Both hydrolytic and transesterification activities of Penicillium expansum lipase are significantly enhanced in ionic liquid [BMIm][PF6]

Article abstract of DOI:10.1016/j.molcatb.2009.11.014

Both hydrolytic and transesterification activities of Penicillium expansum lipase (PEL) were investigated in the ionic liquid [BMIm][PF₆] as well as in organic solvents such as hexane. The initial rate of PEL-catalyzed hydrolysis of p-nitrophen

Full text of DOI:10.1016/j.molcatb.2009.11.014

Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Both hydrolytic and transesterification activities of Penicillium expansum lipase
are significantly enhanced in ionic liquid [BMIm][PF6]
Zhen Yang∗, Kai-Pei Zhang, Ying Huang, Zhi Wang
College of Life Sciences, Shenzhen University, Nanshan District, Shenzhen 518060, Guangdong, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Both hydrolytic and transesterification activities of Penicillium expansum lipase (PEL) were investigated
in the ionic liquid [BMIm][PF6] as well as in organic solvents such as hexane. The initial rate of PEL-
catalyzed hydrolysis of p-nitrophenyl palmitate (pNPP) was 12-fold enhanced in the ionic liquid. The
optimal water content required by PEL in [BMIm][PF6] was 10 times higher relative to that in hexane
due to the greater polarity of the ionic liquid. Direct addition of salt hydrates into the two nonaqueous
reaction media showed different impacts on the enzyme activity, which could be related to the dual
functions of the salt hydrates, i.e., the water buffering effect, and the specific ion effect (Hofmeister
effect). The transesterification activity of PEL was reflected by the yield of producing fatty acid methyl
esters (FAMEs) in the methanolysis of corn oil for 25 h, which was 69.7% in [BMIm][PF6], as compared to
19.4%, 14.0%, and 1.0% obtained in tert-butanol, solvent-free system, and hexane, respectively. The high
production yield of FAMEs obtained in [BMIm][PF6] demonstrates the potential use of ionic liquids as the
reaction media for PEL-catalyzed biodiesel production.
Received 22 September 2009
Received in revised form
17 November 2009
Accepted 27 November 2009
Available online 2 December 2009
Keywords:
Penicillium expansum lipase
Ionic liquid
Water activity
Biodiesel
Hofmeister effect
Jones–Dole viscosity B coefficient
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
biodiesel production is the lipase B from Candida antarctica (CALB),
which has been commercialized by Novozymes as an immobilized
Lipases are triacylglycerol ester hydrolases (EC 3.1.1.3) that
catalyze the hydrolysis and synthesis of esters formed from glyc-
erol and long-chain fatty acids. Being widely found in nature,
many lipases are active in organic solvents where they catalyze
a number of useful reactions including hydrolysis, esterification,
transesterification, kinetic resolution of racemic acids and alcohols,
regioselective acylation of steroids, sugars, and sugar derivatives,
and synthesis of peptides and other chemicals [1,2]. Lipases have
become industrially important, especially in detergent, food, and
pulp and paper industries [3]. More recently, lipases have found an
important application in biodiesel production [4].
As a biodegradable, nontoxic, cleaning, and renewable fuel,
biodiesel is a mixture of fatty acid methyl esters (FAMEs) which
are produced from a broad range of crude oil materials, such
as vegetable oil, animal fats, and waste oil, via transesterifica-
tion of triglycerides with methanol or ethanol. Biodiesel has been
regarded as a promising fuel to be able to partly substitute for
conventional fossil diesel. Being environmentally friendly, lipase-
catalyzed biodiesel production is more advantageous over the
alternative chemical approaches due to the mild reaction condi-
tions and easy product recovery [5]. The lipase that is mostly used in
form (Novozym 435).
As compared to those lipases that have been commonly used and
widely investigated, the lipase used in this study has not gained suf-
ficient interest from researchers in terms of its catalytic properties
in nonaqueous reaction media and its applications in biotransfor-
mations. It was produced from a special strain, PF898, of Penicillium
expansum. This P. expansum lipase (PEL) has been crystallized with
a molecular mass of 28 kDa and it hydrolyzes acylglycerols at the
optimal temperature of 34 ◦C and the optimal pH of 9.5 [6]. Accord-
ing to the cloning and sequencing results [7], this enzyme has a
low protein sequence homology compared with other lipases and a
high percentage of hydrophobic residues (51.4%) in the N-terminal
region compared with other typical lipases. Its application in cat-
alyzing biodiesel production in organic solvents from corn oil [8]
and waste oil [9] has been explored. But investigation regarding
its catalytic properties in organic solvents and ionic liquids has not
been reported.
Ionic liquids (ILs) are organic salts remaining as liquids under
ambient temperatures. Interests in using them as a new type of sol-
vents for biotransformations have been increasing rapidly because
they have the ability of presenting excellent enzyme activity, sta-
bility, and selectivity [10,11]. So far different types of enzymes
have been demonstrated to show catalytic activity in ILs, and it
seems evident that enzymes in ILs basically follow the same cat-
alytic mechanism as in organic solvents [12–14]. Therefore, it is
∗
Corresponding author. Tel.: +86 755 2653 4152; fax: +86 755 2653 4274.
E-mail address: zyang@szu.edu.cn (Z. Yang).
1381-1177/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2009.11.014

24
Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
expected that the fundamental rules governing the enzyme activity
in organic solvents may also apply to that in ionic liquids.
tion in ionic liquids. So far there has been only one report in the
literature regarding the potential use of ionic liquids as the reac-
tion media for enzymatic approach of biodiesel production, which
was demonstrated by using an immobilized C. antarctica lipase B
(CALB) [23]. Our work also aimed to investigate the effect of water,
water activity, and salt hydrates on enzyme activity in ionic liquids
as well as in organic solvents, by using PEL as a model enzyme.
It has been well known that water plays an important role in
affecting enzyme activity in organic media [15]. An enzyme is usu-
ally inactive under completely dry conditions, but addition of a
small amount of water to the nonaqueous reaction medium nor-
mally accelerates the enzymatic reactions, and the major function
of the water is to assure the enzyme of its optimal flexibility. The
key determinant of the enzyme activity is the amount of water
that is bound to the enzyme rather than the total water content in
the reaction system. Depending on the solvent polarity, the reac-
tion system requires different amounts of water for optimizing the
enzyme activity. The fact that an enzyme (such as a lipase) usually
presents a higher activity in a less polar and more hydrophobic sol-
vent [16] can also be explained by the lower ability of such a solvent
to strip off the essential water from the enzyme [15]. It is therefore
necessary to control the water level associated with the enzyme,
2. Experimental
2.1. Materials
Crude lipase from P. expansum (PEL) was kindly donated by
Shenzhen Leveking Bio-engineering Co. Ltd., China. p-Nitrophenyl
palmitate (pNPP), p-nitrophenol (pNP), and the ionic liquid
[BMIm][PF6] (≥97%, HPLC) were purchased from Sigma–Aldrich
China Inc. Methyl palmitate, methyl stearate, methyl oleate, methyl
linoleate, methyl linolenate, and methyl heptadecanoate were of
chromatographic purity also from Sigma–Aldrich China Inc. Refined
corn oil was obtained from a local supermarket in Shenzhen, China.
All other reagents used were of analytical grade from local manu-
facturers in China.
which is characterized by the thermodynamic water activity (aw
)
[17]. Two methods are commonly used to achieve a constant water
activity: (1) separate pre-equilibration of substrate solution and
enzyme preparation with a saturated aqueous solution of a salt,
and (2) direct addition of a salt hydrate to the reaction mixture to
act as a water buffer. Originally discussed by Halling [18], the latter
method provides a convenient means of buffering water activity
and has been widely used in the organic solvent system. Although
the effects of water, water activity, and salt hydrates have been
well demonstrated in conventional organic solvents, it is necessary
to assess whether these effects also apply to the enzymes in ionic
liquids. A few recent reports have covered this aspect [12,19–22].
In this study, the catalytic activity of PEL in both organic sol-
vents (hexane) and ionic liquids ([BMIm][PF6]) was investigated
(see Scheme 1 for the enzymatic reactions). The hydrolytic activity
of PEL was determined by following the formation of the product,
p-nitrophenol (pNP), from hydrolysis of p-nitrophenyl palmitate
(pNPP), and the transesterification activity of PEL in these two
reaction media was evaluated by conducting the PEL-catalyzed
methanolysis of corn oil. The latter reaction was also used to
demonstrate the feasibility of utilizing PEL for biodiesel produc-
2.2. Determination of molar extinction coefficients for pNP in
different buffer solutions
A series of pNP solutions (0–0.01 mM) were prepared in a 50 mM
buffer at a particular pH (phosphate buffer, pH 5.0–8.0; Tris–HCl
buffer, pH 7.0–9.0; barbiturate–HCl buffer, pH 7.5–10.0), and their
absorbances were measured at 410 nm. The molar extinction coef-
ficients of pNP were then determined as slopes derived from the
A410 vs. [pNP] plots.
2.3. Activity assay in aqueous solution
The hydrolytic activity of PEL in aqueous solution was deter-
mined according to the method described in [24], with a slight
modification. The PEL solution (∼0.23 U/ml, based on the activity
Scheme 1. PEL-catalyzed hydrolysis and transesterification reactions used in this study.

Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
25
of catalyzing the hydrolysis of pNPP at 37 ◦C in 50 mM phosphate
buffer, pH 6.0) was obtained as the supernatant after dissolution of
1.5 g of the enzyme powders in 10.0 ml phosphate buffer (50 mM,
pH 8.0) followed by centrifugation. A substrate mixture was pre-
pared by mixing 1 ml of 15.0 mM pNPP in isopropanol and 9 ml of
a 50 mM buffer solution (phosphate buffer, pH 5.0–8.0; Tris–HCl
buffer, pH 7.0–9.0; barbiturate–HCl buffer, pH 7.5–10.0) contain-
ing 0.1% (w/v) arabic gum and 0.4% (w/v) Triton X-100. A cuvette
containing 2.4 ml of this substrate mixture was placed in the Phar-
macia Biotech Ultraspec 2000 UV/Vis spectrophotometer, equipped
with a thermostated cell, for pre-equilibration at 37 ◦C, and the
reaction was started by the addition of 0.1 ml of the PEL solution.
The variation of the absorbance at 410 nm of the assay against a
blank without enzyme was monitored for 2–5 min. The reaction
rate was calculated from the slope of the absorbance vs. the time
curve by using the molar extinction coefficient for pNP, obtained as
measured above.
gas chromatogarphy (Thermal Co., Germany) equipped with
a SUPELCOWAX 10 column (30 m × 0.32 mm × 0.25 ␮m, Agilent
Technologies, USA). The column temperature was hold at 160 ◦C
for 2 min, and raised at 10 ◦C/min to 240 ◦C, where it was then
maintained for 10 min. Nitrogen was used as the carrier gas at
1.5 ml/min. Splitless mode was selected. The temperatures of the
injector and the detector were 250 and 260 ◦C, respectively.
3. Results and discussion
3.1. Effect of pH and buffer type on PEL activity in aqueous
solution
The hydrolytic activity of PEL was determined by spectropho-
tometrically following the formation of the product pNP based on
its strong absorbance at 410 nm. It is necessary to emphasize the
need to take account of the variation of pNP’s extinction coefficient,
used in estimation of the enzyme activity, upon the change in the
buffer conditions [25]. In the present study, it was found that the
molar extinction coefficient of pNP is much more sensitive to the
change in the buffer pH (varying from pH 5.0 to 10.0), showing a
sigmoid-curve relation, and less to the change in the buffer type
(Fig. 1).
Based on the molar extinction coefficients of pNP determined
in different buffer systems, our experiments have demonstrated
that the hydrolytic activity of PEL is very much dependent on
both the type and pH of the buffer used (Fig. 2). Among the three
buffer types tested, phosphate buffer promoted the highest activity,
whereas barbiturate buffer the lowest. In both phosphate and Tris
buffers, the hydrolytic activity of PEL decreased with an increase in
the buffer pH within the pH ranges tested (pH 5–8 for phosphate
buffer and pH 7–9 for Tris buffer). In barbiturate buffer, however,
there seemed to be an abnormal V-shaped relationship between
enzyme activity and the buffer pH within the pH range of 7.5 and
10. The fact that lipase presented different activities in different
buffers does not seem to be uncommon. For instance, activities of
lipase from Penicillium aurantiogriseum have also been reported to
be higher in phosphate buffer than in Tris–HCl buffer [26]. In our
very recent paper regarding the study of Hofmeister effects on alka-
line phosphatase [25], we have also demonstrated that the enzyme
presented a variant activity in the same DEA buffer, depending
on the type and concentration of the salt present. It can be cer-
tain that pH is not the only factor controlling the enzyme activity,
but instead, the species present in the buffer solution, whether in
the form of ions or not, may also be responsible for regulating the
2.4. Activity assay in hexane
According to the method described in [24], a typical hydroly-
sis reaction was started by the addition of 160 mg PEL to 2.0 ml
hexane containing 10 mM pNPP in a 5-ml capped bottle, which
was placed in a shaker and incubated at 37 ◦C with agitation of
200 rpm. Occasionally, a certain amount of water or an inorganic
salt hydrate (0.5 g) was added to control the water content in the
reaction system. Samples (0.1 ml) were taken at intervals (typ-
ically every 10 min) and mixed thoroughly with 2.0 ml of 0.1 M
NaOH solution. After centrifugation, the bottom phase was taken
for absorbance reading at 410 nm. A blank reaction (without addi-
tion of the enzyme) was also followed in parallel. The reaction
rate was calculated from the slope of the absorbance (subtract-
ing the one for the blank reaction) vs. the time curve by using
the molar extinction coefficient for pNP determined in 0.1 M NaOH
as described above. For transesterification of pNPP with methanol,
enzyme powders (160 mg) were added to 2.0 ml hexane containing
10 mM pNPP and 10 mM methanol to start the reaction.
2.5. Activity assay in [BMIm][PF6]
The hydrolytic activity of PEL in this ionic liquid was determined
basically following the method described above for the hexane sys-
tem. But due to the low solubility of pNPP in [BMIm][PF6], the
substrate solution was prepared by first dissolving pNPP in iso-
propanol in a concentration of 25 mM, which was then diluted with
[BMIm][PF6] to a final concentration of 2.3 mM.
2.6. Methanolysis of corn oil
In a typical experiment, the reaction was carried out in a 25 ml
capped flask containing 1 g of the corn oil, two molar equivalent
of methanol, 2 ml of a given solvent, and 100 mg of the PEL. The
reaction flask was placed in a shaker at 40 ◦C with agitation of
220 rpm. Prior to use, all the solvents have been dried over molec-
ular sieves, and no additional water control was conducted during
the reactions. Periodically, samples of 50 ␮l were withdrawn and
centrifuged at 10,000 rpm for 10 min. Then 5 ␮l from the super-
natant was taken and diluted with 295 ␮l of hexane, followed by
addition of 300 ␮l of methyl heptadecanoate (1 mg/ml, as an inter-
nal standard). This mixture was then ready for gas chromatographic
analysis.
2.7. GC analysis
The fatty acyl methyl esters (FAMEs) produced from lipase-
catalyzed methanolysis of corn oil was assayed with a Trace2000
Fig. 1. Dependence of molar extinction coefficient of pNP on both buffer type and
buffer pH. Concentrations for all the buffers were 50 mM.

26
Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
solvent, hexane, as an example. Candida rugosa lipase produced a
lower yield in the esterification of (R,S)-2-substituted-propanoic
acids with 1-butanol in [BMIm][PF6] than in hexane [19], whereas
the same enzyme in [BMIm][PF6] catalyzed a transesterification
reaction between methyl methacrylate and 2-ethyl-1-hexanol at an
initial rate that is 1.5 times faster than the reaction in hexane [31].
For this latter reaction, another lipase from C. antarctica, adsorbed
onto a macroporous acrylic support (Novozym 435), was active in
organic solvents such as hexane and acetonitrile but totally inactive
in all the ionic liquids tested including [BMIm][PF6] [31]; whereas
the same CALB, when immobilized on a different acrylic resin, was
active in [BMIm][PF6], but with an initial rate higher or lower than
that obtained in hexane depending on the water content in the reac-
tion system [12]. The immobilized CALB (Novozym 435) was also
active in catalyzing an esterification of geraniol with acetic acid
in [BMIm][PF6], but the reaction rate was lower than in hexane
[20]. Very recently, de Diego et al. [22] have compared the cat-
alytic efficiency of four free lipases from different microbes (i.e., C.
antarctica lipase A, CALA; C. antarctica lipase B, CALB; Thermomyces
lanuginosus lipase, TLL; and Rhizomucor miehei lipase, RML) in five
different ionic liquids and in hexane. The transesterification activi-
ties of CALB, TLL, and RML were 1.5-, 6.1-, and 1.4-fold increased in
[BMIm][PF6] as compared to that obtained in hexane, respectively,
while CALA was less active in the ionic liquid. All these different
results can probably be attributed to differences in the enzyme
source and preparation, the reaction type, the substrate, and also
the water content in the reaction system. Therefore, our experi-
mental result about PEL offers a new type of lipases that exhibits
a significantly improved activity in ionic liquids as compared to
conventional organic solvents.
Fig. 2. Effect of pH and buffer type on the hydrolytic activity of PEL in aqueous
solution. Concentrations for all the buffers were 50 mM.
enzyme activity, presumably by interacting with either the enzyme
or the substrate, thereby affecting the enzyme conformation and
the substrate binding.
Although P. expansum has been taken as a major source for pro-
ducing alkaline lipases and the lipase from the strain PF898 of this
genus has been shown to hydrolyze acylglycerols at the optimal pH
of 9.5 [6], a neutral [27,28] or slightly basic [26] pH optimum for
lipases produced from a few other strains of the genus Penicillium
has been reported. The pH dependence of PEL shown in Fig. 2 seems
rather abnormal.
3.3. Effect of water content, water activity, and salt hydrates
3.2. Comparing the hydrolytic activity of PEL in both hexane and
[BMIm][PF6]
Fig. 4 confirms that the hydrolytic activity of PEL in [BMIm][PF6]
was sensitive to the amount of water added. For both solvents
[BMIm][PF6] and hexane, there is a similar bell-shaped relation-
ship between enzyme activity and amount of water added to the
reaction system, although the reaction rates obtained in the IL are
much higher. The enzyme requires a greater optimal water con-
tent in the IL system (2.3%, v/v) than that for the hexane system
(0.2%, v/v). Water is a co-substrate for the PEL-catalyzed hydrolysis
reaction, and this can partly account for the initial enhancement
It is interesting that even under a low-water condition, PEL holds
a strong power of catalyzing the hydrolysis of pNPP rather than
the transesterification of the same substrate with methanol. The
enzyme was fairly active when catalyzing the hydrolysis of pNPP
in hexane. When the enzyme powders were suspended in hexane
containing 10 mM pNPP and 10 mM methanol, the initial rate of
pNP formation was 95% of that obtained in the absence of methanol;
and this rate was dropped significantly with gradual addition of
more methanol, presumably due to the methanol-induced inacti-
vation. Therefore, in our study the hydrolytic activity of PEL was
compared in both hexane and [BMIm][PF6].
Fig. 3 depicts the proportional increase of the initial reac-
tion rates of PEL-catalyzed pNPP hydrolysis upon addition of the
enzyme in the two solvent systems. When tested at the same sub-
strate concentration (2.3 mM), the hydrolytic activity of PEL (in the
unit of ␮mol/h/mg) in [BMIm][PF6] was 12.1 times as high as that
in hexane. A low substrate concentration of 2.3 mM was selected
simply because of the low solubility of pNPP in [BMIm][PF6]. A few
enzymes have been reported to show higher activity in ionic liquids
than in conventional organic solvents [22,29,30], but an enhance-
ment in activity for more than 10-fold is rarely reported.
Lipase is the type of enzymes that has been extensively inves-
tigated in ionic liquids and has been reported to present activities
in such reaction media that is greater than, or comparable to, those
obtained in conventional organic solvents. For instance, C. antarc-
tica lipase B (CALB), whether free or immobilized on three different
carriers, exhibited improved ascorbyl oleate synthesis in all the
six ionic liquids tested, with an initial rate up to 9-fold higher
as compared to the reaction conducted in acetone [21]. However,
the results are sometimes contradictory. Take the comparison of
lipase activity in the ionic liquid, [BMIm][PF6], and in the organic
Fig. 3. Comparison of PEL’s hydrolytic activity in hexane and in [BMIm][PF6]. Dif-
ferent amounts of enzyme powders were added to the reaction system containing
2.3 mM pNPP in 2.0 ml of either [BMIm][PF6] or hexane, and the reactions were
carried out at 37 ◦C with agitation of 240 rpm. Both solvents were used directly as
supplied and no additional water was added.

Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
27
enzyme in hexane was most active at very low-water activity (i.e.,
in the presence of CaCl2·2H2O and MgCl2·6H2O, with a water activ-
ity of 0.043 and 0.041 at 30 ◦C, respectively), and then its activity
showed a general trend, although somewhat scattering, of decreas-
ing upon an increase in the water activity (Fig. 5B). This trend is
clearly confirmed by a comparison of the activity data obtained in
hexane with addition of three different hydrate forms of a same salt,
Na2HPO4 (i.e., Na2HPO4·2H2O, Na2HPO4·7H2O, Na2HPO4·12H2O,
see the points b, c, and d on Fig. 5B). In fact, it is not uncommon for a
lipase to be active at a low aw. R. miehei lipase has been reported to
remain highly active at water activity below 0.0001 [34]. However,
the activity of PEL in the IL [BMIm][PF6] did not show an obvious
Fig. 4. Dependence of PEL’s hydrolytic activity in hexane and in [BMIm][PF6] on the
amount of water added to the reaction system. Different amounts of water were
added to the reaction system containing 2.0 ml of 2.3 mM pNPP in the solvent and
160 mg enzyme powders, and the reactions were carried out at 37 ◦C with agitation
of 240 rpm. Both solvents were dried over molecular sieves prior to use.
of the enzyme’s hydrolytic activity with an increase in the water
content. Generally, water also plays an important role in affecting
the enzyme activity in nonaqueous media [15]: essential amount
of water may activate the enzyme by increasing the polarity and
structural flexibility of the enzyme’s active site; but too much
water is harmful to the enzyme by facilitating enzyme aggregation,
thus diminishing the substrate diffusion and eventually leading
to enzyme inactivation. Lumps of enzyme powders have indeed
been observed in our experiments, when a high water content was
applied in the IL. One may notice that the enzyme activity obtained
in Fig. 4 was approximately half of what was obtained in Fig. 3, and
this can be attributed to the relatively low storage stability of the
enzyme: the two experiments were done with more than half a year
apart. Research will be undertaken in our laboratory to improve the
enzyme’s storage stability.
A
previous report by Ulbert et al. [19] has shown
that the optimal water content for C. rugosa lipase var-
ied according to the solvent used, following the order of
[BMIm][BF4] > [BMIm][PF6] > [ONIm][PF6] > toluene > hexane;
which is consistent with the decreasing order of solvent polarity
(the log P values for the above solvents are −2.44, −2.38, −2.19,
2.5, and 3.5, respectively). A more polar solvent has a higher
tendency of stripping off the essential water from the enzyme.
Ionic liquids are highly polar [32,33], and their greater ability to
retain water has been demonstrated by de Diego et al. [22], who
investigated the effect of water content on the water activity of the
reaction system of lipase-catalyzed butyl propionate synthesis in
the ionic liquid [CPMA][MS] at 40 ◦C: the water activity was slowly
increased from 0.2 to about 0.5 when the water content was raised
abruptly from 1 to 8% (v/v). It is therefore understandable that the
IL system requires a higher water content for an optimal enzyme
activity.
Recently, Russell’s group has conducted a successful demonstra-
tion of using salt hydrate pairs to control water activity for enzyme
catalysis in ionic liquids [12]. In our experiment, 13 salt hydrates
with different water activity values at 30 ◦C ranging from 0.041 to
0.85 were selected for the study of their effect on controlling the
water activity as well as the hydrolytic activity of PEL in hexane
and in [BMIm][PF6]. Considering the extremely low activity of PEL
in hexane, a higher substrate concentration was used in this reac-
tion medium (10 mM, instead of 2.3 mM used in [BMIm][PF6]). The
Fig. 5. Correlation between PEL’s hydrolytic activity in [BMIm][PF6] (A) and
hexane (B) and the water activity (aw
) offered by the salt hydrate added to
the reaction system at 30 ◦C. Different salt hydrates (0.5 g) were added to the
reaction system containing 2.0 ml of pNPP solution (10.0 mM in hexane and
2.3 mM in [BMIm][PF6]) and 160 mg enzyme powders, and the reactions were
carried out at 30 ◦C with agitation of 240 rpm. Both solvents were dried over
molecular sieves prior to use. The salt hydrates used and their water activities
at 30 ◦C (obtained from [18] and listed in brackets) are: MgCl2·6H2O (0.041),
CaCl2·2H2O (0.043), NaAc·3H2O (0.3), CuSO4·5H2O (0.35), K4Fe(CN)6·3H2O (0.48),
Na4P2O7·10H2O (0.52), Na2B4O7·10H2O (0.65), ZnSO4·7H2O (0.68), Na2CO3·10H2O
(0.79), Na2SO4·10H2O (0.83), Na2HPO4·2H2O (0.177), Na2HPO4·7H2O (0.65), and
Na2HPO4·12H2O (0.85). The data points labeled with the letter ‘a’ refers to the reac-
tion rate obtained without addition of any salt hydrate, whereas those labeled with
‘b’, ‘c’, ‘d’, ‘e’, and ‘f’ represent the rates obtained with addition of the salt hydrates
Na2HPO4·2H2O, Na2HPO4·7H2O, Na2HPO4·12H2O, CaCl2·2H2O, and MgCl2·6H2O,
respectively.

28
Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
correlation with the water activities presented by the salt hydrates
(Fig. 5A), and the data for the three hydrate forms of Na2HPO4 (see
points b, c, and d on Fig. 5A) has even shown the opposite trend, as
compared to Fig. 5B. The possible reason for this unexpected result
might be related to the specific salt effect.
As the activity of an enzyme in nonaqueous media depends on its
hydration, theenzyme is activewhenthereis amicroaqueous phase
associated with it [15]. The salt hydrates added to the reaction
system can partially penetrate and dissolve in this microaqueous
phase and dissociate into individual cations and anions to function
on the enzyme. It has long been realized that different ions affect
the enzyme performance in aqueous solution following a common
order known as the Hofmeister series [35]. Although the mecha-
nism regarding this Hofmeister effect is still not clearly known, it
is believed to be correlated with the kosmotropic/chaotropic prop-
erties of the ions, which can be characterized by the Jones–Dole
in Fig. 5A. Addition of the higher hydrate form of this salt provides
a higher aw, thus introducing more water onto the enzyme sur-
face where more salt is dissolved and ionized to get direct contact
2−
with the enzyme molecule. As the anion HPO4
is a strong kos-
motrope, it is not surprising for lipase to be more active in the ionic
liquid with addition of a higher hydrate form of Na2HPO4. Inter-
estingly, when PEL was suspended in hexane, the enzyme activity
was well correlated with aw (Fig. 5B) but not with the (B –B+) value
−
(Fig. 6B) of the salt hydrate added, suggesting that in this case, the
water buffering effect offered by the salt hydrates is predominant,
as compared to their specific ion effect.
Actually, the different results obtained in the two different sol-
vent systems can be attributed to the different water requirements
by the solvents: [BMIm][PF6] is a highly polar ionic liquid that
requires a large amount of water, some of which may be attracted
by the enzyme to form a thin layer surrounding the enzyme
molecule; whereas hexane is such a hydrophobic and nonpolar
solvent that there may be only a small amount of water avail-
viscosity B coefficients (B+ for cations and B for anions, avail-
−
able in [36]) (for recent reviews, see [37–39]). In a recent study
of salt effects on the hydrolytic activity of C. rugosa lipase in aque-
ous solution, Salis et al. [40] found that NaSCN is a strong inhibitor
to the enzyme, Na2SO4 acts as an activator, whereas NaCl shows
a quasi-neutral behavior. This has been attributed to the role of
anions in modifying both the surface pH and the tertiary struc-
ture of the enzyme. In particular, a chaotropic anion (such as SCN−)
has a higher tendency of binding to the enzyme surface. As a
result, the two residues (Glu341 and His449) involved in the catalytic
triad mechanism are protonated and hence cannot assist the other
residue Ser209 in the nucleophilic attack of the substrate, leading
to inhibition to the lipase. Previously, the same research group has
already observed a similar effect of anions on the hydrolytic activ-
ity of Aspergillus niger lipase in aqueous solution [41]: the enzyme
activity varied in the order of Br− > Cl− ∼ NO3− > ClO4−, basically
following the Hofmeister series.
If there is indeed a thin layer of water surrounding the enzyme
molecule for its catalysis, it is reasonable to speculate that the
salt hydrates added to the reaction system offers dual func-
tions in affecting the enzyme activity: they can not only exist
as intact molecules in the nonaqueous reaction medium to con-
trol the hydration level of the enzyme (the water activity) via
their hydration/dehydration equilibria; but also penetrate into the
microaqueous phase surrounding the enzyme molecule and dis-
solve and dissociate into their corresponding cations and anions,
exhibiting their specific ion effect (the Hofmeister effect) on the
enzyme. Correspondingly, a reasonable V vs. aw or V vs. (B –B+)
−
relationship would be expected for the former or latter salt hydrate
effect, respectively. For our case of PEL in the ionic liquid, although
the enzyme did not show an obvious V–aw correlation (Fig. 5A), it
did show a clear trend of increasing activity upon an increase in the
(B –B+) value of salts added (Fig. 6A). This is in line with the above
−
literature reports regarding the higher lipase activity in aqueous
solution in the presence of more kosmotropic anions (with higher
2−
and HPO4−) and can be explained by the
B
−
values, such as SO
4
specific ion effect on the surface pH, structure, and catalytic mech-
anism of the enzyme [40]. Our recent study about the Hofmeister
effect on the activity of alkaline phosphatase has provided further
evidence for the idea that kosmotropic cations and anions have
opposite effect on activity [25], supporting the viewpoint that the
(B –B+) value of a salt is a better parameter for quantifying the salt
−
effect [42]. The experimental results obtained in this study have
therefore suggested that for PEL suspended in [BMIm][PF6], it is the
Hofmeister effect rather than the water buffering effect, resulted
from addition of salt hydrates, that has a stronger impact on the
enzyme activity. In fact, the dual functions of the salt hydrates may
work together to affect the enzyme catalysis, and this may offer a
plausible explanationfor thehigherlipaseactivity inthe ionicliquid
containing a higher hydrate form of Na2HPO4, as can be observed
Fig. 6. Correlation between initial rate of PEL-catalyzed hydrolysis of pNPP in
[BMIm][PF6] (A) and hexane (B) and the (B –B+) value of the salt hydrate added
−
to the reaction system. Reaction rates obtained from Fig. 5 were used in this pre-
sentation, and the viscosity B coefficients for both cations (B+) and anions (B ) of
−
the salt hydrates added were obtained from [36] to work out the (B –B+) value for
−
each salt: MgCl2 (−0.39), CaCl2 (−0.289), NaAc (0.161), CuSO4 (−0.17), K4Fe(CN)6
(0.351), ZnSO4 (−0.163), Na2CO3 (0.209), Na2SO4 (0.121), and Na2HPO4 (0.297).

Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
29
able for the enzyme, far insufficient to form a monolayer around
the enzyme molecule. Indeed, the optimal water activity level for
immobilized C. antarctica lipase B to catalyze the esterification of
geraniol in hexane is 0.1, much lower than the value of 0.6 for
the enzyme suspended in [BMIm][PF6] [20]. This is different from
what was observed in another study: the initial rate of esterifica-
tion catalyzed by an immobilized lipase (lipozyme) showed similar
dependence on water activity (with the same optimal aw = 0.5) in
different organic solvents ranging in polarity from pentan-3-one to
hexane [43]. It is worth mentioning that for the above two refer-
ences, water activity was acquired by using the pre-equilibration
method without the involvement of salt hydrates. The fact that
lipase requires a higher aw when suspended in [BMIm][PF6] may be
related to the extremely high polarity and the unique ionic nature
of the ionic liquid.
3.4. Effect of reaction temperature
Variation of enzyme activity in both nonaqueous solvents upon
an increase in the reaction temperature is shown in Fig. 7. A
higher substrate concentration of 10 mM was used in hexane
again because of the extremely low activity of PEL in this reac-
tion medium. Comparing the two sets of initial conversion rates
(i.e., percentages of the substrate converted per hour) supported
the above result that PEL presented a higher activity in [BMIm][PF6]
than in hexane. PEL’s hydrolytic activity in the ionic liquid was opti-
mal at 70 ◦C, higher than the optimal temperature of 34 ◦C obtained
in the hydrolysis of acylglycerols [6]. Higher reaction temperatures
could not be studied in hexane due to its low boiling point (69 ◦C),
after which the solvent evaporates seriously. The data in Fig. 7 were
used to calculate the activation energy of the enzyme, which is
25.8 kJ/mol in hexane and 38.9 kJ/mol in [BMIm][PF6]. It is interest-
ing that although the enzyme was more active in the ionic liquid, its
activation energy in this system was higher. A plausible explana-
tion might be related to the water content in the solvent available
for the hydrolytic reaction.
Fig. 8. PEL-catalyzed methanolysis of corn oil in [BMIm][PF6], tert-butanol, hexane,
and solvent-free system. Please refer to Section 2.6 for experimental details.
production. Fig. 8 shows the production of FAMEs in [BMIm][PF6],
tert-butanol, hexane, and in the solvent-free system. Obviously, the
highest yield was achieved in [BMIm][PF6], whereas hexane was
the solvent to promote the lowest conversion. The yield obtained
in the ionic liquid was 3.6-, 5.0-, and 72.5-fold as high as the one
obtained in tert-butanol, solvent-free system, and hexane, respec-
tively. The extremely low production yield obtained in hexane
may be related to the low ability of hexane to solubilize glycerol
(a by-product) and methanol (a co-substrate). A major problem
for biocatalytic biodiesel production is the accumulation of these
two materials on the surface of the enzyme [8]. Glycerol is so
viscous that it blocks the active site of the enzyme in access
with the substrate; while methanol has a strong power of inac-
tivating enzymes. This latter situation indeed occurred to PEL, as
was demonstrated in our experiments mentioned above. Taking
these into consideration, the ionic liquid [BMIm][PF6] shows its
potential advantage as a solvent for PEL-catalyzed biodiesel pro-
duction.
3.5. PEL-catalyzed methanolysis of corn oil for biodiesel
production
PEL has been utilized by Zong’s group in catalyzing biodiesel
production from corn oil [8] and waste oil [9]. A series of con-
ventional organic solvents (petroleum ether, hexane, cyclohexane,
tert-butanol, and tert-amyl alcohol) have been tested, and tert-amyl
alcohol was found to be the optimal reaction medium for PEL-
mediated biodiesel production presumably due to their good ability
of solubilizing the co-substrate, methanol, and the by-product,
glycerol. So far there has been only one paper dealing with lipase-
catalyzed biodiesel production in ionic liquids [23]. Immobilized C.
antarctica lipase was used to catalyze methanolysis of soybean oil
for biodiesel production in 23 different ionic liquids. [EMIm][TfO]
was found to be the solvent in which the highest yield was achieved,
followed by tert-butanol, a commonly used solvent for lipase-
catalyzed biodiesel production. The conversion in [BMIm][PF6] was
even poorer than that obtained in the solvent-free system.
It has to be mentioned that although our reaction condi-
tions have not been optimized, the yield of FAMEs obtained
in [BMIm][PF6] was comparable to the results presented in the
above three references. Our results of higher production yield in
[BMIm][PF6], as compared to that obtained in the other solvent
systems, have clearly demonstrated that PEL is a promising lipase
used for biodiesel production and that [BMIm][PF6] is a promising
reaction medium for such biotransformations. Our study has also
indicated that as compared to its activity in hexane, PEL is much
more active in [BMIm][PF6], not only in catalyzing hydrolysis but
also in catalyzing transesterification reactions.
The transesterification activity of PEL was investigated by con-
ducting the PEL-catalyzed methanolysis of corn oil for biodiesel
Fig. 7. Dependence of PEL’s hydrolytic activity in hexane and in [BMIm][PF6] on
reaction temperature. Enzyme powders (160 mg) were added to 2 ml of hexane
or [BMIm][PF6] containing 10 or 2.3 mM of pNPP, respectively, and the reaction
mixtures were incubated at different temperatures with agitation of 240 rpm. The
conversion rate refers to the percentage of the substrate being converted per hour.
Both solvents were used directly without drying pretreatment.

30
Z. Yang et al. / Journal of Molecular Catalysis B: Enzymatic 63 (2010) 23–30
4. Conclusions
[5] U. Shimada, Y. Watanabe, A. Sugihara, Y. Tominaga, J. Mol. Catal. B: Enzym. 17
(2002) 133–142.
[6] C. Bian, Y. Yuan, L. Lin, J. Lin, X. Shi, X. Ye, Z. Huang, M. Huang, Biochim. Biophys.
Acta 1752 (2005) 99–102.
[7] L. Lin, B.F. Xie, G.Z. Yang, Q.Q. Shi, Q.Y. Lin, L.H. Xie, X.F. Wu, S.G. Wu, Chin. J.
Biochem. Mol. Biol. 18 (2002) 32–37.
This work has clearly shown that PEL offers a significantly
enhanced activity, both in hydrolysis and transesterification reac-
tions, when assayed in the ionic liquid [BMIm][PF6] relative to
the organic solvent hexane. The higher hydrolytic activity in the
ionic liquid was well demonstrated in the experiments investigat-
ing the effects of enzyme amount (Fig. 3), water content (Fig. 4),
water activity (Fig. 5), and reaction temperature (Fig. 7). We have
also demonstrated the potential use of ionic liquids as the reaction
media for PEL-catalyzed biodiesel production. It is worth noting
that although [BMIm][PF6] is so far still one of the most commonly
used ILs, salts containing PF6− as the anion may be potentially haz-
ardous because their hydrolysis might induce the formation of HF
[44]. This study is aimed to demonstrate that the catalytic activity
of PEL can be significantly improved in ionic liquids. For our later
research regarding IL effects on PEL’s activity and stability, ‘greener’
ionic liquids will be selected.
[8] N. Li, H. Wu, M. Zong, W. Lou, Chin. J. Catal. 28 (2007) 333–338.
[9] N.-W. Li, M.-H. Zong, H. Wu, Process Biochem. 44 (2009) 685–688.
[10] Z. Yang, W. Pan, Enzyme Microb. Technol. 37 (2005) 19–28.
[11] F. van Rantwijk, R.A. Sheldon, Chem. Rev. 107 (2007) 2757–2785.
[12] J.A. Berberich, J.L. Kaar, A.J. Russell, Biotechnol. Prog. 19 (2003) 1029–1032.
[13] T. de Diego, P. Lozano, S. Gmouh, M. Vaultier, J.L. Iborra, Biotechnol. Bioeng. 88
(2004) 916–924.
[14] Z. Yang, Y.-J. Yue, M. Xing, Biotechnol. Lett. 30 (2008) 153–158.
[15] Z. Yang, A.J. Russell, in: A.M.P. Koskinen, A.M. Klibanov (Eds.), Enzymatic Reac-
tions in Organic Media, Blackie Academic & Professional, New York, 1996, pp.
43–69.
[16] C. Laane, S. Boeren, K. Vos, C. Veeger, Biotechnol. Bioeng. 30 (1987) 81–87.
[17] P.J. Halling, Trends Biotechnol. 7 (1989) 50–52.
[18] P.J. Halling, Biotechnol. Tech. 6 (1992) 271–276.
[19] O. Ulbert, T. Fráter, K. Bélafi-Bakó, L. Gubicza, J. Mol. Catal. B: Enzym. 31 (2004)
39–45.
[20] B. Barahona, P.H. Pfromm, M.E. Rezac, Biotechnol. Bioeng. 93 (2006)
318–324.
With regard to the water effect, PEL does require an essential
amount of water for its activity in [BMIm][PF6] as well as in hex-
ane, and the significantly greater polarity of the ionic liquid can
account for the much higher optimal water content it requires.
The essential water can be introduced into the reaction system by
direct addition of water or a salt hydrate. When added to the non-
aqueous reaction system, the salt hydrate may play two roles in
affecting the enzyme activity: water buffering effect and specific
ion effect. To the best of our knowledge, this latter effect caused by
salt hydrates has not been stressed in any studies before. Depend-
ing on the water requirement by the solvent, the former or latter
effect may be dominating. Therefore, care must be taken when try-
ing to use salt hydrates for controlling the water activity in the
biocatalytic ionic liquid system.
[21] M. Adamczak, U.T. Bornscheuer, Process Biochem. 44 (2009) 257–261.
[22] T. de Diego, P. Lozano, M.A. Abad, K. Steffensky, M. Vaultier, J.L. Iborra, J. Biotech-
nol. 140 (2009) 234–241.
[23] S.H. Ha, M.N. Lan, S.H. Lee, S.M. Hwang, Y.-M. Koo, Enzyme Microb. Technol. 41
(2007) 480–483.
[24] G. Pencreac’h, J.C. Baratti, Enzyme Microb. Technol. 18 (1996) 417–422.
[25] Z. Yang, X.-J. Liu, C. Chen, P.J. Halling, Biochim. Biophys. Acta (in press).
[26] V.M.G. Lima, N. Krieger, D.A. Mitchell, J.D. Fontana, Biochem. Eng. J. 18 (2004)
65–71.
[27] T. Mase, Y. Matsumiya, A. Matsuura, Biosci. Biotechnol. Biochem. 59 (1995)
329–330.
[28] M.A.F. Costa, R.M. Peralta, J. Basic Microbiol. 39 (1999) 11–15.
[29] M. Eckstein, M. Sesing, U. Kragl, P. Adlercreutz, Biotechnol. Lett. 24 (2002)
867–872.
[30] S. Lutz-Wahl, E.-M. Trost, B. Wagner, A. Manns, L. Fischer, J. Biotechnol. 124
(2006) 163–171.
[31] J.L. Kaar, A.M. Jesionowski, J.A. Berberich, R. Moulton, A.J. Russell, J. Am. Chem.
Soc. 125 (2003) 4125–4131.
[32] A.J. Carmichael, K.R. Seddon, J. Phys. Org. Chem. 13 (2000) 591–595.
[33] F. van Rantwijk, R.M. Lau, R.A. Sheldon, Trends Biotechnol. 21 (2003)
131–138.
Acknowledgement
[34] R.H. Valivety, P.J. Halling, A.R. Macrae, FEBS Lett. 301 (1992) 258–260.
[35] F. Hofmeister, Arch. Exp. Pathol. Pharmacol. 24 (1888) 247–260.
[36] H.D.B. Jenkins, Y. Marcus, Chem. Rev. 95 (1995) 2695–2724.
[37] P. Bauduin, A. Renoncourt, D. Touraud, W. Kunz, B.W. Ninham, Curr. Opin.
Colloid Interface Sci. 9 (2004) 43–47.
Financial support from the Foundation of the Science and Tech-
nology Program of Shenzhen (Grant No. JC200903120016A) is
gratefully acknowledged.
[38] H. Zhao, J. Mol. Catal. B: Enzym. 37 (2005) 16–25.
[39] Z. Yang, J. Biotechnol. 144 (2009) 12–22.
[40] A. Salis, D. Bilanicˇová, B.W. Ninham, M. Monduzzi, J. Phys. Chem. B 111 (2007)
1149–1156.
References
[41] M.C. Pinna, A. Salis, M. Monduzzi, B.W. Ninham, J. Phys. Chem. B: Lett. 109
(2005) 5406–5408.
[42] J.P. Lindsay, D.S. Clark, J.S. Dordick, Biotechnol. Bioeng. 85 (2004)
553–560.
[43] R.H. Valivety, P.J. Halling, A.R. Macrae, Biochim. Biophys. Acta 1118 (1992)
218–222.
[44] R.P. Swatloski, J.D. Holbrey, R.D. Rogers, Green Chem. 4 (2003) 361–363.
[1] P.D. de María, C. Carboni-Oerleman, B. Tuin, G. Bargeman, A. van der Meer, R.
van Gemert, J. Mol. Catal. B: Enzym. 37 (2005) 36–46.
[2] P. Reis, K. Holmberg, H. Watzke, M.E. Leser, R. Miller, Adv. Colloid Interface Sci.
147–148 (2009) 237–250.
[3] R. Sharma, Y. Chisti, U.C. Banerjee, Biotechnol. Adv. 19 (2001) 627–662.
[4] J.M. Marchetti, V.U. Miguel, A.F. Errazu, Renew. Sust. Energy Rev. 11 (2007)
1300–1311.