Chemists Build a Modular Platform That Sorts the Metabolome by Reactivity

Every cell is a bustling chemical marketplace, and its wares are metabolites: the sugars, fatty acids, amino acids, and countless other small molecules that serve as fuel, signaling messengers, and building blocks. Reading this molecular marketplace accurately is one of the central challenges of modern biology, because metabolites differ wildly in charge, polarity, stability, and abundance. A new study published in Advanced Science introduces a cleverly engineered answer to that challenge: a modular reactivity-encoding platform, or MREP, that sorts the chaotic metabolome into chemically coherent layers before analysis, dramatically expanding what mass spectrometers can see.

The problem the researchers set out to solve is well known to anyone who runs liquid chromatography-mass spectrometry, the workhorse technique of metabolomics. Conventional workflows interrogate a biological sample in a largely undifferentiated way, and the instrument systematically favors molecules that ionize easily. Short-chain fatty acids such as acetic acid and butyric acid, for example, are so poorly ionized and so poorly retained on standard chromatography columns that they often go completely undetected, even when present at biologically meaningful concentrations. The result is a biased, incomplete picture of cellular chemistry, with large regions of chemical space effectively invisible.

Chemists have long tried to fix this with derivatization, the practice of chemically modifying metabolites to make them more detectable. By targeting specific functional groups such as carboxyl, carbonyl, amine, and thiol motifs, derivatization can carve the metabolome into submetabolomes, each with improved ionization and characteristic fragmentation behavior. But most existing approaches label only a single metabolite class at a time, and solid-phase enrichment methods typically require bespoke, custom-synthesized resins tailored to each reactivity type. What has been missing is a unified, modular architecture that can handle many functional groups within a single framework.

MREP fills that gap with a two-step design. In the encoding step, four rationally designed alkyne-tagged probes selectively derivatize their targets: but-3-ynylamine (BYA) labels carboxyl-containing metabolites via carbodiimide-activated acylation, 2,5-dioxopyrrolidin-1-yl 4-ethynylbenzoate (DEB) tags amines, alkyne hydrazide (AHZ) captures carbonyls through nucleophilic addition, and N-propargylmaleimide (NPM) conjugates thiols via Michael addition. Each probe installs a terminal alkyne handle onto the metabolite while preserving its native scaffold. In the capture step, the encoded metabolites are immobilized by copper-catalyzed azide-alkyne cycloaddition, the classic click reaction, onto a single azide-functionalized resin dubbed ACER, for alkyne capture for enrichment and reporter-ion installation.

The ACER resin is where much of the platform’s power resides. After click capture, the resin is washed extensively, physically separating derivatized metabolites from the complex biological matrix, and the bound products are then cleaved under acidic conditions for LC-MS analysis. The resin exhibited a loading capacity of 7.9 nanomoles per milligram, roughly ten times higher than a previously reported chemoselective capture system. Crucially, the same resin works for all four probe types, so researchers no longer need to synthesize a new functionalized support for every metabolite class they wish to interrogate. The embedded valeric amide module also produces a diagnostic reporter ion at m/z 100.0757 in every fragmentation spectrum, a universal signature that flags which features belong to the encoded submetabolomes.

Optimization was a meticulous affair. The team screened amine- and hydrazide-based probes, tested three families of condensation reagents, and varied alkyl linker lengths from three to eight carbons, ultimately selecting the C4 linker for its balance of reactivity and solubility, since the longer C8 variant showed reduced performance toward long-chain fatty acids, likely due to hydrophobic aggregation. Reaction conditions, including probe concentration, temperature, solvent composition, copper catalyst, ligand, and reducing agent, were tuned to work uniformly across all four functional-group channels. Under the final protocol, encoding efficiency exceeded 83.8 percent, capture efficiency exceeded 96.0 percent, and cleavage efficiency exceeded 88.3 percent across a diverse panel of test metabolites.

The sensitivity gains are striking. In their native forms, acetic acid, butyric acid, and succinic acid were undetectable by conventional LC-MS, while phenylalanine, palmitic acid, oleic acid, and eicosapentaenoic acid produced only weak signals. After MREP treatment, all ten carboxyl standards were readily detected in positive ion mode, and across the full panel of carboxyl, amine, carbonyl, and thiol derivatives, limits of detection improved by factors of 20 to 40,000 relative to native analysis. Calibration curves showed excellent linearity over a 1 to 1,000 nanomolar range, and a bifunctional test compound containing both carboxyl and carbonyl groups was quantified accurately through either encoding channel, confirming the platform’s quantitative robustness.

Beyond raw sensitivity, the derivatization imposed a welcome structural order. The triazole-containing derivatives fragmented reproducibly, with the universal reporter ion supplemented by class-specific signatures such as a diagnostic fragment at m/z 237.1352 for carbonyl derivatives and a characteristic neutral loss of 45.0215 daltons for amines. Chromatographic performance also improved, allowing clear separation of positional, epimeric, and hybrid isomers that coeluted in their native forms. To support confident identification, the researchers built an in-house reference library of 115 derivatized standards spanning fatty acids, eicosanoids, bile acids, amino acids, and TCA cycle intermediates, and implemented a local linear regression calibration using nineteen saturated fatty acid calibrants, reducing retention-time deviations to under 5 percent across different gradients, columns, and instruments.

The platform’s real-world credentials were established in a mouse model of alcohol-associated liver disease. Applying MREP to liver tissue and serum from ethanol-fed and pair-fed C57BL/6J mice, the team detected 7,208 putative metabolite features, of which 1,573 in liver and 656 in serum were structurally annotated, with 67 validated against the reference library. Multivariate analysis revealed clear metabolic separation between the dietary groups, and among 191 significantly altered metabolites, the changes tracked known biology: elevated short-chain, monounsaturated, and polyunsaturated fatty acids, increased eicosanoids such as 12,13-DiHOME, raised aromatic amino acids, and decreased pyruvic acid. Notably, several saturated fatty acids and bile acids were detected only under this new framework, and many of the altered metabolites are linked to gut microbial metabolism, consistent with disruption of the gut-liver axis in alcohol-related liver injury.

The authors are candid about limitations: acidic cleavage could in principle affect labile metabolites, though testing showed no significant degradation of the derivatized carboxyl compounds, and densely functionalized molecules may show some cross-reactivity that requires tailored conditions. Even so, the conceptual advance is considerable. By treating chemical encoding not merely as a sensitivity booster but as a structural organizing principle, MREP turns the metabolome’s heterogeneity from an obstacle into a handle. Because the architecture is modular, new probes can extend coverage to additional functional groups through rational design, offering a scalable route toward functionally stratified, high-definition maps of metabolism in health and disease.

Subject of Research: A chemoselective metabolomics platform that stratifies metabolites by functional group reactivity for enhanced LC-MS profiling

Article Title: Chemoselective Metabolomics via a Modular Reactivity‐Encoding Platform

Article References: Tao, X., Zhang, C.-M., Yang, R.-J., Ren, M., Zhong, Z.-L., Wang, K.-H., Hu, R.-X., Liu, J.-Y., Zhou, J.-Y., Li, H.-K., & Wan, J.-B. (2026). Chemoselective Metabolomics via a Modular Reactivity‐Encoding Platform. Advanced Science, 13(56), Article e76599. https://doi.org/10.1002/advs.76599

Image Credits: AI Generated

DOI: 10.1002/advs.76599

Keywords: metabolomics, mass spectrometry, chemical derivatization, click chemistry, LC-MS, submetabolome, fatty acids, bile acids, alcoholic liver disease, solid-phase enrichment, reporter ion, functional group labeling