Reduction of acetophenones with methyl fluorines and a bulky group on the aromatic ring using microorganisms and related enzymes

Article abstract of DOI:10.1016/j.molcatb.2012.06.010

Whole-cell yeasts and mold-catalyzed reduction of two fluorinated acetophenone derivatives with very bulky substituents on ortho position of aromatic ring, (±)-1′-(2-tert-butyl-2-methyl-1,3-benzodioxol-4-yl)- 2′,2′-difluoroethanone and (±)-1′-(2-tert-butyl-2- methyl-1,3-benzodioxol-4-yl)-2′,2′,2′-trifluoroethanone were examined. On the former substrate, Geotrichum candidum NBRC 5767 showed high re-facially selective attack of hydride, while with Pichia angusta JCM 3620, complementary si-facially selective attack proceeded. G. candidum NBRC 5767 was revealed to be potent biocatalyst which provides (1′S)-alcohols from both substrates in a highly facially selective manner. Some unknown reductases were suggested responsible for those reductions, other than so far having been reported acetophenone reductase and trifluoromethyl ketone reductase from G. candidum, comparing the results obtained by applying those enzymes.

Full text of DOI:10.1016/j.molcatb.2012.06.010

Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Reduction of acetophenones with methyl fluorines and a bulky group on the
aromatic ring using microorganisms and related enzymes
Chika Abea, Takahiro Sugawaraa, Takuya Machidaa, Toshinori Higashia, Kengo Hanayaa,
Mitsuru Shojia, Chen Caob, Takuro Yamamotob, Tomoko Matsudab, Takeshi Sugaia,∗
a Department of Pharmaceutical Science, Keio University, 1-5-30, Shibakoen, Minato-ku, Tokyo 105-8512, Japan
b Department of Bioengineering, Tokyo Institute of Technology, 4259, Nagatsuta-cho, Midori-ku, Yokohama 226-8501, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 21 May 2012
Received in revised form 9 June 2012
Accepted 11 June 2012
Available online 19 June 2012
Whole-cell yeasts and mold-catalyzed reduction of two fluorinated acetophenone derivatives
with very bulky substituents on ortho position of aromatic ring,
(
)-1ꢀ-(2-tert-butyl-2-methyl-
1,3-benzodioxol-4-yl)-2ꢀ,2ꢀ-difluoroethanone and ( )-1ꢀ-(2-tert-butyl-2-methyl-1,3-benzodioxol-4-yl)-
2ꢀ,2ꢀ,2ꢀ-trifluoroethanone were examined. On the former substrate, Geotrichum candidum NBRC 5767
showed high re-facially selective attack of hydride, while with Pichia angusta JCM 3620, complemen-
tary si-facially selective attack proceeded. G. candidum NBRC 5767 was revealed to be potent biocatalyst
which provides (1ꢀS)-alcohols from both substrates in a highly facially selective manner. Some unknown
reductases were suggested responsible for those reductions, other than so far having been reported ace-
tophenone reductase and trifluoromethyl ketone reductase from G. candidum, comparing the results
obtained by applying those enzymes.
Keywords:
Asymmetric reduction
Yeast
Fungi
Fluoroketone
Isolated enzymes
Stereochemistry
© 2012 Elsevier B.V. All rights reserved.
1. Introduction
(2R,1ꢀS)-10a. Based on the results from ketones 1, 3, and 5, the effect
of the fluorine-containing substituents in 9 and 11 on reactivity and
Asymmetric reduction of ketones with cultured whole-cell bio-
catalysts, such as strains of Geotrichum candidum (NBRC 4597
and 5767), have been studied intensively [1–3]. Their activity and
enantioselectivity in asymmetric reduction are extremely vari-
able. For example, G. candidum NBRC 4597 reduces acetophenone
(1) and 2ꢀ-fluoroacetophenone (3) to yield (S)-2 and (R)-4 in
the same enantiofacial selectivity. The enantiofacial preference
changes, with increasing the number of fluorine atoms on the
methyl groups. When using difluoro derivative 5 as the start-
ing material, the enantioselectivity in 6 is lost and, eventually, in
2ꢀ,2ꢀ,2ꢀ-trifluoroacetonphenone (7), the selectivity was completely
inversed to give (S)-8 [4,5].
When starting with ( )-9, which contains a unique acetal on the
aromatic ring, the reduction with G. candidum NBRC 4597 was very
slow. This result agreed well with studies showing that the intro-
duction of a bulky substituent on the ortho position suppressed
reduction by this microorganism [6]. The enantiofacially selective
reduction, however, was proceeded by applying cultured cells of
strain G. candidum NBRC 5767 [7] to give mainly (2S,1ꢀS)- and
stereoselectivity was investigated. This report presents the differ-
ence in selectivity of the cultured whole-cell catalyzed reduction of
trifluoromethyl ketone 9 and difluoromethyl ketone 11 from eight
yeasts and fungi strains, including G. candidum. Scheme 1.
2. Experimental
IR spectra were measured as films for oils or KBr disks of solids
on a Jasco FT/IR-410 spectrometer, and as ATR on a Jeol FT-IR
SPX60 spectrometer. 1H NMR spectra were measured in CDCl3
at 270 MHz on a Jeol JNM EX-270 or at 400 MHz on an Agilent
400-MR spectrometer. 13C NMR spectra were measured in CDCl3
at 68 MHz on a Jeol JNM EX-270 or at 100 MHz on an Agilent
400-MR or at 125 MHz on an Agilent INOVA-500 spectrometer.
HPLC data were recorded on Jasco MD-2010 or SHIMADZU SPD-
20A multi-channel detectors. Merck silica gel 60 F254 thin-layer
plate (1.05715, 0.25 mm thickness) was used for thin-layer chro-
matographic analysis. Merck silica gel 60 F254 thin-layer plates
(1.05744, 0.5 mm thickness) and silica gel 60 (spherical and neu-
tral; 100–210 ␮m, 37560-79) from Kanto Chemical Co., Inc. were
used for preparative thin-layer chromatography and column chro-
matography, respectively. Yeast strains are available from Japan
Collection of Microorganisms; Riken Bioresource Center, Planning
∗
Corresponding author. Tel.: +81 3 5400 2665; fax: +81 3 5400 2665.
E-mail address: sugai-tk@pha.keio.ac.jp (T. Sugai).
1381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2012.06.010

C. Abe et al. / Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
87
Section, Research Promotion Division, RIKEN Tsukuba Institute,
3-1-1 Koyadai, Tsukuba, Ibaraki 305-0074, Japan, and to NITE Bio-
logical Resource Center; Department of Biotechnology, National
Institute of Technology and Evaluation, 2-5-8 Kazusakamatari,
Kisarazu, Chiba 292-0818, Japan. Peptone, malt extract and yeast
extract were purchased from Kyokuto Pharmaceutical Industrial
Co., Ltd.
O
OH
a
XH2C
XH2C
(S)-2: X = H
1: X = H
3:
(R)-4:
F
2.1. ( )-1ꢀ-(2-tert-Butyl-2-methyl-1,3-benzodioxol-4-yl)-2ꢀ,2ꢀ-
difluoroethanone
F
(11)
OH
In a similar way for the preparation of 9 [7], a solution of
lithiated form of 2-tert-butyl-2-methyl-1,3-benzodioxole (1.00 g,
5.2 mmol), which was made with n-butyllithium (2.6 M in hexane,
3.0 mL, 1.5 equiv.) and TMEDA (1.1 mL, 1.5 equiv.) was treated with
F2HCCO2Et (1.4 mL, 13 mmol, 2.5 equiv.) at −50 ◦C. Similar workup
[7] and the purification of the residue by silica gel column chro-
matography (100 g) by the elution with hexane/AcOEt = 40:1 to
20:1 provided ketone 11 (1.17 g, 83%) as yellow oil. 1H NMR: ı
F2HC
F2HC
F3C
O
a
(R)-6
F2HC
OH
5
7
1.08 (s, 9H, tert-Bu), 1.62 (s, 3H, Me), 6.48 (t, J2 ,F = 53.6 Hz, 1H, H2ꢀ),
ꢀ
6.85 (dd, J5,6 = 8.2 Hz, J6,7 = 7.6 Hz, 1H, H6), 6.94 (dd, J5,7 = 1.4 Hz,
1H, H7), 7.36 (dd, 1H, H5); 13C NMR: ı 20.2, 24.4, 39.6, 109.1 (t,
(S)-6
J2 ,F = 989.4 Hz, C2ꢀ), 111.6, 113.3, 120.4, 121.4, 126.0, 149.3, 149.3,
ꢀ
184.8 (t, J1 ,F = 96.8 Hz, C1ꢀ), the signals 24.4 included totally three
ꢀ
O
OH
carbons; IR: 2978, 1709, 1460, 1142, 854, 725 cm−1
.
a
F3C
2.2. (2R*,1ꢀR*)-( )-1ꢀ-(2-tert-Butyl-2-methyl-1,3-benzodioxol-4-
yl)-2ꢀ,2ꢀ-difluoroethanol (12a) and
(2R*,1ꢀS*)-( )-1ꢀ-(2-tert-butyl-2-methyl-1,3-benzodioxol-4-yl)-
2ꢀ,2ꢀ-difluoroethanol
(S)-8
(12a)
t-Bu
To a solution of 11 (15.9 mg, 0.059 mmol) in EtOH (590 ␮L) were
treated with NaBH4 (6.7 mg, 0.176 mmol, 3.0 equiv.) at room tem-
perature. The same workup and the purification of the residue
by preparative TLC [developed with hexane/AcOEt = 4:1] to afford
(2R*,1ꢀR*)-12a and (2R*,1ꢀS*)-12a (15.0 mg, 94%) as pale yellow oil
as a mixture. The diastereomeric ratio between (2R*,1ꢀS*)- and
(2R*,1ꢀR*)-12a was determined to be 3:2 judging from their NMR
spectrum: ı 5.88 for (2R*,1ꢀS*)- and ı 5.92 for (2R*,1ꢀR*)-12a. A small
portion was separated by a preparative HPLC [column, Kanto Chem-
ical Co., Inc. Mightysil 60, 1 cm × 25 cm; hexane/AcOEt = 15:1);
flow rate 5 mL/min]: tR (min) = 6.6 [(2R*,1ꢀR*)-12a], 6.8 [(2R*,1ꢀS*)-
12a] and the former was isolated in pure state. (2R*,1ꢀR*)-12a:
1H NMR: ı 1.05 (s, 9H, tert-Bu), 1.55 (s, 3H, Me), 4.86 (ddd,
CH3
2
O
O
a
O
10a
1'
F3C
( )-9
t-Bu
t-Bu
CH3
CH3
OH
OH
O
O
O
O
J1 ,2 = 4.1 Hz, J1 ,F = 9.8, 13.9 Hz, 1H, H1ꢀ), 5.92 (dt, J2 ,F = 55.8 Hz, 1H,
H2ꢀ), 6.70–6.83 (aromatic, 3H); 13C NMR: ı 20.0, 24.5, 39.4, 70.7
ꢀ
ꢀ
ꢀ
ꢀ
F3C
F3C
(t, J1 ,F = 99.8 Hz, C1ꢀ), 108.4, 115.1 (t, J2 ,F = 976.0 Hz, C2ꢀ), 116.5,
119.6, 121.2, 124.4, 145.8, 148.1, the signals 24.5 included totally
three carbons. (2R*,1ꢀS*)-12a: 1H NMR: ı 1.05 (s, 9H, tert-Bu),
ꢀ
ꢀ
(2S,1'S)-10a
t-Bu
(2S,1'R)-10a
t-Bu
1.55 (s, 3H, Me), 4.93 (dt, J1 ,2 = 4.3 Hz, J1 ,F = 10.6 Hz, 1H, H1ꢀ), 5.88
ꢀ
ꢀ
ꢀ
(dt, J2 ,F = 55.9 Hz, 1H, H2ꢀ), 6.70–6.83 (aromatic, 3H); 13C NMR:
ꢀ
CH3
CH3
ı 20.0, 24.4, 39.5, 69.8 (q, J1 ,F = 101.2 Hz, C1ꢀ), 108.3, 115.2 (t,
ꢀ
OH
OH
O
O
J2 ,F = 976.0 Hz, C2ꢀ), 116.5, 119.0, 121.2, 124.3, 145.8, 148.0, the
ꢀ
O
O
signals 24.4 included totally three carbons. HPLC [column, Daicel
Chiralcel OJ-H, 0.46 cm × 25 cm; hexane/i-PrOH = 40:1; flow rate
0.5 mL/min]: tR (min) = 18.3 [(2R,1ꢀR)-12a], 20.4 [(2S,1ꢀS)-12a], 22.4
[(2R,1ꢀS)-12a], 30.5 [(2S,1ꢀR)-12a]. The assignment of relative and
absolute configurations as above were described in detail, in Sec-
tions 2.7 and 2.8.
F3C
F3C
(2R,1'S)-10a
(2R,1'R)-10a
Scheme 1. Reagents and conditions: (a) Geotrichum candidum.
2.3. Screening of microorganisms for the reduction of 9 and 11
The microorganisms from stock culture samples were incubated
in glucose medium [8–10] [containing glucose (5.0 g), peptone

88
C. Abe et al. / Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
Table 1
Whole cell microorganism-catalyzed reduction of ( )-11 and ( )-9.
Entry
Microorganisms (JCM No.)
Xa
Conv.b (%)
Product ratioc
(2R,1ꢀR)
(2S,1ꢀS)
(2R,1ꢀS)
(2S,1ꢀR)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Pichia minuta (3622)
P. minuta (3622)
Trichosporon cutaneum (1534)
T. cutaneum (1534)
Candida floricola (9439)
C. floricola (9439)
Williopsis californica (3600)
W. californica (3600)
Candida nitratophila (9856)
C. nitratophila (9856)
Pichia angusta (3620)
P. angusta (3620)
H
F
H
F
H
F
H
F
H
F
H
F
H
F
H
F
75
80
66
76
22
66
89
19
11
24
23
0
0
10
28
31
9
52
20
4
40
47
1
17
9
31
14
20
14
22
0
8
23
67
41
17
16
27
38
43
35
58
–
11
65
26
50
12
31
13
2
27
9
30
12
30
40
–
13
12
1
–
–
Geotrichum candidum (4597)d
G. candidum (4597)d
G. candidum (5767)d
G. candidum (5767)d
4
3
86
91
14
13
48
49
25
33
50
50
48
42
1
0
1
a
For substituents, as shown in Scheme 2.
b
c
Determined by 1H NMR spectrum of the crude product, as shown in Section 2.3.
Determined by HPLC analysis, as shown in Sections 2.2 and 2.3.
NBRC No.
d
(2.0 g), yeast extract (0.5 g), KH2PO4 (0.3 g), K2HPO4 (0.2 g), at pH
6.5, total volume of 100 mL, in the Erlenmeyer cultivating flasks]
for 2 days at 30 ◦C. Pre-cultivation of G. candidum NBRC 4597 and
5767 were described later in Section 2.6.
2.5. Trifluoromethyl ketone reductase-catalyzed reduction of 9
and 11
Trifluoromethyl ketone reductase was partially purified as fol-
lows, according to the reported procedure [5]. Cell free extract of G.
candidum NBRC 5767 was firstly fractionated with (NH4)2SO4, and
the desired fraction was sequentially submitted to chromatogra-
phy with anion exchange, hydrophobic, hydroxyapatite, and finally
anion exchange column. To a mixture of the enzyme (0.4 units; one
unit of enzyme activity was defined as the micromoles of 8 released
by the reduction of 7 per minute) in HEPES buffer (0.1 M, pH, 7.2,
13.3 mL) were added NADPH (80 mg, 0.10 mmol) and ketone 9 or 11
(10 mg, ca. 0.04 mmol) dissolved in ethanol (0.70 mL). The mixture
was reciprocally shaken at 250 cpm and 40 ◦C for 20 h. The extrac-
tive work-up and evaluation of the products were as in the same
manner described in Section 2.4.
The wet cells in the broth (100 mL) were harvested by cen-
trifugation (3000 rpm) and 500 mg of cells were re-suspended in
phosphate buffer (0.1 M, pH 6.0, 5 mL) in a test tube, together with
a substrate 9 or 11 (25 mg) and glucose (250 mg), and shaken on
a reciprocal shaker (210 cpm) for 2 days at 30 ◦C. The extractive
work-up and the analysis of stereoisomers were performed in the
similar manner as described later in Section 2.6. The conversion
was estimated by the NMR spectra of the crude products. The ratio
of area for signals ı 5.88 [dt, 1H, H2ꢀ for (2R*,1ꢀS*)-12a, 5.92 [dt, 1H,
H2ꢀ for (2R*,1ꢀR*)-12a], and 6.48 (t, 1H, H2ꢀ for 11) were calculated.
Stereoisomeric ratio of 10a was determined HPLC analysis [column,
Daicel Chiralcel OJ-H, 0.46 cm × 25 cm; hexane/i-PrOH = 40:1; flow
rate 0.5 mL/min]: tR (min) = 13.2 [(2R,1ꢀR)-10a], 13.9 [(2R,1ꢀS)-10a],
14.9 [(2S,1ꢀS)-10a], 20.8 [(2S,1ꢀR)-10a] [7]. The results are summa-
rized in Table 1.
Reduction of 9: conversion: 7%, HPLC analysis of 10a: tR
(min) = 13.2 [5.3%, (2R,1ꢀR)-], 13.9 [48.9%, (2R,1ꢀS)-], 14.9 [8.7%,
(2S,1ꢀS)-], 20.8 [37.1%, (2S,1ꢀR)-].
Reduction of 11: conversion: 12%, HPLC analysis of 12a: tR
(min) = 18.3 [10.1%, (2R,1ꢀR)-], 20.4 [11.6%, (2S,1ꢀS)-], 22.4 [50.4%,
(2R,1ꢀS)-], 30.5 [27.9%, (2S,1ꢀR)-].
2.4. Acetophenone reductase-catalyzed reduction of 9 and 11
To a solution of the partially purified acetophenone reductase
from G. candidum NBRC 4597 [11] which contains 64 units (one unit
of enzyme activity was defined as the micromoles of 2 released by
the reduction of 1 per minute) in a HEPES buffer [0.1 M, pH, 7.2, con-
taining DTT (1 mM) and PMSF (1 mM), total 2.53 mL] were added
ketone 9 or 11 (20 mg, ca. 0.07 mmol), NAD+ (30 mg, 0.04 mmol),
and 2-propanol (0.45 mL, 15%, v/v). The mixture was reciprocally
shaken at 250 cpm and 40 ◦C for 20 h and saturated with NaCl.
After filtration with Celite, the filtrate was extracted three times
with AcOEt. The combined extract was washed with brine, dried
over Na2SO4, and concentrated in vacuo. The conversion was esti-
mated in the same manner as Section 2.3. The alcohols were further
purified by preparative TLC [developed with hexane/AcOEt = 4:1)].
Reduction of 9: conversion: 4%, HPLC analysis of 10a: tR
(min) = 13.2 [17.1%, (2R,1ꢀR)-], 13.9 [31.9%, (2R,1ꢀS)-], 14.9 [17.6%,
(2S,1ꢀS)-], 20.8 [33.4%, (2S,1ꢀR)-].
2.6. Whole cell-catalyzed reduction of ( )-11 with G. candidum
NBRC 5767
A loopful of the strain grown on the agar-plate culture was
aseptically inoculated to a glucose medium [glucose (20.0 g), pep-
tone (8.0 g), yeast extract (2.0 g), KH2PO4 (1.2 g), K2HPO4 (0.8 g),
at pH 6.2, total volume of 400 mL] in four 500-mL baffled Erlen-
meyer cultivating flasks. The flask was shaken on a gyratory shaker
(180 rpm) for 24 h at 30 ◦C. The wet cells were harvested by fil-
tration with a rough paper filter and washed with water. The
weight of combined wet cells was ca. 22.2 g from 400 mL of the
broth.
Above-mentioned harvested cells of G. candidum (15.0 g) were
re-suspended in phosphate buffer solution (0.2 M, 50 mL, pH 6.0),
together with a substrate 11 (0.5 g, 1.9 mmol) and glucose (7.5 g),
in a 200 mL round-bottomed flask which was loosely capped with
sterilized filter paper for a plenty of oxygen supply. The mixture was
stirred at 30 ◦C for 24 h. The mixture was centrifuged (3000 rpm) to
Reduction of 11: conversion: 3%, HPLC analysis of 12a: tR
(min) = 18.3 [83.7%, (2R,1ꢀR)-], 20.4 [0.7%, (2S,1ꢀS)-], 22.4 [1.8%,
(2R,1ꢀS)-], 30.5 [13.8%, (2S,1ꢀR)-].

C. Abe et al. / Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
89
remove cell mass and the supernatant was further filtered through
t-Bu
t-Bu
CH3
CH3
a pad of Celite. The filtrate was saturated with NaCl and extracted
with AcOEt. The combined extract was washed with brine, dried
over Na2SO4, and concentrated in vacuo. The residue was purified
by a silica gel column chromatography (30 g). Elution with hex-
ane/AcOEt = 20:1 to 5:1 afforded 12a (330.8 mg, 64%) as colorless
solid. HPLC analysis was performed under the same conditions in
Section 2.2. tR (min) = 18.3 [0.6%, (2R,1ꢀR)-12a], 20.4 [47.8%, (2S,1ꢀS)-
12a], 22.4 [50.3%, (2R,1ꢀS)-12a], 30.5 [1.3%, (2S,1ꢀR)-12a].
OH
O
O
O
a
O
O
F2XC
F2XC
12a: X = H
10a:
( )-11: X = H
( )-9:
F
F
2.7. MTPA esters of (2S,1ꢀS)-1ꢀ-(2-tert-butyl-2-methyl-1,3-
benzodioxol-4-yl)-2ꢀ,2ꢀ-difluoroethanol (12b and
12c)
A small portion of the product mixture in Section 2.3 was
separated by a preparative HPLC using Mightysil 60 under the
conditions in Section 2.2, and the diastereomerically pure (2S,1ꢀS)-
12a was obtained. HPLC analysis using Chiralcel OJ-H under the
conditions in Section 2.2: tR (min) = 18.3 [1.2%, (2R,1ꢀR)-12a], 20.4
[98.8%, (2S,1ꢀS)-12a]. A mixture of this purified alcohol (2.4 mg,
0.01 mmol), (R)-␣-methoxy-␣-trifluoromethylphenylacetyl chlo-
ride [(R)-MTPA chloride, 5.5 mg, 0.03 mmol], pyridine (100 ␮L), and
catalytic amount of DMAP was stirred at room temperature for
30 min. The reaction was quenched by the addition of water, and
the organic materials were extracted with AcOEt. The combined
organic layer was washed with water, hydrochloric acid (1 M), sat-
urated aqueous NaHCO3 solution and brine, dried over Na2SO4, and
concentrated in vacuo. The residue was purified by preparative TLC
[developed with hexane/AcOEt = 4:1)] to afford (S)-MTPA ester 12b
(2.5 mg, 58%) as colorless oil. 1H NMR: ı 1H NMR: ı 1.04 (s, 9H,
P. angusta
H3C
t-Bu
F
H
H
F
O
O
O
H
11
G. candidum
(NBRC 5767)
ꢀ
ꢀ
ꢀ
tert-Bu), 1.41 (s, 3H, Me), 6.13 (dt, J1 ,2 = 5.3 Hz, J2 ,F = 55.3 Hz, 1H,
H2ꢀ), 6.14 (ddd, J1 ,F = 5.9, 10.8 Hz, 1H, H1ꢀ), 6.61 (dd, J5,7 = 1.8 Hz,
ꢀ
J5,6 = 7.5 Hz, 1H, H5), 6.69 (dd, J6,7 = 7.5 Hz, 1H, H6), 6.72 (dd, 1H,
H7).
P. angusta
In the similar manner, (R)-MTPA ester 12c was obtained (2.2 mg,
61%) as colorless oil. 1H NMR: ı 1.05 (s, 9H, tert-Bu), 1.52 (s, 3H, Me),
(No reaction)
6.08 (dt, J1 ,2 = 5.3 Hz, J2 ,F = 55.0 Hz, 1H, H2ꢀ), 6.23 (ddd, J1 ,F = 7.6,
ꢀ
ꢀ
ꢀ
ꢀ
9.8 Hz, 1H, H1ꢀ), 6.70–6.83 (aromatic, 3H).
F
H
F
F
2.8. Authentic samples of the mixture of (2R,1ꢀR)-12a and
(2R,1ꢀS)-12a
O
According to the reported procedure [12], to slurry of magne-
sium turnings (4.2 mg, 0.173 mmol) in anhydrous THF (1.5 mL) was
added chlorotrimethylsilane (44 ␮L, 0.346 mmol) and the mixture
was stirred 5 min. Then a solution of (R)-9 [7] (50.0 mg, 0.173 mmol)
in anhydrous THF (2.2 mL) was added dropwise. After stirring for
2 h at 0 ◦C, the mixture was concentrated in vacuo. The residue
was mixed with hexane and filtered with filter paper. The filtrate
was concentrated in vacuo and the residue was dissolved in MeOH
(2.0 mL). To the solution was added tetra-n-butylammonium flu-
oride (1.0 M solution in THF, 173 ␮L, 0.173 mmol), immediately
followed by NaBH4 (13 mg, 0.346 mmol), and the mixture was
stirred at room temperature for 30 min. The mixture was cooled
with ice-water bath and the reaction was quenched by the addi-
tion of water and acetic acid, and the organic materials were
extracted with AcOEt. The combined organic layer was washed
with water, saturated aqueous NaHCO3 solution and brine, dried
over Na2SO4, and concentrated in vacuo. The residue was purified
by preparative TLC [developed with hexane/AcOEt = 4:1)] to afford
a mixture of (2R,1ꢀR)-12a and (2R,1ꢀS)-12a (22.3 mg, 48%) as color-
less oil. HPLC analysis using Chiralcel OJ-H under the conditions
in Section 2.2: tR (min) = 18.3 [37.8%, (2R,1ꢀR)-12a], 22.4 [62.2%,
(2R,1ꢀS)-12a].
O
H
O
t-Bu
CH3
G. candidum
(NBRC 5767)
9
Scheme 2. Reagents and conditions: (a) cultured whole-cell microoganisms.
3. Results and discussion
3.1. Stereochemical preference of whole-cell biocatalysts on 9
and 11
Results for reduction of ( )-9 and ( )-11 to obtain 10a and
12a, respectively, by the cultured whole-cell biocatalyst are
summarized in Table 1, and characteristic examples are shown in
Scheme 2. For difluoromethyl ketone 11, Pichia angusta JCM 3620
and G. candidum NBRC 5767 showed complementary enantiofacial

90
C. Abe et al. / Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
preference (Table 1, entries 11 and 15) as shown in Scheme 2. G. can-
didum NBRC 5767 showed similar reactivity and selectivity toward
both substrates 9 and 11 (entries 15 and 16), while P. angusta JCM
3620 reduced 9 to a negligible extent (entry 12). The contrast in
behavior between difluoromethyl (11) and trifluoromethyl ketone
(9) with P. angusta (entries 11 and 12) is interesting. The downfield
H2ꢀ chemical shift of 11 (ı: 6.48) compared with that in a simple
non-substituted 2ꢀ,2ꢀ-difluoroacetophenone (ı: 6.29) was exam-
ined. This shift suggests hydrogen bonding between the acetal
oxygen atom and proton in difluoromethyl group to give a very
compact conformation of 11, which would be advantageous for a
carbonyl reductase in P. angusta. In contrast, the trifluoromethyl
group of 9 can rotate freely between CF3−CO, and the CF3 group is
much bulkier, which is a disadvantage to the reductase (Scheme 2).
t-Bu
t-Bu
CH3
CH3
CH3
CH3
2
OH
O
O
O
1'
O
a
O
F2HC
F2HC
F2HC
F2HC
F2HC
F2HC
12a
( )-11
t-Bu
t-Bu
CH3
OH
OH
O
O
O
O
3.2. Reduction of 9 and 11 with enzymes involved in G. candidum
(2S,1'S)-12a
t-Bu
(2S,1'R)-12a
t-Bu
As shown in the entries 13–16 in Table 1, although G. candidum
NBRC 4597 and 5767 are the same species, these possess differ-
ing reactivity to ketones 9 and 11. Two types of enzymes have been
reported [5,11,13] from these strains. The enzyme from G. candidum
NBRC 4597 is an “acetophenone reductase [11],” which is responsi-
ble for the production of (S)-2 from 1. In contrast, the enzyme from
G. candidum NBRC 5767 is a “trifluoromethyl ketone reductase [5],”
which reduces 7 to (S)-8. We became interested in the activity of
the above two reductases, toward 9 and 11.
CH3
OH
OH
O
O
O
O
(2R,1'S)-12a
(2R,1'R)-12a
When using 11 as the substrate, kinetic resolution of the remote
chiral center on the acetal (C2) with acetophenone reductase-
catalyzed reduction proceeded with a preference for (R)-11,
resulting in generation of (2R,1ꢀR)-12a (83.7% of the four stereoiso-
mers, as shown in Section 2.4) via si-facial attack of the hydride
on the carbonyl group. Although the activity of acetophenone
reductase was very low toward both substrates 9 and 11 (4% and
3% conversion, respectively), it contributed to the formation of
(2R,1ꢀR)- and (2S,1ꢀR)-12a from 11 in the whole cell-mediated G.
candidum NBRC 4597 reduction with (entry 13 in Table 1).
For trifluoromethyl ketone reductase, the activity increased
slightly compared with the acetophenone reductase, but was still
low (7% conversion for 9 and 12% for 11, respectively). The products
(2R,1ꢀS)-12a (50.4% of the four stereoisomers, as shown in Section
2.5) from 11 and (2R,1ꢀS)-10a (48.9% of the four stereoisomers) from
9 were obtained through re-facial selectivity. As mentioned ear-
lier, the major products in the whole cell-mediated G. candidum
reduction, however, were almost exclusively the stereoisomers
(2R,1ꢀS) and (2S,1ꢀS) (entries 15 and 16 in Table 1). Upon appli-
cation of trifluoromethyl ketone reductase, other diastereomers
[(2S,1ꢀR)-12a (27.9%) and (2S,1ꢀR)-10a (37.1%)] were formed with
the reverse enantiofacial selectivity. These results suggest that
unknown reductases exist that are responsible for efficient enan-
tiofacially selective reduction of 9 and 11 by G. candidum NBRC
5767.
b, c, d
(S)-MTPA ester 12b
(R)-MTPA ester 12c
( )-11
( )-11
b, c, e
t-Bu
CH3
O
O
HO2C
f
(2S,1'S)-12a
(S)-11
(S)-13
t-Bu
CH3
O
O
O
g, h, i
(2R,1'S)-12a
3.3. Stereochemical assignment of four stereoisomers of 12a
F3C
and
Because the alcohol 12a had not been described previously, the
stereochemistry in its two chiral centers was determined as shown
below, including the Kusumi–Mosher method, for the evidence of
(2R,1'R)-12a
(R)-9
Scheme 3. Reagents and conditions: (a) NaBH4, EtOH, quant; (b) Geotrichum can-
didum NBRC 5767, 64%; (c) chromatographic separation of diastereomers; (d)
(R)-MTPA chloride, DMAP, pyridine; (e) (S)-MTPA chloride, DMAP, pyridine; (f) 1-
Me-AZADO, NaOCl, KBr, CH2Cl2, aq. NaHCO3, quant; (g) Mg, TMSCl, THF; (h) TBAF,
MeOH; and (i) NaBH4, MeOH, 48%.
Table 2
Kusumi–Mosher 1H NMR calculation on MTPA esters.
Proton
ꢀı H [12c–b]
ꢀı H [10c–b] [7]
H5
H6
H7
H2ꢀ
+0.16
+0.08
+0.05
−0.05
+0.23
+0.13
+0.06
–

C. Abe et al. / Journal of Molecular Catalysis B: Enzymatic 82 (2012) 86–91
91
the discussion in Sections 3.1 and 3.2. First, the non-stereoselective
reduction of racemic difluoromethyl ketone ( )-11 with NaBH4
was performed to give four possible stereoisomers of 12a, which
produced four independent peaks on HPLC with a Chiralcel OJ-H
column.
catalytic activity and stereoselectivity. New enzymes are likely to
be isolated from P. angusta JCM 3620 and G. candidum NBRC 5767,
which demonstrated highly selective si-facial and re-facial attacks
of hydrides, respectively, on newly designed substrates. The advan-
tages of these methods, using whole-cell biocatalysts (regeneration
of the reduced form of coenzymes, enzymes immobilized in cell
materials, use of biomass as a sink for substrates and products) or
isolated enzymes (high concentration of enzymes, clean products),
can contribute to the large-scale synthesis of enantiomerically
enriched compounds.
Reduction with G. candidum NBRC 5767 afforded two main iso-
mers of 12a, which were related to the second and third peaks
in the elution order among the four peaks observed using the
Chiralcel OJ-H column. The products were diastereomers, as deter-
mined by NMR. The less polar isomer related to the second peak
was isolated, and derived to the corresponding (S)- and (R)-␣-
methoxy-␣-trifluoromethylphenylacetate (MTPA) esters 12b and
12c, respectively. Upfield signals for H5-7 or aromatic ring and
downfield signal for H2ꢀ of difluoromethyl group in the NMR spectra
of (S)-12b compared to (R)-12c were observed. The two aromatic
rings proximate to each other and a difluoromethyl group is sep-
arated from the MTPA aromatic ring in 12b indicated that the
absolute configuration of secondary alcohol in 12a (C1ꢀ) obtained by
G. candidum-mediated reduction was (S). The differences in chem-
ical shifts together with (S)-10b and (R)-10c [7] are summarized in
Table 2. Thus, the mixture of (2R,1ꢀS)- and (2S,1ꢀS)-12a was revealed
to be the second and third peaks, respectively, on HPLC (Scheme 3).
Determination of the absolute configuration of C-2 of (1ꢀS)-12a
was first attempted, by converting it to the (S)-TBMB carboxylic
acid 13 with known absolute configuration [7] via oxidation of the
secondary alcohol with 1-Me-AZADO [14] and subsequent halo-
form reaction of the CHF2−CO moiety in (S)-11. Enolate formation
prior to halogenation of the difluoromethyl ketone under basic
conditions, however, was very slow. Alternatively, an authentic
sample of trifluoromethyl ketone (R)-9 [7] was treated under reduc-
tive defluorination conditions [12]. Deprotection of the resulting
enol trimethylsilyl ether and subsequent reduction of the result-
ing ketone with NaBH4 provided an authentic mixture of (2R,1ꢀR)-
and (2R,1ꢀS)-12a. This mixture contained the first and third peaks
from the HPLC analysis. Combination of all of this information
allowed unambiguous determination of the retention times of the
four stereoisomers as follows: tR (min) = 18.3 [(2R,1ꢀR)-12a], 20.4
[(2S,1ꢀS)-12a], 22.4 [(2R,1ꢀS)-12a], 30.5 [(2S,1ꢀR)-12a]. The enan-
tiomeric relation between the isomers of the first and second peaks
was further supported by HPLC and NMR analyses of the pure
racemic sample [(2R*,1ꢀR*)-12a], which was obtained from the four
stereoisomers.
Acknowledgments
The authors thank Professors Kaoru Nakamura of Kyoto Univer-
sity, Tohru Yamada of Keio University, Tadashi Ema of Okayama
University, and Dr. Rio Yamanaka of Himeji Dokkyo University for
discussion on the early phase of this research, especially asym-
metric reduction of trifluoromethyl ketone 9. We also thank Dr.
Masaki Handa of Sagami Chemical Research Institute, for his
valuable suggestion on the defluorination of 9. This work was
supported both by a Grant-in-Aid for Scientific Research (No.
23580152) and formation of a research center in cell signal-
ing drug discovery for molecular targeting therapies: matching
fund subsidy 2009–2013 from the Ministry of Education, Culture,
Sports, Science and Technology, Japan, and acknowledged with
thanks.
References
[1] K. Nakamura, R. Yamanaka, T. Matsuda, T. Harada, Tetrahedron: Asymmetry 14
(2003) 2659.
[2] K. Nakamura, T. Matsuda, Curr. Org. Chem. 10 (2006) 1217.
[3] T. Matsuda, R. Yamanaka, K. Nakamura, Tetrahedron: Asymmetry 20 (2009)
513.
[4] K. Nakamura, T. Matsuda, T. Itoh, A. Ohno, Tetrahedron Lett. 37 (1996) 5727.
[5] T. Matsuda, T. Harada, N. Nakajima, T. Itoh, K. Nakamura, J. Org. Chem. 65 (2000)
157.
[6] K. Nakamura, K. Takenaka, M. Fujii, Y. Ida, Tetrahedron Lett. 43 (2002) 3629.
[7] T. Higashi, C. Abe, K. Ninomiya, T. Machida, N. Chishima, S. Taketomi, M. Furuta,
Y. Komaki, Y. Senba, T. Tokuda, M. Shoji, T. Sugai, Adv. Synth. Catal. 352 (2010)
2549.
[8] M. Sakamoto, M. Hamada, T. Higashi, M. Shoji, T. Sugai, J. Mol. Catal. B: Enzym.
64 (2010) 96.
[9] M. Furuta, M. Shoji, T. Sugai, J. Mol. Catal. B: Enzym. 82 (2012) 8.
[10] S. Taketomi, M. Asano, T. Higashi, M. Shoji, T. Sugai, J. Mol. Catal. B: Enzym.,
http://dx.doi.org/10.1016/j.molcatb.2012.01.020, in press.
[11] Y. Nakata, T. Fukae, R. Kanamori, S. Kanamaru, T. Matsuda, Appl. Microbiol.
Biotechnol. 86 (2010) 625.
[12] G.K.S. Prakash, J. Hu, G.A. Olah, J. Fluorine Chem. 112 (2011) 357.
[13] T. Harada, Y. Kubota, T. Kamitanaka, K. Nakamura, T. Matsuda, Tetrahedron Lett.
50 (2009) 4934.
4. Conclusion
Unique biocatalysts consisting of cultivated yeasts and molds
reduced 2ꢀ,2ꢀ,2ꢀ-trifluoro- and 2ꢀ,2ꢀ-difluoroacetophenone deriva-
tives having very bulky ortho-substituents (9 and 11) with high
[14] M. Shibuya, M. Tomizawa, I. Suzuki, Y. Iwabuchi, J. Am. Chem. Soc. 128 (2006)
8412.