Molecular networking (MN) is becoming a standard bioinformatics tool in the metabolomic community. Its paradigm is based on the observation that compounds with a high degree of chemical similarity share comparable MS2 fragmentation pathways. To afford a clear separation between MS2 spectral clusters, only the most relevant similarity scores are selected using dedicated filtering steps requiring time-consuming parameter optimization. Depending on the filtering values selected, some scores are arbitrarily deleted and a part of the information is ignored. The problem of creating a reliable representation of MS2 spectra data sets can be solved using algorithms developed for dimensionality reduction and pattern recognition purposes, such as t-distributed stochastic neighbor embedding (t-SNE). This multivariate embedding method pays particular attention to local details by using nonlinear outputs to represent the entire data space. To overcome the limitations inherent to the GNPS workflow and the networking architecture, we developed MetGem. Our software allows the parallel investigation of two complementary representations of the raw data set, one based on a classic GNPS-style MN and another based on the t-SNE algorithm. The t-SNE graph preserves the interactions between related groups of spectra, while the MN output allows an unambiguous separation of clusters. Additionally, almost all parameters can be tuned in real time, and new networks can be generated within a few seconds for small data sets. With the development of this unified interface (), we fulfilled the need for a dedicated, user-friendly, local software for MS2 comparison and spectral network generation.

Molecular networking is becoming more and more popular into the metabolomic community to organize tandem mass spectrometry (MS) data. Even though this approach allows the treatment and comparison of large data sets, several drawbacks related to the MS-Cluster tool routinely used on the Global Natural Product Social Molecular Networking platform (GNPS) limit its potential. MS-Cluster cannot distinguish between chromatography well-resolved isomers as retention times are not taken into account. Annotation with predicted chemical formulas is also not implemented and semiquantification is only based on the number of MS scans. We propose to introduce a data-preprocessing workflow including the preliminary data treatment by MZmine 2 followed by a homemade Python script freely available to the community that clears the major previously mentioned GNPS drawbacks. The efficiency of this workflow is exemplified with the analysis of six fractions of increasing polarities obtained from a sequential supercritical CO extraction of Stillingia lineata leaves.

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.

In an effort to identify new potent and selective inhibitors of chikungunya virus and HIV-1 and HIV-2 virus replication, the endemic Mascarene species Stillingia lineata was investigated. LC/MS and bioassay-guided purification of the EtOAc leaf extract using a chikungunya virus-cell-based assay led to the isolation of six new (4-9) and three known (1-3) tonantzitlolones possessing the rare C20-flexibilane skeleton, along with tonantzitloic acid (10), a new linear diterpenoid, and three new (11, 13, and 15) and two known (12 and 14) tigliane-type diterpenoids. The planar structures of the new compounds and their relative configurations were determined by spectroscopic analysis, and their absolute configurations were determined through comparison with literature data and from biogenetic considerations. These compounds were investigated for selective antiviral activity against chikungunya virus (CHIKV), Semliki Forest virus, Sindbis virus, and, for compounds 11-15, the HIV-1 and HIV-2 viruses. Compounds 12-15 were found to be the most potent and are selective inhibitors of CHIKV, HIV-1, and HIV-2 replication. In particular, compound 14 inhibited CHIKV replication with an EC50 value of 1.2 μM on CHIKV and a selectivity index of >240, while compound 15 inhibited HIV-1 and HIV-2 with EC50 values of 0.043 and 0.018 μM, respectively. It was demonstrated further that potency and selectivity are sensitive to the substitution pattern on the tigliane skeleton. The cytotoxic activities of compounds 1-10 were evaluated against the HCT-116, MCF-7, and PC3 cancer cell lines.

From a set of 292 Euphorbiaceae extracts, the use of a molecular networking (MN)-based prioritization approach highlighted three clusters (MN1−3) depicting ions from the bark extract of Codiaeum peltatum. Based on their putative antiviral potential and structural novelty, the MS-guided purification of compounds present in MN1 and MN2 afforded two new daphnane-type diterpenoid orthoesters (DDO), codiapeltines A (1) and B (2), the new actephilols B (3) and C (4), and four known 1,4-dioxane-fused phenanthrene dimers (5−8). The structures of the new compounds were elucidated by NMR spectroscopic data analysis, and the absolute configurations of compounds 1 and 2 were deduced by comparison of experimental and calculated ECD spectra. Codiapeltine B (2) is the first daphnane bearing a 9,11,13-orthoester moiety, establishing a new major structural class of DDO. Compounds 1−8 and four recently reported monoterpenyl quinolones (9− 12) detected in MN3 were investigated for their selective activities against chikungunya virus replication and their antipolymerase activities against the NS5 proteins of dengue and zika viruses. Compounds 3−8 exhibited strong inhibitory activities on both dengue and zika NS5 in primary assays, but extensive biological analyses indicated that only actephilol B (3) displayed a specific interaction with the NS5 targets.

Four unprecedented chlorinated monoterpenyl quinolones have been isolated using a molecular networking based prioritisation strategy. The bio-inspired total synthesis of chloroaustralasine A involving a chloroperoxydase-mediated hydroxychlorination is described.

New omics sciences generate massive amounts of data, requiring to be sorted, curated, and statistically analyzed by dedicated software. Data-dependent acquisition mode including inclusion and exclusion rules for tandem mass spectrometry is routinely used to perform such analyses. While acquisition parameters are well described for proteomics, no general rule is currently available to generate reliable metabolomic data for molecular networking analysis on the Global Natural Product Social Molecular Networking platform (GNPS). Following on from an exploration of key parameters influencing the quality of molecular networks, universal optimal acquisition conditions for metabolomic studies are suggested in the present paper. The benefit of data pre-clustering before initiating large datasets for GNPS analyses is also demonstrated. Moreover, an efficient workflow dedicated to Agilent Technologies instruments is described, making the dereplication process easier by unambiguously distinguishing isobaric isomers eluted at different retention times, annotating the molecular networks with chemical formulas, and giving access to semi-quantitative data. This specific workflow foreshadows future developments of the GNPS platform.

Bioassay-guided fractionation of an EtOAc extract of the trunk bark of Sandwithia guyanensis, using a chikungunya virus (CHIKV)-cell-based assay, afforded 17 new diterpenoids 1-17 and the known jatrointelones A and C (18 and 19). The new compounds included two tetranorditerpenoids 1 and 2, a trinorditerpenoid 3, euphoractines P-W (4-11), and euphactine G (13) possessing the rare 5/6/7/3 (4-7), 5/6/6/4 (8-11), and 5/6/8 (13) fused ring skeletons, sikkimenoid E (12), and jatrointelones J-M (14-17) possessing jatropholane and lathyrane carbon skeletons, respectively. Jatrointelones J (14) and M (17) represent the first naturally occurring examples of C-15 nonoxidized lathyrane-type diterpenoids. The structures of the new compounds were elucidated by NMR spectroscopic data analysis. The relative configuration of compound 16 and the absolute configurations of compounds 3-6 and 14 were determined by single-crystal X-ray diffraction analysis. In addition, jatrointelone K (15) was chemically transformed to euphoractine T (8) supporting the biosynthetic relationships between the two types of diterpenoids. Only compound 15 showed a moderate anti-CHIKV activity with an EC value of 14 μM. Finally, using a molecular networking-based dereplication strategy, several close analogues of 12- O-tetradecanoylphorbol-13-acetate (TPA), one of the most potent inhibitors of CHIKV replication, were dereplicated.