Dual Allosteric and Orthosteric Inhibition of 14‐3‐3ζ–α‐Synuclein Interaction: A Multiscale Simulation and Machine Learning Approach

Parkinson's disease (PD) is a progressive neurodegenerative disorder marked by the pathological aggregation of α‐synuclein (α‐syn) and the degeneration of dopaminergic neurons. The interaction between α‐syn and 14‐3‐3ζ has been implicated in modulating α‐syn's stability, localization, and aggregation behavior, rendering it a promising target for therapeutic intervention. In this study, we employed a comprehensive computational pipeline to identify small‐molecule inhibitors capable of disrupting the 14‐3‐3ζ/α‐syn interaction. Structure‐ and ligand‐based virtual screening, followed by toxicity filtering and molecular dynamics simulations, led to the identification of Var84 (orthosteric, ORT) and DB11581 (allosteric, ALO) as candidate inhibitors. A dual‐site inhibition (DUO) approach involving simultaneous binding of both ligands is also investigated. Absolute binding free energy (ABFE) and residence time (RAMD) analyses revealed cooperative binding effects: ALO maintained strong binding across systems, while ORT binding weakened in the DUO system, likely due to increased inter‐protein separation. UMAP clustering and secondary structure analysis indicated that the DUO system preserved helical α‐syn conformations while reducing aggregation‐prone β‐structures. Additionally, supervised machine learning models trained on inter‐protein contact features identified key residue pairs perturbed by ligand binding, corroborating findings from communication network analyses, thereby offering mechanistic insight and a transferable framework for targeting PPIs in neurodegenerative diseases.