PMC

Model for ISRIB’s mechanism of action. ISRIB staples together tetrameric eIF2B(βγδε) subcomplexes, building a more active eIF2B(βγδε)₂ octamer. In turn, the ISRIB-stabilized octamer binds eIF2B(α₂) with greater affinity, enhancing the formation of a fully-active, decameric holoenzyme.

eIF2B structure predicts activity of ISRIB analogs. (A) GEF activity of assembled eIF2B(βγδε) and eIF2B(α₂) in the presence and absence of ISRIB-A19(R,R) and ISRIB-A19(S,S). (B) Stability of decameric eIF2B(δL179A) in the absence of ISRIB (top), presence of ISRIB-A19(S,S) (middle), or presence of ISRIB-A19(R,R) (bottom) as assessed by velocity sedimentation on sucrose gradients. (C) eIF2B GEF activity of assembled eIF2B(βγδε) and eIF2B(α₂) containing a δL179A mutation in the presence and absence of ISRIB-A19(R,R) and ISRIB-A19(S,S). (D) Quantification of eIF2B decamer stability gradients plotted as fraction of eIF2B(βγδε) present in each of lanes 1-13. eIF2B (for comparison from data shown in ), eIF2B(βH188A), eIF2B(βH188Y), eIF2B(βH188F) gradients are plotted in the presence (bottom panel) and absence (top panel) of 500 nM ISRIB. (E, F, G) Stability of decameric eIF2B(βH188A), eIF2B(βH188Y), and eIF2B(βH188F) in the presence of ISRIB as assessed by velocity sedimentation on sucrose gradients.

ISRIB induces dimerization of tetrameric eIF2B subcomplexes. The most abundant 2D class averages from cryoEM imaging of eIF2B(βγδε) in the presence (A) and absence (B) of ISRIB. (C) Characterization of eIF2B(βγδε) by sedimentation velocity analytical ultracentrifugation. eIF2B(βγδε) (1 μM) was analyzed in the presence and absence of 1 μM ISRIB. (D) Mixture of 1 μM eIF2B(βγδε) and 500 nM eIF2B(α₂) characterized by analytical ultracentrifugation in the presence and absence of 1 μM ISRIB. (E) eIF2B(βγδε) (1 μM) characterized by analytical ultracentrifugation in the presence of 1 μM or 10 μM ISRIB. (F) GEF activity of eIF2B(βγδε), here at a higher 100nM concentration to facilitate comparison of 0, 0.2, and 5 μM ISRIB.

Loss- and gain-of-function dimerization mutants resist or bypass the effects of ISRIB. (A) Surface rendering of core eIF2Bβ (blue) and eIF2Bδ (gold) subunits with residues contacting ISRIB highlighted in gray and with dimer interface indicated by dashed line. Interface residues are highlighted in a lighter hue of the colors of the contacting subunits. (B) Open-book view of the dimer-dimer interface, such that each β and δ subunit is rotated by 90˚. βH160, in green, contacts both β’ and δ’; δL179, also in green, contacts both β’ and ISRIB. (C) Characterization of 1 μM eIF2B(βγδε) containing a βH160D mutation in the presence (right) and absence (left) of 1 μM ISRIB by analytical ultracentrifugation. (D) GEF activity of eIF2B(βγδε) containing a βH160D mutation in the presence and absence of ISRIB. (E) Characterization of 1 μM eIF2B(βγδε) containing a δL179V mutation in the presence (right) and absence (left) 1 μM ISRIB by analytical ultracentrifugation. (F) GEF activity of eIF2B(βγδε) containing a δL179V mutation in the presence and absence of ISRIB.

Near-atomic resolution reconstruction of ISRIB-bound eIF2B. (A-C) Three views of cryoEM density for eIF2B(αβγδε)₂, colored in distinct shades for each subunit copy: red for α, blue for β, green for γ, gold for δ, and gray for ε (color code used throughout this manuscript). Density assigned to ISRIB depicted in CPK coloring: oxygens highlighted in red, nitrogens in blue and chlorines in green. The rotational relationships between the views depicted in A, B, and C are indicated. (D) Cross-section of (A), revealing view of the ISRIB binding pocket at the central decamer symmetry interface and density assigned to ISRIB CPK-colored by element. (E) Close-up view of density assigned to ISRIB and its binding pocket in (B) at the intersection of two β and two δ subunits. (F) Two conformers of ISRIB modeled into the density and all residues within a 3.7Å distance from the ligand rendered as sticks.

ISRIB stabilizes decameric eIF2B, accelerating GEF activity. (A) Schematic diagram for three plasmid expression of all five eIF2B genes in E. coli. (B) Characterization of eIF2B(αβγδε)₂ by sedimentation velocity analytical ultracentrifugation and SDS-PAGE followed by Coomassie blue staining. (C) Initial rate of nucleotide exchange (right panel) plotted as a function of substrate concentration. Note that at high eIF2 concentration we reproducibly observed a transient increase in fluorescence that peaked at the 1 min time point (left panel). Such increase was reported previously () and remains unexplained. (D) GEF activity of eIF2B(αβγδε)₂ as measured by unloading of fluorescent GDP from eIF2 in the presence and absence of ISRIB. (E) Absorbance 280 nm trace from an anion exchange column used in the purification of eIF2B in the presence (red) and absence (black) of ISRIB. Peak fractions from the (−) ISRIB purification were analyzed by SDS-PAGE and Coomassie-stained. eIF2B subunits are labeled (α-ε) and an asterisk denotes the presence of a contaminating protein that contributes to peak 1. (F) Stability of eIF2B(αβγδε)₂ was assessed by sedimentation velocity on a 5-20% sucrose gradient in a 400 mM salt buffer. eIF2B(βγδε) and eIF2B(α₂) were combined with and without 500 nM ISRIB. Fractions were analyzed by SDS-PAGE and Coomassie-stained. (G) GEF activity of eIF2B assembled from purified eIF2B(βγδε) and eIF2B(α₂) in the presence and absence of ISRIB. (H) GEF activity of eIF2B(βγδε) in the presence and absence of eIF2B(α₂). (I) GEF activity of eIF2B(βγδε) in the presence and absence of ISRIB.