Fig 1.
(A) The BCL-2-regulated apoptotic pathway involves the following proteins pro-survival BCL-2 proteins, pro-apoptotic BH3-only proteins, and apoptosis effectors BAX/BAK proteins. Lines indicate interactions between molecules. (B) In healthy cells, the pro-survival BCL-2 proteins sequester BAX and BAK proteins so that these effectors cannot initiate apoptosis. Normally, apoptosis occurs indirectly due to the pro-survival BCL-2 proteins being bound by BH3-only proteins, which prevents the pro-survival BCL-2 proteins from binding and inhibiting BAX/BAK. Subsequently, the free BAX/BAK can initiate apoptosis. (C) In cancer cells, the loss of BH3-only proteins or the increase in pro-survival BCL-2 proteins causes BAX/BAK to be directly bound by pro-survival BCL-2 proteins, preventing the initiation of apoptosis.
Fig 2.
Visual representations of the Knob-Socket model.
(A) 3-dimensional model of a socket shown on an α-helix. (B) 2-dimensional representation of a socket formed by the residues X, Y, and H. The 3 residues clearly form a surface that can pack another residue. (C) α-helical lattice showing the sockets formed on an α-helix surface, with numbers representing socket residues. (D) 3-dimensional model of a knob on one helix (upper) packing into a socket on another helix (lower). (E) Geometric model showing the tetrahedral arrangement of the knob-socket complex. (F) 2-dimensional representation of the knob-socket motif with three socket residues X, Y, and H packed against a knob residue B from another helix. (G) 2-dimensional lattice representation of a socket (formed by residues 2, 5, and 6) filled by knob residue B, and a pocket (formed by residues 4, 7, 8, and 11) filled by knob residue B. When a knob residue B packs into 2 contiguous sockets, the overall surface formed by those residues is denoted a pocket.
Table 1.
List of interactions between various pro-survival BCL-2 proteins and the different BH3 ligands examined in this study.
Fig 3.
Model of Mcl-1 general binding.
Only H3, H4, & H5 are shown. The orientation of the α-helices are rotated to match the binding groove and provide a relative idea of the packing cleft’s surface. (A) The knob-socket diagram of Mcl-1, with only the essential parts of the helices included. The orange knobs represent the residues that are conserved across different pro-survival BCL-3 proteins and could be considered important for binding affinity. The blue knobs represent the residues that specific to the pro-survival BCL-3 protein homolog, and most likely contribute to binding specificity. (B) 3D model of Mcl-1 generated from UCSF Chimera (16). The important sockets and knobs have been overlaid. (C) The amino acid sequence of Mcl-1, organized based on the helices H3, H4, and H5.
Fig 4.
Full map showing the interaction between Mcl-1 and Bim. (A) The 2D knob-socket representation of the protein, with the knobs color-coded to identify the helix of origin. (B) A 3D model of the interaction generated by UCSF Chimera (16). (C) The amino acid sequences of all of the helices with corresponding color coding. (D) Side by side, 3D coil representation and 2D helix lattice models of the ligand. In the 3D model, all of the residues that serve as knobs have been shown and labeled.
Table 2.
Minimum and maximum distances both within and between helices provided.
Coordinates for knob residues were calculated by averaging coordinates of Cβ atoms of participating socket residues. Cα atom coordinates were used if residue was a glycine.
Fig 5.
(A-D) Selection of representative ligand helices from pro-survival protein-BH3 domain complexes. (A) BIM helix in complex with Mcl-1 (PDB: 2pqk). (B) BAK in complex with A1 (2voh). (C) BIM L12F in complex with BCL-XL (3io8). (D) MB7 in complex with Mcl-1 (3kz0). (E) Designed helix with specificity for Mcl-1, showing predicted interactions with protein residues. Conserved residues are shown in red and variable residues that increase specificity for Mcl-1 shown in blue. Central glycine is shown in green. Residues outlined in purple represent knobs that project into the sockets of Mcl-1. (F) 3-dimensional representation of peptide with important interacting residues shown (based on BIM).
Fig 6.
A) Consensus models of BAX, BAK, and BH3-only proteins. The essential residues of the BH3 domain for each type of protein are represented on a 2-dimensional lattice of the α-helix. The central glycine is indicated in blue. Blank residues represent non-essential residues that do not participate in the binding interaction between the BH3 peptides and the BCL-2 protein paralog. Residues denoted by ‘X’ represent residues where the consensus residue could not be determined (due to lack of consensus in the chosen set of BH3 peptides). B) Sequence alignments for the unique BH4 ligands of Mcl-1, BFL-1/A1, and BCL-xL are shown. Sequences of point mutants were omitted for clarity. The conserved Gly, Asp, and Leu are highlighted. C) From these sequence alignments as was done in part A, consensus models for BH3 peptides that bind Mcl-1, BFL-1/A1, and BCL-xL. Each 2-dimensional lattice on the bottom row represents the consensus sequence of the BH3 peptides that bind to the respective BCL-2 protein paralog listed in Table 1. The residues shown are those that appeared the most frequently for that BH3 ligand when bound to various pro-survival BCL-2 proteins. Note that the consensus sequence of the BH3 domain ligands is not necessarily the most optimal sequence for binding the BCL-2 protein. These models are only representative of the selected BH3 peptides (Table 1). The number and types of BH3 helices used in the consensus sequence varies for each BCL-2 protein paralog.