Enantiomerically pure cyclopropyl hemiacetals: Lipase-catalyzed synthesis of chiral, nonracemic ester homoenolate equivalents

Article abstract of DOI:10.1021/ol006752f

equation presented Enantiomerically pure cyclopropyl hemiacetals can be obtained by lipase-catalyzed kinetic resolution of their acylated congeners. It is demonstrated that lipases from Candida antarctica and Pseudomonas cepacia show enantiodivergent beha

Full text of DOI:10.1021/ol006752f

ORGANIC
LETTERS
2001
Vol. 3, No. 2
189-191
Enantiomerically Pure Cyclopropyl
Hemiacetals: Lipase-Catalyzed
Synthesis of Chiral, Nonracemic Ester
Homoenolate Equivalents
Bernhard Westermann* and Birte Krebs
UniVersita¨t-Paderborn, Fachbereich fu¨r Chemie und Chemietechnik,
Warburgerstrasse 100, 33098 Paderborn, Germany
bw@fb13n.uni-paderborn.de
Received October 19, 2000
ABSTRACT
Enantiomerically pure cyclopropyl hemiacetals can be obtained by lipase-catalyzed kinetic resolution of their acylated congeners. It is
demonstrated that lipases from Candida antarctica and Pseudomonas cepacia show enantiodivergent behavior toward these substrates.
Subsequent ring opening of these building blocks can be achieved with ZnCl2 leading to chiral, nonracemic r-substituted homoenolate anions.
The exploration of aldol reactions has been the cornerstone
of asymmetric synthesis. In particular, the use of chiral
enolates has been surveyed thoroughly.1 In contrast, chiral
nonracemic homoenolate anions, which are versatile d3
building blocks, have found less abundant applications in
asymmetric synthesis.2 Among the approaches used to obtain
these umpoled reagents, two main strategies have evolved
in recent years: (i) metalation of a heterovinyl group
(offensive strategy) and (ii) metalation of protected and
suitably substituted carbonyl compounds (defensive strategy),
respectively.3,4
availability of appropriate precursors. In studies by Kuwajima
and Nakamura, it has been demonstrated that cyclopropanols
or cyclopropanone hemiacetals can be ring-opened regiose-
lectively to afford keto and ester homoenolates, respectively.5
No racemization was observed when the reaction was carried
out from enantiomerically pure starting compounds 1a as
exemplified in Scheme 1.6 In our recent studies to synthesize
non-natural amino acids of type 4, we became interested in
these building blocks in enantiomerically pure form (1b-
e).7 For this purpose, we envisaged a lipase-catalyzed kinetic
resolution of acylated cyclopropanone hemiacetals (()-5b-
e. Lipases are well-known for their high degree of stereo-
differentiation of secondary alcohols (Kazlauskas rule), but
only a few cases have been reported where tertiary alcohols
have been successfully resolved.8-10 In these cases, one of
A third approach called the direct strategy, however, has
been utilized only to a minor extent, mainly due to the limited
(1) Heathcock, C. H.; Moon Kim, B.; Williams, S. F.; Masamune, S.;
Rathke, M. W.; Weipert, P.; Paterson, I. In ComprehensiVe Organic
Synthesis; Trost, B. M., Fleming, I., Heathcock, C. H., Eds.; Pergamon
Press: Oxford, 1991; Vol. 2, Chapter 1.5-1.9.
(2) Hoppe, D. Angew. Chem. 1984, 96, 930; Angew. Chem., Int. Ed.
Engl. 1984, 23, 932. Hoppe, D.; Hense, T. Angew. Chem. 1997, 109, 2377;
Angew. Chem., Int. Ed. Engl. 1997, 36, 2282.
(5) Kuwajima, I.; Nakamura, E. In ComprehensiVe Organic Synthesis;
Trost, B. M., Fleming, I., Heathcock, C. H., Eds.; Pergamon Press: Oxford,
1991; Vol. 2, Chapter 1.14. Salau¨n, J. Chem. ReV. 1989, 89, 1247.
(6) Nakamura, E.; Kuwajima, I. Tetrahedron Lett. 1987, 28, 337.
Nakamura, E.; Shimada, J. I.; Kuwajima, I. Organometallics 1985, 4, 641.
(7) Westermann, B.; Kortmann, I. Synlett 1996, 665.
(3) Crimmins, M. T.; Nantermet, P. G. Org. Prep. Proc. Int. 1993, 25,
41.
(4) Ahlbrecht, H. Synthesis 1999, 365.
(8) Hagan, D. O.; Zaidi, N. A. Tetrahedron: Asymmetry 1994, 5, 1111.
10.1021/ol006752f CCC: $20.00 © 2001 American Chemical Society
Published on Web 12/22/2000

a diastereomeric mixture of cis/trans isomers (5:10 = 1:1)
in high yields. They can be separated easily by chromatog-
raphy; in general the difference of Rf values is greater than
0.4. We have chosen this approach despite its low cis/trans-
selectivity, but at this stage of our studies we wanted to have
easy access to both diastereomers.
Scheme 1
At the same time we developed a suitable method to
monitor the kinetic resolution by GLC employing cyclodex-
trin-modified capillary columns.15 All substrates could be
baseline separated.
Upon screening of various lipases and esterases, Candida
antarctica lipase B (CALB) and Pseudomonas cepacia lipase
(PCL) proved to be suitable for enzymic resolution of trans
derivatives (()-5b-e.16,17 Unfortunately, the saponified
cyclopropanone hemiacetals 1b-e turned out to be of limited
stability during the lipase-catalyzed reaction; as a conse-
quence, the ring-opened esters 6b-e could be isolated.
However, the desired isolation of both enantiomers of 5b-e
could still be accomplished, because the lipases CALB and
PCL showed enantiodivergent behavior toward (()-5b-e
(Figure 1). In the case of the CALB-catalyzed resolutions,
the substituents was a minor steric hindrance, namely nitrile
or alkine residues. Therefore, we thought it might be possible
to subject cyclopropanone hemiacetals (()-5 to lipase-
catalyzed kinetic resolution with the methylen group exhibit-
ing insignificant steric hindrance, too.11 Another striking
feature in these substrates is that the hydroxy group to be
attacted by the enzyme is part of a hemiacetal. In pioneering
reports, Feringa and Kellogg demonstrated that pyranes
which form stable hemiacetals (from aldehydes) can be
transformed by lipases.12 However, the use of ketone
hemiacetals has been precedentless as we encountered the
stereoselective transformation of (()-5 by lipases. Here we
want to present our results concerning the lipase-catalyzed
resolutions of (()-5b-e and their transformation into chiral-
nonracemic homoenolates 2 as outlined in Scheme 1.
The cyclopropanone hemiacetals 5b-e and their cis-
configured congeners 10b-e are readily prepared via cy-
clopropanation of silyl ketene acetals 9b-e and di-
iodomethane as outlined in Scheme 2.13,14 Methanolysis of
Figure 1. Productive vs nonproductive binding of acylated
cyclopropanone hemiacetals: (1b) enantiopreference of P. cepacia
lipase, (1c) enantiopreference of C. antarctica lipase B toward
substrates (()-5b-e.
the reaction proceeded faster, leading to products in high
yields and high enantiomeric excesses (Table 1). The optical
purity was determined by GLC as stated above and was
Scheme 2
(10) Lalonde, J. Chemtech 1997, (2), 38. We would like to thank Altus
Biologics for providing a sample of C. rugosa esterase.
(11) Barnier, J. P.; Blanco, L.; Rousseau, G.; Guibe-Jampel, E. J. Org.
Chem. 1993, 53, 1570.
(12) Deen, H. v. d.; Cuiper, A. D.; Hof, R. P.; Oeveren, A. v.; Feringa,
B. L.; Kellogg, R. M. J. Am. Chem. Soc. 1996, 118, 3801. Heuvel, M. v.
d.; Cuiper, A. D.; Deen, H. v. d.; Kellogg, R. M.; Feringa, B. L. Tetrahedron
Lett. 1997, 38, 1655.
(13) Matsuzawa, S. W.; Horiguchi, Y.; Nakamura, E.; Kuwajima, I.
Tetrahedron 1989, 45, 349.
(14) Booker-Milburn, K. I.; Thompson, D. F. J. Chem. Soc., Perkin Trans.
1 1995, 2315.
(15) We used a 6-TBDMS-3-butyrate-2-methyl-â-cyclodextrin-modified
GLC column.
(16) CALB has been provided by Novo Nordisk; PCL has been purchased
from Sigma. The bacterial species PC has been renamed recently (Burkhold-
eria cepacia), but we contiunue to use the traditional name.
(17) We screened 11 different lipases and 3 different estereases (from
Altus, Amano, Novo, and Fluka).
the silyl acetal and acetylation with acetic anhydride in
pyridine gave the cyclopropanone hemiacetals 5 and 10 as
(9) Watanabe, N.; Sugai, T.; Ohta, H. Chem. Lett. 1992, 657.
190
Org. Lett., Vol. 3, No. 2, 2001

Table 1. Results of the Enantiodivergent Kinetic Resolution of Cyclopropanone Hemiacetals 5b-e
C. antarctica (+)-5a P. cepacia (-)-5a
c
c
R1
[R]D
% eeb
yield [%]/time [d]
[R]D
eeb
yield [%]/time [d]
butyl
+29.5
+25.9
+22.7
+55.2
>99
>99
>99
>99
84/11
64/7
86/4
98/3
-28.6
-23.2
-21.9
-52.0
98
95
97
96
70/18
34/16
75/36
52/7
pentyl
hexyl
phenyl
a Absolute configuration determined after reaction to 6. b Determined via GLC (6-TBDMS-3-butyrate-2-methyl-â-cyclodextrin). c In CHCl3.
proven to be >99% ee. In the case of PCL-catalyzed
resolutions, the reaction time was prolonged; however almost
no decrease in isolated yields and optical purities, respec-
tively, was observed.
The configuration of the products has been determined
via comparison of known compounds 6; therefore, for (+)-
5b-e a (1S,2S)-configuration and for (-)-5b-e a (1R,2R)-
configuration was established.18
Cis-configured cyclopropanone hemiacetals 10b-e, how-
ever, could not saponified using lipase CALB and PCL,
respectively. Other screened lipases failed, too.15 These
substrates behave in a manner similar to tertiary alcohols,
which have a remote reactivity toward enzyme-catalyzed
saponifications.
To predetermine the stereoselectivity of lipase-catalyzed
reactions of secondary alcohols, the Kazlauskas model has
been endorsed as the most valuable.19 In Figure 1a, the
substrate alignment of a secondary alcohol in the active site
of lipase CALB is shown. To superimpose the experimentally
determined configuration of tertiary alcohols (+) and (-)-
5, respectively, and the active site model, the arrangements
presented in Figures 1b (for CALB) and 1c (for PCL) may
account for the observed enantiodivergent stereoselectivity.
In the context of generating chiral, nonracemic homoeno-
lates 2b-e, 5b-e needed to be reduced first. We chose this
approach due to the instability of the cyclopropanone
hemiacetals (Scheme 1). By reduction with LAH at 0 °C
and a quick workup, 1b-e could be obtained in almost
quantitative yields. Treatment with TMS-triflate gave the
silylated cyclopropanone hemiacetal derivatives without any
racemization. Therefore, we proved that the configuration
of 1b-e is stable.
By stirring silylated hemiacetals in the presence of ZnCl2
and subsequent hydrolysis of intermediary formed homoeno-
lates 2b-e, the ring-opened esters 6b-e could be isolated
enantiomerically pure. These experiments demonstrate that
the homoenolate is formed regioselectively without racem-
ization.
The scope and limitations of this lipase-catalyzed reaction
will be disclosed in due course. In ongoing studies we are
currently exploring conditions to add these nucleophiles to
imines, a reaction which could be termed a “homo-Mannich
reaction“.
Acknowledgment. We are grateful to Prof. W. Ko¨nig,
University of Hamburg, for providing the cyclodextrin-
modified GLC column. Gifts from Amano and Novo Nordisk
and financial support by the DFG (We 1826/2-1) and the
Fonds der chemischen Industrie are highly appreciated.
Supporting Information Available: Characterization
data for compounds 5. This material is available free of
charge via the Internet at http://pubs.acs.org.
(18) Bull, S. D.; Davies, S. G.; Jones, S.; Sanganee, H. J. J. Chem. Soc.,
Perkin Trans. 1 1999, 387.
(19) Kazlauskas, R. J.; Weisfloch, A. N. E.; Rappoport, A. T.; Cuccia,
L. A. J. Org. Chem. 1991, 56, 2656.
OL006752F
Org. Lett., Vol. 3, No. 2, 2001
191