Highly Selective Difluoromethylations of β-Keto Amides with TMSCF2Br under Mild Conditions

Abstract Without employing any transition metal and other additives, efficient methods for selective difluoromethylations of β-keto amides with TMSCF2Br reagent have been developed under mild conditions. This protocol allows a convenient access to various α-difluoromethyl β-keto amides with excellent yields (up to 93%) and high carbon/oxygen (C/O) regioselectivities (up to 99:1). The C/O selectivity of β-keto amides could be easily reversed and controlled by simply changing the base. This protocol can be easily scaled-up and the C-difluoromethylation product could be reduced into CF2H-containing amino alcohol derivatives. Moreover, the first enantioselective electrophilic difluoromethylation of β-keto amides has been achieved by phase-transfer catalysis.

Fluorinated molecules play an increasingly important role within pharmaceuticals and materials science. 1 Among them, the difluoromethyl (CF 2 H) group, an analogue of well-recognized trifluoromethyl (CF 3 ) group, has attracted considerable attention because it can act as a surrogate of a hydroxyl or a thiol group and a lipophilic hydrogen-bond donor. 2 Furthermore, this fascinating group (CF 2 H) could modulate the lipophilicity of a molecule that may increase the drug's metabolic stability. 3 Because of this, the introduction of the difluoromethyl group into leading drug candidates has become a powerful strategy for new drug discovery. On this account, many marketed drugs containing the CF 2 H skeletons, such as soft PDE4 inhibitor roflumilast (medicine for pneumonia) and the proton-pump-inhibiting drug pantoprazole. 4,5 Therefore, it is highly desirable to develop efficient methods to introduce the CF 2 H motif into or-ganic molecules. Until now, numerous methods have been developed to access CF 2 H-containing molecules. 6 Although metal photoredox catalysis and transition-metal catalysis are effective means, 7 the difluorocarbene-mediated electrophilic difluoromethylations are still attractive and straightforward strategies for the preparation of difluoromethylated compounds. 8 However, most of these reactions are focused on the difluoromethylation of heteroatoms, such as O, S, N, P, and Se nucleophiles. 9 However, the difluoromethylation of C-H nucleophiles is sparse, remains an area with limited success. Furthermore, -keto esters are widely used nucleophiles and the development of controlling carbon/oxygen (C/O) regioselectivities were problematic. Recently, efficient difluoromethylations with high C/O selectivities were achieved by the groups led by Hu,10 Shibata,11 Shen, 12 and Liu 13 using different kinds of difluoromethylating reagents. Compared with the -keto esters, -keto amides are still challenging substrates, possibly due to the lower acidity of the -hydrogen, and the reactivity towards difluorocarbene is largely underexplored or unknown. In Hu's work, only one example of -keto amide was showed to react with difluorocarbene, and the yield of the difluoromethylated product was low. 10 Moreover, there are no examples of direct difluoromethylation of the N-H-containing -keto amides as far as we know. Therefore, the development of a general method for efficient difluoromethylation of various -keto amides with readily available reagent is highly desirable. Herein, we reported an efficient method for the highly regioselective difluoromethylations of -keto amides, and the C/O selectivities of -keto amides could be easily reversed and controlled. Moreover, the enantioselective difluoromethylation of -keto amides could be achieved by phase-transfer catalysis (Scheme 1).

Letter Synlett
Scheme 1 Direct difluoromethylation of -keto amides The discovery of simple and efficient systems for efficient difluoromethylation of various C-H nucleophiles with readily available reagent is a highly desirable task in chemistry. 14 To develop a new and convenient method for the Cdifluoromethylation of -keto amides, we investigated the -difluoromethylation of 1-indanone-derived -keto amide 1a using the readily available reagent TMSCF 2 Br, a difluorocarbene reagent originally introduced by Hu 8c,10 as the difluorocarbene precursor. First, we screened several bases with CH 2 Cl 2 in the presence of TMSCF 2 Br. Organic bases such as DBU and Et 3 N were ineffective in this transformation (Table 1, entries 1, 2). We found that inorganic bases such as KOH and NaOH could promote this reaction, but with moderate yields and C/O regioselectivities (  16). To our delight, we found toluene could improve the yield and C/O selectivity obviously (Table 1, entry 17). Other reaction parameters such as temperature, concentration, and other additives were also screened (for more details, see Table S1 in the Supporting Information). Subsequently, further improvement was observed in toluene by using 1.5 equiv of TMSCF 2 Br in the presence of LiOH (3.0 equiv) in 2 mL toluene at 15 °C for 24 h. Under the optimized reaction conditions, the C-difluoromethylation product 2a was obtained with excellent C/O selectivity (C/O = 98:2) in 91% yield (Table 1, entry 18).
With the optimized reaction conditions in hand, the scopes of the difluoromethylation of -keto amides were explored to evaluate the generality of the process (Scheme 2). First, we investigated the substituents at the amide sides. Aniline-derived substrates containing 4′-Me, 4′-MeO, 4′-tBu, and aryl derivatives containing naphthyl and benzyl group could be converted into the corresponding products 2a-f in good yields (85-93%) and up to 99:1 C/O selectivities. It is worth mentioning that we scaled up the -difluoromethylation of -keto amide 1a to the gram scale. Under the standard conditions, 10 mmol of -keto amides (1a) reacted smoothly with 15 mol of TMSCF 2 Br and gave 2.70 g (90% yield) of 2a with 98:2 C/O selectivity after 16 h. Compounds 1g-j, with aliphatic amido groups, affording the corresponding products 2g-j in good yields (92-94%) and excellent C/O selectivities (99:1). Next, we extended the procedure to substrates which have halogen atoms and electron-donating groups on the aromatic ring of the indanone scaffold. Under mild conditions, 1k-m Scheme 2 Substrate scope of the -difluoromethylation of -keto amides. Reagents and conditions (unless otherwise specified): the reactions were performed with -keto amide (0.1 mmol), TMSCF 2 Br (0.13 mmol), LiOH (0.3 mmol), and 2.5 mL toluene at 15 °C for 16 h. C/O regioselectivities were determined by 19 F NMR spectroscopy with trifluorotoluene as the internal standard.

Letter Synlett
which have halogen substituents were nicely converted into the corresponding products in 85-88% yields and excellent C/O selectivities (98:2). Furthermore, when methoxy groups were introduced (1n and 1o), the reactions proceeded smoothly, and the desired products were obtained with 89-93% yields. Interestingly, the five-membered cyclopentanone derived substrate 1p could be converted into the -difluoromethylation products with 79% yield and 90:10 C/O selectivity. Compound 1q, which has methyl and phenyl in the N-position, afforded 2q in 75% yield and 98:2 C/O selectivity under the optimized reaction conditions. Then the scope of 1-tetralone-derived keto amides 1r-u were then examined, and the desired products were obtained with 79-89% yields and up to 96:4 C/O selectivities. To our delight, the six-membered cyclohexanone-derived substrate 1v was nicely converted into the corresponding product 2v with 84% yield and 92:8 C/O selectivity. Ultimately, such a simple and efficient method for -difluoromethylation showed good substrate tolerance for indanone and 1-tetralone-derived -keto amides.
Inspired by these exciting results, we turned our attention to the unique reactivity of TMSCF 2 Br and carried out some comparison experiments using other difluoromethylation reagents B1-4. -Keto amide 1a was selected as a model substrate. As described in Scheme 3, TMSCF 2 Br can difluoromethylate 1a to give product 2a in 90% yield and 98:2 C/O selectivity. However, under the optimized and modified reaction conditions, the S-difluoromethyl sulfonium salt B1, 15 the difluoromethyl ethers B2, 16 the phos-phine ylide type of difluoromethylation reagent B3, 17 and the diethyl (bromodifluoromethyl)-phosphonate B4 18 showed low reactivities towards 1a, and the desired product 2a was formed in low yields (0-31%) and C/O selectivities. These results highlight the unique feature and advantage of TMSCF 2 Br as a privileged difluorocarbene precursor for this C-difluoromethylation of -keto amides.
The controlling of C/O regioselectivities was also significant, and we envisioned that the regioselectivity of TM-SCF 2 Br towards O site could be enhanced under different reaction conditions. For the -keto amide 1a, under standard conditions, only C-difluoromethylated product 2a was formed in 90% yield and 98:2 C/O selectivity. Then we noticed that alcohols can react with :CF 2 directly without predeprotonation. 19 So we used the weakly acidic KHF 2 as the reaction promotor, CH 2 Cl 2 and H 2 O as the mixed solvents, and we were pleased to see that the O-difluoromethylated product 3a was obtained in 83% yield and 8:92 C/O selectivity at 40 ℃ for 12 h. These findings indicated that the C/O selectivity of -keto amides could be easily controlled by simply changing the reaction conditions, it also highlighted TMSCF 2 Br as a unique and versatile difluorocarbene reagent (Scheme 4).

Letter Synlett
The synthetic application of the C-difluoromethylation product 2a was further demonstrated (Scheme 5). To our delight, the CF 2 H-containing amino alcohol derivative 4a could be easily obtained by reduction with LiAlH 4 in THF in 87% yield with 8:1 dr, which highlighted the practicability of this transformation.
To gain insight into the reaction mechanism, some control experiments were performed (Scheme 6). We found that the reaction performed well in an argon atmosphere and in the dark, which suggested the aerobic oxidation 20 and photooxidation 21 were not involved in this reaction. The yield of 2a was not significantly decreased in the presence of radical inhibitor TEMPO and SET inhibitor 1,4-dinitrobenzene, which indicated that a radical process did not occur in this reaction system. Furthermore, no deuterated product was detected when deuterated toluene was used as solvent (see Scheme 6d). These results clearly suggested that the difluorocarbene pathway of the capturing proton from substrate 2 to produce nondeuterated C-difluoromethylation product is predominant.
On the basis of these experiments and the related reported work, we proposed a possible mechanism for the difluoromethylation of -keto amide. For the C-selective difluoromethylation, the CF 2 Br anion was generated via initial desilylation by base (such as HO -), Then -elimination of CF 2 Br anion occurred and it is fragmented to generate difluorocarbene. The protonation of CF 2 Br anion is much slower than its fragmentation to difluorocarbene. 22 Then difluorocarbene was prone to be captured by the carbon site of enolate intermediate 1-1. Based on the Pearson acid-base concept, 23 a strong affinity between hard Li + and the hard oxygen site in enolate 1-1 helped to block the approaching CF 2 carbene from the O-site. Then the final C-difluoromethylation product 2 could be generated by capturing a proton from 1. 11 A plausible mechanism for the O-difluoromethyl-

Letter Synlett
ation with TMSCF 2 Br was also proposed based on our results and the previous studies on reaction of carbenes with hydroxyl groups. 19,24 TMSCF 2 Br could be activated by KHF 2 to form a pentacoordinate silicate intermediate at the oilwater interface. In the organic phase, BrCF 2 K is released from the pentacoordinate silicate intermediate and then splits into KBr and the singlet difluorocarbene. The latter species interacts with two alcohol molecules to form a fivemembered complex with oxonium character 1-3, 25 which eventually undergoes double proton transfer to deliver the difluoromethyl ether and regenerate one O-selective difluoromethylation product 3 (Scheme 7).
Asymmetric phase-transfer catalysis is recognized as an effective and sustainable method, and cinchona alkaloid derived phase-transfer catalysts have been applied to many practical asymmetric synthesis. 26 Considering our interest in the development of efficient and practical enantioselective -functionalizations of -dicarbonyl compounds by phase-transfer catalysis. 27 we sought to develop a convenient method for asymmetric -difluoromethylation of keto amides. First, we investigated the asymmetric -difluoromethylation of 1-indanone-derived -keto amide 1a using cinchona alkaloid derived phase-transfer catalysts (Figure 1 and Table 2). The simple PTC-1 and PTC-2 provided 2a with 6-8% ee in toluene by using 30% K 2 CO 3 aqueous solution as the base (Table 2, entries 2 and 3). PTC-3, which has two hydroxy groups at both C-9-and C-6′-positions, showed poor results (Table 2, entry 4). Then we screened the doubly quaternized catalyst PTC-4, but the enantioselectivity was poor (Table 2, entry 5). The N-oxide PTC-5 and the C-9-hydroxy-protected PTC-6 and PTC-7 afforded 2a with 6-11% ee (Table 2, entries 6-8). Then we turned our attention to the C-2′-arylated PTCs. PTC-8 provided 2a with 24% ee (Table 2, entry 9). PTC-10 improved the enantioselectivities to 33% ee (Table 2, entry 11). Furthermore, PTC-11, containing 3,5-bromo groups in the benzyl position, worked well and gave enantioselectivity of 37% ee, and the enantioselectivity was improved to 45% when the reaction temperature was decreased to -10 °C, but the yield of 2a was significantly decreased to 61% (Table 2, entries 12 and 13). Although the result of the asymmetric -difluoromethylation of -keto amides was not very satisfying, we still developed an enantioselective electrophilic difluoromethylation of -keto amides by phase-transfer catalysis, and no broadly effective strategy for the construction of enantioenriched tertiary and quaternary centers bearing CF 2 H groups has emerged as far as we know. 28 In conclusion, we have developed selective difluoromethylations of -keto amides with TMSCF 2 Br. 29 A variety of -keto amides could be efficiently and selectively transformed into the corresponding C-difluoromethylated products under mild conditions. Moreover, the C/O selectivity of

Letter Synlett
-keto amides could be easily controlled by simply changing the reaction conditions. This protocol can be easily scaled-up, and the C-difluoromethylation product could be reduced into amino alcohol derivatives. Plausible mechanisms and transition state showed that TMSCF 2 Br could be activated by LiOH or KHF 2 to form difluorocarbene, and the addition of LiOH (which not only served as a difluorocarbene activator, but also served as the O-steric blocking site) was crucial in the control of high C/O regioselectivity. Moreover, the first enantioselective electrophilic difluoromethylation of -keto amides under phase-transfer catalysis was investigated. Further work on the application of this useful difluoromethylation for the asymmetric synthesis of natural products and pharmaceuticals are currently underway in our laboratory.