| HPLC method is broadly applied as the prior approach when analyzing the fingerprint of TCM, because of its accuracy, precision and high-efficiency. The near infrared (NIR) spectroscopy has prominent advantages such as convenient, rapid, high-efficient, low-cost, nondestructive, non-consumptive, non-polluting and on-line. The fingerprints of HPLC or NIR are able to provide qualitative and quantitative information of chemical constitutions, so that quality control and evaluation of TCM based on fingerprints will be greatly improved. In this study, identification of geographical origins and multi-component quantitation in Salvia miltiorrhiza var. alba based on HPLC and NIR fingerprints were carried out.Some chemical and physical information can be detected from the fingerprint of TCM, however, biological activities are hardly reflected. Chemometrics is capable of establishing the relationship between chemical or physical information and biological activities. In recent years, the exploration of fingerprint-effect relationship has widely been performed. In this research work, the models for assessing antioxidant activity of S. miltiorrhiza var. alba were constructed on HPLC and NIR fingerprint, and the characteristic peaks or informative spectral regions significantly related with antioxidant activity were also found. These work would provide useful references for quality assurance and evaluation of S. miltiorrhiza var. alba. Our present research work includes the following sections:1. Discrimination of S. miltiorrhiza var. alba according to geographical origins by fingerprint.HPLC fingerprint of S. miltiorrhiza var. alba from different origins was developed successfully and also validated. In addition, the identification model of geographical origin based on HPLC fingerprint was established by principal component analysis (PCA), which was subsequently applied to discrimination of Salvia miltiorrhiza var. alba samples from Taian, Laiwu, Guangrao and Rongcheng.NIR spectra of Salvia miltiorrhiza var. alba samples from4origins were determined and applied to discrimination of geographical origin. Two models for discrimination based on extract solution and powders of Salvia miltiorrhiza var. alba were built by discriminant analysis (DA), respectively. The results of DA based on the two models were also compared. The model based on sample powder achieved accurate rate up to100%for discrimination of geographical origins, while poor results of discrimination with NIR spectra of sample extract solutions were acquired. Therefore, the model based on sample powder may be as an optimal model to discriminate the Salvia miltiorrhiza var. alba samples according to geographical origins. The characteristic spectral regions with marked differences between the NIR spectra of samples from different origins were found and related absorption band assignment were preliminarily given as follows:the region of5750-6100cm"1contains information on the first overtones of C-H in CH or CH2, and5050-5150cm-1region displayed in the second rectangle includes the spectral bands associated with the combination of stretching and deformation of the O-H group; the range of4500-4660cm-1consists of information concerning the combination tone of stretching vibration and bending vibration in N-H (amide Ⅱ/Ⅲ). Besides the regions mentioned above, some spectral bands such as4073cm-1,4422cm-1,4887cm-and5253cm-1also display differences among samples of various origins.2. Quantitative analysis of multi-componentsA new HPLC method under same condition was established to determine the contents of10constituents in samples from4geographical origins. The HPLC method was also validated. The results revealed that the content of each constituent between samples from different origins had great differences.A Partial least squares regression (PLSR) model based on NIR spectra of S. miltiorrhiza var. alba samples was established for determination of the contents of6constituents. The correlation coefficients of calibration set (Rc) and validation set (Rp) for determination of each component content were all above0.85and that of salvianolic acid B or danshensu are above0.98, indicating that the model could be used for determination of the contents of6components.3. Correlative analysis between fingerprint and antioxidant activity.The antioxidant activities of90samples from4geographical origins, Vitamin E and Vitamin C were determined by DPPH method. The relationship between HPLC fingerprint and the corresponding IC50values was modeled by partial least squares regression (PLSR), and the results were as follows:Rc=0.9138, Rp=0.9342, RMSEP (root mean square error of prediction)=0.1571mg-mL-1, Q2(multi-correlation coefficient square of cross-validation)=0.684, indicating that HPLC fingerprint was able to predict the antioxidant activity of Salvia miltiorrhiza var. alba samples. Furthermore, according to the established model, salvianolic acid B was considered as the important ingredient responsible for antioxidant activity, which wasin accordance with traditional knowledge.PLSR was also employed to build the model for evaluation of antioxidant activity (AA) of Salvia miltiorrhiza var. alba samples by NIR spectroscopy (NIRS). The spectral regions were selected by interval PLS (i-PLS) method. The MSC+FD+SG was applied for the spectral pre-processing. The final optimal results of PLSR model showed that correlation coefficients in the calibration set (Rc) and the prediction set (Rp), root mean square error of prediction (RMSEP), residual prediction deviation (RPD), root mean square error of calibration (RMSEC) and root mean square error of cross-validation RMSECV were0.974,0.950,0.163mg mL-1,2.66,0.098mg-mL"1and0.223rng-mL"1, respectively. The results demonstrated that NIRS combined with chemometric methods could be a rapid and nondestructive tool to discriminate geographical origin and evaluate AA of S. miltiorrhiza var. alba. The developed NIRS method might have a potential application to high-throughput screening of a great number of raw S. miltiorrhiza var. alba samples for AA. |
PDF Full Text Request