Thermus thermophilus glycosynthases for the efficient synthesis of galactosyl and glucosyl β-(1→3)-glycosides

Article abstract of DOI:10.1002/ejoc.200500014

Inverting mutant glycosynthases were designed according to the Withers strategy, starting from wild-type Thermus thermophilus retaining Tt-β-Gly glycosidase. Directed mutagenesis of catalytic nucleophile glutamate 338 by alanine, serine, and glycine afforded the E338A, E338S, and E338G mutant enzymes, respectively. As was to be expected, the mutants were unable to catalyze the hydrolysis of the transglycosidation products. In agreement with previous results, the E338S and E338G catalysts were much more efficient than E338A. Moreover, our results showed that these enzymes were inactive in the hydrolysis of the α-D-glycopyranosyl fluorides used as donors, and so suitable experimental conditions, under which the rate of spontaneous hydrolysis of the donor was considerably lower than that of enzymatic transglycosidation, provided galactosyl and glucosyl β-(1→3)-glycosides in yields of up to 90%. The structure of native Tt-β-Gly available in the Protein Data Bank offers a good basis for interpretation of our results by means of molecular modeling. Thus, in the case of the E338S mutant, a lower energy of the system was obtained when the donor and the acceptor were in the right position to form the β-(1→3)-glycosidic bond. Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim, Germany, 2005.

Full text of DOI:10.1002/ejoc.200500014

FULL PAPER
Thermus thermophilus Glycosynthases for the Efficient Synthesis of Galactosyl
and Glucosyl β-(1Ǟ3)-Glycosides
Jullien Drone,[a] Hui-yong Feng,[b] Charles Tellier,[a] Lionel Hoffmann,[a] Vinh Tran,[a]
Claude Rabiller,[a] and Michel Dion*[a]
Keywords: Glycosynthase / Transglycosidation / Thermophilic β-glycosidase / Thermus thermophilus / β-(1Ǟ3)-Disaccha-
rides / Molecular modeling
Inverting mutant glycosynthases were designed according to
the Withers strategy, starting from wild-type Thermus
thermophilus retaining Tt-β-Gly glycosidase. Directed muta-
genesis of catalytic nucleophile glutamate 338 by alanine,
serine, and glycine afforded the E338A, E338S, and E338G
mutant enzymes, respectively. As was to be expected, the
mutants were unable to catalyze the hydrolysis of the trans-
glycosidation products. In agreement with previous results,
the E338S and E338G catalysts were much more efficient
than E338A. Moreover, our results showed that these en-
conditions, under which the rate of spontaneous hydrolysis
of the donor was considerably lower than that of enzymatic
transglycosidation, provided galactosyl and glucosyl β-
(1Ǟ3)-glycosides in yields of up to 90%. The structure of
native Tt-β-Gly available in the Protein Data Bank offers a
good basis for interpretation of our results by means of mol-
ecular modeling. Thus, in the case of the E338S mutant, a
lower energy of the system was obtained when the donor
and the acceptor were in the right position to form the β-
(1Ǟ3)-glycosidic bond.
zymes were inactive in the hydrolysis of the α-
anosyl fluorides used as donors, and so suitable experimental
D
-glycopyr-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
functioned as an acid/base and the other as a nucleophile to
create the ester glycosyl enzyme intermediate (glycosylation
step). As shown in Scheme 1, this activated complex was
then able to react with the nucleophiles present in the reac-
tion medium/water (hydrolysis) and the hydroxy groups of
carbohydrates (transglycosylation). As pointed out by
Withers, the similar transition states (oxocarbenium ions)
involved either in inversion or in retention of glycosidases
offered the potential to convert the latter into the former
by appropriate mutations. Thus, a mutant of a retaining exo
β-glucosidase from Agrobacterium, devoid of any action of
E358 as a nucleophile (E358A mutant), showed a high
transglycosidation activity with use of α-glucosyl fluoride
as a donor in the presence of various acceptors. Good
transglycosidation yields were obtained, since this “nucleo-
phile-free” enzyme was unable to hydrolyze the reaction
product.[7] A similar approach developed at the same time
by Planas et al. on an endo β-glucanase also came to the
same conclusions.[8–12] The “glycosynthase” concept was
born (see Scheme 1), and the development of the idea was
to result in the design of efficient mutants for the synthesis
of oligosaccharides.[13–22] Improvements in the catalytic
properties of glycosynthases from Agrobacterium sp. were
also recently obtained by directed evolution.[23]
Introduction
A number of carbohydrate ligands play important roles
in many biological mechanisms. Their use as drugs and, to
a degree, understanding of their action depend on the avail-
ability of the oligosaccharides involved in the processes, and
so the transfer activity exerted by inexpensive retaining gly-
cosidases has provided an efficient approach for the synthe-
sis of saccharides. One of the drawbacks of this method was
the incomplete regioselectivity, which resulted in mixtures
of regioisomers not easy to separate. Another obstacle in-
volved the hydrolysis of the substrate and of the products,
which remained in competition with transglycosidation.
Moreover, when working with an exo-glycosidase, two types
of glycosides were usually synthesized: those resulting from
the reaction between donor and acceptor and those pro-
duced from the self-condensation of the donor.[1–3] An im-
pressive advancement of the method was achieved after the
elucidation of the mechanism of retaining β-glucosidases by
Withers et al.[4–6] A pair of carboxylic acids acting as cata-
lytic residues were shown to be involved in the process. One
[a] Biotechnologie, Biocatalyse, Biorégulation (UMR CNRS
6204), Faculté des Sciences et des Techniques,
2, rue de la Houssinière, BP 92208, 44322 Nantes, cedex 3,
France,
In our continuing efforts to provide new glycosidases for
the selective synthesis of oligosaccharides, we have cloned
and overexpressed (in E. coli) a β-glycosidase (Tt-β-Gly)
from the thermophilic species Thermus thermophilus.[24,25]
Fax: +33-51-12-56-37
E-mail: michel.dion@univ-nantes.fr
[b] Institute of Biosciences and Biotechnology, University of Sci-
ence and Technology,
Shijiazhuang, Hebei, China
Eur. J. Org. Chem. 2005, 1977–1983
DOI: 10.1002/ejoc.200500014
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1977

J. Drone, H. Feng, C. Tellier, L. Hoffmann, V. Tran, C. Rabiller, M. Dion
FULL PAPER
Scheme 1. Transglycosidation mechanism: A) with retaining glycosidases, and B) with glycosynthases.
This enzyme catalyzed the synthesis of β-(1Ǟ3)-disaccha- fraction from the heated cells was used as a biocatalyst
rides by transglycosidation in reasonable yields (Ϸ 40–50%) without any further purification. Kinetic studies[24] have
with use of o-nitrophenyl β-d-Galp (or Glcp) as a glycosyl shown that the hydrolytic activity with p-nitrophenyl glyco-
donor.[1] However, the enzyme mainly catalyzed the self- sides as substrates is higher for β-d-fucoside than for β-d-
condensation reaction and consequently the yield of the glucoside and β-d-galactoside (the respective values for kcat
/
condensation disaccharide remained relatively moderate. Km are 247, 227, and 14.8 mm–1 s–1).
The aim of this paper is to evaluate mutants, according to
the “glycosynthase ” concept, capable of synthesizing oligo-
saccharides in higher yields. Since only a few glycosidases
exhibit such regioselectivity, an enzyme capable of ef-
ficiently catalyzing the synthesis of the β-(1Ǟ3)-glycosidic
bonds would be of great interest in view of the biological
importance of the saccharides incorporating such struc-
tures. Such a glycosidic bond is present in the Thomsen–
Friedenreich antigen determinant β-Galp-(1Ǟ3)-α-Galp-
NAc-O–Ser and in the antigen H type 1 (l-Fucp-α-(1Ǟ2)-
d-Galp β-(1Ǟ3)-d-Glcp-NAc.[26] In addition, the d-Glcp-β-
(1Ǟ3)-d-Glcp disaccharide is also known to act as an im-
munosuppressant.[27] The Glc-β-(1Ǟ3)-d-Glc linkage is in-
volved in the structure of hexasaccharides acting as elicitors
of the defense response in plants,[28] particularly in to-
bacco.[29]
Although the native enzyme showed good transglycosid-
ation activity, the possibility of improving this property by
site-directed mutagenesis still remained open. More particu-
larly, according to the findings of Withers and Planas,[7,12]
replacement of catalytic nucleophile glutamate 338 by ala-
nine, serine, or glycine should provide glycosidases devoid
of their hydrolytic activity towards the products of the reac-
tions. Furthermore, the inverting mutants thus obtained are
able to accept α-d-activated donors, such as the readily
available α-d-glycopyranosyl fluorides (e.g., α-GlcF and α-
GalF). Moreover, the use of such substrates generally
avoids the concurrent self-condensation of the donor.
Mutants E338A, E338G, and E338S were therefore pro-
duced, and their transglycosidation potentials were tested
with α-GlcF or α-GalF as donors and different acceptors.
Preliminary experiments in the presence of p-nitrophenyl α-
d-glycopyranosides indicated that these donors were not
recognized as substrates for the mutant enzymes, which in
addition did not catalyze the self-condensation of α-glyco-
syl fluorides. This property is probably due to a steric clash
of the axial fluorine at site +2. Furthermore, water was not
accepted as a nucleophile by the three mutants (see
Scheme 2). This rather surprising, but very interesting,
property is highlighted in Figure 1 for the case of the
E338G mutant.
Results and Discussion
The gene tt-β-gly, encoding a β-glycosidase, has been
cloned from Thermus thermophilus and overexpressed into
E. coli.[25] The amount of this enzyme can reach up to 50%
of the total protein fraction in the cells. After lysis of the
cell membranes, simple heating at 70 °C destroys all the gly-
The great similarity between the kinetic hydrolysis curves
cosidasic activities normally present in the mesophilic spe- obtained in the presence or in the absence of the enzyme
cies, apart from that of Tt-β-Gly. In this work, the soluble implied that this mutant had lost its ability to accept water
1978
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1977–1983

Efficient Synthesis of β-(1Ǟ3)-Glycosides
FULL PAPER
Scheme 2. Reactions catalyzed by Tt-β-Gly glycosynthase mutants.
tion could be enhanced relative to that of spontaneous hy-
drolysis of the fluoride. Thus, in a first set of experiments,
the potentials of the three mutants were tested under condi-
tions allowing quantitative comparisons (pH = 7.8 buffer
solutions containing 50 mmolL–1 of glycosyl fluoride as a
donor, 50 mmolL–1 of an acceptor, and 4.1 mg of enzyme
per mmol of donor, see Table 1).
The yields were determined by proton NMR spectra in-
tegration measurements on the crude reaction mixtures ob-
tained at the time at which the donor was almost consumed.
The separation of the components was achieved by silica
gel liquid chromatography and the structures of the prod-
ucts were determined by standard one- and two-dimen-
sional NMR proton and carbon spectra. Whatever the do-
nor and the mutant used, the only disaccharides produced
were of the β-(1Ǟ3) type. Attempts to obtain trisaccharides
by use of double amounts of donor were unsuccessful, di-
saccharides always being synthesized as single products.
Thus, on comparison with native Tt-β-Gly, which also gave
the β-(1Ǟ6)-regioisomer, the three mutants looked com-
Figure 1. Comparison between spontaneous and enzymatically cat-
alyzed hydrolysis of α-glucosyl fluoride {experimental conditions:
E338G glycosynthase,
T = 55 °C, pH = 7.8, [α-GlcF] =
50 mmolL–1}.
as a nucleophile. The same was true for the two other mu-
tants (data not shown).
This property should be usable to improve the transgly-
cosidation yields if the rate of the transglycosidation reac- pletely regioselective. Meanwhile, trisaccharide synthesis
Table 1. Transglycosidation yields (%) obtained with Thermus thermophilus mutant glycosynthases for a donor/acceptor ratio = 1.
E338A
E338S
E338G
Entry
Donor
Acceptor
β-1,3
β-1,x
β-1,3
β-1,x
β-1,3
β-1,x
1
2
3
4
5
6
7
8
α-GlcF
β-d-GlcPh
β-d-GlcPh[c]
β-d-GalPh
β-d-FucpNP
β-CellpNP
β-CellpNP[c]
Cell
46
77
10
0
0
0
0
0
0
86
0
89
96
0
0
77
39
31
0
0
0
37[a]
30
31
35[a]
52[a]
0
0
0
0
50[b]
37[b]
Lac
9
β-d-GlcMe
β-d-GlcPh
β-d-GalPh
β-CellpNP
Mal
10
86
18
0
0
0
20
0
10
11
12
13
α-GalF
13
0
0
0
0
0
0
0
[a] Mixture of β-(1Ǟ6)-regioisomer as a major compound (Ͼ90%) and another unknown compound. [b] Complex mixture of α,β-anom-
eric regioisomers. [c] A three times higher enzyme concentration (relative to the other experiments) was used.
Eur. J. Org. Chem. 2005, 1977–1983
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1979

J. Drone, H. Feng, C. Tellier, L. Hoffmann, V. Tran, C. Rabiller, M. Dion
FULL PAPER
was achieved by use of β-(1Ǟ4)-disaccharides as acceptors, β-d-N-acetylglucopyranoside in good yields from α-GlcF
but the regioselectivity was not so good and several re- and pNP–GlcNAc.[7]
gioisomers were obtained. This was the case, for instance,
In order to provide an explanation of this behavior, we
with p-nitrophenyl β-d-cellobioside (β-CellpNP), from used molecular modeling techniques. Since the X-ray struc-
which equivalent amounts of p-nitrophenyl β-d-glucopyr- ture of the native Tt-β-Gly was available (PDB code: lug6),
anosyl-(1Ǟ3)-cellobioside and its (1Ǟ6)-regioisomer were we performed in silico mutation of glutamate 338 with a
present, in addition to a small amount of a third unknown serine residue to obtain a potentially valid model for the
species. The experimental results collected in Table 1 also E338S mutant enzyme, since only one residue had been re-
indicate that the E338A mutant, although providing single placed, on the assumption that the local environment had
β-(1Ǟ3)-disaccharides (in moderate yields) displays a rather been preserved. All possible side chain orientations for the
narrow tolerance for acceptors. This result, in agreement serine residue were tested and that of lowest energy was
with previous experiments,[13] can be explained by hydro- kept for further calculations. We then docked the ligands
phobicity induced by the methyl of alanine 338. Conversely, into the active site. For this process, we used a three-dimen-
mutation by serine is known to afford much more efficient sional structure that is a covalent intermediate between the
mutants, although the free space between the anomeric car- β-glycosidase of Bacillus polymyxa (PDB code: le4i) and 2-
bon of a substrate and the hydroxymethyl group of serine deoxy-2-fluoroglucose. Because of the close similarity be-
is reduced. The best results with such mutants have been tween these two enzymes (both belonging to Glycoside Hy-
explained in terms of a favorable interaction between the drolase Family 1), superimposition of the complex with the
CH2OH group and the fluorine of the glycosyl fluoride, E338S mutant model, followed by superimposition of α-
which could occur in the glycosylation transition state, re- GlcF with its 2-deoxy-2-fluoroglucose analogue in the com-
sulting in an enhanced transglycosylation reaction rate. This plex, provided a good starting localization for the α-GlcF.
trend was also observed for the Tt-β-Gly E338S mutant, for This position was then optimized by a molecular mechanics
which yields of up to 80% were observed with β-d-GlcPh scheme. In a second step, we placed the second ligand, β-
or β-d-GalPh as acceptors (Table 1, Entries 1 and 3). More- d-GlcPh, roughly inside the E338S mutant model active site
over, this enzyme catalyzed the transglycosidation reaction according to the supposed transglycosidation reaction
with a broad spectrum of acceptors, such as fucosides or β- mechanism. We then performed systematic manual transla-
(1Ǟ4)-disaccharides (Table 1). It was noteworthy that Tt-β- tions and rotations according to the interaction energy be-
Gly mutants behaved similarly to Withers Abg glycosynth- tween β-d-GlcPh and the E338S mutant model. The best
ases. The transglycosidation reaction rates induced by the position found for β-d-GlcPh was that shown in Figure 2.
alanine mutant were always lower than those of the serine According to this result, the need for a phenyl group could
and glycine counterparts. Our preliminary observations
thus suggested the conclusion that Tt-β-Gly glycosynthases
did not catalyze the hydrolysis of α-GlcF or α-GalF. As a
result, we carried out a second series of experiments with
three times the amount of enzyme (12.3 mg of enzyme per
mmol of donor). The results shown in Table 1 clearly indi-
cate an enhancement of the transglycosidation yields. This
was particularly the case with the E338A mutant, for which
a yield of 77% (instead of 46%, Table 1, Entry 2) was ob-
tained with β-d-GlcPh as an acceptor. Similar results were
also observed for the transglycosidation yields in the pres-
ence of β-CellpNP with E338G glycosynthase as a catalyst.
This mutant also produced a nearly quantitative yield (96%
of phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-glucopyranoside)
with the most favorable β-d-GlcPh acceptor (Table 1, En-
try 1). It should be highlighted that these mutants also need
acceptors bearing an aromatic aglycon group (see Table 1,
Entries 1 and 9). Thus, in the presence of the E338S mutant,
the replacement of β-d-GlcPh by β-d-GlcMe resulted in a
dramatic drop in the transglycosylation yield from 77% to
10%. Similarly, free disaccharides such as lactose were not
good acceptors for the mutants. Maltose, an α-(1Ǟ4)-disac-
Figure 2. Molecular modeling, in the active site, of the disposition
charide, was the worst acceptor, with no transglycosylation
product present in the reaction medium. Moreover, pNP–
GlcNAc, although possessing an aromatic aglycon group,
was not recognized by the E338S mutant. This behavior
also showed a significant difference from that of Abg glyco-
of reactants before the transglycosidation reaction catalyzed by
E338S glycosynthase, with α-GlcF as donor and β-GlcPh as ac-
ceptor. An energy minimum was obtained when stacking occurred
between the phenyl group of β-GlcPh and tryptophan 312. More-
over, the relative dispositions of the reactants, stabilized by several
hydrogen bonds, were in agreement with the formation of a β-
synthase E358S, which produced d-glucopyranosyl-(1Ǟ4)- (1Ǟ3) glycosidic bond.
1980 © 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim www.eurjoc.org
Eur. J. Org. Chem. 2005, 1977–1983

Efficient Synthesis of β-(1Ǟ3)-Glycosides
FULL PAPER
correlations). The spectra were recorded with a Bruker AX 500
spectrometer operating at 500 MHz for 1H and 126 MHz for 13C
or on a Bruker AX 400 spectrometer operating at 400 MHz for 1H
and 100.6 MHz for 13C. In all cases, chemical shifts in ppm are
quoted from the resonance of methyl protons of sodium 3-(trimeth-
ylsilyl)propanesulfonate (DSS) used as an internal reference.
be explained by the important stacking between this phenyl
group and the tryptophan 312 residue, which was mainly
responsible for the energy stabilization. Moreover, the rela-
tive positions of the two monosaccharides stabilized by se-
veral hydrogen bonds are consistent with the complete β-
(1Ǟ3)-regioselectivity observed for this reaction. In con-
trast, when the acceptors are cellobiose or lactose the site
+2 is occupied by a glucosyl unit. As a result, a lower con-
formational constraint is probably imposed on the orienta-
tion of the disaccharide within the active site by Trp 312.
This may explain the relatively low regioselectivity observed
with these acceptors.
Determination of Yields by Proton NMR Spectroscopy
Use of E338S and E338G Glycosynthases: The donor (30 μmol) and
the acceptor (30 μmol) were dissolved in ammonium/hydrogen car-
bonate buffer in D2O (pH = 7.8, 510 μL, 150 mmolL–1) and the
mutant solution, prepared as described above (90 μL, 125 μg of
enzyme, 100 mmolL–1 phosphate buffer in H2O, pH = 7.0), was
added. The reactions were allowed to proceed at 55 °C until the
donor had completely disappeared. The reaction mixture was then
subjected to proton NMR spectroscopy. Experiments with three-
fold amounts of enzymatic solutions were carried out similarly, but
the reaction mixture was evaporated under reduced pressure and
redissolved in D2O before NMR analysis.
Conclusion
In summary, we have shown that three Tt-β-Gly glyco-
synthases, designed according to the strategy developed by
Withers, circumvented the drawbacks of wild-type retaining Use of E338A Glycosynthase: The donor (30 μmol) and the ac-
ceptor (30 μmol) were dissolved in ammonium hydrogen carbonate
buffer in D2O (pH = 7.8, 550 μL of 150 mmolL–1), and the mutant
solution, prepared as described above (50 μL, 125 μg of enzyme),
was added. The reactions were allowed to proceed at 55 °C until
the donor had completely disappeared. The reaction mixture was
glycosidases in terms of their transglycosidation ability,
since:
a) no hydrolytic activity towards the product and the sub-
strate was detectable,
b) the regioselectivity was almost total, at least for the syn-
thesis of β-(1Ǟ3)-disaccharides, and
c) no donor self-condensation products were synthesized.
Optimization of the reaction to produce yields of up to
90% was also possible through the use of appropriate
amounts of enzymes to enhance transglycosidation rates
relative to those of spontaneous hydrolysis of the glycosyl
fluorides. An interpretation of the properties of the mutant
glycosynthases based on molecular modeling studies has
also been proposed. Moreover, directed evolution of
Thermus thermophilus glycosidase is currently under investi-
gation in order to increase its catalytic activities and expand
its substrate repertoire.
1
then subjected to H NMR spectroscopy. Experiments with three-
fold amounts of enzymatic solutions were carried out similarly but
the reaction mixture was evaporated under reduced pressure and
redissolved in D2O before NMR analysis.
Kinetics of Hydrolysis of α-Glycosyl Fluorides: Experiments were
performed directly in the NMR tube at 55 °C. 1H NMR spectra
were recorded each hour for 9 h.
Enzymatic Hydrolysis: α-GlcF or -GalF (30 μmol) was dissolved in
ammonium hydrogen carbonate buffer in D2O (pH = 7.8, 510 μL
of 150 mmolL–1). The glycosynthase E338G solution, prepared as
described above (90 μL, 125 μg of enzyme), was then added.
Spontaneous Hydrolysis: α-GlcF or -GalF (30 μmol) was dissolved
in ammonium hydrogen carbonate buffer in D2O (pH = 7.8,
600 μL, 150 mmolL–1).
Experimental Section
General Optimized Procedure for the Synthesis of β-(1Ǟ3) Disac-
charides with Use of the Tt-β-Gly E338S Mutant: α-d-Glucopyr-
anosyl (or -galactopyranosyl) fluoride (18.2 mg, 0.1 mmol) and an
appropriate acceptor (0.1 mmol) were dissolved in ammonium car-
bonate buffer (pH = 7.8, 2 mL, 150 mmolL–1). The Tt-β-Gly
E338S mutant (0.126 mg, 0.9 mL), prepared as described above,
was then added and the mixture was stirred overnight at 55 °C.
After solvent removal under vacuum, the residue was purified by
flash chromatography to afford the corresponding disaccharides.
General: Directed mutagenesis of gene ttβGly was carried out by
PCR with overlapping extension, and cloning was done into the
pBtac2 vector at sites EcoRI-PstI. The primers Nphi1/Nphi2 (5Ј-
gtcacggcgaacggggccgcctac; 5Ј-gtaggcggccccgttcgccgtgac), Nphi3/
Nphi4 (5Ј-gtcacgtctaacggggccgcctac, 5Ј-gtaggcggccccgttagacgtgac),
and
Nphi5/Nphi6
(5Ј-gtcacgggtaacggggccgcctac,
5Јgtaggcggccccgttacccgtgac) were used to generate the mutant en-
zymes E338A, E338S, and E338G, respectively. Overexpression of
the different β-glycosidases was performed in E. coli strain XL1
Blue MRFЈ. After overnight cultivation in LB medium, cells were
harvested, lysed, and heated for 1 h at 70 °C to remove E. coli ther-
molabile proteins. Supernatants of centrifuged extracts were used
as biocatalysts in transglycosylation reactions.
Phenyl β-D-Glucopyranosyl-(1Ǟ3)-β-D-glucopyranoside: Flash
chromatography (AcOEt/MeOH/water, 17:2:1) gave the disaccha-
ride (33 mg, 75%) as a white solid. M.p. 77 °C. [α]2D0 = –39.1 (c =
1.067, H2O). Rf 0.62 (AcOEt/MeOH/water, 7:2:1). HRMS (CI/glyc-
erol) calcd. for C18H27O11 [M + H]+: 419.1475, found 419.1578.
NMR: see Table 2, Table 3, and Table 4.
Chemicals supplied by Aldrich were used without further purifica-
tion. The courses of the reactions were followed by TLC
(Merck F254 precoated silica gel 60 sheets) and proton NMR spec-
troscopy. The components of the reaction mixtures were separated
by silica gel chromatography. Complete analysis of the NMR 1H
and 13C resonances and subsequent structure assignments were car-
ried out by use of standard 2D sequences (COSY HH and HC
Phenyl β-D-Glucopyranosyl-(1Ǟ3)-β-D-galactopyranoside: Flash
chromatography (AcOEt/MeOH/water, 15:2:1) gave the disaccha-
ride (31 mg, 70%) as a white solid. M.p. 142 °C. [α]2D0 = +6.5 (c =
0.467, H2O). Rf = 0.51 (AcOEt/MeOH/water, 7:2:1). HRMS (CI/
glycerol) calcd. for C18H27O11 [M + H]+: 419.1475; found 419.1604.
NMR: see Table 5, Table 6, and Table 7.
Eur. J. Org. Chem. 2005, 1977–1983
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1981

J. Drone, H. Feng, C. Tellier, L. Hoffmann, V. Tran, C. Rabiller, M. Dion
FULL PAPER
Table 2. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-glucopyranoside:
1H NMR (500 MHz, D2O), chemical shifts (δ in ppm).
Table 8. Phenyl β-d-galactopyranosyl-(1Ǟ3)-β-d-glucopyranoside:
1H NMR (500 MHz, D2O), chemical shifts (δ in ppm).
H-1
H-2
H-3
H-4
H-5
H-6A H-6B
H-1
H-2
H-3
H-4
H-5
H-6A H-6B
I
II
5.16
4.78
3.78
3.39
3.86
3.54
3.62
3.43
3.65
3.50
3.77
3.73
3.93
3.93
I
II
5.16
4.70
3.77
3.63
3.87
3.69
3.62
3.93
365
3.73
3.77
nd
3.93
3.80
δ(aromatic) = 7.15 and 7.41 ppm
δ(aromatic) = 7.15 and 7.41 ppm
Table 9. Phenyl β-d-galactopyranosyl-(1Ǟ3)-β-d-glucopyranoside:
1H-1H coupling constants (Hz, solvent D2O).
Table 3. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-glucopyranoside:
1H-1H coupling constants (Hz, solvent D2O).
J1,2
J2,3
J3,4
J4,5
J5,6A
J5,6B
J6A,6B
I
II
7.9
7.7
8.4
9.9
9.0
3.4
9.9
0.8
5.3
3.8
2.2
8.1
–12.4
–11.7
J1,2
J2,3
J3,4
J4,5
J5,6A
J5,6B J6A,6B
I
II
7.9
8.0
9.2
9.4
8.4
9.1
9.9
9.8
5.1
6.1
2.0
2.2
–12.4
–12.3
Table 10. Phenyl β-d-galactopyranosyl-(1Ǟ3)-β-d-glucopyrano-
side: 13C NMR (100.8 MHz, D2O), chemical shifts (δ in ppm).
C-1
C-2
C-3
C-4
C-5
C-6
Table 4. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-glucopyranoside:
13C NMR (126 MHz, D2O), chemical shifts (δ in ppm).
I
II
102.7
106.0
75.3
74.0
87.0
75.3
70.8
71.3
78.4
78.0
63.3
63.8
C-1
C-2
C-3
C-4
C-5
C-6
δ(aromatic) = 119.3, 126.1, 132.7 and 159.2 ppm
I
II
102.7
105.5
75.4
76.2
86.9
78.3
70.7
72.3
78.4
78.7
63.2
63.4
4-Nitrophenyl β-D-Glucopyranosyl-(1Ǟ3)-β-D-fucopyranoside: Flash
δ(aromatic) = 119.3, 126.1, 132.7 and 159.2 ppm
chromatography (AcOEt/MeOH/water, 15:2:1) gave the disaccha-
ride (13 mg, 31%) as a white solid. M.p. 80 °C. [α]2D0 = –32.9 (c =
0.480, H2O). Rf = 0.61 (AcOEt/MeOH/water, 7:2:1). NMR: see
Table 11, Table 12, and Table 13.
Table 5. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-galactopyranoside:
1H NMR (400 MHz, D2O), chemical shifts (δ in ppm).
Table 11. 4-Nitrophenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-fucopyr-
anoside: 1H NMR (400 MHz, [D5]pyridine), chemical shifts (δ in
ppm).
H-1
H-2
H-3
H-4
H-5
3.87 3.71–3.82 3.71–3.82
3.47 3.71–3.82 3.90
H-6A
H-6B
I
II
5.11
4.71
3.98
3.40
3.92
3.52
4.25
3.43
H-1
H-2
H-3
H-4
H-5
H-6A
H-6B
δ(aromatic) = 7.15 and 7.41 ppm
1.58
(CH3)
4.44
I
II
5.61
5.52
4.89
4.13
4.43
4.30
4.27
4.41
4.10
4.08
–
4.61
δ(aromatic) = 7.33 and 8.25 ppm
Table 6. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-galactopyranoside:
1H-1H coupling constants (Hz, solvent D2O)
Table 12. 4-Nitrophenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-fucopyr-
1
anoside: H-1H coupling constants (Hz, solvent [D5]pyridine).
J1,2
J2,3
J3,4
J4,5
J5,6A
J5,6B
J6A,6B
J1,2
J2,3
J3,4
J4,5
J5,6A
J5,6B
J6A,6B
I
II
7.6
7.9
9.8
9.4
3.2
9.7
0.9
9.7
5.5
2.3
6.4
5.0
n.d.
n.d.
I
II
7.7
7.8
8.2
8.0
n.d.
8.7
n.d.
n.d.
6.4 (CH3)
n.d.
–
2.5
–
–11.6
Table 13. 4-Nitrophenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-fucopyr-
anoside: 13C NMR (100.8 MHz, [D5]pyridine), chemical shifts (δ
in ppm).
Table 7. Phenyl β-d-glucopyranosyl-(1Ǟ3)-β-d-galactopyranoside:
13C NMR (100.8 MHz, D2O), chemical shifts (δ in ppm).
C-1
C-2
C-3
C-4
C-5
C-6
C-1
C-2
C-3
C-4
C-5
C-6
I
II
103.2
106.4
72.4
76.0
84.8
78.3
70.9
78.5
77.7
72.2
63.4
63.3
I
II
103.4
108.1
72.2
77.5
85.9
80.1
73.5
80.1
73.7
80.3
18.6
64.5
δ(aromatic) = 119.3, 126.1, 132.7 and 159.2 ppm
δ(aromatic) = 118.6, 127.7, 144.4 and 164.9 ppm
Phenyl β-
D
-Galactopyranosyl-(1Ǟ3)-β-
D
-glucopyranoside: Flash Acknowledgments
chromatography (AcOEt/MeOH/water, 17:2:1) gave the disaccha-
ride (30 mg, 68%) as white needles. M.p. 134 °C. [α]2D0 = –30.5 (c =
Thanks are due to the French Ministry of Education and Research
0.967, H2O). Rf = 0.49 (AcOEt/MeOH/water, 7:2:1). HRMS (CI/ for a grant to one of us (J. D.) and to the CNRS for financially
glycerol) calcd. for C18H27O11 [M]+: 419.1475; found 419,1557.
NMR: see Table 8, Table 9, and Table 10.
supporting this work. D. Maume (Ecole Vétérinaire de Nantes) is
gratefully acknowledged for mass spectra experiments.
1982
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 1977–1983

Efficient Synthesis of β-(1Ǟ3)-Glycosides
FULL PAPER
[17] M. Jahn, H. Chen, J. Muellegger, J. Marles, R. A. J. Warren,
S. G. Withers, Chem. Commun. 2004, 274–275.
[18] M. Moracci, A. Trincone, G. Perugino, M. Ciaramella, M.
Rossi, Biochemistry 1998, 37, 17262–70.
[19] J.-L. Viladot, F. Canals, X. Batllori, A. Planas, Biochem. J.
2001, 355, 79–86.
[20] H. Lin, H. Tao, V. W. Cornish, J. Am. Chem. Soc. 2004, 126,
15051–15059.
[21] O. Nashiru, D. L. Zechel, D. Stoll, T. Mohammadzadeh,
R. A. J. Warren, S. J. Withers, Angew. Chem. Int. Ed. 2001, 40,
417–420.
[22] M. Jahn, D. Stoll, R. A. J. Warren, L. Szabo, P. Singh, H. J.
Gilbert, V. M.-A. Ducros, G. J. Davies, S. J. Withers, Chem.
Commun. 2003, 1327–1329.
[23] S. G. Withers, M. Jahn, in PCT Int. Appl., Wo, 2004, p. 36.
[24] L. Fourage, M. Dion, B. Colas, Glycoconjugate J. 2000, 17,
377–383.
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
V. Chiffoleau-Giraud, P. Spangenberg, M. Dion, C. Rabiller,
Eur. J. Org. Chem. 1999, 757–763.
P. Spangenberg, V. Chiffoleau-Giraud, C. Andre, M. Dion, C.
Rabiller, Tetrahedron: Asymmetry 1999, 10, 2905–2912.
P. Spangenberg, C. Andre, V. Langlois, M. Dion, C. Rabiller,
Carbohydr. Res. 2002, 337, 221–228.
S. G. Withers, K. Rupitz, D. Trimbur, R. A. J. Warren, Bio-
chemistry 1992, 31, 9979–9985.
J. C. Gebler, D. E. Trimbur, R. A. J. Warren, R. Aebersold, M.
Namchuk, S. G. Withers, Biochemistry 1995, 34, 14547–14553.
Q. Wang, R. W. Graham, D. Trimbur, R. A. J. Warren, S. G.
Withers, J. Am. Chem. Soc. 1994, 116, 11594–5.
L. F. Mackenzie, Q. Wang, R. A. J. Warren, S. G. Withers, J.
Am. Chem. Soc. 1998, 120, 5583–5584.
M. Faijes, J. K. Fairweather, H. Driguez, A. Planas, Chem.
Eur. J. 2001, 7, 4651–4655.
[9] M. Faijes, X. Perez, O. Perez, A. Planas, Biochemistry 2003, 42,
13304–18.
[25] M. Dion, L. Fourage, J.-N. Hallet, B. Colas, Glycoconjugate J.
1999, 16, 27–37.
[26] K. Fukase, T. Yasukochi, Y. Suda, M. Yoshida, S. Kusumoto,
Tetrahedron Lett. 1996, 37, 6763–6766.
[10] M. Faijes, T. Imai, V. Bulone, A. Planas, Biochem. J. 2004, 380,
635–641.
[11] J. K. Fairweather, M. Faijes, H. Driguez, A. Planas, ChemBi-
oChem 2002, 3, 866–873.
[12] C. Malet, A. Planas, FEBS Lett. 1998, 440, 208–212.
[13] C. Mayer, D. L. Zechel, S. P. Reid, R. A. Warren, S. G. Withers,
FEBS Lett. 2000, 466, 40–44.
[14] J. F. Tolborg, L. Petersen, K. J. Jensen, C. Mayer, D. L. Jake-
man, R. A. J. Warren, S. G. Withers, J. Org. Chem. 2002, 67,
4143–4149.
[15] G. Perugino, A. Trincone, M. Rossi, M. Moracci, Trends Bi-
otechnol. 2004, 22, 31–37.
[27] Y. Yamamoto, S. Murosaki, H. Kusaka, in: Jpn. Kokai Tokkyo
Koho (Takeda Shokuhin Kogyo K. K., Japan), 1997, p. 13.
[28] P. Albersheim, A. G. Darvill, M. McNeil, B. S. Valent, J. K.
Sharp, E. A. Nothnagel, K. R. Davis, N. Yamazaki, D. J. Gol-
lin, W. S. York, W. F. Dudman, J. E. Darvill, A. Dell, Oligosac-
charins: naturally occurring carbohydrates with biological regu-
latory functions, in Structure and Function of Plant Genomes
(Eds.: O. Ciferri, L. Dure, III), Plenum, New York, 1983,
pp. 293–312.
[29] O. Klarzynski, B. Plesse, J.-M. Joubert, J.-C. Yvin, M. Kopp,
B. Kloareg, B. Fritig, Plant Physiol. 2000, 124, 1027–1037.
Received: January 10, 2005
[16] M. Jahn, J. Marles, R. A. J. Warren, S. G. Withers, Angew.
Chem. Int. Ed. 2003, 42, 352–354.
Eur. J. Org. Chem. 2005, 1977–1983
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1983