Quality evaluation of compounds in leaves of six Taxus species based on UPLC-MS/MS and chemometrics

Abstract

Introduction: Taxus species are used as medicinal plants all over the world. The leaves of Taxus species are sustainable medicinal resources that are rich in taxoids and flavonoids. However, traditional identification methods cannot effectively identify Taxus species on the basis of leaces used as raw medicinal materials, because their appearance and morphological characteristics are almost the same, and the probability of error identification increases in accordance with the subjective consciousness of the experimenter. Moreover, although the leaves of different Taxus species have been widely used, their chemical components are similar and lack systematic comparative research. Such a situation is challenging for quality assessment. Materials and methods: In this study, ultra-high-performance liquid chromatography coupled with triple quadrupole mass spectrometry combined with chemometrics was applied for the simultaneous determination of eight taxoids, four flavanols, five flavonols, two dihydroflavones, and five biflavones in the leaves of six Taxus species, namely, T. mairei, T. chinensis, T. yunnanensis, T. wallichiana, T. cuspidata, and T. media. Chemometric methods, including hierarchical cluster analysis, principal component analysis, orthogonal partial least squares-discriminate analysis, random forest iterative modeling, and fisher linear discriminant analysis, were utilized to differentiate and evaluate the six Taxus species. Results: This proposed method exhibited good linearity (R ² = 0.9999-0.9972) with a lower quantification limits of 0.94-3.05 ng/mL for all analytes. The intra- and inter-day precisions were within 6.83%. Six compounds, namely, 7-xylosyl-10-deacetyltaxol, ginkgetin, rutin, aromadendrin, 10-deacetyl baccatin III, and epigallocatechin, were identified through chemometrics for the first time. These compounds can be used as important chemical markers to distinguish the above six Taxus species rapidly. Conclusion: This study established a method for determination of the leaves of six Taxus species, and revealing the differences in the chemical components of these six Taxus species.

Keywords: Taxus species; chemometrics; flavonoids; quality control; taxoids.

FIGURE 1

The chemical structures of 24 analytes and 3 internal standards. 1. 10-DAB, 2. BAC, 3. 7-xyl-10-DAT, 4. 10-DAT, 5. CE, 6. 7-epi-10-DAT, 7. TAXOL, 8. 7-epi-TAXOL, 9. GC, 10. EGC, 11. C, 12. EC, 13. TAX, 14. ARO, 15. RT, 16. IQC, 17. NFR, 18. QC, 19. QR, 20. AF, 21. DGK, 22. GK, 23. IGG, 24. SDN, IS1.TP, IS2.CAS, IS3.LIQ; The shadow area shows six chemical quality markers with distinguishing radical pharmacophores.

FIGURE 2

UPLC-MS/MS chromatograms of 24 analytes in mixed standard solution. The sequence of 24 analytes in the figure is consistent with the number in Table 1.

FIGURE 3

GC (A), RT (B), TAX (C), SDN (D), and TAXOL (E) are used as examples to clarify the detailed identification processes of flavanols, flavonols, dihydroflavones, biflavones, and taxoids.

FIGURE 4

Comparative analysis of 24 components in 42 leaf samples from the six Taxus species. The mean ± SDs not sharing the same lowercase letters are signifificantly different (p < 0.05). 1: T. mairei, 2: T. chinensis, 3: T. yunnanensis, 4: T. wallichiana, 5: T. cuspidata, 6: T. media.

FIGURE 5

Heatmap and dendrogram of HCA in 42 leaf samples from the six Taxus species. S1-S42 represents 42 leaf samples from the six Taxus species. 1: T. mairei, 2: T. chinensis, 3: T. yunnanensis, 4: T. wallichiana, 5: T. cuspidata, 6: T. media, and 24 analyte names are shown at the right of the figure. The colors from red, yellow and blue represent the content of analytes in the sample from high to low.

FIGURE 6

Scores (A) and Accumulated R²X and Q² with different numbers of principal components (B) of PCA; scores (C) of OPLS-DA and the variable importance predictive value (VIP) of 24 analytes (D) in 42 leaf samples from the six Taxus species.

FIGURE 7

Cumulative number of each variable in the top Nth of 100 RF iterative modeling (A). Frequency of the variable as the most important parameter in 100 RF iterative modeling (B).