Exploration of habitat-related chemomarkers for Magnoliae officinalis cortex applying both global and water-soluble components-based metabolomics method

• Published in: Phytomedicine (Stuttgart) 2022-04, Vol.98, p.153957-153957, Article 153957
• Main Authors: Xue, Zhenzhen, Zhang, Xiaobo, Peng, Huasheng, Zhu, Shoudong, You, Jinwen, Zhou, Tao, Yu, Liying, Song, Chuan, Yang, Bin
• Format: Article
• Language: English
• Subjects: Global components; Habitat-related chemomarkers; Magnoliae Officinalis Cortex; Quality evaluation; Traditional Chinese medicines; Water-soluble components
• ISSN: 0944-7113; 1618-095X
• DOI: 10.1016/j.phymed.2022.153957

•A multi-metabolomics approach was proposed to explore habitat-related markers.•It involved global compounds and water-soluble components analysis.•MOC samples from 21 habitats were analyzed using UPLC/Q-TOF-MS, GC-MS and LC-2ECD.•Support vector machines model could predict MOC samples applying chemomarkers.•Temperature and precipitation were of importance to accumulate chemomarkers. The quality of traditional Chinese medicines (TCMs) has been closely related to their growth regions. The geo-herbalism of TCMs is just like the protected destination of origin on foodstuffs and wines, telling us the specific geographic regions could yield TCMs with superior quality. However, the impact of habitat on TCMs could hardly been indicated in current quality evaluation, defects were as follows: (1) few studies involved the effect of environmental factors, (2) more attentions were paid to several abundant compounds, while global components especially water-soluble compounds were prone to be ignored. A new integrated metabolomics analysis based on global and water-soluble components was proposed aiming to explore habitat-related chemomarkers for TCMs combined with correlation analysis to environmental factors. The geo-herbalism of Magnoliae officinalis cortex (MOC) was studied as an example. Multi-metabolomics approach based on UPLC/Q-TOF-MS and GC-MS combined with LC-2ECD were employed to analyze global components and accurately quantified water-soluble compounds, respectively. Meanwhile, decision tree, partial least squares discriminant analysis (PLS-DA) as well as hierarchical clustering analysis (HCA) heat map was applied to classify different samples and explore habitat-related chemomarkers. In addition, support vector machines model was used to verify the importance of screened out chemomarkers in predicting sample classification, and the impact of environmental factors on the markers were also demonstrated by correlation analysis. By analyzing 148 batches of MOC samples from 21 habitats, 238 variables were picked and 84 of them were identified by UNIFI, meanwhile, seven water-soluble compounds were accurately quantified. Among them, thirteen markers including Var.1, magnolignan E, magnoloside N isomer, α-agarofuran, γ-eudesmol, β-eudesmol, magnolosides A, B, D, F, H, L and M were suggested importance in grouping Chuan-po and the other MOC samples. Support vector machines model also indicated well prediction performance with an accuracy of 96.97%. Most markers be