For my inaugural blog post, I thought
I'd take a look at an oldie, but a goodie–a (semi?) total synthesis from the
groovy 1960s, that of racemic thujopsene. This one comes to you courtesy of the
now-deceased Prof. William G. Dauben (a fellow original Ohioan like myself) and
crew from the University of California at Berkeley. I like to look back at
classic synthesis papers not because I'm old (I'm 32, I'm not that old yet!), but because they often
contain instructional chemistry that's no longer utilized but still very
useful. Thujopsene is a sesquiterpene that really has little bioactivity to
speak of. Being one of the essential oils of some conifer trees, it does have a
nice scent, though.
I put the semi part in there because
it's not clear what the origin of their octalin starting material 1 is. Either way, their first reported
step provides a nice contrast between reaction kinetics and thermodynamics.
Enolate formation* from 1 with tert-butoxide and addition of
iodomethane to 2 provides the less
thermodynamically stable alkene in dimethylated 3 (the alkene conjugated with the ketone takes that title).
Dimethylation at the a-position
leads to the kinetic product. A little redox manipulation followed; Wolff-Kishner
reduction (a reaction analagous to the Shapiro reaction) of ketone 3 followed by allylic oxidation to the
ketone 5 with Cr(VI). The mechanism
of allylic oxidation with chromium is said to proceed via allylic hydride
abstraction, followed by trapping of the allylic cation by the reduced chromium
reagent.
The powerful reducing agent lithium
aluminum hydride was utilized to accomplish a diastereoselective 1,2-reduction
of the enone 5 from the convex face to
give allylic alcohol 6 in a 95:5 d.r.
The authors also attempted Meerwein-Ponndorf reduction of the enone using Al(Oi-Pr)3.
However, nearly a 1:1 ratio of stereoisomers was obtained in this case (58:42).
Honestly, given the choice of reducing agent, it's surprising the molecule
wasn't completely reduced, alkene and ketone. Nowadays, the Luche reduction is
commonly employed when one wants to selectively 1,2-reduce an enone, and this
method makes use of the much less reactive sodium borohydride.
The next step was essentially the crux
of the paper, as they mentioned in the opening paragraphs their recent studies
on cyclopropanation of allylic alcohols (JACS,
1963, 86, 468). Specifically, they found that cis-selectivity was observed, a fact that is crucial to the
stereochemistry obtained in the conversion of 6 to cyclopropyl-containing 7.
Thus, usage of the Simmons-Smith organozinc reagent gave 7 as a single isomer, but in relatively low yield. A look at the
experimental section shows that quite a bit of work was necessary to get the
fully purified material, including a spinning band column distillation, recrystallization
and flash chromatography over alumina.
Jones oxidation of the hydroxyl group
in 7 gave 8 in expectedly high yield, and addition of MeMgBr followed by a
workup with ammonium chloride resulted in facile dehydration to alkene 9, with intermediate alcohol 8 not being isolated. The authors
mention that a characteristic reaction of thujopsene (9) is its rearrangement to the ring-expanded widdrol (10), and that was accomplished by
treatment with oxalic acid in wet ethanol as a final proof of synthesis.
Overall it’s a nice synthesis.
Typically when an organic chemist sees a decalin ring structure with a distal
alkene, they immediately think of employing a Diels-Alder reaction. However, the
presence of the cyclopropyl group necessitates a slightly different synthetic
plan. The cyclopropanation reaction wasn’t quite as efficient as in their less complicated
model systems, but still delivered the product stereospecifically. Perhaps some
of the newer cyclopropanation reagents (rhodium carbenoids, newer organozincs)
could give better yields.
*For a half-page review of
stereoselectivity in the reductive alkylation of bicyclic enones, see this half-page review I made
back in my grad school days.
