Partial least squares-1 (PLS-1), a chemometrics-assisted attenuated total reflectance infrared (ATR-IR) spectroscopy method, for quantitative analysis of caffeic acid and rosmarinic acid contents in Rang Chuet (Thunbergia laurifolia Lindl.) leaf samples was developed and validated. PLS-1 models were performed by using 69 IR spectra of calibration set samples. Leave-one-out cross validation was utilized for model construction. Prediction efficiency of the optimum model was checked from the 24 determination results of test set samples which were not contributed in modelling step. The optimum PLS-1models for quan titative analysis of caffeic acid and rosmarinic acid contents were constructed from the first derivative pretreated IR spectral data in the wave number range 700-1,700, 2,800-3,020 cm-1 and 700-1,700 cm-1, respectively. The final PLS-1 models were validated for linearity and range, accuracy, specificity and precision according to ICH guideline. The validation results were found acceptable in every validation parameters. High-performance liquid chromatography (HPLC) method was used as a standard method for comparison of caffeic acid and rosmarinic acid contents throughout the study. From the results, caffeic and rosmarinic contents in T. laurifolia leaf samples collected from different locations in Thailand ranged from 0.62-1.90 and 1.12-4.56 mg/g of dried weight, respectively. The results of the tested set of T. laurifolia leaf samples obtained from the developed PLS-1 models were not significantly different with those obtained from HPLC method.
Keywords
Thunbergia laurifolia; Caffeic acid; Rosmarinic acid; Attenuated total reflectance infrared spectroscopy; High performance liquid chromatography