Oxazolidinones occupy a special place in modern medicinal chemistry. The five-membered ring, containing both a nitrogen and an oxygen atom adjacent to a carbonyl, appears in approved antibiotics and in countless molecules moving through drug discovery pipelines. Of particular urgency are the 5-(S)-aminomethyl oxazolidinones, scaffolds central to next-generation antibiotics designed to combat multidrug-resistant and extensively drug-resistant strains of Mycobacterium tuberculosis, the pathogen behind one of the deadliest infectious diseases on Earth. As resistance spreads, the demand for efficient ways to assemble these rings has grown correspondingly sharper.

The trouble, historically, has been chirality. Molecules like oxazolidinones are three-dimensional objects, and their biological activity depends exquisitely on the handedness of their stereocentres. Conventional synthetic routes have leaned on the so-called chiral pool strategy, in which enantiopure amino alcohols harvested from natural sources serve as the key starting intermediates. That approach works, but it constrains chemists to the structural inventory of nature and demands lengthy sequences of functional group manipulations. Many methods exist to set the stereocentre at the 4-position of the ring, alpha to nitrogen, yet strategies for installing the 5-stereocentre, alpha to oxygen, have remained underdeveloped, leaving a stubborn gap in the synthetic toolbox.

The Caltech and Pittsburgh teams closed that gap with a two-stage reaction sequence performed by a single engineered enzyme. The cascade begins with aziridination, a transformation in which a nitrogen atom, delivered as a nitrene, is inserted across a carbon-carbon double bond to form a three-membered aziridine ring. The enzyme then guides a ring expansion of that strained intermediate, and the aziridine rearranges into the five-membered oxazolidinone. The result is a direct, enantioselective synthesis of clinically relevant and discovery-stage oxazolidinones starting from the simplest possible feedstocks: plain alkenes.

The choice of catalyst reflects a larger trend in synthetic biology. Haemproteins, enzymes built around an iron-containing porphyrin cofactor, have emerged in recent years as remarkably tunable platforms for carbene and nitrene transfer chemistry, reactions that no natural enzyme performs natively. Under the directed evolution methods pioneered in Arnold’s laboratory, researchers mutate and screen these proteins iteratively until the active site, originally shaped by evolution for tasks such as oxygen insertion, learn to conduct entirely new chemical transformations with high selectivity. In the new work, mutations introduced through directed evolution proved to be the decisive factor in controlling which mirror-image product the reaction delivers.

What makes the achievement stand out within the biocatalysis community is the class of alkene substrates involved. Until now, haemprotein-catalysed nitrene transfer has been largely restricted to conjugated systems such as styrenes, alkenes whose electronic character makes them reactive and easy to control. Unactivated alkenes, the saturated, electronically inert double bonds that pepper the structures of fats, terpenes, and countless pharmaceutical precursors, have resisted this chemistry. By extending nitrene transfer to these unactivated substrates, the new work substantially broadens the reach of enzymatic nitrene chemistry and opens a much wider swath of chemical space to biocatalytic functionalisation.

Behind the laboratory results lies a computational story. The team carried out detailed computational analysis of the reaction mechanism and found that the key mutations installed during directed evolution are directly responsible for the enantioselective formation of the products. In other words, the protein scaffold does not merely accelerate the reaction; specific amino acid substitutions sculpt the active site geometry so that the aziridination and ring expansion proceed with the precise three-dimensional outcome needed for the drug-like scaffold. This mechanistic understanding, developed jointly with Liu’s computational group at Pittsburgh, illustrates how theory and laboratory evolution now reinforce one another in modern enzyme design.

The practical implications are considerable. Because the cascade starts from simple alkenes and delivers enantioenriched oxazolidinones directly, it offers medicinal chemists a shorter, more modular path to analogues of clinically validated antibiotic scaffolds. Speeding access to structural variants matters enormously in anti-infective research, where teams must explore hundreds of derivative molecules to optimise potency, safety, and pharmacokinetics before a candidate can enter development. A route that removes the dependence on chiral-pool amino alcohols and sets the difficult 5-stereocentre in a single enzymatic operation could meaningfully compress discovery timelines for drugs aimed at resistant tuberculosis and beyond.

The study also adds a chapter to the broader narrative of enzyme engineering as a general-purpose tool for chemistry. Over the past two decades, the Arnold laboratory and others have shown that haemproteins can be reprogrammed to catalyse reactions absent from biology, including cyclopropanation, silicon-carbon bond formation, and a widening repertoire of nitrogen-transfer chemistry. Each extension of this platform challenges the traditional boundary between biological and abiological synthesis. The direct construction of oxazolidinone rings from unactivated alkenes now joins that list, and it does so with an added mechanistic account of how engineered mutations translate into stereochemical control.

For a field racing against the spread of antimicrobial resistance, the work carries both immediate and long-term significance. In the near term, the biocatalytic cascade provides a validated route to the exact scaffolds needed for the next generation of tuberculosis therapeutics. Over the longer term, the demonstration that engineered haemproteins can tame unactivated alkenes in enantioselective nitrene transfer suggests that many other transformations once considered the exclusive province of transition-metal catalysis may fall within reach of programmed biology. As the authors note, the chemistry was peer-reviewed and accepted by Nature, and while the published version is an early-release article subject to further editorial refinement, its conclusions are citable and carry a permanent digital identifier, marking a milestone that synthetic chemists and drug hunters alike will be watching closely.

Subject of Research: Biocatalytic enantioselective synthesis of chiral oxazolidinones from unactivated alkenes using engineered haemproteins

Article Title: Chiral oxazolidinones via biocatalytic aziridination of unactivated alkenes

Article References: Li, Z.-Q., Hanley, D., Zhang, Y., Xie, P.-P., Wu, S. J., Qin, Z.-Y., Zhang, C., Alfonzo, E., Li, F.-Z., Brinkman-Chen, S., Liu, P., & Arnold, F. H. (2026). Chiral oxazolidinones via biocatalytic aziridination of unactivated alkenes. Nature. https://doi.org/10.1038/s41586-026-11169-0

Image Credits: AI Generated

DOI: 10.1038/s41586-026-11169-0

Keywords: biocatalysis, oxazolidinones, aziridination, nitrene transfer, directed evolution, haemproteins, antibiotics, tuberculosis, enantioselectivity, unactivated alkenes, enzyme engineering, drug discovery