Differentiation between wild and artificial cultivated Stephaniae tetrandrae radix using chromatographic and flow‐injection mass spectrometric fingerprints with the aid of principal component analysis

Abstract High‐performance liquid chromatographic (HPLC) and flow‐injection mass spectrometric (FIMS) fingerprinting profiles were used to differentiate between wild and artificial cultivated Stephaniae tetrandrae Radix samples. HPLC and FIMS fingerprints of 15 wild S. tetrandrae Radix samples and 12 artificial cultivated S. tetrandrae Radix samples were obtained and analyzed with the aid of principal component analysis (PCA). PCA of the fingerprints showed that the chemical differences between wild and artificial cultivated S. tetrandrae Radix samples could be differentiated by either HPLC or FIMS fingerprints. The HPLC fingerprints provided more chemical information but required longer analytical time compared with FIMS fingerprints. This study indicated that the wild samples contained higher concentrations of almost all of the major compounds than the cultivated samples. Three characteristic compounds which were responsible for the differences between the samples were tentatively identified with the aid of MS data. Furthermore, these three compounds, tetrandrine (TET), fangchinoline (FAN), and cyclanoline (CYC), were quantified. The HPLC and FIMS fingerprints combined with PCA could be used for quality assessment of wild and artificial cultivated S. tetrandrae Radix samples.

it increased zinc content (Lu, Liu, Li, & Fan, 2019). These results reveal the beneficial health effects are closely related to TET, FAN, and alkaloids compounds.
Historically, the dried root of Aristolochia fangchi Y.C.Wu ex L.D.
At present, almost all the S. tetrandrae Radix medicinal plants are wild. With the increasing demand of the herb, the plants were over collected, causing acute shortage of the herb (Huang, Liu, & Li, 2008;Qing & Wang, 2007). A few enterprises and medicine material farms are trying to cultivate S. tetrandrae S. Moore. However, S. tetrandrae Radix cannot be applied in large scale for lack of cultivation techniques. To the best of my knowledge, the quality assessment on artificial cultivated S. tetrandrae Radix has never been reported.
High-performance liquid chromatographic (HPLC) fingerprints (Li, Zhang, & Yang, 2017;Liang et al., 2019;Zhou, Qin, Wang, & Huang, 2018) and determination, identification of active chemical components (Liu et al., 2018;Lu et al., 2015;Sim, Kim, Lee, & Hong, 2013;Wang et al., 2015) of S. tetrandrae Radix and its prescription were reported. However, no study has been reported on comparison between wild and cultivated S. tetrandrae Radix. In our previous studies, chromatographic fingerprints combined with chemometrics have been successfully used to evaluate the quality of S. tetrandrae Radix collected from different habitats (Zhou et al., 2018). On the top of this, we aimed to differentiate between wild and cultivated S. tetrandrae Radix using chromatographic and FIMS fingerprints combined with chemometrics.

TA B L E 1
The detailed information of the tested samples and the concentration values (mg/g) of the three alkaloids in these samples

| Reference standards
Reference standards of fangchinoline (FAN) and tetrandrine (TET) were acquired from National Institutes for Food and Drug Control.
The reference standard of cyclanoline (CYC) and magnoflorine (MAG) was obtained from Chengdu Must Bio-technology Co., LTD and Shanghai Yuanmu Bio-technology Co., LTD, respectively. Their purities were above 98% by HPLC analysis.

| Reagents
Acetonitrile (Tedia) and formic acid (Aladdin) were MS grade. Water was Wahaha pure water (Wahaha Food and Beverage Ltd.). All other regents were of analytical grade.

| Sample preparation
The S. tetrandrae Radix samples were dried and ground to power at 60-mesh particle size and stored at −4°C prior to analysis. Two

| HPLC and FIMS Conditions
The HPLC analysis was performed according to our previously reported laboratory procedure (Zhou et al., 2018) To get rid of the solvent interferences, the chromatographic data used for peak integration were retention time between 2 and 20 min. Eight peaks from each chromatographic fingerprint were selected as common peaks. Then, the absolute peak area was exported from SES and saved in.csv format. Peak 7, the predominant peak in both the WS and CS samples, was selected as the reference peak. Data sets of absolute peak areas of the eight common peaks and the relative ratios of the other seven peaks to the reference peak were both used for PCA. Next, each missing m/z in mass list was filled with zero so that each sample contained 901 data points. Finally, two-dimensional matrix (81 samples × 901 masses) was then export to SIMCA-P for PCA.
Preprocessing in SIMCA-P, prior to PCA, consisted of normalization and mean centering.

| Quantitative analysis
In this study, the contents of TET, FAN, and CYC in 27 batches of S. tetrandrae Radix samples were quantified. The concentration values of three compounds' mean and standard deviation (SD) are displayed in Table 1.
The results showed that TET, FAN, and CYC are the three major alkaloids in the samples. The mean concentration values for TET, FAN, and F I G U R E 3 PCA scores plot (a) and loadings plot (b) for absolute peak areas in HPLC fingerprints of Stephaniae tetrandrae Radix samples CYC were 10.46, 6.80, and 0.27 mg/g in wild samples, respectively. The values were two times higher than that in cultivated samples (p < .01).

| Chromatographic and FIMS fingerprints
Typical chromatographic fingerprints for mixed standards, wild, and cultivated samples are shown in Figure 1

| PCA of chromatographic fingerprints
Principal component analysis concerns a mathematical procedure that transforms a number of possibly correlated variables into a smaller number of uncorrelated variables (Huang et al., 2020;Zhao, Chang, & Chen, 2015). PCA provides visual patterns (Castro-Alba et al., 2019;Chen, Sun, & Ford, 2014;Reale et al., 2020;Sun, Sun, & Han, 2018) which can be understood and accepted easily; furthermore, the results also avoid subjective decisions. In the present study, to assess the resemblance and differences between wild and cultivated samples, a PCA was performed based on the eight common peaks (Figure 1). The absolute peak areas of the eight common peaks in the 27 chromatograms  Figure B). We found that the contents of them were statistically significant (p < .01) in wild S. tetrandrae samples than that in cultivated S. tetrandrae samples. illustrates that peak 1 (CYC) and 4 (TET) are responsible for the separation of WS and CS samples and contribute positively to PC1. Peak 7 (FAN), selected as the reference peak, is located at zero on both x and y axes. These three peaks were found to be the characteristic peaks in differentiation between WS and CS samples, in the present study.

| PCA of FIMS fingerprints
Typical FIMS fingerprint of WS6 and CS6 are shown in Figure 2a Figure 5a shows the PCA scores plot of FIMS F I G U R E 5 PCA scores plot (a) and loadings plot (b) for flow-injection mass spectrometric fingerprints of wild and cultivated Stephaniae tetrandrae Radix samples