Introduction

Multicomponent reactions have been identified as a prominent strategy for the rapid generation of molecular complexity1,2,3,4,5,6. The key to the success of multicomponent reactions is to meticulously ensure the compatibility of each step, allowing the simultaneous addition of all reactants and catalysts at the onset of the reaction. Among numerous processes, transition-metal-catalyzed asymmetric multicomponent radical reactions have gained considerable interest due to their remarkable compatibility with many functional groups and unique reactivity profile7,8,9,10,11,12,13,14,15. To ensure high levels of chemo-, regio-, and enantioselectivity, one variant of this process is to combine reactants possessing different reactivities (Fig. 1a, left). Another type of reaction is to utilize easily available reactants with close similarity in the chemical structures and properties (Fig. 1a, right). The similarity would undoubtedly increase the complexity of reaction pathway, making it difficult to achieve a delicate balance among chemo-, regio-, and stereoselectivity.

Fig. 1: Motivation and development of Cu(I)-catalyzed enantioselective radical 1,2-carboalkynylation of two different terminal alkynes.
figure 1

a Transition-metal catalyzed three-component asymmetric radical reactions. b Challenges for asymmetric radical difunctionalization of two different terminal alkynes. c Cu(I)-catalyzed enantioselective radical 1,2-carboalkynylation with two terminal alkynes. M metal, tBu tert-butyl, Ar aryl, iPr isopropyl, Tf trifluoromethanesulfonyl, Cz carbazolyl.

Conjugated 1,3-enynes are important structural units in organic synthesis, serving as precursors for polysubstituted aromatic rings, biologically active molecules, and organic materials16,17,18,19. Although many strategies have been developed for the construction of 1,3-enynes20,21,22,23,24,25,26,27,28,29, the direct coupling of alkynes to 1,3-enynes via a radical relay pathway has been rarely reported30,31,32,33,34. In this context, Wu30, Koenigs31, and others32 have reported the Pd-catalyzed dicarbofunctionalization of terminal alkynes using secondary or tertiary alkyl iodide, including radical Mizoroki−Heck reactions for the synthesis of 1,3-enynes. Nishikata and co-workers demonstrated a copper-catalyzed tandem radical addition/Sonogashira-coupling process using alkyl bromides and terminal alkynes33. However, the limitation of the above processes stems from the use of the same alkyne in both steps of the tandem process, which ultimately results in the installation of two identical substituents on the enyne motifs. To increase the diversity of 1,3-enynes, the Koh group made a breakthrough using structurally different bromoalkynes as the coupling partners to achieve the 1,2-carboalkynyaltion of terminal alkynes34. Despite these impressive advances, the development of a catalytic version of two similar terminal alkynes to realize highly chemo- and regioselective control is still challenging. Moreover, despite the recent enormous development of radical asymmetric catalysis35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52, the enantioselective radical difunctionalization of two different terminal alkynes with simple radical precursors to give the axially chiral 1,3-enyne frameworks53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82 remains unexploited (Fig. 1b). In addition to innate reactions of terminal alkynes and alkyl halides, such as Glaser-homocoupling83,84,85, halogen atom transfer reaction22,86,87,88,89, and Sonogashira-coupling90,91,92, this reaction suffers from additional challenges: (1) how to achieve the chemoselectivity between vinyl radical I and II, which are generated from the two terminal alkynes93,94,95,96,97 (Fig. 1b); (2) how to ensure the chemo-, and stereoselectivity in the coupling of vinyl radical with the Mn+1L*–acetylide complex (Fig. 1b, yne-Ar and yne-R1). Therefore, developing a new strategy to synthesize axially chiral 1,3-enynes with high chemo-, regio-, and stereoselectivity is urgently needed. To address these challenges, we expected to chemoselectively control the vinyl radical generation to give radical I by regulating the electronic properties of two different terminal alkynes (Fig. 1b, route a). We anticipated that our recently developed copper catalysts with steric crowded chiral multidentate anionic ligands91,92,98,99,100,101,102,103,104,105,106, could selectively promote deprotonation of alkyne with smaller substitute R1 to generate less steric crowded Mn+1L*–acetylide complex yne-R1. At the same time, the thus-generated yne-R1 intermediate with low steric congestion might also prefer to trap vinyl radical I, therefore not only controlling the desired asymmetric process91,98,99,100,101,102,103 but also overcoming the readily occurring side reactions to afford axially chiral 1,3-enynes with high chemo-, regio-, and stereoselectivity.

Herein, we report a copper-catalyzed three-component asymmetric radical 1,2-carboalkynylation using two different terminal alkynes, providing a straightforward route to a diverse array of axially chiral 1,3-enynes (Fig. 1c). This protocol exhibits excellent chemo-, regio-, and stereoselectivity under mild conditions, featuring a broad substrate scope of readily available terminal alkynes and high functional group tolerance, enabling the synthesis of a variety of axially chiral 1,3-enynes. Furthermore, the reaction could be scaled up, and the resulting axially chiral 1,3-enynes could be easily transformed into a series of useful axially chiral 1,3-enyne building blocks.

Results

Reaction development

To examine our aforementioned proposal, aryl alkyne S1 with a large 2-substitutent was selected as the model coupling alkyne, therefore hoping not only to generate the conjugation-stabilized vinyl radical but also to introduce axial chirality. Alkyl alkyne S16, which lacked conjugation effect and would generate thermodynamically unfavorable vinyl radical, was used as the nucleophile to construct 1,3-enyne skeletons. Also, copper salts with chiral ligands would undergo single-electron transfer (SET) with tert-butyl α-bromoisobutyrate C1 to initiate this reaction, giving rise to the alkyl radical. At first, a screening of bidentate neutral ligands, such as bisoxazoline-based ligand L*1 and phosphinooxazoline ligand L*2, with CuTc as the catalyst, indicated no reaction activity, presumably due to the insufficient reducing capability of the copper complex (entries 1–2; Table 1). Thus, we tested chiral anionic oxazoline-derived N,N,N-ligand L*3, which have shown enough reducing capability of copper catalyst to reduce tertiary alkyl bromide in our previous report104,105. Indeed, the desired product 1 was obtained in 28% yield, albeit with low enantioselectivity (10% e.e.) along with the homocoupling side product 1’ in a higher yield than 1, and 14% yield of atom transfer side product 1” (entry 3, and also see Supplementary Table 3), demonstrating the competence of the N,N,N-coordination manifold in promoting the desired reaction. Interestingly, the reaction using the cinchona alkaloid-derived tridentate picolinamide L*4106 still suffered from alkyne homocoupling. To our surprise, tridentate N,N,P-tridentate ligand L*591,107, gave the desired product 1 in 56% e.e., albeit with low yield while inhibiting the alkyne homocoupling process (entry 5). Encouraged by this finding, we next examined the other steric bulk N,N,P-tridentate ligand L*6 and found that the reaction provided product 1 in 77% e.e. but the reaction efficiency remained low (entry 6). Accordingly, we next sought to install a sterically bulky 3,5-disubstituted phenyl ring in L*7, and the e.e. increased to 87% (entry 7). Replacing the key chiral skeleton with cyclohexylenediamine of L*8 significantly improved the yield of 1 to 75% with enhanced enantioselectivity (89% e.e.) with excellent chemo- and regioselectivity. Collectively, these results indicated that the use of bulky N,N,P-tridentate ligand L*8 is crucial for not only selectively generating less steric CuIL*–acetylide complex through deprotonation of smaller-sized alkyne S16 but also effectively interacting vinyl radical I with thus-generated CuIL*–acetylide complex due to low steric congestion, which would exert effective cross-coupling to give 1 while inhibiting side self-coupling and atom transfer processes. Further optimization of conditions, including copper salts, bases, solvents, reaction time and other factors (see Supplementary Tables 17 for the results of condition screening), identified the optimal conditions (entry 9) as follows: S1 (2.0 equiv.), C1 (2.0 equiv.) and S16 (0.05 mmol, 1.0 equiv.) in the presence of CuTc (10 mol%), L*8 (12 mol%), and Cs2CO3 (5.0 equiv.) in Et2O at 10 °C for 5 days under argon. This condition provided 1 in 72% yield and 93% e.e. with excellent chemo- and regioselectivity.

Table 1 Effect of chiral ligands

Substrate scope

Considering the high utility of axially chiral phenol compounds108,109, we first investigated the reaction with 2-oxyaryl alkynes, and those bearing a wide range of functional groups with different electronic properties were readily accommodated in this reaction. As for the scope of 2-oxyaryl alkynes, a series of 6-substituted naphthyl rings were readily accommodated in this reaction (25; Fig. 2a). However, 6- and 7-phenyl-subsituted aryl alkynes suffered from a low conversion to afford the desired products 3 and 6 in low yield, likely due to steric hindrance of the phenyl substituent. Furthermore, various 4- and/or 5-substituted naphthyl rings underwent the reaction smoothly, delivering the corresponding products 79 in 62−70% yield with 91−92% e.e. Likewise, good tolerance of 5- and/or 6-substitution of 2-oxyphenyl alkyne substrates was also observed (10 and 11). In addition, 4,6-disubstituted 2-oxyphenyl alkyne was also applicable to this reaction, affording good enantioselectivity and high yield (12). As for the other ortho-oxygen-substituted alkynes, such as carboxylic, proved to be workable in this reaction, providing the desired products (13 and 14) with high enantioselectivity and moderate yield. However, for ortho-nitrogen-substituted groups like pivaloyamine, the corresponding products 15 was obtained with decreased enantioselectivity under the standard reaction conditions. To further improve enantioselectivity, after systematic optimization efforts (see Supplementary Table 8 for condition optimizations), we found that the more sterically hindered cinchona alkaloid-derived N,N,P-L*11 was able to control enantioselectivity, affording the final product 15 with 70% yield and 86% e.e. (Fig. 2b). As for the ortho-carbon-based alkynes, such as phenyl (S28) and isopropyl (S29), the desired products P1 and P2 were obtained with low yield and e.e. (see Supplementary Fig. 4). These results demonstrated the crucial role of the oxygen- and nitrogen-containing functional groups at the ortho positions. The ortho-(O)PPh2 substituted alkyne S30 failed to afford the desired product possibly due to the catalyst poisoning by the diphenyl phosphine oxide group (see Supplementary Fig. 4).

Fig. 2: Substrate scope for 2-substituted aryl alkynes.
figure 2

a Substrate scope for 2-oxyaryl alkynes. Standard reaction conditions: 2-oxyaryl alkynes (2.0 equiv.), C1 (2.0 equiv.), S16 (0.20 mmol, 1.0 equiv.), CuTc (10 mol%), L*8 (12 mol%), and Cs2CO3 (5.0 equiv.) in Et2O (4.0 mL) at 10 °C for 5 d under argon. b ortho-N-substituted group. Standard reaction conditions: 2-aminoaryl alkyne S15 (2.0 equiv.), C1 (2.0 equiv.), S16 (0.20 mmol, 1.0 equiv.), CuTc (10 mol%), L*11 (12 mol%), and Cs2CO3 (5.0 equiv.) in Et2O (4.0 mL) at r.t. for 5 d under argon. bConducted at r.t. Isolated yield was shown; E.e. values were based on chiral HPLC analysis; FG functional group, Tol tolyl, Me methyl.

We then turned our attention to the scope of radical precursors (Fig. 3a). Numerous α-bromo alkyl and aryl esters including ethyl (C2), cyclopentyl (C3), 4-methoxybenzyl (C4), and tolyl (C5) were effective for the reaction to deliver the corresponding products 1619 with good efficiency and stereoselectivity. Furthermore, the cyclic α-bromoesters (C6) also participated readily in the reaction, producing 20 in moderate yield and excellent enantioselectivity. In addition, tertiary alkyl bromides with α-amide functional group exhibited good reactivity to produce 21 in 78% yield with 99% e.e. It’s worth noting that the Weinreb amide-type bromide C8 was also compatible with the reaction, affording 22 in 30% yield with 92% e.e. However, tertiary α-bromo ketone failed to give the desired product (see Supplementary Fig. 5). In addition, primary and secondary alkyl bromides, such as (bromomethyl)benzene, ethyl 2-bromoacetate, 2-bromoacetonitrile, (3-bromoprop-1-yn-1-yl)trimethylsilane, (bromomethylene)dibenzene, and diethyl 2-bromomalonate as well as Togni reagent, and tosyl chloride were not suitable for the reactions (see Supplementary Fig. 5 for more details). We then proceeded to investigate the scope of alkyne nucleophiles. Regarding the alkyne scope, many alkyl alkynes with aliphatic chains or a variety of functional groups, such as ester (23), ether (24), benzyloxy (25), p-toluenesulfonyloxy (26), carbamate (27) as well as phenylamine (28), underwent the reaction smoothly to generate 2328 in moderate to good yield with excellent chemo-, regio- and enantioselectivity. Notably, the low yield of 2527 is due to the low conversion of the corresponding alkyl alkynes.

Fig. 3: Substrate scope for radical precursors and alkyne nucleophiles.
figure 3

a Scope of radical precursors and alkyl alkyne nucleophiles. Standard reaction conditions: S1 (2.0 equiv.), S1622 (0.20 mmol, 1.0 equiv.), radical precursor C18 (2.0 equiv.), CuTc (10 mol%), L*8 (12 mol%), and Cs2CO3 (5.0 equiv.) in Et2O (4.0 mL) at 10 °C for 5 d under argon. b Scope of aryl alkyne nucleophiles. Standard reaction conditions: S1 (2.0 equiv.), S2327 (0.20 mmol, 1.0 equiv.), C1 (2.0 equiv.), CuTc (10 mol%), L*8 (12 mol%), and Cs2CO3 (5.0 equiv.) in Et2O (4.0 mL) at 10 °C for 5 d under argon. Isolated yield was shown; E.e. values were based on chiral HPLC analysis. Et ethyl, PMB 4-methoxybenzyl, Ac acetyl, Bn benzyl, Ts p-toluenesulfonyl, Ph phenyl, Boc t-butoxycarbonyl.

Unlike alkyl alkynes, aryl alkynes appeared to be more similar to S1 in chemical properties, which should pose a greater challenge in controlling the chemo-, regio-, and stereoselective. Based on radical polarity properties96,97, we proposed that tuning the electronic properties of two similar terminal alkynes may influence the chemoselectivity of the radical addition, in which in-situ generated electrophilic tertiary alkyl radical from α,α-dimethyl-α-bromoesters would easily prefer to attack aryl alkyne S1 than more electron-withdrawing aryl alkynes. In order to realize such transformation with excellent selectivity, aryl alkynes bearing diverse electron-withdrawing groups (chloride, methoxycarbonyl, trifluoromethyl, and cyano) on the phenyl rings were selected as the nucleophile. Indeed, the reaction worked smoothly to afford the corresponding products 2932 in 40–58% yield with 80–83% e.e. (Fig. 3b). Furthermore, heteroaryl alkyne featuring pyridine (33) was suitable substrate to give the desired product in 40% yield with 84% e.e. These results demonstrated the importance of radical polarity to achieve asymmetric multicomponent process involving two similar terminal aryl alkynes with high chemo-, regio-, and stereoselectivity.

Synthetic utility

To demonstrate the synthetic potential of these axially chiral 1,3-enynes products, we first carried out gram-scale reactions with 2-oxyaryl alkyne substrate S1 (Fig. 4a) and still obtained good yield and excellent enantioselectivity (1.96 g, 61% yield, and 92% e.e.). Next, we examined their thermal stability and observed marginal racemization up to 70 °C (see Supplementary Table 9 for more details). As is well known, axially chiral styrene compounds have proven to be a valuable chiral backbone for developing chiral catalysts or ligands in asymmetric catalysis59,64,77,110,111,112,113. To our delight, product 1 could be efficiently converted to corresponding triflate 34 in 41% overall yield with 92% e.e. (Fig. 4b), which has the potential to be a valuable partner in transition metal-catalyzed coupling reactions for constructing variety of axially chiral 1,3-enyne building blocks. Likewise, the configurational stability of product 33 was examined, and only marginal racemization was observed up to 80 °C (see Supplementary Table 10 for more details). The trivalent phosphine-based 35, of which had the potential for asymmetric catalysis, could be easily synthesized by Ni-catalyzed coupling of 34 with Ph2PH at 80 °C in 65% yield and 91% e.e. (Fig. 4c, left)114. Furthermore, Pd-catalyzed coupling of 34 with Ph2P(O)H at 80 °C provided the diphenyl phosphine oxide product 36 in a moderate yield with 90% e.e. (Fig. 4c, right)115. In addition, the absolute structure of product 1b (Fig. 4b, and also see Supplementary Fig. 1) was determined to be Ra by X-ray structural analysis. Finally, we performed a preliminary investigation on the application of axially chiral compounds 35 and 36 as ligand in the Pd-catalyzed asymmetric cross-coupling116. As shown in Fig. 4d, diphenylphosphine 35 could catalyze the Suzuki−Miyaura cross-coupling of 37 and 38 as a potential ligand, affording the coupling product 39 with moderate yield and enantioselectivity, while diphenyl phosphine oxide 36 failed to afford the coupling product. These results demonstrate that the axially chiral 1,3-enyne scaffolds are promising for developing a new class of chiral monophosphine ligand.

Fig. 4: Synthetic utility for the construction of valuable axially chiral reagents.
figure 4

a Gram-scale reactions. b, c Transformation of enantioenriched axially chiral 1,3-enynes. d Chiral styrenes 35 and 36 as potential ligand in the Pd-catalyzed asymmetric cross-coupling. aConditions: (i) LiAlH4, THF, 0 °C, Ar, 4 h; (ii) MeI, NaH, THF, 0 °C, Ar, 4 h; (iii) DIBAL-H, DCM, ‒78 °C to r.t., Ar, 4 h; (iv) Tf2O, pyridine, DCM, 0 °C to r.t., Ar, 4 h. bConditions for 35: Ni(COD)2, dppf, HPPh2, Na2CO3, 1,4-dioxane, 80 °C, Ar, 3 d. cConditions for 36: Pd(OAc)2, dppb, HP(O)Ph2, DIPEA, DMSO, 80 °C, Ar, 3 d. LiAlH4 lithium aluminum hydride, THF tetrahydrofuran, DCM dichloromethane, DMSO dimethyl sulfoxide, COD 1,5-cyclooctadiene, DIBAL-H diisobutylaluminium hydride, dppf 1,1’-bis(diphenylphosphino)ferrocene, dppb 1,4-bis(diphenylphosphino)butane, DIPEA N,N-diisopropylethylamine, HPPh2 diphenylphosphine, HP(O)Ph2 diphenylphosphine oxide.

Mechanistic investigations

Control experiments conducted in the absence of the copper salt, chiral ligand, or base additive confirmed that all of these components were indispensable for the reaction (see Supplementary Table 11 for more details). The preliminary results of deprotonation experiments suggested that the less steric alkyl alkyne S16 was consumed more rapidly than S1 (see Supplementary Figs. 2 and 3), which might probably indicate the sterically bulky CuIL*8 complex would favor deprotonation of less steric alkyne (Fig. 5c). In the presence of either BHT or TEMPO, the reaction was completely inhibited (Fig. 5a). Additionally, the BHT-trapped product 40 was isolated in 77% yield and the TEMPO adducts could be detected by high-resolution mass spectrometry (Fig. 5a, and also see Supplementary Fig. 6). Moreover, introducing an additional radical trapper, such as phenyl diselenide, resulted in the formation of vinyl radical trapped product 41 in 99% yield with 0% e.e. (Fig. 5a)82. These radical inhibition experiment results indicated the formation of an alkyl radical, which preferentially added to aryl alkynes S1, generating the conjugation-stabilized vinyl radical IV (Fig. 5c). An additional control experiment with alkenyl bromide 1” failed to produce any product 1 under the standard conditions (Fig. 5b), indicating that a tandem atom-transfer radical addition/cross-coupling reaction pathway21 is unlikely.

Fig. 5: Mechanistic experiments and proposal.
figure 5

a Radical inhibition experiments with BHT, TEMPO, and phenyl diselenide. b Control experiments of vinyl bromide with S16. c Mechanistic proposal for the formation of axially chiral 1,3-enynes. aThe yield of 41 was calculated based on the amount of phenyl diselenide. BHT butylated hydroxytoluene, TEMPO 2,2,6,6-tetramethyl-1-piperinedinyloxy.

Based on these results as well as previous reports90,91,92,98,99,100,101,102,103, we propose a plausible reaction mechanism, as shown in Fig. 5c. Copper(I) salt first coordinated with L*8 to form the catalytically active species CuIL*8, entering the catalytic cycle. Then, complex CuIL*8 selectively reacted with smaller-sized terminal alkyne S-R1 to generate the less steric CuIL*8–acetylide complex I-R1 in the presence of a base91. Afterwards, complex I-R1 underwent single-electron transfer (SET)98 with alkyl bromide, giving rise to the CuIIL*8–acetylide complex II-R1 and alkyl radical III. The following selectively intermolecular addition of III to the triple bond of alkyne S-Ar afforded the conjugation-stabilized vinyl radical IV22. Next, radical IV interacted with complex II-R1 to deliver the desired axially chiral 1,3-enyne products 133 and regenerated the CuIL*8 complex in the presence of the highly sterically demanded ligand L*8. However, we do not have enough evidence to support the proposed process with excellent control of chemo-, regio-, and stereoselectivity and are currently performing more experimental and theoretical studies to disclose the detailed mechanism.

Discussion

In sum, we have demonstrated a three-component asymmetric radical 1,2-carboalkynylation with two different terminal alkynes and diverse alkyl radical precursors, which could successfully access axially chiral 1,3-enynes with excellent chemo-, regio-, and stereoselectivity under copper catalysis. The key to success lies not only in using the sterically bulky chiral multidentate anionic N,N,P-ligand but also in modulating alkynes with significantly different electronic properties to tune selectivity. This process features good substrate scope and functional group tolerance of terminal alkynes and radical precursors and readily affords an abundance of valuable enantioenriched axially chiral 1,3-enynes, thus providing a robust platform for expedient access to a myriad of chiral styrene building blocks. These results highlight the great potential of strategically devised multidentate anionic ligands for controlling the asymmetric multicomponent radical reactions with excellent selectivity.

Methods

Representative procedure for axially chiral 1,3-enynes

An oven-dried resealable Schlenk tube equipped with a magnetic stir bar was charged with 2-substituted aryl alkynes (0.40 mmol, 2.0 equiv.), alkyne nucleophiles (0.20 mmol, 1.0 equiv.), CuTc (3.8 mg, 0.02 mmol, 10 mol%), L*8 (17.6 mg, 0.024 mmol, 12 mol%), and anhydrous Cs2CO3 (325.8 mg, 1.00 mmol, 5.0 equiv.). The tube was evacuated and backfilled with argon three times. Then Et2O (4.0 mL) was added by syringe under argon atmosphere. Finally, radical precursor (0.40 mmol, 2.0 equiv.) was added into the mixture and the reaction mixture was stirred at 10 °C for 5 d. Upon completion of the reaction (monitored by TLC), the reaction mixture was filtered through a short pad of Celite and washed with EtOAc. The filtrate was concentrated to afford the crude product, which was purified by column chromatography on silica gel to afford the desired product.