Why an A-Loop Phospho-Mimetic Fails to Activate PAK1: Understanding an Inaccessible Kinase State by Molecular Dynamics Simulations

Abstract

Crystal structures of inactive PAK1(K299R) and the activation (A)-loop phospho-mimetic PAK1(T423E) have suggested that the kinase domain is in an active state regardless of activation loop status. Contrary to a large body of literature, we find that neither is PAK1(T423E) active in cells, nor does it exhibit significant activity in vitro. To explain these discrepancies all-atom molecular dynamics (MD) simulations of PAK1(phospho-T423) in complex with ATP and substrate were performed. These simulations point to a key interaction between PAK1 Lys308, at the end of the alphaC helix, and the pThr423 phosphate group, not seen in X-ray structures. The orthologous PAK4 Arg359 fulfills the same role in immobilizing the alphaC helix. These in silico predictions were validated by experimental mutagenesis of PAK1 and PAK4. The simulations explain why the PAK1 A-loop phospho-mimetic is inactive, but also point to a key functional interaction likely found in other protein kinases.

The Cdc42 Effector Kinase PAK4 Localizes to Cell-Cell Junctions and Contributes to Establishing Cell Polarity

“…This is consistent with the notion that PAK4 lacking residues 10–30 in the Cdc42/Rac interactive binding (CRIB) domain is active [16]. Although structural and biochemical analysis suggests that PAK1 activation occurs through activation loop Thr-423 phosphorylation [17], it is notable that PAK4 is constitutively phosphorlyated on Ser-474 [14], and kept in check through the AID. The binding of Cdc42 can serve to activate PAK4 in cells but it is unclear if there is any auto-phosphorylation event associated with this activation [14].…”

PAK family kinases

“…Mutation of certain residues in the AID lead to constitutive PAK1 activation in the absence of Cdc42 36 . However, the phospho-mimetic PAK1(T423E), which has been used by a number of groups, fails to exhibit auto-phosphorylation characteristic of active kinase when expressed in mammalian cells 37 . This is consistent with the observation that PAK1 (T423E) catalytic domain can be tolerated by E. coli , which are growth-sensitive to active forms of PAK1.…”

Chemical states of the N-terminal “lid” of MDM2 regulate p53 binding: Simulations reveal complexities of modulation

“…Our study was inspired by earlier successes in reproducing and successfully predicting various features of the p53-MDM2 interaction, 4,19 and recent demonstrations that a phosphomimetic may display dynamical features that can be quite different from those displayed by a phosphate group. 16 Simulations show that in the WT, the lid indeed has a weak affinity for the p53-binding cleft on the surface of MDM2, as postulated from the experiments. However the lid can also dynamically span several conformational sub-states that can interconvert between open, semi open and closed states (however on the timescales of the simulations we only see the closed to open transition), in agreement with NMR observations.…”

“…Indeed, recent simulation studies combined with experiments have shown that these two could give rise to significant differences. 15,16 In order to probe these differences in the MDM2-p53 system, we carried out detailed molecular dynamics simulation studies and provide hypotheses than can lead to novel experimental approaches to study MDM2 regulation. We have simulated the wild type (WT), S17D and the pS17 states of p53 binding domain of MDM2 (residues 1-119) with the lid in 'open' and 'closed' conformations (see Fig.…”