Improved o-chlorobenzoylformate bioreduction by stabilizing aldo-keto reductase YtbE with additives

Article abstract of DOI:10.1016/j.molcatb.2014.03.006

Asymmetric reduction of methyl o-chlorobenzoylformate (CBFM) using aldo-keto reductase YtbE is a potentially cost-effective and green technology in manufacturing methyl (R)-o-chloromandelate which is a key intermediate for synthesizing (S)-clopidogrel (a

Full text of DOI:10.1016/j.molcatb.2014.03.006

Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Improved o-chlorobenzoylformate bioreduction by stabilizing
aldo-keto reductase YtbE with additives
Yan-Peng Xu, Yue Hugh Guan, Hui-Lei Yu∗, Yan Ni, Bao-Di Ma, Jian-He Xu∗
Laboratory of Biocatalysis and Synthetic Biotechnology, State Key Laboratory of Bioreactor Engineering, East China University of Science and Technology,
Shanghai 200237, People’s Republic of China
a r t i c l e i n f o
a b s t r a c t
Article history:
Asymmetric reduction of methyl o-chlorobenzoylformate (CBFM) using aldo-keto reductase YtbE is a
potentially cost-effective and green technology in manufacturing methyl (R)-o-chloromandelate which
is a key intermediate for synthesizing (S)-clopidogrel (a popular medicine for treating atherosclerosis). At
the moment, large scale application of YtbE has been complicated by uncertain thermal and operational
stabilities. Consequently, we endeavored possible enzyme inactivation mechanism, and showed that (a)
unfolding of YtbE explains enzyme activity loss, and (b) YtbE dimerization has a less significant effect
owing to a small quantity detected. The effects of substrate and temperature on YtbE are mostly upheld
by a one-step inactivation model, whereas the effect of product by a 2-step activity reduction modality.
Partially based on these new understandings, a multi-factor experimental strategy was rationalized for
improving the YtbE stability. For instance, glycerol was introduced to reduce enzyme unfolding whilst
dithiothreitol to suppress its dimerization. This improved substrate conversion from 62.9% to 98.7%, and
from 70.5% to 96.6% at 0.1 M and 1.0 M CBFM, respectively, with YtbE half-life being increased from
46.6 min to 159 min.
Received 13 December 2013
Received in revised form 27 February 2014
Accepted 6 March 2014
Available online 19 March 2014
Keywords:
Aldo-keto reductase
Enzyme stability
Enzyme inactivation
Inactivation mechanism
Bioprocess enhancement
© 2014 Elsevier B.V. All rights reserved.
1. Introduction
using a single or a mixed enzyme system involving invariably a
carbonyl reductase [1,5–8].
Methyl (R)-o-chloromandelate [(R)-CMM] is a key chiral syn-
thon for manufacturing (S)-clopidogrel bisulfate [e.g. trade name
Plavix®], the world second best-selling drug widely administered
to atherosclerotic patients [1]. Regardless of practicability, there are
usually 3 known routes for preparing (R)-CMM. Both the first and
the second routes are common by, first of all, obtaining the precur-
sor chiral (R)-carboxylic acid: (a) enzymatic and enantioselective
hydrolysis of a racemic nitrile compound mixture to produce solely
a required chiral (R)-carboxylic acid [2,3]; or alternatively (b) the
(R)-carboxylic acid is obtained through tedious fractional crystal-
lization from its racemic mixture [4]. The (R)-carboxylic acid from
the either route then undergoes methyl esterification to produce
(R)-CMM. The third route (c), termed as asymmetric reduction, is to
convert the non-chiral compound, CBFM, directly to chiral (R)-CMM
Compared to the first and the second routes, the third route
obviously is more straightforward and so commercially more
desirable. With large-scale application in mind, our group has
recently identified and developed a new aldo-keto reductase, YtbE
(EC1.1.1.X), for realizing this reaction. Compared to the efforts made
previously, our own enzyme YtbE exhibited a high activity toward
the substrate CBFM, showed a remarkable substrate tolerance, and
resulted in excellent chiral selectivity with >99% ee for (R)-CMM [8].
In view of the documented outcome for enduring high substrate
concentration by YtbE, the prospect appears to be positive for syn-
thesizing enantiopure clopidogrel intermediate using this enzyme
on industrial scales. This set of enzyme reactions is composed of
two coupled reactions (see Scheme 1).
However, the inherent, rather generic enzyme instability may
well complicate large scale application using this enzyme. This is
so despite of encouraging advantages of this enzyme in product
stereo-purity, in compelling activity under mild conditions, in high
reaction turnover number, and environmentally friendly [9].
In order to compensate for the loss of YtbE activity during man-
ufacturing processes, an excess amount of this enzyme will be
required and this will inevitably result in a heightened burden
on downstream processing and so on the overall production cost.
∗
Corresponding authors at: East China University of Science and Technology,
Laboratory of Biocatalysis and Synthetic Biotechnology, State Key Laboratory of
Bioreactor Engineering, Shanghai 200237, People’s Republic of China.
Fax: +86 21 6425 0840.
E-mail addresses: huileiyu@ecust.edu.cn (H.-L. Yu), jianhexu@ecust.edu.cn
(J.-H. Xu).
http://dx.doi.org/10.1016/j.molcatb.2014.03.006
1381-1177/© 2014 Elsevier B.V. All rights reserved.

Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
109
2.2. Enzymatic activity assay
Nomenclature
To assay YtbE activity, an appropriate amount of the enzyme was
added to 1 ml of 0.1 M sodium phosphate buffer (pH 7.0) containing
2 mM CBFM and 0.1 mM NADPH at 30 ◦C and then decrease of the
absorbance at 340 nm was monitored by a Beckman DU 730 Nucleic
Acid/Protein Analyzer. One unit of enzyme activity was defined as
oxidizing 1 ␮mol NADPH per minute.
ANS
8-anilino-1-naphthalene-sulphonate A hydropho-
bic fluorescent molecule for probing YtbE confor-
mational change during loss of activity
dithiothreitol reducing agent to prevent YtbE inter-
molecular disulfide bonds
DTT
Glc-Na and Glc-K sodium gluconate and potassium glu-
conate, respectively, Alkalines Na2CO3 and K2CO3
can be used to maintain the pH of reaction
(Scheme 1), but Glc-Na and Glc-K affected YtbE sta-
bility differently.
2.3. Native SDS–PAGE
Gel electrophoresis was performed on 12% SDS–PAGE with
Tris–glycine buffer (pH 8.3), and a stacking gel containing 5% acryl-
amide was also used. All the enzyme samples (unless specified
in figure legend) were processed without 2-mercaptoethanol and
boiling treatments.
YtbE
aldo-keto reductase asymmetrically reducing CBFM
into (R)-CMM
2.4. Detecting YtbE conformational change by UV scanning and
by a fluorescence probe
In order to reveal temperature-caused conformational change
of YtbE during an inactivation process, 0.08 mg/ml YtbE in 10 mM
sodium phosphate buffer (pH 7.0) was incubated at 50 ◦C for a time
period (to be specified in figure legend) and following 4 ◦C for 1 h,
and then 240–400 nm UV spectra were taken using a Beckman DU
730 Nucleic Acid/Protein Analyzer.
Scheme 1. CBFM was directly reduced to chiral (R)-CMM using a mixed enzyme
system involving the reductase YtbE and a glucose dehydrogenase (GDH).
This conceivable configuration shift was investigated further
using a fluorescence probing molecule ANS. This molecule binds to
hydrophobic regions of proteins and therefore exhibits increased
fluorescence intensity upon binding to proteins. Following addition
of 0.1 mM ANS to 100 mM sodium phosphate buffer (pH 7.0) con-
taining 0.12 mg/ml YtbE, incubation at 50 ◦C was for a time period
(to be specified in figure legend) and then at 30 ◦C for 30 min. Mix-
tures of YtbE and (R)-CMM were made by swiftly combining their
concentrated stocks. Upon further fast mixing with an ANS con-
centrated stock, incubation at 30 ◦C was made for 30 min. At the
end of a designated incubation, the ANS fluorescence emission was
scanned in the range 400–600 nm at 380 nm excitation wavelength
using an F-4600 fluorescence spectrophotometer (Hitachi, Japan).
To manage this problem, a proposed solution was to lower the reac-
tion temperature from 30 ◦C to 20 ◦C [8]. Conceivably, this approach
is at the costs of decreased productivity (owing to an increased
reaction time) and increased refrigeration energy.
This present work however sought an alternative approach in
which the loss of enzyme activity could be minimized through
molecular interactions. Based on an understanding of enzyme inac-
tivation mechanism, Chen et al. improved the thermal stability of
maltose-binding protein Heparinase I [10]. Aided by enzyme kinetic
analysis, Dib and coworkers considerably reduced thermal inacti-
vation of d-amino acid oxidase [11]. For the present work, we thus
desire to understand the mechanism of YtbE inactivation and to
quantify its activity loss kinetics.
Two complementary approaches are usually seen for studying
such enzyme interactions. Biochemical analysis could elucidate the
effect of external factors on secondary or tertiary structures of an
enzyme whilst mathematical simulation focuses on quantitative
effects of these factors on enzyme activity [12,13]. Thus, the first
objective of this work was to gain an insight into plausible YtbE
inactivation mechanism using both the biochemical and the kinetic
simulation approaches. Partially guided by these results, the second
objective was to work out an improved YtbE reaction system for
recommending to industrial uses.
2.5. Kinetic and thermodynamic modeling
The enzyme concerned, YtbE, is assumed to take three forms
[15]: the native (N), the unfolded (U), and the fully inactivated (D),
which are connected by 2 first-order inactivation reactions,
k
k
2
1
N−→ U−→ D
(1)
where
dN
= −k1N
dt
(2)
2. Experimental and modeling methods
and
2.1. Materials
dU
dt
= k1N − k2U
(3)
(4)
CBFM (≥98%) was prepared as reported previously [8] and (R)-
CMM (≥99%) was then synthesized through CBFM bioreduction in
our laboratory. ANS was purchased from Tokyo Chemical Industry
Shanghai (China). NADP+ and NADPH were obtained from Roche
Shanghai (China). All other chemicals were of analytical grade. The
enzyme YtbE was produced by Escherichia coli BL21 cells geneti-
cally engineered to synthesize YtbE through plasmid pET28a. This
recombinant YtbE has an N-terminal His-tag and so, as described
before [14], an immobilized metal affinity chromatograph was used
to purify it to a specific enzyme activity of 6.47 U/mg.
and
dD
dt
= k2U
at t = 0, N = N0, U = D = 0; otherwise N + U + D = N0 (5)
The residual enzyme activity is defined as,
(N + ˇU)
Ra
=
(6)
N0

110
Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
Combining Eqs. (2)–(6) and the following integration, we have,
of the purified YtbE concentration (Fig. 1d). Indicatively, accumu-
lation of (R)-CMM during this enzyme reaction plays a pivotal role
in causing inactivation of YtbE. Furthermore, the inactivation of
YtbE in the presence of (R)-CMM exhibited little dependence on
the enzyme concentration (Fig. 1d). At high enzyme concentrations,
enzymes obviously have more chances to form aggregates usu-
ally through forming intermolecular disulphide bonds. This result
therefore strongly argues that YtbE inactivation occurs through
intramolecular interactions rather than intermolecular ones.
t
t
2
ˇ × k1(e−k − e−k
)
1
t
Ra = e−k
+
(7)
1
(k2 − k1)
where N, U and D stand for the concentrations of the enzyme in the
native, unfolded and fully inactivated states, respectively. ki stands
for the rate constant. ˇ stands for activity ratio between U and N.
N0 stands for the initial concentration of the active enzyme.The
fitting with the experimental data were made using the software
OriginPro 8.0.
3.2. YtbE unfolding and dimerization during inactivation
2.6. Site-directed mutagenesis of YtbE
For managing the YtbE instability problem in desired commer-
cial application, it is useful to understand inactivation mechanism
and the consequent alterations to the enzyme configuration [15].
To unravel these myths, we relied on the binding capacity of flu-
orescence probe ANS to the hydrophobic domains of YtbE. When
YtbE is folded or activated, access to these domains by ANS is gen-
erally known to be limited and hence binding capacity can be low.
As the effect of temperature on enzyme unfolding is well known,
the results in Fig. 2a validate the experiment. The ANS fluorescence
emission exhibited a notable blue shift and an increase in intensity
obviously caused by more ANS binding to the enzyme, once again
confirming the marked unfolding of YtbE at 50 ◦C. Fig. 2b clearly
corroborates with our previous experimental deduction that the
product (R)-CMM has a similar effect as temperature.
As a means to modify this enzyme stability, mutation of residue
at Cys183 by Ser on the plasmid pET28a DNA was made using the
QuikChange Site-directed Mutagenesis procedure and confirmed
by sequencing. The two primer sequences are,
5ꢀ-primer: 5ꢀ-TCACACAGAA AGAGTTGATAAGATACTCTCAAAAT
CAGG
3ꢀ-primer:3ꢀ-CCTCCATCTGGATGCCCTGATTTTGAGAGTATCTTATC
2.7. Multi-factors for improving enzyme stability
YtbE, 0.3 mg/ml, mixed with glycerol (10, 20, or 30% w/v), glu-
cose (10, 20, or 30% w/v), Tween 80 (0.002% w/v), KCl (0.1, 0.5, or
1 M), sodium gluconate (30% w/v), potassium gluconate (30% w/v),
or DTT (0.2, 1, or 2 mM), was incubated at 30 ◦C in the presence of
25 mM (R)-CMM for 30 min. This was followed by enzyme activity
assay.
YtbE thermal unfolding at 50 ◦C was further testified through
UV–Vis wavelength scanning [16]. By 30 min at this tempera-
ture there is hardly any detectable YtbE activity (Fig. 1a), but the
protein unfolding evidently has continued (Fig. 2c). After YtbE
inactivation, we did not observe any recovery of the enzymatic
activity. Irreversible change or incorrect refolding of the local struc-
ture around the active site is likely attributed to YtbE irreversible
inactivation. To cite an example, formation of an altered enzyme
conformation causes irreversible thermal inactivation of mitochon-
drial aspartate aminotransferase (mAAT) [17].
Formation of aggregates is a further clue for possibly unravel-
ing the enzyme inactivation mechanism [10,18]. Compared to the
purified YtbE under the reducing condition (Lane 6, Fig. 3), that in
the native SDS–PAGE was a single-banded monomer. This result
confirmed the complete separation of protein aggregates possibly
forming during the expression and purification processes. In this
study, as seen from the native SDS–PAGE (Fig. 3), YtbE undergo-
ing the pursued inactivation formed a dimer and the amount of
this aggregate appears to increase when the presenting time of (R)-
CMM augments. Disappearance of the YtbE dimer in the presence
of 2-mercaptoethanol reassures the role of intermolecular disulfide
bonds in the YtbE inactivation process.
As reported earlier [10,18], the YtbE inactivation using the
dimerization mechanism may be irreversible and formation of
disulfide bonds may effectively quench enzymatic activity usu-
ally if one of the thiol groups is located at or near the active site.
YtbE molecule contains only one cysteine residue, Cys183, which
is located inside the protein such that its dimers cannot be formed
until the enzyme is unfolded [19]. Even though this dimerization
results in irreversible YtbE inactivation, its proportional small pres-
ence (Fig. 3) shows that it cannot be the predominant mechanism
leading to YtbE inactivation. This therefore indicates that the major-
ity of YtbE was inactivated in its monomeric state mostly through
unfolding.
2.8. Enzymatic (R)-CMM synthesis coupled with cofactor
regeneration
The synthesis experiment was carried out at 30 ◦C for 1.5 h with
magnetic stirring in a 10-ml reaction vessel containing 100 mM
CBFM (diluted from 1 M CBFM solution in ethanol), 150 mM glu-
cose (0.27 g), 0.1 mM NADP+, 30 mg YtbE (1.3 U/mg), 5 mg GDH
(15.3 U/mg) (both were crude enzyme powders) and 100 mM phos-
phate buffer (pH 7.0). The pH was automatically adjusted to 7.0 by
titrating 1 M Na2CO3 or K2CO3 solution with Metrohm 902 Titrando
(Switzerland). The reaction progress was monitored via GC mea-
surements (Agilent 6820 gas chromatograph, CP-Chirasil-Dex CB
column, inj.: 280 ◦C, det.: 280 ◦C, col.: 180 ◦C). The residual activity
of YtbE was periodically assayed as described above. The modi-
fied reaction system with various additives added in the reaction
mixture was shown in Table 1.
For whole-cell bioreduction, CBFM (2.0 g, 1 M) was transformed
at 30 ◦C for 5 h in phosphate buffer (10 ml, 100 mM, pH 7.0) con-
taining lyophilized E. coli cells (0.3 g) harboring pET-YtbE-GDH and
glucose (2.7 g, 1.5 M).
3. Results and discussion
3.1. The inactivation of YtbE by temperature, substrate and
product
The following effects on the YtbE enzyme stability were exam-
ined for the coupled enzyme reactions depicted in Scheme 1:
temperature, stirring speed (data not shown), pH (data not shown),
CBFM concentration (the substrate), (R)-CMM concentration (the
product), glucose concentration (the substrate for enzyme GDH)
and sodium gluconate (the product of enzyme GDH). YtbE thermal
inactivation generally follows an exponential trend, and is bear-
able at <40 ◦C and unbearable at >45 ◦C (Fig. 1a). The substrate CBFM
and the product (R)-CMM, especially the latter, exert more negative
effects on the YtbE stability (Fig. 1b and c) and this is so irrespective
3.3. Kinetic modeling of YtbE inactivation process
Mathematical simulation in describing the effects of tempera-
ture and the other factors on YtbE catalytic activity is instrumental

Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
111
Table 1
Asymmetric reduction of CBFM with additivesa.
Entry
Enzyme reaction condition
KCl (M) Glycerol (%)
Enzyme reaction outcome
Half-life of YtbE (min)
DTT (mM)
Conversion of CBFM (%)
1b
2c
3c
4c
5c
6c
0
0
0
0
10
20
20
0
5.50
25.8
34.9
18.5
30.4
28.6
62.9
48.1
54.6
74.3
98.7
94.0
0.5
1.0
0
0
0
1.0
1.0
1.0
1.0
0
a
Reaction conditions: CBFM (1 ml, 100 mM), glucose (150 mM), YtbE powders (20 mg), GDH powders (5.0 mg), phosphate buffer (100 mM, pH 7.0, 9 ml), NADP+ (1 mM),
additives, pH was kept at 7.0, 30 ◦C for 1.5 h.
b
The pH was adjusted at 7.0 with 1 M Na2CO3 and the buffer was sodium phosphate.
The pH was adjusted at 7.0 with 1 M K2CO3 and the buffer was potassium phosphate.
c
for achieving the objective of the present work [20,21]. Although
YtbE inactivation by heat or substrate (CBFM) could be approached
by a first-order kinetic, that induced by product (R)-CMM clearly
showed a non-first-order kinetics. Consequently, an inactivation
pathway encompassing two consecutive, serial first-order reac-
tions was assumed, with the first stage (U) being at the reduced level
of enzyme activity and the second stage (D) being no remaining
activity (Eq. (1)).
is accompanied by an unfolding process and all our results (e.g. to
examine any enzyme recovery possibility by cooling the enzyme
following a warming period) indicated that the YtbE inactivation is
almost certain to be irreversible. Our speculation at present is that
tertiary and/or quaternary structure changes prior to the secondary
structure play important roles in the loss of the enzyme activity
and formation of dimers plays a small part in the YtbE inactivation
process.
The overall curve fitting results show that this simulation
approach has been working (Fig. 1c) and may have reflected, to
some extent, the genuine enzyme inactivation mechanism.
Based on the parallel experimental results of YtbE catalytic
activity and the structural unfolding as probed by the fluorescence
probe and spectrophotometric scanning results, YtbE inactivation
3.4. Search for additives to stabilize YtbE
Since the YtbE inactivation process at least consists of, to a large
extent, enzyme unfolding and, to a less extent, enzyme dimer-
ization, we endeavored to primarily stabilize the conformation or
Fig. 1. Experimental (discrete points) and model-fitted [continuous curves] enzyme inactivation (as residual activity) with time at 0.3 mg/ml YtbE in response to differed factors.
(a) In response to incubation temperatures (ꢀ 30 ◦C; ᭹ 40 ◦C; ꢁ 45 ◦C; ꢂ 50 ◦C), with model fitting by first-order kinetic. (b) In response to substrate CBFM concentration (ꢀ
0 mM; ꢃ 20 mM; ᭹ 30 mM; ꢁ 40 mM; ꢂ 50 mM), with temperature at 30 ◦C and model fitting by first-order kinetic. (c) In response to product (R)-CMM concentration (ꢀ
0 mM; ꢄ 10 mM; ꢃ 20 mM; ꢁ 25 mM; ᭹ 30 mM), with temperature at 30 ◦C and model fitting by Eq. (7). (d) In response to variation of YtbE concentration (ꢀ 0.3 mg/ml; ᭹
1.5 mg/ml; ꢁ 3.0 mg/ml) per se with temperature at 30 ◦C and (R)-CMM concentration at 25 mM.

112
Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
Fig. 3. Native SDS–PAGE of YtbE (0.6 mg/ml) at different inactivation degrees. Lane M,
protein standards; Lane 1, incubation at 50 ◦C for 5 min, Ra (residual activity) 65.1%;
Lane 2, incubation at 50 ◦C for 15 min, Ra 27.6%; Lane 3, incubation at 30 ◦C for 10 min
with 30 mM (R)-CMM, Ra 34.6%; Lane 4, incubation at 30 ◦C for 30 min with 30 mM
(R)-CMM, Ra 14.4%; Lane 5, incubation at 30 ◦C for 60 min with 30 mM (R)-CMM, Ra
5.9%; Lane 6, reducing SDS–PAGE of the protein in lane 5 (adding 2-mercaptoethanol
during sample processing); Lane 7, Ra 100%.
enzyme configuration and, according to the aforementioned inac-
tivation mechanism for YtbE, such agent may well retard the loss
of this enzyme activity [9,22].
We have also noted that certain sugars and polyols can
strengthen the hydrophobic interactions occurring among non-
polar amino acid residues and thus make the protein molecules
more resistant to inactivation surroundings [9]. Furthermore, it has
long been known that a range of low-molecular weight molecules
greatly disfavor protein unfolding owing to induction of prefer-
ential hydration of the protein [23]. Glucose, glycerol, Glc-K and
Glc-Na fall into this category. Glycerol generated the best result,
and this is followed by glucose and Glc-K almost on an equal footing
(Fig. 4). However, the outcome for 30% Glc-Na has been disappoint-
ing (Fig. 4). Comparison between KCl, Glc-K and Glc-Na shows that
potassium ion has a most favorable role in containing the loss of this
Fig. 2. Spectroscopic analysis of the structural changes of YtbE during inactivation
process. (a) Fluorescence spectra of ANS (0.1 mM) in the presence of 0.12 mg/ml
YtbE after incubation at 50 ◦C for different times (black, 0 min; blue, 3 min; green,
6 min; red, 11 min; magenta, 30 min). (b) Fluorescence spectra of ANS (0.1 mM) in
the presence of 0.12 mg/ml YtbE after incubation at 30 ◦C for 30 min with different
concentrations of (R)-CMM (magenta, only ANS; black, 0 mM; blue, 10 mM; green,
20 mM; red, 30 mM). (c) UV–Vis profiles of YtbE after 50 ◦C incubation for different
times (black, 0 min; blue, 10 min; green, 30 min; red, 60 min) followed by 4 ◦C incu-
bation for 1 h. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)
secondarily to minimize the YtbE dimer formation by searching for
inexpensive small molecules to reduce the loss of YtbE activity.
Compared to the control where no additive was employed, the
use of 1 M KCl resulted in a marked reduction in the enzyme cat-
alytic activity (Fig. 4). Our justification is that KCl would neutralize
intra-molecular electrostatic interactions via binding to charged or
dipole groups on the protein molecule and lead to salting-out. A
documented consequence for this approach is compression of the
Fig. 4. Residual activity of YtbE in the presence of 25 mM (R)-CMM after incubation at
30 ◦C for 30 min with different modifications. The concentration of each additive was
optimized for maximizing the residual activity.

Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
113
enzyme activity. Our fluorescent probe ANS analysis indicates that
Glc-K may well bind the YtbE enzyme and consequently inhibit the
unfolding of YtbE whereas Glc-Na facilitates the unfolding of YtbE
in the presence of (R)-CMM (data not shown).
Tween-80 is a nonionic surfactant and can be justified to fall into
this category of de-stabilizing mechanism as well. Unfortunately, it
shows to have no observable impact on the stability of this enzyme.
As the formation of YtbE dimer also has a contribution to
the inactivation of this enzyme, cysteine-to-serine substitution
at Cys183 was introduced. Unfortunately, the mutant was more
inclined to be inactivated. Instead, we tested DTT in the present
experiment. Following incubation with (R)-CMM, DTT indeed
results in a decent level of increment when compared with the
control without any additives (Fig. 4). As the reductant DTT mode-
rates YtbE dimerization, this result not only corroborates the role of
dimerization in quenching this enzyme activity but also supports
our earlier finding that the majority of this enzyme loses its activity
through unfolding mechanism.
Fig. 5. Asymmetric reduction of 1.0 M CBFM by lyophilized cells of recombinant E. coli
(YtbE/GDH). (ꢀ Conversion of the control reaction; ᭹ conversion of the modified
reaction; ꢁ residual activity of YtbE in the control reaction; ꢂ residual activity of YtbE
in the modified reaction). Reaction conditions for the control reaction: CBFM (2.0 g,
1.0 M), glucose (3.0 g, 1.5 M), lyophilized cells of E. coli (pET28a-YtbE-GDH) (30 g/l),
sodium phosphate buffer (100 mM, pH 7.0, 10 ml), the pH was adjusted at 7.0 with
1 M Na2CO3, 30 ◦C. Reaction conditions for modified reaction: CBFM (2.0 g, 1.0 M),
glucose (3.0 g, 1.5 M), lyophilized cells of E. coli (pET28a-YtbE-GDH) (30 g/l), glycerol
(2.0 g, 20%), DTT (1 mM), potassium phosphate buffer (100 mM, pH 7.0, 8 ml), the pH
was adjusted at 7.0 with 1 M K2CO3, 30 ◦C.
Backed by a framework for rationalization, introduction of small,
inexpensive molecules, glycerol, glucose and Glc-K, prepared us for
achieving an improved YtbE-mediated bioreduction reaction in this
work.
3.5. YtbE-catalyzed bioreduction guided by a systematic
stabilization strategy
As a means to manage the daunting pricing regime for the
cofactors and based on the set of results obtained, we thus used
30 g/l recombinant E. coli cells coexpressing YtbE and GDH to carry
out the present asymmetric reduction out of 1 M CBFM. The con-
version of substrate (CBFM) indeed was incremented from 70.5% to
96.6% by simply adding 20% glycerol and 1 mM DTT (Fig. 5). Clearly,
the increased conversion rate mostly results from the improve-
ment of the YtbE stability. This is backed by the observation that
the half-life of YtbE during bioreduction correspondingly increases
from 46.6 min to 159 min. Compared with the conversion rate at
1 M CBFM and using 50 g/l E. coli at 30 ◦C when a conversion rate
of 93% was reported [8], the productivity (g(R)-CMM per gYtbE) was
increased from 3.72 to 6.44. The proposed new strategy of this work
potentially opens a horizon of practical applications.
Thermodynamically it is possible to stoichiometrically convert
a high concentration of CBFM into enantiopure (R)-CMM at 20 ◦C
[8], but kinetically the low reaction rate at such a low temperature
may restrict its commercial uses. Indeed, the large-scale manufac-
turing was greatly impeded by the poor stability of YtbE, especially
under the aforementioned reaction conditions. Taking into account
all the above facets and results, we proposed a systematic stabi-
lization strategy for moderating the activity loss of YtbE during the
targeted bioreduction process. Glycerol, glucose, Glc-K and NADP+
were selected rationally to retard unfolding of YtbE whereas DTT
was employed to suppress dimerization of the same.
In the absence of those small-molecule additives, whilst the
substrate conversion of 62.9% may be accommodated, YtbE half-
life of 5.5 min is clearly detrimental (Entry 1 of Table 1) and this
shows how serious this enzyme inactivation problem has been. In
view of the indigenous presence of glucose to this enzymatic reac-
tion system (see Scheme 1), focus was given to adjusting KCl and
DTT concentrations. The half-life of YtbE in this reaction system
increases significantly from 5.5 min to 25.8 min in the co-presence
of 0.5 M KCl and 1.0 mM DTT (entry 2). Encouragingly, upon increas-
ing of KCl from 0.5 to 1 M KCl, we gained further 9.1 min (from 25.8
to 34.9 min) for the YtbE half-life (entry 3 of Table 1). KCl however
seems to possess a propensity to reduce the reaction rate (entries
1, 2 and 3 of Table 1). Speculatively, a salting-out effect induced by
KCl may have improved the YtbE stability, but KCl may have notice-
ably created mass transfer barrier and reduced the CBFM solubility.
The use of glycerol not only gained on the half-life of this enzyme,
but also significantly improved the kinetics of this reaction system
(entries 4, 5 and 6 of Table 1) and this is consistent with its known
reputation in stabilizing YtbE (Fig. 4).
4. Conclusions
We proposed a systematic strategy for improving the ther-
mal and operational stability of YtbE and achieved successfully in
enhancing the CBFM bioreduction based on guidance provided by
our understandings on the YtbE inactivation mechanisms, which
were gained through biochemical analysis and kinetic modeling.
For the further pilot plant outcome, this established approach has
made a success in that the productivity (g(R)-CMM per gYtbE) was
increased from 3.72 to 6.44, and thus has suggested a commer-
cially viable reduction of the production cost. It is also hoped that
the strategy endeavored which bears elements to base on inacti-
vation mechanism can be not only used for other complex enzyme
systems but also developed further in the time to come.
In addition, the pH was adjusted with 1 M K2CO3 and the sodium
phosphate buffer was replaced by potassium phosphate buffer. Sur-
prisingly, the conversion increased to 74.3% in the presence of 10%
glycerol and further to 98.7% in the presence of 20% glycerol. At
the same time, the half-life of YtbE during bioreduction climbed to
30.4 min with the use 20% glycerol. A high conversion leads to a high
(R)-CMM concentration, but the presence of a high (R)-CMM con-
centration considerably mars the stability of the YtbE. Conceivably,
the stabilizing effect of glycerol on the present enzyme reaction
system has been vital.
Acknowledgments
Thanks to financial supports from National Natural Science
Foundation of China (Nos. 31200050
of Science and Technology, P.R. China (Nos. 2011AA02A210 &
2011CB710800), and Huo Yingdong Education Foundation.
& 21276082), Ministry
References
[1] T. Ema, S. Ide, N. Okita, T. Sakai, Adv. Synth. Catal. 350 (2008) 2039–2044.

114
Y.-P. Xu et al. / Journal of Molecular Catalysis B: Enzymatic 104 (2014) 108–114
[2] A. Glieder, R. Weis, W. Skranc, P. Poechlauer, I. Dreveny, S. Majer, M. Wubbolts,
[12] A. Sadana, J.P. Henley, Biotechnol. Bioeng. 30 (1987) 717–723.
[13] A. Sadana, Biotechnol. Adv. 6 (1988), 349-IN342.
[14] Y. Ni, C.X. Li, H.M. Ma, J. Zhang, J.H. Xu, Appl. Microbiol. Biotechnol. 89 (2011)
1111–1118.
H. Schwab, K. Gruber, Angew. Chem. Int. Ed. 42 (2003) 4815–4818.
[3] C.S. Zhang, Z.J. Zhang, C.X. Li, H.L. Yu, G.W. Zheng, J.H. Xu, Appl. Microbiol.
Biotechnol. 95 (2012) 91–99.
[4] K. Kinbara, Y. Tagawa, K. Saigo, Tetrahedron: Asymmetry 12 (2001) 2927–2930.
[5] T. Ema, N. Okita, S. Ide, T. Sakai, Org. Biomol. Chem. 5 (2007) 1175–1176.
[6] M. Jeong, Y.M. Lee, S.H. Hong, S.Y. Park, I. Yoo, M.J. Han, Biotechnol. Lett. 32
(2010) 1529–1531.
[15] M. Ladero, A. Santos, F. García-Ochoa, Enzyme Microb. Technol. 38 (2006)
1–9.
[16] D.Y. Fu, C. Li, J. Lu, A. Rahman, T.W. Tan, J. Mol. Catal. B: Enzym. 66 (2010)
136–141.
[7] H.M. Ma, L.L. Yang, Y. Ni, J. Zhang, C.X. Li, G.W. Zheng, H.Y. Yang, J.H. Xu, Adv.
Synth. Catal. 354 (2012) 1765–1772.
[8] Y. Ni, J. Pan, H.M. Ma, C.X. Li, J. Zhang, G.W. Zheng, J.H. Xu, Tetrahedron Lett. 53
(2012) 4715–4717.
[9] P.V. Iyer, L. Ananthanarayan, Process Biochem. 43 (2008) 1019–1032.
[10] S. Chen, F.C. Ye, Y. Chen, H.X. Zhao, R. Yatsunami, S. Nakamura, F. Arisaka, X.H.
Xing, Biotechnol. Bioeng. 108 (2011) 1841–1851.
[11] I. Dib, A. Slavica, W. Riethorst, B. Nidetzky, Biotechnol. Bioeng. 94 (2006)
645–654.
[17] J.M. Lawton, S. Doonan, Biochem. J. 334 (1998) 219.
[18] R.W. Lencki, J. Arul, R.J. Neufeld, Biotechnol. Bioeng. 40 (1992) 1421–1426.
[19] J. Lei, Y.F. Zhou, L.F. Li, X.D. Su, Protein Sci. 18 (2009) 1792–1800.
[20] M. Ladero, R. Ferrero, A. Vian, A. Santos, F. Garcia-Ochoa, Enzyme Microb. Tech-
nol. 37 (2005) 505–513.
[21] Z.D. Zhang, Z.M. He, M.X. He, J. Mol. Catal. B: Enzym. 14 (2001) 85–94.
[22] A.M. Klibanov, Adv. Appl. Microbiol. 29 (1983) 28.
[23] C. Ó’Fágáin, Enzym Microb. Technol. 33 (2003) 137–149.