Metabolomic profiling of Zanthoxylum species: Identification of anti-cholinesterase alkaloids candidates

Abstract The isolation of bioactive compounds from natural sources is a key step in drug discovery and development, however, this procedure is usually expensive and difficult due to the complexity and the limited amounts of the metabolites in the extracts. Thus, rational or targeting isolations are becoming more popular to reduce the bottlenecks in bioactive natural products research. In this study, we used a LC-MS-based metabolomic approach and biochemometric statistical tools (PCA and OPLS-DA) to identify potential anti-cholinesterase alkaloids predictors in Zanthoxylum genus (Rutaceae). For this purpose, 41 alkaloid extracts from nine Colombian Zanthoxylum species were screened by UHPLC-UV-HRMS and inhibitory activity against Acetylcholinesterase (AChE) and Butyrylcholinesterase (BChE). Based on the screening results, a multivariate statistical analysis (MVA) and selection of anti-cholinesterase candidates were performed using the S-plot from the OPLS-DA model. The supervised analysis (OPLS-DA) paring the anti-cholinesterase screening and LC-HRMS data showed at least 11 ChE inhibition markers which could have contributed in the differentiation of active and inactive extracts. The predictors were tentatively identified by comparing chromatographic retention times (Rt) and accurate mass and MS2 fragmentation patterns. In general, the inhibition markers correspond to four types of isoquinoline alkaloids: tetrahydroprotoberberines, protoberberines, dihydrobenzophenanthridines and benzophenanthridines. The most active extracts from Z. schreberi and Z. monophylum showed the highest presence of berberine and chelerythrine, previously reported as cholinesterase inhibitors. Thus, to validate the results of the OPLS-DA model, three alkaloids from the bark of Z. schreberi (identified as berberine, chelerythrine and columbamine) were bio-directed isolated, and all of them showed strong inhibition against both enzymes. These findings support our statistical models and contribute to the rational search of anticholinesterase alkaloids. Therefore, LC-MS-based metabolomic approach combined with chemometric statistical analysis are shown as useful tools for the isolation of targeted bioactive natural products, contributing to improve the research and development stages of lead compounds.