Protein Engineering of d-Succinylase from Cupriavidus sp. for d-Amino Acid Synthesis and the Structural Implications

Article abstract of DOI:10.1002/adsc.202100587

d-Amino acids are important chiral building blocks for pharmaceuticals and agrochemicals. Previously, we have used d-Succinylase (DSA) from Cupriavidus sp. P4-10-C and N-succinyl amino acid racemase (NSAR, EC.4.2.1.113) from Geobacillus stearothermophilus NCA1503 to produce d-amino acids via the dynamic kinetic resolution of N-succinyl-dl-amino acids. However, the use of this bioconversion system remains challenging for industrial application due to the insufficient enantioselectivity of DSA toward N-succinyl-d-amino acids. Therefore, we screened DSA mutants for improved enantioselectivity by directed evolution. Several mutants showed improved enantioseletivity compared to wild-type DSA. L182E mutant had superior enantioselectivity, and the thermal stability was also remarkably improved by this single mutation. We solved the crystal structure of the L182E mutant in complex with succinic acids at a resolution 2.0 ?. The mutated residues in all generated mutants that showed improved enantioselectivity (including the substituted Glu182 in the L182E mutant) are found very close to the active site. The solved crystal structure also provides some rationale to explain the higher thermostability of the L182E mutant compared to wild-type DSA. d-phenylalanine and d-tryptophan were produced in high conversion (approximately 90%) with 98.8% ee and 99.6% ee, respectively, using coupled L182E DSA and NSAR with the one-pot enzymatic method. These data suggested that L182E DSA may be a useful biocatalyst for industrial d-amino acids production. (Figure presented.).

Full text of DOI:10.1002/adsc.202100587

RESEARCH ARTICLE
doi.org/10.1002/adsc.202100587
Protein Engineering of d-Succinylase from Cupriavidus sp. for
d-Amino Acid Synthesis and the Structural Implications
Yosuke Sumida,+a, b Masayuki Yamasaki,+c, d Yoshiaki Nishiya,b, e
Shinya Kumagai,f Toshihide Yamada,f and Masayuki Azumaa,*
a
Department of Applied Chemistry and Bioengineering
Graduate School of Engineering, Osaka City University
3-3-138 Sugimoto, Sumiyoshi-ku, Osaka 558-8585, Japan
fax: (+81)-6-6605-3092
phone: (+81)-6-6605-3092
E-mail: azuma@bioa.eng.osaka-cu.ac.jp
Biotechnology Research Laboratory, Toyobo Co., Ltd.
b
10-24 Toyo-cho, Tsuruga, Fukui 914-0047, Japan
Faculty of Agriculture, Ryukoku University
c
1-5 Yokotani, Ohe-cho, Seta, Otsu, 520-2194, Japan
Institute for Frontier Medical Sciences, Kyoto University
d
53 Kawahara-cho, Shogoin, Sakyo-ku, Kyoto 606-8507, Japan
Department of Life Science, Setsunan University
e
17-8 Ikeda-Nakamachi, Neyagawa, Osaka 572-8508, Japan
R&D Group, Iwate Research & Development Center
f
SEKISUI MEDICAL Co., Ltd.
4-115, Matsuo, Hachimantai, Iwate 028-7305, Japan
+
These authors contributed equally this work.
Manuscript received: May 15, 2021; Revised manuscript received: August 9, 2021;
Version of record online: ■■■, ■■■■
Supporting information for this article is available on the WWW under https://doi.org/10.1002/adsc.202100587
Abstract: d-Amino acids are important chiral building blocks for pharmaceuticals and agrochemicals.
Previously, we have used d-Succinylase (DSA) from Cupriavidus sp. P4-10-C and N-succinyl amino acid
racemase (NSAR, EC.4.2.1.113) from Geobacillus stearothermophilus NCA1503 to produce d-amino acids
via the dynamic kinetic resolution of N-succinyl-dl-amino acids. However, the use of this bioconversion
system remains challenging for industrial application due to the insufficient enantioselectivity of DSA toward
N-succinyl-d-amino acids. Therefore, we screened DSA mutants for improved enantioselectivity by directed
evolution. Several mutants showed improved enantioseletivity compared to wild-type DSA. L182E mutant had
superior enantioselectivity, and the thermal stability was also remarkably improved by this single mutation. We
solved the crystal structure of the L182E mutant in complex with succinic acids at a resolution 2.0 Å. The
mutated residues in all generated mutants that showed improved enantioselectivity (including the substituted
Glu182 in the L182E mutant) are found very close to the active site. The solved crystal structure also provides
some rationale to explain the higher thermostability of the L182E mutant compared to wild-type DSA. d-
phenylalanine and d-tryptophan were produced in high conversion (approximately 90%) with 98.8% ee and
99.6% ee, respectively, using coupled L182E DSA and NSAR with the one-pot enzymatic method. These data
suggested that L182E DSA may be a useful biocatalyst for industrial d-amino acids production.
Keywords: amino acid; biocatalysis; crystal structure; dynamic kinetic resolution; enantioselectivity; enzyme
catalysis; protein engineering
Adv. Synth. Catal. 2021, 363, 1–10
1
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
Introduction
neering lipases.[9,10] Herein, we performed protein
engineering of DSA from Cupriavidus sp. P4-10-C.
Chiral amino acids such as d-amino acids are We were able to demonstrate dynamic kinetic resolu-
important compounds for agrochemical and pharma- tion using mutant DSA as shown in Scheme 1. As well
ceutical production. For example, d-phenylalanine and as enantioselectivity, the thermal stability of a bio-
d-tryptophan are used in the synthesis of nateglinide, catalyst is an important factor in its potential applica-
an antidiabetic drug, and tadalafil, a drug for erectile tion in industry. Therefore, we used a thermal stability
dysfunction, respectively.[1–4] Moreover, Adaligil et al. test as a final evaluation step for DSA mutants.
recently discovered a peptide antibiotic composed of Furthermore, we discuss the improvement of the
d-amino acids.[5] Even now, d-amino acids attract enantioselectivity of the generated promising mutant
significant attention as promising raw materials. As a enzyme in relation to the three-dimensional (3D)
result, new bioconversion systems for d-amino acid structure of the enzyme.
production have been developed.[6]
Previously, we reported that d-succinylase (DSA)
from Cupriavidus sp. P4-10-C has high hydrolytic Results and Discussion
activity toward various N-succinyl-d-amino acids.[7]
Planning of Mutagenesis Strategy
Dynamic kinetic resolution of N-succinyl-dl-amino
acids could be achieved by coupling DSA and N- As previously reported, DSA from Cupriavidus sp. P4-
succinyl amino acid racemase (NSAR, EC.4.2.1.113). 10-C is a heterodimer composed of α- and β- subunits
We showed that a one-pot enzymatic method using (Figure 1) and it shares 22%–46% amino acid se-
those enzymes converted 100 mM N-succinyl-dl- quence similarity with Penicillin G acylase and
phenylalanine to d-phenylalanine in 91.1% conversion Cephalosporin C acylases.[8] These acylases are also
with 86.7% ee.[7] However, DSA also has weak heterodimers composed of small and large subunits. In
hydrolytic activity toward N-succinyl-l-amino acids. addition, two of the four deduced catalytic residues of
Therefore, enantiomeric excess of the product by this these acylases were strictly conserved.[8,11,12] However,
conversion system makes it unsuitable for commercial it was difficult to determine the target amino acid of
application. Furthermore, we have investigated bio- the mutation from such limited information. Although
conversion conditions such as the effect of DSA many crystal structures of these acylases had been
concentration on enantiomeric excess in d-phenyl- determined,[13–16] there was almost no structural in-
alanine and d-tryptophan production. However, these formation regarding enantioselectivity. Therefore, we
studies only resulted in slight improvement.[8] In constructed a screening system using semirandom
addition, there was no report that other microorganisms mutagenesis by error-prone PCR as shown in
produce DSA with higher enantioselectivity.
Scheme 2.
At first, we selected a target region to be mutated
Lipase, a widely used resource in the synthesis of
various chiral compounds, is an example of improved based on the available literature.[11,12,17–20] Doggleby
enantioselectivity following bioconversion. However, et al. and Mcdonough et al. had previously showed
the enantioselectivity of lipase is not always perfect, four catalytic residues in the active site of β-lactam
and the level of enantiomeric excess of products is acylases and the catalytic mechanisms of these
often not acceptable for application in the pharmaceut- enzymes.[11,12] Moreover, Deaguero et al. and Morillas
ical industry. In such cases, protein engineering of et al. had also described the function of the mutant,
lipase is often performed. In fact, there have been Penicillin acylase F24β from Esherichia coli and
many reports of improved enantioselectivity in engi- Penicillin acylase F71β from E. coli with improved
Scheme 1. Dynamic kinetic resolution of N-succinyl-dl-amino acids using a combination of mutant DSA and NSAR.
Adv. Synth. Catal. 2021, 363, 1–10
2
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
Figure 1. Primary structure of DSA and a target region for mutagenesis
N-succinyl-l-tryptophan. After the reaction, ninhydrin
solution was added to each reaction mixture. The
substrates do not react with ninhydrin because the
amino group is succinylated, instead, the reaction starts
when the succinyl group is hydrolyzed by DSA,
resulting in a purple reaction solution. Thus, the
amount of product was evaluated semiqualitatively as
a first step (Scheme 2) and mutants expected to
improve enantioselectivity were selected. Conse-
quently, 84 strains out of approximately 5.000 strains
were selected according to the degree of coloring in
the reaction solution.
Next, the 84 strains were screened using thin layer
chromatography (TLC) as a second screening step. The
substrate concentration was increased up to 100 mM
because l-tryptophan spots were not detected when
several strains were evaluated with the same reaction
conditions used in the first screening step. Subse-
quently, 24 strains were selected by visual evaluation.
In addition, we analyzed the DNA sequences of
recombinant plasmids from these 24 strains and found
that Gln72, Arg176, Leu182, Pro184, Asn185, Val286,
Arg305, Leu348, Phe351, Asn461, His518, and
Gly539 were all mutational candidates involved with
Scheme 2. Overview of the protein engineering process in this
study.
diastereoselectivity
and
substrate
specificity, enantioselectivity. We were unable to identify specific
respectively.[17,18] In addition, Tishkov et al. and Done mutations contributing to improvements in enantiose-
et al. reported that M142α, F146α, P22β, Q23β, F24β, lectivity, because some strains contained multiple
V56β, F57β, T68β, F71β, W154β, I177β, L253β, and mutations. For the identification, it is necessary to
F256β are located in regions that interact with the obtain an enzyme having a single mutation. Therefore,
substrate.[19,20] With this in mind, we selected target we used site-directed mutagenesis to introduce muta-
DNA (1,600 bp long) containing both α- and β- tions at each of the aforementioned amino acid
subunits for semirandom mutagenesis (Figure 1 and residues. Although not all designed mutants were
Supporting Information, Figure S1).
obtained, 56 different mutant enzymes were obtained
in this study. The list of mutations is shown in Table 1.
First and Second Screening Steps by Ninhydrin
Reaction
Third Screening Step using HPLC
Expression vectors for random mutagenesis were Next, we performed a quantitative evaluation using
generated by an error-prone PCR (Scheme 2 and HPLC as a third screening step. The protein concen-
Supporting Information, Figure S2). Next, approxi- trations of all prepared cell-free extracts were meas-
mately 5,000 strains carrying a randomly mutated gene ured and N-succinyl-d-tryptophan and N-succinyl-l-
introduced by error-prone PCR were separately cul- tryptophan reactions were performed using a fixed
tured in 96-well deep plates as a first screening step. amount of protein. Reactions with N-succinyl-l-
The same volume of culture broth was then used for tryptophan were performed with 4× the normal protein
assays with 30 mM N-succinyl-d-tryptophan or 30 mM amount to drive the reaction toward l-tryptophan. The
Adv. Synth. Catal. 2021, 363, 1–10
3
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
Table 1. Hydrolysis rate by mutants obtained by site-directed
Table 1. continued
mutagensis.
Strain
N-succinyl-tryptophan
Strain
N-succinyl-tryptophan
l-form
Hydrolysis
Rate (%)
d-form
Hydrolysis
Rate (%)
Hydrolysis
Ratio[a]
(l-form/d-form)
l-form
d-form
Hydrolysis
Ratio[a]
(l-form/d-form)
Hydrolysis
Hydrolysis
Rate (%)
Rate (%)
G539M
G539T
G539N
1.2
7.9
5.1
16.2
42.3
30.0
0.07
0.19
0.17
WT
67.9
38.0
3.4
18.7
10.0
3.5
0.0
1.6
10.0
10.5
2.6
1.7
1.5
7.0
1.7
10.2
10.4
8.4
13.1
8.9
12.8
1.9
4.3
5.6
42.9
4.4
5.2
0.0
0.0
2.5
1.8
11.2
1.7
74.4
31.6
16.1
48.4
51.5
25.9
29.5
37.3
83.9
71.0
30.5
87.9
68.5
42.7
94.3
77.1
56.3
46.5
49.2
55.2
44.3
14.6
38.4
46.9
39.4
45.8
35.7
43.3
40.8
82.2
33.6
58.7
39.6
45.7
39.2
52.6
54.5
59.7
76.7
44.3
46.9
30.1
71.0
28.1
50.6
47.0
46.7
27.9
16.1
22.2
71.9
57.3
22.7
0.91
1.20
0.21
0.39
0.19
0.14
0.00
0.04
0.12
0.15
0.09
0.02
0.02
0.16
0.02
0.13
0.18
0.18
0.27
0.16
0.29
0.13
0.11
0.12
1.09
0.10
0.15
0.00
0.00
0.03
0.05
0.19
0.04
0.12
0.07
0.07
0.14
0.12
0.68
0.66
0.14
0.06
0.41
0.48
0.38
0.16
0.14
0.13
0.19
0.18
0.93
0.30
0.08
R176H
G181W
G181K
G181R
G181D
G181E
L182W
L182S
L182C
L182Y
L182K
L182R
L182D
L182E
L182P
T183P
T183L
T183N
L184P
N185P
N185F
N185S
N185D
V286A
R305T
R305V
R305S
R305G
R305N
L348E
L348P
L348W
L348S
L348T
L348K
L348H
L348Q
L348M
L348C
F351L
F351I
[a] Hydrolysis ratio was calculated using the l- and d-form
hydrolysis rates.
hydrolysis rate of each substrate was calculated from
the HPLC data (Table 1). As a result, we found many
single mutants with reduced activity in the l-form
maintained hydrolytic activity toward the d-form, with
the exception of R176H, V286A, and H518L.
Although the l-form hydrolytic activities of several
mutants, including G181E, R305S, and R305G were
not detected at all, we speculate that was because the
l-form hydrolytic activity was undetectable due to low
DSA content or DSA activity. Therefore, we selected
L182K, L182R, L182E, and R305N, the top four
mutants, as promising candidates based on the follow-
ing criteria; (1) d-form hydrolysis rate is greater than
60% and, (2) the hydrolysis ratio of l- to d-form is
less than 0.05.
Fourth Screening Step by Thermal Stability Evalu-
ation
To determine which of the top four mutants is most
suitable for industrial application, we compared the
thermal stabilities of their crude extracts. All four
°
crude extracts were incubated for 1 h at 68 C, and the
residual activity was compared. As a result, we found
that L182E was the most stable and that it was
surprisingly more stable than the wild-type (WT)
protein (Table 2). Therefore, we compared the thermal
stabilities of recombinant L182E and WT protein.
L182E was purified to homogeneity from recombinant
E. coli DH5α as previously described.[7] Indeed, the
5.6
2.9
3.6
7.6
7.2
52.3
29.3
6.5
1.8
Table 2. Thermal stability of crude extracts prepared from
mutant DSAs.
F351M
F351N
F351Q
N461I
N461F
N461T
N461K
N461R
H518L
G539P
G539V
29.1
13.5
19.4
7.3
6.5
3.6
3.0
3.9
67.0
17.3
1.8
Strain
Residual activity[a]
WT
41%
82%
40%
35%
22%
L182E
L182 K
L182R
R305 N
[a] The residual activity of crude extract was evaluated in
25 mM potassium phosphate buffer (pH 7.0) at 68 C for 1 h.
°
Adv. Synth. Catal. 2021, 363, 1–10
4
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
on d-amino acid production. Recombinant DSA and
NSAR were purified to homogeneity from recombinant
E. coli DH5α as previously described.[7] Next, d-
tryptophan was produced in 1-mL reaction mixture
containing DSA and NSAR as previously described
with some modification.[7] Briefly, NSAR were added
with 540 μg/mL of protein because enantiomeric
excess with this concentration of NSAR with WT DSA
was a little higher than 270 μg/mL NSAR (Table 3).
Consequently, enantiomeric excess using L182E (32
μg/mL) achieved 99.6% ee (Table 3). The enantiomeric
excess achieved with this mutantfigr was much higher
than WT DSA (95.6% ee).[8] In addition, the con-
version rate was also high (90.3%), although the
specific activity of L182E (41.2 U/mg) was slightly
lower than WT DSA (75.2 U/mg). Next, d-phenyl-
alanine production was also performed in 1-mL
reaction mixture containing DSA and NSAR as
previously described.[7] As a result, L182E enantiomer-
ic excess achieved 98.8% ee in both DSA concen-
Figure 2. Thermal stability of WT and L182E DSA. The
enzyme (1.0 mg/mL) in 25 mM potassium phosphate buffer
(pH 7.5) was incubated at different temperature for 1 h. The
remaining activity of WT (closed circles) and L182E (open
circles) are shown.
thermal stability of L182E was significantly higher trations (Table 3). The value of enantiomeric excess
°
compared to WT protein (Figure 2). Even at 65 C, was also much higher compared to previously reported
more than 40% activity remained. In addition, d-form WT DSA data (91.6% ee).[8] In addition, the L182E
hydrolysis rate of L182E (94.3%) was the highest conversion rate (32 μg/mL) also remained high
among the four mutants (Table 1). We also demon- (89.2%). In these one-pot reactions, we showed that
strated the improved enantioselectivity of L182E L182E is a practical biocatalyst that produces d-amino
toward N-succinyl-d-phenylalanine, N-succinyl-d-ser- acids with high enantiomeric excess and high con-
ine, and N-succinyl-d-glutamic acid as well as N- version rates. Furthermore, there is potential to exploit
succinyl-d-tryptophan (Supporting Information, Fig- the increased thermal stability of L182E by using
ure S3). Therefore, we performed d-amino acid pro- higher reaction temperatures. NSAR also has high
duction and X-ray crystallography studies using activity at higher reaction temperatures (Supporting
L182E, the most promising mutant DSA.
Information, Figure S4). This will also reduce con-
version time and production costs.
d-Phenylalanine and d-Tryptophan Production us-
ing WT and L182E DSA
Structural Implications of L182E DSA
A one-pot reaction using DSA and NSAR was The crystal structure of L182E DSA was determined at
conducted to evaluate the enantioselectivity of L182E 2.0 Å in complex with succinic acids (Figure 3a). In
Table 3. Bioconversion of N-succinyl-dl-amino acids by Mutant DSA (L182E) coupled with NSAR.
Substrate
DSA (WT)
(μg/mL)
DSA (L182E)
(μg/mL)
NSAR
(μg/mL)
Conversion
(%)[a]
Enantiomeric
Ref.
excess (% ee)[b]
N-succinyl-dl-
tryptophane
32
32
8
8
–
–
270
540
270
540
540
540
270
270
270
270
270
93.9
94.6
81.8
82.7
90.3
68.6
92.0
87.1
82.4
89.2
64.6
89.4
92.0
94.7
95.6
99.6
99.8
87.1
89.9
91.6
98.8
98.8
8
This study
8
This study
This study
This study
8
–
–
32
8
–
–
–
32
8
–
N-succinyl-dl-
phenylalanine
32
16
8
–
–
8
8
This study
This study
[a] The conversion was determined by HPLC analysis using ODS-3.
[b] The enantiomeric excess was determined by HPLC analysis using CROWNPAK CR (À ).
Adv. Synth. Catal. 2021, 363, 1–10
5
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
Figure 3. Crystal structure of L182E DSA (α-subunit in yellow, β-subunit in red). (a) Overall structure of L182E. (b) Succinic acids
in the surface structure. (c) The putative catalytic site and surrounding resides. Putative catalytic residues are labeled in white.
Residues R305-E182-R350 (colored in green) would contribute to the increased thermal stability of L182E. (d) Sliced view around
the active pocket structure. 7 residues located around protein surface or near the succinic acid in the catalytic pocket (5 from the α-
subunit, and 2 from the β-subunit). In contrast, 3 residues located within the DSA structure. All affected the enantioselectivity of
DSA.
this structure, DSA was found to bind two succinic ure 3a). This may act as the coordinate after DSA
acids, one very close to the active site and the other in catalyzes the substrate. One of the succinic acids binds
the neighboring pocket (Figure 3b). As an overall very close to the predicted four key catalytic amino
structure, DSA shares some similarity with typical acid residues; Ser282 in β-subunit (Ser1β), H304 in β-
acylases of Penicillin acylase from E. coli (PDB code: subunit (His23β), R350 in β-subunit (Arg69β) and
1GK9, with a root-mean-square deviation (RMSD) of N523 in β-subunit (Asn242β). In fact, this succinic
3.02 Å),[11] Cephalosporin acylase from Pseudomonas acid is surrounded by S282, R305, R313 R350, Y433,
sp. GK16 (PDB code: 1OR0 with an RMSD of E460, and N523, which may together form the
3.14 Å), or Cephalosporin acylase from Pseudomonas substrate binding pocket (Figure 3c). The other suc-
sp. N176 (PDB code: 4HST with an RMSD of cinic acid bound in the neighboring pocket indirectly
1.46 Å).[12,13] As the primary structure, DSA shares interacts with surrounding amino acids via water
some similarity (only β-subunits) with Penicillin molecules, which seems to form weak interactions
acylase from E. coli (27% similarity), Cephalosporin with DSA and could attenuate the true substrate
acylase from Pseudomonas sp. GK16 (27% similarity), binding. Again, as an overall structure, the substitution
or Cephalosporin acylase from Pseudomonas sp. N176 of hydrophobic Leu182 for negatively charged Glu182
(29% similarity). Both succinic acids seem to bind enables DSA to maintain strong ionic interactions with
very centrally and deep into the DSA structure (Fig- the α- and β-subunits via R305-E182-R350 (Figure 3c,
Adv. Synth. Catal. 2021, 363, 1–10
6
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
green). These interactions would contribute to the similar to that of lipase, which create an oxyanion hole
increased thermal stability of L182E compared to WT in transition state.[11] That means the reaction mecha-
DSA by connecting α- and β-subunits, as was observed nism of DSA is similar to that of lipase. Alternatively,
in Figure 2. In addition, ten amino acid residues, whose it is reported that enantioselectivity of lipase is thought
mutations had affected the enantioselectivity of DSA to result principally from the conformational require-
(Table 1), cover a broader area around the catalytic ment and interaction in the transition state, not from
pocket surrounding the α- and β-subunits (Figure 3d). the attractive interactions between the enzyme pockets
Five of these residues (G181, L182, T183, L184, and and its substrates.[9] Given that, like lipase, DSA has a
N185 in yellow) are in the loop region of the α-subunit broad substrate specificity and high enantioselectivity,
locate at the, two residues (R305 and L348 in red) are the mechanism of enantioselectivity in DSA may be
in a region of the β-subunit involved in succinic acid similar to lipase. However, further research is required
binding, and the final three residues (F351, N461, and to demonstrate this hypothesis.
G539 in gray) are located inside the molecule. Thus,
there are as many as ten mutations that affect the
enantioselectivity of DSA in a wide range around the
Conclusions
catalyst pocket. These improvements of enantioselec- We performed protein engineering of DSA from
tivity may be due to different factors. Furthermore, Cupriavidus sp. P4-10-C to improve its enantioselec-
there is a case where a mutation near the catalyst tivity toward N-succinyl-d-amino acids. We found that
pocket increases the thermal stability (Figure 2). the L182E mutant, obtained by random mutagenesis
Therefore, the combination of these mutations has followed by site-saturation mutagenesis, showed much
possibility of improvement of enantioselectivity and/or higher enantioselectivity compared to WT DSA. In d-
thermal stability. It seems reasonable to conclude that amino acid production, a one-pot enzymatic method
the three residues (R176, V286, and H518) that did not using L182E produced d-tryptophan and d-phenyl-
improve enantioselectivity are not close to residues alanine with 99.6% ee and 98.8% ee, respectively.
supporting a catalytic pocket function, either directly Furthermore, we found that L182E DSA has a higher
or indirectly. Collectively, they may together attenuate thermal stability compared to the WT protein. There-
the formation of catalytic pocket conformation and fore, L182E DSA could be used as a practical
affect the enantioselectivity.
industrial biocatalyst. The structural studies and dock-
ing simulations could not explain why the L182E
mutant shows improved enantioselectivity, on the other
hand, these data provided a rationale for the improved
Molecular Docking Simulations
To investigate the mechanism behind the improved thermal stability of the L182E mutant. The structural
enantioselectivity of the L182E mutant, we performed data obtained in this study will be useful for further
molecular docking simulations of N-succinyl-d-phe- investigation.
nylalanine and N-succinyl-l-phenylalanine with
L182E. However, the calculated docking scores of N-
succinyl-l-phenylalanine and N-succinyl-d-phenylala-
Experimental Section
nine on L182E were À 7.66 and À 7.72, respectively. In Media and Culture Conditions
addition, the docking scores of N-succinyl-l-phenyl-
For the screening of mutant DSA, Escherichia coli DH5α
carrying the recombinant plasmid constructed by an error-prone
PCR and site-saturated mutagenesis were cultured in LB
alanine and N-succinyl-d-phenylalanine on WT DSA
were À 7.22 and À 7.52, respectively (Supporting
Information, Figure S5). Moreover, the docking scores
°
medium containing 100 mg/L ampicillin at 30 C. For the
of N-succinyl-l-tryptophan and N-succinyl-d-trypto-
phan on L182E were À 6.93 and À 7.97, respectively.
In other words, these data suggest that interactions
between the substrate and the enzyme pocket cannot
be used to explain the high enantioselectivity of the
L182E DSA mutant. The reaction mechanism of DSA
is assumed to be similar to those of Penicillin acylase
and Cephalosporin acylase because the crystal struc-
ture analysis showed that the amino acid residues
corresponding to four catalytic residues reported in
these acylases[11,12] are located near the active center of
DSA. On the other hand, the reaction mechanism of
Penicillin acylase and Cephalosporin acylase, which
hydrolyze amide and ester bonds using a catalytic
serine residue by an acyl enzyme mechanism, is
production of d-amino acids, E. coli DH5α carrying recombi-
nant plasmids for WT DSA, L182E DSA, or NSAR were
cultured in TB medium containing 100 mg/L ampicillin at
°
30 C.
Chemicals
l-amino acids, d-amino acids and other chemicals were
purchased from Nacalai Tesque (Kyoto, Japan). N-succinyl
amino acids were synthesized as previously described.[7]
Ninhydrin reagent was purchased from Tokyo Kasei (Tokyo,
Japan).
Adv. Synth. Catal. 2021, 363, 1–10
7
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
extracts were prepared by sonication. The cell debris was
removed by centrifugation, and the cell-free extract was used
for following reaction. For the analysis of N-succinyl-d-
tryptophan, reaction mixtures comprised of 25 mM potassium
phosphate buffer (pH 7.0) and 33 mM N-succinyl-d-tryptophan
containing 0.5 mg/mL cell-free extract. For N-succinyl-l-
tryptophan analysis, reaction mixtures contained 25 mM potas-
sium phosphate buffer (pH 7.0) and 33 mM N-succinyl-l-
tryptophan with 2.0 mg/mL cell-free extract. Each reaction
Construction of a Mutant DSA Library
The template vector for the random mutagenesis of DSA was
constructed as described in the supporting information, Fig-
ure S2. Random mutagenesis libraries with two mutations per
1.000 bp were generated using the Diversify PCR Random
Mutagenesis Kit (Takara, Kyoto, Japan) according to the
manufacturer’s instructions. Error-prone PCR for random muta-
genesis was performed as described in the supporting informa-
tion, Figure S2. The condition are as follows: 1 cycle, 30 sec at
°
mixture was incubated at 37 C for 16 h. After the reaction, the
°
°
°
94 C; 30 cycles, 30 sec at 94 C, 15 sec at 58 C, 2 min at
hydrolysis rate was analyzed using HPLC as follows. The
product and substrate were assayed by HPLC on an Inertsil
ODS-3 column (GL Science, Tokyo, Japan) using acetonitrile/
water (25/75, v/v, pH 2.3 adjusted with phosphoric acid) as the
°
°
68 C; 1 cycle, 2 min at 68 C. Site-saturation mutagenesis was
performed using the KOD Plus Mutagenesis kit (Toyobo Co.,
Ltd., Osaka, Japan) according to the manufacturer’s instruc-
tions. All oligonucleotides used for mutagenesis are listed in
supporting information, Table S1.
°
mobile phase at a flow rate of 0.4 mL/min at 40 C. Detection
was at 280 nm.
First Screening on 96-Well Plates by Ninhydrin
Reaction
Fourth Screening by Thermal Stability Evaluation
The cell-free extract obtained in the third screening step was
Colonies obtained on LB agar plates were incubated in 96-well
°
incubated at 68 C for 1 h for thermal stability evaluation.
°
deep plates containing 300 μL of LB medium at 37 C for 16 h.
Following incubation, the remaining DSA activity was meas-
ured as previously described.[7]
Cells expressing mutant DSA were harvested by centrifugation,
washed with 25 mM potassium phosphate buffer (pH 7.0), and
resuspended in 600 μL of 25 mM potassium phosphate buffer
(pH 7.0). Resuspended cells (50 μL) were added to the reaction
mixture (50 μL) containing 25 mM potassium phosphate buffer
(pH 7.0) and 60 mM N-succinyl-d-tryptophan or 60 mM N-
Expression and Purification of DSA (WT and
L182E)
Expression and purification of the WT and L182E DSA were
performed as previously described.[7]
°
succinyl-l-tryptophan. The mixture was incubated at 37 C for
16 h. After the reaction, 100 μL of ninhydrin solution (Tokyo
Kasei, Tokyo, Japan) was added to the reaction mixture and
incubated again at room temperature for 15 min before visual-
ization. The amount of product (d-tryptophan or l-tryptophan)
was estimated by visual inspection.
Protein Measurement
Protein concentrations were determined using a protein assay
kit (Bio-Rad, Hercules, CA, USA) using bovine serum albumin
as the standard.
Second Screening on TLC by Ninhydrin Reaction
Colonies were incubated in test tubes containing 5 mL of LB
Assay of the DSA and NSAR Activity
°
medium at 37 C for 16 h. Cells expressing (1 mL) mutant DSA
The DSA and NSAR activity was measured as previously
described.[7]
were harvested by centrifugation, washed with 25 mM potas-
sium phosphate buffer (pH 7.0), and resuspended with 400 μL
of 25 mM potassium phosphate buffer (pH 7.0). Resuspended
cells (20 μL) were added to the reaction mixture (80 μL)
containing 25 mM potassium phosphate buffer (pH 7.0) or
125 mM N-succinyl-d-tryptophan 125 mM N-succinyl-l-trypto-
Dynamic Kinetic Resolution of N-Succinyl-dl-
Amino Acids into d-Amino Acid by DSA (WT and
L182E) and NSAR
°
phan. The mixture was incubated at 37 C for 16 h. After the
reaction, the hydrolysis products (d-tryptophan or l-tryptophan)
were analyzed by TLC. The solvent system consisted of 1-
butanol/acetic acid/water (3/1/1, v/v/v). After chromatography,
the plate was sprayed with ninhydrin reagent (Tokyo Kasei,
Tokyo, Japan) for visualization. The amount of product (d-
tryptophan or l-tryptophan) was estimated by visual evaluation.
A reaction for d-tryptophan production was carried out in 1 mL
of 50 mM Tris-HCl buffer (pH 7.5) containing 100 mM N-
succinyl-dl-tryptophan, 1 mM CoCl2, 8 or 32 μg of WT or
°
L182E DSA, and 270 or 540 μg of NSAR at 40 C for 20 h. A
reaction for d-phenylalanine production was carried out in
1 mL of 50 mM Tris-HCl buffer (pH 7.5) containing 100 mM
N-succinyl-dl-phenylalanine, 1 mM CoCl2, 8, 16 or 32 μg of
°
WT or L182E DSA, and 270 μg of NSAR at 40 C for 20 h.
Third Screening using HPLC
Reactions were stopped by boiling for 3 min. The amount of
remaining substrate was determined by HPLC on the Inertsil
ODS-3 column (GL Science, Tokyo, Japan) using acetonitrile/
water (25/75, v/v, pH 2.3 adjusted with phosphoric acid) as the
Colonies were incubated in test tubes containing 5 mL of LB
°
medium at 37 C for 16 h. Cells expressing (1 mL) mutant DSA
were harvested by centrifugation, washed with 25 mM potas-
sium phosphate buffer (pH 7.0), and resuspended with 1 mL of
25 mM potassium phosphate buffer (pH 7.0). Next, crude
°
mobile phase at a flow rate of 0.4 mL/min at 40 C, and the
yield was calculated. Detection was at 280 nm for N-succinyl-
dl-tryptophan at 254 nm for N-succinyl-dl-phenylalanine.
Adv. Synth. Catal. 2021, 363, 1–10
8
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
asc.wiley-vch.de
Optical purity of d-tryptophan was determined using
a
Acknowledgements
CROWNPAK CR(À ) column (Daicel, Tokyo, Japan) with
methanol/HClO4 (10/90, v/v, pH 2.0) as the mobile phase at a
We thank all the members of our biochemical group at
Biotechnology Research Laboratory. The synchrotron radiation
experiments were performed at beamline BL26B1 of SPring-8,
with the approval of the Japan Synchrotron Radiation Research
Institute (JASRI) (proposal numbers: 2013B1373 and
2014A1308).
°
flow rate of 0.8 mL/min at 15 C and that of d-phenylalanine
was determined using a CROWNPAK CR(À ) column (Daicel,
Tokyo, Japan) with HClO4 (pH 2.0) as the mobile phase at a
°
flow rate of 0.8 mL/min at 15 C. Detection was at 280 nm and
254 nm for dl-tryptophan and dl-phenylalanine, respectively.
Crystallization and Structure Determination
References
Crystallization of WT and L182E DSA were first performed
using the sitting-drop method to test conditions of PACT
premier-, JCSG-plus-, and Structure Screen 1&2- HT-96 Eco
Screens (Molecular Dimensions, Sheffield, UK). Conditions
were then optimized using the hanging-drop method. Finally,
WT DSA was concentrated to 7.7 mg/mL and crystalized in
24% (v/v) PEG1500 and 20% (v/v) glycerol. Crystals were
frozen in liquid-nitrogen before data collection. L182E DSA
was concentrated to 10 mg/mL and crystalized in 11% (v/v)
PEG8000, 0.1 M sodium cacodylate buffer (pH 6.6) and 0.2 M
MgCl2. Crystals were soaked into the same solution with
10 mM succinic acid for 30 min and frozen in liquid-nitrogen.
Data were collected under cryogenic conditions at SPring-8
(Hyogo, Japan), up to 2.0 Å for WT DSA and 1.95 Å for
L182E DSA in complex with succinic acids. Data were
processed by HKL2000 and imported by CCP4-implemented
SCALA. The structure of WT DSA was solved using the
molecular replacement method by CCP4-implemented Phaser
using a truncated structure of 4HST.pdb. The structure was
determined by conducting multicycles of manually building the
structure on Coot and subsequent refinement on CCP4-
implemented Refmac and Phenix.refine. The structure of L182E
DSA was solved using the molecular replacement method by
the CCP4-implmented Phaser using the monomeric structure of
WT DSA and determined as described above. Ramachandran
statistics for the final WT model show 97.4% of residues in the
most favorable regions, 2.60% in allowed regions and 0.02% as
outliers. Ramachandran statistics for the final L182E model
show 97.8% of residues in the most favorable regions, 2.24% in
allowed regions and 0.00% in outliers. The R/free-R values for
the final models of WT and L182E DSA were 0.204/0.252
(PDB: 7EBY) and 0.164/0.208 (PDB: 7EA4), respectively. All
basic statistics are shown in Table S2.
[1] N. Tentolouris, C. Voulgari, N. Katsilambros, Vasc.
Health Risk Manage. 2007, 3, 797–807.
[2] I. W. Campbell, Int. J. Clin. Pract. 2005, 59, 1218–1228.
[3] C. C. Carson, M. Rosenberg, J. Kissel, D. G. Wong, Int.
J. Clin. Pract. 2014, 68, 94–103.
[4] A. Govorov, G. Kasyan, D. Priymak, D. Pushkar, S.
Sorsaburu, Cent. European J. Urol. 2014, 67, 167–177.
[5] E. Adaligil, K. Patil, M. Rodenstein, K. Kumar, ACS
Chem. Biol. 2019, 14, 1498–1506.
[6] S. Nakano, Y. Minamino, F. Hasebe, S. Ito, ACS Catal.
2019, 9, 10152–10158.
[7] Y. Sumida, S. Iwai, Y. Nishiya, S. Kumagai, T. Yamada,
M. Azuma, Adv. Synth. Catal. 2016, 358, 2041–2046.
[8] Y. Sumida, S. Iwai, Y. Nishiya, S. Kumagai, T. Yamada,
M. Azuma, J. Biosci. Bioeng. 2018, 125, 282–286.
[9] Y. Koga, K. Kato, H. Nakano, T. Yamane, J. Mol. Biol.
2003, 331, 585–592.
[10] T. Ema, Y. Nakano, D. Yoshida, S. Kamata, T. Sakai,
Org. Biomol. Chem. 2012, 10, 6299–6308.
[11] H. J. Duggleby, S. P. Tolley, C. P. Hill, E. J. Dodson, G.
Dodson, P. C. Moody, Nature 1995, 373, 264–268.
[12] M. A. McDonough, H. E. Klei, J. A. Kelly, Protein Sci.
1999, 8, 1971–1981.
[13] C. E. McVey, M. A. Walsh, G. G. Dodson, K. S. Wilson,
J. A. Brannigan, J. Mol. Biol. 2001, 313, 139–150.
[14] J. K. Kim, I. S. Yang, S. Rhee, Z. Dauter, Y. S. Lee, S. S.
Park, K. H. Kim, Biochemistry 2003, 42, 4084–4093.
[15] E. Golden, R. Paterson, W. J. Tie, A. Anandan, G.
Flematti, G. Molla, E. Rosini, L. Pollegioni, A. Vrielink,
Biochem. J. 2013, 451, 217–226.
[16] V. Ferrario, M. Fischer, Y. Zhu, J. Pleiss, Sci. Rep. 2019,
9, 12402.
[17] A. L. Deaguero, J. K. Blum, A. S. Bommarius, Protein
Eng. Des. Sel. 2012, 25, 135–144.
Molecular Docking
[18] M. Morillas, C. E. McVey, J. A. Brannigan, A. G.
Ladurner, L. J. Forney, R. Virden, Biochem. J. 2003, 371,
143–150.
[19] S. H. Done, J. A. Brannigan, P. C. Moody, R. E. Hub-
bard, J. Mol. Biol. 1998, 284, 463–475.
Molecular docking studies were performed using the induced fit
docking protocol in the Molecular Operating Environment
(MOE) software ver. 2015.1001 (Chemical Computing Group
Inc., Montreal, Canada) to allow for side chain flexibility. All of
the compounds used for docking process were built and saved
as MOE files. Next, 3D protonation of the proteins were energy
minimized using the default parameters in MOE. The docking
site was defined around the active center. Triangle matcher was
used as a placement method. Rescoring was performed with
London dG and GBVI/WSA. For each ligand, 5 conformations
were allowed to be formed and the top ranked reactive
conformations, based on docking scores, were selected.
[20] V. I. Tishkov, S. S. Savin, A. S. Yasnaya, Acta Naturae
2010, 2, 47–61.
Adv. Synth. Catal. 2021, 363, 1–10
9
© 2021 Wiley-VCH GmbH
��
These are not the final page numbers!

RESEARCH ARTICLE
Protein Engineering of d-Succinylase from Cupriavidus sp.
for d-Amino Acid Synthesis and the Structural Implications
Adv. Synth. Catal. 2021, 363, 1–10
Y. Sumida, M. Yamasaki, Y. Nishiya, S. Kumagai, T.
Yamada, M. Azuma*