Bak regulates mitochondrial morphology and pathology during apoptosis by interacting with mitofusins

Abstract

Mitochondrial injury, characterized by outer membrane permeabilization and consequent release of apoptogenic factors, is a key to apoptosis of mammalian cells. Bax and Bak, two multidomain Bcl-2 family proteins, provide a requisite gateway to mitochondrial injury. However it is unclear how Bax and Bak cooperate to provoke mitochondrial injury and whether their roles are redundant. Here, we have identified a unique role of Bak in mitochondrial fragmentation, a seemingly morphological event that contributes to mitochondrial injury during apoptosis. We show that mitochondrial fragmentation is attenuated in Bak-deficient mouse embryonic fibroblasts, baby mouse kidney cells, and, importantly, also in primary neurons isolated from brain cortex of Bak-deficient mice. In sharp contrast, Bax deficiency does not prevent mitochondrial fragmentation during apoptosis. Bcl-2 and Bcl-XL inhibit mitochondrial fragmentation, and their inhibitory effects depend on the presence of Bak. Reconstitution of Bak into Bax/Bak double-knockout cells restores mitochondrial fragmentation, whereas reconstitution of Bax is much less effective. Bak interacts with Mfn1 and Mfn2, two mitochondrial fusion proteins. During apoptosis, Bak dissociates from Mfn2 and enhances the association with Mfn1. Mutation of Bak in the BH3 domain prevents its dissociation from Mfn2 and diminishes its mitochondrial fragmentation activity. This study has uncovered a previously unrecognized function of Bak in the regulation of mitochondrial morphological dynamics during apoptosis. By this function, Bak may collaborate with Bax to permeabilize the outer membrane of mitochondria, unleashing the apoptotic cascade.

Fig. 1.

Mitochondrial fragmentation precedes cytochrome c release during apoptosis and is inhibited by dominant-negative Drp1 mutant and Bcl-2. (A) Representative cell images showing mitochondrial fragmentation and cyt.c release after azide treatment. HeLa cells were transfected with MitoRed to fluorescently label mitochondria in red. The transfected cells were subjected to control incubation (Control) or 3 h of 10 mM azide treatment (Azide). Cyt.c was stained in green by indirect immunofluorescence. Images of MitoRed-labeled mitochondria (MitoRed) and cyt.c immunofluorescence (Cyt.c) were collected by confocal microscopy. (B) Time courses of mitochondrial fragmentation, cyt.c release, and apoptosis. HeLa cells transfected with MitoRed were treated with 10 mM azide for indicated time and stained for cyt.c immunofluorescence. Percentages of cells showing mitochondrial fragmentation, cyt.c release, and apoptotic morphology were evaluated by cell counting. Data are means ± SD of three separate experiments. (C–E) Effects of VAD, Drp1, dn-Drp1, and Bcl-2 on mitochondrial fragmentation, cyt.c release, and apoptosis. HeLa cells were cotransfected with MitoRed and dn-Drp1, Drp1, or Bcl-2. The cells were then treated with 10 mM azide for 3 h in the presence or absence of 100 μM VAD. Percentages of mitochondrial fragmentation, cyt.c release, and apoptosis in MitoRed-labeled cells were quantified by cell counting. Data are means ± SD of three separate experiments. *, significantly different from the untreated (Control) group; #, significantly different from the treated no-addition (NA) group.

Fig. 2.

Bak (but not Bax) deficiency blocks mitochondrial fragmentation during apoptosis. Wild-type (wt), Bax-knockout (Bax^−/−), Bak-knockout (Bak^−/−), and Bax/Bak double-knockout (DKO) MEFs were subjected to apoptotic treatments with 10 mM azide for 3 h, 1 μM STS for 4 h, or 20 μM cisplatin for 16 h. To evaluate mitochondrial fragmentation, the MEFs were transfected with MitoRed before apoptotic treatments. (A) Cells with mitochondrial fragmentation were examined by fluorescence microscopy and quantified by cell counting. (B) To analyze cyt.c release, MEFs after cisplatin treatment were extracted to collect the cytosolic fraction for immunoblot analysis of cyt.c. (C) To analyze apoptosis, MEFs after cisplatin treatment were stained with Hoechst 33342. Apoptosis was evaluated by counting of the cells with typical apoptotic morphology including cellular and nuclear condensation and fragmentation. Data are presented as means ± SD of four separate experiments. *, significantly different from the untreated control group; #, significantly different from the treated wild-type group.

Fig. 3.

Bcl-2 inhibits mitochondrial fragmentation in Bax-knockout cells but not in Bak-knockout cells. MEF cells of different genotypes wild-type (A), Bax-knockout (B), Bak-knockout (C), and Bax/Bak double knockout (D) were transfected with MitoRed alone (NA) or cotransfected with MitoRed and Bcl-2 (Bcl-2). The cells were then subjected to apoptotic treatment with 1 μM STS. Mitochondrial fragmentation was examined by fluorescence and confocal microscopy and quantified by cell counting. Data are means ± SD of three separate experiments. ∗, significant difference between the nontransfected NA group and the Bcl-2-transfected group.

Fig. 4.

Bak is more effective than Bax in restoring mitochondrial fragmentation in Bax/Bak double-knockout cells. (A) Induction of mitochondrial fragmentation by Bax or Bak transfection in Bax/Bak double-knockout (DKO) MEF cells. DKO cells were cotransfected with MitoRed and GFP, GFP-Bax, or GFP-Bak. At indicated time points, the cotransfected cells were examined by fluorescence and confocal microscopy to evaluate mitochondrial fragmentation. Data are means ± SD of five separate experiments. ∗, significantly different from GFP-transfected group; #, significantly different from GFP-Bax-transfected group. (Inset) Immunoblot analysis of GFP-Bax and GFP-Bak expression by using an anti-GFP antibody. (B) Restoration of azide-induced mitochondrial fragmentation in DKO cells by Bak but not by Bax. DKO cells were cotransfected with MitoRed and GFP, GFP-Bax, or GFP-Bak for 16 h. The cells were then incubated for another 3 h with control medium or 10 mM azide. Mitochondrial fragmentation in transfected cells was examined by fluorescence and confocal microscopy. Data are means ± SD of four separate experiments. ∗, significant difference between the treated (+Azide) and untreated (−Azide) group. (C) Representative images of MitoRed/GFP, MitoRed/GFP-Bax, and MitoRed/GFP-Bak cotransfected cells after azide treatment. Shown are transfected cells with superimposed MitoRed and GFP signals (Upper) and magnified areas with separated MitoRed and GFP signals (Lower).

Fig. 5.

Bak interaction with mitofusins: changes during apoptotic treatment and effects of BH3 mutation. (A) HeLa cells were untreated or treated with 10 mM azide for 3 h. Whole lysates were collected with CHAPS buffer and subjected to immunoprecipitation using an anti-Bak antibody. The resultant immunoprecipitates were analyzed for Fis1, Drp1, Mfn1, Mfn2, and Bak by immunoblotting. (B) HeLa cells were transfected with Myc-Mfn1 or Myc-Mfn2. The cells were untreated or treated with azide to collect whole-cell lysates for immunoprecipitation using an anti-Myc antibody. The resultant immunoprecipitates were analyzed for Bak, Bax, Myc-Mfn1, or Myc-Mfn2 by immunoblotting. (C) Bax/Bak double-knockout MEFs were cotransfected with Mito-Red and Bak, mBak, or empty vector. Cells with mitochondrial fragmentation were examined and counted by fluorescence microscopy. (D and E) Bax/Bak double-knockout MEFs were cotransfected with Bak or mBak and Myc-Mfn1 or Myc-Mfn2. The cells were then untreated or treated with azide to collect whole-cell lysates for immunoprecipitation using an anti-Myc antibody. The resultant immunoprecipitates were analyzed for Bak, mBak, and Myc-Mfn1 or Myc-Mfn2 by immunoblotting. Results in A, B, D, and E are representatives of at least three separate experiments. Data in C are presented as means ± SD of three separate experiments. ∗, significantly different from the empty vector transfection group.