Candida tenuis xylose reductase catalyzed reduction of aryl ketones for enantioselective synthesis of active oxetine derivatives

Article abstract of DOI:10.1002/chir.22082

Candida tenuis xylose reductase shows high catalytic efficiencies in carbonyl reduction of acetophenone and 1-phenyl-1-propanone derivatives. The quite low substrate solubility in aqueous buffer systems is circumvented by addition of methanol or by two-phase solvent systems. In the latter, methanol improves the substrate phase transfer as solvent mediator and leads to reasonable space/time yields. Resulting enantiomerically pure chiral alcohols are key intermediates for synthesis of active pharmaceutical ingredients. (R)-Atomoxetine is exemplarily synthesized in four steps, and the further use for generation of other oxetine derivatives and a polo-like kinase 1 inhibitor are discussed. Copyright

Full text of DOI:10.1002/chir.22082

CHIRALITY (2012)
Candida tenuis Xylose Reductase Catalyzed Reduction of Aryl Ketones
for Enantioselective Synthesis of Active Oxetine Derivatives
1
MICHAEL VOGL,1 REGINA KRATZER,2 BERND NIDETZKY,2 AND LOTHAR BRECKER
*
1Department of Organic Chemistry, University of Vienna, Wien, Austria
2Institute of Biotechnology and Biochemical Engineering, Graz University of Technology, Graz, Austria
ABSTRACT
Candida tenuis xylose reductase shows high catalytic efficiencies in carbonyl
reduction of acetophenone and 1-phenyl-1-propanone derivatives. The quite low substrate solubility
in aqueous buffer systems is circumvented by addition of methanol or by two-phase solvent sys-
tems. In the latter, methanol improves the substrate phase transfer as solvent mediator and leads
to reasonable space/time yields. Resulting enantiomerically pure chiral alcohols are key intermedi-
ates for synthesis of active pharmaceutical ingredients. (R)-Atomoxetine is exemplarily synthesized
in four steps, and the further use for generation of other oxetine derivatives and a polo-like kinase 1
inhibitor are discussed. Chirality 00:000–000, 2012. © 2012 Wiley Periodicals, Inc.
KEY WORDS: aryl ketone reduction; Candida tenuis xylose reductase; enantioselectivity; oxetine;
phase transfer
selective serotonin reuptake inhibitors.21,23 These secondary
alcohols can also be applied to synthesize promising serine/
INTRODUCTION
Chiral secondary alcohols are very suitable key intermediates
for the synthesis of active pharmaceutical ingredients and
hence an important group of substances in pharmaceutical
industry.1–3 It is necessary to use them in a high enantiomeric
purity to exclusively gain the desired most active enantiomer
of the final product. An otherwise obtained racemic mixture
has to be separated into their enantiomers causing the loss of
up to half of the material.4 Alternative application of racemic
mixtures in pharmaceuticals is possible but often leads to a
decrease of therapeutic effect or even to unwanted side effects.5
Therefore, there is a large interest to develop procedures
gaining enantiomerically pure secondary alcohols. An enzyme
catalyzed transformation of prochiral precursors is a sensible
method for this purpose. The disadvantage of gaining only
one stereoisomer of the resulting alcohol can be abolished by
inversion of the desired chiral center during further synthe-
sis.6,7 In particular, enzymes accepting a variety of non-natural
substrates with high transformation rates are interesting
candidates for applications in such enantioselective syntheses.8–12
However, a large number of these prochiral precursors have
a quite non-polar character and, hence, show a very small
solubility in aqueous buffer systems, which are necessary for
biocatalyzed reactions. An alternative use of organic solvents
for biotransformations is limited, in particular, when coenzymes
are involved in the reactions.13 Another approach is the
application of a biphasic solvent system with phase transfer of
substrates and products.14
We describe the application of xylose reductase [E.C.1.1.
1.175] from Candida tenuis (CtXR) to enantioselectively synthe-
size chiral intermediates for the formation of two different
groups of drugs.15–21 This enzyme reduces its natural substrate
xylose to xylitol using NADPH or NADH as cofactor.22 However,
it also accepts the carbonyl functions of several other com-
pounds for reduction. In particular, aromatic a-keto esters and
substituted acetophenones are enantioselectively transferred to
the corresponding secondary alcohols.15,16 The poor water
solubility of the aryl ketones is circumvented by a biphasic setup
of the transformations.14 The resulting chiral key intermediates
gained are exemplarily used to obtain oxetines, which are potent
© 2012 Wiley Periodicals, Inc.
threonine kinase polo-like kinase
1 (PLK1) inhibitor
candidates.20
MATERIALS AND METHODS
Chemicals
If not otherwise indicated, all reagents were obtained from commercial
suppliers (Sigma-Aldrich-Fluka, Vienna, Austria or Acros Organics, Geel,
Belgium) and used without further purification or drying. The recombinant
CtXR has been prepared as described earlier.17,35 TLC was performed with
Merck silica gel 60 F254 pre-coated plates. Silica gel column chromatography
was carried out using silica gel 60 M (40–63 mm, Macherey-Nagel, Düren; Germany).
NMR Spectroscopy
NMR spectra were recorded on a Bruker Avance DRX-400 (Bruker,
Rheinstetten, Germany) at 400.13 MHz. Measurements were performed
at a temperature of 298.1 K using the Bruker Topspin 1.3 software. The
compounds were dissolved in a 700 ml CDCl3 and transferred into 5 mm
high-precision NMR sample tubes. The solvent signals of CDCl3 at
7.26 ppm (1H) and 77.16 ppm (13C) were used as shift reference.
UV Spectrometry to Measure CtXR Catalytic Efficiency
The catalytic efficiency was performed at a Beckman DU-800 spectrometer
(Beckman Coulter, Krefeld, Germany) observing the decreasing of the
NADH signal at 340 nm. Initial rate measurements of compounds 1–4 and
6–10 are carried out using reported protocols.16,18
Analytical Chromatography
The enantiomeric excess is determined by HPLC analysis using a
reversed phase CHIRALPAK AD-RH column from Daicel (purchased at
VWR International, Vienna, Austria) on a LaChrom HPLC system
(Merck-Hitachi) equipped with an L-7400 UV-detector (280 nm). Best
resolutions of the (R)- and (S)-alcohols were obtained with acetonitrile
and water (20:80, by volume) as eluent at a flow rate of 0.5 ml/min and
a temperature of 40 ꢀC.
*Correspondence to: Lothar Brecker, Department of Organic Chemistry,
University of Vienna, Währinger Straße 38, A-1090 Wien, Austria.
E-mail: lothar.brecker@univie.ac.at
Received for publication 19 March 2012; Accepted 16 May 2012
DOI: 10.1002/chir.22082
Published online in Wiley Online Library
(wileyonlinelibrary.com).

VOGL ET AL.
2,4,6-Tribromo acetophenone (4). Compound 4 is prepared from
Solution Experiments
40 ml 2,4,6-tribromobenzene, following the procedure for synthesis of
compound 2. The yield is 0.1 g (1%) 4. 1H NMR (400 MHz, CDCl3), d,
ppm: 2.66 (s, 3H, Me), 7.74 (s, 2H, Ph); 13C NMR (400 MHz, CDCl3), d,
ppm: 30.2 (q), 118.2 (2C, s), 121.4 (s), 134.0 (2C, d), 142.4 (s), 199.3 (s).
2,4-Dichloro acetophenone (2) in buffer/methanol mixtures. Mix-
tures of methanol and potassium phosphate buffer (50mM) are prepared
with 5, 10, 15, 20, 30, and 40% methanol (v/v), respectively. We use 10ml
of each mixture to solve 2,4-dichloro acetophenone (2, 2.5 mM) and NADH
(3 mM). CtXR (0.11 units) is added, and after 12 and 14h, two samples are
taken. They are extracted with DCM and concentrated under reduced
pressure. Identification and quantification of the resulting (S)-1-(20,40-
dichloro phenyl) 1-ethanol (5) are determined by H NMR (Table 2). H
NMR (400 MHz, CDCl3), d, ppm: 1.45 (d, J = 6.39 Hz, 3H, Me), 5.22
(q, J = 6.39Hz, 1H, CHOH), 7.25 (1H, m, Ph), 7.32 (1H, m, Ph), 7.57
(1H, m, Ph); 13C NMR (400MHz, CDCl3), d, ppm: 23.6 (q), 66.5 (d), 127.4
(d), 127.5 (s), 129.1 (d), 132.2 (d), 133.4 (d), 141.7 (s).
1-(20,40-Dichloro phenyl) 1-propanone (6). Compound
6 is
prepared from 10ml 1,3-dichlorobenzene and propionyl chloride (1.1 ml,
10mmol), following the procedure for synthesis of compound 2. The yield
1
1
1
is 0.53 g (26%) 6. H NMR (400 MHz, CDCl3), d, ppm: 1.20 (t, J = 7.2, 3H,
Me), 2.92 (q, J = 7.2, 2H, CH2), 7.09 (m, 1H, Ph), 7.21 (m, 1H, Ph), 7.51
(m, 1H, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 11.2 (q), 35.8 (t), 127.1
(d), 128.4 (s), 130.0 (d), 133.4 (d), 134.5 (d), 137.8 (s), 201.4 (s).
3-Chloro-1-phenyl 1-propanone (7). Compound 7 is prepared
from 5 ml benzene and 3-chloropropanoyl chloride (0.65 ml, 6.8 mmol),
following the procedure for synthesis of compound 2. The yield is
0.27 g (24%) 7. 1H NMR (400 MHz, CDCl3), d, ppm: 3.46 (t, J = 6.9 Hz,
2H, CH2), 3.93 (t, J = 6.9 Hz, 2H, CH2Cl), 7.48 (m, 2H, Ph), 7.59 (m, 1H,
Ph), 7.96 (m, 2H, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 38.7 (t),
42.3 (t), 128.0 (2C, d), 128.7 (2C, d), 133.5 (d), 136.3 (s), 196.7 (s).
2,4-Dichloro acetophenone (2) distribution in two-phase
systems. A solution of 1.0 ml 2,4-dichloro acetophenone (2, 1.0 to 4.0M)
in hexane, ethyl acetate, ether, toluene, or cyclohexane, respectively, is pre-
pared. They are intensively stirred for 24 h with 3 ml of a 50 mM potassium
phosphate buffer. Samples (0.5 ml) are taken after 10 and 24 h, respectively,
from the aqueous layers after a rapid layer separation. The respective
aqueous layer is extracted three times with methylenchloride. The extracts
are combined and directly concentrated under reduced pressure. Remaining
residues are dissolved in CDCl3 and analyzed by H NMR to determine the
amount of 2,4-dichloro acetophenone (2). No detectable amounts of 2 have
been recorded.
An amount of 0.4 ml methanol is added to the original biphasic systems
containing 2.5 M of compound 2. Resulting mixtures are intensively stirred,
and samples are taken after 30 and 60 min, respectively, from the aqueous
layers after a rapid layer separation. The respective aqueous layer is
extracted and analyzed as described above. 2,4-Dichloro acetophenone
(2) has been detected in ca. 0.05 mM amounts in all aqueous layers.
1-(2,4-Dichlorophenyl)-3-(N-methylamino)-propan-1-one (8)34. 2,4-
Dichloro acetophenone (2, 0.2g, 1.0 mmol), methylamine hydrochloride
(0.5 g, 5 mmol), paraformaldehyde (0.25 g, 1.7 mmol), and concentrated
HCl (0.1 g) are heated in 15ml ethanol in a sealed flask at 110 ꢀC for 24h.
After cooling to room temperature, the solvent is removed and 25ml of ethyl
acetate was added. The resulting suspension is vigorously stirred at room
temperature for 6 h, filtered, and washed with ethyl acetate to afford 0.12 g
(50%) (8). 1H NMR (400MHz, CDCl3), d, ppm: 2.32 (s, 1H, NH), 2.68
(s, 3H, Me), 3.34 (t, J = 5.86 Hz, 2H, NCH2), 3.45 (t, J = 5.86 Hz, 2H, CCH2O),
7.40 (dd, J = 2.07/8.62 Hz, 1H, Ph), 7.55 (d, J = 2.07 Hz, 1H, Ph), 7.61
(d, J= 8.62Hz, 1H, Ph).
1
CtXR transformation in a two-phase system with methanol as
phase connector. A solution of 2.0 ml 2,4-dichloro acetophenone (2, 2.0
or 4.0 M) in cyclohexane, hexane, ethyl acetate, or toluene, respectively, is
prepared. They are intensively stirred for 24 h with 0.4 ml methanol and
2.0ml of a 50 mM potassium phosphate buffer containing 10mM NADH
and 0.55 units CtXR. Two samples (0.5 ml) are taken after 10 and 24h,
respectively, from the organic layer and concentrated under reduced
pressure. Identification and quantification of the resulting (S)-1-(20,40-
dichloro phenyl) 1-ethanol (5) are determined by 1H NMR (Table 3). Chemical
shifts are the same as those listed in distribution in two-phase systems.
3-(20-Chloro phenyl)-3-oxo propionitrile (9)25. A 0.33 M LDA
solution in 6 ml abs. THF is mixed and cooled under argon to ꢁ78ꢀC, before
0.1 ml acetonitrile (1.9 mmol) are added dropwise. After this the mixture is
stirred for 1 h at ꢁ78 ꢀC, a solution of methyl 2-chlorobenzoate (0.28 g,
1.7 mmol) in 0.5ml THF is added and stirred for 1 h. The reaction is allowed
to reach room temperature within 2 h and then quenched with water. The
mixture was acidified to pH4, extracted with ether, dried over Na2SO4,
and concentrated. The residue is purified by flash chromatography on silica
gel with ethyl acetate/petrolether 3:7 to afford 0.25g (85%) 9. 1H NMR
(400 MHz, CDCl3), d, ppm: 4.12 (s, 2H, CH2), 7.37 (m, 1H, Ph), 7.43 (m,
1H, Ph), 7.47 (m, 1H, Ph), 7.60 (m, 1H, Ph); 13C NMR (400 MHz, CDCl3),
d, ppm: 32.8 (t), 113.4 (s), 127.4 (s), 130.3 (d), 130.9 (d), 131.6 (d), 133.7
(d), 135.5 (s), 189.3 (s).
Synthesis and Spectroscopic Data
2-Chloro acetophenone (1). Compound 1 has been synthesized as
reported earlier by Vogl et al.16 1H NMR (400 MHz, CDCl3), d, ppm:
2.61 (s, 3H, Me), 7.28 (m, 1H, Ph), 7.35 (m, 1H, Ph), 7.38 (m, 1H, Ph),
7.51 (m, 1H, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 30.6 (q), 126.8
(s), 129.3 (d), 130.6 (d), 131.2 (s), 131.9 (d), 139.0 8 (s), 200.3 (s).
3-(20,40-Dichloro phenyl)-3-oxo propionitrile (10)25. Compound
10 is prepared from methyl 2,4-dichlorobenzoate (0.3 g, 1.5mmol),
following the procedure for synthesis of compound 9. The yield is 0.22g
(67%). 1H NMR (CDCl3) d 3.84 (2H, s), 7.28 (1H, m), 7.33 (1H, m), 7.54
(1H, m); 13C NMR (CDCl3) d 32.7 (t), 116.5 (s), 127.8 (s), 128.9 (d), 130.6
(d), 132.5 (d), 133.2 (s), 135.6 (s), 186.4 (s).
2,4-Dichloro acetophenone (2). AlCl3 (4.2 g, 32mmol) is mixed with
40 ml 1,3-dichlorobenzene, stirred, and cooled to 0–5 ꢀC. Then acetyl
chloride (1.7ml, 24mmol) is added dropwise. The mixture is stirred for
further 30min and heated to 50ꢀC. After completion of the reaction (DC
control), 50ml 2 M aqueous HCl and 150 ml ice water are added. The
mixture is extracted with CH2Cl2. The organic layer is washed with brine,
NaHCO3 solution and brine, dried over MgSO4, and concentrated. Result-
ing crude product is purified by flash chromatography on silica gel with
3-(20,60-Dichloro phenyl)-3-oxo propionitrile (11)25. Compound
11 is prepared from methyl 2,6-dichlorobenzoate (0.3 g, 1.5mmol),
following the procedure for synthesis of compound 9. The yield is 0.22g
(67%). 1H NMR (400MHz, CDCl3), d, ppm: 3.88 (s, 2H, CH2), 7.23 (m, 1H,
Ph), 7.28 (m, 2H, Ph); 13C NMR (400MHz, CDCl3), d, ppm: 32.7 (t), 115.2
(s), 128.9 (2C, s), 130.6 (2C, d), 132.5 (d), 135.6 8 s), 186.4 (s).
1
petrol ether/ethylacetate 9:1 to afford 2.57 g (57%) 2. H NMR (400 MHz,
CDCl3), d, ppm: 2.61 (3H, s), 7.27 (1H, m, Ph), 7.39 (1H, m, Ph), 7.50 (1H,
m, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 30.5 (q), 127.2 (d), 130.4 (d),
130.6 (d), 132.8 (s), 137.1 (s), 137.5 (s), 199.3 (s).
(S)-3-Chloro-1-phenyl 1-propanol (12). The transformation is
performed at 25ꢀC in 50ml aq. potassium phosphate buffer (50 mM,
pH7.0), which contained 10% methanol, 0.50 mmol NADH, and 0.37 mmol
compound 7. Recombinant CtXR (2.2 units) is added, and the transformation
is allowed to proceed for 24 h. After TLC indicated a wide transformation, the
solution is extracted three times with 20 ml DCM. The organic extracts are
combined, washed with brine, and dried over Mg2SO4. After removal of the
solvent in vacuum, the remaining residue was purified by flash
2-Bromo acetophenone (3). Compound 3 has been synthesized as
reported earlier by Vogl et al.16 1H NMR (400 MHz, CDCl3), d, ppm:
2.61 (s, 3H, Me), 7.27 (m, 1H, Ph), 7.35 (m, 1H, Ph), 7.44 (m, 1H, Ph),
7.59 (m, 1H, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 30.3 (q), 118.9
(d), 127.4 (s), 128.8 (d), 131.7 (d), 133.8 (d), 141.5 (s), 201.3 (s).
Chirality DOI 10.1002/chir

Candida tenuis XYLOSE REDUCTASE CATALYZED REDUCTION OF ARYL KETONES FOR ENANTIOSELECTIVE
SYNTHESIS OF ACTIVE PHARMACEUTICAL INGREDIENTS
chromatography on silica gel with DCM to afford 50 mg (80%) 12. 1H NMR
(400 MHz, CDCl3), d, ppm: 2.10 (m, 1H, CH2), 2.25 (m, 1H, CH2), 3.56 (m,
1H, CH2Cl), 3.75 (m, 1H, CH2Cl), 4.65 (dd, J= 4.75/8.55 Hz, 1H, CHOH),7.29
(m, 1H, Ph), 7.37 (m, 4H, Ph); 13C NMR (400 MHz, CDCl3), d, ppm: 41.3 (t),
41.7 (t), 71.2 8d), 125.7 (2C, d), 127.8 (d), 128.6 (2C, d), 143.6 (d).
(R)-Atomoxetine (R-15). (S)-3-Chloro-1-phenylpropan-1-ol (12, 30mg,
0.16 mmol) and ortho-cresol (20mg, 0.18 mmol) are dissolved in 50ml dry
THF. Diethylazodicarboxylat (30 mg, 0.17mmol) and triphenylphosphine
(0.05 g, 0.19mmol) are added to the mixture, which is then stirred for 16h
at room temperature. THF is removed on a rotary evaporator, and the
residue is extracted three times with hexane. The combined organic
extracts are concentrated and purified by flash chromatography on silica
gel with hexane/ether 4:1 to afford 0.029 g (70%, 0.09 mmol) of crude (R)-
1-chloro-3-(20-methylphenyloxy)-3-phenyl propane. It is directly dissolved
in 5 ml ethanol, and 0.5 ml aq. MeNH2 (40%) is added. The mixture is heated
in a closed vessel at 120ꢀC for 12h then cooled to room temperature and
poured onto water. After extraction with ether, the organic fraction is
washed with brine, dried over MgSO4, and concentrated in vacuum. The
remaining residue is purified by chromatography DCM/methanol 10:1 to
afford 0.015g (60%) (R)-atomoxetine (R-15) as hydrochloride. 1H NMR
(CDCl3) d 2.10 (s, 1H, NH), 2.17 (m, 2H, CH2), 2.26 (s, 3H, Me-Ph), 2.76
(s, 3H, NMe), 3.60 (m, 2H, CH2N), 5.12 (m, 1H, CHO–), 6.49 (m, 1H, Ph),
6.72 (m, 1H, Ph), 6.89 (m, 1H, Ph), 7.06 (m, 1H, Ph), 7.25 (m, 5H, Ph).
Fig. 1. Aryl-substituted acetophenones (a) as well as various substituted
acetophenones and phenyl propanones (b), which have been tested for their
acceptance as substrate by CtXR (Table 1).
(20,40-dichloro phenyl) 1-ethanol (5) has a very high enan-
tiomeric enrichment with ee > 98%, determined by chroma-
tography and as mosher ester with NMR spectroscopy.16
However, the potential use as chiral building block is lim-
ited, because it does not provide good possibilities for fur-
ther derivatization, except on the hydroxyl function.
A product oriented design of the substrate enables a wide
variety of reactions for consecutive transformations resulting
in enantiomerically pure target compounds. Slight elongation
of the short alkyl chain carrying an additional functionality
allows such optimized further synthesis. In order to assure
the CtXR acceptance of such substrates, we mainly tested
compounds with halogens in ortho- and para-positions of the
aromatic moiety. All these compounds are accessible in very
short syntheses.24,25 The catalytic activity of CtXR to reduce
them is, however, limited by the substrate size and by the
influences of additional functionalities on binding. Table 1
lists the catalytic activities of CtXR catalyzed reduction of
selected model compounds, which are shown in Figure 1a
and b. The substrates with functional groups in the alkyl chain
are, however, not reduced by the enzyme.
RESULTS AND DISCUSSION
Large Substrate Acceptance of CtXR
We recently described C. tenuis xylose reductase (CtXR) to
accept a couple of ortho-halogen substituted acetophenones as
well as aromatic a-keto esters with remarkably high transforma-
tion rates.15,16 This very good acceptance can rather be
explained by the electronic structure of the substrates than by
selective docking to the active site of the enzyme.16 We now
focus on the transformation of slightly larger acetophenone
derivatives with respect to their applicability in subsequent
synthesis (Scheme 1).
The electronic influence of the ring substituent has been
enlarged by the insertion of two or more electron-withdrawing
groups in ortho- and para-positions to the carbonyl group, while
the methyl group remained unchanged. The yet best accepted
2-chloro acetophenone (1) is altered to 2,4-dichloro acetophenone
(2), and 2-bromo acetophenone (3) is changed to 2,4,6-
tribromo acetophenone (4, Fig. 1a), which are both easily
available by Friedel–Crafts acylations.16,24 However, the
trisubstituted substrate 4 carrying most electron-withdrawing
groups is nearly not accepted by the CtXR. Its aromatic ring
seems to be too space consuming to fit into the binding side,
likely caused by the relative large bromine atoms. Quite
contrary is the acceptance of 2,4-dichloro acetophenone (2),
which is transformed by CtXR even three times faster than 2-
chloro acetophenone (1) and eight times faster compared to the
natural substrate xylose. The aromatic ring with two chlorine
atoms obviously fits well into the active site, and the summed
electronic effect of the halogen atoms in ortho- and para-
positions causes the fast reduction. The resulting (S)-1-
Obviously, two chlorine atoms in the ortho-positions as shown
in compound 11 hinder the substrate to bind in a productive
mode. Furthermore, space requirement and binding behavior
of the functionalized alkyl chains in the active site have varying
influences on the transformation. In all compounds with 2,4-
disubstituted aromatic moieties (8, 10), the alkyl chain seems
TABLE 1. Catalytic efficiency of CtXR catalyzed transforma-
tions of the investigated substrates
Substratea
Catalytic efficiency (molꢁ1 sꢁ1
)
1
340.3
1128.6
260.9
0.3
0.8
37.0
0.8
76.5
3.5
0.01
2
3
4
6
7
8
9
10
11
Scheme 1. Stereoselective CtXR catalyzed reduction of substituted
alkylphenones resulting in enantiomerical pure secondary alcohols. X = H,
electron-withdrawing groups.
aStructures are shown in Figure 1.
Chirality DOI 10.1002/chir

VOGL ET AL.
to compete against the larger aromatic ring for the limited space
hence only applicable in laboratory scale, but for synthetic
application in larger scale, the space/time yield is, however,
by far too low.26
in the active site. The electronic effect of the chlorine atoms is
hence limited by suboptimal substrate binding resulting in
reasonable bad catalytic activity as shown for compounds 8,
10, and 11 (Fig. 1b)
In another approach, we applied two-phase systems with a
water non-miscible organic solvent containing the substrate as
well as an aqueous buffer solution with the CtXR and the
coenzyme. The organic layer is a reservoir for substrate and
product, and the reaction occurs in aqueous medium. Phase
transfer of the substrate and product ensures a constant
substrate concentration at the enzyme.27 Distribution of 2,
4-dichloro acetophenone (2) between the two phases has been
tested with respect to maximal soluble concentrations in
both layers and the time to reach the distribution equilibrium.
Cyclohexane, hexane, ether, dichloromethane (DCM), toluene,
and ethyl acetate have been tested with varying concentrations
of 2. Having a concentration of 4.0 M 2 in the organic layer,
it takes ca. 12h to reach the distribution equilibrium. The
reduction would hence be controlled by phase transfer
leading to very long reaction times. Apart from the resulting
inadmissible low space/time yield, the CtXR lifetime is too low
for such application, because after 24h, the denaturation
leads to a noticeable loss of catalytic activity. Furthermore,
accumulated products might be distinctly subjected to
undesired consecutive reactions in such long reaction times.
A small amount of methanol added to the two-phase system
accelerates the phase transfer, and the equilibrium distribution
is reached within ca. 60min. Methanol acts as a connector
between the two phases similar to phase transfer catalysts.27,28
The CtXR catalytic activity and lifetime are, however, influ-
enced by the organic layer in some cases. In particular, DCM
and ethyl acetate lead to decomposition of the protein, and
toluene causes a fast loss of the enzymatic activity. The other
tested solvents do not cause these intensive effects. The small
amount of added methanol does also not notably influence the
enzyme activity and stability. In a solvent system of cyclohexane
(4ml), water (1ml), and methanol (0.5 ml), the time until
makeable loss of activity is recognized decreases to 18h.
Performing the reduction of 2,4-dichloro acetophenone (2)
using a 4 M solution in cyclohexane, hexane, or ethyl acetate
leads to the first occurrence of the (S)-1-(20,40-dichloro phenyl)
1-ethanol (5) after 60 min. After 6 h, up to 10 mmol of the
substrate has been transformed, and 5 is accumulated in the
respective organic layer (Table 3). The CtXR shows still a
reasonable activity at this time. When a 2 M concentration of
2 in the organic layer is used, it takes 3 h to detect small
amounts of the product in the organic layers. These data
indicate the phase transfer of one of the reaction rate controlling
We circumvent this problem by use of substrates with only
one or no chlorine at the aromatic ring, which are accepted with
a quite reasonable catalytic activity as listed for compounds 7
and 9 in Table 1. The transformation rates of these substrates
are up to 4.5 times smaller compared to those of model
compound 1.16 They are therefore still quite good substrates
for further synthetic applications. This consecutive reaction
might reduce the yield of the desired products of CtXR
reduction. The similar compound 6 is only barely accepted,
probably caused by its lack to act as hydrogen bridge acceptor
at the terminal end of the short side chain.
Increasing the solubility of substrates. The natural and
required environment of active CtXR is an aqueous buffer
solution containing the substrate and the cofactor. We use a
50 mM potassium phosphate buffer, which is a good solvent for
NADH and for the natural substrate xylose. However, the quite
non-polar acetophenones and phenyl-propanones are only soluble
in small concentrations below 0.1 mM, which makes a reasonable
scale-up process for a synthetic application by far impossible.
Addition of methanol to the aqueous buffer, however,
increases the amount of dissolved substrates. 3-Chloro-1-phenyl
1-propanone (7) can be solved to a concentration of about 5 mM
at a methanol extent of 20% v/v. Up to this extent, the organic
solvent does not influence the enzymatic activities and (S)-3-
chloro-1-phenyl 1-propanol (12) is gained with good transfor-
mation rates. Higher amounts of methanol, however, lead to a
distinct decrease of the transformation rate and a decomposi-
tion of the enzyme at ca. 30% v/v (Table 2). Such approach is
TABLE 2. Influence of the methanol amount in water on the
CtXR catalyzed reduction of 3-chloro-1-phenyl 1-propanone (7)
to the corresponding (S)-3-chloro-1-phenyl 1-propanol (12)
Conversion after
12 h (%)
Conversion after
14 h (%)
MeOH (v/v)
5
84
75
90
80
17
0
90
87
97
86
40
0
10
15
20
30
40
TABLE 3. Time and concentration depending on reduction of 2,4-dichloro acetophenone (2) to (S)-1-(20,40-dichloro phenyl) 1-eth-
anol (5) in two-phase system (water/organic solvent 1:1, 4.4 ml [v/v]) including 10 vol% methanol as solvent mediator; 0.55 units a
CtXR and 10 M NADH are in the aq. layer
Cylcohexane
2 Mb
Hexane
Ethyl acetate
2 Mb
Toluene
4 Mb
2 Mb
4 Mb
4 Mb
2 Mb
4 Mb
30 min
1 h
3 h
6 h
48 h
<0.5c
<0.5
3.2
3.6
4.0
<0.5
3.2
4.4
6.4
8.8
<0.5
<0.5
<0.5
<0.5
1.2
<0.5
2.4
2.8
4.0
4.4
<0.5
<0.5
4.0
5.6
6.0
<0.5
4.8
5.2
10.8
12.0
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
<0.5
aOne unit CtXR converts 3.0 mM of (2) in 60 s.
bConcentration (M) of (2) in the organic layer.
cConcentration (mM) of (5) in the organic layer.
Chirality DOI 10.1002/chir

Candida tenuis XYLOSE REDUCTASE CATALYZED REDUCTION OF ARYL KETONES FOR ENANTIOSELECTIVE
SYNTHESIS OF ACTIVE PHARMACEUTICAL INGREDIENTS
parameters, which, however, allows an application of this
method to gain larger amounts of the products.
(7) is easily available in a Friedel–Crafts acylation of benzene
with 3-chloro propanoyl chloride.24 After its CtXR catalyzed
reduction to 12, the resulting enantiomeric pure (S)-alcohol
is transferred to the chiral (R)-ether by an Sn2-type reaction
with inversion of the configuration. For that purpose,
diethylazodicarboxylate is used to etherify the alcohol with
o-cresol in a Mitsunobu-type reaction. The resulting (R)-1-(30-
chloro-10-phenyl propoxy)-2-methyl benzene is gained in
good yield29,30 and directly applied in the final direct amina-
tion with methylamine to gain the desired (R)-atomoxetine
(R-15).31 This short synthesis allows to gain the pharma-
ceutically active compound in good yields and very high
enantiomeric excess (Fig. 3).
Use of key intermediates for synthesis of pharmaceutical
drugs. The enantiomerically pure secondary alcohols result-
ing from CtXR catalyzed transformation (Scheme 1) can be used
for a chiral synthesis of several pharmaceutical active com-
pounds. This approach allows to circumvent the use of unde-
sired metal catalysts for enantioselective synthesis as well as
kinetic resolution steps, which lead to loss of material. For
example, (S)-1-(20-chloro phenyl) 1-ethanol (13) has recently
been used as chiral key intermediate for the synthesis of
thiophene-benzimidazole and thiophene-imidazopyridine deri-
vatives, which are promising serine/threonine kinase PLK1
inhibitor candidates.20 PLK1 is a regulator of mitotic progression
and cell division in eukaryotes and has a good antitumor
efficacy.23 Therefore, the CtXR catalyzed reduction of 2-chloro
acetophenone (1) to (S)-1-(20-chloro phenyl) 1-ethanol (13)
can help as an additional step in the fight against cancer.
Another potential use of some enantiomerically pure
secondary alcohols is the synthesis of oxetines, which are
selective serotonin and norepinephrine reuptake inhibitors.
They are used as potent antidepressants in several pharmaceu-
ticals.21,29 Most commonly used members of this group are
N-methyl-3-(40-trifluoromethyl phenoxy)-3-phenyl propanamine
(fluoxetine, 14) and N-methyl-3-(20-methyl phenoxy)-3-
phenylpropanamine (atomoxetine, 15, Fig. 2). Several synthetic
methods to gain these active agents as racemic mixtures and
enantiomerical pure compounds are known.21,29 However, only
atomoxetine is sold as pure enantiomer in (R)-configuration,
while fluoxetine is disposed as racemic mixture.
Further synthetic approaches for oxetines. Based on this
(R)-atomoxetine (R-15) synthesis and good availability of the
further applicable aceto- and propiophenones, a couple of other
short synthetic routes to oxetine derivatives are possible.
In particular, the other enantiomer of atomoxetine (15) might
be also available from 12 by an etherification in a coupling
with 2-chloro toluene using NaH as base.31,32 This Williamson
ether type reaction leads to complete retention of the chiral
center. Final methylation with methylamine provides the
(S)-atomoxetine (Fig. 4).
These two ways can be applied for the synthesis of all
oxetines using the respective phenols and chlorobenzenes.
Both enantiomers of the widely used fluoxetine (14)31,32 as
well as of norfluoxetine and nisoxetine are, hence, also avail-
able in four-step syntheses. The ketones 8–10 bearing other
functionalities in the alkyl chain are now also in discussion for
possible short oxetine syntheses. All four compounds can be
transferred to the (S)-alcohols CtXR catalyzed reduction
(Table 1; Fig. 2). Mitsunobu-type reaction and Williamson ether
synthesis can again be used to generate both enantiomers,
respectively. In case of using compound 8 as starting material,
which can be reduced to the (S)-alcohol, the desired products
are directly available by the formation of the ethers. Using
compounds 9 or 10 reduction lead to (S)-3-(20-chloro phenyl)-3-
hydroxy propionitrile or (S)-3-(20,40-dichloro phenyl)-3-hydroxy
propionitrile, respectively. After etherification, the cyanides have
to be reduced, protected by a tert-butyloxycarbonyl group,
methylated, and deprotected (Fig. 4). These three- and six-
step syntheses are also valuable alternatives to generate
oxetines enantioselectively in larger amounts for pharmaceutical
use.29–34
We exemplarily use (S)-3-chloro-1-phenyl 1-propanol (12) for
short and efficient enantioselective syntheses of (R)-atomoxetine
(R-15) (Fig. 3). The substrate 3-chloro-1-phenyl 1-propanone
Fig. 2. Structure of racemic fluoxetine (14) and (R)-atomoxetine (R-15).
Fig. 3. Four-step synthesis of (R)-atomoxetine (R-15). In the first step, a Friedel–Crafts acylation leads to the basic carbon skeleton in 7. In the second step, CtXR
catalyzed reduction forms the (S)-3-chloro-1-phenylpropan-1-ol (12) in high enantiomeric excess (ee > 98%). Then a Mitsunobu-type reaction with o-cresol leads to
etherification under inversion of the chiral center. Finally, a direct amination forms the desired (R)-atomoxetine (R-15).
Chirality DOI 10.1002/chir

VOGL ET AL.
Fig. 4. Overview of synthetic routes for enantioselective oxetine synthesis using enantioselective CtXR catalyzed reduction. The well and inexpensive available
ketones are reduced in a first step. Then etherification with diethylazodicarboxylat and a substituted phenol cause inversion of the stereocenter and lead to the
(R) configured series (i). Alternatively, an ether coupling with a substituted chlorophenol via NaH used as base gains the (S) configured analogs (ii). Apart from
direct amination of the chloro alkyls (R2 = Cl) shown in Figure 3, the presence of a methyl amine; R2 = N–CH3) directly leads to the products (iii). Alternatively,
the nitrile is reduced, protected with a tert-butyloxycarbonyl group, methylated, and deprotected (iv) to gain the final products.
6. Eames J. Synthesis of chiral alkanols by resolution and inversion meth-
ods. Sci Synth 2007;36:341–421.
CONCLUSIONS
Candida tenuis xylose reductase (CtXR) is a highly potent
7. But TYS, Toy PH. The Mitsunobu reaction: origin, mechanism, improve-
enzyme with a large substrate acceptance. In particular, several
acetophenone and phenyl propanone derivatives are reduced
with reasonable to excellent catalytic efficiencies. Therefore,
it is applicable to produce several enantiomerically pure key
intermediates for pharmaceutical drug synthesis. Low water
solubility of the substrates can be circumvented by addition of
up to 30% (v/v) methanol, which does not influence the enzyme
activity. Alternatively, a two-phase system with an organic
phase serving as reservoir for the organic substrates and
products can be used. For an acceleration of phase transfer,
again methanol can be used. This method does also not
decrease the CtXR lifetime and allows to reduce the substrates
in a sensible space/time yield. Resulting chiral products can be
used as intermediates in short syntheses to gain pharmaceutical
active products like oxetines or thiophene-benzimidazoles and
thiophene-imidazopyridines.
ments, and applications. Chem Asian J 2007;2:1340–1355.
8. May SW. Applications of oxidoreductases. Curr Opin Biotechnol
1999;10:370–375.
9. Devaux-Basseguy R, Bergel A, Comtat M. Potential applications of NAD
(P)-dependent oxidoreductases in synthesis: a survey. Enzym Microb
Tech 1997;20:248–258.
10. Hahn V, Mikolasch A, Wende K, Bartrow H, Lindequist U, Schauer F.
Synthesis of model morpholine derivatives with biological activities by
laccase-catalysed reactions. Biotechnol Appl Biochem 2009;54:187–195.
11. Hoelsch K, Weuster-Botz D. New oxidoreductases from cyanobacteria:
exploring nature’s diversity. Enzym Microb Tech 2010;47:228–235.
12. Mu XQ, Xu Y, Yang M, Sun ZH. New kinetic and thermodynamic
approach of one-pot synthesis of chiral alcohols by oxidoreductase from
Candida parapsilosis. Process Biochem 2011;46:233–239.
13. Grunwald J, Wirz B, Scollar PM, Klibanov MA. Asymmetric oxidoreductions
catalyzed by alcohol dehydrogenase in organic solvents. J Am Chem Soc
1986;108:6732–6734.
14. Rosche B, Breuer M, Hauer B, Rogers PL. Biphasic aqueous/organic
biotransformation of acetaldehyde and benzaldehyde by Zymomonas
mobilis pyruvate decarboxylase. Biotechnol Bioeng 2004;86:788–798.
ACKNOWLEDGMENT
15. Kratzer R, Nidetzky B. Identification of Candida tenuis xylose reductase
as highly selective biocatalyst for the synthesis of aromatic a-hydroxy
esters and improvement of its efficiency by protein engineering. Chem
Commun 2007;10:1047–1049.
We thank Susanne Felsinger for the kind help with the
measurement of NMR spectra.
16. Vogl M, Kratzer R, Nidetzky B, Brecker L. Candida tenuis xylose reductase
catalyzed reduction of acetophenones: the effect of ring-substituents on
catalytic efficiency. Org Biomol Chem 2011;9:5863–5870.
LITERATURE CITED
1. Stymiest LJ, Bagutski V, French MR, Aggarwal KV. Enantiodivergent
conversion of chiral secondary alcohols into tertiary alcohols. Nature
2008;456:778–782.
17. Mayr P, Brueggler K, Kulbe KD, Nidetzky B. D-Xylose metabolism by
Candida intermedia: isolation and characterisation of two forms of aldose reduc-
tase with different coenzyme specificities. J Chromatogr B 2000;737:195–202.
2. Nakamura K, Yamanaka R, Matsuda T, Harada T. Recent developments in
asymmetric reduction of ketones with biocatalysts. Tetrahedron:
Asymmetry 2003;14:2659–2681.
18. Kratzer R, Leitgeb S, Wilson KD, Nidetzky B. Probing the substrate binding
site of Candida tenuis xylose reductase (AKR2B5) with side-directed
mutagenesis. Biochem J 2006;393:51–58.
3. Stewart JD. Dehydrogenases and transaminases in asymmetric synthesis.
Curr Chem Biol 2001;5:120–129.
4. Sih CJ, Wu SH. Resolution of enantiomers via biocatalysis. Top Stereochem
19. Kratzer R, Pukl M, Egger S, Vogl M, Brecker L, Nidetzky B. Enzyme iden-
tification and development of a whole-cell biotransformation for asymmetric
reduction of o-chloroacetophenone. Biotechnol Bioeng 2011;108:797–803.
1989;19:65–125.
20. Sato Y, Onozaki Y, Sugimoto T, Kurihara H, Kamijo K, Kadowaki C,
Tsujino T, Watanabe A, Otsuki S, Mitsuya M, Iida M, Haze K, Machida
T, Nakatsuru Y, Komatani H, Kotani H, Iwasawa Y. Imidazopyridine
derivatives as potent and selective polo-like kinase (PLK) inhibitors.
Bioorg Med Chem Lett 2009;19:4673–4678.
5. Strümper D, Gogarten W, Durieux EM, Hartleb K, Van Aken H, Marcus
AEM. The effects of S(+)-ketamine and racemic ketamine on uterine
blood flow in chronically instrumented pregnant sheep. Anesth Analg
2004;98:497–502.
Chirality DOI 10.1002/chir

Candida tenuis XYLOSE REDUCTASE CATALYZED REDUCTION OF ARYL KETONES FOR ENANTIOSELECTIVE
SYNTHESIS OF ACTIVE PHARMACEUTICAL INGREDIENTS
21. Castelli E, Lo Monaco G, Mantovani S, Daverio P, Riva P, Vailati A,
30. Srebnik M, Ramachandran PV, Brown CH. Chiral synthesis via
Bianchi S. Process for the preparation of atomoxetine hydrochloride.
U.S. Patent No. 4,018,895.
22. Neuhauser W, Haltrich D, Kulbe KD, Nidetzky B. NAD(P)H-dependent
aldose reductase from the xylose-assimilating yeast Candida tenuis.
Biochem J 1997;326:683–692.
organoboranes. 18 Selective reductions. 43. Diisopinocampheylchloroborane
as an excellent chiral reducing reagent for the synthesis of halo alcohols of
high enantiomeric purity. A highly enantioselective synthesis of both optical
isomers of tomoxetine, fluoxetine, and nisoxetine.
1988;53:2916–2920.
J Org Chem
23. Molloy B, Schmiegel K. Arloxyphenylpropylamines. U.S. Patent No. 4,314,081.
24. Judefind WL, Reid EE. The identification of acids v para halogen phenacyl
esters. J Am Chem Soc 1920;42:1043–1055.
31. Li Y, Li Z, Li F, Wang Q, Tao F. Preparation of polymer-supported
Ru-TsDPEN catalysts and use for enantioselective synthesis of (S)-fluoxetine.
Org Biomol Chem 2005;3:2513–2518.
32. Mitchell D, Koenig MT. Synthesis of R- and S- fluoxetine, norfluoxetine
and related compounds from styrene oxide. Synthetic Comm
1995;25:1231–1238.
33. Kamal A, Khanna GBR, Ramu R. Chemoenzymatic synthesis of both
enantiomers of fluoxetine, tomoxetine and nisoxetine: lipase-catalyzed
resolution of 3-aryl-3-hydroxypropanenitriles. Tetrahedron-Asymmetry
2002;18:2039–2051.
34. Liu D, Gao W, Wang C, Zhang X. Practical synthesis of enantio-
pure g-amino alcoholes by rhodium-catalyzed asymmetric hydroge-
nation of b-secondary-amino ketones. Angew Chem Int Ed
2005;44:1687–1689.
25. Takahashi T, Sakuraba A, Hirobashi T, Shibata T, Hirose M, Haga Y,
Nonoshita K, Kanno T, Ito J, Iwaasa H, Kanatani A, Fukami T, Sato N.
Novel potent neuropeptide Y-Y5 receptor antagonists: synthesis and
structure-activity relationships of phenylpiperazine derivatives. Bioorg
Med Chem 2006;14:7501–7511.
26. Serebryakova TL, Zorin AN, Karyakin AA. Improvement of hydrogenase
enzyme activity by water-miscible organic solvents. Enzyme Microb Tech
2009;44:329–333.
27. Makosza M. Phase-transfer catalysis. A general green methodology in
organic synthesis. Pure Appl Chem 2000;72:1399–1403.
28. Yadav DG, Insight into green phase transfer catalysis. Top Catal
2004;29:145–161.
35. Kratzer R, Egger S, Nidetzky B. Whole-cell bioreduction of aromatic
a-keto esters using Candida tenuis xylose reductase and Candida boidinii
formate dehydrogenase co-expressed in Escherichia coli. Biotechnol
Bioeng 2008;110:1094–1101.
29. Liu H-L, Hoff BH, Anthonsen T. Chemoenzymatic synthesis of the
non-tricyclic antidepressants fluoxetine, tomoxetine and nisoxetine. J Chem
Soc, Perkin Trans 1 2000;11:1767–1769.
Chirality DOI 10.1002/chir