Chemoenzymatic synthesis of (S)-duloxetine using carbonyl reductase from Rhodosporidium toruloides

Article abstract of DOI:10.1016/j.bioorg.2016.02.002

A chemoenzymatic strategy was developed for (S)-duloxetine production employing carbonyl reductases from newly isolated Rhodosporidium toruloides into the enantiodetermining step. Amongst the ten most permissive enzymes identified, cloned, and overexpressed in Escherichia coli, RtSCR9 exhibited excellent activity and enantioselectivity. Using co-expressed E. coli harboring both RtSCR9 and glucose dehydrogenase, (S)-3-(dimethylamino)-1-(2-thienyl)-1-propanol 3a was fabricated with so far the highest substrate loading (1000 mM) in a space-time yield per gram of biomass (DCW) of 22.9 mmol L^-1 h^-1 g DCW^-1 at a 200-g scale. The subsequent synthetic steps from RtSCR9-catalyzed (S)-3a were further performed, affording (S)-duloxetine with 60.2% overall yield from 2-acethylthiophene in >98.5% ee.

Full text of DOI:10.1016/j.bioorg.2016.02.002

Bioorganic Chemistry 65 (2016) 82–89
Contents lists available at ScienceDirect
Bioorganic Chemistry
journal homepage: www.elsevier.com/locate/bioorg
Chemoenzymatic synthesis of (S)-duloxetine using carbonyl reductase
from Rhodosporidium toruloides
⇑
Xiang Chen, Zhi-Qiang Liu, Chao-Ping Lin, Yu-Guo Zheng
Key Laboratory of Bioorganic Synthesis of Zhejiang Province, College of Biotechnology and Bioengineering, Zhejiang University of Technology, Hangzhou 310014, China
Engineering Research Center of Bioconversion and Biopurification of Ministry of Education, Zhejiang University of Technology, Hangzhou 310014, China
a r t i c l e i n f o
a b s t r a c t
Article history:
A chemoenzymatic strategy was developed for (S)-duloxetine production employing carbonyl reductases
from newly isolated Rhodosporidium toruloides into the enantiodetermining step. Amongst the ten most
permissive enzymes identified, cloned, and overexpressed in Escherichia coli, RtSCR9 exhibited excellent
activity and enantioselectivity. Using co-expressed E. coli harboring both RtSCR9 and glucose dehydroge-
nase, (S)-3-(dimethylamino)-1-(2-thienyl)-1-propanol 3a was fabricated with so far the highest substrate
loading (1000 mM) in a space-time yield per gram of biomass (DCW) of 22.9 mmol LÀ1 hÀ1 g DCWÀ1 at a
200-g scale. The subsequent synthetic steps from RtSCR9-catalyzed (S)-3a were further performed,
affording (S)-duloxetine with 60.2% overall yield from 2-acethylthiophene in >98.5% ee.
Ó 2016 Elsevier Inc. All rights reserved.
Received 15 November 2015
Revised 5 February 2016
Accepted 8 February 2016
Available online 8 February 2016
Keywords:
Chemoenzymatic
Duloxetine
Carbonyl reductases
Asymmetric synthesis
(S)-3-(dimethylamino)-1-(2-thienyl)-1-pro
panol
1. Introduction
One-step asymmetric reduction has been established for
accessing enantioenriched thiophene-ring containing alcohols
(S)-duloxetine [CymbaltaÒ, (S)-N-methyl-3-(1-naphthaleny
loxy)-3-(2-thienyl) propanamine] is a versatile pharmaceutical in
the treatment of major depressive disorders, stress urinary incon-
tinence, and obsessive compulsive disorder [1–3]. The increasing
demand has resulted in the quest for the effective synthesis routes
to manufacture (S)-duloxetine, particularly in the enantioselective
preparation. Optically active thiophene-ring containing alcohols
with Li(ent-Chirald)2AlH2, oxazaborolidine or Ru catalysts [19–
24], however, it prevalently provides moderate enantioselectivity
and leads to by-products. In contrast to traditional chemical
method, biocatalytic asymmetric synthesis of enantiopure thiop
hene-ring containing alcohols is less well documented, though
the use of carbonyl reductases to afford chiral pharmaceutical
intermediates is widely proposed [25–30].
constitute key building blocks, in fact enantiopure
hols, -chloro alcohols, b-hydroxy esters, and b-hydroxynitrile
are frequently employed [4,5].
c
-amino alco-
To date, several microorganisms and enzymes, including Can-
dida tropicalis [31], Candida viswanathii [32], Saccharomyces cere-
visiae [33], carbonyl reductases from Candida albicans XP1463
[34], Exiguobacterium sp. F42 [35], Rhodotorula glutinis CY12 [36],
and Chryseobacterium sp. CA49 [37] have been described as the bio-
catalysts for the asymmetric synthesis of chiral thiophene-ring
containing alcohols with high stereoselectivity, yet with low sub-
strate loading (50 g LÀ1) and long reaction time (24 h). Therefore,
an efficient strategy to achieve the enantiodetermining step would
be advantageous.
Herein, we develop an effective and straightforward route to
synthesize (S)-duloxetine starting from 2-acetylthiophene in only
four steps, employing carbonyl reductases in the enantioselective
procedure with high substrate concentration (1000 mM), yield
(92.1%), and enantioselectivity (99.9% ee). This highly efficient
chemoenzymatic pathway appears to have potential for industrial
application for (S)-duloxetine production.
c
Despite the existence of numerous pathways for the enantiose-
lective process, kinetic resolution and asymmetric reduction can be
highlighted [6–12]. The kinetic resolution is a major approach for
the synthesis of chiral thiophene-ring containing alcohols, whereas
the maximum theoretical yield 50% is too low [13–15]. The resolu-
tion of racemic 3a via diastereomeric salt formation with (S)-
mandelic acid has been used for an industrial-scale production
coupled with the racemization of the antipode, while it required
four steps from 2a and the yield (<70%) was insufficient to accom-
plish an economic and environmental balance (Scheme 1) [16–18].
⇑
Corresponding author at: Key Laboratory of Bioorganic Synthesis of Zhejiang
Province, College of Biotechnology and Bioengineering, Zhejiang University of
Technology, Hangzhou 310014, China.
E-mail address: zhengyg@zjut.edu.cn (Y.-G. Zheng).
http://dx.doi.org/10.1016/j.bioorg.2016.02.002
0045-2068/Ó 2016 Elsevier Inc. All rights reserved.

X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
83
(HCHO)n,
HN(CH3)2·HCl,
Me2CHOH
NaBH4,
MeOH
N
N
S
S
S
0 oC, 90%
reflux for 6 h, 94%
HCl
O
2a
O
1
OH
3a
t-BuOMe,
HCl, 83%
(S)-mandelic acid 7,
t-BuOMe, 41%
Chemo- or biocatalysts
OH
NaOH
N
NH
N
S
O2C
Ph
S
S
92%
OH
OH
OH
(S)-3a
(R)-3a
(S)-3a.(S)-7
Scheme 1. The industrial route for (S)-3a production and asymmetric reduction of 2a.
molecular weight of RtSCR9 was around 27 kDa (Fig. 1), and the
2. Results and discussion
quaternary structure was tetramer.
2.1. Identification and sequence analysis of recombinant carbonyl
reductases
2.3. Characterization of recombinant carbonyl reductases
More than 200 microbial strains from soil samples were
screened for the asymmetric reduction of thiophene-ring contain-
ing ketone 2a. Strain ZJB2014212 was found to display activity and
high enantioselectivity (>95% ee). Combination of physiological,
biochemical, and genetic tests, it was identified as Rhodosporidium
toruloides ZJB2014212 (Table S1, Figs. S1 and S2), and deposited in
China Center for Type Culture Collection (CCTCC No: M 2014613)
for further studies.
To explore the robust carbonyl reductases for the target reac-
tions, ten most permissive open reading frames encoded rtscr
genes (designed as RtSCR1 to RtSCR10) from ZJB2014212 were
identified through BLAST-P searches on the genome of R. toruloides
NP11 combined with the distinct sequence motifs of BgADH2 (Gen-
RtSCR9 activity was improved 11 and 5-fold in response to dif-
ferent pHs and temperatures, and the highest activity was found at
pH 7.0, 35 °C (Fig. S5). RtSCR9 has the potential for prolonged incu-
bation due to its stability at 35 °C for 48 h. Although RtSCR9 dis-
played a remarkably higher activity in the presence of Fe2+ and
Mn2+, the metal chelator EDTA slightly reduced the activity, indica-
tive of the non-metal dependence of RtSCR9, different from the
most representive carbonyl reductase LbADH (Table S4) [43].
Importantly, RtSCR9 preferred NADPH over NADH in terms of the
Km value against NADPH was more than 200-fold lower than that
bank: YP_004348055.1),
a known active carbonyl reductase
toward 2a [25,38]. Notably, RtSCR1 (Genbank: EMS20030.1),
RtSCR2 (Genbank: EMS18263.1), RtSCR3 (Genbank: EMS21819.1),
RtSCR4 (Genbank: EMS18956.1), RtSCR5 (Genbank: EMS21415.1),
and RtSCR6 (Genbank: EMS21411.1) are annotated as short-chain
dehydrogenase/reductase; RtSCR7 (Genbank: EMS18506.1) and
RtSCR9 (Genbank: EMS18622.1) belong to glucose/ribitol dehydro-
genase; RtSCR8 (Genbank: EMS22288.1) is commented as 3a(or
20b)-hydroxysteroid dehydrogenase, and RtSCR10 (Genbank:
EMS20237.1) is putative 3-hydroxyacyl-CoA-dehydrogenase. The
sequence identities among RtSCRs are low (only 28–32%), while
the conserved motifs are consistent, including ‘‘Rossmann-fold”
domain, cofactor-binding area and active site (Fig. S3) [39].
2.2. Preparation and screening of recombinant carbonyl reductases
Fig. 1. SDS-PAGE analysis of purified RtSCR9. Lane 1, Marker; Lane 2, RtSCR9
supernatant; Lane 3, purified RtSCR9.
RtSCRs were cloned, and heterologously expressed in Escheri-
chia coli. The sodium dodecyl sulfatepolyacrylamide gel elec-
trophoresis (SDS-PAGE) analysis revealed bands with molecular
weights ranging from 26 to 35 kDa, which were consistent with
the theoretical molecular weights (Fig. S4). Most of the enzymes
were expressed in soluble and active, except for RtSCR6 and
RtSCR10 with low protein expression level. RtSCRs were screened
against 2a through the formation of corresponding (S)-3a. Gratify-
ingly, RtSCR9 displayed excellent enantioselectivity (>99% ee) and
conversion (>99%). Immobilized metal chelate affinity chromatog-
raphy was selected for purification of RtSCR9 expressed as His6-
tagged proteins [40]. The native molecular mass of RtSCR9 deter-
mined on Discovery BIO GFC 150 column was calculated at
123.7 kDa. Like known carbonyl reductases [41,42], the subunit
Table 1
Apparent kinetic parameters of RtSCR9.a
Substrate Km
(mM)
Vmax
mol minÀ1 mgÀ1
kcat
kcat/Km
(
l
)
(sÀ1
)
(mMÀ1 sÀ1
)
NADH
NADPH
2a
9.94
0.048
0.52
–
–
7.1
–
–
14.6
–
–
28.1
a
Reaction conditions: substrate 2a (50 mM), NADH (0.1–10 mM) or NADPH
(0.01–0.5 mM), purified RtSCR9 (0.1 mg mLÀ1), pH 7.0, and 35 °C toward NAD(P)H;
2a (0.1–50 mM), purified RtSCR9 (0.1 mg mLÀ1), NADPH (0.5 mM), pH 7.0, and 35 °C
against 2a.

84
X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
Table 2
toward NADH (Table 1). High turnover numbers (TONs) was found
at 21694, rivaling that of some most active SCRs reported to date
for prochiral ketones transformations [44,45].
Substrate specificity of RtSCR9.a
Entry
Substrate
Con. (mM)
Conv. (%)
ee (%)
1
2
3
4
5
6
7
8
2b
2b
2c
2c
2d
2d
2e
2e
50
250
50
250
50
250
50
250
90
87
>99(S)b
>99(S)b
>99(S)b
>99(S)b
>99(S)c
>99(S)c
>99(S)b
>99(S)b
2.4. Substrate specificity
>99
>99
>99
>99
>99
>99
To check whether RtSCR9 could accept other substrates, a rep-
resentative set of thiophene-ring containing ketones, including
3-(methylamino)-1-(2-thienyl)-1-propanonehydrochloride 2b, 3-
chloro-1-(2-thienyl)-1-propanone 2c, 3-oxo-3-(2-thienyl) propi-
onic acid ethyl ester 2d, and 3-oxo-3-(2-thienyl)propanenitrile 2e
were then synthesized through Mannich reaction [21], or Friedel-
Crafts acylation (Scheme 2) [46], and subjected to the reaction con-
ditions. Interestingly, RtSCR9 was not restricted to 2a, and it
worked well against other key building blocks of (S)-duloxetine
related to 2a, following the anti-Prelog’s rule (Table 2). Of the four
substrates tested, all led to excellent enantioselectivity (>99% ee).
Remarkably, the conversion of 2c, 2d, and 2e were higher than
toward 2b. This might be explained by the electron-withdrawing
substituents in the side of the carbonyl group which usually
increase the electrophilicity of carbonyl carbon and enhance the
activities of enzymes [47,48].
Reaction conditions: substrate 2b–2e (50–250 mM), glucose (278 mM), NADP+
a
(0.05 mM), purified RtSCR9 (0.1 mg mLÀ1), and GDH (0.1 mg mLÀ1), pH 7.0, 35 °C for
4 h.
b
Determined by HPLC analysis.
Determined by GC analysis.
c
2.6. Optimization of reaction conditions
To further employ the newly constructed E. coli/RtSCR9-GDH in
the practical synthesis of (S)-3a and minimize the costs, we main-
tained high substrate loading (1000 mM), keeping NADP+/2a ratio
at 0.75 lmmol/(mmol) and optimizing a series of reaction param-
eters. The optimal pH and temperature were observed at pH 7.0
and 30 °C (Fig. 3). The impact of glucose/2a ratio was subsequently
evaluated and revealed an increased conversion as the glucose/2a
was raised from 0.5 to 6 mmol/(mmol) (Fig. 3c). Overall, a 6-fold
excess of glucose over 2a was found to be optimum for maximizing
production of (S)-3a. Different cell dosages (1–30 g CDW LÀ1) were
investigated as well. At a cell loading of 10 g CDW LÀ1, co-
expressing E. coli/RtSCR9-GDH was found to give quantitative con-
version (Fig. 3d). When the biocatalyst dosage exceeded
20 g CDW LÀ1, the conversion was slightly impaired. This might
be due to the increased viscosity hindered the transportation
between the molecules in the reaction media.
2.5. Co-expression of RtSCR9 and GDH genes
As the expensive cofactor dependency and low substrate toler-
ance of carbonyl reductases are the major impediments to the
practical applications [49,50], ‘‘enzyme-coupled” cofactor recycling
using glucose dehydrogenase (GDH) and glucose was first selected
for the NADPH regeneration to alleviate the reversibility and
strengthen the thermodynamic driving force [51–53]. In addition
to the possibility without additional cofactor in the asymmetric
reduction working with co-expression E. coli [54,55], ‘‘designer
cell” harboring RtSCR9 and GDH gene (Genbank: ACB59697.1)
was then constructed (Fig. 2). The cells were cultivated easily to
a high density (6.8–7.5 g cell dry weight (CDW) LÀ1) in a shaking
flask, and showed an activity of 1369.8 U g CDWÀ1 using 2a as sub-
strate. Different concentrations (250–1500 mM) of 2a were tested
using the co-expressed cell in the presence or absence of additional
NADP+. Conversion was slightly decreased without NADP+ at a 2a
concentration of 1000 mM (Table 3), likely because the NADP(H)
in the host was insufficient with high substrate loading. Almost
2.7. Biocatalytic process and scale-up
To telescope the catalytic property of the co-expressed E. coli/
RtSCR9-GDH, the process of asymmetric reduction of 2a was mon-
itored under the optimized conditions. The excellent conversion
(>99%) could be achieved in the presence of up to 1000 mM 2a
within 240 min, and the plateau emerged after 120 min (Fig. 4).
To further test the potential of RtSCR9 for industrial application,
E. coli/RtSCR9-GDH-catalyzed asymmetric reduction of 2a was per-
formed at a 200-g scale with a 1000 mM substrate loading over a
240 min period. As a result, 2a was almost completely converted,
and the desired (S)-3a was obtained by a simple extraction/distil-
lation process in a yield of 92.1% with 99.9% ee, representing an
outstanding space-time yield per gram of biomass (DCW) of
22.9 mmol LÀ1 hÀ1 gÀ1 DCW.
complete conversion was found using
mmol/mmol) ratio of 0.75 and excess of the cofactor was not
necessary. The further increase of 2a loading to 1500 mM resulted
in decrease of conversion but had no influence on
a
NADP+/substrate
(l
a
enantioselectivity.
2.8. Fabrication of (S)-6 from the RtSCR9-catalyzed product (S)-3a
To assess the feasibility of the ‘‘chemoenzymatic strategy”, sub-
sequent synthetic steps were carried out from the RtSCR9-
catalyzed product, (S)-3a. The preparation of respective 4 was
implemented by using 1-fluoronaphthalene via nucleophilic aro-
matic substitution. Potassium benzoate (0.2 equiv) was selected
as the catalyst. Sodium hydride (1.2 equiv) was added to create
an alkaline environment, facilitating the nucleophilic attack on 1-
fluoronaphalene (1.2 equiv), as outlined in Scheme 3. Petroleum
ether was utilized to remove 1-fluoronaphalene and could be
reused through distillation. The resulting aqueous phase contain-
ing 4 was taken up into toluene and an yield of 92% was obtained
for 4 from (S)-3a. Notably, increased reaction temperature or pro-
longed reaction time led to a decrease of the yield.
Scheme 2. General pathway for the synthesis of thiophene-ring containing
ketones. Reagents and conditions: (i) 2b: (HCHO)n, H2NCH3ÁHCl, ethanol, reflux
for 9 h with 1, 71%; (ii) 2c: AlCl3, Cl(CO)CH2CH2Cl, dichloromethane, À5 °C, 1 h with
2, 91%; (iii) 2d: NaH, DEK, toluene, reflux for 2 h with 1, 89%; (iv) 2e: CH3CN,
CH3CH2C(CH3)2OK, THF, rt, 24 h with 3, 71%.

X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
85
(a)
(b)
I
Nco
T7-1
1
RtSCR9
lac1
d III
Hin
I
Nde
T7-2
pCDFDuet-1-RtSCR9-GDH
5196bp
GDH
CDF ori
I
Xho
Sm
Fig. 2. Schematic presentation of co-expressed plasmid pCDFDuet-1-RtSCR9-GDH (a) and SDS-PAGE analysis (b). Lane 1, Marker; Lane 2, E. coli/pCDFDuet-1; Lane 3 E. coli/
pET28a-GDH; Lane 4, E. coli/pET28a-RtSCR9; Lane 5, soluble fraction of E. coli/pCDFDuet-1-RtSCR9-GDH.
Table 3
the highest yield (92.1%) and enantioselectivity (99.9% ee). Che-
Asymmetric reduction of 2a using co-expressed E. coli/RtSCR9-GDH with and without
additional NADP+.a
moenzymatic strategy was successfully accomplished together
with the subsequent nucleophilic aromatic substitution and N-
demethylation from enzymatic product (S)-3a. The final product
(S)-6 was obtained with the overall yield of 60.2% from 2-
acethylthiophene through four steps in excellent enantioselectivity
(>98.5% ee), representing a practical route to (S)-duloxetine.
4. Material and methods
Entry
Con. (mM)
NADP+
(lmmol/mmol 2a)
Conv. (%)
ee (%)
4.1. General
1
2
3
4
5
6
7
8
9
250
250
500
–
0.75
–
0.75
–
0.75
1
–
0.75
1
>99
>99
>99
>99
94
>99
>99
64
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
>99(S)
The glucose dehydrogenase was obtained as previously
described [45]. The pGEM-T (Promega, Beijing, China) and pET28a
(Novagen, Darmstadt, Germany) were selected for cloning and
expression. The pCDFDuet-1 (Novagen, Darmstadt, Germany) was
used as the co-expression plasmid. T4 DNA ligase, restriction
enzymes, and DNA polymerases were purchased from Vazyme
(Nanjing, China). NAD(P)H and NAD(P)+ were obtained from Roche
(Karlsruhe, Germany). Antibiotics ampicillin, kanamycin, strepto-
mycin, and inducer isopropylthio-b-galactoside (IPTG) were pro-
vided from Sango Biotech (Shanghai, China). The remaining
chemicals were gained from commercial sources and used without
further purification.
500
1000
1000
1000
1500
1500
1500
85
85
10
a
Reaction conditions: substrate 2a (250–1500 mM), glucose (6 mmol/(mmol
2a)), NADP+ (0–1
lmmol/(mmol 2a)), sonicated cells of E. coli/RtSCR9-GDH
(equivalent to 10 g CDW LÀ1), 30 °C for 8 h.
1H and 13C NMR were recorded on a Bruker AVANCE III (Bruker,
Switzerland). Mass spectra (MS) were analyzed on an Agilent 6210
TOF LC/MS (Agilent, USA). Analytical thin layer chromatography
(TLC) was performed on aluminum-backed 0.2 mm thick silica
gel 60 F254 plates. Absorption measurements were determined on
a SpectraMaxM5 (Molecular Devices, CA, USA). Optical rotator dis-
persion were carried out by Rudolph Autopol IV polarimeter
(Rudolph, USA). The ee values of bioproducts were determined on
Model LC-20AT HPLC (Shimadu, Japan) or GC-14C (Shimadu,
Japan). The native molecular mass of enzyme was evaluated on a
N-Demethylation was subsequently conducted by protecting 4
as a carbamate intermediate 5. The N-substituted group of 5 was
cleaved through alkaline hydrolysis. Different reaction conditions
were evaluated for these two consecutive steps, and it was found
that subjecting 4 to 1.2 equiv phenyl chloroformate in the presence
of N,N-diisopropylethylamine followed by aqueous workup, pro-
viding the desired intermediate 5. This compound (5) was taken
to the next step without purification to liberate the free amine
upon basification with aqueous solution of sodium hydroxide,
obtaining (S)-6 in 75.6% yield, slightly higher than that using ethyl
chroloformate as the protecting agent [56]. The ee of (S)-6 was
determined to be 98.5% by HPLC.
Discovery BIO GFC 150 (300 Â 7.8 mm, 3
lm) column (Sigma–
Aldrich, USA) equilibrated with 150 mM phosphate (pH 7.0) using
cytochrome c (12.4 kDa), ovalbumin (43 kDa), bovine serum albu-
min (66 kDa), and yeast alcohol dehydrogenase (150 kDa) as stan-
dard proteins.
3. Conclusions
In summary, we developed the most effective route for (S)-
duloxetine production. By utilizing RtSCR9 mined from newly iso-
lated R. toruloides ZJB2014212 for the asymmetric synthesis strat-
4.2. Cloning and expression of recombinant carbonyl reductase genes
Total RNA from R. toruloides ZJB2014212 was isolated using TRI-
zol Reagent (Ambion, USA). 1 mg of RNA was used for cDNA syn-
thesis employing the ReverTra Ace qPCR RT Kit (ToYoBo, Japan).
egy in the enantiodetermining step, the
c-amino alcohol, (S)-3a,
was obtained with high substrate loading (1000 mM) in so far

86
X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
(a)
(b)
120
100
80
60
40
20
0
120
100
80
60
40
20
0
120
120
100
80
60
40
20
0
Conv.
Conv.
ee
ee
100
80
60
40
20
0
20
25
30
35
40
45
6.0
6.5
7.0
7.5
pH
8.0
8.5
9.0
Temperature (oC)
(c)
(d)
120
100
80
60
40
20
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
120
100
80
60
40
20
0
Conv.
Conv.
ee
ee
0
1
2
3
4
5
6
7
8
9
0
5
10
15
20
25
-1
30
mmol G/(mmol 2a)
Cell dosage (g CDW L
)
Fig. 3. Effects of pH (a), temperature (b), glucose concentration (c), and cell dosage (d) on the asymmetric synthesis of (S)-3a using co-expressed E. coli/RtSCR9-GDH.
The rtscr genes encoding RtSCR1–RtSCR10 were amplified by
reverse transcription polymerase chain reaction (RT-PCR) using
primers with restriction sites as listed in Table S2. The DNA frag-
ments were inserted into pGEM-T to obtain pGEM-RtSCRs. The lin-
ear pET28a and target fragments were prepared by double-
digestion of pET28a and pGEM-RtSCRs (Nde I and Hind III for
RtSCR1 and RtSCR9; Nco I and Hind III for RtSCR2, RtSCR3, RtSCR4,
RtSCR5, RtSCR7, and RtSCR8; Nco I and Xho I for RtSCR6; Nde I and
Xho I for RtSCR10). Linear DNAs were linked to pET28a and trans-
formed into E. coli BL21 (DE3). Protein expression was preformed
as described previously [25], and checked on SDS-PAGE.
1500
1250
1000
750
500
250
0
100
80
60
40
20
0
Conversion
ee
Concentration of (S)-3a
4.3. Substrate synthesis
0
60
120
180
240
Time (min)
300
360
420
480
4.3.1. 3-(Dimethylamino)-1-(2-thienyl)-1-propanone hydrochloride
2a
Fig. 4. Time course of asymmetric reduction of 2a using co-expressed E. coli/
RtSCR9-GDH at a 1000 mM substrate loading. All experiments were assayed in
triplicate.
Yield: 94%, 19.4 g, white solid. 1H NMR (500 MHz, DMSO-d6):
d = 10.96 (s, 1H), 8.08 (m, 2H), 7.30 (dd, J = 6.5, 2.5 Hz, 1H), 3.60
Scheme 3. Chemical synthesis of (S)-duloxetine from the RtSCR9-catalyzed enantiopure (S)-3a.

X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
87
(t, J = 7.4 Hz, 2H), 3.37 (d, J = 7.4 Hz, 2H), 2.78 (s, 6H). 13C NMR
(126 MHz, DMSO-d6): d = 189.58, 142.76, 135.51, 134.09, 128.83,
51.41, 42.07, 33.39. MS (ESI): m/z 184.1 [M+HÀHCl]+.
volume of 0.5 mL was shaken at 30 °C with 200 rpm. After 8 h, the
obtained solution was adjusted to pH 11–12 using 6 M NaOH,
extracted with ethyl acetate and analyzed by HPLC.
4.3.2. 3-(Methylamino)-1-(2-thienyl)-propanone hydrochloride 2b
Yield: 71%, 14.8 g, slightly yellow powder. 1H NMR (500 MHz,
D2O): d = 7.99 (d, J = 3.9 Hz, 1H), 7.95 (d, J = 4.9 Hz, 1H), 7.28 (m,
1H), 3.56 (t, J = 6.2 Hz, 2H), 3.46 (t, J = 6.1 Hz, 2H), 2.81 (s, 3H),
2.62 (s, 1H). 13C NMR (126 MHz, D2O): d = 195.33, 144.26, 138.90,
137.61, 131.67, 46.69, 37.05, 35.76. MS (ESI): m/z 170.1 [M
+HÀHCl]+.
4.5. Enzyme activity
The recombinant E. coli cells of RtSCR9 were harvested and dis-
rupted by sonication and centrifuged for 20 min at 4 °C and
9000 rpm. The soluble fraction was applied to the Ni-NTA column
with equilibrating buffer (20 mM sodium dihydrogen phosphate,
300 mM sodium chloride, pH 8.0) followed by washing with bind-
ing buffer (20 mM imidazole, 20 mM sodium dihydrogen phos-
phate, 300 mM sodium chloride, pH 8.0) to remove the unbound
and weakly bound proteins. His-tagged target protein was eluted
with washing buffer containing 500 mM imidazole [40]. Enzyme
activity was determined spectrophotometrically by measuring
the consumption of NAD(P)H at 340 nm on SpectraMax M5. One
4.3.3. 3-Chloro-1-(2-thienyl)-1-propanone 2c
Yield: 91%, 15.9 g, colorless oil. 1H NMR (500 MHz, DMSO-d6):
d = 8.03 (m, 2H), 7.27 (m, 1H), 3.92 (t, J = 6.3 Hz, 2H), 3.50 (t,
J = 6.3 Hz, 2H). 13C NMR (126 MHz, DMSO-d6): d = 189.81, 143.37,
135.24, 133.82, 128.77, 40.92, 39.43. MS (ESI): m/z 175.1 [M+H]+.
unit (U) of enzyme activity was defined as 1 lmol of NAD(P)H
4.3.4. Ethyl 3-oxo-3-(2-thienyl)propanoate 2d
depleted per minute at 35 °C. Each mixture is consisted of sub-
strate (50 mM), NAD(P)H (0.5 mM), and purified RtSCR9
(0.1 mg mLÀ1) in phosphate buffer (100 mM, pH 7.0).
Yield: 89%, 17.6 g, colorless liquid. 1H NMR (500 MHz, DMSO-
d6): d = 8.07 (dd, J = 4.9, 1.2 Hz, 1H), 8.00 (dd, J = 3.8, 1.1 Hz, 1H),
7.27 (m, 1H), 4.12 (m, 4H), 1.18 (t, J = 7.1 Hz, 3H). 13C NMR
(126 MHz, DMSO-d6): d = 185.99, 167.22, 142.93, 136.06, 134.87,
128.87, 60.69, 45.57, 13.94. MS (ESI): m/z 199.1 [M+H]+.
The effects of pH and temperature on the activity and stereose-
lectivity of RtSCR9 were examined under the standard assay condi-
tions at different pHs (4.0–10.5) and temperatures (25–65 °C). The
thermostability of RtSCR9 was evaluated by preincubating the
enzyme in phosphate buffer (100 mM, pH 7.0) at different temper-
atures (35–65 °C) for required time intervals. The effect of metal
ions and EDTA on RtSCR9 activity was measured by adding metal
ion or EDTA–Na2 (2 mM) in the reaction mixture for 30 min at
35 °C.
The kinetic analysis of RtSCR9 was performed by measuring the
activity on the constant substrate concentration (50 mM) with dif-
ferent concentrations of NADH (0.1–10 mM) or NADPH (0.01–
0.5 mM). Similarly, the apparent Km value for substrate was exam-
ined by altering the concentration of substrate (0.1–50 mM) at a fix
NADPH concentration (0.5 mM). The apparent Km and Vmax values
of RtSCR9 were calculated using nonlinear regression to the
Michaelis–Menten equation.
4.3.5. 3-Oxo-3-(2-thienyl)propanenitrile 2e
Yield: 71%, 10.7 g, brown solid. 1H NMR (500 MHz, DMSO-d6):
d = 8.12 (d, J = 4.2 Hz, 1H), 7.98 (d, J = 2.6 Hz, 1H), 7.29 (t,
J = 4.4 Hz, 1H), 4.70 (s, 2H). 13C NMR (126 MHz, DMSO-d6):
d = 182.28, 140.94, 136.56, 135.15, 128.97, 115.44, 29.73. MS
(ESI): m/z 151.1 [M+H]+.
4.3.6. Racemic alcohols 3a–3e
3a. Yield: 7.4 g, 87%, white powder. 1H NMR (500 MHz, DMSO-
d6): d = 7.37 (dd, J = 4.9, 1.1 Hz, 1H), 6.94 (m, 2H), 5.80 (s, 1H), 4.87
(m, 1H), 2.30 (m, 2H), 2.13 (s, 6H), 1.80 (m, 2H). 13C NMR
(126 MHz, DMSO-d6): d = 150.74, 126.42, 123.81, 122.52, 67.41,
55.98, 45.18, 36.94. MS (ESI): m/z 186.2 [M+H]+. 3b. Yield: 7.9 g,
86%, white powder. 1H NMR (500 MHz, DMSO-d6): d = 7.36 (dd,
J = 4.9, 1.0 Hz, 1H), 6.93 (m, 2H), 4.92 (t, J = 6.3 Hz, 1H), 4.40 (m,
2H), 2.59 (m, 2H), 2.25 (m, 3H), 1.81 (m, 2H). 13C NMR (126 MHz,
DMSO-d6): d = 150.88, 126.43, 123.77, 122.42, 67.63, 48.28,
38.48, 35.70. MS (ESI): m/z 172.1 [M+H]+. 3c. Yield: 7.2 g, 82%,
oil. 1H NMR (500 MHz, DMSO-d6): d = 7.41 (dd, J = 3.9, 2.4 Hz,
1H), 6.98 (m, 2H), 5.82 (t, J = 4.4 Hz, 1H), 4.96 (dd, J = 8.7, 4.6 Hz,
1H), 3.75 (m, 1H), 3.65 (ddd, J = 10.7, 6.7, 5.5 Hz, 1H), 2.13 (m,
2H). 13C NMR (126 MHz, DMSO-d6): d = 149.63, 126.35, 124.28,
122.98, 65.43, 41.98, 40.43. MS (ESI): m/z 177.1 [M+H]+. 3d. Yield:
6.7 g, 66.7%, colorless liquid. 1H NMR (500 MHz, DMSO-d6):
d = 7.40 (dd, J = 4.9, 1.4 Hz, 1H), 6.97 (ddd, J = 8.4, 3.8, 2.1 Hz, 2H),
5.90 (d, J = 5.2 Hz, 1H), 5.21 (dt, J = 8.2, 5.0 Hz, 1H), 4.06 (t,
J = 7.1 Hz, 2H), 2.70 (m, 2H), 1.19 (q, J = 6.7 Hz, 3H). 13C NMR
(126 MHz, DMSO-d6): d = 170.15, 149.06, 126.58, 124.43, 123.07,
65.51, 59.88, 44.55, 14.05. MS (ESI): m/z 201.1 [M+H]+. 3e. Yield:
6.9 g, 91%, oil. 1H NMR (500 MHz, DMSO-d6): d = 7.45 (dt, J = 3.9,
1.9 Hz, 1H), 7.06 (m, 1H), 7.01 (dd, J = 5.0, 3.5 Hz, 1H), 6.36 (s,
1H), 5.15 (t, J = 5.7 Hz, 1H), 2.96 (m, 2H). 13C NMR (126 MHz,
DMSO-d6): d = 147.30, 126.75, 124.95, 123.66, 118.34, 64.70,
27.81. MS (ESI): m/z 154.1 [M+H]+.
4.6. Substrate specificity
The asymmetric reduction of varied thiophene-ring containing
ketones 2b–2e, related to 2a, to the corresponding enantioenriched
alcohols 3b–3e was investigated by using RtSCR9 as catalyst. Each
reaction mixture comprising phosphate buffer (100 mM, pH 7.0),
2b–2e (50–250 mM), glucose (278 mM), NADP+ (0.05 mM), puri-
fied RtSCR9 (0.1 mg mLÀ1), and GDH (0.1 mg mLÀ1) in a total vol-
ume of 10 mL was shaken at 35 °C with 200 rpm. Substrate 2c–
2e was dissolved in DMSO prior to dilution into buffer to give a
final DMSO concentration of 5% (v/v). After 4 h, the obtained mix-
ture of 2b was adjusted to pH 11–12 using 6 M NaOH, and
extracted with ethyl acetate. In contrast, the resulting mixture of
2c–2e was directly extracted with ethyl acetate. The final products
were analyzed by HPLC or GC (Table S5).
4.7. Co-expression of RtSCR9 and GDH genes
The pGEM-T-GDH was obtained and digested with Nde I and
Xho I restriction endonucleases as described previously [29]. The
resulting fragment was inserted into the MCSI site of pCDFDuet-
4.4. Screening of recombinant carbonyl reductases
1 vector to provide recombinant plasmid pCDFDuet-1-GDH.
RtSCR9 gene digested from pGEM-T-RtSCR9 with Nco I and Hind
III restriction endonucleases was then introduced into the MCS II
site of pCDFDuet-1-GDH vector, resulting in plasmid pCDFDuet-
1-RtSCR9-GDH. This co-expressed plasmid harboring RtSCR9 and
GDH genes was transformed into E. coli BL21 (DE3) and grown in
RtSCRs were screened against 2a through the formation of cor-
responding (S)-3a. Reaction mixture comprising phosphate buffer
(100 mM, pH 7.0), 2a (50 mM), glucose (278 mM), NAD(P)+
(0.05 mM), and sonicated cells of RtSCRs (10 g CDW LÀ1) in a total

88
X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
LB/streptomycin (50
genes were confirmed by SDS-PAGE.
l
g mLÀ1). The expression levels of the two
120.18, 107.24, 73.68, 55.13, 45.19, 36.28. MS (ESI): m/z 312.2 [M
+H]+. [ 25 = + 136.8° (c = 10 mg/mL, methanol).
a]
D
4.8. Optimization of reaction conditions in the asymmetric synthesis of
(S)-3a using co-expressed E. coli/RtSCR9-GDH
4.10.2. (S)-N-methyl-3-(1-naphthalenyloxy)-3-(2-thienyl)
propanamine 6
A mixture of residue 4 obtained above in toluene (250 mL) was
stirred for 10 min. The temperature was increased at 45 °C and
diisopropylethyl-amine (7.5 g) was added with stirring for further
10 min. Subsequently, phenyl chloroformate (52.8 g, 330 mmol)
was added dropwise with stirring at 55 °C for 2 h. Then, the mix-
ture was cooled to 25 °C and 1% sodium bicarbonate (150 mL)
was added and stirred for 10 min. The organic layer was succes-
sively washed with 0.5 N hydrochloric acid solution, 1% sodium
bicarbonate and water. The organic phase was concentrated and
dissolved in DMSO (300 mL) at 45 °C followed by adding sodium
hydroxide (68 g/100 g water). The basic mixture was stirred at
55 °C for 8 h. The obtained mixture was treated as described above,
yielding the free base (S)-6. Yield: 56.8 g, 75.6%, amber oil. 1H NMR
(500 MHz, DMSO-d6): d = 9.40 (s, 1H), 8.28 (m, 1H), 7.86 (m, 1H),
7.53 (m, 2H), 7.46 (dd, J = 4.5, 3.5 Hz, 2H), 7.34 (t, J = 8.0 Hz, 1H),
7.29 (m, 1H), 7.09 (d, J = 7.7 Hz, 1H), 6.99 (dd, J = 5.0, 3.5 Hz, 1H),
6.19 (dd, J = 7.4, 5.5 Hz, 1H), 3.07 (m, 2H), 2.58 (m, 4H), 2.41 (qd,
J = 10.8, 5.5 Hz, 1H). 13C NMR (126 MHz, DMSO-d6): d = 152.17,
143.59, 134.08, 127.43, 126.75, 126.41, 126.09, 125.83, 125.41,
121.63, 120.49, 107.53, 72.61, 45.00, 34.36, 32.26. MS (ESI): m/z
Several reaction parameters (e.g., glucose/substrate ratio, bio-
catalyst loading, reaction pH, temperature, etc.) affected the bio-
transformation
were
evaluated
by
adding
different
concentrations of glucose (0.5–9 mmol glucose/(mmol 2a)) and
sonicated cell dosages (1–30 g DCW LÀ1). The effects of reaction
pH and temperature on the bioreduction were performed at vari-
ous temperatures (20–45 °C) and different pHs (6.0–9.0). Unless
otherwise stated, each assay was stirred at 30 °C for 4 h.
4.9. Time-course and scale-up of the biocatalytic process
Reaction components 2a (1000 mM), glucose (6000 mM),
NADP+ (0.75 mM), sonicated cells of co-expressed E. coli/RtSCR9-
GDH (10 g DCW LÀ1) were combined in a total volume of 100 mL
phosphate buffer (100 mM, pH 7.0) and proceeded at 30 °C with
200 rpm shaking. The pH was kept constant at 7.0 using 2 M NaOH
during the reaction. Aliquot of the obtained mixture was with-
drawn at predetermined times, followed by adjusting pH to 11–
12. The residual was extracted with ethyl acetate and analyzed
by HPLC.
298.1 [M+H]+. [ 25 = + 93.3° (c = 10 mg/mL, methanol).
a]
D
In a 2-L bioreactor, 2a (1000 mM), glucose (6000 mM), NADP+
(0.75 mM), and sonicated cells of co-expressing E. coli/RtSCR9-
GDH (10 g DCW LÀ1) were mixed in a total volume of 1 L phos-
phate buffer (100 mM, pH 7.0). The resulting mixture was stirred
at 30 °C with 200 rpm shaking for 240 min. The pH was maintained
at 7.0 with 2 M NaOH during the reaction. After reaching comple-
tion, the pH was adjusted to 11–12 followed by extraction using
ethyl acetate. The combined organic phases were dried over anhy-
drous sodium sulfate and concentrated to give (S)-3a. Yield: 92.1%.
1H NMR (500 MHz, DMSO-d6): d = 7.36 (dd, J = 4.9, 1.2 Hz, 1H), 6.93
(m, 2H), 5.79 (s, 1H), 4.86 (m, 1H), 2.30 (qt, J = 12.0, 7.1 Hz, 2H),
2.12 (s, 6H), 1.79 (m, 2H). 13C NMR (126 MHz, DMSO-d6):
d = 150.72, 126.40, 123.79, 122.51, 67.39, 55.96, 45.15, 36.92.
Acknowledgements
This study was financially supported by the National Basic
Research Program of China (973 Program) (No. 2011CB710800),
and Natural Science Foundation of Zhejiang Province, China (No.
R3110155).
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bioorg.2016.02.
002.
[a
]
D
25 = À4.9° (c = 10 mg/mL, ethyl acetate).
4.10. Preparation of (S)-6 from the RtSCR9-catalyzed product (S)-3a
References
4.10.1. (S)-N,N-dimethyl-3-(1-naphthalenyloxy)-3-(2-thienyl)
propanamine 4
[1] D.J. Goldstein, C. Mallinckrodt, Y.L. Lu, M.A. Demitrack, Duloxetine in the
treatment of major depressive disorder: a double-blind clinical trial, J. Clin.
Psychiatr. 63 (2002) 225–231.
To a solution of (S)-3a (55.5 g, 300 mmol) dissolved in dry
DMSO (120 mL) was carefully added sodium hydride (60% in min-
eral oil, 14.4 g, 360 mmol) at ambient temperature in portions with
stirring for 30 min. Potassium benzoate (9.6 g, 60 mmol) was com-
bined with an additional stirring for 30 min. Then, 1-
fluoronaphthalene (52.5 g, 360 mmol) was added dropwise and
the reaction mixture was heated at 65 °C for 8 h. After complete
reaction, the residual was slowly poured into the ice water
(300 mL) followed by adding acetic acid to pH 4.8. The mixture
was extracted with petroleum ether (3 Â 300 mL). The residual
aqueous phase was treated with 50% sodium hydroxide to pH
11–12, and extracted by toluene (3 Â 300 mL). The combined
toluene phase was washed with 1% sodium bicarbonate, dried over
anhydrous sodium sulfate, and concentrated to afford product 4.
Yield: 85.8 g, 92%, amber oil. 1H NMR (500 MHz, DMSO-d6):
d = 8.26 (m, 1H), 7.84 (dt, J = 6.7, 2.9 Hz, 1H), 7.52 (m, 2H), 7.43
(dd, J = 8.4, 4.6 Hz, 2H), 7.34 (m, 1H), 7.22 (m, 1H), 7.04 (d,
J = 7.6 Hz, 1H), 6.97 (td, J = 4.7, 2.9 Hz, 1H), 5.92 (d, J = 5.9 Hz,
1H), 2.38 (m, 2H), 2.33 (m, 1H), 2.13 (s, 6H), 2.09 (m, 1H). 13C
NMR (126 MHz, DMSO-d6): d = 152.74, 144.84, 134.11, 128.84,
128.13, 127.39, 126.57, 126.30, 125.91, 125.49, 125.27, 121.57,
[2] J.C. Chappell, G. Eisenhofer, M.J. Owens, H. Haber, D.R. Lachno, R.A. Dean, M.P.
Knadler, C.B. Nemeroff, M.I. Mitchell, M.J. Detke, S. Iyengar, B. Pangallo, E.D.
Lobo, Effects of duloxetine on norepinephrine and serotonin transporter
activity in healthy subjects, J. Clin. Psychopharm. 34 (2014) 9–16.
[3] D.T. Wong, F.P. Bymaster, D.A. Mayle, L.R. Reid, J.H. Krushinski, D.W. Robertson,
Ly248686,
a new inhibitor of serotonin and norepinephrine uptake,
Neuropsychopharmacology 8 (1993) 23–33.
[4] A.D.J. Calow, E. Fernandez, A. Whiting, Total synthesis of fluoxetine and
duloxetine through an in situ imine formation/borylation/transimination and
reduction approach, Org. Biomol. Chem. 12 (2014) 6121–6127.
[5] E.N. Xun, Z. Wang, J.M. Zhao, J.X. Guo, Kinetic resolution of (R, S)-2-(2-chloro-1-
hydroxyethyl) thiophene via immobilizing lipase from Alcaligenes sp. onto
magnetic nanoparticles, J. Chem. Technol. Biotechnol. 90 (2015) 492–499.
[6] R.K. Rej, T. Das, S. Hazra, S. Nanda, Chemoenzymatic asymmetric synthesis of
fluoxetine, atomoxetine, nisoxetine, and duloxetine, Tetrahedron-Asymmetry
24 (2013) 913–918.
[7] R.C. Simon, F.G. Mutti, W. Kroutil, Biocatalytic synthesis of enantiopure
building blocks for pharmaceuticals, Drug Discov. Today Technol. 10 (2013)
e37–e44.
[8] P. Motloch, I. Valterova, M. Kotora, Enantioselective allylation of thiophene-2-
carbaldehyde: formal total synthesis of duloxetine, Adv. Synth. Catal. 356
(2014) 199–204.
[9] K. Sakai, R. Sakurai, A. Yuzawa, Y. Kobayashi, K. Saigo, Resolution of 3-
(methylamino)-1-(2-thienyl)propan-1-ol,
a
new key intermediate for
duloxetine, with (S)-mandelic acid, Tetrahedron-Asymmetry 14 (2003)
1631–1636.

X. Chen et al. / Bioorganic Chemistry 65 (2016) 82–89
89
[10] J. Majer, P. Kwiatkowski, J. Jurczak, Highly enantioselective Friedel-crafts
reaction of thiophenes with glyoxylates: formal synthesis of duloxetine, Org.
Lett. 11 (2009) 4636–4639.
[35] M. Wada, A. Yoshizumi, Y. Furukawa, H. Kawabata, M. Ueda, H. Takagi, S.
Nakamori, Cloning and overexpression of the Exiguobacterium sp. F42 gene
encoding
a
new short chain dehydrogenase, which catalyzes the
[11] H.Y. Bae, J.H. Sim, J.W. Lee, B. List, C.E. Song, Organocatalytic enantioselective
decarboxylative aldol reaction of malonic acid half thioesters with aldehydes,
Angew. Chem. Int. Ed. 52 (2013) 12143–12147.
stereoselective reduction of ethyl 3-oxo-3-(2-thienyl)propanoate to ethyl
(S)-3-hydroxy-3-(2-thienyl)propanoate, Biosci. Biotechnol. Biochem. 68
(2004) 1481–1488.
[12] Y. Suzuki, M. Iwata, R. Yazaki, N. Kumagai, M. Shibasaki, Concise
enantioselective synthesis of duloxetine via direct catalytic asymmetric aldol
reaction of thioamide, J. Org. Chem. 77 (2012) 4496–4500.
[13] H.L. Liu, B.H. Hoff, T. Anthonsen, Chemo-enzymatic synthesis of the
antidepressant duloxetine and its enantiomer, Chirality 12 (2000) 26–29.
[14] A. Kamal, G.B.R. Khanna, R. Ramu, T. Krishnaji, Chemoenzymatic synthesis of
duloxetine and its enantiomer: lipase-catalyzed resolution of 3-hydroxy-3-(2-
thienyl) propanenitrile, Tetrahedron Lett. 44 (2003) 4783–4787.
[36] C.G. Tang, H. Lin, C. Zhang, Z.Q. Liu, T. Yang, Z.L. Wu, Highly enantioselective
bioreduction of N-methyl-3-oxo-3-(thiophen-2-yl) propanamide for the
production of (S)-duloxetine, Biotechnol. Lett. 33 (2011) 1435–1440.
[37] Z.Q. Ren, Y. Liu, X.Q. Pei, H.B. Wang, Z.L. Wu, Bioreductive production of
enantiopure (S)-duloxetine intermediates catalyzed with ketoreductase
ChKRED15, J. Mol. Catal. B – Enzym. 113 (2015) 76–81.
[38] Z. Zhu, S. Zhang, H. Liu, H. Shen, X. Lin, F. Yang, Y.J. Zhou, G. Jin, M. Ye, H. Zou, Z.
K. Zhao,
A multi-omic map of the lipid-producing yeast Rhodosporidium
[15] A. Kamal, M.S. Malik, A.A. Shaik, S. Azeeza, Lipase mediated resolution of
gamma-azidoalcohols in aqueous and organic media: synthesis of (R)- and (S)-
fluoxetine and duloxetine, J. Mol. Catal. B – Enzym. 58 (2009) 132–137.
[16] S.I.Y.S.M. Abramov, Process for the preparation of (S)-(À)-N,N-dimethyl-3-(2-
thienyl)-3-hydroxypropananine, a duloxetine intermediate, US7560573, 2009.
[17] Y. Fujima, M. Ikunaka, T. Inoue, J. Matsumoto, Synthesis of (S)-3-(N-
methylamino)-1-(2-thienyl)propan-1-ol: revisiting Eli Lilly’s resolution-
racemization-recycle synthesis of duloxetine for its robust processes, Org.
Process Res. Dev. 10 (2006) 905–913.
toruloides, Nat. Commun. 3 (2012) 1112.
[39] C. Filling, K.D. Berndt, J. Benach, S. Knapp, T. Prozorovski, E. Nordling, R.
Ladenstein, H. Jornvall, U. Oppermann, Critical residues for structure and
catalysis in short-chain dehydrogenases/reductases, J. Biol. Chem. 277 (2002)
25677–25684.
[40] Z.Q. Liu, P.J. Baker, F. Cheng, Y.P. Xue, Y.G. Zheng, Y.C. Shen, Screening and
improving the recombinant nitrilases and application in biotransformation of
iminodiacetonitrile to iminodiacetic acid, PLoS ONE 8 (2013) e67197.
[41] N. Itoh, S. Kariya, J. Kurokawa, Efficient PCR-based amplification of diverse
alcohol dehydrogenase genes from metagenomes for improving biocatalysis:
screening of gene-specific amplicons from metagenomes, Appl. Environ.
Microbiol. 80 (2014) 6280–6289.
[42] A. Pennacchio, V. Sannino, G. Sorrentino, M. Rossi, C.A. Raia, L. Esposito,
Biochemical and structural characterization of recombinant short-chain NAD
(H)-dependent dehydrogenase/reductase from Sulfolobus acidocaldarius highly
enantioselective on diaryl diketone benzil, Appl. Microbiol. Biotechnol. 97
(2013) 3949–3964.
[43] K. Niefind, J. Müller, B. Riebel, W. Hummel, D. Schomburg, The crystal structure
of R-specific alcohol dehydrogenase from Lactobacillus brevis suggests the
structural basis of its metal dependency, J. Mol. Biol. 327 (2003) 317–328.
[44] L.J. Wang, C.X. Li, Y. Ni, J. Zhang, X. Liu, J.H. Xu, Highly efficient synthesis of
chiral alcohols with a novel NADH-dependent reductase from Streptomyces
coelicolor, Bioresour. Technol. 102 (2011) 7023–7028.
[18] X. Chen, Y.G. Zheng, Z.Q. Liu, L.H. Sun, Stereoselective determination of 2-
benzamidomethyl-3-oxobutanoate
and
methyl-2-benzoylamide-3-
hydroxybutanoate by chiral high-performance liquid chromatography in
biotransformation, J. Chromatogr. B 974 (2015) 57–64.
[19] J. Deeter, J. Frazier, G. Staten, M. Staszak, L. Weigel, Asymmetric-synthesis and
absolute stereochemistry of Ly248686, Tetrahedron Lett. 31 (1990) 7101–7104.
[20] D. Liu, W.Z. Gao, C.J. Wang, X.M. Zhang, Practical synthesis of enantiopure
gamma-amino alcohols by rhodium-catalyzed asymmetric hydrogenation of
beta-secondary-amino ketones, Angew. Chem. Int. Ed. 44 (2005) 1687–1689.
[21] E. Buitrago, H. Lundberg, H. Andersson, P. Ryberg, H. Adolfsson, High
throughput screening of
a catalyst library for the asymmetric transfer
hydrogenation of heteroaromatic ketones: formal syntheses of (R)-fluoxetine
and (S)-duloxetine, ChemCatChem 4 (2012) 2082–2089.
[22] V. Ratovelomanana-Vidal, C. Girard, R. Touati, J.P. Tranchier, B. Ben Hassine, J.P.
Genet, Enantioselective hydrogenation of beta-keto esters using chiral
diphosphine-ruthenium complexes: optimization for academic and
industrial purposes and synthetic applications, Adv. Synth. Catal. 345 (2003)
261–274.
[45] Z.Y. You, Z.Q. Liu, Y.G. Zheng, Characterization of a newly synthesized carbonyl
reductase and construction of a biocatalytic process for the synthesis of ethyl
(S)-4-chloro-3-hydroxybutanoate with high space-time yield, Appl. Microbiol.
Biotechnol. 98 (2014) 1671–1680.
[23] R.D. Mahale, S.P. Chaskar, K.E. Patil, G.C. Maikap, M.K. Gurjar, Corey-Itsuno
reduction of ketones: a development of safe and inexpensive process for
synthesis of some API intermediates, Org. Process Res. Dev. 16 (2012) 710–713.
[24] J.N. Zhou, Q. Fang, Y.H. Hu, L.Y. Yang, F.F. Wu, L.J. Xie, J. Wu, S.J. Li, Copper(II)-
catalyzed enantioselective hydrosilylation of halo-substituted alkyl aryl and
heteroaryl ketones: asymmetric synthesis of (R)-fluoxetine and (S)-duloxetine,
Org. Biomol. Chem. 12 (2014) 1009–1017.
[25] X. Chen, Z.Q. Liu, J.F. Huang, C.P. Lin, Y.G. Zheng, Asymmetric synthesis of
optically active methyl-2-benzamido-methyl-3-hydroxy-butyrate by robust
short-chain alcohol dehydrogenases from Burkholderia gladioli, Chem.
Commun. 51 (2015) 12328–12331.
[46] A. Kamal, M.S. Malik, A.A. Shaik, S. Azeeza, Synthesis of enantiomerically pure
gamma-azidoalcohols by lipase-catalyzed transesterification, Tetrahedron-
Asymmetry 19 (2008) 1078–1083.
[47] D.M. Zhu, B.E. Rios, J.D. Rozzell, L. Hua, Evaluation of substituent effects on
activity and enantioselectivity in the enzymatic reduction of aryl ketones,
Tetrahedron-Asymmetry 16 (2005) 1541–1546.
[48] G.C. Xu, H.L. Yu, X.Y. Zhang, J.H. Xu, Access to optically active aryl halohydrins
using a substrate-tolerant carbonyl reductase discovered from Kluyveromyces
thermotolerans, ACS Catal. 2 (2012) 2566–2571.
[49] H. Groger, F. Chamouleau, N. Orologas, C. Rollmann, K. Drauz, W. Hummel, A.
Weckbecker, O. May, Enantioselective reduction of ketones with ‘‘Designer
cells” at high substrate concentrations: highly efficient access to
functionalized optically active alcohols, Angew. Chem. Int. Ed. 45 (2006)
5677–5681.
[26] W.P. Dijkman, D.E. Groothuis, M.W. Fraaije, Enzyme-catalyzed oxidation of 5-
hydroxymethylfurfural to furan-2,5-dicarboxylic acid, Angew. Chem. Int. Ed.
53 (2014) 6515–6518.
[27] N. Reinhardt, J. Fischer, R. Coppi, E. Blum, W. Brandt, B. Drager, Substrate
flexibility and reaction specificity of tropinone reductase-like short-chain
dehydrogenases, Bioorg. Chem. 53 (2014) 37–49.
[50] F. Hollmann, I.W.C.E. Arends, K. Buehler, Biocatalytic redox reactions for
organic synthesis: nonconventional regeneration methods, ChemCatChem 2
(2010) 762–782.
[28] B.H. Hoff, E. Sundby, Preparation of pharmaceutical important fluorinated 1-
arylethanols using isolated enzymes, Bioorg. Chem. 51 (2013) 31–47.
[29] Z.Q. Liu, J.J. Ye, Z.Y. Shen, H.B. Hong, J.B. Yan, Y. Lin, Z.X. Chen, Y.G. Zheng, Y.C.
Shen, Upscale production of ethyl (S)-4-chloro-3-hydroxybutanoate by using
carbonyl reductase coupled with glucose dehydrogenase in aqueous-organic
solvent system, Appl. Microbiol. Biotechnol. 99 (2015) 2119–2129.
[30] J.A. Friest, Y. Maezato, S. Broussy, P. Blum, D.B. Berkowitz, Use of a robust
dehydrogenase from an archael hyperthermophile in asymmetric catalysis-
dynamic reductive kinetic resolution entry into (S)-profens, J. Am. Chem. Soc.
132 (2010) 5930–5931.
[51] S. Ludeke, M. Richter, M. Muller, Stereoselective synthesis of three isomers of
tert-butyl
5-hydroxy-4-methyl-3-oxohexanoate
through
alcohol
dehydrogenase-catalyzed dynamic kinetic resolution, Adv. Synth. Catal. 351
(2009) 253–259.
[52] S. Kara, J.H. Schrittwieser, F. Hollmann, M.B. Ansorge-Schumacher, Recent
trends and novel concepts in cofactor-dependent biotransformations, Appl.
Microbiol. Biotechnol. 98 (2014) 1517–1529.
[53] M. Kataoka, K. Kita, M. Wada, Y. Yasohara, J. Hasegawa, S. Shimizu, Novel
bioreduction system for the production of chiral alcohols, Appl. Microbiol.
Biotechnol. 62 (2003) 437–445.
[31] P. Soni, U.C. Banerjee, Biotransformations for the production of the chiral drug
[54] H.M. Ma, L.L. Yang, Y. Ni, J. Zhang, C.X. Li, G.W. Zheng, H.Y. Yang, J.H. Xu,
Stereospecific reduction of methyl o-chlorobenzoylformate at 300 g LÀ1
without additional cofactor using a carbonyl reductase mined from Candida
glabrata, Adv. Synth. Catal. 354 (2012) 1765–1772.
[55] N.D. Shen, Y. Ni, H.M. Ma, L.J. Wang, C.X. Li, G.W. Zheng, J. Zhang, J.H. Xu,
Efficient synthesis of a chiral precursor for angiotensin-converting enzyme
(ACE) inhibitors in high space-time yield by a new reductase without external
cofactors, Org. Lett. 14 (2012) 1982–1985.
(S)-duloxetine catalyzed by
a novel isolate of Candida tropicalis, Appl.
Microbiol. Biotechnol. 67 (2005) 771–777.
[32] S. Pankaj, H. Kansal, U.C. Banerjee, Optimization of process parameters for the
production of carbonyl reductase by Candida viswanathii in a laboratory-scale
fermentor, J. Ind. Microbiol. Biotechnol. 35 (2008) 167–173.
[33] O. Zhimin, Z. Haibing, T. Lan, Z. Wei, Y. Gensheng, Asymmetric synthesis of
duloxetine intermediate (S)-(À)-3-N-methylamino-1-(2-thienyl)-1-propanol
using immobilized Saccharomyces cerevisiae in liquid-core sodium
alginate/chitosan/sodium alginate microcapsules, Bioprocess Biosyst. Eng. 37
(2014) 2243–2250.
[56] A. Traff, R. Lihammar, J.E. Backvall,
A chemoenzymatic dynamic kinetic
resolution approach to enantiomerically pure (R)- and (S)-duloxetine, J. Org.
Chem. 76 (2011) 3917–3921.
[34] Y.J. Wang, X.Q. Liu, X. Luo, Z.Q. Liu, Y.G. Zheng, Cloning, expression and
enzymatic characterization of an aldo-keto reductase from Candida albicans
XP1463, J. Mol. Catal. B – Enzym. 122 (2015) 44–50.