Strategy of mass defect filter combined with characteristic fragment analysis for the chemical profiling of Dictamnus dasycarpus Turcz. From multiple regions

Abstract

In this study, a comprehensive data filtering and identification strategy was developed. The five-point mass defect filtering (MDF) screening was combined with characteristic fragments determined based on representative reference standards and literature to identify quinoline alkaloids and limonoids in the root bark of Dictamnus dasycarpus Turcz. (DD) extract. Compared with typical characterization method, by using the established five-point MDF model, a large number of background interference ions were filtered out with an efficiency of 82 % (from 6280 to 1136). It’s the first time to construct a characteristic fragments database based on six substructures of quinoline alkaloids and limonoids in DD. Finally, by using CLogP values and dipole moments to determine the elution order of isomers, a total of 113 compounds were identified efficiently and accurately, including 95 quinoline alkaloids and 18 limonoids. Additionally, 33 compounds from DD were identified for the first time, and 23 were regarded as potential novel compounds for the first time (7 furoquinoline alkaloids, 11 quinolinone alkaloids, 1 quinolone, 1 A.D-ring open-loop limonoid-1, and 3 A.D-ring open-loop limonoids-2). Untargeted metabolomics combined with machine learning-based chemometrics was applied to analyze the metabolite profiles of 72 batches of DD from 12 production regions. Finally, a total of 27 differential metabolites (21 quinoline alkaloids and 6 limonoids) were identified by the PLS-DA analysis. The results indicated that this method was an efficient, accurate, and promising approach for classifying and exploring compounds in the complex system of natural products, providing a basis for evaluating the quality of DD from different sources.