Proper execution of cell death ensures normal biological processes, and its dysregulation causes human illness, ranging from cancer to neurodegenerative disorders. Apoptosis is a regulated form of cell death. Impairment of apoptosis is not only central to cancer development but also renders tumors refractory to cytotoxic therapy. Hence, further elucidation of the apoptotic signaling framework will not only help understand how cancer cells escape apoptotic checkpoints but also contribute to the development of rationally designed targeted cancer therapy. The BCL-2 family proteins govern cell death-versus-survival decisions at the mitochondria and can be divided into three subfamilies: (1) multidomain antiapoptotic BCL-2, BCL-XL and MCL-1; (2) multidomain proapoptotic BAX and BAK; and (3) proapoptotic BH3-only molecules (BH3s). BH3s relay upstream apoptotic signals to promote apoptosis by either activating BAX/BAK or inactivating BCL-2/BCL-XL/MCL-1. Genetic loss-of-function studies reveal an essential axis of upstream "activator" BH3s and downstream BAX/BAK in activating mitochondrion-dependent apoptosis. In response to apoptotic signals, the "activator" BH3s, including BID, BIM and PUMA, trigger the homo-oligomerization of BAX and BAK to permeabilize mitochondria, leading to the efflux of cytochrome c to the cytosol for caspase activation. We recently discovered a novel mechanism by which BH3s activate BAX/BAK- dependent mitochondrial permeabilization. We propose to elucidate this novel mechanism and reclassify BH3s based on their mechanisms in initiating cell death. Activation of BAX/BAK by BH3s not only triggers APAF-1-mediated caspase activation but also initiates caspase- independent cell death. Further characterization of this novel form of cell death holds promises for the future development of anti-cancer therapeutics that targets cancer cells with defective caspase activation. We are also pursing whether BAX and BAK reside in different protein complexes to differentially regulate caspase-dependent and -independent cell death programs. Overall, our goal is to build a comprehensive proapoptotic BCl-2 signaling network in activating mitochondrion-dependent cell death programs, which offers common ground for therapeutic interventions.