Structure-Based In-Silico Evaluation of Arbutin against the Invasion– Metastasis Targets MRCK-β, 14-3-3σ and Heparanase: Molecular Docking, Protein–Ligand Interaction and ADMET Profiling

AbstractBackground: Arbutin (hydroquinone-O-β-D-glucopyranoside) is a phenolic glycoside usedwidely as a skin-depigmenting agent, but its potential to engage the protein machinery of tumourinvasion is undefined. We evaluated, by a reproducible structure-based workflow, whetherarbutin binds three mechanistically distinct drivers of cutaneous tumour invasion and comparedit with the topical agent 5-fluorouracil.Methods: α- and β-arbutin, 5-fluorouridine and 5-fluorouracil were docked (AutoDock Vina1.2.7) against five human crystal structures: MRCK-β (PDB 5OTE), 14-3-3σ (7OBC, 7OBX,8RRM) and heparanase (7PR7). Each ligand–receptor pair was docked in triplicate withindependent seeds; the protocol was validated by re-docking the co-crystallised ligands. Bestposes were profiled with PLIP, and physicochemical, drug-likeness and ADMET properties werecomputed with RDKit and ADMET-AI.Results: Re-docking reproduced BDP-00008900 (RMSD 0.54 Å) and fusicoccin-A (1.34 Å),validating the protocol. Both arbutin anomers bound all three target classes with predictedaffinities of -6.0 to -6.6 kcal/mol, essentially invariant across structurally unrelated sites, andoutscored 5-fluorouracil at every target by 1.0–2.0 kcal/mol. In the heparanase catalytic cleft α-arbutin formed ten hydrogen bonds, including a direct contact with the catalytic nucleophileGlu343; in 14-3-3σ it occupied the phosphoserine-coordinating sub-site; and in MRCK-β itadopted a canonical type-I kinase orientation, although the designed inhibitor BDP-00008900scored 2.5 kcal/mol stronger. Arbutin was predicted to be highly polar with low passivepermeability, no P-glycoprotein or cytochrome-P450 liability, and a markedly more favourabletoxicity profile than the fluoropyrimidines (predicted DILI 0.14 vs 0.81; LD₅₀ ≈ 2,700 vs 436mg/kg).Conclusion: Arbutin engages all three invasion targets with chemically specific interactions,most compellingly the heparanase catalytic cleft, whose extracellular localisation matchesarbutin’s poorly permeable physicochemistry and the requirements of a topical agent. Thesecomputational predictions warrant biochemical and cellular validation.