Discovery of diverse human BH3-only and non-native peptide binders of pro-apoptotic BAK indicate that activators and inhibitors use a similar binding mode and are not distinguished by binding affinity or kinetics

ABSTRACT Apoptosis is a programmed form of cell death important for the development and maintenance of tissue homeostasis. The BCL-2 protein family controls key steps in apoptosis, dysregulation of which can lead to a wide range of human diseases. BCL-2 proteins comprise three groups: anti-apoptotic proteins, pro-apoptotic proteins, and BH3-only proteins. BAK is one of two pro-apoptotic proteins, and previous work has shown that binding of certain BH3-only proteins such as truncated BID (tBID), BIM, or PUMA to BAK leads to mitochondrial outer membrane permeabilization, the release of cytochrome c, and ultimately cell death. This process, referred to as activation , involves the BH3-stimulated conversion of BAK from monomer to dimer and then to oligomers that promote membrane disruption. Crystal structures of putative intermediates in this pathway, crosslinking data, and in vitro functional tests have provided insights into the activation event, yet the sequence-function relationships that make some but not all BH3-only proteins function as activators remain largely unexamined. In this work, we used computational protein design, yeast surface-display screening of candidate BH3-like peptides, and structure-based energy scoring to identify ten new binders of BAK that span a large sequence space. Among the new binders are two peptides from human proteins BNIP5 and PXT1 that promote BAK activation in liposome assays and induce cytochrome-c release from mitochondria, expanding current views of how BAK-mediated cell death may be triggered in cells. High-resolution crystal structures and binding experiments revealed a high degree of similarity in binding geometry, affinity, and association kinetics between peptide activators and inhibitors, including peptides described previously and those identified in this work. We propose a model for BAK activation that is based on differential engagement of BAK monomers vs. the BAK activation transition state that integrates our observations with previous reports of BAK binders, activators, and inhibitors.