Engineering of glucoside acceptors for the regioselective synthesis of β-(1→3)-disaccharides with glycosynthases

Article abstract of DOI:10.1016/j.carres.2008.07.018

Glycosynthase mutants obtained from Thermotoga maritima were able to catalyze the regioselective synthesis of aryl β-d-Galp-(1→3)-β-d-Glcp and aryl β-d-Glcp-(1→3)-β-d-Glcp in high yields (up to 90 %) using aryl β-d-glucosides as acceptors. The need for an aglyconic aryl group was rationalized by molecular modeling calculations, which have emphasized a high stabilizing interaction of this group by stacking with W312 of the enzyme. Unfortunately, the deprotection of the aromatic group of the disaccharides was not possible without partial hydrolysis of the glycosidic bond. The replacement of aryl groups by benzyl ones could offer the opportunity to deprotect the anomeric position under very mild conditions. Assuming that benzyl acceptors could preserve the stabilizing stacking, benzyl β-d-glucoside firstly assayed as acceptor resulted in both poor yields and poor regioselectivity. Thus, we decided to undertake molecular modeling calculations in order to design which suitable substituted benzyl acceptors could be used. This study resulted in the choice of 2-biphenylmethyl β-d-glucopyranoside. This choice was validated experimentally, since the corresponding β-(1→3) disaccharide was obtained in good yields and with a high regioselectivity. At the same time, we have shown that phenyl 1-thio-β-d-glucopyranoside was also an excellent substrate leading to similar results as those obtained with the O-phenyl analogue. The NBS deprotection of the S-phenyl group afforded the corresponding disaccharide quantitatively.

Full text of DOI:10.1016/j.carres.2008.07.018

Carbohydrate Research 343 (2008) 2939–2946
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Engineering of glucoside acceptors for the regioselective
synthesis of b-(1?3)-disaccharides with glycosynthases
*
Zsuzanna Marton, Vinh Tran, Charles Tellier, Michel Dion, Jullien Drone, Claude Rabiller
Biotechnologie, Biocatalyse, Biorégulation (UMR CNRS 6204), Université de Nantes, Faculté des Sciences et des Techniques, 2, Rue de la Houssinière, BP 92208, F-44322 Nantes, France
a r t i c l e i n f o
a b s t r a c t
Article history:
Glycosynthase mutants obtained from Thermotoga maritima were able to catalyze the regioselective syn-
Received 18 March 2008
Received in revised form 13 July 2008
Accepted 15 July 2008
thesis of aryl b-
D-Galp-(1?3)-b-D-Glcp and aryl b-D-Glcp-(1?3)-b-D-Glcp in high yields (up to 90 %) using
aryl b- -glucosides as acceptors. The need for an aglyconic aryl group was rationalized by molecular mod-
D
eling calculations, which have emphasized a high stabilizing interaction of this group by stacking with
W312 of the enzyme. Unfortunately, the deprotection of the aromatic group of the disaccharides was
not possible without partial hydrolysis of the glycosidic bond. The replacement of aryl groups by benzyl
ones could offer the opportunity to deprotect the anomeric position under very mild conditions. Assum-
ing that benzyl acceptors could preserve the stabilizing stacking, benzyl b-D-glucoside firstly assayed as
acceptor resulted in both poor yields and poor regioselectivity. Thus, we decided to undertake molecular
modeling calculations in order to design which suitable substituted benzyl acceptors could be used. This
Available online 13 September 2008
Keywords:
Glycosidases and glycosynthases
Molecular modeling
b-(1?3)-Disaccharides
study resulted in the choice of 2-biphenylmethyl b-
imentally, since the corresponding b-(1?3) disaccharide was obtained in good yields and with a high reg-
ioselectivity. At the same time, we have shown that phenyl 1-thio-b- -glucopyranoside was also an
D-glucopyranoside. This choice was validated exper-
D
excellent substrate leading to similar results as those obtained with the O-phenyl analogue. The NBS
deprotection of the S-phenyl group afforded the corresponding disaccharide quantitatively.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
separate and in competition between the transglycosylation reac-
tion and the hydrolysis of the substrate and of the product. In order
Carbohydrate ligands play an important role in a number of bio-
logical mechanisms. The understanding of their action and, conse-
quently, their use as drugs depend on the availability of the
oligosaccharides involved in the processes. The difficulties encoun-
tered in building a glycosidic bond regioselectively and stereose-
lectively are well known. Despite the considerable development
of efficient methods in this field, the synthesis by means of conven-
tional chemical methods suffers from the need for cumbersome
protection–deprotection sequences. In order to avoid this disad-
vantage, enzymatic methods have been proposed in the last decade
considering the regio- and stereo-selectivities usually induced by
biocatalysts. This is the case for glycosyl transferases, which cata-
lyze such a synthesis with both high yield and high selectivity.
Unfortunately, the need for costly nucleotide donors greatly limits
their application for large-scale production of saccharides. In an-
other approach, the transfer activity exerted by ‘retaining’ glycosi-
dases also provided an efficient alternative, since these enzymes
accommodate much simpler and readily available donors. Prob-
lems encountered in this case were incomplete regioselectivity,
which resulted in mixtures of regioisomers that were difficult to
to improve the properties of the glycosidases, two strategies were
undertaken.
The first one, using directed mutagenesis was developed by
Withers and co-workers with exoglycosidases1–7 and by Planas
and co-workers with endoglycosidases.8–12 The mutant enzymes
thus produced were called glycosynthases because they were de-
void of hydrolytic activity but maintained a high transglycosyla-
tion ability. Thus, chemo-enzymatic synthesis of b-(1?3)-
oligosaccharides was achieved using the ‘glycosynthase methodol-
ogy’.13 We have also recently validated this approach with a b-
D-
glycosidase (TtbGly) from Thermus thermophilus,14 an enzyme
cloned and overexpressed in our laboratory.15 Thus, E338G or
E338S mutants allowed the synthesis of phenyl b-
b- -Glcp and phenyl b- -Galp-(1?3)-b- -Glcp in yields of up to
85% starting from -glycosyl fluorides as donors and phenyl b-
glucopyranoside (bGlcC6H5) as an acceptor (see Scheme 1).
D-Glcp-(1?3)-
D
D
D
a
D-
The second strategy developed recently by us was based on
directed evolution.16,17 Our results indicated that some mutant
enzymes had completely lost their hydrolytic activity while
keeping the transglycosidase one. Such enzymes were called
transglycosidases.
A disadvantage of the Ttbgly glycosidases is the need for a phe-
nyl ring at the (+2) sub-site to ensure the correct positioning of the
* Corresponding author. Tel.: +33 (0)2 51 12 57 32; fax: +33 51 12 56 37.
E-mail address: claude.rabiller@univ-nantes.fr (C. Rabiller).
0008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.07.018

2940
Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
E338G mutant TtβGly
OH OH
OH
OH
OH OH
glycosidase from
O
O
O
Thermus thermophilus
O
HO
HO
+
O
O
HO
O
HO
HO
OH
OH
OH
OH
Yield 84%
F
Scheme 1. Regioselective synthesis of phenyl b-
thermophilus.
D-Galp-(1?3)-b-D-Glcp catalyzed by the glycosynthase E338G obtained by mutagenesis of TtbGly glycosidase from Thermus
acceptor in the active site.14 Furthermore, this substituent has to
be eliminated in order to subsequently functionalize the anomeric
position. Previous attempts to achieve this step were unsuccessful,
since the chemical deprotection of the aglycon by hydrolysis was
in competition with the breakdown of the glycosidic bond. This pa-
per deals with the design of suitable acceptors bearing readily
removable anomeric substituents and able to induce the same reg-
ioselectivity. For this purpose, several acceptor subtrates like aryl
requirements, we have tested several glucosides acting as accep-
tors in the glycosylation reaction mediated by the glycosynthase
E338G using
a-galactopyranosyl fluoride and a-glucopyranosyl
fluoride as donors. The list of the acceptors used in this work is
shown in Scheme 2, and the results obtained are summarized in
Table 1.
Our primary strategy was to replace the phenyl group by an-
other substituted aromatic ring that is able to allow the grafting
of a saccharide on a polymeric framework. For example, 4-nitro-
phenyl or 4-cyanophenyl groups could be hydrogenated to provide
amines that could be used for this purpose according to well-
known methodologies. Moreover, O-4-nitrophenyl group, a better
leaving group than O-phenyl one, could be hydrolyzed without
cleavage of the b-(1?3) glycosidic linkage.
and benzyl b-
of their aglycon part to be further eliminated. Selection of suitable
b- -glucopyranosides was previously based on molecular model-
D-glucopyranosides, were tested regarding the ability
D
ing calculations on Ttbgly enzyme/substrates complexes.
2. Results and discussion
As indicated by our previous results, the replacement of the b-
O-phenyl group by a b-4-nitrophenyl analogue had no significant
effect on either the regioselectivity or the glycosylation yields
(see Table 1, entries 1–3). Unfortunately, although the reduction
of the nitro group was quite easy to achieve, the further use of
the amino group resulted in bad grafting yields of the disaccharide
on proteins. In other respects, the removal of O-4-nitrophenyl
group via hydrolysis could not be performed without partial break-
down of the b-(1?3) disaccharide bond. The use of 4-cyanophenyl
Therapeutic applications of oligosaccharides are determined by
their multi-grafting on proteins. For this purpose, the functionali-
zation of the reactive free anomeric carbon is often used. Mean-
while, the glycosynthase E338G (Ttbgly from Thermus
thermophilus) acted as an efficient catalyst for the regioselective
synthesis of the b-(1?3)-disaccharides when the acceptor beared
a b-O-aryl group. The phenyl group used in our previous studies
was shown to be non-easily removable. Thus, two strategies could
be considered in order to produce useful oligosaccharides: either to
find an aromatic group bearing chemical functions further usable
for the grafting with proteins or to design aglycon aromatic groups
easier to remove than the phenyl group. According to these
b-D-glucopyranoside 1c as an acceptor led to similar results
although the glycosylation reaction rate was lower in that case.
Consequently, a higher ratio of [donor]/[1c] or [1c]/[[donor] had
to be used since the rate of spontaneous hydrolysis of galactosyl
OH
OH
HO
HO
Compound
R
O
O
HO
HO
R
1a 2a 3a 4a 5a
1b 2b 3b
1c 2c 3c 4c 5c
1d 2d 3d 4d 5d
C6H5-O
4-NO2-C6H5-O
4-CN-C6H5-O
C6H5-CH2-O
OH
OH
F
1
E338G mutant TtβGly
glycosidase from
Thermus thermophilus
1e 2e 3e
1f 2f 3f
O-H C
2
C6H5-S
OH
OH
HO
OH
HO
O
R
OH
O
HO
O
Deprotection
HO
O
O
HO
O
2
OH
HO
OH
OH
OH
OH
2 : β-Gal-(1→3)-Glc,3 : β-Gal-(1→6)-Glc
4 : β-Glc-(1→3)-Glc,5 : β-Glc-(1→6)-Glc
6 : β-Gal-(1→3)-Glc
OH
OH
O
Deprotection
O
HO
HO
3
HO
O
OH
OH
OH
7 : β-Gal-(1→6)-Glc
Scheme 2. Reactions, substrates and compounds.

Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
2941
Table 1
Experimental conditions and b-(1?3) disaccharide yields using various acceptors 1 and a glycosynthase (E338G) obtained from Thermus thermophilus glycoside hydrolase as a
catalyst
Entry
Donor
Acceptor
[Donor] (mmol /L)
[Acceptor] (mmol/L)
b-(1?3)-Disaccharide
b-(1?6)-Disaccharide
Transglycosylation yields (%)
1
2
3
4
5
6
7
8
9
Gal-F
Glc-F
Gal-F
Gal-F
Gal-F
Glc-F
Glc-F
Gal-F
Glc-F
Gal-F
Gal-F
Gal-F
1a
1a
1b
1c
1c
1c
1c
1d
1d
1e
1f
50
50
50
50
100
50
100
100
100
50
25
30
50
50
50
50
50
50
50
50
50
25
25
15
100
100
100
100
100
100
100
55
56
84
96
94
0
0
0
0
0
0
0
45
44
16
4
86*
96*
84
45
85
45
95
67
82
70
81
94
10
11
12
1f
6
*
See Ref. 14.
Table 2
fluoride became competitive (see Table 1, entries 4–6). The non-
commercially available monosaccharide 1c was obtained with
the standard sugar chemistry procedures: coupling of 4-hydrox-
Enzyme–acceptor docking energy (kcal/mol)
Acceptor
Energy
Variation
ybenzonitrile with 2,3,4,6-tetraacetyl-a-D-glucopyranosyl bromide
1a
1f
1d
1e
ꢀ46.6
ꢀ44.4
ꢀ39.7
ꢀ62.7
Ref.
+2.2
+6.9
ꢀ16.1
in the presence of silver carbonate, followed by Zemplén deacety-
lation. Unfortunately, the reduction of the nitrile group of saccha-
rides 1c, 2c or 4c proved to be very difficult to achieve in good
conditions. Moreover, this first strategy suffered from the presence
of an aromatically substituted cycle that may induce undesirable
immunogenic responses, when these molecules are used as thera-
peutic agents. So, the need for saccharides deprotected at the
anomeric position remained very important since the functionali-
zation at this position is quite easy via the synthesis of anomeric
amino derivatives.
The complex with 1a was chosen as a reference. Its replacement by 1f had no
significant effect on energy docking, while 1d induced much more significant rise in
docking energy. Conversely, 1e led to a severe docking energy decrease. In this case,
the increase in ligand–enzyme contact surface only partly explains this strong
affinity between the two partners.
b-(1?3)-disaccharide. It is likely that the introduction of a methy-
lene group between the anomeric position and the phenyl group
promoted an increased conformational mobility that induced a re-
laxed regioselectivity. This disappointing result prompted us to use
molecular modeling techniques in order to choose a suitable
benzylic acceptor. For this purpose, we screened in silico a set of
acceptors like 2-, 3-, or 4-hydroxybenzyl , 2-, 3- or 4-methylbenzyl
According to these requirements, the use of 4-methoxyphenyl
b-D-glucopyranoside as an acceptor was assayed, since the result-
ing disaccharide could be deprotected using ceric ammonium ni-
trate.18 Unfortunately, this acceptor was a poor substrate for our
mutant enzyme, thus resulting in very low transglycosylation
yields. Similarly, benzyl derivatives looked like very attractive can-
didates due to an easy deprotection of the b-O-benzyl aglycon by
catalytic reduction under very mild conditions. Moreover, the pres-
ence of the aromatic group separated by a single methylene group
from the anomeric carbon could preserve the stabilizing stacking
with W312 of Ttbgly glycosynthases. So, we synthesized benzyl
b-D-glucopyranosides and 4- or 2-biphenylmethyl b-D-glucopyr-
anosides. Calculations showed that hydroxybenzyl derivatives fit
more or less correctly in the active site, while the introduction of
methylbenzyl ones resulted in higher energy interactions. The
same was true for 4-biphenylmethyl b-D-glucopyranoside, while
b-
lyzed by the wild type Ttbgly glycosidase in the presence of
p-nitrophenyl b- -glucopyranoside (pNPbGlc) as a donor and
D
-glucopyranoside 1d using the transglycosylation reaction cata-
an unexpected high stabilization was obtained with the 2-biphe-
nylmethyl analogue 1e in which the second phenyl group
produced a stabilizing stacking with W312 of the mutant glyco-
synthase (see Table 2, Fig. 1 and Section 3).
The preparation of 1e by means of the enzymatic approach used
for 1d was not possible due to the very poor solubility of 2-biphe-
nylmethanol in water. Thus, the synthesis of 1e was achieved using
conventional sugar chemistry. Coupling of 2-biphenylmethanol
D
benzyl alcohol as an acceptor. The kinetic study of the correspond-
ing reaction indicated that theoretical yields of about 40% could be
reached. This very easy one-step enzymatic reaction was preferred
in that case to the conventional chemical approach, since benzyl
alcohol was a good substrate for the enzyme. The ability of 1d to
act efficiently in the glycosylation reactions is tested as described
in Scheme 2. Unfortunately, 1d was also shown to be a poor sub-
strate, since the yields were lower than those previously obtained
and the regioselectivity was also poor (see Table 1, entry 8). This
was particularly the case for the synthesis of the b-(1?3)-disac-
charide 2d, which was obtained in similar amounts to its
b-(1?6)-regioisomer 3d. The same was true when using Glc-F as
a donor (see Table 1, entry 9). Despite this disappointing result,
we have undertaken the deprotection of b-O-benzyl aglycon by
catalytic hydrogenation of the mixture of regioisomers 2d and
3d. This reaction afforded quantitatively the corresponding
with
2,3,4,6-tetra-O-acetyl-1-O-trichloroacetimidoyl-a-D-gluco-
pyranose was catalyzed by means of trifluoroboride etherate. Sub-
sequent deacetylation in the presence of ammoniac in methanol
afforded 1e. The glycosylation reaction then performed in the pres-
ence of Gal-F as a donor in the same conditions as those previously
used, resulted in dramatic improvement of the b-(1?3) regioselec-
tivity of the E338G [relative percentages of b-(1?3) 2e regio-iso-
mer versus b-(1?6) 3e: 84/16, see Table 1, entry 10]. This result
constituted an additional indication of the need for the stabiliza-
tion of the acceptor by stacking a phenyl group of the latter with
W312. The deprotection of the biphenylmethyl at O-2 group,
achieved via Pd/C catalyzed hydrogenolysis, afforded quantita-
mixture of the free disaccharides, b-D-Galp-(1?3)-D-Glcp 6 and
b- -Galp-(1?6)- -Glcp 7. Correspondingly, molecular modeling
D
D
tively the corresponding b-D-Galp-(1?3)-D-Glcp 6. The attachment
calculations revealed that the benzylic acceptor 1d was unable to
fit properly in the active site in order to synthesize the desired
of a readily removable functional group to a glycosidase substrate
facilitated a rational control of the regioselectivity. The same

2942
Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
Figure 1. Rational design of an acceptor in the transglycosylation reaction catalyzed by TtbGly glycosynthase. Phenyl b-
biphenylmethyl b-
-glucopyranoside (*1e, *white sticks) are acting as acceptors in the glycosylation reaction. They both present one aromatic ring strongly stacked with
D
-glucopyranoside (*1a, *black sticks) and 2-
D
W312 (spheres) for the stabilization of the ligand fragment at sub-site (+2) with respect to b-(1?3) linkage formation (glycosyl enzyme inside the cleft in black ball and stick).
The catalytic site is visualized in mesh representation. Important catalytic residues (E164, N282) are also shown (ball and stick, medium, and light, respectively).
approach, previously reported by Lairson et al. in the case of glyco-
syltransferases resulted in similar results.19
3. Experimental
As an alternative, we also tested phenyl 1-thio-b-D-glucopyran-
3.1. General procedures
oside 1f, which is known to be easily deprotected via hydrolysis of
the thioglycosidic bond under very mild conditions. In this way, we
verified, by means of molecular modeling calculations that accep-
tor 1f is fitted into approximately the same position as its oxygen-
ated analogue. As expected, the replacement of the oxygen atom by
a sulfur did not induce any important topological modifications of
the position of the reactants in the active site, thus indicating that
the synthesis of the corresponding b-(1?3)-regioisomer should be
preferred. Once again, this prediction was quite satisfactory in
accordance with experimental data, since the use of 1f as an accep-
tor led to similar results to those obtained with phenyl glucoside
1a (Table 1, entries 1 and 11) although very minor amounts of
the b-(1?6) regioisomer were present. It should be noted that phe-
Chemicals supplied by Aldrich were used without further puri-
fication. The courses of the reactions were followed by means of
TLC (precoated Silica Gel 60 sheets Merck F254) and proton NMR
spectroscopy. The components of the reaction mixtures were sep-
arated by means of silica gel chromatography. Analysis of the NMR
1H and 13C resonances and subsequent structure assignments were
made by using standard 2D sequences (COSY HH HOHAHA and HC
correlations). The spectra were recorded with a Bruker DRX500
spectrometer operating at 500 MHz for 1H and at 126 MHz for
13C or with a Bruker AX400 spectrometer operating at 400 MHz
for 1H and at 100.6 MHz for 13C. Due to their solubility in water,
the spectra of the saccharides were recorded in D2O, and the chem-
ical shifts (in ppm) were quoted from the resonance of methyl pro-
tons of sodium 3-(trimethylsilyl)-propansulfonate (DSS) used as an
internal reference. For each compound, the proton chemical shifts
and coupling constants are given in Table 3, and the carbon chem-
ical shifts are given in Table 4.
nyl 1-thio-b-D-glucopyranoside 1f was already successfully used as
an acceptor in transglycosylation reactions catalyzed by exoglyco-
sidases.20 Substantial improvement of the yield (81–94%) was ob-
tained when working with a ratio of [donor]/[acceptor] = 2 (see
Table 1, entries 11 and 12). The deprotection of the thiophenyl
group by the action of N-bromosuccinimide in water afforded
quantitatively b-D-Galp-(1?3)-Glcp 6.
3.2. Molecular modeling calculations
In conclusion, we have shown that substrate engineering can be
a complementary approach to enzyme engineering to afford high
yields and regioselectivity in glucosidation reactions catalyzed by
All molecular modeling visualizations and calculations were
performed on an SGI workstation with Accelrys tools (InsightII
software and CFF91 force field). For the enzyme, the crystal struc-
ture of the native TtbGly (pdb:1ug6) and the starting modeled
glycosynthases. Thus, two acceptors, 2-biphenylmethyl b-
D-gluco-
pyranoside 1e and phenyl 1-thio-b- -glucopyranoside 1f were cho-
D
sen for their ability to produce regioselectively the corresponding
b-(1?3)-disaccharides in good yields. In both cases, the removal
of the aglycon afforded quantitatively b-D-Galp-(1?3)-D-Glcp 6.
complex with a DP3 oligomer b-
D-Galp-(1?3)-b-D-Glcp-(1?4)-D-
Glcp has already been described.16
Molecular modeling calculations were performed in two steps.
The first consisted in generating all probable conformations of sub-
strates mentioned above, regardless of the catalytic space con-
Moreover, this choice was supported by molecular modeling calcu-
lation, which proved to constitute a rational approach for the de-
sign of suitable reactants.

Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
2943
Table 3
1H NMR data of compounds described (500 MHz), solvent D2O, T = 35 °C
Comp.
H-1
H-2
H-3
H-4
H-5
H-6
H-60
Ph-CH2
Others
1c
1d
—
—
—
5.22
4.55
4.45
3.51
3.37
5.05
3.61
3.53
5.08
3.63
3.45
5.16
3.67
3.45
3.57
3.75
3.97
4.02
3.93
3.78
4.23
—
7.76, 7.22 (aromatics)
7.40–7.55 (aromatics)
2.00, 2.01, 2.02, 2.05 (CH3)
7.29–7.50 (aromatics)
7.3–7.7 (aromatics)
7.40, 7.57 (aromatics)
7.40–7.55 (aromatics)
—
7.23, 7.76, 7.76, 7.77 (aromatics)
—
7.26, 7.26, 7.26, 7.77 (aromatics)
—
4.97, 4.78
4.56, 4.80
Tetraacetyl 1e
1e
1f
2b
—
—
I
II
I
II
I
II
I
II
I
II
I
4.29
4.78
5.28
4.70
5.24
4.70
5.24
4.78
4.35
4.40
4.31
4.42
4.81
4.67
5.23
4.66
4.64
3.38a
3.35
3.87
3.63
3.81
3.63
3.81
3.38
3.27
3.50
3.28
3.52
3.53
3.57
3.71
3.43
3.60
3.23
3.52
3.90
3.69
3.88
3.69
3.87
3.53
3.55
3.67
3.54
3.65
3.78
3.66
3.90
3.73
3.67
3.34a
3.41
3.65
3.93
3.63
3.94
3.62
3.41
n.d.
3.90
n.d.
3.91
n.d.
3.25
3.47
3.67
3.73
3.69
3.73
3.70
3.49
3.34
3.40
3.36
3.39
n.d.
3.78
3.71
3.78
n.d.
3.77
3.80
3.77
3.72
3.81
n.d.
3.82
n.d
3.89
n.d
3.75
3.72
n.d.
3.65
3.88
3.95
3.80
3.93
3.77
3.93
3.93
n.d.
n.d
n.d.
n.d.
3.72
n.d.
4.71, 4.87
—
—
—
—
—
—
—
4.70, 4.86
—
4.71, 4.87
—
—
—
—
—
—
2c
4c
2d
2e
2f
7.40–7.55 (aromatics)
—
7.35–7.65 (aromatics)
—
7.35–7.60 (aromatics)
II
3.91
3.51
3.52
3.92
n.d.
—
—
—
—
6 (
a,b)
I
a
3.86
3.48
n.d.
3.83
3.89
n.d.
I b
II
3J1,2
3J2,3
3J3,4
3J4,5
3J5,6
3J5,6
2J6,6’
2JCH
2 -Ph
1c
1d
—
—
—
—
—
I
II
I
II
I
II
I
II
I
II
I
II
7.4
7.9
7.8
7.8
9.8
7.9
7.7
7.9
7.7
7.8
7.9
7.8
7.8
7.8
7.8
10.0
7.9
3.8
8.0
7.6
9.4
9.1
9.3
n.d
8.3
8.4
9.9
9.3
9.9
9.3
9.5
8.8
n.d.
8.8
n.d.
8.8
9.9
9.7
9.2
9.8
9.4
7.9
9.3
n.d.
8.8
9.0
3.4
8.6
3.4
8.6
9.1
8.3
n.d.
8.3
n.d.
8.8
3.4
9.0
8.3
3.4
9.9
1.1
9.3
n.d
9.8
9.9
0.8
9.8
0.9
9.9
9.8
9.6
3.4
9.6
3.4
9.6
n.d.
9.9
9.9
n.d.
5.7
5.6
2.5
2.1
5.6
5.3
3.8
5.4
4.1
5.1
5.1
2.3
n.d.
2.3
n.d
1.1.
n.d.
5.0
5.3
n.d
2.2
1.8
4.9
5.4
2.4
2.2
8.1
2.3
8.0
2.3
2.3
4.4
n.d.
4.4
n.d
n.d.
n.d.
2.5
2.3
n.d
ꢀ12.3
ꢀ12.3
ꢀ12.2
ꢀ12.2
ꢀ12.7
ꢀ12.4
ꢀ12.1
ꢀ12.4
ꢀ12.4
ꢀ12.4
ꢀ12.4
ꢀ12.2
n.d
—
8.9 (aromatics)
n.d.
n.d.
n.d.
n.d.
n.d
—
8.9 (aromatics)
—
8.9 (aromatics)
—
n.d
—
n.d.
—
n.d.
—
—
ꢀ12.3
ꢀ11.7
ꢀ11.7
ꢀ
Tetraacetyl 1e
1e
1f
2b
—
—
—
—
—
—
2c
4c
2d
2e
2f
ꢀ11.7
—
ꢀ12.2
ꢀ11.7
—
ꢀ12.2
n.d.
—
n.d
n.d.
n.d
ꢀ12.3
ꢀ12.3
n.d
6 (
a,b)
I
a
Ib
II
—
—
—
—
Chemical shifts in ppm, ref: sodium 3-(trimethylsilyl)-propansulfonate, coupling constants J in Hz. Tetraacetyl 1e, solvent CDCl3, ref: TMS.
a
May be reversed.
Table 4
13C NMR data of the compounds described (125 MHz), solvent D2O, T = 35 °C)
Comp.
C-1
C-2
C-3
C-4
C-5
C-6
Ph-CH2
Others
1c
1d
—
—
—
—
—
I
II
I
II
I
II
I
II
I
II
I
II
102.1
100.3
89.4
100.8
90.0
104.7
106.0
102.0
106.0
101.9
105.5
102.6
100.4
102.9
100.8
89.7
72.0
72.2
71.5
72.1a
74.4
76.9
74.0
75.4
74.0
75.2
76.1
72.4
71.0
72.8
71.0
70.8
74.0
78.0
76.4
74.0
78.2
74.9
68.8
75.0
80.0
87.9
75.3
86.0
75.2
86.7
78.2
78.1
72.4
78.3
72.4
88.3
75.3
85.3
87.6
75.4
75.5
68.7
73.1
68.9a
72.2
72.2
71.3
70.8
71.4
70.5
72.3
n.d.
78.9
75.0
71.9
75.2
82.6
79.0
78.0
78.6
78.1
78.5
78.7
74.3
n.d.
63.2
59.9
62.0
60.3
63.6
63.7
63.8
63.3
63.8
63.1
63.4
60.2
61.1
60.0
61.1
63.5
63.8
63.4
63.6
63.6
—
122.5 (CN) ; 119.7, 137.7, 107.7, 162.8 (aromatics)
70.4
68.6
71.1
—
—
—
—
—
—
Tetraacetyl 1e
1e
1f
20.8, 20.9 (CH3) ; 170.9,170.5,169.6, 169.5 (CO); 127.5 ꢀ 142.3 (aromatics)
127.6, 127.9, 128.4, 128.8, 129.3, 130.1, 130.4, 133.6, 140.4, 142.4 (aromatics)
134.7, 134.3,132.0, 130.8 (aromatics)
2b
—
2c
4c
2d
2e
2f
107.8, 119.8, 137.4, 162.8 (aromatics), 122.6 (CN)
119.7, 137.3, 107.7, 162.7 (aromatics); 122.4 (CN)
70.4
—
68.8
—
—
—
—
—
—
(Aromatics)
—
68.6
n.d.
74.6
n.d.
127.6, 127.8, 128.5, 128.8, 129.4, 130.1, 130.6, 133.5, 140.4, 142.5 (aromatics)
—
130.8, 132.0, 134.3, 134.7 (aromatics)
—
—
—
—
68.6
73.9
71.3
71.0
71.0
71.3
82.3
78.0
74.0
78.2
73.6
105.9
94.7
98.4
6 (
a,b)
I
a
I b
II
106.1
Chemical shifts in ppm, ref : sodium 3-(trimethylsilyl)-propansulfonate. Tetraacetyl 1e solvent CDCl3, ref : TMS.
a
May be reversed.

2944
Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
straints: all linkage bonds yielding substrate flexibility were sys-
tematically scanned at regular intervals (10°) according to the pro-
tocol already described.21
3.4. Enzymatic synthesis of benzyl b-D-glucopyranoside 1d
To a soln of pNPbGlc (500 mg, 1.66 mmol) in a solvent com-
posed of 20% DMF and 80% sodium phosphate buffer (55 mL,
0.05 M, pH 7), were successively introduced benzylic alcohol
The second step consisted in locating these substrates in their
right place in the catalytic site for subsequent complex energy
minimization. From previous analysis of the TtbGly/DP3 substrate
complex16, it was possible to deduce structural information for
specific docking at each sub-site. Sub-site (ꢀ1), deep in the cata-
lytic site, was characterized by a very strong docking energy
involving all possible stabilizing factors. Contrary to this first case,
sub-site (+1) presented a weak docking mainly due to Van der
Waals terms. Finally, sub-site (+2) also had a rather weak docking
partly due to W312 stacking (with a sugar or aglycon aromatic
ring). Thus, it could be deduced that the sugar ring initially located
at subsite (+1) had a significant degree of freedom without notable
consequence in the total enzyme/substrate docking energy. Con-
versely, W312 should have a pivotal influence on the stabilization
of the acceptor moiety. Accordingly, all conformations of acceptors
were tested by first superimposing the sugar ring atoms on those
of the corresponding modeled ring at sub-site (+1) with b-(1?3)
disposition with sub-site (ꢀ1). All complexes presenting high steric
hindrance between enzyme and substrate were discarded. The few
remaining solutions were minimized with all heavy atoms of the
enzyme fixed. The same energy minimization protocol (steepest
descent algorithm, 10,000 iterations) was applied to all calcula-
tions to evaluate the capability of each of these conformations to
adapt to severe catalytic cavity constraints preserving the
b-(1?3) selectivity, while using (when possible) the stacking
opportunity between the aglycon moiety and W312. Due to these
drastic energy minimization conditions, these results should be
used qualitatively, and it is generally admitted that this interaction
energy is a relevant criterion of comparison when the total
enzyme/substrate contact surface is quite comparable. In practice,
this condition is only fulfilled between some compounds. For
example (see Table 2), the replacement of phenyl glucoside (1a)
by phenyl thioglucoside (1f) led to an increase of interaction
energy of only 2.2 kcal/mol (not significant within the protocol
(1.6 mL, 1.79 g, 16.6 mmol) and Ttbgly glycosidase sol. (333 lL)
prepared as previously described.15 The mixture was stirred at
30 °C for 2.0 h (increased reaction time resulted in a fast hydrolysis
of 1d). The reaction mixture was then quenched by warming at
100 °C for 5 min. Thus, water was distilled under diminished pres-
sure, and the resulting mixture was washed with CHCl3 in order to
remove DMF and excess benzylic alcohol. Next, saccharide 1d was
purified on a silica gel column (5:2:2 BuOH–EtOH–water, Rf 0.68).
After collection of the fractions containing 1d, and elimination of
the solvent, pure 1d was obtained (135 mg, yield 30%, 1H and 13C
NMR spectra of 1d, see Tables 3 and 4). The present spectra were
identical to those already published for authentic samples.22
3.5. Synthesis of 4-cyanophenyl 2,3,4,6-tetra-O-acetyl-b-D-
glucopyranoside
To 2,3,4,6-tetra-O-acetyl-a-D-glucopyranosyl bromide (5.2 g,
12.7 mmol), prepared according to a standard procedure23 and dis-
solved in 1:1 MeCN–CH2Cl2 (55 mL), molecular sieves (1 g, 3 Å)
were added. The mixture was stirred in the dark for 30 min at rt
and silver carbonate (6.46 g, 23.4 mmol) and 4-cyanophenol
(2.28 g, 19.2 mmol) were successively added. The reaction mixture
was stirred in the dark at rt for 12 h (completion of the reaction fol-
lowed by TLC of the remaining bromosugar). The reaction mixture
was then filtered over Celite, and the filtrate was concentrated over
diminished pressure. Dissolution of the oily product in CH2Cl2 re-
sulted in the precipitation of AgBr as a white solid, which was fil-
tered over Celite. Solvent evaporation resulted in an oily residue,
which was purified by silica gel chromatography (3:2 petroleum
ether–EtOAc) resulting in 4-cyanophenyl 2,3,4,6-tetra-O-acetyl-b-
D
-glucopyranoside as a syrup (3.1 g, 7.0 mmol).The present 1H
and 13C NMR spectra (Tables 3 and 4) were identical to those val-
used here). When 2-biphenylmethyl b-
D
-glucopyranoside (1e)
ues previously published for 4-cyanophenyl 2,3,4,6-tetra-O-acetyl-
was tested, the interaction energy was much more favorable
(ꢀ16.1 kcal/mol). Although a part of the gain was probably due
to a greater contact surface, it seems reasonable to consider that
the stacking of one phenyl ring with W312 significantly contrib-
uted to stabilize this acceptor in the catalytic site.
b-
3.6. Synthesis of 4-cyanophenyl b-
Zemplén deacetylation23 of 4-cyanophenyl 2,3,4,6-tetra-O-acet-
D
-glucopyranoside.24
D
-glucopyranoside 1c
yl-b- -glucopyranoside (3.1 g, 7.0 mmol) afforded 1c, which was
D
3.3. Kinetic study of the synthesis of benzyl
obtained as white crystals after recrystallization in ethanol (1.8 g,
6.4 mmol). The NMR parameters (see Tables 3 and 4) and physical
constants were consistent with the structure of 4-cyanophenyl b-
b-D-glucopyranoside 1d using 4-nitrophenyl b-D-glucopyranoside
and benzylic alcohol as substrates for Ttbgly
D
-glucopyranoside 1c, mp 185–187 °C, ½a D20
ꢀ92.3, (c 1.5, water
ꢁ
Sodium phosphate buffer (0.74 mL, 0.05 M, pH 7.0) was lyoph-
ilized, dissolved in D2O (2 mL), and lyophilized once more. The
resulting salts were re-dissolved into 0.74 mL D2O. Then, 0.2 mL
lit.25 mp 188–189 °C, ½a 2D0
ꢀ91.9, (c 1.5, water).
ꢁ
3.7. 2-Biphenylmethyl b-D-glucopyranoside 1e
of DMF containing 9 mg (30
lmol) of pNPbGlc and 32 mg
(300 mol) of benzylic alcohol were successively added. The
l
3.7.1. 2-Biphenylmethyl 2,3,4,6-tetra-O-acetyl-b-
D-
resulting soln was immediately filtered, transferred to an NMR
glucopyranoside
tube, and a 1H spectrum of the mixture was recorded at 30 °C.
2,3,4,6-Tetra-O-acetyl-1-O-trichloroacetimidoyl-
a-
D-glucopyra-
Then, 6
l
L of the enzymatic Ttbgly sol.15 was added, and the reac-
nose (prepared according to the standard procedure,23 724 mg, 1
.47 mmol), 2-biphenylmethanol (271 mg, 1.47 mmol), dry CH2Cl2
(7.5 mL), and BF3(Et)2O (19 drops) were successively introduced
in a reactor under a dry nitrogen flow. The reaction was allowed
to proceed for 2 h and the reaction mixture was quenched by add-
ing EtOAC (37 mL). Then, this soln was washed three times with a
saturated sodium hydrogencarbonate sol. (3 ꢂ 15 mL) and with
distilled water (15 mL). The organic layer was dried over sodium
sulfate, and the solvent was removed under diminished
pressure. Purification of the product thus obtained via silica gel
chromatography (2:3 AcOMe–petroleum ether, Rf 0.48) afforded
tion was allowed to proceed in the magnet of the spectrometer at
30 °C. Using this procedure, it was possible to perform the first
measurements at about 4 min after the introduction of the enzyme
in the NMR tube. Concerning the kinetics measurements, standard
conditions were used (PW = 30°, NS = 32, acquisition time =
0.344 s), and the integrals were measured using the Brucker stan-
dard program. The accuracy of the measurements of the concentra-
tions was estimated at about 5% for the major products. This study
showed that a 40 % maximum concentration was obtained for
benzyl b-D-glucopyranoside 1d after 2 h.

Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
2945
peracetylated 1e as an oily product (236 mg, 0.46 mmol, yield
31%). Anal. Calcd for C25H28O10: C, 61.48; H, 5.74. Found: C,
3.11. Synthesis of 4-nitrophenyl b-
D-glucopyranoside 2b
D
-galactopyranosyl-(1?3)-b-
61.36; H, 5.82. NMR data for 2-biphenylmethyl 2,3,4,6-tetra-O-
acetyl-b-
D-glucopyranoside, see Tables 3 and 4.
a-
D-Galactopyranosyl fluoride (18.2 mg, 0.1 mmol) and 4-nitro-
phenyl b- -glucopyranoside 1b (0.1 mmol) were dissolved into
D
2 mL of 150 mmol/L of ammonium carbonate buffer (pH 7.8). Then,
0.9 mL of Tt-b-Gly E338G sol., prepared as already described,15 was
added and the mixture was stirred overnight at 55 °C. After solvent
elimination under diminished pressure, the residue purified by sil-
ica gel chromatography (7:2:1 AcOEt–MeOH–water) gave 2b
(39 mg, 84%) as white solid. Rf 0.49 (eluent above). Calcd for
C18H25NO13: C, 46.65; H, 5.40. Found: C, 46.80; H, 5.26. 1H and
13C NMR parameters (see Tables 3 and 4) were identical to those
previously reported.26
3.7.2. 2-Biphenylmethyl b-
D
-glucopyranoside 1e
To a soln of peracetylated 1e (236 mg, 0.46 mmol) in MeOH
(5 mL) was added concentrated ammonia (36%, 0.11 mL). The reac-
tion was allowed to proceed at room temperature for 6 h. Then, the
solvent was removed by distillation under diminished pressure,
and ether was added in order to precipitate 1e, which was obtained
as white crystals (105 mg, 0.33 mmol). mp 173–175 °C, ½a D20
ꢀ7.8
ꢁ
(c 3.0, water), Anal. Calcd for C17H20O6: C, 63.75; H, 6.25. Found:
C, 63.69; H, 6.35. The NMR data of 1e are given in Tables 3 and 4.
3.12. 4-Cyanophenyl b-
glucopyranoside 2c and 4-cyanophenyl b-
(1?3)-b- -glucopyranoside 4c
D
-galactopyranosyl-(1?3)-b-
D-
3.8. Synthesis of phenyl 1-thio-b-D-glucopyranoside 1f
D-glucopyranosyl-
D
The synthesis of 1f was achieved via deacetylation of commer-
cial (Aldrich) phenyl 2,3,4,6-tetra-O-acetyl-1-thio-b- -glucopyran-
oside in the presence of catalytic amounts of NaOMe in MeOH.23
D
a-
D-Galactopyranosyl or
a-
D
-glucopyranosyl fluoride (34 mg,
-glucopyranoside 1c (28 mg,
0.2 mmol) and 4-cyanophenyl b-
D
0.1 mmol) were dissolved in hydrogencarbonate buffer
(150 mmol/L, pH 7.8). Then, glycosynthase E338G sol. (0.9 mL),
prepared as previously described,15 was added and the reaction
was allowed to proceed at 55 °C until complete consumption of
the fluoride (12 h). After removal of the solvent under diminished
pressure, purification by silica gel chromatography (5:2:2 BuOH–
EtOH–water, Rf of 2c 0.55 and Rf of 4c 0.64) afforded 38 mg of pure
3.9. Mixture of benzyl b-
D-galactopyranosyl-(1?3 and 1?6)-b-
D
-glucopyranosides 2d and 3d
a
-
D
-Galactopyranosyl fluoride (36 mg, 0.2 mmol, prepared
according to Malet and Planas12) and benzyl b-
D
-glucopyranoside
1d (27 mg, 0.1 mmol) were dissolved in hydrogencarbonate buffer
(150 mmol/L, pH 7.8). Then, glycosynthase E338G sol. (0.9 mL),
prepared as previously described,15 was added and the reaction
was allowed to proceed at 55 °C until the complete consumption
of the fluoride (12 h). After removal of the solvent under dimin-
ished pressure, purification by silica gel chromatography (5:2:2
BuOH–EtOH–water, Rf 0.52) afforded a mixture of disaccharides
2d and 3d, 55% and 45%, respectively, 29 mg, overall yield 67%.
Anal. (mixture of regioisomers) Calcd for C19H28O11: C, 52.78; H,
6.48. Found: C, 52.69; H, 6.59). Some fractions enriched in disac-
charide 2d allowed the NMR structural identification of this
2c (yield 85 %, mp 162–163 °C, ½a D20
ꢀ34.3 (c 2.8, water). Calcd for
ꢁ
C19H25NO11: C, 51.47; H, 5.64. Found: C, 51.67; H, 5.49) and 42 mg
of pure 4c (yield 95%, mp 166–168 °C, ½a D20
ꢀ24.8, c 2.4, water.
ꢁ
Calcd for C19H25NO11: C, 51.47; H, 5.64. Found: C, 51.39; H, 5.71).
3.13. 2-Biphenylmethyl b-D-galactopyranosyl-(1?3)-b-D-
glucopyranoside 2e
a-
D-Galactosyl fluoride (83 mg, 0.46 mmol) and 2-biphenylm-
ethyl b- -glucopyranoside 1e (91 mg, 0.27 mmol) were dissolved
D
compound (see Tables 3 and 4). The structure of benzyl b-
D-Galp-
in ammonium hydrogencarbonate buffer (50 mmol/L, pH 7.8,
8 mL). Then, E338G glycosynthase sol. (3 mL), prepared as previ-
ously described,15 was added. The reaction was allowed to proceed
at 55 °C for 6.5 h. At this time, the Gal-F had completely disap-
peared. After elimination of the solvent under diminished pressure,
flash silica gel chromatography of the mixture (28.3:10:1:1.7:4.1
CHCl3–MeOH–AcOH–water–petroleum ether) gave fractions con-
taining a mixture of two regioisomers as revealed by the 1H NMR
spectrum (80 mg, 86% and 14%, respectively, overall yield 70%). A
further purification by silica gel chromatography (same eluent as
above) afforded first, fractions containing the minor disaccharide
(Rf 0.75) as a major compound, which was identified as the b-
(1?6) regioisomer 3e on the basis of the 13C NMR spectrum of
the mixture. Further fractions contained only the b-(1?3) regio-
isomer 2e (Rf 0.62, identified on the basis of the 1H and 13C NMR
spectra, see Tables 3 and 4). After distillation of the solvents under
diminished pressure and washing the residue with CH2Cl2
(3 ꢂ 10 mL) in order to remove traces of acetic acid, 2e (38 mg)
(1?6)-b-
D
-Glcp 3d was determined on the basis of the 13C NMR
spectrum of the mixture of regioisomers: in the typical
60–65 ppm zone for the CH2OH resonances, three peaks were pres-
ent, two of them with equal intensity were easily attributed to 2d
while the single third one belonged to 3d, thus confirming the 1?6
linkage.
3.10. Mixture of benzyl b-
glucopyranosides 4d and 5d
D
-glucopyranosyl-(1?3 and 1?6)-b-
D
-
a-
D-Glucopyranosyl fluoride (36 mg, 0.2 mmol) and benzyl b-
D
-
glucopyranoside 1d (27 mg, 0.1 mmol) were dissolved in
hydrogencarbonate buffer (150 mmol/L, pH 7.8). Then, 0.9 mL of
glycosynthase E338G sol., prepared as previously described,15
was added and the reaction was allowed to proceed at 55 °C until
complete consumption of the fluoride (12 h). After removal of the
solvent under diminished pressure, purification by silica gel chro-
matography (5:2:2 BuOH–EtOH–water, Rf 0.50) afforded a mixture
of disaccharides 4d and 5d (56% and 44%, respectively, 37 mg over-
was obtained as a syrup (½a D20
ꢀ34.3 (c 2.8, water). Anal. Calcd
ꢁ
for C23H30O11: C, 57.26; H, 6.22. Found: C, 57.17; H, 6.31).
all yield 82%, Anal. (mixture of regioisomers) Calcd for C19H28O11
:
C, 52.78; H, 6.48. Found: C, 52.61; H, 6.61. Some fractions enriched
in disaccharide 4d allowed the NMR structural identification of this
compound. The 1H and 13C NMR spectra were consistent with the
Glc(1?3)-Glc structure of 4d (see Tables 3 and 4).The structure of
3.14. Synthesis of phenyl b-D-galactopyranosyl-(1?3)-1-thio-b-
D-glucopyranoside 2f
a-
D-Galactosyl fluoride (150 mg, 0.8 mmol) and phenyl 1-thio-
benzyl b-D-Glcp(1?6)-b-D-Glcp 5d was established on the same
b-
D-glucopyranoside 1f (220 mg, 0.8 mmol) were dissolved in
basis as already described for the structure determination of 3d
(see above).
20 mL of ammonium hydrogencarbonate buffer (150 mmol/L, pH
7.8). Then 9 mL of the E338G glycosynthase sol., prepared as

2946
Z. Marton et al. / Carbohydrate Research 343 (2008) 2939–2946
previously described,15 was added. The reaction was allowed to
proceed at 55 °C for 7 h while stirring. At this time, the Gal-F had
completely disappeared. Then, the solvent was removed under
diminished pressure, and the products were purified by silica gel
flash chromatography (9:1 BuOH–water). The disaccharide frac-
tions were collected and after removal of the solvent, 281 mg of
a mixture of the two regioisomers were obtained (96% and 4%,
respectively, as determined on the basis of the 1H NMR spectra,
overall yield 81%). A further purification performed with silica
gel chromatography using the same eluent as above yielded
205 mg of pure 2f (Rf 0.46) and 58 mg of a mixture of 2f and 3f
(Rf of 3f 0.60). Anal. Calcd for C18H26O10S (2f): C, 49.77; H, 5.99.
Found: C, 49.91; H, 5.81. The 1H and 13C NMR spectra of 2f were
EtOAc–MeOH–water) was obtained nearly quantitatively (yield
94 %, 32 mg, Anal. Calcd for C12H22O11: C, 42.11; H, 6.43. Found:
C, 42.02; H, 6.32). Comparison with the literature data for b-
D-
Galp-(1?3)-
D
-Glcp28 allowed the structure elucidation of 6. The
assignments of the NMR resonances of 6 are given in Tables 3
and 4.
Acknowledgments
The authors thank the French Ministry of Education and Re-
search, the Centre National de la Recherche Scientifique, and the
Region des Pays de la Loire (VANAM Program) for financially sup-
porting this work.
identical as those previously described for phenyl b-D-Galp-
(1?3)-1-thio-b-D
-Glcp.27 The structure of 3f was determined on
References
the basis of the 13C NMR spectrum of the mixture of 2f and 3f
(CH2OH resonances).
1. Mackenzie, L. F.; Wang, Q.; Warren, R. A. J.; Withers, S. G. J. Am. Chem. Soc. 1998,
120, 5583–5584.
2. Jahn, M.; Stoll, D.; Warren, R. A. J.; Szabo, L.; Singh, P.; Gilbert Harry, J.; Ducros
Valerie, M. A.; Davies Gideon, J.; Withers Stephen, G. Chem. Commun. 2003,
1327–1329.
3. Jahn, M.; Chen, H.; Muellegger, J.; Marles, J.; Warren, R. A. J.; Withers, S. G.
Chem. Commun. 2004, 274–275.
4. Mayer, C.; Zechel, D. L.; Reid, S. P.; Warren, R. A.; Withers, S. G. FEBS Lett. 2000,
466, 40–44.
5. Tolborg, J. F.; Petersen, L.; Jensen, K. J.; Mayer, C.; Jakeman, D. L.; Warren, R. A.
J.; Withers, S. G. J. Org. Chem. 2002, 67, 4143–4149.
6. Viladot, J.-L.; Canals, F.; Batllori, X.; Planas, A. Biochem. J. 2001, 355, 79–86.
7. Nashiru, O.; Zechel, D. L.; Stoll, D.; Mohammadzadeh, T.; Warren, R. A. J.;
Withers, S. G. Angew. Chem., Int. Ed. 2001, 40, 417–420.
8. Faijes, M.; Fairweather, J. K.; Driguez, H.; Planas, A. Chem. Eur. J. 2001, 7, 4651–
4655.
9. Faijes, M.; Perez, X.; Perez, O.; Planas, A. Biochemistry 2003, 42, 13304–13318.
10. Faijes, M.; Imai, T.; Bulone, V.; Planas, A. Biochem. J. 2004, 380, 635–641.
11. Fairweather, J. K.; Faijes, M.; Driguez, H.; Planas, A. ChemBioChem 2002, 3, 866–
873.
12. Malet, C.; Planas, A. FEBS Lett. 1998, 440, 208–212.
13. Hrmova, M.; Imai, T.; Rutten, S. J.; Fairweather, J. K.; Pelosi, L.; Bulone, V.;
Driguez, H.; Fincher, G. B. J. Biol. Chem. 2002, 277, 30102–30111.
14. Drone, J.; Feng, H.-y.; Tellier, C.; Hoffmann, L.; Tran, V.; Rabiller, C.; Dion, M.
Eur. J. Org. Chem. 2005, 1977–1983.
15. Dion, M.; Fourage, L.; Hallet, J.-N.; Colas, B. Glycoconjugate J. 1999, 16, 27–37.
16. Feng, H.-Y.; Drone, J.; Hoffmann, L.; Tran, V.; Tellier, C.; Rabiller, C.; Dion, M. J.
Biol. Chem. 2005, 280, 37088–37097.
3.15. Hydrogenolysis of benzyl and of 2-biphenylmethyl
disaccharides
(a) Synthesis of mixtures of b-
1?6)- -glucopyranoside 6 and 7 by hydrogenolysis of ben-
zyl b- -Galp-(1?3 and 1?6)-b- -Glcp 2d and 3d.
D-galactopyranosyl-(1?3 and
D
D
D
The mixture of disaccharides 2d and 3d (43.2 mg, 0.1 mmol,
55% and 45%, respectively) was dissolved in 10 mL of MeOH.
The catalyst (Pd/C, 50 mg) and H2 were introduced in the
flask. The mixture was stirred at rt for 2 h. After filtration
of the catalyst and elimination of the solvent under dimin-
ished pressure,
a white solid was obtained (33 mg,
0.096 mmol, yield 96%). The 1H NMR spectrum revealed
the presence of the two b-(1?3) b-(1?6) regioisomers 6
and 7 (relative percentages 55:45). Comparison with the lit-
erature data for b-D-Galp-(1?3)-D
-Glcp28 and with the main
product obtained from the hydrolysis of 2f (see below)
allowed the structure elucidation. The attribution of the
NMR resonances of 6 is given in Tables 3 and 4.
(b) Synthesis of b-
6 by hydrogenolysis of 2-biphenylmethyl b-
-Glcp 2e. The same procedure as above was used, but with a
8 h reaction time. Thus, hydrogenation of 2e (24 mg,
0.50 mmol) produced the free disaccharide (16 mg,
D
-galactopyranosyl-(1?3)-
D
-glucopyranoside
17. Osanjo, G.; Dion, M.; Drone, J.; Solleux, C.; Tran, V.; Rabiller, C.; Tellier, C.
Biochemistry 2007, 46, 1022–1033.
18. Fukuyama, T.; Laird, A. A.; Hotchkiss, L. M. Tetrahedron Lett. 1985, 26, 6291–
6292.
19. Lairson, L. L.; Watts, A. G.; Wakarchuk, W. W.; Withers, S. G. Nat. Chem. Biol.
2006, 2, 724–728.
D
-Galp-(1?3)-b-
D
6
20. Withers, S. G.; Jahn, M. Engineered glycosidases with thioglycoligase activity
and their use for synthesis of thioglycosides. In PCT Int. Appl., 2004, 36 pp.
CODEN PDXXD2 WO 2004024908 A1 20040325 CAN 140:283382 AN 2004
252631 CAPLUS.
0.47 mmol, yield 94%). The NMR spectra were identical to
those described below (see Section 3.16).
21. Andre, G.; Buleon, A.; Juy, M.; Aghajari, N.; Haser, R.; Tran, V. Biopolymers 1999,
49, 107–119.
3.16. Synthesis of b-D-galactopyranosyl-(1?3)-D-
glucopyranoside 6 by hydrolysis of phenyl b-
D
-
22. Meyer, B.; Stuike-Prill, R. J. Org. Chem. 1990, 55, 902–906.
23. Lindhorst, T. K. Essentials of Carbohydrate Chemistry and Biochemistry; Wiley-
VCH Verlag GmbH: Wienheim (Germany), 2000.
24. Smits, E.; Engberts, J. B. F. N.; Kellogg, R. M.; van Doren, H. A. J. Chem. Soc., Perkin
Trans. 1 1996, 2873–2877.
galactopyranosyl-(1?3)-1-thio-b-D-glucopyranoside
2f in the presence of NBS.29
25. Nath, R. L.; Rydon, H. N. Biochem. J. 1954, 57, 1–10.
26. Murata, T.; Akimoto, S.; Horimoto, M.; Usui, T. Biosci., Biotechnol., Biochem.
1997, 61, 1118–1120.
27. Prade, H.; Mackenzie, L. F.; Withers, S. G. Carbohydr. Res. 1998, 305, 371–381.
28. Petzelbauer, I.; Reiter, A.; Splechtna, B.; Kosma, P.; Nidetzky, B. Eur. J. Biochem.
2000, 267, 5055–5066.
The b-(1?3) disaccharide 2f (43.4 mg, 0.1 mmol) was dissolved
in water (8 mL). After addition of 2 equiv of NBS (71.2 mg,
0.4 mmol), a white precipitate appeared. After redissolving of the
later, the reaction was allowed to proceed for 30 min at rt. Then,
water was eliminated under diminished pressure, and the
disaccharide purified using silica gel chromatography (7:2:1
29. Motavia, M. S.; Marcussen, J. Carbohydr. Chem. 1995, 14, 1279–1294.