Structural and Computational Insight into the Catalytic Mechanism of Limonene Epoxide Hydrolase Mutants in Stereoselective Transformations

Article abstract of DOI:10.1021/jacs.7b10278

Directed evolution of limonene epoxide hydrolase (LEH), which catalyzes the hydrolytic desymmetrization reactions of cyclopentene oxide and cyclohexene oxide, results in (R,R)- and (S,S)-selective mutants. Their crystal structures combined with extensive theoretical computations shed light on the mechanistic intricacies of this widely used enzyme. From the computed activation energies of various pathways, we discover the underlying stereochemistry for favorable reactions. Surprisingly, some of the most enantioselective mutants that rapidly convert cyclohexene oxide do not catalyze the analogous transformation of the structurally similar cyclopentene oxide, as shown by additional X-ray structures of the variants harboring this slightly smaller substrate. We explain this puzzling observation on the basis of computational calculations which reveal a disrupted alignment between nucleophilic water and cyclopentene oxide due to the pronounced flexibility of the binding pocket. In contrast, in the stereoselective reactions of cyclohexene oxide, reactive conformations are easily reached. The unique combination of structural and computational data allows insight into mechanistic details of this epoxide hydrolase and provides guidance for future protein engineering in reactions of structurally different substrates.

Full text of DOI:10.1021/jacs.7b10278

Subscriber access provided by READING UNIV
Article
Structural and Computational Insight into the Catalytic Mechanism of
Limonene Epoxide Hydrolase Mutants in Stereoselective Transformations
Zhoutong Sun, Lian Wu, Marco Bocola, H.C. Stephen Chan, Richard
Lonsdale, Xu-Dong Kong, Shuguang Yuan, Jiahai Zhou, and Manfred T. Reetz
J. Am. Chem. Soc., Just Accepted Manuscript • DOI: 10.1021/jacs.7b10278 • Publication Date (Web): 12 Dec 2017
Downloaded from http://pubs.acs.org on December 12, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Structural and Computational Insight into the Catalytic Mech-
anism of Limonene Epoxide Hydrolase Mutants in Stereose-
lective Transformations
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Zhoutong Sun†#, Lian Wu‡#, Marco Bocola§ξ, H.C. Stephen Chan ¶, Richard Lonsdale§ξ, XuꢀDong Kong
‡, Shuguang Yuan ¶*, Jiahai Zhou‡*, and Manfred T. Reetz†§ξ*
†Tianjin Institute of Industrial Biotechnology, Chinese Academy of Sciences, 32 West7th Avenue, Tianjin Airport Economic
Area, Tianjin 300308, China
‡State Key Laboratory of Bioꢀorganic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese
Academy of Sciences, Shanghai 200032, China
§MaxꢀPlanckꢀInstitut für Kohlenforschung, KaiserꢀWilhelmꢀPlatz 1, 45470 Mülheiman der Ruhr, Germany
ξFachbereich Chemie der PhilippsꢀUniversität, HansꢀMeerweinꢀStrasse, 35032 Marburg, Germany
¶Laboratory of Physical Chemistry of Polymers and Membranes, Ecole Polytechnique Fédérale de Lausanne (EPFL),CH B3
495 (Bâtiment CH) Station 6, CHꢀ1015 Lausanne, Switzerland
ABSTRACT: Directed evolution of limonene epoxide hydrolase (LEH), which catalyzes the hydrolytic desymmetrization reactions of
cyclopentene oxide and cyclohexene oxide, results in (R,R)ꢀ and (S,S)ꢀselective mutants. Their crystal structures combined with extensive
theoretical computations shed light on the mechanistic intricacies of this widely used enzyme. From the computed activation energies of
various pathways, we discover the underlying stereochemistry for favorable reactions. Surprisingly, some of the most enantioselective
mutants that rapidly convert cyclohexene oxide do not catalyze the analogous transformation of the structurally similar cyclopentene oxide,
as shown by additional Xꢀray structures of the variants harboring this slightly smaller substrate. We explain this puzzling observation on
the basis of computational calculations which reveal a disrupted alignment between nucleophilic water and cyclopentene oxide due to the
pronounced flexibility of the binding pocket. In contrast, in the stereoselective reactions of cyclohexene oxide, reactive conformations are
easily reached. The unique combination of structural and computational data allows insight into mechanistic details of this epoxide hydroꢀ
lase and provides guidance for future protein engineering in reactions of structurally different substrates.
with the aim of learning more about this widely used enzyme. The
hydrolytic desymmetrization reactions of the cyclic epoxides 1
(dubbed CYO1) and 3 (dubbed CYO3), respectively, were used as
the model LEHꢀcatalyzed transformations (Scheme 1). Wildtype
(WT) LEH is a poor catalyst in both reactions (13% ee favoring
(R,R)ꢀ2, and 4% ee favoring (S,S)ꢀ4). These transformations were
previously employed as the experimental platform for developing
INTRODUCTION
In studies focusing on the elucidation of enzyme mechanisms,
stereoselectivity has traditionally served as a unique probe.1 More
recently, the rapidly growing area of directed evolution of stereꢀ
oselective enzymes as practical catalysts in organic chemistry and
biotechnology continues to generate a huge set of potentially
useful data,2 which in principle can be used to deepen our
knowledge of the intricacies of enzyme mechanisms. In endeavors
of this kind, crystal structures of enantioselective mutants,3 ideally
complemented by kinetic studies and molecular dynamics (MD)
computations, are particularly revealing. Such an approach is
especially powerful when the directed evolution study includes Xꢀ
ray structures of both (+)ꢀ and (ꢀ)ꢀselective mutants which enable
access to both enantiomeric products.3a Any notable increase or
decrease in activity (in the extreme case complete shutdown) also
constitutes a handle for studying mechanisms.
Scheme 1. LEHꢀcatalyzed hydrolytic desymmetrization of epoxꢀ
ides 1 and 3.
Epoxide hydrolases (EHs) occur widely in many organisms
and plants, catalyzing the hydrolysis of epoxides with formation
of the corresponding 1,2ꢀdiols. The respective biological functions
vary according to the organism of origin, which include biosynꢀ
thesis of natural products, detoxification of toxic epoxides, and
cellular signaling.4 LEHs have also been used extensively in
organic chemistry for stereoselective synthesis of chiral vicinal
diols (and also of epoxides).5
Scheme 2. The proposed catalytic mechanism of LEH6a,11
In the present study, we focus on limonene epoxide hydrolase
LEH)6 and utilize extensive information obtained from stereoseꢀ
lective mutants, kinetics, Xꢀray analyses and MD computations
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 9
Table 1. The best LEH variants as catalysts in the hydrolytic desymmetrization of substrates 1 and 3.
1
2
3
4
1
3
Variant
Sequence
Source
ee%
c%
ee%
c%
5
6
7
8
WT
13(R,R) 99
4(S,S)
99
99
83
97
99
97
98
SZ92
L74F/M78F/L103V/L114V/I116V/F139V/L147V
L74F/M78V/I80V/L114F
72(S,S)
ꢀ
6
92(S,S)
96(R,R)
99(S,S)
97(R,R)
95(S,S)
97(S,S)
SCSM/ISM3a
SZ338
SZ348
SZ529
SZ718
SZ719
ꢀ
I80Y/L114V/I116V
ꢀ
≤5
8
TCSM7a
TCSM/ISM7a
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
M32V/M78V/I80V/L114F
10(R,R)
92(S,S)
92(S,S)
L74F/I80F/L114V/I116V/F139V
L74F/M78F/I80F/L114V/I116V/F139V
56
85
ISM/
This study
Figure 1. The overall structure and flexibility of LEH. (A) Cartoon structure of LEH. The α-helixes and β-strands were colored by red and
blue respectively. Yellow spheres indicate the location of catalytic site. (B) The sequence secondary structure of LEH. (C) The intrinsic
flexibility of LEH calculated from PCA analysis. The cyan vector length correlates with the domainꢀmotion scale. Yellow spheres indicate
the location of catalytic site.
optimal directed evolution strategies which ensure the generation
of highꢀquality mutant libraries with minimal screening effort3a,7
(bottleneck of directed evolution).8 Among them, single code
saturation mutagenesis (SCSM)3a based on only one amino acid as
building block in saturation mutagenesis at relatively large ranꢀ
domization sites, and triple code saturation mutagenesis (TCSM)7a
employing three amino acids, combined with iterative saturation
mutagenesis (ISM),9 were systematically explored. TCSM proved
to be the most efficient method,7a leading to highꢀquality mutant
libraries which require limited screening.8 Relevant is a recent
Rosettaꢀbased study on computationally designed LEH mutants as
catalysts in the desymmetrization of cyclopentene oxide (1) and
cyclohexene oxide (3), requiring the screening of only 37 variants
with the identification of mutants showing moderate to proꢀ
nounced enantioselectivity.10
cyclohexene oxide 3 (CYO3), but not of cyclopentene oxide 1
(CYO1).3a In principle, CYO1 is more strained than CYO3 and
can be expected to be hydrolyzed faster. Although the crystal
structures of WT LEH and several mutants have been determined,
which in some cases help to explain the source of stereoselectivity
qualitatively, they are of limited use in revealing the observed
drastic switches in substrate specificity.3a As already pointed out,
previous directed evolution studies led to moderate to excellent
improvements in (R,R)ꢀ as well as (S,S)ꢀselectivity for both subꢀ
strates (Table 1). To understand why some LEH variants are
specifically active toward CYO1, whereas others are selective for
CYO3, and why stereoselectivity is achieved, we crystalized three
LEH mutants (SZ338, SZ348 and SZ529) in complex with CYO1
or its diol product. Coupled with new mutagenesis experiments,
molecular docking, MD simulation as well as quantum mechaniꢀ
cal (QM) calculations, the overall results shed new light on mechꢀ
anistic details of LEHꢀcatalyzed
Surprisingly, in our study some of the evolved highly stereoseꢀ
lective LEH mutants were found to catalyze the rapid reaction of
ACS Paragon Plus Environment

Page 3 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. Xꢀray electron density map of LEH variants complex for (A) SZ529, (B) SZ348 and (C) SZ338. Black grid: 2FoꢀFc at 1.2σ level.
Green grid: FoꢀFc at 3.2σ level. Red grid: FoꢀFc at ꢀ3.2σ level. White sticks: side chain of key residues in the catalytic site of LEH. The
inactive substrate CYO1 (yellow ballsꢀandꢀsticks) is in multipleꢀconformation state in the complex structures of SZ529ꢀCYO1 and SZ348ꢀ
CYO1, as well as in the resolved complex structures of SZ338ꢀCYO1.
stereoselectivity transformations. The basic mechanistic features
derived from an earlier structural investigation6a are shown in
Scheme 2, which are supported by a recent QM study.11
MD trajectory of Apo WT LEH (pdb: 1NU3)6a indicates that these
three motifs can undergo noticeable movements, that change both
the shape and volume of the catalytic zone (Figure 1 and movie
S1). This phenomenon was also confirmed by the rootꢀmeanꢀ
square fluctuation (RMSF) of LEH, which was calculated from
the MD simulation trajectory (Figure S1).
RESULTS AND DISCUSSION.
Newly Evolved LEH Mutants Improve Enantioselectivity in
Reactions of Substrate CYO1. In order to answer the puzzling
question why some LEH mutants are completely inactive against
CYO1, yet drastically enhancing enantioselectivity in the desymꢀ
metrization of CYO3, we screened several mutant libraries previꢀ
ously evolved for CYO3 (Table S1). Several variants were found
to be active against CYO1, and had already been identified in the
original libraries using CYO3 as substrate, e.g., SZ80
(L114V/I116V/F139V). However, their enantioselectivities were
found to be poor (66% ee in favor of (S,S)ꢀ2, 80% conversion in
the case of SZ80). To optimize enantioselectivity, SZ80 was used
as a template for iterative saturation mutagenesis (ISM)9 at the 5ꢀ
residue randomization site lining the binding pocket
L74/M78/I80/L103/L147 using SCSM based on phenylalanine as
the only building block. Two optimal mutants SZ718
In the complex structures of SZ529ꢀCYO1 and SZ348ꢀCYO1
(Figure 2), we found that the inactive substrate CYO1 is quite
flexible in the binding site, adopting two different conformations
which fit both 2FoꢀFc and FoꢀFc map very well. Interestingly, a
conserved water molecule between D132 and Y53 (Figure 2B
and 2C), which is considered to be essential for the catalytic
step,6a has been expelled from the catalytic center (Figure 2A). In
our previously published crystal structures of SZ338 and SZ348,
the OsubstrateꢀOD101 and OsubstrateꢀOY53 distances appear too short
(2.2ꢀ2.4Å) for a typical hydrogenꢀbond interaction (2.7ꢀ3.2Å).3a,6a
Considering the low occupancy of CYO1 in the asymmetric units
of SZ338, we therefore reꢀrefined the crystal structures of SZ338ꢀ
CYO1 (Figure 2C) introducing multiple orientations for the
CYO1 molecule. In all crystal structures, the CYO1 molecules in
various orientations fit the electron density maps perfectly in both
SZ338ꢀCYO1 and SZ348ꢀCYO1. No close contact is observed for
both crystal water and CYO1 molecules in the catalytic regions.
Moreover, the highly conserved water molecules in both SZ338ꢀ
CYO1 and SZ348ꢀCYO1 structures are maintained between D132
and Y53 (Figure 2B and 2C). Furthermore, the orientations of
CYO1 in SZ338ꢀCYO1, SZ348ꢀCYO1 and SZ529ꢀCYO1 are
different from each other, confirming that CYO1 rotates freely in
the binding pocket. Such high flexibility of CYO1 hinders perfect
alignment of activated water and substrate in the transition state,
as also indicated in the crystal structures, is responsible for its
inactivity towards SZ338, SZ348 and SZ529.
(L74F/I80F/L114V/I116V/F139V)
and
SZ719
(L74F/M78F/I80F/L114V/I116V/F139V) were discovered, showꢀ
ing a selectivity of 92% ee (S,S) and conversion amounting to
56~85%. More importantly, these two variants can also hydrolyze
the larger CYO3 with eeꢀvalues of 95% and 97% respectively,
yielding (S,S)ꢀ4. For comparison purpose, the present and previꢀ
ous results are summarized in Table 1.
Crystal Structures and MD Simulations Reveal Flexibility of
LEH. In order to throw light on the origin of enhanced LEH
enantioselectivity, we determined the crystal structures of mutants
SZ529 and SZ348 (Table S2), each bound with inactive substrate
CYO1 (pdb: 5YAO，5YNG). We also reꢀsolved the previously
reported complex structure of SZ338 with CYO1 as inert “guest”
in the binding pocket (pdb: 5YQT)3a by removing all suspicious
water molecules and adding multiple ligand conformations. The
overall structure of LEH comprises four αꢀhelices and six βꢀ
strands, which fold into a compact catalytic pocket (Figure 1).
The catalytic trait of LEH is composed of Y53, N55, R99, D101
and D132 (Figure 2). Both Y53 and N55 are located in β1,
whereas both R99 and D101 reside in β4. D132 is in β6. Moreoꢀ
ver, we found three flexible regions within the vicinity of the
substrate binding pocket, the Cꢀterminal loop (C loop), helix 4
(H4) and loop A. Principal component analysis (PCA)12 on the
To further confirm these assessments, we performed extensive
allꢀatom MD simulations for the complexes SZ338ꢀCYO1,
SZ338ꢀCYO3, SZ348ꢀCYO1, SZ348ꢀCYO3, SZ529ꢀCYO1,
SZ529ꢀCYO3, SZ719ꢀCYO1 and SZ719ꢀCYO3. The CYO3
complexes were constructed by molecular docking. It is worth
noting that mutants SZ338, SZ348 and SZ529 are either inactive
or show extremely low activity toward CYO1. In all cases, CYO1
reꢀorientates
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. The Substrate binding modes and proteinꢀligand interaction fingerprint for (A) SZ529ꢀCYO3, (B) SZ719ꢀCYO1 and (C) SZ719ꢀ
CYO3. Dash: hydrogenꢀbond interactions. Yellow ballsꢀandꢀsticks: substrate molecule. White sticks: side chains of key residues in the
catalytic pocket. Proteinꢀligand interaction fingerprint analysis (IFP) are indicated by cyan radar plots. The statistical interaction frequency
in the IFP plot was obtained from calculating the proteinꢀligand interaction frequency in the MD simulations.
Interestingly, we also found that the size and shape of the
catalytic site in each mutant are noticeably different. The volumes
of the catalytic regions are 107 ± 1 Å3, 98 ± 1 Å3 and 102 ± 1 Å3 for
mutants SZ338, SZ348 and SZ529 respectively, whereas the
volume of CYO1 is only 62 ± 1 Å3. Such noticeable differences
are responsible for the CYO1 flexibilities in both crystal strucꢀ
tures and allꢀatom MD simulations (Table S3). However, the
volume of SZ719 is much more limited, with a value of 86 ± 1 Å3
which can accommodate CYO1 in a more stable conformation.
The computed volume of the CYO3 molecule amounts to ~75 ± 1
Å3, which results in a much more stable interaction within all
mutants.
Combining the observations from both crystal structures and
MD simulations, we conclude that the volume of the inactive LEH
variants are too big to stabilize a proper near attack conformation
(NAC)13 for the smaller CYO1 substrate prior to reaching the
transition state. In such an environment, CYO1 continuously
alters its conformations in the catalytic site. However, with a
Figure 4. The calculated activation energy barrier and substrateꢀ
limited space, as in the case of mutant SZ719, both CYO1 and
water distances of SZ719. (A) The activation energy barrier of
CYO3 are held in active poses, leading to (S,S)ꢀ2 and (S,S)ꢀ4,
SZ719ꢀCYO1. Distance in blue was calculated by QM methods.
respectively.
(B) Distances of C1ꢀO2 and C2ꢀO2 along MD simulations of
Shedding Further Light on the Active Substrate Binding
Mode of LEH. In order to elaborate the atomic details of active
substrateꢀLEH interactions, we performed molecular docking and
allꢀatom MD simulations and proteinꢀligand interaction fingerꢀ
print analysis (IFP) (Figure 3 and Figure S4) for active subꢀ
strates in complex with LEH, including the complexes SZ338ꢀ
CYO3, SZ348ꢀCYO3, SZ529ꢀCYO3, SZ719ꢀCYO1 and SZ719ꢀ
CYO3. In each case, the oxygen atom in the epoxide consistently
establishes a strong hydrogenꢀbond interaction with the protonatꢀ
ed
SZ719ꢀCYO1. (C) The activation energy barrier of SZ719ꢀCYO3.
Distance in blue was calculated by QM methods. (D) Distances of
C1ꢀO2 and C2ꢀO2 along MD simulations of SZ719ꢀCYO3
itself freely in the trajectories. This is suggested by the higher
rootꢀmeanꢀsquare deviations (RMSD) of both protein and that of
CYO1 (Figure S2 and S3, movie S2-S4). In contrast, SZ719 is
active on CYO1, which remains stable in the catalytic zone with a
very low RMSD value (Figure S3, movie S5). Identically, all
studied mutants are active on CYO3, which also exhibits a more
confined orientation in each case (Figure S3, movies S6-9).
ACS Paragon Plus Environment

Page 5 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 5. The basis of LEH catalytic mechanism and stereoselectivity. (A) When the substrate approaches LEH, domain H4, C loop and
loop A can undergo noticeable change, opening a space for the substrate entrance. A highly conserved water molecule is stabilized by a
dedicated hydrogenꢀbond network, formed by Y53, N55 and D132. (B) Once the substrate sits in the catalytic site, the conserved water
molecule can attack the epoxide ring in two different ways: at carbon C1 and C2. (C) If the distance between the water oxygen atom and
C1 (d1) is shorter than that of C2 (d2), formation of (R,R)ꢀproducts is favored, due to the lower activation energy. If d1>d2, LEH will
produce (S,S)ꢀproducts.
D101,6a which is involved in a dedicated hydrogenꢀbond network
with R99, D132, W130, Y53, N55 and a conserved water moleꢀ
cule W1. Among the LEH variants, different mutations in the
vicinity of catalytic sites alter the hydrophobic interaction proꢀ
files. In SZ338, hydrophobic interactions are found between
CYO3 and L35ꢀL103ꢀW130ꢀF134, whereas CYO3 interacts with
Y80ꢀL103ꢀL114ꢀF134 in SZ348. In SZ529, identical interactions
with L103ꢀI116ꢀF134ꢀF139 are also observed. Both CYO1 and
CYO3 are hydrolyzed in SZ719. The IFPs of SZ719ꢀCYO1 and
SZ719ꢀCYO3 are similar to each other, both of them interact with
F78ꢀF80ꢀL103ꢀV116ꢀF134ꢀV139, although the interaction freꢀ
quency of each residue varies slightly. Interestingly, the number
of residues involved in the IFP of SZ719 (7 residues) is much
higher than that of SZ338, SZ348 and SZ529 (5 residues). This is
mainly because SZ719 has a much smaller catalytic site (volume
= 86 ± 1 Å3) than any other variants (volumes = 98 ꢀ 107 ± 1 Å3).
This enables SZ719 to undergo more compact interactions with
substrates.
values<10ꢀ100. This shows that (nearly) perfect alignment of the
nucleophilic water and substrate, determined by the mutations, is
necessary for a smooth reaction to occur with high enantioselecꢀ
tivity.
The activation energies from QM calculations show that they
are correlated with these distances. The shorter distance d2 (O2ꢀ
C2) generates a lower activation energy (Ea): 13.4 kcal/mol for
SZ179ꢀCYO1 and 14.1 kcal/mol for SZ719ꢀCYO3, whereas the
longer distance d1 (O2ꢀC1) results in much higher activation
energies: 19.3 kcal/mol for SZ179ꢀCYO1 and 20.5 kcal/mol for
SZ719ꢀCYO3. The catalytic processes and covalent bond changes
can be visualized in movies S10-12.
We extended our QM studies to three other systems: SZ338,
SZ348 and SZ529. None of them can hydrolyze CYO1 because of
their high flexibilities in the catalytic binding pocket, although
SZ529 shows a very low activity (c%=0.08). Thus, we only simuꢀ
lated these three variants bound with respect to CYO3. Similar to
SZ719, SZ348 (Figure S5) also yields (S,S)ꢀ4. In this case nucleꢀ
ophilic water W1 favors attack at the C2 (d2=3.30 Å) atom in
CYO3 with an Ea of 14.3 kcal/mol, whereas the Ea is as high as
18.6 kcal/mol if W1 attacks C1 (d1=4.36 Å). In contrast, W1 in
SZ338 (Figure S5) prefers to attack C1 (d1=3.28Å), with an Ea of
12.9 kcal/mol yielding (R,R)ꢀ4, whereas the Ea is as high as 16.5
kcal/mol if W1 attacks C2 (d2=3.45 Å). Identically, SZ529 also
leads to preferential reaction at C1 (d1=3.35 Å) with an Ea of 11.0
kca/mol, whereas Ea is as high as 15.7 kcal/ml if W1 attacks C2
(d2=3.65Å).
Revealing the Source of Stereoselectivity. Quantum mechanics,
especially the density functional theory (DFT), is an accurate and
efficient method for the study of enzyme mechanisms.14 In order
to understand the origin of evolved enantioselectivity, a large
activeꢀsite model (280ꢀ300 atoms) based on crystal structures was
designed for a QM study. In this model, we include a highly
conserved water molecule, the substrate, and residues R99, W130,
Y53, D132, V116, F134, N55, D101, V114, F80, L103, F78,
V143, V139 as well as F74 in the corresponding positions of
mutant SZ719. The geometry of each catalytic site was submitted
to geometry optimization by QM method prior to transition state
scan. Mutant SZ719 is capable of hydrolyzing both CYO1 and
CYO3, yielding (S,S)ꢀ2 and (S,S)ꢀ4 molecules, respectively
(Scheme 1 and Figure 4). All LEH variants, in principle, can
generate two different stereoisomers (Schemes 1 and 2). If the
highly conserved water molecule W1 (Figure 4) attacks the C2
carbon of the substrate, (S,S)ꢀ2 or (S,S)ꢀ4 are generated for CYO1
and CYO3, respectively. In contrast, if W1 (Figure 4) attacks the
C1 carbon of the substrate, the enantiomers (R,R)ꢀ2 or (R,R)ꢀ4 are
generated. We succeeded in gaining insight into the origin of
stereoselectivity as follows: In both complexes SZ179ꢀCYO1 and
SZ719ꢀCYO3, QM calculations and statistics on the distances of
d1 (O2ꢀC1) (corresponding to formation of (R,R)ꢀ2) and d2 (O2ꢀ
C2) (corresponding to formation of (S,S)ꢀ4) from MD simulations
indicate that d1 is noticeably longer than d2, with statistical pꢀ
It is well known that the “angle of attack” (AA) in SN2 reacꢀ
tions is critical for a smooth transformation to occur.14 Therefore,
we calculated the AA for both QM (Table S4) and MD simulaꢀ
tions (Figure S6). In the present case involving an epoxide, the
AA is defined by three atoms (Figure S6A): the oxygen from
highly conserved water (Ow), reactive carbon (C1/C2) in the
substrate and the oxygen atom (O1) in the substrate. The AA for
each active substrate in the QM calculation is within the range of
145ꢀ155°, indicating that OwꢀC1/C2ꢀO1 are almost on the same
line which is known to be most favorable for a SN2 reactions. This
observation is also in a good agreement with another previously
reported SN2 enzyme process in a haloalkane dehalogenase proꢀ
cess.15 To further confirm our results, we also considered the
statistics of our MD simulations for AA (Figure S6). The AA for
a favored reaction always shows a value facilitating the SN2 cataꢀ
lytic reaction: SZ338ꢀCYO3 (148 ± 0.1°), SZ348ꢀCYO3 (147 ±
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 9
0.6°), SZ529ꢀCYO3 (149 ± 0.1°), SZ719ꢀCYO1 (148 ± 0.1°) and
into a solution containing 10% glycerol, 2.4 M sodium/potassium
phosphate, 0.1 M TrisꢀHCl (pH 8.5). Diffraction data of mutant
SZ348ꢀCYO1 complex was collected at the wavelength of 1.5418
Å on an Raxis IV++ imaging plate detector. The SZ529 crystals
were grown in 1.4 M sodium citrate and 0.1 M Hepes (pH 7.0) by
the sittingꢀdrop vapor difusion method at 4 °C. The proteins at 20
mg/ml were mixed with the reservoir solution at a ratio of 1:1 in a
final volume of 4ꢁL. The complex crystals of SZ529ꢀCYO1were
made by the similar method by soaking the SZ529 crystals in the
buffer 1.6 M sodium citrate and 0.1 M Hepes (pH 7.0), together
with 10 mM CYO1 for 5ꢀ10 min. Diffraction data of mutant
SZ529ꢀCYO1 was collected at the wavelength of 0.93911 Å on
SSRF beamline 17B of the National Center for Protein Science
Shanghai (China). All data collection was performed at 100 K. All
data sets were indexed, integrated, and scaled using the HKL2000
package.17 The complex structures of SZ348ꢀCYO1and SZ529ꢀ
CYO1 were solved by molecular replacement method using the
program PHASER18 and the structure of apoꢀSZ348 LEH (PDB
code 5CF2) as a search model. Rounds of automated refinement
were performed with PHENIX19 and the models were extended
and rebuilt manually with COOT.20 The structures of SZ348ꢀ
CYO1, and SZ529ꢀCYO1complex have been refined to 2.50, 2.60
Å, respectively. Considering the low occupancy of CYO1 in the
asymmetric units of SZ338ꢀCYO1, we reꢀrefined this crystal
structure by introducing multiple orientations for CYO1 molecule.
SZ719ꢀCYO3 (150 ± 0.1°). These observations are also in good
1
2
3
4
5
6
7
8
agreement with previous work.11
The differences in Ea originate from the catalytic pathways
with respect to the binding conformation of the substrate (Figure
4 and movies S10-12). Specifically, if oxygen atom O2 attacks
the distant C1 carbon atom, it undertakes a longer pathway with
an unfavorable “angle of attack” (AA). Meanwhile, this oxygen
atom would bypass C2 in vicinity and experience additional steric
hindrance. In contrast, attacking the neighboring C2 directly
results in a better AA with minimal steric effect from C1.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Conclusions and Perspectives. In summary, we have applied a
stereochemical probe for gaining insight into the substrate stereꢀ
oselectivity of limonene epoxide hydrolase (LEH). Based on
crystal structures and mutagenesis, the QM and MD calculation
have shed light on the mechanistic intricacies of this enzyme.
When substrates approach LEH, they can induce large fluctuation
in H4, C loop and loop A, which opens space for the entrance of
the respective compounds. Subsequently, an activated highly
conserved water molecule, hydrogenꢀbonded with Y53, N55 and
D132, can undergo nucleophilic attack in two different ways:
reaction either at C1 or at C2 of the epoxide moiety (Figure 5). If
the distance between water and C1 (d1) is shorter than that of C2
(d2), LEH favors the formation of the (R,R)ꢀproducts. This stereꢀ
oselectivity is supported by the QMꢀcomputed lower activation
energy.
The statistics for data collection and crystallographic refineꢀ
ment are summarized in Table S2.
We also found that different point mutations of LEH could
lead to the rearrangement of the highly conserved water, resulting
in (S,S)ꢀproducts. A relatively compact binding site (e.g. that of
SZ719) can stabilize both the smaller substrate CYO1 and the
larger substrate CYO3. In contrast, larger binding sites (e.g. those
of SZ338, SZ348 and SZ529) destabilize the relatively smaller
substrate CYO1 in the near attack conformation (reactive conforꢀ
mation),16 which leads to LEH inactivity towards the smaller
substrate CYO1.
Protein, ligand structure preparations. All protein and ligand
structures were prepared in Schrodinger Maestro software.21
Protein-ligand docking. Docking a ligand to the receptor was
performed using Glide.22 Cubic boxes centered on the ligand mass
center with a radius 8 Å for all ligands defined the docking bindꢀ
ing regions. Flexible ligand docking was executed for all strucꢀ
tures. Twenty poses per ligand out of 20,000 were included in the
postꢀdocking energy minimization. The best scored pose for the
ligand was chosen as the initial structure for MD simulations.
Our interaction fingerprint (IFP) analysis has depicted resiꢀ
dues frequently involved in proteinꢀligand interactions. These
residues play an essential role in maintaining a specific substrate
conformation which corresponds to a specific product. The Xꢀray
structural results and the insights gained by the theoretical analyꢀ
sis can be expected to be of significant use in future protein engiꢀ
neering of selective LEH mutants.
Molecular dynamics simulations. All unbiased MD simulations
were performed in Gromacs 5.1.4.23 All amino acid residues of
the protein were modelled according to their protonation states at
neutral pH. Residue D101 was found in a protonated state. The
protein was centered in a water box with a distance of 11 Å away
from the protein. The total number of atoms was approximately
40800: 48 Na+ and 39 Cl− ions, and about 12,800 water moleꢀ
cules. The Amber99SB*ꢀILDN force field24 was assigned to the
protein, water and ions, while the ligands were treated by the
Amber GAFF2 force field25 through the ACPYPE26 tool. The
ligands were submitted to the GAUSSIAN 09 program27 for
structure optimization at the HartreeꢀFock 6ꢀ31G* level prior to
the generation of force field parameters. All bond lengths of
hydrogen atoms in the system were constrained using Mꢀ
SHAKE.28 A 10 Å cutoff was used for Van der Waals and shortꢀ
range electrostatic interactions. The whole system was heated
linearly at constant volume (NVT ensemble, 1bar) from 0 to 310K
over 400 ps. Ten nanoseconds equilibration was performed at
constant pressure and temperature (NPT ensemble; 310 K, 1 bar)
using the NoseꢀHoover coupling scheme with two temperature
groups. Longꢀrange electrostatic interactions were computed by
particle mesh Ewald (PME) summation. Finally, a 300 ns MD
simulation with a time step of 2.0 fs were performed for Apo
LEH, SZ338ꢀCYO1, SZ338ꢀCYO3, SZ348ꢀCYO1, SZ348ꢀ
CYO3, SZ529ꢀCYO1, SZ529ꢀCYO3, SZ719ꢀCYO1 and SZ719ꢀ
CYO3. The MD simulations results were analyzed in Gromacs23
and VMD.29 The binding site volume was calculated by POVME
2.0 software.30 The interaction fingerprint (IFP) between protein
Materials and Methods
Materials. KOD Hot Start DNA Polymerase was obtained from
Novagen. The oligonucleotides were synthesized by Life Techꢀ
nologies. Plasmid preparation kit was ordered from Zymo Reꢀ
search and PCR purification kit was bought from QIAGEN. The
mutant was sequenced in GATC Biotech. All commercial chemiꢀ
cals were purchased from SigmaꢀAldrich. Lysozyme and DNase I
were purchased from AppliChem.
Primers design, library construction and screening. Please see
our previous work.3a,7a
X-ray structure determination The SZ348 variant was crystalꢀ
lized in 2.4 M sodium/potassium phosphate, 0.1 M TrisꢀHCl (pH
8.5), using the sittingꢀdrop vapor diffusion method at 18 °C.
Proteins (at 16 mg/mL) were mixed in a 1:1 ratio with the reserꢀ
voir solution in a final volume of 4 ꢁL and equilibrated against the
reservoir solution. The complex crystals of SZ348ꢀCYO1 was
made by soaking the crystals of SZ348 in sodium/potassium
phosphate, 0.1 M TrisꢀHCl (pH 8.5) containing 10 mM of CYO1
and 2.5% (v/v) of acetonitrile as coꢀsolvent for 5ꢀ10 min. All
crystals were flashꢀcooled in liquid nitrogen after being dipped
ACS Paragon Plus Environment

Page 7 of 9
Journal of the American Chemical Society
and ligand was done with PLIP tool.31 More details about IFP can
“Stereochemistry”, in New Comprehensive Biochemistry,
Tamm, C. (ed.), Vol. 3, Elsevier, Amsterdam, 1982.
be found in our previous work.32 Figures were prepared in
PyMOL and Inkscape.33
1
2
3
4
5
6
7
8
2. Recent reviews of directed evolution9c: (a) Alcalde, M. (ed.)
Directed Enzyme Evolution: Advances and Applications,
Springer, Stuttgart, 2017. (b) Reetz, M. T. Directed Evolu-
tion of Selective Enzymes: Catalysts for Organic Chemistry
and Biotechnology, WileyꢀVCH, Weinheim, 2016. (c)
Bommarius, A. S. Annu. Rev. Chem. Biomol. Eng. 2015, 6,
319ꢀ345. (d) Currin, A.; Swainston, N.; Day P. J.; Kell, D. B.
Chem. Soc. Rev. 2015, 44, 1172ꢀ1239. (e) Renata, H.; Wang,
Z. J.; Arnold, F. H. Angew. Chem. Int. Ed. 2015, 54, 3351ꢀ
3367. (f) Denard, C. A.; Ren H.; Zhao, H. Curr. Opin. Chem.
Biol. 2015, 25, 55ꢀ64. (g) Gillam, E. M. J.; Copp, J. N.;
Ackerley, D. F. (eds.), Methods in Molecular Biology, Vol.
1179: Directed Evolution Library Creation: Methods and
Protocols, ed. Humana, Totowa, 2nd edn, 2014. (h) Goldꢀ
smith, M.; Tawfik, D. S. Methods Enzymol. 2013, 523, 257ꢀ
283.
The quantum mechanics surface scan. All calculations were
done in GTKDynamo tool,34 a PyMOL interface plugin, using
ORCA35 program as the QM engine. QM calculations were perꢀ
formed using the hybrid density function B3LYP. Residues within
5 Å around the substrate, including Y53, W130, N55, D132, F134,
I116, R99, F139, D101, L103, V80, L35 and L74 (a total of 280ꢀ
300 atoms), were included in the QM calculation. The backbones
of the residues were restrained with a force constant of 4000
kcal·molꢀ1·Å2. For each step in the reaction catalyzed by LEH, we
performed relaxed potential energy surface (PES) scans, from
which we can reconstruct the reaction mechanisms as reported
previously.36 All fixed points were located and characterized using
an analogous version of the micro/macro iteration algorithm37
with convergence criterion of 0.01 kcal·mol−1·Å−1. The final
Hessian was computed to check the existence of a single imagiꢀ
nary frequency (saddle point) or no imaginary frequency (miniꢀ
mum). Steepest descent path calculations were performed for each
step so as to verify that the transition states connect the related
reactant and product states. Maximum interaction was set to 5,000
steps.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
3. Directed evolution studies of stereoselective enzymes featurꢀ
ing Xꢀray structures of evolved mutants: (a) Sun, Z.; Lonsꢀ
dale, R.; Kong, X.ꢀD.; Xu, J.ꢀH.; Zhou, J.; Reetz, M. T. An-
gew. Chem. Int. Ed. 2015, 54, 12410−12415. (b) Noey, E. L.;
Tibrewal, N.; JiménezꢀOsés, G.; Osuna, S.; Park, J.; Bond, C.
M.; Cascio, D.; Liang, J.; Zhang, X.; Huisman, G. W.; Tang,
Y.; Houk, K. N. Proc. Natl. Acad. Sci. U.S.A. 2015, 112,
E7065ꢀE7072. (c) Brenna, E.; Crotti, M.; Gatti, F. G.; Monti,
D.; Parmeggiani, F.; Powell, R. W.; Santangelo, S.; Stewart,
J. D. Adv. Synth. Catal. 2015, 357, 1849ꢀ1860. (d) Pompeu,
Y. A.; Sullivan, B.; Stewart, J. D. Adv. Synth. Catal. 2013, 3,
2376ꢀ2390. (e) Walton, A. Z.; Conerly, W. C.; Stewart, J. D.
ACS Catal. 2011, 1, 989ꢀ993. (f) Reetz, M. T.; Bocola, M.;
Wang, L.ꢀW.; Sanchis, J.; Cronin, A.; Arand, M.; Zou, J.;
Archelas, A.; Bottalla, A.ꢀL.; Naworyta, A.; Mowbray, S. L.
J. Am. Chem. Soc. 2009, 131, 7334ꢀ7343. (g) Kim, J.; Seok,
S.ꢀH.; Hong, E.; Yoo, T. H.; Seo, M.ꢀD.; Ryu, Y. Appl. Mi-
crobiol. Biotechnol. 2017, 101, 2333ꢀ2342. (h) Pang, P.;
Cao, L.ꢀC.; Liu, Y.ꢀH.; Xie, W.; Wang, Z. J. Struct. Biol.
2017, 198, 154ꢀ162. (i) Wang, Z.; Zhou, S.; Zhang, S.;
Zhang, S.; Zhu, F.; Jin, X.; Chen, Z.; Xu, X. Sci. Rep. 2017,
7 (1), 4007.
ASSOCIATED CONTENT
Supporting Information. Selectivity data for additional mutants;
data collection and refinement statistics for Xꢀray diffraction
experiments; MD simulations and QM calculation data. This
material is available free of charge via the Internet at
http://pubs.acs.org.
AUTHOR INFORMATION
* Corresponding Author
shuguang.yuan@gmail.com;
jiahai@mail.sioc.ac.cn;
reetz@mpiꢀmuelheim.mpg.de;
#Authors contribute equally to this work.
Notes
4. (a) Morisseau, C.; Hammock, B. D. Annu. Rev. Pharmacol.
Toxicol. 2005, 45, 311ꢀ313. (b) ElꢀSherbeni, A. A.; ElꢀKadi,
A. O. S. Arch. Toxicol. 2014, 88, 2013ꢀ2031. (c) Godoy, P.;
Aranda, M. Arch Toxicol. 2015, 89, 2463ꢀ2464. (d) Kodani,
S. D.; Hammock, B. D. Drug Metab. Disp. 2015, 43, 788ꢀ
802. (e) SaenzꢀMendez, P.; Katz, A.; PerezꢀKempner, M. L.;
Ventura, O. N.; Vazquez, M. Proteins 2017, 85, 720ꢀ730.
5. Reviews of EHs in organic synthesis: (a) Archelas, A.;
Furstoss, R. Curr. Opin. Chem. Biol. 2001, 5, 112ꢀ119. (b)
Steinreiber, A.; Faber, K. Curr. Opin. Biotechnol. 2001, 12,
552ꢀ558. (c) van Loo, B.; Kingma, J.; Arand, M.; Wubbolts,
M. G.; Janssen, D. B. Appl. Environ. Microbiol. 2006, 72,
2905ꢀ2917. (d) Widersten, M.; Gurell, A.; Lindberg, D. Bio-
chim. Biophys. Acta. 2010, 1800, 316ꢀ326. (e) Kotik, M.;
Archelas, A.; Wohlgemuth, R. Curr. Org. Chem. 2012, 16,
451ꢀ482. (f) Schober, M.; Faber, K. Trends Biotechnol. 2013,
31, 468ꢀ478. (g) Saini, P.; Sareen, D. Mol. Biotechnol. 2017,
59, 98ꢀ116.
The authors declare no competing financial interests.
ACKNOWLEDGMENT
We thank the staff of beamline BL17B of the National Center for
Protein Science Shanghai (China) for access and help with the Xꢀ
ray data collection. Manfred T. Reetz thanks the MaxꢀPlanckꢀ
Society and the LOEWE Research cluster SynChemBio for genꢀ
erous support. Zhoutong Sun thanks the CAS Pioneer Hundred
Talent Program (Type C) (reference number 2016ꢀ053) for initial
start support. Jiahai Zhou thanks the Strategic Priority Research
Program (B) of the Chinese Academy of Sciences
(XDB20000000) and the Science and Technology Commission of
Shanghai Municipality (15JC1400403) for fund support. The
molecular modelling and molecular dynamics simulation were
performed at the Interdisciplinary Centre for Mathematical and
Computational Modeling in Warsaw (GB71ꢀ3 and GB70ꢀ3). We
also thank Dr. José Fernando Ruggiero Bachega for helpful adꢀ
vices on QM calculations.
6. (a) Arand, M.; Hallberg, B. M.; Zou, J.; Bergfors, T.; Oesch,
F.; van der Werf, M. J.; de Bont, J. A.; Jones, T. A.;
Mowbray, S. L. EMBO 2003, 22, 2583ꢀ2592. (b) van der
Werf, M. J.; Overkamp, K. M.; de Bont, J. A. M. J. Bacteriol.
1989, 180, 5052ꢀ5057. (c) van der Werf, M. J.; de Bont, J. A.
M.; Swarts, H. J. Tetrahedron: Asymmetry 1999, 10, 4225ꢀ
4230.
ABBREVIATIONS
Ea, activation energy; MD, molecular dynamics; PCA, principle
component analysis; QM, quantum mechanics.
REFERENCES
1. (a) Koshland, D. E. Jr. Biol. Rev. 1953, 28, 416ꢀ436. (b)
7. (a) Sun, Z.; Lonsdale, R.; Wu, L.; Li, G.; Li, A.; Wang, J.;
Zhou, J.; Reetz, M. T. ACS Catal. 2016, 6, 1590ꢀ1597. (b)
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 9
Zheng, H.; Reetz, M. T. J. Am. Chem. Soc. 2010, 132,
P. S. J. Med. Chem. 2004, 47, 1739ꢀ1749.
15744ꢀ15751.
23. Pronk, S.; Pall, S.; Schulz, R.; Larsson, P.; Bjelkmar, P.;
Apostolov, R.; Shirts, M. R.; Smith, J. C.; Kasson, P. M.; van
der Spoel, D.; Hess, B.; Lindahl, E. Bioinformatics 2013, 29,
845ꢀ854.
24. LindorffꢀLarsen, K.; Piana, S.; Palmo, K.; Maragakis, P.;
Klepeis, J. L.; Dror, R. O.; Shaw, D. E. Proteins 2010, 78,
1950ꢀ1968.
25. Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.;
Case, D. A. J. Comput. Chem. 2004, 25, 1157ꢀ1174.
26. Sousa da Silva, A. W.; Vranken, W. F. BMC Res. Notes.
2012, 5, 367.
27. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.
M., B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.;
Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.;
Sonnenberg, J. L.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda,
R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.;
Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, Jr., J. A.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothꢀ
ers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.;
Normand, J.; Raghavachari, K.; Rendell, A.; Burant, J. C.;
Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.; Millam, J.
M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo,
C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev,
O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.;
Martin, R. L.; Morokuma, K.; Zakrzewski, V. G.; Voth, G.
A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.; Daniels, A.
D.; Farkas, Ö.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.;
Fox, D. J. Gaussian 09, Revision A.1; Gaussian, Inc.: Wallꢀ
ingford CT, 2009.
28. Barth, E.; Kuczera, K.; Leimkuhler, B.; Skeel, R. D. J. Com-
put. Chem. 1995, 16, 1192ꢀ1209.
29. Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph. Mod-
el. 1996, 14, 33ꢀ38.
30. Durrant, J. D.; Votapka, L.; Sorensen, J.; Amaro, R. E. J.
Chem. Theory Comput. 2014, 10, 5047ꢀ5056.
31. Salentin, S.; Schreiber, S.; Haupt, V. J.; Adasme, M. F.;
Schroeder, M. Nucleic Acids Res. 2015, 43, W443ꢀ447.
32. Chan, H. C.; Filipek, S.; Yuan, S. Sci. Rep. 2016, 6, 34736.
33. (a) Yuan, S.; Chan, H. C.; Filipek, S.; Vogel, H. Structure
2016, 24, 2041ꢀ2042. (b) Yuan, S.; Chan, H. C. S.; Hu, Z.
WIREs. Comput. Mol. Sci. 2017, 7:e1298.
34. Bachega, J. F.; Timmers, L. F.; Assirati, L.; Bachega, L. R.;
Field, M. J.; Wymore, T. J. Comput. Chem. 2013, 34, 2190ꢀ
2196.
35. Neese, F. WIREs. Comput. Mol. Sci. 2012, 2, 73ꢀ78.
36. Kanaan, N.; Crehuet, R.; Imhof, P. J. Phys. Chem. B. 2015,
119, 12365ꢀ12380.
37. Marti, S.; Moliner, V.; Tunon, I. J. Chem. Theory. Comput.
2005, 1, 1008ꢀ1016.
1
2
3
4
5
6
7
8
8. (a) Reymond, J.ꢀL. (ed.) Enzyme Assays ꢀ High throughput
Screening, Genetic Selection andFingerprinting, Wileyꢀ
VCH, Weinheim, 2006. (b) AcevedoꢀRocha, C. G.; Agudo,
R.; Reetz, M. T. J. Biotechnol. 2014, 191, 3ꢀ10.
9. (a) Practical tips and guidelines when applying iterative
saturation mutagenesis (ISM): AcevedoꢀRocha, C. G.; Hoeꢀ
benreich, S.; Reetz, M. T. Methods Mol. Biol. 2014, 1179,
103ꢀ128. (b) Reetz, M. T.; Carballeira, J. D. Nat. Protocols.
2007, 2 (4), 891ꢀ903. (c) Review of directed evolution of steꢀ
reoselective enzymes with emphasis on ISM: Reetz, M. T.
Angew. Chem. Int. Ed. 2011, 50, 138ꢀ174.
10. Wijma, H. J.; Floor, R. J.; Bjelic, S.; Marrink, S. J.; Baker,
D.; Janssen, D. B. Angew. Chem. Int. Ed. 2015, 54, 3726ꢀ
3730.
11. (a) Lind, M. E.; Himo, F. Angew. Chem. Int. Ed. 2013, 52,
4563ꢀ4567. (b) Hopmann, K. H.; Hallberg, B. M.; Himo, F. J.
Am. Chem. Soc. 2005, 127, 14339ꢀ14347. (c) Lind, M. E. S.;
Himo, F. ACS Catal. 2016, 6, 8145ꢀ8155.
12. Bakan, A.; Dutta, A.; Mao, W.; Liu, Y.; Chennubhotla, C.;
Lezon, T. R.; Bahar, I. Bioinformatics 2014, 30, 2681ꢀ2683.
13. Bruice, T. C.; Lightstone, F. C. Acc. Chem. Res. 1999, 32,
127ꢀ136.
14. Lodola, A.; Mor, M.; Zurek, J.; Tarzia, G.; Piomelli, D.;
Harvey, J. N.; Mulholland, A. J. Biophys. J. 2007, 92, L20.
15. Hur, S.; Kahn, K.; Bruice, T. C. Proc. Natl. Acad. Sci. U.S.A.
2003, 100, 2215.
16. Selected recent reviews of QM/MM of enzymes: (a) Senn, H.
M.; Thiel, W. Angew. Chem. Int. Ed. 2009, 48, 1198. (b)
Siegbahn, P. E. M.; Himo, F. WIREs Comput. Mol. Sci.
2011, 1: 323ꢀ336. (c) van der Kamp, Mulholland, A. J.
Biochemistry 2013, 52, 2708ꢀ2728. (d) Warshel, A. Angew.
Chem. Int. Ed. 2014, 53, 10020ꢀ10031. (e) Osuna, S.;
JimenezꢀOses, G.; Noey, E. L.; Houk, K. N. Acc. Chem. Res.
2015, 48, 1080ꢀ1089. (f) Jindal, G.; Warshel, A. J. Phys.
Chem. B. 2016, 120, 9913. (g) Sousa, S. F.; Ribeiro, A. J.
M.; Neves, R. P. P.; Bras, N. F.; Cerqueira, N. M. F. S. A.;
Fernandes, P. A.; Ramos, M. J. WIREs Comput. Mol. Sci.
2017, 7: e1281.
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
17. Otwinowski, Z.; Minor, W. Method. Enzymol. 1997, 276,
307ꢀ326.
18. McCoy, A. J.; GrosseꢀKunstleve, R. W.; Adams, P. D.;
Winn, M. D.; Storoni, L. C.; Read, R. J. J. Appl. Crystallogr.
2007, 40, 658ꢀ674.
19. Adams, P. D.; Afonine, P. V.; Bunkoczi, G.; Chen, V. B.;
Davis, I. W.; Echols, N.; Headd, J. J.; Hung, L. W.; Kapral,
G. J.; GrosseꢀKunstleve, R. W.; McCoy, A. J.; Moriarty, N.
W.; Oeffner, R.; Read, R. J.; Richardson, D. C.; Richardson,
J. S.; Terwilliger, T. C.; Zwart, P. H. Acta Crystallogr. D.
Biol. Crystallogr. 2010, 66, 213ꢀ221.
20. Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Acta
Crystallogr. D. Biol. Crystallogr. 2010, 66, 486ꢀ501.
21. Schrödinger, L.; Schrödinger, LLC: 2015.
22. Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T. A.;
Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E. H.;
Shelley, M.; Perry, J. K.; Shaw, D. E.; Francis, P.; Shenkin,
ACS Paragon Plus Environment

Page 9 of 9
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
For Table of Contents Only
ACS Paragon Plus Environment