Acid sphingomyelinase is indispensable for UV light-induced Bax conformational change at the mitochondrial membrane.

Ultraviolet light-induced apoptosis can be caused by DNA damage but also involves immediate-early cell death cascades characteristic of death receptor signaling. Here we show that the UV light-induced apoptotic signaling pathway is unique, targeting Bax activation at the mitochondrial membrane independent of caspase-8 or cathepsin D activity. Cells deficient in acid sphingomyelinase (ASMase) do not show UV light-induced Bax activation, cytochrome c release, or apoptosis. In ASMase-deficient cells, the apoptotic UV light response is restored by stable or transient expression of human ASMase. Bax conformational change in ASMase(-/-) cells is also caused by synthetic C(16)-ceramide acting on intact cells or isolated mitochondria. The results suggest that UV light-triggered ASMase activation is essentially required for Bax conformational change leading to mitochondrial release of pro-apoptotic factors like cytochrome c and Smac.

Ultraviolet light induces a complex transcriptional cellular response that is similar to that of tumor promotors and mitogens yet also includes the induction of growth arrest and apoptosis (1,2). UV light-induced apoptosis seems to represent a controlled scavenging mechanism that protects cells from malignant transformation. UV light triggers a variety of signaling pathways, including nuclear DNA damage and activation of the tumor suppressor gene p53 (3) or triggering of cell death receptors either directly (4,5) or by autocrine release of death ligands leading to mitochondrial damage and cytochrome c release (6,7), which leads to the activation of executioner caspases and eventually to apoptosis.
The primary mechanism of UV light-induced apoptosis was originally thought to be the signaling induced by DNA damage and p53 (3), which could be promoted by UV light-induced increase of reactive oxygen species (8). However, the UV response also entails a number of p53-independent signaling pathways (9 -13). For example, exposure to UV light induces clustering and internalization of cell surface death receptors, like TNF 1 -R1. This may result not only in activation of the c-Jun N-terminal protein kinase cascade but also in caspase-8 activation (4,5). Caspase-8 can cleave Bid to form truncated (t)Bid that then translocates from the cytosol to mitochondrial membranes causing cytochrome c release (14). Furthermore, like various other apoptotic stimuli, including CD95, UV light activates acid as well as neutral sphingomyelinases (15). Although the activation of sphingomyelinases (SMases) by apoptotic stimuli is well established and evidence has been accumulated for their role in apoptosis (16 -19), the precise molecular mechanisms of SMase pro-apoptotic action remained ill-defined and in dispute (20).
Despite this variety of established UV light-induced signaling pathways, there are gaps in our understanding of the rapid induction of mitochondrial pro-apoptotic function that appears to be an important hallmark of the UV light response (21). The pro-apoptotic activity of mitochondria is known to be regulated by the family of Bcl2 proteins (22)(23). The Bcl2 family consists of anti-apoptotic members such as Bcl2 but also includes proapoptotic members such as Bax. Bax can induce the release of cytochrome c from mitochondria and represents an important relay of the mitochondrial apoptotic pathway and subsequent activation of caspases (24). In apoptotic cells, Bax is translocated from the cytosol to the mitochondria, undergoes a conformational change, oligomerizes, and inserts into the mitochondrial outer membrane by which it triggers the cytochrome c release and gains its pro-apoptotic activity (25)(26). The proapoptotic action of Bax is antagonized by Bcl2 and Bcl-X L , which inhibit the release of cytochrome c from mitochondria (23). In unstimulated cells, Bax is a soluble monomeric protein present in the cytosol or loosely associated with mitochondria (27). On isolated mitochondria, Bax monomers do not trigger the release of cytochrome c, whereas Bax oligomers do (27). In order to oligomerize, Bax is suggested to undergo first a conformational change, which can be demonstrated in cell lysates using detergents like Triton X-100 (28). Immunoprecipitation and immunofluorescence analysis with conformationally sensitive antibodies revealed conformational change of Bax in apoptotic cells stimulated by different apoptotic agents including CD95 ligand (29), TNF (30), protein kinase C inhibitors like staurosporine (STS) (14,31), anti-microtubule agents like Epothilon B analogues (32), and fludarabine/dexamethasone (33). Thus, the induction of Bax conformational change is viewed as a key regulatory step initiating the mitochondrial apoptotic pathway. Upon apoptotic induction, Bax conforma-tional change takes place as an early event and is suggested to be a result of the action of BH3 (Bcl2 homology)-only proapoptotic Bcl2 proteins such as Bim (32), Bid (14), and Bad (34). Although BH3-only pro-apoptotic Bcl2 members induce the conformational change of Bax, anti-apoptotic Bcl2 family proteins like Bcl2, Bcl2 homolog adenovirus-encoded gene products E1B 19K, and Bcl-X L are known to inhibit this process (26).
The precise UV light-mediated pro-apoptotic mechanisms acting upstream of and targeting mitochondria are less well understood. Here we provide evidence that UV light rapidly induces Bax conformational change, which is accompanied by the mitochondrial release of pro-apoptotic factors such as cytochrome c and Smac. We show that the UV light-induced Bax activation pathway can be dissected from that triggered by CD95 ligation or STS. UV light-induced Bax activation is not sensitive to inhibition of either caspases-8 and -3 or cathepsin D. By using two types of ASMase-deficient cells from Niemann Pick disease type A patients, we found that ASMase is indispensable for UV light-induced Bax activation. In these genetically defined ASMase-deficient cells, Bax activation is restored by stable or transient expression of human ASMase as well as by exogenous ceramide. The activation of Bax by ceramide is abrogated in Bcl2-overexpressing cells, indicating that ceramide itself does not exert direct mitochondrial release activity. Finally, Bax activation by ASMase and ceramide was observed in isolated mitochondria but not in cytosolic fractions, localizing the critical step of UV light-induced Bax conformational change at the mitochondrial membrane.
Immunoblotting and Antibodies-Rabbit polyclonal antisera specific for human caspase-9 and -8, human Bax, Bid, and mouse monoclonal antibodies against Bax (clone 6A7 and clone 3) (28), cytochrome c, Smac, and Bcl2 were obtained from BD Biosciences. Goat anti-ASMase antiserum was obtained from Santa Cruz Biotechnology. Horseradish peroxidase conjugates of anti-rabbit and anti-mouse IgG (Bio-Rad) were used as secondary antibodies, and signals were detected by ECL (Amersham Biosciences).
Cell Fractionation and Cell-free Systems-The subcellular fractionation was performed as described (14,31). Briefly, cells were washed two times with ice-cold PBS, resuspended in 1 ml of ice-cold cytosolic buffer (50 mM PIPES, pH 7.0, 50 mM KCl, 2 mM MgCl 2 , 5 mM EGTA, 10 M cytochalasin B, 220 mM mannitol, 68 mM sucrose, and protease inhibitors; complete mixture, Roche Diagnostics), and incubated for 20 min on ice. Cells were then cracked by passing through a 27-gauge needle or by Dounce homogenizer with 20 strokes, and cell breakage was verified microscopically using trypan blue exclusion. Cell lysates were centrifuged at 800 ϫ g for 20 min at 4°C. The resulting pellets contain nuclei, cellular debris, and intact cells, whereas the supernatants contain the cytosol, including mitochondria. Supernatants were centrifuged for an additional 10 min at 800 ϫ g and at 4°C to remove any residual cellular debris. Mitochondria were separated by spinning at 6000 ϫ g for 20 min at 4°C and then washed by pre-chilled cytosolic buffer. The supernatant was additionally centrifuged at 16,000 ϫ g for 20 min at 4°C to remove any residual mitochondria, and the resulting cytosolic fraction was recovered. Mitochondria were then lysed by CHAPS and centrifuged at 14,000 ϫ g at 4°C, and supernatants (mitochondrial fraction) were stored at Ϫ20°C. Whole cell extracts were prepared by lysing 10 7 cells in 1 ml of CHAPS lysis buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 1% CHAPS, protease inhibitors; complete mixture) on ice for 30 min. The crude lysate was then centrifuged at 14,000 ϫ g for 20 min at 4°C, and the supernatants were stored at Ϫ80°C. The purity of mitochondria was monitored by Western blot analysis using antibodies specific for poly(ADP-ribose) polymerase, cytochrome c, cathepsin D, and actin.
Immunoprecipitation-Whole cell extract (10 7 cells ml Ϫ1 ) and mitochondrial or cytosolic fractions derived from 5 ϫ 10 7 cells were used for immunoprecipitation. The KCl concentration of the lysates was adjusted to 150 mM, and all samples were brought to a final volume of 500 l of CHAPS lysis buffer with a 1% final concentration of CHAPS. Samples were rotated for 12 h at 4°C with 6 g of monoclonal anti-Bax 6A7 antibody (Pharmingen). Antigen-antibody complexes were immobilized by rotation for 2 h at 4°C with GammaBind G-Sepharose (Amersham Biosciences). The complexes were pelleted (1 min, 14,000 ϫ g) and the supernatant removed. The complexes were then washed three times with the same buffer used for the immunoprecipitation (CHAPS lysis buffer, 150 mM KCl) and subjected to SDS-PAGE and immunoblotted as described above.
Immunofluorescence-Cells were either exposed to UV light (10 mJ cm (2)Ϫ1 ) or incubated with 0.5 M STS or ceramide (20 M) and washed twice with cold PBS. Cells were then fixed with 3% paraformaldehyde for 20 min, permeabilized with 0.1% saponin in PBS for 10 min, and blocked with 3% bovine serum albumin, 0.005% sodium azide, 4 l ml Ϫ1 gelatin (45% Teleostean gelatin), and 0.1% saponin in PBS for 30 min. For immunostaining, cells were incubated with specific primary antibodies for 1 h, washed with 0.1% saponin in PBS, and then incubated for 30 min with secondary antibodies conjugated with Alexa Fluor 568 and/or Alexa Fluor 488 (Molecular Probes, Leiden, Netherlands). Nuclei were counterstained with Hoechst 33258 (10 g ml Ϫ1 PBS), mounted on glass slides, and examined using a fluorescence microscope.
SMase Assays-Activation of neutral and acidic SMase after UV treatment was measured as described recently (39), with minor modifications. Briefly, cells were lysed in two distinct buffer systems as described, and equal amounts of protein (between 15 and 50 g) were added to 52.25 l of reaction buffer (250 mM sodium acetate, 1 mM EDTA (pH 5.0)) for ASMase measurements. [N-methyl-14 C]Sphingomyelin (56 mCi mmol Ϫ1 ) was added (1.1 Ci ml Ϫ1 final concentration), and the reaction mixtures were incubated at 37°C for 1 h. SMase activity was quantified by counting released [ 14 C]phosphorylcholine using a ␤-counter.
DNA Constructs and RNA Interference-Open reading frames of genes encoding human Bcl2 and Smac were amplified by PCR using Bcl2 primers containing HindIII/ApaI restriction sites and Smac primers containing XhoI/BamHI restriction sites and HeLa cDNA library as template. PCR products were subjected to restriction digestion with the corresponding restriction enzymes, and the digested DNA fragments were ligated to pEGFP-C3 (N-terminal fusion of GFP to Bcl2, GFP-Bcl2) and pEGFP-N3 (C-terminal fusion of GFP to Smac, Smac-GFP), respectively, using T4 ligase (New England Biolabs). The resulting plasmids were transformed into Escherichia coli (one shot, Invitrogen), and transformants were selected, and isolated recombinant DNA plasmids were examined by DNA sequence analysis. Double-stranded (ds) small interfering RNAs (siRNA) of ASMase (ASM-siRNA1, ID 8725; ASM-siRNA2, ID 8630) and control (scrambled) siRNA were obtained from Ambion (Ambion Europe, Huntingdon, UK). Each siRNA (final concentration, 10 nM) was transiently transfected into 1 ϫ 10 6 cells using Lipofectamine (Invitrogen), according to the manufacturer's instructions.

UV Light Induces Bax Conformational Change in HeLa
Cells-The release of cytochrome c from mitochondria has been reported as an early event in the apoptotic process induced by UV light treatment (7,21). To investigate the mechanisms causing cytochrome c release, the mitochondrial apoptotic signaling cascade induced by UV light was analyzed in HeLa cells (Fig. 1). Bax activation/conformational change was detected as an immediate-early event within 0.5 h following UV light ex- HeLa cells were treated with UV light (10 mJ cm (2)Ϫ1 ) and subsequently incubated for the indicated times in culture medium. A, activated and total Bax was detected in total cell lysates by immunoprecipitation (IP) with the conformation-specific antibody 6A7, followed by Western blotting (IB) using anti-human Bax antiserum. Cytochrome c was detected in cytosolic and mitochondrial extracts by Western blotting. B, activated Bax and total Bax were determined in the cytosolic and mitochondrial fractions by immunoprecipitation/Western blotting. C, HeLa cells treated with UV light (10 mJ cm (2)Ϫ1 ) were incubated for 0.5 h and co-immunostained with Alexa Fluor 488-conjugated 6A7 anti-Bax antibody (green) and Alexa Fluor 568-conjugated anti-cytochrome c antibody (red). Nuclei are stained by Hoechst 33258 and appear blue. IF, immunofluorescence. posure (Fig. 1A). Bax conformational change was detected by combined immunoprecipitation/Western blot analysis of Bax using a conformation-specific antibody (clone 6A7 specific for the N-terminal epitope 12-24 of Bax) that recognizes Bax only in its active conformation (28). The total amount of Bax protein remained unchanged, suggesting that the increasing fraction of active Bax protein does not correspond to newly synthesized protein. Concomitantly, pro-apoptotic proteins like cytochrome c were released from mitochondria into the cytosol after 60 min, which was paralleled by diminished levels of cytochrome c detected in mitochondrial fractions (Fig. 1A).
To determine the precise subcellular site of UV light-induced Bax conformational change, we performed combined immunoprecipitation/Western blot analysis of Bax in cytosolic and mitochondrial fractions. In accordance to previous reports (14), Bax is found in both the cytosolic and mitochondrial fractions of untreated HeLa cells (Fig. 1B). After UV light treatment, Bax translocation from cytosol to mitochondria occurred only after 120 min, indicating that Bax conformational change precedes Bax translocation. Activated Bax, however, could be discerned in the mitochondrial but not in cytosolic fractions 15 min after UV light treatment by 6A7 antibody. These results indicated that UV light induces the conformational change of only that fraction of Bax that is already sequestered by mitochondrial membranes. This notion could be confirmed by immunofluorescence-based co-localization studies. 30 min after UV light exposure, HeLa cells co-stained with Alexa 488-conjugated 6A7 anti-Bax antibody (conformation specific) (green) and Alexa 568-conjugated anti-cytochrome c antibody (red) showed colocalization of activated Bax and cytochrome c in a distinct punctated mitochondrial pattern (Fig. 1C). These results confirm that UV light-induced Bax activation only occurs at mitochondrial membranes but not in the cytosol.
UV Light Induces Bax Conformational Change Independent of Caspases-8 and -3-In general, apoptotic stimuli can activate Bax through caspase-dependent and -independent signaling cascades. CD95-induced Bax conformational change is mediated through tBid previously cleaved by caspase-8 (29), whereas STS activates the mitochondrial apoptotic pathway independent of caspase activity (40). In order to examine the role of caspases in UV light-induced Bax activation, Jurkat cells were stimulated with UV light, STS, and anti-CD95. All three stimuli induced caspase-8 processing, Bid cleavage, Bax conformational change, and efflux of pro-apoptotic proteins such as cytochrome c and Smac into the cytosol ( Fig. 2A, lanes  1-4). Pretreatment of Jurkat cells with caspase inhibitors completely blocked both caspase-8 activation and the generation of tBid induced by any of the three stimuli ( Fig. 2A, lanes 5-8). Bax conformational change induced by CD95 was abrogated after caspase inhibition. In contrast, UV light-and STS-induced Bax conformational change remained unaffected ( Fig.  2A, lanes 6 and 7). Likewise, CD95-induced cytochrome c and Smac release was blocked by caspase inhibition, whereas UV light-and STS-induced release of cytochrome c and Smac was not impaired by Z-VAD-FMK and DEVD-CHO ( Fig. 2A, compare lane 8 with lanes 6 and 7). Proper action of caspase inhibitors is demonstrated by the lack of DNA fragmentation in response to all three stimuli (Fig. 2B, compare lanes 1-4 with  lanes 5-8). These results indicate that UV light-induced Bax conformational change is a caspase-8 and -3 independent event, and it is not specific to HeLa cells.
UV Light-induced Acid Sphingomyelinase Activity Is Associated with Bax Conformational Change, Cytochrome c Release, and Cell Death-SMases have been suggested to play a role in apoptosis induced by UV light (15). Specifically, an ASMase has been reported to respond to apoptotic stimulation (18) and therefore is a candidate enzyme that could be involved in UV light-induced apoptosis. As shown in Fig. 3A, UV light treatment indeed leads to increased ASMase activity. To investigate the involvement of ASMase in UV light-induced apoptotic response, RNA interference was used to reduce the expression of ASMase. Compared with ASMase activity in control cells or cells transfected with a control siRNA, basal ASMase activity was reduced by ϳ70% in cells transfected with ASM-siRNAs. In ASM-siRNA2-transfected HeLa cells, UV light treatment led to marginal, if any, enhancement of ASMase activity, which did not even reach basal ASMase activity levels of untreated control cells (Fig. 3A). The down-regulation of ASMase by ASM-siRNA could be confirmed by immunofluorescence analysis. To transfected cells were then exposed to UV light. After 5 min, cellular extracts were prepared and analyzed for ASMase activity. B, HeLa cells were transiently transfected with Smac-GFP visualizing transfected cells and co-transfected with ASM-siRNA or control scrambled-siRNA. ASMase was detected by goat anti-ASMase, primary antibody, and Alexa Fluor 568-conjugated secondary antibody (red). Nuclei were stained with Hoechst 33258. C, HeLa cells transfected with siRNAs as in B were exposed to UV light. After 1 h, activated Bax was detected by conformation-specific primary antibody (6A7) and Alexa Fluor 568-conjugated goat anti-mouse antibody. D, transfected HeLa cells were exposed to UV light and incubated for 4 h. Cytochrome c was detected by mouse anti-cytochrome c antibody and Alexa Fluor 568-conjugated goat anti-mouse secondary antibody. E, HeLa cells were transfected as in B, left untreated, or exposed to UV light and incubated for 4 and 8 h. Apoptosis was microscopically determined by evaluating Ͼ300 cells by the presence of cytosolic Smac-GFP/cytochrome c associated with condensed/fragmented nuclei. The percentage of apoptotic cells was calculated relative to Smac-GFP-positive cells. IF, immunofluorescence. ptotic UV light response. To assess the consequences of ASMase down-regulation for apoptosis, at least 300 transfected HeLa cells were microscopically scrutinized for the presence of activated Bax, release of cytochrome c, and condensed nuclei. As depicted in Fig. 3E, only 40% of ASM-siRNA-transfected HeLa cells showed signs of apoptosis 8 h after UV light exposure, whereas almost all control cells were apoptotic, indicating that down-regulation of ASMase by RNA interference protects HeLa cells from UV light-induced apoptosis.
UV Light Fails to Induce Bax Conformational Change and Cytochrome c Release in ASMase-deficient Cells-To provide more pertinent evidence for the possible involvement of ASMase in UV light-induced Bax activation, we employed genetically determined, ASMase-deficient cells derived from patients suffering inherited Niemann Pick type A disease (NPD), where a defined splicing site mutation in the ASMase gene causes almost complete loss of ASMase enzymatic activity (35). Epstein-Barr virus-transformed B-lymphoblasts (MS1418) (36) as well as fibroblasts from patients suffering NPD type A were used to examine the apoptotic response to UV light. As shown in Fig. 4A, UV light fails to induce Bax conformational change in ASMase-deficient lymphoblasts. Correspondingly, neither release of cytochrome c or Smac nor caspase-9 processing were detected in UV light-treated MS1418 lymphoblasts. In contrast, NPD lymphoblasts stably transfected with an ASMaseexpressing vector (ASM-MS1418) responded to UV light with Bax conformational change, cytochrome c/Smac release, and caspase-9 processing/activation (Fig. 4A, lanes 3 and 4). Similarly, immunofluorescence analysis of NPD lymphoblasts revealed that UV light induces Bax conformational change, as well as cytochrome c release and fragmentation of nuclei (blue) only in ASMase-transfected cells (ASM-MS1418) but not in ASMase-deficient MS1418 cells (Fig. 4, B and C).
To rule out any possible cell type-specific effects secondary to inherited ASMase deficiency, UV light-mediated Bax activation pathways were investigated in ASMase-deficient NPD type A fibroblasts. As shown in Fig. 4D, UV light fails to induce Bax conformational change and release of cytochrome c and Smac in NPD fibroblasts, whereas the STS-induced mitochondrial apoptotic pathway was completely functional. By transient ASMase expression in NPD fibroblasts, the capability of UV light to induce Bax conformational change was restored (Fig. 4E) when NPD fibroblasts were co-transfected with an ASMase expression plasmid along with an eGFP-C3 control vector to visualize transfected cells. UV light-induced Bax conformational change was only observed in transfected NPD fibroblasts. Approximately 80% of UV light-treated fluorescent NPD fibroblasts stained positive for activated Bax (Fig. 4F), indicating that overexpression of ASMase restored the proapoptotic UV light response. Accordingly, cytochrome c release was also restored in fluorescent NPD fibroblasts (green cells) (Fig. 4G). The release of cytochrome c is indicated by the change from a punctuated (mitochondrial) staining pattern to a diffuse (cytosolic) staining. Together, these results provide genetic evidence for the essentiality of ASMase in the UV lightinduced Bax activation pathway and subsequent pro-apoptotic signaling.
Ceramide Induces Bax Conformational Change and Cytochrome c Release in ASMase-deficient Cells-Ceramide, the reaction product of UV light-induced ASMase, is a well known stimulus of apoptotic cell responses. The next set of experiments was designed to show that ceramide serves as an initiator of Bax conformational change and cytochrome c release from mitochondria. Like UV irradiation, ceramide induces Bax conformational change in HeLa cells as shown by immunofluorescence analysis (Fig. 5A). Furthermore, Western blot anal-ysis revealed that C 16 -ceramide in a dose-dependent manner induced Bax conformational change, release of cytochrome c/Smac, and processing of caspase-9 (Fig. 5B). In contrast, dh-C 16 -ceramide at 50 M failed to initiate any apoptotic response in HeLa cells. Notably, ceramide induced Bax conformational change and cytochrome c/Smac release in ASMasedeficient NPD lymphoblasts (Fig. 5C), indicating that the defective UV light response in ASMase-deficient MS1418 cells can be bypassed by the reaction product of ASMase.
Ceramide Induces Bax Conformational Change at Mitochondrial Membranes in Cell-free Systems-As indicated by Bax activation and release of cytochrome c/Smac, UV light-induced pro-apoptotic signaling clearly targets the mitochondria. Therefore, we next addressed the effects of UV light and ceramide on isolated mitochondria. As shown in Fig. 6A, UV light failed to induce Bax conformational change or cytochrome c release in isolated mitochondria. In contrast, incubation of isolated mitochondria with C 16 -ceramide but not dh-C 16 -ceramide led to conformational change of Bax (Fig. 6B). Notably, C 16 -ceramide did not activate Bax in the cytosol. As expected, C 16 -ceramide induced Bax conformational change and release of cytochrome c and Smac from mitochondria isolated from ASMase-deficient MS1418 cells (Fig. 6C). The finding that ceramide failed to induce Bax conformational change in the cytosolic fractions of HeLa cells (Fig. 6B) indicated the requirement of mitochondria for ceramide induction of Bax activation. In addition, ceramide was able to induce Bax conformational change in intact mitochondria but not in membrane-free mitochondrial protein lysates (Fig. 6D, compare lanes 2 and 4), indicating the critical role of mitochondrial membranes or membrane-associated factors for ceramide action on Bax. It is important to note that Bax conformational change was inducible in membrane-free mitochondrial lysates by Triton X-100 (Fig. 6D, lane 6), a nonionic detergent known to induce Bax conformational change (28), indicating the presence of intact, inducible Bax. The observation that C 16 -ceramide but not dh-C 16 -ceramide induced Bax conformational change as well as release of cytochrome c/Smac suggested a specific action of ceramide. To provide a second line of evidence for the specificity of ceramide, Bcl2-overexpressing HeLa cells were used for further investigations. Bcl2 is known to localize in mitochondria and to antagonize Bax activation (26). As shown in Fig. 7A, C 16 -ceramide was unable to induce Bax conformational change or release of cytochrome c/Smac from mitochondria isolated from Bcl2-overexpressing HeLa cells, indicating that ceramide action is not secondary to nonspecific perturbance of the membrane integrity. The amounts of Bax were equal in mitochondrial fractions of HeLa and Bcl2-HeLa cells suggesting that overexpression of Bcl2 does not affect the basal level of Bax associated with mitochondria (Fig. 7A). Similar results were obtained when HeLa cells were transiently transfected with a GFP-Bcl2 fusion protein expressing vector (Fig. 7B). Bcl2 localizes to mitochondria and the endoplasmic reticulum and is known to inhibit Bax and other pro-apoptotic factors interacting with mitochondrial membranes but does not affect signaling events upstream of mitochondria (29,41). Immunofluorescence analysis revealed that both UV light and ceramide induced Bax activation in untransfected cells but failed to do so in GFP-Bcl2-transfected cells. This suggests that ceramide has no direct mitochondrial releasing activity but rather promotes and acts through Bax conformational change. DISCUSSION UV light is known to induce a cell death cascade involving mitochondria, which eventually leads to apoptosis. However, the precise initiating apoptotic mechanisms upstream of mitochondria remained obscure. By using a conformation-specific anti-Bax antibody that recognizes activated Bax, we elucidated the immediate-early events of UV light-induced apoptosis that target mitochondria for release of cytochrome c and other proapoptotic factors. UV-irradiated cells show rapid conformational change of Bax, followed by release of cytochrome c and Smac and eventually by apoptosis. Strikingly, the UV light-induced mitochondrial pro-apoptotic events absolutely require previous activation of ASMase.
Our finding that UV light does not promote Bax activation on isolated mitochondria but rather needs intact cells suggests that an ASMase-dependent signaling cascade upstream of mitochondria is activated by UV light to mediate mitochondrial FIG. 4. Essentiality of ASMase for UV light-induced Bax conformational change and cytochrome c/Smac release. A, NPD lymphoblasts (MS1418) and NPD lymphoblasts stably expressing ASMase (ASM-MS1418) were treated with UV light and incubated for 4 h, and total cell lysates and cytosolic extracts were analyzed by combined immunoprecipitation/Western blotting for activated Bax, cytochrome c, Smac, and caspase-9 processing, respectively. Re-probing for actin ensured equal loading of cytosolic extracts. The asterisk indicates nonspecific bands recognized by polyclonal antibody against caspase-9. B, MS1418 and ASM-MS1418 lymphoblasts were treated with UV light and analyzed after 2 h for activated Bax by monoclonal anti-activated Bax antibody (6A7) using second reagent Alexa Fluor 568-conjugated goat anti-mouse IgG antibody (red). C, cytochrome c was detected 4 h after UV light treatment of MS1418 and ASM-MS1418 lymphoblasts by mouse anti-cytochrome c primary antibody using Alexa Fluor 488-conjugated goat anti-mouse IgG antibody (green) as a second reagent. D, NPD fibroblasts type A (NPD-A) were treated with UV light (10 mJ cm (2)Ϫ1 ) or with STS (0.5 M). After 4 h of incubation total cell lysates and cytosolic extracts were analyzed for activated Bax and cytochrome c/Smac, respectively. E-G, NPD-A fibroblasts were transiently co-transfected with pEGFP-C3/pEFBOS or with pEGFP-C3/pEFBOS-ASMase expression vector (50). 24 h after transfection cells were treated with UV light and incubated for an additional 4 h. Activated Bax (E) was detected by mouse anti-activated Bax antibody (6A7) and Alexa Fluor 568-conjugated goat anti-mouse IgG antibody. The percentage of cells showing Bax activation (F) was microscopically determined by evaluating 300 fluorescent cells for the presence of activated Bax and in addition for cytosolic cytochrome c and condensed/fragmented nuclei. Cytochrome c (G) was detected in transfected NPD fibroblasts by primary mouse anti-cytochrome c and secondary Alexa Fluor 568-conjugated goat anti-mouse IgG antibody (red). IF, immunofluorescence. modification within a very short period of time, i.e. 15-30 min after UV light treatment. As for other apoptotic stimuli, Bax conformational change is known to be initiated by BH3-only pro-apoptotic Bcl2 family members (29 -34). In addition, it has been reported previously that UV irradiation leads to clustering and internalization of cell surface receptors, including death receptors of the TNF receptor family (4). Thus, on theoretical grounds it can be envisioned that UV light-induced clustering of a death domain-containing member of the TNF receptor family could trigger activation of caspase-8 that in turn may cause cytochrome c release through formation of tBid (29,30). However, unlike CD95, UV light-induced Bax conformational change in the presence of caspase inhibitors as well as in the absence of tBid, indicating that UV light-induced Bax activation involves a caspase-8-and tBid-independent mechanism. Western blot analysis of other BH3-only pro-apoptotic Bcl2 family members such as Bad, Bim, and Noxa did not indicate any interactions with Bax or mitochondria within 2 h after UV light treatment (data not shown). In addition, examination by specific inhibitors of other established UV light signaling pathways, including reactive oxygen species production or c-Jun N-terminal protein kinase activation, excluded FIG. 6. Ceramide induces Bax conformational change in isolated mitochondria. A, HeLa cells or isolated mitochondria were exposed to UV light, and mitochondria were analyzed for activated Bax and cytochrome c release. B, mitochondria isolated from HeLa cells were incubated for 1 h at 30°C with 2 M C 16ceramide or 2 M dh-C 16 -ceramide. Mitochondria were pelleted, lysed by CHAPS, and analyzed for activated Bax. Supernatants were analyzed for released cytochrome c and Smac and activated Bax. C, mitochondria isolated from MS1418 and ASM-MS1418 lymphoblasts (5 ϫ 10 7 ) were incubated for 1 h at 30°C with 2 M C 16 -ceramide. Mitochondria were pelleted, lysed by CHAPS, and analyzed for activated Bax. Supernatants were analyzed for released cytochrome c and Smac. D, cytosolic extracts of HeLa cells and isolated mitochondria as well as membrane-free mitochondrial (CHAPS) lysates were prepared and incubated for 1 h at 30°C with C 16 -ceramide or for 1 h on ice with Triton X-100 (1%) and analyzed for activated and total Bax. IP, immunoprecipitate; IB, immunoblot.
FIG. 5. Ceramide induces Bax conformational change and cytochrome c release. A, HeLa cells were treated with UV light (10 mJ cm (2)Ϫ1 ) or with C 16 -ceramide (50 M) and incubated for 4 h. Activated Bax was detected by monoclonal 6A7 antibody. B, HeLa cells were treated and incubated with increasing concentrations of C 16 -ceramide or dh-C 16ceramide for 4 h. Total cell lysates and cytosolic extracts were analyzed for activated Bax and cytochrome c, Smac, and caspase-9 processing, respectively. Reprobing for actin ensured equal loading of cytosolic extracts. C, MS1418 lymphoblasts were treated with C 16 -ceramide (50 M) for 4 h. Activated Bax and cytochrome c were detected by specific primary antibodies and Alexa Fluor 568-(red) or Alexa Fluor 488-conjugated (green) secondary goat anti-mouse IgG antibodies, respectively. IF, immunofluorescence. the involvement of these pathways in UV light-induced mitochondrial Bax activation. 2 Our finding that ASMase is activated by UV light is in line with previous studies indicating that exposure of cells to UV light activates SMases (15,42), which is followed by a rapid increase in intracellular ceramide significantly above basal levels of untreated cells. Although none of these reports suggested SMase to play a role in UV light-induced apoptosis at the level of mitochondria, we now provide evidence for the essentiality of SMases in UV light-induced Bax activation. The idea that ASMase may play a critical role in UV light-induced Bax activation was corroborated by use of siRNAs targeting ASMase expression in HeLa cells and genetically defined ASMase-deficient cells derived from NPD patients. Whereas STS was able to induce Bax conformational change and cytochrome c release in NPD fibroblasts, UV light completely failed to promote Bax conformational change or any other apoptotic mitochondrial event. Similarly, down-regulation of ASMase expression in HeLa cells by RNA interference reproduced a defective mitochondrial apoptotic pathway, which was observed with cells derived from patients with inherited ASMase deficiency. Transient as well as stable expression of functional ASMase in NPD fibroblasts or NPD lymphoblasts, respectively, restored the capability of UV light to induce Bax conformational change, cytochrome c release, and apoptosis. These data provide genetic evidence for the essential role of ASMase in UV light-triggered apoptotic signaling pathways targeting the mitochondria. In addition, pepstatin A, an inhibitor of cathepsin D (40), did not abrogate UV light-induced cell death (data not shown). As shown by others (40), pepstatin A did inhibit STSinduced Bax activation and cytochrome c release, indicating that STS activates apoptotic signaling mechanisms distinct from those induced by UV light.
How can ASMase activity be linked to pro-apoptotic mitochondrial events? Two forms of ASMase have been described, one showing an endo-lysosomal localization and the other secretory form is targeted to the plasma membrane (43). The secretory form of ASMase has been implicated in stress-associated cell death signaling by producing ceramide in the plasma membrane (44). However, a translocation of ASMase to mitochondria has never been reported. This raises the question of how ceramide produced at the plasma membrane or within the endo-lysosomal compartment may trigger apoptotic mitochondrial events. It has been shown earlier by Lipsky and Pagano (45) that exogenous ceramide exposed to intact cells rapidly distributes from the plasma membrane to different organelles, including mitochondria. Indeed, ceramide can be intracellularly transported by lipid flux routes such as the endocytosis of plasma membrane lipids occurring by vesicle-mediated mechanisms (46) or by the mitochondria-associated endoplasmic reticulum subcompartment (MAM fraction) (47), which provides a network for targeting of lipids. Furthermore, a ceramide transfer protein (CERT) has been described recently (48) that is able to propel ceramide transport between different organelles. These observations suggest that ceramide produced in a specific subcellular site may end up in the mitochondrial membrane through vesicle-mediated mechanisms. This notion is supported by the recent finding that accumulation of ceramide can be detected in mitochondrial membranes after UV light-induced SM hydrolysis (42). The idea that ceramide may rather act directly on mitochondrial membranes to promote Bax activation was addressed by adding synthetic ceramides to either intact cells or isolated mitochondria. Treatment of HeLa cells with ceramide resulted in Bax conformational change, which was not observed with HeLa cells overexpressing Bcl2 (Fig. 5). Strikingly, addition of synthetic C 16 -ceramide to isolated mitochondria sufficed to stimulate Bax conformational 2 H. Kashkar and M. Krönke, unpublished observations. FIG. 7. Bcl2 controls ceramide action on mitochondria. A, mitochondrial and cytosolic fractions isolated from HeLa and Bcl2-HeLa cells (5 ϫ 10 7 ) were incubated for 1 h at 30°C with 2 M ceramide analogues. Mitochondria were pelleted and lysed by CHAPS, and Bax and Bcl2 were detected in mitochondrial fractions by specific antibodies. Cytochrome c and Smac were detected in supernatants by Western blot analysis using specific antibodies. B, HeLa cells were transiently transfected with GFP-Bcl2 expression vector and incubated for 24 h. Cells were treated with UV light or C 16 -ceramide and incubated for 4 h. Activated Bax was detected by primary monoclonal antibody (6A7) and Alexa Fluor 568-conjugated secondary goat anti-mouse IgG antibody (red). change and subsequent cytochrome c/Smac release without any additional cytosolic component (Fig. 6B). This finding suggests but does not formally prove that ceramide itself rather than a derivative thereof causes alterations of the mitochondrial membrane that promote the activation of Bax. Ceramide-induced Bax conformational change and cytochrome c/Smac release are completely abrogated in mitochondria from Bcl2-HeLa cells, excluding the possibility that ceramide action on Bax and cytochrome c release is secondary to the general perturbance of membrane integrity.
Ceramide was found to induce Bax conformational change in isolated mitochondrial fractions but not in cytosolic fractions or mitochondrial protein lysates, strongly suggesting the need of the mitochondrial membrane to support Bax action (Fig. 6B). Our findings are consistent with a previous report (49) that suggested Bax conformational change to be the result of Baxlipid membrane cooperation. Ceramide might modulate the function or conformation of Bax, for example by generating lipid microdomains in the outer mitochondrial membrane and/or by modifying the interaction of other proteins possibly masking the immunoreactive residues of activated Bax. Ceramide-induced remodeling of the membrane by exposing breaches can also allow transient interactions to occur with the lipid bilayer that facilitate Bax conformational change in a similar fashion to liposomes containing cardiolipin (49).