Enzymatic synthesis of oligo-d-galactofuranosides and l-arabinofuranosides: From molecular dynamics to immunological assays

Article abstract of DOI:10.1039/b926988f

D-Galactofuranosyl-containing conjugates are ubiquitous in many pathogenic microorganisms, but completely absent from mammals. As they may constitute interesting pharmacophores, recent works have been dedicated to their preparation. Besides well-reported

Full text of DOI:10.1039/b926988f

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Enzymatic synthesis of oligo-D-galactofuranosides and L-arabinofuranosides:
from molecular dynamics to immunological assays†‡
Ilona Chlubnova´,a,b,c Dominik Filipp,d Vojtech Spiwok,c Hana Dvorˇa´kova´,e Richard Daniellou,*a,b
Caroline Nugier-Chauvin,*a,b Blanka Kra´lova´c and Vincent Ferrie`resa,b
Received 23rd December 2009, Accepted 15th February 2010
First published as an Advance Article on the web 11th March 2010
DOI: 10.1039/b926988f
D-Galactofuranosyl-containing conjugates are ubiquitous in many pathogenic microorganisms, but
completely absent from mammals. As they may constitute interesting pharmacophores, recent works
have been dedicated to their preparation. Besides well-reported chemical procedures, enzymatic
approaches are still limited, mainly due to the lack of the corresponding biocatalysts. Based on the
similarity between chemical structures, the arabinofuranosyl hydrolase Araf 51 from Clostridium
thermocellum was expected to recognize both the L-Araf motif and its D-Galf analogue. Molecular
dynamics and STD-NMR were firstly used to confirm this hypothesis and increase our knowledge of
the active site. Interestingly, this arabinofuranosidase was not only able to hydrolyze galactosyl
derivatives, but was also really efficient in catalyzing oligomerisations using p-nitrophenyl furanosides as
donors. The structures of the products obtained were determined using mass spectrometry and NMR.
Amongst them, all the possible regioisomers of di-arabino and -galactofuranosides were synthesized,
and the ratio of each regioisomer was easily tuned with respect to the reaction time. Especially, the
galactofuranobioside displaying the biologically relevant sequence b-D-Galf -(1,6)-b-D-Galf was
enzymatically prepared for the first time. All fractions going from di- to penta-arabino- and
galactofuranosides were tested for their ability in eliciting the production of TNF-a. Interesting
immunological properties were observed with arabinofuranosides as short as three sugar residues.
b-(1,5)- and b-(1,6)-linked D-Galf residues. Moreover, the latter
has been described to be essential for the growth and survival
of mycobacteria.3 Both AG and LAM have critical roles in the
pathogenicity of mycobacterial diseases, including tuberculosis.2,4
More precisely, inhibition of enzymes that assemble these polysac-
charides prevents proliferation of mycobacteria.5 Noteworthily,
these hexofuranosides are not present in mammals. Therefore,
structurally well-defined synthetic analogues are of great interest
for the development of new pharmacophores and new therapies.6–9
Some of them have been proven to present antigenic properties. For
instance, b-D-Galf O-linked oligosaccharides, found in Penicillium
and Aspergillus species, appeared to be immunodominant.10,11
Many efforts were then dedicated to the chemical synthesis of
fragments of natural galactofuranose-containing glycosides.1,12
Disaccharidic sequences such as b-D-Galf -(1,5)-b-D-Galf and b-
D-Galf -(1,6)-b-D-Galf ,13–18 as well as trisaccharidic units of the
galactan produced by M. tuberculosis19 have been synthesized.
However, the number of methods involving biocatalysts is still
quite limited, probably due to the low availability of the re-
quired specific enzymes. Lowary and colleagues have recently
studied a galactofuranosyl transferase and performed a small
elongation of a galactofuranosyl chain.20 Nevertheless, and to
the best of our knowledge, no galactofuranosyl hydrolase has
been involved in chemo-enzymatic approaches. In this context,
and owing to the structural similarity between residues of D-
Galf and L-Araf (Fig. 1), biocatalyzed furanosylation has been
recently developed. The wild type arabinofuranosyl hydrolase
Abf D3 (E.C.3.2.1.55) from Thermobacillus xylanolyticus was
obviously efficient enough to recognize and transfer 4-nitrophenyl
Introduction
D-Galactose is by far the most widespread hexose found in
the furanose form in naturally occurring glycoconjugates from
many bacteria, protozoa, fungi, plants and archaebacteria.1 One
impressive example is that of Mycobacterium tuberculosis. Despite
several decades of successful chemotherapeutic treatment, this
microorganism has re-emerged through multidrug resistance to
become a major cause of mortality, with an annual rate of
approximately three million deaths over the world. The cell wall
complex in this species is largely composed of two polysaccharides,
a lipoarabinomannan (LAM) and an arabinogalactan (AG), in
which all of the D-galactose and the D-arabinose residues are
present as five-membered rings.2 The AG portion of mycobac-
teria comprises a linear chain of approximately 30 alternating
aEcole Nationale Supe´rieure de Chimie de Rennes, CNRS, UMR 6226,
Avenue du Ge´ne´ral Leclerc, CS 50837, 35708 Rennes Cedex 7, France.
E-mail: richard.daniellou@ensc-rennes.fr, caroline.nugier@ensc-rennes.fr;
Fax: (+) 33-2-23-23-80-46; Tel: (+) 33-2-23-23-80-96
bUniversite´ europe´enne de Bretagne, France
cDepartment of Biochemistry and Microbiology, Institute of Chemical
Technology of Prague, Technika´ 3, 166 28 Prague 6, Czech Republic
dLaboratory of Immunobiology, Institute of Molecular Genetics AS CR,
Videnska 1083, 142 20 Prague 4, Czech Republic
eLaboratory of NMR spectroscopy, Institute of Chemical Technology of
Prague, Technika´ 5, 166 28 Prague 6, Czech Republic
† This paper is part of an Organic & Biomolecular Chemistry web theme
issue on biocatalysis.
‡ Electronic supplementary information (ESI) available: NMR spectra of
compounds 3–13. See DOI: 10.1039/b926988f
2092 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
with explicit water in a periodic box, several molecules of water
being changed for Na+ in order to keep the box neutral. Finally,
after optimization of the geometry, the simulations of molecular
dynamics (MD) were first performed with fixed chosen atoms
(restrained) for 0.2 + 0.2 ns, in order to equilibrate each system,
and then followed by a 5 ns unrestrained MD. Final pictures are
depicted in Fig. 2. In both cases, the xylopyranoside derivative
was rather flexible in the +1 subsite. Its major interaction with
the protein seemed to occur through stacking with a tryptophan
residue, thus confirming results previously obtained during the
structural determination and explaining the substrate versatility
naturally observed for such enzymes. Consequently, we inferred
that the +1 subsite was likely to interact with a panel of aglycons,
such as the furanosyl entity or aromatic rings. On the contrary,
the Araf part of the natural ligand was quite stable during
the simulation. This behaviour was mainly ascribed to a strong
anchorage of the arabinosyl unit in the -1 subsite, due to the
participation of multiple hydrogen bonds with at least six different
amino acid residues. Interestingly, the Galf residue also stabilized
in a similar position, thanks to an equivalent network of direct
polar interactions, despite increased steric hindrance. Still, it is
noteworthy that the hydroxymethyl group finally settled in a non-
polar cavity. Moreover, there was no indication that water binds
to this cavity when an arabinoside substrate is in the active site of
Araf 51. However, such computer modelling clearly demonstrated
that the enzyme could be able to recognize the Galf motif,
potentially even better than the Araf disaccharide. Therefore, this
hypothesis encouraged us to pursue our study.
Fig. 1 Structural similarities between pNP-L-Araf 1 and pNP-D-Galf 2.
L-arabinofuranoside 1 (pNP-Araf ), but also the corresponding
D-galactofuranoside 2 (pNP-Galf ) to either pyranosyl acceptors
in transglycosylation reactions or to pNP-Galf itself in auto-
condensation dimerizations.21,22 The versatility of this enzyme was
further extended to the preparation of non-natural disaccharides
obtained from 5-deoxy-L-arabinosyl and D-fucofuranosyl, as well
as 6-fluoro-D-galactofuranosyl donors.23,24 Despite these interest-
ing results towards the synthesis of (1,2)- and (1,3)-difuranosides,
further biocatalysts with quite different reactivity (regioselectivity,
capability of oligomerization and/or branching) deserve to be
studied and developed.
In this paper, we present our recent exploration of another
a-L-arabinofuranosyl hydrolase, Araf 51, from Clostridium ther-
mocellum.25 This particular enzyme, belonging to the GH51
family, is known to remove a-(1,2)-, a-(1,3)- and a-(1,5)-linked
L-arabinofuranosyl moieties from arabinans and xylans. We were
thus really inclined to evaluate the potential synthetic ability
of this enzyme for the preparation of both L-arabinofuranosyl-
and D-galactofuranosyl-containing oligosaccharides. Molecular
modelling was first performed to anticipate how the target pentose
and hexose moieties fit within the active site of the enzyme.
Subsequently, the scope of interactions evaluated in silico was
completed in aqueous solution by a Saturation Transfer Difference
NMR (STD-NMR) study. The Araf 51-catalyzed oligomeriza-
tions using either 1 or 2 as favoured substrates were finally
performed and the resulting oligofuranosides were examined as
potent immunostimulating agents.
Results and discussion
Unlike Abf D3, the arabinofuranosidase Araf 51 has never been
used for synthetic purposes up to now, and even pNP-Galf 2 has
never been reported to act as a substrate for this enzyme. However,
since the amino acid sequences of these two biocatalysts share 26%
of their identity, and since a similar active site has been revealed
through X-ray determination of both crystal structures,25,26 we
were likely to envisage Araf 51 as a strong candidate for the
enzymatic synthesis of oligogalactofuranosides.
Molecular dynamics and STD-NMR
Firstly, we compared the in silico behaviour of the natural ligand
a-L-arabinofuranosyl-(1,4)-a-D-xylopyranoside (Araf -Xylp) to
that of its Galf analogue (Galf -Xylp). Therefore, one subunit
of the homohexameric enzyme containing the co-crystallized
substrate was extracted from the 2C8N pdb file. The 6-C side chain
of one of the furanosyl parts was modified manually to obtain the
analogue. Both complexes of enzyme and ligand were surrounded
Fig. 2 Molecular modelling of arabinofuranoside (2A) and galactofura-
noside (2B) in the active site or Araf 51 (the circle highlights the interesting
tryptophan residue).
In order to complement our in silico findings, we also con-
sidered a physical method. Saturation Transfer Difference NMR
(STD-NMR) nowadays constitutes a reliable and cheap tool for
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 2092–2102 | 2093
©

View Article Online
observing the interactions between ligand and protein.27 Par-
ticularly, this method allows the quantification of the relative
importance of the interaction between each proton of a substrate
and the active site. Because the previously used disaccharides were
not commercially available, experiments were performed using
both furanosides 1 and 2. To ensure observation of the STD effect,
the E173A mutant25 was preferred to the wild type Araf 51, in
order to prevent hydrolysis of the substrates and, therefore, for
simplifying NMR analysis. The resultant effects were expressed as
percentages, giving 100% to be the highest signal, i.e. the proton in
the meta position of the aromatic aglycon (Fig. 3). The pNP rings
in both 1 and 2 showed the strongest signals and, therefore, were in
close contact with the active site of Araf 51, surely through a strong
p–stacking interaction involving the tryptophan residue previously
shown as responsible for the recognition of the xylopyranoside
ring. Probably despite this particular effect caused by the artificial
aglycon, the sugar ring displayed moderate signals. Nevertheless,
all protons of the furanosyl moiety show STD effects. Particularly,
1-H with its 12 or 11% effect for 1 and 2, respectively, appeared
to have good interaction with residue(s) in the active site. Even
more interestingly, the hindered 6-C side chain of 2 exhibited the
same value as the hydroxymethyl group of 1 and did not disturb
at all the way the sugar fit into the protein. Consequently, STD
data complemented well in silico modelling and strengthened the
thought that both molecules 1 and 2 exhibited a similar pattern
of interactions with the arabinofuranosidase Araf 51, and could
be engaged in biocatalytic reactions. However, these two studies
needed to be complemented by the measurement of the specificity
of the enzymes towards both substrates.
Kinetic parameters of Araf 51
The kinetic parameters of Araf 51 were evaluated by looking at the
hydrolytic activity of the enzyme towards the Galf derivative 2.
Whereas Araf 51 is a thermophilic enzyme with Topt ~ 82 ◦C, using
pNP-Araf 1 as a substrate,25 we determined its activity at 25 ◦C, for
practical reasons. The value obtained for Km for the hydrolysis of
1 under standard conditions (100 mM PBS pH 7, 1 mg mL-1 BSA)
was 0.15 mM (0.25 mM at 37 ◦C). Optimum pH was determined
for this reaction using both 0.65 and 1 mM as the substrate
concentrations, and thus, appeared to be in the range of 7 to 8.
Evaluation of Km for the hydrolysis of 2 under the same conditions
showed a significantly greater value of 53 mM, similar to the
one previously observed during Abf D3-catalyzed hydrolysis.23
Moreover, Araf 51-mediated hydrolysis of 2 was characterized by
a significantly lower activity (Vm = 0.32 mmol min-1) compared
with that of 1 (Vm = 2.7 mmol min-1). Despite a 6000-fold lower
specificity (Vm/Km), the D-galactofuranoside 2 is still recognized
as a substrate by the enzyme, as previously shown through the
molecular modelling and the STD-NMR study. Nevertheless,
measurement of the kinetic parameters clearly showed that Galf
derivative 2 can be hydrolyzed, demonstrating that it is in a
favourable position in the active site of Araf 51, and might act
as an efficient donor in auto-condensation reactions.
Analytical study of the auto-condensation reactions
Furanosides 1 and 2 (5 mM) were incubated with Araf 51 (2376
UI) at 60 ◦C and pH 7.4. Reactions were quenched by enzyme
◦
denaturation at 100 C for three minutes before HPLC analysis.
All products were separated thanks to size exclusion chromatog-
raphy, and also detected and quantified by UV absorbance at
280 nm. Newly observed peaks at shorter retention times than
those corresponding to the substrates 1 (77.7 min) and 2 (78.7
min) were isolated. Mass spectrometric analyses allowed simple
identification of four fractions, respectively containing penta- (10
min), tetra- (13–15 min), tri- (18–20 min) and difuranosides (30–
55 min). In a second experiment, aliquots were withdrawn at
increasing reaction times (2, 5, 10, 15 and 20 min) in order
to evaluate the impact of time on the ratio of all synthesized
oligofuranosides. Starting from arabinoside 1, the HPLC profile
revealed that the substrate was completely consumed after less
than 10 min (Fig. 4). Moreover, the overall conversion yield
Fig. 4 Kinetics of the pNP-Araf 1 auto-condensation products catalyzed
by Araf 51. The conversions were calculated as a percentage of the initial
quantity of 1 from 0 to 45 min.
Fig. 3 STD-NMR spectra of 1 (3A) and 2 (3B).
2094 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
observed for the desired auto-condensation reaction reached its
maximum at 35% after approximately 5 min. Interestingly, the di-,
tri- and tetrasaccharides were produced right from the beginning
of the Araf 51-catalyzed incubation with maximum conversions
of 22.1, 8.4 and 3.6%, respectively, within less than 5 min of
reaction. Thereafter, decrease of the auto-condensation products
were observed, suggesting the straightforward partial hydrolysis
of these oligofuranosides.
As far as the pNP-Galf 2 auto-condensation reaction was
concerned, consumption of the substrate was nearly complete after
10 min (Fig. 5). Maximum conversion was indeed observed from
5 min and, thanks to mass spectrometric analysis, three fractions
corresponding to di-, tri- and tetrasaccharides were identified,
representing 34.6, 8.6 and 1.1% yield, respectively. Thereafter,
the amount of disaccharide significantly decreased, while those
of the upper tri- and tetrasaccharides still increased to reach a
plateau at nearly 10 and 4%, respectively. These analytical results
clearly showed different behaviour between the arabinose and
the galactose series towards Araf 51. They could be ascribed
(i) to a very interesting ability of the Araf 51 to recognize the
galactofuranoside 2 and to use it as a donor in the glycosylation
reaction, and (ii) to a lower capability of the biocatalyst for
hydrolyzing the synthesized galactofuranosides, thus resulting in
higher conversion yields starting from 2.
spectrometry and their structures were finally assessed through 1D
and 2D NMR analysis.
Scale-up was first performed from the arabinoside 1. In accor-
dance with the analytical studies, arabinofuranobiosides consti-
tuted the major fraction (22.6%), followed by the trisaccharides
(7.9%) and the tetrasaccharides (3.1%) (Table 1, entries 1–3).
Moreover, even in low yield, we were also able to isolate pen-
tasaccharides (entry 4), and prove their presence thanks to mass
spectrometry. A more detailed analysis of the NMR data of the
fraction containing the disaccharides permitted us to characterize
the three different and possible regioisomers 3, 4 and 5. Because the
spectra were similar to those previously published,21 disaccharide
3, which was obtained with the best isolated yield of 12.6%,
was unambiguously identified as the a-(1,2)-linked furanobioside.
This result already constitutes a significant improvement on the
enzymatic synthesis of this compound, as it was only isolated
in 5.8% when using Abf D3. Subsequently, the small coupling
constants observed between 1b-H and 2b-H (J1b,2b = 1.8 and
1.2 Hz for 4 and 5, respectively) clearly indicated that the newly
formed glycosidic bonds exhibited a-L-stereochemistry. As 4 is
concerned, its 13C NMR spectra showed a downfield shift of the
3a-C (Dd = +3.8 ppm, compared to 3a-C of 1) of the reducing
arabinofuranosyl moiety, as well as intense three-bond coupling
in the 1H–13C (HMBC) spectra between 3a-C (d = 81.9 ppm) and
1b-H (d = 5.06 ppm). This indicated that the linkage is (1,3).
For disaccharide 5, similar results were obtained, but involving
5a-C (Dd = +4.0 ppm compared to 1, d = 66.6 ppm) and 1b-
H (d = 4.92 ppm). These data demonstrated the presence of an
a-L-(1,5) bond between both furanosyl residues. Consequently,
the three possible isomers could be synthesized thanks to the
use of the Araf 51 biocatalyst. With structural data in hand, we
also established the evolution profile according to reaction time
(Fig. 6). Thus, between 2 and 10 min, the ratio of 3, 4 and 5
was significantly modified. Indeed, the (1,2)-linked disaccharide
3 was kinetically produced after 2 min, but disappeared after
10 min. On the other side, its (1,5)-linked counterpart was the
major one after this time. As a consequence, and with regard
to the observed regioselectivity of the reaction, as well as the
amount obtained for each disaccharide, we can easily deduce that
the a-L-(1,2) linkage is kinetically formed (but rapidly hydrolyzed
thereafter), and that the a-L-(1,5), and the a-L-(1,3) but to a lesser
extent, are thermodynamically favored. Such kinetic observations
are in complete accordance with the hydrolytic preferences of
Araf 51 towards arabinoxylane in nature, and emphasize the fact
that frequently, information on the hydrolytic substrate specificity
of glycosidases can be very valuable in order to predict the
regioselectivity of transglycosylation reactions.28
Fig. 5 Kinetics of the pNP-Galf 2 auto-condensation products catalyzed
by Araf 51. The conversion ratios were calculated as a percentage of the
initial quantity of 2 from 0 to 45 min.
Preparative scale of enzymatic reaction
Analytical results further needed to be sharpened in order to
precisely elucidate the number of regioisomers formed, as well
as the nature of the glycosidic linkages. When auto-condensation
reactions were performed on larger scale (100 mM of substrate,
i.e. 30 mg of 1 or 100 mg of 2) but with lower quantity of the
biocatalyst, times had to be adjusted to 25 and 90 min, respectively,
in order to reach the maximum conversion. After denaturation of
the enzyme at 100 ◦C for 10 min, products were analyzed by usual
TLC methods (see the ESI‡) and carefully separated thanks to the
use of P2 biogel. Such size-exclusion chromatography allowed us
to rapidly quantify the amounts of the different oligosaccharides,
once again going from di- to pentafuranosides. Sugars with
similar retention properties were then pooled together and finally
lyophilized. Purified compounds were then identified using mass
Having elucidated the structures of disaccharides and their
behaviour under the described conditions, we expected Araf 51
to be able to produce all of the trisaccharidic regioisomers, but
in various amounts. In this context, three different trifuranosides
6–8 were chromatographically separated and isolated in 2.7, 0.6
and 2.2% yield, respectively. On the assumption of small coupling
constants between 1b-H and 2b-H (J1b,2b = 1.2–1.8 Hz), and also
between 1c-H and 2c-H (J1c,2c = 1.2 Hz), 1,2-trans configurations
between arabinosyl entities were easily established. The nature
1
of the linkages was subsequently deduced from 13C and H–13C
NMR spectra. The arabinotrioside 6 was demonstrated to have
two (1,2) linkages, since the two carbon atoms exhibited signals
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 2092–2102 | 2095
©

View Article Online
Table 1 Auto-condensation of 1
Entry
Products
Yield (%)
Elucidated structures (yield, %)
1
2
3
4
Disaccharides
22.6
7.96
3.1
3 (12.6)
6 (2.7)
4 (2.2)
7 (0.6)
5 (5.1)
8 (2.2)
Trisaccharides
Tetrasaccharides
Pentasaccharides
0.1
of oligogalactofuranosides starting from donor 2 (Table 2). As
expected, disaccharides represented the major part of the isolated
compounds, with a yield of 34.5%, even higher than the one
obtained in the arabinose series (entry 1). Tri- and tetrasaccharides
were also obtained but in somewhat lower yields (entries 2 and
3). Only small amounts of derivatives with five sugar units were
detected by mass spectrometry (entry 4). Unsurprisingly with
regards to the structural similarity between L-Araf and D-Galf ,
1H NMR spectra of all of the compounds showed signals for 1-H
with small J1,2 coupling constants close to 1.2 Hz, representative
of a 1,2-trans b-D-glycodic linkages. Moreover, all of the possible
regioisomers, i.e. (1,2), (1,3), (1,5) and (1,6), could be observed
in the galactofuranobiosides fraction (entry 1). Notably, 9, which
was previously enzymatically synthesized in an isolated yield of
6.0% using Abf D3, was obtained, herein, in 23.9% yield.22 All of
the other structures were determined through careful analysis of
the 13C and the HMBC NMR spectra. Although the majority
of these compounds were isolated in moderate quantities, the
Araf 51-biocatalyzed reaction gave 12 in 7.8% yield. Moreover,
it is important to note the influence of reaction time on the ratios
of 9–12 (Fig. 7). Indeed, the (1,2)-linked digalactoside 9 was
rapidly formed but also quickly disappeared from the reaction
mixture. On the contrary, (1,3), (1,5) and (1,6) disaccharides were
concomitantly obtained and their amounts remained constant
under a large range of time. It is noteworthy that 12, the major
thermodynamic disaccharide, displays the important sequence b-
D-Galf -(1,6)-b-D-Galf . This is found in many pathogenic microor-
ganisms, and was here enzymatically prepared for the first time.
Finally, amongst the variety of possible oligogalactofuranosides,
three trisaccharides were clearly characterized: (i) 13 presented a
linear chain built up with (1,2) glycosidic bonds, (ii) 14 exhibited
Fig. 6 HPLC profiles showing the modulation of the ratio of disaccha-
rides 3, 4 and 5 in the presence of 2376 units of Araf 51.
with a downfield shift compared to 1 (for 2a-C, Dd = +3.7 ppm,
and for 2b-C, Dd = +4.5 ppm, Table 1). They also showed an
intense correlation in the HMBC experiment with 1b-H (d = 5.19
ppm) and 1c-H (d = 5.02 ppm), respectively. The structure of 7
was identified as the (1,5) linear analogue as its 5a-C and 5b-C
showed upshifted signals (Dd = +4.1 ppm), and correlations with
1b-H (d = 4.93 ppm) and 1c-H (d = 4.92 ppm). Finally, 8 was
deduced to be the branched (1,2)- and (1,5)-triarabinofuranoside
using a similar reasoning, and the signals corresponding to 2a-C
(Dd = +3.7 ppm) and 5a-C (Dd = +3.3 ppm). Therefore, Araf51
showed very interesting versatility, even for the production of both
linear and branched oligoarabinofuranosides.
Looking at the selectivity of Araf 51 during reactions involving
pNP-Araf 1, we were really confident in isolating a great variety
2096 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 2 Auto-condensation of 2
Entry
Products
Yield (%)
Elucidated structures (yield, %)
1
2
3
4
Disaccharides
34.5
4.6
2.5
0.1
9 (23.9)
13 (1.2)
10 (0.6)
14 (0.8)
11 (1.5)
15 (0.3)
12 (7.8)
Trisaccharides
Tetrasaccharides
Pentasaccharides
mycobacteria species.12 Therefore, the synthesis of small and well-
defined oligofuranosides with immuno-modulatory properties
represents a promising challenge.29,30 To assess the biological prop-
erties of the enzymatically synthesized neofuranosides described
in this study, we incubated a murine macrophage cell line Raw 264
with either oligo-arabino or galactofuranosides. Their stimulatory
effects measured by the TNF-a production were compared to
those of a lipopolysaccharide (LPS) used as a positive control
(Fig. 8). It is of note that trace levels of contaminating LPS in the
samples used, measured by limulus assay, had no significant effect
on the stimulatory activity of neofuranosides, as the addition of
LPS inhibitor polymyxin B to cell cultures resulted in production
of comparable levels of TNF-a (data not shown).
As depicted in Fig. 8 and in sharp contrast to LPS, none
of the galactofuranosides were able to produce any significant
response, even when used at concentrations up to 1 mg mL-1.
This observation may be assigned to both the short length of the
oligosaccharide chains and to their high structural heterogeneity
in the mixtures used. Importantly, significant TNF-a responses
were elicited upon stimulation with oligoarabinofuranosides when
used at concentrations of 100 ng mL-1 and higher. Cytokine
responses were concentration- and length of the oligosaccharide-
dependent (see columns Tri-, Tetra- and Penta-Ara). Stimulatory
effects comparable to the Tri-Ara mixture were observed using
isolated 6, 7 and 8. These data strongly suggest that the nature
Fig. 7 HPLC profiles showing the modulation of ratio of disaccharides
9–12 in the presence of 2376 units of Araf 51.
(1,2) and (1,3) glycosidic linkages, and (iii) 15, a branched (1,2) and
(1,6) galactotrioside, resulted from the two favoured regiochemical
preferences of Araf 51.
Biological evaluation
The presence of galactofuranosyl residues in M. tuberculosis
constitutes an interesting and original target in the fight against
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 2092–2102 | 2097
©

View Article Online
600 MHz for 1H and 125 MHz for 13C. Chemical shifts are
given in d-units (ppm) measured downfield from Me4Si. Coupling
constants J are given in Hz. Electrospray-ionisation mass spectra
(ESI-MS) were recorded on samples dissolved in MeOH injected
in a volume of 2–5 mL into a flow (100 mL min-1) of MeOH
using a Q-TOF Micro (Waters, USA). Sample cone voltage was
42 V and the source temperature was 150 ◦C. Measurements were
performed in positive ([M + Na]+ ion detection) mode in the range
of 100–1000 Da.
Molecular dynamics
All simulations were performed in GROMACS 3.3.2 package.34
A single subunit of arabinofuranosidase complexed with b-L-
Araf -(1,3)-b-D-Xylp (PDB ID: 2C8N) was taken as the starting
structure.25 The protein was modelled using the AMBER99SB
force field.35 Protonation states of dissociable residues were
assigned using the output from the H++ tool.36 Force field
parameters of b-L-Araf -(1,3)-b-D-Xylp were generated using
Glycam server.37 Force field parameters of b-D-Galf -(1,3)-b-D-
Xylp were generated from those of b-L-Araf -(1,3)-b-D-Xylp by
prolonging the hydroxymethyl moiety (missing parameters were
taken from N-acetylneuraminic acid). Partial charges were taken
from b-D-Galf -OMe calculated using the RESP38 method at the
HF/6-31G* level of theory using Gaussian03 package.39 Glycam
topologies were then converted to GROMACS format using a
modified amb2gmx.pl script (it was necessary to modify the script
because in the original it ignores periodic functions in dihedral
terms if they are multiplied by negative coefficients).40 The starting
structure of the complex with b-L-Araf -(1,3)-b-D-Xylp was taken
from the crystal structure. An additional moiety of the complex
with b-D-Galf -(1,3)-b-D-Xylp was fitted manually. Structures were
solvated by 13069 TIP3P41 water molecules and seven sodium ions
to neutralize its net charge. The system was energy minimized
and then equilibrated by 400 ps molecular dynamics simulation
with harmonic restraints applied to non-hydrogen atoms of the
protein (200 ps strong and 200 ps weak restraints). Finally, a 5 ns
unrestrained molecular dynamics simulation was performed.
Fig. 8 Production of TNF-a cytokine upon macrophage stimulation
with enzymatically synthesized oligoarabinofuranosides (Ara) and oli-
gogalactofuranosides (Gal). (* Effect of LPS at 1000 ng mL-1 was not
assessed.)
of the glycosidic linkages between furanosyl units, i.e. (1,2), (1,3)
and/or (1,5), had no significant impact on the production of TNF-
a. Thus, even if immune activities of polysaccharides containing
furanosyl entities were already reported,31,32 this is the first time
to the best of our knowledge that such short oligofuranosides,
mimics of natural oligosaccharides derived from pathogens, are
able to induce a cytokine response in immunocompetent cells.
Conclusions
Besides the (1,2) and (1,3) linkages already yielded by the
arabinofuranosyl hydrolase Abf D3, the linkages (1,5) and (1,6),
reported to be present in some clinically important pathogens, were
obtained thanks to the assistance of Araf 51. These results also
confirmed preferences of this easily available enzyme to transfer
the furanosyl residue kinetically onto the 2-OH and thermo-
dynamically to the primary hydroxyl function of the furanosyl
acceptor, whatever the Araf or the Galf series. Moreover, the
regioselectivity of the reaction was modulated with respect to
reaction time. It also has to be pointed out that the isolated yields
were significantly improved in all cases and thus, this enzymatic
procedure is now at least as competitive as the corresponding
multistep chemical reactions. Interestingly, Araf 51 appeared to
be much more efficient as a synthetic tool than its counterpart
Abf D3: (i) the ratio of hydrolysis over auto-condensation of
substrates as well as reaction times being reduced, (ii) auto-
condensation reactions being not limited to the preparation of
disaccharides. These results encourage us to pursue the study
of the synthetic potency of Araf 51 and will lead us to the
evaluation of its versatility towards original furanosyl-based
substrates. Finally, immunostimulating responses obtained with
short oligoarabinofuranosides are very promising, and this opens
very interesting opportunities in this field,33 especially for further
development on poly-L-arabinan directly extracted from biomass.
STD-NMR experiments
Samples were prepared in 0.5 mL of D2O and contained ~30 mmol
of pNP-sugars. STD-NMR spectra were recorded on a Bruker
Avance III 400 MHz spectrometer. Processing of all data was
performed on a PC with Bruker Topspin v2.0 software. After the
determination of the optimal conditions, i.e. temperature, delay
between pulse (d20) and molecular ratio (protein : ligand), STD-
NMR experiments were performed at 283 K as followed. The
protein (1 : 100 ratio) was saturated on resonance at 0.7 ppm and
off resonance at 40 ppm with a cascade of 40 selective gaussian-
shaped pulses of 50 ms duration with a 100 ms delay between
each pulse. The total duration of the saturation time was set to
2 s. A total of 256 scans/STD-NMR experiment was acquired.
A WATERGATE sequence was used to suppress residual HOD
signal. A spin lock filter with strength of 5 kHz and duration of
10 ms was also applied to suppress the protein background. A
similar experiment with no enzyme was used as a reference in
order to verify the absence of STD effects in these experimental
conditions. Intensities of all STD effects were calculated though
Experimental
General procedures
1H, 13C, HSQC, HMBC and COSY NMR spectra were recorded
on a Bruker 600 spectrometer equipped with a cryoprobe at
2098 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
integrals over the respective signals in 1H NMR reference spectra.
The largest STD effect in each spectrum was set to 100% and
relative intensities were determined, as common for non-refined
STD effects. Hence, sufficient comparisons of relative STD effects
between sugars were possible, but absolute binding intensities
could not be determined.
after 25 min or 90 min in the case of 1 and 2, respectively, by
enzyme denaturation at 100 ◦C for 10 min. The reaction products
were separated by gel permeation chromatography on P-2 Bio-
Gel (Bio-Rad) using an FPLC system consisting of a solvent
delivery system Biologic F40 DuoFlow, Biologic QuadTec UV-Vis
Detector and Biologic BioFrac Fraction Collector (all Bio-Rad).
Deionized filtrated (0.22 mm PVDF membrane, Millipore) water
was used as a mobile phase with a flow rate of 0.15 mL min-1.
Separation was monitored by UV absorbance at 280 and 405 nm
by operating software Biologic DuoFlow. Collected fractions
were lyophilized (FreeZone Freeze Dry System, Labconco) and
submitted to structural analyses.
Production of the enzyme and determination of its kinetic
parameters
Recombinant arabinofuranosidase Araf 51 expressed from the
plasmid-borne Araf 51 gene was produced and purified from
Escherichia coli cells as described previously.25 The hydrolytic
activity of Araf 51 was quantified by incubation of the enzyme
with 1 or 2 (5 mM) in 100 mM potassium phosphate buffer pH 7.4
at 60 ◦C. Continuous release of p-nitrophenol was measured at
405 nm (Microplate Spectrophotometer PowerWave XS/XS2,
BioTek) and data evaluated with Gen5 Data Analysis Software
(BioTek). One unit of activity corresponds to the amount of
enzyme releasing 1 mmol of p-nitrophenol per minute. Initial rate
conditions and suitable substrate concentrations (0.03–3 mM and
0.625–40 mM for 1 and 2, respectively) were used in order to
determine the kinetic parameters Km and kcat.
p-Nitrophenyl a-L-arabinofuranosyl-(1,2)-a-L-arabinofuranoside
(3). 1H NMR (600 MHz, D2O): d = 8.11 (2H, d, J = 9.6, Hm
C6H4), 7.08 (2H, d, J = 9.6, Ho C6H4), 5.81 (1H, d, J1a,2a = 1.5,
1a-H), 5.08 (1H, d, J1b,2b = 1.5, 1b-H), 4.30 (1H, dd, J2a,3a = 4.0,
2a-H), 4.09 (1H, dd, J3a,4a = 6.2, 3a-H), 4.02 (1H, ddd, J4a,5a
=
3.3, J4a,5¢a = 5.1, 4a-H), 4.00 (1H, dd, J2b,3b = 3.3, 2b-H), 3.87 (1H,
ddd, J4b,5b = 3.3, J3b,4b = J4b,5¢b = 6.2, 4b-H), 3.79 (1H, dd, 3b-H),
3.68 (1H, dd, J5a,5¢a = 12.8, 5a-H), 3.59 (1H, dd, 5¢a-H), 3.55 (1H,
dd, J5b,5¢b = 12.8, 5b-H), 3.30 (1H, dd, 5¢b-H) ppm. 13C NMR
(125 MHz, D2O): d = 161.4 (Cipso C6H4), 142.3 (Cp C6H4), 126.1
(Cm C6H4), 116.7 (Co C6H4), 107.6 (1b-C), 104.5 (1a-C), 87.3 (2a-
C), 84.2 (4b-C, 4a-C), 81.3 (2b-C), 76.6 (3b-C), 74.5 (3a-C), 61.0
(5b-C), 60.4 (5a-C) ppm. MS (ESI): m/z calcd for C16H21NNaO11
[M + Na]+ 426.1012; found 426.0892.
Analytical scale of auto-condensation reactions
Reactions were performed in buffered conditions (100 mM
potassium phosphate buffer, pH 7.4) at 60 ◦C with shaking (INC
Orbital Mixing Chilling/Heating Plate, Torrey Pines Scientific)
in the presence of pNP-glycosides (50 mM) and 2376 U of
enzyme. The total reaction volume was 360 mL. Aliquots (20 mL)
of the reaction mixture were withdrawn at different times and
the reaction was quenched by enzyme denaturation at 100 ◦C
for 3 min. Samples were mixed with 50 mL of deionized water,
passed through an ultrafiltration membrane (VectaSpin Micro-
tubes, MWCO 12 kDa, Whatman) at 20000 ¥ g for 30 min (Mikro
22R, Hettich) to remove protein, and applied in a volume of
25 mL to a calcium carbohydrate HPLC column. The HPLC
system was consisting of a solvent delivery system 600 LCD HPLC
Pump (Waters), an UV/VIS 486 Tunable Absorbance Detector
(Waters) and Differential Refractometer RIDK 101 (Laboratorn´ı
Prˇ´ıstroje Praha), equipped with a column (Supelcogel Ca, 30 cm ¥
7.8 mm, Supelco) and an appropriate guard column (Supelcogel
Ca and C611, 5 cm ¥ 4.6 mm, Supelco) maintained at 80 ◦C
during the analysis. Deionized and filtrated (0.22 mm PVDF
membrane, Millipore) water was used as a mobile phase in isocratic
mode with a constant flow rate of 0.5 mL min-1. The substrate,
products of auto-condensation and p-nitrophenol were detected
and quantified by UV absorbance at 280 nm.
p-Nitrophenyl a-L-arabinofuranosyl-(1,3)-a-L-arabinofuranoside
(4). 1H NMR (600 MHz, D2O): d = 8.11 (2H, d, J = 9.6, Hm
C6H4), 7.08 (2H, d, J = 9.6, Ho C6H4), 5.73 (1H, d, J1a,2a = 1.2, 1a-
H), 5.06 (1H, d, J1b,2b = 1.8, 1b-H), 4.43 (1H, dd, J2a,3a = 2.4, 2a-H),
4.13 (1H, ddd, J4a,5a = 3.0, J4a,5¢a = 5.4, 4a-H), 4.03 (1H, dd, J3a,4a
5.4, 3a-H), 4.00 (1H, dd, J2b,3b = 3.0, 2b-H), 3.88 (1H, ddd, J4b,5b
=
=
3.0, J3b,4b = J4b,5¢b = 6.0, 4b-H), 3.80 (1H, dd, 3b-H), 3.71 (1H, dd,
J5a,5¢a = 12.6, 5a-H), 3.67 (1H, dd, J5b,5¢b = 12.0, 5b-H), 3.63 (1H,
dd, 5¢a-H), 3.54 (1H, dd, 5¢b-H) ppm. 13C NMR (125 MHz, D2O):
d = 161.1 (Cipso C6H4), 142.3 (Cp C6H4), 126.1 (Cm C6H4), 116.8
(Co C6H4), 107.2 (1b-C), 105.6 (1a-C), 84.4 (4a-C), 84.0 (4b-C),
81.9 (3a-C), 81.2 (2b-C), 79.5 (2a-C), 76.6 (3b-C), 61.1 (5b-C),
60.8 (5a-C) ppm. MS (ESI): m/z calcd for C16H21NNaO11 [M +
Na]+ 426.1012; found 426.1002.
p-Nitrophenyl a-L-arabinofuranosyl-(1,5)-a-L-arabinofuranoside
(5). 1H NMR (600 MHz, D2O): d = 8.11 (2H, d, J = 9.6, Hm
C6H4), 7.07 (2H, d, J = 9.6, Ho C6H4), 5.71 (1H, d, J1a,2a = 1.8, 1a-
H), 4.92 (1H, d, J1b,2b = 1.2, 1b-H), 4.27 (1H, dd, J2a,3a = 3.0, 2a-H),
4.13 (1H, ddd, J4a,5a = 3.0, J4,5¢a = 5.4, 4a-H), 4.00 (1H, dd, J3a,4a
5.4, 3a-H), 3.95 (1H, dd, J2b,3b = 3.0, 2b-H), 3.92 (1H, ddd, J4b,5b
=
=
3.0, J3b,4b = J4b,5¢b = 6.0, 4b-H), 3.78 (1H, dd, 3b-H), 3.75 (1H,
dd, J5a,5¢a = 12.0, 5a-H), 3.65 (2H, dd, J5b,5¢b = 12.0, 5b-H, 5¢a-H),
3.54 (1H, dd, 5¢b-H) ppm. 13C NMR (125 MHz, D2O): d = 161.2
(Cipso C6H4), 142.3 (Cp C6H4), 126.1 (Cm C6H4), 116.7 (Co C6H4),
107.4 (1b-C), 105.4 (1a-C), 84.0 (4b-C), 83.7 (4a-C), 81.0 (2a-C),
80.9 (2b-C), 76.5 (3b-C, 3a-C), 66.6 (5a-C), 61.2 (5b-C) ppm. MS
(ESI): m/z calcd for C16H21NNaO11 [M + Na]+ 426.1012; found
426.1663.
Preparative scale reactions
Exper◦iments were performed in 100 mM phosphate buffer, pH 7.4
at 60 C in the presence of pNP-glycosides (100 mM) with 1440
U of Araf 51 in a total reaction volume of 1.05 mL for reaction
with 1, and with 4800 U of enzyme in a total reaction volume of
3.5 mL in the case of 2. Reaction mixtures were monitored by TLC
(Kieselgel 60 F254 (Merck), ethyl acetate–acetic acid–water, 7 : 2 : 2)
and visualized under UV light and by exposure to 2% orcinol in
20% H2SO4–ethanol. Auto-condensation reactions were stopped
p-Nitrophenyl a-L-arabinofuranosyl-(1,2)-a-L-arabinofuranosyl-
(1,2)-a-L-arabinofuranoside (6). 1H NMR (600 MHz, D2O): d =
8.11 (2H, d, J = 9.6, Hm C6H4), 7.08 (2H, d, J = 9.6, Ho C6H4),
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 2092–2102 | 2099
©

View Article Online
5.82 (1H, d, J1a,2a = 1.8, 1a-H), 5.19 (1H, d, J1b,2b = 1.2, 1b-H), 5.02
(1H, d, J1c,2c = 1.8, 1c-H), 4.30 (1H, dd, J2a,3a = 3.6, 2a-H), 4.12
(1H, dd, J3a,4a = 6.0, 3a-H), 4.03 (1H, ddd, J4a,5a = 3.0, J4a,5¢a = 5.4,
4a-H), 4.00 (1H, dd, J2b,3b = 3.6, 2b-H), 3.97 (1H, dd, J2c,3c = 3.0,
(1H, dd, 6¢b-H) ppm. 13C NMR (125 MHz, D2O): d = 161.6 (Cipso
C6H4), 142.6 (Cp C6H4), 126.7 (Cm C6H4), 117.3 (Co C6H4), 108.4
(1b-C), 105.1 (1a-C), 88.5 (2a-C), 83.5 (4b-C), 83.2 (4a-C), 82.0
(2b-C), 77.2 (3b-C), 75.4 (3a-C), 70.9 (5b-C), 70.5 (5a-C), 63.3
(6b-C), 63.2 (6a-C) ppm. MS (ESI): m/z calcd for C18H25NNaO13
[M + Na]+ 486.1224; found 486.2127.
2c-H), 3.92 (2H, m, 3b-H, 4c-H), 3.86 (1H, ddd, J3b,4b = 3.6, J4b,5b
=
3.0, J4b,5¢b = 5.4, 4b-H), 3.80 (1H, dd, J3c,4c = 6.0, 3c-H), 3.68 (2H,
m, 5a-H, 5c-H), 3.56 (3H, m, 5¢a-H, 5b-H, 5¢c-H), 3.47 (1H, dd,
J5b,5¢b = 12.0, 5¢b-H) ppm. 13C NMR (125 MHz, D2O): d = 161.3
(Cipso C6H4), 142.3 (Cp C6H4), 126.1 (Cm C6H4), 116.7 (Co C6H4),
107.7 (1c-C), 106.2 (1b-C), 104.5 (1a-C), 88.0 (2b-C), 87.2 (2a-C),
84.4 (4a-C), 84.2 (3b-C), 82.8 (4b-C), 81.2 (2c-C), 76.5 (3c-C),
75.0 (4c-C), 74.5 (3a-C), 66.8 (5c-C), 60.6 (5a-C, 5b-C) ppm. MS
(ESI): m/z calcd for C21H29NNaO15 [M + Na]+ 558.1435; found
558.2307.
p-Nitrophenyl b-D-galactofuranosyl-(1,3)-b-D-galactofuranoside
(10). 1H NMR (600 MHz, D2O): d = 8.10 (2H, d, J = 9.6, Hm
C6H4), 7.08 (2H, d, J = 9.6, Ho C6H4), 5.69 (1H, d, J1a,2a = 1.2,
1a-H), 5.05 (1H, d, J1b,2b = 1.8, 1b-H), 4.41 (1H, dd, J2a,3a = 3.0,
2a-H), 4.14 (1H, dd, J3a,4a = 6.0, 3a-H), 4.06 (1H, dd, J4a,5a = 3.6,
4a-H), 3.99 (1H, dd, J2b,3b = 3.6, 2b-H), 3.93 (1H, dd, J3b,4b = 6.2,
3b-H), 3.80 (1H, dd, J4b,5b = 4.2, 4b-H), 3.79 (1H, ddd, 5a-H),
3.68 (1H, ddd, J5b,6¢b = 7.8, J5b,6¢b = 4.8, 5b-H), 3.50 (4H, m, 6a-H,
6¢a-H, 6b-H, 6¢b-H) ppm. 13C NMR (125 MHz, D2O): d = 161.6
(Cipso C6H4), 142.3 (Cp C6H4), 126.0 (Cm C6H4), 116.8 (Co C6H4),
107.2 (1b-C), 105.6 (1a-C), 83.3 (4a-C), 82.9 (4b-C), 82.6 (3a-C),
81.3 (2b-C), 79.5 (2a-C), 76.6 (3b-C), 70.5 (5b-C), 70.4 (5a-C), 62.7
(6b-C, 6a-C) ppm. MS (ESI): m/z calcd for C18H25NNaO13 [M +
Na]+ 486.1224; found 486.2245.
p-Nitrophenyl a-L-arabinofuranosyl-(1,5)-a-L-arabinofuranosyl-
(1,5)-a-L-arabinofuranoside (7). 1H NMR (600 MHz, D2O): d =
8.11 (2H, d, J = 9.6, Hm C6H4), 7.08 (2H, d, J = 9.6, Ho C6H4),
5.72 (1H, d, J1a,2a = 1.2, 1a-H), 4.93 (1H, d, J1b,2b = 1.2, 1b-H), 4.92
(1H, d, J1c,2c = 1.2, 1c-H), 4.27 (1H, dd, J2a,3a = 3.0, 2a-H), 4.13
(1H, ddd, J4a,5a = 3.0, J4a,5¢a = 5.4, 4a-H), 4.04 (1H, ddd, J4b,5b
=
3.0, J4b,5¢b = 6.0, 4b-H), 3.99 (1H, dd, J3a,4a = 6.0, 3a-H), 3.96 (2H,
m, 2c-H, 2b-H), 3.93 (1H, ddd, J3c,4c = 3.6, J4c,5c = 3.0, J4c,5¢c = 6.0,
4c-H), 3.84 (1H, dd, J2b,3b = 3.0, J3b,4b = 6.0, 3b-H), 3.78 (1H, dd,
J2c,3c = 3.0, J3c,4c = 6.0, 3c-H), 3.76 (1H, dd, J5a,5¢a = 11.8, 5¢a-H),
3.67 (2H, m, 5a-H, 5c-H), 3.64 (2H, m, 5b-H, 5¢b-H), 3.54 (1H,
dd, J5a,5b = 12.6, 5¢c-H) ppm. 13C NMR (125 MHz, D2O): d =
161.3 (Cipso C6H4), 142.3 (Cp C6H4), 126.1 (Cm C6H4), 116.2 (Co
C6H4), 107.6 (1b-C), 107.4 (1c-C), 105.4 (1a-C), 83.9 (4c-C), 83.7
(4a-C), 82.4 (4b-C), 80.8 (2a-C, 2b-C, 2c-C), 76.5 (3a-C, 3b-C, 3c-
C), 66.7 (5a-C, 5b-C), 61.1 (5c-C) ppm. MS (ESI): m/z calcd for
C21H29NNaO15[M+Na]+ 558.1435; found 558.2434.
p-Nitrophenyl b-D-galactofuranosyl-(1,5)-b-D-galactofuranoside
(11). 1H NMR (600 MHz, D2O): d = 8.10 (2H, d, J = 9.0, Hm
C6H4), 7.08 (2H, d, J = 9.0, Ho C6H4), 5.65 (1H, d, J1a,2a = 1.8,
1a-H), 5.06 (1H, d, J1b,2b = 1.2, 1b-H), 4.25 (1H, dd, J2a,3a = 3.6,
2a-H), 4.09 (1H, dd, J3a,4a = 6.6, 3a-H), 4.03 (1H, dd, J4a,5a
=
3.6, 4a-H), 3.98 (1H, dd, J2b,3b = 3.6, 2b-H), 3.92 (2H, m, 3b-H,
4b-H), 3.82 (1H, ddd, J5a,6a = 7.4, J5a,6¢a = 5.4, 5a-H), 3.68 (1H,
ddd, J5b,6¢b = 7.8, J5b,6b = 4.2, 5b-H), 3.60 (2H, m, 6a-H, 6¢a-H),
3.56 (1H, dd, J6b,6¢b = 12.0, 6b-H), 3.50 (1H, dd, 6¢b-H) ppm. 13
C
NMR (125 MHz, D2O): d = 161.6 (Cipso C6H4), 142.3 (Cp C6H4),
126.0 (Cm C6H4), 116.8 (Co C6H4), 107.0 (1b-C), 105.2 (1a-C), 83.2
(4a-C), 82.7 (4b-C or 3b-C), 81.2 (2a-C, 2b-C), 76.5 (3a-C), 75.9
(3b-C or 4b-C), 75.1 (5a-C), 70.5 (5b-C), 62.7 (6b-C), 60.8 (6a-C)
ppm. MS (ESI): m/z calcd for C18H25NNaO13 [M + Na]+ 486.1224;
found 486.3852.
p-Nitrophenyl a-L-arabinofuranosyl-(1,5)-[a-L-arabinofurano-
syl-(1,2)]-a-L-arabinofuranoside (8). 1H NMR (600 MHz, D2O):
d = 8.11 (2H, d, J = 9.6, Hm C6H4), 7.08 (2H, d, J = 9.6, Ho
C6H4), 5.82 (1H, d, J1a,2a = 1.2, 1a-H), 5.08 (1H, d, J1c,2c = 1.2,
1c-H), 4.91 (1H, d, J1b,2b = 1.2, 1b-H), 4.29 (1H, dd, J2a,3a
=
p-Nitrophenyl b-D-galactofuranosyl-(1,6)-b-D-galactofuranoside
(12). 1H NMR (600 MHz, D2O): d = 8.09 (2H, d, J = 9.0, Hm
C6H4), 7.07 (2H, d, J = 9.0, Ho C6H4), 5.66 (1H, d, J1a,2a = 1.8,
1a-H), 4.82 (1H, d, J1b,2b = 1.8, 1b-H), 4.25 (1H, dd, J2a,3a = 3.6,
2a-H), 4.06 (1H, dd, J3a,4a = 6.0, 3a-H), 3.97 (1H, dd, J4a,5a = 3.6,
4a-H), 3.90 (1H, dd, J2b,3b = 3.0, 2b-H), 3.85 (2H, m, 3b-H, 5a-H),
3.0, 2a-H), 4.15 (2H, m, 3a-H, 4a-H), 4.00 (1H, dd, J2c,3c = 3.0,
2c-H), 3.95 (1H, dd, J2b,3b = 3.0, 2b-H), 3.90 (1H, ddd, J3c,4c = 3.6,
J4c,5c = 3.0, J4c,5¢c = 6.0, 4c-H), 3.88 (1H, ddd, J4b,5b = 3.0, J4b,5¢b
=
6.0, 4b-H), 3.78 (2H, m, 3b-H, 3c-H), 3.75 (1H, dd, J4a,5a = 4.8,
J5a,5¢a = 12.0, 5a-H), 3.66 (2H, m, 5c-H, 5¢a-H), 3.54 (2H, m,
5b-H, 5¢c-H), 3.46 (1H, dd, J4b,5¢b = 12.0, J5b,5¢b = 12.0, 5¢b-H). 13
C
3.71 (1H, dd, J4b,5b = 4.8, 4b-H), 3.66 (1H, dd, J6a,6¢a = 10.8, J5a,6¢a
=
NMR (125 MHz, D2O): d = 161.3 (Cipso C6H4), 142.3 (Cp C6H4),
126.1 (Cm C6H4), 116.7 (Co C6H4), 107.5 (1c-C), 107.4 (1b-C),
104.4 (1a-C), 87.2 (2a-C), 84.1 (4b-C), 83.9 (4c-C), 82.5 (4a-C),
81.4 (2c-C), 81.0 (2b-C), 76.6 (3b-C), 76.5 (3c-C), 74.7 (3a-C),
66.0 (5a-C), 61.1 (5c-C), 61.0 (5b-C) ppm. MS (ESI): m/z calcd
for C21H29NNaO15 [M + Na]+ 558.1435; found 558.1612.
4.2, 6a¢-H), 3.59 (1H, ddd, J5b,6¢b = 7.2, J5b,6b = 4.2, 5b-H), 3.42
(3H, m, 6a-H, 6b-H, 6¢b-H) ppm. 13C NMR (125 MHz, D2O):
d = 161.4 (Cipso C6H4), 142.2 (Cp C6H4), 126.1 (Cm C6H4), 116.7
(Co C6H4), 107.7 (1b-C), 105.3 (1a-C), 84.0 (4a-C), 83.0 (4b-C),
81.2 (2a-C), 81.0 (2b-C), 76.7 (3b-C), 76.5 (3a-C), 70.7 (5b-C),
68.9 (5a-C), 68.6 (6a-C), 62.7 (6b-C) ppm. MS (ESI): m/z calcd
for C18H25NNaO13 [M + Na]+ 486.1224; found 486.1533.
p-Nitrophenyl b-D-galactofuranosyl-(1,2)-b-D-galactofuranoside
(9). 1H NMR (600 MHz, D2O): d = 8.10 (2H, d, J = 9.0, Hm
C6H4), 7.08 (2H, d, J = 9.0, Ho C6H4), 5.73 (1H, d, J1a,2a = 1.8,
1a-H), 5.06 (1H, d, J1b,2b = 1.2, 1b-H), 4.27 (1H, dd, J2a,3a = 4.8,
2a-H), 4.20 (1H, dd, J3a,4a = 7.2, 3a-H), 3.99 (1H, dd, J2b,3b = 3.6,
2b-H), 3.96 (1H, dd, J4a,5a = 3.6, 4a-H), 3.93 (1H, dd, J3b,4b = 6.2,
3b-H), 3.79 (1H, dd, J4b,5b = 3.4, 4b-H), 3.72 (1H, ddd, J5a,6a = 7.4,
J5a,6¢a = 5.0, 5a-H), 3.60 (1H, ddd, J5b,6¢b = 7.8, J5b,6b = 4.2, 5b-H),
3.50 (2H, m, 6a-H, 6¢a-H), 3.24 (1H, dd, J6b,6¢b = 12.0, 6b-H), 3.17
p-Nitrophenyl b-D-galactofuranosyl-(1,2)-b-D-galactofuranosyl-
(1,2)-b-D-galactofuranoside (13). 1H NMR (600 MHz, D2O): d =
8.12 (2H, d, J = 9.0, Hm C6H4), 7.10 (2H, d, J = 9.0, Ho C6H4),
5.74 (1H, d, J1a,2a = 1.2, 1a-H), 5.16 (1H, d, J1b,2b = 1.8, 1b-H), 5.02
(1H, d, J1c,2c = 1.8, 1c-H), 4.27 (1H, dd, J2a,3a = 2.4, 2a-H), 4.20
(1H, dd, J3a,4a = 6.0, 3a-H), 4.07 (1H, dd, J3b,4b = 6.0, 3b-H), 4.02
(1H, dd, J2b,3b = 3.6, 2b-H), 3.97 (1H, dd, J2c,3c = 3.6, 2c-H), 3.95
(2H, m, 3c-H, 4a-H), 3.86 (1H, dd, J4c,5c = 4.8, 4c-H), 3.80 (1H,
2100 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
dd, J4b,5b = 4.8, 4b-H), 3.71 (1H, ddd, 5a-H), 3.55 (6H, m, 5b-H,
5c-H, 6a-H, 6¢a-H, 6b-H, 6¢b-H) 3.22 (1H, dd, J6c,6¢c = 12.0, 6c-H),
3.18 (1H, dd, 6¢c-H) ppm. 13C NMR (125 MHz, D2O): d = 161.4
(Cipso C6H4), 142.2 (Cp C6H4), 126.1 (Cm C6H4), 116.6 (Co C6H4),
107.5 (1c-C), 106.7 (1b-C), 104.5 (1a-C), 88.1 (2a-C, 2b-C), 83.2
(4c-C), 82.5 (4a-C), 81.8 (4b-C), 81.3 (2c-C), 76.7 (3c-C), 75.2 (3b-
C), 74.7 (3a-C), 70.7 (5a-C), 70.0 (5b-C), 69.9 (5c-C), 62.7 (6a-C,
6b-C, 6c-C) ppm. MS (ESI): m/z calcd for C24H35NNaO18 [M +
Na]+ 648.1752; found 648.1912.
supernatants were collected and cytokine production was assessed
by TNF-a-specific ELISA (Mouse TNF-a ELISA Ready-SET-Go
kit, eBioscience).
Acknowledgements
I.C. thanks the Czech Ministry of Education for a grant
(MSM:6046137305). Financial support by the Ministe`re Franc¸ais
des Affaires Etrange`res, the Programme Hubert Curien
(Barrande), the Agence Nationale de la Recherche (JCJC06_
p-Nitrophenyl b-D-galactofuranosyl-(1,2)-b-D-galactofuranosyl-
(1,3)-b-D-galactofuranoside (14). 1H NMR (600 MHz, D2O): d =
8.12 (2H, d, J = 9.6, Hm C6H4), 7.10 (2H, d, J = 9.6, Ho C6H4),
5.72 (1H, d, J1a,2a = 1.8, 1a-H), 5.17 (1H, d, J1b,2b = 1.8, 1b-H),
5.01 (1H, d, J1c,2c = 1.8, 1c-H), 4.44 (1H, dd, J2a,3a = 2.4, 2a-H),
4.15 (1H, dd, J3a,4a = 6.0, 3a-H), 4.08 (3H, m, 3b-H, 4a-H, 4b-H),
4.01 (1H, dd, J2b,3b = 3.6, 2b-H), 3.97 (1H, dd, J2c,3c = 3.6, 2c-H),
3.93 (1H, dd, J3c,4c = 6.0, 3c-H), 3.86 (1H, dd, J4c,5c = 4.8, 4c-H),
3.80 (2H, m, 5a-H, 5b-H), 3.71 (2H, m, 5c-H, 6a-H), 3.61 (5H,
m, 6¢a-H, 6b-H, 6¢b-H, 6c-H, 6¢c-H) ppm. 13C NMR (125 MHz,
D2O): d = 161.4 (Cipso C6H4), 142.2 (Cp C6H4), 126.1 (Cm C6H4),
116.8 (Co C6H4), 107.7 (1c-C), 105.9 (1a-C), 105.7 (1b-C), 88.2
(2b-C), 83.6 (4a-C, 4b-C), 83.3 (4c-C), 82.5 (3a-C), 81.2 (2c-C),
79.4 (2a-C), 76.6 (3c-C), 75.4 (3b-C), 70.7 (5b-C), 70.6 (5c-C),
70.2 (5a-C), 62.8 (6a-C, 6b-C, 6c-C) ppm. MS (ESI): m/z calcd for
C24H35NNaO18 [M + Na]+ 648.1752; found 648.2551.
140075) and the GDR-RDR2 “Aller vers une chimie eco-
compatible” are gratefully acknowledged, as well as Prof G. Davies
for supplying the plasmid of Araf 51.
´
Notes and references
1 P. Peltier, R. Euzen, R. Daniellou, C. Nugier-Chauvin and V. Ferrie`res,
Carbohydr. Res., 2008, 343, 1897–1923.
2 P. J. Brennan, Tuberculosis, 2003, 83, 91–97.
3 F. Pan, M. Jackson, Y. Ma and M. McNeil, J. Bacteriol., 2001, 183,
3991–3998.
4 D. Chatterjee and K.-H. Khoo, Glycobiology, 1998, 8, 113–120.
5 C. E. Barry, Biochem. Pharmacol., 1997, 54, 1165–1172.
6 J. B. Houseknecht and T. L. Lowary, Curr. Opin. Chem. Biol., 2001, 5,
677–682.
7 P. Peltier, R. Daniellou, C. Nugier-Chauvin and V. Ferrie`res, Org. Lett.,
2007, 9, 5227–5230.
8 G. Eppe, P. Peltier, R. Daniellou, C. Nugier-Chauvin, V. Ferrie`res and
S. P. Vincent, Bioorg. Med. Chem. Lett., 2009, 19, 814–816.
9 P. Peltier, J.-P. Gue´gan, R. Daniellou, C. Nugier-Chauvin and V.
Ferrie`res, Eur. J. Org. Chem., 2008, 5988–5994.
10 S. Notermans, G. H. Veeneman, C. W. E. M. Van Zuylen, P. Hoogerhout
and J. H. Van Boom, Mol. Immunol., 1988, 25, 975–979.
11 E. A. Leitao, V. C. B. Bittencourt, R. M. T. Haido, A. P. Valente, J.
Peter-Katalinic, M. Letzel, L. M. de Souza and E. Barreto-Bergter,
Glycobiology, 2003, 13, 681–692.
12 M. L. Richards and T. L. Lowary, ChemBioChem, 2009, 10, 1920–1938.
13 Y. J. Lee, D. B. Fulse and K. S. Kim, Carbohydr. Res., 2008, 343, 1574–
1584.
14 A. K. Pathak, V. Pathak, L. Seitz, J. A. Maddry, S. S. Gurcha, G. S.
Besra, W. J. Suling and R. C. Reynolds, Bioorg. Med. Chem., 2001, 9,
3129–3143.
p-Nitrophenyl b-D-galactofuranosyl-(1,6)-[b-D-galactofuranosyl-
(1→2)]-b-D-galactofuranoside (15). 1H NMR (600 MHz, D2O):
d = 8.12 (2H, d, J = 9.0, Hm C6H4), 7.10 (2H, d, J = 9.0, Ho C6H4),
5.74 (1H, d, J1a,2a = 1.8, 1a-H), 5.05 (1H, d, J1c,2c = 1.8, 1c-H), 4.82
(1H, d, J1b,2b = 1.8, 1b-H), 4.26 (1H, dd, J2a,3a = 2.4, 2a-H), 4.20
(1H, dd, J3a,4a = 6.0, 3a-H), 3.98 (1H, dd, J2c,3c = 3.6, 2c-H), 3.97
(1H, dd, J4a,5a = 3.6, 4a-H), 3.91 (1H, dd, J3c,4c = 6.0, 3c-H), 3.89
(1H, dd, J2b,3b = 3.6, 2b-H), 3.85 (2H, m, 3b-H, 5a-H), 3.78 (1H,
dd, J4c,5c = 4.8, 4c-H), 3.71 (1H, dd, J4b,5b = 4.8, 4b-H), 3.65 (1H,
dd, J6a,6¢a = 10.2, J5a,6a = 4.8, 6a-H), 3.60 (2H, m, 5b-H, 5c-H),
3.49 (1H, m, 6b-H), 3.45 (1H, m, 6¢a-H), 3.39 (1H, m, 6¢b-H), 3.19
(1H, dd, J6c,6¢c = 12.0, J5c,6c = 4.2, 6c-H), 3.19 (1H, dd, J5c,6¢c = 7.8,
6¢c-H) ppm. 13C NMR (125 MHz, D2O): d = 161.4 (Cipso C6H4),
142.2 (Cp C6H4), 126.1 (Cm C6H4), 116.6 (Co C6H4), 107.9 (1c-C),
107.6 (1b-C), 104.4 (1a-C), 88.1 (2a-C), 83.0 (4b-C), 82.9 (4c-C),
82.3 (4a-C), 81.4 (2c-C), 80.9 (2b-C), 76.7 (3b-C), 76.5 (3c-C), 74.7
(3a-C), 70.7 (5c-C), 70.2 (5b-C), 68.5 (6a-C), 68.4 (5a-C), 62.7 (6b-
C, 6c-C) ppm. MS (ESI): m/z calcd for C24H35NNaO18 [M + Na]+
648.1752; found 648.1688.
15 G. Zhang, M. Fu and J. Ning, Tetrahedron: Asymmetry, 2005, 16, 733–
738.
16 J. T. Smoot, P. Pornsuriyasak and A. V. Demchenko, Angew. Chem.,
Int. Ed., 2005, 44, 7123–7126.
17 Y. Bai and T. L. Lowary, J. Org. Chem., 2006, 71, 9672–9680.
18 M. Fu, G. Zhang and J. Ning, Carbohydr. Res., 2005, 340, 25–30.
19 L. Gandolfi-Donadio, C. Gallo-Rodriguez and R. M. de Lederkremer,
Carbohydr. Res., 2008, 343, 1870–1875.
20 N. L. Rose, G. C. Completo, S. J. Lin, M. McNeil, M. M. Palcic and
T. L. Lowary, J. Am. Chem. Soc., 2006, 128, 6721–6729.
21 C. Remond, R. Plantier-Royon, N. Aubry, E. Maes, C. Bliard and M. J.
O’Donohue, Carbohydr. Res., 2004, 339, 2019–2025.
22 C. Remond, R. Plantier-Royon, N. Aubry and M. J. O’Donohue,
Carbohydr. Res., 2005, 340, 637–644.
Immunostimulation assay
23 R. Euzen, G. Lopez, C. Nugier-Chauvin, V. Ferrie`res, D. Plusquellec,
C. Re´mond and M. O. Donohue, Eur. J. Org. Chem., 2005, 4860–4869.
24 G. Lopez, C. Nugier-Chauvin, C. Re´mond and M. O’Donohue,
Carbohydr. Res., 2007, 342, 2202–2211.
25 E. J. Taylor, N. L. Smith, J. P. Turkenburg, S. D’Souza, H. J. Gilbert
and G. J. Davies, Biochem. J., 2006, 395, 31–37.
26 G. Pae¨s, L. K. Skov, M. J. O’Donohue, C. Re´mond, J. S. Kastrup, M.
Gajhede and O. Mirza, Biochemistry, 2008, 47, 7441–7451.
27 M. Mayer and B. Meyer, Angew. Chem., Int. Ed., 1999, 38, 1784–1788.
28 S. Mala´, H. Dvora´kova´, R. Hrabal and B. Kra´lova´, Carbohydr. Res.,
1999, 322, 209–218.
29 M. Joe, D. Sun, H. Taha, G. C. Completo, J. E. Croudace, D. A.
Lammas, G. S. Besra and T. L. Lowary, J. Am. Chem. Soc., 2006, 128,
5059–5072.
Stock solutions of oligosaccharides were prepared in Iscove’s
Modified Dulbecco’s medium (IMDM) (Gibco). LPS (E. coli
0127:B8, Sigma) was used as a positive control, negative controls
had no oligosaccharide added. The murine macrophage cell line
RAW264.7 (obtained from ATCC, USA) was maintained in
IMDM supplemented with 10% fetal bovine serum (FBS) (Gibco)
at 37 ◦C and 5% CO2. Cells were harvested by centrifugation and
resuspended in the same media to a final concentration of 5 ¥
105 cells mL-1. Thereafter, 5 ¥ 104 RAW264.7 cells in 100 mL of
IMDM with 10% FBS per well were incubated with the indicated
concentration of oligosaccharides at 37 ◦C and 5% CO2. After 18 h,
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 2092–2102 | 2101
©

View Article Online
30 A. Tzianabos, J. Y. Wang and D. L. Kasper, Carbohydr. Res., 2003, 338,
2531–2538.
31 F. Dourado, P. Madureira, V. Carvalho, R. Coelho, M. A. Coimbra,
M. Vilanova, M. Mota and F. M. Gama, Carbohydr. Res., 2004, 339,
2555–2566.
39 M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb,
J. R. Cheeseman, J. J. A. Montgomery, T. Vreven, K. N. Kudin, J. C.
Burant, J. M. Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci,
M. Cossi, G. Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M.
Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T.
Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X. Li, J. E.
Knox, H. P. Hratchian, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo,
R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi,
C. Pomelli, J. W. Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth,
P. Salvador, J. J. Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D.
Daniels, M. C. Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S.
Clifford, J. Cioslowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz,
I. Komaromi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y.
Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson,
W. Chen, M. W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Inc.,
Wallingford CT, 2004.
32 Q. Dong, J. Yao, J.-n. Fang and K. Ding, Carbohydr. Res., 2007, 342,
1343–1349.
33 R. P. McGeary, C. Olive and I. Toth, J. Pept. Sci., 2003, 9, 405–
418.
34 D. Van Der Spoel, E. Lindahl, B. Hess, G. Groenhof, A. E.
Mark and H. J. Berendsen, J. Comput. Chem., 2005, 26, 1701–
1718.
35 V. Hornak and C. Simmerling, Proteins: Struct., Funct., Genet., 2003,
51, 577–590.
36 J. C. Gordon, J. B. Myers, T. Folta, V. Shoja, L. S. Heath and A.
Onufriev, Nucleic Acids Res., 2005, 33, W368–371.
37 K. N. Kirschner, A. B. Yongye, S. M. Tschampel, J. Gonza´lez-Outeirin˜o,
C. R. Daniels, B. L. Foley and R. J. Woods, J. Comput. Chem., 2008,
29, 622–655.
40 http://chemistry.csulb.edu/ffamber/tools.html.
41 H. J. C. Berendsen, J. P. M. Postma, W. van Gunsteren and J. Hermans,
in In Intermolecular Forces, ed. B. Pullman, Reidel, Dordrecht, 1981,
pp. 331-342.
38 C. I. Bayly, P. Cieplak, W. Cornell and P. A. Kollman, J. Phys. Chem.,
1993, 97, 10269–10280.
2102 | Org. Biomol. Chem., 2010, 8, 2092–2102
This journal is
The Royal Society of Chemistry 2010
©