Separation of enantiomers by extraction based on lipase-catalyzed enantiomer-selective fluorous-phase labeling

Article abstract of DOI:10.1002/1521-3773(20010702)40:13<2492::AID-ANIE2492>3.0.CO;2-8

No chromatography is necessary to separate a racemic alcohol into its enantiomers. A highly fluorinated acyl residue was tranferred in an enantiomer-selective manner onto a racemic alcohol in the presence of a lipase [Eq. (1)]. The labeled enantiomer was separated from the unlabeled one by a simple but very efficient partition between fluorous and organic phases.

Full text of DOI:10.1002/1521-3773(20010702)40:13<2492::AID-ANIE2492>3.0.CO;2-8

COMMUNICATIONS
Crystal structure analysis data for 1: C40H50N2O2ClFe, Mr ˆ 682.12, triclinic,
Separation of Enantiomers by Extraction Based
on Lipase-Catalyzed Enantiomer-Selective
Fluorous-Phase Labeling**
Å
P1, a ˆ 10.738(1), b ˆ 13.195(1), c ˆ 14.061(1) Š, a ˆ 75.19(1), b ˆ 79.71(1),
3
g ˆ 77.80(1)8, Vˆ 1866.1(3) Š3, Z ˆ 2, 1calcd ˆ 1.214 Mgm
;
m(MoKa) ˆ
0.511 mm 1, F(000) ˆ 726; 18080 reflections collected at 100(2) K; 9891
independent reflections; GOF ˆ 0.990; R ˆ 0.0508; wR2 ˆ 0.1059. Crystal
Benno Hungerhoff, Helmut Sonnenschein, and
Fritz Theil*
Å
structure analysis data for 2: C40H50N2O2BrFe, Mr ˆ 726.58, triclinic, P1,
a ˆ 10.4544(9), b ˆ 14.737(1), c ˆ 24.387(2) Š, a ˆ 86.88(2), b ˆ 84.89(2),
3
g ˆ 87.58(2)8, Vˆ 3734.0(5) Š3, Z ˆ 4, 1calcd ˆ 1.292 Mgm
;
m(MoKa) ˆ
1.509 mm 1, F(000) ˆ 1524; 29363 reflections collected at 100(2) K; 12766
independent reflections; GOF ˆ 0.949; R ˆ 0.0552; wR2 ˆ 0.1030. Crystal
structure data analysis for 3: C40H50N2O2IFe, Mr ˆ 773.57, monoclinic, P21/c,
a ˆ 13.497(1), b ˆ 26.326(2), c ˆ 11.672(1) Š, b ˆ 112.99(2)8, Vˆ
Lipase-mediated kinetic resolution of racemic alcohols and
their esters by esterification or hydrolysis, respectively, is a
well-established method for the preparation of enantiomeri-
cally pure or enriched building blocks.[1] Lipases are cheap
biocatalysts; the reactions can be run with standard equip-
ment and are highly selective in many cases. However, there is
one major drawback of this type of biotransformation, which
affords one of the enantiomers as an alcohol and the other one
as the corresponding carboxylate: The products must be
separated by chromatography. This separation step may not
be a serious problem on the laboratory scale. However, on a
large scale in the pharmaceutical industry, a chromatographic
step might be the reason this method is not considered to be a
useful access to enantiomerically pure intermediates. Until
now, there has been no general solution to overcome this
disadvantage.
On the other hand, remarkable progress has been made for
the extractive separation of homogeneous catalysts,[2] re-
agents, and products[3] equipped with perfluorinated auxiliary
groups. This methodology is based on partitioning between
the organic and fluorous phases in order to improve the
recovery of the homogeneous catalyst and the isolation of
products from the reaction mixture.
From the progress made in performing reactions in fluorous
media and/or improving workup procedures by the introduc-
tion of a fluorous phase,[4] the following question arises: Is it
possible to apply a highly fluorinated acyl donor to a lipase-
catalyzed kinetic resolution of a racemic alcohol? Such an
acyl donor should promote lipase-mediated enantiomer-
selective acyl transfer onto the faster-reacting enantiomer
and, thereby, simultaneously label it. This labeled enantiomer
with a ªteflon ponytailº[2b] could then be recognized selec-
tively by a fluorous phase, to allow the extractive separation
of the fluorinated and nonfluorinated enantiomers between a
fluorous and an organic solvent.
For a successful realization of this principle a suitable
acyl donor is required. This reagent should be accepted
by the lipase forming the reactive acyl enzyme, that sub-
sequently reacts in an enantiomer-selective manner with a
racemic alcohol. Furthermore, the transferred acyl residue
should have a sufficient fluorine content to allow selec-
tive separation of the fluorinated ester from the nonfluori-
3
1
3817.9(6) Š3,
Z ˆ 4,
1calcd ˆ 1.346 Mgm
;
m(MoKa) ˆ 1.238 mm
,
F(000) ˆ 1596; 38047 reflections collected at 100(2) K; 12090 independent
reflections; GOF ˆ 1.007; R ˆ 0.039; wR2 ˆ 0.072. Crystallographic data
(excluding structure factors) for the structures reported in this paper have
been deposited with the Cambridge Crystallographic Data Centre as
supplementary publication no. CCDC-159894 (1), CCDC-159895 (2), and
CCDC-159896 (3). Copies of the data can be obtained free of charge on
application to CCDC, 12 Union Road, Cambridge CB21EZ, UK (fax:
(‡44)1223-336-033; e-mail: deposit@ccdc.cam.ac.uk).
Received: March 13, 2001 [Z16765]
[1] a) S. Koch, R. H. Holm, R. B. Frankel, J. Am. Chem. Soc. 1975, 97, 6714;
b) D. P. Riley, D. H. Busch, Inorg. Chem. 1984, 23, 3235; c) K. L.
Kostka, B. G. Fox, M. P. Hendrich, T. J. Collins, C. E. F. Rickard, L. J.
Wright, E. Münck, J. Am. Chem. Soc. 1993, 115, 6746; d) H. Keutel, I.
Käpplinger, E.-G. Jäger, M. Grodzicki, V. Schünemann, A. X. Traut-
wein, Inorg. Chem. 1999, 38, 2320; e) E.-G. Jäger, H. Keutel, Inorg.
Chem. 1997, 36, 3512; f) D. Nicarchos, A. Kostikas, A. Simopoulos, D.
Coucouvanis, D. Piltingsrud, R. E. Coffman, J. Chem. Phys. 1978, 69,
4411.
[2] a) M. D. Snodin, L. Ould-Moussa, U. Wallmann, S. Lecomte, V.
Bachler, E. Bill, H. Hummel, T. Weyhermüller, P. Hildebrandt, K.
Wieghardt, Chem. Eur. J. 1999, 5, 2554; b) R. M. Buchanan, S. L.
Kessel, H. H. Downs, C. G. Pierpont, D. N. Hendrickson, J. Am. Chem.
Soc. 1978, 100, 7894; c) W. O. Koch, V. Schünemann, M. Gerdan, A. X.
Trautwein, H.-J. Krüger, Chem. Eur. J. 1998, 4, 1255.
[3] a) P. Chaudhuri, C. N. Verani, E. Bill, E. Bothe, T. Weyhermüller, K.
Wieghardt, J. Am. Chem. Soc. 2001, 123, 2213; b) H. Chun, C. N.
Verani, P. Chaudhuri, E. Bothe, E. Bill, T. Weyhermüller, K. Wieghardt,
Inorg. Chem., in press.
[4] C. N. Verani, S. Gallert, E. Bill, T. Weyhermüller, K. Wieghardt, P.
Chaudhuri, Chem. Commun. 1999, 1747.
[5] We have also recorded the Mössbauer spectrum of 3 at 298 K; a single
quadrupole doublet is observed (d ˆ 0.22 mms 1, jDEQ jˆ 2.21 mms 1).
This clearly establishes that the increasing magnetic moment at
temperatures >100 K for 3 (Figure 2) is not due to a spin crossover
St ˆ 1ꢁ2 !St ˆ 3ꢁ2.
[6] D. Sellmann, S. Emig, F. W. Heinemann, Angew. Chem. 1997, 109, 1808;
Angew. Chem. Int. Ed. Engl. 1997, 36, 1734.
[7] C. K. Jörgensen, Struct. Bonding (Berlin) 1966, 1, 234.
[*] Dr. F. Theil, Dr. B. Hungerhoff
ASCA GmbH
Richard-Willstätter-Strasse 12, 12489 Berlin (Germany)
Fax : (‡49)30-6392-4103
E-mail: theil@asca-berlin.de
Dr. H. Sonnenschein
Institut für Nichtklassische Chemie an der Universität Leipzig
Permoserstrasse 15, 04303 Leipzig (Germany)
[**] This work was supported by the Deutsche Forschungsgemeinschaft
(grant: TH 562/3-1) and the Fonds der Chemischen Industrie.
2492
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4013-2492 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 13

COMMUNICATIONS
nated alcohol in an appropriate biphasic organic/fluorous
system.
vinyl acetate in tert-butyl methyl ether in the presence of
Pseudomonas sp. lipase reached 50% conversion after
44 hours to yield the corresponding (R)-acetate and (S)-3
with ee values of >99 and 93%, respectively.[8] This demon-
strates that the reactivity of perfluoroester 2 is, at least, in the
same range as that of vinyl acetate.
In the next step, we looked for a suitable biphasic fluorous/
organic solvent system to separate ester (R)-4 from alcohol
(S)-3 whilst avoiding chromatography. Screening of organic
solvents immiscible with n-perfluorohexane such as cyclo-
hexane, toluene, THF, and methanol showed that the biphasic
mixture methanol/n-perfluorohexane was the system of
choice and allowed a very efficient separation of the
compounds rac-3 and rac-4, which were used as model
substances. When an equimolar mixture of rac-3 and rac-4
was extracted at least five times with n-perfluorohexane, rac-3
remained in the organic phase and rac-4 moved into the
fluorous phase. The organic phase was contaminated with less
than 1% of rac-4 and the fluorous phase contained less than
1% of rac-3.
After identification of the appropriate fluorous/organic
biphasic system for the separation of (R)-4 and (S)-3, the
reaction mixture was worked up as follows (Scheme 2): The
enzyme was filtered off, the solvent was removed under
reduced pressure, and the products were separated by
partition between methanol and n-perfluorohexane. This
method yielded (S)-3 with an ee value of 99% in the organic
phase and (R)-4 with an ee value of 98% in the fluorous phase,
which also contained the excess of the acylating agent 2.
Comparison of the results obtained by conventional
chromatographic separation and extractive separation shows
that there is almost no difference between both procedures
regarding the purity and the yield of the products. After the
extractive workup (S)-3 was contaminated with a trace (not
more than 1%) of (R)-4.[9] The ee value of 98% determined
after saponification of (R)-4 to (R)-3, represents an impurity
of not more than 1% of (S)-3 in the fluorous phase. In
addition, saponification of the mixture of (R)-4 and 2 to yield
(R)-3 allows the almost quantitative recovery of the fluori-
nated carboxylic acid in solid form as the lithium salt.
This newly developed methodology for the separation of
enantiomers, which combines a lipase-catalyzed kinetic res-
olution with fluorous-phase labeling and separation of the
products in a biphasic fluorous/organic solvent system, should
also be applicable for the enantiomer-selective hydrolysis or
alcoholysis of fluorous-phase labeled esters of racemic
alcohols and for the enantiomer-selective alcoholysis of esters
of racemic carboxylic acids with highly fluorinated alcohols.[10]
We are currently investigating further applications of this
newly developed strategy.
Esters of the structure CF3(CF2)n(CH2)mCOOR (where
R ˆ vinyl or 2,2,2-trifluoroethyl, for example) were designed
as suitable irreversible or quasi-irreversible acyl donors.[1c]
A
spacer of one or two methylene groups (m ˆ 1 or 2) should be
necessary to exclude a nonselective chemical acylation of the
alcohol. The commercially available alcohol 1[5] was selected
as a useful starting material for the synthesis of the ester 2,[6]
envisaged as a potential acyl donor and synthesized in high
yield by the method shown in Scheme 1.
Scheme 1. Synthesis of the highly fluorinated acyl donor 2. a) 4-
CH3C6H4SO2Cl, b) LiBr, c) Mg, d) CO2, e) PCl5, f) CF3CH2OH, pyridine.
In order to demonstrate the feasibility of the principle, we
chose 1-phenylethanol (rac-3) as the alcohol to be resolved.
After screening of lipases and solvents, lipase B from Candida
antarctica (CAL-B) in acetonitrile turned out to be a useful
biocatalytic system with ester 2 as the acylating agent; the
resolution of the enantiomers of rac-3 into (R)-4 and (S)-3
proceeded with high efficacy within 19 hours (Scheme 2). The
corresponding ester with one less CH2 group (m ˆ 1) was not
suitable as an acyl donor.
Scheme 2. Lipase-catalyzed kinetic resolution of rac-3 and subsequent
extractive separation of the products with a biphasic fluorous/organic
solvent system.
Experimental Section
A solution of rac-3 (1.22 g, 10 mmol) in MeCN (65 mL) was treated with
ester 2 (8.61 g, 15 mmol) and CAL-B (Chirazyme L-2, c.-f., lyo., Roche
Diagnostics, Mannheim; 2.00 g). The reaction mixture was stirred at
ambient temperature until the conversion reached approximately 50%
(19 h). The enzyme was filtered off and the solid residue was washed with
acetone (2 Â 40 mL). The combined filtrates were evaporated under
reduced pressure and the residue was dissolved in MeOH (25 mL). The
In order to determine the yields and enantiomeric excesses
(ee)[7] of the products, the reaction mixture was worked up
conventionally with purification by flash column chromatog-
raphy to yield (R)-4 (46% yield, ee > 99%) and (S)-3 (41%
yield, ee > 99%). In comparison, the resolution of rac-3 with
Angew. Chem. Int. Ed. 2001, 40, No. 13
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4013-2493 $ 17.50+.50/0
2493

COMMUNICATIONS
resulting solution was extracted with n-C6F14 (6 Â 25 mL). The organic
phase was concentrated to dryness to yield (S)-3 (0.59 g, 48% yield, 99%
ee) with approximately 1% of (R)-4. From the fluorous phase a mixture of
(R)-4 (ee 98%) and the excess of 2 (8.50 g) was isolated.
Betzemeier, M. Cavazini, S. Quici, P. Knochel, Tetrahedron Lett. 2000,
41, 4343 ± 4346.
[3] a) D. P. Curran, M. Hoshino, J. Org. Chem. 1996, 61, 6480 ± 6481; b) P.
Wipf, J. T. Reeves, Tetrahedron Lett. 1999, 40, 4649 ± 4652; c) S. Röver,
P. Wipf, Tetrahedron Lett. 1999, 40, 5667 ± 5670.
[4] a) D. P. Curran, Angew. Chem. 1998, 110, 1231 ± 1255; Angew. Chem.
Int. Ed. 1998, 37, 1174 ± 1196; b) U. Diederichsen, Nachr. Chem. Tech.
Lab. 1999, 47, 805 ± 809.
Saponification of (R)-4: The mixture of (R)-4 and 2 was dissolved in THF
and water (1/1, 40 mL) together with LiOH (0.64 g, 26.7 mmol), and the
mixture was heated to reflux for 3 h. Subsequently, the reaction mixture
was diluted with cyclohexane (100 mL), cooled to 08C, and filtered. The
filter cake was washed with a mixture of cyclohexane (100 mL) and tert-
butyl methyl ether (30 mL). The filtrate was concentrated to dryness to
yield (R)-3 (0.57 g, 47%, 98% ee). The remaining solid filter cake (7.35 g,
98%) consists of the lithium salt of the perfluorinated carboxylic acid.
[5] One can expect that the currently very high price of the alcohol
(104.10 E for 50 g, Avocado/ABCR) should go down significantly if
there is a need for bulk quantities.
1
[6] B.p. 578C (1  10 3 mbar); H NMR (300 MHz, CDCl3): d ˆ 2.52 (tt,
2
1J ˆ 8.4, J ˆ 8.0 Hz, 2H), 2.76 (t, J ˆ 8.4 Hz, 2H), 4.42 (q, J ˆ 8.4 Hz,
2H).
Received: January 25, 2001
[7] Before determination of the ee value, (R)-4 was saponified to the
corresponding alcohol (R)-3. The ee values of the enantiomeric
alcohols (R)- and (S)-3 were determined by HPLC on a Chiralcel OJ
column (250 Â 4.6 mm; eluent: n-heptane/n-propanol 95/5; flow rate:
1 mLmin 1; UV detection at 254 nm). The absolute configurations of
the enantiomeric alcohols were assigned by comparison with com-
mercially available authentic samples.
Revised: March 12, 2001 [Z16506]
[1] a) C.-S. Chen, C.-J. Sih, Angew. Chem. 1989, 101, 711 ± 724; Angew.
Chem. Int. Ed. Engl. 1989, 28, 695 ± 707; b) W. Boland, C. Fröûl, M.
Lorenz, Synthesis 1991, 1049 ± 1072; c) K. Faber, S. Riva, Synthesis
1992, 895 ± 910; d) E. Santaniello, P. Ferraboschi, P. Grisenti, Enzyme
Microb. Technol. 1993, 15, 367 ± 382; e) F. Theil, Chem. Rev. 1995, 95,
2203 ± 2227; f) U. T. Bornscheuer, R. J. Kazlauskas, Hydrolases in
Organic Synthesis, Wiley-VCH, Weinheim, 1999; g) G. Carrea, S.
Riva, Angew. Chem. 2000, 112, 2312 ± 2341; Angew. Chem. Int. Ed.
2000, 39, 2226 ± 2254.
[8] K. Laumen, D. Breitgoff, M. P. Schneider, J. Chem. Soc. Chem.
Commun. 1988, 1459 ± 1461.
[9] Extraction was carried out in an ordinary separation funnel. The
residual 1% of alcohol (S)-3 left in the fluorous phase could either be
caused by contamination of the glassware (for example, a stopcock) or
by incomplete separation of the phases. However, a reextraction of the
fluorous phase with methanol should remove the trace of (S)-3.
[10] We have not tested whether esterases or proteases will also accept
long-chain fluoroesters.
Â
Â
[2] a) I. T. Horvath, J. Rabai, Science 1994, 266, 72 ± 75; b) J. A. Gladysz,
Science 1994, 266, 55 ± 56; c) B. Cornils, Angew. Chem. 1997, 109,
2147 ± 2149; Angew. Chem. Int. Ed. Engl. 1997, 36, 2057 ± 2059; d) B.
2494
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4013-2494 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 13