The appeal of 3,3-diaryloxetanes stems from their role as bioisosteres of diaryl ketones. Bioisosteric replacement is a cornerstone strategy in drug design: chemists swap one functional group for another that mimics its key electronic and spatial characteristics while improving properties such as solubility, stability or potency. Diaryl ketones, the motif in which two aromatic rings flank a central carbonyl group, appear in numerous marketed pharmaceuticals, including the anti-inflammatory drug ketoprofen, the lipid-lowering agent fenofibrate, the antidiabetic compound adiporon and the antiarrhythmic amiodarone. The polarized carbon–oxygen double bond of the ketone generates a substantial dipole moment, typically around 2.7 to 3.0 Debye, which shapes how the molecule interacts with biological targets through electrostatic contacts and molecular recognition.

Oxetanes reproduce much of this electronic character through an entirely different structural principle. The ring oxygen withdraws electron density through sigma-bond polarization, and the two polarized carbon–oxygen bonds of the strained ring collectively generate a dipole of roughly 1.93 Debye, comparable in effect to the carbonyl dipole. Replacing a flat, planar carbonyl with a tetrahedral oxetane therefore preserves overall polarity while introducing a three-dimensional centre into the molecule, a change that can dramatically alter binding geometry and metabolic fate. Crucially, however, oxetanes and ketones differ fundamentally in reactivity. The carbonyl carbon of a ketone is intrinsically electrophilic because of its low-lying pi-star orbital, whereas oxetanes lack a pi system altogether, and their polarity is distributed across the ring. No carbon on an oxetane behaves like a carbonyl carbon, which means the two motifs can be electronically similar yet chemically orthogonal, an ideal combination for drug design.

Despite this conceptual elegance, synthetic access to 3,3-diaryloxetanes has lagged far behind. The dominant approach, pioneered by James Bull and co-workers, relies on Lewis acid-promoted Friedel–Crafts reactions that proceed through oxetanyl cation intermediates. These reactive intermediates are inherently unstable, and the strongly acidic conditions required restrict the method largely to electron-rich arenes, with competing ring-opening and rearrangement pathways eroding yields and selectivity. Recent iron-catalysed variants have broadened the substrate scope somewhat, but the fundamental limitations of cation chemistry remain. An alternative strategy developed by Phil Baran uses decarboxylative coupling of oxetane-3-carboxylic acid derivatives, generating tertiary benzylic radicals that can engage in cross-coupling. However, these radicals are prone to oxidation to carbocations or reduction to carbanions through radical-polar crossover, and the method demands multistep precursor synthesis and costly starting materials, limiting its practicality and scalability.

The Ritter group’s solution reimagines the problem through the lens of transition-metal catalysis. The chemists hypothesized that converting 3-aryl oxetan-3-ols into electrophiles suitable for nickel-catalysed cross-coupling would provide a modular route, provided they could overcome a formidable obstacle: oxidative addition at the sterically congested carbon–oxygen bond of a fully substituted oxetanyl centre. Such bonds combine steric hindrance with limited intrinsic polarization, disfavouring the oxidative addition step that initiates most cross-coupling cycles. Their answer was to install a tetrafluoropyridyl ether as the leaving group. This electron-deficient aromatic unit polarizes the carbon–oxygen bond and stabilizes the departing tetrafluoropyridone through pi delocalization, striking an optimal balance between reactivity and stability. Unlike strongly electron-withdrawing leaving groups such as triflates, which promote elimination and solvolysis when tertiary benzylic cations form, the tetrafluoropyridyl unit enables controlled activation while its low nucleophilicity and weak basicity suppress ring-opening side reactions.

The practical execution is strikingly simple. In the first step, an aryl bromide is lithiated with n-butyllithium and added to 3-oxetanone, and the resulting alkoxide is trapped in the same pot with pentafluoropyridine to form the oxetanyl tetrafluoropyridyl ether. This one-pot sequence delivered a model substrate in 89 percent yield on a 6.2-gram scale, and the addition of sodium hydride before the trapping step boosted yields by roughly 46 percent. The resulting ethers are bench-stable reagents that can be stored at room temperature for at least six months without detectable decomposition, a critical attribute for a method intended for widespread adoption. Grignard reagents can substitute for the aryllithiums, though the ether formation then works better as a separate step.

The second step is a nickel-catalysed Suzuki-type coupling with arylboronic esters. Using 10 mole percent Ni(cod)2, potassium phosphate as base and isobutanol as an additive in tetrahydrofuran at 30 degrees Celsius, the model coupling afforded the 3,3-diaryloxetane product in 99 percent yield. Control experiments confirmed that every component matters: removing the nickel catalyst or the base eliminated product formation entirely, while omitting the alcohol additive dropped the yield to 73 percent, likely because the alcohol promotes transmetalation of the boronic ester. Notably, palladium catalysts and common nitrogen and phosphine ligands such as dtbbpy and dppf suppressed the reaction, suggesting that strongly coordinating ligands inhibit formation of the reactive nickel species, whereas pi-accepting olefin additives such as 4-methoxystilbene maintained efficiency and may stabilize the catalytically active complex.

The substrate scope demonstrates the modularity that has been missing from this field. Both electron-rich and electron-deficient aryl bromides were converted to the ether precursors in 80 to 95 percent yields on 5-millimole scale, and heteroaromatic motifs including indole, benzothiophene, benzofuran, furan and quinoline were all incorporated. On the coupling side, electron-poor arylboronic esters, heteroaryl boronates and even alkenyl boronic esters performed smoothly, a striking contrast to Friedel–Crafts chemistry, where such substrates are typically unreactive because reactivity there depends on arene nucleophilicity rather than metal-catalysed processes. A methyl ester-bearing product was obtained in 95 percent yield on gram scale. Late-stage diversification succeeded on complex molecules including fenbufen, estrone, tocopherol, flurbiprofen and pyriproxyfen, and the method extended to 3,3-diaryl azetidines, although cyclobutane analogues remain out of reach because their tetrafluoropyridyl ethers are too unstable to isolate.

The synthetic payoff is best illustrated through drug analogues. The team prepared oxetane versions of scaffolds related to ketoprofen, oxybenzone, fenofibrate and a tyrosine kinase inhibitor, and assembled an oxetane analogue of adiporon, a drug candidate for type 2 diabetes, in 65 percent yield after deprotection and amide formation. Most dramatically, a KIT inhibitor intermediate was reached in 20 percent overall yield over two steps from 3-oxetanone, whereas a previous route required four steps and delivered only 0.1 percent overall yield. A second intermediate was obtained in 58 percent yield over two steps compared with 22 percent across four steps previously, and it can be advanced to a complement C1s inhibitor by established procedures.

Mechanistic studies support a conventional but hard-won catalytic cycle. Stoichiometric and catalytic nickel both delivered high yields, and experiments with the one-electron oxidant ferrocenium showed that nickel(I) cannot sustain turnover, implicating nickel(0) as the active species. The radical scavenger BHT had little effect, while TEMPO completely inhibited the reaction with quantitative recovery of starting material, consistent with a polar pathway free of radicals, though TEMPO is also known to oxidize low-valent nickel directly. A Hammett analysis gave a shallow negative reaction constant of minus 0.49, indicating only weak dependence on aryl electronics. Together the data point to a Ni(0)/Ni(II) manifold of oxidative addition, transmetalation and reductive elimination, distinct from the Ni(I)/Ni(III) radical pathways common in modern cross-coupling. The work establishes tetrafluoropyridone as a leaving group worth exploring more broadly and hands medicinal chemists a programmable, two-step entry to benzophenone bioisosteres that were previously among the most difficult scaffolds to access.

Subject of Research: Modular nickel-catalysed synthesis of 3,3-diaryloxetanes as ketone bioisosteres from 3-oxetanone

Article Title: Modular synthesis of 3,3-diaryloxetanes

Article References: Sun, Q., Liu, J., & Ritter, T. (2026). Modular synthesis of 3,3-diaryloxetanes. Nature Synthesis. https://doi.org/10.1038/s44160-026-01183-5

Image Credits: AI Generated

DOI: 10.1038/s44160-026-01183-5

Keywords: oxetanes, bioisosteres, nickel catalysis, cross-coupling, 3-oxetanone, drug discovery, medicinal chemistry, tetrafluoropyridyl ether, oxidative addition, synthetic methodology, diaryl ketones, heterocycles