Cell culture, irradiation and FB₁-treatment

HeLa cells (ATCC) were cultured in low glucose Dulbecco's Modified Eagle's Medium (DMEM from Gibco BRL) containing 10% fetal bovine serum (FBS), penicillin (50 units/ml), streptomycin (50 µg/ml) and 2 mM glutamine. 1 h prior to irradiation, HeLa cells were changed to medium containing 0.2% FBS. BAEC were cultured until confluent in low glucose DMEM supplemented with 10% normal calf serum, 1 ng/ml basic FGF (Scios Inc.), penicillin (50 units/ml), streptomycin (50 µg/ml) and 2 mM glutamine. After confluence, BAEC were maintained in DMEM with 5% heat inactivated normal calf serum. At 24 h before irradiation, medium was changed to DMEM with 0.2% human albumin (HA) containing 50 µg/ml pentosan polysulfate (PPS; Sigma). FB₁ (Sigma or Biomol) was solubilized in phosphate-buffered saline (PBS; 10 mM KH₂PO₄, 150 mM NaCl, 3 mM NaH₂PO₄·7H₂O) at a concentration of 5 mM and added to cells at final concentration of 10–50 µM FB₁. Note commercially-available FB₁ is a biologic isolate from Fusarium verticillioides and Fusarium proliferatum that displays batch-to-batch variation. Hence, each batch must be tested empirically for effectiveness. Irradiation was carried out at 22°C using a Cs-137 irradiator (Shepherd Mark-I, model 68, SN 643) at a dose rate of 240 cGy/min.

Antibodies and Immunoblotting

Anti-Bax (N20, Santa Cruz; 0.4 µg/ml or YTH-6A7, Trevigen; 0.2 µg/ml), anti-Bak (NT, Upstate Biotechnology; 2 µg/ml), anti-Bcl-xL (Transduction Laboratories; 0.5 µg/ml), anti-BID (AF860 R&D; 1 µg/ml), anti-BIM (Calbiochem; 1∶1000), anti-PUMA (N-terminal, Sigma; 1 µg/ml), anti-GST (Z-5, Santa Cruz; 0.2 µg/ml), anti-Fis1 (Imgenex; 1∶1000), anti-COX II (12C4, Molecular Probes; 0.2 µg/ml), anti-COX IV (20E8, Molecular Probes; 0.2 µg/ml), anti-PDH-E1α (9H9, Molecular Probes; 0.2 µg/ml), anti-Hsp60 (LK-1, Calbiochem; 0.05 µg/ml), anti-KDEL (10C3, Abcam; 1∶50), anti-VDAC [31HL(Ab-3), Calbiochem; 2 µg/ml], anti-Climp63 (G1/296, Alexis; 1 µg/ml), anti-Calnexin (H-70, Santa Cruz; 0.2 µg/ml) or anti-metaxin (Transduction laboratories; 0.5 µg/ml) was used as indicated. Proteins were separated on a 12–15% SDS-PAGE gel and transferred to 0.2 µm PVDF membrane (Biorad). Then the blot was incubated with 5% Blotto (Biorad) in TBST (20 mM Tris–HCl, pH 7.5, 500 mM NaCl, 0.1% Tween 20) for 1 h at room temperature. After brief washing with TBST, the blot was incubated with primary antibodies in TBST containing 2.5% BSA for 16 h at 4°C. This was followed by extensive washing with TBST and incubation with the appropriate secondary antibodies conjugated to HRP (Amersham) in TBST containing 5% Blotto for 1 h at room temperature. After extensive washing with TBST, immunolabeled proteins were revealed using ECL kit (Amersham).

Apoptosis assay

Apoptotic HeLa cells were stained with bis-benzimide Trihydrochloride (Hoechst #33258; Sigma) and quantified by fluorescence microscopy, as described [32].

Caspase assay

Caspase 3 activity was measured fluorospectrometrically using Z-DEVD-AFC as substrate as described [33].

Cytochrome c release in intact cells

Cytosolic cytochrome c content was measured according to Newmeyer [34] by either immunoblotting using mouse monoclonal anti-cytochrome c antibody (7H8.2C12, PharMingen; dilution 1∶500) or by ELISA (Human Cytochrome c Quantikine ELISA Kit, R&D).

Isolation of subcellular fractions from HeLa cells and BAEC

Mitochondria and endoplasmic reticulum from HeLa cells and BAEC were isolated by differential centrifugation via a discontinuous sucrose gradient as described [35], with modification. All procedures were conducted at 4°C. For these studies, cell monolayers were washed twice and scraped into ice-cold PBS. Cells were collected by centrifugation at 500×g for 10 min, then washed and resuspended in SHE buffer [250 mM sucrose, 10 mM HEPES-KOH pH 7.4, 1 mM EGTA, protease inhibitor cocktail (Roche)], and homogenized twice by 20 strokes in a loose-fitting Dounce homogenizer. Cell debris and nuclei were pelleted by centrifugation at 800×g for 5 min. The postnuclear supernatant was collected and centrifuged at 10,000×g for 10 min and the pellet containing heavy membrane fractions (P₁₀) was resuspended in 20 ml SHE buffer and centrifuged through an underlayed cushion of 5 ml HE medium containing 750 mM sucrose at 10,000×g for 30 min. The resulting mitochondrial-enriched pellet was resuspended in SHE buffer and further purified by a second discontinuous sucrose gradient centrifugation as above. MAM-free mitochondria were isolated by Percoll gradient according to the method of Vance [14] with slight modification. For these studies, heavy membrane fractions (P₁₀) were layered on top of 20 ml of medium consisting of 2.2 ml of 2.5 M sucrose, 6.55 ml of Percoll and 12.25 ml of 10 mM HEPES, pH 7.4, 1 mM EGTA, and centrifuged at 100,000×g for 1 h in a Beckman Ti-55.2 rotor. Two bands were recovered from the gradient; a lower band corresponding to mitochondria (R_f: 0.48) and an upper band containing MAM (R_f: 0.23) [36]. Mitochondria were collected using a Pasteur pipette, diluted 10-fold with SHE buffer and washed twice by centrifugation at 10,000×g for 10 min to remove the Percoll. The MAM fraction was diluted with SHE buffer and pelleted by centrifugation at 100,000×g for 1 h in a Beckman Ti-55.2 rotor. To isolate the ER-enriched membrane fraction, the supernatant of the P₁₀ fraction was centrifuged at 100,000×g for 1 h. Fractionation was monitored by immunoblotting using mouse monoclonal anti-COX II (mitochondrial marker) and rabbit polyclonal anti-Climp63 or anti-Calnexin (ER Markers).

Bax insertion status by alkali extraction of the mitochondria

For measurement of Bax insertion, mitochondria, subjected to alkali extraction by incubation in freshly prepared 0.1 M Na₂CO₃ (pH 11.5) for 60 min at 4°C, were pelleted by centrifugation at 100,000×g for 30 min at 4°C [37], [38]. Supernatant (alkaline-sensitive fraction) and mitochondrial pellet (alkaline-resistant fraction) were separated, mitochondrial membranes were lysed in RIPA buffer, and fractions were analyzed by immunoblotting using anti-Bax antibody.

Bax oligomerization status by gel filtration

Gel filtrations were performed at 4°C on a 1.5×55 cm Sephacryl 200 column equilibrated in 25 mM HEPES-KOH pH 7.4, 137 mM NaCl, 2 mM EDTA, 10% Glycerol, 1% or 2% (w/v) CHAPS and calibrated with thyroglobulin (669 KD), ferritin (440 KD), aldolase (158 KD), bovine serum albumin (67 KD), ovalbumin (43 KD), chymotrypsinogen A (25 KD), and ribonuclease A (13.7 KD) (Amersham Biosciences). A 1.5 ml sample was loaded onto the column and fractions of 1 ml were collected at a flow rate of 0.5 ml/min. Aliquots of 400 µl from the fractions were analyzed by Western blotting after 20% TCA precipitation. Samples were separated on 12% or 15% SDS-PAGE under reducing conditions and transferred to 0.2 µm PVDF membranes (Biorad). Proteins were detected with the specific antibodies as indicated in the figures, and the blots were developed with the ECL system from Amersham Pharmacia Biotech.

Ceramide synthase activity assay

For these studies, 75×10⁶ cells were pelleted, washed once with cold PBS, and resuspended in 300 µl of homogenization buffer [25 mM HEPES pH 7.4, 5 mM EGTA, 50 mm NaF, protease inhibitor cocktail (Roche)]. Cells were disrupted on ice by sonication, and lysates centrifuged at 800×g for 5 min. The postnuclear supernate was centrifuged at 100,000×g for 1 h and the microsomal pellet (P₁₀₀) was resuspended into 1 ml of homogenization buffer. To measure ceramide synthase activity in sub-cellular fractions, mitochondrial-enriched and endoplasmic reticulum-enriched fractions were prepared as described above. Membranes were prepared fresh daily. The assay for ceramide synthase (EC 2.3.1.24) activity was based on the procedure of Harel and Futerman [39]. 75 µg protein were incubated in a 1 ml reaction mixture containing 2 mM MgCl₂, 20 mM HEPES pH 7.4, 20 µM defatted BSA (Sigma), and varying concentrations (0.2–20 µM) of dihydrosphingosine (sphinganine; Biomol), 70 µM unlabeled palmitoyl-coenzyme A (Sigma), and 3.6 µM (0.2 µCi) [1-¹⁴C]palmitoyl-coenzyme A (55 mCi/mmol; American Radiolabeled Chemicals). Dihydrosphingosine was dried under N₂ from a stock solution in 100% ethanol and dissolved by sonication in reaction mixture prior to addition of protein. The reaction was started by addition of palmitoyl-coenzyme A at 37°C, and after 1 h stopped by extraction of lipids using 2 ml of chloroform∶methanol (1∶2, v/v). 500 µl of lower phase was dried under N₂ and dihydroceramide resolved by thin layer chromatography using a solvent system of chloroform∶methanol∶3.5 N ammonium hydroxide (85∶15∶1). Radiolabeled dihydroceramide was identified by iodine vapor staining based on comigration with ceramide standards (Type III, Sigma), and quantified by liquid scintillation counting. The velocity of the reaction was linear for at least 2 h, and the amount of palmitoyl-coenzyme A consumed did not exceed 5% of total. Ceramide synthase activity was unaffected by CHAPS (0.5–4%) or Triton X-100 (0.05–1%), except at the highest concentrations where both detergents attenuated activity 20–30%.

Lipid analysis

Ceramide mass was quantified by the diacylglycerol kinase assay as described [32]. Cholesterol was quantified using Amplex Red Cholesterol Assay Kit according to manufacturer's instruction (Invitrogen). Neutral sphingolipids and gangliosides were analyzed by HPTLC or immune-HPTLC as described [40] with modification. Sphingomyelin was analyzed according to Mallikarjuneswara et al. [41]. For these studies, 15.3 µg MCRM was extracted with CHCl₃∶MeOH∶HCl (100∶100∶1 v/v/v), and the extract was washed once with an equal volume of pre-equilibrated upper phase, followed by resolution on Silica gel 60 Å HPTLC plates (Whatman) using CHCl₃∶MeOH∶H₂0 (65∶25∶4) as solvent. Sphingomyelin was visualized with 7% phosphomolybdic acid in ethanol (Sigma) followed by charring, and quantified by comparison to a concomitantly-run standard curve of known quantities of sphingomyelin.

Cellular ceramide localization by confocal microscopy

HeLa cells, grown on coverslips in 6-well plates, were washed 3 times with cold PBS, and fixed with 4% Formalin in PBS for 15 min. Cells were then permeabilized by incubation with-20°C acetone/methanol (1∶1) for 10 min after 3 times of washing with PBS. After blocking with 3% goat serum and 3% FBS diluted in PBS for 16 h at 4°C, cells were incubated with anti-CoxI-Alexa 488 conjugated (Invitrogen; 1∶75) diluted in cold wash buffer (1% FBS, 0.025% Tween20 in PBS) at 22°C for 45 min. Cells were then washed 3 times for 5 min with cold wash buffer followed by incubation with monoclonal anti–ceramide (Alexis; 1∶30) at 22°C for 45 min. After washing 3 times with cold wash buffer, cells were incubated with Cy3-conjugated goat anti-mouse IgM (Jackson ImmunoResearch; 1∶100) at 22°C. After 45 min, cells were washed 3 times with cold wash buffer, and mounted on slides using the ProLong Antifade Kit (Invitrogen). When indicated, Mitotracker Red CMXRos (50 nM) (Invitrogen) was added to the culture medium at 37°C for 30 min to stain mitochondria before fixation. Images were acquired with a Leica TCS SP2 AOBS confocal microscope equipped with a 63×1.4 NA OIL DIC D Objective combined with 4× scan zoom. Co-localization analysis was performed with MetaMorph 7.5 software. For ER staining, anti-KDEL (Abcam; 1∶50) and anti–ceramide (Alexis; 1∶20) were employed in combination with Texas red-conjugated goat anti-mouse IgG (Jackson ImmunoResearch; 1∶100) and Cy2-conjugated goat anti-mouse IgM (Jackson ImmunoResearch; 1∶100), respectively.

Preparation of recombinant full length Bax and BaxΔC

Purified recombinant full length Bax was kindly provided by Dr. Richard Youle [see [42]]. Mouse BaxΔC with or without a GST tag was prepared as described [16]. Briefly, 50 ml of an overnight culture of E. coli BL21 expressing pGEX-4T-2 BaxΔC was inoculated into 1000 ml of LB-medium containing 50 µg/ml of ampicillin and grown at 37°C to an A₆₀₀ 0.6–0.8. After addition of 0.1 mM IPTG, cells were grown for 2 h at 25°C and collected by centrifugation. Cell pellets were resuspended in 30 ml lysis buffer (50 mM Tris pH 8.0, 150 mM NaCl, 1% Tween 20, 0.1% 2-mercaptoethanol, 5 mM EDTA, 1 mM PMSF, protease inhibitor cocktail) and disrupted by sonication. The cell lysate was centrifuged at 28,000×g for 20 min at 4°C. GST-BaxΔC was purified using the Bulk GST Purification Kit (Amersham). GST tag was removed from the recombinant GST–BaxΔC proteins bound to glutathione-Sepharose 4B resin by thrombin (10 units, Amersham) according to manufacturer's instructions. Released BaxΔC protein was purified on a Mono-Q Sepharose column using a linear gradient of 0.5 M NaCl, pH 8.0 and dialyzed in dilution buffer (20 mM HEPES pH 7.5, 100 mM KCl, 20 mM MgCl₂, 1 mM EDTA) for 16 h at 4°C. Ultrafree Centrifuge Filter Devices (Millipore) were used to concentrate purified protein. Recombinant BaxΔC was stored in dilution buffer.

Preparation of C₁₆-ceramide and C₁₆-dihydroceramide

C₁₆-ceramide (Biomol) or C₁₆-dihydroceramide (Toronto Research Chemical Inc.) was dissolved in ethanol∶dodecane (98∶2, v/v) at 37°C with occasional vortexing, and further diluted to a 1% final solvent concentration with experimental reaction buffers.

Bax insertion into and cytochrome c release from HeLa cell mitochondria ex vivo

Bax or tBid (R&D Systems, Inc.) in Bax dilution buffer [20 mM HEPES pH 7.4, 10 mM KCl, 20 mM MgCl₂, 1 mM EDTA] was incubated with 50 µg of the P₁₀-mitochondria-enriched fraction in the presence or absence of C₁₆-ceramide in 50 µl MSB buffer [400 mM mannitol, 50 mM Tris-HCl pH 7.2, 4 mM MgCl₂, 10 mM KH₂PO₄, 50 µM rotenone, 5 mM succinate, 5 mg/ml fatty acid free bovine serum albumin (BSA; Sigma)] for 30 min or 1 h at 30°C. Reaction samples were centrifuged at 14,000×g for 5 min, and supernatants and mitochondrial pellets were analyzed for cytochrome c release. For measurement of Bax insertion, mitochondrial pellets were subjected to alkali extraction as described above and alkaline-sensitive and alkaline-resistant fractions were analyzed by immunoblotting using anti-Bax or anti-GST antibodies.

Isolation of mitochondria from mouse liver

Mouse liver mitochondria were isolated as described [16] with modification. Livers were surgically removed from 8 week old C57BL/6 black mice (Jackson Laboratory, Bar Harbor, ME) sacrificed by CO₂ asphyxiation under protocol 09-08-016 in compliance with the guidelines of the Institutional Animal Care and Use Committee (IACUC) of MSKCC. Briefly, isolated liver tissue was washed twice and minced in ice-cold Buffer A (0.25 M sucrose, 10 mM HEPES pH 7.4, 0.5 mM EGTA). Minced tissue was then re-washed in Buffer A and homogenized (0.3 g of liver/ml of buffer) on ice for 3 min using a glass homogenizer with a loose-fitting pestle. Homogenates were centrifuged at 600×g for 15 min, and the resulting pellets containing nuclei and undisrupted cells were discarded. Supernatants were centrifuged at 9750×g for 15 min, and the resulting pellet was resuspended in Buffer A and centrifuged at 9750×g for 10 min. The pellet was washed 3 more times in Buffer A and once in Buffer B (0.25 M sucrose, 10 mM HEPES pH 7.4). The final mitochondrial pellet was resuspended in Buffer B at 50 mg protein/ml. All steps were carried out at 0–4°C. Mice were housed at the animal core facility of Memorial Sloan-Kettering Cancer Center. This facility is approved by the American Association for Accreditation of Laboratory Animal Care and is maintained in accordance with the regulations and standards of the United States Department of Agriculture and the Department of Health and Human Services, National Institutes of Health. The Institutional Animal Care and Use Committee at Memorial Sloan-Kettering Cancer Center approved these mice experiments under Protocol # 09-08-016.

Uptake of C₁₆-ceramide or C₁₆-dihydroceramide by isolated mouse liver mitochondria

Isolated mitochondria (1 µg/µl) were incubated with C₁₆-ceramide or C₁₆-dihydroceramide (0–5 µM; 1% final solvent concentration) in a KCl-based medium [20 mM HEPES, pH 7.4, 150 mM KCl, 25 mM NaHCO₃, 1 mM MgCl₂, 3 mM NaH₂PO₄, 250 mM sucrose, 1 mM glutamate/malate, 5 mg/ml fatty acid free BSA] for 5 min at 37°C. Thereafter, samples were centrifuged at 10,000×g for 10 min at 4°C, the mitochondrial pellet was resuspended in 1×PBS buffer, and ceramide levels in mitochondria (600 µg) were quantified by the diacylglycerol kinase assay.

Formation of MCRMs within and cytochrome c release from mouse liver mitochondria ex vivo

Recombinant full length Bax or BaxΔC in Bax dilution buffer was incubated with or without C₁₆-ceramide with isolated mitochondria (1 mg/ml) in KCl-based medium for 5 or 15 min at 37°C. Reaction samples were centrifuged at 14,000×g for 5 min at 4°C and the supernatant containing proteins released from isolated mitochondria and pellet containing retained mitochondrial proteins were analyzed for cytochrome c. For MCRM isolation, the mitochondrial pellet was subjected to 5–30% continuous sucrose density gradient after incubation with 0.05% Triton X-100 as below.

Isolation of MCRM by sucrose density gradient

Isolation of HeLa and mouse liver MCRM was based on concepts derived from isolation of detergent-insoluble raft-like plasma membrane macrodomains [43]. Identification of HeLa and mouse liver MCRM initially involved extensive detergent dose titration. Mitochondria (3.3 mg/ml), isolated in SHE buffer, were incubated on ice for 30 min with Triton X-100 (0.01–1%) or CHAPS (0.1–2%), and solubilized mitochondrial proteins were separated from insoluble proteins by centrifugation at 14,000×g for 5 min at 4°C. Extent of solubilization of the inner mitochondrial membrane protein COXIV was confirmed by Western blot. The lowest concentrations that induced complete COXIV solubilization, 0.05% Triton X-100 for mouse liver mitochondria and 0.15% Triton X-100 or 1% CHAPS for HeLa mitochondria, were selected for MCRM isolation. For these studies, mitochondria (3.3 µg/µl) were treated on ice with detergent for 30 min in 1×MBS buffer (25 mM MES pH 6.5, 150 mM NaCl), followed by homogenization with 20 strokes of a loose-fitting dounce homogenizer or by sonication (3×10 sec). The homogenate was adjusted with 80% sucrose in MBS buffer to 40% final sucrose concentration and transferred to a 12 ml ultracentrifuge tube. A continuous gradient consisting of 4 ml of 5% and 6 ml of 30% sucrose, or a discontinuous gradient consisting of 5 ml of 5% and 5 ml of 30% sucrose was layered on top of the homogenate, and samples were ultracentrifuged at 182,000×g for 16 h in Beckman SW41 rotor at 4°C. As with isolation of plasma membrane platforms, this procedure yielded an opaque detergent insoluble band at fraction 5–7 of the gradient. One ml fractions were collected from atop the gradient and profiled for proteins by Western blotting after 20% TCA precipitation, and for lipids as in Experimental Procedures. For mini-discontinuous sucrose gradient, 80 µl of homogenates adjusted to 40% sucrose were transferred to an ultracentrifuge tube and overlaid successively with 80 µl of 30% sucrose and 80 µl of 5% sucrose. Samples were centrifuged at 182,000×g for 18 h in Beckman TLA 100 rotor at 4°C, and three 80 µl fractions were collected from the top. Pellets were resuspended in 80 µl MBS by brief sonication.

Staining and visualization of MCRMs

MitoTracker was used to outline mitochondrial anatomical boundaries for ease of co-localization, and hence cannot bestow an MCRM artifact. In studies examining MCRM formation upon radiation in HeLa cells, mitochondria were stained by incubating live cells with 100 nM MitoTracker Deep Red 633 for 30 min at 37°C as per manufacturer instructions before mitochondrial isolation. However, for ex vivo studies with isolated mitochondria examining the effects of ceramide and Bax on MCRM formation, the isolated mitochondria were stained either before or after fixation. When staining live mitochondria prior to fixation, mitochondria were pre-incubated with 200 nM MitoTracker Deep Red 633 in 2.5% FBS and 2.5% goat serum in PBS for 1 h on ice. MitoTracker is an organic dye with a charge that directs its binding to membranes with a potential, i.e. live mitochondria. However, its hydrophobic nature also permits binding membranes lacking a potential after fixation, if incubation is prolonged. Therefore, the fixed staining approach was used for the majority of ex vivo experiments for ease of experimental procedure. For staining of fixed mitochondria, mitochondria were fixed for 45 min with 2% formaldehyde, pelleted by centrifugation at 6000×g for 5 min, followed by brief washing with PBS, and incubation with 200 nM MitoTracker Deep Red 633 in 2.5% FBS and 2.5% goat serum in PBS for 16 h at 4°C. Mitochondria, stained with MitoTracker, were washed once in Wash Buffer (1% FBS, 0.025% Tween 20 in 1×PBS), and incubated with anti-ceramide (Alexis, MID15B4; 1∶30) and/or anti-Bax (6A7; 1∶50, Calbiochem or Δ21; 1∶10, Santa Cruz) antibodies in Wash Buffer for 45 min. Mitochondria were then washed, followed by incubation with Cy3-conjugated goat anti-mouse IgM (Jackson ImmunoResearch; 1∶100) and/or Cy2-conjugated goat anti-mouse IgG (Jackson ImmunoResearch; 1∶100) or Alexa Fluor 488-conjugated goat anti-rabbit IgG (Molecular Probes; 1∶100) secondary antibodies, respectively. After 45 min, mitochondria were washed 3 times, and then once in 0.025% Tween 20 in PBS. All staining steps were at 4°C. After staining, mitochondrial pellets were resuspended in 20 µl 0.025% Tween 20 in PBS. 5 µl of mitochondrial suspension was spread on slide using a cover slip and mounted using the ProLong Gold Antifade Reagent (Invitrogen). Images were acquired with a Leica TCS AOBS SP2 confocal microscope equipped with a 100×1.4 NA OIL DIC D Objective combined with 2× scan zoom. Co-localization analysis was performed with MetaMorph 7.5 software.

Statistical Analysis

Statistical analysis was performed by Student's t-test or linear regression analysis with 95% confidence estimation for ED₅₀.

Mitochondrial ceramide elevation precedes Bax activation in irradiated HeLa cells

Morphologic evidence of apoptosis was detected in 5–15 Gy irradiated HeLa cells by 36 h and became significant at 48 h (Figure 1A; p<0.001 vs. control). Increased cytosolic effector caspase 3 activity, measured fluorometrically by cleavage of the substrate Z-DEVD-AFC, was similarly detected by 30 h post-irradiation, peaking at 36 h, remaining elevated for at least 72 h after 10 Gy (see Figure 2E for 36 h data). Cytochrome c release from mitochondria into cytosol was detected at 30–32 h after 10 Gy by Western analysis (Figure 1B) and was maximal at 36 h. Consistent with the cytochrome c release profile, Bax insertion into the MOM also began at 30 h, peaking at 34 h after 10 Gy (Figure S1A). In agreement with previous studies [44], Bax insertion occurred without apparent translocation from cytosol (not shown), but rather from a Bax pool loosely attached to unirradiated HeLa mitochondria but not inserted into MOM [37], [45]. This pool, which is readily removed by 100 mM Na₂CO₃ washing and does not spontaneously induce MOMP [45], [46], represents the large majority of endogenous Bax co-isolated with HeLa mitochondria (estimated at ∼80% in our preparations based on semi-quantitative densitometry normalized for protein recovery; n = 8). At 34 h post-irradiation, the fraction of Bax inserted into the MOM had increased 4–5-fold (Figure S1A, p<0.001 vs. control unirradiated, n = 6). Bax multimerization into low molecular weight (not shown) and higher molecular weight oligomers (M_r 360 KD) occurred concomitant with Bax insertion (Figure S1B). Whereas mitochondrial Bax was primarily monomeric in unirradiated HeLa cells (n = 4), a majority of Bax (estimated at ∼60% of the total by densitometry) redistributed into high molecular weight oligomeric fractions post-irradiation (Figure S1B). In contrast, Bak and VDAC were found in both fractions in unirradiated controls, and as previously reported [37] redistributed minimally upon irradiation (Figure S1B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Ionizing radiation-induced ceramide elevation is confined to mitochondria.

(A) Time- and dose-dependent induction of apoptosis in response to ionizing radiation. Morphologic changes of nuclear apoptosis were detected using the DNA-specific fluorochrome bis-benzimide. Data (mean±SEM) are collated from 3 experiments analyzing 500 cells per point. (B) Time-dependent cytochrome c (Cyt. c) release after 10 Gy. Cytosolic fractions of HeLa cells, collected at the indicated times post-irradiation were analyzed by immunoblotting using α-Cyt. c. Data are from 1 representative of 3 studies. (C) Increased CS activity in isolated mitochondria (MT) after 20 Gy. CS activity was measured at 28 h post-irradiation. Data are from 2 experiments. (D) Radiation increases mitochondrial ceramide content. At 33 h post-irradiation, ceramide was quantified by the diacylglycerol kinase assay in mitochondrial-enriched and ER-enriched fractions from HeLa cells. Inset: The ER-enriched fraction was devoid of mitochondrial contamination based on western blotting with anti-COXII (mitochondrial marker), while the mitochondrial fraction was 10.9±1.7% (mean±SE) ER based on anti-Climp63 (ER marker) blotting. Data (mean±SE) are from 4 experiments performed in triplicate. *, p<0.05 vs. control. (E) Ionizing radiation increases CS activity in MAM and ceramide levels in MAM-free mitochondria (MT). HeLa cells were harvested 33 h post-10 Gy and organelles isolated as in Materials and Methods. ER (P100) was fractionated by differential centrifugation and mitochondria and MAM within the heavy membrane fraction (P10) were further separated from each other by 30% Percoll gradient. Upper panel; CS activity was measured in each fraction using sphinganine and palmitoyl-CoA as substrates as in Materials and Methods. Middle panel: purity of ER, Mitochondria and MAM fractions was analyzed by immunoblotting with antibodies to Calnexin (ER marker) and COXII (MT marker). Based on anti-Calnexin blotting, Percoll-purified mitochondria were 3–4% contaminated with ER. Lower panel; Ceramide levels were quantified by diacylglycerol kinase assay as in Materials and Methods. (F) Co-localization of mitochondrial ceramide and COXI after 20 Gy. At 33 h post-irradiation, HeLa cells were stained with anti-ceramide IgM (red) and anti-COXI IgG (green). Images were acquired with a Leica TCS AOBS SP2 confocal microscope equipped with a 63×1.4NA OIL DIC D objective combined with 4× scan zoom, and co-localization (yellow) was analyzed with MetaMorph 7.5 software. Scale bar; 10 µm. Inset images (rectangles) represent 4× magnification of left upper (0 Gy) and left lower (20 Gy) regions for observation of co-localization. Data are from 1 of 5 experiments.

https://doi.org/10.1371/journal.pone.0019783.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. FB₁ prevents radiation-induced MOMP in HeLa cells.

(A) FB₁ blocks mitochondrial ceramide generation. HeLa cells were irradiated (IR) and treated with 15 µM FB₁ 20 h post-irradiation. Ceramide in isolated mitochondria was quantified by diacylglycerol kinase assay at 36 h post-irradiation. Data (mean±SEM) are from 2 experiments performed in triplicate. *, p<0.05 vs. control; **, p<0.01 vs. irradiated. (B) FB₁ prevents radiation-induced Bax insertion into the MOM. Alkali-resistant mitochondrial fractions containing inserted Bax were isolated after 34 h from HeLa cells irradiated with 20 Gy and treated with 25 µM FB₁ or 75 nM ISP-1 at 20 h post-irradiation. COXII was used as mitochondrial loading control. Data are from 1 of 4 studies. (C) FB₁ blocks radiation-induced Bax oligomerization. At 34 h post-irradiation, mitochondrial proteins from HeLa cells treated as in (B) were separated by gel filtration. Data are from 2 studies. (D) FB₁ attenuates radiation-induced cytochrome c release in HeLa cells and BAEC. HeLa cells were irradiated with 10 Gy, and 15 µM FB₁ was added 20 h post-irradiation. BAEC cells were treated with 25 µM FB₁ 1 h before irradiation with 5 Gy. 36 h (HeLa) and 12 h (BAEC) post-irradiation, cytosolic fractions were analyzed by immunoblotting using mouse monoclonal anti-Cyt. c and mouse monoclonal anti-tubulin antibodies. Data are from 1 of 2 studies in HeLa and BAEC each. (E) FB₁ attenuates radiation-induced caspase activity. FB₁ was added to cells 20 h after 10 Gy and caspase activity measured at 36 h post-irradiation using the fluorogenic caspase substrate Z-DEVD-AFC. Data (mean±SEM) are from 1 of 2 investigations performed in triplicate.

https://doi.org/10.1371/journal.pone.0019783.g002

Ceramide elevation preceded Bax activation. Ceramide levels increased up to 2.6-fold of baseline (350 pmols/10⁶ cells) beginning at 28 h post-20 Gy and remained elevated until at least 36 h post irradiation (not shown; p<0.005 at all times vs. control). Whereas sphingomyelinase activity was not elevated post-irradiation, similar to published data [47], CS activity increased in the post-nuclear membrane fraction (P₁₀₀) at 28 h in the current studies and remained elevated until at least 32 h [47]. Ionizing radiation increased CS maximal reaction velocity (V_max) of the mitochondrial-enriched fraction specifically, as determined by Michaelis-Menten kinetics (Figure 1C), whereas ER CS activity was not increased (see Figure 1E). Consistent with this observation, the cellular ceramide increase upon irradiation appeared confined to the mitochondrial-enriched fraction, with no change in irradiated ER (Figure 1D).

Previous studies showed mitochondria prepared by differential centrifugation on discontinuous sucrose gradients co-purify with mitochondrial-associated membrane (MAM), an ER-like structure that links to mitochondria via specific MOM contact sites and acts as a conduit for transfer of ER and/or MAM metabolites into mitochondria [48]. We, therefore, separated MAM from mitochondria to further explore the site of radiation-induced CS activity. Precedence exists for transfer of lipids from MAM to MOM as phosphatidylserine (PS), synthesized exclusively in MAM via PS synthases, is transported to MOM, then the inner mitochondrial membrane, where it serves as the sole source of mitochondrial phosphatidylethanolamine, via decarboxylation [49], [50]. Here we show that HeLa cell MAM, but not ER or mitochondria, is the exclusive site of radiation-induced CS activation (Figure 1E, upper panel) with apparent subsequent transfer of ceramide to mitochondria (Figure 1E, lower panel), as 75% of the ceramide mass increase is observed within MAM-free mitochondria. Similar transfer of MAM ceramide to MOM was recently reported by Colombini and co-workers [51]. In contrast, the level of total mitochondrial phospholipid and of sphingosine-1-phosphate, a ceramide metabolite with second messenger function, remained unchanged (not shown). A similar high baseline ceramide concentration and stimulated elevation were recently detected by mass spectroscopy in MAM-free mitochondria isolated from untreated and phorbol ester-treated MCF-7 breast cancer cells [52].

Further evidence for preferential ceramide elevation in mitochondria post-irradiation was derived from confocal microscopic double immunostaining using anti-ceramide IgM, and anti-COXI IgG as mitochondrial marker (Figure 1F). The anti-ceramide antibody used is highly specific, distinguishing C₁₆-ceramide from C₁₆-dihydroceramide (Figure S2), which differs from ceramide by a double bond in the sphingoid base backbone. While only 18.4±2.7% (mean±SE) of ceramide co-localized with mitochondria (Figure 1F, upper panels, yellow merged signal) in unirradiated cells, the level of ceramide co-localized to mitochondria increased to 31.2±3.3% (Figure 1F, bottom panels) (p<0.001; n = 20) at 33 h post-irradiation. Similar co-localization was displayed with Mitotracker (not shown). Alternately, no radiation-induced co-localization of anti-ceramide with ER-marker (anti-KDEL) was detected (Figure S3). Taken together, these studies indicate that ceramide increases preferentially within HeLa cell mitochondria during evolution of radiation-induced apoptosis.

The natural CS inhibitor, Fumonisin B₁, inhibits MOMP

Fumonisin B₁ (FB₁) is a natural competitive CS inhibitor [53] with an IC₅₀ two orders of magnitude left-shifted to the K_m for sphingoid base acylation [53]. A short treatment of HeLa cells with 15 µM FB₁, beginning 20 h post-irradiation, did not affect basal ceramide levels but abolished radiation-induced mitochondrial ceramide elevation (Figure 2A). FB₁ also reduced Bax insertion into the MOM (Figure 2B). In five experiments, FB₁ inhibited Bax insertion by 83±5% (p<0.01 vs. irradiated) and nearly completely inhibited formation of high molecular weight Bax oligomers (Figure 2C). Inhibition of radiation-induced Bax insertion and oligomerization by FB₁ reduced cytochrome c release into cytosol by 72% (Figure 2D, left panel). Similar reduction was observed in bovine aortic endothelial cells (BAEC) (Figure 2D, right panel), previously reported to undergo radiation-induced CS activation [54]. Consequently, FB₁ also attenuated caspase 3 activation (67% at 30 µM; Figure 2E) and inhibited apoptosis assessed by bis-benzimide staining in HeLa cells (not shown) and BAEC [54]. Similar results were obtained using ISP-1 (Figure 2B and not shown), an inhibitor of serine palmitoyl transferase, the enzyme catalyzing synthesis of the sphingoid base substrate for de novo ceramide synthesis via CS [55]. These data define a role for ceramide generated via CS in the mitochondrial phase of radiation-induced apoptosis in HeLa cells.

C₁₆-ceramide activates Bax within isolated HeLa mitochondria

To provide direct evidence that ceramide effects Bax-mediated MOMP, isolated HeLa mitochondria were treated with natural C₁₆-ceramide (0.05–1 µM), resulting in dose-dependent cytochrome c release (Figure 3A). In contrast, treatment with anti-apoptotic sphingosine-1-phosphate (0.05–1 µM) failed to induce cytochrome c release under the same conditions (not shown). 1 µM C₁₆-ceramide also rendered the endogenous Bax known to co-purify with HeLa mitochondria (termed attached Bax; [37], [46]) resistant to alkali extraction (Figure 3B), indicating attached Bax had inserted into the MOM. The ED₅₀ for cytochrome c release of approximately 0.2 µM C₁₆-ceramide (Figure 3A) and peak response at 0.5 µM (Figure 3A) were right-shifted from the ED₅₀ for Bax insertion (approximately 0.05 µM C₁₆-ceramide with maximal insertion at 0.12 µM) (Figure S4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Effect of C₁₆-ceramide on MOMP in isolated HeLa mitochondria.

(A) Ceramide induces cytochrome c release from isolated HeLa mitochondria. C₁₆-ceramide (0–1 µM) was incubated with HeLa mitochondria (1 µg/µl) in MSB buffer. After 1 h at 30°C, samples were centrifuged at 14,000×g for 5 min at 4°C to separate released (supernatant) and retained (pellet) mitochondrial proteins, and analyzed for cytochrome c release by immunoblotting using anti-Cyt.c and anti-COXII as loading control. Data are from 1 of 4 investigations. (B) Ceramide induces insertion of endogenous Bax into HeLa mitochondrial membranes. Isolated mitochondria were incubated with 1 µM C₁₆-ceramide and mitochondrial pellets were collected after incubation as in (A). Attached and inserted Bax were separated by alkali extraction of mitochondrial pellets as in Figure 2B and analyzed by Western blot with anti-Bax and anti-COXII as loading control. Data are from 1 of 4 investigations. (C) FB₁ inhibits tBid-induced cytochrome c release. Isolated HeLa mitochondria were incubated with 0.25–12.5 ng of caspase-8 cleaved human Bid for 30 min and cytochrome c release was analyzed as in (A). Data represent one of three similar studies. (D) FB₁ inhibits BaxΔC-induced cytochrome c release. HeLa mitochondria, replete (control) or depleted of ceramide (from 35 µM FB₁-pretreated cells), were incubated with BaxΔC (0–1 µM) for 30 min and cytochrome c release was analyzed as in (A). The left panel shows the cytochrome c content of mitochondria isolated from FB₁-pretreated cells was not different than that from untreated HeLa cells. Data are from 1 of 5 investigations.

https://doi.org/10.1371/journal.pone.0019783.g003

To provide additional evidence for ceramide function in Bax insertion, HeLa cells were pre-treated for a prolonged time (24 h) with high FB₁ doses (35–50 µM), reducing baseline mitochondrial ceramide content from 250±40 to 80±40 pmols/0.05 mg mitochondrial protein (mean±SD) without affecting mitochondrial cytochrome c content (Figure 3D, left panel). Sensitivity of ceramide-depleted HeLa mitochondria towards tBid-induced cytochrome c release, which depends on insertion/activation of the endogenous Bax associated with isolated mitochondria [44], was markedly reduced (Figure 3C). Consistent with published literature [44], tBid-induced dose-dependent insertion of attached endogenous Bax into the MOM (not shown) and cytochrome c release, which was maximal at 2.5 ng tBid/50 µg mitochondrial protein (Figure 3C, left). tBid-induced cytochrome c release from ceramide-depleted mitochondria, however, was attenuated at least 5-fold (Figure 3C, right), accompanied by inhibition of Bax insertion (not shown).

As there are two distinct events in the Bax activation process in which lipids direct Bax insertional activation, the initial tethering of Bax to the MOM via the α9 C-terminal helical domain and the activation of the lipidic pore through the α5-α6 hairpin, studies were designed to evaluate ceramide impact specifically on the lipidic pore. For these studies we employed BaxΔC, which lacks amino acids 172–192 including the C-terminal TM domain [56]. Similar to what we observed with tBid, cytochrome c induced release by BaxΔC was reduced >1 log in ceramide-depleted mitochondria compared to control HeLa mitochondria (Figure 3D). Ceramide depletion also blocked Bax insertion into the MOM of mitochondria isolated from FB₁-treated BAEC (not shown), and prevented cytochrome c release therefrom (Figure S5, lower panel). Together these data indicate ceramide is a significant regulator of Bax-mediated MOMP.

Ceramide induces formation of a mitochondrial ceramide-rich macrodomain (MCRM)

The above studies are consistent with either mitochondrial ceramide acting, like Bax, to directly release cytochrome c, or alternatively as a facilitator of Bax action. To resolve this issue, we employed mouse liver mitochondrial preparations, which, in contrast to HeLa mitochondria, are isolated without attached Bax [46] (Figure S6). Hence, they can serve as a model to examine autonomous ceramide action. Treatment of mouse hepatic mitochondria with recombinant BaxΔC induced concentration-dependent cytochrome c release (Figure 4A, upper panel). As little as 0.05 µM BaxΔC was effective, and 1 µM BaxΔC was maximal, and equivalent to 150 µM CaCl₂, a maximally-effective concentration for permeability transition-induced release [46]. Neither stimulus impacted the content of the MOM protein VDAC (Figure 4A, upper panel, lower lane) or the soluble matrix protein Hsp-60 (Figure S7), indicating specificity of the cytochrome c release process. In contrast, C₁₆-ceramide alone had no impact on cytochrome c release up to 50 µM (Figure 4A, lower panel), consistent with ceramide not functioning as a non-specific “detergent” effecting cytochrome c release [note ceramide is a non-swelling amphiphile and hence by definition not a detergent [57]]. Furthermore, when C₁₆-ceramide was added to mouse hepatic mitochondria prior to 0.05 µM BaxΔC, a minimally-effective dose for BaxΔC alone, cytochrome c release was dramatically enhanced. Combination of 1–5 µM C₁₆-ceramide and 0.05 µM BaxΔC yielded cytochrome c-releasing capacity equivalent to 1 µM BaxΔC alone, a dose-modifying factor of 20-fold. This effect appeared specific for ceramide, as the ceramide precursor C₁₆-dihydroceramide was ineffective (Figure S8). Further, the profile of proteins released from mitochondria in response to 50 nM BaxΔC plus C₁₆-ceramide was nearly identical to that released by 1 µM BaxΔC alone measured by Coomassie blue staining (Figure S9) and confirmed by mass spectrometry, consistent with ceramide acting via the Bax pore. Similar results were obtained combining C₁₆-ceramide with full-length recombinant Bax albeit with lower efficacy for cytochrome c release (not shown), confirming published data using isolated rat liver mitochondria [15].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Ceramide is non-apoptogenic by itself and facilitates Bax-induced MOMP.

(A) Ceramide facilitates Bax-induced cytochrome c release from isolated mouse liver mitochondria ex vivo. Recombinant BaxΔC (0–1 µM, upper panel), or C₁₆-ceramide (0–50 µM) plus 0.05 µM BaxΔC (lower panel) were incubated with isolated mouse liver mitochondria (1 µg/µl) in KCl buffer for 5 min at 37°C, supernatants and pellets were separated by centrifugation at 14,000×g for 5 min at 4°C, and cytochrome c release was measured as in Figure 3A. The outer mitochondrial membrane protein VDAC was used as loading control. Data are from 3 studies. (B) Ceramide induces insertion of recombinant BaxΔC into isolated mouse liver mitochondria. Mitochondria were treated as in (A) and attached and inserted BaxΔC analyzed as in Figure 3B. Data are from 2 independent studies. (C) Full-length (FL) Bax co-localizes with MCRMs induced by exogenous C₁₆-ceramide. Mouse liver mitochondria were incubated in MSB-based medium in the presence or absence of 50 nM full-length Bax for 5 min at 30°C. Thereafter, 0 or 1 mM C₁₆-ceramide [1% final solvent concentration (ethanol∶dodecane, 98∶2 v/v)] was added to the mixture for an additional 10 min. Mitochondria were fixed and stained with MitoTracker (blue), and ceramide and Bax were localized using anti-ceramide IgM (red) or anti-Bax IgG (green), respectively. Images, acquired with a Leica TCS AOBS SP2 confocal microscope equipped with a 100×1.4NA OIL DIC D objective combined with 2× scan zoom, were analyzed with MetaMorph 7.5 software. Control IgM and IgG did not yield detectable signals (not shown). Scale bar; 1 µm. Data represent 1 of 3 similar studies. (D) BaxΔC co-localizes with MCRMs induced by exogenous C₁₆-ceramide. Experiments were performed as in (C) using 5 mM C₁₆-ceramide [1% final solvent concentration (ethanol∶dodecane, 98∶2 v/v)] in the presence or absence of 50 nM recombinant BaxΔC. Scale bar; 1 µm. Data represent 1 of 4 similar studies. (E) Biophysical isolation of mouse liver MCRMs. After incubation with 50 nM BaxΔC and 5 µM C₁₆-ceramide, mitochondria (3.3 mg/ml), pelleted as in Figure 4A were resuspended in cold MBS buffer containing 0.05% Triton X-100. After 30 min on ice, mitochondria were homogenized with 20 strokes of a loose-fitting dounce homogenizer. The mitochondrial homogenate was adjusted to 40% final sucrose concentration and subjected to 5–30% continuous sucrose density gradient as in Materials and Methods. 400 ml of each 1 ml fraction were concentrated by 20% TCA precipitation, and proteins were resolved on a 15% SDS-PAGE gel and identified by immunoblot analysis using antibodies to the indicated proteins. Ceramide was measured using 400 µl aliquots by diacylglycerol kinase assay. Data are from 3 independent studies.

https://doi.org/10.1371/journal.pone.0019783.g004

Enhanced BaxΔC function in the presence of 1–5 µM C₁₆-ceramide, which increased mitochondrial ceramide content only 2–3 fold (Figure S8), appeared to reflect both increased Bax insertion and effectiveness. Figure 4B shows that at 0.05 µM BaxΔC, half the mitochondrial-associated BaxΔC is inserted, and that C₁₆-ceramide confers complete insertion, effectively doubling the quantity of inserted BaxΔC. However, as the dose-modifying factor for cytochrome c release is 20-fold, these studies insinuate increased apoptogenic effectiveness of inserted Bax in the presence of ceramide. In contrast, C₁₆-dihydroceramide did not impact BaxΔC insertion (Figure S8). Hence, ceramide is not an autonomous apoptotic factor per se but rather serves to enhance mitochondrial Bax insertion and function.

To ascertain whether ceramide elevation in mitochondria results in formation of a ceramide-rich macrodomain into which Bax inserts and oligomerizes, we employed confocal microscopy using intact mitochondria isolated from mouse liver. Detection of MCRMs and co-localization of ceramide with Bax were performed using Cy3 and Cy2, probes that emit at 566 nm and 510 nm, respectively. Co-localization analysis of confocal microscopic images can be affected by the resolution limit, defined by the standard equation: 0.4×λ_em/NA = R (λ_em: emission wavelength, R: resolution, NA: numerical aperture; 1.4 in this study). Based on this equation, the calculated resolutions for Cy3 and Cy2 are 162 nm and 146 nm, respectively, readily permitting co-localization analysis using these probes in ≈1 µM-sized mitochondria.

After addition of C₁₆-ceramide, and/or recombinant full-length (FL) Bax (Figure 4C) or BaxΔC (Figure 4D), mitochondria were stained with MitoTracker (blue), and mitochondrial ceramide and Bax were localized using anti-ceramide IgM (red) or anti-Bax IgG (green), respectively. Control IgM and IgG did not yield detectable signals (see Figure 5A). Consistent with the Western analysis (Figure S6), Figures 4C and 4D show that endogenous Bax was not immunodetected in naïve mitochondria (top lanes), while barely detectable ceramide-containing domains were observed at low incidence (top lanes). Very little added recombinant full length Bax or BaxΔC co-localized with these small ceramide-rich domains (second lanes in Figures 4C and 4D, respectively). Addition of exogenous C₁₆-ceramide alone to naïve mitochondria resulted in formation of large MCRMs in 65–75% of mitochondria (third lanes and bottom lanes in Figures 4C and 4D), with recombinant full length Bax or BaxΔC co-localizing in ∼70% of these newly-formed platforms (identified as whitish-yellow merged signals in the bottom lanes). Further, more than half of the C₁₆-ceramide uptake, 60±5% of the total, went to manufacturing these newly-formed MCRMs. MCRM formation was not an artifact of fixation preceding staining as reported for the Forssman antigen on the apical surface of MDCKII cells [58], as reversing this order did not impact MCRM detection (Figure S10).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Ceramide induces formation of a mitochondrial ceramide-rich macrodomain (MCRM).

(A) Ionizing radiation (10 Gy) induces co-localization of endogenous Bax with MCRMs in HeLa cells. Mitochondria were isolated from HeLa cells 34 h after irradiation and immunostained as described in Supporting Information Text S1. Data represent typical stainings from 1 of 4 similar studies in which 2000 mitochondria were analyzed each. (B) Addition of exogenous C₁₆-ceramide induces co-localization of endogenous full-length Bax with MCRMs in HeLa cells. Mitochondria were isolated from HeLa cells using percoll gradient and treated with ceramide as Figure 3A. After 30 min incubation, mitochondria were fixed and stained with MitoTracker (blue), while ceramide and Bax were localized using anti-ceramide IgM (red) or anti-Bax IgG (green), respectively. Control IgM and IgG did not yield detectable signals (not shown). These data represent 1 of 3 similar studies. (C) Bax translocates into a radiation-generated HeLa MCRM. Upper panel: 34 h post-irradiation, HeLa mitochondria were isolated as in Materials and Methods and incubated with 0.15% Triton X-100 in MBS buffer for 30 min on ice. 40 µl mitochondrial homogenate (3.3 µg/µl) were subjected to 5–30% mini-discontinuous sucrose density gradient centrifugation as described in Materials and Methods. 20 µl aliquots of 80 µl fractions were analyzed by immunoblotting using the indicated antibodies. The protein level of each fraction was assessed using the Bio-Rad D_c protein assay kit (PE, Pellet). Data are from 1 of 4 studies, consisting of 2 independent gradients per study. The gradient shown displays our clearest example of Bax translocation into light membranes. Lower panel: Bax in each fraction, revealed by immunoblotting and quantified using NIH Image software, was normalized to protein content for all 8 gradients. (D) MCRM Bax exists as high molecular weight oligomers. Mitochondria from 10 Gy-irradiated HeLa cells, disrupted by either (a) 1% CHAPS and sonication or (b) dounce homogenization in 0.15% Triton X-100, were subjected to 5–30% discontinuous sucrose gradient for MCRM isolation as in Experimental Procedures. Light (MCRM; fractions 6,7) and heavy fractions (solubilized proteins; fractions 11,12) were analyzed by gel filtration on Sephacryl S-200 column as in Figure 2C. 500 µl of each eluted fraction were concentrated by 20% TCA precipitation for immunoblotting. Data are from 3 independent studies.

https://doi.org/10.1371/journal.pone.0019783.g005

We next isolated these MCRMs as a detergent-resistant light membrane fraction by continuous sucrose density flotation, a technique commonly employed to biophysically separate GEM platforms from bulk plasma membrane [18]. Figure 4E (middle panel, upper lane) confirms that Bax is absent from naïve mitochondria, and that upon incubation with BaxΔC, most of the BaxΔC (71%) resolves with the high-density bulk membrane-containing fraction (upper panel, upper lane, fractions 10–12). This compartment contains 57% of endogenous mitochondrial ceramide (upper panel, lower lane), and most of the MOM proteins VDAC (56%), Bak (89%), Bcl-xL (97%), and Metaxin (98%). Addition of 5 µM C₁₆-ceramide, which alone does not effect cytochrome c release, doubled mitochondrial ceramide content from 936 to 1812 pmol/5 mg mitochondrial protein, largely within light sucrose density membranes (middle/lower panels, lower lanes, fractions 1–9), consistent with the confocal data showing that the majority of ceramide added goes to formation of MCRMs. In contrast to studies using BaxΔC alone, upon co-addition of BaxΔC and ceramide, a combination that releases most mitochondrial cytochrome c (see Figure 4A), 68% of BaxΔC resolves with the MCRMs (lower panel, upper lane, fractions 1–9). C₁₆-dihydroceramide did not induce BaxΔC translocation into the light membrane fraction (Figure S10). A substantial portion of VDAC and a modest amount of Bak (lower panel, middle lanes), cholesterol (Figure S11) and various glycosphingolipids (not shown) also redistributed into this compartment upon C₁₆-ceramide treatment, while Bcl-xL and several other membrane proteins remained within the high-density fraction (Figure 4E and Figure S11). Taken together, these data indicate that two distinct approaches, biophysical isolation and confocal microscopy, identify a ceramide-rich macrodomain where Bax and other MOMP-inducing proteins associate. How MCRMs bestow Bax pore-forming capability is not addressed in this study, although the data are consistent with MCRMs constituting an essential element of the proteolipid Bax pore that appears to regulate MOMP [7], [9].

Ionizing radiation generates an endogenous MCRM in HeLa mitochondrial membranes

A similar approach was used to demonstrate MCRMs in irradiated HeLa mitochondria, which engage endogenous full-length Bax to mediate MOMP. While mitochondria isolated from unirradiated HeLa cells are decorated with attached but uninserted endogenous Bax (see Figure 3B, Figure S4), only 17±4% of these mitochondria manifest ceramide-rich domains, which tend to be small and contain minimal Bax by confocal microscopy (Figure 5A, upper lanes, n = 4). However, at 33 h post-10 Gy, MCRMs were detected in 41±4% of the total mitochondrial population (p<0.001). Up to 80% of the ceramide synthesized de novo (in 5 experiments 66±13% mean±SD) upon irradiation was utilized to construct these MCRMs. Hence MCRM ceramide rather than free ceramide would appear to be responsible for functionalizing Bax. Furthermore, the localization of Bax in MCRMs increased 14-fold after irradiation compared to that in the small ceramide-containing domains of control mitochondria (p<0.001), with 50% of the MCRMs displaying a Bax co-signal. Addition of 1 µM C₁₆-ceramide to generate MCRMs in mitochondria isolated from control unirradiated HeLa cells, which maximally releases cytochrome c (see Figure 3A), mimicked ionizing radiation conferring Bax co-localization (typical image depicted in Figure 5B) in 56% of HeLa MCRMs (p<0.001 vs. control 0 µM ceramide).

Consistent with these confocal observations, biophysical isolation of ceramide-rich domains by either discontinuous (identified as the 5% sucrose light membrane fraction; Figure 5C) or continuous (Figure S12) sucrose density flotation 34 h post-10 Gy revealed that 80.4±12.0% of the radiation-induced mitochondrial ceramide elevation reported in Figure 1E was confined to the HeLa light membrane fraction post-radiation (not shown). These data are consistent with results where exogenous C₁₆-ceramide added to isolated mouse liver mitochondria largely localized to this fraction (Figure 4E). Cholesterol was similarly concentrated in ceramide-rich light membranes post-radiation (Figure 5C). While a portion of Bak (20.2±1.1%) and VDAC (16.7±1.6%) constitutively associated with HeLa ceramide-rich light membranes, all other mitochondrial proteins tested, including Bcl-xL, were largely absent (Figure 5C and Figure S12). Little of the Bax (∼5–6%) associated with mitochondria in unirradiated HeLa cells resolved with ceramide-poor light membranes. However, after irradiation a substantial portion of the Bax pool (46%) localized to ceramide-rich light membranes. In eight experiments, we observed a 6.7±0.5 fold increase in light membrane-associated Bax (Figure 5C, lower panel; p<0.001). Thus two separate but complementary approaches, sucrose density floatation and confocal microscopy, confirm radiation-induced generation of mitochondrial MCRMs that coordinates Bax insertion/oligomerization required for MOMP.

HeLa cell MCRMs contain high molecular weight oligomerized endogenous Bax

Because Bax oligomerization is an essential element of MOMP induction, the multimerization status of endogenous Bax within MCRMs was assessed post-irradiation. To assure that Bax multimerization within MCRMs was not an artifact of membrane lipid isolation, MCRMs were initially isolated either using 1–2% CHAPS, which does not impact low molecular weight homodimerization [37], [59], or 0.15% Triton X-100, known to induce low molecular weight homodimerization [37], [59]. After MCRMs were isolated, the molecular weight profile of associated Bax was defined by size exclusion chromatography. Isolation in CHAPS (Figure 5D, upper panels) revealed that MCRMs contained high molecular weight Bax oligomers exclusively, while the heavy fraction contained Bax monomers. A similar profile was identified using 0.15% Triton X-100 (Figure 5D, lower panels), except the molecular weight of Bax was slightly higher in these heavy fractions consistent with Bax-Bax homotypic interaction, as reported when using Triton X-100 [37]. VDAC and Bak were also found in MCRM high molecular weight oligomers, and in both high and low molecular weight fractions of heavy membranes. These studies define the MCRM as a site of preferential Bax oligomerization.