Four-membered heterocycles occupy a privileged position in drug discovery. Their compact geometry forces substituents into well-defined three-dimensional arrangements, and they can improve the solubility, metabolic stability and binding properties of drug candidates without adding excessive molecular weight. Oxetanes and azetidines, the oxygen- and nitrogen-containing members of this family, have been studied extensively and appear in numerous clinical and marketed compounds. Thietanes, their sulfur-bearing cousins, have received far less attention, largely because reliable methods to make them have lagged behind. Yet the examples that do exist suggest real promise: sulfur analogues of spirocyclic uridines have shown extended antiviral activity, and a thietane-containing ketohexokinase inhibitor was recently advanced as a drug candidate, indicating that swapping oxygen for sulfur can meaningfully change and sometimes improve biological behavior.

The traditional route to a thietane analogue of an oxetane-containing molecule has been de novo synthesis: chemists would have to rebuild the entire molecule from scratch around the new sulfur ring, a laborious process that can take many steps and often fails on complex, functionalized substrates. Skeletal editing offers a fundamentally different philosophy. Instead of constructing molecules ring by ring, it treats an existing molecule as an editable document, deleting, inserting or substituting individual atoms late in the synthesis. This approach has gathered enormous momentum in recent years, with landmark demonstrations of carbon-to-nitrogen transmutations in azaarenes, nitrogen insertion into carbocycles, and photocatalytic oxygen-atom transmutation of oxetanes. The new work extends this logic to a direct oxygen-to-sulfur single-atom swap in a saturated four-membered ring, a transformation that had remained elusive.

The reagent choice is central to the reaction’s practicality. Thiourea is inexpensive, stable, easy to handle and rich in nucleophilic sulfur, making it an attractive sulfur donor compared with the malodorous or hazardous sulfide reagents that often plague sulfur chemistry. In the reported transformation, thiourea participates in a cascade that begins with activation of the oxetane’s strained C-O bonds. The ring then opens in a regioselective fashion, an intermediate rearranges, and an intramolecular cyclization closes the chain back into a four-membered ring, this time containing sulfur in place of oxygen. The net result is a clean atom exchange performed in a single operation, without isolating any of the intermediates along the way.

Combined experimental and computational studies underpin the mechanistic picture. Density functional theory calculations, directed by Yao Fu’s group, mapped the Gibbs free-energy profile of the cascade, showing how the reaction navigates the ring-opening and ring-closing steps and why the pathway favors thietane formation over competing outcomes. Computational analysis of both the ring-opening and ring-closing events helped rationalize the regioselectivity of the process, a critical question because unsymmetrical oxetanes could in principle open at either C-O bond. The interplay between experiment and theory, the authors note, provides a coherent account of this O-to-S skeletal editing process and should guide the design of related atom-swap reactions on other ring systems.

The scope of the reaction is one of its most striking features. The method tolerates a broad range of functional groups and applies across a wide collection of oxetane substrates, including 2-substituted, 3-substituted and spirocyclic variants. Particularly compelling is the demonstration that readily available oxetane building blocks, many of them commercial reagents used routinely in medicinal chemistry programs, can be converted in one step into thietanes that would otherwise require multivariate synthesis. This converts the extensive commercial inventory of oxetanes into a ready-made feedstock for thietane exploration, dramatically lowering the barrier to comparing oxygen and sulfur analogues side by side in biological assays.

Crucially, the chemistry is not confined to simple model systems. The team showed that the reaction works in late-stage settings on structurally complex, drug-like molecules, editing pharmaceutically relevant oxetanes into thietanes without disturbing the rest of the molecule. This late-stage compatibility is the acid test for any skeletal editing method, because real drug candidates carry a dense array of functional groups, heterocycles and stereocenters that a new reaction must leave untouched. The broad functional group tolerance reported here suggests that medicinal chemists could take an advanced oxetane-containing lead and rapidly generate its thietane counterpart, along with analogues, to probe how the atom swap affects potency, selectivity and pharmacokinetic properties.

The significance for drug discovery goes beyond convenience. Replacing oxygen with sulfur changes a molecule’s physicochemical profile in several ways at once: sulfur is larger and more polarizable, the C-S bonds are longer, the ring geometry shifts subtly, and the sulfur atom introduces different hydrogen-bonding and metabolic liabilities and opportunities. These changes can alter how a compound binds to its protein target, how it partitions into membranes and how it is cleared by the liver. Because thietanes remain underexplored relative to oxetanes and azetidines, systematic access to thietane analogues of known oxetane-containing compounds gives researchers a new axis of exploration in lead optimization, one that previously demanded a full resynthesis campaign for every comparison.

The work also fits into a broader movement sometimes described as single-atom logic for heterocycle editing. Recent years have seen oxygen-to-nitrogen swaps in tetrahydrofurans, oxygen-to-sulfur substitutions in oxazoles and isoxazoles, oxygen-atom replacement in non-strained cyclic ethers, and photocatalytic transmutation reactions of oxetanes, each expanding the repertoire of editable transformations. The new oxetane-to-thietane reaction adds a saturated, highly strained four-membered ring to this growing list and demonstrates that even ring strain, which typically drives oxetanes toward destructive ring-opening reactions, can be harnessed productively within an editing cascade. The authors’ mechanistic framework, connecting activation, regioselective opening, rearrangement and cyclization, offers a template that other laboratories can adapt to different heteroatoms and ring sizes.

Challenges naturally remain. The reaction, like much of the skeletal editing field, is at an early stage, and questions of scale, cost on process chemistry scale, and performance across the full diversity of drug-like scaffolds will require further study. The work was supported by the University of Science and Technology of China and the National Natural Science Foundation of China, and parts of the computational and automated experimentation were performed on the Chinese Academy of Sciences’ robotic AI-Scientist platform, reflecting the increasing integration of automation and computation in reaction discovery. Even so, the conceptual achievement is clear: a molecule’s heteroatom, the element that most directly shapes its biological personality, can now be exchanged in a strained four-membered ring in a single step. For a field racing to explore underexplored chemical space with limited synthetic resources, that is a genuinely enabling capability, and it signals that the era of treating drug molecules as editable frameworks rather than fixed structures is accelerating rapidly.

Subject of Research: One-step skeletal editing conversion of oxetanes to thietanes using thiourea

Article Title: One-step skeletal editing of oxetanes to thietanes using thiourea

Article References: Luo, L., Liu, D., Tian, D., Xiao, X., Chen, G., Zhang, L., Wang, J., Wang, Y., Wang, X., Fu, Y., & Zhang, H.-J. (2026). One-step skeletal editing of oxetanes to thietanes using thiourea. Nature Synthesis. https://doi.org/10.1038/s44160-026-01189-z

Image Credits: AI Generated

DOI: 10.1038/s44160-026-01189-z

Keywords: skeletal editing, oxetanes, thietanes, thiourea, heterocycles, medicinal chemistry, organic synthesis, atom swap, drug discovery, reaction mechanism, DFT calculations, four-membered rings

The story begins with the molecule itself. 5-Azacytidine is a mimic of cytidine, one of the four canonical letters of RNA, but with a crucial difference: the carbon at position 5 of the pyrimidine ring is replaced by nitrogen, converting the base into 5-azacytosine, a member of the 1,3,5-triazine family. This single atom swap is what gives the drug its therapeutic power. Once incorporated into RNA, the analogue disrupts RNA metabolism and traps methyltransferase enzymes, ultimately reactivating tumor suppressor genes that cancer cells had silenced. Approved by the FDA in 2004 under the brand name Vidaza, the compound has become a mainstay of treatment for patients with myelodysplastic syndromes.

Because the drug was first made synthetically by Czech chemists in 1964 and only isolated from microbial cultures two years later, most researchers assumed its biosynthesis would never be found, or simply never thought to look. The prevailing view treated 5-azacytidine as a synthetic analogue that happened to also occur in nature, rather than as the product of an evolved metabolic pathway. That assumption, the new work shows, hid one of the most unusual enzymatic cascades yet discovered in natural product chemistry.

A team led by Yasushi Ogasawara and Tohru Dairi of Hokkaido University, together with Hiroyuki Morita of the University of Toyama and collaborators across Japan and Taiwan, set out to find the gene cluster. Through comparative genomics and biochemical screening of the producing organism, they pinpointed a set of genes designated azc that is both necessary and sufficient for assembling the triazine nucleobase. The identification immediately posed a puzzle: how does a cell convert ordinary purine and pyrimidine building blocks into a ring system with an extra nitrogen in an arrangement that is otherwise essentially unknown among natural nucleosides?

The answer begins with AzcE, an enzyme the researchers identify as a guanosine triphosphate cyclohydrolase. GTP cyclohydrolases are best known for opening the purine ring of GTP during the biosynthesis of folate and riboflavin, but AzcE repurposes this chemistry to generate 2,5,6-triaminopyrimidin-4(1H)-one, a pyrimidine intermediate carrying three amino groups. Structural analysis of AzcE, determined in complex with zinc and its product, revealed how the enzyme positions the substrate for ring opening, providing the first stepping stone on the path from a standard nucleotide precursor toward the triazine scaffold of the drug.

The centerpiece of the pathway, and of the study, is AzcA, a cupin domain-containing enzyme that performs what the authors describe as a skeletal editing reaction. Rather than building the triazine ring from scratch, AzcA takes the pyrimidine delivered by AzcE and surgically reworks it: the enzyme selectively cleaves the carbon-nitrogen framework of the heterocycle and reassembles it into 6-amino-4-oxo-1,4-dihydro-1,3,5-triazine-2-carboxylic acid. In effect, the protein performs molecular surgery on a ring, cutting specific bonds and stitching the fragments back together with an additional nitrogen inserted into the skeleton. X-ray crystal structures of AzcA bound to manganese and to reaction intermediates, combined with biochemical assays and density functional theory calculations, allowed the team to trace the mechanism in atomic detail, showing how the metal center and active-site residues orchestrate bond cleavage and reformation with remarkable selectivity.

The final step is equally unconventional. The carboxylic acid that AzcA installs must be removed to yield 5-azacytosine, and the enzyme pair AzcB/C accomplishes this through a thiamine pyrophosphate-dependent decarboxylation acting on the alpha-imino carboxylic acid moiety of the AzcA product. Thiamine-dependent enzymes typically handle alpha-keto acids, so acting on an imino acid represents an atypical use of this cofactor class. Cryo-electron microscopy structures of AzcB/C captured with thiamine pyrophosphate and substrate analogues revealed the active-site architecture that stabilizes the reactive intermediates, completing the mechanistic picture of how the triazine base is finished and presumably glycosylated to give the mature nucleoside antibiotic.

What makes the discovery resonate beyond one molecule is the concept of enzymatic skeletal editing itself. Synthetic chemists have in recent years celebrated skeletal editing, the late-stage insertion, deletion, or transmutation of single atoms within a molecular framework, as a frontier strategy for drug discovery. The demonstration that a cupin enzyme performs precisely this kind of single-atom logic on a heterocyclic ring during natural biosynthesis shows that nature arrived at the same idea long ago, and suggests that related enzymes scattered across bacterial genomes may perform similar transformations on other scaffolds. Genome mining guided by the azc gene cluster could therefore uncover new triazine and related azine natural products, or inspire engineered enzymes for late-stage functionalization of existing drugs.

There are also practical implications for the drug itself. Current industrial production of 5-azacytidine relies on multi-step chemical synthesis, and a biosynthetic route defined by just a handful of enzymes offers a potential platform for biocatalytic or fermentative manufacturing, possibly enabling access to analogues that are difficult to make chemically. The work also resolves a long-standing oddity in natural product history: a molecule used daily in oncology wards worldwide turns out to be a genuine bacterial metabolite, forged by enzymes that edit the very skeletons of nucleobases. For a drug whose clinical value stems from impersonating DNA’s own letters, it is fitting that its maker turns out to be a master of rewriting them.

Subject of Research: Enzymatic biosynthesis of the anticancer nucleoside analogue 5-azacytidine

Article Title: Enzymatic skeletal editing reaction forming the 1,3,5-triazine core during biosynthesis of the anticancer nucleoside analogue 5-azacytidine

Article References: Umezawa, A., Tsunoda, T., Nakashima, Y., Wang, Y.-L., Yoneda, T., Maruyama, C., Hsu, N.-S., Matsuyama, T., Miyamoto, K., Sato, H., Shimose, A., Takazawa, R., Hamano, Y., Li, T.-L., Inokuma, Y., Uchiyama, M., Ogasawara, Y., Morita, H., & Dairi, T. (2026). Enzymatic skeletal editing reaction forming the 1,3,5-triazine core during biosynthesis of the anticancer nucleoside analogue 5-azacytidine. Nature Catalysis. https://doi.org/10.1038/s41929-026-01611-x

Image Credits: AI Generated

DOI: 10.1038/s41929-026-01611-x

Keywords: 5-azacytidine, biosynthesis, natural products, enzymes, skeletal editing, 1,3,5-triazine, anticancer drugs, myelodysplastic syndrome, GTP cyclohydrolase, cupin enzyme, thiamine pyrophosphate, enzyme mechanisms