Engineering the large pocket of an (S)-selective transaminase for asymmetric synthesis of (S)-1-amino-1-phenylpropane

Article abstract of DOI:10.1039/d0cy02426k

Amine transaminases offer an environmentally benign chiral amine asymmetric synthesis route. However, their catalytic efficiency towards bulky chiral amine asymmetric synthesis is limited by the natural geometric structure of the small pocket, representing a great challenge for industrial applications. Here, we rationally engineered the large binding pocket of an (S)-selective ?-transaminase BPTA fromParaburkholderia phymatumto relieve the inherent restriction caused by the small pocket and efficiently transform the prochiral aryl alkyl ketone 1-propiophenone with a small substituent larger than the methyl group. Based on combined molecular docking and dynamic simulation analyses, we identified a non-classical substrate conformation, located in the active site with steric hindrance and undesired interactions, to be responsible for the low catalytic efficiency. By relieving the steric barrier with W82A, we improved the specific activity by 14-times compared to WT. A p-p stacking interaction was then introduced by M78F and I284F to strengthen the binding affinity with a large binding pocket to balance the undesired interactions generated by F44. T440Q further enhanced the substrate affinity by providing a more hydrophobic and flexible environment close to the active site entry. Finally, we constructed a quadruple variant M78F/W82A/I284F/T440Q to generate the most productive substrate conformation. The 1-propiophenone catalytic efficiency of the mutant was enhanced by more than 470-times in terms ofk_cat/K_M, and the conversion increased from 1.3 to 94.4% compared with that of WT, without any stereoselectivity loss (ee > 99.9%). Meanwhile, the obtained mutant also showed significant activity improvements towards various aryl alkyl ketones with a small substituent larger than the methyl group ranging between 104- and 230-fold, demonstrating great potential for the efficient synthesis of enantiopure aryl alkyl amines with steric hindrance in the small binding pocket.

Full text of DOI:10.1039/d0cy02426k

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Engineering the large pocket of an (S)-selective
transaminase for asymmetric synthesis of (S)-1-
amino-1-phenylpropane†
Cite this: Catal. Sci. Technol., 2021,
11, 2461
a
a
a
a
b
Youyu Xie, Feng Xu, Lin Yang, He Liu, Xiangyang Xu,
a
a
Hualei Wang * and Dongzhi Wei
*
Amine transaminases offer an environmentally benign chiral amine asymmetric synthesis route. However,
their catalytic efficiency towards bulky chiral amine asymmetric synthesis is limited by the natural
geometric structure of the small pocket, representing a great challenge for industrial applications. Here, we
rationally engineered the large binding pocket of an (S)-selective ω-transaminase BPTA from
Paraburkholderia phymatum to relieve the inherent restriction caused by the small pocket and efficiently
transform the prochiral aryl alkyl ketone 1-propiophenone with a small substituent larger than the methyl
group. Based on combined molecular docking and dynamic simulation analyses, we identified a non-
classical substrate conformation, located in the active site with steric hindrance and undesired interactions,
to be responsible for the low catalytic efficiency. By relieving the steric barrier with W82A, we improved
the specific activity by 14-times compared to WT. A π–π stacking interaction was then introduced by M78F
and I284F to strengthen the binding affinity with a large binding pocket to balance the undesired
interactions generated by F44. T440Q further enhanced the substrate affinity by providing a more
hydrophobic and flexible environment close to the active site entry. Finally, we constructed a quadruple
variant M78F/W82A/I284F/T440Q to generate the most productive substrate conformation. The
1
-propiophenone catalytic efficiency of the mutant was enhanced by more than 470-times in terms of
Received 19th December 2020,
Accepted 28th January 2021
kcat/K , and the conversion increased from 1.3 to 94.4% compared with that of WT, without any
M
stereoselectivity loss (ee > 99.9%). Meanwhile, the obtained mutant also showed significant activity
improvements towards various aryl alkyl ketones with a small substituent larger than the methyl group
ranging between 104- and 230-fold, demonstrating great potential for the efficient synthesis of
enantiopure aryl alkyl amines with steric hindrance in the small binding pocket.
DOI: 10.1039/d0cy02426k
rsc.li/catalysis
present stringent stereoselectivity, no requirement of external
cofactor supplementation, a broad substrate spectrum, and
Introduction
Approximately 95% of drugs are predicted to be chiral by
020 and share a market of almost $5b in chiral ligands and
high stability, which makes them attractive for industrial
process development.
7
–9
2
1
,2
agrochemicals. Chiral amines play a fundamental role in a
series of pharmaceutical drugs, and approximately 40% of
Transaminases require pyridoxal 5′-phosphate (PLP) as a
coenzyme that mediates the transamination of prochiral keto
acids or ketones by transferring an amine group from an
3
them are estimated to contain a chiral amine functionality.
1
0
The asymmetric synthesis of chiral amine drug intermediates
by enzymatic catalysis strategies has attracted growing
attention due to the social and environmental demand for
amino donor. The active site of (S)-selective transaminases
is composed of a small and a large binding pocket (SBP and
1
1
LBP, respectively) in the dimerization of monomers. In
most cases, the large pocket can accommodate substituents
4
–6
green processes in the chemical industry.
Transaminases
1
2
with a rather broad size distribution, while the small pocket
1
3
is restricted to accept only a methyl group. The steric
constraint in the active site, especially that in the SBP, has
been the crucial limiting factor for the production of various
bulky chiral amines, representing a great challenge for
industrial synthetic applications.
a
State Key Laboratory of Bioreactor Engineering, New World Institute of
Biotechnology, East China University of Science and Technology, Shanghai 200237,
People's Republic of China. E-mail: hlwang@ecust.edu.cn, dzhwei@ecust.edu.cn
b
Zaozhuang Jienuo Enzyme Co., Ltd., Shandong Province 277116, PR China
To boost the synthesis utility of ω-TAs toward bulky
prochiral ketones, protein engineering has been widely
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
d0cy02426k
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2461–2470 | 2461

View Article Online
Paper
Catalysis Science & Technology
carried out to overcome the limitation of the binding pocket
geometry by using directed evolution and semi-rational/
conversion of the target ketone 2b (10 mM) was only 1.3% at
48 h, indicating that the ethyl group of the model substrate
was indeed not easily accommodated by the ω-transaminase
SBP due to the limitation of the natural SBP geometry.
rational
design
approaches
for
different
amine
1
4,15
transaminases.
An impressive example was highlighted
by Merck and Co. and Codex, applying the direct evolution of
an (R)-selective ω-transaminase ATA-117 for enlarging the
LBP and SBP to the highly efficient asymmetric synthesis of
Molecular modeling of the BPTA active site
Molecular docking and MD simulations were performed to
analyze the interaction between the ketone and the residues
in the active site. We used the ω-transaminase crystal
structure with the PMP cofactor from Chromobacterium
1
4
bulky chiral sitagliptin.
Regarding rational (S)-TA
engineering, most studies focused on engineering the LBP
and SBP to modify the acceptance of bulky substrates.
1
6
Dourado et al.
engineered the LBP and SBP of the
2
6
violaceum (PDB ID: 6s4g) with 58.84% identity as a template
for comparative modeling. The substrate 2b was docked into
the BPTA active site with the PMP cofactor, and the top-
ranked structure with the lowest binding energy was chosen
as the initial conformation for the MD simulation (Fig. 1A).
During 50 ns of MD simulation, the result showed that the
substrate could not form a stable attack conformation in the
active site and would move out of the BPTA active site after
transaminase Vibrio fluvialis TAm and achieved a high
catalytic activity for the chiral synthesis of (1S)-1-(1,1′-
1
7
biphenyl-2-yl)ethanolamine. Midelfort et al. engineered the
LBP and SBP of Vibrio fluvialis TAm and improved the
catalytic conversion of (R)-ethyl-5-methyl-3-oxooctanoate by
1
8
6
0-times. Han et al. expanded the bulky substrate specificity
for aryl alkylamines and alkylamines of ω-transaminase OATA
by rational LBP remodeling. However, successful examples of
SBP engineering to accommodate bulky substrates are still
8.3 ns. The distance between the PMP's exocyclic nitrogen
and the ketone's carbonyl carbon atom (d1) showed two
major increases from 0 to 8.3 ns and 8.3 to 47.4 ns. The first
increase occurred at 8.3 ns, where d1 increased to 7.5 Å, and
the second increase at 47.4 ns with d1 sharply increasing to
1
5,18–24
needed to be developed.
Most residues in the SBP were
adjacent to PLP and stayed at the interface of the
homodimer, which played an important role in the
stabilization and stereoselectivity of the ω-transaminase. In
most cases, modifying the residues in the SBP would result
>
10.0 Å (Fig. 1A). This phenomenon was similar to that
1
3
described by Voss et al. and indicated that the pocket in the
active site of BPTA failed to generate sufficient affinity to
stabilize the Michaelis complex between the ketone and
2
5
in catalytic activity loss toward prochiral ketones. Therefore,
it is rather difficult to improve the catalytic efficiency for aryl
alkyl substrates with alkyl chains larger than a methyl group
by engineering the SBP.
2
7
PMP, which was necessary to activate the transamination
catalysis. The conformation before 8.3 ns in the MD
simulation represents a plausible attack conformation that
could initiate external aldimine formation. Therefore, we
extracted the average structure before 8.3 ns in the MD
simulation for subsequent analysis.
In this study, we rationally engineered the LBP of an (S)-ω-
transaminase BPTA from Paraburkholderia phymatum to
address the challenge of aryl alkyl substrates being generally
hardly accommodated in the SBP by relieving steric
hindrance and strengthening the interaction upon substrate
binding in the LBP.
Rational design of the large pocket of the active site
Based on the modeling results, the small BPTA binding
pocket consisted of F44, L81, F110, Y175, and G341, while
the large one was constructed by M78, W82, A253, I284,
R438, and T440 (Fig. S2†). As Scheme 1 shows, the classical
nucleophilic attack conformation indicated that a small
substituent of the ketone is located in the SBP and the large
substituent adjacent to the carbonyl group stayed in the LBP.
Results and discussion
Initial screening of ω-transaminases
Although transaminases have a broad substrate range, most
of them accept only substituents no larger than a methyl
group at the position adjacent to the carbonyl moiety of the
1
4
ketone substrate. To identify a potential candidate for the
asymmetric synthesis of bulky chiral pharmaceutical
intermediates, we selected (S)-1-amino-1-phenylpropane (1b)
However, the docking results showed
a non-classical
substrate conformation in the active site, with the phenyl
moiety pointing out of the LBP, while the small substituent
pointed towards the LBP (Fig. 1A). This non-classical
conformation seemed to be the main reason for the low
(precursor of corticotropin-releasing factor type I antagonist)
as the model substrate for initial screening, which is
generally difficult to accommodate by the SBP of most
ω-transaminases due to the small ethyl group substituent of
the prochiral ketone. A total of 72 in-house natural ω-TAs
1
3,28
catalytic activity towards the model substrate.
The interactions between the substrate and residues in
the active site were analyzed to further explore the reason for
the abnormal conformation mentioned above. As shown in
Fig. 1B, the indole moiety of W82 showed a huge steric
hindrance with ketones, which hampered their efficient
binding to the SBP and LBP in the active site. Meanwhile, the
side chain of F44 generated an undesired π–π stacking
(not supplied) were screened in the thermodynamically
favored kinetic mode using an (S)-amine as a donor and
pyruvate as an acceptor. The results showed that only one
constructed ω-transaminase BPTA from Paraburkholderia
phymatum STM815 (accession number: WP_012402885.1)
could be determined with poor detectable activity, and the
2462 | Catal. Sci. Technol., 2021, 11, 2461–2470
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Fig. 1 Docking model of 1-propiophenone in the BPTA active site. (A) The residues of the active site in the docking model. The green and orange
sticks represent the substrate ketone and PMP, respectively. The yellow dotted line between the exocyclic nitrogen of PMP and the carbonyl
carbon atom of the ketone represents the distance. The brown dotted line between the phenyl group of F44 and the phenyl moiety of the
substrate represents the π–π stacking interaction between them. (B) Side view of the ketone substrate phenyl group interaction visualization with
the W82 indolyl group, shown as a blue CPK representation. The yellow dotted line represents the distance between the exocyclic nitrogen of
PMP and the carbonyl carbon atom of the ketone. Orange sticks indicate the internal aldimine. (C) White, orange, and green sticks indicate the key
mutation residues in the docking conformation of variant W82A/M78F/I284F/T440Q, the internal aldimine, and the ketone, respectively.
interaction with the ketone, and the stacking interaction
prevented the phenyl group from being properly
accommodated in the LBP. We speculated that the steric
hindrance of W82 and undesired interaction of F44 with the
substrate might result in poor catalytic efficiency (Fig. 1A).
To alleviate the steric hindrance and maintain the
hydrophobic environment at the same time, hydrophobic
and smaller substituents (FLIVAG) were selected to replace
the W82 residue with the phenyl group. The results showed
that there are two positive mutations W82L and W82A with
better catalytic activity, improving by 3.89 and 14.00-times in
compared to WT (Table 1), respectively. Both variants
k
cat/K
M
showed a sharply decreased K . Among them, the KM value
M
of W82A decreased from 92.86 to 11.00 mM, demonstrating
significant improvement in affinity with the substrate by the
W82A mutation. Considering that F44 in the SBP was at the
interface of the dimer, there was no particular rule to modify
this position, and the saturated mutation of F44 was carried
out. Unfortunately, the saturation mutations at F44 showed
no positive results compared to WT towards 2b, and most of
the variants showed no detectable catalytic activity (Fig. S3†).
This result indicated that it would be critical to engineer the
position at the interface, which is correlated to enzyme
1
3
stability and enantioselectivity.
To eliminate the
unfavorable interaction produced by F44 in the SBP, the
residues in the LBP (M78, W82, A253, I284, R438, and T440)
in the opposite direction (LBP) were investigated for their
potential in improving the binding interaction with the
phenyl group of the ketone in the LBP. Among them, residue
R438 plays an important role in dual substrate recognition,
and is responsible for forming a salt bridge with the
Scheme 1 The classical substrate binding conformation in the active
site using L-Ala as an amine donor and 1-propiophenone as a bulky
ketone acceptor, as a model. (A) Binding of the amine donor L-Ala to
the PLP form of the ω-TAs (E-PLP). (B) Binding of the substrate ketone
acceptor 1-propiophenone to the pyridoxamine 5′-phosphate form of
the ω-TAs (E-PMP).
29
carboxylate group of the donor and acceptor substrate,
3
0,31
while A253 participates in the binding of PLP.
Therefore,
these two conserved sites were excluded from our
consideration for modification.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2461–2470 | 2463

View Article Online
Paper
Catalysis Science & Technology
Table 1 Kinetic parameters towards 1-propiophenone in kinetic resolution mode
−1
−1 −1
Variant
K
M
[mM]
kcat [s
]
kcat/K
M
[M
s
]
Fold increase in kcat/K
M
BPTA-WT
W82A
W82L
W82A/M78F
92.86 ± 2.16
11.00 ± 0.24
30.10 ± 1.17
4.90 ± 0.44
6.00 ± 0.59
7.94 ± 0.26
3.14 ± 0.19
5.32 ± 0.93
3.48 ± 0.48
1.67 ± 0.11
0.68 ± 0.03
1.12 ± 0.07
0.85 ± 0.01
3.31 ± 0.04
1.73 ± 0.15
1.37 ± 0.05
4.92 ± 0.78
1.96 ± 0.14
4.42 ± 0.33
5.71 ± 0.54
7.27 ± 0.17
101.82 ± 5.59
28.29 ± 2.36
681.07 ± 16.99
288.33 ± 5.63
172.54 ± 4.38
1566.88 ± 48.69
368.42 ± 14.43
1270.11 ± 12.21
3419.16 ± 45.63
1.00
14.01
3.89
93.70
39.67
23.74
215.56
50.68
174.73
470.38
W82A/I284F
W82A/T440Q
M78F/W82A/I284F
W82A/I284F/T440Q
M78F/W82A/T440Q
M78F/W82A/I284F/T440Q
The residues M78 and I284 were located directly opposite
to F44 (Fig. 1A). We speculated that introducing an opposite
π–π stacking interaction would be helpful for the phenyl
substituent of the substrate binding in the LBP. Thus, we
replaced M78 and I284 with hydrophobic and aromatic
substituents (WF) using W82A as a template to construct
double variants. The mutation results showed that the
catalytic efficiency of W82A/M78F and W82A/I284F increased
by 93.07 and 39.67-times (Table 1) in terms of kcat/KM
may be advantageous for improving affinity and maintaining
the flexible property of the loop for efficient diffusion of the
bulky substrate. According to the hydropathy index of 20 amino
acids, arginine, lysine, and histidine are basic amino acids,
while glutamic acid and aspartic acid are acidic amino acids,
all of which should be excluded for its strongest hydropathy
index of the side chains. Then, a smart library (MILVACNQ)
was constructed based on W82A and the mutation results
showed that only one mutant W82A/T440Q exhibited positive
catalytic activity with a 23.74-fold improvement (Table 1) in
3
2
M
compared to that of WT, respectively. The K values of W82A/
M78F and W82A/I284F decreased by 19.10-times and 15.47-
times relative to that of WT, respectively. The lower KM and
significant improvement in the kcat values of W82A/M78F and
W82A/I284F demonstrated that the introduced π–π stacking
interaction was important to enhance the affinity with the
phenyl group of the substrate and to improve the catalytic
efficiency.
The residue T440 was located in a flexible loop, close to
the entrance of the active site (Fig. 2). Considering the polar
property of T440, which may prevent the diffusion of the
hydrophobic substrate into the active site, a relatively more
flexible and (or) less sterically hindered residue than T440
M
kcat/K compared to WT. Compared with other residues
included in the smart library (MILVACN) constructed, residue
Q may occupy more space; however, it was the most flexible
residue among them. We speculated that the flexible property
of Q440 was more helpful for substrate diffusion than for
relieving steric hindrance. The results indicated that the
diffusion of the substrate in the active site played a key role in
the catalytic activity of the ω-TA enzymes.
Based on the above results, four sites were identified to be
responsible for the catalytic activity of BPTA, and the mutant
obtained demonstrated improved catalytic activity towards
the substrate ketone. Considering the potential positive
additive effects between these positive mutations, triple and
quadruple iterative combinatorial mutageneses were carried
out to further improve the enzyme activity of BPTA. The
results of the mutation are summarized in Table 1. The
results showed that all the variants presented positive
synergistic effects among the selected residues. Among them,
the best triple mutant was M78F/W82A/I284F with a 215.56-fold
increase in kcat/KM relative to that of WT, while the most
effective variant obtained was quadruple mutant M78F/W82A/
I284F/T440Q, which sharply increased enzyme activity by
470.38-times in terms of kcat/KM compared to WT, exhibiting
commendably positive additive effects. The K value of W82A/
M
M78F/I284F/T440Q significantly decreased from 92.86 to 1.67
mM, which indicated that W82A/M78F/I284F/T440Q sharply
promoted the affinity with the substrate ketone.
Molecular dynamics (MD) simulation was applied to
interpret the interaction between the variants and the
substrate (Fig. S4†). The complex of variant W82A and the
substrate ketone was investigated during a 50 ns MD
simulation. Despite the distance fluctuation between the
ketone and PMP during MD simulation, the ketone remained
Fig. 2 A perspective of T440 in the active site. The structure of the
enzyme is represented by a rainbow-colored cartoon, the substrate
binding pocket is represented by a white-colored surface, and the
cofactor PMP and the substrate are represented by orange and green
sticks, respectively. The residues in position T440 are represented by
blue sticks.
2464 | Catal. Sci. Technol., 2021, 11, 2461–2470
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
stable in the active site, and the average distance d1 was 5.7
Å compared to 6.4 Å of WT during the MD simulation (Fig.
S5A and B†). This indicated that by relieving the steric
hindrance by the smaller hydrophobic substituent, the
capacity to accommodate the phenyl group increased. In
differences of the amino acid locations in the LBP between
those in the engineering studies and BPTA in our study,
multiple amino acid sequence alignment was performed
3
6
(Table S2 and Fig. S6†). The result showed that BPTA
shared relatively small similarities with the TAs with 37.1%
(Ruegeria sp.TM1040), 33.9% (Vibrio fluvialis) and 62.2%
similarity (Chromobacterium violaceum), respectively. Three of
the four mutations identified in our study shared similar
locations with the TAs mentioned above, such as W82 in
3FCR (corresponding to Y59), W82 and I284 in 4E3Q
(corresponding to W57 and I259, respectively), and T440 in
4A6R (corresponding to C418) (Fig. S6†), demonstrating that
these locations are hotspots in engineering (S)-TAs for
accommodating bulky substrates. However, the mutated
residues obtained in our study were basically different from
those of the TAs referred above. Due to the diverse substrates
employed, it was hard to exactly predict the positive residues
to be mutated in a given situation. Except for the three
similar locations, M78 was a new hotspot identified in our
study which was not reported before. Among the three
double mutants constructed on the base of W82A (Table 1),
the mutant W82A/M78F showed the highest activity with up
to 6.7 fold improvement compared to W82A. Meanwhile,
3
3
addition, the MMPBSA-based binding free energy (ΔΔGW82A
)
−
1
was calculated to be −3.42 kcal mol compared to 1.24 kcal
−
1
mol for WT, demonstrating an increased affinity for the
substrate ketone. These results were in line with the kinetic
analysis. The complex conformation of variant M78F/W82A/
I284F/T440Q during MD simulation showed that the
substrate ketone was stable in the active site and the average
distance between the PMP's exocyclic nitrogen and the
ketone's carbonyl carbon atom was 4.86 Å, which was closer
to the PMP's exocyclic nitrogen. Significantly, the phenyl and
alkyl groups of the substrate were beneficially accommodated
in the large pocket and small pocket (Fig. S5C†) in the right
form of the classical Michaelis complex as described by
2
7
Cassimjee et al., which was helpful for a nucleophilic attack
between PMP's exocyclic nitrogen and the ketone's carbonyl
carbon atom in the transaminase. As shown in Fig. 1C, the
side chains of M78F and I284F pointed towards the phenyl
group of the substrate and formed a tractive π–π stacking
interaction with the phenyl group, which further
strengthened the interaction with the phenyl group to pull it
appropriately back to the LBP. In addition, the
W82A/I284F and W82A/T440Q gave
improvement, respectively. This newly identified residue M78
may be potential hotspot in engineering TAs for
a 2.9 and 1.7-fold
a
−
1
ΔΔGW82A/M78F/I284F/T440Q value was −36.33 kcal mol , which
indicates that W82A/M78F/I284F/T440Q indeed improved the
affinity with the substrate and finally remodeled the
conformation into a productive conformation with the ethyl
group normally accommodated in the SBP and the phenyl
group positioned in the LBP.
accommodating aryl alkyl substrates.
Activity improvements towards bulky substrate analysis
To explore the synthetic and deracemization utility of
BPTAM78F/W82A/I284F/T440Q for bulky substrates, the catalytic
activities toward aryl alkyl amines (compounds 1, using ‘1’ as
a prefix, such as ‘1a, 1b, 1c…) (Scheme 2) and corresponding
ketones (compounds 2, using ‘2’ as a prefix, such as ‘2a, 2b,
2c…) were investigated using pyruvic acid as the acceptor
and L-Ala as the amine donor, respectively.
The kinetic resolution results showed that WT exhibited
modest catalytic activity toward the majority of the bulky
amines, with which the aryl moiety was generally
By combining molecular docking and MD simulation
analysis, the key mutations (M78F, W82A, I284F and T440Q)
of BPTA were identified to be responsible for accommodating
aryl alkyl substrates, and they were all located in the LBP of
the active site. At present, many significant advances have
been achieved by engineering (S)-selective TAs for bulky
1
2,13,15,16,20,34,35
15
substrates.
For example, Pavlidis et al. have
successfully obtained an (S)-selective TA mutant Y59W/Y87F/
Y152F/P423H from Ruegeria sp. TM1040, which showed an
8
900-fold improvement towards the bulky substrate (S)-1,3-
1
2
diphenylpropan-1-amine. Nobili et al.
identified two
mutants (F85L/V153A and Y150F/V153A) from Vibrio fluvialis
showing a 30-fold increased activity in the conversion of (S)-
phenylbutylamine and (R)-phenylglycinol, respectively. Voss
1
3
et al. obtained two double mutants (F88L/C418(G/L)) from
Chromobacterium violaceum which showed >200-fold
improved activities in the conversion of 1-phenylbutylamine.
Most of the engineering studies on TAs were focused on
modifying the LBP and SBP. The modification studies
mentioned above modified either both the LBP and SBP or
only the SBP to achieve the enhancement of the catalytic
efficiency towards bulky substrates. In contrast, we reshaped
the LBP to improve the conversion efficiency towards the SBP
accommodating aryl alkyl substrates. In order to explore the
Scheme 2 Chiral amine donors used in this study.
Catal. Sci. Technol., 2021, 11, 2461–2470 | 2465
This journal is © The Royal Society of Chemistry 2021

View Article Online
Paper
Catalysis Science & Technology
Table
2 Activity improvements of BPTAM78F/W82A/I284F/T440Q
group as the small substituent, which indicated that the
volume of the LBP of BPTA was large enough to synthetically
accommodate the substrate ketones with an aryl group.
However, when it comes to bulky substrates with two bulky
substituents, such as aryl alkyl ketones (2b–d), indanone (2g),
and tetralone (2h), which were aryl substrates with a small
substituent bulkier than the methyl group, WT showed
decreasing activities ranging from 0.01 to 0.15-fold. It was
demonstrated that it is difficult for WT to accommodate aryl
alkyl ketones with longer alkyl side chains in the SBP.
Fortunately, the variant BPTAM78F/W82A/I284F/T440Q showed
significant activity improvements towards the majority of
bulky ketones compared to WT (Table 3). Among these aryl
alkyl ketones, the variant W82A/M78F/I284F/T440Q showed a
sharp improvement in activity for the bulky substrate ketones
with two bulky substituents (i.e., 104.0, 230.0, and 224.8-fold
reaction increases for 2b, c, and d, respectively). This
indicated that the mutant W82A/M78F/I284F/T440Q
improved the catalytic activity towards aryl alkyl substrates,
especially the substrate ketones with longer alkyl side chains
that are accommodated in the SBP.
towards various amines in kinetic resolution mode
a
Relative reactivity [%]
Amines
BPTA
BPTAM78F/W82A/I284F/T440Q
Fold-increase
b
1
1
1
1
1
1
1
1
1
1
a
b
c
d
e
f
g
h
i
100.00
118.42 ± 4.44
92.98 ± 2.11
19.10 ± 1.12
16.52 ± 3.21
95.03 ± 1.66
116.69 ± 5.55
112.28 ± 3.38
104.05 ± 8.74
127.51 ± 1.11
114.05 ± 6.18
1.18
10.46
8.97
7.34
1.24
1.02
1.33
1.40
1.30
0.97
8.89 ± 0.34
2.13 ± 0.18
2.25 ± 0.06
76.53 ± 1.56
114.48 ± 2.35
84.38 ± 1.36
74.12 ± 3.54
98.42 ± 2.37
117.13 ± 5.56
j
a
The relative activity was determined from the initial reaction rate
(i.e., conversion <10%) compared with that of the WT enzyme. The
reaction conditions were as follows: 15% DMSO, 50 mM amine
donor, 50 mM pyruvic acid, and 0.2 μM purified enzyme in 100 mM
Tris-HCl buffer (pH = 8.0) containing 0.1 mM PLP at 30 °C for 10
b
min; analyzed using HPLC. The initial reaction rate was 386.9 μM
−1
min
.
accommodated in the LBP (Table 2). However, when the
length of the alkyl chain of the substrates was longer than
that of the methyl group (1b, c, and d), WT showed poor
catalytic performance with a minimum relative activity as low
as 2.13% compared to 1a. In contrast, the variant M78F/
W82A/I284F/T440Q showed a substantial improvement in
activity toward most of the substrates except 1j (0.97-fold).
Intriguingly, the mutant presented sharp catalytic
improvements towards bulky amines with longer alkyl
chains: 1b (10.46-fold), 1c (8.97-fold), and 1d (7.34-fold),
which indicated that engineering the LBP of the enzyme
improved the catalytic activity toward bulky amines,
especially aryl alkyl amines with long alkyl chains.
Cascade strategies for chiral amine asymmetric synthesis
As a commonly used amino donor for most transaminases,
alanine has been widely used in the asymmetric synthesis of
37
chiral amines. However, due to the low equilibrium constant
−3 38
of alanine (Keq < 10 ), the equilibration problem of the
transaminases is the main challenge for industrial process
requirements. In addition, the co-product pyruvate generated in
the amination process inhibits amine production. Multiple
cascade strategies have already been established to efficiently
39,40
replace the transaminase equilibrium shift.
To evaluate the
practical utility of variant BPTAM78F/W82A/I284F/T440Q towards the
asymmetric synthesis of chiral amine 1b with L-alanine as an
amine donor, we investigated different cascade reaction systems
to displace the unfavorable equilibration in the asymmetric
synthesis of (S)-1-amino-1-phenylpropane. As shown in Fig. 3D,
the synthetic efficiency of chiral amine 1b in all the cascade
The specific activities toward aryl alkyl ketones were
investigated and showed a similar trend compared to the
catalytic activity improvements in resolution mode (Table 2).
WT showed relatively higher activity for 2e, 2f, 2i, and 2j;
these substrates were all aryl alkyl ketones with a methyl
Table 3 Activity improvements of BPTAM78F/W82A/I284F/T440Q towards various acceptors
a
Relative reactivity [%]
Ketones
BPTA
BPTAM78F/W82A/I284F/T440Q
Fold-increase
b
2
2
2
2
2
2
2
2
2
2
a
b
c
d
e
f
g
h
i
100.00
408.25 ± 12.28
602.14 ± 24.65
311.70 ± 13.83
211.64 ± 14.55
616.25 ± 34.41
360.14 ± 9.32
65.48 ± 3.43
41.77 ± 1.83
520.05 ± 26.10
677.30 ± 35.52
4.08
104.00
230.00
224.75
2.60
1.03
4.36
3.83
1.42
5.79 ± 0.35
1.36 ± 0.63
0.94 ± 0.12
237.33 ± 11.21
350.84 ± 15.66
15.01 ± 2.31
10.91 ± 0.57
366.08 ± 13.21
689.14 ±19.87
j
0.98
a
The relative activity was determined from the initial reaction rate (i.e., conversion <10%) relative to that of the WT enzyme. The reaction
conditions were as follows: 15% DMSO, 50 mM acceptor, 250 mM L-Ala, and 2 μM purified enzyme in 100 mM Tris-HCl buffer (pH = 8.0)
b
−1
containing 0.1 mM PLP at 30 °C for 6 h; analyzed using HPLC. The initial reaction rate was 0.9736 μM min
.
2466 | Catal. Sci. Technol., 2021, 11, 2461–2470
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Fig. 3 Comparison of the asymmetric synthesis of (S)-1-amino-1-phenylpropane using the BPTAM78F/W82A/I284F/T440Q mutation and WT. (A)–
C) present the asymmetric synthesis of a chiral amine by ω-TA BPTA combined with the PDC, LDH and LAADH cascade reaction system,
respectively. (D) Comparison of the conversion rates under various cascade reaction conditions by WT and the mutation M78F/W82A/I284F/T440Q.
(
systems with WT was inferior and showed no further obvious
increase during the whole reaction after 4 h, and the highest
conversion was only 1.3% at 48 h in the PDC cascade system.
The conversion and reaction rates of the prochiral ketone
catalyzed by the final variant BPTAM78F/W82A/I284F/T440Q
significantly improved. The conversion rates of the LDH and
LAADH cascade reaction systems were 88.1% and 63.4%,
respectively. The variant BPTAM78F/W82A/I284F/T440Q combined
with the PDC reaction system showed a comparatively higher
conversion of 94.4% at 48 h. This result showed that the
results of our mutation experiment showed that W82A plays
a key role in relieving the steric barrier to accept and catalyze
the target aryl alkyl ketone, 1-propiophenone, and M78F,
I284F, and T440Q play an important role in improving the
binding affinity and generation of the classical productive
conformation in the active site. The rational engineering of
the LBP successfully accommodated aryl alkyl substrates with
bulkier small substituents, which are generally located in the
SBP, and the utility of the asymmetric synthesis of chiral
bulky amines was significantly improved relative to WT
without any stereoselectivity loss. The rational engineering
approach adopted in this study could potentially facilitate
active site engineering to create an ω-transaminase variant
that can accommodate structurally diverse substrates.
quadruple
mutant
BPTAM78F/W82A/I284F/T440Q
indeed
conspicuously improved the catalytic efficiency for the
asymmetric synthesis of a chiral amine compared with WT.
The reduction of substrate ketone 2b was conducted using
the mutant W82A/M78F/I284F/T440Q and PDC coupling
cascade system at a substrate concentration of 10 mM on a Experimental
1
00 mL scale for 48 h at 30 °C. The reaction mixture was
Chemicals
extracted and evaporated achieving 88.6% yield. The
compound was characterized by
chloroform-d) δ 7.29–7.22 (m, 4H), 7.19–7.15 (m, 1H), 3.73 (t,
J = 6.9 Hz, 1H), 1.63 (m, 2H), 1.54 (br, 2H), 0.80 (t, J = 7.4 Hz,
1
(S)-α-Methylbenzylamine, (S)-phenylpropylamine, (S)-2-methyl-1-
phenyl-1-propanamine,
methylphenyl)ethylamine, (S)-(−)-1-(4-methoxyphenyl)ethylamine,
(S)-(+)-1-aminoindane, (S)-1,2,3,4-tetrahydro-1-naphthalenamine,
H NMR ((400 MHz,
(S)-1-phenylbutan-1-amine,
(S)-1-(4-
1
3
3
H)) (Fig. S7†), and
chloroform-d) δ 146.5, 128.5 (2C), 127.0, 126.5 (2C), 57.9,
2.5, 11.1) (Fig. S8†). The enantioselective analysis was
C NMR spectroscopy ((100 MHz,
(
S)-(−)-1-methyl-3-phenylpropylamine,
amine, phenoxyacetone, 4-phenylbutanone, acetophenone,
propiophenone, isobutyrophenone, butyrophenone, 4′-
methylacetophenone, 4′-methoxyacetophenone, 1-indanone,
-tetralone, and pyridoxal 5-phosphate hydrate (PLP) were
(S)-1-phenoxypropan-2-
3
carried out by HPLC, and the result is shown in Fig. S9.† We
could observe only one peak at 37.20 min in the chiral HPLC,
demonstrating the high stereoselectivity (up to >99.9%
enantiomeric excess).
1
purchased from Sigma-Aldrich (St. Louis, USA). Other reagents
used in this study were commercially available.
Conclusions
Mutagenesis, overexpression, and purification
In this study, we rationally engineered an (S)-selective
ω-transaminase BPTA to effectively synthesize the bulky aryl
alkyl chiral amine (S)-1-amino-1-phenylpropane, which is not
accommodated by selected natural transaminases. The
An ω-transaminase BPTA (Gene ID: CP001044.1, https://www.
ncbi.nlm.nih.gov/nucleotide)
cloned
from
Burkholderia
phymatum STM815 was ligated into the pET28a (+) expression
vector by double digestion with the Ned I and Hind III restriction
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2461–2470 | 2467

View Article Online
Paper
Catalysis Science & Technology
4
3
enzymes (Thermo Fisher Scientific, USA) and transformed into
Escherichia coli BL21 (DE3) cells (Thermo Fisher Scientific, USA).
The single point mutation of BPTA was carried out using a
QuickChange Lightning site-directed mutagenesis kit (Agilent
Technologies Co.), according to the manufacturer's
instructions. The primers for the mutation (Table S1†) were
designed using the primer design program of Agilent (http://
www.agilent.com). The amplicons were generated by PCR under
the following thermal cycling conditions: 3 min at 95 °C,
followed by 35 cycles at 95 °C for 15 s, annealing at 55 °C for 20
s, extension at 72 °C for 6 min, and a final extension step at 72
swissmodel.expasy.org/). Molecular docking was carried out
using the Autodock 4.2.6 suite software with the standard
free-energy scoring function and the Lamarckian algorithm
1
6
(LGA). The center of the grid box was defined in the N atom
of the amino group in the PMP cofactor. The number of runs
of LGA for the substrate ketone (1-propiophenone) and
enzyme (BPTA) was 200. The population was set to 150, the
maximum number of energy evaluations was 2 500 000, and
the maximum number of generations was 27 000. The
molecular dynamics simulation was carried out using Amber
16 with GAFF and the ff99SB force field, and the docking
complex was placed in a solvated box using the TIP3PBOX
model with a spacing distance of 10 Å. Counter ions were
added to neutralize the charge in the solvated complex
system. Two energy minimizations were performed first,
followed by the NVT and NPT ensemble using Langevin
17
°C for 10 min. The amplicons were pooled and extracted using
a FavorPrep™ GEL/PCR purification kit (Favorgen Biotech
Corp., Taiwan). The mutagenesis was confirmed by DNA
sequencing by Sangon Biotech.
−
1
We used lysogeny broth (LB) containing 50 μg mL
4
4
kanamycin for the BPTA overexpression culture, incubating
overnight for 12 h at 37 °C with shaking at 200 rpm. When
the OD600 reached 0.6, isopropyl-β-D-thiogalactoside (IPTG)
was added to the culture at a final concentration of 0.1 mM,
followed by further incubation for 20 h at 20 °C with shaking
at 200 rpm. The culture broth was centrifuged at 10 000g for
dynamics.
Kinetic analysis
The kinetic parameters of WT and the variants towards (S)-1-
amino-1-phenylpropane were obtained according to a pseudo
kinetic model built with a fixed concentration of the co-
2
0 min at 4 °C to pellet the cells, which were then
resuspended in Tris-HCl buffer (pH = 8.0, 20 mM, containing
.1 mM PLP). The pellets were disrupted by ultrasonication
45
substrate using the method of Lineweaver and Burk with
slight modifications. The KM and kcat values were calculated
from the slopes and y-intercepts of the double reciprocal
plots with various amine concentrations ranging between 0.1
and 2.5 mM. The purified enzymes were saturated with 0.1
mM PLP during purification.
0
at 4 °C, and then centrifuged at 13 000g for 30 min at 4 °C to
harvest crude enzymes in the supernatant.
The His-tagged BPTA enzyme purification was carried out
using an AKTA Explore 100 system (GE Healthcare, USA) using a
His-trap column (GE Healthcare) eluted with a linear gradient
of imidazole (ranging between 10 mM and 0.5 M). The purified
enzymes were concentrated and desalted using ultrafiltration
Substrate specificity
The effects of various amino donors were investigated
Scheme 2). The reaction conditions were as follows: 15%
DMSO, 50 mM (S)-amine, and 50 mM pyruvic acid in 100
mM Tris-HCl buffer (pH = 8.0) containing 0.1 mM PLP at 30
(
(Millipore Co., Billerica, USA). The resulting samples were
stored at −80 °C in 30% glycerol stock. The protein
41
concentration was determined by the Bradford assay.
°
C for 10 min; analyzed using HPLC. Similarly, to assess the
spectrum of the acceptor, the following reaction conditions
were used to determine the activities: 15% DMSO, 50 mM
ketone, and 250 mM L-Ala (L-alanine) in 100 mM Tris-HCl
buffer (pH = 8.0) containing 0.1 mM PLP at 30 °C for 6 h;
analyzed using HPLC. All assays were performed in triplicate.
Enzyme activity analysis
Unless otherwise specified, the BPTA enzyme assays towards
S)-1-amino-1-phenylpropane were all performed using an
(
4
2
acetophenone photometric assay with slight modifications.
The best determination wavelength for (S)-1-amino-1-
phenylpropane and the corresponding ketone was
ascertained by full wavelength scanning (Fig. S1†). The assay
was performed at 30 °C for 10 min in a total of 200 μL (UV-
Str 96-well plate, 655 801, Greiner Bio-one) containing 2.5
mM (S)-1-amino-1-phenylpropane (amino donor), 2.5 mM
pyruvic acid (amino acceptor), and 15% DMSO. The
production of 1-propiophenone was determined at 242 nm.
One unit of BPTA enzyme activity was defined as the amount
Asymmetric synthesis of (S)-1-amino-1-phenylpropane
To improve the reaction efficiency for the asymmetric
synthesis of (S)-1-amino-1-phenylpropane, different cascade
reaction systems were investigated. Reaction conditions 1: 10
mM 1-propiophenone, 250 mM L-Ala, 0.25 g BPTA dry cell
power, 0.125 g pyruvate decarboxylase (PDC) dry cell power,
and 15% DMSO. Reaction conditions 2: 10 mM
1-propiophenone, 250 mM L-Ala, 0.25 g BPTA dry cell power,
of enzyme that converted
phenylpropane into acetophenone per minute.
1
μmol of (S)-1-amino-1-
0
.125 g lactate dehydrogenase (LDH) dry cell power, 0.125 g
+
glucose dehydrogenase (GDH), 0.25 mM NAD , 0.25 g
glucose, and 15% DMSO. Reaction conditions 3: 10 mM
Bioinformatic analysis
1-propiophenone, 250 mM L-Ala, 0.25 g BPTA dry cell power,
The models of WT and the variants were constructed by the
searching algorithm of the Swiss modeling website (http://
0.125 g L-amino acid dehydrogenase (LAADH) dry cell power,
0.125 g glucose dehydrogenase (GDH), 0.25 mM NAD , 0.25 g
+
2468 | Catal. Sci. Technol., 2021, 11, 2461–2470
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
glucose, 25 mM ammonium formate, and 15% DMSO. Three
reactions were performed in 100 mM Tris-HCl buffer (pH =
6 M. Truppo, J. M. Janey, B. Grau, K. Morley, S. Pollack, G.
Hughes and I. Davies, Catal. Sci. Technol., 2012, 2, 1556–1559.
7 P. P. Taylor, D. P. Pantaleone, R. F. Senkpeil and I. G.
Fotheringham, Trends Biotechnol., 1998, 16, 412–418.
8 M. S. Malik, E.-S. Park and J.-S. Shin, Appl. Microbiol.
Biotechnol., 2012, 94, 1163–1171.
9 F. Dumeignil, M. Guehl, A. Gimbernat, M. Capron, N. L.
Ferreira, R. Froidevaux, J.-S. Girardon, R. Wojcieszak, P. Dhulster
and D. Delcroix, Catal. Sci. Technol., 2018, 8, 5708–5734.
10 Á. Mourelle-Insua, D. Méndez-Sánchez, J. L. Galman, I.
Slabu, N. J. Turner, V. Gotor-Fernández and I. Lavandera,
Catal. Sci. Technol., 2019, 9, 4083–4090.
11 F. Guo and P. Berglund, Green Chem., 2017, 19, 333–360.
12 A. Nobili, F. Steffen-Munsberg, H. Kohls, I. Trentin, C.
Schulzke, M. Hohne and U. T. Bornscheuer, ChemCatChem,
2015, 7, 757–760.
13 M. Voss, D. Das, M. Genz, A. Kumar, N. Kulkarni, J. Kustosz,
P. Kumar, U. T. Bornscheuer and M. Höhne, ACS Catal.,
8
.0) containing 1 mM PLP in a total volume of 10 mL at 30
°C for 48 h at 250 rpm. The 100 mL preparation scale
conditions were as follows: 10 mM 1-propiophenone, 250
mM L-Ala, 2.5 g BPTA dry cell power, 1.25 g pyruvate
decarboxylase (PDC) dry cell power, and 15% DMSO, and the
reaction time was 48 h at 30 °C. After the reaction was
completed, the pH of the mixture was adjusted to 3 to
dissolve (S)-1-amino-1-phenylpropane as much as possible,
and then the mixture was extracted with ethyl acetate three
times. After the extraction, the pH of the mixture was
adjusted to 13 to dissolve the ketones, and amine (S)-1-
amino-1-phenylpropane was extracted three times with ethyl
acetate. Finally, the extractions were collected and evaporated
under vacuum at 40 °C to obtain the purified amine. The
compound was analyzed using enantioselective analysis, as
1
13
well as H NMR and C NMR spectroscopy.
2
018, 8, 11524–11533.
1
4 C. K. Savile, J. M. Janey, E. C. Mundorff, J. C. Moore, S. Tam,
W. R. Jarvis, J. C. Colbeck, A. Krebber, F. J. Fleitz, J. Brands,
P. N. Devine, G. W. Huisman and G. J. Hughes, Science,
HPLC analysis
HPLC analyses were carried out on an Agilent 1290 U-HPLC
system equipped with a ZORBAX SB-AQ C18 column (4.6 ×
2
010, 329, 305–309.
2
3
0
50 mm × 5 μm) (Agilent Technologies Co., USA), heated to
7 °C using various ratios of methanol and water containing
.1% trifluoroacetic acid (TFA) at 1 mL min . The UV
1
1
1
5 I. V. Pavlidis, M. S. Weiss, M. Genz, P. Spurr, S. P. Hanlon, B.
Wirz, H. Iding and U. T. Bornscheuer, Nat. Chem., 2016, 8,
−
1
1
076–1082.
detection was performed at 210 nm. The details of the
elution conditions are described in Table S2.† The
quantitative chiral analyses of the amine products were
performed using a CROWNPAK CR (+) column (4.6 × 150 mm
6 D. F. Dourado, S. Pohle, A. T. Carvalho, D. S. Dheeman, J. M.
Caswell, T. Skvortsov, I. Miskelly, R. T. Brown, D. J. Quinn
and C. C. Allen, ACS Catal., 2016, 6, 7749–7759.
7 K. S. Midelfort, R. Kumar, S. Han, M. J. Karmilowicz, K.
McConnell, D. K. Gehlhaar, A. Mistry, J. S. Chang, M.
Anderson, A. Villalobos, J. Minshull, S. Govindarajan and
J. W. Wong, Protein Eng., Des. Sel., 2013, 26, 25–33.
×
5 μm) (Daicel Co., Japan), heated to 25 °C with 90% HClO
4
−
1
(
pH = 1.5) and 10% methanol at a flow rate of 1 mL min .
The peak signal was detected by UV spectrophotometry at
20 nm.
2
1
1
2
8 S. W. Han, E. S. Park, J. Y. Dong and J. S. Shin, Adv. Synth.
Catal., 2015, 357, 2712–2720.
Conflicts of interest
9 E. S. Park, S. R. Park, S. W. Han, J. Y. Dong and J. S. Shin,
Adv. Synth. Catal., 2014, 356, 212–220.
0 K. S. Midelfort, R. Kumar, S. Han, M. J. Karmilowicz, K.
McConnell, D. K. Gehlhaar, A. Mistry, J. S. Chang, M.
Anderson and A. Villalobos, Protein Eng., Des. Sel., 2013, 26,
There are no conflicts to declare.
Acknowledgements
2
5–33.
This work was supported by the National Natural Science
Foundation of China (No. 21776084 and 22078097) and the
Fundamental Research Funds for the Central Universities.
2
2
1 A. Nobili, F. Steffen-Munsberg, H. Kohls, I. Trentin, C.
Schulzke, M. Höhne and U. T. Bornscheuer, ChemCatChem,
2
015, 7, 757–760.
2 M. Voss, D. Das, M. Genz, A. Kumar, N. Kulkarni, J. Kustosz,
P. Kumar, U. T. Bornscheuer and M. Höhne, ACS Catal.,
2018, 8, 11524–11533.
23 M. S. Weiß, I. V. Pavlidis, P. Spurr, S. P. Hanlon, B. Wirz, H.
Iding and U. T. Bornscheuer, Org. Biomol. Chem., 2016, 14,
10249–10254.
24 M. S. Weiß, I. V. Pavlidis, P. Spurr, S. P. Hanlon, B. Wirz, H. Iding
and U. T. Bornscheuer, ChemBioChem, 2017, 18, 1022–1026.
25 S. W. Han, E. S. Park, J. Y. Dong and J. S. Shin, Adv. Synth.
Catal., 2015, 357, 1732–1740.
26 F. Ruggieri, J. C. Campillo-Brocal, S. Chen, M. S. Humble, B.
Walse, D. T. Logan and P. Berglund, Sci. Rep., 2019, 9, 16946.
Notes and references
1
2
3
4
5
Z. S. Seddigi, M. S. Malik, S. A. Ahmed, A. O. Babalghith and
A. Kamal, Coord. Chem. Rev., 2017, 348, 54–70.
A. Yohannes, X. Feng and S. Yao, J. Chromatogr. A,
2
020, 1627, 461395.
S. A. Kelly, S. Pohle, S. Wharry, S. Mix, C. C. R. Allen, T. S.
Moody and B. F. Gilmore, Chem. Rev., 2018, 118, 349–367.
S. Wenda, S. Illner, A. Mell and U. Kragl, Green Chem.,
2
011, 13, 3007–3047.
J. F. Rocha, A. F. Pina, S. F. Sousa and N. M. Cerqueira,
Catal. Sci. Technol., 2019, 9, 4864–4876.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol., 2021, 11, 2461–2470 | 2469

View Article Online
Paper
Catalysis Science & Technology
2
2
2
3
3
7 K. E. Cassimjee, B. Manta and F. Himo, Org. Biomol. Chem.,
015, 13, 8453–8464.
8 J.-S. Shin and B.-G. Kim, Biosci., Biotechnol., Biochem.,
001, 65, 1782–1788.
9 A. Dawood, M. Weiß, C. Schulz, I. Pavlidis, H. Iding, R. de Souza
and U. Bornscheuer, ChemCatChem, 2018, 10, 3943–3949.
36 X. Robert and P. Gouet, Nucleic Acids Res., 2014, 42,
W320–W324.
37 I. Slabu, J. L. Galman, R. C. Lloyd and N. J. Turner, ACS
Catal., 2017, 7, 8263–8284.
38 T. Börner, G. Rehn, C. Grey and P. Adlercreutz, Org. Process
Res. Dev., 2015, 19, 793–799.
39 D. Koszelewski, K. Tauber, K. Faber and W. Kroutil, Trends
Biotechnol., 2010, 28, 324–332.
2
2
0 B. Manta, K. E. Cassimjee and F. Himo, ACS Omega, 2017, 2,
8
90–898.
1 K. Hirotsu, M. Goto, A. Okamoto and I. Miyahara, Chem.
Rec., 2005, 5, 160–172.
40 M. Hoehne, S. Kuehl, K. Robins and U. T. Bornscheuer,
ChemBioChem, 2008, 9, 363–365.
3
3
2 J. Kyte and R. F. Doolittle, J. Mol. Biol., 1982, 157, 105–132.
3 Y. Westermaier, S. Ruiz-Carmona, I. Theret, F. Perron-Sierra,
G. Poissonnet, C. Dacquet, J. A. Boutin, P. Ducrot and X.
Barril, J. Comput.-Aided Mol. Des., 2017, 7, 757–760.
41 M. M. Bradford, Anal. Biochem., 1976, 72, 248–254.
42 S. Schätzle, M. Höhne, E. Redestad, K. Robins and U. T.
Bornscheuer, Anal. Chem., 2009, 81, 8244–8248.
43 T. Schwede, J. Kopp, N. Guex and M. C. Peitsch, Nucleic Acids
Res., 2003, 31, 3381–3385.
44 S.-W. Han, J. Kim, H.-S. Cho and J.-S. Shin, ACS Catal.,
2017, 7, 3752–3762.
45 H. Lineweaver and D. Burk, J. Am. Chem. Soc., 1934, 56,
658–666.
3
4 B.-K. Cho, H.-Y. Park, J.-H. Seo, J. Kim, T.-J. Kang, B.-S. Lee
and B.-G. Kim, Biotechnol. Bioeng., 2008, 99, 275–284.
5 M. Genz, C. Vickers, T. Van Den Bergh, H.-J. Joosten, M.
Dörr, M. Höhne and U. Bornscheuer, Int. J. Mol. Sci.,
3
2
015, 16, 26953–26963.
2470 | Catal. Sci. Technol., 2021, 11, 2461–2470
This journal is © The Royal Society of Chemistry 2021