4 as the major product (2.2:1 ratio of R to ꢀ, 86% yield)
overriding the substrate’s diastereoselective preference.8
After significant experimentation with Brønsted and Lewis
acids it was determined that the rearrangement of the
R-epoxide 4 was optimally achieved using the in situ derived
aluminum phenoxide generated through the combination of
4-bromophenol and trimethyl aluminum (3:1) in dichlo-
romethane at 0 °C.9 This reagent, when combined with 4,
provides the carbocyclic core of caryolanemagnolol in a
single operation, yielding bromide 11 in 69% yield (Scheme
2). To verify the structural assignment the X-ray crystal
A second Suzuki reaction between bromide 14 and pinacol
ester 15 (prepared in three steps from 4-allylanisole, Sup-
porting Information) using conditions developed by Fu and
co-workers formed the required biaryl bond, generating 16
in 71% yield (Scheme 3).12 Simultaneous removal of the
Scheme 3. Completion of the Synthesis of Caryolanemagnolol
Scheme 2. Synthesis of Bromide 14
structure of the 3,5-dinitrobenzoyl ester of 11 was obtained.10
Due to the inability of 2,4-dibromophenol to effectively
participate in the rearrangement reaction, the bromination
of 11 was required for subsequent palladium bond-forming
reactions. Preliminary attempts at halogenation of the aryl
group occurred with competitive oxidation of the secondary
alcohol, forming the corresponding ketone of 11. As a result,
protection of the secondary alcohol of 11 was required. An
acetate was appended under standard conditions followed
by installation of a second bromide using bromine and
sodium acetate at elevated temperature (60 °C) to generate
the dibromide 12 in 77% yield over two steps (Scheme 2).
With the bromine atoms installed, sequential Suzuki reactions
were used to form the allyl appendage and the biaryl bond.11
In the majority of allylation reactions examined, Suzuki cross
coupling led to diallylation of 12. Using pinacol allylboronate
13, tetrakis(triphenylphosphine)palladium catalysis, and close
monitoring of the reaction, however, compound 14 could
be isolated in 72% yield.
acetate and carbamoyl groups was achieved with lithium
aluminum hydride providing caryolanemagnolol in 71%
yield. While the synthetic material matched the spectra for
the isolated caryolanemagnolol it should be noted that the
original isolation report incorrectly stated the chemical
shifts.13
The synthesis of clovanemagnolol was achieved over six
steps similar to those used for the synthesis of caryolane-
magnolol (Scheme 4). Unfortunately, the aluminum phe-
noxide reagent used in the caryolanemagnolol sequence failed
to convert epoxide 5 to the carbocyclic clovane core structure.
Therefore, starting from the recrystallized, commercial
caryophyllene ꢀ-oxide (5) the rearrangement was achieved
using diphenyl phosphate and 4-bromophenol providing the
clovane core in 35% yield. Of the Brønsted and Lewis acids
examined diphenyl phosphate proved optimal. Regrettably,
as a result of a competing elimination reaction providing
clovene, the rearrangement was less efficient than the reaction
forming the caryolane core.5,14 Conversion of bromide 17
to clovanemagnolol was achieved using the same set of
transformations used in the caryolanemagnolol synthesis. The
spectral properties of natural and synthetic clovanemagnolol
fully match.1,13
(8) (a) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J.-R.; Shi, S. J. Am.
Chem. Soc. 1997, 119, 11224. (b) Tian, H.; She, X.; Shu, L.; Yu, H.; Shi,
Y. J. Am. Chem. Soc. 2000, 122, 11551.
(9) (a) Naruse, Y.; Esaki, T.; Yamamoto, H. Tetrahedron Lett. 1988,
29, 1417. (b) Naruse, Y.; Esaki, T.; Yamamoto, H. Tetrahedron 1988, 44,
4747.
(10) X-ray crystal structure of the 3,5-dinitrobenzoyl ester of 11:
(11) (a) Miyaura, N.; Suzuki, A. Chem. ReV. 1995, 95, 2457. (b)
Sambasivarao, K.; Manoranjan, B.; Vrajesh, S. Synlett 2005, 1877.
(12) (a) Fu, G. C. Acc. Chem. Res. 2008, 41, 1555. (b) Littke, A. F.;
Dai, C.; Fu, G. C. J. Am. Chem. Soc. 2000, 122, 4020.
(13) Please see the Supporting Information for spectral comparisons of
isolated caryolanemagnolol and clovanemagnolol to synthetic material. In
addition, optical rotations of synthetic material were larger than those
reported for the isolated compounds.
1306
Org. Lett., Vol. 12, No. 6, 2010