Calpain-cleaved Type 1 Inositol 1,4,5-Trisphosphate Receptor (InsP3R1) Has InsP3-independent Gating and Disrupts Intracellular Ca2+ Homeostasis*

The type 1 inositol 1,4,5-trisphosphate receptor (InsP3R1) is a ubiquitous intracellular Ca2+ release channel that is vital to intracellular Ca2+ signaling. InsP3R1 is a proteolytic target of calpain, which cleaves the channel to form a 95-kDa carboxyl-terminal fragment that includes the transmembrane domains, which contain the ion pore. However, the functional consequences of calpain proteolysis on channel behavior and Ca2+ homeostasis are unknown. In the present study we have identified a unique calpain cleavage site in InsP3R1 and utilized a recombinant truncated form of the channel (capn-InsP3R1) corresponding to the stable, carboxyl-terminal fragment to examine the functional consequences of channel proteolysis. Single-channel recordings of capn-InsP3R1 revealed InsP3-independent gating and high open probability (Po) under optimal cytoplasmic Ca2+ concentration ([Ca2+]i) conditions. However, some [Ca2+]i regulation of the cleaved channel remained, with a lower Po in suboptimal and inhibitory [Ca2+]i. Expression of capn-InsP3R1 in N2a cells reduced the Ca2+ content of ionomycin-releasable intracellular stores and decreased endoplasmic reticulum Ca2+ loading compared with control cells expressing full-length InsP3R1. Using a cleavage-specific antibody, we identified calpain-cleaved InsP3R1 in selectively vulnerable cerebellar Purkinje neurons after in vivo cardiac arrest. These findings indicate that calpain proteolysis of InsP3R1 generates a dysregulated channel that disrupts cellular Ca2+ homeostasis. Furthermore, our results demonstrate that calpain cleaves InsP3R1 in a clinically relevant injury model, suggesting that Ca2+ leak through the proteolyzed channel may act as a feed-forward mechanism to enhance cell death.

from intracellular stores (20 -22). InsP 3 R1 is also a substrate for calpain, which sequentially cleaves the protein into 200-, 130-, and 95-kDa carboxyl-terminal fragments (23)(24)(25). However, the specific calpain cleavage site that generates the stable 95-kDa fragment has not been established, and the functional consequences of proteolysis on channel activity are unknown. We hypothesized that calpain-mediated proteolysis, which removes the amino-terminal ligand binding domain and a large portion of the regulatory domain, generates a channel with InsP 3 -independent gating that leaks Ca 2ϩ from ER stores.
In the present study we have identified the unique calpain cleavage site of InsP 3 R1 and investigated the electrophysiological and functional properties of the truncated carboxyl-terminal channel. Using a combination of single-channel nuclear patch clamp electrophysiology and single-cell Ca 2ϩ imaging, we have determined that the cleaved channel is constitutively active in the absence of InsP 3 , although its gating retains sensitivity to [Ca 2ϩ ] i . Constitutive channel activity accounts for an observed reduction in the Ca 2ϩ content of the ER lumen in cells expressing recombinant calpain-cleaved InsP 3 R1. Importantly, we provide evidence of calpain-mediated proteolysis of InsP 3 R1 in an in vivo model of ischemic brain injury. These data highlight the important functional consequences of calpain-mediated channel proteolysis, which may critically disrupt intracellular Ca 2ϩ homeostasis, particularly under pathologic conditions where other Ca 2ϩ regulatory proteins are also compromised.

EXPERIMENTAL PROCEDURES
Materials-Unless otherwise noted, all chemical reagents were purchased from Sigma.
Antibodies-The anti-InsP 3 R1 polyclonal antibody targeted against the 20 carboxyl-terminal residues of rat InsP 3 R1 was generously provided by Dr. Suresh Joseph (Thomas Jefferson University) (26). The antibody to calpain-cleaved spectrin (Ab38) was generously provided by Dr. Robert Siman (University of Pennsylvania). Anti-calreticulin polyclonal antibody was purchased from Thermo Scientific Pierce Antibodies (PA3-900). Alexa-488 conjugated secondary antibody used for immunofluorescence was purchased from Invitrogen. The antibody against calpain-cleaved InsP 3 R1 (Ab2054) was generated against the peptide ASAATRKAC and is described under "Results." Cerebellar Microsome Isolation-Male Long-Evans rats were deeply anesthetized with pentobarbital (200 mg/kg) and decapitated. The brain was extracted, and the cerebellum was dissected and homogenized in cold MSHE buffer: 220 mM mannitol, 70 mM sucrose, 2 mM K-HEPES (pH 7.4), 0.5 mM EGTA with 0.1% fatty-acid free bovine serum albumin (BSA), and protease inhibitor mixture. The homogenate was centrifuged at 650 ϫ g for 10 min at 4°C to remove nuclei (P1). The supernatant from the preceding fraction (S1) was centrifuged again at 8000 ϫ g for 10 min at 4°C to separate cytoplasm and microsomes from mitochondria and synaptosomes. The remaining supernatant (S2) was centrifuged at 100,000 ϫ g for 1 h in a micro-ultracentrifuge (ThermoFisher Scientific). The pellet from the final spin (P3) containing microsomes was resuspended in either MSHE buffer without BSA for Western blot or digest buffer for subsequent in vitro digest, snap-frozen, and stored at Ϫ20°C. InsP 3 R1 Peptide Generation-The amino acid sequences of rat InsP 3 R1 from residues 1582-1932 or from residues 1932-2257 were fused to an amino-terminal glutathione S-transferase (GST) using the linker sequence LEVLFQGP and cloned into pGEX expression vectors. The fusion protein was grown in BL21 Escherichia coli and purified using glutathione-Sepharose 4B (GE Healthcare). The GST tag was removed from the InsP 3 R1 1582-1932-amino acid peptide after purification using PreScission Protease, which recognizes the sequence LEV-LFQ2GP and cleaves between the Gln and Gly residues (GE Healthcare).
In Vitro Caspase-3 and Calpain-1 Digests of Microsomes and InsP 3 R1 Peptide-For in vitro caspase-3 digest of cerebellar microsomes, freshly isolated microsomes were resuspended in lysis buffer (50 mM Tris-HCl (pH 7.4), 5 mM EDTA, 1 mM DTT (Fisher), with 1% Triton X-100) and incubated at 4°C for 90 min before snap-freezing. For subsequent digest, lysed microsomes were diluted to 1 g/l in digest buffer (100 mM Tris-HCl (pH 7.4), 2 mM EDTA, and 20 mM DTT). Samples were preincubated with 1 M calpastatin peptide (EMD Biosciences) to inhibit endogenous calpain activity and then digested with recombinant active human caspase-3 (0.05 units/l; Millipore) for 4 h at 37°C. Loading buffer was added to the digest to stop the reaction.
For in vitro calpain digest of cerebellar microsomes, freshly isolated microsomes were resuspended in a modified digest buffer (25 mM HEPES (pH 7.4), 250 mM sucrose, 1 mM EDTA) and snap-frozen. For subsequent digest, microsomes were diluted to 1 g/l in complete digest buffer (25 mM HEPES (pH 7.4), 250 mM sucrose, 1 mM EDTA, 2 mM DTT, and 1 mM CaCl 2 ). Samples were preincubated with 1 mM benzyloxycarbonyl-VAD-fluoromethyl ketone (Promega) to inhibit endogenous caspase activity and then digested with 0.5 M calpain-1 (-calpain) from human erythrocytes (EMD Biosciences) at 4°C for various times. Loading buffer was added to the digest to stop the reaction.
For in vitro calpain digest of the GST-InsP 3 R1 fusion peptides, purified peptide was diluted in digest buffer (25 mM HEPES (pH 7.5), 250 mM NaCl, 2 mM DTT, 1 mM CaCl 2 with 1.2% CHAPS). Samples were digested with 0.1 M calpain-1 at 4°C for various times. Adding loading buffer to the digest terminated the reaction.
Edman Amino-terminal Sequencing to Determine Cleavage Site-Amino-terminal automated Edman sequencing was performed on an Applied Biosystems 494 Protein Sequencer using standard programs and reagents (Wistar Institute, Proteomics Core Facility). Samples were excised from stained PVDF membranes, wetted with MeOH, and sonicated in MilliQ water for 5 min. Samples were then removed and placed in MeOH before loading into the upper sample cartridge unit. The cartridge was assembled, and the sample was dried with argon, inserted into the instrument, and processed with a standard pulsed liquid method for 8 cycles.
Plasmids-The pIRES2-EGFP (Clontech) expression vector for wild-type (wt) rat InsP 3 R1 was previously generated (27). To construct the expression vector for calpain-cleaved InsP 3 R1 (capn-InsP 3 R1; Fig. 3A), we amplified the 3Ј-portion of rat InsP 3 R1 cDNA sequence corresponding to the stable carboxyl-terminal fragment of the channel generated by calpain proteolysis (2.6 kb) using primers 5Ј-GCGAATTATGGCAT-CTGCTGCCACCAG-3Ј and 5Ј-GTCTCTAGAAAATGTAC-TTAAGCGCACAT-3Ј (underlined regions indicate EcoRI and XbaI restriction sites). PCR was performed using Pfu Turbo DNA polymerase (Stratagene) and the capn-InsP 3 R1 PCR product was subcloned into the pIRES2-EGFP expression vector using EcoRI and XbaI restriction sites. EGFP was expressed with InsP 3 R1 constructs from a single bicistronic mRNA using an internal ribosome entry site in the pIRES2-EGFP expression vector.
Cell Culture, DNA Transfection, and Lysate Preparation-Neuro-2A (N2a) cells were cultured in Minimum Essential Medium (Invitrogen) supplemented with 10% fetal bovine serum and maintained at 37°C in a humidified incubator with 5% CO 2 . N2a cells were grown on polystyrene multiwell plates (Western blot), polystyrene tissue culture flasks (electrophysiology), or 20 ϫ 50-mm glass coverslips (single-cell Ca 2ϩ imaging) and transfected using Lipofectamine 2000 (Invitrogen). For Western blot, cells were harvested by trypsinization and resuspended in homogenization buffer (50 mM Tris, 150 mM NaCl, 2 mM EGTA) and sonicated. Lysates were treated with SDS loading buffer, boiled, and analyzed on SDS-PAGE gels. Where indicated, capn-InsP 3 R1-transfected cells were solubilized and denatured in buffer supplemented with 5 M urea or 1% SDS.
Single-channel Electrophysiology-Nuclei from N2a cells were isolated 24 h post-transfection as previously described (28 -31). Briefly, nuclei were mechanically isolated and plated onto a glass-bottomed dish with standard bath solution (140 mM KCl, 10 mM HEPES, 0. Where indicated, 10 M InsP 3 with or without 100 g/ml heparin were included in the pipette solution. Data were acquired at room temperature using an Axopatch 200A amplifier (Axon Instruments) and analyzed as previously described (28).
Single-cell Ca 2ϩ Imaging-To measure [Ca 2ϩ ] i , N2a cells were plated onto glass coverslips (Warner Instruments) and transfected. Six hours post-transfection, cells on coverslips were secured in a perfusion chamber and mounted on the stage of an inverted microscope (Nikon Eclipse TE2000). Cells were loaded with Fura-2-AM (Molecular Probes; 2.5 M) for 30 min at room temperature in Ca 2ϩ -containing extracellular solution (137 mM NaCl, 2 mM KCl, 2 mM CaCl 2 , 10 mM HEPES (pH 7.3)) supplemented with 1% BSA. Fura-2 was alternately illuminated at 340/380 nm, and fluorescence intensity was filtered at 510 nm. Data were collected and recorded as described previously (29,31). Cells were perfused with 2 mM Ca 2ϩ extracellular solution to establish base-line [Ca 2ϩ ] i before ionomycin (Invitrogen; 2 M) was applied in 0-Ca 2ϩ extracellular solution (137 mM NaCl, 2 mM KCl, 0.5 mM EDTA, 0.5 mM EGTA, 10 mM HEPES (pH 7.3)) to measure Ca 2ϩ release from intracellular stores. At the end of the experiment, Mn 2ϩ was used to quench Fura-2 fluorescence (2 mM Ca 2ϩ extracellular solution supplemented with 10 mM MnCl 2 and 10 M ionomycin). The remaining background fluorescence after Mn 2ϩ quench was subtracted during analysis.
To measure ER lumenal Ca 2ϩ concentration ([Ca 2ϩ ] ER ), transfected cells (6 h post-transfection) were loaded with the low affinity Ca 2ϩ indicator Mag-Fura-2-AM (Invitrogen; 10 M) for 30 min at 37°C in HEPES-HBSS buffer (Hanks' balanced salt solution supplemented with 10 mM HEPES, 4.2 mM NaHCO 3 , 1.8 mM CaCl 2 , and 0.8 mM MgCl 2 (pH 7.3)) with 1% BSA. After loading, cells were permeabilized with 10 g/ml digitonin for 2 min in MgATP-free cytoplasm-like medium (140 mM KCl, 20 mM NaCl, 1 mM EGTA, 0.375 mM CaCl 2 (final concentration Ϸ 70 nM), 20 mM PIPES (pH 7.3)). Cells were perfused with cytoplasm-like medium for 30 min to wash out digitonin and deplete ER Ca 2ϩ stores. ER store loading was activated by the addition of 1.5 mM MgATP to the perfusate to stimulate SERCA-mediated Ca 2ϩ uptake. After filling, the passive ER Ca 2ϩ leak was evaluated by measuring [Ca 2ϩ ] ER during exposure to 10 M cyclopiazonic acid (Calbiochem) in the absence of MgATP. Mag-Fura-2 excitation and emission were monitored as described above. Rates for ER loading and release were calculated by fitting individual single-cell responses using single exponential functions and determining the mean of those rates.
Changes in [Ca 2ϩ ] i and [Ca 2ϩ ] ER are presented as changes in fluorescence ratio. Dye calibration was achieved by applying experimentally determined constants to the equation Macros used for analysis were custom macros written for IGOR Pro (WaveMetrics).
Cardiac Arrest Model-Male Long-Evans rats weighing 300 -350 g (Harlan Laboratories, Inc.) were subjected to asphyxial cardiac arrest followed by cardiopulmonary resuscitation and post-cardiac arrest temperature regulation as previously described (32)(33)(34). Rats were anesthetized, orotracheally intubated, and mechanically ventilated. Temperature was maintained between 37.0 and 37.5°C. To initiate cardiac arrest, rats were chemically paralyzed using intravenous vecuronium (2 mg/kg), and asphyxia was induced by discontinuing mechanical ventilation. Cessation of arterial pulse pressure and reduction of mean arterial pressure to Յ20 mm Hg was used to confirm circulatory arrest, which usually occurred within 3-4 min. After the 7-min asphyxia, mechanical ventilation resumed with 100% O 2 , intravenous epinephrine (0.005 mg/kg) and HCO 3 Ϫ (1.0 mEq/kg) were administered, and external chest compressions were performed (350 -400 compressions/min). After the return of spontaneous circulation, rats were maintained on mechanical ventilation for 1 h. An intraperitoneal telemetric temperature probe was surgically implanted in the abdomen (Data Systems International). Rats were then extubated and transferred to a computer controlled temperature regulation chamber, which telemetrically monitored intraperitoneal temperature. Post-cardiac arrest body temperature was maintained between 36.5 and 37.5°C using software-driven relays connected to a heat lamp, water misters, and a cooling fan (32,35). This study was approved by the University of Pennsylvania Institutional Animal Care and Use Committee.
Immunohistochemical Staining of Cerebellum-At the indicated time points after cardiac arrest (24 or 48 h), rats were anesthetized with pentobarbital (200 mg/kg) and transcardially perfused with cold phosphate-buffered saline (PBS (pH 7.4)) followed by 4% paraformaldehyde in 0.1 M phosphate buffer. Brains were extracted, post-fixed in 4% paraformaldehyde overnight, and cryoprotected using serial incubations in 0.1 M phosphate buffer containing 10, 20, and 30% sucrose. Serial sagittal sections through the cerebellum (40 m) were cut on a freezing sliding microtome and stored in cryoprotectant (0.1 M phosphate buffer with 30% sucrose and 30% glycerol) at Ϫ20°C for future use.
Immunohistochemistry to identify calpain-cleaved InsP 3 R1 (Ab2054) required heat-induced epitope retrieval. Cerebellar tissue sections were removed from cryoprotectant, rinsed in distilled H 2 O, and incubated in citrate buffer (10 mM sodium citrate with 0.05% Tween 20 (pH 6.0)) at 95°C for 20 min. After heat-induced epitope retrieval, tissue was removed from heat and rinsed in PBS (pH 7.2). Standard immunohistochemical staining procedures followed.
Floating cerebellar brain sections were rinsed in PBS (pH 7.2) and blocked using PBS with 4% normal goat serum and 0.5% Triton X-100. Sections were incubated in primary antibody (Ab38 or Ab2054) diluted in block solution overnight at 4°C. Sections were rinsed with PBS and incubated with Alexa 488 fluorescent secondary antibody (Invitrogen). Finally, sections were rinsed with PBS, counterstained with Hoechst (5 g/ml in PBS) to label nuclei, mounted on slides, and cover-slipped under Fluormount-G (Electron Microscopy Sciences).
For Fluoro-Jade labeling, paraformaldehyde-fixed sections were mounted onto gelatin-coated slides and dried. Tissue was treated with 100% ethanol, 70% ethanol, and distilled H 2 O rinses. Hydrated tissue was incubated with 0.06% potassium permanganate for 10 min, rinsed with distilled H 2 O, and stained with 0.0004% Fluoro-Jade B (Millipore) in 0.1% acetic acid for 20 min. Stained tissue was rinsed with distilled H 2 O, dried at 37°C, dehydrated with xylene. and cover-slipped using Permount (Fisher). Immunofluorescence was studied at 100ϫ using an epifluorescence microscope (Leica DM4500B).
Analyses and Statistics-Data are presented as the mean Ϯ S.E., and statistical significance of differences between the means was assessed using either unpaired t tests or analysis of variance for repeated measures using Barlett's test for equal variances and a Bonferroni correction. Differences between means were accepted as statistically significant at the 95% level (p Ͻ 0.05).

RESULTS
Identifying the Calpain Cleavage Site of InsP 3 R1-InsP 3 R1 is a known substrate of calpain that sequentially cleaves the protein into 200-, 130-, and 95-kDa carboxyl-terminal fragments, with the 95-kDa fragment being the stable, predominant cleavage product (Fig. 1B) (23)(24)(25). Although the 95-kDa calpainderived fragment is similar in size to the caspase-derived fragment, the cleavage sites are different. To determine the calpain cleavage site of InsP 3 R1, we first examined the relative sizes of the caspase-and calpain-cleaved carboxyl-terminal fragments of InsP 3 R1 using in vitro digests of rat cerebellar tissue, which is enriched in InsP 3 R1. We performed in vitro digests of cerebellar microsomes with either recombinant, constitutively active caspase-3, or exogenous calpain-1. A Western blot of the digest products using a carboxyl-terminal InsP 3 R1 antibody demonstrated loss of the full-length protein and generation of stable carboxyl-terminal fragments of ϳ95 kDa (Fig. 1C). Shown here for the first time, calpain cleavage of InsP 3 R1 generates a slightly smaller fragment than the fragment generated by caspase cleavage.
Caspase cleavage of InsP 3 R1 is known to occur at the carboxylterminal end of the 1888 DEVD consensus sequence (Fig. 1A) (17). Based on the relative sizes of the caspase and calpain InsP 3 R1 cleavage products, we deduced that the calpain cleavage site responsible for generating the 95-kDa stable fragment was less than 30 residues downstream of the caspase-3 cleavage site. To identify the site we utilized a GST fusion peptide containing residues 1583-1932 of rat InsP 3 R1, which includes the caspase-3 cleavage site (Fig. 1A). As a control, we used a GST fusion peptide containing an adjacent sequence within InsP 3 R1 (residues 1932-2257) that does not include the caspase-3 cleavage site (supplemental Fig. 1A). We digested both GST fusion peptides with exogenous calpain-1, separated reaction products using SDS-PAGE, and transferred them to PVDF membranes for protein staining. Protein staining of digest products identified a single proteolytic fragment derived from the 1582-1932-amino acid peptide but not the 1932-2257 amino acid peptide, confirming that the calpain cleavage site is between residues 1582 and 1932 (supplemental Fig. 1B). The addition of the calpain inhibitor calpastatin blocked generation of the proteolytic fragment produced by calpain digest of InsP 3 R1 peptide 1582-1932 amino acid (Fig. 1D). Eight rounds of Edman degradation and amino-terminal sequencing of the proteolytic fragment from the calpain-digested GST fusion peptide returned four potential sequences, only one of which ( 1918 ASAATRKA) was downstream of the caspase-DEVD site (Fig. 1E). Based on the location of this sequence in InsP 3 R1, the predominant calpain cleavage product is expected to have a molecular mass of 94.84 kDa. This fragment is 3.02 kDa smaller than the caspase-cleaved fragment, which is consistent with Western blot comparisons of calpain and caspase in vitro digests of microsomal InsP 3 R1 (Fig. 1C). These data demonstrate that calpain cleaves rat InsP 3 R1 between residues 1917 and 1918 to generate the 95-kDa carboxyl-terminal fragment. The sequences surrounding the cleavage site are highly conserved in InsP 3 R1 across species, although it is not present in type 2 or type 3 receptors.
Calpain-cleaved InsP 3 R1 Antibody-To facilitate investigations of calpain-cleaved InsP 3 R1, we generated an antibody specific to the calpain-derived carboxyl-terminal fragment of the channel using a neoepitope specific approach (36). The peptide ASAATRKAC, corresponding to the new amino terminus of InsP 3 R1 after calpain cleavage, was synthesized and covalently linked to a keyhole limpet hemocyanin to elicit an immune response ( Fig. 2A; Covance ImmunoTechnologies). A New Zealand White rabbit (2054) was immunized by subcutaneous injections of the peptide, and serum ELISA demonstrated a 1:843,000 titer at the time of test-bleed 1.
We tested specificity of Ab2054 for the 95-kDa calpain-derived InsP 3 R1 fragment by Western blot of rat cerebellar microsomes after in vitro digestion. Using a carboxyl-terminal InsP 3 R1 antibody, we detected full-length, caspase-digested and calpain-digested InsP 3 R1 samples (Fig. 2B). Using Ab2054 on the same blot, we detected no signal in undigested microsomes or microsomes digested with caspase-3, but a single band at ϳ95 kDa was observed in microsomes digested with calpain-1, indicating the antibody reacts exclusively with the stable 95-kDa calpain-cleaved form of InsP 3 R1 (Fig. 2C). These data validate Ab2054 as specific antibody for calpain-cleaved InsP 3 R1.
Expression of Recombinant InsP 3 R1 Constructs-To examine the consequences of calpain cleavage of InsP 3 R1 on channel function, we generated a truncated cDNA of the protein (⌬1-1917) corresponding to the stable carboxyl-terminal fragment derived from calpain cleavage (capn-InsP 3 R1). We used fulllength wt rat InsP 3 R1 as a control (Fig. 3A).
For all functional experiments, we studied wt-and capn-InsP 3 R1 expressed transiently in N2a mouse neuroblastoma cells. We chose N2a cells because they are neural in tissue origin and have high transfection efficiency. To examine the time course of expression of recombinant rat InsP 3 R1 in N2a cells, we transiently transfected cells with either wt-InsP 3 R1 or capn-InsP 3 R1 and harvested cells 6, 16, 24, or 48 h later. Western blotting of whole-cell lysates using the carboxyl-terminal InsP 3 R1 antibody demonstrated expression of both wt-and capn-InsP 3 R1 by 6 h post-transfection. For capn-InsP 3 R1, expression was maximal at 24 and decreased by 48 h post-transfection, whereas expression of the wt channel continued to increase up to 48 h, suggesting toxicity or cellular regulation of gene expression associated with capn-InsP 3 R1 (Fig. 3B). In lysates from wt-InsP 3 R1 cells, we occasionally observed lower molecular weight fragments that are possibly proteolytic degradation products. In lysates from cells expressing capn-InsP 3 R1, we observed high molecular weight smears that are likely aggregates of capn-InsP 3 R1. Similar high molecular weight smears have been observed when expressing the caspase-3 cleaved form of InsP 3 R1 (21). These smears were eliminated by harvesting capn-InsP 3 R1-transfected cells in lysis buffer containing either 5 M urea or 1% SDS (supplemental Fig. 2).
InsP 3 -independent Gating of Capn-InsP 3 R1-To determine whether capn-InsP 3 R1 still functioned as an ion channel, we examined single-channel activity in native ER membranes. We utilized patch clamp electrophysiology of the outer membrane of isolated N2a cell nuclei to study the channel activities of recombinant InsP 3 R1 constructs (15). To maximize the likelihood of recording recombinant channels, we performed patch clamp experiments at 24 h post-transfection, when expression of capn-InsP 3 R1 was maximal. We monitored channel activity with the pipette solution (cytoplasmic side) containing 2 M Ca 2ϩ and 0.5 mM ATP. In untransfected cells, we observed endogenous InsP 3 R with a single-channel conductance of 395 Ϯ 12 picosiemens with saturating 10 M InsP 3 in the pipette solution (supplemental Fig. 3, A and B). In nuclei from cells transfected with recombinant rat wt-InsP 3 R1, we observed channel activities that had a smaller single-channel conduc-tance (231 Ϯ 10 picosiemens; supplemental Fig. 3C). Despite their smaller single-channel conductance, we identified these channels as InsP 3 R based on their dependence on InsP 3 for activation (n ϭ 10; Fig. 4A) and their sensitivity to competitive inhibition by heparin (100 g/ml in pipette solution; n ϭ 5; Fig.  4, A, C, and D). With 10 M InsP 3 in the pipette solution, the recombinant wt channels had an open probability (P o ) of 0.69 Ϯ 0.08 (n ϭ 8; Fig. 4, A, C, and D). In cells transfected with capn-InsP 3 R1, we observed channel activities with a single-channel conductance (230 Ϯ 10 picosiemens) similar to the recombinant wt-InsP 3 R1 (supplemental Fig. 3D). Strikingly, we observed these channels even in the absence of InsP 3 in the pipette solution. Exposed to pipette solution containing no InsP 3 , 2 M Ca 2ϩ , and 0.5 mM ATP, the spontaneously active capn-InsP 3 R1 channels had high P o in either the absence (P o ϭ 0.82 Ϯ 0.04, n ϭ 9) or presence (P o ϭ 0.74 Ϯ 0.09, n ϭ 5) of heparin (Fig. 4, B-D). These results indicate that a truncated InsP 3 R1 corresponding to the carboxyl-terminal calpain cleavage fragment forms a functional ion channel with InsP 3 -independent gating.
Ca 2ϩ Regulation of Capn-InsP 3 R1 Gating-Channel activity of InsP 3 R1 is both activated and inhibited by Ca 2ϩ , resulting in a biphasic, bell-shaped P o dependence on [Ca 2ϩ ] i (15). Although InsP 3 regulation of the truncated channel was absent, we asked whether any Ca 2ϩ regulation of capn-InsP 3 R1 remained. Accordingly, we next examined channel activity at low Ca 2ϩ concentration (70 nM), typical of resting cytoplasmic levels. With 10 M InsP 3 and 70 nM free Ca 2ϩ in the pipette solution, P o of wt-InsP 3 R1 was 0.016 Ϯ 0.004 (n ϭ 8; Fig. 5, A and C), significantly lower than the observed P o in 2 M Ca 2ϩ (P o ϳ 0.7; p Ͻ 0.001). With no InsP 3 and 70 nM free Ca 2ϩ included in the pipette solution, P o of capn-InsP 3 R1 was 0.26 Ϯ 0.06 (n ϭ 5; Fig. 5, B and C), also significantly lower than the observed P o in 2 M Ca 2ϩ (P o ϳ 0.8; p Ͻ 0.001) but significantly higher than that of wt-InsP 3 R1 at low Ca 2ϩ (p Ͻ 0.001). These  data suggest that capn-InsP 3 R1 may constitute an ER Ca 2ϩ leak permeability in resting cells. To determine whether high Ca 2ϩ still inhibits capn-InsP 3 R1, we examined channel activity at 25 M Ca 2ϩ , which corresponds to pathologic neuronal cytoplasmic levels immediately after ischemia (37,38). With 10 M InsP 3 and 25 M free Ca 2ϩ in the pipette solution, P o of wt-InsP 3 R1 was 0.22 Ϯ 0.06 (n ϭ 8; Fig. 5, A and C). In the absence of InsP 3 , with heparin (100 g/ml) and 25 M free Ca 2ϩ in the pipette solution, P o of capn-InsP 3 R1 was 0.39 Ϯ 0.08 (n ϭ 8; Fig. 5, B and C). P o of wt-and capn-InsP 3 R1 channels in 25 M Ca 2ϩ was significantly lower than the corresponding P o of each channel in 2 M Ca 2ϩ (p Ͻ 0.001). P o of capn-InsP 3 R1 at 25 M Ca 2ϩ was not, however, significantly different from P o of wt-InsP 3 R1 at that Ca 2ϩ concentration. These data demonstrate that Ca 2ϩ regulation of channel activity is preserved in capn-InsP 3 R1.
Decreased Intracellular Ca 2ϩ Stores in Cells Expressing Capn-InsP 3 R1-To investigate the functional consequences of InsP 3 R1 proteolysis on cellular Ca 2ϩ homeostasis, we first used single-cell Ca 2ϩ imaging to examine [Ca 2ϩ ] i and estimate total Ca 2ϩ content of intracellular stores. We identified individual transfected cells by EGFP expression driven by the pIRES2-EGFP expression vector. To avoid potential complications from compensatory changes in gene expression, we performed Ca 2ϩ imaging at an early time point (6 h) after transfection.
[Ca 2ϩ ] i was monitored with Fura-2 in transiently transfected N2a cells expressing wt-InsP 3 R1, capn-InsP 3 R1, or a control plasmid (EGFP). We utilized the Ca 2ϩ ionophore ionomycin (2 M) to liberate Ca 2ϩ from intracellular stores in the absence of extracellular Ca 2ϩ and measured the resulting increase in  [Ca 2ϩ ] i (Fig. 6A). We used ionomycin instead of a sarcoplasmic/endoplasmic reticulum Ca 2ϩ -ATPase (SERCA) inhibitor because it produces a much faster leak and consequent rise in [Ca 2ϩ ] i , providing a better estimate of the Ca 2ϩ content of intracellular stores. Expression of capn-InsP 3 R1 did not effect resting [Ca 2ϩ ] i in N2a cells (Fig. 6B). However, expression of capn-InsP 3 R1 significantly decreased the ionomycin-induced peak [Ca 2ϩ ] i (252 Ϯ 15 nM) compared with control cells (540 Ϯ 30 nM) or cells overexpressing wt-InsP 3 R1 (520 Ϯ 60 nM; p Ͻ 0.001; Fig. 6C). These data indicate that capn-InsP 3 R1 causes partial depletion of intracellular Ca 2ϩ stores, consistent with the hypothesis that calpain-cleaved InsP 3 R1 is a Ca 2ϩ leak channel.
Decreased ER Ca 2ϩ Loading in Cells Expressing Capn-InsP 3 R1-To directly examine the ER Ca 2ϩ leak caused by truncated InsP 3 R1, we monitored [Ca 2ϩ ] ER in N2a cells using Mag-Fura-2 at 6 h post-transfection. We permeabilized the plasma membranes of Mag-Fura-2-loaded EGFP-expressing cells by a brief exposure to digitonin under microscopic observation. After depleting ER stores by washing with ATP-free cytoplasmlike medium for 30 min, we initiated Ca 2ϩ filling of ER stores by perfusing permeabilized cells with saturating 1.5 mM MgATP to stimulate the SERCA pump. After a steady-state [Ca 2ϩ ] ER was reached, we observed the ER Ca 2ϩ leak by inhibiting the pump with cyclopiazonic acid (10 M) and removing ATP ( To compare the kinetics of ER Ca 2ϩ loading and passive leak in cells expressing various InsP 3 R1 constructs, we determined the average rate of ER loading and release by fitting each respective portion of single-cell responses with a single exponential equation. Expression of capn-InsP 3 R1 reduced the rate of ER Ca 2ϩ loading compared with cells expressing wt-InsP 3 R1, although the rate was similar to that observed in EGFP-expressing cells (Fig. 7C), suggesting that SERCA activity was sufficient to overcome the Ca 2ϩ leak through the truncated channel. Expression of capn-InsP 3 R1 and, to a lesser extent, wt-InsP 3 R1 significantly increased the ER Ca 2ϩ leak rate compared with control (p Ͻ 0.001; Fig. 7D). Taken together, these data demonstrate that calpain-cleaved InsP 3 R1 alters ER Ca 2ϩ homeostasis.
Calpain-mediated Proteolysis of InsP 3 R1 after in Vivo Ischemia-To determine whether calpain proteolysis of InsP 3 R1 occurs in vivo in a clinically relevant injury model, we examined rat cerebellum after 7 min of asphyxial cardiac arrest. We chose this model because cerebellar Purkinje neurons express high levels of InsP 3 R1 and are selectively vulnerable to transient global ischemia caused by cardiac arrest (39). We first performed Western blot analysis of rat cerebellar microsomes from naïve animals or from animals resuscitated from cardiac arrest. Using the carboxyl-terminal InsP 3 R1 antibody, which reacts with both unproteolyzed and cleaved InsP 3 R1, we observed evidence of InsP 3 R1 proteolysis and the appearance of a faint ϳ95-kDa fragment at both 1 and 24 h after return of spontaneous circulation (Fig. 8A). Using Ab2054, which specifically recognizes the stable 95-kDa carboxyl-terminal fragment generated by calpain proteolysis, we identified calpain-cleaved InsP 3 R1 in both post-cardiac arrest samples (Fig. 8B).
To determine which cell populations within the cerebellum contained calpain-cleaved InsP 3 R1, we performed immunohistochemical analysis of rat cerebellum 24 and 48 h after cardiac arrest using naïve animals as a control (Fig. 8C). Immunolabeling with Ab38, which reacts with calpain-cleaved spectrin, demonstrated robust calpain activity within subpopulations of Purkinje neurons at 24 h post-cardiac arrest, which decayed but was still evident at 48 h post-cardiac arrest. Using Ab2054, we detected calpain-cleaved InsP 3 R1 in subpopulations of Purkinje neurons in a distribution similar to that of calpain-cleaved spectrin (Fig. 8C). The intensity of Ab2054 staining was comparable at 24 and 48 h after cardiac arrest, suggesting that the calpain-derived InsP 3 R1 fragment is more stable than the spectrin fragment.
Finally, to evaluate the relationship between calpain cleavage of InsP 3 R1 and cell death, we used Fluoro-Jade B to identify degenerating neurons in the cerebellum (Fig. 8C). At both time points examined after cardiac arrest, we observed selective Fluoro-Jade labeling of subpopulations of Purkinje neurons. The intensity of Fluoro-Jade staining was greater at 48 h compared with 24 h post-cardiac arrest and was accompanied by a shrunken somatic morphology consistent with neurodegeneration. The morphological changes and increased Fluoro-Jade intensity in Purkinje cells at 48 h post-injury suggests that both ] ER , shows decreased ER loading in cells expressing capn-InsP 3 R1 (one-way analysis of variance; **, p Ͻ 0.001). C, shown is a summary of averaged single-cell ER loading rates in response to MgATP stimulation of SERCA in transfected cells (unpaired t tests; *, p Ͻ 0.05; **, p Ͻ 0.01). D, a summary of averaged single-cell ER leak rates in response to cyclopiazonic acid inhibition of SERCA in transfected cells is shown. Expression of capn-InsP 3 R1 and to a lesser extent wt-InsP 3 R1 increase leak rate compared with control (unpaired t tests with unequal variance; **, p Ͻ 0.001).
enhanced calpain activity and cleavage of InsP 3 R1 precede neurodegeneration. We did not observe any Ab38, Ab2054, or Fluoro-Jade labeling in naïve tissue. Together, these data demonstrate that calpain cleaves InsP 3 R1 in vivo in a clinically relevant model of brain ischemia and that calpain-mediated proteolysis of InsP 3 R1 is an early molecular event in the injury cascade.

DISCUSSION
The present study is the first functional investigation of calpain-cleaved InsP 3 R1. Based on our results from single-channel electrophysiology and Ca 2ϩ imaging experiments, we conclude that capn-InsP 3 R1 forms a dysregulated channel with InsP 3independent gating that functions as a leak channel in the ER. Moreover, evidence of calpain-mediated InsP 3 R1 proteolysis in the brain after cardiac arrest demonstrates that InsP 3 R1 proteolysis is a clinically relevant cellular pathway that is active in a neurodegenerative disease.
Calpain Proteolyzes InsP 3 R1 at a Unique Cleavage Site-Previous reports (23)(24)(25) and data presented here (Fig. 1) establish that InsP 3 R1 is a proteolytic target of calpain that cleaves the channel at multiple sites to form 200-, 130-, and 95-kDa carboxyl-terminal fragments. Calpains, unlike caspases, do not use consensus sequences for target recognition and proteolysis. Our data show that calpain cleaves InsP 3 R1 between residues 1917 and 1918 to generate the stable 95-kDa channel fragment (Fig. 1, D and E). The InsP 3 R1 calpain cleavage site and surrounding residues are highly conserved in mouse and human homologs and are unique to the type 1 isoform of the channel. Our cleavage specific antibody (Ab2054; Fig. 2) targeting the new amino terminus of InsP 3 R1 after calpain proteolysis provides a useful tool for studying the truncated form of the channel in multiple species, including human. 3 To begin characterizing the functional consequences of calpain proteolysis of InsP 3 R1, we utilized a recombinant channel construct corresponding to the 95-kDa carboxyl-terminal fragment of InsP 3 R1 produced by calpain cleavage (Fig. 3). We expected the recombinant truncated InsP 3 R1 to tetramerize in the ER, as previous studies demonstrated that the transmembrane domains and carboxyl-terminal tail of the channel are sufficient for ER localization and oligomerization (40,41). We had attempted to study capn-InsP 3 R1 on a null-background using the InsP 3 R-deficient DT40 cell line. The extremely low transfection efficiency of these cells requires working in stable lines. However, we were unable to generate stable InsP 3 R-deficient DT40 lines expressing capn-InsP 3 R1 despite our laboratory and others being able to generate stable lines expressing the similarly sized caspase-3 cleaved form of the channel (data not shown) (20,22). The apparent toxicity associated with expression of capn-InsP 3 R1 in the DT40 cell system suggests that the truncated channel may cause cell death or attenuate proliferation. Instead of using stable lines, we studied capn-InsP 3 R1 in transiently transfected N2a cells at early time points after transfection to minimize the impact of potentially confounding compensatory changes as seen in stable lines overexpressing wt-InsP 3 R1 (42)(43)(44). Furthermore, studying capn-InsP 3 R1 in transiently transfected cells expressed under the CMV promoter strongly increases the probability that the recombinant truncated subunits form homotetramers. Thus, heteroligomerization of truncated subunits with endogenous, full-length InsP 3 R1 was unlikely in the studies presented here. However, understanding the possible existence and implications of heteroligomerization of truncated and full-length InsP 3 R requires additional studies.
Our approach to studying the functional consequence of calpain proteolysis of InsP 3 R1 using capn-InsP 3 R1 does have limitations. Foremost, it is unknown if the amino terminus of InsP 3 R1 completely dissociates from the channel domain after proteolysis. Limited digestion of cerebellar microsomes with trypsin results in channel fragments that remain associated via noncovalent or indirect interactions (45,46). The trypsinized channel also remains functional, demonstrated by InsP 3 -induced Ca 2ϩ release from microsomes (46). Thus, expression of capn-InsP 3 R1 may not accurately reflect the conformation that exists after proteolysis in vivo. Caspase-3 proteolysis of InsP 3 R1, however, results in loss of InsP 3 -mediated Ca 2ϩ release from microsomes in a manner that corresponds to the percentage of digestion (17). The same is likely true for calpaincleaved InsP 3 R1. By studying capn-InsP 3 R1, we are likely exam-ining a model of an InsP 3 R1 channel in which all four subunits have been completely proteolyzed by calpain. Understanding the behavior of InsP 3 R1 channels with both intact and calpaincleaved subunits warrants future investigation.
Calpain-cleaved InsP 3 R1 Has InsP 3 -independent, Ca 2ϩ -dependent Channel Gating-To explore how calpain cleavage affects InsP 3 R1 channel function, we examined the singlechannel properties of recombinant capn-InsP 3 R1 using nuclear patch clamp electrophysiology. Recordings from N2a nuclei expressing recombinant InsP 3 R1 revealed channels with smaller single-channel conductance than that of endogenous InsP 3 R. This difference is potentially mediated by interactions with endogenous proteins, which may modulate recombinant rat InsP 3 R1 activity differently than endogenous mouse InsP 3 R. The low conductance of recombinant InsP 3 R1 was cell typespecific, as we did not observe such reduction in conductance when the channels were expressed in HEK293 cells. 3 Nevertheless, it provided a means to distinguish recombinant and endogenous InsP 3 R channels in N2a cells. In examining capn-InsP 3 R1 specifically, we observed activity of single channels in patches in the absence of InsP 3 , indicating that calpain cleavage does not eliminate channel function but instead leads to constitutive InsP 3 -independent gating (Fig. 4B). Previous studies of the caspase-3 cleaved form of InsP 3 R1 similarly suggested that the proteolyzed channel retained activity (21,22).
In addition to InsP 3 , Ca 2ϩ is the most important ligand for InsP 3 R1 (15). Even in saturating concentrations of InsP 3 , [Ca 2ϩ ] i greater than 100 nM is required for the InsP 3 R1 channel to be appreciably activated (15). Unlike the InsP 3 binding site in InsP 3 R1, which is known and missing in the carboxyl-terminal fragment generated by calpain proteolysis, the location of functionally important Ca 2ϩ binding sites within the primary sequence of InsP 3 R1 is largely unknown (15). Therefore, although we expected that capn-InsP 3 R1 activity would be InsP 3 -independent if it formed a functional channel, we made no a priori assumptions about Ca 2ϩ regulation of capn-InsP 3 R1 activity. To experimentally determine if the Ca 2ϩ requirement for channel activation was retained in capn-InsP 3 R1, we also studied recombinant channel gating at non-optimal Ca 2ϩ concentrations (70 nM and 25 M). As expected, the P o of wt-InsP 3 R1 was significantly decreased at 70 nM and 25 M Ca 2ϩ compared with channel P o at optimal 2 M Ca 2ϩ (Fig. 5). Notably, the constitutively active capn-InsP 3 R1 channel demonstrated a similar behavior (Fig. 5). This result suggests that the carboxyl-terminal part of InsP 3 R1 beyond residue 1917 contains functionally relevant Ca 2ϩ binding sites involved in channel activation and inhibition. The P o of capn-InsP 3 R1 at 70 nM Ca 2ϩ was, however, significantly greater than that of wt-InsP 3 R1 at this Ca 2ϩ concentration. This difference may reflect loss of the high affinity, InsP 3 -regulated Ca 2ϩ binding site in the truncated channel (15). However, because of the complexity of Ca 2ϩ regulation of InsP 3 R channel activity, involving multiple functional Ca 2ϩ binding sites (47), future studies are required to determine what elements of Ca 2ϩ regulation are modified in the truncated channel. These investigations may provide critical clues needed to identify additional Ca 2ϩ binding sites within the primary InsP 3 R sequence.
Expression of capn-InsP 3 R1 Decreases Ca 2ϩ Content of Intracellular Ca 2ϩ Stores-How does InsP 3 -independent activity of capn-InsP 3 R1 affect cellular Ca 2ϩ regulation? Using changes in [Ca 2ϩ ] i to indirectly measure the Ca 2ϩ content of intracellular stores, we observed that expression of capn-InsP 3 R1 significantly reduced ionomycin-releasable Ca 2ϩ compared with wt-InsP 3 R1 and EGFP controls, although it did not completely deplete intracellular stores (Fig. 6, A and C). The kinetics of the ionomycin response in capn-InsP 3 R1-expressing cells was also slower than in EGFP wt-InsP 3 R1 controls, which suggests a smaller Ca 2ϩ driving force consistent with a smaller intracellular Ca 2ϩ store. On the other hand, expression of capn-InsP 3 R1 was not associated with an increased basal [Ca 2ϩ ] i , suggesting that Ca 2ϩ transport mechanisms were able to compensate for enhanced ER Ca ϩ leak through the cleaved channel, at least at the early times studied after transfection (Fig. 6B). In ER Ca 2ϩ imaging experiments, expression of capn-InsP 3 R1 significantly reduced the steady-state Ca 2ϩ loading capacity of the ER compared with wt-InsP 3 R1 and EGFP controls (Fig. 7, A and B). The steady-state [Ca 2ϩ ] ER represents the equilibrium between the passive Ca ϩ leak and Ca 2ϩ uptake by SERCA. Under steadystate filling conditions, the enhanced Ca 2ϩ leak induced by expression of capn-InsP 3 R1 exceeds the SERCA-mediated Ca 2ϩ uptake rate, resulting in lower [Ca 2ϩ ] ER . The small effect of capn-InsP 3 R1 expression on the ER-filling rate may reflect the negligible driving force for Ca 2ϩ leak under these conditions (Fig. 7C). Of note, the ER leak rate in capn-InsP 3 R1 was considerably higher than that observed in control or wt-InsP 3 R1-expressing cells (Fig. 7D). Taken together, our results suggest that capn-InsP 3 R1 acts as a Ca 2ϩ leak channel that perturbs normal ER Ca 2ϩ homeostasis.
The results from nuclear patch clamp electrophysiology and Ca 2ϩ imaging experiments together elucidate the physiologic relevance of calpain-mediated InsP 3 R1 proteolysis. Evidence that InsP 3 -independent gating of capn-InsP 3 R1 remains Ca 2ϩdependent may explain why expression of capn-InsP 3 R1 induces only a moderate, albeit significant decrease in intracellular and ER Ca 2ϩ stores, at least at the early time point examined after transfection (6 h). The electrophysiology data suggest that resting [Ca 2ϩ ] i is insufficient to fully activate the truncated channel, thus reducing the Ca 2ϩ leak. It is interesting to speculate that this result may also account for conflicting findings regarding the effects of caspase-3 cleavage of InsP 3 R1 on Ca 2ϩ homeostasis. Previous studies have reported that expression of caspase-cleaved InsP 3 R1, which is only 26 residues longer than the calpain cleaved form, either depletes (21) or does not deplete (22) Ca 2ϩ stores in the resting state. Although it was agreed that caspase-3-cleaved InsP 3 R1 represents a leak channel, the size and impact of that leak remains disputed. It is possible that different [Ca 2ϩ ] i in the two studies resulted in channels with different P o , which would likely lead to distinct effects on cellular Ca 2ϩ homeostasis. Electrophysiological recordings of caspase-cleaved InsP 3 R1, similar to those performed here for capn-InsP 3 R1, may clarify the functional consequences of caspase-mediated channel proteolysis.
InsP 3 R1 Is Cleaved by Calpain after Ischemic Brain Injury-Under pathologic conditions, particularly those associated with elevated [Ca 2ϩ ] i and impaired ATP-dependent Ca 2ϩ removal mechanisms, as in ischemia, calpain cleavage of InsP 3 R1 may be an important mechanistic component of cell death. Brain ischemia and reperfusion dramatically disrupt neuronal Ca 2ϩ homeostasis, and there is compelling evidence for a causal role of both Ca 2ϩ overload and consequent pathologic calpain activation in ischemic neurodegeneration (3,39). As both a Ca 2ϩ regulatory protein and calpain substrate, InsP 3 R1 is at a critical intersection between protease activation and disruption of cellular Ca 2ϩ homeostasis during the molecular injury cascade. Here, we demonstrate that InsP 3 R1 proteolysis occurs in vivo in an animal model of ischemic brain injury (Fig. 8). In this cardiac arrest model, calpain cleavage of InsP 3 R1 occurred in cerebellar Purkinje neurons, which are selectively vulnerable to post-ischemic neurodegeneration (39). Moreover, identification of calpain-cleaved InsP 3 R1 in cerebellar microsomes as early as 1 h after cardiac arrest and reperfusion demonstrates that channel proteolysis is an early event in the cell death cascade rather than merely reflecting broad cellular degradation. The persistence and stability of calpain cleaved InsP 3 R1 at the latest time point examined after cardiac arrest (48 h) also suggests that the leaky, proteolyzed channel may act as a feed-forward mechanism for increased Ca 2ϩ overload and calpain activation that eventually leads to neuronal death. In addition to InsP 3 R1, a number of Ca 2ϩ regulatory proteins are also calpain substrates, including the N-methyl-D-aspartate receptor, plasma membrane Ca 2ϩ ATPase, Na ϩ /Ca 2ϩ exchanger, L-type Ca 2ϩ channel, SERCA, and ryanodine receptor (3). Thus, calpain cleavage of InsP 3 R1 may be an important component of a broader pathway of calpain-mediated disruption of neuronal Ca 2ϩ homeostasis in neurodegenerative diseases associated with Ca 2ϩ dysregulation.
Conclusions-In summary, our results indicate that calpain proteolysis of InsP 3 R1 creates an ER Ca 2ϩ release channel that is InsP 3 -independent and constitutively active. Expression of the truncated channel reduces the content of intracellular Ca 2ϩ stores, which may have detrimental effects on Ca 2ϩ signaling and buffering under pathologic conditions. Evidence of calpain cleavage of InsP 3 R1 in neurons after cardiac arrest provides a potential mechanism to account for decreased InsP 3 binding (10, 11), depletion of ER Ca 2ϩ stores (13), and disruption of Ca 2ϩ homeostasis reported in previous studies of in vivo ischemia and reperfusion. Together, these results provide important insights into a molecular pathway that may act as a feed-forward mechanism to enhance cell death. Furthermore, the results presented here identify a novel target for therapeutic intervention after brain ischemia or in other neurodegenerative disorders associated with Ca 2ϩ dysregulation and protease activation.