Synthesis of diglycosylated (di)sulfides and comparative evaluation of their antiproliferative effect against tumor cell lines: A focus on the nature of sugar-recognizing mediators involved

Article abstract of DOI:10.1016/j.carres.2019.107740

A mini-library of symmetrical and unsymmetrical diglycosyl (di)sulfides, containing D-galactose, L-fucose and N-acetyl glucosamine units, were synthesized and tested for the antiproliferative activity against cervix carcinoma (HeLa) and melanoma (A375) tumor cell lines as well as healthy fibroblasts (HDF). Comparative analysis of results seems to indicate that the most relevant antiproliferative effect is not primarily influenced by interactions with galectins, as the most cytotoxic compound observed for HeLa and A375 is not a ligand for such receptors. The most active molecules against HeLa and A375 lines also exhibited a good selectivity, showing a low toxicity to HDF cells. Obtained results offer useful indications for future design of structurally simple antitumor molecules based on sugar moieties with bridging sulfur atoms.

Full text of DOI:10.1016/j.carres.2019.107740

Accepted Manuscript
Synthesis of diglycosylated (di)sulfides and comparative evaluation of their
antiproliferative effect against tumor cell lines: A focus on the nature of sugar-
recognizing mediators involved
Sonia Di Gaetano, Emiliano Bedini, Alfredo Landolfi, Emilia Pedone, Luciano Pirone,
Michele Saviano, Serena Traboni, Domenica Capasso, Alfonso Iadonisi
PII:
S0008-6215(19)30326-X
DOI:
https://doi.org/10.1016/j.carres.2019.107740
Article Number: 107740
Reference:
CAR 107740
To appear in: Carbohydrate Research
Received Date: 25 May 2019
Revised Date: 29 June 2019
Accepted Date: 1 July 2019
Please cite this article as: S. Di Gaetano, E. Bedini, A. Landolfi, E. Pedone, L. Pirone, M. Saviano, S.
Traboni, D. Capasso, A. Iadonisi, Synthesis of diglycosylated (di)sulfides and comparative evaluation of
their antiproliferative effect against tumor cell lines: A focus on the nature of sugar-recognizing mediators
involved, Carbohydrate Research (2019), doi: https://doi.org/10.1016/j.carres.2019.107740.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
organic
diglycosyl
D-Gal, L-Fuc,
D-GlcNAc
synthesis
(di)sulfides
MINI-LIBRARY
comparative
of
evaluation
antiproliferative
effect
indication of
possible
tumor cells examined:
sugar-recognizing
mediators
HeLa, A375

ACCEPTED MANUSCRIPT
Synthesis of diglycosylated (di)sulfides and comparative evaluation of their antiproliferative
effect against tumor cell lines: a focus on the nature of sugar-recognizing mediators involved
Sonia Di Gaetano,a Emiliano Bedini,b Alfredo Landolfi,b Emilia Pedone,a Luciano Pirone,a Michele
Saviano,c Serena Traboni,b Domenica Capasso, d, * and Alfonso Iadonisi b, *
aInstitute of Biostructures and Bioimmaging, CNR Via Mezzocannone 16, 80134 Naples (Italy)
bDepartment of Chemical Sciences, University of Naples Federico II, Via Cinthia 4, 80126 Naples (Italy)
cIstituto di Crystallography, CNR, Via Amendola 122/o - 70126 Bari (Italy).
dDepartment of Pharmacy, University of Naples Federico II, Via Mezzocannone 16, 80134 Naples (Italy).
*Corresponding author at Dipartimento di Scienze Chimiche Università di Napoli Federico II, Via
Cinthia 4, 80126 Naples (Italy). E-mail: iadonisi@unina.it
*Corresponding author at Dipartimento di Farmacia Università di Napoli Federico II Via
Mezzocannone 16, 80134 Naples (Italy). E-mail: domenica.capasso@unina.it
Key-words: galectins, glycosyl sulfide, glycosyl disulfide, tumour cells, anti-proliferative activity.
Abstract
A mini-library of symmetrical and unsymmetrical diglycosyl (di)sulfides, containing D-galactose,
L-fucose and N-acetyl glucosamine units, were synthesized and tested for the antiproliferative
activity against cervix carcinoma (HeLa) and melanoma (A375) tumor cell lines as well as healthy
fibroblasts (HDF). Comparative analysis of results seems to indicate that the most relevant
antiproliferative effect is not primarily influenced by interactions with galectins, as the most
cytotoxic compound observed for HeLa and A375 is not a ligand for such receptors. The most
active molecules against HeLa and A375 lines also exhibited a good selectivity, showing a low
toxicity to HDF cells. Obtained results offer useful indications for future design of structurally
simple antitumor molecules based on sugar moieties with bridging sulphur atoms.
1. Introduction
Over the last years, a number of sparse data in the literature evidenced a growing interest towards
biological properties of pseudodisaccharide structures with one or two bridging sulfur atoms
between anomeric positions of both saccharide residues. An in vivo study with murine models
demonstrated that the symmetrical digalactosyl sulfide 1 (Figure 1) exhibits anti-tumor properties
against B16F10 melanoma and 4T1 orthotopic breast cancer models [1], an activity exerted through
several mechanisms. In this regard, the topical role played by galectin-1, a member of a broad class
of receptors recognizing β-galactoside moieties, was evidenced, as tumor cells depleted of this
galectin were found to be much less affected by 1.
In another research line, it was shown that symmetrical digalactosyl sulfides equipped with
aromatic substituents at C-3 of both sugar moieties can be potent inhibitors of galectin-3 [2], and
numerous structural variants of this core have been synthesized and tested for their inhibition
activity [3-7].

ACCEPTED MANUSCRIPT
These combined data support the idea that an effective antitumor agent with a sulfurated
pseudosaccharide structure should be a good ligand of galectins, that are indeed well known for
their multiple implications in tumor progression [8].
Some years ago we reported a synthetic approach to symmetrical diglycosyl disulfides and
examination of their cytotoxicity towards two tumor cell lines (melanoma and hepatocarcinoma cell
lines, A375 and HepG2, respectively) [9]. That investigation revealed an interesting cytotoxic
activity dependent on the nature of the appended sugars; in particular, for both cell lines the best
activity was observed with symmetrical disulfides containing N-acetyl glucosamine (GlcNAc), L-
fucose, and D-lactose. As some of these sugars (GlcNAc and fucose) are not expected to be good
ligands for galectins, a galectin-independent mechanism should account in this case for the
observed activity.
In order to gain further data to define a possible structure-activity relationship for such sulfurated
pseudodisaccharides, we undertook a general project aimed at clarifying the effect of several
structural parameters, such as the symmetry of the molecules, the specific nature of the residues and
the number of bridging sulfur atoms, on the cytotoxic activity against several relevant tumor
models. To this end, we wish herein to report a first comparative analysis within a mini-library of
four representative molecules towards two different tumor cell lines. The mini-library (Figure 1)
was composed of digalactosyl sulfide 1, the corresponding disulfide 2, and the unsymmetrical
sulfides 3 and 4, both containing a single GlcNAc unit, but differentiated for the presence of either a
D-galactose or a L-fucose residue, respectively.
The mini-library was designed on the basis of several combined premises such as: a) digalactosyl
sulfide 1 is the only derivative for which, as above discussed [1], an antitumor activity was reported
in in vivo experiments on murine models; b) comparison of galactosyl sulfide 1 and disulfide 2
allows to evidence the impact of the specific nature of the sulfurated bridge on their relative
activity; c) the structure of unsymmetrical sulfides was selected in such a way to have a derivative
with a potential affinity for galectin (compound 3) and another lacking this property (compound 4),
because of the presence or the absence of a β-galactoside moiety, respectively; d) the selection of
the saccharide residues present in 3 and 4 was driven taking into account both the nature of the
monosaccharides in the glycosyl disulfides endowed with the best anti-proliferative action against
tumor cells (see above) [9] and considerations of synthetic ease. In particular, D-GlcNAc and L-
fucose monosacchrides were preferred over lactose (a disaccharide) because in preliminary
experiments this latter was found to dispay a lower propensity to act either as a thioglycosyl donor
or acceptor.
Figure 1. Structures of diglycosyl sulfides or disufides investigated

ACCEPTED MANUSCRIPT
A comparative investigation on the cytotoxic activity of compounds 1-4 towards different tumor
cell lines, integrated with the evaluation of their affinity towards galectins, was expected to give
interesting indications on several issues such as the dependence of the activity on the bridging
functionality (sulfide or disulfide), the role played by galectins in the antiproliferative activity or the
feasible involvement of different sugar-recognizing proteins in the mechanism of action.
2 Results and discussion
2.1 Synthesis of compounds
Synthesis of 2 was performed with a methodology reported by us [9]. Compounds 1, 3 and 4 were
instead synthesized through approaches herein described.
Symmetrical sulfide 1 was prepared by adapting a previously reported methodology aimed at the
synthesis of thioglycosides [10]. As shown in Scheme 1, this is based on an initial fast conversion of
peracetylated galactose 5 into galactosyl iodide 6 through exposure to a slight excess of iodine and
polymethyl hydrosiloxane (PMHSA) at reflux for a few minutes [11, 12]. Crude glycosyl iodide 6
obtained from a simple liquid-liquid extraction, was then exposed to 0.5 equivalent thiourea in
acetonitrile at 60 °C. Under these conditions, half the amount of 6 was converted into glycosyl
isothiouronium 7 in short times, whereas unreacted residual 6 was sufficiently stable to survive
unaffected as also evidenced by TLC. Upon direct addition of TEA into the mixture, glycosyl
thiolate 8 was quickly generated from 7, and entrapped in situ by electrophilic iodide 6.
The resulting peracetylated sulfide 9 was isolated in an excellent overall yield (71% from 5) and
deprotected through a Zemplen deacetylation to afford 1. The herein proposed approach to key
intermediate 9 can be a valuable alternative to a method described in the literature[13, 14]which is
based on the double glycosidation of S2- dianion in benzene-containing solvent mixtures.
Scheme 1. Synthesis of galactosyl sulfide 1

ACCEPTED MANUSCRIPT
As to the synthesis of unsymmetrical sulfides 3 (Scheme 2) and 4 (Scheme 3), we used a common
general scheme based on a key coupling step committing an appropriate GlcNAc donor and either a
galactosyl or a fucosyl thiolate suitably generated in situ.
Scheme 2. Synthesis of asymmetrical sulfide 3
To this aim, we took advantage of straightforward approaches available for accessing the requisite
building-blocks. In particular, the acetyl-protected glycosyl chloride 10 (Schemes 2 and 3) was
selected as a suitable GlcNAc precursor owing to its easy one-step preparation from cheap N-acetyl
glucosamine [15]; nucleophilic galactosyl and fucosyl thiolates 12 and 15 were instead generated
through in situ reduction of the corresponding disulfides 11 and 14 (Scheme 2 and 3, respectively),
prepared according to a previously reported procedure [9].
Scheme 3. Synthesis of asymmetrical sulfide 4.
The key coupling steps were performed upon exposure of disulfide 11 or 14 to sodium borohydride
in acetonitrile for 30 minutes at rt, and subsequent addition of chloride 10 (Scheme 2 and 3,
respectively). Optimization studies evidenced that addition of a further aliquot of sodium
borohydride (30 minutes after the GcNAc donor addition) was found beneficial in both syntheses

ACCEPTED MANUSCRIPT
for improving yields. Unsymmetrical sulfides 13 and 16 were indeed isolated in 60-62% yield
(Scheme 2 and 3, respectively). Standard Zemplen deacetylation afforded desired deprotected
unsymmetrical sulfides 3 and 4 in almost quantitative yield.
2.2 Biological evaluation
To evaluate the effect of compounds 1-4 on tumor cell proliferation, melanoma (A375) and cervix
carcinoma (HeLa) cells were incubated with the molecules at 0.2, 1 and 5 mM concentrations for 48
h. The HDF cells were used as normal cellular model to evaluate the selectivity of action of
examined molecules towards tumor cell in comparison with normal cells [16].
Analysis of the effect on tumor cells evidenced that 1 induces a dose-independent inhibition of
HeLa proliferation (about 25%) and does not have significant effect on A375 cells. Compound 2
results substantially inactive at all tested concentrations on both cell lines. In contrast, asymmetrical
sulfides 3 and 4 displayed a dose-dependent effect in both cases.
Compound 3 is formally obtained from 1 with the replacement of a galactose residue with a
GlcNAc unit; such a structural change results in a more active molecule that reduces HeLa and
A375 proliferation of 50% and 25%, respectively, at 5 mM. Further replacement of the galactose
unit in 3 with a L-fucose residue provides the most active molecule of the mini-library, 4, which
displays a 36% and 53% of HeLa proliferation inhibition at 1mM and 5 mM, respectively, and 28%
and 41% on A375 at 1 and 5 mM, respectively. Noteworthy, 1, 3 and 4 do not affect normal
fibroblasts (Figure 2, C), showing a selectivity towards tumor cells, whereas 2 is active only to HDF
but at the highest tested concentration.
A
140.0
120.0
100.0
*
80.0
60.0
40.0
20.0
0.0
0.2mM
1mM
*
*
*
*
5mM
B
120.0
100.0
80.0
60.0
40.0
20.0
0.0
*
0.2mM
1mM
5mM

ACCEPTED MANUSCRIPT
C
140.0
120.0
100.0
80.0
60.0
40.0
20.0
0.0
1 mM
*
5 m M
Control
1
2
3
4
Figure 2. Proliferation assay on HeLa (A), A375 (B), HDF (C) cells. Cells were treated with molecules at indicated
concentrations or vehicle, and incubated for 48 h. Proliferation was evaluated using MTT assay. The results are
presented as the percentage of proliferating cells respect to the control (vehicle treated cells) and are expressed as means
± SE of three independent experiments performed in triplicate. Statistical significance was analyzed using Student’s t
test, unpaired, two-sided (* p < 0.05)
.
Overall, it should be outlined that compounds 3, 4 (on HeLa and A375) perform better than 1, the
only antitumor compound of the series reported to date in in vivo experiments [1]. In agreement
with our data, 1 was reported not to be especially cytotoxic and its in vivo activity was indeed
demonstrated to be related with anti-migratory and anti-angiogenic activity [1]. Such an activity
was accounted for by the ability of 1 to bind galectins as supported by knock-down experiments.
The potential of examined cell lines to be involved in galectin-mediated processes was confirmed
by western blotting analysis of representative galectin-3 and -9; as shown in Figure 4, A375 and
HeLa cells expressed galectin-3 and not galectin-9.
As a matter of fact, a general trend observed with A375 and HeLa cells is the higher dose-
dependent activity displayed by mixed sulfides 3 and 4, the former being the only expected galectin
ligand (because of its β-galactoside moiety). Comparative results at 5 mM concentration evidence
similar activity of 3 and 4, whereas 4 performs even better at 1 mM.
Collected data from A375 and HeLa cells seem to suggest that a galectin-independent mechanism
should play a major role in explaining the antiproliferative effect of the most active compounds 3
and 4. To support this view, both 3 and 4 were found not to be interacting with galectin 3, as shown
in ITC experiments (see Supporting Information). In addition, the inactive compound 1 was
reported to be a good ligand for several galectins (including galectin-3 and 9) [17].
On this basis, the hypothesis of the involvement of alternative sugar-recognizing proteins might be
considered. The observed activity of molecules containing GlcNAc and L-fucose on cancer cells,
such as A375 and HeLa, could be justified by the interaction with different proteins implicated in
glycan metabolism, such as N-acetylglucosaminyltransferase-V (GnT-V). Indeed, numerous studies
pointed to this protein, an enzyme that acts on different substrates including GlcNAc, as one of the
most characterized cancer-associated glycosyltransferases [18]. GnT-V transfers a GlcNAc residue
from donor UDP-GlcNAc to the α-(1-6)-linked mannose of N-glycans to form a β-(1-6) linkage
[19]. On the other hand, fucosylation plays important roles in signal transduction pathways.
Although detailed mechanisms are still unclear, increased levels of fucose are frequently found in
the serum and urine of cancer patients, and increased fucosylation is found in cancer cells [20].

ACCEPTED MANUSCRIPT
In addition to these considerations, at this stage it is pertinent to remind that in our previous studies
on symmetrical diglycosyl disulfides, the best anti-proliferative activity against A375 and HepG2
tumor lines was recorded with the N-acetyl glucosamine disulfide, followed by the lactose and L-
fucose derivatives, these latter with a comparable activity [9]. Among these active disulfides, only
in the lactose derivative is present the β-galactoside moiety which should be recognized by
galectins.
1
2
3
Gal 3
Actin
Gal3 Gal9
1. HeLa
2. A375
3. HDF
+
+
+
-
-
+
1
2
3
Gal 9
Actin
Figure 4: Protein lysates derived from different cell lines (HeLa, A375, HDF) were investigated by
western blotting to analyse the presence of galectins using anti-gal3 or anti-gal9 antibody. Actin
was used as the normalization control. The blot is representative of 3 experiments with comparable
result.
These combined data support the idea that activity of diglycosyl disulfides is rather sugar-
dependent, but reasonably not critically related to the interaction of the molecules with galectins.
3. Conclusions
In this investigation we have pursued a combined approach based on organic synthesis and cellular
biology to shed light on mechanisms underlying the antiproliferative activity against tumor lines of
diglycosylated structures with one or two bridging sulfur atoms. For this purpose, a minilibrary of
four structurally varied diglycosyl (di)sulfides was rationally designed and synthesized through
streamlined approaches.
Resulting derivatives 1-4 were tested for their antiproliferative activity towards HeLa, A375, as
well as HDF healthy cells. Comparative analysis of antiproliferative action exhibited a similar
profile with HeLa and A375 cells, with a dose-indipendent effect of compounds 1-2, and a dose-
dependent for 3-4, which proved the most active molecules at 5 mM.
The antiproliferative effect does not seem to be primarily influenced by interaction of molecules
with galectins. Indeed, the most cytotoxic compound observed for HeLa and A375 is not a ligand
for such receptors, whereas inactive compound 1 was found to be a ligand for several galectins [17].
Remarkably, with both tumor cell lines, compounds with a higher anti-proliferative activity than
compound 1, the only anti-tumor agent tested in in vivo experiments to date, were disclosed.
Therefore, these results strongly encourage a further effort in the search for ever more effective
anti-tumor compounds, also in view of the low cost and the relative synthetic ease in the assembly
of the pseudodisaccharide molecules in object. Evidences herein reported indicate that sugars other
than galactose should be considered in the future design of potentially active compounds.

ACCEPTED MANUSCRIPT
4. Experimental section
4.1 General
1H and 13C NMR were recorded with a Brucker 400 MHz device. Data for 1H NMR are reported as
follows: chemical shift (δ ppm), multiplicity (s=singlet, d=doublet, t=triplet, q=quartet,
m=multiplet), coupling constant (Hz), integration, signal assignation.
4.2.1 Synthesis of 2,3,4,6-Tetra-O-acetyl-β-D-galactopyranosyl 2,3,4,6-Tetra-O-acetyl-1-thio-β-D-
galactopyranoside 9.
To a solution of penta-O-acetyl galactose (320 mg, 0.82 mmol) in anhydrous DCM (4 mL) was
added I2 (229 mg, 0.90 mmol), and polymethylhydrosiloxne (PMHS) (54 µL, 0.90 mmol -SiH
units) (caution: exothermic reaction). The system was refluxed until complete consumption of the
starting material (5–10 min as monitored by TLC, eluent: hexane/ethyl acetate 1:1). The mixture
was then diluted with DCM and the organic phase washed with aq. sodium carbonate containing
sodium thiosulfate (this latter is added portionwise until consumption of residual iodine in the
organic phase). The aqueous phase was re-extracted with DCM, and collected organic phases were
dried with anhydrous sodium sulfate, filtrated and concentrated. Thiourea (31 mg, 0.41 mmol) was
added to the crude residue and the mixture suspended in acetonitrile (4 mL), and then heated to 60
°C. After 15 min TLC analysis evidenced the presence of comparable amounts of residual glycosyl
iodide and a very polar product. The vessel was cooled to room temperature and triethylamine (0.22
mL, 1.6 mmol) was added. The mixture was left under stirring for two hours and then concentrated
in vacuo. The residue was dissolved in DCM, and the organic phase washed with water. The
aqueous phase was then re-extracted with DCM and combined organic phases were dried with
sodium sulfate, filtrated and concentrated. Silica gel flash-chromatography of the residue (eluent:
hexane/ethyl acetate from 6:4 to 3:7) yielded sulfide 9 [13, 14] as an oil (202 mg, 71 % overall
yield). 1H NMR (400 MHz, CDCl3)
δ 5.44 (d, 1H, J = 2.8 Hz, H-4), 5.39 (t, 1H, J = 10.0 Hz, H-2),
5.06 (dd, 1H, J = 2.8 and 10.0 Hz, H-3), 4.88 (d, 1H, J = 9.6 Hz, H-1), 4.25-4.10 (m, 2H, 6-H2),
4.05 (bt, 1H, J = 6.4 Hz, H-5), 2.19, 2.08, 2.04, 1.97 (4 x s, 12H, 4 x COCH3). 13C NMR (100 MHz,
CDCl3)
δ 170.2, 170.1, 169.9, 169.6 (-COCH3), 81.1 (C-1), 75.4, 71.5, 69.7, 67.0, 60.8; 20.8, 20.7,
20.6, 20.5 (4 x –COCH3). Anal. Calcd. for C28H38O18S: C, 48.41; H, 5.51. Found: C, 48.55; H, 5.45.
MALDI HRMS m/z [M + Na]+ calc. for (C28H38O18SNa) 717.1677, found 717.1678.
4.2.2 Synthesis of unsymmetrical peracetylated sulfide 2,3,4,6-Tetra-O-acetyl-
β-D-
galactopyranosyl 2-deoxy-2-acetammido-3,4,6-tri-O-acetyl-1-thio- -D-glucopyranoside13.
β
NaBH4 (20 mg, 0.51 mmol) was added under argon to a solution of galactosyl disulphide 11 [9]
(123 mg, 0.17 mmol) in anhydrous acetonitrile (4 mL). After stirring for 30 minutes at rt, glycosyl
chloride 10 [15] (124 mg, 0.34 mmol) was added. After further 30 minutes, a new aliquot of NaBH4
(20 mg, 0.51 mmol) was added, and the mixture was then left under overnight stirring at rt. After
quenching the reaction with acetic acid, the mixture was concentrated in vacuo. The residue was
dissolved in DCM, and the organic phase washed with water. The aqueous phase was then re-
extracted with DCM and combined organic phases were dried with sodium sulfate, filtrated and
concentrated. Silica gel flash-chromatography of the residue (eluent: hexane/ethyl acetate from 1:4
to 0:1) yielded unsymmetrical sulfide 13 as a foam (146 mg, 62 % yield). [α]D – 10.8 (c 1.15,

ACCEPTED MANUSCRIPT
CHCl3). 1H NMR (400 MHz, CDCl3)
δ 6.06 (bd, 1H, J = 9.2 Hz, NH-2 GlcNAc), 5.37 (bd, 1H, J =
3.2 Hz, H-4 Gal), 5.15 (t, 1H, J = 10.0 Hz), 5.10 (t, 1H, J = 10.0 Hz), 5.05-4.95 (overlapped signals,
2H), 4.81 (d, 1H, J = 10.0 Hz, H-1), 4.75 (d, 1H, J = 10.0 Hz, H-1), 4.15 (m, 1H, H-2 GlcNAc),
5.05-4.95 (overlapped signals, 4H), 3.87 (t, 1H, J = 6.4 Hz, H-5 Gal), 3.62 (m, 1H, H-5 Gal), 2.09,
2.01, 2.00, 1.99, 1.97 (x2), 1.94, 1.90 (7 x s, 24H, 8 x COCH3). 13C NMR (100 MHz, CDCl3)
δ
170.6, 170.4, 170.2, 170.0, 169.9, 169.8, 169.5, e 169.1 (-COCH3), 81.3 and 81.1 (C-1), 75.9, 74.6,
73.7, 71.6, 68.2, 67.1, 62.1, 61.2, 52.8, 22.8, 20.5-20.3 (-COCH3). Anal. Calcd. for C28H39NO17S:
C, 48.48; H, 5.67. Found: C, 48.35; H, 5.70. MALDI HRMS m/z [M + Na]+ calc. for
(C28H39NO17S) 716.1836, found 716.1838.
4.2.3 Synthesis of 2,3,4-Tri-O-acetyl-β-L-fucopyranosyl 2-deoxy-2-acetammido-3,4,6-tri-O-acetyl-
1-thio- -D-glucopyranoside 16.
β
Compound 16 was synthesized with a similar procedure as for 13, adopting fucosyl disulfide 14 (92
1
mg, 0.15 mmol) in place of galactosyl disulfide 11. [α]D – 6.3 (c 1.15, CHCl3). H NMR (400
MHz, CDCl3)
δ
5.82 (bd, 1H, J = 8.4 Hz, NH-2 GlcNAc), 5.30-5.20 (overlapped signals, 3H),
5.08 (t, 1H, J = 10.0 Hz, H-4 GlcNAc), 4.92 (dd, 1H, J = 3.2 and 10.4 Hz, H-3 Fuc), 4.92 (d, 1H, J
= 10.4 Hz, H-1 GlcNAc), 4.57 (d, 1H, J = 10.0 Hz, H-1 Fuc), 4.24 (dd, 1H, J = 5.2 and 12.8 Hz, H-
6a GlcNAc), 4.06 (dd, 1H, J = 2.4 and 12.8 Hz, H-6b GlcNAc), 3.94 (bq, 1H, J = 10.4 Hz, H-2
GlcNAc), 3.87 (bq, 1H, J = 6.4 Hz, H-5 Fuc), 3.51 (m, 1H, H-5 GlcNAc), 2.17, 2.07, 2.03, 2.03,
2.01, 1.94, 1.91 (7 x s, 21H, 7 x COCH3), 1.19 (d, 3H, J = 6.4 Hz, H3-6 Fuc).13C NMR (100 MHz,
CDCl3)
δ
170.8, 170.7, 170.6, 170.4, 170.1, 170.0, 169.3 (-COCH3); 81.3, 81.0 (C-1 GlcNAc and
Fuc), 76.1, 73.7, 73.6, 72.4, 70.3, 68.3, 67.4, 62.2, 53.8; 23.2, 20.7-20.4 (7 x –COCH3), 16.4 (C-6).
Anal. Calcd. for C26H37NO15S: C, 49.13; H, 5.87. Found: C, 49.25; H, 5.90. MALDI HRMS m/z [M
+ Na]+ calc. for (C26H37NO15SNa) 658.1782, found 658.1787.
4.2.4 General procedure for Zemplen deacetylation.
To a solution of peracetylated diglycoside sulfide (0.2-0.4 mmol) in MeOH was added a solution of
sodium methoxide in methanol (0.1 eq). The mixture was left under stirring at rt until detection via
TLC (eluent: ethyl acetate) of the reaction completion (generally reactions took 2-4 hours). The
mixture was treated with Amberlyst H+ resin until neutrality, and supernatant liquid was transferred
via pipette to another vessel. Resin was washed with further methanol which was then transferred to
the same vessel. Removal of methanol in vacuo provided deprotected sulfides.
β-D-galactopyranosyl 1-thio-β-D-galactopyranoside (1):1H NMR (400 MHz, D2O)
δ 4.86 (d,
1H, J = 9.6 Hz, H-1), 4.04 (d, 1H, J = 3.2 Hz, H-4), 3.90-3.75 (overlapped signals, 3H, H-5 and H2-
6), 3.74 (dd, 1H, J = 3.2 and 9.6 Hz, H-3), 3.68 (t, 1H, J = 9.6 Hz, H-2). 13C NMR (100 MHz, D2O)
δ
83.5 (C-1), 79.0, 73.9, 69.6, 68.8, 61.2. Anal. Calcd. for C12H22O10S: C, 40.22; H, 6.19. Found:
C, 40.10; H, 6.25. MALDI HRMS m/z [M + Na]+ calc. for (C12H22O10SNa) 381.0831, found
381.0835.
β-D-galactopyranosyl 2-deoxy-2-acetammido-1-thio-β-D-glucopyranoside (3): 1H NMR (400
MHz, D2O)
δ 4.85 (d, 1H, J = 10.4 Hz, H-1 GlcNAc), 4.64 (d, 1H, J = 10.0 Hz, H-1 Gal), 3.86 (d,
1H, J = 3.2 Hz, H-4 Gal), 3.82 (dd, 1H, J = 1.6 and 12.4 Hz, H-6b GlcNAc), 3.73-3.60 (overlapped
signals, 6H), 3.55 (dd, 1H, J = 3.2 and 9.6 Hz, H-3 Gal), 3.50-3.42 (overlapped signals, 2H), 3.38

ACCEPTED MANUSCRIPT
(m, 1H, H-5 GlcNAc), 1,94 (s, 3H, COCH3). 13C NMR (100 MHz, D2O):
δ 174.4, 83.3, 81.7, 79.8,
78.9, 74.9, 73.8, 69.6, 69.3, 68.6, 61.2, 60.8, 54.4, 22.1. Anal. Calcd. for C14H25NO10S: C, 42.10; H,
6.31. Found: C, 42.05; H, 6.35. MALDI HRMS m/z [M + Na]+ calc. for (C14H25NO10SNa)
422.1097, found 422.1100.
β-L-fucopyranosyl 2-deoxy-2-acetammido-1-thio-β-D-glucopyranoside (4): 1H NMR (400
MHz, D2O)
δ 4.81 (d, 1H, J = 10.4 Hz, H-1 GlcNAc), 4.37 (d, 1H, J 9.6 Hz, H-1 Fuc,), 3.79 (dd,
1H, J 2.0, 12.4 Hz, H-6a GlcNAc), 3.72 (q, 1H, J 6.4 Hz, H-4 Fuc), 3.70-3.65 (overlapped signals,
2H), 3.62 (dd, 1H, J = 5.2 and 12.4 Hz, H-6b GlcNAc), 3.57-3.50 (overlapped signals, 2H), 3.48-
3.35 (overlapped signals, 3H), 1.93 (s, 3H, COCH3), 1.14 (d, 3H, J = 6.4 Hz, H3-6 Fuc). 13C NMR
(100 MHz, D2O) δ 174.3, 84.8, 82.5, 79.4, 75.1 (x2), 73.7, 71.3, 69.4, 69.2, 60.6, 54.6, 22.1, 15.7.
Anal. Calcd. for C14H25NO9S: C, 43.86; H, 6.57. Found: C, 43.75; H, 6.65. MALDI HRMS m/z [M
+ Na]+ calc. for (C14H25NO9SNa) 406.1148, found 406.1154.
4.3 Biological evaluation
4.3.1 Cell lines and culture conditions
Human cervix carcinoma (HeLa), human melanoma (A375) cell lines (ATCC U.S.), and human
dermal fibroblasts (HDF), gently provided by Dr Annalisa Tito (Arterra Bioscience) were grown in
DMEM supplemented with 10% fetal bovin serum (FBS), 2.5 mM glutamine, 100 U/mL penicillin
and 100 µg/mL streptomycin (Euroclone, Milano, Italy) [21]. The cells were maintained in
humidified air containing 5% CO2 at 37 °C.
4.3.2 Proliferation assay
Cells were plated at density: 2x103 HeLa, 5x103, A-375, 7x103 and HDF per 100 µL medium in 96-
well plates and incubated with molecules, previously solubilized in 10% DMSO or with vehicle, at
indicated concentrations. After 48 h incubation at 37 °C, cell viability was assessed by performing
the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Sigma Aldrich, Italy)
reduction inhibition assay [22]. Cytotoxicity experiments were independently performed at least
three times. Statistical significance between groups was assessed by Student’s t-test. Data are
expressed as means standard error (S.E.). P values < 0.05 were considered to be statistically
significant.
4.3.3 Western Blot
Whole cell lysates of different cell lines were obtained by using lysis buffer (50 mM Hepes pH 7.4,
50 mM NaCl, 1% Triton) supplemented with protease inhibitor cocktail (Roche Milano, Italy) [23].
Briefly, cells were incubated in lysis buffer for 30 minutes on ice and centrifuged at 13,000 rpm for
30 min. The supernatant was removed and its protein concentrations was determined by Bradford
method using Bio-Rad reagent (Bio-Rad Laboratories, Italy). Successively, lysates were loaded on a
SDS-polyacrylamide gel electrophoresis (SDS-PAGE) then transferred to a PVDF membrane (Bio-
Rad). The membrane was incubated with the primary antibody (anti-gal3, anti-gal9 and actin
respectively from R&D and Sigma-Aldrich) for 1 h at RT and with a secondary antibody
conjugated with horseradish peroxidase. Signals obtained with the enhanced chemiluminescence

ACCEPTED MANUSCRIPT
detection system (Euroclone) were visualized and acquired with ChemiDoc XRS System (Bio-Rad
Laboratories, Italy) and analysed with the QuantityONE software.
4.3.4 ITC studies
The interactions of Gal3 CRD with 3 and 4 were investigated by isothermal titration calorimetry
using an ITC200 calorimeter (MicroCal/GE Healthcare) [24]. ITC experiments were conducted in a
buffer containing 20 mM Tris-HCl, 150 mM NaCl, mM DTT pH 7.0. 3 (2 mM) and 4 (2 mM) were
titrated into 20ꢀµM of Gal3 CRD. Heat of reaction was measured at 22ꢀ°C for 27 injections. To
minimize the contribution of heat of dilution to the measured heat change, 3 and 4 were prepared in
the same buffer. Lactose (500 µM) was used as positive control. All data were analyzed and fitted
using the Microcal Origin version 7.0 software package. Dissociation constants were determined by
fitting the data using a one-set-of-site-binding model. ITC runs were repeated twice to evaluate the
reproducibility of the results.
References
[1] K. Ito, S.A. Scott,S. Cutler S, L.F. Dong, J. Neuzil, H. Blanchard, S.J. Ralph, Thiodigalactoside
inhibits murine cancers by concurrently blocking effects of galectin-1 on immune dysregulation,
angiogenesis and protection against oxidative stress, Angiogenesis 14 (2011) 293-307.
[2] I. Cumpstey , A. Sundin, H. Leffler, U.J. Nilsson, C2-Symmetrical thiodigalactoside bis-
benzamido derivatives as high-affinity inhibitors of Galectin-3: efficient lectin inhibition through
double arginine–arene interactions, Angew. Chem. Int. Ed. 44 (2005) 5110-5112.
[3] K. Peterso, R. Kumar, O. Stenström, P. Verma, P.R. Verma, M. Håkansson, B. Kahl-Knutsson,
F. Zetterberg, H. Leffler, M. Akke, D.T. Logan, U.J. Nilsson, Systematic tuning of fluoro-galectin-3
interactions provides thiodigalactoside derivatives with single-digit nM affinity and high selectivity,
J. Med. Chem. 61 (2018) 1164-1175.
[4] T.J. Hsieh, H.Y. Lin, Z. Tu, T.C. Lin, S.C. Wu, Y.Y. Tseng, F.T. Liu, S.T.D. Hsu, C.H. Lin,
Dual thio-digalactoside-binding modes of human galectins as the structural basis for the design of
potent and selective inhibitors, Sci. Rep. 6 (2016); 29457.
[5] V.K. Rajput, A. MacKinnon, S. Mandal, P. Collins, H. Blanchard, H. Leffler, T. Sethi, H.
Schambye, B. Mukhopadhyay, U.J. Nilsson, A selective galactose–coumarin-derived galectin-3
inhibitor demonstrates involvement of galectin-3-glycan interactions in a pulmonary fibrosis model,
J. Med. Chem. 59 (2016) 8141-8147.
[6] F.R. Zetterberg, K. Peterson, R.E. Johnsson, T. Brimert, M. Håkansson, D.T. Logan, H. Leffler,
U.J. Nilsson, Monosaccharide derivatives with low‐nanomolar lectin affinity and high selectivity
based on combined fluorine–amide, phenyl–arginine, sulfur–π, and halogen bond interactions,
ChemMedChem. 13 (2018) 133-137.
[7] A.L. Noresson, O. Aurelius, C.T. Oberg, O. Engstrom, A.P. Sundin, M. Hakansson, O.
Stenstrom, M. Akke, D.T. Logan, H. Leffler, U.J. Nilsson, Designing interactions by control of

ACCEPTED MANUSCRIPT
protein-ligand complex conformation: tuning arginine-arene interaction geometry for enhanced
electrostatic protein-ligand interactions, Chem. Sci. 9 (2018) 1014-1021.
[8] F.T. Liu, G.A. Rabinovich, Galectins as modulators of tumour progression, Nat Rev Cancer 5
(2005) 29-41.
[9] M. Adinolfi, D. Capasso, D. Di Gaetano, A. Iadonisi, L. Leone, A. Pastore, A straightforward
synthetic access to symmetrical glycosyl disulfides and biological evaluation thereof, Org. Biomol.
Chem. 9 (2011) 6278-6283.
[10] S. Valerio, A. Iadonisi, A. Ravidà, Novel approaches for the synthesis and activation of thio-
and selenoglycoside donors, J. Org. Chem. 72 (2007) 6097-6106.
[11] M. Adinolfi, A. Iadonisi, A. Ravidà, M. Schiattarella, Efficient and direct synthesis of
saccharidic 1,2-ethylidenes, orthoesters, and glycals from peracetylated sugars via the in situ
generation of glycosyl iodides with I2/Et3SiH, Tetraedron Lett. 44 (2003) 7863-7866.
[12] M. Giordano, A. Iadonisi, Polymethylhydrosiloxane (PMHS): a convenient option for synthetic
applications of the iodine/silane combined reagent. Straightforward entries to 2-hydroxyglycals and
useful building-blocks of glucuronic acid and glucosamine, Eur. J. Org. Chem. (2013) 125-131.
[13] J. Bogusiak, W. Szeja, Synthesis of bis(2,3,4,6-tetra-O-acetyl-beta-D-glucopyranosyl) and
(2,3,4,6-tetra-O-acetyl-beta-D-galactopyranosyl) sulfide under phase-transfer catalysis, Carbohydr.
Res. 141 (1985) 165-167.
[14] F. Chretien, P. Di Cesare, B. Gross. Synthesis of thiodisaccharides using phase-transfer
catalysis, J. Chem. Soc., Perkin Trans I. (1988) 3297-3000.
[15] T.V. RajanBabu, T.A. Ayers, G.A. Halliday, K.K. You, J.C. Calabrese. Carbohydrate
phosphinites as practical ligands in asymmetric catalysis: Electronic effects and dependence of
backbone chirality in Rh-catalyzed asymmetric hydrogenations. Synthesis of R- or S-amino acids
using natural sugars as ligand precursors, J. Org. Chem. 62 (1997) 6012-6028.
[16] C. Vicidomini, F. Cioffi, K. Broersen, V. Roviello, C. Riccardi, D. Montesarchio, D. Capasso,
S. Di Gaetano, D. Musumeci, G.N. Roviello, Benzodifurans for biomedical applications: BZ4, a
selective anti-proliferative and anti-amyloid lead compound, Future Med. Chem. 11 (2019) 285-
302.
[17] S. André, K. E. Kover, H.-J. Gabius, L. Szilagyi, Thio- and selenoglycosides as ligands for
biomedically relevant lectins: valency–activity correlations for benzene-based dithiogalactoside
clusters and first assessment for (di)selenodigalactosides, Biorg. Med. Chem. Lett. 25 (2015), 931-
935.
[18] M.G. Shoreibah, O. Hindsgaul, M. Pierce, Purification and characterization of rat-kidney UDP-
N-acetylglucosamine-alpha-6-D-mannoside beta-1,6-N-acetylglucosaminyl transferase, J. Biol.
Chem. 267 (1992) 2920-2927.
[19] J. Gu, A. Nishikawa, N. Tsuruoka, M. Ohno, N. Yamaguchi, K. Kangawa, N. Taniguchi,
Purification and characterization of UDP-N-acetylglucosamine-alpha-6-D-mannoside beta-1,6-N-

ACCEPTED MANUSCRIPT
acetylglucosaminyl transferase (N-acetylglucosaminyl transferase V) from a human lung-cancer
cell line, J. Biochem. 113 (1993) 614-619.
[20] N.G. Sawke, G.K. Sawke, Serum fucose level in malignant diseases, Indian J. Cancer. 47
(2010) 452-457.
[21], A. Del Gatto, L. Pirone, D. Comegna, L. Zaccaro, M. Saviano, B. Arcà, D. Capasso, E.
Pedone, A selective alpha(v)beta(5) integrin antagonist hidden into the anophelin family protein
cE5 from the malaria vector Anopheles gambiae, Peptide Sci. 110 (2018) e24054
[22] J. Amato, A. Pagano, D. Capasso, S. Di Gaetano, M. Giustiniano, E. Novellino, A. Randazzo,
B. Pagano, Targeting the BCL2 gene promoter G-quadruplex with a new class of furopyridazinone-
based molecules, ChemMedChem. 13 (2018) 406-410.
[23] D. Capasso, S.Di Gaetano, V. Celentano, D. Diana, L. Festa, R. Di Stasi, L. De Rosa, R.
Fattorusso, L.D. D’Andrea, Molecular Biosystems. 13 (2017) 1619-1629.
[24] N. Balasco, L. Pirone, G. Smaldone, S. Di Gaetano, L. Esposito, E. Pedone, L. Vitagliano,
Biochim Biophys Acta. 1844 (2014) 1289-1298.

ACCEPTED MANUSCRIPT
Rapid synthesis of a minilibrary of four representative diglycosyl (di)sulfides
Comparative evaluation of anti-proliferative activity toward tumor cell lines
Evidence of negligible role of galectins in the antiproliferative mechanism