1The properties of ryanodine-sensitive Ca2+ stores in CA1 pyramidal cells were investigated in rat hippocampal slices by using whole-cell patch-clamp recordings combined with fura-2-based fluorometric digital imaging of cytoplasmic Ca2+ concentration ([Ca2+]i).
2Brief pressure applications of caffeine onto the somata of pyramidal cells caused large transient increases in [Ca2+]i (Ca2+ transients) of 50–600 nm above baseline.
3The Ca2+ transients evoked by caffeine at −60 mV were not associated with an inward current, persisted after blocking voltage-activated Ca2+ currents and were completely blocked by bath-applied ryanodine. Similar transients were also evoked at +60 mV. Thus, these transients reflect Ca2+ release from intracellular ryanodine-sensitive Ca2+ stores.
4The Ca2+ transients evoked by closely spaced caffeine pulses rapidly decreased in amplitude, indicating progressive depletion of the Ca2+ stores. The amplitude of the Ca2+ transients recovered spontaneously with an exponential time constant of 59 s. Recovery was accelerated by depolarization-induced elevations in [Ca2+]i and blocked by cyclopiazonic acid (CPA) and thapsigargin, indicating that store refilling is mediated by endoplasmic reticulum Ca2+-ATPases.
5Even without prior store depletion the caffeine-induced Ca2+ transients disappeared after 6 min exposure to CPA, suggesting that ryanodine-sensitive Ca2+ stores are maintained at rest by continuous Ca2+ sequestration.
6Caffeine-depleted Ca2+ stores did not refill in Ca2+-free saline, suggesting that the refilling of the stores depends upon Ca2+ influx through a ‘capacitative-like’ transmembrane influx pathway operating at resting membrane potential. The refilling of the stores was also blocked by Ni2+ and gallopamil (D600).
7Elevations of basal [Ca2+]i produced by bath-applied KCl markedly potentiated (up to 6-fold) the caffeine-induced Ca2+ transients. The degree of potentiation was positively related to the increase in basal [Ca2+]i. The Ca2+ transients remained potentiated up to 9 min after reversing the KCl-induced [Ca2+]i increase. Thus, the ryanodine-sensitive Ca2+ stores can ‘overcharge’ when challenged with an increase in [Ca2+]i and slowly discharge excess Ca2+ after basal [Ca2+]i returns to its resting level.
8Pressure applications of caffeine onto pyramidal cell dendrites evoked local Ca2+ transients similar to those separately evoked in the respective somata. Thus, dendritic ryanodine-sensitive Ca2+ stores are also loaded at rest and can function as independent compartments.
9In conclusion, the ryanodine-sensitive Ca2+ stores in hippocampal pyramidal neurones contain a releasable pool of Ca2+ that is maintained by a Ca2+ entry pathway active at subthreshold membrane potentials. Ca2+ entry through voltage-gated Ca2+ channels transiently overcharges the stores. Thus, by acting as powerful buffers at rest and as regulated sources during activity, Ca2+ stores may control the waveform of physiological Ca2+ signals in CA1 hippocampal pyramidal neurones.
The ryanodine-sensitive intracellular Ca2+ stores are known to play a major role in excitation–contraction coupling in skeletal and cardiac muscle (see McPherson & Campbell, 1993; Sutko & Airey, 1996). Much less is known about the function of these stores in central neurones, despite the fact that all known members of the ryanodine receptor (RyR) family, namely, skeletal muscle type (type I), cardiac muscle type (type II) and brain type (type III), are abundantly expressed in the central nervous system (Furuichi, Furutama, Hakamata, Nakai, Takeshima & Mikoshiba, 1994).
According to another hypothesis (see Simpson, Challiss & Nahorski, 1995), neuronal ryanodine-sensitive Ca2+ stores may act as a buffering system for intracellular Ca2+ ions (Ca2+i). Firstly, KCl-induced increases in [Ca2+]i in bullfrog sympathetic neurones, were substantially attenuated after depletion of ryanodine-sensitive Ca2+ stores by prolonged caffeine application (Friel & Tsien, 1992). Secondly, blockers of sarcoplasmic–endoplasmic reticulum Ca2+-ATPases (SERCAs), prolonged the depolarization-induced increases in dendritic [Ca2+]i in rat neocortical layer V pyramidal neurones in slices (Markram, Helm & Sakmann, 1995). Thus, sequestration of cytosolic Ca2+ by intracellular Ca2+ stores may contribute substantially to Ca2+ clearance in peripheral and central neurones.
Using fluorometric digital imaging of [Ca2+]i combined with whole-cell patch-clamp recordings in rat hippocampal slices, we have investigated the dynamic properties of the ryanodine-sensitive Ca2+ stores in CA1 pyramidal cells. These neurones express the three types of RyRs and, compared with other central neurones, have the highest level of the brain-type RyRs (Furuichi et al. 1994). Moreover, in these neurones the RyRs are expressed in the axon, soma and dendrites, including the dendritic spines (Sharp, McPherson, Dawson, Aoki, Campbell & Snyder, 1993) and thus occupy strategically important positions for synaptic signalling and integration.
The composition of the standard saline was (mm): 125 NaCl, 2.5 KCl, 2 CaCl2, 1 MgCl2, 1.25 NaH2PO4, 26 NaHCO3 and 20 glucose (pH7.4 when bubbled continuously with 95% O2–5% CO2). To block voltage-gated sodium channels 0.5 μm tetrodotoxin (TTX) was added to the standard saline in all experiments. In Ca2+-free saline the CaCl2 was replaced with 1 mm MgCl2 and 0.1 mm EGTA was added.
The standard pipette (intracellular) solution contained (mm): 130 CsCl, 20 tetraethylammonium chloride (TEACl), 4 Mg-ATP, 0.2 Na-GTP, 100–200 μm fura-2 pentapotassium salt and 10 Hepes (pH 7.3, adjusted with CsOH). In several experiments 140 mm KCl and 10 mm NaCl were used instead of CsCl and TEACl, as indicated.
All chemicals were purchased from Sigma. D600 (gallopamil) was a gift from Knoll (Ludwigshafen, Germany).
Electrophysiological recordings and drug application
Experiments were performed at room temperature (22–24°C) in an experimental set-up equipped with an upright microscope (Axioscope FS; for details see also Eilers, Schneggenburger & Konnerth, 1995). CA1 pyramidal neurones in hippocampal slices were visually identified according to their characteristic shape and their location within the cell body layer (Konnerth, 1990; Garaschuk et al. 1996). Before establishing the whole-cell recording configuration the membrane surface was cleaned, if necessary, using procedures described previously (Edwards et al. 1989). Patch pipettes were made of borosihcate glass (Hilgenberg GmbH, Malsfeld, Germany) and coated with a silicon-based resin (General Electrics, Ruesselsheim, Germany). They had a resistance of 2–3 MΩ when filled with the intracellular solution. Membrane currents were recorded with the EPC-9 patch-clamp amplifier (HEKA, Lambrecht, Germany). The holding potential was set, if not otherwise indicated, at −60 mV. Data acquisition and pulse protocols for controlling drug application (see below) were performed using ‘Pulse’ software (HEKA). The pipette series resistance was less than 10–12 MΩ. In some experiments it was compensated up to 60% following standard procedures.
Caffeine was pressure applied using a Picospritzer II (General Valve, Fairfield, NJ, USA) from a fine pipette (6–12 MΩ) filled with standard saline in which 10 or 20 mm NaCl was substituted with 20 or 40 mm caffeine, respectively. The tip of the caffeine application pipette was placed near (∼5 μm) a deliberately chosen primary target neurone (cell 1 in Fig. 1B) and, for a given caffeine concentration in the pipette, the shortest pulse duration (2–8s) evoking a reproducible response was selected. We estimate that the true concentration of caffeine reaching this cell was less than 10 mm. To obtain 80 mm KCl-containing saline, 80 mm NaCl was substituted for KCl in the standard perfusion saline.
Fluorometric Ca2+ measurements
Fluorometric Ca2+ measurements were performed with fura-2 (pentapotassium salt; Molecular Probes) added to the pipette solution at a concentration of 100–200 μm and loaded into the cell via the patch pipette. Alternatively, cells were loaded with the acetoxymethyl ester of fura-2 (fura-2 AM; Molecular Probes). For this purpose slices were incubated for 10 min in the oxygenated saline containing fura-2 AM (10–15 μm) at 36°C. A reasonably good loading of neuronal somata located in the top layers of the slice was routinely achieved.
Fluorometric measurements were done using a variable-scan digital imaging system (T.I.L.L. Photonics GmbH, München, Germany) with ‘Image-8’ software (University of Saarland, Homburg, Germany). Digital fluorescence images of the slice were obtained using paired exposures to 360 and 380 nm excitation wavelength. Following the establishment of the whole-cell configuration, the loading of the cell with fura-2 was monitored as a gradual increase in the Ca2+-independent fluorescence signal. Measurements were started only after equilibration of the fura-2 concentration between the pipette and all cellular compartments of interest (e.g. soma and dendrites), as indicated by stable maximal intensity signals at 360 nm excitation wavelength. In general, the time taken to reach this equilibrium was about 10 min.
Routinely, fluorescence values representing spatial averages from up to three defined pixel regions were transferred on-line through a serial port to another personal computer and displayed there as a function of time together with the whole-cell current recording (Eilers et al. 1995). In most cases the regions of interest were placed on somata of different pyramidal cells. Thus, each experiment provided simultaneous measurements from up to three different neurones. The background fluorescence was measured in a separate pixel region near the cell and used for an on-line correction of the cellular fluorescence signals. The calibration constants Keff (effective binding constant), Rmin (fluorescence ratio at zero Ca2+) and Rmax (limiting fluorescence ratio at high Ca2+) were obtained from in vivo calibration experiments (Neher, 1989; Filers et al. 1995). Typical calibration parameters found for Keff, Rmin and Rmax were 1048, 0.67 and 4.2 nm, respectively. In some instances (see Fig. 1), consecutive paired exposures to 360 and 380 nm were used to construct background-corrected digital fluorescence images. They were displayed on-line in a pseudocolour mode on a monitor and stored on hard disk for later analysis.
Data are given as means ± standard deviation (s.d.).
Resting basal [Ca2+]i was 49.5 ± 14.4 nm (n= 76 cells from 33 slices) in pyramidal cells loaded with fura-2 AM and 42.5 ± 19.7 nm (n= 52 cells from 42 slices) in whole-cell clamped pyramidal cells loaded with fura-2 via the recording patch pipette. The difference between the mean values was not significant (P > 0.25, Student's two-tailed t test).
Caffeine-induced increases in [Ca2+]i in somata of pyramidal cells
Figure 1 illustrates fluorometric measurements obtained from neurones in the CA1 pyramidal cell layer (Fig. 1A), which were loaded with fura-2 AM (Fig. 1B) Brief (3s) pressure application of caffeine rapidly evoked a marked increase in [Ca2+]i (Ca2+ transient) in the neurones immediately adjacent to the application pipette (Fig. 1D). The spatially averaged somatic [Ca2+]i values obtained from four neurones in the field of view (delineated in Pig. 1B) were plotted as a function of time in Fig. 1C. In these cells, numbered 1-4, caffeine increased [Ca2+]i by 156, 136, 127 and 80 nm, respectively. These peak values were attained during caffeine application and prolongation of the caffeine pulse did not cause a further rise in [Ca2+]i (not shown). This suggests that the Ca2+ transients were the maximal responses to the caffeine concentrations that stimulated the cells. Expectedly, neurones that were closer to the application pipette (e.g. cell 1 in Fig. 1) displayed larger and faster caffeine-evoked Ca2+ transients than those located downstream (e.g. cell 4 in Fig. 1).
The amplitudes of caffeine-induced Ca2+ transients in 105 pyramidal neurones located in the immediate vicinity of the caffeine application pipette ranged from 50 to 600 nit above basal [Ca2+]i with few exceptions exceeding this range (mean 253 ± 269, n= 105 cells from 85 slices). The large scatter in amplitudes of these Ca2+ transients probably reflects the variation in the concentrations of applied caffeine that actually stimulated the cells (compare cell 1 and cell 4 in Fig. 1).
Despite these differences in amplitude, the caffeine-evoked Ca2+ transients were highly robust and reproducible in most cells tested. In control experiments we found that only 11 out of 133 cells (n= 9 slices), which experienced caffeine concentration sufficient to evoke Ca2+ transients in surrounding cells, failed to respond to caffeine applications (not shown). One of these eleven cells became responsive after increasing the duration of application from 2-5 to 3s suggesting that slightly higher concentrations of caffeine were necessary to stimulate these cells.
Caffeine induces Ca2+ release from internal stores
To identify the origin of the caffeine-induced Ca2+ transients we first examined whether these signals are associated with influx of extracellular Ca2+ ions. We tested this question in whole-cell clamped pyramidal cells in conditions which isolated Ca2+ currents (see Methods). As shown in Fig. 2A (left panel), a 200 ms depolarizing pulse from −60 to 0 mV evoked an inward current and a large increase in [Ca2+]i By contrast, the caffeine-induced Ca2+ transient was not associated with any inward current. Perfusing the slice with nominally Ca2+-free saline for about 4 min completely blocked the depolarization-induced inward current and the associated Ca2+ transient, while the caffeine-induced Ca2+ transient was preserved. Similar observations were made in seven experiments, supporting the notion that the caffeine-induced Ca2+ transients reflect the release of Ca2+ from internal stores.
Another test for the internal origin of these Ca2+ transients is illustrated in Fig. 2B. The pyramidal cells were held at the approximate Ca2+ equilibrium potential (+60 mV) to block transmembrane Ca2+ influx. Application of caffeine induced Ca2+ transients in all cases (n= 6 neurones) at this potential. Their amplitudes exceeded those of the transients recorded at resting membrane potential (Fig. 2B). This increase in amplitude is related to the increased level of [Ca2+]i following Ca2+ entry through voltage-activated Ca2+ channels when stepping from −60 to +60 mV (see below for further explanation).
Ryanodine sensitivity of caffeine-induced Ca2+ release
The effect of ryanodine (20 μm) on caffeine-induced Ca2+ transients was examined in intact pyramidal cells loaded with fura-2 AM (n= 19). At this concentration, ryanodine was shown to bind to RyR channels and ‘lock’ them in an open subconducting state, leading to depletion of the ryanodine-sensitive Ca2+ stores (Rousseau, Smith & Meissner, 1987; Bezprozvanny, Watras & Ehrlich, 1991). As illustrated in Fig. 3A, the drug completely abolished the caffeine-induced Ca2+ transients in less than 10 min in all cells tested. Further, the effects of ryanodine on depolarization-versus caffeine-induced Ca2+ transients were compared in patch-clamped neurones loaded with fura-2 (n= 2). As shown in Fig. 3B, bath application of ryanodine completely abolished the caffeine-induced Ca2+ transients within 8 min, but not those evoked by depolarizing voltage pulses.
Taken together the data indicate that caffeine releases Ca2+ from ryanodine-sensitive Ca2+ stores in CA1 pyramidal cells. The fact that in virtually all pyramidal cells examined, Ca2+ transients were readily induced by caffeine prior to activation of any voltage-gated Ca2+ currents (e.g. Fig. 3A), indicates that even at rest the stores contain a releasable pool of Ca2+ ions.
Caffeine-induced Ca2+ release evokes a K+ current
In pyramidal neurones dialysed with an intracellular solution containing K+ as the main cation, the caffeine-induced Ca2+ transients were accompanied by outward membrane currents (Fig. 4; n= 4). The onset and peak of these outward currents were coincident with those of the associated Ca2+ transients. Such outward currents were never seen when the standard Cs+-containing intracellular solution was used (e.g. Figs 2 A, 3 B and 14 A). These observations demonstrate that Ca2+ release from ryanodine-sensitive stores increases [Ca2+]i near the plasma membrane to a level that is sufficient to activate Ca2+-activated K+ channels.
Depletion and spontaneous refilling of ryanodine-sensitive Ca2+ stores
During a paired-pulse caffeine-application protocol, when caffeine was applied again shortly after a conditioning application, the evoked Ca2+ transient was diminished in amplitude (Fig. 5A). Recovery of the Ca2+ transient occurred spontaneously and gradually over a course of 2 min. The pooled data obtained from fifty-six paired-pulse caffeine applications to twenty-nine pyramidal cells is plotted in Fig. 5B. Fitting a single exponential function to these data (Fig. 5B. continuous line) yielded a recovery time constant of 59 s. Extrapolating the fitted function to an interpulse interval of Δt= 0 (Fig. 5B, dashed line) yielded a Ca2+ transient of zero amplitude.
A likely explanation for these observations is that the conditioning caffeine pulse empties the ryanodine-sensitive Ca2+ stores almost completely. After caffeine application the ryanodine-sensitive Ca2+ stores refill spontaneously, albeit slowly. Consistent with this notion, multiple closely spaced (e.g. 0.1 Hz) applications of caffeine elicited progressively smaller Ca2+ transients (Fig. 6A). Apparent depletion of internal Ca2+ stores was seen in all cells exposed to repetitive caffeine applications (46 intact and 9 patch-clamped pyramidal cells).
It may be argued that when exposed to a high caffeine concentration or to repeated caffeine applications, the Ca2+ release mechanism undergoes a form of desensitization, which renders it unresponsive to caffeine for some time. To distinguish between store depletion and desensitization, the experiment described in Fig. 6 was performed in sixteen intact pyramidal cells loaded with fura-2 AM. Immediately after applying a series of caffeine pulses, which led to rundown of the Ca2+ transient, KCl was pressure applied from a nearby pipette (containing 80 mm KCl). In all cases, the KCl-induced rise in [Ca2+]i (presumably due to depolarization-induced Ca2+ entry through voltage-activated channels), was followed by a prompt recovery of the response to caffeine (Fig. 6B). A similar effect was seen when [Ca2+]i, was elevated by a depolarizing voltage pulse in whole-cell clamped neurones (see below Fig. 14; n= 5 cells).
These data suggest that caffeine applications empty, rather than desensitize, the ryanodine-sensitive Ca2+ stores and that store refilling is accelerated by increasing the availability of cytosolic Ca2+.
CPA and thapsigargin block refilling of ryanodine-sensitive Ca2+ stores
Uptake of cytosolic Ca2+ into endoplasmic reticulum stores is mediated by Ca2+-ATPases (SERCAs; see Tsien & Tsien, 1990; Pozzan, Rizzuto, Volpe & Meldolesi, 1994), known to be blocked reversibly by cyclopiazonic acid (CPA; Seidler, Jona, Vegh & Martonosi, 1989; Markram et al. 1995) and irreversibly by thapsigargin (Thastrup, Cullen, Drobak, Hartley & Dawson, 1990; Pozzan et al. 1994). To test the role of this uptake mechanism in the filling of neuronal ryanodine-sensitive Ca2+ stores, we examined the effects of these drugs on the spontaneous recovery of the caffeine-induced Ca2+ transients following repetitive caffeine applications that caused store depletion. As shown in Fig. 7, recovery in control saline was complete several minutes after the last caffeine pulse. However, when this protocol was repeated during exposure to 20 or 30 μm CPA (Fig. 7A; n= 6 cells) or to 1 or 3 μm thapsigargin (Fig. 7B; n= 12 cells), the Ca2+ transients failed to recover in all cells tested. Recovery occurred after wash of CPA (Fig. 7A). These findings demonstrate that refilling of ryanodine-sensitive Ca2+ stores is mediated by SPRCAs.
Interestingly, even without prior depletion of the Ca2+ stores by caffeine, exposure of the pyramidal cells to CPA for > 7 min completely abolished the caffeine-induced Ca2+ transients (Fig. 8; n= 6 out of 6 cells). The transients recovered after wash of CPA. These data suggest that the ryanodine-sensitive Ca2+ stores continuously leak Ca2+ and that the loss of Ca2+ is counterbalanced by continuous active uptake of Ca2+ into the stores.
Refilling of ryanodine-sensitive Ca2+ stores depends on extracellular Ca2+ influx
In order to characterize further the process of Ca2+ uptake into ryanodine-sensitive Ca2+ stores, we examined whether extracellular Ca2+ is required for spontaneous refilling of the stores. Perfusing the slices with Ca2+-free saline caused a small reduction (15.1 ± 11.3 nm; n= 6) in basal [Ca2+]i As shown in Fig. 9, the decrease in the magnitude of the Ca2+ transients induced by repetitive caffeine applications was markedly accelerated in Ca2+-free saline (containing 0.1 mm EGTA), suggesting a more rapid depletion of the ryanodine-sensitive Ca2+ stores. Furthermore, recovery of the Ca2+ transient was mostly blocked in this condition (6 out of 6 cells), suggesting that the stores do not refill when extracellular Ca2+ is removed. Indeed, reperfusion with standard saline prompted the recovery of the Ca2+ transient.
To test further the contribution of extracellular Ca2+ to the repletion of ryanodine-sensitive Ca2+ stores, we examined separately the effects of two broad-spectrum Ca2+ channel antagonists, namely Ni2+ and D600, on this process (Fig. 10). Adding either 2 mm Ni2+ (Fig. 10A) or 0.5 mm D600 (not shown) to the saline completely abolished KCl-induced [Ca2+]i increases, presumably by blocking voltage-activated Ca2+ channels in the plasmalemma. This effect was associated with a very small decrease in basal [Ca2+]i (3.0 ± 2.4 nm in Ni2+, n= 4; 8.2 ± 7.7 nm in D600, n= 9). Yet in all experiments the drugs almost completely blocked the recovery of the Ca2+ transient following store depletion by repetitive caffeine applications (Fig. 10). In contrast, recovery of the Ca2+ transient was not blocked by the selective L-type Ca2+ channel blocker nitrendipine at the concentration of 10 μm (n= 5 cells, not shown).
Taken together, the data suggest that the repletion of the ryanodine-sensitive Ca2+ stores depends on Ca2+ influx through Ni2+- and D600-sensitive, dihydropyridine-insensitive Ca2+ channels operating at resting potential.
The filling state of the ryanodine-sensitive Ca2+ stores depends on basal [Ca2+]i
In cerebellar Purkinje cells the amplitudes of caffeine -induced Ca2+ transients strongly depend on the basal [Ca2+]i (Kano et al. 1995). We tested whether ryanodine-sensitive Ca2+ stores in CA1 pyramidal neurones are similar in this respect. Caffeine-induced Ca2+ transients were induced before and after increasing basal [Ca2+]i with bath-applied KCl (20 or 40 mm). In different experiments, the KCl-elevated basal [Ca2+]i ranged from 200 to 640 nm. A representative experiment is illustrated in Fig. 11A. The initially low level of basal [Ca2+]i (42 nm) was associated with a caffeine-induced Ca2+ transient of 196 nm. After basal [Ca2+]i was elevated to 284 nm by 40 mm KCl, the Ca2+ transient increased to 683 nm (i.e. 348% of control). The decay of the large Ca2+ transient was followed by a noticeable undershoot in basal [Ca2+]i (Fig. 11A).
Similar results were obtained in twenty-six pyramidal neurones. These are summarized in the histogram in Fig. 11B, showing clearly that the amplitude of caffeine-induced Ca2+ transients increases with basal [Ca2+]i. One possible explanation for this relation is that elevated [Ca2+]i sensitizes the RyRs, so that a greater fraction of stored Ca2+ is released by a given caffeine pulse (Bezprozvanny et al. 1991; Kano et al. 1995). Alternatively, raising [Ca2+]i may enhance Ca2+ uptake into the stores so that more Ca2+ is available for release by a caffeine stimulus. To distinguish between these two explanations, the caffeine pulse was applied after the KCl-elevated [Ca2+]i returned to its initial level. As shown in Fig. 12A (upper panel), the evoked Ca2+ transient was still supranormal compared with its control. Indeed, recovery of the augmented Ca2+ transient to baseline value (Fig. 12A, lower panel) lagged 6–9 min behind the peak of the KCl-induced increase in [Ca2+]i (Fig. 12B).
Furthermore, when the experimental protocols described in Figs 11 and 12 were applied sequentially to the very same cell, we observed a marked similarity between caffeine-mediated Ca2+ transients evoked immediately after similar KCl applications and transients evoked in conjunction with KCl applications (Fig. 13). The ratio of the corresponding amplitudes of the caffeine-mediated Ca2+ transients evoked at high basal levels of [Ca2+]i and the amplitudes of the Ca2+ transients evoked just after the KCl applications was 1.14 ± 0.46 (n= 7).
These results suggest that ryanodine-sensitive Ca2+ stores in CA1 pyramidal cells have a high storage capacity and that the amount of Ca2+ sequestered by the stores is directly proportional to the availability of cytosolic Ca2+ ions. Thus, transient or sustained elevations of [Ca2+]i seem to lead to an increased amount of calcium ions stored inside the store.
Depolarization-induced Ca2+ transients are not dependent on the filling state of the ryanodine-sensitive Ca2+ stores
In some neurones Ca2+ entry through voltage-gated Ca2+ channels was shown to induce release of Ca2+ from ryanodine-sensitive Ca2+ stores (Friel & Tsien, 1992; Hua et al. 1993; Llano et al. 1994). To test for the presence of CICR, we examined whether depolarization-evoked Ca2+ transients are affected by depletion of ryanodine-sensitive Ca2+ stores (Fig. 14). If CICR contributes to the depolarization-induced Ca2+ transients, then these transients should be smaller after store depletion.
In five patch-clamped, fura-2-loaded pyramidal cells, Ca2+ transients were evoked by 800 ms depolarizing pulses. A representative experiment is illustrated in Fig. 14A. It can be seen that the transients evoked before and after store depletion were nearly the same. One may be concerned, however, by the relatively slow sampling rate (1 Hz) or relatively high intracellular fura-2 concentrations (up to 200 μm) used in these experiments. Therefore, we repeated these experiments in cells loaded with fura-2 AM using pressure KCl applications (80 mm KCl in the application pipette) and 4 Hz sampling rate. (Fig. 14B). Also in this case, in which we estimated the intracellular fura-2 concentration to be around 50 μm, the depletion of the caffeine-sensitive stores by six repetitive caffeine applications had virtually no effect on the amplitude of depolarization-induced Ca2+ transients. The mean amplitude of the KCl-mediated Ca2+ transient after the depletion of the stores related to control was 1.01 ± 0.16 (57 experiments in 36 cells). However, when the data were grouped with respect to the amplitude of KCl-mediated Ca2+ transients in control, we found a small effect of store depletion for the amplitudes of more than 200 nm (0.89 ± 0.10, n= 16) and no effect for the amplitudes of less than 200 nm (1.05 ± 0.16, n= 41). There was a significant difference between these two groups of data (P < 0.001, Student's two-tailed t test).
Taken together, these data suggest that CICR from ryanodine-sensitive Ca2+ stores does not contribute notably to the small depolarization-induced Ca2+ transients (amplitude < 200 nm) but may contribute significantly to the Ca2+ transients with amplitudes above 200 nm.
Caffeine evokes local Ca2+ transients in the dendrites of pyramidal cells
To examine the distribution of ryanodine-sensitive Ca2+ stores in CA1 pyramidal cells, caffeine was locally applied to different parts of the apical dendrites as well as to the soma. An exemplary experiment is illustrated in Fig. 15A. Caffeine applied either to the soma (left panel) or to the apical dendrite (right panel) induced a Ca2+ transient spatially restricted to the region under the application pipette. The signals evoked in the dendrites (Fig. 15B) had amplitudes that were similar to those seen in the soma.
Similar results were obtained at nine application sites in six out of six cells tested. Thus ryanodine-sensitive Ca2+ stores also preside in the dendrites of CA1 pyramidal cells and can generate independent, spatially delimited Ca2+ transients in response to local caffeine application.
The purpose of this study was to gain insight into the functional properties of ryanodine-sensitive Ca2+ stores in CA1 hippocampal pyramidal neurones. Using the RyR agonist caffeine as a pharmacological tool, we found that functional Ca2+ stores reside in the somata and dendrites of CA1 pyramidal cells. They have a high capacity for sequestering Ca2+ and are partially filled at rest. Filling of the stores is mediated by SERCAs and it is controlled by levels of cytosolic Ca2+. Depleted by agonist application, ryanodine-sensitive Ca2+ stores refill spontaneously utilizing a Ca2+ entry pathway operating at resting membrane potential.
Ryanodine-sensitive Ca2+ stores in hippocampal pyramidal cells
Several lines of evidence support the notion that caffeine triggered the release of Ca2+ from intracellular ryanodine-sensitive Ca2+ stores. Firstly, the caffeine-induced Ca2+ transients were not associated with an inward transmembrane current. Secondly, these transients persisted in Ca2+ -free saline, while voltage-activated Ca2+ currents disappeared. Thirdly, caffeine also evoked Ca2+ transients in cells which were held at +60 mV (approximate Ca2+ reversal potential), at which transmembrane Ca2+ influx is supposedly negligible. Finally, the action of caffeine was completely blocked by 10 or 20 μm ryanodine. We have previously shown (Garaschuk et al. 1996) that intracellular dialysis with 20 μm Ruthenium Red, known to block Ca2+ release from ryanodine-sensitive stores (Henzi & MacDermott, 1992; Ehrlich, Kaftan, Bezprozvannaya & Bezprozvanny, 1994), also blocked the Ca2+ responses to caffeine in CA1 pyramidal neurones.
It should be stressed, that in the present study caffeine evoked large Ca2+ responses even without prior ‘loading’ of intracellular Ca2+ stores via activation of voltage-activated Ca2+ channels. Likewise, in a recent study in cultured mouse hippocampal neurones (Seymour-Laurent & Barish, 1995), pressure-applied caffeine unconditionally evoked conspicuous Ca2+ transients. Thus, the ryanodine-sensitive Ca2+ stores in hippocampal pyramidal cells contain a releasable pool of Ca2+ even at rest. This conclusion is incongruent with a previous study in cultured hippocampal neurones (Shmigol et al. 1994), which suggested that ryanodine-sensitive Ca2+ stores are empty at rest. The apparent discrepancy may arise from the slow bath-application technique used in the latter study. Indeed, we also found that inclusion of 20 mm caffeine in the perfusing saline evoked much smaller Ca2+ responses (O. Garaschuk, Y. Yaari & A. Konnerth, unpublished observations). These data suggest that the rate of Ca2+ clearance from the cytosol may be nearly as fast as the rate of Ca2+ release by bath-applied caffeine. Therefore, rapid caffeine applications, as used in this study, are more appropriate to resolve functional states of ryanodine-sensitive Ca2+ stores.
When pyramidal neurones were dialysed with K+ -containing intracellular solution designed to spare K+ channel permeability, caffeine-induced Ca2+ transients concurred with an outward current. This current, not seen in neurones dialysed with Cs+, TEA-containing saline, most probably represents a K+ current activated by the elevated [Ca2+]i. A caffeine-induced Ca2+ -activated K+ current was shown earlier in peripheral and central neurones including hippocampal pyramidal cells (Sah & McLachlan, 1991; Uneyama, Munakata & Akaike, 1993). The absence of a noticeable delay between the rise in [Ca2+]i and the activation of the outward current suggests that Ca2+ is released near the plasmalemma. Indeed, neuronal endoplasmic reticulum may extend very close to this membrane (Henkart, Landis & Reese, 1976).
Release and sequestration of Ca2+ by ryanodine-sensitive Ca2+ stores
When two identical caffeine pulses were applied in rapid succession, the second caffeine-evoked Ca2+ transient was smaller than the first. Likewise, a short series of caffeine pulses resulted in a progressive run-down of the evoked Ca2+ responses. Subsequent recovery of these responses occurred spontaneously over 3–4 min (with a time constant of 59 s, assuming mono-exponential recovery process), but could be accelerated by interposing a depolarizing voltage or KCl pulse. Thus the use-dependent decrease of the Ca2+ transients most probably reflects depletion of the ryanodine-sensitive Ca2+ stores, rather than desensitization of the caffeine-binding RyRs.
Exposure of the pyramidal cells to the SERCA blockers thapsigargin and CPA abolished refilling of the depleted Ca2+ stores. This finding suggests that Ca2+ sequestration in ryanodine-sensitive Ca2+ stores is mediated predominantly by SERCAs. Even without prior store depletion, the responses to caffeine were abolished in CPA-containing saline, suggesting that ryanodine-sensitive Ca2+ stores are spontaneously leaking, and that maintenance of these stores in a filled state requires continuous Ca2+ pumping by the SERCAs.
Raising basal [Ca2+]i with high-K+ saline augmented the size of the caffeine-induced Ca2+ transients, as recently shown also in cerebellar Purkinje neurones (Kano et al. 1995). One explanation for this finding is that cytosolic Ca2+ interacts co-operatively with caffeine at the ryanodine receptor to augment the open probability of the receptor channel (Bezprozvanny et al. 1991), thereby increasing the release of stored Ca2+. Alternatively, elevated basal [Ca2+]i may enhance Ca2+ sequestration into the stores, so that more Ca2+ is available for release by caffeine. Our data showing that the caffeine-induced Ca2+ transients were similarly augmented immediately after basal [Ca2+]i declined to control value, and remained elevated up to 6 min thereafter (Figs 12 and 13), favours the enhanced Ca2+ sequestration hypothesis.
Extracellular Ca2+ entry into ryanodine-sensitive Ca2+ stores
Perfusing the slices with Ca2+ -free saline or with saline solutions containing the Ca2+ channel antagonists Ni2+ or D600, prevented the spontaneous refilling of depleted ryanodine-sensitive Ca2+ stores. This finding seen in both fura-2 AM-loaded and in whole-cell patch-clamped (holding potential, −60 mV) pyramidal neurones, demonstrates that store refilling requires a Ca2+ influx pathway that is active at the resting membrane potential.
To account for this Ca2+ influx, we hypothesize the existence of a class of Ca2+ channels in the plasmalemma, activated by depletion of ryanodine-sensitive Ca2+ stores. This mechanism would be analogous to the ‘capacitative’ Ca2+ entry in non-neural cells, activated by agonist-induced depletion of inositol 1,4,5-trisphosphate-sensitive Ca2+ stores (see Tsien & Tsien, 1990; Berridge, 1995). Interestingly, Ni2+ (2 mm), which blocked refilling of depleted ryanodine-sensitive Ca2+ stores, is an effective blocker of capacitative Ca2+ entry in non-neural cells (Tsien & Tsien, 1990; Hoth & Penner, 1993).
Capacitative Ca2+ entry in some non-neural cells is manifested by a small inward current and an associated rise in [Ca2+]i (Hoth & Penner, 1993). In CA1 pyramidal cells we did not detect a significant inward current or change in [Ca2+]i associated with the refilling of depleted ryanodine-sensitive Ca2+ stores. Possibly the route of capacitative-like Ca2+ entry is spatially restricted to cellular microdomains, where the endoplasmatic reticulum encroaches on the plasmalemma, rendering it invisible to the indicator dye (Putney, 1986; Tsien & Tsien, 1990).
In addition to the capacitative-like Ca2+ entry pathway, operating at the resting membrane potential, voltage-gated Ca2+ channels represent a powerful mechanism responsible for the activity-dependent refilling of ryanodine-sensitive Ca2+ stores. Indeed, ryanodine-sensitive Ca2+ stores in hippocampal pyramidal cells refilled promptly following the activation of voltage-gated Ca2+ channels. This finding is in agreement with properties of caffeine-sensitive stores found in other types of peripheral and central neurones (Brorson et al. 1991; Friel & Tsien, 1992; Shmigol et al. 1994).
In conclusion, the refilling of the ryanodine-sensitive Ca2+ stores in hippocampal pyramidal cells critically depends upon transmembrane Ca2+ entry. Two distinct mechanisms, the newly identified capacitative-like Ca2+ entry pathway and voltage-activated Ca2+ channels, are utilized for the Ca2+ entry at rest and during neuronal activity.
Role of ryanodine-sensitive Ca2+ stores in shaping physiological Ca2+ transients
The scheme in Fig. 16 represents a hypothetical framework for the results of the present study. It shows that high capacity ryanodine-sensitive Ca2+ stores residing in the endoplasmic reticulum communicate bidirectionally with the cytosol via the SERCAs on the one hand and the ‘leak’ and RyR channels on the other hand.
The SERCAs pump cytosolic Ca2+ into the ryanodine-sensitive Ca2+ stores, thereby creating a Ca2+ concentration gradient across the endoplasmic reticulum membrane. Their activity appears to be regulated by cytosolic [Ca2+]i. At rest, basal [Ca2+]i is low and the stores are only partially filled (Fig. 16). Elevation in [Ca2+]i as caused by transmembrane Ca2+ entry through voltage-gated Ca2+ channels, enhances SERCA activity and increases the amount of stored Ca2+. Through this action the ryanodine-sensitive Ca2+ stores may act as Ca2+ sinks, participating in the clearance of excess cytosolic Ca2+ associated with neuronal activity (e.g. Markram et al. 1995).
The ryanodine-sensitive Ca2+ stores constantly leak Ca2+ back into the cytosol. Therefore, physiological or pharmacologically induced decreases in SERCA activity lead to slow dissipation of the Ca2+ gradient across the endoplasmic reticulum membrane and to store depletion. The leak of Ca2+ does not show up as an increase in basal [Ca2+]i, indicating that it is effectively removed out of the cell or buffered within the cell by other mechanisms (not illustrated in Fig. 16). In contrast, stimulation of RyR channels leads to a rapid discharge of stored Ca2+, which appears as a large transient increase in [Ca2+]i. This Ca2+ transient occurs near the plasmalemma, because it triggers an outward K+ current via Ca2+ -activated K+ channels (Fig. 16). The size of the Ca2+ transient reflects the amount of Ca2+ stored prior to RyR stimulation.
Subsequent replenishment of the stores is achieved by the activity of the SERCAs and depends critically on extracellular Ca2+ entry through putative ‘capacitative’ Ca2+ channels in the plasmalemma (Fig. 16).
In this study, depleting or blocking the ryanodine-sensitive Ca2+ stores (with caffeine or ryanodine) did not significantly alter the amplitude and waveform of depolarization-induced Ca2+ transients (Figs 3 and 14). This would suggest that CICR does not contribute to these Ca2+ transients under our experimental conditions. It should be noted, however, that our experimental conditions were not optimal for evoking the CICR process. In our experiments we evoked rather small depolarization-induced Ca2+ transients (up to 500 nm) resulting from relatively small transmembrane Ca2+ currents (0.5–1.5 nA). In skeletal muscle, for example, the threshold for CICR was found to be 1.5–2 μm (see Henzi & MacDermott, 1992) and the CICR in cerebellar Purkinje cells (Llano et al. 1994) was evoked by the transmembrane Ca2+ currents of 4–8 nA giving rise to Ca2+ transients with amplitudes reaching 1–2.5 μm. Moreover, no CICR was found in cerebellar Purkinje cells when the amplitude of depolarization-induced Ca2+ transient was less that 500 nm (Llano et al. 1994; Kano et al. 1995).
It can be concluded, therefore, that in central neurones CICR does not contribute significantly to depolarization-induced Ca2+ transients evoked by low frequency activity. However, according to a recent study (Alford et al. 1993) CICR from ryanodine-sensitive Ca2+ stores seems to generate most (65%) of the Ca2+ transient in CA1 pyramidal cell dendrites during tetanic activation of excitatory synapses. Thus, CICR may be a major source for the Ca2+ transients associated with intense neuronal discharges.
We thank Nicole Rothgerber and Erie Eilers for excellent technical help. This work was supported by grants from the Deutsche Forschungsgemeinschaft (SFB 246), the Bundesministerium für Bildung, Wissenschaft, Forschung und Technologie and the German–Israeli Foundation for Scientific Research and Development (GIF).