Nucling recruits Apaf-1/pro-caspase-9 complex for the induction of stress-induced apoptosis.

Nucling is a novel protein isolated from murine embryonal carcinoma cells with an up-regulated expression during cardiac muscle differentiation. We show here that Nucling was up-regulated by proapoptotic stimuli and important for the induction of apoptosis after cytotoxic stress. We further demonstrated that overexpressed Nucling was able to induce apoptosis. In Nucling-deficient cells, the expression levels of Apaf-1 and cytochrome c, which are the major components of an apoptosis-promoting complex named apoptosome, were both down-regulated under cellular stress. A deficiency of Nucling also conferred resistance to apoptotic stress on the cell. After UV irradiation, Nucling was shown to reside in an Apaf-1/pro-caspase-9 complex, suggesting that Nucling might be a key molecule for the formation and maintenance of this complex. Nucling induced translocation of Apaf-1 to the nucleus, thereby distributing the Nucling/Apaf-1/pro-caspase-9 complex to the nuclear fraction. These findings suggest that Nucling recruits and transports the apoptosome complex during stress-induced apoptosis.

cal agents, and oxidative stress can also induce apoptosis.
These proapoptotic stimuli bring about organellar stress, affecting the nucleus, peroxisome, lysosome, or Golgi apparatus. Most of the apoptosis-inducing signals from these organelles converge at mitochondria. Mitochondria release proteins that promote cell death after cellular stress (4). One of these proteins is cytochrome c, which forms a complex with the cytoplasmic protein Apaf-1 and pro-caspase-9, leading to the activation of caspase-9. Caspase-9, in turn, activates caspase-3, the protease that cleaves the majority of caspase substrates during apoptosis. Mitochondria also release apoptosis-inducing factor (AIF) 1 and endonuclease G, which appear to kill cells independently of caspases. Therefore, mitochondria are thought to be a central regulatory element in stress-induced apoptosis (5).
Nucling was originally isolated from murine embryonal carcinoma cells as a protein, the expression of which was upregulated during cardiac muscle differentiation (6). A bovine homolog of Nucling, named ␤CAP73, was isolated and characterized as a novel regulator of ␤-actin assembly (7,8). Recently, we reported that Nucling down-regulates expression of the antiapoptotic molecule, galectin-3, through interference with nuclear factor-B (NF-B) signaling (9). It has been reported that galectin-3 is translocated to the perinuclear membranes after exposure to a variety of apoptotic stimuli and becomes abundant in mitochondria, where it prevents mitochondrial damage and cytochrome c release (10). These findings led us to investigate whether Nucling is involved in any of the apoptotic signaling pathways. To address this issue, we performed molecular biological analyses, including gene knock-out experiments.
In this report, we show that Nucling is a regulatory molecule for stress-induced apoptosis, which interacts with the Apaf-1/ pro-caspase-9 complex, thereby acting as a stabilizer for the apoptosome. Nucling was shown to be able to induce apoptosis in mammalian cells and could promote the caspase cascade. Nucling Ϫ/Ϫ cells showed resistance to cellular stress. Up-regulation of Apaf-1, release of cytochrome c, and activation of caspase-9 induced by cytotoxic stress were not observed in Nucling Ϫ/Ϫ cells. Two-dimensional native/denaturing-PAGE analysis revealed that Nucling assembles with the Apaf-1/pro-caspase-9 complex in vivo. These findings suggest that Nucling acts as a regulatory factor for stress-induced apoptosis, sustaining the expression level of Apaf-1 by interacting with the Apaf-1/pro-caspase-9 complex. Moreover, this assemblage composed of Nucling, Apaf-1, and caspase-9 was observed in both cytosol and nuclear fractions. Furthermore, we confirmed that Nucling was required for the translocation of Apaf-1 to the nucleus after proapoptotic stress, suggesting that Nucling is a transporter of the Apaf-1/caspase-9 complex to the nucleus.

EXPERIMENTAL PROCEDURES
Cell Lines and Culture Conditions-COS7 cells, HeLa cells, and mouse embryonal fibroblast (MEF) cells were used in this study. Cells were maintained in DMEM with 100 units/ml of penicillin, 100 g/ml of streptomycin, 5 M mercaptoethanol, and 10% (v/v) fetal calf serum. They were cultured in 10 ml of medium in 95-mm plastic tissue culture plates at 37°C in an atmosphere of 5% CO 2 /95% air in a humidified incubator. For routine propagation, cultures were split, and the growth medium was replenished every three to four days.
Construction of Expression Vectors-The cDNA fragment of Nucling was tagged with the sequence encoding the Flag peptide using the vector pFlag-CMV2 (Kodak).
Transfection of COS7 Cells with Plasmid DNA-Plasmid DNA was transiently transfected into COS7 cells using the Effecten reagent (Qiagen, Hilden, Germany) according to the manufacturer's instructions. For immunofluorescence microscopic analysis, a sample of COS7 cells was transferred to an eight-well chamber slide for tissue culture and incubated at 37°C overnight before transfection.
Confocal Immunofluorescence Microscopic Analysis-After 24 h of transfection with Flag-Nucling expression plasmid vector or control vector (pFlag-BAP), cells were washed with PBS and incubated with propidium iodine (PI; 3,8-diamino-5-(3-(diethyl-methylamino)propyl)-6phenyl phenanthridinium diiodide; Sigma) for 20 min at room temperature. After nuclear staining of dead cells with PI, cells were washed with PBS three times and then fixed with 4% paraformaldehyde in PBS for 10 min at room temperature. For blocking, the cells were incubated with 3% bovine serum albumin in PBS for 1 h. After five washes with PBS, cells were incubated with anti-Flag M5 antibody for 2 h at room temperature. After five washes with PBS, the slides were incubated with FITC-conjugated secondary antibodies against mouse IgG (Amersham Biosciences) for 1 h. After five washes with PBS, the slides were mounted with anti-fade solution, sealed, and examined using a confocal laser scanning microscope and software (Leica TCS NT, Heidelberg, Germany).
Cell Death Assays-PI staining assays were performed using standard protocols (Oncor). TdT-mediated dUTP-biotin nick end labeling (TUNEL) assays were performed using the In Situ Cell Death Detection kit, Fluorescein (Roche) according to the manufacturer's instructions. Cell death was measured by lactate dehydrogenase (LDH) release assay or trypan blue exclusion assay. For the LDH release assay, the culture medium was centrifuged to remove detached cells, and the supernatants were used to determine the quantitative LDH activity with a Cytotoxicity Detection kit (Roche) according to the manufacturer's instructions. For total LDH activity, cell lysate was prepared from control cells in the culture by adding a 1/100 volume of 10% Triton X-100 into the culture medium, incubated at room temperature for 15 min, and then centrifuged at 800 ϫ g for 10 min. For the trypan blue exclusion assay, detached COS7 cells transfected with pFlag-Nucling, or pEGFP-C1 in culture medium, were concentrated by centrifugation (800 ϫ g for 5 min) and resuspended in 100 l of cold PBS before being incubated with 0.4% trypan blue solution (Sigma) for 10 min. More than 100 cells were scored on a hemocytometer.
Western Blot Analysis-Cellular extracts were prepared as described previously (6). Antibodies reactive to cytochrome c (BD Biosciences), Apaf-1 (Chemicon), caspase-9 (Cell Signaling), caspase-3 (BD Biosciences), ␤-actin (Sigma-Aldrich), and the middle portion of Nucling (Nucl.mid) (6) were used in this study. Western blot analysis was carried out according to standard procedures using an ECL detection kit (Amersham Biosciences) or AP detection kit (Roche). Quantification was performed by comparing densitometric scanning readings using NIH-Image v1.63 software; numbers (arbitrary units) represent values corrected for loading with the data reprobed by ␤-actin.
Northern Blot Analysis-After treatment, COS7 cells were washed with PBS and pelleted by centrifugation. Total RNA was isolated from cells using ISOGEN as described by the manufacturer (Nippon Gene, Toyama, Japan). RNA was fractionated on 2.2 M formaldehyde/1.2% agarose gels and transferred overnight onto Hybond N nylon membranes (Amersham Biosciences) in 10ϫ SSC. The RNA was cross-linked to the membrane using a UV cross-linker (Amersham Biosciences) before hybridization. A specific probe was generated by labeling the cloned cDNA fragment of full-length Nucling with [␣-32 P]dCTP (NEN, Boston, MA) using Ready-To-Go DNA Labeling Beads (-dCTP) (Amersham Biosciences). After overnight hybridization at 42°C, the filters were washed once in 2ϫ SSC for 10 min (23°C) and twice in 0.1ϫ SSC for 15 min (68°C), covered in plastic wrap, and exposed to Kodak X-Omat AP film at Ϫ70°C for 3-24 h.
Generation of Nucling Ϫ/Ϫ Mice-The genomic DNA containing the Nucling gene was isolated from a 129/Sv mouse genomic library. The targeting vector was constructed by inserting a PGK-neo-poly(A) cassette into a HindIII site of the exon containing the leucine zipper motif region (6) of the Nucling gene. The targeting vector thus contained 0.5and 5.4-kb regions of homology in the 5Ј and 3Ј region of the neomycinresistance marker, respectively. The maintenance, transfection, and selection of embryonic stem cells were performed as described (11). The mutant embryonic stem cells were microinjected into C57BL/6 blastocysts, and the resulting male chimeras were mated with female C57BL/6 mice. Heterozygous offspring were intercrossed to produce homozygous mutant animals. All mice were maintained in a specific pathogen-free animal facility.
Reverse Transcription-Polymerase Chain Reaction (RT-PCR)-Total RNA was purified from cells using ISOGEN. RT was performed using 2 g of total RNA (at 42°C for 2 h) in a 20-l reaction volume containing oligo(dT) primer and Superscript II reverse transcriptase (Invitrogen). The PCR primers, designed based on the published sequence of cytochrome c, and Nucling were 5Ј-CGAATTAAAAATGGGTGATGTTGA-A-3Ј (cytochrome c, sense), 5Ј-GTGGAATTACTCATTAGTAGCCTTTT-TAAG-3Ј (cytochrome c, antisense), 5Ј-TGATCACCCAGGACCCGGAA-GTTACC-3Ј (Nucling, sense), and 5Ј-GGTGCTCTTTGAGGGCGAGGA-AGTG-3Ј (Nucling, antisense). The primers 1 and 14 designed by Honarpour et al. (12) were used for the Apaf-1 message. For PCR, 2 l of cDNA was used in a 50-l reaction mixture containing 0.5 mM primers, deoxynucleotide triphosphates, and TaqDNA polymerase, using the cycling profile of 45 s at 94°C, 1 min at 56°C, and 1 min at 72°C for 24 cycles for cytochrome c and 28 cycles for Apaf-1 with a final extension at 72°C for 10 min. The cycling profile for Nucling was 1 min at 95°C, 1 min at 56°C, and 2 min at 72°C for 32 cycles with a final extension at 72°C for 10 min. PCR products were analyzed on 2% agarose gel. RT-PCR of glyceraldehyde-3-phosphate dehydrogenase (6) or ␤-actin was used as a control.
Two-dimensional Native/Denaturing-PAGE Analysis of Protein Complexes-Wild-type or Nucling Ϫ/Ϫ MEF cell lysates were fractionated into cytosol or nuclear fraction as described before (6). The fractions were analyzed directly by native-PAGE (13). In short, cell lysates were resolved onto 2-15% precast gels (Daiichi Pure Chemicals Co. Ltd., Tokyo, Japan). For Western blotting, the native gel was soaked in blot buffer (20 mM Tris-base, 150 mM glycine, and 0.08% SDS) for 10 min, and denatured proteins were transferred to polyvinylidene difluoride membranes (Millipore, Bedford, MA) in the same buffer using the semi-dry blotting technique. Immunodetection was carried out according to standard procedures and was visualized by the ECL method (Amersham Biosciences). For two-dimensional gel analysis, individual lanes were cut out from the first-dimension native gel and layered on top of a 15-25% gradient resolving gel, and a 7% stacking gel was poured over and around the native gel slice.

RESULTS
Nucling Is a Potent Promoter of Apoptosis-During the course of our previous study to identify Nucling (6), we had been examining the functional role of this molecule and noticed that the transfection efficiency of the Nucling-expressing plasmid into mammalian culture cells was very low. In addition, the overexpressed Nucling brought about nuclear deformation or fragmentation (data not shown). These findings led us to suspect that Nucling might be able to induce apoptosis. To investigate this possibility, we performed PI staining and TUNEL staining on transfected cells.
PI has been used as a marker of cell damage. We first checked whether the nuclei of cells could be stained with PI without permeabilization by immunofluorescence microscopy. An ϳ10-fold excess of cells transfected with pFlag-Nucling were stained with PI compared with control cells transfected with pFlag-BAP (data not shown). In addition, we observed a strong correlation between the Flag-Nucling-expressing cells (Fig. 1Aa) and PI-positive cells (Fig. 1, Ab and Ac), whereas no correlation was found between Flag-BAP-expressing cells (control; Fig. 1Ad) and PI-positive cells (Fig. 1, Ae and Af). TUNEL assay also supported the possibility of apoptosis. We found a strong correlation between the Flag-Nucling-expressing cells (Fig. 1Ah) and TUNEL-positive cells (Fig. 1, Ag and Ai), whereas no correlation was found between Flag-BAP-expressing cells (control; Fig. 1Ak) and TUNEL-positive cells (Fig. 1, Aj and Al). To confirm these results concerning cell damage, we investigated the effect of Nucling on cell survival in COS7 cells detected using the transient transfection assay with the LDH release assay. Transfection of the pFlag-Nucling-expressing plasmid in COS7 cells was performed. The LDH release activity in the culture medium was calculated every 24 h for 2 days after transfection. Transfection of pFlag-Nucling led to an in-crease in LDH activity compared with that of the pFlag-vector control (mock) at 24 and 48 h after transfection (Fig. 1B, columns). At 48 h, an ϳ20% increase in LDH activity was observed in Nucling-transfected cells compared with mock-transfected cells. This percentage was in good accord with the estimated transfection efficiency (ϳ20%, data not shown) of the Nucling expression vector. These results indicate that Nucling possesses an intrinsic cell death-promoting activity.
To investigate whether the caspase system is important for the activity of Nucling promoting cell death, we performed a zVAD-fmk inhibition assay. Cell death-inducing activity of the Nucling protein was clearly reduced by zVAD-fmk, a pancaspase inhibitor (Fig. 1B, Ⅺ). In addition, the zDQMD-fmk (an inhibitor specific for caspase-3 and caspase-6) treatment assay for the Nucling-overexpressing cells revealed that Nucling can promote cell death by activating caspase-3 or caspase-6, because this treatment clearly suppressed the cell death-inducing activity of Nucling (Fig. 1C). These results suggest that Nu-FIG. 1. Nucling promotes cell death with caspase activation. A, Nucling overexpression promoted cell death. In most of the cells overexpressing Nucling (anti-Flag, green in a and red in h), PI staining (red in b and yellow in c), or TUNEL staining (green in g and yellow in i) was also observed. On the other hand, we observed no correlation between Flag-BAP expression (green in d and f, red in k and l) and PI staining (red in e and f) or TUNEL staining (green in j). Scale bar represents 10 m. B, cell death-inducing activity of Nucling was caspase-dependent. The LDH activity in culture medium was measured at 24 or 48 h after transfection of the expression vector into COS7 cells. pFlag-vector (mock) or pFlag-Nucling was used as the expression vector. Triplicates of semiconfluent cells on a 10-cm culture dish were transfected with (ϩ) or without (Ϫ) zVAD-fmk treatment. The values corresponding to the Flag-Nucling transfectants without zVAD-fmk (24 and 48 h) were statistically compared with those of the mock cells. ‫,ء‬ p Ͻ 0.05; ‫,ءء‬ p Ͻ 0.01. C, Nucling promotes cell death by activating caspase-3 or caspase-6. The HeLa cells were transfected with pEGFP-C1 (mock) or pFlag-Nucling full-length (Nucling). Triplicates of the semiconfluent cells on a 10-cm culture dish were transfected with (ϩ) or without (Ϫ) zDQMD-fmk treatment. Trypan blue exclusion assay was performed to enumerate dead detached cells.
cling may be a member of the caspase signaling pathways.
Proapoptotic Stimuli Up-regulate Nucling Expression-To assess whether Nucling is a part of the apoptosis signaling pathways, we investigated whether endogenous Nucling expression was regulated by several proapoptotic stimuli. Northern blot analyses revealed that Nucling expression was induced in COS7 cells (Fig. 2Aa) and HeLa cells (data not shown) by tumor necrosis factor-␣ (TNF-␣)/cycloheximide (CHX) (lane 3). Weak up-regulation by H 2 O 2 stress was also observed (lane 2). RT-PCR also revealed that Nucling can be induced by Adriamycin treatment or heat shock stress (Fig. 2Ab). To reconfirm these findings, an immunofluorescence analysis was performed with Nucl.mid antibody to detect the endogenous Nucling protein. We could detect many apoptotic cells by immunofluorescence analysis after TNF-␣/CHX treatment for 24 h. Most of the TUNEL-positive cells after TNF-␣/CHX treatment expressed endogenous Nucling (Fig. 2B, upper panel). The same results were obtained using H 2 O 2 (1 mM) in COS7 cells. Most of the TUNEL-positive cells following H 2 O 2 stress expressed endogenous Nucling as well (lower panel). To obtain the overall induction profile of endogenous Nucling expression during cellular stress, we examined expression patterns under different degrees of apoptotic stimulation. Serial concentrations of H 2 O 2 were used to induce stepwise stimuli for apoptosis in COS7 cells. Immunofluorescent staining using Nucl.mid antibody was performed to reveal the expression pattern of endogenous Nucling. At 24 h after trypsinization for subculture without any other forms of apoptotic stress, we observed spots of staining in nuclei in some of the cells (Fig. 2C, 0 mM). At 48 h after the subculture, we could not observe any immunodetection cells with Nucl.mid antibody (data not shown). Mild apoptotic stress (0.5 mM H 2 O 2 stress) induced Nucling expression in the nucleus (Fig. 2C, arrowheads), whereas the staining intensity and the increase in positive cell numbers became more evident under severe apoptotic stress (1 mM H 2 O 2 stress) (representatives are indicated by arrows in Fig. 2C). These results suggest that endogenous Nucling is induced by apoptotic stress in a dose-dependent manner.
Nucling Is Required for UV Irradiation-induced Apoptosis-To further address the issue of the physiological function of Nucling in vivo, we next generated Nucling Ϫ/Ϫ mutant mice. We constructed a targeting vector designed to insert the neomycin-resistance cassette into the exon encoding the LZ motif (Fig. 3A). Germline transmission was achieved from two independent clones and confirmed by Southern blot analysis (Fig. 3B). Both heterozygous and homozygous Nucling Ϫ/Ϫ mice are viable and fertile. Northern blot analysis of total RNA extracted from adult (8-week) wild-type (WT), heterozygous, and homozygous hearts and skeletal muscle using the XbaI fragment of Nucling cDNA as a probe confirmed the absence of Nucling mRNA transcripts in homozygous-deficient mice (Fig.  3C). Thus, our targeting strategy resulted in a null allele for the Nucling gene.
To obtain direct evidence that Nucling deficiency leads to defects in the regulatory mechanism for stress-induced apoptosis, we investigated the cell-death response to UV irradiationinduced cellular stress. Exposure to excessive UV irradiation is known to cause apoptosis in murine fibroblasts (14). In addition, we further reported that Nucling down-regulated the antiapoptotic factor, galectin-3, via the modification of NF-B activation (9). The activation of NF-B is known to be induced by UV irradiation, followed by the up-regulation of galectin-3 transcription. At first, we prepared MEFs from WT or Nucling Ϫ/Ϫ embryos. Western blot analysis was performed to check the expression of Nucling in WT MEFs. We could detect a distinct Nucling band in WT MEFs but not in Nucling Ϫ/Ϫ MEFs (Fig. 3D). Next, we investigated the reactivity of Nucling Ϫ/Ϫ MEFs to a proapoptotic stress using UV irradiation. UV irradiation induced an increase in cell death in WT MEFs as compared with Nucling Ϫ/Ϫ MEFs (Fig. 4). Although the total amount of LDH in Nucling Ϫ/Ϫ MEFs was less than that in WT MEFs, significant levels of cell killing were observed, arguing that UV irradiation-induced apoptosis has a partial Nucling dependence. Therefore, we assumed that Nucling regulates one or some of the many apoptotic pathways induced by cytotoxic stress from UV irradiation.
Nucling May Play a Critical Role in the Expression of Apoptosome-related Molecules-To elucidate the molecular mechanism underlying the apoptosis-promoting activity of Nucling, we investigated expression levels or activation patterns of several candidate molecules related to stress-induced apoptosis. We compared first the expression levels of Apaf-1 between WT and Nucling Ϫ/Ϫ mice in MEFs treated with H 2 O 2 (Fig. 5A) or UV irradiation (Fig. 5B). At first, we determined whether any defect occurred in the activation of caspase-9 in Nucling Ϫ/Ϫ MEFs. Immunoblot analysis revealed that not only the proenzyme of caspase-9 (pro-caspase-9) but also some of the processed type of caspase-9 (p37) does exist in WT MEFs, but no processing was observed in Nucling Ϫ/Ϫ MEFs under normal culture conditions (Fig. 5, A and B, lanes 1 and 2). In addition, this tendency was also observed in response to the apoptotic stimulus of H 2 O 2 (Fig. 5A, lanes 3 and 4) or UV irradiation (Fig.  5B, lanes 5 and 6). The apoptotic pathway mediated by caspase-9 is initiated by the release of cytochrome c from mitochondria (15,16). Therefore, we examined the expression level of cytochrome c in these MEFs. Exposure of WT MEFs to H 2 O 2 caused a large increase in cytoplasmic cytochrome c (Fig.  5A, lane 3). The same result was observed using UV irradiation. In contrast, the apoptotic stress of H 2 O 2 or UV irradiation did not cause cytochrome c up-regulation in Nucling Ϫ/Ϫ MEFs. In addition, the expression level of cytochrome c in whole-cell lysate from Nucling Ϫ/Ϫ MEFs was also reduced after H 2 O 2 treatment or UV irradiation (Fig. 5, A and B). Western blot analysis revealed that the total level of cytochrome c in the normal culture without proapoptotic stress was almost the same between WT and Nucling Ϫ/Ϫ MEFs (Cyto c (whole), lanes  1 and 2). In contrast, cytochrome c expression was remarkably down-regulated in Nucling Ϫ/Ϫ MEFs compared with WT MEFs after the proapoptotic stress (Fig. 5, lanes 3 and 4 in A, lanes  3-6 in B). These results indicate that Nucling is essential for the up-regulation of cytochrome c in response to apoptotic stimuli but not for the release of cytochrome c from mitochondria. Western blot analysis also revealed that Apaf-1 was strikingly down-regulated in its expression in Nucling Ϫ/Ϫ MEFs after proapoptotic stress (H 2 O 2 in Fig. 5A and UV irradiation in Fig. 5B). In contrast, the expression level of AIF, a mitochondrial apoptosis-inducing factor, was the same in both WT and Nucling Ϫ/Ϫ MEFs (Fig. 5B). Active caspase-9 (p37) was not detectable in Nucling Ϫ/Ϫ MEFs (Fig. 5, A and B). Pro-caspase-3 (apoptosis executioner regulated by caspase-9) remained in its inactive form in Nucling Ϫ/Ϫ MEFs (Fig. 5A, lane 4). This observation can be explained by the absence of p37 in Nucling Ϫ/Ϫ mice. These results strongly suggest that Nucling may be an apoptosis-promoting factor specifically regulating the Apaf-1/ cytochrome c/caspase-9 apoptosome pathway after cellular stress. Semiquantitative RT-PCR analysis revealed that the transcriptional levels of both cytochrome c and Apaf-1 expression were not decreased in Nucling Ϫ/Ϫ MEFs under proapoptotic conditions as shown in Fig. 5C. Both the 0.3-kb fragment of cytochrome c and the 0.5-kb fragment of Apaf-1 were amplified from the RNAs of WT and Nucling Ϫ/Ϫ MEFs. We could not observe any obvious change in the amplification reactions between these MEFs in response to the proapoptotic stimuli. Consequently, we postulated that the down-regulation of cytochrome c in Nucling Ϫ/Ϫ MEFs comes from the post-transcriptional regulation of the apoptosome molecules, Apaf-1 and cytochrome c, after proapoptotic stress. Therefore, we focused on investigating the regulatory role of Nucling for the apoptosome components in terms of their interactions at the protein level.
Nucling Plays a Critical Role in the Nuclear Translocation of Apaf-1 after Proapoptotic Stress-To confirm whether Nucling intrinsically up-regulates Apaf-1, a double immunostaining assay was performed after transfection of the Nucling expression vector into COS7 cells. The intensity of red-stained endogenous Apaf-1 was up-regulated in most of the Nucling-overexpressing cells (green in Fig. 6A). Very interestingly, the staining pattern of Apaf-1 in the Nucling-overexpressing cells was different from that of nontransfected cells. Ectopic expression of Nucling induced translocation of Apaf-1 from cytoplasm to nucleus (Fig.  6A). To investigate whether Nucling physiologically regulates the nuclear translocation of Apaf-1, we checked the expression profiles of Apaf-1 in WT and Nucling Ϫ/Ϫ MEFs after UV irradiation (Fig. 6B). We first confirmed the nuclear redistribution of Apaf-1 in WT MEFs after proapoptotic stress as reported previously (17). We found that cytochrome c redistributes 8 h after UV irradiation, forming a ring structure (indicated by an arrow in Fig. 6Bg). In addition, we detected a diffuse cytochrome c pattern in the nucleus (arrowhead in Fig. 6Bg) at the time point when Ͼ40% of cells were clearly apoptotic (data not shown). In contrast, Apaf-1 was translocated to the nucleus in most of the cytochrome c redistributed cells (yellow arrows in Fig. 6Bh). In Nucling Ϫ/Ϫ MEFs, up-regulation of cytochrome c in the cytoplasm was observed as well (arrow in Fig. 6Bj). However, nuclear localization of Apaf-1 in these cells was not observed, and the expression pattern of Apaf-1 mostly overlapped that of cytochrome c (arrow in Fig. 6Bl). These results strongly suggest that Nucling might directly regulate the redistribution of Apaf-1 into the nucleus after proapoptotic stress.
Nucling Assembles with Apaf-1/Pro-caspase-9 Complex in Vivo-On the basis of the finding that Nucling regulated the expression pattern of apoptosome molecules at the protein level, we focused on whether Nucling directly interacts with Apaf-1, caspase-9, or cytochrome c. First, the interaction between Nucling and Apaf-1 was investigated with coimmunoprecipitation assays. As shown in Fig. 7A, endogenous Nucling in COS7 cells was coimmunoprecipitated with Flag-tagged Apaf-1 by anti-Flag antibody. To visualize a protein complex containing Nucling or Apaf-1 directly, the cytosol fraction ( Fig.  7B) or nuclear fraction (data not shown) was separated under nondenaturing conditions. Immunoblots of these native gels revealed Nucling or Apaf-1 immunoreactivity within several distinct complexes in fractions from UV-irradiated WT MEFs. Immunodetection with antibodies directed against Nucling (Fig. 7B, lane 7) or Apaf-1 (Fig. 7B, lane 3) identified a common complex band of ϳ260 kDa in size (Fig. 7, Bb and BbЈ, marked with arrowheads in lanes 3 and 7Ј). Additionally, a complex of 500 kDa was detected with Nucling antibody (Fig. 7Ba in lanes  7 and 7Ј). A band of ϳ120 kDa was also detected with high intensity by Apaf-1 (ϳ120 kDa) antibody (lane 3). Two bands of ϳ160 kDa (Fig. 7Bc) and 80 kDa (Fig. 7Bd) were also detected Cell lysates were fractionated into whole-cell lysate for Apaf-1, caspase-9, pro-caspase-3, and cytochrome c (Cyto c (whole) or cytosol lysate for cytochrome c (Cyto c (cytosol)) and ␤-actin. Fractions were blotted and probed with an antibody to Apaf-1, caspase-9, cytochrome c, or caspase-3. Reprobing with an antibody to ␤-actin served as a loading control for the cytosolic fractions. Similar results were obtained with two individual lines of Nucling Ϫ/Ϫ MEFs. Bar graphs show the compiled means Ϯ S.E. of densitometric scanning of three experiments for Apaf-1, cytochrome c (whole), or procaspase-3, quantified by NIH-Image software. Data were normalized to the density of untreated WT MEFs for the Apaf-1/␤-actin ratio or the density of untreated Nucling Ϫ/Ϫ MEFs for the cytochrome c/␤-actin and pro-caspase-3/␤-actin ratios, respectively. The values corresponding to the H 2 O 2 -treated WT MEFs were statistically compared with those of the H 2 O 2 -treated Nucling Ϫ/Ϫ MEFs. ‫,ء‬ p Ͻ 0.05. B, Nucling is crucial for Apaf-1, caspase-9, and cytochrome c expression but not for AIF under cytotoxic stress induced by UV irradiation. WT or Nucling Ϫ/Ϫ MEFs were cultured for 24 h after UV irradiation as indicated. Cell lysates were fractionated into a mitochondrial fraction for AIF and cytosol lysate for cytochrome c (Cyto c), Apaf-1, and ␤-actin. Fractions were blotted and probed with an antibody to Apaf-1, caspase-9, cytochrome c, or AIF. Reprobing with an antibody to ␤-actin served as a loading control for the cytosolic fractions. Similar results were obtained with two individual lines of Nucling Ϫ/Ϫ MEFs. Bar graphs show the compiled means Ϯ S.E. of densitometric scanning of three experiments for Apaf-1 or cytochrome c (whole), quantified by NIH-Image software. Data were normalized to the density of untreated WT MEFs for the Apaf-1/␤-actin ratio or the density of untreated Nucling Ϫ/Ϫ MEFs for the cytochrome c/␤-actin ratio. The values corresponding to the UV-irradiated WT MEFs were statistically compared with those of the UV-irradiated Nucling Ϫ/Ϫ MEFs. ‫,ء‬ p Ͻ 0.05; ‫,ءء‬ p Ͻ 0.01. C, Nucling does not affect expression levels of Apaf-1 and cytochrome c transcripts. RT-PCR was carried out to confirm the presence of transcripts. Data are representative of at least three separate experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase.
with Nucling (160 kDa) antibody (lanes 7 and 7Ј). As a next step, cytosol or nuclear proteins from UV-irradiated WT MEFs were fractionated into native complexes under nondenaturing conditions (first dimension), and then subsequently separated into their individual components (second dimension) by placing a native gel slice at a horizontal position as a stack above an SDS-PAGE denaturing gel (Fig. 7C). Second-dimension gels were also transferred and immunoblotted, confirming that the highest density band in lane 3 of Fig. 7Be was free Apaf-1 (120 kDa, Fig. 7, Ce and CeЉ, marked with * in Fig. 7C). We could also detect a free Nucling band only in the nuclear fraction (160 kDa, marked with **). In addition, we confirmed the presence of Nucling, Apaf-1, and pro-caspase-9 (50 kDa) in a protein complex of ϳ260 kDa in the native first-dimension gel (Fig. 7,  b and bЈ, marked with arrowheads). The results strongly suggest that Nucling directly interacts with Apaf-1/pro-caspase-9 complex in vivo. Moreover, we detected cytochrome c (15 kDa) and Apaf-1 spots in the lane derived from the 50-kDa native protein complex (Fig. 7Cd) in the cytosol fraction. In the nu-clear fraction, Apaf-1 was detected at the same position as the complex (Fig. 7CdЉ). We also observed that an ϳ500 kDa complex contained Nucling, Apaf-1, and pro-caspase-9 in the nuclear fraction (Fig. 7CaЉ). Taken together, Nucling is considered to exist in vivo as a component of an assemblage with Apaf-1 and pro-caspase-9 during the course of stress-induced apoptosis. DISCUSSION Nucling has a unique cellular distribution, being expressed in the perinucleus and cytoplasm in mammalian cells. Ectopic induction of Nucling promoted apoptotic cell death in mammalian cells. To investigate the physiological function of Nucling for apoptosis, we generated Nucling Ϫ/Ϫ mutant mice. The mutant mice exhibited a high incidence of inflammatory lesions in preputial glands. Cells around the lesions showed resistance to apoptosis (9). On the basis of these findings and the data showing the resistance to apoptosis of MEFs from the mutant mice, we suggest that Nucling plays a crucial role in the apoptotic reaction after cytotoxic stress.
Mitochondria play a key role in the commitment of cells to apoptosis through the release of cytochrome c and many apoptosis-inducing factors from the intermembrane space into the cytosol (18 -25). Several proapoptotic stimuli induced endogenous Nucling expression in the nucleus, followed by the cytoplasm (Fig. 2). We also observed that Nucling was a crucial molecule for Apaf-1, cytochrome c up-regulation, and caspase-9 activation but not for AIF induction after exposure to a proapoptotic stimulus such as UV irradiation or H 2 O 2 stress. These findings indicate that Nucling acts as a component of the mitochondrial apoptotic pathways, especially of the cytochrome c/Apaf-1/caspase-9 apoptosome pathway. Furthermore, a lack of Nucling expression conferred on MEF cells resistance to apoptosis after cytotoxic stress from UV irradiation (Fig. 4).
It was reported that low levels of or a deficiency in Apaf-1 protein can determine sensitivity to apoptosis downstream of mitochondrial events, suggesting that regulation of Apaf-1 may be important for apoptotic processes (26,27). Our findings directly show the presence of a novel regulatory mechanism for Apaf-1 expression at the protein level. We confirmed that the inhibition of cytochrome c release in Nucling Ϫ/Ϫ cells comes from the post-translational down-regulation of the apoptosome molecules, Apaf-1 and cytochrome c, after a proapoptotic stress (Fig. 5). During stress-induced apoptosis, caspase activation requires a large number of post-translational events, including translocation to other organelles (28). This is also the case in the translocation of caspase-9 or Apaf-1 to the nucleus (17,29).
Here we propose that Nucling may be a key molecule for the retention of the caspase-9/Apaf-1 complex, its translocation to the nucleus, and its activation. We also confirmed that endogenous Nucling assembles with the Apaf-1/pro-caspase-9 complex in vivo in both the nuclear and the cytoplasmic fractions (Fig. 7). The affinity of this interaction may be weak, because a large amount of free Apaf-1 was detected in Fig. 7C, lane e. In addition, Nucling itself distributes actively to the perinucleus (6). We also found that Nucling is essential for the redistribution of Apaf-1 into the nucleus (Fig. 6). On the basis of these observations, Nucling is a strong candidate for the shuttle molecule in the translocation of caspase-9/Apaf-1 to the nucleus after stress-induced apoptosis.
It is well known that proapoptotic stimuli trigger the release of cytochrome c from mitochondria, which forms the complex with Apaf-1. In this context, the two-dimensional PAGE analysis, as shown in Fig. 7C, revealed that an Apaf-1/cytochrome c complex was present in the cytosol as an ϳ55-kDa protein assembly. On the other hand, cytochrome c was not detected in the nuclear fraction in the same analysis. This result may FIG. 6. Nucling regulates nuclear translocation of Apaf-1 after proapoptotic stress. A, COS7 cells were transfected with pFlag-Nucling and incubated with zVAD-fmk (100 M) for 18 h. Endogenous Apaf-1 was detected with polyclonal antibody against Apaf-1 and Texas Red-conjugated secondary antibody (red). Ectopic Flag-Nucling was detected by FITC-conjugated monoclonal antibody against Flag epitope (green). Two representative, double-staining images are shown. Bars, 20 m. B, WT and Nucling Ϫ/Ϫ MEFs were cultured for 18 h after UV irradiation (0 for control and 200 J/m 2 ), followed by immunofluorescent staining. Cytochrome c was detected with primary antibody against cytochrome c and FITC-conjugated secondary antibody (a, d, g, and j). Apaf-1 was detected with primary antibody against Apaf-1 and Texas Red-conjugated secondary antibody (b, e, h, and k). c, f, i, and l are merged images of the left two panels. Scale bars represent 10 m.
indicate that cytochrome c is released from Apaf-1 at the time of nuclear translocation.
As described previously, Nucling Ϫ/Ϫ mice displayed frequent inflammatory lesions (9) but no other defects similar to those of Apaf-1 Ϫ/Ϫ or caspase-9 Ϫ/Ϫ mice, including forebrain hyperpla-sia (11,30,31). Although this phenotypic discrepancy might come from the existence of unknown redundant molecules or pathways, and Nucling may not be essential for apoptosis during neural development, we observed distinct differences between WT and Nucling Ϫ/Ϫ mice concerning the apoptotic FIG. 7. Nucling assembles with Apaf-1 and pro-caspase-9. A, Nucling interacts with Apaf-1. COS7 cells were transfected with the indicated expression vector (Flag-Apaf-1 or Flag-vector). After 18 h, cellular lysates were prepared and incubated with the anti-Flag antibody for immunoprecipitation (IP). Immunoprecipitates were subjected to immunoblotting (IB) using anti-Nucling antibody (anti-Nucl.mid). upper panel, expression levels of endogenous Nucling in preimmunoprecipitated (pre-IP) cellular lysates were confirmed by IB with anti-Nucl.mid antibody (middle panel). Expression of Flag-Apaf-1 was confirmed by IB with anti-Flag antibody (lower panel). B, several Nucling-containing complexes or Apaf-1-containing complexes were observed in UV-irradiated WT MEFs but not in Nucling Ϫ/Ϫ MEFs on native gel electrophoresis. Fifty g of the cytosol fraction was subjected to nondenaturing electrophoresis (first dimension). Proteins were transferred and immunoblotted with an anti-Nucl.mid antibody for Nucling or anti-Apaf-1 antibody. Arrows indicate the prominent Nucling (a-d) or Apaf-1 (bЈ and e)-containing bands. The right panel (lane 5Ј to lane 8Ј) is an enlarged image of part of the left panel ranging from lane 5 to lane 8. A common complex band of ϳ260 kDa in lanes 3 and 7Ј is marked with arrowheads. C, apoptosome components were assembled in the complex containing Nucling. Cytosol and nuclear fractions were analyzed as the second dimension. Immunoblot of the second-dimension gel of cytosol or nuclear protein complexes from the UV-irradiated WT MEFs revealed the presence of Nucling (160 kDa), Apaf-1 (120 kDa), and pro-caspase-9 (50 kDa) in a protein complex of ϳ260 kDa in the native first-dimension gel (lanes b and bЈ, marked with arrowheads). A band of putative free Apaf-1(‫)ء‬ or free Nucling ‫)ءء(‬ is marked with asterisks. Arrows a-e and aЉ-eЉ correspond to arrows a-e of B, respectively. responsiveness under cellular stress. In particular, the expression levels of Apaf-1, cytochrome c, and caspase-9 in MEF cells under several forms of cellular stress differed strikingly between WT and Nucling Ϫ/Ϫ strains. In fact, down-regulation of Apaf-1 was also observed in several Nucling Ϫ/Ϫ tissues including kidney, spleen, and lung but not in brain (data not shown). There might be an alternative molecule(s) in place of Nucling in the brain or neural development. There might also be a distinct signal transduction pathway for stress-induced apoptosis, different from that for developmental apoptosis. Or other apoptosome-independent mitochondrial apoptosis-inducing factors, such as AIF or endonuclease G, may be prominent in the development of Nucling Ϫ/Ϫ mice. Actually, it has been reported that there must be an Apaf-1-independent pathway for apoptosis triggered by cytotoxic stress (32)(33)(34). Our findings may also support the hypothesis that the "regulation of Apaf-1 expression may be a new regulatory mechanism developed in postmitotic cells to prevent an irreversible commitment to die after the release of cytochrome c", proposed by Sanchis et al. (34) recently.
We reported previously that Nucling negatively regulated the antiapoptotic molecule galectin-3 and NF-B activation (9). NF-B is known to be most commonly involved in suppressing apoptosis by transactivating the expression of antiapoptotic genes (35). From these findings, we concluded that a stressinduced factor, Nucling, promotes apoptosis by regulating three pathways, apoptosome up-regulation, galectin-3 downregulation, and NF-B inactivation.
This report shows that Nucling is the regulator of Apaf-1 expression and plays an important role in the regulation of stress-induced apoptosis.