An original chemoenzymatic route for the synthesis of β-D- galactofuranosides using an α-l-arabinofuranosidase

Article abstract of DOI:10.1016/j.carres.2005.01.016

Galactofuranose is a widespread component of cell wall polysaccharides in bacteria, protozoa and fungi, but is totally absent in mammals. Importantly, galactofuranose is a key constituent of major cell envelope polysaccharides in pathogenic mycobacteria.

Full text of DOI:10.1016/j.carres.2005.01.016

Carbohydrate
RESEARCH
Carbohydrate Research 340 (2005) 637–644
An original chemoenzymatic route for the synthesis of
b-D-galactofuranosides using an a-L-arabinofuranosidase
a
Caroline Remond, Richard Plantier-Royon, Nathalie Aubry and
b
a
´
Michael J.O ꢀDonohuea,*
aLaboratoire de Technologie Enzymatique et Physico-chimie des Agroressources, UMR URCA/INRA FARE, 8, rue Gabriel Voisin,
BP 316, F-51688 Reims, France
b
´ ´ ´
Laboratoire Reactions Selectives et Applications, UMR URCA/CNRS 6519, Faculte des Sciences, BP 1039, F-51687 Reims, France
Received 20 December 2004; accepted 20 January 2005
Abstract—Galactofuranose is a widespread component of cell wall polysaccharides in bacteria, protozoa and fungi, but is totally
absent in mammals.Importantly, galactofuranose is a key constituent of major cell envelope polysaccharides in pathogenic mycobac-
teria.In this respect, galactofuranose-based glycoconjugates are interesting target molecules for drug design. O-Glycosidases and nota-
bly b-D-galactofuranosidases could be useful tools for the chemoenzymatic synthesis of galactofuranosides, but to date no studies of
this type have been reported.Here we report the use of a GH 51 a-L-arabinofuranosidase for the synthesis of b-D-galactofuranosides.
We have demonstrated that this enzyme can catalyse both the autocondensation of p-nitrophenyl-b-D-galactofuranoside and the trans-
galactofuranosylation of benzyl a-D-xylopyranoside, forming p-nitrophenyl b-D-galactofuranosyl-(1!2)-b-D-galactofuranoside and
benzyl b-D-galactofuranosyl-(1!2)-a-D-xylopyranoside, respectively.Both reactions were very regiospecific and the reaction involving
benzyl a-D-xylopyranoside afforded very high yields (74.8%) of the major product. To our knowledge, this demonstration of chemo-
enzymatic synthesis of galactofuranosides constitutes the very first use of an O-glycosidase for the synthesis of galactofuranosides.
Ó 2005 Elsevier Ltd. All rights reserved.
Keywords: a-L-Arabinofuranosidase; Transglycosylation; O-Glycosidase; GH 51; Galactofuranose; Mycobacteria
1. Introduction
of cell wall glycoconjugates of bacteria,2 protozoa3 and
fungi.4 Importantly, b-D-galactofuranose and/or a-D-
arabinofuranose are cell wall constituents in several
pathogenic species.These include Mycobacterium tuber-
culosis and leprae, the causative agents of tuberculosis
and leprosy, respectively, and Trypanosama cruzi, the
protozoan responsible for Chagaꢀs disease.
In nature, arabinose is frequently present in the furanose
form.Notably, a-L-arabinofuranose is abundant in
plant cell walls,1 whereas a-D-arabinofuranose is charac-
teristic of the cell wall glycoconjugates of several pro-
karyotes.While being absent in plants and mammals,
galactose in the furanose form is also widespread in
nature. b-D-Galactofuranose is an abundant constituent
Over the last decade or so, the alarming progression
of multidrug-resistant mycobacterial strains (MDRTB)
around the world has been the subject of major public
concern.5 MDRTB strains are defined as strains that
are resistant to at least isoniazid and rifampicin,
two first-line drugs currently used for the treatment of
tuberculosis.6 The former, isoniazid, and another
first-line drug ethambutol affect the biosynthesis of the
Abbreviations: AbfD3, a-L-arabinofuranosidase D3; pNP, p-nitrophe-
nol; pNPGalf, p-nitrophenyl b-D-galactofuranoside; pNPAraf, p-
nitrophenyl a-L-arabinofuranoside; pNPXylp, p-nitrophenyl-b-D-xylo-
pyranoside; pNP b-D-Glcf, p-nitrophenyl-b-D-glucofuranoside; pNP
b-D-Xylf, p-nitrophenyl-b-D-xylofuranoside; pNP b-D-Galp, p-nitro-
phenyl-b-D-galactopyranoside; PhCH2 a-D-Xylp, benzyl a-D-xylopyr-
anoside; PhCH2 b-L-Arap, benzyl b-L-arabinopyranoside.
arabinogalactan–peptidoglycan complex (MAPc),7–9
a
major component of the mycobacterial cell envelope.
MAPc is composed of a galactan polymer, which is
*
Corresponding author.Tel:. +33 326 355 365; fax: +33 326 355 369;
e-mail: michael.odonohue@univ-reims.fr
0008-6215/$ - see front matter Ó 2005 Elsevier Ltd.All rights reserved.
doi:10.1016/j.carres.2005.01.016

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
638
made of D-galactofuranose monomers, which are inter-
linked via (1!5) and (1!6) glycosidic bonds, and an
arabinan polymer that is mainly composed of (1!5)-
linked D-arabinofuranose residues.The arabinan part
of MAPc is itself linked to mycolic acids.A second
major cell envelope polysaccharide is lipoarabinomannan
(LAM) that also contains arabinose and galactose in
the furanose form.This polymer, localised in the cell
membrane, extends out towards and contacts MAPc.10
Because of the absence of galactofuranose in humans,
galactofuranose-based conjugates are both highly
immunogenic and resistant to endogenous human
glycosidases.Therefore, MAPc and LAM-associated
galactans constitute attractive targets for new drug
development.To this end, two strategies can be envis-
aged: (i) the synthesis of galactofuranose analogues that
could act as inhibitors of the biosynthetic enzymes
UDP-galactopyranose mutase11 and galactofuranosyl
transferases12 or (ii) the synthesis of galactofuranosides
that could be used for the elaboration of vaccines.With
regard to the inhibition of mycobacterial cell wall
biosynthesis, the production of b-D-galactofuranosyl-
(1!4)-a-L-rhamnopyranosyl octyl analogues has been
described.13 Although these compounds proved to be
effective acceptors for mycobacterial glycosyl transfer-
ases, in vitro inhibition tests revealed that they only
moderately inhibited the growth of two mycobacterial
strains.13 So far, no reports of new galactofuranose-
based vaccines have been reported.However, it is
known that cell-surface bacterial polysaccharides often
elicit immune responses in humans.Therefore, it is
expected that certain mycobacterial cell wall polysaccha-
rides, or motifs thereof, could constitute effective vac-
cines.Indeed, to combat other pathogenic bacteria
such as Streptococcus pneumoniae type 23F14 or
Haemophilus influenzae type b,15 the synthesis and use
of oligosaccharide motifs for vaccine development
has already proved successful.
To obtain relevant oligosaccharides for drug design it
is necessary to identify an appropriate strategy.A
straightforward approach consists of the extraction via
chemical and/or enzymatic means and purification of
bacterial polysaccharides.A second widely used
approach is the chemical synthesis of specific oligosac-
charides.Although several reports have described or-
ganic syntheses of galactofuranose-based compounds,
generally speaking, the chemical synthesis of oligosac-
charides remains a challenge.13,16–21 This is because
appropriate, reactive and anomerically-pure monomers
must first be prepared and then linked together in a ste-
reo- and regiospecific fashion via glycosidic bonds.
Increasingly, to simplify some or all of the synthetic
steps for the production of oligosaccharides, enzymatic
or chemoenzymatic strategies are being adopted.22,23
Generally, by default, O-glycosidases and glycosyl trans-
ferases catalyse the modification of glycosidic bonds in a
stereospecific manner and, in addition, are often regio-
specific.Therefore, the use of such enzymes for synthesis
can eliminate lengthy protection/deprotection proce-
dures and drastically reduce both the time required
and the use of organic solvents and other undesirable
chemicals.23,24
Glycosyl transferases (EC 3.2.4) catalyse regioselec-
tive glycosylation and afford high-transfer yields.25,26
However, the high costs of the nucleotide–sugar sub-
strates and, until recently, the low availability of these
enzymes are often perceived as limiting factors.In addi-
tion to hydrolysis, many so-called ꢁretainingꢀ O-glycosi-
dases (EC 3.2.1) also possess the ability to catalyse the
formation of oligosaccharides.Generally, glycosidase-
mediated synthesis is performed either by exploiting
the ability of these enzymes to perform reverse hydroly-
sis (thermodynamic control) or by employing activated
substrates, which by generating a steady-state concen-
tration of the glycosyl-enzyme intermediate, exert a
kinetic control over the transglycosylation reaction.
Most frequently, the activated substrates are p-nitro-
phenyl glycosides and the enzymes used are either gluco-
sidases or galactosidases.Intriguingly, to date, although
b-D-galactofuranosidase (EC 3.2.1.146) activities have
been identified,27,28 no genes have been isolated and
expressed and no transgalactofuranosylation activities
have been associated with these enzymes.
Until now, most studies describing the enzymatic pro-
duction of oligosaccharides have been concerned with
the synthesis of oligopyranosides.However, recently
two reports, one from our group29 and the other from
Sakamoto et al.,30 have demonstrated that both an arab-
inofuranosidase and an arabinanase can perform trans-
glycosylation in the presence of an appropriate donor
sugar.With regards to our study, we showed that the
a-L-arabinofuranosidase from Thermobacillus xylanily-
ticus (AbfD3) is able to catalyse the autocondensation
of a-L-arabinofuranose or b-D-xylopyranose, as well as
the synthesis of benzyl a-L-arabinofuranosyl-(1!2)-a-
D-xylopyranoside.Here, taking advantage of the similar
spatial orientation of the C-2, C-3 and C-4–OH groups
in a-L-arabinofuranose and b-D-galactofuranose, we
have explored the ability of AbfD3 to perform transga-
lactofuranosylation using p-nitrophenyl b-D-galactofur-
anoside (pNPGalf) as the donor sugar.
2. Results and discussion
2.1. Synthesis of p-nitrophenyl b-D-galactofuranoside
Despite several attempts, we were unable to perform
the synthesis of pNPGalf according to the previously
described protocol.31 Therefore, the intermediate
1,2,3,5,6-tetra-O-acetyl-D-galactofuranose was prepared
according to a more recent procedure as described in

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
639
Scheme 1. Synthesis of the p-nitrophenyl b-D-galactofuranoside donor.
Scheme 1.32,33 Introduction of the p-nitrophenyl group
onto the anomeric position was achieved with reason-
able yield via a classical method using p-nitrophenol
and BF3ÆEt2O in dichloromethane.Deprotection of
the acetyl groups was carried out using sodium meth-
anolate 1 M solution in methanol as described.31 The
overall yield for the synthesis from D-galactose was
16%.
major product was purified for further analysis by
NMR.Analysis of the Rf values of the minor products
led us to speculate that these were trisaccharides, but
further verification was not performed due to the low
availability of these compounds.Analysis of the 13C
NMR data for the major product revealed a downfield
shift of the C-2 of the D-galactofuranoside moiety from
81.7 (for pNP-Galf) to 88.5 ppm. Moreover, the 13C–1H
(HMBC) spectrum showed an intense three-bond
coupling between C-10 (108.4 ppm) and H-2, clearly
indicating that this compound was p-nitrophenyl
b-D-galactofuranosyl-(1!2)-b-D-galactofuranoside (1)
(Scheme 2).Importantly, no correlations were observed
between C-10 and H-3, H-5 or H-6, confirming the pres-
ence, within the major product fraction, of only one
(1!2)-linked regioisomer.Therefore the AbfD3-cataly-
sed autocondensation of pNPGalf was quite regioselec-
tive.Interestingly, we have previously shown that the
autocondensation of pNPAraf or pNPXylp occurs with
only moderate regioselectivity (60–70% of p-nitrophenyl
a-L-arabinofuranosyl-(1!2)-a-L-arabinofuranoside and
50% each of p-nitrophenyl b-D-xylopyranosyl-(1!2)-b-
D-xylopyranoside and p-nitrophenyl b-D-xylopyrano-
syl-(1!3)-b-D-xylopyranoside),29 although a (1!2)-
linked compound was always present among the reac-
tion products.
2.2. Hydrolytic activity of AbfD3 on p-nitrophenyl
b-D-galactofuranoside
In order to investigate the specificity of AbfD3 towards
p-nitrophenyl b-D-galactofuranoside (pNPGalf), this
compound was used as a substrate for hydrolysis.
Specific activity measured in the presence of pNPGalf
(0.64 IU/mg) was much lower than that measured
for p-nitrophenyl a-L-arabinofuranoside (pNPAraf,
465 IU/mg), but was similar to that measured for
p-nitrophenyl b-D-xylopyranoside (pNPXylp, 0.292 IU/
mg).Attempts to determine the kinetic parameters
for AbfD3-catalysed hydrolysis of pNPGalf were not
pursued because KM values were estimated to be extre-
mely high (>50 mM) and hence incompatible with the
solubility and availability of the substrate.This indi-
cates that the presence of the extra hydroxymethyl
group (C-6) is extremely detrimental for substrate bind-
ing in the À1 (donor) subsite.In a previous study,
despite being a poor substrate for AbfD3-catalysed
hydrolysis, pNPXylp was shown to act as both donor
and acceptor in AbfD3-mediated transglycosylation
reactions.29 Therefore, the low but measurable activity
on pNPGalf was taken to be an encouraging indication
of the ability of this sugar to act as a donor for
transglycosylation.
2.3. Synthesis of galactofuranobioside with AbfD3
For initial tests, pNPGalf was incubated with AbfD3
and the reactions products were examined by TLC.
One major and two minor products were detected.The
Scheme 2.

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
640
2.4. Transgalactofuranosylation of various acceptors
using pNPGalf
90
80
70
60
50
40
30
20
10
0
As well as pNPAraf and pNPXylp, several other mono-
saccharide derivatives (p-nitrophenyl a-D-xylopyrano-
side, pNP a-D-Xylp; p-nitrophenyl b-D-xylofuranoside,
pNP b-D-Xylf; p-nitrophenyl b-D-glucofuranoside, pNP
b-D-Glcf; p-nitrophenyl a-L-arabinopyranoside, pNP a-
L-Arap; benzyl a-D-xylopyranoside, PhCH2 a-D-Xylp;
benzyl b-L-arabinopyranoside, PhCH2 b-L-Arap; p-
nitrophenyl b-D-galactopyranoside, pNP b-D-Galp)
were tested for their ability to act as acceptors for
AbfD3-catalysed transgalactofuranosylation.Control
reactions, in which pNPGalf was omitted, were per-
formed in order to monitor autocondensation reactions.
The products of all the reactions were examined by TLC
analysis.Most of the monosaccharide derivatives tested
were not subject to autocondensation and could there-
fore only act as acceptors for transgalactofuranosyl-
ation.In the majority of cases, one reaction product
was detected, although in reactions involving pNP
b-D-Xylf or pNP a-D-Xylp, two and three products,
respectively, were formed.Overall, despite the fact that
poor yields precluded further structural analysis of all
but one of the products, it is reasonable to suppose that
they were all galactofuranose-containing disaccharides.
These results indicate that the specificity in the acceptor
(+1) subsite of AbfD3 is rather large.In the light of
structural data available for another GH 51 arabino-
furanosidase from Geobacillus stearothermophilus (GS-
Abf)34 such permissiveness is expected.The crystal
structure of this enzyme complexed with Abf Araf-
a(1!3)-Xylp revealed that the xylose residue at the +1
subsite formed only one hydrogen bond, compared to
the multiple hydrogen bonding at the À1 subsite.Unlike
all of the other monosaccharide derivatives tested,
PhCH2 a-D-Xylp led to the appearance of one very in-
tense spot on TLC.Being sufficiently abundant, this
product was therefore purified for analysis (see below).
Like (1), the purification of this product was character-
ised by a disappointingly low yield (15%).In the case of
pNPAraf and pNPXylp, autocondensation was the dom-
inant reaction and so the major products formed were
arabinobiosides and xylobiosides, respectively.
0
100
200
300
400
500
Time (min)
Figure 1. Kinetics of the formation of p-nitrophenyl b-D-galactofur-
anosyl-(1!2)-b-D-galactofuranoside (m) and benzyl a-D-xylopyranos-
yl-(1!2)-b-D-galactofuranoside (h) obtained through AbfD3-
catalysed transglycosylation reactions using pNPGalf as the donor
sugar.Yields are calculated as a percentage of the initial quantity of
pNPGalf.
Time-course analysis revealed that AbfD3-catalysed
autocondensation of pNPGalf was relatively slow, since
maximum disaccharide yield was obtained after 2 h (Fig.
1).In comparison, in the presence of pNPAraf and
pNPXylp, maximum yields (8.3% for p-nitrophenyl
a-L-arabinofuranosyl-(1!2)-a-L-arabinofuranoside and
11.8% for p-nitrophenyl b-D-xylopyranosyl-(1!2 or
1!3)-b-D-xylopyranosides) were achieved after 30 s
and 10 min, respectively (Table 1).29 After reaching
maximum yield (20.8%), secondary hydrolysis was
responsible for the progressive disappearance of (1).
No disaccharide was detectable after 8 h.Overall, these
results reflect the poor binding (high KM) of galactofura-
nose in the À1 (donor) subsite of AbfD3 and reveal that
the catalytic efficiency (kcat/KM) of AbfD3-medi-
ated hydrolysis of (1) is lower than that of the homo-
disaccharides p-nitrophenyl b-D-xylopyranosyl-(1!2 or
1!3)-b-D-xylopyranosides and p-nitrophenyl a-L-arabi-
nofuranosyl-(1!2)-a-L-arabinofuranoside.Despite
reaching 20.8% yield during transglycosylation, the final
yield of compound (1) after purification was only 6%.
Table 1. Maximal yields for the AbfD3-catalysed synthesis of disac-
charides in presence of various acceptor and donor molecules at 60 °C
Donor
Acceptor Maximal Linkages References
formed
2.5. Structural characterisation of a galactofuranosyl-
xylopyranoside
yields
pNPGalf pNPGalf 20.8
b-(1!2)
b-(1!2)
This study
This study
pNPGalf PhCH2
a-D-Xylp
pNPAraf PhCH2
74.8
7
Analysis of the 13C NMR data for the product of the
reaction involving pNPGalf and PhCH2 a-D-Xylp re-
vealed a downfield shift of C-2 of the D-xylopyrano-
side moiety from 73.5 (for PhCH2-Xylp) to 78.4 ppm.
Moreover, a 13C–1H (HMBC) spectrum showed an in-
tense three-bond coupling between C-10 (109.0 ppm)
and H-2, while revealing no correlation between C-10
and H-3 or H-4.Overall, these data show that the
29
29
´
Remond et al.
a-(1!2)
a-D-Xylp
pNPAraf pNPAraf
´
Remond et al.
8.3
a-(1!2)
a-(1!3)
a-(1!5)
b-(1!2)
b-(1!3)
pNPXylp pNPXylp 11.8
Re´mond et al.29
Yields are calculated as a percentage of the initial quantity of donor.

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
641
The chemoenzymatic method uses an a-L
-arabinofura-
nosidase and the experimental rationale is based upon
the fact that b-D-Galf is a C-5 hydroxymethylated ana-
logue of a-L-Araf.
Galactofuranosyl-containing oligosaccharides are
potentially useful compounds for the development of
new vaccines or chemotherapeutic strategies for the pre-
vention or treatment of tuberculosis and other serious
pathologies.In mycobacterial cell walls, galactofurano-
syl units are linked to each other via b-(1!5) or b-
(1!6) bonds and to rhamnose through b-(1!4) bonds.
In some fungal cell walls D-Galf is interlinked via b-
(1!2) and b-(1!3) bonds39 and elsewhere in nature it
can be linked to other sugar moieties such as Man or
Glc via b-(1!3) bonds.40–43 Although very regiospecific,
the transglycosylation reactions in this study disappoint-
ingly led to the formation of b-(1!2) linkages.Never-
theless, it might be interesting to test whether either of
the two b-(1!2)-linked disaccharides could inhibit en-
zymes involved in galactofuran biosynthesis in Myco-
bacteria.In this respect, the aryl groups on 1 and 2
(pNP and benzyl groups, respectively), might be useful.
In the case of n-octyl-5-(a-D-arabinofuranosyl)-b-D-
galactofuranoside derivatives, it was demonstrated that
the hydrophobic group blocking the reducing end of
disaccharides presented better biological activity against
mycobacterial arabinosyl transferases than their less
hydrophobic analogues.44
As a preliminary study, the results obtained for
AbfD3 are extremely encouraging and open several
exciting prospects for further work.Concerning the reg-
ioselectivity, potentially more useful b-(1!5) or b-
(1!6) linkages between Galf residues might be obtained
in a number of ways: (i) through enzyme engineering of
AbfD3; (ii) via a thioligase approach45 or (iii) by using
other arabinofuranosidases displaying other hydrolytic
specificities.Exploration of each of these approaches is
currently underway.
Scheme 3.
AbfD3-catalysed transgalactofuranosylation of PhCH2
a-D-Xylp occurs in a highly regioselective manner and
leads to the production of benzyl b-D-galactofuranosyl-
(1!2)-a-D-xylopyranoside (2) (Scheme 3).A time-course
analysis of this reaction revealed that, like for compound
1, maximum yield of 2 was reached after 2 h.However, in
this case transglycosylation yield was much higher
(74.8%) (Fig.1 ).Significantly, this represents the highest
yield obtained so far for any AbfD3-catalysed transgly-
cosylation reaction, being 10-fold higher than that ob-
tained for the transarabinosylation of PhCH2 a-D-Xylp
(7%) (Table 1).29 This indicates that in the presence of
the covalently linked AbfD3-Galf intermediate, PhCH2
a-D-Xylp is a much better acceptor than water.One
might speculate that this situation arises because of a
specific configuration of the donor sugar that favours
PhCH2 a-D-Xylp-mediated deglycosylation, coupled to
a favourable interaction of PhCH2 a-D-Xylp with AbfD3
in the +1 subsite.Moreover, the reaction was highly reg-
ioselective because, like in the previously described trans-
arabinosylation of PhCH2 a-D-Xylp,29 only one (1!2)-
linked product was generated.Previously, others have
suggested that there is a causal relationship between
the bond specificity of a O-glycosidase working in hydro-
lytic mode and regiospecificity during transglycosyl-
ation.35 Likewise, taken together, our findings infer that
in hydrolytic mode AbfD3 should preferentially cleave
(1!2) linkages.In our previous study, based on the find-
ings of others,36–38 we suggested that the high regioselec-
tivity of reactions involving PhCH2-Xylp might be
correlated to the a configuration of the anomeric carbon
in this sugar derivative.These present data add further
weight to this argument, but do not provide confirmation
because the equivalent b anomer was not tested.Further-
more, the fact that more products are generated when
PhCH2a–D-Xylp is replaced by pNP a-D-Xylp (three in
the former and one in the latter), implies that the aglycon
group might also influence regiospecificity.After achiev-
ing maximum yield, 2 was subject to secondary hydroly-
sis and, like 1, was totally degraded after 8 h.
Finally, taken together our present and previous29,46
data indicate that AbfD3 is a useful and quite versatile
tool for chemoenzymatic syntheses involving furano-
sides.Therefore, it might be interesting to extend the
rationale of this study towards D-fucofuranose, which
is 6-deoxygalactofuranose.
3. Experimental
3.1. Enzyme and substrates
Recombinant arabinofuranosidase (AbfD3) was pro-
duced in Escherichia coli and purified as previously
described.47 The hydrolytic activity of AbfD3 was quan-
tified after incubation of the enzyme (0.1 mL corre-
sponding to 0.1 IU) with different pNP glycosides
(0.9 mL, 5 mM in 50 mM sodium acetate, pH 5.8) at
2.6. Concluding remarks
Herein, we report the first use of an O-glycosidase for
the synthesis of galactofuranose-based disaccharides.

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
642
60 °C.Continuous release of pNP was measured at
401 nm.One unit of activity corresponds to the amount
of enzyme releasing 1 lmol of pNP per minute.
Benzyl a-D-xylopyranoside and benzyl b-L-arabino-
pyranoside were obtained by a previously described
method.48 pNP b-D-Xylp, pNP a-D-Xylp, pNP a-L-Arap
and pNP b-D-Galp were purchased from Sigma–Aldrich
(Saint-Quentin Fallavier, France). pNP a-L-Araf,29,46
pNP b-D-Xylf29 and pNP b-D-Gluf were synthesised
on a multigram scale according to the method previ-
ously described for the monosaccharides. pNP b-D-Galf
was obtained according to the synthetic approach in
Scheme 1.
a-D-Xylp.All reactions were performed in 50 mM so-
dium acetate buffer, pH 5.8 and at 60 °C.Incubation
periods were variable and reactions were stopped by en-
zyme denaturation at 100 °C for 10 min.The reaction
products were examined by TLC using Kieselgel 60
F254 aluminium-backed sheets (E.Merck) and 7:2:2
EtOAc–AcOH–water as the mobile phase.Detection
of products was achieved at 100 °C using 0.2% v/v orci-
nol in H2SO4 (20% v/v).Larger amounts of compounds
1 and 2 were prepared in an identical manner by scaling
up reactions by a factor of 40 (12 mL final volume).In
this case, the compounds were purified by flash chroma-
tography on silica gel columns (E.Merck) using 7:1
EtOAc–MeOH as the mobile phase.
3.1.1. p-Nitrophenyl b-D-glucofuranoside. White nee-
dles (35% overall yield from 1,2-O-isopropylidene-a-D-
glucofuranose, in a four-step synthesis): mp 118–
For time-course analyses, reactions were stopped after
different incubation periods by the addition of an equal
volume of veronal buffer, pH 11 (6.008 g citric acid,
3.893 g KH2PO4, 1.769 g H3BO3, 5.266 g sodium barbi-
tal in 1 L adjusted at pH 11 with NaOH).Afterwards,
product yield was monitored and quantified by
high-performance anion-exchange chromatography
(HPAEC).
22
119 °C (EtOAc); ½aꢀ À204 (c 1.00, methanol); 1H
D
NMR (CD3OD): d 8.21 (d, 2-H, J3 ,2 9.3 Hz, H-30, H-
0
0
50), 7.18 (d, 2-H, J2 ,3 9.3 Hz, H-20, H-60), 5.63 (s,
1-H, H-1), 4.34 (s, 1-H, H-2), 4.31–4.27 (m, 2-H, H-3,
H-4), 3.91 (ddd, 1-H, J5,4 8.5, J5,6b 5.9, J5,6a 3.0 Hz,
H-5), 3.70 (dd, 1-H, J6a,6b 11.5, J6a,5 3.0 Hz, H-6a),
3.54 (dd, 1-H, J6b,6a 11.5, J6b,5 5.9 Hz, H-6b); 13C
NMR (CD3OD): d 163.3 (C-10), 143.4 (C-40), 126.6 (C-
30, C-50), 117.4 (C-20, C-60), 107.8 (C-1), 84.4 (C-4),
82.5 (C-2), 76.9 (C-3), 71.4 (C-5), 65.1 (C-6). Anal.
Calcd for C12H15O8N: C, 47.84; H, 5.02; N, 4.65.
Found: C, 47.97; H, 4.99; N, 4.61.
0
0
3.2.1. p-Nitrophenyl b-D-galactofuranosyl-(1!2)-b-D-
galactofuranoside (1). 1H NMR (D2O): d 8.16 (d, 2-
H, J 9.2 Hz, H-300, H-500), 7.11 (d, 2-H, J 9.2 Hz, H-200,
H-600), 5.78 (d, 1-H, J1,2 1.9 Hz, H-1), 5.13 (d, 1-H,
J1 ,2 1.7 Hz, H-10), 4.31 (dd, 1-H, J2,3 4.4, J2,1 1.9 Hz,
0
0
H-2), 4.24 (dd, 1-H, J3,4 7.2, J3,2 4.4 Hz, H-3), 4.06
(dd, 1-H, J2 ,3 3.6, J2 ,1 1.7 Hz, H-20), 4.00 (ddd, 1-H,
0
0
0
0
0
0
0
0
3.1.2. p-Nitrophenyl b-D-galactofuranoside. White nee-
dles (16% overall yield from D-galactose, in a five-step
J4,3 7.2, J4,5 3.6 Hz, H-4), 3.97 (dd, 1-H, J3 ,4 6.2, J3 ,2
0
3.6 Hz, H-30), 3.83 (dd, 1-H, J4 ,3 6.6, J4 ,5 3.4 Hz, H-
40), 3.76 (ddd, 1-H, J5,6b 7.4, J5,6a 5.0, J5,4 3.6 Hz, H-
0
0
0
synthesis): mp 153–154 °C (EtOAc), lit.31 152–154 °C;
22
½aꢀ À205 (c 1.00, MeOH), lit.31 [a]D À203 (c 1.0,
0
0
0
0
0
0
5), 3.66 (ddd, 1-H, J5 ,6 b 7.7, J5 ,6 a 4.6, J5 ,4 3.4 Hz,
H-50), 3.56 (dd, 1-H, J6a,6b 11.8, J6a,5 5.0 Hz, H-6a),
3.53 (dd, 1-H, J6b,6a 11.8, J6b,5 7.4 Hz, H-5a), 3.29 (dd,
D
MeOH); 1H NMR (CD3OD): d 8.18 (d, 2-H, J3 ,2
0
0
9.2 Hz, H-30, H-50), 7.19 (d, 2-H, J2 ,3 9.2 Hz, H-20,
H-60), 5.64 (d, 1-H, J1,2 2.0 Hz, H-1), 4.28 (dd, 1-H,
J2,3 4.2, J2,1 2.0 Hz, H-2), 4.18 (dd, 1-H, J3,4 6.5, J3,2
4.2 Hz, H-3), 4.08 (dd, 1-H, J4,3 6.5, J4,5 3.0 Hz, H-4),
3.75 (dt, 1-H, J5,6a = J5,6b 6.4, J5,4 3.0 Hz, H-5), 3.61
(d, 2-H, J6,5 6.4 Hz, H-6a, H-6b); 13C NMR (CD3OD):
d 163.5 (C-10), 143.5 (C-40), 126.6 (C-30, C-50), 117.6 (C-
20, C-60), 107.7 (C-1), 85.5 (C-4), 83.4 (C-2), 78.1 (C-3),
72.0 (C-5), 64.2 (C-6). Anal. Calcd for C12H15O8N: C,
47.84; H, 5.02; N, 4.65. Found: C, 48.11; H, 4.89; N,
4.54.
0
0
1-H, J6 a,6 b 11.6, J6 a,5 4.6 Hz, H-60a), 3.29 (dd, 1-H,
0
0
0
0
J6 b,6 a 11.6, J6 b,5 7.7 Hz, H-60b); 13C NMR (D2O) d
161.6 (C-100), 142.6 (C-400), 126.7 (C-300, C-500), 117.3
(C-200, C-600), 108.4 (C-10), 105.1 (C-1), 88.5 (C-2), 83.5
(C-40), 83.2 (C-4), 82.0 (C-20), 77.2 (C-30), 75.4 (C-3),
70.9 (C-50), 70.5 (C-5), 63.3 (C-60), 63.2 (C-6). ESI
HRMS: m/z calcd for [C18H25NO13Na]+: 486.1224.
Found: 486.1213.
0
0
0
0
3.2.2. Benzyl b-D-galactofuranosyl-a-D-xylanopyranoside
(2). 1H NMR (D2O): d 7.45–7.36 (m, 5-H, H-aro-
matic), 5.03 (d, 1-H, J1 ,2 1.9 Hz, H-10), 4.94 (d, 1-H,
J1,2 3.7 Hz, H-1), 4.73 (d, 1-H, Jgem 11.9 Hz, H-benzyl),
4.56 (d, 1-H, Jgem 11.9 Hz, H0-benzyl), 4.09 (dd, 1-H,
0
0
3.2. Transglycosylation reactions
For analytic purposes, small amounts of compounds 1
and 2 were obtained by incubating 11 IU of AbfD3 with
5 mM pNPGalf alone (compound 1), or in combination
with an equimolar amount of PhCH2 a-D-Xylp (com-
pound 2), in a final reaction volume of 0.3 mL. Reac-
tions involving pNPGalf and other sugar acceptors
were prepared in a similar way by replacing PhCH2
J2 ,3 4.3, J2 ,1 1.9 Hz, H-20), 3.99 (dd, 1-H, J3 ,4 7.1,
J3 ,2 4.3 Hz, H-30), 3.73–3.63 (m, 4-H, H-3, H-4, H-
50, H-5a), 3.62 (dd, 1-H, J4 ,3 7.1, J4 ,5 3.3 Hz, H-40),
3.53 (t, 1-H, J5b,5a = J5b,4 10.4 Hz, H-5b), 3.50 (dd, 1-
0
H, J2,3 9.6, J2,1 3.7 Hz, H-2), 3.43 (dd, 1-H, J6 a,6 b
11.6, J6 a,5 8.0 Hz, H-60a), 3.24 (dd, 1-H, J6 b,6 a 11.6,
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
643
13
J6 b,5 3.9 Hz, H-60b); C NMR (D2O): d 136.0 (C-100),
128.9 (C-300, C-500), 128.6 (C-200, C-600), 128.3 (C-400),
109.0 (C-10), 96.3 (C-1), 82.0 (C-40), 81.1 (C-20), 78.4
(C-2), 76.2 (C-30), 71.9 (C-3), 69.8 (C-50), 69.1 (C-4),
68.7 (CH2 benzyl), 62.7 (C-5), 60.8 (C-60).ESI HRMS:
m/z calcd for [C18H26O10Na]+: 425.1424. Found:
424.1422.
0
0
6.Centers for disease control and prevention. Morbid.
Mortal. Wky. Rep. 1993, 42, 1–8.
7.Winder, F.G.In
The Biology of Mycobacteria; Ratledge,
C., Stanford, J., Eds.; Academic: London, 1982; Vol. 1, pp
417–521.
8. Mikusova, K.; Slayden, R. A.; Besra, G. S.; Brennan, P. J.
Antimicrob. Agents Chemother. 1995, 39, 2484–2489.
9. Wheeler, P.R;. Anderson, P.M.
451–457.
Biochem. J. 1996, 318,
10. Crick, D. C.; Mahapatra, S.; Brennan, P. J. Glycobiology
2001, 11, 107R–118R.
3.3. Analytical methods
11. Weston, A.; Stern, R. J.; Lee, R. E.; Nassau, P. M.;
Monsey, D.; Martin, S. L.; Scherman, M. S.; Besra, G. S.;
Duncan, K.; Mcneil, M. R. Tuber. Lung Dis. 1998, 78,
123–131.
To monitor the production of 1 and 2, HPAEC
equipped with tandem with amperometric detection
(PAD) (Dionex Corporation, CA, USA) was performed.
Products of transglycosylation were separated on a
chromatographic system consisting of a CarboPac PA1
precolumn and a CarboPac PA1 analytical column
(Dionex, USA) to which a gradient of 0.3 M sodium
acetate in 0.1 M NaOH was applied at a flow rate of
1 mL/min.
1H and 13C NMR analyses for the disaccharides were
performed in D2O at 20 °C on a Bruker DRX spectro-
meter at 500 and 125 MHz, respectively.Chemical shifts
were referenced to the signal of the methyl group of
internal acetone (1H NMR d = 2.22 ppm, 13C NMR
d = 31.5 ppm).
Mass analysis of disaccharides was performed using
electrospray-ionisation technology and Q-TOF mass
analysis (ES-Q-TOF, Micromass, UK).ESI-Q-TOF
data were acquired in positive mode ([M+Na]+ ion
detection).The cone voltage was set to 30 V and the
source temperature was 80 °C.
12. Mikusova, K.; Yagi, T.; Stern, R.; McNeil, M. R.; Besra,
J. Biol. Chem. 2000,
G.S;. Crick, D.C;. Brennan, P.J.
275, 33890–33897.
13. Pathak, A. K.; Besra, G. S.; Crick, D.; Maddry, J. A.;
Morehouse, C. B.; Suling, W. J.; Reynolds, R. C. Bioorg.
Med. Chem. 1999, 7, 2407–2413.
14. Alonso de Velasco, E.; Verheul, A. F.; van Steijn, A. M.;
Dekker, H. A.; Feldman, R. G.; Fernandez, I. M.;
Kamerling, J. P.; Vliegenthart, J. F.; Verhoef, J.; Snippe,
H. Infect. Immun. 1994, 62, 799–808.
15. Pillai, S.; Ciciriello, S.; Koster, M.; Eby, R. Infect. Immun.
1991, 59, 4371–4376.
16. van Heeswijk, W. A. R.; Visser, H. G. J.; Vliegenhart, J. F.
G. Carbohydr. Res. 1977, 59, 81–86.
17. Velty, R.; Benvegnu, T.; Gelin, M.; Privat, E.; Plusquellec,
D. Carbohydr. Res. 1997, 299, 7–14.
18. Randell, K. D.; Johnston, B. D.; Brown, P. N.; Pinto, B.
M. Carbohydr. Res. 2000, 325, 253–264.
19. Randell, K. D.; Johnston, B. D.; Lee, E. E.; Pinto, B. M.
Tetrahedron: Asymmetry 2000, 11, 207–222.
20. Randell, K. D.; Johnston, B. D.; Pinto, B. M. Carbohydr.
Res. 2000, 326, 145–150.
21. Bamhaoud, T.; Sanchez, S.; Prandi, P. Chem. Commun.
2000, 8, 660.
22.Thiem, J. FEMS Microbiol. Rev. 1995, 16, 193–211.
Acknowledgements
23.Sears, P;. Wong, C. H.
2350.
24.Wymer, N;.Toone, E.J.
110–119.
25. Crout, D.H.G;. Vic, G.
98–111.
Science 2001, 291, 2344–
Curr. Opin. Chem. Biol. 2000, 4,
Curr. Opin. Chem. Biol. 1998, 2,
We thank H.Baillia for technical help with NMR work,
D.Harakat for the ESI-TOF analyses and F.Cherouv-
rier and L.Hachadi for help with the synthesis of pNP
glycofuranosides.The present work was partially sup-
ported by the GLYCOVAL project (EUROPOLꢀAGRO-
26.Withers, S.G. Carbohydr. Polym. 2001, 44, 325–337.
27. Rietschel-berst, M.; Jentoft, N. H.; Rick, P. D.; Pletcher,
C.; Fang, F.; Gander, J. E. J. Biol. Chem. 1977, 252, 3219–
3226.
´ ´
Conseil General de la Marne).
28. Ramli, N.; Fujinaga, M.; Tabuchi, M.; Takegawa, K.;
Iwahara, S. Biosci. Biotechnol. Biochem. 1995, 59, 1856–
1860.
References
´
29.Re mond, C.; Plantier-Royon, R.; Aubry, N.; Maes, E.;
1. Joseleau, J. P.; Comtat, J.; Ruel, K. Chemical Structure of
Xylans and Their Interaction in Plant Cell Walls.In
Xylans and Xylanases; Visser, J., Beldman, G., Someren,
K.-V., Voragen, A., Eds.; Elsevier: Amsterdam, 1992; pp
1–15.
Bliard, C.; OꢀDonohue, M.J. Carbohydr. Res. 2004, 339,
2019–2025.
30. Sakamoto, T.; Fujita, T.; Kawasaki, H. Biochim. Biophys.
Acta 2004, 1674, 85–90.
31. Varela, O.; Marino, C.; de Lederkremer, R. M. Carbo-
hydr. Res. 1986, 155, 247–251.
2.Brennan, P.J;.Nikaido, H.
29–63.
Ann. Rev. Biochem. 1995, 64,
`
32.Ferrie res, V.; Gelin, M.; Boulch, R.; Toupet, L.; Plus-
quellec, D. Carbohydr. Res. 1998, 314, 79–83.
3.De Lederkremer, R.M;.Colli, W.
547.
Glycobiology 1995, 5,
`
33.Ferrie res, V.; Bertho, J.-N.; Plusquellec, D. Carbohydr.
Res. 1998, 311, 25–35.
4. Notermans, S.; Veeneman, G. H.; Van Zuylen, C. W. E.
M.; Hoogerhout, P.; Van Boom, J. H. Mol. Immunol.
1988, 25, 975–979.
34.Ho¨vel, K.; Shallom, D.; Niefind, K.; Belakhov, V.;
Shoham, G.; Baasov, T.; Shoham, Y.; Schomburg, D.
EMBO J. 2003, 22, 4922–4932.
5.Coker, R.J. Trop. Med. Int. Health 2004, 9, 25–40.

´
C. Remond et al. / Carbohydrate Research 340 (2005) 637–644
644
35. Mala, S.; Dvorakova, H.; Hrabal, R.; Kralova, B.
Carbohydr. Res. 1999, 322, 209–218.
36. Fessner, W.-D.; Juarez-Ruiz, J. M. Can. J. Chem. 2002,
80, 739–742.
43. De Lederkremer, R. M.; Lima, C.; Ramirez, M. I.;
Ferguson, M. A.; Homans, S. W.; Thomas-Oates, J. J.
Biol. Chem. 1991, 266, 23670–23675.
44. Pathak, A. K.; Pathak, V.; Suling, W. J.; Gurcha, S. S.;
Morehouse, C. B.; Besra, G. S.; Maddry, J. A.; Reynolds,
R.C. Bioorg. Med. Chem. 2002, 10, 923–928.
37.Nilsson, K.G.I.
103.
Carbohydr. Res. 1987, 167, 95–
38. Zeng, X.; Yoshino, R.; Murata, T.; Ajisaka, K.; Usui, T.
Carbohydr. Res. 2000, 325, 120–131.
45. Jahn, M.; Marles, J.; Warren, R. A.; Withers, S. G.
Angew. Chem., Int. Ed. 2003, 42, 352–354.
´
39. Prieto, A.; Bernabe, M.; Leal, J. A. Mycol. Res. 1995,
99(PN Part 1), 69–75.
46.Re mond, C.; Ferchichi, M.; Aubry, N.; Plantier-Royon,
R.; Portella, C.; OꢀDonohue, M.J. Tetrahedron Lett. 2002,
43, 9653–9655.
47. Debeche, T.; Cummings, N.; Connerton, I.; Debeire, P.;
OꢀDonohue, M.J. Appl. Environ. Microbiol. 2000, 66,
1734–1736.
40.Pederson, L.L;.Turco, S.J.
259–266.
Cell. Mol. Life Sci. 2003, 60,
41. Whitfield, C.; Richards, J. C.; Perry, M. B.; Clarke, B. R.;
Maclean, L.L. J. Bacteriol. 1991, 173, 1420–1431.
42. Moody, S. F.; Handman, E.; McConville, M. J.; Bacic, A.
J. Biol. Chem. 1993, 268, 18457–18466.
48. Helm, R. F.; Ralph, J.; Anderson, L. J. Org. Chem. 1991,
56, 7015–7021.