Introduction

1,2,3-Triazoles play a crucial role in drug discovery, biorthogonal chemistry, synthetic chemistry, and materials sciences1,2,3,4,5,6,7,8. Since being simultaneously reported by the groups of Sharpless9,10 and Meldal11, copper-catalyzed azide-alkyne cycloaddition, also known as click reaction12,13,14,15,16,17,18,19, has emerged as a highly efficient and biocompatible strategy for synthesizing 1,2,3-triazoles. Despite the well-explored racemic version, the synthesis of chiral triazoles remains challenging mainly due to no sp3 stereogenic center in the triazole skeleton and the linear geometry of the azide and alkyne20,21. In this context, Zhou22,23,24,25,26, Fossey27,28,29, Topczewski30,31, and others32,33,34,35,36,37,38,39 have employed strategies such as (dynamic) kinetic resolution or desymmetrization to achieve a series of enantioselective copper-catalyzed click reaction of azides and terminal alkynes, which generated enantioenriched 1,4-disubstituted triazoles regioselectively (Fig. 1a). Notably, Zhou and co-workers used 1-iodoalkynes as special internal alkynes to prepare the corresponding chiral 1,4,5-trisubstituted triazoles24. In 2021, Topczewski et al. discovered enantioselective nickel-catalyzed click reaction with internal alkynes for dynamic kinetic resolution of allylic azides40, which gave the expected 1,4,5-trisubstituted triazoles with up to 86% ee (Fig. 1b). Recently, significant progress has been made in the atroposelective click reaction of β-naphthol-derived internal alkynes with azides, independently disclosed by Li41, Xu42, and Cui43,44. In these protocols, a range of axially chiral 1,4,5-trisubstituted triazoles were prepared by employing precious rhodium or iridium catalyst (Fig. 1c). Although current methods could provide diverse access to chiral 1,4-disubstituted and 1,4,5-trisubstituted triazoles, there were a few kind of asymmetric click reactions and no reports on efficient strategy for synthesis of chiral 1,5-disubstituted triazoles, which are core units of various bio-active molecules45,46,47,48,49,50. Therefore, exploring a kind of efficient catalytic system to achieve asymmetric click reaction especially for synthesizing chiral 1,5-disubstituted triazoles, is still highly desirable.

Fig. 1: Strategies for catalytic asymmetric synthesis of chiral 1,2,3-triazoles.
figure 1

a Synthesis of chiral 1,4-disubstituted triazoles by enantioselective Cu-catalyzed Click reaction. b Synthesis of chiral 1,4,5-trisubstituted triazoles by enantioselective Ni-catalyzed Click reaction. c Synthesis of chiral 1,4,5-trisubstituted triazoles by atroposelective Rh/Ir-catalyzed Click reaction. d This work: synthesis of chiral 1,5-disubstituted triazoles (unexplored) and 1,4,5-trisubstituted triazoles by enantioselective Cu-catalyzed azidation/Click cascade. Cat. catalyst, L ligand, Cu copper, Ni nickel, Rh rhodium, Ir iridium.

Organic azides, as one of the key starting materials for click reactions, have been extensively studied for their asymmetric synthesis51,52,53,54,55,56,57,58,59. Among these, hypervalent iodine-based azidating reagents were often selected as the azido sources60,61,62,63,64,65,66 due to their advantages of mild properties and high selectivity. In previous research, benziodazolone-based azidating reagents were identified as efficient azido transfer reagents that could significantly improve reaction outcomes by adjusting the substituents64. In continuation of our interest in hypervalent iodine chemistry64,65,66,67,68,69, we report herein an efficient enantioselective copper-catalyzed azidation/click cascade reaction of N-propargyl-β-ketoamides and azidobenziodazolone 1d. This method demonstrates broad substrate scope, easy operation, and mild reaction conditions. Importantly, this work represents catalytic asymmetric example of combining azidation and click reactions to construct chiral triazoles, particularly chiral 1,5-disubstituted triazoles (Fig. 1d).

Results

We studied the model reaction of N-propargyl-β-ketoamide 2a with azidobenziodoxolone (1a) in the presence of the catalyst prepared in situ from 10 mol % of Cu(MeCN)4PF6 and 12 mol % of Box ligand to optimize the reaction conditions (Table 1, see the details of optimization in Supplementary Tables 14). When the reaction was carried out in 1,2-dichloroethane at room temperature, the desired chiral triazole 3a was observed. Screening a series of Box ligands revealed that using L8 as the optimal ligand afforded 3a in 62% yield with 92% ee (Table 1, entries 1−10). After evaluating various solvents, it was found that dichloromethane could improve the yield and ee value of 3a (70% yield, 94% ee; Table 1, entry 11). A range of azidating reagents were subsequently tested. In comparison to 1a and azidodimethylbenziodoxole (1b), azidobenziodazolones (1c1g) were more suitable for the reaction, yielding product 3a with higher yields even at a reduced catalyst loading of 5 mol % (Table 1, entries 11−17). These results implied that the substituent on the nitrogen atom could significantly influence the reaction efficiency, with N-benzyl-azidobenziodazolone (1d) identified as the optimal azido source, affording product 3a in 99% yield with 94% ee (Table 1, entry 14).

Table 1 Optimization of the reaction conditionsa

With the optimized reaction conditions in hand, we evaluated the scope of the substrates bearing terminal alkyne unit (Fig. 2). When R2 was a benzyl group, all of the substrates were compatible with the reaction conditions to deliver the desired chiral 1,5-disubstituted triazoles 3 in high yields with high to excellent ee values. For example, the substrates bearing diverse substituents at the different sites of the aryl ring, including electron-withdrawing groups (5-Br, 4-Br, 6-Br, 7-Br, 5-Cl, 6-F, 6-CN, 6-CF3), electron-donating groups (4-OMe, 5-OMe, 6-OMe, 7-OMe, 5-Me) and electron-neutral group (H) were well tolerated under the conditions, affording products 3a3n in 75−99% yields and 85−95% ees. Moreover, substrate possessing 4,5-di-OMe on the aryl ring and naphthalene-derived substrate were also successfully employed in the process, generating 3o (81% yield, 80% ee) and 3p (93% yield, 91% ee), respectively. Subsequently, the influence of the N-substituent (R2) was investigated. Groups such as 1-naphthylmethyl, 1,1-diphenylmethyl, and even methyl were tolerated in this protocol, leading to the formation of the corresponding products 3q3s in 80−98% yields with 88−95% ees, while tert-butyl amide only gave the product 3t with 78% ee.

Fig. 2: Scope for the synthesis of chiral 1,5-disubstituted triazoles.
figure 2

Reaction conditions: 2 (0.1 mmol), 1d (0.15 mmol), Cu(MeCN)4PF6 (5 mol%), L8 (6 mol%), CH2Cl2 (2 mL), 25 °C; isolated yields; ee values were determined by HPLC analysis.

When six-membered ring β-ketoamide 2u and seven-membered ring β-ketoamide 2v were used as substrates, the corresponding product 3u and 3v were obtained with poor ee values, respectively. Unprotected N-propargyl-β-ketoamide 2w and O-propargyl-β-ketoester 2x were not compatible with the reaction, only generating the corresponding azide 3w’ and 3x’, respectively. These results showed the limitations of substrate scope.

As is well known, copper-catalyzed click reactions are traditionally limited to terminal alkynes14,17,18, except for a few special examples24,70,71. After investigating the scope of substrates bearing terminal alkyne units, we explored the possibilities for the reaction of internal alkynes (Fig. 3). Using 4a as the substrate, the cycloaddition step indeed did not proceed completely under the standard conditions mentioned above. Fortunately, the expected 1,4,5-trisubstituted triazoles 5a could be prepared in 75% yield and 93% ee just by simply raising the temperature to 40 °C and extending the reaction time to 96 h. Substrates with different groups (R1) in the backbone got the desired products 5b5e in 62−70% yields with 83−94% ees.

Fig. 3: Scope for the synthesis of chiral 1,4,5-trisubstituted triazoles.
figure 3

Reaction conditions: 4 (0.1 mmol), 1d (0.15 mmol), Cu(MeCN)4PF6 (5 mol %), L8 (6 mol %), CH2Cl2 (2 mL); the mixture was stirred at 25 °C for 48 h, then turned to 40 °C for more 48 h; isolated yields; ee values were determined by HPLC analysis. bThe absolute configuration was established to be R by the single-crystal X-ray structure analysis.

When the substituents (R2) of the triple bond were aryl rings with diverse functional groups in different positions, the corresponding products 5f5u could be formed in 36−75% yields with 90−94% ees. To our delight, the protocol was also applicable to the substrates with heteroaryl framworks such as 3-pyridyl and 6-quinolyl, affording the expected products 5v (60% yield, 90% ee) and 5w (60% yield, 94% ee), respectively. Furthermore, substrates with alkyl units such as methyl, ethyl and cyclohexyl in the triple bond could be converted to the corresponding products 5x5z with excellent enantiocontrol. Notably, the absolute configuration of enantiopure 5b, 5d, and 5x was established to be R by single-crystal X-ray structure analysis (for details, see Supplementary Notes). The stereochemical assignment of all other products was also made by analogy.

To demonstrate the synthetic utility of this protocol, a gram scale reaction was conducted with 2a on a 3 mmol scale (1.15 g) under standard conditions, affording the desired product 3a (1.21 g) in 95% yield without loss of enantioselectivity (Fig. 4a).

Fig. 4: Gram-scale reaction and mechanistic studies.
figure 4

a Gram-scale reaction. b Radical inhibition experiments. c Nonlinear effect of catalyst. d Investigation of intermediate azide. e The role of triple bond in enantiocontrol. f Proposed mechanism. TEMPO: 2,2,6,6-tetramethylpiperidoxyl, BHT: butylated hydroxytoluene, DPE: 1,1-diphenylethylene.

A series of control experiments were conducted to gain a preliminary understanding of the reaction mechanism. The reaction was inhibited completely by adding radical scavengers such as 2,2,6,6-tetramethyl-1-piperidinyloxy (TEMPO), 2,6-di-tert-butyl-4-methylphenol (BHT), or 1,1-diphenylethene (DPE) (Fig. 4b). These results revealed that the reaction pathway probably involves a radical process. A linear correlation was observed between the ee values of L8 and the enantioenrichment of the corresponding products (Fig. 4c). The absence of a nonlinear effect in this reaction suggested that a single catalyst is likely involved in the enantiodetermining transition state72. To confirm that the reaction is performed sequentially in the steps of azidation and cycloaddition, we terminated the reaction of 4x with 1d to get the intermediate azide 6 with 95% ee, which could be converted to the desired product 5x in 75% yield with 94% ee under standard reaction conditions. However, the cycloaddition step did not proceed effectively under conditions without a catalyst or with only Cu(MeCN)4PF6 as the catalyst (Fig. 4d). These results indicated that the Cu/L8 catalyst probably involved in both steps of the reaction, with the azidation being the enantio-determining step. To verify the role of the triple bond in enantiocontrol, several control compounds were chosen to perform the asymmetric azidation reaction (Fig. 4e). Ketoester 7a and ketoamide 7b, both with no propargyl group, got the corresponding products 8a8b with moderate ee values. Compounds 7c and 7d, using propyl instead of propargyl, could get the azidating products 8c8d with reduced ee values in comparison with those of the analogous triazoles (3a and 3q), respectively. These data implied that the triple bond in the substrate might behave as a directing group in enantiocontrol of azidation step.

To gain insights into the mechanisms of azido transfer and the origin of the enantioselectivity induced by the Box ligand, we carried out DFT calculations to propose a catalytic model (Fig. 4f). The bidentate ligand L8, azido anion, and a carbonyl group of substrate coordinate to the Cu(III) center to form a complex with a distorted tetrahedral coordination geometry as the intermediate. The favored conformation of the intermediate (int-I) exhibits an energy that is 8.34 kcal/mol lower than that of the other side (int-II). This discrepancy is primarily attributed to the steric hindrance caused by the carbonyl group of substrate and the tert-butyl group of ligand L8. Such hindrance leads to an increase in the angle between the exocyclic amide carbonyl and the five-membered ring from 54° (int-I) to 71° (int-II). The enlargement of this angle results in a deterioration of conjugation, which not only elevates the energy level but also makes the enol double bond more reactive due to the loss of conjugation. Additionally, the enlargement of the angle causes the groups attached to the amide, such as the alkynyl group in this case, to rotate towards the backside of the α-carbon, thereby obstructing the approach of azide.

Even though the exact mechanism remains unclear, a tentative catalytic mechanism was proposed according to the experimental results and previous related reports (Fig. 4f)52,73,74. Initially, copper ion coordinates with Box ligand to get the chiral catalyst A, which activates azidobenziodazolone 1d to afford azido-Cu(III) species B73,74. Intermediate B exhibits radical properties and can dissociate into an azido radical. Intermediate B then reacts with substrate 4x to get intermediate C, where copper probably coordinates with both a carbonyl group of the substrate and an azido anion. Intermediate C, in its favored conformation int-I, undergoes stereoselective reductive elimination to afford the chiral intermediate azide D. Copper species would coordinate with both the azido group and triple bond in D to render intermediate E, which probably proceeds the click cycloaddition to form intermediate F75,76. Finally, F releases product 5x along with the regeneration of catalyst A.

Discussion

In conclusion, we have developed an enantioselective copper-catalyzed azidation/click cascade reaction of N-propargyl-β-ketoamides using our previously developed azidobenziodazolone as the efficient azido source. This protocol afforded an efficient and mild route to synthesize chiral triazoles, especially chiral 1,5-disubstituted triazoles, which have not been obtained via reported asymmetric click reactions. Moreover, high functional group tolerance, scalability in synthesis, and easily operated reaction conditions further demonstrated the synthetic utility of this strategy. Initial mechanistic studies revealed that the copper/Box catalyst participates in both azidation and click steps. Further mechanistic investigation and expansion of the azidation/click cascade strategy to access more chiral triazoles are ongoing in our laboratory.

Methods

General procedure for the synthesis of racemic 3a–3v

To a mixture of Cu(MeCN)4PF6 (0.005 mmol, 5 mol %), substrates 2a–2v (0.10 mmol), 1d (0.15 mmol) in dry dichloromethane (2 mL) under nitrogen atmosphere, the reaction system was stirred at 25 °C for 48 h. The disappearance of substrates 2a2v was monitored by TLC, indicating complete consumption. Finally, the crude product was purified by silica gel flash chromatography to afford the desired product rac-3arac-3v.

General procedure for the synthesis of racemic 5a–5z

To a mixture of Cu(MeCN)4PF6 (0.005 mmol, 5 mol %), substrates 4a–4z (0.10 mmol), 1d (0.15 mmol) in dry dichloromethane (2 mL) under nitrogen atmosphere, the reaction system was stirred at 40 °C for 48 h. The disappearance of substrates 4a4z was monitored by TLC, indicating complete consumption. Finally, the crude product was purified by silica gel flash chromatography to afford the desired products rac-5arac-5z.

General procedure for the synthesis of chiral 3a–3v

After stirring a mixture of Cu(MeCN)4PF6 (0.005 mmol, 5 mol %) and L8 (0.006 mmol, 6 mol %) in dry dichloromethane (1 mL) at 25 °C under nitrogen atmosphere for 1.5 h, substrates 2a–2v (0.10 mmol) and 1d (0.15 mmol) in dry dichloromethane (1 mL) were added. The reaction mixture was then stirred at 25 °C under a nitrogen atmosphere. After 48 h, the substrates 2a–2v disappeared completely (monitored by TLC). Finally, the crude product was purified by silica gel flash chromatography to afford the desired products 3a–3v.

General procedure for the synthesis of chiral 5a-5z

After stirring a mixture of Cu(MeCN)4PF6 (0.005 mmol, 5 mol %) and L8 (0.006 mmol, 6 mol %) in dry dichloromethane (1 mL) at 25 °C under nitrogen atmosphere for 1.5 h, substrates 4a–4z (0.10 mmol) and 1d (0.15 mmol) in dry dichloromethane (1 mL) were added. The reaction mixture was stirred at 25 °C under a nitrogen atmosphere. After 48 h, when the substrates 4a–4z disappeared (monitored by TLC), the reaction mixture was stirred at 40 °C for an additional 48 h. Finally, the crude product was purified by silica gel flash chromatography to afford the desired product 5a–5z.