Investigating Substrate Scope and Enantioselectivity of a Defluorinase by a Stereochemical Probe

Article abstract of DOI:10.1021/jacs.7b06019

The possibility of a double Walden inversion mechanism of the fluoracetate dehalogenase FAcD (RPA1163) has been studied by subjecting rac-2-fluoro-2-phenyl acetic acid to the defluorination process. This stereochemical probe led to inversion of configuration in a kinetic resolution with an extremely high selectivity factor (E > 500), showing that the classical mechanism involving S_N2 reaction by Asp110 pertains. The high preference for the (S)-substrate is of synthetic value. Wide substrate scope of RPA1163 in such hydrolytic kinetic resolutions can be expected because the reaction of the even more sterically demanding rac-2-fluoro-2-benzyl acetic acid proceeded similarly. Substrate acceptance and stereoselectivity were explained by extensive molecular modeling (MM) and molecular dynamics (MD) computations. These computations were also applied to fluoroacetic acid itself, leading to further insights.

Full text of DOI:10.1021/jacs.7b06019

Subscriber access provided by UNIV OF NEWCASTLE
Article
Investigating Substrate Scope and Enantioselectivity
of a Defluorinase by a Stereochemical Probe
Jian-bo Wang, Adriana Ilie, SHUGUANG YUAN, and Manfred T. Reetz
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 25 Jul 2017
Downloaded from http://pubs.acs.org on July 25, 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 8
Journal of the American Chemical Society
1
2
3
4
5
6
Investigating Substrate Scope and Enantioselectivity of a
Defluorinase by a Stereochemical Probe
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†,Ⱶ
†,Ⱶ
Ш
,†,Ⱶ
Jianꢀbo Wang, Adriana Ilie, Shuguang Yuan*, and Manfred T. Reetz*
†
Department of Chemistry, PhilippsꢀUniversity, HansꢀMeerweinꢀStrasse 4, 35032 Marburg, Germany,
Ⱶ
MaxꢀPlanckꢀInstitut für Kohlenforschung, KaiserꢀWilhelmꢀPlatz 1, 45470 Muelheim, Germany,
Ш
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: The possibility of a double Walden inversion mechanism of the fluoracetate dehalogenase FAcD (RPA1163) has
been studied by subjecting racꢀ2ꢀfluoroꢀ2ꢀphenyl acetic acid to the defluorination process. This stereochemical probe led to inverꢀ
sion of configuration in a kinetic resolution with an extremely high selectivity factor (E >500), showing that the classical mechaꢀ
nism involving S 2 reaction by Asp110 pertains. The high preference for the (S)ꢀsubstrate is of synthetic value. Wide substrate
N
scope of RPA1163 in such hydrolytic kinetic resolutions can be expected because the reaction of the even more sterically demandꢀ
ing racꢀ2ꢀfluoroꢀ2ꢀbenzyl acetic acid proceeded similarly. Substrate acceptance and stereoselectivity were explained by extensive
molecular modelling (MM) and molecular dynamics (MD) computations. These computations were also applied to fluoroacetic
acid itself, leading to further insights.
1
0
INTRODUCTION
acid (Scheme 1a). A recent study provided further illuminatꢀ
ing details, revealing a subtle asymmetry in the crystal strucꢀ
ture of the apoꢀRPA1163 homodimer which plays a role in the
Fluorinated organic compounds are used in a vast number of
industrial applications, ranging from polymers and plantꢀ
protecting agents to therapeutic drugs. Not only the developꢀ
ment of efficient synthetic techniques, but also the invention
of ecologyꢀdriven degradation processes constitute prominent
11
1
observed “halfꢀofꢀthe sites” reactivity. This means that at any
instance only one of the two protomers participates in subꢀ
strate binding.
2
3
tasks. Due to the exceptionally high dissociation energy of CꢀF
bonds, ranging up to 130 kcal/mol, the establishment of mild
degradation methods is challenging. In all of these endeavors,
synthetic reagents and manꢀmade catalysts dominate current
methodology, while biocatalysts are just beginning to play a
4
role. It can be expected that the study of substrate scope couꢀ
pled with knowledge of the enzyme mechanism will help in
paving the way to novel (industrial) applications.
Metabolism of organofluorine compounds such as Fꢀ
containing therapeutic drugs is another important research
5
area. Enzymes called defluorinases are known to cleave CꢀF
6
bonds under ambient conditions, one of the first types being
Scheme 1. a) Widely accepted mechanism of the defluori-
nation of fluoroacetic acid (FAc) catalyzed by RPA1163 as
7
fluoroacetate dehalogenases (FAcD). These enzymes catalyze
8,10
the hydrolysis of the very stable fluoroacetic acid (FAc) with
formation of glycolic acid. FAc is one of the most toxic comꢀ
pounds to mammals, yet it is used as a commercial rodenticide
in a number of countries, including the USA.
the fluoroacetate dehalogenase (FAcD).
zymatic defluorination of FAc by an intramolecular S
reaction followed by an S 2 transformation.
b) Possible en-
N
N
Several early studies showed that the mechanism of a typiꢀ
While the intermediacy of the FAcꢀcomplexes and the reꢀ
cently proposed domainꢀspecific dynamics are indisputaꢀ
cal FAcD involves a classical S 2 reaction in which an asparꢀ
N
10,11
tate (Asp105) functions as the nucleophile with direct expulꢀ
ble,
nism
the possibility of a double Walden inversion mechaꢀ
involving two successive nucleophilic substitution
8
12,13
sion of fluoride, analogous to homologous haloalkane
9
dehalogenases, which however do not accept FAc. Subseꢀ
reactions with the intermediacy of an αꢀlactone playing a
pivotal role cannot be rigorously excluded. Relevant is the
observation that the solvolysis of chiral αꢀbromopropionic acid
under basic conditions with formation of lactic acid occurs
quently, in a study on the other defluorinase RPA1163, the
details of this accepted mechanism were illuminated by mapꢀ
ping the reaction coordinates of defluorination using Xꢀray
crystallography, thereby providing snapshots which picture the
free enzyme, enzymeꢀFAc complex, enzymeꢀglycolyl covalent
complex, and enzymeꢀproduct complex harboring glycolic
14
with retention of configuration. In related chemistry, it has
been demonstrated that the deamination of (R)ꢀalanine by the
action of HNO with formation of (R)ꢀlactic acid also proceeds
2
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 2 of 8
1
8
via an αꢀlactone followed by an S 2 reaction with overall
retention of configuration. Indeed, this is a general synthetic
Sihꢀformula, the selectivity factor, reflecting the relative rate
of reaction of (S)ꢀ versus (R)ꢀ1, proved to be exceptionally
large, E = >500. The stereochemical outcome proves that (S)ꢀ1
reacts with inversion of configuration leading to the sole forꢀ
mation of (R)ꢀ2. These results allow for three conclusions:
N
1
5a
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
method, (S)ꢀamino acids leading to the respective (S)ꢀhydroxy
1
5bꢀc
carboxylic acids.
αꢀLactones are now accepted as reactive
16
intermediates in these and other reactions, and some have
even been isolated.
1
7
Following this line of thought, fluoroacetic acid could be
activated in the enzyme’s binding pocket by deprotonation of
the carboxyl which then induces an intramolecular nucleoꢀ
philic substitution reaction with formation of a reactive αꢀ
lactone (Scheme 1b). The usual fluoride anion stabilization
can be expected in this scenario. In a second step, the intermeꢀ
•
•
The fluoroacetate dehalogenase RPA1163 is capable of
accepting a rather bulky substrate.
The mechanism in which an aspartate expels fluoride directꢀ
ly according to Scheme 1a operates, and not the alternative
overall Walden retention process by way of two inversion
events (Scheme 1b).
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
diate αꢀlactone would undergo an S 2 reaction by Asp110,
N
followed by the usual hydrolysis of the covalent enzymeꢀester
intermediate.
•
Hydrolytic kinetic resolution is possible with an unusually
high degree of enantioselectivity, which is of interest in synꢀ
thetic organic chemistry.
Since FAc is not chiral, such a reaction mode cannot be
proven or disproven on the basis of the presently available
mechanistic data. In order to test this mechanistic alternative,
we designed a stereochemical probe by resorting to the chiral
substrate 2ꢀfluoroꢀ2ꢀphenyl acetic acid (1) (Scheme 2). Since
it is sterically more demanding than FAc, the question of
substrate acceptance by RPA1163 had to be addressed first. In
principle both (R)ꢀ and (S)ꢀenantiomers could be tested in
separate experiments. However, should this substrate be acꢀ
cepted by wildtype RPA1163, then only one of the two enanꢀ
tiomers might react in a stereoselective hydrolytic kinetic
resolution, which would provide the necessary stereochemical
evidence for mechanistic conclusions. High enantioselectivity
at 50% conversion in this particular case and in similar transꢀ
formations of related racemates would also be of preparative
significance in synthetic organic chemistry, enabling asymꢀ
metric biocatalytic access to chiral fluoro compounds. To date
only the sterically small unsubstituted FAc has been used as a
Kinetics. Using racꢀ1 as substrate and RPA1163 as the fluoroꢀ
acetate defluorinase, the following kinetic parameters were measꢀ
ured:
ꢀ
1
ꢀ
k
cat
ꢀ1
= 61.3±0.5 min ; K = 1.7±0.2 mM; k /K = 601 sec
m cat m
1
M
These results can be compared to the kinetic parameters of the
natural substrate FAc. Due to the difficulty in determining the
concentrations of FAc and glycolic acid in the aqueous medium,
we were unable to obtain kinetic data as in the case of substrate 1.
Fortunately, in a previous study steady state kinetics were perꢀ
formed by using a microcalorimetric method, which provided the
10
following data
:
ꢀ1
ꢀ1 ꢀ1
kcat = 6.7±0.6 min ; K = 3.3±0.2 mM; k /K = 33 sec M
m
cat
M
8
,10,11
substrate in RPA1163 catalysis.
While the results of the two determinations may not be preciseꢀ
ly comparable, they indicate that the sterically demanding nonꢀ
natural substrate (S)ꢀ1 is actually more active, a surprising findꢀ
ing. These results suggest that fluoroacetate may not be the natuꢀ
ral substrate, but some as yet unknown compound.
Scheme 2. Hydrolytic kinetic resolution of rac-1 catalyzed
by the fluoroacetate dehalogenase RPA1163.
Another Chiral Substrate. In order to test whether wildtype
RPA1163 also accepts other even more sterically demanding
substrates, racꢀ3 was subjected to hydrolytic kinetic resolution
under comparable reaction conditions (Scheme 3). As before, a
similar reaction occurred, showing the same mode of stereoselecꢀ
tivity likewise with a very high selectivity factor of E > 500. This
finding suggests once more that RPA1163 is a fluroacetate
dehalogenase with potentially broad substrate acceptance.
In the present study, we report that the fluoracetate dehaloꢀ
genase FAcD (RPA1163) is indeed capable of not only acceptꢀ
ing compound 1, but also a sterically more demanding subꢀ
strate. This sets the stage for hydrolytic kinetic resolution. In
these reactions we discovered an unusually high degree of
stereoselectivity favoring the transformation of (S)ꢀ1, which
reveals the synthetic potential of RPA1163. Also serving as a
stereochemical probe, the reaction demonstrates that αꢀ
lactones are not involved as shortꢀlived intermediates. We also
performed molecular dynamics (MD) computations with the
monomer and the dimer, which shed light on the origin of
stereoselectivity.
Scheme 3. Hydrolytic kinetic resolution of rac-3 catalyzed
by the fluoroacetate dehalogenase RPA1163.
Molecular Dynamics Simulations. In order to gain insight inꢀ
to the origin of enantioselectivity, we performed extensive molecꢀ
ular modelling (MM) and molecular dynamics (MD) simulations
starting from the RPA1163 crystal structures. The overall
Results and discussion
10
Substrate acceptance and kinetic resolution of rac-1. Upꢀ
on subjecting racꢀ1 to RPA1163ꢀcatalysis, we discovered that
the reaction occurs rapidly, showing for the first time that
RPA1163 is capable of accepting bulky substrates, not just the
sterically small FAc. The transformation stops completely at
50% conversion, the only new product being (R)ꢀ2, leaving
behind unreacted (R)ꢀ1 in a ratio of 1:1 (Scheme 2). Using the
RPA1163 structure (pdb: 5SWN) comprises eleven αꢀhelices
and eight βꢀstrands, which fold into a compact catalytic pocket
(Figure S1 and S2). It is a dimer. The catalytic machinery
RPA1163 is composed of D110, D134, and H280, all three resiꢀ
dues being located in loops: D110 is situated between β5 and
helix3, D134 is situated between β6 and helix4, and H280 is
situated between β8 and helix11 (Figure S1B). Loop B (residues
ACS Paragon Plus Environment

Page 3 of 8
Journal of the American Chemical Society
G249ꢀT259) is unresolved in most solved crystal structures (Fig-
the substrate interactions (Figure S1C, S2B and S4). These comꢀ
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
ure S3), which indicates high flexibility of this region. Moreover,
we find that loop A (residues M145ꢀP166) is flexible in the monꢀ
omer, although it is less so in the dimer (Figure S1 and S2). We
computationally “repaired” this region and used the repaired
model for all computational tasks.
putational findings are consistent with the stereochemical results.
In order to further confirm the favorable binding modes of both
RPA1163ꢀ(S)ꢀ1 and RPA1163ꢀ(R)ꢀ2 obtained from docking, we
performed a 200 ns molecular dynamics (MD) simulation for each
complex model in both dimer and monomer states. The interacꢀ
tions in dimer and monomer states proved to be identical. These
computations revealed Hꢀbond interactions, salt bridges, πꢀπ
stacking and hydrophobic contacts (Figure S4) which are all in
line with the initial model derived only from docking computaꢀ
tions. The interactions in the dimer systems are stronger than that
in the monomer state, as indicated by interaction fingerprinting
(Figure S4).
To understand how substrate and product interact with the naꢀ
tive RPA1163, we computationally constructed three complex
models in the dimer: RPA1163ꢀ(S)ꢀsubstrate complex (RPA1163ꢀ
(S)ꢀ1), RPA1163ꢀ(R)ꢀsubstrate complex (RPA1163ꢀ(R)ꢀ1) and the
RPA1163ꢀ(R)ꢀproduct complex (RPA1163ꢀ(R)ꢀ2), all of which
were obtained using proteinꢀligand docking. Two dominant bindꢀ
ing modes for both RPA1163ꢀ(S)ꢀ1 and RPA1163ꢀ(R)ꢀ1 occur
during docking procedures (Figure 1). In the first binding mode
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Based on the experimental and computational findings, we proꢀ
pose the catalytic mechanism of RPA1163 as follows (Figure 2):
(A) In the APO RPA1163, two loops (loop A and loop B) next to
the catalytic pocket are very flexible and the dynamics allows the
size of the pocket to vary extensively. H280, which stacks beꢀ
tween D110 and D134, is protonated (calculated pKa = 8.6±0.2).
(B) The motion of the substrate into RPA1163 leads to the shift of
loop A and loop B, thereby producing large space in the binding
pocket. This explains why the sterically demanding substrate 1
can readily enter and bind, necessary for catalysis. (C) After (S)ꢀ1
adopts an energetically favored pose at the catalytic site, D110
acts as the nucleophile, attacking the Fꢀbearing Cꢀatom with
inversion of configuration and expulsion of fluoride which is
stabilized by H155. (D) An intermediate covalent enzyme ester
complex forms, which becomes protonated by H280. (E) Hydroꢀ
lytic ester cleavage takes place by the attack of nucleophilic waꢀ
ter, which is activated by H155, leading to product (R)ꢀ2. In this
process D110 becomes a free protonated residue, while (R)ꢀ2
adopts a new pose on the way to departing from the binding pockꢀ
et. (F) Finally, with the fluctuations of loop A and loop B, the
product departs.
(
Figure 1A), which is more favorable than the other according to
the docking scoring function, the Fꢀatom points to the opposite
position of base D109 and is close to H155. This special position
leads to (R)ꢀ2 after the S 2 reaction. In the other binding mode,
N
the Fꢀatom faces D109 directly (Figure 1B). This second binding
mode is identical to the first binding mode of RPA1163ꢀ(S)ꢀ1
(
Figure 1D). We then applied 4x20 ns MD simulations to each
binding mode for Poisson–Boltzmann and surface area continuum
1
9
solvation (MM/PBSA). The results of relative energy calculaꢀ
tions, widely used for ligand relative binding affinity evaluations,
are shown in Figure 1.
Finally, the question whether fluoroacetic acid (FAc) reacts
similarly was addressed by introducing 200 ns MD simulations
for RPA1163ꢀfluoroacetate (RPA1163ꢀFAc) and its correspondꢀ
ing product complex RPA1163ꢀhydroxyacetate (RPA1163ꢀHAc).
The initial binding poses were first constructed by docking, and
proved to be essentially identical to a recently solved RPA1163
11
mutant crystal structure (pdb: 5SWN, Figure S5). Since FAc is
sterically much smaller than substrate 1, the interactions between
RPA1163 and FAc are lower in number than those computed for
the RPA1163ꢀ(S)ꢀ1 complex (Figure S6). Only three interactions
were found: A salt bridge with R111 as well as halogen interacꢀ
tions with H155 and W156. With the introducing of an additional
hydroxyl group in the product, a few more interactions occur in
the RPA1163ꢀHAc complex, including less frequent Hꢀbond
interactions with K152, W185 and Y219. The frequency of the
salt bridge interaction with R111 is lower than that in the
RPA1163ꢀFAc complex.
Figure 1. Two dominant binding modes of RPA1163 substrates
and their relative ligand binding energies. The substrates are
shown in green ballꢀandꢀsticks. The Fꢀatoms are colored in yelꢀ
S1
low. (A) The binding mode 1 of (S)ꢀ1, with ꢁG =ꢀ60.1±1.1 (ꢀ
3
5.4±0.7, monomer) kcal/mol. (B) The binding mode 2 of (S)ꢀ1,
S2
with ꢁG = ꢀ28.1±0.4 (ꢀ14.4±0.9, monomer) kcal/mol. (C) The
binding mode 1 of (R)ꢀ1, with ꢁG =ꢀ35.7±0.7 (ꢀ16.4±0.8, monꢀ
omer) kcal/mol. (D) The binding mode 2 of (R)ꢀ1, with ꢁG = ꢀ
To illustrate the intrinsic property of RPA1163 and the catalytic
mechanism, we performed an additional MD simulation for the
native APO RPA1163. The rootꢀmeanꢀsquare deviation (RMSD)
in the case of APO RPA1163 is notably higher than those of both
RPA1163ꢀ(S)ꢀ1 and RPA1163ꢀ(R)ꢀ2 (Figure S7 & S8), indicating
its pronounced flexibility. We then performed the principle comꢀ
ponent analysis (PCA), an efficient and reliable method to reveal
the most important motions in proteins, specifically for revealing
R1
S2
1
7.5±1.1 (ꢀ6.4±0.9, monomer) kcal/mol.
Accordingly, the binding energy of the (S)ꢀ1 binding mode 1
(
(
Figure 1A), in a pose leading to (R)ꢀ2, is clearly more favorable
ꢁG =ꢀ60.1±1.1 kcal/mol) than that of either (S)ꢀ1 in binding
20
S1
the RPA1163 MD simulation trajectory. The PCA analysis
mode 2 (ꢁG =ꢀ28.1±0.4 kcal/mol) or bound (R)ꢀ1 (ꢁG =ꢀ
S2
R1
indicates that loop A and loop B are capable of undergoing draꢀ
matic shifting, altering the size of catalytic site to a notable extent
3
5.7±0.7 kcal/mol, ꢁG =ꢀꢀ17.5±1.1 kcal/mol). We also calculatꢀ
R2
ed the MM/PBSA for RPA1163 monomer which is in line with
that of the dimer system. Interestingly, the calculated ꢁG is higher
than its corresponding dimer value. This is probably because loop
A is much less flexible in dimer than that in monomer, stabilizing
(
Figure S1C, S2B and movie 1,2). The rootꢀmeanꢀsquare fluctuaꢀ
tion (RMSF) and the Bꢀfactor values of each residue, calculated
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 8
from the three simulated systems (APO RPA1163, RPA1163ꢀ(S)ꢀ
1 and RPA1163ꢀ(R)ꢀ2), lend further support to this conclusion
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. The catalytic mechanism of RPA1163. (A) In the APO RPA1163, two loops (loop A and loop B) next to the catalytic pocket are
very flexible and the size of the pocket can be fluctuated to a large extent. (B) The entry of substrate into RPA1163 leads to the inward
shift of loop A and outward shift of loop B, resulting in a large space in the binding pocket. This facilitates the substrate molecule entering
the catalytic region. (C) After substrate situating at the catalytic site, D110 acts as a nucleophile, attacking the methenyl group connected
with F atom. This leads to the leaving of F atom from the substrate. (D) A covalent intermediate complex forms (E) The covalent intermeꢀ
diate is hydrolyzed by a water molecule. D110 becomes a free deprotonated residue. The hydroxyl group from water is transferred to the
product. (F) Finally, with the fluctuations of loop A and loop B, the product leaves catalytic pocket.
(
Figure S7B & S8B). While loop C fluctuates noticeably more in
as a catalyst in stereoselective transformations which allow access
to enantiomerically pure αꢀfluoro carboxylic acids from racemic
precursors.
APO RPA1163, loop A undergoes only mild shifting for all simuꢀ
lated systems.
Conclusion and Perspectives. In conclusion, we have deꢀ
signed and applied a stereochemical probe for gaining insight into
the substrate scope and stereoselectivity of the fluoroacetate
dehalogenase RPA1163. Using racꢀ2ꢀfluoroꢀ2ꢀphenyl acetic acid
in a kinetic resolution, inversion of configuration of the favored
MATERIALS AND METHODS
Bacterial strains, enzymes, DNA and chemicals. The expresꢀ
sion strain was E. coli BL21 (DE3). All chemicals were purchased
from TCI EUROPE N.V. Endonuclearase NdeI and SalI were
obtained from New England Biolabs. The gene of RPA1163 was
synthesized by the company Lifetechnology and the DNA code
was optimized for E. coli expression. It was used as template to
amplify the target gene. The PCR products were digested by
endonuclease (NdeI and SalI) then were inserted into the
NdeI/SalI sites of pET28a to get the expressed plasmid (named
DF/pET28a).
(
S)ꢀsubstrate was observed in a rapid reaction. This demonstrates
for the first time that RPA1163 is capable of accepting bulky
substrates other than the sterically small fluoroacetic acid, and
that a synthetically useful level of stereoselectivity is possible (E
>
500 in favor of the transformation of the (S)ꢀenantiomer). Thus,
this process constitutes a simple way to access enantiomerically
pure (R)ꢀ2ꢀfluoroꢀ2ꢀphenyl acetic acid. Surprisingly, the kinetics
show that this nonꢀnatural compound is more reactive than the
natural sterically small substrate FAc. A second even more bulky
substrate, racꢀ2ꢀfluoroꢀ2ꢀbenzyl acetic acid, behaves similarly.
The stereochemical results lend support to the previously postuꢀ
Protein expression, purification and reaction conditions.
The expressed plasmid DF/pET28a was transformed into BL21
(DE3), single colonies were picked into the Lumbria Broth which
contains 50 ꢂg*mLꢀ1 Kanamycin, then were incubated in 37℃
overnight. About 1% seeding cultures were transferred into the
TB culture containing 50 ꢂg*mLꢀ1 Kanamycin, the expressed
system was incubated in 37℃ for about 2ꢀ3 hours. After the
OD600 nm had reached 0.6ꢀ0.8, 1 mM IPTG was added to induce
10,11
lated mechanism,
thereby excluding the possibility of an αꢀ
lactone as a reactive intermediate. Extensive MM and MD comꢀ
putations provide insight into the origin of stereoselectivity. Finalꢀ
ly, our study also points to the potential application of RPA1163
ACS Paragon Plus Environment

Page 5 of 8
Journal of the American Chemical Society
the protein expression, and the temperature was lowered to 25℃.
lowest DOPE (Discrete Optimized Protein Energy) score was
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
After about 20 hours, the cells were gathered by centrifuging, and
were then washed two times with TrisꢀH2SO4 buffer (pH=7.5).
To purify the protein, the cells were lyzed by sonicating, then
were purified by NiꢀNTA. To use the whole cells to catalyze the
reaction, the cells were dispersed in TrisꢀH SO buffer, the dry
weight of cells amounted to10ꢀ12 g*Lꢀ1, which were then tested
with 5 mM, 10 mM, 15 mM or 20 mM substrate 1 and 10 mM, 15
mM or 20 mM substrate 3. The total volume of the reaction sysꢀ
tem was 500ꢂL, which was incubated at 30℃,1000 rpm for at
least 2 hours.
chosen for protein construction.
Protein structure preparations. All protein models were preꢀ
pared in Schrödinger suite software under OPLS_2005 force
2
4
25
field . Hydrogen atoms were added to repaired crystal structures
2
6
according to the physiological pH (7.5) with the PROPKA tool
2
7
in Protein Preparation tool in Maestro to optimize the hydrogen
bond network. Constrained energy minimizations were conducted
on the fullꢀatomic models, with heavy atom coverage to 0.4 Å.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Reaction protocol and kinetics measurement. Concentrated
HCl was used to quench the reaction (1/10), then 700 ꢂL ethyl
acetate was added to extract the substrates and product. 500 ꢂL
organic layer was obtained after centrifuging. Then 300 ꢂL methꢀ
anol were added to the system. To derivate the substrates and the
products, 30ꢂL of 2M TMSCHN2 (trimethylsilyl diazomethane in
diethyl ether) was added to the above mixture; 30℃, 1000 rpm for
Ligand structure preparations. All ligand structures were
2
7
produced in Schrödinger Maestro software . The LigPrep module
in Schrodinger software was introduced for geometric optimizaꢀ
tion by using OPLS_2005 force field. The ionization states of
28
ligands were calculated with Epik tool employing Hammett and
Taft methods in conjunction with ionization and tautomerization
2
8
tools .
3
0 min, then GC (Gas Chromatography) was used directly to
analyze the samples. GC conditions: 50℃ꢀ220℃, 5℃/min, 220℃
for 5 minutes. Column: Hydrodex βꢀTBDAC, 25m*0.25mm ID.
When the concentration of the substrates 1 was 5 mM, 10 mM, 15
mM and 20 mM, the conversion ratios were about 50%, 49%,
Protein-ligand docking. The docking of a ligand to the recepꢀ
2
9
tor was performed using Glide . Cubic boxes centered on the
ligand mass center with a radius 8 Å for all ligands defined the
docking binding regions. Flexible ligand docking was executed
for all structures. 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.
4
9
5
8%, 45%, the eeꢀvalues for unreacted substrates were >99%,
8%, 93%, 91%; for substrate 2, the conversion ratios were about
0% when the concentration were 10 mM and 15 mM. However
when the concentration of substrate 2 was 20 mM, the conversion
ratio decreased to 35%. In all cases, only one product was proꢀ
duced. To measure the kinetics of substrate 1, 1 ꢂM purified
RPA1163 was used, different concentrations of substrate were
tested in 10 minutes, then concentrations of the products were
calculated by the peaks’ areas in the GC.
Molecular dynamics simulations. All unbiased MD simulaꢀ
3
0
tions were performed in Gromacs 5.1.4 . All amino acid residues
of the protein were modelled according to their protonation state
at neutral pH. The protein was centered in a water box with a
distance of 12 Å away from the protein. The total number of
atoms was approximately 43, 000: 47 Na and 41 Cl ions, and
about 13,000 water molecules. Amber99SB*ꢀILDNP force field
Confirming the enantioselectivity of the reaction. The reacꢀ
tion system was amplified to 1 mL, and 50 reactions were perꢀ
formed in parallel. The system included 15 mM of substrate 1, 10
g*Lꢀ1 cells, incubated at 30℃, 1000 rpm for 2 h. Then 10%
concentrated HCl was added for quenching. The reacted reaction
mixtures were combined, and the total mixture extracted by addꢀ
ing 50 mL of ethyl acetate. After esterifying derivation by
+
−
31
was assigned to the protein, water and ions, while the ligands
were treated by Amber GAFF2 force field through ACPYPE
3
2
33
3
4
tool. The ligands were submitted to GAUSSIAN 09 program for
structure optimization at HartreeꢀFock 6ꢀ31G* level prior to the
generation of force field parameters. All bond lengths of hydrogen
atoms in the system were constrained using MꢀSHAKE Van der
Waals and shortꢀrange electrostatic interactions were cut off at 10
Å. The whole system was heated linearly at constant volume
TMSCHN leading to the respective methylesters, the sample was
2
3
5
subjected to column chromatography (Et O: nꢀhexane: formic
2
acid = 1 : 0.1: 0.01 v/v), leading to the isolation of (R)ꢀmandelic
acid, as shown by optical rotation measurement. The eeꢀvalue for
(NVT ensemble, 1bar) from 0 to 310K over 400 ps. Ten nanosecꢀ
the unreacted substrates is under these conditions 33%, [α]D22 = ꢀ
20
onds equilibration was performed at constant pressure and temꢀ
perature (NPT ensemble; 310 K, 1 bar) using the NoseꢀHoover
coupling scheme with two temperature groups. Longꢀrange elecꢀ
trostatic interactions were computed by particle mesh Ewald
3
8.8 . Both substrate and product were analyzed by NMR specꢀ
troscopy to confirm their structures (Figure S6 and S7). For subꢀ
1
1
strate 1: H NMR (300 MHz, CDCl3) δ 10.90 (s, H), 7.55 – 7.37
1
3
(m, 5H), 5.83 (d, J = 47.4 Hz, 1H). C NMR (75 MHz, CDCl3) δ
174.54 (d, J = 27.9 Hz), 133.57 (d, J = 20.5 Hz), 130.07 (d, J = 2.1
Hz), 129.04 (s), 126.81 (d, J = 6.1 Hz), 88.92 (d, J = 186.9 Hz).
F NMR (282 MHz, CDCl3) δ ꢀ181.50 (d, J = 15.0 Hz). NMR
data is consistent with reported data. NMR data of product (R)ꢀ2
H NMR (300 MHz, acetoneꢀd6) δ 7.72ꢀ7.48 (m, 2H), 7.42 –
.26 (m, 3H), 5.21 (s, 1H). C NMR (75 MHz, acetoneꢀd6) δ
74.44 (s), 140.79 (s), 129.11 (s), 128.76 (s), 127.57 (s), 73.50 (s).
NMR data is consistent with reported data. [α]D22 = ꢀ123.24
(PME) summation. Finally, a 200 ns MD simulations with a time
step of 2.0 fs were performed for APO RPA1163, RPA1163ꢀSUB,
RPA1163ꢀPRO, RPA1163ꢀFAc and RPA1163ꢀHAc. The MD
1
9
30
36
2
1
simulations results were analyzed in Gromacs and VMD . The
PCA analysis have been done in VMD. Figures were prepared in
1
:
37,38
13
PyMOL and Inkscape
.
7
1
2
2
MM/PBSA relative binding energy calculation. 4x20 ns reꢀ
3
0
23
strained MD simulations were performed in Gromacs 5.1.4 for
each specific binding mode. Small force constant (1 kcal/mol/Å2)
was added to each ligand to keep a specific binding pose. All
generated trajectories (a total of 400 frames for each case) were
used for the relative binding energy evaluation. The MM/PBSA
calculations were finished using MMPBSA.py script in Amberꢀ
(
10 mg/mL in acetone), ee=99%.
Loop filling and refinements. Since the loop B was missing in
24
the complex structure, loop refinement protocol in Modeller
V9.10 was used to fulfill and refine this area. A total of 20,000
loops were generated for RPA1163 and a conformation with the
39
Tools 16 .
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 8
Interaction fingerprint calculations. The interaction fingerꢀ
Badet, B.; Copin, E. J. Fluorine Chem. 2005, 126, 424. (d). Murꢀ
phy, C. D. Biotechnol. Lett. 2010, 32, 351.
7) Goldman, P. Science 1969, 164, 1123.
8) (a) Liu, J.ꢀQ.; Kurihara, T.; Ichiyama, S.; Miyagi, M.; Tsunasawa,
S.; Kawasaki, H.; Soda, K.; Esaki, N. J. Biol. Chem. 1998, 273,
30897. (b) Kurihara, T.; Esaki, N. The Chemical Record 2008, 8,
40
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
print between protein and ligand was done with PLIP tool . We
first extracted snapshots of the final 20ns MD simulation. We then
used PLIP to convert protein−ligand coordinates into a bitꢀstring
fingerprint (TIFP) registering the corresponding molecular interꢀ
action pattern. TIFP fingerprints have been calculated for 500
protein−ligand complexes for each case, enabling a broad comparꢀ
ison of relationships between interaction pattern similarities and
(
(
6
7.
(9) Reviews of dehalogenases: (a) van Hylckama, J. E. T.; Tang, L.;
Spelberg, J. H. L.; Smilda, T.; Poelarends, G. J.; Bosma, T.; van
Merode, A. E. J.; Fraaije, M. W.; Janssen, D. B. J. Bacteriol. 2001,
183, 5058. (b) de Jong, R. M.; Dijkstra, B. W. Curr. Opin. Struct.
Biol. 2003, 13, 722. (c) Koudelakova, T.; Bidmanova, S.; Dvorak,
P.; Pavelka, A.; Chaloupkova, R.; Prokop, Z.; Damborsky, J. Bio-
tech. J. 2013, 8, 32. (d) Nagata, Y.; Ohtsubo, Y.; Tsuda, M. Appl.
Microbiol. Biotechnol. 2015, 99, 9865.
(10) Chan, P. W.; Yakunin, A. F.; Edwards, E. A.; Pai, E. F. J. Am.
Chem. Soc. 2011, 133, 7461.
(11) Kim, T. H.; Mehrabi, P.; Ren, Z.; Sljoka, A.; Ing, C.; Bezginov, A.;
Ye, L.; Pomès, R.; Prosser, R. S.; Pai, E. F. Science 2017, 355,
eaag2355.
41,19b
ligand or binding site pairwise similarities
. In this work, we
kept the default parameters of PLIP and focused on three types of
interactions: Hꢀbond interaction, πꢀstacking and hydrophobic
contact.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ASSOCIATED CONTENT
The Supporting Information is available free of charge on the
ACS
Publications
website.
(
12) (a) Walden, P. Ber. Dtsch. Chem. Ges. 1893, 26, 210. (b) Walden,
P. Ber. Dtsch. Chem. Ges. 1886, 29,133. (c) Walden, P. Ber. Dtsch.
Chem. Ges. 1899, 32, 1855.
Calculated models, movies and NMR data of related compounds.
(
13) Hughes, E. Trans. Faraday Soc. 1938, 34, 202.
(14) (a) Rørdam, H. N. K. J. Chem. Soc. (Resumed) 1928, 2447. (b)
Winstein, S.; Grunwald, E.; Ingraham, L. J. Am. Chem. Soc. 1948,
70, 821.
AUTHOR INFORMATION
Corresponding Author
(
15) (a) Aitken, R. A.; Meehan, A.; Power, L. A. Synthesis 2015, 47,
557. (b) Moumne, R.; Lavielle, S.; Karoyan, P. J. Org. Chem.
reetz@mpiꢀmuelheim.mpg.de; shuguang.yuan@gmail.com
1
2006, 71, 3332. (c) Pitt, N.; Gani, D. Tetrahedron Lett. 1999, 40,
3811. (d) Jia, C.; Yang, K.ꢀW.; Liu, C.ꢀC.; Feng, L.; Xiao, J.ꢀM.;
Zhou, L.ꢀS.; Zhang, Y.ꢀL. Bioorg. Med. Chem. Lett. 2012, 22, 482.
16) (a) Rodriquez, C.; Williams, I. J. Chem. Soc., Perkin Trans. 2 1997,
953. (b) Buchanan, J. G.; Charlton, M. H.; Mahon, M. F.; Robinson,
J. J.; Ruggiero, G. D.; Williams, I. H. J. Phys. Org. Chem. 2002, 15,
Notes
The authors declare no competing financial interests.
(
ACKNOWLEDGMENT
Support from the MaxꢀPlanckꢀSociety is gratefully acknowledged.
6
42. (c) Buchanan, J. G.; Diggle, R. A.; Ruggiero, G. D.; Williams,
I. H. Chem. Commun. 2006, 1106. (d) Buchanan, J. G.; Ruggiero,
G. D.; Williams, I. H. Org. Biomol. Chem. 2008, 6, 66.
REFERENCES
(17) (a) Wheland, R.; Bartlett, P. D. J. Am. Chem. Soc. 1970, 92, 6057.
b) Adam, W.; Liu, J.ꢀC.; Rodriguez, O. J. Org. Chem. 1973, 38,
269. (c) Coe, P. L.; Sellars, A.; Tatlow, J. C.; Whittaker, G.; Fieldꢀ
(
2
(
1) (a) Banks, R. E.; Smart, B. E.; Tatlow, J. Organofluorine Chemis-
try: Principles and Commercial Applications, Springer Science &
Business Media, 2013. (b) Ojima, I. Fluorine in Medicinal Chemis-
try and Chemical Biology, John Wiley & Sons, 2009. (c) Reddy, V.
P. Organofluorine Compounds in Biology and Medicine, Elsevier,
Amsterdam, 2015. (d) Zhou, Y.; Wang, J.; Gu, Z.; Wang, S.; Zhu,
W.; Aceña, J. L.; Soloshonok, V. A.; Izawa, K.; Liu, H. Chem. Rev.
ing, H. C. J. Chem. Soc., Chem. Commun. 1982, 362.
18) Chen, C. S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am. Chem.
Soc. 1982, 104, 7294.
(
(
19) (a) Genheden, S.; Ryde, U. Expert Opin. Drug Discov. 2015, 10,
4
49. (b) Chan, H. S.; Filipek, S.; Yuan, S. Sci. Rep. 2016, 6, 34736.
(
20) Effenberger, F.; Oßwald, S. Tetrahedron: Asymmetry 2001, 12,
2
016, 116, 422. (e) Ni, C.; Hu, J. Chem. Soc. Rev. 2016, 45, 5441.
(f) Fujiwara, T.; O’Hagan, D. J. Fluorine Chem. 2014, 167, 16. (g)
279.
Gouverneur. V. Fluorine in Pharmaceutical and Medicinal Chemis-
try, Imperial College Press, Singapore, 2012.
(21) Mizuta, S.; Stenhagen, I. S.; O’Duill, M.; Wolstenhulme, J.;
Kirjavainen, A. K.; Forsback, S. J.; Tredwell, M.; Sandford, G.;
Moore, P. R.; Huiban, M. Org. lett. 2013, 15, 2648.
(2) Reviews of synthetic methods for the production of organofluorine
(
(
(
22) Zhu, S. F.; Chen, C.; Cai, Y.; Zhou, Q. L. Angew. Chem., Inter. Ed.
compounds: (a) O'Hagan, D. Chem. Soc. Rev. 2008, 37, 308 (b)
.
2
008, 47, 932.
Furuya, T.; Kamlet, A. S.; Ritter, T. Nature 2011, 473, 470. (c)
Preshlock, S.; Tredwell, M.; Gouverneur, V. Chem. Rev. 2016, 116,
23) Szeleczky, Z.; Bagi, P.; Pálovics, E.; Fogassy, E. Tetrahedron:
Asymmetry 2014, 25, 1095.
24) Eswar, N.; Webb, B.; MartiꢀRenom, M. A.; Madhusudhan, M. S.;
Eramian, D.; Shen, M. Y.; Pieper, U.; Sali, A. Current protocols in
protein science 2007, Chapter 2, Unit 29.
25) Shivakumar, D.; Williams, J.; Wu, Y. J.; Damm, W.; Shelley, J.;
Sherman, W. J. Chem. Theory Comput. 2010, 6, 1509.
26) Sondergaard, C. R.; Olsson, M. H.; Rostkowski, M.; Jensen, J. H. J.
Chem. Theory Comput. 2011, 7, 2284.
27) Schrödinger, L.; Schrödinger, LLC: 2015.
(28) Greenwood, J. R.; Calkins, D.; Sullivan, A. P.; Shelley, J. C. J.
Comput. Aided Mol. Des. 2010, 24, 591.
29) 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, P. S. J. Med. Chem.
2004, 47, 1739.
30) 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, 845.
7
19. (d) Ma, J.ꢀA.; Li, S. Org. Chem. Frontiers 2014, 1, 712. (e)
Wu, J. Tetrahedron Lett. 2014, 55, 4289. (f) Lin, X.; Weng, Z.
Dalton Trans. 2015, 44, 2021. (g) Shibata, N. Bull. Chem. Soc. Jp.
2
016, 89, 1307. (h) Zard, S. Z. Org. Biomol. Chem. 2016, 14, 6891.
(
(
(
(i) Schimler, S. D.; Cismesia, M. A.; Hanley, P. S.; Froese, R. D.;
Jansma, M. J.; Bland, D. C.; Sanford, M. S. J. Am. Chem. Soc.
2017, 139, 1452.
(
3) (a) Murphy, C. D. Appl. Microbiol. Biotechnol. 2016, 100, 2617.
(b) Rotander, A.; Kärrman, A.; van Bavel, B.; Polder, A.; Rigét, F.;
Auðunsson, G. A.; Víkingsson, G.; Gabrielsen, G. W.; Bloch, D.;
Dam, M. Chemosphere 2012, 86, 278.
(
(4) Reviews of fluorinases: (a) O’Hagan, D.; Deng, H. Chem. Rev.
2
2
014, 115, 634. (b) Thuronyi, B. W.; Chang, M. C. Acc. Chem. Res.
015, 48, 584. (c) Senn, H. M. Front. Chem. 2014, 2. (d) Eustáquio,
A. S.; O’Hagan, D.; Moore, B. S. J. Nat. Prod. 2010, 73, 378.
(5) Murphy, C. D.; Sandford, G. Expert Opin. Drug Metab. Toxicol.
2015, 11, 589.
6) (a) Au, K. G.; Walsh, C. T. Bioorg. Chem. 1984, 12, 197. (b)
Kurihara, T.; Yamauchi, T.; Ichiyama, S.; Takahata, H.; Esaki, N. J.
Mol. Catal. B: Enzym. 2003, 23, 347. (c) Natarajan, R.; Azerad, R.;
(
(
(
31) Aliev, A. E.; Kulke, M.; Khaneja, H. S.; Chudasama, V.; Sheppard,
T. D.; Lanigan, R. M. Proteins 2014, 82, 195.
ACS Paragon Plus Environment

Page 7 of 8
Journal of the American Chemical Society
(
(
(
32) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D.
A. J. Comput. Chem. 2004, 25, 1157.
33) Sousa da Silva, A. W.; Vranken, W. F. BMC Res. Notes 2012, 5,
67.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
3
34) 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.; Brothers, 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.: Wallingford CT, 2009.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(
(
(
(
35) Barth, E.; Kuczera, K.; Leimkuhler, B.; Skeel, R. D. J. Comput.
Chem. 1995, 16, 1192.
36) Humphrey, W.; Dalke, A.; Schulten, K. J. Mol. Graph Model 1996,
14, 33.
37) Yuan, S.; Chan, H. C.; Filipek, S.; Vogel, H. Structure 2016, 24,
2
041.
38) Yuan, S.; Chan, H. C. S.; Hu, Z. Wiley Interdisciplinary Reviews:
Computational Molecular Science 2017, e1298.
(39) D.A. Case, R. M. B., D.S. Cerutti, T.E. Cheatham, III, T.A. Darden,
R.E. Duke, T.J. Giese, H. Gohlke,; A.W. Goetz, N. H., S. Izadi, P.
Janowski, J. Kaus, A. Kovalenko, T.S. Lee, S. LeGrand, P. Li, C.;
Lin, T. L., R. Luo, B. Madej, D. Mermelstein, K.M. Merz, G.
Monard, H. Nguyen, H.T. Nguyen, I.; Omelyan, A. O., D.R. Roe,
A. Roitberg, C. Sagui, C.L. Simmerling, W.M. BotelloꢀSmith, J.
Swails,; R.C. Walker, J. W., R.M. Wolf, X. Wu, L. Xiao, P.A.
Kollman; University of California: San Francisco, 2016.
(40) Salentin, S.; Schreiber, S.; Haupt, V. J.; Adasme, M. F.; Schroeder,
M. Nucleic Acids Res. 2015, 43, W443.
(41) Marcou, G.; Rognan, D. J. Chem. Inf. Model. 2007, 47, 195.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 8
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
For Table of Contents Only
ACS Paragon Plus Environment