Synthesis of a Spirocyclic Seco Structure of the Principal Vetiver Odorant Khusimone

Abstract

AbstractThe three‐dimensional structure of tricyclic compounds with a zero‐bridge to one bridgehead atom is determined by the underlying spirocyclic framework. (–)‐Khusimone (1), the principal odorant of vetiver oil (content up to 2 %), is such a tricyclic norsesquiterpene, and dissection of the 7,8‐bond between the methylene and the gem‐dimethyl unit of 1 results in a spirocycle extending over an almost identical molecular volume and shape. Given that the “vetiver rule” postulates that one α‐branched carbonyl osmophore in a certain spatial distance to a bulky moiety is responsible for the odor of vetiver, a 7,8‐seco structure of 1 could prove or disprove these structural requirements. Therefore, (4R*,5R*)‐7‐isopropyl‐4‐vinylspiro[4.4]nonan‐1‐one [(4R*,5R*)‐2] was synthesized in a 10‐step synthetic sequence commencing with Steglich esterification of allyl alcohol (15) and isovaleric acid (14). Ireland–Claisen rearrangement of the formed allyl isovalerate (16) with subsequent lithium aluminum hydride reduction of resulting γ,δ‐unsaturated acid 18, Appel bromination of corresponding alcohol 19, and ozonolysis provided 4‐bromo‐3‐isopropylbutanal (23) in 19 % overall yield as a building block for the projected spiroannulation reactions. Although attempts on cyclopentanone (7) failed, cyclopent‐2‐en‐1‐one, via its TMS‐trapped lithium 3‐vinylcyclopent‐1‐enolate 12, turned out to be a successful starting material. Evans' variant of the Mukaiyama aldol reaction with 1‐trimethylsiloxy‐3‐vinylcyclopent‐1‐ene (12) in the presence of BF3·OEt2, followed by palladium‐catalyzed conjugate tin hydride reduction of resulting enone 28 provided 2‐(4′‐bromo‐3′‐isopropylbutyl)‐3‐vinylcyclopentanone (29) in 50 % overall yield as the anlation precursor. LDA‐mediated 5‐exo‐tet cyclization of 29 concluded the synthesis of the racemic 7,8‐seco‐/6‐epi‐7,8‐secokhusimone mixture (4R*,5R*)‐2, which possessed a floral, rosy, green, geranium‐like odor with a threshold of 42.0 ng L–1 air, which is 10 times less intense than that of (–)‐khusimone (1). Most importantly, seco structure (4R*,5R*)‐2 did not display any woody nor any vetiver character, which proves the postulated vetiver rule wrong.