Fig. 5: Changes in domain communications upon active mutations in MEK1.

a Domain partition of MEK1 protein, including the positions of mutations (A52V, S218Sp/S222Sp, and E203K). b Different views of the MEK1 structure. The N-terminal lobe (N-lobe) contains one core kinase (gray) and two conserved α-helices (blue). The C-terminal lobe (C-lobe) contains three core kinase domains (gray and black), an activation segment (orange), and a proline-rich loop (green). c Distribution of learned edges between residues in the MD simulations of WT, A52V, S218Sp/S222Sp, and E203K MEK1. d Distribution of learned edges between domains in the MD simulations of WT, A52V, S218Sp/S222Sp, and E203K MEK1. In e, the interaction graph is mapped from the learned edges of active mutant MEK1. The size of a node represents the number of learned edges that directly connect to the node. The thickness of an edge represents the strength of the interaction. The arrows denote the directionality of a learned edge. In f, the allosteric pathways start from N221 in the activation segment and lead to the αA-helix and the proline-rich loop in the S218Sp/S222Sp MEK1 (left). On the right, the allosteric pathways start from R201 (near E203K) and lead to the αC-helix and the proline-rich loop in the E203K MEK1. The size of a node represents the number of learned edges that directly connect to the node. The thickness of an edge represents the strength of the interaction. We used N221 and R201 (near E203K) as the starting points, and the residues in the αA/αC helices and the proline-rich loop as the ending points to present the pathways.