Highly efficient resolution of mandelic acid using lipase from Pseudomonas stutzeri LC2-8 and a molecular modeling approach to rationalize its enantioselectivity

Article abstract of DOI:10.1016/j.molcatb.2013.10.026

Mandelic acid, a key precursor of chiral synthons, was successfully acylated in diisopropyl ether. The reaction was catalyzed by the lipase from Pseudomonas stutzeri LC2-8, and vinyl acetate was employed as acyl donor. Under the optimized reaction conditions, a resolution of 180 mM (55 g/L) mandelic acid was achieved. (S)-O-Acetyl mandelic acid was enantioselectivity formed in >99% ee at a yield close to the maximum theoretical value for kinetic resolution (50%). The highly substrate tolerable and enantioselective nature of lipase LC2-8 suggests that it is of great potential for the practical resolution of racemic mandelic acid. Additionally, the high enantiopreference of lipase LC2-8 toward (S)-mandelic acid in acetylation was also rationalized through molecular docking and molecular dynamics simulations.

Full text of DOI:10.1016/j.molcatb.2013.10.026

Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Highly efficient resolution of mandelic acid using lipase from
Pseudomonas stutzeri LC2-8 and a molecular modeling approach to
rationalize its enantioselectivity
a
a
b
a
a,∗
Yan Cao , Shanshan Wu , Jiahuang Li , Bin Wu , Bingfang He
a
College of Biotechnology and Pharmaceutical Engineering, Nanjing University of Technology, 30 Puzhunan Road, Nanjing 211816, People’s Republic of
China
State Key Laboratory of Pharmaceutical Biotechnology, College of Life Sciences, Nanjing University, Nanjing 210093, People’s Republic of China
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 2 August 2013
Received in revised form 31 October 2013
Accepted 31 October 2013
Available online 8 November 2013
Mandelic acid, a key precursor of chiral synthons, was successfully acylated in diisopropyl ether. The
reaction was catalyzed by the lipase from Pseudomonas stutzeri LC2-8, and vinyl acetate was employed as
acyl donor. Under the optimized reaction conditions, a resolution of 180 mM (55 g/L) mandelic acid was
achieved. (S)-O-Acetyl mandelic acid was enantioselectivity formed in >99% ee at a yield close to the max-
imum theoretical value for kinetic resolution (50%). The highly substrate tolerable and enantioselective
nature of lipase LC2-8 suggests that it is of great potential for the practical resolution of racemic mandelic
acid. Additionally, the high enantiopreference of lipase LC2-8 toward (S)-mandelic acid in acetylation was
also rationalized through molecular docking and molecular dynamics simulations.
Keywords:
Lipase
Pseudomonas stutzeri
Mandelic acid
Resolution
Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.
Molecular modeling
1
. Introduction
Enantiomers of mandelic acid and their derivatives have been
using docking simulations, Santaniello succeeded in explaining
the enantioselectivity of the Burkholderia cepacia lipase-catalyzed
transesterification of aromatic primary alcohols with vinyl esters
with different chain lengths [8]. Based on molecular dynamics sim-
ulation, using methyl mandelate as a “model” substrate, Yu’s group
has successfully shed light on the source of enantioselectivity mod-
ified and the trade-off of enantioselectivity and activity in directed
evolution of an esterase from Rhodobacter sphaerioides [9,10]. Thus,
molecular modeling seems to be a good choice to rationalize the
enantioselectivity of the lipase-catalyzed acylation of mandelic
acid.
Recently, we screened and characterized lipase LC2-8 from
Pseudomonas stutzeri [11]. This lipase exhibited significant solvent
tolerance and good enantioselectivity toward 1-phenylethanol. In
the present work, lipase LC2-8 was used in the resolution of man-
delic acid, with vinyl acetate as acyl donor (Fig. 1). Lipase LC2-8
showed high substrate concentration tolerance and excellent enan-
tiopreference toward (S)-mandelic acid. In addition, we developed
a molecular model and explained the strict enantioselectivity of
lipase LC2-8 toward mandelic acid.
considered as important substances because they are utilized
extensively for synthetic purposes as well as in stereochemical
investigations. Mandelic acid enantiomers are employed in the
resolution of racemic alcohols and amines [1,2]. In addition, (R)-
mandelic acid is used as a versatile intermediate for preparation of
semisynthetic cephalosporins, penicillins, anti-tumor agents and
anti-obesity agents [3]. In the last decade, the use of lipases in the
enantioselective resolution of racemic mandelic acid for obtaining
chiral enantiomers has received considerable attention. Although
a high enantiomeric excess (ee value) can be obtained in some
reports [4–7], the relatively low substrate concentration, usually
below 50 mM, restricts its practical application. Therefore, devel-
oping the efficient resolution process and improving its space–time
yield become imperative.
With the rapid development of computer simulations at
present, biomolecular modeling provides a promising way to
understand the mechanism and selectivity of enzyme-catalyzed
reactions and to guide the modification of enzymes. For example,
1.1. Biological and chemical materials
∗ Corresponding author. Tel.: +86 25 58139902; fax: +86 25 58139902.
E-mail address: bingfanghe@njut.edu.cn (B. He).
Lipase LC2-8 from P. stutzeri LC2-8 was screened and charac-
terized in our previous report [11]. The strain of P. stutzeri LC2-8
1
381-1177/$ – see front matter. Crown Copyright © 2013 Published by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.molcatb.2013.10.026

Y. Cao et al. / Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
109
Fig. 1. Resolution of (R,S)-mandelic acid catalyzed by lipase LC2-8.
is currently deposited at the China Center for Type Culture Collec-
1.5. Construction of the homology model of lipase LC2-8
tion (Wuhan, China) with accession number CCTCC M 2010279. The
nucleotide sequence of lipase LC2-8 has been assigned a GenBank
accession number of JN681265.
Mandelic acid was purchased from Merck (Germany). p-
Nitrophenyl palmitate (p-NPP) and high-performance liquid
chromatography (HPLC)-grade solvents were purchased from
Sigma (USA). All other chemicals were of analytical grade.
Lipase LC2-8 is composed of 311 amino acid residues including
a predicted 24-amino acid signal peptide. All computational meth-
ods were carried out using DS2.5 (Accelrys Software Inc., San Diego,
CA) [14]. A search of the NCBI database was done using the Basic
Local Alignment Search Tool (BLAST) [15]. Automated sequence
alignment between lipase LC2-8 and the template sequence was
carried out. The BLAST search showed significant sequence identity
(78.9%) between lipase LC2-8 and Pseudomonas aeruginosa lipase
1.2. Preparation of crude lipase LC2-8
(
Protein Data Bank (PDB) code: 1EX9) possessing the highest bit
score of 450.7. Unanimously, the crystal structure of PAL (resolu-
A˚ ) was selected as the template. The generated structure
Strain LC2-8 was cultured in lipase-producing medium consist-
ing of (w/v) 1.0% yeast extract, 0.8% glucose, 0.2% K HPO , 0.05%
tion 2.5
2
4
MgSO ·7H O, 0.05% (v/v) Triton X-100, and 0.5% (v/v) sunflower
was improved by subsequent refinement of the loop conforma-
tions by assessing the compatibility of an amino acid sequence to
known PDB structures using the Protein Health module [14]. The
geometry of loop regions was corrected using the Refine Loop. The
three-dimensional structure of lipase LC2-8 obtained was further
optimized by energy minimization. The steepest-descent method
was set at 500 steps for first minimization, and was followed
by 1000 steps of the conjugate gradient method. The programs
PROCHECK [16] and Profiles-3D [17] were used to validate the
accuracy of the refined model of lipase LC2-8.
4
2
oil, at a pH of 8.0. The incubations were carried out with shaking
◦
at 180 rpm at 30 C. After 48 h of growth, the culture supernatant
was collected by centrifugation. Chilled acetone was added under
◦
magnetic stirring to the crude lipase at 0 C for 4 h until the ratio
was 0.8:1 (v/v). The precipitate was obtained by centrifugation
◦
at 10,000 × g and 4 C for 30 min. Finally, the precipitate was air-
dried and lipase LC2-8 powder was obtained. The lipase activity
was determined using a modified spectrophotometric method with
p-NPP as substrate [12].
1.3. Resolution of mandelic acid catalyzed by lipase LC2-8
1.6. Docking and molecular dynamics simulations
The resolution of (R,S)-mandelic acid was carried out in iso-
It is known that the transesterification mechanism is based
propyl ether containing 300 mM vinyl acetate and 30 mM mandelic
acid. The reaction was carried out with shaking at 180 rpm, and was
on the acylation and deacylation of the catalytic triad of serine,
and involves two tetrahedral intermediates. The first intermedi-
ate results from the nucleophilic attack of the catalytic serine on
the acylating substrate, and forms the acyl–enzyme complex. The
second derives from the nucleophilic attack of the acyl acceptor
substrate on the acyl–enzyme complex, and leads to the release
of the ester product [18]. Therefore, the Ser82 residue is covalently
bound to the acyl group in order to mimic the acyl–enzyme complex
◦
catalyzed by 10 mg/mL (30 U/mL) lipase LC2-8 powder at 30 C. A
control reaction was done by performing the above procedure in
the absence of enzyme.
1
.4. Enantiomeric analysis of products by HPLC
The concentrations of (R,S)-mandelic acid, (S)-, and (R)-
[
8,19,20].
The acylation of Ser82 was constructed using the Build fragment
O-acetyl mandelic acid were determined by HPLC (Dionex
P680) using a Chiracel OD-H column (250 mm × 4.6 mm) and n-
hexane/isopropanol/TFA (94/6/0.2, v/v/v) as mobile phase, which
was introduced at a flow rate of 1 mL/min. The wavelength of the
UV detector was set at 228 nm. The enantiomeric excess (eep, ees)
and conversion yield (c) were calculated using the corresponding
peak areas:
module of DS2.5. All hydrogen atoms were added in their theo-
retical positions. The catalytic residues Asp228 and His250 were
assigned as deprotonated. After all preparations, the acyl–enzyme
complex was run through 1000 steps of conjugate gradient energy
minimization with the backbone atoms of the enzyme fixed, using
the CHARMM force field.
The docking experiment was performed using the protocol of
Dock Ligand (LibDock) (Receptor–Ligand Interactions module in
DS2.5) [21]. The acyl–enzyme complex was used as receptor and
the active site side chain residues were defined as the binding site
sphere. The ligands with optimized structures were docked into
the active site of the enzyme. The generated conformations were
manually analyzed, and the one corresponding to the lowest energy
and the highest consensus score was selected for the subsequent
energy minimizations.
Enantiomeric excess,
[
[
(S)-O-acetyl mandelic acid] − [(R)-O-acetyl mandelic acid]
(S)-O-acetyl mandelic acid] + [(R)-O-acetyl mandelic acid]
eep =
[
[
(R)-mandelic acid] − [(S)-mandelic acid]
(S)-mandelic acid] + [(R)-mandelic acid]
ees =
Conversion yield [13],
ees
c =
Considering that the docking algorithm treats the protein as a
rigid body, a careful post-docking optimization was carried out on
the enzyme–substrate complexes, in order to take into account
a potential induced fit effect (little displacements in the protein
ees + eep
All reported data are averages of experiments performed at least in
triplicate.

1
10
Y. Cao et al. / Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
Table 1
Table 2
Effect of organic solvents on the acylation of (R,S)-mandelic acid catalyzed by lipase
Effect of temperature on the acylation of (R,S)-mandelic acid catalyzed by lipase
LC2-8.
LC2-8.
◦
Solvent
Log P
Conversion yield (%)
eep (%)
Temperature ( C)
Conversion yield (%)
eep (%)
Diisopropyl ether
Diethyl ether
THF
1.90
0.85
0.5
49
40.5
45.3
99
99
95.3
25
30
35
49
49
49
99
99
99
4
0
51.7
93.4
Reactions were carried out in various organic solvents (2 mL) with (R,S)-mandelic
acid (30 mM), vinyl acetate (300 mM), and lipase LC2-8 (20 mg) at 30 C for 15 h.
◦
Reactions were carried out in diisopropyl ether (2 mL) with (R,S)-mandelic acid
30 mM), vinyl acetate (300 mM), and lipase LC2-8 (20 mg) at 30 C for 15 h.
◦
(
structure due to the presence of the ligand) [20]. The energy mini-
mization consisted of three steps as follows: (1) a fix constraint was
applied on the protein backbone and all the remaining parts of the
system were tethered with a harmonic restraint; (2) the protein
backbone was kept fix, the side chains were kept tethered and the
docked ligand was free to move; (3) only the backbone was kept fix,
with side chains and the ligands free to move. The Conjugate Gra-
dient algorithm was used in the three steps, with root mean square
enzyme is slightly more flexible compared with its state at low
temperature. Therefore, lipase probably loses enantiomeric speci-
ficity at high temperature as a result of the increased flexibility
◦
[
26]. Thus, 35 C was considered as the optimum temperature for
the resolution.
2.2. Effect of substrate molar ratio and mandelic acid
concentration on the efficiency of resolution
−
1
−1
deviation (rmsd) gradients of 0.01, 0.001and 0.0001 kcal mol A˚
,
respectively.
Many investigators hypothesize that increasing the amount of
acyl donor could aid in forming the acyl–enzyme intermediate and
accelerate the reaction rate [27]. As Fig. 2 shows, the conversion
yield increased with increasing molar ratio of vinyl acetate to (R,S)-
mandelic acid up to 8:1, and further increase in the molar ratio of
two reactants did not significantly increase the conversion yield of
mandelic acid. On the other hand, no change in the enantioselectiv-
ity of lipase LC2-8 occurred as the substrate molar ratio increased,
indicating the rigorous enantioselectivity of lipase LC2-8 toward
mandelic acid. Therefore, the optimal substrate molar ratio of vinyl
acetate to mandelic acid for the reaction was thought to be 8:1.
A production process with high yield and high conversion is
of considerable economic significance. In the actual catalytic pro-
cess, enzyme activity is frequently inhibited by the existence of
its own substrate at relatively high concentrations. The effect of
mandelic acid concentration was also investigated. As shown in
Fig. 3, although the conversion yield decreased when mandelic acid
concentration increased from 30 to 180 mM, the acylation rate of
mandelic acid sharply rose even at the same enzyme loading, with-
out effect on the high entantioseletivity of lipase LC2-8. A further
increase to higher concentrations of mandelic acid (e.g., 220 mM)
led to a dramatic decrease of the conversion rate, presumably
via substrate inhibition. After the optimization of enzyme load-
ing (20 mg/mL of final concentration), the conversion yield reached
49.8% with an eep of >99% (data not shown).
In order to evaluate the dynamics of the docked substrates,
further 510 ps MD trajectory was executed on the corresponding
complexes with CHARMm force field, adopting a 12 A˚ nonbound
spherical cut-off, using the isothermal-isochoric ensemble (NVT)
and with distance-dependent dielectric implicit solvent model,
maintaining the backbone fix. In particular, the stability of the
essential hydrogen bond interactions and the critical distances for
reactivity were verified. The trajectories were divided in 5 ps of
heating from 50 to 300 K, 5 ps of equilibration at 300 K and 500 ps
of production at 300 K. Frames were registered every 500 fs. Only
production phases were considered for analysis.
2
. Results and discussion
2.1. Effect of organic solvents and temperatures on the
enantioselectivity of resolution
Enzymatic reactions in organic solvent are known to be highly
sensitive to the solvent selection, especially in hydrophilic solvent
systems. Generally, comparable to the hydrophobic counterparts,
the hydrophilic solvents could more easily strip the “essential
water” bound to the lipase, which is necessary to preserve the flex-
ibility of the enzyme conformation; this phenomenon deactivates
the lipase [22]. However, our previous investigations demonstrated
that lipase LC2-8 showed significant stability in both hydrophobic
and hydrophilic solvent [11]. Since mandelic acid has poor solubil-
ity in hydrophobic solvents, we employed weakly hydrophilic or
hydrophilic solvents such as diisopropyl ether, diethyl ether, and
tetrahydrofuran (THF) as media for resolution of mandelic acid. As
shown in Table 1, diisopropyl ether gave the best result with regard
to enantioselectivity (eep 99%) and reaction rate (15 h for 49% con-
version). When the reaction was carried out in THF, a relatively low
ee value (95.3%) was obtained. This lower value may be due to the
interaction of polar solvent with enzyme that might have resulted
in small changes in the lipase active site [23]. Diisopropyl ether was
therefore chosen as medium for the resolution of mandelic acid in
the subsequent experiments.
2.3. Time course of kinetic resolution of mandelic acid
We monitored the time course of the enantioselective acylation
of mandelic acid catalyzed by lipase LC2-8 under the optimized
conditions. The results show that (S)-O-acetyl mandelic acid was
obtained in 99.8% ee, and the conversion reached 49.15% within
15 h (Fig. 4). There are a number of reports on lipase mediated
hydrolysis of racemic mandelic acid esters as well as the acylation
of racemic mandelic acid [3–7]. Han reported that the enantiose-
lectivity ratio for (R)-mandelic acid (E) was drastically increased
from 29.2 to more than 300 upon immobilization of Burkholde-
ria sp. GXU56 lipase, using octyl sepharose CL-4B as support
[5]. Yu’s group enhanced the enantioselectivity of esterase from
Rhodobacer sphaeroides in hydrolytic kinetic resolution of methyl
mandelate by directed evolution [9]. In contrast, the ability of
lipase from P. stutzeri LC2-8 to efficiently catalyzed the resolution
of (R,S)-mandelic acid demonstrates its intrinsic enantioselectivity
to mandelic acid. Moreover, the lipase from Burkholderia ambifaria
YCJ01 reported by our group also displayed a remarkable enantio-
selectivity toward mandelic acid, while the substrate concentration
In an enzymatic reaction, the temperature significantly influ-
ences the activity, enantioselectivity, and stability of a biocatalyst
[
24,25]. The effect of reaction temperature on enantioselectivity of
mandelic acid was investigated. As shown in Table 2, lipase LC2-8
maintained strict enantioselectivity at temperatures between 25
◦
and 35 C; the eep was higher than 99%. With further increase
◦
of temperature up to 40 C, a decrease in enantioselectivity was
observed. Generally, at higher temperature, the structure of the

Y. Cao et al. / Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
111
Fig. 2. Effect of substrate mole ratio on the acylation of (R,S)-mandelic acid catalyzed by lipase LC2-8. Reactions were carried out in diisopropyl ether (2 mL) with lipase LC2-8
◦
(
20 mg) at 30 C for 15 h. The concentration of (R,S)-mandelic acid was kept constant (30 mmol), whereas the concentration of vinyl acetate was varied from 30 to 300 mM.
Symbols: (᭹) conversion yield; (ꢀ) ees; (ꢀ) eep.
was about 30 mM [7]. From the viewpoint of practical applica-
tion, high substrate concentration would be beneficial for practical
application of an enzymatic process because it could improve the
space–time yield and reduce the cost of product isolation to a large
extent. In the present study, the concentration of (R,S)-mandelic
acid reached 180 mM (55 g/L), which is significantly higher than
those in other reports which expounded the resolution of mandelic
acid with a substrate concentration below 50 mM [3–7]. The highly
substrate tolerance and enantioselective nature of lipase LC2-8 sug-
gests that it is of great potential for the practical resolution of
racemic mandelic acid.
with a high percentage of residues in the favored and allowed cat-
egories more plausibly represents the folding of the protein. Thus,
PROCHECK validates the folding integrity of our model and indi-
cates that the model structure derived from the template (1EX9) is
of higher quality in terms of protein folding. The Profile-3D score of
the model is 178.23 against 197.15, the maximum expected score.
The built model was also evaluated by superimposing the template
with the crystal structure; the root-mean-square deviation (rmsd)
for the superimposition based on C␣ atom positions was found to
be 0.391 A˚ . The active site residues were compared and superim-
posed with the template structure. The conserved catalytic residues
(
Ser82, Asp228, and His250) in the lipase LC2-8 model have orien-
tations and locations similar to those of the residues in the template
1EX9). The oxyanion hole consists of two residues whose backbone
amide protons stabilize the tetrahedral intermediate formed dur-
ing the reaction mechanism. In lipase LC2-8, the oxyanion hole is
probably formed by the main chain nitrogens of Met16 and His83.
His83 is next to the catalytic Ser, whereas Met16 is located in a loop
after ␤-strand 3, next to a Gly residue.
2
.4. Construction of the molecular model and molecular docking
(
to rationalize the enantioselectivity
The calculated Ramachandran plot suggests that 91.97% of the
residues in the derived model are either in the most favored or
in the additional allowed regions. A model structure of protein
Fig. 3. Effect of mandelic acid concentration on the acylation of (R,S)-mandelic acid
catalyzed by lipase LC2-8. Reactions were carried out in diisopropyl ether (2 mL)
with lipase LC2-8 (20 mg) at 30 C for 15 h. The molar ratio of vinyl acetate to man-
Fig. 4. Time course of resolution of (R,S)-mandelic acid catalyzed by lipase LC2-
8. Reaction was carried out in 2 mL diisopropyl ether with (R,S)-mandelic acid
(180 mM), vinyl acetate (1440 mM), and lipase LC2-8 (40 mg) at 30 C. Symbols:
(᭹) conversion yield; (ꢀ) ees; (ꢀ) eep.
◦
◦
delic acid was kept constant (8:1). Symbols: (᭹) conversion yield; (ꢀ) ees; (ꢀ) eep;
(ꢁ) the amount of transformed mandelic acid.

1
12
Y. Cao et al. / Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
Table 3
On the basis of the modeled structure, molecular docking
Distances of hydroxyl groups of mandelic acid from the N␧ atom of His250
His250:N␧) of lipase LC2-8 and acetate carbonyl carbon (Ace:C) of vinyl acetate
orienting to Ser82 after docking.
of (R,S)-mandelic acid to lipase LC2-8 was undertaken. Some
reports indicate that protein conformational rearrangements may
be needed for better ligand affinity, but if the protein distortion
is too extensive, the orientation of some essential residues may
be disturbed, leading to loss of activity [20]. In our case, the con-
(
Complexes
dH N (Å)
dO C (Å)
S
R
2.56
4.57
2.929
5.362
formational changes of lipase LC2-8 were minimal (rmsd values
about 0.9 A˚ ) after docking. The docking results show that both enan-
tiomers could access the catalytic center of lipase LC2-8, while
the acetate oxygen atom (namely Ace:O) of vinyl acetate orienting
to Ser82 forms hydrogen bonds with the backbone NH of Met16
and His83 in the oxyanion hole (Fig. 5) in both complexes. These
hydrogen bonds may help to stabilize transition-state complex for
the second tetrahedral intermediate. The alcohol hydroxyl of (S)-
mandelic acid is located close to the active site, with distances of
tetrahedral intermediate between (R)-mandelic acid and acetyl-
lipase is difficult.
It is well known that applying molecular dynamics on com-
plexes obtained by docking would be efficient to refine the models,
accounting for the flexibility of both receptor and ligand [29]. To
complement our study of docking experiments, we carried out
510 ps of molecular dynamics simulations on the docking com-
plexes after optimization. The stability of the lipase LC2-8 structure
all along the trajectory was monitored with the RMSD of the back-
2
.929 A˚ from Ace:C and 2.56 A˚ from His250:N␧ (Fig. 5 and Table 3).
These results are in accordance with the required conditions [28]
for the complex to be productive (i.e., the respective distances to
catalytic His and Ser are <4 A˚ ), suggesting that acylation may occur
on the alcohol hydroxyl of (S)-mandelic acid. In the case of (R)-
mandelic acid, it distorts the orientation of the imidazole ring of the
catalytic histidine, thereby breaking the hydrogen bond between
the imidazole and the oxygen of the substrate (Fig. 5) and preven-
ting the protonation of His250. The alcohol hydroxyl is localized
at distances of 5.362 and 4.57 A˚ from the Ace:C and the His250:N␧,
respectively. Such distances relative to the histidine and the acetate
exceed the maximal distance criterion for acylation that allows pro-
ton transfer and nucleophilic attack. Therefore, forming the second
bone atoms referring to the initial structure of the protein. During
the MD production phase, the rmsd values were about 0.8–1.5 A˚ ,
indicating the protein did not undergo significant distortions and
the systems were stable throughout simulations. The essential
hydrogen bonds between the acetate carbonyl oxygen of the serine-
bound acetate (Ace:O) and the oxyanion hole residues, Met16: HN
and His83: HN, were present in most of trajectory frames, which
indicated the acetate was in a catalytically competent orienta-
tion. The trajectories were also analyzed basing on the mandelic
acid alcohol hydroxyl from the catalytic residue His250 and the
Fig. 5. Models of the binding between lipase LC2-8 and (R,S)-mandelic acid obtained from the LibDock approach. (A) The (S)-mandelic acid–lipase LC2-8 complex. (B) The
R)-mandelic acid–lipase LC2-8 complex. (C) The hydrogen bonding interactions of (S)-mandelic acid with catalytic residues and key residues of lipase LC2-8. (D) The hydrogen
bonding interactions of (R)-mandelic acid with catalytic residues and key residues of lipase LC2-8.
(

Y. Cao et al. / Journal of Molecular Catalysis B: Enzymatic 99 (2014) 108–113
113
enantioselectivity of lipase LC2-8 in the chiral resolution of man-
delic acid. The analysis of molecular dynamics simulations on the
docking complexes reinforces the conclusions about enantiose-
lectivity of lipase LC2-8 toward mandelic acid. These results may
contribute to the design and engineering of other lipases, especially
those with improved enantioselectivity, for chiral resolution.
Acknowledgements
Financial supports for this research from the National Program
on Key Basic Research Project (2011CB710800) and the National
High Technology Research and Development Key Program of China
(2012AA022200). We also acknowledge the support of the projects
funded by PCSIRT (IRT1066) and PADD.
References
[
[
1] J.K. Whitesell, D. Reynolds, J. Org. Chem. 48 (1983) 3548–3551.
2] S.P. Zingg, E.M. Arnett, A.T. McPhail, A.A. Bothner By, W.R. Gilkerson, J. Am.
Chem. Soc. 110 (1988) 1565–1580.
[
[
3] G.D. Yadav, P. Sivakumar, Biochem. Eng. J. 19 (2004) 101–107.
4] J.M. Palomo, G. Fernandez Lorente, C. Mateo, C. Ortiz, R. Fernandez Lafuente,
J.M. Guisan, Enzyme Microb. Technol. 31 (2002) 775–783.
[
[
[
5] H.N. Wei, B. Wu, Appl. Biochem. Biotechnol. 149 (2008) 79–88.
6] N. Queiroz, M.G. Nascimento, Tetrahedron Lett. 43 (2002) 5225–5227.
7] C.J. Yao, Y. Cao, S.S. Wu, S. Li, B.F. He, J. Mol. Catal. B: Enzym. 85–86 (2013)
105–110.
[
[
8] E. Santaniello, S. Casati, P. Ciuffreda, G. Meroni, A. Pedretti, G. Vistoli, Tetrahe-
dron: Asymmetry 20 (2009) 1833–1836.
9] J.B. Ma, L. Wu, F. Guo, J.L. Gu, X.L. Tang, L. Jiang, J. Liu, J.H. Zhou, H.W. Yu, Appl.
Microbiol. Biotechnol. 97 (2013) 4897–4906.
[
10] F. Guo, H.M. Xu, H.N. Xu, W.H. Yu, Appl. Microbiol. Biotechnol. 97 (2013)
355–3362.
3
[
[
[
[
11] Y. Cao, Y. Zhuang, C.J. Yao, B. Wu, B.F. He, Biochem. Eng. J. 64 (2012) 55–60.
12] U.K. Winkler, M. Stuckmann, J. Bacteriol. 138 (1979) 663–670.
13] Y.H. Zhang, Z.Y. Li, C.J. Yuan, Tetrahedron Lett. 43 (2002) 3247–3249.
14] Accelrys Software Inc., Discovery Studio Modeling Environment, Release 2.5,
Accelyrs Software Inc., San Diego, 2009.
15] S.F. Altschul, T.L. Madden, A.A. Schäffer, J. Zhang, Z. Zhang, W. Miller, D.J. Lipman,
Nucleic Acids Res. 25 (1997) 3389–3402.
16] R.A. Laskowski, M.W. MacArthur, D.S. Moss, J.M. Thornton, J. Appl. Crystallogr.
Fig. 6. The distances from mandelic acid alcohol to the Ace:C atom and the His250:
N␧ atom along the trajectory.
[
[
serine-bound acetate. Results were shown in Fig. 6. The alcohol
hydroxyl of (S)-mandelic acid maintained close to the catalytic
residues at distances inferior to 3.2 A˚ all along the trajectories.
For (R)-mandelic acid, the alcohol hydroxyl was too far from the
catalytic residues (about 4.5–6.5 A˚ ). These results from molecular
dynamics simulations on the docking complexes reinforce our con-
clusions about enantioselectivity of lipase LC2-8 toward mandelic
acid.
26 (1993) 283–291.
[17] J. Bowie, R. Luthy, D. Eisenberg, Science 253 (1991) 164–170.
[18] A.B.C. Simas, A.A.T.D. Silva, A.G. Cunha, R.S. Assump c¸ ão, L.V.B. Hoelz, B.C. Neves,
T.C. Galvão, R.V. Almeida, M.G. Albuquerque, D.M.G. Freire, R.B. de Alencastro,
J. Mol. Catal. B: Enzym. 70 (2011) 32–40.
[
[
[
19] B. Christelle, B.D.O. Eduardo, C. Latifa, M. Elaine-Rose, M. Bernard, R.-H. Evelyne,
G. Mohamed, E. Jean-Marc, H. Catherine, J. Biotechnol. 156 (2011) 203–210.
20] E.B. De Oliveira, C. Humeau, L. Chebil, E.R. Maia, F. Dehez, B. Maigret, M. Ghoul,
J.M. Engasser, J. Mol. Catal. B: Enzym. 59 (2009) 96–105.
21] S.N. Rao, M.S. Head, A. Kulkarni, J.M. LaLonde, J. Chem. Inf. Model. 47 (2007)
2159–2171.
[
[
22] A.M. Klibanov, Trends Biotechnol. 15 (1997) 97–101.
23] D.H. Yu, Z. Wang, L.F. Zhao, Y.M. Cheng, S.G. Cao, J. Mol. Catal. B: Enzym. 48
3
. Conclusion
(
2007) 64–69.
In this paper, lipase LC2-8 demonstrated excellent enantioselec-
[24] A. Wolff, L. Zhu, Y.W. Wong, A.J.J. Straathof, J.A. Jongejan, J.J. Heijnen, Biotechnol.
Bioeng. 62 (1999) 125–134.
tive acylation toward (S)-mandelic acid and afforded a conversion
yield of 49.15% and an eep value of 99%. The reaction concentration
of (R,S)-mandelic acid reached 180 mM (55 g/L), which is markedly
higher than previously reported. In addition, the mechanism of
catalysis of the resolution of mandelic acid with high enantio-
selectivity by lipase LC2-8 was analyzed. Molecular docking was
used to obtain the models of interaction between (R,S)-mandelic
acid and lipase LC2-8. The results reasonably explain the high
[
25] S.Y. Han, Z.Y. Pan, D.F. Huang, M. Ueda, X.N. Wang, Y. Lin, J. Mol. Catal. B: Enzym.
9 (2009) 168–172.
[26] D.Y. Kwon, Y.J. Hong, S.H. Yoon, J. Agric. Food Chem. 48 (2000) 524–530.
5
[
[
[
27] G. Yang, J.P. Wu, G. Xu, L.R. Yang, Appl. Microbiol. Biotechnol. 81 (2009)
47–853.
28] C. Palocci, M. Falconi, S. Alcaro, A. Tafi, R. Puglisi, F. Ortuso, M. Botta, L.
Alberghina, E. Cernia, J. Biotechnol. 128 (2007) 908–918.
8
29] H. Alonso, A.A. Bliznyuk, J.E. Gready, Med. Res. Rev. 26 (2006) 531–568.