Pseudomonas aeruginosa homoserine lactone triggers apoptosis and Bak/Bax‐independent release of mitochondrial cytochrome C in fibroblasts

Pseudomonas aeruginosa use N‐(3‐oxododecanoyl)‐homoserine lactone (C12) as a quorum‐sensing molecule to regulate gene expression in the bacteria. It is expected that in patients with chronic infections with P. aeruginosa, especially as biofilms, local [C12] will be high and, since C12 is lipid soluble, diffuse from the airways into the epithelium and underlying fibroblasts, capillary endothelia and white blood cells. Previous work showed that C12 has multiple effects in human host cells, including activation of apoptosis. The present work tested the involvement of Bak and Bax in C12‐triggered apoptosis in mouse embryo fibroblasts (MEF) by comparing MEF isolated from embryos of wild‐type (WT) and Bax−/−/Bak−/− (DKO) mice. In WT MEF C12 rapidly triggered (minutes to 2 h): activation of caspases 3/7 and 8, depolarization of mitochondrial membrane potential (Δψmito), release of cytochrome C from mitochondria into the cytosol, blebbing of plasma membranes, shrinkage/condensation of cells and nuclei and, subsequently, cell killing. A DKO MEF line that was relatively unaffected by the Bak/Bax‐dependent proapoptotic stimulants staurosporine and etoposide responded to C12 similarly to WT MEF: activation of caspase 3/7, depolarization of Δψmito and release of cytochrome C and cell death. Re‐expression of Bax or Bak in DKO MEF did not alter the WT‐like responses to C12 in DKO MEF. These data showed that C12 triggers novel, rapid proapoptotic Bak/Bax‐independent responses that include events commonly associated with activation of both the intrinsic pathway (depolarization of Δψmito and release of cytochrome C from mitochondria into the cytosol) and the extrinsic pathway (activation of caspase 8). Unlike the proapoptotic agonists staurosporine and etoposide that release cytochrome C from mitochondria, C12's effects do not require participation of either Bak or Bax.

The goal of the present study was to test whether Bak and Bax were involved in mediating apoptotic responses of mouse embryo fibroblasts (MEF) to C12. We used MEF for these studies for two reasons. First, fibroblasts are relevant biologically since it is expected that in patients with chronic infections with P. aeruginosa biofilms local [C12] will be high and, since C12 is lipid soluble, it will diffuse from the airways across and into the epithelium and underlying fibroblasts. Second, MEF are advantageous experimentally since MEF from wild-type (WT) and Bak −/− /Bax −/− (DKO) mice are readily available, and complete knockout of these genes in airway epithelia using siRNA or similar method is extremely difficult to achieve. Bak/Bax are commonly involved in apoptosis, particularly in the release of cytochrome C from mitochondria to the cytosol to trigger downstream activation of caspases 3/7 and 9 (Degenhardt et al., 2002;Mikhailov et al., 2003;Oakes et al., 2003;Scorrano et al., 2003;Wang et al., 2011). Schwarzer et al. (2012) previously showed that C12 rapidly triggered events commonly associated with activation of apoptosis (membrane blebbing, activation of caspases 3/7, 8 and 9; depolarization of mitochondria and release of cytochrome C). Recent experiments (Valentine et al., 2013a) that were reported while work for this article was in progress showed that C12 caused ZVAD-blocked cleavage and increased activity of caspase 3/7 and killing in MEF over 4-8 h. These experiments showed that these C12-triggered proapoptotic events were totally dependent on IRE1α and XBP-1 in MEF. The present experiments performed direct tests of the roles of Bak and Bax in C12-triggered apoptosis in MEF. We measured/assayed cellular and nuclear morphology, activities of caspases 3/7 and 8, cell killing and three different aspects of mitochondrial function/morphology in MEF from WT and DKO mice and also in DKO MEF that had been transduced to re-express Bak and/or Bax. The results indicated that C12 triggers multiple proapoptotic events in MEF, including release of cytochrome C from mitochondria into the cytosol, and Bak or Bax were not required for any of these effects. As far as we can tell, C12 is a unique bacterial product in that it is a membrane-permeant small molecule that triggers cytochrome C release from mitochondria without the involvement of Bak or Bax.

Effects of C12 on caspase 3/7 and cell morphology in WT MEF
We tested for C12-induced apoptosis in WT MEF by performing assays of caspases, membrane blebbing and cellular and nuclear shrinkage. In dose-response experiments conducted after 1 h exposure, C12 increased caspase 3/7 activity at 10 and 50 μM in WT MEF, but not at 1 μM (Fig. 1). Control, untreated cells showed typical, flattened morphology with oval nuclei and BCECF distributed throughout the cytosol and nuclei. 50 μM C12 caused WT cells and nuclei to shrink/condense, and plasma membrane had multiple blebs containing BCECF, indicating continuity with the cytosol. These effects were very prominent by 2 h of treatment (Fig. 2).

Effects of C12 on caspases 3/7 and 8 and cell killing in WT, DKO and Bak or Bax-corrected MEF
The roles of Bax and Bak in apoptotic responses were tested using MEF from WT and DKO mice (Wang et al., 2011). DKO MEF were also transduced with retrovirus expressing Bak or Bax as a further test of the role of Bak and Bax in the MEF responses to C12. We first tested for Bak and Bax expression in the cell lines. Western blots showed that WT MEF and the GFP-transduced MEF (used as control for transductions) and Bak-and/or Baxcorrected DKO MEF expressed Bak and Bax as predicted (Fig. 3). C12 caused activation of caspases 8 and 3/7 after 2 h in both WT and DKO MEF (Fig. 4). Further, re-expression of Bak or Bax in DKO MEF did not change caspase 3/7 responses to C12 (Fig. 5). Staurosporine, a commonly used proapoptotic agonist, was also tested to assure that the DKO and Bak/Bax-corrected DKO MEF were behaving as predicted . Staurosporine had no effect on caspase 3/7 in DKO MEF, but did activate caspase 3/7 activity in DKO + Bak and DKO + Bax MEF. We also tested whether the antiapoptotic Bcl-2 affected C12-induced apoptosis. WT MEF were transduced with Bcl-2 using a retroviral vector (Supplementary Data Fig. S1A). However, C12-induced activation of caspase 3/7 occurred similarly in WT MEF and in WT MEF expressing either GFP (as control for transduction) or BCl-2 (Supplementary Data Fig. S2B).
Measurements of cell death (uptake of propidium iodide, PI) after 24 h treatment with C12 were also used to test the roles of Bak/Bax. PI responses to the common proapoptotic agonist etoposide were used to assure WT and DKO and Bak/Bax-corrected DKO MEF were behaving as predicted. Preliminary dose-response experiments showed that 50 μM C12 caused roughly 20% killing, and 100 and 150 μM C12 caused equivalent 70-75% killing of WT MEF (Zhao and Li, data not shown). We used 100 μM MEF were grown on 12-well plates for 2 days, and growth medium was changed to Ringer's solution. C12 was added at the concentrations shown for 1 h, followed by preparation for measurement of caspase 3/7 activities in RLU. The threshold concentration for C12-triggered activation of caspase was between 1 and 10 μM. Average ± SD, n = 3 experiments. *P < 0.05 for comparisons of control versus C12.
C12 for cell killing experiments. Consistent with results measuring caspase 3/7 activity, C12 caused PI uptake (killing) of both WT and DKO MEF, and re-expression of Bak and/or Bax in DKO MEF did not change killing responses (Fig. 6A). Etoposide caused predictable cell killing in WT MEF and in DKO MEF that expressed Bak or Bax, but much less in DKO and DKO + GFP MEF (Fig. 6B). C12 also caused roughly equivalent, dosedependent killing of WT and Bax −/− /Bak −/− cells in another cell line, NIH3T3 MEF ( Fig. S2A and B). Overall, these data showed that C12 caused activation of caspase 3/7  Fluorescence images of blue Hoechst-stained nuclei and green BCECF-stained cytosol were overlain using Adobe Photoshop. Control, untreated cells showed typical, flattened morphology with oval nuclei and BCECF distributed throughout the cytosol and nuclei. Compared to control conditions, C12-treated WT cells and nuclei were obviously shrunken, and plasma membrane had multiple blebs (arrows) containing BCECF. Images typical of > 15 each in three different coverglasses each.   4. Effects of C12 on caspases 3/7 and 8 in WT and DKO MEF. WT and DKO MEF were exposed to 50 μM C12 for 2 h, then prepared for assays of caspase 8 (A) and caspase 3/7 (B). C12 activated caspases 3/7 and 8 in WT and in DKO MEF. Average ± SD, n = 3 experiments each. *P < 0.05 for comparisons between C12-treated versus control.

Testing roles of Bax and Bak in C12-induced depolarization of Δψmito and release of cytochrome C from mitochondria in MEF
Imaging of the voltage-sensitive dye JC-1 was used to test for Bak/Bax-dependent effects of C12 in mitochondria of MEF. JC1 accumulated in WT MEF in punctate patterns throughout the cytosol consistent with mitochondrial localization. Depolarization of Δψmito was assessed from the increase in fluorescence as JC1 was released from mitochondria into the cytosol and nucleus. In WT MEF C12 caused a slow increase of JC1 fluorescence over the course of 15 min (Fig. 7A), similar to effects in airway epithelial cells (Schwarzer et al., 2012). Complete depolarization of Δψmito was triggered by FCCP (10 μM) treatment at the end of each experiment (Fig. 7A). C12 also caused depolarization of Δψmito in DKO MEF (Fig. 7B) that was very similar to that in WT MEF (Fig. 7A). Averages from these time-course experiments are summarized in Fig. 8: C12 caused 70% depolarization of Δψmito (i.e. steady-state values in response to C12 compared to effects of FCCP 100%) in WT and DKO MEF.
We performed both immunofluorescence and ELISA assays to test whether C12's proapoptotic effects included Bak/Bax-dependent release of cytochrome C from mitochondria into the cytosol (Goldstein et al., 2005). Typical images of control and C12-treated MEF are shown in Fig. 9. Under control conditions WT MEF exhibited char-acteristic cytochrome C staining throughout the cytosol and typical oval nuclei with smooth boundaries that excluded cytochrome C. In control conditions DKO MEF exhibited less distinct and more punctate staining of cytochrome C-labelled mitochondria, consistent with the mitochondrial morphology described in previous work (Karbowski et al., 2006). Cytochrome C was largely Fig. 5. Effects of C12 on caspase 3/7 in WT and DKO MEF, and the (lack of) effect of Bak and Bax expression. WT and DKO MEF stably expressing GFP, Bak or Bax were treated with either 50 μM C12 or 10 μM staurosporine for 2 h, followed by measurement of caspase 3/7 activity. C12 activated caspase 3/7 in WT and in DKO MEF, whether or not they expressed GFP or Bak or Bax. Staurosporine activated caspase 3/7 in WT and in DKO MEF that expressed Bak or Bax, but not in DKO (parental) or DKO+GFP MEF. Caspase activities in RLU were normalized in each comparison to values observed in the absence of C12 (control). Data are average ± SD (n = 3 experiments). *P < 0.05 for comparisons between C12-treated or staurosporine-treated versus control.

Fig. 6.
Effects of C12 and etoposide on killing (PI uptake) in WT and DKO MEF and effects of Bak and Bax expression. WT and DKO MEF and DKO MEF transduced with GFP, Bak or Bax were treated with 100 μM C12 or 10 μM etoposide for 24 h, followed by treatment with PI and FACS analysis of PI uptake. A. C12 elicited PI uptake (compared to untreated controls) in WT and in DKO or DKO+GFP, DKO+Bak, DKO+Bax MEF. B. Etoposide increased PI uptake in WT and in DKO MEF that expressed Bak or Bax, but much less in DKO or DKO+GFP. Data in A and B are average ± SD (n = 3 experiments). *P < 0.05 for comparisons between C12-or etoposide-treated and controls. excluded from the nuclei of both WT and DKO MEF in control conditions. In WT and DKO MEF C12 caused cytochrome C staining to become diffuse in both the cytosol and nuclei (see arrows). Nuclei of C12-treated WT and DKO MEF were often shrunken and fragmented (arrowheads). These results indicated cytochrome C had been released from mitochondria into the cytosol and had diffused into the nuclei of WT and DKO MEF.
Release of cytochrome C from mitochondria into cytosol and nucleus of MEF was quantified from these and other immunofluorescence images of cytochrome C staining. We used images like those shown in Fig. 9 to measure the ratio of intensity of cytochrome C staining in the nucleus compared to that in the cytosol of WT and DKO MEF, i.e. [(nuclear intensity -background)/cytosol intensity -background)], in control conditions and during C12 treatment. As summarized in Fig. 10, C12 caused large and roughly equivalent increases in the nuclear/ cytosol distribution of cytochrome C in WT and DKO MEF. These data indicated that cytochrome C was being promi-nently released from mitochondria into the cytosol (and diffusion into the nuclei) of WT and DKO MEF.
ELISA assays for cytochrome C were performed on cytosolic and mitochondrial fractions from WT and DKO MEF exposed to 50 μM C12 for 1 h. The same fractions were also analysed for the presence of mitochondria by performing Western blots with antibodies to the mitochondrial marker VDAC (voltage-dependent anion channel; Huang et al., 2013) and actin, as a loading control. C12 increased cytosol/mitochondria ratio of cytochrome C in WT and DKO MEF (Fig. 11A). Western blots showed that VDAC was present in the mitochondrial fractions but not in the cytosolic fractions of all cells under all conditions (Fig. 11B). Actin staining indicated that equal amounts of protein were loaded to each lane. Thus, Western blots were consistent with the idea that the C12induced increase in cytosol/mitochondria cytochrome C ratio in WT and DKO MEF resulted from release of cytochrome C into the cytosol and not from an artefact resulting from ineffective cell fractionation or homogenization of mitochondria. Immunomicroscopy and ELISA data were therefore consistent in showing that C12 caused cytochrome C release from mitochondria into the cytosol in WT and DKO cells.

Discussion
An important conclusion from this study was that C12 rapidly triggered multiple events associated with apoptosis  in MEF. C12 caused apoptosis-characteristic activation of caspases 3/7 and 8, release of cytochrome C from mitochondria into the cytosol, shrinkage of cells and nuclei and fragmentation of nuclei and blebbing of plasma membranes. A striking aspect of these effects of C12 was the rapidity of the proapoptotic responses: C12-induced depolarization of mitochondrial membrane potential began within 5 min and was complete within 20 min; release of cytochrome C from mitochondria into the cytosol and activation of executioner caspase 3/7 occurred within 60 min. In contrast, many other proapoptotic agonists require hours or even days for apoptosis to become manifest (Goldstein et al., 2000;2005). C12 appears also to trigger similar apoptosis in other cell types, including mast cells, macrophages, neutrophils, airway epithelial cells and mammary epithelial cells (Tateda et al., 2003;Li et al., 2004;Kravchenko et al., 2006;Shiner et al., 2006;Schwarzer et al., 2012).
C12 induced killing of WT and DKO MEF after 24 h. Although PI uptake could have resulted from either necrosis or apoptosis, PI uptake also occurred in response to the proapoptotic agonist etoposide, a result consistent with the idea that C12-induced PI uptake may have similarly resulted from apoptosis. While C12-induced cell killing varied between 1.5-and 8-fold over control levels (Fig. 6), caspase activations varied between 8-and 25-fold activation over controls. This apparent quantitative discrepancy may result from differences in the methods. Caspase measurements were performed after 2 h on one set of cells while cell killing was performed after 24 h on cells that had different passage number and cell density, both of which could affect responses. Further tests of the quantitative relationship between C12-induced caspase activation and cell killing will require performing experiments over similar time-courses on cells in which growth patterns were con-  trolled. Despite this quantitative inconsistency in killing versus caspase assays, the data were clear in showing that C12 induced very rapid activation of proapoptotic responses.
A second major conclusion was that C12's proapoptotic effects occurred without the involvement of Bak or Bax. WT and DKO MEF exhibited equivalent depolarization of mitochondrial membrane potential and release of cytochrome C into the cytosol. C12 also induced equivalent activation of executioner caspase 3/7 and cell killing in WT and DKO MEF. Further, expressing Bak or Bax in DKO MEF did not increase C12's ability to activate caspase 3/7 or to kill MEF. In contrast, effects of staurosporine and etoposide, known activators of the intrinsic pathway to apoptosis (Degenhardt et al., 2002;Mikhailov et al., 2003) were dependent on Bak or Bax. Thus, staurosporine activated caspase 3/7 in WT and DKO MEF expressing Bak or Bax, but not in DKO MEF. And etoposide caused cell killing in WT and DKO MEF expressing Bak or Bax, but not in DKO MEF. Experiments on WT and DKO NIH3T3 cells (Supplementary Data Fig. S2) confirmed dose-dependent cell killing by C12 in WT and DKO MEF used for all the other experiments (Fig. 6A).
These results showing no involvement of proapoptotic Bak/Bax in C12-triggered apoptosis were consistent with the finding that overexpression of anti-apoptotic Bcl-2 in WT MEF also did not alter C12's ability to activate caspase 3/7. A previous study showed that C12-triggered depolarization of Δψ mito and activation of caspases 3 and 8 and cell killing was reduced in Jurkat cells that overexpressed Bcl-2 (Tateda et al., 2003), consistent with the idea that Bcl-2 and related proteins including Bak and Bax were intimately involved in C12-triggered apoptosis. Different cell lines or methods or even genes in apparently matched cell lines may explain such apparent discrepancies. Recent work (Valentine et al., 2013a) showed that C12-induced apoptosis in MEF required IRE1α and XBP1, also consistent with the idea that something besides Bak and Bax was mediating C12-induced apoptosis.
The most striking and important results of this study were that C12 caused cytochrome C release from mitochondria into the cytosol equally well in WT and DKO MEF. It is generally thought that following activation by upstream BH3-only proteins and inactivation of Bcl-2 and Bcl-xl, Bak and Bax oligomerize in the outer membrane of mitochondria and facilitate release of cytochrome C into the cytosol (Bender and Martinou, 2013). C12's ability to release cytochrome C from mitochondria into the cytosol without the involvement of Bak and Bax is, as far as we can tell, novel for bacteria-derived products. The triterpenoid plant derivative betulinic acid causes similar release of cytochrome C from mitochondria and activation of apoptosis, an effect that appears to be mediated by the opening of the mitochondrial permeability transition pore or PTP (Mullauer et al., 2009;. The role of the PTP in C12-induced apoptosis should be tested. Although it is clear that C12's proapoptotic effects are Bak/Bax-independent, the specific molecules or pathways triggered by C12 remain unknown. C12 clearly affects mitochondria and downstream caspase 3/7, so it seems likely that C12 activates molecule(s) in the intrinsic pathway to apoptosis in MEF. In addition, C12 appears also to activate the extrinsic pathway (via caspase 8 activation), implicating a role for C12 activating a receptor like the TNF receptor or TNFR-like receptor (Speeckaert et al., 2012;de Wilt et al., 2013). Previous studies showed that C12 triggers events often associated with receptor- Fig. 11. Effects of C12 on cytosol/mitochondria cytochrome C in WT and DKO MEF: ELISA assays. WT and DKO MEF grown on plastic dishes were incubated with Ringer's (control) or Ringer's + 50 μM C12 (C12) for 1 h followed by preparation of cytosolic and mitochondrial fractions and assays for cytochrome C (by ELISA) and VDAC and actin (by Western blot). A. C12 increased cytosolic/mitochondrial cytochrome C ratio for WT and DKO MEF. Average ± SD, n = 3 separate samples. *P < 0.05 for C12 versus control. B. Western blots showed VDAC in mitochondrial but not cytosolic fractions; actin was present prominently in cytosolic and less in mitochondrial fractions. Results typical of two similar experiments. activation: changes in activities of Stat 3 (Li et al., 2004), PPARγ (Jahoor et al., 2008), Ca 2+ and cAMP signalling (Schwarzer et al., 2010;, Ca 2+ -triggered phosphorylation (Vikström et al., 2009; and activation of IRE1α, splicing of XBP1 mRNA and production of XBP1s protein (Valentine et al., 2013a). However, the roles of and connections among these potential receptors, pathways and molecules to activation of the wide array of C12induced proapoptotic responses remain unknown. A recent study showed that C12's effects to inhibit NF-κB activation were blocked by several triazolo[4,3a]quinolones with nM affinities, indicating that C12 elicits at least some of its effects selectively (Valentine et al., 2013b). Further tests of these compounds on C12's activation of proapoptotic and inflammatory responses may provide insights into whether C12's multiple effects result from one or multiple activation steps. Although the molecular mechanisms remain to be determined, it seems possible that C12, with its high lipid solubility, activates a pattern associated molecular pattern (PAMP) receptor or damage-associated molecular pattern receptor (DAMP) (Kumar et al., 2011) in the cytosol. In lungs or other tissues infected with P. aeruginosa biofilms, where local [C12] > 10 μM, epithelia, dendritic cells, white cells, fibroblasts, neurones, smooth muscle and endothelial cells, may all exhibit increased proapoptotic signalling. Such effects may contribute to the persistence of P. aeruginosa infections in CF lungs.

Reagents
Unless otherwise specified, reagents and chemicals were obtained from Sigma. C12 was dissolved in DMSO and frozen in separate vials and then thawed for single experiments. Preliminary experiments showed that C12 lost potency with repeated thaw-freeze-thaw cycles.

Generation and culture of WT and DKO and Bax/Bak-corrected MEF
Uncloned population of wild-type and Bak −/− Bax −/− (DKO) MEF immortalized by SV40 large T antigen expression have been described previously (Wang et al., 2011). DKO MEF were resistant to the apoptosis stimuli etoposide. Bax/Bak-corrections of DKO MEF were performed. The cDNA of murine Bak or murine Bax was subcloned into the retroviral expression vector pBABE-IRES-EGFP with the marker protein enhanced green fluorescent protein (EGFP) expressing from an internal ribosomal entry site (IRES). To generate retrovirus, the package cell line human embryonic kidney (HEK)-293T was transfected with the plasmid carrying the respective gene of interest or the empty vector along with the retroviral helper plasmids pUVMC and pMDG2.0 using the jetPRIME® transfection reagent (Polyplus-transfection; New York, NY). Medium containing retrovirus was collected 48-72 h following transfection and supplemented with 10 μg ml −1 polybrene (Sigma; St. Louis, MO) to enhance infection efficiency. Bak, Bax or the control GFP alone was expressed in DKO MEF by retroviral infection. Infected cells were again subcloned by limited dilution to identify cells that expressed high levels of GFP.

Assay for caspases 3/7 and 8
Caspase 3/7 and caspase 8 activities were measured by cell-based homogeneous luminescent assays (Caspase-Glo, Promega, Madison, WI), in which a specific substrate that contains the tetrapeptide DEVD (specific for caspase 3/7) and LETD (specific for caspase 8) was cleaved by the activated caspases from the cells to release aminoluciferin reacting with the luciferase and resulting in the production of light. WT and DKO MEF were plated on a clear-bottom, white 96-well plate in 100 μl media per well for 4-5 days until they were confluent. During the experiment, cells were treated with drugs in the 37°C incubator for 60-120 min, depending on the experiment, or were left untreated as controls. On the same plate, some wells without cells but 100 μl of the same media served as blanks. After treatment, 100 μl reagent was added to each well with cells (treated or controls) and their media, or blank (media only). The plate was incubated at room temperature for 1 h on a shaker, and the end-point luminescence was measured in a plate-reading luminometer (LmaxII 384, Molecular Devices, Sunnyvale, CA). Data were background (blank) subtracted and averaged.

Cell viability assay
The indicated cell lines were plated (2 × 10 4 ) in each well of a 48-well tissue culture plate and grown for 24 h. The different cell lines were treated with various concentrations of C12 or with 10 μM etoposide for a further 24 h, then incubated with Trypsin-EDTA (Mediatech) followed by harvesting in the presence of 1.0 μg ml −1 propidium iodide (PI). Cell viability was determined by PI exclusion method using flow cytometry analysis (FACScalibur, Beckon Dickinson; San Jose, CA). Briefly, on a dot plot of forward scatter (FSC) versus FL3, the events with normal cell size (high FSC values) and low PI staining (small FL3 values) were defined as viable cells. The viability was determined as the percentage of viable events in total events, and the percentage of cell death represented 100 minus the cell viability percentage.

Cell morphology in live MEF
MEF grown on coverslips were treated in media in the incubator for 1 h with the nuclear stain Hoechst 33342 (5 μM) and BCECF/AM (2 μM), which becomes cleaved and retained in the cytosol. MEF were then rinsed with Ringer's solution and mounted in a chamber that allowed observation of cells on the stage of an inverted fluorescence microscope (Nikon T2000). Cells were then incubated with Ringer's or Ringer's + 50 μM C12 for 2 h and observed using DIC bright-field and green and blue fluorescence. Blue and green images were overlain using Adobe Photoshop to permit observation of cytosol and nuclei to compare to DIC images.

Measuring Δψmito using imaging microscopy of JC1
For imaging experiments to measure mitochondrial membrane potential (Δψmito), cells were incubated with growth media containing the Δψmito probe JC-1 (10 μM) for 10 min at room temperature, and then washed three times with Ringer's solution to remove the extra dye. Dye-loaded cells were mounted onto a chamber on the stage of either a wide field imaging microscope or a spinning disk confocal imaging microscope. Cells were maintained at room temperature during the experiment. Treatments were made by diluting stock solutions into Ringer's solution at the concentrations stated in the text. Fluorescence imaging measurements of Δψmito were performed using equipment and methods that have been reported previously Hybiske et al., 2007;Schwarzer et al., 2008). Briefly, a Nikon Diaphot inverted microscope was used with a 40× Neofluar objective (1.4 NA). A CCD camera collected emission images (em: > 510 nm for JC-1) during excitation [490 ± 5 nm for JC-1] using a filter wheel (Lambda-10, Sutter Instruments, Novato, CA). Axon Imaging Workbench 5.0 (Axon Instruments, Foster City, CA) controlled both light exposure and collection of data. Images were corrected for background (region without cells). Data have been reported as relative ratios. Maximum JC1 ratios were measured at the end of experiments by treating cells with 10 μM FCCP to completely depolarize Δψmito. Minimum JC1 fluorescence was defined as the starting value, and maximal depolarization was defined as fluorescence measured in presence of FCCP.

Cytochrome C, nuclear and cellular morphology in fixed and stained MEF
MEF grown on coverglasses were left untreated or incubated with C12 (50 μM) for times mentioned in the text; cells were then rinsed with PBS, fixed for 30 min with 2% paraformaldehyde in PBS at 4°C, rinsed with PBS and permeabilized with 0.2% Triton X-100 (in PBS) for 15 min. After blocking with 2% BSA-PBS for 60 min, cells were incubated for 1 h at room temperature with an anti-cytochrome C (Invitrogen, Grand Island, NY) followed with thorough rinses with PBS (+0.05% Tween-20). Finally, cells were incubated for 1 h with an Alexa Fluor 546 anti-rabbit -or Alexa Fluor 488 anti-mouse secondary antibodies (Invitrogen, Grand Island, NY) followed by 10 min incubation with 1 μM Hoechst 33342. Fluorescence images were obtained using an inverted confocal microscope with a 63× objective (Zeiss 710). Quantification of resulting cytochrome C distribution was performed on these images using NIH Image. Using Adobe Photoshop, we outlined images of nuclei in blue images. These outlines were then overlaid on the cytochrome C (green images). Equal-sized areas of the outlined nuclei and cytosol (area outside the nuclei) were corrected for background (area without cells) and the nuclear/cytosol ratio was calculated for each cell: (nucleusbackground)/(cytosol -background). At least 10 cells in three fields from each slide were measured and averaged for quantification of nuclear/cytosol ratio for cytochrome C.

ELISA assays of cytosolic and mitochondrial cytochrome C
ELISA assays of cytosolic cytochrome C were performed to test whether C12 altered cytosol/mitochondrial distribution of WT and DKO MEF, which were grown on plastic plates, rinsed with Ringer's and exposed for 1 h to Ringer's alone or Ringer's + 50 μM C12. Cells were then homogenized and processed using differential centrifugation to isolate cytosolic and mitochondrial fractions as previously described (Frezza et al., 2007). Protein content of fractions was measured by Bradford and equal amounts of protein per well were assayed by Quantikine ELISA (R&D Systems, Minneapolis, MN) to quantify rat/mouse cytochrome C content. Aliquots of the same fractions were also analysed in Western blots using anti-VDAC (voltage-dependent anion channel, Huang et al., 2013) and anti-actin antibodies (Cell Signaling Technology, Danvers, MA, and MP Biomedicals, Solon, OH) to determine whether the cytosolic fractions had contamination from mitochondria that occurred during preparation of the cells and to assure equivalent protein loading into the lanes of the gel.

Statistics
Quantitative data were presented as averages ± SD, where n = number of different biological samples or experiments. Statistical comparisons between C12 treatment and control (untreated) were determined using t-tests for paired or unpaired samples as appropriate; P < 0.05 was considered statistically significant.
Oakland Research Institute for help in using their Zeiss 710 LSM. We also think Drs Eileen White and Wei-Xing Zong for providing cell lines.