An S-selective lipase was created by rational redesign and the enantioselectivity increased with temperature

Article abstract of DOI:10.1002/anie.200500971

(Chemical Equation Presented) Higher activity with larger pockets: The figure shows a superposition of intermediates that occur in acyl transfer to (S)-1-phenylethanol catalyzed by Candida antarctica lipase B (CALB). Wild-type CALB cannot accomodate the p

Full text of DOI:10.1002/anie.200500971

Communications
Protein Engineering
substituent toward the active-site entrance. For the slow-
reacting enantiomer, the medium-sized substituent points
toward the active-site entrance and the large substituent is
placed in the stereospecificity pocket. Although the large
substituent is not easily fitted in the stereospecificity pocket,
this is the catalytically active binding mode for the slow-
reacting enantiomer. The largest substituent that is well
accommodated in the stereospecificity pocket is an ethyl
group. This steric limitation makes CALB very selective
toward secondary alcohols with a medium-sized substituent
not larger than an ethyl group and a large substituent bigger
than an ethyl group. When both substituents are larger than
an ethyl group, the selectivity is low and the catalytic activity
An S-Selective Lipase Was Created by Rational
Redesign and the Enantioselectivity Increased
with Temperature**
Anders O. Magnusson, Mohamad Takwa,
Anders Hamberg, and Karl Hult*
Lipases have characteristics that make them attractive tools
for the synthesis of fine chemicals. They catalyze acyl-transfer
reactions and are stable in organic solvents and at high
temperatures. The possibility of modifying the performance
of enzymes by biomolecular methods gives the opportunity to
create tailor-made catalysts for each desired application. The
high enantioselectivity of lipases toward secondary alcohols is
[
11]
drops drastically. To overcome this limitation in size and to
increase the number of secondary alcohols that are good
substrates for CALB, we earlier redesigned the volume of the
stereospecificity pocket. Trp104, which forms the bottom of
the stereospecificity pocket, was replaced by smaller amino
acids. The redesigned stereospecificity pockets were shown to
readily accommodate much larger substituents than the
stereospecificity pocket of the wild-type lipase. The mutant
with the largest stereospecificity pocket, Trp104Ala, had a
[
1]
a property often used in organic synthesis. The enantiose-
lectivity follows the Kazlauskas rule which normally predicts
[
2]
an R selectivity. Structural explanations now exist for the
[
3]
strong chiral preference of many lipases. Efforts to modify
their properties have resulted in large improvements of the
catalyst, such as increased stability and activity, but few
ꢀ
1
ꢀ1
specificity constant (k /K ) of 830 s m toward 5-nonanol,
cat
M
[
4,5]
[12]
articles report large effects on the enantioselectivity.
The
5500 times greater than that of the wild-type enzyme.
largest changes in enantioselectivity of lipases concern
substrates with a chiral center on the acyl chain of the
A larger stereospecificity pocket should affect the enan-
tioselectivity of the lipase. Chiral secondary alcohols might
easily position their large substituent in the redesigned
stereospecificity pocket of the Trp104Ala mutant, which
would lead to a low enantioselectivity. Computer modeling
was used to compare the binding of secondary alcohols in the
active site of the Trp104Ala mutant and wild-type CALB.
Energy minimizations were performed on the tetrahedral
intermediate of both lipase variants with the butanoate ester
of (R)- and (S)-1-phenylethanol covalently bound to the
catalytic serine (Figure 1). The R enantiomer was similarly
positioned in both wild-type CALB and the Trp104Ala
mutant; the large substituent (phenyl group) was oriented
toward the active-site entrance and the medium-sized sub-
stituent (methyl group) was positioned in the stereospecificity
pocket. The S enantiomer had different orientations in the
two enzyme variants. In the wild-type lipase the slow-reacting
enantiomer could not be accommodated in a catalytically
active tetrahedral intermediate, as the phenyl group was too
large to fit in the stereospecificity pocket. This finding is in
accordance with a low reaction rate for the S enantiomer and
a high enantioselectivity favoring the R enantiomer. How-
ever, the redesigned stereospecificity pocket of the
Trp104Ala mutant comfortably accommodated the phenyl
group.
[
6]
substrate. We redesigned the active site of Candida antarc-
tica lipase B (CALB) to modify its specificity toward secon-
dary alcohols. The enantioselectivity was greatly changed and
an S-selective enzyme was created, mainly as a result of the
increased activity toward the slow-reacting enantiomer. The
mutant had very unusual behavior: the enantioselectivity
increased strongly with temperature, and thermodynamic
analysis showed that the altered enantioselectivity was
dominated by entropy. In dynamic kinetic resolution, lipases
[
7]
can be used to produce close to 100% of the R enantiomer.
The creation of an S-selective lipase affords the possibility of
achieving high yields of the S enantiomer.
CALB is a very robust and efficient catalyst that shows a
[
8]
high regio- and enantioselectivity. The active site contains
the catalytic triad Ser-His-Asp, an oxyanion hole that
stabilizes the transition states, and a cavity called the stereo-
[
9]
specificity pocket. The strong R selectivity of CALB toward
secondary alcohols is explained by the different binding
[
10,11]
modes of the enantiomers in the stereospecificity pocket.
The fast-reacting enantiomer positions its medium-sized
substituent in the stereospecificity pocket and its large
[*] A. O. Magnusson, M. Takwa, A. Hamberg, Prof. K. Hult
Royal Institute of Technology (KTH)
Department of Biochemistry, School of Biotechnology
AlbaNova University Center
We tested the low enantioselectivity expected for the
Trp104Ala mutant by performing acyl-transfer reactions
from vinyl butanoate to four secondary alcohols in cyclohex-
ane. With 3-methyl-2-butanol, the selectivity for the R enan-
tiomer dropped by a factor of 160 as a result of the mutation
10691 Stockholm (Sweden)
Fax: (+46)8-5537-8468
E-mail: kalle@biotech.kth.se
(Table 1, entries 6 and 7). With 2-hexanol, the R selectivity of
[
**] This work was supported by the Swedish Research Council and the
Erasmus program. The graphical material presenting the active
sites of the lipases was prepared by Linda Fransson.
several hundred for the wild-type enzyme changed into a low
S selectivity for the mutant (Table 1, entry 5). Thus methyl,
isopropyl, and butyl groups are almost equally well-accom-
modated in the stereospecificity pocket of the Trp104Ala
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
4
582
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200500971
Angew. Chem. Int. Ed. 2005, 44, 4582 –4585

Angewandte
Chemie
mutant. Interestingly, while wild-type CALB has a very high
R selectivity, the mutant has a moderate S selectivity toward
the two secondary alcohols with the bulky phenyl group as the
large substituent (Table 1, entries 3 and 4). The phenyl group
can probably fit well in the space liberated by the Trp104Ala
mutation.
The enantioselectivity of wild-type CALB is very sensitive
[
13]
to the solvent used in the incubation. The effects of the
solvent and temperature on the enantioselectivity of the
Trp104Ala mutant were investigated. We resolved 1-phenyl-
ethanol in three solvents and at temperatures between 10 and
7
08C (Figure 2). Both the solvent and the temperature had
Figure 1. The active site of wild-type CALB (left) and Trp104Ala
mutant (right) with the butanoate ester of (R)-1-phenylethanol (top)
and (S)-1-phenylethanol (bottom) covalently bound to the catalytic
serine in the tetrahedral reaction intermediate. The substrate is pre-
sented with a stick model and amino acid 104 with a space-filling
model in white. The R enantiomer has a similar configuration in the
wild-type CALB and Trp104Ala mutant: the large substituent (phenyl)
points toward the active-site entrance and the medium-sized substitu-
ent (methyl) is positioned in the stereospecificity pocket. In the wild-
type CALB, the S enantiomer cannot position its phenyl group in the
stereospecificity pocket, and not all the hydrogen bonds required for
catalysis can be formed. In the Trp104Ala mutant, the phenyl group is
comfortably accommodated in the space liberated by the mutation in
the stereospecificity pocket.
Figure 2. The S selectivity increased with temperature in the resolution
of 1-phenylethanol catalyzed by the Trp104Ala mutant of CALB in ace-
tonitrile (*), cyclohexane (~), and cis-decalin (
). Trend lines are plot-
&
ted according to Equation (1) using the values of the entropy and
enthalpy terms in Table 1.
large effects on the enantioselectivity. The S selectivity
increased to a large extent with the molecular size of the
solvent. An increase in the temperature also favored the
S enantiomer, which resulted in the unusual behavior of an
increasing enantioselectivity with temperature. These effects
gave an S selectivity of 44 for the Trp104Ala mutant toward
1
-phenylethanol at 698C in cis-decalin. Wild-type CALB is
Table 1: Thermodynamic components for acyl-transfer reactions from vinyl butanoate to secondary alcohols in various solvents catalyzed by wild-type
°
°
CALB and the Trp104Ala mutant. Differential entropy (D DS ) and enthalpy (D DH ) and their standard errors were determined from the linear
SꢀR
SꢀR
ꢀ1
°
°
regression of R”lnE versus T . The entropic co ntribution ( TD DS ), Gibbs free energy (D DG ), and the enantioselectivity (E) were calculated
SꢀR
SꢀR
for 303 K. The preferred enantiomer (Pe) and the racemic temperature (T ) are also presented.
R
[
a]
°[a]
°[a]
°
°
Entry
CALB
variant
Sec. alcohol
Solv.
Pe
E
D
DG
TD DS
[kJmol
D
DH
D DS
SꢀR
TR
[K]
SꢀR
SꢀR
SꢀR
ꢀ
1
ꢀ1
ꢀ1
ꢀ1
ꢀ1
[kJmol
]
]
[kJmol
]
[JK mol
]
1
2
3
Trp104Ala
Trp104Ala
Trp104Ala
CH CN
S
S
S
3.8
13
7.9
ꢀ3.3
ꢀ6.4
ꢀ5.2
21
34
35
18ꢁ2
70ꢁ8
112ꢁ1
116ꢁ5
250
250
260
3
[
b]
Deca
27.5ꢁ0.3
30ꢁ2
[
c]
cHex
cHex
[
c]
c]
4
5
Trp104Ala
Trp104Ala
S
S
12
ꢀ6.3
ꢀ1.9
26
23
20ꢁ1
21ꢁ2
87ꢁ4
77ꢁ5
230
280
[
cHex
2.2
[
[
c]
c]
6
7
Trp104Ala
wild-type
cHex
cHex
R
R
2.8
470
2.6
15
26
7.8
28ꢁ2
23ꢁ2
84ꢁ5
26ꢁ7
330
900
[a] Calculated for 303 K from the linear relation of Equation (1). [b] cis-Decalin. [c] Cyclohexane.
Angew. Chem. Int. Ed. 2005, 44, 4582 –4585
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4583

Communications
used in dynamic kinetic resolution to achieve the R enan-
components agree well with the values obtained for the
resolution of other secondary alcohols catalyzed by the same
[
14]
tiomer, and by using the Trp104Ala mutant in the same
processes a high yield and enantiomeric excess could be
achieved for the S enantiomer. The CALB mutant also has a
higher activity (Table 2) and stability compared to those of
subtilisin, which has been used in dynamic kinetic resolution
[
17]
enzyme. For the resolutions catalyzed by the Trp104Ala
mutant, the enthalpic contribution was not significantly
different from that of the wild-type enzyme. However, the
entropic contribution increased so much that the enantiose-
lectivity was eliminated or reversed. An enantioselectivity
dominated by the entropy afforded a very
[
15]
to achieve the S enantiomer.
unusual temperature dependence that has
rarely been seen before. The enantiose-
lectivity of CALB Trp104Ala increased
strongly with temperature: in the resolution
of 1-phenylethanol in cis-decalin an E-value
of 44 in favor of the S enantiomer was
Table 2: Apparent kinetic constants, specificity constants, preferred enantiomer (Pe), and E value of the
wild-type CALB and Trp104Ala mutant for the acylation of the pure enantiomers of 1-phenylethanol in
cyclohexane at 308C with 500 mm vinyl butanoate.
[18]
app
cat
app
M
[a]
CALB
Substrate
k
K
k
cat/KM
Pe
E
ꢀ1
ꢀ1
ꢀ1 [a]
variant
enantiomer
[s
]
[mM]
[s
M ]
Wild-type
Trp104Ala
R
S
570
0.00053
61
71
9300
0.0075
R
1300000
reached at 698C (Figure 2). The racemic
[19]
temperature (T = D DH/D DS),
at
R
SꢀR
SꢀR
which the enantioselectivity is 1, was 330 K
for this mutant toward 3-methyl-2-butanol;
below this temperature the mutant was R-
selective and above it, S-selective. A race-
mic temperature outside the experimental
range should only be used to compare
R
S
4.4
34
29
34
150
1000
S
6.6
app
app
[
a] Calculated from kcat and KM
.
The Trp104Ala variant showed the very unusual behavior
of increasing enantioselectivity with temperature for all the
substrates tested, except for 3-methyl-2-butanol where the
selectivity changed from the R to the S enantiomer. The
selectivity between the two enantiomers depends on the
difference in activation free energy (D DG ), which can be
divided into its enthalpic (D DH ) and entropic compo-
nents (D DS ) according to Equation (1), where R is the gas
constant and T is the temperature in Kelvin.
enthalphic and entropic contributions. Resolutions catalyzed
by the Trp104Ala mutant have a much lower racemic
temperature (Table 1) than those catalyzed by the wild-type
lipase. This corresponds to the three to four times higher
entropy contribution of the mutant.
The enantioselectivity toward 1-phenylethanol changed
from a strong R selectivity with wild-type CALB to an
S selectivity with the Trp104Ala mutant. The Trp104Ala
mutation increases the size of the stereospecificity pocket,
which should facilitate the binding of the S enantiomer in an
orientation conducive to catalysis. This would lead to a higher
ratio of productive binding of the S enantiomer in the
Trp104Ala mutant compared to that in the wild-type
CALB, and accordingly to a higher reaction rate. The
°
SꢀR
°
SꢀR
°
SꢀR
°
°
°
E ¼ eꢀ
DSꢀRDG =RT
¼ eðꢀDSꢀRDH
þTDSꢀRDS Þ=RT
ð1Þ
The value of the thermodynamic components can be
calculated by determining the enantioselectivity (E) at
several temperatures. This was done for the Trp104Ala
mutant with 1-phenylethanol in three solvents (see Figure 2)
and with three other secondary alcohols in cyclohexane. The
enantioselectivity of the wild-type enzyme was determined
for only one of the substrates at several temperatures; with
the other substrates the R selectivity was too large to be
accurately determined by resolution. Equation (1) was rear-
ranged into Equation (2) and, if one assumes that the
thermodynamic components are constant within the exper-
imental temperature range, there is a linear correlation
app
app
apparent kinetic constants (kcat and kM ) were determined
for wild-type CALB and Trp104Ala mutant for the acylation
of (R)- and (S)-1-phenylethanol with vinyl butanoate
(
500 mm) in cyclohexane at 308C (Table 2). The specificity
constant (k /K ) and the enantioselectivity (E) were calcu-
cat
M
lated using the kinetic constants determined for both enzyme
variants (Table 2).
The enantioselectivity was changed by a factor of 8300000
by the carefully selected point mutation in the stereospeci-
ficity pocket of CALB, which is more than any other example
found in the literature. Rational design has been used to
create variants of a phosphotriesterase with 6900000-fold
difference in enantioselectivity, but the largest difference
compared to the wild-type enzyme was 18000-fold. The
huge change in enantioselectivity by the Trp104Ala mutation
of CALB was mainly achieved by increasing the apparent
ꢀ
1 [16]
between RlnE and T . A good linearity was also seen,
which justified this assumption. The differential entropy and
enthalpy terms and their standard errors were calculated
(
Table 1) by linear regression to Equation (2).
[
5]
°
°
D
DG
D
SꢀR
DH
SꢀR
°
ð2Þ
R ln E ¼ ꢀ
¼ ꢀ
þ DSꢀRDS
T
T
app
catalytic constant (kcat ) toward the slow-reacting enantiomer.
In all the resolutions tested, the R enantiomer was favored
The apparent catalytic constant toward (S)-1-phenylethanol
was 64000 times larger for the mutant than for the wild-type
CALB, while the same constant decreased by a factor of 130
toward the R enantiomer. However, the mutation did not
°
by the enthalpy (D DH > 0) and the S enantiomer by the
SꢀR
°
entropy (D DS > 0). The resolution of 3-methyl-2-butanol
SꢀR
catalyzed by wild-type CALB afforded the R enantiomer, as
the enthalpic contribution was larger than the entropic one
app
affect the apparent Michaelis constant (kM ) to a large extent.
(
Table 1, entry 7). The magnitudes of the thermodynamic
The apparent Michaelis constants for the Trp104Ala mutant
4
584
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2005, 44, 4582 –4585

Angewandte
Chemie
were almost equal toward the R and S enantiomers, and they
both decreased by a factor of two compared to those of the
wild-type enzyme.
[15] M.-J. Kim, Y. I. Chung, Y. K. Choi, H. K. Lee, D. Kim, J. Park, J.
Am. Chem. Soc. 2003, 125, 11494.
[
16] P. L. A. Overbeeke, J. Ottosson, K. Hult, J. A. Jongejan, J. A.
Duine, Biocatal. Biotransform. 1999, 17, 61.
In summary, the redesigned stereospecificity pocket of
Candida antarctica lipase B (CALB) was able to accommo-
date much larger groups than that of the wild-type lipase. This
change transformed the strongly R-selective wild-type CALB
into an S-selective mutant. The S selectivity increased with
temperature and it was dominated by entropy. For 1-phenyl-
ethanol the enantioselectivity changed by a factor of 8300000
as a result of the mutation. This was mainly achieved by an
increased reaction rate toward the S enantiomer, which
resulted in a very effective catalyst with a high specificity
constant of 1000 s m . The S selectivity of the Trp104Ala
mutant could be increased to a respectable value of E = 44
when the reaction was conducted in cis-decalin at 698C. The
altered enantioselectivity of CALB is a demonstration of the
possibilities offered by protein redesign and shows the
importance of the entropy contribution in enzyme catalysis.
[
[
17] J. Ottosson, L. Fransson, K. Hult, Protein Sci. 2002, 11, 1462.
18] a) R. S. Phillips, Enzyme Microb. Technol. 1992, 14, 417; b) K.
Watanabe, T. Koshiba, Y. Yasufuku, T. Miyazawa, S.-I. Ueji,
Bioorg. Chem. 2001, 30, 65; c) P. Lopez-Serrano, M. A. Wegman,
F. van Rantwijk, R. A. Sheldon, Tetrahedron Asymmetry 2001,
12, 235; d) B. Galunsky, S. Ignatova, V. Kasche, Biochim.
Biophys. Acta 1997, 1343, 130.
[
19] R. S. Phillips, Trends Biotechnol. 1996, 14, 13.
ꢀ1
ꢀ1
Received: March 16, 2005
Published online: June 23, 2005
Keywords: enantioselectivity · enzyme catalysis · hydrolases ·
.
protein engineering · thermodynamics
[
[
1] T. Ema, Curr. Org. Chem. 2004, 8, 1009.
2] R. J. Kazlauskas, A. N. E. Weissfloch, A. T. Rappaport, L. A.
Cuccia, J. Org. Chem. 1991, 56, 2656.
[
3] a) M. Cygler, P. Grochulski, R. J. Kazlauskas, J. D. Schrag, F.
Bouthillier, B. Rubin, A. N. Serreqi, A. K. Gupta, J. Am. Chem.
Soc. 1994, 116, 3180; b) A. Kovac, H. Scheib, J. Pleiss, R. D.
Schmid, F. Paltauf, Eur. J. Lipid Sci. Technol. 2000, 102, 61;
c) R. V. Muralidhar, R. R. Chirumamilla, R. Marchant, V. N.
Ramachandran, O. P. Ward, P. Nigam, World J. Microbiol.
Biotechnol. 2002, 18, 81.
[
4] H. Scheib, J. Pleiss, P. Stadler, A. Kovac, A. P. Potthoff, L.
Haalck, F. Spener, F. Paltauf, R. D. Schmid, Protein Eng. 1998,
11, 675.
[
5] M. Chen-Goodspeed, M. A. Sogorb, F. Wu, F. M. Raushel,
Biochemistry 2001, 40, 1332.
[
6] a) Y. Koga, K. Kato, H. Nakano, T. Yamane, J. Mol. Biol. 2003,
331, 585; b) D. Zha, S. Wilensek, M. Hermes, K.-E. Jaeger, M. T.
Reetz, Chem. Commun. 2001, 24, 2664.
[
[
7] B. A. Persson, A. L. E. Larsson, M. Le Ray, J.-E. Baeckvall, J.
Am. Chem. Soc. 1999, 121, 1645.
8] a) E. M. Anderson, K. M. Larsson, O. Kirk, Biocatal. Biotrans-
form. 1998, 16, 181; b) D. Rotticci, J. Ottosson, T. Norin, K. Hult,
Methods Biotechnol. 2001, 15, 261.
[
9] a) J. Uppenberg, N. Oehrner, M. Norin, K. Hult, G. J. Kleywegt,
S. Patkar, V. Waagen, T. Anthonsen, T. A. Jones, Biochemistry
1995, 34, 16838; b) C. Orrenius, F. Haeffner, D. Rotticci, N.
Ohrner, T. Norin, K. Hult, Biocatal. Biotransform. 1998, 16, 1.
[
[
10] F. Haeffner, T. Norin, K. Hult, Biophys. J. 1998, 74, 1251.
11] D. Rotticci, F. Haeffner, C. Orrenius, T. Norin, K. Hult, J. Mol.
Catal. B 1998, 5, 267.
[
[
[
12] A. O. Magnusson, J. C. Rotticci-Mulder, A. Santagostina, K.
Hult, ChemBioChem 2005, 6, 1051.
13] J. Ottosson, L. Fransson, J. W. King, K. Hult, Biochim. Biophys.
Acta 2002, 1594, 325.
14] B. Martin-Matute, M. Edin, K. Bogar, J.-E. Baeckvall, Angew.
Chem. 2004, 116, 6697; Angew. Chem. Int. Ed. 2004, 43, 6535.
Angew. Chem. Int. Ed. 2005, 44, 4582 –4585
www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4585