Fig 1: Signalling profile of B-017.a Immunoblot analysis showing the effect of B-017 on the signalling of mitochondrial-induced cell death, characterised by BAX insertion into the mitochondrial outer membrane. Wild-type mouse embryonic fibroblasts (WT MEFs, 3 × 106 cells) were exposed to staurosporine (STS) (2 µM) and B-017 (146 µM), vehicle (Veh.), or the negative control peptide (Ctrl., 146 µM). To assess BAX insertion, mitochondria were isolated and subjected to alkaline extraction to remove loosely attached BAX. (n = 4 independent experiments, repeated-measures ANOVA, two-sided, data are presented as mean ± SD). b Immunoblot analyses showing the impact of B-017 on the 1st activation step of BAX using the 6A7 antibody to detect early conformational changes in BAX. STS, BNIP3-Peptide1–49 (Peptide1–49), PUMA and BIM were used as stimulation agents; the negative control peptide (NCP) served as a control. (n = 3 independent experiments, Supplementary Fig. 14b). c Immunoblot analyses showing impact of B-017 on the 2nd activation step, characterised by BAX insertion into the mitochondrial outer membrane (MOM). To assess the percentage of BAX that is inserted into the MOM, mitochondria were treated with strong alkali, which extracts only loosely attached BAX, not that which is inserted. STS, Peptide1–49, PUMA, BIM, and BIM-SAHB variants were used as stimulating agents; NCP served as a control. (Representative immunoblots of n = 4 and n = 5 independent experiments). d Response to B-017 in terms of late-stage cell death signalling, characterised by caspase 3/7-activity, in STS-treated WT MEFs (left) and MEFs lacking BAX, BAK, BAX/BAK, or the anti-apoptotic protein BCL-2 (right). (WT MEFs n = 3 independent experiments, Knockout MEFs n = 3 experiments with four technical replicates, two-way ANOVA, data are presented as mean ± SD). e Cellular response to B-017 treatment in different knockout MEFs exposed to cell death signalling stimulation using live cell impedance measurements. Bax−/−, Bak−/−, and Bcl-2−/− MEFs were treated simultaneously with 100 nM STS and B-017 (52 µM) or vehicle after reaching a cell index of 1, indicating a stable adhesion to the microtiter plate. (n = 3 independent experiments, two-way ANOVA, data are presented as mean ± SD). Source data are provided as a Source Data file.
Fig 2: Identification of B-017, a BNIP3 antagonist peptide, by reverse engineering.a HDOCK docking of human (h)BAX (PDB code 2K7W) and BNIP3-peptide1–49 (hit 3 of top 10 optimal conformations; core binding residues are highlighted in magenta). b Immunoblot analyses showing the impact of BNIP3-peptide1-49 on the 1st (left) and 2nd (right) activation step of BAX. BIM and PUMA were used as BAX activators. (n = 5 independent experiments, representative immunoblots). c Identification of crucial N-terminal amino acid (aa) residues required for binding using a protein-peptide microarray. Fluorescence-labelled recombinant mouse (m)BNIP3 was incubated with a library of 65 synthesised peptides representing N-terminal truncations of the BNIP3 sequence MSQSGEENLQGSWVELHFSN. (n = 3 identical subarrays with mouse IgG controls). d Substitution analysis using fluorescence-labelled recombinant mBNIP3 and a library of 379 synthesised peptides. Single residues of the BNIP3 sequence MSQSGEENLQGSWVELHFSN were exchanged for 20 aa. (n = 3 identical subarrays with mouse IgG controls). e Substitution analysis with fluorescence-labelled recombinant mBAX and a library of 156 synthesised peptides. Single residues of the BNIP3 sequence WVELHFSN were exchanged for 20 aa. f Fluorescence anisotropy assay with titration of increasing recombinant hBNIP3 concentrations to 100 nM Cy5.5-B-017 (left). (Data points represent the mean ± SD, n = 3 biological replicates). Microscale thermophoresis dose-response curve of BAX binding to Cy5.5-labelled peptide. A concentration of 200 nM fluorescence-labelled peptide was incubated with serial dilutions of BAX (12.2–25,000 nM). Normalised fluorescence changes were plotted as a function of protein concentration. Data points represent the mean ± SD (n = 3 biological replicates). Solid red lines show the fitted binding curves (right). g Structure prediction of B-017 interaction with hBAX, using AlphaFold2. h Interactions between B-017 and recombinant hBAX, mapped by photo-cross-linking BAX:B-017-BpA complex with BpA, and identification of contacts by mass spectrometry (left). Cross-links in BAX were mapped to the BAX model (PDB code: 4S0O) (right). i Interactions between B-017 and BAX/BNIP3 mapped by photo-cross-linking BNIP3/BAX:B-017-BpA complex and identification of contacts by mass spectrometry (left). Cross-links in BAX and BNIP3 mapped to the docking model of the homology model of BNIP3 and BAX (PDB code: 4S0O) (right). Source data are provided as a Source Data file.
Fig 3: Cellular, computational and in vitro assessment of BNIP3 binding to BAX.a Close proximity of BNIP3 to BAX or BAK1 in mouse embryonic fibroblasts (WT MEFs) (top) and in Bak knockout (Bak−/−) or Bax knockout (Bax −/−) MEFs (bottom). Interactions of BNIP3 and BAX or BAK1 are indicated by red dots with white arrows. Antibodies against BAX (2D2), BAK1 (4C2) and BNIP3 (1C8) were used along with a pair of secondary antibodies conjugated to complementary oligonucleotides in an in situ proximity ligation assay (PLA). (Nuclei stained with DAPI, blue; scale bars 30 µm and 10 µm, n = 3 independent experiments). b Close proximity of BNIP3 and BAX in isolated male mouse Bnip3+/+ cardiomyocytes (mCM) (left) and in CM of male Bnip3-3xFlag-in (Flag-Bnip3ki/ki) mice (right). (n = 3 independent experiments). c Fluorescence-labelled recombinant mouse (m)BNIP3 incubated with a library of 13 synthesised BAX peptides immobilised on microarrays to reveal potential interaction sites in BAX. The colour coding ranges from white (low or no intensity), through yellow (medium intensity), to red (high intensity) (top). Structure of mBAX (PDB 4S0O) in ribbon representation with BNIP3 interaction sites coloured blue (α5), green (α6), yellow (α7) and orange (α8) (bottom). (n = 3 identical subarrays with mouse IgG controls). d Modeller algorithm was used to predict the BNIP3 homology model based on BAX (PDB code 2K7W) as a template, followed by energy minimisation using NAMD 2.9 and CHARMM36 force field. The resulting model agreed well with the circular dichroism (top) and intramolecular cross-linking data obtained with recombinant human BNIP3 (hBNIP3). The DSSO cross-links identified in recombinant hBNIP3 by mass spectrometry were mapped onto the homology model (orange lines), supporting the predicted compact three-dimensional structure (bottom). e HDOCK docking of the hBNIP3 homology model and hBAX (PDB code 2k7w, model 1), with a docking score of −253.86 and a confidence score of 0.8887 (hit 1 of the top 10 optimal conformations). The predicted intermolecular hydrophobic interactions were mapped onto the docking model (magenta lines) (left). Furthermore, two hydrogen bonds were predicted (3D detail map) (right). Source data are provided as a Source Data file.
Supplier Page from Creative BioMart for Recombinant Human BNIP3 Protein, GST-tagged, a.a.1-194