Cytosolic zinc release and clearance in hippocampal neurons exposed to glutamate – the role of pH and sodium

Abstract

Although Zn²⁺ homeostasis in neurons is tightly regulated and its destabilization has been linked to a number of pathologies including Alzheimer's disease and ischemic neuronal death, the primary mechanisms affecting intracellular Zn²⁺ concentration ([Zn²⁺]_i) in neurons exposed to excitotoxic stimuli remain poorly understood. The present work addressed these mechanisms in cultured hippocampal neurons exposed to glutamate and glycine (Glu/Gly). [Zn²⁺]_i and [Ca²⁺]_i were monitored simultaneously using FluoZin-3 and Fura2-FF and intracellular pH (pH_i) was studied in parallel experiments using 2',7'-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein. Glu/Gly applications under Na⁺-free conditions (Na⁺ substituted with N-methyl-D-glucamine⁺) caused Ca²⁺ influx, pH_i drop, and Zn²⁺ release from intracellular stores. Experimental maneuvers resulting in a pH_i increase during Glu/Gly applications, such as stimulation of Na⁺-dependent pathways of H⁺ efflux, forcing H⁺ efflux via gramicidin-formed channels, or increasing extracellular pH counteracted [Zn²⁺]_i elevations. In the absence of Na⁺, the rate of [Zn²⁺]_i decrease could be correlated with the rate of pH_i increase. In the presence of Na⁺, the rate of [Zn²⁺]_i decrease was about twice as fast as expected from the rate of pH_i elevation. The data suggest that Glu/Gly-induced cytosolic acidification promotes [Zn²⁺]_i elevations and that Na⁺ counteracts the latter by promoting pH_i-dependent and pH_i-independent mechanisms of cytosolic Zn²⁺ clearance.

Introduction

Increasing evidence indicates that destabilization of Zn²⁺ homeostasis can play a major role in neuronal death in vitro and in vivo (Yokoyama et al. 1986; Choi et al. 1988; Tonder et al. 1990; Koh et al. 1996; Aizenman et al. 2000; Lee et al. 2000; Sheline et al. 2000; Medvedeva et al. 2009; Sensi et al. 2009). Chelation of extracellular Zn²⁺ was neuroprotective in animal models of stroke (Koh et al. 1996; Lee et al. 2002) and global brain ischemia (Calderone et al. 2004). Recently, Zn²⁺ influx was found to promote spreading depression in brain slices (Dietz et al. 2008) and elevated extracellular Zn²⁺ levels have also been implicated in the formation of Alzheimer beta-amyloid plaques (Bush 2003; Sensi et al. 2009). However, the mechanisms maintaining Zn²⁺ homeostasis in neurons remain poorly understood. Two families of Zn²⁺ transporters are recognized, ZnT with the systematic name of SLC30 (Palmiter and Huang 2004) and Zip with the systematic name of SLC39 (Eide 2006). Among the ZnT transporters expressed in the brain, the best characterized is localized in the plasma membrane ZnT-1 (Palmiter and Findley 1995). The remaining ZnTs are found in intracellular compartments (Beyersmann and Haase 2001; Eide 2006), usually associated with endosomes, Golgi apparatus, or endoplasmic reticulum (Cousins et al. 2006), or in the case of ZnT-3, synaptic vesicles (Palmiter 2004). The molecular mechanism of ZnT-mediated transport involves an exchange of Zn²⁺ for H⁺ or K⁺ (Guffanti et al. 2002; Chao and Fu 2004; Ohana et al. 2009).

Transcripts of several Zip family members, Zip1, Zip9, Zip10, and Zip14, are robustly expressed in the mouse brain (Gyulkhandanyan et al. 2006) and Zip1 and Zip3 were recently found in neurons (Qian et al. 2011). While the cellular localization of the Zip proteins is yet unclear, in non-neuronal cells, Zip2 and Zip4 have been found in plasma membranes (Gaither and Eide 2000; Dufner-Beattie et al. 2003). The mechanism of Zn²⁺ transport by Zip2 involves a Zn²⁺-HCO₃⁻ co-transport (Gaither and Eide 2000) and the Zip proteins generally mediate Zn²⁺ influx (Eide 2006; Sensi et al. 2009). For example, Zip4 mediates zinc influx leading to lysosomal Zn²⁺ sequestration (Emmetsberger et al. 2010).

Early studies suggested that plasmalemmal Na⁺/Ca²⁺ exchangers may transport Zn²⁺ (Sensi et al. 1997; Cheng and Reynolds 1998) and the inhibition of the Na⁺/Ca²⁺ exchange by Zn²⁺ was also reported (Colvin 1998). However, Ohana et al. (2004) described what appears to be a genuine Na⁺/Zn²⁺ exchanger in HEK 293 cells and in cultured cortical neurons. Also in cultured cortical neurons, Qin et al. (2008) described distinct mechanisms of Na⁺-dependent Zn²⁺ efflux that were inhibited by 10 μM La³⁺ and Ca²⁺-free medium. None of the Na⁺/Zn²⁺ exchangers has been yet cloned.

Activation of glutamate receptors in cultured neurons leads to intracellular Zn²⁺ concentration ([Zn²⁺]_i) elevations representing a Ca²⁺-dependent Zn²⁺ release from intracellular stores (Sensi et al. 2003; Dineley et al. 2008). The present work tested the impact of Na⁺ and pH_i on Glu/Gly-induced [Zn²⁺]_i elevations. It was found that the [Zn²⁺]_i elevations critically depend on the Glu/Gly-induced pH_i drop and that the Na⁺-induced pH_i increase plays an important role in cytosolic Zn²⁺ clearance.

Materials and Methods

Measurement of Zn²⁺ concentration in Locke's buffer

Locke's buffer contained (in mM) NaCl (157.6), KCl (2.0), KHCO₃ (3.6), MgCl₂ (1.0), CaCl₂ (1.3), HEPES (10), glucose (5), and pH 7.4 adjusted with Tris. To measure [Zn²⁺], the buffer was supplemented with 10 nM FluoZin-3 tetrapotassium salt, and F488 fluorescence (488 nm excitation, 510 nm emission) was monitored at 25°C using a PTI QuantaMaster spectrofluorometer (Photon Technology International, Inc., Birmingham, NJ). Minimal F488 (Fmin), and maximal F488 (Fmax) values were measured after adding 100 μM N,N,N',N'-tetrakis(2-pyridalmethyl)ethylenediamine (TPEN) or 100 μM ZnSO₄, respectively, to the buffer. The formula [Zn²⁺] = Kd (F488 – Fmin) / (Fmax – F488) with the Kd value that was assumed to be 15 nM (Gee et al. 2002) was used to calculate Zn²⁺ concentration in the buffer.

The Fura-2FF isosbestic point and the effects of pH on Fura-2FF fluorescence

A solution of Fura-2FF K⁺ salt (1 μM) in a buffer containing (in mM) KCl (159.6), KHCO₃ (3.6), HEPES (10), and pH 7.4 adjusted with Tris was supplemented with 1 mM EGTA, or 1 mM CaCl₂, and/or 0.1 mM ZnCl₂ to obtain a Ca²⁺/Zn²⁺-free, a Ca²⁺-containing, a Zn²⁺-containing, or a Ca²⁺ and Zn²⁺-containing solution of Fura-2FF, respectively. Using the PTI QuantaMaster spectrofluorometer (Birmingham, NJ), excitation spectra (300 – 400 nm excitation, 515 nm emission) were obtained at 25°C. The excitation wavelength at which the spectra of Ca²⁺-containing and Zn²⁺-containing solutions intersected the spectrum of Ca²⁺/Zn²⁺-free solution was taken as the respective isosbestic point.

To determine the impact of the pH drop from 7.4 to 5.5 on Fura-2FF fluorescence, analogous (300 – 400 nm excitation, 515 nm emission) spectra of 1 μM Fura-2FF (K⁺ salt) dissolved in the above described but nominally Ca²⁺-free (not supplemented with EGTA), or supplemented with 30 μM CaCl₂ buffer were obtained at pH 7.4 and 5.5. First the spectrum at pH 7.4 was measured, than 1 μl of 5M HCl per 1 ml of the solution was added to the cuvette, which dropped the pH to 5.5 as determined using a pH meter and the spectrum was measured again.

Primary cultures of mouse hippocampal neurons

The cultures were plated using a SPOT™ culture kit from the University of Illinois at Chicago Research Resources Center (http://www.rrc.uic.edu/portal/SPOT_Culture_Kit). The kit contains cryopreserved hippocampal neurons from embryonic day 16 c57/bl/6 mice. The neurons were cultured in circular spots of about 4 mm diameter in the center 35 mm plastic dishes (BD Falcon™ 353001), according to the manufacturer's instructions. For the ion imaging experiments, the SPOT™ cultures were plated in a center of 25 mm glass coverslips (Fisher Scientific 12-545-86) placed in 35 mm dishes. After plating, the cells were cultured at 37°C in 2 ml of Neurobasal Medium (Gibco Invitrogen, Carlsbad, CA) supplemented with 2% B-27 (Gibco Invitrogen, Carlsbad, CA) and 2 mM glutamine in a humidified incubator (MCO-5M, Sanyo, Wood Dale, IL) maintaining 5% CO₂ and 5% O₂ balanced with N₂. Two days after plating, glial proliferation was curtailed by adding 10 μg/ml of 5-fluoro-2'-deoxyuridine (FURD) to the cultures. The next day and once a week thereafter, half the culture medium was replaced using an analogous medium but containing 2 mM glutamax (Gibco Invitrogen, Carlsbad, CA) instead of glutamine and not supplemented with FURD. Cells were used for experiments after 13 days in vitro.

Monitoring Fura2-FF and FluoZin-3 fluorescence in neurons

Coverslips with SPOT™ cultures were placed in custom-made 50 μl imaging chambers and the cells were loaded at 37°C with 0.1 μM FluoZin-3 AM and 0.1 μM Fura-2FF AM for 5 – 6 minutes. The concentrations of the indicators and the time of loading were adjusted such that at the end of the experiments during in situ calibration, fluorescent signals did not exceed the dynamic range of the camera.

Imaging chambers with loaded cells were placed on a Zeiss Axiovert 100 microscope stage and prior to monitoring fluorescence, a Hoffman modulation contrast image of the cells was taken using a Zeiss LD A-Plan 20×/0.3 HMC objective. Fluorescence was monitored using a Zeiss Fluar 20×, NA 0.75 objective and an Attofluor digital imaging system (Atto Instruments, Rockville, MD). Superfusion media were delivered directly onto the cells via an 8-channel manifold (MPRE-8, Cell MicroControls, Norfolk, VA) with a computer-controlled flow using an 8-channel valve switch (cFlow8, Cell MicroControls). Temperature was maintained at 37°C using a bipolar temperature controller (TC2BIP, Cell MicroControls).

For time-lapse imaging, cells were exposed every 5 seconds to a sequence of 340, 380, 488 and 356 nm excitation and the images of fluorescence emitted at >520 nm (F340, F380, F488, and F356) were saved on a computer hard-drive for off-line analysis. Cells were superfused at a rate of 0.5 ml/min, initially with Locke's buffer and then with the experimental media. Glutamate receptors were activated using Mg-free Locke's buffer supplemented with 100 μM glutamate and 10 μM glycine (Gly/Glu). Na⁺-free buffers were made by replacing NaCl with chloride salts of N-methyl-D-glucamine (NMDG) or lithium or cesium, as indicated in the text. In Ca²⁺- and Zn²⁺-free buffers, CaCl₂ was omitted and 100 μM EGTA was added. To remove Zn²⁺ but not Ca²⁺ from the buffers, 1 mM CaEDTA was used.

All fluorescence data were background-corrected by subtracting the fluorescence measured in cell-free areas from the fluorescence measured in the regions of interest positioned on neuronal somata. At the end of each experiment, the maximal F340/F380 ratio (Rmax) of Fura-2FF and the maximal F488 (Fmax) of FluoZin-3 was measured during “Ca saturation” and “Zn saturation,” respectively. During Ca saturation, Locke's buffer was supplemented with 10 μM ionomycin, 10 μM TPEN, 2 μM carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) and 10 mM CaCl₂. During Zn saturation, Locke's buffer was supplemented with 20 μM 1-hydroxypyridine-2-thione zinc salt and 100 μM ZnCl₂.

In some experiments, an attempt was made to calibrate FluoZin-3 fluorescence in terms of [Zn²⁺]_i. In these experiments, before performing Zn saturation to measure Fmax, minimal F488 (Fmin) was measured. To this end, the cells were exposed to 10 μM TPEN. [Zn²⁺]_i was calculated as described above (measurement of Zn²⁺ concentration in Locke's buffer). In a majority of the experiments, the FluoZin-3 and Fura-2FF data were routinely expressed as a percentage of Fmax and Rmax, respectively, for the reasons explained in the text.

To compare the rates of cytoslolic Zn²⁺ clearance following 5 min exposure to Glu/Gly under Na⁺-free conditions, F488 data were normalized according to the formula: normalized F488 = (F488 – F488_basal)*100/(F488_5min – F488_basal), where F488_basal is the fluorescence measured before Glu/Gly application and F488_5min is the fluorescence measured at the end of the 5th minute of exposure to Glu/Gly, just before the Zn²⁺ clearance-inducing media described in the text were applied. The rates of F488 drop were calculated with the linear regression method using the linear portion of the data.

Monitoring intracellular pH (pH_i) in neurons

Coverslips with SPOT™ cultures were placed in custom-made 50 μl imaging chambers and the cells were loaded for 5 minutes at 37 °C with 1 μM BCECF-AM. After the loading, the cells were superfused and the fluorescence was measured using the above described instruments. For time-lapse imaging, cells were exposed every 5 seconds to 488 and 440 nm excitation and the images of fluorescence emitted at >520 nm (F488 and F440) were saved on a computer hard-drive for off-line analysis. In most experiments, the background-corrected F488/F440 ratio was converted to pH_i values. To this end, a four point in situ calibration in the pH range of 5.8 – 8.0 was performed at the end of the experiments. The pH calibrating solutions contained 5μM monensin, 5μM nigericin, 2μM FCCP, 134.2 mM K-gluconate, 25.4 mM KCl, 1.3 mM CaCl₂, 1 mM MgCl₂, 3.6 mM KHCO₃, and 10 mM 2-[N-morpholino]ethanesulfonic acid (MES), or 10 mM PIPES or 10 mM HEPES. MES was used to adjust the pH of the calibrating solutions to a range of 5.5 to 6.5; PIPES to a range of 6.6 –7.0 and HEPES to a range of 7.2 to 8.0. Using SigmaPlot 10.0 software (Systat Software Inc., Richmond, CA), the data were fitted to the four-parameter equation of a sigmoidal curve: F488/F440 = y + a /(1+ exp(−(pH−x)/b)). The rates of pH_i increase were calculated with the linear regression method using the linear portion of the data.

Statistical analysis

A SigmaStat 3.5 package (Systat Software Inc., Richmond, CA) was used for statistical analysis. Normally distributed data were analyzed using one-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls test. Data that were not normally distributed were analyzed using a Kruskal-Wallis one-way ANOVA on ranks followed by Dunn's test.

Reagents

Fura-2FF AM and Fura-2FF (K⁺ salt) were obtained from Teflabs (Austin, TX, USA). Neurobasal medium, B-27 supplement, BCECF-AM, FluoZin-3 AM and FluoZin-3 tetrapotassium salt were from Invitrogen (Carlsbad, CA, USA). All other reagents were from Sigma-Aldrich (St Louis, MO, USA) unless otherwise stated.

Results

Simultaneous monitoring of FluoZin-3 and Fura-2FF fluorescence

Fig. 1 illustrates the data generation and processing method used in this report. At the beginning of each experiment, the neuronal morphology of the cells was verified using the Hoffman modulation contrast (Fig. 1a, upper panel). At the end of each experiment, the maximal Fura-2FF F340/F380 ratio (Rmax) was measured in each cell after saturating Fura-2FF with Ca²⁺ (Ca saturation), which was followed by measuring the maximal FluoZin-3 F488 signal (Fmax) after saturating FluoZin-3 with Zn²⁺ (Zn saturation). The lower panel of Fig. 1a shows a F488 image of the cells captured during Zn saturation. The cells were exposed for 4 minutes to 100 μM glutamate and 10 μM glycine (Glu/Gly) under Na⁺-and Zn²⁺-free conditions (Na⁺ substituted with NMDG⁺ and extracellular zinc chelated by CaEDTA). Upon Glu/Gly application, both F340/F380 ratio and F488 began to increase. When 10 μM TPEN (a plasma membrane-permeable chelator) was added to chelate the intracellular zinc, F488 promptly dropped in all cells but the response in F340/F380 ratio varied greatly among the cells. In 50 neurons tested (3 experiments), the pattern of F340/F380 ratio response could be divided into three categories: category 1 (17 neurons, 34%) – the F340/F380 ratio failed to drop or kept increasing; category 2 (14 neurons, 28%) – the F340/F380 ratio initially dropped but then started to increase; category 3 (19 neurons, 38%) – the F340/F380 ratio dropped and remained low. Exemplar cells representing the three categories are indicated in the lower panel of Fig. 1a, and data from these cells are presented in Fig. 1c. Cells A (red) and B (green) represent category 1, cell C (blue) represents category 2, and cell D (orange) category 3.

Figure 1.

Simultaneous monitoring of FluoZin-3 and Fura-2FF fluorescence in single hippocampal neurons. (a) Upper panel, Hoffman modulation contrast image of 14 hippocampal neurons cultured for 14 days. Bar = 100 μm. Lower panel, F488 fluorescence during Zn saturation in the same neurons. Indicated A, B, C, D are four exemplar neurons that are discussed in the text. (b) Fura-2FF F340/F380 ratio and Fluo-Zin-3 fluorescence (F488) data from all 14 neurons shown in (a). The neurons were exposed to 100 μM glutamate and 10 μM glycine (Glu/Gly) under Na⁺-free conditions with Na⁺ substituted with N-methyl-D-glucamine (NMDG⁺) and in the presence of 1 mM CaEDTA to chelate extracellular Zn²⁺. Where indicated, 10 μM TPEN was added to chelate intracellular Zn²⁺. At the end of the experiment, Fura-2FF was saturated with Ca²⁺ to measure the maximal F340/F380 ratio (Rmax) followed by saturation of FluoZin-3 with Zn²⁺ to measure the maximal F488 fluorescence (Fmax). (c) Fluorescence data from the four exemplar cells indicated in (a). (d) Normalized Fura-2 FF and FluoZin-3 data expressed as a percentage of Rmax and Fmax, respectively.

Since the Glu/Gly-induced FluoZin-3 signal increase took place in the absence of extracellular Zn²⁺ (in the presence of 1 mM CaEDTA), the [Zn²⁺]_i elevations represent Zn²⁺ release from intracellular stores (Fig. 1c lower panel). Dineley et al. (2008) recently reported that such an intracellular Zn²⁺ release depends on the Glu/Gly-induced Ca²⁺ influx. Therefore, one would expect that cells showing a faster rate and/or onset of [Ca²⁺]_i (F340/F380 ratio) elevation would show a faster rate of [Zn²⁺]_i (F488) increase. However, looking at the raw FluoZin-3 F488 data, this was not the case. Note that in cell A (red) and cell B (green), there is an F340/F380 ratio data overlap, which suggests that [Ca²⁺]_i increases in both cells at the same rate. However, the rate of F488 increase is much faster in cell B than in cell A, suggesting that more Zn²⁺ is released in cell B (Fig. 1c). On the other hand, the F488 traces in cells A, C, and D overlap despite the very different patterns of F340/F380 ratio increases in these cells (Fig. 1c), suggesting that the rates of [Zn²⁺]_i and [Ca²⁺]_i elevations in these cells are independent from each other. Interestingly, Fmax values measured during Zn saturation greatly differed among the cells. For example, Fmax was much higher in cell B than in cell A (Fig. 1c bottom panel). Since FluoZin-3 fluorescence intensity during Fmax measurement is a function of FluoZin-3 concentration, the data indicate that cell B loaded more FluoZin-3 than cell A. Such uneven FluoZin-3 loading among the cells creates an artifact. The cells that loaded more of the indicator elevated F488 at a faster rate, which could be misinterpreted to indicate that [Zn²⁺]_i elevates faster in these cells. This artifact could be normalized for by expressing F488 data as a percentage of Fmax. After such normalization, the rates of FluoZin-3 fluorescence increase could be related to the onset and/or rate of [Ca²⁺]_i increase, the fastest in cells A and B and the slowest in cell D (Fig. 1d).

In neurons in which Glu/Gly-induced elevations of Fura-2FF signal were small (did not exceed 30% of Rmax), TPEN applications promptly decreased the signal, indicating that Zn²⁺ rather than Ca²⁺ contributed to the signal (for example cells C and D in Fig. 1c). Since in this case the Fura-2FF signal represented primarily [Zn²⁺]_i elevations, no attempt was made to calibrate the Fura-2FF data in terms of [Ca²⁺]_i. Instead, to relate the Fura-2FF signal to the maximal signal, the F340/F380 data were expressed as a percentage of Rmax (Fig. 1d upper panel). When Zn saturation was performed after Ca saturation, the Fura-2FF F340/F380 signal dropped to about 30% of Rmax (Fig. 1b,c,d upper panel). This drop was expected because Devinney II et al. (2005) already demonstrated that Zn²⁺ and Ca²⁺ affect the Fura2-FF excitation spectrum differently, which is confirmed by the data shown in Supplementary Fig. 1a. The F340/F380 ratio of Fura-2FF saturated with Zn²⁺ in vitro (Supplementary Fig. 1b), similarly as the one measured in vivo (Fig. 1d), can reach only about 30% of the ratio measured when the indicator is saturated with Ca²⁺. These data suggest that when the Fura-2FF signal exceeds 30% of Rmax, the signal results primarily from Ca²⁺ binding to the indicator. This can explain why Zn²⁺ chelation with TPEN fails to affect the signal when the latter exceeds 30% of Rmax (Fig. 1c, cells A and B).

The expression of FluoZin-3 data as a percentage of Fmax does not take into account additional artifacts associated with the use of a nonratiometric [ion] indicator. Namely, fluorescence intensity is also affected by cell swelling/shrinking or dye bleaching or leakage from the cells during experimentation. Such artifacts can be accounted for by measuring fluorescence at the isosbestic point excitation of a ratiometric indicator, which provides a measure of indicator concentration independent of intracellular [ion]. Using the approach described in the Methods, the Ca²⁺ isosbestic point of Fura-2FF excitation was determined to be 356 nm (Supplementary Fig. 1a). To account for the artifacts associated with cell swelling, dye leakage, or bleaching, we considered expressing FluoZin-3 fluorescence data as a FluoZin-3 F488/Fura-2FF F356 ratio. This method would generate valid results only if both indicators are loaded, metabolized, and/or leaked at the same rates. If this were the case, there would be a very good correlation between FluoZin-3 Fmax and Fura-2FF F356 measured during Zn saturation. Fig. 2 shows that although on average the cells that loaded more FluoZin-3 also loaded more Fura-2FF, the correlation between Fura-2FF F356 and FluoZin-3 Fmax was not perfect, r²=0.483. Therefore, the idea of using FluoZin-3 F488/Fura-2FF F356 ratio as an index of [Zn²⁺]_i was abandoned.

Figure 2.

Correlation between FluoZin-3 Fmax and Fura-2FF fluorescence excited at the Ca²⁺ isosbestic point (F356) and measured during Zn saturation in 47 hippocampal neurons (3 coverslips).

Glu/Gly-induced [Zn²⁺]_i elevations depend on Ca²⁺ and Na⁺

Free Zn²⁺ available to the high affinity Zn²⁺ indicator FluoZin-3 in Locke's buffer was estimated to be 38 nM (for details see Methods). However, the extracellular Zn²⁺ did not affect the Glu/Gly-induced [Zn²⁺]_i elevations because chelation with 1 mM CaEDTA had no effect on Glu/Gly-induced elevations of the FluoZin-3 signal (Fig. 3a,c,f). Glu/Gly-induced [Zn²⁺] and [Ca²⁺]_i elevations were eliminated by preventing Ca²⁺ influx (Fig. 3b,e), which confirms the recent data of Dineley et al. (2008), indicating that exposure of neurons to Glu/Gly causes Zn²⁺ release from intracellular stores and that the Zn²⁺ release is triggered by Ca²⁺ influx. The Glu/Gly-induced [Ca²⁺]_i elevations were greatly enhanced when extracellular Na⁺ was substituted with NMDG⁺ (Fig. 3a,e). This outcome was expected, as under such conditions the Glu/Gly-induced destabilization of Ca²⁺ homeostasis is exacerbated by a number of factors. These factors include a lack of Ca²⁺ extrusion by plasmalemmal Na⁺/Ca²⁺ exchangers, a lack of competition between Na⁺ and Ca²⁺ for NMDA and other Na⁺- and Ca²⁺-permeable channels, and an increased electrochemical driving force for Ca²⁺ influx due to a lack of Na⁺-dependent depolarization of the plasma membrane (Mattson et al. 1989; Kiedrowski 1999). Interestingly, when Na⁺ was present in the medium, the Glu/Gly-induced elevation of FluoZin-3 F488 signal was also greatly decreased (Fig. 3d,e,f), and further experiments were designed to address the mechanisms responsible for this outcome.

Figure 3.

Glu/Gly-induced [Zn²⁺]_i elevations are Ca²⁺- and Na⁺-dependent. The effects of Glu/Gly on Fura-2FF (red) and FluoZin-3 (blue) fluorescence in hippocampal neurons were monitored as explained in Fig. 1, under Na⁺-free conditions with Na⁺ substituted with NMDG⁺ (a–c) or in the presence of 158 mM Na⁺ (d). In (a) and (d), 1.3 mM CaCl₂ was present in the medium. In (b), Ca²⁺ was omitted from the medium and 0.1 mM EGTA was added. In (c), the medium contained 1.3 mM CaCl₂ and 1 mM CaEDTA. The y-axes and the 5-min time bar apply to all panels. The data are means ± standard error from the indicated number of neurons monitored in a single experiment. All experiments were repeated at least 3 times with similar results. Panels (e) and (f) show Fura-2FF and FluoZin-3 signals, respectively, measured at the end of the 5th minute of Glu/Gly application. The data are means ± standard errors from 47 – 99 individual neurons (3 to 5 separate experiments). ** p < 0.01 versus Na⁺-free conditions; Kruskal-Wallis one-way ANOVA on Ranks followed by Dunn's test.

Na⁺-dependent [Zn²⁺]_i clearance

An addition of Na⁺ to cells that have been exposed to Glu/Gly under Na⁺-free conditions (NMDG⁺ was removed when Na⁺ was added) caused a prompt decrease of the FluoZin-3 F488 signal (Fig. 4a blue trace), whereas the response in Fura-2FF F340/F480 ratio greatly varied among the cells (Fig. 4a red trace). The Na⁺-induced F488 decrease represented a [Zn²⁺]_i decrease rather than a signal decrease due to cell swelling or dye leakage because a simultaneously monitored Fura-2FF F356 signal, the Ca²⁺ isosbestic point, was not affected (Fig. 4a green trace). Whereas the Fura2-FF F356 signal was not affected by Na⁺, there was a marked, 30 – 40%, drop in this signal during the first 4 minutes of Glu/Gly application. Since [Zn²⁺]_i was increasing during that time (blue trace), one might be concerned that the drop in F356 was caused by Zn²⁺ binding to Fura-2FF. Such a decrease in F356 signal due to Zn²⁺ binding to Fura-2FF is unlikely, however, because the Zn²⁺ isosbestic point of Fura-2FF was found to be 364 nm (Supplementary Fig. 1a). Consequently, Zn²⁺ binding to Fura-2FF increases the fluorescence excited at the wavelengths lower than 364 nm, namely the F356 – the Ca²⁺ isosbestic point. Since the latter was decreasing, the decrease could not be caused by Zn²⁺ binding to the indicator. To further verify this line of reasoning, it was tested whether Zn²⁺ chelation with TPEN affects the F356 signal in neurons exposed to Glu/Gly. As shown in Fig. 4b, TPEN failed to affect Fura-2FF F356 signal (green trace) but as expected, caused a prompt drop in the FluoZin-3 F488 signal (blue trace). The data confirmed that the drop in the F356 signal during Glu/Gly exposure could not be caused by Zn²⁺ binding to Fura-2FF.

Figure 4.

Effects of Na⁺ and TPEN on FluoZin-3 F488 (blue), Fura-2FF F340/F380 ratio (red), and the Fura-2FF Ca²⁺ isosbestic point, F356 (green). (a) Hippocampal neurons co-loaded with FluoZin-3 and Fura-2FF were exposed to Glu/Gly under Na⁺-free conditions (Na⁺ substituted with NMDG⁺) followed by an application of 158 mM Na⁺ and removal of NMDG⁺. The data are means ± standard error from 17 neurons monitored in a single experiment. The experiment was repeated 5 times with similar results. (b) The cells were treated as in (a) except that 10 μM TPEN was applied instead of Na⁺. The data are means ± standard error from 22 neurons monitored in a single experiment. The experiment was repeated 5 times with similar results.

It was then considered that the F356 decrease could be caused by the major pH_i drop taking place under such conditions (Kiedrowski 1999) and the impact of pH on the Fura-2FF F356 signal was examined. To this end, excitation spectra of Fura-2FF at pH 7.4 and 5.5 in vitro were obtained. The data show that the Fura-2FF fluorescence intensity excited at 356 nm was slightly higher at pH 5.5 than at pH 7.4 (Fig. 5), although the pH drop greatly compromised the ability of Fura-2FF to report [Ca²⁺] increase (Fig. 5, inset). Since the F356 signal was not decreased by the pH drop, the data argue against the idea that the decrease of the F356 signal during the first 4 minutes of Glu/Gly application (Fig. 4a,b) could be caused by a pH_i decrease.

Figure 5.

Excitation spectra of Fura-2FF solution under nominally Ca²⁺-free conditions (−Ca, blue), in the presence of 30 μM CaCl₂ (+Ca, red), at pH 7.4 (solid lines), and pH 5.5 (dashed lines). Note that acidification does not decrease the isosbestic point fluorescence (F356). Inset, acidification compromises the ability of Fura-2FF to report [Ca²⁺] increase. To calculate the F340/F380 ratio the data shown in the main figure were used. Note that the F340/F380 ratio is greatly increased by 30 μM Ca²⁺ (Ca+) at pH 7.4 but not at pH 5.5. The assay was repeated 3 times with similar results.

Most likely, the F356 signal decrease represents a Fura-2FF leakage. Similar leakage was observed by others (Vander Jagt et al. 2008; Medvedeva et al. 2009). It appears that this leakage occurs through pannexin-1 hemichannels which open following NMDA receptor activation and permeate a number of large molecules, including calcein (Thompson et al. 2008). One may envision that FluoZin-3 also leaks through these channels. Although such FluoZin-3 leakage is not apparent from the data (Fig. 4a,b), one has to consider that the basal FluoZin-3 F488 was so low in these experiments that any additional decrease of the signal could not be detected. Moreover, when [Zn²⁺]_i starts to elevate, the detection of any F488 signal drop due to FluoZin-3 leakage is complicated by the simultaneous F488 signal increase due to Zn²⁺ binding to the indicator.

Additional experiments were performed to better understand the relationships between [Ca²⁺]_i and [Zn²⁺]_i changes during Glu/Gly applications. In these experiments, the neurons were exposed to Glu/Gly for 20 minutes. During the first 8 minutes, the medium was Na⁺-free, during the next 4 minutes, Na⁺ (158 mM) was present in the medium, and during the last 8 minutes, the medium was Na⁺-free again. Fig. 6a shows data from a representative experiment. As in the experiment shown in Fig. 1, there was a great variability in Fura-2FF signal changes (Fig. 6a upper panel). The pattern of signal changes could be divided into four categories. Of 131 cells tested in 7 experiments, in 53 cells (40%) the Fura-2FF signal increased all the way to Rmax during the first Na⁺-free period and remained at this level for the rest of the experiment (Fig. 6a upper, red). In 20 cells (15%), the Fura-2FF signal increased (often to Rmax) during the first Na⁺-free period, but then decreased when Na⁺ was added and remained low during the second Na⁺-free period (Fig. 6a upper, green). In 25 cells (19%), the Fura-2FF signal increased to variable levels during the first Na⁺-free period, then decreased when Na⁺ was added and increased again when Na⁺ was removed (Fig. 6a upper, blue). In 33 cells (25%), the Fura-2FF signal showed a moderate (about 50% of Rmax) increase during the first Na⁺-free period, decreased when Na⁺ was added and remained low when Na⁺ was removed (Fig. 6a upper, orange). To simplify the picture, average Fura-2FF and FluoZin-3 data from the representative experiment are shown in the right panel of Fig. 6a. In over 53% of neurons, the Fura-2FF signal remained low (below 30% of Rmax) during the second Na⁺-free period (green and orange traces). The persistence of low [Ca²⁺]_i in these cells could be explained in terms of a decreased Ca²⁺ influx, increased Ca²⁺ efflux, or both. Since an increased Ca²⁺ efflux under Na⁺-free conditions (when plasmalemmal Na⁺/Ca²⁺exchangers do not remove Ca²⁺ from the cells) is unlikely, the persistence of low [Ca²⁺]_i can be best explained in terms of a decreased Ca²⁺ influx as the NMDA currents are known to run down (Rosenmund and Westbrook 1993; Li et al. 2002).

Figure 6.

Effects of Na⁺ and DTDP on FluoZin-3 and Fura-2FF fluorescence, and on pH_i. (a) The neurons were exposed to Glu/Gly and 1 mM CaEDTA under Na⁺-free conditions (Na⁺ substituted with NMDG⁺). Where indicated, 158 mM Na⁺ (Na) was transiently applied while NMDG⁺ was removed. The patterns of Fura-2FF fluorescence change could be divided into four color-coded categories that are discussed in the text. The left panel shows Fura-2FF and FluoZin-3 fluorescence changes in individual neurons and the right panel shows average fluorescence changes in neurons representing each category. The experiment was repeated 7 times with similar results. (b) Experiment analogous to the one shown in (a) except that 100 μM DTDP was added during the second Na⁺-free period, where indicated. The data are means ± standard error from 16 neurons. The experiment was repeated 4 times with similar results. (c) Experiment analogous to the one shown in (b) except that pH_i (BCECF F488/F440 ratio) was monitored. The data are means ± standard error from 23 neurons. The experiment was repeated 3 times with similar results.

The bottom panel of Fig. 6a shows that the FluoZin-3 signal increased during the first Na⁺-free period and decreased following Na⁺ application, as expected. Interestingly, during the second Na⁺-free period, the FluoZin-3 signal remained low in all neurons despite the above discussed variability in the Fura-2FF signal among the neurons. Since virtually all cells failed to increase [Zn²⁺]_i during the second Na⁺-free period (Fig. 6a bottom), it appears that either the intracellular store from which Zn²⁺ was released during the first Na⁺-free period had been depleted or that the mechanism involved in Zn²⁺ release is inactive during the second Na⁺-free period. To clarify this, it was tested whether an oxidizing agent, 2,2'dithiopyridine (DTDP) known to induce Zn²⁺ release from internal stores (Aizenman et al. 2000), is able to elevate [Zn²⁺]_i during the second Na⁺-free period. As shown in Fig. 6b, DTDP application promptly increased the FluoZin-3 signal, indicating that prior to DTDP addition the intracellular stores still contained Zn²⁺.

pH_i increase plays a role in Na⁺-dependent [Zn²⁺]_i clearance

A pH_i drop is expected to promote Zn²⁺ release from intracellular thiols (Li et al. 1954) and other ligand types as well as promote Zn²⁺ efflux from organelles, for example from the Golgi apparatus, by activating H⁺/Zn²⁺ exchange (Ohana et al. 2009). Therefore, it was considered that the difference in [Zn²⁺]_i elevation between the first and the second Na⁺-free period (Fig. 6b) could be explained in terms of a different pH_i. To this end, using BCECF fluorescence, pH_i was monitored under analogous experimental conditions. Glu/Gly application profoundly decreased BCECF F488/F440 ratio, indicating a pH_i decrease, and Na⁺ application caused the pH_i to increase. However, during the second Na⁺-free period, the Glu/Gly-induced pH_i drop was less profound than during the first Na⁺-free period (Fig. 6c). This result may be interpreted to indicate that the pH_i decrease during the second Na⁺-free period was insufficient to elevate [Zn²⁺]_i. No pH_i drop was associated with DTDP application (Fig. 6c), confirming that the mechanism of [Zn²⁺]_i elevation induced by this agent represents oxidation (Aizenman et al. 2000), not a pH_i decrease.

To clarify the role of pH_i in [Zn²⁺]_i changes, the relationships between pH_i and [Zn²⁺]_i were studied in a greater detail. To this end, the effects of Na⁺ versus Li⁺ and Cs⁺ on the rates of [Zn²⁺]_i drop and pH_i increase were compared. The efficacy of these cations in elevating pH_i (Fig. 7a) and clearing [Zn²⁺]_i (Fig. 7b) was similar, namely, Na⁺>Li⁺>Cs⁺. This result confirms that pH_i fluctuations may play a role in the mechanism of [Zn²⁺]_i changes. To further explore this possibility, gramicidin was used as an experimental tool. Gramicidin is an antibiotic that forms channels that permeate small monovalent but not divalent cations (Hladky and Haydon 1972) and therefore when pH_i is low, gramicidin is expected to increase it by promoting H⁺ efflux. Indeed, as shown in Fig. 7c, the rate of pH_i elevation increased when gramicidin was added. The rate of pH_i elevation increased even more when gramicidin was co-applied with Cs⁺ (NMDG⁺ was removed when Cs⁺ was added). Gramicidin also greatly increased the rate of pH_i elevation when extacellular pH was increased from 7.4 to 9.0 (Fig. 7c). As shown in Fig. 7d, the experimental maneuvers leading to the pH_i increase also resulted in [Zn²⁺]_i clearance and the efficacy of these maneuvers in provoking [Zn²⁺]_i decrease and pH_i increase was similar, namely, gramicidin + pH 9.0 > gramicidin + Cs⁺ > pH 9.0 alone > gramicidin alone.

Figure 7.

The relationships between pH_i and [Zn²⁺]_i in neurons exposed to Glu/Gly. (a) Effects of Na⁺, Li⁺, and Cs⁺ on pH_i in neurons treated with Glu/Gly. During the Na⁺-free period, Na⁺ was substituted with NMDG⁺. The left panel shows averages from 17 to 24 neurons monitored in a single experiment. The right panel shows average rates of pH_i increase calculated from three such experiments. **p < 0.01, *p < 0.05; one-way ANOVA followed by Student–Newman–Keuls test. (b) Analogous experiments as those shown in (a) except that [Zn²⁺]_i was monitored. **p < 0.01, one-way ANOVA followed by Student– Newman–Keuls test. (c) Analogous experiments as those shown in (a) except that the effects of cesium ± gramicidin (Gram; 5 µM) or pH 9.0 ± gramicidin on the rate of pH_i increase were studied. *p < 0.05; Kruskal–Wallis one-way ANOVA on Ranks followed by Dunn's test. (d) Analogous experiments as those shown in (c) except that [Zn²⁺]_i was monitored. **p < 0.01, one-way ANOVA followed by Student–Newman– Keuls test.

The faster rate of gramicidin-induced pH_i increase in the presence of Cs⁺ than in the presence of NMDG⁺ was expected. The gramicidin-formed channels do not permeate NMDG⁺. Therefore, K⁺ plus H⁺ efflux through the channels generates electrical potential that hampers exit of these ions. This does not occur in the presence of Cs⁺ (or another cation that permeates gramicidin channels) because the electrical potential generated by the K⁺ plus H⁺ efflux is dissipated by the Cs⁺ influx via the gramicidin channels.

The data shown in Fig. 7 indicate that pH_i fluctuations play a role in [Zn²⁺]_i changes. However, should the pH_i increase be the sole mechanism responsible for the clearance of cytosolic Zn²⁺, there should be a perfect correlation between the rate of pH_i increase and the rate of [Zn²⁺]_i drop. As shown in Fig. 8a, this was indeed the case for all experimental conditions except for the Na⁺ application. The rate of Na⁺-induced [Zn²⁺]_i decrease was about twice as fast as expected from the rate of the Na⁺-induced pH_i increase. This suggests that while the pH_i increase plays a role in Na⁺-dependent [Zn²⁺]_i decrease, Na⁺ also promotes additional pH_i-independent mechanisms of cytosolic Zn²⁺clearance.

Figure 8.

Na⁺-dependent [Zn²⁺]_i clearance involves pH_i-dependent and pH_i-independent mechanisms. (a) Correlation between the rate of pH_i increase and the rate of [Zn²⁺]_i decrease using the data from Fig. 7(a–d) (right panels). The linear regression with r² = 0.968 is calculated for all the data except Na⁺ (the red point). Note that the rate of Na⁺-induced [Zn²⁺]_i drop is about twice as fast as could be predicted based on the rate of Na⁺-induced pH_i increase. (b) The Na⁺-dependent [Zn²⁺]_i clearance (Na, blue trace) is not inhibited by 10 µM La³⁺ (Na + La³⁺, green trace) and is fully active under Ca-free conditions (Na Ca-free, red trace). The experimental details are the same as in Fig. 4(a). The data are averages from 18 to 23 neurons. The experiments were repeated four times with similar results.

In cortical neurons, Qin et al. (2008) recently described a Na⁺-dependent mechanism of Zn²⁺ efflux, which requires the presence of extracellular Ca²⁺ and is inhibited by 10 μM La³⁺. To test whether this mechanism plays a role in Na⁺-induced cytosolic Zn²⁺ clearance, the impact of 10 μM La³⁺ and the Ca²⁺-free medium on the rate of Na⁺-dependent [Zn²⁺]_i decrease was tested. However as shown in Fig. 8b, the rate of the Na⁺-dependent [Zn²⁺]_i decrease was not affected by either treatment.

Glu/Gly-induced [Zn²⁺]_i elevations do not exceed 20 nM

In an attempt to calibrate FluoZin-3 signals in terms of [Zn²⁺]_i, possible complications from overloading neurons with indicator (Dineley et al. 2002) were minimized by loading the neurons with very low concentrations of FluoZin-3 AM (0.1 μM) for only 5–6 minutes. Using the calibration approach described in Methods, [Zn²⁺]_i elevations appeared to reach the low nanomolar range (Fig. 9). However, as mentioned earlier, FluoZin-3 most likely leaks from the Glu/Gly-treated neurons. Therefore, Fmax values measured at the end of the experiments and used to calibrate the data are likely underestimated. Nevertheless, such calibration defines the upper limit of [Zn²⁺]_i increase in these experiments. While the exact value of [Zn²⁺]_i elevation is not known, one can be confident that it does not exceed 20 nM (Fig. 9).

Figure 9.

FluoZin-3 fluorescence data calibrated for [Zn²⁺]_i. Experimental conditions were analogous to those described in Fig. 1. The data are means ± standard error from 20 cells monitored in a single experiment. The experiment was repeated 5 times with similar results.

Discussion

Earlier work showed that reversal of the Na⁺ concentration gradient across the plasma membrane greatly enhances [Zn²⁺]_i elevations resulting from the applications of extracellular Zn²⁺ (Sensi et al. 1997; Cheng and Reynolds 1998), suggesting that Na⁺/Zn²⁺ exchanger may mediate Zn²⁺ influx under such conditions. More recently, Ohana et al. (2004) reported that under physiological conditions when extracellular Na⁺ is high and intracellular low, such Na⁺/Zn²⁺ exchanger extrudes Zn²⁺ from the cells. Additional mechanisms of Na⁺-dependent Zn²⁺ efflux that are inhibited by La³⁺ and require extracellular Ca²⁺ for operation were also characterized (Qin et al. 2008), but those do not seem to play a role in the mechanism of Na⁺-dependent Zn²⁺ clearance reported here (Fig. 8b).

The major finding of the present report is that Na⁺-dependent [Zn²⁺]_i clearance may result not only from activation of plasmalemmal Na⁺/Zn²⁺ exchangers but also from Na⁺-induced pH_i increase. The pathways of Na⁺-dependent pH_i regulation in neurons are well recognized and include plasmalemmal Na⁺/H⁺ exchangers (Tolkovsky and Richards 1987; Raley-Susman et al. 1991) and Na⁺-dependent Cl⁻-HCO₃⁻ exchangers (Schwiening and Boron 1994). Application of Na⁺, Li⁺ or Cs⁺ to neurons that had been exposed to Glu/Gly under Na⁺-free conditions (Na⁺ substituted with NMDG⁺) may also increase pH_i due to Na⁺, Li⁺ or Cs⁺ influx and K⁺ plus H⁺ efflux via NMDA and other channels that permeate these cations - a mechanism analogous to the above discussed acceleration of pH_i increase by gramicidin co-applied with Cs⁺.

The Glu/Gly-induced pH_i drop depends on Ca²⁺ influx because activation of NMDA receptors under Ca²⁺-free conditions fails to induce any major drop in pH_i (Hartley and Dubinsky 1993; Irwin et al. 1994; Kiedrowski 1999; Wu et al. 1999). Since the present data show the correlation between [Zn²⁺]_i and pH_i (Fig. 8a), one may conclude that the pH_i drop plays a critical role in the mechanism of Ca²⁺-dependent [Zn²⁺]_i elevation in this experimental model.

Several mechanisms may participate in the acidification-induced increase of [Zn²⁺]_i. For example, Zn²⁺ may be displaced from a number of intracellular binding sites as intracellular Zn²⁺ is buffered by interactions of the ion with a variety of ligand types, including sulfur from cysteine (glutathione, metallothionein), nitrogen from histidine, and oxygen from, for example, glutamate or aspartate (Berg and Shi 1996), and also phosphate (Binder et al. 2001). Moreover, cytosolic acidification favors Zn²⁺ release from organelles that use the H⁺/Zn²⁺ exchange to store Zn²⁺ (Ohana et al. 2009). These organelles are expected to release Zn²⁺ when pH_i drops sufficiently. The resulting [Zn²⁺]_i increase is expected to persist at low pH_i because low pH_i compromises ligand-dependent Zn²⁺ buffering (Li et al. 1954; Sensi et al. 2003). On the other hand, a pH_i increase favors Zn²⁺ re-loading to the organelles, which is supported by the data showing that re-loading of the organelles with Zn²⁺ takes place at pH 7.4 but not 6.0 (Colvin 2002).

Whereas the present author observed only minor [Zn²⁺]_i increases in neurons exposed to Glu/Gly and Na⁺ (Fig. 3d), Dineley et al. (2008) found major statistically significant elevations of FluoZin-3 signal in cortical neurons exposed to Glu/Gly despite the presence of Na⁺ in the medium. This discrepancy could be due in part to the fact that in the cortical neurons used by Dineley et al. (2008), [Ca²⁺]_i increased to Fura-2FF-saturating levels within 3 minutes of Glu/Gly application, but much smaller increases of the Fura-2FF signal were observed in the majority of hippocampal neurons used in the present study (Fig. 3d). As the release of Zn²⁺ from intracellular stores is triggered by Ca²⁺ influx (Dineley et al. 2008), one would expect that neurons that increase [Ca²⁺]_i sooner and faster also show a faster rate of [Zn²⁺]_i elevations. Another reason that the FluoZin-3 signal increase in Fig. 3d is small is that the data are expressed as the percentage of Fmax. The practical advantage of such data presentation is that it normalizes for differences in FluoZin-3 loading among the cells. These differences, if overlooked, may lead to data misinterpretation because as shown in Fig. 1c, [Zn²⁺]_i seemingly elevates at a faster rate in cells that loaded higher levels of FluoZin-3. Another reason for not using the common way of expressing FluoZin-3 data as the F488/F488basal ratio was that this author intended to measure the FluoZin-3 Fmax at the end of the experiments. Therefore, the cells needed to be loaded with very low concentrations of FluoZin-3 (see Methods) so that the F488 light intensity during the Fmax measurements was within the dynamic range of the camera. After such limited FluoZin-3 loading, F488basal was very low, close to or at background levels, and subtracting the background F488 from such low F488basal values often yielded zero. Loading more FluoZin-3 to the cells (to increase F488basal) resulted in Fmax values being higher than the maximal fluorescence intensity that could be measured using our Attofluor instrument. Presenting F488 data as a percentage of Fmax (to normalize for differences in FluoZin-3 loading) as we did is not an ideal solution either. As indicated in the Results, the Fmax measured at the end of the experiments is most likely decreased due to dye leakage similar to that observed for Fura-2FF (Fig. 4a,b, green traces). Moreover, phenomena such as bleaching and/or cell swelling also decrease Fmax. These problems could be avoided by using a ratiometric Zn²⁺ indicator with a high affinity for zinc. Unfortunately, the ratiometric Zn²⁺ indicator we had access to, zinbo-5 (Taki et al. 2004), could not be used to monitor [Zn²⁺]_i in this project. Zinbo-5 binds Zn²⁺ and permeates the plasma membrane. We found that it acts as an ionophore transporting Zn²⁺ across the plasma membrane according to the ion concentration gradient (data not shown). A similar problem was reported for another fluorescent Zn²⁺ indicator, zinquin (Snitsarev et al. 2001).

The current work is a follow up to our earlier report on the impact of Na⁺ on Glu/Gly-induced [Ca²⁺]_i elevations (Kiedrowski 1999). At the time of that publication, Ca²⁺-dependent and Glu/Gly-induced Zn²⁺ release from intracellular stores was not known. Therefore, the extent to which Fura-2 signals in that study were affected by [Zn²⁺]_i versus [Ca²⁺]_i elevations was not addressed. Since Glu/Gly-induced [Ca²⁺]_i elevations in neurons saturate Fura-2, the present study employed a low affinity Ca²⁺ indicator, Fura-2FF, which also binds Zn²⁺ with high affinity (Hyrc et al. 2000). The data indicate that Glu/Gly-induced [Zn²⁺]_i elevations did not exceed 20 nM (Fig. 9). However, even much higher Zn²⁺ concentrations elevated the F340/F380 ratio to no more than 30% of Rmax (Fig. 1d; Supplementary Fig. 1b). The data indicate that the increase of Fura-2FF F340/F380 ratio above 30% of Rmax represents [Ca²⁺]_i elevations.

In conclusion, the present study found that the Glu/Gly-induced pH_i decrease plays a critical role in Ca²⁺-dependent [Zn²⁺]_i elevations and suggests that the major extra- and intracellular acidification taking place in the ischemic brain (Smith et al. 1986) may destabilize Zn²⁺ homeostasis. The Na⁺-dependent Zn²⁺ clearance may be important for the restoration of Zn²⁺ homeostasis during brain recovery from ischemia.

Abbreviations used

BCECF

2',7'-bis-(2-carboxyethyl)-5(6)-carboxyfluorescein

[Ca²⁺]_i and [Zn²⁺]_i

intracellular Ca²⁺ and Zn²⁺ concentration

CaEDTA

calcium EDTA

DTDP

2,2'dithiopyridine

FCCP

carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone

F340, F356, F380, F440, F488

fluorescence emitted after 340, 356, 380, 440, and 488 nm excitation

Fmin and Fmax

minimal and maximal F488

FURD

5-fluoro-2'-deoxyuridine

Glu/Gly

cell exposure to 100 μM glutamate plus 10 μM glycine

MK-801

dizocilpine

NMDG

N-methyl-D-glucamine

pH_i

intracellular pH

Rmax

maximal F340/F380 ratio

TPEN

N,N,N′,N′-tetrakis(2-pyridylmethyl)ethylenediamine