Impact of ionic liquid physical properties on lipase activity and stability

Article abstract of DOI:10.1021/ja028557x

Lipase activity and stability was investigated in dialkylimidazolium and pyrrolidinium-based ionic liquids with a variety of anions including hexafluorophosphate, acetate, nitrate, methanesulfonate, trifluoroacetate, and trifluoromethylsulfonate. The initial rate of lipase-catalyzed transesterification of methyl methacrylate in these ionic liquids and several organic solvents was examined as well as the polytransesterification of divinyl adipate and 1,4-butanediol. Free lipase (Candida rugosa) catalyzed the transesterification of methyl methacrylate in 1-butyl-3-methylimidazolium hexafluorophosphate at a rate 1.5 times greater than in hexane. However, no detectable activity was observed in all the hydrophilic ionic liquids studied. Methods of enzyme stabilization including adsorption, PEG-modification, and immobilization in polyurethane foam were ineffective in improving enzymatic activity in the hydrophilic ionic liquids. Polytransesterifications performed in 1-butyl-3-methylimidazolium hexafluorophosphate using Novozym 435 produced polyesters with weight average molecular weights limited to 2900 Da due to precipitation of the polymer. Solvatochromic studies and partition coefficient measurements suggest that ionic liquids are more polar and hydrophilic than organic solvents such as hexane, acetonitrile, and tetrahydrofuran. Stability studies indicate that lipases exhibit greater stability in ionic liquids than in organic solvents including hexane.

Full text of DOI:10.1021/ja028557x

Published on Web 03/13/2003
Impact of Ionic Liquid Physical Properties on Lipase Activity
and Stability
Joel L. Kaar,†,‡ Anita M. Jesionowski, Jason A. Berberich, Roger Moulton, and
‡
†,§
|
,
†,‡
Alan J. Russell*
Contribution from the McGowan Institute for RegeneratiVe Medicine, Suite 200,
1
00 Technology DriVe, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15219,
Department of Chemical and Petroleum Engineering, 1249 Benedum Hall,
UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15261, Department of Bioengineering,
7
49 Benedum Hall, UniVersity of Pittsburgh, Pittsburgh, PennsylVania 15261, and
SACHEM Inc., 821 East Woodward Street, Austin, Texas 78704
Received September 16, 2002; E-mail: russellaj@msx.upmc.edu
Abstract: Lipase activity and stability was investigated in dialkylimidazolium and pyrrolidinium-based ionic
liquids with a variety of anions including hexafluorophosphate, acetate, nitrate, methanesulfonate,
trifluoroacetate, and trifluoromethylsulfonate. The initial rate of lipase-catalyzed transesterification of methyl
methacrylate in these ionic liquids and several organic solvents was examined as well as the polytrans-
esterification of divinyl adipate and 1,4-butanediol. Free lipase (Candida rugosa) catalyzed the trans-
esterification of methyl methacrylate in 1-butyl-3-methylimidazolium hexafluorophosphate at a rate 1.5 times
greater than in hexane. However, no detectable activity was observed in all the “hydrophilic” ionic liquids
studied. Methods of enzyme stabilization including adsorption, PEG-modification, and immobilization in
polyurethane foam were ineffective in improving enzymatic activity in the hydrophilic ionic liquids.
Polytransesterifications performed in 1-butyl-3-methylimidazolium hexafluorophosphate using Novozym 435
produced polyesters with weight average molecular weights limited to 2900 Da due to precipitation of the
polymer. Solvatochromic studies and partition coefficient measurements suggest that ionic liquids are more
polar and hydrophilic than organic solvents such as hexane, acetonitrile, and tetrahydrofuran. Stability
studies indicate that lipases exhibit greater stability in ionic liquids than in organic solvents including hexane.
Introduction
Thermal stability and the nonexistence of vapor pressure,
which eliminates volatile organic compound (VOC) emissions,
make ionic liquids an environmentally attractive alternative to
conventional organic solvents. Recoverability and recyclability
also make ionic liquids a practical solution to industrial
environmental concerns.1,2 That said, little is known about the
biological consequences of increased concentrations of ionic
liquids in the environment.
As solvents for chemical processing, ionic liquids exhibit
excellent physical characteristics including the ability to dissolve
polar and nonpolar organic, inorganic, and polymeric com-
pounds.1,3 The chemical and physical properties of ionic liquids
may be modified by altering the cation, anion, and attached
substituents. This feature plays a key role in manipulating the
solvent properties of an ionic liquid, which allows ionic liquids
to be designed for specific reaction systems.3
The use of ionic liquids to replace organic solvents in
biocatalytic processes has recently gained much attention. Cull
-
-
Figure 1. Anion X1 is available as hexafluorophosphate [PF ], acetate
6
-
-
-
-
[
CH3CO2 ], nitrate [NO3 ], and trifluoroacetate [CF3CO2 ]. Anion X2 is
available as hexafluorophosphate [PF6 ], acetate [CH3CO2 ], nitrate [NO3 ],
trifluoroacetate [CF3CO2 ], trifluoromethylsulfonate [CF3SO3 ], and meth-
anesulfonate [CH3SO3 ].
-
-
-
-
-
-
et al.4 used the ionic liquid 1-butyl-3-methylimidazolium
hexafluorophosphate, [bmim][PF6], for the two-phase biotrans-
formation of 1,3-dicyanobenzene to 3-cyanobenzamide and
3
-cyanobenzoic acid using the nitrile hydratase from Rhodo-
coccus R312. This work established ionic liquids as a potential
alternative to organic solvents for multiphase biotransformations.
The first reported enzyme-catalyzed reaction in an ionic liquid,
performed in our laboratory, involved the thermolysin-catalyzed
synthesis of Z-aspartame in [bmim][PF ]. This work showed
5
6
enzyme activity approached that seen in conventional organic
solvents and that the enzyme maintained very high stability in
the ionic liquid. Since this work, there have been many reports
of enzymatic catalysis in ionic liquids, including the anhydrous
†
McGowan Institute for Regenerative Medicine.
Department of Chemical and Petroleum Engineering, University of
‡
Pittsburgh.
(
3) Brennecke, J. F.; Maginn, E. J. AIChE J. 2001, 47, 2384-2389.
4) Cull, S. G.; Holbrey, J. D.; Vargas-Mora, V.; Seddon, K. R.; Lye, G. J.
Biotechnol. Bioeng. 2000, 69, 227-233.
§
Department of Bioengineering, University of Pittsburgh.
(
|
SACHEM Inc.
(
(
1) Gordon, C. M. Applied Catalysis A: General 2001, 222, 101-117.
(5) Erbeldinger, M.; Mesiano, A. J.; Russell, A. J. Biotecnol. Prog. 2000, 16,
1129-1131.
2) Holbrey, J. D.; Seddon, K. R. Clean Products Process. 1999, 1, 223-236.
10.1021/ja028557x CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 4125-4131
9
4125

A R T I C L E S
Kaar et al.
Figure 2. Lipase-catalyzed transesterification of methyl methacrylate with 2-ethylhexanol to 2-ethylhexyl methacrylate.
L-8, lyophilizate (Thermomyces lanuginosus), and Chirazyme L-9,
lyophilizate (Mucor miehei) were purchased from BioCatalytics
(Pasadena, CA). Novozym 435 (Candida antartica, type B) was a gift
from Novo Nordisk (Denmark). All enzymes were used without further
purification. Hypol 3000 prepolymer was purchased from Hampshire
Chemical (Lexington, MA). 3,3,4,4,5,5,6,6-octafluorooctan-1,8-diol
(OFOD) was purchased from Oakwood Products (West Columbia, SC)
whereas all other fluorinated diols were purchased from Lancaster
Synthesis (Windham, NH). DVA was a gift from Union Carbide
(Dandbury, CT). L-62 Pluronic surfactant was supplied as a gift by
BASF (Parsippany, NJ). The [bmim][PF ] used in all polytransesteri-
6
fication reactions was synthesized following the procedure described
8
by Huddleston and co-workers. For all transesterification reactions,
ionic liquids were synthesized and purified by SACHEM Inc. (Austin,
TX) and used as supplied. All other reagents were purchased from
Aldrich Chemicals (St. Louis, MO) and were of the highest purity
available.
Characterization of Ionic Liquids. Determination of Octanol-
Water Partition Coefficients (log P). Partition coefficients for ionic
liquids were determined as a ratio of ionic liquid concentration in the
octanol phase to the ionic liquid concentration in the aqueous phase
(eq 1). The logarithm of the partition coefficient is referred to as the
log P value
Figure 3. Lipase-catalyzed polytransesterification of divinyl adipate and
,4-butanediol.
1
transesterification of ethyl butanoate and butanol catalyzed by
Candida antarctica lipase type B (CaLB) in 1-butyl-3-meth-
ylimidazolium with hexafluorophosphate and tetrafluoroborate
[Solute]Octanol Phase
Partition Coefficient (P) )
(1)
[
Solute]Aqueous Phase
6
anions and the R-chymotrypsin catalyzed transesterification
reaction of N-acetyl-L-phenylalanine ethyl ester with 1-propanol
The “shake flask” method was used for the determination of all log P
values. The imidazolium ring present in the dialkylimidazolium based
9
in 1-butyl-3-methylimidazolium and 1-octyl-3-methylimidazo-
7
lium cations in conjunction with the hexafluorophosphate anion.
ionic liquids absorbs strongly at approximately 211 nm and can be
quantified using UV spectrometry.10 Water saturated with octanol and
octanol saturated with water were used in all experiments. For each
ionic liquid, the maximum wavelength (λmax) due to the absorbance of
the imidazolium ring was verified in both phases. The volume ratio of
the octanol phase to aqueous phase was varied in an attempt to
determine the necessary ratio to force the solute into the hydrophobic
octanol phase to a detectable level. A 100:1 volume ratio was used for
the partitioning of all ionic liquids. In a 200-mL separatory funnel,
ionic liquid was added to the saturated phases so that its concentration
did not exceed 10 mM in either phase. Inversion of the separatory funnel
was carried out for 5 min at an approximate rate of 20 inversions/min.
The phases were then recovered and centrifuged to eliminate emulsions
formed during partitioning. The ionic liquid concentration was deter-
mined in each phase using standard curves. A Perkin-Elmer Lambda
In this paper, we discuss the practicality of replacing
conventional biocatalytic solvents with ionic liquids consisting
of dialkylimidazolium and pyrrolidinium cations with a wide
range of anions including hexafluorophosphate, acetate, nitrate,
methanesulfonate, trifluoroacetate, and trifluoromethylsulfonate
(Figure 1).
Herein, we describe the first attempt of enzyme catalysis in
ionic liquids containing pyrrolidinium cations. We also compare
the polarity and hydrophibicity, via solvatochromic and oc-
tanol-water partition coefficient analysis respectively, of ionic
liquids to those of organic solvents used in biocatalytic reactions.
The lipase-catalyzed transesterification of methyl methacrylate
and 2-ethylhexanol (Figure 2) and the polytransesterification
of divinyl adipate (DVA) and 1,4-butanediol (BD) (Figure 3)
were selected as model reactions to determine the activity of
enzymes as a function of ionic liquid physical properties.
4
5 UV/VIS Spectrophotometer was used for the log P analysis.
Solvatochromic Characterization of Ionic Liquids. The protocol
N
used for measuring E
T
polarity values is described by Fletcher and
11
co-workers. Reichardt’s dye, (2,6-diphenyl-4-(2,4,6-triphenylpyridin-
io)phenolate), was dissolved in the ionic liquid and centrifuged to
Materials and Methods
Materials. Lipase L-1754 (Candida rugosa) and porcine pancreatic
lipase were purchased from Sigma Chemicals (St. Louis, MO).
Chirazyme L-2, lyophilizate (Candida antarctica) type B, Chirazyme
(
8) Huddleston, J. G.; Willauer, H. D.; Swatloski, R. P.; Visser, A. E.; Rogers,
R. D. Chem. Comm. 1998, 16, 1756-1766.
(9) Leo, A.; Hansch, C.; Elkins, D. Chem. ReV. 1971, 71, 525-616.
10) Anthony, J. L.; Maginn, E. J.; Brennecke, J. F. J. Phys. Chem. B 2001,
105, 10 942-10 949.
(
(
6) Lau, R. M.; van Rantwijk, F.; Seddon, K. R.; Sheldon, R. A. Org. Lett.
000, 2, 4189-4191.
7) Laszlo, J. A.; Compton, D. L. Biotechnol. Bioeng. 2001, 75, 181-186.
2
(11) Fletcher, F. A.; Storey, I. A.; Hendricks, A. E.; Pandey, Sh.; Pandey, Si.
(
Green Chem. 2001, 3, 210-215.
4126 J. AM. CHEM. SOC.
9
VOL. 125, NO. 14, 2003

Ionic Liquid Physical Properties on Lipase Activity
A R T I C L E S
remove any residual dye particulates. Solvents were then analyzed via
a scan of the visible spectrum using a Perkin-Elmer Lambda 45 UV-
vis Spectrophotometer. The position of the absorption band was
weight range of 580-66 000 Da. Due to error in GPC measurements,
all molecular weight results were rounded to the nearest hundred.
Attempts to Improve Enzyme Activity in Ionic Liquids. Synthesis
of Polyethylene Glycol (PEG)-Modified Lipase. Free lipase (Candida
correlated to an EN
value (eqs 2 and 3)
T
2
rugosa) was dissolved in an aqueous buffer (10 mM Tris, 5 mM CaCl ,
2
8591
pH 7.5) at a concentration of 1 mg/mL. PEG-NCO was added in excess
to the lipase solution at a molar ratio of 1:100 lipase to PEG-NCO to
ensure complete lipase modification. The reaction mixture was mixed
for 30 min followed by lyophilization.14 No further purification to
remove the excess PEG was attempted. The freeze-dried enzyme was
used in its crude form.
ET(solvent)[kcal/mol] )
(2)
(3)
λmax[nm]
ET(solvent) - 30.7
N
E
T(solvent) )
3
2.4
Polyurethane-Lipase Foam Synthesis. The synthesis of lipase-
containing (Candida rugosa) polyurethane foams was accomplished
using the procedure described by LeJeune and Russell. The poly-
urethane foams were prepared with an aqueous buffer (50 mM Bis-
2
Tris Propane, 5 mM CaCl , pH 7.5) containing 1% (w/w) L-62 Pluronic
surfactant. Lipase (1 g) was then added to the aqueous buffer (2.5 mL).
Hypol 3000 (2.5 g), a toluene diisocyanate based pre-polymer, was
added to the aqueous lipase solution and vigorously mixed for 30 s
using a 2500 rpm hand drill for sufficient mixing. The lipase-
immobilized foam was allowed to dry overnight at ambient temperature
and pressure and stored at 4 °C when not being used. In preparation
for use in transesterification reactions, the foams were ground into small
particle size pieces using a handmade drill bit wrapped in sandpaper
The normalized scale ranges from water being the most polar (λmax
)
)
4
53 nm, EN
25 nm, EN
T
) 1.0) to tetramethylsilane being the least polar (λmax
1
5
9
T
) 0.0). All ionic liquids were dried in a vacuum oven
(
26 in Hg vac.) at 70 °C for several days prior to solvatochromic
analysis.
Enzyme Activity in Ionic Liquids. Lipase-Catalyzed Trans-
esterification of Methyl Methacrylate and 2-Ethylhexanol. The
substrates methyl methacrylate and 2-ethylhexanol were dissolved in
the solvent at a concentration of 200 mM. Reactions were performed
through the addition of solvent containing methyl methacrylate (1 mL,
2
00 mM) and of 2-ethylhexanol (1 mL, 200 mM) to lipase in a 5-mL
Kimble glass vial. The enzyme suspension was sonicated for ap-
proximately 15 s. The reaction vials were then immediately placed in
a constant temperature shaker set at 30 °C and 300 rpm. When
performing reactions in organic solvents, aliquots (1.0 µL) were
removed from the reaction mixture using a 1-µL Hamilton syringe.
When performing reactions in ionic liquids, 2-ethylhexyl methacrylate
was recovered via liquid extraction using hexane. In both cases, samples
were withdrawn at specified time intervals for measurement of initial
reaction rate. Formation of 2-ethylhexyl methacrylate was analyzed
using a Perkin-Elmer Autosystem gas chromatograph (GC). The GC
was fitted with an Alltech EC 1000 capillary column (30 m × 0.53
mm × 1.0 µm) and was operated with a 1:4 split ratio using helium as
the carrier gas. The injector and detector temperatures were set to 300
(grit no. 180). The drill was operated at low rpm during the grinding
process in attempt to maintain constant particle size.
Enzyme Stability in Ionic Liquids. Lipase was incubated in the
organic solvent or ionic liquid (1 mL) at a given temperature (30 °C
or 50 °C). The solvent was then extracted and the lipase was assayed
for activity using a Lipase Diagnostic kit (Sigma 800-B). Typically,
assays were performed with lipase, in the presences of substrate (3
mL), incubated at 50 °C and shaken at 300 rpm for 2 h. Titration with
sodium hydroxide (0.05 N) was performed to determine the amount of
acid produced.
Results and Discussion
°
C. The oven program consisted of an initial temperature of 100 °C
Solvent selection for enzymatic reactions can be severely
constrained due to the inherent ability of solvents to deactivate
enzymes, as is the case with hydrophilic solvents. Hildebrandt
solubility parameters (δ), dielectric constants (ꢀ), and dipole
moments (µ) are all methods of predicting enzyme activity.
However, the most common method of predicting enzyme
activity is through the use of the octanol-water partition
that was maintained for 2 min after which the temperature was increased
at a rate of 25 °C/min to a final temperature of 160 °C and maintained
for 3 min. The retention time of 2-ethylhexyl methacrylate was 5.1
min. A calibration curve of peak area versus the concentration of
2
-ethylhexyl methacrylate in the presence of the substrates was created
1
2,13
for each organic solvent and ionic liquid studied.
No internal
standard was used; however, the GC was calibrated daily.
Lipase-Catalyzed Polytransesterification of DVA and BD. Typi-
cally, [bmim][PF ] was placed in a Wheaton vial followed by the
6
addition of DVA and BD monomers. The vials were then placed in a
1
6,17
coefficient.
Solvatochromic parameters represent a simple
16
and common means for assessing solvent polarity.
To understand how the physical properties of ionic liquids
influence enzyme activity and stability, one must first understand
how the physical properties of the ionic liquids vary with
structure. We therefore investigated the polarity and hydropho-
bicity of a variety of ionic liquids.
Physical Characterization of Ionic Liquids. A partition
coefficient is a measure of how a solvent partitions between an
aqueous and organic phase, thus serving as an indication of
solvent hydrophobicity. Laane reported that solvents with a log
P less than two are considered hydrophilic in nature and tend
to be unfavorable for enzymatic reactions. In contrast, solvents
with a log P of greater than four generally support enzymatic
activity. Although log P values are a common method of
New Brunswick Scientific Series 25 incubator/shaker set at 50 °C and
2
50 rpm for 30 min to allow for solubilization of the monomers in the
ionic liquid. Once the monomers were solubilized, polymerization was
initiated by the addition of enzyme. The polyester product precipitated
from the ionic liquid and was easily separated. The polyester product
was then dissolved in tetrahydrofuran for molecular weight analysis
via gel permeation chromatography (GPC). A Waters 150CV GPC
equipped with a refractive index detector was used for the molecular
weight analysis. Tetrahydrofuran was used as the mobile phase at a
flow rate of 1.0 mL/min and 35 °C. The GPC contains three columns
in series to achieve sufficient separation in the molecular weight range
of 500-30 000 Da. The first two columns in the series, PL-gel Mixed-E
Polymer Laboratories, contained mixed porosity. The third column, a
Waters Ultrastyragel column, had a uniform pore size of 500 Å. A
calibration curve of the log of molecular weight versus retention time
was constructed using 11 polystyrene standards within the molecular
(
14) Drevon, G. F.; Hartleib, J.; Scharff, E.; R u¨ terjans, H.; Russell, A. J.
Biomacromolecules 2001, 2, 664-671.
(
15) LeJeune, K. E.; Russell, A. J. Biotechnol. Bioeng. 1996, 51, 450-457.
(
12) Kamat, S.; Critchley, G.; Beckman, E. J.; Russell, A. J. Biotechnol. Bioeng.
995, 46, 610-620.
13) Kamat, S.; Barrera, J.; Beckman, E. J. Biotechnol. Bioeng. 1992, 40, 158-
66.
(16) Laane, C.; Boeren, S.; Vos, K.; Veeger, C. Biotechnol. Bioeng. 1987, 30,
1
81-87.
(
(17) Pogorevc, M.; Stecher, H.; Faber, K. Biotechnol. Lett. 2002, 24, 857-
1
860.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 14, 2003 4127

A R T I C L E S
Kaar et al.
Table 1. Log P and Reichardt’s Dye Polarity (ENT) Values for Ionic Liquids and Select Organic Solvents
solvent
EN
T
log P
1
1
1
1
1
1
1
1
1
-butyl-3-methylimidazolium acetate
-butyl-3-methylimidazolium nitrate
-butyl-3-methylimidazolium trifluoroacetate
-butyl-3-methylimidazolium hexafluorophosphate
-methyl-1-(-2-methoxyethyl)pyrrolidinium acetate
-methyl-1-(-2-methoxyethyl)pyrrolidinium nitrate
-methyl-1-(-2-methoxyethyl)pyrrolidinium trifluoroacetate
-methyl-1-(-2-methoxyethyl)pyrrolidinium trifluoromethylsulfonate
-methyl-1-(-2-methoxyethyl)pyrrolidinium methanesulfonate
0.57
0.65
0.63
0.67
0.52
0.84
0.37
0.91
0.78
0.0091
-2.77 ( 0.11
-2.90 ( 0.01
-2.39 ( 0.27
9
9
3.516
hexane
tetrahydrofuran (THF)
acetonitrile
1
16
0.207
0.49
0.4719
-0.33
16
biocompatibility assessment of organic solvents, it is important
to note that this method does not always accurately predict
enzyme activity.16,17
forces. Solvents that demonstrate the ability to hydrogen bonding
present the potential to interfere with enzyme structure.
Results of our solvatochromic analysis of [bmim][PF ] agree
6
N
T
To our knowledge, these results are the first determinations
of log P for ionic liquids. Huddleston and co-workers recently
reported that there is a distinct correlation between the partition-
ing of several substituted-benzene derivatives in a 1-butly-3-
methylimidizolium hexafluorophosphate/water system and their
respective log P values.8 However, they do not explicitly
measure, calculate, or approximate the log P values of the ionic
liquid 1-butly-3-methylimidizolium hexafluorophosphate in a
with the reported E (Table 1). Our results indicate that
replacement of the [PF ] anion with the [CH CO ], [NO ], and
6
3
2
3
N
[CF CO ] anions resulted in E T’s of 0.57, 0.65 and 0.63
3
2
respectively (Table 1). This small change in the polarity series
1
8
is consistent with previous work. The ionic liquids [mmep]-
[CH CO ], [mmep][NO ], [mmep][CF CO ], [mmep][CF SO ],
3
2
3
3
2
3
3
N
and [mmep][CH SO ] have E T’s of 0.52, 0.84, 0.37, 0.91, and
3
3
0.78 respectively (Table 1). The polarity of the ionic liquids
containing the pyrrolidinium-based cation appear to be strongly
dependent on the anion. For comparison with common organic
1
-octanol/water system. The experimentally measured values
of log P for the ionic liquids tested ranged between -2.39 and
2.90 (Table 1). For comparison, the log P values for hexane,
acetonitrile, and tetrahydrofuran are 3.5, -0.33, and 0.49
N
T
-
solvents, the reported E for hexane, acetonitrile, and tetra-
19
hydrofuran are 0.009, 0.47, and 0.207 respectively. With the
16
N
respectively. Using the guidelines detailed by Laane and co-
workers,16 this would indicate that the ionic liquids are highly
hydrophilic in nature and would likely inactivate enzymes.
Because the pyrrolidinium-based ionic liquids did not have any
characteristic peaks in the UV spectra, the log P values could
not be determined for these ionic liquids.
exception of [mmep][CF CO ], the E T’s for all ionic liquids
3
2
studied were greater than those of acetonitrile and conventional
organic solvents used in biocatalysis. The polar nature of ionic
liquids presents an ideal reaction media for performing biotrans-
formations involving highly polar substrates such as carbohy-
drates, which contain multiple hydroxyl functionalities.
Solvatochromic parameters provide a simple method to
determine the polarity of a solvent. The E T empirical polarity
Enzyme Activity in Ionic Liquids. Almost all of the
enzymatic reactions performed to date are in dialkylimidazolium
based ionic liquids with the hexafluorophosphate and tetrafluo-
roborate anions. Only recently have other cations and anions
N
scale is based on the shift of the charge-transfer absorption band
of a solvatochromic probe in the presence of a solvent. Changes
in the position of the charge-transfer absorption band within
the visible spectrum are due to hydrogen bonding between the
solvent and the phenoxide oxygen atom present in Reichardt’s
2
4-26
began to be studied.
Our interest is in expanding the
portfolio of potential biocatalytic solvents by studying a variety
of dialkylimidazoium and N-methyl-N-(-2-methoxyethyl) pyr-
rolidinium-based ionic liquids.
N
dye. An E T value is then determined as a function of the
18,19
position of the charge-transfer absorption band.
Initial reaction rates of the lipase-catalyzed transesterification
of methyl methacrylate and 2-ethylhexanol were measured in
several ionic liquids and organic solvents to determine the effect
of the ionic liquid on enzyme activity. Our experience with this
reaction system, combined with an extensive data set, makes
this an ideal model system. The selection of lipase for these
experiments was based on prior work in which several lipases
including Amano lipases G (Penicillium camembertii), AY
There are only a few reported cases of the solvatochromic
characterization of ionic liquids.1
1,18,20-23
Results of solvato-
chromic analysis of [bmim][PF6] using Reichardt’s dye reported
N
in the literature indicate that dry [bmim][PF6] has an E T value
N
very similar to that of ethanol (E T ) 0.67) and other short
chain primary alcohols.1
1,18,20,23,24
This value is sensitive to the
20,21
presence of water in the ionic liquid.
Muldoon and co-
1
8
N
workers suggest E T’s for ionic liquids are dependent upon
the strength of hydrogen bonding between the cation of the ionic
liquid and the phenoxide group present in Reichardt’s dye. High
(Candida rugosa), AK (Pseudomonas fluorescens), MAP (Mu-
cor jaVanicus), FAP (Rhizopus oryzae), PS (Pseudomonas
cepacia), and Sigma lipases L1754 (Candida rugosa), L0763
N
E T indicates that ionic liquids exhibit strong hydrogen bonding
(Chromobacterium Viscosum), L3001 (wheat germ) were screened
for activity in hexane via the lipase-catalyzed transesterification
(
18) Muldoon, M. J.; Gordon, C. M.; Dunkin, I. R. J. Chem. Soc., Perkin Trans.
2
001, 433-435.
(19) Reichardt, C. Chem. ReV. 1994, 94, 2319-2358.
(
(
(24) Kaftzik, N.; Wasserscheid, P.; Kragl, U. Org. Process Res. DeV. A 2002,
6, 553-557.
20) Baker, S. N.; Baker, G. A.; Bright, F. V. Green Chem. 2002, 4, 165-169.
21) Aki, S. N. V. K.; Brennecke, J. F.; Samanta, A. Chem. Commun. 2001,
(25) Eckstein, M.; Sesing, M.; Kragl, U.; Adlercreutz, P. Biotechnol. Lett. 2002,
24, 867-872.
4
13-414.
(
(
22) Dzyuba, S. V.; Bartsch, R. A. Tetrahedron Lett. 2002, 43, 4657-4659.
23) Carmichael, A. J.; Seddon, K. R. J. Phys. Org. Chem. 2000, 13, 591-595.
(26) Lozano, P.; De Diego, T.; Carrie, D.; Vaultier, M.; Iborra, J. L. Chem.
Commun. 2001, 23, 1529-1533.
4128 J. AM. CHEM. SOC.
9
VOL. 125, NO. 14, 2003

Ionic Liquid Physical Properties on Lipase Activity
A R T I C L E S
of methyl methacrylate with 2-ethylhexanol. The transesterifi-
cations were performed under the same conditions, 30 °C and
Table 2. Monomer Solubilities at 50 °C in [bmim][PF6]
maximum solubility
in [bmim][PF
300 rpm, as were used in this study eliminating any temperature
monomer
divinyl adipate (DVA)
,4-butanediol (BD)
2,2,3,3-tetrafluoro-1,4-butanediol (TFBD)
6
]
or mixing effects. Results of the lipase screen indicated that
3.45M
0.98M
1
Candida rugosa lipase exhibited the highest activity and was
1
2,13
not soluble
0.20M
2.23M
therefore chosen as the model enzyme for this study.
2
3
,2,3,3,4,4-hexafluoro-1,5-pentanediol (HFPD)
,3,4,4,5,5,6,6-octafluorooctan-1,8-diol (OFOD)
In [bmim][PF6], free lipase (Candida rugosa) catalyzed the
transesterification at an initial rate of 6.75 µM/hr/mg-enzyme,
1
.5 times faster than the reaction in hexane (3.90 µM/hr/mg-
enzyme). Although [bmim][PF6] showed promise as an effective
solvent for the transesterification, free Candida rugosa lipase
was unfortunately inactive in all the other ionic liquids
investigated ([bmim][CH3CO2], [bmim][NO3], [bmim][CF3-
CO2], [mmep][CH3CO2], [mmep][NO3], [mmep][CH3SO3],
the solubility of several monomers was assessed in [bmim][PF6],
including several fluorinated diols (Table 2).
Polymerizations catalyzed by Novozym 435 were performed
using DVA and each of the soluble diols. Polytransesterifications
were performed for 24 h after which time precipitation of the
polymer from the ionic liquid was observed. The reaction
between DVA and BD provided the highest molecular weight
oligomers (MW ) 2900 Da, PDI ) 1.20). An enzyme screen
was also performed for the DVA/BD reaction. The use of lipases
from Thermomyces lanuginosus and Mucor miehei produced
oligomers with lower molecular weights. All other lipases
investigated failed to produce polymer in this reaction system.
Polytransesterifications between DVA and 2,2,3,3,4,4-hexafluoro-
1,5-pentanediol (HFPD) and DVA and 3,3,4,4,5,5,6,6-octafluo-
rooctan-1,8-diol (OFOD) produced oligomers with similar
molecular weights, 1300 and 1000 Da respectively. It is possible
in all of the described polymerizations that low molecular weight
oligomers remained dissolved in the reaction medium and
therefore resulted in fractionation of the polymer.
[mmep][CF3CO2], and [mmep][CF3SO3]). The reaction also did
not proceed in the more polar organic solvents THF or
acetonitrile.
Water was added to ionic liquids, 25 to 100 µL H2O/mL
solvent, to activate the free lipase in [bmim][NO3] and
[mmep][CH3CO2]. Water stripping from an enzyme by polar
solvents is a common cause of deactivation. However, hydration
of the enzyme failed to improve activity in either of the ionic
liquids making water stripping an unlikely mechanism of
deactivation. The addition of the hydrophilic ionic liquids
[
[
bmim][NO3] and [bmim][CH3CO2] as cosolvents with [bmim]-
PF6] was also investigated as a means of maintaining enzyme
activity while enhancing the ability to dissolve polar substrates.
Transesterification reactions were performed using free lipase
(
Candida rugosa) in mixtures of [bmim][NO3]/[bmim][PF6] and
After our initial report,28 Uyama and co-workers recently
described a polycondensations of dicarboxylic acid diesters with
1,4-butanediol in [bmim][PF6] and [bmim][BF4] catalyzed with
free lipase (Candida antarctica). Polycondensation reactions
performed at ambient conditions in [bmim][PF6] generated
oligomers (Mn ) 350 Da), although oligomers with increased
molecular weights (Mn ) 1500 Da) were formed at reduced
pressure and increased polymerization times.
Attempts to Improve Enzyme Activity in Ionic Liquids.
Having shown free lipase to be inactive in the presence of all
[bmim][CH3CO2]/[bmim][PF6]. Concentrations of [bmim][PF6]
in the cosolvent mixtures were varied at 25, 50, and 75 (v/v
). No lipase activity was observed in any of the mixed ionic
%
liquid mixtures indicating the use of [bmim][PF6] as a cosolvent
with more hydrophilic ionic liquids did not prevent the
inactivation of free lipase.
Our results suggest that enzyme activity in ionic liquids is
anion dependent. Anions such as [NO3], [CH3CO2], and [CF3-
CO2] are more nucleophilic than [PF6] and may coordinate more
strongly to positively charged sites in the lipase’s structure
2
9
ionic liquids with the exception of [bmim][PF ], several methods
6
2
7
causing conformation changes in the enzyme’s structure.
Similar results observed with [mmep][NO3], [mmep][CH3CO2],
mmep][CF3CO2], [mmep][CF3SO3], and [mmep][CH3SO3] in
of enzyme stabilization were investigated to determine if
conventional activating techniques would be useful. Adsorption
onto an acrylic resin, PEG-modification, and immobilization
via multipoint attachment in polyurethane foam are three
methods that have been shown to prevent deactivating confor-
mational changes in harsh environments and enhance enzyme
[
which all anions are strong nucleophiles further suggest anion
nucleophilicity is essential to enzyme activity. However, due
to little work with pyrrolidinium-based ionic liquids, the
possibility that the pyrrolidinium cation inhibits enzyme activity
must not be ruled out. This new data calls into question the
broad utility of ionic liquids with enzymes unless the deactivat-
ing mechanism can be understood and reversed. Naturally, we
are now exploring such mechanisms.
It would appear that although there is much enthusiasm about
enzymatic catalysis in ionic liquids, only one such liquids
supports significant activity with lipases. Having demonstrated
that lipase maintains a high level of activity in [bmim][PF6], a
natural extension was to perform an enzymatic polymerization
in the ionic liquid. The lipase-catalyzed polytransesterification
of DVA and BD in [bmim][PF6] was studied to determine if
the polymerizations would proceed in this environment. Initially,
1
5,30,31
activity.
Transesterifications catalyzed by the modified
enzymes were performed to determine the effectiveness of the
activating methods in ionic liquids.
Candida antarctica lipase, type B adsorbed onto a macroporous
3
0
acrylic support (Novozym 435) was used to catalyze trans-
esterification in several ionic liquids and organic solvents. In
hexane and acetonitrile, initial rates of 3.68 and 8.85 µM/hr/
mg-enzyme were observed while in THF no lipase activity was
observed. In [bmim][PF ],[bmim][CH CO ], [bmim][NO ],
6
3
2
3
(28) Mesiano, A. J.; Beckman, E. J.; Russell, A. J. Abstract #279k, presented
at the National Meeting of the American Institute of Chemical Engineers
November 2000.
(
29) Uyama, H.; Tetsufumi, T.; Kobayashi, S. Polymer J. 2002, 34, 94-96.
30) Arroyo, M.; Sanchez-Montero, J. M.; Sinisterra, J. V. Enzyme Microb.
Technol. 1999, 24, 3-12.
(
(27) Sheldon, R. A.; Lau, R. M.; Sorgedrager, M. J.; van Rantwijk, F.; Seddon,
(31) Inada, Y.; Takahashi, K.; Yoshimoto, T.; Ajima, A.; Matsushima, A.; Saito,
K. R. Green Chem. 2002, 4, 147-151.
Y. Trends Biotechnol. 1986, 4, 190-194.
J. AM. CHEM. SOC.
9
VOL. 125, NO. 14, 2003 4129

A R T I C L E S
mmep][CH3CO2], [mmep][NO3], [mmep][CH3SO3] no activity
Kaar et al.
[
was detected.
PEG-modification of enzymes is another common technique
for stabilizing enzymes in denaturing environments. Drevon and
co-workers reported the ability to prepare a PEG-diisopropyl-
fluorophosphatase (DFPase) conjugate. Their results indicate
that a 100% extent of PEG-modification was achieved meaning
that all of the native DFPase was modified. Matrix-assisted laser
desorption/ionization mass spectrometry (MALDI MS) analysis
confirmed that two to seven PEG chains attached per molecule
1
4
of DFPase. It has also been reported that in some organic
solvents, PEG-modified enzymes exhibit increased stability
when compared to their untreated form.3 PEG-modification of
Candida rugosa lipase was used to stabilize the enzyme in ionic
liquids. Although PEG-modified lipase catalyzed the trans-
esterification in hexane (0.75 µM/hr/mg-enzyme), no improved
lipase activity was observed in [bmim][PF6], [bmim][NO3], and
1
Figure 4. Stability of Novozym 435 in tetrahydrofuran (O) vs [bmim]-
[PF6] (0) and procine pancreatic lipase in tetrahydrofuran (b) vs [bmim]-
[PF ] (9) at 50 °C.
6
[
mmep][CH3CO2] compared to the use of free Candida rugosa
lipase.
With adsorption and PEG-modification being unsuccessful
in stabilizing lipase in ionic liquids, covalent immobilization
was considered as another technique for enzyme stabilization.
Immobilization of free lipase (Candida rugosa) within poly-
urethane foam offers multi-point covalent attachment of the
enzyme throughout the polymer matrix.14 Drevon and Russell
reported the ability to immobilized DFPase in polyurethane
foam. After thorough rinsing with distilled water, only 1% (w/
w) DFPase leached from the enzyme-containing polyurethane
foams. This suggests that approximately 100% of DFPase is
irreversibly immobilized in the polyurethane foams during
synthesis. Additionally, the enzyme-containing polyurethane
foams retained 67% of the relative activity of native DFPase
when assayed against diisopropylfluorophosphate (DFP) in
buffer.32 Polyurethane immobilized lipase catalyzed the trans-
esterification in [bmim][PF6] at an initial rate of 0.93 µM/hr/
mg-enzyme, however, no detectable activity was observed in
Figure 5. Stability of Novozym 435 intetrahydrofuran (2) vs [bmim][PF6]
(
(
[) and porcine pancreatic lipase intetrahydrofuran (4) vs [bmim][PF6]
]) at 50 °C in the presence of divinyl adipate and 1,4-butanediol.
relative activity measurements indicate the portion of lipase
irreversibly inactivated as a result of incubation in ionic liquid.
However, the stability studies do not demonstrate the stability
of lipase in ionic liquids to catalyze a reaction. After 48 h
suspended in [bmim][PF6] at 50 °C, Novozym 435 and porcine
pancreatic lipase retained approximately 67% and 94% of their
original activities respectively (Figure 4). This represents an
improvement over activity retentions of Novozym 435 (21%)
and porcine pancreatic lipase (61%) in tetrahydrofuran. Interest-
ingly, Novozym 435 lost no activity during the initial 50 min
of incubation in [bmim][PF6]. Incubation of Novozym 435 and
porcine pancreatic lipase in [bmim][PF6] over a two week period
demonstrate that both lipases retained appreciable activity. After
[bmim][CH3CO2], [bmim][NO3], [mmep][CH3CO2], and [mmep]-
[CH3SO3] using polyurethane immobilized lipase thus demon-
strating that this activating technique was ineffective.
Sheldon and co-workers reported similar results using im-
mobilized forms of Candida antarctica lipase type B. Conver-
sions of less than 5% were measured in a 24 h period in
[
bmim][NO3] using cross-linked enzyme crystals and cross-
2
7
linked enzyme aggregates. The lack of immobilized lipase
activity using a different lipase than was used in this work
supports our conclusion that the described activation techniques
would most likely prove ineffective in enhancing activity of
other lipases in ionic liquids. Clearly, for the broad range of
ionic liquids tested, ionic liquids are not the panacea for
nonaqueous biocatalysis that we had hoped.
4
8 h incubation in [bmim][PF6] at 50 °C with DVA and BD
substrates present, Novozym 435 and porcine pancreatic lipase
retained 50% and 80% activity respectively (Figure 5).
High levels of activity retention upon incubation in [bmim]-
[PF6] for 120 h at 70 °C indicated both Novozym 435 and
porcine pancreatic lipase (20% and 48% respectively) have
increased thermal stability in the ionic liquid in comparison to
octane (3% and 35%).
The stability of Novozym 435 was further investigated in
the other ionic liquids via the same method (Figures 6 and 7).
Incubation in [mmep][NO3] and [bmim][NO3] for 24 h at 30
Enzyme Stability in Ionic Liquids. The stability of No-
vozym 435 and porcine pancreatic lipase in [bmim][PF6] was
investigated. The goal of the stability studies was to determine
permanent, or irreversible, inactivation effects on lipase when
in ionic liquids. This was achieved by incubating the lipase in
the ionic liquid for a specified period of time after which the
enzyme was isolated via extraction of ionic liquid, diluted with
water, and assayed. Extraction of the ionic liquid was performed
using a pipet with care taken to minimize enzyme loss. The
°
C resulted in 59% and 0% activity retention. Conversely, results
of incubation in [mmep][CH3SO3], [bmim][CH3CO2], and
mmep][CH3CO2] under the same conditions demonstrate
[
increases in enzyme activity once re-suspended in water. More
specifically, [mmep][CH3CO2], [bmim][CH3CO2], and [mmep]-
(32) Drevon, G. F.; Russell, A. J. Biomacromolecules 2000, 1, 571-576.
4
130 J. AM. CHEM. SOC. VOL. 125, NO. 14, 2003
9

Ionic Liquid Physical Properties on Lipase Activity
A R T I C L E S
Sheldon and co-workers recently studied the stability of free
lipase (Candida antarctica), Novozym 435, cross-linked lipase
crystals, and cross-linked lipase aggregates in several ionic
liquids. They reported that after 100 hrs of incubation in [bmim]-
[PF6] at 80 °C, an increase in free lipase activity (120%) was
observed. Incubation of Novozym 435 also resulted in increased
activity (350%) after 40 h of incubation in [bmim][PF6] at 80
°
C in comparison to untreated free lipase. They suggest that
the ionic liquid coated and thus protected the layer of essential
27
water surrounding the lipase, but it is hard to explain increases
in activity of the free lipase through a merely protective
mechanism. We believe that one must invoke a permanent
activating conformational change or an increase in active site
concentration to explain this unusual data.
Figure 6. Stability of Novozym 435 at 30 °C in various ionic liquids
(
(
[bmim][PF6] (]), [bmim][NO3] (0), [bmim][CH3CO2] (4), [mmep][NO3]
9), [mmep][CH3CO2] (2), and [mmep][CH3SO3] (b)). Error bars represent
Conclusions
deviation from the mean for two separate experiments.
Our previous work has already demonstrated the feasibility
of ionic liquids as media for biocatalysis. Solvatochromic
analysis and partition coefficients suggest ionic liquids are highly
polar and hydrophilic in nature in comparison to organic solvents
such as hexane, acetonitrile, and tetrahydrofuran. In this study,
we examined the lipase-catalyzed transesterification of methyl
methacrylate and the polytransesterification of divinyl adipate
and 1,4-butanediol in a variety of hydrophobic and hydrophilic
ionic liquids. Although the activity of free lipase in [bmim]-
[PF6] was greater than in hexane, no reaction occurred in the
hydrophilic ionic liquids investigated. Conventional methods
of enzyme stabilization including adsorption, PEG-modification,
and covalent immobilization in polyurethane foams proved
ineffective in the hydrophilic ionic liquids. Polytransesterifica-
tions performed in [bmim][PF6] produced polyesters of limited
molecular weight due to precipitation of the polymer. Stability
studies indicate that reversible inactivation of lipase is observed
when the enzyme is suspended in [bmim][CH CO ], [mmep][CH -
Figure 7. Stability of Novozym 435 at 30 °C in various organic solvents
(butanol (]), tetrahydrofuran (0), acetonitrile (2), dimethyl sulfoxide (9),
and hexane ([)). Error bars represent deviation from the mean for two
separate experiments.
[CH3SO3] suspended - Novozym 435 exhibited 297%, 202%,
3
2
3
and 176% relative activity when returned to water, respectively.
For comparison, under the same conditions, no activity loss was
observed in hexane while Novozym 435 retained 73% of its
activity in tetrahydrofuran. Activity retentions of 68% and 55%,
respectively, were observed for incubation in acetonitrile and
dimethyl sulfoxide, whereas in butanol the enzyme was
completely deactivated. Interestingly, although these liquids
support no enzyme activity, they cause irreversible activation
of the immobilized enzyme. One plausible explanation is that
these liquids swell the acrylic resin causing previously inac-
cessible enzyme to be accessed once returned to water. Another
possibility is that the ionic liquid exposed enzyme has a modified
tertiary structure with increased activity in water. Although this
seems less likely we cannot rule out that possibility.
2 3 3
CO ], and [mmep][CH SO ], whereas incubation in [bmim]-
[NO ] and [mmep][NO ] irreversibly inactivates lipase. To use
3
3
hydrophilic ionic liquids as solvents for biocatalytic transforma-
tions, the mechanism of inactivation must be more clearly
understood.
Acknowledgment. We would like to thank SACHEM Inc.
for partially funding this work. In specific, we are grateful Dr.
Richard Goodin and Rosemary Hoffman from SACHEM Inc.
for their discussions and contributions involving ionic liquids.
In addition, this work was partially funded by research grant
from the EPA (R-82813101-0).
JA028557X
J. AM. CHEM. SOC.
9
VOL. 125, NO. 14, 2003 4131