Flexibility of lipase brought about by solvent effects controls its enantioselectivity in organic media

Article abstract of DOI:10.1246/bcsj.76.399

The behavior of the enantioselectivity of Candida rugosa lipase was studied in the esterification of 2-(4-substituted phenoxy)propionic acids with 1-butanol in aliphatic, aromatic, and ethereal solvents, (cyclohexane, heptane, toluene, benzene, isooctane, dibutylether, etc.). Changing the solvent from cyclohexane to tert-butyl methyl ether, the isotropic signal increased quickly and the spectral line narrowed in width. The enzyme enantioselectivity in organic solvents was mainly controlled by its flexibility. The enantioselectivity of lipase in organic solvents was closely correlated with the lipase flexibility brought about by the cooperative solvent effects rather than with a sole solvent property, e.g., dielectric constant and hydrophobicity.

Full text of DOI:10.1246/bcsj.76.399

c. Jpn.
© 2003 The Chemical Society of Japan
Bull. Chem. Soc. Jpn., 76, 399–403 (2003)
399
e Chemical
ciety of Ja-
n
e Chemical
ciety of Ja-
n
Flexibility of Lipase Brought About by Solvent Effects Controls Its
Enantioselectivity in Organic Media
Solvent Effects on Lipase Enantioselectivity
S. Ueji et al.
Shin-ichi Ueji,* Tomohiko Taniguchi, Takashi Okamoto,† Keiichi Watanabe,† Yasuhito Ebara, and
Hitoshi Ohta††,#
03
9
3
Division of Natural Environment and Bioorganic Chemistry, Faculty of Human Development and Sciences,
Kobe University, Nada, Kobe 657-8501
SJA8
09-2673
258
†The Graduate School of Science and Technology, Kobe University, Nada, Kobe 657-8501
††Department of Physics, Faculty of Science, Kobe University, Nada, Kobe 657-8501
02
(Received September 5, 2002)
02
The behavior of the enantioselectivity of Candida rugosa lipase was investigated in the esterification of 2-(4-substi-
tuted phenoxy)propionic acids with 1-butanol in aliphatic, aromatic, and ethereal solvents. The observed enantioselec-
tivity (E value) was greatly affected by the solvent nature. Two key solvent characteristics, dielectric constant (ε) and hy-
drophobicity (log P), failed to explain the solvent effects on the E value. On the other hand, the variation of the E value
was found to be successfully correlated with the lipase flexibility brought about by the solvent effects, the flexibility of
which was estimated by the ESR measurements of the spin-labeled lipase. Although organic chemists should often un-
dertake a solvent screening step in the biocatalytic asymmetric resolutions, the obtained correlation will provide a useful
guide in the solvent optimization step.
0.3
Since the pioneering study showing enzymatic activity even
Results and Discussion
in organic solvents by Klibanov et al.,1 a great number of
papers have reported that the enantioselectivity and catalytic
activity of enzymes can be affected by the nature of the organic
solvents used.2 Indeed, the enzymatic properties have been
changed profoundly on switching from one solvent to another.
A striking example of the solvent effects is the solvent-induced
inversion in the enantioselectivity of enzyme-catalyzed reac-
tions,3 although the stereochemical preference of enzyme is
considered to be its intrinsic property. Therefore, the choice of
the reaction medium (solvent engineering) provides an attrac-
tive strategy for researchers seeking the improvement of the
enzyme enantioselectivity, because an ultimate goal, especially
for organic chemists, is to control rationally the enantioselec-
tivity as a function of the reaction conditions. In spite of nu-
merous studies of the solvent effects on enzyme properties, an
understanding of what characteristics of the solvents are essen-
tial for the determination of the enzyme enantioselectivity is
still elusive
As a model reaction, we chose the esterification of 2-(4-sub-
stituted phenoxy)propionic acids with 1-butanol catalyzed by
Candida rugosa lipase MY in aliphatic, aromatic, and ethereal
organic solvents (Scheme 1); in the hydrophilic solvents such
as tetrahydrofuran, lipase MY displayed extremely low reac-
tivity and poor enantioselectivity. Lipase MY used here was
semi-purified by dialyzing and lyophilizing from crude materi-
al. The MALDI-TOF MS spectrum of the semi-purified lipase
showed a single parent peak, m/z 60.2 kD, which is consistent
with the molecular weight of Candida rugosa lipase.4 In the
actual reactions, 0.3 vol% of water was added to the reaction
medium, because lipase with the flexible conformation in-
duced by water added should be affected by the solvent nature
like a dielectric constant; without water, the enantioselectivity
was poor and insensitive to the change of the solvents (data not
shown). In a model reaction, lipase catalyzed preferentially
the R enantiomer of all the substrates used.
Here, we report that the enantioselectivity for the lipase-cat-
alyzed reaction in various organic solvents is closely correlated
with the lipase flexibility estimated by the ESR measurement.
Furthermore, the specific solvent effect, such as the formation
of solvent-substrate complex, apart from altering the lipase
flexibility, may be detected on the basis of the correlation ob-
tained.
Table 1 summarizes the results of the variation of the en-
antioselectivity (E value5) in the esterification of the substrates
1–3 by changing the solvents. It can be seen from the data of
Table 1 that the solvent nature affects significantly the enantio-
selectivity of lipase. The observed E values appear to show
some correlation with two key solvent characteristics, dielec-
tric constant (ε) and hydrophobicity (log P), as listed in Table
2. The plots of the E value for 1 against ε and log P are illus-
trated in Figs. 1 and 2, respectively; some scatter diagrams
were obtained. In a few studies, the behavior of the enzyme
enantioselectivity in organic solvents has been found to corre-
# Molecular Photoscience Research Center, Kobe University,
Nada, Kobe 657-8501

400 Bull. Chem. Soc. Jpn., 76, No. 2 (2003)
Solvent Effects on Lipase Enantioselectivity
Scheme 1. Lipase-catalyzed esterification of 2-(4-substituted phenoxy)propionic acids 1–5 with 1-butanol in aliphatic, aromatic,
and ethereal solvents.
Table 1. Effects of Organic Solvents on the Enantioselectivity in the Lipase-Catalyzed Esterification of the Substrates 1–3
with 1-Butanol
Substrate
Entry Solvent
1
2
3
Convn./% ee/%
E
Convn./% ee/%
E
Convn./% ee/%
E
1
2
3
4
5
6
7
8
9
Cyclohexane
Hexane
Heptane
Toluene
Benzene
Chlorobenzene
Isooctane
Dibutyl ether
Diisopropyl ether
t-Butyl methyl ether
43.6
43.0
38.6
40.5
38.6
44.2
35.7
43.7
35.6
34.5
61.2
66.3
71.4
54.5
61.9
62.0
80.0 13.9
77.9 14.8
88.2 25.8
93.8 51.4
6.5
8.0
9.3
5.5
6.2
5.8
42.6
46.0
39.4
44.6
33.1
36.2
45.3
37.4
39.0
26.2
61.1
61.6
66.4
42.8
51.5
48.6
66.6
72.3
5.8
7.0
7.5
4.0
4.0
3.4
8.6
9.5
36.6
36.5
36.7
38.6
46.4
46.7
39.0
36.9
37.2
26.2
61.8
63.2
65.7
56.3
46.2
44.2
74.6 10.9
77.1 12.7
75.0 10.8
84.1 15.5
6.0
6.3
7.0
4.4
4.8
4.5
74.6 10.9
84.5 16.0
10
Table 2. Solvent Property (Dielectric Constant: ε and Hydrophobicity: log P) and the Value of
Hi/(Ha + Hi) as a Direct Measure of the Lipase Flexibility Estimated from ESR Spectrum
Solvent property
Lipase flexibility
Entry
Solvent
εa)
log Pb)
3.2
Hi/(Ha + Hi)
0.36
1
2
3
4
5
6
7
8
9
Cyclohexane
Hexane
Heptane
Toluene
Benzene
Chlorobenzene
Isooctane
Dibutyl ether
Diisopropyl ether
t-Butyl methyl ether
2.22
1.88
1.92
2.38
2.28
5.62
1.95
3.08
3.88
4.50
3.5
4.0
0.41
0.43
0.47
0.54
0.62
0.63
0.68
0.72
2.5
2.0
2.8
4.5
2.9
1.9
1.4
10
0.80
a) Taken from Ref. 19. b) Taken from Ref. 20.
late with one of two solvent parameters, ε and log P.6 It is rea-
sonable, however, to expect that the variation of the E values is
brought about by the cooperative effects of these solvent prop-
erties.7 Also, it is known that enzyme should become more
flexible in solvents with high ε and/or large log P than in those
with low ε and/or small log P.6,8 From these facts, the cooper-
ative solvent effects on lipase are assumed to bring about the
alteration of its flexibility, sometimes accompanying a confor-
mational change. In other words, we consider that the cooper-
ative solvent effects on lipase may converge on its flexibility.
This assumption prompted us to estimate the lipase flexibili-
ty by the ESR measurements of the spin-labeled lipase under
the same medium conditions as the model reaction. The active
site (serine) of lipase was spin-labeled with 1-oxy-2,2,6,6-
tetramethyl-4-piperidyl-ethoxyphosphorofluoridate (TEMPO-
4-EPF). Figure 3 shows the typical ESR spectra of the spin-
labeled lipase suspended in aliphatic, aromatic, and ethereal
solvents, the spectrum of which is composed of two compo-
nents, the isotropic signal (Hi) and the anisotropic signal (Ha)
(see the spectrum in cyclohexane). In Fig. 3, on changing the
solvent from cyclohexane to t-butyl methyl ether, the isotropic
signal (Hi) was found to arise quickly and the spectral line nar-
rowed in width. This change of the ESR spectrum signifies
that the conformation of the lipase’s active site around the spin
label becomes more flexible.9,10 According to our method,11
the degree of the lipase flexibility was estimated by the change
in the ratio of Hi to (Hi + Ha), where Hi and Ha represent the
peak heights of the corresponding signal (Fig. 3). Table 2 also
includes the value of Hi/(Hi + Ha) as a direct measure of the
lipase flexibility altered by the change of the solvents.
To test the validity of the Hi/(Hi + Ha) value instead of the
solvent parameters, the E values for 1 and 2 listed in Table 1

S. Ueji et al.
Bull. Chem. Soc. Jpn., 76, No. 2 (2003) 401
Fig. 1. Plot of the E value in lipase-catalyzed esterification
of 1 against dielectric constant (ε). Numbering (1–10)
indicates the corresponding solvents (see Table 1).
Fig. 3. Typical ESR spectra of the spin-labeled lipase sus-
pended in aliphatic, aromatic, and ethereal solvents.
Fig. 2. Plot of the E value in the lipase-catalyzed esterifica-
tion of 1 against hydrophobicity (log P). Numbering
(1–10) indicates the corresponding solvents (see Table 1).
are plotted as a function of the Hi/(Hi + Ha) value in Fig. 4.
As judged from Fig. 4, the variation of the E value is domi-
nantly controlled by the lipase flexibility brought about by the
cooperative solvent effects, except for the case of the aromatic
solvents. Also, for the other two substrates, 4 with the bulky
substituent (X = t-Bu) and 5 with the small substituent (X =
H), the dependence of the lipase flexibility due to the solvent
effects upon the E value is almost the same as the profile
shown in Fig. 4, although the solvent sensitivity of the E value
is relatively small (Table 3). This small solvent sensitivity can
probably be attributed to a loose (X = H) or a tight (X = t-Bu)
accommodation of the substrate into the lipase's active site.
Thus, for a combination of lipase MY and the substrates 1–5
studied here, the enantioselectivity is found to be correlated
Fig. 4. Correlation between the Hi/(Ha + Hi) value as a
direct measure of lipase flexibility and the E value in the
lipase-catalyzed esterification of 1 (ꢀ) and 2 (ꢁ). Num-
bering (1–10) indicates the corresponding solvents (see
Table 1).
with the increased flexibility of lipase brought about by the
solvent effects.
In contrast to our findings, there is a possibility that the in-

402 Bull. Chem. Soc. Jpn., 76, No. 2 (2003)
Solvent Effects on Lipase Enantioselectivity
Table 3. Effects of Organic Solvents on the Enantioselectivity in the Lipase-Catalyzed Esterification
of the Substrates 4 and 5 with 1-Butanol
Substrate
Entry
Solvent
4
5
Convn./%
37.2
ee/%
44.2
41.5
50.2
39.3
41.6
43.9
47.2
55.6
68.2
70.1
E
Convn./%
33.9
ee/%
46.9
47.4
50.1
26.5
24.8
27.1
48.2
49.4
64.5
67.2
E
1
2
3
4
5
6
7
8
9
Cyclohexane
Hexane
Heptane
Toluene
Benzene
Chlorobenzene
Isooctane
Dibutyl ether
Diisopropyl ether
t-Butyl methyl ether
3.2
3.2
4.1
3.0
3.0
4.0
3.9
5.4
7.9
8.2
3.5
4.0
3.9
2.1
1.9
2.0
4.2
4.1
6.8
7.1
38.2
42.8
41.7
35.7
36.2
43.2
44.6
38.3
45.1
35.5
32.3
34.1
35.2
45.8
40.9
38.3
10
22.1
20.9
method.17
creased flexibility of enzyme causes an acceleration of the
reactivity of the incorrectly binding enantiomer relative to that
of the correctly binding counterpart, thus leading to the loss of
the E value. The direction in the effect of the enzyme flexibili-
ty on its enantioselectivity seems to depend on a special rela-
tionship between the structure of a given substrate and the
stereochemical environment around the active site of a given
enzyme. In any event, it is noteworthy that the enzyme enanti-
oselectivity in organic solvents is mainly controlled by its flex-
ibility and that the obtained correlation (Fig. 4) permits a pre-
diction of the solvent effects on the E value.
As to the observed deviation of the aromatic solvents corre-
sponding to numbers 4, 5, and 6 in Fig. 4, one of the possible
explanations is a particular contribution due to the OH≥π
association12,13 between the carboxylic proton of the substrate
and the π electrons of the aromatic solvents or the local sol-
vent-enzyme interaction at the close vicinity of enzyme active
site reported by Nakamura et al.14
The direct measurements of enzyme flexibility in organic
solvents are limited and are performed by solid state NMR,15
ESR,9 and time-resolved fluorescence anisotropy studies.16
Among them, to our knowledge, only one study showed that
the increased enantioselectivity of subtilisin Carlsberg in
organic solvents was correlated with its increased flexibility.16
Our results obtained for lipase are consistent with those for
subtilisin.
Lipase-Catalyzed Esterification. The substrates 1–5 (0.036
mmol) and 1-butanol (1.08 mmol, 30 mol amt.) were dissolved in
an organic solvent (2 mL). To the solution, 0.3 vol% of water was
added, followed by ultrasonic dispersion, and then the semipuri-
fied lipase (2 mg) was added. The suspension was shaken (170
strokes/min) at 37 °C. The E value was calculated from the enan-
tiomeric excess (ee) for the butyl ester produced, according to the
literature.5 The ee was measured by HPLC on a chiral column
(Chiralcel OK, from Daicel Chemical Industries Co. Ltd., Japan).
MALDI-TOF MS and ESR Spectra. The MALDI-TOF MS
spectrum of the sample prepared in a sinapic acid (matrix) was ob-
tained with a Shimadzu AXIMA-CFR. The active site (serine) of
the semi-purified lipase was spin-labeled with 1-oxy-2,2,6,6-tetra-
methyl-4-piperidyl-ethoxyphosphorofluoridate (TEMPO-4-EPF)
purchased from SIGMA, according to the procedure reported by
Morrisett and Broomfield.18 It can be assumed that the spin label
has attached to the active site, because the spin-labeled lipase
showed a clear decrease in the enzymatic activity for the esterifi-
cation of 1. Typically about 35% of the active sites is considered
to be labeled as calculated from the residual enzymatic activity.
All the ESR measurements were carried out at room temperature
(ca. 25 °C) on a Bruker EMX081 spectrometer at X-band frequen-
cy in organic solvents containing 0.3 vol% of water.
We thank Meito Sangyo Co., Ltd., for its generous gifts of
lipases.
In conclusion, the enantioselectivity of lipase in organic sol-
vents is found to be closely correlated with the lipase flexibili-
ty brought about by the cooperative solvent effects rather than
with a sole solvent property such as ε or log P. The success of
the correlation should be of interest to organic chemists, be-
cause somewhat rational approaches to the solvent choice as
the reaction medium can be made for enantioselective synthe-
sis by use of enzyme.
References
1
a) A. Zaks and A. M. Klibanov, Science, 224, 1249 (1984).
b) A. Zaks and A. M. Klibanov, J. Biol. Chem., 263, 8017 (1988).
For recent reviews, see: a) C. R. Wescott and A. M.
2
Klibanov, Biochim. Biophys. Acta, 1206, 1 (1994). b) G. Garrea,
G. Ottolina, and S. Riva, Trends Biotechnol, 13, 63 (1995). c) G.
Garrea and S. Riva, Angew. Chem., Int. Ed., 39, 2226 (2000). d)
A. M. Klibanov, Nature, 409, 241 (2001).
Experimental
Materials. Lipase MY was supplied from Meito Sangyo Co.,
Ltd., and was semi-purified by dialyzing and lyophilizing from the
crude material. Organic solvents were purchased from Wako Pure
Chemical Industries, Ltd., Japan. Racemic 2-(4-substituted phen-
oxy)propionic acids 1–5 were prepared by the reaction of the cor-
responding 4-substituted phenol and ethyl 2-bromopropionate
(Tokyo Kasei Kogyo Co., Ltd., Japan), according to a known
3
a) S. Ueji, R. Fujino, N. Okubo, T. Miyazawa, S. Kurita,
M. Kitadani, and M. Muromatsu, Biotechnol. Lett., 14, 163
(1992). b) S. Tawaki and A. M. Klibanov, J. Am. Chem. Soc., 114,
1882 (1992). c) Y. Hirose, K. Kariya, J. Sasaki, Y. Kurono, H.
Ebike, and K. Achiwa, Tetrahedron Lett., 33, 7157 (1992).
4
a) N. Tomizuka, Y. Ota, and K. Yamada, Agr. Biol. Chem.,
30, 1090 (1966). b) R. J. Kazlauskas, Tibtech, 12, 464 (1994).

S. Ueji et al.
Bull. Chem. Soc. Jpn., 76, No. 2 (2003) 403
5
C.-S. Chen, Y. Fujimoto, G. Girdaukas, and C. J. Sih, J.
12 M. Oki, H. Iwamura, T. Onoda, and M. Iwamura, Tetra-
hedron, 24, 1905 (1968).
Am. Chem. Soc., 104, 7294 (1982).
6
For example, see: a) P. A. Fitzpatrick and A. M. Klibanov,
13 S. Ueji, K. Nakatsu, H. Yoshida, K. Kinoshita, Tetrahedron
Lett., 23, 1173 (1982).
14 K. Nakamura, Y. Takebe, T. Kitayama, and A. Ohno, Tetra-
hedron Lett., 31, 4941 (1991).
15 P. A. Burke, R. G. Griffin, and A. M. Klibanov, Biotechnol.
Bioeng., 42, 87 (1993).
J. Am. Chem. Soc., 113, 3166 (1991). b) E. Rubio, A. Fernandez-
Mayorales, and A. M. Klibanov, J. Am. Chem. Soc., 113, 695
(1991). c) A. Ducret, T. Trani, and R. Lortie, Enzyme Microb.
Technol., 22, 212 (1998).
7
For example, the solvent effects may be represented by αε
+ βlog P + γ (other factors such as dipole moment), where α, β
and γ are adjustment parameters depending on the reaction condi-
tions such as enzyme, substrate, temperature, etc.
16 J. Broos, A. J. W. G. Visser, J. F. J. Engbersen, W.
Verboom, A. van Hoek, and D. N. Reinhoudt, J. Am. Chem. Soc.,
117, 12657 (1995).
8
R. Affleck, C. H. Haynes, and D. S. Clark, Proc. Natl.
Acad. Sci., U. S. A., 89, 5167 (1992).
R. Affleck, Z.-F. Xu, V. Suzawa, K. Focht, D. S. Clark, and
17 D. T. Witiak, T. C.-L. Ho, and R. E. Hackney, J. Med.
Chem., 11, 1086 (1968).
18 J. D. Morrisett, C. A. Broomfield, and J. Broomfield, J.
Med. Chem., 247, 7224 (1972).
19 C. Reichardt, “Solvents and Solvent Effects in Organic
Chemistry,” VCH, Weinheim (1988).
9
J. S. Dordick, Proc. Natl. Acad. Sci., U. S. A., 89, 1100 (1992).
10 S. Ueji, M. Nishimura, R. Kudo, R. Matsumi, K. Watanabe,
and Y. Ebara, Chem. Lett., 2001, 912.
11 K. Watanabe, T. Yoshida, and S. Ueji, Chem. Commun.,
2001, 1260.
20 C. Laane, S. Boeren, K. Vos, and C.Veeger, Biotechnol.
Bioeng., 30, 81 (1987).