Chemoenzymatic synthesis of (R)- and (S)-propranolol using an engineered epoxide hydrolase with a high turnover number

Article abstract of DOI:10.1016/j.molcatb.2015.10.005

Simultaneous synthesis of the R- and S-enantiomers of biologically active propranolol, a typical β-blocker, was achieved in high optical purity via an epoxide hydrolase-catalyzed resolution of racemic α-naphthyl glycidyl ether (rac-1). A preparative resolution of 100 g/L rac-1 was accomplished with high enantioselectivity (E = 92) using a variant of Bacillus megaterium epoxide hydrolase (BmEH_F128T). A biphasic system (isopropyl ether/isooctane/aqueous) was used, in which the product 3-(1′-naphthyloxy)-propane-1,2-diol (2) precipitated instantly, facilitating its separation and increasing the enantiopurity of (R)-2 (>99.5% ee). This enzymatic resolution had a total turnover number of 70,000, affording enantiopure epoxide (S)-1 (>99% ee) and 1,2-diol (R)-2 (>99% ee) in 45.3% and 42.4% yields, respectively. (R)-2 and (S)-1 were subsequently converted to (R)- and (S)-propranolol (3) (>99% ee) in overall yields of 31.4% and 44.8%. To the best of our knowledge, this is the best case for enzymatic resolution of rac-1 using an epoxide hydrolase, giving high space-time yields [136 g L^-1 days^-1 for (S)-1 and 139 g L^-1 days^-1 for (R)-2] under mild reaction conditions. It provides a new and eco-friendly route that complements other methods using organometallic catalysts.

Full text of DOI:10.1016/j.molcatb.2015.10.005

Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Chemoenzymatic synthesis of (R)- and (S)-propranolol using an
engineered epoxide hydrolase with a high turnover number
Xu-Dong Konga,b, Hui-Lei Yu , Sheng Yang , Jiahai Zhou , Bu-Bing Zeng , Jian-He Xu
a
c
b
d
a,∗
a
State Key Laboratory of Bioreactor Engineering, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China
Key Laboratory of Bio-organic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 345 Lingling Road,
Shanghai 200032, China
b
c
Laboratory of Molecular Microbiology, Institute of Plant Physiology and Ecology, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences,
3
00 Fenglin Road, Shanghai 200032, China
Shanghai Key Laboratory of New Drug Design, East China University of Science and Technology, 130 Meilong Road, Shanghai 200237, China
d
a r t i c l e i n f o
a b s t r a c t
Article history:
Simultaneous synthesis of the R- and S-enantiomers of biologically active propranolol, a typical ␤-blocker,
was achieved in high optical purity via an epoxide hydrolase-catalyzed resolution of racemic ␣-naphthyl
glycidyl ether (rac-1). A preparative resolution of 100 g/L rac-1 was accomplished with high enantiose-
Received 8 August 2015
Received in revised form 7 October 2015
Accepted 9 October 2015
lectivity (E = 92) using a variant of Bacillus megaterium epoxide hydrolase (BmEHF128T). A biphasic system
ꢀ
(
isopropyl ether/isooctane/aqueous) was used, in which the product 3-(1 -naphthyloxy)-propane-1,2-
Keywords:
diol (2) precipitated instantly, facilitating its separation and increasing the enantiopurity of (R)-2 (>99.5%
ee). This enzymatic resolution had a total turnover number of 70,000, affording enantiopure epoxide
Bacillus megaterium epoxide hydrolase
Biocatalytic resolution
Biphasic reaction
Naphthyl glycidyl ether
Chemoenzymatic synthesis
Beta-blocker
(S)-1 (>99% ee) and 1,2-diol (R)-2 (>99% ee) in 45.3% and 42.4% yields, respectively. (R)-2 and (S)-1 were
subsequently converted to (R)- and (S)-propranolol (3) (>99% ee) in overall yields of 31.4% and 44.8%. To
the best of our knowledge, this is the best case for enzymatic resolution of rac-1 using an epoxide hydro-
−1
−1
−1
−1
lase, giving high space-time yields [136 g L days for (S)-1 and 139 g L days for (R)-2] under mild
reaction conditions. It provides a new and eco-friendly route that complements other methods using
organometallic catalysts.
Propranolol
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
allowed the large scale enzymatic production of enantiopure epox-
ides with potential industrial applications [11–13]. Among which,
Deregnaucourt et al. developed a liquid–liquid two-phase method-
ology to improve the Aspergillus niger epoxide hydrolase-catalyzed
resolution of styrene oxide derivatives, affording a very high sub-
strate loading of 1.8 M (360 g/L) [11]. However, the biocatalytic
resolution of bulky epoxides (e.g., phenyl glycidyl ether derivatives)
with high substrate loading and easy separation remains challeng-
ing. In our previous study, the substrate specificity of an epoxide
hydrolase from Bacillus megaterium (BmEH) was engineered with
structure-guided redesign [14]. The resultant variants of BmEH
have a diverse substrate spectrum and can be used for the kinetic
resolution of various bulky aryl glycidyl ethers, which are unreac-
tive with other reported EHs. Amongst them, BmEHF128T was found
to have an activity of 4.0 U/mg cell free extract (CFE) and a good
enantioselectivity (E > 50) for ␣-naphthyl glycidyl ether (1).
Optically active pharmaceuticals play an important role in mod-
ern medicine, because they tend to have less side effects and higher
efficacy than their racemic mixtures. Various ␤-blockers, antidia-
betic and anti-HIV drugs can be synthesized from chiral epoxide
precursors [1–4] and in the past few decades, both chemical and
biological processes for their preparation have been developed
[
5–8].
Hydrolytic resolutions of racemic epoxides using epoxide
hydrolases (EHs) have been widely studied as these enzymes
are cofactor-independent, relatively stable and are suitable for a
wide spectrum of substrates [9,10]. These enzymes catalyze the
hydrolytic ring opening of epoxides in an enantioselective way,
forming optically active epoxides and vicinal diols. Several exam-
ples of heterologous overexpression and mutagenesis of EHs have
The bulky epoxide rac-1 is a precursor of the classic ␤-blocker,
ꢀ
propranolol [1-(1-methylethylamino)-3-(1 -naphthyloxy) propan-
2
-ol, 3], for which the S-enantiomer is 100 times more potent than
∗ Corresponding author.
E-mail address: jianhexu@ecust.edu.cn (J.-H. Xu).
the R-enantiomer in blocking beta adrenergic receptors. The R-
http://dx.doi.org/10.1016/j.molcatb.2015.10.005
1
381-1177/© 2015 Elsevier B.V. All rights reserved.

2
76
X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
enantiomer is also known to act as a contraceptive [15,16]. Several
biocatalytic routes have been reported for the preparation of enan-
tiopure precursors of 3 by using lipase or ketone reductase [17–19],
however, epoxide hydrolases were rarely reported to possess a
comparative activity for bulky epoxides like rac-1 [20,21].
Herein, we report as the first example an epoxide hydrolase-
catalyzed preparative bioresolution of rac-1 at a high substrate
loading of 100 g/L and with a total turnover number (TTN) of up
to 70,000. The resultant (R)-2 and (S)-1 were chemically converted
to (R)- and (S)-3 by functional group transformation in excellent
overall yields (Fig. 1).
(1000 rpm). The samples of 100 ␮L were withdrawn and mixed
with 400 ␮L of methanol to terminate the reaction. The resulting
solution was analyzed by reverse-phase HPLC (C18 column) with a
mobile phase of MeOH/H O = 90/10 (v/v) and detected at 280 nm.
2
The initial rate was determined by monitoring the concentration
change of substrate. Kinetic parameters were obtained by non-
linear fitting of initial rates at varied substrate concentrations to
Michaelis–Menten equation with Origin 8.0.
2.4. Selection of organic solvent
Hydrolysis reactions of rac-1 were routinely performed in 15-mL
2
. Experimental
sealed tubes. Lyophilized cell-free extract (3.2 mg) was suspended
in 1.6 mL of potassium phosphate buffer (100 mM, pH 7.0, 0.02%
2.1. General
◦
Tween-80), and rehydrated at 25 C for 10 min. After 0.4 mL of
organic phase containing 250 mM rac-1 was added, the mixture
was shaken for 1 h at 180 rpm. The reaction was terminated by
extraction with ethyl acetate, and the resultant organic phase was
␣
-Naphthol and epichlorohydrin used in the synthesis of
␣
-naphthyl glycidyl ether were purchased from Aladdin (Shang-
ꢀ
hai, China). Enantiopure (R)-1, (S)-1, and (R)-3-(1 -naphthyloxy)
propane-1,2-diol (2) used in this study were the products of
our previous work [14]. All the other chemicals were obtained
commercially and of reagent grade. HPLC was performed on
dried over Na SO4 and analyzed by chiral HPLC to determine the
2
conversion and enantiomeric excesses of epoxide and diol. The
sample of aqueous phase was withdrawn and the residual activity
of enzyme was assayed to test its stability in presence of var-
ious organic solvents. To analyze the solubility of epoxide and
diol in organic solvents, rac-1 or (R)-2 was added to saturation
®
Shimadzu LC2010A HT (Kyoto, Japan) using CHIRALCEL OD-H
1
column (0.46 cm × 25 cm). H NMR measurements were recorded
on a Bruker 400 MHz spectrometer (Massachusetts, USA). Optical
rotations were measured on an automatic polarimeter (Rudolph
Research Autopol V, New Jersey, USA) in a 10 cm (10 mL) cell.
rac-␣-Naphthyl glycidyl ether (1) was synthesized from ␣-
naphthol and epichlorohydrin [20]. To a solution of ␣-naphthol
and their concentrations were determined by HPLC. Chiral HPLC
conditions: Daicel CHIRALCEL® OD-H, hexane/i-PrOH = 85/15 (v/v);
flow rate, 1 mL/min; ꢀ = 280 nm; retention time, (R)-1 = 9.7 min, (S)-
1
= 10.9 min, (R)-2 = 17.2 min, (S)-2 = 18.9 min.
(
14.4 g, 100 mmol) in 25 mL epichlorohydrin (320 mmol), sodium
2
.5. Effect of organic-aqueous phase volume ratio
hydroxide (200 mmol, 50% in water) was added very carefully
at 60 C under mechanical stirring. After the starting material
◦
The bioresolution of rac-1 was performed in a biphasic reaction
was consumed (as monitored by TLC), the mixture was extracted
with ethyl acetate (3 × 50 mL). The combined organic phase
was washed with brine (3 × 30 mL), dried over Na SO , filtered,
system with different volume ratios of the organic phase to the
aqueous phase (o/w). Varied volume ratios (o/w), from 0:10 (with
1
loadings of 10 g/L and 50 g/L, respectively. The weight ratio of sub-
strate and enzyme (lyophilized cell-free extract) was kept constant
at 10 (w/w). To a 10-mL sealed bottle, 20 mg of rac-1 and 2.0 mg
of lyophilized cell-free extract were added into 2 mL of biphasic
solvents with varied volume ratios (o/w) as described above. The
2
4
0% DMSO as co-solvent) to 4:1, were tested at overall substrate
and concentrated under vacuum. The residue was purified by
flash chromatography (silica gel) using EtOAc/petroleum ether
(
1:10, v/v) as an eluent, affording epoxide 1 as light yellow oil.
1
rac-␣-Naphthylglycidyl ether (1): yield, 76%; H NMR (400 MHz,
CDCl ), ı/ppm: 8.28–8.30 (m, 1H), 7.76–7.78 (m, 1H), 7.43–7.47 (m,
3
1
3
H), 7.33 (t, J = 8.0 Hz, 1H), 6.75 (d, J = 7.6 Hz, 1H), 4.32 (dd, J = 2.8,
0.8 Hz, 1H), 4.06 (dd, J = 5.6, 10.8, 1H), 3.42–3.43 (m, 1H), 2.89 (t,
◦
mixtures were agitated at 25 C with a magnetic stirrer bar and
samples were withdrawn at 1 h. The conversion, ees (enantiomeric
excess of substrate) and eep (enantiomeric excess of product) were
determined by chiral HPLC to show the effect of phase volume ratio
on the activity and enantioselectivity of bioresolution.
J = 4.8 Hz, 1H), 2.78 (dd, J = 2.4, 4.8 Hz, 1H).
2
.2. Expression and purification of BmEHF128T
As described previously [14], BmEHF128T was overexpressed in
Escherichia coli BL21(DE3) and the cells containing the expressed
recombinant protein were harvested by centrifugation at 6000 × g
for 10 min. Then the cell pellets were re-suspended in a potassium
phosphate buffer (50 mM, pH 7.0) followed by disruption with a
high-pressure homogenizer. Cell debris was removed by centrifu-
gation (31,000 × g, 40 min). The supernatant was lyophilized to
afford crude enzyme powder (cell-free extract). For protein purifi-
cation, the supernatant was loaded onto a nickel affinity column
2.6. Preparative bioresolution of rac-1
For the up-scaled bioresolution of epoxide 1, the reaction was
performed in a 250-mL three-neck flask. Recombinant whole cells
(1.0 g dry cell weight) harboring BmEHF128T was added to 50 mL of
potassium phosphate buffer (100 mM, pH 7.0) and rehydrated at
◦
25 C for 10 min. Another 50 mL of buffer containing 0.8% Tween-
80 (w/v) was mixed with 100 mL of organic phase (diisopropyl
ether/isooctane = 65/35, v/v) containing 20 g of rac-1 and stirred
vigorously in the flask, resulting in a water/oil emulsion. The rehy-
drated biocatalyst was added before the reaction mixture was
(
GE Healthcare) and eluted with 10–250 mM imidazole solution
containing 500 mM NaCl and 50 mM potassium phosphate (pH 8.0).
The pure enzyme was identified by SDS-PAGE and dialyzed against
potassium phosphate buffer (50 mM, pH 7.0).
mechanically agitated at 180 rpm for 8 h at room temperature
◦
(
25 C). Once the ee value of residual epoxide reached 99% (S), the
2
.3. Determination of kinetic parameters
reaction mixture was centrifuged to separate the organic phase
containing the residual (S)-epoxide, the aqueous phase and the pre-
cipitate containing the (R)-diol and cell pellets. The diol precipitate
was washed with water and petroleum ether in turn to remove the
residual epoxide and recrystallized with ethanol, giving (R)-2 as
a white crystal. The organic layer was dried over Na SO , filtered
The purified enzyme was mixed with potassium phosphate
buffer (100 mM, pH 7.0, with 10% DMSO and 0.02% Tween-80)
containing (R)-1 or (S)-1 at concentrations ranged from 0.15 to
5
◦
.0 mM, and the reaction was performed at 30 C with shaking
2
4

X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
277
◦
Fig. 1. Chemoenzymatic routes for the synthesis of propranolol enantiomers (R)- and (S)-3. a) BmEHF128T, isopropyl ether/isooctane/buffer = 32.5/17.5/50 (v/v), 25 C, 8 h; the
◦
epoxide and diol were separated by centrifugation. b) 33% HBr/HOAc, 50 C, 1 h; K2CO3/CH3OH, r.t., 30 min. c) Isopropylamine, reflux, 24 h.
Table 1
and the solvent was evaporated, the column chromatography of
the residue afforded (S)-1 as a light yellow oil.
a
Kinetic parameters of BmEHF128T towards (R)- and (S)-1.
Substrate kcat (s−1) KM (mM) kcat/KM (s−1
M
−1
)
Eb [=(kcat/KM)R/(kcat/KM)S]
(
S)-␣-Naphthyl glycidyl ether (1): yield, 45.3% (9.06 g); 99.1%
ee; [␣]25D: +37.0 (c 1.0, methanol), Lit. [15] [␣]25D: +32.9 (c
5
(
R)-1
264
3.54
0.72
0.53
3.7 × 10
55
1
8
.0, methanol) with >95% ee. 1H NMR (400 MHz, CDCl ), ı/ppm:
(S)-1
6.7 × 103
3
.28–8.30 (m, 1H), 7.75–7.78 (m, 1H), 7.44–7.47 (m, 3H), 7.40 (t,
a
The kinetic parameters were determined with substrate concentrations of
J = 7.9 Hz, 1H), 6.71 (d, J = 7.6 Hz, 1H), 4.29 (dd, J = 2.9, 11.0 Hz, 1H),
0.15–5 mM for (R)- and (S)-1.
b
The theoretical E value was calculated based on the kinetic parameters.
4
2
.03 (dd, J = 5.6, 11.0, 1H), 3.40–3.41 (m, 1H), 2.88 (t, J = 4.5 Hz, 1H),
.76 (dd, J = 2.6, 4.8 Hz, 1H).
ꢀ
(
R)-3-(1 -Naphthyloxy)-propane-1,2-diol (2): yield, 42.4% (9.24
g); >99.5% ee; [␣]25D: –8.3 (c 1.2, ethanol), Lit. [22] [␣]25D: –7.6
(
c 1.2, ethanol) with 97% ee. 1H NMR (500 MHz, CDCl ), ı/ppm:
3
CDCl ), ı/ppm: 8.28–8.30 (m, 1H), 7.75–7.78 (m, 1H), 7.44–7.47 (m,
8
.20–8.22 (m, 1H), 7.80–7.82 (m, 1H), 7.45–7.49 (m, 3H), 7.37 (t,
J = 8.0 Hz, 1H), 6.84 (d, J = 7.5 Hz, 1H), 4.27–4.28 (m, 1H), 4.23–4.24
m, 2H), 3.94 (dd, J = 3.6, 11.2 Hz, 1H), 3.86 (dd, J =n5.2, 11.2 Hz, 1H),
3
3
1
H), 7.41 (t, J = 7.9 Hz, 1H), 6.73 (d, J = 7.6 Hz, 1H), 4.30 (dd, J = 2.9,
1.0 Hz, 1H), 4.03 (dd, J = 5.6, 11.0, 1H), 3.41–3.42 (m, 1H), 2.88 (t,
(
2
J = 4.5 Hz, 1H), 2.77 (dd, J = 2.6, 4.8 Hz, 1H).
.21 (s, 2H).
ꢀ
(
R)-1-(1-Methylethylamino)-3-(1 -naphthyloxy) propan-2-ol
(
[
3): overall yield from
-1, 31.4% (1.2 g); >99.5% ee; [␣]25D = +10.2 (c 1.58, EtOH), lit.
2
.7. Synthesis of (R)- and (S)-propranolol (3) from (R)-2 and
1
15] [␣]25D = +9.82 (c 1.6, EtOH) with >95% ee. H NMR (400 MHz,
(
S)-1
CDCl ) ı/ppm: 8.23–8.26 (m, 1H), 7.78–7.81 (m, 1H), 7.47–7.48 (m,
3
3
H), 7.35–7.43 (m, 1H), 6.80 (d, J = 7.5, 1H), 4.27–4.31 (m, 1H), 4.20
The (R)-2 obtained from bioresolution (5 g, 22.9 mmol) was
◦
(dd, J = 5.2, 9.4, 1H), 4.13 (dd, J = 5.2, 9.4, 1H), 3.72 (s, 2H), 3.07 (dd,
J = 3.5, 12.2, 1H), 2.90–3.00 (m, 2H), 1.17–1.19 (m, 6H).
treated with 33% HBr-HOAc (20 mL) at 50 C for 1 h. The reaction
mixture was poured slowly into ice water, neutralized with aque-
ous NaOH (1.0 M) and was extracted with ethyl acetate (3 × 50 mL).
The combined organic layers were washed with brine, dried over
Na SO , and filtered in turn. After removal of the solvent, the
ꢀ
(
S)-1-(1-Methylethylamino)-3-(1 -naphthyloxy) propan-2-ol
(3): overall yield from rac-1, 44.8% (6.4 g); 99.1% ee; [␣]25D = –10.5
1
(c 1.5, EtOH), lit. [15] [␣]25D = –9.7 (c 1.5, EtOH) with >95% ee. H
2
4
NMR (400 MHz, CDCl ) ı/ppm: 8.15–8.17 (m, 1H), 7.67–7.69 (m,
1
4
residue was treated with K CO3 (5.0 g) in 50 mL methanol for
3
2
H), 7.30–7.36 (m, 3H), 7.22–7.24 (m, 1H), 6.67 (d, J = 7.6 Hz, 1H),
.03–4.10 (m, 2H), 3.97 (dd, J = 4.8, 9.0, 1H), 2.67–2.87 (m, 5H), 1.12
3
0 min at room temperature. The reaction mixture was then con-
centrated, dissolved with 100 mL distilled water and extracted
with ethyl acetate (3 × 50 mL). The combined organic layers were
washed with brine and dried over Na SO . After removal of the sol-
(d, J = 6.2 Hz, 6H).
2
4
vent, the crude product was purified by flash chromatography on
silica gel (eluted with EtOAc/petroleum ether = 1/10), giving (R)-1
3. Results and discussion
(
3.67 g, 80% yield) as a light yellow oil.
The resultant (R)-1 (1.0 g) and (S)-1 (5.0 g) were then refluxed
3
.1. Catalytic properties of BmEHF128T
in isopropylamine, respectively, for 24 h (the reaction progress
was monitored by TLC) followed by evaporation of the residual
isopropylamine, affording 1.20 g of (R)-3 (>99.5% ee) and 6.40 g
of (S)-3 (99.1% ee) as white crystals after recrystallization with
hexane/ethyl acetate. HPLC conditions for (R)- and (S)-3: Daicel
BmEH variants for the bioresolution of aryl glycidyl ethers were
previously prepared in our laboratory [14]. BmEHF128T has a very
high activity and good enantioselectivity towards ␣-naphthyl gly-
cidyl ether (1) (Table 1). This enantioselectivity can be attributed to
the higher kcat for (R)-1 compared with (S)-1. A calculation revealed
®
CHIRALCEL OD-H, hexane/i-PrOH/triethylamine = 90/10/0.1; flow
rate, 1 mL/min; ꢀ = 280 nm; retention time (R)-3 = 10.0 min, (S)-
>
99.5% regioselectivity for the less substituted ␤-carbon (terminal
3
= 17.2 min.
R)-␣-Naphthyl glycidyl ether (1): overall yield from rac-1,
carbon) of the epoxide ring of (R)-1 or (S)-1 (Supporting infor-
mation). Therefore, BmEHF128T is a promising biocatalyst for the
preparative bioresolution of rac-1 at high space-time yield (STY).
(
3
[
3.9% (3.67 g); >99.5% ee; [␣]25D = –36.5 (c 1.5, methanol), Lit. [15]
1
␣]25D = –33.9 (c 1.55, methanol) with >95% ee. H NMR (400 MHz,

2
78
X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
Fig. 2. Effect of organic phase solvent on the bioresolution of rac-1 in a biphasic system. ᭿ Enantiomeric excess of epoxide (S)-1; ᭿ enantiomeric excess of diol (R)-2; ᭿
conversion of rac-1; ᮀ residual activity of the enzyme in aqueous phase. The activity of enzyme added at the beginning of the reaction was set as 100%. The reactions were
carried out for 1 h at an overall substrate load of 50 mM and biocatalyst load of 1.6 gCFE/L with a phase ratio of 1/4 (o/w).
3
.2. Choice of a solvent for composing the biphasic system
ture of 60/40 (v/v) was chosen as the organic phase of the biphasic
system, which afforded a compromise of substrate loading, enzyme
activity and stability.
Epoxides usually have poor aqueous solubility [5], therefore,
biphasic reaction conditions are frequently used to increase the
substrate loading. The logP value is the logarithm of the par-
tition coefficient of a compound in the mixture of 1-octanol
and water, which was often used to predict the biocompatibil-
ity of organic solvents [23]. The logP values for methyl tert-butyl
ether (MTBE), dichloromethane (DCM), diisopropyl ether, chloro-
form, butyl acetate, toluene, cyclohexane and isooctane are 1.15,
3.3. Optimization of phase volume ratio
The phase ratio of a two-phase system will influence the relative
concentrations of substrate and enzyme on the biphasic interface
and the specific surface area for mass transportation. In our study,
the phase ratio was initially investigated at fixed loadings of biocat-
alyst and substrate. In the case of 10 g/L rac-1 vs 1 gCFE/L, biphasic
systems with varied phase ratios from 4:1 to 0:10 (o/w) were tested.
As shown in Table 3, the optimal phase ratio was 1/2 or 1/4 (o/w)
under the aforementioned conditions. Comparatively, for bipha-
sic systems using phase ratios other than 1/2 or 1/4 (o/w), lower
catalytic rates were observed and the enantioselectivities were sig-
nificantly lower. For comparison, the reaction was also performed
with 10% (v/v) of DMSO as a monophasic cosolvent (o/w = 0/10),
where the initial rate was higher at the beginning because of the
higher substrate concentration in the aqueous phase. However,
this high substrate concentration also accelerated the inactivation
of the enzyme, which slowed down the rate in the later reaction
period.
When the substrate loading was increased to 50 g/Ltotal with a
constant substrate/catalyst (S/C) ratio (10, w/w), the phase ratios
of 1:2 and 1:1 (o/w) were found to be the best. Therefore, a higher
proportion of the organic phase was needed for higher substrate
loading, which may be explained in two ways: first, a higher loading
would concentrate rac-1 in the organic phase and at the biphasic
interface. Therefore, the increased phase ratio (o/w) could dilute
the substrate and lower its inactivation effect on the enzyme. How-
ever, the optimal substrate concentration was 20–40 g/Lorg in the
organic phase in the case of 10 g/Ltotal substrate (1 g/Ltotal biocata-
lyst) or 100–150 g/Lorg in the case of 50 g/Ltotal substrate (5 g/Ltotal
biocatalyst). This leads to the second point, where better tolerance
of the enzyme against the substrate and organic solvent might
be achieved at higher biocatalyst loading. For phase ratios over
1:1 (o/w), a low enantioselectivity at either 10 or 50 g/Ltotal sub-
strate loading was observed, which might be caused by the mass
transfer limitation due to the lower substrate and higher enzyme
concentrations on the biphasic interface. Eventually, the phase ratio
of 1:1 (v/v) was chosen for subsequent reactions with 50 g/Ltotal or
higher substrate loading.
1
.19, 1.68, 1.76, 1.77, 2.68, 3.39 and 4.82, respectively (data from
SciFinder database). The effects of the aforementioned solvents on
the activity, enantioselectivity and stability of BmEHF128T can be
probed by the conversion, ees, eep, and the residual activity of the
enzyme after 1 h of reaction in the presence of these solvents. As
shown in Fig. 2, the biocatalyst was inactivated with DCM and chlo-
roform, with no diol and very little residual activity detected. In the
other solvents, a good enantioselectivity was always observed, with
higher activities in diisopropyl ether and cyclohexane. In addition,
better stability of enzyme was obtained in high logP solvent like
isooctane (4.82), toluene (2.68) and cyclohexane (3.39). Even more
than 100% of the residual activity was observed with isooctane,
which is probably caused by a more active conformation of enzyme
induced by isooctane. However, due to the low solubility of rac-1
in isooctane (as described below) and the low water miscibility of
isooctane, a much lower concentration of rac-1 would be diffused
into the aqueous phase of water/isooctane biphasic system, which
resulted in a relative low conversion (Fig. 2).
In consideration of the demand of high substrate loading in
the industrial biocatalysis processes, the solubilities of rac-1 and
(
R)-2 in diisopropyl ether, cyclohexane, isooctane and the aqueous
phase were tested (Table 2). (R)-2 had poor solubility in organic
and aqueous solvents, causing precipitation during the reaction as
previously observed [14]. In contrast, rac-1 had a good solubility in
diisopropyl ether, but a low solubility in cyclohexane and isooctane,
and was insoluble in the aqueous phase. Since the enzyme has bet-
ter stability in isooctane, a mixed organic phase of isooctane and
diisopropyl ether was also tested (Table 2). As the proportion of
diisopropyl ether was increased, the solubilities of rac-1 and (R)-2
and the enzyme activity increased, although the enzyme stabil-
ity decreased. A highly efficient bioresolution of rac-1 requires an
organic phase with a substrate loading of ≥100 g/L. Conversely, the
lower solubility of (R)-2 could facilitate the downstream separation
of epoxide and diol. Therefore, the diisopropyl ether/isooctane mix-

X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
279
Table 2
Effect of the organic phase composition on the substrate/product solubility and bioresolution efficiency.
Solvent
Solubility
ees (%)
eep (%)
Conv. (%)
Residual activity (%)a
rac-1(g/L)
(R)-2 (g/L)
Isooctane
29
46
76
146
0.05
0.24
0.65
1.35
2.22
4.33
0.09
0.11
24.2
72
98
19.8
42.1
41.7
43.5
43.5
43.8
40.4
n.d.
116
87.2
76.2
68.3
53.9
58.0
48.0
n.d.
2
4
6
8
0/80 (D/I)b
0/60 (D/I)
0/40 (D/I)
0/20 (D/I)
98.9
>99.5
>99.5
>99.5
99.4
>99.5
n.d.
71.4
76.8
76.8
77.6
67.6
miscible
miscible
101
Diisopropyl ether
Cyclohexane
Aqueous phase
N.D.c
n.d.
d
a
◦
The residual activity of enzyme in the aqueous phase after 1 h of reaction at 25 C. The activity of enzyme added at the beginning was set as 100%.
b
c
(
D/I) is the volumetric ratio of diisopropyl ether and isooctane.
Not detectable.
Not determined.
d
Table 3
a
Effect of phase volume ratio on the bioresolution of rac-1 at 10 and 50 g/Ltotal loads.
Phase ratio (o/w)
10 g/Ltotal
50 g/Ltotal
ees, 10 min(%)b
ees, 3 h(%)
Conv. 3 h(%)
Ec
ees, 4 h(%)
Conv. 4 h(%)
E
0
0
0
0
0
0
0
:10d
1:09
1:04
1:02
1:01
2:01
4:01
33.0
19.2
21.9
18.9
13.8
8.0
90.8
70.7
99.1
98.4
83.5
71.6
31.3
48.3
46.0
51.7
49.9
45.8
45.6
24.7
>200
23
142
>200
>200
27
n.d.e
n.d.
93.6
99.5
98
n.d.
n.d.
n.d.
n.d.
>200
164
>200
68
49.1
51.7
50.2
50.7
n.d.
93
n.d.
5.8
58
n.d.
a
Reaction conditions: substrate/catalyst (w/w) = 10/1; total volume = 2 mL; organic phase was diisopropyl ether/isooctane = 60/40; aqueous phase was 100 mM potassium
◦
phosphate buffer (pH 7.0) with 0.02% (w/v) of Tween-80, 25 C, magnetically stirred at 120 rpm. The subscript total means the concentration in total volume, and org means
in organic phase.
b
The ee of (S)-1 in the reaction mixture at a specific time.
c
E value at the end of reaction (3 h for 10 g/Ltotal or 4 h for 50 g/Ltotal) was calculated according to E = ln[(1−ees)(1−c)]/ln[(1+ees)(1−c)].
d
DMSO at 10% (v/v) was used as a cosolvent.
e
n.d. = not determined.
bioresolution of rac-1 at 50 g/Ltotal and 100 g/Ltotal were efficiently
completed within 2 h and 4 h, with ≥98% ees. Furthermore, the
application potential of BmEH
was also tested in the form of a
F128T
whole cell biocatalyst, which took merely 20% longer time to finish
the reaction, as compared with the same dry weight of CFE. This
implies a good balance between the mass-transfer limitation and
the higher stability of whole cell biocatalyst.
A preparative bioresolution with 20 g of rac-1 was carried out
at 100 g/Ltotal (or 0.5 M) substrate loading with 1 g dry cell weight
(DCW) of E. coli cells harboring BmEHF128T, biphasic solvents at
a phase ratio of 1/1 (o/w), a binary organic phase composed of
diisopropyl ether/isooctane (D/I = 65/35, v/v) and 200 g/L (1.0 M in
organic phase) of rac-1. As a result, an ees of 99% was achieved
in 8 h and the diol (R)-2 precipitated during the reaction (Fig. 4a).
The formation rate of (S)-2 (minor enantiomer) increased signifi-
cantly after 6 h and the R-enantiomer precipitated instantly, giving
an excellent enantiopurity of (R)-2 (>99.5% ee) after the post-
reaction separation. In contrast, the ee of the diol left in the organic
phase was only 13.4% (S). (S)-1 and (R)-2 could be simultaneously
isolated in 45.3% and 42.4% yields from rac-1, respectively, both
with excellent enantiopurity (>99% ee, Fig. 4b). The bioresolution
Fig. 3. Effect of the substrate and biocatalyst loadings on bioresolution effi-
ciency. The composition of the organic phase of 65/35 was used to dissolve
2
00 g/Lorg of rac-1 in the organic phase. ᭿ 50 g/Ltotal substrate vs. 5 gCFE/Ltotal
enzyme; ꢀ 100 g/Ltotal substrate vs. 5 gCFE/Ltotal enzyme; ꢁ 100 g/Ltotal substrate
vs. 5 gDCW/Ltotal lyophilized cell.
3.4. Preparative bioresolution of rac-1 at 100 g/Ltotal
An ideal biocatalytic process needs a high STY and turnover
−
1
−1
described herein provides STYs of 136 g L days for (S)-1 and
number to be cost-effective. Industrial processes usually have a
biocatalyst loading of ≤5 g/L. This is to avoid the high viscosity
of the aqueous phase, resistance of mass-transfer and difficulties
in product recovery [24]. To determine the highest sustainable
substrate loading at the maximal biocatalyst loading of 5 gCFE/L, a
substrate loading of 50 g/Ltotal and 100 g/Ltotal (or 100 & 200 g/Lorg
in the organic phase) were tested in a 10 mL biphasic system, with
−
1
−1
1
39 g L days for (R)-2, and TTN of the enzyme was as much as
7
0,000, which is outstanding when compared with other biocat-
alytic or chemical methods (Table 4).
Finally, enantiomerically pure (S)-1 and (R)-2 were chemically
converted into the corresponding (R)-3 and (S)-3, as shown in Fig. 1.
(R)-2 (5.0 g) was treated with HOAc/HBr, then converted into epox-
ide (R)-1 (3.67 g) by addition of a base [28], giving 80% yield with
no loss of enantiopurity [(R)-1 >99% ee, Fig. 4b]. (S)-1 and (R)-1
0.2% (w/v) of Tween-80 in the aqueous phase to achieve a bet-
ter dispersion of the biphasic system [25]. As shown in Fig. 3, the

2
80
X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
Fig. 4. Bioresolution of rac-1 at 100 g/Ltotal. a) Changes in molar concentration of epoxide and diol enantiomers in the course of the bioresolution: ᭿ (S)-1; ᮀ (R)-1; ᭹ (S)-2; ꢁ
R)-2. b) HPLC spectra: 1) rac-1 before kinetic resolution; 2) Reaction mixture at the end of the bioresolution; 3) (S)-1 separated with flash column chromatography; 4) (R)-2
(
purified with filtration and recrystallization; 5) Conversion of (R)-2 to (R)-1. The reaction volume was 200 mL, with 20 g of rac-1 and 1 g dry cell weight of BmEHF128T whole
◦
cell. The reaction was performed at 25 C, 180 rpm, with a phase ratio (o/w) of 1/1. The organic phase was diisopropyl ether/isooctane = 65/35 (v/v). The aqueous phase was
1
00 mM potassium phosphate buffer (pH 7.0) supplemented with 0.2% (w/v) Tween-80.
Table 4
a
Comparison of biocatalytic or ogano-metal catalyzed processes for asymmetric synthesis of precursors of (S)-propranolol (3).
Substrate load (g/L) Catalyst load (g/L) ee (%) Yield (%) STYb (g L days
−1
−1
)
TTN (mol/mol) Ref.
Entry Catalyst
1
2
3
4
5
6
Co (III) salen complex
Co (III) salen complex
Pichia mexicana reductase (cell)
Mucor javanicus lipase (CFE)
Pseudomonas cepacia lipase (CFE) n.a.
BmEHF128T(cell) 100
660
80
0.002
50
1.21
3.93
n.a.c
50
n.a.
5
96
98
95
>99
96
45
49
86
44
50
45.8
297
59
< 0.01
4.5
n.a.
136
1,650
50
n.a.
n.a.
n.a.
Khan et al. [26]
Kureshy et al.[27]
Sinisterra et al. [19]
Taneja et al.[17]
Kamal et al.[18]
This study
d
>99
70,000
a
The product of entries 1, 2 and 6 is (S)-naphthyl glycidyl ether (1). The product of entries 3, 4 and 5 is (R)-1-chloro-3-(1-naphthyloxy)-2-propanol, which is also a precursor
of (S)-propranolol (3).
b
STY was calculated by the following equation: product weight/(reaction time × reaction volume).
c
The data were not available from the corresponding literature.
d
The number was calculated based on the purification yield of BmEHF128T, which is ∼25 mg (7.14 × 10−4 mmol) of pure enzyme per gram dry cell weight. The molecular
weight of BmEHF128T is 35 kDa.
were then refluxed with isopropylamine, followed by recrystalliza-
tion, giving (R)- and (S)-3 in 44.8% and 31.4% yields (both >99% ee),
respectively.
of Science and Technology (No. 11431921600) and DSM Research,
The Netherlands (No. ACER 9808-0126) for financial supports.
Appendix A. Supplementary data
4
. Conclusions
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.molcatb.2015.10.
005.
In summary, we have developed an enzymatic resolution of
rac-1 using BmEHF128T, an engineered epoxide hydrolase of high
turnover number (∼70,000). Various organic solvents for a bipha-
sic system were screened and the phase ratio (v/v) was optimized.
The best results were obtained with diisopropyl ether/isooctane
References
(
60/40, v/v) and a volume ratio of organic/aqueous of 1/1. Dur-
[
1] N. Kasai, T. Suzuki, Y. Furukawa, J. Mol. Catal. B-Enzym. 4 (1998) 237–252.
ing the bioresolution, diol (2) precipitated instantly because of
poor solubility, facilitating its separation and improving the enan-
tiopurity of (R)-2 (>99.5% ee). Under the optimal conditions, the
bioresolution was scaled up to 100 g/Ltotal substrate loading afford-
ing enantiopure epoxide (S)-1 (>99% ee) and diol (R)-2 (>99% ee)
in 45.3% and 42.4% yields, respectively. Furthermore, both the R-
and S-enantiomers of propranolol were successfully synthesized in
[2] M. Widersten, A. Gurell, D. Lindberg, BBA-Gen. Subjects 1800 (2010) 316–326.
[
[
[
3] W. Xu, J.H. Xu, J. Pan, Q. Gu, X.Y. Wu, Org. Lett. 8 (2006) 1737–1740.
4] S. Hwang, C.Y. Choi, E.Y. Lee, J. Ind. Eng. Chem. 16 (2010) 1–6.
5] W.J. Choi, Appl. Microbiol. Biotechnol. 84 (2009) 239–247.
[6] B. Seisser, I. Lavandera, K. Faber, J.H.L. Spelberg, W. Kroutil, Adv. Synth. Catal.
349 (2007) 1399–1404.
[
[
7] A. Riera, M. Moreno, Molecules 15 (2010) 1041–1073.
8] G. De Faveri, G. Ilyashenko, M. Watkinson, Chem. Soc. Rev. 40 (2011)
1722–1760.
4
4.8% and 31.4% yields and >99% ee, respectively. This chemoenzy-
[9] S. Wu, A. Li, Y.S. Chin, Z. Li, ACS Catalysis 3 (2013) 752–759.
10] L.F. Solares, C. Mateo, Curr. Org. Chem. 17 (2013) 744–755.
11] J. Deregnaucourt, A. Archelas, F. Barbirato, J.M. Paris, R. Furstoss, Adv. Synth.
Catal. 349 (2007) 1405–1417.
[
[
matic route to (R)- and (S)-propranolol displayed high STYs under
mild reaction conditions, with obvious merits of cost-effectiveness
and eco-friendlyness.
[12] M.T. Reetz, H. Zheng, Chembiochem 12 (2011) 1529–1535.
[
[
13] H. Zheng, M.T. Reetz, J. Am. Chem. Soc. 132 (2010) 15744–15751.
14] X.-D. Kong, Q. Ma, J. Zhou, B.-B. Zeng, J.-H. Xu, Angew. Chem. Int. Ed. 53 (2014)
6641–6644.
15] H.S. Bevinakatti, A.A. Banerji, J. Org. Chem. 56 (1991) 5372–5375.
16] S.P. Panchgalle, R.G. Gore, S.P. Chavan, U.R. Kalkote, Tetrahedron Asymmetry
Acknowledgements
[
[
Thanks are given to the National Natural Science Foundation
of China (Nos. 21276082 and 21536004), Ministry of Science and
Technology, PR China, (No. 2011CB710800), Shanghai Commission
20 (2009) 1767–1770.
[17] M. Kapoor, N. Anand, S. Koul, S.S. Chimni, K.S. Manhas, C. Raina, R. Parshad,
S.C. Taneja, G.N. Qazi, Bioorg. Chem. 31 (2003) 259–269.

X.-D. Kong et al. / Journal of Molecular Catalysis B: Enzymatic 122 (2015) 275–281
281
[
[
[
18] A. Kamal, M. Sandbhor, A. Ali Shaik, Bioorg. Med. Chem. Lett. 14 (2004)
[23] P.F. Gong, J.H. Xu, Enzyme Microb. Technol. 36 (2005) 252–257.
[24] S. Luetz, L. Giver, J. Lalonde, Biotechnol. Bioeng. 101 (2008) 647–653.
[25] Y.Y. Liu, J.H. Xu, Y. Hu, J. Mol. Catal. B-Enzym. 10 (2000) 523–529.
[26] A. Sadhukhan, N.-U.H. Khan, T. Roy, R.I. Kureshy, S.H.R. Abdi, H.C. Bajaj, Chem.
Eur. J. 18 (2012) 5256–5260.
[27] M. Kumar, R.I. Kureshy, A.K. Shah, A. Das, N.-U.H. Khan, S.H.R. Abdi, H.C. Bajaj,
J. Org. Chem. 78 (2013) 9076–9084.
4
581–4583.
19] F.M.N. Lagos, C. Del Campo, E.F. Llama, J.V. Sinisterra, Enzyme Microb.
Technol. 30 (2002) 895–901.
20] J. Zhao, Y.-Y. Chu, A.-T. Li, X. Ju, X.-D. Kong, J. Pan, Y. Tang, J.-H. Xu, Adv. Synth.
Catal. 353 (2011) 1510–1518.
[
[
21] Y. Xu, J.-H. Xu, J. Pan, Y.-F. Tang, Biotechnol. Lett. 26 (2004) 1217–1221.
22] S.E. Schaus, B.D. Brandes, J.F. Larrow, M. Tokunaga, K.B. Hansen, A.E. Gould,
M.E. Furrow, E.N. Jacobsen, J. Am. Chem. Soc. 124 (2002) 1307–1315.
[28] B. Jiang, Z.L. Chen, M. Shen, Z.M. Wang, M. Shi, Chin. J. Chem. 21 (2003)
789–792.