ABSTRACT
Cassiae Semen (CS) is consumed as fried tea or medicinal food in Asian areas. Its two commercial forms, namely raw and fried CS, exert different clinical applications, in which fried CS is commonly applied as a functional tea for losing weight. To prevent confusion in the use of the two forms of CS, a comprehensive strategy by combining plant metabolomics and spectrum-effect relationship analyses was developed for the fast and efficient discrimination of raw and fried CS, and further for the discovery of the potential hypoglycemic metabolites of CS to control its quality. First, the plant metabolic profiling of raw and processed samples was performed by UHPLC-QTOF-MS/MS. A total of 1111 differential metabolites were found to well distinguish the raw and fried CS after analyzing by MPP software. Subsequently, α-glucosidase inhibition of raw and fried CS was investigated. As a result, fried CS demonstrated much stronger α-glucosidase inhibition activity than the raw sample. By analyzing the spectrum-effect relationship with GRA, BCA, and PLSR, 14 potential hypoglycemic-related compounds were discovered. As anticipated, they were also found as the differential metabolites of the raw and fried samples with a potential hypoglycemic effect, which might be beneficial for the quality control of CS tea. Additionally, molecular docking analysis was conducted to reveal the underlying inhibition mechanisms of the four most critical constituents, including physcion, chrysoobtusin, aurantio-obtusin, and obtusifolin. This study provides a powerful tool for the discrimination of processed samples and fast screening of the active constituents in complex natural products on a high-throughput basis.