|Home | About | Journals | Submit | Contact Us | Français|
Human embryonic stem (hES) cells activate a rapid apoptotic response after DNA damage but the underlying mechanisms are unknown. A critical mediator of apoptosis is Bax, which is reported to become active and translocate to the mitochondria only after apoptotic stimuli. Here we show that undifferentiated hES cells constitutively maintain Bax in its active conformation. Surprisingly, active Bax was maintained at the Golgi rather than at the mitochondria, thus allowing hES cells to effectively minimize the risks associated with having pre-activated Bax. After DNA damage, active Bax rapidly translocated to the mitochondria by a p53-dependent mechanism. Interestingly, upon differentiation, Bax was no longer active and cells were not acutely sensitive to DNA damage. Thus, maintenance of Bax in its active form is a unique mechanism that can prime hES cells for rapid death, likely to prevent the propagation of mutations during the early critical stages of embryonic development.
Human embryonic stem (hES) cells have the unique capability of differentiating into all cell types, leading to the development of an entire organism (Jaenisch and Young, 2008; Thomson et al., 1998). As the integrity of ES cells is critical for the developing embryo, these cells have likely evolved mechanisms that detect and respond rapidly to adverse stimuli. Indeed, hES cells have been shown to be highly sensitive to DNA damage, but the molecular mechanisms underlying this rapid death remain unclear (Filion et al., 2009; Grandela et al., 2008; Momčilović et al., 2009; Qin et al., 2007).
Caspases are critical mediators of apoptosis in mammalian cells and a key protein that controls their activation is Bax, a proapoptotic member of the Bcl-2 family (Tait and Green, 2010; Youle and Strasser, 2008). In healthy cells, Bax has been reported to be predominantly cytosolic, present in an inactive conformation. However, apoptotic stimuli such as DNA damage activate signaling pathways that result in the induction of BH3-only family proteins that promote a change in Bax conformation (Brunelle and Letai, 2009; Chipuk et al., 2010). Activated Bax then translocates to the mitochondria where it induces cytochrome c release and apoptosome-mediated caspase activation, resulting in apoptotic cell death (Dewson and Kluck, 2009; Taylor et al., 2008). Thus, Bax activation is a critical event in the commitment of cells to apoptosis.
Considerable details of the mechanism by which Bcl-2 family proteins regulate Bax activation have recently emerged. Bax can be directly activated via binding to a subset of BH3-only proteins known as “activators” (e.g. Bim, Bid, and Puma). However, the antiapoptotic members of the Bcl-2 family (e.g. Bcl-2, Bcl-XL) can also bind to Bax and prevent its function. A second subset of BH3-only proteins known as “sensitizers” (e.g. Bad, Noxa, Bmf) bind to and inactivate the antiapoptotic Bcl-2 family proteins and therefore, may also be needed to enable efficient Bax activation (Giam et al., 2008; Shamas-Din et al., 2011). Importantly, structural studies have shown that binding of the BH3-only activators to Bax induces a conformational change in Bax that exposes the N-terminus and mobilizes the C-terminus to insert into the mitochondrial outer membrane. These conformational changes enable Bax to oligomerize and form pores in the mitochondrial membrane that trigger the release of cytochrome c (Gavathiotis and Walensky, 2011). Adding to this complexity, Bax activation can also be regulated in cells by cytosolic factors such as Ku70, a protein also involved in non-homologous end joining DNA repair, that binds to inactive Bax and prevents its activation (Cohen et al., 2004; Gomez, 2007). Apoptotic stimuli result in the acetylation of Ku70 and the disruption of the Bax-Ku70 complex, facilitating Bax activation (Cohen et al., 2004; Li et al., 2007; Subramanian et al., 2005).
While the main components of the apoptotic pathway have been identified, exactly how this pathway is regulated in various primary cells remains unclear. Here, we examined the apoptotic pathway in hES cells and report a unique mechanism engaged by hES cells that can prime them to undergo rapid apoptosis in response to genotoxic damage.
To examine the apoptotic pathway in hES cells, we treated them with various stimuli that induce apoptosis. We found that DNA damage activates very rapid death in H9 hES cells, with all cells dying by 5 hours after etoposide treatment (Figures 1A, B). This death was apoptotic as it resulted in caspase-3 activation and was completely prevented by the addition of the pan-caspase inhibitor z-VAD-fmk (Figures 1A, B, C). hES cells were remarkably more sensitive to etoposide than fibroblasts in a time- and dose-dependent manner (Figure S1A). Furthermore, this sensitivity of hES cells was selective to DNA damage. While exposure to either etoposide or doxorubicin induced rapid apoptosis, tunicamycin and taxol treatments, which induce apoptosis by ER stress and microtubule stabilization respectively, did not induce hES cell death within this timeframe (Figure 1D).
hES cell death in response to DNA damage was also p53-dependent, as phosphorylated p53 (Ser15) accumulated after etoposide treatment (Figure S1B) and shRNA-mediated knock down of p53 completely blocked hES cell death (Figure 1E, Figure S1C). p53 can induce cell death either by localizing to the nucleus and activating proapoptotic genes such as Noxa or Puma (Nakano and Vousden, 2001; Oda et al., 2000; Yu et al., 2001) or by a faster, transcription independent mechanism whereby it translocates to the mitochondria and directly interacts with members of the Bcl-2 family to promote cytochrome c release (Chipuk et al., 2004; Mihara et al., 2003). Following DNA damage, p53 was detectable only in the nuclei of etoposide-treated hES cells (Figure 1F). Furthermore, addition of the protein synthesis inhibitor cycloheximide prevented the etoposide-induced death of hES cells (Figure S1D). Thus, the rapid apoptotic death of hES cells was likely mediated by a protein synthesis-dependent activity of p53, rather than by direct translocation to the mitochondria.
Bax and Bak are the two main proapoptotic members of the Bcl-2 family that directly mediate mitochondrial permeabilization during apoptosis. We found etoposide-induced hES cell death to be Bax-dependent but Bak-independent (Figures 2A, B, Figure S2A), indicating that Bax is the key inducer of apoptosis in these hES cells. Typically, in untreated cells, Bax is present in its inactive, monomeric conformation in the cytosol (Hsu, 1997). However, after the induction of apoptosis, Bax undergoes a conformational change, oligomerizes, and inserts into the outer mitochondrial membrane where it promotes the release of cytochrome c which results in caspase activation (Antonsson, 2001; Wolter et al., 1997; Youle and Strasser, 2008). To determine the kinetics of Bax activation in hES cells, we immunolabeled untreated and etoposide treated hES cells using the Bax 6A7 antibody, which detects Bax only in its active conformation (Hsu and Youle, 1997). Surprisingly, we found Bax to be in its active state in the untreated hES cells (Figures 2C). These results were unexpected because thus far Bax activation has been seen only in cells that are actively undergoing apoptosis (Youle and Strasser, 2008) (Figure S2B). However, we found no evidence of cell death in untreated hES cells that have activated Bax (Figure 2C and Figure S2C). We additionally confirmed that the Bax 6A7 staining was specific for Bax as shRNA-mediated knock down of Bax decreased active Bax staining in hES cells (Figure 2D). Consistent with the observation that Bax is maintained in an active state, active Bax was immunoprecipitated with the 6A7 antibody in untreated hES cells (Figure 2E). In contrast, in a human fibroblast line (CCL-151), the active form of Bax could only be detected upon the induction of apoptosis (Figure 2E). The observation that hES cells have Bax in its active conformation provides insight into how these cells are able to rapidly induce apoptosis after DNA damage. However, this raised the question of how hES cells are capable of maintaining Bax in its active state without spontaneously activating apoptosis.
In cells undergoing apoptosis active Bax localizes to the mitochondria to promote caspase activation (Brunelle and Letai, 2009; Gross et al., 1999; Hsu, 1997; Wolter et al., 1997). Based on our observation that hES cells are Bax 6A7 positive, we asked whether active Bax was localized to the mitochondria of untreated hES cells. Unexpectedly, we found active Bax to be localized not at the mitochondria but at the Golgi in the untreated hES cells (Figure 3A and Figures S3A, B, C). Co-immunolabeling of hES cells with Bax 6A7 and TGN46, a protein maintained at the trans-Golgi network, showed that active Bax is specifically localized to the trans-Golgi compartment; no colocalization was seen with the cis-Golgi marker GM130 (Figures 3A, B). Consistent with the observation that active Bax was localized at the Golgi, treatment of hES cells with Brefeldin A, a known Golgi complex disassembly agent (Fujiwara et al., 1988), altered Bax localization in hES cells (Figure S3C). Importantly, in addition to our findings that the 6A7 antibody labeled Bax at the Golgi, we found ectopically expressed GFP-Bax to be localized to the trans-Golgi compartment in hES cells, in contrast to the cytosolic localization that is seen in other cell types (Figure 3C). To confirm Bax localization to the Golgi using biochemical techniques, we also conducted subcellular fractionation of hES cells by sucrose gradient ultracentrifugation. Our results show strong localization of Bax in the Golgi-enriched fractions of hES cells. In contrast, no Bax was detected in the Golgi fractions in HeLa cells (Figure 3D). To examine whether active Bax at the Golgi was also seen in other hES cell lines, we examined the status of Bax in a panel of four hES and four non-hES cell lines. Indeed, we found Bax to be in an active conformation and localized to the Golgi in three of the four hES cell lines we tested (H9, H7, CHB6; see Discussion section). Importantly, no active Bax was detected in any of the non-hES cells tested (Figure 3E). Thus, multiple experimental approaches show that activated Bax is found at the Golgi of healthy, untreated hES cells.
We noticed that a subset (approximately 30–35%) of untreated hES cells did not have Bax in its active conformation. Since the cell cycle distribution in a population of hES cells is reported to be mostly in S phase (65%), followed by G0-G1 (20%) and G2-M (15%) (Becker et al., 2006), we investigated whether the presence of active Bax in hES cells was correlated with the cell cycle state. We labeled untreated hES cells with Bax 6A7, EdU (S phase marker), and DAPI (for DNA) and conducted FACS analysis to identify the population of cells with active Bax. Our results showed that Bax 6A7 positive cells were almost exclusively (greater than 95%) in the S phase of the cell cycle (Figure 4A).
As hES cells have the potential to be differentiated into all cell lineages, we examined whether the constitutively active status of Bax seen in undifferentiated hES cells changes with differentiation. Differentiation of hES cells can be achieved by removing cells from feeder layers or matrixes and allowing them to grow in suspension on a non-adherent surface (Odorico et al., 2001). hES cells begin to differentiate under these conditions and form aggregates called embryoid bodies which can be dissociated for analysis. Interestingly, Bax was found in its active conformation only in undifferentiated hES cells; Bax was no longer present in its active state after just two days of differentiation (Figure 4B and Figure S4). Consistent with the idea that the presence of active Bax in undifferentiated hES cells primes them for rapid apoptosis, the two-day differentiated hES cells that do not have active Bax were also no longer acutely sensitive to DNA damage (Figures 4C, D). These results illustrate how the apoptotic machinery undergoes dynamic changes to set apoptotic thresholds even at the earliest stages of hES cell differentiation.
Confocal analysis using triple labeling with Bax 6A7, TGN46 (Golgi), and Tom20 (mitochondria) antibodies confirmed active Bax to be localized at the Golgi in untreated hES cells (Figure 5A). Next, we examined whether active Bax changes localization after DNA damage. As early as 3 hours after etoposide treatment Bax 6A7 co-localized with the mitochondrial marker Tom20, indicating that active Bax translocates from the Golgi to the mitochondria in hES cells undergoing apoptosis (Figure 5B and Figure S5). Bax was localized to distinct focal points at the mitochondria during apoptosis, as previously described (Karbowski et al., 2002; Nechushtan et al., 2001). Bax translocation to the mitochondria was also confirmed with another Bax antibody (2D2) that showed a similar pattern of Bax localization to mitochondrial foci upon etoposide treatment in hES cells (Figure 5C). As our results show that apoptosis after DNA damage in hES cells was p53-dependent, we investigated whether p53 was necessary for the translocation of active Bax from the Golgi to mitochondria. Indeed, we found active Bax to be maintained at the Golgi after etoposide treatment in the p53-knock down hES cells, indicating that p53 was required for the Golgi-to-mitochondria translocation of active Bax after DNA damage (Figures 5D, E).
The antiapoptotic members of the Bcl-2 family such as Bcl-2 and Bcl-XL have been shown to inhibit Bax-induced mitochondrial permeabilization (Youle and Strasser, 2008). We examined whether these proteins were important for maintaining Bax at the Golgi and restricting its translocation to the mitochondria in untreated hES cells. Addition of ABT-737, an inhibitor of Bcl-2 and Bcl-XL (Oltersdorf et al., 2005), did not change localization of active Bax at the Golgi (Figure 6A). Consistent with this observation, we did not detect either Bcl-2 or Bcl-XL at the Golgi in hES cells (Figure 6B). Treatment with ABT-737 alone also did not induce hES cells to undergo apoptosis. However, ABT-737 addition increased the rate of etoposide-induced hES cell death, with nearly 60% of cells dying by 2 hours after etoposide treatment (Figure 6C). These results suggest that inactivation of Bcl-2 and Bcl-XL alone is not sufficient to trigger Bax translocation to the mitochondria in the absence of an apoptotic stimuli in hES cells.
Previous studies have found Bax to be maintained in an inactive form in a complex with Ku70 (Cohen et al., 2004; Subramanian et al., 2005). A peptide containing the Ku70 binding domain (Bax inhibitory peptide, BIP) potently binds inactive Bax and prevents its activation (Yoshida et al., 2004). Because Ku70 cannot bind active Bax, BIP is unable to block apoptosis once Bax is activated. We investigated the ability of BIP to block apoptosis in hES cells following DNA damage as an alternative method of confirming that Bax is pre-activated in hES cells. As anticipated, BIP could effectively inhibit apoptosis in HeLa cells but not in hES cells where Bax is already in its active conformation (Figures 7A, B).
In response to DNA damage, Ku70 becomes acetylated by histone acetyl-transferases, resulting in the release of Bax and the subsequent translocation of Bax to mitochondria (Cohen et al., 2004; Subramanian et al., 2005). Because we found that Bax was in its active conformation in untreated hES cells, we examined the acetylation status of Ku70 in these cells. Our immunoprecipitation results show that Ku70 is constitutively acetylated in the untreated hES cells (Figure 7C). In contrast, Ku70 was acetylated only after DNA damage in HEK293 and HeLa cells (Figure 7C, data not shown). Consistent with the model that acetylated Ku70 can no longer bind Bax, we found no interaction between Ku70 and Bax in untreated hES cells (Figure 7D). These results in hES cells are distinct from those in HEK293 and HeLa cells, where Bax co-immunoprecipitated with Ku70. The interaction between Bax and Ku70 was disrupted only after the acetylation of Ku70 as a consequence of DNA damage in these cells (Figures 7C, D, data not shown). These results identify an interesting feature of hES cells in which Ku70 is constitutively acetylated in the absence of DNA damage, and therefore incapable of sequestering Bax.
Here we describe the striking observation that healthy undifferentiated hES cells maintain Bax in its pre-activated state at the Golgi. This is in contrast to other cell types in which Bax is typically present in an inactive form in the cytosol. Importantly, as early as 3 hours after DNA damage, active Bax translocates from the Golgi to the mitochondria in a p53-dependent manner. We propose that by sequestering active Bax at the Golgi, hES cells are able to effectively minimize the risks associated with having pre-activated Bax yet maintain their ability to respond rapidly to genotoxic insults.
Detailed biochemical and structural studies on the mechanism of how Bax is activated have emerged in recent years. However, exactly how Bax activation is regulated in various primary cells remains largely unexplored. This is particularly important as different cell types may modify their apoptotic threshold in order to function in, and adapt to, different physiological conditions. Typically, Bax is activated by the BH3-only proteins in a rapid sequence of events. The direct binding of the activator BH3-only proteins to Bax induces a conformation change in the N-terminus (recognized by the 6A7 antibody) which is followed immediately by the mobilization of the C-terminal transmembrane domain and the insertion of Bax on the mitochondrial outer membrane (Walensky and Gavathiotis, 2011). Our results show that these steps of Bax activation are uncoupled in hES cells with the 6A7 positive form of Bax stabilized without mitochondrial translocation. Whether the maintenance of active Bax is dependent on some BH3-only proteins in hES cells is not known. The full repertoire of BH3-only proteins expressed in hES cells is unknown and the simultaneous knock down of multiple members may be needed to discern their importance in inducing Bax activation. Recently, a constitutively active isoform of Bax (Baxβ) was detected in various cell lines. However, in contrast to our results in hES cells, the Baxβ isoform was found at the mitochondria and only detected in the presence of proteasome inhibitors in the cell lines analyzed (Fu et al., 2009).
A recent study found Bax to shuttle between the cytosol and mitochondria in healthy cells and that Bcl-XL is important for the retrotranslocation of Bax from the mitochondria to cytosol (Edlich et al., 2011). We specifically examined whether Bcl-2 and Bcl-XL were important for maintaining active Bax at the Golgi and preventing its spontaneous translocation to mitochondria. While both Bcl-2 and Bcl-XL are known to promote hES cell survival (Ardehali et al., 2011; Bai et al., 2012), our results show that inhibition of these proteins alone did not change the status of active Bax or trigger apoptosis in hES cells. The association of Bax with Ku70 has been reported as one mechanism by which Bax is maintained in an inactive conformation (Cohen et al., 2004; Subramanian et al., 2005). Importantly, we found that Ku70 is constitutively acetylated and therefore unable to bind Bax in hES cells. These results contrast the findings in other cell types where the Ku70-Bax complex is disrupted only when Ku70 is acetylated by histone acetyltransferases (e.g. p300/CBP, PCAF) in response to apoptotic stimuli (Cohen et al., 2004; Subramanian et al., 2005). The observation that Ku70 is constitutively acetylated and unable to bind to Bax provides insight into how Bax is available for activation in hES cells.
Our results also highlight the fact that the apoptotic machinery undergoes dynamic changes even at early stages of differentiation. While undifferentiated hES cells have constitutively active Bax and undergo rapid apoptosis in response to DNA damage, just 2 days of differentiation induced significant changes such that Bax was no longer active, and the cells were no longer highly sensitive to DNA damage. This could be manifested with even greater complexity in vivo as cells during early embryogenesis undergo rapid proliferation and differentiation. While most of our studies were done in the H9 hES cell line, we also examined the status of Bax in a panel of other hES cells. Like the H9 hES cells, the H7 and CHB6 hES cells also exhibited 6A7 positive Bax staining at the Golgi (Figure 3E). However, the staining intensity was weaker in the H7 and CHB6 hES cells (requiring confocal imaging) as compared to the H9 hES cells, and negative in the H1 hES cells. Thus, we anticipate that there may be differences in the status of the apoptosis machinery between various hES cell lines potentially as a consequence of their derivation from the inner cell mass at different time points during blastocyst outgrowth.
Recent evidence demonstrates that sustained treatment of ES cells with DNA damaging agents can predispose ES cells to karyotypic abnormalities (Bueno et al., 2009; Stambrook and Tichy, 2010). Therefore, the ability to maintain genomic integrity in ES cells is of critical importance to prevent propagation of mutations to the developing embryo and reduce the risk of carcinogenesis. Because hES cells are critical for embryogenesis, yet need to be sensitive to genomic perturbations, their ability to engage a rapid death after DNA damage is likely a key evolutionary adaptation for optimal organismal development.
Human embryonic stem cell lines H9 (WA09) and H7 (WA07), H1 (WA01) were obtained from WiCell Research Institute (Wisconsin). CHB6 hES cells were obtained from Dr. George Daley (Children’s Hospital, Boston). Cells were seeded as undifferentiated colonies on plates coated with fibronectin (Sigma), maintained at 37°C and 5% CO2, and fed mouse embryo fibroblasts (MEF) conditioned human ES culture medium plus bFGF daily. For CHB6 hES cells and certain experiments that required larger number of cells, hES cells were seeded on matrigel and fed daily with mTeSR media (Stem Cell Technologies). Differentiation of hES cells was induced by plating 4-day-old colonies onto non-adherent plates without fibronectin and fed non-conditioned hES cell media daily. All experiments with hES cells were conducted with the H9 cell line. As indicated, the experiment in Figure S3C was conducted in the H7 cell line.
The human lung fibroblast cells (CCL-151), HEK293T cells, and HeLa cells were received from ATCC and maintained and cultured as recommended by the supplier in modified Ham’s F12 (F-12K) medium supplemented with 15% FBS (CCL-151) and Dulbecco’s Modified Eagle’s Medium (DMEM) supplemented with 10% FBS and 1% L-glutamine (HEK293 T cells). Human dermal fibroblasts (HDF UNCH1) were received from the Tissue Culture Facility at UNC and maintained in DMEM/F12 + 5% FBS.
All treatments were performed on day 4–5 for hES cells or at 80% confluency for fibroblasts. Etoposide (Sigma) was added at a final concentration of 20 μM to hES cells and fibroblasts unless otherwise stated. Staurosporine (Sigma) was added to a final concentration of 200 nM. The pan-caspase inhibitors Z-VAD-FMK (MP Biomedicals) and Q-VD-OPH (SM Biochemicals) were added to the cells at a concentration of 50 μM (Z-VAD) and 25 μM (QVD-OPH). Cycloheximide (Sigma) was added at a final concentration of 1 μg/ml for 5 hrs. The cell impermeable dyes propidium iodide and sytox green (Molecular Probes) were added to hES cells at a final concentration of 1 μg/ml (propidium iodide) and 10 nM (sytox green) and incubated at 37ºC in the dark for 20 min. Membrane permeable peptides, negative control and Bax inhibiting peptide (BIP) were purchased from Calbiochem. The stock solutions were prepared in DMSO and the peptides were added directly into medium (final concentration: 200 μM) at 37 °C 1hr before apoptosis treatments. ABT-737 was purchased from Selleck Chemicals LLC (Houston TX, USA) and stock solutions were prepared in DMSO. hES cells were treated with etoposide (20 μM) for 5 hours and HeLa cells for 16–24 hours. Cell survival was assessed at multiple time points by morphology (by counting the total number of cells attached with intact morphology in a defined area at each timepoint). Wherever indicated, cell death was also quantitated by counting apoptotic nuclei (condensed or fragmented) stained with Hoechst Staining or with propidium iodide dye exclusion. Golgi was disrupted using Brefeldin A (5 μg/ml; Sigma) for 15 minutes at 37ºC. All experiments were repeated at least 3 times and error bars indicate standard deviation.
To generate lentiviral-mediated knock down lines of hES cells, day 3 hES cell colonies were transduced with p53, Bax or Bak lenti-shRNAs (Open Biosystems) for 10 hours. After 10 hours, the cells were allowed to recover for 12–14 hours in fresh conditioned medium. Selection with puromycin (10 μg/ml; Sigma) was started 24 hours after the initiation of the transduction and was continued for 4 days, with replacement of the conditioned medium and puromycin every day. Transduction efficiency was validated by Western blotting on healthy, recovered hES cell colonies.
To visualize GFP-tagged Bax, 3-day colonies of hES cells were transfected with 2 μg of hBaxC3-EGFP (Addgene) with FuGENE HD transfection reagent (Roche). EdU labeling and detection was performed using a flow cytometry kit according to manufacturer’s instructions (Click-iT EdU Alexa Fluor 594, Invitrogen, Carlsbad, CA). DAPI (2 μg/ml; Sigma, St. Louis, MO) was used to measure the DNA content. Samples were subjected to FACS analysis as measured on a Dako CyAN ADP instrument. For immunofluorescence, cells were fixed with 4% paraformaldehyde for 30 min at 4ºC and permeabilized in 0.3% Triton X-100 (or 0.04% Saponin in Supplementary Fig. S3B) for 1 hr at room temperature. After blocking, cells were then treated with primary and secondary antibodies using standard methods. Immunoprecipitation experiments were performed in order to detect the active form of Bax (using CHAPS buffer), the acetylation status of Ku70 (using Triton X-100 buffer), and the interaction status between Bax and Ku70 (using CHAPS buffer).
To separate the Golgi from the mitochondrial fraction we performed subcellular fractionation as previously described (Jesch and Linstedt, 1998). Briefly, hES cells were collected by centrifugation (2,000 rpm) for 5 min, washed once in PBS (containing 1 mM EDTA), washed once in sucrose buffer (250 mM sucrose, 10 mM triethylamine pH 7.4, 1 mM EDTA, with protease inhibitors), homogenized by 20 passages through a 25-gauge needle in 0.5 ml homogenization buffer (50 mM NaCl, 10 mM triethylamine, pH 7.4, 1 mM EDTA, with protease inhibitors), and centrifuged at 1,000 × g for 2 min. The postnuclear supernatant was adjusted to 1.6 M sucrose (1.1 ml), layered on 2 M sucrose (0.5 ml), covered with three sucrose layers (1.1 ml each: 1.4 M, 1.2 M, 0.8 M), and centrifuged at 105,000 × g for 105 min in a MLS-50 rotor (Beckman), and 0.4-ml fractions were collected from the top. All steps were carried out at 0–4°C. Immunoblotting was used to assay each fraction for the presence of Bax (N20), Golgi (TGN46) and mitochondria (VDAC).
Images were acquired using a Hamamatsu ORCA-ER digital B/W CCD camera mounted on a Leica inverted fluorescence microscope (DMIRE 2). The software used for image acquisition was Metamorph version 5.0 (Universal Imaging Corporation). Confocal images were taken on an Olympus FV1000 confocal microscope.
Additional detailed description of the methods is included in the Supplementary Information.
We thank members of the Deshmukh Lab for helpful discussions and critical review of this manuscript. We also thank Dr. Adam Listedt, Dr. Terry Magnuson, Dr. Vladmir Ghukasyan and Dr. Lucy Williams for technical advice. This work was supported by NIH grants GM078366 (to MD) and UNC’s University Cancer Research Fund (to MD). RD was supported by the Helen Lyng White Fellowship. VG was supported by the American Heart Association and American Brain Tumor Association postdoctoral fellowships. hES cells were maintained in the UNC Human Embryonic Stem Cell Core Facility. Confocal imaging was supported by Core 5 of NINDS Center Grant P30 NS04892.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.