Enhancing the enantioselectivity of CALB by substrate imprinting: A combined experimental and molecular dynamics simulation model study

Article abstract of DOI:10.1016/j.molcatb.2012.04.017

Kinetic resolution of pentan-2-ol by CALB catalyzed enantioselective transesterification, with various alkylpropanoate acyl donors, was studied in a solid-gas reactor. Results show that the leaving alkoxy group influences the enantiomeric ratio of the reaction. Resolution of pentan-2-ol with methyl propanoate gives an enantiomeric ratio of 62. Esters with longer linear alkyl chains, from ethyl to pentyl propanoate give higher enantiomeric ratios, comprised between 103 and 117. Enantiopure ester (R)-1-methylpentyl propanoate increases the enantiomeric ratio to 140 compared with E = 120 for the racemic mixture. In contrast, enantiopure (S)-1-methylpentyl propanoate decreases the enantiomeric ratio to 72. Our data support the notion of an imprinting effect or ligand-induced enzyme memory caused by the shape of the leaving alcohol. To simulate the imprinting effect caused by the alkoxy part of the acyl donor, molecular modeling studies were performed with both (R)- and (S)-enantiopure 1-methylpentyl propanoate. To investigate how the first step of the reaction, through the first tetrahedral intermediate, affects the enzyme conformation depending on the enantiopure ester substrate used, 20 ns molecular dynamics simulations were carried out. Clustering analysis was done to study relevant conformations of the systems. Differences in the global conformation of the enzyme between systems with R or S enantiomers were not observed. Interestingly however, orientation of the partially buried side chain for Ile285 was affected. This could explain the increased enantiomeric ratio observed with the substrate ester (R)-1-methylpentyl propanoate due to an improved (R)-pentan-2-ol/enzyme interaction.

Full text of DOI:10.1016/j.molcatb.2012.04.017

Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
Contents lists available at SciVerse ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Enhancing the enantioselectivity of CALB by substrate imprinting: A combined
experimental and molecular dynamics simulation model study
L. Chaputa,b, Z. Marton , P. Pineau , L. Domon , V. Tran , M. Graber
a
a
a
b
a,∗
a
Université La Rochelle, UMR 6250 LIENSs CNRS – ULR, Pôle Sciences et Technologies, Bâtiment Marie Curie, Avenue Michel Crépeau, 17042 La Rochelle, France
Unité Biotechnologie, Biocatalyse et Biorégulation (U3B), UMR CNRS 6204 – Faculté des Sciences et des Techniques de Nantes, 2, rue de la Houssinière, BP 92208 44322 Nantes Cedex
, France
b
3
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 5 May 2012
Kinetic resolution of pentan-2-ol by CALB catalyzed enantioselective transesterification, with vari-
ous alkylpropanoate acyl donors, was studied in a solid-gas reactor. Results show that the leaving
alkoxy group influences the enantiomeric ratio of the reaction. Resolution of pentan-2-ol with methyl
propanoate gives an enantiomeric ratio of 62. Esters with longer linear alkyl chains, from ethyl to pentyl
propanoate give higher enantiomeric ratios, comprised between 103 and 117. Enantiopure ester (R)-1-
methylpentyl propanoate increases the enantiomeric ratio to 140 compared with E = 120 for the racemic
mixture. In contrast, enantiopure (S)-1-methylpentyl propanoate decreases the enantiomeric ratio to 72.
Our data support the notion of an imprinting effect or “ligand-induced enzyme memory” caused by the
shape of the leaving alcohol.
Keywords:
Enantioselectivity
Imprinting effect
Kinetics
Candida antarctica lipase B
Molecular modeling
To simulate the imprinting effect caused by the alkoxy part of the acyl donor, molecular modeling stud-
ies were performed with both (R)- and (S)-enantiopure 1-methylpentyl propanoate. To investigate how
the first step of the reaction, through the first tetrahedral intermediate, affects the enzyme conformation
depending on the enantiopure ester substrate used, 20 ns molecular dynamics simulations were carried
out. Clustering analysis was done to study relevant conformations of the systems. Differences in the global
conformation of the enzyme between systems with R or S enantiomers were not observed. Interestingly
however, orientation of the partially buried side chain for Ile285 was affected. This could explain the
increased enantiomeric ratio observed with the substrate ester (R)-1-methylpentyl propanoate due to an
improved (R)-pentan-2-ol/enzyme interaction.
©
2012 Elsevier B.V. All rights reserved.
1
. Introduction
intermediate and the release of the first product, an alcohol formed
with the alkoxy group. In the second step, the chiral alcohol inter-
acts with the acyl–enzyme to form a new binary complex and the
second product ester.
In the active site of CALB, the acyl and alcohol substrate moi-
eties bind in a hairpin orientation. It is thus not surprising that the
acyl chain of the ester interacts with the alcohol and consequently
influences the chiral discrimination of alcohols. However it is unex-
pected that the leaving group, which is the alkoxy part of the ester,
also influences the enantiomeric ratio.
The hypothesis of “molecular imprinting effect” has been pro-
posed by several authors to explain such modification of enzyme
selectivity or activity [2–4]. In these studies the enzyme is brought
together with an “imprinting” molecule (substrate mimics for
example) and then co-immobilized on solid support or by sol–gel
encapsulation. The enzyme molds its active site structure around
the imprint molecule and remains “trapped” in this conformation
until the substrate enters.
Lipases are ˛/ˇ hydrolases (EC 3.1.1.3) which catalyze the
hydrolysis of triglycerides in vivo. They can also form ester bonds
under reverse hydrolytic conditions, which enables them to cat-
alyze esterification and transesterification reactions. Furthermore,
they are enantioselective catalysts useful in the synthesis of phar-
maceutical intermediates and fine chemicals. Lipase B from Candida
antarctica, CALB, has found widespread applications in the enan-
tioselective synthesis of bioactive molecules and in the resolution
of racemic mixtures, due to its high stability in organic media and
its large-scale availability [1].
CALB catalyzed resolution of secondary alcohols by transester-
ification with esters as acyl donors, occurs through a Ping Pong
Bi Bi mechanism, which includes two steps. The first is the acyla-
tion of the enzyme by the ester substrate, to yield the acyl–enzyme
∗ Corresponding author. Tel.: +33 5 46 45 86 30; fax: +33 5 46 45 82 65.
E-mail address: mgraber@univ-lr.fr (M. Graber).
In 2000, Lee et al. [5] described a new approach to enhance
lipase enantioselectivity by the “substrate matching” strategy.
1
381-1177/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2012.04.017

5
6
L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
Resolution of three different secondary alcohols with three dif-
ferent acyl donors by lipase-catalyzed transesterification with
C. antarctica and Pseudomonas cepacia lipases were performed.
Results demonstrated that the enantioselectivity of lipases was
maximized by using acyl donor and alcohol substrates which
matched well. The hypothesis of the “enzyme memory” induced by
the acyl donors active site moulding in the first step of the reaction
was proposed. This work differs from previous studies in that the
buffer (pH 7.5, 10 mM), and dry Chromosorb P AW DMCS (1 g)
was added to the solution. The amount of immobilized enzyme
was determined by measuring absorbance at 280 nm, by taking a
molar extinction coefficient equal to 40,690 M cm . After vigor-
ous shaking, the preparation was left for 1 week under vacuum and
−
1
−1
over P O5 at room temperature.
2
2.3. Enzymatic reactions
“
imprinting” molecule is consistently in contact with the enzyme
◦
during the reaction, because it is the first substrate of the reaction.
Here we used this “imprinting” method and experimentally
highlight the significant influence of the alkoxy part of the ester acyl
donor on the enantiomeric ratio, for the resolution of pentan-2-ol
by CALB. We then established the full kinetic model for a Ping Pong
Bi Bi mechanism with two competing chiral alcohol substrates, in
order to verify that the differences in enantiomeric ratio, obtained
with different acyl donors, did not simply arise from differences
in reaction rates occurring during the acylation step, with the dif-
ferent esters. Our data from both experimental and kinetic studies
support the hypothesis of molecular imprinting. We then looked
for structural changes using molecular modeling methods.
Molecular modeling is a useful tool to provide a rational expla-
nation of experimental data. In 2010, Lousa et al. provided a
structural explanation for the imprinting effect [6] observed with
pre-treated subtilisin by co-lyophilization with an inhibitor in the
active site, using a molecular modeling approach. Results showed
that in the presence of the inhibitor, the active site was maintained
in an open conformation which was stable in hexane solvent, in
contrast to simulation with “untreated” enzyme. Here, 20 ns molec-
ular dynamics simulations were carried out to study how the first
step of the reaction, through the first tetrahedral intermediate,
affects the enzyme conformation, depending on the enantiopure
ester substrate used.
Initial rate of reaction measurements were performed at 70 C
in a solid-gas reactor as previously described [10]. Thermody-
namic activities for ester and alcohol substrates were respectively
aester = 0.1 and aalcohol = 0.05. Reactions were carried out in anhy-
drous conditions. The amount of enzyme comprised between 20
and 200 mg, depending on the acyl donor used. The total flow was
−
1
equal to 900 ␮mol mol
.
2.4. GC analysis
Quantitative analysis of reaction products were conducted using
a 7890 GC system from Agilent for the analysis of ester products
◦
◦
−1
(R)-1- and (S)-1-methylbutyl propanoate (55 C 15 min, 3 C min ,
◦
−1
85 C 5 min), at a flow rate of 1.5 ml min with a Chirasil-Dex CB
(25 m, 0.25 mm i.d., 0.25 ␮m ˇ-cyclodextrin, Chrompack, France)
column. Products were detected by FID and quantified using HP
Chemstation software.
2.5. Enantioselectivity measurements
Enantiomeric ratio values for the different kinetics reported in
Section 4, were obtained in our laboratory, by measuring the ratio
of initial reaction rates for ester products synthesis [11], in a contin-
uous solid-gas reactor with different acyl donors, and immobilized
CALB, as previously described [9,12].
2
. Experimental
3. Computational methods
2.1. Chemicals
3.1. Setup of the system
Substrates and other chemicals were purchased from
The starting CALB enzyme was the R = 1.55 A˚ crystallographic
Sigma–Aldrich–Fluka Chemical Co. They were of the highest
purity available (98% minimum) and checked by gas chromatog-
raphy before use. Substrates were dried by distillation under
argon prior to use and stored under argon atmosphere and over
molecular sieves. Solvents were purchased from Carlo Erba.
Racemic 1-methylpentyl propanoate was synthesized from the
corresponding alcohol and propanoic anhydride in pyridine at
room temperature [7].
structure solved by Uppenberg et al. [13] (PDB entry 1TCA). To
evaluate the effect of the ester substrate on the enzyme struc-
ture during the first step of the reaction path, the two tetrahedral
intermediates, obtained in the reaction with R or S 1-methylpentyl
propanoate were modelized. The choice of studying the intermedi-
ates, instead of free substrates, in the active site was done in order
to prevent the substrates from getting out of active site, observed
several times in the case of subtilisin by Lousa et al. [6]. Further-
more, the formation of the tetrahedral intermediate may have
more impact on the structure conformation, because its formation
requires the crossing of the energy barrier. Thus, three systems
were modelized: free enzyme, enzyme with R and S tetrahedral
intermediates.
A transition state analog crystal structure, obtained with phos-
phonate irreversible inhibitor (PDB entry 1LBS) was used to build
the tetrahedral part of the reaction intermediate, to allow for the
correct location of the central part of the tetrahedral intermediate.
The acyl part is a propanoyl group. The negatively charged oxy-
gen was oriented toward the oxyanion hole to establish hydrogen
bonds with Thr40 and Gln106.
(
R)-1-methylpentyl propanoate was obtained by enzymatic
resolution from vinyl propionate and hexan-2-ol using CALB
®
◦
Novozym 435 in heptane solvent at 35 C, eep enantiomeric
excess of ester product was 99.3%. Enriched hexan-2-ol in S
form taken from the previous reaction was used after purifica-
tion by chromatography on silica gel (eluent EP/AcOEt: 95/5), then
esterification with anhydride propionic was done to obtain 1-(S)-
methylpentyl propanoate.
2
.2. Enzyme used for kinetic studies
CALB was produced in the methylotropic yeast Pichia pastoris
and was expressed extracellularly and purified from the medium
by hydrophobic interaction chromatography, followed by gel fil-
tration [8,9]. Enzyme adsorption was performed onto 60/80 mesh
Chromosorb P AW DMCS (acid washed dimethylchlorosilanized)
NAMD 2.7 program and the CHARMM22 all-atom force field
were used. Calculations were done in an explicit water box (model
TIP3P) with boundary conditions (15 A˚ between the enzyme and the
edge of the box). A waterbox was used for the calculations because
it is in accordance with the CHARMM force field parameters def-
inition. It is supposed to increase the flexibility of the enzyme,
(
Varian, France). In a typical adsorption procedure for solid/gas
catalysis, enzyme (0.106 mg) was dissolved in sodium phosphate

L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
57
Table 1
therefore offering the possibility to make enzyme “deformation”
easier, due to the ester substrate in the active site in the 20 ns
trajectories done for each system. The thermodynamic ensemble
is “isotherm-isobar” (NTP). The timestep was 2 fs and the SHAKE
algorithm was used to freeze bonds involving hydrogen atoms.
Force field parameters for the tetrahedral intermediate were
taken from the literature [14]. These parameters were obtained
from ab initio calculations and were specifically developed for
CHARMM22 force field. Other parameters required for modeling
the alkyl side chains of alcohols were defined by homology with
available CHARMM22 parameters.
Enantiomeric ratio for CALB catalyzed transesterification involving pentan-2-ol with
different alkyl propanoate esters, in solid-gas reactor at 70 C.
◦
Acyl donor ester
E
O
O
O
O
O
O
O
Methyl propanoate
62
O
O
O
O
O
O
O
Ethyl propanoate
Propyl propanoate
Butyl propanoate
Pentyl propanoate
117
103
115
105
51
First, water molecules surrounding the enzyme were minimized
by 5000 iterations of the conjugate gradient, then the whole system
was minimized with 10,000 iterations using the same algorithm.
The heat steps were carried out in 600 ps, starting from 50 K and
going up to 300 K, with a temperature incrementation of 1 K every
−1
4
ps. An harmonic constraint of 5 kcal mol
was set up on the
enzyme. The equilibration step is the succession of four short
dynamics of 200 ps with a decreasing harmonic constraint (5, 3,
−1
0
.5 and 0.1 kcal mol ) followed by one nanosecond without con-
straint. Then, the production dynamic lasted for 20 ns.
1
1
-Methylpropyl propanoate
3.2. Clustering analysis
-(± )-Ethylbutyl propanoate
84
Clustering analysis provides a good overview of enzyme confor-
mations. 2000 structures, extracted from the productive dynamics
one every 10 ps), were used for the dynamic analysis. Using the
VMD program [15], RMSD (root-mean-square-deviation) matrices
were calculated for the 2000 structures, one diagonal matrix with
a size of 2000 by 2000 was obtained for each of the three systems.
RMSD was calculated on the backbone for residues 35–80,
O
O
O
(
1
1
-(± )-Methylpentyl propanoate
-(R)-Methylpentyl propanoate
122
140
72
O
O
O
1
00–240 and 260–290. Residues far from the active site were not
included, because their mobility is not supposed to influence the
active site conformation, which is the region putatively involved
in the imprinting effect. In addition, terminal regions are highly
mobile (and far from the active site for CALB) and over-influenced
the RMSD matrix, therefore, they were not included. In the manner,
we aimed to obtain conformational information specific to the rest
of the structure and more particularly near the active site.
In a second step, RMSD matrix was calculated to focus on the
residues of the active site. RMSD were based on every heavy atom,
including side chains, of residues 103, 104, 106, 224, 187, 40, 42, 47,
1-(S)-Methylpentyl propanoate
esters with branched alkyl chains were also evaluated. The race-
mate 1-methylpropyl propanoate gives an E value of 51, which
is quite close to the E value (62) for methyl propanoate. Higher
enantiomeric ratios were obtained with the longer racemate 1-
ethylbutyl propanoate (ratio of 84) and with 1-methylpentyl
propanoate, (ratio of 122).
2
78, 282, 285, and the backbone of residue 105 (due to the fact that
Enantiomeric ratios with the enantiopure (R)-1-methylpentyl
propanoate and (S)-1-methylpentyl propanoate, were measured
and found to be 140 and 72, respectively. Thus, the enantiomeric
form of the chiral ester substrate is essential for determining
the enantioselectivity of the reaction. CALB displays enantiopref-
erence for the R alcohol [9,17], and using an ester with the R
chiral form for the alkyl part increases the enantioselectivity com-
pared to the racemate. In contrast, an ester with the S chiral
form for the alkyl part, results in decreased enantioselectivity.
Therefore, the more the alkyl part of the chiral ester resem-
bles the preferred enantiomer, (R)-pentan-2-ol, the higher the
enantioselectivity attained. A similar observation was made by
González-Sabína et al. [18], who obtained higher enantioselectivity
when the alkoxy group of the acyl donor was structurally close to
the amine to be resolved. As a consequence, an improved resolution
of (± )-cis-2-phenylcyclopentanamine was obtained with the acyl
donor phenylmethyl ethanoate (E value = 37), compared to ethyl
ethanoate (E value = 17).
the side chain changes for each system for this residue). This RMSD
matrix was then used to process a hierarchical ascendant classifica-
tion (HAC). The Ward method [16] was applied to the agglomerative
steps used to build the dendrogram. Clustering analysis was carried
out using the R statistical software. The average structure of the
two most representative structures from productive dynamics was
chosen. Then, the closest structures to the average structure were
extracted and used for analysis.
4
. Results and discussion
4.1. Experimental results
Enantiomeric ratios experimentally determined for the reso-
lution of pentan-2-ol by transesterification with various alkoxy
propanoates as acyl donors are presented in Table 1. Resolution
of pentan-2-ol with methyl propanoate displays an enantiomeric
ratio of 62. Esters with longer linear alkyl chains, from ethyl
to pentyl propanoate give higher enantiomeric ratios up to 117.
Enantiomeric ratios equal to 117, 103, 115 and 105 were found
for ethyl propanoate, propyl propanoate, butyl propanoate and
pentyl propanoate, respectively. Enantiomeric ratios for chiral
4.2. Kinetic equation study
To confirm whether the differences in enantioselectivity are
due to an “imprinting effect” caused by the leaving alcohol, it was

5
8
L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
Fig. 1. Cleland representation. The enzyme (E), the acyl enzyme (F), the ester (A),
the (R) and (S) alcohols (respectively B and C), the leaving alcohol product (P), the
(R) and (S) ester products (respectively Q and S).
necessary to check that these differences, obtained using differ-
ent acyl donors, do not simply arise from differences in reaction
rates occurring during the acylation step with the different esters.
The complete kinetic model for the Ping Pong Bi Bi mechanism,
with two competing chiral alcohol substrates, was established. The
enantiomeric ratio was then expressed as a function of individ-
ual catalytic rate constants of the reaction, in order to investigate
whether the catalytic rate constants involved in the acylation step
influence the E value.
Fig. 2. King–Altman representation. The enzyme (E), the acyl enzyme (F), the ester
A), the (R) and (S) alcohols (respectively B and C), the leaving alcohol product (P),
the (R) and (S) ester products (respectively Q and S).
(
R
init.
the forward velocities for the R and S ester products synthesis (v
and v Ri nit.) were obtained:
R
4
.2.1. Kinetic model determination
The enantiomeric ratio is defined as the ratio of specificity con-
v
n4[A][B]
init.
=
=
(4.1)
(4.2)
[
E]t
d [A][C] + d [A][B] + d13[C] + d14[B] + d15[A]
8
9
stants for R and S enantiomers, according to the following formula:
E = (k /K )/(k /K ). The kinetic parameter determination is
R
cat
R
M
S
cat
S
M
S
init.
v
n8[A][C]
d [A][C] + d [A][B] + d13[C] + d14[B] + d15[A]
straightforward in the case of a monosubstrate reaction following
the classic Michaelis mechanism. The transesterification studied
here corresponds to a much more complex kinetics system. It
involves a first substrate ester and two competing second sub-
strates, R and S forms of the secondary alcohol. It obeys a Ping Pong
Bi Bi mechanism. Classic kinetic experiments provide apparent con-
stants KM and Vmax, which are dependent on the catalytic rate
constant of the first step of the mechanism. The Michaelis–Menten
constant for the R alcohol KR is thus equal to k (k + k )/k (k + k )
[
E]t
8
9
The complete formulas, with detailed values of ni and di are
reported in Appendix A.
4
.2.2. Kinetic parameter determination
The Cleland nomenclature [21] allowed the calculation of the
Michaelis–Menten constant KM for (R) and (S) alcohols (respec-
2
3
4
3
4
2
tively KR and K ) and maximum reaction rates VR and V
.
max
S
S
M
M
M
max
(
Fig. 2), where k depends on the leaving alcohol in the first reaction
2
The parameters determination was based on coefficients from the
global equations (4.1) and (4.2).
step. This observation could explain the enantiomeric ratio modifi-
cation observed when the leaving alkoxy group changes. Therefore,
the relationship between the acylation step rate and enantiomeric
ratio is worth considering. Here, the full kinetic model for a Ping
Pong Bi Bi mechanism involving one ester and two competiting
alcohols substrates was defined. The kinetic equation was calcu-
lated using the King–Altman method and specificity constants were
determined with the Cleland method.
We focused on the resolution of a racemic mixture of R and
S enantiomeric forms of pentan-2-ol, through acyl transfer from
an ester substrate. In the reaction model a second pathway for
the reaction with the second enantiomer was added, as shown in
Figs. 1 and 2, for Cleland and King–Altman representations.
The kinetic profiles of multisubstrate systems can be resolved
using the King–Altman method [19,20], which, in the present
work, has been devised as an interactive web form by BioKin Ltd.
R
S
V
k k
V
k k
2 6
max
2
4
R
cat
max
S
cat
=
= k
=
= k
(4.3)
(4.4)
[
E]t
k4 + k2
[E]t
k6 + k2
k2 (k + k )
k2 (k + k )
−5
R
M
−3
4
4
S
6
K
=
K
=
M
k (k + k )
k5 (k + k )
3
2
6
2
The four parameters KR , K , V
R
R
and VS enabled the calcula-
M
M
max max
tion of the enantiomeric ratio E, according to the following formula
E = (k /K )/(k /K ):
R
cat
R
M
S
cat
S
M
k3 k4 (k−5 + k6)
k5 k6 (k−3 + k4)
E =
(4.5)
Thus, it appears that the E value is not controlled by catalytic
rate constants involved in the acylation step, the first part of the
reaction (k , k , k , k ), (c.f. Figs. 1 and 2), indicating that the
nature of the leaving alcohol did not influence the enantiomeric
ratio E, through kinetic effects.
(
available at: http://www.biokin.com/king-altman/). It was used to
1
−1
2
−2
obtain the reaction velocity as a function of individual catalytic rate
constants presented above in Cleland and King–Altman represen-
tations.
The model provides complex equations, whose detailed expres-
sion is given in Appendix A. Equations were simplified by
considering the system in the absence of products P, Q and S. Indeed,
all reaction velocities were measured under conditions of initial
rate of reaction, i.e. with negligible product concentrations. Thus,
In addition, the ratio of initial reaction rates vR /vS was equal
init. init.
to:
R
init.
S
init.
v
v
n4[A][B]
n6[A][C]
−5
k k (k + k ) [B]
3
4
6
=
=
(4.6)
k5 k6 (k−3 + k4) [C]

L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
59
Therefore, when reaction velocity is measured under conditions
of initial rate, where B and C are enantiopure alcohols in racemic
mixture, then the ratio of forward velocities vR /vS is equal to:
init. init.
R
v
v
−5
k k (k + k )
init.
3
4
6
=
= E
(4.7)
S
init.
k5 k6 (k−3 + k4)
R
S
Thus, the enantiomeric ratio E is equal to the ratio v /v in con-
ditions of initial rate (insignificant concentration of products) and
with a racemic mixture of alcohols at the initial step of the reaction.
We can conclude that measuring the ratio of initial reaction rates
R
S
v /v is a valid method to determine the enantiomeric ratio E.
R
S
Similarly, the relationship between the ratio v /v in conditions
of initial reaction and E was demonstrated by Chen [11] in the case
of the simple Michaelis Menten model. Furthermore, Chen’s propo-
sition remains correct in case of the Ping Pong Bi Bi system with R
or S as competitive alcohol substrates.
We have confirmed here, then, that (1) E values can be correctly
determined by measuring the ratio of initial reaction rates for enan-
tiopure ester synthesis, (2) E values do not depend on catalytic rate
constants involved in the acylation step.
Fig. 3. View of the active site for the six clusters obtained for the three studied
systems after alignment based on the heavy atoms of residues 103, 104, 106, 224,
187, 40, 42, 47, 278, 282, 285 and the backbone of residues 105. The color code for
the clusters is Clust1 for Enz-IT-S in blue, Clust2 for Enz-IT-S in light blue, Clust1 for
Enz-IT-R in red, Clust2 for Enz-IT-R in orange, Clust1 for enzyme free in gray and
Clust2 for enzyme free in dark gray. (For interpretation of the references to color in
this figure legend, the reader is referred to the web version of the article.)
4
.3. Molecular modeling results
The results presented above suggest that there is an imprint-
ing effect: the first substrate of the reaction and in particular
the alkoxy part of the ester causes a conformation change of
the enzyme, which is “memorized” by the enzyme and modi-
fies its ability to discriminate between enantiomers of the second
alcohol substrates. Interesting results, concerning the sensitivity
of the enantioselectivity in relation to the enantiomeric form of
the leaving alcohol, indicate that the imprinting effect involves
modifications near the active site (Table 1): E obtained with (R)-
1 s) [24]. Experimental results show that buried side chains rotate
very slowly compared with the time scale of molecular dynam-
ics. NMR was used by Skrynnikov et al. [25] to quantitate slow
hydrogen–deuterium exchange processes at methyl-containing
side chains in proteins. This method was also applied to the study
of ms time scale side-chain dynamics of methionine residues in a
buried cavity. These authors observed that the methionine residues
1
-methylpentyl propanoate as acyl donor is equal to 140, whereas
were sensitive to an exchange event with a rate of the order of
it is equal to 72 with (S)-1-methylpentyl propanoate. Our attempts
to confirm this hypothesis by molecular modeling are presented
below.
In the first clustering study we focused on global enzyme confor-
mation. None of the clusters display global enzyme conformation
rearrangement. The only difference between clusters concerns the
position of a helix above the active site: results reveal flexibility of
the alpha helix 5, which has already been described by Skjøt et al.
1200 s−1 at 20 ◦C and that the corresponding motions may be linked
to a process which allows entry and exit of ligands to and from
the cavity. Similar NMR studies on a protease, by Ishima et al.
[26], demonstrated that the hydrogen–deuterium exchange time
of buried methyl side chains was above 1 ms.
The side chain of residue Ile285 is partially buried, as it is ori-
ented toward the top of the stereospecificity pocket in the active
site, and near the side chain of the alcohol. The time scale of Ile285
side chain rotation can thus be considered to be around 1 ms. Pre-
viously the same order of magnitude was obtained for pentan-2-ol
[
22] and is not due to the presence of the substrate in the active
site.
We then focused on amino acids near the active site, in order
to compare active site structures of free enzyme, enzyme + R ester
and enzyme + S ester first tetrahedral intermediates. Cluster analy-
sis based on amino acids of the active site was done as described in
Section 3. Amino acids alignment gave good superimposed struc-
tures, including side chains orientation (c.f. Fig. 3).
3
00
250
Interestingly however, the orientation of the side chain of
residue Ile285 was different for the cluster 2 of TI-R. This may be
due to the specific constraint generated by the alcohol enantiomer
on the side chain orientation which pointed toward Ile285. In the
2
00
150
◦
case of the R enantiomer, the side chain of Ile285 rotated by 120
in around 10 ns (Fig. 4). The cluster analysis was consistent with
this fact, and split the trajectory into two dominant clusters, one
before the rotation, and the second after it. Side chain orientation
in the TI-S system is the same as that observed with free enzyme.
Residue Ile 285 belongs to ˛ helix 10. Marton et al. demonstrated
that mutations of residues Leu282 and Ile282 of the ˛ helix 10
affected enantioselectivity [23]. Thus, side chain rotation of Ile285
may also influence the enantioselectivity of the reaction.
1
00
50
0
0
5
10
15
20
Time (ns)
A major question concerns the timescale of side chain rota-
tion, such as of the branched-chain of Ile285. The time scale for
Fig. 4. Diheral angle (atoms CA–CB–CG1–CD) defining the side chain rotation of
residue Ile285. Solid line for the system with the R enantiomer and dashed line with
the S enantiomer.
−
4
the rotation of a buried side chain can be very large (10
to

6
0
L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
−
1
vR
[E]t
transesterification in a solid-gas reactor [17]: kcat equal to 800 s
NR
D
=
=
(A.3)
(A.4)
−1
and 17 s for (R)- and (S-pentan-2-ol respectively. The similarity
of these two time scales namely, side chain rotation and sub-
strate catalysis, is consistent with the hypothesis of “imprinting
effect”. Ile285 side chain rotation provides a much more suitable
active site shape for interacting with the substrate (R)-enantiomer
alcohol. This is exactly what is experimentally observed: enan-
tiopreference for the R form of pentan-2-ol increases when the
vS
NS
D
[
E]t
Expression of numerators and denominator, NR, N , and D:
S
NR = +n [Q][P] + n [C][Q] + n [B][S] + n [A][B]
(A.5)
(A.6)
1
2
3
4
(
R)-1-methylpentyl propanoate is used as acyl donor.
NS = +n5[S][P] + n [C][Q] + n7[B][S] + n [A][C]
6
8
Other authors also performed molecular modeling studies to
explore enzyme structural changes upon imprinting. Rich and
Dordick [27] obtained an increase of subtilisin catalytic rate and
also a better control of enzyme substrate specificity, by means
of lyophilizing subtilisin in presence of different nucleophile sub-
strates, as “imprinters”. By molecular dynamic simulations it was
shown that structural changes in the catalytic triad occurred during
imprinting, that may contribute to imprinting-induced substrate
selectivity.
D = +d [P][S] + d [Q][P] + d [C][S] + d [C][Q] + d5[B][S]
1
2
3
4
+ d [B][Q] + d [A][P] + d [A][C] + d [A][B] + d10[S]
7
6
8
9
+d11[P] + d12[Q] + d13[C] + d14[B] + d15[A]
(A.7)
n1 = −k−1k−2k−3k−4k−5 − k−1k−2k−3k−4k+6
(A.8)
(A.9)
n2 = −k−1
n3 = +k−1
n4 = +k+1
n5 = −k−1
k
k
k
k
k
k
k
k
k
k
k+5k − k+2
+6
k
k
k
k
k
k
k
k+5k
k−5k
−3
+3
+2
−2
−4
+4
+3
−3 −4
+6
5
. Conclusion
k−5k + k+2
−6
k
(A.10)
(A.11)
(A.12)
(A.13)
(A.14)
(A.15)
+3 +4
−6
Experimental results shown here demonstrate that using an
k+4k−5 + k+1
k
k
k
+2 +3 +4 +6
ester with an adequate alkoxy group is an efficient method to
enhance enantioselectivity. Generally, an alkoxy group larger than
ethyl increased enantioselectivity. Furthermore, the resolution of
pentan-2-ol was sensitive to the chirality of the alkoxy group of the
ester. Thus, (R)-1-methylpentyl propanoate increased enantiose-
lectivity compared to the racemic mixture, whereas S enantiomer
decreased enantioselectivity compared with the racemic mixture.
The comprehensive study of the full kinetics for the Ping Pong
Bi Bi mechanism, with three substrates, one ester and two compet-
itive R and S alcohols, allowed us to confirm that the experimental
method, based on initial rate measurements employed here for
the determination of enantiomeric ratio, is relevant. In particular,
it excluded the hypothesis that enantiomeric ratio modifications
observed in the experimental results could arise from a kinetic
model pitfall.
Finally, molecular dynamics simulations were performed to dis-
criminate between conformational changes caused by both (R)-
and (S)-enantiopure 1-methylpropyl propanoate. It appears that
the R enantiomer causes the rotation of the side chain of residue
Ile285, which appears to have an effect on subsequent discrimina-
tion between secondary alcohol enantiomers. If this model study
can be extrapolated for general use, molecular imprinting by the
first substrate would offer the possibility of controlling enantios-
electivity for a second substrate and thus provide a new tool for
biocatalyst engineering.
−3 −5 −6
k
k
− k−1
k
−2 +4 −5 −6
k
k
n6 = +k−1
−3 −4
k
k+5k
+6
+ k+2
−3 −4
k
k+5k
+6
−6
n7 = −k−1
k
+4
k−5k − k+2
−6
k
k−5k
+3
+3 +4
n8 = +k+1k+2k−3k+5k+6 + k+1k+2k+4k+5k+6
d1 = k−2
k
k
−5 −6
k
+ k−2
k
k
k
k
k
k
k
k
−3 −5 −6
k
k
+ k−1
k
k
k
+4
−2 +4 −6
+
k
1k
k
k
(A.16)
(A.17)
(A.18)
(A.19)
(A.20)
(A.21)
(A.22)
(A.23)
−
−2 −3 −6
d2 = k−2
k
−3 −4 +6
k
k
+ k−2
−3
−3
−4
+3
+3
−2
−3
k
−4k−5 + k
−1 −2 −4 +6
k
k
k
+
k
1k
−2 −4
k
k−5
−
d3 = k−1
k
+4
k+5k + k−1
−6
k+5k + k+2
−6
k
k
k
k+5k
+4 −6
+ k+2
k
−3
k
+5 −6
k
d4 = k−1
k
−4
k+5k + k+2
+6
k+5k + k−1
+6
k
k+5
−3 −4
+ k+2k−3k−4k+5
Acknowledgements
d5 = k−1
k
+3
k−5k + k+2
k−5k + k−1
−6
k
k
−6
+3 +4 −6
This study was supported by the French ANR (National Research
Agency) through the EXPENANTIO project (program CP2D). CINES
and IDRIS from GENCI (Grand Equipement National de Calcul
Intensif) are acknowledged for giving access to super-computing
facilities. The manuscript was corrected by a native English speak-
ing scientific translator (http://traduction.lefevere-laoide.net)
+
k
2k
k
k
+
+3 +4 −6
d6 = k−1
k
+3 −4 +6
k
k
+ k−1
k
−4k−5 + k
+2 +3 −4 +6
k
k
k
+
k
2k
+3 −4
k
k−5
+
Appendix A.
d7 = k+1
k
−2
k
k
+ k+1
k+4k−5 + k+1
k
k
k
+4 +6
−2 −3 +6
Speed rate equation for synthesis of R and S enantiomer products
v and v :
R
S
+ k+1k−2k−3k−5
d[Q]
dt
vR
=
= k+4[FB] − k−4[Q][E]
= k+6[FC] − k−6[S][E]
(A.1)
(A.2)
d8 = k+1
k
+4
k+5k + k+1
k+5k + k+1
+6
k
k
k+5
+6
+2 +4
d[S]
dt
S
v =
+ k+1
k
+2
k
−3 +5
k

L. Chaput et al. / Journal of Molecular Catalysis B: Enzymatic 84 (2012) 55–61
61
References
d9 = k+1
k
+3
k
+4 +6
k
+ k+1
k
+3
k
+4k−5 + k
+1 +2 +3 +6
k
k
k
[1] M.T. Reetz, Curr. Opin. Chem. Biol. 6 (2002) 145–150.
[
2] J.-Y. Yan, Y.-J. Yan, J.-K. Yang, L. Xu, Y. Liu, Proc. Biochem. 44 (2009)
128–1132.
+
k
1k
k
k−5
(A.24)
(A.25)
(A.26)
(A.27)
(A.28)
(A.29)
(A.30)
+
+2 +3
1
[
[
3] J. Yang, L. Liu, X. Cao, Enzyme Microb. Technol. 46 (2010) 257–261.
4] G. Hellner, Z. Boros, A. Tomin, L. Poppe, Adv. Synth. Catal. 353 (2011)
2481–2491.
d10 = k−1
k
+4
k−5k + k−1
−6
k
−3
k
−2
k
−3
k
−3
k
+3
k
+2
k−5k + k+2 k−5k
k
−6 +4 −6
[
[
[
5] D. Lee, Y.K. Choi, M.-J. Kim, Org. Lett. 2 (2000) 2553–2555.
6] D. Lousa, A.M. Baptista, C.M. Soares, Protein Sci. 20 (2011) 379–386.
7] J.C. Rotticci-Mulder, M. Gustavsson, M. Holmquist, K. Hult, M. Martinelle, Pro-
tein Expr. Purif. 21 (2000) 386–392.
+
k
2k k−5k
+
−3
−6
[
[
8] A. Ducret, M. Trani, R. Lortie, Enzyme Microb. Technol. 22 (1998) 212–216.
9] V. Leonard, S. Lamare, M.-D. Legoy, M. Graber, J. Mol. Catal. B: Enzym. 32 (2004)
53–59.
d11 = k−1
k
−2
k
+4 +6
+ k−1
k
k
+4k−5 + k
−1 −2 −3 +6
k
k
k
[
[
[
[
10] F. Létisse, S. Lamare, M.-D. Legoy, M. Graber, Biochim. Biophys. Acta 1652 (2003)
27–34.
11] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C.J. Sih, J. Am. Chem. Soc. 104 (1982)
+
k
1k
−2 −3
k−5
k
−
7294–7299.
12] V. Leonard-Nevers, Z. Marton, S. Lamare, K. Hult, M. Graber, J. Mol. Catal. B:
Enzym. 59 (2009) 90–95.
13] J. Uppenberg, M.T. Hansen, S. Patkar, T.A. Jones, Structure 2 (1994) 293–308.
d12 = k−1
k
−3
k
−4 +6
+ k−1
k
k
−4k−5 + k
+2 −3 −4 +6
k
k
k
+
k
2k
−3 −4
k−5
k
+
[14] N. Otte, M. Bocola, W. Thiel, J. Comput. Chem. 30 (2009) 154–162.
[
[
[
15] W. Humphrey, A. Dalke, K. Schulten, J. Mol. Graph. 14 (1996) 33–38.
16] J.H. Ward, J. Am. Stat. Assoc. 58 (1963) 236–244.
17] L. Chaput, Y.-H. Sanejouand, A. Balloumi, V. Tran, M. Graber, J. Mol. Catal. B:
Enzym. 76 (2012) 29–36.
18] J. González-Sabína, V. Gotor, F. Rebolledo, Tetrahedron: Asymmetry 15 (2004)
481–488.
19] E.L. King, C. Altman, J. Phys. Chem. 60 (1956) 1375–1378.
20] I.H. Segel, Enzyme Kinetics, Wiley, New York, 1993, pp. 606–625.
21] W. Cleland, Biochim. Biophys. Acta 67 (1963) 104–137.
[22] M. Skjøt, L. De Maria, R. Chatterjee, A. Svendsen, S.A. Patkar, P.R. Østergaard, J.
Brask, Chembiochem 10 (2009) 520–527.
[23] Z. Marton, V. Léonard-Nevers, P.-O. Syrén, C. Bauer, S. Lamare, K. Hult, V. Tran,
M. Graber, J. Mol. Catal. B: Enzym. 65 (2010) 11–17.
24] S.J. Benkovic, S. Hammes-Schiffer, Science 301 (2003) 1196–1202.
25] N.R. Skrynnikov, F.A.A. Mulder, B. Hon, F.W. Dahlquist, L.E. Kay, J. Am. Chem.
Soc. 123 (2001) 4556–4566.
d13 = k−1
k
+4
k+5k + k−1
+6
k+5k + k+2
+6
k
k
k
k+5k
+4 +6
[
+
k
2k k+5k
+
−3
+6
[
[
[
d14 = k−1
k
+3
k
k
+ k−1
k+4k−5 + k+2
k
k
+4 +6
+3 +4 +6
+
k
2k
k
k−5
+
+3 +4
[
[
d15 = k+1
k
+2
k
k
+ k+1
k+4k−5 + k+1
k
k
+4 +6
k−5
+2 −3
+2 −3 +6
[
26] R. Ishima, J.M. Louis, D.A. Torchia, J. Am. Chem. Soc. 121 (1999) 11589–11590.
[27] J.O. Rich, J.S. Dordick, J. Am. Chem. Soc. 119 (1997) 3245–3252.
+
k
1k
k
+