Pd(II)-Catalyzed Enantioselective C–H Activation of Cyclopropanes

Abstract

Systematic ligand development has led to the identification of novel mono-N-protected amino acid ligands for Pd(II)-catalyzed enantioselective C–H activation of cyclopropanes. A diverse range of organoboron reagents could be used as coupling partners, and the reaction was found to proceed under mild conditions. These results provide a new retrosynthetic disconnection for the construction of enantioenriched cis-substituted cyclopropanecarboxylic acids.

Recently, Pd-catalyzed asymmetric C–H activation reactions have been demonstrated through the use of a chiral auxiliary¹ or chiral ligand.^2–6 Spectroscopic and crystallographic investigations have provided valuable insights into the process by which [Pd(II)– mono-N-protected amino acid] catalysts asymmetrically cleave prochiral C–H bonds.² Nevertheless, achieving high levels of enantioselectivity in these reactions remains a significant challenge, largely due to the paucity of suitable ligand scaffolds capable of effecting stereoinduction during C–H cleavage. In our previous work, high ee was obtained in the desymmetrization of prochiral aryl C–H bonds (up to 95%, Scheme 1), and promising initial results were also found in asymmetric alkyl C–H activation (up to 37% ee) by using [Pd(II)-mono-N-protected amino acid] catalysts.^2a

Scheme 1.

Desymmetrization of Prochiral C–H Bonds

Encouraged by these precedents, we sought to develop enantioselective C–H activation reactions of cyclopropanes.⁷ Owing to the prominence of enantiopure cyclopropanes in natural products and pharmaceuticals, a diverse collection of transition metal–mediated transformation have been developed for their synthesis.⁸ Herein, we report a complementary method, which constitutes the first example of the enantioselective cyclopropyl C–H activation/organoboron cross-coupling (Scheme 2).^9,10 A diverse collection of aryl-, alkyl-, and vinylboron coupling partners were compatible with these reaction conditions. Systematic ligand tuning has led to the development of a protocol that gives high levels of stereoinduction under mild conditions. This reaction provides a versatile route for the synthesis of cis-substituted chiral cyclopropane carboxylates.

Scheme 2.

Asymmetric Cyclopropane C–H Activation

Based on our recent success in utilizing acidic N-arylamides as weakly coordinating directing groups for a diverse range of alkyl and aryl C–H functionalization reactions,^11,12 we first sought to establish a robust reaction to cross-couple the amide derivative of 1-methylcyclopropanecarboxylic acid (1) with phenylboronic acid pinacol ester (Ph–BPin) in the absence of a chiral ligand. Extensive screening revealed that a 2:1 mixture of mono- and di-arylated products (1a) could be obtained in 91% yield at 100 °C. Gratifyingly, aryl-, alkyl- and vinylboron reagents were all suitable coupling partners (Table 1). Importantly, this is the first example of Pd(II)-catalyzed cross-coupling of alkyl C–H bonds with vinyl boron reagents (1c and 1d). The use of boronic acid pinacol esters (BPin) and NaHCO₃ were crucial for arylation and vinylation, while potassium trifluoroborate salts (BF₃K) and Li₂CO₃ were optimal for alkylation. The presence of 40 mol% of DMSO was found to promote arylation and vinylation (1a–1d),¹³ while the addition of DMF as a cosolvent was beneficial for alkylation (1e–h). Importantly, even when the temperature was lowered to 40 °C, substrate 1 could still be arylated without a major decline in yield (78%, mono:di = 2.7:1).

Table 1.

Racemic Cross-Coupling of Cyclopropyl C–H Bonds with Organoboron Reagents^a

^aThe mono:di ratio was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as an internal standard.

^bConditions: 0.1 mmol of substrate, 10 mol% Pd(OAc)₂, 2.0 equiv of Ar–BPin, 1.5 equiv of Ag₂CO₃, 3.0 equiv of NaHCO₃, 0.5 equiv of BQ, 5 equiv of H₂O, 40 mol% DMSO, 0.5 mL of t-AmylOH, 100 °C, N₂, 12 h.

^c0.1 mmol of substrate, 10 mol% Pd(OAc)₂, 1.2 equiv of vinyl–BPin, 1.5 equiv of Ag₂CO₃, 3.0 equiv of NaHCO₃, 0.5 equiv of BQ, 40 mol% DMSO, 0.5 mL of THF, 100 °C, N₂, 6 h.

^d0.1 mmol of substrate, 10 mol% Pd(OAc)₂, 2.0 equiv of alkyl–BF₃K, 1.5 equiv of Ag₂CO₃, 3.0 equiv of Li₂CO₃, 0.5 equiv of BQ, 0.1 mL of DMF, 0.5 mL of THF, 100 °C, N₂, 12 h.

With the mild cross-coupling protocol at 40 °C in hand, we proceeded to examine systematically mono-N-protected amino acid ligands in an effort to develop an enantioselective protocol (Table 2). We initially focused on screening mono-N-protected L-leucine and found that carbamate groups gave superior ee and monoselectivity, compared with amide groups. The monoselectivity was also improved proportionally with the ee (for complete ligand screening data, see SI). Based on this observation, we further optimized conditions using Fmoc-Leu-OH as the ligand and discovered that 5 mol% catalyst and 10 mol% ligand loadings at 40 °C gave the highest ee (50%). The ee dropped to 12% when temperature was raised to 70 °C. Increasing the catalyst loading to 10 mol% gave improved yield (60%) but decreased the ee (40%); importantly, adding the catalyst and ligand in two batches gave high yield (74%) while maintaining the ee (48%). The addition of DMSO improved the yield but led to erosion of the ee, presumably because DMSO is capable of competing with the ligand for coordination to Pd. The presence of H₂O enhanced the yield (likely by promoting transmetallation)7 without reducing the ee. Of the various carbamate protecting groups that were tested, 2,2,2-trichloro-tert-butyloxycarbonyl (TcBoc) afforded the best ee (78%) and yield (47%).

Table 2.

Screening of Ligand Protecting Groups^a,^b

^aConditions (unless otherwise specified): 0.1 mmol of substrate, 5 mol% Pd(OAc)₂, 10 mol% ligand, 1.0 equiv of Ph–BPin, 1.0 equiv of Ag₂CO₃, 3.0 equiv of NaHCO₃, 0.5 equiv of BQ, 5 equiv of H₂O, 0.5 mL of t-AmylOH, 40 °C, N₂, 12 h.

^bThe yield was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as an internal standard. Stereochemical assignment is tentative.

We subsequently investigated the effect of the amino acid backbone. Although TcBoc-Leu-OH gave the highest ee, we instead focused on Fmoc-protected amino acids due to their commercial availability (Table 3). As expected, achiral Fmoc-Gly-OH (L1) gave a racemic mixture of products. The carboxylic acid moiety was found to be essential for stereoinduction, as Fmoc-alanine methyl ester (L3) gave no ee. Fmoc-protected amino acids containing hydrophobic alkyl chains (L2, L4–L6) gave ee values between 43 and 50%. Intriguingly, coordinating functional groups on the side chain such as an ester (L7), thioether (L8), and ether (L9) gave improved ee, between 65 and 73%; however, the conversion dropped to below 40% in each case. We then screened amino acids with aryl side chains (L10, L11). To our delight, Fmoc-Phe-OH and its derivatives gave improved ee values of 68% and above, with Fmoc-Tyr(t-Bu)-OH (L11) giving 80% ee and 49% product yield. Fmoc-Trp(Boc)-OH (L12) also gave 73% ee. These combined findings signaled to us that an aryl group on the amino acid side chain was crucial for obtaining high ee.

Table 3.

Screening of the Amino Acid Side Chains^a,^b

^aThe conditions are identical to Table 2.

^bThe yield was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as an internal standard. Stereochemical assignment is tentative.

Having established that both the TcBoc protecting group and phenylalanine backbone were beneficial for enantioselectivity, we synthesized a series of TcBoc-protected amino acids (L13–L18). We confirmed that TcBoc-Phe-OH (L17) gave better ee (85%) than those with alkyl side chains (L13–L15) (Table 4). TcBoc-PhG-OH (L16) and TcBoc-MePhe-OH (L18), both of which possess an aryl group, however, gave significantly lower ee. Further optimization of the protecting group on phenylalanine was carried out. While retaining the CCl₃ moiety present in TcBoc, we varied the two alkyl groups and found that L23 and L24 improved the ee to 90 and 91%, respectively. Subsequently, the newly designed protecting group (PG7) in L23 was installed on commercially available phenylalanine derivatives (L25–L27). Substitution on the phenyl ring was found to have a modest effect on the enantioselectivity, with L27 improving the ee to 93%. To investigate in more detail whether the CCl₃ moiety of the TcBoc group has a dominant effect on the enantioselectivity, we extensively screened a variety of sterically hindered protecting groups with phenylalanine (see SI), however, only 48–62% ee was obtained. The CCl₃ moiety presumably serves not only as a sterically bulky group, but also tunes the electronic properties of the nitrogen atom through its electron-withdrawing character.

Table 4.

Systematic Tuning of the Amino Acid Ligands^a,^b

^aThe conditions are identical to Table 2.

^bThe yield was determined by ¹H NMR analysis of the crude product using CH₂Br₂ as an internal standard. Stereochemical assignment is tentative.

With the optimized reaction conditions in hand, we performed enantioselective C–H/organoboron cross-coupling of cyclopropane 1 with Ph–BPin, 1-cyclohexenyl–BPin and n-butyl–BF₃K (Table 5). The reactants (excluding the substrate) were added in two batches, using 5 mol% catalyst and 10 mol% ligand (L27) in each batch to give the optimal yield and ee. The addition of the reactants in a single batch resulted in inferior and inconsistent results. The apparent dependence of the ees on the concentration of catalysts remains to be investigated. Phenylated product 1a was obtained in 81% yield and 91% ee. The cross-coupling of 1-cyclohexenyl- and n-butyl-boron reagents required elevated temperatures of 50 and 70 °C to obtain appreciable product formation, which decreased the ee values to 82% and 62%, respectively. Primary alkyl, iso-propyl, and cyclopentyl groups at the α-position of the cyclopropane were tolerated, giving good ee values (2a–7a). β-Benzyl ethers (8a) and γ-phthalimide-protected amines (9a) were compatible, as was α-substitution with an aryl group (10a, 11a). Substitution of the aryl ring with electron-withdrawing halide groups suppressed competitive ortho-C(aryl)–H functionalization. The chiral cycloproprane products could also undergo further C–H coupling reactions to give cis-1,2,3-substituted cyclopropanes under the same conditions in the absence of ligands, albeit in low yields (20–38%). Unfortunately, substrates containing an α-hydrogen atom or α-heteroatoms gave poor yields and ee at 40 °C. Detailed mechanistic studies through spectroscopic and crystallographic analyses, as well as further optimization of the ligand and the reaction conditions are underway to solve these problems.¹⁴

Table 5.

Asymmetric Cyclopropane C–H Functionalization^a,^b

^aConditions: (First batch) 0.1 mmol of substrate, 5 mol% Pd(OAc)₂, 10 mol% ligand, 1.0 equiv of Ph–BPin, 0.75 equiv of Ag₂CO₃, 2.0 equiv of NaHCO₃, 0.25 equiv of BQ, 3 equiv of H₂O, 0.5 mL of t-AmylOH, 40 °C, N₂, 6 h. (Second batch) 5 mol% Pd(OAc)₂, 10 mol% ligand, 0.5 equiv of Ph–BPin, 0.75 equiv of Ag₂CO₃, 1.0 equiv of NaHCO₃, 0.25 equiv of BQ, 1 equiv of H₂O, 0.2 mL of t-AmylOH, 40 °C, N₂, 6 h.

^bIsolated yield. Stereochemical assignment is tentative.

^cConditions: (First batch) 0.1 mmol of substrate, 5 mol% Pd(OAc)₂, 10 mol% ligand, 0.5 equiv of vinyl–BPin, 0.75 equiv of Ag₂CO₃, 2.0 equiv of NaHCO₃, 0.25 equiv of BQ, 0.5 mL of THF, 50 °C, N₂, 6 h. (Second batch) 5mol% Pd(OAc)₂, 10 mol% ligand, 0.5 equiv of vinyl–BPin, 0.75 equiv of Ag₂CO₃, 1.0 equiv of NaHCO₃, 0.25 equiv of BQ, 0.2 mL of THF, 50 °C, N₂, 6 h.

^dConditions: (First batch) 0.1 mmol of substrate, 5 mol% Pd(OAc)₂, 10 mol% ligand, 1.0 equiv of n-Bu–BF₃K, 0.75 equiv of Ag₂CO₃, 1.5 equiv of Li₂CO₃, 0.25 equiv of BQ, 3 equiv of H₂O, 0.5 mL of THF, 70 °C, N₂, 6 h. (Second batch) 5 mol% Pd(OAc)₂, 10 mol% ligand, 0.5 equiv of n-Bu–BF₃K, 0.75 equiv of Ag₂CO₃, 0.75 equiv of Li₂CO₃, 0.25 equiv of BQ, 0.2 mL of THF, 70 °C, N₂, 6 h.

In summary, the first example of enantioselective C–H activation of cyclopropanes was achieved through systematic tuning of the mono-N-protected amino acid ligand and reaction conditions. Enantioselective C–H/R–BX_n cross-coupling with aryl-, vinyl- and alkylboron reagents provides a new disconnection for the synthesis of cis-substituted chiral cyclopropanecarboxylic acids. Studies to expand the substrate scope and to extend this methodology to other prochiral methyl and methylene C–H bonds are ongoing in our laboratory.