A synthesis of the banana weevil pheromone component (1S,3R,5R,7S)-sordidin starting from (S)-2-(2-ethyl-1,3-dioxolanyl)propan-1-ol and (R)-4-hydroxypentan-2-one is described. No epimerization was observed for the final, intramolecular, acid-catalyzed cyclization step using aqueous oxalic acid. The banana, which is an important food crop in tropical and subtropical areas of the world, is threatened by disease and predators. 1 The banana weevil, Cosmopolites sordidus (Germar), is a serious pest insect for this plant. The weevil elicits a male-produced aggregation pheromone 2 of which the main component, (+)-sordidin, was proven to be (1S,3R,5R,7S)-1-ethyl-3,5,7-trimethyl-2,8-dioxabicyclo[3.2.1]octane (1a). 3-5
Figure 1Oehlschlager and co-workers 6,7 conducted field studies with synthetic compounds and showed the pheromone to consist of a mixture of the diastereomers 1a-d in approximately the ratio given in Figure 1. Additional syntheses of sordidin or its stereoisomers have been published. [8][9][10][11] Recently, two asymmetric syntheses of (+)-sordidin (1a) have been reported 12,13 The need for a more practical synthesis of the pheromone enantiomers seemed apparent to us at the time our project was initiated, and our goal was a synthesis of the major component 1a, without having to resort to preparative GC for isolation of the final product.A synthesis of sordidin (1a) requires control of three stereogenic centers in the open-chain precursor 2 (Scheme 1). Furthermore, the subsequent cyclization should take place without epimerization at C-7 (see Figure 1 for numbering), which has been a problem. 5 Our approach is illustrated retrosynthetically in Scheme 1, depicting the preparation of building blocks (R)-3 and (R)-4 as our primary objective. We intended to metalate the bromide (R)-3 and react it with the ketone (R)-4, thereby creating the third stereogenic center of 2.
Scheme 1With the purpose of finding satisfactory conditions for the metalation of (R)-3 some metalation reactions were carried out on the racemic compound. First we attempted a bromine-lithium exchange reaction with t-BuLi in THF at -78°C, and the subsequent reaction with 4-methylpentan-2-one (5) as a model substance gave the alcohol 6 in 42% yield ( Figure 2); replacing THF with other solvents caused no improvement in the yield, and neither did the use of Barbier conditions. The lithium derivative was also formed from 3 and lithium metal in the presence of naphthalene as catalyst, and reaction with 5 furnished a mixture of 6 (41%) and the dimer 7 (30%). We then turned to the Grignard reagent derived from 3, prepared in THF as solvent and 1,2-dibromoethane as entrainer, 14 but reaction with 5 gave a mixture of products containing the alcohol 6 (10%) and the dimer 7 (20%). Apparently, a more reactive magnesium was needed and we resorted to Rieke magnesium, prepared by reduction of magnesium dichloride immediately before use. 15 In this case the Grignard reaction afforded 6 in 52% yield together with a small amount of the dimer (5%). Other products ...