Structural Insight into Enantioselective Inversion of an Alcohol Dehydrogenase Reveals a "polar Gate" in Stereorecognition of Diaryl Ketones

Article abstract of DOI:10.1021/jacs.8b08640

Diaryl ketones are important building blocks for synthesizing pharmaceuticals and are generally regarded as "difficult-to-reduce" ketones due to the large steric hindrance of their two bulky aromatic side chains. Alcohol dehydrogenase from Kluyveromyces polyspora (KpADH) has been identified as a robust biocatalyst due to its high conversion of diaryl ketone substrate (4-chlorophenyl)(pyridine-2-yl)ketone (CPMK) with a moderate R-selectivity of 82% ee. To modulate the stereoselectivity of KpADH, a "polarity scanning" strategy was proposed, in which six key residues inside and at the entrance of the substrate binding pocket were identified. After iterative combinatorial mutagenesis, variants Mu-R2 and Mu-S5 with enhanced (99.2% ee, R) and inverted (97.8% ee, S) stereoselectivity were obtained. The crystal structures of KpADH and two mutants in complex with NADPH were resolved to elucidate the evolution of enantioselective inversion. Based on MD simulation, Mu-R2-CPMK_ProR and Mu-S5-CPMK_ProS were more favorable in the formation of prereaction states. Interestingly, a quadrilateral plane formed by α-carbons of four residues (N136, V161, C237, and G214) was identified at the entrance of the substrate binding pocket of Mu-S5; this plane acts as a "polar gate" for substrates. Due to the discrepancy in charge characteristics between chlorophenyl and pyridine substituents, the pro-S orientation of CPMK is defined when it passes through the "polar gate" in Mu-S5, whereas the similar plane in wild-type is blocked by several aromatic residues. Our result paves the way for engineering stereocomplementary ADH toward bulky diaryl ketones and provides structural insight into the mechanism of stereoselective inversion.

Full text of DOI:10.1021/jacs.8b08640

Article
pubs.acs.org/JACS
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
Structural Insight into Enantioselective Inversion of an Alcohol
Dehydrogenase Reveals a “Polar Gate” in Stereorecognition of
Diaryl Ketones
Jieyu Zhou,† Yue Wang,† Guochao Xu,† Lian Wu,‡ Ruizhi Han,† Ulrich Schwaneberg,§ Yijian Rao,†
Yi-Lei Zhao,⊥ Jiahai Zhou,*,‡ and Ye Ni*,†
†Key Laboratory of Industrial Biotechnology, Ministry of Education, School of Biotechnology, Jiangnan University, Wuxi, 214122
Jiangsu, China
‡State Key Laboratory of Bio-organic and Natural Products Chemistry, Center for Excellence in Molecular Synthesis, Shanghai
Institute of Organic Chemistry, Chinese Academy of Sciences, Shanghai 200032, China
§Institute of Biotechnology, RWTH Aachen University, Worringerweg 3, 52074 Aachen, Germany
⊥State Key Laboratory of Microbial Metabolism, School of Life Sciences and Biotechnology, Shanghai Jiao Tong University,
Shanghai 200240, China
S
* Supporting Information
ABSTRACT: Diaryl ketones are important building blocks for synthesizing
pharmaceuticals and are generally regarded as “difficult-to-reduce” ketones
due to the large steric hindrance of their two bulky aromatic side chains.
Alcohol dehydrogenase from Kluyveromyces polyspora (KpADH) has been
identified as a robust biocatalyst due to its high conversion of diaryl ketone
substrate (4-chlorophenyl)(pyridine-2-yl)ketone (CPMK) with a moderate
R-selectivity of 82% ee. To modulate the stereoselectivity of KpADH, a
“polarity scanning” strategy was proposed, in which six key residues inside
and at the entrance of the substrate binding pocket were identified. After
iterative combinatorial mutagenesis, variants Mu-R2 and Mu-S5 with
enhanced (99.2% ee, R) and inverted (97.8% ee, S) stereoselectivity were
obtained. The crystal structures of KpADH and two mutants in complex
with NADPH were resolved to elucidate the evolution of enantioselective
inversion. Based on MD simulation, Mu-R2−CPMKProR and Mu-S5−CPMKProS were more favorable in the formation of
prereaction states. Interestingly, a quadrilateral plane formed by α-carbons of four residues (N136, V161, C237, and G214) was
identified at the entrance of the substrate binding pocket of Mu-S5; this plane acts as a “polar gate” for substrates. Due to the
discrepancy in charge characteristics between chlorophenyl and pyridine substituents, the pro-S orientation of CPMK is defined
when it passes through the “polar gate” in Mu-S5, whereas the similar plane in wild-type is blocked by several aromatic residues.
Our result paves the way for engineering stereocomplementary ADH toward bulky diaryl ketones and provides structural insight
into the mechanism of stereoselective inversion.
2-yl)ketone (CPMK) is an alternative method for obtaining
chiral intermediates of bepotastine. Due to the steric hindrance
of two bulky aromatic substitutes, asymmetric reduction of
diaryl ketones is often difficult for chemists and biologists.6−8
Recently, a few studies on chemo- or bioreduction of diaryl
ketones have been reported. Asymmetric transfer hydro-
genation of CPMK in 40.8% ee (R)9 and 60.6% ee (S)10 has
been achieved using ruthenium complex catalysts. Sporobolo-
myces salmonicolor carbonyl reductase (SsCR) is the first
heterologously expressed diaryl ketone dehydrogenase.11 SsCR
and its mutant Q245P catalyzed the enantioselective reduction
of (4-chlorophenyl)(phenyl)methanone at 78% ee (R) and
INTRODUCTION
■
Diaryl alcohols are important building blocks in the synthesis
of pharmaceuticals.1,2 Among them, (S)-(4-chlorophenyl)-
(pyridin-2-yl) methanol [(S)-CPMA] is an important inter-
mediate for synthesizing the antiallergy drug bepotastine
(formerly designated as bepotastine besilate, (+)-(S)-4-{4-
[(4-chlorophenyl)(2-pyridyl)methoxy]piperidino}butyric acid
monobenzenesulfonate).3,4 Bepotastine is a second-generation
histamine H1 receptor antagonist possessing high binding
selectivity and no side effects such as sedation and
cardiotoxicity.5 The synthetic method to produce bepotastine
is patent-protected, and requires a resolving agent, (2R,3R)-2-
hydroxy-3-(4-methoxyphenyl)-3-(2-nitro-5-chlorophenylthio)-
propionic acid, which is expensive and not readily available.
However, asymmetric reduction of (4-chlorophenyl)(pyridine-
Received: August 22, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Residues chosen for polar scanning in the homology model of KpADH. (A) Residues within the substrate binding pocket. (B) Residues at
the entrance of the substrate binding pocket. Residues are automatically assigned different colors.
76% ee (S), respectively.12 Cryptococcus sp. whole-cell was
employed as a catalyst for producing S-CPMA with excellent
99% ee in hydrophilic ionic liquid, whereas only 10 mM CPMK
could be tolerated.13 Apparently, the poor substrate tolerance
and undesirable stereoselectivity of reported catalysts could not
meet the requirements for pharmaceutical intermediates.
Alcohol dehydrogenases (ADHs, EC 1.1.1.X, X = 1 or 2) are
nicotinamide adenine dinucleotide (phosphate) [NAD(P)+]-
dependent enzymes that catalyze the reduction of carbonyl
compounds to their corresponding alcohols. The bioreduction
of ketones is generally regarded as a green method that
provides efficient access to optically active alcohols.14
However, the application of ADHs is often hindered by their
insufficient enantioselectivity and low activity.15−17 According
to Prelog’s rule, hydrides are delivered from the Re- or Si-face
in the asymmetric reduction of prochiral ketones to generate
(S)- or (R)-configured alcohol products.18 Engineering
enzymatic enantioselectivity is often challenging due to the
complicated stereorecognition mechanism.19−22 Optimized
codon usage in directed evolution experiments has emerged
as a powerful means of engineering enantioselectivity.23 For
instance, using a triple code saturation mutagenesis, Sun and
co-workers evolved highly R- and S-selective variants of
TbSADH toward difficult-to-reduce ketones, such as tetrahy-
drofuran-3-one.24 However, in most cases, semirational design
based on crystal structure or a homology model is more
suitable for enantioselectivity engineering due to the lack of
high-throughput screening methods for chirality.25 There have
been several reports on engineering enantioselectivity of
alcohol dehydrogenases, mainly focusing on aromatic ketones
with only one aryl substitute. For instance, Patel and co-
workers obtained five mutants with improved ee toward 4-
phenyl-2-butanone from a comprehensive mutant library.26 Li
and co-workers found that aromatic and halogen substitutes
played critical roles in stereoselectivity of SDRs (short-chain
dehydrogenases/reductases), which was inverted from Prelog
to anti-Prelog.21
KpADH for inverted stereoselectivity from R to S toward diaryl
ketone substrates. Key residues lining the substrate binding
pocket were identified based on polarity scanning. Enantio-
complementary variants were obtained after iterative combi-
natorial mutagenesis, which could be applied in the
preparation of a variety of S- and R-diaryl alcohols. Crystal
structures and computational data explained the stereoselective
inversion of KpADH.
RESULTS AND DISCUSSION
■
Identification of Core Amino Acids. In our previous
study, variant S237A was identified by random mutagenesis;
this variant exhibited improved activity and enantioselectivity
(96.1%, R) compared with WT (wild-type) KpADH.27 This
result suggests that protein engineering could be used to
modulate its enantioselectivity. The homology model of
KpADH was constructed based on the crystal structure of
Saccharomyces cerevisiae NADPH-dependent methylglyoxal
reductase GRE2 (PDB 4PVC) (EC 1.1.1.283).28 KpADH
possesses a superlarge substrate binding cavity with a
calculated volume of 214.8 Å3, while the solvent-excluded
volume of CPMK is 159.8 Å3 (computed by Accelrys
Discovery Studio software)29 based on its homology model.
After eliminating the catalytic triad (S126−Y164−K168), 11
residues (V84, Y127, A128, M131, C165, P195, S196, F197,
T215, S237, and Q238) within the binding site pocket (Figure
1A) and five residues (F161, F86, Q136, E214, and I218) at
the entrance of the pocket (Figure 1B) were selected for
further mutagenesis study.
Initially, we attempted to identify key residues responsible
for enantioselective recognition. Screening semisaturation/
saturation mutant libraries of 16 selected residues is, however,
an enormous task even when a simplified NDT codon is
used.30 Based on the smallest alphabets proposed by Sun et
al.,31 saturation mutagenesis with single amino acid alphabets
was attempted to reshape the binding pocket. All 16 residues
were grouped based on their positions in the binding pocket,
then each group of residues was simultaneously mutated to
valine or alanine. However, most variants were inactivated after
the mutation. This unsuccessful attempt indicates that the
stereoselectivity of KpADH cannot be altered by simply
reshaping the substrate binding pocket due to the large volume
of its substrate pocket. A number of reports have demonstrated
that “stereomatching” strategies such as alanine scanning are
effective in identifying residues responsible for catalytic
selectivity,32,33 especially when the substrate contains sub-
Recently, an alcohol dehydrogenase from Kluyveromyces
polyspora (KpADH) was discovered by genome mining.
KpADH exhibited a moderate ee of 82.5% (R), with excellent
substrate tolerance (>100 mM) toward diaryl ketones. A
carbonyl group-dependent colorimetric high-throughput
screening method was developed, and a beneficial variant
S237A with improved R-selectivity of 96.1% was identified
from a random mutagenesis library, indicating the evolvability
of enantioselectivity.27 Herein, we attempted to engineer
B
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
stituents that differ substantially in steric hindrance.34 In
CPMK, chlorobenzene and pyridine groups are noncoplanar
and could swing slightly around a pro-chiral carbon axis. As
calculated by Discovery studio,29 volumes of chlorobenzene
and pyridine rings are estimated to be 111.2 Å3 and 104.8 Å3,
respectively (Figure S1). The small difference in volume
suggests that it is difficult to impact enantioselectivity by
altering steric hindrance of residues alone.
Table 1. Saturation Mutagenesis of Six Key Residues for
Evolving Enantioselectivity of KpADH
decreased/inverted variants
ee (%)
enhanced variants
ee (%)
Q136N
Q136S
S196G
F161V
E214G
E214S
E214N
T215V
S237C
S237N
55.4
68.6
47.5
72.6
−3.29
11.8
33.4
44.3
43.8
73.4
S196V
S196W
S196P
E214V
E214I
T215S
T215A
S237A
S237Y
S237W
98.2
96.1
93.5
92.8
93.0
91.6
93.3
96.1
93.9
94.3
Polarity Scanning of 16 Core Residues. In addition to
steric effects, intermolecular forces (including ion-induced
dipole force, ion−dipole force, hydrogen bonding, and van der
Waals force) also play a key role in molecular recognition
through ligand−protein binding.35 Formation of intermolecu-
lar forces is closely related to the polarity of two molecules.
The positive end of a polar molecule will attract the negative
end of another molecule and thereby affect their position; thus
the orientation of polar CPMK could be affected by its
surrounding amino acids. Additionally, polarity at the entrance
of the binding pocket could influence enzymatic selectivity or
activity.36 Therefore, a “polarity scanning” strategy was
proposed to use typical polar/nonpolar amino acids as a
sieve to eliminate residues with little effect on catalytic
stereoselectivity by a single mutation. It is presumed that the
polarity changes in the binding pocket could adjust the
intermolecular force and thus affect the orientation of CPMK.
According to the hydropathy index of 20 amino acids,37
arginine (−4.5) is a basic amino acid possessing the strongest
polar side chain, followed by basic lysine (−3.9), and neutral
asparagine (−3.5). Since the polarity of acidic and basic amino
acids may be reduced or even eliminated by the formation of
salt bonds, polar and neutral asparagine (−3.5) with a short
side chain was selected as the “polar sieve”. Simultaneously,
valine (4.2) was considered an appropriate “nonpolar sieve”
after comprehensive evaluation of its hydrophobicity, mobility,
and steric hindrance.
Sixteen residues were subjected to “polarity scanning” by
single mutation, and the effect of polarity change on
stereoselectivity was determined using CPMK as a model
substrate. The ee values of more than 90% or less than 75%
were defined as the thresholds to select residues sensitive to
polarity change. Six residues were obtained including Q136,
F161, S196, E214, T215, and S237 (Figure S2). When these
residues were mutated to polar Asn, decreased ee values were
observed at Q136N (55.4%), E214N (33.4%), and S237N
(73.4%); the same trend was evident when F161 and T215
were mutated into nonpolar Val. In addition, two markedly
improved R-selective mutants S196V (98.2% ee) and E214V
(92.8% ee) were identified. It is interesting that the
stereoselectivity of E214V (92.8%, R) and T215V (43.8%,
R) shows the opposite trend, even though the “inside site” 215
and the “entrance site” 214 are adjacent. These six residues
exhibited a prominent influence on stereoselectivity and were
thus subjected to further saturation and iteratively combina-
torial mutagenesis.
(124.5 Å3),38 both variants showed a decrease in R-selectivity.
A similar phenomenon was observed at sites 214 and 237.
Variants E214G (3.3%, S), E214N (33.4%, R), and E214S
(11.8%, R) exhibited a similar tendency in reversed stereo-
selectivity, whereas the stereoselectivity of S237A, S237T, and
S237W was higher than 90%. T215S (91.6%, R), which was
different from E214S (11.8%, R), was selected due to its
enhanced R-selectivity; this was similar to the opposite trend
observed with E214V and T215V. Residue S196 was picked for
further saturation mutagenesis due to the excellent enantiose-
lectivity of S196V (98.2% ee, R). Mutation of S196 to other
nonpolar amino acids such as proline (93.5% ee, R) and
tryptophan (96.1% ee, R) also resulted in improved ee, and
reduced enantioselectivity was observed only for S196G
(47.5% ee, R). For F161, all variants showed no significant
improvement compared with WT, and F161V displayed the
lowest R-selectivity of 72.6% (Figure S3).
Iterative Combinatorial Mutagenesis for Evolving
Stereoselectivity. Iterative combinatorial mutagenesis
(ICM) libraries were constructed using a data-based
“Progressive Optimal Combination” (dPOC) approach. First,
double-mutation libraries were constructed based on the 10
decreased/inverted single mutations listed in Table 1. For 95%
coverage, theoretically only 111 transformants need to be
screened for stereoselectivity inversion (Figure S4A). The best
mutant, E214G/S237C (Mu-S2), led to pronounced inverted
stereoselectivity in favor of the S-configuration of 51.8% ee.
Other variants including S196G/E214S (32.3% ee, S), Q136N/
E214G (34.2% ee, S), E214S/S237C (34.4% ee, S), and
F161V/E214S (31.2% ee, S) were also identified, indicating
positive synergistic effects among the selected residues (Table
S1). Next, double mutants that exhibited S-preference were
selected to construct triple mutagenesis libraries (Figure S4B).
More than half of the mutants containing E214G or E214S
reached over 80% ee (S), whereas mutants without these two
mutations achieved merely 60% ee (S). Our results suggest that
residue 214 plays an important role in the stereoinversion of
KpADH, and synergistic effects of other residues are also
indispensable. In progressive screening of quadruple muta-
genesis libraries, Q136N/S196G/E214G/S237C (Mu-S4a)
(92.2% ee, S) and F161V/S196G/E214G/S237C (Mu-S4b)
(94.5% ee, S) were obtained (Figure S4C). After four rounds of
ICM, Q136N/F161V/S196G/E214G/S237C (Mu-S5) with
97.8% ee (S) was attained. The optical purity of the S-CPMA
product reached above 99% ee after recrystallization.39
Saturation Mutagenesis of Six Key Residues. To fully
investigate the effect of Q136, F161, S196, E214, T215, and
S237, saturation mutagenesis (SM) was performed using the
NNK codon. Mutants that exhibited over 90% or less than
75% ee were selected for further iterative combinations (Table
1). For residue 136, over half of its SM library exhibited 75−
80% ee. Among the others, Q136N (55.4%, R) and Q136S
(68.6%, R) with decreased ee were selected. Despite the
appreciable difference in the volume of Ser (99.1 Å3) and Asn
Combinatorial libraries were also constructed based on the
ten single mutants with enhanced stereoselectivity (Table 1).
Fortunately, several excellent R-selective variants were
C
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Overall crystal structures of WT (5Z2X), Mu-R2 (5ZED), and Mu-S5 (5ZEC). (A) 3D Structure alignment of WT (green), Mu-R2
(purple blue), and Mu-S5 (pink). Images B, C, and D represent detailed views of circled areas in image A: main interactions between loop 82−96
and site 214 in WT, Mu-R2, and Mu-S5, respectively. Images E, F, and G represent the solvent surface profiles of WT, Mu-R2, and Mu-S5,
respectively.
identified in the first round of double mutagenesis (Figure
S4D). Compared with moderate R-selectivity of WT (82.5%),
the best variant, E214V/T215S (Mu-R2), exhibited over 99%
ee (R) and was also selected for further investigation. It is
noteworthy that E214V and T215S are not the best among
single mutants with enhanced R-selectivity (Table 1). The
excellent stereoselectivity of Mu-R2 could therefore be
attributed to a synergistic effect.
to be 6.2 Å (Figure 2B−D). In WT, the carboxylic acid group
of E214 could form hydrogen bonds with one of the loop
residues among N85, F86, and G87, resulting in an attractive
force on loop 82−96. This result suggests that the crystal of
WT KpADH was obtained under closed-binding pocket status
as the substrate could not enter the catalytic center when the
loop was bound with E214. In Mu-R2 and Mu-S5, neither
E214V nor E214G could serve as a hydrogen donor to bind
the loop (Figure 2C,D). As shown in Figure 2E−G, the surface
of the enzyme is hydrophilic, while the catalytic center is a
strongly hydrophobic region. Variants Mu-S5 and Mu-R2
(Figures 2F,G) possess a semiopened substrate binding pocket,
which renders them more easily able to accept bulky substrates
with excellent affinity and catalytic efficiency.
Molecular Dynamic Simulations of Prereaction State.
The prereaction states observed in computational simulations
could reflect early stereoselective recognition of enzymes.41
Based on the reaction mechanism of short chain dehydrogen-
ase, Ser126 stabilizes the substrate, Tyr164 functions as the
catalytic base, and the pKa of the hydroxyl group of Tyr164
could be lowered by Lys168. The reaction is initiated by
proton transfer from Tyr164−OH to the carbonyl oxygen
atom of CPMK. Then, the hydrogen atom of NADPH is
transferred to the carbonyl carbon, and the carbonyl group of
CPMK is converted into a hydroxyl group.42 Therefore, there
were two distances that could be used to evaluate whether the
enzyme−substrate complex could be developed into the
prereaction state. One is defined by the distance between the
carbonyl oxygen (O14) of CPMK and the H atom of Tyr164−
OH, designated as d(O14CPMK−H9Y164), indicating the
formation of the hydrogen bond and the protonation of
CPMK. The other is defined by the distance between the
carbonyl carbon (C7) of CPMK and the hydrogen atom (H4)
at C4 of NADPH, designated as d(C7CPMK−H4NPH),
indicating the process of nucleophilic attack (Figure S5).
d(O14CPMK−H9Y164) ≤ 3.4 Å and d(C7CPMK−H4NPH) ≤ 4.5 Å
could be used to represent formation of the hydrogen bond
Crystal Structures of WT KpADH, Mu-R2, and Mu-S5.
To probe the molecular mechanism of stereoselective
inversion, we crystallized WT, Mu-R2, and Mu-S5 of
KpADH. After screening 1300 crystallization conditions,
crystals were obtained only when NADPH was added. The
structures of three protein−NADPH complexes were solved by
molecular replacement using the protein structure of Gre2
(PDB 4PVC)28 as a model and refined to resolutions of 1.98,
2.20, and 1.78 Å. The asymmetric unit of each variant consists
of two monomers in the P61 space group (Table S2).The
three-dimensional (3D) structure of KpADH displayed high
similarity with SsCR (PDB 1ZZE)11 with an RMSD of 1.67 Å
on main-chain atoms, despite their low sequence identity of
27%. The crystal structure of KpADH represents the first
resolved structure of alcohol dehydrogenase with high catalytic
efficiency and enantioselectivity toward diaryl ketones. The
overall folding of KpADH is consistent with structures of other
members of the SDR family.11,25i,40 KpADH has two distinct
domains. One is an NADPH binding domain consisting of a
standard Rossmann motif of a large twisted β-sheet surrounded
by seven α-helices, and the other is a variable C-terminal
domain responsible for substrate binding. The NADPH
molecule fits the electron density map perfectly in all
protein−NADPH complexes.
Based on structure alignment, a significant difference in loop
82−96 of WT and its variants was noticed (Figure 2A).
Compared with WT, a dislocation of the loop in Mu-R2 and
Mu-S5 is represented by the position of an α-carbon at the F86
site, and the offset distance after the mutation was determined
D
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. Docking poses of CPMK in WT (A), Mu-R2 (B), and Mu-S5(C). Interactions between CPMKProR and Mu-R2 (D) and CPMKProS and
Mu-S5 (E) in prereaction states. Conformation maps of Mu-R2 and Mu-S5 in prereaction state simulations (F).
Figure 4. Typical catalytic conformations of (A) WT-CPMKProR and (B) Mu-S5-CPMKProS based on MD analysis; green sticks represent aromatic
amino acids around CPMK; arrows indicate the entry direction of CPMK; the gray plane in part A represents the entrance blocked by aromatic
amino acids of WT KpADH; the blue and red planes in (B) represent the polar entrance of Mu-S5; purple sticks represent NADPH, and black
dashed lines indicate the direction of hydrogen attack; the mutation sites are marked in red.
and the process of hydride transfer, respectively.43 Therefore,
conformations meeting both constraints could be regarded as a
sign of prereaction states. Additionally, the substrate in a
prereaction state of the KpADH−CPMK complex could be
defined as CPMKProS and CPMKProR according to the
configuration of final alcohol products.
CPMKProR. For Mu-S5, the distance between H4NPH and
C7CPMK is too large for the nucleophilic reaction to proceed.
For Mu-R2, it is difficult to form a stable hydrogen bond
between Tyr164−OH and CPMK in the enzyme−substrate
complex systems (Figure 3F). Remarkably, the proportions of
“catalytic” conformations with both d(O14CPMK−H9Y164) ≤
3.4 Å and d(C7CPMK−H4NPH) ≤ 4.5 Å were 4.56% in Mu-R2−
CPMKProR, more than eight times that observed in Mu-S5−
CPMKProR (0.52%). Similarly, the “catalytic” conformations in
Mu-S5−CPMKProS were 7.64%, approximately four times that
observed in Mu-R2−CPMKProS (1.84%). The above results
demonstrate that Mu-R2−CPMKProR and Mu-S5−CPMKProS
are more favorable in the formation of prereaction states to
give corresponding (R)- and (S)-CPMA.
The structural characteristics of prereaction states of Mu-
R2−CPMKProR and Mu-S5−CPMKProS were further studied.
Major interactions of both systems were calculated as shown in
Figure 3D,E, respectively. In Mu-R2−CPMKProR, the chlor-
obenzene ring of CPMK was mainly maintained through a π−
alkyl hydrophobic interaction with A128, F161, Y127, and
F197, while a similar π−alkyl hydrophobic interaction was
formed between E214V and the pyridine group. In Mu-S5−
CPMKProR, none of the five mutation sites had a direct
interaction with the substrate (except van der Waals force),
demonstrating that polarity change at these sites plays
Docking studies using CDOCKER (Discovery Studio 4.0)
were performed with WT, Mu-R2, and Mu-S5 with CPMK as a
substrate. For WT, the energy of loop 82−96 needs to be
lowered by loop refinement tools so that CPMK can be docked
into the active site. Docking poses with the lowest CDOCKER
energy (Table S3) were retained, and they were all in line with
the experimentally observed stereoselectivity (Figure 3A−C).
One interesting finding is that the substrate-binding pockets of
Mu-R2 and Mu-S5 were more open than that of WT, whereas
two variants showed higher stereoselectivity. A comparison of
Figure 3, parts A and B, demonstrates that the polarity of WT
and Mu-R2 varies greatly at the entrance. A less polar entrance
of Mu-R2 is favorable for the preferential entry of
chlorobenzene rings. Molecular dynamic simulations were
then performed based on the docking studies. The orientation
of CPMKProS was observed in the stable prereaction states of
Mu-S5, while CPMKProR was observed in the stable prereaction
states of Mu-R2 (Figure 3F). However, it is difficult to sustain
a stable prereaction state of Mu-R2−CPMKProS or Mu-S5−
E
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Table 2. Kinetic Parameters of WT KpADH and Its Variants
mutational sites
Km (mM) Vmax (μmol·min−1·mg−1
)
Kcat (s−1
)
K
cat/Km (s−1·mM−1
)
ee (%)
config (R/S)
WT
none
0.410
0.160
0.520
0.204
0.922
1.08
0.632
0.861
0.840
1.82
0.720
1.12
0.574
0.973
0.702
17.9
17.2
11.1
41.3
8.78
30.9
9.32
10.2
19.8
27.0
21.3
25.2
17.5
24.8
21.3
11.8
11.4
7.32
27.3
5.81
22.6
6.16
6.76
15.8
21.6
17.0
20.1
11.7
16.5
14.2
28.9
71.5
14.1
134
82.5
55.4
3.29
42.9
47.5
72.6
51.8
88.0
79.2
92.2
94.5
97.8
92.8
91.6
99.2
R
R
S
R
R
R
S
S
S
S
S
Q136N
E214G
S237C
S196G
F161V
Mu-S2
Mu-S3a
Mu-S3b
Mu-S4a
Mu-S4b
Mu-S5
E214V
T215S
Mu-R2
Q136N
E214G
S237C
S196G
F161V
6.30
20.9
9.74
7.85
18.8
11.8
23.6
17.9
14.2
16.9
20.3
E214G/S237C
Q136N/E214G/S237C
Q136N/S196G/S237C
Q136N/S196G/E214G/S237C
F161V/S196G/E214G/S237C
Q136N/F161V/S196G/E214G/S237C
E214V
S
R
R
R
T215S
E214V/T215S
Figure 5. Enantioselectivity of WT KpADH and its variants in the asymmetric reduction of various aromatic ketones 1a−10a.
significant roles in substrate recognition while a hydrophobic
interaction is important for the formation of a stable
prereaction.
polarity change at E214V made a major contribution to the
stereoselective improvement as discussed above. The calcu-
lated total electrostatic interaction between site 214 and
CPMK was −6.60 kcal/mol for WT−CPMKProR (E214) and
−2.59 kcal/mol in Mu-R2−CPMKProR (V214).
The mechanism of stereoselective inversion was explored by
comparing average conformation of WT−CPMKProR and Mu-
S5−CPMKProS. For Mu-S5, a quadrilateral plane with a
dihedral angle of 0.31° could be formed when connecting α-
carbons of four mutation sites, N136, V161, C237, and G214
(Figure S7B). This plane has a special charge distribution due
to the different polarities of these amino acids (Figure S7D).
CPMK needs to traverse this plane to access the catalytic
Constrained MD Simulations Analysis Reveals Evolu-
tion of Stereoselective Inversion. To better understand the
evolution of stereoselective inversion, constrained MD
simulations with prereaction conformations of WT−
CPMKProR, Mu-R2−CPMKProR, and Mu-S5−CPMKProS as
initial input were performed. Harmonic restraint was created
on O14CPMK, H9Y164, and C4NPH to allow certain freedom in
space, and thereby reduced computational time. After MD
simulation, average conformations of three protein−ligand
complexes were extracted and analyzed (Figure S6).The
F
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
center, which resembles passing through a “polar gate” (Figure
4B and Figure S7B). For CPMK, the chlorine-substituent
endows the chlorobenzene ring electronegativity, while the
surface of the basic pyridine ring on the other side of the
carbonyl group was positively charged. The catalytic
orientation of CPMK is presumed to be defined when
traversing the “polar gate”, due to its electrostatic attraction
and repulsion forces. In addition, an alkyl hydrophobic
interaction between V161 and chlorine-substituent and a
π−π interaction between Y164 and chlorobenzene rings could
be formed (Figure S7B). The nonbinding interaction network
formed by five mutation sites results in a prominent synergistic
effect on the evolution of S-selectivity.
while its dipole moment (4.88 D) was clearly different from
that of 5a. Inverted selectivity for these two ketones is most
likely caused by the reduced space steric effect of 4a and
enhanced polarity of 6a, which affect the intermolecular forces.
In addition, docking results based on the crystal structure of
WT show that sites F197 and Y127 can form π−alkyl
hydrophobic interactions with the chlorine atom in 3a and the
bromine atom in 5a. For 4a or 6a, however, no interaction
between WT and the substituents (fluorine or methoxy) was
observed (Figure S10), which may be responsible for the
relatively lower enantioselectivity compared with 3a and 5a.
Substrates 7a−9a are ketone substrates containing only one
aryl group. For mutant Mu-S5, configurations of products 7b−
9b were different. It was therefore presumed that para-chlorine
substituents on benzene rings are critical for chiral recognition.
Compared with 7a, the opposite orientation was observed for
substrates 8a and 9a under a catalytic state. Molecular docking
analysis showed that hydrophobic interactions could be formed
between Y127 and chlorine as well as NADPH and the
chlorobenzene ring for substrates 8a and 9a, which favors the
formation of (S)-8b and (R)-9b (Figure S11). Mu-S3, Mu-S4a,
Mu-S4b, and Mu-S5 exhibited excellent ee (>99%) toward 9a.
As a β-aromatic ketone ester, 10a has been widely investigated
as a substrate in biocatalysis.44 All variants displayed the same
R-selectivity as WT, and the highest ee value (92.5% ee) was
attained by Mu-R2.
Synthesis of Optically Pure R- and S-CPMA. The gram-
scale synthesis of R- and S-CPMA was carried out using
lyophilized recombinant E. coli cells overexpressing KpADH
variants (Mu-R2 or Mu-S5) and BmGDH for NADH
recycling. In a preliminary study, substrate inhibition was
observed at CPMK concentrations higher than 100 mM. Thus,
substrate fed-batch culture was adopted for substrate loading of
500 mM. For Mu-R2, CPMK was added at 0, 0.5, 1, 1.5, 2.0,
and 4.0 h. A complete conversion of CPMK was achieved in 6
h (Figure S12A), and 1.62 g R-CPMA was obtained at a high
optical purity (99.1% ee, [α]2D5 = −118.4).
For Mu-S5, extended substrate feeding intervals allowed a
complete reaction due to the relatively lower catalytic
efficiency of Mu-S5 compared with Mu-R2 (Figure S12B).
After 24 h, 500 mM CPMK was fully converted, and 1.61 g of
S-CPMA at excellent optical purity (97.6% ee, [α]2D5 = 117.1)
was obtained. NMR data and chiral HPLC chromatograms are
shown in Figure S13.
Conclusions and Perspectives. Here, we attempted to
engineer KpADH for inverted stereoselectivity toward diaryl
ketone substrates. A “polarity scanning” strategy was proposed
to identify key residues responsible for stereorecognition in the
substrate binding pocket and entrance tunnel. Based on
saturation mutagenesis of six key residues, iterative combina-
torial mutagenesis was performed, and two excellent variants,
Mu-R2 (99.1% ee, R) and Mu-S5 (97.4% ee, S), were obtained.
Though KpADH variants with excellent R- and S-selectivity
were obtained, the homology modeling is of limited use in
revealing the mechanisms of stereoselective inversion.
In WT, a similar quadrilateral plane consisting of Q136,
F161, E214, and S237 is, however, blocked by several aromatic
amino acids (F161, F197, and F236) (Figure 4A and Figure
S7A). Therefore, only one side of the aromatic ring (mostly
chlorobenzene) of CPMK could enter when approaching the
catalytic center (Figure 4A and Figure S7C). Compared with
Mu-S5, more nonbonded interactions could be formed in WT
(Figure S8), which favors a stronger affinity and a lower Km
toward the substrate (Table 2). Based on calculation in
Accelrys Discovery Studio (Figure S9), the biggest difference
in nonbonded interaction energy between WT and Mu-S5 was
observed at site 214, representing the importance of E214G in
stereoselective inversion. In the average conformation
extracted, nonbonded interactions could be formed between
Y164, F161, F197, C165, and the chlorobenzene ring. In
addition, π−alkyl hydrophobic interactions could form
between chlorine and residue Y127 in the binding pocket.
These results explain the R-selectivity observed for WT.
Enantioselectivity toward Various Aromatic Ketones.
Enantioselectivity of WT KpADH and its variants (Table 2)
toward 10 ketone substrates are illustrated in Figure 5 and
Table S4. All combinatorial mutants, with either inverted or
enhanced selectivity, showed a synergistic effect on all diaryl
ketones except the halogen-free substrate 2a, for which poor
stereoselectivity was observed. This result demonstrates that
the para-substituents on the benzene ring are critical in the
chiral recognition of enzymes. For 3a, the pyridine ring is
replaced by a benzene ring. Although all combinatorial mutants
showed similar selectivity as that of 1a, Mu-S5 and Mu-R2
exhibited slightly lower ee values of 91.7% (R-3b) and 96.4%
(S-3b). Compared with 1a (CPMK), the reversed config-
uration of 3a is mainly due to the changes in group priority
based on the Cahn−Ingold−Prelog rule.18 The effect of
various para-substituted benzene rings on enantioselectivity of
KpADH was evaluated using 3a−6a. Although polarities and
molecular volumes of these diaryl ketones are different, all
substituents including halogens and methoxy are electron-
withdrawing groups and show electronegativity. Variant Mu-S5
showed the same R-selectivity for all above diaryl ketones,
demonstrating that polar interactions are the major factor in its
stereorecognition. WT and variants with single mutations
exhibited opposite selectivity in the reduction of 4a (F-) and
6a (CH3O-) compared with that of 3a (Cl-) and 5a (Br-). To
understand the effect of various substituents on stereo-
selectivity, physical properties of these diaryl ketones were
measured using Gaussian software. Using 5a as a reference,
there was a small difference in the dipole moments between 5a
(2.95 D) and 4a (3.02 D); the difference in the volume of 5a
(164.4 Å3) and 4a (147.5 Å3), however, was significant. In
contrast, the volume of 6a (168.3 Å3) was close to that of 5a,
The crystals of WT and two variants (Mu-S5 and Mu-R2) in
complex with NADPH were resolved, and the structures were
further analyzed by prereaction state analysis and constrained
MD simulation. Based on computational data, polarity changes
at the substrate entrance are the predominant factor for the
increased enantioselectivity of Mu-R2. In MD simulation
analysis, a “polar gate” with a special charge distribution was
observed in Mu-S5, which is formed by N136, V161, C237,
G
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
proteins. The complexes were placed in a truncated octahedral box of
water molecules, extending 8.0 Å along each dimension. The total of
11 counterions, Na+, were added to neutralize the system. Then the
systems were minimized by the steepest descent minimization of 1000
steps followed the conjugate gradient minimization of 9000 steps,
heated from 0 to 300 K at constant volume in 50 ps, and equilibrated
for another 50 ps without any restraints. After an additional 10 ns of
and G214. Remarkably, a pro-S catalytic configuration of
CPMK is defined when traversing the “polar gate”. The energy
calculation by MD simulation proved the importance of E214
in stereoselectivity alteration, which is consistent with the
experiments. This study provides efficient approaches for
engineering stereocomplementary alcohol dehydrogenases
toward bulky diaryl ketones and elucidates its mechanism of
stereoselective inversion.
NPT simulation, constraints of d(O14CPMK−H9Y164) and d(C7CPMK
−
H4NPH) were added by using a harmonic vibrational potential with a
force constant of 500 kcal/(mol·Å2) and the O···H and C···H balance
distance at 1.7 Å. The distance-restricted systems were reminimized,
heated, and equilibrated with the same procedure. Then the distance
constraints were removed, and the PRS structures were re-
equilibrated for the trajectory production. Finally, multiple 10 ns
trajectories were collected for further PRS analysis. In the MD
simulations, the Particle Mesh Ewald (PME) method was employed
for long-range electrostatic interactions.
MATERIALS AND METHODS
■
Microorganism and Chemicals. Recombinant E. coli BL21
(DE3) strains harboring pET28-KpADH and pET28-BmGDH have
been constructed in a previous study.27 (4-Chlorophenyl)(pyridin-2-
yl)methanone (CPMK), all other reagents and solvents of analytical
grade, and biochemical reagents were obtained from Sinopharm
Chemical Reagent Co. Ltd.
Constrained Molecular Dynamics Simulation. The simula-
tions were executed in the simulation module of the Discovery Studio
4.0.29 Conformations of WT−CPMKProR, Mu-R2−CPMKProR, and
Mu-S5−CPMKProS of the prereaction state were treated with
Mutagenesis. Cloning and recombinant expression of KpADH
was performed as described in our previous study.27 To engineer the
enantioselectivity, polarity scanning using asparagine and valine as
“polar sieve” residues were employed toward residues around
substrate-binding pocket of KpADH. Variants were generated using
whole-plasmid mutagenesis protocol. Saturation mutagenesis was then
performed on beneficial sites using the NNK codon. Mutants
exhibiting over 90% or less than 75% ee were subjected to further
iterative combinatorial mutagenesis. Equal moles of mutant plasmids
were mixed and amplified using whole-plasmid PCR with KOD high-
fidelity polymerase. The PCR procedures were as follows: 95 °C for 2
min, followed by 15 cycles of 94 °C for 30 s, 55 °C for 30 s, and 68 °C
for 2 min and 30 s. After DpnI digestion at 37 °C for 30 min, the PCR
products were transformed into competent E. coli BL21 (DE3) cells
and plated onto LB agar plates supplemented with 50 μg/mL
kanamycin.
Library Screening. Variant colonies were inoculated into 300 μL
of LB medium containing 50 μg/mL kanamycin in 96-well plates and
grown at 37 °C for 12 h. Then 50 μL of cell culture was transferred
into 450 μL of fresh LB medium containing 50 μg/mL kanamycin and
incubated for another 2 h. Enzyme expression was induced by
addition of IPTG to a final concentration of 0.4 mM. After 5 h of
incubation at 25 °C, cells were harvested by centrifugation, then lysed
by addition of 1000 U of lysozyme and shaking at 37 °C for 1 h. The
plates were centrifuged at 4000 × g and 4 °C for 15 min, and 10 μL of
supernatant was transferred to a 96-well plate. The activity was
determined by adding 190 μL of enzymatic assay solution comprised
of 1 mM CPMK, 100 mM potassium phosphate buffer (pH 7.0), and
1 mM NADPH. The decrease in absorbance of NADPH at 340 nm
was monitored by a microplate reader (Biotek, USA).
Molecular Docking. The homology model of KpADH was built
based on the crystal structure of Saccharomyces cerevisiae NADPH-
dependent methylglyoxal reductase GRE2 (EC 1.1.1.283) (Protein
Data Bank Code 4PVC),28 which shares 53% amino acid sequence
identity with KpADH. Molecular docking was performed using
Discovery studio 4.0.29 Receptor proteins were prepared by
eliminating all bound water molecules. Hydrogen atoms were then
added to the protein, which was optimized by applying a CHARMm
force field.45 The active pocket was defined as amino acid residues
enclosed within a 5 Å radius sphere centered on a cavity. Amino acids
in or near the active pocket were selected to create protein
conformations and refine side chains. The obtained conformations
with hydrogen bond interactions between carbonyl oxygen and
Tyr164, as well as a less than 4.5 Å distance between carbonyl carbon
and NADPH-C4 atoms were considered as reasonable.
Molecular Dynamic Simulations of Prereaction State. The
initial coordinates used in the PRS (prereaction state) modeling were
taken from the CDOCKER results. The parameters of NADPH were
generated with the RESP46 method and RHF/6-31*47 calculation as
in the literature,48 and those of water molecules were assigned with
the TIP3P model. The force field parameters of the CPMK substrate
were prepared with the Antechamber package, and the general Amber
force field (GAFF) and ff03.r1 force field were applied for the
CHARMm force field. Harmonic restraint was created at O14CPMK
,
H9Y164, and C4NPH. Explicit solvent was applied to the system under
periodic boundary conditions, and then two stages of steepest descent
energy minimization were conducted. The system was heated from 0
to 300 K in 50 ps, and then a 1 ns equilibration process was
performed to thermally equilibrate the molecules of the systems. The
hydrogen bonds were fixed using the SHAKE algorithm during MD.
Finally, a 20 ns MD simulation was performed on the whole system at
300 K.
Enantioselectivity of KpADH and Mutants toward Prochiral
Ketones. Enantioselectivity of KpADH and mutants toward 10
prochiral ketones was determined by monitoring the reduction of
these aromatic ketones in an NADPH regeneration system containing
purified enzyme and glucose dehydrogenase (GDH) from Bacillus
subtilis. The reaction mixture (500 μL) included 0.25 mM NADP+, 10
mM ketone/ketonic ester, 100 μg of purified mutant or WT KpADH
enzyme, 15 mM glucose, and 2 U of GDH in 100 mM phosphate
buffer (pH 7.0). After 6 h of reduction, the reaction medium was
extracted with ethyl acetate. The ee value and concentration of
produced alcohols were determined by HPLC or GC equipped with
chiral columns as summarized in Table S5. The absolute configuration
and identity of diaryl alcohol products were confirmed by comparing
with refs 6 and 49.
Asymmetric Reduction of CPMK. Frozen cell pellets of KpADH
variants and BmGDH were mixed at ratio of 1:2 (25 U of KpADH
variants and 50 U of BmGDH) and suspended in sodium phosphate
buffer (100 mM, pH 7.0). The reactions consisting of 10 mL of cell
suspension mentioned above, 500 mM CPMK, and 1.5-fold glucose
were performed in a 20 mL bioreactor at 30 °C and 200 rpm. The pH
was automatically adjusted to 7.0 by titration with 1.0 M NaOH
solution. The progress of the asymmetric reduction was monitored
using Agilent 1100 HPLC equipped with a Chiralcel OB-H column
(0.46 mm × 250 mm, 5 μm, Diacel, Japan). HPLC was performed at
254 nm using hexane/ethanol (95:5, v/v) as eluent at a flow rate of
1.0 mL/min and 30 °C.
Protein Crystallization. Crystals of WT and variants Mu-S5 and
Mu-R2 in complex with NADPH were obtained at 18 °C using the
sitting drop vapor diffusion technique, with the initial condition of
mixing 1 μL of the 10 mg/mL protein sample and 1 mM NADPH
with an equal volume of mother liquor (0.2 M ammonium acetate, 0.1
M HEPES, pH 7.0, 31% PEG 3350 for WT−NADPH complex; 0.2 M
ammonium acetate, 0.1 M BIS-TRIS, pH 6.5, 27% PEG 3350 for Mu-
S5−NADPH complex; 0.1 M BIS-TRIS, pH 6.5, 27% PEG3350 for
Mu-R2−NADPH complex). Crystals grew up in about 3 days. Then,
they were transferred into a cryoprotectant (reservoir solution
supplemented with 10% (v/v) glycerol) and flash cooled in liquid
nitrogen. All data were collected at the Shanghai Synchrotron
Radiation Facility, Shanghai Institute of Applied Physics, Chinese
Academy of Sciences, using beamline BL17B, BL18U. Data
H
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
processing and scaling were performed using the HKL3000 package.50
All crystal structures were determined by molecular replacement
method with PHENIX51 using coordinates of Saccharomyces cerevisiae
NADPH-dependent methylglyoxal reductase (PDB ID 4PVC,
sequence identity is 53%) as the search model. PHENIX Phaser-
MR (simple one-component interface) was used for building the
initial model. Then initial model and phase information were
transferred to native data set. Other parts of the structure were
manually built in COOT52 and iteratively refined in PHENIX Refine.
In the final model, more than 99.8% of residues fell in the favored
region in the Ramachandran plot, and the final Rwork/Rfree are
0.1589/0.1904, 0.1697/0.2150, and 0.1473/0.1780 for WT KpADH−
NADPH, Mu-R2−NADPH, and Mu-S5−NADPH, respectively. Data
collection and refinement statistics are listed in Table S2. The atomic
coordinates of WT and variants Mu-R2 and Mu-S5 in complex with
NADPH have been deposited in the Protein Data Bank (PDB) under
accession numbers 5Z2X, 5ZED, and 5ZEC.
from BL17B, BL18U at Shanghai Synchrotron Radiation
Facility (China) for assistance during X-ray data collection.
ABBREVIATIONS
■
MD, molecular dynamics; ICM, iterative combinatorial
mutagenesis; dPOC, data based Progressive Optimal Combi-
nation; PRS, prereaction state
REFERENCES
■
(1) Devalia, J. L.; De Vos, C.; Hanotte, F.; Baltes, E. Allergy 2001,
56, 50−7.
(2) Schmidt, F.; Stemmler, R. T.; Rudolph, J.; Bolm, C. Chem. Soc.
Rev. 2006, 35, 454−470.
(3) Roszkowski, A. P.; Govier, W. M. Pharmacologist 1959, 1, 60−
78.
(4) Jen, W. S.; Truppo, M. D.; Amos, D.; Devine, P.; McNevin, M.;
Biba, M.; Campos, K. Org. Lett. 2008, 10, 741−744.
(5) (a) Yato, N.; Murata, T.; Saito, N.; Sakai, A.; Kikuchi, M.;
Tsuzurahara, K.; Marita, H. Nippon Yakurigaku Zasshi 1997, 110, 19−
29. (b) Kida, T.; Fujii, A.; Sakai, O.; Iemura, M.; Atsumi, I.; Wada, T.;
Sakaki, H. Exp. Eye Res. 2010, 91, 85−89. (c) Morita, K.; Urabe, K.;
Moroi, Y.; Koga, T.; Furue, M. J. Dermatol. 2002, 29, 709−712.
(d) Protzko, E. E.; Gomes, P. J.; Williams, J. I.; Gow, J. A.;
McNamara, T. R. J. Allergy Clin. Immunol. 2009, 123, S50−58.
(e) Simons, E. R.; Simons, K. J. J. Allergy Clin. Immunol. 2011, 128,
1139−1341.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.8b08640.
■
S
Selectivity data for additional mutants and data
collection and refinement statistics for X-ray diffraction
experiments (PDF)
Crystallographic structures of WT, Mu-R2, and Mu-S5
KpADH in complex with NADPH (CIF)
(6) Truppo, M. D.; Pollard, D.; Devine, P. Org. Lett. 2007, 9, 335−
338.
(7) Mahdy, A. M.; Webster, N. R. Anaesth. Crit. Care. Pa. 2017, 18
(4), 210.
AUTHOR INFORMATION
■
(8) Sun, Z.; Li, G.; Ilie, A.; Reetz, M. T. Tetrahedron Lett. 2016, 57,
3648−51.
Corresponding Authors
*yni@jiangnan.edu.cn
*jiahai@mail.sioc.ac.cn
(9) Wang, B.; Zhou, H.; Lu, G.; Liu, Q.; Jiang, X. Org. Lett. 2017, 19,
2094−7.
ORCID
(10) Chen, C.; Reamer, R. A.; Chilenski, J. R.; McWilliams, C. J. Org.
Lett. 2003, 5, 5039−42.
Ulrich Schwaneberg: 0000-0003-4026-701X
Yi-Lei Zhao: 0000-0003-4687-7847
Ye Ni: 0000-0003-4887-7517
Author Contributions
All authors have given approval to the final version of the
manuscript.
Notes
(11) Kamitori, S.; Iguchi, A.; Ohtaki, A.; Yamada, M.; Kita, K. J. Mol.
Biol. 2005, 352, 551−558.
(12) Li, H.; Zhu, D.; Hua, L.; Biehl, E. R. Adv. Synth. Catal. 2009,
351, 583−588.
(13) Xu, J.; Zhou, S.; Zhao, Y.; Xia, J.; Liu, X.; Xu, J.; He, B.; Wu, B.;
Zhang, J. Chem. Eng. J. 2017, 316, 919−27.
(14) Agudo, R.; Roiban, G. D.; Reetz, M. T. J. Am. Chem. Soc. 2013,
135, 1665−8.
The authors declare no competing financial interest.
(15) Zhu, D.; Yang, Y.; Buynak, J. D.; Hua, L. Org. Biomol. Chem.
2006, 4, 2690−5.
ACKNOWLEDGMENTS
(16) Zhu, D.; Yang, Y.; Majkowicz, S.; Pan, T. H.; Kantardjieff, K.;
Hua, L. Org. Lett. 2008, 10, 525−528.
■
Y.N. thanks the National Natural Science Foundation of China
(21776112), Natural Science Foundation of Jiangsu Province
(BK20150003), and Six Talent Peaks Project of Jiangsu
Province (2015-SWYY-008) for funding support. G.X. thanks
the National Natural Science Foundation of China
(21506073) and Natural Science Foundation of Jiangsu
Province (BK20171135) for funding support. Y.R. thanks
Natural Science Foundation of Jiangsu Province
(BK20160167) and the Fundamental Research Funds for the
Central Universities (JUSRP51712B) for funding support. J.Z.
thanks the Strategic Priority Research Program (B) of the
Chinese Academy of Sciences (XDB20000000) and the
Science and Technology Commission of Shanghai Municipal-
ity (15JC140040) for funding support. Funding from national
first-class discipline program of Light Industry Technology and
Engineering (LITE2018-07), the Program of Introducing
Talents of Discipline to Universities (111-2-06) and the
Collaborative Innovation Center of Jiangsu Modern Industrial
Fermentation are gratefully acknowledged. We thank the staffs
(17) Reetz, M. T. J. Am. Chem. Soc. 2013, 135, 12480−96.
(18) Prelog, V. Pure Appl. Chem. 1964, 9, 119−130.
(19) Zhu, D.; Hua, L. Pure Appl. Chem. 2010, 82, 117−128.
(20) Li, Z.; Liu, W.; Chen, X.; Jia, S.; Wu, Q.; Zhu, D.; Ma, Y.
Tetrahedron 2013, 69, 3561−4.
(21) Li, A.; Ye, L.; Yang, X.; Yang, C.; Gu, J.; Yu, H. Chem. Commun.
(Cambridge, U. K.) 2016, 52, 6284−7.
(22) Li, G.; Wang, J.; Reetz, M. T. Bioorg. Med. Chem. 2018, 26,
1241−51.
(23) Directed evolution for engineering enzyme enantioselectivity:
(a) Jaeger, K. E.; Eggert, T. Curr. Opin. Biotechnol. 2004, 15, 305−
313. (b) Williams, G. J.; Woodhall, T.; Farnsworth, L. M.; Nelson, A.;
Berry, A. J. Am. Chem. Soc. 2006, 128, 16238−47. (c) Reetz, M. T.;
Bocola, M.; Wang, L.; Sanchis, J.; Cronin, A.; Arand, M.; Zou, J.;
Archelas, A.; Bottalla, A.; Naworyta, A.; Mowbray, S. L. J. Am. Chem.
Soc. 2009, 131, 7334−43. (d) Reetz, M. T. J. Org. Chem. 2009, 74,
5767−78. (e) Giger, L.; Caner, S.; Obexer, R.; Kast, P.; Baker, D.;
Ban, N.; Hilvert, D. Nat. Chem. Biol. 2013, 9, 494−498. (f) Roiban, G.
D.; Reetz, M. T. Chem. Commun. (Cambridge, U. K.) 2015, 51, 2208−
24. (g) Sun, Z.; Wikmark, Y.; Backvall, J. E.; Reetz, M. T. Chem. - Eur.
I
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
J. 2016, 22, 5046−54. (h) Li, A.; Ilie, A.; Sun, Z.; Lonsdale, R.; Xu, J.
H.; Reetz, M. T. Angew. Chem., Int. Ed. 2016, 55, 12026−9. (i) Li, G.;
Wang, J. B.; Reetz, M. T. Bioorg. Med. Chem. 2018, 26, 1241. (j) Li,
G.; Furst, M.; Mansouri, H. R.; Ressmann, A. K.; Ilie, A.; Rudroff, F.;
Mihovilovic, M. D.; Fraaije, M. W.; Reetz, M. T. Org. Biomol. Chem.
2017, 15, 9824−9.
(44) Deng, J.; Yao, Z.; Chen, K.; Yuan, Y. A.; Lin, J.; Wei, D. J.
Biotechnol. 2016, 217, 31−40.
(45) Momany, F. A.; Rone, R. J. Comput. Chem. 1992, 13, 888−900.
(46) Cornell, W.; Cieplak, P.; Bayly, C. I.; Kollman, P. A.; et al. J.
Am. Chem. Soc. 1995, 117, 5179−5197.
(47) Stoll, H.; Wagenblast, G.; Preuss, H. J. Am. Chem. Soc. 1978,
100 (24), 7742−7743.
(24) Sun, Z.; Lonsdale, R.; Ilie, A.; Li, G.; Zhou, J.; Reetz, M. T. ACS
Catal. 2016, 6, 1598−605.
(48) Ryde, U. Proteins: Struct., Funct., Genet. 1995, 21, 40−56.
(49) Chen, C.; Reamer, R. A.; Chilenski, J. R.; McWilliams, C. J. Org.
Lett. 2003, 5, 5039−42.
(25) (a) Agudo, R.; Roiban, G. D.; Reetz, M. T. ChemBioChem
2012, 13, 1465−73. (b) Pennacchio, A.; Sannino, V.; Sorrentino, G.;
Rossi, M.; Raia, C. A.; Esposito, L. Appl. Microbiol. Biotechnol. 2013,
97, 3949−64. (c) Wechsler, C.; et al. ChemBioChem 2015, 16, 2580−
4. (d) Wijma, H. J.; Floor, R. J.; Bjelic, S.; Marrink, S. J.; Baker, D.;
Janssen, D. B. Angew. Chem., Int. Ed. 2015, 54, 3726−30. (e) Sun, Z.;
Lonsdale, R.; Wu, L.; Li, G.; Li, A.; Wang, J.; Zhou, J.; Reetz, M. T.
ACS Catal. 2016, 6, 1590−7. (f) Li, A.; Ye, L.; Yang, X.; Yang, C.; Gu,
J.; Yu, H. Chem. Commun. (Cambridge, U. K.) 2016, 52, 6284−7.
(g) Qin, F.; Qin, B.; Mori, T.; Wang, Y.; Meng, L.; Zhang, X.; Jia, X.;
Abe, I.; You, S. ACS Catal. 2016, 6, 6135−40. (h) Ebert, M. C. C. J.
C.; Pelletier, A. N. Curr. Opin. Chem. Biol. 2017, 37, 89−96. (i) Liu, J.;
Kuan, Y.; Tsou, Y.; Lin, T.; Hsu, W.; Yang, M.; Lin, J.; Wang, W. Sci.
Rep. 2018, 8, 2316. (j) Chen, X.; Zhang, H.; Feng, J.; Wu, Q.; Zhu, D.
ACS Catal. 2018, 8, 3525−31. (k) Chen, F.-F.; et al. ACS Catal. 2018,
8, 2622−8.
(50) Otwinowski, Z.; Minor, W. Methods Enzymol. 1997, 276, 307−
326.
(51) Adams, P. D.; Grosse-Kunstleve, R. W.; Hung, L. W.; Ioerger,
T. R.; McCoy, A. J.; Moriarty, N. W.; Read, R. J.; Sacchettini, J. C.;
Sauter, N. K.; Terwilliger, T. C. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2002, 58, 1948−54.
(52) Emsley, P.; Cowtan, K. Acta Crystallogr., Sect. D: Biol.
Crystallogr. 2004, 60, 2126−32.
(26) Patel, J. M.; Musa, M. M.; Rodriguez, L.; Sutton, D. A.; Popik,
V. V.; Phillips, R. S. Org. Biomol. Chem. 2014, 12, 5905−10.
(27) Zhou, J.; Xu, G.; Han, R.; Dong, J.; Zhang, W.; Zhang, R.; Ni,
Y. Catal. Sci. Technol. 2016, 6, 6320−27.
(28) Guo, P.; Bao, Z.; Ma, X.; Xia, Q.; Li, W. Biochim. Biophys. Acta,
Proteins Proteomics 2014, 1844, 1486−92.
̀
(29) Dassault Systemes BIOVIA Discovery Studio Modeling Environ-
̀
ment, Release 2017, Dassault Systemes, San Diego, 2016.
(30) Sun, Z.; Lonsdale, R.; Li, G.; Reetz, M. T. ChemBioChem 2016,
17, 1865−72.
(31) Sun, Z.; Lonsdale, R.; Kong, X. D.; Xu, J. H.; Zhou, J.; Reetz,
M. T. Angew. Chem., Int. Ed. 2015, 54, 12410−5.
(32) Marrone, L.; Viswanatha, T. Biochim. Biophys. Acta, Protein
Struct. Mol. Enzymol. 1997, 2, 263−277.
(33) Acevedo-Rocha, C. G.; Ferla, M.; Reetz, M. T. Methods Mol.
Biol. 2018, 1685, 87−128.
̇
(34) Gudiukaite, R.; et al. Appl. Biochem. Biotechnol. 2016, 178, 654.
(35) Moy, V. T.; Florin, E. L.; Gaub, H. E. Science 1994, 266, 257−9.
(36) Li, G.; Yao, P.; Gong, R.; Li, J.; Liu, P.; Lonsdale, R.; Wu, Q.;
Lin, J.; Zhu, D.; Reetz, M. T. Chem. Sci. 2017, 8, 4093−9.
(37) Kyte, J.; Doolittle, R. F. J. Mol. Biol. 1982, 157, 105−132.
(38) Zamyatnin, A. A. Annu. Rev. Biophys. Bioeng. 1984, 13, 145−65.
(39) Ha, T. H.; Suh, K.; Lee, G. S. B. Bull. Korean Chem. Soc. 2013,
34, 549−552.
(40) Pennacchio, A.; Sannino, V.; Sorrentino, G.; Rossi, M.; Raia, C.
A.; Esposito, L. Appl. Microbiol. Biotechnol. 2013, 97, 3949−64.
(41) (a) Klinman, J. P. CRC Crit Rev. Biochem. 1981, 10, 39−78.
(b) Kitson, T. M. Biochem. J. 1987, 248, 989−991. (c) Pei, X.; Erixon,
K. M.; Luisi, B. F.; Leeper, F. J. Biochemistry 2010, 49, 1727−1736.
(d) Chen, X.; Shi, T.; Wang, X.; Wang, J.; Chen, Q.; Bai, L.; Zhao, Y.
ACS Catal. 2016, 6, 4369−4378. (e) Shi, T.; Liu, L.; Tao, W.; Luo, S.;
Fan, S.; Wang, X.; Bai, L.; Zhao, Y. ACS Catal. 2018, 8, 4323−4332.
(42) Reaction mechanism of short-chain dehydrogenases: (a) Mc-
Kinley-McKee, J. S.; Winberg, J. O.; Pettersson, G. Biochem Int. 1991,
̈
5, 879−885. (b) Jornvall, H.; Persson, B.; Krook, M.; Atrian, S.;
Gonzalez-Duarte, R.; Jeffery, J.; Ghosh, D. Biochemistry 1995, 34,
6003−6013. (c) Filling, C.; Berndt, K. D.; Benach, J.; Knapp, S.;
̈
Prozorovski, T.; Nordling, E.; Ladenstein, R.; Jornvall, H.;
Oppermann, U. J. Biol. Chem. 2002, 277, 25677−25684.
(43) (a) Maria-Solano, M. A.; Romero-Rivera, A.; Osuna, S. Org.
Biomol. Chem. 2017, 15, 4122−4129. (b) Agarwal, P. K.; Webb, S. P.;
Hammes-Schiffer, S. J. Am. Chem. Soc. 2000, 122, 4803−4812.
(c) Billeter, S. R.; Webb, S. P.; Agarwal, P. K.; Iordanov, T.; Hammes-
Schiffer, S. J. Am. Chem. Soc. 2001, 123, 11262−11272.
J
DOI: 10.1021/jacs.8b08640
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX