A major goal in natural product discovery programs is to rapidly dereplicate known entities from complex biological extracts. We demonstrate here that molecular networking, an approach that organizes MS/MS data based on chemical similarity, is a powerful complement to traditional dereplication strategies. Successful dereplication with molecular networks requires MS/MS spectra of the natural product mixture along with MS/MS spectra of known standards, synthetic compounds, or well-characterized organisms, preferably organized into robust databases. This approach can accommodate different ionization platforms, enabling cross correlations of MS/MS data from ambient ionization, direct infusion, and LC-based methods. Molecular networking not only dereplicates known molecules from complex mixtures, it also captures related analogs, a challenge for many other dereplication strategies. To illustrate its utility as a dereplication tool, we apply mass spectrometry-based molecular networking to a diverse array of marine and terrestrial microbial samples, illustrating the dereplication of 58 molecules including analogs.

A new cyclic hexapeptide, baceridin (1), was isolated from the culture medium of a plant-associated Bacillus strain. The structure of 1 was elucidated by HR-HPLC-MS and 1D and 2D NMR experiments and confirmed by ESI MS/MS sequence analysis of the corresponding linear hexapeptide 2. The absolute configurations of the amino acid residues were determined after derivatization by GC-MS and Marfey's method. The cyclopeptide 1 consists partially of nonribosomal-derived D- and allo-D-configured amino acids. The order of the D- and L-leucine residues within the sequence cyclo(-L-Trp-D-Ala-D-allo-Ile-L-Val-D-Leu-L-Leu-) was assigned by total synthesis of the two possible stereoisomers. Baceridin (1) was tested for antimicrobial and cytotoxic activity and displayed moderate cytotoxicity (1-2 μg mL(-1)) as well as weak activity against Staphylococcus aureus. However, it was identified to be a proteasome inhibitor that inhibits cell cycle progression and induces apoptosis in tumor cells by a p53-independent pathway.

In addition to the known spirangiens A (Ia) and B (Ib), eleven new analogues G‐Q are isolated.

In addition to spirangiens A (1) and B (2), 11 new spirangiens have been isolated from the myxobacterium Sorangium cellulosum (strain So ce90). Spirangiens A and B consist of a polyketide backbone with a spiroketal core, 13 stereocenters, and a conjugated pentaene chromophore with a terminal carboxylic acid. The structures of the derivatives were elucidated on the basis of NMR and MS data. The stereochemistry was determined by 1D and 2D NMR experiments as well as by biosynthetic studies. Four derivatives 6–9 differ in the cis/trans double‐bond geometry of the pentaene chromophore side‐chain and four of them (4–6 and 10) are characterized by the absence of methoxy groups. Derivative 13 does not exhibit the spiroketal core due to reduction of the acetalic carbonyl group. In addition, 11 and 12 exhibit the opposite configuration at C‐20 compared with the known spirangiens A and B. Derivative 3 contains an additional hydroxy group at the end of the side‐chain. All the new spirangien derivatives, as well as spirangiens A and B, were tested in cytotoxicity assays against two different cell lines and showed strong cytotoxic activity. Even though spirangien A was identified as the most cytotoxic compound, other derivatives, such as 3, 4, 6, and 10, showed significantly higher activity against the KB‐3‐1 cell line than against the nontransformed cells (L‐929).