Screening of minor bioactive compounds from herbal medicines by in silico docking and the trace peak exposure methods

• A strategy for screening trace enzyme inhibitors in herbal medicines was proposed.
• UF-LC-QTOF MS was used to preliminarily find the α-glucosidase binders.
• Docking analysis was adopted to predict the potential trace inhibitors.
• Segment and exposure approach based LC-QTOF MS was applied to identify trace compounds.
• 11 Compounds of Ginkgo biloba extract were identified as high-quality α-glucosidase inhibitors.

Screening of high potent enzyme inhibitors from herbal medicines is always lacking of efficiency due to the complexity of chemicals. The constituents responsible for the holistic effect may be trace-level chemicals, but these chemicals were covered by highly abundant compositions. To challenge this bottleneck, a strategy for screening minor bioactive compounds was proposed. It generally included four steps, (1) preliminarily find the enzyme binders by ultrafiltration; (2) optimise and predict the potential inhibitors by docking analysis; (3) selectively identify and prepare trace compounds by segment and exposure approach; (4) validate the activity and summarize the structure-activity relationship. As a case study, α-glucosidase (AGH) and Ginkgo biloba extract were used as the experimental materials. By comprehensive screening, 11 trace flavones were screened out and identified as strong AGH inhibitors. Among them, AGH inhibitory activities of syringetin and sciadopitysin were reported for the first time. Their IC50 values were 36.80 and 8.29 μM, respectively, which were lower than that of a positive control acarbose. In addition, the AGH inhibitory activities of the flavonoids could be ranked, based on a decreased order, as biflavone, flavone, flavone glycoside, flavone biglycoside. The strategy is expected to be practical and useful for screening bioactive molecules from herbal medicines, especially for the minor compounds, which will definitely accelerate the discovery of new drug candidates.