Abstract

Perennial plant Gentiana lutea L. is used worldwide for the preparation of pharmaceutical and food products. Health benefits of G. lutea roots are associated with the presence of major bitter-tasting secoiridoid gentiopicroside and xanthone isogentisin. The aim was to optimize the heat-assisted extraction of gentipicroside (GP), isogentisin (ISG) and total phenolics (TP) from G. lutea roots and develop models with high accuracy and prediction capacity by response surface methodology (RSM) and artificial neural networks (ANN). Extracts were prepared according to central composite design. Significant independent variables which were previously identified by Plackett-Burman screening design were varied at five levels - temperature (20−80 °C), time (8−180 min), solid-to-solvent ratio (1:10−1:50) and ethanol concentration (10–70 %). Contents of GP, ISG (by HPLC-DAD) and TP (by Folin–Ciocalteu method) were analyzed. The optimal conditions for the extraction were temperature of 65 °C, time of 129.08 min, solid-to-solvent ratio of 1:40, and ethanol concentration 49.33 %. Under these conditions, experimentally obtained results for GP (18.03 mg/g dw), ISG (8.15 mg/g dw) and TP (17.46 mg of gallic acid equivalents/g dw) content were in agreement with the values predicted by RSM and ANN. Comparison of models through the coefficient of determination (R2) and the root mean square error (RMSE) showed that ANN approach was superior to RSM in predicting and modelling GP, ISG and TP content, simultaneously. Effective heat-assisted extraction method for the extraction of GP, ISG and TP from the roots of G. lutea was designed, and models with high accuracy and good prediction capacity were developed.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call