Enantioselective nucleophilic opening of meso epoxides by organolithium reagents

Article abstract of DOI:10.1055/s-1998-5741

Aryl lithium reagents, complexed with (-)-sparteine, react enantioselectively with cyclic meso epoxides, to afford chiral aryl cyclanols. The enantiomeric excess, though moderate (27-87%), is the best in the literature for such a reaction. Activation by B

Full text of DOI:10.1055/s-1998-5741

October 1998
SYNLETT
1165
Enantioselective Nucleophilic Opening of meso Epoxides by Organolithium Reagents
Alexandre Alexakis*, Emmanuel Vrancken, Pierre Mangeney
Department of Organic Chemistry, Université P. et M. Curie, T44-45, case 183, 4 place Jussieu, Paris 75005, France
Fax (int.) 33 1 44 27 75 67; e-mail alexakis@moka.ccr.jussieu.fr
Received 9 July 1998
Abstract : Aryl lithium reagents, complexed with (-)-sparteine, react
enantioselectively with cyclic meso epoxides, to afford chiral aryl
cyclanols. The enantiomeric excess, though moderate (27-87%), is the
best in the literature for such a reaction. Activation by BF .OEt is
3
2
needed, and is compatible with a diamine such as (-)-sparteine.
The enantioselective desymmetrization of achiral meso epoxides is an
interesting way to create new stereogenic centers. The topic has recently
1
been covered by a comprehensive review, where it appears that carbon
nucleophiles, such as common organometallic reagents (RLi, RMgX,
R CuLi, R Al, RZnX ...), are the least efficient for undergoing this
2
3
asymmetric process. For example, the best result was attained by
2
Tomioka et al according to the following equation.
It is already known that in the absence of the Lewis acid, the
nucleophilic opening of such unreactive epoxides is a very slow
reaction. Activation by coordination of a Lewis acid, such as BF .OEt ,
3
2
3
to the oxygen atom of the epoxide is required.
4
Although we have been working in this field since a long time , we
report herein our results which bring additional informations to the
above reported ones. There are several efficient chiral ligands for
organolithium reagents, but none of them matches (-)-sparteine in term
5
of enantioselectivity. At first glance, it seems incongruous to mix a
strong Lewis base (sparteine is a diamine) with a strong Lewis acid
(BF .OEt ). Competition for the Lewis acid should be in favor of the
diamine moiety instead of the oxygen of the epoxide ! Despite this fact,
we reasoned that an equilibrium may allow the reaction to proceed, and
this is indeed the case.
Variation of the number of equivalents of (-)-sparteine with respect to
the organolithium reagent gave lower ee's. The optimum is a 1:1 ratio. If
the PhLi, LiBr complex is used, additional (-)-sparteine is needed, and
the ee is as high as without LiBr. However, the yield is lower due to the
formation of bromocyclohexanol, by competitive nucleophilic attack of
the Br anion. In fact, the most important finding concerns the nature of
the organic group of the organolithium reagent which has a profound
effect. Only an aromatic group gave a significant enantiomeric excess of
aryl cyclohexanol. The aggregation state of the organometallic species
should play the most crucial role.
3
2
When 1.5 equivalents of BF .OEt were added dropwise to an ethereal
3
2
solution of 2 equivalents of salt-free PhLi, complexed to 2 equivalents of
(-)-sparteine, and one equivalent of cyclohexene oxide, a rapid reaction
takes place at -78°C, to give in quantitative yield trans-phenyl-
cyclohexanol (this stoichiometry corresponds to that used by Tomioka et
2
31
6
al). The enantiomeric excess was measured by our P NMR method
and the absolute configuration was determined to be 1R,2S by
Since optically pure phenylcyclohexanol is a known useful chiral
auxiliary , we screened other aryllithium reagents which would afford
other arylcyclohexanols. These compounds may, in turn, serve as new
chiral auxiliaries. These results shown in Table 2 are obtained according
to the following equation:
7
polarimetry and comparison of the optical rotation with the known
2
value. Exactly the same result was achieved in toluene as solvent, a
result which is in contrast to Tomioka's report where an increase from
2
38 to 47% ee was observed. In the absence of BF .OEt no reaction
3
2
was observed at -78°C. Increasing the temperature to 0°C, for 3 days,
gave a 34% yield of trans-phenylcyclohexanol of 5% ee.
This promising result was further examined with various stoichiometries
of the reactants and with other organolithium compounds and our results
are compiled in Table 1.

1166
LETTERS
SYNLETT
Increasing the steric effects using anthracenyllithium (entry 11) resulted
in a lower ee.
Finally, we examined the behaviour of other epoxides with our best
reagent, 1-naphthyllithium. Although the ee's are lower, it is interesting
to note that a rather flat epoxide, such as cyclopentene oxide gave a
result as good as cycloheptene oxide or norbonene oxide.
In summary, although the enantiomeric excesses are moderate, they are
among the best reported for this type of reaction. We believe that the
enantioselective nucleophilic opening of meso epoxides, assisted by a
strong Lewis acid, may be improved by screening new ligands, even if
they are a priori considered as strong Lewis bases. Work is in progress
toward this end.
References and Notes
(1) Hodgson, D.M.; Gibbs, A.R.; Lee, G.P. Tetrahedron 1996, 52,
14361-14384. The review includes hetero-nucleophiles as well as
the particular case of TMS-CN, which is a C-nucleophile (see
Cole, B.M.; Shimizu, K.D.; Krueger, C.A.; Harrity, J.P.A.;
Snapper, M.L.; Hoveyda, A.H. Angew. Chem., Int. Ed. Engl. 1996,
35, 1668-1671.
(2) a) Mizuno, M.; Kanai, M.; Iida, A.; Tomioka, K. Tetrahedron:
Asymmetry. 1996, 7, 2483-2484. b) Mizuno, M.; Kanai, M.; Iida,
A.; Tomioka, K. Tetrahedron 1997, 53, 10699-10708.
(3) a) Yamaguchi, M.; Hirao, I. Tetrahedron Lett. 1983, 24, 391-394.
b) Eis, M.J.; Wrobel, J.E.; Ganem, B. J. Am. Chem. Soc. 1984,
106, 3693-3694. c) Ghribi, A.; Alexakis, A.; Normant, J.F.
Tetrahedron Lett. 1984, 25, 3075-3078. d) Lipshutz, B.H.; Parker,
D.A.; Kozlowski, J.A.; Ngyuen, S.L. Tetrahedron Lett. 1984, 25,
5959-5962.
(4) Koenig, I. Diplôme d'Etudes Approfondies, Université P. Et M.
Comparison of the three isomers of Tol-Li shows the strong influence of
an ortho substituent (entry 2) on increasing the enantioselectivity.
However, an ortho methoxy substituent has a detrimental effect due to
the internal coordination of the metal by the oxygen (entry 6). The
electronic nature of the para substituent plays a moderate role, since the
ee's are in the same range 43-53%. The same may be roughly said about
Curie, 1991, Paris.
(5) Hoppe, D.; Hense, T. Angew. Chem. Int. Ed. Engl. 1997, 36, 2282-
2316.
(6) a) Alexakis, A.; Mutti, S.; Mangeney, P.; Normant, J.F.
Tetrahedron: Asymmetry. 1990, 1, 437-440. b) Alexakis, A.;
Frutos, J.; Mutti, S.; Mangeney, P. J. Org. Chem. 1994, 59, 3326-
3334.
the meta substituent, although we may note the 60% ee of the CF
3
moiety.
(7) Whitesell, J.K. Chem. Rev. 1992, 92, 953-964
1-Naphthyllithium is an extremely interesting case where a high ee
value (85-87%) is obtained only in toluene as a solvent (entry 10). This
is the sole example in our study where such a strong solvent effect is
(8) Basavaiah, D.; Rao, P.D. Tetrahedron: Asymmetry. 1994, 5, 223-
234 and references cited therein.
observed (22% ee in Et O). It is interesting to note that this
organolithium reagent is reported by Tomioka to give only 15% ee. The
(9) Typical Procedure. To a cooled (-78°C) solution of (-)-sparteine
2
2
(0.97 mL, 4.2 mmol) in 10 mL of dry Et O, under an argon
2
higher ee value of 87% was obtained on a large scale experiment (5g of
final material), showing that the method has some preparative value.
Moreover, only one equivalent of organolithium reagent was used.
atmosphere, was added dropwise an ethereal solution of salt-free
PhLi (5.88 mL of a 0.68 N solution, 4 mmol). After stirring for 30
min, cyclohexene oxide (202 µL, 2 mmol) was injected via

October 1998
SYNLETT
1167
syringe. A solution of BF .Et O (0.38 mL, 3 mmol) in 5 mL Et O
was slowly added (over 5 min) in order to maintain the
temperature at -78°C. After stirring for 1 h, the reaction was
afford 343 mg of (1R,2S)-2-phenyl-1-cyclohexanol as white
3
2
2
1
needles. mp 57°C. [α] -24 (c 1.26, benzene). H NMR (400
D
MHz, CDCl ) δ 1.27-2.17 (m, 9H), 2.42 (m, 1H), 3.66 (m, 1H),
3
13
quenched with 10 mL of 5% aqueous H SO . After standard
7.20-7.38 (m, 5H). C NMR 25.11, 26.06, 33.41, 34.43, 53.18,
2
4
work-up, the crude product was purified by column
chromatography on silica gel (eluent pentane : Et O = 80 : 20) to
74.29, 126.83, 128.12, 128.77, 143.71.
2