Bcl-2-like protein 13 is a mammalian Atg32 homologue that mediates mitophagy and mitochondrial fragmentation

Damaged mitochondria are removed by mitophagy. Although Atg32 is essential for mitophagy in yeast, no Atg32 homologue has been identified in mammalian cells. Here, we show that Bcl-2-like protein 13 (Bcl2-L-13) induces mitochondrial fragmentation and mitophagy in mammalian cells. First, we hypothesized that unidentified mammalian mitophagy receptors would share molecular features of Atg32. By screening the public protein database for Atg32 homologues, we identify Bcl2-L-13. Bcl2-L-13 binds to LC3 through the WXXI motif and induces mitochondrial fragmentation and mitophagy in HEK293 cells. In Bcl2-L-13, the BH domains are important for the fragmentation, while the WXXI motif facilitates mitophagy. Bcl2-L-13 induces mitochondrial fragmentation in the absence of Drp1, while it induces mitophagy in Parkin-deficient cells. Knockdown of Bcl2-L-13 attenuates mitochondrial damage-induced fragmentation and mitophagy. Bcl2-L-13 induces mitophagy in Atg32-deficient yeast cells. Induction and/or phosphorylation of Bcl2-L-13 may regulate its activity. Our findings offer insights into mitochondrial quality control in mammalian cells.

M itochondria are subcellular organelles that produce energy through oxidative phosphorylation. Dysregulated mitochondrial activity results in generation of reactive oxygen species as a by-product of oxidative phosphorylation, which cause damage to DNA and proteins 1 . Thus, mitochondrial quality control is essential for normal cellular functions. Macroautophagy (hereafter referred to autophagy) is responsible for mitochondrial quality control 1 . There are two types of autophagy, non-selective and selective autophagy. Non-selective autophagy sequesters bulk cytoplasm and organelles engulfed by isolation membrane as cargos to autophagosomes 2 . These then undergo fusion with lysosomes, allowing degradation of the cargo. In contrast, selective autophagy targets specific proteins or organelles as cargos, such as mitochondria and peroxisomes. The degradation of damaged mitochondria is mediated by a selective type of autophagy, mitophagy 3 . Dysregulation of mitophagy is implicated in the development of neurodegenerative diseases, such as Alzheimer's disease and Parkinson's disease as well as metabolic diseases, heart failure and ageing 3 .
Mitochondrial morphologies change continuously through actions of fission and fusion (collectively termed mitochondrial dynamics). In yeast 4 and mammalian cells 5 , mitophagy is reported to be preceded by mitochondrial fission, which divides elongated mitochondria into pieces of manageable size for engulfment by isolation membrane. To date, more than 30 autophagy-related (Atg) genes have been identified, which function as molecular machinery for autophagy 2 . In yeast, Atg32 is essential for mitophagy and functions as a receptor of mitophagy through its interaction with Atg8 and Atg11 (ref. 6,7). It has a single transmembrane domain in the C-terminal fifth of the protein, spanning outer mitochondrial membrane (OMM) and contains a WXXI motif, which binds to Atg8. Based on amino acid similarity, Atg32 has no mammalian homologue.
In mammals, mitophagy is involved in mitochondria elimination from reticulocytes, which is mediated by NIP3-like protein X (NIX, also known as BNIP3L) 8 . It is also reported that FUNDC1, localized in OMM, is a receptor for hypoxia-induced mitophagy 9 . The OMM kinase, phosphatase and tensin homolog (PTEN)-induced putative kinase protein 1 (PINK1) and the cytosolic E3 ubiquitin ligase Parkin, the mutations of which are causative for hereditary Parkinson's disease, are known to mediate mitophagy to eliminate damaged mitochondria in many types of cells 10 . Parkin is expressed in most of adult tissues, but some fetal tissues and cell lines including HeLa cells show little or no endogenous Parkin expression [11][12][13] . Parkindeficient mice show only mild phenotypes 14 . Thus, it is reasonable to assume that there may be an unknown receptor for mitophagy in mammalian cells. Here, we show that Bcl2-L-13 induces mitochondrial fragmentation and mitophagy in mammalian cells and can function as a mitophagy receptor when it is expressed in yeast.

Results
Identification of Bcl2-L- 13. In this study, we hypothesized that a mammalian mitophagy receptor will share the following molecular features with Atg32: mitochondrial localization; WXXL/I motifs; acidic amino acid clusters; and single membrane-spanning topology. Using this molecular profile of Atg32 as a search tool, we screened UniProt database (http://www.uniprot.org/) for novel Atg32 functional homologues and identified Bcl-2-like protein 13 (Bcl2-L- 13).
Mouse Bcl2-L-13 gene (Bcl2l13) encodes a protein of 434 amino acids, which contains a C-terminal single transmembrane domain 15 (Fig. 1a). Bcl2-L-13 is expressed in all tissues and cell lines tested including HeLa cells and localized in mitochondria. It has been known that Bcl2-L-13 contains four conserved Bcl-2 homology domains (BH1-4) and triggers cell death independently of the BH domains. In contrast to the previous report, overexpression of Bcl2-L-13 did not induce the activation of caspase 3 in HEK293 cells (Fig. 1b). Sequence alignment of mouse Bcl2-L-13 revealed 2 WXXL/I motifs at positions 147-150 and 273-276 (Fig. 1a,c).
Bcl2-L-13 is an outer mitochondrial membrane protein.
To confirm that Bcl2-L-13 is an integral mitochondrial membrane protein, the mitochondrial fraction from HEK293 cells was treated with Na 2 CO 3 , which results in release of soluble and peripheral proteins 18,19 (Fig. 1g). Bcl2-L-13, the outer-membrane protein, Tom20 and the inner-membrane protein, Tim23 were retained in the pellet, whereas cytochrome c was released to the supernatant, suggesting Bcl2-L-13 is an integral mitochondrial membrane protein. Treatment of mitochondrial fraction with proteinase K resulted in almost complete digestion of Bcl2-L-13 and Tom20, but a significant amount of Tim23 remained (Fig. 1h). These results indicate that Bcl2-L-13 is an integral OMM protein. Since the anti-Bcl2-L-13 antibody used in this study recognizes amino acids near the centre of Bcl2-L-13, Bcl2-L-13 has its N-terminus exposed to the cytosol and C-terminus in the intermembrane space.
The relationship between Bcl2-L-13 and Drp1. Dynaminrelated protein 1 (Drp1) is known to be the master regulator of mitochondrial fission 1 . Mitochondrial fission 1 (Fis1) is a partner protein of Drp1. On the other hand, mitofusins, Mfn1 and Mfn2, together with optic atrophy protein 1 (Opa1) are the core components of the mitochondrial fusion machinery. Knockdown of Drp1 resulted in elongation of mitochondria and overexpression of HA-Bcl2-L-13 was still able to induce mitochondrial fragmentation in Drp1 knockdown cells (Fig. 4a-c). However, the ratio of cells with fragmented Phosphorylation of Drp1 on Ser637 inhibits mitochondrial fission, while phosphorylation of that on Ser616 promotes mitochondrial fission 20 . Overexpression of Bcl2-L-13 increased the amount of phosphorylated Drp1 on Ser637 and decreased that on Ser616 (Fig. 4d). On the other hand, knockdown of Bcl2-L-13 increased the amount of phosphorylation on Ser616, but not on Ser637, when compared with control siRNA cells. Dimer formation of the dynamin-related proteins such as Drp1 increases their GTPase activity 20 . Overexpression of Bcl2-L-13 led to a decrease in dimer formation of Drp1 (Fig. 4e). These data suggest that Drp1 appeared to compensate Bcl2-L-13-induced changes in mitochondrial morphology. Thus, Drp1 is not essential for Bcl2-L-13-induced mitochondrial fragmentation.
Changes in Bcl2-L-13 expression had no effect on the protein level of Fis1 or the molecules involved in fusion such as Mfn1, Mfn2 and Opa1 (Fig. 4d,f). In addition, HA-Bcl2-L-13 was able to induce mitochondrial fragmentation in Bak or Bax knockdown cells (Fig. 4g,h).
Bcl2-L-13 induces mitophagy. Then, we investigated the role of Bcl2-L-13 in mitophagy. Conversion of LC3-I to LC3-II is an essential step for autophagosome formation 21 . HA-Bcl2-L-13 increased LC3-II protein level (left top blot and right upper panel in Fig. 5a). To assess the autophagy flux, we treated cells with a lysosomal inhibitor, bafilomycin A1. Treatment with bafilomycin A1 led to an increase in protein level of LC3-II. For quantitative analysis, we exposed the transfer membrane to the film for a shorter period (the second blot from the top blot in Fig. 5a). The level of LC3-II in HA-Bcl2-L-13 overexpressing cells was higher than that in cells transfected with empty vector or Bcl2-L-13 W273A I276A after treatment with bafilomycin A1 (right lower panel in Fig. 5a). This indicates that the increase in LC3-II level in HA-Bcl2-L-13 overexpressing cells is not because of reduced autophagosome turnover, but increased autophagic flux. We produced a stable cell line of HEK293 expressing mitochondrial targeted Keima (mKeima), a coral-derived acid-stable lysosomal proteases-resistant fluorescent protein 22 . We used the excitation of mKeima at 559 nm that causes emission when the molecule is in acidic compartments such as lysosome as well as in neutral compartments 22 (Fig. 5b). We transfected the cells with HA-Bcl2-L-13 and GFP-LC3 and incubated with protease inhibitors to derive sufficient number of autophagosomes or autolysosomes for analysis. Bcl2-L-13 increased the number of LC3-and mKeima-positive dots (Fig. 5c). Next, mKeima-expressing HEK293 cells were transfected with HA-Bcl2-L-13 and stained with LysoTracker Green before microscopic analysis (Fig. 5d,e). HA-Bcl2-L-13 increased the number of LysoTracker-and mKeima-positive dots. These results indicate that mitochondria were engulfed in autophagosomes or autolysosomes. Ultrastructural analysis of HA-Bcl2-L-13 expressing HEK293 cells revealed double-membrane vacuoles containing a single mitochondrion-like structure, but little cytoplasm (Fig. 5f). HA-Bcl2-L-13 significantly reduced mitochondrial DNA amount (Fig. 5g). These indicate that Bcl2-L-13 induced mitophagy. The mutations in the LIR domain reduced mitophagic activity (Fig. 5a,c-g). The mutants in the BH domains were unable to induce mitochondrial fragmentation and subsequent mitophagy (Fig. 5d,e). Thus, Bcl2-L-13 induces mitophagy through the interaction with LC3.
To examine whether Bcl2-L-13-induced mitophagy is coupled to the PINK1/Parkin-mediated pathway, we expressed HA-Bcl2-L-13 in Parkin knockdown HEK293 cells harbouring mKeima.  Bcl2-L-13 induced a similar level of mitophagy in Parkin knockdown HEK293 and control cells, as indicated by the number of LysoTracker-and mKeima-positive dots (Fig. 6a). Furthermore, Bcl2-L-13 was able to induce mitophagy in HeLa cells, which reportedly lack a functional Parkin gene 12 (Fig. 6b). It has been reported that the mitochondria were maintained after adding CCCP in HeLa cells, whereas few mitochondria remained detectable in Parkin expressing HeLa cells, assessed by immunocytochemistry using anti-Tom20 antibody 17 . We confirmed the effect of Parkin on CCCP-treated HeLa cells (Fig. 6c). Similar selective mitochondrial elimination by CCCP treatment was observed in Bcl2-L-13 expressing HeLa cells. These indicate that Parkin is not necessary for Bcl2-L-13 to induce mitophagy. Bcl2-L-13 functions as a mitophagy receptor in yeast. Atg32 is essential for mitophagy under respiratory conditions 6,7 . Yeast cells expressing the mitochondrial matrix-targeted dehydrofolate reductase-mCherry protein (mito-dihydrofolate reductase (DHFR)-mCherry) were grown under respiratory conditions 23 . It has been reported that mitophagy in yeast is strongly activated in cells at stationary phase, which seems to be triggered by oxidative stress 6 . Upon mitophagy, this fusion protein is transported and processed to generate free mCherry in the vacuole. In yeast lacking Atg32, the processing of mito-DHFR-mCherry barely occurred. We replaced the transmembrane domain in Bcl2-L-13 with mitochondrial tail-anchor (TAmito) domain derived from an authentic outer membrane protein to facilitate mitochondrial localization of the expressed protein in yeast 23 . Surprisingly, expression of Bcl2-L-13 (1-407)-TAmito in atg32D yeast generated free mCherry, suggesting partial restoration of the ability to induce mitophagy, but the LIR mutant did not (Fig. 7a). Furthermore, Bcl2-L-13 failed to induce mitophagy in atg7D (Fig. 7b), suggesting Bcl2-L-13-induced mitophagy is mediated through known autophagy molecular machinery. These results indicate that Bcl2-L-13 can substitute for Atg32 in yeast. To confirm that Bcl2-L-13 induces mitophagy in mammalian cells, we generated a fusion construct of mCherry and Tom22 to target mCherry to mitochondria, and transfected HEK293 cells with mCherry-Tom22 and wildtype or the indicated mutant HA-Bcl2-L-13. Overexpression of Bcl2-L-13 generated free mCherry in HEK293 cells (Fig. 7c).
The protein level of processed mCherry in cells transfected with Bcl2-L-13 was more than that in Bcl2-L-13 W273A I276A overexpressing cells.
Knockdown of Bcl2-L-13 attenuated CCCP-induced fragmentation and mitophagy, indicating that endogenous Bcl2-L-13 plays an important role in CCCP-induced mitochondrial fragmentation and mitophagy. Then, we attempted to elucidate the activation mechanism of Bcl2-L-13 to induce mitochondrial fragmentation and mitophagy. Atg32 is temporally upregulated to induce mitophagy and subsequently degraded 6,7 . The Atg11-Atg32 interaction is believed to be the initial molecular step, in which the autophagic machinery recognizes mitochondria as a cargo. Phosphorylation of the Ser-114, close to Atg8-interaction site (amino acid residue 86-89), in Atg32 mediates the Atg11-Atg32 interaction and mitophagy 24,25 . Thus, induction of the protein and/or post-translational modification of Bcl2-L-13 might regulate its function. The protein level of Bcl2-L-13 was increased 1 h after CCCP treatment compared with control, then returned to baseline thereafter (Fig. 8d). Overexpression of Bcl2-L-13 induced its Ser/Thr phosphorylation (Fig. 8e). We mutated Ser272 to Ala in Bcl2-L-13, which is close to the second LIR motif. The phosphorylation level of Bcl2-L-13 S272A was significantly attenuated compared with control. Bcl2-L-13 S272A showed less ability for binding with LC3 (Fig. 1e). Overexpression of Bcl2-L-13 S272A induced mitochondrial fragmentation, but decreased the LC3-II protein level and the number of LC3 dots colocalized with ATP synthase (Fig. 8f-h). The amount of processed mitochondrial targeted mCherry in Bcl2-L-13 S272A overexpressing cells was less than that in cells transfected with wild-type Bcl2-L-13 (Fig. 7c). Another possible mechanism to regulate the function of Bcl2-L-13 will be ubiquitination. CCCP induced ubiquitination of mitochondria in Parkin overexpressing cells as previously reported 26 , but not in Bcl2-L-13 overexpressing cells (Fig. 9), excluding the involvement of ubiquitination in the regulation mechanism underlying the functions.

Discussion
Although Atg32 is essential for as mitophagy in yeast 6,7 , no mammalian homologue has been identified. Here, we demonstrate that Bcl2-L-13 is involved in mitophagy as well as mitochondrial fragmentation. Most surprisingly, Bcl2-L-13 exhibited the ability to compensate the function of Atg32 in yeast.
In contrast to our results, Kataoka et al. 15 reported that overexpression of Bcl2-L-13 resulted in caspace-3 activation and cytochrome c release from mitochondria in HEK293T cells. It has been reported that Bcl2-L-13 shows no interaction with either anti-apoptotic (Bcl-2, Bcl-xL, Bcl-w, A1, MCL-1, E1B-19K and BHRF1) or pro-apoptotic (Bax, Bak, Bik, Bid, Bim and Bad) members of the Bcl-2 family, even though it has BH1-4 domains 15 . Although we do not know the exact reason for the discrepancy, some non-specific effects of its overexpression may result in the induction of apoptosis.
Our results indicate that Bcl2-L-13 is localized to OMM and interacts with LC3 through the conserved LIR sequence. In yeast, Atg32 interacts with Atg8 and Atg11 (refs 6,7), suggesting Atg32 recruits Atg8 and Atg11 and they and other core Atg proteins cooperatively generate isolation membranes surrounding mitochondria. Interestingly, there is no known mammalian homologue of Atg11 (ref. 27). Although the detailed mechanism how Bcl2-L-13 mediates mitophagy is not entirely understood, we can speculate that Bcl2-L-13 recruits LC3 to the surface of mitochondria, leading to the formation of mitochondria-specific autophagosomes (mitophagosomes). Bcl2-L-13 may bind to an unidentified mammalian homologue of Atg11 or may play a role as a scaffold in a similar fashion with Atg11. Mitophagy is closely linked to mitochondrial dynamics. Thus, it is possible that Bcl2-L-13-induced mitophagy is a consequence of mitochondrial fragmentation. However, our results showing that the mutation in the LIR motif attenuated mitophagy without affecting mitochondrial fragmentation and that Bcl2-L-13 induced mitophagy in Atg32-deficient yeast strongly indicate that Bcl2-L-13 has the dual effects. The molecular mechanisms of Bcl2-L-13-mediated mitochondrial fragmentation remain to be elucidated. Mitochondria exist largely as an extended, reticular structure. For engulfment of mitochondria by isolation membranes, the mitochondrial size will be an issue. In yeast, Atg11 recruits Dnm1 (a yeast homologue of Drp1), the fission machinery to drive mitophagy, suggesting that Atg 11 plays a role as a scaffold that recruits the fission components in addition to its role in connecting the damaged mitochondria with the autophagy machinery and mitochondrial fragmentation and mitophagy occur in a coordinated manner 4 . Thus, the mitochondria destined for degradation will be selected first, and then the fission machinery would be recruited to drive the separation of these mitochondrial from the mitochondrial network and finally the fragmented mitochondria are degraded mediated through autophagy 4 . Since Bcl2-L-13 has the dual effects, the coordination will operate more efficiently. We showed that BH1-4 motifs are involved in Bcl2-L-13-induced fragmentation. Identification of a binding protein to the BH domains in Bcl2-L-13 may elucidate a molecular mechanism underlying the coordinated interaction between fission and mitophagy.
CCCP is the most popular stimulus to induce mitochondrial damage in cell biological research, which induces mitochondrial fragmentation and mitophagy 17 , although the ability of CCCP to induce mitophagy is not so potent as it decreases the total amount of mitochondrial protein level in cells 28 (Fig. 8d). We showed that Bcl2-L-13 is essential for CCCP-induced mitochondrial fragmentation and mitophagy, suggesting the physiological importance of Bcl2-L-13. This leads to the question of how Bcl2-L-13 is related to known pathways for mitochondrial fragmentation and mitophagy. Our results indicate that Drp1 is not essential for Bcl2-L-13-induced mitochondrial fragmentation. However, we found the existence of elongated mitochondria and reduced ratio of cells with fragmented mitochondria in Drp1 knockdown cells expressing Bcl2-L-13. These results suggest that there may be some interaction between the two pathways. In addition, we showed that Parkin is not necessary for Bcl2-L-13 to induce mitophagy. The Parkin-deficient mice have only modest phenotypes 14,29 and Parkin is not expressed in all cell types, ARTICLE suggesting the presence of alternate pathways of mitophagy. NIX and FUNDC1 mediate mitophagy in mammalian cells 8,9,30 . These are involved in specific types of mitophagy; the former for mitochondrial elimination from reticulocytes, the latter for hypoxia-induced mitophagy. However, we are not able to exclude the possibility that Bcl2-L-13 can cooperate with NIX and/or FUNDC1 for mitophagy. Germ-line gene ablation of Bcl2-L-13 will provide its in vivo physiological role in mitochondrial quality and quantity control. The molecular mechanism underlying the activation of Bcl2-L-13 to induce mitochondrial fragmentation and mitophagy remains to be elucidated. The induction of the protein might regulate mitochondrial fragmentation and/or mitophagy. Furthermore, the Ser272 phosphorylation of Bcl2-L-13 regulates mitophagy, but not mitochondrial fragmentation. When mitochondrial fragmentation is induced, unknown kinase(s) activates the Bcl2-L-13 by phosphorylation of its Ser residue close to the LIR domain to allow recruitment of mitophagy machinery.
Dysregulation of mitophagy is implicated in the development of many chronic diseases including neurodegenerative diseases, metabolic diseases and heart failure 3 . Our study will provide a novel insight into molecular mechanisms of the pathogenesis of such diseases.
Cell culture and transfection. HEK293A and HeLa cells were purchased from ATCC. HEK293A cells were grown in Dulbecco's modified Eagle's medium (Sigma) supplemented with 10% fetal bovine serum (Sigma) and 1% penicillinstreptomycin (Life Technologies) at 37°C under 5% CO 2 . Transient transfections were performed using Lipofectamine 2000 (Life Technologies) or ScreenFect A (Wako) according to the manufacturer's instructions. After 48 h of transfections, cells were subjected to analysis, unless otherwise indicated. Mitophagy was induced in HEK293A and HeLa cells by the treatment with 5 and 10 mM of CCCP for the indicated time, respectively. Control cells were treated with the vehicle, dimethylsulphoxide (DMSO).
Plasmid constructions. Constructs were obtained by conventional restriction enzyme-based cloning. Site-directed mutagenesis was performed using the QuikChange Site-Directed Mutagenesis Kit according to the supplier's instructions (Agilent Technologies).
cDNA encoding LC3B or Bcl2-L-13 was obtained by RT-PCR from C57B/6J mouse heart total RNA with LC3B forward (5 0 -GAATTCATGCCGTCCGAGA AGACCTTC-3 0 ) and reverse (5 0 -GAATTCGTCCGCTGGTAACATCCCTT-3 0 ) primers, and Bcl2-L-13 forward (5 0 -AGCAGAATTCATGGCGTCCTCTACGAC TGC-3 0 ) and reverse (5 0 -AGCAGAATTCTTACTTTCTTCTTAAAGCCAGT Yeast transfected with wild-type or LIR mutant HA-Bcl2-L-13 or empty vector were collected at the indicated time points after induction of mitophagy and subjected to western blotting for mCherry. The red arrows depict free mCherry generated by mitophagy. Yeast strains are wild-type (WT) or atg32D in (a) or atg32D or atg7D derivatives in (b) expressing a mitochondrial matrix-targeted DHFR-mCherry. Quantitative analysis for free mCherry in (a) is shown in the right graph (n ¼ 3). The value for wild-type yeast strain was set equal to 1. Results are shown as the mean±s.e.m. *Po0.05 versus all other groups. (c) HEK293 cells were transfected with mCherry-Tom22 and wild-type or the indicated mutant HA-Bcl2-L-13. Seventy-two h after transfection, cells were lysed and subjected to western blotting for mCherry. Processed mCherry-Tom22 was detected as free mCherry. Quantitative analysis for free mCherry is shown in the right graph (n ¼ 3). The value for vector transfected cells was set equal to 1. Results are shown as the mean ± s.e.m. *Po0.05 versus all other groups. A one-way ANOVA followed by Tukey-Kramer's post hoc test was used for statistical analysis.
All plasmid constructs were verified by restriction digestion and/or DNA sequencing. pMT-mKeima-Red was purchased from MBL. pEGFP-LC3 was obtained from Prof. Noboru Mizushima 21 .
Establishment of mKeima stable expression cell line. The pMT-mKeima-Red plasmid was transfected into HEK293A cells using calcium phosphate method. After 48 h, the cells were passaged and 1 mg ml À 1 G418 for selection was added 24 h later. After 14 days, the single colonies were transferred to 24-well plates and expanded.
Yeast two-hybrid assay. Bcl2-L-13 or its mutants containing W147A L150A or W273A I276A substitution and LC3B were cloned into pGADT7 and pGBKT7, respectively. The cloned constructs were co-introduced into AH109 yeast strain using lithium acetate/polyethylene glycol with herring testis carrier DNA. The transformants were spotted on agar plates containing a synthetic dropout medium (Clontech) lacking Leu and Trp for maintenance of the plasmids, and those additionally lacking His and Ade to suppress background, and grown at 30°C for 4 days.
HEK293A cells were transfected with 250 ng of mCherry-Tom22 and 750 ng of wild-type or the HA-Bcl2-L-13 mutants. Seventy-two h after transfection, cells were lysed and subjected to western blotting for mCherry. Processed mCherry-Tom22 was detected as 30 kDa protein band of free mCherry.
Live cell and immunofluorescence microscopy. HEK293A cells stably expressing mKeima were seeded at 5.0 Â 10 4 ml À 1 in a glass-based dish (Iwaki) and transfected 24 h later using ScreenFect A according to the manufacturer's instructions. The cells were imaged 48 h after transfection, otherwise indicated, by FV1000-D (Olympus) equipped with cell culture incubator. To visualize the colocalization of autophagosomes and mitochondria, cells transfected with pEGFP-LC3 and Bcl2-L-13 constructs were treated with E64d (10 mg ml À 1 ) and pepstatin A (10 mg ml À 1 ) (Peptide Institute) for 4 h prior to analysis. We excited mKeima at 559 nm and collected the emission from 570 to 670 nm to monitor mitochondria.
For immunostaining, cells were plated on sterile coverslip (22 mm diameter). After treatment, cells were fixed with 4% formaldehyde at 37°C for 10 min, permeabilized with 0.2% Triton X-100 for 15 min, and then blocked with 2% bovine serum albumin for 1 h at room temperature (RT). For immunostaining of endogenous LC3, cells were fixed and permeabilized with methanol for 10 min at À 20°C. Cells were incubated with primary antibodies overnight at 4°C followed by staining with secondary antibodies for 1 h at RT. After washing, cells were mounted with Vectashield mounting medium (VECTOR Laboratories) and analysed by LSM 510 (Zeiss) or FV1000-D.
To visualize the mitochondria or lysosomes, cells were stained with 100 nM MitoTracker Deep Red FM (Molecular Probes), 100 nM MitoTracker Green or 50 nM LysoTracker Green (Molecular Probes) for 30 min before confocal microscopic analysis or fixation. To evaluate mitochondrial membrane potential, cells were stained with 100 nM TMRE (Molecular Probes) for 30 min before confocal microscopic analysis. HEK293A cells expressing FLAG-Parkin or HA-Bcl2-L-13 were incubated with 10 mM CCCP for 3 h. Cells were stained antibodies for Tom20 and ubiquitin. Scale bar, 10 mm.