Structural representations of the wild-type Abl1 kinase core in either the DFG-in (left, PDB 6XR6) or DFG-out (right, PDB 6XR7) states as elucidated through nuclear magnetic resonance experiments.

For ease of visualization, key residues are represented as sticks (Asp381, red; Phe382 green, both indicated by broad arrows of matching colors; Lys271, dark blue; Glu286, dark red). Additional elements important for kinase function and involved in the DFG flip are colored in pink (P-Loop, residues 249-255), orange (B-Loop, residues 275-278), red (αC-Helix, residues 280-293), green (hydrophobic C-spine), and cyan (activation loop, residues 385-400).

Progress coordinates and auxiliary datasets tracked in the WE simulations of the Abl1 kinase core DFG flip.

(A): Structural representation of the observables tracked by each progress coordinate in three contrasting DFG conformational states (DFG-in, DFG-inter, and DFG-out). References that introduced each chosen PC are listed in the bottom right section of the panel. (B): Evolution of auxiliary datasets 1 and 2 (used to track the torsions associated with Asp381/Phe382 flipping) and progress coordinates 1-3 (used to discretize the conformational space populated by the WE simulations and direct walkers towards the DFG-out state). Data were taken from two representative trajectories of the DFG flip in the wild-type Abl1 kinase core simulated with the WE methodology.

Summary of the conformational space explored in the wild-type Abl1 kinase core WE simulations and pathways identified for the DFG flip.

(A): Bi-dimensional projection of the Asp381 and Phe382 torsion angles (tracked by the dihedrals defined in Figure 2) for the entire dataset resulting from the WE simulations. Each dot represents a walker (individual simulation). (B): Summary of the pathways identified in WE simulations of the wild-type Abl1 kinase core DFG flip, with accompanying probabilities for each identified state. (C): Visual representation of average DFG conformations deviating from the common DFG-inter conformation identified for each pathway.

Detailed analysis of a successful DFG flip event in wild-type Abl1 kinase simulated with the weighted ensemble method.

(A) Representative frames from the selected concerted DFG flip trajectory arranged in a timeline showcasing important interactions and rearrangements for the flip. Frames are ordered from left to right, top to bottom. (B) Evolution of selected observables shown in (A) during the DFG flip trajectory.

Correlation between the Lys271/Glu286 salt bridge and DFG torsions in WE simulations of the wild-type Abl1 kinase core DFG flip.

(A): Visual representation of the distances between the terminal carbon of Glu286 and the charged nitrogen of Lys271 in two distinct DFG conformations (starting conformation on the left; and a DFG flip intermediate conformation on the right). (B): Bi-dimensional projection of the correlation between Lys271/Glu286 distances and Asp381 (left) or Phe382 (right) torsions for the entire WE simulation dataset. Each dot corresponds to an individual simulation (walker), and dots are colored based on the inverse of their relative probabilities. Starting, intermediate, and target states are labeled in the plots.

Results of weighted ensemble simulations of the DFG flip in the Abl1 kinase core and Abl1’s drug-resistant variants.

(A) Structural representation of the Abl1 kinase core and the positions of wild-type residues Glu255 and Thr315 and their variants, Val/Lys255 and Ile315. For ease of visualization, key residues are represented as sticks (Asp381, red; Phe382, green; Glu255, orange; Thr315, cyan; Val255, yellow; Lys255, blue; Ile315, magenta). Additional elements important for kinase function and involved in the DFG flip are colored in pink (P-Loop, residues 249-255), orange (B-Loop, residues 275-278), red (αC-Helix, residues 280-293), and cyan (activation loop, residues 385-400). (B) Tri-dimensional projection of the values of the torsion angles used to define and visualize the state of the DFG motif in Abl1 kinase in the weighted ensemble simulations (X and Y) and of Progress Coordinate 3 (PC3, Z). Color bars define the probability associated with each walker, and are a surrogate for the free energy barrier between states.

Contributions of different electrostatic interactions to DFG flip events in Abl1 kinase and its drug-resistant variants.

(A) Representative frames showcasing relevant interactions and residue positions for complete (top), half (middle), and frustrated (bottom) DFG flip events taken from the weighted ensemble simulations of each Abl1 variant. Distances between reference atoms are depicted as yellow dotted lines, which are changed to magenta dotted lines if residues form a salt bridge interaction. Distances are given in angstroms. (B) Evolution of selected observables showcased in (A) from representative trajectories for each variant.

Comparison between intermediate conformations adopted by the Abl1 kinase core and a drug-resistant variant during a DFG flip event.

(A) Structural representations of the (left) wild-type and (right) Glu255Lys Thr315Ile Abl1 kinase cores. Residues relevant for inducing or stabilizing either conformation are shown as sticks, accompanied by distance measurements (in Angstroms, shown as yellow dotted lines) to highlight the differences between conformations. (B) Evolution of distances and angles shown in A during a successful DFG flip trajectory (for wild-type Abl1) or a frustrated DFG flip trajectory (for the Glu255Lys Thr315Ile variant). (C) Kernel density estimation of distances shown in B. Trajectories were chosen as representatives from the weighted ensemble simulation results.

Comparison between intermediate conformations adopted by the Abl1 kinase core and by a drug-resistant variant during a DFG flip event.

(A) Structural representation of the (left) wild-type and (right) Glu255Val Thr315Ile Abl1 kinase cores. Residues relevant for inducing or stabilizing either conformation are shown as sticks, accompanied by distance measurements (in Angstroms, shown as yellow dotted lines) to highlight the differences between either conformation. (B) Evolution of distances and angles shown in A during a successful DFG flip trajectory (for wild-type Abl1) and for a half DFG flip trajectory (for the Glu255Val Thr315Ile variant). (C) Kernel density estimation of distances shown in (B). Trajectories were chosen as representatives from the weighted ensemble simulation results.

Conservation analysis of the charge composition of residues in the αC-Helix or P-Loop of the Abl1, Ancestral AS, and Src kinases.

(A) Three-dimensional structures of evolutionarily-linked kinase cores annotated with spheres marking residues within or in the close vicinity of either the P-Loop (residues 249-255 in Abl1) or αC-Helix (residues 249-255) whose interactions are thought to be important for DFG flip events. (B) Alignment of the first 100 residues of the three evolutionarily-linked kinase domains with significantly different activation free energy barriers. (C) Presumed effect of different charge compositions on the backbone dynamics of the Abl1, Src, and Abl1 Glu255Lys Thr315Ile kinase cores. Colored arrows represent repulsive or attractive forces between structural elements, and thin arrows illustrate movements potentially caused by the observed charge compositions. For ease of visualization, all residue positions were annotated for the DFG-in state of wild-type Abl1 (PDB 6XR6).