Enantiodivergent Synthesis of Allenes by Point‐to‐Axial Chirality Transfer

Abstract An enantiodivergent method for the synthesis of multiply substituted allenes is described. Highly enantioenriched, point‐chiral boronic esters were synthesized by homologation of α‐seleno alkenyl boronic esters with lithiated carbamates and eliminated to form axially chiral allene products. By employing either oxidative or alkylative conditions, both syn and anti elimination could be achieved with complete stereospecificity. The process enables the synthesis of either M or P allenes from a single isomer of a point‐chiral precursor and can be employed for the enantioselective assembly of di‐, tri‐, and tetrasubstituted allenes.


Allenesareversatilefunctionalgroupsthatcanbeemployed
in aw ide range of chemical transformations. [1] Theu nique pattern of reactivity displayed by allenes stems from the consecutive arrangement of two orthogonal p-bonds-a feature that also results in axial chirality.C hiral non-racemic allenes are extremely valuable intermediates in synthesis and have additional applications in medicinal chemistry,materials science,and catalysis. [2] Consequently,considerable effort has been placed on the development of asymmetric methods for allene synthesis. [3,4] Although various elegant strategies have been reported, the most general method involves nucleophilic substitution of enantioenriched propargylic electrophiles (Scheme 1A). [5] Whilst broad-ranging,t his method can result in lower enantiospecificity for tetrasubstituted allenes, for which only ah andful of enantioselective preparative methods exist. [6] Afurther limitation of such strategies is that if the opposite enantiomer of the allene is desired, then the synthesis must be repeated by first preparing the enantiomeric propargylic electrophile. [7] We envisaged ac omplementary strategy in which an alkenyl boronic ester bearing an a-leaving group (LG 1 )i s homologated with an enantioenriched lithium carbenoid (Scheme 1B). [8] As tereospecific elimination process would then convert the resulting point-chiral intermediate into an axially chiral allene. [9] In such aprocess,there are two critical points at which selectivity must be controlled. First, the boronate complex must undergo the desired 1,2-metallate rearrangement in which LG 2 is displaced, rather than ap otentially competing rearrangement in which the vinylic leaving group (LG 1 )i se xpelled. [10] Second, to obtain complete transfer of stereochemical information, the elimination process must proceed with very high stereospecificity. [11] The choice of vinylic leaving group (LG 1 )iscritical to controlling the selectivity in both of these key steps.W ewere attracted to selenium as we hoped that its relatively poor leaving group ability would enable it to act as aspectator group during the lithiation-borylation process and then, upon activation, to undergo elimination (Scheme 1C). Moreover,w eh ave recently shown that b-seleno boronic esters can undergo selective anti elimination in the presence of base or syn elimination upon oxidation to the corresponding selenoxide. [12] Using this strategy,either enantiomer of agiven chiral allene could be obtained from as ingle intermediate in ahighly divergent manner. Herein, we describe the successful Scheme 1. Previous work and our strategy for the enantiodivergent synthesis of allenes.
implementation of this strategy and its application to the enantiodivergent synthesis of di-, tri-and tetrasubstituted allenes.
We commenced our study with alkenyl boronic ester 1, which was readily prepared as the pure Z isomer in three steps from benzaldehyde. [13] Boronate complex formation was carried out with lithiated carbamate 2,followed by promotion of the 1,2-metallate rearrangement by addition of magnesium bromide in methanol and warming to 40 8 8C[ Eq. (1)]. [14] Pleasingly,n oc ompeting reactions involving displacement of the vinylic selenide were observed, and allylic boronic ester 3 was obtained in 91 %yield and 99:1 e.r.
With highly enantioenriched material in hand, we turned our attention to the development of ap rotocol for syn elimination. Upon treatment of aT HF solution of boronic ester 3 with m-CPBAa tÀ45 8 8C, we obtained the desired allene product (P)-4 in 41 %yield along with alcohol 5 in 41 % yield, which results from competing oxidation of the CÀB bond ( Table 1, entry 1). We were delighted to find that (P)-4 was formed with complete enantiospecificity,i ndicating that the point-to-axial chirality transfer had occurred with high fidelity.R educing the amount of m-CPBAf rom 2equiv to 1.2 equiv led to as mall improvement in selectivity for elimination over oxidation, and (P)-4 was obtained in 53 % yield (Table 1, entry 2). Lowering the temperature to À78 8 8C had very little impact, but carrying out the elimination at 0 8 8C led to improved selectivity in favour of (P)-4 ( Table 1, entries 3a nd 4). Remarkably,w hen the elimination was performed at room temperature,the desired allene (P)-4 was formed as the sole product in 88 %y ield with complete enantiospecificity (Table 1, entry 5).
We next focused our attention on developing aprocedure for enantiospecific anti elimination. Upon employing sodium methoxide in THF,w ewere disappointed to obtain (M)-4 in low yield and modest enantiospecificity (6 %yield, 44 %e .s.) along with as ignificant quantity of alkenyl selenide 6 ( Table 2, entry 1). This result suggests that the selenide is too poor al eaving group to compete with facile basemediated allylic protodeborylation. We rationalized that if we could convert the selenoether into abetter leaving group, the desired elimination process might become the dominant pathway.T ot est this hypothesis, 3 was transformed into the corresponding selenonium salt by alkylation with MeOTf followed by addition of sodium bicarbonate ( Table 2, entry 2). Pleasingly,t hese conditions resulted in clean elimination to form (M)-4 in 59 %y ield and significantly improved the enantiospecificity (89 %e .s.). Employing aqueous bicarbonate provided (M)-4 in an improved yield of 85 %w ith the same enantiospecificity (Table 2, entry 3). We evaluated ar ange of different aqueous bases (see the Supporting Information for full details) and found that in all cases,(M)-4 was produced with similar or reduced enantiospecificity ( Table 2, entries 4-6). When we carried out the elimination with sodium bicarbonate in methanol, we obtained the desired allene product (M)-4 with almost complete enantiospecificity (98 %e .s.) in 79 %yield (Table 2, entry 7). Finally, performing the reaction with ar educed quantity of MeOTf (2 equiv) provided (M)-4 in 83 %yield and 98 %e.s.( Table 2, entry 8).
With optimized conditions for homologation and enantiodivergent elimination in hand, we set out to investigate the generality of the process,i nitially focusing our attention on variations of the alkene partner ( Table 3). Introduction of an electron-rich methoxy substituent was well tolerated, and both enantiomers of the corresponding allene 8 were obtained  in 98:2 e.r. and in excellent yields.E lectron-deficient and sterically encumbered alkene partners also smoothly underwent the desired chemistry,providing allene products (P)-10, (M)-10,( P)-12,a nd (M)-12 all in excellent yields and with very high enantioselectivity.A na lkene partner with an aliphatic side chain (synthesized in two steps from 1-pentyne) underwent efficient lithiation-borylation (71 %y ield, 98:2 e.r.) and, after enantiodivergent elimination, provided allene products (P)-14 and (M)-14 in 77 %a nd 82 %y ield, respectively,with complete enantiospecificity.W enext investigated variations of the lithium carbenoid partner.W ef ound that boronic ester 1 could be efficiently homologated with an enantioenriched carbenoid containing asilyl ether to form 15, which underwent enantiospecific elimination affording (P)-16 (74 %y ield, 98:2 e.r.) and (M)-17 (76 %y ield, 98:2 e.r.). [15,16] Homologation of alkenyl partner 1 with al ithiated carbamate derived from (À)-citronellol provided boronic ester 18 in 70 %yield as asingle diastereoisomer.Elimination of this intermediate under either oxidative or alkylative conditions provided access to either diastereoisomer of the resulting allene (P,S)-19 or (M,S)-20.S imilarly,e mploying al ithiated carbamate containing two additional stereogenic centres enabled the highly diastereoselective synthesis of allenes (P,S,R)-22 and (M,S,R)-23.The modular nature of this synthesis is particularly noteworthy;e ach alkenyl boronic ester could be paired with as eries of different carbamates, rapidly building up alarge library of enantioenriched allenes.
We next moved on to investigate the synthesis of trisubstituted allenes (Table 4). Accordingly,w ecarried out the one-carbon homologation of 1 with an enantiopure lithiated secondary benzylic carbamate. [8c] Ther esulting tertiary allylic boronic ester 24 was obtained in 88 %y ield as as ingle enantiomer.S ubjecting this intermediate to our optimized reaction conditions for syn and anti elimination provided access to either enantiomer of the corresponding trisubstituted allene,(M)-25 (87 %yield, 99:1 e.r.) and (P)-25 (84 %yield, 99:1 e.r.). This approach was also successful with cyclic benzylic and aliphatic lithium carbenoids,p roviding allenes (M)-27,( P)-27,( M)-29,a nd (P)-29 all with very high yields and excellent enantioselectivity.W en ext explored the introduction of an additional substituent on the alkene substrate.A llylic boronic ester 30 was synthesized in very high enantioselectivity by one-carbon homologation of atetrasubstituted alkenyl boronic ester.E nantiodivergent elimination afforded allenes (P)-31 and (M)-31 in good yields with very high enantiospecificity.H omologation of at etrasubstituted styrenyl boronic ester afforded enantioenriched tertiary boronic ester 32 in 69 %yield and 97:3 e.r. Both oxidative and alkylative elimination of 32 proceeded smoothly;h owever, we were surprised to find that both reactions generated the same enantiomer of allene product 33 with high enantiospecificity.
To determine whether this elimination proceeded by a syn or an anti pathway,i tw as necessary to determine the absolute stereochemistry of allene 37.Asthis compound was an oil, we were unable to establish its absolute configuration by X-ray crystallographic analysis.W et herefore simulated the electronic circular dichroism spectrum (ECD) of (P)-37 at the CAM-B3LYP/6-311(d,p) level of theory. [17] Thes imulated ECD spectrum for (P)-37 was ag ood match for the experimental spectrum, enabling us to determine that both oxidative and alkylative elimination proceeded via a syn mechanism (Scheme 2a). This inversion of selectivity in the alkylative elimination of 36 and 32 is likely due to the fact that the conformation necessary for anti elimination results in significant A-1,3 strain between the tetravalent boron centre and the bulky phenyl substituent (Scheme 2b). [18] In this case, an alternative syn elimination pathway with fewer unfavourable steric interactions is preferred. We rationalized that the isomeric allylic boronic ester 38 ought to develop significantly less A-1,3 strain and might therefore undergo enantiodivergent elimination. Pleasingly,t his proved to be the case,a nd using this approach, both (M)-37 and (P)-37 were efficiently synthesized with high enantiospecificity.
In conclusion, we have developed an ew method for the enantiodivergent synthesis of allenes by point-to-axial chirality transfer. Homologation of alkenyl boronic esters with enantioenriched lithium carbenoids followed by syn or anti Table 4: Enantiodivergent synthesis of tri-and tetrasubstituted allenes.
[c] Reaction conditions for the anti elimination:1,2-Selenoboronic ester (1 equiv), MeOTf (2-5 equiv), CH 2 Cl 2 ,RT, 16 h; then NaHCO 3 (20 equiv), CH 2 Cl 2 /MeOH, RT,3h. Scheme 2. Determination of the absolute configuration for allene (P)-37 and rationalization of the observed selectivity.a )Experimental ECD spectrum (solid line): 0.23 mm in MeOH, RT.S imulated ECD spectrum (dashed line) calculated at the CAM-B3LYP/6-311(d,p) level of theory.b)Rationalization of the syn elimination, which is driven by A-1,3 strain. elimination provided efficient access to either enantiomer of the resulting allene products.T he method is extremely general, enabling the highly convergent synthesis of di-, tri-, and even tetrasubstituted allenes bearing arange of different aromatic and aliphatic groups.This method serves as auseful alternative to nucleophilic addition to propargylic electrophiles and will find widespread use for the synthesis of chiral, non-racemic allenes.