Diels–Alder Reactions of 1-Alkoxy-1-amino-1,3-butadienes: Direct Synthesis of 6-Substituted and 6,6-Disubstituted 2-Cyclohexenones and 6-Substituted 5,6-Dihydropyran-2-ones

Abstract

We report the cycloaddition reactions of 1-alkoxy-1-amino-1,3-butadienes. These doubly activated dienes are prepared on a multigram scale from crotonic acid chloride and its derivatives. The dienes undergo Diels–Alder (DA) and hetero-Diels–Alder (HDA) reactions under mild reaction conditions with a variety of electron-deficient dienophiles to afford cycloadducts in good yields with excellent regioselectivities. The hydrolysis of the DA cycloadducts provides 6-substituted and 6,6-disubstituted 2-cylohexenones, which are versatile building blocks for complex molecule synthesis. The corresponding HDA cycloadducts afford 6-substituted 5,6-dihydropyran-2-ones.

Graphical Abstract

The Diels–Alder (DA) reaction is one of the most important transformations in organic chemistry, providing direct access to six-membered cyclic compounds in a regio- and stereocontrolled manner with up to four chiral centers.¹ The power of the DA reaction is evident from its indispensable role in the synthesis of numerous complex molecules.² Of special importance in the development of this reaction has been the advent of a suite of heteroatom-substituted dienes, which not only are more reactive but also yield a wide range of functionalized building blocks for chemical synthesis.³ The introduction of Danishefsky’s diene (1, Scheme 1a), for example, enabled the facile synthesis of various 4,4-disubstituted cyclohexenones (and further substituted derivatives thereof), which paved the way to many intricate natural products.⁴ The development of the 1-amino-derivatives of this diene (i.e., 3, Scheme 1b), which is considerably more reactive, opened further opportunities in synthesis,^5–7 including the development of enantioselective DA reactions.⁸ Given the importance of 6,6-disubstituted cyclohexanone cores (5) as building blocks for the synthesis of complex molecules⁹ and the paucity of methods to access them, we investigated various additional heteroatom-substituted butadienes and their cycloadditions and report here the results of our studies on the synthesis and DA and hetero-Diels–Alder (HDA) reactions of 1-alkoxy-1-amino-1,3-butadienes.

Scheme 1.

Activated Butadienes for Diels–Alder Reactions

The synthesis of 6,6-disubstituted cyclohexenones (5) via a DA cycloaddition requires either vinyl ketene (6) or its formal equivalent (Scheme 1c). To realize this capability, several 1,1-dialkoxybutadienes have been developed and examined (7a) in cycloaddition reactions.¹⁰ Notably, Sustmann reported that whereas 1,1-dimethoxybutadiene gave the expected cycloadducts with highly electron-deficient dienophiles such as dimethyl 2,3-dicyanomaleate, its reactions with common dienophiles, such as methyl acrylate, acrylonitrile, fumaro- and maleonitrile, dimethyl fumarate, and dimethyl maleate, gave no cycloadducts and only polymeric materials.^10d Among the 1,1-dialkoxybutadienes, the most important is Brassard’s diene (7b, Scheme 1d). Although used widely for HDA and Mukaiyama aldol reactions, its successful use in DA reactions is primarily with quinone or doubly activated dienophiles.¹¹ Additionally, the cycloadducts it generates are necessarily more highly oxygenated, giving 3-alkoxycyclohexenone products, the masked form of 1,3-cyclohexanediones, rather than 2-cycohexenones. The related 1-alkoxy-1-aminobutadiene (cf. 8), which is expected to be even more reactive, has seen limited use for DA reactions. Indeed, the reaction of 8b with dimethyl acetylenedicarboxylate did not afford the expected DA adduct, instead giving a product (9) “with a substitution pattern incompatible with the normal Diels–Alder pathway”.¹²

We reasoned that the poor DA reactivity of 1-alkoxy-1-aminobutadienes such as 8 was likely due to steric interactions that disfavor the s-cis rotamer that is required for DA reactions, instead allowing alternate reaction paths (Scheme 1e).¹³ Given this background of literature reports, we investigated the oxazolidine-fused butadiene 10, wherein the N and O atoms are linked through a two-carbon unit, thereby obviating the steric issues. The desired diene was synthesized in good yield through a simple protocol starting with Woollaston’s route to α,β-unsaturated oxazoline 12a (Scheme 2).¹⁴ This oxazoline was then converted into the desired diene in two steps via the formation of the oxazolinium salt followed by deprotonation with NaHMDS. Through this route, we prepared both the base diene 10 and the gem-dimethyl-substituted diene 13. An alternate synthesis of the diene was also developed to overcome the long reaction times and the difficult isolation procedure, especially the distillation of the thermally unstable oxazolines 12. Crotonyl chloride was reacted with N-methylethanolamine, and the resulting amide 14 was treated with triflic anhydride, which induced the desired cyclization to give oxazolonium triflate salt 15. Deprotonation of 15 with NaHMDS then proceeded cleanly to give the desired diene in 71% overall yield from crotonyl chloride. Whereas the diene is unstable in aqueous solutions of pH <10, we found that it can be subjected to a 2 M NaOH/H₂O solution with no degradation. By quenching the reaction with such a solution, all polar nonvolatiles can be removed by extraction, and the desired diene can be obtained pure without the need for distillation. This improved route is shorter and affords the diene in high yield, requiring no distillation or columns. Importantly, intermediate 15 is stable for an extended period of time, even when stored at room temperature. The improved route was used to prepare over 15 g of salt 15 and 4 g of diene 10 in a single pass.

Scheme 2.

Synthesis of Oxazolidine-Fused Butadienes

The initial studies were aimed at assessing the cycloaddition capability of the new dienes. Upon heating a solution of diene 10 and methacrolein in toluene to 60 °C for 2 h, the diene was fully consumed and yielded a 3:1 mixture of two products, as observed by NMR. The major product was the expected cycloadduct, and the minor product was tentatively assigned to be the HDA adduct.^15a The major product was unstable to silica gel but could be hydrolyzed to give the desired 6,6-disubstituted cyclohexanone 17a (Scheme 3). The analogous reaction with the gem-dimethylated diene 13 gave a cycloadduct (cf. 16, 30%) that was column-stable, allowing the confirmation of its structure. However, the DA reaction proceeded significantly more slowly, so diene 13 was not further investigated.^15b

Scheme 3.

Diels–Alder Reactions of Diene 10 with Dienophiles

^aDA reactions run in a sealed tube. ^bExpected cycloadduct not formed. ^cMixture of keto and enol forms.

Various parameters were examined to improve the reaction outcome with diene 10. When carried out in toluene at room temperature, the reaction required 10 h to go to completion and gave a similar ratio of the two products. In hydrogen-bond donor solvents (e.g., t-BuOH), the reaction rate of the HDA reaction increased, and the reaction gave a lower proportion of the desired DA cycloadduct. The best outcome, albeit by a small margin, was obtained when the reaction was performed in benzene. Upon optimization, the DA reaction and the hydrolysis could be performed in a single procedure that afforded ketone 17a in 70% isolated yield.

To evaluate the generality of the protocol, we reacted diene 10 with several common dienophiles (Scheme 3). Ethyl- and n-butyl-acroleins reacted analogously to methacrolein and afforded the respective 6,6-disubstituted 2-cyclohexenones in good yields. We were delighted to find that even tiglic aldehyde participated in the cycloaddition to give, after hydrolysis, trisubstituted cyclohexenone 17d. The reactions with acrylonitrile and methyl acrylate proceeded well, as did the reaction with methyl maleate. Unfortunately, the reaction with methyl vinyl ketone gave no cycloadduct 17e.^15c

The useful reactivity shown by diene 10 in DA reactions with traditional dienophiles motivated us to examine its reactions with nitroalkenes (Scheme 4). Whereas nitroethylene is reported to react at room temperature with highly active dienes like cyclopentadiene, the DA reaction of β-arylnitro-ethylenes generally requires higher temperatures or special activation modes.¹⁶ In light of this limitation, we were delighted to observe that the oxazolidine-fused butadiene 10 rapidly reacted at room temperature with β-nitrostyrene to give a cycloadduct (cf. 18), which upon quenching with aqueous oxalic acid gave the expected 6-nitro-substituted cyclohexenone 19a in 75% yield.¹⁷ Several additional β-arylnitro-ethylenes and two β-alkyl-substituted nitroethylenes were subjected to the cycloaddition/hydrolysis protocol, and all gave the cyclohexanone products in good to excellent yields. Nitroethylenes with aryl units possessing donor groups or withdrawing groups worked equally well, as did naphthyl- and heteroaryl-substituted nitroethylenes. The two alkyl-substituted β-nitroalkene products are also noteworthy, in particular, the spiro-fused bicyclic compound 19i, which was formed in 78% yield. The present method offers a simple route to various 6-nitrocyclohexenones, the chemistry of which appears to have been scarcely investigated.¹⁸

Scheme 4.

Diels–Alder Reactions of Diene 10 with Nitroalkenes

^aYield in parentheses is the NMR yield of the cycloadduct (18).

We next turned our attention to the preparation and DA reactivity of more substituted analogs of diene 10 (Scheme 5). Three different dienes were synthesized using the first protocol described above, starting with the requisite acid chlorides. The procedures transferred well and enabled the synthesis of gram quantities of the different dienes, which were isolated as colorless liquids that were stored under an inert atmosphere. The dienes reacted with several common dienophiles to afford, after the in situ hydrolysis of the cycloadducts, the expected cyclohexanone products in good overall yields (Scheme 6).¹⁹ Given the robustness of diene preparation and the generality of the DA reactions, the present method provides facile access to various functionalized mono- and bicyclic systems that should prove to be of value in complex molecule synthesis.

Scheme 5.

Synthesis of Substituted Oxazolidine-Butadienes

Scheme 6.

Diels–Alder Reactions of Substituted Oxazolidine–Butadienes with Various Dienophiles

To further expand the scope of the cycloadditions of diene 10, we examined its HDA reaction with aldehydes, which would provide a simple and direct route to 6-substituted dihydro-2-pyrones. This subunit is found in many bioactive natural products and consequently is the subject of much synthesis work.²⁰ As previously noted, we had observed the formation of a labile side product that was presumed to be the HDA adduct. To capitalize on this observation, we carried out the reaction of 10 with benzaldehyde (PhH, 60 °C) and were delighted to observe the clean formation of cycloadduct 28, as confirmed by NMR. As the cycloadduct proved labile to isolation, the reaction was directly quenched with aqueous oxalic acid, which promoted its hydrolysis to afford the α,β-unsaturated δ-lactone product 29a in 70% yield. Given the simplicity of the procedure, we examined the HDA reaction of 10 with several common aldehydes and found the process to be useful for both electron-poor and electron-rich aromatic aldehydes (Scheme 7). Aliphatic aldehydes were unreactive under the conditions used.

Scheme 7.

HDA Reaction of Diene 10 with Aromatic Aldehydes

The breadth of facile reactions observed with diene 10 and its more substituted derivatives motivated us to benchmark its reactivity against other highly reactive dienes, such as Danishefsky’s diene (1), 1-amino-3-siloxybutadiene (3), and its carbamate derivative (30). The kinetic measurements were carried out at 60 °C in C₆D₆, and the product concentrations were monitored by ¹H NMR. The second-order rate constant for the reaction between diene 10 and diethyl fumarate in benzene was determined to be 2.6 × 10⁻⁴ M⁻¹ s⁻¹ (Table 1).^15b For diene 1 and carbamate diene 30, the rate constants are 4.1 × 10⁻⁵ M⁻¹ s⁻¹ and 3.5 × 10⁻⁵ M⁻¹ s⁻¹, respectively. Also listed are the reported rate constants for the reaction between the 1-amino-3-siloxy diene 3 and diethyl fumarate at 17 °C and with methacrolein at 17 and 60 °C.²¹ The results show that whereas Danishefsky’s diene 1 and carbamate diene 30 react with fumarate at approximately the same rate, diene 10 reacts nearly seven times faster. All three dienes reacted two to three times faster in chloroform. Interestingly, although dienes 3 and 10 have similar heteroatom substituents, the latter is considerably less reactive, likely due to the steric hindrance from the cis-oriented oxygen.

Table 1.

Rate Constants for DA Reactions of Some Reactive Dienes

entry	diene	dienophile	temperature (°c)	k2 (m−1s−1)	relative rate
1 a			60	2.6 × 10−4	7.1
2 a			60	4.1 × 10−5	1.1
3 b,c	1		17	3.6 × 10−6	0.1
4 a			60	3.5 × 10−5	1.0
5 a,b			17	1.5 × 10−3	42.8
6 a,b	3		17	2.0 × 10−3	57.1
7 b,d	3		60	2.0 × 10−2	571.4
8 b,c	3		17	1.2 × 10−2	342.9

^aRun in C₆D₆

^bValues from Kozmin et al.²¹

^cRun in CDCl₃

^dRun in Tol-d₈.

To get further insight into the relative reactivities of the dienes, we determined the activation parameters for the DA reactions of diethyl fumarate with dienes 1 and 10 (Figure 1). As expected, the activation energy (E_a) for the reaction with Danishefsky’s diene was found to be substantially larger than that with diene 10. Arrhenius plots extrapolated from the kinetic data indicate a much larger difference in the relative reactivities of dienes 1 and 10 at room temperature.^15b Interestingly, above 140 °C, diene 1 is predicted to react faster with diethyl fumarate than diene 10.

Figure 1.

Arrhenius plots and activation parameters for the reaction of dienes 1 and 10 with diethyl fumarate in toluene; [diene]0 = 0.2 M, [dienophile]0 = 0.6 M. Rate constants for 1 measured at 50, 60, and 70 °C. Rate constants for 10 measured at 40, 50, and 60 °C.

As the previously described results demonstrate, 1-amino-1-oxobutadienes represent an important addition to the family of reactive, heteroatom-substituted dienes. The parent diene can be synthesized in one step from a stable triflate salt precursor, and it and all related dienes can be prepared on a multigram scale. The new dienes undergo DA reactions with a broad range of dienophiles to afford, after in situ hydrolysis, a variety of 6-substituted 2-cyclohexenones, which should prove to be versatile building blocks for the synthesis of complex molecules. The HDA reactions of the parent diene with aldehydes give direct access to 6-substituted 5,6-dihydro-2-pyrones. Kinetics experiments indicate that the new diene, despite its added steric interactions, is significantly more reactive than other highly active dienes such as Danishefsky’s diene, especially at lower temperatures. Further expansion of the chemistry of these dienes, especially the development of enantioselective DA or HDA reactions or reactions with other heterodienophiles, is expected to greatly enhance their usefulness in chemical synthesis.