The chemoselective cyclisation of unsymmetrical γ-diketones to cyclopentenones by directed aldol reaction using a magnesium chelate

Abstract

The feasibility of synthesising disubstituted cyclopent-2-enones chemoselectively from unsymmetrical γ-diketones, using an aldol reaction directed by magnesium chelation, has been studied. Treatment of 3,3-dimethylhexane-2,5-dione 15 with aqueous sodium hydroxide (0.5 mol dm^–3) gives a mixture of 3,5,5-trimethyl-12 and 3,4,4-trimethylcyclopent-2-enone 14 in an isomer ratio of 2.2 : 1. Insertion of an α-methoxycarbonyl grouping as a control element allows formation of a magnesium chelate 17 when treated with magnesium methoxide, and the major product is then mainly the undehydrated aldol 21. This, when treated with aqueous sodium hydroxide (0.5 mol dm^–3) to dehydrate, hydrolyse and decarboxylate, gives the 3,5,5-trimethyl- and 3,4,4-trimethylcyclopent-2-enones in a nearly chemospecific ratio of 1 : 49. When 3-methoxycarbonyl-4,4-dimethylhexane-2,5-dione 16 is treated with aqueous sodium hydroxide (0.5 mol dm^–3), omitting the magnesium methoxide treatment, the corresponding ratio of cyclopentenones was still an interesting 1 : 7.3. Treatment with sodium methoxide in methanol gives, by contrast, γ-lactone 27.

Treatment of undecane-4,7-dione 34 with aqueous sodium hydroxide (0.5 mol dm^–3) at reflux gives 2,3-dipropylcyclopent-2-enone 36 and 3-butyl-2-ethylcyclopent-2-enones 35 in a ratio of almost 1 : 1. Treatment of 6-methoxycarbonylundecane-4,7-dione 33 with magnesium methoxide in methanol gives undehydrated aldol which when treated with aqueous sodium hydroxide gives the dipropylcyclopent-2-enone 36 and 3-butyl-2-ethylcyclopent-2-enones 35 in 9 : 1 ratio. Conversely, the 5-methoxycarbonyldione 32 gives a corresponding ratio of 36 to 35 of 1 : 9. The best ratios attained, 1 : 15 for 36 and 35 from 32 and 20 : 1 for 36 and 35 from 33, were when magnesium methoxide in refluxing benzene or toluene were employed. There is still a strong chemoselective effect when the magnesium treatment is omitted.

Preliminary examination of the corresponding cyclohex-2-enone systems gave poorer chemoselectivities when either procedure was employed. 6-Methoxycarbonyldodecane-5,9-dione 52 gave 2,3-dipropyl-54 and 3-butyl-2-ethyl-cyclohex-2-enone 55 in a ratio of ∼4 : 1 whilst the 5-methoxycarbonyldodecane-4,8-dione 53 gave a 1: ∼4 ratio.