Chemical taxonomy of tree peony species from China based on root cortex metabolic fingerprinting

Abstract

The section Moutan of the genus Paeonia consists of eight species that are confined to a small area in China. A wide range of metabolites, including monoterpenoid glucosides, flavonoids, tannins, stilbenes, triterpenoids, steroids, paeonols, and phenols, have been found in the species belonging to section Moutan. However, although previous studies have analyzed the metabolites found in these species, the metabolic similarities that can be used for the chemotaxonomic distinction of section Moutan species are not yet clear. In this study, HPLC–DAD-based metabolic fingerprinting was applied to the classification of eight species: Paeoniasuffruticosa, Paeoniaqiui, Paeoniaostii, Paeoniarockii, Paeoniajishanensis, Paeoniadecomposita, Paeoniadelavayi, and Paeonialudlowii. In total, of the 47 peaks that exhibited an occurrence frequency of 75% in all 23 tree peony samples, 43 of these metabolites were identified according to their retention times and UV absorption spectra, together with combined HPLC–QTOF-MS. These data were compared with reference standard compounds. The 43 isolated compounds included 17 monoterpenoid glucosides, 11 galloyl glucoses, 5 flavonoids, 6 paeonols and 4 phenols. Principal component analysis (PCA), and hierarchical cluster analysis (HCA), showed a clear separation between the species based on metabolomics similarities and four groups were identified. The results exhibited good agreement with the classical classification based on the morphological characteristics and geographical distributions of the subsections Vaginatae F.C. Stern and Delavayanae F.C. Stern with the exception of P. decomposita, which was found to be a transition species between these two subsections. According to their metabolic fingerprinting characteristics, P. ostii and P. suffruticosa can be considered one species, and this result is consistent with the viewpoint of medicinal plant scientists but different from that of classical morphological processing. Significantly large variations were obtained in the metabolic profiles of P. delavayi, whereas no significant difference was found between P. delavayi and P. ludlowii. This indicates that these two species have a close genetic relationship. In conclusion, the combination of HPLC–DAD and multivariate analyses has great potential for guiding future chemotaxonomic studies to examine the potential pharmaceutical value of the effective constituents of tree peony species and appears to be able to clarify the confusion and skepticism associated with the reported morphology- and molecular phylogenetics-based taxonomy of tree peonies.