Merging C–H and C–C Activation in Pd(II)-Catalyzed Enantioselective Synthesis of Axially Chiral Biaryls

Introduction

One of the ultimate goals of organic synthesis is the efficient, selective, and atom-economical construction of molecular complexity from simple and readily available feedstocks. In this context, the direct transformations of carbon–hydrogen (C–H) and carbon–carbon (C–C) bonds, main constituents of organic compounds, are highly important. Over the past decades, transition-metal-catalyzed C–H activation has been established as an effective and straightforward synthetic strategy for building molecular complexity.^1–12 Similar to C–H activation, transition-metal-catalyzed C–C activation, which enables direct editing of the molecular carbon skeleton, has also been identified as an attractive approach for molecular editing.^1–19 Despite the undisputed significance of both strategies, these two cutting-edge synthetic techniques almost evolved separately and tandem C–H/C–C functionalizations continue to be underdeveloped. The key challenge of this tandem transformation lies in ensuring these two approaches do not interfere with each other. Recently, increasing interests among the synthetic community have been attracted to merging C–H functionalization and C–C bond cleavage into a single chemical process, which opens a new avenue for the synthesis of complex molecular architectures that are otherwise difficult to access.^20–22

Three-membered cyclopropane motifs, a class of useful synthetic entities, are well-known applicable building blocks for C–C bond activation.^23–25 Since 2010, cyclopropanes, such as vinylcyclopropanes (VCPs)^26,27 and cyclopropanols,^28,29 have emerged as versatile synthons for C–H/C–C activation reactions. Generally, two different strategies based on the pathways of C–C cleavage in the catalytic system could be expected for the activation of cyclopropanes. One strategy is the merging of C–H/C–C bond cleavage in the same catalytic cycle (Scheme 1a, path A). The other involves the combination of two distinct catalytic cycles: a C–C cleavage catalytic cycle to generate the coupling partners that merges with the C–H activation catalytic cycle (Scheme 1a, path B). Recent advances highlight the challenging process of the merged catalytic cycles (Scheme 1b). For example, early reports have demonstrated that VCPs were compatible coupling partners, merging C–H/C–C activations with the aid of transition-metal catalysts to provide the allylated products.^30–34 The use of cyclopropanols as useful building blocks for the synthesis of β-aryl ketones via Rh(III)-catalyzed alkylation of (hetero)arenes has also been achieved.^35,36 However, these pioneering works were confined to the use of Mn,³² Ru,³³ Rh,^30,35–39 and Co^31,40 catalysts and generally proceeded by C–H/C–C bond cleavage occurring within the same catalytic cycle. Palladium catalysts, one of the most widely used transition-metal catalysts in C–H activation reactions, have not yet been explored in this emerging field. Of particular note is that, to date, the asymmetric version of tandem C–H/C–C activation reactions remains an extremely rare explored field, probably because of the difficulty in finding a proper catalytic system to control enantioselectivity within the accompanying process of C–H activation and C–C cleavage.^41–45 These limitations encouraged us to develop a new, highly enantioselective tandem C–H/C–C functionalization system using readily available feedstocks via divergent strategies.

• Download figure
• Download PowerPoint

Biaryl atropisomers are ubiquitous in natural products, pharmaceuticals, ligands, and catalysts.^46–50 Their unique architecture and board applications have stimulated the rapid evolution of strategies to access these structures.^51–57 Particularly, transition-metal-catalyzed asymmetric C–H functionalization has emerged as a powerful and economical strategy to access such scaffolds with high enantiopurity.^58–71 Despite cyclopropanes playing unique and important roles in enantioselective ring-opening reactions to access various valuable functionalized targets, the success of using these reactive synthons in asymmetric synthesis is limited to the generation of molecules with point chirality.²⁴ To date, the highly efficient synthesis of axially chiral biaryl compounds via a merged enantioselective C–H functionalization and C–C cleavage of cyclopropanes has not yet been reported. Because of our continuous interests in the synthesis of axially atropisomers via asymmetric C–H activation reactions,^72–74 we herein report the first Pd(II)-catalyzed atroposelective synthesis of axially chiral biaryls by merging C–H/C–C functionalization with cyclopropane derivatives (Scheme 1c).

Experimental Methods

General procedure for Pd(II)-catalyzed atroposelective C–H/C–C activation with VCPs

To an oven-dried 50 mL Schlenk tube were added substrate rac- 1 (0.10 mmol), VCPs 2 (0.25 mmol), Pd(OAc)₂ (2.2 mg, 0.01 mmol), L3 (5.6 mg, 0.03 mmol), benzoquinone (BQ; 10.8 mg, 0.1 mmol), trifluoroacetic acid (TFA; 15 μL, 0.2 mmol), Ag₂O (34.8 mg, 0.15 mmol), and acetic acid (HOAc) (1.0 mL). The mixture was stirred for 30 h at 70 °C followed by cooling. The resulting mixture was filtered through a celite pad and concentrated in vacuo. The residue was purified by preparative TLC using hexane/ethyl acetate (EtOAc) as the eluent to afford a mixture of E/Z diastereomers. A vigorously stirred Raney-Ni catalyst in 0.5 mL anhydrous tetrahydrofuran (THF) was added to the resulting E/Z diastereomers in an oven-dried flask in anhydrous THF (0.5 mL). The reaction vessel was filled with hydrogen atmosphere (H₂ balloon) and vigorously stirred at r.t. The reaction was monitored by TLC. After completion, the solution was diluted with CH₂Cl₂ and filtered through a celite pad. The organic layer was concentrated in vacuo. The residue was purified by preparative TLC using hexane/EtOAc as the eluent to afford the alkylated product 3.

General procedure for Pd(II)-catalyzed atroposelective C–H/C–C activation with cyclopropanols

To an oven-dried 50 mL Schlenk tube were added substrate rac- 1 (0.10 mmol), cyclopropanols 4 (0.3 mmol), Pd(OAc)₂ (2.2 mg, 0.01 mmol), L-tert-leucine (3.9 mg, 0.03 mmol), BQ (10.8 mg, 0.1 mmol), NaOEt (13.6 mg, 0.2 mmol), and 2,2,2-trifluoroethanol (TFE):HOAc = 9:1 (1.0 mL). The mixture was stirred for 30 h at 60 °C, followed by cooling. The resulting mixture was filtered through a celite pad and the organic layer was concentrated in vacuo. The residue was purified by preparative TLC using hexane/EtOAc as the eluent to afford the product 5. More experimental details and characterization are available in the Supporting Information.

Results and Discussion

We envisioned that assembly of VCPs and biaryls with the assistance of palladium catalyst and ligand offers a promising option for the construction of axial chirality in a single step. Notably, such reactions have three key features: First, the inherent strain of VCPs provides the thermodynamic driving force for C–C bond cleavage. Second, the increased rotation barrier by introducing a bulky group via asymmetric C–H functionalization enhances the atropostability. Third, undoubtedly the most challenging, metal palladium must serve as an assembly center of biaryls, VCPs, and chiral ligands to enable the overall transformation. With these considerations in mind, we initiated our investigation with the reaction of rac- 1a with VCP 2a in the presence of 10 mol % Pd(OAc)₂ and 30 mol % L-tert-leucine as a transient directing group (TDG).^75–81 As expected, the reaction proceeded smoothly in HOAc, affording a mixture of E/Z diastereomers. To simplify the separation process, the obtained mixture was hydrogenated to give the alkylation product 3a in 30% yield with 96% ee in a two-step sequence (Table 1, entry 4). The addition of carboxylic acids proved to be critical for the transformation (entries 5–8). TFA significantly improved the yield and slightly increased the enantioselectivity, probably due to the promotion of the recycling of the transient auxiliary and the favorable cleavage of the C–C bond of 2a (entry 8, 55%, 98% ee). After surveying the efficacy of various oxidants, Ag₂O was the most efficient (entry 9, 62%, 97% ee). Ag₂O might play a dual role, functioning as both a co-oxidant and a heteronuclear active species to facilitate C–H cleavage.⁸² Next, we evaluated the effects of chiral amino acids and their derivatives (entries 9–12). We were delighted to find that L-isoleucine-derived diethyl amide L3 was particularly suited for this transformation, giving 3a in 72% yield with 99% ee (entry 11).

Table 1 | Optimization of Reaction Conditions^a

Entry	Solvent	Additive	Ligand	Yield(%)b	eec
1	THF	—	L1	—	—
2	TFE	—	L1	Trace	—
3	1,1,1,3,3,3-Hexafluoro-2-propanol (HFIP)	—	L1	Trace	—
4	HOAc	—	L1	30	96
5	HOAc	AdCO2H	L1	34	96
6	HOAc	PivOH	L1	39	95
7	HOAc	(BnO)2PO2H	L1	47	98
8	HOAc	TFA	L1	55	98
9d	HOAc	TFA	L1	62	97
10d	HOAc	TFA	L2	49	99
11d	HOAc	TFA	L3	72	99
12d	HOAc	TFA	L4	67	99

^aReaction conditions: rac-1a (0.1 mmol), dimethyl 2-vinylcyclopropane-1,1-dicarboxylate 2a (0.25 mmol), Pd(OAc)2 (10 mol %), ligand (30 mol %), BQ (1.0 equiv), solvent (1.0 mL), additive (2.0 equiv), 70 °C, air, 30 h

^bIsolated yield.

^cThe ee value was determined by chiral high-performance liquid chromatograph (HPLC).

^dAg₂O (0.15 mmol).

With the optimal conditions in hand, we next investigated the scope of this reaction. As shown in Table 2, this C–H/C–C activation reaction turned out to be widely applicable for a range of VCPs and biaryls. VCPs with different esters were well tolerated, affording the corresponding products in moderate to good yields with excellent enantioselectivities ( 3b– 3d, 47–77%, 96–99% ee). Various functional groups, such as fluoro, chloro, methyl, and methoxy, on a phenyl moiety (Ar₁) were compatible, giving the desired products with excellent enantioselectivities ( 3e– 3k, up to >99% ee). Pentasubstituted diphenyl 1n also reacted smoothly to afford the desired product 3n in 60% yield with >99% ee. Biaryl 1o containing a pyrenyl group could also be tolerated, giving the desired product 3o in 62% yield and 95% ee. Intriguingly, biaryl aldehyde 1p containing five-membered heteroarene was compatible in this reaction, affording 3p in 78% yield with 95% ee. Moreover, desymmetrization of the proaxial biaryl aldehyde was also achieved, giving 3q in excellent enantioselectivity, albeit with low yield (37%, 95% ee). It is important to note that this method is capable of effecting dynamic kinetic resolution to obtain biaryls bearing substituents at both the 6- and 2′- positions with excellent selectivities ( 3r– 3t). The absolute configuration of 3o was unambiguously confirmed by the single-crystal X-ray analysis, and those of other axially chiral biaryls were assigned by analogy.

Table 2 | Scope of Atroposelective C–H/C–C Activation with VCPs^a

^aReaction conditions: rac- 1 (0.1 mmol), VCP 2 (0.25 mmol), Pd(OAc)₂ (10 mol %), L3 (30 mol %), BQ (1.0 equiv), HOAc (1.0 mL), TFA (0.2 mmol), Ag₂O (0.15 mmol), 70 °C under air for 30 h.

Encouraged by the positive results obtained from the Pd(II)-catalyzed atroposelective C–H/C–C activation with VCPs, we decided to explore the possibility of extending this strategy to access more structurally diverse axially chiral biaryls. We surmised that cyclopropanols could also serve as efficient and reactive partners in our catalytic system because of strain release upon exposure to transition-metal catalysts. After systematic optimizations, we were delighted to find that the reaction of rac- 1a and cyclopropanols 4a occurred smoothly under the following conditions: 10 mol % Pd(OAc)₂, 30 mol % L-tert-leucine ( L1), 1.0 equiv BQ, and 2 equiv NaOEt in TFE/HOAc at 60 °C for 30 h ( Supporting Information Table S5). A mixture of alkylation product 5a and alkenylation product 5a′^a was obtained in 66% total yield ( 5a: 5a′ = 1.7:1; 5a, 98% ee; 5a′, 86% ee). The formation of alkylation and alkenylation products might be ascribed to the competitive pathway between β-H elimination and protodepalladation, in sharp contrast to the results in Rh(III) catalysis.^35,36 The scope of this transformation was then examined (Table 3). The reaction was compatible with different substituents on the phenyl ring, such as methyl, fluoro, and chloro, showing that the electronic properties of substituents do not have obvious influence on the reactivity and selectivity ( 5b/ 5b′– 5f/ 5f′, 58–70% yields, 85–99% ee). Notably, biaryls containing heteroaromatics, such as benzofuran and benzothiophene, were also viable under slightly modified conditions, and only alkylated products were obtained with excellent enantioselectivities ( 5g– 5j). Various other cyclopropanols bearing aryl, alkyl, and heterocyclic substituents were also compatible ( 5k– 5r).

Table 3 | Scope of Atroposelective C–H/C–C Activation with Cyclopropanols^a

^aReaction conditions: rac- 1 (0.1 mmol), cyclopropanol 4 (0.3 mmol), Pd(OAc)₂ (10 mol %), L1 (30 mol %), BQ (0.1 mmol), TFE:HOAc = 9:1 (1.0 mL), NaOEt (0.2 mmol), 60 °C under air for 30 h.

^bWithout NaOEt, HFIP/HOAc (0.8 mL/0.2 mL) as the solvent.

^cWithout NaOEt.

Subsequently, we demonstrated the applications of this protocol by performing gram-scale reactions and synthetic transformations (Scheme 2). First, a 5 mmol scale reaction of rac- 1a and 2a was conducted under standard conditions, and the reaction proceeded smoothly to produce the desired product 3a (74%, 99% ee, 1.56 g). Treatment of 3a with LiBH₄ generated multi-hydroxyl biaryl 6a in 78% yield with 97% ee. Interestingly, chiral biaryl bearing a ten-membered ring was prepared in a two-step sequence ( 6b, 37% overall yield, >99% ee). Moreover, the obtained 3a could also be oxidized successively to give the corresponding axially chiral acid 6c in 81% yield without erosion of the enantioselectivity. To further highlight the potential practicality of this method, another gram-scale reaction was carried out by the reaction of rac- 1v with cyclopropanol 4h. The alkylated product 5k was isolated in 69% yield with 97% ee. Oxidation of 5k provided the carboxylic acid 6d in 68% yield with 97% ee. The abovementioned methods provide a facile synthetic route to enantiopure axially chiral aldehydes and carboxylic acids, which might offer promising opportunities for asymmetric catalysis.^83,84

• Download figure
• Download PowerPoint

To gain insights into the mechanism, several experiments were conducted. First, we performed the reaction of rac- 1a with cyclopropanol 4b or enone 4b′ under the standard conditions, respectively. The desired products 5f and 5f′ were obtained in both reactions with comparable outcomes (Scheme 3a, with 4b: 5f, 48%, 98% ee; 5f′, 22%, 94% ee and with 4b′: 5f, 38%, 99% ee; 5f′, 17%, 90% ee). When a control reaction of rac- 1v with cyclopropanol 4b or enone 4b′ was conducted, similar results were obtained (Scheme 3b, with 4b: 66%, 97% ee and with 4b′: 59%, 92% ee). Meanwhile, when competitive reaction of cyclopropanol 4b and deuterated phenyl vinyl ketone 4c′ with rac- 1v was performed, a mixture of 5k and deuterated 5k was afforded, with the formation of deuterated 5k more favorable (Scheme 3c, 56% yield, 64% D). These results suggested that the enone species might be involved in the reaction (Schemes 3a–3c). To further probe the reaction mechanism, deuterium-labeling studies were carried out using deuterated solvents. As shown in Scheme 3d, 72% deuterium incorporation was observed with 4b, and similarly, 74% deuterium incorporation was observed when using 4b′ as the coupling partner. The control experiment revealed that no deuterium incorporation occurred when 5k was simply heated under the above deuteration conditions. These results clearly indicated that the alkylated product was generated via a protodepalladation mechanism rather than a reductive elimination pathway (for proposed mechanism, see Scheme 4). We next explored the role of cyclopropanol in the reaction to better understand this catalytic system (Scheme 3e). Intriguingly, when cyclopropanol 4b was subjected to the standard conditions in the absence of biaryl substrate, three products, including the expected enone 4b′ and the unexpected ketones 7a and 7b, were obtained. We were aware that the cyclopropanol 4b was completely consumed around 12 h, while the ketone 7a increased throughout the reaction. The yield of 4b′ increased to ∼40% at 6 h followed by an obvious decrease, while the conversion to 7b gradually increased over time but stopped around 8 h at ∼30% yield. To confirm the active species of the transformation, we also studied the reaction of biaryl rac- 1a with 7a and 7b, but no desired product was observed (Scheme 3f). Together, these results not only suggested the involvement of the in situ formed enone 4b′ in the catalytic system but also provided important support for the formation of enone coupling partners via C–C cleavage without the involvement of biaryl aldehydes.

• Download figure
• Download PowerPoint

Based on the above observations, two distinct but closely related pathways of the C–H/C–C activation with cyclopropanes are proposed, as shown in Scheme 4. Both pathways start with imine formation by the condensation of the biary aldehydes with TDGs. Due to the steric interaction, imine B is prone to react with Pd(II) to form an enantioenriched palladacycle C via C–H activation. Next, coordination and insertion of VCP 2 into C gives intermediate D, which would undergo β-C elimination to generate E. Protonolysis of intermediate E by HOAc would release the allylation product F with the regeneration of TDG1 and Pd(II) species to initiate the next catalytic cycle. The allylation product F was subjected to Raney-Ni/H₂ to give the desired product 3. The reaction of cyclopropanol, however, most likely proceeds through the formation of enone 4′ from cyclopropanol 4 in an independent catalytic cycle involving ligand exchange/β-C elimination/β-H elimination ( 4 →  J →  K →  4′). As the experimental results in Scheme 3f show, enone 4′ was produced under the standard conditions and became the most likely reactive species for the corresponding transformation. The in situ-generated enone 4′ coordinated with palladacycle C to produce G, which underwent a subsequent migratory insertion to give intermediate H. Intermediate H could either undergo a β-H elimination to give I followed by the hydrolysis step to obtain 5′, or alternatively undergo protodepalladation to produce 5 in the presence of HOAc. Then regenerated Pd(II) species and chiral TDG would enter the next catalytic cycle. Another catalytic cycle involving the coordination of 4 with intermediate C, followed by a β-C elimination/reductive elimination pathway could be ruled out by deuterium incorporation experiments (Scheme 3d).

• Download figure
• Download PowerPoint

Conclusion

We have successfully developed the first Pd-catalyzed catalytic atroposelective C–H/C–C functionalization reactions merging C–H and C–C bond cleavage. This novel strategy streamlines the asymmetric synthesis of axially chiral biaryl compounds using readily available cyclopropane-based motifs. A broad range of axially chiral biaryls were synthesized with these strategies, affording axially chiral biaryls in synthetically useful yields with excellent enantioselectivities (up to >99% ee). These reactions can easily be scaled. Further elaborations of the resulting biaryl atropisomers enable the access of various axially chiral biaryls with high enantiopurity. Further applications of this asymmetric C–H/C–C activations strategy are ongoing.

Supporting Information

Supporting Information is available and includes X-ray crystal structure for 3o (CCDC 1964884) (CIF).

Conflict of Interest

There is no conflict of interest to report.