Introduction

Axially chiral scaffolds are widely found in natural products, chiral catalysts, ligands, and materials1,2,3,4,5. Over the past decades, great efforts have been made for the synthesis of atropisomers featuring abundant skeletons, and biaryl derivatives are undoubtedly the most widely studied6,7,8,9,10,11,12,13. Recently, several challenging skeletons of axially chiral compounds have attracted considerable interest. For example, axially chiral styrenes are one challenging family of chiral compounds due to their relatively lower rotational barrier between different atropisomers compared to their diaryl counterparts14,15,16 (Fig. 1a). Despite significant achievements in the field, the investigations of axially chiral alkenes have focused on the C − C and C-N stereogenic axes. In this context and as part of our continued research interest in the C-B axial chirality17,18,19,20, we aimed to develop an alkenylboron scaffold based on the 1,2-azaborine units for expending the structural diversity of axial chirality (Fig. 1a). The interesting structure, derived from the replacement of the C = C bond of all-carbon aromatic rings with a B-N bond, can maintain the aromatic character and show great potential in functional materials, ligands and medicinal chemistry21,22,23,24,25,26,27 (Fig. 1b). For the designed alkenylboron atropisomers, their synthesis may be more challenging than the counterpart styrene atropisomers owing to the lower rotational barrier resulting from the fact that the Csp2-B bond is longer than the Csp2-Csp2 bond28,29,30,31.

Fig. 1: Designed nickel-catalyzed atroposelective radical alkylarylation of ethynyl-azaborines.
figure 1

a The design of C-B axially chiral alkenylborons. b Isoelectronic relationship between C = C and B-N bonds. c Diverse transformations via α-boryl carbon radical intermediates. d This work: nickel-catalyzed atroposelective relayed reductive coupling via α-borylvinyl radical. DPBNA = diphenyl-substituted 1,5-diaza-2,6- diboraanthracene.

Recently, tremendous progress has been made in the transition metal-catalyzed difunctionalization of alkynes via an aryl-stabilized vinyl radical intermediate, which serves as a powerful strategy to access multi-substituted alkenes32,33,34,35,36,37,38,39,40,41,42. Noteworthily, such a strategy has been successfully applied for the construction of axially chiral styrenes by Zhang43 and Liu44 groups respectively. Considering the feasibility of the process, we envisioned that an atroposelective difunctionalization of ethynyl-azaborines through an azaborine-stabilized vinyl radical intermediate could be an efficient approach to construct the designed C-B axially chiral alkenylborons. However, compared to the extensive studies on the α-borylalkyl radicals45,46,47,48,49,50 (Fig. 1c, left), the research on α-borylvinyl radicals is very rare and undeveloped51, which seems to totally be ignored by chemical community (Fig. 1c, right). Although α-borylvinyl radicals can theoretically serve as a fascinating intermediate to construct multi-substituted alkenylborons, only one example from iron-catalyzed radical addition of ethynylboronic acid pinacol ester involves such radical intermediate so far51. Herein, we report an atroposelective radical alkylarylation of ethynyl-azaborines to enable the axially chiral alkenylborons (Fig. 1d). Mechanistically, the catalytic protocol involves alkyl radical addition to the ethynyl moiety and cross-coupling of the resulting α-borylvinyl radical with nickel complex. This chemistry enables a straightforward access to the challenging axially chiral alkenylborons under mild conditions in generally high enantioselectivity and excellent stereoselectivity.

Results

Reaction conditions optimization

To overcome the above-mentioned challenges, we firstly designed and synthesized the sterically hindered ethynyl-azaborine 1a52,53,54. Then, ethynyl-azaborine 1a, 3-iodoanisole (2a) and tert-butyl iodide (3a) were employed to investigate the envisaged atroposelective radical relayed reductive coupling with NiBr2·DME as catalyst and tetrakis(dimethylamino)ethylene (TDAE) as reductant. To our delight, when pybox L1 was used as chiral ligand, the reaction could proceed smoothly (Table 1, entry 1), delivering the desired axially chiral alkenylboron 4a in 39% NMR yield and −30% enantiomeric excess (ee). The preliminary result encouraged us to find a more efficient ligand for this transformation. Other pybox ligands L2-L5 could efficiently promote this transformation in moderate enantioselectivities (Table 1, entries 2-5), and chlorine-substituted pybox ligand L4 was determined to be the best one (Table 1, entry 5, 48% yield and 86% ee). Chiral quinolin-2-yl pybox L6, bisoxazoline ligand L7, and diamine ligand L8 proved largely ineffective in this reaction (Table 1, entries 6-8). The mixed solvent (2-MeTHF/DCE) could slightly improve the yield and enantioselectivity (Table 1, entry 9), and the better enantioselectivity (89% ee) was obtained through decreased reaction temperature (Table 1, entry 10). The optimal reaction condition was furnished when MgCl2 was used as an additive (Table 1, entry 11, 72% yield and 92% ee). In addition, other Ni catalysts, such as NiCl2·DME and NiBr2, afforded inferior results (Table 1, entries 12 and 13).

Table 1 Optimization of reaction conditionsa

Substrate scopes

With the optimized asymmetric radical relayed reductive coupling conditions in hand, we first evaluated the substrate scope of aryl iodides (Fig. 2). Generally, aryl iodides bearing electron-donating, -neutral and -withdrawing substituents proceeded smoothly under the standard conditions, furnishing the corresponding axially chiral alkenylborons 4a-4v with good to excellent enantioselectivities (83–93% ee). The lower to moderate yields for some axially chiral alkenylboron products were mainly due to the hydroalkylation process of ethynyl-azaborines51,55. The reaction showed good halogen compatibility, not only ethynyl-azaborine 1a but also aryl iodines (2c, 2i-2l and 2o), which offers valuable handles for further late-stage functionalization. Importantly, sensitive functional groups on the benzene ring, such as cyano (4l and 4m), aldehyde (4n), ester (4o-4s), alkene (4r), ketone (4t), trifluoromethyl (4u), and trifluoromethoxy (4v), also worked well. The absolute configuration of 4m was determined by X-ray crystallographic analysis. Notably, aryl iodides derived from L(-)-borneol, geraniol and gemfibrozil were also well compatible, and furnished the desired three-component coupling products 4q4s in moderate yields with good to excellent diasteroselective (4q, >20:1 dr) and enantioselectivities (89% ee and 92% ee). 2-Naphthalene iodine is also a good substrate for this atroposelective radical relayed reductive coupling reaction (4w). Heterocyclic derivatives, such as thiophene and pyridine, were also suitable candidates for this reductive coupling to give axially chiral alkenylborons 4x and 4y with good to excellent enantioselectivities (82% and 91% ee).

Fig. 2: Substrate scope of aryl iodides.
figure 2

Reaction conditions: 1a (0.1 mmol, 1.0 equiv), 2a (1.3 equiv), 3a (3.0 equiv), NiBr2·DME (10 mol%), ligand (10 mol%), TDAE (2.1 equiv), MgCl2 (1.0 equiv) in 1.0 mL of 2-MeTHF and 0.1 mL of DCE under argon at −5 oC for 24 h.

Next, the scope of ethynyl-azaborines in this protocol was also examined (Fig. 3). 1,2-benzazaborine ring of ethynyl-azaborines bearing alkyls, halides, and phenyl could all undergo this atroposelective coupling reaction to deliver the corresponding axially chiral alkenylborons 4z-4ae in 55%–74% yields with 89%–94% ee. Notably, when the bromo group on ethynyl-azaborine was replaced by phenyl group, the reaction also performed smoothly under the standard conditions, providing the corresponding product 4af in moderate yield and good enantioselectivity (41% yield, 83% ee). The installations of isopropyl or phenyl to the nitrogen atom had little effect on the enantioselectivities (4ag and 4ai, 89% and 92% ee) for this protocol. Of note, further reduction of the steric hindrance of ethynyl-azaborine would decrease the enantioselectivity of the target product (4ah, 88% yield, 81% ee). Finally, the reaction of other unactivated tertiary alkyl iodides also worked well to deliver the desired axially chiral alkenylboron products 4aj-4al in 42%–61% yields with 84%–87% ee.

Fig. 3: Substrate scope of ethynyl-azaborines and alkyl iodides.
figure 3

Reaction conditions: 1a (0.1 mmol, 1.0 equiv), 2a (1.3 equiv), 3a (3.0 equiv), NiBr2·DME (10 mol%), ligand (10 mol%), TDAE (2.1 equiv), MgCl2 (1.0 equiv) in 1.0 mL of 2-MeTHF and 0.1 mL of DCE under argon at −5 oC for 24 h.

Downstream applications and transformations

The potential value of this reaction was further illustrated by synthetic transformations (Fig. 4). The resulting axially chiral alkenylboron products have been pre-installed with a bromo group, providing a handle for later functionalizations, which could undergo Negishi and Sonogashira coupling to achieve alkylation, arylation and alkynylation (5, 6 and 7). In addition, the new chiral (P, olefin) ligand 8 was successfully synthesized via in situ axially chiral lithium intermediate with high retention of the enantiopurity.

Fig. 4: Synthetic transformations.
figure 4

a Pd-catalyzed Negishi cross-coupling of 4a and alkyl zinc reagent. b Pd-catalyzed Negishi cross-coupling of 4a and aryl zinc reagent. c Pd-catalyzed Sonogashira cross-coupling of 4a and aryl alkyne. d Phosphonation reaction of 4a via in situ axially chiral lithium intermediate.

Mechanism investigations

To elucidate the radical process of this reductive coupling, control experiments were carried out (Fig. 5a). Firstly, TEMPO as a radical scavenger was added to the three-component reaction, and the formation of coupling product 11 was completely inhibited. Meanwhile, the addition of 1,1-diphenylethylene almost interrupted this reductive coupling reaction, whereas affording compound 13, which may be derived from the addition of tert-butyl radical to 1,1-diphenylethylene. Fortunately, when the reaction of ethynyl-azaborine 10 was carried out in the presence of BrCCl3, the corresponding vinyl-CCl3 adduct (14) was detected by HRMS. Notably, the reaction of ethynyl-azaborine 10 with a smaller steric hindrance (ethyl group on N atom) in the absence of aryl iodide gave double addition products 15a and 15b as well as self-coupling product 15c. Moreover, hydrogenation of α-borylvinyl radical intermediates was detected by GC-MS in the process of substrate investigation. These experiments suggest that an α-borylvinyl radical is involved in this coupling reaction. Then, several control experiments with nickel species were also carried out (Fig. 5b). Firstly, the stoichiometric reaction of aryl iodide, Ni(COD)2 and 4,4’-dtbbpy furnished Ar-Ni(II)-I complex 16. Only trace amount of three-component coupling product 17 was obtained by mixing complex 16, ethynyl-azaborine 9 and tert-butyl iodide, while more obvious product 17 could be observed when reductant TDAE were added. Moreover, when the catalytic amount of the complex 16 was employed in the cross-over reaction with 3-iodoanisol, compounds 11 and 17 were obtained in 75% yield and 6% yield, respectively. Overall, these results suggest that the reduction of complex 16 is necessary for this catalytic cycle.

Fig. 5: Mechanistic experiments.
figure 5

a Verification of the radical pathway. b Control reaction with Ar-Ni(II)-I complex 13. TEMPO = 2,2,6,6-tetramethyl-1-piperinedinyloxy.

Based on the results of the above experiments and previous literatures43,56,57,58, we proposed a Ni(I)/Ni(II)/Ni(III) cycle for the asymmetric radical relayed reductive coupling (Fig. 6, left): the chiral Ni(I) species A generated in situ undergoes oxidative addition into the aryl iodine 2a to afford the Ar-Ni(III)L* intermediate B, which is reduced by TDAE to produce the Ar-Ni(I)L* complex C. The activation of alkyl iodide 3a by the Ar-Ni(I)L* complex C generates an alkyl radical D and a Ar-Ni(II)L* species E. Alkyl radical D undergoes regioselective addition to ethynyl-azaborine 1a to afford the α-borylvinyl radical F, which can combine with Ar-Ni(II)L* species E to deliver the Ni(III) intermediate G. The excellent E-stereoselectivity of alkene may stem from steric hindrance32,33,34,35,36,37,38,39,40. The final reductive elimination of the Ni(III) intermediate G yields axially chiral alkenylboron 4a and regenerates the chiral Ni(I) catalyst. However, Ni(0)/Ni(I)/Ni(II)/Ni(III) catalytic cycle can’t completely rule out (Fig. 6, right), which involves the oxidative addition of Ni(0) species to aryl iodides.

Fig. 6: Possible reaction mechanism.
figure 6

Possible reaction mechanism about Ni(I)/Ni(II)/Ni(III) catalytic cycle and Ni(0)/Ni(I)/Ni(II)/Ni(III) catalytic cycle. L = ligand.

In conclusion, we have developed a nickel-catalyzed atroposelective radical relayed reductive coupling reaction, leading to the formation of the challenging C-B axially chiral alkenylborons. The method features mild conditions, high enantioselectivity and excellent stereoselectivity. The key mechanistic feature of the process is the generation and transformation of a boron-stabilized vinyl radical. It is expected that the concept obtained here will encourage the development of more asymmetric transformations around the multifunctional α-borylvinyl radical intermediates.

Methods

General procedure for the synthesis of C-B axially chiral alkenylborons

In a glove box, a 10 mL Schlenk tube was charged with NiBr2·DME (0.01 mmol, 10 mol%), ligand (0.01 mmol, 10 mol%), the bromination of B-ethynyl-2,1-borazaronaphthalene (0.1 mmol, 1.0 equiv), aryl iodide (0.13 mmol, 1.3 equiv), alkyl iodide (0.3 mmol, 3 equiv), MgCl2 (0.1 mmol, 1.0 equiv), 2-Me-THF (1 mL) and DCE (0.1 mL). The resulting mixture was stirred at room temperature for 1 min. Then TDAE (0.21 mmol, 2.1 equiv) was added. The reaction tube was taken out of the glove box and reacted at −5 oC for 24 h. Upon completion, proper amount of silica gel was added to the reaction mixture. After removal of the solvent, the crude reaction mixture was purified on silica gel to afford 4.