Structure Searching Comes to Metabolomics
A new web tool, StructureMASST, makes public metabolomics data searchable by structure rather than a single reference spectrum
A web-based search tool called StructureMASST has made public metabolomics data searchable by chemical structure, allowing users to trace molecules and substructures across large-scale mass spectrometry repositories. Developed by an international team led by researchers at UC San Diego and UC Riverside, the platform connects chemical names, SMILES strings, or substructure patterns to public tandem mass spectrometry (MS/MS) data and associated sample metadata.
The tool builds on earlier MASST and FASST infrastructure, but shifts the query from individual reference spectra to molecular structures. Its precomputed knowledge base contains more than 1.2 billion MS/MS matches, 1.56 million reference spectra, and 420 million metadata links across repositories including GNPS/MassIVE, MetaboLights, Metabolomics Workbench, and NORMAN/DSFP.
“Search engines allow you to input text and quickly retrieve all the information associated with it because the entire worldwide web has been indexed,” said Pieter C. Dorrestein, who also directs the UC San Diego Collaborative Mass Spectrometry Innovation Center, in the team’s press release. “We do essentially the same thing that these web search engines have done, but for molecules.”
Instead of requiring users to select a single representative spectrum, StructureMASST retrieves available spectra for a molecule or substructure and runs multi-MASST searches across public data. Returned matches can be linked to metadata such as organism, tissue or biofluid, health condition, environment, ion form, and collision energy. The interface also supports exact structure, substructure, modification-tolerant, and blind analog searches.
The resulting searches can connect molecules to biological, clinical, and environmental contexts, from candidate biomarkers to drug exposure and microbial metabolism. “It will tell you what organs it’s found in, which organisms can produce it, what health conditions it’s associated with, and what molecules are connected to it,” Dorrestein added
How Protein Modifications Change Small-Molecule Binding
Chemical proteomics shows that phosphorylation and N-linked glycosylation can reshape small-molecule binding across hundreds of proteins
A chemical proteomics study from The Scripps Research Institute has shown that post-translational modifications can reshape how hundreds of proteins engage small molecules in cells. The work suggests that ligandability is a dynamic property of protein state, rather than a fixed feature of a given target.
The team used broad-spectrum photoaffinity probes to map small-molecule–protein interactions under altered phosphorylation and N-linked glycosylation states. In phosphorylation experiments, MDA-MB-231 cells were treated with staurosporine, while N-linked glycosylation was perturbed in HEK293T cells using tunicamycin. Probe-labeled proteins were enriched and quantified by tandem mass tag–based proteomics, with binding-site mapping used to locate modification-sensitive pockets.
Across the experiments, the researchers identified 235 phosphorylation-dependent and 225 glycosylation-dependent ligandability changes among thousands of probe-enriched proteins. These targets spanned enzymes, transporters, kinases, transcription factors, epigenetic regulators, and other protein classes, with many lacking well-annotated ligands. The changes were largely independent of protein abundance, pointing to altered molecular recognition rather than expression differences.
Binding-site analysis showed that many modification-sensitive changes occurred near functional regions, including active sites, binding pockets, and protein–protein interfaces. The effects were not limited to direct blocking of a pocket: in some cases, altered modification state appeared to change protein interactions or conformation. KRAS provided one drug-relevant example, with phosphorylation changing engagement near the switch-II region and altering how the protein responded to KRAS inhibitors.
For drug discovery, the work highlights the value of profiling targets in native cellular contexts, where modification state, binding partners, and conformation can shape whether a pocket is available to small molecules.
Ancient Proteins from Homo Erectus Teeth
LC–MS/MS analysis of minimally sampled teeth identifies AMBN variants in Middle Pleistocene fossils from China
Ancient enamel proteins recovered by liquid chromatography–tandem mass spectrometry (LC–MS/MS) have provided rare molecular evidence from six Homo erectus teeth in China, showing how paleoproteomics can reach fossils where ancient DNA is unlikely to survive.
Led by researchers at the Institute of Vertebrate Paleontology and Paleoanthropology, the study used minimally invasive acid etching to sample enamel from specimens dated to around 400,000 years ago without visibly damaging tooth morphology. The team analyzed teeth from Zhoukoudian, Hexian, and Sunjiadong, with LC–MS/MS runs split across two laboratories and searched using MaxQuant, PEAKS Online, and pFind. The resulting enamel proteomes contained 650–3,457 peptides per specimen from 6–11 enamel-related proteins, with degradation patterns and elevated deamidation supporting their ancient origin.
The main evolutionary signal came from two amino acid variants in ameloblastin (AMBN), an enamel protein recovered from all six H. erectus specimens. To reduce the risk of false assignments, the team required support from multiple peptides, manual inspection of tandem mass spectra, and confirmation across the three search pipelines.
One variant, AMBN(A253G), had not previously been detected in other human lineages or primates, making it a potential molecular marker for these Middle Pleistocene East Asian H. erectus populations. The second, AMBN(M273V), had been considered Denisovan-associated, but its recovery in all six teeth suggests that the variant may instead trace back to a Middle Pleistocene H. erectus population in East Asia.
The study also points to a wider role for enamel proteomics in hominin evolution, offering a way to compare populations and lineages that remain beyond the reach of ancient DNA.
A Mineral Filter for Dissolved Organic Matter
The study suggests pH-driven mineral sorting can shape whether dissolved carbon becomes microbial fuel or longer-lived environmental carbon
Iron oxide minerals may help determine whether dissolved organic matter becomes microbial food or longer-lived environmental carbon, according to a study of soil-derived organic matter exposed to goethite. Rather than simply reducing dissolved organic matter, goethite adsorption shifted which molecular classes remained in solution for microbial degradation.
The team extracted dissolved organic matter from forest soil, exposed it to goethite at pH 4.5 or 6.5, and incubated the original and mineral-fractionated samples with native soil microbes for 63 days. Molecular and microbial changes were tracked using ultraviolet-visible spectroscopy, fluorescence spectroscopy, Fourier-transform ion cyclotron resonance mass spectrometry (FT-ICR MS), and 16S rRNA gene sequencing.
FT-ICR MS showed that goethite preferentially adsorbed high-molecular-weight, aromatic, and humified compounds, including lignin-like, tannin-like, unsaturated, and condensed aromatic structures. More labile components, including proteins, lipids, carbohydrates, and lower-molecular-weight molecules, were enriched in the remaining solution, with stronger fractionation at lower pH.
That sorting altered biodegradation. Dissolved organic matter fractionated at pH 6.5 showed the greatest cumulative dissolved organic carbon loss, reaching about 63 percent by Day 63. The pH 4.5 fractionated sample degraded more rapidly at first, reaching about 52 percent loss by Day 49, before declining as the readily degradable pool was depleted.
The microbial data pointed to a sequence in substrate use, with protein- and lipid-like compounds consumed first, followed by quinone-like molecules and, later, humic-like substances such as lignins. For iron-rich environments, the results suggest that pH-driven mineral sorting may shape whether dissolved carbon remains mobile, becomes microbial fuel, or moves toward longer-term stabilization.
(Mass) Spectacular and Strange
Stevia’s "Sweet Spot"
Stevia may be sold as a simple “natural sweetener,” but the leaf chemistry behind that sweetness is surprisingly complex. A University of Toyama study suggests that the balance of compounds linked to cleaner, more sugar-like sweetness is shaped not only by sweetness-related genes, but by where those genes are active inside the leaf.
To connect genetic variation with leaf chemistry, the team built a chromosome-scale reference genome for Stevia rebaudiana and combined population genomics with single-nucleus RNA sequencing and imaging mass spectrometry. Imaging mass spectrometry helped map the spatial distribution of steviol glycosides – the intensely sweet diterpenoid compounds behind stevia’s sweetness – while the sequencing data connected those chemical patterns to gene activity.
The analysis highlighted UGT76G glycosyltransferase genes involved in enhancing sweetness, as well as UGT91D4, a gene linked to high-value steviol glycosides such as rebaudioside D and rebaudioside M. UGT91D4 was active only in certain mesophyll and epidermal cells, which may help explain why the most desirable sweet compounds accumulate at relatively low levels.
“Thus, the flavor profile of stevia is determined not just by its genes, but by precisely where those genes are activated,” said Professor Tsubasa Shoji in the team’s press release.
Newsletters
Receive the latest analytical science news, personalities, education, and career development – weekly to your inbox.
