Using ionic liquid [EMIM][CH3COO] as an enzyme-'friendly' co-solvent for resolution of amino acids

Article abstract of DOI:10.1016/j.tetasy.2006.09.009

An ionic liquid (IL), 1-ethyl-3-methylimidazolium acetate [EMIM][CH₃COO], was used in 0-4.0 M (～60% IL, v/v), as a nonvolatile organic medium for the enzymatic resolution of amino acids. When dl-phenylalanine methyl ester was studied as a model substrate, high enantiomeric excesses (ee) of l-amino acid were obtained in all ionic concentrations; however, lower yields were observed at high IL concentrations. This IL is more enzyme-'friendly' than the hydrophilic organic solvent acetonitrile and those ILs containing chaotropic anions (such as [EMIM][OTs]). Among three proteases and two lipases investigated, lyophilized Bacillus licheniformis protease exhibited the best enantioselectivity and activity. Highly enantioselective resolutions were also produced for several other amino acids in 2.0 M IL. Interestingly, high ee were also found in deuterium oxide (D₂O) rather than in ordinary water, and a further enhancement was achieved with the co-existence of [EMIM][CH₃COO]. The heavy water effect was explained in terms of protein stabilization by D₂O. The secondary structural changes of enzyme in various media were interpreted by the second derivatives of FT-IR spectra.

Full text of DOI:10.1016/j.tetasy.2006.09.009

Tetrahedron: Asymmetry 17 (2006) 2491–2498
Using ionic liquid [EMIM][CH COO] as an enzyme-‘friendly’
3
co-solvent for resolution of amino acids
*
Hua Zhao, Lee Jackson, Zhiyan Song and Olarongbe Olubajo
Chemistry Program, Savannah State University, Savannah, GA 31404, USA
Received 25 July 2006; accepted 7 September 2006
Abstract—An ionic liquid (IL), 1-ethyl-3-methylimidazolium acetate [EMIM][CH
3
COO], was used in 0–4.0 M (ꢀ60% IL, v/v), as a non-
volatile organic medium for the enzymatic resolution of amino acids. When DL-phenylalanine methyl ester was studied as a model sub-
strate, high enantiomeric excesses (ee) of L-amino acid were obtained in all ionic concentrations; however, lower yields were observed at
high IL concentrations. This IL is more enzyme-‘friendly’ than the hydrophilic organic solvent acetonitrile and those ILs containing cha-
otropic anions (such as [EMIM][OTs]). Among three proteases and two lipases investigated, lyophilized Bacillus licheniformis protease
exhibited the best enantioselectivity and activity. Highly enantioselective resolutions were also produced for several other amino acids in
2
.0 M IL. Interestingly, high ee were also found in deuterium oxide (D O) rather than in ordinary water, and a further enhancement was
2
achieved with the co-existence of [EMIM][CH
secondary structural changes of enzyme in various media were interpreted by the second derivatives of FT-IR spectra.
2006 Elsevier Ltd. All rights reserved.
3 2
COO]. The heavy water effect was explained in terms of protein stabilization by D O. The
ꢀ
1
. Introduction
might dissolve certain enzymes, different enzyme activities
2
1
are observed using a pure or very concentrated IL as the
reaction medium can be disadvantageous because the high
IL viscosity usually leads to a slow dissolution of sub-
strate(s) and a reduction of reaction rates.
Ionic liquids (ILs) are relatively new organic solvents con-
sisting of ions. Their increasing applications in organic reac-
tions including biocatalysis (see reviews ) are not solely
due to their favorably low volatility thus potentially being
1
–6
‘
green’, but more importantly, owe to their unique solvent
To overcome these drawbacks, we have been exploring the
uses of aqueous solutions of hydrophilic ILs as enzymatic
properties, such as tunable polarity, viscosity, and hydro-
phobicity through modifying the combination of cations
and anions.
reactions are carried out in hydrophobic ILs (e.g., PF
or (CF SO ) N salts) (see more discussion in our recent
review ). These hydrophobic ILs can be as advantageous
as hydrophobic organic solvents in enzymatic reactions,
particularly their lower tendency to strip the ‘essential’
2
2–24
media.
In diluted IL aqueous solutions, the enzyme
7
–10
At present, most IL-mediated enzymatic
is typically well hydrated; also, the medium viscosity and
solution dissolution ability (towards substrates) are adjust-
able through varying the IL content. However, since
hydrophilic ILs dissociate into cations and anions in water,
the overall solvent properties (such as polarity) do not
always account for the enzymatic behaviors in these media.
In other words, the mechanism of solvent effect on the
ꢁ
6
ꢁ
3
2 2
1
1
12–14
1
5,16
water from enzymes.
The enzyme activity, stability,
and enantioselectivity have been interpreted in terms of IL
enzyme function in IL aqueous solutions is different from
1
17,18
22–24
properties including polarity, hydrogen-bond basicity,
that in nearly dry ILs. Our recent studies
suggested
1
9
5
anion nucleophilicity, and viscosity. However, since the
enzyme is suspended rather than dissolved in hydrophobic
ILs, its activity in ILs is expected to be lower than that in
aqueous solutions due to similar observations in hydropho-
that the individual ion’s kosmotropicity, following the
Kosmotropes are strongly hydrated species and thus called water
ꢁ
2ꢁ
‘
structure-makers’. Kosmotropic ions include CH
3
COO , SO4
,
1
2,20
bic organic solvents.
Although some hydrophilic ILs
2+
2+
+
+
ꢁ
2ꢁ
HPO4 , Mg , Ca , Li , H , and OH . Chaotropes are weakly
hydrated species and thus called water ‘structure-breakers’. They include
(guanidinium), and
(CH ) N (tetramethylammonium). The kosmotropicity measures how
ꢁ
ꢁ
+
ꢁ
ꢁ
+
+
+
SCN , I , NO , BF4 , Cs , K , (NH
2
)
3
C
3
*
Corresponding author. Tel.: +1 912 353 5290; fax: +1 912 691 6839;
e-mail: zhaoh@savstate.edu
3
4
strong a species (ion or neutral compound) is as a kosmotrope.
0957-4166/$ - see front matter ꢀ 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tetasy.2006.09.009

2492
H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
Hofmeister ion series of protein stabilization,11,25–27 is
responsible for the enzyme stability and enantioselectivity
in IL aqueous solutions. The general conclusion was that
ILs consisting of a kosmotropic anion and chaotropic cat-
ion can stabilize or even activate the enzyme in aqueous
solutions. One explanation of the kosmotropic effect is that
although both a kosmotropic cation and an anion (e.g.,
ee
100
yield
80
ee
2
+
2ꢁ
60
E
Ca and SO ) strongly salt out nonpolar groups, the
4
2
+
cation (Ca ) strongly salts in (and interacts with) the
peptide group (causing destabilization) while the anion
yield
4
0
0
0
2ꢁ
28
ðSO4 Þ does not.
Through previous studies,22,23 we have identified several
room-temperature ILs (such as 1-ethyl-3-methylimidazo-
2
E
+
lium [EMIM] salts of acetate, trifluoroacetate, ethyl sul-
fate, and citrate) that are considered enzyme-‘friendly’ in
terms of enabling high enzyme activities and enantioselec-
2.0 M [EMIM][OAc]
2.0 M acetonitrile
tivities. It is because each of these ILs contains a kosmo-
tropic anion and a chaotropic cation.29 In addition,
Figure 1. Comparing the effect of IL and organic solvent on the resolution
of DL-phenylalanine (lyophilized Bacillus licheniformis protease, 40 min
reaction time, 30 ꢁC, the error bars indicating percentage errors).
[
EMIM][CH COO] has a relatively low viscosity (deter-
3
à
mined in this study as 11.2 cP, see Section 4.6), and does
not contain a fluorine atom as many other ILs do (thus
being ‘greener’). For these reasons, we focused on the use
of this IL in a current study for the enzymatic resolution
of amino acids. It is well recognized that chiral amino acids
are becoming increasingly important intermediates for
heavy water on enzyme activity, with or without the
co-presence of an IL.
3
1–33
chiral drug synthesis.
2. Results and discussion
Meanwhile, we are very interested in examining the effect
2.1. Solvent effect on the enzymatic reaction
of heavy water (deuterium oxide, D O) on enzymatic reso-
2
lution. Heavy water (92%D) has been known to be lethal to
tadpoles, guppy fish, and worms although a low concentra-
The enzymatic resolution of DL-phenylalanine methyl ester
was performed in 2.0 M [EMIM][CH COO] and 2.0 M
3
4
3
tion (30%) seems safe for tadpoles. However, an early
3
5
acetonitrile, respectively (Fig. 1). It is not surprising that
low enantioselectivity and activity were observed in the
organic solution (especially at high concentrations). The
reason is that polar organic solvents tend to deactivate
study suggested that the rate of ovalbumin denaturation
by urea in heavy water is one-third that in ordinary water;
it was suspected that the strong hydrogen bonds with deu-
3
6
terium stabilize the protein. Another study indicates that
a protein named tubulin is unstable when existing in solu-
4
0,41
enzymes due to their interactions with proteins.
However, considerably higher ee and yield (or in terms of
enantiomeric ratio, E) were obtained in 2.0 IL solution,
suggesting that this IL does not have a strong interaction
with the protein and can stabilize the enzyme.
tions, but could be stabilized by D O against inactivation
2
at both 4 and 37 ꢁC; it was explained that D O might be
2
involved in the conformational step that influences the
hydrophobic forces. The hydrogen bonds are essential for
enzymes to maintain the protein tertiary structures. Hydro-
gen bond with deuterium is slightly stronger than that with
a regular hydrogen, therefore, the rigidity of most protein
structures can be increased by using D O instead of H O,
In order to understand the structural changes of proteins in
the above media, we measured the infrared spectra of a
protease in these solvents and calculated their second deriv-
atives (Fig. 2). It is well established that the second deriv-
2
2
especially at higher temperatures, as demonstrated by
ꢁ
1
3
7
atives of the IR spectra between 1600 and 1700 cm (so-
Cioni and Strambini. They also suggested that the folded
state is more stable in D O than in H O based on the in-
§
called the amide I band) can reveal the protein secondary
2
2
structures (a-helix, b-sheets, b-turns, and nonordered or
verse relationship between structural flexibility and stabil-
4
2,43
3
8
irregular structures).
The a-helical structures are
ity to thermal denaturation. However, a contradictory
result based on the Gibbs energy of protein unfolding at
ꢁ
1
normally observable between 1650 and 1658 cm , and
b-sheets are between 1620 and 1640 cm
ꢁ
1 43
.
As illustrated
2
5 ꢁC demonstrated that the stabilities of ribonuclease A
in Figure 2, after 40 min of incubation at 30 ꢁC, the prote-
in D O and H O are quite close; however, the stabilities
2
2
ase in 2.0 M [EMIM][CH COO] contains essentially the
3
of horse cytochrome c and hen egg lysozyme in H O are
2
3
9
same structural elements (a-helix, b-sheets, and b-turns)
as that in water, suggesting that the protein’s secondary
structures were well maintained in the IL solution. How-
ever, when the enzyme was kept in 2.0 M acetonitrile for
higher than their respective ones in D O.
In this
2
paper, we discuss some preliminary data of the effect of
à
The viscosities of other common ILs are usually much higher, for
3
0
§
example, [EMIM][BF
25 ꢁC).
4
] = 43 cP (30 ꢁC) and [EMIM][Tf
2
N] = 28 cP
This band is due to the in-plane C@O stretching vibration of the amide
group, coupled with C–N stretching and in-plane N–H bending.
4
2
(

H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
2493
tropicity on enzyme activity has been discussed in detail
in our above-mentioned studies.
water
2
2
4
.0 M ACN
.0 M IL (water)
.0 M IL (water)
2.2. Effect of IL concentrations
1
671
1640
β-turn
To further examine the compatibility of [EMIM][CH3-
COO] with proteases, we conducted the enantioseparation
of DL-phenylalanine methyl ester in different concentra-
tions of this IL. The IL concentrations studied range from
1
669
1690
0
(pure water) to 4.0 M (ꢀ60% IL, v/v). As illustrated in
1
623
1663
β-sheet
1
675
Figure 4, at 40 min of reaction time, very high ee (>90%)
of L-amino acid could be achieved in all IL solutions (up
to 4.0 M). Considering the margins of error, the ee does
1
636 1647
or turn
1
628
-sheets
β
β-sheets
1653
α-helix
1684
not seem to strongly depend on the IL concentration,
44,45
which implies that unlike many inorganic salts
or other
1
600
1620
1640
1660
1680
1700
22,46
hydrophilic ILs,
[EMIM][CH COO] does not strongly
3
-
1
Wavenumber (cm )
interact with proteins and destabilize the enzyme at rela-
tively high concentrations. This observation further con-
firms that this organic salt is an enzyme-‘friendly’ IL.
Figure 2. The second derivative spectra (FT-IR) of Bacillus licheniformis
protease after incubation in various media for 40 min at 30 ꢁC (the peak
2
4
assignment was based on the literature as discussed in our recent paper,
ꢁ1
Figure 4 also suggests that the yield of L-phenylalanine is
independent of the IL content at low concentrations (from
while assigning the 1671 cm peak as a b-turn was based on a different
7
0
article ).
0
to 1.0 M), but decreases dramatically when the IL con-
tent is greater than 2.0 M (suggesting slower reaction rates
and lower enzyme activities). At first, we suspected that it
was because of mass-transfer limitations caused by the high
viscosity of concentrated IL solutions. Therefore, we mea-
sured the viscosities of IL solutions at 30 ꢁC (Fig. 5). The
viscosity of IL solution was found to be almost linearly
increasing with the IL concentration. However, the yield
was not linearly decreasing with the IL concentration, espe-
cially in the 0–1.0 M range. On the other hand, the increase
of viscosity from 0.5 to 4.0 M is less than 0.2 cP, which is a
very small increment when compared with the viscosity of
pure IL (11.2 cP). Therefore, it is not very likely that the
viscosity is the major factor in determining the enzyme
activity in IL solutions (at least up to 4.0 M). We also
4
0 min, most peaks basically disappeared indicating that
the protein structures were seriously disrupted by the
organic solvent.
Figure 3 further illustrated the effect of two different ILs:
one bearing a kosmotropic anion (CH COO ) while the
other containing a chaotropic anion (OTs ). A better
enzymatic resolution (higher ee and E values) was achieved
in [EMIM][CH COO] solution than in [EMIM][OTs] solu-
tion. This observation is consistent with our previous stud-
ies:
ꢁ
29
3
ꢁ
3
1
1,22–24
kosmotropic anions stabilize the enzyme while
chaotropic ones destabilize it. Since [EMIM][CH COO]
3
has a unique combination of a kosmotropic anion and a
chaotropic cation, it is quite favorable for the protein sta-
bilization and enzyme activation. The effect of IL kosmo-
ee
yield
95.5
99.0
9
7.6
94.2
100.0
93.7
93.3
ee
E
80.0
1
00.0
93.7
8
9.0
8
6
4
2
0.0
0.0
0.0
0.0
60.0
4
0.0
0.0
31.0
2
1
7.3
0.0
0
M
0.5 M
1.0 M
2.0 M
3.0 M
4.0 M
0
.0
[
EMIM][CH COO] concentration
3
0
.5 M [EMIM][OAc] 0.5 M [EMIM][OTs]
Figure 3. Effect of IL kosmotropicity on the resolution of DL-phenyl-
alanine (lyophilized Bacillus licheniformis protease, 40 min reaction time,
Figure 4. Effect of [EMIM][CH
of L-phenylalanine (lyophilized Bacillus licheniformis protease, 40 min
reaction time, 30 ꢁC, the error bars indicating percentage errors).
3
COO] concentration on the ee and yield
30 ꢁC, the error bars indicating percentage errors).

2494
H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
1
.04
1
more water molecules to hydrate themselves, leading to a
dehydrated’ enzyme and lower enzyme flexibility. A simi-
‘
lar observation was found in the case of amino acids (as
model compounds of proteins): high concentrations of kos-
4
7
motropic salts dehydrate amino acids. When the IL con-
centration is very high, severe dehydration may lead to
0
0
0
0
.96
.92
.88
.84
4
2
protein unfolding. Figure 2 demonstrated that when the
protease was incubated in 4.0 M IL for 40 min, the charac-
teristic protein peaks are very weak compared with that in
water, suggesting a considerable structural change of pro-
tein molecules. Meanwhile, since water is also the substrate
in this reaction, a high IL concentration results in a low
4
8
0.0
1.0
2.0
3.0
4.0
5.0
water activity, causing a slow substrate (amino acid ester)
conversion. In summary, a high IL concentration may
change the native environment of the enzyme, or induce
the substrate ground-state stabilization.
[
EMIM][CH COO] concentration (M)
3
Figure 5. Viscosities of [EMIM][CH
3
COO] aqueous solutions at 30 ꢁC.
Dupont49 proposed a different explanation for the enzyme
stabilization by ILs using the structural organization of
imidazolium ILs. Based on a number of experiments, he
suggested the structure of pure 1,3-dialkylimidazolium
ILs (in solid, liquid, or even gas phase) to be similar to
those of hydrogen-bonded polymeric supramolecules. In
solutions of ILs, the hydrogen-bonded nano-structures
with polar and nonpolar regions can be formed. Therefore,
the structured IL network embraces the hydrated enzyme,
protecting it from losing essential water and from thermal
denaturation. This theory has been used to explain the
enzyme-stabilization by ILs. In our case, at higher IL
concentrations, the structure of IL network is more
enhanced and the enzyme is more tightly ‘included’ in the
network. As a result, the enzyme is less accessible by the
substrate, resulting in a lower yield.
conducted the same reaction in pure IL at 30 ꢁC, and no
considerable resolution was detected within 24 h. This is
consistent with the decreasing trend of L-yield with the
IL concentration. To further examine the mass-transfer
concerns, we compared the kinetic resolution at two differ-
ent substrate concentrations in 4.0 M [EMIM][CH COO]
3
(
Fig. 6). The L-yield is almost doubled when the substrate
concentration is doubled, suggesting that the overall reac-
tion rate is proportional to the substrate concentration
(
i.e., r / [S]). Therefore, the mass transfer due to viscosity
5
0
is not the limiting step (at least up to 4.0 M IL). However,
in other more viscous media, the viscosity may impact the
enzyme activity. In addition, Figure 6 also indicates that
5
the enzyme selectivity is lower when the substrate concen-
tration is increased. This could be explained by the reactiv-
ity–selectivity principle: the substrate is more reactive at a
higher concentration, thus a lower selectivity is expected.
Overall, considering that high ee and moderately high yield
could be achieved in 2.0 M [EMIM][CH COO], and the
3
One explanation of the decreasing yield (Fig. 4) is that high
concentrations of hydrophilic ILs (especially those carrying
strongly hydrated ions such as acetate) tend to ‘absorb’
2.0 M (ꢀ25%, v/v) solution has a relatively strong dissolu-
tion power towards hydrophobic substrates, we continued
to investigate this concentration for further study of other
enzymes and amino acids. One might suspect the advan-
tage of using 2.0 M IL instead of water as a reaction med-
ium, since the same high resolution could be achieved in
water. However, pure water may not be suitable for the res-
olution of hydrophobic substrates although it is perfectly
fine for phenylalanine methyl ester, since it is water soluble.
By using a reasonable amount of IL, we can increase the
dissolution ability of the media, especially when the hydro-
phobic substrates are present.
9
9.0
ee
1
00.0
90.7
yield
8
6
4
2
0.0
0.0
0.0
0.0
2.3. IL effect on different enzymes
38.2
In order to compare the enantioselectivities of different
19.7
enzymes in [EMIM][CH COO], we selected several other
3
enzymes frequently employed in enzymatic resolutions,
including two more proteases (alcalase and Amano P6)
and two lipases (PPL and Novozyme 435/immobilized
CAL-B). These enzymes were chosen based on the follow-
ing considerations: alcalase showed high enantioselectivi-
ties in the kinetic resolution of amino acids conducted in
0
.0
10
20
PA ester concentration (mg/mL)
5
1–55
56
organic solutions
and in low-IL solutions; Amano
Figure 6. Effect of substrate concentration (phenylalanine methyl ester,
or PA ester) on ee and L-yield at 30 ꢁC (40 min of reaction time, 4.0 M
P6 maintained high stabilities in 0.7 M solutions of pyr-
idinium and imidazolium based ILs including [EMIM][CH3-
[
3
EMIM][CH COO]).

H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
2495
COO];22 a highly enantioselective resolution of amino
Table 2. Enzymatic resolution of DL-amino acid esters in 2.0 M
57
[EMIM][CH COO] at 30 ꢁC
acids was achieved in phosphate buffer using PPL; Novo-
zyme 435 (or CAL-B) has been widely applied in many
enzymatic reactions performed in ILs, with high enzyme
3
Amino acid ester
Optimum
reaction
time (h)
ee (%)
(L-acid)
Yield (%)
(L-acid)
1
9,58–62
activities being reported.
However, low enzyme
enantioselectivities of alcalase and Novozyme 435 were
also seen in some IL systems.
Phenylalanine methyl ester
(in 2.0 M acetonitrile)
Phenylalanine methyl ester
0.33
1.5
2.0
1.0
70.0
96.3
98.3
95.5
33.9
>99
>99
>99
6
2
The data listed in Table 1 clearly illustrate that three pro-
teases produced higher ee than two lipases. Among the pro-
teases, the resolution catalyzed by the lyophilized protease
(2.0 M IL in H
Phenylalanine methyl ester
2.0 M IL in D O)
2
O)
(
2
4
-Chlorophenylalanine
ethyl ester
(
Bacillus licheniformis) exhibited the best enantioselectivity
and yield. In concentrated [EMIM][CH COO] solutions,
3
Methionine methyl ester
Phenylglycine methyl ester
p-Hydroxyphenylglycine
methyl ester
0.33
24
3.0
91.4
94.8
81.6
31.0
36.4
16.1
the lyophilized protease may not be completely solvated
as in water because the kosmotropic anion (acetate) tends
2
9,47
to pull water molecules away to hydrate itself
and thus
6
3
allow the protein to refold. Therefore, being similar to the
situation of lyophilized enzymes in organic solvents, pro-
teins are more rigid in IL solutions and aggregates of pro-
tein molecules may exist, which allows for possible control
of substrate specificity and protects enzymes from thermal
Note: 0.5 mg of Bacillus licheniformis protease lyophilizate was used in all
reactions.
much higher ee were observed in 2.0 M [EMIM][CH COO]
3
1
4,64–66
or chemical denaturation.
On the other hand, alca-
solutions (of H O: 96.3% ee, or in D O: 98.3 % ee; the
2
2
lase is an aqueous form of B. licheniformis protease. Com-
pared with lyophilized enzyme, the alcalase molecules have
a higher flexibility before dissolving in IL solutions. Once
alcalase is mixed with IL solutions, this enzyme produces
fewer aggregates than the lyophilized enzyme, which per-
mits protein molecules having more freedom to unfold in
a high IL environment. The third protease (Amano P6),
however, is a different type of protease obtained from
Aspergillus melleus, which might be liable from its different
activity. The poor resolutions by lipases might be due to
the insolubility of enzymes in IL aqueous solutions, caus-
ing poor interactions between substrates and enzyme active
sites, especially when the lipase is immobilized (Novozyme
heavy water effect is explained in Section 2.5). In general,
moderately high to very high resolutions of amino acids
were obtained in 2.0 M [EMIM][CH COO] solutions at
3
different optimum reaction times. These results suggest that
this IL system could be adopted by other similar enzymatic
reactions.
2.5. Heavy water effect of water on the enzymatic reaction
Figure 7 and Table 2 showed that the ee of L-amino acid
achieved in D
even slightly higher than) their respective ones in H
O solutions are very comparable with (or
2
O
2
4
35). Furthermore, our results indicate that the optimal IL
concentration for one particular enzyme (lyophilized prote-
ase) may not always be the best concentration for other en-
zymes, because different biological macromolecules are
stabilized by different degrees of ion kosmotropicity.44,67,68
ee
yield
9
8.2
96.6
92.6
93.8
100.0
Table 1. Enzymatic resolution of phenylalanine in 2.0 M [EMIM][CH
COO] by various enzymes at 30 ꢁC
3
-
8
6
0.0
0.0
Enzyme
Optimum reaction
time (min)
ee (%)
(L-acid)
Yield (%)
(L-acid)
Lyophilized protease
Alcalase
Amano protease P6
PPL
90
20
40
20
20
96.3
78.6
69.8
47.6
1.4
>99
45.2
26.7
11.3
7.6
40.0
Novozyme 435
20.0
Note: The amount of enzymes used is as follows: alcalase 20 mg, Novo-
zyme-435 1.0 mg and other enzymes 0.5 mg.
0.0
2
.4. Enzymatic resolution of different amino acids
H O
D O
2.0 M IL
2.0 M IL
2
2
(
H O)
(D O)
2
2
In order to demonstrate the optimum conditions (2.0 M IL
and lyophilized protease) on more substrates, we further
conducted the resolution reactions of several other amino
acids. Table 2 indicates that at optimum reaction times, a
poor resolution (70.0% ee and 33.9% yield) of phenyl-
alanine was produced in 2.0 M acetonitrile solution, while
Reaction medium
Figure 7. Effect of heavy water on the enzymatic resolution of phenyl-
alanine methyl ester (IL is [EMIM][CH COO], lyophilized Bacillus
licheniformis protease, 1 h reaction time, 30 ꢁC, the error bars indicating
3
percentage errors).

2496
H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
reaction was performed in a basic buffer, the H/D exchange
of the imidazolium cation could occur at the C-2 posi-
7
4
1
75
tion. Through H NMR determination, Nguyen et al.
observed such a H/D exchange occurring in imidazolium
cations. As a result, the imidazolium cation becomes par-
tially deuterated, and thus its hydration behavior and kos-
motropicity in D O are expected to be different from those
2
in H O.
2
3. Conclusion
The lyophilized B. licheniformis protease showed a very
high enantioselectivity and activity in up to 4.0 M
water
heavywater
[
EMIM][CH COO], qualifying this IL as an enzyme-
3
‘
friendly’ solvent. The high enantioselectivity is almost
2.0 M IL (heavy water)
independent of the IL concentration, while lower yields
were observed in high ionic media. High enantioselective
separations of several other amino acids were also observed
in 2.0 M [EMIM][CH COO]. Deuterium oxide (D O) is
1600
1620
1640
1660
1680
1700
-
1
Wavenumber (cm )
3
2
able to stabilize the enzyme, especially with the presence
of [EMIM][CH COO], producing high ee and yield.
Figure 8. The second derivative spectra (FT-IR) of Bacillus licheniformis
protease after incubation in various media (H O or D O) for 40 min at
0 ꢁC.
3
2
2
3
4
. Experimental
solutions (with or without [EMIM][CH COO]), implying
3
4.1. Materials
that the enzyme is stabilized by D O. Furthermore,
2
enhanced ee was observed with the addition of IL, suggest-
The following chemicals and enzymes were purchased
from Sigma–Aldrich: 1-ethyl-3-methylimidazolium bromide
ing that [EMIM][CH COO] activates the enzyme. A similar
3
observation was found in the D O solution of another IL
2
(
[EMIM]Br), 1-ethyl-3-methylimidazolium tosylate ([EMIM]-
[OTs]), silver acetate, B. licheniformis protease lyophilizate
subtilisin Carlsberg, 12 U/mg, product number 85968),
named [EMIM][5-APA] (where 5-APA is 5-aminopentano-
6
9
ate). However, a lower yield was obtained in pure D O
2
(
comparing with that in pure H O, indicating a slower reac-
2
alcalase (B. licheniformis protease, P2.4 U/g, product
number P4860), procine pancreas lipase (PPL, 30–90
U/mg using triacetin, product number L3126), Novozyme
tion rate in D O. It is known that deuterium oxide
2
decreases cellular metabolism because its higher molecular
mass (than H O) reduces the reaction rates. The rate reduc-
2
435 immobilized on acrylic resin (lipase B from Candida
tion is due to two reasons: (1) the enzyme is more rigid in
3
7
antarctica so-called CAL-B, product number 537322), DL-
phenylalanine methyl ester hydrochloride, DL-methionine
methyl ester, 4-chlorophenylalanine ethyl ester, DL-p-
hydroxyphenylglycine methyl ester, deuterium oxide
D O than in H O, and (2) the H/D exchange of peptide
2
2
7
0
NH protons modifies the protein properties. However,
very comparable yields were achieved in both H O and
2
D O with the presence of IL (2.0 M [EMIM][CH COO]).
2
3
(99.96 atom% D), and other reagents. Protease P ‘Amamo’
6
6
(produced from a selected strain of Aspergillus melleus,
0 U/mg) was a kind gift from the Amano Enzyme USA.
As shown in Figure 8, the second derivative spectra of pro-
tease in H O, D O and 2.0 M IL (in D O) are not quite dis-
2
2
2
tinguishable, suggesting that all these solvents are capable
of preserving the protein secondary structures and enabling
high enzyme activities. The quantitative analyses of second
4.2. IL preparations
7
0
derivative spectra by Dong et al. suggest that although
the H/D exchange does not change the secondary struc-
tures of proteins, the structural elements (a-helix, b-sheets,
b-turns and unordered) of proteins in H O and D O are
1-Ethyl-3-methylimidazolium acetate ([EMIM][CH COO])
was synthesized by a slow addition of an aqueous
[EMIM]Br solution into an equimolar Ag(CH COO) solu-
3
3
tion. The reaction was covered by the aluminum foil to pre-
vent the photodegradation of silver acetate. The reaction
mixture was stirred at room temperature for 2 h, followed
by a removal of AgBr precipitate through filtration. Char-
coal was added to the filtrate to remove color and impuri-
ties overnight. After filtering off the charcoal, water was
removed from the filtrate through rotary evaporation
under vacuum at 60 ꢁC. The resulting IL is a slightly vis-
cous and colorless liquid at room temperature. The absence
2
2
different, possibly resulting in different enzyme conforma-
5
9,71
tion and catalytic activity.
On the other hand, we also suspected that the hydration of
+
[
EMIM] in D O might be different from that in H O,
2 2
+
affording different kosmotropicity. The [EMIM] cation
in D O may experience H/D exchange. The C-2 hydrogen
2
is the most acidic hydrogen (pK = 21–23) among all
a
ꢁ
+
hydrogens on the imidazolium ring, and may be deproto-
of Br and Ag in the IL was examined by 0.1 M AgNO
3
7
2–74
nated under basic conditions.
Because our enzymatic
and 0.1 M HCl solutions, respectively. The IR, HPLC, and

H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
2497
1
H NMR measurements confirmed that the prepared IL
contains no measurable impurities, including water.
tween two CaF windows (Aldrich, 25 mm · 2mm). A Shi-
2
madzu FT-IR 8300 equipped with a dTGS detector was
used to measure the infrared spectra (% transmittance)
ꢁ
1
4
.3. Enzymatic resolution
through averaging 32 scans at 2 cm resolution (using
the Happ-Genzel apodization). The instrument was con-
trolled by the Shimadzu software IRsolution 1.20, and
second derivative spectra were calculated via the 9-point
Savitsky-Golay function provided by the same software.
DL-Amino acid (10 mg) was dissolved in 1.0 mL solvent
consisting of an IL and 0.2 M NaHCO buffer. Immedi-
ately, 0.5 mg of enzyme was added to the reaction mixture
at time zero. The reaction was shaken and maintained at
3
3
0 ± 1 ꢁC. The samples were withdrawn from the reaction
mixture periodically and analyzed by a chiral HPLC. All
experiments were run in duplicates. The averaged values
are reported.
Acknowledgments
This project was supported by NSF HBCU-UP 2006
Research Program, and a CASTME Research Project
under the Title III grant from the Department of Education.
4
.4. HPLC analysis
A Schimadzu LC-10AT HPLC is equipped with a SPD-
0A UV–Vis dual wavelength detector, and a Crownpak
CR(+) chiral column (150 mm · 4.0 mm, particle size
lm). The flow rate was 1.0 mL/min of 0.1 M HClO solu-
tion. The detection wavelength is 254 nm. The typical elut-
ing sequence of four isomers is in the increasing retention
times of D-acid < L-acid < D-ester < L-ester.
1
References
5
1. van Rantwijk, F.; Madeira Lau, R.; Sheldon, R. A. Trends
Biotechnol. 2003, 21, 131–138.
. Park, S.; Kazlauskas, R. J. Curr. Opin. Biotechnol. 2003, 14,
32–437.
. Kragl, U.; Eckstein, M.; Kaftzik, N. Curr. Opin. Biotechnol.
002, 13, 565–571.
4
2
3
4
4
2
4
.5. Calculations of ee, L-yield and E
. Sheldon, R. A.; Lau, R. M.; Sorgedrager, M. J.; van
Rantwijk, F.; Seddon, K. R. Green Chem. 2002, 4, 147–151.
5. Yang, Z.; Pan, W. Enzyme Microb. Technol. 2005, 37, 19–
28.
6. Ionic Liquids in Synthesis; Wasserscheid, P., Welton, T., Eds.;
Wiley-VCH: Weinheim, 2003.
The ee of L-amino acid (so-called ‘eep,’ ee of product) was
calculated from the HPLC integration area as (L area ꢁ D
area)/(L area + D area) · 100%. The percentage yield of
L-acid (shorten as yield or L-yield, its maximum is 100%
for a complete conversion of L-ester) was calculated from
comparing the current area of L-acid with that of complete
conversion of L-ester. The L-acid area for the complete con-
version of L-ester was determined individually for each
sample by HPLC, typically at 2–3 h of extended reaction
time. The enantiomeric ratio (E) was calculated from the
7
. Seddon, K. R. J. Chem. Technol. Biotechnol. 1997, 68, 351–
56.
3
8
9
. Zhao, H.; Malhotra, S. V. Aldrichim. Acta. 2002, 35, 75–83.
. Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.;
Watanabe, M. J. Phys. Chem. B 2004, 108, 16593–16600.
1
0. Tokuda, H.; Hayamizu, K.; Ishii, K.; Susan, M. A. B. H.;
Watanabe, M. J. Phys. Chem. B 2005, 109, 6103–6110.
7
6
following formula as defined by Chen et al.
11. Zhao, H. J. Mol. Catal. B: Enzym. 2005, 37, 16–25.
2. Zaks, A.; Klibanov, A. M. Proc. Natl. Acad. Sci. U.S.A.
985, 82, 3192–3196.
1
ln½1 ꢁ cð1 þ eeðPÞÞꢂ
E ¼
1
ln½1 ꢁ cð1 ꢁ eeðPÞÞꢂ
13. Bell, G.; Halling, P. J.; Moore, B. D.; Partridge, J.; Rees, D.
G. Trends Biotechnol. 1995, 13, 468–473.
14. Gupta, M. N. Eur. J. Biochem. 1992, 203, 25–32.
where c = 1 ꢁ (A + B)/(A + B ) and ee(P) = (P ꢁ Q)/
0
0
(
P + Q). A and B are concentrations of a pair of enantio-
1
1
1
1
1
5. Zaks, A.; Klibanov, A. M. J. Biol. Chem. 1988, 263, 3194–
mers, A and B are their initial concentrations.
3201.
0
0
6. Laane, C.; Boeren, S.; Vos, K.; Veeger, C. Biotechnol. Bioeng.
1
987, 30, 81–87.
4
.6. Measurement of viscosity
7. Anderson, J. L.; Ding, J.; Welton, T.; Armstrong, D. W. J.
Am. Chem. Soc. 2002, 124, 14247–14254.
8. Swatloski, R. P.; Spear, S. K.; Holbrey, J. D.; Rogers, R. D.
J. Am. Chem. Soc. 2002, 124, 4974–4975.
9. Kaar, J. L.; Jesionowski, A. M.; Berberich, J. A.; Moulton,
R.; Russell, A. J. J. Am. Chem. Soc. 2003, 125, 4125–4131.
20. Zaks, A.; Klibanov, A. M. J. Biol. Chem. 1988, 263, 8017–
8021.
The viscosity of IL or IL aqueous solution was determined
by a Cannon–Fenske Routine (CFR) viscometer. The vis-
cometer was incubated in a water bath maintained at
3
0 ± 1 ꢁC. The distilled water was used as a standard com-
77
pound (g = 0.7977 at 30 ꢁC). Triplicates were measured
for each sample. The relative errors are less than 3%. The
viscosity of pure [EMIM][CH COO] (dried) was deter-
mined as 11.2 ± 0.3 cP.
21. van Rantwijk, F.; Secundo, F.; Sheldon, R. A. Green Chem.
3
2
006, 8, 282–286.
2
2
2
2. Zhao, H.; Olubajo, O.; Song, Z.; Sims, A. L.; Person, T. E.;
Lawal, R. A.; Holley, L. A. Bioorg. Chem. 2006, 34, 15–25.
3. Zhao, H.; Campbell, S.; Jackson, L.; Song, Z.; Olubajo, O.
Tetrahedron: Asymmetry 2006, 17, 377–383.
4. Zhao, H.; Campbell, S.; Solomon, J.; Song, Z.; Olubajo, O.
Chin. J. Chem. 2006, 24, 580–584.
4
.7. FT-IR measurements
The method is a modification of a literature method:71
0
.5 mg of enzyme was incubated in 1.0 mL of 0.2 M NaH-
CO aqueous solution of IL or organic solvent at 30 ꢁC.
25. Collins, K. D.; Washabaugh, M. W. Quart. Rev. Biophys.
1985, 18, 323–422.
3
The solution was periodically withdrawn and placed be-

2498
H. Zhao et al. / Tetrahedron: Asymmetry 17 (2006) 2491–2498
2
6. Bostr o¨ m, M.; Williams, D. R. M.; Ninham, B. W. Curr. Opin. 54. Chen, S.-T.; Chen, S.-Y.; Hsiao, S.-C.; Wang, K.-T. Biotech-
Colloid Interface Sci. 2004, 9, 48–52.
nol. Lett. 1991, 13, 773–778.
2
2
7. Wiggins, P. M. Microbiol. Rev. 1990, 54, 432–449.
8. Nandi, P. K.; Robinson, D. R. J. Am. Chem. Soc. 1972, 94,
55. Zhao, H.; Luo, R. G.; Wei, D.; Malhotra, S. V. Enantiomer.
2002, 7, 1–3.
1
299–1308.
56. Zhao, H.; Malhotra, S. V. Biotechnol. Lett. 2002, 24, 1257–
1260.
57. Houng, J.-Y.; Wu, M.-L.; Chen, S.-T. Chirality 1998, 8, 418–
422.
2
3
9. Zhao, H. J. Chem. Technol. Biotechnol. 2006, 81, 877–891.
0. Huddleston, J. G.; Visser, A. E.; Reichert, W. M.; Willauer,
H. D.; Broker, G. A.; Rogers, R. D. Green Chem. 2001, 3,
1
56–164.
58. Husum, T. L.; Jorgensen, C. T.; Christensen, M. W.; Kirk, O.
Biocatal. Biotransform. 2001, 19, 331–338.
59. De Diego, T.; Lozano, P.; Gmouh, S.; Vaultier, M.; Iborra,
J. L. Biomacromolecules 2005, 6, 1457–1464.
60. Kim, K. W.; Song, B.; Choi, M. Y.; Kim, M. J. Org. Lett.
2001, 3, 1507–1509.
3
3
1. Rhodes, V. Chim. Oggi (Chem. Today) 2002, 20.
2. Chorghade, M. S.; Gurjar, M. K.; Talukdar, A. Chim. Oggi
(
Chem. Today) 2002, 20.
3
3
3. Stinson, S. C. Chem. Eng. News 2001, 79, 79–97.
4. Science News. 1933, October 14 Issue. Online, http://
www.sciencenews.org/articles/20031018/timeline.asp (accessed
April 2006).
61. Park, S.; Kazlauskas, R. J. J. Org. Chem. 2001, 66, 8395–
8401.
3
3
5. Maybury, R. H.; Katz, J. J. Nature 1956, 177, 629–630.
6. Chakrabarti, G.; Kim, S.; Gupta, M. L., Jr.; Barton, J. S.;
Himes, R. H. Biochemistry 1999, 38, 3067–3072.
7. Cioni, P.; Strambini, G. B. Biophys. Chem. 2002, 82, 3246–
62. Liu, Y.-Y.; Lou, W.-Y.; Zong, M.-H.; Xu, R.; Hong, X.; Wu,
H. Biocatal. Biotransform. 2005, 23, 89–95.
63. Schiffer, C. A.; Dotsch, V. Curr. Opin. Biotechnol. 1996, 7,
428–432.
64. Dabulis, K.; Klibanov, A. M. Biotechnol. Bioeng. 1993, 41,
566–571.
65. Dordick, J. S. Biotechnol. Prog. 1992, 8, 259–267.
66. Yang, L.; Dordick, J. S.; Garde, S. Biophys. J. 2004, 87, 812–
821.
67. Wiggins, P. M. Cell. Mol. Biol. 2001, 47, 735–744.
68. von Hippel, P. H.; Schleich, T. The effects of neutral salts on
the structure and conformational stability of macromolecules
in solution. In Structure and Stability of Biological Macro-
molecules; Timasheff, S. N., Fasman, G. D., Eds.; Marcel
Dekker: New York, 1969; pp 417–574.
69. Zhao, H.; Jackson, L.; Song, Z.; Olubajo, O. Tetrahedron:
Asymmetry 2006, 17, 1549–1553.
70. Dong, A.; Matsuura, J.; Allison, S. D.; Chrisman, E.;
Manning, M. C.; Carpenter, J. F. Biochemistry 1996, 35,
1450–1457.
71. Lau, R. M.; Sorgedrager, M. J.; Carrea, G.; van Rantwijk, F.;
Secundo, F.; Sheldon, R. A. Green Chem. 2004, 6, 483–487.
72. Amyes, T. L.; Diver, S. T.; Richard, J. P.; Rivas, F. M.; Toth,
K. J. Am. Chem. Soc. 2004, 126, 4366–4374.
73. Magill, A. M.; Yates, B. F. Aust. J. Chem. 2004, 57, 1205–
1210.
3
3
3
4
3
253.
8. Tang, K. E. S.; Dill, K. A. J. Biomol. Struct. Dyn. 1998, 16,
97–411.
9. Makhatadze, G. I.; Clore, G. M.; Gronenborn, A. M. Nat.
Struct. Biol. 1995, 2, 852–855.
3
0. Chin, J. T.; Wheeler, S. L.; Klibanov, A. M. Biotechnol.
Bioeng. 1994, 44, 140–145.
4
4
1. Klibanov, A. M. Trends Biotechnol. 1997, 15, 97–101.
2. Carpenter, J. F.; Prestrelski, S. J.; Dong, A. Eur. J. Pharm.
Biopharm. 1998, 45, 231–238.
4
4
4
4
3. Surewicz, W. K.; Mantsch, H. H.; Chapman, D. Biochemistry
1
993, 32, 389–394.
4. von Hippel, P. H.; Schleich, T. Acc. Chem. Res. 1969, 2, 257–
65.
2
5. Biotransformations in Organic Chemistry; Faber, K., Ed.;
Springer: New York, 2000.
6. Turner, M. B.; Spear, S. K.; Huddleston, J. G.; Holbrey, J.
D.; Rogers, R. D. Green Chem. 2003, 5, 443–447.
7. Zhao, H. Biophys. Chem. 2006, 122, 157–183.
8. Magnusson, D. K.; Bodley, J. W.; Evans, D. F. J. Solution
Chem. 1984, 13, 583–587.
4
4
4
5
9. Dupont, J. J. Braz. Chem. Soc. 2004, 15, 341–350.
0. Lozano, P.; De Diego, T.; Gmouh, S.; Vaultier, M.; Iborra,
J. L. Biocatal. Biotransform. 2005, 23, 169–176.
1. Kijima, T.; Ohshima, K.; Kise, H. J. Chem. Technol.
Biotechnol. 1994, 59, 61–65.
74. Dupont, J.; Spencer, J. Angew. Chem., Int. Ed. 2004, 43,
5296–5297.
75. Nguyen, H.-P.; Kirilov, P.; Matondo, H.; Baboul e` ne, M. J.
Mol. Catal. A. Chem. 2004, 218, 41–45.
76. Chen, C.-S.; Fujimoto, Y.; Girdaukas, G.; Sih, C. J. J. Am.
Chem. Soc. 1982, 104, 7294–7299.
77. Lide, D. R. CRC Handbook of Chemistry and Physics, 84th
ed.; CRC Press: New York, 2003; Section 6, p 3.
5
5
5
2. Chen, S.-T.; Wang, K.-T.; Wong, C.-H. J. Chem. Soc., Chem.
Commun. 1986, 1514–1516.
3. Chen, S.-T.; Huang, W.-H.; Wang, K.-T. Chirality 1994, 6,
5
72–576.