Rational design and synthesis of optimized glycoclusters for multivalent lectin-carbohydrate interactions: Influence of the linker arm

Article abstract of DOI:10.1002/chem.201200010

The design of multivalent glycoclusters requires the conjugation of biologically relevant carbohydrate epitopes functionalized with linker arms to multivalent core scaffolds. The multigram-scale syntheses of three structurally modified triethyleneglycol analogues that incorporate amide moiety(ies) and/or a phenyl ring offer convenient access to a series of carbohydrate probes with different water solubilities and rigidities. Evaluation of flexibility and determination of preferred conformations were performed by conformational analysis. Conjugation of the azido-functionalized carbohydrates with tetra-propargylated core scaffolds afforded a library of 18 tetravalent glycoclusters, in high yields, by Cu^I-catalyzed azide-alkyne cycloaddition (CuAAC). The compounds were evaluated for their ability to bind to PA-IL (the LecA lectin from the opportunistic pathogen Pseudomonas aeruginosa). Biochemical evaluation through inhibition of hemagglutination assays (HIA), enzyme-linked lectin assays (ELLA), surface plasmon resonance (SPR), and isothermal titration microcalorimetry (ITC) revealed improved and unprecedented affinities for one of the monovalent probes (K_d=5.8 μM) and also for a number of the tetravalent compounds that provide several new nanomolar ligands for this tetrameric lectin. Copyright

Full text of DOI:10.1002/chem.201200010

DOI: 10.1002/chem.201200010
Rational Design and Synthesis of Optimized Glycoclusters for Multivalent
Lectin–Carbohydrate Interactions: Influence of the Linker Arm
Samy Cecioni,[a, b] Jean-Pierre Praly,[a] Susan E. Matthews,[c] Michaela Wimmerovꢀ,[d]
Anne Imberty,*[b] and Sꢁbastien Vidal*[a]
Abstract: The design of multivalent
glycoclusters requires the conjugation
of biologically relevant carbohydrate
epitopes functionalized with linker
arms to multivalent core scaffolds. The
multigram-scale syntheses of three
structurally modified triethyleneglycol
analogues that incorporate amide
by conformational analysis. Conjuga-
tion of the azido-functionalized carbo-
hydrates with tetra-propargylated core
scaffolds afforded a library of 18 tetra-
valent glycoclusters, in high yields, by
CuI-catalyzed azide–alkyne cycloaddi-
tion (CuAAC). The compounds were
evaluated for their ability to bind to
PA-IL (the LecA lectin from the op-
portunistic pathogen Pseudomonas aer-
uginosa). Biochemical evaluation
through inhibition of hemagglutination
assays (HIA), enzyme-linked lectin
assays (ELLA), surface plasmon reso-
nance (SPR), and isothermal titration
microcalorimetry (ITC) revealed im-
proved and unprecedented affinities
for one of the monovalent probes
(Kd =5.8 mm) and also for a number of
the tetravalent compounds that provide
several new nanomolar ligands for this
tetrameric lectin.
moietyACHTUNGTRENNUNG(ies) and/or a phenyl ring offer
convenient access to a series of carbo-
hydrate probes with different water
solubilities and rigidities. Evaluation of
flexibility and determination of pre-
ferred conformations were performed
Keywords: carbohydrates
·
click
chemistry · glycoclusters · lectin ·
multivalency
Introduction
Carbohydrate–lectin interactions[1–6] play a prominent role
in various biological processes such as fertilization, cell–cell
communication, viral and bacterial infection, cancer meta-
stasis, and inflammation. The biomolecular interactions in-
volved at the monovalent level between a carbohydrate and
its specific lectin are usually in the millimolar range.[3]
Nature uses the so-called “cluster glycoside effect”[7,8]
through multivalent interactions with multiple copies of car-
bohydrate epitopes, lectins, or both to reach higher affinities
and more specific biomolecular interactions. The concept of
multivalency[9–12] has therefore emerged as a powerful ap-
proach in both glycochemistry and glycobiology to design
potent lectin ligands to prevent viral or bacterial infection
for instance through anti-adhesive strategies.[13,14] Neverthe-
less, the precise mode of action of such multivalent glyco-
conjugates is still not completely understood both at the bio-
molecular and biological level. The thermodynamics of such
multivalent interactions is highly challenging and important
contributions have been made over the past decades using
isothermal titration microcalorimetry[15,16] (ITC) among
other bioanalytical techniques.[17]
[a] Dr. S. Cecioni, Dr. J.-P. Praly, Dr. S. Vidal
Institut de Chimie et Biochimie
Molꢀculaires et Supramolꢀculaires
Laboratoire de Chimie Organique 2
Glycochimie, UMR 5246, CNRS,
Universitꢀ Claude Bernard Lyon 1
43 Boulevard du 11 Novembre 1918
69622 Villeurbanne (France)
Fax : (+33)472-432-752
E-mail: sebastien.vidal@univ-lyon1.fr
[b] Dr. S. Cecioni, Dr. A. Imberty
Centre de Recherche sur les
Macromolꢀcules Vꢀgꢀtales (CNRS)
affiliated with Universitꢀ Joseph Fourier
Grenoble and ICMG
BP 53, 38041 Grenoble (France)
[c] Dr. S. E. Matthews
School of Pharmacy
University of East Anglia
Norwich, NR4 7TJ (UK)
[d] Dr. M. Wimmerovꢁ
Central European Institute of Technology
Masaryk University
62500 Brno (Czech Republic)
Bacterial lectins are involved in the early stages of infec-
tion, and competing with their binding to human glycoconju-
gates is an appealing alternative or complement to antibiotic
treatment.[13,18] PA-IL (LecA) from Pseudomonas aeruginosa
is a tetrameric calcium-dependent lectin with specificity for
a-galactose-terminated oligosaccharides[19–21] and for b-gal-
actosides with hydrophobic aglycon.[22] The inhibition of PA-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201200010. It includes experi-
mental procedures for all new compounds, SPR sensorgrams and ITC
data for monovalent and multivalent ligands, molecular modeling in-
formation, characterization with copies of H and 13C NMR spectra
1
for all new compounds, and crystallographic data for compound 1e.
6250
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

FULL PAPER
IL by simple monosaccharides was shown to be efficient for
decreasing the population of P. aeruginosa in infected lungs
for both mice models and human patients with cystic fibro-
sis.[23,24] To obtain highly efficient competitors, we and others
have synthesized multivalent galactosylated glycococlusters
based on dendrimers,[22,25] calix[4]arenes,[26] calix[6]arenes,[27]
fullerenes,[28,29] resorcin[4]arenes,[30] porphyrins,[27] b-pep-
toids,[27] and poly(phenylacetylene).[31] Among these mole-
cules, the 1,3-alternate conformer of calix[4]arene demon-
strated a dramatic increase in affinity, and a chelate aggrega-
tive binding mode was suggested on the basis of molecular
modeling.[26] Recently, this hypothesis was further supported
by the observation of well-defined nanometric fibers of
lectin/glycocluster complexes through atomic force micros-
copy (AFM) investigations.[32] Tetravalent-based scaffolds
therefore appear to be particularly appropriate for high-af-
finity binding to PA-IL.
a too rigid linker may not allow the second epitope to reach
its binding pocket on the multivalent lectin in a chelate
binding mode.
The general formula of the glycoclusters presented here
(Scheme 1A) is based on a multivalent central core scaffold
with an EG3 spacer arm as previously described[26]
As a rational continuation of the project, second-genera-
tion tetravalent glycoclusters are reported herein, with the
aim of optimizing the thermodynamic contributions, that is,
enthalpy and entropy of binding, in the design of new opti-
mized linkers. The design of high-affinity ligands for lectins
requires the best possible match between valency and topo-
logical display of the carbohydrate epitopes that surround
the multivalent core scaffold. Both of these parameters have
been extensively examined in the literature,[10–12] and the in-
fluence of the linker arm that connects the carbohydrate to
the central core unit has also been previously ad-
dressed.[33–43] The design of spacer arms that present differ-
ent rigidities with similar lengths would provide the basis for
a careful analysis of the parameters that govern multivalent
carbohydrate–lectin interactions. Gains in enthalpy would
arise from modifications of the structure of the carbohydrate
epitope through additional stabilizing contacts, but increased
affinities for lectins can also be obtained with more rigid
linker arms, which decrease the entropic cost upon binding.
Oligo(ethyleneglycol)s are typical spacer arms used for
the design of multivalent glycoconjugates; they have the ad-
vantage of being water soluble, of various lengths, and with
an alcohol moiety that can be either easily derivatized to al-
ternative functional groups or even directly involved in a gly-
cosidic bond for connection to the carbohydrate epitope.
Triethyleneglycol (EG3) is a very common member of this
family, and numerous examples have been previously de-
scribed in the literature.[26,44–50] This linker arm is easy to in-
troduce through glycosylation but also provides improved
water solubility for the resulting multivalent ligands synthe-
sized. The sp3 hybridization of carbon and oxygen atoms
along its chain provides a high freedom of movement for
the terminal carbohydrate epitopes.
Scheme 1. Structure of the designed linker arms and their respective
characteristics. EG=ethyleneglycol, NAz=N-azidoacetyl, Gly=glycine,
Ph=phenyl.
(Scheme 1B). The introduction of one amide or two amide
moieties into the EG3 chain of the linker arm (Scheme 1B
and C) should provide more rigid chains.[51] The C=O and
NH groups of the amide moieties could also interact with
the side chain of the lectin polypeptide, thus leading to po-
tentially more favorable enthalpic contributions. The intro-
duction of a phenyl aromatic group (Scheme 1E) modifies
the general structure of an EG3-type chain and also its
length. This hydrophobic moiety was selected based on ex-
perimental results that demonstrate that a strong gain in af-
finity for PA-IL can be obtained with a b-galactoside that
contains an aromatic aglycon.[22,52]
Linker arms with sp2-hybridized atoms in a chain analo-
gous to EG3 in terms of length can be used to obtain glyco-
clusters that present more rigidity than the parent EG3-
based structures. Such glycoconjugates would provide a rela-
tive entropic gain and therefore better binding properties to-
wards the lectin. Nevertheless, the introduction of rigidifying
motifs in the linker arm should be accurately balanced since
Results and Discussion
Synthesis: The different carbohydrate probes were synthe-
sized on a multigram scale in a limited number of steps. We
recently reported a high-yield glycosylation protocol that
started from 1,2,3,4,6-penta-O-acetyl-b-d-galactopyranose
under activation with tin(IV) and silver(I) salts.[53,54] This re-
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6251

S. Vidal, A. Imberty et al.
visited protocol affords the desired glycosides in high yields,
complete stereoselectivity, and short reaction times, but also
starts from a nonactivated galactose derivative in compari-
son to glycosyl bromides or trichloroacetimidates. This effi-
cient glycosylation methodology has been applied to the
preparation of b-d-galactopyranoside 1b from the commer-
cially available chlorinated diethyleneglycol and subsequent
azidation (Scheme 2). The azidoacetyl group was then intro-
duced by reduction, acylation, and azidation to afford 1c
after a single chromatographic purification step. To observe
an unbiased effect of the multivalent presentation of carbo-
hydrate epitopes, it was of prime importance to prepare suit-
able monovalent references that included the epitope and
the whole linker arm. Therefore, the CuI-catalyzed azide–
alkyne cycloaddition[55–57] (CuAAC), under microwave irra-
diation, of 1c with propargyl acetate afforded the desired
1,4-substituted 1,2,3-triazole 2 in a short reaction time. The
subsequent solvolysis of the ester groups afforded the hy-
droxylated derivative 3.
Similarly, the glycosylation of 2-chloroethanol and subse-
quent azidation afforded 4, which was then reduced to the
amine. The crude intermediate was treated with N-chloroa-
cetylglycine under N-hydroxybenzotriazole/1-ethyl-3-(3-di-
methylaminopropyl) (HOBt/EDCI) activation to obtain 5,
which was then converted to the corresponding azide 1d.
CuAAC of 1d with propargyl acetate afforded the acetylat-
ed cycloadduct 6, and deprotection of the acetate groups
yielded the desired hydroxylated galactoside 7.
Scheme 3. Syntheses of the aromatic monovalent probe. Reagents and
conditions: a) Ac2O, C5H5N, 4-dimethylaminopyridine (DMAP), RT,
16 h; b) H2, Pd/C 10%, CH2Cl2, RT, 16 h; c) BrCH2COBr, Et3N, CH2Cl2,
RT, 2 h; d) NaN3, nBu4NI, DMF, 508C, 16 h, 74% over four steps;
e) propargyl acetate, CuI, iPr2NEt, DMF, 1108C, microwaves, 15 min,
98%; and f) MeOH, H2O, Et3N, RT, 16 h, 95 %.
ꢀ
bond along with an overall planar geometry of the Ph NAz
linker arm (Figure 1). The three-dimensional crystal packing
ꢀ
is achieved through weak C H···O intermolecular interac-
tions. All bond lengths and angles are within the expected
values. The aromatic ring is tilted between the benzene ring
and the mean plane of the sugar cycle (through C1 and C4)
with an angle of 34.88. A distance of 10.8 ꢃ was measured
between the anomeric carbon atom and the terminal azido
nitrogen atom.
The commercially available 4-nitrophenyl b-d-galactopyr-
anoside 8 was converted into the azido-functionalized deriv-
ative 1e in four steps with only one chromatographic purifi-
cation and in 77% overall yield (Scheme 3). Similarly, the
CuAAC with propargyl acetate afforded 9, and the hydroxy-
lated derivative 10 was then obtained by solvolysis of the
ester-protecting groups.
Single crystals suitable for X-ray crystallography were ob-
tained by crystallization of 1e, and analysis of the data col-
lected displayed the expected Z configuration of the amide
With azido-functionalized precursors in hand, conjuga-
tions of the carbohydrate probes 1c, 1d, 1e, or 1a[29] were
then performed by CuAAC with several propargylated core
scaffolds to afford acetylated glycoclusters (Scheme 4). The
tetra-propargylated methyl glucoside 11G[58] was obtained
by Williamson etherification of methyl a-d-glucopyranoside.
Scheme 2. Syntheses of the mono-amide and bis-amide derivatives. Reagents and conditions: a) HOCH2CH2OCH2CH2Cl, SnCl4, CF3CO2Ag, CH2Cl2,
RT, 2 h; b) NaN3, nBu4NI, DMF, 858C, 16 h, 57% for 1b and 78% for 4 over two steps; c) H2, Pd/C 10%, CH2Cl2, RT, 16 h; d) BrCH2COBr, Et3N,
CH2Cl2, RT, 12 h; e) NaN3, nBu4NI, DMF, 808C, 16 h, 57% over three steps; f) propargyl acetate, CuI, iPr2NEt, DMF, 1108C, microwaves, 15 min, 72%
for 2 and 99% for 6; g) MeOH, H2O, Et3N, RT, 16 h, 94% for 3 and 78% for 7; h) HOCH2CH2Cl, SnCl4, CF3CO2Ag, CH2Cl2, RT, 2 h; i) H2, Pd/C 10%,
CH2Cl2, RT, 20 h; j) N-chloroacetylglycine, EDCI, HOBt, CH2Cl2/DMF, RT, 16 h, 54% over two steps; and k) NaN3, nBu4NI, DMF, 808C, 16 h, 78%.
6252
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
Table 1. Biochemical evaluation of the monovalent probes for their bind-
ing to PA-IL.
HIA
ELLA
SPR
Ligand
MIC [mm]
IC50 [mm]
IC50 [mm]
b-d-GalOMe[a]
GalEG3
GalEG2NAz
GalEGGlyNAz
GalPhNAz
5000
10000
10000
2500
127.5
220
350
300
46
n.d.[b]
63.5
58.3
60
[a]
22
3
7
10
250
2.6
[a] Synthesized from the literature[27] and see Scheme 5 for structural de-
scription. [b] n.d.=not determined.
resonance (SPR). In the HIA assays, ligands 22,[26] 3, and 7
that bore aliphatic linker arms displayed weak minimum in-
hibitory concentrations (MIC), whereas the p-amidophenyl-
based ligand 10 demonstrated a significant improvement
with submillimolar MIC.
Figure 1. ORTEP view of the asymmetric unit of the unit cell of 1e show-
ing 30% probability displacement ellipsoids.
Evaluation of IC50 values (concentration that gives 50%
inhibition) by other biochemical techniques (ELLA and
competitive SPR; Table 1) highlighted similar profiles with
comparable inhibition potencies for aliphatic compounds 22,
3, and 7, and strong improvement in the inhibition potency
of the aromatic ligand 10 (Figure 2). These results were of
great interest in evaluating the influence of different linker
arms on the binding properties of multivalent glycoclusters.
The linkers of molecules 22, 3, and 7 did not appear to
strongly influence the binding of the monovalent probe,
thereby indicating that the linker is not likely to interact
with the surface. In contrast, the aromatic linker of 10 ap-
peared to provide an optimized structure from the glycomi-
metic approach. We therefore had in hand linkers with very
different properties, even at the monovalent compound
level.
Further studies were undertaken by using isothermal titra-
tion microcalorimetry (ITC) to measure precisely the affini-
ties between monovalent glycoclusters and PA-IL and to de-
termine the enthalpy and entropy contributions of the bind-
ing energy. Compounds 22, 3, and 7 have submillimolar-
range dissociation constant (KD), and the titrations were
conducted with an excess amount of ligand and fixed stoi-
chiometry as recommended for low-affinity systems[60] (see
the figures in the Supporting Information). The higher-affin-
ity aromatic ligand 10 provided a well-defined sigmoid
during titration (Figure 3) with a stoichiometry value of
0.82, close to the value of n expected from structural data.
The introduction of one amide residue in 3 compared to
22 did not modify significantly the binding to PA-IL with KD
for the monovalent ligand 3 of 107 mm (Table 2). The intro-
duction of a second amide motif was slightly more detrimen-
tal and resulted in a KD value of 181 mm for compound 7.
Overall, the affinities of 22, 3, and 7 for PA-IL are in the
same range. Titration of PA-IL with the aromatic ligand 10
yielded the determination of a KD value of 5.8 mm, which is
an unprecedented affinity for a monovalent glycomimetic
toward this bacterial lectin.
The conjugations to prepare 12aG, 12cG, 12dG, and 12eG
proceeded with modest yields (about 50%) when consider-
ing the overall yield of the reaction but represented an aver-
age yield of 85% per reactive site. Couplings of the glyco-
sides with the tetra-propargylated Zn–porphyrin 11P[27,59]
also proceeded smoothly under microwave activation with
no exchange of the chelated metallic ion to yield acetylated
glycoclusters 14aP,[27] 14dP, and 14eP in high yields. The
synthesis of topoisomerical tetra-propargylated calix[4]ar-
enes was previously described,[26] and CuAAC coupling with
azido-functionalized carbohydrates provided various con-
formers of tetravalent calix[4]arene-based glycoclusters.
Thus, cone tetra-propargylated calix[4]arene 11C yielded
acetylated glycoclusters 16aC,[26] 16dC, and 16eC, whereas
the 1,3-alternate conformer 11A provided glycoclusters
18aA,[26] 18bA, 18cA, 18dA, and 18eA with high yields,
and the partial cone conformer 11PC afforded acetylated
compounds 20aPC, 20dPC, and 20ePC. These calixarene-
based atropoisomers were stable during CuAAC conjuga-
tion under microwave irradiation since no modification of
the conformation were observed by NMR spectroscopic
analyses and characteristic 1H signals were conserved for
each conformation. The deprotection method for the solvol-
ysis of the acetate groups by using MeOH/H2O/Et3N was
carefully chosen because it creates byproducts that are
easily eliminated under vacuum and therefore does not re-
quire purification of the final hydroxylated glycoclusters. All
acetylated and hydroxylated glycoclusters were fully charac-
terized by NMR spectroscopy (1D and 2D) and high-resolu-
tion mass spectrometry (HRMS).
Interaction of the monovalent references with PA-IL: To
determine the influence and contribution of the spacer on
the affinity for the studied lectin PA-IL, all monovalent ref-
erences were assessed in the interaction study. In a first
series of experiments (Table 1), the ability of monovalent
probes to inhibit binding of PA-IL to glycosylated surfaces
was evaluated by hemagglutination inhibition assays (HIA),
enzyme-linked lectin assays (ELLA), and surface plasmon
When looking at the thermodynamic contributions, it can
be noted that enthalpic contributions increase along with
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6253

S. Vidal, A. Imberty et al.
Scheme 4. Syntheses of 18 glycoclusters with rigidified linker arms.
6254
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
Figure 3. Raw ITC data (top) obtained by injections of 10 (700 mm) in
a solution containing PA-IL (82 mm). Corresponding integrated titration
curve (bottom).
Table 2. Isothermal titration microcalorimetry (ITC) for the monovalent
probes and PA-IL.
[b]
n[a]
ꢀDH[b]
ꢀTDS
KD
&
Figure 2. ELLA (top) and SPR (bottom) inhibition plots: ) b-d-
Ligand
~
!
*
^
GalOMe, ) 22, ) 3, ) 7, ) 10. Errors bars represent standard devia-
G
]
U
[mm]
tion over triplicates (ELLA only).
b-d-GalOMe[c,d]
GalEG3
GalEG2NAz
GalEGGlyNAz
GalPhNAz
1[e]
39
15
14
26
43
23
70
[f]
22
3
7
1[e]
36ꢁ1
150ꢁ1
107ꢁ7
181ꢁ12
5.8ꢁ0.9
1[e]
48.8ꢁ0.7
64ꢁ11
53ꢁ2
1[e]
10
0.81ꢁ0.01
the hypothesized rigidity from compound 22 to ligands 3
and then 7. This improvement in enthalpy could be ex-
plained by additional contacts between amido moieties of
the linker arm and the lectin polypeptide backbone. Indeed,
there is a 12 kJmolꢀ1 increase in enthalpy (DH) between li-
gands 22 and 3 and a further 15 kJmolꢀ1 rise between com-
pounds 3 and 7, which could account for one and then two
[a] Stoichiometry. [b] Standard deviation is given over 2 or 3 experi-
ments. [c] Only one experiment performed. [d] Data from the litera-
ture.[27] [e] Fitting with constant stoichiometry determined with p-nitro-
phenyl-b-d-galactopyranoside. [f] Data from the literature.[26]
ꢀ
additional N H···O hydrogen bonds. Meanwhile, entropic
dicates that the aromatic ring interacts directly with the PA-
IL protein surface in the proximity of the binding site.[22]
cost (ꢀTDS) increases by 12 kJmolꢀ1 between 22 and 3 and
by 17 kJmolꢀ1 between 3 and 7. Even though supplementary
hydrogen bonds or an increase in water solvation could ac-
count for the increase in entropic cost, this behavior remains
quite intriguing, as it is frequently assumed that rigidifica-
tion of the structure would diminish the loss of entropy in
the bound state.
Molecular modeling of linker flexibility: To further investi-
gate the conformational freedom and geometry of the rigidi-
fied monovalent ligands, a molecular modeling procedure
was developed based on previously described methodolo-
gy.[61] The ligands were built in silico; the rotatable bonds
were defined as well as a set of geometrical descriptors for
analyzing the key structural features of the monovalent li-
gands (Scheme 5).
Centroid S was defined by the galactose ring atoms and
centroid T by the triazole ring atoms. Plane V was com-
posed of centroids S and T and the endocyclic oxygen atom.
Plane T was built on the three atoms of the triazole ring and
The higher affinity of compound 10 relative to 22 is due
to a much stronger enthalpy of binding (17 kJmolꢀ1 more
than 22) accompanied by a limited additional entropic cost
(9 kJmolꢀ1). The incorporation of an anomeric aromatic
moiety is known to create additional hydrophobic contacts
that result in higher desolvation of both ligand and receptor
upon binding. This is in agreement with recent data that in-
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6255

S. Vidal, A. Imberty et al.
Figure 4. Modeled probability as a function of distance between S and T
(centroids).
Figure 5. Modeled probability as a function of height between S and
plane T.
Scheme 5. Top: Definition of rotatable bonds and torsional increments
during systematic conformational search. Bottom: Geometrical parame-
ters measured for all conformations obtained from each ligands (e.g.,
ligand 3 presented here).
plane N by using an axis perpendicular to the centroid T
and the C5 atom of the triazole ring.
For each ligand, a systematic search was performed in the
conformational space defined by the rotatable bonds. All
conformations with a relative energy cutoff of 20 kcalmolꢀ1
were analyzed. The geometrical properties were calculated
by taking into account the probability of each conformation
as a function of its relative energy according to Boltzmann
equations [Eqs. (1) and (2)] in which pi is the probability of
existence of one given conformation, Ei is the energy of the
conformation, and k is the Boltzmann constant. This calcula-
tion allowed us to plot the probability of existence of con-
formations as a function of the geometrical parameters (Fig-
ures 4, 5, and 6).
Figure 6. Modeled probability as a function of height between S and
plane N.
6256
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
ꢀEi
ðkT
Þ
ligands are very efficient for the inhibition of agglutination
of erythrocytes.
e
ð1Þ
ð2Þ
pi ¼
Q ¼
Q
X
ꢀEi
ðkT
Þ
Analysis of ELLA results: All glycoclusters were soluble at
the concentration required for the ELLA assays, except for
13eG, that is, GalPhNAz conjugated to the tetra-propargy-
lated glucose core (Figure 7). Modest b values (9<b<57)
were obtained for GalPhNAz multivalent ligands (15eP,
17eC, 21ePC, and 19eA), but these compounds are never-
theless efficient inhibitors since the monovalent GalPhNAz
reference ligand 10 already displayed a low IC50 value of
46 mm. As previously observed in the HIA evaluation, the
porphyrin-based glycoclusters 15aP and 15dP displayed
b values above 400 and displayed the best IC50 values for
competitive binding to PA-IL as measured by ELLA.
Comparison of the b values obtained for the glucose-
based glycoclusters 13aG, 13cGm and 13dG illustrates the
influence of the linker arm, as observed from the HIA data,
and the mono-amide linker arm displayed the best binding
towards PA-IL among this subfamily of three glycoclusters.
Although the IC50 values measured in the ELLA assay for
these glucose-based glycoclusters are very similar (2 to
9 mm), the analysis of the b values demonstrates the influ-
ence of the linker arm on the improvement of binding prop-
erties relative to the monovalent ligands.
i e
Using the energy-weighted values (Boltzmann probabili-
ty), the distance between centroids S and T—that is, the
elongation of the molecule—varies between 5 and 14 ꢃ for
monovalent ligands 22, 3, and 7 (Figure 4). Small differences
are observed for the preferred shape with compound 22
being more elongated than the other two. In contrast, exten-
sion of compound 10 closely centered around a value of
10 ꢃ.
The heights between centroid S and plane T or plane N
also showed comparable behavior for ligands 22, 3, and 7
but with a slight preference for negative values (Figures 5
and 6). Positive and negative values indicate that the sugar
could move to both sides of the planes without high energy
penalties. For compound 10, a distribution centered on two
symmetric values of ꢀ10 and 9 ꢃ was observed.
These results indicate a high flexibility for compounds 22,
3, and 7 with little difference between them, which stands in
contradiction to the initial hypothesis for the chemical
design of the second-generation ligands. On the other hand,
ligand 10 clearly shows less flexibility and adopts a more re-
strained conformational behavior by populating a limited
number of conformational families.
Analysis of SPR results: The GalPhNAz glycoclusters were
not sufficiently soluble for SPR analyses. The IC50 values ob-
tained for the glycoclusters are in the low to submicromolar
range (Figure 7). The b values were rather modest and usu-
ally below 60. The best multivalent ligands identified by
SPR analysis were the 1,3-alternate calixarene-based glyco-
clusters 19aA and 19bA with IC50 values of 0.5 and 0.6 mm,
respectively. The shorter EG2-based linker was still capable
of inhibiting the adhesion of PA-IL onto the galactose-
coated gold surfaces used in the SPR experimental setup.
Although the rational design of the rigidified second-gen-
eration glycoclusters appeared promising, most of these mul-
tivalent ligands displayed potent binding properties for PA-
IL with IC50 values in the low micromolar to nanomolar
range as determined by ELLA and SPR. The improvement
in binding due to multivalency was remarkable for some of
the best molecules with b values of close to 100, but these
remained comparable to the first-generation glycoclusters
with EG3-based linker arms.
Among the second-generation multivalent ligands, several
structures displayed interesting inhibition features. The por-
phyrin-based glycocluster 15dP demonstrated strong and
promising potency both in HIA and ELLA. Calixarene-
based glycoclusters 17dC, 21dPC, and 19dA that bore the
same bis-amide EGGlyNAz linker arm yielded interesting
increases in the inhibition potency by ELLA. In HIA, these
three multivalent ligands demonstrated smaller improve-
ments.
Evaluation of the second-generation glycoclusters (HIA,
ELLA, and SPR): The second-generation glycoclusters pro-
posed here have been synthesized based on previously de-
scribed multivalent core scaffolds but with optimized linker
arms. Their evaluation as ligands of PA-IL was achieved
through a set of complementary bioanalytical techniques
(HIA, ELLA, SPR, and ITC; Tables 3 and 4). An improve-
ment value (b) of MIC or IC50 values was calculated versus
the appropriate monovalent reference for each analytical
technique.
Analysis of HIA results: During the HIA experiments, the
first observation was the haemolysis caused by the PhNAz-
based glycoclusters (15eP, 17eC, 21ePC, 19eA, and 13eG)
that incorporated the aromatic linker arm. The b values ob-
tained for the other glycoclusters were below 40, thereby
highlighting a modest improvement of inhibition relative to
the monovalent ligands. The glucose-cored glycoclusters
13aG and 13dG did not inhibit the agglutination of red
blood cells by the lectin (MIC>2 mm). Nevertheless, the
GalEG2NAz glycocluster 13cG displayed an MIC value of
250 mm. The number of amide bonds present in the linker
arm therefore affects the inhibition properties towards PA-
IL. The linker incorporating one amide bond can inhibit the
lectin-promoted agglutination, whereas linkers that incorpo-
rate either none or two amide bonds are not active. The por-
phyrin-based glycoclusters 15aP and 15dP display b values
of 159 and 2500, respectively, thus suggesting that these two
Evaluation of the second-generation glycoclusters (ITC):
ITC studies were then performed for the systematic analysis
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6257

S. Vidal, A. Imberty et al.
Table 3. Biochemical evaluation of glycoclusters.
Core
15aP[b]
15dP
15eP
HIA
MIC [mm]
ELLA
IC50 [mm]
SPR
IC50 [mm)]
Linker
b[a]
159
2500
–
b[a]
b[a]
GalEG3
GalEGGlyNAz
GalPhNAz
63
1
0.5
0.7
2
440
429
23
1.4
45
37.5
–
1.6
haemol.[d]
n.s.[e]
17aC[b]
17dC
17eC
GalEG3
GalEGGlyNAz
GalPhNAz
290
250
34
10
–
42
21
0.8
5.2
14
57
2.5
29
–
–
n.s.[e]
n.s.[e]
haemol.[e]
21aPC
21dPC
21ePC
GalEG3
GalEGGlyNAz
GalPhNAz
500
250
20
10
–
46
7
0.9
4.8
43
51
1.7
42
54
–
1.1
haemol.[d]
n.s.[e]
19aA[b]
19cA
19dA
19eA
19bA
GalEG3
500
500
625
20
20
4
–
–
36
7
14
5
6.1
0.5
1.0
1.2
0.8
0.6
144
58
50
3.5
105
GalEG2NAz
GalEGGlyNAz
GalPhNAz
50
21
9
haemol.[d]
>2000
[c]
GalEG2
15
15
13aG
13cG
13dG
13eG
GalEG3
>2000
250
–
40
–
9
2
4
24
175
75
2.8
1.9
23
31
35
–
GalEG2NAz
GalEGGlyNAz
GalPhNAz
>2000
1.7
haemol.[d]
–
n.s.[e]
–
n.s.[e]
[a] Improvement in IC50 relative to the appropriate monovalent reference for each linker arm. [b] Data from the literature.[27] [c] Monovalent reference
used for that glycocluster is compound 22. [d] haemol.=visual observation of red-cell haemolysis. [e] n.s.=not soluble under the experimental conditions
required, even in the presence of 5% of DMSO.
of the binding of glycoclusters to PA-IL. Unfortunately, the
issue of solubility was even more pronounced in the micro-
calorimetry experimental setup, and only the more soluble
compounds could be tested (Table 3). Glycoclusters built
with GalPhNAz (15eP, 17eC, 21ePC, 19eA, and 13eG) and
ligands based on porphyrin (15dP and 15eP) and on the
cone conformer of calix[4]arene (17dC and 17eC) were not
soluble enough for ITC measurements.
Soluble glycoclusters evaluated by ITC displayed KD
values in the nanomolar range with correspondingly high
b values (b>357) (Table 4). The glycoclusters with glucose
scaffold 13cG and 13dG bound less tightly to PA-IL with
KD values of 800 and 1720 nm. However, two situations
should be distinguished. First, the GalEGGlyNAz-based
multivalent ligands 21dPC and 19dA displayed an increase
in the binding stoichiometry (n) relative to the first-genera-
tion EG3-based glycoclusters (21aPC and 19aA). These
data suggest that these rigidified structures were not able to
interact simultaneously with four PA-IL monomers but
rather between two or three monomers. This feature was ac-
companied by lower enthalpic gains and entropic costs.
Changes in stoichiometry would imply different binding
mechanisms and therefore the interpretation of the linker
arm influence on the binding to PA-IL strictly based on the
thermodynamic profiles would be less appropriate. The glu-
cose-based glycocluster 13dG displayed a higher stoichiome-
try than ligand 13cG. The latter gave a KD value of 0.8 mm
6258
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
Table 4. Thermodynamic parameters of glycoclusters upon binding to
PA-IL (ITC) (T=298 K). (See Table 3 for the structures of the cores and
linkers.)
n[a]
ꢀDH[b]
ꢀTDS
[kJmolꢀ1
KD
b[c]
[kJmolꢀ1
]
G
]
[mm][b]
15aP[e]
15dP
0.46ꢁ0.04
60ꢁ7
23
0.33ꢁ0.06
0.4ꢁ0.1
451
not soluble
not soluble
34
not soluble
not soluble
60
15eP
17aC[f]
17dC
0.33ꢁ0.01
71ꢁ6
357
17eC
21aPC[f]
21dPC
21ePC
19aA[f]
19cA
0.26ꢁ0.01
0.38ꢁ0.01
98ꢁ9
78ꢁ2
0.200ꢁ0.005
0.239ꢁ0.003
750
753
40
not soluble
65
119
0.24ꢁ0.03
0.24ꢁ0.02
0.39ꢁ0.03
104ꢁ1
158ꢁ2
91ꢁ2
0.176ꢁ0.006
0.09ꢁ0.02
0.20ꢁ0.04
852
1138
891
19dA
19eA
19bA
13aG
13cG
53
not soluble
110
0.20ꢁ0.01
147ꢁ2
0.25ꢁ0.09
595
poor quality fit[d]
0.32ꢁ0.01
0.44ꢁ0.05
129.4ꢁ0.1
95
69
0.80ꢁ0.02
1.72ꢁ0.5
225
106
13dG
13eG
102ꢁ10
not soluble
[a] Stoichiometry. [b] Standard deviation is given over 2 or 3 experiments.
[c] Improvement of affinity relative to the appropriate monovalent refer-
ence (linker arm). [d] Heat signal increased during first injections.
[e] Data from the literature.[27] [f] Data from the literature.[26]
that was two times lower but remained a ligand with an in-
termediate affinity for PA-IL.
The second situation includes the second-generation li-
gands that display a 1:4 binding stoichiometry (nꢂ0.25) to
PA-IL monomers, thus indicating that all galactose residues
are engaged in a lectin binding site. The core scaffold that
resulted in higher affinity was the 1,3-alternate calix[4]arene.
With a stoichiometry of 0.24 and a KD value of 90 nm, ligand
19cA (Figure 8) showed a significant rise in enthalpic contri-
bution relative to first-generation ligand 19aA. This increase
must be related to the increase observed during the evalua-
tion of the monovalent reference 3. Interestingly, the en-
tropic cost of 19cA binding with PA-IL was also increased
significantly but did not counterbalance the enthalpic gain
as observed for other glycoclusters. The resulting improved
affinity of 90 nm corresponds to a 1138-fold increase relative
to its monovalent probe 3.
Figure 7. ELLA (top) and SPR (bottom) inhibition plots for the 1,3-alter-
~
*
nate calixarene core scaffold: ) GalEG3NAz 19aA, ) GalEG2NAz
!
&
^
19cA, ) GalEGGlyNAz 19dA, ) GalPhNAz 19eA, ) GalEG2NAz
19bA. Error bars represent standard deviation over triplicates.
Plotting relative potencies (b) versus b-GalOMe for each
of the glycoclusters (Figure 9) highlighted that the partial
cone and 1,3-alternate conformers of calix[4]arenes were the
most efficient core scaffolds. Among them, the 1,3-alternate
calix[4]arene that bore EG2NAz linker 19cA was about
twice the potency of the second-best glycocluster 19aA.
Multivalent ligands based on the cone conformer of calix[4]-
arene or on the glucose scaffold displayed modest potencies
independently of the structures of the linkers.
Figure 8. Column representation of the relative potency (b) of glycoclus-
ters versus b-d-GalOMe against PA-IL as measured by ITC.
Optimization of the structure of the linker arms yielded
the highest-affinity ligand 19cA for this bacterial lectin. By
changing the structure of the first-generation ligand 19aA to
the optimized glycocluster 19cA, it provided a twofold in-
crease in affinity towards PA-IL with only a subtle modifica-
tion of its molecular structure. Introduction of a second
amide moiety in the linker arm (i.e., glycocluster 19dA) did
not further improve the binding to PA-IL, but the observed
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6259

S. Vidal, A. Imberty et al.
gands 21aPC and 19dA displayed the same KD value and
also a comparable thermodynamic profile. This type of
graphical representation of the thermodynamic parameters
of the biomolecular interaction provides a rapid and general
overview of a large set of ligands to better identify the
common binding features among them. At a given tempera-
ture (298 K), dissociation constant frontiers (from millimolar
to micro- and nanomolar) illustrated the phenomenon some-
times referred as “enthalpy/entropy compensation.” This ap-
parently linear correlation between enthalpy and entropy
has been greatly discussed based on calorific capacities and
solvent reorganization.[62–65] The statistical significance of
this phenomenon is still an area of investigation. The pres-
ent studies illustrate that high affinity can be attained by
using different strategies when designing multivalent glyco-
clusters.
Conclusion
In conclusion, a series of three azido-functionalized carbo-
hydrate derivatives were prepared that incorporated ana-
logues of triethyleneglycol as a linker arm. The selected
structures incorporated one or two amide groups in a triethy-
leneglycol moiety for increased rigidity of the chain. A
phenyl aromatic ring was also introduced and displayed un-
precedentedly high affinity as a monovalent ligand of PA-
IL. The synthetic routes are based on a combination of effi-
cient and reliable glycosylation of peracetylated galactose,
amide bond formation, and azidation. The azido-functional-
ized derivatives were then conjugated by azide–alkyne
“click” chemistry methodology to either propargyl acetate
to obtain a monovalent probe or to tetra-propargylated core
scaffolds (porphyrin, three topologically isomeric calixar-
enes, and glucose) to afford the corresponding multivalent
glycoclusters.
Binding towards PA-IL was investigated by using four
bioanalytical techniques (HIA, ELLA, SPR, and ITC).
Some variability was observed when comparing data from
different techniques as observed previously by us and
others.[8,27,66] Differences appear in the scale of weak and
strong ligands since ELLA and ITC appeared more power-
ful than SPR to discriminate between the most efficient li-
gands. Some differences are also observed in the order of ef-
ficiency. The reason for such differences was due to the dif-
ferent conditions of measurement (e.g., buffer, concentra-
tion) and the fact that different physical phenomena were
measured (homogeneous solution versus surface adhesion).
Nevertheless, the different methods used are in broad agree-
ment and it was found that the higher efficiency of the 1,3-
alternate calixarene represents the best galactose-presenting
scaffold. Evaluation of the monovalent reference probes
showed comparable inhibition potencies and affinities,
except for the phenyl-containing compound, which is the
highest-affinity ligand for PA-IL reported so far. The ther-
modynamic signatures displayed unexpected differences.
Development of a molecular modeling approach provided
Figure 9. Raw ITC data (top) obtained by injecting a solution of 19cA
(120 mm) in a solution containing PA-IL (50 mm). Corresponding integrat-
ed titration curve (bottom).
KD value was almost identical to the first-generation EG3-
based glycocluster 19aA.
Finally, glycocluster 19bA that bore the EG2-based linker
arm displayed only a slight affinity decrease for PA-IL with
a KD value of 250 nm. The free energy was composed of
a stronger enthalpic contribution balanced by a higher en-
tropic cost relative to the EG2-based glycocluster 19aA. The
length of the linker arm could therefore be reduced to only
two ethyleneglycol units while maintaining the simultaneous
binding of four PA-IL monomers (n value of 0.20).
Plotting the enthalpic (DH) versus entropic (ꢀTDS) con-
tributions measured by ITC for each multivalent glycoclus-
ter underlined the differences in the thermodynamic profiles
related to the modifications of the ligandꢄs chemical struc-
ture (Figure 10). Ligands 21dPC and 19bA have almost
identical affinities (0.24 and 0.25 mm, respectively) but very
different thermodynamic signatures. On the other hand, li-
Figure 10. Enthalpic contribution related to entropic cost.
6260
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
General procedure for deacetylation of carbohydrates (method B): The
acetylated glycoside was suspended in distilled MeOH, ultrapure water,
and ultrapure triethylamine (4:1:1, v/v/v). The mixture was stirred under
argon at room temperature for 2 to 4 d. Solvents were evaporated, co-
evaporated with toluene, and the resulting white foam was dissolved in
ultrapure water (5 mL) and freeze-dried to afford pure hydroxylated gly-
coconjugates.
access to the geometrical parameters and flexibility of the
monovalent ligands.
During the linker-arm optimization, solubility appeared to
be a major issue for several synthesized multivalent glyco-
clusters. This study clearly illustrates the complexity of pre-
dicting the influence of “rigidifying” the structure on the
thermodynamic parameters. Indeed, significant improve-
ment in the affinity of the second-generation rigidified mul-
tivalent ligands arose from a strong increase in enthalpic
contribution rather than a decrease in entropic cost as ini-
tially hypothesized. Among rigidified glycoclusters, the
shifts in stoichiometry highlighted variations in binding
mechanisms, which made comparisons of thermodynamic
patterns inappropriate.
The present study resulted in optimized molecular archi-
tectures as high-affinity ligands of PA-IL for potential thera-
peutic applications as anti-adhesive drugs that target Pseu-
domonas aeruginosa bacterial infections. When comparing
those results with other multivalent ligands developed
against this bacterial target,[22,25] we have reported here very
potent structures with the advantages of relative low molec-
ular weights and straightforward synthetic strategies.
Binding studies
Hemagglutination inhibition assays (HIA): Hemagglutination inhibition
assays (HIA) were performed in U-shaped 96-well microtitre plates.
Rabbit erythrocytes were purchased from Biomꢀrieux and used without
further washing. Erythrocytes were diluted to a 4% solution in NaCl
(150 mm). PA-IL solutions of 2 mgmLꢀ1 were prepared in TRIS-HCl
20 mm (TRIS=tris(hydroxymethyl)aminomethane), NaCl 100 mm, and
CaCl2 100 mm. The hemagglutination unit (HU) was first obtained by the
addition of the 4% erythrocyte solution (25 mL) to aliquots (25 mL) of se-
quential (twofold) lectin dilutions. The mixture was incubated at 258C
for 60 min. The HU was measured as the minimum lectin concentration
required to observe hemagglutination. For the following lectin-inhibition
assays, lectin concentrations of 4 HU were used. For PA-IL, this concen-
tration was found to be 8 mgmLꢀ1. Subsequent inhibition assays were
then carried out by the addition of lectin solution (12.5 mL, at the re-
quired concentration) to sequential dilutions (25 mL) of glycoclusters, mo-
nomer molecules, and controls. These solutions were incubated at 258C
for 2 h, then 4% erythrocyte solution (12.5 mL) was added, followed by
an additional incubation at 258C for 30 min. The minimum inhibitory
concentration for each molecule was determined for each duplicate.
Determination of lectin concentration by using ELLA: 96-Well microtiter
plates (Nunc Maxisorb) were coated with a-PAA-Gal (PAA=polyacryl-
amide) for PA-IL (Lectinity Holding, Inc.): 100 mL of 5 mgmLꢀ1 in car-
bonate buffer, pH 9.6 for 1 h at 378C, then blocking at 378C for 1 h with
100 mL per well of 3% (w/v) bovine serum albumin (BSA) in phosphate
buffer solution (PBS). Lectin solutions (100 mL) were diluted (1:2) start-
ing from 30 mgmLꢀ1. After 1 h incubation at 378C and three washes with
T-PBS (PBS that contained 0.05% Tween 20), horseradish peroxidase
(HRP)–streptavidin conjugate (100 mL; dilution 2:5000; Boehringer–
Mannheim) was added and left for 1 h at 378C. Coloration was devel-
oped by using 100 mL per well of 0.05m phosphate/citrate buffer that con-
tained o-phenylenediamine dihydrochloride (0.4 mgmLꢀ1) and urea hy-
drogen peroxide (0.4 mgmLꢀ1) (OPD kit, Sigma–Aldrich) for 15 min and
stopped with sulfuric acid (50 mL, 30%). Absorbance was then read at
490 nm using a microtiter plate reader (BioRad 680). The concentration
of biotinylated lectins was determined by plotting the relative absorbance
versus lectin concentration. The concentration that led to the highest re-
sponse in the linear area was selected as the standard lectin concentra-
tion for the subsequent inhibition experiments. The final concentrations
were 0.5 mgmLꢀ1 for PA-IL.
Experimental Section
Synthesis
Materials and methods: All reagents were commercial and used without
further purification. Solvents were distilled over CaH2 (CH2Cl2) or Mg/I2
(MeOH). All the reactions were performed under an argon atmosphere.
Reactions under microwave activation were performed using a Biotage
Initiator system. NMR spectra were recorded at 293 K, unless otherwise
stated, using a 400 MHz spectrometer. Shifts are referenced relative to
deuterated solvent residual peaks. Complete signal assignments from 1D
and 2D NMR spectroscopy were based on COSY, HSQC, and HMBC
correlations. High-resolution (HR-ESI-QToF) mass spectra were record-
ed using a Bruker MicroToF-Q II XL spectrometer. Thin-layer chroma-
tography (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) followed by heating. Silica gel column chromatography was per-
formed with silica gel Si-60 (40–63 mm).
Determination of inhibition potency (IC50) by using ELLA: ELLAs were
conducted using 96-well microtiter plates (Nunc Maxisorb) coated with
a-PAA-Gal for PA-IL (Lectinity Holding, Inc.): 100 mL of 5 mgmLꢀ1 in
carbonate buffer, pH 9.6 for 1 h at 378C, then blocking at 378C for 1 h
with 100 mL per well of 3% (w/v) BSA in PBS. Inhibitor solutions
(50 mL) were submitted to serial dilutions (1:3) with PBS-BSA 0.3% (w/
v). Then, biotinylated lectin solution (50 mL at the appropriate concentra-
tion) was added in each well and the plates were incubated for 1 h at
378C. After three washings with T-PBS (50 mL, 5 min.), horseradish per-
oxidase (HRP)–streptavidin conjugate (100 mL; dilution 2:5000; Boeh-
ringer–Mannheim) was added and left for 1 h at 378C. After three more
washings, coloration was developed using 100 mL per well of 0.05m phos-
phate/citrate buffer that contained o-phenylenediamine dihydrochloride
The numbering of molecules is written using a lower-case letter for the
spacer arm used (a–d) but also a capital letter (G for glucose, P for por-
phyrin, A for the 1,3-alternate, C for the cone, and PC for the partial
cone conformers of calixarenes) to provide some information about the
general structure of the glycocluster.
General procedure for 1,3-dipolar cycloadditions (method A): The alkyne
derivative, copper iodide, N,N-diisopropylethylamine (DIPEA), and
azido derivative in DMF were introduced into a Biotage Initiator 2–5 mL
vial. The solution was sonicated for 30 s. The vial was sealed with
a septum cap and heated at 1108C for 15 min under microwave irradia-
tion (solvent absorption level: high). If the product was partially soluble
in water, the crude mixture was concentrated and co-evaporated with tol-
uene three times before flash chromatography. If the product was not
soluble in water, the mixture was diluted with EtOAc (250 mL). The or-
ganic layer was washed with 150 mL portions of 1n HCl, saturated
NaHCO3, water, EDTA 0.1m, water, and brine successively. The organic
layer was dried (Na2SO4), filtered, and evaporated. The crude product
was purified by flash silica gel column chromatography to afford the de-
sired cycloadducts.
) ) (OPD kit,
(0.4 mgmLꢀ1 and urea hydrogen peroxide (0.4 mgmLꢀ1
Sigma–Aldrich) for 15 min and stopped with sulfuric acid (50 mL, 30%).
Absorbance was then read at 490 nm using a microtiter plate reader
(BioRad 680) and transformed in inhibition percentage with the help of
positive and negative controls. Plots of inhibition percentage versus in-
hibitor concentration and sigmoidal fitting provided the IC50 determina-
tion.
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6261

S. Vidal, A. Imberty et al.
Isothermal titration microcalorimetry (ITC): Recombinant lyophilized
PA-IL was dissolved in buffer (0.1m TRIS-HCl, 6 mm CaCl2, pH 7.5) and
degassed. Protein concentration (between 50 and 120 mm depending on
the ligand affinity) was checked by measurement of optical density by
using a theoretical molar extinction coefficient of 28000. Carbohydrate li-
gands were dissolved directly into the same buffer, degassed, and placed
in the injection syringe (concentration range : monovalent 2–0.7 mm,
multivalent 0.2–0.12 mm). ITC was performed using a VP-ITC MicroCa-
lorimeter from MicroCal Incorporated. PA-IL was placed into the
1.4478 mL sample cell, at 258C. Titration was performed with 10 mL in-
jections of carbohydrate ligands every 300 s. Data were fitted using the
“one-site model” using MicroCal Origin 7 software according to standard
procedures. Fitted data yielded the stoichiometry (n), the association con-
stant (KA), and the enthalpy of binding (DH). Other thermodynamic pa-
rameters (i.e., changes in free energy DG and entropy DS) were calculat-
ed from the equation DG=DHꢀTDS=ꢀRTlnKa in which T is the abso-
lute temperature and R=8.314 Jmolꢀ1 Kꢀ1. Two or three independent ti-
trations were performed for each ligand tested.
ring), plane T (plane from 3 atoms of the triazole ring), plane N (plane
from two extremities of the T centroid normal and the C5 atom of the
triazole ring), plane V (plane from centroid S, centroid T, and the endo-
cyclic oxygen of the carbohydrate). Calculations of the geometrical pa-
rameters were performed by using these references in the Sybyl spread-
sheet.
X-ray crystallography: A crystal suitable for X-ray crystallography was
selected and mounted using an Xcalibur k-geometry diffractometer (Agi-
lent Technologies UK Ltd) equipped with an Eos CCD detector and
using MoKa radiation (l =0.71073 ꢃ). Intensities were collected at room
temperature by means of the CrysalisPro software. Reflection indexing,
unit-cell parameters refinement, Lorentz-polarization correction, peak in-
tegration, and background determination were carried out with the Cry-
salisPro software. An analytical absorption correction was applied using
the modeled faces of the crystal.[66] The structures were solved by direct
methods with SIR97[67] and the least-squares refinement on F2 was ach-
ieved with the CRYSTALS software.[68] All non-hydrogen atoms were re-
fined anisotropically. The hydrogen atoms were all located in a difference
map, but those attached to carbon atoms were repositioned geometrical-
ly. The hydrogen atoms were initially refined with soft restraints on the
Surface plasmon resonance (SPR): SPR inhibition experiments were per-
formed using a Biacore 3000 instrument at 258C. Measurements were
carried out on two channels with two immobilized sugars: a-l-fucose
(channel 1) and a-d-galactose (channel 2). Immobilization of sugars was
performed at 258C using running buffer (HBS) at 5 mLminꢀ1. Immobili-
zation on each channel (CM5 Chip) was performed independently as fol-
lows. First, the channel was activated by injecting a fresh mixture of
EDC/NHS (35 mL, 420 s). Then a solution of streptavidin (100 mgmLꢀ1 in
AcONa pH 5 buffer) was injected (50 mL, 600 s). The remaining reactive
species were quenched by injecting ethanolamine (1m, 35 mL, 420 s) into
the solution. Finally, a solution of the desired biotinylated-polyacryl-
amide–sugar (lectinity, 200 mgmLꢀ1) was coated onto the surface (50 mL,
600 s) through streptavidin–biotin interaction. This procedure led to
804 RU (resonance units) (fucoside) and 796 RU (galactoside) of immo-
bilized sugars on channels 1 and 2, respectively. Inhibition experiments
were performed with the galactosylated channel 2 and plots represent
substracted data (channel 2ꢀchannel 1). The running buffer for PA-IL
experiments was HEPES 10 mm, NaCl 150 mm, CaCl2 10 mm, Tween P20
0.005%, pH 7.4. Inhibition studies consisted of the injection (150 mL,
10 mLminꢀ1, dissociation 120 s) of incubated (>1 h, RT) mixtures of PA-
IL (5 mm) and various concentrations of inhibitor (twofold cascade dilu-
tions). For each inhibition assay, PA-IL (5 mm) without inhibitor was in-
jected to observe the full adhesion of the lectin onto the sugar-coated sur-
face (0% inhibition). The CM5 chip was fully regenerated by successive
injections of d-galactose (2ꢅ30 mL, 100 mm in running buffer). Binding
was measured as RU over time after blank subtraction, and data were
then evaluated using the BIAevaluation Software version 4.1. For IC50
evaluation, the response (Reqꢀfitted) was considered to be the amount of
lectin bound to the carbohydrate-coated surface at equilibrium in the
presence of a defined concentration of inhibitor. Inhibition curves were
obtained by plotting the percentage of inhibition against the inhibitor
concentration (on a logarithmic scale) by using Origin 7.0 software (Ori-
ginLab Corp.), and IC50 values were extracted from sigmoidal fit of the
inhibition curve.
ꢀ
bond lengths and angles to regularize their geometry (C H in the range
0.93 to 0.98 ꢃ) and Uiso(H) (in the range 1.2 to 1.5 times Ueq of the
parent atom), after which the positions were refined with riding con-
straints. The drawings were achieved with Diamond 3.2 software.[70]
CCDC-845766 (1e) contains the supplementary crystallographic data for
this paper. These data can be obtained free of charge from The Cam-
bridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_re-
quest/cif.
Acknowledgements
The authors thank the Universitꢀ Claude Bernard Lyon 1 and the CNRS
for financial support. S.C. thanks the Rꢀgion Rhꢆne-Alpes (Cluster de
Recherche Chimie) and the CNRS (Programme Interdisciplinaire:
Chimie pour le Dꢀveloppement Durable) for additional funding. Dr. F.
Albrieux, C. Duchamp, and N. Henriques are gratefully acknowledged
for mass spectrometry analyses. X-ray crystallographic analyses were per-
formed by Dr. E. Jeanneau at the Centre de Diffractomꢀtrie Henri Long-
chambon of Universitꢀ Claude Bernard Lyon 1. Financial help is ac-
knowledged from Vaincre la Mucoviscidose and GDR Pseudomonas. S.V.
and A.I. are grateful to financial support from the COST Action CM-
1102 MultiGlycoNano. A.I. acknowledges funding from ANR Glycoas-
terix. M.W. acknowledges the Czech Ministry of Education
(MSM0021622413, ME08008). Part of the research has received funding
from the European Communityꢄs Seventh Framework Programme under
grant agreement no. 205872.
Molecular modeling: Construction of monovalent references 22, 3, 7, and
10 was performed with Tripos Sybyl 7.3 (Tripos Associates, St. Louis,
MO). Atomic partial charges were then calculated (MOPAC/MNDO) for
the linker and set up according to Tripos[69] for the galactose moieties.
Linkers and carbohydrates epitopes were connected and the charges
were derived and symmetrized to obtain a neutral global charge. The re-
sulting molecule was then energy-minimized by using the conjugate gra-
dient method and TRIPOS Force Field with addition of carbohydrate pa-
rameters. This minimization led to 14.4, 14.9, 15.2, and 14.6 kcalmolꢀ1 for
22, 3, 7, and 10 respectively. For each compound, the torsion angle F was
checked to be close to 3008 (exo-anomeric effect). A systematic search
was performed for each compound with 9, 8, 7, and 4 rotatable bonds for
22, 3, 7, and 10, respectively, and a 20 kcalmolꢀ1 energetic cutoff. These
systematic searches led to 48548, 44324, 26988, and 7624 reachable con-
formers for 22, 3, 7, and 10, respectively. Geometric references and
planes were defined for each molecule as follows: S (centroid of all 6
atoms of the carbohydrate ring), T (centroid of all 5 atoms of the triazole
[1] C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357–2364.
[2] R. A. Dwek, Chem. Rev. 1996, 96, 683–720.
[3] H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637–674.
[4] J. C. Paulson, O. Blixt, B. E. Collins, Nat. Chem. Biol. 2006, 2, 238–
248.
[5] A. Varki, Glycobiology 1993, 3, 97–130.
[6] A. Varki, R. D. Cummings, J. D. Esko, H. H. Freeze, P. Stanley,
C. R. Bertozzi, G. W. Hart, M. E. Etzler, Essentials of glycobiology,
Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N. Y.,
2009, p. .
[7] Y. C. Lee, R. T. Lee, Acc. Chem. Res. 1995, 28, 321–327.
[8] J. J. Lundquist, E. J. Toone, Chem. Rev. 2002, 102, 555–578.
[9] M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998,
110, 2908–2953; Angew. Chem. Int. Ed. 1998, 37, 2754–2794.
[10] N. Jayaraman, Chem. Soc. Rev. 2009, 38, 3463–3483.
[11] M. Lahmann in Architectures of Multivalent Glycomimetics for
Probing Carbohydrate-Lectin Interactions/Glycoscience and Microbi-
6262
www.chemeurj.org
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 6250 – 6263

Lectin–Carbohydrate Interactions
FULL PAPER
al Adhesion, Vol. 288 (Eds.: T. K. Lindhorst, S. Oscarson), Springer,
2009, pp. 17–65.
[12] R. J. Pieters, Org. Biomol. Chem. 2009, 7, 2013–2025.
[13] N. Sharon, Biochim. Biophys. Acta Gen. Subj. 2006, 1760, 527–537.
[14] A. Imberty, Y. M. Chabre, R. Roy, Chem. Eur. J. 2008, 14, 7490–
7499.
[15] T. K. Dam, C. F. Brewer, Chem. Rev. 2002, 102, 387–430.
[16] T. K. Dam, C. F. Brewer, Biochemistry 2008, 47, 8470–8476.
[17] O. Sulak, E. Lameignere, M. Wimmerova, A. Imberty in Specificity
and Affinity Studies in Lectin/Carbohydrate Interactions, Vol. 35 The
Royal Society of Chemistry, 2009, pp. 357–372.
[18] M. Touaibia, R. Roy, Mini-Rev. Med. Chem. 2007, 7, 1270–1283.
[19] B. Blanchard, A. Nurisso, E. Hollville, C. Tꢀtaud, J. Wiels, M.
Pokornꢁ, M. Wimmerovꢁ, A. Varrot, A. Imberty, J. Mol. Biol. 2008,
383, 837–853.
[20] G. Cioci, E. P. Mitchell, C. Gautier, M. Wimmerova, D. Sudakevitz,
S. Pꢀrez, N. Gilboa-Garber, A. Imberty, FEBS Lett. 2003, 555, 297–
301.
[21] N. Gilboa-Garber, Methods Enzymol. 1982, 83, 378–385.
[22] R. U. Kadam, M. Bergmann, M. Hurley, D. Garg, M. Cacciarini,
M. A. Swiderska, C. Nativi, M. Sattler, A. R. Smyth, P. Williams, M.
Cꢁmara, A. Stocker, T. Darbre, J.-L. Reymond, Angew. Chem. Int.
Ed. 2011, 50, 10631–10635.
[23] C. Chemani, A. Imberty, S. de Bentzman, P. Pierre, M. Wimmerovꢁ,
B. P. Guery, K. Faure, Infect. Immun. 2009, 77, 2065–2075.
[24] H.-P. Hauber, M. Schulz, A. Pforte, D. Mack, P. Zabel, U. Schu-
macher, Int. J. Med. Sci. 2008, 5, 371–376.
[25] Y. M. Chabre, D. Giguꢇre, B. Blanchard, J. Rodrigue, S. Rocheleau,
M. Neault, S. Rauthu, A. Papadopoulos, A. A. Arnold, A. Imberty,
R. Roy, Chem. Eur. J. 2011, 17, 6545–6562.
[41] Z. Zhang, E. A. Merritt, M. Ahn, C. Roach, Z. Hou, C. L. M. J. Ver-
linde, W. G. J. Hol, E. Fan, J. Am. Chem. Soc. 2002, 124, 12991–
12998.
[42] P. I. Kitov, J. M. Sadowska, G. Mulvey, G. D. Armstrong, H. Ling,
N. S. Pannu, R. J. Read, D. R. Bundle, Nature 2000, 403, 669–672.
[43] S. Andrꢀ, B. Liu, H.-J. Gabius, R. Roy, Org. Biomol. Chem. 2003, 1,
3909–3916.
[44] C. Bouillon, A. Meyer, S. Vidal, A. Jochum, Y. Chevolot, J.-P. Cloar-
ec, J.-P. Praly, J.-J. Vasseur, F. Morvan, J. Org. Chem. 2006, 71,
4700–4702.
[45] Y. Chevolot, C. Bouillon, S. Vidal, F. Morvan, A. Meyer, J. P. Cloar-
ec, A. Jochum, J. P. Praly, J. J. Vasseur, E. Souteyrand, Angew.
Chem. 2007, 119, 2450–2454; Angew. Chem. Int. Ed. 2007, 46, 2398–
2402.
[46] J. Li, S. Zacharek, X. Chen, J. Wang, W. Zhang, A. Janczuk, P. G.
Wang, Bioorg. Med. Chem. 1999, 7, 1549–1558.
[47] F. Morvan, A. Meyer, A. Jochum, C. Sabin, Y. Chevolot, A. Imberty,
J.-P. Praly, J.-J. Vasseur, E. Souteyrand, S. Vidal, Bioconjugate Chem.
2007, 18, 1637–1643.
[48] N. Nagahori, S.-I. Nishimura, Biomacromolecules 2001, 2, 22–24.
[49] Z. Szurmai, L. Szabo, A. Liptak, Acta Chim. Hung. 1989, 126, 259–
269.
[50] C.-S. Tsai, T.-B. Yu, C.-T. Chen, Chem. Commun. 2005, 4273–4275.
[51] G. Fischer, Chem. Soc. Rev. 2000, 29, 119–127.
[52] N. Garber, U. Guempel, A. Belz, N. Gilboa-Garber, R. J. Doyle,
Biochim. Biophys. Acta Gen. Subj. 1992, 1116, 331–333.
[53] J. L. Xue, S. Cecioni, L. He, S. Vidal, J.-P. Praly, Carbohydr. Res.
2009, 344, 1646–1653.
[54] S. Cecioni, M. Almant, J.-P. Praly, S. Vidal in Synthesis of Azido-
Functionalized Carbohydrates for the Design of Glycoconjugates,
ˇ
[26] S. Cecioni, R. Lalor, B. Blanchard, J.-P. Praly, A. Imberty, S. E. Mat-
thews, S. Vidal, Chem. Eur. J. 2009, 15, 13232–13240.
[27] S. Cecioni, S. Faure, U. Darbost, I. Bonnamour, H. Parrot-Lopez, O.
Roy, C. Taillefumier, M. Wimmerovꢁ, J.-P. Praly, A. Imberty, S.
Vidal, Chem. Eur. J. 2011, 17, 2146–2159.
[28] J.-F. Nierengarten, J. Iehl, V. Oerthel, M. Holler, B. M. Illescas, A.
Munoz, N. Martin, J. Rojo, M. Sanchez-Navarro, S. Cecioni, S.
Vidal, K. Buffet, M. Durka, S. P. Vincent, Chem. Commun. 2010, 46,
3860–3862.
Vol. 1 (Ed. P. Kovꢁc), CRC Press, 2012, pp. 175–180.
[55] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless, Angew.
Chem. 2002, 114, 2708–2711; Angew. Chem. Int. Ed. 2002, 41, 2596–
2599.
[56] C. W. Tornøe, C. Christensen, M. Meldal, J. Org. Chem. 2002, 67,
3057–3064.
[57] M. Meldal, C. W. Tornøe, Chem. Rev. 2008, 108, 2952–3015.
[58] F. Perez-Balderas, J. Morales-Sanfrutos, F. Hernandez-Mateo, J.
Isac-Garcꢈa, F. Santoyo-Gonzalez, Eur. J. Org. Chem. 2009, 2441–
2453.
[59] Y. Liu, C.-F. Ke, H.-Y. Zhang, J. Cui, F. Ding, J. Am. Chem. Soc.
2008, 130, 600–605.
[29] S. Cecioni, V. Oerthel, J. Iehl, M. Holler, D. Goyard, J.-P. Praly, A.
Imberty, J.-F. Nierengarten, S. Vidal, Chem. Eur. J. 2011, 17, 3252–
3261.
[30] Z. H. Soomro, S. Cecioni, H. Blanchard, J.-P. Praly, A. Imberty, S.
Vidal, S. E. Matthews, Org. Biomol. Chem. 2011, 9, 6587–6597.
[31] I. Otsuka, B. Blanchard, R. Borsali, A. Imberty, T. Kakuchi, Chem-
BioChem 2010, 11, 2399–2408.
[60] W. B. Turnbull, A. H. Daranas, J. Am. Chem. Soc. 2003, 125, 14859–
14866.
[61] A. Imberty, S. Pꢀrez, M. Hricovꢈni, R. N. Shah, J. P. Carver, Int. J.
Biol. Macromol. 1993, 15, 17–23.
[32] D. Sicard, S. Cecioni, M. Iazykov, Y. Chevolot, S. E. Matthews, J.-P.
Praly, E. Souteyrand, A. Imberty, S. Vidal, M. Phaner-Goutorbe,
Chem. Commun. 2011, 47, 9483–9485.
[33] D. Deniaud, K. Julienne, S. G. Gouin, Org. Biomol. Chem. 2011, 9,
966–979.
[34] V. Wittmann, S. Seeberger, Angew. Chem. 2004, 116, 918–921;
Angew. Chem. Int. Ed. 2004, 43, 900–903.
[35] S. Andrꢀ, Z. Pei, H.-C. Siebert, O. Ramstrçm, H.-J. Gabius, Bioorg.
Med. Chem. 2006, 14, 6314–6326.
[62] K. N. Houk, A. G. Leach, S. P. Kim, X. Zhang, Angew. Chem. 2003,
115, 5020–5046; Angew. Chem. Int. Ed. 2003, 42, 4872–4897.
[63] L. Liu, Q.-X. Guo, Chem. Rev. 2001, 101, 673–696.
[64] E. Grunwald, C. Steel, J. Am. Chem. Soc. 1995, 117, 5687–5692.
[65] A. Cornish-Bowden, J. Biosci. 2002, 27, 121–126.
[66] R. C. Clark, J. S. Reid, Acta Crystallogr. Sect. A 1995, 51, 887–897.
[67] A. Altomare, M. C. Burla, M. Camalli, G. L. Cascarano, C. Giaco-
vazzo, A. Guagliardi, A. G. G. Moliterni, G. Polidori, R. Spagna, J.
Appl. Crystallogr. 1999, 32, 115–119.
[36] E. Fan, E. A. Merritt, C. L. M. J. Verlinde, W. G. J. Hol, Curr. Opin.
Struct. Biol. 2000, 10, 680–686.
[68] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J.
Watkin, J. Appl. Crystallogr. 2003, 36, 1487.
[37] E. Fan, Z. Zhang, W. E. Minke, Z. Hou, C. L. M. J. Verlinde,
W. G. J. Hol, J. Am. Chem. Soc. 2000, 122, 2663–2664.
[38] P. I. Kitov, H. Shimizu, S. W. Homans, D. R. Bundle, J. Am. Chem.
Soc. 2003, 125, 3284–3294.
[39] E. A. Merritt, Z. Zhang, J. C. Pickens, M. Ahn, W. G. J. Hol, E. Fan,
J. Am. Chem. Soc. 2002, 124, 8818–8824.
[69] A. Imberty, E. Bettler, M. Karababa, K. Mazeau, P. Petrova, S.
Pꢀrez in, Building Sugars: The Sweet Part of Structural Biology
(Eds.: M. Vijayan, N. Yathindra, A. S. Kolaskar), Indian Academy
of Sciences and Universities Press, Hyderabad, 1999, pp. 392–409.
[70] Diamond version 3.2, M. Berndt, K. Brandenburg, H. Putz, 1999,
Bonn (Germany), Crystal Impact GbR.
[40] D. Schwefel, C. Maierhofer, J. G. Beck, S. Seeberger, K. Diederichs,
H. M. Mçller, W. Welte, V. Wittmann, J. Am. Chem. Soc. 2010, 132,
8704–8719.
Received: January 2, 2012
Published online: April 4, 2012
Chem. Eur. J. 2012, 18, 6250 – 6263
ꢂ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6263