Xin-Xin
Liu
a,
Shaohong
Wang
b,
Yuan
Zhang
a,
Fan
Yang
a,
Hong-Jie
Miao
a,
Peng-Chen
Ma
a,
Li-Na
Guo
*a and
Pin
Gao
*a
aDepartment of Chemistry, School of Chemistry, Xi'an Key Laboratory of Sustainable Energy Material Chemistry and MOE Key Laboratory for Nonequilibrium Synthesis and Modulation of Condensed Matter, Xi'an Jiaotong University, Xi'an 710049, China. E-mail: gaopin@xjtu.edu.cn
bHubei Key Laboratory of Energy Storage and Power Battery, School of Optoelectronic Engineering, Hubei University of Automotive Technology, Shiyan 442002, PR China. E-mail: wangshaohong@huat.edu.cn
First published on 4th August 2025
Here, we report a copper-catalyzed C–C bond cleavage of cyclic oxaziridines and a C(sp3)–N coupling reaction with amines. The reaction, which proceeds under mild conditions and features simple operation and an ideal 100% atom economy, enables the synthesis of a wide variety of functionalized amides in moderate to excellent yields. Density functional theory (DFT) calculations reveal that copper initially catalyses N–O bond cleavage to generate an N-radical species, which in turn induces β-scission of the C–C bond by producing an amidoalkyl radical species. This reactive amidoalkyl radical species is subsequently trapped by amine nucleophiles, furnishing the final products after a C(sp3)–N reductive elimination process.
Green foundation1. We report an efficient copper-catalyzed protocol for the C–C bond cleavage of cyclic oxaziridines and the subsequent C(sp3)–N coupling with amines, which proceeds under mild, redox-neutral conditions with 100% atom economy.2. The reaction demonstrates excellent functional group compatibility, allowing smooth amidoalkylation of structurally diverse nitrogen nucleophiles to afford the corresponding 6-aminohexanamides in moderate to excellent yields. Additionally, it enables ring-opening/amination transformations for both strained and unstrained cyclic systems. 3. In future work, we intend to explore efficient coupling reactions between various nucleophiles and cyclic oxaziridines with 100% atom economy. |
Additionally, Hu, MacMillan and others have made significant progress in decarboxylative amination of masked alkyl carboxylic acids such as alkyl NHPI esters and iodomesitylene dicarboxylates using synergetic photoredox/copper-catalytic reaction systems (Scheme 1b).6 Diacyl peroxides,7 alkylsilyl peroxides,8 cyclobutanone oxime esters,9 alkylthianthrenium salts10 and other compounds were subsequently proven to serve as efficient sources of alkyl radicals. Under thermal conditions, transition-metal catalysis or irradiation, they can generate the corresponding alkyl radicals, which would then undergo the C(sp3)–N cross-coupling reaction (Scheme 1c). Despite these remarkable developments, most of the above precursors have functional groups that need to be removed during the process of generating the alkyl radical. It is still highly desirable to develop more efficient precursors to provide alkyl radicals and achieve the direct construction of C(sp3)–N bonds with 100% atom economy.
Oxaziridines, by virtue of their weak N−O bond and strained three-membered heterocyclic ring, display abnormal stability and distinctive reactivities, and have been used as powerful and elegant oxygen- and nitrogen-transfer agents for a broad array of nucleophiles.11 Asymmetric transformations including oxidation and amination have also seen significant progress.11,12 Moreover, oxaziridines could undergo various intramolecular radical rearrangements, isomerization and cyclization under transition-metal catalysis or photochemical conditions.11,13 During these transformations, carbon radicals have been shown to be generated from nitrogen radicals via the selective cleavage of C–C bonds.13a,14 Based on this concept, the Maruoka group independently developed efficient protocols for generating alkyl radicals from oxaziridines, which were trapped by various 1,1-diarylethylenes to yield Heck-type products.15 Moreover, these alkyl radicals could also successfully participate in copper-catalyzed decarboxylative C(sp2)–C(sp3) coupling of α,β-unsaturated carboxylic acids.16 However, to the best of our knowledge, only these two types of radical acceptors have been successfully applied to capture the alkyl radicals, so it remains both necessary and highly desirable to investigate more efficient nucleophiles for such transformations and broaden the frontier of oxaziridine chemistry.
Our group has a strong interest in radical-mediated carbon–carbon bond cleavage and subsequent functionalization transformations. A wide range of nitrogen nucleophiles has been employed to trap the alkyl radicals generated from iminyl-radical-initiated carbon–carbon bond cleavage, providing a useful approach for C(sp3)–N bond formation.9a Based on this discovery and previous reports, we set out to pursue the challenging C(sp3)–N bond coupling of cyclic oxaziridines with amines, aiming at incorporating the important alkyl-substituted amide moieties into structurally diverse nitrogen-containing molecules, especially bioactive natural products and drugs, with 100% atom economy (Scheme 1d).
Our investigation began with an examination of the reaction between oxaziridine 1a and aniline 2a in the presence of copper(I) catalysts and 1,10-phen at 60 °C for 8 h. To our delight, the reaction occurred and gave the desired N-cyclohexyl-6-(phenylamino)hexanamide 3a in 25% yield with 100% atom economy (Table 1, entry 1). Other solvents, including EtOH, toluene and acetonitrile were less effective for this reaction (entries 2–4). The effect of different nitrogen ligands including dtbpy, terpy, 5,6-di-Me-phen and 3,4,7,8-tetra-Me-phen was then examined, and the choice of 3,4,7,8-tetra-Me-phen afforded the highest chemical yield (entries 5–8). Notably, even when the reaction temperature was lowered from 60 °C to room temperature, the reaction continued to proceed smoothly, affording 3a in a satisfactory yield of 46% (entry 9). Screening of copper catalysts revealed CuSCN to be the most effective catalyst (entries 10–13), whereas other catalysts such as PdCl2 and CoBr2 were found to be inert for this transformation (entry 14). To our delight, the yield of 3a was improved from 58% to 79% when the ratio of 1a to 2a was optimized at 2:
1 (entry 15). Control experiments indicated both the copper catalyst and ligand to be essential for this reaction (entries 16–17). In summary, we established 5.0 mol% CuSCN, 10 mol% 3,4,7,8-tetra-Me-phen in 2.0 mL of CH3OH at rt for 8 h as the optimal conditions.
Entry | Catalyst | Ligand | Solvent | Temp. (°C) | Yield (%) |
---|---|---|---|---|---|
a Reaction conditions: 1a (0.2 mmol, 1.0 equiv.), 2a (0.4 mmol, 2.0 equiv.), catalyst (5.0 mol%), ligand (10 mol%), solvent (2.0 mL), rt, 8 h, under N2. b Isolated yield. c 1a (0.4 mmol, 2.0 equiv.), 2a (0.2 mmol, 1.0 equiv.). | |||||
1 | CuI | 1,10-phen | CH3OH | 60 | 25 |
2 | CuI | 1,10-phen | EtOH | 60 | 11 |
3 | CuI | 1,10-phen | Toluene | 60 | 0 |
4 | CuI | 1,10-phen | CH3CN | 60 | 0 |
5 | CuI | dtbpy | CH3OH | 60 | 15 |
6 | CuI | terpy | CH3OH | 60 | 20 |
7 | CuI | 5,6-Di-Me-phen | CH3OH | 60 | 38 |
8 | CuI | 3,4,7,8-tetra-Me-phen | CH3OH | 60 | 45 |
9 | CuI | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 46 |
10 | CuBr | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 30 |
11 | CuBr2 | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 23 |
12 | Cu(OTf)2 | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 15 |
13 | CuSCN | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 58 |
14 | PdCl2 or CoBr2 | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 0 |
15c | CuSCN | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 79 |
16c | None | 3,4,7,8-tetra-Me-phen | CH3OH | rt | 0 |
17c | CuSCN | None | CH3OH | rt | 0 |
With the optimal reaction conditions in hand, we set out to evaluate the generality and limitations of this ring-opening/amination reaction with 2-cyclohexyl-1-oxa-2-azaspiro [2.5]octane 1a as the model substrate (Scheme 2). The results showed that aryl amines bearing electron-withdrawing groups (Cl, Br, I, CF3, CN, etc.) on the para-position of the phenyl ring all reacted smoothly and delivered the desired products 3b–3h in moderate to high yields. However, the electron-rich aromatic amines 2i and 2j were less efficient under the standard reaction conditions, giving 3i and 3j in lower yields along with some unidentified side products, probably due to the electron-rich aromatic amines being more easily oxidized in this system. Remarkably, the amination appeared to be insensitive to steric effects (2k–2p), providing the corresponding 6-aminohexanamides 3k–3p in 45–95% yields. Disubstituted aryl amines and 2-aminonaphthalene 2q–2t showed good performance levels under the optimized conditions. To our delight, this protocol could be extended to heterocyclic amine nucleophiles such as 6-aminobenzothiazole 2u and 2-aminopyridine 2v, providing the products 3u and 3v in 58 and 60% yields, respectively. Upon extending the reaction time and elevating the reaction temperature, both pyrrolo[2,1-F][1,2,4]triazin-4-amine 2w and 4-aminoquinazoline 2x also showed good reaction efficiency, delivering the desired compounds 3w and 3x in excellent yields. In addition, weakly nucleophilic aromatic N-heterocycles such as indole 2y and carbazole 2z proved to be amenable substrates, delivering the anticipated products 3y and 3z in satisfactory yields. Unfortunately, no desired product was formed when diphenylamine (2aa) was used as the nucleophile. The more challenging aliphatic amine 2ab—which has rarely been investigated, primarily due to the issues of E2 side reactions and over-alkylation—also underwent amidoalkylation, affording a 60% yield. In addition, benzyl carbamate, benzamide and 2-thiophenecarboxamide were also suitable substrates, affording the desired products 3ac–3ae in 47–72% yields. However, presumably due to its weak nucleophilicity, tosylamide (2af) was found to be incompatible with this reaction. Finally, drugs including emtricitabine and apixaban were also engaged in this amidoalkylation reaction to afford the desired derivatives 3ag and 3ah in satisfying yields, which clearly highlighted the potential utility of this protocol for late-stage modification of complex bioactive molecules.
Subsequently, we evaluated the scope of oxaziridines 1 using aniline 2a as the model nucleophile under this copper-catalytic system (Scheme 3). Satisfactorily, the 4-, 5-, 7-, 8-, and 15-membered cyclic oxaziridines furnished the corresponding amides 4a–4e in moderate to good yields via ring opening of the cycloalkyl moiety of the oxaziridine. Oxaziridine 1f, bearing an oxygen atom on the cycloalkane ring, was well tolerated in the reaction, affording 4f in 70% yield. To our delight, oxaziridine 1g derived from 4-phenylcyclohexanone also furnished the target product, namely 4g, in a high yield of 72%. Interestingly, when 2-cyclohexyl-4-methyl-1-oxa-2-azaspiro[2.5]octane 1h was subjected to the standard conditions, two types of 6-aminohexanamides, 4h and 4h′, were detected in a ratio of more than 2:
1. This phenomenon can be explained by the regioselectivity of the ring-opening amination reaction. The reaction preferentially occurred at the carbon atom adjacent to the methyl substituent, leading to the formation of a relatively stable alkyl radical for the construction of the C–N bond. Benzocyclohexanone-derived oxaziridine could efficiently participate in the C–C bond cleavage, affording the expected product 4i in 45% yield. Luckily, when the substituent on the nitrogen was changed to a tosyl group, the anticipated product 4j could still be obtained, albeit in relatively low yield. In contrast, replacing the nitrogen protecting group with n-hexyl precluded the formation of ring-opening amination product 4k. Furthermore, when the acyclic oxaziridine 1l (derived from methyl isopropyl ketone) was used in the C–C cleavage/amination reaction, isopropylamine 4l was obtained in a yield of 63%.
The utility of this protocol was illustrated by performing a gram-scale synthesis (Scheme 4). When the reaction of 1a with 2a was conducted on a 4.5 mmol scale, 3a was obtained in 70% yield, showing a yield and selectivity comparable to those for the small-scale reactions.
To shed light on the reaction mechanism, control experiments were carried out (Scheme 5a and b). When 2.0 equiv. of TEMPO or BHT was added under the standard reaction conditions, the reaction was suppressed completely, and the alkyl-TEMPO adduct S1 was detected using HRMS. These results suggested a probable involvement of an alkyl radical intermediate in this transformation.
Density functional theory (DFT) calculations were performed to learn more about the reaction mechanism, and the resulting Gibbs free energy profile for the reaction is shown in Fig. 1. Initially, according to the proposed mechanism, oxaziridine 1a coordinates to active catalyst 1SM, producing the Cu(I) species 1IM1 (3.7 kcal mol−1), with this process consistent with previous calculations by Maruoka.151IM1 then undergoes N–O bond cleavage via the transition state 1TS1 at the closed shell with an energy barrier of 12.7 kcal mol−1, affording the d9-centered Cu(II) species 3IM2; this cleavage is the rate-determining step. 3IM2 is understood to result from homolysis of the N–O bond, accompanied by electron transfers from Cu to the O atom. Subsequently, penta-coordinated copper complex 3IM3 (−7.3 kcal mol−1) is formed with a release of 7.1 kcal mol−1 of energy. The spin density analysis carried out on O, N and Cu atoms further indicated the N-radical species 3IM3 to be more stable than the O-radical species 3IM2. Finally, the N-radical species 3IM3 induces β-scission via3TS2 with an energy barrier of 12.4 kcal mol−1. Compared to the values for 3IM3, the spin density on N of 3TS2 is decreased by 0.46 and that on C is simultaneously increased to 0.60, affording the amidoalkyl radical intermediate 3IM4 with −8.6 kcal mol−1 energy. The d8-metal-centered Cu(III) species 1IM5 (−11.6 kcal mol−1) is formed by a thermodynamically favorable transmetalation between MeOH and 3IM4, and a classical reductive elimination of 1IM5 is exothermic, specifically with a change in energy of −45.6 kcal mol−1, leading to the final product 3a.
The optimization process of experimental conditions, specific procedures for the reaction, relevant characterization data of products, and data from the calculation section are all included in the Supplementary Information files. See DOI: https://doi.org/10.1039/d5gc03001c
This journal is © The Royal Society of Chemistry 2025 |