Ubiquitous Calpains Promote Caspase-12 and JNK Activation during Endoplasmic Reticulum Stress-induced Apoptosis*

Ubiquitously expressed μ- and m-calpain proteases are implicated in development and apoptosis. They consist of 80-kDa catalytic subunits encoded by the capn1 and capn2 genes, respectively, and a common 28-kDa regulatory subunit encoded by the capn4 gene. The regulatory subunit is required to maintain the stability and activity of μ- and m-calpains. Accordingly, genetic disruption of capn4 in the mouse eliminated both ubiquitous calpain activities. In embryonic fibroblasts derived from these mice, calpain deficiency correlated with resistance to endoplasmic reticulum (ER) stress-induced apoptosis, and this was directly related to a calpain requirement for activation of both caspase-12 and the ASK1-JNK cascade. This study provides compelling genetic evidence for calpain's role in caspase-12 activation at the ER, and reveals a novel role for the ubiquitous calpains in ER-stress induced apoptosis and JNK activation.

Calpains are a family of Ca 2ϩ -dependent intracellular cysteine proteases. By cleaving their protein substrates, ubiquitously expressed -calpain and m-calpain are implicated in a wide variety of biological functions including cell migration, cell cycle regulation, differentiation, and apoptosis (reviewed in Ref. 1). Bothand m-calpains are heterodimers, consisting of a distinct large 80-kDa catalytic subunit, encoded by the genes capn1 and capn2, respectively, and a common small 28-kDa regulatory subunit encoded by the capn4 gene. The small subunit is essential to calpain activities, as shown by in vitro biochemical studies where a 25-amino acid truncation at the C terminus abolished all detectable calpain activity (2). This provided the rationale for the first reported capn4 knock-out mouse model, which interrupted the coding sequences in exon 9 (3,4) and was predicted to truncate 38 C-terminal amino acids from the small subunit. The resulting hypothetical small subunit was not detectable, the steady-state levels of -80, m-80 catalytic subunits were reduced, and no calpain activity was observed (4). More recently, we have developed a conditionally targeted capn4 locus by inserting loxP sites into intron 8 and the noncoding region of exon 11. 2 This allows conditional knock-out of capn4 by Cre-mediated recombination. Although a hypothetical small subunit protein with a 60-amino acid deletion at the C terminus might still be produced, no detectable small subunit was detected, probably because of destabilization of the hypothetical truncated protein. Expression of the large sub-units was also greatly diminished, supporting the proposed role for the small subunit in stabilizing large subunits. As expected, mouse embryonic fibroblasts (MEFs) 3 from these knockouts also lacked any detectable ubiquitous calpain activity. 2 Many reports on calpain function are based on using small molecule inhibitors, which lack specificity. In contrast, this genetic knock-out model is completely selective, and therefore provides a powerful tool to address the physiological functions of the ubiquitous calpains.
Ubiquitousand m-calpains have been suggested to participate in apoptosis by cleaving either pro-apoptotic or anti-apoptotic proteins like p53, Bcl-2, and Bax, depending on the nature of the stimuli and type of cells involved (6 -8). Among these studies, m-calpain was proposed to play a role in a distinct apoptotic pathway initiated by ER stress (9). The ER plays key roles in protein biosynthesis, modification, folding, and trafficking, and it is also the major pool for calcium storage. Perturbation of ER homeostasis abolishes protein folding, and the consequent accumulation of unfolded proteins in the ER imposes so called ER stress on the cell. The ER membrane residing proteins IRE1, Perk, and ATF6 are ER stress sensors, which are maintained in an inactive monomeric state by binding to the chaperone protein Bip. During ER stress, the release of Bip activates these sensors to convey their protective responses, known as the unfolded protein response (UPR). This includes enhancing the transcription of genes encoding ER chaperones and the attenuation of general protein synthesis (10). Autophosphorylated IRE1 possesses an endoribonuclease activity that alternatively splices the mRNA encoding the transcription factor XBP-1, thus turning on chaperone genes (11). The activated Perk kinase phosphorylates eIF-2␣ and inhibits general protein synthesis (12). Also, the unfolded proteins trigger ATF6 cleavage by the Site-1 and Site-2 proteases to release its cytoplasmic domain, which then enters the nucleus and activates Bip transcription (13).
ER stress can be induced by pharmacological agents such as the calcium ionophore A23187, the Ca 2ϩ pump inhibitor thapsigargin, the N-linked glycosylation inhibitor tunicamycin, the ER to Golgi transport inhibitor brefeldin A, and inhibitors of disulfide bond formation such as dithiothreitol (14). Excessive ER stress can trigger cellular apoptosis through the activation of caspase-12, which resides on the outside of ER membrane, and is cleaved and activated during ER stress (15). As an initiator caspase, caspase-12 triggers the activation of caspases-9, -7, and -3 in a cytochrome c and Apaf-1-independent manner (16,17). Caspase-12 knock-out cells are resistant to ER stress-induced apoptosis (15). In vitro biochemical and calpain inhibitor studies in glial cells sug-* This work was supported by an operating grant from the Canadian Institutes of Health Research. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. □ S The on-line version of this article (available at http://www.jbc.org) contains supplemental data. 1  gested that m-calpain is responsible for the cleavage of caspase-12 during ER stress (9).
Besides its role in ER stress induced apoptosis, calpain might also participate in a mitochondrial-dependent apoptotic pathway by cleaving and thus releasing apoptosis-inducing factor (AIF) from the inner mitochondrial membrane (18 -20). Methamphetamine induced apoptosis of striatal glutamic acid decarboxylase containing neurons through the rapid and simultaneous activation of calpain and caspase-12, and release of AIF and cytochrome c from the mitochondria (18). Calpain activation was persistent under those conditions, so it might have further participated in other downstream apoptotic signaling events. Surprisingly, the activation of caspase-9 in this system was not synergistically accelerated, but delayed (18). In vitro studies suggested that calpain-mediated cleavage of an N-terminal AIF domain is required to allow its release from Bid-permeabilized mitochondria (19). Calpains might also mediate cross-talk between mitochondrial and ER apoptotic pathways depending on the nature of the apoptotic stimuli and the cell type. In mouse pancreatic ␤-islet cells, apoptosis induced by type 2 ryanodine receptor-mediated calcium overloading of mitochondria relies on calpain-10 up-regulation and activation (21). Interestingly, polymorphisms in calpain-10 are linked to human susceptibility to type II diabetes.
The c-Jun N-terminal kinase (JNK) pathway is also activated in response to ER stress by IRE1, a serine/threonine kinase, which gets autophosphorylated during UPR. Activated IRE1 recruits TNF receptor-associated factor 2 (TRAF2), which in turn activates the apoptosis signal-regulating kinase ASK1, which in turn activates the MKK4/JNK cascade (22)(23)(24). Although the precise role of JNK in ER stress is not defined, prolonged JNK activation has pro-apoptotic functions downstream of various cell death-inducing stimuli, including TNF␣, H 2 O 2 , and DNA-damaging agents (25,26). Interestingly, protein-tyrosine phosphatase 1B (PTP1B) knock-out cells, which are not capable of activating the IRE1-JNK pathway after ER stress are resistant to apoptosis (27).
In this study, we investigated the role of the ubiquitous calpains in ER stress. MEFs from conditional capn4 knock-out animals were used to establish a role for calpain in two key ER stress apoptosis pathways. We show that calpain-deficient MEFs were resistant to thapsigargin (TG) and tunicamycin (TN)-induced apoptosis. This resistance correlated with a defect in the activation of the caspase-12, -9, -3 cascade and the ASK1/MKK4/JNK/c-Jun cascade. These observations provide compelling genetic and biochemical evidence for a central role for calpain in the conversion of Ca ϩ2 signals from the stressed ER to the caspase-12 apoptotic pathway and JNK activation.
Conditional capn4-targeted Allele-The generation of mice harboring a targeted floxed (loxP-flanked) capn4 allele is described elsewhere. 2 Briefly, calpain small subunit coding sequences in exons 9, 10, and 11 were flanked by two loxP sites. An internal ribosome entry site (IRES)-lacZ reporter and PGK promoter-Neo selectable gene cassette were inserted after the coding sequences in exon 11 and before the second loxP site to permit X-gal staining and G418 selection. Cre recombinasemediated excision of sequences between the loxP sites deleted the intervening sequences as well as the lacZ/Neo cassette. We refer to this allele as capn4 PZ/PZ before Cre-mediated excision, and capn4 P/P after excision. In order to achieve ubiquitous loss of calpain activity, these mice were crossed with a transgenic line which expresses Cre recombinase under the control of the cytomegalovirus (CMV) promoter, referred to as cmv-cre (29). The capn4 P/P ; cmv-cre embryos derived from the capn4 ϩ/P ; cmv-cre x capn4 PZ/PZ breeding pairs died at approximately embryonic day (E) 11.5.
Cell Culture-Primary fibroblast cultures were established from E10.5 embryos. The yolk sacs, heads, and internal organs were isolated and used for genotyping by Southern blot hybridization. Carcasses were treated with 0.25% collagenase in 20% fetal bovine serum in phosphatebuffered saline for 30 min at 37°C. Cells were then washed twice with media and plated in high glucose Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Primary MEFs were used in experiments before the third passage. Spontaneously immortalized MEFs were established as previously described (30). Briefly, primary MEFs were seeded in 60-mm plates at a density of 3 ϫ 10 5 cells/plate and maintained at that density by passage every 3 days. Immortalized MEFs were established between the 13th and 16th passages.
Lentivirus-mediated Gene Transduction-Lentiviral vectors (pWPXLD, pWPI) were kindly provided by Dr. Trono at University of Geneva. The IRES sequence was excised from pWPI and inserted before the EGFP sequence in pWPXLD using blunt-end ligation. The murine capn4 cDNA was cloned using RT-PCR (forward primer ATCCGACTCAGCTGC-GAT; reverse primer TTGAGCTCTGGCTGCTGA) and inserted at a PmeI site upstream of the IRES. Lentivirus was produced by transfecting 293T cells with pWPXLD and the packaging plasmids pMD2G and pCMV-dR8.74 as described (31). The multiplicity of infection (MOI) was determined by infecting MEFs with different titrations of virus-containing supernatant and measuring the percentage of GFP-positive cells using flow cytometry. To rescue calpain activity, immortalized capn4 P/P ; cre MEFs were transduced with lentiviruses encoding recombinant mouse capn4.
Apoptosis Analysis-MEFs were seeded at 3 ϫ 10 5 cells per 60 mm plate and allowed to adhere overnight. On the following day, cells were treated with 3 g/ml tunicamycin (TN) or 5 M thapsigargin (TG) for the indicated periods. Mock treated cells were exposed to equivalent concentrations of the vehicle, Me 2 SO. Apoptosis was measured by 0.4% trypan blue exclusion assays and flow cytometry analysis of PE-conjugated Annexin-V and 7-AAD staining, according to the suppliers instructions. Alternatively, apoptotic cells were quantified by measuring the fraction of cells with a sub-G1 content of PI-staining DNA (32).
Analysis of Lysosome Stability-Cells were exposed to 5 g/ml acridine orange (AO) for 15 min under standard culture conditions, then released by trypsin treatment and washed twice with medium. Red fluorescence was measured immediately by flow cytometry. Cells with decreased red fluorescence (pale cells) were gated, and their percentages were determined.
Western Blotting-Cell monolayers were washed twice with cold phosphate-buffered saline, then incubated for 15 min on ice in lysis buffer (50 mM HEPES, pH 7.6, 150 mM NaCl, 1% Triton X-100, 5 mM EDTA, 10 mM 2-mercaptoethanol, 0.1 mM phenylmethylsulfonyl fluoride, 10 g/ml of leupeptin). Cell lysates were scraped from the dishes, transferred to Eppendorf tubes, and centrifuged at 13,000 rpm for 20 min to remove insoluble material. Protein content was determined using the Bradford method. Samples (20 g) were separated by SDS-PAGE and transferred to polyvinylidene difluoride membranes. Immunoblotting was performed using specific primary antibodies, and detected using horseradish peroxidase-coupled secondary antibodies.
Release of Cytochrome c-2 ϫ 10 6 cells were harvested and washed twice in ice-cold phosphate-buffered saline. The cells were then spun at 200 ϫ g for 5 min. The cell pellet was suspended in 600-l extraction buffer containing 220 mM mannitol, 250 mM sucrose, 50 mM PIPES-KOH (pH 7.4), 50 mM EGTA, 2 mM MgCl 2 , 1 mM dithiothreitol, and protease inhibitors. After 30 min of incubation on ice, cells were treated with a Dounce homogenizer for 80 strokes and spun for 60 min at 14,000 ϫ g. Supernatants were analyzed for cytochrome c release by Western blotting (33).
In Vitro Analysis ofand m-calpain Activities-Calpain activities in cell extracts were detected by casein zymography in a standard Trisglycine system (34). 40 g of protein was resolved on 8% non-denaturing polyacrylamide gels containing 1.5 mg/ml casein. Calpain was activated by incubating the gels overnight with 5 mM Ca 2ϩ , and its activity was assessed by visualization of the casein-cleared regions of the gel after staining with Coomassie Brilliant Blue.
Confocal Fluorescence Microscopy Analysis-To examine m-calpain subcellular localization, NIH3T3 cells were transfected with an m-80-EGFP fusion protein-encoding plasmid (pEGFP-m-80) using Lipofectamine 2000 (Invitrogen). After TG or TN treatment, cells were stained with ER-Tracker Red and imaged by confocal fluorescence microscopy using the green and red channels to visualize m80-EGFP and the ER, respectively. Co-localization was quantified using the Pearson's correlation coefficient, which was calculated using the Image-Pro Plus 5.1 software package (Media Cybernetic Inc).
Statistical Analysis-Results were expressed as means Ϯ S.D. of at least three independent experiments. Statistical analysis was performed using Student's t test, with level of significance set at p Ͻ 0.05.

Capn4 P/P Cells Lack Calpain Activity and Have Reduced Expression of Calpain Subunit
Proteins-Spontaneously immortalized MEFs were established from capn4 ϩ/P ; cre and capn4 P/P ; cre embryos, and calpain activity was analyzed by in vitro casein zymograms (Fig. 1A). As expected,and m-calpain activity was detected in capn4 ϩ/P ; cre MEFs, but not in capn4 P/P ; cre cells. However, when the capn4 P/P ; cre MEFs were transduced with a lentivirus encoding the mouse capn4 cDNA, bothand m-calpain activity were restored. Western blotting analysis of these same cell lysates revealed the expected elimination of small subunit expression in capn4 P/P ; cre MEFs, but we also noted that steadystate levels of both -80 and m-80 were dramatically reduced; and this defect was also rescued by transduction with the capn4 lentivirus (Fig. 1B).
ER Stress Correlated with Calpain ER Translocation and Activation-Calpain inhibitors have been reported to inhibit ER stress-induced apoptosis by inhibiting caspase-12 activation (9). Calpain is widely distributed in the cell cytosol and is associated with membranes; and it might undergo further concentration at membranous structures and in the nucleus in response to specific signals such as calcium (35). Calpain has been localized at the plasma membrane, with cytosolic membrane surfaces, and within the lumens of the ER and Golgi apparatus (36). In primary MEFs, inhibition of N-linked glycosylation in the ER with tunicamycin (TN), or inhibition of the ER-localized Ca 2ϩ -ATPase with thapsigargin (TG) caused an elevation in cytoplasmic Ca 2ϩ levels by about 20%; but no difference was observed between capn4 ϩ/PZ ; cre and capn4 P/P ; cre MEFs (data not shown). To track calpain localization during TG/TN-triggered ER stress, we performed confocal fluorescence microscopy analysis of NIH3T3 fibroblasts transfected with an m-80-EGFP fusion protein-encoding plasmid (28). Recombinant m-80-EGFP was widely dispersed in the cells; however, upon TN or TG treatment, we observed a statistically significant increase in calpain ER localization as assessed by increased spectral overlap of m-80-EGFP and ER-Tracker Red (Fig. 2). This indicated that calpain was recruited to the ER in conjunction with calcium release (Fig. 2B).   right panels). B, degree of co-localization was calculated as the Pearson's correlation factor. The significance between TG and control is p Ͻ 0.001, and between TN and control is p Ͻ 0.002. These data are derived from four independent experiments.
Calpain large subunits have a half-life of 5 days, but they are quickly degraded after activation (1). We therefore assessed the stability of calpain large subunits after TG or TN treatment as an indication of their degree of activation. In primary capn4 ϩ/P ; cre MEFs, m-80 subunit levels declined during TG or TN treatment (Fig. 3). Although there was substantially less m-80 subunit in primary capn4 P/P ; cre MEFs, that amount was further diminished after TM or TG treatment (Fig. 3). The -80 subunit was not as easily detected in capn4 P/P ; cre MEFs; however, in capn4 ϩ/P ; cre MEFs, we determined that the -80 subunit also diminished during TN or TG treatment (Fig. 3). This disappearance of calpain catalytic subunits is consistent with the hypothesis that they are activated and subsequently degraded during ER stress.
Calpain-deficient MEFs Are Resistant to Thapsigargin-induced Cell Death-To establish whether calpain is involved in the ER stress pathway, we compared TN or TG-induced apoptosis in MEFs with or without calpain. We first used primary MEFs (P-MEFs) from capn4 ϩ/P ; cre  . Calpain-deficient MEFs were resistant to ER stress-induced cell death. A, cell death after TG or TN treatment was measured by trypan blue exclusion assays of primary P-capn4 ϩ/P ; cre and P-capn4 P/P ; cre MEFs, as well as immortalized I-capn4 P/P ; cre MEFs and those cells after rescue with a capn4-expressing lentivirus (I-rescue). B, apoptosis was also assessed by flow cytometry after staining with 7-AAD for necrotic cells and annexin-V for apoptotic cells. Cells in the 7-AAD-negative and annexin-V-positive quadrant were considered apoptotic. Cells were assessed 24 h after exposure to TG or TN, as indicated. and capn4 P/P ; cre sibling embryos to perform these apoptosis studies. Similar results were then obtained using spontaneously immortalized capn4 P/P ; cre MEFs (I-MEFs) and their lentiviral-capn4 rescued counterparts (Fig. 4). After 3 g/ml TN or 5 M TG treatment for 24 h, apoptosis was measured by trypan blue staining of total dead cells (Fig.  4A), or annexin-V and 7-AAD staining of apoptotic cells (Fig. 4B). Fewer dead cells were seen in primary capn4 P/P ; cre MEFs compared with the primary capn4 ϩ/P ; cre MEFs in response to either TN or TG treatment. Fewer dead cells were also observed in the immortalized capn4 P/P ; cre MEFs compared with their capn4 rescued counterparts after TG incubation. Apoptosis was also independently assessed by flow cytometry, where cells with a sub-G1 content of PI-staining DNA were considered apoptotic (Fig. 5). This analysis also showed that calpain deficiency in either primary or immortalized MEFs correlated with resistance to ER stress-induced apoptosis. These observations support a permissive role for calpain in ER stress-induced apoptosis.
Calpain-deficient Cells Display Defective ER Stress-induced Activation of the Caspase-12, -9, -3 Cascade-Caspase-12 plays a pivotal role in ER stress-induced apoptosis (15). Like other initiator caspases, activation of pro-caspase-12 requires the cleavage of a short inhibitory peptide. Active caspase-12 can then directly activate downstream caspase-9 independently of cytochrome c release from the mitochondria (9, 16). Active caspase-9 in turn activates the executioner caspase-3, leading to apoptosis. Caspase-12-null cells were resistant to ER stress-induced apoptosis, and in vitro experiments showed that m-calpain could mediate conversion of pro-caspase-12 into an active form by cleavage of the N-terminal pro-peptide regulatory sequence (9). We therefore asked if ER stress-induced caspase-12 cleavage and activation was compromised in calpain-deficient cells. In primary capn4 ϩ/P ; cre MEFs, the 55-kDa caspase-12 was quickly cleaved in response to TN treatment, but such cleavage was delayed in the capn4 P/P ; cre MEFs (Fig. 6A). This was also seen after TG treatment of primary MEFs (Fig. 6B). In the immortalized capn4 PZ/PZ MEFs, caspase-12 cleavage was apparent as early as 6 h and essentially complete after 12 h of TG treatment (Fig. 6C, left lanes). In contrast, TG-induced caspase-12 cleavage was significantly attenuated in the immortalized capn4 P/P ; cre cells (Fig. 6C, middle lanes). Interestingly, the rescued immortalized capn4 P/P ; cre MEFs demonstrated considerably lower levels of caspase-12 expression, and this was further diminished upon TG treatment (Fig. 6C, right lanes). Collectively, these results strongly support a role for calpain in ER stress-induced caspase-12 activation.
Because active caspase-12 can directly cleave and activate caspase-9, which in turn can activate caspase-3, we next explored the requirement for calpain in ER stress-induced activation of caspase-9 and caspase-3, as well as cleavage of the caspase-3 substrate, PARP. In the primary capn4 ϩ/P ; cre MEFs, caspase-3 was quickly activated after TN treatment, while it appeared much later and to a lesser extent in the capn4 P/P ; cre MEFs (Fig. 6A). As expected, PARP cleavage followed precisely the activation kinetics of caspase-3. During TG treatment, weaker activation of the caspase-9/caspase-3/PARP cascade was observed in primary capn4 P/P ; cre MEFs relative to capn4 ϩ/P ; cre MEFs (Fig. 6B). In immortalized capn4 PZ/PZ MEFs, which retain normal calpain small subunit expression and calpain activity, pro-caspase-9 was cleaved to the active form in response to TG treatment (Fig. 6C, left lanes). In the immortalized capn4 P/P ; cre MEFs, only a small amount of cleaved caspase-9 was seen (Fig. 6C, middle lanes). In immortalized capn4 P/P ; cre MEFs that had been rescued with the capn4 lentivirus (I-rescue) we observed a partial rescue of cleaved caspase-9, and this was apparent as early as 12 h after TG exposure (Fig. 6C, right lanes). The kinetics of caspase-3 cleavage and activation correlated with those of caspase-9 activation in these immortalized MEFs, with maximal cleavage of both caspase-3 and PARP at 24 h in calpain-expressing cells, compared with 30 h in the calpain-deficient cells (Fig. 6C ).
Calpain Does Not Regulate the Activation of Bcl-2 Family Members by Direct Cleavage during ER Stress-Although caspase-12 is thought to be the initiator of ER stress-induced apoptosis, it was recently recognized that activation of ER residing Bax/Bak proteins occurs upstream of caspase-12 cleavage, because caspase-12 was not cleaved in Bax/Bak double knock-out cells (37,38). Although the mechanism of Bax/Bak activation is unknown, a role for BH3-only proteins has been proposed (39). Our data strongly suggested that calpain has a direct role in cleaving caspase-12; however, it was still possible that calpain has additional upstream roles, perhaps at the level of Bax/Bak activation. Calpain has been implicated in the regulation of several different Bcl-2 family members through direct cleavage during apoptosis (1). For example, calpainmediated cleavage of Bcl-xl in glial cells is proposed to convert an antiapoptotic molecule into a pro-apoptotic one during ER stress-induced apoptosis (9). We therefore analyzed the protein levels of members of the Bcl-2 family in immortalized capn4 P/P ; cre cells and their capn4 rescued counterparts during ER stress induced apoptosis (Fig. 7). However, we did not observe any reproducible differences in either the steady state or post TG-treatment levels of Bax, Bad, Bak, Puma, Bok, Bik, Bim, Bcl-2, Bcl-X L, and Bcl-X S/L (normal and deaminated forms) (40). This suggested thatand m-calpain did not directly cleave these proteins in MEF cells during ER stress.
Calpain Potentiates Sustained ER Stress-induced Activation of the ASK1 MAP Kinase Pathway-ER stress can induce JNK activation through activation of IRE1. Interestingly, JNK activation was undetectable in primary capn4 P/P ; cre MEFs whereas both the p46 and p54 JNK isoforms were activated in primary capn4 ϩ/P ; cre MEFs (Fig. 8A). Consistent with this, phosphorylation of the downstream JNK target c-Jun, and the upstream JNK activator MKK4 were both compromised in cal-pain-deficient primary capn4 P/P ; cre MEFs. Lastly, we examined the MKK4 kinase, ASK1 (41). Phosphorylation at the ASK1 activation loop threonine 845, which is diagnostic for the active form of ASK1, was induced by TG in primary capn4 ϩ/P ; cre MEFs but not in capn4 P/P ; cre MEFs (Fig. 8A). A similar defect in the activation of ASK1, MKK4, JNK, and c-Jun was observed after TN treatment in primary capn4 P/P ; cre MEFs (Fig. 8B). This defect in the AKT/MKK4/JNK/c-Jun ER stress cascade was also apparent in immortalized capn4 P/P ; cre MEFs, where it was effectively rescued using the capn4 lentivirus (Fig. 8C ). These observations clearly implicate calpain as an important participant in the upstream regulation of the ER stress-induced ASK1-JNK pathway. Other major signaling pathways including ERK, p38, NFB, and Akt were not obviously affected by calpain deficiency in these immortalized MEFs (Fig. 8C ).
Calpain Deficiency Does Not Affect Transcriptional Responses to ER Stress-We next examined some of the major transcriptional responses to TG-triggered ER stress, including genes encoding the chaperone protein Bip, and the transcription factors Xbp-1 and Chop. Immortalized capn4 P/P ; cre MEFs and their capn4 rescued counterparts displayed no differences in Bip, Xbp-1, or Chop transcript levels before or after TG treatment; nor was there any difference in the levels of the alternatively spliced Xbp-1 transcripts (Fig.  9A). This indicated that calpain did not affect IRE1 activation and its subsequent release from the ER because mRNA levels of the ATF6 target gene Bip, and the Perk target gene Chop were indistinguishable between immortalized capn4 P/P ; cre MEFs and their capn4 rescued counterparts (Fig. 9A). However, we did note that resting Bip protein levels (and possibly mRNA levels) were slightly higher in immortalized capn4 P/P ; cre MEFs after transduction with the capn4 rescue lentivirus (Fig. 9, A and B). No difference was observed between the primary capn4 ϩ/P ; cre and capn4 P/P ; cre MEFs (Fig. 9C ). FIGURE 6. Calpain-deficient MEFs displayed defective ER stress-induced activation of the caspase-12, -9, -3 cascade. A, cleavage of caspase-12, caspase-3 and PARP in primary P-capn4 ϩ/P ; cre and P-capn4 P/P ; cre MEFs during TN-induced apoptosis. B, cleavage of caspase-12, caspase-9, caspase-3, and PARP in primary P-capn4 ϩ/P ; cre and P-capn4 P/P ; cre MEFs during TG-induced apoptosis. C, cleavage of caspase-12, caspase-9, caspase-3, and PARP in immortalized I-capn4 PZ/PZ , I-capn4 P/P ;cre and rescued (I-rescue) MEFs during TG-induced apoptosis. Cells were either untreated (C) or treated for the indicated times (6, 12, 24, and 30 h). Western blotting of equivalent amounts of soluble cell lysates was performed with antibodies to the indicated proteins to reveal the positions of pro-or active caspase-12, -9, or -3, ␤ actin or PARP.

DISCUSSION
In this study, we provide novel genetic evidence that calpain participates in the apoptosis response to ER stress. This analysis validates the proposed involvement of calpain in activation of caspase-12, and it also provides novel evidence implicating calpain in activation of ASK1 and the JNK stress response pathway. Although calpain has been implicated in cleavage and regulation of Bcl-2 family members, our analysis does not support such a role in the context of ER stress in MEFs.
MEFs from the conditional capn4 knock-out mice represent a powerful in vitro system to explore the cellular functions of calpain. The capn4 knock-out MEFs lacked any detectableand m-calpain activities or small subunit expression, and this activity and expression was rescued upon lentiviral transduction with capn4 cDNAs (Fig. 1). This genetic model system is superior to interfering RNA methods, which do achieved complete inhibition of calpain expression, and protease inhibitors which may lack specificity or the ability to completely inhibit calpain activity.
Upon exposure to calcium in vitro, calpain subunits undergo autocleavage, which facilitates activation, and ultimately leads to calpain degradation (1). In vivo, ER stress causes calcium efflux from the ER, and this has been reported to activate calpain (42)(43)(44). It has also been suggested that this might involve calpain translocation to the ER (35). Calpain was reported to have a wide subcellular distribution including the ER and perinuclear regions (35,36). In our study, TN or TG treatment quickly elevated intracellular free calcium levels in NIH3T3 fibroblasts, and this was followed by accumulation of m-calpain at the ER within 2 h (Fig. 2). This suggested that calpain translocation to the ER might be linked to regionally elevated calcium concentrations. We also found that prolonged treatment with TN or TG correlated with deple-tion of -80 and m-80 protein levels (Fig. 3). Since activated calpain rapidly undergoes auto-proteolysis, the observed accumulation of calpain at the ER and subsequent depletion of large subunits during ER stress argues that calpain is massively activated at the ER in response to calcium release. A pro-apoptotic role for calpain at the ER was strongly suggested by the observation that ER stress-induced apoptosis was significantly attenuated in calpain-deficient MEFs (Figs. 4 and 5). The increased ER localization of calpain after TN or TG treatment might give it access to ER-residing proteins, including caspase-12. Calpain inhibitors were previously shown to attenuate ER stress-induced caspase-12 activation, and in vitro evidence for calpain-mediated cleavage and activation of caspase-12 provided an attractive model to explain this relationship (9). Our observations in calpain-deficient MEFs now provide compelling genetic data to support this hypothesis. Lack of calpain correlated with a defect in TG-or TN-induced activation of caspase-12, -9, and -3, and cleavage of PARP (Fig. 6). Thus, in response to ER stress, calpain might be locally activated by calcium efflux, allowing it to cleave and activate caspase-12. Activated caspase-12 might then directly activate caspase-9 (independently of the mitochondria) and that in turn might directly activate caspase-3, (9,16,17).
Bcl-2 family members are also involved in regulating the ER stressinduced apoptotic response (45). Translocation of Bcl-2 to the ER protected against ER stress-induced apoptosis (46). On the other hand, caspase-12 was not cleaved and activated in Bax/Bak double knock-out MEFs, but this cleavage could be induced by the expression of ER- targeted Bak (37,38). Bax activation during ER stress might involve displacement of Bcl-2 by BH3-only proteins, such as Bad (47) or Bim (39), leading to the activation of caspase-12. Calpains have also been implicated in cleavage of Bcl-2 family members including Bax (8,48,49) and Bcl-2 (9). This raised the possibility that activation of caspase-12 might involve a role for calpain in cleavage of Bcl-2 family members. However, our analysis of a number of these proteins failed to reveal any difference in their steady state levels or ER stress-induced levels in calpain-deficient cells.
Our data support the hypothesis that calpain might serve as a link from Bax activation to caspase-12 activation. ER stress-activated Bax might promote calcium release from the ER by blocking the inhibitory role of Bcl-x L on InsP 3 R mediated calcium release (50) thus activating calpain. This supports the idea of calpains functioning downstream of Bax. An important caveat to consider in this analysis is that human caspase-12 is apparently not involved in apoptosis; rather it has a role in modulating endotoxin responsiveness and cytokine release (51). Caspase-4 has recently been implicated in ER stress-induced apoptosis in human cells (52). It will therefore be important to examine the potential role of calpain in regulating caspase-4 at the ER.
Calpain might also be involved in other apoptotic mechanisms that cooperate with ER stress-induced signals. However, calpain-deficient cells showed normal induction of p53 expression in response to the DNA damaging agent camptothecin, and this was not altered by TGinduced ER stress (supplementary data). We also did not observe any TG-induced lysosome leakage in calpain deficient cells (supplementary data).
The Bax/caspase-12 pathway and the protective UPR are apparently independent ER stress responses because activation of the Perk pathway was not affected in Bax/Bak double knock-out cells (38). In our analysis, calpain-deficient cells had largely normal IRE1 and Perk functions based upon the induction of Bip and Chop, and the observed appearance of alternatively spliced Xbp-1 (Fig. 9). However, we did observe a striking defect in the JNK pathway in calpain-deficient cells (Fig. 8), and this could be traced upstream to a defect in ASK1 activation. Interestingly, JNK activation shows some mechanistic linkage with caspase-12 cleavage. Active IRE1 recruits TRAF2, which in turn activates the ASK1-MKK4-JNK cascade. However, TRAF2 can also interact with caspase-12 and weakly induce caspase-12 cleavage and activation (53). This might also explain the weak cleavage of caspase-12 in the calpain-deficient cells (Fig. 6).
Because calpain-deficient cells also have a defect in ASK-JNK activation ( Fig. 8) but retain normal IRE1 activation (Fig. 9), calpain might promote recruitment of ASK1 to the ER by TRAF2. However, it remains to be determined what the putative targets of calpain are which might lead to ASK1 activation during ER stress. The mechanism of ASK1 activation has been extensively studied in the context of other stress pathways, and potential regulators include PP5, (41) thioredoxin, 14-3-3 proteins (54, 55), calmodulin-dependent protein kinase IV (5), and reactive oxygen species (54). These agents, especially those with apparent inhibitory roles in ASK1 regulation, would therefore represent attractive potential calpain targets during the ER stress response for future studies.
In conclusion, using gene targeted capn4 knock-out MEFs we have conclusively demonstrated an important pro-apoptotic role for the ubiquitous calpains in the ER stress pathway. Our analysis suggests that Bax-induced calcium release from the ER leads to the recruitment and activation of calpain, which is responsible for the activation of caspase-12, and subsequent activation of caspase-9 and caspase-3. We also provide novel evidence implicating calpain in the activation of ASK1, and subsequent activation of MKK4 and JNK. It remains to be determined what upstream calpain targets are responsible for ASK1 activation.