel-mediated Zn²⁺ influx can occur in the presence of Ca²⁺

While isolating the contribution of Zn²⁺ to voltage-gated inward currents required the removal of other mono- and divalent cations – including Ca²⁺, in particular – from the extracellular bathing solution, it is important to determine whether voltage-gated Ca²⁺ channel-mediated Zn²⁺ entry can occur in physiological solutions. First, the effect of extracellular Ca²⁺ on Zn²⁺ current was examined. When inward currents were generated by voltage steps in the presence of 2 mm Zn²⁺, extracellular Ca²⁺ up to 2 mm had no effect on steady-state current amplitude (n = 7–14 cells per Ca²⁺ concentration; Fig. 6A). At 10 mm Ca²⁺, the net amplitude of inward current doubled compared to the level of current recorded in Zn²⁺ alone, but it was not as high as the currents measured in the presence of 2 mm Ca²⁺ alone (n = 4; Fig. 6A), suggesting that the presence of Zn²⁺ impeded Ca²⁺ permeation. To confirm that Zn²⁺ could block cortical neuronal voltage-gated Ca²⁺ current as reported previously in other systems (see Introduction), the effect of Zn²⁺ on voltage-gated inward currents generated in the presence of 2 mm Ca²⁺ was tested. In contrast to the effect of Ca²⁺ on Zn²⁺ current, Zn²⁺ exerted a dose-dependent depression of Ca²⁺ current, with an approximate IC₅₀ of 210 μm (n = 4; Fig. 6B).

Figure 6.

Zn²⁺ competed with Ca²⁺ for voltage-gated Ca²⁺ channel permeation

A, inward currents, evoked by voltage steps from −70 to +10 mV, were recorded from cells in the presence of 2 mm Zn²⁺ plus varying concentrations of Ca²⁺ (○; n = 7–14 cells per condition). Steady-state current amplitude is plotted relative to control currents evoked in the presence of 2 mm Zn²⁺ alone (left). Relative current amplitude recorded in the presence of 2 mm Ca²⁺ and no Zn²⁺ is plotted for comparison (□; n = 4). Points represent means ± s.e.m. Representative traces (right) show currents recorded in the presence of 2 mm Zn²⁺ with or without 2 mm Ca²⁺; a representative trace from another neurone in the same culture dish, recorded in the presence of 2 mm Ca²⁺ alone, is illustrated for comparison. Dotted lines indicate the zero current level. * Significant difference between current amplitude in the indicated condition and control currents evoked in the presence of Zn²⁺ alone; † significant difference in current amplitudes recorded in the presence of 2 mm Ca²⁺ with versus without 2 mm Zn²⁺ (P < 0·05, one-way ANOVA with Bonferroni's t test). B, inward currents, evoked by voltage steps from −70 to 0 mV, were recorded from 4 cells in the presence of 2 mm Ca²⁺ plus varying concentrations of Zn²⁺. Steady-state current amplitude is plotted relative to currents evoked in the presence of 2 mm Ca²⁺ alone (left). Points represent means ± s.e.m. The points were fitted by a curve described by the equation:

where IC₅₀ = 210 μm and m = −1·2. Superimposed traces from a representative cell illustrate the effects of 300 μm Zn²⁺ on current recorded in the presence of 2 mm Ca²⁺ (right).

Since it is impossible to assign to Zn²⁺ a defined portion of the inward current evoked in the presence of extracellular Ca²⁺ plus Zn²⁺, we turned to Zn²⁺-selective fluorimetric imaging to verify that Zn²⁺ can permeate voltage-gated Ca²⁺ channels in the presence of physiological extracellular cation concentrations. Cortical neurones were loaded with the Zn²⁺-selective fluorescent probe Newport Green, which is capable of discriminating Zn²⁺ from other divalent cations, including Ca²⁺ and Mg²⁺ (Canzoniero et al. 1999; Sensi et al. 1999). When Newport Green-loaded neurones were exposed to depolarising conditions (60 mm KCl, 2 min) in the presence of physiological extracellular cation concentrations plus 300 μm Zn²⁺, a marked increase in [Zn²⁺]_i was detected. The addition of nimodipine (10 μm) substantially attenuated the KCl-induced [Zn²⁺]_i increase, confirming that Zn²⁺ entry occurred predominantly through L-type Ca²⁺ channels (Fig. 7).

Figure 7.

Voltage-gated Ca²⁺ channels mediated depolarisation-induced rises in [Zn²⁺]_i in physiological saline

A, cortical neuronal cultures loaded with Newport Green and bathed in Hepes-buffered saline (see Methods) were exposed for 2 min to either high-potassium + 300 μm Zn²⁺ (control) or high-potassium + 300 μm Zn²⁺ + 10 μm nimodipine (high-potassium solutions contained 60 mm KCl, 65 mm NaCl, and were otherwise identical to the Hepes-buffered saline). MK-801 (10 μm) and NBQX (10 μm) were present throughout experiments to block glutamate receptor activation by endogenous glutamate release. Pseudo-colour images from a representative experiment depict changes in [Zn²⁺]_i according to the Newport Green fluorescence values indicated in the scale. The fluorescence signal attributed to increased [Zn²⁺]_i was completely quenched by the addition of the selective membrane-permeable Zn²⁺ chelator, 100 μm TPEN (not shown). Scale bar = 50 μm. B, each trace represents the change in [Zn²⁺]_i observed in the same experiments described in A (control, n = 57 cells; nimodipine (Nimo), n = 52). Fluorescence values were converted to Zn²⁺ concentrations by a calibration performed at the end of the experiment (see Methods).

Zn²⁺ current is enhanced by extracellular acidity

Lowering extracellular pH (pH_o) suppresses HVA Ca²⁺ channel currents carried by Ca²⁺, Ba²⁺, or monovalent cations in a variety of cell types including hippocampal neurones (e.g. Tombaugh & Somjen, 1996), and we examined the effect of this manoeuvre on inward currents carried by 1 mm Ca²⁺ or 500 μm Zn²⁺. In these experiments, intracellular pH was buffered to 7·4 with 10 mm Hepes (Table 1). Consistent with previous studies, lowering pH_o from 7·4 to 6·4 reversibly decreased Ca²⁺ current amplitude at 0 mV by 29 ± 2 % (n = 7). In contrast, the same drop in pH_o reversibly enhanced Zn²⁺ current amplitude at 0 mV by 92 ± 10 % (n = 5; Fig. 8A), and raising pH_o from 7·4 to 8·4 resulted in near-complete suppression of the Zn²⁺ current. Extracellular pH affected Zn²⁺ current in a graded fashion with an approximate pK of 7·4 (Fig. 8C). Compared to currents at pH_o 7·4, Zn²⁺ currents recorded at pH_o 6·4 were similarly sensitive to blockade by 10 μm Gd³⁺ (currents were suppressed by 86 ± 5 %; n = 4; Fig. 8B) and had a similar time constant of Zn²⁺ current activation (Table 2).

Figure 8.

Extracellular acidity enhanced Zn²⁺ current

A, representative traces elicited by a voltage step to 0 mV in the presence of 500 μm Zn²⁺ show reversible changes in steady-state current amplitude at pH_o 6·4 compared to pH_o 7·4. Cells were exposed to a new extracellular buffer solution for 30–60 s between trials. This change in pH_o resulted in no change in the linearity of membrane currents throughout the range of voltages used by the leak subtraction protocol (see Methods; data not shown). B, representative traces recorded as in A at pH_o 6·4 show reversible blockade by 10 μm Gd³⁺. C, steady-state Zn²⁺ current amplitudes at various test pH_o values are plotted relative to control currents at pH_o 7·4. Voltage steps to 0 mV were delivered to cells in the presence of 500 μm Zn²⁺; a cell was perfused with a new extracellular buffer solution for 30–60 s between trials. Points represent means ± s.e.m. of 3–6 cells per condition and were fitted with a curve described by the equation:

where (I/I_7·4)_max = 2·0, pK = 7·4, and m = −1·7.

We examined the effects of varying pH_o on the I–V relationship of the Zn²⁺ current. A decrease in pH_o was accompanied by an increase in inward steady-state Zn²⁺ current amplitude at all membrane potentials as well as a positive shift in reversal potential (n = 12; Fig. 9A) from 29 mV at pH_o 7·4 to 59 mV at pH_o 6·4 (approximated by fitting curves to the I–V data; see Methods), although the voltage dependence of the estimated steady-state Zn²⁺ current activation was similar in both conditions, with V_h being −13 mV at pH_o 7·4 and −11 mV at pH_o 6·4 (Fig. 9B). These data suggest that the enhanced Zn²⁺ current observed at the lower pH_o was not due to an increase in the activity of free Zn²⁺ available for voltage-gated Ca²⁺ channel permeation, as an increase in [Zn²⁺]_o sufficient to double peak inward current would be expected to cause a considerable shift in V_h (see Fig. 4). Consistent with the hypothesis that the observed relationship between Zn²⁺ current amplitude and pH_o was not explained by changes in Zn²⁺ activity, the amount of zinc recoverable from solutions, which were prepared in the same plastic containers used in experiments and exposed to cultured cortical neurones, was no different at pH 6·4 than at 7·4 (Table 3).

DISCUSSION

We demonstrate here that extracellular Zn²⁺ carried a voltage-activated inward current in murine cortical neurones exhibiting properties consistent with mediation by HVA Ca²⁺ channels. Under the recording conditions utilised, only extracellular Zn²⁺ or intracellular Cl⁻ were available to carry inward current, and current amplitude increased with increasing extracellular ZnCl₂. Consistent with the propensity of divalent cations to screen negative membrane surface charges (Hille, 1992), increasing [Zn²⁺]_o increased the threshold and peak potentials of the current's I–V relationship to a degree consistent with similar observations of Zn²⁺ currents in Helix neurones (Oyama et al. 1982). Relative to Ca²⁺ currents in these neurones, Zn²⁺ current was smaller and showed slower activation kinetics, and unlike Ca²⁺ currents, Zn²⁺ current did not exhibit voltage-dependent steady-state inactivation. In addition, Zn²⁺ was able to permeate voltage-gated Ca²⁺ channels in the presence of Ca²⁺. Unexpectedly, Zn²⁺ current amplitude increased with decreasing pH_o, whereas HVA Ca²⁺ currents were suppressed by the same manoeuvre. Zn²⁺ current enhancement at lower pH_o was not accompanied by changes in activation kinetics or in the voltage sensitivity of channel gating.

Zn²⁺ current was mediated by voltage-gated Ca²⁺ channels

Supporting a role for HVA Ca²⁺ channels in mediating voltage-activated Zn²⁺ current, the I–V relationship of the current exhibited a profile typical for these channels (Fig. 1), and the current was sensitive to L- and N-type voltage-gated Ca²⁺ channel blockers (Figs 5 and 7). Nimodipine blocked a majority of Zn²⁺ entry into cortical neurones as measured both by electrophysiology and fluorescent dye imaging, suggesting that L-type channels carry the bulk of voltage-activated Zn²⁺ influx. ω-Conotoxin GVIA blocked a smaller proportion of the current, suggesting a smaller role for N-type channels. These data do not permit a calculation of the relative permeabilities of L- versus N-type channels to Zn²⁺, as the magnitude of the pharmacological effects is also affected by differences in the population sizes and membrane distribution of these channel subtypes.

As noted above, previous electrophysiological studies of the effects of Zn²⁺ on mammalian voltage-gated Ca²⁺ channels have focused on the ability of Zn²⁺ to serve as a channel blocker (Winegar & Lansman, 1990; Büsselberg et al. 1992). However, Zn²⁺ permeation and blockade of current carried by other ions through voltage-gated Ca²⁺ channels may occur simultaneously, as has been proposed, for instance, in the interactions of Mg²⁺ (Ascher & Nowak, 1988) and Zn²⁺ (Christine & Choi, 1990) with NMDA receptor-gated channels. Winegar & Lansman (1990) showed that Zn²⁺ antagonises Ba²⁺ currents through single dihydropyridine-sensitive Ca²⁺ channels in mouse myotubes in a voltage-dependent manner, suggesting that Zn²⁺ binds in or near the channel pore. This blockade was characterised by an increase in open channel noise, presumably reflecting fast, unresolved blocking and unblocking of the channel, and calculations based on the kinetics of the blockade suggested that Zn²⁺ binds to open channels and has an exit rate from the pore faster than that of Mg²⁺, Cd²⁺, or La³⁺. In addition, although the entry rate of Zn²⁺ into the Ca²⁺ channel pore was linearly dependent on [Zn²⁺]_o, the exit rate was independent of Zn²⁺ concentration, suggesting, in agreement with data in the present study, that as Zn²⁺ exits a Ca²⁺ channel pore, it leaves from the cytoplasmic face.

Extracellular acidity enhanced Zn²⁺ current

In this study, decreasing pH_o was found to have opposite effects on inward currents carried by 500 μm Zn²⁺ and by 1 mm Ca²⁺. The enhancement of Zn²⁺ current observed at acidic pH was not accompanied by a shift in the voltage sensitivity of the current (Fig. 9), as was seen when [Zn²⁺]_o was increased (Fig. 4), arguing against the possibility that this enhancement was due to an increase in free Zn²⁺. In agreement with that observation, the amount of zinc recoverable from recording solutions was not affected by pH (Table 3).

The ability of decreasing pH_o to suppress HVA Ca²⁺ channel currents carried by Ca²⁺, Ba²⁺, or monovalent cations is well established, and appears to reflect some combination of two underlying mechanisms. First, experiments on single dihydropyridine-sensitive voltage-gated Ca²⁺ channels have suggested the existence of a low-conductance state for L-type channels that is stabilised by protonation and destabilised by the binding of permeant cations (Prod'som et al. 1987; Pietrobon et al. 1988). Second, H⁺ may screen negative surface charges and thus alter the voltage dependence of channel gating (e.g. Iijima et al. 1986; Prod'som et al. 1989).

Additional studies will be needed to delineate the mechanisms responsible for the paradoxical effect of lowering pH_o to enhance Zn²⁺ current through HVA Ca²⁺ channels. Perhaps H⁺ permeates voltage-gated Ca²⁺ channels, and increased Zn²⁺ permeation does not account for the increased current amplitude. Arguing against this possibility is the observation that in the presence of monovalent cations as charge carriers, H⁺ affects single L-type Ca²⁺ channels without apparently interacting with the channel pore (Prod'som et al. 1987). Some controversy surrounds this point, as recent work has shown that the sensitivity of L-type Ca²⁺ channels to pH_o can be altered by mutagenesis of the glutamate residues that form the high-affinity Ca²⁺ interaction site in the channel pore (Chen et al. 1996), suggesting that pore interactions may indeed account for the ability of H⁺ to elicit changes in channel conductance. However, single channel studies have revealed no direct competition between permeant cations and H⁺ for entry into the channel (Pietrobon et al. 1988; Prod'som et al. 1989); even if such competition were possible, Zn²⁺, with a relatively high affinity for voltage-gated Ca²⁺ channel pore binding sites, might be expected to exclude H⁺ from the pore even more effectively than other permeant cations.

Alternatively, extracellular H⁺ may allosterically influence the permeability of voltage-gated Ca²⁺ channels (Prod'som et al. 1987, 1989; Pietrobon et al. 1988, 1989), as is the case, for instance, for NMDA receptor-gated channels (Tang et al. 1990; Traynelis & Cull-Candy, 1990). A H⁺-induced conformational change in voltage-gated Ca²⁺ channels may increase their permeability to Zn²⁺ while decreasing their permeability to Ca²⁺ and monovalent cations. The increase in reversal potential observed in Zn²⁺ currents at lower pH_o is consistent with an increase in permeability to Zn²⁺ relative to permeant intracellular cations, which, in these experiments, may include Na⁺ or Li⁺ (see Fig. 1 legend and Table 1). The fact that the net membrane conductance g (sum of all conductances when multiple permeant ions are present; see Methods), as estimated from curve fits to I–V data (Fig. 9A), did not change with a change in pH_o is consistent with a simple model in which a small absolute increase in Zn²⁺ conductance is counterbalanced by a comparable decrease in absolute conductance to permeant intracellular cations, although other scenarios cannot be excluded. Regardless of the exact underlying mechanism, the observed ability of extracellular acidity to enhance voltage-gated Ca²⁺ channel permeability to Zn²⁺ suggests that H⁺ may have more complex effects on ion permeation through voltage-gated Ca²⁺ channels than previously thought.

Physiological significance

Both electrophysiological and fluorimetric evidence is presented supporting the notion that Zn²⁺ can permeate voltage-gated Ca²⁺ channels in the presence of physiological amounts of other cations, including Ca²⁺. First, unlike the effect of Ca²⁺ on voltage-gated Ca²⁺ channel-mediated Ba²⁺ currents (Hille, 1992), no blockade of Zn²⁺ current was evident at low [Ca²⁺]_o; net inward current amplitude in the presence of 2 mm extracellular Zn²⁺ was unaffected by up to equimolar amounts of extracellular Ca²⁺. When [Ca²⁺]_o exceeded [Zn²⁺]_o, at least some of the inward current was probably carried by Ca²⁺, but the presence of Zn²⁺ clearly impeded voltage-gated Ca²⁺ channel permeability to Ca²⁺. As in other cell types, Zn²⁺ blocked cortical neuronal voltage-gated Ca²⁺ currents. Second, neurones loaded with the fluorescent Zn²⁺ probe Newport Green exhibited depolarisation-evoked increases in [Zn²⁺]_i in the presence of physiological concentrations of extracellular cations plus 300 μm extracellular Zn²⁺. Depolarisation-evoked Zn²⁺ influx was largely inhibited by nimodipine, implicating L-type Ca²⁺ channels as the major mediators of membrane permeability to Zn²⁺. In these experiments, [Ca²⁺]_o exceeded [Zn²⁺]_o by sixfold.

The demonstration that murine neuronal voltage-gated Ca²⁺ channels can carry an inward Zn²⁺ current supports the notion that blockade of these channels may have therapeutic utility in pathological conditions, such as cardiac arrest or sustained seizures, where excessive Zn²⁺ influx may contribute to neuronal death (Choi & Koh, 1998). Although the absolute permeability of voltage-gated Ca²⁺ channels to Zn²⁺ is low compared to Ca²⁺, these channels exhibited much less voltage-dependent steady-state inactivation when Zn²⁺, instead of Ca²⁺, was used as the charge carrier (Fig. 2). These data suggest that voltage-gated Ca²⁺ channels could indeed mediate the sustained increases in [Zn²⁺]_i observed in conditions where Zn²⁺ is toxic to cortical neurones (Canzoniero et al. 1999). In addition, data presented here raise the novel possibility that the neurotoxic contribution of Zn²⁺ influx through HVA Ca²⁺ channels may be specifically enhanced by increases in extracellular acidity caused by the build-up of brain lactate associated with ischaemic conditions.