Importance of topology for glycocluster binding to Pseudomonas aeruginosa and Burkholderia ambifaria bacterial lectins

Article abstract of DOI:10.1039/c5ob01445j

Pseudomonas aeruginosa (PA) and Burkholderia ambifaria (BA) are two opportunistic Gram negative bacteria and major infectious agents involved in lung infection of cystic fibrosis patients. Both bacteria can develop resistance to conventional antibiotherap

Full text of DOI:10.1039/c5ob01445j

Organic &
Biomolecular Chemistry
View Article Online
View Journal
PAPER
Importance of topology for glycocluster binding
to Pseudomonas aeruginosa and
Cite this: DOI: 10.1039/c5ob01445j
Burkholderia ambifaria bacterial lectins†
Caroline Ligeour,a Lucie Dupin,b Anthony Angeli,a Gérard Vergoten,d
Sébastien Vidal,c Albert Meyer,a Eliane Souteyrand,b Jean-Jacques Vasseur,a
Yann Chevolot*b and François Morvan*a
Pseudomonas aeruginosa (PA) and Burkholderia ambifaria (BA) are two opportunistic Gram negative bac-
teria and major infectious agents involved in lung infection of cystic fibrosis patients. Both bacteria can
develop resistance to conventional antibiotherapies. An alternative strategy consists of targeting virulence
factors in particular lectins with high affinity ligands such as multivalent glycoclusters. LecA (PA-IL) and
LecB (PA-IIL) are two tetravalent lectins from PA that recognise galactose and fucose respectively. BambL
lectin from BA is trimeric with 2 binding sites per monomer and is also specific for fucose. These three
lectins are potential therapeutic targets in an anti-adhesive anti-bacterial approach. Herein, we report the
synthesis of 18 oligonucleotide pentofuranose-centered or mannitol-centered glycoclusters leading to
tri-, penta- or decavalent clusters with different topologies. The linker arm length between the core and
the carbohydrate epitope was also varied leading to 9 galactoclusters targeting LecA and 9 fucoclusters
targeting both LecB and BambL. Their dissociation constants (Kd) were determined using a DNA-based
carbohydrate microarray technology. The trivalent xylo-centered galactocluster and the ribo-centered
fucocluster exhibited the best affinity for LecA and LecB respectively while the mannitol-centered
decafucocluster displayed the best affinity to BambL. These data demonstrated that the topology and
nature of linkers were the predominant factors for achieving high affinity rather than valency.
Received 15th July 2015,
Accepted 18th September 2015
DOI: 10.1039/c5ob01445j
www.rsc.org/obc
site respectively. LecA is galactose specific with an association
Introduction
constant (Ka) of 3.4 × 104 M−1 while LecB is fucose specific
Pseudomonas aeruginosa (PA) and Burkholderia ambifaria (BA) with a Ka of 1.6 × 106 M−1. LecA has a cobblestone shape
are opportunistic pathogens involved in nosocomial diseases where the binding sites are located at the corner of the rec-
and they especially affect cystic fibrosis patients. PA has two tangle (32 and 70 Å away on the width and length, respect-
soluble lectins LecA and LecB involved in its virulence.1 Both ively).2,3 Due to this configuration, several studies suggest that
lectins are homotetrameric with one binding site per galactoclusters can reach simultaneously two binding sites
monomer and one or two calcium ions involved in the binding along the small side of one tetramer leading to a glycoside
cluster effect.3,4 LecB has more or less a tetrahedral spatial dis-
tribution of the binding sites that are more difficult to reach
aInstitut des Biomolécules Max Mousseron (IBMM), UMR 5247 CNRS, Université de
simultaneously and very few glycoside cluster effects have been
Montpellier, ENSCM, place Eugène Bataillon, CC1704, 34095 Montpellier Cedex 5,
reported so far with LecB.5,6 Furthermore, its affinity for mono-
France. E-mail: morvan@univ-montp2.fr
bUniversité de Lyon, Institut des Nanotechnologies de Lyon UMR CNRS 5270,
valent fucose is already high with a Kd in the micromolar
range.1,2
Ecole Centrale de Lyon, 36 avenue Guy de Collongue, 69134 Ecully, France.
E-mail: yann.chevolot@ec-lyon.fr
cInstitut de Chimie et Biochimie Moléculaires et Supramoléculaires (ICBMS),
BA expresses a fucose-binding lectin called BambL. Simi-
larly to LecB, it has a very high affinity for fucose (a Kd of
Laboratoire de Chimie Organique 2 – Glycochimie, UMR 5246, CNRS, Université
0.96 µM).7 It is a crown shaped trimeric lectin with two
Claude Bernard Lyon 1, 43 Boulevard du 11 Novembre 1918, 69622 Villeurbanne,
France
binding sites per monomer. The six binding sites are on the
same side of the protein with a inter binding site distance
ranking between 18 and 36 Å. So far very little fucoclusters tar-
dUnité de Glycobiologie Structurelle et Fonctionnelle (UGSF), UMR 8576 Université
de Lille 1, Cité Scientifique Avenue Mendeleiev, Bat C9, 59655 Villeneuve d’Ascq
cedex, France
geting BambL have been synthesized and no glycoside cluster
†Electronic supplementary information (ESI) available: NMR data and HPLC of
effect has been reported.8,9 Nevertheless, its topology is favor-
glycoclusters. See DOI: 10.1039/c5ob01445j
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
able to the glycoside cluster effect with binding sites that can hydroxyl was phosphitylated using phosphoramidite chemistry
be reached by multivalent ligands with a topology in close with an alkyne linker leading to scaffolds bearing three or five
relationship with the AB5 shiga toxin complexes.
alkynes. In addition, two alkyne functions were introduced on
The use of glycoclusters exhibiting high affinity for the tar- each hydroxyl of the reduced mannosyl platform leading to a
geted lectins has emerged as a promising strategy to fight decaalkyne scaffold. Then, the conjugation between the poly-
against such pathogens.5,10 Since there is no pressure on the alkyne scaffolds and the carbohydrate azides (galactoside 16 or
bacteria this strategy should not induce a resistance pheno- fucoside 17) occurred through copper-catalyzed azide–alkyne
menon. However, the design of good candidates remains cycloaddition (CuAAC) affording the expected glycoclusters
empirical since the precise mode of recognition of multivalent exhibiting from three to ten galactose or fucose residues.
glycoclusters by their bacterial lectin partners is still not fully Finally, a DNA sequence was incorporated for immobilization
understood.
on a DNA microarray of the final glycocluster leading to a gly-
coarray. Kd values of the glycoclusters toward LecA, LecB or
BambL were determined using this glycoarray.
Results and discussion
Building block synthesis
We have previously reported the syntheses of glycoclusters
with a hexose-platform exhibiting either galactose11–13 or
fucose12,14 epitopes for interactions with LecA and LecB
respectively. To gain more insight into the effect of topology
and valency on the affinity of the corresponding glycoclusters
toward LecA and LecB, we have designed two new families of
glycoclusters. The first one used furanose-based platforms
such as arabinose, ribose and xylose (3a–c, Fig. 1) and the
second one used an acyclic azido-mannitol (10, Fig. 1). The
furanose platforms exhibit three hydroxyl functions which are
differently orientated in the space according to the nature of
the furanose to study the effect of the orientation of the 2′-
and 3′-hydroxyl groups of the furanose moiety. The mannitol
core 10 exhibits five hydroxyl functions on a flexible linear
scaffold. The impact of rigidity and multivalence on the
affinity for LecA and LecB can therefore be evaluated using
these different scaffolds. The affinity of the fucoclusters toward
BambL was also evaluated since this lectin from Burkholderia
ambifaria recognizes fucose moieties.
The different building blocks required for their further assem-
bly to obtain the desired glycoclusters were synthesized: the
three ribo, arabino and xylo-furanoside propargyl platforms
3a–c, the 6-azido-6-deoxymannitol 10, azide 1115 and alkyne
1216 solid supports, three different alkyne phosphoramidites
exhibiting one 1315 and 1412 or two 1517 alkynes and β-D-galac-
toside 1618 or α-L-fucoside 1719 azide derivatives (Fig. 1).
For the synthesis of the three different furanose scaffolds,
we used ribo-, arabino- and xylo-furanosyl uracil since it is easy
to introduce an alkyne function on the N3 of uracil by selective
alkylation.20 Commercially available ribo-1 and arabino-2
uracils were treated with propargyl bromide and potassium
carbonate in DMF at 60 °C overnight to give compounds 3a
and 3b respectively exhibiting N3-propargyl for subsequent
immobilization on the azide solid support 11 (Scheme 1).
Since xylo-furanosyl uracil is not commercial, N3-propargyl
xylo-furanosyl uracil 3c was synthesized in two steps by glycosy-
lation of the 1,2,3,5-tetra-O-acetyl-D-xylofuranose with uracil,
using bis(trimethylsilyl)acetamide and trimethylsilyl trifluoro-
methanesulfonate in acetonitrile. In accord with Baker’s
rule,21 the direct condensation of the suitably protected 2′-O-
acyl-D-xylofuranose 4 and uracil led to the formation of
β (trans-1′,2′) nucleoside 5. The resulting peracetyl xylouracil
According to our strategy, the polyhydroxylated platforms
3a–c and 10 were immobilized on a solid support and each
Scheme 1 Synthesis of ribo-, arabino- and xylo-furanosyl uracils 3a–c.
(a) BrCH2CCH, K2CO3, DMF, 60 °C, overnight; (b) N,O-Bis(trimethylsilyl)
acetamide, TMSOTf, CH3CN, reflux, 4 h 30 min; (c) MeOH, rt.
Fig. 1 Structure of building blocks used for the synthesis of the
glycoclusters.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Organic & Biomolecular Chemistry
Paper
5 was alkylated as described above. The deprotection of the under acidic conditions and reduction using sodium boro-
acetyl groups on the xylofuranoside was achieved by addition hydride23 afforded 6-azido-6-deoxy-D-mannitol 10.
of methanol in the basic medium to give 3c.
Synthesis of glycocluster oligonucleotide conjugates
The linear platform 10 was obtained from methyl α-D-
mannopyranoside 6 (Scheme 2). The primary 6-hydroxyl group
was selectively tosylated in pyridine. Then, the tosyl group was
replaced by an azide function by treatment with NaN3 to afford
8.22 Acetal 8 was converted to the corresponding hemiacetal 9
Ribo-, arabino-, xylo-furanosyl uracil-centered glycol-oligo-
nucleotides were synthesized using the building blocks
reported in Fig. 1. The first step was the immobilization of
scaffolds 3a–c on azide solid support 11 by CuAAC under
microwave assistance (Scheme 3). The resulting tri-hydroxyl
scaffolds 18a–c were then phosphorylated with alkyne phos-
phoramidites 13 or 14 using a DNA synthesizer with a triple
coupling step. Then, the oligonucleotide was elongated and
labelled with a fluorescent dye (Cy3) using phosphoramidite
chemistry. The resulting trialkyl furanoside oligonucleotide
conjugates 19a–c and 20a–c were deprotected and released
from the solid support by ammonia treatment leading to the
trialkyl furanoside oligonucleotide conjugates 21a–c and 22a–c
exhibiting different furanoside scaffolds (i.e. ribo-, arabino- or
xylo) with alkyne linkers of different lengths (pent-4-ynyl phos-
phodiester 21 or propargyl diethyleneglycyl phosphodiester 22)
in solution. Crude materials were analyzed and characterized
by HPLC and MALDI-TOF spectrometry. Around 100 nmol of
crude material was conjugated with either galactoside 16 or
fucoside 17 azide by CuAAC using copper nanopowder24 under
Scheme 2 Synthesis of 6-azido-6-deoxymannitol 10. (a) TsCl, pyridine,
22 h, 0 °C, 65%; (b) NaN3, DMF, 2 h, 80 °C, 84%; (c) HCl, H2O, dioxane,
2 h, reflux, 46%; (d) NaBH4, EtOH, 1 h, rt, 82%.
Scheme 3 Synthesis of the triglycocluster oligonucleotides G1a–c, G2a–c, F1a–c and F2a–c. (a) 3a–c, CuSO4, Na ascorbate, MeOH/H2O (1 : 1),
MW, 60 °C, 1 h; (b) Solid Phase Oligonucleotide Synthesis (SPOS) with 13 or 14, then SPOS with A, T, C, G, Cy3; (c) NH4OH; (d) Cu0, 16 or 17.
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Scheme 4 Synthesis of the penta- and deca-glycocluster oligonucleotides G3a–c and F3a–c. (a) SPOS with A, T, C; (b) 10, CuSO4, Na ascorbate,
MeOH/H2O (1 : 1), MW, 65 °C, 1 h; (c) SPOS with 3; (d) 16 or 17, CuSO4, Na ascorbate, dioxane/H2O (1 : 1), MW, 70 °C, 1 h; (e) SPOS with A, T, C, G,
Cy3, (ii) NH4OH.
microwave assistance. Finally, a short ammonia treatment was Cy3. A final treatment with ammonia led to four pentaglyco-
applied to remove the acetyl groups. The twelve resulting glyco- clusters G3a–b, F3a–b and two decaglycoclusters G3c, F3c.
clusters exhibiting three galactosides (G1a–c and G2a–c) or
The affinity (Kd) of the eighteen new glycoclusters display-
three fucosides (F1a–c and F2a–c) with different scaffolds ing 3, 5 or 10 galactose or fucose moieties with different topol-
(ribo-, arabino- and xylo-furanose) and with short (pro, G1) or ogies and conjugated to different DNA sequences (Table S1†)
long (EG2M, G2) linkers were purified by C18 HPLC.
was determined by using a microarray (Table 1). The 9 galacto-
For the mannitol-centered oligonucleotides, the synthetic clusters can only target LecA while the 9 fucoclusters can
strategy was changed since difficulties in characterizing the target both LecB and BambL. The microarray was elaborated
final glycocluster oligonucleotide conjugates were encountered by DNA Directed Immobilization (DDI).26,27 Each glycocluster
due to their high molecular weights. A short DNA sequence labelled with Cy3 and bearing a specific DNA sequence was
was synthesized on the alkyne solid support 12 and then the hybridized specifically with an immobilized complementary
mannitol azide 10 was immobilized by CuAAC with copper DNA sequence on the array.28,29 Fluorescence scanning of the
sulfate and sodium ascorbate under microwave assistance Cy3 signal allows verifying glycocluster immobilization with a
(Scheme 4). The resulting pentahydroxyl compound 28 was deviation lower than 14% illustrating similar surface densities
phosphorylated with the three different phosphoramidites for all immobilized glycoclusters. Microstructured glass slides,
11–13 with a triple coupling step to obtain respectively 29a–c. bearing 40 microwells, were used as an array of 64 spots/micro-
Both galactoside and fucoside azide derivatives were conju- well allowing 40 independent experiments on one single slide.
gated on a solid support by CuAAC using CuSO4 and sodium In each microwell, two glycoclusters to be tested were immobi-
ascorbate to afford 30a–c and 31a–c respectively. A few beads lized and incubated with increasing concentrations (one con-
were withdrawn and treated with ammonia for deprotection centration per well) of Alexa-647-labelled lectin (LecA, LecB or
and release. After evaporation, the sample was analyzed by BambL). Scanning at 635 nm provided the binding isotherms
HPLC and MALDI-TOF showing the efficient introduction of for each glycocluster/lectin binding event. Kd values were
carbohydrate moieties on the polyalkyne scaffolds. Then, the obtained using the linear regression described in the Experi-
remaining oligonucleotide was elongated and labeled with mental section.30 The background noise of the Alexa-647 fluo-
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Organic & Biomolecular Chemistry
Paper
brings an additional benefit allowing high binding with only
trivalent clusters.
Table 1 Kd values of the DNA-glycoclusters towards LecA, LecB or
BambL determined by using a DNA directed immobilization glycoarray25
In the case of LecB, glycoclusters with ribo and xylo cores
exhibited the best affinities (56 and 62 nM). A stronger affinity
was found with the shortest Pro linker (F1a and F1c) than with
the EG2M linker (F2a and F2c). Surprisingly, for the glycoclus-
ters with an arabinose core (F1b–F2b) or with the mannitol
core (F3a–F3b), the effect of the linker on the binding was not
significant. Interestingly, the glycoclusters with an arabinose
core were less prone to bind LecB. Finally the decavalent
cluster (F3c) bound LecB better than the pentavalent clusters
but lower than the trivalent clusters. One can imagine that the
rigidity of the furanose cores has an entropic benefit in the
interaction of the corresponding clusters leading to better
binding.
For BambL, all fucoclusters displayed strong binding with
only a few differences between them (Kd from 13.8 to 29.6 nM).
The increase of the length of the linker from Pro to EG2M was
slightly detrimental for the furanose cores in contrast to the
mannitol ones. Ribo and xylo-centered clusters were found to
be better ligands toward BambL than toward arabino-centered
ones. Interestingly, the decafucocluster (F3c) displayed the
strongest affinity (Kd = 13.8 nM) and this effect can be
explained by the fact that the CRDs of BambL are localized on
the same side leading to easy and simultaneous (multivalent)
access to several fucoses of the cluster in a “glycoside cluster
effect” fashion.
In order to understand why glycoclusters with an arabino-
furanose core have a higher dissociation constant (Kd) against
LecA and LecB compared to ribofuranose and xylofuranose,
preliminary molecular modeling studies have been performed.
For each lectin, the glycoclusters exhibiting the best affinity
have been chosen: G2a, G2b G2c and F1a, F1b F1c. Quantum
semi empirical calculations using the PM6 Hamiltonian39 and
the solvation model SM5.4/A40 showed that β-D-ribofuranose,
β-D-arabinofuranose and β-D-xylofuranose exhibit a very similar
heat of formation both under vacuum (ΔHf) and in aqueous
solution (ΔHaq) (Table 2). From a structural point of view, the
main differences among the three sugars lie in the chiralities
of carbon 2 and carbon 3 of the furanose ring (Table 2) while
that of carbons 1 and 4 is the same (R).
According to these findings, the differences in glycocluster
interactions with LecA and LecB are supposed to be related to
branches originating from C2 and C3. Accordingly, in silico
molecular modelling was performed and the following empiri-
cal potential energies of interaction were obtained (Table 3).
They clearly show that arabinofuranose derivatives (G2b and
F1b) are the weakest ligands as found from their Kd values.
When docking was applied with the epitopes on carbons 4 and
2 involved in the CRD, the empirical potential energies of
interaction (ΔE) were not in agreement with the Kd values.
These data strongly suggest that the interaction with lectin
occurred with epitopes on carbons C2 and C3.
Kd (nM)
Glyco-a
Clusters
LecA
LecB
BambL
G1a
G2a
G1b
G2b
G1c
G2c
G3a
G3b
G3c
F1a
F2a
F1b
F2b
F1c
F2c
F3a
F3b
F3c
RiboPro
RiboEG2M
AraPro
AraEG2M
XyloPro
XyloEG2M
Mannipro
ManniEG2M
Manni2
RiboPro
RiboEG2M
AraPro
AraEG2M
XyloPro
XyloEG2M
ManniPro
ManniEG2M
Manni2
66
54
64
62
54
49
66
78
50
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
—
56
73
81
76
62
93
91
95
84
—
—
—
—
—
—
—
—
—
20.1
21.2
25.1
29.6
20.5
26.5
22.8
18.2
13.8
a Name of the core (ribose, arabinose, xylose or mannitol) with the
nature of linker L (Pro or EG2M) and manni2 for mannitol with two
glycosides per hydroxyl.
rescence signal was very low (below 20 a.u.) illustrating that
Alexa-647-labelled lectin was not adsorbed nonspecifically on
the surface.
The 9 galactoclusters tested exhibited the same order of
magnitude of binding strength with Kd values between 49 and
78 nM for LecA. For each trimeric cluster, the hydrophilic and
flexible EG2M linker (G2a–c) seems to be slightly preferred to
the short hydrophobic Pro linker (G1a–c) especially for the
ribose core. The opposite was observed for the mannitol core
(G3a vs. G3b). It was previously observed that the length of a
linker could have a strong effect on binding. Thus dimeric
galactoclusters with a linker differing from only four methyl-
enes exhibit a four-fold difference of binding.31
Regardless of the linker or the valence (3 or 5), the xylose
core seems to be preferred. Then, with a slightly lower affinity,
arabinose and ribose scaffolds showed comparable binding
properties. The decavalent glycocluster G3c displayed a similar
Kd value (50 nM) than the best trimeric glycocluster G2c (49
nM). These data could be explained by some steric hindrances
leading to a lower accessibility of the ten galactoside motifs.
Furthermore, we have previously observed that mannose-
centered galactoclusters exhibiting four phenyl galactoside
motifs with Pro and EG2M linkers have a Kd value of 57 and
39 nM32 respectively, showing also the benefit of a longer
linker. Thus, the affinity of the xylose centered trigalactocluster
G2c to the EG2M linker exhibiting a Kd value of 49 nM is not
so different, taking into account that only three motifs are
involved in the construct.
These data showed that the “aromatic effect” has a strong
consequence on the high binding of glycoclusters, as already
observed by others and us.13,32–38 Nevertheless, the topology
In the case of LecA interaction, the difference of the ener-
gies of interaction could be tentatively explained by the posi-
tion of the galactose with the linker on carbon 4 that brought
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
Table 2 ΔHf and ΔHaq and absolute configuration of carbon 2 and 3 obtained by quantum semi empirical calculations using PM6 Hamiltonian and
SM5.4/A solvation model
Absolute configuration
Furanose
ΔHf, kcal mol−1
ΔHaq, kcal mol−1
C2
C3
β-D-Ribo
β-D-Arabino
β-D-Xylo
−231.8
−230.9
−229.7
−252.0
−251.2
−251.9
R
S
R
S
S
R
The situation of LecB is slightly different. In the case of F1a
and F1c, the free fucose, on the C4 branch, is located at an
equal distance from the two binding sites. F1a (Fig. 3a) exhi-
bits strong intramolecular hydrogen bonds, a pseudo stacking
interaction with the uracil ring and some interactions with
residues A1, Q3 and L76. The free fucose in F1c (Fig. 3b) also
takes a central position and hence displays strong interactions
with N34, K62 and Q64. In the case of F1b, the free fucose is
oriented towards the triazole moiety of branch on C2 leading
to some interactions with P73 and S74 (Fig. 3c).
Table 3 Empirical potential energy of interaction (ΔE) and experimental
dissociation constants (Kd)
Complex
ΔE (kcal mol−1
)
Kd (nM)
LecA–G2a
LecA–G2b
LecA–G2c
LecB–F1a
LecB–F1b
LecB–F1c
−255.8
−198.9
−244.4
−232.7
−192.0
−222.0
54
62
49
56
81
62
an additional stabilization by interaction with the lectin. For
LecA–G2a, the free ending galactose interacts with the R48
residue of LecA. In the case of LecA–G2c (Fig. 2a), the inter-
action takes place with P38 and T39 (Fig. 2b). Finally for LecA–
G2b, no interaction occurs (Fig. 2c).
Fig. 2 LecA in complex with (a) G2a, (b) G2c and (c) G2b.
Fig. 3 LecB in complex with (a) F1a, (b) F1c and (c) F1b.
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Organic & Biomolecular Chemistry
Paper
394A DNA synthesizer). Reactions under microwave activation
were performed on a Biotage Initiator system or an Anton Paar
Conclusion
In the present report, 18 glycoclusters were synthesized in Monowave 300 system. Solutions of Cap A, Cap B and iodide
order to target two lectins of PA (LecA and LecB) and one were purchased from Link Technologies as well as the com-
lectin from BA (BambL). These three lectins are virulence mercial solid-supports. Solutions of TCA and CH3CN for DNA
factors involved in the adhesion and/or in biofilm formation. synthesis were purchased from Biosolve. Cy3-amidite was pur-
Their binding to the 18 glycoclusters was evaluated using the chased from GE Healthcare. All glycooligonucleotides were
DDI-based glycocluster array allowing for the determination of purified and analysed by C18 reversed-phase HPLC (Macherey-
their Kd values which were below 100 nM. An in silico mole- Nagel, Nucleodur 4.6 × 75 mm, 3 mM) on a Dionex Ultimate
cular modeling study suggested that, for furanose-centered 3000 system with a Reodyn injector and a detector UV DAD
glycoclusters, epitopes borne on carbons 2 and 3 are involved 3000. Oligonucleotides were dosed by UV-Vis spectropho-
in the CRD of LecA or LecB while the free epitope on carbon 4 tometry at 260 nm on
for ribose and xylose interacts with the lectin leading to an spectrophotometer.
additional stabilization. The topology and nature of linkers
a Varian Cary 300 Bio UV-Vis
General procedure for N3-alkylation of uracyls 1 and 2
were the predominant factors for achieving high affinity rather
than valency. LecB and BambL showed different bonding pro- A solution of propargyl bromide (1.18 eq.) was added to a solu-
files toward the set of 9 fucosylated clusters suggesting that tion of furanosyl uracil 1 or 2 (1.0 eq.) and potassium carbon-
the topology of the binding sites leads to an additional ate (1.08 eq.) in DMF, and the system was stirred at 60 °C
(additional with respect to the saccharide residues) selectivity overnight. The reaction mixture was filtered and evaporated to
toward the cluster structure (topology and linker). These dryness. The crude product was filtered on silica gel column
selectivity profiles can be viewed as a fingerprint of the lectin. chromatography (5% MeOH in AcOEt) to afford the desired
These results demonstrated that high avidity of a glycocluster compounds 3a–b.
to a lectin is a delicate fine-tuning of its topology.
1-(β-D-Ribofuranosyl)-3-(prop-2-yn-1-yl)uracil 3a. Obtained
as a pale yellow solid (702 mg, 83%) following the general pro-
cedure for alkylation: uridine 1 (730 mg, 3.0 mmol), propargyl
bromide (420 mg, 3.5 mmol), potassium carbonate (450 mg,
Experimental
General methods
1
3.2 mmol), DMF (30 mL). H NMR (300 MHz, CD3OD) δ 8.08
(d, J = 8.1 Hz, 1H, H6), 5.95 (d, J = 4.2 Hz, 1H, H1′), 5.82 (d, J =
8.1 Hz, 1H, H5), 4.68 (dd, J = 2.4, 1.6 Hz, 2H, CH2CuCH),
4.22–4.16 (m, 2H, H2′, H3′), 4.06–4.03 (m, 1H, H4′), 3.76–3.92
(dd AB system, J = 9.0, 2.1 Hz, 2H, H5′), 2.56 (t, J = 2.4 Hz, 1H,
CuCH). 13C NMR (75 MHz, CD3OD) δ 163.9, 151.9, 141.3,
102.0, 91.7, 86.4, 79.1, 75.9, 71.9, 71.1, 62.2, 31.1. HR-ESI-QToF
MS (positive mode): m/z calcd for C12H15N2O6 [M + H]+
283.0930 found 283.0926.
All reagents for synthesis were commercial and used without
further purification. Solvents were commercial and used
without further purification. All moisture sensitive reactions
were performed under an argon atmosphere. NMR solvents
were purchased from Euriso-Top. NMR spectra were recorded
at 293 K using a 200, 300, 400 or a 600 MHz spectrometer
(Bruker). Shifts (δ) are referenced relative to deuterated solvent
residual peaks and expressed in parts per million (ppm).
Coupling constants are expressed in hertz (Hz). The following
abbreviations are used to explain the observed multiplicities:
s (singlet), d (doublet), t (triplet), q (quartet), dd (doublet of
doublets), m (multiplet), sbr (broad singlet). High-resolution
(HR-ESI-QToF) mass spectra were recorded using a Q-Tof
Micromass spectrometer. MALDI-TOF mass spectra were
recorded on a Voyager mass spectrometer (Perspective Biosys-
tems, Framingham, MA) equipped with a nitrogen laser. Thin-
layer chromatography (TLC) was carried out on aluminum
sheets coated with silica gel 60 F254 (Merck). TLC plates were
inspected by UV light (l = 254 nm) and developed by treatment
with a mixture of 10% H2SO4 in EtOH/H2O (1 : 1 v/v), vanillin,
KMnO4 or phosphomolybdic acid followed by heating. Silica
gel column chromatography was performed with silica gel Si
60 (40–63 mm). Reverse phase chromatography was performed
with a C18 flash column.
1-(β-D-Arabinofuranosyl)-3-(prop-2-yn-1-yl)uracil
Obtained as a pale yellow solid (761 mg, 90%) following the
general procedure for alkylation: arauridine (730 mg,
3b.
2
3.0 mmol), propargyl bromide (420 mg, 3.5 mmol), potassium
carbonate (450 mg, 3.2 mmol), DMF (30 mL). 1H NMR
(300 MHz, CD3OD) δ 7.88 (d, J = 8.1 Hz, 1H, H6), 6.16 (d, J =
4.2 Hz, 1H, H1′), 5.76 (d, J = 8.1 Hz, 1H, H5), 4.66 (dd, J = 2.4,
1.2 Hz, 2H, CH2CuCH), 4.19 (dd, J = 4.2, 2.8 Hz, 1H, H2′), 4.08
(t, J = 3.3 Hz, 1H, H3′), 3.97–3.93 (m, 1H, H4′), 3.86–3.76 (dd
JAB, J = 11.7, 3.0 Hz, 2H, H5′), 2.56 (t, J = 2.4 Hz, 1H, CuCH).
13C NMR (75 MHz, CD3OD) δ 162.9, 152.3, 141.5, 95.1, 87.3,
85.4, 77.9, 76.4, 75.8, 70.7, 61.3, 29.8. HR-ESI-QToF MS (posi-
tive mode): m/z calcd for C12H15N2O6 [M + H]+ 283.0930 found
283.0930.
1-(2′,3′,5′-Tri-O-acetyl-β-D-xylofuranosyl)uracil
5. Uracil
(262 mg, 2.3 mmol, 1.2 eq.) and N,O-bis(trimethylsilyl)aceta-
mide (1.1 mL, 4.5 mmol, 2.2 eq.) were added to a solution of
1,2,3,5-tetra-O-acetyl-D-xylofuranose 4 (637 mg, 2.0 mmol, 1.0
eq.) in CH3CN (25 mL) and refluxed until the suspension
Glycooligonucleotide synthesis
The oligonucleotide synthesis on a solid support was per- became a clear solution. Then, the reaction mixture was
formed on a DNA synthesizer Applied Biosystems (381A or cooled to 0 °C and TMSOTf (511 mg, 2.3 mmol, 1.2 eq.) was
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
added dropwise, followed by refluxing for 4.5 h. CH3CN was General procedure for elongation of DNA sequences and
evaporated under reduced pressure. To the residue was added labelling with Cy3
saturated NH4Cl solution and the mixture was extracted with
CH2Cl2. The organic layer was dried with Na2SO4, evaporated
and nucleoside 5 was purified by silica gel column chromato-
The DNA sequences were synthesized on the solid-supported
scaffolds (LCAA-CPG, 500 Å) at the 1 mmol scale on a DNA
synthesizer (ABI 394) by standard phosphoramidite chemistry.
For the coupling step, benzylmercaptotetrazole was used as an
graphy (AcOEt : cyclohexane 8 : 2) to afford the title compound
5 (488 mg, 65%). 1H NMR (300 MHz, CDCl3) δ 9.82 (s, 1H,
activator (0.3 M in anhydrous CH3CN), and commercially avail-
NH), 7.46 (d, J = 8.1 Hz, 1H, H6), 5.96 (d, J = 2.4 Hz, 1H, H1′),
able nucleoside phosphoramidites (0.1
M in anhydrous
5.74 (d, J = 8.1 Hz, 1H, H5), 5.30 (dd, J = 3.9, 1.5 Hz, 1H, H3′),
5.13 (t, J = 2.1 Hz, 1H, H2′), 4.46 (q, J = 9.6, 5.4 Hz, 1H, H4′),
4.27 (d, J = 5.7 Hz, 2H, H5′). 13C NMR (100 MHz, CDCl3)
δ 169.4, 168.1, 168.0, 162.3, 149.3, 138.1, 101.8, 87.4, 78.6,
77.7, 73.3, 59.9, 20.0, 19.7, 19.6. HR-ESI-QToF MS (positive
mode): m/z calcd for C15H19N2O9 [M + H]+ 371.1091 found
371.1092.
1-(β-D-Xylofuranosyl)-3-(prop-2-yn-1-yl)uracil 3c. A solution
of propargyl bromide (76 mg, 0.64 mmol, 1.18 eq.) was added
to a solution of 2′,3′,5′-tri-O-acetyl xylouridine 5 (200 mg,
0.54 mmol, 1.00 eq.) and potassium carbonate (373 mg,
2.7 mmol, 5.00 eq.) in DMF (5 mL). The system was stirred at
60 °C overnight and then methanol was added (300 mL). The
reaction mixture was filtered and evaporated to dryness. The
crude product was purified by silica gel column chromato-
graphy (AcOEt : cyclohexane 8 : 2) to afford the title compound
3c (98 mg, 65%). 1H NMR (300 MHz, CD3OD) δ 7.86 (d, J =
8.4 Hz, 1H, H6), 5.69 (d, J = 3.3 Hz, 1H, H1′), 5.67 (d, J = 8.1 Hz,
1H, H5), 4.55 (d, J = 2.4 Hz, 2H, CH2CuCH), 4.22 (m, 1H, H4′),
4.04–3.99 (m, 2H, H2′, H3′), 3.84 (d, J = 5.4 Hz, 2H, H5′), 2.45 (t,
J = 2.4 Hz, 1H, CuCH). 13C NMR (75 MHz, CD3OD): δ 30.87,
61.13, 71.87, 76.42, 79.08, 82.46, 85.49, 94.16, 100.71, 141.72,
151.73, 164.03. HR-ESI-QToF MS (positive mode): m/z calcd for
C12H15N2O6 [M + H]+ 283.0930 found 283.0930.
CH3CN) were introduced with a 20 s coupling time and Cy3
amidite (0.06 M in anhydrous CH3CN) with a 180 s coupling
time. The capping step was performed with acetic anhydride
using commercial solution (Cap A, Ac2O/pyridine/THF,
10 : 10 : 80; Cap B, 10% N-methylimidazole in THF) for 15 s.
Each oxidation was performed with commercial solution of
iodide (0.1 M in THF/pyridine/H2O, 78 : 20 : 2) for 15 s. Detrity-
lation was performed with 3% TCA in CH2Cl2 for 35 s.
Immobilization on solid support 11 or 12 of compounds 3a–c
or 10 by CuAAC
An aqueous solution of modified carbohydrates 3a–c or 10
(0.1 M, 50 mL, 5 eq.), freshly prepared aqueous solutions of
CuSO4 (0.1 M, 4 mL, 0.4 eq.) and sodium ascorbate (0.5 M,
4 mL, 2 eq.), TEAAc buffer (50 mL), water (17 mL) and MeOH
(125 mL) were added to 1 mmol of solid support 11 or 12. The
resulting mixture was heated at 65 °C for 1 h. The solution was
removed, and CPG beads were washed with H2O (3 × 2 mL),
MeOH (3 × 2 mL) and CH3CN (3 × 2 mL), and dried.
General procedure for introduction of phosphoramidites
13–15 on furanoside and reduced mannoside scaffolds
Solid-supported furanoside 18a–c and mannitol 28 scaffolds
(1 mmol) were treated by phosphoramidite chemistry, on a
DNA synthesizer, with phosphoramidites 13, 14 or 15. Only
coupling and oxidation steps were performed. For the coupling
step, benzylmercaptotetrazole was used as an activator (0.3 M
in anhydrous CH3CN) and phosphoramidites 13–15 (0.3 M
with the platform 18a–c and 0.5 M with the platform 28 in
anhydrous CH3CN) were introduced three times with a 180 s
coupling time. Oxidation was performed with 0.1 M commer-
cial solution of iodide for 15 s.
6-Azido-6-deoxy-D-mannose 9. To
a solution of methyl
6-azido-6-deoxy-α-D-mannopyranoside 822 (379 mg, 1.7 mmol,
1 eq.) in H2O/dioxane (15 mL, 1 : 2) was added a solution of
12 N HCl (4 mL). The mixture was stirred under reflux for 4 h
and then a saturated solution of sodium carbonate was added
to increase the pH to 7. The crude product was concentrated
in vacuo, diluted with MeOH and filtered. The filtrate was con-
centrated and purified by silica gel column chromatography
(EtOAc) to afford the desired compound 9 α : β mixture
(252 mg, 72%) as a pale yellow oil. Analyses were in agreement
with the literature.41
General procedure for deprotection of solid-supported alkyne
oligonucleotides
6-Azido-6-deoxy-D-mannitol 10. To a solution of 6-azido-6-
deoxy-D-mannose 9 (60 mg, 0.29 mmol, 1 eq.) in EtOH (8 mL)
was added NaBH4 (66 mg, 1.75 mmol, 6 eq.). The mixture was
stirred for 2 h at room temperature and ion exchange resin
(DOWEX-50W X8, H+ form) was then added to decrease the pH
to 3. The resin was then removed by filtration. The filtrate was
concentrated and co-evaporated with MeOH to give the desired
compound 10 after silica gel chromatography (49 mg, 88%).
1H NMR (300 MHz, CD3OD) δ 3.83–3.61 (m, 5H), 3.55 (dd, J =
12.6, 2.7 Hz, 1H), 3.39 (dd, J = 6.6, 12.6 Hz, 1H). 13C NMR
The CPG beads bearing modified oligonucleotides were trans-
ferred to a 4 mL screw top vial and treated with 2 mL of con-
centrated aqueous ammonia for 15 h at room temperature and
warmed to 55 °C for 2 h. The supernatants were withdrawn
and evaporated.
General procedure for introduction of galactoside 16 or
fucoside 17 derivatives by CuAAC reaction
In solution. To a solution of 5′-fluorescent-3′-alkyne oligo-
(100 MHz, D2O) δ 73.0, 71.8, 71.5, 71.0 (4 × CHOH), 65.1 nucleotides 21a–c and 22a–c (100 nmol) were added galacto-
(CH2OH), 55.8 (CH2N3). HR-ESI-QToF MS (negative mode): m/z side 16 or fucoside 17 derivatives (3 eq. per alkyne function,
calcd for C6H12N3O5 [M − H]−: 206.0777 found 206.0777.
0.1 M in MeOH or dioxane for 16), 1 mg of Cu(0) nanopowder,
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Organic & Biomolecular Chemistry
Paper
0.1 M triethylammonium acetate buffer, pH 7.7 (25 mL), water night. Then, slides were washed for 30 min at 70 °C in 0.1%
and MeOH or dioxane to obtain a final volume of 250 mL SDS and rinsed with water.
(water/MeOH, 1 : 1 v/v). The resulting mixture was heated at
Blocking non-specific adsorption was performed using
65 °C for 1 h. EDTA (0.1 M, 100 mL) was added to the mix- Bovine Serum Albumin (BSA). This step was performed before
tures, and after centrifugation, the supernatants were with- DDI of the glycoclusters and repeated before incubation with
drawn to eliminate Cu(0) and desalted by size exclusion the lectins. The slides were incubated with a 4% BSA solution;
chromatography on NAP10. After evaporation, the 5′-fluo- in 1× PBS, pH 7.4 for 2 h. The slides were washed in 1× 0.05%
rescent 3′-acetyl-glycomimetic oligonucleotides were treated PBS-Tween20 (3 × 3 min), in 1× PBS (3 × 3 min) and rinsed
with concentrated aqueous ammonia for 2 h at room tempera- with water.
ture to remove acetyl groups and evaporated. Final compounds
were purified by C18 reversed phase HPLC.
Immobilization of the glycoclusters was accomplished
using DNA directed Immobilization (DDI).26 Two µL of a
On solid support. A solution of galactoside 16 or fucoside mixture containing two different glycoclusters (1 µM per glyco-
17 derivatives (5 eq. per alkyne function, 0.1 M in MeOH or cluster in 1× PBS pH 7.4) was poured at the bottom of each
dioxane for 16), freshly prepared aqueous solutions of CuSO4 microwell and incubated (3 h, 37 °C, in a water saturated
(0.1 M, 0.4 eq.) and sodium ascorbate (0.5 M, 2 eq.), and chamber). The slides were washed in 2× SSC , 0.1% SDS at
TEAAC buffer (50 mL) were added to the solid supported oligo- 51 °C for 1 min, 2× SSC for 5 min and rinsed with water.
nucleotides 29a–c (1 mmol). The resulting mixture was heated Correct immobilization of the Cy3 labeled glycoclusters was
at 70 °C for 1 h. The solution was removed, and CPG beads ensured by scanning the slides at 532 nm (excitation of Cy3).
were washed with H2O (3 × 2 mL), MeOH (3 × 2 mL) and The fluorescence signal of each glycocluster was determined
CH3CN (3 × 2 mL), and dried. Finally, the DNA sequence was as the average of the mean fluorescence signal of 32 spots. The
elongated and labelled with Cy3.
surface densities of the glycoclusters deviated by less than
14%.
Lectins LecA, LecB and BambL (a gift from Dr Anne
Imberty, CERMAV) were labeled with Alexa 647 according to
Microarray production and Kd microarray assays
The procedure has been described by Goudot et al.25 3′-Amino- the supplier protocol. The degrees of labeling per lectin (multi-
modified oligonucleotides were purchased from Eurogentec meric form) were 0.30, 0.34 and 0.24 moles of Alexa-647 per
and chemicals from Sigma-Aldrich. An Alexa Fluor® 647 mole of lectin respectively.
Microscale Protein Labelling Kit was purchased from Invitro-
gen. Fluorescence scanning was performed using an Axon labeled lectin ranking from 0.01 nM to 2.0 µM for LecA and
microarray scanner, and the GenePix 4100A software package. BambL (or 1 nM to 2.4 µM for LecB) were prepared in 1× PBS
Microarray production. Nexterion Glass slides from (pH 7.4) in order to draw isotherm curves of 20 discrete
Kd determination. Solutions of 2% BSA, 7.5 µM CaCl2 and
D
Schott AG were used. The slides were structured by adapting points. In each microwell, 2 µL of solution was added and
the protocol of Mazurczyk et al.42 leading to slides bearing incubated for 3 h at 37 °C in a water saturated chamber. Then,
40 microwells (square 3 × 3 mm, depth: 102
4.5 mm spacing between each well).
1 µm, with slides were washed for 5 min with 1× 0.02% PBS-Tween20 and
rinsed with water. An average of 32 spots per glycocluster of
The resulting glass slides were silanized according to the the mean fluorescence signal at 635 nm (FI) was calculated.
gas phase protocol43 leading to a tert-butyl ester functionalized For each glycocluster, Kd values were determined by the inter-
surface. Deprotection of the acid function was accomplished cept with the y-axis using the linear regression: [LecA]/FI =
in glacial formic acid for 7 h. The slides were washed in 1/FImax × [LecA] + Kd/FImax
.
dichloromethane and water (10 min, under ultrasound). Acti-
vation of the carboxylic function was performed using diiso-
In silico molecular docking
propylcarbodiimide and N-hydroxysuccinimide solution at The three dimensional structures of LecA and LecB were
0.1 M in tetrahydrofuran, overnight at room temperature. The retrieved from the Protein Data Bank (http://www.rcsb.org)
slides were washed in THF followed by dichloromethane under the PDB codes 2VXJ and 1UVZ respectively. As an
(10 min, ultrasounds).
example, the building procedure of the lectin (LecA)–ligand
Next, four amino-modified oligonucleotides (25 µM, PBS (G2b) complex is depicted as follows: the galactose endings of
10× pH 8.5) were covalently immobilized in an array fashion the two glycocluster branches linked on C2 and C3 of the ara-
using a Scienion sciFLEXARRAYER s3 system piezoelectric. binofuranose ring are brought close to the crystallographic
Amino-modified oligonucleotides were spotted at 25 µM for galactose ligands (Fig. S1†). The terminal sugars of the two
LecB and at 0.5 µM for LecA and BambL in 10× PBS (pH 8.5). glycocluster branches are removed and the chemical bond
The immobilized sequences were 5′-GTG AGC CCA GAG GCA with the galactose moieties of LecA is built (Fig. S2†). LecA
GGG-(CH2)7-NH2, 5′-GTG GAG GCA CCA AGC TTT-(CH2)7-NH2, and its two galactose groups are considered as aggregates and
5′-CCA AGC GAG GTG GCA TTT-(CH2)7-NH2 and 5′-CCA GCA the complex is optimized (Fig. S2†). A classical Monte Carlo
GTG GAG AGC TTT-(CH2)7-NH2. In each microwell, 64 spots of conformational searching procedure is then performed as
two sequences (2 × 32 spots) were spotted under a water satu- described using the BOSS software.44 The constraints have
rated atmosphere. Water was allowed to evaporate gently over- been removed and the whole structure was relaxed. In the
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.

View Article Online
Paper
Organic & Biomolecular Chemistry
minimization procedures, the spectroscopic empirical energy 13 B. Gerland, A. Goudot, C. Ligeour, G. Pourceau, A. Meyer,
function SPASIBA and the corresponding parameters are
used.45–47 A typical lowest energy structure for the complex
LecA–G2b is shown in Fig. S3.†
S. Vidal, T. Gehin, O. Vidal, E. Souteyrand, J. J. Vasseur,
Y. Chevolot and F. Morvan, Bioconjugate Chem., 2014, 25,
379–392.
In the same way, an empirical potential energy of inter- 14 B. Gerland, A. Goudot, G. Pourceau, A. Meyer, S. Vidal,
action ΔE for the lectin–ligand complexes using the simple
expression:
E. Souteyrand, J.-J. Vasseur, Y. Chevolot and F. Morvan,
J. Org. Chem., 2012, 77, 7620–7626.
15 G. Pourceau, A. Meyer, J. J. Vasseur and F. Morvan, J. Org.
Chem., 2009, 74, 6837–6842.
ΔEinteraction ¼ Ecomplex ꢀ Eprotein ꢀ Eligand
16 J. Lietard, A. Meyer, J. J. Vasseur and F. Morvan, J. Org.
Chem., 2008, 73, 191–200.
was evaluated using the same force field.
17 C. Ligeour, A. Meyer, J. J. Vasseur and F. Morvan,
Eur. J. Org. Chem., 2012, 1851–1856.
18 S. Cecioni, J. P. Praly, S. E. Matthews, M. Wimmerova,
A. Imberty and S. Vidal, Chem. – Eur. J., 2012, 18, 6250–
6263.
19 F. Morvan, A. Meyer, A. Jochum, C. Sabin, Y. Chevolot,
A. Imberty, J. P. Praly, J. J. Vasseur, E. Souteyrand and
S. Vidal, Bioconjugate Chem., 2007, 18, 1637–1643.
20 M. Nakane, S. Ichikawa and A. Matsuda, J. Org. Chem.,
2008, 73, 1842–1851.
Acknowledgements
This work was financially supported by ANR-12-BSV5-0020
“GLYCOMIME”. Plateforme NanoLyon is acknowledged for its
technical support. C. L and A. A. thank the University of
Montpellier for a research studentship. F.M. is a member of
Inserm.
21 B. R. Baker, in The Ciba Foundation Symposium on the
Chemistry and Biology of the Purine, ed. G. E. W. Wolsten-
holme and C. M. O’Connor, John Wiley & Sons, Chichester,
UK, 1957, pp. 120–133.
References
1 N. Gilboa-Garber, Methods Enzymol., 1982, 83, 378–385.
2 A. Imberty, M. Wimmerova, E. P. Mitchell and N. Gilboa- 22 J. J. Reina, O. S. Maldonado, G. Tabarani, F. Fieschi and
Garber, Microbes Infect., 2004, 6, 221–228. J. Rojo, Bioconjugate Chem., 2007, 18, 963–969.
3 D. Sicard, S. Cecioni, M. Iazykov, Y. Chevolot, 23 L. Chaveriat, I. Stasik, G. Demailly and D. Beaupere, Tetra-
S. E. Matthews, J. P. Praly, E. Souteyrand, A. Imberty, hedron: Asymmetry, 2006, 17, 1349–1354.
S. Vidal and M. Phaner-Goutorbe, Chem. Commun., 2011, 24 Since the catalytic form of copper is Cu(I), Cu(0) was used
47, 9483–9485.
as a nanopowder to ensure a large specific surface area
allowing its oxidation by oxygen contained in solvents.
4 S. Cecioni, R. Lalor, B. Blanchard, J. P. Praly, A. Imberty,
S. E. Matthews and S. Vidal, Chem. – Eur. J., 2009, 15, 25 A. Goudot, G. Pourceau, A. Meyer, T. Gehin, S. Vidal,
13232–13240.
J. J. Vasseur, F. Morvan, E. Souteyrand and Y. Chevolot,
Biosens. Bioelectron., 2013, 40, 153–160.
5 S. Cecioni, A. Imberty and S. Vidal, Chem. Rev., 2015, 115,
525–561.
6 K. Buffet, E. Gillon, M. Holler, J. F. Nierengarten,
A. Imberty and S. P. Vincent, Org. Biomol. Chem., 2015, 13,
6482–6492.
26 Y. Chevolot, C. Bouillon, S. Vidal, F. Morvan, A. Meyer,
J. P. Cloarec, A. Jochum, J. P. Praly, J. J. Vasseur and
E. Souteyrand, Angew. Chem., Int. Ed., 2007, 46, 2398–2402.
27 R. Wacker and C. M. Niemeyer, ChemBioChem, 2004, 5,
453–459.
7 A. Audfray, J. Claudinon, S. Abounit, N. Ruvoen-Clouet,
G. Larson, D. F. Smith, M. Wimmerova, J. Le Pendu, 28 J. Zhang, G. Pourceau, A. Meyer, S. Vidal, J.-P. Praly,
W. Romer, A. Varrot and A. Imberty, J. Biol. Chem., 2012,
287, 4335–4347.
E. Souteyrand, J.-J. Vasseur, F. Morvan and Y. Chevolot,
Biosens. Bioelectron., 2009, 24, 2515–2521.
8 C. Ligeour, A. Audfray, E. Gillon, A. Meyer, N. Galanos, 29 J. Zhang, G. Pourceau, A. Meyer, S. Vidal, J. P. Praly,
S. Vidal, J. J. Vasseur, A. Imberty and F. Morvan, RSC Adv.,
2013, 3, 19515–19524.
E. Souteyrand, J. J. Vasseur, F. Morvan and Y. Chevolot,
Chem. Commun., 2009, 6795–6797.
9 B. Richichi, A. Imberty, E. Gillon, R. Bosco, I. Sutkeviciute, 30 S. Park and I. Shin, Org. Lett., 2007, 9, 1675–1678.
F. Fieschi and C. Nativi, Org. Biomol. Chem., 2013, 11, 31 F. Pertici, N. J. de Mol, J. Kemmink and R. J. Pieters, Chem.
4086–4094.
– Eur. J., 2013, 19, 16923–16927.
10 R. J. Pieters, Org. Biomol. Chem., 2009, 7, 2013–2025.
11 G. Pourceau, A. Meyer, Y. Chevolot, E. Souteyrand,
J. J. Vasseur and F. Morvan, Bioconjugate Chem., 2010, 21,
1520–1529.
32 F. Casoni, L. Dupin, G. Vergoten, A. Meyer, C. Ligeour,
T. Gehin, O. Vidal, E. Souteyrand, J. J. Vasseur, Y. Chevolot
and F. Morvan, Org. Biomol. Chem., 2014, 12, 9166–9179.
33 N. Garber, U. Guempel, A. Belz, N. Gilboa-Garber and
R. J. Doyle, Biochim. Biophys. Acta, 1992, 1116, 331–333.
12 B. Gerland, A. Goudot, G. Pourceau, A. Meyer, V. Dugas,
S. Cecioni, S. Vidal, E. Souteyrand, J. J. Vasseur, Y. Chevolot 34 R. U. Kadam, M. Bergmann, M. Hurley, D. Garg,
and F. Morvan, Bioconjugate Chem., 2012, 23, 1534–1547.
M. Cacciarini, M. A. Swiderska, C. Nativi, M. Sattler,
Org. Biomol. Chem.
This journal is © The Royal Society of Chemistry 2015

View Article Online
Organic & Biomolecular Chemistry
Paper
A. R. Smyth, P. Williams, M. Camara, A. Stocker, T. Darbre 40 G. D. Hawkins, C. J. Cramer and D. G. Truhlar, J. Phys.
and J. L. Reymond, Angew. Chem., Int. Ed., 2011, 50, 10631–
10635.
35 J. Rodrigue, G. Ganne, B. Blanchard, C. Saucier,
Chem., 1996, 100, 19824–19839.
41 D. C. M. Kong and M. von Itzstein, Carbohydr. Res., 1997,
305, 323–329.
D. Giguere, T. C. Shiao, A. Varrot, A. Imberty and R. Roy, 42 R. Mazurczyk, G. E. Khoury, V. Dugas, B. Hannes,
Org. Biomol. Chem., 2013, 11, 6906–6918.
E. Laurenceau, M. Cabrera, S. Krawczyk, E. Souteyrand,
J. P. Cloarec and Y. Chevolot, Sens. Actuators, B, 2008, 128,
552–559.
36 R. U. Kadam, D. Garg, J. Schwartz, R. Visini, M. Sattler,
A. Stocker, T. Darbre and J. L. Reymond, ACS Chem. Biol.,
2013, 8, 1925–1930.
43 M. Phaner-Goutorbe, V. Dugas, Y. Chevolot and
E. Souteyrand, Mater. Sci. Eng., C, 2011, 31, 384–390.
37 M. L. Gening, D. V. Titov, S. Cecioni, A. Audfray,
A. G. Gerbst, Y. E. Tsvetkov, V. B. Krylov, A. Imberty, 44 W. L. Jorgensen and J. Tirado-Rives, J. Comput. Chem.,
N. E. Nifantiev and S. Vidal, Chem. – Eur. J., 2013, 19, 9272–
9285.
38 C. Ligeour, L. Dupin, A. Marra, G. Vergoten,
2005, 26, 1689–1700.
45 P. Derreumaux and G. Vergoten, J. Chem. Phys., 1995, 102,
8586–8605.
A. Meyer, A. Dondoni, E. Souteyrand, J. J. Vasseur, 46 G. Vergoten, I. Mazur, P. Lagant, J. C. Michalski and
Y. Chevolot and F. Morvan, Eur. J. Org. Chem., 2014,
7621–7630.
39 J. J. P. Stewart, J. Mol. Model, 2007, 13, 1173–1213.
J. P. Zanetta, Biochimie, 2003, 85, 65–73.
47 P. Lagant, D. Nolde, R. Stote, G. Vergoten and M. Karplus,
J. Phys. Chem. A, 2004, 108, 4019–4029.
This journal is © The Royal Society of Chemistry 2015
Org. Biomol. Chem.