Metabolomics experiments can employ non-targeted tandem mass spectrometry to detect hundreds to thousands of molecules in a biological sample. Structural annotation of molecules is typically carried out by searching their fragmentation spectra in spectral libraries or, recently, in structure databases. Annotations are limited to structures present in the library or database employed, prohibiting a thorough utilization of the experimental data. We present a computational tool for systematic compound class annotation: CANOPUS uses a deep neural network to predict 1,270 compound classes from fragmentation spectra, and explicitly targets compounds where neither spectral nor structural reference data are available. CANOPUS even predicts classes for which no MS/MS training data are available. We demonstrate the broad utility of CANOPUS by investigating the effect of the microbial colonization in the digestive system in mice, and through analysis of the chemodiversity of different Euphorbia plants; both uniquely revealing biological insights at the compound class level.

G protein-independent, arrestin-dependent signaling is a paradigm that broadens the signaling scope of G protein-coupled receptors (GPCRs) beyond G proteins for numerous biological processes. However, arrestin signaling in the collective absence of functional G proteins has never been demonstrated. Here we achieve a state of “zero functional G” at the cellular level using HEK293 cells depleted by CRISPR/Cas9 technology of the Gs/q/12 families of Gα proteins, along with pertussis toxin-mediated inactivation of Gi/o. Together with HEK293 cells lacking β-arrestins (“zero arrestin”), we systematically dissect G protein- from arrestin-driven signaling outcomes for a broad set of GPCRs. We use biochemical, biophysical, label-free whole-cell biosensing and ERK phosphorylation to identify four salient features for all receptors at “zero functional G”: arrestin recruitment and internalization, but—unexpectedly—complete failure to activate ERK and whole-cell responses. These findings change our understanding of how GPCRs function and in particular of how they activate ERK1/2.

Computational approaches such as genome and metabolome mining are becoming essential to natural products (NPs) research. Consequently, a need exists for an automated structure-type classification system to handle the massive amounts of data appearing for NP structures. An ideal semantic ontology for the classification of NPs should go beyond the simple presence/ absence of chemical substructures, but also include the taxonomy of the producing organism, the nature of the biosynthetic pathway, and/or their biological properties. Thus, a holistic and automatic NP classification framework could have considerable value to comprehensively navigate the relatedness of NPs, and especially so when analyzing large numbers of NPs. Here, we introduce NPClassifier, a deep-learning tool for the automated structural classification of NPs from their counted Morgan fingerprints. NPClassifier is expected to accelerate and enhance NP discovery by linking NP structures to their underlying properties.

This report describes the first application of the novel NMR-based machine learning tool “Small Molecule Accurate Recognition Technology” (SMART 2.0) for mixture analysis and subsequent accelerated discovery and characterization of new natural products. The concept was applied to the extract of a filamentous marine cyanobacterium known to be a prolific producer of cytotoxic natural products. This environmental Symploca extract was roughly fractionated, and then prioritized and guided by cancer cell cytotoxicity, NMR-based SMART 2.0, and MS2-based molecular networking. This led to the isolation and rapid identification of a new chimeric swinholide-like macrolide, symplocolide A, as well as the annotation of swinholide A, samholides A–I, and several new derivatives. The planar structure of symplocolide A was confirmed to be a structural hybrid between swinholide A and luminaolide B by 1D/2D NMR and LC-MS2 analysis. A second example applies SMART 2.0 to the characterization of structurally novel cyclic peptides, and compares this approach to the recently appearing “atomic sort” method. This study exemplifies the revolutionary potential of combined traditional and deep learning-assisted analytical approaches to overcome longstanding challenges in natural products drug discovery.

The cyclic depsipeptide FR900359 (FR), isolated from the tropical plant Ardisia crenata, is a strong and selective inhibitor of Gq proteins, making it an indispensable pharmacological tool to study Gq-related processes, as well as a promising drug candidate. Gq inhibition is a novel mode of action for defense chemicals and crucial for the ecological function of FR, as shown by in vivo experiments in mice, its affinity to insect Gq proteins, and insect toxicity studies. The uncultured endosymbiont of A. crenata was sequenced, revealing the FR nonribosomal peptide synthetase (frs) gene cluster. We here provide a detailed model of FR biosynthesis, supported by in vitro enzymatic and bioinformatic studies, and the novel analogue AC-1, which demonstrates the flexibility of the FR starter condensation domains. Finally, expression of the frs genes in E. coli led to heterologous FR production in a cultivable, bacterial host for the first time.

Gallinamide A, originally isolated with a modest antimalarial activity, was subsequently reisolated and characterized as a potent, selective, and irreversible inhibitor of the human cysteine protease cathepsin L. Molecular docking identified potential modifications to improve binding, which were synthesized as a suite of analogs. Resultingly, this current study produced the most potent gallinamide analog yet tested against cathepsin L (10, K i = 0.0937 ± 0.01 nM and k inact /K i = 8 730 000). From a protein structure and substrate preference perspective, cruzain, an essential Trypanosoma cruzi cysteine protease, is highly homologous. Our investigations revealed that gallinamide and its analogs potently inhibit cruzain and are exquisitely toxic toward T. cruzi in the intracellular amastigote stage. The most active compound, 5, had an IC 50 = 5.1 ± 1.4 nM, but was relatively inactive to both the epimastigote (insect stage) and the host cell, and thus represents a new candidate for the treatment of Chagas disease.

A majority of Ardisia species harbour Burkholderia sp. bacteria within specialized leaf nodules. The bacteria are transmitted hereditarily and have not yet been cultured outside of their host. Because the plants cannot develop beyond the seedling stage without their symbionts, the symbiosis is considered obligatory. We sequenced for the first time the genome of Candidatus Burkholderia crenata (Ca. B. crenata), the leaf nodule symbiont of Ardisia crenata. The genome of Ca. B. crenata is the smallest Burkholderia genome to date. It contains a large amount of insertion sequences and pseudogenes and displays features consistent with reductive genome evolution. The genome does not encode functions commonly associated with plant symbioses such as nitrogen fixation and plant hormone metabolism. However, we identified unique genes with a predicted role in secondary metabolism in the genome of Ca. B. crenata. Specifically, we provide evidence that the bacterial symbionts are responsible for the synthesis of compound FR900359, a cyclic depsipeptide with biomedical properties previously isolated from leaves of A. crenata.

Somatic gain-of-function mutations of GNAQ and GNA11, which encode α subunits of heterotrimeric Gαq/11 proteins, occur in about 85% of cases of uveal melanoma (UM), the most common cancer of the adult eye. Molecular therapies to directly target these oncoproteins are lacking, and current treatment options rely on radiation, surgery, or inhibition of effector molecules downstream of these G proteins. A hallmark feature of oncogenic Gαq/11 proteins is their reduced intrinsic rate of hydrolysis of guanosine triphosphate (GTP), which results in their accumulation in the GTP-bound, active state. Here, we report that the cyclic depsipeptide FR900359 (FR) directly interacted with GTPase-deficient Gαq/11 proteins and preferentially inhibited mitogenic ERK signaling rather than canonical phospholipase Cβ (PLCβ) signaling driven by these oncogenes. Thereby, FR suppressed the proliferation of melanoma cells in culture and inhibited the growth of Gαq-driven UM mouse xenografts in vivo. In contrast, FR did not affect tumor growth when xenografts carried mutated B-RafV600E as the oncogenic driver. Because FR enabled suppression of malignant traits in cancer cells that are driven by activating mutations at codon 209 in Gαq/11 proteins, we envision that similar approaches could be taken to blunt the signaling of non-Gαq/11 G proteins.