Mutations in the stereospecificity pocket and at the entrance of the active site of Candida antarctica lipase B enhancing enzyme enantioselectivity

Article abstract of DOI:10.1016/j.molcatb.2010.01.007

Two different parts of Candida antarctica lipase B (stereospecificity pocket at the bottom of the active site and hydrophobic tunnel leading to the active site) were redesigned by single- or double-point mutations, in order to better control and improve enzyme enantioselectivity toward secondary alcohols. Single-point isosteric mutations of Ser47 and Thr42 situated in the stereospecificity pocket gave rise to variants with doubled enantioselectivity toward pentan-2-ol, in solid/gas reactor. Besides, the width and shape of the hydrophobic tunnel leading to the active site was modified by producing the following single-point mutants: Ile189Ala, Leu278Val and Ala282Leu. For each of these variants a significant modification of enantioselectivity was observed compared to wild-type enzyme, indicating that discrimination of the enantiomers by the enzyme could also arise from their different accessibilities from the enzyme surface to the catalytic site.

Full text of DOI:10.1016/j.molcatb.2010.01.007

Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Mutations in the stereospecificity pocket and at the entrance of the active site of
Candida antarctica lipase B enhancing enzyme enantioselectivity
Z. Martona, V. Léonard-Neversa, P.-O. Syrénb, C. Bauera, S. Lamarea, K. Hultb, V. Trancc, M. Grabera,∗
a UMR CNRS 6250 LIENSs, Université de La Rochelle, Bâtiment Marie Curie, Avenue Michel Crépeau, 17042 La Rochelle cedex 1, France
b Department of Biochemistry, School of Biotechnology, Royal Institute of Technology, AlbaNova University Center, SE-106 91 Stockholm, Sweden
c UMR CNRS 6204 U3B, Faculté des Sciences et Techniques, 2 rue de la Houssinière, 44322 Nantes cedex 3, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 18 January 2010
Two different parts of Candida antarctica lipase B (stereospecificity pocket at the bottom of the active site
and hydrophobic tunnel leading to the active site) were redesigned by single- or double-point mutations,
in order to better control and improve enzyme enantioselectivity toward secondary alcohols. Single-point
isosteric mutations of Ser47 and Thr42 situated in the stereospecificity pocket gave rise to variants with
doubled enantioselectivity toward pentan-2-ol, in solid/gas reactor. Besides, the width and shape of
the hydrophobic tunnel leading to the active site was modified by producing the following single-point
mutants: Ile189Ala, Leu278Val and Ala282Leu. For each of these variants a significant modification of
enantioselectivity was observed compared to wild-type enzyme, indicating that discrimination of the
enantiomers by the enzyme could also arise from their different accessibilities from the enzyme surface
to the catalytic site.
Keywords:
Lipase B from Candida antarctica
Stereoselective catalysis
Protein engineering
Stereospecificity pocket
Substrate accessibility to the active site
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
a medium substituent not larger than an ethyl group and a large
substituent bigger than an ethyl group.
It is now well-known that enantioselectivity of lipases can be
modified by structural factors of the substrates and of the enzyme.
In this context, directed mutagenesis has been used a lot to improve
lipases enantioselectivity and to increase the understanding of
how these enzymes solve the task of distinguishing between enan-
tiomers [1–4]. Especially, lipase B from Candida antarctica (CALB)
has been studied a lot by this way [5,6]. In this present work
our goal has been to better understand structural determinants of
CALB enantioselectivity toward secondary alcohols and to obtain
more-selective mutants. Crystal structure analysis of CALB reveals
a rather narrow and deep active site with a small stereospecificity
pocket flanked by residues Thr42, Ser47 and Trp104, which is at the
origin of its enantioselectivity [7,8]. The strong R-selectivity of CALB
toward secondary alcohols is explained by the different binding
modes of the enantiomers in the stereospecificity pocket [9,10]. The
fast-reacting enantiomer positions its medium-sized substituent in
the stereospecificity pocket and its large substituent toward the
active site entrance. The slow-reacting enantiomer has to get into
position in the active site in the opposite way, in order to form a
catalytically active binding mode, although the largest substituent
is not easily fitted in the stereospecificity pocket. This steric limi-
tation makes CALB very selective toward secondary alcohols with
Nevertheless, Rotticci et al. showed that steric interaction in the
stereospecificity pocket is not the only important factor for the dis-
crimination between enantiomers. These authors produced CALB
variants with nearly isosteric mutations in the stereospecificity
pocket (Thr42Val, Ser47Ala and Thr42Val/Ser47Ala) and obtained
an increased enantioselectivity towards 1-bromo-2-octanol and
1-chloro-2-octanol, when the single- or double-point mutation
Ser47Ala was used [6]. This effect was ascribed to a decrease
of electrostatic repulsion between the halogen atom of the sub-
strate and the stereospecificity pocket induced by the mutation.
In the present paper, the effect of the same isosteric mutations
in the stereospecificity pocket was further studied in a solid–gas
bioreactor, for the CALB-catalyzed resolution of linear secondary
alcohols (Scheme 1). Moreover, for these variants, variations in
activation enthalpy and entropy were measured, and the effect of
water on selectivity was studied, in the same type of reactor. This
reactor includes the enzyme at a solid state forming a fixed bed,
and gaseous substrates and nitrogen as carrier gas percolating the
enzymatic bed. The reactor presents the advantage to indepen-
dently control the thermodynamic activities of the components
present in the reaction medium by varying the partial pressure
of each compound in the carrier gas that percolates through the
catalytic bed [11]. Moreover this type of reactor permits the deter-
mination of the specific water addition in the medium, by adding
water in the gaseous stream at a controlled thermodynamic activity
[12,13].
∗
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).
1381-1177/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.01.007

12
Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
Scheme 1.
Finally, we focused on the narrow and short tunnel leading to
was left for 1 week under vacuum and over P2O5 at room temper-
ature.
CALB active site, and bounded by four hydrophobic residues, Ile189,
Leu278, Ala282 and Ile285. As substrates have to cross this tun-
nel before the formation of the second tetrahedral intermediate,
we tested the ability of these hydrophobic amino acids to play a
role in the discrimination between enantiomers. By using different
single-point mutations, the access of linear secondary alcohols to
the active site was made more difficult by increasing the volume
of Ala282 at the active site entrance (variant Ala282Leu), or easier
by decreasing the volume of Leu278 or Ile189 (variants Leu278Val
and Ile189Ala). The only study mentioning the discrimination of
the enantiomers through their channeling from the enzyme sur-
face to the catalytic triad, concerns Burkholderia cepacia lipase in
the case of the resolution of (R, S)-2-bromophenyl acetic acid ethyl
ester by transesterification in the presence of octanol [14]. Such
an investigation has never been performed in case of CALB lipase
and is presented in this paper, by studying the enantioselectivity
of CALB variants with single-point mutations at the entrance of the
active site crevice.
2.2.2. Mutants of the entrance crevice
The three mutants Ile189Ala, Leu278Val and Ala282Leu were
obtained through site-directed mutagenesis [18]. The pGAPZ␣ B
plasmid was amplified by polymerase chain reaction (PCR) with the
mutagenic primers designedtobenon-overlappingatthe ends [19].
The product of PCR was analysed on 1% agarose gel and was used
to transform E.coli XL-1 blue cells. Colonies containing the wished
mutation were identified by sequencing and used to produce the
plasmid in large quantity (i.e. >20 ␮g) using QIAPrep Spin miniprep
(QIAGEN GmbH, Hilden, Germany). Plasmids were linearized and
used to transform P. pastoris SMD 1168 H by electroporation. Trans-
formants of P. pastoris were cultured in order to produce CALB
mutants. The enzyme contained in the supernatant was purified
on butyl sepharose FF column. Purified protein fractions were sub-
jected to buffer exchange to 50 mM potassium phosphate buffer,
pH 7.5 and freeze dried.
0.53 mg enzyme was dissolved in 1 mL of MilliQ water, and dry
Chromosorb P AW DMCS (1 g) was added to the solution. After vig-
orous shaking, the preparation was left for 1 week under vacuum
and over P2O5 at room temperature.
2. Experimental
2.1. Chemicals
2.3. Experimental setup for solid–gas catalysis
Substrates and other chemicals were purchased from
Sigma–Aldrich-Fluka Chemical Co. They were of the highest
purity available (98% minimum) and checked by gas chro-
matography 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
The bioreactor used in this study has already been described
in a previous publication [20]. Thermodynamic experiments were
run from 308 to 363 K with immobilized enzyme preparation
(10–200 mg). All other experiments were made at 318 K. The total
flow rate passing through the reactor was set to 800 ␮mol/min for
thermodynamic study, 1000 ␮mol/min for entrance crevice muta-
tions effects study and 1150 ␮mol/min for water effects study.
Methyl propanoate and secondary alcohols thermodynamic activi-
ties in the solid/gas reactor were fixed at 0.1 and 0.05, respectively,
except for experiments with heptan-3-ol and octan-3-ol, for which
thermodynamic activities for alcohol was fixed at 0.1 and 0.2 for
methyl propanoate.
Erba.
1-Methylpropylpropanoate,
1-methylbutylpropanoate,
1-methylpentylpropanoate, 1-ethylpentylpropanoate and 1-
ethylhexylpropanoate were synthesized from corresponding
alcohol and propanoic acid in presence of p-toluene sulfonic
acid and molecular sieves [15]. 1-Methylhexylpropanoate and
1-methylheptylpropanoate were obtained from the corresponding
alcohol and propanoic anhydride in pyridine at room temperature
[16]. High purity water was obtained with the MilliQ system
(Millipore, France).
2.4. Chromatographic assays
2.2. Enzyme
For the solid/gas system analyses, the vapour phase leaving the
bioreactor was sampled by using a loop (0.25 mL) on a six-way
valve (Valco) maintained at 175 or 220 ◦C. Samples were automati-
cally injected into the split injector of a gas chromatograph (Agilent
6890N) (split 2:1) equipped with a flame ionization detector (FID)
for the detection of all products. The column was a Chirasil-DEX CB
(25 m, 0.25 mm i.d., 0.25 ␮m ␤-cyclodextrin; Chrompack, France).
Separation methods depended on the substrate used. Column tem-
perature was 55 ◦C for 20 min (1.5 mL/min, 180 ◦C, 180 ◦C) with
butan-2-ol, 55 ◦C for 15 min, 3 ◦C/min, 85 ◦C (1.5 mL/min, 180 ◦C,
180 ◦C) with pentan-2-ol, or 50 ◦C for 30 min (1.8 mL/min, 250 ◦C,
250 ◦C) when water was added, 60 ◦C for 15 min, 3 ◦C/min, 90 ◦C for
5 min (2 mL/min, 180 ◦C, 180 ◦C) with hexan-2-ol and hexan-3-ol,
60 ◦C for 15 min, 3 ◦C/min, 90 ◦C for 5 min (2 mL/min, 200 ◦C, 200 ◦C)
with heptan-2-ol and heptan-3-ol, 70 ◦C for 15 min, 3 ◦C/min,
100 ◦C for 5 min (2 mL/min, 200 ◦C, 200 ◦C) with octan-2ol, and 90 ◦C
2.2.1. Wild-type (WT) and stereospecific pocket mutants
WT CALB and stereospecificity pocket variants (Ser47Ala,
Thr42Val and Ser47Ala–Thr42Val) were produced in the methy-
lotropic yeast Pichia pastoris. These lipases were expressed
extracellularly and purified from the medium by hydrophobic
interaction chromatography, followed by gel filtration [6,17].
Enzyme adsorption was performed onto 60–80 mesh Chro-
mosorb
P AW DMCS (acid washed dimethylchlorosilanized)
(Varian, France). In a typical adsorption procedure for solid/gas
catalysis, WT enzyme (0.106 mg) or variant (0.85 mg) was dis-
solved in sodium phosphate 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 40690 M−1 cm−1. After vigorous shaking, the preparation

Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
13
Table 1
Thermodynamic components of the enantiomeric ratio, E, and activity A (␮mol of product min−1 mg−1 of free enzyme) for acylation of pentan-2-ol with methyl propanoate
as acyl donor, catalyzed by WT Candida antarctica lipase B and the three variants of the stereospecificity pocket. All experiments were performed at aw equal to 0 in the
gaseous stream of the solid/gas reactor. For measurements of both E- and A- values the thermodynamic activity of methyl propanoate and alcohol were fixed at 0.1 and 0.05,
respectively.
b
c
d
c
Type of enzyme
A
E at 318 Ka
ꢀR−SꢀG* at 318 K [kJ mol−1
]
ꢀR−SꢀH* [kJ mol−1
]
−TꢀR−SꢀS* [kJ mol−1
]
ꢀR−SꢀS* [J K−1 mol−1
]
Ne
WTf
Thr42Val
Ser47Ala
Thr42Val–Ser47Ala
30
101
236
198 19
98 13
5
6
−12.2 0.1
−14.4 0.2
−14 0.6
−18 1.2
−11 0.7
−20 1.0
−11 0.7
6
1.3
−3 0.5
0.9
−1 0.6
−19
11
4
2
3
2
18
16
11
13
7.5
10
2.3
6
−19
−12.1 0.4
3
a
Standard deviations were calculated with three values of E, and correspond to 95% confidence intervals.
Values calculated from −RT Ln E. Standard deviations were obtained from the standard deviations of E.
b
c
ꢀR−SꢀH* and ꢀR−SꢀS* were calculated by using the least square method, with between 11 and 18 measurement of E as a function of T. Standard deviations correspond
to 99% confidence intervals.
d
−TꢀR−SꢀS* values were calculated at 318 K. Standard deviations were obtained from standard deviations of ꢀR−SꢀS*.
e
Total number of E measurements.
f
Results of this line have already been published in a previous publication [12].
for 17 min (2 mL/min, 200 ◦C, 200 ◦C) for octan-3-ol. Nitrogen was
used as the carrier gas with a constant flow rate (between 1.5 and
2 mL/min according to experiments). Hydrogen and air were sup-
plied to the FID at 35 and 350 mL/min respectively. Quantitative
data were obtained by integration on an Agilent 3396 Series III
integrator or on GC Chemstation Rev B.03.02. The external calibra-
tions of the substrates were carried out by programming a range of
their partial pressures in the bioreactor, and by gas chromatography
analysis. For the products (methanol and chiral esters), an internal
calibration was carried out by using methyl propanoate and corre-
sponding alcohol as internal standards. For accurate determination
of E-values the vapour phase leaving the bioreactor was condensed.
Elution peaks of (R)-pentan-2-ol and (S)-pentan-2-ol were iden-
tified by using commercial pure (R)-pentan-2-ol. Retention time
of the reaction products was obtained by adding chemically syn-
thesized esters (R, S), as described in Section 2.1. The absolute
configuration of the reaction products was established by esteri-
fication of commercial pure alcohol enantiomers.
and without solvents. The Standard Dynamics Cascade Protocol of
DS 1.6 was used with a time step of 1 fs. The complex was energy
minimized (2500 Steepest descent and 2500 adobted Basis NR)
and temperature of the system was raised from 50 to 400 K over
2 ps. After a short equilibration (1 ps) a production run was carried
out for 10 ps at 400 K. Backbone atoms were kept fixed during
simulation, in order to keep the original structural integrity.
3. Results and discussion
3.1. The effect of two point mutations in the stereospecificity
pocket of CALB
Acyl-transfer reaction rates, in case of catalysis by native
or modified CALB, were experimentally measured with methyl
propanoate as acyl donor and pentan-2-ol as racemic nucleophile,
in a continuous solid gas reactor at 318 K, at water activity close to
zero. The acyl-transfer activity decreased by factors 3 and 4, when
using respectively Ser47Ala and Thr42Val mutants, compared to
WT CALB, for which a value of 30 ␮mol of product min−1 and mg−1
of pure and free enzyme was found. It decreased by a factor 13 when
using double mutant Ser47Ala–Thr42Val. These results can be com-
pared with the ones obtained by Rotticci et al. with tributyrin as
substrate in aqueous media. In this case activity decreased by fac-
tors 1.8, 6 and 3 compared to WT CALB when using respectively
Ser47Ala, Thr42Val and Ser47Ala–Thr42Val mutants [6].
The enantioselectivity of these different enzymes toward
pentan-2-ol was also measured in the solid gas reactor. The E-
value significantly increased when using Ser47Ala and Thr42Val
mutants, reaching respectively 198 and 236, instead of 100 with WT
enzyme [12]. On the other hand, double mutant Ser47Ala–Thr42Val
presented an enantioselectivity toward pentan-2-ol very close to
that of the WT enzyme. Furthermore, a thermodynamic analysis
of enantioselectivity was performed by considering both acti-
vation enthalpy and entropy (Table 1). It can be seen that the
preferred enantiomer (R)-pentan-2-ol, was favored by enthalpy
(ꢀR−SꢀH* < 0) but disfavored by entropy (ꢀR−SꢀS* < 0) in case
of WT enzyme and Ser47Ala variant. In case of Thr42Val and
Ser47Ala–Thr42Val variants, enantioselectivity is favored both by
enthalpy and entropy.
2.5. Determination of the enantiomeric ratio
The enantiomeric ratio, E, was calculated from ees and eep
[21,22]. The E-values were based on the average of three
measurements for each condition; in all cases the standard devi-
ation was less than 10%. The equations RT ln E = −ꢀR−SꢀG* and
ꢀ
R−SꢀG* = ꢀR−SꢀH* − TꢀR−SꢀS* were used to calculate enthalpic
and entropic components of E [23,24].
2.6. Shape of active site and entrance tunnel
The crystal structure of native CALB was obtained from the
Protein Data Bank [25] (PDB entry: 1TCA; resolution 1.55 Å [26]).
All molecular modeling visualizations, mutations and calculations
were performed with CHARMM force field [27] using the program-
package Discovery Studio (DS) version 1.6 or 2.1 (Accelrys Inc.).
Systematic rotamer search was performed on crystal structure
using Richardson rotamer library [28]. Conolly surface [29] of the
sum of all possible rotamers for a residue was calculated at a probe
radius of 1.4 Å.
Structure of second tetrahedral intermediate formed (R)-
heptan-3-ol:complex
shown in Fig. 4, was obtained from
obtained with (S)-hexan-3-ol:complex
ethylbutylpropanoate, by manual modifications followed by
Steepest Descent minimizations (RMSD ≤ 0.01). This last structure
was obtained from crystal structure of CALB by manual construc-
tion, after removal of crystallographic water molecules, followed
by MD simulation carried out by adopting a 12 Å non-bound spher-
ical cut-off, using the isothermal–isochoric ensemble (NVT) [30]
enzyme-(R)-1-ethylpentylpropanoate
structure previously
enzyme-(S)-1-
These different results show that single- or double-point muta-
tions in the stereospecificity pocket have very significant effects
both on activity and selectivity of CALB. Moreover, thermodynamic
analysis reveals that these effects are complex, as mutations lead to
modifications of both activation enthalpy and entropy differences
between R- and S-enantiomers. The importance of the entropy con-
tribution in defining enantioselectivity is once again shown [31],
but the exact molecular mechanism of how entropy and flexibility
are related to enzyme activity and selectivity is still elusive.
a

14
Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
Fig. 1. Stereospecificity pocket of CALB: (a) WT; and (b) double mutant Thr42ValSer47Ala.
In the present study, the effects of mutations cannot be ascribed
a pronounced effect on CALB activity and enantioselectivity. E-
value increased from 100 at water activity close to zero, in case
of the resolution of pentan-2-ol at 318 K, to a maximum of 320, at
water activity of 0.2. In this part, we were intending to compare
the effect of water on the enantioselectivity toward pentan-2-ol
at 318 K, of Thr42Val, Ser47Ala and Thr42Val–Ser47Ala variants
with the results previously obtained for the WT enzyme. Results
are shown in Fig. 2. First, enantioselectivity increases in all cases
up to aw = 0.1 for Ser47Ala, aw = 0.08 for Thr42Val, and aw = 0.3
for Thr42Val–Ser47Ala to reach the respective following maxima:
E = 449, 567 and 289. Above aw values corresponding to maximal E,
enantioselectivity decreases, after a plateau for Ser47Ala variant in
the range of aw from 0.1 to 0.3.
to modifications of the size of the stereospecificity pocket, as thre-
onine and valine on the one side are isosteric, and serine and
alanine on the other side, have similar molar volumes. Therefore
the mutations did not lead to major changes in the size of the stere-
ospecificity pocket (Fig. 1). Counterexample can be found in the
literature in which mutation effects is directly related to a change
of the stereospecificity pocket size: the Trp104Ala mutation of CALB
increases very significantly the size of the stereospecificity pocket,
and this redesigned pocket was found to accommodate much larger
groups than the WT enzyme. This change transformed the strongly
R-selective WT CALB into an S-selective mutant for the resolution
of 1-phenylethanol [32].
Besides, the change in hydrophobicity of the stereospecificity
pocket of CALB caused by the Thr42Val and the Ser47Ala muta-
tions cannot either be considered as a reason for variation of
enantioselectivity toward pentan-2-ol, as it was shown that sub-
strate selectivity of these mutants was very similar to that of the
WT enzyme, for the acylation of symmetrical secondary alcohols
(propan-2-ol, pentan-3-ol and heptan-4-ol) [18]. On the contrary,
in case of resolution of halohydrins through CALB-catalyzed acy-
lation, the change of hydrophobicity promoted by Thr42Val and
Ser47Ala mutations affected the CALB selectivity toward the halo-
drins, by decreasing electrostatic repulsion between substrates and
enzyme [6].
So it appears that the enhancement of E promoted by water is
significantly reduced in case of double-point variant compared to
WT enzyme, whereas it is strongly enhanced in case of single-point
variants.
Previously, the increase of enantioselectivity of WT CALB for
pentan-2-ol at low water activity, has been interpreted by specif-
ically binding of a water molecule in the stereospecificity pocket,
leading to a reduction of its size and thereby rendering binding
of S-enantiomer still more difficult compared to the R-form [12].
The increase of enantioselectivity toward pentan-2-ol afforded
by single-point mutations (Thr42Val and Ser47Ala) may originate
from modification of the H-bond network between amino acid
residues in the vicinity of the stereospecificity pocket, and differ-
ent spatial degrees of freedom and mobility of these residues and
also of substrate substituent positioned in this pocket. The higher
E-value for the single-point mutants compared to that of the WT
could arise from higher mobility of the non-mutated residue, which
now is not hydrogen bonded and has more mobility. If it moves
more than in the WT, the specificity pocket will be smaller, then
the S-enantiomer will not fit so well and E goes up. The H-bond
network progress and mobility of the residues located in the stere-
ospecificity pocket, when using variants Thr42Val and Ser47Ala,
compared to WT CALB, will be the topic of future studies by molec-
ular dynamics simulation.
3.2. The effect of water on enantioselectivity of the
stereospecificity pocket variants
Fig. 2. Influence of thermodynamic water activity, aw on enantioselectivity at 318 K
for the acylation of pentan-2-ol with methyl propanoate as acyl donor, catalyzed by
WT CALB and three variants Thr42Val, Ser47Ala and Thr42ValSer47Ala, in continu-
ous solid/gas reactor.
The effect of water thermodynamic activity (aw) on WT CALB-
catalyzed acyl-transfer reaction was previously extensively studied
by using a solid/gas reactor [12,13]. It was shown that water had

Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
15
Fig. 3. Shape of the entrance tunnel. (a) WT with occupancy of hydrophobic residues: Ile189, Leu278, Ala282 and Ile285 (surface of all rotamers); (b) mutant Ala282Leu; (c)
mutant Leu278Val; and (d) mutant Ile189Ala; occupancy of mutants in black grid.
The result obtained with double-point variant Thr42Val–Ser47Ala
showing a significant decrease of water effect, is in accordance
with this hypothesis, as the redesigned stereospecificity pocket
is not water-attractive anymore, because both mutations lead to
the replacement of a polar hydroxyl group by an apolar sub-
stituent (methyl or hydrogen atom). This fact can be confirmed
by calculating the water dissociation constant with specificity
pocket (Kd), thanks to a Hill equation analysis for aw < 0.1. Kd for
Thr42Val–Ser47Ala variant was found to be equal to 0.1, which
is much higher than the value of 0.03 previously obtained with
WT enzyme. The Hill coefficient for double mutant (h = 0.55) was
consistent with one water molecule bound in the stereospecificity
pocket as for the WT enzyme [12]. Finally, the fact that maximal E-
value was obtained at higher water activity for Thr42Val–Ser47Ala
variant compared to WT is also consistent with a lower affinity for
water.
On the contrary, in case of single-point variants, the Hill equa-
tion analysis gave much lower Kd values than that obtained for
the WT enzyme: Kd = 0.004 and 0.0002 for Ser47Ala and Thr42Val
respectively, with Hill coefficients being respectively equal to 1.45
and 2.24. These results are in agreement with a high affinity
for water for the stereospecificity pocket of single-point vari-
ants, leading to a strong enhancement of E-value when going
from water activity close to zero to 0.1 for Ser47Ala or 0.08 for
Thr42Val.
3.3. The effect of mutations at the entrance of the active site of
CALB
The crystallographic structure of CALB shows that the substrate
binding pocket is an elliptical, steep funnel. The pocket constitu-
tion is similar to other lipases with a hydrophilic bottom of the
funnel, which is formed by the catalytic Ser105 and residues of
the oxyanion hole, Thr40 and Gln106 [33]. Near the surface the
pocket is bordered by hydrophobic residues, which interact with
organic solvents, but unlike to lipases, CALB have no lid and is
not submitted to interfacial activation. The active site crevice can
be partitioned into two sides, an acyl side and an alcohol side,
where the corresponding parts of the substrate will be located dur-
ing catalysis [7]. The alcohol side is bounded by four hydrophobic
residues, Ile189, Leu278, Ala282 and Ile285. They form a narrow
tunnel, which must be crossed by substrate before the forma-
tion of the second tetrahedral intermediate. The volumes occupied
by these residues were calculated by systematic conformational
search. This study showed that these residues have side chains
free to rotate and have no steric interaction with neighboring
residues (Fig. 3a). Consequently a concerted motion of the side
chains cannot be observed, whereas such was the case with B.
cepacia lipase, when the substrate passes though the bottleneck
formed by Val266 and Leu17 [14]. However Fig. 3a suggests that
these four hydrophobic residues, Ile189, Leu278, Ala282 and Ile285
can affect substrate trajectory. Mutations were considered in order
to understand the role of these residues in the selectivity of CALB.
To avoid electrostatic effects, only mutations in another aliphatic
hydrophobic residue were considered. The volume of residue 282
was expended by changing native Ala into Leu, and volumes of
These very high E-values are promoted by water at very low
activity, where water acts mainly through specifically located
effects. At higher water activity, entropic effects coming from water
molecules have been shown to become predominant and can lead
to a decrease of E [13].

16
Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
Table 2
Enantiomeric ratio Ea and activity A (␮mol of product min−1 and mg−1 of free enzyme) for acylation of chiral secondary alcohols with methyl propanoate as acyl donor at
318 K, in solid/gas reactor, catalyzed by WT Candida antarctica lipase B and variants of CALB obtained by mutations of amino acids at the entrance of the active site crevice.
All experiments were performed at aw equal to 0 in the gaseous stream of the solid/gas reactor. Thermodynamic activity of methyl propanoate and alcohol were fixed at 0.1
and 0.05, respectively, except for heptan-3-ol and octan-3-ol (0.2 and 0.1), for measurements of both E-and A-values.
Structure
WT
Ala282Leu
Leu278Val
Ile189Ala
E
A
E
A
E
A
E
A
4.6 0.1
n.d.
4.9 0.1
20.7
5.8 .0.2
20.5
6.9 0.4
12.3
100
297
6
9
17.7
13.5
11.7
6.7
130
228
337
1
8
2
22.6
26.1
20.9
11.1
9.8
98
8
7
29.1
18.2
11.1
7.2
87
1
6
12.6
10.8
4.8
216
211
476 30
435 18
299 18
437 25
400 10
298 11
227
83
3
3
4.3
78
5
8.7
103
8
6.4
45
3
7
2.8
318 15
523 31
2.4
208 14
245 15
2.2
203 15
210 15
1.8
192
1.7
3.1
4.1
3.3
199 14
2.0
a
Standard deviations were calculated with 5–12 values of E from two condensates, and correspond to 95% confidence intervals.
positions 278 and 189 were decreased by respectively changing
Leu into Val and Ile into Ala (Fig. 3b–d). Ile189Ala, Leu278Val and
Ala282Leu mutants were produced and their activity and enan-
tioselectivity toward different chiral alcohols tested in solid/gas
bioreactor.
jectory of the S-enantiomer, as this substrate has to enter the
active site with its large substituent in the front, in order to react.
Indeed, according to the model proposed by Haeffner et al., the
S-enantiomer has to put its large substituent into the stereospeci-
ficity pocket, located at the bottom of the active site, in order to
form catalytically active second tetrahedral intermediate [9]. This
last covalent intermediate is considered to be a good model of the
transition state intermediate in case of CALB-catalyzed resolution
of chiral nucleophile by acyl-transfer [9].
As shown in the following discussion, this postulate provides
a rational explanation for most results. In some cases however,
other structural explanations have to be given, like a modification of
the R-enantiomer trajectory also, an effect of the mutation on the
structure and energy state of second tetrahedral intermediate or
an interaction between acyl part and residue 189, which are very
close together. In any case, a complete structural explanation of
the effect of mutations on E very likely involves a combination of
these different individual effects. Molecular dynamics simulations
are currently in progress to study substrate trajectory in WT and
variants of CALB.
According to the above postulate, Leu278Val mutant is supposed
to make access of S-enantiomer to the active site easier, because of
the smaller size of Val compared to Leu. This hypothesis is veri-
fied (lower E-value than WT enzyme) for hexan-2-ol, heptan-2-ol,
octan-2-ol, heptan-3-ol and octan-3-ol. So, if the large substituent
of linear secondary alcohol includes four or more C, then the pos-
tulate is verified. If the large substituent includes three C or less,
the observed effect is not in agreement with the postulate: higher
E-value than WT enzyme (butan-2-ol and hexan-3-ol) or similar
E-value (pentan-2-ol). In these cases, an easier access to the active
site for the R-enantiomer could explain the effect.
Acyl-transfer reaction rates were experimentally measured
with methyl propanoate as acyl donor and eight different linear
secondary alcohols ranging from four to eight carbons as racemic
nucleophile, at 318 K and water activity close to zero. Results
showed that activity increased by a factor 1.1–2, when using
Ala282Leu variant compared to WT CALB, except for heptan-3-ol,
for which it decreased; activity was similar or slightly increased
when using Leu278Val variant, except for hexan-3-ol and heptan-
3-ol, for which it decreased, and activity decreased with Ile189Ala
mutant. As an example, in case of hexan-2-ol as nucleophile, a value
of 13.5 ␮mol of product min−1 mg−1 of pure and free enzyme was
found with WT CALB, which was doubled to 26.1 with Ala282Leu,
increased to 18.2 with Leu278Val and decreased to 10.8 with
Ile189Ala.
Table 2 shows the results for the enantiomeric ratio E for
acylation of the eight chiral alcohols (butan-2-ol, pentan-2-ol,
hexan-2-ol, heptan-2-ol, octan-2-ol, hexan-3-ol, heptan-3-ol and
octan-3-ol), catalyzed by WT CALB or the different variants of CALB
with mutations at the entrance of the active site crevice. It appears
that all mutations have a significant effect on E, promoting an
increase of E up to a factor 1.5 or a decrease up to a factor 2.2.
Nevertheless the upward or downward trend for E is variable for
each mutation, depending on the length of the large substituent
of the chiral alcohol. So another time, the analysis of the results
for E variations is complex and needs to consider many different
structural causes.
To analyze these data, the following postulate can be formu-
lated: modifications of the entrance crevice to the active site
by mutations of amino acids, will predominantly affect the tra-
Ile189Ala mutant is also supposed to ease access of S-
enantiomer to the active site. This hypothesis is verified (lower
E-value than WT enzyme) for all alcohols except butan-2-ol. So if

Z. Marton et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 11–17
17
enzyme activity and selectivity, and remain to be accurately eluci-
dated.
Acknowledgements
Financial support by the French National Agency for Research
ANR (Chimie et procédés pour le développement durable) and the
Swedish Research Council is gratefully acknowledged. We also like
to thank Philippe Pineau for fruitful help in chemical synthesis of
esters.
References
[1] M. Holmquist, M. Martinelle, P. Berglund, I.G. Clausen, S. Patkar, A. Svendsen,
K. Hult, J. Protein Chem. 12 (1993) 749.
[2] H. Scheib, J. Pleiss, P. Stadler, A. Kovac, A.P. Potthoff, L. Haalk, F. Spener, F. Paltauf,
R.D. Schmid, Protein Eng. 11 (1998) 675–682.
[3] F. Secundo, G. Carrea, C. Tarabiono, P. Gatti-Lafranconi, S. Brocca, M. Lotti, K.-E.
Jaeger, M. Puls, T. Eggert, J. Mol. Catal. B: Enzym. 39 (2006) 166–170.
[4] F. Bordes, E. Cambon, V. Dossat-Létisse, I. André, C. Croux, J.-M. Nicaud, A. Marty,
ChemBioChem 10 (2009) 1705–1713.
[5] A. Magnusson, K. Hult, M. Holmquist, J. Am. Chem. Soc. 123 (2001) 4354–4355.
[6] D. Rotticci, J.C. Rotticci-Mulder, S. Denman, T. Norin, K. Hult, ChemBioChem 2
(2001) 766–770.
[7] J. Uppenberg, N. Ohrner, M. Norin, K. Hult, G.J. Kleywegt, S. Patkart, V. Waagen,
T. Anthonsen, T.A. Jones, Biochemistry 34 (1995) 16838–16851.
[8] C. Orrenius, F. Haeffner, D. Rotticci, N. Ohrner, T. Norin, K. Hult, Biocatal. Bio-
transfor. 16 (1998) 1–15.
[9] F. Haeffner, T. Norin, K. Hult, Biophys. J. 74 (1998) 1251–1262.
[10] D. Rotticci, F. Haeffner, C. Orrenius, T. Norin, K. Hult, J. Mol. Catal. B: Enzym. 5
(1998) 267–272.
[11] S. Lamare, M.-D. Legoy, M. Graber, Green Chem. 6 (2004) 445–458.
[12] V. Léonard, L. Fransson, S. Lamare, K. Hult, M. Graber, ChemBioChem 8 (2007)
1–7.
Fig. 4. Structure of WT and Ala282Leu CALB with (R)-1-ethylpentylpropanoate
covalently attached to Ser105 (second tetrahedral intermediate formed with (R)-
heptan-3-ol), obtained by MD simulation described in Section 2.6; occupancy of
mutants in black grid, (R)-1-ethylpentylpropanoate in dark grey.
the large substituent of linear secondary alcohol includes three or
more C, then the postulate is verified.
For the Ala282Leu variant, our postulate implies a more diffi-
cult access of the S-enantiomer to the active site, because of the
steric hindrance of the crevice caused by Leu282 replacing Ala. The
postulate is verified (higher E-value than WT enzyme) for butan-
2-ol, pentan-2-ol and hexan-3-ol. So, if the large substituent of
linear secondary alcohol includes three C or less, then the postu-
late is verified. If the large substituent includes more than three C
(hexan-2-ol, heptan-2-ol, octan-2-ol, heptan-3-ol, octan-3-ol), the
observed effect is opposite to the one predicted by the postulate:
lower E-value than WT enzyme. In these cases, a more difficult
access to the active site for the R-enantiomer could explain the
effect or also an interaction between the large substituent of the R-
enantiomer and the residue Leu282, during formation of the second
tetrahedral intermediate. Fig. 4 shows an example of such an inter-
action in case of the second tetrahedral intermediate formed with
(R)-heptan-3-ol:(R)-1-ethylpentylpropanoate covalently attached
to Ser105, with WT or Ala282Leu CALB.
[13] V. Léonard-Nevers, Z. Marton, S. Lamare, K. Hult, M. Graber, J. Mol. Catal. B:
Enzym. 59 (2009) 90–95.
[14] D. Guieysse, C. Salagnad, P. Monsan, M. Remaud-Simeon, V. Tran, Tetrahedron
Asymm. 14 (2003) 1807–1817.
[15] C.E. Rehberg, Org. Syn. Coll. 3 (1955) 146.
[16] M.B. Smith, J. March, Advanced Organic Chemistry: Reactions, Mechanisms and
Structure, 5th ed., Wiley, New York, 2001.
[17] J.C. Rotticci-Mulder, M. Gustavsson, M. Holmquist, K. Hult, M. Martinelle, Pro-
tein Expression Purif. 21 (2001) 386–392.
[18] A.O. Magnusson, J.C. Rotticci-Mulder, A. Santagostino, K. Hult, Chembiochem 6
(2005) 1051–1056.
[19] L. Zheng, U. Baumann, J.L. Reymond, Nucleic Acids Res. 32 (2004) e115.
[20] S. Lamare, M.D. Legoy, Biotechnol. Bioeng. 45 (1995) 387–397.
[21] C.-S. Chen, Y. Fujimoto, G. Girdaukas, C.J. Sih, J. Am. Chem. Soc. 104 (1982)
7294–7299.
[22] J.L.L. Rakels, A.J.J. Straathof, J.J. Heijnen, Enzyme Microb. Technol. 15 (1993)
1051–1056.
[23] R.S. Phillips, Enzyme Microb. Technol. 14 (1992) 417–419.
[24] R.S. Phillips, Trends Biotechnol. 14 (1996) 13–16.
[25] H.M. Berman, J. Westbrook, Z. Feng, G. Gilliland, T.N. Bhat, H. Weissig, I.N.
Shindyalov, P.E. Bourne, Nucleic Acids Res. 28 (2000) 235–242.
[26] J. Uppenberg, M.T. Hansen, S. Patkart, T.A. Jones, Structure 2 (1994) 293–308.
[27] A.D. MacKerell Jr., D. Bashford, M. Bellott, R.L. Dunbrack Jr., J.D. Evanseck, M.J.
Field, S. Fischer, J. Gao, H. Guo, S. Ha, D. Joseph-McCarthy, L. Kuchnir, K. Kucz-
era, F.T.K. Lau, C. Mattos, S. Michnick, T. Ngo, D.T. Nguyen, B. Prodhom, W.E.
Reiher III, B. Roux, M. Schlenkrich, J.C. Smith, R. Stote, J. Straub, M. Watan-
abe, J. Wiorkiewicz-Kuczera, D. Yin, M. Karplus, J. Phys. Chem. B 102 (1998)
3586–3616.
[28] S.C. Lovell, J.M. Word, J.S. Richardson, D.C. Richardson, Proteins 40 (2000)
389–408.
[29] M.L. Connolly, Science 221 (1983) 709–713.
[30] H.J.C. Berendsen, J.P.M. Postma, W.F. Van Gunsteren, A. Dinola, J.R. Haak, J.
Chem. Phys. 81 (1984) 3684–3690.
[31] J. Ottosson, L. Fransson, K. Hult, Protein Sci. 11 (2002) 1462–1471.
[32] A.O. Magnusson, M. Takwa, A. Hamberg, K. Hult, Angew. Chem. Int. Ed. 44 (2005)
4582–4585.
4. Conclusion
In this paper, thanks to single-point mutations, new amino acid
positions important for CALB enantioselectivity were identified at
the entrance of the active site and offered altered enantioselectivity.
Moreover, the crucial role of residues situated in the stereospeci-
ficity pocket was confirmed and resolution of pentan-2-ol could be
improved, by using variants of this pocket. The present investiga-
tion is a new demonstration of the possibilities offered by protein
redesign, for improving kinetic resolution using CALB. It also high-
lights that entropic phenomena, including enzyme and substrate
flexibility, are crucial for understanding molecular mechanism of
[33] J. Pleiss, M. Fischer, R.F. Schmid, Chem. Phys. Lipids 93 (1998) 67–80.