Abstract
Alzheimer's disease (AD) is a multifactorial neurodegenerative disorder characterized by cholinergic dysfunction, amyloid-β aggregation, mitochondrial stress, and aberrant kinase activity. Carotenoids, naturally occurring pigments with antioxidant and neuroprotective properties, have emerged as promising candidates for AD intervention. In this study, we performed a systematic stepwise computational screening of a large carotenoid library (n = 1191) to identify multitarget candidates against AD-related proteins. The workflow consisted of predefined ADMET filtering (oral absorption > 90%, Caco-2 > 0.9, logBB > -1, and absence of major CYP inhibition and toxicity alerts), reducing the dataset to 61 compounds, followed by multi-target molecular docking against AChE, BChE, BACE-1, MAO-B, and GSK3-β. Compounds were ranked using an aggregated mean docking score across all five targets, and the top-performing candidate was subjected to detailed mechanistic analyses. Hopkinsiaxanthin emerged as the highest-ranked multitarget carotenoid and was further evaluated using frontier molecular orbital (FMO) analysis, pharmacophore modeling, 100 ns molecular dynamics (MD) simulations, MM/PBSA binding free energy calculations, and per-residue decomposition. Docking predicted favorable estimated binding affinities toward all targets. MD simulations confirmed stable receptor-ligand complexes with low RMSD values (0.278-0.285 nm). MM/PBSA analysis indicated favorable binding free energies, particularly for GSK3-β (-22.73 kcal/mol) and AChE (-21.50 kcal/mol). Per-residue decomposition identified key hotspot residues driving stabilization. Overall, this structured computational framework identifies Hopkinsiaxanthin as a promising multitarget scaffold and supports its prioritization for experimental validation in AD models.