Article pubs.acs.org/accounts
Cite This: Acc. Chem. Res. 2018, 51, 2036−2046
Copper-Catalyzed Radical Relay for Asymmetric Radical Transformations Published as part of the Accounts of Chemical Research special issue “Hydrogen Atom Transfer”. Fei Wang, Pinhong Chen, and Guosheng Liu*
Acc. Chem. Res. 2018.51:2036-2046. Downloaded from pubs.acs.org by UNIV OF SUNDERLAND on 09/24/18. For personal use only.
State Key Laboratory of Organometallic Chemistry, Center for Excellence in Molecular Synthesis, Shanghai Institute of Organic Chemistry, University of Chinese Academy of Sciences, Chinese Academy of Sciences, 345 Lingling Road, Shanghai 200032, P. R. China CONSPECTUS: The direct transformation of C−H bonds into diverse functional groups represents one of the most atom- and step-economical strategies for organic synthesis and has received substantial attention over the last few decades. Despite recent advances, asymmetric C−H bond functionalizations are less developed, especially asymmetric oxidations of sp3 C−H bonds. Inspired by enzyme (e.g., P450) catalysis, chemists have made great efforts to develop non-enzymatic systems for enantioselective oxidations of sp3 C−H bonds. However, the involvement of highly reactive radical intermediates makes enantioselective transformations extremely challenging. In this Account, we present our recent studies on the enantioselective induction of prochiral benzylic radicals using a chiral bisoxazoline (Box)/Cu catalytic system. This reaction system was developed on the basis of our extensive studies of copper-catalyzed intermolecular alkene difunctionalizations, such as azidotrifluoromethylations, trifluoromethylcyanations, and trifluoromethylarylations. In these reactions, the proposed catalytic cycle starts from the oxidation of the Cu(I) species by the activated Togni-I reagent (via a Lewis acid/base interaction with a silyl reagent or arylboronic acid) through a single electron transfer process. The generated CF3 radical can efficiently add to the alkene, and the resultant carbon-centered radical is subsequently trapped by an active Cu(II) species bearing a nucleophile (e.g., an N3, CN, or Ar moiety) to form a new C−heteroatom or C−C bond and regenerate the Cu(I) catalyst. Kinetic studies of the trifluoromethylarylation of alkenes support a Cu(I/II/III) catalytic cycle in which the carbon radical reacts with the Cu(II) species to form a highly reactive Cu(III) intermediate and its reductive elimination contributes to the final bond formation. This assumption inspired us to explore asymmetric radical transformations by introducing chiral ligands. Enantioselective cyanations and arylations of benzylic radicals have been demonstrated in the presence of chiral Box/Cu(I) catalysts, and a series of asymmetric difunctionalizations of styrenes have been successfully achieved. In addition, by means of the same benzylic radical trapping process, enantioselective decarboxylative cyanations have been demonstrated using a cooperative photocatalysis and copper catalysis system. Compared with radical addition and decarboxylative processes, hydrogen atom abstraction (HAA) provides direct and facile access to benzylic radicals. By using bisbenzenesulfonimidyl radical for HAA, our group has developed an enantioselective cyanation of benzylic C−H bonds via a copper-catalyzed radical relay, and excellent reactivity and enantioselectivity were achieved in the presence of chiral Box/Cu(I) catalysts. In addition, a regioselective benzylic C−H bond arylation proceeding through a similar process was also developed, providing simple access to 1,1-diarylalkanes. Notably, alkyl arenes were used as the limiting reagent in these reactions, which allowed the late-stage functionalization of sp3 C−H bonds in complex molecules, including natural products, pharmaceuticals, and agrochemicals. numerous documented reviews.2 Nevertheless, analogous asymmetric C−H functionalizations are much less developed, especially for sp3 C−H bonds.3,4 Enzymatic oxygenases and halogenases, such as cytochrome P450 and non-heme iron enzymes, usually exhibit excellent reactivity in regio- and enantioselective functionalizations of sp3
1. INTRODUCTION The direct oxidative functionalization of C−H bonds is among the most efficient methods for incorporating functional groups into organic molecules, and such reactions represent a formidable challenge in organic synthesis.1 In recent years, some strategies for the direct functionalization of C−H bonds have been developed. For instance, C−H bond activation by transition metal catalysis has been proven to be a powerful method for achieving this goal and has since become the focus of © 2018 American Chemical Society
Received: June 5, 2018 Published: September 5, 2018 2036
DOI: 10.1021/acs.accounts.8b00265 Acc. Chem. Res. 2018, 51, 2036−2046
Article
Accounts of Chemical Research C−H bonds.5 In this case, mechanistically, a metal−oxo intermediate at the core of these enzymes can abstract a hydrogen atom in an sp3 C−H bond to form a metal hydroxide and an alkyl radical intermediate. Subsequently, C−X bonds (X = OH, halide) are formed through a radical rebound process (Scheme 1 shows hydroxylation as an example). Inspired by
Scheme 2. Copper-Catalyzed ARTs through a Radical Relay Process
Scheme 1. Oxidation of sp3 C−H Bonds by Cytochrome P450 or Non-heme Iron Enzymes
species, and subsequent reductive elimination from the CuIII center would lead to the enantioselective formation of the sp3 C−FG bond.16 More importantly, this strategy was successfully applied to the asymmetric functionalization of sp3 C−H bonds, in which the benzylic radical was formed via HAA of a benzylic C−H bond. In these reactions, the alkyl arene was used as the limiting reagent, which allows the late-stage C−H functionalization of bioactive molecules.
enzymatic catalysis, chemists have designed a series of catalysts that mimic the core motif of P450, such as metalloporphyrin-6 and tetradentate aminopyridine-coordinated metal (e.g., Mn, Fe, etc.) complexes.7 However, these metal catalysts are often less enantioselective than the enzyme catalysis.8 For example, Groves and co-workers developed a novel manganese-catalyzed method for the azidation of aliphatic C−H bonds, but they reported only one asymmetric example, which occurred with moderate enantioselectivity.9 As a key step, enantioselective control of the highly reactive carbon radical is crucial to the design of ideal asymmetric radical C−H functionalization reactions, but this remains a challenging objective. Chemists have addressed this problem by employing chiral organocatalysts or Lewis acids with chiral ligands.10 Notably, transitionmetal-catalyzed redox systems also play an eminent role in this field. For example, Fu and co-workers developed a Ni-catalyzed strategy for enantioselective cross-couplings11 in which the enantioselective induction of the radical species was controlled by the formation of an R−Ni(III)−R′ species from an alkyl radical (R·) and R′Ni(II); subsequently, as the enantiodetermining step, reductive elimination from the Ni(III) center led to enantiomerically enriched products.12 Following a similar pathway, enantioselective cross-coupling reactions of α-halo esters/amides were recently realized by the use of Fe,13 Co,14 and Cu15 catalysts. However, this appealing strategy has not yet been successfully applied to asymmetric functionalizations of sp3 C−H bonds. In this Account, we summarize our recent studies on coppercatalyzed asymmetric radical transformations (ARTs). Notably, these reactions proceed through a radical relay rather than a radical rebound pathway, i.e., the initially generated radical undergoes hydrogen atom abstraction (HAA) or addition to the alkene to generate a new carbon-centered radical (Scheme 2). For example, in the asymmetric difunctionalization of alkenes, an electrophilic reagent X−Y oxidizes a chiral L*Cu(I) species to produce a heteroatom-centered radical (Y·) and a L*CuIIX species. The radical Y· then adds to a styrene to generate a benzylic radical; meanwhile, transmetalation (or ligand exchange) between the L*CuIIX species and a nucleophilic reagent [FG] leads to the L*CuIIFG species. The benzylic radical can be trapped by L*CuIIFG to form an alkyl−CuIIIFG
2. COPPER-CATALYZED INTERMOLECULAR DIFUNCTIONALIZATION OF ALKENES As part of our ongoing research on transition-metal-catalyzed difunctionalizations of alkenes, 17 we reported the first palladium-catalyzed intramolecular trifluoromethylarylation of activated alkenes.18 Simultaneously, allylic C−H trifluoromethylation reactions using copper catalysts were independently developed by the groups of Buchwald, Wang, and Liu,19 and an alkyl radical species was proposed as a key intermediate in these reactions. Inspired by these pioneering works, we found that the radical species can be selectively captured by nucleophiles in the presence of a copper catalyst (Scheme 3A). For instance, the copper-catalyzed intermolecular azidotrifluoromethylation of alkenes using the less reactive Togni-I reagent as the CF3 source and TMSN3 as the nucleophile has been developed (Scheme 3B).20 The reaction exhibited a broad substrate scope. For example, styrenes, mono- and disubstituted aliphatic alkenes, electron-deficient alkenes, and cyclic alkenes were suitable substrates for the reaction (e.g., 2a−h). Regiospecific addition of the CF3 radical to the terminal carbon of the alkene was observed in all cases, as this addition gave the most stable alkyl radical intermediate. In addition, trifluoromethylated alkyl nitriles 3 and isocyanides 4 could also be obtained with TMSCN and TMSNCS as the nucleophilic reagents under modified conditions (Scheme 3C).21,22 Notably, the silyl reagents are crucial to these reactions, and other inorganic nucleophiles, such as NaN3, NaCN, and KSCN, are ineffective. Furthermore, our catalytic systems could also be used for the trifluoromethylation of alkynes, allenes, and enynes.23 The first intermolecular trifluoromethylarylation of alkenes was achieved with arylboronic acids as the aryl source (Scheme 4A).24 The reaction featured a broad substrate scope of aryl- and alkyl-substituted alkenes as well as arylboronic acids. Moreover, reactions of cyclic alkenes gave the desired products with excellent diastereoselectivities (>20:1). Notably, only the Togni-I reagent was suitable for the reaction, and other CF3 reagents such as the Togni-II reagent and Umemoto’s reagent were ineffective. Other arylboron reagents (e.g., ArBpin and 2037
DOI: 10.1021/acs.accounts.8b00265 Acc. Chem. Res. 2018, 51, 2036−2046
Article
Accounts of Chemical Research
The 19F NMR studies combined with the above control experiments revealed that mutual activation between the arylboronic acid and the Togni-I reagent was crucial to the success of the trifluoromethylarylation reaction (Scheme 4B). This unique interaction enhanced the oxidative potential of the Togni-I reagent, facilitating the oxidation of the Cu(I) catalyst to give the CF3 free radical and the Cu(II) species through a single electron transfer (SET) process. As shown in Scheme 5, the CF3
Scheme 3. Intermolecular Trifluoromethylation of Alkenes
Scheme 5. Possible Mechanism for the Trifluoromethylarylation of Alkenes
radical quickly added to the alkene, giving an alkyl radical intermediate (Int. III); meanwhile, the alkoxy anion generated from the reduced Togni-I reagent acted as a strong base to activate ArB(OH)2, promoting the transmetalation to deliver the ArCu(II) species.25 These two species were then combined to provide Cu(III) species Int. IV, and reductive elimination from this species led to products 5.16,26 This catalytic cycle was supported by the kinetics studies, Hammett plot, and competitive reactions. Importantly, this discovery of Cu-assisted C−Ar bond formation reflects a promising approach to achieving asymmetric radical transformations by introducing a chiral ligand, which inspired us to explore this possibility.27
Scheme 4. Intermolecular Trifluoromethylarylation of Alkenes
3. ENANTIOSELECTIVE COPPER-CATALYZED DIFUNCTIONALIZATION OF STYRENES Initially, we examined the feasibility of asymmetric versions of the above copper-catalyzed trifluoromethylation reactions (Scheme 6A). In the presence of bisoxazoline (Box) ligand L1, reactions of 1i with different nucleophiles smoothly afforded the corresponding products 2i−4i in good yields. Unlike the low enantioselectivities observed for the azidation (17% ee) and thiocyanation (
