Dependency of the hydrogen bonding capacity of the solvent anion on the thermal stability of feruloyl esterases in ionic liquid systems

Article abstract of DOI:10.1039/c1gc15115k

Three feruloyl esterases, EC 3.1.1.73, (FAEs), namely FAE A from Aspergillus niger (AnFaeA), FAE C from Aspergillus nidulans (AndFaeC), and the FAE activity in a commercial β-glucanase mixture from Humicola insolens (Ultraflo L) were tested for their ability to catalyse esterification of sinapic acid with glycerol in four ionic liquid (IL) systems. The IL systems were systematically composed of two selected pairs of cations and anions, respectively: [BMIm][PF₆], [C₂OHMIm][PF₆], [BMIm][BF₄], and [C₂OHMIm][BF₄]. AnFaeA had activity in [PF₆]^--based ILs, whereas the AndFaeC and the FAE in Ultraflo L had no appreciable activities and were generally unstable in the IL systems. FAE stability in the IL systems was apparently highly dependent on enzyme structure, and notably AnFaeA's similarity to IL-compatible lipases may explain its stability. The thermal stability of AnFaeA was higher in buffer than in the IL systems, but at 40 °C and below there was no significant difference in AnFaeA stability between the buffer and the [PF₆] ^--based systems: AnFaeA was stable in the [BMIm][PF₆] and [C₂OHMIm][PF₆] systems for 2 h at 40 °C. However, the IL anion had a major effect on stability: [BF₄]^- caused rapid inactivation of AnFaeA, while [PF₆]^- did not. The cation did not have a similar effect. These observations could be explained in terms of the hydrogen bonding capacity of IL cations and anions via COSMO-RS simulations.

Full text of DOI:10.1039/c1gc15115k

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
Green Chemistry
Cite this: Green Chem., 2011, 13, 1550
www.rsc.org/greenchem
PAPER
Dependency of the hydrogen bonding capacity of the solvent anion on the
thermal stability of feruloyl esterases in ionic liquid systems
Birgitte Zeuner,a Tim Sta˚hlberg,b Olivier Nguyen van Buu,b Andreas Jonas Kunov-Kruse,b Anders Riisagerb
and Anne S. Meyer*a
Received 29th January 2011, Accepted 4th April 2011
DOI: 10.1039/c1gc15115k
Three feruloyl esterases, EC 3.1.1.73, (FAEs), namely FAE A from Aspergillus niger (AnFaeA),
FAE C from Aspergillus nidulans (AndFaeC), and the FAE activity in a commercial b-glucanase
mixture from Humicola insolens (Ultraflo L) were tested for their ability to catalyse esterification
of sinapic acid with glycerol in four ionic liquid (IL) systems. The IL systems were systematically
composed of two selected pairs of cations and anions, respectively: [BMIm][PF6],
[C2OHMIm][PF6], [BMIm][BF4], and [C2OHMIm][BF4]. AnFaeA had activity in [PF6]--based
ILs, whereas the AndFaeC and the FAE in Ultraflo L had no appreciable activities and were
generally unstable in the IL systems. FAE stability in the IL systems was apparently highly
dependent on enzyme structure, and notably AnFaeA’s similarity to IL-compatible lipases may
explain its stability. The thermal stability of AnFaeA was higher in buffer than in the IL systems,
but at 40 ◦C and below there was no significant difference in AnFaeA stability between the buffer
and the [PF6]--based systems: AnFaeA was stable in the [BMIm][PF6] and [C2OHMIm][PF6]
systems for 2 h at 40 ◦C. However, the IL anion had a major effect on stability: [BF4]- caused rapid
inactivation of AnFaeA, while [PF6]- did not. The cation did not have a similar effect. These
observations could be explained in terms of the hydrogen bonding capacity of IL cations and
anions via COSMO-RS simulations.
ionic liquids (ILs). Recently, modifications of hydroxycinna-
mates have been performed in order to change their physico-
Introduction
Ferulic acid esterases (FAEs; EC 3.1.1.73) are accessory plant
chemical and functional properties via such enzyme catal-
cell wall-degrading enzymes, which catalyse the hydrolysis of
ysed esterification reactions. These types of reactions may
the ester bond between ferulic acid and the monosaccharide to
take place either through addition of aliphatic alcohols to
which it is covalently linked; in arabinoxylans the ferulic acid is
increase lipophilicity4 or by addition of carbohydrates.5 The
bound via O-5 bonds to arabinose, whereas linkages of ferulic
reported FAE catalysed esterification reactions are mainly direct
esterifications of hydroxycinnamic acids or transesterification
reactions of their esters with primary alcohols, e.g. 1-butanol,
glycerol. However, also reactions with a number of monosac-
acid to the O-2 of arabinose and O-6 of galactose have been
shown in pectins.1 FAEs do however also show activity towards
simpler esters (e.g. methyl esters) and to a varying degree also
towards esters of other hydroxycinnamates such as sinapic acid
(SA), caffeic acid (CA), and p-coumaric acid (pCA),2 and the
sub-type FAE nomenclature is based on the specific activity on
these latter substrates.3
FAEs can also be brought to catalyse the (trans)esterification
reaction in solvents that favour synthesis over hydrolysis, i.e.
systems with low water content such as organic solvents or
charides, glycosides, and arabino-oligosaccharides have been
demonstrated.4,5,6 The reactions have been accomplished in mi-
croemulsions of organic solvents with low water levels (<5%),4,5
in nearly solvent-free systems,6a or in a single organic solvent.6b
Hatzakis and Smonou7 have shown the ability of an FAE from
Humicola insolens to catalyse the transesterification of vinyl
acetate with a number of secondary alcohols in a solvent-free
system.
aCenter for BioProcess Engineering, Department of Chemical and
Biochemical Engineering, Technical University of Denmark, Building
229, DK-2800, Kgs. Lyngby, Denmark. E-mail: am@kt.dtu.dk
During the last decade, the interest in performing enzyme-
catalysed (trans)esterification reactions in ionic liquids (ILs) has
increased rapidly. This interest has mainly been motivated by a
desire to replace volatile organic solvents with non-volatile ILs,
which also have the advantage of allowing increased enzyme
bCentre for Catalysis and Sustainable Chemistry, Department of
Chemistry, Technical University of Denmark, Building 207, DK-2800,
Kgs. Lyngby, Denmark
1550 | Green Chem., 2011, 13, 1550–1557
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
enantio-selectivity and encompass the possibility of solvent
tailoring for implementation of new, efficient reaction regimes
due to their unconventional solvent properties.8 To date, most
of the enzyme-catalysed reactions in IL systems have been done
with lipases, especially lipase B from Candida antarctica (CaLB).
Stability of the enzymes in the IL matrix is vital when performing
these reactions. Although the enzyme stability issue has been
addressed in some cases,9 only Ulbert et al.10 and Lou and
Zong11 have tested the lipase stability (a lipase from Candida
rugosa and CaLB, respectively) at more than one temperature.
Hence, knowledge of enzyme thermal stability in IL systems is
scarce. Only a single study has shown the ability of FAE A from
Aspergillus niger (AnFaeA) to catalyse the (trans)esterification
of sinapic acid and methyl sinapate with glycerol in an IL-water
system using [C2OHMIm][PF6] and [C5O2MIm][PF6] with up to
30% of water.12
It is an obvious premise for successful catalysis that the
enzyme is active in the IL. However, it may be hypothesized
that the activity and stability of enzymes will be affected by
the reaction temperature and that the stability may vary for
different enzyme protein structures. Different ILs may moreover
affect the enzyme stability differently. The objective of this work
was to elucidate the effect of enzyme structure (and origin)
and IL nature on the thermal stability and activity of selected
FAEs in IL systems. This was done by determining the synthetic
activity and thermal stability of three different FAEs in IL-buffer
systems using a carefully selected series of four ILs with pairwise
similar anions and cations of varying hydrophobicity and polar-
ity, namely [C2OHMIm][PF6], [C2OHMIm][BF4], [BMIm][PF6],
and [BMIm][BF4] (Fig. 1). The choice of [C2OHMIm][PF6]
was based on the results previously obtained with FAE;12
the [BMIm][PF6] and [BMIm][BF4] ILs were included because
they are currently used widely in enzyme catalysis research
– although not with FAEs – whereas [C2OHMIm][BF4] was
included to complete the series. Furthermore, the study aims to
explore the use of the quantum chemistry-based COSMO-RS
method for explaining and predicting the effect of given ILs
on enzyme stability to provide a first foundation for predicting
the optimal IL for a particular FAE-catalysed esterification
reaction.
Materials and methods
Materials
Feruloyl esterase type A from Aspergillus niger and Ul-
traflo L (the latter is a commercial b-glucanase mixture
from Humicola insolens) were provided by Novozymes A/S
(Bagsværd, Denmark). Feruloyl esterase type C from As-
pergillus nidulans (AN5267.2) was produced by fermentation
essentially as described by Stratton et al.13 The Pichia pas-
toris clone transformed with the feruloyl esterase gene was
obtained from the Fungal Genetics Stock Center as described
by Bauer et al.14 Methyl sinapate (methyl 3-(4-hydroxy-3,5-
dimethoxyphenyl)prop-2-enoate) was purchased from Apin
Chemicals (Abingdon, UK). Anhydrous glycerol was pur-
chased from AppliChem (Darmstadt, Germany). LC-MS grade
methanol for HPLC analysis was purchased from Fischer
Scientific (Loughborough, UK). Diethyl ether was purchased
from Merck (Darmstadt, Germany) and MgSO4 from Riedel-de
Hae¨n (Seelze, Germany). All other chemicals, including sinapic
acid (3-(4-hydroxy-3,5-dimethoxyphenyl)prop-2-enoic acid) and
the ionic liquids [BMIm][BF4] (1-butyl-3-methyl imidazolium
tetrafluoroborate) and [BMIm][PF6] (1-butyl-3-methyl imida-
zolium hexafluorophosphate) (purity ≥97%), were purchased
from Sigma–Aldrich (St. Louis, MO, USA).
Preparation of [C2OHMIm][PF6] and [C2OHMIm][BF4]
Ionic liquids 1-(2-hydroxyethyl)-3-methylimidazolium hexaflu-
orophosphate, [C2OHMIm][PF6], and 1-(2-hydroxyethyl)-3-
methylimidazolium tetrafluoroborate, [C2OHMIm][BF4], were
prepared according to the method described by Branco et al.15
Feruloyl esterase activity assay on methyl sinapate
The activity assay was based on the one proposed by Juge et
al.16 In brief, 50 mL enzyme solution was added to 500 mL of a
preheated (37 ◦C) solution of 1 mM methyl sinapate (MSA) in
100 mM MOPS (3-(N-morpholino)propanesulfonic acid) buffer
(pH 6.0) to start the reaction. After 10 min of reaction at 37
◦C, the reaction was stopped by adding 200 mL glacial acetic
acid. Amounts of substrate (MSA) and product (sinapic acid;
SA) were analysed by RP-HPLC (see below). Activity (U) is
expressed as the amount of enzyme required to release 1 mmol
of sinapic acid per minute at 37 ◦C and pH 6.0.
Thermal stability
To mimic the system used for the esterification reaction, the
FAEs were incubated in an IL system containing 15% (v/v)
enzyme solution diluted in 100 mM MOPS buffer (pH 6.0). Each
reaction was conducted in an Eppendorf tube at 700 rpm and
at 30 ◦C, 40 ◦C, 50 ◦C, or 60 ◦C in a Thermomixer (Eppendorf,
Hamburg, Germany) for up to 2 h. After 0, 10, 20, 30, 60, and
120 min of incubation, enzyme samples (50 mL; 11.2 mU) were
taken out and tested in the feruloyl esterase activity assay at
50 ◦C. Residual activity was also measured in the esterification
reaction in [C2OHMIm][PF6] (see below) after 0, 30, and 60 min
of incubation in [C2OHMIm][PF6] at 40 ◦C and 50 ◦C. Control
samples were run without enzyme and no reaction was detected.
Fig. 1 Structures of the ionic liquid cations and anions used in this
study: 1-butyl-3-methylimidazolium ([BMIm]+), 1-(2-hydroxyethyl)-3-
methylimidazolium ([C2OHMIm]+), hexafluorophosphate ([PF6]-), and
tetrafluoroborate ([BF4]-).
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 1550–1557 | 1551
©

View Article Online
Enzyme catalysed esterification of glycerol with sinapic acid
profiles, and thus only a single one was used here. s-profiles
were generated with COSMOtherm (COSMOlogic (Leverkusen,
Germany))18 using the BP-TZVP parameterisation.
Using the results obtained by Vafiadi et al.12 in the esterification
of glycerol with sinapic acid using FAE from A. niger, the
esterification activity of each of the FAEs was tested in a system
with 2.5 M glycerol, 0.02 M sinapic acid (solubilised in the
IL from its solid form), and 15% (v/v) enzyme solution in
100 mM MOPS buffer (pH 6.0) to give 56 mU. Hence, all
reaction mixtures contained 15% (v/v) aqueous buffer, 18%
(v/v) glycerol, and 67% (v/v) IL. The total reaction volume
was 600 mL, and the reaction took place in an Eppendorf tube
kept at 1400 rpm and 40 ◦C (30 ◦C when using BF4-based ILs),
in a thermomixer (lower temperature than Vafiadi et al.12 to
ensure enzyme stability). The reaction was stopped by extracting
substrate and product with ethyl acetate (40 mL sample in 1
mL ethyl acetate) for 4 min at 40 ◦C and 1400 rpm. After
evaporation of the ethyl acetate extract, the remaining solids
were re-dissolved in 1 mL of a 1 : 1 methanol-water solution
and analysed by RP-HPLC (see below). Control samples were
run without enzyme or without glycerol, and no conversion
was detected. Conversion yields were calculated from the molar
amount of glycerol sinapate formed compared to the molar
amount of sinapic acid originally present.
Comparison of FAE sequence and structure
Initial sequence similarity searches were performed with
protein-protein BLAST using the PSI-BLAST algorithm
(http://blast.ncbi.nlm.nih.gov/Blast.cgi).19 Structure similarity
searches for homologous proteins were performed by HHPred
(http://toolkit.tuebingen.mpg.de/hhpred)20 using default set-
tings.
Statistics
One-way ANOVA for determination of statistical significance
was made in Minitab 16 (Minitab Inc., State College, PA, USA).
Statistical significance was established at p < 0.05.
Results and discussion
Effect of FAE structure on activity in IL-buffer systems
Among the four ILs used, [BMIm][PF6], [C2OHMIm][PF6],
[BMIm][BF4], and [C2OHMIm][BF4] (Fig. 1), the system with
[BMIm][PF6] formed two phases due to the water-immiscibility
of this IL. Enzyme activity was found in the aqueous phase of
this two-phase system only (data not shown). The other three
ILs were water-miscible and only one phase was formed in each
of these IL reaction systems.
The AndFaeC was inactivated immediately in the
[C2OHMIm]+-based IL systems, and in less than 10 min
in the [BMIm]+-based IL systems. Consequently, only minor
esterification activity, i.e. ~1% conversion, was seen and only in
the [BMIm][PF6] system (Table 1). However, in MOPS buffer,
pH◦6.0, the AndFaeC was completely stable for more than 2 h at
40 C (data not shown). This indicated that the AndFaeC was
very sensitive to the IL environment. Similarly, the FAE activity
in the Ultraflo L preparation was inactivated immediately in the
[C2OHMIm]+-based IL systems, and in less than 10 min in the
[BMIm][BF4] system (Table 1). In [BMIm][PF6], however, the
FAE activity present in the Ultraflo L was stable throughout the
30 min of reaction (data not shown). This stability was possibly
due to the enzyme being present in the aqueous phase in this
two-phase system, rather than in the IL matrix. Despite being
Analysis of reaction components by RP-HPLC
Quantitative analyses were made by RP-HPLC using a Chem-
station 1100 series, Hewlett Packard equipped with a C18
column (150 mm ¥ 4.6 mm, 3 mm; Phenomenex (Torrance, CA,
USA)) with DAD detection of the substrate and product with
quantification at 320 nm. Pure sinapic acid and methyl sinapate
were used as external standards. Elution was conducted at 30
◦C with [methanol] : [10% (v/v) acetic acid in water] ([4] : [6]) as
the mobile phase at a flow rate of 0.5 mL min-1 for analysis
of the hydrolysis reaction and 0.3 mL min-1 for analysis of the
esterification reaction. Samples were filtered through a 0.2 mm
syringe tip filter prior to analysis.
Generation of ionic liquid r-profiles by COSMO-RS
The COSMO-file for [C2OHMIm]+ was made with Gaussian03
(Gaussian Inc. (Wallingford, CT, USA)).17 The structure was
first optimised with the semi-empirical PM3 method and then
refined with B3LYP, first with the basis set 6-31 and then
with 6-311++G(d,p). After optimisation the two conformers
of [C2OHMIm]+ were almost identical with regards to s-
Table 1 Conversion (%) of sinapic acid to glycerol sinapate after 30 min of reaction at 40 ◦C for [PF6]--based ILs and 30 ◦C for [BF4]--based ILs in
an ionic liquid-buffer (15% v/v) system by 56 mU FAE from three different sources: FAE A from Aspergillus niger (AnFaeA), FAE from Humicola
insolens found as a side activity in the commercial b-glucanase mixture Ultraflo L, and FAE C from Aspergillus nidulans (AndFaeC). The number
of phases in each IL-buffer system is determined by visual detection. If complete inactivation has taken place during the reaction this is indicated as
follows: i0: complete inactivation occurs immediately (less than 30 s); i10: complete inactivation takes place within 10 min
Ionic liquid
No. of phases
AnFaeA
Ultraflo L
AndFaeC
[BMIm][PF6]
[C2OHMIm][PF6]
[BMIm][BF4]
2
1
1
1
13 3%a,x
21 2%b,x
0.9 0.1%c,x
0%c,x (i10)
1.0 0.1%a,y
0%b,y (i0)
1.1 0.1%a,y (i10)
0%b,y (i0)
0%b,y (i10)
0%b,x (i0)
0%b,y (i10)
[C2OHMIm][BF4]
0%b,x (i0)
Superscript letters a–c indicate significant difference (one-way ANOVA; p < 0.05) between reaction outcomes in different ILs for each enzyme, and
letters x and y indicate significant difference between enzymes for each IL.
1552 | Green Chem., 2011, 13, 1550–1557
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
stable, the esterification activity of the FAE activity in Ultraflo
L was very low (1% conversion; Table 1).
water-miscible IL to which AnFaeA is sensitive. This anion effect
is discussed further below.
In contrast, the AnFaeA exhibited significant esterification
activity and catalysed 13% conversion in [BMIm][PF6] and
21% conversion in [C2OHMIm][PF6] (Table 1). The former is
remarkable as Vafiadi et al.12 reported that no esterification
activity was found in [BMIm][PF6], using the same enzyme and
a similar system. As pointed out by Park and Kazlauskas,21
impurities in the IL and a resulting shift in pH may be the
reason for contradictory results. AnFaeA was stable throughout
the reaction time in these two [PF6]--based IL-buffer systems,
but the enzyme was apparently sensitive to the [BF4]- systems.
Consequently, only low and insignificant activity, 1% conver-
sion, was obtained in [BMIm][BF4], where the enzyme had only
36% residual activity after 10 min (Fig. 2a). No esterification
activity was observed in [C2OHMIm][BF4] (Table 1), where the
enzyme was completely inactivated within 10 min. This indicated
that it is the [BF4]- anion rather than the single-phase system or
The significant difference in esterification activity of AnFaeA
observed between [BMIm][PF6] and [C2OHMIm][PF6] is most
likely due to differences in water-miscibility and thus water
activity as well as viscosity between the two ILs. Generally,
water-miscible ILs exhibit a lower water activity, aw, at a
given water content than water-immiscible ILs do.22 Thus, the
supposedly lower aw of the water-miscible [C2OHMIm][PF6]
system may in part account for the higher esterification activity.
However, at this high water content (15% v/v) even the more
hydrophilic [BF4]--based systems have an aw fairly close to 1.23
Similarly, the aw of the system containing the more hydrophobic
[C2OHMIm][PF6] will therefore not be dramatically different
from the aw of the water-immiscible systems which reach
maximum aw = 1 at much lower water contents.24 Even with
18% (v/v) glycerol in the reaction mixture, the effect of the
differential viscosity between the ILs (67% (v/v) of the reaction
mixture) cannot be ignored, and the markedly higher viscosity of
[BMIm][PF6] compared to [C2OHMIm][PF6] is likely to cause
lower reaction rates through mass transfer limitations.25 The
interface formed in the two-phase system with [BMIm][PF6],
the aqueous enzyme solution, and glycerol, may also induce
lowered mass transfer and decrease the esterification rate. For
an IL-buffer batch system it is therefore advisable to use an IL
which is water-miscible and still contains the stabilising [PF6]-
anion, e.g. [C2OHMIm][PF6]. From these results it can therefore
be concluded that the AnFaeA exhibited potential for being
used for enzyme catalysed esterification reactions in an IL-buffer
system, whereas AndFaeC and Ultraflo L did not.
All three FAEs have pH optima in the pH 5–6 range (Table
2). AnFaeA and Ultraflo L have higher temperature optima,
55–60 ◦C and 60–65 ◦C, respectively, than AndFaeC (37 ◦C)
(Table 2). AndFaeC was also found to be less thermally stable
in MOPS buffer (pH 6.0) than the other two FAEs (data not
shown). It was thus to be expected that AnFaeA and Ultraflo
L would be more stable, at least in the [BMIm][PF6] system
(Table 1).
Remarkably, the feruloyl esterase activity in Ultraflo L had
the lowest affinity towards MSA, the substrate towards which
AnFaeA and AndFaeC had the highest affinity. Since the FAEs
were dosed according to their hydrolytic activity on MSA, this
cannot explain the low esterification activity of Ultraflo L in
[BMIm][PF6]. However, if the FAE from H. insolens found in
Ultraflo L is similar to the one found in Pentopan 500 BG (also
a Novozymes blend), which Hatzakis and Smonou7 found to
have affinity for secondary alcohols, this may in part explain why
Fig. 2 Thermal stability of FAE A from A. niger (AnFaeA) in IL
systems: Residual activity of AnFaeA at: (a) 30 ◦C, (b) 40 ◦C, (c) 50 ◦C,
and (d) 60 ◦C when incubated in 100 mM MOPS buffer pH 6.0 (ꢀ; open
squares), [BMIm][PF ] ( ; filled circles), [C OHMIm][PF ] (ꢀ; open
᭹
6
2
6
triangles), and [BMIm][BF4] (᭜; filled diamonds) for up to 120 min as
compared to hydrolytic activity at 0 min of incubation. Thermal stability
was also tested in [C2OHMIm][BF4], but inactivation was immediate
(took place in less than 30 s) and for simplicity the data are thus not
included in the figure.
Table 2 Properties of ferulic acid esterase A from A. niger (AnFaeA),26 ferulic acid esterase from Humicola insolens found in Ultraflo L,27 and ferulic
acid esterase C from A. nidulans (AndFaeC)14
AnFaeA
Ultraflo L
AndFaeC
Number of amino acidsa
pH optimum
Temperature optimum
Substrate affinity
260
5
273
249
6.1
~5–6
55–60 ◦
C
60–65 ◦
C
37 ◦
C
MSA > MFA > MpCA
MCA > MFA > MpCA > MSA
MSAb > MFA
MSA: methyl sinapate; MFA: methyl ferulate; MpCA: methyl p-coumarate; MCA: methyl caffeate.a In the final enzyme, without signal peptide.
b MSA tested in this paper; MCA and MpCA not tested.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 1550–1557 | 1553
©

View Article Online
the activity is lower in the current system. It cannot be excluded
that AndFaeC and the feruloyl esterase activity in Ultraflo L
might express higher esterification activity at lower water content
than AnFaeA, which has been found to have its optimum at the
15% (v/v) used here. However, since AndFaeC and the feruloyl
esterase activity in Ultraflo L also showed low stability in the IL
systems (Table 1), it seems more relevant to study the structural
differences between the three enzymes.
denaturation, i.e. re-activation, of AnFaeA, for example upon
transfer to an aqueous buffer system for activity measurement,
would introduce errors in the thermal stability results, the
thermal stability of AnFaeA was also determined as residual
esterification activity in the [C2OHMIm][PF6] system upon
incubation in [C2OHMIm][PF6] for up to 1 h (Fig. 3). The data
confirmed that there was no significant difference in thermal
stability of AnFaeA when measured as residual hydrolytic
activity and when measured as residual esterification activity. It
can thus be concluded that the temperature-induced inactivation
of AnFaeA in [C2OHMIm][PF6] is not reversible, and that the
thermal stability could be assessed based on hydrolytic activity
as well as esterification activity.
Most importantly, the three FAEs differ in their number
of amino acids (aa.), indicating differences in their overall
structure. It has already been established that AnFaeA (260
aa.) has sequence and structure similarities to fungal lipases,
especially the open form of the lipases from Rhizomucor miehei
(37% sequence identity) and Thermomyces lanuginosis (30%
sequence identity).28 These particular lipases have previously
been found to work fairly well in ionic liquid systems, albeit not
as well as the very robust lipase B from Candida antarctica.29
In contrast, AndFaeC (249 aa.) shows a sequence similarity to
other (feruloyl) esterases (PSI-BLAST) and when predicting a
structure of the enzyme by homology modelling using HHPred,
the best match is the ferulic acid esterase domain of the
cellulosomal xylanase Z (XynZ) in Clostridium thermocellum
(sequence identity 21%; E-value 1.6 ¥ 10-26). Hermoso et
al.28 found that XynZ has poor homology in both sequence
and structure with AnFaeA. Although AnFaeA and the FAE
domain of XynZ differ in overall structure, their Ser-Asp-His
catalytic triads in the active site are identical and they both
present a long and narrow substrate-accommodating cavity in
contrast to the wide and short ones generally found in lipases.28
Thus, AnFaeA is similar in catalytic mechanism to the FAEs, but
similar in structure to some of the fungal lipases – a feature that
may explain its higher stability in IL-buffer systems as compared
to the other FAEs tested here.
Fig. 3 Residual activity of FAE A from A. niger (AnFaeA) after
incubation in [C2OHMIm][PF6] for up to 1 h at 40 ◦C measured
as hydrolytic activity (᭡; filled triangles) and as esterification activity
in [C2OHMIm][PF6] (ꢁ; filled squares), and at 50 ◦C measured as
hydrolytic activity (ꢀ; open triangles) and as esterification activity
in [C2OHMIm][PF6] (ꢀ; open squares) as compared to the respective
activities at 0 min of incubation.
The ferulic acid esterase from H. insolens found in Ultraflo L
(273 a.a.)27b showed sequence similarity to feruloyl and acetyl
xylan esterases (PSI–BLAST), and structurally it also has some
similarity to the FAE domain of XynZ from C. thermocellum,
albeit less than AndFaeC (sequence identity 16%; E-value 3.4 ¥
10-23). In conclusion, the fact that AnFaeA has a structure more
similar to IL-compatible lipases than the other FAEs does seem
to be determining for its ability to work well in IL systems.
◦
◦
At 30 C and 40 C, AnFaeA retained full activity for over
2 h in [BMIm][PF6] and [C2OHMIm][PF6] and there was no
significant difference between the stability in the [PF6]--based IL
systems and stability in the pH 6.0 buffer (Fig. 2a,b). However,
at 50 ◦C AnFaeA was more stable in buffer (kD = 0.0012)
than in [BMIm][PF6] (kD = 0.0191) and [C2OHMIm][PF6] (kD =
0.0105). For more than 30 min of incubation, AnFaeA showed
significantly higher residual activity in [BMIm][PF6], but after
2 h the activity was significantly higher in [C2OHMIm][PF6].
After 2 h at 50 ◦C, the residual activity of AnFaeA was 84% in
buffer, 26% in [C2OHMIm][PF6], and 11% in [BMIm][PF6] (Fig.
2c). At 60 ◦C, inactivation was rapid in all media: complete
inactivation took place within 10 min in [BMIm][PF6], within
20 min in [C2OHMIm][PF6] (kD = 0.184), and within 1 h in buffer
(kD = 0.0646) (Fig. 2d).
Effect of IL nature on enzyme thermal stability
Only the FAE A from A. niger showed appreciable esterification
activity in the IL-buffer systems, and was thus chosen for thermal
stability tests at 30–60 ◦C in [BMIm][PF6], [C2OHMIm][PF6],
[BMIm][BF4], and [C2OHMIm][BF4] containing 15% (v/v)
buffer.
Thermal stability has been assessed by determining residual
activity in an aqueous medium after incubation in ILs. It
has been found that CaLB to some extent (33–73%) refolded
upon addition of excess water after having been incubated in
denaturing ILs.30 Later, the same group found denaturation of
CaLB to be irreversible in another denaturing IL, and concluded
that enzyme inactivation in ILs includes a first reversible step
and a second irreversible one.31 The irreversibility of enzyme
denaturation in IL systems may (in part) be caused by aggre-
gation of the denatured enzymes.32 Since any reversibility in the
In contrast, AnFaeA was highly unstable in the [BF4]--based
◦
◦
IL systems. Even at 30 C and 40 C inactivation of AnFaeA
was significant in [BMIm][BF4]: the ◦residual activity was 36%
after 10 min and 8% after 2 h at 30 C (Fig. 2a; kD = 0.0174),
and 31% after 10 min and 7% after 2 h at 40 ◦C (Fig. 2b; kD
= 0.0179). At 50 ◦C, complete inactivation took place within
10 min (Fig. 2c). In the [C2OHMIm][BF4] system, AnFaeA was
1554 | Green Chem., 2011, 13, 1550–1557
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
completely inactivated in less than 30 s (data not shown). The
results emphasise that just by changing the anion from [PF6]-
to [BF4]- while maintaining the same cations, the effect of the
IL on AnFaeA stability changes dramatically. As previously
found for lipases,9a,33 the anion has the dominant effect on FAE
stability, and the choice of IL is thus of crucial importance when
designing an enzyme-IL system for FAE catalysed esterification
reactions. Importantly, the effect of [BF4]- anion in particular
has been subject to some debate, since some studies have
found [BMIm][BF4] to have a significantly negative effect on
the stability of CaLB (Novozym 435),9b,34 whereas Lau et al.9a
reported full activity and no enzyme dissolution using the same
CaLB and [BMIm][BF4]. Again, different levels of IL purity may
explain these contradictory results.21 For FAEs, however, [BF4]-
does seem to have a significant, detrimental effect.
In a recent review, Zhao35 listed a number of IL properties
that are likely to influence enzyme stability and activity in IL
media, namely polarity, hydrogen bond basicity, nucleophilicity,
hydrophobicity, viscosity, and in aqueous IL systems also ion
kosmotropicity. Even if no solid prediction tool for enzyme-IL
compatibility has been developed due to the complex nature
of IL-enzyme systems, there seems to be general consensus
that low hydrogen bond basicity and nucleophility as well as
high viscosity and especially high hydrophobicity all favour
enzyme stability and activity in ILs.35 Increasing viscosity
may however also decrease the reaction rate of the enzyme.25
Similarly, it has been established that the relationship between
enzyme activity and hydrophobicity follows a bell-shaped curve:
activity increases with hydrophobicity up to a certain point,
but then decreases due to a more thermodynamic ground state
stabilization of the substrates and thus less tendency to react.36
No clear relationship between enzyme activity and polarity has
been established, although one study indicated that the reaction
rate is higher in more polar ILs.37 In general, it is believed
that the anion has a larger effect on enzyme stability than the
cation does,31 and that large anions with delocalised negative
charge on more atoms are generally more stabilising.38 Larger
anions have also been suggested to be less destabilising because
they are sterically demanding, in that their size would require
many hydrogen bonds in the protein matrix to be broken in
order to form a few new ones, making such an interaction less
favourable.9a
Although hydrophobicity, nucleophilicity, and hydrogen bond
basicity are separate properties that should not be confused
with each other, the general picture is that the more hydrophilic
IL anions are also the ones with the higher nucleophilicity
and hydrogen bond basicity, rendering them more likely to
interact with positively charged sites in the enzyme thus caus-
ing it to change conformation – and vice versa.35 Generally,
hydrophobic ILs with [PF6]- and [Tf2N]- (bis(trifluoromethane-
sulfonyl)amide) anions are stabilising while hydrophilic ILs with
[NO3]-, [lactate]-, [TfO]- (trifluoromethanosulfonate), [EtSO4]-,
and in some cases also [BF4]- are destabilising to enzymes.9 The
thermal stability results obtained in this work (Fig. 2) support
this hypothesis and are the first to be presented for FAEs.
In aqueous systems of hydrophilic ILs such as the [BF4]--based
systems used here, where the IL is hydrated and dissociates into
individual ions, the importance of having a chaotropic cation
and especially a strong kosmotropic anion has been pointed
out.39 Therefore, the chaotropic nature of the [BF4]- anion may
also play an important role in the FAE inactivation in the
IL-buffer system. The same tendency has not been observed
for hydrophobic ILs which have low solubility in water and
therefore limited ion dissociation,35 explaining why the [PF6]-
anion does not cause enzyme inactivation despite its highly
chaotropic nature.40
Cation hydrophobicity has also been found have an effect on
enzymes since enzyme enantioselectivity and stability decreases
with decreasing alkyl chain length on imidazolium cations
from [OMIm]+ to [BMIm]+.11 This effect may, however, also be
confounded with the viscosity effect since a longer alkyl chain
length in substituents on the imidazolium cation results in higher
viscosity in the range mentioned above (4–8 carbons).41 This
cation effect may be restricted to very similar cations like the
ones used in the studies mentioned above. At least, it does not
extend to the ones used in this work: [BMIm][PF6] has higher
viscosity and hydrophobicity than [C2OHMIm][PF6], but this
does not affect the stability of AnFaeA in the two media. There
may have been an effect for Ultraflo L (see Table 1), but this
may also be explained by the two-phase system formed with
[BMIm][PF6], and this difference in water-miscibility is indeed
the major cation effect observed in this work.
Potential of the COSMO-RS method for explaining and
predicting FAE stability in ILs
The quantum chemistry-based method COSMO-RS has been
introduced as a fast way of performing semi-quantitative tasks
such as solvent screening, e.g. for substrate solubility, especially
for complex media like ILs since it has the advantage over the
classical group contribution methods (e.g. UNIFAC) that it is
not limited to interpolation and partial extrapolation based on
available data.18 Therefore, its ability to explain the effects of
the four different ILs used in this study on FAE stability has
beentested. The COSMO-RSmethodcalculates the polarisation
or screening charge density (SCD), s, which can be seen as a
local measure of polarity for each molecule. The frequencies
of screening charge densities ranging from s = -3 e nm-2 to
s = 3 e nm-2 on the [BMIm]+ and [C2OHMIm]+ cations and
[PF6]- and [BF4]- anions – also known as s-profiles – have been
plotted in Fig. 4 along with the s-profile for FAE’s ‘natural’
solvent, water, for comparison. The hydrogen bonding threshold
is sHB
1 e nm-2 only surface segments with an s-value beyond
=
0.79 e nm-2, but as hydrogen bonding is weak below
1
e nm-2 are considered strongly polar and potentially hydrogen
bonding.18 From the s-profiles it is seen that the peak SCD of
[BF4]- is found outside the hydrogen bonding limit, whereas the
peak SCD of the more hydrophobic [PF6]- is inside this limit
(Fig. 4). Thus, the destabilising effect of [BF4]- on the FAEs
reported here can be explained by the tendency of [BF4]- to act
as a hydrogen bond acceptor and thus interact with the enzyme
and disturb its hydrogen bond-based structure. It should also
be noted that the SCD peak at s > 1 e nm-2 is much larger for
[BF4]- than for water, thus explaining why the IL anion has a
destabilising effect while water does not. Although the difference
in water-miscibility between the two cations is harder to account
for by the s-profiles (the contribution of the hydroxyl group in
[C2OHMIm]+ can be seen in the range from 0.5 to 1 e nm-2), the
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 1550–1557 | 1555
©

View Article Online
stabilisation methods such as formation of cross-linked enzyme
aggregates (CLEAs) on FAE stability, since this has been found
to have a positive effect on CaLB stability in ILs.31
The COSMO-RS method proved to be a useful tool for
explaining the effect of the four ILs studied in this work
on AnFaeA stability in terms of hydrogen bonding capacity.
Given the importance of hydrogen bonds in maintaining enzyme
structure, the COSMO-RS method may thus prove to enable
the prediction of useful ILs for enzymatic reactions in terms of
stability. Further tests need to be made on other ILs in order to
establish the actual potential, but the four ILs used in this study
are good representatives of the ILs commonly used in the field.
Although other aspects of IL nature such as water miscibility
may not be predicted by COSMO-RS, the method can still be
used for initial IL screening to give an indication of substrate
solubility and enzyme stability in the IL medium.
References
Fig. 4 Sigma(s)-profiles of the cations and anions in the ILs used:
[BMIm]+ (conformer 0 ( ; solid red line) and conformer 1 (---; short
maroon dashes)), [C2OHMIm]+ (— —; long black dashes), [PF6]- (– –
–; blue dashes), and [BF4]-( ◊ ◊ ◊ ·; purple dots). Water ( solid grey line)
is included for comparison. The vertical lines indicate the hydrogen
bonding (HB) limits, i.e. screening charge densities outside these lines
indicate the tendency to form hydrogen bonds. s-Profiles are shown for
two conformers 0 and 1 of [BMIm]+; COSMOtherm estimates that in this
IL-buffer system (15% (v/v) water), 85.0% of the [BMIm]+ molecules
are in the conformer 0 form, while 10.2% are in the conformer 1 form.
The two conformers generated for [C2OHMIm]+ had almost identical
s-profiles; thus, only one is shown.
1 I. Colquhoun, M. C. Ralet, J. F. Thibault, C. B. Faulds and G.
Williamson, Carbohydr. Res., 1994, 263, 243–256.
2 D. W. S. Wong, Appl. Biochem. Biotechnol., 2006, 133, 87–112.
3 V. F. Crepin, C. B. Faulds and I. F. Connerton, Appl. Microbiol.
Biotechnol., 2004, 63, 647–652.
4 E. Topakas, C. Vafiadi, H. Stamatis and P. Christakopoulos, Enzyme
Microb. Technol., 2005, 36, 729–736.
5 C. Vafiadi, E. Topakas, E. J. Bakx, H. A. Schols and P. Christakopou-
los, Molecules, 2007, 12, 1367–1375.
6 (a) T. Tsuchiyama, T. Sakamoto, T. Fujita, S. Murata and H.
Kawasaki, Biochim. Biophys. Acta, Gen. Subj., 2006, 1760, 1071–
1079; (b) M. Mastihubov
Borekova´, J. Mol. Catal. B: Enzym., 2006, 38, 54–57.
a´, V. Mastihuba, D. Bilancova´ and M.
7 N. S. Hatzakis and I. Smonou, Bioorg. Chem., 2005, 33, 325–337.
8 (a) F. van Rantwijk and R. A. Sheldon, Chem. Rev., 2007, 107, 2757–
2785; (b) H. Zhao, G. A. Baker, Z. Song, O. Olubajo, T. Crittle and
D. Peters, Green Chem., 2008, 10, 696–705.
9 (a) R. M. Lau, M. J. Sorgedrager, G. Carrea, F. van Rantwijk, F.
Secundo and R. A. Sheldon, Green Chem., 2004, 6, 483–487; (b) S.
H. Lee, Y. M. Koo and S. H. Ha, Korean J. Chem. Eng., 2008, 25(6),
1456–1462.
fact that no major difference in the effect on AnFaeA stability
was observed between [BMIm]+ and [C2OHMIm]+ is consistent
with the two cations having similar s-profiles in the hydrogen
bond donor range (s < -1 e nm-2).
Other studies have also found COSMO-RS to be a useful
tool for solvent screening in IL-based biocatalysis. For example,
Lue et al.42 studied enzymatic flavonoid acylation with CaLB
in numerous different ILs successfully screened for flavonoid
solubility with COSMO-RS and found the method useful for
choosing an IL giving the right balance between substrate
solubility and enzyme stability.
10 O. Ulbert, K. Belafi-Bako, K. Tonova and L. Gubicza, Biocatal.
Biotransform., 2005, 23(3/4), 177–183.
11 W. Y. Lou and M. H. Zong, Chirality, 2006, 18, 814–821.
12 C. Vafiadi, E. Topakas, V. R. Nahmias, C. B. Faulds and P.
Christakopoulos, J. Biotechnol., 2009, 139, 124–129.
13 J. Stratton, V. Chiruvolu and M. Meagher, in Pichia Protocols, ed.
D. Higgins and J. Cregg, Humana Press, Totowa, 1998, vol. 103, pp.
107–120.
14 S. Bauer, P. Vasu, S. Persson, A. J. Mort and C. R. Somerville, Proc.
Natl. Acad. Sci. U. S. A., 2006, 103(30), 11417–11422.
15 L. C. Branco, J. N. Rosa, J. J. Moura Ramos and C. A. M. Afonso,
Chem.–Eur. J., 2002, 8(16), 3671–3677.
Conclusions
Of the three FAEs tested here, AnFaeA showed the best potential
for being used for esterification in an IL-buffer system, whereas
AndFaeC and Ultraflo L were practically inactive in the systems.
AnFaeA has a structure more similar to IL-compatible lipases
than the other FAEs, and this seems to be determining for its
ability to work well in IL systems. The stability of AnFaeA in
the IL-buffer system was found to be anion dependent: AnFaeA
was rapidly inactivated in the [BF4]--based IL systems, but stable
in the [PF6]--based ones, which were only significantly more
16 N. Juge, G. Williamson, A. Puigserver, N. J. Cummings, I. F.
Connerton and C. B. Faulds, FEMS Yeast Res., 2001, 1, 127–132.
17 M. J. Frisch, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R.
Cheeseman, J. A. Jr. Montgomery, T. Vreven, K. N. Kudin, J.
C. Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J. E. Knox, H. P. Hratchian, J. B. Cross, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A.
D. Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck,
K. Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul,
S. Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-
Laham, C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill,
B. Johnson, W. Chen, M. W. Wong, C. Gonzalez, and J. A. Pople,
◦
destabilising than buffer at 50 C and above. The [PF6]--based
IL systems can thus be used with success at temperatures up to 40
◦C. Whether the water-miscible [C2OHMIm][PF6] or the water-
immiscible [BMIm][PF6] should be used depends on the nature
of the desired system – in some cases a two-phase system may
be preferable. It may be valuable to assess the effect of simple
1556 | Green Chem., 2011, 13, 1550–1557
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Gaussian 03W, Revision B.04, Gaussian, Inc.: Pittsburgh PA15106,
2003.
29 T. De Diego, P. Lozano, M. A. Abad, K. Steffensky, M. Vaultier and
J. L. Iborra, J. Biotechnol., 2009, 140, 234–241.
30 R. A. Sheldon, R. M. Lau, M. J. Sorgedrager, F. van Rantwijk and
K. R. Seddon, Green Chem., 2002, 4, 147–151.
31 F. van Rantwijk, F. Secundo and R. A. Sheldon, Green Chem., 2006,
8, 282–286.
32 D. Sate, M. H. A. Janssen, G. Stephens, R. A. Sheldon, K. R. Seddon
and J. R. Lu, Green Chem., 2007, 9, 859–867.
33 J. L. Kaar, A. M. Jesionowski, J. A. Berberich, R. Moulton and A. J.
Russell, J. Am. Chem. Soc., 2003, 125, 4125–4131.
34 S. H. Ha, S. H. Lee, D. T. Dang, M. S. Kwon, W. J. Chang, Y. J. Yu,
I. S. Byun and Y. M. Koo, Korean J. Chem. Eng., 2008, 25(2), 291–
294.
35 H. Zhao, J. Chem. Technol. Biotechnol., 2010, 85, 891–907.
36 H. Zhao, G. A. Baker, Z. Song, O. Olubajo, L. Zanders and S. M.
Campbell, J. Mol. Catal. B: Enzym., 2009, 57, 149–157.
37 S. Park and R. J. Kazlauskas, J. Org. Chem., 2001, 66, 8395–8401.
38 J. Gorke, F. Srienc and R. J. Kazlauskas, Biotechnol. Bioprocess Eng.,
2010, 15, 40–53.
18 A. Klamt, COSMO-RS. From quantum chemistry to fluid phase
thermodynamics and drug design. Elsevier, Amsterdam, 2005.
19 S. F. Altschul, T. L. Madden, A. A. Scha¨ffer, J. Zhang, Z. Zhang, W.
Miller and D. J. Lipman, Nucleic Acids Res., 1997, 25(17), 3389–3402.
20 J. So¨ding, Bioinformatics, 2005, 21(7), 951–960.
21 S. Park and R. J. Kazlauskas, Curr. Opin. Biotechnol., 2003, 14, 432–
437.
22 S. Garcia, N. M. T. Lourenco, D. Lousa, A. F. Sequeira, P. Mimoso,
J. M. S. Cabral, C. A. M. Afonso and S. Barreiros, Green Chem.,
2004, 6, 466–470.
23 I. Bou Malham, P. Letellier and M. Turmine, Talanta, 2007, 72,
155–164.
24 J. L. Anthony, E. J. Maginn and J. F. Brennecke, J. Phys. Chem. B,
2001, 105, 10942–10949.
25 M. Eckstein, P. Wasserscheid and U. Kragl, Biotechnol. Lett., 2002,
24, 763–767.
26 C. B. Faulds and G. Williamson, Microbiology, 1994, 140, 779–787.
27 (a) C. B. Faulds, A. I. Sancho and B. Bartolome´, Appl. Microbiol.
Biotechnol., 2002, 60, 489–493; (b) M. Maranta and K. Brown, US
Pat., 0151026 A1, 2009.
28 J. A. Hermoso, J. Sanz-Aparicio, R. Molina, N. Juge, R. Gonzalez
and C. B. Faulds, J. Mol. Biol., 2004, 338, 495–506.
39 H. Zhao, J. Mol. Catal. B: Enzym., 2005, 37, 16–25.
40 H. Zhao, J. Chem. Technol. Biotechnol., 2006, 81, 877–891.
41 H. Tokuda, K. Hayamizu, K. Ishii, M. A. B. H. Susan and M.
Watanabe, J. Phys. Chem. B, 2005, 109, 6103–6110.
42 B. M. Lue, Z. Guo and X. Xu, Process Biochem., 2010, 45, 1375–1382.
This journal is
The Royal Society of Chemistry 2011
Green Chem., 2011, 13, 1550–1557 | 1557
©