2,2-difluorovinyl benzoates for diverse synthesis of gem-difluoroenol ethers by Ni-catalyzed cross-coupling reactions

gem-Difluoroalkene is a bioisostere of carbonyl group for improving bioavailability of drug candidates. Herein we develop structurally diverse 2,2-difluorovinyl benzoates (BzO-DFs) as versatile building blocks for modular synthesis of gem-difluoroenol ethers (44 examples) and gem-difluoroalkenes (2 examples) by Ni-catalyzed cross coupling reactions. Diverse BzO-DFs derivatives bearing sensitive functional groups (e.g., C = C, TMS, strained carbocycles) are readily prepared from their bromodifluoroacetates and bromodifluoroketones precursors using metallic zinc as reductant. With Ni(COD)2 and dppf [1,1’-bis(diphenylphosphino)ferrocene] as catalyst, reactions of BzO-DFs with arylboronic acids and arylmagnesium/alkylzinc reagents afforded the desired gem-difluoroenol ethers and gem-difluoroalkenes in good yields. The Ni-catalyzed coupling reactions features highly regioselective C(vinyl)–O(benzoate) bond activation of the BzO-DFs. Results from control experiments and DFT calculations are consistent with a mechanism involving initial oxidative addition of the BzO-DFs by the Ni(0) complex. By virtue of diversity of the BzO-DFs and excellent functional group tolerance, this method is amenable to late-stage functionalization of multifunctionalized bioactive molecules.

While the current cross-coupling protocols are effective for the synthesis of gem-difluoroalkenes, other moieties such as gemdifluoroenol ethers which are also found in some pharmaceutical compounds remains less accessible [36][37][38][39][40] . In this regard, Katz and co-workers reported the preparation of stable potassium trifluoroborate gem-difluoroenol ethers, which would deliver the gem-difluoroenol ethers by cross-coupling with aryl halides under palladium catalysis 41 . To our knowledge, methods producing gem-difluorovinyl building blocks with immense structural diversity for the gem-difluoroenol ethers and gem-difluoroalkenes synthesis are rare. Herein we report a convenient synthesis of structurally diverse 2,2-difluorovinyl benzoates (BzO-DFs) by zinc treatment of the abundant bromodifluoroacetates and bromodifluoroketones as starting materials. The BzO-DFs serve as versatile building blocks for synthesis of diverse gem-difluoroenol ethers and gem-difluoroalkenes by Ni-catalyzed arylation and alkylation using nucleophiles such as arylboronic acids, arylmagnesium bromides, alkylzinc bromides, potassium alkyltrifluoroborates. Encouragingly, cross electrophile coupling reactions involving alkyl chlorides were also feasible under Ni catalysis (Fig. 1c).

Results
Synthesis and scope of BzO-DFs. To begin, 2-bromo-2,2difluoroacetate (1.0 mmol) was treated with benzoyl chloride (0.5 mmol) and zinc powder (2.0 equiv) in a dioxane/MeCN (4:1 v/v) mixture under a N 2 atmosphere, and 3a was obtained in 77% yield (see Supplementary Information for optimization study). For structure characterization, the 19 F NMR spectrum of 3a featured two characteristic doublet signals at −116 and −117 ppm, which is consistent with the gem-difluoroenol ether structure. The molecular structures of the BzO-DF derivatives 4s and 5c have been unambiguously established by X-ray crystallography.
Mechanistic studies. Underlying the favorable reactivity of the BzO-DFs for gem-difluoroenol ethers synthesis is the regioselective C(vinyl)-O(benzoate) bond activation. Compared to the C (acyl)/C(alkyl)-O bonds activation, the observed selectivity may be attributed to the better leaving group reactivity of the benzoates 47 . In this regard, we performed a correlation study using a series of BzO-DFs (3a/3b/3c/3d/3f/3h/3i), which bear a series of substituted benzoic acids as leaving groups.  19 F NMR with PhCF 3 as an internal standard. Plotting log(k FG /k H ) with the Hammett σ(para) constants resulted a linear plot with a slope of +0.82 (R 2 = 0.82) of log(k FG / k H ) against the σ(para) constants. This finding indicated a faster reaction rate is promoted by electron-withdrawing substituents on the benzoates. This implies that partial negative charge buildup at the transition state of the oxidative addition.
In terms of product formation, it was found that the Nicatalyzed arylation of 3a-3f exhibited rather optimal yields (>80%) when the pKa values are within the range of 4.14-4.57. Lower yields were observed for 3d (LG = 4-chlorobenzoic acid) and 3e (LG = 4-bromobenzoic acid) despite of the lower pKa values of 3.97. (Fig. 5b). As expected, for the BzO-DF with adamantanecarboxylic acid (pKa = 4.86) as the leaving group, no coupling product was obtained under the standard conditions (see Supplementary Information for details). The benzoates were indeed the leaving group for the C(vinyl)-O activation; this is verified by the isolation of the benzoic acids in 76% yield after treating the reaction crude with HCl (aq) (Fig. 5c, eq. 1).
Regarding to the mechanism, Ni(0) complexes are known to effect C(sp 2 )-halide bond activation initiated by single-electron transfer mechanism. It appeared plausible that the C(vinyl)-O (benzoate) activation may proceed via initial single-electron reduction of the BzO-DFs by the Ni(0) complex, leading to carboradicals formation via the C-O bond fragmentation. In this work, the radical mechanism was scrutinized by using TEMPO as radical trap. When the "3a+6c" reaction was performed in the presence of TEMPO (2.0 equiv), the 7c formation was completed suppressed with 61% recovery of the starting 3a (Fig. 5c, eq. 2). Apparently, no TEMPO-radical adducts were detected by GC-MS and LC-HRMS analysis of the reaction mixture. The lack of product formation could be attributed to the oxidation of the active Ni(0) complexes by the TEMPO, resulting in the loss of the catalytic activity. Consistent with this notion, when BHT (2.0 equiv) was used in place of TEMPO as radical trap, effective formation of 7c (85%) was detected by NMR analysis (Fig. 5b, eq. 3). Furthermore, when 3a and 6c were subjected to the Nicatalyzed conditions in the presence of α-cyclopropylstyrene (2.0 equiv) as radical probe, 7c was formed in 75% yield and the radical probe was completely recovered unchanged (Fig. 5c, eq.  4). On the basis of these findings, pathways involving radical formation is untenable.
We performed density-functional theory (DFT) calculation to delineate the mechanistic details of the reaction. We chose the "3a +6c" reaction as a model for our DFT study. Based on our ligand screening study, we believed that the dppf-ligated Ni(0) species A is the active species to initiate the coupling reaction. Oxidative addition of 3a should pass through a barrier of~27.2 kcal mol −1 . This step is found to be the rate-determining step 51,52 . As shown in Fig. 6, oxidative addition of 3a to form complex B is nearly a thermal neutral process. Yet, the major driving force of the oxidative addition appears to be the ligand exchange reaction of the benzoate with K 2 CO 3 to form complex C, which is highly exergonic with a drop of the relative Gibbs free energy from 2.0 kcal mol −1 to −12.4 kcal mol −1 . Subsequent transmetalation with arylboronic acid 6c takes place via a six-membered ring transition state (TS C-D ) to generate intermediate D 53,54 . The transmetalation is followed by spontaneous reductive elimination to give Ni  Finally, E reacts with 3a through ligand exchange to give product 7c with the regeneration of the active species A, which is very facile with low energy barriers. In summary, we developed a modular synthesis of gemdifluoroenol ethers and gem-difluoroalkenes by the Ni-catalyzed coupling with structurally diverse BzO-DFs. The BzO-DFs can be easily prepared from abundant bromodifluoroketones and bromodifluoroacetates. Exhibiting remarkable functional groups tolerance, this Ni-catalyzed protocol permits easy installation of a gem-difuorovinyl group to bioactive molecules. This reaction should supplement the conventional strategies in exploiting difluorovinyl moieties in drug design and development. Control experiments and DFT calculations revealed that the reaction is initiated by regioselective oxidative addition of the BzO-DFs by the Ni(0) complex. The intermediate vinylnickel(II) benzoate complexes readily exchanged with organometallic nucleophiles enabling a versatile C(sp 2 )-C(sp 2 ) and C(sp 2 )-C(sp 3 ) coupling manifolds.

Methods
General procedure for preparation of BzO-DFs (3)(4)(5). To a 25 mL roundbottom flask with a magnetic stir bar was added acyl chloride (1.0 equiv, 0.5 mmol), Zn powder (2.0 equiv, 1.0 mmol), MeCN (0.5 mL) and dioxane (1.5 mL) in a nitrogen-filled glovebox. The reaction mixture was then sealed with a rubber septum and removed from the glovebox. To the mixture, 1 (2.0 equiv, 1.0 mmol) was diluted with dioxane (0.5 mL) and added with a syringe pump over 20 min. After stirring the reaction mixture for further 12 h at 30°C, the reaction mixture was filtered, and the filtrate was concentrated in vacuo. The residue was purified by flash column chromatography (n-hexane:ethyl acetate = 200:1-50:1) to give the desired 2,2-difluorovinyl benzoates.

Data availability
All data are available from the corresponding authors upon reasonable request. The Xray crystallographic coordinates for structures reported in this study have been deposited at the Cambridge Crystallographic Data Centre (CCDC), under deposition numbers 2008864 (4s), 1921828 (5c), and 1921829 (9a). These data can be obtained free of charge from The Cambridge Crystallographic Data Cenrte via www.ccdc.cam.ac.uk/ data_request/cif".   Fig. 6 Gibbs free energy profile calculated for the Ni-catalyzed reaction of 3a and 6c. The relative Gibbs energies and Electronic energies (in parentheses) are given in kcal/mol. ARTICLE NATURE COMMUNICATIONS | https://doi.org/10.1038/s41467-020-20725-9