Increased activity of enzymatic transacylation of acrylates through rational design of lipases

Article abstract of DOI:10.1016/j.molcatb.2009.11.016

A rational design approach was used to create the mutant Candida antarctica lipase B (CALB, also known as Pseudozyma antarctica lipase B) V190A having a k_cat three times higher compared to that of the wild type (wt) enzyme for the transacylation of the industrially important compound methyl methacrylate. The enzymatic contribution to the transacylation of various acrylates and corresponding saturated esters was evaluated by comparing the reaction catalysed by CALB wt with the acid (H₂SO₄) catalysed reaction. The performances of CALB wt and mutants were compared to two other hydrolases, Humicola insolens cutinase and Rhizomucor mihei lipase. The low reaction rates of enzyme catalysed transacylation of acrylates were found to be caused mainly by electronic effects due to the double bond present in this class of molecules. The reduction in rate of enzyme catalysed transacylation of acrylates compared to that of the saturated ester methyl propionate was however less than what could be predicted from the energetic cost of breaking the π-system of acrylates solely. The nature and concentration of the acyl acceptor was found to have a profound effect on the reaction rate.

Full text of DOI:10.1016/j.molcatb.2009.11.016

Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Increased activity of enzymatic transacylation of acrylates through rational
design of lipases
Per-Olof Syréna, Ebba Lindgrena,1, Hans Wolfgang Hoeffkenb, Cecilia Brannebyb,2
,
Steffen Maurerb, Bernhard Hauerb,3, Karl Hulta,∗
a Royal Institute of Technology (KTH), School of Biotechnology, Department of Biochemistry, AlbaNova University Centre, SE-106 91 Stockholm, Sweden
b BASF SE, GVF/E - A030, D-67056 Ludwigshafen, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
Available online 2 December 2009
A rational design approach was used to create the mutant Candida antarctica lipase B (CALB, also known as
Pseudozyma antarctica lipase B) V190A having a kcat three times higher compared to that of the wild type
(wt) enzyme for the transacylation of the industrially important compound methyl methacrylate. The
enzymatic contribution to the transacylation of various acrylates and corresponding saturated esters was
evaluated by comparing the reaction catalysed by CALB wt with the acid (H2SO4) catalysed reaction. The
performances of CALB wt and mutants were compared to two other hydrolases, Humicola insolens cutinase
and Rhizomucor mihei lipase. The low reaction rates of enzyme catalysed transacylation of acrylates were
found to be caused mainly by electronic effects due to the double bond present in this class of molecules.
The reduction in rate of enzyme catalysed transacylation of acrylates compared to that of the saturated
ester methyl propionate was however less than what could be predicted from the energetic cost of
breaking the ␲-system of acrylates solely. The nature and concentration of the acyl acceptor was found
to have a profound effect on the reaction rate.
Keywords:
Lipase
CALB
Point mutations
Kinetics
Proficiency
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Lipase catalysed transacylation of acrylates has been described
in the literature. In a pioneering work regioselective acrylation
Acrylates are industrially important compounds that have a
wide range of commercial applications. The ␣,␤-double bond
present in these molecules can be used for radical polymerization
to yield final products such as paints, plastics and adhesives to be
used for example in the automotive industry. Traditionally, esters
of acrylic acid and methacrylic acid are made chemically at elevated
temperature, with polymerization inhibitors and with sulphuric
acid as catalyst [1,2]. Biocatalysis is a “green” option to traditional
chemistry and could provide sustainable routes to acrylic esters
starting from renewable sources such as glucose and lactate [3,4].
Enzymes are attractive since they allow the use of mild reaction
conditions, often displaying high activity as well as high regio-
/stereo selectivity [5,6]. Especially lipases are of industrial interest
due to their high stability and activity in organic media.
of sterically hindered diols was performed using a lipase from
Chromobacterium viscosum [7]. Lipase B from Candida antarctica
(CALB, also known as Pseudozyma antarctica lipase B, EC 3.1.1.3)
was shown to be a superior catalyst out of 19 enzymes investigated
in the enzyme catalysed transacylation of methyl acrylate with 1-
undecanol [8]. The industrial interest in enzymatic acrylation is
shown in the number of patents on the subject. Enzyme catalysed
regioselective transacylation of acrylates with glucosides, with the
aim of producing polymeric sugar acrylates, as well as monoacryla-
tion of polyols has been described by BASF [9,10]. One limitation of
enzyme catalysed acrylation is the low reaction rates compared to
the rates for corresponding saturated esters. Attempts have been
made to increase the reaction rate of enzymatic acrylation, mainly
by the choice of solvent and by optimisation of the water activity
[11,12]. The source of the low rates has to our knowledge not been
discussed.
In this work we have employed a rational design approach
based on computer modelling with the aim of increasing the enzy-
matic activity (kcat) for methyl methacrylate. The analysis of the
CALB mutants D134N, I189A, I189V, V190A, V190L, and I285W
gave valuable insight into the problems associated with enzymatic
acrylation. Humicola insolens cutinase (EC 3.1.1.74) and Rhizomu-
cor miehei lipase (EC 3.1.1.3) were compared to CALB regarding
∗
Corresponding author. Tel.: +46 8 5537 8364; fax: +46 8 5537 8468.
E-mail address: kalle@biotech.kth.se (K. Hult).
Present address: Leco Corporation Svenska AB, Lövängsvägen 6, SE-194 45 Upp-
1
lands Väsby, Sweden.
2
Present address: Cambrex Karlskoga AB, SE-691 85 Karlskoga, Sweden.
Present address: Institute of Technical Biochemistry, University of Stuttgart,
3
Allmandring 31, D-70569 Stuttgart, Germany.
1381-1177/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2009.11.016

4
P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
performance in enzymatic acrylation. The enzymatic contribution
to the transacylation of methyl acrylate and derivatives as well as
to corresponding saturated esters was evaluated.
Alcohols are known to inhibit lipases by forming a dead-end
complex with the enzyme [13]. The protein engineering approach
to enhance enzymatic activity was compared to a reaction engi-
neering approach. The nature and concentration of the alcohol acyl
acceptor was found to have a profound effect on the reaction rate
for enzyme catalysed acrylation.
sion, the wt and mutated pGAPz␣␤ plasmids were harvested by
the QIAprep Miniprep System (QIAGEN GmbH, Hilden, Germany).
Around 20–40 ␮g plasmid DNA was linearized by Bsp HI overnight
at 37 ◦C and concentrated by EtOH precipitation according to stan-
dard protocols. Approximately 10 ␮g of the linearized plasmid was
electroporated into freshly made Pichia pastoris SMD1168H compe-
tent cells using the following settings: 1.5 kV, 400 ꢀ, 25 ␮F. Ice-cold
1 M sorbitol was immediately added after transformation. After 2 h
at 30 ◦C, half of the transformation mixture was plated on YPDS-
plates (1% yeast extract, 2% peptone, 2% dextrose (d-glucose), 1 M
sorbitol, 2% agar) containing 100 ␮g/mL zeocin. The other half of
the mixture was diluted with YPD-media (1% yeast extract, 2%
peptone, 2% dextrose (d-glucose)) and was allowed to shake at
200 rpm for 3 h at 30 ◦C after which plating was performed (on
YPDS-plates containing 100 ␮g/mL zeocin). Around 10 colonies
were selected and grown in 10 mL YPD for approximately 36 h. After
centrifugation at 4000 rpm, 4 ◦C, for 10 min using a Sorvall Super
T21 centrifuge, the culture supernatants were analysed by SDS-gel
electrophoresis (NuPAGE® 10% Bis-Tris gel, Invitrogen, Carlsbad,
CA, USA) using MOPS as running buffer (MOPS 50 mM; Tris base
50 mM; SDS 0.1%; EDTA 1 mM; pH 7.7). The validity of the C. antarc-
tica lipase B protein band was established by performing Western
blot analysis using antibodies against CALB as previously described
[14].
2. Experimental
2.1. Chemicals and enzymes
Methyl propionate, methyl isobutyrate, methyl trimethylac-
etate, methyl acrylate, methyl crotonate, methyl methacrylate,
1-propanol, 2-butanol, acetonitrile (for fluorescence) and dode-
cane of ≥99% purity and H2SO4 ≥ 95% purity were purchased
from Sigma–Aldrich (St. Louis, MO, USA). Methyl ␣-chloro acrylate
(≥99%), stabilised with hydroquinone, was from Acros Organ-
ics (Morris Plains, NJ, USA). Diisopropyl ether (≥99%) was from
AppliChem (Darmstadt, Germany). 2-Ethyl-1-hexanol was from
BASF (Ludwigshafen, Germany).
H. insolens cutinase wt was a kind gift from Novozymes
(Bagsvaerd, Denmark). Immobilised R. miehei lipase was purchased
from Sigma–Aldrich. Restriction enzymes were from Fermentas (St.
Leon-Rot, Germany).
2.3. Protein expression and purification
Positive Pichia pastoris clones expressing CALB variants were
grown on YPD-agar plates containing 100 ␮g/mL zeocin. After 2
days, colonies were picked and grown for 24 h in 5 mL YPD at 30 ◦C,
200 rpm. A fraction of the pre-cultures (to obtain a 1:100 dilution)
were transferred to 50–500 mL BMGY (1% yeast extract, 2% pep-
tone, 100 mM potassium phosphate, 1.34% yeast nitrogen base with
ammonium sulphate without amino acids, 0.4 mg/L (4 × 10−5%)
biotin, 1% glycerol, pH 6.0). The main cultures were allowed to grow
for 3–4 days at 30◦ at 200–260 rpm after which the optical den-
sity (OD600) typically reached ≥40. The cultures were centrifuged
at 4000 rpm, 4 ◦C for 20 min. Protein purification was performed
using hydrophobic interaction chromatography (HIC) as previously
described [16] with minor modifications. Ammonium acetate (9 M)
was slowly added under stirring to the supernatant to a final con-
centration of 0.8 M. The feed-stock was filtered through a 0.2 ␮M
bottle-top filter (Sartorius AG, Goettingen, Germany) and loaded
onto an XK-16 column packed with 15 mL butyl sepharose FF and
connected to an ÄKTA explorer (GE Healthcare, Uppsala, Sweden).
The equilibration buffer was 50 mM potassium phosphate, 0.8 M
ammonium acetate, pH 6. A linear gradient to 50 mM potassium
phosphate, pH 6 was used for elution and finally milliQ was used
to elute remaining protein from the column. Fractions giving raise
to an A280-signal were pooled and concentrated (by a factor of 10)
with 50 mM potassium phosphate, pH 7.5 by using Centricon Cen-
trifugal filters with an MWCO of 5 kDa (Millipore, MA, USA). The
purity of the CALB variants was assessed by SDS polyacrylamide
gel electrophoresis. Purification typically resulted in 10–20 mg pure
protein/L media.
2.2. Preparation of C. antarctica lipase B wt and mutants
The preparation of the CALB wt (also known as Pseudozyma
antarctica lipase B) construct in a pGAPz␣␤ plasmid and its
transformation into the expression host Pichia pastoris has been
described previously in detail [14]. The QuikChange® protocol from
Stratagene (La Jolla, CA, USA) was used to introduce mutations.
The mutagenic primers where designed to be non-overlapping
in the ends [15] and were ordered from Thermo Fisher Scientific
(Ulm, Germany). The DNA-sequences in the 5ꢀ–3ꢀ direction for the
forward (f) and reverse (r) primers are given below. Introduced
alterations in the wt sequence are in uppercase letters for clarity.
D134N f
5ꢀ-ggcctttgcgcccAactacaagggcaccgtcctcg-3ꢀ
D134N r
5ꢀ-ggtgcccttgtagtTgggcgcaaaggccataagtcgatcg-3ꢀ
I189V f
5ꢀ-cgaccgacgagGtcgttcagcctcaggtgtcc-3ꢀ
I189V r
5ꢀ-cctgaggctgaacgaCctcgtcggtcgccg-3ꢀ
I189A f
5ꢀ-cggcgaccgacgagGCcgttcagcctcaggtgtcc-3ꢀ
I189A r
5ꢀ-cctgaggctgaacgGCctcgtcggtcgccgagtagagg -3ꢀ
V190A f
5ꢀ-cgacgagatcgCtcagcctcaggtgtcc-3ꢀ
V190A r
5ꢀ-cctgaggctgaGcgatctcgtcggtcg-3ꢀ
V190L f
5ꢀ-cgaccgacgagatcCttcagcctcaggtgtccaactcg-3ꢀ
V190L r
5ꢀ-cctgaggctgaaGgatctcgtcggtcgccgagtagagg-3ꢀ
I285W f
2.4. Protein immobilization and active site titration
5ꢀ-gctgcagccTGGgtggcgggtccaaagcagaactgc-3ꢀ
I285W r
Lipase immobilization and active site titration have previously
been described by our group [17]. C. antarctica lipase B wt and
mutants as well as H. insolens cutinase were immobilized on Accurel
MP 1000 < 1500 ␮m (Accurel Systems, Sunnyvale, CA, USA). The
supernatant was confirmed to be devoid of lipase activity after
immobilization. The lipase-containing polypropylene beads were
vacuum-dried overnight and put in a desiccator under LiCl(s) to
ensure a water activity of 0.1.
5ꢀ-ggacccgccacCCAggctgcagctgccggcgccaggagc-3ꢀ
The PCR reactions were analysed on a 1% agarose gel for the
presence of products. After DpnI digestion, the mutated plasmid
DNA as well as the wt plasmid DNA was transformed by heat-shock
into E. coli XL1Blue cells according to the QuikChange protocol.
All mutations were confirmed by sequencing. Prior to expres-

P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
5
The suicide inhibitor 4-methylumbelliferyl hexylphosphonate
was added to 10–50 mg polypropylene beads containing lipase
in acetonitrile. The final concentration of the inhibitor in the
experiments was 50 ␮M and the total volume was 1 mL. The fluo-
rescence intensity was analysed by adding 100 ␮L sample solution
to 900 ␮L buffer containing 100 mM Tris–HCl, 1 mM CaCl2, pH
8.0 and by measuring the signal on a luminescence spectrome-
ter (Luminescence Spectrometer Model LS-50B, PerkinElmer, MA,
USA). Excitation was at 360 nm and the emission wavelength was
445 nm. The amount of active sites were calculated from a stan-
dard curve of the linear relationship between fluorescense intensity
and the concentration of the leaving group 4-methylumbelliferone.
Resulting protein load on the support for C. antarctica lipase B was
found to be (% weight/weight): 0.5% for wt, 0.5% for D134N, 0.3%
for I189A, 0.6% for I189 V, 0.2% for V190A, 1% for V190L and 0.8%
for I285W. The corresponding numbers were 0.5% for H. insolens
cutinase and 0.2% for R. miehei lipase.
2-butanol was used as acyl acceptor at a concentration of 470 mM.
Typically, the acyl donor concentration was varied from 50 mM to
bulk-conditions (when the acyl donor is the solvent). The appar-
ent kcat values were calculated by fitting the experimental data to
pseudo one-substrate Michaelis–Menten kinetics using non-linear
regression in the solver function of Microsoft Excel 2003. At least
6 points were used to create a Michaelis–Menten curve. The rate
constants were measured using the acyl donor as solvent (from at
least 2 separate reactions). Turnover of substrate never exceeded
10% conversion and the concentrations of active sites were taken
into consideration when calculating apparent rate constants. The
temperature was kept constant at 30 ◦C and the shaking speed was
100 rpm using an HLC heat-block. Typically 10 mg Accurel MP 1000
with immobilised enzyme (approximately 1 nmol active site) was
used for each reaction mixture.
Apparent KM values were either determined from
Michaelis–Menten curves or estimated from known absolute
kcat/KM values and known rates in bulk (which was then set to
represent kcat).
2.5. Kinetic measurements
The relative initial rate of CALB catalysed transacylation of
methyl methacrylate as a function of the mole fraction of 1-
propanol was performed in
Corresponding experiments were performed using 2-butanol and
2-ethyl-1-hexanol as acyl acceptors.
All reactions were performed in glass tubes with screw-lids. All
glassware was dried at 250◦ prior to use and chemicals were dried
by molecular sieves (3 A, Sigma–Aldrich).
a
solvent free system at 30 ◦C.
The relative substrate specificities were measured by com-
petition experiments where several acyl donors were used
simultaneously in the same reaction tube. The temperature was set
to 30 ◦C using a water bath and magnetic stirring was used. Typi-
cally, the acyl donor concentration was in the range of 50–200 mM
and the acyl acceptor 1-propanol had a concentration of 100 mM.
Dodecane was used as internal standard at a concentration of
2.5 mM. Diisopropyl ether was chosen as solvent, being moder-
ately hydrophobic and able to dissolve all substrates and products.
The total reaction volume was 2 mL. The relative specificities were
not dependent on the alcohol concentration used. As catalysts,
8% vol./vol. H2SO4 (resulting in relative second-order rate con-
stants) or 10 mg beads with immobilized lipase (CALB wt) were
used (resulting in relative kcat/KM values). Samples from the reac-
tion mixtures were taken at regular time intervals and diluted 10
times in diisopropyl ether. The amount of product (propyl esters)
was determined by GC-analysis using a HP 5890 Series II Gas
Chromatograph (Agilent Technologies, Santa Clara, CA, USA). GC-
columns were Varian Chrompack Capillary column, WCOT fused
silica, 25 m × 0.32 mm in diameter, CP Chirasil-DEX CB, DF = 0.25
and Chrompack WCOT fused silica 25 m × 0.32 mm in diameter, CP-
SIL 5CB, DF = 1.2 respectively. The initial rate was calculated from
data of product concentration as a function of time. Only data cor-
responding to a conversion of 10% or less were considered. The
relative kcat/KM values were calculated from Eq. (1) where A and B
are the two acyl donors being compared and vA and vB the ini-
tial rates for the production of the corresponding propyl esters.
The same equation is valid for the acid catalysed reaction if the
second-order rate constants are used.
Background reactions were in all cases found to be negligible.
2.6. Computer modelling on C. antarctica lipase B
The structure of the tetrahedral intermediate as a model of the
transition state (TS) for acylation of methyl methacrylate catal-
ysed by C. antarctica lipase B was prepared using YASARA version
8.9.23 [18]. The PDB-file 1TCA was used as starting structure. Sugars
were deleted, all missing hydrogens were added and the hydrogen
bonding network was optimised by repeated minimizations and
short dynamics runs (1200 fs at 298 K) using the Amber99 force
field [19] and by first keeping the protein backbone fixed. Free
methyl methacrylate was built and energy minimised in YASARA
using the AUTOSMILES approach [20]. The tetrahedral intermedi-
ate was then created by the covalent attachment of the carbonyl
carbon in methyl methacrylate to the catalytic S105 in CALB. The
resulting tetrahedral intermediate was minimized and molecular
dynamics was run on the covalent enzyme–substrate complex for
2 ps at 298 K using Amber99 and PME for long range electrostatics
[21]. This was repeated several times and the molecular dynam-
ics simulation was allowed to proceed until (1) the oxyanion hole
(Q106 and T40) was clearly established and contributing with at
least two hydrogen bonds to the (former) carbonyl oxygen and (2)
the hydrogen bonds in the catalytic triad formed (D187 hydrogen
bonded to H224 and H224 sharing a hydrogen bond between the
ester oxygen in methyl methacrylate and the O␥ of the catalytic ser-
ine). The catalytic histidine was in its protonated state. To explore
the conformational space the dihedral angle between the carbonyl
oxygen, carbonyl carbon, and the unsaturated ␣- and ␤-carbons
in the acrylate substrate in TS was manually varied followed by
minimization and molecular dynamics without constraints at 298 K
(for 2 ps). The aim of the simulations was to get a visual picture of
potential problematic interactions in TS.
(kcat/kM
(kcat/kM
)
)
vA
vB
B
[ ]
A
=
×
(1)
A
[ ]
B
Absolute kcat/KM values for methyl propionate were determined
for the various lipases using 25 mM 1-propanol as acyl acceptor at
30 ◦C. The concentrations of active sites determined by active site
titration were taken into account. The known relative kcat/KM values
were then used to obtain the absolute kcat/KM values at 25 mM 1-
propanol for the other substrates.
3. Results and discussion
The apparent kcat values and rate constants for enzymatic
transacylation of methyl propionate, methyl acrylate, methyl
methacrylate and methyl ␣-chloro acrylate were measured at
30◦ in diisopropyl ether as solvent with 25 mM 1-propanol as
nucleophile and 2.5 mM dodecane as internal standard. Optionally,
3.1. Comparison of acid and enzyme catalysed transacylation of
acrylates and corresponding saturated esters
To evaluate the contribution of lipases in transacylation reac-
tions of acrylates in relation to methyl propionate the enzyme

6
P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
catalysed reaction was compared to the reaction catalysed by
H2SO4. The apparent second-order rate constant for the enzyme
(kcat/KM) was compared to the corresponding second-order rate
constant for the acid. The relative second-order rate constants
for H2SO4-catalysed transacylation of methyl acrylate, methyl
methacrylate (relative second-order rate constant k = 1), methyl
crotonate, methyl ␣-chloro acrylate as well as for the corresponding
saturated esters are given in Table 1 together with the correspond-
ing relative kcat/KM values for C. antarctica lipase B (with the value
for methyl methacrylate also set to 1).
Methyl propionate was chosen as reference substrate for the
lipase catalysed transacylation reactions since it is the correspond-
ing saturated ester of methyl acrylate and is more similar to natural
substrates. The reaction rate for acid catalysed transacylation of
methyl methacrylate was 5–6 times lower than for methyl acry-
late. The corresponding figure for CALB was a 17-fold reduction.
This finding is encouraging since it indicates that the enzymatic
transacylation of methyl methacrylate could in theory be improved
at least three times if the normal capacity of the enzyme could be
used. We believe that sterical hindrance around the ␣-carbon of
the substrate as well as electronic effects are important. Acrylates
have been shown to predominantly exist in two flat conformations,
s-cis and s-trans (which indicates the relative orientation of the two
double bonds) [22]. Substrate geometrical effects can be seen in the
fact that the enzyme specificity for methyl isobutyrate was a magni-
tude larger than for the iso-steric compound methyl methacrylate.
In the acid catalysed reaction methyl trimethylacetate displayed
a twofold higher reactivity than methyl methacrylate whereas the
enzyme had very low specificity towards this bulky substrate. Inter-
estingly, the enzyme specificity for methyl crotonate was higher
than for methyl methacrylate. This was also true for methyl ␣-
chloro acrylate, indicative of an electronic effect since the size of the
chlorine is close to that of the methyl group. The relative enzymatic
contribution is graphically visualized in Fig. 1. It can be concluded
that CALB is not working at its maximal capacity for substrates
with sterical hindrance at the ␣-position when taking the intrin-
sic chemical reactivities of the substrates into account. It should
also be noted that the relative enzymatic contribution to cataly-
sis for methyl propionate and methyl acrylate were essentially the
same.
3.2. Molecular model of the transition state of CALB-catalysed
acylation of methyl methacrylate
The tetrahedral intermediate has been shown to be a good
approximation of the transition state (TS) for lipase catalysed acy-
lation [23]. Fig. 2 shows a graphical representation of the TS for
Fig. 1. Relative enzymatic contributions were calculated from the apparent second-
order rate constants for enzyme (CALB wt) and acid (H2SO4) catalysed transacylation
(kcat/KM and k respectively). The enzymatic contribution to the transacylation of
methyl propionate was used as reference (relative contribution set to 0). Negative
relative energies are associated with higher degree of enzymatic contribution ꢁsub
=
Methyl propionate
.
Acid-Enzyme
substrate
ꢁ
; ꢁprop = ꢁ
Acid-Enzyme

P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
7
Fig. 2. A minimised snapshot of the tetrahedral intermediate from the dynamics
simulation of covalently attached methyl methacrylate (MMA) in CALB. Residues
targeted for mutation are shown (D134, I189, V190, I285). The catalytic triad (S105,
D187, H224) and oxyanion hole (T40, Q106) are shown. Atoms are coloured accord-
ing to type.
CALB wt-catalysed acylation of methyl methacrylate (one energy
minimised snapshot from the molecular dynamics simulation).
The ␣-methyl group in methyl methacrylate is in proximity of
I189 and V190 (closest distance between carbon atoms are 3.6 Å in
both cases). From this simple model it can be concluded that the
mutations I189 V, I189A, V190A and V190L could relieve potential
sterical clashes in TS. D134N changes the electronic environment
around the substrate and I285W could potentially introduce ␲-
stacking interactions.
3.3. Performances of lipases in enzymatic acrylation
The performances of CALB wt and mutants, H. insolens cuti-
nase and R. miehei lipase were analysed in transacylation reactions
of acrylates and corresponding saturated compounds. Substrate
specificities are given in Table 2. The first observation is that the
specificities were low and very far from the diffusion limit of
1010
M
−1 min−1 [24]. The specificities dropped one order of mag-
nitude for methyl acrylate compared to methyl propionate and
two orders of magnitude for methyl methacrylate compared to
methyl propionate for all enzymes tested. The enzyme specificity
for methyl isobutyrate was one order of magnitude larger than for
the isosteric methyl methacrylate. We believe that the flat s-cis/s-
trans acrylate conformers interact unfavourably with the active
site of lipases. Methyl isobutyrate is not forced to be in a planar
arrangement. H. insolens cutinase seemed to have a higher intrinsic
specificity for acrylates compared to the other enzymes investi-
gated. R. miehei lipase displayed a kcat/KM ratio of methyl acrylate
and methyl methacrylate of two, which is lower than the relative
chemical reactivity of these substrates. Clearly, three methyl groups
on the ␣-position as in methyl trimethylacetate was too bulky for
CALB.
Enzyme substrate specificities (kcat/KM) tell something about
both binding and catalysis and corresponds to the observed rate
constant at very low substrate concentrations. For industrial pro-
cesses, maximal rates are of high interest since they indicate how
many molecules that can be converted into product per unit time
and per enzyme molecule under saturating conditions. By using
bulk-conditions (i.e. using the acyl donor as solvent) such max-
imal rates can be estimated (in cases where substrate inhibition
is not predominant). Rate constants for enzymatic transacylation
of acrylates and for methyl propionate are given in Table 3. The
reactions were performed in bulk with 25 mM 1-propanol as acyl
acceptor. The rate constant for methyl acrylate was observed to be

8
P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
Table 3
Rate of enzymatic transacylation in pure acyl donor using 25 mM 1-propanol as acyl acceptor at 30 ◦C.
Catalyst
rate [min−1
]
rate [min−1
]
rate [min−1
]
rate [min−1
]
CALB wta
I189Aa
1600
300
800
1300
1000
3000
580
170
240
240
170
35
60
30
25
85
80
50
290
n.m.b
n.m.b
600
I189Va
V190Aa
Cutinasec
RMLd
280
120
a
Candida antarctica lipase B.
Not measured.
Humicola insolens cutinase.
Rhizomucor miehei lipase.
b
c
d
Table 4
Apparent KM values using 25 mM 1-propanol and diisopropyl ether as solvent at 30 ◦C.
Catalyst
KMapp [M]
KMapp [M]
KMapp [M]
1.5b
>13
1.4
10b
KMapp [M]
CALB wta
I189Aa
0.2b
0.4
0.2
0.9
0.1
0.3
0.6b
2
0.6
1.5
0.2
0.4
0.1b
n.m.c
n.m.c
1b
0.05
0.1
I189Va
V190Aa
Cutinased
RMLe
0.6
1.3
The apparent KM values were estimated from known absolute kcat/KM values and the rate constants from reactions where the acyl donor was used as solvent or from
Michaelis-Menten data as indicated.
a
Candida antarctica lipase B.
b
KM values from Michaelis–Menten kinetics using 25 mM 1-propanol and diisopropyl ether as solvent at 30 ◦C.
c
Not measured.
Humicola insolens cutinase.
Rhizomucor miehei lipase.
d
e
around 6 times lower than for methyl propionate (with the excep-
tion of R. miehei lipase where the decrease in rate was 85 times).
Thestabilising ␲-systempresent in methylacrylate is wortharound
17 kJ/mol in diethyl ether [25]. The conjugated ␲-system has to be
broken during the course of the reaction so the observed rate for
transacylation of methyl acrylate is higher than expected (17 kJ/mol
would correspond to a 1000-fold reduction in rate). The conversion
of methyl methacrylate to propyl methacrylate proceeded slower
by an order of magnitude compared to methyl acrylate, indica-
tive of steric hindrance. Interestingly, R. miehei lipase displayed a
slightly higher rate for the transacylation of methyl methacrylate as
compared to methyl acrylate. Methyl ␣-chloro acrylate displayed
a higher rate compared to methyl methacrylate indicative of an
electronic effect.
The apparent estimated KM values for the acrylates and the ref-
erence compound methyl propionate are given in Table 4. It is
seen that poor substrate binding is a problem in enzymatic acryla-
tion. This is clearly illustrated for methyl methacrylate with a KM
of 1.5 M for CALB wt and 10 M for the CALB V190A mutant (pure
methyl methacrylate has a concentration around 9 M). Clearly, the
mutations I189A and V190A were deleterious for substrate binding.
Apparent kcat values for CALB-catalysed transacylation of acry-
lates are given in Table 5 with the values using 2-butanol as acyl
acceptor given within brackets. CALB V190A showed three times
higher reactivity towards methyl methacrylate compared to wt.
By using the activated ester vinyl methacrylate and 1-propanol,
the apparent kcat value increased 10-fold compared to methyl
methacrylate for both CALB wt and the V190A mutant. It should be
noted that the apparent KM value for vinyl methacrylate was around
10 M for both wt and mutant using 1-propanol. The observed
rate increase was thus a “real” effect and not caused by success-
ful competition of the activated ester with the alcohol inhibitor.
The acylation step of enzymatic acrylation was thus concluded to
be more rate-limiting using 1-propanol as acyl acceptor. When
using 1-hexanol as acyl acceptor the rate of C. antarctica lipase B
catalysed transacylation of methyl methacrylate was essentially
unchanged. For 2-butanol, the effect of switching to the activated
ester vinyl methacrylate had a less profound effect with only mod-
erate increase in reaction rate compared to methyl methacrylate.
This is indicative that deacylation has become more important and
consistent with the fact that 2-butanol is a less potent nucleophile
compared to 1-propanol. Interestingly by using 2-butanol as a
nucleophile, the apparent KM value for methyl methacrylate for the
V190A mutant dropped to around 4 M, similar to the value for wt
using 2-butanol (data not shown). Saturation was now achievable.
This illustrates the importance of alcohol inhibition. The alcohol
competes with the acyl donor for the lipase active site and thus
affects the observed apparent KM value. A less potent inhibitor
would lead to a decrease in KMapp
.
3.4. Reaction media engineering
Fig. 3 shows the measured relative initial rate for the transacy-
lation of methyl methacrylayte catalysed by C. antarctica lipase B as
a function of the mole fraction of 1-propanol used in the reaction
mixture. The reaction rate varied up to 50 times with an optimal

P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
9
Table 5
Apparent kcat values for enzymatic transacylation of acrylates from pseudo one-substrate Michaelis–Menten kinetics using 25 mM 1-propanol as acyl acceptor in diisopropyl
ether at 30 ◦C. Corresponding apparent kcat values for 470 mM 2-butanol as acyl acceptor are given within brackets.
Catalyst
kcat [min−1
]
kcat [min−1
]
kcat [min−1
]
kcat [min−1
]
CALB wta
V190Aa
460
56 (40)b
140
660
860 (60)b
n.m.c
170 (109)b
2600 (329)b
a
Candida antarctica lipase B.
Values within brackets are apparent kcat values using 470 mM 2-butanol as acyl acceptor.
Not measured.
b
c
Table 6
concentration of 1-propanol close to 25 mM (0.2 mol%, the value
that was used in the transacylation reactions above). For the more
sterically hindered primary alcohol 2-ethyl-1-hexanol and the sec-
ondary alcohol 2-butanol, the optimal concentrations were found
to be around 4–5 mol.% and the reaction rate varied up to 3 times
depending on the alcohol concentration (with a similar profile as
the curve shown in Fig. 3, data not shown).
Activation energies and binding energies relative to methyl propionate for Candida
antarctica lipase B wt calculated from the data given in Section 3.3 at 30 ◦C.
ꢂꢂG‡kcat
ꢂꢂG‡kcat/KM
ꢂꢂGb
[kJ/mol]a
[kJ/mol]b
[kJ/mol]c
3.1
5.5
−2.8
3.5. Factors that affect enzymatic acrylation
As was shown in Sections 3.3 and 3.4 steric hindrance, electronic
effects and the nature of the acyl acceptor affects enzymatic acry-
lation. Poor substrate binding can be beneficial for the value of kcat
if the transition state for the weak binding substrate is unaffected.
This so-called ground state destabilisation (GSD) and circe effect
(the case where one part of the substrate provide binding energy
for the distortion of another part) has been widely discussed in the
literature [26], for example in the context of the source of the rate
enhancement for orotidine-5-monophosphate decarboxylase (EC
4.1.1.23) [27,28]. Substrate binding energies ꢂGb and activation
energies ꢂG‡kcat and ꢂG‡kcat/KM for the lipase catalysed transacy-
lation of the different acrylates relative to methyl propionate are
shown in Table 6. It can be concluded that the transition state
for the different acrylates experience different amount of stabil-
isation. Ground state destabilisation was concluded to be of less
importance.
The influence of electrostatics on enzymatic acrylation was
analysed by making the CALB mutation D134N. This mutation
changed the charge distribution in the active site around the acry-
late substrate. The effect of this mutation was an increased relative
specificity of methyl propionate compared to methyl acrylate by a
factor of 2. The absolute rate, corrected for the number of active
sites, dropped around 2-fold for methyl propionate and 10-fold for
8.4
6.1
12.6
3.2
−5.1
1.8
a
Activation energies relative to methyl propionate based on apparent kcat values
using 25 mM 1-propanol as acyl acceptor.
Activation energies based on apparent kcat/KM values using 25 mM 1-propanol
and relative to methyl propionate.
b
c
Relative substrate binding energies compared to methyl propionate based on
apparent KM values.
methyl methacrylate compared to the corresponding rates for CALB
wt. This indicates that electronic effects may be important. Simi-
larly, the idea behind the CALB I285W mutation was to introduce
␲-stacking. However no rate enhancement for the transacylation
of methyl methacrylate was observed.
The nature and concentration of the acyl acceptor had a big
impact on the rate for the enzyme catalysed reaction. For the sec-
ondary alcohol 2-butanol concentrations up to 470 mM could be
used without significant alcohol inhibition. For 1-propanol, the
optimal concentration was 25 mM. In the literature, 2-octanol was
found to have an inhibition constant 10 times higher compared to
1-octanol in CALB-catalysed transacylation in heptane [13].
4. Conclusions
Lipases have not evolved to deal with planar substrates such
as acrylates. The poor binding of acrylates is accompanied with a
decrease in rate. Clearly, sterical and electronic effects contribute to
the different amount of TS stabilisation experienced by the different
acrylates.
There are several potential ways of increasing reaction rates.
Elevated temperature is one, catalyst design is another and opti-
mising the reaction media is a third. Rational enzyme design
is competitive with directed evolution approaches in the case
of improving enzyme activity for methyl methacrylate. Although
potential screening protocols exist like the Purpald method [29],
the expected increase in rate of created positive mutants lie within
Fig. 3. Relative initial rate of CALB wt-catalysed transacylation of methyl methacry-
late as a function of the mole fraction of acyl acceptor (1-propanol).

10
P.-O. Syrén et al. / Journal of Molecular Catalysis B: Enzymatic 65 (2010) 3–10
the error margin of the screening protocol. Our rational design
approach was successful and resulted in a CALB V190A mutant with
a kcat 3-fold higher than CALB wt for the transacylation of methyl
methacrylate. For industrial purposes kcat values are of higher inter-
est than kcat/KM [30].
The importance of the reaction media was shown by the fact
that the reaction rate of lipase catalysed transacylation of methyl
methacrylate varied up to 50 times with the concentration of 1-
propanol. The combination of enzyme rational design and reaction
media engineering can thus lead to additional improvements.
[6] U.T. Bornscheuer, K. Buchholz, Eng. Life Sci. 5 (2005) 309–323.
[7] A.B. Hajjar, P.F. Nicks, C.J. Knowles, Biotechnol. Lett. 12 (1990) 825–830.
[8] S. Warwel, G. Steinke, M.R. Klaas, Biotechnol. Tech. 10 (1996) 283–286.
[9] D. Boeckh, B. Hauer, D. Haering, US Patent 2,006,035,341 (A1) (2006).
[10] W. Paulus, B. Hauer, D. Haring, F. Dietsche, US Patent 2,006,030,013 (A1) (2006).
[11] M. Nordblad, P. Adlercreutz, J. Biotechnol. 133 (2008) 127–133.
[12] M. Nordblad, P. Adlercreutz, Biotechnol. Bioeng. 99 (2008) 1518–1524.
[13] M. Martinelle, K. Hult, Biochim. Biophys. Acta, Protein Struct. Mol. Enzymol.
1251 (1995) 191–197.
[14] M.W. Larsen, U.T. Bornscheuer, K. Hult, Protein Expr. Purif. 62 (2008) 90–97.
[15] L. Zheng, U. Baumann, J.L. Reymond, Nucleic Acids Res. 32 (2004) e115.
[16] J.C. Rotticci-Mulder, M. Gustavsson, M. Holmquist, K. Hult, M. Martinelle, Pro-
tein Expr. Purif. 21 (2001) 386–392.
[17] A.O. Magnusson, J.C. Rotticci-Mulder, A. Santagostino, K. Hult, ChemBioChem
6 (2005) 1051–1056.
[18] E. Krieger, T. Darden, S.B. Nabuurs, A. Finkelstein, G. Vriend, Proteins Struct.
Funct. Bioinf. 57 (2004) 678–683.
Acknowledgement
[19] J. Wang, P. Cieplak, P.A. Kollman, J. Comput. Chem. 21 (2000) 1049–1074.
[20] A. Jakalian, D.B. Jack, C.I. Bayly, J. Comput. Chem. 23 (2002) 1623–1641.
[21] U. Essmann, L. Perera, M.L. Berkowitz, T. Darden, H. Lee, L.G. Pedersen, J. Chem.
Phys. 103 (1995) 8577–8593.
BASF (Ludwigshafen, Germany) is greatly acknowledged for
financial support.
[22] T. Tsuji, H. Ito, H. Takeuchi, S. Konaka, J. Mol. Struct. 475 (1999) 55–63.
[23] C.-H. Hu, T. Brinck, K. Hult, Int. J. Quantum Chem. 69 (1998) 89–103.
[24] A. Fersht, Structure and Mechanism in Protein Science, first ed., W.H. Freeman
and Company, USA, 1999.
References
[1] T. Ohrui, Y. Sakakibara, Y. Aono, M. Kato, H. Takao, T. Kawaguchi, US Patent
3,875,212 (1975).
[25] P-O. Syrén, K. Hult, Manuscript.
[26] A. Warshel, J. Biol. Chem. 273 (1998) 27035–27038.
[2] N. Andoh, I. Nishiwaki, M. Arakawa, US Patent 4,329,492 (1982).
[3] A.J.J. Straathof, S. Sie, T.T. Franco, L.A.M. van der Wielen, Appl. Microbiol.
Biotechnol. 67 (2005) 727–734.
[4] P.D. Bloom, P. Venkitasubramanian, US Patent Application 20,090,018,300
(2009).
[27] A. Warshel, J. Florián, M. Strajbl, J. Villà, ChemBioChem 2 (2001) 109–111.
[28] S. Hur, T.C. Bruice, Proc. Natl. Acad. Sci. U.S.A. 99 (2002) 9668–9673.
[29] T.S. Wong, U. Schwaneberg, R. Sturmer, B. Hauer, M. Breuer, Comb. Chem. High
Throughput Screen. 9 (2006) 289–293.
[30] R.J. Fox, M.D. Clay, Trends Biotechnol. 27 (2009) 137–140.
[5] H.E. Schoemaker, D. Mink, M.G. Wubbolts, Science 299 (2003) 1694–1697.