Phosphoinositides protonation states dictate AKT1 binding with membrane

Abstract

Phosphoinositides are key regulators of membrane-associated signaling, yet their electrostatic properties are commonly represented by a single nominal charge in molecular simulations and structural models. Using the essential protein kinase AKT1 as a model system, we show that phosphoinositide protonation microstates, both the degree and positional placement of protons, govern peripheral membrane-protein recognition. By integrating quantum-mechanics-derived headgroup parameters, microsecond-scale all-atom molecular dynamics, and umbrella-sampling free-energy calculations with the available solid-state NMR-derived lipid populations via Bayesian/Maximum Entropy reweighting, we account explicitly for the thermodynamic ensemble of phosphoinositide protonation states. This population-weighted free-energy framework provides a quantitative mechanistic rationale for the strict specificity of wild-type AKT1 toward PI(3, 4, 5)P3 and elucidates how the oncogenic sentry mutation (E17K) alters this specificity to enable high-affinity PIP2 binding. Furthermore, Proximity-based Labeling of Membrane-Associated Proteins (PLiMAP) assays experimentally confirm that AKT1 pleckstrin homology domain binding to distinct phosphoinositide species is differentially sensitive to pH. Together, our findings establish a direct connection between phosphoinositide protonation equilibria, binding energetics, and AKT1 lipid specificity. Beyond AKT1, these results demonstrate that the dynamic protonation microstates of anionic phospholipids act as a functional regulatory layer governing peripheral membrane-protein recruitment across diverse cellular microenvironments. Significance Statement In physiological conditions, AKT1 protein is highly specific to PI(3,4,5)P3 lipids and completely excludes PI(4,5)P2 lipids. However, a single charge reversal (oncogenic) mutation of sentry glutamate (E17K) on the membrane adaptor Pleckstrin Homology domain of AKT1 enhances its PIP2 binding remarkably. Our simulation data reveal that AKT1-PHD dissociation free-energies have remarkable variations with changes in lipid protonation state. By integrating our atomic simulation readouts with available ensemble-averaged low-resolution information from experiments using machine learning methods, we could explain the highly enhanced PIP2 specificity for the pathological mutant. The work highlights the degenerate roles of secondary messenger lipids as a function of membrane environments.