Multi-scale In Silico and Biochemical Evaluation of Natural Bisbenzylisoquinoline Alkaloids as Aldose Reductase Inhibitors

Abstract

Aldose reductase (AR) is the rate-limiting enzyme of the polyol pathway and a validated target for preventing micro- and neurovascular complications of diabetes. Here, we combined multi-scale in-silico analyses with biochemical testing to evaluate five commercially available bisbenzylisoquinoline alkaloids—cepharanthine, dauricine, isotetrandrine, fangchinoline and sinomenine—as potential AR inhibitors. Density-functional optimization, structure-based docking and 500 ns molecular-dynamics simulations revealed that the macrocyclic scaffolds of cepharanthine (ΔG_dock =  − 8.4 kcal mol⁻¹) and dauricine (− 9.7 kcal mol⁻¹) fully occupy the Phe122-Trp219-Trp111 aromatic cage and lock AR into a single, deep free-energy basin, whereas sinomenine explores a broad landscape. MM/PBSA calculations on the 150–200 ns of each trajectory ranked binding free energies as dauricine ≈ isotetrandrine ≈ cepharanthine < sinomenine < fangchinoline, with van-der-Waals forces dominating. ADMET profiling predicted high gastrointestinal absorption across the series but flagged a potential hERG potassium-channel liability for the four macrocycles. Enzyme-kinetic assays corroborated the computational hierarchy: cepharanthine, dauricine and isotetrandrine inhibited recombinant AR with IC₅₀ values of 4.25 ± 0.42, 5.38 ± 0.22 and 6.65 ± 0.40 µM, respectively, compared with 2.36 ± 0.32 µM for quercetin. Lineweaver–Burk and Michaelis–Menten analysis showed mixed inhibition for cepharanthine (K_i = 3.71 µM) and non-competitive inhibition for dauricine (K_i = 4.63 µM) and isotetrandrine (K_i = 6.99 µM). Fangchinoline and sinomenine were an order of magnitude weaker (IC₅₀ = 37–57 µM). Taken together, these data position cepharanthine and dauricine as mechanistically validated, hit-stage starting points for next-generation AR inhibitors, and identify isotetrandrine as an allosteric back-up scaffold. More broadly, the study illustrates a transparent, reproducible computational–experimental workflow for prioritizing structurally complex natural products against redox enzymes implicated in diabetic pathology.