Imaging free zinc in synaptic terminals in live hippocampal slices

Pergamon

PII:

Neuroscience Vol. 79, No. 2, pp. 347–358, 1997 Copyright ? 1997 IBRO. Published by Elsevier Science Ltd Printed in Great Britain. All rights reserved 0306–4522/97 $17.00+0.00 S0306-4522(96)00695-1

IMAGING FREE ZINC IN SYNAPTIC TERMINALS IN LIVE HIPPOCAMPAL SLICES T. BUDDE,*Q A. MINTA,† J. A. WHITE‡ and A. R. KAY*§ *Department of Biological Sciences, University of Iowa, Iowa City, IA 52242, U.S.A. †Texas Fluorescence Laboratories, Austin, TX 78747, U.S.A. ‡Department of Biomedical Engineering, Boston University, Boston, MA 02215, U.S.A. Abstract––Some glutamatergic synapses in the mammalian central nervous system exhibit high levels of free ionic zinc in their synaptic vesicles. The precise role of this vesicular zinc remains obscure, despite suggestive evidence for zinc as a neuromodulator. As a step towards elucidating the role of free zinc in the brain we have developed a method for imaging zinc release in live brain slices. A newly synthesized zinc-sensitive fluorescent probe, N-(6-methoxy-8-quinolyl)-p-carboxybenzoylsulphonamide (TFLZn), was used to monitor intracellular zinc in live rat hippocampal slices. The dye loaded into the zinc-rich synaptic vesicles of the mossy fibre terminals in the hippocampal formation. Direct electrical stimulation of the mossy fibre pathway diminished the fluorescence in the mossy fibre terminals, consistent with a stimulus-dependent zinc release. The synaptic release of zinc was followed by the rapid replenishment of the zinc levels in vesicles from an as yet unidentified intracellular zinc source. Furthermore, we present evidence that zinc may play a role in a form of long-term potentiation exhibited by the mossy fibre pathway. ? 1997 IBRO. Published by Elsevier Science Ltd. Key words: zinc, hippocampus, synapses, imaging, synaptic vesicles, mossy fibres.

A high concentration of free ionic zinc in synaptic vesicles is a prominent feature of some of the excitatory pathways in the mammalian forebrain.7,11 The anatomical evidence for this first arose from dithiozone staining, which delineates areas rich in free zinc.20 There are two pools of intracellular zinc. The first pool contains zinc that is tightly complexed to proteins and thus is invisible to most cytological stains. The second, a pool of free ionized zinc, is found within vesicles containing glutamate and/or aspartate and can be visualized, histochemically.8 Despite the wealth of information on zinc’s topography, and role in metalloenzymes, the role of free intravesicular zinc remains enigmatic. Exogenouslyapplied zinc has a multiplicity of actions on voltagegated and ligand-gated channels, and although this has been suggestive of a modulatory role for zinc in the CNS,28 there has been no clear demonstration that zinc indeed subserves such a function under normal physiological conditions. On the basis of the zinc blockade of the N-methyl--aspartate receptor, a QPresent address: Insitut fu¨r Physiol., Otto-von-Guericke Universita¨t, 39120 Magdeburg, Germany §To whom correspondence should be addressed. Abbreviations: DEDTC, diethyldithiocarbamate; EDTA, ethylenediaminetetra-acetate; HEPES, N-2-hydroxyethylpiperazine-N*-2-ethanesulphonic acid; SL, stratum lucidum; SP, stratum pyramidale; TFLZn, N-(6-methoxy8-quinoyly)-p-carboxybenzoylsulphonamide; TSQ, N(6-methoxy-8-quinolyl)-p-toluensulphonamide.

neuroprotective role for zinc has been suggested,22 however, there is also evidence that zinc may become neurotoxic when concentrations become excessive.5 The free zinc pool constitutes 5–15% of the total cellular zinc, the balance being tightly associated with macromolecules. The high concentration of the ion in synaptic vesicles (200–300 µM)7 and the energetic cost of maintaining it argues for a significant role in the neuronal economy. The role of zinc in catalysis or structuring protein architecture is not controversial. The same cannot be said for the pool of free zinc, as both its role in neurons and the rationale for the division between zinc-containing and zinc-free glutamatergic neurons has proved difficult to pinpoint. Experiments in hippocampal slices first demonstrated that zinc is released synaptically in a calciumdependent fashion.2,14 Both experiments relied on measuring the amount of zinc released into the extracellular solution during a bout of stimulation. Although clearly demonstrating the efflux of Zn2+, the methods have limited temporal resolution and give no information about the spatial aspects of release. Additional evidence for zinc release comes from experiments demonstrating a reduction of chelatable zinc, as judged by histochemical stains, after seizures provoked by intrahippocampal kainate.9,27 At the ultrastructural level the progress of zinc through the presynaptic terminal and into the synaptic cleft, has been determined by tracking the course of Zn2+ precipitated by sulphide.21

347

348

T. Budde et al.

Subsequently, experiments were extended to whole animals employing push-pull cannulae to measure zinc release brought about by paroxysmal stimulation, at defined loci in the hippocampus.1 The introduction of the fluorescent calcium probes13 greatly facilitated the estimation of the dynamics of intracellular calcium, and similar benefits could be expected to accrue from zinc-sensitive probes.31 Here we describe the use of a new zincsensitive probe TFLZn, which is based on N-(6methoxy-8-quinolyl)-p-toluensulphonamide10 (TSQ) to image intracellular levels of zinc in the mossy fibre boutons of live hippocampal slices. TFLZn is water soluble and exhibits very high selectivity for zinc, as compared to other cations. Its affinity for zinc (Kd220 µM) is appropriate for the expected free zinc concentration in the mossy fibre boutons (200– 300 µM).7 In this communication we have used the projection of the dentate granule cells to area CA3 in the hippocampus, the mossy fibre pathway, to visualize the dynamics of zinc release, as the terminals are large (23 µm) and rich in intravesicular zinc. Furthermore, zinc release has been demonstrated in hippocampal slices,3,14 and has been linked directly to the mossy fibre neuropil.1

EXPERIMENTAL PROCEDURES

All experiments were carried out in accordance with NIH and the University of Iowa animal care and use guidelines. Rats (Long–Evans, 150–250 g, Harlan) were decapitated after deep carbon dioxide anaesthesia. The brain was rapidly removed and slices were prepared by cutting the hippocampus on a McIlwain chopper (2350 mm). The slices were then placed in an interface chamber at room temperature and held for at least 1 h prior to the experiment. Slices were loaded for 1–1.5 h at room temp with 0.25 mM TFLZn (Texas Fluorescence Labs, Austin, TX). Experiments were carried out in a temperature controlled bath (32)C, Medical Systems, Greenvale, NY) on a Nikon Diaphot inverted microscope with a Nikon 10# Fluor objective. The slices were secured with a net and were stimulated (200–500 µA) with bipolar stainless steel electrodes (tip diameter 0.1 mm and 0.25 mm separation, Rhodes Medical Instruments, Woodland Hills, CA) placed over the band of fluorescence in the hilus (see Fig. 1a). Extracellular field potentials were monitored with glass electrodes coupled to an Axoclamp 2A and digitized by a Labmaster A/D using pClamp software (Axon Instruments, Foster City, CA). The recording electrode was placed on the border between the stratum pyramidale (SP) and the stratum lucidum (SL) in area CA3. To control for movement during an experiment bright-field images were taken at the beginning and end of the experiment. Excitation was provided by a pulsed UV laser (337 nm, 3 ns pulse duration, Laser Sciences, Cambridge, MA). The diffused and collimated light was reflected by a dichroic mirror (Omega, Brattleboro, VT 400 DCLP) and the emitted light was transmitted through a 510 nm edge-pass filter (Omega 510 WB 40) onto the faceplate of an intensified CCD camera (ICCD 2525FS, Videoscope, Herndon, VA). Images were acquired at a variable rate (30–0.3 frames/s) by a frame grabber (Recognition Technology Incorporated, Westborough, MA). To conserve dye, the extracellular solution (total volume 25 ml) was recirculated by a peristaltic pump, and continuously bubbled with 95%O2/5% CO2.

The recording solution contained (mM): NaCl 124, KCl 4, CaCl2 3, MgCl2 2, NaHCO3 23, EDTA 1, -glucose 10; pH 7.4. The spectral characteristics of TFLZn were measured in a solution with the following composition: 120 KCl, 20 NaCl, 10 HEPES (pH 7.25) on an SLM Aminco 4800C spectrofluorometer. Results are reported as mean&S.D. (n=number of observations) unless otherwise stated. RESULTS

Properties of TFLZn TFLZn (see Fig. 1a) was synthesized as a water soluble analogue of the Zn2+-specific probe TSQ10 In the absence of Zn2+, TFLZn exhibited little fluorescence, while in the presence of Zn2+ (100 µM) the fluorescence emission was increased some 100-fold. The excitation and emission spectra have peaks at 360 nm and 498 nm, respectively, and show little shifts with increasing Zn2+ concentrations. Calcium or magnesium (1 mM) gave no appreciable fluorescence above background levels, nor did the metals cadmium(II), cobalt(II), copper(II) or iron(III) at concentrations of 100 µM (Fig. 1b). Indeed copper quenched the fluorescence induced by zinc. The fluorescence emission as a function of Zn2+ concentration is illustrated in Fig. 1a. Half-maximal intensity was achieved at a concentration of 20 µM. TFLZn has a lower affinity than another recently synthesized Zn2+-probe, zinquin (Kd21 µM),31 making it less likely to chelate Zn2+ from proteins and metallothioneins. TFLZn labels free zinc-rich boutons Rat hippocampal slices incubated in saline containing TFLZn (0.25–0.5 mM) accumulate the dye in regions that are known to contain high concentrations of free zinc. In particular the mossy fibre neuropil of the hilus and CA3 are clearly evident in fluorescent images (Fig. 2a). Higher magnification fluorescence images of slices of the hippocampus confirm the localization of zinc staining to mossy fibre boutons (Fig. 2b). The small size (2–5 µm), punctate distribution and bead-like arrangement of the fluorescence are consistent with their identification as mossy fibre boutons. Acutely isolated CA3 neurons can be dissociated with mossy fibre terminals attached to the proximal dendrite.15 Neurons were acutely isolated from the CA3 sector of the hippocampus, and loaded with TFLZn. In a number of cases putative mossy fibre terminals had elevated fluorescence, and one such cell is depicted in Fig. 2c. Superfusion of hippocampal slices with a solution containing 0.25 mM TFLZn led to an increased fluorescence in the mossy fibres and hilus, accompanied by a smaller increase in fluorescence in other parts of the slice. The fluorescence increased in an exponential fashion with a time constant of 5.6&0.5 min (n=6), in both the region of mossy fibre terminals in SL and in the cell body layer in SP,

Imaging synaptic zinc release

349

Fig. 1. Properties of TFLZn. a) Normalized fluorescent intensity of TFLZn as a function of Zn2+ concentration (n=5), in a solution containing (mM) 120 KCl, 20 NaCl, 10 HEPES, 0.25 TFLZn (pH 7.25) Insets: chemical structure of TFLZn and its excitation-emission spectrum. b) Metal sensitivity of TFLZn fluorescence measured in the same solution as in A. (Mean&S.E.M., n=3). Data expressed as a fraction of the 100 µM Zn2+ solution. EDTA (2 mM). Excitation 360 nm, emission 510 nm.

however, the fluorescent intensity in the SL was 79&18% (n=30) greater than that in the SP. Once loaded with TFLZn the ratio of the fluorescence in SL to SP remained constant (Fig. 3a). Application of diethyldithiocarbamate (DEDTC; 1 mM), a membrane permeant zinc chelator,4 to slices that had been preloaded with TFLZn led to a reduction in the fluorescence of the mossy fibres, supporting the association between intracellular zinc and TFLZn as the source of the fluorescence

(Fig. 3b). Loading slices in 1 mM DEDTC for 1 h prior to TFLZn application reduced the ratio of SL/SP fluorescence to 1.19&0.04 (n=3) as opposed to 1.79&0.18 (n=30) in control slices. TFLZn appeared to be translocating into intracellular compartments, as was demonstrated by the inability of extracellularly applied Ca-EDTA (1 mM), which chelates zinc with higher affinity than calcium19 and is membrane impermeant, to diminish the fluorescence. TFLZn exists in solution as an

350

T. Budde et al.

Fig. 2. a) Labelling of the mossy fibre pathway with TFLZn. Left, CA3 region; Right, dentate gyrus. Bright-field images of the same field shown below. Stimulating electrodes visible in the left image. Slice loaded with 0.25 mM TFLZn. Scale bar=200 µm. b) Fluorescent structures within the stratum lucidum. Scale bar=10 µm. c) Acutely isolated CA3 neuron. Top, bright-field; Bottom, fluorescence of same field. Scale bar=10 µm.

Imaging synaptic zinc release

351

Fig. 3. a) Loading slices with TFLZn. Fluorescence intensity in SL (/) and SP (,), both normalized by the maximum fluorescence in the SL, and the ratio of the fluorescence intensity in SL to that in SP (4 SL/SP) as a function of time. The arrow indicates when TFLZn (0.25 mM) was introduced. b) DEDTC (1 mM) quenches fluorescence in the stratum lucidum (points labelled ‘‘wash in’’). ‘‘Preincubated’’ slice was exposed to 1 mM DEDTC for 1 h prior to loading with TFLZn.

equilibrium between a charged deprotonated form and an uncharged protonated form. It is the latter which probably moves across the membrane. Once inside zinc-containing vesicles the probe associates with zinc. On removing the probe from the extracellular solution the fluorescence declines slowly indicating that molecule exists in a tight, but slowly reversible association with zinc (Fig. 5a). In unstimulated preloaded slices no difference was observed in the rate of decline of fluorescence for 16.6 Hz and 0.03 Hz rates of laser illumination, suggesting that TFLZn is resistant to bleaching. The decline in fluorescence is probably not the result of spontaneous synaptic release, as holding the slices in low calcium did not decrease the rate of decline of the signal.

Two different dye-loading conditions were used in the experiments reported here. TFLZn at a concentration of 0.25 mM was maintained in the bath solution (‘‘bath-loaded’’) throughout the course of the experiment or the slices were preloaded with TFLZn at a concentration of 0.25 mM, then the slices were transferred to a solution lacking TFLZn (‘‘preloaded’’). The loading of slices was essentially complete after 1 h, whereafter the level of fluorescence in unstimulated slices remained constant. Control experiments Control experiments were performed on slices not exposed to TFLZn to test if changes in background fluorescence might interfere with the stimulation

352

T. Budde et al.

Fig. 4. Effect of TFLZn on mossy fibre field potential amplitude. The average (&S.E.M.) normalized field potential amplitude (top left inset, indicates the method of measurement) of bath-loaded (n=7) and preloaded slices (n=7, 10–20 min after a 1 h incubation in TFLZn). Insets: Field potentials under the two conditions. In the absence of TFLZn the field potential amplitudes exhibited a similar time-course to that of the preloaded slices.

experiments. Stimulation of unloaded slices at a rate of 10 Hz for 20 min led to an increase in the observed fluorescence (<10%), with the background signal being approximately 12% of the fluorescence in the presence of TFLZn. The increase probably results from a change in the NADH/NAD+ ratio.26 The amplitude of the field potential evoked by mossy fibre stimulation in CA3 was determined while recirculating 25 ml of control saline. The amplitude of the response in preloaded slices did not decline over a period of 1 h (Fig. 4). In similar experiments with 0.5 mM TFLZn the amplitude of the field potential tripled after 30 min and then declined in amplitude reaching the control

value some 60 min after commencing the application of TFLZn (Fig. 4). The increase in field potentials has been noted previously with the zinc chelator DEDTC and was attributed to an unidentified effect on the postsynaptic cell.4 Note that in the experiments that follow, optical recordings were commenced after 60 min of incubation in TFLZn. Electrical stimulation causes zinc release In experiments addressing zinc release, the fluorescent swath of mossy fibre terminals in the SL was imaged in a preloaded slice while simulating the

Fig. 5. a) Stimulus-dependent decline in TFLZn fluorescence (- normalized by initial value) in a preloaded slice. The intensity of the signal in the region-of-interest designated in the pseudocolour image (b) was measured by averaging 256 video frames accumulated in response to laser pulses delivered every 100 ms. The average intensity of the pixels in the designated region of interest is plotted as a function of time. A 10 Hz electrical stimulus (50 µs pulse duration) was applied during the period indicated by the shaded band. Application of tetrodotoxin (0.5 µM) prevented the stimulus-induced acceleration of the decline of fluorescence. The fluorescent intensity of unstimulated, preloaded control slices as function of time is designated by blue circles (n=6 slices, mean&S.E.M.). The lower pseudocolour map shows the fluorescent intensity measured along a line through the region-of-interest, as a function of time (0 to 35 min., top to bottom). The stimuli epochs are indicated by the grey bars alongside the pseudocolour map. Scale bar=100 µm. c) Stimulation of mossy fibre pathway releases zinc from mossy fibre terminals in a calcium-dependent fashion. Recording commenced in low calcium solution (0.25 mM CaCl2 and 3.75 mM MgCl2). Electrical stimulation delivered to mossy fibre pathway at the time designated by the grey boxes (50 µs pulse duration, 10 Hz). After the first bout of stimulation the external solution was replaced with normal saline. The box indicates the region-of-interest sampled in the graph. The red line is a linear regression of the 20 min control period. d) The pseudocoloured images were acquired at "2, 1, 4, 7 and 10 min (top to bottom) from the start of the second stimulus epoch. Scale bar=100 µm.

Imaging synaptic zinc release

mossy fibre pathway and monitoring the extracellular electrical response in the SP of area CA3. The fluorescence in the SL declined slowly during a control period when stimuli were delivered every 45 s. Stimulating the pathway at 10 Hz led to a stimulusdependent acceleration of the rate of decline of the fluorescence, that could be blocked by 500 nM tetro-

353

dotoxin (Fig. 5a, see below for control experiments). Reduction of the extracellular calcium concentration (0.25 mM) prevented the stimulus-induced decline in the fluorescence intensity (Fig. 5c). Restoration of the extracellular calcium to control levels reinstated the stimulus dependent decline in the fluorescence. Given the intracellular location of TFLZn and its

Fig. 5.

354

T. Budde et al.

Fig. 6. Variations in the change of TFLZn fluorescence in response to electrical stimulation under different TFLZn loading conditions. ÄS was calculated as the slope of the normalized fluorescence signal in the SL, during the stimulus epoch minus its slope during the preceding 5 min control period. The slopes were estimated by linear regression. The dashed lines indicates two S.Ds of the ÄS for control periods, estimated using two adjacent periods of 5 min duration. (n=32, preloaded; n=14, bath-loaded).

localization in hippocampal regions rich in vesicular zinc it seems likely that the stimulus triggered decline in fluorescence intensity results from the fusion of synaptic vesicles with the presynaptic plasma membrane and the discharge of the TFLZn-zinc complex into the extracellular space. Quantification of zinc release To quantify the stimulus-dependent change in mossy fibre fluorescence, a linear regression was performed on 5 min epochs of data. The change of slope (ÄS) was expressed as the difference between the slope of the stimulus period and the 5 min control period immediately prior to stimulation. A decrease in ÄS (i.e. a more negative value) indicates a faster rate of decline of fluorescence. The criterion levels (dotted vertical lines in Fig. 6) for determining if significant increases or decreases in ÄS had occurred, were estimated by measuring ÄS during control epochs, and setting the criterion levels to two S.Ds of the control observations. Histograms of ÄS show that there is considerable variation in the efficacy of stimulation from slice to slice. Under preloaded conditions, 46.9% of the slices show no detectable

decline in ÄS, while 53.1% exhibit a decrease and 0% an increase (n=32). The variation in the stimulusdependent change in fluorescence may result from variations in the proportion of mossy fibre axons that were transected during slice preparation. In thin slices it is difficult to ensure that the mossy fibres run intact within the confines of the slice, because the angle of the mossy fibre pathway with respect to the septo-temporal axis varies along this axis of the hippocampus.12 In contrast to slices that were preloaded with TFLZn, bath-loaded slices showed little stimulusdependent decline in the mossy fibre fluorescence (Fig. 6). 85.7% of the slices showed no detectable change in ÄS, while 0% exhibited a decrease and 14.3% an increase (n=14). TFLZn may depress synaptic activity pre- or postsynaptically, however, the lack of change in ÄS is not due to an effect on synaptic release, as both the presynaptic volley and postsynaptic response were not depressed by the application of TFLZn (data not shown). We hypothesize that the slower decline in bath-loaded slices resulted from an active replenishment of the intravesicular zinc. In the preloaded experiments once the vesicles discharge their contents they cannot be restocked with TFLZn, so the experiments in these cases measure the rate of release. In cases where the probe is present in the bath throughout the experiment, the change of fluorescence represents the sum of the release of the TFLZn-zinc complex to the extracellular space and possible restoration of zinc levels in the terminals. An alternate explanation for the stimulusdependent decline in fluorescence is that stimulation may change the chemical environment in the vesicles (e.g., pH), quenching TFLZn fluorescence. In this case the stimulus-dependent decline in fluorescence should occur in both bath- and pre-loaded slices. Since it does not, this explanation is unlikely. Zinc chelation and mossy fibre potentiation The extracellular field potential was monitored in the SP throughout the experiment. The 10 Hz stimulus lead to a decline in the field potential amplitude immediately after the tetanus (Fig. 7). Thereafter the amplitude rose to a level exceeding that of the prestimulus field potential, reaching a maximum some 8 min after the end of the stimulus. Potentiation was more common in preloaded (57%, n=30) and control slices (43%, n=7) than bath-loaded slices (20%, n=10). Potentiation was considered ‘‘significant’’ if the field potential amplitude exceeded the prestimulus amplitude plus two S.Ds. The continuous presence of TFLZn in the bath, also suppressed the post-tetanic depression. Of the pre-loaded slices, those that exhibited a significant decrease in ÄS, showed a greater tendency to potentiate (77%, n=13), than the slices that did not show a significant decrease in fluorescence (41%, n=17). We have not

Imaging synaptic zinc release

355

Fig. 7. Potentiation of mossy fibre field potentials by a 10 Hz stimulus of 8 min duration. The amplitude of the field potential was measured as indicated in the inset in Fig. 4. The average (&S.E.M.) normalized amplitude of the field potential is plotted as a function of time for three cases: Preloaded slices that exhibited a statistically significant decline in ÄS (-, n=13) and those that did not (,, n=17), and bath-loaded slices (4, n=10). Inset shows the field potentials of a preloaded slice that was potentiated by the tetanus.

followed the time-course of the potentiation in detail, however, in two control slices the potentiation was maintained for 1 h.

In the cases where the first stimulus did not lead to a significant decline in ÄS, the second stimulus was ineffective in all cases.

Stimulus dependence of zinc release

Release of zinc by neurotoxins

There are indications that zinc release is only recruited at elevated levels of stimulation.1,2 To investigate this, four different rates of stimulation were employed (1, 2, 5 and 10 Hz, n=8 in all cases) with a stimulus epoch of 5 min and a period of 10 min between stimuli. Only the 10 Hz stimulus gave rise to a ÄS significantly different from baseline values (three out of eight slices). The limited dynamic range and noise of the CCD camera, restricted the sensitivity of our measurements. It is thus not possible from our experiments to determine if zinc release is invoked only above a certain frequency of stimulation or if zinc release occurs whenever glutamate is released. Does the 8 min 10 Hz stimulus deplete zinc stocks? To explore this issue experiments were performed where slices which were preloaded with TFLZn, and were then subjected to two bouts of 10 Hz (8 min duration) stimulation, separated by a rest period of 10 min. In the cases (four out of nine) where the first stimulus led to a significant decline in ÄS the second stimulus also evoked a significant decline in ÄS. The average ÄS for the first and second tetani were; "0.0156&0.0063 and "0.0163&0.0089/min (n=4), respectively. This result demonstrates that the stimulus was not sufficient to deplete zinc stores and that an initial stimulus did not depress subsequent release.

If intravesicular zinc is replenished from intracellular stores, it might be expected that bath-loaded slices would show a stimulus-dependent decline in TFLZn fluorescence at elevated levels of stimulation and after a greater latency than preloaded slices as the intracellular zinc supplies become depleted. To test this prediction slices were chemically stimulated with kainate, which has been shown to bleach slices of histochemically reactive zinc.9,27 Application of kainate plus 4-aminopyridine, to increase the excitability, to slices preloaded with TFLZn led to a decline in fluorescence in the mossy fibre fluorescence (Fig. 8a). In bath-loaded slices the excitotoxin led to a slower decline of fluorescence (Fig. 8b, timeconstants: preload,: 10&4 min, n=3; bath-load, 58&17 min, n=4) and with a longer latency to the onset of release (preloaded 7&2 min, bath-loaded 23&5 min). This result is consistent with the stimulus-dependent reloading of synaptic vesicles with zinc. Nanomolar á-latrotoxin brings about the rapid exocytosis of neurotransmitter in a calciumindependent fashion, through the binding of á-latrotoxin to synaptotagmin on the presynaptic terminal.23 Application of 12 nM á-latrotoxin to preloaded hippocampal slices lead to the decline of

356

T. Budde et al.

Fig. 8. Changes in TFLZn fluorescence evoked by different excitotoxins. Decline in mossy fibre fluorescence on the application of kainate (10 µM) and 4-aminopyridine (1 mM) in pre- (a) and bath- (b) loaded slices, (mean&S.E.M.) of three and four experiments, respectively. Application of toxin indicated by shading. Solid lines represent exponential functions (a=1, b=2 exponentials) fitted to normalized fluorescent intensity in the stratum lucidum in control slices (n=3 in both cases). Normalized fluorescent intensity is plotted. c) Decline of mossy fibre fluorescence induced by the application of á-latrotoxin (12 nM). Application of toxin indicated by shading. Absolute fluorescent intensity is plotted. Insets: field potentials measured during the course of á-latrotoxin application.

fluorescence in the mossy fibres with a roughly exponential time-course to 17% of the initial value (Fig. 8c). The accelerated loss of fluorescence is consistent with the á-latrotoxin-induced exocytosis. The evoked field potentials declined in parallel with the fluorescence and also reached a non-zero steady state after 1 h (Fig. 8c). The inability of á-latrotoxin to quench the fluorescence and abolish the field potential might result from the limited access of the high molecular weight toxin to the centre of the slice or from the proteolysis of the toxin by extracellular proteases.

DISCUSSION

On the basis of the topography of areas highlighted by TFLZn we conclude that it stains synaptic vesicles with a high free zinc concentration. The intracellular location of the dye is underscored by the inability of extracellular EDTA to quench the fluorescence. The fluorescence in the mossy fibre band declines in a stimulus-dependent fashion, consistent with the release of the dye and zinc from synaptic vesicles. The absence of a stimulus-dependent decline in the mossy fibre terminal fluorescence in bath-loaded

Imaging synaptic zinc release

slices suggests that zinc levels are rapidly replenished in synaptic vesicles. The continuous presence of EDTA in the saline precludes the extracellular space as a source for the zinc. The time-course of zinc replenishment is consistent with experiments on the recycling of synaptic vesicles25 and the acceleration of zinc uptake by hippocampal slices after electrical stimulation.14 Moreover, TSQ and nuclear magnetic resonance experiments have indicated the presence of an intracellular free zinc pool that is increased after glutamate stimulation.29 We suggest that pools of tightly bound zinc, perhaps chelated by metallotheins,6 could serve as sources for recharging the vesicles. It should be noted that this pool of zinc, because of the high affinity of metallothioneins for zinc, would be ‘‘invisible’’ to TFLZn. We propose that stimulation might liberate zinc from metallothioneins by activating high affinity zinc transporters on synaptic vesicles, the modification metallothionein or by pumping the metallothionein-zinc complex into the synaptic vesicles, where the low pH would release zinc.18 Potentiation of the mossy fibre pathway by a 10 Hz stimulus, has been recently observed by others with a 1.5 min 10 Hz stimulus.16 Our observation that the continuous presence of TFLZn in the bath blocks potentiation, suggests that zinc release may be essential for this form of plasticity. Notice that EDTA does not block the potentiation. The failure of this zinc chelator to prevent potentiation may be accounted for by the slowness of the off-rate for calcium, retarding the interception of zinc in the synaptic cleft. Preloading does not prevent potentiation and this is consistent with the low affinity of TFLZn which would favour the dissociation of the zinc-TFLZn complex once it is released. In contrast to these results, in the conventional paradigm of mossy fibre long-term potentiation, the membranepermeant zinc chelator, 1,2-diethyl-3-hydroxypyridin-4-one, did not disrupt mossy fibre long-term potentiation.30 Most of the current generation of fluorescent ionsensitive probes do not penetrate well into the depths of the adult mammalian brain. For example, isolated cells accumulate Fura-2 AM well, but it is not very effective in loading mature mammalian brain slices.24 In contrast TFLZn penetrates thoughout the depth of mature brain slices and loads zinc-rich neuronal processes. The low molecular weight of TFLZn (418 g/mole) and its water solubility probably accounts for its effectiveness. A further problem with AM-loading is that it is relatively inefficient in loading intracellular compartments such as synaptic vesicles, as the probe is typically entrapped in the cytoplasm. In contrast TFLZn readily enters synaptic

357

vesicles, because of the membrane permeability of the protonated form and the specific affinity of the probe for zinc. The development of fluorimetric zinc probes opens the way to the investigation of zinc physiology in live preparations including intact animals. Moreover, it allows the estimation of approximate zinc concentrations at a subcellular level and the investigation of the mode and tempo of zinc release, at a faster rate than dialysis experiments or superfusate analysis. The low affinity of TFLZn for zinc is something of an advantage as it provides little competition for the native high affinity zinc-binding proteins in the cell and would not be expected to leach zinc from these proteins. Zinc affects a plethora of channels, indeed so many that identifying its precise role has been particularly problematic. An added impetus to understanding zinc’s physiology was provided by the recent finding that free zinc at physiologically reasonable concentrations can precipitate human Aâ amyloid protein, the major constituent of the plaques in Alzheimer’s disease and thus may play a role in its pathophysiology.3 Zinc has also been implicated in the aetiology of excitotoxic cell death in seizures,22,29 a condition in which vesicular zinc is vigorously released27 and stroke.17 The methodology used in this report opens the way to dynamically modifying and monitoring zinc release and may be useful in making sense of the role of this enigmatic ion in cortical physiology. CONCLUSIONS

We have demonstrated that the water-soluble zincsensitive probe TFLZn, highlights intravesicular freezinc in mossy fibre terminals and can be employed to visualize the mobilization of synaptic zinc. Stimulation of the mossy fibre pathway reduces the intensity of the fluorescence associated with the terminals in a calcium- and stimulus-dependent fashion, consistent with the synaptic release of the TFLZn-zinc complex to the extracellular space. Evidence is provided that synaptic vesicles reload with zinc from histochemically cryptic intracellular stores and that synapticallyreleased zinc participates in a form of synaptic potentiation provoked by a sustained 10 Hz stimulus to the mossy fibre pathway. Acknowledgements—We are indebted Dr C. J. Frederickson for his participation in some preliminary experiments and advice during the course of this study. We thank Dr D. J. Krupa for helpful comments on an earlier versions of this manuscripts, J. Richardson and T. Stricker for technical assistance, Dr M. J. Welsh for the use of his spectrofluorometer, and Dr W. G. Regehr for suggesting the experiment in Fig. 2c. Supported by grants from NINDS and ONR (to ARK).

REFERENCES

1.

Anikstejn L., Charton G. and Ben-Ari Y. (1987) Selective release of endogenous zinc from the hippocampal mossy fibers in situ. Brain Res. 404, 58–64.

358 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31.

T. Budde et al. Assaf S. Y. and Chung S.-H. (1984) Release of endogenous Zn2+ from brain tissue during activity. Nature 308, 734–735. Bush A. I., Pettingell W. H., Multhaup G., Paradis M. D., Vonsattel J.-P., Gusella J. F., Beyreuther K., Masters C. L. and Tanzi R. E. (1994) Rapid induction of Alzheimer Aâ amyloid formation by zinc. Science 265, 1464–1467. Danscher G., Shipley M. T. and Andersen P. (1975) Persistent function of mossy fiber synapses after metal chelation with DEDTC (Antabuse). Brain Res. 85, 522–526. Duncan M. W., Marini A. M., Watters R., Kopin I. J. and Markey S. P. (1992) Zinc, a neurotoxin to cultured neurons, contaminates cycad flour prepared by traditional Guamanian methods. J. Neurosci. 12, 1523–1537. Ebadi M. (1991) Metallothioneins and other zinc-binding proteins in brain. Meth. Enzymol. 205, 363–387. Frederickson C. J. (1989) Neurobiology of zinc and zinc-containing neurons. Int. Rev. Neurobiol. 31, 145–238. Frederickson C. J. and Danscher G. (1990) Zinc containing neurons in hippocampus and related CNS structures. In Progress in Brain Research (eds Storm-Mathisen J., Zimmer J. and Ottersen O. P.), Vol. 83, pp. 71-84. Elsevier Science Publishers, Amsterdam. Frederickson C. J., Hernandez M. D., Goik S. A., Morton J. D. and McGinty J. F. (1988) Loss of zinc staining from mossy fibers during kainic acid induced seizures: a histofluoresence study. Brain Res. 446, 383–386. Frederickson C. J., Kasarkis E. J., Ringo D. and Fredrickson R. E. (1987) A quinoline fluorescence method for visualizing and assaying the histochemically reactive zinc (bouton zinc) in the brain. J. Neurosci. Meth. 20, 91–103. Frederickson C. J. and Moncrieff D. W. (1994) Zinc-containing neurons. Biol. Signals 3, 127–139. Gaarskjaer F. B. (1978) Organization of the mossy fiber system of the rat studied in extended hippocampi. J. comp. Neurol. 178, 73–88. Grynkiewicz G., Poenie M. and Tsien R. Y. (1985) A new generation of Ca2+ indicators with greatly improved fluorescence properties. J. biol. Chem. 260, 3440–3450. Howell G. A., Welch M. G. and Frederickson C. J. (1984) Stimulation-induced uptake and release of zinc in hippocampal slices. Nature 308, 736–738. Kay A. R. and Wong R. K. S. (1986) Isolation of neurons suitable for patch-clamping from the adult mammalian central nervous systems. J. Neurosci. Meth. 16, 227–238. Kobayashi K., Manabe T. and Takahashi T. (1996) Presynaptic long-term depression at the hippocampal mossy fiber-CA3 synapse. Science 273, 648–650. Koh J. Y., Suh S. W., Gwag B. J., He Y. Y., Hsu C. Y. and Choi D. W. (1996) The role of zinc in selective neuronal death after transient global cerebral ischemia. Science 272, 1013–1016. Maret W. (1994) Oxidative metal release from metallothionein via zinc-thiol/disulphide interchange. Proc. natn. Acad. Sci. U.S.A. 91, 237–241. Martell A. M. and Smith R. M. (1995) Critically Selected Stability Constants of Metal Complexes. 2nd edn, N.I.S.T., Gaithersburg, MD. Maske H. (1955) U } ber den topochemischen nachweis von zink im Ammonshorn verschiedener sa¨ugestiere. Naturwissenschaften 42, 424. Perez-Clausell J. and Danscher G. (1986) Release of zinc sulphide accumulations into synaptic clefts after in vivo injection of sodium sulphide. Brain Res. 362, 358–361. Peters S., Koh J. and Choi D. W. (1987) Zinc selectively blocks the action of N-methyl--aspartate on cortical neurons. Science 236, 589–592. Petrenko A. G., Perin M. S., Davletov B. A., Ushkaryov Y. A., Geppert M. and Su¨dhof T. C. (1991) Binding of synaptotagmin to the alpha-latrotoxin receptor implicates both in synaptic vesicle exocytosis. Nature 353, 65–68. Regehr W. G. and Tank D. W. (1991) Selective Fura-2 loading of presynaptic terminals and nerve cell processes by local perfusion in mammalian brain slice. J. Neurosci. Meth. 37, 111–119. Ryan T. A., Reuter H., Wendland B., Schweizer F. E., Tsien R. W. and Smith S. J. (1993) The kinetics of synaptic vesicle recycling measured at single presynaptic boutons. Neuron 11, 713–724. Sick T. J. and Rosenthal M. (1989) Indo-1 measurements of intracellular free calcium in the hippocampal slice: Complications of labile NADH fluorescence. J. Neurosci. Meth. 28, 125–132. Sloviter R. S. (1985) A selective loss of hippocampal mossy fiber Timm stain accompanies granule cell seizure activity induced by perforant path stimulation. Brain Res. 330, 150–153. Smart T. G., Xie X. and Krishek B. J. (1994) Modulation of inhibitory and excitatory amino acid receptor ion channels by zinc. Prog. Neurobiol. 42, 393–441. Weiss J. H., Hartley D. M., Koh J. and Choi D. W. (1993) AMPA receptor activation potentiates zinc neurotoxicity. Neuron 10, 43–49. Xie X. and Smart T. G. (1994) Modulation of long-term potentiation in rat hippocampal pyramidal neurons by zinc. Pflu¨gers. Arch. 427, 481–486. Zalewski P. D., Forbes I. J. and Betts W. H. (1993) Correlation of apoptosis with change in intracellular labile Zn(II) using Zinquin [(2-methyl-8-p-toluenesulphonamido-6-quinolyloxy)acetic acid] a new specific fluorescent probe for Zn(II). Biochem. J. 296, 403–408. (Accepted 11 December 1996)