Symmetric dithiodigalactoside: Strategic combination of binding studies and detection of selectivity between a plant toxin and human lectins

Article abstract of DOI:10.1039/c0ob01235a

Thioglycosides offer the advantage over O-glycosides to be resistant to hydrolysis. Based on initial evidence of this recognition ability for glycosyldisulfides by screening dynamic combinatorial libraries, we have now systematically studied dithiodigalactoside on a plant toxin (Viscum album agglutinin) and five human lectins (adhesion/growth-regulatory galectins with medical relevance e.g. in tumor progression and spread). Inhibition assays with surface-presented neoglycoprotein and in solution monitored by saturation transfer difference NMR spectroscopy, flanked by epitope mapping, as well as isothermal titration calorimetry revealed binding properties to VAA (K _a: 1560 ± 20 M^-1). They were reflected by the structural model and the affinity on the level of toxin-exposed cells. In comparison, galectins were considerably less reactive, with intrafamily grading down to very minor reactivity for tandem-repeat-type galectins, as quantitated by radioassays for both domains of galectin-4. Model building indicated contact formation to be restricted to only one galactose moiety, in contrast to thiodigalactoside. The tested glycosyldisulfide exhibits selectivity between the plant toxin and the tested human lectins, and also between these proteins. Therefore, glycosyldisulfides have potential as chemical platform for inhibitor design.

Full text of DOI:10.1039/c0ob01235a

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 5445
www.rsc.org/obc
PAPER
Symmetric dithiodigalactoside: strategic combination of binding studies and
detection of selectivity between a plant toxin and human lectins
a
b
c,d
e
c
Sonsoles Mart ´ı n-Santamar ´ı a, Sabine Andr e´ , Eliza Buzamet, R e´ mi Caraballo, Gloria Fern a´ ndez-Cureses,
f
f
f
a
f
Maria Morando, Jo a˜ o P. Ribeiro, Karla Ram ´ı rez-Gualito, Beatriz de Pascual-Teresa, F. Javier Ca n˜ ada,
Margarita Men e´ ndez, Olof Ramstr o¨ m, Jes u´ s Jim e´ nez-Barbero,* Dolores Sol ´ı s and
c,d
e
f
c,d
b
Hans-Joachim Gabius
Received 22nd December 2010, Accepted 21st April 2011
DOI: 10.1039/c0ob01235a
Thioglycosides offer the advantage over O-glycosides to be resistant to hydrolysis. Based on initial
evidence of this recognition ability for glycosyldisulfides by screening dynamic combinatorial libraries,
we have now systematically studied dithiodigalactoside on a plant toxin (Viscum album agglutinin) and
five human lectins (adhesion/growth-regulatory galectins with medical relevance e.g. in tumor
progression and spread). Inhibition assays with surface-presented neoglycoprotein and in solution
monitored by saturation transfer difference NMR spectroscopy, flanked by epitope mapping, as well as
-
1
isothermal titration calorimetry revealed binding properties to VAA (K : 1560 ± 20 M ). They were
a
reflected by the structural model and the affinity on the level of toxin-exposed cells. In comparison,
galectins were considerably less reactive, with intrafamily grading down to very minor reactivity for
tandem-repeat-type galectins, as quantitated by radioassays for both domains of galectin-4. Model
building indicated contact formation to be restricted to only one galactose moiety, in contrast to
thiodigalactoside. The tested glycosyldisulfide exhibits selectivity between the plant toxin and the tested
human lectins, and also between these proteins. Therefore, glycosyldisulfides have potential as chemical
platform for inhibitor design.
Introduction
such as concanavalin A, the selection is shifting to medically
relevant proteins, that is, to distinct biohazardous toxins and
human lectins. Technically, the aim of identifying promising
anti-lectin lead compounds prompted the introduction of high-
The growing insights into the molecular basis of physiological
processes guide the selection of new targets for drug design.
Owing to the unsurpassed capacity of glycans to store biological
information, the development of a wide array of respective
receptors (lectins) and their dynamic interplay triggering specific
cellular responses e.g. in growth regulation, the design of com-
pounds that potently interfere with lectin-glycan recognition has
2
throughput screening and diverse types of library approaches.
Along these lines, we have targeted on two such classes of
galactoside-binding lectins (toxic AB-type lectins, here Viscum
album agglutinin (VAA), and the family of adhesion/growth-
regulatory galectins with its three groups of proto-, chimera-
1
become a challenge. Initially mostly addressed for plant lectins
3
and tandem-repeat-type proteins ) and applied the dynamic
combinatorial library approach based on thiol-disulfide exchange
4
a
with thioglycosides. As previously demonstrated, the dynamic
Departamento de Qu ´ı mica, Facultad de Farmacia, Universidad San Pablo
CEU, Boadilla del Monte, 28668, Madrid, Spain
library approach is indeed a viable route to identifying new
disulfide ligands for lectins. Nevertheless, the main objective of
b
4d
Institute of Physiological Chemistry, Faculty of Veterinary Medicine,
Ludwig-Maximilians-University Munich, Veterin a¨ rstr. 13, 80539, Munich,
Germany
this study has not been to further explore this methodology, but
rather to evaluate the properties of a ligand that was previously
identified.
In principle, S-glycosides have the pharmacodynamic advantage
of being resistant to hydrolytic cleavage. However, the substitution
engenders enhancement of flexibility and alteration of the topol-
ogy (C–S/C–O bond lengths: 1.78 A vs. 1.41 A; C–S–C/C–O–C
bond angles: 98 vs. 117 ). These differences from O-glycosides
put bioactivity of these derivatives into question. However, by
c
Instituto de Qu ´ı mica F ´ı sica Rocasolano, Consejo Superior de Investigaciones
Cient ´ı ficas, Serrano 119, 28006, Madrid, Spain
d
Centro de Investigaci o´ n Biom e´ dica en Red de Enfermedades Respiratorias
(
CIBERES), Bunyola, Mallorca, Illes Balears, Spain
e
Department of Chemistry, KTH-Royal Institute of Technology, Teknikrin-
gen 30, 10044, Stockholm, Sweden
˚
˚
f
Chemical and Physical Biology, Centro de Investigaciones Biol o´ gicas,
◦
◦
5
Consejo Superior de Investigaciones Cient ´ı ficas, Ramiro de Maeztu 9, 28040,
Madrid, Spain
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5445

View Article Online
testing thiodigalactoside (TDG) in inhibition assays, the S-
glycosidic bond has been shown to interact with AB-type plant
toxins and galectins, with a tendency for an increased entropic
penalty compensated by an enthalpic gain relative to lactose in the
Assaying ligand properties on the toxin
A convenient test method is the analysis of the extent of
lectin binding to surface-presented glycans in the absence and
presence of compounds examined for inhibitory activity. Using
lactose as interaction partner by adsorbing a lactose-bearing
neoglycoprotein to the surface of microtiter plate wells, binding of
biotinylated VAA was first ascertained to be saturable and blocked
by presence of cognate sugar but not mannose or glucose. The
dependence on binding the carbohydrate specifically was further
substantiated by lack of VAA interaction with a neoglycoprotein
constituted by the same carrier protein (bovine serum albumin) but
presenting mannose as sugar headgroup. Next, conditions were
defined to ensure signal intensity in the linear range by varying
the lectin concentration in order to reach optimal sensitivity.
The ensuing inhibitory titrations at constant lectin concentration
with increasing concentrations of tested compounds resulted in
a reduction of the extent of binding of the labelled toxin to the
ligand-bearing matrix (Fig. 1). TDG was a slightly more efficient
inhibitor than lactose when expressed in the measured IC50-values
6
thermodynamic balance sheet. Moving beyond thioglycosides,
the synthesis of glycosyldisulfides had been motivated to make
effective glycosyl donors and a new class of test compounds
7
available. In this respect, our screening of disulfide-forming
thioglycoside libraries with the toxin from mistletoe had indeed
indicated a potential of dithiodigalactoside (DTDG) to protect
4d
a human cell line from toxin binding. Also, the asymmet-
ric disulfide of N-acetylated galactosamine and glucosamine,
a derivative of di-N-acetylated lactose, appeared to harbour
4d
reactivity for galectin-3. These initial results on bioactivity of
a diglycosyldisulfide prompted us to systematically investigate the
reactivity profile of DTDG relative to TDG and lactose. In detail,
we addressed the following issues:
a.) to determine its binding capacity to the mistletoe toxin in
two assay systems of increasing biorelevance,
b.) to provide a structural model for binding, and
c.) to determine the relative reactivity to human galectins from
each of the three groups listed above, to build a structural model
for a representative complex and to run assays with the two
separate domains of a tandem-repeat-type family member, i.e.
galectin-4.
(
Table 1). Thus, the solid-phase inhibition assay was suitable to
determine reactivity and to provide a comparative measure. Tested
under these conditions, DTDG obviously had inhibitory capacity
(
Fig. 1). It was about 2–3-fold lower than for TDG (Table 1).
Technically, ligand properties of the synthetic product were
assessed by spectrophotometric and radioactivity-based inhibition
assays, epitope mapping and competitive binding analysis in sat-
uration transfer difference (STD) NMR spectroscopy, isothermal
titration calorimetry (ITC) and fluorescent cell assays, flanked by
computational docking to visualize a model for the formation of
lectin-glycosyldisulfide complexes.
Results and discussion
Synthesis
The symmetrical disulfide was obtained in a three-step procedure
using a solution of iodine in ethanol in the last step, as outlined in
Scheme 1. This compound was used to probe for its ligand proper-
ties on the plant toxin in biochemical, ITC, NMR spectroscopical
and cell biological assays.
Fig. 1 Titration curves illustrating the course of inhibition of binding
of biotinylated VAA (3 mg ml ) to lactosylated neoglycoprotein in a
-1
solid-phase assay by increasing concentrations of thiodigalactoside (ꢀ),
dithiodigalactoside (ꢁ) and lactose (᭺).
◦
Scheme 1 Reagents and conditions: a) HSAc, BF
446 | Org. Biomol. Chem., 2011, 9, 5445–5455
3
·Et
2 2 2 2 2
O, CH Cl , 0 C - rt, 24 h (89%); b) NaOMe, MeOH, 2 h (97%); c) H O, I -EtOH.
5
This journal is © The Royal Society of Chemistry 2011

View Article Online
Table 1 Relative inhibitory potency of lactose (Lac), thiodigalactoside
TDG) and dithiodigalactoside (DTDG) on lectin binding to a surface-
(
a
immoblized neoglycoprotein (expressed as IC50-value, in mM)
Lectin
Lac
TDG
DTDG
-
1
VAA (3 mg ml )
galectin-1 (20 mg ml )
0.6
0.8
1.6
3.5
4.0
1.2
0.4
0.5
1.1
1.8
2.3
1.4
1.1
4.8
5.4
> 10
> 10
> 10
-
1
-1
galectin-3 (15 mg ml )
-1
galectin-4 (5 mg ml )
-1
galectin-7 (30 mg ml )
-1
galectin-8 (0.1 mg ml )
a
Titrations were performed using a constant coating concentration of
neoglycoprotein (0.25 mg/well) with eight concentrations of sugar in
triplicates and up to six independent series, reaching an upper limit of
1
3.7% for the standard deviation (for exemplary titration curve, please see
Fig. 1).
Fig. 2 Calorimetric titration of VAA (A) and human galectin-4 (B) with
thiodigalactoside (ꢀ) and dithiodigalactoside (ꢁ). Representative plots of
the heat released per mole of ligand injected in the course of titrations as a
function of the sugar/protein molar ratio. Solutions of VAA at 337.5 mM
(ꢀ)/243 mM (ꢁ) (in terms of AB heterodimer concentration) and of
galectin-4 at 117 mM (ꢀ)/154.7 mM (ꢁ) were used in these titrations.
Solid lines correspond to the best fit of the experimental data using a
one-set-of-sites model.
These results intimated a direct competition between the natural
b-galactoside presented by the neoglycoprotein and the tested
thiocompounds. To independently verify this assumption we next
ran direct competition assays in solution, monitored by STD
NMR spectroscopy.
In this setting, lactose was the reporter ligand, whose contact
profile, which underlies magnetization transfer from the saturated
protein to distinct ligand protons, is known. A reduction of
mate sugar moiety of disaccharides is known to lose its mobility
10
8
to varying degrees. As exemplarily shown by the strong signals
in the case of DTDG (Fig. 3 left), ligand protons at positions 2, 3
and 4 in the galactose pyranose ring appear to be in close vicinity
to the protein in all three sugars, irrespective of the nature of the
glycosidic bond. This experimental observation agrees with the
the intensity of these signals of lactose in the presence of
an inhibitor reflects the loss of lectin-lactose contact and the
relative inhibitory capacity of the tested compound. The respective
titrations confirmed the inhibitory activity of DTDG and yielded a
relative potency of about 0.5–0.7fold for DTDG when compared
to TDG. Since these experiments further documented complex
formation with the disulfide, we proceeded to perform isothermal
titration calorimetry for direct affinity measurement, using TDG
as positive control.
6i
conclusions derived from chemical-mapping studies, and can be
readily reconciled with the model of the complex obtained by in
silico docking (Fig. 4). The predicted binding patterns for TDG
and DTDG were nearly indistinguishable. Main interactions at the
Tyr-sites for both thiocompounds are hydrogen bonds involving
hydroxyl groups at positions O3 (with Asp235 and Asn256) and O4
Representative ITC curves for both compounds are shown
in Fig. 2A. Titration data obtained with both DTDG and
TDG could be fitted to a one-set-of-sites model with association
(with Gln238) as well as CH-p interactions of pyranose hydrogens
H3, H4 and H5 with Tyr249. The second galactose unit can
-
1
-1
constants, K , of 1560 ± 20 M (DG = -18.16 ± 0.03 kJ mol )
1
a
-
-1
and 2600 ± 100 M (DG = -19.4 ± 0.1 kJ mol ), respectively.
A reliable determination of the binding stoichiometry for this
range of affinities would require running the titrations at protein
concentrations over the natural solubility level of VAA. Therefore,
only total estimates of enthalpy/entropy changes per B chain,
-
1
-1
-1
-
1
50.3 ± 0.4 kJ mol /-108 ± 1 J mol
K
(DTDG) and -53 ±
(TDG), were derived from the
-
1
-1
-1
kJ mol /-110 ± 30 J mol
K
9
analysis. ITC data obtained for lactose under similar conditions
were consistent with two sets of sites with K
3
-1
a
of 1.1 ¥ 10 M
4
-1
and 1.6 ¥ 10 M , assigned to 2g (Tyr-sites) and 1a (Trp-sites)
subdomains of the B-chains, respectively, the accessibility to the
Trp-sites being restricted by tetramer formation. The fact that
data obtained for DTDG and TDG could be fitted to a one-set-
of-sites model imply that if both Tyr- and Trp-sites are operative
for binding the thioderivatives, they must exhibit similar affinities.
At any rate, DTDG binds to VAA with lower affinity than TDG.
Having herewith determined the affinity of DTDG to the
toxin, STD NMR spectroscopy was used for epitope mapping
by recording proton signals of the ligand when separately running
experiments with lactose, TDG and DTDG. The primary contact
of disaccharides with VAA is the galactose unit, and the penulti-
Fig. 3 Illustration of relative STD signal intensities of protons for the
galactose moiety of dithiodigalactoside binding to the Tyr site of VAA (left)
and to human galectin-1 (right). The 500 MHz off-resonance (above) and
the STD (below) spectra are presented in the lower panel. The STD signals
(
even amplified 128 fold) for the hgal-1 complex are very weak indicating
that the binding affinity is very low. This feature strongly contrasts with the
observation for the VAA complex (amplified 8 fold). The asterisk denotes
the presence of a small amount of ethanol in the sample, which does not
give STD signal.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5447

View Article Online
Fig. 4 Structural model of the complex of VAA (Tyr site, surface rep-
resentation) with thiodigalactoside (green) and dithiodigalactoside (grey),
based on docking analysis and the experimental input from STD NMR
spectroscopy. While one galactose unit makes several contacts (hydrogen
bonds with Asp235 and Asn256, and CH-p interactions with Tyr249 (with
its van der Waals surface emphasized by the dots representation)), the
second moiety contributes weakly to complex stabilization with Gln238 in
the case of TDG. The van der Waals surfaces are emphasized with the grid
and the dots.
contact Glu238 as part of TDG but not of DTDG (Fig. 4). After
having herewith characterized the ligand properties of DTDG
and presented a structural model of the complex, it was an open
question whether the disulfide can protect human cells from toxin
binding, hereby increasing the biorelevance of testing.
In analogy to the solid-phase test system, concentration de-
pendence and inhibition were first ascertained for VAA binding
to cells, here B-lymphoblastic Croco II cells (Fig. 5A, B). The
comparative study of the inhibitory potency of galactose and
lactose (Fig. 5C) and TDG/DTDG (Fig. 5D) under identical
conditions revealed a graded potency, starting with TDG as
frontrunner, and underlined the activity of DTDG. Since the
glycomic profile and spatial parameters of glycan presentation can
differ among various cell lines, we ran the respective assays with
a line of another histogenetic origin, i.e. colon adenocarcinoma
cells, and obtained a similar grading (Fig. 5E, F). In aggregate,
these results are in accord with the affinity determinations and
prove that DTDG can impair carbohydrate-dependent toxin
binding to two types of human cells. Since this lectin’s b-trefoil
folding is different from the architecture of the b-sandwich-type
carbohydrate recognition domain of galectins, which also have
extended sites for ligand contacts beyond the galactose core, as
inferred by binding assays, NMR spectroscopical monitoring and
Fig. 5 Semilogarithmic representation of fluorescent surface staining
subsequent to VAA binding of cells of the human B-lymphoblastoid
-1
line Croco II (A–D; 0.5 mg ml VAA in B–D) and of the human colon
-1
adenocarcinoma line SW480 (E, F: 2 mg ml VAA). The control value of
cell positivity by the second-step reagent in the absence of lectin is given
as grey-shaded area, the 100%-value (lectin staining in the absence of
inhibitor) as thick black line. Numbers characterizing staining (percentage
of positive cells/mean fluorescence intensity) in each panel are always given
in the order of listing (from top to bottom). A: dependence of staining from
-
1
-1
increasing the concentration of VAA from 0.2 mg ml to 0.5 mg ml , 2 mg
6,11
-1
-1
molecular modelling, it is an open question whether DTDG
is a galectin ligand. Moreover, intrafamily sequence divergence
around the carbohydrate recognition site in galectin genes may
lead to differential reactivity. Thus, we tested galectins from each
of the three structural subgroups (proto-type with galectins-1
and -7, chimera-type with galectin-3 and tandem-repeat-type with
galectins-4 and -8) to address this additional issue.
ml and 4 mg ml . B: inhibition of VAA-dependent staining by increasing
the concentration of lactose from 1 mM to 2 mM, 5 mM and 10 mM
(
1
(
numbers correspond to control value, presence of 10 mM lactose and the
00%-value). C, D: staining parameters in the presence of 5 mM inhibitor
C: lactose, galactose; D: TDG, DTDG). E, F: staining parameters of the
colon cancer cells in the presence of 5 mM inhibitor (E: lactose, DTDG;
F: TDG, galactose).
Assaying ligand properties on human galectins
6
the known reactivity of galectins-1, -3 and -7 and extended the
The conditions of the solid-phase assay were first optimized for
each galectin. The obtained results with TDG were in line with
reactivity profile to the tandem-repeat-type galectins-4 and -8
(Table 1). When introducing DTDG to the assays, non-uniform
5
448 | Org. Biomol. Chem., 2011, 9, 5445–5455
This journal is © The Royal Society of Chemistry 2011

View Article Online
responses at comparatively high concentrations were registered
Table 1). Galectins-1 and -3 were sensitive to DTDG presence,
(
albeit weakly in relation to the plant toxin, whereas the other
galectins reacted poorly with the disulfide (Table 1). The use of
tumor cell extracts as source for glycoproteins instead of the
neoglycoprotein did not alter this result, excluding any ligand-
type-related bias. For example, galectin-3 binding was reduced by
the three tested compounds with IC50-values of 0.8 mM (TDG),
1
.2 mM (lactose) and 4.4 mM (DTDG), as similarly seen with the
neoglycoprotein (Table 1).
These results were corroborated by STD NMR spectroscopical
series, which revealed a rather weak or no activity for DTDG
for the galectins. Guiding model building to explain the negative
impact of the disulfide linkage on galectin binding, the STD-
based epitope mapping for TDG showed intense signals for the
H4, H5 and H6 protons, the characteristic signature for galectin
contact (Fig. 3 right). These data, together with the crystal
structures of the galectin of toad ovary and human galectin-1
12
in complex with TDG, provided the input and quality control
for molecular modeling studies. Similar to lactose, TDG thus
engages both moieties of the disaccharide for binding to galectin-
1
. This pattern changes due to the presence of the disulfide in
the glycosidic linkage. One sugar residue still enters the binding
site and maintains the contact pattern, whereas the second would
be completely exposed to the solvent, with arrest of the angle of
◦
the S–S bond at about 90 (Fig. 6A). The structural constellation
remained rather stable over the entire 6-ns period of the MD
simulation. Of note, the analysis of the MD trajectories, after
superimposition of the two galectin-1/ligand complexes, enabled
identification of a water molecule, which could take the place of the
OH2 group of TDG in mediating the hydrogen-bonding network
with Arg48/Glu71 shown in Fig. 6B. Overall, the disclosed lack of
contact building beyond the interacting galactose of the symmetric
diglycosyldisulfide will then, apparently, set strict limits to the
affinity. Evidently, the disulfide bonding, itself being subject to a
conformational arrest, impairs the potential for interacting with
galectins.
This impairment was also seen on the level of cell binding, where
galectins interact with complex glycans of natural glycoconjugates,
as documented by cytofluorimetric analysis with galectin-3 and
colon adenocarcinoma cells (Fig. 7A, B) as well as for galectin-4
and pancreatic adenocarcinoma cells (Fig. 7C–F). In the latter
case, the apparent strongly negative impact on inhibitory potency
was further confirmed by ITC (Fig. 2B). Addition of DTDG
to the galectin-4-containing solution triggered only negligible
heat generation, comparable to the signal of ligand dilution
Fig. 6 Illustration of the models of the average structures of the complexes
of human galectin-1 with dithiodigalactoside (A) based on MD simu-
lations showing the key intermolecular contacts. Different perspectives
in the full and expanded views are offered for the sake of clarity. (B)
Superimposition of the complexes with TDG (thiodiglycoside in white)
and DTDG (diglycosyldisulfide in magenta), also revealing the presence of
the water molecule that, in the second complex, substitutes the OH2 of the
second galactose moiety of TDG to mediate a hydrogen-bonding network
with Arg48 and Glu71. Intermolecular hydrogen bonds are highlighted
with lines.
(
Fig. 2B). The titration profiles also highlight the conspicuous
reactivity difference to the toxin shown in panel A. Titrations
with TDG and galectin-4 as positive control yieled an association
presenting agarose beads and a surface-presented glycoprotein
(asialofibrin films) with iodinated protein using the two galectin-4
domains separately. These experiments will also inform us about
the binding properties of the two, structurally distinct domains
of galectin-4. Extent of binding of the labelled lectin domains
was measured in presence of increasing concentrations of lactose,
TDG and DTDG. Apparent association constants were calculated
from the plot of the reciprocal of the lectin fraction bound to the
-
1
-1
constant of 1210 ± 30 M (DG = -17.53 ± 0.07 kJ mol ) and
-
1
estimated enthalpy/entropy values of -39.9 ± 0.8 kJ mol /-75 ±
1
-1
3
J mol-
K
based on a one-set-of-sites model (assuming two
equivalent and independent sites per molecule), and no evidence
for cooperativity. Lactose binds similarly with a comparatively
-
1
-1
lower affinity (K
a
= 585 ± 8 M ; DG = -15.74 ± 0.03 kJ mol ) due
1
-
to an increased entropic penalty (DH = -43 ± 2 kJ mol , DS =
93 ± 7 J mol K ) (not shown).
-
1
-1
13
-
ligand vs. the inhibitor concentration. A representative plot of
Finally, to track down ligand properties of DTDG at a very
sensitive level, we performed radiobinding assays to lactose-
data for the C-terminal domain is given in Fig. 8. In relation to
the ITC data for the full-length form of galectin-4, the association
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5449

View Article Online
Fig. 7 Semilogarithmic representation of fluorescent surface staining upon binding of galectin-3 (A, B) and galectin-4 (C–F) of cells of the human colon
INK4a
adenocarcinoma line SW480 (A, B) and the pancreatic adenocarcinoma line Capan-1 with functional tumor suppressor p16
(C–F). For details on
-1
presentation, please see legend to Fig. 5. A, B: staining parameters in the presence of 2 mM inhibitor for galectin-3 at 10 mg ml (A: lactose, galactose;
-1
B: TDG, DTDG). C, D: staining parameters in the presence of 2 mM inhibitors for galectin-4 at 20 mg ml (C: lactose, galactose; D: TDG, DTDG). E,
F: staining parameters in the presence of 20 mM DTDG (E) and in the presence of decreasing concentrations of TDG (F: 20 mM, 10 mM, 5 mM).
constants for lactose were in the same range with values of 300 ±
4
DTDG reactivity. By going up to 10 mM DTDG, lectin binding
to the matrix-presented ligand was reduced by about 25%, from
which an upper limit of the association constant at <30 M
could be estimated for both domains. Thus, DTDG is definitely
only a very weak ligand for galectin-4, on the level of both
domains.
0 M- and 440 ± 60 M for the N- and C-terminal domains,
1
-1
-
1
respectively. Thus, the fit for the one-set-of-sites model of ITC
data is attributable to this similarity. The affinity increase for
-
1
TDG was also determined, with K -values of 730 ± 90 M and
a
-
1
6
70 ± 50 M , respectively. This set-up enabled to trace even minor
5
450 | Org. Biomol. Chem., 2011, 9, 5445–5455
This journal is © The Royal Society of Chemistry 2011

View Article Online
1
13
(
0.040–0.063 mm). H-NMR and C-NMR spectra were recorded
on a Bruker Avance 400 spectrometer at 400 (100) MHz and/or
Bruker Avance DMX 500 at 500 (125) MHz, respectively. Com-
pound 1 was purchased from Sigma.
2
,3,4,6-Tetra-O-acetyl-1-S-acetyl-1-thio-b-D-galactopyranose
4b,19a
19b
(adapted from
) (2)
1
5
,2,3,4,6-Penta-O-acetyl-b-D-galactopyranose
.13 mmol; please see Scheme 1) and thioacetic acid (1.52
(1)
(2
g,
◦
ml, 21.51 mmol) were dissolved in dry CH
and BF
a period of 15 min. The reaction mixture was allowed to warm
to ambient temperature and stirred under nitrogen protection
for 24 h. The reaction mixture was then poured onto ice-cold
2
Cl
2
(25 ml) at 0 C,
3
·Et
2
O (3.889 ml, 31.51 mmol) was added dropwise over
water, and extracted with CH
were successively washed with NaHCO
over MgSO , and filtered. The crude product was purified by
2
Cl
2
. The combined organic layers
3
, water and brine, dried
Fig. 8 Inhibitory potency of lactose (᭺), thiodigalactoside (ꢀ) and
dithiodigalactoside (ꢁ) on the binding of the C-terminal domain of
galectin-4 in a solid-phase radioassay. The extent of binding of the
I-labelled domain to asialofibrin films (100 mg) presented on the plastic
microwell surface was measured in the absence and presence of different
4
column chromatography (eluent: hexane/EtOAc 3 : 1) to give
2,3,4,6-tetra-O-acetyl-1-S-acetyl-1-thio-b-D-galactopyranose
125
19b
◦
1
(2) (1.85 g, 4.56 mmol, yield 89%). mp 117 C. H NMR
(
(
1
CDCl
t, 1H, J = 10.1 Hz, H-2), 5.25 (d, 1H, J = 10.2 Hz, H-1), 5.11 (dd,
H, J = 3.4, 9.8 Hz, H-3), 4.03–4.16 (m, 3H, H-5, H-6, H-6¢), 2.39
s, 3H, SCOCH3), 2.08, 2.03, 2.02, 2.01 (12H, 4 s, 4 OCOCH3);
3
, 500 MHz): d (ppm) 5.45 (d, 1H, J = 3.4 Hz, H-4), 5.32
concentrations of the tested sugar compounds.
Conclusions
(
1
3
C NMR (CDCl
3
, 125 MHz): d (ppm) 192.1, 170.3, 170.1, 169.9,
Work with dynamic combinatorial thioglycoside libraries had
indicated potential for DTDG to react with an AB-type plant
1
9.5, 80.6, 75.0, 71.9, 67.2, 66.3, 61.2, 30.8, 20.7, 20.6, 20.6, 20.5.
+
4d
MS (ESI): m/z = 429.06 [M + Na ], C16
22
H O10S; found 429.00.
toxin. The results reported here define DTDG as ligand in
different biochemical assays, characterize its respective properties
in structural terms and document its capacity to protect human
cells from toxin binding. The same panel of strategically combined
experimental and computational methods delineated a significant
difference in reactivity to human galectins, notably with intrafam-
ily grading. Glycosyldisulfides are thus a new substance platform
to accomplish selective inhibition of this class of endogenous
lectins and a plant AB-type toxin. The computation of a structural
model revealed a lack of contact of the second galactose moiety in
the case of the disulfide. Exploration of additional substitutions
19b
1
-Thio-b-D-galactopyranose sodium salt (3)
Compound 2 (1.85 g, 4.56 mmol, yield 89%) was dissolved in a
MeOH solution (15 ml) of NaOMe (271 mg, 5.02 mmol). The
chemical reaction in the resulting mixture quantitatively yielded
1
product 3. H-NMR (500 MHz, D
2
O): d = 4.49 (d, J = 9.1 Hz, 1H,
H-1), 3.92 (d, J = 3.6 Hz, 1H, H-4), 3.74 (dd, J = 11.2, 7.6 Hz, 1H,
H-6), 3.61–3.69 (m, 2H, H-5, H-6¢), 3.58 (dd, J = 9.5, 3.6 Hz, 1H,
13
H-3), 3.31 (t, J = 9.2 Hz, 1H, H-2). C NMR (D
2
O, 125 MHz): d
(ppm) 85.8, 80.1, 77.4, 75.1, 71.0, 62.7.
14
at the 2¢- and 3¢-positions, e.g. 2¢-fucosylation, the introduction
2g,8,15
of anomeric and aglyconic extensions
and the conjugation
19c
1,1¢-Dithio-b-D-digalactopyranoside (4)
to certain scaffolds, e.g. starburst glycodendrimers, cyclodextrin,
16
glycocyclophanes or cyclic neoglycopeptides, are routes to
further increase the level of affinity and selectivity. This also
applies to capitalize on intrafamily differences among galectins,
when aiming to block a certain protein in a functional context
such as tumor progression or intergalectin competition during
Compound 3 (0.2 g, 0.92 mmol) was dissolved in water (3 ml),
and a saturated ethanolic solution of iodine was added dropwise
until a persistent yellow color appeared in the mixture. The solvent
was evaporated, the crude product was dissolved in MeOH and
precipitated by the addition of an excess of EtOAc. Filtration of
17
19c
growth regulation. Having herewith revealed the bioactivity of
the symmetric dithiodigalactoside, it is an emerging question
whether the respective selenides, whose conformational flexibility
the white solid yielded pure compound 4 in good yield (0.16 g,
0.41 mmol, yield 89%). The procedure was repeated twice to
1
remove remaining traces of compound 3. H NMR (D O, 500
2
18
can resemble that of thioglycosides, will be lectin ligands.
MHz): d (ppm) 4.54 (d, 2H, J = 9.7 Hz, H-1, H¢-1), 3.97 (d, 2H,
J = 3.26 Hz, H-4, H¢-4), 3.83 (t, 2H, J = 9.7 Hz, H-2, H¢-2), 3.72–
3
.80 (m, 4H, H-5, H¢-5, H-6, H¢-6), 3.72 (dd, 2H, J = 3.1, 12.3 Hz,
Experimental
13
H-6¢, H¢-6¢), 3.69 (dd, 2H, J = 3.3, 9.5 Hz, H-3, H¢-3); C NMR
(D O, 125 MHz): d (ppm) 89.9, 79.6, 73.9, 68.8, 68.7, 61.1.
Materials and general methods
2
All commercial reagents and solvents were used as received.
Chemical reactions were monitored with thin-layer chromatog-
raphy using precoated silica gel 60 (0.25 mm thickness) plates.
Flash column chromatography was performed on silica gel 60
Lectin purification and quality controls
Purification of the lectins started from extracts of dried mistletoe
leaves (VAA) and bacterial pellets after recombinant production
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5451

View Article Online
(
human full-length galectins and the two separate galectin-4
the momomer of galectin-4 and of the covalently linked AB
heterodimer of VAA was used as input in the fitting procedures.
domains) and included affinity chromatography on lactosylated
Sepharose 4B, obtained after divinyl sulfone activation, as central
20
step. Purity was routinely ensured by one- and two-dimensional
gel electrophoresis and mass spectrometry, quaternary structure by
gel filtration and activity by i) haemagglutination using trypsin-
treated and glutaraldehyde-fixed rabbit erythrocytes, ii) toxicity
tests on human tumor lines (Croco II, SW480) and iii) anoikis
assays on a responsive cell line (Capan-1). Biotinylation of
the lectins with the N-hydroxysuccinimide ester derivative of
biotin (Sigma, Munich, Germany) was performed under activity-
preserving conditions and was followed by mass spectrometric
product analysis to quantify label incorporation and identify sites
STD NMR experiments
All experiments were performed at a lectin concentration of 50 mM
on a Bruker Avance DRX 500 MHz NMR spectrometer equipped
with a 5 mm TXI inverse probe head achieving protein saturation
by 40 selective 70 dB Gaussian pulses of 50 ms duration and
21
8,25
a total number of 360 (VAA) or 720 (galectins) scans.
The
ligand to protein ratio was set to 80 : 1, competition experiments
between lactose and the derivatives were performed up to reaching
equimolar concentrations. All details on epitope mapping and
competition can be obtained upon request from the authors.
22
of conjugation.
Quantum mechanical methods, partial charges, and molecular
mechanics force field
Inhibition spectrophotometric assay
Neoglycoprotein (bovine serum albumin free of contaminating
Starting from the geometry of the crystallographic structure of lac-
tose in the PDB file 1GZW, and of TDG in the PDB file 1A78,
glycoproteins after conjugation of the p-isothiocyanatophenyl
26
12a
23
derivative of lactose ) was adsorbed to the surface of microtiter
we constructed structures for TDG and DTDG, which were
subjected to a full geometry optimization by quantum mechanical
plate wells (0.25 mg/well in 50 ml) from phosphate-buffered saline
◦
at 4 C overnight, and residual sites on the plastic surface for
27
calculations based on the Gaussian03 program using the B3LYP
◦
binding proteins were then saturated for 1 h at 37 C with albumin
method and a 6-31G(d) basis set. The resulting wavefunctions
were used to calculate electrostatic potential-derived (ESP) charges
(
100 ml of a 1% solution (w/v); Biomol, Hamburg, Germany).
Protein extracts from SW480 colon adenocarcinoma cells were
prepared by sonication of cells suspended in 50 mM Tris/HCl
buffer (pH 8.0) containing 150 mM NaCl and 1% (v/v) Nonidet
P-40. Aliquots of the solution with (glyco)proteins (supernatant
after centrifugation; 0.5 mg/well) were kept in microtiter plate
wells overnight for coating as described above. Incubation with
lectin-containing solution in the absence and presence of tested
28
employing the RESP methodology, as implemented in the
Amber 9 suite of programs (http://amber.scripps.edu/). For
the calculation of dihedral parameters for the torsional barriers
around the bonds involving the S atom in TDG and DTDG we
used the methyl derivatives Gal-S-Me and Gal-S-S-Me. These
values were calculated so as to reproduce, in the AMBER force
field, the energy values calculated ab initio in Gaussian03 (keyword
◦
compounds was performed for 1 h at 37 C and the extent of lectin
◦
SCAN) upon rotation of the bond in 20 steps. The remaining
binding was spectrophotometrically determined after colorimetric
detection of biotin with streptavidin-peroxidase conjugate (Sigma)
bonded and nonbonded parameters were assigned, by analogy or
through interpolation from those already present in the AMBER
database, in a way consistent with the Glycam-04 force field
for carbohydrates (http://glycam.ccrc.uga.edu), which has been
tested widely and used successfully in simulations of the type
described herein.
23
as described. Assays were routinely run in triplicates in up to six
independent series with eight concentrations of sugar, standard
deviations not exceeding 13.7%.
Isothermal titration calorimetry
Docking analysis
◦
The calorimetric titrations were performed at 25 C with a Micro-
6
i
cal MCS-ITC microcalorimeter (Microcal LLC., Northampton,
MA, USA). VAA samples were exhaustively dialyzed against
Docking of TDG and DTDG into the binding sites of VAA and
26
29
human galectin-1 was based on AutoDock 4.2 application. We
first validated AutoDock by testing its ability to predict the lactose-
binding mode seen in the crystal structure of human galectin-1
(1GZW). Crystallographic water molecules that were identified as
important for binding were kept within the docking calculations.
From a preceding conformational analysis, different conformers
were selected as starting geometries to be docked. Evaluation of
the finally docked structures indicated that AutoDock was able to
predict the bound-state conformer. The same protocol was applied
to TDG and DTDG: the Lamarckian genetic algorithm (LGA)
implemented in AutoDock was used, by randomly changing the
torsion angles and overall orientation of the molecule (100 runs
with 2 500 000 evaluations and a population size of 150 were
selected). The mutation and crossover rates were set to 0.80
and 0.02, respectively, and the maximal number of generations
was 27 000; elitism was set to 1 and the local search frequency
to 0.06. Search parameters were chosen from those previously
5
mM sodium phosphate buffer, pH 7.2, containing 0.2 M NaCl
(
PBS), whereas PBS containing 4 mM b-mercaptoethanol was
used for equilibration in the case of human galectin-4. For
calorimetric titrations of galectin-4 with DTDG, the reducing
agent was excluded from the last dialysis buffer in order to
avoid reduction of the glycosyldisulfide. Sugar solutions were
prepared using the last dialysis buffer. Titrations were performed
by stepwise injections of ligand-containing solution into the
reaction cell loaded with the protein solution at concentrations of
117–337.5 mM, depending on the lectin. VAA concentrations were
routinely determined using a colorimetric assay ascertained to be
applicable for AB toxins, theoretical molar adsorption coefficients
24
were used for galectin-4. The heat developed by ligand dilution
was determined separately and subtracted from the total heat
produced following each injection. Titration data were analyzed
using the Microcal-ITC Origin software. The concentration of
5
452 | Org. Biomol. Chem., 2011, 9, 5445–5455
This journal is © The Royal Society of Chemistry 2011

View Article Online
established for docking of carbohydrates, and the grid box was
restricted around galectin-1’s contact site for lactose. From careful
analysis of the docking results, the binding pose with the lowest
docking energy was selected as the starting point for performing
MD simulations in explicit water. All details on calculations are
available upon request.
aliquots of cell suspensions of the same passage were run in
triplicates, with standard deviation not exceeding 15.5%.
Inhibition radioassays
The two separate carbohydrate recognition domains of human
galectin-4 were radioiodinated using IODO-GEN (Pierce Euro-
13
chemie), as described previously. Specific activities of 0.57 and
Molecular dynamics simulations
0
.35 mCi/mg were obtained for the N- and C-terminal domains,
1
25
For each of the two free ligands 6-ns MD simulations were carried
out employing the calculated parameters, using Amber 9. Each
respectively. These
I-labelled proteins were first tested for
maintained binding capacity and then for optimal test conditions.
For the N-domain, the amount of labelled protein bound to
agarose beads presenting lactose (Sigma) was measured. Aliquots
˚
ligand was placed in a 8 A-deep truncated octahedral box of
explicit TIP3P water molecules. The equilibration phase started
with energy minimisation of the solvent, followed by an energy
minimisation of the entire system without restraints. The system
was then heated up from 10 K to 298 K during 100 ps with weak
restraints on the solute (positional restraints to a-carbon atoms),
followed by a 100 ps MD simulation at constant temperature and
pressure (1 atm), without restrictions. Equilibrated structures were
used as starting points for 6-ns production trajectories, performed
at constant pressure (1 atm) and temperature (288 K), as in the
NMR experiments, controlled by the Langevin thermostat with a
1
25
of 10 ml of sedimented resin were incubated with the I-protein-
containing solution (15000 cpm) in the absence or presence
of different concentrations of test compound (final volume:
50 ml). The buffer used was PBS containing 0.1% Tween-20 and
0.4 mM b-mercaptoethanol. The inhibitory capacity of DTDG
was evaluated in parallel experiments carried out in the absence of
◦
b-mercaptoethanol. After 2 h at 20 C with regular shaking, the
samples were centrifuged and the radioactivity of the supernatant
counted. The C-terminal domain was assayed using 100 mg-
asialofibrin films on the surface of plastic microwells essentially
-
1
collision frequency of 1.0 ps . During the simulations, SHAKE
algorithm was turned on and applied to all hydrogen atoms. A
13
as described, except that the plate surface was treated with a
◦
˚
2.5% solution of Tween-20 for 1 h at 37 C to preclude adsorption
cut-off of 10 A for all nonbonded interactions was adopted. An in-
tegration time step of 2 fs was employed, and periodic boundaries
conditions were applied throughout. During all simulations, the
of protein and the buffer used was PBS containing 0.1% Tween-
20 and 0.4 mM b-mercaptoethanol. Again, the reducing agent
was omitted when evaluating the ligand properties of DTDG.
For the two systems, i.e. lactose-agarose and asialofibrin films,
carbohydrate-independent binding was determined in the presence
of 0.1 M lactose and subtracted from all measurements.
30
Particle Mesh Ewald (PME) method was used to compute long-
range electrostatic interactions. Minimisation, equilibration and
production phases were carried out by using the sander module,
while analyses of the simulations were performed with ptraj mod-
ule (http://amber.scripps.edu/). The visualization of trajectories
was performed using VMD software (http://www.ks.uiuc.edu/).
Data were processed and plotted using R software (http://www.r-
project.org/). Concerning the conformational behavior of the
ligands during the 6-ns MD simulations, exo-anomeric regions
were sampled for both thioglycosidic torsions, while w angle
around the C5–C6 bond took the expected gg and gt orientations.
The set of charges and parameters described before was employed
for the MD simulations of complexes with galectin-1. Starting
geometries were selected from docking studies as the poses with
the lowest docking energy, in full agreement with the NMR data.
Acknowledgements
This work has been generously supported by grants CTQ2009-
08536, SAF2008-00945, BFU2006-10288 and BFU2009-10052
from the Spanish Ministry, an EC Marie Curie Research Training
Network grant (contract no. MRTN-CT-2005-019561), the EC
research program GlycoHIT (contract ID 260600), the Verein
zur F o¨ rderung des biologisch-technologischen Fortschritts in der
Medizin e. V. and the CIBER of Respiratory Diseases (CIBERES),
an initiative from the Spanish Institute of Health Carlos III
(ISCIII). We thank Drs B. Friday and S. Namirha for inspiring
discussions as well as M. V. L o´ pez-Moyano for excellent technical
assistance.
6
-ns MD simulations of the sugar/receptor complex were carried
out following a protocol identical to that described above for the
free ligands.
Inhibition cell assay
References
Cells of the B-lymphoblastic line Croco II, the colon adenocarci-
noma line SW480 and the pancreatic adenocarcinoma line Capan-
1
H.-J. Gabius, ed. The Sugar Code. Fundamentals of glycosciences, Wiley-
VCH, Weinheim, 2009.
(a) F. Schweizer and O. Hindsgaul, Curr. Opin. Chem. Biol., 1999, 3,
INK4a
1
reconstituted for tumor suppressor p16
status were cultured
2
and processed for cell staining in a two-step procedure using a
streptavidin/R-phycoerythrin conjugate (Sigma), as described in
detail in our previous study on dynamic combinatorial thiogly-
2
2
91–298; (b) P. M. St. Hilaire and M. Meldal, Angew. Chem., Int. Ed.,
000, 39, 1162–1179; (c) A. Barkley and P. Arya, Chem.–Eur. J., 2001, 7,
555–563; (d) O. Ramstr o¨ m and J.-M. Lehn, Nat. Rev. Drug Discovery,
2002, 1, 26–36; (e) S. Andr
Gabius and J. P. Kamerling, Bioorg. Med. Chem. Lett., 2007, 17, 793–
4d
e´ , C. E. P. Maljaars, K. M. Halkes, H.-J.
coside libraries to maintain identical conditions. Decrease of
fluorescent staining (as percentage of positive cells and intensity)
served as indicator of the inhibitory capacity of the tested
compounds. Up to five series of independent measurements on
7
1
98; (f) M. J. Borrok and L. L. Kiessling, J. Am. Chem. Soc., 2007, 129,
2780–12785; (g) S. Andr e´ , D. Gigu e` re, T. K. Dam, C. F. Brewer, H.-J.
Gabius and R. Roy, New J. Chem., 2010, 34, 2229–2240.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5453

View Article Online
3
4
(a) M. R. Hartley and J. M. Lord, Biochim. Biophys. Acta, 2004, 1701, 1–
4; (b) H.-J. Gabius, Crit. Rev. Immunol., 2006, 26, 43–79; (c) F. Stirpe
and M. G. Batelli, Cell. Mol. Life Sci., 2006, 63, 1850–1866; (d) A.
Villalobo, A. Nogales-Gonz a´ lez and H.-J. Gabius, Trends Glycosci.
Glycotechnol., 2006, 18, 1–37; (e) R. Schwartz-Albiez, in The Sugar
Code. Fundamentals of glycosciences, ed. H.-J. Gabius, Wiley-VCH,
Weinheim, 2009, pp. 447–467.
15 (a) R. T. Lee, H.-J. Gabius and Y. C. Lee, J. Biol. Chem., 1992, 267,
23722–23727; (b) R. T. Lee, H.-J. Gabius and Y. C. Lee, Carbohydr. Res.,
1994, 254, 269–276; (c) M. Kolympadi, M. Vontanella, C. Venturi, S.
Andr e´ , H.-J. Gabius, J. Jim e´ nez-Barbero and P. Vogel, Chem.–Eur. J.,
2009, 15, 2861–2873.
16 (a) S. Andr e´ , P. C. Ortega, M. A. Perez, R. Roy and H.-J. Gabius,
Glycobiology, 1999, 9, 1253–1261; (b) S. Andr e´ , H. Kaltner, T. Furuike,
S.-I. Nishimura and H.-J. Gabius, Bioconjugate Chem., 2004, 15, 87–98;
(c) R. Leyden, T. Velasco-Torrijos, S. Andr e´ , S. Gouin, H.-J. Gabius
and P. V. Murphy, J. Org. Chem., 2009, 74, 9010–9026; (d) S. Andr e´ , O.
Renaudet, I. Bossu, P. Dumy and H.-J. Gabius, J. Pept. Sci., 2011, 17,
427–437.
17 (a) J. Kopitz, C. von Reitzenstein, S. Andr e´ , H. Kaltner, J. Uhl, V.
Ehemann, M. Cantz and H.-J. Gabius, J. Biol. Chem., 2001, 276,
35917–35923; (b) S. Rorive, N. Belot, C. Decaestecker, F. Lefranc, L.
Gordower, S. Micik, C.-A. Maurage, H. Kaltner, M.-M. Ruchoux, A.
Danguy, H.-J. Gabius, I. Salmon, R. Kiss and I. Camby, Glia, 2001,
33, 241–255; (c) S. Langbein, J. Brade, J. K. Badawi, M. Hatzinger,
H. Kaltner, M. Lensch, K. Specht, S. Andr e´ , U. Brinck, P. Alken and
H.-J. Gabius, Histopathology, 2007, 51, 681–690; (d) O. Roda, E. Ort ´ı z-
Zapater, N. Mart ´ı nez-Bosch, R. Guti e´ rrez-Gallego, M. Villa-Perell o´ ,
C. Ampurdan e´ s, H.-J. Gabius, S. Andr e´ , D. Andreu, F. X. Real and
P. Navarro, Gastroenterology, 2009, 136, 1379–1390; (e) H. Sanchez-
Ruderisch, C. Fischer, K. M. Detjen, M. Welzel, A. Wimmel, J. C.
Manning, S. Andr e´ and H.-J. Gabius, FEBS J., 2010, 277, 3552–3563.
18 F. Strino, J.-H. Lii, C. A. K. Koppisetty, P.-G. Nyholm and H.-J. Gabius,
J. Comput.-Aided Mol. Des., 2010, 24, 1009–1021.
19 (a) K. Czifr a´ k and L. Soms a´ k, Carbohydr. Res., 2009, 344, 269–277;
(b) R. R. Schmidt and M. Stumpp, Liebigs Ann. Chem., 1983, 1249–
1256; (c) M. J. Kiefel, R. M. Thomson, M. Radovanovic and M. von
Itzstein, J. Carbohydr. Chem., 1999, 18, 937–959.
20 (a) H.-J. Gabius, H. Walzel, S. S. Joshi, J. Kruip, S. Kojima, V. Gerke,
H. Kratzin and S. Gabius, Anticancer Res., 1992, 12, 669–676; (b) A.
Beer, S. Andr e´ , H. Kaltner, M. Lensch, S. Franz, K. Sarter, C. Schulze,
U. S. Gaipl, P. Kern, M. Herrmann and H.-J. Gabius, Cytometry, Part
A, 2008, 73A, 139–147.
21 (a) H.-J. Gabius, R. Engelhardt, F. Cramer, R. B a¨ tge and G. A.
Nagel, Cancer Res., 1985, 45, 253–257; (b) H.-J. Gabius, F. Darro, M.
Remmelink, S. Andr e´ , J. Kopitz, A. Danguy, S. Gabius, I. Salmon and
R. Kiss, Cancer Invest., 2001, 19, 114–126; (c) S. Andr e´ , H. Sanchez-
Ruderisch, H. Nakagawa, M. Buchholz, J. Kopitz, P. Forberich, W.
Kemmner, C. B o¨ ck, K. Deguchi, K. M. Detjen, B. Wiedenmann, M.
von Knebel Doeberitz, T. M. Gress, S.-I. Nishimura, S. Rosewicz and
H.-J. Gabius, FEBS J., 2007, 274, 3233–3256; (d) D. Sol ´ı s, M. J. Mat e´ ,
M. Lohr, J. P. Ribeiro, L. L o´ pez-Merino, S. Andr e´ , E. Buzamet, F. J.
Ca n˜ ada, H. Kaltner, M. Lensch, F. M. Ruiz, G. Haroske, U. Wollina,
M. Kloor, J. Kopitz, J. L. S a´ iz, M. Men e´ ndez, J. Jim e´ nez-Barbero, A.
Romero and H.-J. Gabius, Int. J. Biochem. Cell Biol., 2010, 42, 1019–
1029.
22 (a) H.-J. Gabius, B. Wosgien, M. Hendrys and A. Bardosi, Histochem-
istry, 1991, 95, 269–277; (b) D. K u¨ bler, C.-W. Hung, T. K. Dam, J.
Kopitz, S. Andr e´ , H. Kaltner, M. Lohr, J. C. Manning, L. He, H.
Wang, A. Middelberg, C. F. Brewer, J. Reed, W.-D. Lehmann and H.-J.
Gabius, Biochim. Biophys. Acta, 2008, 1780, 716–722.
23 (a) H.-J. Gabius, S. Bodanowitz and A. Schauer, Cancer, 1988, 61,
1125–1131; (b) S. Andr e´ , C. Unverzagt, S. Kojima, M. Frank, J. Seifert,
C. Fink, K. Kayser, C.-W. von der Lieth and H.-J. Gabius, Eur. J.
Biochem., 2004, 217, 118–134.
24 M. Jim e´ nez, S. Andr e´ , H.-J. Gabius and D. Sol ´ı s, Anal. Biochem., 2000,
284, 418–420.
25 (a) J. Jim e´ nez-Barbero, E. Dragoni, C. Venturi, F. Nannucci, A. Ard a´ ,
M. Fontanella, S. Andr e´ , F. J. Ca n˜ ada, H.-J. Gabius and C. Nativi,
Chem.–Eur. J., 2009, 15, 10423–10431; (b) F. J. Mu n˜ oz, J. I. Santos, A.
Ard a´ , S. Andr e´ , H.-J. Gabius, J. V. Sinisterra, J. Jim e´ nez-Barbero and
M. J. Hern a´ iz, Org. Biomol. Chem., 2010, 8, 2986–2992.
26 M. F. L o´ pez-Lucendo, D. Sol ´ı s, S. Andr e´ , J. Hirabayashi, K.-i. Kasai,
H. Kaltner, H.-J. Gabius and A. Romero, J. Mol. Biol., 2004, 343,
957–970.
27 Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schlegel,
G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgomery, Jr.,
T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S. Iyengar, J.
Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani, N. Rega,
G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K. Toyota,
R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O.
Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian, J.
1
(a) O. Ramstr o¨ m and J.-M. Lehn, ChemBioChem, 2000, 1, 41–48;
(
b) Z. Pei, R. Larsson, T. Aastrup, H. Anderson, J.-M. Lehn and O.
Ramstr o¨ m, Biosens. Bioelectron., 2006, 22, 42–48; (c) T. Hotchkiss,
H. B. Kramer, K. J. Doores, D. P. Gamblin, N. J. Oldham and B. G.
Davis, Chem. Commun., 2005, 4264–4266; (d) S. Andr e´ , Z. Pei, H.-C.
Siebert, O. Ramstr o¨ m and H.-J. Gabius, Bioorg. Med. Chem., 2006,
1
4, 6314–6326; (e) R. Caraballo, M. Sakulsombat and O. Ramstr o¨ m,
Chem. Commun., 2010, 46, 8469–8471.
5
6
(a) B. Aguilera, J. Jim e´ nez-Barbero and A. Fernandez-Mayoralas,
Carbohydr. Res., 1998, 308, 19–27; (b) F. Strino, J.-H. Lii, H.-J.
Gabius and P. G. Nyholm, J. Comput.-Aided Mol. Des., 2009, 23, 845–
8
52.
(a) A. De Waard, S. Hickman and S. Kornfeld, J. Biol. Chem., 1976,
51, 7581–7587; (b) C. P. Sparrow, H. Leffler and S. H. Barondes, J.
2
Biol. Chem., 1987, 262, 7383–7390; (c) R. Ramkumar, A. Surolia and
S. K. Podder, Biochem. J., 1995, 308, 237–241; (d) D. Gupta, M. Cho,
R. D. Cummings and C. F. Brewer, Biochemistry, 1996, 35, 15236–
1
5243; (e) F. P. Schwarz, H. Ahmed, M. A. Bianchet, L. M. Amzel and
G. R. Vasta, Biochemistry, 1998, 37, 5867–5877; (f) S. Bharadwaj, H.
Kaltner, E. Y. Korchagina, N. V. Bovin, H.-J. Gabius and A. Surolia,
Biochim. Biophys. Acta, 1999, 1472, 191–196; (g) K. Bachhawat-Sikdar,
C. J. Thomas and A. Surolia, FEBS Lett., 2001, 500, 75–79; (h) N.
Ahmad, H.-J. Gabius, H. Kaltner, S. Andr e´ , I. Kuwabara, F.-T. Liu,
S. Oscarson, T. Norberg and C. F. Brewer, Can. J. Chem., 2002, 80,
1
096–1104; (i) M. Jim e´ nez, S. Andr e´ , C. Barillari, A. Romero, D.
Rognan, H.-J. Gabius and D. Sol ´ı s, FEBS Lett., 2008, 582, 2309–
312.
(a) B. G. Davis, S. J. Ward and P. M. Rendle, Chem. Commun., 2001,
2
7
8
1
2
2
89–190; (b) L. Szil a´ gyi, T.-Z. Illy e´ s and P. Herczegh, Tetrahedron Lett.,
001, 42, 3901–3903; (c) L. Szil a´ gyi and O. Varela, Curr. Org. Chem.,
006, 10, 1745–1770.
(a) J. P. Ribeiro, S. Andr e´ , F. J. Ca n˜ ada, H.-J. Gabius, A. Butera, R.
J. Alves and J. Jim e´ nez-Barbero, ChemMedChem, 2010, 5, 415–419;
(
b) J. P. Ribeiro, D. T. Carvalho, S. Andr e´ , F. J. Ca n˜ ada, R. J. Alves,
H.-J. Gabius and J. Jim e´ nez-Barbero, R. C. Chimie, 2011, 14, 96–
01.
1
9
M. Jim e´ nez, S. Andr e´ , H.-C. Siebert, H.-J. Gabius and D. Sol ´ı s,
Glycobiology, 2006, 16, 926–937.
1
0 (a) M. Gilleron, H.-C. Siebert, H. Kaltner, C.-W. von der Lieth, T.
Ko zˇ a´ r, K. M. Halkes, E. Y. Korchagina, N. V. Bovin, H.-J. Gabius
and J. F. G. Vliegenthart, Eur. J. Biochem., 1998, 252, 416–427; (b) J.
M. Alonso-Plaza, M. A. Canales, M. Jim e´ nez, J. L. Roldan, A.
Garcia-Herrero, L. Itturino, J. L. Asensio, F. J. Ca n˜ ada, A. Romero,
H.-C. Siebert, S. Andr e´ , D. Sol ´ı s, H.-J. Gabius and J. Jim e´ nez-Barbero,
Biochim. Biophys. Acta, 2001, 1568, 225–236; (c) M. C. Fernandez-
Alonso, F. J. Ca n˜ ada, D. Sol ´ı s, X. Cheng, G. Kumaran, S. Andr e´ ,
H.-C. Siebert, D. R. Mootoo, H.-J. Gabius and J. Jim e´ nez-Barbero,
Eur. J. Org. Chem., 2004, 1604–1613.
1 (a) H.-C. Siebert, S. Andr e´ , S.-Y. Lu, M. Frank, H. Kaltner, J. A. van
Kuik, E. Y. Korchagina, N. V. Bovin, E. Tajkhorshid, R. Kaptein, J.
F. G. Vliegenthart, C.-W. von der Lieth, J. Jim e´ nez-Barbero, J. Kopitz
and H.-J. Gabius, Biochemistry, 2003, 42, 14762–14773; (b) A. M. Wu,
T. Singh, J.-H. Liu, M. Krzeminski, R. Russwurm, H.-C. Siebert, A.
M. J. J. Bonvin, S. Andr e´ and H.-J. Gabius, Glycobiology, 2007, 17,
1
1
1
65–184; (c) M. Krzeminski, T. Singh, S. Andr e´ , M. Lensch, A. M.
Wu, A. M. J. J. Bonvin and H.-J. Gabius, Biochim. Biophys. Acta, 2011,
810, 150–161.
1
2 (a) M. A. Bianchet, H. Ahmed, G. R. Vasta and L. M. Amzel, Proteins:
Struct., Funct., Genet., 2000, 40, 378–388; (b) K. A. Stannard, P. M.
Collins, K. Ito, E. M. Sullivan, S. A. Scott, E. Gabutero, I. Darren
Grice, P. Low, U. J. Nilsson, H. Leffler, H. Blanchard and S. J. Ralph,
Cancer Lett., 2010, 299, 95–110.
3 (a) A. Rivera-Sagredo, D. Sol ´ı s, T. D ´ı az-Mauri n˜ o, J. Jim e´ nez-Barbero
and M. Mart ´ı n-Lomas, Eur. J. Biochem., 1991, 197, 217–228; (b) D.
Sol ´ı s, A. Romero, H. Kaltner, H.-J. Gabius and T. D ´ı az-Mauri n˜ o, J.
Biol. Chem., 1996, 271, 12744–12748.
1
1
4 O. E. Galanina, H. Kaltner, L. S. Khraltsova, N. V. Bovin and H.-J.
Gabius, J. Mol. Recognit., 1997, 10, 139–147.
5
454 | Org. Biomol. Chem., 2011, 9, 5445–5455
This journal is © The Royal Society of Chemistry 2011

View Article Online
B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B.
Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A.
Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M.
W. Wong, C. Gonzalez and J. A. Pople, Gaussian, Inc., Wallingford CT,
2004.
28 I. Bayly, P. Cieplak, W. Cornell and P. Kollman, J. Phys. Chem., 1993,
97, 10269–10280.
29 G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart,
R. K. Belew and A. J. Olson, J. Comput. Chem., 1998, 19, 1639–
1662.
30 T. A. Darden, D. York and L. G. Pedersen, J. Chem. Phys., 1993, 98,
10089–10092.
This journal is © The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5445–5455 | 5455