Virtual screening of protein tyrosine phosphatase 1B inhibitors based on natural products

Protein tyrosine phosphatase 1B (PTP1B) is one of the targets of type II diabetes, screening PTP1B inhibitors is of great significance. Structure-based virtual screening against a library of natural products containing 42 296 molecules was conducted to determine the occurrence of PTP1B inhibitors by molecular docking method. Firstly, the active sites of PTP1B complex crystal structure (PDB code: 1XBO) were analyzed and 7 amino acid residues, Arg254, Gln262, Tyr46, Asp181, Ser216, Phe182, and Arg221, were identified as the active pocket. Before docking, all the molecules were filtered according to the Lipinski's Rule of Five. Then, the screening was carried out based on the LibDock module and CDOCKER module, and 11 top-scored compounds were screened out as virtual hits. Of which 3 molecules, namely para-benzoquinone compound 7, isocoumarins derivative 10 and Clavepictine analogue 11, were determined with low toxicity ultimately according to the predictive ADME simulation and predictive toxic simulation. Binding model analysis revealed that these 3 candidate compounds are all good drug-like PTP1B inhibitors, of which the PTP1B inhibitory activity of compound 10 and 11 haven't been reported before, of which in vitro PTP1B enzyme inhibition of compound 10 was tested with IC₅₀ values of (74.58±1.23) μmol/L, which is potential for the treatment of type II diabete.