Immobilization of feruloyl esterases in mesoporous materials leads to improved transesterification yield

Article abstract of DOI:10.1016/j.molcatb.2011.05.002

Feruloyl esterases are used in biocatalysis for refinement of hydroxycinnamic acids, a group of compounds with antioxidant and antibacterial properties where modification of solubility is necessary for the compounds to be of interest in different commercial products. In industrially feasible and efficient processes, immobilization of enzymes is often required for sufficient enzyme stability and to enable recovery. In recent years, mesoporous materials have become popular as immobilization support due to advantages such as high protein loading capacity and enhanced enzyme activity because of confinement into pores. We used mesoporous silica, for the first time, as immobilization support for feruloyl esterases. The crude enzyme preparation Depol 740L was adsorbed into two SBA-15 mesoporous silica materials of different pore size and the effects of the immobilization on transesterification of methyl ferulate with 1-butanol into butyl ferulate were studied, tested in a reaction system based on 92.5% 1-butanol and 7.5% MOPS buffer (pH 6.0). Immobilization in mesoporous silica with larger pore size (9 nm) showed higher protein loading and higher specific activity compared to immobilization with smaller pore size (5 nm). Importantly, adsorption into mesoporous silica changed the product specificity of the enzymes to favor transesterification and decrease the rate of hydrolysis compared to free enzymes. The immobilized enzyme had a butyl ferulate yield of up to 90%, significantly higher compared to free enzymes. Additionally, the immobilized enzymes showed an excellent operational stability and reusability, retaining ≥70% of the initial activity after 6 sequential runs, each lasting 6 days. Consequently, we show that mesoporous silica is a robust immobilization support for feruloyl esterases to be used in the development of biocatalysts for customization of the antioxidant properties of hydroxycinnamic acids.

Full text of DOI:10.1016/j.molcatb.2011.05.002

Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Immobilization of feruloyl esterases in mesoporous materials leads to improved
transesterification yield
a
b
a,∗
Christian Thörn , Hanna Gustafsson , Lisbeth Olsson
a
Chalmers University of Technology, Department of Chemical and Biological Engineering, Industrial Biotechnology, SE-412 96 Gothenburg, Sweden
Chalmers University of Technology, Department of Chemical and Biological Engineering, Applied Surface Chemistry, SE-412 96 Gothenburg, Sweden
b
a r t i c l e i n f o
a b s t r a c t
Article history:
Feruloyl esterases are used in biocatalysis for refinement of hydroxycinnamic acids, a group of com-
pounds with antioxidant and antibacterial properties where modification of solubility is necessary for
the compounds to be of interest in different commercial products. In industrially feasible and efficient
processes, immobilization of enzymes is often required for sufficient enzyme stability and to enable
recovery. In recent years, mesoporous materials have become popular as immobilization support due to
advantages such as high protein loading capacity and enhanced enzyme activity because of confinement
into pores. We used mesoporous silica, for the first time, as immobilization support for feruloyl esterases.
The crude enzyme preparation Depol 740L was adsorbed into two SBA-15 mesoporous silica materials
of different pore size and the effects of the immobilization on transesterification of methyl ferulate with
1-butanol into butyl ferulate were studied, tested in a reaction system based on 92.5% 1-butanol and
7.5% MOPS buffer (pH 6.0). Immobilization in mesoporous silica with larger pore size (9 nm) showed
higher protein loading and higher specific activity compared to immobilization with smaller pore size
Received 21 December 2010
Received in revised form 23 March 2011
Accepted 2 May 2011
Available online 8 May 2011
Keywords:
Feruloyl esterase
Mesoporous materials
Enzymes
Immobilization
Biocatalysis
(5 nm). Importantly, adsorption into mesoporous silica changed the product specificity of the enzymes
to favor transesterification and decrease the rate of hydrolysis compared to free enzymes. The immo-
bilized enzyme had a butyl ferulate yield of up to 90%, significantly higher compared to free enzymes.
Additionally, the immobilized enzymes showed an excellent operational stability and reusability, retain-
ing ≥70% of the initial activity after 6 sequential runs, each lasting 6 days. Consequently, we show that
mesoporous silica is a robust immobilization support for feruloyl esterases to be used in the development
of biocatalysts for customization of the antioxidant properties of hydroxycinnamic acids.
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
Applications of FAEs have been found within biofuel production,
agriculture and pulp and paper industry [3]. Hydroxycinnamic
acids have been shown to have tumor suppressing [4], antioxi-
dant [5] and antibacterial properties [6]. To find applications in
cosmetics, food products or in therapeutics it is often desirable
to change properties of the hydroxycinnamic acids. The carboxylic
acid group can be used, through esterification, both to add a func-
tional group or to change the solubility by adding a group that
makes the whole molecule more hydrophobic/hydrophilic depend-
ing on the application. It is important to have biosynthetic tools
available since the esterification reactions with hydroxycinnamic
acids are difficult to perform effectively with classical chemi-
cal synthesis [3]. Consequently, there is an interest in utilizing
FAEs as biosynthetic tools for modification of hydroxycinnamic
acids. There are a few examples of FAEs used for esterification
Feruloyl esterases (FAE), is a subclass (E.C. 3.1.1.73) of car-
boxylic ester hydrolases that catalyze the hydrolysis of ester
linkages in plant cell wall materials, releasing ferulic acid (FA)
and other hydroxycinnamic acids. Feruloyl esterases show a great
diversity in the substrate specificity which lead to a first classi-
fication based on substrate specificity on the hydrolysis of four
synthetic hydroxycinnamic acid methyl esters [1]. More recently
a new classification based on sequence descriptors has been
suggested to better describe this diverse group of enzymes [2].
Abbreviations: BFA, butyl ferulate; CLEA, cross-linked enzyme aggregate; FA,
ferulic acid; FAE, feruloyl esterase; MFA, methyl ferulate; MPS, mesoporous silica.
(
and transesterification) of hydroxycinnamic acids focusing on
different model reactions such as: first and secondary alcohols
7,8], glycerol [9,10], saccharides [11,12]. Most of the examples
use microemulsions as reaction system both for solubilizing the
∗ Corresponding author. Tel.: +46 31 772 3805; fax: +46 31 772 3801.
E-mail addresses: christian.thorn@chalmers.se (C. Thörn),
hanna.gustafsson@chalmers.se (H. Gustafsson),
[
lisbeth.olsson@chalmers.se (L. Olsson).
1
381-1177/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2011.05.002

5
8
C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
substrate and hosting the enzymes. When microemulsions are
used, the composition of the solution is restricted to a narrow
range that allows formation of the microemulsion. There are
also examples of lipases active on hydroxycinnamic acids [13],
though the yield when running esterification with ferulic acid
is generally low due to specific residues on the benzene ring
O
O
CH3OH
A
HO
[
14–16].
Low water content prevents unwanted hydrolysis reactions and
O
O
O
creates favorable conditions for esterification or transesterification
reactions. Therefore, many biosynthetic reactions using esterases
or lipases must be run in organic solvents. Substrate solubiliza-
tion and diffusion properties are also important factors that limit
the choices of solvents. However, enzymes will not be as active in
organic solvents as in aqueous solutions, since the organic solvent is
unnatural to enzymes and can cause denaturation [17]. Therefore,
enzyme stability and reusability needs attention and immobiliza-
tion of enzymes is a way of dealing with this issue. Immobilization
can also be a mean of controlling the enzyme activity and selectiv-
ity. To the best of our knowledge, FAEs have only been immobilized
using cross-linked enzyme aggregates (CLEA) [10,18,19]. However,
the reusability of FAE CLEAs (both pure FAEs and crude enzyme
preparation with FAE activity) has been limited, since the enzymes
lost all their activity after being reused 2–5 times (depending on
the reaction conditions).
During the last years, immobilization of enzymes into meso-
porous materials and in particular mesoporous silica (MPS) has
become increasingly popular. Recent reviews point out that meso-
porous materials (mostly commonly mesoporous silica SBA-15
or MCM-41) allow high protein adsorption capacity, enhanced
enzyme stability, good diffusion properties and easy recovery in
immobilization applications [20–22]. The support material itself
has great structural stability in a broad range of pH and tem-
perature. There are extensive possibilities for functionalization
of the surface of the material and the pore size is tunable over
a range (2–50 nm) that matches the sizes of most enzymes.
Additionally, the confining of the enzymes into the mesoporous
material can increase the stability of the enzyme, as well as
altering the enzymatic activity compared to free enzymes in the
bulk solution. Confining of enzymes has also been shown to
create protein–surface interactions that alter enzymatic proper-
ties and give rise to a specific activity significantly higher than
that of free enzyme [23]. Suggested mechanisms include that the
active site is stabilized in a more active state or that the packed
environment inside the pores resembles that inside a cell. In
summary, mesoporous materials as immobilization support have,
through enzyme confinement and surface functionalization, the
potential to improve control over the enzymatic activity with
regards to substrate specificity and fulfill most requirements of
a functional biocatalyst. However, the main limitation is still that
enzymes immobilized into mesoporous materials through physi-
cal adsorption are prone to leaching. The most common strategies
to overcome this problem involve covalent binding of the enzyme
to the surface or partial closing of the pore entrances which, how-
ever, adds complexity and overall cost to the biocatalyst production
OH
O
H2O
HO
C
OH
O
B
OH
H2O
CH OH
3
O
OH
Fig. 1. (A) Transesterification of MFA with 1-butanol generating BFA and methanol.
(B) Hydrolysis of MFA generating ferulic acid and methanol (natural reaction at high
water contents). (C) Esterification of FA with 1-butanol generating BFA and water.
pound. The activity and reusability of the immobilized enzymes
were studied. Additionally, product selectivity of the immobilized
enzymes was compared with that of free enzyme, with evalua-
tion of how the immobilization alters the preference for the side
reaction (hydrolysis) that is inherent when working with FAEs
(Fig. 1B).
2. Experimental
2.1. Materials
Ferulic acid (FA), methyl ferulate (MFA) and the other hydrox-
ycinnamic acids and methyl esters were purchased from Apin
Chemicals Ltd., UK. MOPS was purchased from Amresco Inc., USA.
1-Butanol (Cat. No. 101990) was purchased from MERCK. Methanol
(Cat. No. 34860) was purchased from Sigma–Aldrich. Bradford
reagent was purchased from Bio-Rad Laboratories. Butyl ferulate
(BFA) was provided as a gift by Evangelos Topakas, National Tech-
nical University of Athens, Greece.
2.2. Enzymes
Depol 740L, the crude enzyme preparation used in this study
was provided as a free sample from Biocatalysts Ltd., UK. Depol
740L is a multi-componentpreparation shown to have several other
enzymatic activities [19], i.e. out of the total protein concentration
measured in this study, only a few percent is likely to be feru-
loyl esterases. The control experiments with free enzyme was done
using Depol 740L, that was washed several times with MOPS buffer
using centrifugal filters (Amicon Ultra – 0.5 ml 10 K ultracel mem-
brane). No leakage of active enzymes could be detected using the
centrifugal filters. Free enzyme reactions were stopped by rapidly
[
24].
Until now, immobilization of FAEs has not been tested in
mesoporous materials. Therefore, this work aims to explore immo-
bilization of FAE into MPS and study how the activity is affected
by confining the enzymes into silica nanopores. A crude enzyme
preparation (Depol 740L, Biocatalysts Ltd., UK) was selected,
shown to have FAE activity and a high total protein concentra-
tion [19]. The immobilization was performed into MPS, of two
different pore sizes, through physical adsorption. We chose a
model reaction for the transesterification of hydroxycinnamic acid
esters: transesterification of methyl ferulate with 1-butanol into
butyl ferulate (Fig. 1A), that results in a more hydrophobic com-
◦
putting the sample at −20 C.
2.3. Enzyme assay – substrate profile
The substrate profile assay for the hydrolysis of the 4 synthetic
hydroxycinnamic acid methyl esters, was adapted from previous

C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
59
Table 1
publications [19,25]. In short, substrate solution was prepared by
individually dissolving the esters, MFA (15 mM), methyl sinapate
MSA; 15 mM), methyl p-coumarate (MpCA; 5 mM) and methyl caf-
feate (MCA; 15 mM), in MOPS buffer (100 mM, pH 6.0) and DMSO
10% v/v). The reaction was started by mixing 100 ␮l enzyme solu-
Material properties characterized by nitrogen-adsorption.
(
Sample
Pore diameter
nm)
BET surface
area (m /g)
Pore volume
2
3
(
(cm /g)
(
MPS-5 (SBA-15)
MPS-9 (SBA-15)
5.01
9.32
924
528
0.73
1.17
tion with 300 ␮l substrate-buffer in an eppendorf tube that was put
◦
onto a heating block and incubated at 50 C for 5 min. Reaction was
stopped with 400 ␮l ice-cold 10% acetic acid (FA or MFA stability
◦
2× 30 min. After adding the enzyme preparation (7.2 ml Depol
are not affected by the acetic acid) and the tube stored at −20 C
◦
7
40L/g MPS) the solution was magnetically stirred at 4 C for 22 h.
until analysis.
Thereafter, the MPS was washed (10 min centrifugation at 4000 × g)
3
times with 10 ml MOPS buffer. The amount of immobilized
2.4. Quantification of ferulic acid and its esters
enzyme was estimated by removing the MPS by centrifugation and
measuring the residual total protein concentration of the super-
natant with a Bradford assay, using BSA as standard [28]. The final
pellet was dried in vacuum (CHRIST RCV 2-18) without added
heating. The obtained powder was used in the transesterification
experiments.
FA, MFA and BFA concentrations were determined by HPLC
(
Dionex–Ultimate 3000) using a Kinetex 2.6 u C18 100 × 4.6 mm
column. Absorbance was measured at 300 nm with a PDA detec-
tor. Samples were diluted 2–10 times in methanol:dH O (1:1 v/v)
2
prior to analysis. Elution was done with methanol:dH O (7:3 v/v)
2
at flow rate of 1 ml/min. Calibration curves was established by
running FA, MFA and BFA solutions at known concentrations
2.8. Transesterification of MFA with 1-butanol into BFA
(
0.01–2 mM). Transesterification yield was calculated as the molar
In the transesterification of MFA with 1-butanol, MFA (20 mM)
amount of generated BFA compared to the initial amount of MFA
and expressed as percentage. Unit (U) was defined as amount of
enzyme transforming 1 ␮mol MFA to BFA/min. The product selec-
tivity was defined by the BFA/FA-molar ratio (the concentration of
produced BFA divided by the concentration of produced FA). Tests
were done to ensure that FA, MFA and BFA were fully stable (data
not shown) for at least one week in both the methanol (at 4 C) and
butanol (37 C) solutions used in the experiments. The sum of the
molar amounts of all reactants at the end of the reaction was always
within a 10% error margin compared to the starting molar amount
of the substrate.
was dissolved in 1-butanol:MOPS buffer solutions with solvent
ratios of 100/0–85/15% (v/v). The reaction was started by adding
MPS particles with immobilized enzyme (10 mg/ml reaction solu-
tion) to the reaction solution in an Eppendorf tube which was put
◦
onto a shaking incubator (IKA KS 4000, 37 C, 250 rpm). The reaction
◦
was ended by removing the MPS particles by centrifugation (4 min,
◦
18,000 × g) and transferring the supernatant to a new tube, stored
◦
at −20 C until analysis. Control experiments without enzyme were
done to ensure that FA and MFA do not adsorb onto the synthe-
sized MPS nor that their stability is affected by the presence of
MPS.
2.5. Synthesis of mesoporous silica
2.9. Reusability/stability experiments
SBA-15 of two different pore sizes was synthesized using pro-
The reusability experiments were run as the transesterification
tocols adapted from Zhao et al. [26,27]. The triblock copolymer
Pluronic P123 (EO20PO70EO20) was used as a structure directing
agent and tetraethyl orthosilicate (TEOS) as a silica source. In a
typical synthesis, 4.0 g of P123 was dissolved in 120 g of 2 M HCl
experiments described above but after the reaction was stopped,
the pellet was dried in vacuum and the reaction restarted with fresh
reaction solution (repeated throughout each run). No washing was
done in between the runs to minimize potential enzyme leakage.
Since it is difficult to separate free enzymes from a solution, the
control samples with free enzyme were pre-incubated without sub-
strate (for 3, 6, 12, 19, 26, 33 days) in 750 ␮l reaction solution buffer
◦
and 30 g of deionized water and vigorously stirred at 35 C for 2 h.
◦
8
.5 g of TEOS was added and the solution was stirred at 35 C for
additional 24 h. The gel mixture was transferred to stainless steel
pressure autoclaves in Teflon containers and aged for 24 h at 80 or
◦
◦
at 37 C and started by adding 250 ␮l 80 mM MFA 1-butanol:MOPS
1
50 C, depending on the desired pore diameter. The solid precip-
buffer at the given time points.
itate was recovered by vacuum filtration, washed with deionized
water and dried. Finally, to remove the template, the product was
calcinated by increasing the temperature from room temperature
3. Results and discussion
◦
◦
up to 500 C during 8 h followed by heating at 500 C for another
6
h.
3.1. Simple immobilization of Depol 740L onto MPS
2
.6. Characterization of mesoporous silica
During biocatalysis, enzymes are often subject to unnatural con-
ditions and immobilization is thereby a mean of protecting and
enhancing the reusability of the enzymes. It is also known that
immobilization can alter both stability and activity of enzymes. In
this study, FAEs were immobilized into MPS for the transesterifica-
tion of MFA with 1-butanol. The crude enzyme preparation Depol
740L, was immobilized into SBA-15 mesoporous silica through
physical adsorption. In order to evaluate how the pore size affect
the protein loading capacity and the activity of the immobilized
enzymes MPS of two different pore sizes (5 nm, MPS-5 and 9 nm,
MPS-9; Table 1) were selected.
The pore diameter and surface areas were measured by a Micro-
metrics Tristar. The pore diameter distributions were determined
using the BJH (Barrett–Joyner–Halenda) methods based on the des-
orption isotherms, and the surface area were determined using the
Brunauer–Emmett–Teller (BET) method (material characteristics
are summarized in Table 1). The hexagonal structure of SBA-15 was
confirmed by TEM imaging on a JEM-1200 EX II TEM operated at
1
20 kV.
2
.7. Immobilization of enzymes
Immobilization into MPS was followed over time to determine
the time frame required for adsorption to get close to equilibrium
and no more protein being adsorbed. The immobilization was rapid
with most protein being immobilized during the first 10 min (Fig. 2).
Mesoporous silica and MOPS buffer (20 ml buffer/g MPS) was
mixed in a Falcon tube and sonicated (Elmasonic S30H) for

6
0
C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
4
3
2
1
0
100
80
60
40
20
0
0
50
100
150
200
0
5
10
15
20
25
Time (h)
Time (h)
Fig. 3. Transesterification of MFA (20 mM) into BFA over time in 7.5% water for (᭹)
MPS-5D, (ꢀ) MPS-9D, (ꢀ) free enzyme (amount of free enzyme corresponds to the
amount immobilized into MPS-5D). Data shown is the average of 2 replicates.
Fig. 2. Residual protein concentration throughout the immobilization. (᭹) MPS-5D,
(ꢀ) MPS-9D. The first data point was measured before adding MPS. The last data
point is before the washing step.
However, the immobilization was run for 22 h since the time scale
of protein rearrangement onto surfaces can be in order of hours
resulted in a significantly higher transesterification yield for MPS-
9D compared to MPS-5D (Fig. 3). The FAEs in MPS-9D are also 3.5
times more active than MPS-5D in terms of specific transesterifi-
cation activity (Table 2). Assuming that most proteins are located
on the surface outside of MPS-5D (as discussed in Section 3.1), one
could speculate that it would be less crowded for MPS-9D where
a larger proportion is located inside the pores. More space around
the enzyme seems to be favorable for the specific enzymatic activ-
ity and the transesterification activity detected for MPS-9D is also
higher per amount of support material. Enhanced enzymatic activ-
ity after immobilization into MPS, has previously been associated
with unpredictable conformational changes in the enzyme struc-
ture caused by confining the enzyme [32]. The mechanisms behind
the changes in activity are not very well understood and consider-
able amount of effort is put into getting a better understanding on
how confining can be used in a rational way to control and direct
the enzymatic activity [33].
Comparing the initial activity of MPS-9D with the one of the free
enzyme, 40% of the specific BFA activity was lost (Table 2). However,
MPS-9D supports a significantly higher overall BFA yield up to 90%
and does not have the same tendency of declining yield in the end of
the reaction (Fig. 3). Additionally, if the specific transesterification
activity would have been calculated at one of the later time points,
the advantage of running the transesterification with immobi-
lized enzymes would be even clearer. Previous publications report
conversions up to 78–97% for esterification/transesterification of
hydroxycinnamic butyl esters [18,19]. In these reports, conversion
is defined as the amount of substrate reacted compared to the initial
amount of substrate and therefore includes any amount of gener-
ated hydrolysis product. The corresponding conversion for MPS-9D
is 95%.
[
29].
Since Depol 740L contains a mixture of proteins, a substrate
profile (for the hydrolysis of four synthetic hydroxycinnamic acid
methyl esters) was performed on the supernatant of the immobi-
lization reaction after 22 h. No distinct change in substrate profile
was observed in the residual protein solution using MPS-5. This
indicates that no selective immobilization of potential FAE isoen-
zymes occurred (only data for MFA hydrolysis is shown in Table 2).
Furthermore, the proportion of immobilized proteins (39% before
washing) is reasonably correlated with the proportion of immobi-
lized FAEs (45% MFA activity lost). For MPS-9 no hydrolytic activity
could be detected in the supernatant for MSA nor MCA and only
a few percent activity was left for MFA and MpCA, whereas the
protein measurement showed 83% immobilized before washing.
The low residual activity indicates that a majority of the feruloyl
esterases in Depol 740L were immobilized into MPS-9. Since the
total protein immobilized was lower compared to removed activity,
the immobilization was partially selective towards FAEs (assuming
free enzymes are not inactivated during the immobilization).
MPS-9 was able to support twice the amount of protein com-
pared to MPS-5 (Table 2). However, if considering the pore volume
(
Table 1), MPS-9 should only support 60% more protein than MPS-5.
Also considering that the surface area is almost a factor 2 larger for
MPS-5, it is likely that pore size is the limiting factor for immobiliza-
tion in MPS-5 and most of the proteins are located on outside the
pores. Low protein loading after immobilization, compared to the
large available surface area and pore volume, has previously been
linked to the fact that MPS with smaller pore sizes are more prone
to larger proteins getting stuck in the pore entrance, subsequently
blocking the whole pore [30]. It is also believed the density of
adsorbed proteins is higher close to the pore entrance, making uti-
lization of the whole pore volume difficult [31]. Nevertheless, using
a multi-component preparation and without information regard-
ing size or dimensions of the proteins in Depol 740L, it is difficult
to fully interpret the immobilization degree.
Esterification of FA directly to BFA was also tested, but the BFA
yield of this reaction was just above the detections limit (data not
shown). This result rules out the possibility that MFA would be
initially hydrolyzed to FA and then in a second step esterified to
BFA (Fig. 1B and C).
3.3. Effect of water content on transesterification activity
3.2. Transesterification yield over time higher with larger pore
size
The amount of water present in the reaction solution is known
to influence the activity of enzymes, especially in organic solvents
where a small amount of water usually is required for the enzymes
to be active [34]. Therefore, the transesterification reaction was
Initially, the transesterification activity of the immobilized
enzyme particles was evaluated and measured over time and

C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
61
8
6
4
2
0
0
0
0
0
A
BFA
MFA
FA
0
5
10
15
Retention time
Water content (% v/v)
Fig. 5. Typical HPLC chromatogram illustrating the higher BFA/FA molar ratio of
immobilized enzyme (solid line) compared to free enzyme (dashed line). Retention
times for FA, MFA and BFA were 1.0, 1.5 and 3.0 min, respectively.
80
B
6
4
2
0
0
0
0
generally run at water contents around 2% for optimal esterification
activity [25].
In similar transesterification experiments but with free enzyme,
the profile turns out noticeably different with a minimum BFA
yield at 5% water content (Fig. 4B). Generally, increased water
content should lower the transesterification activity and increase
the hydrolytic activity [11]. Additionally, decreased transesterifi-
cation activity with increase water content has been shown for
lipases in a broad range of organic solvents [35]. However, we
observed that the transesterification yield increased at water con-
tents between 5 and 15%, which might be related to an increase
in the hydration of the enzyme. Unexpectedly, as the activity
increases, the increase in hydrolytic activity is lower relative to
the increase in transesterification activity (further discussed in
Section 3.4). Turning the attention to the immobilized enzyme, it
is generally believed that hydration of enzymes is closely linked
to the enzymatic activity and it has also been shown that pro-
teins inside mesoporous silica are more strongly hydrated [36,37].
Consequently, the immobilized enzymes can obtain the neces-
sary hydration at a lower water content, which would explain
why the transesterification activity has no minimum at water
content 5%.
0
5
10
15
Water content (% v/v)
Fig. 4. Transesterification of MFA to BFA tested at varying water content using
(
(
A) (᭹) MPS-5D (136 h) and (ꢀ) MPS-9D (22 h). (B) Free enzyme control, after
ꢀ) 46 h and (ꢁ) 168 h. At water contents >10% the water:1-butanol solution
is not fully miscible. The indicated times represents the length of the reaction
and a shorter time was selected for MPS-9D to be more comparable with MPS-
5
bars.
D. Data in both graphs are averages of 3 replicates with standard deviation
tested at varying water (MOPS buffer) content. Both the MPS-5D
and MPS-9D have a similar BFA yield profile (Fig. 4A) with low
yield at low water content (0–2.5%) and relatively stable yield
as the water content increases (≥5%). Additionally, MPS-9D con-
sistently has a higher yield compared to MPS-5D at the tested
water contents, which further confirms that the larger pore size
is indeed favorable for the specific transesterification activity. Pre-
vious biosynthetic experiments using FAEs in microemulsions were
3.4. Changed product selectivity
Naturally FAE hydrolyzes the ester bond to yield FA and an
alcohol. However, in this study, it was observed that the immobi-
lized enzymes consistently generated less FA (higher BFA/FA ratio)
than the free enzymes (Fig. 5). This difference was most evident
Table 2
Degree of immobilization based on residual protein concentration.
Protein immobilized (%)
Lost MFA
activity (%)
Protein load (mg/gMPS)
Specific
transesterification
activity (mU/mgprotein)
Transesterification
activity/support
material (mU/gMPS)
a
b
Before wash.
After wash.
MPS-5D (MPS-5)
MPS-9D (MPS-9)
Non-porous Si
Free enzyme
39
83
31
68
29
n.r.
45
94
32
n.r.
34
73
31
n.r.
2.2
7.8
9.3
13
0.07
0.56
0.29
n.r.
36
n.r.c
a
MFA hydrolysis activity on the residual supernatant after immobilization (before washing).
Transesterification activities are derived from the first time points of Fig. 3.
Not relevant.
b
c

6
2
C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
8
6
4
2
0
0
0
0
0
0
0
0
0
0
1
00
75
50
A
A
25
0
1
st
2nd
3rd
4th
5th
6th
0
2
4
6
8
10
12
14
16
#
run
Water content (v/v %)
8
6
4
2
B
1
00
B
7
5
2
5
0
5
0
0
5
10
15
20
25
30
35
0
50
100
150
200
Time (days)
Time (h)
Fig. 7. Reuseability experiments with 7.5% (v/v) water using (A) immobilized
enzyme with each run lasting 6 days, (black) MPS-5D, (white) MPS-9D and non-
porous silica (grey). (B) Free enzyme. 100% activity is defined as the activity of the
first run/first time point. Data shown are averages of 4 replicates with standard
deviation bars.
Fig. 6. BFA/FA molar ratio (A) at varying water content for (᭹) MPS-5D, (ꢀ) MPS-
D, free enzyme (ꢀ) after 46 h and (ꢁ) after 168 h. (B) Different time points for (᭹)
MPS-5D, (ꢀ) MPS-9D, free enzyme (ꢀ) at 7.5% (v/v) water. The first two data points
of MPS-5D were below the detection limit. Data in both graphs are averages of 2–3
replicates.
9
inside of the pores is significantly different to that of the sur-
rounding bulk. However, since there is no difference in BFA/FA
ratio between MPS-5D and MPS-9D, it is more likely that the effect
is a general feature of the enzyme being close to a silica surface
rather than a confining effect of the pore. This hypothesis was
tested by immobilizing Depol 740L onto non-porous silica parti-
cles. The BFA/FA ratio using the non-porous particles was the same
as for MPS-9D at a water content of 7.5% (data not shown). Still,
it remains to determine if the change in BFA/FA ratio is true in
other solvents and also if it can be generalized to other transester-
ification reactions using FAEs. Such studies would aid in deducing
whether the change in ratio comes from potential changes in the
structure of the enzyme due to interactions with the silica sur-
face or if the effect originates from water and butanol molecules
being arranged favorably both inside the pores and on the outer
surface of the particles. One should also consider that it can be a
challenge to draw conclusions when using a crude enzyme prepa-
ration since there can be effects of having other proteins and
other active enzymes present. Using pure enzyme, it would also
be more relevant to assess parameters such as pH, temperature
and surface modifications that could enhance both the immobiliza-
tion but also potentially allow increased control of the enzymatic
activity.
at water contents of 5–10% where the ratio was 4–9 times higher
for immobilized enzymes compared to free enzymes (Fig. 6A). The
BFA/FA ratios in the reactions with MPS-5D and MPS-9D were
closely aligned at the different water contents. Thus the higher
BFA/FA ratio is independent of pore size and more likely related
to the enzyme being adsorbed to the silica surface. The BFA/FA
ratio showed to be relatively stable over time (Fig. 6B). However,
MPS-9D showed a slow decrease in ratio at higher yields. A possi-
ble explanation for this could be that at higher yields, where the
transesterification has slowed down due to substrate limitation,
FA can still be generated through hydrolysis of BFA. This could also
explain the decrease in yield for the free enzyme (Fig. 3) in the
end of the reaction, where selectivity is changed towards hydroly-
sis.
It has been reported that short alcohols (ethanol and butanol)
interact with silanol groups in a combination of hydrogen bonds
and van der Waals interactions that allow the alcohol to form
chain like structures inside the nanochannels of mesoporous silica
(
3.1 nm in diameter) [38]. A previous publication also demonstrates
that water is adsorbed/arranged along silica surfaces in SBA-15
through hydrogen bonding [39]. Furthermore, the arrangement of
water molecules inside the pore changes the stability and move-
ment dynamics of the proteins [40]. Therefore, the environment

C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
63
3.5. Reusable and stable biocatalyst
References
[
[
1] V.F. Crepin, C.B. Faulds, I.F. Connerton, Appl. Microbiol. Biotechnol. 63 (2004)
47–652.
2] D.B.R.K.G. Udatha, I. Kouskoumvekaki, L. Olsson, G. Panagiotou, Biotechnol. Adv.
(2010), doi:10.1016/j.biotechadv.2010.09.003.
The MPS with immobilized Depol 740L were relatively stable
6
throughout 6 runs (Fig. 7A), retaining 70% and 96% of their activi-
ties compared to the first run for MPS-5D and MPS-9D, respectively.
The better reusability of MPS-9D can partly be explained by the fact
that the reaction had reached a BFA yield (∼85%), where a decrease
in activity will not be as easily noticeable as for MPS-5D with a sig-
nificantly lower yield (15–20%). This could also explain the relative
high reusability of the non-porous silica (∼70% in yield) where the
proteins also could be assumed to have adsorbed more strongly
due to a larger available surface area per protein, caused by a sig-
nificantly smaller particles size of the non-porous silica compare to
MPS-5D. Nevertheless, another explanation to the observed differ-
ence in reusability between the two mesoporous silicas, can be that
the enzymes in MPS-9D are conformationally stabilized (assum-
ing a majority of the enzymes are located inside the pores) by
protein–surface interactions with the cylindrical shaped pore wall
confining the enzyme, as described previously for other enzymes
[
[
3] E. Topakas, C. Vafiadi, P. Christakopoulos, Process Biochem. 42 (2007) 497–509.
4] A. Murakami, Y. Nakamura, K. Koshimizu, D. Takahashi, K. Matsumoto, K. Hag-
ihara, H. Taniguchi, E. Nomura, A. Hosoda, T. Tsuno, Y. Maruta, H.W. Kim, K.
Kawabata, H. Ohigashi, Cancer Lett. 180 (2002) 121–129.
5] H. Kikuzaki, M. Hisamoto, K. Hirose, K. Akiyama, H. Taniguchi, J. Agric. Food
Chem. 50 (2002) 2161–2168.
[
[
6] S. Naz, S. Ahmad, S.A. Rasool, S.A. Sayeed, R. Siddiqi, Microbiol. Res. 161 (2006)
43–48.
[7] S. Giuliani, C. Piana, L. Setti, A. Hochkoeppler, P.G. Pifferi, G. Williamson, C.B.
Faulds, Biotechnol. Lett. 23 (2001) 325–330.
[
[
8] N.S. Hatzakis, I. Smonou, Bioorg. Chem. 33 (2005) 325–337.
9] M. Tsuchiyama, T. Sakamoto, T. Fujita, S. Murata, H. Kawasaki, Biochim. Biophys.
Acta 1760 (2006) 1071–1079.
[
[
[
[
[
[
[
[
10] C. Vafiadi, E. Topakas, V.R. Nahmias, C.B. Faulds, P. Christakopoulos, J. Biotech-
nol. 139 (2009) 124–129.
11] E. Topakas, C. Vafiadi, H. Stamatis, P. Christakopoulos, Enzyme Microb. Technol.
36 (2005) 729–736.
12] C. Vafiadi, E. Topakas, P. Christakopoulos, Carbohydr. Res. 341 (2006)
1992–1997.
[
37,41]. Immobilization of Depol 740L in MPS also considerably
13] N.T.L. Chigorimbo-Murefu, S. Riva, S.G. Burton, J. Mol. Catal. B: Enzym. 56 (2009)
277–282.
14] B. Guyot, B. Bosquette, M. Pina, J. Graille, Biotechnol. Lett. 19 (1997)
improves the reusability compared to CLEAs of Depol 740L where
all enzymatic activity was lost after 2 runs (6 days/run in a n-
hexane/1-butanol/water micro emulsion) [19].
Control experiments were done where the reactions were
stopped after 3 days and the supernatant was incubated further
529–532.
15] H. Stamatis, V. Sereti, F.N. Kolisis, J. Mol. Catal. B: Enzym. 11 (2001)
323–328.
16] M.H. Katsoura, A.C. Polydera, L.D. Tsironis, M.P. Petraki, S.K. Rajacic, A.D. Tse-
lepis, H. Stamatis, New Biotechnol. 26 (2009) 83–91.
17] A.M. Klibanov, Nature 409 (2001) 241–246.
3
days longer. No activity could be detected (data not shown) in
the supernatant indicating that the slow decrease in activity of the
immobilized enzymes is mainly due to inactivation of the enzyme
and not leakage. It also suggests that the interaction to the silica
surface through physical adsorption (no covalent linkages) is suf-
ficient for a stable immobilization under the conditions tested in
this study. Immobilization in MPS through adsorption are gener-
ally prone to leaching and usually require either covalent linkage
of the enzyme or partial closure of the pore opening for sufficient
enzyme recovery [22]. However, the free enzyme control is also
quite stable in the 1-butanol:MOPS solution, retaining 75% of its
initial activity after more than 30 days pre-incubation without sub-
strate (Fig. 7B). Nevertheless, MPS should be considered as suitable
support material for immobilization of FAEs with a simple immo-
bilization protocol and satisfying reusability.
[18] C. Vafiadi, E. Topakas, A. Alissandratos, C.B. Faulds, P. Christakopoulos, J.
Biotechnol. 133 (2008) 497–504.
[
19] C. Vafiadi, E. Topakas, P. Christakopoulos, J. Mol. Catal. B: Enzym. 54 (2008)
5–41.
20] S. Hudson, J. Cooney, E. Magner, Angew. Chem. Int. Ed. 47 (2008) 8582–8594.
3
[
[21] C. Ispas, I. Sokolov, S. Andreescu, Anal. Bioanal. Chem. 393 (2009) 543–554.
[
[
22] C.H. Lee, T.S. Lin, C.Y. Mou, Nano Today 4 (2009) 165–179.
23] C.H. Lei, Y. Shin, J.K. Magnuson, G. Fryxell, L.L. Lasure, D.C. Elliott, J. Liu, E.J.
Ackerman, Nanotechnology 17 (2006) 5531–5538.
[24] M. Hartmann, D. Jung, J. Mater. Chem. 20 (2010) 844–857.
[
[
[
[
25] E. Topakas, H. Stamatis, M. Mastihobova, P. Biely, D. Kekos, B.J. Macris, P. Chris-
takopoulos, Enzyme Microb. Technol. 33 (2003) 729–737.
26] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka, G.D.
Stucky, Science 279 (1998) 548–552.
27] D.Y. Zhao, Q.S. Huo, J.L. Feng, B.F. Chmelka, G.D. Stucky, J. Am. Chem. Soc. 120
(
1998) 6024–6036.
28] M.M. Bradford, Anal. Biochem. 72 (1976) 248–254.
[29] A.M. Moulin, S.J. O’Shea, R.A. Badley, P. Doyle, M.E. Welland, Langmuir 15 (1999)
776–8779.
30] H. Takahashi, B. Li, T. Sasaki, C. Miyazaki, T. Kajino, S. Inagaki, Chem. Mater. 12
2000) 3301–3305.
31] M. Hartmann, Chem. Mater. 17 (2005) 4577–4593.
8
[
[
4
. Conclusion
(
MPS proves to be a robust immobilization support for Depol
40L due to good reusability and stability of the immobilized
[32] C.H. Lei, Y. Shin, J. Liu, E.J. Ackerman, Nano Lett. 7 (2007) 1050–1053.
[33] A.K. Dunker, A. Fernandez, Trends Biotechnol. 25 (2007) 189–190.
[34] A. Zaks, A.M. Klibanov, J. Biol. Chem. 263 (1988) 8017–8021.
[35] L.F. Garcia-Alles, V. Gotor, Biotechnol. Bioeng. 59 (1998) 684–694.
7
enzymes and can thereby be considered as a functional biocata-
lyst for modification of hydroxycinnamic acid esters. The simple
immobilization protocol in combination with a crude enzyme
preparation shows the potential of the biocatalyst in indus-
trial applications. The immobilization leads to altered enzymatic
activity towards transesterification of butyl ferulate while less
hydrolysis product is generated, resulting in higher transesterifica-
tion yield compared to free enzymes. Hopefully, this work can aid
in further developing FAE biocatalysts that generally exhibit high
transesterification activity also in other solvents and for a broad
range of reactions, e.g. hydrophilic modifications with glycerol or
saccharification of other hydroxycinnamic acid esters.
[36] V. Kurkal, R.M. Daniel, J.L. Finney, M. Tehei, R.V. Dunn, J.C. Smith, Biophys. J. 89
2005) 1282–1287.
37] R. Ravindra, Z. Shuang, H. Gies, R. Winter, J. Am. Chem. Soc. 126 (2004)
2224–12225.
(
[
1
[38] T. Kamijo, A. Yamaguchi, S. Suzuki, N. Teramae, T. Itoh, T. Ikeda, J. Phys. Chem.
A 112 (2008) 11535–11542.
[
39] B. Grunberg, T. Emmler, E. Gedat, J. Shenderovich, G.H. Findenegg, H.H. Limbach,
G. Buntkowsky, Chem. Eur. J. 10 (2004) 5689–5696.
[40] G. Schiro, M. Sclafani, F. Natali, A. Cupane, Eur. Biophys. J. Biophys. 37 (2008)
543–549.
[
41] Y. Urabe, T. Shiomi, T. Itoh, A. Kawai, T. Tsunoda, F. Mizukami, K. Sakaguchi,
Chembiochem 8 (2007) 668–674.
Christian Thörn received his M.Sc. in biotechnology engi-
neering at Chalmers University of Technology, Gothen-
burg after doing his thesis on metabolic engineering
in yeast for production of bioactive seleno-compounds,
within the group of Industrial Biotechnology. Thereafter,
he started his Ph.D. project regarding enzyme immobi-
lization in mesoporous materials. The project aim to use
feruloyl esterases for biocatalytic applications and study
how the enzymatic activity can be controlled by the use
of a mesoporous immobilization support.
Acknowledgements
We would like to thank the Swedish research council for funding
this project (349-2007-8680), which is part of the Linnaeus Centre
for Bio-inspired Supramolecular Function and Design–SUPRA. We
also thank Prof. Paul Christakopoulos and Dr. Evangelos Topakas
(
National Technical University of Athens, Greece) for many fruitful
discussions regarding the work in this study.

6
4
C. Thörn et al. / Journal of Molecular Catalysis B: Enzymatic 72 (2011) 57–64
Hanna Gustafsson received her M.Sc. in biotechnol-
Lisbeth Olsson received her Ph.D. degree at University
of Lund in applied microbiology in 1994. In the period
1994–2007, she worked at the Technical University of
Denmark in the field of microbial cell factories and fer-
mentation technology. Since 2008, she is professor in
bioprocess technology at Chalmers University of Technol-
ogy heading the Industrial Biotechnology. The Industrial
Biotechnology group works on designing enzymes and
microorganisms to be used in sustainable bioprocesses.
Our primarily interest is on enzymes that act on plant cell
wall materials, with the aim to degrade, modify or upgrade
such material.
ogy engineering at Chalmers University of Technology,
Gothenburg after doing her diploma work on carotenoids
as natural colorants in food, at AB Nordbakels and within
the group of Food Science. Thereafter, she started her Ph.D.
project regarding enzyme encapsulation in mesoporous
materials. The project aim is to synthesize mesoporous
materials with varying properties and study how these
materials affect specific enzymes encapsulated inside the
pores.