Hydroclassified Combinatorial Saturation Mutagenesis: Reshaping Substrate Binding Pockets of KpADH for Enantioselective Reduction of Bulky-Bulky Ketones

Article abstract of DOI:10.1021/acscatal.8b02286

A hydroclassified combinatorial saturation mutagenesis (HCSM) strategy was proposed for reshaping the substrate binding pocket by dividing 20 amino acids into four groups based on their hydrophobicity and size. These smart HCSM libraries could significantly reduce screening effort especially for the simultaneous mutagenesis of three or more residues and lacking high throughput screening methods. Employing HCSM strategy, the stereoselectivity of KpADH, an alcohol dehydrogenase from Kluyveromyces polysporus, was efficiently improved to 99.4% ee. (4-Chlorophenyl)(pyridin-2-yl)methanone (CPMK), generally regarded as a "hard-to-reduce" ketone, was used as a model substrate, and its corresponding chiral alcohol products could be utilized as antihistamine precursors. The best variant 50C10 displayed higher binding affinity and catalytic efficiency toward CPMK with K_M/k_cat of 59.3 s^-1·mM^-1, 3.51-fold that of KpADH. Based on MD simulations, increased difference between two binding pockets, enhanced hydrophobicity, and π-π and halogen-alkyl interactions were proposed to favor the enantioselective recognition and substrate binding in 50C10. Substrate spectrum analysis revealed that 50C10 exhibited improved enantioselectivity toward diaryl ketones especially with halo- or other electron-withdrawing groups. As much as 500 mM CPMK could be asymmetrically reduced into chiral diaryl alcohols with ee of 99.4% and a space-time yield of 194 g·L^-1·d^-1 without addition of external NADP⁺. This study provides an effective mutagenesis strategy for the protein engineering of substrate specificity and enantioselectivity.

Full text of DOI:10.1021/acscatal.8b02286

Research Article
pubs.acs.org/acscatalysis
Cite This: ACS Catal. 2018, 8, 8336−8345
Hydroclassified Combinatorial Saturation Mutagenesis: Reshaping
Substrate Binding Pockets of KpADH for Enantioselective Reduction
of Bulky−Bulky Ketones
Guo-Chao Xu,† Yue Wang,† Ming-Hui Tang,† Jie-Yu Zhou,† Jing Zhao,‡ Rui-Zhi Han,† and Ye Ni*,†
†Key Laboratory of Industrial Biotechnology, Ministry of Education, School of Biotechnology, Jiangnan University, Wuxi 214122,
Jiangsu, China
‡Key Laboratory of Systems Microbial Biotechnology, Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences,
Tianjin 300308, China
S
* Supporting Information
ABSTRACT: A hydroclassified combinatorial saturation
mutagenesis (HCSM) strategy was proposed for reshaping
the substrate binding pocket by dividing 20 amino acids into
four groups based on their hydrophobicity and size. These
smart HCSM libraries could significantly reduce screening
effort especially for the simultaneous mutagenesis of three or
more residues and lacking high throughput screening
methods. Employing HCSM strategy, the stereoselectivity of
KpADH, an alcohol dehydrogenase from Kluyveromyces
polysporus, was efficiently improved to 99.4% ee. (4-
Chlorophenyl)(pyridin-2-yl)methanone (CPMK), generally
regarded as a “hard-to-reduce” ketone, was used as a model
substrate, and its corresponding chiral alcohol products could be utilized as antihistamine precursors. The best variant 50C10
displayed higher binding affinity and catalytic efficiency toward CPMK with KM/kcat of 59.3 s−1·mM−1, 3.51-fold that of
KpADH. Based on MD simulations, increased difference between two binding pockets, enhanced hydrophobicity, and π−π and
halogen−alkyl interactions were proposed to favor the enantioselective recognition and substrate binding in 50C10. Substrate
spectrum analysis revealed that 50C10 exhibited improved enantioselectivity toward diaryl ketones especially with halo- or other
electron-withdrawing groups. As much as 500 mM CPMK could be asymmetrically reduced into chiral diaryl alcohols with ee of
99.4% and a space−time yield of 194 g·L−1·d−1 without addition of external NADP+. This study provides an effective
mutagenesis strategy for the protein engineering of substrate specificity and enantioselectivity.
KEYWORDS: alcohol dehydrogenase, hydroclassified amino acids, directed evolution, enantioselectivity, (R)-CPMA
with two similar aryl substituents remains challenging for not
only chemocatalysts but also biocatalysts, especially at high
substrate loading.6 Only SSCR from Sporobolomyces salmoni-
color and commercial ketoreductase KRED124 have been
reported with 88−94% ee at 10 g·L−1 (4-chlorophenyl)-
(pyridin-2-yl)methanone (CPMK, 1a) in depletion of 1 g·L−1
NADP+.7 Previously, KpADH has been identified from
Kluyveromyces polysporus by genome database mining with
82.5% (R) ee in the reduction of 100 mM CPMK.8 The
KpADH displays the highest kcat/KM among all the reported
reductases, indicating it is a promising candidate for further
protein engineering.
INTRODUCTION
■
Optically active diarylmethanols are important building blocks
and structural motifs of pharmaceuticals, such as antihistamine,
diuretic, antiepileptic, asthma, and antidepressant drugs
(Scheme 1).1 Moreover, via SN2 substitution at the C−O
bond, chiral diarylmethanols could be facilely converted into
diarylmethane derivatives without loss of optical purity.2 As a
result, efficient synthesis of chiral diarylmethanols, typically
through asymmetric addition of aryl nucleophiles benzalde-
hyde or asymmetric transfer hydrogenation (ATH) of the
corresponding diaryl ketones, is of special interest.3
Compared with ATH using precious metals, such as iridium,
rhodium, and ruthenium as catalysts, carbonyl reductases
mediated reactions have many inherent merits such as
environmental compatibility, high enantioselectivity, and mild
reaction conditions.4 Thus, asymmetric bioreduction has
attracted ever-increasing attention, and various reductases/
dehydrogenases have been validated at large scale.5 However,
the enantioselective reduction of bulky−bulky diaryl ketones
Directed evolution is a commonly practiced approach for
protein engineering as an alternative to rational design.9 Site-
directed saturation mutagenesis is a major strategy performed
Received: June 22, 2018
Revised: July 26, 2018
© XXXX American Chemical Society
8336
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Scheme 1. Stereoview of the Homology Structure of KpADH and Asymmetric Reduction of CPMK into (R)-CPMA Catalyzed
by KpADH. (A) Homology Structure of KpADH; (B) Asymmetric Reduction of CPMK into (R)-CPMA; (C) Residues Lining
a
the Substrate Binding Pockets
a
Catalytic triad illustrated in green stick, NADPH in yellow stick, CPMK in magenta stick, residues of substrate binding pockets in cyan ball and
stick.
Scheme 2. Hydroclassified Combinatorial Saturation Mutagenesis Alphabets and the Screening Effort Required for
Combinatorial Saturation Mutagenesis of Two or Three Residues Employing NNK/S, 22c-Trick, and HCSM Strategies
a
a
Dashed lines, 2 residues; solid lines, 3 residues; red lines, NNK/S; purple lines, 22c-trick; green lines, HCSM. Number of variants is calculated
according to the formula L = −V × Ln(1 − F), where L is the number of variants or library size, V is the theoretical number of variants, and F is the
expected coverage of library, e.g., 0.95 for 95% coverage.
at sites lining the binding pockets of enzymes to improve their
limited substrate scope, insufficient activity, and poor regio- or
stereoselectivity.10 Although numerous successes in obtaining
enzymes with improved catalytic properties have been
achieved by directed evolution, many uncertainties still persist
to be clarified, for example, how to group a variety of residues
into randomization site, how to choose optimal upward
pathways, etc.11 As the number of residues in a randomization
site increases to five or more, the screening effort for 95%
library coverage, as calculated by CASTER based on the
Patrick/Firth algorithm, is astronomically high.12 Therefore,
the construction of libraries for multiple sites with reasonable
size is still challenging, especially in the engineering of certain
properties, such as stereoselectivity, for which a high-
throughput screening method is not available.13 To solve the
labor-intensive bottleneck, three tricks have been proposed to
create “smart” mutant libraries: (1) binary patterning method
employing polar/nonpolar residues, (2) splitting up a large
number of sites into smaller ones followed by iterative
saturation mutagenesis, and (3) use of reduced amino acid
alphabets.14 NNK (32:20, codons:amino acids), NDT (12:12),
or 22c-trick (22:20) degeneracy have been commonly adopted
in directed evolution to search the whole protein sequence
space.15 However, the concurrent oversampling increases as
the residues in randomization site increase from one residue to
more, and the inevitable codon bias emerges with the
redundant degenerated codons. In the simultaneous mutation
of three residues as illustrated in Scheme 2, the coverage rates
of 2000−3000 variants (generally accepted screening work for
enantioselectivity engineering) are 14% and 38% for NNK and
8337
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
22c-trick, respectively, far below 95% library coverage. Two
new approaches, the smallest amino acids alphabet and triple-
codon saturation mutagenesis, have been developed to
construct smaller but smarter libraries for reduced screening
work.16 Nevertheless, the diversity is limited and most of the
variants might be deactivated, especially in the case of
simultaneous mutation of five or more sites.17
Herein, a new strategy, hydroclassified combinatorial
saturation mutagenesis (HCSM), was proposed to facilely
engineer the enantioselectivity of an alcohol dehydrogenase
from Kluyveromyces polysporus (KpADH) (Scheme 2) based on
the homology model. Residues lining the binding pockets were
divided into groups and subjected to combinatorial saturation
mutagenesis employing reduced amino acids alphabet based on
hydrophobicity and steric hindrance. With regard to the
simultaneous mutation of three residues, only 2588 variants are
required for 95% coverage. This study provides evidence for
the feasibility of this HCSM strategy by enhancing the
enantioselectivity of KpADH from 82.5% to 99.4%. The
beneficial variant of KpADH also exhibited high catalytic
efficiency and could be applied in the preparation of optically
active diaryl alcohols at as high as 500 mM substrate loading.
Table 1. Alanine/Tyrosine Scanning of KpADH toward
CPMK
residue
ee of alanine mutation (%)
ee of tyrosine mutation (%)
F86
82.7 0.3
78.7 0.6
82.5 0.6
88.6 0.3
85.3 0.7
78.3 0.8
67.6 0.5
75.5 0.3
74.7 0.5
94.2 0.5
79.8 0.3
96.1 0.6
79.3 0.2
82.1 0.5
76.8 0.3
75.6 0.5
82.5 0.6 (77.9 0.5)
82.9 0.3 (82.8 0.2)
a
Y127
A128
M131
P133
Q136
F161
C165
S196
F197
E214
S237
Q238
T234
I308
a
86.7 0.5
82.7 0.3
79.7 0.3
67.5 0.3
88.1 0.5
74.7 0.3
86.6 0.5
93.8 0.2
94.1 0.4
87.7 0.5
83.7 0.3
73.4 0.4
a
Number in the parentheses denotes the ee of mutants with
phenylalanine mutation.
exhibit higher enantioselectivity than mutations into smaller
amino acids. Likewise, variants with mutation at PocL residues
into smaller amino acids might exhibit higher enantioselectivity
than mutations into larger amino acids.
RESULTS AND DISCUSSION
■
Alanine and Tyrosine Scanning for Hotspots. A
proposed mechanism for asymmetric carbonyl reduction
catalyzed by short chain dehydrogenase/reductase (SDR) has
been well reviewed with Ser-Tyr-Lys as catalytic triad.18
Catalytic residues Ser and Tyr are responsible for the
stabilization of alkoxide formed on hydride transfer from
NADPH and nucleophilic attack, while Lys is essential for the
stabilization of substrate and activation of water. It has been
recognized that there are small and large binding pockets (PocS
and PocL) in SDR which could discriminate and accommodate
prochiral ketones with substituents differing in size.19 Physical
and chemical properties of PocS and PocL, determined by their
residues constitution, are vital for catalytic properties,
especially substrate specificity and enantioselectivity. A
homology model of KpADH was constructed employing
crystal structures of methylglyoxal/isovaleraldehyde reductase
from Saccharomyces cerevisiae (51% identity, PDB: 4PVC) and
ketoreductase 1 from Candida glabrata (50% identity, PDB:
5B6K) as templates.20 Evaluation of the homology model of
KpADH reveals that over 99.4% of residues are located in the
allowable regions in the Ramachandran plot (Figure S1). The
proposed substrate binding pockets of KpADH are made up of
16 residues according to its homology model, including V84,
F86, Y127, A128, M131, P133, Q136, F161, C165, S196,
F197, E214, S237, T234, Q238, and I308 (Table 1).
Simultaneous saturation mutagenesis of all the residues in
substrate binding pockets is obviously beyond reach. More-
over, most of the above 16 residues locate in the loop region
and might not be correctly assigned into PocS or PocL according
to the homology model. As a result, we attempted to identify
the hotspots by alanine and tyrosine scanning and then classify
them into PocS or PocL, which were presumed to accommodate
2′-pyridyl and 4′-chlorophenyl groups of CPMK, respectively.
Mutation of PocS residues into larger amino acids might cause a
reduction in the size of PocS and thus increased volumetric
difference between PocS and PocL, whereas mutation into
smaller amino acids could lead to enlarged PocS and decreased
difference between two pockets. Presumably, variants with
mutations at PocS residues into larger amino acids might
Preliminary screening was therefore performed by mutation
of each residue into alanine and tyrosine (namely alanine and
tyrosine scanning method),21 due to their moderate hydro-
phobicity (+1.8 for alanine and −1.3 for tyrosine), different
size, and additional hydroxyl group of tyrosine. Scanning
results revealed that 11 residues participate in influencing the
enantioselectivity of KpADH toward CPMK, while the other
five residues V84, F86, A128, P133, and T234 were eliminated
due to their unstable ee of 15−83% (V84, data not shown) or
similar ee values compared with KpADH (Table 1). As shown
in Table 1, variants with mutations at residues Y127, Q136,
F161, and S196 exhibited similar enantioselectivity between
alanine and tyrosine mutations. As a result, they were not
selected as hotspots, in spite of their different enantioselectivity
in comparison with KpADH. The ee values of I308A and I308Y
were 76.8% and 73.4% respectively, lower than 82.5% ee of
KpADH. Thus, site 308 was also not chosen as a hotspot.
Mutations at M131, C165, F197, E214, S237, and Q238
resulted in highly improved ee. Hence, these six sites were
selected as mutational hotspots. The alanine mutants M131A
(88.6%), F197A (94.2%), and S237A (96.1%) exhibited higher
ee values than their corresponding tyrosine mutants M131Y
(86.7%), F197Y (86.6%), and S237Y (94.1%), respectively.
The results suggest that M131, F197, and S237 could be vital
sites in PocL due to their enhanced stereoselectivity when
mutated into smaller amino acids. The alanine mutants C165A
(75.5%), E214A (79.8%), and Q238A (79.3%) displayed lower
ee values than their corresponding tyrosine mutants C165Y
(88.1%), E214Y (93.8%), and Q238Y (87.7%), respectively. It
indicates that C165, E214, and Q238 could be key sites in PocS
since higher ee values were achieved when mutated into larger
amino acids. Consequently, six hotspots were identified to play
critical roles in manipulating the enatioselectivity of KpADH,
and they were grouped into PocS and PocL residues for further
study.
Development and Screening of HCSM Libraries.
Considering that the enantioselectivity could be manipulated
8338
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Figure 1. Screening results of HCSM libraries on large and small substrate binding pockets. (A) PocL-I, (B) PocL-II, (C) PocL-III, (D) PocL-IV, (E)
PocS-I, (F) PocS-II, (G) PocS-III, (H) PocS-IV, and (I) combinatorial mutagenesis.
by more than one residue in substrate binding pockets,
combinatorial mutagenesis is often performed. To conduct the
combinatorial saturation mutagenesis with high efficacy,
hydroclassified amino acid alphabet was proposed by
classifying the codons into four groups based on the codon
bias of E. coli, hydrophobicity, and size of amino acids. As
illustrated in Scheme 2, 24 codons were chosen for 20 amino
acids and were classified into Group I (designated as
“Hydrophobic group”, VYG), Group II (“Hydrophilic
group”, VRT), Group III (“Mixed group”, VWA), and Group
IV (“Aromatic group”, TDS). Specifically, Group I codes for
threonine, methionine, alanine, valine, proline, and leucine.
Group II codes for aspartate, asparagine, serine, glycine,
histidine, and arginine. Group III codes for lysine, isoleucine,
glutamate, valine, glutamine, and leucine. Group IV codes for
tyrosine, tryptophan, phenylalanine, cysteine, leucine, and stop
codons. Saturation mutagenesis employing this hydroclassified
amino acid alphabet is named as hydroclassified combinatorial
saturation mutagenesis (HCSM). Compared with other
combinatorial mutagenesis strategies for engineering enantio-
selectivity such as triple-codon saturation mutagenesis
(TCSM), double-codon saturation mutagenesis (DCSM),
and single-codon saturation mutagenesis (SCSM), HCSM is
a hexatruple-codon saturation mutagenesis method in which
each site is mutated into six other amino acids. It is therefore
supposed to be more efficient since more codons are employed
in randomizing the mutation sites (TCSM > DCSM >
SCSM).22 Most importantly, in HCSM, all hotspots could be
simultaneously mutated into hydrophobic (Group I), hydro-
philic (Group II), or aromatic (Group IV) amino acids, as well
as the mixed hydrophobic and hydrophilic codons in Group
III, which could conduce to the evaluation of possible
synergistic effects of hydrophobicity and hydrophilicity.
Consequently, HCSM is advantageous for developing libraries
with high diversity in volume, hydrophobicity, substituents,
and electricity.
Results of HCSM Libraries in PocL. Furthermore, the HCSM
libraries of KpADH at the above six hotspots in PocS and PocL
were constructed. For M131, F197, and S237 in PocL, all three
sites were combinatorial saturation mutated adopting the
8339
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Table 2. Kinetic Parameters of KpADH and Variants toward CPMK
variant
library
mutation
ee (%)
KM (mM)
kcat (s−1
)
k
cat/KM (s−1·mM−1
)
KpADH
3D3
3H8
6C12
43D11
45A3
47G9
49A7
49D6
50C10
51E10
M131V
C165F
F197T
E214Y
S237A
Q238C
−
PocL-I
PocL-I
PocL-I
PocS-IV
PocS-IV
PocS-IV
Comb. Lib.
Comb. Lib.
Comb. Lib.
Comb. Lib.
−
82.5 0.5
98.2 0.3
97.8 0.4
98.5 0.3
97.4 0.5
97.5 0.3
97.7 0.5
98.9 0.2
98.9 0.3
99.4 0.2
98.7 0.4
87.4 0.6
89.1 0.3
87.4 0.5
94.2 0.3
96.1 0.3
88.4 0.3
0.78 0.07
3.84 0.11
1.95 0.10
2.75 0.15
1.51 0.08
0.68 0.05
1.46 0.09
1.88 0.07
2.62 0.08
0.35 0.05
6.52 0.13
1.14 0.07
0.52 0.04
8.12 0.13
0.80 0.06
0.38 0.04
3.42 0.14
13.2 0.3
8.23 0.17
21.3 0.4
12.7 0.5
15.8 0.4
14.9 0.5
13.2 0.4
28.4 0.5
22.0 0.3
20.8 0.3
55.4 1.2
9.39 0.21
9.26 0.22
6.80 0.19
14.9 0.6
15.4 0.7
10.2 0.3
16.9
2.14
10.9
4.62
10.46
21.91
9.04
15.1
8.40
59.4
8.50
8.24
17.8
0.84
18.6
40.5
2.98
M131A/F197T/S237A
M131A/S237A
M131V/F197T/S237A
C165F/E214Y/Q238C
C165F/E214Y
C165W/E214Y/Q238C
M131V/F197T/E214Y/S237A
C165W/E214Y/S237A/Q238C
C165F/E214Y/S237A
M131A/C165F/F197T/E214Y/S237A
M131V
C165F
F197T
E214Y
S237A
a
b
SDM
SDM
SDM
SDM
SDM
SDM
Q238C
a
b
Comb. Lib.: combinatorial mutagenesis library. SDM: site-directed mutagenesis.
hydroclassified amino acid alphabet (Groups I, II, III, and IV)
to generate HCSM libraries of PocL-I, PocL-II, PocL-III, and
PocL-IV, respectively. Theoretically, at least 647 variants are
required for each library to achieve 95% coverage according to
the formula in Scheme 2. In total, at least 2588 variants are
required for four libraries of PocL. This library size is still too
large for enantioselectivity analysis by chiral HPLC. As a result,
two rounds of screening were adopted to isolate KpADH
variants with both high activity and enantioselectivity. First,
deactivated variants were eliminated by high throughput
screening on activity. Then, biotransformation employing the
active variants was conducted for enantioselectivity analysis. As
shown in Figure 1B, 98% of the variants of PocL-II were
deactivated, and only 15 variants displayed 82−94% ee and less
than 50% activity of KpADH, indicating hydrophilic muta-
genesis was not advantageous for the enantioselective binding
and reduction of CPMK. Mutation of residues in PocL into
aromatic amino acids resulted in less than 82% ee for all
variants (Figure 1D), demonstrating that the decreased
difference between size of PocS and PocL was unfavorable for
the discrimination of 2′-pyridyl and 4′-chlorophenyl groups,
which is in coincidence with the above hypothesis. Most
variants of PocL-I and PocL-III were still active and exhibited
higher ee, especially using Group I codons (Figure 1A, C). The
three best variants in PocL-I, 3D3, 3H8, and 6C12, exhibited ee
of 98.2%, 97.8%, and 98.5% respectively, significantly higher
than that of KpADH (82.5% ee) (Table 2). Compared with
PocL-II and PocL-III, hydrophobic mutagenesis could alleviate
the severe deactivation caused by hydrophilic mutagenesis to
some extent. Consequently, hydrophobic interaction was
important for the large substrate binding pocket. All the
variants with over 95% ee in four HCSM libraries of PocL were
rescreened and sequenced as shown in the Supporting
Information. The mutational sites in 3D3, 3H8, and 6C12
are M131A/F197T/S237A, M131A/S237A, and M131V/
F197A/S237A, respectively.
specific activity toward CPMK than KpADH were selected for
enantioselectivity analysis. Almost all the variants in PocS-I,
PocS-II, and PocS-III displayed lower ee than KpADH as
illustrated in Figure 1E−G. Unlike the importance of
hydrophobic interaction observed in PocL, steric hindrance
plays a critical role in PocS. It is noted that mutation of C165,
E214, and Q238 into aromatic amino acids employing Group
IV codons resulted in significantly improved ee (Figure 1H).
The three best variants in PocS-IV, 43D11 (C165F/E214Y/
Q238C), 45A3 (C165F/E214Y), and 47G9 (C165W/E214Y/
Q238C), exhibited ee of 97.4%, 97.5%, and 97.7%, respectively.
As expected, mutation of residues in PocS into larger amino
acids would decrease the size of PocS and further increase the
discrepancy between PocS and PocL, which is favorable for the
enantioselective recognition of prochiral ketones.
Results of Combinatorial Mutagenesis of PocL and PocS.
Combinatorial mutageneses at hotspots of PocS and PocL
were performed by mixing 3D3, 5H4, 6C12, 43D11, 45A3,
and 47G9 as templates. About 250 variants in the
combinatorial mutagenesis library were isolated and screened
for improved activity and enantioselectivity. About 50%
variants displayed improved activity as compared to KpADH
and enantioselectivity of over 90% ee (Figure 1I and Table
S11). The ee values of 49A7 (M131V/F197T/E214Y/S237A),
49D6 (C165W/E214Y/S237A/Q238C), 50C10 (C165F/
E214Y/S237A), and 51E10 (M131A/C165F/F197T/E214Y/
S237A) were 98.9%, 98.9%, 99.4%, and 98.7%, higher than the
variants isolated from PocS or PocL libraries. Remarkably,
50C10 exhibited the highest ee of 99.4% which satisfies the
optical purity requirement of pharmaceutical intermediates.23
Although some variants also display high enantioselectivity,
such as 50D8 (M131A/C165F/F197T/E214Y/S237A/
Q238C) with 99.5% ee, they were eliminated due to lower
specific activities than KpADH.
Kinetic Parameters and MD Simulation Analysis. To
get insight into the effect of substrate binding pockets on the
improved enantioselectivity, all the beneficial variants were
purified to homogeneity, and their kinetic parameters were
determined. Several single mutants including M131V, C165F,
F197T, E214Y, S237A, and Q238C were also constructed. The
KM, Vmax, and kcat/KM of KpADH were 0.78 mM, 19.8 μmol·
Results of HCSM Libraries in PocS. Likewise, four HCSM
libraries (PocS-I, PocS-II, PocS-III, and PocS-IV) of C165, E214,
and Q238 of PocS were also developed and screened as shown
in Figure 1E−H and Tables S3−S10. No severe deactivation
was found in all four libraries. Variants with similar or higher
8340
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Figure 2. Analysis of substrate binding pockets and interactions between ligand and protein in 20 ns MD simulation of KpADH and 50C10. (A)
Substrate binding pocket of KpADH. (B) Substrate binding pocket of 50C10. (C) Interactions formed in the average conformation of KpADH−
CPMK. (D) Interactions formed in the average conformation of 50C10-CPMK. CPMK is depicted in magenta, NADPH in yellow, catalytic triad in
green, residues of binding pockets in cyan. Hydrogen bond is illustrated in green dotted line, π−π stacking interaction in magenta line, halogen−
alkyl interaction in pink line.
Figure 3. Substrate spectrum of KpADH and 50C10 toward various prochiral ketones.
min−1·mg−1, and 16.9 s−1·mM−1, respectively. Variants of PocL
displayed decreased kcat/KM as compared with KpADH due to
their higher KM (Table 2). Among M131V, F197T, and S237A,
mutation at F197 exhibited a negative effect on substrate
binding affinity, and the KM of F197T was as high as 8.12 mM,
over 10 times that of KpADH. Similarly, the KM of 3D3 and
3H8 was also higher than that of KpADH. The kcat/KM of
43D11, 45A3, and 47G9 were 10.4, 22.0, and 9.06 s−1·mM−1,
respectively. The decreased kcat/KM of 43D11 and 47G9 could
be ascribed to the mutation of Q238C since the KM and kcat of
C165F, E214Y, and Q238C were 0.52 mM and 9.26 s−1, 0.80
mM and 14.8 s−1, and 3.42 mM and 10.2 s−1. A similar
phenomenon was also found with 49A7, 49D6, and 51E10.
Although the kcat values of 49A7, 49D6, and 51E10 were 28.4,
22.0, and 55.4 s−1, much higher than 13.2 s−1 of KpADH, the
increased KM led to lower kcat/KM, considering all of them
contain F197T and (or) Q238C. Both F197T and Q238C play
important roles in influencing enantioselectivity; the ee values
of 3D3 and 47G9 were slightly higher than those of 3H8 and
45A3 as shown in Table 2. The kcat/KM of the best variant
50C10 (C165F/E214Y/S237A) was 59.3 s−1·mM−1, 3.51-fold
that of KpADH, indicating the two-round screening was
effective in identifying variants with high activity and
enantioselectivity. The ee values of C165F, E214Y, and
8341
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
S237A were 89.1%, 94.2%, and 96.1%, while their kcat/KM
values were 17.8, 18.6, and 40.5 s−1·mM−1. The results indicate
synergistic effects existed among C165, E214, and S237, since
50C10 displayed higher ee and kcat/KM than three single
mutants C165F, E214Y, and S237A.
Asymmetric Preparation of Chiral (R)-1b Employing
Variant 50C10. Considering its highest enantioselectivity and
catalytic efficiency, 50C10 was applied in asymmetric
reduction of CPMK (1a) for producing (R)-CPMA ((R)-1b)
at high substrate loading (Figure 4). Glucose dehydrogenase
To understand the molecular mechanism of improved
enantioselectivity, the crystal structure of KpADH in complex
with NADPH was resolved by X-ray diffraction and was
deposited in the Protein Data Bank (PDB) under accession
number of 5Z2X (unpublished data). KpADH is a homodimer
enzyme made up of 12 β-sheet surrounded by 10 α-helix
(Figure S2). The homology model of 50C10 was constructed
by virtual mutation of KpADH. As illustrated in Figure 2, the
substrate binding pocket of 50C10 is enlarged, and the surface
becomes more hydrophobic than that of KpADH. Further-
more, substrate CPMK (1a) was docked into KpADH and
50C10, and the protein−ligand complexes were subjected to
20 ns molecular dynamic (MD) simulation (Figures S3−S4).
The average conformations of KpADH−CPMK and 50C10−
CPMK under catalytic state were extracted. The binding free
energy between 50C10 and CPMK in the average
conformation was calculated to be −103.7 kcal·mol−1, which
is −31.2 kcal·mol−1 lower than that of KpADH−CPMK. The
significantly reduced substrate binding energy could explain
the decreased KM and increased kcat of 50C10. Interactions
analysis reveals that only weak interactions could be formed
between F197/V84 and CPMK in KpADH (Figure 2C).
However, in 50C10, more nonbonded interactions emerged,
including halogen−alkyl interactions between M131/S237A
and chlorine-substitute of CPMK and π−π interactions
between E214Y and pyridyl of CPMK as well as C165F and
phenyl of CPMK (Figure 2D, Figure S5−S6). In addition, the
increased difference between size of PocS and PocL is also
important for the enantioselective recognition. The above
analysis indicates that the stereoselectivity of the newly evolved
50C10 was enhanced by reshaping its substrate binding pocket
employing HCSM strategy.
Substrate Spectrum of KpADH and 50C10. Various
prochiral ketones with similar substituents as CPMK (1a) were
chosen to evaluate the substrate specificity of KpADH and
50C10 as illustrated in Figure 3. For the reduction of
chloroacetophenone derivatives (2a−5a), the specific activities
and enantioselectivities of 50C10 were relatively lower than
those of KpADH and in an order of 4′-chloroacetophenone >
3′-chloroacetophenone > 2′-chloroacetophenone. A similar
phenomenon was also observed in the reduction of different
acetylpyridines (6a−8a), except for 1-(pyridin-4′-yl)ethanone
(8a). These results confirmed the importance of 2′-pyridyl and
4′-chlorophenyl groups in influencing the recognition and
binding of 1a by KpADH. Eight diaryl ketones (9a−16a) were
also selected to investigate the substrate profiles. Unlike
chloroacetophenones or acetylpyridines, 50C10 displayed
improved specific activities and enantioselectivities to almost
all tested diaryl ketones. The specific activities of 50C10
toward 9a, 11a, 12a, 15a, and 16a were 1.02−7.54-fold that of
KpADH. The ee values of 50C10 were significantly increased,
especially toward 12a (97.5%, S), 13a (99.9%, S), 14a (99.9%,
S), 15a (87.7%, S), and 16a (99.9%, S) with large electron-
withdrawing substituents at para position, much higher than
12.3%−88.4% (S) of KpADH. Consequently, variant 50C10
could be potentially applied in the synthesis of optically pure
diaryl alcohols with halogen or other electron-withdrawing
substituents.
Figure 4. Asymmetric preparation of (R)-CPMA employing 50C10 at
⧫
■
●
different substrate loadings: ( ) 100 mM, ( ) 200 mM, ( ) 500
mM, (↓) addition of 200 mM and 100 mM substrate at 2.0 and 6.0 h.
Reactions were performed with 0.30 and 0.10 g dry cell weight of
50C10 and GDH, 2.0−10.0 mmol CPMK dissolved in 2 mL ethanol,
and 3.0−7.5 mmol glucose in 18 mL Tris-HCl (pH 7.5, 100 mM),
magnetically agitated at 200 rpm and 30 °C and maintained at pH 7.5
using 1.0 M NaOH. Samples were withdrawn at different time
intervals and analyzed by HPLC equipped with chiral OB-H column
at 254 nm.
from Bacillus megaterium (BmGDH) was introduced for
cofactor regeneration. Tris-HCl (pH 7.5, 100 mM) was used
as reaction buffer in which 50C10 and BmGDH retained
relatively matched activities. As an initial attempt, 100 mM 1a
(21.7 g·L−1) with 10% ethanol as cosolvent was completely
converted into (R)-1b within 2 h. Then, the substrate
concentration was increased to 200 mM 1a (43.4 g·L−1)
with the same amount of enzymes. The conversion rate
reached 72.1% at 2 h, and a full conversion was achieved at 6 h
with 99.5% ee and no external NADP+. Considering the
superior enantioselectivity and substrate tolerance of 50C10
compared with other reported diaryl ketone reductases, much
higher substrate loading was also investigated.7 Due to the low
solubility of 1a, fed-batch of substrate was employed. As shown
in Figure 4, 4.0, 4.0, and 2.0 mmol of 1a were added into the
reaction mixture (a final concentration of 500 mM) at 0, 2, and
6 h. After 12 h of agitation, 1a was fully reduced with
conversion of 99.7% and 99.4% ee. After extraction with ethyl
acetate and evaporation under vacuum, about 1.94 g of white
powder of (R)-1b was obtained with an isolation yield of
88.6%. The space−time yield reached 194 g·L−1·d−1 without
assistance of external NADP+. The product was confirmed by
NMR analysis (Figure S37−S38): 1H NMR (400 MHz,
CDCl3) δ 5.34 (s, 1H), 5.74 (s, 1H), 7.14 (d, J = 7.9 Hz, 1H),
7.23 (dd, J1 = 7.1 Hz, J2 = 5.1 Hz, 1H), 7.29−7.38 (m, 4H),
7.65 (td, J1 = 7.7 Hz, J2 = 1.7 Hz, 1H), 8.58 (d, J = 4.8 Hz,
1H); 13C NMR (75 MHz, CDCl3) δ 160.93, 148.00, 141.79,
137.12, 133.47, 128.68, 128.37, 122.65, 121.19, 77.66, 77.23,
76.81, 74.58. To the best of our knowledge, this is the first
report on the bioreductive synthesis of chiral antihistamine
precursor (R)-CPMA with >99% ee and over 100 g·L−1
substrate loading.
8342
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Activity Assay and Enantioselectivity Determination.
The activity of KpADH variants was spectrophotometrically
determined according to the optical changes of NADPH at 340
nm and 30 °C. Assay mixture contained 2.0 mM CPMK, 0.5
mM NADPH, and 10 μL enzyme in PBS buffer (100 mM, pH
6.0). One unit of activity was defined as the amount of enzyme
required for the depletion of 1.0 μmol NADPH at the above-
mentioned condition. Enantioselectivity was calculated accord-
ing to the peak area of (R)- and (S)-CPMA determined by
chiral HPLC equipped with Chiralcel OB-H column (4.6 mm
× 250 mm × 5 μm, Daicel Chiral Technologies Co., Ltd.) as
previously reported. All the assays were performed in triplicate.
Construction and Screening of HCSM Libraries.
Overlap extension PCR was adopted to obtain the whole
KpADH product with three mutational sites. Furthermore, the
products were ligated into linearized pET28a by ClonExpress
II (Vazyme Inc.). The resultant recombinant plasmids
containing multiple mutations at M131/F197/S237 or
C165/E214/Q238 were transformed into E. coli BL21(DE3).
Different pairs of primers were used to form different HCSM
libraries, and all the primers are listed in Table S2. About 600
monocolonies in each HCSM library were screened to achieve
95% coverage. Then all the strains were inoculated into 96-
deep well plates and cultivated as above-mentioned for the
expression of KpADH variants. After centrifugation, cells were
freeze−thawed at −80 °C followed by lysozyme and DNaseI
treatment to obtain crude extracts. Two rounds of screening
were adopted. First, the activities of all the variants were high
throughout determined as above-mentioned. Then, the
enantioselectivities of variants with similar or higher activity
than KpADH were determined.
Protein Purification and Preparation of Crude
Enzyme Powder. KpADH and variants with His-tag were
purified to homogeneity by nickel-affinity chromatography as
previously reported and analyzed by SDS-PAGE. Fermentation
of 50C10 was carried out with 2 × LB as fermentation
medium. Cells harboring 50C10 were harvested, washed with
physiological saline, and disrupted with nanohomogenize
machine (ATS AH-BASICI). The crude extract was obtained
by centrifugation and lyophilized under vacuum (SCIENTZ-
10N) to form the crude enzyme powder, which was stored at 4
°C for further use.
Determination of Kinetic Parameters and Substrate
Profiles. Kinetic parameters of KpADH and variants were
determined. The specific activities of each variant toward 0.1−
5.0 mM 1a were assayed employing the above-mentioned
protocol. KM, Vmax, and kcat were calculated according to the
Lineweaver−Burk plot. Substrate profiles of KpADH and
50C10 were measured toward prochiral ketones 1a−16a. The
final substrate concentration of each substrates was 2.0 mM.
The analytical methods for enantioselectivity of all tested
substrates are summarized in Table S12. All the activities and
enantioselectivities were measured in triplicate.
CONCLUSIONS AND PERSPECTIVES
■
In summary, hydroclassified combinatorial saturation muta-
genesis (HCSM) was proposed and applied in the directed
evolution of enantioselectivity of alcohol dehydrogenase
KpADH. Using this smart HCSM strategy, the library size
could be significantly reduced especially with simultaneous
mutation of three or more residues. The interactions including
hydrophobicity and steric hindrance in the substrate binding
pocket could be properly balanced employing the HCSM
strategy by dividing 20 amino acids into four groups. Six
hotspots critical for the enantioselectivity were identified in the
substrate binding pockets of KpADH by alanine/tyrosine
scanning. After two rounds of screening, a highly enantiose-
lective variant 50C10 was identified among 5500 variants.
50C10 displayed high binding affinity and catalytic efficiency
with kcat/KM of 59.3 s−1·mM−1, 3.51-fold that of KpADH.
Hydrophobic interaction in large substrate binding pocket and
steric hindrance in small substrate binding pocket were
proposed to be the main mechanisms for the improved
enantioselectivity as proved by MD simulation. Substrate
spectrum analysis revealed that 50C10 preferred diaryl ketones
and displayed improved enantioselectivity, especially toward
diaryl ketones with halogen or other electron-withdrawing
substituents. Variant 50C10 also demonstrated high efficacy
and substrate tolerance in the preparation of chiral diaryl
alcohols and could convert as much as 0.5 M (108.5 g·L−1) (4-
chlorophenyl)-(pyridin-2-yl)-methanone with 99.4% ee and
space−time yield of 194 g·L−1·d−1 without external NADP+.
This newly proposed HCSM strategy could be potentially
applied in the directed evolution of activity, substrate
spectrum, and enantioselectivity especially when lacking of
high throughput screening methods. This study also provides a
highly efficient and enantioselective ketoreductase for the
preparation of chiral diaryl alcohols.
MATERIALS AND METHODS
■
General Remarks. Plasmid pET28-kpadh was constructed
in our previous work. Prochiral ketones, (4′-chlorophenyl)-
(pyridin-2-yl)methanone (1a), 4′-chloroacetophenone (2a),
2′-chloroacetophenone (3a), 3′-chloroacetophenone (4a), 2-
chloro-1-(4′-chlorophenyl)ethanone (5a), 1-(pyridin-2′-yl)-
ethanone (6a), 1-(pyridin-3′-yl)ethanone (7a), 1-(pyridin-4′-
yl)ethanone (8a), benzophenone (9a), (pyridin-2′-yl)-
phenylmethanone (10a), (4′-fluorophenyl)phenylmethanone
(11a), (4′-chlorophenyl)phenylmethanone (12a), (4′-
bromophenyl)phenylmethanone (13a), (4′-nitrophenyl)-
phenylmethanone (14a), (4′-methoxyphenyl)-
phenylmethanone (15a), and 1,2-diphenylethanone (16a)
were purchased from Aladdin Inc.. NADPH was bought
from Merck Inc.
Site-Directed Mutagenesis and Protein Expression.
Site-directed mutagenesis of alanine/tyrosine scanning was
conducted by whole plasmid PCR using Phanta Super-Fidelity
DNA Polymerase (Vazyme Inc.). Primers are listed in Table
S1. PCR procedure was predenaturation at 96 °C for 5 min, 15
cycles of denaturation at 98 °C for 20 s, annealing at 50−65 °C
for 20 s and elongation at 68 °C for 210 s, and final elongation
at 68 °C for 10 min. The resultant PCR products were digested
with DpnI to remove the parental pET28-kpadh. Furthermore,
5.0 μL of digested products were transformed into E. coli
BL21(DE3). Expression of the KpADH variants was induced
with 0.2 mM IPTG at 25 °C for 6 h as previously reported.
Asymmetric Preparation of Chiral (R)-CPMA. Asym-
metric reduction of 1a for the preparation of (R)-CPMA was
performed with 0.30 and 0.10 g of crude enzyme powder of
50C10 and GDH, 2.0−10.0 mmol 1a dissolved in 2 mL of
ethanol, and 3.0−7.5 mmol glucose in 18 mL of Tris-HCl (pH
7.5, 100 mM). Reactions were magnetically agitated at 200
rpm and 30 °C and maintained at pH 7.5 using 1.0 M NaOH.
Samples were withdrawn from the reaction mixture at different
time points, and the conversion and enantioselectivity were
analyzed as above-mentioned. After the reaction, the mixture
8343
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
(2) (a) Botta, M.; Summa, V.; Corelli, F.; Di Pietro, G. D.;
Lombardi, P. Synthesis of Aryl 2-Benzofuranyl- and 2-Indolyl(aryl)-
carbinols of High Enantiomeric Purity via Palladium-catalyzed
Heteroannulation of Chiral Arylpropargylic Alcohols. Tetrahedron:
Asymmetry 1996, 7, 1263−1266. (b) Jen, W. S.; Truppo, M. D.;
Amos, D.; Devine, P.; McNevin, M.; Biba, M.; Campos, K. Copper-
catalyzed Synthesis of Enantioenriched Tetraarylethanes. Org. Lett.
2008, 10, 741−744.
(3) (a) Hashiguchi, S.; Fujii, A.; Takehara, J.; Ikariya, T.; Noyori, R.
Asymmetric Transfer Hydrogenation of Aromatic Ketones Catalyzed
by Chiral Ruthenium(II) Complexes. J. Am. Chem. Soc. 1995, 117,
7562−7563. (b) Yang, H. L.; Huo, N. N.; Yang, P.; Pei, H.; Lv, H.;
Zhang, X. M. Rhodium Catalyzed Asymmetric Hydrogenastion of 2-
Pyridine Ketones. Org. Lett. 2015, 17, 4144−4147. (c) Touge, T.;
Nara, H.; Fujiwhara, M.; Kayaki, Y.; Ikariya, T. Efficient Access to
Chiral Benzhydrols via Asymmetric Transfer Hydrogenation of
Unsymmetrical Benzophenones with Bifunctional Oxo-tethered
Ruthenium Catalysts. J. Am. Chem. Soc. 2016, 138, 10084−10087.
(4) (a) Bornscheuer, U. T.; Huisman, G. W.; Kazlauskas, R. J.; Lutz,
S.; Moore, J. C.; Robins, K. Engineering the Third Wave of
Bocatalysis. Nature 2012, 485, 185−194. (b) Honig, M.;
Sondermann, P.; Turner, N. J.; Carreira, E. M. Enantioselective
Chemo- and Biocatalysis: Partners in Retrosynthesis. Angew. Chem.,
Int. Ed. 2017, 56, 8942−8973. (c) Sheldon, R. A.; Woodley, J. M.
Role of Biocatalysis in Sustainable Chemistry. Chem. Rev. 2018, 118,
801−838.
(5) (a) Ma, S. K.; Gruber, J.; Davis, C.; Newman, L.; Gray, D.;
Wang, L.; Grate, J.; Huisman, G. W.; Sheldon, R. A. A Green-by-
Design Biocatalytic Process for Atorvastatin Intermediate. Green
Chem. 2010, 12, 81−86. (b) Savile, C. K.; Janey, J. M.; Mundorff, E.
C.; Moore, J. C.; Tam, S.; Jarvis, W. R.; Colbeck, J. C.; Krebber, A.;
Fleitz, F. j.; Brands, J.; Devine, P. N.; Huisman, G. W.; Hughes, G. J.
Biocatalytic Asymmetric Synthesis of Chiral Amines from Ketones
Applied to Sitagliptin Manufacture. Science 2010, 329, 305−309.
(6) (a) Chartrain, M.; Lynch, J.; Choi, W. B.; Churchill, H.; Patel, S.;
Yamazaki, S.; Volante, R.; Greasham, R. Asymmetric Bioreduction of
a Bisaryl Ketone to its Corresponding (S)-Bisaryl Alcohol, by the
Yeast Rhodotorula pilimanae ATCC 32762. J. Mol. Catal. B: Enzym.
2000, 8, 285−288. (b) Ni, Y.; Zhou, J. Y.; Sun, Z. H. Production of a
Aey Chiral Intermediate of Betahistine with a Newly Isolated
Kluyveromyces sp. in an Aqueous Two-Phase System. Process Biochem.
2012, 47, 1042−1048. (c) Xu, J. X.; Zhou, S. Y.; Zhao, Y. J.; Xia, J.;
Liu, X. Y.; Xu, J. M.; He, B. F.; Wu, B.; Zhang, J. F. Asymmetric
Whole-Cell Bioreduction of Sterically Bulky 2-Benzoylpyridine
Derivatives in Aqueous Hydrophilic Ionic Liquid Media. Chem. Eng.
J. 2017, 316, 919−927.
(7) (a) Truppo, M. D.; Pollard, D.; Devine, P. Enzyme-Catalyzed
Enantioselective Diaryl Ketone Reductions. Org. Lett. 2007, 9, 335−
338. (b) Li, H. M.; Zhu, D. M.; Hua, L.; Biehl, E. R. Enantioselective
Reduction of Diaryl Ketones Catalyzed by a Carbonyl Reductase from
Sporobolomyces salmonicolor and its Mutant Enzymes. Adv. Synth.
Catal. 2009, 351, 583−588.
(8) Zhou, J. Y.; Xu, G. C.; Han, R. Z.; Dong, J. J.; Zhang, W. G.;
Zhang, R. Z.; Ni, Y. Carbonyl Group-Dependent High-Throughput
Screening and Enzymatic Characterization of Diaromatic Ketone
Reductase. Catal. Sci. Technol. 2016, 6, 6320−6327.
(9) (a) Davids, T.; Schmidt, M.; Bottcher, D.; Bornscheuer, U. T.
Strategies for the Discovery and Engineering of Enzymes for
Biocatalysis. Curr. Opin. Chem. Biol. 2013, 17, 215−220. (b) Reetz,
M. T. Biocatalysis in Organic Chemistry and Biotechnology: Past,
Present, and Future. J. Am. Chem. Soc. 2013, 135, 12480−12496.
(c) Denard, C. A.; Ren, H. Q.; Zhao, H. M. Improving and
Repurposing Biocatalysts via Directed Evolution. Curr. Opin. Chem.
Biol. 2015, 25, 55−64. (d) Arnold, F. H. Directed Evolution: Bringing
New Chemistry to Life. Angew. Chem., Int. Ed. 2018, 57, 4143−4148.
(10) (a) Reetz, M. T.; Bocola, M.; Carballeira, J. D.; Zha, D.; Vogel,
A. Expanding the Range of Substrate Acceptance of Enzymes:
Combinatorial Active-Site Saturation Test. Angew. Chem., Int. Ed.
2005, 44, 4192−4196. (b) Bougioukou, D. J.; Kille, S.; Taglieber, A.;
was extracted three times with ethyl acetate. The organic phase
was combined and dried over anhydrous Na2SO4. Further-
more, the organic phase was evaporated under vacuum to
obtain (R)-1b as white powder.
Homology Modeling and Molecular Docking Anal-
ysis. Protein modeling was implemented with EasyModeller
4.0 using protein structures 4PVC and 5B6K as templates and
verified with SAVES (Structure Analysis and Verification
Server version 4). All docking calculations were accomplished
with AutoDock Vina 1.1, with the docking algorithm that took
account of ligand flexibility but kept the protein rigid. Docking
runs were carried out using standard parameters of the
program for interactive growing and subsequent scoring.
Molecular dynamic simulation was carried out employing the
NAMD module of Discover Studio 3.5. CHARMm force field
and temperature of 300 K were adopted, and the simulation
time was set as 20 ns. Stereoviews were constructed using
Pymol.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.8b02286.
■
S
Supporting tables and figures for primers, creation and
screening of HCSM libraries, enantioselectivity analysis,
original data of kinetic parameters and substrate profiles,
1H NMR and 13C NMR of synthesized (R)-CPMA, and
crystal structure of KpADH (5Z2X) (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: yni@jiangnan.edu.cn.
■
ORCID
Ye Ni: 0000-0003-4887-7517
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We are grateful to the National Natural Science Foundation of
China (21506073, 21776112), the Natural Science Foundation
of Jiangsu Province (BK20150003, BK20171135), Six Talent
Peaks project of Jiangsu Province (2015-SWYY-008), the
Program of Introducing Talents of Discipline to Universities
(111-2-06), the national first-class discipline program of Light
Industry Technology and Engineering (LITE2018-07), and a
project funded by the Priority Academic Program Develop-
ment of Jiangsu Higher Education Institutions for the financial
support of this research.
REFERENCES
■
(1) (a) Barouh, V.; Dall, H.; Patel, D.; Hite, G. Stereochemical
Aspects of Antihistamine Action. 4. Absolute Configuration of
Carbinoxamine Antipodes. J. Med. Chem. 1971, 14, 834−836.
(b) Meguro, K.; Aizawa, M.; Sohda, T.; Kawamatsu, Y.; Nagaoka,
A. New 1,4-Dihydropyridine Derivatives with Potent and Long-
Lasting Hypotensive Effect. Chem. Pharm. Bull. 1985, 33, 3787−3797.
(c) Casy, A. F.; Drake, A. F.; Ganellin, C. R.; Mercer, A. D.; Upton, C.
Stereochemical Studies of Chiral H-1 Antagonists of Histamine: The
Resolution, Chiral Analysis, and Biological Evaluation of Four
Antipodal Pairs. Chirality 1992, 4, 356−366. (d) Farina, V.; Reeves,
J. T.; Senanayake, C. H.; Song, J. J. Asymmetric Synthesis of Active
Pharmaceutical Ingredients. Chem. Rev. 2006, 106, 2734−2793.
8344
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345

ACS Catalysis
Research Article
Reetz, M. T. Directed Evolution of an Enantioselective Eoate-
Reductase: Testing the Utility of Iterative Saturation Mutagenesis.
Adv. Synth. Catal. 2009, 351, 3287−3305.
Analysis of 3,000 Mutations. Biotechnol. Bioeng. 2014, 111, 2380−
2389.
(18) (a) Filling, C.; Berndt, K. D.; Benach, J.; Knapp, S.;
̈
Prozorovski, T.; Nordling, E.; Ladenstein, R.; Jornvall, H.;
(11) (a) Neylon, C. Chemical and Biochemical Strategies for the
Randomization of Protein Encoding DNA Sequences: Library
Construction Methods for Directed Evolution. Nucleic Acids. Res.
2004, 32, 1448−1459. (b) Shivange, A. V.; Marienhagen, J.;
Mundhada, H.; Schenk, A.; Schwaneberg, U. Advances in Generating
Functional Diversity for Directed Protein Evolution. Curr. Opin.
Chem. Biol. 2009, 13, 19−25. (c) Reetz, M. T.; Krebs, G. P. L.
Challenges in the Directed Evolution of Stereoselective Enzymes for
Use in Organic Chemistry. C. R. Chim. 2011, 14, 811−818. (d) Feng,
X. J.; Sanchis, J.; Reetz, M. T.; Rabitz, H. Enhancing the Efficiency of
Directed Evolution in Focused Enzyme Libraries by the Adaptive
Substituent Reordering Algorithm. Chem. - Eur. J. 2012, 18, 5646−
5654. (e) Reetz, M. T. Laboratory Evolution of Stereoselective
Enzymes as a Means to Expand the Toolbox of Organic Chemists.
Tetrahedron 2012, 68, 7530−7548.
Oppermann, U. Critical Residues for Structure and Catalysis in
Short-Chain Dehydrogenases/Reductases. J. Biol. Chem. 2002, 277,
25677−25684. (b) Deng, J.; Yao, Z. Q.; Chen, K. L.; Yuan, Y. A.; Lin,
J. P.; Wei, D. Z. Towards the Computational Design and Engineering
of Enzyme Enantioselectivity: A CaseStudy by a Carbonyl Reductase
from. J. Biotechnol. 2016, 217, 31−40.
(19) (a) Schlieben, N. H.; Niefind, K.; Muller, J.; Riebel, B.;
̈
Hummel, W.; Schomburg, D. Atomic Resolution Structures of R-
Specific Alcohol Dehydrogenase from Lactobacillus brevis Provide the
Structural Bases of its Substrate and Cosubstrate Specificity. J. Mol.
́
Biol. 2005, 349, 801−813. (b) Noey, E. L.; Tibrewal, N.; Jimenez-
́
Oses, G.; Osuna, S.; Park, J. Y.; Bond, C. M.; Cascio, D.; Liang, J.;
Zhang, X. Y.; Huisman, G. W.; Tang, Y.; Houk, K. N. Origins of
Stereoselectivity in Evolved Ketoreductases. Proc. Natl. Acad. Sci. U. S.
A. 2015, 112, E7065−E7072.
(12) (a) Lutz, S.; Patrick, W. M. Novel Methods for Directed
Evoltuion of Enzymes: Quality, not Quantity. Curr. Opin. Biotechnol.
2004, 15, 291−297. (b) Pines, G.; Pines, A.; Garst, A. D.; Zeitoun, R.
I.; Lynch, S. A.; Gill, R. T. Codon Compression Algorithms for
Saturation Mutagenesis. ACS Synth. Biol. 2015, 4, 604−614.
(13) (a) Reetz, M. T. Laboratory Evolution of Stereoselective
Enzymes: A Profilic Source of Catalysts for Asymmetric Reactions.
Angew. Chem., Int. Ed. 2011, 50, 138−174. (b) Reetz, M. T.;
Kahakeaw, D.; Lohmer, R. Addressing the Numbers Problem in
Directed Evolution. ChemBioChem 2008, 9, 1797−1804.
(20) (a) Guo, P. C.; Bao, Z. Z.; Ma, X. X.; Xia, Q.; Li, W. F.
Structural Insights into the Cofactor-Assisted Substrate Recognition
of Yeast Methylglyoxal/isovaleraldehyde Reductase Gre2. Biochim.
Biophys. Acta, Proteins Proteomics 2014, 1844, 1486−1492. (b) Qin, F.
Y.; Qin, B.; Mori, T.; Wang, Y.; Meng, L. X.; Zhang, X.; Jia, X.; Abe,
I.; You, S. Engineering of Candida glabrata Ketoreductase 1 for
Asymmetric Reduction of Alpha-halo Ketones. ACS Catal. 2016, 6,
6135−6140.
(21) Cunningham, B. C.; Wells, J. A high-Resolution Epitope
Mapping of hGH-Receptor Interactions by Aanine-Scanning Muta-
genesis. Science 1989, 244, 1081−1085.
(14) (a) Kamtekar, S.; Schiffer, J. M.; Xiong, H.; Babik, J.; Hecht, M.
H. Protein Design by Binary Patterning or Polar and Nonpolar Amino
Acids. Science 1993, 262, 1680−1685. (b) Roy, S.; Hecht, M. H.
Cooperative Thermal Denaturation of Proteins Designed by Binary
Patterning of Polar and Nonpolar Amino Acids. Biochemistry 2000,
39, 4603−4607. (c) Sun, Z. T.; Salas, P. T.; Siirola, E.; Lonsdale, R.;
Reetz, M. T. Exploring Productive Sequence Space in Directed
Evolution Using Binary Patterning versus Conventional Mutagenesis
Strategies. Bioresour. Bioprocessing 2016, 3, 44. (d) Li, G. Y.; Reetz,
M. T. Learning Lessons from Directed Evolution of Stereoselective
Enzymes. Org. Chem. Front. 2016, 3, 1350−1358. (e) Reetz, M. T.
Combinatorial Libraries Reloaded. Isr. J. Chem. 2018, 58, 52−60.
(15) (a) Hughes, M. D.; Nagel, D. A.; Santos, A. F.; Sutherland, A.
J.; Hine, A. V. Removing the Redundancy from Randomized
GeneLibraries. J. Mol. Biol. 2003, 331, 973−979. (b) Reetz, M. T.;
Wu, S. Greatly Reduced Amino Acid Alphabets in Directed Evolution:
Making the Right Choice for Saturation Mutagenesis at Homologous
Enzyme Position. Chem. Commun. 2008, 43, 5499−5501. (c) Reetz,
M. T.; Prasad, S.; Carballeira, J. D.; Gumulya, Y.; Bocola, M. Iterative
Saturation Mutagenesis Accelerates Laboratory Evolution of Enzyme
Stereoselectivity: Rigorous Comparison with Traditional Methods. J.
Am. Chem. Soc. 2010, 132, 9144−9152. (d) Kille, S.; Acevedo-Rocha,
C. G.; Parra, L. P.; Zhang, Z. G.; Opperman, D. J.; Reetz, M. T.;
Acevedo, J. P. Reducing Codon Redundancy and Screening Effort of
Combinatorial Protein Libraries Created by Saturation Mutagenesis.
ACS Synth. Biol. 2013, 2, 83−92.
(22) (a) Sun, Z. T.; Lonsdale, R.; Li, G. Y.; Reetz, M. T. Comparing
Different Strategies in Directed Evolution of Enzyme Stereo-
selectivity: Single- versus Double-Codon Saturation Mutagenesis.
ChemBioChem 2016, 17, 1865−1872. (b) Sun, Z. T.; Wikmark, Y.;
Backvall, J. E.; Reetz, M. T. New Concepts for Increasing the
Efficiency in Directed Evolution of Stereoselective Enzymes. Chem. -
Eur. J. 2016, 22, 5046−5054.
(23) Luetz, S.; Giver, L.; Lalonde, J. Engineered Enzymes for
Chemical Production. Biotechnol. Bioeng. 2008, 101, 647−653.
(16) (a) Sun, Z. T.; Lonsdale, R.; Kong, X. D.; Xu, J. H.; Zhou, J. H.;
Reetz, M. T. Reshaping an Enzyme Binding Pocket for Enhanced and
Inverted Stereoselectivity: Use of Smallest Amino Acid Alphabets in
Directed Evolution. Angew. Chem., Int. Ed. 2015, 54, 12410−12415.
(b) Sun, Z. T.; Lonsdale, R.; Wu, L.; Li, G. Y.; Li, A. T.; Wang, J. B.;
Zhou, J. H.; Reetz, M. T. Structure-Guided Triple Code Saturation
Mutagenesis: Efficient Tuning of the Stereoselectivity of an Epoxide
Hydrolase. ACS Catal. 2016, 6, 1590−1597. (c) Sun, Z. T.; Lonsdale,
R.; Ilie, A.; Li, G. Y.; Zhou, J. H.; Reetz, M. T. Catalytic Asymmetric
Reduction of Difficult-to-Reduce Ketones: Triple Code Saturation
Mutagenesis of Alcohol Dehydrogenase. ACS Catal. 2016, 6, 1598−
1605.
(17) Zhao, J.; Kardashliev, T.; Ruff, A. J.; Bocola, M.; Schwaneberg,
U. Lessons from Diversity of Directed Evolution Experiments by an
8345
DOI: 10.1021/acscatal.8b02286
ACS Catal. 2018, 8, 8336−8345