The proapoptotic protein BNIP3 interacts with VDAC to induce mitochondrial release of endonuclease G

Abstract

BNIP3 is a proapoptotic protein that induces cell death through a mitochondria-mediated pathway. We reported previously that mitochondrial localization of BNIP3 and translocation of EndoG from mitochondria to the nucleus are critical steps of the BNIP3 pathway. It is not clear, however, that how BNIP3 interacts with mitochondria. Here we show that expression of BNIP3 resulted in mitochondrial release and nuclear translocation of EndoG. Incubation of a recombinant GST-BNIP3 protein with freshly isolated mitochondria led to the integration of BNIP3 into mitochondria, reduction in the levels of EndoG in mitochondria and the presence of EndoG in the supernatant that was able to cleave chromatin DNA. Co-immunoprecipitation and mass spectrometry analysis reveals that BNIP3 interacted with the voltage-dependent anion channel (VDAC) to increase opening probabilities of mitochondrial permeability transition (PT) pores and induce mitochondrial release of EndoG. Blocking VDAC with a VDAC antibody largely abolished mitochondrial localization of BNIP3 and prevented EndoG release. Together, the data identify VDAC as an interacting partner of BNIP3 and support endonuclease G as a mediator of the BNIP3 pathway.

Figure 1. Expression of BNIP3 results in EndoG translocation.

A, SH-SY5Y human neuroblastoma cells cells transfected with the full length BNIP3 or the mutant BNIP3ΔTM plasmids were immune-labeled with antibodies to BNIP3 and EndoG and counterstained with Hoechst 33342. Majority of BNIP3 expressing cells (82%) showed nuclear translocation of EndoG, while only 12% of cells expressing BNIP3ΔTM showed EndoG translocation. B and C, Western blot analysis of EndoG in mitochondrial and nuclear samples prepared from SH-SY5Y cells transfected with the indicated plasmids. Results shown represent the mean ±SD for combined data from 4 independent experiments. **, p<0.01.

Figure 2. BNIP3-induced mitochondrial release of EndoG.

Freshly isolated mitochondria were incubated with the indicated recombinant proteins for 1 h and then separated into mitochondrial and supernatant fractions. In the group treated with both Bcl-2 and BNIP3, both were used at 0.5 µM. A and B, Western blotting analysis of BNIP3 and EndoG in the mitochondrial and supernatant fractions. Loading controls were performed with Cox IV and HSP60 antibodies. Results shown represent the mean ±SD for combined data from three independent experiments. ΔTM ^#, = BNIP3ΔTM; *, p<0.05; **, p<0.01. C, Amino acid sequence of EndoG. The tryptic peptide identified from the supernatant samples by mass fingerprinting is indicated in bold and underlined. D, Western blot analysis of cytochrome c and AIF. Both were not detectable in the supernatant samples.

Figure 3. EndoG released from isolated mitochondria following BNIP3 induction is enzymatically active.

Freshly isolated mitochondria were incubated with various recombinant proteins for 1 h at 37°C. The supernatants were collected and incubated with HEK293 nuclei for 2 h at 37°C. The DNA samples were loaded on a horizontal 1.8% agarose gel for DNA laddering assay. Shown is a representative image selected from three independent experiments.

Figure 4. BNIP3 induces MPT by interacting with VDAC.

A, BNIP3-induced MPT. Calcein release was presented as the percentage of calcein escaped from mitochondria during the 30-minute incubation. **, p<0.01 when compared with the control. B, Inhibition of PT pores reduced BNIP3-induced EndoG release. Cyclosporine A (2 µM) was added to mitochondria in CFS buffer 10 min before the BNIP3-GST protein at different final concentrations as indicated. BNIP3, CoxIV and EndoG in mitochondrial and supernatant fractions were detected by Western blotting. C, Distribution of EndoG was determined by analyzing the total intensities of the EndoG bands. D, Identification of VDAC as an interacting partner of BNIP3. After mitochondria-BNIP3 incubation, interacting proteins of BNIP3 were precipitated with a BNIP3 antibody and analyzed by Mass spectrometry fingerprinting. The amino acid sequence of VDAC with the tryptic peptide identified from the immunoprecipitates is labelled in bold and underlined. E Western bloting analyses show interaction between BNIP3 and VDAC. a) Levels of GST-BNIP3 in the mitochondrial fraction after mitochondria-BNIP3 incubation as determined by Western blotting. b) After co-immunoprecipitation with a BNIP3 antibody, immunoblotting with a VDAC antibody reveals the presence of VDAC in the BNIP3 immunoprecipitates. c) BNIP3 was present in the precipitates when a VDAC antibody was used for co-immunoprecipitation. Shown are samples from three independent experiments. Cell lysates without BNIP3 treatment were used as controls.

Figure 5. Blocking VDAC abolishes BNIP3-induced mitochondrial release of EndoG.

A, Mitochondria (1 mg/ml) were incubated with a recombinant GST-BNIP3 in the presence of 0.3 mg/ml of an anti-VDAC antibody or normal rabbit IgG in control. Incubation with the trunked BNIP3 (BNIP3ΔTM) was used as a control. The samples were separated into mitochondrial and supernatant fractions by centrifugation. a), Blocking VDAC prevented BNIP3 from integration into mitochondria. b) Analysis of EndoG from the supernatant showed that levels of EndoG released from mitochondria was reduced by the VDAC antibody blocking. c) Co-immunoprecipitation with a BNIP3 antibody to pull down BNIP3-interacting proteins showed that blocking VDAC abolished the integration of BNIP3 with mitochondria. B, Quantification of BNIP3 integration with mitochondria and mitochondrial release of EndoG affected by the VDAC antibody blocking. Results shown represent the mean ±SD for combined data from three independent experiments. **, p<0.01.