Characterization of chemo- and regioselectivity in enzyme-catalyzed consecutive hydrolytic deprotection of methyl acetyl derivatives of 1-β-O-acyl glucuronides

Article abstract of DOI:10.1016/j.molcatb.2010.12.014

Methyl acetyl derivatives of 1-β-O-(o-, m-, or p-phenyl)benzoyl glucuronides 2a-c are fully deprotected by a one-pot consecutive enzyme-catalyzed hydrolytic reaction to afford 4a-c, without isolation of the O-deacetylated derivatives 3a-c. A lipase AS Amano from Aspergillus niger (LAS) and a carboxylesterase from Streptomyces rochei (CSR) showed high chemoselectivity toward the O-deacetylation of the o- and m-isomers, respectively. Chemoselective O-deacetylation of the p-isomer was promoted only in the presence of both enzymes. A lipase type B from Candida antarctica (CALB) was effective for the subsequent enzymatic hydrolysis of the methyl esters of 3a-c. LAS exhibited also regioselective 3-O-deacetylation activity to afford the corresponding 2,4-di-O-acetyl intermediates 5a-c, for which CSR showed higher O-deacetylation activity than that for 2a-c. In kinetic studies using p-nitrophenyl ester substrates, LAS exhibited a broader acyl preference, the octanoyl ester being most effectively hydrolysed, whereas CSR exhibited the highest hydrolytic activity toward the acetyl ester. LAS and CSR play complementary as well as synergistic roles in the O-deacetylation of 2 bearing R groups of different steric bulkiness.

Full text of DOI:10.1016/j.molcatb.2010.12.014

Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
Contents lists available at ScienceDirect
Journal of Molecular Catalysis B: Enzymatic
journal homepage: www.elsevier.com/locate/molcatb
Characterization of chemo- and regioselectivity in enzyme-catalyzed consecutive
hydrolytic deprotection of methyl acetyl derivatives of 1-␤-O-acyl glucuronides
Akiko Baba, Tadao Yoshioka∗
Hokkaido Pharmaceutical University School of Pharmacy, 7-1 Katsuraoka-cho, Otaru, Hokkaido 047-0264, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Methyl acetyl derivatives of 1-␤-O-(o-, m-, or p-phenyl)benzoyl glucuronides 2a–c are fully deprotected
by a one-pot consecutive enzyme-catalyzed hydrolytic reaction to afford 4a–c, without isolation of the
O-deacetylated derivatives 3a–c. A lipase AS Amano from Aspergillus niger (LAS) and a carboxylesterase
from Streptomyces rochei (CSR) showed high chemoselectivity toward the O-deacetylation of the o- and
m-isomers, respectively. Chemoselective O-deacetylation of the p-isomer was promoted only in the pres-
ence of both enzymes. A lipase type B from Candida antarctica (CALB) was effective for the subsequent
enzymatic hydrolysis of the methyl esters of 3a–c. LAS exhibited also regioselective 3-O-deacetylation
activity to afford the corresponding 2,4-di-O-acetyl intermediates 5a–c, for which CSR showed higher
O-deacetylation activity than that for 2a–c. In kinetic studies using p-nitrophenyl ester substrates, LAS
exhibited a broader acyl preference, the octanoyl ester being most effectively hydrolysed, whereas CSR
exhibited the highest hydrolytic activity toward the acetyl ester. LAS and CSR play complementary as
well as synergistic roles in the O-deacetylation of 2 bearing R groups of different steric bulkiness.
© 2011 Elsevier B.V. All rights reserved.
Received 12 October 2010
Received in revised form
28 December 2010
Accepted 28 December 2010
Available online 7 January 2011
Keywords:
1-␤-O-Acyl glucuronides
Chemoenzymatic synthesis
Enzyme-catalyzed hydrolytic deprotection
Chemoselectivity
Regioselective 3-O-deacetylation
1. Introduction
relationships for the electrophilicity of their 1-␤-O-acyl linkages
to predict the electrophilic reactivity [13–15].
It is important to understand pathways for the metabolic
activation of drugs to reactive metabolites and subsequent irre-
versible covalent binding to target tissue macromolecules, as
these processes are implicated in adverse drug reactions. There-
fore, in discovery and development of safer drugs, efficient
and widely applicable synthetic methodologies are required for
preparing these reactive metabolites as standard samples and
materials for toxicological and pharmacokinetic studies. 1-␤-O-
Acyl glucuronides (␤GAs), a group of the above-mentioned reactive
metabolites, are in general electrophilic species capable of bind-
ing to tissue proteins [1–4], possibly leading to adverse reaction of
the parent carboxylic acid drugs [5–7]. For example, some non-
steroidal anti-inflammatory drugs have been withdrawn due to
their toxicities [8–10]. However, specific target macromolecules to
which ␤GAs could covalently bind, toxicological consequences of
the covalent binding, and relationships between the electrophilic-
ity of ␤GAs and the toxicity of the parent carboxylic acid drugs
remain largely unknown.
Among the enzymes screened under the conditions shown in
Scheme 1, a lipase from A. niger (LAS) and a carboxylesterase from
S. rochei (CSR) showed hydrolytic activity for the deprotection of
the O-acetyl groups of the methyl acetyl precursor 2 to the corre-
sponding methyl ester derivative 3, whereas a lipase type B from C.
antarctica (CALB) and an esterase from porcine liver (PLE) showed
hydrolytic activity toward the methyl ester group of 3 to afford the
desired ␤GA 4. So far the chemoenzymatic method has been suc-
cessfully applied to synthesis of 15 kinds of ␤GAs derived from
six aromatic and nine aralkyl carboxylic acids [11–15]. LAS and
CSR showed a good chemoselectivity toward the O-acetyl group
among the three types of the ester functions of the compound
2: the O-acetyl group, the 1-␤-O-acyl linkage, and the carboxyl
methyl ester. However, the structure of R group of 1-␤-O-acyl
linkage in a given compound 2 has been proved to considerably
affect the chemoselectivity as well as the catalytic activity of these
enzymes. For example, toward compounds 2 with phenyl and
benzyl groups as the R group, LAS did not show good chemose-
lectivity toward the O-acetyl groups and unfavorably cleaved the
1-␤-O-acyl linkage to liberate the parent benzoic and phenylacetic
acids, respectively, whereas CSR showed much higher chemos-
electivity toward these substrates to afford the corresponding
fully O-deacetylated compounds 3 in good yields [12]. Further-
more, LAS exhibited high chemoselective O-deacetylation activity
toward 2 with a bulky o-(anilino)phenyl group as the R group,
We have recently established
a methodology for chemo-
enzymatic synthesis of ␤GAs with exclusive ␤-selectivity
(Scheme 1) [11–13], and derived quantitative structure–activity
∗
Corresponding author. Tel.: +81 134 62 1894; fax: +81 134 62 5161.
E-mail address: yoshioka@hokuyakudai.ac.jp (T. Yoshioka).
1381-1177/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcatb.2010.12.014

A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
75
Scheme 1. Chemoenzymatic synthetic route to ␤GAs 4. (i) RCO2Cs, DMSO; (ii) LAS and/or CSR, 20% (v/v) DMSO in citrate buffer (pH 5.0); (iii) CALB, 20% (v/v) DMSO in citrate
buffer (pH 5.0).
whereas CSR did not exhibit good hydrolytic activity toward this
substrate [12]. In addition, only concurrent use of LAS and CSR syn-
ergistically showed good chemoselective O-deacetylation activities
toward 2 with p-biphenyl and ␣,␣-diethyl- and ␣,␣-dimethyl-(p-
phenyl)benzyl groups as the R group [12,13]. For these reasons,
LAS and CSR are likely to play complementary and also synergistic
roles in the hydrolytic deprotection of the O-acetyl group of 2 but
the relationship between their hydrolytic activity and the struc-
ture of the R group has yet to be determined. Furthermore, LAS
is different from CSR in that LAS is prone to afford some partially
O-deacetylated intermediates in the reaction course.
There are several reviews for applications of hydrolytic enzymes
to the regioselective O-deacylation of peracylated carbohydrates
[16–19]. Among the enzymes, A. niger lipase has been reported
to catalyze regioselective hydrolyses of anomeric O-acetyl esters
of per-O-acetylated ␤-d-glucopyranose [20,21], ␤-d-ribofuranose
[22] and oligosaccharides [23,24]. LAS-catalyzed chemoselective
O-deacetylation of 2 to afford 3, leaving the 1-␤-O-acyl linkage
intact, might strongly depend on the bulkiness of the R group of
2. On the other hand, regioselective O-deacetylation activity has
not been reported for CSR-catalyzed ester hydrolysis to the best of
our knowledge.
It is, therefore, important and of interest to study the acyl pref-
erence and the influence of the steric effect of the R groups on LAS-
and CSR-catalyzed chemo- and regioselective O-deacetylation of
2, and thereby to gain insight into the enzymatic characteristics
in the chemo- and regioselectivity. It has to be confirmed also the
acyl preference with respect to the steric effect of R groups not
yet available in the literature. Furthermore, a consecutive one-pot
deprotection procedure from 2 to 4, without isolation of 3, seems to
be an intriguing synthetic approach. Among the abovementioned
15 kinds of ␤GAs, the methyl acetyl derivatives of regioisomeric
␤GAs 2a–c derived from o-, m- and p-phenylbenzoic acids 1a–c
(Scheme 2) have been selected as model substrates, because the
compounds 2a–c were O-deacetylated under different conditions
using LAS, CSR, and concurrent use of LAS and CSR, respectively
[12,13]. In this study, we report (1) the chemo- and regioselectiv-
ity of LAS and CSR toward 2a–c, to characterize the effects of the
R group, (2) acyl preferences in LAS- and CSR-catalyzed hydrolysis
using several p-nitrophenyl esters 7a–e (Scheme 2) by analysing
the reaction kinetics, and (3) a one-pot deprotection procedure
from 2a–c to the corresponding 1-␤-O-acyl glucuronides 4a–c.
2. Experimental
2.1. General methods and materials
Lipases AS Amano from A. niger (LAS) and from C. antarctica
type B (CALB), a carboxylesterase from S. rochei (CSR), and o-,
m-, and p-phenylbenzoic acids were obtained from Wako Pure
Chemical Industries (Osaka, Japan). Novozym 435 (an immobi-
lized form of CALB) was purchased from MIK Pharmaceuticals Co.,
Ltd. (Tokyo, Japan) and Lipozyme® CALBL was kindly gifted from
Novozymes Japan Ltd. (Chiba, Japan). The methyl acetyl deriva-
tives 2a–c were synthesized according to our previous papers
[11–13]. p-Nitrophenyl esters 7a, 7c, and 7d were obtained from
Nacalai tesque Inc. (Kyoto, Japan); the propionyl [25] and ben-
zoyl [26] esters 7b and 7e were synthesized according to reported
procedures. Amberlite XAD-4 was obtained from the Organo Cor-
poration (Tokyo, Japan) and used after grinding (80–200 mesh). All
other chemicals were commercial products. 1H and 13C NMR were
recorded using a JNM-AL400 (JEOL) and the chemical shifts are
presented as ı-values in ppm with reference to the residual sol-
vent signals [27] of d6-DMSO (2.49 ppm for 1H NMR and 39.50 ppm
for 13C NMR) or CD3OD (3.30 ppm for 1H NMR and 49.00 ppm for
13C NMR). MS and HRMS were measured by electron impact (EI)
ionization using a Hitachi M-2000 spectrometer.
2.1.1. General incubation conditions for enzymatic hydrolyses
The reaction mixtures were incubated at pH 5.0 to minimize a
non-enzymatic intramolecular acyl migration of partially and fully
O-deacetylated intermediates formed from 2a–c. The progress of
enzyme-catalyzed hydrolyses were monitored using HPLC. Unless
otherwise indicated, the incubation was performed in 20 mM
sodium citrate buffer (pH 5.0) containing 20% (v/v) DMSO as a co-
solvent and at 40 0.1 ◦C for LAS and CALB (including Lipozyme®
CALBL and Novozym 435) and at 50 0.1 ◦C for CSR, respectively.
Among several cosolvents tested, DMSO was the best cosolvent for
LAS and CSR using 2a and 2b as substrates, respectively. The con-
centration of LAS and CSR was 10 mg/mL of the incubation mixture.
The incubation with both enzymes was performed at 40 0.1 ◦C;
the optimal temperature of LAS toward 2a was 40 ◦C and the activity
was gradually deactivated at 50 ◦C, whereas the catalytic efficiency
of CSR was optimal at around 50 ◦C [12]. Incubation was started by
the addition of substrate in DMSO solution to the incubation mix-
ture. Hydrolytic activity of LAS and CSR toward 5a–c was presented
MeO2C
AcO
AcO
R
O
O
O
O
O2N
AcO
O
2a; o-isomer
2b; m-isomer
2c; p-isomer
7a; R = CH3
7b; R = CH2CH3 7e; R = C6H5
7c; R = (CH2)2CH3
7d; R = (CH2)6CH3
Scheme 2. Substrates used for LAS- and CSR-catalyzed hydrolysis.

76
A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
as an initial rate of decrease in the substrate concentration. For
the kinetic experiments on the hydrolysis of 7a–e, kinetic param-
eters (Km and Vmax) were obtained from double-reciprocal plots
of the initial rate data for LAS with 7a and 7b and for CSR with
7a–c. For other p-nitrophenyl ester substrates, pseudo first-order
rate constants were determined at low substrate concentrations.
The rate of enzymatic hydrolysis of 7a–e was corrected to account
for non-enzymatic hydrolysis. Kinetic measurement of the hydrol-
ysis of 3c, with initial concentration of 0.1 mM, was performed
using Lipozyme® CALBL, Novozyme 435, or CALB, at 100 ␮l of the
enzyme solution, 25 mg of the immobilized enzyme, or 0.8 mg
of the enzyme per mL of incubation mixture, respectively. The
hydrolytic reaction obeys a pseudo-first order reaction and the rate
constants obtained for Lipozyme® CALBL, Novozym 435, and CALB
were 2.38 0.02, 1.18 0.10, and 3.81 0.01 h−1, respectively. For
preparative purposes, the initial concentrations of 2a–c were in the
range of 0.4–0.6 mM.
2.2.2. Methyl 1-O-(m-phenyl)benzoyl-ˇ-d-glucopyranuronate
(3b)
A solution of 2b (15.7 mg, 30.5 ␮mol) in 51 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 50 ◦C using CSR (510 mg). After 3 h, the conversion
yield was 98%, as determined by HPLC analysis. The product was
quantitatively extracted with EtOAc (50 mL × 3), and then EtOAc
was evaporated in vacuo. The residue was dissolved in 4 mL of 20%
(v/v) CH3CN containing 0.01% (v/v) AcOH and applied to an Amber-
lite XAD-4 column (2.5 g, height 6 cm). Product 3b was eluted with
30% (v/v) CH3CN containing 0.01% AcOH. Yield (11.5 mg, 97%). 1H
NMR (400 MHz, CD3OD) ı: 3.51–3.65 (m, 3H, C2-H, C3-H, and C4-H),
3.76 (s, 3H, CO2CH3), 4.06 (d, J 9.5 Hz, 1H, C5-H), 5.77 (d, J 7.8 Hz, 1H,
C1-H), 7.37 (t, J 7.6 Hz, 1H, aromatic H), 7.46 (t, J 7.1 Hz, 2H, aromatic
H), 7.58 (t, J 7.8 Hz, 1H, aromatic H), 7.65–7.67 (m, 2H, aromatic H),
7.89–7.91 (m, 1H, aromatic H), 8.05–8.08 (m, 1H, aromatic H), 8.32
(t, J 1.5 Hz, 1H, aromatic H). 13C NMR (100 MHz, CD3OD) ı: 52.9, 73.0
(C4), 73.7 (C2), 77.3 (C5), 77.4 (C3), 96.3 (C1), 128.1, 129.0, 129.3,
129.8, 130.1, 130.3, 131.1, 133.3, 141.2, 143.1, 166.5, 170.8. HR MS
(EI): calcd. for C20H20O8 [M+]: 388.1156; found 388.1162. MS (EI):
m/z 388 (0.2%), 198 (25), 181 (8), 152 (9), 105 (100).
2.1.2. HPLC analysis of enzymatic hydrolyses
To analyse the time courses of the enzymatic hydrolyses of
2a–c, 3a–c, 5a–c, and 7a–e, at appropriate intervals, aliquots of
the reaction mixture were diluted with HPLC mobile phase, to
stop the enzymatic reaction. The HPLC samples were injected
onto a Shimadzu HPLC that was equipped with a reversed-phase
Symmetry C18 column (5 ␮m, 4.6 mm × 150 mm, Waters), a C-R8A
Chromatopac data processor, a Shimadzu SPD-10A VP UV detector
anda ShimadzuCTO-10ASVPcolumn ovenat 30 ◦C. Thecolumn was
eluted with HPLC mobile phase at a flow rate of 0.7 mL/min with
detection at 260 nm for 2a–c, 3a–c, and 5a–c and 310 nm for 7a–e,
respectively. The mobile phase was aqueous CH3CN containing
10 mM tetra-n-butylammonium bromide and 50 mM ammonium
acetate buffer (pH 4.5) with variable concentrations (%, v/v) of
CH3CN; 35% for 3a–c, 45% for 5a–c and 7a, 50% for 2a, 7b,c, and
7e, 55% for 2b,c and 70% for 7d.
2.2.3. Methyl 1-O-(p-phenyl)benzoyl-ˇ-d-glucopyranuronate
(3c)
A solution of 2c (16.4 mg, 31.9 ␮mol) in 80 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS (800 mg) and CSR (800 mg). After
2 h, the conversion yield was 95%, as determined by HPLC analysis.
The product was quantitatively extracted with EtOAc (50 mL × 3),
and then EtOAc was evaporated in vacuo. The residue was dis-
solved in 4 mL of 20% (v/v) CH3CN containing 0.01% (v/v) AcOH and
applied to an Amberlite XAD-4 column (2.5 g, height 6 cm). Product
3c was eluted with 30% (v/v) CH3CN containing 0.01% AcOH. Yield
(11.7 mg, 95%). 1H NMR (400 MHz, CD3OD) ı: 3.51–3.65 (m, 3H,
C2-H, C3-H, and C4-H), 3.76 (s, 3H, CO2CH3), 4.06 (d, J 9.5 Hz, 1H,
C5-H), 5.77 (d, J 7.8 Hz, 1H, C1-H), 7.38–7.42 (m, 1H, aromatic H),
7.45–7.49 (m, 2H, aromatic H), 7.67–7.70 (m, 2H, aromatic H), 7.76
(d, J 8.5 Hz, 2H, aromatic H), 8.17 (d, J 8.5 Hz, 2H, aromatic H). 13C
NMR (100 MHz, CD3OD) ı: 52.9, 73.0 (C4), 73.7 (C2), 77.3 (C5), 77.4
(C3), 96.2 (C1), 128.2, 128.3, 129.2, 129.4, 130.1, 131.6, 166.4, 170.8.
HR MS (EI): calcd. for C20H20O8 [M+]: 388.1156; found 388.1128.
MS (EI): m/z 388 (2%), 198 (100), 181 (82), 152 (40).
2.1.3. Computational chemistry
To obtain the optimized structures of 2a–c, calculations were
performed with a molecular mechanics model with MMFF, a
semiempirical AM1 and PM3 models, and Hartree-Fock 3-21G and
6-31+G* models using Spartan ‘08 (Wave Function, Inc.).
2.2. Compound characterization
2.2.4. Methyl
1-O-(o-phenyl)benzoyl-2,4-di-O-acetyl-ˇ-d-glucopyranuronate
(5a)
2.2.1. Methyl 1-O-(o-phenyl)benzoyl-ˇ-d-glucopyranuronate
(3a)
A solution of 2a (17.5 mg, 34.0 ␮mol) in 85 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS (170 mg, 2.0 mg/mL of incubation
mixture). After 0.5 h, the product 5a was quantitatively extracted
with EtOAc (50 mL × 2), and then EtOAc was evaporated in vacuo.
The residue was dissolved in 4 mL of 40% (v/v) CH3CN containing
0.01% (v/v) AcOH and applied to an Amberlite XAD-4 column (2.5 g,
height 6 cm). Product 5a was eluted with 50% (v/v) CH3CN contain-
ing 0.01% AcOH. Yield (13.0 mg, 81%). 1H NMR (400 MHz, d6-DMSO)
ı: 2.04 (s, 3H, OCOCH3), 2.05 (s, 3H, OCOCH3), 3.62 (s, 3H, CO2CH3),
3.96 (dt, J 6.1 and 9.5 Hz, 1H, C3-H), 4.45 (d, J 9.5 Hz, 1H, C5-H),
4.79 (t, J 9.5 Hz, 1H, C4-H), 4.86 (dd, J 8.1 and 9.5 Hz, 1H, C2-H),
5.86 (d, J 6.1 Hz, 1H, OH), 5.93 (d, J 8.1 Hz, 1H, C1-H), 7.26 (dd, J 2.0
and 6.8 Hz, 2H, aromatic H), 7.36–7.41 (m, 3H, aromatic H), 7.44
(dd, J 1.2 and 7.8 Hz, 1H, aromatic H), 7.54 (dt, J 1.2 and 7.8 Hz, 1H,
aromatic H), 7.68 (dt, J 1.2 and 7.8 Hz, 1H, aromatic H), 7.72 (dd, J
1.2 and 7.8 Hz, 1H, aromatic H). 13C NMR (100 MHz, d6-DMSO) ı:
20.5, 20.6, 52.4, 70.0 (C3), 71.3 (C4), 71.7 (C5), 72.1 (C2), 91.5 (C1),
127.3, 127.6, 128.1, 128.3, 128.5, 129.3, 131.1, 132.4, 140.0, 142.4,
165.0, 167.4, 169.2, 169.5. HR MS (EI): calcd. for C24H24O10 [M+]:
A solution of 2a (14.1 mg, 27.4 ␮mol) in 68 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS (680 mg). After 5 h, the conversion
yield was 98%, as determined by HPLC analysis. The product was
quantitatively extracted with EtOAc (50 mL × 3), and then EtOAc
was evaporated in vacuo. The residue was dissolved in 4 mL of 20%
(v/v) CH3CN containing 0.01% (v/v) AcOH and applied to an Amber-
lite XAD-4 column (2.5 g, height 6 cm). Product 3a was eluted with
30% (v/v) CH3CN containing 0.01% AcOH. Yield (10.3 mg, 97%). 1H
NMR (400 MHz, CD3OD) ı: 3.30 (dd, J 8.1 and 9.5 Hz, 1H, C2-H), 3.44
(t, J 9.5 Hz, 1H, C3-H or C4-H), 3.53 (t, J 9.5 Hz, 1H, C3-H or C4-H), 3.78
(s, 3H, CO2CH3), 3.93 (d, J 9.5 Hz, 1H, C5-H), 5.54 (d, J 8.1 Hz, 1H, C1-
H), 7.30–7.41 (m, 6H, aromatic H), 7.45 (dt, J 1.2 and 7.8 Hz, 1H,
aromatic H), 7.60 (dt, J 1.2 and 7.8 Hz, 1H, aromatic H), 7.93 (dd,
J 1.2 and 7.8 Hz, 1H, aromatic H). 13C NMR (100 MHz, CD3OD) ı:
52.9, 72.8 (C4), 73.6 (C2), 77.2 (C5), 77.3 (C3), 96.1 (C1), 128.2, 128.3,
129.2, 129.8, 129.9, 131.1, 132.0, 133.0, 142.1, 144.5, 168.0, 170.7.
HR MS (EI): calcd. for C20H20O8 [M+]: 388.1156; found 388.1140.
MS (EI): m/z 388 (1%), 198 (100), 181 (98), 152 (46).

A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
77
472.1367; found 472.1348. MS (EI): m/z 472 (1%), 440 (3), 198 (28),
181 (100), 152 (18).
(C1), 128.2, 128.3, 129.1, 129.5, 130.1, 131.7, 141.0, 147.9, 166.3,
169.8, 172.0. HR MS (EI): calcd. for C22H22O9 [M+]: 430.1263; found
430.1273. MS (EI): m/z 430 (7%), 198 (98), 181 (100), 152 (54).
2.2.5. Methyl
1-O-(m-phenyl)benzoyl-2,4-di-O-acetyl-ˇ-d-glucopyranuronate
(5b)
2.2.8. One-pot synthesis of
1-O-(o-phenyl)benzoyl-ˇ-d-glucopyranuronate (4a)
A solution of 2b (17.9 mg, 34.8 ␮mol) in 87 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS (174 mg, 2.0 mg/mL of incubation
mixture). After 17 min, the product 5b was quantitatively extracted
with EtOAc (50 mL × 2), and then EtOAc was evaporated in vacuo.
The residue was dissolved in 4 mL of 40% (v/v) CH3CN contain-
ing 0.01% (v/v) AcOH and applied to an Amberlite XAD-4 column
(2.5 g, height 6 cm). Product 5b was eluted with 50% (v/v) CH3CN
containing 0.01% AcOH. Yield (13.6 mg, 83%). 1H NMR (400 MHz,
d6-DMSO) ı: 2.03 (s, 3H, OCOCH3), 2.07 (s, 3H, OCOCH3), 3.61 (s,
3H, CO2CH3), 4.06 (dt, J 5.9 and 9.5 Hz, 1H, C3-H), 4.57 (d, J 9.5 Hz,
1H, C5-H), 4.86 (t, J 9.5 Hz, 1H, C4-H), 5.02 (dd, J 8.1 and 9.5 Hz, 1H,
C2-H), 5.91 (d, J 5.9 Hz, 1H, OH), 6.08 (d, J 8.1 Hz, 1H, C1-H), 7.44 (t, J
7.1 Hz, 1H, aromatic H), 7.53 (t, J 7.1 Hz, 2H, aromatic H), 7.66–7.73
(m, 3H, aromatic H), 7.93 (d, J 7.8 Hz, 1H, aromatic H), 8.03 (d, J
7.8 Hz, 1H, aromatic H), 8.18 (t, J 1.5 Hz, 1H, aromatic H). 13C NMR
(100 MHz, d6-DMSO) ı: 20.5, 20.6, 52.4, 69.7 (C3), 71.4 (C4), 71.8
(C5), 72.2 (C2), 91.9 (C1), 126.8, 127.5, 128.1, 128.4, 128.8, 129.1,
129.8, 132.5, 138.8, 140.9, 163.8, 167.4, 169.5, 169.6. HR MS (EI):
calcd. for C24H24O10 [M+]: 472.1367; found 472.1364. MS (EI): m/z
472 (7%), 440 (8), 198 (60), 181 (100), 152 (29).
A solution of 2a (16.3 mg, 31.6 ␮mol) in 60 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS (600 mg). After 5 h, CALB (60 mg
in 1 mL of water) was added at one portion and incubated for fur-
ther 1.0 h. The reaction mixture was then acidified with 2 M HCl
to adjust pH to about 2.0. After filtration, the filtrate was applied
to an Amberlite XAD-4 column (2.5 g, height 6 cm). After washing
with 50 mL of 0.01% aqueous AcOH and then 10%(v/v) CH3CN con-
taining 0.01% AcOH, product 4a was eluted with 20% (v/v) CH3CN
containing 0.01% AcOH. Yield (11.3 mg, overall yield of 95%). The
HPLC retention time and the NMR data were identical with those
of previously synthesised 4a with a two-step procedure [13].
2.2.9. One-pot synthesis of
1-O-(m-phenyl)benzoyl-ˇ-d-glucopyranuronate (4b)
A solution of 2a (15.7 mg, 30.5 ␮mol) in 60 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 50 ◦C using CSR (600 mg). After 3.5 h, CALB (60 mg
in 1 mL of water) was added at one portion and incubated for fur-
ther 1.0 h. The reaction mixture was then acidified with 2 M HCl
to adjust pH to about 2.0. After filtration, the filtrate was applied
to an Amberlite XAD-4 column (2.5 g, height 6 cm). After washing
with 50 mL of 0.01% aqueous AcOH and then 10% (v/v) CH3CN con-
taining 0.01% AcOH, product 4b was eluted with 25% (v/v) CH3CN
containing 0.01% AcOH. Yield (10.6 mg, overall yield of 93%). The
HPLC retention time and the NMR data were identical with those
of previously synthesised 4b with a two-step procedure [13].
2.2.6. Methyl
1-O-(p-phenyl)benzoyl-2,4-di-O-acetyl-ˇ-d-glucopyranuronate
(5c)
A solution of 2c (19.8 mg, 38.5 ␮mol) in 96 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using LAS and CSR (each 960 mg). After
10 min, the product 5c was quantitatively extracted with EtOAc
(50 mL × 2), and then EtOAc was evaporated in vacuo. The residue
was dispersed in 20 mL of 20% (v/v) CH3CN containing 0.01% (v/v)
AcOH and applied to an Amberlite XAD-4 column (2.5 g, height
6 cm). Elution was performed stepwise from 20, 30, 40 to 50% (v/v)
CH3CN containing 0.01% (v/v) AcOH. Product 5c was eluted with
50% (v/v) CH3CN containing 0.01% AcOH. Yield (9.5 mg, 52%). 1H
NMR (400 MHz, d6-DMSO) ı: 2.03 (s, 3H, OCOCH3), 2.07 (s, 3H,
OCOCH3), 3.51 (s, 3H, CO2CH3), 4.06 (t, J 9.3 Hz, 1H, C3-H), 4.56
(d, J 9.3 Hz, 1H, C5-H), 4.85 (t, J 9.3 Hz, 1H, C4-H), 5.01 (t, J 8.5 Hz,
1H, C2-H), 5.92 (brs, 1H, OH), 6.08 (d, J 8.1 Hz, 1H, C1-H), 7.45 (t, J
6.8 Hz, 1H, aromatic H), 7.52 (t, J 7.8 Hz, 2H, aromatic H), 7.77 (d, J
8.1 Hz, 2H, aromatic H), 7.89 (d, J 8.3 Hz, 2H, aromatic H), 8.01 (d, J
8.1 Hz, 2H, aromatic H). 13C NMR (100 MHz, d6-DMSO) ı: 20.5, 20.6,
52.4, 69.8 (C3), 71.4 (C4), 71.8 (C5), 72.2 (C2), 91.7 (C1), 126.8, 127.0,
127.2, 128.6, 129.1, 130.1, 138.5, 145.6, 163.8, 167.4, 169.4, 169.6.
HR MS (EI): calcd. for C24H24O10 [M+]: 472.1367; found 472.1355.
MS (EI): m/z 472 (4%), 440 (12), 198 (35), 181 (100), 152 (21)
2.2.10. One-pot synthesis of
1-O-(p-phenyl)benzoyl-ˇ-d-glucopyranuronate (4c)
A solution of 2c (26.0 mg, 50.5 ␮mol) in 100 mL of sodium citrate
buffer (20 mM, pH 5.0) containing 20% (v/v) DMSO was incubated
under stirring at 40 ◦C using CSR (1.00 g) and LAS (1.00 g). After
2.5 h, the reaction mixture was heated in a boiling water to keep
temperature of the reaction mixture at least more than 70 ◦C for 30 s
to deactivate LAS. CALB (100 mg in 2 mL of water) was then added at
one portion and incubated for further 1.5 h. The reaction mixture
was then acidified with 2 M HCl to adjust pH to about 2.0. After
filtration, the filtrate was applied to an Amberlite XAD-4 column
(2.5 g, height 6 cm). After washing with 50 mL of 0.01% aqueous
AcOH and then 10%(v/v) CH3CN containing 0.01% AcOH, product
4c was eluted with 25% (v/v) CH3CN containing 0.01% AcOH. Yield
(17.4 mg, overall yield of 92%). The HPLC retention time and the
NMR data were identical with those of previously synthesised 4c
with a two-step procedure [12].
2.2.7. Methyl 1-O-(p-phenyl)benzoyl-4-mono-O-acetyl-ˇ-d-
glucopyranuronate
3. Results and discussion
(6c)
3.1. Chemoselectivity of LAS and CSR in hydrolysis of 2a–c
In the fraction eluted with 40% (v/v) CH3CN containing 0.01%
(v/v) AcOH from the XAD-4 column applied the above-mentioned
reaction mixture, another product 6c was found. Yield (2.5 mg,
15%). 1H NMR (400 MHz, CD3OD) ı: 2.09 (s, 3H, OCOCH3), 3.63–3.69
(m, 1H, C2-H), 3.68 (s, 3H, CO2CH3), 3.79 (t, J 9.3 Hz, 1H, C3-H),
4.27 (d, J 10.0 Hz, 1H, C5-H), 4.92 (t, J 9.5 Hz, 1H, C4-H), 5.81 (d,
J 8.1 Hz, 1H, C1-H), 7.38–7.41 (m, 1H, aromatic H), 7.47 (t, J 7.5 Hz,
2H, aromatic H), 7.66–7.69 (m, 2H, aromatic H), 7.77 (d, J 8.5 Hz, 2H,
aromatic H), 8.17 (d, J 8.5 Hz, 2H, aromatic H). 13C NMR (100 MHz,
CD3OD) ı: 20.7, 53.2, 73.1 (C4), 73.7 (C2), 74.1 (C5), 74.9 (C3), 95.7
The compounds 2a–c, from which the corresponding 1-␤-O-acyl
glucuronides 4a–c had been synthesized previously with exclusive
␤-selectivity [12,13], were selected as the model compounds to
examine in detail the chemo- and regioselectivity of LAS and CSR
in the enzymatic O-deacetylation.
As shown in Table 1, both the enzymes exhibited different
hydrolytic activities toward substrates 2a–c; the chemoselective
O-deacetylation of 2a–c depended significantly on the position of
the phenyl substituent. Fig. 1 shows the typical time courses of the

78
A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
Fig. 1. Time courses of LAS- and CSR-catalyzed hydrolytic reactions of 2a–c. (A) LAS toward 2a; (B) CSR toward 2a; (C) LAS toward 2b; (D) CSR toward 2b; (E) LAS toward 2c;
(F) CSR toward 2c. Symbols are the concentrations of liberated parent carboxylic acids 1a–c (ꢀ), substrates 2a–c (᭹), fully O-deacetylated derivatives 3a–c (ꢀ), 2,4-di-O-acetyl
derivatives 5a–c (ꢁ), and mono-O-acetyl derivatives (ꢂ). Initial concentration of substrates was 0.4 mM. Enzyme concentration was 10 mg/mL of incubation mixture.
enzymatic O-deacetylation of 2a–c. LAS and CSR exhibited the high-
est chemoselective O-deacetylation activity toward 2a (Fig. 1A) and
2b (Fig. 1D) to yield 3a and 3b in 97% yield in 5 h and in 97% yield
in 3 h, respectively. The chemoselectivity worsened in the oppo-
site combination (Fig. 1B and C). With LAS, 2b liberated the parent
carboxylic acid 1b, other than the desired product 3b, as a major
product in 55% yield in 4 h without time-lag (Fig. 1C), indicating that
LAS hydrolysed not only the O-acetyl groups but also 1-␤-O-acyl
ester linkage of 2b. As shown in Fig. 1E with LAS, 2c also liber-
ated the parent carboxylic acid 1c with accumulation of partially
O-deacetylated intermediates (2,4-di-O-acetylated and 4-mono-
O-acetylated compounds 5c and 6c, respectively); the structural
assignment of 5c and 6c is described later in this paper. With CSR,
2a was hydrolysed to 3a (45% yield in 3 h) but with concomitant
formation of a measurable amount of 1a (Fig. 1B), whereas 2c was
hardly hydrolysed to afford only 3c in 4% yield in 4 h (Fig. 1F).
The chemoselective O-deacetylation of 2c was satisfactory, in a
synergistic manner, only by concurrent use of LAS and CSR; 3c was
obtained in 95% yield in 2 h (Table 1 and Fig. 2A). Fig. 2B clearly
shows the inhibition of CSR-catalyzed O-deacetylation of 2b by the
presence of 2c, albeit a poor substrate for CSR as shown in Fig. 1F.
This result indicates that 2c could occupy the binding site(s) of CSR
but the p-phenyl substituent might force 2c into an unfavorable
orientation for the enzymatic hydrolysis in the binding site.

A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
79
Fig. 2. Time courses of (A) chemoenzymatic O-deacetylation of 2c (C0 0.4 mM) with concurrent use of LAS and CSR and (B) CSR-catalyzed hydrolysis of 2b (C0 0.2 mM) in the
absence and presence of 2c (C0 0.2 mM). Symbols in (A) were the concentrations of 1c (ꢀ), 2c (᭹), 3c (ꢀ), and 5c (ꢁ). Symbols in (B) were the concentrations of 2b in the
absence (♦) and presence (ꢃ) of 2c and 2c in the absence (ꢀ) and presence (᭹) of 2b. Enzyme concentration was 10 mg/mL of incubation mixture.
Table 1
and their structures were characterized. For the isolation of max-
imum amounts of 5a and 5b by retarding the reaction velocities,
the concentrations of LAS and CSR, respectively, were reduced to
Yields of the major products obtained from LAS- and CSR-catalyzed hydrolytic reac-
tions of 2a–c.
2 mg/mL of incubation mixture. 1H and 13C NMR spectra of 5a–c
indicate four carbonyl carbons and two acetyl methyl carbons with
attached hydrogens. These results, together with MS data, indicate
that the compounds 5a–c are di-O-acetyl derivatives of 3a–c. In
the 1H NMR spectra of 5a–c, signals at ı ca. 5.9, exchangeable with
D2O, were assigned to the proton of the OH group at the 3-position,
as determined from the 1H-1H COSY spectra (data not shown), in
which the protons of C1-H through C5-H on the pyranose ring could
be easily assigned. Comparing the C1-H through C5-H of 5a–c with
those of 2a–c, the upfield shift observed for C3-H of 5a–c could
be due to 3-O-deacetylation. In addition, large downfield shifts
(1.3–1.5 ppm) were observed for both C2-H and C4-H of 5a–c, com-
pared with those of 4a–c; this downfield shifts [28,29] also support
the 2,4-di-O-acetyl structures. The structures are also supported by
the downfield shifts (2.5–2.7 ppm) of C2 and C4 observed for 5a–c, a
typical deacetylation shift [30] at 3-O-acetyl group of 2a–c. The iso-
lation yields of 5a, 5b and 5c were 81, 83 and 52%, respectively. In
the isolation of 5c, a more polar product 6c (15%) was also obtained.
6c was confirmed to be 4-mono-O-acetyl derivative of 4c by com-
paring its spectral data with those of 2c and 5c, and by analysing
the 1H-1H COSY spectrum of 6c (data not shown). A major path-
way for the enzymatic O-deacetylation of 2c to 4c, through major
intermediates 5c and 6c, is shown in Scheme 3.
Substrate
Enzyme
Time (h)
Products and yield (%)
LASa
5
3
3
0.5
4
3
2
0.3
4
4
3ac (97%)
CSRa
3ac (45%) and 1ad (15%)
3ac (81%) and 1ad (16%)
5ae (81%)
2a
LASa + CSRa
LASb
LASa
3bc (35%) and 1bd (55%)
3bc (97%)
CSRa
2b
LASa + CSRa
LASb
3bc (79%) and 1bd (19%)
5be (83%)
LASa
3cc (50%), 1cd (23%) and 6c f (23%)
3cc (4%)
CSRa
2c
LASa + CSRa
LASa + CSRa
2
0.2
3cc (95%)
5ce (52%) and 6cf (15%)
a
Enzymes was used at concentration of 10 mg/mL of incubation mixture.
LAS was used at concentration of 2.0 mg/mL of incubation mixture.
3a–c are the corresponding fully O-deacetylated compounds derived from 2a–c.
1a–c are the corresponding carboxylic acids, namely, o-, m-, and
p-phenylbenzoic acids, respectively.
b
c
d
e
f
5a–c are the 2,4-di-O-acetyl compounds derived from 2a–c.
6c is the 4-mono-O-acetyl compound derived from 2c.
3.2. Regioselectivity of LAS in O-deacetylation of 2a–c
In LAS-catalyzed O-deacetylation of 2a–c, as mentioned above
for 2c, partially O-deacetylated intermediates 5a–c and 6c were as
well detected. To investigate the regioselectivity of LAS-catalyzed
O-deacetylation of 2a–c, the major intermediates 5a–c that accu-
mulated in the early stage of the hydrolytic reactions were isolated
All the enzymatic hydrolyses in this study were performed
at pH 5.0 to minimize a non-enzymatic intramolecular acyl
migration [31,32] of 3a–c and 5a–c as well as other interme-
MeO2C
MeO2C
O
AcO
O
AcO
i
O
AcO
O
HO
AcO
AcO
O
O
2c
5c
i
MeO2C
MeO2C
AcO
HO
O
HO
i
O
HO
O
O
HO
O
HO
6c
O
3c
Scheme 3. Chemo- and regioselectivity in LAS-catalyzed O-deacetylation of 2c to 3c through intermediates 5c and 6c. (i) LAS, 20% (v/v) DMSO in citrate buffer (pH 5.0).

80
A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
Table 2
diates formed. The half-life times of 2,4-di-O-acetyl derivatives
5a–c under the conditions without enzyme at 40 ◦C were
over 13 h (data not shown), strongly indicating that these
compounds are directly formed through LAS-catalyzed regios-
elective 3-O-deacetylation of 2a–c. In literature, regioselective
2-O-deacetylation has been reported for methyl and benzyl 2,3,4-
tri-O-acetyl-␤-d-galactopyranosiduronates catalyzed by acylase I
[33]. In addition to enzymatic regioselective O-deacetylation of
primary ester of fully O-acetylated sugars [34–36] and glycosides
[37–41], regioselective O-deacetylation has been reported for 3-
O-deacetylation of secondary esters of methyl and 1-thiomethyl
2,3,4,6-tetra-O-acetyl-␤-d-glucopyranosides by a lipase from A.
niger [42,43] and a cellulose acetate esterase from Neisseria
sicca SB [44], 2-, 3- or 5-O-deacetylation of methyl 2,3,5-tri-O-
acetyl-d-arabinofuranoside by several enzymes [45], and 2- or
3-O-deacetylation of fully O-acetylated adenosine derivative by an
esterase from porcine liver [46]. In these reports, regioselective
hydrolysis of the secondary O-acetyl ester has been observed for
substrates without anomeric ester group. To our knowledge, our
finding would be the first report for LAS-catalyzed regioselective 3-
O-deacetylation toward the methyl acetyl derivatives of 1-␤-O-acyl
glucuronides 2a–c.
Comparison of hydrolytic activities of LAS and CSR toward 2a–c and 5a–c.
Substrate
Initial velocitya
With LASb
With CSRd
2a
2b
2c
5a
5b
5c
11.7 0.3
25.0 0.2
22.4 0.2
0.40 0.01
1.5 0.1
0.62 0.02
1.9 0.1
0.016 0.002
205
365
3
3
0.28 0.05
4.3 0.1
a
Values were represented as mM/h at the initial concentrations of substrates
(0.4 mM).
b
The concentration of the enzyme (LAS or CSR) was 10 mg/mL of incubation
mixture.
could be effectively hydrolysed to the product 3c by CSR, whose
initial rate toward 5c was approximately 15-fold higher than that
of LAS (Table 2).
3.4. One-pot procedure for enzymatic deprotection of 2a–c to
4a–c
To reduce the time and costs, a one-pot procedure for the
enzyme-catalyzed consecutive hydrolytic deprotection of 2a–c to
4a–c, without isolation of 3a–c, was next examined. Because the
hydrolysis of the methyl ester of 2 prior to the O-deaceylation did
not give a good result [11] and CALB showed higher chemoselec-
tive hydrolytic activity than PLE toward the methyl ester [12], we
tried a one-pot procedure using LAS and/or CSR as the enzyme(s)
for the first step of O-deacetylation followed by addition of CALB as
the enzyme for the second step of methyl ester hydrolysis. There-
fore, a prerequisite for this one-pot procedure is that the enzyme(s)
used for the first O-deacetylation step shows no hydrolytic activity
toward the final products 4a–c. As shown in Scheme 4, 2a and 2b
were successfully converted to 4a and 4b in good yields (95 and 93%,
respectively) using LAS and CSR as the first enzyme, respectively.
LAS and CSR did not show significant hydrolytic activity toward
4a and 4b, respectively. In the case of 2c, because LAS, but not
CSR, showed hydrolytic activity toward 4c with a half-life of ca.
1.2 h under the incubation conditions (data not shown), the reac-
tion mixture of 2c with both LAS and CSR was heated at least more
than 70 ◦C for 30 sec to deactivate LAS before the addition of CALB;
in consequence 4c was obtained in overall yield of 92%. Besides
CALB, the hydrolytic activity of commercially available lipases of
Lipozyme® CALBL and Novozym 435 was also examined using 3c
as a model substrate. The hydrolysis of the methyl ester of 3c
smoothly the proceeded according to first-order reaction kinetics
3.3. Hydrolytic activity of LAS and CSR toward 2a–c and 5a–c
As described above, no significant accumulation of partially
O-deacetylated intermediates was observed in CSR-catalyzed O-
deacetylation of 2a–c to 3a–c. This result indicates that CSR could
hydrolyse partially O-deacetylated intermediates such as 5a–c
much more effectively than the initial substrates 2a–c. To examine
the possibility, initial velocities of hydrolytic reactions catalyzed
by CSR as well as LAS were determined using 2a–c and 5a–c as
substrates. Table 2 summarizes the initial velocities of LAS- and
CSR-catalyzed decrease in the concentration of the substrates 2a–c
and 5a–c.
LAS exhibited higher hydrolytic activity toward 2a–c than
the 2,4-di-O-acetyl intermediates 5a–c. Conversely, CSR exhibited
much higher catalytic activity toward 5a–c than 2a–c, as expected.
This result explains the reason because accumulation of 5a–c was
hardly observed during CSR-catalyzed O-deacetylation of 2a–c. In
the case of 2c, the rate of LAS-catalyzed O-deacetylation of 2c was
shown to dramatically decrease due probably to increasing con-
centrations of the intermediates 5c and 6c (see Fig. 1E), indicating
a competitive inhibition of enzyme activity by poorer substrates 5c
and 6c. The synergistic acceleration of the O-deacetylation of 2c to
3c with concurrent use of LAS and CSR was, therefore, reasonably
explained by thinking that the intermediates 5c (and probably 6c)
MeO2C
MeO2C
O
AcO
O
HO
LAS/CSR
O
AcO
HO
O
AcO
OH
O
O
2a-c
3a-c
CALB
4a (with LAS; overall 95%)
4b (with CSR; overall 93%)
4c (with LAS and CSR; overall 92%)
HO2C
HO
HO
O
O
O
OH
4a-c
Scheme 4. One-pot synthetic procedure from 2a–c to 4a–c.

A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
81
Table 3
Kinetic parameters obtained for LAS- and CSR-catalyzed hydrolysis of 7a–e.
Ester
With LAS
With CSR
a
b
a
b
Vmax
Km
Vmax/Km
Vmax
Km
Vmax/Km
7a
7b
7c
7d
7e
1.2
2.0
4.2
0.29
0.17
0.76
0.19
0.021
0.12
0.62
3.45
6.3
0.31
11.5
c
c
–
–
–
–
–
0.27
0.0061
c
c
c
c
4.2d
–
–
0.0003d
c
c
c
c
–
0.0093d
–
–
<0.0001d
a
Expressed as ␮mol of liberated p-nitrophenol/min/mg of enzyme.
mM.
Values not determined.
Values from the apparent first-order rate constants.
b
c
d
Fig. 3. Optimized structures of 2a–c calculated by the Hartree-Fock 6-31 + G* models.
and the desired product 4c was obtained in good yields (more than
98%) with chemoselectivity comparable with that of CALB. On the
basis of the rate constants, the hydrolytic activities of 1 mL of the
enzyme solution of Lipozyme® CALBL and 1 g of the Novozym 435
were estimated to be equivalent to that of 6.2 and 12.3 mg of CALB,
respectively.
that LAS has a binding site large enough to be occupied by 2b and 2c
and a catalytic activity toward the 1-␤-O- acyl linkages. In compari-
son with LAS, CSR showed a narrower acyl preference (Table 3), the
acetyl ester 7a being most efficiently catalyzed. The Km value of CSR
for 7a was much lower than the value of LAS and the Vmax/Km value
of CSR for 7a was much higher than that of LAS. Furthermore, the
Vmax/Km values of CSR for other substrates decreased with increas-
ing numbers of methylene chains of 7b–d, probably due to the
increasing steric bulkiness of the acyl groups. The finding that CSR
hardly hydrolysed benzoyl ester 7e suggests that the binding site
of CSR should be narrower than that of LAS. This might in part
explain the high chemoselectivity of CSR in the O-deacetylation
of 2 bearing, as R substituents, phenyl and benzyl groups in the
previous paper [17], or m-(phenyl)benzoyl group 2b in this study;
however, the reasons for the low chemoselective O-deacetylation
activity toward 2a and the extremely low O-deacetylation activity
toward 2c remain unexplained.
Fig. 3 shows the optimized structures of 2a–c calculated with
Hartree-Fock 6-31 + G* model. Similar structures [15] were also
obtained with other models used (data not shown). As expected, the
o-phenyl substituent of 2a, in contrast with 2b and 2c, covered the
1-␤-O-acyl linkage; the o-phenyl substituent might consequently
prevent access to the catalytic sites and protect the linkage from
LAS- and CSR-catalyzed hydrolysis. Because LAS showed hydrolytic
activity toward a medium-chain fatty acid ester (7d), the 1-␤-O-
acyl linkages of 2b and 2c also might be oriented to the catalytic
site of LAS and be subsequently hydrolysed, resulting in a lowered
chemoselectivity of LAS toward 2b and 2c. On the other hand, due
to the hypothesized binding site of CSR, the 1-␤-O-acyl linkages of
2b and 2c could neither be oriented to the catalytic site of CSR nor
be hydrolysed by CSR.
3.5. Acyl preference of LAS- and CSR-catalyzed hydrolysis of 7a–e
Although A. niger lipase generally cleaved anomeric acetyl esters
of fully O-acetylated sugars regioselectively [20–24], to our knowl-
edge there have been no reports on acyl preference of LAS and
CSR-catalyzed ester hydrolytic reaction. Since acyl preferences of
LAS and CSR are thought to be an important factor [13] determining
the chemoselectivity toward 2, acyl preferences were investigated
using p-nitrophenyl esters 7a–e (shown in Scheme 2).
The enzymatic hydrolysis of 7a–e was performed under the
same reaction conditions used for the substrates 2a–c, and the lib-
erated p-nitrophenol was determined by HPLC. Km and Vmax values
of LAS and CSR were obtained for 7a–b and 7a–c, respectively. Since
the solubility of substrates (7c–e with LAS and 7d–e with CSR) in the
incubation mixture was too low to obtain the Km and Vmax values,
apparent first-order rate constants (alternatives to Vmax/Km values)
were determined at a low substrate concentration. The obtained
kinetic parameters are listed in Table 3.
LAS exhibited hydrolytic activity toward all substrates 7a–e
with the medium-chain fatty acid ester 7d being the most effi-
ciently hydrolysed, as derived from the Vmax/Km values. The
finding that LAS has a broad acyl preference including the benzoyl
group (Table 3) might in part explain the low chemoselective O-
deacetylation of 2 bearing phenyl and benzyl moieties as R groups
[12] as well as the m- and p-isomers 2b and 2c in this study.
Conversely, the finding that LAS showed high chemoselective O-
deacetylation only toward the o-phenyl isomer 2a, together with
the findings [11,12] that LAS also showed high chemoselective O-
deacetylation toward 2 with R groups such as o-(anilino)phenyl
and o-(2,6-dichloroanilino)benzyl groups, indicates that sterically
bulky R groups, especially as ortho substituents, might prevent
the 1-␤-O-acyl linkage from LAS-catalyzed hydrolysis. The poor
chemoselectivity for 2b and 2c might be rationalized by considering
4. Conclusion
We investigated LAS- and CSR-catalyzed chemoselective and
regioselective O-deacetylation of 2a–c. The position of the phenyl
substituent of 2a–c strongly affected the chemoselectivity; high
chemoselectivity of LAS and CSR was observed toward ortho-isomer
2a and meta-isomer 2b, respectively. The chemoselectivity wors-
ened in the opposite combination. These data indicate that LAS and

82
A. Baba, T. Yoshioka / Journal of Molecular Catalysis B: Enzymatic 69 (2011) 74–82
CSR play complementary roles in chemoselective O-deacetylation
of 2 bearing R groups of different steric bulkiness. Based on the cat-
alytic efficiency of the enzymes toward p-nitrophenyl esters 7a–e,
LAS wasshown to have abroad acylpreferencethat was sufficientto
hydrolyse the benzoyl ester 7e. In contrast, CSR was shown to have
a smaller binding site that was almost completely limited to the
acetyl ester 7a. These characteristics, together with the computa-
tionally optimized structures, could largely explain the differences
observed in the chemoselective O-deacetylation of 2a–c. LAS, but
not CSR, exhibited regioselective 3-O-deacetylation activity toward
2a–c to yield the corresponding 2,4-di-O-acetyl derivatives 5a–c,
among which 5a and 5b were obtained in good yields. CSR, differ-
ing from LAS, hydrolysed 5a–c much more effectively than 2a–c. A
one-pot procedure for the synthesis of ␤GAs 4a–c from 2a–c were
performed with LAS and/or CSR for the deprotection of the O-acetyl
group followed by CALB for the deprotection of the carboxyl methyl
ester. Application of these enzymes to hydrolytic deprotection for
the synthesis of other glycosides with an alkaline-labile glycosidic
bonding is currently under investigation.
[14] A. Baba, T. Yoshioka, Chem. Res. Toxicol. 22 (2009) 1998–2008.
[15] T. Yoshioka, A. Baba, Chem. Res. Toxicol. 22 (2009) 1559–1569.
[16] U.T. Bornscheuer, R.J. Kazlauskas, Hydrolases in Organic Synthesis: Regio- and
Stereoselective Biotransformations, Wiley-VCH, Weinheim, 1999, 131–147.
[17] K. Drauz, H. Waldmann, in: K. Drauz, H. Waldmann (Eds.), Enzyme Catalysis in
Organic Synthesis: A Comprehensive Handbook, vol. III, 2nd ed., Wiley-VCH,
Weinheim, 2002, pp. 1369–1380.
[18] N.B. Bashir, S.J. Phythian, A.J. Reason, S.M. Roberts, J. Chem. Soc., Perkin Trans.
1 (1995) 2203–2322.
[19] D. Kadereit, H. Waldmann, Chem. Rev. 101 (2001) 3367–3396.
[20] J.-F. Shaw, A.M. Klibanov, Biotechnol. Bioeng. 29 (1987) 648–651.
[21] M. Filice, R. Fernandez-Lafuente, M. Terreni, J.M. Guisan, J.M. Palomo, J. Mol.
Catal. B: Enzym. 49 (2007) 12–17.
[22] W.J. Hennen, H.M. Sweers, Y.-F. Wang, C.-H. Wong, J. Org. Chem. 53 (1988)
4939–4945.
[23] G.-T. Ong, K.-Y. Chang, S.-H. Wu, K.-T. Wang, Carbohydr. Res. 265 (1994)
311–318.
[24] A. Giordano, A. Trincone, Tetrahedron Lett. 43 (2002) 4939–4942.
[25] D.F. Detar, C. Delahunty, J. Am. Chem. Soc. 105 (1983) 2734–2739.
[26] J.F. Kirsch, W. Clewell, A. Simon, J. Org. Chem. 33 (1968) 127–132.
[27] H.E. Gottlieb, V. Kotlyar, A. Nudelman, J. Org. Chem. 62 (1997) 7512–7515.
[28] K. Yoshiomoto, Y. Itatani, K. Shibata, Y. Tsuda, Chem. Pharm. Bull. 28 (1980)
208–219.
[29] P.-E. Jansson, L. Kenne, E. Schweda, J. Chem. Soc., Perkin Trans. 1 (1987)
377–383.
[30] H. Komura, A. Matsuno, Y. Ishido, K. Kushida, K. Aoki, Carbohydr. Res. 65 (1978)
271–277.
Appendix A. Supplementary data
[31] J.M. Sugihara, Adv. Carbohydr. Chem. Biochem. 8 (1953) 1–44.
[32] A.H. Haines, Adv. Carbohydr. Chem. Biochem. 33 (1976) 11–109.
[33] C. Vogel, S. Kramer, A.-J. Ott, Liebigs Ann/Recueil (1997) 1425–1428.
[34] T. Horrbin, C.H. Tran, D. Crout, J. Chem. Soc., Perkin Trans. 1 (1998) 1069–1080.
[35] M. Terreni, R. Salvetti, L. Linati, R. Fenandez-Lafuente, G. Fernández-Lorente, A.
Bastida, J.M. Guisan, Carbohydr. Res. 337 (2002) 1615–1621.
[36] T.-C. Chien, J.-W. Chern, Carbohydr. Res. 339 (2004) 1215–1217.
[37] K.-F. Hsiao, H.-J. Lin, D.-L. Leu, S.-H. Wu, Bioorg. Med. Chem. Lett. 4 (1994)
1629–1632.
[38] S. Inigo, M.T. Porro, J.M. Montserrat, L.E. Iglesias, A.M. Iribarren, J. Mol. Catal. B:
Enzym. 35 (2005) 70–73.
[39] M. Mastihubová, J. Szemesová, P. Biely, Bioorg. Med. Chem. 14 (2006)
1805–1810.
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.molcatb.2010.12.014.
References
[1] E.M. Faed, Drug Metab. Rev. 15 (1984) 1213–1249.
[2] L. Sphan, L.Z. Benet, Drug Metab. Rev. 24 (1992) 5–47.
[3] R.B. van Breeman, C. Fenselau, Drug Metab. Dispos. 13 (1985) 318–320.
[4] P.C. Smith, L.Z. Benet, A.F. McDonagh, Drug Metab. Dispos. 18 (1990) 639–644.
[5] A.S. Kalgutker, I. Gardner, R.S. Obach, C.L. Shaffer, E. Callegari, K.R. Henne, A.E.
Mutlib, D.K. Dalvie, J.S. Lee, Y. Nakai, J.P. ODonnell, J. Boer, S.P. Harriman, Curr.
Drug Metab. 6 (2005) 161–225.
[6] A.S. Kalgutkar, R.S. Obach, T.S. Maurer, Curr. Drug Metab. 8 (2007) 407–447.
[7] C. Skonberg, J. Olsen, K.G. Madsen, S.H. Hansen, M.P. Grillo, Exp. Opin. Drug
Metab. Toxicol. 4 (2008) 425–438.
[40] T. Bavaro, S. Rocchietti, D. Ubiali, M. Filice, M. Terreni, M. Pregnolato, Eur. J. Org.
Chem. (2009) 1967–1975.
[41] M.B. Sabaini, M.A. Zinni, M. Mohorcˇicˇ, J. Friedrich, A.M. Iribarren, L.E. Iglesias, J.
Mol. Catal. B: Enzymatic 62 (2010) 225–229.
[42] K.-F. Hsiao, S.-H. Wu, K.-T. Wang, Bioorg. Med. Chem. Lett. 3 (1993) 2125–2128.
[43] H.-P. Chen, K.-F. Hsiao, S.-H. Wu, K.-T. Wang, Biotechnol. Lett. 17 (1995)
305–308.
[44] K. Moriyoshi, H. Yamanaka, T. Ohmoto, T. Ohe, K. Sakai, Biosci. Biotechnol.
Biochem. 69 (2005) 1292–1299.
[8] B.K. Park, M. Pirmohamed, N.R. Kitteringham, Chem. Res. Toxicol. 11 (1998)
969–988.
[9] D.A. Smith, E.F. Schmid, Curr. Opin. Drug Discov. Dev. 9 (2006) 38–46.
[10] D.A. Smith, R.S. Obach, Chem. Res. Toxicol. 22 (2009) 267–279.
[11] A. Baba, T. Yoshioka, Org. Biomol. Chem. 4 (2006) 3303–3310.
[12] A. Baba, T. Yoshioka, J. Org. Chem. 72 (2007) 9541–9549.
[13] A. Baba, T. Yoshioka, Chem. Res. Toxicol. 22 (2009) 158–172.
[45] S.J. Jun, M.S. Moon, S.H. Lee, C.S. Cheong, K.S. Kim, Tetrahedron Lett. 46 (2005)
5063–5065.
[46] Z. Car, V. Petrovic´, S. Tomic´, J. Carbohydr. Chem. 25 (2006) 713–723.