A Catalytic Intermolecular Formal Ene Reaction between Ketone-Derived Silyl Enol Ethers and Alkynes

Abstract

A catalytic formal ene reaction between ketone-derived silyl enol ethers and terminal alkynes is described. This transformation is uniquely capable of bimolecular assembly of 2-siloxy-1,4-dienes and can be used to access β,γ-unsaturated ketones containing quaternary carbons in the α-position.

Intermolecular alkenylation of enolates or their surrogates offers direct and convenient access to β,γ-unsaturated ketones, versatile building blocks that contain two orthogonal functionalization sites. Over 30 years ago, Migita reported pioneering examples of such sp²–sp³ coupling processes, which relied on a Pd(0)-catalyzed alkenylation of in situ-generated tri-n-butyltin enolates with substituted vinyl bromides.¹ Subsequent works from other laboratories further demonstrated the power of Pd(0)- and Ni(0)-catalyzed alkenylations in the synthesis of β,γ-unsaturated ketones,^2–8 although these reactions are typically limited to aryl alkyl or alkenyl alkyl ketones. Intermolecular alkenylations of dialkyl ketones bearing two unactivated enolizable positions are much less represented.^9,10 In this context, selective formation of quaternary centers poses a particular challenge, and only isolated examples can be found in the literature.^7,11 It is noteworthy that analogous intramolecular alkenylations are much more common and have found a number of applications in synthesis, likely owing to high levels of regiocontrol enforced by kinetic selectivity during cyclization.¹² Here we demonstrate a new catalytic formal ene reaction between silyl enol ethers and terminal alkynes that is uniquely capable of bimolecular assembly of 2-siloxy-1,4-dienes and can be used to access β,γ-unsaturated ketones containing quaternary carbons in the α-position.

During our work on the synthesis of paxilline indole diterpenes, we identified an indium(III) bromide-mediated alkenylation of a silyl enol ether with a terminal alkyne as suitable means for assembly of the requisite β,γ-unsaturated ketone.¹³ We became intrigued by the idea of performing intermolecular alkenylations in a catalytic fashion with retention of the silyl enol ether functionality in the product.¹⁴ Electrophilic activation of alkynes toward nucleophilic attack by silyl enol ethers has found extensive application in α alkenylation of ketones. In his pioneering work, Conia described the first, to our knowledge, examples of such reactivity.¹⁵ A number of reports of related Al(III)-mediated¹⁶ and transition-metal-catalyzed^17–22 intramolecular alkenylations of ketone-derived enoxysilanes followed in subsequent years, including some asymmetric variants.²³ However, relevant intermolecular alkenylations of silyl enol ethers find very little precedent,^13,24,25 and the only available catalytic intermolecular process requires the use of 1,3-dicarbonyls as activated precursors²⁶ (Scheme 1). In the latter example, facile enolization of the substrates enables the ene-like reaction with alkynes. A similar process between ketone-derived silyl enol ethers and alkynes is considerably more challenging, as it requires a net proton transfer from the allylic position of the enoxysilane to the former alkyne fragment that has never been observed in a bimolecular setting.²⁷ Furthermore, the propensity of silyl enol ethers derived from unsymmetrical dialkyl ketones to isomerize under acidic conditions^28,29 poses a potential issue of site selectivity. Finally, the product of the alkenylation, a 2-siloxy-1,4-diene, would itself represent a substrate for the reaction, posing an inherent issue of chemoselectivity. In particular, application of such a reaction to the selective construction of quaternary centers would require preferential alkenylation of fully substituted enoxysilane starting materials in the presence of less substituted enoxysilane products.

Scheme 1.

Comparison of Relevant Intermolecular Alkenylations with Current Work

With these considerations in mind, we set out to investigate the catalytic formal ene reaction between silyl enol ether 1 and 1-octyne (Table 1; see Supporting Information (SI) for additional details). Gold(I) complexes¹⁹ proved to be poor catalysts for the desired transformation with the best-performing combination of a NHC-derived precatalyst and a halidescavenging additive providing only small amounts of diene 2 and the hydrolyzed product 3 (entry 1). Application of zinc(II) bromide²² was unsuccessful (entry 2). Conditions similar to those previously employed by Nakamura²⁶ afforded only traces of enoxysilane 2 (entry 3). However, changing the counter-anion in the indium(III)-based catalyst produced instructive trends. While chloride (entry 4) offered only marginal improvement over triflate, bromide (entry 5) allowed for the efficient alkenylation of substrate 1, and only small amounts of ketone 3, allene 4,³⁰ and regioisomer 5 were observed. Application of indium(III) iodide and introduction of halide-scavenging additives were less efficient (entries 6 and 7). Catalytic alkenylation in the presence of indium(III) bromide could be performed at temperatures as low as 50 °C, which decreased the content of allene 4 and regioisomer 5 (entry 8). Attempted alkenylation in the presence of catalytic amounts of hydrogen bromide (generated in situ, entry 9) as well as triflic acid (entry 10) did not lead to product formation, confirming the catalytic role of the metal derivative. The robust TIPS-enol ethers³¹ performed better than more labile TMS and TBS derivatives. Interestingly, attempted alkenylation of isomeric silyl enol ether 6 afforded siloxydiene 2 as a major product, suggesting isomerization of 6 to 1 under the reaction conditions.

Table 1.

Development of the Catalytic Formal Ene Reaction

Entry	Conditionsa	2(%)b	3(%)b	4(%)b,c	5(%)b
1	5 mol%IPrAuCI,5 mol% NaBArF, 65℃	11	8	10	<1
2	10 mol% ZnBr2, 65℃	4	5	1	<1
3	5 mol% In(OTf)3, 65℃	<1	1	<1	<1
4	5 mol% InCI3, 65℃	5	9	<1	<1
5	5 mol% InBr3, 65℃	72	5	6	2
6	5 mol% InI3, 65℃	36	6	7	<1
7	5 mol% InBr3, 5 mol% NaBArF,65℃	40	11	3	<1
8d	5 mol% InBr3,50℃	65	9	3	<1
9	5 mol% MeOH, 8 mol% TMSBr, 65℃	<1	<1	<1	<1
10	5 mol% TfOH, 65℃	<1	<1	<1	<1
Attempted alkenylation of silyl enol ether 6:

^a0.25–0.5 mmol of 1, 1.5 equiv of alkyne, 0.25–0.5 mL of (CH₂Cl)₂, 20–28 h.

^bBased on internal standard and determined by 1H NMR.

^cdr = 1:1.

^d65% isolated yield of 2.

It is important to note that the product of double alkenylation was not observed during these studies. Alkenylation of the fully substituted enoxysilane 1 must therefore proceed considerably faster under our conditions than alkenylation of product 2. Although speculatively, we attribute the observed selectivity to the difference in the conformational preferences of the bulky siloxy group.³² To demonstrate the application of such unusual selectivity, the synthesis of diketone 9 was undertaken (Scheme 2). Starting diketone 8 contains three reactive α-positions, C1, C8, and C10, and selective alkenylation at C8 represents a considerable challenge. To our delight, subjection of the bis-silyl enol ether derived from diketone 8 to our alkenylation conditions resulted in a selective reaction of the fully substituted siloxyalkene fragment, delivering monoalkenylated product 9 after a mild hydrolytic workup.

Scheme 2.

Synthesis of Diketone 9

With a functional set of conditions in hand, we investigated the preliminary scope of this catalytic formal ene reaction. We found that alkyl acetylenes with aromatic, halide, and ether functionalities as well as conjugated alkynes successfully participate in the alkenylation process (dienes 10–14 in Table 2). The reaction also tolerated a Lewis basic ester group (diene 15), although more forcing conditions were required in this case.³³ Trimethylsilylacetylene and internal alkynes did not participate in the alkenylation. The formal ene reaction was effective at forming quaternary centers containing four non-methyl substituents (dienes 16 and 17), and various substitution patterns of the cyclopentene ring were tolerated (dienes 18–21). Synthesis of dienes 19–21 demonstrates a convenient approach to diastereoselective vicinal difunctionalization of a simple alkenone³⁴ and can be useful in the context of natural product synthesis.¹³ Cyclohexanone-derived silyl enol ethers also underwent successful alkenylation, and corresponding ketones 22 and 23 could be readily isolated after a mild hydrolytic workup.³⁵ Furthermore, acyclic silyl enol ethers were suitable substrates for this formal ene reaction (e.g., see diene 24). Because the resulting 2-siloxy-1,4-dienes often proved hydrolytically unstable relative to their cyclic counterparts, the reaction mixtures were subjected to a mild hydrolytic workup, delivering the corresponding β,γ-unsaturated ketones (products 25–30). In almost all cases, the desired 1,1-disubstituted alkenes were produced in a highly regioselective manner, and only small quantities of corresponding 1,2-disubstituted isomers were observed.

Table 2.

Preliminary Substrate Scope of the Alkenylation^a

^aTypically, starting silyl enol ethers were mixtures of isomers; see SI for details. 1 equiv of silyl enol ether, 1.5 equiv of alkyne, 10 mol% of InBr₃, (CH₂Cl)₂ (1 M in enoxysilane), 65 °C, ca. 24 h; rr ≥ 10:1 (except 15, rr 9:1; 24, rr 7:1). Siloxydienes (except 13, 14, and 24) contained ca. 5 mol% of inseparable allenes.

^bHeated to 50 °C.

^cHeated to 80 °C.

^d5 mol% of InBr₃.

^eNeat. ^f5 M in silyl enol ether.

^gHeated for ca. 72 h.

^hHeated for ca. 48 h.

ⁱAfter hydrolysis.

We propose that the mechanism of this catalytic formal ene reaction involves initial nucleophilic attack of enoxysilane 1 on the alkyne–indium(III) bromide complex³⁶ to form alkenylindium 31 (or the corresponding divinylindium derivative; Scheme 3).³⁷ The formation and stereochemistry of intermediate 31 or related species are supported by observation of deuterated ketone 34 upon treatment of the reaction mixture with CD₃OD.³⁴ Similar alkenylindium derivatives have been previously isolated and characterized by Baba.²⁴ Intermediate 31 can undergo protodemetalation with indium(III) bromide–ketone complex 32 to produce ketone 3 and enolate 33, respectively.^26,38 Participation of ketone 3, which is formed early and whose concentration remains constant during the reaction,³⁹ is supported by the observation that an additive of ketone 35 undergoes silylation in situ to form enoxysilane 36.³⁴ Subsequent silylation of enolate 33 is expected to afford the desired siloxydiene 2.

Scheme 3.

Proposed Mechanism of the Alkenylation

In summary, we disclose a new catalytic intermolecular ene reaction between ketone-derived silyl enol ethers and terminal alkynes. Among the features of this alkenylation are selective formation of highly sterically encumbered quaternary centers and direct access to 2-siloxy-1,4-dienes and corresponding β,γ-unsaturated ketones with synthetically challenging substitution patterns. The reaction has its limitations and future work will focus on expanding the substrate scope and further improving the efficiency of the process. However, this transformation has no alternatives among the current methods and can be expected to find broad application in natural product synthesis.