A Lewis Base Catalysis Approach for the Photoredox Activation of Boronic Acids and Esters

Abstract We report herein the use of a dual catalytic system comprising a Lewis base catalyst such as quinuclidin‐3‐ol or 4‐dimethylaminopyridine and a photoredox catalyst to generate carbon radicals from either boronic acids or esters. This system enabled a wide range of alkyl boronic esters and aryl or alkyl boronic acids to react with electron‐deficient olefins via radical addition to efficiently form C−C coupled products in a redox‐neutral fashion. The Lewis base catalyst was shown to form a redox‐active complex with either the boronic esters or the trimeric form of the boronic acids (boroxines) in solution.

Abstract: We report herein the use of ad ual catalytic system comprising aL ewis base catalyst such as quinuclidin-3-ol or 4-dimethylaminopyridine and ap hotoredox catalyst to generate carbon radicals from either boronic acids or esters.T his system enabled awide range of alkylboronic esters and aryl or alkylb oronic acids to react with electron-deficient olefins via radical addition to efficiently form C À Cc oupled products in aredox-neutral fashion. The Lewis base catalyst was shown to form ar edox-active complex with either the boronic esters or the trimeric form of the boronic acids (boroxines) in solution.
Carbon-centered radicals are as ynthetically powerful class of reactive intermediates. [1] They are particularly attractive in the context of C À Cb ond-forming reactions, [2] overcoming problems often associated with two-electron processes. [3] By enabling visible-light-promoted single electron transfer, photoredox catalysis has become am ethod of choice for the single-electron reduction or oxidation of organic substrates and allows to generate open-shell intermediates in amild and selective fashion. [4] Ar ange of reductive or oxidative carbon radical precursors are now available to generate carbon radicals in the context of ap hotocatalytic cycle. [5] Oxidative carbon radical precursors are often anionic species suffering from poor solubility in common organic solvents.F or example,e xtensively studied organoborates [6] possess an electron-rich B(sp 3 )m oiety that can be subjected to single-electron oxidation, leading to an eutral carbon radical after CÀBbond cleavage (Scheme 1A).
Despite their ubiquity as reagents in organic synthesis [7] and in biologically active molecules, [8] theuse of boronic acid derivatives to generate carbon-centered radicals remains underexplored. [9] Owing to their high oxidation potentials, they have received much less attention in this regard, with few reports making use of strong stoichiometric oxidants or anodic oxidation. [10] We recently demonstrated that benzyl boronic esters can undergo single-electron oxidation under photoredox conditions when their vacant porbital is engaged in ad ative bond with the norbital of as toichiometric Lewis base (LB) additive (Scheme 1B). [11] Lewis base catalysis was introduced as aconcept by Denmark to enhance the reactivity of electrophilic n*, p*, and s*orbitals. [12] Based on this knowledge,w eh ypothesized that the use of ac atalytic amount of an organic Lewis base would be av iable option for the photoredox activation of boronic acids and esters. [13] Herein, we describe adual catalytic method to effectively form alkyl and aryl radicals from aw ide array of boronic esters and acids by direct photoredox single-electron oxidation under mild and safe conditions,without the requirement for stoichiometric activators or oxidants.T hese reactive species were further engaged in intermolecular C À Cb ondforming processes to deliver desirable C(sp 3 )ÀC(sp 3 )a nd C(sp 2 )ÀC(sp 3 )bonds in ar edox-neutral fashion.
Thea ddition of electron-rich carbon-centered radicals onto electron-deficient olefins,a lso known as Giese-type addition, [14] is an interesting method to form C À Cb onds in ar edox-neutral fashion and can also be used to assess the presence of the postulated radical intermediates. [15] We initially subjected model boronic ester 1a to an excess of methyl acrylate (2a)i nt he presence of 1.5 equiv of 4-dimethylaminopyridine (DMAP) as an additive and the photoredox catalyst PC(1), which we have already shown to be quenched by DMAP-activated 1a. [11] Irradiation of this mixture with blue LEDs for 24 hr eadily led to the coupling product 3aa in 86 %y ield. Reducing the DMAP catalyst loading to 20 mol %s till provided 3aa in 75 %y ield (Scheme 2), with the remaining mass balance resulting from oligomerization due to multiple acrylate additions.Pleased by this level of catalytic activity,wedecided to investigate other Lewis bases.
According to Denmarkstheory,n-n* interactions are the most productive type of activation for aLewis base catalyst to be active, [12] so ar ange of commercial neutral Lewis bases with an available non-bonding norbital were screened at 20 mol %l oading. Strongly nucleophilic [16] quinuclidinederived bases such as quinuclidin-3-ol and quinuclidine were identified as productive catalysts,l eading to the formation of 3aa in 80 %( 75 %u pon isolation) and 77 % yield, respectively.P hosphine-derived Lewis bases were also investigated, with triphenylphosphine (PPh 3 )s howing good activity.C ontrol experiments revealed the necessity of blue light irradiation, photocatalyst, Lewis base,and methanol for the successful conversion of boronic esters (see the Supporting Information for the full optimization and control experiments).
With optimized reaction conditions in hand, we assessed the scope with electron-deficient alkenes 2a-2r (Scheme 3). Aside from methyl acrylate, tert-butyl and benzyl acrylate are also suitable coupling partners (3aa-3ac). Methyl vinyl ketone was identified as the best coupling partner,w ith the conjugate addition product being isolated in 82 %yield (3ad). Pleasingly,a crolein and acrylonitrile coupling products (3ae and 3af)were also obtained in high yields,thereby expanding the range of functional groups tolerated with this method. gem-Disubstituted olefins also reacted in ar adical conjugate addition, and methyl methacrylate (3ah), aconjugated lactam (3ai), and two cyclic enones (3aj and 3ak)w ere selectively coupled in 58-68 %y ield. Interestingly,2 -a nd 4-vinylpyridines were successfully alkylated at the b-carbon atom (3am and 3an), providing examples of reactions with challenging N-heteroaromatic compounds. [17] These results could be extended to a2 -pyridyl-containing 1,1-disubstituted olefin (3ao), showcasing the possibility to generate pheniramine analogues and the potential application of the method for antihistaminic drug discovery. [18] Finally,f lavone natural products can also be alkylated, albeit in lower yield (3aq and 3ar).
We next turned our attention to establishing the scope with respect to boronic ester coupling partners with methyl vinyl ketone 2d (Scheme 4). Primary benzylic pinacol esters were selectively coupled (3ad-3ed)i nt he presence of quinuclidin-3-ol as the Lewis base catalyst. Interestingly, a-heteroatom-substituted primary alkyl boronic esters were also coupled in high yields (3fd-3hd,8 6-91 %), with triphenylphosphine proving to be the most efficient catalyst for the a-amino products 3gd and 3hd.M ore sterically demanding secondary benzylic esters required the use of DMAP as the Lewis base catalyst, highlighting the effect of the steric hindrance on the required initial complexation between boronic ester and Lewis base.W hereas methyl (3id and 3jd)a nd benzyl (3kd)s ubstituents were well tolerated, the presence of larger isopropyl (3ld)orphenyl (3md)groups led to less efficient coupling.L astly,t ertiary boronic esters were explored (3nd-3pd). Despite their well-known difficulty to be efficiently engaged in metal-catalyzed cross-couplings, [14a, 15, 19] DMAP allowed for clean activation to form quaternary carbon centers in respectable yields even from commercial and less activated tBuBpin (3pd).
Aryl boronic esters,o nt he other hand, were found to be substantially less reactive than their activated alkyl counterparts.W ei nitially observed only low reactivity after 24 ho f irradiation, and therefore surveyed different aryl-substituted B(sp 2 )s pecies to find that aryl boronic acids were more reactive than the corresponding pinacol, glycol, neopentyl, and catechol esters (see the Supporting Information). Our experience with the Lewis acidity of aryl boronic acids led us  to propose that the reactive species in solution was more likely to be the trimeric boroxine than the monomeric species. [20] This was confirmed by NMR experiments showing the complexation of quinuclidin-3-ol with boroxine instead of the corresponding free boronic acid (see the Supporting Information). This finding led us to screen as eries of commercially available boronic acids in this reaction (Scheme 5).
Despite the usually harsh reaction conditions employed to oxidize aryl boronic acids, [10b] we found that alarge number of electron-rich aryl boronic acids could be successfully coupled to 2d under extremely mild and redox-efficient conditions. Thecouplings of aryl boronic acids with nitrogen (5cd-5ed), oxygen (5ad and 5fd), and sulfur (5bd and 5gd)substituents on the ring all proceeded in good to excellent yields.Oxygencontaining heterocycles derived from catechol could be incorporated into the substrates (5hd and 5id), and unprotected 5-and 6-indoyl boronic acids (5jd and 5kd)were also successfully functionalized in the presence of nucleophilic NH and C3 centers.T he enhanced reactivity observed with boroxines relative to boronic esters encouraged us to attempt using unactivated alkyl boronic acids as starting materials. Primary alkyl boronic acids were successfully coupled (5md-5od)a long with secondary alkyl derivatives (5pd and 5qd), showcasing the usefulness of this method to generate functional unstabilized alkyl radicals.
[5a] Secondary a-amino boronic acids derived from amino acids [7d] were also well tolerated, with proline-derived 5rdas well as the peptide drug ixazomib transformed in high yield (5sd), illustrating the potential application to late-stage functionalization.
According to NMR studies,afast, dynamic equilibrium is established between the boroxine 6a' derived from boronic acid 6a or boronic ester 6b and the Lewis base catalyst (LB) in the reaction solvent mixture (see the Supporting Information). Cyclic voltammetry measurements informed us that complex 7 can be single-electron-oxidized( E 1/2 (1a-DMAP) =+0.81 Vv s. SCE) within the reductive quenching cycle of PC(1) (E 1/2 (Ir III* ) =+1.2 Vv s. SCE). [21] Thec arbon radical thus generated (8)u ndergoes ar adical addition with 10 to form the intermediate radical 11,w hich can then be reduced and quenched by aproton from methanol to provide coupling product 13 (Scheme 6). [6c] Theresulting methanolate can then be used to regenerate the LB from 9.
In conclusion, we have developed anew set of photoredox reaction conditions taking advantage of the Lewis acidity of boronic esters and boroxines (from boronic acids) to generate primary,s econdary,a nd tertiary alkyl or aryl radicals.T hese intermediates were engaged in redox-neutral C À Cc ouplings with electron-deficient olefins,f orming ar ange of new C(sp 3 )ÀC(sp 3 )a nd C(sp 2 )ÀC(sp 3 )c ross-coupled products. Over 50 structurally and functionally diverse products were successfully synthesized. This new activation method should enable the use of boronic acids and esters in awide range of other radical-based reactions.