ATP Regulation of Type-1 Inositol 1,4,5-Trisphosphate Receptor Activity Does Not Require Walker A-type ATP-binding Motifs*

ATP is known to increase the activity of the type-1 inositol 1,4,5-trisphosphate receptor (InsP3R1). This effect is attributed to the binding of ATP to glycine rich Walker A-type motifs present in the regulatory domain of the receptor. Only two such motifs are present in neuronal S2+ splice variant of InsP3R1 and are designated the ATPA and ATPB sites. The ATPA site is unique to InsP3R1, and the ATPB site is conserved among all three InsP3R isoforms. Despite the fact that both the ATPA and ATPB sites are known to bind ATP, the relative contribution of these two sites to the enhancing effects of ATP on InsP3R1 function is not known. We report here a mutational analysis of the ATPA and ATPB sites and conclude neither of these sites is required for ATP modulation of InsP3R1. ATP augmented InsP3-induced Ca2+ release from permeabilized cells expressing wild type and ATP-binding site-deficient InsP3R1. Similarly, ATP increased the single channel open probability of the mutated InsP3R1 to the same extent as wild type. ATP likely exerts its effects on InsP3R1 channel function via a novel and as yet unidentified mechanism.

and Ca 2ϩ , also regulates InsP 3 R as do numerous kinases, phosphatases, and protein-binding partners (7)(8)(9)(10). This intricate network of regulation allows InsP 3 R activity to be finely tuned by the local cytosolic environment (9). As a result, InsP 3 -induced Ca 2ϩ signals can exhibit a wide variety of spatial and temporal patterns, which likely allows Ca 2ϩ to control many diverse cellular processes.
Modulation of InsP 3 -induced Ca 2ϩ release (IICR) by ATP and other nucleotides provides a direct link between intracellular Ca 2ϩ signaling and the metabolic state of the cell. Metabolic fluctuations could, therefore, impact Ca 2ϩ signaling in many cell types given that InsP 3 R are expressed in all cells (11,12). Consistent with this, ATP has been shown to augment IICR in many diverse cell types including primary neurons (13), smooth muscle cells (14), and exocrine acinar cells (15) as well as in immortalized cell lines (16 -18). The effects of ATP on InsP 3 R function do not require hydrolysis because non-hydrolyzable ATP analogues are as effective as ATP (7,14). ATP is thought to bind to distinct regions in the central, coupling domain of the receptors and to facilitate channel opening (2,19). ATP is not required for channel gating, but instead, increases InsP 3 R activity in an allosteric fashion by increasing the open probability of the channel in the presence of activating concentrations of InsP 3 and Ca 2ϩ (7,8,20).
Despite a wealth of knowledge regarding the functional effects of ATP on InsP 3 R function, there is relatively little known about the molecular determinants of these actions. ATP is thought to exert effects on channel function by direct binding to glycine-rich regions containing the consensus sequence GXGXXG that are present in the receptors (2). These sequences were first proposed to be ATP-binding domains due to their similarity with Walker A motifs (21). The neuronal S2 ϩ splice variant of InsP 3 R1 contains two such domains termed ATPA and ATPB. A third site, ATPC, is formed upon removal of the S2 splice site (2,22). The ATPB site is conserved in InsP 3 R2 and InsP 3 R3, while the ATPA and ATPC sites are unique to InsP 3 R1. Our prior work examining the functional consequences of mutating these ATP-binding sites has yielded unexpected results. For example, mutating the ATPB site in InsP 3 R2 completely eliminated the enhancing effects of ATP on this isoform while mutating the analogous site in InsP 3 R3 failed to alter the effects of ATP (23). This indicated the presence of an additional locus for ATP modulation of InsP 3 R3. In addition, mutation of the ATPC in the S2 Ϫ splice variant of InsP 3 R1 did not alter the ability of ATP to modulate Ca 2ϩ release, but instead impaired the ability of protein kinase A to phosphorylate Ser-1755 of this isoform (22).
The ATPA and ATPB sites in InsP 3 R1 were first identified as putative nucleotide-binding domains after the cloning of the full-length receptor (24). Early binding experiments with 8-azido-[␣-32 P]ATP established that ATP cross-linked with receptor purified from rat cerebellum at one site per receptor monomer (19). Later, more detailed, binding experiments on trypsinized recombinant rat InsP 3 R1 showed cross-linking of ATP to two distinct regions of the receptor that corresponded with the ATPA and ATPB sites (17). We and others (16,22,23) have also reported the binding of ATP analogues to purified GST fusions of small regions of InsP 3 R1 surrounding the ATPA and ATPB sites. It is widely accepted, in the context of the sequence similarity to Walker A motifs and biochemical data, that the ATPA and ATPB sites are the loci where ATP exerts its positive functional effects on InsP 3 R1 function (1)(2)(3)16). Furthermore, the higher affinity of the ATPA site to ATP is thought to confer the higher sensitivity of InsP 3 R1 to ATP versus InsP 3 R3, which contains the ATPB site exclusively (25,26). The purpose of this study, therefore, was to examine the contributions of the ATPA and ATPB sites to ATP modulation of the S2 ϩ splice variant of InsP 3 R1. We compared the effects of ATP on InsP 3 R1 and on ATP-binding site mutated InsP 3 R1 using detailed functional analyses in permeabilized cells and in single channel recordings. Here we report that InsP 3 R1 is similar to InsP 3 R3 in that ATP modulates IICR even at maximal InsP 3 concentrations and that neither the ATPA nor the ATPB site is required for this effect.

Site-directed Mutagenesis
An expression construct harboring cDNA for rat InsP 3 R1 was used as templates for mutagenesis. A two-step QuikChange (Stratagene, La Jolla, CA) mutagenesis strategy (27) was used to create the ⌬ATPA and ⌬ATPB mutants. The ATPA mutant was generated by introducing G5318C, G5324C, G5330C, G5333C, and G5339C point mutations and the ⌬ATPBB mutant was made by introducing G6047C, G6053C, and G6062C point mutations into the InsP 3 R1 cDNA. These mutations code for amino acid substitutions: G1773A, G1775A, G1777, G1778A, and G1780A for ⌬ATPA and G2016A, G2018, and G2021 for ⌬ATPB. Correct incorporation of the mutations was confirmed by DNA sequencing (GeneWiz, South Plainfield, NJ).

Creation of Stable InsP 3 R1-expressing DT40-3KO Cell Lines
Wild type and mutated InsP 3 R1 were linearized with NruI and introduced into DT40-3KO cells by nucleofection using solution T and program B23 as per the manufacturer's instructions (Amaxa, Cologne, Germany). The cells were incubated in growth medium for 24 h prior to dilution in selection medium containing 2 mg/ml Geneticin (Invitrogen). Cells were then seeded into 96-well tissue culture plates at ϳ1000 cells/well and incubated in selection medium for at least 7 days. Wells exhibiting growth after the selection period were picked for expansion.

Permeabilized Cell Ca 2؉ Measurements
Measurements of ER luminal Ca 2ϩ were performed essentially as described previously (22,23,28). InsP 3 R1-expressing stable DT40-3KO cells were loaded with 20 M furaptra-AM (Teflabs, Austin, TX) at 39°C for 30 min in a HEPES-buffered physiological saline solution (HEPES-PSS) containing 5.5 mM glucose, 137 mM NaCl, 0.56 mM MgCl 2 , 4.7 mM KCl, 1 mM Na 2 HPO 4 , 10 mM HEPES (pH 7.4), 1.2 mM CaCl 2 , and 1% (w/v) bovine serum albumin. Furaptra-loaded cells were permeabilized by superfusion for 1-2 min with 40 M ␤-escin in intracellular medium (ICM) containing 125 mM KCl, 19 mM NaCl, 10 mM HEPES, 1 mM EGTA (pH 7.3). Permeabilized cells were then washed in ICM without ␤-escin for 15 min to facilitate removal of cytosolic dye, and the cells were then superfused in ICM containing 1.4 mM MgCl 2 , Na 2 ATP, and 0.650 mM CaCl 2 (free [Ca 2ϩ ] of 200 nM, calculated using Maxchelator) to load the intracellular stores. The free [Ca 2ϩ ] was subsequently maintained at a constant 200 nM throughout all experimental maneuvers. The free Ca 2ϩ concentration was verified by fluorescent measurement of free Ca 2ϩ in solutions. Prior to application of InsP 3 , the cells were superfused in ICM without MgCl 2 for 1 min to disable SERCA activity. The unidirectional flux of Ca 2ϩ upon InsP 3 application was then monitored in the same solution containing various concentrations of InsP 3 and ATP by monitoring the emission of the dye above 505 nm following excitation at 340 and 380 nm (exposure for 20 ms, once per second) using a TILL Photonics imaging system. Following removal of InsP 3 , refilling of the stores to allow repeated stimulations was accomplished by superfusion of ICM containing MgCl 2 and ATP. Ca 2ϩ release events were averages of the 30 -50 cells in a field of view. Rates of Ca 2ϩ release were estimated from these average responses by fitting the initial 30-s period of decreasing fluorescence to a single exponential function (GraphPad Prism, San Diego, CA).

Single InsP 3 R1 Channel Measurements in Isolated DT40 Nuclei
Nuclear Preparations-Isolated DT40 nuclei were prepared by homogenization. Homogenization buffer (HB) contained 250 mM sucrose, 150 mM KCl, 3 M ␤-mercaptoethanol, 10 mM Tris, 1 mM phenylmethylsulfonyl fluoride, pH 7.5 with a complete protease inhibitor tablet (Roche, Nutley, NJ). Cells were washed and resuspended in HB prior to nuclear isolation using a RZR 2021 homogenizer (Heidolph Instruments) with 25 strokes at 1200 rpm. 3 l of nuclear suspension were placed in 3 ml of bath solution (BS), which contained 140 mM KCl, 10 mM HEPES, 500 M BAPTA, and 246 nM free Ca 2ϩ , pH 7.1. Nuclei were allowed to adhere to a plastic culture dish for 10 min.
Patch Clamp Experiments-Single InsP 3 R1 currents were measured in the on-nucleus voltage clamp configuration using PClamp 9 and an Axopatch 200B amplifier (MDS Analytical Technologies, Toronto, Canada). Currents were measured at 20 kHz sampling rate and filtered at 5 kHz with a low-pass 4-pole bessel filter. Pipette solution (PS) contained 10 M InsP 3 , 140 mM KCl, 10 mM HEPES, 100 M BAPTA, 200 nM free Ca 2ϩ , and 0 -5 mM ATP, pH 7.1. Measurements were made at Ϫ100 mV following establishment of a Ͼ5 G⍀ seal. Data for currentvoltage relationships were collected by stepping voltage from 0 mV to potentials between Ϫ100 and 100 mV in 10 mV increments. Typically, 1 out of every 5 patches contained reproducible single InsP 3 R1 channel activity. No activity was evident in patches from DT40-3KO nuclei under standard pipette conditions with 5 mM ATP (40 patches, 407 min of total recording time). Additionally, there was no single channel activity in nuclei from cells expressing InsP 3 R1 in the absence of InsP 3 in the recording pipette (37 patches, 412 min of total recording time).
Data Analysis-Analyses were performed using the event detection protocol in Clampfit 9. Channel openings were detected by half-threshold crossing criteria. We assumed that the number of channels in any particular patch is represented by the maximum number of discrete stacked events observed during the experiment. Even at low P o , stacking events were evident (data not shown). Only patches with 1 apparent channel were considered for analysis of open probability, and mean open and closed times. The slope conductances were determined from the linear fits of the current-voltage relationships where g ϭ I k /(V Ϫ V k ). Equation parameters were estimated using a non-linear, least squares algorithm.

ATP Enhances InsP 3 -induced Ca 2ϩ Release in DT40-3KO
Cells Expressing Rat InsP 3 R1-We generated a DT40 triple knock-out (DT40-3KO) cell line stably expressing the rat S2 ϩ InsP 3 R1 splice isoform (DT40-InsP 3 R1) containing only the putative ATPA and ATPB sites, to study the effects of ATP on this isoform expressed in isolation. These cells allow the unambiguous examination of isoform-dependent InsP 3 R function because all three endogenous InsP 3 R isoforms have been deleted via homologous recombination (29). Furthermore, stable cell lines harboring mutated receptors allow the examination of the effects of removing putative sites for modulation (28).
We employed ER calcium measurements in permeabilized DT40-InsP 3 R1 cells to examine the effect of ATP on IICR in the presence of various InsP 3 and ATP concentrations. Similar to other cell lines expressing predominantly InsP 3 R1 (16) and DT40 cells expressing chicken InsP 3 R1 in isolation (30), ATP clearly potentiates IICR in DT40-InsP 3 R1 cells (Fig. 1). Ca 2ϩ release rates were measured over a range of InsP 3 concentrations in the presence or absence of 5 mM ATP. As shown in Fig. 1, ATP was required for the maximal rate of IICR in DT40-InsP 3 R1 cells. This effect of ATP on IICR from InsP 3 R1 is similar to results obtained from DT40 cells expressing rat InsP 3 R3 in isolation, but distinctly different from cells expressing InsP 3 R2 (23). While InsP 3 R1 and InsP 3 R3 are regulated by ATP at maximal InsP 3 concentrations, InsP 3 R2 is only modulated by ATP at submaximal InsP 3 .
Our prior results also established that InsP 3 R2 was more sensitive to ATP than InsP 3 R3 (23). To determine how InsP 3 R1 compared with the other two isoforms in this regard, we next determined the sensitivity of InsP 3 R1 to ATP modulation by stimulating IICR at a set InsP 3 concentration over a range of ATP concentrations. Under these conditions, ATP increased InsP 3 R1 activity with an EC 50 of ϳ150 M (Fig. 2). The sensitivity of InsP 3 R1 to ATP is therefore lower than that of InsP 3 R2, but higher than that of InsP 3 R3.
The ATPA and ATPB Sites Are Not Required for ATP Modulation of InsP 3 R1-The ATPB site is the only ATP-binding site conserved in all three mammalian InsP 3 R isoforms (2). Even though this site is conserved in InsP 3 R2 and InsP 3 R3, the effects of mutating the sites in these isoforms were dramatically different. Abrogation of the ATPB site in InsP 3 R2 eliminated the enhancing effects of ATP, but ATP was able to modulate InsP 3 R3 even after the ATPB site was disrupted (23). We next generated a stable cell line expressing ATPB-mutated InsP 3 R1 (DT40-InsP 3 R1-⌬ATPB) to test whether this site was required for ATP modulation of InsP 3 R1. In our earlier studies, mutation of the central glycine to alanine was sufficient to completely abrogate specific TNP-ATP binding to a GST fragment con- taining the ATPB site (22). In the present studies, all three glycine residues present in the ATPB site (Gly-2016, Gly-2018, Gly-2021) were mutated to alanine residues to ensure that ATP binding would be eliminated unquestionably. Cells expressing this mutated receptor were tested for ATP modulation in the permeabilized DT40 cells under identical experimental conditions as described for the wild type receptor. Similar to experiments with wild type InsP 3 R1, ATP (5 mM) increased the rate of IICR in DT40-InsP 3 R1-⌬ATPB stimulated with 10 M InsP 3 (Fig. 3).
The results presented in Fig. 3 suggest that the ATPB site is not required for ATP modulation of InsP 3 R1. One possible explanation for this observation is that the presence of the ATPA site, still present in the mutated receptor, might mask any possible effects of the ATPB site mutation. This would be expected based on results from bilayer measurements of the opisthotonos mutant InsP 3 R1 (InsP 3 R1-opt) (25). This spontaneously mutated receptor is devoid of a stretch of amino acids (Gly-1732 to Gln-1839), which contains the ATPA (Gly-1773 to Gly-1780) as well as the PKA phosphorylation site (Ser-1755) (31). When this mutated receptor was incorporated into bilayers it exhibited a diminished sensitivity to ATP compared with the wild type receptor (25). This result led the authors to conclude that the ATPA site conferred the high sensitivity of InsP 3 R1 to ATP whereas the ATPB site conferred a lower sensitivity to ATP in the absence of the ATPA site.
We tested whether the ATPA site is indeed determining the functional effects of ATP in the DT40-InsP 3 R1-⌬ATPB cells, using a stable cell line expressing InsP 3 R1 mutated at both the ATPA and ATPB sites (DT40-InsP 3 R1-⌬ATPA/B). Again, mutation of 2 glycine residues to alanine in a GST fragment harboring the ATPA was earlier shown to completely abrogate TNP-ATP binding to the ATPA site (22). In the present study, all glycine residues (5 in total) in the motif were mutated to alanine. This mutant receptor is, therefore, devoid of any putative ATP-binding sites. As shown in Fig. 4, however, ATP was able to increase the rate of IICR from permeabilized DT40-InsP 3 R1-⌬ATPA/B. This effect was evident over a range of InsP 3 concentrations similar to the effects of ATP on IICR from DT40-InsP 3 R1 and DT40-InsP 3 R1-⌬ATPB cells. ATP also increased IICR in DT40-InsP 3 R1-⌬ATPA/B with an EC 50 of 100 M, which was similar to the sensitivity of wild type InsP 3 R1 (EC 50 of 150 M). These results indicate that the known ATPbinding sites in InsP 3 R1 are not required for ATP modulation of this isoform. This is in contrast to the results described above for InsP 3 R1-opt and could reflect the different methodologies to obtain the data (lipid bilayers measurements versus ER Ca 2ϩ measurements). Alternatively, the results may suggest a role for the other residues removed in the InsP 3 R1-opt in conferring ATP sensitivity of InsP 3 R1. It is important to note, however, that the opt mutation had numerous effects on the single channel properties of InsP 3 R1. For example, changes in the single channel conductance, Ca 2ϩ regulation, and the sensitivity to ATP were apparently induced by the opt deletion (25).
ATP Enhances the Single Channel Open Probability of InsP 3 R1-⌬ATPA/B-Mutating the ATPA and ATPB sites in InsP 3 R1 did not alter the ability of ATP to regulate IICR from permeabilized cells expressing this mutant. Macroscopic meas-  urements from pools of permeabilized cells, as described above, may not provide the resolution to detect subtle changes in ATP modulation of InsP 3 R1-⌬ATPB. Because ATP is thought to exert effects on InsP 3 R activity by altering the channel open probability (7,8), we next compared the effects of ATP on the single channel properties of wild type and InsP 3 R1-⌬ATPA/B.
We utilized the nuclear membrane patch-clamp methodology developed by Foskett and co-workers (1,32,33) to analyze the effects of ATP on wild type and ⌬ATPA/B-mutated InsP 3 R1. In this configuration, high resistance seals are made between the pipette and the outer nuclear membrane. Nuclei were isolated from DT40 cell lines stably expressing InsP 3 R1 as described under "Experimental Procedures." InsP 3 R single channels were readily detected in these preparations in the presence of 10 M InsP 3 , ϳ200 nM free Ca 2ϩ and 5 mM ATP. No channel activity was evident if InsP 3 was excluded from the pipette solution. Similarly, there was no detectable activity in nuclei prepared from DT40-3KO cells, indicating that the single channels were, indeed, recorded as a consequence of InsP 3 R activity. Fig. 5 shows representative 200-ms sweeps from patch recordings of nuclei prepared from DT40-InsP 3 R1 and DT40-InsP 3 R1-⌬ATPA/B cells in the presence of 10 M InsP 3 and at various concentrations of ATP. Similar to prior results (7,8), no channel activity was stimulated by ATP in the absence of InsP 3 in 40 separate patches (Ͼ400 min of recording time). This indicates that ATP alone is not sufficient to stimulate channel activity. Channel open probability (P o ) induced by 10 M InsP 3 was, however, increased by more than 5-fold with the inclusion of 5 mM ATP in the pipette solution (Fig. 5). Similar to the effects of ATP on IICR in the permeabilized cell experiments above, 5 mM ATP increased the P o of InsP 3 R1-⌬ATPA/B channels to the same extent as wild type (Fig. 5). Wild type and mutated InsP 3 R1 P o was also increased by ATP in a concentration-dependent manner in both preparations. The major effect of increasing [ATP] on both the WT and ⌬ATPA/B receptor was to decrease the mean closed times (Fig. 5F). Increasing [ATP] had no effect on the mean open times of either receptor (Fig. 5F). Disruption of these sites had no additional observable effects on channel function as evidenced by similar single channel conductances of the two receptors (Fig.  6). These results clearly demonstrate enhancing effects of ATP on InsP 3 R1-⌬ATPA/B single channel function and further strengthen our contention that the Walker A-type ATP-binding motifs in InsP 3 R1 are dispensable for channel modulation by ATP.
Unlike the opt mutant, therefore, the targeted deletion of the ATPA site had no effect on the single channel properties of InsP 3 R1. One explanation for the lack of an effect in the ⌬ATPA/B receptor is that the mutations did not completely eliminate ATP binding. This is not likely given our prior results examining TNP-ATP binding to GST fusion proteins of InsP 3 R1 fragments containing the ATPA or ATB sites (22). Fusion proteins containing wild type ATPA or ATPB sequences readily bound TNP-ATP, while purified fragments harboring mutated Walker-A motifs completely eliminated TNP-ATP binding. Another possibility is that some ATP binding is contained in the region deleted in the opt mutant. Again, this is improbable since the binding experiments described above were performed on purified InsP 3 R1 fragments that spanned the opt deletion region. This indicates that ATP binding to this region of InsP 3 R1 is dependent on the presence of the ATPA site. Despite clear, demonstrable binding, the present data lead us to the conclusion that this binding is unlikely to mediate any functional effect.

DISCUSSION
InsP 3 R are regulated by multiple cytosolic factors including ATP and related nucleotides (34). The combined contributions of InsP 3 , Ca 2ϩ , ATP, and other interacting regulatory mechanisms likely facilitates a multitude of possible activation states of the receptor. InsP 3 R activity and the resulting Ca 2ϩ signals can, therefore, be tightly controlled and tuned to the specific needs of a given cell. A complete understanding of the molecular underpinnings of these regulatory mechanisms will permit insight into the manner in which cells control intracellular Ca 2ϩ signaling via modulation of InsP 3 R. This study, combined with the results from our prior work (22,23), allows the first direct comparison of the effects of ATP on mammalian wild type versus ATP-binding site-mutated InsP 3 R family members expressed in isolation. The three isoforms exhibit some similarities in their respective responses to ATP, but there are also striking differences with regard to the manner in which ATP modulates InsP 3 R function. For example, while Ca 2ϩ release activity is increased by ATP in all three isoforms, the major defining feature is that ATP regulates IICR from InsP 3 R1 and InsP 3 R3 at all InsP 3 concentrations, but augments IICR from InsP 3 R2 exclusively at submaximal [InsP 3 ].
Another major isoform-dependent difference is the range of ATP sensitivities displayed by the three isoforms. InsP 3 R2 exhibits the highest sensitivity with an EC 50 for ATP of ϳ40 M (23), followed by InsP 3 R1 (EC 50 ϳ 150 M, Fig. 2), and InsP 3 R3 (EC 50 ϳ 400 M (23)). The differences in the relative sensitivities of the three InsP 3 R isoforms to ATP would be important in cell types that express a significant majority of one isoform over the others. In fact, our prior work has demonstrated that the high sensitivity of InsP 3 R2 to ATP dominates in cell types that predominantly express this isoform including isolated mouse pancreatic acinar cells and cultured AR42J cells (15). Cells that express mostly InsP 3 R3, such as pancreatic acinar cells from InsP 3 R2 knock-out mice and RinM5F cells, exhibit a 10-fold higher EC 50 for ATP, indicating sensitivities more in line with that of InsP 3 R3 (15).
All mammalian cells express some amount of InsP 3 R1 (11). Neurons are unique in expressing this isoform in a large majority over InsP 3 R2 and InsP 3 R3 with InsP 3 R1 being better than 95% of total neuronal InsP 3 R (11). There are a range of reported ATP sensitivities for InsP 3 R1. Purified cerebellar InsP 3 R, which is almost exclusively InsP 3 R1, is reported to display a high sensitivity to ATP (half-maximal value of 40 M) (7). More recent experiments under similar conditions with recombinant InsP 3 R1 expressed in Sf9 cell nuclei displayed sensitivity to ATP more in line with the present data (half-maximal value of 150 M in (26) and a half-maximal value of 240 M in Ref. 25). Our experimental design is similar to these bilayers studies in that ATP sensitivity was determined at set InsP 3 and Ca 2ϩ concentrations. An additional report established a dissociation coefficient of 270 M for ATP activation of Xenopus InsP 3 R1 using nuclear patch clamp recordings (8). The sensitivity to ATP in this study was established by analyzing the effects of ATP on the Ca 2ϩ -dependent activation of InsP 3 R single channel activity.
The reported sensitivities of InsP 3 R1 to ATP are well below the presumed total ATP concentration in healthy neurons, which ranges from 3-10 mM (35). The majority of cytosolic ATP is expected to be in complex with Mg 2ϩ . As has been noted previously, however, the reported sensitivities of InsP 3 R1 to ATP are more in line with the predicted levels of free ATP in cells (8). This distinction is important considering that, while MgATP was as effective as free ATP in enhancing InsP 3 R activity in some studies (7,16), it was ineffective in others (8,20). Even if the sensitivity of InsP 3 R1 allows it to be maximally stimulated by physiological concentrations of ATP, cellular ATP levels have been shown to fall below 100 M in cerebral cortex tissue in response to acute hypoxic treatments (36). InsP 3 R1 activity would therefore be expected to be impaired under these conditions.
The major finding presented here, is that the known ATPbinding domains present in InsP 3 R1 are not the major determinants of ATP modulation of this isoform. This result is in contrast to the case of InsP 3 R2 in which the ATPB site is required for ATP modulation (23). The identity of the functionally relevant ATP-binding domains in InsP 3 R1 and InsP 3 R3, therefore, remain to be determined. The most complete analysis of ATP binding to InsP 3 R1 demonstrated that 8-azido-[␣ 32 -P]ATP cross-linked with InsP 3 R1 in at least two distinct regions of trypsin-digested InsP 3 R1 (17). Similarly, the same study reported that ATP cross-linked with at least one region of InsP 3 R3. The identified regions corresponded to the areas containing the ATPA and ATPB sites in InsP 3 R1 and the ATPB site in InsP 3 R3 leading the authors to conclude that these were the only ATP-binding sites present in the receptors (17). The results were not, however, confirmed by performing the ATP cross-linking assays on mutated receptors. These types of experiments will need to be attempted to determine if additional ATP-binding sites are present on ATPA-and ATPB-deficient InsP 3 R1 and InsP 3 R3. Of note, recent work has demonstrated that incubation of InsP 3 R1 with Ca 2ϩ to promote receptor activation causes large pockets of the receptor to become exposed (37). These structural changes in response to activation may lead to the exposure of additional ATP-binding sites. Cross-linking 8-azido-[␣ 32 -P]ATP to InsP 3 R1 and InsP 3 R3 under these activating conditions may help uncover additional ATP-binding sites not apparent in the original binding studies. Because the ATPB mutated InsP 3 R2 is insensitive to ATP modulation, a functional comparison of chimerical InsP 3 R may also yield insights into the regions required for ATP modulation. Regions of InsP 3 R1 could be transposed onto the ATPB-mutated InsP 3 R2 backbone to identify areas of the receptor that rescue ATP sensitivity. These approaches may also help determine if the functionally relevant ATP-binding domain is present in the InsP 3 R1 sequence or contained in one of the many protein-binding partners of the receptor.
The ultimate identification of the functionally important ATP-binding site(s) in InsP 3 R1 would likely not be a trivial undertaking, but could yield insight into novel mechanisms of ATP modulation of ion channels besides InsP 3 R. Most notably, members of the related ryanodine receptor (RyR) family are known to be modulated by ATP (38). Similar to the InsP 3 R, ATP analogues can cross-link with distinct regions of RyR1 (39). Furthermore, electron-spin resonance (ESR) spectroscopy on skeletal muscle RyR also demonstrated ATP binding (40). In this study, the authors determined that a maximum of 8 spin-labeled ATP analogues bound to the tetrameric receptor, thus predicting two ATP-binding sites per RyR monomer. Sequence analysis of the RyR predicts at least 8 putative ATP-binding domains (41). If there are indeed two ATP-binding sites per monomer as suggested by the ESR studies, then a subset of these motifs or alternatively, equally possible, none of the motifs at all may confer ATP sensitivity. Similar to our current study, mutations in putative ATP-binding sites in RyR1 did not alter perturb the effects of ATP on channel function (41). Given the functional and structural homology of InsP 3 R with RyR, the identification of novel ATP-binding domains in InsP 3 R may shed light on the molecular determinants of ATP regulation of RyR.
In summary, this study represents the completion of mutagenic analyses of the known ATP-binding sites in the InsP 3 R family. To date, only two of the known ATP-binding sites have been assigned a function. ATP binding to the ATPB site of InsP 3 R2 confers ATP sensitivity to this isoform (23) and ATP binding to the ATPC site of the S2 Ϫ isoform of InsP 3 R1 facilitates PKA phosphorylation of Ser-1755 of this splice variant (22). The mechanistic details and functional consequences of ATP modulation of InsP 3 R are isoform-dependent. Any role of ATP in regulating InsP 3 -induced Ca 2ϩ signaling in cells will, therefore, be dependent on which isoforms are expressed. The emphasis of further work will be to identify the ATP-binding sites responsible for ATP modulation of InsP 3 R1 and InsP 3 R3.