Reductive Coupling of Acrylates with Ketones and Ketimines by a Nickel‐Catalyzed Transfer‐Hydrogenative Strategy

Abstract Nickel‐catalyzed coupling of benzyl acrylates with activated ketones and imines provides γ‐butyrolactones and lactams, respectively. The benzyl alcohol byproduct released during the lactonization/lactamization event is relayed to the next cycle where it serves as the reductant for C−C bond formation. This strategy represents a conceptually unique approach to transfer‐hydrogenative C−C bond formation, thus providing examples of reductive heterocyclizations where hydrogen embedded within an alcohol leaving group facilitates turnover.

The identification of catalytic paradigms for the direct and atom-economical assembly of C À Cb onds is ak ey goal of organic chemistry.Within this context, transfer-hydrogenative C À Cb ond formation has emerged as ap owerful platform for reaction design. Forexample,hydrogen borrowing allows the direct a-alkylation of carbonyl compounds with alcohols by ac atalytic dehydrogenation/condensation/reduction sequence (Scheme 1a). [1] Therelated Guerbet reaction effects the dehydrative union of two alcohols,t hus providing an efficient method to upgrade bioethanol to butanol (Scheme 1b). [2] Krische and co-workers have pioneered transferhydrogenative alcohol CÀHfunctionalizations as exemplified by processes where alcohol dehydrogenation drives the reductive generation of nucleophilic metal allyls in advance of carbonyl addition (Scheme 1c). [3] Each of these reaction classes merges redox events with C À Cb ond formation, thus avoiding stepwise generation of reactive functionality and enhancing substantially atom economy.Assuch, new transferhydrogenative CÀCbond-forming strategies are likely to find wide utility in reaction design.
Our studies in this area were initiated by considering synthetic entries to g-butyrolactones and lactams, [4][5][6][7] which are versatile intermediates as well as core motifs in an array of natural products.A na ppealing,y et unrealized approach to these compounds resides in metal-catalyzed reductive coupling of either ac arbonyl or imine with an acrylate to afford a g-amino or g-hydroxy ester,w hich upon cyclization would provide the target (Scheme 1d). This disconnection requires the identification of as trategy which enables reductive CÀC bond formation, but avoids nonproductive reduction of the starting materials.W er easoned that these criteria might be fulfilled by coupling the release of the reductant to the formation of either the lactone or lactam, thereby minimizing nonproductive background reduction events.Such aproposition appears practically challenging,h owever,asimple solution is availed by harnessing the native reducing power of the alcohol released upon cyclization to drive turnover. In this way,t he alcohol byproduct from one cycle is relayed to the next, where it then serves as the reductant for C À Cb ond formation. Herein, as proof-of-concept, we show that lactones and lactams can be generated by nickel-catalyzed union of activated ketones and ketimines,r espectively,w ith O-benzyl acrylates.T his approach provides unique examples of reductive heterocyclizations where hydrogen embedded within an alcohol leaving group facilitates catalytic turnover, [8] thus adding anew vista to the wider area of transfer-hydrogenative CÀCbond formation. [1][2][3] In early studies,w ea ssayed aw ide range of latetransition-metal systems for the reductive coupling of isatin 1a and ethyl acrylate (2a, R = Et;T able 1). At 150 8 8Ci n PhMe,and with 10 mol %benzyl alcohol as the initiator (see Scheme 1d), the combination of 7.5 mol %N i(cod) 2 and 15 mol %P ( o-OMeC 6 H 4 ) 3 provided the target lactone 3a in 19 %y ield, with unreacted starting material accounting for the mass balance ( Table 1, entry 1). Here,a ccording to our reaction design, ethanol released during the first turnover must then function as the reductant for subsequent cycles. Based on this we considered whether more easily oxidized alcohol-based leaving groups might provide increased efficiencies. [9] Ultimately,t his led to the reaction conditions outlined in entry 3, which use 300 mol %benzyl acrylate (2b, R = Bn) as the reaction partner,a nd generate 3a in 84 % yield. Some turnover was observed in the absence of the initiating alcohol (entry 4), likely facilitated by hydrolytic release of BnOH from benzyl acrylate under the reaction conditions.This generates acrylic acid as abyproduct, acomponent which control experiments found to be inhibitory to the reductive lactonization process. [10] Lower loadings of either the benzyl alcohol initiator or the nickel pre-catalyst resulted in diminished efficiencies (entry 5), and use of stoichiometric BnOH also resulted in al ower yield (entry 6). This latter result highlights the benefits of coupling reductant release to turnover. 3a was generated in 58 %yield when the reaction was run with only 100 mol % 2b (entry 7). An ickel(0) pre-catalyst is essential for efficient reactivity; nickel(II) systems (e.g.e ntry 8) or commonly employed transfer-hydrogenation catalysts,s uch as [IrCp*Cl 2 ] 2 (entry 9), were completely ineffective. [11] Thes cope of the process with respect to the isatin component is outlined in Table 2. Avariety of electronically distinct systems (1a-j)p articipated to provide the target spirocyclic systems 3a-j in moderate to excellent yield. The protocol shows useful functional-group tolerance,w ith both esters (3h)a nd methoxy (3d)s ubstituents surviving,d espite the established lability of these functionalities under nickel-(0)-catalyzed conditions. [12] Processes involving disubstituted acrylates required the addition of Mg(OTf) 2 as aLewis-acidic co-catalyst. [13] By using this modification, reductive coupling of 1a with a-methyl (2c)a nd a-phenyl (2d)b enzyl acrylate provided the targets 3k and 3l,respectively,inhigh yield and as single diastereomers (> 20:1 d.r.). Ther elative stereochemistries of 3kand 3lwere assigned by X-ray diffraction. [14] Interestingly,these products possess opposite relative configurations. b-Substituted acrylates also participate,s uch that targets 3m and 3n were formed in 77 and 73 %y ield, respectively.Inthe latter case,the Lewis acid co-catalyst was not required, likely due to the high electrophilicity of the acrylate partner,dibenzylfumarate 2f.
Our observations are that isatins are privileged substrates for this reductant relay process.N evertheless,w eh ave [a] Yield determined by 1 HNMR analysis using 1,4-dinitrobenzene as an internal standard. [b] Using 100 mol %benzyl acrylate (2b). cod = 1,5cyclooctadiene, Cp* = C 5 Me 5 . established that, in certain cases,o ther classes of 1,2dicarbonyl also participate,t hus suggesting potentially wider applications of the strategy.For example,benzil systems 4a-c generated the corresponding monocyclic lactones 5a-c in modest to very good yield (Scheme 2). Cyclic system 6 was also ac ompetent reaction partner, generating lactone 7 in 52 %yield when Mg(OTf) 2 was used as co-catalyst. As far as we are aware,t he examples in Table 2a nd Scheme 2a re the first catalytic reductive lactonizations which harness carbonyls and unfunctionalized acrylate esters.Existing noncatalytic protocols use exogenous stoichiometric reductants, [5] whereas catalytic approaches require alcohols as the starting material, and, in turn, mandate prior reduction of the carbonyl partner. [4] To probe the mechanism of the process as eries of experiments was undertaken. When deuterio-2b,w hich incorporates deuterium at the benzylic positions,was exposed to the optimized reaction conditions,40% deuterium transfer to C2 of deuterio-3a was observed (Scheme 3a). Significant deuterium incorporation was also found at C2',t hus indicating that the nickel(0) system can also activate the N-benzylic position. [15] For 1a to 3a (84 %Y ield), GCMS analysis of the crude reaction mixture revealed the concomitant formation of benzaldehyde in 78 %yield. These observations show that the benzyloxy unit of the acrylate partner (2b)a cts as the reductant for C À Cbond formation. Under optimized reaction conditions we have confirmed that benzyl acrylates are most effective (Scheme 3b). Other systems with either primary or secondary alcohol based leaving groups,such as methyl, ethyl, and cyclohexyl acrylate,a lso enabled turnover, but provided 3a in significantly diminished yields.C onversely,p henyl and tert-butyl acrylate,w hich release "non-oxidizable" phenol or tBuOH, did not allow turnover, with the yield of 3a limited to the loading of the benzyl alcohol initiator (10 mol %). Overall, these observations are consistent with the reductive formation of g-hydroxy ester 9,inadvance of lactonization to give 3a (Scheme 3c). Intermediate 9 might arise by either ac arbonyl reduction/conjugate addition pathway (Path a) [16] or an oxidative coupling/reduction sequence (Path b). [17,18] Tw ok ey observations provide circumstantial support for Path a: 1) an adjacent acidifying group is required on the carbonyl partner [19] and 2) products of oxidative coupling with the benzaldehyde byproduct are not formed. [20] Thebeneficial effects of Mg(OTf) 2 in certain cases would be consistent with Lewis acid activation of the acrylate for conjugate addition. Exposure of 8 (the reduced form of 1a)t ot he optimized reaction conditions,with either 2b or 2c,generated 3a in high yield (Scheme 3d). Lactone formation from 8 in the absence of the nickel catalyst was feasible,b ut resulted in low conversion to 3a (15 %y ield). Thus,i fP ath ai so perative, the nickel catalyst must play an intimate role in enhancing the CÀCb ond-forming event. One possibility is that oxidative addition of nickel(0) into the C3ÀHb ond of 8 generates anickel enolate,aprocess which has been suggested in other contexts. [21] Exposure of 8 to benzaldehyde (100 mol %) under standard catalytic conditions (in the absence of acrylate) resulted in a3 5% yield of 1a,t hus showing that reduction of 1a is reversible.Because of this,initial oxidation of 8 to 1a in advance of spirolactonization by Path bc annot be ruled out. As already discussed, either nickel(II) systems Scheme 2. Reductivecoupling of benzyl acrylate with activated ketones.

Angewandte Chemie
Communications or commonly employed ruthenium-and iridium-based transfer-hydrogenation catalysts do not promote the reaction, thus supporting ar ole for the nickel(0) system beyond simply effecting transfer hydrogenation of 1a.
According to the mechanistic blueprint outlined in Scheme 1d,o ther classes of process might be achievable using areductant relay approach. Although further expansion of the strategy will require the identification of new catalysts and/or fragment coupling steps,w ew ere keen to uncover additional processes which might be achieved using the nickel(0) system presented here.S pecifically,w ee nvisaged that a-oxo imines might couple with acrylates to provide lactams.T his proposition was appealing because only sparse reports document the use of stoichiometric metallic reductants to achieve this seemingly simple process,a nd no catalytic approaches are available. [7] Pleasingly,w hen the Np-methoxyphenyl imine 10 a was exposed to the reaction conditions optimized for lactonization, spirocycliclactam 11 a was generated in 68 %y ield (Table 3). Further evaluation revealed that this lactamization process has similar scope to the lactonization methodology.Indeed, electronically diverse isatin-based imines (10 b-e)a ll engaged in smooth reductive coupling to provide lactam targets 11 b-e in good to excellent yield. Extension of the protocol to the imine derived from benzil 4b provided monocyclic system 12 in 68 %y ield;t he alternate lactone product was not observed. We also investigated ao ne-pot imine formation/lactamization sequence (Scheme 4). Exposure of 1a to p-methoxyaniline under acidic conditions generated imine 10 a.R emoval of the volatile components was followed by direct addition of the reagents required for reductive lactamization, allowing at elescoped synthesis of 11 a in 50 %y ield over the one-pot, threecomponent process.
In summary,w ed emonstrate au nique approach to transfer-hydrogenative CÀCb ond formation, wherein the native reducing power of an alcohol released upon either lactonization or lactamization is used to drive catalytic turnover. This approach provides an interesting example of an atom-economical methodology,h ighlighting how an otherwise wasted byproduct can be used productively.T he studies described herein encompass the first catalytic methods for accessing lactones and lactams by the direct reductive coupling of carbonyls and imines,r espectively,w ith unfunctionalized acrylates.Future studies will seek to identify other catalyst systems which can promote the stereocontrolled coupling of awider range of reaction partners.