A novel rearrangement of 2-bromo-1-(bromomethyl)ethyl esters into the corresponding 2-oxopropyl derivatives is reported. The mechanism of this transformation was studied by means of 18 O-and 2 H-labeling experiments. The reaction proceeds through a transient dioxolane intermediate that is hydrolyzed to form the corresponding 2-oxopropyl acetate.As part of our ongoing interest in the synthesis of bioactive compounds, 1 we recently developed an original approach for the synthesis of milnacipran, 2 an antidepressant currently on the market for the treatment of major depressive disorders. Our strategy involved the synthesis of a halo lactone related to 1 (Scheme 1) as a key intermediate that was subsequently transformed into the target drug. The results obtained in the context of that study prompted us to try to develop a general and alternative route to halo lactone 1 by intramolecular cyclization of substituted 2-bromo-1-(bromomethyl)ethyl phenylacetate (2a). However, various attempts to induce the formation of the lactone ring by carbocyclization under basic conditions systematically led to an unexpected rearrangement to the keto ester 3a (Scheme 1), which contained an extra oxygen atom and no bromine atoms. Intrigued by this peculiar behavior of compound 2a, we decided to investigate the scope and mechanism of this novel rearrangement.
Scheme 1 Reaction of 2-bromo-1-(bromomethyl)ethyl phenylacetate (2a) on treatment with a baseWe synthesized a series of halo esters 2 from the corresponding carboxylic acids by esterification with 1,3-dibromopropan-2-ol under dehydrating conditions, from the corresponding methyl esters by transesterification under acidic conditions, or from the corresponding acyl chlorides (Scheme 2).
Scheme 2 Synthesis of halo ester 2Before attempting to extend the scope of our novel rearrangement, we first optimized the reaction conditions for ester 2a by changing the solvent, base, and reaction temperature. Attempts to perform the reaction using 2.2 equivalents of sodium hydride in warm dimethyl sulfoxide (Table 1, entry 1) gave only a moderate yield of product 3a (54%); this could be slightly improved by performing the reaction in refluxing tetrahydrofuran, although the yield (62%) was still less than satisfactory (entry 2). We nevertheless selected tetrahydrofuran as the solvent and we examined the effects of other bases. The use of lithium diisopropylamide in refluxing tetrahydrofuran gave disappointing results, as no traces of the desired keto ester 3a were detected (entry 3). The use of potassium tert-butoxide in tetrahydrofuran at room temperature gave the expected compound in a poor yield (21%; entry 4); the best results were obtained with this base in refluxing tetrahydrofuran, which gave a 92% yield of keto ester 3a (entry 5).Having optimized the reaction conditions, we next focused on the use of other substrates. Substitution of the aryl moiety appeared to have little or no effect on the overbase base Ph O O Br Ph O O