Mechanism-Guided Engineering of ω-Transaminase to Accelerate Reductive Amination of Ketones

Article abstract of DOI:10.1002/adsc.201500211

Asymmetric reductive amination of ketones using ω-transaminases (ω-TAs) offers a promising alternative to the chemocatalytic synthesis of chiral amines. One fundamental challenge to the biocatalytic strategy is the very low enzyme activities for most ketones compared with native substrates (i.e., <1% relative to pyruvate). Here we have demonstrated that a single point mutation in the active site of the (S)-selective ω-TA from Ochrobactrum anthropi could induce a remarkable acceleration of the amination reaction without any loss in stereoselectivity and enzyme stability. Molecular modeling of quinonoid intermediates, alanine scanning mutagenesis and kinetic analysis revealed that the W58 residue acted as a steric barrier to binding and catalytic turnover of ketone substrates. Removal of the steric strain by W58L substitution, which was selected by partial saturation mutagenesis, led to dramatic activity improvements for structurally diverse ketones (e.g., 340-fold increase in k_cat/K_M for acetophenone). The W58L mutant afforded an efficient synthesis of enantiopure amines (i.e., >99% ee) using isopropylamine as an amino donor.

Full text of DOI:10.1002/adsc.201500211

FULL PAPERS
DOI: 10.1002/adsc.201500211
Mechanism-Guided Engineering of w-Transaminase to Accelerate
Reductive Amination of Ketones
+
a
+a
a
a,
Sang-Woo Han, Eul-Soo Park, Joo-Young Dong, and Jong-Shik Shin *
a
Department of Biotechnology, Yonsei University, Shinchon-Dong 134, Seodaemun-Gu, Seoul 120-749, South Korea
Fax: (+82)-2-362-7265; phone: (+82)-2-2123-5884; e-mail: enzymo@yonsei.ac.kr
+
These authors contributed equally to this work.
Received: March 1, 2015; Published online: May 13, 2015
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201500211.
Abstract: Asymmetric reductive amination of ke- acted as a steric barrier to binding and catalytic turn-
tones using w-transaminases (w-TAs) offers a promis- over of ketone substrates. Removal of the steric
ing alternative to the chemocatalytic synthesis of strain by W58L substitution, which was selected by
chiral amines. One fundamental challenge to the bio- partial saturation mutagenesis, led to dramatic activi-
catalytic strategy is the very low enzyme activities ty improvements for structurally diverse ketones
for most ketones compared with native substrates (e.g., 340-fold increase in k /K for acetophenone).
cat
M
(
i.e., <1% relative to pyruvate). Here we have dem- The W58L mutant afforded an efficient synthesis of
onstrated that a single point mutation in the active enantiopure amines (i.e., >99% ee) using isopropyl-
site of the (S)-selective w-TA from Ochrobactrum amine as an amino donor.
anthropi could induce a remarkable acceleration of
the amination reaction without any loss in stereose-
lectivity and enzyme stability. Molecular modeling of Keywords: asymmetric amination; chiral amines;
quinonoid intermediates, alanine scanning mutagene- molecular modeling; protein engineering; w-transa-
sis and kinetic analysis revealed that the W58 residue minases
[
6]
Introduction
amines from prochiral ketones using w-TAs. Incon-
sistent with the industrial need, w-TAs have not
Asymmetric synthesis of chiral compounds via biocat- evolved to exploit ketones as native amino acceptors
alytic strategies has attracted growing attention as an but rather a limited range of keto acids (i.e., pyruvate
[
7]
alternative to chemocatalytic methods due to environ- and glyoxylate) as observed with other types of
[
8]
mental and social demands for green processes in the transaminases.
This biologically driven substrate
[
1]
chemical industry. To harness the full capacity of bias leads to very low activities of all known w-TAs
biocatalysts as synthetic toolkits, enzyme properties for most ketones (i.e., <1% relative to pyruvate). For
such as stability, substrate range and optimal reaction example, the (S)-selective w-TA from Ochrobactrum
conditions should be compatible with a manufacturing anthropi (OATA) exhibits only 0.03% activity for ace-
setting. However, the design of a biocatalytic pro- tophenone (i.e., a typical ketone substrate) relative to
cess is often limited by sophisticated enzyme proper- pyruvate (see Table S1 in the Supporting Informa-
ties evolved for biological fitness that is incongruous tion). Such a low substrate reactivity should be com-
with industrial demands. A typical example is w- pensated by high enzyme dosage to achieve reasona-
transaminase (w-TA) which possesses a unique cata- ble reaction rates, which poses a significant challenge
lytic capability affording the stereoselective transfer for implementing a scalable process.
[2]
[
3]
of an amino group from primary amines to carbonyl
In addition to the poor kinetic performance for ke-
compounds without an external cofactor such as tones, asymmetric synthesis of chiral amines using w-
[4]
NADH. Spurred by the versatile utility of chiral TAs has suffered from an unfavorable reaction equili-
[
6a]
amines as essential building blocks for a number of brium. Recent studies to cope with the thermody-
[
5]
pharmaceuticals and fine chemicals, there has been namic limitation have proved that an equilibrium shift
a massive research effort to synthesize enantiopure can be efficiently driven by several reaction engineer-
1732
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1732 – 1740

FULL PAPERS
Mechanism-Guided Engineering of w-Transaminase
ing strategies employing enzymatic or physicochemi- Results and Discussion
[
9]
cal removal of a co-product. However, only a few
examples of tackling the kinetic limitation are availa- Structural Modeling of Quinonoid Intermediates
[6d,10]
ble.
To boost the synthetic utility of w-TAs, it is
imperative to develop a facile strategy for the crea- w-TA exploits pyridoxal 5’-phosphate (PLP) as a pros-
tion of enzyme variants displaying desirable kinetic thetic group to mediate the transfer of an amino
[
4]
properties with ketones. In this regard, Savile et al. group. The whole reaction consists of two half reac-
demonstrated an expansion of substrate specificity of tions (i.e., oxidative deamination of an amino donor
(
R)-selective w-TA from Arthrobacter sp. (ARTA) by and reductive amination of an amino acceptor), which
protein engineering involving 27 amino acid substitu- involves multiple reaction intermediates
[6d]
[12]
tions. Contrary to expectations, the engineered var- (Scheme 1). The PLP form of the enzyme is present
iant of ARTA showed a modest activity improvement as a Schiff base formed with an active site lysine (i.e.,
toward acetophenone (i.e., only 46% increase; Sup- internal aldimine). Transaldimination with an incom-
porting Information, Table S2). Moreover, the mas- ing amino donor converts the internal aldimine to an
sive mutations led to a significant loss in the stereose- external aldimine (i.e., a Schiff base between PLP
lectivity toward some ketones (e.g. benzylacetone; and the amino donor). Abstraction of a C-a hydrogen
Supporting Infromation, Table S3). This led us to ex- from the external aldimine produces an unstable car-
plore the feasibility of whether active site redesign in- banionic intermediate that is converted to a ketimine
volving minimal mutations could fulfill a desirable im- after protonation at the C-4’ position. Hydrolysis of
provement in amination kinetics for structurally di- the ketimine leads to the formation of a ketone prod-
verse ketones without deteriorating essential parental uct and a pyridoxamine 5’-phosphate (PMP) form of
properties such as stereoselectivity. To this end, the enzyme (i.e., E-PMP). The reverse reaction (i.e.,
OATA was chosen for protein engineering owing to from E-PMP to the internal aldimine) represents the
[
11]
its exceptionally high activity for isopropylamine
which is an ideal amino donor for industrial applica-
reductive amination of acetophenone.
It is generally accepted that the carbanionic inter-
[6d]
tions.
Instead of time-consuming iterative library mediate is the most unstable reaction intermediate
generation and selection, we sought to pinpoint a key which should be resonance-stabilized by a quinonoid
[
13]
residue by mechanism-based structural modeling and form (Scheme 2). Thus, effective stabilization of the
optimize the hot spot by saturation mutagenesis. quinonoid in the active site is known to be pivotal in
[
13]
the transaminase reactions. This led us to posit that
Scheme 1. Detailed reaction pathway showing the oxidative deamination of a-methylbenzylamine.
Scheme 2. Resonance structures of the carbanionic intermediate.
Adv. Synth. Catal. 2015, 357, 1732 – 1740
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1733

Sang-Woo Han et al.
FULL PAPERS
Figure 1. Molecular model of the quinonoid intermediates in the active site of OATA. Thick sticks represent the quinonoid
intermediate. Thin sticks represent nine active site residues (i.e., Y20, L57, W58, F86, Y151, V154, A230, I261 and T324).
W58 is shown in a CPK representation. A) Acetophenone-quinonoid intermediate where the substrate moiety is colored in
yellow. Color use of the active site residues is consistent with that of labels. B) Pyruvate-quinonoid intermediate. R417 is
a conserved active site residue that forms hydrogen bonds with a carboxylate of the keto acid substrate. The active site argi-
[14]
nine is known to undergo a gross conformational change, depending on the hydrophobic type of the incoming substrate.
the very low activity of w-TAs toward ketones, rela- except for A230, and assessed whether the model pre-
tive to native amino acceptors, would result from diction was in accordance with experimental results
a failure in the effective stabilization of the quinonoid (Table 1). Indeed, W58A mutation led to an abrupt
intermediate. Therefore, we performed in silico mod- activity improvement for acetophenone (i.e., 41-fold
eling of the two quinonoids formed with acetophe- increase) and the substitution with alanine elsewhere
none and pyruvate to compare differences in the did not elicit such an activity increase except for
structural environments. The molecular model identi- modest activity improvements by V154A and I261A.
fied six active site residues (i.e., L57, W58, Y151, To visualize whether the W58A mutation relieved the
A230, I261 and T324) whose side chains lay within steric strain, we performed molecular modeling of the
a 3 Š distance from the substrate moiety of the qui- acetophenone-quinonoid in the W58A mutant (Sup-
nonoid formed with acetophenone (Figure 1A). porting Informationb, Figure S1). The model predic-
Among the nearby residues, W58 was found to engen- tion supported that the activity improvement came
der a severe steric interference with the phenyl sub-
stituent of acetophenone. In contrast, no steric inter-
Table 1. Effect of alanine scanning mutagenesis of active site
ference of the quinonoid formed with pyruvate was
[
a]
residues on the amination of acetophenone.
observed against the proximal active site residues
Figure 1B). The modeling results suggest that the Mutation Reaction rate (”10 mMmin ) Fold increase
[b]
À2
À1
(
steric strain caused by W58 destabilizes the quinonoid
intermediate formed with a bulky ketone and thereby
leads to the drastic reduction in the reactivity of ke-
tones.
none
8.26Æ0.90
3.52Æ0.25
4.78Æ0.41
340Æ37
–
Y20A
L57A
W58A
F86A
Y151A
V154A
I261A
0.4
0.6
41
ꢀ0
0.01
1.7
2.9
0.5
[c]
n.d.
0.12Æ0.01
13.6Æ0.1
23.1Æ0.1
3.93Æ0.06
Alanine Scanning Mutagenesis
Structural models of the two quinonoid intermediates T324A
predict that substitution of W58 with a smaller resi-
due would relieve the steric strain and consequently
induce an activity enhancement for acetophenone. To
test this notion, we performed alanine scanning muta-
genesis of active site residues, shown in Figure 1
[a]
Reaction conditions: 10 mM acetophenone and 10 mM
l-alanine in phosphate buffer (50 mM, pH 7) at 378C.
Reaction rate represents the initial rate per 1 mM
enzyme.
[
[
b]
c]
n.d.: not detectable.
1734
asc.wiley-vch.de
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1732 – 1740

FULL PAPERS
Mechanism-Guided Engineering of w-Transaminase
from removal of the steric strain observed with the of acetophenone more selectively than its reverse re-
wild-type enzyme.
action. Considering the increases in both reaction
rates, we chose the W58L mutant (OATAW58L) for fur-
ther study.
Partial Saturation Mutagenesis of W58
Taken together with the modeling results, the alanine Activity Improvements for Various Ketones
scanning mutagenesis indicated that W58 imposed
a steric penalty on the quinonoid bearing a bulky sub- To assess the synthetic utility of OATAW58L, we exam-
stituent in the ketone moiety. Fine-tuning of the hot ined whether the W58L substitution improved the
spot was performed on the basis of structural consid- enzyme activity toward structurally diverse ketones
erations by partial saturation mutagenesis with hydro- using isopropylamine (2) as an amino donor
phobic residues smaller than tryptophan (i.e., (Table 2). OATAW58L showed dramatic activity im-
W58MVIL). Molecular modeling predicted that these provements toward all the ketones tested [i.e., nine
substitutions allowed a reduction in the steric strain aryl alkyl ketones (1a–i) and six alkyl ketones (1j–o)]
on the acetophenone-quinonoid as observed with the with 41- to 760-fold increases in the reaction rates.
W58A substitution (Supporting Information, Fig- The activity improvements were high enough to bring
ure S2). Hydrophilic residues were not taken into ac- about substantial reactivities of ketones that were
count for the substitution because W58 forms a hydro- inert with the wild-type enzyme (i.e., 1c, 1g and 1h).
phobic patch with neighboring residues (i.e., M54, In contrast to the engineered ARTA, OATAW58L did
L57, V233, I261 and I380; Supporting Information, not show any loss in stereoselectivity for the amina-
Figure S3). Enzyme activities of the resulting mutants, tion of all the ketones listed in Table 2 [i.e., ee values
including W58A, were compared with those of the of the produced (S)-3a–o >99%], including benzyl-
wild-type enzyme in the reaction with acetophenone acetone (1f). It is notable that the W58L mutation af-
and l-alanine as well as in its reverse reaction with fords efficient amination of bulky ketones carrying
(
(
S)-a-methylbenzylamine [(S)-a-MBA] and pyruvate even a naphthyl (1i) and an n-hexyl (1l) group.
Figure 2). We found a positive correlation between
the enzyme activities measured with the two substrate
pairs (Pearson correlation coefficient=0.86), indicat- Kinetic Analysis of the W58L Mutant
ing that the mutations benefit both reactions presum-
ably because the two reactions follow the same reac- To provide a mechanistic understanding on how the
tion pathway but run in the opposite direction. It is W58L mutation altered the catalytic properties, we
notable that all the mutations lead to much higher- carried out a kinetic analysis using four typical sub-
fold increases in the amination of acetophenone than strates (Table 3). OATAW58L showed a 55.7-fold in-
those in the deamination of (S)-a-MBA. This result crease in the turnover number (i.e., k ) for 1a, which
cat
indicates that the mutations accelerate the amination is consistent with the modeling results (i.e., reduction
in the steric strain on the quinonoid as shown in the
Supporting Information, Figure S2). Moreover, the
mutation promoted the formation of a Michaelis com-
plex (i.e., 6.1-fold tighter binding), seemingly because
binding of 1a to the active site would be facilitated by
the W58L mutation. Indeed, docking simulation sug-
gested that steric interference between W58 and the
phenyl substituent of 1a was found to be severe in the
Michaelis complex (Supporting Information, Fig-
ure S4). The facilitated binding in concert with the ac-
celerated turnover led to a 340-fold increase in the
specificity constant (i.e., k /K ) of OATA for 1a.
cat
M
W58L
It is remarkable that kcat of OATAW58L for 1a corre-
sponds to 20% of that of OATA for pyruvate.
In contrast to 1a, the W58L mutation resulted in
much weaker binding to pyruvate (i.e., 17.5-fold in-
Figure 2. Partial saturation mutagenesis of the W58 residue.
Fold increases represent the initial reaction rates of the mu-
tants normalized by those of the wild-type enzyme (i.e..
crease in K ). This result is ascribable to the loss of
M
a potential hydrogen bond donor, that is, the indole
group of W58, capable of interacting with the carbox-
À2
À2
À1
8.26”10 Æ0.90”10
and 118Æ2 mMmin
per 1 mM
enzyme for substrate pairs of acetophenone:l-alanine and ylate of pyruvate (Figure 1B). Proximity of W58 to
(
S)-a-MBA:pyruvate, respectively).
the carboxylate of pyruvate suggests that W58, in ad-
Adv. Synth. Catal. 2015, 357, 1732 – 1740
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1735

Sang-Woo Han et al.
FULL PAPERS
Table 2. Activity improvements of OATAW58L toward various ketones.
[a]
À1
Ketones
Substituents
R
Reaction rate (mMh )
OATA
Fold increase
1
2
R
OATAW58L
1a
1b
1c
1d
1e
1f
-C H
-CH3
1.92Æ0.02
0.02Æ0.01
316Æ10
5.2Æ0.2
0.13Æ0.01
162Æ8
165
260
>130
167
200
86
6
5
5
5
-C H
-CH CH
6
2
3
[b]
-C H
-(CH ) CH
n.d.
6
2
2
3
-C H -p-CH
-CH3
-CH3
-CH3
0.97Æ0.02
0.55Æ0.01
10.3Æ0.1
6
4
3
-C H -p-OCH
3
110Æ2
6
4
-(CH ) C H
885Æ5
2
2
6
5
[b]
1g
n.d.
7.6Æ0.1
>760
[b]
1h
n.d.
2.1Æ0.1
>210
1
1
1
1
1
1
1
i
j
k
l
m
n
o
-1-naphthyl
-(CH ) CH
-CH3
-CH3
-CH3
-CH3
-CH3
-CH3
-CH3
0.82Æ0.02
1.6Æ0.1
150Æ10
445Æ15
895Æ40
670Æ30
250Æ55
195Æ5
183
278
309
167
81
2
2
3
3
3
-(CH ) CH
2.9Æ0.1
2
3
-(CH ) CH
4.0Æ0.2
2
5
-CH(CH3)2
3.1Æ0.1
-CH CH(CH )
0.40Æ0.01
487
41
2
3
2
-cyclopropyl
0.32Æ0.01
13.0Æ0.1
[
a]
Reaction rates represent initial rates per 1 mM enzyme. Reaction conditions: 10 mM ketone and 10 mM isopropylamine in
phosphate buffer (50 mM, pH 7) containing 15% v/v DMSO at 378C.
n.d.: not detectable. Considering the detection limit, the reaction rate was less than 0.01 mMh .
[
b]
À1
[a]
Table 3. Kinetic parameters toward typical substrates.
[b]
Substrate
Rate constants (OATA, OATAW58L
)
Fold increase in k /KM
cat
À3 À1
À1 À1
KM [mM]
kcat [”10 s ]
k /K [M s ]
cat M
1
a
110Æ4
18Æ1
9.7Æ0.4
540Æ10
0.088
22000
160
30
340
0.41
170
8.6
pyruvate
0.12Æ0.01
150Æ16
470Æ10
2.1Æ0.04
0.73Æ0.09
53Æ2
2600Æ70
24000Æ2000
9700Æ300
19000Æ500
20000Æ1000
9300Æ600
9000
27000
180
(
S)-3a
2
21
[
[
a]
Kinetic parameters represent the apparent rate constants determined at a fixed concentration of cosubstrate.
Rate constants are mean values determined from three independent experiments.
b]
dition to R417, plays a role as a hydrogen bond OATAW58L exhibited tighter binding to 2 (i.e., 8.9-fold
donor. Nevertheless, OATAW58L performed a faster decrease in K ), implying that even a methyl substitu-
M
catalytic turnover of pyruvate than OATA did (i.e., ent of an amino donor might undergo steric interfer-
7
.1-fold increase in k ). Taken together with the in- ence with W58. The kcat values for both amino donors
cat
crease in kcat for 1a, this result suggests that the W58L were not substantially altered in contrast to the re-
mutation affords acceleration of the catalytic turnover markable increases in kcat for amino acceptors, which
irrespective of the hydrophobic nature of amino ac- indicates that the W58L mutation promotes the cata-
ceptors. Despite the accelerated turnover, the drastic lytic turnover for the amination step selectively over
decrease in the binding affinity led to 60% reduction the deamination step as observed in Figure 2. As
in k /K for pyruvate by the W58L mutation.
a result, increases in k /K for the amino donors are
cat
M
cat M
The facilitated binding of OATAW58L to the bulky solely attributed to the promotion in the binding step.
substrate (i.e., 1a) was even more reinforced for (S)- Taken together with the dramatic increase in k /KM
cat
3
a (i.e., 207-fold decrease in K ). Interestingly, for 1a, the 8.6-fold increase in k /K for 2 renders
M
cat
M
1736
asc.wiley-vch.de
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1732 – 1740

FULL PAPERS
Mechanism-Guided Engineering of w-Transaminase
Table 4. Asymmetric synthesis of chiral amines using
OATAW58L in comparison with the wild-type enzyme.
[
a]
Substrate Reaction Conversion
Product
[b]
time [h]
[%]
OATA OATAW58L
(% ee)
1
1
1
1
1
d
f
i
k
7
5
15
10
15
7
91
93
75
93
91
(S)-3d (>99)
(S)-3f (>99)
(S)-3i (>99)
(S)-3k (>99)
(S)-3l (>99)
26
13
14
30
l
[
a]
Conversion was calculated by analyzing (S)-amine pro-
duced from ketone. Reaction conditions: same as those in
Figure 3. Reactions were carried out at 300 torr vacuum.
The ee values represent the ones obtained with
[b]
OATAW58L
.
Figure 3. Asymmetric synthesis of (S)-3a using OATAW58L in
sure drove the conversion to reach 94% at 7 h with
comparison with the wild-type enzyme. Reaction conditions:
À1
1
a (50 mM), 2 (500 mM), w-TA (5 mgmL ), PLP (0.1 mM), >99% ee of (S)-3a using OATAW58L. In contrast, the
DMSO (15% v/v) and phosphate buffer (50 mM, pH 7.0) in engineered ARTA afforded only 19% conversion at
a 2 mL reaction mixture.
7 h under the same reaction conditions (Supporting
Information, Figure S7).
To prove the synthetic applicability of OATAW58L to
OATAW58L highly promising for scalable amination of the preparation of diverse chiral amines, we per-
ketones using 2 as an amino donor.
formed asymmetric aminations of five additional ke-
tones (Table 4). Consistent with the activity improve-
ments shown in Table 2, OATAW58L afforded much
higher conversions [i.e., >90%, except for 1i, within
Asymmetric Synthesis of Chiral Amines
15 h with >99% ee of the produced (S)-amines] than
To evaluate how much the W58L mutation improved OATA did.
the practical utility of OATA, time-course monitoring
One of the ketones (i.e., 1f) was subjected to
of the asymmetric synthesis of (S)-3a from 1a and 2 a preparative-scale synthesis in a 100-mL reaction
was performed at atmospheric pressure in the pres- mixture composed of 1f (1.5 g, 10 mmol), 2 (8.6 mL,
ence of DMSO (15% v/v) to solubilize the ketone 100 mmol), DMSO (15 mL) and OATAW58L (5 mmol)
substrate (Figure 3). As expected, OATAW58L afforded at 460 torr. Conversion reached 92% at 18 h with
a dramatic improvement in the amination of 1a with- >99% ee of (S)-3f. Product isolation using cation-ex-
out any loss in stereoselectivity [i.e., ee of produced change column chromatography and structural char-
(
S)-3a >99%]. No activity loss was observed for both acterizations of the resulting (S)-3f were performed
OATA and OATAW58L throughout the reaction, indi- (see the Supporting Information), leading to recovery
cating that the W58L mutation did not affect the of pure (S)-3f (1.22 g, 81.3% isolated yield, >99%
enzyme stability (Supporting Information, Figure S5). ee).
In contrast, the engineered ARTA showed a signifi-
cant decrease in the enzyme stability compared with
the wild-type ARTA under the same reaction condi- Conclusions
tions (Supporting Information, Figure S6).
Despite the high operational stability of OATAW58L
,
The motivation of this study was conceived with the
the reaction progress at atmospheric pressure reached notion that the metabolic conversion of ketones to
a plateau at 3 h [i.e., 39% conversion of 1a to (S)-3a] primary amines would be irrelevant to enhancing the
and then further increase in the conversion was only biological fitness of microorganisms that have evolved
2% until 7 h (Figure 3). We ascribed this to an unfav- w-TAs. Therefore, we posited that endowing naturally
orable reaction equilibrium. Based on a previous occurring w-TAs with a decent activity for ketones by
study, the equilibrium constant for the 1a:2 reaction redesigning the native active site would not be chemi-
in the absence of DMSO was calculated to be 5.86” cally unachievable but could have been biologically
À2
10
which did not allow the concentration of pro- untried. Indeed, our results clearly show that the sub-
duced (S)-3a to exceed 25.9 mM at the given reaction strate bias of a native w-TA is amenable to engineer-
[15]
conditions. Indeed, an equilibrium shift by remov- ing (i.e., from keto acids to ketones) even by a single
ing a co-product (i.e., acetone) at the reduced pres- point mutation. The W58 residue of OATA, which
Adv. Synth. Catal. 2015, 357, 1732 – 1740
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1737

Sang-Woo Han et al.
FULL PAPERS
serves as a crucial steric barrier to the amination of 48C) and the resulting cell pellet was resuspended in 15 mL
resuspension buffer (50 mM Tris-HCl, pH 7, 50 mM NaCl,
ketones, was found to be well conserved among (S)-
1
0
mM EDTA, 1 mM b-mercaptoethanol, 0.1 mM PMSF and
.5 mM PLP). The cell suspension was then subjected to ul-
selective w-TAs (Supporting Information, Table S4).
Among the fourteen w-TAs used for multiple se-
quence alignment, W58 was perfectly conserved in
eleven w-TAs and was replaced by tyrosine, which
trasonic disruption and centrifuged (13,000”g, 60 min, 48C)
to remove cell debris.
Protein purification was carried out on ¾KTAprime plus
[
16]
also could act as a hydrogen bond donor,
in the
(
GE Healthcare, Piscataway, USA). The cell-free extract
other three w-TAs. This implies that the role of W58
was loaded on an HisTrap HP column (GE Healthcare) and
to the tight interaction with the carboxylate of native the His -tagged w-TA was eluted by an elution buffer
6
keto acid substrates, as observed with aromatic amino (20 mM sodium phosphate, 0.5M sodium chloride, 0.5 mM
[
8]
acid transaminase, is universally played among the PLP, pH 7.4) with a linear gradient of imidazole (0.05–
0
.5M). Removal of imidazole was carried out by a HiTrap
desalting column (GE Healthcare) using an elution buffer
50 mM sodium phosphate, 0.15M sodium chloride and
.5 mM PLP, pH 7). When necessary, the enzyme solution
w-TAs.
(
0
Experimental Section
was concentrated using an ultrafiltration kit (Ultracel-30)
purchased from Millipore Co. (Billerica, USA).
Molar concentrations of the purified w-TAs were deter-
mined by measuring UV absorbance at 280 nm (UV-
Chemicals
Butyrophenone, 4-methyl-2-pentanone, (S)-a-propylbenzyl-
amine, (R)-a-propylbenzylamine, (S)-2-aminohexane, (R)-2-
aminohexane, (S)-2-aminooctane, (R)-2-aminooctane, (S)-1-
cyclopropylethylamine and (R)-1-cyclopropylethylamine
were purchased from Alfa Aesar (Ward Hill, USA). Pyruvic
acid was obtained from Kanto Chemical Co. (Tokyo, Japan).
Isopropylamine was purchased from Junsei Chemical Co.
1
650PC, Shimadzu Co., Japan). Molar extinction coefficients
of the homodimeric w-TAs were obtained by protein extinc-
tion coefficient calculator (http://www. biomol.net/en/tools/
proteinextinction.htm) and were used for concentration de-
termination (78,076M cm for the wild-type OATA and
its variant carrying a single point mutation of L57A, F86A,
V154A, I261A or T324A substitution; 67,076M cm for
OATA variants carrying a single point mutation of W58A,
W58M, W58V, W58I or W58L substitution; 75,095M cm
for OATA variants carrying a single point mutation of
À1
À1
À1
À1
(
Tokyo, Japan). l-Alanine was purchased from Acros Or-
ganics Co. (Geel, Belgium). Dimethyl sulfoxide (DMSO)
and n-hexane were purchased from Duksan Pure Chemicals
Co. (Ansan, Korea). All other chemicals were purchased
from Sigma Aldrich Co. (St. Louis, USA). Materials used
for the preparation of culture media including yeast extract,
tryptone and agar were purchased from Difco. (Spark,
USA).
À1
À1
À1
À1
Y20A or Y151A substitution; 101,885M cm for ARTA ;
1
À1
À1
02,135M cm for the engineered ARTA).
Measurement of Enzyme Activity
All enzyme assays were carried out at 378C and pH 7
(50 mM phosphate buffer). Standard substrate conditions
for activity assay were 10 mM (S)-3a and 10 mM pyruvate.
Typical reaction volume was 50 mL and the enzyme reaction
was stopped after 30 min by adding 10 mL HCl solution
Site-Directed Mutagenesis of w-TA
Variants of OATA carrying a single point mutation were
created by a QuikChange Lightning site-directed mutagene-
sis kit (Agilent Technologies Co.) according to the instruc-
tion manual. Mutagenesis primers shown in the Supporting
Information, Table S5 were designed using a primer design
program (http://www.agilent.com). The template used for
the mutagenesis PCR was pET28-OATA which was previ-
(
5N). Acetophenone produced from the reactions was ana-
lyzed by HPLC. For initial rate measurements, reaction con-
versions were limited less than 10%.
In the reactions using acetophenone:l-alanine and (S)-
3
a:pyruvate substrate pairs to compare activities of OATA
[
12]
ously constructed.
by DNA sequencing.
Intended mutagenesis was confirmed
W58AMVIL, initial rate measurements were carried out by
HPLC analysis of produced (S)-3a and acetophenone, re-
spectively. All the substrate concentrations used in both re-
actions were 10 mM. The initial rate measurements were
carried out separately three times.
Preparation of Purified w-TA
Overexpression of His -tagged w-TAs was carried out as de-
6
To determine activities of OATA and OATAW58L toward
various ketones, a reaction mixture containing 10 mM
ketone, 10 mM isopropylamine, 15% v/v DMSO and 50 mM
phosphate buffer (pH 7) was used. To determine initial rates
from three independent measurements, produced amines
were analyzed by chiral HPLC after derivatization using
Marfeyꢁs reagent.
[17]
scribed elsewhere with minor modifications.
OATA, ARTA and the engineered ARTA, we used expres-
To obtain
sion vectors harboring the w-TA genes that were construct-
[12,15]
ed in the previous studies.
To obtain variants of OATA,
expression vectors constructed in this study were used. Es-
cherichia coli BL21(DE3) cells carrying the expression vec-
tors [i.e., pET28a(+)] harboring the w-TA gene were culti-
vated in LB medium (typically 1 L) containing 50 mgmL
kanamycin at 378C. IPTG was added to the culture broth
À1
Molecular Modeling
(
final concentration=0.1 mM) around 0.4 OD600 to induce
protein expression and the cells were allowed to grow for
0 h. The culture broth was centrifuged (10,000”g, 10 min,
Molecular modeling was performed with the Discovery
Studio package (version 3.5.0, Accelrys, USA) using the
CHARMm force field. A homology model of OATA was
1
1
738
asc.wiley-vch.de
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1732 – 1740

FULL PAPERS
Mechanism-Guided Engineering of w-Transaminase
constructed using the Modeler module (version 9.8). X-ray
phate buffer (pH 7) containing 15% v/v DMSO. For an
equilibrium shift by removing acetone, the reaction mixture
was incubated at 300 torr vacuum. Aliquots of the reaction
mixture (50 mL) were taken at predetermined reaction times
and mixed with 10 mL HCl solution (5N) to stop the reac-
tion. The reaction mixtures were subjected to HPLC analy-
sis for measurement of conversion and enantiomeric excess.
structures of the four (S)-selective w-TAs from Paracoccus
[18]
denitrificans (PDB ID: 4GRX), Chromobacterium viola-
[
19]
[20]
ceum (4A6T), Mesorhizobium loti (3GJU) and Rhodo-
[20]
bacter sphaeroides (3I5T)
were used as templates. The
four w-TAs assume a homodimeric structure where both
active-site arginines form inward conformations (i.e., point-
ing away from the solvent side), except for 4GRX where
one subunit harbors an outward-pointing active-site argi-
nine. The outward-pointing arginine of 4GRX was set to be
conserved in the homology model, leading to a dimeric
structure of OATA where each subunit has a different con-
formation of the active-site arginine (i.e., an inward-pointing
arginine in one subunit and an outward-pointing arginine in
the other subunit). To construct a holoenzyme structure, the
PLP moiety was copied from 4A6T. Sequence alignment for
the homology modeling was carried out using BLOSUM62
as a scoring matrix with a default setting (À900 gap open
penalty; À50 gap extension penalty). Ramachandran phi-psi
analysis showed that only 0.4% non-glycine residues lay in
the disallowed region, indicative of the validity of the ho-
mology model.
Preparative-Scale Synthesis and Isolation of (S)-3f
Preparative-scale synthesis of (S)-3f was carried out at 378C
under magnetic stirring in a round-bottom flask charged
with 100 mL reaction mixture containing 1f (1.5 g,
1
0 mmol), 2 (8.6 mL, 100 mmol), PLP (2.7 mg), DMSO
(
(
15 mL), OATAW58L (5 mmol) and potassium phosphate
50 mM, pH 7.0). The reaction mixture was incubated at
3
00 torr vacuum with a condenser running at 158C. When
the conversion of (S)-3f reached 92%, the reaction mixture
was subjected to product isolation.
The pH of the reaction mixture was adjusted to 1.0 by
adding 5N HCl for protein precipitation and then the result-
ing solution was filtered through a glass-fritted filter funnel
to remove the protein precipitate. The filtrate solution was
loaded on a glass column packed with Dowex 50WX8
cation-exchange resin (40 g), followed by washing with
water (200 mL), and then elution was done with 150 mL of
Active site models of OATA variants were constructed by
amino acid substitution on the homology model of the wild-
type OATA, followed by energy minimization (2,000 steps;
dielectric constant=4) of the mutation site as well as the
proximal residues within 3 Š from W58 (i.e., M54, S55, L57,
S59, G229, V233, I261, R266, I380 and C426) until the RMS
1
0% ammonia solution. The elution fractions were extracted
with n-hexane (3”300 mL). The organic extractants were
pooled and evaporated at 508C and 0.25 bar.
À1
gradient reached 0.1 kcalmol Š.
Docking simulations were accomplished using the
CDOCKER module under a default setting (2,000 steps at HPLC Analysis
7
00 K for a heating step; 5,000 steps at 300 K for a cooling
All the HPLC analyses were performed on a Waters HPLC
system (Milford, USA). Analysis of acetophenone was per-
formed using a Sunfire C18 column (Waters Co.) with iso-
cratic elution of 60% methanol/40% water/0.1% trifluoro-
step; 8 Š grid extension) within the active site defined by
the Binding-Site module. The active site used for the dock-
ing simulation with l-alanine or (S)-3a was the one placed
in the subunit harboring the inward- or outward-pointing ar-
ginine, respectively. The most stable docking pose showing
a spatial orientation of the substrate capable of subsequent
catalytic conversion was chosen and then was subjected to
stepwise chemical modifications to generate an external ald-
imine and a quinonoid structure, each modification followed
by energy minimization of K287, PLP and the substrate
moiety (2,000 steps; dielectric constant=4).
À1
acetic acid at 1 mLmin . UV detection was done at 254 nm.
Retention time of acetophenone was 3.8 min. Quantitative
chiral analyses of alanine and amines were carried out using
a Crownpak CR(À) column (Daicel Co., Japan) or a Sunfire
C18 column after chiral derivatization with Marfeyꢁs re-
[22]
agent. Details of the HPLC analysis are described in the
Supporting Information.
Structural Characterization of (S)-3f
Kinetic Analysis
The purified (S)-3f was structurally characterized by
H NMR, C NMR, IR and LC/MS. H NMR and C NMR
spectra were recorded on an Avance II FT-NMR spectrome-
ter (Bruker Co.) using tetramethylsilane as a standard. IR
spectrum was obtained by a Vertex70 FTIR spectrometer
To obtain apparent kinetic parameters, a pseudo-one-sub-
strate kinetic model was used under the fixed concentration
1
13
1
13
[
21]
of the cosubstrate. Initial rate data were fitted to a Mi-
chaelis–Menten equation, and the KM and kcat values were
calculated from the slopes and y intercepts of the double-re-
ciprocal plots. Rate constants were determined from three
independent initial rate measurements over the concentra-
tion ranges of substrate and cosubstrate listed in the Sup-
porting Information, Table S6. HPLC analytes used in the
initial rate measurements are also presented in Table S6.
(
Bruker Co.). Mass spectral data were obtained with a LTQ
Orbitrap Mass Spectrometer (electrospray ionization, posi-
tive ion mode, Thermo Fisher Scientific Inc.). Details of the
structural data are provided in the Supporting Information.
Small-Scale Asymmetric Synthesis of Chiral Amines
The reaction volume for the small-scale asymmetric synthe- Acknowledgements
sis was 2 mL and the reaction mixture was incubated at
3
0
78C. Reaction conditions were 50 mM ketone, 500 mM 2,
.1 mM PLP and 5 mgmL purified w-TA in 50 mM phos-
We acknowledge financial support from the Advanced Bio-
mass R&D Center (ABC-2011-0031358) through the Nation-
À1
Adv. Synth. Catal. 2015, 357, 1732 – 1740
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
1739

Sang-Woo Han et al.
FULL PAPERS
al Research Foundation of Korea funded by the Ministry of
Education, Science and Technology.
[9] a) F. G. Mutti, C. S. Fuchs, D. Pressnitz, J. H. Sattler, W.
Kroutil, Adv. Synth. Catal. 2011, 353, 3227–3233; b) M.
Hçhne, S. Kühl, K. Robins, U. T. Bornscheuer, Chem-
BioChem 2008, 9, 363–365; c) S. Schätzle, F. Steffen-
Munsberg, A. Thontowi, M. Hçhne, K. Robins, U. T.
Bornscheuer, Adv. Synth. Catal. 2011, 353, 2439–2445;
d) J. S. Shin, B. G. Kim, Biotechnol. Bioeng. 1999, 65,
206–211; e) M. D. Truppo, J. D. Rozzell, N. J. Turner,
Org. Process Res. Dev. 2010, 14, 234–237; f) P. Tufves-
son, J. S. Jensen, W. Kroutil, J. M. Woodley, Biotechnol.
Bioeng. 2012, 109, 2159–2162; g) K. E. Cassimjee, C.
Branneby, V. Abedi, A. Wells, P. Berglund, Chem.
Commun. 2010, 46, 5569–5571; h) A. P. Green, N. J.
Turner, E. OꢁReilly, Angew. Chem. Int. Ed. 2014, 53,
10714–10717.
References
[
1] a) A. Schmid, J. S. Dordick, B. Hauer, A. Kiener, M.
Wubbolts, B. Witholt, Nature 2001, 409, 258–268; b) S.
Wenda, S. Illner, A. Mell, U. Kragl, Green Chem. 2011,
1
3, 3007–3047; c) C. A. Denard, J. F. Hartwig, H. Zhao,
ACS Catal. 2013, 3, 2856–2864; d) B. M. Nestl, S. C.
Hammer, B. A. Nebel, B. Hauer, Angew. Chem. 2014,
1
26, 3132–3158; Angew. Chem. Int. Ed. 2014, 53, 3070–
3
095.
[
2] P. Tufvesson, J. Lima-Ramos, J. S. Jensen, N. Al-Haque,
[10] a) K. S. Midelfort, R. Kumar, S. Han, M. J. Karmilo-
wicz, K. McConnell, D. K. Gehlhaar, A. Mistry, J. S.
Chang, M. Anderson, A. Villalobos, J. Minshull, S. Go-
vindarajan, J. W. Wong, Protein Eng. Des. Sel. 2013, 26,
25–33; b) K. E. Cassimjee, M. S. Humble, H. Land, V.
Abedi, P. Berglund, Org. Biomol. Chem. 2012, 10,
5466–5470.
[11] E. S. Park, J. Y. Dong, J. S. Shin, Org. Biomol. Chem.
2013, 11, 6929–6933.
[12] E. S. Park, M. Kim, J. S. Shin, Appl. Microbiol. Biotech-
W. Neto, J. M. Woodley, Biotechnol. Bioeng. 2011, 108,
1479–1493.
[
3] a) M. T. Reetz, Angew. Chem. 2011, 123, 144–182;
Angew. Chem. Int. Ed. 2011, 50, 138–174; b) U. T.
Bornscheuer, G. W. Huisman, R. J. Kazlauskas, S. Lutz,
J. C. Moore, K. Robins, Nature 2012, 485, 185–194;
c) N. J. Turner, Nat. Chem. Biol. 2009, 5, 567–573.
4] a) M. S. Malik, E. S. Park, J. S. Shin, Appl. Microbiol.
Biotechnol. 2012, 94, 1163–1171; b) R. C. Simon, N.
Richter, E. Busto, W. Kroutil, ACS Catal. 2013, 4, 129–
[
nol. 2012, 93, 2425–2435.
1
43; c) M. Hçhne, S. Schätzle, H. Jochens, K. Robins,
[13] a) M. D. Toney, Arch. Biochem. Biophys. 2014, 544,
119–127; b) A. C. Eliot, J. F. Kirsch, Annu. Rev. Bio-
chem. 2004, 73, 383–415.
U. T. Bornscheuer, Nat. Chem. Biol. 2010, 6, 807–813.
5] a) T. C. Nugent, Chiral Amine Synthesis: Methods, De-
velopments and Applications, Wiley-VCH, Weinheim,
[
[14] V. N. Malashkevich, J. J. Onuffer, J. F. Kirsch, J. N. Jan-
2
010; b) M. Breuer, K. Ditrich, T. Habicher, B. Hauer,
sonius, Nat. Struct. Biol. 1995, 2, 548–553.
[15] E.-S. Park, M. S. Malik, J.-Y. Dong, J.-S. Shin, Chem-
CatChem 2013, 5, 1734–1738.
M. Kesseler, R. Sturmer, T. Zelinski, Angew. Chem.
2004, 116, 806–843; Angew. Chem. Int. Ed. 2004, 43,
788–824; c) D. Ghislieri, N. J. Turner, Top. Catal. 2014,
57, 284–300.
[16] K. Okada, K. Hirotsu, H. Hayashi, H. Kagamiyama,
Biochemistry 2001, 40, 7453–7463.
[
6] a) D. Koszelewski, K. Tauber, K. Faber, W. Kroutil,
Trends Biotechnol. 2010, 28, 324–332; b) M. Hçhne,
U. T. Bornscheuer, ChemCatChem 2009, 1, 42–51; c) D.
Koszeleiwski, I. Lavandera, D. Clay, G. M. Guebitz, D.
Rozzell, W. Kroutil, Angew. Chem. 2008, 120, 9477–
[17] E. S. Park, J. S. Shin, Adv. Synth. Catal. 2014, 356,
3505–3509.
[18] C. Rausch, A. Lerchner, A. Schiefner, A. Skerra, Pro-
teins 2013, 81, 774–787.
[19] M. S. Humble, K. E. Cassimjee, M. H¼kansson, Y. R.
Kimbung, B. Walse, V. Abedi, H. J. Federsel, P. Ber-
glund, D. T. Logan, FEBS J. 2012, 279, 779–792.
[20] F. Steffen-Munsberg, C. Vickers, A. Thontowi, S. Schät-
zle, T. Tumlirsch, M. S. Humble, H. Land, P. Berglund,
U. T. Bornscheuer, M. Hçhne, ChemCatChem 2013, 5,
150–153.
9
480; Angew. Chem. Int. Ed. 2008, 47, 9337–9340;
d) 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,
G. J. Hughes, Science 2010, 329, 305–309.
[
7] a) J. S. Shin, B. G. Kim, J. Org. Chem. 2002, 67, 2848–
2
853; b) U. Kaulmann, K. Smithies, M. E. B. Smith,
H. C. Hailes, J. M. Ward, Enzyme Microb. Technol.
007, 41, 628–637.
[21] E. S. Park, J. S. Shin, Appl. Environ. Microbiol. 2013,
79, 4141–4144.
[22] a) C. B’Hymer, M. Montes-Bayon, J. A. Caruso, J. Sep.
Sci. 2003, 26, 7–19; b) R. Bhushan, H. Brückner,
Amino Acids 2004, 27, 231–247.
2
[
8] K. Hirotsu, M. Goto, A. Okamoto, I. Miyahara, Chem.
Rec. 2005, 5, 160–172.
1740
asc.wiley-vch.de
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2015, 357, 1732 – 1740