Lifeguard Inhibits Fas Ligand-mediated Endoplasmic Reticulum-Calcium Release Mandatory for Apoptosis in Type II Apoptotic Cells*

Death receptors are members of the tumor necrosis factor receptor superfamily involved in the extrinsic apoptotic pathway. Lifeguard (LFG) is a death receptor antagonist mainly expressed in the nervous system that specifically blocks Fas ligand (FasL)-induced apoptosis. To investigate its mechanism of action, we studied its subcellular localization and its interaction with members of the Bcl-2 family proteins. We performed an analysis of LFG subcellular localization in murine cortical neurons and found that LFG localizes mainly to the ER and Golgi. We confirmed these results with subcellular fractionation experiments. Moreover, we show by co-immunoprecipitation experiments that LFG interacts with Bcl-XL and Bcl-2, but not with Bax or Bak, and this interaction likely occurs in the endoplasmic reticulum. We further investigated the relationship between LFG and Bcl-XL in the inhibition of apoptosis and found that LFG protects only type II apoptotic cells from FasL-induced death in a Bcl-XL dependent manner. The observation that LFG itself is not located in mitochondria raises the question as to whether LFG in the ER participates in FasL-induced death. Indeed, we investigated the degree of calcium mobilization after FasL stimulation and found that LFG inhibits calcium release from the ER, a process that correlates with LFG blockage of cytochrome c release to the cytosol and caspase activation. On the basis of our observations, we propose that there is a required step in the induction of type II apoptotic cell death that involves calcium mobilization from the ER and that this step is modulated by LFG.


Death receptors are members of the tumor necrosis factor receptor superfamily involved in the extrinsic apoptotic pathway. Lifeguard (LFG) is a death receptor antagonist mainly expressed in the nervous system that specifically blocks Fas ligand (FasL)-induced apoptosis.
To investigate its mechanism of action, we studied its subcellular localization and its interaction with members of the Bcl-2 family proteins. We performed an analysis of LFG subcellular localization in murine cortical neurons and found that LFG localizes mainly to the ER and Golgi. We confirmed these results with subcellular fractionation experiments. Moreover, we show by coimmunoprecipitation experiments that LFG interacts with Bcl-X L and Bcl-2, but not with Bax or Bak, and this interaction likely occurs in the endoplasmic reticulum. We further investigated the relationship between LFG and Bcl-X L in the inhibition of apoptosis and found that LFG protects only type II apoptotic cells from FasLinduced death in a Bcl-X L dependent manner. The observation that LFG itself is not located in mitochondria raises the question as to whether LFG in the ER participates in FasL-induced death. Indeed, we investigated the degree of calcium mobilization after FasL stimulation and found that LFG inhibits calcium release from the ER, a process that correlates with LFG blockage of cytochrome c release to the cytosol and caspase activation. On the basis of our observations, we propose that there is a required step in the induction of type II apoptotic cell death that involves calcium mobilization from the ER and that this step is modulated by LFG.
Death receptors (DRs) 4 and their ligands are widely expressed. They can induce either cell survival (1), differentiation (2), or cell death (3). In the latter context, DR stimulation induces the formation of the death-inducing signaling complex (DISC), which leads to caspase activation and nuclear fragmentation.
Fas is a DR expressed in various tissues, including the CNS (4,5). Cells can be classified by their response to FasL. In type I apoptotic cells, caspase-8 activation is sufficient to induce the activation of executor caspase-3 (6), whereas in type II apoptotic cells, the engagement of these caspases requires a feedback activation loop through the mitochondria (7). Neuronal cells are mainly type II apoptotic cells.
Bcl-2 family members play a key role in regulating cell death (8). The balance between the pro-apoptotic members Bax and Bak and the anti-apoptotic members Bcl-X L and Bcl-2 controls the permeabilization of the mitochondria. Moreover, Bcl-X L and Bcl-2 are also localized at the ER (9, 10), where they participate in regulating calcium signals, which are also involved in apoptosis regulation (11,12).
DR activity can be modulated by a group of proteins called DR antagonists, such as FLIP (13), FAIM-L (Fas apoptotic inhibitory molecule) (14), and LFG. LFG is also known as FAIM2 or transmembrane Bax inhibitory motif (TMBIM) 2 (15). It was discovered in a screening for genes that conferred resistance to Fas-induced apoptosis (16). LFG is highly, although not exclusively, expressed in the nervous system, especially in the hippocampus and the cerebellum (17). Although its location in lipid rafts in the plasma membrane has been reported (18), other studies have described its presence at the Golgi (19). LFG expression is influenced by the PI3K/Akt pathway (20), and the down-regulation of this protein renders cerebellar granule neurons sensitive to Fas-induced apoptosis (21). LFG also has cytoprotective effects in models of ischemia (22). Despite all these findings, the mechanism of action underlying LFG-mediated protection is poorly understood.
Here we performed a study to elucidate the detailed subcellular distribution of LFG. We demonstrate that it localizes mainly at the ER membrane of neurons, whereas no co-localization with mitochondria markers was found. We also studied the relation of LFG with Bcl-2 family members, showing that it interacts with Bcl-X L and Bcl-2 but not with Bax or Bak. In addition, we also demonstrate that LFG overexpression protects type II apoptotic cells, but not type I, from FasL-induced apoptosis and that this protection depends on the endogenous expression of Bcl-X L . Additionally, LFG was found to modulate calcium release from the ER after FasL treatment. Our results suggest a novel function of LFG in the ER relevant to Fas-mediated cell death.

Experimental Procedures
Reagents-Recombinant human FasL expressing plasmid was a generous gift from Pascal Schneider (23). Fluorogenic caspase substrate Ac-DEVD-afc was purchased from Calbiochem/Merck Biosciences. Unless otherwise specified, all biochemical reagents were supplied by Sigma-Aldrich.
For RNAi experiments, constructs were generated in pSUPER.retro.puro (OligoEngine) using specific oligonucleotides targeting Bcl-X L and LFG sequences as follows: shBcl-X L , GATTGCAAGTTGGATGGCC; and shLFG, GGAGAGAA-GGATGCTC. The control scrambled sequence was GGTCC-CTTTCTCTGTAGT. Adaptors to clone the oligonucleotides into the BglII/HindIII sites of pSUPy.retro.puro were added as required. Oligonucleotides were obtained from Sigma-Aldrich. Lentiviral constructs were generated by digesting pSUPER-sh with EcoRI and ClaI to replace the H1 promoter with the H1-shRNA cassette in pLVTHM.
Lentiviral Production-HEK293T cells were seeded at a density of 2.5 ϫ 10 6 cells in 0.1% collagen-coated 100-mm dishes. The following day, cells were transfected with 20 g of pLVTHM-derived constructs, 13 g of pSPAX2, and 7 g of pM2G. The transfection was routinely performed by the calcium phosphate transfection method. Cells were allowed to produce lentiviruses for 48 h. After 48 h, the medium was centrifuged at 1200 ϫ g for 5 min, and the supernatant was filtered using 45-m filters. Lentiviruses were concentrated at 50,000 ϫ g for 90 min and then resuspended in 20 l of PBS containing 1% BSA. Lentiviruses were stored at Ϫ80°C. Biological titers of the viral preparations expressed as a number of transducing units/ml were determined by transducing HEK293T cells in limiting dilutions. After 48 h of incubation, the percentage of GFP-positive cells was counted, and viruses at 5 ϫ 10 8 transducing units/ml were used in the experiments.
Female C57BL/6 mice were killed and manipulated following the experimental protocol approved by the Vall d'Hebron Institutional Review Board. Embryonic cerebral cortices were dissected from mouse embryos at day 16. Cells were counted and resuspended in DMEM with glutamine supplemented with 5% heat-inactivated FBS and 5% heat-inactivated fetal horse serum, 20 units/ml penicillin, and 20 g/ml streptomycin. Cells were seeded in 25 mg/ml poly-L-lysine-coated plates at a density of 1.6 ϫ 10 5 cells/cm 2 . Cell culture plates were kept at 37°C in a humidified incubator with 5%CO 2 /95%air.
Cell Transfection and Infection-HEK293T, HEK293, and SK-N-AS cells were transfected with the desired expression plasmids using Lipofectamine 2000 (Invitrogen), following the manufacturer's instructions.
For lentiviral-based knockdown experiments, SK-N-AS cells were seeded in 60-mm plates at a density of 1 ϫ 10 5 cells/ml. Titrated lentiviruses were added to the medium when seeding, and medium was changed after 24 h. Transduction efficiency was monitored by direct observation of GFP-positive cells.
Western Blot-Cells were harvested and rinsed once with icecold 1ϫ PBS, pH 7.2, and lysed in immunoprecipitation lysis buffer (20 mM Tris, pH 7.4, 140 mM NaCl, 10% glycerol, 2 mM EDTA, 1 mM EGTA, and 1% Triton X-100) supplemented with 1ϫEDTA-free Complete protease inhibitor mixture (Roche). They were then centrifuged at 16,000 ϫ g at 4°C for 30 min, and the supernatants were collected. Protein concentration was quantified by a modified Lowry assay (DC protein assay; Bio-Rad). The cell lysates obtained were resolved by SDS-PAGE and transferred onto PVDF Immobilon-P membranes (Millipore). After blocking with 1ϫ TBS, 0.1% Tween 20 containing 5% nonfat dry milk for 1 h at room temperature, membranes were probed with the appropriate primary antibodies. They were then incubated for 1 h with the appropriate specific peroxidaseconjugated secondary antibody. Membranes were developed using the EZ-ECL chemiluminescence detection kit (Biological Industries).
The following primary antibodies were used: anti-FLAG Immunoprecipitation-Samples were lysed in immunoprecipitation lysis buffer as described above, and lysates were quantified by a modified Lowry assay (DC protein assay; Bio-Rad). 1 mg of total protein was adjusted with immunoprecipitation buffer to achieve a concentration of 1 mg/ml. A total of 20 l of anti-FLAG M2-agarose-coupled antibody was added to each sample and incubated overnight at 4°C in an orbital shaker. Beads were then washed twice with immunoprecipitation buffer and three more times with 1ϫ TBS. They were then eluted for 30 min at 4°C with 80 l of 1ϫ TBS containing 100 ng/ml of 3ϫ FLAG competitor peptide. After a short spin, supernatants were carefully removed, Laemmli buffer was added, and SDS-PAGE was performed.
Endogenous Immunoprecipitation of Bcl-X L -Cerebellum was dissected from mice adult brain and homogenized in immunoprecipitation lysis buffer, and protein concentration was monitored as described above. 1 mg of protein in 1 ml was subjected to immunoprecipitation. Lysates were precleared by 30 min of incubation with 20 l of suspended protein G-agarose conjugated and 0.25 g/l of control mouse IgG. Afterward, 1 g of Bcl-X L antibody (Sigma) or control mouse IgG (Santa Cruz Biotechnology) were added to each sample and incubated 2 h at 4°C in an orbital shaker. Then 20 l of agarose-conjugated protein G suspension was added and incubated overnight in an orbital shaker. Beads were then washed five times in immunoprecipitation buffer. Acidic elution in citrate 0.1 M was performed for 5 min and stopped by 1 M Tris-HCl in a 1:3 proportion. After a short spin, supernatants were carefully removed, Laemmli buffer was added, and SDS-PAGE was performed.
Subcellular Fractionation-Subcellular fractionation was performed as previously described (24). Briefly, adult mouse brain was homogenized in buffer containing 10 mM HEPES, pH 7.4, 2 mM EDTA, 0.32 M sucrose, and protease inhibitor mixture (Roche). Cell homogenates were centrifuged sequentially at 600 ϫ g for 10 min to remove nuclei and unbroken cells, 3000 ϫ g for 10 min to pellet plasma membrane and heavy intracellular membranes, 15,000 ϫ g for 15 min to yield the light membranes, and 100,000 ϫ g for 1 h to yield the microsomal (pellet) and cytosolic (supernatant) fractions. Nuclei were extracted from the 600 ϫ g pellet with centrifugations at 500 ϫ g for 15 min. The supernatants were centrifuged twice at each speed, and the pellets were washed twice by resuspension in homogenization buffer and recentrifugation. Pellets were lysed in SET buffer and quantified by a modified Lowry assay. Laemmli buffer was added, and SDS-PAGE was performed Cytochrome c Release from the Mitochondria-1 ϫ 10 6 SK-N-AS cells were resuspended in a buffer containing 220 mM mannitol, 70 mM sucrose, 50 mM HEPES-KOH, pH 7.2, 10 mM KCl, 5 mM EGTA, 2 mM MgCl 2 , and 0.025% digitonin and kept on ice for 5 min. Lysed cells were centrifuged (16,000 ϫ g; 5 min; 4°C), and the supernatant was retained as the cytosolic fraction. The release of cytochrome c from the mitochondria to the cytosol was assessed by Western blot.
Caspase Activity-After the indicated treatments, cells were rinsed once with PBS and resuspended in lysis buffer containing 20 mM HEPES/NaOH, pH 7.2, 10% sucrose, 150 mM NaCl, 10 mM DTT, 5 mM EDTA, 1% Triton X-100, 0.1% CHAPS, and 1ϫ EDTA-free Complete protease inhibitor mixture (Roche). The lysates were cleared by centrifugation at 16 000 ϫ g at 4°C for 10 min, and supernatants were quantified by the Bradford method (Bio-Rad). Assays were performed in triplicate using 25 g of protein in the same specific lysis buffer supplemented with 50 M of the fluorogenic substrate Ac-DEVD-afc. After incubation for 1 h at 37°C, plates were read in a fluorimeter using a 360-nm (40 nm bandwidth) excitation filter and a 530-nm (25 nm bandwidth) emission filter.
Cell Viability Assays-For apoptotic nuclear morphology, the cells were seeded in 24-well plates and treated as indicated. They were then fixed with 2% paraformaldehyde and stained with 0.05 g/ml of Hoechst 33258 for 30 min at room temperature in a buffer composed of 2% paraformaldehyde and 1% Triton X-100. Condensed or fragmented nuclei were counted as dead cells, as described previously (25). The determination of cell death by chromatin condensation was performed in blind testing, counting at least 1000 cells for each data point, and was repeated at least three times in independent experiments.
Immunofluorescence-Embryonic day 16 cortical neurons at 1 day in vitro were transduced or not with LFG-FLAG-overexpressing lentiviral particles. Seventy-two hours later they were rinsed with PBS at room temperature and fixed in 4% paraformaldehyde/PBS for 30 min at room temperature. Next, they were washed twice with PBS and subsequently permeabilized and blocked with 5% bovine serum albumin and 0.1% Triton X-100 in PBS for 90 min at room temperature. The cells were incubated overnight with the indicated antibodies, rinsed three times with PBS, and incubated with secondary antibodies as indicated for 1 h at room temperature protected from light. Confocal micrographs were obtained using a FluoView1000 spectral confocal microscope.
The following primary antibodies were used: anti-FLAG Image Analysis-Co-localization was quantified in 10 -15 randomly chosen cells from each sample using ImageJ (JaCoP) software to calculate Manders' co-localization coefficients (M1 and M2) and Pearson's correlation coefficient (R r ) (26). Threshold values were determined using the method described by Costes (27), incorporated in the JaCoP software. Only pixels whose red and green intensity values were both above their respective thresholds were considered to be pixels with co-localized probes. M1 and M2, corresponding to channel 1 (LFG) or channel 2 (other markers), were then calculated as the fractions of total fluorescence in the region of interest that occurred in these "co-local" pixels (with a higher value indicating more co-localization). The M1 values shown represent mean percentages of co-localization.
Cytosolic Calcium Measurements-Cells were incubated in prewarmed control buffer (140 mM NaCl, 3.6 mM KCl, 1.2 mM MgSO 4 , 1 mM CaCl 2 , 20 mM HEPES sodium salt, and 5 mM glucose, pH 7.4) in the presence of 5 M fluo-4 acetoxymethyl ester (F-14201; Molecular Probes) for 30 min at 37°C, protected from light. Buffer with the dye was washed, and fresh prewarmed control buffer was added.
The cells were excited at 488 nm using an argon/krypton laser confocal system (Yokogawa CSU10, Nipkow spinning disk) set on an Olympus IX70 (VoxCell Scan; Visitech) with a 60ϫ oil immersion objective lens. Emission was captured by a digital CCD camera (ORCA, Hamamatsu, Japan) and analyzed using VoxCell Scan Software (Visitech).
After measurement, each cell in the captured fields was analyzed separately. Cells with a fold-induction of calcium release Ն 1.2 were considered for further analysis. The cells were normalized at the time point they began to react after Fc-FasL treatment.
Statistical Analysis-All the experiments were repeated at least three times. The values are expressed as means Ϯ S.D.
LFG Is Localized along the Endocytic Pathway-Because the localization of LFG appears to be vesicular, we sought to study the localization of LFG along the endocytic pathway in greater depth. LFG co-localization with markers of proteins of early, late, and recycling endosomes, lysosomes, as well as markers of the ER and trans-Golgi network, was analyzed ( Fig. 2A)  and TGN38 (M1: 65.9 Ϯ 8.2; R r : 0.754 Ϯ 0.059) (Fig. 2B). Altogether, these results indicate that most cellular LFG in neurons is located at the ER and, to a lesser extent, at the Golgi and vesicular membranes.
LFG Interacts with Bcl-X L -LFG is a member of the TMBIM family of proteins (28). Some of these members have the capacity to modulate the activity of Bcl-2 family members (29), such as Bcl-X L and Bcl-2, which are found in mitochondria but also in the ER. We hypothesized that LFG also modulates the activity of Bcl-2 family members in a similar manner. Therefore, we sought to determine LFG interaction with Bcl-2 family proteins. In this regard, we observed that LFG co-immunoprecipitated with Bcl-X L and Bcl-2 but not with Bax or Bak (Fig. 3A).
Because Bcl-X L and LFG are highly expressed in the CNS (30, 31), we focused our attention on the LFG-Bcl-X L relationship. To further characterize this interaction, we constructed truncated forms of LFG (Fig. 3B). The first 101 amino acids of the LFG sequence, corresponding to the N-ter unfolded domain, were subcloned and fused to a 3ϫFLAG tag in a pcDNA3 expression vector to form the 3ϫFLAG-⌬C-LFG construct. The rest of the sequence, corresponding to the transmembrane domain with the C-ter tail, was subcloned along with a FLAG tag and a methionine residue at the start to initiate the open reading frame, forming the ⌬N-LFG-FLAG vector. These constructs and LFG-FLAG were overexpressed, along with 3ϫHA-Bcl-X L . FLAG immunoprecipitation and HA detection revealed that LFG interacts with Bcl-X L through its C-ter region (Fig. 3C). Endogenous co-localization of LFG with Bcl-X L in cortical neurons was also studied by immunofluorescence. We observed LFG co-localization with Bcl-X L ( Fig. 3D; M1: 47.6 Ϯ 2.9; R r : 0.558 Ϯ 0.048), located at the cytosol excluding the nucleus, with a stronger staining in the perinuclear region, thereby suggesting ER co-localization. To further confirm the observation that LFG interacts with Bcl-X L and to rule out possible bias in overexpression experiments, we performed an immunoprecipitation of endogenous Bcl-X L in lysate from adult mice cerebellum, in which both proteins are highly expressed. We observed that LFG is co-immunoprecipitated with Bcl-X L (Fig. 3E). To study in more detail the localization of LFG-Bcl-X L interaction, we performed a subfractionation assay in lysate from adult mice brain. LFG is detected in the heavy and light membrane fractions (Fig. 3F), further confirming our results about LFG localization in the ER and endocytic vesicles membranes. Moreover, Bcl-X L is also detected in the heavy and light membranes fraction. Thus, all these data indicate that LFG interacts with Bcl-X L , and this interaction is likely to occur in the ER membrane.

Although LFG Is Not Localized at the Mitochondria, It Protects Type II Cells from FasL-induced Cell Death-LFG prevents
FasL-induced cell death in neurons and neuronal-like cells among others (16,20), all of them classified as type II apoptotic cells. Nonetheless, there is no direct evidence that LFG protects type I cells against FasL-induced apoptosis. To address this question, we performed a comparative study of LFG inhibition of FasL-induced apoptosis in type I and neuronal-like type II apoptotic cells. To this aim, we overexpressed this protein in parallel with other anti-apoptotic proteins (Bcl-X L and FLIP) in HEK293 and SK-N-AS cells as models of type I and type II apoptotic cells, respectively. SK-N-AS cells are neuroblastomaderived cells that are sensitive to Fas-L-induced apoptosis and behave as type II apoptotic cells as expected by their neuronal origin. Bcl-X L inhibits FasL-induced apoptosis in type II cells but not in type I. FLIP, which competes with caspase-8 for recruitment to the DISC after DR stimulation (32), protects against DR-induced apoptosis in type I and II cells.
Our results show that FasL-induced cell death in SK-N-AS cells, as measured by apoptotic nuclei counting, was inhibited by Bcl-X L and LFG overexpression at 24 and 48 h after treatment ( Fig. 4C and data not shown). Such inhibition was not observed in HEK293 cells, where only FLIP overexpression prevented apoptosis (Fig. 4D). Moreover, LFG overexpression, as well as that of Bcl-X L , decreased caspase-3 activation in SK-N-AS (Fig. 4A) cells but not in HEK293 cells (Fig. 4B). The efficiency of overexpression of the DR antagonists was assessed by Western blot (Fig. 4G). Altogether, these data demonstrate that LFG overexpression protects type II apoptotic cells, but not type I apoptotic cells, against Fas-induced cell death.

Requirement of Endogenous Bcl-X L for LFG Anti-apoptotic Effects-Our findings that LFG protects type II apoptotic cells from Fas-induced cell death and that it interacts with
Bcl-X L suggested that one of these molecules might depend on the other to exert its anti-apoptotic effects. To study this hypothesis, we used SK-N-AS cells, which express both LFG and Bcl-X L endogenously.
Cells were transduced with lentiviral particles carrying LFG, Bcl-X L , or Scrambled shRNA, and after 3 days of transduction they were infected again with overexpression lentiviral particles for LFG or Bcl-X L for 3 additional days. Knockdown efficiency was determined by Western blot (Fig. 5C). LFG and Bcl-X L overexpression reduced caspase 3 activity and the percentage of apoptotic nuclei after Fas-L treatment (Fig. 5, A and B), in consonance with previous results. However, when Bcl-X L was down-regulated, LFG overexpression was unable to protect SK-N-AS cells. This was not the case in the opposite situation; thus Bcl-X L overexpression maintained its anti-apoptotic activity despite endogenous LFG down-regulation (Fig. 5, A and B). On the basis of these observations, we conclude that endogenous levels of Bcl-X L are required for LFG anti-apoptotic action, whereas protection from FasL-induced cell death by Bcl-X L overexpression does not require LFG.
LFG Overexpression Down-regulates Calcium Release from the ER after FasL Stimulation-Calcium transfer from the ER to mitochondria is a crucial step in various forms of cell death (33,34), including FasL-induced apoptosis, as demonstrated in experiments in Jurkat cells (35). Bcl-X L exerts anti-apoptotic effects in the ER by blocking calcium release (36,37). In the same way, BI-1, also known as TMBIM5, modulates calcium release from the ER (29) and has an anti-apoptotic role. Because LFG is also a member of the TMBIM family, and we have demonstrated that it exerts its anti-apoptotic effects in a Bcl-X L -dependent manner, we hypothesized that it may be involved in the regulation of calcium release from the ER after Fas engagement.
After FasL treatment, SK-N-AS cells released calcium from the ER (Fig. 6A), as demonstrated by the lack of calcium release after depletion of the ER calcium content by pretreatment with the sarcoplasmic/endoplasmic reticulum calcium-ATPase inhibitor thapsigargin (Fig. 6A). This is the first report involving calcium mobilization from the ER after FasL stimulation in cells from neuronal lineage. In LFG-overexpressing SK-N-AS cells, calcium release from the ER was significantly decreased (Fig. 6,  A and D), and as expected and previously described (36 -38), Bcl-X L overexpression dramatically abrogated calcium release (Fig. 6, A and D). On the other hand, LFG or Bcl-X L downregulation did not significantly alter calcium release from the ER (Fig. 6B); nevertheless, these cells showed a tendency to reach higher rates of calcium release than the scrambled control. Strikingly, Bcl-X L overexpression in shLFG-transduced cells was unable to inhibit calcium release from the ER (Fig. 6C). This indicates that Bcl-X L effects on the ER are dependent on LFG expression, and because these cells are resistant to FasLinduced apoptosis (Fig. 5, A and B), it suggests that the Bcl-X L anti-apoptotic effect in the mitochondria is enough to protect from FasL-induced apoptosis. LFG overexpression in shBcl-X L -transduced cells was unable to prevent calcium release from the ER (Fig. 6C), in accordance with previous results that showed a requirement of Bcl-X L for LFG anti-apoptotic effects. To rule out that the observed effects are not due to changes in ER luminal calcium concentration, we compared the baseline levels of each condition. Baseline levels of intracellular calcium fluorescence were similar in all groups of treatment (data not shown). Our data are in agreement with previously reported studies that point out Bcl-X L regulation of ER calcium efflux, rather than decreasing intraluminal calcium content, as the mechanism involved in its anti-apoptotic role in the ER (37,38). . LFG interacts with Bcl-X L . A, HEK293T cells were transfected with LFG-FLAG expression plasmid and co-transfected with 3ϫHA-Bcl-X L , Bcl-2, 3ϫHA-Bax, or 3ϫHA-Bad as indicated. FLAG-tagged proteins were immunoprecipitated from whole cell extracts with anti-FLAG resin, followed by immunoblotting HA, Bcl-2, and FLAG antibodies. B, schematic representation of LFG-FLAG, ⌬N-LFG-FLAG, and 3ϫFLAG-⌬C-LFG truncated forms. Green boxes represent putative transmembrane regions, and red boxes the FLAG tag. A fragment of the primary sequence of LFG is shown highlighting the specific amino acid where ⌬N-LFG-FLAG ends and 3ϫFLAG-⌬C-LFG starts. C, HEK293T cells were transfected as indicated, and FLAG-tagged proteins were immunoprecipitated from whole cell extracts with anti-FLAG resin, followed by immunoblotting HA and FLAG antibodies. D, representative confocal images of cortical neurons. Cortical neurons were fixed, permeabilized, and immunostained with anti-LFG (green), anti-Bcl-X L (red), and Hoechst (purple). The lower right panel shows a merge of all channels. Scale bar, 10 m. The values of all Manders' co-localization coefficients and Pearson correlation coefficients are detailed under "Results." E, adult mice cerebellum was homogenized and lysed. Endogenous Bcl-xL was immunoprecipitated, and LFG interaction was assessed by immunoblotting. F, adult mice brain was homogenized and lysed. Subfractionation assay was performed, and the different fractions were analyzed by Western blot. LFG and Bcl-xL presence was assessed in each fraction by immunoblotting, and correct purification of each fraction was assessed by immunoblotting with specific markers: anti-Rab5 (early endocytic vesicles), anti-calnexin (ER marker), anti-cytC (mitochondria marker), anti-GluR2 (plasma membrane marker), anti-H3 (nucleus marker), and anti-GADPH (cytosol marker). N, nucleus; HM, heavy membranes; LM, light membranes; Mi, microsomal fraction; Cy, cytosolic fraction; IP, immunoprecipitation.
In summary, LFG and Bcl-X L overexpressing cells showed a marked reduction in calcium release after FasL treatment when compared with sh scrambled-transduced cells (Fig. 6A), whereas down-regulation of LFG or Bcl-X L did not induce a significant increase (Fig. 6B). Overexpression of LFG in shBcl-X L -transduced cells showed no effect, in the same way as Bcl-X L overexpression in shLFG-transduced cells (Fig. 6C). Taken together, these results suggest that LFG anti-apoptotic effects are mediated by inhibition of calcium release from the ER and that Bcl-X L is essential for this step. Notably, Bcl-X L effects on calcium release also appear to be dependent on LFG expression.
FasL-induced Calcium Release from the ER Is Mediated by Caspase-8 Activation in Neuron-like Cells-It has been reported that ER calcium release is mediated by caspase-8 activation and the cleavage of its substrate Bap31 in Jurkat cells after Fas-L stimulation (35,39). We decided to explore this pathway in our neuronal-like model SK-NA-S cells. Effectively, we observed that ER calcium release after FasL treatment was dependent on caspase-8 activation, because no calcium release from the ER was detected after FasL stimulation when cells were pretreated with the caspase-8 inhibitor IETD (Fig. 7A). We also observed the cleavage of procaspase-8 in our conditions after FasL treatment (Fig. 7B). More importantly, we were able to detect the cleavage of Bap31 in its p20 fragment, which in turn participates in the cytochrome c release from the mitochondria by enhancing calcium signals (40) (Fig. 7C). The observed cleavage was caspase-8-mediated, because its inhibition by IETD or the competitive antagonist, FLIP, resulted in no cleavage of Bap31. These data indicate that in neuronal-like cells, caspase-8 activation after FasL stimulation is necessary for ER calcium release.
LFG Overexpression Inhibits Cytochrome c Release from the Mitochondria after FasL Stimulation-Because calcium mobilization from the ER to mitochondria has been described as an important step in the inducement of mitochondria permeabilization (35), we sought to study the effects of LFG in this process. After 6 h of FasL treatment, cytochrome c could be detected in the cytosolic fraction of sh scrambled-transduced cells (Fig.  7D). As has already been described, Bcl-X L overexpression inhibited mitochondria permeabilization, as it does FLIP. According to the results for calcium release from the ER, LFGtransduced cells showed a marked reduction of cytochrome c release to the cytosol after Fas engagement, suggesting that LFG reduction on ER calcium release may be the mechanism through which LFG performs its anti-apoptotic effect.
On the other hand, down-regulation of Bcl-X L showed no difference in the amount of cytochrome c release compared with sh scrambled control cells (Fig. 7D), and LFG overexpres-sion in shBcl-X L -transduced cells was unable to inhibit mitochondria permeabilization, further confirming the dependence of LFG effects on Bcl-X L endogenous expression. In shLFGtransduced cells, cytochrome c was detected in the cytosol as in sh scrambled-transduced cells, and Bcl-X L overexpression in these cells blocked cytochrome c release, suggesting that Bcl-X L inhibitory function on mitochondria permeabilization is able to protect from cell death despite the calcium release from the ER. In summary, LFG overexpression reduces cytochrome c release from the mitochondria. However, LFG failed to inhibit this processes when Bcl-X L was down-regulated, highlighting the importance of the association between both proteins on LFG anti-apoptotic function.

Discussion
The protein LFG was discovered during screening for genes that confer resistance to Fas-induced cell death (16). Its predicted structure contains seven hypothetical transmembrane domains and one unfolded region at the N-ter of the protein.
However, a functional role of the structural domains of LFG has not yet been described. To date, the mechanism by which LFG protects against Fas-mediated apoptosis is still unknown. It has been proposed that this protein acts upstream of caspase-8 activation because it is able to inhibit the activity of this caspase. The finding that LFG interacts with Fas receptor in the lipid raft has revealed this union as the most plausible mechanism of protection (17). However, no differences in the formation of DISC or in the agonist binding to the receptor have been found when LFG is overexpressed (20).
Studies addressing LFG localization show that it is expressed in postsynaptic sites and dendrites (18). However, a clear punctate pattern not only in dendrites but also in the cell body is detectable when using LFG antibodies for immunofluorescence. We observed that LFG is not localized at the mitochondria. Instead, we demonstrate that LFG localizes mainly in the ER and Golgi apparatus. In agreement with this finding, a study addressing other TMBIM family members reported that LFG also localizes at the membranes of the trans-Golgi complex (19,28).
LFG is highly homologous to BI-1 (15). BI-1 blocks the intrinsic pathway that protects cells from DNA-damaging agents and ER stress inducers (41,42). It interacts with Bcl-X L in the ER and regulates calcium signaling (29). Fas-mediated apoptosis shares some features with intrinsic death pathways, such as ER stress (43,44), and promotes a rise in intracellular calcium (45,46), which mediates some apoptotic features attributed to FasL-induced cell death (35). The cross-talk between the ER and mitochondria and the calcium efflux between them are Scale bar, 10 m. G, immunoblot analysis was performed to assess LFG, Bcl-X L and FLIP overexpression in HEK293 cells. *, short exposure; **, long exposure. Naphtol Blue staining was performed to confirm equal loading. FIGURE 5. Endogenous Bcl-X L is essential for LFG protection against FasL-induced cell death. A, SK-N-AS cells were transduced with lentiviral particles carrying shRNA against LFG, Bcl-X L , or scrambled (Scr) for 3 days and then transduced with LFG-or Bcl-X L -overexpressing lentiviral particles for 3 additional days as indicated. The cells were treated with Fc-FasL (100 ng/ml) or left untreated (UT) for 6 h. DEVDasa activity was assessed (one-way ANOVA test). ***, p Յ 0.001. B, SK-N-AS cells were transduced as in A. The cells were treated with Fc-FasL (100 ng/ml) or left untreated for 24 h. The percentage of apoptosis was assessed by counting cells with apoptotic nuclei (one-way ANOVA test). ***, p Յ 0.001; *, p Յ 0.05. C, immunoblot analysis was performed to assess LFG and Bcl-X L down-regulation in SK-N-AS cells. Naphtol Blue staining was performed to confirm equal loading. important steps in Fas apoptotic signaling, at least in the immune system (47). Nonetheless, nothing is known in other systems. Thus, we propose that LFG exerts its anti-apoptotic effects by modulating mitochondria permeabilization via an ER-dependent mechanism in the nervous system. LFG inhibits calcium release from the ER after FasL treatment. Accordingly, a marked reduction in cytochrome c release, and caspase-8 activation is observed in LFG-transduced cells. However, we observed that LFG effects are Bcl-X L -dependent. Bcl-X L overexpression also inhibits calcium release from the ER, and cytochrome c release. When LFG is down-regulated, Bcl-X L is still able to inhibit mitochondria permeabilization and caspases activation; nevertheless, calcium release from the ER remains unaffected.
On the basis of these observations, we suggest that LFG may play a role in the regulation of calcium release from the ER after an apoptotic stimulus. Blockade of the calcium efflux from ER to the mitochondria would be a limiting step to inhibit the cytochrome c release and subsequent activation of caspase-3, thereby preventing apoptosis (Fig. 8).
Taking into the account that LFG did not influence ligand binding to Fas-receptor or DISC formation (16), it was suggested that LFG might act at the level of caspase-8 activation. Nevertheless, we report data that contradicts this view. First, we report that LFG is only able to protect type II apoptotic cells, whereas it fails to inhibit apoptosis in type I apoptotic cells. If LFG was able to inhibit caspase-8 activation directly, it should be able to inhibit type-I Fas-induced apoptosis, in the same way as other caspase-8 inhibitors, such as FLIP. In addition, LFG lose its protective effects when Bcl-X L is down-regulated. Because our data suggest that the interaction between both proteins is essential for LFG function, is highly unlikely that LFG may exert its main effect at the receptor level, where Bcl-X L is not localized. If LFG effects were located at the DISC level, effects on blockade of calcium release from the ER and prevention of cytochrome c release from mitochondria should be observed despite Bcl-X L down-regulation.
Bcl-X L plays a central role in the regulation of FasL-induced apoptosis in type II cells. It is located at both the mitochondria and the ER (48,49), and its anti-apoptotic function through FIGURE 6. LFG inhibits calcium release from the ER after Fas stimulation. A, SK-N-AS cells were transduced with empty plasmid and LFG-and Bcl-X Loverexpressing lentiviral particles for 3 days or treated with thapsigargin (Tg; 1 M) for 30 min prior to Fc-FasL (200 ng/ml) treatment. Intracellular calcium mobilization was assessed as detailed under "Experimental Procedures." B, SK-N-AS cells were transduced with lentiviral particles carrying shRNA against LFG, Bcl-X L , or scrambled (Scr) for 3 days. Cells were treated with Fc-FasL (200 ng/ml), and intracellular calcium mobilization was assessed. C, SK-N-AS cells were transduced with lentiviral particles carrying shRNA against LFG, Bcl-X L , or scrambled for 3 days and then transduced with LFG-or Bcl-X L -overexpressing lentiviral particles for 3 additional days as indicated. The cells were treated with Fc-FasL (200 ng/ml), and intracellular calcium mobilization was assessed. D, maximum calcium release of SK-N-AS transduced cells, calculated as the average of maximum calcium release of at least 10 individual cells per group (one-way ANOVA test). *, p Յ 0.05. The cells were treated with Fc-FasL (100 ng/ml) or left untreated (UT) for 6 h, and then Western blot was used to assess the cleavage of procaspase-8 (B) or Bap31 (C). Tubulin was used as loading control. D, SK-N-AS cells were transduced with lentiviral particles carrying shRNA against LFG, Bcl-X L , or scrambled (Scr) for 3 days and then transduced with LFG-or Bcl-X L -overexpressing lentiviral particles for 3 additional days as indicated. Cells were treated with Fc-FasL (100 ng/ml) or left untreated for 4 h. Cytochrome c (Cyt C) release to the cytosol was assessed by Western blot. Naphtol Blue staining was used as a loading control. FIGURE 8. Schematic representation of Fas apoptotic signaling pathway and the hypothetic steps inhibited by LFG. After Fas stimulation, caspase-8 cleaves Bap31 and tBid. p20 fragment induces Ca 2ϩ efflux from the ER to the mitochondria, which in conjunction with tBid induces mitochondrial permeabilization and cytochrome c release. Although Bcl-X L is able to inhibit ER calcium release and mitochondria permeabilization directly, LFG is only able to block the ER step.
both organelles is well documented. Strikingly, Bcl-X L effects on ER calcium signaling appear to be dependent on LFG endogenous expression.
In summary, our results reveal a hitherto undescribed step in the extrinsic apoptotic signaling pathway in cells from neuronal lineage. Calcium mobilization from the ER has been shown to be relevant in FasL apoptotic signaling. We demonstrate that LFG modulates calcium release from the ER after FasL stimulation and inhibits FasL-induced apoptosis in neuron-like cells that are type II apoptotic cells regarding the way they subside apoptosis. On the basis of our observations, we propose that LFG protects against FasL-induced apoptosis by modulating calcium release from the ER.