MAPL SUMOylation of Drp1 Stabilizes an ER/Mitochondrial Platform ...

Article

MAPL SUMOylation of Drp1 Stabilizes an ER/ Mitochondrial Platform Required for Cell Death Graphical Abstract

Authors Julien Prudent, Rodolfo Zunino, Ayumu Sugiura, Sevan Mattie, Gordon C. Shore, Heidi M. McBride

Correspondence [email protected]

In Brief Prudent et al. show that MAPL SUMOylates Drp1 at the ER/mitochondria contact sites during cell death. This mitochondrial SUMOylation is involved in the stabilization of an ER/mitochondrial signaling platform required for mitochondrial constriction, calcium flux, cristae remodeling, and an efficient cytochrome c release downstream of BAX/BAK activation.

Highlights d

MAPL and mitochondrial SUMOylation are required for an efficient cell death

d

MAPL SUMOylates Drp1 at the ER/mitochondria interface during apoptosis

d

SUMOylated Drp1 functionally stabilizes an ER/mitochondrial signaling platform

d

Ca2+ transfer from ER is required for cristae remodeling and cytochrome c release

Prudent et al., 2015, Molecular Cell 59, 941–955 September 17, 2015 ª2015 Elsevier Inc. http://dx.doi.org/10.1016/j.molcel.2015.08.001

Molecular Cell

Article MAPL SUMOylation of Drp1 Stabilizes an ER/Mitochondrial Platform Required for Cell Death Julien Prudent,1 Rodolfo Zunino,1 Ayumu Sugiura,1 Sevan Mattie,1 Gordon C. Shore,2 and Heidi M. McBride1,* 1Montreal

Neurological Institute, McGill University, 3801 University Ave, Montreal, QC H3A 2B4, Canada Medical Sciences Building, Department of Biochemistry, McGill University, Montreal, QC H3G 1Y6, Canada *Correspondence: [email protected] http://dx.doi.org/10.1016/j.molcel.2015.08.001 2McIntyre

SUMMARY

There has been evidence that mitochondrial fragmentation is required for apoptosis, but the molecular links between the machinery regulating dynamics and cell death have been controversial. Indeed, activated BAX and BAK can form functional channels in liposomes, bringing into question the contribution of mitochondrial dynamics in apoptosis. We now demonstrate that the activation of apoptosis triggers MAPL/MUL1-dependent SUMOylation of the fission GTPase Drp1, a process requisite for cytochrome c release. SUMOylated Drp1 functionally stabilizes ER/mitochondrial contact sites that act as hotspots for mitochondrial constriction, calcium flux, cristae remodeling, and cytochrome c release. The loss of MAPL does not alter the activation and assembly of BAX/BAK oligomers, indicating that MAPL is activated downstream of BAX/BAK. This work demonstrates how interorganellar contacts are dynamically regulated through active SUMOylation during apoptosis, creating a stabilized platform that signals cytochrome c release.

INTRODUCTION It has long been known that the process of intrinsic programmed cell death occurs by the regulated release of cytochrome c from the intermembrane space of the mitochondria into the cytosol. Cytochrome c is a critical component of the apoptosome that activates the caspase cleavage cascade within the executioner phase of cell death. The release of cytochrome c is considered a ‘‘point of no return,’’ so it is important that we understand the mechanisms that regulate this critical event. The formation of a large channel, or pore, in the mitochondrial outer membrane is mediated by the activation, insertion, and assembly of the proapoptotic BAX and BAK proteins (Chipuk et al., 2010). The biophysical properties of the insertion and assembly events have been worked out within reconstituted systems with high resolution (Chi et al., 2014). However, there is emerging evidence that additional forces may be required to facilitate the formation of

functional BAX or BAK pore complexes within intact mitochondria in vivo. Moreover, the bulk of cytochrome c within the intermembrane space is bound to cardiolipin at the inner membrane and locked within the cristae with limited access to the boundary of the outer membrane (Scorrano et al., 2002). Therefore, the events required within intact mitochondria are certainly more complex than within a purified system. Mitochondrial fragmentation and cristae remodeling events, for example, accompany the pro-apoptotic release of cytochrome c, triggered by the recruitment and stabilization of the fission GTPase Dynamin-related protein 1 (Drp1) on the mitochondria (Youle and Strasser, 2008). The intermembrane space GTPase OPA1 was shown to disassemble from its oligomeric state during apoptosis to release the cristae junctions (Frezza et al., 2006). The apoptotic machinery has been shown to intersect with the fusion/fission machinery in a number of model systems (Autret and Martin, 2009). However, the field remains divided on the mechanistic contribution of mitochondrial morphology to the process of BAX activation and cytochrome c release during cell death (Parone et al., 2006; Sheridan et al., 2008). Nevertheless, it is clear that activated BAX colocalizes with Drp1 at mitochondrial foci (Karbowski et al., 2002), and it has been shown using cell-free systems that the negative curvature induced at sites of mitochondrial constriction facilitates efficient BAX insertion into the outer membrane (Montessuit et al., 2010). In addition, changes in the lipid composition of the mitochondrial membrane appear to play a role in the efficiency of BAX/BAK-induced cytochrome c release (Chipuk et al., 2012; Kushnareva et al., 2012; Lucken-Ardjomande et al., 2008). Consistent with this, it has been shown that BAX pore formation in mitochondria occurs in a two-step process, with a lag period following the initial activation and insertion of BAX into the outer membrane, perhaps involving outer membrane protein(s) or lipid transitions in addition to Drp1 (Kushnareva et al., 2012). Together these data have shown that the architecture of the mitochondrial membranes contributes to the efficiency of BAX-mediated cytochrome c release and the apoptotic program. However, the molecular machinery and mechanisms that regulate these processes are unclear. Mitochondrial division during apoptosis is mediated by Drp1, a target of multiple post-translational modifications, including SUMOylation (Elgass et al., 2013). Small ubiquitin-like modifier 1 (SUMO1) is a 100 aa protein that is a member of the ubiquitin family and is covalently conjugated to protein

Molecular Cell 59, 941–955, September 17, 2015 ª2015 Elsevier Inc. 941

Figure 1. MAPL Is Required for Apoptosis (A) TOM20 staining of YFP-SUMO1-transfected cells treated with non-targeting (NT) or MAPL siRNA, and with DMSO or 1 mM STS for 3 hr (h). Arrows indicate sites of YFP-SUMO1 recruitment to mitochondria during cell death. Scale bars: 10 mm (top) and 5 mm (bottom). (B) Histogram depicting the percent of apoptotic transfected cells with YFP-SUMO1 foci accumulation at mitochondria. HeLa cells were silenced with NT, pooled si-MAPL or the single si-MAPL-12, and rescued, as indicated, with MAPL-FLAG in control or STS-treated cells. (mean ± SD; n = 3 independent experiments, quantifying at least 30 cells from three coverslips [100 cells total] within each experiment). (legend continued on next page)

942 Molecular Cell 59, 941–955, September 17, 2015 ª2015 Elsevier Inc.

substrates through an enzymatic cascade including an E1 heterodimer, an E2 ligase (Ubc9), and an E3 ligase that provides specificity for the reaction (Flotho and Melchior, 2013). While the majority of identified SUMO substrates reside within the nucleus, roles for SUMOylation within the cytosol have recently become apparent (Wasik and Filipek, 2014). We previously reported that the sites of division during apoptosis are accompanied by a BAX/BAK-dependent increase in the SUMOylation of Drp1 (Wasiak et al., 2007). Subsequent to our study, others have shown an involvement of SUMOylation in cell death models, also linking this to Drp1 conjugation (Fu et al., 2014; Guo et al., 2013). In our search for proteins that regulate mitochondrial fission, we identified a mitochondrial anchored RING-finger containing protein, MAPL, whose overexpression led to hyper-fragmentation (Neuspiel et al., 2008). We found that MAPL possessed efficient SUMO E3 ligase activity in vitro and in vivo, promoting the covalent SUMOylation of Drp1 (Braschi et al., 2009). SUMOylation is a transient, reversible process in which protein deSUMOylation is catalyzed by a family of seven conserved cysteine proteases (Sentrin proteases [SenPs]) (Mukhopadhyay and Dasso, 2007). We previously identified SenP5 as a protease that cleaves SUMOylated mitochondrial substrates, including Drp1, playing a specific role at the G2/M transition (Zunino et al., 2007, 2009). To further define the contribution of mitochondrial SUMOylation during cell death, we now demonstrate that MAPL-dependent SUMOylation of Drp1 stabilizes its oligomeric form on the mitochondria, a process enhanced upon an apoptotic trigger. Oligomeric Drp1 is required to generate an interorganellar platform between the endoplasmic reticulum (ER) and mitochondria that facilitates calcium flux between the organelles, driving cristae remodeling and OPA1 complex disassembly. We thus define the molecular mechanism by which SUMOylation functions in apoptosis, providing a switch mechanism to link the activation and assembly of BAX/BAK with the release of cytochrome c. RESULTS MAPL and Mitochondrial SUMOylation Are Required for Programmed Cell Death Drp1 is one of the key components of the mitochondrial fission machinery and is involved in the execution of apoptosis (Frank et al., 2001; Martinou and Youle, 2011). We previously demonstrated that Drp1 was SUMOylated in a BAX/BAK-dependent manner at a point after BAX activation but before cytochrome c release (Wasiak et al., 2007). We later identified MAPL as a

mitochondrial SUMO E3 ligase that SUMOylates Drp1 to control mitochondrial dynamics (Braschi et al., 2009). Therefore, we here examined the role of MAPL and subsequent mitochondrial SUMOylation during cell death using siRNA. We first tested whether transfected YFP-SUMO1 was stably recruited to mitochondria within apoptotic cells in the presence or absence of MAPL. Confocal microscopy analysis revealed a few YFPSUMO1 puncta colocalizing with mitochondria in untreated cells, consistent with the few fission events in progress at the time of fixation (Figure 1A). However, upon induction of cell death (1 mM staurosporin [STS] for 3 hr), we observed a clear accumulation of YFP-SUMO1 at foci along mitochondria tubules (Figures 1A and 1B). In contrast to control cells, we observed no accumulation of YFP-SUMO1 at the mitochondria in cells lacking MAPL, treated or not with STS (Figures 1A, 1B, and S1A). These results were confirmed by biochemical fractionation of either STS-untreated or STS-treated cells in the presence or absence of MAPL. Indeed, fractionation revealed that endogenous SUMO1 conjugates on mitochondria were significantly increased in control apoptotic cells (Figure 1C). Silencing of MAPL in HeLa cells was efficient, and the increase of endogenous SUMO1 conjugates at mitochondria after STS treatment was no longer observed (Figure 1C). There was a reduction of Drp1 at mitochondrial membrane observed in si-MAPL cells, consistent with the decreased stability when SUMOylation is lost (Harder et al., 2004; Zunino et al., 2007). To test the specificity of the siRNA SMARTpool used in our experiments, we performed a rescue experiment using an individual siRNA: si-MAPL-12 (Figures 1B and S1A). Si-MAPL-12 also leads to a significant decrease of the YFP-SUMO1 accumulation at mitochondria during apoptosis (25.5% ± 2.3% to 80.4% ± 11% in control cells compared to 22.8% ± 2.8% to 31.2% ± 3.5% in si-MAPL cells and from 28.4% ± 4.3% to 19.6% ± 6.3% in si-MAPL-12). Moreover, overexpression of MAPL-FLAG in HeLa cells silenced with the si-MAPL-12 and treated with STS led to the rescue of the phenotype confirming its specificity. These results show that MAPL plays a crucial role in the SUMOylation of mitochondrial proteins during programmed cell death. To determine the effect of MAPL on the apoptotic program, we examined the cleavage of the pro-caspase 3 as a hallmark of apoptosis. As shown in Figure 1D, there was a slight reduction of the total cellular SUMOylation upon MAPL knockdown in the absence of STS treatment. Upon STS treatment, a significant decrease in the amount of cleaved caspase 3 was observed, and the smear of total SUMO conjugates was lost following MAPL silencing, indicating that MAPL is required for an efficient execution of apoptosis. Similar results were obtained using the

(C) Control or STS-treated HeLa cells, silenced or not for MAPL, were fractionated to isolate mitochondrial and cytosol fractions. Western blot shows increased endogenous SUMO1 conjugation on mitochondrial targets upon STS treatment, which is lost in the absence of MAPL. Expression of Drp1 was analyzed. HSP60 was used as loading control. (D–F) Western blots showing a decrease in the total amount of cellular SUMO1 conjugates and cleaved caspase 3 in HeLa cells silenced for MAPL following treatment with (D) 1 mM STS for 3 hr. (E) 3 mM camptothecin for 12 hr. (F) tBid infection for 15 hr. (G) Propidium iodide stain to monitor apoptotic DNA fragmentation was analyzed by FACS. MAPL-silenced HeLa cells presented a delay in the peak of fragmented DNA compared to control under three different cell death inducers: STS, camptothecin and etoposide (n = 2 independent experiments). See also Figure S1.


(legend on next page)


DNA-damaging agents camptothecin and etoposide, showing significant reductions in the amount of the cleaved caspase 3 and the total endogenous SUMO conjugates (Figures 1E and S1B). Employing a viral expression vector, we then addressed the effect of MAPL on tBid-induced cell death, the expression of which specifically activates BAX and BAK at mitochondria (Li et al., 1998; Luo et al., 1998). tBid-induced cell death also showed MAPL dependence, since si-MAPL-treated cells showed a decrease in cleaved caspase 3 and SUMO1 conjugates compared to control cells (Figure 1F). In addition, by performing fluorescence-activated cell sorting (FACS) analysis of cleaved nuclear DNA stained with propidium iodide, we observed a decrease in the peak of fragmented DNA in response to three different cell death triggers: STS, camptothecin, and etoposide (Figure 1G). Finally, overexpression of MAPL-FLAG in STS-treated HeLa cells silenced with si-MAPL-12 rescued the effect of MAPL knockdown on the amount of cleaved caspase 3 (Figure S1C). It has been shown that Drp1 must be dephosphorylated during cell death for efficient recruitment (Cereghetti et al., 2008; Cribbs and Strack, 2007). This is due, in part, to the death-induced degradation of the mitochondrial protein kinase A anchoring protein AKAP1, leading to the loss of the kinase PKA on mitochondria (Kim et al., 2011). We directly evaluated the effect of MAPL silencing during cell death on AKAP1 and P-Drp1 S637 by western blot analysis. Both in control and in si-MAPL cells, AKAP1 was efficiently degraded, which correlated with the decrease of the P-Drp1 S637 level during cell death (Figure S1D). These data demonstrate that MAPL is responsible for the induction of SUMOylation during cell death, an event downstream Drp1 dephosphorylation. Targeting the SUMO-Protease SenP5 to Mitochondria Phenocopies MAPL Silencing during Cell Death We, and others, have previously shown that MAPL (also called MULAN/MUL1, GIDE, and HADES) may possess ubiquitin E3 ligase activity (Bae et al., 2012; Braschi et al., 2009; Jung et al., 2011; Li et al., 2015; Li et al., 2008; Lokireddy et al., 2012; Yun et al., 2014; Zhang et al., 2008). Therefore, although we observed clear reduction in total SUMO conjugates in the absence of

MAPL, it was important to exclude a specific requirement for MAPL during cell death through an indirect ubiquitination activity. To test whether the effect of MAPL on the death program is linked to its SUMOylation activity, we ectopically targeted a SUMO protease, SenP5, to the mitochondrial surface. We have shown previously that SenP5 deSUMOylates a broad spectrum of mitochondrial substrates, including Drp1, to maintain mitochondrial morphology (Zunino et al., 2007). We observed that deleting the amino-terminal nuclear targeting signal of SenP5 led to the spontaneous recruitment of YFP-tagged SenP5 to the mitochondria (SenP5D1-125-YFP) (Figure 2A). We therefore generated stably expressing COS7 cells containing either the empty vector or vector containing SenP5D1-125-YFP. Consistent with the role of SUMOylation in promoting Drp1 stability and mitochondrial fission, cells expressing SenP5D1-125YFP showed a significant increase in hyperfused mitochondria (Figures 2A and S2A). Examination of the total cellular SUMOylation within these stable cell lines revealed a significant reduction in the high-molecular-weight SUMO conjugates in cells expressing SenP5D1-125-YFP (Figure 2B). There was also a mild reduction in total Drp1 protein levels, and upon longer exposures, the Drp1 oligomer formation was destabilized (Figure 2B). It has also been suggested that the fusion machinery, and in particular Mitofusin2 (Mfn2), may be directly affected by MAPL (Lokireddy et al., 2012; Yun et al., 2014). However, under our conditions, MAPL silencing or expression of SenP5D1-125-YFP led toward a decrease in Mfn2 protein levels that may reflect compensatory effects to counteract the loss in Drp1 function (Figure S2B). This indicates that MAPL is not required for Mfn2 turnover. We next evaluated the capacity of cells stably expressing SenP5D1-125 to execute apoptosis after treatment with camptothecin. As shown in Figure 2C, cells expressing SenP5D1-125YFP resulted in a significant reduction in caspase 3 cleavage, and its substrate Poly [ADP-ribose] polymerase 1, compared to control cells (Figure 2C). These results showed that targeting a SUMO protease to the mitochondrial surface phenocopies the loss of MAPL during cell death, further confirming a requirement for mitochondrial SUMOylation during apoptosis.

Figure 2. Targeting SenP5 to Mitochondria Phenocopies Silencing of MAPL (A) COS7 cells transfected with SenP5-YFP or SenP5D1-125-YFP (MitoSenP5-YFP) were analyzed by fluorescence microscopy following incubation with Mitotracker Red 633. Enlarged boxes show mitochondrial targeting of SenP5D1-125-YFP. Scale bars: 10 mm. (B) COS7 cells stably expressing either empty pcDNA3 vector or vector containing SenP5D1-125-YFP were processed for SDS-PAGE and western blot using indicated antibodies. TOM20 is used as loading control. (C) Western blot of COS7 cells stably expressing vector alone or SenP5D1-125-YFP, treated with 3 mM camptothecin, show a decrease in the amount of cleaved PARP and cleaved caspase 3 compared to control. (D) Immunofluorescence of HeLa cells treated with NT or SenP3 siRNA, transfected with YFP-SUMO1, and treated with STS. Mitochondria were stained with antiTOM20 antibody. Arrows indicate sites of SUMO1-YFP recruitment to mitochondria during cell death. Scale bars: 10 mm (left) and 5 mm (right). (E) Western blot examining HeLa cells silenced for SenP3 following STS treatment. STS treatment increased the total SUMO2/3 conjugates upon loss of SenP3; however, no variation of cleaved caspase 3 was observed compared to control. Drp1 and MAPL expression are shown and HSP70 was used as loading control. (F) Immunofluorescence of HeLa cells treated with si-NT, -MAPL, or -Drp1 were transfected with GFP-SenP3 and treated with STS. Drp1 and TOM20 were immunolabeled, and GFP-SenP3 was not seen to translocate to mitochondria during cell death, nor to alter Drp1 recruitment. Long exposures of GFP-SenP3 in the cytosol are shown in the inset. Scale bars: 10 mm (left) and 5 mm (right). (G) Western blot analysis of cells treated, as indicated, expressing either GFP or GFP-SenP3. Total SUMO2/3 conjugates are decreased in HeLa cells expressing GFP-SenP3, with increased conjugation following STS treatment. Expression of GFP-SenP3 led to a decrease in cleaved caspase 3 compared to control. MAPL expression is shown, and TOM20 was used as loading control. See also Figure S2.


SenP3 Does Not Interfere with SUMOylation Activity of MAPL during Cell Death Consistent with our observation that the ectopic targeting of SenP5 to mitochondria delayed cell death, previous work has demonstrated the loss of SenP2 also led to the stable association of Drp1 with mitochondria, increasing fragmentation and cell death (Fu et al., 2014). However, contradictory evidence has also been presented where the loss of a SUMO protease SenP3 stabilized SUMO2/3 conjugated forms of Drp1, which inhibited cytochrome c release in a model of neuronal ischemia (Guo et al., 2013). In our model system, the loss of SenP3 had no impact of SUMO1 punctae localized at mitochondria after STS treatment (Figures 2D, S2C, and S2D). The loss of SenP3 was efficient (Figure S2E), showing a further increase in SUMO2/3 conjugates on isolated mitochondria and in wholecell extracts upon death induction (Figures 2E and S2F). However, this increase in SUMO2/3 conjugation did not affect the efficiency of cell death, as caspase 3 cleavage was similar in both STS-treated cells (Figure 2E). These results suggest that while SenP3 may affect mitochondrial SUMO2/3 conjugates, it does not play a role in death induction within this cell model system. We next analyzed the effect of ectopic overexpression of SenP3-GFP in HeLa cells. Confocal microscopy analysis showed that SenP3-GFP was predominantly expressed in the nucleus and was not recruited to mitochondria upon STS-treatment or/and silencing of MAPL or Drp1 (Figures 2F, S2G, and S2H). As expected, this overexpression of SenP3-GFP led to a decrease in SUMO2/3 conjugates within whole-cell extracts in steady state (Figure 2G). Overexpression of SenP3 did not alter the STS-induced stimulation of SUMO2/3 conjugation, consistent with its presence within the nucleus (Figure 2G). We observed a mild decrease in the activation of SUMO1 conjugates during death. Consistent with this, there was a marked decreased in the amount of the cleaved caspase 3 compared to control (Figure 2G). Taken together, the manipulation of the SUMO E3 ligase MAPL and two SUMO proteases SenP5 and SenP3 demonstrate that (1) mitochondrial SUMOylation is increased during cell death and (2) SUMOylation is a critical step in apoptosis. MAPL Is Required Downstream of BAX/BAK Activation and Oligomerization We next mapped the function of MAPL relative to BAX activation using the monoclonal conformation-specific antibody (BAX6A7), which detects activated BAX (Hsu and Youle, 1997). The total amount of cellular BAX was unchanged upon loss of MAPL (Figure 3A, top panels); however, the quantity of immunoprecipitated, activated BAX following STS treatment was actually higher in si-MAPL cells, clearly demonstrating that BAX activation does not require MAPL (Figure 3A, lower panel). To test whether BAX insertion into the mitochondrial membrane may have been altered, we performed an alkali carbonate extraction experiment in cells treated or not with si-MAPL. Again, we observed no change in the fraction of activated BAX that was integrated into the membrane pellet upon a death trigger (Figure 3B). Lastly, we performed a crosslinking experiment to analyze potential changes in BAX oligomerization. Addition of

the crosslinker BMH revealed a laddering of BAX on reducing gels upon addition of STS. Consistent with its activation and insertion, we also saw no change in the oligomerization of BAX in cells lacking MAPL (Figure 3C). In addition, the loss of MAPL did not affect the oligomerization of BAK at the mitochondria (Figure 3D). We next employed confocal microscopy to monitor BAX/BAK recruitment and activation at the mitochondria and to quantify cytochrome c release within intact cells. In untreated cells, cytochrome c was localized within mitochondria, and antibodies against activated BAX and BAK revealed faint cytosolic staining. Following STS treatment, activated BAX and BAK were clearly detected at punctate sites on the mitochondria in both control and si-MAPL-treated cells (Figures 3E–3G). However, in STStreated cells lacking MAPL where BAX or BAK were actively recruited, cytochrome c release occurred only in 50% of the cells relative to control (Figures 3F and 3H). We further confirmed the reduction in cytochrome c release using a biochemical fractionation of cells treated with STS, showing a 2-fold reduction in the release of cytochrome c in cells lacking MAPL (Figures S3A and S3B). A similar result was obtained in cells expressing mitochondrial SenP5 (Figure S3C). These results map the role of mitochondrial SUMOylation at a point downstream of BAX/BAK activation and oligomerization, highlighting the requirement for SUMOylated substrates in the release of cytochrome c. Drp1 Is the Primary Substrate for MAPL-Induced SUMOylation during Cell Death Although we observe a clear increase in many MAPL-dependent SUMO conjugates upon mitochondria during cell death, only Drp1 has been identified as a direct substrate (Braschi et al., 2009). To confirm whether Drp1 is the major substrate of MAPL during cell death, we examined the effect of Drp1 knockdown on the YFP-SUMO1 status following STS treatment (Figures 4A, 4B, and S4A). Overexpression of YFP-SUMO1 leads to the accumulation of SUMO1 puncta at mitochondria in control cells treated with STS (from 19.1% ± 3.6% of cells showing YFPSUMO1 at mitochondria in steady state compared to 79% ± 3.6% of STS-treated cells). However, the loss of either MAPL or Drp1 similarly blocked the accumulation of YFP-SUMO1 foci on the mitochondria of STS-treated cells (from 19.8% ± 6.7% to 25% ± 8.6% in si-MAPL cells and from 15.6% ± 8.2% to 27% ± 3.4% in si-Drp1 cells). The efficiency of silencing is shown in Figure S4B. This demonstrates that Drp1 is either the central substrate for MAPL-induced SUMOylation or is required for the stabilization of YFP-SUMO1 conjugates within these sites. We next performed a biochemical fractionation of STS-treated cells silenced for Drp1, and we again confirm the increase in endogenous mitochondrial SUMO1 conjugates during the death treatment (Figure 4C). In contrast, mitochondria isolated from STS-treated cells lacking Drp1 showed a clear and significant reduction of SUMO1 conjugates at mitochondria. Moreover, the decrease of mitochondrial SUMOylation in cells lacking Drp1 was accompanied by a reduction in the execution of apoptosis, as shown by a decrease in cytosolic cleaved caspase 3 (Figures 4C and 4D). To directly examine the SUMOylation of Drp1, we performed affinity chromatography experiments from


Figure 3. Mitochondrial SUMOylation Is Required for Cytochrome c Release Downstream of BAX/BAK Activation (A) si-NT- or si-MAPL-silenced HeLa cells were treated with STS, and total cell lysates were analyzed by western blot for BAX protein level analysis (upper panel). HSP70 is used as loading control. Activated BAX was then immunoprecipitated using the activated anti-BAX 6A7 antibody (lower panel). (B) Biochemical analysis of mitochondrial BAX insertion in control or MAPL-silenced HeLa cells following incubation with STS. In STS-treated conditions BAX was found in the alkali-resistant mitochondrial pellet in the presence or absence of MAPL. VDAC1 and SDHA were used as control. (C) Mitochondria isolated from si-NT- or si-MAPLtreated HeLa cells treated or not with STS were cross-linked with 10 mM 1,6-Bis(maleimido)hexane (BMH) where indicated, and (C) oligomerized BAX complexes were analyzed by western blot. HSP60 was used as loading control. * and ** show BAX monomer and BAX complexes, respectively. (D) BAK complexes were analyzed by western blot. HSP60 was used as loading control. * and *** show BAK monomer and BAK complexes, respectively. ** denotes an intrachain cross-link of the inactive BAK conformer. (E) Immunofluorescence imaging of si-NT- or siMAPL-silenced HeLa cells incubated with DMSO or STS. BAX activation and recruitment to mitochondria was visualized using the 6A7 antibody, and cytochrome c was stained for analysis by confocal microscopy. Zoomed areas highlight cells where active BAX was recruited to mitochondria. In control cells, cytochrome c is diffuse, reflecting release from the mitochondria. In cells lacking MAPL, BAX is recruited, but there is a clear delay in cytochrome c release. Scale bars: 10 mm (left) and 5 mm (right). (F) Quantification of cells in (E) exhibiting mitochondrial BAX recruitment where cytochrome c is released. (Mean ± SD; n = 3 independent experiments, 100 cells in each case.) (G) As in (E), with an analysis of activated BAK and cytochrome c. Scale bars: 10 mm (left) and 5 mm (right). (H) Quantification of (G) (mean ± SD; n = 3 independent experiments, 100 cells in each case). See also Figure S3.

si-MAPL-treated cells overexpressing His6-SUMO1. Isolation of the total cellular SUMO1 conjugates using this approach revealed a significant decrease in Drp1-SUMO1 conjugates in cells lacking MAPL during cell death (Figure 4E). Our previous examination suggested that Drp1 may be engaged in highly stable oligomeric complexes during cell death (Wasiak et al., 2007). To test whether SUMOylation may be required for these changes, we employed crosslinking strategies. Indeed, upon STS treatment, we observed the stabilization of Drp1 dimers, trimers, and higher order structures, which were decreased by 50% in cells silenced for MAPL (Figure 4F). Even in the absence of STS, the loss of MAPL inhibits the formation of Drp1 dimers. This extends our previous findings and indicates

that Drp1 SUMOylation acts to stabilize and trap assembled Drp1 oligomers on the mitochondrial membrane. MAPL Stabilizes Mitochondria-ER Contact Sites Requisite for Calcium Flux Although these data establish a requirement for Drp1-mediated SUMOylation by MAPL in cell death, it remained unclear why this event would be required for cytochrome c release. It has recently been established that Drp1-mediated constriction of mitochondria in steady-state fission events occurs at sites of ER contact, both in yeast and mammals (Friedman et al., 2011; Korobova et al., 2013; Murley et al., 2013). Therefore, we next explored whether SUMOylated Drp1 may stabilize sites of interorganellar


Figure 4. Drp1 Is the Major Target of MAPL SUMOylation during Cell Death (A) Immunofluorescence of si-NT-, -MAPL-, or -Drp1-silenced HeLa cells were transfected with YFP-SUMO1, treated with STS, and labeled with anti-TOM20 antibodies. Arrows indicate sites of YFP-SUMO1 recruitment to mitochondria during cell death, which are lost in the absence of MAPL or Drp1. Scale bars: 10 mm (left) and 5 mm (right). (B) Quantification of DMSO- or STS-treated, YFP-SUMO1-transfected cells in (A) exhibiting YFP-SUMO1 accumulation at mitochondria (mean ± SD; n = 3 independent experiments, 100 cells in each case). (C) Mitochondrial fractionation of si-NT- or siDrp1-silenced HeLa cells, treated as indicated. Western blot shows a loss in STS-induced mitochondrial endogenous SUMO1 conjugates in absence of Drp1 relative to control cells. HSP60 was used as loading control. (D) Quantification of the ratio of cytosolic cleaved caspase 3 from (C). (E) Ni:NTA-agarose isolation of total His6-SUMO1 conjugates from transfected HeLa cells (siNT or siMAPL), treated with STS, were probed with antiDrp1 antibodies. The total amount of Drp1 and its SUMOylated form are decreased in MAPLsilenced cells during cell death. Actin is used as loading control. (F) Death-induced oligomerization of Drp1 is dependent upon MAPL. Mitochondria isolated from control or si-MAPL-silenced HeLa cells treated or not with STS were cross-linked with 10 mM BMH where indicated, and Drp1 complexes were analyzed by western blot. VDAC1 was used as loading control. * and ** show Drp1 monomer and Drp1 complexes, respectively. Lower panel shows quantification of the ratio of crosslinked Drp1 complexes. See also Figure S4.

contact sites that may facilitate the influx of calcium during cell death (de Brito and Scorrano, 2010; Giorgi et al., 2011). We first examined ER-mitochondria contact sites using immunofluorescence analysis (Figure 5A) where the silencing of MAPL showed a 13% decrease in contacts in healthy cells (Mander’s coefficient: 0.76 ± 0.02 and 0.67 ± 0.02, respectively, for control and si-MAPL cells) (Figure 5B). While significant, this moderate effect suggests that MAPL is not an essential tethering factor required for contact formation, but it rather may stabilize these contacts. Indeed, Mfn2 knockdown, used as positive control for the ER and mitochondria tethering, leads to a decrease of 21% (0.61 ± 0.01) in our system (Figure 5B) (de Brito and Scorrano, 2008). The efficiency of silencing is shown in Figure S5A.

To further corroborate a link between MAPL and the establishment of ER/ mitochondrial contacts, we examined the localization of MAPL-FLAG within the mitochondria. Expression of MAPLFLAG promotes mitochondrial fragmentation; however, a closer inspection revealed that MAPL-FLAG was enriched at sites of mitochondrial constriction that were ‘‘crossed’’ by the ER tubules, placing the SUMOylation machinery directly at the ER/mitochondrial constriction sites (Figure 5C). To examine the functional consequences of this change, we analyzed the capacity of silenced cells to transfer Ca2+ from the ER to mitochondria upon histamine stimulation using a mitochondrial calcium probe, Rhod-2 (Scorrano et al., 2003). In control HeLa cells, histamine rapidly increased mitochondrial calcium levels, reflecting the transfer of Ca2+ from the ER via the inositol-3,4,5-triphosphate receptor (IP3R) to mitochondria. Silencing MAPL led to a significant reduction of the maximal amplitude peak induced by histamine (Figures 5D and 5E)


(maximal F/F0 control cells: 1.36 ± 0.015 versus 1.22 ± 0.008 for cells lacking MAPL), reflecting a reduction by almost 40%. There was no change in the load of Ca2+ in the ER, since inhibition of the SERCA pump with thapsigargin (Prudent et al., 2013) showed similar activation of the cytosolic Ca2+ probe FluoForte (Figures 5F and 5G). We also analyzed the linear fit of the curves upon thapsigargin treatment and observed a decrease of 44% ± 5.6% in the relative mitochondrial calcium uptake velocity (Figure S5B). Taken together, these results strongly suggest that MAPL is involved in the stability of the ER/mitochondrial contact sites, and mitochondrial Ca2+ uptake defects are direct consequences of this function.

to the influx of calcium. Quantitative electron microscopy analysis of healthy cells and those treated with STS showed a clear reduction in class I (swollen cristae) and class II (absent cristae) mitochondria when MAPL was absent (Figures 6F, 6G, S6F, and S6G) (Scorrano et al., 2002). Importantly, silencing of Drp1 or inhibiting MCU also reduced the appearance of class I and II mitochondria (Figures 6G, S6F, and S6G). These data demonstrate that the SUMOylation of Drp1 by MAPL during cell death stabilizes ER mitochondrial contact sites that facilitate calcium uptake into the mitochondria, a process that triggers the remodeling of the cristae and efficient release of cytochrome c.

MAPL-Mediated ER/Mitochondrial Contact Sites Facilitate Calcium Flux and Cristae Remodeling during Cell Death While MAPL is required to stabilize ER/mitochondrial contacts in steady state, we next examined this function during a tBid death trigger. In control cells, we observed a 15% increase in ER/mitochondrial contacts following tBid treatment, consistent with previous EM quantification during serum-deprivationand tunicamycin-induced death paradigms (Csorda´s et al., 2006). However, using Mander’s coefficient to quantify the contacts in tBid-treated cells showing active mitochondrial Bax recruitment revealed a reduction by 14% in cells lacking MAPL (Figures 6A, 6B, S6A, and S6B). Importantly, a direct examination of the mitochondrial calcium uptake during STSinduced cell death by live imaging microscopy was performed using the quantitative calcium probe RCamP targeted to the mitochondrial matrix (Figures 6C and 6D). Cells lacking MAPL initiated some calcium flux; however, there was a clear divergence where control cells showed a more rapid and constant increase in calcium uptake. By 3 hr of treatment, there was a 50% decrease in the total mitochondrial calcium load into cells lacking MAPL (Figure S6C). We tested whether these changes may be indirectly due to alterations in the expression of the mitochondrial calcium uniporter (MCU) and/or its regulators MICU1/2 (Baughman et al., 2011; De Stefani et al., 2011; Perocchi et al., 2010). Western blots confirmed that these proteins were unchanged in the absence of MAPL, in steady state, and upon cell death (Figure S6D). In addition, the morphology of the mitochondria may impact contacts with the ER (Filadi et al., 2015); however, in the absence of MAPL, the mitochondria are not fragmented, showing mild hyperfusion (data not shown). Together, these data are consistent with a requirement for MAPL in the efficient assembly of ER contact sites facilitating calcium entry into mitochondria. Our earlier work showed a requirement for Drp1, along with a sensitivity to the MCU inhibitor Ru360, in the remodeling of the cristae during calcium-induced death triggered by expression of the ER-targeted BH3-only protein Bik (Germain et al., 2005). An established index of cristae remodeling is the disassembly of OPA1 oligomers (Frezza et al., 2006; Jiang et al., 2014). Indeed, the loss of MAPL led to the retention of OPA1 oligomers upon treatment with STS, particularly the high-molecular-weight complexes (Figures 6E and S6E). This shows that at least one of the functions of the stabilized ER/mitochondrial contacts is to drive the opening of the cristae in response

Recruitment of Activated BAX to ER Contacts Stabilized by SUMOylated Drp1 We next used confocal imaging and confirmed that the accumulation of YFP-SUMO1 colocalized with mitochondria-ER contact sites in cells treated with tBid (65% ± 4.9% and 58% ± 4.9% of total YFP-SUMO1 foci colocalized at mitochondria and at mitochondria/ER contact sites, respectively) (Figures 7A, 7B, and S7A). In cells lacking MAPL, there were no YFP-SUMO1 foci, and the ER contacts were decreased. In addition, quadruple staining showed a colocalization between Drp1 and YFPSUMO1 precisely at the sites of mitochondrial constriction and ER contact (Figures 7C and S7B). Finally, to test whether BAX was recruited to the stabilized apoptotic interorganellar ER/mitochondrial platform, we imaged YFP-SUMO1 puncta with activated BAX. The data show that 62% of the YFP-SUMO1 foci localized on the mitochondria colocalize with activated BAX (Figures 7D, 7E, and S7C). Importantly, in the absence of MAPL, BAX was still actively recruited to the mitochondria, but the contacts were decreased and cell death was delayed (Figures 7D and 7E). This suggests that BAX initiated the stabilization of ER/mitochondrial contacts through the activation of MAPL-mediated Drp1 SUMOylation. These constricted contact sites reflect ‘‘hotspots’’ for calcium flux that facilitates cristae remodeling required for BAX/BAKmediated cytochrome c release (Figure 7F). DISCUSSION In this study, we demonstrate that the covalent conjugation of Drp1 by MAPL-dependent SUMOylation acts to stabilize sites of mitochondrial constriction and ER contact within intact cells. These sites of constriction are required for calcium flux into mitochondria, which is important to drive the disassembly of OPA1 oligomers leading to the remodeling of the cristae and efficient cytochrome c release. Consistent with biophysical approaches using purified components in liposomes, we also observe that activated BAX and BAK oligomers form in the absence of MAPL, acting upstream in the initiation of ER/mitochondrial signaling platforms. Given the complexity of the mitochondrial inner membrane, it is perhaps not surprising that there are a critical number of molecular events required for the release of cytochrome c. Many have addressed each of the concepts individually: for example, that calcium is required for PTP transition (Pacher and Hajnoćzky, 2001), OPA1 is required for cristae opening (Frezza et al., 2006), or cardiolipin oxidation is required


Figure 5. MAPL Silencing Leads to a Decrease in ER-Mitochondria Contact Sites (A) Immunofluorescence of si-NT- or si-MAPL-silenced HeLa cells transfected with the ER marker and ER-DsRed2 and stained with anti-TOM20 antibody. Zoomed areas highlight ER-mitochondria contact sites. Scale bars: 10 mm (left) and 5 mm (right). (B) Histograms representing the abundance of the ER-mitochondria contact sites using the overlapping Mander’s coefficient. Mfn2 silencing effect on Mander’s coefficient in HeLa cells is used as a positive control. (Mean ± SEM; n = 3 independent experiments, 30 cells in each condition.) (C) Immunofluorescence of HeLa cells transfected with ER-DsRed2 and MAPL-FLAG. Cells were stained with anti-TOM20 and FLAG antibodies. (C0 –C0 0 0 ) Zoomed areas highlight the localization of MAPL-FLAG in ER-mitochondria contact sites. Scale bars: 10 mm (left) and 2 mm (right). (legend continued on next page)


Figure 6. MAPL Stabilizes an ER/Mitochondrial Platform Required for Mitochondrial Calcium Uptake and Cristae Remodeling during Cell Death (A) Immunofluorescence of si-NT- or si-MAPLsilenced HeLa cells infected with AD-tBid for 9 hr. Mitochondria and ER were stained with antiTOM20 and anti-KDEL antibodies. Activation of BAX in tBid-treated cells was shown using the antiBAX-6A7 antibody. Zoomed areas highlight ERmitochondria contact sites. Scale bars: 10 mm (left) and 2 mm (right). (B) Quantification of ER-mitochondria contact sites following BAX activation from (A) using the overlapping Mander’s coefficient. (Mean ± SEM; n = 3 independent experiments, 30 cells in each experiment.) (C) Direct visualization of mitochondrial calcium uptake during STS-induced cell death using the mito-RCamP probe. Images from si-NT- or siMAPL-silenced HeLa cells were acquired every 10 min, and 1 mM STS was added after 40 min, as indicated. Representative images at T0, T20, and T180 min after STS injection. Heatmap false colors are shown. Scale bars: 10 mm. (D) Quantification of mitochondrial Ca2+ uptake from (C) presented as the ratio of F/F0 RCamP intensities (AU; mean ± SEM; n = 3 biological replicates, 10 cells in each case). (E) OPA1 disassembly is delayed in si-MAPLtreated cells. Mitochondria isolated from control or si-MAPL-silenced HeLa cells treated or not with STS were cross-linked with 10 mM BMH, where indicated, and OPA1 complexes were analyzed by western blot. VDAC1 was used as loading control. * and ** show OPA1 monomer and OPA1 complexes, respectively. (F) Ultrastructural analysis of mitochondria within STS-treated cells highlighting the three different classes of mitochondrial cristae shape in HeLa cells treated with 1 mM STS + 10 mM Z-VAD for 3 hr (scale bars: 100 nm). (G) Quantification of cristae shape in mitochondria from si-NT-, -MAPL-, -Drp1-, and -MCU-silenced cells treated with STS and Z-VAD (mean ± SEM; n = 2 independent experiments, si-NT +STS, siMAPL+STS, si-Drp1+STS, si-MCU+STS: 547, 781, 263, 304 mitochondria from 5–10 cells were analyzed, respectively). See also Figure S6.

to release cytochrome c from the inner membrane (Kagan et al., 2005). However, our work now unifies the field by bringing together these events for the efficient execution of apoptosis. The data indicate that it is not necessarily mitochondrial division

that is required for BAX/BAK-mediated cytochrome c release, but rather the stabilization of a specific interorganellar platform that facilitates membrane curvature and metabolite flux. This is consistent with previous work showing that hyperfragmented

(D) Representative trace of mitochondrial Ca2+ uptake in cells loaded with the mitochondrial Ca2+-sensitive dye Rhod-2 AM following stimulation with 10 mM histamine (F/F0 Rhod-2, AU). (E) Quantification of mitochondrial Ca2+ uptake in (D) assessed by analysis of the maximal mitochondrial Ca2+ peak obtained (max F/F0 rhod-2 AM) (mean ± SEM, 10 cells per field analyzed, three fields per experiment, n = 3 independent experiments). (F) Representative trace of ER Ca2+ stores in cells loaded with the cytoplasmic Ca2+-sensitive dye FluoForte and stimulated with 10 mM thapsigargin (F/F0 Fluoforte, AU). (G) Quantification of the mean maximal amplitude of the thapsigargin-induced Ca2+ peak (max F/F0 Fluoforte) (mean ± SD, four fields per experiment, n = 3 biological replicates). See also Figure S5.


Figure 7. BAX and Drp1 Colocalize with SUMO1 Foci at the ER/Mitochondria Contact Sites during Cell Death (A) Immunofluorescence of si-NT- or si-MAPLsilenced HeLa cells transfected with the ER maker, ER-DsRed2 and YFP-SUMO1, and infected with AD-tBid for 9 hr. Mitochondria were stained with anti-TOM20 antibody. Zoomed area highlights ERmitochondria contact sites. Arrows indicate sites of YFP-SUMO1 recruitment to the ER-mitochondria contact, which is MAPL dependent. Scale bars: 5 mm (left) and 10 mm (right). (B) Quantification of YFP-SUMO1 foci at mitochondria and at mitochondria/ER contact sites during tBid-induced cell death from (A) (mean ± SD; n = 3 independent experiments, 30 cells in each experiment). (C) Drp1 is recruited to ER/mitochondrial contacts during cell death. Si-NT- or si-MAPL-silenced HeLa cells transfected with ER-DsRed2 and YFPSUMO1 were infected with AD-tBid for 9 hr. Mitochondria and Drp1 were stained. Zoomed area highlights the ER-mitochondria contact sites which colocalize with Drp1. Arrows indicate sites of YFP-SUMO1 and Drp1 recruitment to the ERmitochondria contact. Scale bars: 5 mm (left) and 10 mm (right). (D) Cells treated as in (A)–(C) were stained with anti-TOM20 and anti-BAX 6A7 antibodies. Arrows indicate sites of YFP-SUMO1 to mitochondria which co-localize with active BAX. Scale bars: 5 mm (left) and 10 mm (right). (E) Quantification of (D) traducing the colocalization between YFP-SUMO1-positive foci and activated BAX form in whole cells or specifically at mitochondria (mean ± SD; n = 3 independent experiments, 30 cells in each experiment). (F) A model illustrating the action of mitochondrial SUMOylation in the ER-mitochondria contact sites during cell death. MAPL-dependent SUMOylation of Drp1 stabilizes a mitochondrial/ER platform facilitating mitochondrial calcium uptake, cristae remodeling, and cytochrome c release. In the absence of MAPL, ER-mitochondria sites are destabilized and mitochondrial BAX/BAK are recruited and assembled, but they fail to function in the release of cytochrome c. See also Figure S7.

mitochondria do not support BAX-induced cytochrome c release, since the ER contacts would not be maintained (Renault et al., 2015; Szabadkai et al., 2004). We also envision that stabilized ER/mitochondrial platforms could facilitate the transfer of lipids like hexadecenal, or ceramide, into the mitochondria, both shown to facilitate BAX/ BAK-induced cytochrome c release (Chipuk et al., 2012; Colombini, 2010). Lastly, MAPL expression is controlled by the FoxC2 transcriptional machinery, and its mRNA is upregulated during muscle wasting, where it was shown to promote cell death (Lokireddy et al., 2012). MAPL/MUL1 protein levels were increased, along with increases in Drp1, Fis1, and Bnip3 within cultured models of muscle wasting. Consistent with a role for MAPL in the SUMOylation and activation of Drp1, the authors observed

increased mitochondrial fragmentation and cell death. However, they attributed this phenotype to the destabilization of Mfn2 (Lokireddy et al., 2012). We have not observed any significant changes in Mfn2 levels upon manipulation of MAPL, leading us to consider that the SUMOylation of Drp1 could drive the death phenotypes in their model. We propose that increased mitochondrial SUMOylation, driven either by the loss of SenP2, SenP5, or the overexpression of MAPL, leads to Drp1-dependent interorganellar platforms stabilization, providing a functional ‘‘landing pad’’ for what may normally be sub-lethal levels of activated BAX, tipping the cell toward death. For now, we have precisely mapped the role for mitochondrial SUMOylation in apoptosis, defining a new function for Drp1 in stabilizing ER contacts critical for the remodeling of the mitochondria and release of cytochrome c.


EXPERIMENTAL PROCEDURES Immunofluorescence Microscopy After 24 hr of silencing in 6-well plates, cells were seeded into 24-well plates upon glass coverslips placed within each well. Transfection and experimental treatments were performed within these wells. For immunofluorescence, cells were fixed in 5% paraformaldehyde (PFA) in PBS at 37 C for 15 min and then washed three times with PBS, followed by quenching with 50 mM ammonium chloride in PBS. After three washes in PBS, cells were permeabilized in 0.1% Triton X-100 in PBS, followed by three washes in PBS. Then the cells were blocked with 10% fetal bovine serum (FBS) in PBS, followed by incubation with primary antibodies in 5% FBS in PBS, for 1 hr at RT. After three washes with 5% FBS in PBS, cells were incubated with appropriate secondary antibodies (1:1,000) for 1 hr at RT. After three washes in PBS, coverslips were mounted onto slides using Dako fluorescence mounting medium (Dako). Stained or live cells were imaged using a 603 or a 1003 objective NA1.4 on an Olympus IX83 inverted microscope with appropriate lasers using Yokogawa spinning disc system microscope coupled to a Neo camera (Andor). To analyze the interaction between mitochondria and ER in healthy cells, a 0.2 mm z axis image series of cells expressing ER-DsRed2 and labeled for TOM20 were obtained and stacked. The background signals were removed and stacks were thresholded manually using ImageJ. Interaction/colocalization was quantified by Mander’s colocalization coefficient using ImageJ software. To determine the interaction between mitochondria and ER during cell death, a 0.2 mm z axis image series of cells stained with antibodies against TOM20 and KDEL were obtained and stacked. Cells were also labeled with anti-BAX-6A7 to selectively examine cells where BAX was active and recruited at mitochondria. Analysis was performed as described above. To evaluate the localization of YFP-SUMO1-positive foci recruited to mitochondria and mitochondria/ER contact points during cell death induced by 9 hr tBid, a 0.2 mm z axis image series of YFP-SUMO1 expressing cells labeled for TOM20 and KDEL were obtained and stacked. First, a mask was created between the mitochondria and ER channels to highlight an interaction map between TOM20 and KDEL channels using the ImageJ ‘‘colocalisation’’ plug-in. Then, the nucleus signal was removed and the SUMO1 foci were identified. Following this, the mitochondria/ER profile mask was merged with the YFPSUMO1 channel. Total and SUMO1-positive foci in contact with mitochondria or mitochondria/ER profile were quantified. Intracellular Calcium Analysis To measure maximal mitochondrial calcium uptake, HeLa cells were cultured on Nunc Lab-Tek chambered 8-well cover glass (Thermo Scientific). HeLa cells were washed three times in a balanced salt solution buffer (BSS) (121 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl2, 6 mM NaHCO3, 5.5 mM D-glucose, 2 mM CaCl2, and 25 mM HEPES [pH 7.3]) and incubated with 2.5 mM of the mitochondrial calcium probe Rhod-2 AM and 0.002% pluronic acid in BBS at 37 C. Cells were then washed three times in BBS and de-esterified in BSS buffer at RT. Fluorescence values were collected every 2 s, and at 20 s, 10 mM histamine final in BSS was injected for further imaging every 2 s for 4 min using the 303 objective of the Olympus IX83 inverted microscope Yokogawa spinning disc system microscope coupled with a Neo camera, exciting with 594 nm laser. Determination of the ER Ca2+ content of cells was performed as previously described (Prudent et al., 2013). Briefly, HeLa cells were washed in Ca2+-free BSS and incubated with 5 mM of the cytosolic calcium indicator FluoForte for 1 hr at 37 C. Cells were then washed in Ca2+-free BSS, and fluorescence values were collected every 2.5 s during 4 min using an inverted IX81 microscope coupled with the FV1000 confocal scanning microscope, excited with the 488 nm laser. After 20 s of measurement, thapsigargin was injected at 10 mM final. To analyze the effect of MAPL in mitochondrial calcium uptake during cell death, si-NT- or si-MAPL-silenced HeLa cells were cultured in MaTek dishes and transfected with the new genetically mito-RCamP (Akerboom et al., 2013) Ca2+ indicator probe. 24 hr following transfection, the medium was replaced by DMEM without phenol red, and live cell imaging was performed on positive transfected RCamP cells using the Viva View FL incubator microscope (Olympus). Images were acquired every 10 min using the 594 nm

channel. After 30 min of acquisition, 1 mM STS final was injected and images were acquired for additional 180 min. The mitochondrial calcium pool was analyzed by the F/F0 ratio, and a new cell ROI was designed for each time point. Image acquisition was carried out at the same gain, laser intensity, and exposure time for each experimental condition and corresponding control. Movies and images were analyzed using ImageJ Software. Statistical Analysis Errors bars displayed on graphs represent the means ± SD (or ± SEM when specified) of at least three independent experiments. Statistical significance was analyzed using unpaired Student’s t test. * p < 0.05, ** p < 0.01, and *** p < 0.001 were considered significant. SUPPLEMENTAL INFORMATION Supplemental Information includes seven figures and Supplemental Experimental Procedures and can be found with this article online at http://dx.doi. org/10.1016/j.molcel.2015.08.001. AUTHOR CONTRIBUTIONS J.P. and R.Z. performed the experiments; J.P. and S.M. performed the transmission electron microscopy; A.S. participated in data analysis; G.C.S. provided critical reagents and contributed intellectually to the development of this project; J.P. and H.M.M. designed the experiments and wrote the manuscript. ACKNOWLEDGMENTS The authors thank Liqun Xu (University of Ottawa) for technical help with the FACS analysis and Gyorgy Hajnoczky (Thomas Jefferson) and Loren Looger (HHMI Janelia Farms) for sharing the M-RCaMP plasmid. H.M.M. is a recipient of a Canada Research Chair in Mitochondrial Cell Biology. This work was funded by the Canadian Institutes of Health Research (CIHR) MOP#68833 to H.M.M. and a CCSRI Innovation grant to H.M.M. G.C.S. was supported by a CIHR MOP#6192. J.P. is supported by a CIHR postdoctoral fellowship (MFE-140925), and A.S. is supported by JSPS Postdoctoral fellowship for Research Abroad (Japan Society for the Promotion of Science). The authors declare they have no conflict of interest. Received: March 13, 2015 Revised: June 11, 2015 Accepted: August 4, 2015 Published: September 17, 2015 REFERENCES Akerboom, J., Carreras Calderoń, N., Tian, L., Wabnig, S., Prigge, M., Tolo¨, J., Gordus, A., Orger, M.B., Severi, K.E., Macklin, J.J., et al. (2013). Genetically encoded calcium indicators for multi-color neural activity imaging and combination with optogenetics. Front. Mol. Neurosci. 6, 2. Autret, A., and Martin, S.J. (2009). Emerging role for members of the Bcl-2 family in mitochondrial morphogenesis. Mol. Cell 36, 355–363. Bae, S., Kim, S.Y., Jung, J.H., Yoon, Y., Cha, H.J., Lee, H., Kim, K., Kim, J., An, I.S., Kim, J., et al. (2012). Akt is negatively regulated by the MULAN E3 ligase. Cell Res. 22, 873–885. Baughman, J.M., Perocchi, F., Girgis, H.S., Plovanich, M., Belcher-Timme, C.A., Sancak, Y., Bao, X.R., Strittmatter, L., Goldberger, O., Bogorad, R.L., et al. (2011). Integrative genomics identifies MCU as an essential component of the mitochondrial calcium uniporter. Nature 476, 341–345. Braschi, E., Zunino, R., and McBride, H.M. (2009). MAPL is a new mitochondrial SUMO E3 ligase that regulates mitochondrial fission. EMBO Rep. 10, 748–754. Cereghetti, G.M., Stangherlin, A., Martins de Brito, O., Chang, C.R., Blackstone, C., Bernardi, P., and Scorrano, L. (2008). Dephosphorylation by


calcineurin regulates translocation of Drp1 to mitochondria. Proc. Natl. Acad. Sci. USA 105, 15803–15808. Chi, X., Kale, J., Leber, B., and Andrews, D.W. (2014). Regulating cell death at, on, and in membranes. Biochim. Biophys. Acta 1843, 2100–2113.

Jiang, X., Jiang, H., Shen, Z., and Wang, X. (2014). Activation of mitochondrial protease OMA1 by Bax and Bak promotes cytochrome c release during apoptosis. Proc. Natl. Acad. Sci. USA 111, 14782–14787.

Chipuk, J.E., Moldoveanu, T., Llambi, F., Parsons, M.J., and Green, D.R. (2010). The BCL-2 family reunion. Mol. Cell 37, 299–310.

Jung, J.H., Bae, S., Lee, J.Y., Woo, S.R., Cha, H.J., Yoon, Y., Suh, K.S., Lee, S.J., Park, I.C., Jin, Y.W., et al. (2011). E3 ubiquitin ligase Hades negatively regulates the exonuclear function of p53. Cell Death Differ. 18, 1865–1875.

Chipuk, J.E., McStay, G.P., Bharti, A., Kuwana, T., Clarke, C.J., Siskind, L.J., Obeid, L.M., and Green, D.R. (2012). Sphingolipid metabolism cooperates with BAK and BAX to promote the mitochondrial pathway of apoptosis. Cell 148, 988–1000.

Kagan, V.E., Tyurin, V.A., Jiang, J., Tyurina, Y.Y., Ritov, V.B., Amoscato, A.A., Osipov, A.N., Belikova, N.A., Kapralov, A.A., Kini, V., et al. (2005). Cytochrome c acts as a cardiolipin oxygenase required for release of proapoptotic factors. Nat. Chem. Biol. 1, 223–232.

Colombini, M. (2010). Ceramide channels and their role in mitochondria-mediated apoptosis. Biochim. Biophys. Acta 1797, 1239–1244.

Karbowski, M., Lee, Y.J., Gaume, B., Jeong, S.Y., Frank, S., Nechushtan, A., Santel, A., Fuller, M., Smith, C.L., and Youle, R.J. (2002). Spatial and temporal association of Bax with mitochondrial fission sites, Drp1, and Mfn2 during apoptosis. J. Cell Biol. 159, 931–938.

Cribbs, J.T., and Strack, S. (2007). Reversible phosphorylation of Drp1 by cyclic AMP-dependent protein kinase and calcineurin regulates mitochondrial fission and cell death. EMBO Rep. 8, 939–944. Csorda´s, G., Renken, C., Va´rnai, P., Walter, L., Weaver, D., Buttle, K.F., Balla, T., Mannella, C.A., and Hajnoćzky, G. (2006). Structural and functional features and significance of the physical linkage between ER and mitochondria. J. Cell Biol. 174, 915–921.

Kim, H., Scimia, M.C., Wilkinson, D., Trelles, R.D., Wood, M.R., Bowtell, D., Dillin, A., Mercola, M., and Ronai, Z.A. (2011). Fine-tuning of Drp1/Fis1 availability by AKAP121/Siah2 regulates mitochondrial adaptation to hypoxia. Mol. Cell 44, 532–544.

de Brito, O.M., and Scorrano, L. (2008). Mitofusin 2 tethers endoplasmic reticulum to mitochondria. Nature 456, 605–610.

Korobova, F., Ramabhadran, V., and Higgs, H.N. (2013). An actin-dependent step in mitochondrial fission mediated by the ER-associated formin INF2. Science 339, 464–467.

de Brito, O.M., and Scorrano, L. (2010). An intimate liaison: spatial organization of the endoplasmic reticulum-mitochondria relationship. EMBO J. 29, 2715– 2723.

Kushnareva, Y., Andreyev, A.Y., Kuwana, T., and Newmeyer, D.D. (2012). Bax activation initiates the assembly of a multimeric catalyst that facilitates Bax pore formation in mitochondrial outer membranes. PLoS Biol. 10, e1001394.

De Stefani, D., Raffaello, A., Teardo, E., Szabo`, I., and Rizzuto, R. (2011). A forty-kilodalton protein of the inner membrane is the mitochondrial calcium uniporter. Nature 476, 336–340.

Li, H., Zhu, H., Xu, C.J., and Yuan, J. (1998). Cleavage of BID by caspase 8 mediates the mitochondrial damage in the Fas pathway of apoptosis. Cell 94, 491–501.

Elgass, K., Pakay, J., Ryan, M.T., and Palmer, C.S. (2013). Recent advances into the understanding of mitochondrial fission. Biochim. Biophys. Acta 1833, 150–161.

Li, W., Bengtson, M.H., Ulbrich, A., Matsuda, A., Reddy, V.A., Orth, A., Chanda, S.K., Batalov, S., and Joazeiro, C.A. (2008). Genome-wide and functional annotation of human E3 ubiquitin ligases identifies MULAN, a mitochondrial E3 that regulates the organelle’s dynamics and signaling. PLoS ONE 3, e1487.

Filadi, R., Greotti, E., Turacchio, G., Luini, A., Pozzan, T., and Pizzo, P. (2015). Mitofusin 2 ablation increases endoplasmic reticulum-mitochondria coupling. Proc. Natl. Acad. Sci. USA 112, E2174–E2181. Flotho, A., and Melchior, F. (2013). Sumoylation: a regulatory protein modification in health and disease. Annu. Rev. Biochem. 82, 357–385. Frank, S., Gaume, B., Bergmann-Leitner, E.S., Leitner, W.W., Robert, E.G., Catez, F., Smith, C.L., and Youle, R.J. (2001). The role of dynamin-related protein 1, a mediator of mitochondrial fission, in apoptosis. Dev. Cell 1, 515–525. Frezza, C., Cipolat, S., Martins de Brito, O., Micaroni, M., Beznoussenko, G.V., Rudka, T., Bartoli, D., Polishuck, R.S., Danial, N.N., De Strooper, B., and Scorrano, L. (2006). OPA1 controls apoptotic cristae remodeling independently from mitochondrial fusion. Cell 126, 177–189. Friedman, J.R., Lackner, L.L., West, M., DiBenedetto, J.R., Nunnari, J., and Voeltz, G.K. (2011). ER tubules mark sites of mitochondrial division. Science 334, 358–362. Fu, J., Yu, H.M., Chiu, S.Y., Mirando, A.J., Maruyama, E.O., Cheng, J.G., and Hsu, W. (2014). Disruption of SUMO-specific protease 2 induces mitochondria mediated neurodegeneration. PLoS Genet. 10, e1004579. Germain, M., Mathai, J.P., McBride, H.M., and Shore, G.C. (2005). Endoplasmic reticulum BIK initiates DRP1-regulated remodelling of mitochondrial cristae during apoptosis. EMBO J. 24, 1546–1556. Giorgi, C., Wieckowski, M.R., Pandolfi, P.P., and Pinton, P. (2011). Mitochondria associated membranes (MAMs) as critical hubs for apoptosis. Commun. Integr. Biol. 4, 334–335. Guo, C., Hildick, K.L., Luo, J., Dearden, L., Wilkinson, K.A., and Henley, J.M. (2013). SENP3-mediated deSUMOylation of dynamin-related protein 1 promotes cell death following ischaemia. EMBO J. 32, 1514–1528. Harder, Z., Zunino, R., and McBride, H. (2004). Sumo1 conjugates mitochondrial substrates and participates in mitochondrial fission. Curr. Biol. 14, 340–345. Hsu, Y.T., and Youle, R.J. (1997). Nonionic detergents induce dimerization among members of the Bcl-2 family. J. Biol. Chem. 272, 13829–13834.

Li, J., Qi, W., Chen, G., Feng, D., Liu, J., Ma, B., Zhou, C., Mu, C., Zhang, W., Chen, Q., et al. (2015). Mitochondrial outer-membrane E3 ligase MUL1 ubiquitinates ULK1 and regulates selenite-induced mitophagy. Autophagy. Published online May 27, 2015. Lokireddy, S., Wijesoma, I.W., Teng, S., Bonala, S., Gluckman, P.D., McFarlane, C., Sharma, M., and Kambadur, R. (2012). The ubiquitin ligase Mul1 induces mitophagy in skeletal muscle in response to muscle-wasting stimuli. Cell Metab. 16, 613–624. Lucken-Ardjomande, S., Montessuit, S., and Martinou, J.C. (2008). Contributions to Bax insertion and oligomerization of lipids of the mitochondrial outer membrane. Cell Death Differ. 15, 929–937. Luo, X., Budihardjo, I., Zou, H., Slaughter, C., and Wang, X. (1998). Bid, a Bcl2 interacting protein, mediates cytochrome c release from mitochondria in response to activation of cell surface death receptors. Cell 94, 481–490. Martinou, J.C., and Youle, R.J. (2011). Mitochondria in apoptosis: Bcl-2 family members and mitochondrial dynamics. Dev. Cell 21, 92–101. Montessuit, S., Somasekharan, S.P., Terrones, O., Lucken-Ardjomande, S., Herzig, S., Schwarzenbacher, R., Manstein, D.J., Bossy-Wetzel, E., Basan˜ez, G., Meda, P., and Martinou, J.C. (2010). Membrane remodeling induced by the dynamin-related protein Drp1 stimulates Bax oligomerization. Cell 142, 889–901. Mukhopadhyay, D., and Dasso, M. (2007). Modification in reverse: the SUMO proteases. Trends Biochem. Sci. 32, 286–295. Murley, A., Lackner, L.L., Osman, C., West, M., Voeltz, G.K., Walter, P., and Nunnari, J. (2013). ER-associated mitochondrial division links the distribution of mitochondria and mitochondrial DNA in yeast. eLife 2, e00422. Neuspiel, M., Schauss, A.C., Braschi, E., Zunino, R., Rippstein, P., Rachubinski, R.A., Andrade-Navarro, M.A., and McBride, H.M. (2008). Cargo-selected transport from the mitochondria to peroxisomes is mediated by vesicular carriers. Curr. Biol. 18, 102–108.


Pacher, P., and Hajnoćzky, G. (2001). Propagation of the apoptotic signal by mitochondrial waves. EMBO J. 20, 4107–4121. Parone, P.A., James, D.I., Da Cruz, S., Mattenberger, Y., Donze´, O., Barja, F., and Martinou, J.C. (2006). Inhibiting the mitochondrial fission machinery does not prevent Bax/Bak-dependent apoptosis. Mol. Cell. Biol. 26, 7397–7408. Perocchi, F., Gohil, V.M., Girgis, H.S., Bao, X.R., McCombs, J.E., Palmer, A.E., and Mootha, V.K. (2010). MICU1 encodes a mitochondrial EF hand protein required for Ca(2+) uptake. Nature 467, 291–296. Prudent, J., Popgeorgiev, N., Bonneau, B., Thibaut, J., Gadet, R., Lopez, J., Gonzalo, P., Rimokh, R., Manon, S., Houart, C., et al. (2013). Bcl-wav and the mitochondrial calcium uniporter drive gastrula morphogenesis in zebrafish. Nat. Commun. 4, 2330. Renault, T.T., Floros, K.V., Elkholi, R., Corrigan, K.A., Kushnareva, Y., Wieder, S.Y., Lindtner, C., Serasinghe, M.N., Asciolla, J.J., Buettner, C., et al. (2015). Mitochondrial shape governs BAX-induced membrane permeabilization and apoptosis. Mol. Cell 57, 69–82. Scorrano, L., Ashiya, M., Buttle, K., Weiler, S., Oakes, S.A., Mannella, C.A., and Korsmeyer, S.J. (2002). A distinct pathway remodels mitochondrial cristae and mobilizes cytochrome c during apoptosis. Dev. Cell 2, 55–67.

Szabadkai, G., Simoni, A.M., Chami, M., Wieckowski, M.R., Youle, R.J., and Rizzuto, R. (2004). Drp-1-dependent division of the mitochondrial network blocks intraorganellar Ca2+ waves and protects against Ca2+-mediated apoptosis. Mol. Cell 16, 59–68. Wasiak, S., Zunino, R., and McBride, H.M. (2007). Bax/Bak promote sumoylation of DRP1 and its stable association with mitochondria during apoptotic cell death. J. Cell Biol. 177, 439–450. Wasik, U., and Filipek, A. (2014). Non-nuclear function of sumoylated proteins. Biochim. Biophys. Acta 1843, 2878–2885. Youle, R.J., and Strasser, A. (2008). The BCL-2 protein family: opposing activities that mediate cell death. Nat. Rev. Mol. Cell Biol. 9, 47–59. Yun, J., Puri, R., Yang, H., Lizzio, M.A., Wu, C., Sheng, Z.H., and Guo, M. (2014). MUL1 acts in parallel to the PINK1/parkin pathway in regulating mitofusin and compensates for loss of PINK1/parkin. Elife 3, e01958. Zhang, B., Huang, J., Li, H.L., Liu, T., Wang, Y.Y., Waterman, P., Mao, A.P., Xu, L.G., Zhai, Z., Liu, D., et al. (2008). GIDE is a mitochondrial E3 ubiquitin ligase that induces apoptosis and slows growth. Cell Res. 18, 900–910.

Scorrano, L., Oakes, S.A., Opferman, J.T., Cheng, E.H., Sorcinelli, M.D., Pozzan, T., and Korsmeyer, S.J. (2003). BAX and BAK regulation of endoplasmic reticulum Ca2+: a control point for apoptosis. Science 300, 135–139.

Zunino, R., Schauss, A., Rippstein, P., Andrade-Navarro, M., and McBride, H.M. (2007). The SUMO protease SENP5 is required to maintain mitochondrial morphology and function. J. Cell Sci. 120, 1178–1188.

Sheridan, C., Delivani, P., Cullen, S.P., and Martin, S.J. (2008). Bax- or Bakinduced mitochondrial fission can be uncoupled from cytochrome C release. Mol. Cell 31, 570–585.

Zunino, R., Braschi, E., Xu, L., and McBride, H.M. (2009). Translocation of SenP5 from the nucleoli to the mitochondria modulates DRP1-dependent fission during mitosis. J. Biol. Chem. 284, 17783–17795.
