Cβ-Selective Aldol Addition of d -Threonine Aldolase by Spatial Constraint of Aldehyde Binding

Article abstract of DOI:10.1021/acscatal.1c01348

d-Threonine aldolase (DTA) is a useful biocatalyst that reversibly converts glycine and aldehyde to β-hydroxy-α-d-amino acid. However, low activity and poor diastereoselectivity limit its applications. Here we report DTA from Filomicrobium marinum (FmDTA) that shows much higher activity and Cβ-stereoselectivity in d-threonine production compared with those of other known DTAs. We determine the FmDTA structure at a 2.2 ? resolution and propose a DTA catalytic mechanism with a kernel of the Lys49 inner proton sink and metal ion in the aldol reaction cycle. The enzyme is rationally engineered to have high Cβ-stereoselectivity based on spatial constraint at the anti-specific aldehyde position in the mechanism, and the rational strategy is further applied to other DTAs for syn-production. The final FmDTAG179A/S312A variant exhibits a near-perfect 99.5% de value for d-threonine and maintains the de value above 93% even under kinetically unfavorable conditions. This study demonstrates how a detailed understanding of the reaction mechanism can be used for rational protein engineering.

Full text of DOI:10.1021/acscatal.1c01348

pubs.acs.org/acscatalysis
Research Article
Cβ-Selective Aldol Addition of D‑Threonine Aldolase by Spatial
Constraint of Aldehyde Binding
Sung-Hyun Park,# Hogyun Seo,# Jihye Seok, Haseong Kim, Kil Koang Kwon, Soo-Jin Yeom,*
Seung-Goo Lee,* and Kyung-Jin Kim*
Cite This: ACS Catal. 2021, 11, 6892−6899
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: D-Threonine aldolase (DTA) is a useful biocatalyst that
reversibly converts glycine and aldehyde to β-hydroxy-α-D-amino acid.
However, low activity and poor diastereoselectivity limit its applications.
Here we report DTA from Filomicrobium marinum (FmDTA) that shows
much higher activity and Cβ-stereoselectivity in D-threonine production
compared with those of other known DTAs. We determine the FmDTA
structure at a 2.2 Å resolution and propose a DTA catalytic mechanism with
a kernel of the Lys49 inner proton sink and metal ion in the aldol reaction
cycle. The enzyme is rationally engineered to have high Cβ-stereoselectivity
based on spatial constraint at the anti-specific aldehyde position in the
mechanism, and the rational strategy is further applied to other DTAs for
syn-production. The final FmDTAG179A/S312A variant exhibits a near-perfect
99.5% de value for D-threonine and maintains the de value above 93% even
under kinetically unfavorable conditions. This study demonstrates how a
detailed understanding of the reaction mechanism can be used for rational protein engineering.
KEYWORDS: D-threonine aldolase, stereoselectivity, β-hydroxy-α-amino acid, catalytic mechanism, protein engineering
low-specificity L -threonine aldolase (LTA, EC
INTRODUCTION
■
4.1.2.48).10,16−19 However, there have been no engineering
efforts exerted on the D-threonine aldolase (DTA, EC
4.1.2.42). Unlike L-threonine or L-allo-threonine aldolase (EC
4.1.2.5 or 4.1.2.49),20 there are no Cβ-specific DTA enzymes
known. Although the crystal structure of the DTA from
Alcaligenes xylosoxidans (AxDTA) has been reported,21 the
underlying principle of the low Cβ-stereospecificity of DTAs
has still not been elucidated. Herein, we provide a study on
rational design protein engineering to tailor the Cβ-selectivity
of DTAs, which comprises the multidisciplinary works of
enzyme identification, structural analysis, and catalytic
mechanisms.
Aldolases transform the carbon skeleton into a larger molecule
by forming a carbon and carbon bond (C−C) and play critical
roles in organic matter cycling in nature. The reaction can be
used to produce and reconstruct the building blocks of many
high-value compounds under green conditions.1,2 In particular,
the aldol addition reaction enabled by enzymes is of interest in
chemical industries due to the chiral-selectivity of the
biocatalyst.3,4
Pyridoxal-5′-phosphate (PLP)-dependent threonine aldo-
lases (TAs) catalyze the C−C bond formation reaction
between the α-carbon (Cα) of glycine and acetaldehyde.
The catalytic promiscuity of aldehyde selection has made the
enzymes an important method for the synthesis of β-hydroxy-
α-amino acids, which provide a versatile platform for various
high-value compounds, such as active pharmaceutical ingre-
dients, antibiotics, or agrochemical building blocks.5−9
However, the insufficient diastereoselectivity of TAs at the
Cβ restricts the biocatalyst-based method for β-hydroxy-α-
amino acids.2,10−15
RESULTS
■
Identification of FmDTA. To identify an enzyme with
high aldol activity and Cβ-selectivity, we performed a
phylogenetic analysis of DTAs and selected four DTA
candidates from Achromobacter insolitus (AiDTA), Bordetella
Received: March 24, 2021
Revised: May 14, 2021
Published: May 28, 2021
A number of studies have attempted to control Cβ-
specificity by changing reaction conditions to favor kinetic or
thermodynamic control/influence, but altering the regulated
conditions inevitably limited the performance and usage of
TAs.11−13 Thus, protein engineering approaches have been
carried out to overcome the Cβ-stereoselectivity problem of
https://doi.org/10.1021/acscatal.1c01348
© 2021 American Chemical Society
ACS Catal. 2021, 11, 6892−6899
6892

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 1. Identification of FmDTA. (A) The unrooted maximum likelihood phylogenetic tree of DTAs. The known DTAs and the newly selected
DTAs in this study are labeled with black and blue colors, respectively. The clades are distinguished with different background colors. The
bootstrap values in 1000 replication are shown at the branching edges and also expressed with red circles with lightness. The nodes of the alphabet
indicate its sequence accession code in the NCBI database: (a) WP_050728312, (b) WP_183007039, (c) WP_081695032, (d) WP_005017435,
and (e) WP_164855593. (B−F) All of the reactions were performed with 100 mM glycine, 100 mM acetaldehyde, and 50 mM buffer conditions
for 10 min. (B) Specific activities of DTAs. Standard deviations are shown as bars (n = 3). (C) The diastereomeric excess achieved by the DTAs.
Standard deviations are shown as bars (n = 3). (D) Temperature effect on FmDTA activity. The total product concentration is shown with the sum
of D-threonine and D-allo-threonine concentration. Standard deviations are shown as bars (n = 3). (E) Metal ion preference of FmDTA. Standard
deviations are shown as bars (n = 3). The concentration-dependent data were fitted with the modified Briggs−Haldane equation. (F) The effect of
PLP concentration on enzyme activity (circles) and de value for selectivity (triangles). Standard deviations are shown as bars (n = 3). The activity
data were fitted with the Hill equation.
hinzii (BhDTA), Oligotropha carboxidovorans (OcDTA), and
Filomicrobium marinum (FmDTA), which show a low phylo-
gram relationship (Figure 1A, Figure S1, Tables S1-2). We
then measured the specific aldol activities of the selected
enzymes and the aldol activities with those of AxDTA, which is
the most extensively studied enzyme among DTAs, using a pH
range of 5.5−10.0.11,13,22−25 Considering that acetaldehyde is
the basic form of the aldehyde substrate that exacerbates the
selectivity problem of DTA,22−25 we used acetaldehyde as a
substrate. Although all DTAs showed an optimal pH in the
alkaline pH ranging from 7.5 to 9.0, three DTAs, such as
BhDTA, AiDTA, and FmDTA, exhibited higher activities than
those of AxDTA (Figure 1B). In particular, FmDTA showed
4.7-fold higher aldol activity than that of AxDTA under
optimal pH conditions (Figure 1B). Moreover, the enzyme
exhibited a superb de (46−51%) for D-threonine, depending
on the pH conditions, compared with that of other DTAs with
up to 30% de value (syn) (Figure 1C). We then investigated
the detailed enzyme characteristics of FmDTA. When we
measured the enzyme activity at various temperatures, FmDTA
showed a maximum activity at 35 °C and maintained its
specific activity around 150 μmol min−1 mg−1 over the broad
temperature range from 25 to 50 °C (Figure 1D). FmDTA
showed a clear dependency on divalent metal ions, such as
Mn2+, Mg2+, Co2+, and Fe2+, and the highest activity was
observed for Mn2+ (Figure 1E). FmDTA also required
approximately 100 μM extrinsic PLP for the maximized
activity; however, the enzyme showed higher syn-selectivity
under a lower PLP concentration (Figure 1F).
Crystal Structure of FmDTA. Although the diastereose-
lectivity of FmDTA was relatively higher than that of other
DTAs we tested, the product still has a racemic problem when
using the DTA-based method. We then determined the crystal
structure of FmDTA to elucidate the molecular mechanism of
the enzyme and to provide a rational strategy increasing the
Cβ-specificity. The crystals belonged to the P21212 space group
with unit cell dimensions of a = 71.0, b = 178.1, and c = 65.5
(Table S3). The asymmetric unit contained a dimeric form of
FmDTA, and the size-exclusion chromatographic data also
revealed that the protein functions as a dimer in solution
(Figure S2). The structure of FmDTA exhibits a canonical
alanine racemase fold: the (α/β)8-barrel-like domain was
decorated with a subsidiary part (β-decoration) at the N-,C-
termini, which comprises an α-helix, a 310-helix, and eight
continuously connected tangled β-strands (Figure S3).
Although the dimeric structure of FmDTA exhibits a root-
mean-square deviation value of 0.78 with the Cα-coordinates
of AxDTA, the spatial coordinates are somewhat different,
6893
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 2. Structure and key residues of FmDTA. (A) Overall dimeric structure of FmDTA. The structure of FmDTA is shown as cartoon (upper)
and surface (lower) models. (B) Active site of FmDTA. The key residues at the active site are shown as magenta sticks and superposed with the
AxDTA cartoon model with a Mn2+ (PDB code 4V15). Key residues of FmDTA and AxDTA are distinguished with colors of magenta and gray,
respectively. Hydrogen bonds and coordination interaction are indicated by yellow dotted lines. The simulated-annealing composite OMIT map for
PLP contoured at 1.0σ is shown with the cyan mesh. (C, D) Site-directed mutagenesis. The enzyme activities of the variants were measured at the
pH 9.0 condition with 100 mM glycine, 100 mM acetaldehyde, 0.1 mM PLP, and 0.1 mM MnCl2, and 0.015 mg mL−1 enzyme. The relative activity
is shown with bar graph, and the corresponding de value is shown with triangles. Standard deviations are shown as bars (n = 3).
which is likely attributed to the differences in their enzymatic
characteristics (Figure S4).
(Figure 2C). However, alanine racemases, which share a fold
that is similar to DTAs but does not adopt a quinonoid
formation during enzyme catalysis, have an arginine residue at
the corresponding position of Gln239 of FmDTA.26,27
Structural comparisons with other PLP-dependent enzymes
at the corresponding position implied that Gln239 is related
with the protonation state of N1-PLP and quinonoid
intermediate formation (Figure S5, Note S1). The complete
loss of activity of the FmDTAQ239A variant confirmed that the
Gln239 residue is heavily involved in enzyme catalysis (Figure
2C). Moreover, we found several unique residues in the second
layer of the PLP-binding site of FmDTA, such as Gly145,
His288, Asp294, and Ser312 (Figure 2B). Substitutions of
these residues with the corresponding residues of AxDTA
showed a 25−80% decrease in activity (Figure 2D), which
indicates that these unique residues might contribute to the
higher catalytic potential of FmDTA rather than that of
AxDTA.
Lys49-PLP internal aldimine is located at the active site
cavity that forms at the dimerization interface (Figure 2A,B).
The PLP coenzyme is surrounded by residues that are
presumed to be involved in enzyme catalysis and/or
substrate/metal ion binding (Figure 2B). The pyridoxal moiety
of PLP is interposed in the ring structures of Tyr177 and
His47, and the residues of Gln71 and Ser90 are also in contact
with the moiety (Figure 2B). The phosphate moiety forms
hydrogen bonds with the hydroxyl groups of Thr223, Ser242,
and Tyr250 (Figure 2B). His339 and Asp341 are highly
conserved residues for the coordination of Mn2+ in AxDTA,
and His183 is located close to the position of the metal ion
(Figure 2B).21 The residue involvement in enzyme activity was
assessed by alanine-scanning site-directed mutagenesis, and the
replacements resulted in complete loss or a dramatic decrease
of aldol addition activities (Figure 2C). Interestingly, the
Gln239 residue is located in the vicinity of the pyridine
nitrogen of PLP (N1-PLP) with a distance of 3.1 Å by
providing a polar carboxamide moiety to the nitrogen atom
Catalytic Mechanism of FmDTA with an Inner Proton
Sink. Comprehensive understanding of the catalytic mecha-
nisms of DTAs is crucial for determining Cβ-stereospecificity,
6894
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 3. Catalytic residues and pyridine-ring-extended plane in FmDTA. (A) The critical steps in the proposed mechanism showing circular
electron-relay by the proton sink. PLP and PLP* connote the PLP structure with an amine group and its quinonoid form, respectively. The whole
mechanism is described in Scheme S1. (B) Mirror-like catalytic site superposition between FmDTA and an LTA from Aeromonas jandaei (PDB
code 3WGB). The residues and the glycine-PLP external aldimine in the LTA are shown as yellow sticks. Lys49 and its surrounding residues in
FmDTA are shown with magenta sticks. The pyridine-ring-extended plane and the possible D-specific position of aldehyde are shown and labeled.
Two residues of FmDTA and AxDTA close to the aldehyde position are shown with magenta and gray sticks, respectively. The Mn2+ ion is
prepared from the structure of AxDTA. The small figure is a top view of the large figure, and the green arrow indicates the possible position of the
free Lys49.
because the positional relationship between the aldehyde and
glycine substrates determines the stereospecificity of the
enzyme. A previous study on AxDTA suggested the retro-
aldol reaction (reverse reaction) mechanism of DTA, where
His193 drives water-mediated deprotonation of the Cβ-
hydroxyl group of a substrate.15,21 The mechanism is quite
similar to a mechanism found in LTAs and the redesigned
alanine racemases, by using the histidine residue as a
protonation and deprotonation mediator.18,28 The proposed
mechanism, however, is in conflict with our mutational study
where the FmDTAH183A variant showed a detectable activity
(Figure 2C). Production of a significant amount of D-threonine
and D-allo-threonine with a higher concentration of the variant
indicates that the histidine residue does not participate directly
in enzyme catalysis (Figure S6). In fact, His183 is involved in
the binding of PLP by stabilizing the crucial Tyr250 residue
(Figure S7), which seems to explain the reduced activity of the
variant. Moreover, considering the pKa of biprotonated
histidine,29,30 the proton-received histidine may not be stable
at the optimum pH values for DTA (Figure 1B).
It is of interest that, besides His183, there is no acidic
protein residue that can give a proton to the oxygen of
aldehyde within the active site (Figure 2B). We thus assumed a
mechanism that excludes the external proton-withdrawing
residue, reassessing the functions of the metal ion and Lys49
(Figure 3A, Scheme S1). The lysine-mediated Cα-deprotona-
tion is a reaction trigger mechanism commonly found in the
PLP-dependent enzymes, such as aminotransferases, tyrosine
phenol-lyases, and O-phospho-L-seryl-tRNASec:L-selenocys-
tein-tRNA synthase (Note S1).31−33 In the newly proposed
mechanism, proton abstraction at the Cα by Lys49 inner
proton sink initiates enzyme catalysis (Figure 3A, Scheme S1).
When the resulting carbanion attacks aldehyde, the divalent
metal ion stabilizes the oxyanion state of the Cβ-oxygen
(Figure 3A, Scheme S1), as a rate limiting step.11 Then, Cβ-
oxygen receives a proton through a proton-for-proton swap at
Cα, where nucleophilic oxygen attacks the other proton of Cα
and the carbon regains a proton from the sink (Figure 3A,
Scheme S1). An interesting observation is that His47 of
FmDTA blocks the corresponding position of the aldimine-
forming lysine of LTA in a mirror-like catalytic site
6895
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
Figure 4. Rational design for the syn-specific DTA. The middle panel shows 3D configurations, and the left and right panels show a 2D diagram for
the syn- and anti-positions of acetaldehyde, respectively. The FmDTA residues and the PLP-Lys49 internal aldimine are shown with magenta sticks
and labeled. The Mn2+ ion is prepared from the structure of AxDTA. G179A and acetaldehyde are prepared by PyMol manipulation and shown as a
surface and stick model, respectively. The side-chain of G179A is shown with a block diagram in the two outside panels.
Figure 5. Near-perfect Cβ-stereoselectivity of the FmDTA variants. (A) Relative activities and de values (syn) of the engineered enzymes. Standard
deviations are shown as bars (n = 3). (B) Rational design application to other DTAs. Standard deviations are shown with bars (n = 3). (C−E)
Changes of accumulated products and de values by the FmDTA variants. Pink and blue areas indicate concentrations of D-threonine and D-allo-
threonine, respectively. The reactions were performed with 500 mM glycine, 100 mM acetaldehyde, 0.1 mM PLP, 0.1 mM MnCl2, and enzyme
concentrations of 0.01, 0.05, or 0.09 mg mL−1, respectively. Standard deviations are shown (n = 3). (F−H) Changes of accumulated products and
de values by the FmDTA variants over the long period of time. The reactions were performed under the same conditions as above, but the
temperature was set at 15 °C in view of the volatility of acetaldehyde. The decreased de values are fitted to the logistic equation, and the graph is
shown with a gray line. Standard deviations are shown (n = 3).
superposition, indicating that DTA has a different mechanism
compared to those of LTAs (Figure 3B), whereupon when
external aldimine is formed, the free and uncharged lysine may
be positioned to a limited space located under the pyridine-
6896
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
ring-extended plane close to the potential position of Cα
(Figure 3B). Asp313 seems to aid the proton abstraction by
the sink, and complete activity abolishment of the
FmDTAD313A variant and restriction of Lys49 positioning
strongly support the proton sink concept (Figures 2C and 3B).
This proposed mechanism explains the metal-dependency of
DTA and the flow of protons that transform aldehyde into the
Cβ-hydroxyl group of a product. According to the mechanism,
DTA shows a unique enzymatic aldol addition model where
only the metal ion and Cα with the sink wait for the aldehyde
substrate to be trapped, which will be described in the next
section.
FmDTAG179A and the FmDTAG179A/S312A variants over time,
the variants maintained high de values of over 93% and 95%,
respectively (Figure 5C−E). Since it is known that one of the
drawbacks of TAs is a decrease in the Cβ-stereoselectivity over
time,11,34 16 h reactions were conducted to validate that the
FmDTA variants maintain the high selectivity without kinetic
control. Although the overall conversion was not affected, the
wild-type showed a severe decrease of Cβ-stereoselectivity by a
17.4% change in the de value, indicating the kinetic effect in
the 4 h reaction (Figure 5F), whereas the Cβ-stereoselectivities
of FmDTAG179A and FmDTAG179A/S312A were slightly reduced
by 1.8% and 2.4% over the long period of time (Figure 5G,H).
The effect of our rational design was further demonstrated with
the same experiments using BhDTA and the BhDTAG192A
variant (Figure S8).
We further tested if the variants can be utilized for other
aldehydes. We first measured D-hydroxynorvaline production
using a propionaldehyde substrate. For the 16 h reaction, the
wild-type produced both syn- and anti-forms of D-hydrox-
ynorvaline while the peak corresponding to the anti-form was
negligible in the reactions of FmDTAG179A and
FmDTAG179A/S312A (Figure S9). It has been known that
DTAs tend to produce syn-form products when bulky aromatic
substrates are used,13,25 which leads us to presume that the syn-
specific active sites of these variants might prefer benzaldehyde
to acetaldehyde as a substrate. When benzaldehyde was used as
a substrate, FmDTAWT showed a much higher Cβ-stereo-
selectivity with a de value (syn) of 94.5% than when using
acetaldehyde, and the syn-specific FmDTAG179A and
FmDTAG179A/S312A variants exhibited even higher de values
(Table S5). Surprisingly, the catalytic ability of the variants
increased 1.40 and 2.61 times compared to that of the wild-
type (Table S5). These results are consistent with the
presumption that there is a relationship between the
mechanistic positioning of aldehyde and the rate of bond
formation, further emphasizing that the rationale of the DTA
variants is suitable for aldehyde substrates of various sizes.
Rational Engineering for High Cβ-Stereoselectivity.
Our proposed mechanism suggests that the aldehyde substrate
is coordinated to a metal ion and is located on the quinonoid
structure opposite the inner proton sink (Figure 3). Here, we
can expect that the aldehyde substrate in a plane configuration
of the formyl group has two possible positions for its R-group,
considering the interaction with the quinonoid intermediate.
When the R-group is located toward the inner position of the
cavity, the anti-form of its product (D-allo-threonine) is formed
(Figure 4). On the contrary, when the R-group is located
toward the outer position of the cavity, the syn-form of its
product (D-threonine) is formed (Figure 4). We then designed
the FmDTAG179A variant based on the rationale that an
artificial spatial constraint might hinder the placement of the
R-group at the inner position of the cavity (Figure 4). The
variant showed a dramatically increased de value of 95.0% for
the syn-diastereomer, although its activity decreased to 22.6%
of the wild-type (Figure 5A). Surprisingly, mutations of the
corresponding glycine residue of other DTAs, such as BhDTA,
AiDTA, OcDTA, AxDTA, and DrDTA, to alanine resulted in a
dramatic increase of the syn-specificity (Figure 5B). These
results confirm that the rationale can also be applied to other
DTAs, which further supports our proposed reaction
mechanism and the aldehyde binding mode of DTA. We
also observed the Arg147 and Ser312 residues in the vicinity of
the aldehyde binding site (Figures 2B and 3), and mutations of
these residues to alanine resulted in one-third of the activity
and the increased de value compared with that of the wild-type
(Figure 5A), which indicates that these two residues affect the
positioning of aldehyde and consequently the stereospecificity
determination.
We then attempted to combine the G179A mutation with
the R147A and S312A mutations to achieve even higher syn-
specific variants. Unfortunately, the FmDTAR147A/G179A variant
showed a decreased de value, which implies that the R147A
mutation abolished the hindrance effect of the G179A
mutation (Figure 5A). However, the FmDTAG179A/S312A variant
exhibited a much higher de value of 99.5% (Figure 5A).
Kinetic analysis of the FmDTAG179A/S312A variant showed a
similar level of Km and a reduced kcat, compared with those of
the wild-type (Table S4). These results imply that the
mutations only affected the direction of the substrate binding
without changing its affinity for the substrate, and the position
of the bound substrate in the variant is slightly different from
that needed for optimal enzyme catalysis. Next, we measured
the D-threonine production by the FmDTA variants for 4 h to
monitor the changes in the Cβ-stereoselectivity during the
extended catalytic processes. We adjusted enzyme quantities to
secure an activity level of the wild-type for each of the variants
to compare the selectivity while minimizing the kinetic effect.
Although the de values (syn) decreased in both the
CONCLUSIONS
■
In this study, the FmDTA enzyme with high activity and
selectivity for D-threonine production was identified. On the
basis of the structural and biochemical analyses of FmDTA, we
proposed a catalytic mechanism of DTA, in which the divalent
metal ion plays an important role in the common PLP-
dependent mechanism. Comprehensive understanding of the
catalytic mechanism of DTA provided us with a strategy for
rational protein engineering, by which we achieved the
FmDTA variants having near-perfect Cβ-stereoselectivity of a
de value of 99.5% for D-threonine. We also demonstrate that
the rational design is also applicable to other DTA
homologues. This is the first case of asymmetric bioconversion
of short-chain aldehydes into β-hydroxy-α-D-amino acids by an
unnatural syn-specific DTA (Table S6). Furthermore, the
FmDTAG179A/S312A variant showed enhanced catalytic activity
and Cβ-stereoselectivity against the benzaldehyde substrate.
This study on DTAs shows how mechanism-based rational
engineering can control enzyme catalysis and can be utilized
for β-hydroxy-α-D-amino acid conversion.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/acscatal.1c01348.
■
sı
6897
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
The structure of FmDTA is deposited in Protein Data Bank
with an accession code of 7DIB. The source data underlying
Figures 1, 2, and 5a are associated with raw data. The data sets
generated and analyzed during the current study are available
from the corresponding authors upon request.
Methods and supporting note (Note S1) and additional
results presented in tables (Tables S1−S6), figures
(Figures S1−S10), and scheme (Scheme S1) (PDF)
AUTHOR INFORMATION
Corresponding Authors
Kyung-Jin Kim − School of Life Sciences, BK21 FOUR KNU
Creative BioResearch Group, KNU Institute for
■
ACKNOWLEDGMENTS
■
This work was supported by the C1 Gas Refinery Program
( N R F - 2 0 1 8 M 3 D 3 A 1 A 0 1 0 5 5 7 3 2 a n d N R F -
2018M3D3A1A01056181) and the Basic Science Research
Program (NRF-2020R1A4A1018332) funded by the Ministry
of Science and ICT, and the Korea Research Institute of
Bioscience and the Biotechnology (KRIBB) Research Initiative
Program (KGM5402113).
Microorganisms, Kyungpook National University, Daegu
41566, Republic of Korea; orcid.org/0000-0002-0375-
2561; Phone: +82-53-950-5377; Email: kkim@knu.ac.kr
Seung-Goo Lee − Synthetic Biology and Bioengineering
Research Center, Korea Institute of Bioscience and
Biotechnology (KRIBB), Daejeon 34141, Republic of Korea;
Department of Biosystems and Bioengineering, KRIBB
School, University of Science and Technology (UST), Daejeon
34113, Republic of Korea; orcid.org/0000-0003-4539-
9812; Phone: +82-42-860-4373; Email: sglee@kribb.re.kr
Soo-Jin Yeom − School of Biological Sciences and Technology,
Chonnam National University, Gwangju 61186, Republic of
Korea; Phone: +82-62-530-1911; Email: soojin258@
chonnam.ac.kr
REFERENCES
■
(1) Windle, C. L.; Mu
̈
ller, M.; Nelson, A.; Berry, A. Engineering
aldolases as biocatalysts. Curr. Opin. Chem. Biol. 2014, 19, 25−33.
(2) de Berardinis, V.; Guérard-Hélaine, C.; Darii, E.; Bastard, K.;
Hélaine, V.; Mariage, A.; Petit, J.-L.; Poupard, N.; Sánchez-Moreno,
I.; Stam, M.; Gefflaut, T.; Salanoubat, M.; Lemaire, M. Expanding the
reaction space of aldolases using hydroxypyruvate as a nucleophilic
substrate. Green Chem. 2017, 19, 519−526.
Authors
(3) Clapés, P.; Fessner, W. D.; Sprenger, G. A.; Samland, A. K.
Recent progress in stereoselective synthesis with aldolases. Curr. Opin.
Chem. Biol. 2010, 14, 154−167.
Sung-Hyun Park − Synthetic Biology and Bioengineering
Research Center, Korea Institute of Bioscience and
Biotechnology (KRIBB), Daejeon 34141, Republic of Korea;
Department of Biosystems and Bioengineering, KRIBB
School, University of Science and Technology (UST), Daejeon
34113, Republic of Korea
Hogyun Seo − School of Life Sciences, BK21 FOUR KNU
Creative BioResearch Group, KNU Institute for
Microorganisms, Kyungpook National University, Daegu
41566, Republic of Korea; Pohang Accelerator Laboratory,
Pohang University of Science and Technology, Pohang 37673,
Republic of Korea
(4) Wechsler, C.; Meyer, D.; Loschonsky, S.; Funk, L. M.;
Neumann, P.; Ficner, R.; Brodhun, F.; Muller, M.; Tittmann, K.
̈
Tuning and Switching Enantioselectivity of Asymmetric Carboligation
in an Enzyme through Mutational Analysis of a Single Hot Spot.
ChemBioChem 2015, 16, 2580−2584.
(5) Schmidt, N. G.; Eger, E.; Kroutil, W. Building bridges:
Biocatalytic C−C-bond formation toward multifunctional products.
ACS Catal. 2016, 6, 4286−4311.
(6) Yoo, C. H.; Yook, J. S.; Lee, W. H.; Lee, J. S.; Lee, J. C.; Lee, C.
M.; Lee, T. S. Novel method for producing cefbuperazone
intermediate. KR patent 2009, KR 100898103 B1.
Jihye Seok − School of Life Sciences, BK21 FOUR KNU
Creative BioResearch Group, KNU Institute for
Microorganisms, Kyungpook National University, Daegu
41566, Republic of Korea
Haseong Kim − Synthetic Biology and Bioengineering
Research Center, Korea Institute of Bioscience and
Biotechnology (KRIBB), Daejeon 34141, Republic of Korea;
Department of Biosystems and Bioengineering, KRIBB
School, University of Science and Technology (UST), Daejeon
34113, Republic of Korea
(7) Liu, J. Q.; Odani, M.; Dairi, T.; Itoh, N.; Shimizu, S.; Yamada, H.
A new route to L-threo-3-[4-(methylthio)phenylserine], a key
intermediate for the synthesis of antibiotics: recombinant low-
specificity D-threonine aldolase-catalyzed stereospecific resolution.
Appl. Microbiol. Biotechnol. 1999, 51, 586−591.
(8) Duckers, N.; Baer, K.; Simon, S.; Gröger, H.; Hummel, W.
̈
Threonine aldolases-screening, properties and applications in the
synthesis of non-proteinogenic beta-hydroxy-alpha-amino acids. Appl.
Microbiol. Biotechnol. 2010, 88, 409−424.
(9) Miller, M. J.; Mattingly, P. G.; Morrison, M. A.; Kerwin, J. F., Jr
Synthesis of. beta.-lactams from substituted hydroxamic acids. J. Am.
Chem. Soc. 1980, 102, 7026−7032.
Kil Koang Kwon − Synthetic Biology and Bioengineering
Research Center, Korea Institute of Bioscience and
Biotechnology (KRIBB), Daejeon 34141, Republic of Korea
(10) Chen, Q.; Chen, X.; Feng, J.; Wu, Q.; Zhu, D.; Ma, Y.
Improving and inverting Cβ-stereoselectivity of threonine aldolase via
substrate-binding-guided mutagenesis and a stepwise visual screening.
ACS Catal. 2019, 9, 4462−4469.
Complete contact information is available at:
https://pubs.acs.org/10.1021/acscatal.1c01348
(11) Fesko, K.; Reisinger, C.; Steinreiber, J.; Weber, H.; Schurmann,
̈
M.; Griengl, H. Four types of threonine aldolases: similarities and
differences in kinetics/thermodynamics. J. Mol. Catal. B: Enzym.
2008, 52, 19−26.
Author Contributions
#S.-H.P. and H.S. contributed equally to this work. S.-J.Y., S.-
G.L., and K.-J.K. conceived the project. S.-H.P., H.K., K.K.K.,
and S.-J.Y. performed the biochemical experiments. H.S. and
J.S. performed the structural determination and the rational
design. S.-H.P. and H.S. analyzed the data. S.-G.L., K.-J.K., S.-
H.P., H.S., and S.-J.Y. wrote the paper.
(12) Franz, S. E.; Stewart, J. D. Threonine aldolases. Adv. Appl.
Microbiol. 2014, 88, 57−101.
(13) Kimura, T.; Vassilev, V. P.; Shen, G. J.; Wong, C. H. Enzymatic
synthesis of β-hydroxy-α-amino acids based on recombinant D-and L-
threonine aldolases. J. Am. Chem. Soc. 1997, 119, 11734−11742.
(14) Steinreiber, J.; Fesko, K.; Mayer, C.; Reisinger, C.; Schurmann,
̈
Notes
M.; Griengl, H. Synthesis of γ-halogenated and long-chain β-hydroxy-
α-amino acids and 2-amino-1, 3-diols using threonine aldolases.
Tetrahedron 2007, 63, 8088−8093.
The authors declare the following competing financial
interest(s): 10-2019-0131991 and PCT/KR2020/01429.
6898
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899

ACS Catalysis
pubs.acs.org/acscatalysis
Research Article
(15) Fesko, K. Threonine aldolases: perspectives in engineering and
screening the enzymes with enhanced substrate and stereo
specificities. Appl. Microbiol. Biotechnol. 2016, 100, 2579−2590.
(16) Fesko, K. Comparison of L-Threonine Aldolase Variants in the
Aldol and Retro-Aldol Reactions. Front. Bioeng. Biotechnol. 2019, 7,
119.
(17) Qin, H. M.; Imai, F. L.; Miyakawa, T.; Kataoka, M.; Kitamura,
N.; Urano, N.; Mori, K.; Kawabata, H.; Okai, M.; Ohtsuka, J.; Hou,
F.; Nagata, K.; Shimizu, S.; Tanokura, M. L-allo-threonine aldolase
with an H128Y/S292R mutation from Aeromonas jandaei DK-39
reveals the structural basis of changes in substrate stereoselectivity.
Acta Crystallogr., Sect. D: Biol. Crystallogr. 2014, 70, 1695−1703.
(18) di Salvo, M. L.; Remesh, S. G.; Vivoli, M.; Ghatge, M. S.;
Paiardini, A.; D’Aguanno, S.; Safo, M. K.; Contestabile, R. On the
catalytic mechanism and stereospecificity of Escherichia coli L-
threonine aldolase. FEBS J. 2014, 281, 129−145.
phenol-lyase from X-ray structures of quinonoid intermediates. J. Biol.
Chem. 2008, 283, 29206−29214.
(33) Palioura, S.; Sherrer, R. L.; Steitz, T. A.; Söll, D.; Simonovic, M.
The human SepSecS-tRNASec complex reveals the mechanism of
selenocysteine formation. Science 2009, 325, 321−325.
(34) Xu, L.; Wang, L. C.; Xu, X. Q.; Lin, J. Characteristics of L-
threonine transaldolase for asymmetric synthesis of β-hydroxy-α-
amino acids. Catal. Sci. Technol. 2019, 9, 5943−5952.
(19) Gwon, H. J.; Baik, S. H. Diastereoselective synthesis of L-threo-
3,4-dihydroxyphenylserine by low-specific L-threonine aldolase
mutants. Biotechnol. Lett. 2010, 32, 143−149.
(20) Riario-Sforza, G.; Pagani, R.; Marinello, E. Threonine aldolase
and allothreonine aldolase in rat liver. Eur. J. Biochem. 1969, 8, 88−92.
(21) Uhl, M. K.; Oberdorfer, G.; Steinkellner, G.; Riegler-Berket, L.;
Mink, D.; van Assema, F.; Schurmann, M.; Gruber, K. The crystal
̈
structure of D-threonine aldolase from Alcaligenes xylosoxidans
provides insight into a metal ion assisted PLP-dependent mechanism.
PLoS One 2015, 10, e0124056.
(22) Hirato, Y.; Tokuhisa, M.; Tanigawa, M.; Ashida, H.; Tanaka,
H.; Nishimura, K. Cloning and characterization of d-threonine
aldolase from the green alga Chlamydomonas reinhardtii. Phytochem-
istry 2017, 135, 18−23.
(23) Liu, J. Q.; Odani, M.; Yasuoka, T.; Dairi, T.; Itoh, N.; Kataoka,
M.; Shimizu, S.; Yamada, H. Gene cloning and overproduction of low-
specificity D-threonine aldolase from Alcaligenes xylosoxidans and its
application for production of a key intermediate for parkinsonism
drug. Appl. Microbiol. Biotechnol. 2000, 54, 44−51.
(24) Kataoka, M.; Ikemi, M.; Morikawa, T.; Miyoshi, T.; Nishi, K.;
Wada, M.; Yamada, H.; Shimizu, S. Isolation and characterization of
D-threonine aldolase, a pyridoxal-5′-phosphate-dependent enzyme
from Arthrobacter sp. DK-38. Eur. J. Biochem. 1997, 248, 385−393.
(25) Chen, Q.; Chen, X.; Cui, Y.; Ren, J.; Lu, W.; Feng, J.; Wu, Q.;
Zhu, D. A new d-threonine aldolase as a promising biocatalyst for
highly stereoselective preparation of chiral aromatic β-hydroxy-α-
amino acids. Catal. Sci. Technol. 2017, 7, 5964−5973.
(26) Major, D. T.; Gao, J. A combined quantum mechanical and
molecular mechanical study of the reaction mechanism and alpha-
amino acidity in alanine racemase. J. Am. Chem. Soc. 2006, 128,
16345−16357.
(27) Chan-Huot, M.; Dos, A.; Zander, R.; Sharif, S.; Tolstoy, P. M.;
Compton, S.; Fogle, E.; Toney, M. D.; Shenderovich, I.; Denisov, G.
S.; Limbach, H. H. NMR studies of protonation and hydrogen bond
states of internal aldimines of pyridoxal 5′-phosphate acid-base in
alanine racemase, aspartate aminotransferase, and poly-L-lysine. J. Am.
Chem. Soc. 2013, 135, 18160−18175.
(28) Seebeck, F. P.; Hilvert, D. Conversion of a PLP-dependent
racemase into an aldolase by a single active site mutation. J. Am.
Chem. Soc. 2003, 125, 10158−10159.
(29) Li, S.; Hong, M. Protonation, tautomerization, and rotameric
structure of histidine: a comprehensive study by magic-angle-spinning
solid-state NMR. J. Am. Chem. Soc. 2011, 133, 1534−1544.
(30) Hansen, A. L.; Kay, L. E. Measurement of histidine pKa values
and tautomer populations in invisible protein states. Proc. Natl. Acad.
Sci. U. S. A. 2014, 111, E1705−E1712.
(31) Kirsch, J. F.; Eichele, G.; Ford, G. C.; Vincent, M. G.;
Jansonius, J. N.; Gehring, H.; Christen, P. Mechanism of action of
aspartate aminotransferase proposed on the basis of its spatial
structure. J. Mol. Biol. 1984, 174, 497−525.
́
́
́
(32) Milic, D.; Demidkina, T. V.; Faleev, N. G.; Matkovic-Calogovic,
D.; Antson, A. A. Insights into the catalytic mechanism of tyrosine
6899
https://doi.org/10.1021/acscatal.1c01348
ACS Catal. 2021, 11, 6892−6899