Introduction

Spiro[2.3]hexane consists of a three-membered ring and a four-membered ring sharing a single spiro atom. The two numbers in “[2.3]” indicate the numbers of atoms in the two rings excluding the shared spiro atom. The entire scaffold therefore contains six ring atoms.

These structures possess a high degree of saturation, a well-defined three-dimensional geometry, and restricted conformational flexibility. Replacing carbon atoms with oxygen or nitrogen can further modify molecular polarity, basicity, hydrogen-bonding capacity, and the spatial orientation of substituents. These characteristics give spiro[2.3]hexanes the potential to replace commonly used carbocycles or nitrogen- and oxygen-containing heterocycles.

However, the practical utility of a scaffold in medicinal chemistry depends not only on its structural novelty, but also on whether it can be prepared reliably, rapidly, and as a series of related analogues. A study published in Angewandte Chemie International Edition in 2026 systematically prepared nine classes of spiro[2.3]hexane core structures through reactions of three four-membered-ring-containing sulfonium salts with alkenes, carbonyl compounds, and imines. The study also evaluated the potential of selected scaffolds as bioisosteres.

1 Why Spiro[2.3]hexanes Have Historically Been Underexplored

1.1 Structural value does not necessarily translate into improved drug properties

Increasing the proportion of sp³-hybridized carbon atoms in a molecule generally helps reduce molecular planarity and may improve solubility, conformational properties, and target selectivity. “Escape from flatland” has therefore become an important concept in modern drug design.

However, this represents a statistical trend rather than a universal rule applicable to every molecule. Whether a spirocyclic scaffold improves drug-like properties also depends on the lipophilicity, ionization state, hydrogen-bonding characteristics, metabolic liabilities, and target-binding mode of the complete molecule. The practical value of spiro[2.3]hexane lies in providing a compact and conformationally restricted structural option with substituent vectors that differ from those of conventional aromatic rings and six-membered rings. It does not guarantee higher biological activity or improved pharmacokinetic properties.

1.2 The principal challenge is the difficulty of preparing analogue series

Traditional methods generally begin with a four-membered ring bearing an exocyclic double bond, followed by formation of the three-membered ring through one of the following reactions:

① Cyclopropanation to form carbocyclic spiro[2.3]hexanes;

② Epoxidation to form spirocycles containing an oxirane ring;

③ Aziridination to form spirocycles containing an aziridine ring.

Although these routes can provide the desired structures, the substituents on the three-membered ring are generally already incorporated into the exocyclic alkene precursor. Each change in substitution may therefore require the preparation of a new exocyclic alkene, making the rapid synthesis of a series of structurally related compounds inefficient.

Strain-release reactions of bicyclo[1.1.0]butanes and azabicyclo[1.1.0]butanes have improved synthetic flexibility for certain structures. However, these approaches require the in situ preparation of highly reactive organolithium intermediates and cannot access five of the nine target scaffolds, specifically those containing a cyclopropane three-membered ring or an oxetane four-membered ring.

2 Core Design of the New Method

2.1 Preinstalling the four-membered ring in a sulfonium salt

The researchers designed three sulfonium salts containing:

① A cyclobutane ring;

② An oxetane ring;

③ An N-p-toluenesulfonyl-protected azetidine ring, abbreviated as N-Ts-protected azetidine.

The role of these sulfonium salts is to convert the three types of four-membered rings into isolable, weighable, and storable reagents. Upon addition of a strong base, the sulfonium salts undergo deprotonation to generate four-membered-ring-containing sulfur ylides in situ.

Sulfur ylides are zwitterionic species in which adjacent atoms formally bear negative and positive charges. Their carbanionic carbon center can attack an alkene, carbonyl compound, or imine, followed by intramolecular ring closure and elimination of a sulfide to form a three-membered ring. These transformations belong to the Johnson–Corey–Chaykovsky class of reactions.

This design separates two different requirements:

① The sulfonium salt provides sufficient storage and handling stability;

② The sulfur ylide provides the reactivity required for ring formation during the reaction.

2.2 Preparation of the three sulfonium salts

The researchers first prepared aryl sulfides containing a cyclobutane, oxetane, or azetidine unit. The sulfides were then selectively oxidized to sulfoxides using N-bromosuccinimide, abbreviated as NBS, providing the intermediates in yields of 55%–91%.

The sulfonium triflates initially formed through the interrupted Pummerer reaction were hygroscopic, waxy materials. Following anion exchange with lithium tetrafluoroborate, sulfonium tetrafluoroborates 8–10 could be prepared on a multigram scale and isolated as free-flowing solids that were stable toward air and moisture.

The resulting sulfonium salts could be prepared on a multigram scale and isolated as stable, free-flowing solids. From a practical synthetic perspective, this feature is as important as reaction yield: researchers can prepare the sulfonium salts in advance and directly weigh them out when needed, rather than regenerating difficult-to-store, highly reactive precursors for each reaction.

3 How the “3 × 3” Reaction Design Covers Nine Scaffold Classes

The three sulfonium salts determine the composition of the four-membered ring, whereas the three classes of electrophilic reaction partners determine the composition of the three-membered ring.

Note: The names in the table indicate the heteroatom composition of the core ring systems. In the original study, the nitrogen-containing four-membered-ring sulfonium salt and the corresponding directly obtained products were N-Ts-protected. The corresponding free nitrogen-containing scaffolds or other N-substituted derivatives could be obtained only after deprotection and further derivatization.

The overall reaction relationships can be represented in simplified form as follows:

Four-membered ring—CH—S⁺(Ar¹)(Ar²) · BF₄⁻

                         │

              Deprotonation with a strong base

                         ↓

Four-membered ring—C⁻—S⁺(Ar¹)(Ar²)

                    Sulfur ylide

In the sulfur ylide, the carbon atom bears a formal negative charge, while the sulfur atom bears a formal positive charge, making the molecule overall electrically neutral. This carbon atom is located within the four-membered ring and becomes the spiro atom of the spiro[2.3]hexane in the subsequent reaction.

Sulfur ylide + electron-deficient alkene

→ Cyclopropane-type spiro[2.3]hexane + Ar¹SAr²

Sulfur ylide + aldehyde or ketone

→ Oxirane-type spiro[2.3]hexane + Ar¹SAr²

Sulfur ylide + imine

→ Aziridine-type spiro[2.3]hexane + Ar¹SAr²

Here, Ar¹SAr² represents the neutral diaryl sulfide that is eliminated during ring formation.

The key feature of this design is the clear division of structural variation:

① Changing the sulfonium salt changes whether the four-membered ring contains carbon, oxygen, or nitrogen;

② Changing the class of electrophilic reaction partner changes whether the three-membered ring contains carbon, oxygen, or nitrogen;

③ Changing the substituents on the electrophilic reaction partner modifies the peripheral substitution of the three-membered ring.

The nine scaffold classes therefore no longer require nine independent synthetic routes. Instead, they are obtained through combinations based on the same sulfur-ylide-mediated ring-forming principle.

4 Reaction Conditions and Ring-Formation Mechanism

4.1 The sulfur ylide must be generated in situ

The researchers used the reaction between cyclobutane sulfonium salt 8 and α-(trifluoromethyl)styrene as the model reaction for condition optimization. The preferred conditions were:

① Tetrahydrofuran, abbreviated as THF, as the solvent;

② A reaction temperature of 0 °C;

③ Lithium bis(trimethylsilyl)amide, abbreviated as LiHMDS, as the base;

④ 1.3 equivalents of the sulfonium salt;

⑤ 1.6 equivalents of LiHMDS;

⑥ 1.0 equivalent of the alkene.

The desired product was obtained in 95% yield under these model conditions. When potassium tert-butoxide was used instead, the yield decreased to 62%. With N,N-diisopropylethylamine or sodium hydride, the starting materials were largely recovered.

A key operational feature is that the electrophilic substrate is present in the reaction mixture before the base is added to generate the sulfur ylide. This allows the newly formed ylide to react rapidly with the electrophilic substrate, reducing decomposition and other side reactions.

In the substrate-scope studies, the general conditions typically employed 1.6 equivalents of LiHMDS in THF, with the reaction initiated at 0 °C and then allowed to warm to room temperature for approximately 1 hour. The reagent ratios and operating conditions were adjusted for certain substrates and should be confirmed against the original paper and its Supporting Information.

4.2 Ring formation proceeds through nucleophilic addition and intramolecular substitution

Using an electron-deficient alkene as an example, and based on the general mechanism of the Johnson–Corey–Chaykovsky reaction together with the substrate electronic effects observed in this study, the reasonably proposed ring-forming pathway can be summarized in three steps.

Step 1: Deprotonation of the sulfonium salt to generate the sulfur ylide

A strong base removes a proton from the carbon atom within the four-membered ring that is directly bonded to sulfur, producing a carbanion–sulfonium sulfur ylide:

Four-membered ring—CH—S⁺(Ar¹)(Ar²) · BF₄⁻

                         │ Strong base

                         ↓

Four-membered ring—C⁻—S⁺(Ar¹)(Ar²)

                    Sulfur ylide

After deprotonation, the carbon atom within the four-membered ring is converted from CH to C⁻. This carbon bears a formal negative charge, while the sulfur atom bears a formal positive charge, and the sulfur ylide is overall electrically neutral. This carbon atom subsequently becomes the spiro atom of the spiro[2.3]hexane.

Step 2: Nucleophilic addition of the sulfur ylide to an electron-deficient alkene

The carbanionic carbon center of the sulfur ylide attacks an alkene bearing an electron-withdrawing group. A general activated alkene can be represented as follows:

Sulfur ylide + CH₂=C(R)—EWG

                     ↓

Formation of a new C—C bond and a zwitterionic intermediate

Here, EWG represents an electron-withdrawing group. The sulfur ylide generally attacks the β-carbon of the alkene, while the electrons of the alkene π bond shift toward the α-carbon adjacent to the electron-withdrawing group, generating a new carbanionic center.

Step 3: Intramolecular substitution and ring closure

The newly formed carbanion in the zwitterionic intermediate undergoes intramolecular nucleophilic substitution at the four-membered-ring carbon bearing the sulfonium group. Formation of the new C—C bond is accompanied by elimination of the neutral diaryl sulfide Ar¹SAr², completing closure of the three-membered ring:

Zwitterionic intermediate

          │ Intramolecular nucleophilic substitution

          ↓

Cyclopropane-type spiro[2.3]hexane + Ar¹SAr²

The overall process can therefore be summarized as:

Sulfur-ylide formation

→ Nucleophilic addition to an electron-deficient alkene

→ Formation of a zwitterionic intermediate

→ Intramolecular substitution with elimination of a neutral sulfide

→ Cyclopropane-type spiro[2.3]hexane

Reactions with carbonyl compounds and imines follow a similar Johnson–Corey–Chaykovsky ring-forming logic. The sulfur ylide undergoes nucleophilic addition to the C=O bond of a carbonyl compound or the C=N bond of an imine, generating an oxygen- or nitrogen-containing zwitterionic intermediate. Subsequent intramolecular nucleophilic substitution and elimination of a neutral diaryl sulfide afford oxirane-type or aziridine-type spiro[2.3]hexanes, respectively.

4.3 Alkene electronic properties determine reaction efficiency

Electron-deficient alkenes perform substantially better than electron-rich alkenes. Compatible substrates include α-(trifluoromethyl)styrenes, acrylates, acrylamides, vinyl sulfoxides, and selected styrene derivatives bearing electron-withdrawing groups.

These observations further support a mechanism in which the sulfur ylide first undergoes nucleophilic addition to the alkene, rather than one in which carbene insertion serves as the principal ring-forming pathway. Electron-withdrawing groups reduce the electron density of the alkene and make it more susceptible to attack by the sulfur ylide. Consequently, although the method is applicable to several classes of alkenes, it is not equally effective for all conventional alkenes. Alkenes with insufficient electrophilicity may fail to trap the sulfur ylide before it decomposes.

5 Structural Contributions of the Three Classes of Electrophilic Reaction Partners

5.1 Alkenes provide cyclopropane-type products

Electron-deficient alkenes react with the three sulfonium salts to form spirocycles in which a three-membered carbocycle is connected to a cyclobutane, oxetane, or azetidine ring.

The principal significance of this reaction class is that the substituents on the cyclopropane ring can be adjusted directly by changing the alkene, without the need to resynthesize four-membered-ring precursors bearing different exocyclic double bonds. The study also incorporated a leflunomide-related structural fragment, demonstrating that the reaction can accommodate substrates with medicinal-chemistry-relevant features.

5.2 Aldehydes and ketones provide oxirane-type products

Sulfur ylides react with aldehydes or ketones to form spiro[2.3]hexanes containing an oxirane ring. The study covered a range of aldehydes and ketones and incorporated a fenofibrate-related structure into the products.

The oxirane ring may serve either as the final scaffold or as a functional group for subsequent nucleophilic ring-opening reactions. However, because these products contain both a three-membered ring and a four-membered ring, their chemical stability, storage requirements, and selectivity in subsequent ring-opening reactions must be evaluated for each individual compound.

5.3 Imines provide aziridine-type products

Imines react with the three sulfonium salts to provide aziridine-containing spirocycles, including the previously underreported 5-oxa-1-azaspiro[2.3]hexane and 1,5-diazaspiro[2.3]hexane systems.

The substituent on the imine nitrogen can be readily varied, allowing relatively complex structural fragments to be incorporated into the products. The researchers prepared a spiroaziridine containing a celecoxib-related fragment, as well as a product containing both celecoxib- and fenofibrate-related fragments.

6 What Questions Were Addressed by the Bioisosteric Evaluation?

Bioisosterism is not simply a matter of two-dimensional structural similarity. Whether a new scaffold can replace an existing one must be evaluated by considering molecular volume, shape, polarity, dipole moment, hydrogen-bonding characteristics, substituent orientation, and the role of the scaffold at a specific biological target.

The researchers used three sequential steps:

① Comparison of the physicochemical and three-dimensional properties of the core scaffolds;

② Prediction of the interactions between specific compounds and a selected target;

③ In vitro testing of prioritized compounds.

6.1 Unsupervised clustering generated a piperidine-replacement hypothesis

The researchers compared the nine spiro[2.3]hexane cores with more than 70 heterocycles commonly used in medicinal chemistry. All structures were first subjected to density functional theory, abbreviated as DFT, geometry optimization to obtain energy-optimized three-dimensional conformations and related quantum-chemical properties. DFT-derived results were then combined with RDKit calculations or descriptor extraction to obtain molecular weight, heteroatom count, calculated logP, topological polar surface area, dipole moment, asphericity, and deviation from the plane of best fit, among other descriptors.

Principal component analysis, abbreviated as PCA, was subsequently used to reduce the dimensionality of the dataset, followed by unsupervised clustering with the k-medoids method. Five clusters were obtained. Eight of the spiro[2.3]hexane cores were assigned to the same cluster and occupied regions close to heterocycles such as isoxazole and pyridine.

Piperidine and these spirocycles were not assigned to the same cluster, but they were located relatively close to one another in three-dimensional PCA space. The 5-azaspiro[2.3]hexane core showed similarities to piperidine in properties such as molecular volume and dipole moment, while possessing greater ring strain. On this basis, the researchers proposed that it could potentially serve as a bioisosteric replacement for piperidine.

This result can only be used to propose candidate structures. It does not independently demonstrate that the two scaffolds are interchangeable in all drug molecules. The clustering analysis compares the overall properties of isolated core structures, whereas specific biological activity depends on the structure of the complete molecule.

6.2 The pethidine model demonstrates that the side chain is equally important

The researchers selected pethidine, a μ-opioid receptor agonist containing a 4-substituted piperidine, as a model. The piperidine core was replaced with 5-azaspiro[2.3]hexane, and the ester side chain was systematically varied.

The first two 5-azaspiro[2.3]hexane analogues had predicted pAct values of 6.54 and 5.92, respectively. Replacing the ethyl ester with a methyl or isopropyl ester gave predicted values of 6.50 and 6.48, respectively. Introduction of an azetidinyl ester increased the predicted pAct to 7.22, whereas the corresponding oxetanyl ester analogue did not show comparable predicted activity.

A higher pAct value indicates stronger activity predicted by the model. However, the value of 7.22 corresponds to a complete molecule bearing a particular side chain and is not an intrinsic, fixed property of the 5-azaspiro[2.3]hexane scaffold itself. These results show that scaffold replacement can provide a new spatial arrangement and physicochemical profile, but the ultimate target interaction is jointly determined by the spirocycle, ester group, and side chain.

6.3 In vitro results provided preliminary support

The researchers selected three compounds with relatively high predicted activity and included one spiro[2.3]hexane analogue predicted to be inactive as a control. The compounds were evaluated in human neuroblastoma SH-SY5Y cells expressing the μ-opioid receptor using an EnSpire platform and a label-free optical biosensor-based in vitro binding assay.

The test concentration range was 0.03–150 μmol/L. The authors reported that all four compounds showed micromolar binding activity of approximately 10–39 μmol/L in the label-free assay. The azetidinyl ester analogue with the highest predicted pAct performed best, with a value of approximately 10 μmol/L.

The experiment supports the possibility that 5-azaspiro[2.3]hexane can retain receptor interactions in this specific pethidine–μ-opioid receptor model. However, the results do not justify the following conclusions:

① That the new compounds have been proven to exhibit the same agonist activity as pethidine;

② That 5-azaspiro[2.3]hexane can replace piperidine in all drug molecules;

③ That the new compounds possess better selectivity, safety, or in vivo efficacy;

④ That the scaffold has resolved safety concerns associated with pethidine metabolites.

The control compound predicted to be inactive also produced a measurable effect, indicating that the predictive model may generate false-negative results for novel scaffolds that are poorly represented in the training data. Computational methods can be used to prioritize compounds for testing, but they cannot replace experimental evaluation.

7 Classification and Research Applications of Representative Chemicals Related to Spiro[2.3]hexane Scaffold Construction

Note: The tables below list reaction reagents directly related to the study, candidate substrates of the same general types, and related synthetic building blocks. Unless otherwise specified, inclusion in the tables does not indicate that the particular product or grade has been experimentally validated under the conditions reported in the original paper. Specific applicability, reaction yield, stabilizer treatment, and operating conditions should be determined with reference to the Supporting Information of the original paper and actual experimental results.

Table 1. Reagents for Sulfonium Salt Preparation and Sulfur Ylide Reactions

Table 2. Electrophilic Alkene Substrates

Table 3. Carbonyl Compounds, Imines, and Four-Membered Heterocyclic Building Blocks

Table 4. Drug-Related Structures and Compounds Used to Validate Complex Fragments

Note: The products listed above are representative Aladdin products related to scientific research and formulation studies. Additional product specifications, grades, and Certificates of Analysis may be retrieved from the Aladdin website by searching by product name, CAS number, or catalog number.

References

[1] Natho P, Vicenti A, Mastrolorito F, et al. Novel Sulfonium Reagents for the Modular Synthesis of Spiro[2.3]Hexanes and Heteroatom-Containing Analogues: Synthesis, Application, and Evaluation as Bioisosteres. Angewandte Chemie International Edition, 2026, 65: e21633. DOI: 10.1002/anie.202521633.

[2] Natho P, Colella M, Luisi R. Strained Spiro Heterocycles as Potential Bioisosteres: An Update on the Synthesis of Heteroatom-Containing Spiro[2.3]Hexanes and Spiro[3.3]Heptanes. Chemical Communications, 2025, 61: 6579–6594. DOI: 10.1039/D5CC00656B.

[3] Lovering F, Bikker J, Humblet C. Escape from Flatland: Increasing Saturation as an Approach to Improving Clinical Success. Journal of Medicinal Chemistry, 2009, 52: 6752–6756. DOI: 10.1021/jm901241e.

[4] Meanwell N A. Synopsis of Some Recent Tactical Application of Bioisosteres in Drug Design. Journal of Medicinal Chemistry, 2011, 54: 2529–2591. DOI: 10.1021/jm1013693.

[5] Ammara, Asim Mansha, Samreen Gul Khan, et al. A Review on the Sulfur Ylide-Mediated Corey–Chaykovsky Reaction: A Powerful Approach to Natural Product Synthesis. RSC Advances, 2025, 15: 37125–37151. DOI: 10.1039/D5RA04921K.

For more related articles, see below:

Oxetane: Property-Window Optimization and a Building-Block Selection Guide (Tables 1–4)

Haloheterocycles and Cross-Coupling: A Research-Oriented Selection Framework from Substrate Identification to Bond-Forming Routes (Including Product Navigation and Tables 1–5)

Spirocyclic Building Blocks for Scaffold Assembly

Innovations in the design of stereospecific drug molecular structures: Spirocyclic Scaffolds

Cyclic isomers--Azabicyclic molecular building blocks to aid drug design

One Atom Can Change a Drug’s Fate: Atom-Level Knobs and a Functional-Group Toolbox for Medicinal Chemistry (Methyl / Halogen Bonding / 3D Building Blocks / Late-Stage Fluorination + Product-Selection Tables)

The Role of 7-Membered Nitrogen Heterocycles in Drug Discovery: Microstate Management, Conformational Bias, and Developability Trade-offs (with Research Selection Navigator and Product Tables 1–3)

Pyridinones as “Property Knobs” in Drug Design: From Tautomers and H-Bonding Fingerprints to Product Selection (Tables A–C)

Piperazine Selection Guide: Using a “Six-Membered Diaza Ring” to Make Salt Forms and Linking Strategies More Controllable (Appendix Tables 1–4 Product Navigator)

Why Do We So Often Add a “Morpholine Ring” to Molecules? — Definitions, Structural Features, and a Selection Guide to Morpholine/Thiomorpholine (with Tables 1–4)

Piperidine and Its Derivatives: Controllable Design of Charge, Conformation, and Connecting Exit with Product Navigation (Tables 1–5)