Integrating metabolite-based molecular networking with database matching and LC-MS-guided targeted isolation for the discovery of novel chemical constituents: application to Euphorbia helioscopia L.

Abstract

Molecular networking (MN) analysis facilitates the targeted discovery of novel constituents and enhances the understanding of natural products. While various molecular networks could reduce the effects of redundant nodes, current research is still limited by the interference from the same and coeluted metabolites, including isotopic peaks, a variety of adduct ions, in-source fragmentations, and dehydration. This research proposes a novel strategy: stratified precursor lists (SPLs)-guided Metabolite-Based Molecular Networking (MBMN), which ensures a high-quality MS² spectrum for each metabolite precursor due to the absence of retention time overlap with other coeluted metabolites, and each node represents a unique metabolite. By collecting over 40 MS² databases from multiple online platforms and public databases, an integrated MS² database (IM2DB) containing more than two million MS² fragmentation data was constructed. In addition, a customized MS¹ database (M1DB) of reported compounds was also created. Nodes representing known compounds were annotated compared to the IM2DB and M1DB. Combining with MBMN analysis significantly enhances compound identification and characterization, thereby facilitating the discerning of potential novel constituents. To demonstrate the applicability of this strategy, we selected Euphorbia helioscopia L. as an example. 135 nodes were annotated, and three reference nodes were obtained. From the selected 35 target nodes, 10 purified compounds were isolated and elucidated. Among these, three were identified as novel compounds, while the remaining nine were discovered for the first time in Euphorbia helioscopia L. By using this strategy, we can effectively minimize the interference from redundant nodes and discover potentially new compounds.