Merging Asymmetric [1,2]-Additions of Lithium Acetylides to Carbonyls with Type II Anion Relay Chemistry

Abstract

An enantioselective three-component coupling reaction has been developed, enabling the union of a variety of lithium acetylides and electrophiles exploiting an achiral linchpin via an anionic reaction cascade. This Type II Anion Relay Chemistry tactic is initiated via an enantioselective [1,2]-carbonyl addition exploiting BINOL catalysis to access an enantioenriched alkoxide intermediate. Migration of charge across the linchpin via a [1,4]-Brook rearrangement with electrophile capture affords a three-component propargyl ether adduct. Herein, we report the development, scope, and limitations of this reaction sequence.

Graphical Abstract

The challenge of organic synthesis is not only the ability to synthesize a target but also the efficiency with which the target is constructed. In this regard, Anion Relay Chemistry (ARC) comprises a one-pot, three-component union tactic in which a linchpin component sequentially forms a C-C bond to each of two other components via an anionic reaction cascade.¹ This tactic permits the assembly of complex polyketide fragments from simple building blocks, thereby minimizing the number of synthetic steps required to achieve the synthetic target. Accordingly, our group has exploited ARC to permit the total syntheses of several natural products, including (−)-nahuoic acid Ci,² (−)-mandelalide A,³ and (−)-enigmazole A.⁴ The Type II ARC protocol is initiated by [1,2]-addition of an organolithium reagent to linchpin 1, followed in turn by a [1,4]-Brook rearrangement⁵ and electrophile capture to afford the three-component adduct 2 (Figure 1a). Here, diastereoselectivity arises via Felkin-Anh control with the absolute configuration set by the α-methyl stereogenic center present in 1.^6,7

Figure 1.

(a) Prior ARC Work: Diastereoselectivity; (b) Prior Asymmetric [1,2]-Addition of Lithium Acetylides to Ketones; (c) Proposed ARC Tactic

While stereoretentive and diastereoselective processes have been successfully incorporated into the ARC protocol, an enantioselective variant has yet to be developed. We reasoned that this could be achieved via initiation of the ARC sequence with an asymmetric [1,2]-carbonyl addition. Asymmetric organometallic [1,2]-carbonyl additions commonly employ Zn acetylides,⁸ although Sn, B, Li, and other metals have been employed.⁹ This approach has been exploited by others, including Marek¹⁰ and Johnson,¹¹ to access three-component couplings initiated by a Zn acetylide. However, these examples rely on a [1,2]-Brook rearrangement of the [1,2]-addition intermediate, whereas our envisioned strategy would comprise a [1,4]-Brook rearrangement; a protocol for the [1,4]-Brook rearrangement of Zn alkoxides remains elusive. Thus, our focus has been on lithium acetylides.

The enantioselective [1,2]-carbonyl addition of a lithium acetylide was first disclosed by Mukaiyama¹² and later improved by the Merck¹³ group in the synthesis of the HIV-1 reverse transcriptase inhibitor efavirenz. However, these early examples required several equivalents of ligand to impart high enantioselectivity. Recently, Nakajima reported a catalytic method for the asymmetric [1,2]-addition of lithium acetylides to ketones to yield enantioenriched propargyl alcohols utilizing BINOL derivative (R)-6 (Figure 1b).¹⁴ We sought to utilize (R)-6 in conjunction with a lithium acetylide to access enantioenriched alkoxide 8 via an asymmetric [1,2]-addition to achiral linchpin 7 to obtain the enantioenriched three-component adduct 9 after [1,4]-Brook rearrangement and electrophile capture (Figure 1c).

Our efforts were first directed toward evaluation of the enantioselectivity of [1,2]-addition of lithium phenyl acetylide to linchpin 10 to afford propargyl alcohol 11 (Table 1). The reaction was initiated by deprotonation of phenylacetylene in the presence of (R)-6 at −78 °C, followed by the addition of linchpin 10 over a period of 10 min. The rate of ketone addition was examined first. Decreased enantioselectivity was observed with rapid dropwise addition of linchpin 10, with no significant difference in enantioselectivity observed when linchpin 10 was added over a 30 min period. Next, the influence of silyl substitution on the linchpin was assessed by comparing TMS and TBS linchpins 10 and 12, respectively. Greater enantioselectivity was observed for [1,2]-addition to 10 than 12. Importantly, reduction of alkyne stoichiometry to 1 equiv afforded excellent enantioselectivity.

Table 1.

Optimization of Enantioselectivity Utilizing ARC Linchpin^a

^an-BuLi (2.2 equiv); acetylene (2.0 equiv); (R)-6 (10 mol %); 10 (1 equiv); HCl (1.5 equiv); [Si] = TMS (11a); TBS (11b).

^b10 added over 10 min;

Having optimized the enantioselectivity of the [1,2]-addition, we turned our attention to the full ARC sequence. The temperature dependence of the [1,4]-Brook rearrangement and electrophile capture was examined first (Table 2). In these experiments, benzyl bromide (BnBr) was added to the reaction mixture after completion of the [1,2]-addition at −78 °C. Upon addition of BnBr (2.0 equiv), the mixture was warmed to a range of temperatures from −78 to +50 °C and held at the corresponding temperature for 1 h before the reaction was quenched. We anticipated the formation of products 13 and 14, whereby [1,4]-Brook rearrangement is followed either by alkylation or protonation, respectively. The [1,4]-Brook rearrangement was observed at temperatures as low as −30 °C. The highest ratio of alkylation to protonation product (13/14) was observed with immediate transfer of the reaction vessel to a water bath at room temperature. Allowing the reaction to gradually warm to room temperature was deleterious, as was either addition of HMPA, or reduction of the BnBr stoichiometry. Additionally, maintaining the temperature of the reaction at −30 °C, the lowest temperature at which [1,4]-Brook rearrangement was observed, for 24 h did not increase the formation of 13.

Table 2.

Optimization of ARC Sequence^a

entry	variation (BnBr addition)	1H NMR ratio (13:14)
1	maintained at −78 °C	b
2	warmed to −50 °C	b
3	warmed to −30 °C	54:46
4	warmed to 0 ° C	59:41
5	warmed to rt	75:25
6	warmed to 50 ° C	63:37
7	warmed to −30 °C, 24 h	23:77
8	gradually warmed to rt	69:31
9	HMPA added	60:40
10	1.0 equiv of BnBr	67:33

^an-BuLi (1.2 equiv); acetylene (1.0 equiv); (R)-6 (10 mol %); 10 (1 equiv); BnBr (2 equiv); 10 added over 10 min.

^bOnly 11 observed.

With optimized conditions for the full ARC sequence in hand, the nucleophile scope was investigated. Reactions were performed with linchpin 10 and BnBr as the electrophile (Scheme 1). The influence of steric and electronic factors of the lithium acetylide on enantioselectivity was clear. Electron-rich acetylides afforded greater enantioselectivity than electron-poor acetylides. This trend is most evident in three-component adducts 16a-16c, wherein enantioselectivity and yield are positively correlated with the electron-donating character of the para substituent.¹⁵ Perhaps of greater significance is the influence of steric encumbrance of the nucleophile on enantioselectivity. Bulky nucleophiles such as 15d_i (TBS acetylene) and 15d_ii (TMS acetylene) afforded 16d with modest to moderate enantiomeric ratios of 72:28 and 90:10, respectively. This steric influence was further exemplified in the comparison of cyclic and linear aliphatic lithium acetylides, wherein a greater enantiomeric ratio was observed for 16f than 16e. Pleasingly, the use of benzyl propargyl ether was well tolerated, providing access to diversifiable three-component adduct 16g in good yield and excellent enantioselectivity.

Scheme 1. Nucleophile Scope^a.

^an-BuLi (1.2 equiv); 15 (1.0 equiv); (R)-6 (10 mol %); 10 (1.0 equiv); BnBr (2.0 equiv); 10 added over 10 min; TBAF deprotection of crude product. ^bTBS acetylene (15d_i). ^cTMS acetylene (15d_ii).

Next, the scope of electrophiles was evaluated using lithium phenyl acetylide as the nucleophile and ketone linchpin 10 to afford 18 (Scheme 2). Benzyl and allyl bromide electrophiles provided the desired three-component adducts 18a and 18b in good yield. Diyne product 18c was obtained from the corresponding silyl propargyl bromide after TBAF deprotection, albeit in modest yield. While an excellent yield of 18d was achieved using MeI, a poor yield was observed for 5-bromo-1-pentene, likely due to the competing elimination reaction. Epoxide electrophiles were found to be viable, affording adducts 18f-18h in modest to good yield, permitting inclusion of an additional stereocenter. In each case, the product was observed by ¹H/¹³C NMR as a single diastereomer. Benzaldehyde was also successfully employed, proceeding with high enantioselectivity (99:1 er) to yield 18i as an expected mixture of diastereomers (1.5:1).

Scheme 2. Electrophile Scope^a.

^an-BuLi (1.2 equiv); acetylene (1.0 equiv); (R)-6 (10 mol %); 10 (1.0 equiv); El (2.0 equiv); 10 added over 10 min; crude TBAF deprotection. ^bNo TBAF used. ^c(S)-epichlorohydrin was used as electrophile 17f.

An effective cascade reaction should afford the final product in greater yield than that obtained by the sequential execution of the constituent reactions. The ARC cascade was therefore divided into a two-pot sequence for comparison. The first step of the ARC sequence is [1,2]-addition of lithium phenylacetylide to linchpin 10 to arrive at alkoxide 19. A [1,4]-Brook rearrangement of alkoxide 19 is then thermally triggered with subsequent electrophile capture terminating the standard one-pot ARC sequence (Scheme 3). Alternatively, alkoxide 19 can be acidified at −78 °C to afford alcohol 11. In the two-pot protocol, purified alcohol 11 enters the ARC sequence via deprotonation to 19. The one-pot and two-pot methods were compared for the ARC sequence employing linchpin 10 with phenylacetylene and benzaldehyde serving as the nucleophile and electrophile, respectively, to afford three-component adduct 18i. A drastic difference in the product distribution was observed between the one- and two-pot methods. Whereas the one-pot protocol provided 18i in good yield (73%), the two-pot process provided only trace 18i with most of the recovered material consisting of quenched product 14 (89%) and benzyl alcohol (48%). We suspect the reduction of benzaldehyde to benzyl alcohol occurs via a single-electron transfer (SET) process that is favored by the two-pot reaction sequence.¹⁶ The role of the two-pot approach in favoring a SET pathway is unclear but may be a consequence of a different aggregation state that occurs via access of 19 by an alternate route. Thus, this one-pot asymmetric three-component union tactic circumvents the undesired reactivity and is more efficient than the sum of its parts.

Scheme 3. Comparison of One-Step and Two-Step ARC Approaches^a.

^aTBAF deprotection of products. ^bn-BuLi (1.2 equiv), 10 (1.0 equiv) THF, −78 °C 2 h; PhCHO (2.0 equiv). ^c11 (1.0 equiv), n-BuLi (1.2 equiv), THF, −78 °C 30 min, then PhCHO (2.0 equiv); both (R)-6 and no catalyst yielded identical results.

Having demonstrated the generality of the one-pot protocol with respect to acetylides and electrophiles, we examined the viability of aldehyde linchpins 24a and 24b to afford secondary alcohol 25 with phenylacetylene and BnBr serving as the nucleophile and electrophile, respectively (Scheme 4). Pleasingly, good yields and moderate enantioselectivities were achieved employing aldehyde linchpins 24a and 24b. As opposed to ketone linchpins 10 and 12 (Table 1), the size of the silyl substituent did not affect the enantiomeric ratio of the corresponding three-component adducts 25. Thus, either silyl ether product can be obtained at no expense of enantioselectivity, permitting convenient integration into a protecting group strategy.

Scheme 4. Aldehyde Linchpin^a.

^an-BuLi (1.2 equiv); acetylene (1.0 equiv); (R)-6 (10 mol %); 24 (1.0 equiv); BnBr (2.0 equiv); 24 added over 10 min; crude TBAF deprotection. ^b24a. ^c24b.

The greatest limitation to the generality of this tactic with respect to enantioselectivity is derived from the steric factors of the acetylene. Although many lithium acetylide substrates enabled excellent enantioselectivity with linchpin 10, synthetically useful silyl acetylenes afforded moderate enantioselectivity for three-component adduct 16d (Scheme 1). To overcome this limitation, the relationship between catalyst loading and enantioselectivity was investigated (Figure 2). First, we examined the ARC sequence with highly sterically encumbered TBS acetylene 15d_i and linchpin 10 to afford three-component adduct 16d. Acetylene 15d_i was chosen because it is the least ideal nucleophile examined herein. A logarithmic relationship between enantiomeric excess (% ee) and catalyst loading (mol % (S)-6) was observed, reaching a maximum of 80%ee (90:10 er) with 80 mol % (S)-6. Alternatively, excellent enantioselectivity (97:3 er) can be achieved for the same synthon type using TMS acetylene 15b_ii. Generally, most carbon substituted terminal acetylenes are less sterically hindered than a TBS group and as such, this method holds the promise of excellent enantioselectivity for a broad range of terminal acetylenes. In addition, (S)-6 was employed instead of (R)-6 to demonstrate the interchangeable nature of the catalyst to afford the opposite enantiomer.

Figure 2.

Catalyst loading experiments with hindered silyl acetylenes. (a) Conditions: n-BuLi (1.2 equiv); acetylene (1.0 equiv); 10 (1.0 equiv); electrophile (2.0 equiv); 10 added over 10 min; crude TBAF deprotection;

In summary, we disclose here a new method for the preparation of enantioenriched three-component union adducts from achiral components via an ARC tactic initiated by an organocatalytic asymmetric [1,2]-addition of a lithium acetylide. Excellent enantioselectivities were observed in the coupling of a variety of acetylenes and electrophiles with ketone linchpin 10. Reduced enantioselectivity observed with highly sterically encumbered acetylenes can be attenuated with increased catalyst loading to permit good to excellent enantioselectivity. This tactic enables effective diversity-oriented synthesis by virtue of not only a three-component nature but also the control over absolute configuration that is exerted via the application of either (S)-Ph₂BINOL-6 or (R)-Ph₂BINOL-6.