Bioreductive synthesis of perfluorinated chiral alcohols

Article abstract of DOI:10.1016/j.tetlet.2006.05.061

Perfluorinated chiral alcohols are interesting building blocks for pharmaceuticals and agrochemicals. Different chiral (R)- and (S)-configured perfluorinated alcohols were produced by asymmetric reduction of the corresponding ketones. Commercially available alcohol dehydrogenases were used as catalyst in combination with different cofactor regenerating systems. High selectivities of >99% were observed in most cases. The results also demonstrate the influence of the CF₃ group on reactivity and enantioselectivity of alcohol dehydrogenases.

Full text of DOI:10.1016/j.tetlet.2006.05.061

Tetrahedron Letters 47 (2006) 4803–4806
Bioreductive synthesis of perfluorinated chiral alcohols
Thomas C. Rosen,* Ralf Feldmann, Pascal Dunkelmann and Thomas Daußmann
¨
Julich Chiral Solutions GmbH, Prof.-Rehm-Str. 1, D-52428 Ju¨lich, Germany
Received 28 April 2006; revised 9 May 2006; accepted 10 May 2006
Abstract—Perfluorinated chiral alcohols are interesting building blocks for pharmaceuticals and agrochemicals. Different chiral (R)-
and (S)-configured perfluorinated alcohols were produced by asymmetric reduction of the corresponding ketones. Commercially
available alcohol dehydrogenases were used as catalyst in combination with different cofactor regenerating systems. High selectiv-
ities of >99% were observed in most cases. The results also demonstrate the influence of the CF3 group on reactivity and enantio-
selectivity of alcohol dehydrogenases.
Ó 2006 Elsevier Ltd. All rights reserved.
Fluorinated compounds are important building blocks
for biologically active compounds, such as pharmaceuti-
cals and agrochemicals, due to the unique properties of
the fluorine substituent.1 Because of the need of chiral
molecules in new drug developments, the preparation
of chiral fluorinated building blocks is of particular
interest. As example ethyl (R)-3-hydroxy-4,4,4-trifluoro-
butyrate was used as an intermediate for the antide-
pressant Befloxatone.2 Methods for the stereoselective
synthesis of chiral fluorinated alcohols have been de-
scribed.3 Especially, asymmetric reduction of prochiral
fluorinated ketones is an efficient approach. As reagents,
chiral boranes,3 and chiral aluminium hydride reagents,4
or transition metal-catalyzed hydrogenation5 were used.
Also bioreductive techniques were examined. Whole
cells such as baker’s yeast were applied for the reduction
of a range of fluorinated ketones.6 Still, most of these
methods yield the products with enantiomeric excess less
than 99%. Isolated alcohol dehydrogenases (ADH) are
usually more selective than whole cell systems such as
baker’s yeast, which usually contains different enzymes
with opposite stereoselectivity. We recently produced
different chiral fluorinated alcohols as building blocks
for early stage developments of new pharmaceuticals.
These products were synthesized by asymmetric reduc-
tion of perfluorinated ketones with both (R)- and
(S)-selective commercially available alcohol dehydro-
genases. The following alcohol dehydrogenases were
used as isolated enzymes: alcohol dehydrogenase from
Lactobacillus brevis (ADH LB),7 alcohol dehydrogenase
from Thermoanaerobacter spec. (ADH T),8 and two
alcohol dehydrogenases from Rhodococcus species
(ADH RS1 and RS2).9 ADH LB is the only (R)-selective
enzyme among these biocatalysts.
Initially, the activity of these ADHs for the conversion
of the perfluorinated aliphatic and aromatic ketones as
well as a b-ketoester (1–5) was measured using a photo-
metrical assay (Table 1).10 The results revealed good
activities of ADH LB and T for the conversion of all
tested perfluorinated ketones. Only for the sterically
demanding ketone, 1-(9-anthryl)-2,2,2-trifluoroetha-
none (4), no activity was found. Low activity was found
for ADH RS2. ADH RS1 showed only activity for the
aliphatic ketones 1, 2 and 5. Comparison of the activities
with the nonfluorinated ketone shows that replacement
of a CH3 by a CF3 group leads to significantly decreased
activity. The effect is very significant for ADH RS1 (en-
tries 5 and 6). This indicates the larger steric demand of
the trifluoromethyl group. Interestingly, higher activities
were found for 3,3,4,4,4-pentafluorobutanone (2)
compared to 1,1,1-trifluoroacetone (1) and 2-butanone,
indicating higher reactivity of the carbonyl group.
After measuring the photometrical activity based on
NAD(P)H consumption, small scale conversions were
made to examine yield and selectivity for the ADH-cat-
alyzed reduction of the perfluorinated ketones 1–5. The
hydride donating cofactors NADH or NADPH were
regenerated by different methods.
*
Corresponding author. Tel.: +49 2461 980 403; fax: +49 2461 980
410; e-mail addresses: t.rosen@julich.com; t.rosen@gmx.net
0040-4039/$ - see front matter Ó 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetlet.2006.05.061

4804
T. C. Rosen et al. / Tetrahedron Letters 47 (2006) 4803–4806
Table 1. Results of photometrical activity of different isolated ADHs for the reduction of perfluorinated ketones
Entry
R1
Rf
Relative activity of ADHs in %
ADH LBa
ADH Tb
ADH RS1c
ADH RS2d
1
2
3
4
5
6
7
8
9
CH3
CH3
CH3
CH3
CH2–CO2Et
CH2–CO2Et
Ph
Ph
9-Anthryl
CF3
CH3
C2F5
C2H5
CF3
CH3
CF3
CH3
CF3
61
73
133
94
6.75
148
101
100
1.25
26
100
65
72
10
21
28
52
1.8
1.4
10
30
10
2
296
<1
15
10
2.8
n.d.
6.5
6.8
89
4.6
23
6.9
<1
a Characteristic activity is 1000 U/mL with the standard substrate acetophenone.
b Characteristic activity is 1000 U/mL with the standard substrate acetone.
c Characteristic activity is 100 U/mL with the standard substrate p-chloro acetophenone.
d Characteristic activity is 59 U/mL with the standard substrate chloro acetone.
Since ADH LB and ADH T tolerate high concentra-
tions of 2-propanol, the cofactor can be regenerated
by addition of an excess of this alcohol (10–40 vol %).
In a typical experiment, 2 mmol (200 mM) of the corre-
sponding ketones 1–5 and 0.001 mmol (0.1 mM) of
NADP+ were dissolved in a mixture of 6–8 mL Tris/
HCl-buffer (pH 7.0). 2-Propanol (2–4 mL) was added
so that a reaction volume of 10 mL is obtained. The
reaction was started by the addition of a crude cell-free
extract of the corresponding enzyme (1 U/mL based on
results of the photometrical assay). The reaction was
stirred for 24 h at 30 °C. Alternatively, a second enzy-
matic system such as glucose dehydrogenase (GDH)
from Bacillus subtilis (GDH BS) or Bacillus megaterium
(GDH BM) and glucose was used as cofactor regenera-
tion system. Formate dehydrogenase from Candida
boidinii (FDH CB) and sodium formate was used in con-
versions with ADH RS1. For these experiments, GDH
BM/BS (1 U/mL) or FDH CB (1 U/mL) was added to
the reaction mixture. In case of GDH, 3.5 mmol
(350 mM) glucose and 3.0 mmol (300 mM) CaCO3 were
added, for FDH 5.0 mmol (500 mM) sodium formate
was added. Results of the small conversions are summa-
rized in Table 2 (Scheme 1).
rinated acetophenone, it was expected that ADH T and
RS2 would yield the (R)-enantiomer. Models were devel-
oped to explain the unexpected stereoselectivity for the
reduction of 3 with aluminium reagents.4b,c From these
models, it can be concluded that the transition state for
the ADH T-catalyzed reaction is influenced by the steric
demand of the CF3 group leading to an inverse selectivity.
Interestingly, this does not influence the selectivity of
ADH LB. No conversion of 1-(9-anthryl)-2,2,2-trifluoro-
ethanone (4) was observed for all enzymes. Steric hin-
drance is expected to be problematic. Finally, ethyl
4,4,4-trifluoro-3-oxobutyrate (5) was reacted with differ-
ent ADHs. All ADHs showed very high enantioselectivity
of >99%. ADH LB yielded the corresponding (S)-enan-
tiomer as determined by optical rotation.14 All other
enzymes gave the corresponding (R)-isomer. Reactions
with ADH LB or T and 40% 2-propanol gave only low
conversions after 24 h. Full conversion was observed
after longer reaction times.
The results were used to produce (S)- and (R)-1,1,1-tri-
fluoro-2-propanol ((S)- and (R)-6) on a 100 g scale for
an early stage pharmaceutical development project.
Both enantiomers were produced before according to
the literature by asymmetric reduction of 1,1,1-trifluoro-
acetone,3c,6b or by resolution reactions.15 Unfortunately,
cofactor regeneration with 2-propanol was not possible
because of the similar boiling point. (R)-1,1,1-Triflu-
oro-2-propanol ((R)-6) was produced using ADH LB
and GDH BM for cofactor regeneration.16 The reaction
was performed at 15 °C, which led to a slightly im-
proved selectivity from 94% (Table 2, entry 2) to 97%.
(S)-1,1,1-Trifluoro-2-propanol ((S)-6) was produced
using ADH RS1 and FDH CB for cofactor regenera-
tion. The reaction was also performed at 15 °C, which
resulted in an ee of 93% as observed in the small scale
preparations (Table 2, entry 1).17
Both aliphatic ketones, 1,1,1-trifluoroacetone (1) and
3,3,4,4,4-pentafluorobutanone (2), were converted by all
of the tested ADHs. Nevertheless, enantioselectivity of
(S)-selective ADHs was only 80–93% for 1. Using ADH
LB as catalyst, the corresponding (R)-configured alcohol
was obtained with ee 94% and 98% dependent on the
cofactor regeneration.11 If the perfluorinated group is
increased from CF3 to C2F5, ADH LB and ADH RS1
would give the corresponding alcohol with high ee of
>99%. Only low selectivities were found for the reduction
of the unfluorinated 2-butanone: ADH LB 97% ee (2-pro-
panol), ADH T 34% ee (2-propanol), ADH RS1 62% ee
(FDH CB).12 The larger steric demand of the trifluoro-
methyl group decreases enzymatic activity and increases
the selectivity of the ADHs. Enzymatic activity for the
conversion of 1,1,1-trifluoroacetophenone (3) was only
found for ADH LB, T and RS2. ADH LB and T gave
both the (S)-enantiomer, with very high selectivities of
ee >99%.13 From the results for the conversion of unfluo-
The results demonstrate the broad applicability of alco-
hol dehydrogenases for the production of perfluorinated
chiral alcohols by asymmetric reduction. It was also
shown that the introduction of a perfluorinated residue
influences the activity and selectivity of analogous
reactions.

Table 2. Results of the asymmetric reduction of perfluorinated ketones with different isolated recombinant alcohol dehydrogenases
Entry R1
Rf
ADH LB
ADH T
ADH RS1
ADH RS2
NADPH
regeneration
Conversiona eeb
%
%
NADPH
regeneration
Conversiona eeb
%
%
NADH
regeneration
Conversiona eeb
NADH
regeneration
Conversiona eeb
%
%
%
%
1c
2c
3c
4d
5d
6c
7c
8e
CH3
CF3
CF3
20%
n.d.
100
53
87
64
0
98 (R) 20%
iPrOH
n.d.
83 (S)
80 (S)
FDH CB
NaHCO2
100
93 (S) GDH BM
90
90
iPrOH
Glucose
CH3
GDH BM
Glucose
94 (R) GDH BM
8
Glucose
CH3
C2F5 GDH BM
Glucose
>99 (R)
FDH CB
NaHCO2
100
n.t.
n.t.
n.t.
n.t.
90f
>99 (S)
n.t.
12
Ph
CF3
CF3
CF3
CF3
30%
iPrOH
>99 (S)
99 (S)
n.d.
30%
89
55
0
>99 (S)
>99 (S)
n.d.
GDH BM
Glucose
98 (R)
n.d.
iPrOH
Ph
GDH BS
Glucose
GDH BS
Glucose
9-Anthryl
9-Anthryl
30%
iPrOH
30%
iPrOH
GDH BM
glucose
0
GDH BS
Glucose
0
n.d.
GDH BS
Glucose
0
n.d.
CH2CO2Et CF3
40%
iPrOH
23
>99 (S)
40%
iPrOH
4
>99 (R) GDH BS
>99 (R) GDHBM
23
>99 (R)
Glucose
Glucose
n.t. – not tested.
n.d. – not determined.
a Conversion after 24 h, determined by GC.
b Determined by chiral GC.
c Concentration of the ketone, 200 mM.
d Concentration of the ketone, 100 mM.
e Concentration of the ketone, 250 mM.
f Concentration of the ketone, 50 mM.

4806
T. C. Rosen et al. / Tetrahedron Letters 47 (2006) 4803–4806
8. Daußmann, T.; Hennemann, H.-G. Patent PCT/EP2005/
Alcohol
dehydrogenase
OH
O
OH
006034.
or
+
R f
R
9. Julich Chiral Solutions GmbH, a Codexis company.
10. The enzymatic assay for the activity measurement of the
alcohol dehydrogenases was done as follows: 20 lL of a
solution of NAD(P)H (10 mM in potassium phosphate
buffer, pH 7.0) was added to a solution of 970 lL of
50 mM potassium phosphate buffer, pH 7.0, and 10 mM
of the corresponding ketone at 30 °C. The reaction was
started by adding 10 lL of the corresponding enzyme
solution (diluted to 0.5–1.5 U/mL). The cuvettes were
thermostated and the activity was measured at 30 °C. The
consumption of NAD(P)H was measured by UV pho-
tometry at 340 nm. The activities were calculated as
enzyme units (U, i.e., lmol/min) by using a molar
Rf
R
R f
R
NAD(P) , 30 °C
Cofactor regeneration
1, R = CH , R = CF
3
2, R = CH , R = C F
3, R = Ph, R = CF
4, R = 9-Anthryl, R = CF
3
f
3
f
2
3
5
f
f
3
5, R = CH CO Et, R = CF3
2
2
f
Scheme 1. Reduction of perfluorinated ketones with alcohol dehydro-
genases.
References and notes
1. (a) Biomedical Effects of Fluorine Chemistry; Filler, R.,
Kobayashi, Y., Eds.; Kodama Ltd/Elsevier: Tokyo/
Amsterdam, 1982; (b) Fluorine-Containing Molecules;
Liebman, J. F., Greenberg, A., Dolbier, W. R., Jr., Eds.;
VCH: New York, 1988; (c) Smart, B. E. In Organofluorine
Chemistry: Principles and Commercial Applications;
Banks, R. E., Smart, J. B. E., Tatlow, W., Eds.; Plenum
Press: New York, 2004; pp 57–88; (d) Smart, B. E.
J. Fluorine Chem. 2001, 109, 3.
extinction coefficient of 6220 MÀ1 cmÀ1
.
11. Absolute configuration was determined by optical rotation
values in comparison to the value given in Ref. 3d,14.
12. Not published results of Julich Chiral Solutions GmbH.
13. Absolute configuration was determined by optical rota-
tion: +13.4 (c 19, ethanol), compare also Talma, A. G.;
Jouin, P.; De Vries, J. G.; Troostwijk, C. B.; Werumeus
Buning, G. H.; Waninge, J. K.; Visscher, J.; Kellogg, R.
M. J. Am. Chem. Soc. 1985, 107, 3981; and Kasai, M.
J. Org. Chem. 1983, 48, 459.
2. Rabasseda, X.; Sorbera, L. A.; Castaner, J. Drugs Future
1999, 24, 1057.
14. Seebach, D.; Renaud, P.; Schweizer, W. B.; Zuger, M.;
¨
3. (a) Hanzawa, Y.; Kawagoe, K.; Kobayashi, Y. Chem.
Pharm. Bull. 1987, 35, 2609; (b) Bravo, P.; Resnati, G.
Tetrahedron: Asymmetry 1990, 10, 661; (c) Ramachan-
dran, P. V.; Teodorovic, A. V.; Brown, H. C. Tetrahedron
1993, 49, 1725; (d) Ramachandran, P. V.; Teodorovic, A.
V.; Gong, B.; Brown, H. C. Tetrahedron: Asymmetry 1994,
5, 1075; (e) Ramachandran, P. V.; Gong, B.; Brown, H. C.
J. Org. Chem. 1995, 60, 41; (f) Haufe, G. J. Fluorine Chem.
2004, 125, 875.
4. (a) Giacomelli, G.; Menicagli, R.; Lardicci, L. J. Org.
Chem. 1973, 38, 2370; (b) Nasipuri, D.; Bhattacharya, P.
K. J. Chem. Soc., Perkin Trans. 1 1977, 576; (c) Giaco-
melli, G. P.; Menicagli, R.; Lardicci, L. J. Am. Chem. Soc.
1975, 97, 4009.
Brienne, M.-J. Helv. Chim. Acta 1984, 67, 1843.
15. (a) Feigl, D. M.; Mosher, H. S. J. Org. Chem. 1968, 33,
4242; (b) Rotticci, D.; Ha¨ffner, F.; Orrenius, C.; Norrin,
T.; Hult, K. J. Mol. Catal. B: Enzym. 1998, 5, 267; (c)
Crawford, J. W. C. J. Chem. Soc. C 1967, 22, 2332.
16. In 7 L of 100 mM Tris buffer (100 mM, pH 7.5), 167 mL
1,1,1-trifluoroacetone, 35 kU ADH LB (5 U/mL), 14 kU
GDH BM (2 U/mL), 455 g glucose (300 mM) and 0.55
NADP+ (0.1 mM) were dissolved and stirred for 70 h at
15 °C. The pH was adjusted by titration with 4 M NaOH.
The reaction was stopped by ultrafiltration. After extrac-
tion the product was purified by distillation. The product
(R)-6 formed an azeotrope with tert-butyl methyl ether
(75:25 by NMR) at 73 °C.
5. Pirkle, W. H.; Hauske, J. R. J. Org. Chem. 1977, 42, 2436.
6. (a) Kitazume, T.; Ikekawa, N. Chem. Lett. 1983, 237; (b)
Bucciarelli, M.; Forni, A.; Moretti, L.; Torre, G. Synthesis
1983, 897; (c) Kitazume, T.; Ishikawa, N. Chem. Lett.
1984, 587; (d) Zhang, J.; Duetz, W. A.; Witholt, B.; Li, Z.
Chem. Commun. 2004, 2120.
7. (a) Hummel, W. Adv. Biochem. Eng. 1997, 58, 145; (b)
Hummel, W.; Riebel, B. Patent U.S. 6,225,099 and EP
0796914 A2.
17. In 7 L of 100 mM Tris buffer (100 mM, pH 7.0), 167 mL
1,1,1-trifluoroacetone, 35 kU ADH RS1 (5 U/mL), 21 kU
FDH CB (3 U/mL), 238 g sodium formate (500 mM) and
1 g NAD+ (0.2 mM) were dissolved and stirred for 140 h
at 15 °C. The pH was adjusted by titration with formic
acid. The reaction was stopped by ultrafiltration. After
extraction the product was purified by distillation. The
product (S)-6 formed an azeotrope with tert-butyl methyl
ether (75:25) at 73 °C.