ATP Regulation of Type 1 Inositol 1,4,5-Trisphosphate Receptor Channel Gating by Allosteric Tuning of Ca2+ Activation*

Inositol 1,4,5-trisphosphate (InsP3) mobilizes intracellular Ca2+ by binding to its receptor (InsP3R), an endoplasmic reticulum-localized Ca2+ release channel. Patch clamp electrophysiology of Xenopus oocyte nuclei was used to study the effects of cytoplasmic ATP concentration on the cytoplasmic Ca2+ ([Ca2+] i ) dependence of single type 1 InsP3R channels in native endoplasmic reticulum membrane. Cytoplasmic ATP free-acid ([ATP] i ), but not the MgATP complex, activated gating of the InsP3-liganded InsP3R, by stabilizing open channel state(s) and destabilizing the closed state(s). Activation was associated with a reduction of the half-maximal activating [Ca2+] i from 500 ± 50 nm in 0 [ATP] i to 29 ± 4 nm in 9.5 mm [ATP] i , with apparent ATP affinity = 0.27 ± 0.04 mm, similar to in vivo concentrations. In contrast, ATP was without effect on maximum open probability or the Hill coefficient for Ca2+activation. Thus, ATP enhances gating of the InsP3R by allosteric regulation of the Ca2+ sensitivity of the Ca2+ activation sites of the channel. By regulating the Ca2+-induced Ca2+ release properties of the InsP3R, ATP may play an important role in shaping cytoplasmic Ca2+ signals, possibly linking cell metabolic state to important Ca2+-dependent processes.

Modulation of free cytoplasmic Ca 2ϩ concentration ([Ca 2ϩ ] i ) is a ubiquitous cellular signaling system. In many cell types, binding of ligands to plasma membrane receptors activates the hydrolysis of phosphatidylinositol 4,5-bisphosphate by membrane-bound phospholipase C, generating inositol 1,4,5trisphosphate (InsP 3 ). 1 InsP 3 causes the release of Ca 2ϩ from the endoplasmic reticulum (ER) by binding to its receptor (InsP 3 R), which itself is a Ca 2ϩ channel (1)(2)(3). Complex control of Ca 2ϩ release through the InsP 3 R by various intracellular factors, including cooperative activation by InsP 3 (4 -8) and biphasic feedback from the permeant Ca 2ϩ ion (6, 8 -11) generates intricate [Ca 2ϩ ] i signals that can be manifested temporally as repetitive spikes or oscillations, with frequencies often tuned to the level of stimulation, and spatially as propagating waves or highly localized events (2,12,14) and display properties of "adaptation" and "quantal release," which are poorly understood (15). Several types of InsP 3 R as products of different genes with alternatively spliced isoforms have been identified and sequenced (16,17). The InsP 3 Rs have about 2700 amino acid residues in InsP 3 binding, regulatory (modulatory) and transmembrane channel domains (16 -18). The sequences of the regulatory domains of all InsP 3 R isoforms include putative ATP-binding site(s) (17). ATP was shown to bind to the InsP 3 R (19) and regulate InsP 3 R-mediated Ca 2ϩ release (20 -24), although the detailed mechanisms of this regulation remain unclear.
Several studies have demonstrated that mitochondria and the ER are in close physical and functional proximity in many cell types, including neurons (24 -27). [Ca 2ϩ ] i signals generated by InsP 3 R-mediated Ca 2ϩ release from the ER appear to be rapidly and efficiently transmitted to mitochondria (28 -30), acutely affecting mitochondrial functions (31)(32)(33), including ATP synthesis (34). It is unknown, however, whether the communication between these two organelles is reciprocal. ATP release from mitochondria, globally into the cytoplasm and locally into the vicinity of the InsP 3 R channels that are in close apposition, may provide a signaling pathway for communication from the mitochondria back to the ER. Thus, regulation of the InsP 3 R by ATP could have considerable significance for intracellular signaling, particularly if the channel is sensitive to ATP levels in normal physiological as well as pathological conditions, including ischemia.
Most previous studies of ATP regulation of the InsP 3 R have been limited to indirect measurements, i.e. Ca 2ϩ fluxes or concentrations, to infer InsP 3 R channel activity, because the intracellular location of the Ca 2ϩ release channel has limited its accessibility to electrophysiological approaches. Furthermore, only a limited range of [Ca 2ϩ ] i was examined in previous studies, despite the fact that the InsP 3 R is intricately regulated by [Ca 2ϩ ] i (6, 8 -11) and that the primary known regulator of the channel, InsP 3 , mediates its effects by modulating the [Ca 2ϩ ] i dependence of channel gating (8). Therefore, in the present study, we have systematically investigated the effects of cytoplasmic ATP concentration on the [Ca 2ϩ ] i response of single InsP 3 R channels. We applied the patch clamp technique to isolated Xenopus oocyte nuclei (35)(36)(37) to study the single channel activities of the type 1 InsP 3 R (InsP 3 R-1), the major brain isoform (38,39), in its native ER membrane environment under rigorously defined conditions on both the cytoplasmic and luminal sides of the channel. Our results demonstrate that cytoplasmic ATP free acid, but not cytoplasmic MgATP complex, activates the gating of the InsP 3 R primarily by allosteric regulation of the [Ca 2ϩ ] i sensitivity of the Ca 2ϩ activation sites of the channel. By regulating the Ca 2ϩ -induced Ca 2ϩ release properties of the InsP 3 R, ATP may play an important role in shaping the extent and duration of [Ca 2ϩ ] i signals, possibly linking cell metabolic state to important Ca 2ϩ -dependent process including synaptic plasticity.

EXPERIMENTAL PROCEDURES
Patch Clamping the Oocyte Nucleus-Patch clamp experiments were performed as described in Refs. 8, 35, 36, and 40. Briefly, stage V or VI oocytes were opened mechanically just prior to use. The nucleus was separated from the cytoplasm and transferred to a dish on the stage of a microscope for patch-clamping. The oocyte expresses only a single InsP 3 R isoform (type 1) and lacks other (e.g. ryanodine receptor) Ca 2ϩ release channels (41). Experiments were done in "on-nucleus" configuration, with the solution in the perinuclear lumen between the outer and inner nuclear membranes in apparent equilibrium with the bath solution (35) and with the cytoplasmic aspect of the InsP 3 R channel facing into the patch pipette. Following standard conventions, the applied potential is that of the pipette electrode minus the reference bath electrode (positive current flows from pipette outward). Experiments were performed at room temperature with the pipette electrode at ϩ20 mV relative to the reference bath electrode.
Data Acquisition and Analysis-Single channel currents were amplified with an Axopatch-1D amplifier (Axon Instruments, Foster City, CA) with antialiasing filtering at 1 kHz, transferred to a Power Macintosh 8100 via an ITC-16 interface (Instrutech Corp., Great Lake, NY), digitized at 5 kHz, and written directly onto the hard disc by PulseϩPulseFit software (HEKA Elektronik, Lambrecht, Germany). Data were analyzed to identify channel opening and closing events and evaluate channel open probabilities using MacTac 3 (Bruxton, Seattle, WA). Each data point shown is the mean of results from at least four separate patch clamp experiments performed under the same conditions. Error bars indicate the S.E. Theoretical curves were fitted to experimental data using Igor Pro 3 (WaveMetrics, Lake Oswego, OR).
Solutions for Patch Clamp Experiments-All patch clamp experiments were performed with solutions containing 140 mM KCl and 10 mM HEPES with pH adjusted to 7.1 using KOH.

Activation of Channel
Gating by ATP-To examine the effects of ATP on the permeation and gating properties of the InsP 3 R, we included 10 M InsP 3 , 250 nM free Ca 2ϩ , and 0.5 mM free ATP in the pipette solution. Under these conditions, the endogenous Xenopus type 1 InsP 3 R channel exhibited channel conductance properties and kinetics similar to those observed previously under similar conditions (8,40). The channels gated with a moderately high open probability (P o ) of ϳ0.5 (Fig. 1A).
In similar experiments employing pipette solutions that lacked ATP, the InsP 3 R channel P o was significantly lower (ϳ0.2) in either the absence or presence of 3 mM Mg 2ϩ (Fig. 1, B and C).
To determine whether the ATP activation of the channel P o was mediated by MgATP, which could suggest a role for ATP hydrolyisis or phosphorylation, similar experiments were undertaken with 3 mM total Mg 2ϩ and 0.5 mM total ATP in the pipette. Under these conditions, [MgATP] is approximately 0.5 mM, and the free Mg 2ϩ concentration ([Mg 2ϩ ] i ) and the cytoplasmic free ATP concentration ([ATP] i ) were calculated to be 2.5 and 0.012 mM, respectively. Nevertheless, P o remained low (Fig. 1D). The low P o in the presence of MgATP ( Fig. 1) was solely caused by ATP complexation by Mg 2ϩ , since it was fully reversed by adding more ATP to the pipette solution to restore [ATP] i (Fig. 1E). Thus, MgATP has no effect, stimulatory or inhibitory, on InsP 3 R activity. We previously demonstrated that the P o of the Xenopus type 1 InsP 3 R is independent of [Mg 2ϩ ] i up to 9.5 mM (40). Taken together, these results suggest that ATP free acid (ATP 3Ϫ or ATP 4Ϫ ) was the relevant ionic species and that ATP hydrolysis was not involved in the stimulation of InsP 3 R channel gating.
Effects of ATP on the [Ca 2ϩ ] i Dependence of InsP 3 R Gating-InsP 3 activates the InsP 3 R by modulating the sensitivity of the channel to [Ca 2ϩ ] i (8). To determine the mechanism of ATP activation of the InsP 3 R channel gating, we investigated in detail the effects of ATP on the channel kinetics of the InsP 3 R over a wide range of [Ca 2ϩ ] i . A systematic series of patch clamp experiments were performed using pipette solutions containing various [Ca 2ϩ ] i with 0.5 mM ATP alone, 3 mM Mg 2ϩ alone, 0.5 mM ATP, and 3 mM Mg 2ϩ (calculated [ATP] i ϭ 0.012 mM; calculated [Mg 2ϩ ] i ϭ 2.5 mM) or no ATP or Mg 2ϩ . To avoid possible effects of Ca 2ϩ on InsP 3 binding, a functionally saturating InsP 3 concentration of 10 M was used (8). The [Ca 2ϩ ] i sensitivity of the InsP 3 R in the absence of cytoplasmic free ATP was biphasic (Fig. 2) and could be well fitted with a biphasic Hill equation similar to the following one previously derived for the InsP 3 R in the presence of 0.5 mM cytoplasmic free ATP (8).
Similar results were obtained independent of the presence or absence of either Mg 2ϩ or MgATP in the pipette solutions. This result indicates that the InsP 3 R can achieve a maximum open probability P max of 0.79 in the absence of cytoplasmic free ATP, a level of activity very similar to the P max of 0.81 found in the presence of 0.5 mM cytoplasmic free ATP (8). Thus, ATP does not activate the channel by increasing P max . The Hill coefficient for Ca 2ϩ activation H act was 2.4 Ϯ 0.6 in the absence of free ATP, similar to H act ϭ 1.9 Ϯ 0.3 in the presence of 0.5 mM free ATP. This result suggests that Ca 2ϩ probably activates the InsP 3 R via the same cooperative process in either the presence or absence of cytoplasmic free ATP. Thus, ATP does not activate the channel by modulating H act . The observed activation of the InsP 3 R by cytoplasmic free ATP ( Fig. 1) was associated with a reduction of the half-maximal activating [Ca 2ϩ ] i (K act ) from 500 Ϯ 50 nM in the absence of free ATP to 190 Ϯ 20 nM in the presence of 0.5 mM free ATP. Thus, ATP activates the channel by sensitizing it to Ca 2ϩ . ATP therefore enhances Ca 2ϩ -induced Ca 2ϩ release (CICR) by the InsP 3 R.
Because InsP 3 activates channel gating by modifying the [Ca 2ϩ ] i inhibition phase of the channel [Ca 2ϩ ] i dependence (8), we were also interested in examining the effects of ATP on this aspect of the response. However, investigations of InsP 3 R channel activity in the absence of free ATP at [Ca 2ϩ ] i which inhibit channel gating (Ͼ20 M; Ref. 8) were not possible because of the unavailability of a Ca 2ϩ chelator with the appropriate Ca 2ϩ affinity. In our previous experiments that examined the effects of high [Ca 2ϩ ] i on InsP 3 R channel gating (8), ATP was used as the Ca 2ϩ chelator for buffering [Ca 2ϩ ] i at high [Ca 2ϩ ] i . The data we were able to obtain in the absence of ATP in the present study indicated that P o began to decrease as [Ca 2ϩ ] i was increased beyond 10 M, but the inhibitory half-maximal [Ca 2ϩ ] i , K inh , or Hill coefficient, H inh , could not be determined accurately (Fig. 2).   (Fig. 5).

Effects of ATP on InsP 3 R Channel Kinetics-Analysis
[ATP] i decreased K act of the InsP 3 R by over an order of magnitude, from 500 Ϯ 50 nM in 0 mM [ATP] i to 29 Ϯ 4 nM in 9.5 mM [ATP] i . The effects of [ATP] i on K act of the InsP 3 R were analyzed by fitting the data with a modified Michaelis-Menten equation (Fig. 6).  ity of the stimulatory effects we observed for free ATP, we also investigated the effects of adenosine, AMP, ADP, GTP, and UTP. Each nucleotide was present as 0.5 mM free nucleotide in the absence of Mg 2ϩ . The [Ca 2ϩ ] i was fixed at 220 Ϯ 15 nM, because the P o is very sensitive to activation by free ATP at this [Ca 2ϩ ] i (Fig. 2). Using the channel P o (0.14) in the absence of any nucleotide as the reference, the relative P o of the channel was determined in the presence of the various nucleotide species (Fig. 7). Similar to the MgATP complex, free UTP, ADP, and adenosine had no effects on the P o of the InsP 3 R (p Ͼ 0.05). In contrast, free ATP, AMP, and GTP each activated the InsP 3 R (p Ͻ 0.05). Free ATP had the greatest effect, more than tripling the P o , whereas both free AMP and GTP doubled the channel P o . DISCUSSION We previously described the detailed permeation and gating properties of the Xenopus type 1 InsP 3 R channel by patch clamp studies of isolated oocyte nuclei (8,35,36,40). Under optimal conditions, gating of the channel is robust, with maximum open probability of ϳ80% over a wide range of [Ca 2ϩ ] i (8). Importantly, the gating of the channel is regulated by both The results from the present study suggest that ATP stimulates gating of the InsP 3 R by modulating the Ca 2ϩ sensitivity of the Ca 2ϩ activation sites.
Regulation of InsP 3 -mediated Ca 2ϩ Release by ATP-We performed a systematic investigation of the effects of nucleotides on gating of the Xenopus type 1 InsP 3 R Ca 2ϩ release channel. Our study focused on the effects of ATP on single channel activity, and included additional examination of effects of other nucleotides for comparison. Stimulation of InsP 3 R by nucleotides in the presence of InsP 3 has been previously reported (20 -23, 42, 43). The majority of published studies measured the effects of nucleotides on InsP 3 -induced Ca 2ϩ fluxes either from intracellular stores in permeabilized cells or into lipid vesicles in vitro. ATP stimulation of InsP 3 -induced Ca 2ϩ fluxes was reported in all these studies, but the considerable qualitative as well as quantitative variability among them makes comparison with our results difficult. Maximum stimulation by ATP of Ca 2ϩ release by the InsP 3 R has been reported to range from 1.5-to 2-fold (20 -22) the activity with InsP 3 alone. Considerable discrepancies exist regarding the ATP concentration required for maximal stimulation, ranging from 10 M (20) to 1 mM (21,22). In terms of the nucleotide specificity, ADP was reported to be as potent as ATP (21,22) or to be only 40% as effective (20) or ineffective (24); AMP has been reported to stimulate to ϳ70% (21,22) or 10% (20) of the ATP stimulation level; GTP was ineffective (21) or stimulated to 30% of the ATP stimulation level (22). Regarding the specific ATP species, MgATP complex was as effective as free ATP (20) or only 50% as effective (21).
Some reports indicate that the effects of ATP on InsP 3 R activity are biphasic, being stimulatory at low [ bilayer experiments, ATP concentrations of Ͼ5 mM inhibited channel activity with a half-maximal concentration of 11 mM (23). In contrast, recent Ca 2ϩ flux studies found no inhibitory effects of ATP at 5 (24) or 10 mM (42) on InsP 3 -induced Ca 2ϩ release from the ER. The results from our nuclear patch clamp experiments are in agreement with these latter studies. We detected no inhibitory effects of cytoplasmic ATP up to 10 mM. It has been suggested that the inhibitory effects of high [ATP] i are caused by competitive inhibition of InsP 3 binding to the InsP 3 R (23). Therefore, the different observations may be caused by different [InsP 3 ] i used in these studies, Ͻ5 M in studies reporting ATP inhibition (20 -23) and Ͼ5 M in studies reporting no ATP inhibition (Refs. 24 and 42 and this study).
Stimulation of Single Channel Gating of the InsP 3 R by ATP-Many of the discrepancies in the results of Ca 2ϩ flux studies may have been caused by species differences, different concentrations of InsP 3 , or other important parameters, including Ca 2ϩ and other divalent cations (i.e. Mg 2ϩ ) whose concentrations in the vicinity of the InsP 3 R might not have been adequately controlled. Furthermore, the measurements of Ca 2ϩ fluxes involved populations of unknown numbers and multiple types of InsP 3 R, which may also have contributed to the discrepant results (42). Importantly, the effects of nucleotides on the activity of the InsP 3 R on the single molecule level can only be inferred from these studies. There has been only one previous detailed investigation of the effects of ATP on single channel activity of the InsP 3 R (23). In the presence of 0.2 M [Ca 2ϩ ] i and 2 M InsP 3 , ATP enhanced the P o of canine cerebellum type 1 InsP 3 R reconstituted from microsomes into artificial planar lipid bilayers with a binding coefficient of 40 M and Hill coefficient of 1. The MgATP complex was as effective as free ATP, whereas GTP was only 20% as effective and AMP was ineffective.
The effects of nucleotides on the Xenopus type 1 channel observed by nuclear patch clamp in the present study have some similarities but also differ in several important respects. In agreement with the results from the bilayer study, the Hill coefficient for ATP activation was also 1 in our study, although the apparent binding coefficient K ATP for free ATP was ϳ300 M in our study, compared with 40 M in the bilayer study. In both studies, ATP activation of channel gating was associated with stabilization of the open state and destabilization of the closed state. Both studies concluded that ATP hydrolysis was not involved, since nonhydrolyzable analogs were as effective as ATP in the bilayer study. In contrast to the lipid bilayer results, the MgATP complex had no stimulatory effects in our study; furthermore, free GTP and AMP had potent stimulatory effects (both about 60% of the ATP stimulation level) in the present study. The difference in nucleotide specificity observed in the two studies is difficult to reconcile. It may be caused by the different sources (canine versus Xenopus), lipid environment (artificial lipid bilayers versus native ER), and/or the isolation/reconstitution protocols. Of particular interest is our finding that MgATP was without effect on the channel activity. This distinction is important; whereas the MgATP concentration in the cytoplasm is in the range of 3-8 mM (44 -48), the cytoplasmic free ATP concentration is in the range of 400 -600 M. The apparent affinity of 40 M for ATP stimulation of the InsP 3 R in bilayers and the apparent effectiveness of MgATP substituting for ATP in InsP 3 R stimulation would suggest that ATP (free or complexed with Mg 2ϩ ) plays no role in channel modulation under most conditions, since the total concentration of free and magnesium-complexed ATP always far exceeds its apparent affinity for the channel. In contrast, and as discussed in more detail below, our determination that only free ATP is effective, with an apparent affinity (apparent K D ϳ 270 M) that is nearly coincident with levels in the cytoplasm, suggests that physiological modulation of ATP levels in cells will have profound effects on Ca 2ϩ release activity of the InsP 3 R by modifying its CICR properties. Of note, a recent study using permeabilized lymphocytes that were engineered to express only the type 1 InsP 3 R found that free ATP enhanced InsP 3 -induced Ca 2ϩ release with an apparent K D of 390 M (42), in quite good agreement with our channel results.
A second important distinction between our results and those of the bilayer study concerns the mechanism by which ATP stimulates channel gating. Our results demonstrate that ATP binding to the InsP 3 R increases the Ca 2ϩ sensitivity of the Ca 2ϩ activation site of the channel. The bilayer experiments were carried out at a single [Ca 2ϩ ] i (0.2 M). The P o observed in 2 M InsP 3 increased from ϳ0.01 in the absence of ATP to ϳ0.07 in 1 mM ATP (the order of magnitude difference in absolute values of P o observed in the two studies may be due to an InsP 3 insensitivity of the reconstituted channels in the bilayer study (8)). It was concluded that ATP binding to InsP 3 R increases the intrinsic efficacy of InsP 3 to activate the InsP 3 R (23). However, our results demonstrate that ATP does not affect P max . The increased P o observed in the bilayer study can be accounted for by the increase in affinity of the Ca 2ϩ activation site observed in the present study (Fig. 5).

Regulation of the [Ca 2ϩ ] i Dependence of InsP 3 R Gating by ATP-
The present study represents the first systematic investigation of the effects of ATP on single channel activity of the InsP 3 R over a wide range of [Ca 2ϩ ] i (10 nM to 20 M). Our results reveal the major mechanism by which elevations of cytoplasmic free ATP stimulate gating of the InsP 3 R, by demonstrating that ATP increases the affinity of the Ca 2ϩ activating site of the channel specifically. ATP decreased the halfmaximal activating [Ca 2ϩ ] i (K act ), without affecting the maximum P o . Although channel P o decreased at low [ATP] i , this could be fully reversed by increased [Ca 2ϩ ] i . Therefore, ATP is not a necessary agonist for activation of the InsP 3 R, but rather it is an allosteric regulator, tuning the efficacy of [Ca 2ϩ ] i to stimulate the activity of the InsP 3 -liganded InsP 3 R over a limited range of [Ca 2ϩ ] i (10 nM to 1 M as shown in Fig. 5). We previously showed that InsP 3 activates the InsP 3 R solely by tuning the half-maximal inhibitory [Ca 2ϩ ] i (K inh ) of the channel, whereas activation of the InsP 3 R by [Ca 2ϩ ] i is unaffected (8). InsP 3 is therefore a regulator of Ca 2ϩ -inhibition of Ca 2ϩ release, whereas ATP is a regulator of CICR. Thus, the effect of free ATP on the activation of the InsP 3  inhibition. This mode of regulation is therefore analogous to that of InsP 3 . Further studies are required to define the effects of free ATP on Ca 2ϩ inhibition of the channel and the relationship of those effects to that of InsP 3 . Nevertheless, it is interesting to consider the similarities involved in the regulation of channel gating by ATP and InsP 3 . As mentioned earlier, InsP 3 is a channel activator because it decreases the affinity of the Ca 2ϩ inhibition site of the channel. It is important to note, however, that it is not the absolute magnitude of the Ca 2ϩ affinity of that site that is critical for InsP 3 R activity, but rather its relationship to that of the Ca 2ϩ activation site.
The higher Ca 2ϩ affinity of the inhibition site keeps the channel inactive in the absence of InsP 3 . However, the Ca 2ϩ affinity of the inhibition site becomes less than that of the activation site when the channel binds InsP 3 . It follows, therefore, that an alternate mechanism to activate the channel would be to increase the Ca 2ϩ affinity of the activation site. Because this is a major effect of ATP, we speculate that suitable conditions could be defined in which channel gating could be activated by an increase in [ATP] i without any change in [InsP 3 ] i . Because the relative affinities for Ca 2ϩ of the InsP 3 R activation and inhibition sites is the critical factor in determining the level of channel activity, it follows that allosteric regulation of antagonistic Ca 2ϩ -binding sites by ATP and InsP 3 , by together tuning the Ca 2ϩ dependence of channel gating, render the Ca 2ϩ dependence of Ca 2ϩ release by the InsP 3 R a dynamic property, dependent upon stimulus intensity and cell metabolic state.
Physiological  (44,45) and total ATP (46 -48) concentrations in cells are each in the range of 5-10 mM. Assuming equal concentrations of each, the free ATP concentration is calculated to vary roughly from 420 M (5 mM total Mg 2ϩ ) to 540 M (8 mM total Mg 2ϩ ). For comparison, the apparent affinity of the ATP-binding site on the InsP 3 R was determined in the present study to be 270 M. Interestingly, not only is the apparent affinity of the ATP-binding site on the InsP 3 R coincident with the normal cytoplasmic free ATP concentration, but changes in free ATP concentration are very sensitive to changes in total ATP concentration in the cytoplasm. For example, at 5 mM total Mg 2ϩ , an increase of total ATP concentration from 5 to 5.5 mM (10% change) will result in a 1.5-fold increase in the concentration of free ATP (from ϳ420 M to ϳ600 M). Thus, relatively small changes in total cytoplasmic ATP can have pronounced effects on free ATP concentration and, therefore, on the CICR properties of the InsP 3 R. Thus, our results suggest that the channel is poised in vivo to respond to changes in the free ATP concentration, for example those that may occur during ischemia. Therefore, the nucleotide sensitivity may enable Ca 2ϩ release properties of the InsP 3 R to be tuned to the metabolic state of the cell.
It has become evident from several recent studies that mitochondria and the ER form a tightly coupled, complex signaling unit. Imaging studies have revealed that mitochondria are in close physical proximity to the ER (25), especially to sites of Ca 2ϩ release (26). In cerebellar Purkinje cells, ER cisternae containing high densities of InsP 3 R-1 are often wrapped around or closely apposed to mitochondria (49,50). A physiological implication of this structural arrangement is that it enables the rise in [Ca 2ϩ ] i caused by agonist-stimulated InsP 3 R activity to be effectively transmitted as a transient increase in mitochondrial matrix [Ca 2ϩ ] that closely parallels the [Ca 2ϩ ] i rise, due to the locally high [Ca 2ϩ ] i in the microdomain of the release channels and rapid uptake of released Ca 2ϩ by the mitochondria (28 -30, 33, 51). The resulting changes in mitochondrial matrix [Ca 2ϩ ] affect the mitochondrial membrane potential (32) and the activities of the mitochondrial dehydrogenases that are crucial in ATP synthesis and intracellular ATP levels (27,31,33,34,52). The cytosolic ATP concentration in turn affects processes that contribute to [Ca 2ϩ ] i regulation, including InsP 3 -induced Ca 2ϩ release (Refs. 20 -24 and 43 and this study), passive leak from Ca 2ϩ stores (53), plasma membrane store-operated Ca 2ϩ entry (24,54,55), and Ca 2ϩ extrusion and uptake into the ER by Ca 2ϩ -ATPases (24,56,57). Thus, in addition to buffering [Ca 2ϩ ] i directly by active Ca 2ϩ sequestration and export (58 -62), mitochondria indirectly participate in intracellular Ca 2ϩ signaling, using cytosolic ATP as a global cytoplasmic messenger. We speculate that the close physical proximity of mitochondria and ER may enable local changes in ATP concentration, due to release from mitochondria into the microdomains of close ER-mitochondria apposition, to rapidly effect local InsP 3 R-mediated Ca 2ϩ release. Of significance, the ATP released by mitochondria is free ATP, the InsP 3 R ligand, not MgATP (63). Thus, communication between these two organelles may be two-way, with local Ca 2ϩ release as the currency of communication from ER to mitochondria and local ATP release providing the cross-talk from mitochondria to ER. The phosphoinositide signaling system is highly expressed throughout the brain (64,65). Recent observations suggest that Ca 2ϩ release from type 1 InsP 3 receptors is involved in nerve growth (66) and synaptic plasticity, including long term potentiation (67-69) and depression (LTD) (67, 67, 70 -72). Disruption of the mouse InsP 3 R-1 gene eliminates LTD in the cerebellum (70), and the competitive InsP 3 R inhibitor heparin blocks LTD in the neocortex (73). Metabotropic glutamate receptors (mGluR), which couple to the InsP 3 signaling pathway, have been implicated in synaptic plasticity (65,69,73), and mice with targeted disruption of mGluR1 show impaired LTD (74,75). InsP 3 -mediated LTD in Purkinje cell dendrites was recently shown to be spatially restricted to sites where both mGluR and InsP 3 R are located (71). Ca 2ϩ influx through Nmethyl-D-aspartate receptors and voltage-gated channels is considered to be of major importance in synaptic plasticity (65). Synaptic plasticity in several different brain regions requires both Ca 2ϩ entry and mGluR/InsP 3 R (72,76,77), and it been suggested that Ca 2ϩ influx might serve to trigger Ca 2ϩ release by CICR (65,69,78), with InsP 3 Rs therefore playing a critical role in amplifying the Ca 2ϩ influx signal (65). Importantly, [Ca 2ϩ ] i signaling in nonexcitable cells is also associated with both Ca 2ϩ release from stores and Ca 2ϩ influx, and Ca 2ϩ influx has been demonstrated to play a similar role in amplifying and modifying InsP 3 -mediated [Ca 2ϩ ] i signals (79). By demonstrating that the Ca 2ϩ sensitivity of CICR by the InsP 3 R-1 can be regulated, our data raise the possibility that synaptic plasticity and other cellular processes involving InsP 3 Rs may be modulated by physiological stimuli that impinge on the Ca 2ϩ sensitivity of the release channel. The results of the present study suggest that cytoplasmic ATP, and therefore the metabolic status of the cell, may be relevant in this respect. Of note, phospholipase C activity in brain tissue is enhanced in response to ischemia (80), and exposure of hippocampal slices to anoxia (81) or 2-deoxyglucose (13), manipulations expected to alter cytoplasmic ATP concentrations, induce long term potentiation. Directed studies of the role of cell metabolic state and cytoplasmic ATP concentrations in controlling InsP 3 -mediated Ca 2ϩ release will be required to determine the relevance of ATP regulation of InsP 3 R gating for synaptic plasticity as well as other cellular processes.