Soft Enolization of 3-Substituted Cycloalkanones Exhibits Significantly Improved Regiocontrol vs Hard Enolization Conditions

Abstract

Soft enolization conditions are revealed to be markedly better than the typically applied hard enolization protocols for regioselective enoxysilane formation from unsymmetrical 3-substituted cycloalkanones. Five-, six-, and seven-membered cycloalkanones each with either 3-methyl, 3-isopropyl, or 3-phenyl substituents were investigated and, in all but one case, regioselectivities were ≥11:1 for enolization away from the substituent. The selectivities observed are significantly improved over those obtained using strong anionic bases, and are complementary to the regiospecific enoxysilane formation derived from cycloalkenone conjugate addition/enolate silylation.

Graphical Abstract

Enoxysilanes are ubiquitous, high-value intermediates for organic synthesis.¹ They feature in powerful carbon–carbon bond forming reactions such as Mukaiyama aldol² or Mukaiyama–Michael³ additions, and also behave as weak nucleophiles toward strongly electrophilic reagents (e.g. Br₂, m-CBPA, PhSCl, etc.).⁴ Alternatively, enoxysilanes can be treated with certain nucleophiles, such as MeLi, LiNH₂, or fluoride sources to generate at will a reactive enolate species, often the same one from which they were initially formed.⁵ Oxidation can also lead to the net dehydrogenation of the parent ketone, with palladium-catalyzed⁶ and hypervalent-iodine-mediated⁷ processes commonly used to generate enones. Enoxysilanes also undergo productive oxidative couplings to generate new C–C bonds.⁸ Another valuable transformation is the direct conversion to alkenyl triflates with preservation of alkene position and geometry, thus permitting a range of cross-coupling reactions.⁹ Finally, direct cross-coupling of enoxysilanes can be achieved.¹⁰

Ketone-derived enoxysilanes are typically prepared via three major reaction types: (1) from the corresponding ketone by enolization; (2) from an α-heteroatom-substituted ketone by reduction; or (3) from an α,β,-unsaturated enone via conjugate reduction/addition. While the latter two processes are regiospecific, the first protocol is the most common, and can have issues of regiochemical control for unsymmetrical ketones with similar substitution patterns on each side of the carbonyl carbon.

It is well-known that regioselectivity of enolization can in some cases be dictated by the choice of reaction conditions and the steric environment of the substrate. For instance, treatment of 2-substituted cycloalkanones under hard enolization conditions (e.g., LDA; TMSCl) selectively generates the kinetic, less substituted enoxysilane product, whereas with treatment under equilibrating conditions (e.g., Et₃N, TMSCl, DMF, heat) the thermodynamic, more substituted, enoxysilane is formed preferentially.¹ These selective methods are limited to unsymmetrical ketones whose substitution levels differ at the α-carbons.

A survey of the literature shows that the protocol most used for the formation of the Δ^1,n-enoxysilanes from 3-substituted cycloalkanones is treatment with a strong base (most often LDA or LiTMP) to generate an enolate followed by trapping with a chlorotrialkylsilane (Scheme 1a). Although this procedure typically favors enolization away from the 3-substituent, regioisomeric mixtures of Δ^1,n (2) and Δ^1,2 (3) products are formed, often with low to moderate selectivity, and are typically inseparable by silica gel chromatography.^11,12 It is well known that the Δ^1,2 enoxysilane can generally be obtained regiospecifically from the corresponding α,β-unsaturated enone via conjugate addition of an organocuprate/hydride reagent or dissolving metal reduction, with subsequent trapping with chlorotrialkylsilane reagents (Scheme 1b.).^13,14 This communication describes a highly regioselective synthesis of enoxysilanes derived from 3-substituted cyclic ketones, wherein the application of experimentally simple soft enolization conditions vastly improves the regioselectivity “away” from the 3-substituent, compared with the more standard hard enolization conditions (Scheme 1c). As a result, it complements the regiospecific formation of enoxysilanes formed by enone conjugate addition/reduction. We demonstrate that this soft enolization protocol works for varying ring sizes and with substituents of different steric environments at C3. This advance is especially important in light of the difficulty of separation of the regioisomeric enoxysilanes—owing in equal measures to their similar polarities and instabilities—which leads to the formation of isomeric products in the subsequent step.

Scheme 1.

Methods for synthesizing TMS-alkenyl ethers from 3-substituted cyclohexanones

In the context of recent work in our laboratory toward the synthesis of the pseudopterosins,¹⁵ we learned that (R)-3-methylcyclohexanone underwent a highly regioselective Mukaiyama-type aldol condensation with acetaldehyde when the enoxysilane was formed in situ using soft enolization conditions (Eq. 1).

(1)

A careful study of the enolization portion of this reaction (Table 1) was undertaken. Application of the hindered strong base LiTMP with TMSCl was only moderately regioselective (3:1 rr, entry 1); this result was consistent with the ratio previously reported using LDA and TMSCl with the same substrate;^11b more importantly, it served as a benchmark against which to compare future results. The initial enoxysilane formation corresponding to the Mukaiyama aldol condensation in Eq.1 was highly regioselective: the combination of Hünig’s base (diisopropylethylamine) and TMSOTf led to a 13.5:1 mixture in favor of the Δ^1,6 product (entry 2). The reactivity and selectivity were sensitive to the identity of the amine base. The less bulky triethylamine was not as selective as diisopropylethylamine (7.5:1 rr, entry 3); however, the very hindered PMP (1,2,2,6,6-pentamethylpiperidine) proved marginally less selective (entry 4). The amidine base DBU required heating for effective reaction and resulted in a decreased regioselectivity of 2.2:1 (entry 5). Unsurprisingly, the conditions that we identified in entry 2 are just as successful with TBSOTf, thus yielding the TBS alkenyl ether with a 14:1 rr (entry 6).¹⁶

Table 1.

Initial studies of the regioselectivity in enoxysilane formation^a

^a0.5 mmol ketone, [ketone] = 0.2 M, −78 °C, 1 h, 3.5 equiv. base, 3 equiv. Lewis acid.

^bDetermined by ¹H analysis of unpurified reaction mixture.

^c1.4 equiv. BuLi, 1.5 equiv. TMP, 1.4 equiv. TMSCl.

^dPMP = 1,2,2,6,6-pentamethylpiperidine.

^eHeating to 40 °C was required for reactivity, 1.2 equiv. DBU, 1.1 equiv. TMSCl.

With effective conditions (i-Pr₂NEt, 3.5 equiv and TMSOTf, 3.0 equiv at −78 °C in CH₂Cl₂) established, cyclopentanones, cyclohexanones, and cycloheptanones¹⁷ with methyl, phenyl, and isopropyl C3-substituents were evaluated (Table 2). In the cyclopentanone series, an increase in regioselectivity was observed with an increase in the bulk of the substituent.^18,19 For the six- and seven-membered ring ketones, universally high selectivities (≥12:1) were observed. In all cases except for 3-methylcyclopentanone, the chromatographically isolated yields of pure material (as regioisomeric mixtures) were high, and closely mirrored the yields calculated by NMR spectroscopy with integration against an internal standard. To ensure that soft enolization was advantageous with all ring sizes, we also evaluated standard hard enolization conditions for 3-methylcyclopentanone and 3-methylcycloheptanone. The regioselectivity regioselectivity was again greatly enhanced upon switching from hard enolization to soft enolization, (1.5:1 to 5.8:1 and 2.6:1 to >20:1), respectively. The near perfect selectivities in all three cycloheptanone cases is noteworthy.¹⁷

Table 2.

Scope of regioselective enoxysilane formation^a

^a0.5 mmol scale; [10] = 0.2 M; yields reported as a mixture of regioisomers; regioselectivies determined by ¹H NMR analysis of unpurified reaction mixtures.

^bIsolated yield reported after column chromatography on basic alumina.

^cNMR yield reported using 1,3,5-trimethoxybenzene (20 mol%) as an internal standard.

Of course, in nearly all instances, enoxysilane formation would be coupled with a subsequent synthetically useful transformation.¹ We therefore document the use of these intermediates in (a) Mukaiyama aldol addition (Scheme 2a)²; (b) enolate regeneration and alkylation (Scheme 2b)⁵; and dehydrogenation (Scheme 2c)⁷. The products of each of these reactions is obtained in high yield over the two steps, wherein chromatographic purification of the enoxysilane was not required.

Scheme 2.

Synthetic utility of enoxysilane intermediates

The selectivity observed in enoxysilane formation was surprising, given the seemingly small amount of steric differentiation, especially in methyl-substituted cases. However, the simplest explanation for the selectivity is a steric argument, which is consistent with moderate increases in selectivity as the 3-substituent size increases,^18,19 and with the notable increase in selectivity upon going from triethylamine to Hünig’s base. Because of the ubiquity of Hünig’s base, we opted against evaluating multiple more hindered, less omnipresent bases, given the already quite remarkable selectivities observed.

Given the well-established ability to generate 3-substituted Δ^1,2-enoxysilanes regiospecifically from conjugate reactivity of cycloalkenones, it is noteworthy that highly selective access to the other regioisomer of enoxysilanes derived from C3-substituted cycloalkanones has been largely unattainable until now. We have identified operationally trivial conditions for the selective formation of this regioisomer with good scope across five-, six-, and seven-membered cycloalkanones. Further, we know that this reaction works well on a preparative scale, having conducted the reaction sequence shown in Eq. 1 on an 87 mmol scale.¹⁵ The simple soft enolization protocol described here should, as a result, be readily adopted in chemical synthesis.