It is a common problem in natural product therapeutic lead discovery programs that despite good bioassay results in the initial extract, the active compound(s) may not be isolated during subsequent bioassay-guided purification. Herein, we present the concept of bioactive molecular networking to find candidate active molecules directly from fractionated bioactive extracts. By employing tandem mass spectrometry, it is possible to accelerate the dereplication of molecules using molecular networking prior to subsequent isolation of the compounds, and it is also possible to expose potentially bioactive molecules using bioactivity score prediction. Indeed, bioactivity score prediction can be calculated with the relative abundance of a molecule in fractions and the bioactivity level of each fraction. For that reason, we have developed a bioinformatic workflow able to map bioactivity score in molecular networks and applied it for discovery of antiviral compounds from a previously investigated extract of Euphorbia dendroides where the bioactive candidate molecules were not discovered following a classical bioassay-guided fractionation procedure. It can be expected that this approach will be implemented as a systematic strategy, not only in current and future bioactive lead discovery from natural extract collections but also for the reinvestigation of the untapped reservoir of bioactive analogues in previous bioassay-guided fractionation efforts.

Covering: 2013 to 2019 The exploration of the chemical diversity of extracts from various biological sources has led to major drug discoveries. Over the past two decades, despite the introduction of advanced methodologies for natural product (NP) research (e.g., dereplication and high content screening), successful accounts of the validation of NPs as lead therapeutic candidates have been limited. In this context, one of the main challenges faced is related to working with crude natural extracts because of their complex composition and the inadequacies of classical bioguided isolation studies given the pace of high-throughput screening campaigns. In line with the development of metabolomics, genomics and chemometrics, significant advances in metabolite profiling have been achieved and have generated high-quality massive genome and metabolome data on natural extracts. The unambiguous identification of each individual NP in an extract using generic methods remains challenging. However, the establishment of structural links among NPs via molecular network analysis and the determination of common features of extract composition have provided invaluable information to the scientific community. In this context, new multi-informational-based profiling approaches integrating taxonomic and/or bioactivity data can hold promise for the discovery and development of new bioactive compounds and return NPs back to an exciting era of development. In this article, we examine recent studies that have the potential to improve the efficiency of NP prioritisation and to accelerate the targeted isolation of key NPs. Perspectives on the field's evolution are discussed.

Natural products represent an inexhaustible source of novel therapeutic agents. Their complex and constrained three-dimensional structures endow these molecules with exceptional biological properties, thereby giving them a major role in drug discovery programs. However, the search for new bioactive metabolites is hampered by the chemical complexity of the biological matrices in which they are found. The purification of single constituents from such matrices requires such a significant amount of work that it should be ideally performed only on molecules of high potential value (i.e., chemical novelty and biological activity). Recent bioinformatics approaches based on mass spectrometry metabolite profiling methods are beginning to address the complex task of compound identification within complex mixtures. However, in parallel to these developments, methods providing information on the bioactivity potential of natural products prior to their isolation are still lacking and are of key interest to target the isolation of valuable natural products only. In the present investigation, we propose an integrated analysis strategy for bioactive natural products prioritization. Our approach uses massive molecular networks embedding various informational layers (bioactivity and taxonomical data) to highlight potentially bioactive scaffolds within the chemical diversity of crude extracts collections. We exemplify this workflow by targeting the isolation of predicted active and nonactive metabolites from two botanical sources (Bocquillonia nervosa and Neoguillauminia cleopatra) against two biological targets (Wnt signaling pathway and chikungunya virus replication). Eventually, the detection and isolation processes of a daphnane diterpene orthoester and four 12-deoxyphorbols inhibiting the Wnt signaling pathway and exhibiting potent antiviral activities against the CHIKV virus are detailed. Combined with efficient metabolite annotation tools, this bioactive natural products prioritization pipeline proves to be efficient. Implementation of this approach in drug discovery programs based on natural extract screening should speed up and rationalize the isolation of bioactive natural products.