The generation of α-imino metal carbenes is arguably the most important aspect of metal carbene chemistry attributed to their abundant carbene transfer reactions in the synthesis of valuable N-containing molecules [1-12]. Among these, gold-catalyzed alkyne amination reactions via α-imino gold carbene complexes have proved to be a powerful and intriguing method in organic synthesis because of their high regio- and stereoselectivity for rapid assembly of a variety of worthwhile N-heterocycles from readily available alkynes. In the past decades, a series of amination reagents, including nitrogen ylides [13-20], sulfilimines [21-27], isoxazoles [28-37], anthranils [38-46] and others [47-52], have been nicely exploited in the gold-catalyzed intermolecular alkyne amination reactions (Scheme 1a). Since the pioneer work on the gold-catalyzed azide-alkyne cyclization reaction via α-imino gold carbenes in 2005 by Toste and co-workers [53], an increasing number of protocols involving the generation of α-imino gold carbenes through gold-catalyzed amination reactions based on azides have been developed in the past decades owing to the ready availability of the azide precursors, environmental friendliness and good compatibility with functional groups of these reactions [53-68]. However, such protocols are only limited to intramolecular generation of α-imino gold carbenes probably due to the weak nucleophilicity of azides, and intermolecular generation of α-imino gold carbenes by the use of azides as amination reagents has not been evoked.

In 2015, our group described the first gold-catalyzed amination reactions by employing the benzyl azides and indolyl azides as a new type of amination reagents for the intermolecular generation of α-imino gold carbenes, leading to the facile formation of functionalized 2-aminoindoles and 3-amino-β-carbolines in generally high yields via formal [3 + 2] and [4 + 2] annulations, respectively [69,70]. On the basis of this work, in 2016, we further realized a gold-catalyzed intermolecular amination reaction initiated aza-Nazarov cyclization reaction based on 3-en-1-ynamides and a gold-catalyzed intermolecular tandem ynamide amination/C—H functionalization by both using the benzyl azides as amination reagents, thus allowing the flexible synthesis of synthetically useful 2-aminopyrroles and 2-aza-1,3-butadienes, respectively [71,72]. Inspired by the above results and by our recent study on developing ynamide chemistry for N-heterocycle synthesis [73-91], we envisioned that the use of allyl azides might act as amination reagents to react with ynamides for the intermolecular generation of alkene-tethered α-imino gold carbenes, followed by intramolecular cyclopropanation, eventually leading to valuable 3-azabicyclo[3.1.0] scaffold (Scheme 1b), which is widely existing in pharmaceutical molecules and natural products (Fig. 1) [92-101]. However, realizing this intermolecular annulation is highly challenging: (i) Whether the allyl azide is nucleophilic enough to inter-molecularly attack the ynamide; (ii) how to prevent the possible triazole formation via direct azide-alkyne [3 + 2] cycloaddition [102-105] and (iii) how to control the regioselectivity and diastereoselectivity. Herein, we disclose such a gold-catalyzed ynamide amination initiated cyclopropanation by employing the allyl azide as amination reagent, thus representing the first use of allyl azide as a novel amination reagent. This protocol enables the efficient and flexible synthesis of a wide range of 3-azabicyclo[3.1.0]hex-2-enes in generally good to excellent yields. In addition, the formed 3-azabicyclo[3.1.0]hex-2-enes can be readily converted into valuable γ-lactams [106-108] through a facile hydrolysis of the amide moiety under the acidic conditions. Furthermore, a mechanistic rationale for this α-imino gold carbene-involved intermolecular annulation reaction is strongly supported by theoretical calculations.

We commenced our investigation by examining the reaction between readily prepared ynamide 1a and allyl azide 2a in DCE at room temperature using IPrAuNTf₂ as gold catalyst. To our delight, the desired fused cyclopropanation product 3a could be obtained in 30% yield with 60% of ynamide 1a recovered (Table 1, entry 1). Encouraged by this preliminary attempt, we then increased the reaction temperature to 80 ℃, and found that 3-azabicyclo[3.1.0]hex-2-ene 3a could be obtained in 85% NMR yield with the full conversion of 1a (Table 1, entry 2). Subsequently, solvent screening showed that typical solvents such as toluene and MeCN failed to improve the reaction yield (Table 1, entries 3 and 4). Various gold catalysts, including BrettPhosAuNTf₂, Au(Ⅲ), Ph₃PAuNTf₂ and Cy-JohonPhosAuNTf₂, were then investigated but also failed to further improve the yield, and significant formation of hydrolysis product 3aa was observed in these cases (Table 1, entries 5–8). Of note, only triazole formation (97%) via azide-alkyne [3 + 2] cycloaddition was observed under rhodium catalysis, which was similar to Huang's protocol (Table 1, entry 9) [102]. In addition, the use of other representative transition metal catalysts such as AgNTf₂ and Zn(OTf)₂ also failed to deliver the desired product 3a, and hydrolysis product 3aa was obtained as the main product in these cases (Table 1, entries 10 and 11). Finally, it was found that the reaction time could be shortened to 60 h by increasing the catalyst amount to 10 mol%, but no obvious improvement of the yield was observed (Table 1, entry 12). Importantly, excellent diastereoselectivities (> 20:1) were achieved in all these cases.

Having established the optimized reaction conditions (Table 1, entry 2), we next explored the substrate scope of this gold-catalyzed intermolecular amination reaction of ynamides with allyl azides. As shown in Scheme 2, we initially investigated the scope of a wide array of substituted ynamides 1 under the optimized conditions. It was found that ynamides with various substitutions at the para position of the phenyl ring were suitable substrates to furnish the corresponding 3-azabicyclo[3.1.0]hex-2-ene products 3a–3f in 64%–88% yields, and ynamides bearing electron-deficient groups gave better yields than the ones containing the electron-rich groups on the aryl. Ynamides bearing the disubstituted aryl, such as 3,5-dichloro and 3-chloro-5-bromo substitution, were well tolerated in this annulation, affording the desired 3-azabicyclo[3.1.0]hex-2-enes 3g and 3h in excellent yields. In addition, the reaction was also extended to the naphthyl and furyl substituted ynamides, and the expected cyclopropanation products 3i and 3j were formed in both of 66% yields. Finally, ynamides with diverse R² substituents (ⁿBu, Bn) were able to undergo smooth annulation to deliver the desired products 3k in 78% yield and 3l in 84% yield, respectively. Prominently, excellent diastereoselectivities (dr > 20:1) were obtained in all these cases, and no triazole formation via azide-alkyne [3 + 2] cycloaddition was detected in such a gold catalysis.

Next, we explored the reaction scope of various allyl azides 2. As depicted in Scheme 3, different aryl substituted allyl azides tethered with either electron-donating or -withdrawing substituents on different position of the phenyl ring, such as Me, ⁱPr, OMe, OTBS, F, Cl, and Br, were well tolerated in this intermolecular amination reaction with ynamide 1a to generate the desired 3-azabicyclo[3.1.0]hex-2-ene products 3m–3t in generally good to excellent yields. Besides, the reaction proceeded smoothly with disubstituted aryl allyl azides, providing the corresponding bicyclic cyclopropanation products 3u–3w in 79%–83% yields. 1-Naphthyl-substituted allyl azide also provided an excellent universal fit for this annulation reaction, affording the expected product 3x in 80% yield. Apart from the aryl substituted allyl azides, the alkyl substituted allyl azide was proved to be an appropriate substrate to produce the alkyl substituted cyclopropane compound 3y in 79% yield. Finally, the allyl azide containing 1,2-disubstituted olefin moiety could also be smoothly converted into the corresponding multisubstituted 3-azabicyclo[3.1.0]hex-2-ene 3z bearing two vicinal all carbon quaternary stereocenters which was difficult to construct by traditional methods in organic synthesis. Again, excellent diastereoselectivities (dr > 20:1) were achieved in all cases. The relative configuration of product 3r was verified by X-ray crystallographic analysis and that of others was assigned analogously (Fig. 2) [106-108]. Thus, this intermolecular protocol provides a highly efficient and flexible route for the construction of valuable 3-azabicyclo[3.1.0] scaffold. Our attempts to extend the reaction to the chiral amine-derived ynamide for the asymmetric synthesis has been unsuccessful, and attempts to develop the direct asymmetric catalysis by employing Shi's chiral N-heterocyclic carbene (NHC)-derived chiral gold catalysts also failed [109-115].

Further preparative scale reaction and synthetic application of the as-synthesized 3-azabicyclo[3.1.0]hex-2-enes 3 were then explored (Scheme 4). Initially, preparative scale could be achieved for the synthesis of 3a (530 mg, 62%) with a catalyst loading of 3 mol%. Then, synthetic applications of 3-azabicyclo[3.1.0]hex-2-ene 3l were investigated. For example, hydrolysis of the amide moiety of 3-azabicyclo[3.1.0]hex-2-ene 3l via the facile use of HCl (aq.) led to the cyclopropane-fused γ-lactam 4 in 66% yield, and subsequent reduction could afford pyrrolidine 5, which is broadly existed in natural products and bioactive molecules [106-108]. Moreover, 3l could be readily transferred into the corresponding pyrrolidine 6 in 80% yield with excellent diastereoselectivity (>20:1) via selective ring opening of cyclopropane and debenzylation process under the reductive condition by using Pd/C and HCO₂NH₄. The structures of γ-lactam 4 and pyrrolidine 6 were confirmed by X-ray diffraction analysis (Fig. 3).

On the basis of our previous studies and density functional theory (DFT) calculations [69-72,109], a plausible reaction mechanism for this gold-catalyzed intermolecular amination of ynamides with allyl azides is proposed. As shown in Scheme 5, ynamide 1a is first activated by the gold catalyst to afford intermediate A, which undergoes regioselective attack by the allyl azide 2a via transition state TS with a free energy barrier of 19.6 kcal/mol, leading to the formation of vinyl gold intermediate B. Subsequently, intermediate B overcomes a free energy barrier of 5.7 kcal/mol via TS with a release of N₂, resulting in the generation of α-imino gold carbene intermediate C. Then, an intramolecular attack occurs between the gold carbene and the alkene moiety within intermediate C to deliver the carbocation intermediate D with a free energy barrier of 7.9 kcal/mol via TS. Subsequent intramolecular cyclopropanation of the intermediate D occurs almost without barriers to yield the cyclopropane intermediate E. Finally, intermediate E gives rise to the desired 3-azabicyclo[3.1.0]hex-2-ene 3a via a substrate exchange. The whole process proceeds smoothly and highly exergonically with a release of free energy amounting to 88.3 kcal/mol.

In summary, we have developed an efficient gold-catalyzed intermolecular amination reaction of ynamides with allyl azides, leading to the flexible synthesis of a range of synthetically useful 3-azabicyclo[3.1.0]hex-2-ene derivatives in good to excellent yields with excellent diastereoselectivities. Importantly, this protocol represents the first use of allyl azide as an efficient amination reagent in gold-catalyzed alkyne amination reactions. Of note, the as-synthesized 3-azabicyclo[3.1.0]hex-2-enes can be readily converted into valuable γ-lactams via hydrolysis of the amide moieties under the acidic conditions, thus fulfilling the facile removing of the directing group of ynamides. Moreover, theoretical calculations provide further evidence for the proposed α-imino gold carbene-involved mechanism. Further investigation of the asymmetric version of this annulation and other related intermolecular alkyne aminations are ongoing in our lab.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

CRediT authorship contribution statement

Chong-Yang Shi: Writing – original draft, Validation, Investigation, Data curation. Jian-Xing Gong: Validation, Investigation, Data curation. Zhen Li: Investigation, Data curation. Chao Shu: Investigation, Data curation. Long-Wu Ye: Writing – review & editing, Writing – original draft, Validation, Funding acquisition, Conceptualization. Qing Sun: Writing – review & editing, Funding acquisition. Bo Zhou: Writing – review & editing. Xin-Qi Zhu: Writing – review & editing, Conceptualization.

Acknowledgments

We are grateful for financial support from the National Natural Science Foundation of China (Nos. 22125108, 22331004 and 22121001), Yunnan Normal University, Applied Basic Research Projects of Yunnan Province (No. 202101AT070217), the Jiangxi Provincial Natural Science Foundation (No. 20224BAB213013), and the Jiangxi Provincial Department of Education in Science and Technology Program Project (No. GJJ210906). We also thank Mr. Zanbin Wei from Xiamen University for assistance with X-ray crystallo-graphic analysis.

Supplementary materials

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.109895.