Angewandte
Chemie
with lithium TMS acetylide to give the propargylic alcohol 16.
Finally the remaining epoxides were installed in one step
using the Shi asymmetric epoxidation,[26] thus providing the
polyepoxide 18.
The formation of the hexacarbonyl dicobalt complex from
18 proceeded smoothly to give the intermediate 12, which was
submitted to the same reaction conditions used for the shorter
homologue 7 (Scheme 7). As expected, the cyclization
Scheme 5. Retrosynthetic analysis of teurilene (1) by an epoxide-open-
ing cascade triggered by a Nicholas reaction.
Having established the expediency of our approach for
the cascade cyclization, we decided to move forward and
synthesize 1 by using our developed methodology. As
depicted in the retrosynthetic analysis (Scheme 5), we envis-
aged that 1 might be secured through a sequence of reactions
starting from the cobalt hexacarbonyl complex 12. It must be
emphasized that as a result of the Cs symmetry of 1, the
election of the chiral auxiliary in the asymmetric epoxidation
reactions was to establish the suitable relative configuration
in the polyepoxide 12. For the synthesis of this key inter-
mediate we have used a slightly different route (Scheme 6).
By employing the bidirectional strategy, the known diol 13
was synthesized in 55% overall yield after four steps.[23] The
diol 13 was monoprotected as a silyl ether, thus leaving the
bis(homoallylic) alcohol ready for homologation. By using
the protocol developed by our group we obtained the a,b-
unsaturated ester 14 in one step and excellent yield.[24]
Reduction of the ester group was performed with aluminium
hydride to provide the allylic alcohol. The first chiral epoxide
was generated through the Katsuki–Sharpless asymmetric
epoxidation[25] with subsequent protection with Boc to afford
the compound 15. The silyl group was removed under
standard reaction conditions and the resulting alcohol was
oxidized to the corresponding aldehyde, which was coupled
Scheme 7. Reagents and conditions: a) [Co2(CO)8], CH2Cl2, RT, quant.;
b) 1. SiO2, CH2Cl2, RT, 76 h;[21] 2. CAN, acetone, 08C, 75% after 2
steps; c) 1. K2CO3, MeOH, RT, 24 h; 2. HgSO4, H2SO4 3m, RT, 2 h,
87% after 2 steps; d) K2CO3, MeOH, RT, 24 h, 75% (2 iterations);
e) 1. CH3PPh3+BrÀ, nBuLi, THF, 08C to RT; 2. AD-mix b, CH3SO2NH2,
tBuOH/H2O (1:1), 08C, 75% after 2 steps; f) 1. MsCl, Py, CH2Cl2, 08C;
2. K2CO3, MeOH, 08C to RT, 90% yields after 2 steps; g) Allylmagne-
sium chloride, CuI, THF, 08C to RT, 93% yields. Ms=methanesul-
fonyl.
process provided the three THF rings in excellent yield
(75% after two steps). The cascade is high yielding, averaging
91% yield per epoxide. Recently an interesting methodology
based on an SN2 strategy was disclosed, thus leading to three
THF rings in one step in a slightly lower overall yield.[27]
Intriguingly, compared to 7, the complex 12 proved to be less
reactive to the silica gel, thus needing a longer reaction time
to consume the starting material. Although the cascade
proceeded efficiently and followed the same stereochemical
course as 7, the reaction gave a 1:1 epimeric mixture of 19 and
20 at the propargylic position.[28] Interestingly, neither
extending the reaction time nor changing the acid strength
improved the isomeric ratio or the overall yield. From
a synthetic standpoint, the epimeric mixture does not
represent a problem for the synthesis of 1 because the
isomerization of the propargylic ether can be done at a later
stage of the synthetic route (see below).
Scheme 6. Reagents and conditions: a) 1. TBSCl, imidazole, CH2Cl2,
=
RT, 96% (3 iterations); 2. SO3·Py, Et3N, DMSO, CH2Cl2, Ph3P C(CH3)-
CO2Et, RT, 88%; b) 1. AlCl3, LiAlH4, Et2O, 08C, 94%; 2. Ti(OiPr)4,
(+)-DET, tBuOOH, 4 ꢃ M.S., CH2Cl2, À208C; 3. DMAP, Boc2O,
toluene, RT, 98% after 2 steps; c) 1. TBAF, THF, RT, 86%; 2. SO3·Py,
Et3N, DMSO, CH2Cl2, RT; 3. TMS acetylene, nBuLi, THF/toluene (1:2),
À788C, 70% after 2 steps; d) oxone, K2CO3, DMM/CH3CN/buffer
pH 10.5 (2:1:1), 17, 08C, 81%. TBS=tert-butyldimethylsilyl,
(+)-DET=(+)-diethyl l-tartrate, TBAF=tetrabutylammoniumfluoride,
DMM=dimethoxymethane.
The isomers 19 and 20 were treated under basic conditions
to remove the trimethylsilyl group on the alkyne and the Boc
group present at the primary alcohol.[29,30] The next step was
the hydration of the terminal alkyne to afford the methyl
ketones 21 and 22, respectively, in excellent yield after two
steps. At this stage of the synthesis, we decided to execute the
isomerization of 22 into the correct isomer 21. For this
Angew. Chem. Int. Ed. 2013, 52, 3659 –3662
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3661