SmI₂‐Catalyzed Coupling of Alkyl Housane Ketones and Alkenes in an Approach to Norbornanes

Abstract

Cross‐coupling strategies involving strain release have gained significant recent attention for the construction of complex molecular frameworks, particularly in the context of preparing bioisosteres for medicinal chemistry. While the reactivity of cyclopropanes and bicyclo[1.1.0]butanes (BCBs) has been extensively studied, higher homologues are emerging as valuable substrates for synthesis. For example, methods for the fragmentation and coupling of bicyclo[2.1.0]pentane, or housane, ketones show promise but are currently limited in substrate scope. Here, we describe a mild, atom‐economical, samarium diiodide (Sml₂)‐catalyzed fragmentation and coupling of alkyl and aryl housane ketones with alkenes that grants access to functionalized norbornane structural motifs, not easily accessible by classical cycloaddition approaches, and with considerable potential for further manipulation.

Samarium diiodide (Sml₂) is an effective catalyst for the mild, atom‐economical coupling of bicyclo[2.1.0]pentane, or housane, ketones, and electron‐deficient alkenes to give functionalized norbornane structural motifs not easily accessible by classical cycloaddition approaches. In addition to aryl‐substituted housane ketones, harder‐to‐reduce alkyl housane ketones are also excellent substrates for the radical cross‐coupling.

Introduction

Synthetic methods that exploit the strain energy of small rings to drive the construction of complex molecular frameworks have attracted intense recent interest.^{[ 1} , ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ^{12 ]} In particular, strained cyclopropanes and bicyclo[1.1.0]butanes (BCBs) have captured the attention of synthetic chemists due to their unique reactivity and the potential for generating reactive intermediates through fragmentation reactions (Figure 1a);^{[ 7} , ⁸ , ⁹ , ¹⁰ , ¹¹ , ^{12 ]} over 700 reports focusing on the opening of these ring systems have been published in the last 5 years. Recent advances in the catalysis of ring‐opening reactions of strained cycles, such as cyclopropanes and BCBs, have utilized metal catalysts and organocatalysts and embraced new mechanistic manifolds.^{[ 1} , ² , ³ , ⁴ , ⁵ , ^{6 ]} These additions to the synthetic toolbox are often selective, operate under mild conditions, and can be fine‐tuned to control reaction outcome. Thus, substrates containing strained ring systems are now indispensable intermediates in synthesis, allowing efficient access to an expanding range of functionalized molecular scaffolds (Figure 1a). The further development of efficient catalytic processes, driven by the rupture of strained rings, promises very different routes to new and established, important compound classes, thus accelerating discovery programs underpinned by synthesis and screening.

Figure 1.

a) Selected products of strain‐release approaches using cyclopropanes and [1.1.0]‐bicyclobutanes (BCBs). b) Bioactive compounds featuring the norbornane ring system. c) Photocatalytic intermolecular couplings of housane ketones. d) Our strategy to access norbornane derivatives by SmI₂‐catalyzed intermolecular coupling of housane ketones and alkenes. Gdm = guanidinium; PG = protecting group; EnT = energy transfer.

The norbornane (bicyclo[2.2.1]heptane) ring system is a classical molecular architecture,^{[ 13} , ¹⁴ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ^{19 ]} now gaining attention in medicinal chemistry due to its rigidity and three‐dimensionality; norbornanes are conformationally locked and present their substituents in a defined manner, thus making them valuable scaffolds for designing bioactive molecules.

For example, norbornane‐based cationic antimicrobial peptides have shown potent antibacterial activity for methicillin‐resistant Staphylococcus aureus (MRSA),^{[ 18 ]} and the natural product nanolobatolide has promising neuroprotective properties (Figure 1b).^{[ 19 ]} Traditional approaches to the norbornane ring system are dominated by Diels–Alder cycloadditions, followed by reduction of the residual alkene, and often require forcing reaction conditions.^{[ 20} , ^{21 ]} Hence, complementary access to the norbornane ring system, through the catalytic ring‐opening reactions of strained ring precursors, would represent a valuable new synthetic strategy, delivering new, differently substituted norbornane architectures that would be otherwise difficult to obtain through traditional Diels–Alder approaches due to electronic mismatching.^{[ 22} , ^{23 ]}

The housane (bicyclo[2.1.0]pentane) ring system possesses strain energy higher than that of a cyclopropane ring and similar to that of the BCB system.^{[ 24} , ²⁵ , ^{26 ]} Thus, housanes are attractive potential substrates for strain‐release strategies. However, there are currently few synthetic approaches that exploit housane substrates.^{[ 27 ]} Of particular note, Brown and coworkers recently exploited energy transfer (EnT) catalysis to fragment naphthyl housane ketones and generate aza‐norbornane products through couplings with imines (Figure 1c).^{[ 28 ]} With a focus on constructing aza‐norbornanes, couplings with alkenes involved aza‐housanes, with the presence of a naphthyl ring system essential for efficient EnT.

Here, we describe a mild, atom‐economical, samarium diiodide (Sml₂)‐catalyzed fragmentation and coupling of alkyl and aryl housane ketones with alkenes that enables access to functionalized norbornane structural motifs with considerable potential for further manipulation (Figure 1d). The coupling proceeds through a radical relay mechanism triggered by single‐electron transfer (SET) from the SmI₂ catalyst.^{[ 29} , ³⁰ , ³¹ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ^{44 ]} Of note, the use of Sml₂ as a catalyst lends the approach significant substrate scope, in that typically unreactive alkyl housane ketones are excellent substrates.

Results and Discussion

Reaction Optimization and Tolerance

We began by examining the coupling of housane ketone 1a and acrylonitrile, using SmI₂ (25 mol%), at −10 °C (Table 1, entry 1). Pleasingly, the targeted norbornane 2a was formed in quantitative yield with modest diastereocontrol (2.7:1, exo/endo). Further screening of conditions showed that a lower catalyst loading (15 mol%) could be employed in this system without a drop in yield (entry 2). Three‐day‐old SmI₂ performed reasonably well under the reaction conditions, with only a moderate decrease in yield observed (entry 3). Control experiments (entries 4–7), highlighted the essential role of Sm(II), as removal or replacement of the SmI₂ catalyst with SmI₃ failed to lead to product formation. The decrease in the amount of recovered starting material observed in the presence of Sm(III) suggests that a Lewis acid‐mediated decomposition process is out‐competed by the desired coupling.

Table 1.

Reaction optimization and control experiments.

Entry	Deviation from above	1H NMR yield (%)
1	None	>99 (2.7:1 d.r.)
2	SmI2 (15 mol%)	>99 (2.8:1 d.r.)
3	3 days old SmI2 (15 mol%)	61 (2.6:1 d.r.)
4	SmI2 (15 mol%) + no acrylonitrile	n.d. (50% rsm)
5	No SmI2	n.d. (>99% rsm)
6	SmI3 (15 mol%) instead of SmI2	n.d. (80% rsm)
7	SmI3 (15 mol%) + no acrylonitrile	n.d. (75% rsm)

Reaction conditions: 1a (0.1 mmol, 1 equiv), acrylonitrile (2.5 equiv), SmI₂ (25 mol%), in THF (0.1 M) under N₂ (−10 °C, 30 min). The yield and diastereoisomeric ratio were determined by ¹H qNMR analysis of the crude reaction mixture with CH₂Br₂ as internal standard (0.05 mmol). rsm = recovered starting material.

Next, the reaction tolerance was evaluated using the optimal conditions (Scheme 1 and Table 1, entry 2). Higher reaction concentration gave a moderate decrease in yield of 2a; however, lowering the concentration had no impact. The use of higher temperature led to lower product yield. While lower temperature proved detrimental to conversion; extended reaction times would likely improve conversion at −20 °C.

Scheme 1.

Reaction tolerance screening. See Supporting Information for experimental details.

Batch‐to‐batch inconsistencies in the quality of Sm metal used to prepare SmI₂ can affect its reactivity.^{[ 45 ]} To evaluate the impact that this would have on catalysis, an old batch of Sm metal was reactivated by heating the metal under vacuum with vigorous agitation and storage under an inert atmosphere prior to the preparation of SmI₂.^{[ 45 ]} While SmI₂ prepared from this reactivated metal was a competent catalyst, product yield was diminished, thus highlighting the importance of SmI₂ quality when working in catalysis. The addition of a proton source, H₂O (5 µL), led to recovery of starting material with no product detected. We next probed the sensitivity of catalysis to oxygen. Removal of the N₂ inlet to the reaction vessel led to a moderate decrease in yield; the addition of air (1 mL) led to a further decrease in product formation. Finally, a lower catalyst loading (5 mol%) was found to return only traces of product.

Reaction Scope

We evaluated the scope of the SmI₂‐catalyzed coupling of housane ketones with alkenes using the optimal conditions (Table 1, entry 2). Regarding the carbonyl substituent, an array of housane ketones was subjected to the coupling with acrylonitrile (Scheme 2). Housane ketones bearing primary alkyl substituents afforded norbornane ketones 2a−i in good to excellent yield. More hindered housane ketones bearing secondary alkyl substituents were also competent substrates for the catalytic coupling providing 2j−l in good yield. However, a tertiary butyl ketone substrate gave 2m in a much lower yield.

Scheme 2.

Scope of the SmI₂‐catalyzed coupling of alkyl and aryl housane ketones with acrylonitrile. Reaction conditions: 1 (0.2 mmol, 1 equiv), acrylonitrile (2.5 equiv), SmI₂ (15 mol%), in THF (0.1 M) under N₂ (−10 °C, 30 min). The isolated yield is reported unless otherwise stated. The diastereoisomeric ratio was determined by ¹H qNMR analysis of the crude reaction mixture with CH₂Br₂ as an internal standard (0.05 mmol).

The reaction is not limited to alkyl ketones; aryl housane ketones afforded the norbornane products in good yields (2n–p). Aryl groups bearing electron‐withdrawing (2o) and electron‐donating (2p) substituents were compatible, as were heteroaryl groups; the corresponding norbornane ketones 2q and 2r were obtained in moderate yield. The catalytic coupling showed good functional group compatibility in tolerating ether (2e, 2l, 2p), alkenyl (2h), acetal (2i), fluoro (2o), indolyl (2f), oxazolyl (2g), furanyl (2q), and thienyl (2r) motifs in the housane ketone substrates. A housane derived from oxaprozin, a nonsteroidal anti‐inflammatory drug (NSAID), proved an effective partner furnishing norbornane ketone product 2g in 85% yield. Of note, a substrate bearing a terminal alkene underwent smooth coupling to give 2h despite the potential for the putative ketyl radical intermediate (see Scheme 5)^{[ 46 ]} to undergo 5‐exo‐trig cyclization.^{[ 47 ]}

Scheme 5.

Gibbs free energy profiles (∆G, in kcal mol⁻¹) for the SmI₂‐catalyzed coupling of 1a with acrylonitrile. For clarity, only structures leading to the major exo‐diastereoisomer are shown. Geometry optimizations were carried out at the PBE0‐D3(BJ) level using Stuttgart‐Köln ECPs and associated valence basis sets for Sm and I, with cc‐pVDZ basis sets for the remaining elements. Subsequent single‐point energy calculations were performed at the THF(PCM)‐PBE0‐D3(BJ) level, incorporating the DKH Hamiltonian, with the SARC basis set for Sm, the Jorge basis set for I and cc‐pVTZ basis sets for the remaining elements.

In all cases, the couplings proceeded with complete regiocontrol and with moderate diastereocontrol—norbornanes were obtained in approximately 3:1 dr. Products can be epimerized to give single diastereosiomers. For example, the crude mixture of diastereoisomers of 2a could be directly converted to a single exo‐diastereoisomer of the corresponding norbornane ketoacid in (81% overall yield) upon basic epimerization/nitrile hydrolysis (vide infra). Of note, exo‐diastereoisomers were the major products when employing alkyl housane ketones (products 2a−m), while endo‐diastereoisomers were formed in excess from the couplings of aryl housane ketones (2n−r). The relative stereochemistry of the major diastereoisomers of 2j‐exo and 2o‐ endo was confirmed by X‐ray crystallographic analysis.^{[ 48 ]}

We next sought to engage alternative electron‐deficient alkenes in the coupling with alkyl housane ketone 1a (Scheme 3). In addition to acrylonitrile, acrylates proved competent partners in the SmI₂‐catalyzed coupling, delivering norbornane ketones 2s–z in moderate to good yield. The successfully engaged acrylate partners highlight the compatibility of the process with alkynyl (2u), trifluoromethyl (2v), ether (2w), amino (2x), and chloro (2y) motifs. Vinyl sulfone and vinyl sulfonate esters also engaged as effective coupling partners, forming 2aa–2ac in good yield. Finally, alternative alkyl housane ketone 1b underwent smooth coupling with two acrylates and a vinyl sulfonate ester to give norbornanes 2ad–2af in high yield. While the couplings with acrylates, vinyl sulfones, and vinyl sulfonates proceeded with little diastereocontrol, products were amenable to epimerization to give single diastereoisomers (vide infra).

Scheme 3.

Scope of the SmI₂‐catalyzed coupling of alkyl housane ketones 1a and 1b with electron‐deficient alkenes. Reaction conditions: 1a,b (0.2 mmol, 1 equiv), alkene (2.5 equiv), SmI₂ (15 mol%), in THF (0.1 M) under N₂ (−10 °C, 30 min). The isolated yield is reported unless otherwise stated. The diastereoisomeric ratio was determined by ¹H qNMR analysis of the crude reaction mixture with CH₂Br₂ as an internal standard (0.05 mmol).

To further probe the scope of the reaction, bridgehead‐substituted housane ketone 1s was prepared and coupled under our standard catalytic conditions to furnish norbornane products 2ag–2aj in excellent yield (Scheme 4), indicating that tertiary alkyl radicals are competent intermediates in the intermolecular couplings. The structure of 2ai‐exo was confirmed by X‐ray crystallographic analysis.^{[ 48 ]}

Scheme 4.

Scope of the SmI₂‐catalyzed coupling of bridgehead‐substituted 1s with electron‐deficient alkenes. Reaction conditions: 1s (0.2 mmol, 1 equiv), alkene (2.5 equiv), SmI₂ (15 mol%), in THF (0.1 M) under N₂ (−10 °C, 30 min). The isolated yield is reported unless otherwise stated. The diastereoisomeric ratio was determined by ¹H qNMR analysis of the crude reaction mixture with CH₂Br₂ as an internal standard (0.05 mmol).

Computational Studies

Computational studies (PBE0‐D3(BJ) level) were performed to probe the mechanism of the SmI₂‐catalyzed coupling between model housane ketone 1a and acrylonitrile (see Supporting Information for computational details).^{[ 7} , ¹⁰ , ³⁰ , ⁴⁹ , ^{50 ]} The computed potential energy surface (Scheme 5) shows that the ketone undergoes an inner‐sphere SET with an activation energy of 15.5 kcal mol⁻¹, leading to the formation of a ketyl radical intermediate Int I. Then, Int I undergoes facile fragmentation, driven by the relief of strain in the housane, via transition state TS I (15.9 kcal mol⁻¹), yielding Int II. Subsequently, Int II is trapped by acrylonitrile through transition state TS II, with barrier energies of 4.2 and 0.4 kcal mol⁻¹ for the endo and exo pathways, respectively. This step is followed by the rebound, ring‐closing reaction of the radical adduct intermediate Int III to give Sm(III)‐bound ketyl radical Int IV, which then undergoes back electron transfer, regenerating Sm(II) and yielding the final products. The computational results suggest that the overall barrier for housane ketone coupling is only slightly higher than that calculated for the coupling of highly strained bicyclo[1.1.0]butane (BCB) ketones.^{[ 49 ]}

Conformational Analysis

The use of rigid, sp³‐rich carbocyclic motifs as replacements for benzene rings in biologically active molecules is now an established strategy in drug discovery.^{[ 51} , ^{52 ]} While recent work by Grygorenko and coworkers compared 1,4‐disubstituted norbornanes to para‐substituted benzenes,^{[ 53 ]} other substitution patterns on the norbornane ring system have yet to be considered as rigid alternatives to substituted benzenes with very different exit vectors. Motivated by this, we compared the X‐ray stuctures of 2j‐exo and 2o‐endo —representative products of our catalytic process—to that of a known 2‐ketobenzonitrile 3 (Scheme 6a).^{[ 48} , ^{54 ]} As expected, the C1─C2 bond length (r) was longer (1.557 and 1.546 Å) for both norbornane diastereoisomers, when compared with the analogous bond in 3 (1.409 Å). Furthermore, the φ1 and φ2 bond angles in both diastereoisomers were slightly smaller than in 3 (120.4° and 121.5°, respectively). Also, as expected, the torsion angle ϕ (the angle between the planes defined by the R¹−C¹−C² and C¹−C²−R² groups) is very different for the three compounds; ϕ = 2.9° in 3, while much larger angles, ϕ = 69.0° and ϕ = 48.3°, were measured for the norbornanes 2o‐endo and 2j‐exo .

Scheme 6.

a) Exit vector analysis of representative product diastereoisomers. b) Scale‐up SmI₂‐catalyzed alkene insertion into alkyl housane ketones and norbornane product manipulation. Reaction conditions: I – 2a (1.0 equiv), KOH (9.0 equiv), EtOH:H₂O (1:1, 0.25 M), Δ. II – 4 (1.0 eq), MeI (1.6 equiv), K₂CO₃ (2.0 equiv), DMF (0.2 M). III – step II, then, N₂H₄•H₂O (5.0 equiv), DME (0.2 M), Δ. IV – step II, then, NH₂OH•HCl (8.0 equiv), NH₄OAC (8.0 equiv), EtOH:H₂O (1:1, 0.08 M), Δ, then I₂ (50 mol%), MeCN (0.1 M), Δ. V – 4 (1.0 equiv), DPPA (1.5 equiv), Et₃N (1.5 equiv), PhMe (0.1 M), Δ, then 4‐OMe‐benzyl alcohol (3.0 equiv), PhMe (0.1 M), Δ. VI – Performed on exo isomer only; 2a (1.0 equiv), [RuCl₂(p‐cymene)]₂ (1.5 mol%), (CH₂O) _n (9.0 equiv), PhMe:H₂O (1:1, 0.25 M), Δ. VII – 2u (1.2 equiv), 3‐bromoprop‐2‐yn‐1‐ol (1.0 equiv), CuBr (10 mol%), ⁿ BuNH₂ (1.2 equiv), Na‐ascorbate (1.0 eq), EtOH (0.1 M). VIII – 2a (1.0 equiv), ^t BuOK (2.0 equiv), ^t BuOH (0.5 M), THF (0.1 M). Isolated yield is based on conversion of the exo‐isomer. XI – 2u (1.0 equiv), TsN₃ (1.0 equiv), CuTc (10 mol%), PhMe (0.1 M).

Manipulation of Norbornane Ketones

The SmI₂‐catalyzed coupling of housane 1a with acrylonitrile was carried out on a larger scale (5.0 mmol) to give 1.07 g of 2a (79% yield, 2.7:1 mixture of exo:endo diastereoisomers) (Scheme 6b). Epimerization of the products of catalysis allows convenient access to single diastereoisomers. For example, straightforward hydrolysis of the nitrile motif in 2a allowed access to the carboxylic acid 4 in quantitative yield and as a single exo diastereoisomer. Diastereomerically pure acid 4 was converted to the corresponding methyl ester 2s‐exo that was then converted to either the dihydropyridazinone 5, whose structure was confirmed by X‐ray crystallographic analysis,^{[ 48 ]} or to amide regioisomers 6 and 7 after Beckmann rearrangement. The diastereoisomerically pure acid 4 was also subjected to a Curtius rearrangement to generate exo‐carbamate 8. The exo isomer of 2a underwent selective nitrile reduction, under transfer hydrogenation conditions, to generate primary alcohol 9. Norbornane 2u bearing a propargyl ester group was prepared on a 1.0 mmol scale and underwent Cadiot–Chodkiewicz cross‐coupling, to afford diyne 10,^{[ 55 ]} and copper‐catalyzed azide‐alkyne cycloaddition to generate triazole 12.^{[ 56 ]} Finally, a diastereoisomeric mixture of 2a was converted to cyclic enamide 11 upon treatment with base. The structure of 11 was confirmed by X‐ray crystallographic analysis.^{[ 48 ]}

Conclusion

Samarium diiodide (Sml₂) is an effective catalyst for the mild, atom‐economical coupling of bicyclo[2.1.0]pentane, or housane, ketones, and electron‐deficient alkenes to give functionalized norbornane structural motifs not easily accessible by classical cycloaddition approaches. In addition to aryl‐substituted housane ketones, harder‐to‐reduce alkyl housane ketones are also excellent substrates for the radical cross‐coupling. The products of catalysis have considerable potential for further manipulation to give diastereoisomerically pure norbornane derivatives and norbornane‐fused heterocycles. The catalytic coupling constitutes a promising platform for the construction of sp³‐rich, conformationally defined, 3D architectures for use, for example, in medicinal chemistry.

Supporting Information

Experimental procedures, characterization data, computational results, X‐ray data, and NMR spectra for all new compounds. The authors have cited additional references within the Supporting Information.

Conflict of Interests

The authors declare no conflict of interest.

Supporting information

Supporting Information

Roy D., Mansell J. I., Barison G., Yu S., Katavic R., Romano C., Kaltsoyannis N., Procter D. J., Angew. Chem. Int. Ed. 2025, 64, e202512018. 10.1002/anie.202512018

Data Availability Statement

The data that support the findings of this study are available in the Supporting Information of this article.