Catalytic enantioselective halocyclization of olefinic substrates attracted much attention over the past few years. The resulting enantioenriched heterocycles form the core of many natural products and pharmaceutically important intermediates. In addition, the halogen atoms can be manipulated readily to give various functional building blocks. Most of the reported catalytic enantioselective halocyclization reactions involve the use of a nitrogenor an oxygen-containing nucleophilic partner that has an N H or a O H moiety (e.g., amides or alcohols/carboxylic acids, respectively). The catalysts can then interact with the R H functionalities (R=N, O) through hydrogen bonds and/or salt bridges. Such interactions allow for an effective enantioinduction, resulting in highly enantioselective transformations. 1,3-Dicarbonyl compounds 1, which contain an acidic proton in a-position to the carbonyl funtionalities, are also potential nucleophiles for halocyclizations (when R is an olefinic side chain), furnishing highly functionalized cyclic ethers 3 (Scheme 1). However, this type of nucleophile is also highly susceptible to halogenation in a-position to the carbonyl functionalities, which is a well-known process. The resulting a-halogenated dicarbonyl compound 2 is not a good halogenating reagent and not usually used in the halocyclization of olefinic substrate 1. Nonetheless, herein we report an efficient, catalytic, and highly enantioselective bromocyclization of olefinic 1,3-dicarbonyl compounds, giving rise to highly functionalized furans (see below). A protocol that employs an amine–thiocarbamate as catalysts was developed for the chemoselective bromination at the olefinic moiety over the carbon atom in a-position to the carbonyl functionalities. At the outset of this study, we screened several catalysts that are commonly used in halocyclizations. Olefinic 1,3dicarbonyl compound 1a and N-bromosuccinimide (NBS) were used as substrate and stoichiometric source of the halogen atom, respectively. Reactions with simple catalysts, including DABCO (4), DBU, and DMAP, gave significant amounts of 2-bromo 1,3-dicarbonyl compound 2a, while the reaction with NEt3 gave a mixture of unidentified products (Table 1, entries 1–4). In particular, when DBU was used as the catalyst, cyclized product 3a was not obtained and 2a was isolated in 98% yield. During further investigations, we found that catalysts with Lewis basic sulfur atoms preferably mediate the bromination at the olefinic moiety instead of at the carbon atom in aposition to the carbonyl functionalities, and thus facilitate the cyclization process. The reaction with triphenylphosphine sulfide (5) gave the desired product 3a with only trace amounts of 2a (Table 1, entry 5). When the reaction was catalyzed by thiocarbamate 6, 3a was obtained as the only product in 78% yield (Table 1, entry 6). To our delight, 3a was obtained in an excellent yield (92%) and with only a negligible amount of 2a when amine–thiocarbamate catalyst 7a was used (Table 1, entry 7). Encouraged by these preliminary results, we proceeded to optimize the reaction conditions and to develop the asymmetric version of the reaction. A systematic screening of various halogen sources and solvents showed that toluene and NBS gave the best results (see the Supporting Information for details). The influence of the reaction temperature was also examined. At 60 8C, the reaction of 1a with quinidine-derived amine–thiocarbamate 7a gave product 3a in 34% yield and 16% ee (Table 2, entry 1, footnote c). The ee value and yield improved when the reaction was conducted at 40 8C (Table 2, entry 1). Unexpectedly, reactions with thiocarbamate catalysts that contain other cinchona alkaloid frameworks, including quinine, cinchonine, and cinchonidine, were sluggish and no Scheme 1. Comparison of previous and present studies.
Read full abstract