Portal de Periódicos da CAPES

Differential contribution of Puma and Noxa in dual regulation of p53-mediated apoptotic pathways

2006; Springer Nature; Volume: 25; Issue: 20 Linguagem: Inglês

10.1038/sj.emboj.7601359

1460-2075

Tsukasa Shibue, Saori Suzuki, Hideaki Okamoto, Hiroki Yoshida, Yusuke Ohba, Akinori Takaoka, Tadatsugu Taniguchi,

Cancer-related Molecular Pathways

Article5 October 2006free access Differential contribution of Puma and Noxa in dual regulation of p53-mediated apoptotic pathways Tsukasa Shibue Tsukasa Shibue Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, JapanPresent address: Whitehead Institute for Biomedical Research, Cambridge, MA, USA Search for more papers by this author Saori Suzuki Saori Suzuki Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Department of Plastic and Reconstructive Surgery, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Hideaki Okamoto Hideaki Okamoto Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Saga Medical School, Saga, Japan Search for more papers by this author Hiroki Yoshida Hiroki Yoshida Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Saga Medical School, Saga, Japan Search for more papers by this author Yusuke Ohba Yusuke Ohba Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, JapanPresent address: Division of Pathophysiological Science, Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan Search for more papers by this author Akinori Takaoka Akinori Takaoka Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Tadatsugu Taniguchi Corresponding Author Tadatsugu Taniguchi Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Tsukasa Shibue Tsukasa Shibue Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, JapanPresent address: Whitehead Institute for Biomedical Research, Cambridge, MA, USA Search for more papers by this author Saori Suzuki Saori Suzuki Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Department of Plastic and Reconstructive Surgery, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Hideaki Okamoto Hideaki Okamoto Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Saga Medical School, Saga, Japan Search for more papers by this author Hiroki Yoshida Hiroki Yoshida Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Saga Medical School, Saga, Japan Search for more papers by this author Yusuke Ohba Yusuke Ohba Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, JapanPresent address: Division of Pathophysiological Science, Graduate School of Medicine, Hokkaido University, Sapporo, Hokkaido, Japan Search for more papers by this author Akinori Takaoka Akinori Takaoka Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Tadatsugu Taniguchi Corresponding Author Tadatsugu Taniguchi Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan Search for more papers by this author Author Information Tsukasa Shibue1, Saori Suzuki1,2, Hideaki Okamoto3, Hiroki Yoshida3, Yusuke Ohba1, Akinori Takaoka1 and Tadatsugu Taniguchi 1 1Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan 2Department of Plastic and Reconstructive Surgery, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, Tokyo, Japan 3Division of Molecular and Cellular Immunoscience, Department of Biomolecular Sciences, Saga Medical School, Saga, Japan *Corresponding author. Department of Immunology, Graduate School of Medicine and Faculty of Medicine, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan. Tel.: +81 3 5841 3375/73; Fax: +81 3 5841 3450; E-mail: [email protected] The EMBO Journal (2006)25:4952-4962https://doi.org/10.1038/sj.emboj.7601359 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The activation of tumor suppressor p53 induces apoptosis or cell cycle arrest depending on the state and type of cell, but it is not fully understood how these different responses are regulated. Here, we show that Puma and Noxa, the well-known p53-inducible proapoptotic members of the Bcl-2 family, differentially participate in dual pathways of the induction of apoptosis. In normal cells, Puma but not Noxa induces mitochondrial outer membrane permeabilization (MOMP), and this function is mediated in part by a pathway that involves calcium release from the endoplasmic reticulum (ER) and the subsequent caspase activation. However, upon E1A oncoprotein expression, cells also become susceptible to MOMP induction by Noxa, owing to their sensitization to the ER-independent pathway. These findings offer a new insight into differential cellular responses induced by p53, and may have therapeutic implications in cancer. Introduction An appropriate response to stress stimuli is crucial for the prevention of cellular transformation as well as the maintenance of normal tissue function. The tumor suppressor p53 plays central roles in the organization of stress responses (Ko and Prives, 1996). The p53 protein accumulates promptly in stressed cells, and is activated as a transcription factor. Activated p53 in turn induces stress responses mainly through the transcriptional regulation of effector molecules, each of which is involved in the execution of a specific response (Vousden and Lu, 2002; Oren, 2003). p53 activation can result in the elicitation of two opposing responses, namely, eliminating a cell by activating the apoptotic pathway or supporting the preservation of a cell by arresting the cell cycle (Vousden and Lu, 2002; Oren, 2003). In effect, the balance between p53-induced apoptosis and cell cycle arrest should be important for the simultaneous achievement of eliminating cancerous cells and saving normal cells. The outcome of p53 activation can be affected by the cell state as well as the cell type (Vousden and Lu, 2002; Oren, 2003). It has been shown that the apoptotic response is favorably selected in cells expressing oncoproteins such as adenovirus E1A. Mouse embryonic fibroblasts (MEFs) normally undergo p53-mediated cell cycle arrest in response to stresses caused by DNA damage or serum deprivation, whereas MEFs expressing E1A undergo p53-mediated apoptosis under the same conditions (Lowe et al, 1993a). Moreover, several types of cell, including thymocytes (Clarke et al, 1993; Lowe et al, 1993b) and epithelial stem cells in the crypts of the small intestine (Merritt et al, 1994), are ready to apoptose in a p53-dependent manner upon DNA damage, even in the absence of oncoproteins. To date, it remains largely unclear how contextual factors such as the state and the type of cell affect the progression of the p53-mediated apoptotic response. The molecular pathways of p53-mediated apoptosis are still not fully understood. Several proapoptotic molecules have been shown to be transcriptionally induced by p53; moreover, the transcription-independent role of p53 in the promotion of apoptosis has also been described (Zamzami and Kroemer, 2005). The contribution of each factor to the entire p53-mediated apoptotic response is still under extensive investigation. Among the proapoptotic transcriptional targets of p53, Puma (Nakano and Vousden, 2001; Yu et al, 2001) and Noxa (Oda et al, 2000) belong to the BH3-only group of the Bcl-2 family, and are considered to indirectly induce mitochondrial outer membrane permeabilization (MOMP), known to be induced by the activation of Bax and Bak (Letai et al, 2002): Puma and Noxa interfere with the interaction of prosurvival Bcl-2 family members with the proapoptotic Bax and Bak, through their direct binding to the prosurvival members. Indeed, this BH3-only subgroup is termed by Korsmeyer and colleagues as ‘sensitizers’ vis-à-vis ‘activators’ that directly engage Bax and Bak (Letai et al, 2002) for their activation and subsequent MOMP induction (Wei et al, 2001). The contributions of Puma and Noxa to p53-mediated apoptosis have been studied using gene-targeting strategies in mice (Shibue et al, 2003; Villunger et al, 2003). Both these factors play crucial roles in the DNA-damage-induced apoptosis of MEFs expressing E1A, whereas only Puma is involved in the X-ray-irradiation-induced apoptosis of thymocytes (Shibue et al, 2003; Villunger et al, 2003). It has been reported that Puma interacts with various prosurvival Bcl-2 family members, whereas Noxa selectively interacts with Mcl-1 and A1 of this family (Chen et al, 2005). The important roles of Puma and Noxa in p53-mediated apoptosis, together with the observation that the progression of p53-mediated apoptosis is affected by the cellular context, prompted us to further study apoptotic pathways induced by Puma and/or Noxa. We report here the presence of a hitherto unrecognized pathway of p53-mediated apoptosis involving calcium release from the endoplasmic reticulum (ER), which is activated by Puma but not Noxa. Our results may also offer a molecular basis for differential cellular responses to stress stimuli, which are commonly mediated by p53. Results Distinct proapoptotic functions of Puma and Noxa We first generated NIH3T3 fibroblasts that constitutively express adenovirus E1A, cellular oncoprotein E2F1 and papilloma virus E7 (hereafter referred to as E1A-3T3, E2F1-3T3 and E7-3T3 cells, respectively), all of which interfere with the mechanism of retinoblastoma protein-mediated cell cycle regulation, and may therefore alter the fate of cellular responses (Lavia et al, 2003). The effects of Puma and Noxa on these generated and control NIH3T3 cells were studied by the retroviral expression of these genes. Interestingly, Noxa did not induce the death of control NIH3T3, E2F1-3T3 or E7-3T3 cells, but did induce that of E1A-3T3 cells 24 h after its expression, whereas Puma induced the death of all cells at the same time point (Figure 1A). It is worth noting that the level of ectopically expressed Noxa is even lower in E1A-3T3 cells than in control NIH3T3 cells (Supplementary Figure 1A). Therefore, the specific function of Noxa in E1A-3T3 cells likely reflects a change in the cellular state induced by E1A, but not the level of Noxa expression; E1A expression results in the sensitization of cells to Noxa-induced apoptosis. E1A-dependent sensitization to Noxa-induced apoptosis, as well as the differential effects of Puma and Noxa on cells that do not express E1A, was similarly observed in MEFs (Supplementary Figure 2A). Figure 1.Noxa- and Puma-induced apoptosis in the absence and presence of oncoproteins. (A) Noxa- and Puma-induced apoptosis of NIH3T3 cells expressing various oncoproteins. E1A-3T3, E2F1-3T3 and E7-3T3 cells as well as NIH3T3 cells expressing control vectors (pBabe and pLR) were infected with the retrovirus expressing human (pMx-hNoxa) or mouse Noxa (pMx-mNoxa), human Puma (pMx-hPuma) or the control retrovirus (pMx). Testing with a GFP-expressing pMx construct indicated that the efficiencies of pMx-derived retrovirus infection of control and oncoprotein-expressing NIH3T3 cells were consistently about 80 and 70%, respectively, at 24 h after infection. Values shown are means±s.d. of triplicate samples. (B) Oligomerization of Bax (left panel) and Bak (right panel). Mitochondria-enriched heavy-membrane fractions were prepared from NIH3T3 cells infected with the indicated retrovirus. Obtained protein samples were subsequently treated with the chemical crosslinker 1,6-bismaleimidohexane (BMH) or dimethylsulfoxide (DMSO), and analyzed by immunoblotting. Markers Mo, Di, Tr and Te represent the sizes of monomer, dimer, trimer and tetramer, respectively. (*) An intramolecularly crosslinked Bak monomer. (C, D) MOMP induced by Noxa and Puma in control NIH3T3 cells and E1A-3T3 cells. Membrane insertion of Bax (C) and cytosolic release of cytochrome c (cyt. c) (D) were analyzed. In panel C, cell homogenates were treated with alkali, and subsequently separated by centrifugation. Bax molecules residing in the cytosol or loosely attaching to the membrane are separated into the supernatant (S) fraction, whereas membrane-inserted Bax molecules are separated into the pellet (P) fraction. In panel D, the supernatant and pellet fractions obtained by the digitonin treatment and subsequent centrifugation were analyzed, in which supernatant corresponds to the cytosolic fraction and pellet corresponds to the membrane and nuclear fraction. Download figure Download PowerPoint To gain further insights into the proapoptotic functions of Puma and Noxa in the above cells, each expressing a distinct oncoprotein, we analyzed events associated with MOMP, which has been suggested to be the rate-limiting step in the mitochondrial pathway of apoptosis (Green and Kroemer, 2004). We first examined the oligomerization of Bax and Bak, one of the hallmarks of the progression of MOMP (Wei et al, 2001). Perhaps surprisingly, we found that Bax is substantially oligomerized in E1A-3T3 cells but not in E2F1-3T3, E7-3T3 or control NIH3T3 cells (Figure 1B). Although Noxa could not induce Bax oligomerization in NIH3T3 cells, it did enhance the oligomerization of Bax in E1A-3T3 cells (Figure 1B). Noxa failed to induce Bak oligomerization in either of these cells, suggesting that Noxa selectively acts on Bax (Figure 1B). Further analysis of the activation of Bax, using an antibody that specifically detects the activation-related conformational change of the Bax protein (Hsu and Youle, 1997), revealed that E1A expression induces this change, which is further enhanced by the expression of Noxa (Supplementary Figure 3). In contrast, Puma effectively induced the oligomerization of Bax and Bak even in the absence of E1A expression (Figures 1B). These observations are consistent with those of other studies showing that Puma functions independently of the cell type by interfering with more Bcl-2 prosurvival members than does Noxa (Chen et al, 2005), and also raise the question of why Puma is more potent than Noxa in inducing apoptosis. We next studied the membrane insertion of Bax, as well as the cytosolic release of cytochrome c, the events directly associated with MOMP (Liu et al, 1996; Goping et al, 1998). As expected from the above results, Puma but not Noxa induced the membrane insertion of Bax and the cytosolic release of cytochrome c in NIH3T3 cells within 20 h after its expression (Figure 1C and D). Interestingly, both the membrane insertion of Bax and the cytosolic release of cytochrome c were almost undetectable in E1A-3T3 cells, unless Noxa or Puma was expressed (Figure 1C and D). Thus, we interpret that Bax oligomerization detected in E1A-3T3 cells is insufficient to induce Bax membrane insertion for apoptosis and that Noxa and Puma can induce apoptosis in these cells by further promoting the activation of Bax (i.e., enhancing oligomerization to allow its membrane insertion) to trigger cytochrome c release. Although the oligomerization of neither Bax nor Bak was observed in control NIH3T3 cells, Bax and Bak were effectively activated for MOMP induction upon Puma expression (Figure 1B–D). These results in toto suggest that, in addition to the common apoptotic pathway induced by Puma and Noxa, Puma may activate an additional pathway that would converge to MOMP during p53-dependent apoptosis. Differential requirement of caspases for MOMP induction Caspases, a group of proteases essential for apoptosis, have been generally supposed to be activated in the downstream of MOMP (Degterev et al, 2003). However, it has also been indicated that several members of the caspase family can be activated before MOMP, and play crucial roles in the progression of MOMP, at least in some forms of apoptosis (Lassus et al, 2002; Marsden et al, 2002). On the basis of these reports, we next examined the role of caspases in Noxa-induced and Puma-induced apoptotic responses. Treatment with zVADfmk, a ubiquitous inhibitor of caspases, markedly decreased the rate of cell death induced by Puma in control NIH3T3 cells, as well as that induced by Noxa or Puma in E1A-3T3 cells (Figure 2A). Interestingly, however, the membrane insertion of Bax and the cytosolic release of cytochrome c were inhibited only in the Puma-induced apoptosis of control NIH3T3 cells (Figure 2B and C). These results in toto suggest that the induction of MOMP is indeed mediated by two pathways. That is, in control NIH3T3 cells, the induction of MOMP is mediated, at least in part, by a zVADfmk-sensitive pathway, which is activated by Puma but not Noxa, whereas in E1A-3T3 cells a zVADfmk-insensitive MOMP-inducing pathway, which is activated by both Noxa and Puma, operates affectively. Time-course analysis of cell viability using another caspase inhibitor, Q-VD-OPh, revealed that the Puma-induced death of control NIH3T3 cells is almost completely inhibited up to 72 h after infection with a Puma-expressing retrovirus, whereas the Noxa- or Puma-induced death of E1A-3T3 cells gradually progresses even in the presence of this inhibitor (Supplementary Figure 1B). These results can be interpreted as follows. In the former case, control NIH3T3 cells remain almost completely alive up to 72 h because MOMP progression is impaired or delayed by Q-VD-OPh treatment. In contrast, MOMP occurs even in the presence of Q-VD-OPh in the latter cases; hence, E1A-3T3 cells eventually start dying, which is consistent with the report showing that MOMP reduces cell viability even in the absence of massive caspase activation (Ekert et al, 2004). These results collectively suggest the fundamental difference between zVADfmk-sensitive and -insensitive pathways of MOMP induction. Figure 2.zVADfmk-sensitive and -insensitive pathways of MOMP induction. (A) Effect of zVADfmk on Puma- and Noxa-induced cell death in control NIH3T3 cells and E1A-3T3 cells. Under these experimental settings, zVADfmk treatment did not have clear cytotoxic effects. Values shown are means±s.d. from triplicate samples. (B, C) Inhibitory effect of zVADfmk on MOMP. The effects of zVADfmk treatment on membrane insertion of Bax (B) and cytochrome c release (C) were analyzed. In panel C, the amount of cytochrome c detected in the supernatant fractions of Noxa- or Puma- expressing E1A-3T3 cells was reproducibly increased by the zVADfmk treatment. Because the amount of cytochrome c detected in the pellet fraction decreased to a similar extent regardless of zVADfmk treatment, difference in the supernatant fraction is presumably due to the inhibitory effect of zVADfmk on the progression of apoptosis, that is, zVADfmk-treated cells do not readily lose their plasma membrane integrity and thereby they should retain released cytochrome c molecules in the cytosol for longer periods than untreated cells. (D) Puma-induced Bax membrane insertion in wild-type (WT) and Apaf1-deficient (Apaf1−/−) MEFs. WT and Apaf1−/− MEFs were prepared from the same litter. Notably, Puma-induced membrane insertion of Bax occurred in an Apaf-1-independent but caspase-dependent manner. Download figure Download PowerPoint To delineate further the Puma-activated, zVADfmk-sensitive MOMP induction pathway, we expressed Puma in MEFs deficient in the Apaf1 gene (Apaf1−/− MEFs), in which the activation of post-MOMP apoptotic events is severely impaired (Yoshida et al, 1998). As shown in Figure 2D, the Puma-induced membrane insertion of Bax was inhibited in both wild-type and Apaf1−/− MEFs following zVADfmk treatment, indicating that a caspase(s) that contributes to the insertion of Bax is activated by Puma upstream of Apaf-1, namely, in the absence of post-MOMP apoptotic events. In contrast, Puma-induced apoptosis was almost completely suppressed in Apaf1−/− MEFs at least 24 h after the expression of Puma (Supplementary Figure 2B), collectively indicating that Puma activates a caspase(s) involved upstream of MOMP induction, whereas Apaf-1 and other caspases (such as caspase-3) are required for post-MOMP apoptotic events. Involvement of caspase-12 in Puma-induced apoptosis To identify the caspase(s) responsible for the induction of MOMP, we used peptide-based inhibitors that show specificity for several members of the caspase family (Garcia-Calvo et al, 1998). Among these, only zWEHDfmk inhibited Puma-induced Bax membrane insertion in NIH3T3 cells (Figure 3A). Indeed, whereas zWEHDfmk interfered with Bax membrane insertion and cell death induced by Puma in NIH3T3 cells, it failed to do so in the case of Noxa- and Puma-induced apoptosis of E1A-3T3 cells (Figure 3B and Supplementary Figure 4A). Furthermore, protease activity that cleaves the W-E-H-D peptide was induced by Puma in NIH3T3 cells, but not by Noxa in E1A-3T3 cells (Supplementary Figure 4B), indicating the specific involvement of W-E-H-D peptide-cleaving caspase(s) in the former case. Figure 3.Essential role of caspase-12 in Puma-induced MOMP and apoptosis. (A) Effects of various peptide-based inhibitors of caspases on Puma-induced Bax membrane insertion. NIH3T3 cells expressing Puma were treated with DMSO, a pan-caspase inhibitor zVADfmk, or various peptide-based inhibitors that specifically inhibit several members of the caspase family (from zWEHDfmk to zLEEDfmk). (B) Effect of zWEHDfmk on Noxa- and Puma-induced cell death in control NIH3T3 cells and E1A-3T3 cells. (C) Caspase-12 (casp-12) processing during Puma-induced apoptosis of NIH3T3 cells and Noxa-induced apoptosis of E1A-3T3 cells. Caspase-12 in the total cell lysate of indicated cells was detected by immunoblotting using the anti-caspase-12 rat monoclonal antibody. The processed form of caspase-12, a possible intermediate for the conversion of the precursor (procasp-12) into the active form, was detected specifically during Puma-induced apoptosis of NIH3T3 cells. Cells were harvested at the indicated times after infection with the pMx-derived retrovirus. (D–F) Effect of shRNA-mediated knock-down of caspase-12 on Puma-induced apoptosis of NIH3T3 cells. NIH3T3 cells were infected with lentivirus that expresses either scramble shRNA or shRNA targeting caspase-12. Three different caspase-12-targeting shRNA sequences (sh casp-12 A–C) were tested, and each resulted in 64, 83 and 63% reduction in the level of procaspase-12 expression, respectively (D). The effects of caspase-12-targeting shRNA on Puma-induced cell death (E) and Bax membrane insertion (F) were also analyzed. In panels B and E, values shown are means±s.d. from triplicate samples. Download figure Download PowerPoint The W-E-H-D peptide is a substrate of caspase-1 (human and mouse), -4 and -5 (human) (Thornberry et al, 1997). These three caspases form a subgroup in the caspase family together with caspase-11 and -12 (mouse), all of which share structural and functional similarities (Degterev et al, 2003). As caspase-1 is not expressed in NIH3T3 cells (Li et al, 1995), remaining candidates in this murine cell line are caspase-11 and -12. For caspase-11, the L-E-H-D peptide is a more preferable substrate than the W-E-H-D peptide (Kang et al, 2000), but zLEHDfmk showed no inhibitory effect on the Puma-induced membrane insertion of Bax (Figure 3A). We therefore focused on caspase-12, and found that this caspase is indeed processed for activation in NIH3T3 cells expressing Puma (Figure 3C). As caspase-12 predominantly resides on the cytoplasmic surface of the ER membrane (Nakagawa et al, 2000), this suggests that caspase-12 is activated by Puma on ER. We further tried to suppress the expression of caspase-12 in NIH3T3 cells using several different short hairpin RNA (shRNA) sequences that specifically interfere with the caspase-12 expression (Figure 3D), and found that Puma-induced cell death was suppressed by up to 50% 24 h after the expression of Puma, which is accompanied by an impaired Bax membrane insertion (Figure 3E and F). Moreover, Puma-induced cell death was partly impaired by caspase-12 deficiency in MEFs (Supplementary Figure 2C). These data collectively indicate the essential role of caspase-12 in Puma-induced MOMP and apoptosis in these fibroblasts, at least at their onset. Puma-mediated calcium release from ER We then determined how Puma, which is generally considered to reside on mitochondria, activates caspase-12 presumably on the ER membrane. The most established function of caspase-12 is its contribution to the apoptotic pathway induced by ‘ER stresses’ (Orrenius et al, 2003), which include all the insults that interfere with the physiological functions of ER, such as the deregulated mobilization of calcium through the ER membrane and the accumulation of misfolded proteins in the ER lumen (Orrenius et al, 2003). Accordingly, we considered the possibility that Puma also perturbs ER function and thereby activates caspase-12. Calcium release from ER and the resultant increase in cytosolic free-calcium concentration ([Ca2+]c) is commonly triggered by most forms of ER stress stimuli, and is supposed to play central roles in the ER-initiated apoptotic response (Breckenridge et al, 2003). Assuming that Puma activates the apoptotic response through the perturbation of ER function, Puma-induced apoptosis should be affected by the interference of this calcium release. The release of calcium from ER is primarily achieved through two different types of channel, the inositol 1,4,5-triphosphate receptor (InsP3R) and ryanodine receptor (RyR) families (Breckenridge et al, 2003). We therefore attempted treatment with chemical inhibitors specific for each type of calcium channel, and found that xestospongin C (xestC) and 2-aminoethoxy-diphenylborate (2-APB), both of which are blockers of InsP3Rs, interfered with the Puma-induced apoptosis of NIH3T3 cells (Figure 4A). In contrast, dantrolene, a specific inhibitor of RyRs, did not exert significant effects (Figure 4A). In addition, Puma-induced caspase-12 processing and Bax membrane insertion were both specifically inhibited by the treatment with xestC or 2-APB (Figure 4B and Supplementary Figure 5A). Moreover, shRNA-mediated suppression of type I InsP3R expression caused partial inhibition of Puma-induced death of NIH3T3 cells (Figure 4C and D). We also investigated the link between Puma and InsP3Rs by the immunoprecipitation analysis. As shown in Figure 4E, Puma was not clearly co-precipitated with InsP3Rs; however, in accordance with the previous reports (Chen et al, 2004; White et al, 2005), co-precipitation was observed between InsP3Rs and the prosurvival Bcl-2 family members Bcl-2 or Bcl-XL (Figure 4E). It may be worth noting that levels of this coprecipitation were significantly lower when Puma is expressed (Figure 4E), suggesting the possibility that Puma affects the function of InsP3Rs indirectly through its interaction with prosurvival Bcl-2 family members. These results collectively support the notion that Puma induces calcium release from ER through InsP3Rs, which promotes subsequent activation of caspase-12 and further progression of the apoptotic response. Figure 4.InsP3R-mediated calcium release from ER during Puma-induced apoptosis. (A, B) Effects of the blockade of ER calcium channels on Puma-induced apoptosis of NIH3T3 cells. Puma-induced cell death (A) and caspase-12 processing (B) in NIH3T3 cells were analyzed in the presence of the blockers of InsP3R (xestC and 2-APB) or dantrolene, the blocker of ryanodine receptor (RyR). (C, D) Impairment of Puma-induced apoptosis by the shRNA-mediated knock-down of type I InsP3R expression. NIH3T3 cells were infected with lentivirus that expresses either scramble shRNA or shRNA targeting type I InsP3R. Three different type I InsP3R-targeting shRNA sequences (sh InsP3RI A–C) were tested, and each resulted in 6, 84 and 60% reduction in the level of type I InsP3R expression, respectively (C). The effects of these shRNA on Puma-induced cell death (D) were also analyzed. (E) Interaction between Bcl-2 family proteins and InsP3Rs. NIH3T3 cells infected with Puma-expressing, Noxa-expressing or control retrovirus were analyzed by immunoprecipitation. The lysate from each sample was immunoprecipitated with an antibody to type I InsP3R (InsP3RI) (upper panel) or an antibody to type III InsP3R (InsP3RIII) (lower panel), and subsequently analyzed by immunoblotting. In panels A and D, values shown are means±s.d. from triplicate samples. Download figure Download PowerPoint To examine directly the effect of Puma on the mobilization of calcium, we used yellow cameleon 3.60 (YC3.60), a cytosolic calcium indicator based on the principle of fluorescent resonance energy transfer (Nagai et al, 2004). As shown by the increased emission ratio of YC3.60, [Ca2+]c was elevated by Puma in NIH3T3 cells, whereas Noxa did not significantly change [Ca2+]c in E1A-3T3 cells (Supplementary Figure 6A). Puma-induced increase in [Ca2+]c was significantly suppressed by the treatment with xestC, supporting the notion that this [Ca2+]c change was caused by the release of ER calcium through InsP3Rs (Supplementary Figure 6B). Moreover, [Ca2+]c change was still observed in the presence of zVADfmk, indicating that calcium mobilization occurs in the upstream of caspases activation (Supp