The role of conformational flexibility of enzymes in the discrimination between amino acid and ester substrates for the subtilisin-catalyzed reaction in organic solvents

Article abstract of DOI:10.1016/j.bioorg.2004.05.001

To investigate how the conformational flexibility of subtilisin affects its ability to discriminate between enantiomeric amino acid and ester substrates for the subtilisin-catalyzed reaction in an organic solvent, the flexibility around the active site and the surface of subtilisin was estimated from the mobility of a spin label bound to subtilisin by ESR spectroscopy. Many studies on enzyme flexibility focus on the active site. Both the surface and active site flexibility play an important role in the enantioselectivity enhancement of the enzyme-catalyzed reaction. It was found, however, that the different behavior observed for the enantioselectivity between the amino acid and ester substrates could be correlated with the flexibility around the surface rather than the flexibility at the active site of subtilisin. In other words, for the ester substrates, the greater flexibility around the surface of subtilisin induced by a conformational change resulting from the presence of an additive such as DMSO is essential for the enantioselectivity enhancement. This model is also supported by the Michaelis-Menten kinetic parameters for each enantiomeric substrate. Our findings provide insight into the enantioselectivity enhancement for the resolution of enantiomers for enzyme-catalyzed reactions in organic solvents.

Full text of DOI:10.1016/j.bioorg.2004.05.001

Bioorganic Chemistry 32 (2004) 504–515
www.elsevier.com/locate/bioorg
The role of conformational Xexibility of enzymes
in the discrimination between amino acid and
ester substrates for the subtilisin-catalyzed
reaction in organic solvents
Keiichi Watanabe,a Takashi Yoshida,b and Shin-ichi Uejia,b,¤
a Graduate School of Science and Technology, Kobe University, Nada, Kobe 657-8501, Japan
b Division of Natural Environment and Bioorganic Chemistry, Faculty of Human Development and Sciences,
Kobe University, Nada, Kobe 657-8501, Japan
Received 26 January 2004
Available online 6 July 2004
Abstract
To investigate how the conformational Xexibility of subtilisin aVects its ability to discrimi-
nate between enantiomeric amino acid and ester substrates for the subtilisin-catalyzed reaction
in an organic solvent, the Xexibility around the active site and the surface of subtilisin was esti-
mated from the mobility of a spin label bound to subtilisin by ESR spectroscopy. Many studies
on enzyme Xexibility focus on the active site. Both the surface and active site Xexibility play an
important role in the enantioselectivity enhancement of the enzyme-catalyzed reaction. It was
found, however, that the diVerent behavior observed for the enantioselectivity between the
amino acid and ester substrates could be correlated with the Xexibility around the surface
rather than the Xexibility at the active site of subtilisin. In other words, for the ester substrates,
the greater Xexibility around the surface of subtilisin induced by a conformational change
resulting from the presence of an additive such as DMSO is essential for the enantioselectivity
enhancement. This model is also supported by the Michaelis–Menten kinetic parameters for
each enantiomeric substrate. Our Wndings provide insight into the enantioselectivity enhance-
ment for the resolution of enantiomers for enzyme-catalyzed reactions in organic solvents.
 2004 Elsevier Inc. All rights reserved.
¤ Corresponding author. Fax: +81-78-803-7761.
E-mail address: ueji@kobe-u.ac.jp (S.-i. Ueji).
0045-2068/$ - see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bioorg.2004.05.001

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
505
Keywords: Enantioselectivity; Enzyme Xexibility; Subtilisin; Organic solvent; ESR spectroscopy
1. Introduction
One of the most challenging problems in organic syntheses is the production of
the enantiomerically pure compounds such as pharmaceutical and agricultural chem-
icals because only one of the enantiomers is generally biologically active. Recently,
biocatalysts have received much attention as enantioselective catalysts for the resolu-
tion of racemic mixtures [1]. Enzyme-catalyzed reactions can normally be performed
under much milder conditions (i.e., room temperature and near neutral pH) than
their chemical counterparts. It is also known that many enzymes like lipases or the
proteases (e.g., subtilisin and chymotrypsin) display enzymatic activity in organic sol-
vents [2]. Hence, the use of the enzymatic reactions is becoming increasingly preva-
lent in organic syntheses. For the last decade, many strategies have been developed to
achieve the high enantioselectivity including the use of additives in the reaction
medium [3], diVerent organic solvents [4], changing the reaction temperature [5], and
protein engineering by site-directed mutagenesis [6] or by the direct chemical modiW-
cation of the enzyme [7]. Among these strategies, the addition of a compound or a
solvent to the reaction medium is the most popular because of its simplicity. In a
recent review, various additives have been successfully used to achieve high enanti-
oselectivity and/or high activity of enzymes [8].
For enzyme-catalyzed reactions in organic solvents, however, enantioselectivity is
not always high enough to obtain chiral compounds, probably because these target
compounds are not the natural substrates for the enzyme with structures signiWcantly
diVerent from those of the natural ones. These non-physiological substrates require
suYcient Xexibility and/or a conformational change of the enzyme before binding
can take place at the active site. Thus, the eVect of a change in the Xexibility and/or
the conformation of the enzyme on the enantioselectivity is important to understand
so that high enantioselectivity can be achieved for the reactions using these non-nat-
ural substrates such as pharmaceutical and agricultural chemicals.
There is clearly a diVerence in the origin of the enhanced enantioselectivity for
amino acid and ester substrates in subtilisin-catalyzed reactions that use additives
in an organic solvent. The diVerent behavior is discussed in terms of the relation-
ship between the kinetic parameters for each enantiomer and the conformational
Xexibility of the active site and the surface of subtilisin. The Xexibility of the active
site and surface of subtilisin is estimated from the mobility of the spin label bound
to subtilisin by ESR spectroscopy. For model reactions, two popular additives,
water or dimethyl sulfoxide (DMSO), were selected to elucidate the eVect of the
conformational Xexibility of subtilisin on its enantioselectivity. The addition of
DMSO, which may correspond to greater Xexibility around the surface as well as
the active site of subtilisin, is found to improve the enantioselectivity and/or the
enzymatic activity for subtilisin-catalyzed reactions. The greater Xexibility of the
enzyme may be the critical step for generation of its enantioselectivity and/or

506
K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
activity, especially for the ester substrates. This is the Wrst report that the Xexibility
around the surface of subtilisin brought on by additives is involved in the control
of the enantioselectivity, although there are a few reports describing the relation-
ship between the enantioselectivity and the Xexibility in the active site of the
enzyme [9].
2. Materials and methods
2.1. Materials
Subtilisin Carlsberg (serine protease from Bacillus licheniformis, EC 3.4.21.14) was
purchased from Sigma Chemical and used without further puriWcation. Isooctane, 1-
butanol, and dimethyl sulfoxide were purchased from Wako Pure Chemical Indus-
tries, Japan. Organic solvents were dried over molecular sieves for more than 24h
before use. Each enantiomer of Z-Ala-OEt 1 [10], ethyl 2-(4-ethylphenoxy)propio-
nate 2, and methyl mandelate 3 [10] was prepared according to known methods. The
synthesis, resolution, and esteriWcation of 2 were carried out by literature procedures
[3b,10,11].
2.2. Subtilisin-catalyzed transesteriWcation in isooctane containing the additives
In a typical subtilisin-catalyzed transesteriWcation, each enantiomer of substrates
1–3 (0.01–0.025 mmol) and 1-butanol (0.06–0.15 mmol, 6 equiv.) was dissolved in dry
isooctane (2ml). To the solution, a small amount of DMSO or water was added, fol-
lowed by ultrasonic dispersion, and then subtilisin (10mg) was added. The suspen-
sion was shaken (170strokes/min) at 37 °C. At an appropriate time interval, aliquots
were withdrawn and the supernatant was analyzed by HPLC on a chiral column
(Daicel Chiralcel OJ) to determine the initial rate. All the experimental data points
gave a straight line with a correlation coeYcient 10.98.
2.3. Michaelis–Menten kinetic parameters
The kinetic study was carried out by measuring the initial rates of the transesteriW-
cation of each enantiomer of substrates 1 and 2. The amount of each enantiomer was
varied from 0.006 to 0.05mmol and the amount of 1-butanol was 0.15mmol. The
enantiomer and 1-butanol were dissolved in dry isooctane (2 ml). To the solution, a
small amount of DMSO or water was added, followed by the ultrasonic dispersion,
which was followed by the addition of subtilisin (10 mg). At an appropriate time
interval, aliquots were withdrawn and the supernatant was analyzed by HPLC on a
chiral column (Daicel Chiralcel OJ) to determine the initial rate at each substrate
concentration. The values of Km and Vmax were obtained from the Michaelis–Menten
equation. All the experimental data points obtained from the Lineweaver–Burk plot
gave a straight line with a correlation coeYcient 10.97. The values were reproducible
to about §5% on repeated runs.

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
507
2.4. ESR measurement
Subtilisin was spin-labeled at the active site Ser-221 with 1-oxy-2,2,6,6-tetra-
methyl-4-piperidinyl ethoxyphosphoroXuoridate (TEMPO-4-EPF) purchased from
Sigma Chemical as described by the literature [12]. Met spin-labeled subtilisin was
prepared by protecting the active site Ser-221 with the inhibitor (PMSF), and then
labeling at Met residues with the synthesized compound, 4-(2-iodoacetamide)-
TEMPO. The spin-labeled subtilisin was separated from the free label by the dialysis
for 10 h, followed by the lyophilization (¡50°C, 24h). A control sample of spin-
labeled subtilisin in the appropriate solvent was used to tune the ESR signal. Subse-
quently, a new sample was placed into the sample cavity, and the signal was quickly
Wne tuned after the sample reached the room temperature. For the same solvent con-
ditions of the subtilisin-catalyzed transesteriWcation, all the ESR spectra were
recorded at room temperature on a Bruker EMX081 spectrometer with a microwave
power of 6.3mW, a microwave frequency of 9.6GHz, and a sweepwidth of 300G
(resolution 1024 points).
3. Results and discussion
3.1. Additive eVects on the enantioselectivity and activity of subtilisin for amino acid or
ester substrates in transesteriWcation reactions
To investigate the enantioselectivity for subtilisin-catalyzed transesteriWcation in
organic solvents, we chose the amino acid derivative (Z-Ala-OEt 1) or the
carboxylic esters, ethyl 2-(4-ethylphenoxy)propionate 2, and methyl mandelate 3,
as substrates (Scheme 1). For these reactions, subtilisin used preferentially the
L enantiomer of 1, the S enantiomer of 2, and the R enantiomer of 3. Substrates 2
and 3 are important intermediates for useful agricultural and pharmaceutical
chemicals. The reactions were carried out in dry isooctane containing a small
amount of water [13] or dimethyl sulfoxide (DMSO) [3c,14] as additives because
these additives are known to improve the enantioselectivity and/or enzymatic activ-
ity. Moreover, in hydrophilic or aromatic solvents such as diethyl ether and tetra-
hydrofuran, or benzene, subtilisin displayed extremely low enzymatic activity (data
not shown).
Table 1 summarizes the initial rates for each enantiomer of 1–3 and the enantiose-
lectivity (ratio of the initial rate of the correctly binding enantiomer to the incorrectly
binding one: VL/VD for 1, VS/VR for 2, and VR/VS for 3) observed for the subtilisin-
catalyzed transesteriWcation in isooctane containing various amount of water or
DMSO. In the absence of additives, substrates 1–3 were not processed by subtilisin.
As shown in Table 1, when a small amount of water was added to the reaction
medium, the initial rate between the amino acid substrate and the ester substrate
diVers. For the model reaction using 1, the addition of water brought about the high
enantioselectivity with an increase of the initial rate for the correctly binding L enan-
tiomer, whereas the subtilisin-catalyzed reaction of 2–3 displayed extremely low

508
K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
Scheme 1.
enzymatic activity (data not shown). The initial rate for 2–3 was too slow to obtain
reproducible values.
Changing the additive from water to DMSO improved the enantioselectivity and
enzymatic activity with all substrates 1–3. As shown in Table 1, when DMSO was
added to the reaction medium, the initial rate for the correctly binding enantiomers of
1–3 increased, as compared with that for the incorrect ones. As a result, there was an
enhancement of the enantioselectivity. For the transesteriWcation of the correctly bind-
ing L enantiomer of 1, the addition of DMSO produced a signiWcant increase in the
initial rate, as compared with that for substrates 2–3. Upon the addition of the opti-
mum amount of DMSO (0.30vol%) to the reaction using 1, subtilisin displayed excel-
lent enantioselectivity VL/VD D26. The marked enhancement of the enantioselectivity
was conWrmed in model reactions using other amino acid substrates (for example, VL/
VD D105 for Z-Met-OEt, VL/VD D171 for Z-Phe-OEt, and VL/VD D928 for Z-Leu-
OEt). For all the reactions, a steep decrease in the initial rate was observed by addition
of an excess of DMSO (0.60vol%), although with the addition of more water (0.45–
1.0vol%) to these reactions, the enantioselectivity and the initial rate for each enantio-
mer of 1 remained almost constant VL/VD D ca. 85, VL D ca. 30, and VD Dca. 0.35 (data
not shown). DMSO may cause a conformational change of subtilisin.
3.2. Relationship between the enantioselectivity and activity and the conformational
Xexibility of subtilisin estimated from the ESR measurement of a spin-labeled enzyme
For subtilisin-catalyzed transesteriWcation in an organic solvent, the variation of
the enantioselectivity and/or enzymatic activity by the presence of additives is found

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
509
Table 1
The variation of the initial rate for each enantiomer and the enantioselectivity for the subtilisin-catalyzed
transesteriWcation of 1–3 with 1-butanol in isooctane with additives
Substrate Additivea Amount
(vol%)
Initial rate (nmol h¡1 mg¡1
)
Enantioselectivityb
Vcorrectly biding enantiomer Vincorrectly biding enantiomer
1
Water
0
0
10
24
32
0
0
0.35
0.38
—
—
69
84
0.150
0.300
0.450
DMSO
0
0
20
257
212
151
0
0
2.8
10
14
8.8
—
7.1
26
15
17
—
0.150
0.300
0.375
0.450
0.600
0
2
3
DMSO
DMSO
0
0
0
—
0.150
0.300
0.375
0.450
0.500
0.600
0.21
1.0
15
20
14
0.16
0.77
3.2
2.1
2.1
1.0
1.3
1.3
4.7
9.5
6.7
1.4
1.4
0
0
0
—
0.150
0.300
0.375
0.450
0.500
0.600
6.8
24
37
21
3.5
1.2
5.2
5.2
4.0
1.9
0
5.7
4.6
7.1
5.3
1.8
—
0
a For the addition of water, the subtilisin-catalyzed reaction of 2–3 displayed extremely low enzymatic
activity.
b Ratio of the initial rate of the correctly binding enantiomer to the incorrectly binding one.
to be signiWcantly dependent on the structure of the substrate. One could speculate
that the observed diVerence in the enantioselectivity and/or enzymatic activity
between the amino acid and ester substrates is primarily attributed to the additive-
induced Xexibility of subtilisin. Thus, in order to investigate how the conformational
Xexibility of subtilisin aVects the discrimination between the amino acid and ester
substrates, the mobility of a spin label bound to the active site Ser-221 of subtilisin
was examined in isooctane containing the various amounts of water or DMSO by
ESR spectroscopy. Fig. 1 shows the ESR spectra in isooctane containing 0 and
0.45 vol% of DMSO, in which two parts of the spectrum are arbitrarily labeled Ha
and Hi, respectively. According to our method [3c,4d], the degree of the enzyme's
Xexibility can be monitored by the change in the ratio of the signal height, Hi/
(Ha + Hi), because each signal Ha and Hi represents the anisotropy and the isotropy
of the spin label, respectively. Thus, the increase in the Hi/(Ha + Hi) value reXects an
increase in the mobility of the spin label. Fig. 2A shows the variation of the

510
K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
Fig. 1. The ESR spectra of the spin-labeled subtilisin in isooctane; (A) with no additive and (B) in the pres-
ence of DMSO (0.45 vol%).
Hi/(Ha + Hi) value for the spin label bound to Ser-221 as a function of the amount of
two additives, water and DMSO. The increase in the amount of each additive is
found to increase the Xexibility around the subtilisin's active site. Unfortunately,
however, as judged from the result that two additives gave the change of the Hi/
(Ha + Hi) value in the same manner, the observed diVerence in the enantioselectivity
and/or enzymatic activity between the amino acid and ester substrates cannot be
explained by the Xexibility around the active site of subtilisin.
To investigate the change in the subtilisin's Xexibility outside the active site, we
measured the ESR spectra of the spin-labeled subtilisin bound to the Met residues
(119, 125, 135, or 199), which are found on the surface of subtilisin. The Met spin-
labeled subtilisin was prepared by protecting the active site Ser-221 with the inhibi-
tor (PMSF), in order not to label Met-222 in the active site of subtilisin by the Met
spin label (see Section 2). The MALDI-TOF MS spectrum of the Met spin-labeled

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
511
Fig. 2. The variation of the Hi/(Ha + Hi) value for the spin-labeled subtilisin as a function of the amount of
two additives (᭺, water; ᭹, DMSO). (A) Hi/(Ha + Hi) value for the spin-label bound to Ser-221, (B) Hi/
(Ha + Hi) value for the spin-label bound to Met around the surface of subtilisin.
subtilisin suggested that only one spin label reacted with subtilisin, although the
position of the spin label (Met 119, 125, 135, or 199) was not determined. Thus,
taking into consideration that the spin label combined with the Met residues at ran-
dom, the Hi/(Ha + Hi) value represents the average Xexibility around the four Met
residues.
The variation of the Hi/(Ha + Hi) value as a function of the amount of two addi-
tives is depicted in Fig. 2B. In the absence of additives, subtilisin displays little Xexi-
bility around its surface (Hi/(Ha + Hi)D 0.4) as well as in its active site (see Fig. 2A;
Hi/(Ha + Hi) D0.1). This Wnding suggests that the subtilisin molecule is very rigid with
little Xexibility in isooctane without the additives. In fact, as seen in Table 1, the reac-
tion condition without the additives causes a serious decrease in the enzymatic activ-
ity. This decrease is likely due to inability of substrate to Wt into the rigid binding
pocket of subtilisin because of the steric bulk of the substrate.
Interestingly, the addition of 0.15vol% of DMSO produced greater Xexibility
around the surface of subtilisin (Hi/(Ha + Hi)D0.69) than is observed for the addition
of water (Hi/(Ha + Hi)D0.54) (Fig. 2B). However, there was little diVerence in the
Xexibility around the active site of subtilisin using either water or DMSO as additives
(see Fig. 2A; Hi/(Ha + Hi)D 0.25 for water, Hi/(Ha + Hi)D0.29 for DMSO). With the
addition of more than 0.30vol% of DMSO, the isotropic signal (Hi) became too satu-
rated to obtain the reliable Hi/(Ha + Hi) value. With water, the Hi/(Ha + Hi) value
changed in proportion to the amount of water added to the isooctane. This result can
be explained by the assumption that the addition of DMSO has a denaturing ability
on subtilisin which brings about the large change in the Xexibility and/or the confor-
mation of subtilisin through the speciWc interaction between the surface of the subtil-
isin and the ionic SDO group of DMSO. In view of this assumption, the larger
Xexibility around the surface rather than the active site of subtilisin is required to
overcome the steric problems of the ester substrate Wtting into the binding pocket of

512
K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
subtilisin. The Xexibility seems to be essential for a stable association between
substrate and enzyme. Therefore, the larger Xexibility caused by the addition of
DMSO is assumed to produce the increase in enzymatic activity in model reactions
using 2–3. In the absence of DMSO or water, subtilisin displayed extremely low
activity due to the loss of Xexibility. The critical diVerence in the variation of the
enantioselectivity and the enzymatic activity for the amino acid and ester substrates
may be due to the Xexibility around the surface of subtilisin.
3.3. EVect of the conformational Xexibility on the discrimination between enantiomers
of amino acid and ester substrates as assessed by kinetic studies
To elucidate the mechanistic details of the enhancement of the enantioselectivity
and/or the enzymatic activity brought about by the conformational Xexibility of sub-
tilisin, Michaelis–Menten kinetic parameters were determined for the subtilisin-cata-
lyzed transesteriWcation of 1 and 2 in isooctane containing a small amount of water
(Table 2) or DMSO (Tables 3 and 4). The data show that for the correctly binding L
enantiomer of 1 (Tables 2 and 3), the increase in the amount of additives, corre-
sponding to an increased Xexibility of subtilisin, brings about an increase in Vmax/Km
value where the Km value is almost unchanged and the Vmax value increases. The
small eVect of the subtilisin Xexibility on the Km value can probably be attributed to
the ready accommodation of the L enantiomer in the active site of subtilisin, because
Table 2
The Michaelis–Menten parameters for the subtilisin-catalyzed transesteriWcation of substrate 1 with 1-
butanol in isooctane containing a small amount of water
ConWguration
Water (vol%)
Km (mM)
Vmax (mmol h¡1 mg¡1
)
Vmax/Km (h¡ 1
)
L
0.150
0.300
0.450
0.150
0.300
0.450
15.9
14.0
13.7
4.21 £ 10¡5
1.3 £ 10¡4
L
9.37 £ 10¡5
3.3 £ 10¡3
L
1.19 £ 10¡4
4.3 £ 10¡3
a
a
a
D
D
D
151
151
1.13 £ 10¡5
3.7 £ 10¡5
l.14 £ 10¡5
3.8 £ 10¡5
a The initial rate for the D enantiomer of 1 is too slow to obtain the reproducible value.
Table 3
The Michaelis–Menten parameters for the subtilisin-catalyzed transesteriWcation of substrate 1 with 1-
butanol in isooctane containing a small amount of DMSO
ConWguration
DMSO (vol%)
Km (mM)
Vmax (mmol h¡1 mg¡1
)
Vmax/Km (h¡1
)
L
0.150
0.300
0.375
0.450
0.150
0.300
0.375
0.450
14.3
11.6
11.6
12.0
40.9
30.6
23.9
38.0
7.74 £ 10 ¡5
8.51 £ 10¡4
7.04 £ 10¡4
5.10 £ 10¡4
2.54 £ 10¡5
7.30 £ 10¡5
8.32 £ 10¡5
7.67 £ 10¡5
2.7 £ 10¡3
3.7 £ 10¡2
3.0 £ 10¡2
2.1 £ 10¡2
3.1 £ 10¡4
1.2 £ 10¡3
1.7 £ 10¡3
1.0 £ 10¡3
L
L
L
D
D
D
D

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
513
Table 4
The Michaelis–Menten parameters for the subtilisin-catalyzed transesteriWcation of substrate 2 with 1-
butanol in isooctane containing a small amount of DMSO
ConWguration
DMSO (vol%)
Km (mM)
Vmax (mmol h¡1 mg¡1
)
Vmax/Km(h¡1
)
S
S
S
S
R
R
R
R
0.300
0.375
0.450
0.500
0.300
0.375
0.450
0.500
109
87.4
68.1
80.1
157
147
143
151
7.51 £ 10¡5
8.77 £ 10¡5
1.10 £ 10¡4
9.71 £ 10¡5
2.51 £ 10¡5
4.31 £ 10¡5
3.56 £ 10¡5
2.84 £ 10¡5
3.4 £ 10¡4
5.0 £ 10¡4
8.1 £ 10¡4
6.1 £ 10¡4
8.0 £ 10¡5
1.4 £ 10¡4
1.2 £ 10¡4
9.4 £ 10¡5
the L enantiomer has the correct conWguration for the shape of the active site. Thus,
the greater Xexibility is not required for the formation of the acyl-enzyme in the reac-
tion using the L enantiomer. In addition, the unchanged Km value suggests that the
addition of the denaturing additive DMSO does not alter the shape of the active site.
As for the reactivity of the acyl-enzyme, however, the increase in the subtilisin Xexi-
bility favors the preferred orientation of the amino acid side chains to promote eVec-
tively the reaction of the acyl-enzyme with 1-butanol. Hence, the Vmax value increases
with the increased Xexibility of subtilisin.
In contrast to the kinetic parameters measured for the L enantiomer of 1, there is a
marked diVerence in the Km value for the D enantiomer when water and DMSO are
used as additives. The data in Tables 2 and 3 show that the addition of DMSO,
resulting in a greater Xexibility around the surface of subtilisin, produces a 0.2-fold
decrease in Km and a 7-fold increase in Vmax than those measured when water is used
as the additive. The greater Xexibility around the surface of subtilisin will enable the
accommodation of the D enantiomer of 1 into the active site and stabilize the com-
plex between the substrate and subtilisin. The small Xexibility brought about by addi-
tion of water is not suYcient to provide a stable association between the
enantiomer and subtilisin, thereby resulting in the large Km value (Table 2).
D
For substrate 2, the dependence of Km and Vmax on subtilisin Xexibility showed the
characteristic proWle, as compared with that of substrate 1 (Table 4). As can be seen
in Table 4, the increase in subtilisin Xexibility caused by addition of DMSO produced
a decreased Km value and an increased Vmax value for the correctly binding S enan-
tiomer. The large eVect of subtilisin Xexibility on Km and Vmax for the S enantiomer
of 2 can be also attributed to accommodation of the substrate into the active site
through the induced Wt for the substrate. In sharp contrast to the S enantiomer, the
Km and Vmax values for the incorrectly binding R enantiomer of 2 were not sensitive
to the increased Xexibility of subtilisin. This is probably because the R enantiomer is
the non-natural substrate with a structure signiWcantly diVerent from that of the nat-
ural one and has an incorrect conWguration for the shape of the active site of subtili-
sin. Thus, the diVerent enzymatic activities (Vmax/Km) for each enantiomer caused by
the Xexibility eVect produces the high enantioselectivity ((Vmax/Km)S/(Vmax
/
Km)R)D 6.8 for 0.45vol% DMSO).

514
K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
4. Conclusions
Our observation is the Wrst report that the Xexibility of subtilisin can have an eVect
on the discrimination between amino acid and ester substrates for subtilisin-cata-
lyzed reactions in organic solvents. This work is supported by a kinetic study as well
as a study of the Xexibility in the active site and the surface of the enzyme estimated
from the ESR spectroscopy. The conformational Xexibility of subtilisin appears to be
correlated with the enhancement of enantioselectivity, although the Xexibility eVect
depends markedly on the structure of the substrate used. In particular, for the ester
substrates, the large Xexibility around the surface of subtilisin with a partial confor-
mational change induced by the presence of a denaturing additive such as DMSO is
found to be signiWcantly responsible for the enantioselectivity enhancement based on
the increase in enzymatic activity for the correctly binding enantiomer. These Wnd-
ings should be of interest to organic chemists as a strategy to improve the enantiose-
lectivity and/or the enzymatic activity for ester substrates used as pharmaceutical and
agricultural chemicals.
Acknowledgments
The authors thank Prof. H. Ohta (Department of Physics, Faculty of Science,
Kobe University), Miss C. Urakawa, and Miss J. Yoshikawa for the helpful sugges-
tions regarding the acquisition of the ESR spectra.
References
[1] (a) G. Carrea, S. Riva, Angew. Chem. Int. Ed. 39 (2000) 2226–2254;
(b) A.M. Klibanov, Nature 409 (2001) 241–246;
(c) F. Bordusa, Chem. Rev. 102 (2002) 4817–4867.
[2] (a) A. Zaks, A.M. Klibanov, Proc. Natl. Acad. Sci. USA 82 (1985) 3192–3196;
(b) C.-S. Chen, C.J. Sih, Angew. Chem. Int. Ed. 28 (1989) 695–707;
(c) J.S. Dordick, Enzyme Microb. Technol. 11 (1989) 194–211.
[3] (a) Y.L. Khmelnitsky, S.H. Welch, D.S. Clark, J.S. Dordick, J. Am. Chem. Soc. 116 (1994) 2647–
2648;
(b) T. Okamoto, S. Ueji, Chem. Commun. (1999) 939–940;
(c) K. Watanabe, T. Yoshida, S. Ueji, Chem. Commun. (2001) 1260–1261;
(d) S. Ueji, M. Nishimura, R. Kudo, R. Matsumi, K. Watanabe, Y. Ebara, Chem. Lett. (2001) 912–
913;
(e) S. Ueji, H. Sakamoto, K. Watanabe, T. Okamoto, S. Mori, Bull. Chem. Soc. Jpn. 75 (2002) 2239–
2240.
[4] (a) P.A. Fitzpatrick, A.M. Klibanov, J. Am. Chem. Soc. 113 (1991) 3166–3171;
(b) S. Parida, J.S. Dordick, J. Org. Chem. 58 (1993) 3238–3244;
(c) G. Carrea, G. Ottolina, S. Riva, Trends Biotechnol. 13 (1995) 63–70;
(d) S. Ueji, T. Taniguchi, T. Okamoto, K. Watanabe, Y. Ebara, H. Ohta, Bull. Chem. Soc. Jpn. 76
(2003) 399–403.
[5] (a) R.S. Phillips, Trends Biotechnol. 14 (1996) 13–16;
(b) B. Galunsky, S. Ignatova, V. Kasche, Biochim. Biophys. Acta 1343 (1997) 130–138;
(c) Y. Yasufuku, S. Ueji, Bioorg. Chem. 25 (1997) 88–99.

K. Watanabe et al. / Bioorganic Chemistry 32 (2004) 504–515
515
[6] (a) C.R. Wescott, A.M. Klibanov, Biochim. Biophys. Acta 1206 (1994) 1–9;
(b) M.T. Reetz, A. Zonta, K. Schimossek, K. Liebeton, K.-E. Jaeger, Angew. Chem. Int. Ed. 36 (1997)
2830–2832;
(c) M.T. Reetz, K.-E. Jaeger, Chem. Eur. J. 6 (2000) 407–412.
[7] (a) Y. Okahata, Y. Fujimoto, K. Ijiro, J. Org. Chem. 60 (1995) 2244–2250;
(b) A. Fishman, S. Basheer, S. Shatzmiller, U. Cogan, Biotechnol. Lett. 20 (1998) 535–538;
(c) S. Ueji, H. Tanaka, T. Hanaoka, A. Ueda, K. Watanabe, K. Kaihatsu, Y. Ebara, Chem. Lett. (2001)
1066–1067;
(d) S. Ueji, A. Ueda, H. Tanaka, K. Watanabe, T. Okamoto, Y. Ebara, Biotechnol. Lett. 25 (2003) 83–
87.
[8] F. Theil, Tetrahedron 56 (2000) 2905–2919.
[9] (a) J. Broos, A.J.W.G. Visser, J.F.J. Engbersen, W. Verboom, A. Hoek, D.N. Reinhoudt, J. Am. Chem.
Soc. 117 (1995) 12657–12663;
(b) J. Broos, Biocat. Biotrans. 20 (2002) 291–295.
[10] M.K. Dhaon, R.K. Olsen, K. Ramasamy, J. Org. Chem. 47 (1982) 1962–1965.
[11] D.T. Witiak, T.C.-L. Ho, R.E. Hackney, J. Med. Chem. 11 (1968) 1086–1089.
[12] (a) J.D. Morrisett, C.A. BroomWeld, J. Biol. Chem. 247 (1972) 7224–7231;
(b) R. AZeck, Z.-F. Xu, V. Suzawa, K. Focht, D.S. Clark, J.S. Dordick, Proc. Natl. Acad. Sci. USA 89
(1992) 1100–1104.
[13] (a) A. Zaks, A.M. Klibanov, J. Biol. Chem. 263 (1988) 8017–8021;
(b) H. Kitaguchi, I. Itoh, M. Ono, Chem. Lett. (1990) 1203–1206;
(c) R. Bovara, G. Carrea, G. Ottolina, S. Riva, Biotechnol. Lett. 15 (1993) 937–942;
(d) K. Kawashiro, H. Sugahara, S. Sugiyama, H. Hayashi, Biotechnol. Lett. 17 (1995) 1161–1166.
[14] (a) J. Boutelje, M. Hjalmarsson, K. Hult, M. Lindback, T. Norin, Bioorg. Chem. 16 (1988) 364–375;
(b) K. Watanabe, S. Ueji, Biotechnol. Lett. 22 (2000) 599–603;
(c) K. Watanabe, S. Ueji, Biocat. Biotrans. 19 (2001) 281–299;
(d) K. Watanabe, S. Ueji, Soc. J. Chem. Perkin Trans. 1 (2001) 1386–1390.