Chemo-enzymatic route for (R)-terbutaline hydrochloride based on microbial asymmetric reduction of a substituted α-chloroacetophenone derivative

Article abstract of DOI:10.1016/j.molcatb.2012.01.020

To synthesize (R)-terbutaline hydrochloride, a potent β₂- adrenoceptor-stimulating agent, asymmetric reduction of a substituted α-chloroacetophenone derivative with cultured whole-cell biocatalyst of the yeast Williopsis californica JCM 3600 was developed as the key reaction. The reduction proceeded by a si-facial attack of hydride in a highly enantioselective manner. Co-factor generation was enhanced by applying glycerol as the carbon source.

Full text of DOI:10.1016/j.molcatb.2012.01.020

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Chemo-enzymatic route for (R)-terbutaline hydrochloride based on microbial
asymmetric reduction of a substituted ␣-chloroacetophenone derivative
Shohei Taketomia, Masayoshi Asanob, Toshinori Higashia, Mitsuru Shojia, Takeshi Sugaia,∗
a Department of Pharmaceutical Science, Keio University, 1-5-30 Shibakoen, Minato-ku, Tokyo 105-8512, Japan
b Department of Chemistry, Keio University, 3-14-1 Hiyoshi, Kohoku-ku, Yokohama 223-8522, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 11 February 2012
To synthesize (R)-terbutaline hydrochloride, a potent ␤2-adrenoceptor-stimulating agent, asymmetric
reduction of a substituted ␣-chloroacetophenone derivative with cultured whole-cell biocatalyst of the
yeast Williopsis californica JCM 3600 was developed as the key reaction. The reduction proceeded by a
si-facial attack of hydride in a highly enantioselective manner. Co-factor generation was enhanced by
applying glycerol as the carbon source.
Keywords:
Asymmetric reduction
Yeast
© 2012 Elsevier B.V. All rights reserved.
Lipase
Hydrolysis
1. Introduction
2. Experimental
Among a number of aromatic aminoalcohols bearing phenolic
hydroxy groups, terbutaline (1a, as hydrochloride salt) has been
developed as a potent ␤2-adrenoceptor-stimulating agent [1,2].
The activity is predominantly ascribable to (R)-1a, while the (S)-
isomer has been reported to show side effects, causing airway
hyper-reactivity [1]. To date, much effort has been devoted to
produce (R)-isomer, either by fractional crystallization after deriva-
tization to a chiral carboxylate salt [2], or by enzyme-catalyzed
asymmetric hydrocyanation of an aldehyde precursor [3]. Addition-
ally, enantiomerically enriched forms of aromatic aminoalcohols
have a wide variety of uses in synthetic organic chemistry, espe-
cially in asymmetric syntheses [4], so the development of new
methods of preparation is important. We planned to synthesize
(R)-1a by nucleophilic displacement of the halogen atom with tert-
butylamine on chlorohydrin (R)-3, whose two phenolic hydroxy
groups are protected as methoxymethyl (MOM) group. Protection
of the hydroxy groups is necessary, to avoid the oxidation of resorci-
nol moiety in the step of the introduction of nitrogen atom. For this
purpose, MOM group seems to be advantageous, which is cleav-
able under acidic conditions and enables simultaneous formation
of the requisite hydrochloride salt in (R)-1a. The stereochemistry
of the secondary alcohol involved in (R)-3 would be established by
microbial reduction of the corresponding ketone 2a (Scheme 1).
All mps are uncorrected. IR spectra were measured as films
for oils or KBr disks of solids on a Jasco FT/IR-410 spectrome-
ter, and as ATR on a Jeol FT-IR SPX60 spectrometer. 1H NMR
spectra were measured at 270 MHz on a Jeol JNM EX-270 or at
400 MHz on a VARIAN 400-MR spectrometer. 13C NMR spectra
were measured at 63 MHz on a Jeol JNM EX-270 or at 100 MHz
on a VARIAN 400-MR or at 125 MHz on a VARIAN INOVA-500
spectrometer. HPLC data were recorded on Jasco MD-2010 or SHI-
MADZU SPD-20A multi-channel detectors. Optical rotation values
were recorded on a Jasco P-1010 polarimeter. Merck silica gel
60 F254 thin-layer plate (1.05715, 0.25 mm thickness) was used
for thin-layer chromatographic analysis. Merck silica gel 60 F254
thin-layer plates (1.05744, 0.5 mm thickness) and silica gel 60
(spherical and neutral; 100–210 ␮m, 37560-79) from Kanto Chem-
ical Co., Inc. were used for preparative thin-layer chromatography
and column chromatography, respectively. Racemic terbutaline
sulfate, 2-bromo-1-(3ꢀ,5ꢀ-dihydroxyphenyl)ethanone 4a and 1-
(3ꢀ,5ꢀ-diacetoxyphenyl)ethanone 5a were purchased from Wako
Pure Chemical Industries, Ltd. (No. 208-12033), ChemPacific Co.
(No. 32927), and Tokyo Chemical Industry Co. Ltd. (No. D1978),
respectively. Yeast strains are available from Japan Collection
of Microorganisms; Riken Bioresource Center, Planning Section,
Research Promotion Division, RIKEN Tsukuba Institute, Tsukuba,
Ibaraki, Japan, and to NITE Biological Resource Center; Department
of Biotechnology, National Institute of Technology and Evaluation,
Kisarazu, Chiba Japan. Microorganisms were pre-incubated in com-
mon medium [peptone (2%), yeast extract (0.5%), KH2PO4 (0.3%),
K2HPO4 (0.2%) and pH 6.5] with carbon sources as indicated in each
section. Peptone, malt extract and yeast extract were purchased
∗
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.
doi:10.1016/j.molcatb.2012.01.020

84
S. Taketomi et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
provided the desired product (692 mg, 17%), and the combined
OH
H2
N
yield was 88%. Mp 68.0–68.5 ◦C; 1H NMR (400 MHz, CDCl3): ı 2.31
Cl
HO
ꢀ
ꢀ
ꢀ
(s, 6H, Ac), 4.38 (s, 2H, H2), 7.18 (t, J2 ,4 = 2.2 Hz, 1H, H4 ), 7.58 (d,
t-Bu
2H, H2ꢀ,6ꢀ). Calculated from 1H NMR spectrum, this was contami-
ꢀ
ꢀ
nated with ␣,␣-dibromo derivatives (8%) [ı 6.56 (t, J2 ,4 = 2.0 Hz, 1H,
H4ꢀ), 7.69 (d, 2H, H2ꢀ,6ꢀ)], and was employed for next step without
further purification.
OH
(R)-1a
terbutaline hydrochloride
2.3. 2-Chloro-1-(3ꢀ,5ꢀ-diacetoxyphenyl)ethanone (2b)
To a solution of above-mentioned 5b (3.43 g, 92% purity,
10.0 mmol) in CH3CN (100 mL) was added a solution of NaCl (4.12 g,
71.0 mmol, 7.1 equiv.) in water (34.3 mL), and the resulting mixture
was vigorously stirred under reflux for 2.5 h. After cooling, the mix-
ture was concentrated in vacuo, and the residue was diluted with
brine and extracted with AcOEt three times. The combined organic
layer was washed with saturated aqueous Na2S2O3 solution and
brine, dried over anhydrous Na2SO4, and concentrated in vacuo.
The residue was purified by recrystallization from ether to afford
2b (1.71 g, 63%) as colorless fine needles, and further chromato-
graphic purification of the concentrated mother liquor provided
the desired product (455 mg, 17%). The combined yield was 80%.
Mp 92.0–92.5 ◦C; 1H NMR (400 MHz, CDCl3): ı 2.31 (s, 6H, Acꢀ), 4.63
O
OH
MOMO
Cl
MOMO
Cl
OMOM
2a
OMOM
(R)-3
Scheme 1. Microbial asymmetric reduction of a substituted ␣-chloroacetophenone
derivative 2a toward (R)-terbutaline hydrochloride (1a).
ꢀ
ꢀ
(s, 2H, H2), 7.19 (t, J2 ,4 = 2.0 Hz, 1H, H4 ), 7.56 (d, 2H, H2 ,6 ); 13C
NMR (CDCl3, 125 MHz): ı 21.0 (×2), 45.8, 119.1 (×2), 121.1, 135.9,
151.4 (×2), 168.7 (×2), 189.3; IR: 872, 922, 1018, 1122, 1184, 1369,
1443, 1593, 1705, 1763 cm−1. Anal. Calcd for C12H11ClO5: C 53.25,
H 4.10; found: C 53.37, H 4.17.
ꢀ
ꢀ
from Kyokuto Pharmaceutical Industrial Co., Ltd. The growth was
monitored by occasionally observing the OD (660 nm) of the broth
on SP-300 spectrophotometer, Optima Inc., under suitably diluted
conditions so that the OD was shown between 0.15 and 0.50. The
ODs in the text and Sections 2.8 and 2.9 are the recalculated values.
2.1. 2-Chloro-1-[3ꢀ,5ꢀ-bis(methoxymethoxy)phenyl]ethanone
2.4. Candida antarctica lipase B-catalyzed hydrolysis of 2b
(2a)
To a mixture of 2b (297 mg, 1.1 mmol), toluene (3 mL), and phos-
phate buffer (pH 7.0, 0.2 M, 15 mL) was added C. antarctica lipase
B (Novozym 435, 300 mg). The mixture was vigorously stirred for
15.5 h at 50 ◦C under Ar. The mixture was diluted with AcOEt, fil-
tered through a Celite pad and the aqueous layer was extracted
four times with AcOEt. The combined organic layer was washed
with saturated aqueous NaHCO3 and brine, dried over Na2SO4, and
concentrated in vacuo. The residue was purified by silica gel column
chromatography (6 g). Elution with hexane/AcOEt = 4:1 afforded 2c
To a solution of 4a (2.00 g, 8.7 mmol) in CH3CN (86 mL) were
added NaCl (2.78 g, 47.6 mmol) at room temperature. After cool-
ing to 0 ◦C, MOMCl (1.58 mL, 21.7 mmol) and i-Pr2NEt (4.08 mL,
23.4 mmol) were added dropwise to the mixture. The mixture was
stirred overnight at room temperature. The reaction was quenched
with phosphate buffer (pH 7.0, 0.5 M). The mixture was saturated
with NaCl and extracted with AcOEt. The organic layer was washed
with brine, dried over anhydrous Na2SO4, and concentrated in
vacuo. The residue was charged on a silica gel column (100 g). Elu-
tion with hexane–AcOEt (5:1) afforded 2a (1.96 g, 83%) as colorless
needles, mp 47.0–47.5 ◦C. This was unstable upon contact with met-
als such as spatula, and was directly employed for the next step.
(160 mg, 78%) as pale yellow needles, mp 150.5–151.0 ◦C. 1H NMR
ꢀ
ꢀ
ꢀ
(400 MHz, CD3OD): ı 4.81 (s, 2H, H2), 6.51 (t, J2 ,4 = 2.0 Hz, 1H, H4 ),
6.87 (d, 2H, H2ꢀ,6ꢀ); 13C NMR (CD3OD, 125 MHz): ı 46.1, 106.3 (×2),
107.6, 136.2, 158.8 (×2), 191.8; IR: 795, 849, 1011, 1057, 1165, 1277,
1338, 1477, 1593, 1689, 3321 cm−1. Anal. Calcd for C8H7ClO3: C
51.49, H 3.78; found: C 51.53, H 3.74.
1H NMR (400 MHz, CDCl3): ı 3.48 (s, 6H, OCH3), 4.65 (s, 2H, H2),
ꢀ
ꢀ
ꢀ
5.18 (s, 4H, CH3O CH2 ), 6.96 (t, J2 ,4 = 2.2 Hz, 1H, H4 ), 7.24 (d,
2H, H2ꢀ,6ꢀ); IR: 666, 681, 692, 831, 922, 1040, 1088, 1150, 1324,
1453, 1606, 1707, 2826, 2937, 3079 cm−1. When bromoketone 4b
was contaminated, a signal of ı 4.37 (s, 2H, H2) appeared in 1H NMR
spectrum.
2.5. 2-Chloro-1-[3ꢀ,5ꢀ-bis(methoxymethoxy)phenyl]ethanone
(2a)
2.2. 2-Bromo-1-(3ꢀ,5ꢀ-diacetoxyphenyl)ethanone (5b)
To a solution of 2c (148.0 mg, 0.79 mmol) in anhydrous CH3CN
(8.0 mL) was added MOMCl (159 mg, 1.97 mmol, 2.5 equiv.) at 0 ◦C,
and then i-Pr2NEt (296 mg, 2.29 mmol, 2.9 equiv.) was added drop-
wise at 0 ◦C. The mixture was stirred for 1 h at 0 ◦C. Then it was
further stirred at room temperature, and finally for 4 h at 40 ◦C. The
reaction was quenched by the addition of phosphate buffer (pH
7.0, 0.2 M) and the mixture was concentrated in vacuo. The residue
was extracted with AcOEt. The organic layer was washed with
brine, dried over anhydrous Na2SO4, and concentrated in vacuo.
The residue was charged on a silica gel column (5.0 g). Elution with
hexane/AcOEt = 3:1 afforded 2a (138.0 mg, 63%) as colorless nee-
dles. The physical properties and spectral data were identical with
those in Section 2.1.
To a solution of 5a (3.00 g, 12.7 mmol) in AcOH (10.4 mL) was
added Br2 (2.44 g, 15.3 mmol, 1.2 equiv.) in AcOH (3.4 mL) drop-
wise over 30 min at 20 ◦C. The mixture was stirred for 30 min. The
reaction was quenched by the addition of ice and the mixture was
extracted with AcOEt three times. The combined organic layer was
washed successively with water twice, saturated aqueous NaHCO3
solution, water, saturated aqueous Na2S2O3 solution, and brine.
The extract was dried over anhydrous Na2SO4, and concentrated
in vacuo. The residual brown solid was recrystallized from ether
to give 5b (2.84 g, 71%) as slightly yellow needles. Further chro-
matographic purification of the concentrated mother liquor also

S. Taketomi et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
85
2.6. ( )-2-Chloro-1-[3ꢀ,5ꢀ-bis(methoxymethoxy)phenyl]ethanol
(3)
OD of 18 at 36 h. Cells (wet, 4.0 g) were obtained from the broth
(100 mL). A mixture of carbon source with glucose (2%) and glyc-
erol (2%) resulted in a slower growth. The conventional incubation
showed an OD (8.0) after reaching the stationary phase, while under
the aerated conditions, the OD was 18.7 at 36 h to give cells (wet,
3.8 g/100 mL).
To a solution of 2a (47.0 mg, 0.17 mmol) in anhydrous MeOH
(260 ␮L) was added NaBH4 (4.4 mg, 0.12 mmol, 0.71 equiv.) at 0 ◦C.
The mixture was stirred for 5 h. The reaction was quenched with
phosphate buffer (pH 7.0, 0.5 M) and the mixture was extracted
with AcOEt. The organic layer was washed with water, dried over
anhydrous Na2SO4, and concentrated in vacuo. The residue was
purified by preparative TLC (developed with hexane/AcOEt = 2:1)
to afford ( )-3 (36.6 mg, 77%). 1H NMR (400 MHz, CDCl3): ı 3.46 (s,
6H, OCH3), 3.61 (dd, J1,2a = 8.9 Hz, J2a,2b = 11.2 Hz, 1H, H2a), 3.72 (dd,
J1,2b = 3.3 Hz, 1H, H2b), 4.81ꢀ(dd, 1H, H1), 5.14 (s, 4H, CH3O CH2 ),
2.9. (R)-2-Chloro-1-[3ꢀ,5ꢀ-bis(methoxymethoxy)phenyl]ethanol
(3)
Harvested cells (wet, 6.0 g) of W. californica, from incubation
with glycerol for 37 h at 30 ◦C (OD = 20.0), were re-suspended in
phosphate buffer (pH 6.5, 0.1 M, 30 mL) and glycerol (600 mg).
The broth was divided into two test tubes, and each 75 mg
(0.27 mmol) of 2a was added to the tube. Two tubes were
shaken on a reciprocal shaker (210 cpm) at 30 ◦C for 20 h. At 3 h,
an insoluble precipitate was dissolved by the addition of EtOH
(0.5 mL) for each test tube. After extractive workup as described
in Section 2.7, the residue (150 mg) was charged on a silica
gel column (10 g). Elution with hexane/AcOEt = 3:1 afforded (R)-3
(118 mg, 80%). [␣]D25–25.8 (c 1.00, EtOH). Its 98.4% ee was deter-
mined by HPLC analysis: tR (min) = 32.3 [0.8%, (S)-3], 38.8 [99.2%,
(R)-3]. The spectral data were identical with those in Section
2.6.
ꢀ
ꢀ
6.71 (t, J2 ,4 = 2.1 Hz, 1H, H4 ), 6.73 (d, 2H, H2 ,6 ); 13C NMR (CDCl3,
100 MHz): ı 50.8, 56.1 (×2), 73.9, 94.4 (×2), 104.6 (×2), 107.3,
142.4, 158.4 (×2); IR: 650, 697, 854, 922, 1026, 1083, 1145, 1289,
1458, 1599, 2956 cm−1. Anal. Calcd for C12H17ClO5: C 52.09, H
6.19; found: C 51.89, H 6.20. HPLC [column, CHIRALCEL® OD-
H, 0.46 cm × 25 cm; hexane-isopropyl alcohol (15:1); flow rate
0.5 mL/min; detected at 208 nm]: tR (min) = 32.3, 38.8.
ꢀ
ꢀ
2.7. Screening of microorganisms for the reduction of 2a
The microorganisms were incubated in the common medium
(100 mL) with glucose as carbon source (5%) in a 500-mL baffled
Erlenmeyer cultivating flask. The flasks were shaken on a gyra-
tory shaker (180 rpm) for 48 h at 30 ◦C. The cells were harvested
by centrifugation (3000 rpm, 1700 × g) for 15 min, and portion of
cells (wet, ca. 500 mg) were re-suspended in phosphate buffer
solution (pH 6.5, 0.1 M, 10 mL) in a test tube, together with 2a
(20.0 mg) and glucose (500 mg). The test tubes were shaken on
a reciprocal shaker (210 cpm) for 48 h at 30 ◦C. The progress of
the reduction was monitored by a TLC analysis: Rf for 2a: 0.55;
3: 0.45 (developed with hexane/AcOEt = 2:1). The broth was cen-
trifuged (5000 rpm, 1200 × g) for 15 min. The precipitated cells
were extracted three times with acetone. The supernatant in the
step of centrifugation was saturated with NaCl and extracted with
AcOEt four times. The combined extract was concentrated in vacuo.
The residue was diluted with brine and extracted with AcOEt
three times. The extract was dried over anhydrous Na2SO4 and
concentrated in vacuo. The conversion was estimated by com-
paring the area of signals in 1H NMR: ␦ 3.72 (dd, 1H, H2b for
3), 4.65 (s, 2H, H2 for 2a), 6.71 (t, 1H, H4ꢀ for 3) and 6.96 (t,
1H, H4ꢀ for 2a) of this crude product. The residue was purified
by preparative TLC (developed with hexane/AcOEt = 2:1) to afford
3, whose ee was determined by HPLC analysis as described in
Section 2.6.
2.10. (R)-[3ꢀ,5ꢀ-Bis(methoxymethoxy)phenyl]oxirane (10)
To a solution of (R)-3 (807 mg, 2.91 mmol) in ether (27.0 mL)
were added TBAI (42.7 mg, 0.12 mmol, 0.04 equiv.) and aqueous
NaOH solution (2 M, 13.5 mL). The mixture was stirred for 4 h at
room temperature and quenched with phosphate buffer (pH 7.0,
0.5 M). The ether layer was separated and the aqueous layer was
extracted with AcOEt. The combined organic layer was washed with
brine, dried over anhydrous Na2SO4, and concentrated in vacuo. The
residue was charged on a silica gel column (32 g). Elution with hex-
ane/AcOEt = 5:1 afforded (R)-10 (667 mg, 96%). This was employed
for next step without further purification. 1H NMR (400 MHz,
CDCl3): ı 2.69 (dd, J1,2a = 2.5 Hz, J2a,2b = 5.4 Hz, 1H, H1a), 3.03 (dd,
J1,2b = 3.9 Hz, 1H, H2b), 3.40 (s, 6H, OCH3), 3.74 (dd, ꢀ1H, H1), 5.08
ꢀ
ꢀ
ꢀ
(s, 4H, CH3O CH2 ), 6.57 (d, J2 ,4 = 2.4 Hz, 2H, H2 ,6 ), 6.60 (t, 1H,
H4ꢀ); IR: 693, 836, 922, 1024, 1143, 1212, 1294, 1378, 1457, 1597,
2827 cm−1
.
2.11. (R)-1-[3ꢀ,5ꢀ-Bis(methoxymethoxy)phenyl]-2-(tert-
butylamino)ethanol (1b)
2.8. Pre-incubation of Williopsis californica JCM 3600
W. californica was inoculated in a common medium. With glu-
cose as carbon source (5%, w/v), the gyrotary-shaken culture (30 ◦C,
180 rpm) showed OD (660 nm) = 0.7 with an early logarithmic
growing phase at 12 h. After being its OD to be 8.8 at 24 h, the
growth become slower, but finally the OD reached to be 18 after
72 h, by way of 8.5 at 36 h. Cells (wet, 2.5 g/100 mL) could be recov-
ered by centrifugation (3000 rpm, 1700 × g) for 15 min at the stage
of 36 h incubation. The growth was accelerated, by adding antifoam
(Dow-Corning, FS antiform AFE emulsion, diluted 10 times with
water, 0.5 mL) to the broth (100 mL) and replacement of conven-
tional urethane-foam stopper with perforated filter paper cap when
OD reached 2.3 at 12 h. The amount of wet cells was increased to be
3.8 g from 100 mL of the broth at 33 h. The change of carbon source
from glucose (5%) to glycerol (2%, w/v) brought about a moderate
growth with an OD of 12 at 36 h under the stationary phase. Aer-
ated condition as mentioned above also promoted the growth with
To a solution of (R)-10 (25.8 mg, 0.11 mmol) in anhydrous MeOH
(2.0 mL) was added tert-butylamine (500 ␮L, 4.73 mmol, 43 equiv.).
The mixture was stirred for 24 h at 50 ◦C. Then another portion
of tert-butylamine (400 ␮L, 3.78 mmol, 34 equiv.) was added, and
the mixture was further stirred for 5 h. The mixture was concen-
trated in vacuo, and the residue was purified by preparative TLC
(developed with EtOH) to afford (R)-1b (26.6 mg, 77%). 1H NMR
(400 MHz, CD3OD): ı 1.11 (s, 9H, tert-butyl), 2.65 (dd, J1,2a = 4.4 Hz,
J2a,2b = 11.7 Hz, 1H, H2a), 2.71 (dd, J1,2b = 8.8 Hz, 1H, H2b), 4.61 (dd,
1H, H1), 3.43 (s, 6H, OCH3), 5.15 (s, ꢀ4H, CH3O CH2 ), 6.62 (t,
ꢀ
ꢀ
ꢀ
ꢀ
J2 ,4 = 2.0 Hz, 1H, H4 ), 6.71 (d, 2H, H2 ,6 ). Regioisomer 11 on which
amine attacked benzylic position: ı 1.02 (s, 9H, tert-butyl), 3.35 (dd,
J1,2a = 8.8 Hz, J2a,2b = 10.9 Hz, 1H, H2a), 3.43 (s, 6H, OCH3), 3.51 (dd,
J1,2b = 4.9 Hz, 1H, H2b), 3.84ꢀ(dd, 1H, H1), 5.14 (s, 4H, CH3O CH2 ),
ꢀ
ꢀ
ꢀ
ꢀ
6.58 (t, J2 ,4 = 2.2 Hz, 1H, H4 ), 6.73 (d, 2H, H2 ,6 ).

86
S. Taketomi et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
2.12. (R)-2-(N-tert-butylamino)-1-(3ꢀ,5ꢀ-dihydro-
O
O
xyphenyl)ethanol hydrochloride (1a)
HO
Br
MOMO
X
a
Alcohol (R)-1b (26.6 mg, 0.085 mmol) was treated with a mix-
ture of concentrated HCl (12 M) and ethanol (1:5, 1.5 mL), and
stirred for 24 h at room temperature. The reaction mixture was
concentrated in vacuo to afford (R)-1a (22.0 mg, quant, 99.4% ee).
[␣]D25–46.7 (c 1.10, MeOH) [lit. [3] [␣]D20–32.5 (c 0.76, MeOH)].
1H NMR (400 MHz, CD3OD): ı 1.38 (s, 9H, tert-butyl), 2.97 (dd,
OH
OMOM
2a: X = Cl
4a
4b:
Br
J1,2a = 10.7 Hz, J2a,2b = 12.7 Hz, 1H, H2a), 3.10 (dd, J1,2b = 2.9 Hz, 1H,
ꢀ
ꢀ
ꢀ
H2b), 4.75 (dd, 1H, H3), 6.21 (t, J2 ,4 = 2.4 Hz, 1H, H4 ), 6.37 (d,
2H, H2ꢀ,6ꢀ); 13C NMR (CD3OD, 100 MHz): ı 25.8 (×3), 49.8, 58.1,
70.7, 103.2, 105.2 (×2), 144.7, 159.8 (×2). NMR spectra were
in good accordance with those of commercially available terbu-
taline sulfate. IR: 849, 1049, 1159, 1383, 1459, 1605, 2876, 2979,
O
O
AcO
X
AcO
Cl
c
3399 cm−1
.
OAc
OAc
5a: X = H
2b
b
2.13. (R)-N-[2-acetoxy-2-(3ꢀ,5ꢀ-diacetoxyphenyl)ethyl]-N-tert-
butylacetamide (12)
b:
Br
O
To a solution of (R)-1a (3.9 mg, 0.042 mmol) in pyridine (0.5 mL)
was added acetic anhydride (0.5 mL, 5.3 mmol) and DMAP (0.2 mg),
and the mixture was stirred for 16 h at room temperature. The
reaction mixture was quenched by the addition of ice and the mix-
ture was extracted with AcOEt three times. The combined organic
layer was washed successively with water, HCl (1 M), water, satu-
rated aqueous NaHCO3 solution, and brine. The extract was dried
over anhydrous Na2SO4, and concentrated in vacuo. The resulted
oily residue was purified by preparative TLC (developed with hex-
ane/AcOEt = 1:1) to afford pure 12 (0.4 mg, 7%, 99.4% ee). This was
analyzed by HPLC [column, CHIRALCEL® OD-H, 0.46 cm × 25 cm;
hexane/isopropyl alcohol = 10:1, flow rate 0.5 mL/min, detected at
212 nm]: tR (min) = 29.7 [0.3%, (S)-12], 46.8 [99.7%, (R)-12]. 1H
NMR (400 MHz, CDCl3): ı 1.44 (s, 9H, tert-butyl), 2.08 (s, 3H,
NCOCH3), 2.19 (s, 3H, 2-OCOCH3) 2.27 (s, 6H, Ar OCOCH3) 3.53
RO
Cl
d
OR
2c: R = H
a:
MOM
MOM = methoxymethyl
e
Scheme 2. Reagents and conditions: (a) MOMCl, NaCl, i-Pr2NEt, CH3CN (83%); (b)
Br2, AcOH; (c) NaCl, CH3CN, H2O (65%, two steps from 5a); (d) C. antarctica lipase B
(Novozym 435), pH 7.0 (78%); and (e) MOMCl, i-Pr2NEt, CH3CN (63%).
(dd, J1a,2 = 3.4 Hz, J1a,1b = 16.1 Hz, 1H, H1a), 3.72 (dd, J1b,2 = 10.3 Hz
ꢀ
ꢀ
ꢀ
1H, H1b), 5.87 (dd, 1H, H2), 6.91 (t, J2 ,4 = 2.0 Hz, 1H, H4 ), 6.93 (d,
and we optimized the reaction conditions. The limited dose of i-
Pr2NEt (2.9 equiv.) in the treatment with MOMCl (2.5 equiv.) was
the key for successful protection (63%).
2H, H2ꢀ,6ꢀ).
3. Results and discussion
Recently, a number of biocatalysts for the asymmetric reduc-
tion of substituted acetophenones have been studied [5–20]. We
applied set of incubated microorganisms involving yeasts and
fungi, which had been used for the reduction of ketone 6 with
an aromatic ring [21] (Scheme 3). In our hand, out of six strains
which showed a certain progress of the reduction, two exhibited the
desired enantiofacial selectivity as shown in Table 1. The si-facial
attack took place to give (R)-3. Moran and co-workers also reported
a si-facial attack on similarly substituted ␣-haloacetophenone 8 to
give (R)-9 [12]. Another high, but reverse re-facial selectivity was
shown in three strains.
Our first attempt was the direct synthesis of 2a from
commercially available 4a. The desired product was, however, con-
taminated with certain amount of the corresponding bromide 4b,
whose strongly alkylating property was deleterious as an inhibitory
effect in the subsequent microbial reduction. Due to the very
similar chromatographic behavior between 2a and 4b, and poor
crystalline nature of 2a (mp 47.0–47.5 ◦C), we switched the syn-
thetic route by way of another chloride 2c with higher melting
point (150.5–151.0 ◦C) as shown in Scheme 2. Cheaper starting
material 5a was treated with bromine in acetic acid to give ␣-
bromoketone 5b. The bromine atom was substituted with chlorine
by applying NaCl in aqueous acetonitrile. The next requisite task
was the hydrolysis of two phenolic acetates. Under basic condi-
tions such as K2CO3/MeOH, the disappearance of 2b was very fast,
but some over-reaction was accompanied with the formation of
more polar by-products. In turn, under strongly acidic conditions
(hydrochloric acid), the hydrolysis was very slow. As this substrate
has ␣-haloketone moiety which is susceptible to the nucleophilic
displacement, biocatalytic transformation seemed to be advanta-
geous under mild conditions.
In Pichia minuta, Trichosporon cutaneum, and W. californica,
which had high enantiofacial preference on 6, for the present sub-
strate 3, the attack of hydride occurred from the same enantioface
as illustrated in Fig. 1.
The selected W. californica JCM 3600 showed high catalytic
activity, however, there was observed a severe problem. The enan-
tiomeric excess of the product (R)-3 fluctuated (80–99% ee) even
under the same incubation conditions, and we could not find
out any reason in the incubation by applying conventional glu-
cose medium for pre-incubation and reaction. In the incubated
whole-cell microorganism, total enantioselectivities is the sum of
contrasting enantiofacial preference shown by plural enzymes. For
enhancing the major reductase with desired enantioselectivity, we
embarked upon the optimization of the incubation and reduction
conditions.
For that purpose, C. antarctica lipase B (Novozymes, Novozym
435) was very effective catalyst to give 2c (78%). MOM protection
of the two phenolic hydroxy groups in 2c was the next task. The
reactive ␣-chloroketone itself worked as a good alkylating agent,

S. Taketomi et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
87
Table 1
Screening on the reduction of 2a.
Microorganism
(JCM No.)
Conversion (%)a
Recovery of 2a (%)
3b isolated yield (%)
Abs. config.
%Ee
Candida kefyr
(21874)
Candida
66
8
35
90
50
5
S
S
96.7
83.8
nitratophila
(9856)
Pichia farinosa
15
12
86
85
74
13
10
12
69
S
R
S
45.1
69.7
90.4
(10896)c
Pichia minuta
(3622)
Trichosporon
cutaneum
(1534)
Williopsis
californica
(3600)
97
3
94
R
99.3
a
Determined by 1H NMR analysis of the crude product. For detail, see Section 2.7.
See Section 2.7.
NBRC No.
b
c
First, we found that the growth strongly depended upon the
effect by the addition of glycerol showed a certain time lag, we sus-
pected that the carbonyl reductases responsible for the reduction
of 2a were induced. Then, at the pre-incubation stage, the growth
and the enzyme activity were examined by changing the carbon
source [glucose (5%), glycerol (2%), and glucose (2%) + glycerol (2%)].
Although growth was slightly slower after pre-incubation with
glycerol (2%) compared with the other two conditions involving
glucose, in all three cases the OD eventually equally reached ca. 18
under intensively aerobic conditions as mentioned above. By apply-
ing whole-cell biocatalysts of W. californica harvested through the
optimized pre-incubation conditions with glycerol, the reduction
of 2a by adding glycerol proceeded very smoothly to provide (R)-3
with 98.4% ee in an 80% yield (Scheme 3).
We should mention some comments on the low concentration
(5.0 g/L) of substrate 2a. Due to the hydrophobic property of 2a,
its solubility in the broth was very low. When the charge of sub-
strate is increased (10–20 g/L), certain amount remained intact as
the precipitates in the incubation mixture. An increased concen-
tration of 2a by the addition of water-miscible solvent resulted
in the poor enantioselectivity, as the chloroketone structure itself
worked as an inhibitor for the reductase with si-facial preference.
Such inhibitory effect was more prominent with the bromoketone
4b. When the substrate 2a was incubated in the presence of 4b
(1.7 g/L), the reduction of 2a became very slow and the ee of (R)-
3a was as low as 63%. Those observations recommend the keeping
the substrate concentration to be low, according to the elaborated
experimental procedure (Section 2.9).
supply of air during the pre-incubation, and the growth was
promoted under aerobic conditions by applying antifoam and a
perforated cap (see Section 2.8). Cells (wet, 3.8 g/100 mL of broth)
were harvested at OD (660 nm) = 18 after 36 h. On the other hand,
pre-incubation with a limited air supply retarded growth. For
example, it dropped when the incubation flasks were capped with
conventional urethane-foam stoppers. An apparent log phase was
observed between 9 and 24 h, but after that, the growth slightly
slowed down, finally reaching OD = 18 (3.0 g/100 mL of broth) after
36 h.
Among glucose, glycerol [22], and 2-propanol as the external
carbon sources for the regeneration of NAD(P)H, only the addition
of glycerol (2%, w/v) remarkably promoted the reduction. As the
O
OH
MOMO
Cl
MOMO
Cl
a
OMOM
2a
OMOM
3
O
OH
O
O
O
O
S
S
a
W. californica
P. minuta
i-Pr
i-Pr
6
7
H
H
O
O
Cl
O
S
r
P
-
i
O
OH
O
M
Cl
Cl
O
O
M
b
H
H
6
M
O
O
O
M
2a
O
O
O
8
9
T. cutaneum
Scheme 3. Reagents and conditions: (a) whole-cell microorganisms, see Table 1;
and (b) Rhodotorula glutinis CCT 2182 [12] (98%, >99% ee).
Fig. 1. Enantiofacial preference of microbial reductions on ketones.

88
S. Taketomi et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 83–88
B-mediated hydrolysis under mild conditions enabled the choice of
easily available and cheap starting material 5a, whose two phenolic
hydroxy groups were protected as acetates.
Acknowledgments
This work was supported by “SORST: Solution-Oriented
Research for Science and Technology” of Japan Science and Technol-
ogy Corporation, and we express our sincere thanks to Professors
Keisuke Suzuki, Takashi Matsumoto, Ken Ohmori of Tokyo Insti-
tute of Technology. This work was also supported by “High-Tech
Research Center” Project for Private Universities: matching fund
subsidy 2006–2011 from the Ministry of Education, Culture, Sports,
Science and Technology, Japan, and acknowledged with thanks. We
acknowledge Novozymes Japan for the generous gift of Novozym
435. M. A. thanks Japan Student Service Organization, for the
exemption from repayment upon graduation for graduate school
students with outstanding results (2007).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2012.01.020.
References
[1] B. Waldeck, Eur. J. Phamacol. 445 (2002) 1.
[2] K. Wetterlin, J. Med. Chem. 15 (1972) 1182.
[3] F. Effenberger, J. Jäger, J. Org. Chem. 62 (1997) 3867.
[4] D.J. Ager, I. Prakash, D.R. Schaad, Chem. Rev. 96 (1996) 835.
[5] K. Nakamura, R. Yamanaka, T. Matsuda, T. Harada, Tetrahedron: Asymmetry 14
(2003) 2659.
[6] K. Nakamura, T. Matsuda, Curr. Org. Chem. 10 (2006) 1217.
[7] T. Matsuda, R. Yamanaka, K. Nakamura, Tetrahedron: Asymmetry 20 (2009)
513.
[8] J. Goswami, R.L. Bezbaruah, A. Goswami, N. Borthakur, Tetrahedron: Asymme-
try 12 (2001) 3343.
[9] N.A. Salvi, S. Chattopadhyay, Tetrahedron 57 (2001) 2833.
[10] J.V. Comasseto, L.H. Andrade, A.T. Omori, L.F. Assis, A.L.M. Porto, J. Mol. Catal.
B: Enzyme 29 (2004) 55.
[11] R.N. Patel, A. Goswami, L. Chu, M.J. Donovan, V. Nanduri, S. Goldberg, R. John-
ston, P.J. Siva, B. Nielsen, J. Fan, W. He, Z. Shi, K.Y. Wang, R. Eiring, D. Cazzulino,
A. Singh, R. Mueller, Tetrahedron: Asymmetry 15 (2004) 1247.
[12] H. Antunes, L.C. Fardelone, J.A.R. Rodrigues, P.J.S. Moran, Tetrahedron: Asym-
metry 15 (2004) 2615.
[13] A.L. Kamble, P. Soni, U.C. Banerjee, J. Mol. Catal. B: Enzyme 35 (2005) 1.
[14] L.H. Andrade, J.V. Comasseto, D.F. Rodrigues, V.H. Pellizari, A.L.M. Porto, J. Mol.
Catal. B: Enzyme 33 (2005) 73.
[15] J.V. Comasseto, L.F. Assis, L.H. Andrade, I.H. Schoenlein-Crusius, A.L.M. Porto, J.
Mol. Catal. B: Enzyme 39 (2006) 24.
[16] D. Pollard, M. Truppo, J. Pollard, C. Chen, J. Moore, Tetrahedron: Asymmetry 17
(2006) 554.
[17] L.C. Fardelone, A.R. Rodriquez, P.J.S. Moran, J. Mol. Catal. B: Enzyme 39 (2006)
9.
[18] W. Yang, J.-H. Xu, H. Xie, Y. Xu, G. Zhao, G.-Q. Lin, Tetrahedron: Asymmetry 17
(2006) 1769.
Scheme 4. Reagent and conditions: (a) Williopsis californica, pH 6.5, glycerol (80%,
98.4% ee); (b) aq. NaOH (2 M), TBAI, Et2O (96%); (c) tert-BuNH2, MeOH; (d) aq. HCl
[79%, two steps from (R)-10]; and (e) Ac2O, pyridine, DMAP.
The requisite (R)-3 was converted into epoxide (R)-10 with
a high yield of 96%, by the aid of a catalytic amount of TBAI
under basic conditions. The attack of tert-butylamine proceeded
predominantly (9:1) on the terminus of the epoxy ring over the
regioisomer 11. Acid treatment of the purified isomer to remove the
MOM protective groups worked very well to provide terbutaline
hydrochloride (R)-1a (Scheme 4) in 77% in two steps from (R)-10.
The ee of (R)-1a was confirmed to be 99.4% by the HPLC analysis of
(R)-12.
[19] E. Kyslíková, P. Babiak, H. Maresová, A. Palyzová, J. Hájecek, P. Kyslík, J. Mol.
Catal. B: Enzyme 67 (2010) 266.
[20] B.A. Barros-Filho, F.M. Nunes, M.C.F. de Oliveira, L.G. Lemos, M.C. de Mattos, G.
de Gonzalo, V. Gotor-Fernández, V. Gotor, J. Mol. Catal. B: Enzyme 65 (2010)
37.
[21] C. Hiraoka, M. Matsuda, Y. Suzuki, S. Fujieda, M. Tomita, K. Fuhshuku, R. Obata,
S. Nishiyama, T. Sugai, Tetrahedron: Asymmetry 17 (2006) 3358.
[22] L.H. Andrade, L. Piovan, M.D. Pawquini, Tetrahedron: Asymmetry 20 (2009)
1521.
4. Conclusion
In this study, an expeditiousapproachfor terbutalinehydrochlo-
ride was examined. Preparation of highly enantiomerically
enriched key intermediate (R)-3 was established, based on W.
californica-catalyzed asymmetric reduction. C. antarctica lipase