A Modular Approach to a Library of Semi-Synthetic Fucosylated Chondroitin Sulfate Polysaccharides with Different Sulfation and Fucosylation Patterns

Article abstract of DOI:10.1002/chem.201603525

Fucosylated chondroitin sulfate (fCS)—a glycosaminoglycan (GAG) found in sea cucumbers—has recently attracted much attention owing to its biological properties. In particular, a low molecular mass fCS polysaccharide has very recently been suggested as a s

Full text of DOI:10.1002/chem.201603525

DOI: 10.1002/chem.201603525
Full Paper
&
Polysaccharide Chemistry
A Modular Approach to a Library of Semi-Synthetic Fucosylated
Chondroitin Sulfate Polysaccharides with Different Sulfation and
Fucosylation Patterns
Antonio Laezza,[a] Alfonso Iadonisi,[a] Anna V. A. Pirozzi,[b] Paola Diana,[b] Mario De Rosa,[b]
Chiara Schiraldi,[b] Michelangelo Parrilli,[c] and Emiliano Bedini*[a]
Abstract: Fucosylated chondroitin sulfate (fCS)—a glycosami-
noglycan (GAG) found in sea cucumbers—has recently at-
tracted much attention owing to its biological properties. In
particular, a low molecular mass fCS polysaccharide has very
recently been suggested as a strong candidate for the devel-
opment of an antithrombotic drug that would be safer and
more effective than heparin. To avoid the use of animal
sourced drugs, here we present the chemical transformation
of a microbial sourced unsulfated chondroitin polysaccharide
into a small library of fucosylated (and sulfated) derivatives
thereof. To this aim, a modular approach based on the differ-
ent combination of only five reactions was employed, with
an almost unprecedented polysaccharide branching by O-
glycosylation as the key step. The library was differentiated
for sulfation patterns and/or positions of the fucose branch-
es, as confirmed by detailed 2D NMR spectroscopic analysis.
These semi-synthetic polysaccharides will allow a wider and
more accurate structure–activity relationship study with re-
spect to those reported in literature to date.
Introduction
and Fuc branches are sulfated to a various extent. Different va-
rieties of sea cucumbers from the seawaters of different geo-
graphical zones produce a fCS polysaccharide with a different
sulfation pattern at the Fuc and/or GalNAc sites.[4] The sulfation
profiles found up to now are depicted in Figure 1. It is worth
noting that in all fCSs studied up to now, the presence of two
or even more different sulfation patterns on the Fuc branches
of the same polysaccharide have been detected.
Fucosylated chondroitin sulfate (fCS) is a glycosaminoglycan
(GAG) found up to now exclusively in sea cucumbers (Echinoi-
dea, Holothuroidea). Its polysaccharide structure is usually con-
stituted of a repeating unit with N-acetyl-d-galactosamine
(GalNAc) and fucosylated d-glucuronic acid (GlcA) residues
linked together through alternating b-1!3 and b-1!4 glycosi-
dic bonds. The fucosylation decoration very often consists of
a single l-fucose (Fuc) branch per repeating unit, linked at po-
sition O-3 of the GlcA units through an a-configured glycosidic
bond,[1] but some slightly different structural features have
been very recently suggested for fCS extracted from Aposticho-
pus japonicus, Actinopyga mauritania, and Ludwigothurea
grisea. In the former two cases, a random fucosylation at either
the GlcA O-3 or GalNAc O-4/O-6 position has been proposed,[2]
whereas in the latter fCS an a-Fuc-(1!3)-Fuc disaccharide
branch was found.[3] Both the GlcA-GalNAc linear backbone
[a] A. Laezza, Prof. A. Iadonisi, Dr. E. Bedini
Figure 1. Sulfation patterns found in natural fCSs.
Department of Chemical Sciences, University of Naples Federico II
Complesso Universitario Monte S. Angelo, via Cintia 4, 80126 Naples (Italy)
E-mail: ebedini@unina.it
fCS shows interesting activity in biological events related to
inflammation, angiogenesis, cellular growth, cancer metastasis,
fibrosis, virus infection, hyperglycemia, atherosclerosis, and,
above all, coagulation and thrombosis.[1] Interestingly, its anti-
coagulant and antithrombotic activity seems to be driven by
both a serpin-dependent mechanism—mediated by antithrom-
bin (AT) and heparin cofactor II (HC-II)[5]—and a serpin-inde-
[b] A. V. A. Pirozzi, P. Diana, Prof. M. De Rosa, Prof. C. Schiraldi
Department of Experimental Medicine, Second University of Naples
via de Crecchio 7, 80138 Naples (Italy)
[c] Prof. M. Parrilli
Department of Biology, University of Naples Federico II
Complesso Universitario Monte S. Angelo, via Cintia 4, 80126 Naples (Italy)
Supporting information for this article and ORCID(s) for the authors are
available on the WWW under http://dx.doi.org/10.1002/chem.201603525.
Chem. Eur. J. 2016, 22, 1 – 13
1
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
pendent one, which makes fCS active also on AT- and HC-II-
free plasmas.[6] Furthermore, fCS seems to retain its activity
also when it is orally administered.[7]
Results and Discussion
The production of unsulfated chondroitin (CS-0) from several
The most widespread and long-term used anticoagulant microbial sources has recently been accomplished.[17] As with
drug is unfractionated heparin, which unfortunately display our recent works on the semi-synthesis of non-animal sourced
several limitations such as risks of bleeding, platelet aggrega- chondroitin sulfate polysaccharides,[18] the CS-0 starting materi-
tion, and contamination with toxic agents, as well as a short al for the preparation of the fCS and fC library was obtained
half-life and a reduced activity on congenital or acquired an- by fed-batch fermentation of Escherichia coli K4, followed by
tithrombin deficient patients.[8] The introduction of low- and microfiltration, protease treatment, diafiltration, and mild hy-
ultra-low-molecular-mass heparins has alleviated these prob- drolysis to eliminate both lipopolysaccharide (LPS) contami-
lems to some extent, nonetheless the development of even nants and unwanted fructosyl residues attached at position C-
more effective and safer antithrombotic drugs is still a hot 3 of some GlcA units.[19] After further purification steps using
topic.[9] To this aim, the anticoagulant and antithrombotic ethanol and repeated alcoholic precipitations, the unsulfated,
properties displayed by fCS have attracted much attention, defructosylated polysaccharide was obtained in 90% purity
even if some drawbacks must be taken into account: the grade, as evaluated by NMR analysis and capillary electropho-
severe regulations for animal-derived drugs and the high varia- resis.[20] The residual endotoxin content was evaluated by
bility of the sulfation patterns in polysaccharides extracted using the limulus amebocyte lysate (LAL) test and found to be
from animal tissues, which require strict control of the fCS sul- lower than 0.1 EUmgꢀ1 [21]
fation level. Indeed, a too highly sulfated GAG such as per-O- The transformation of CS-0 into eight different fC and fCS
.
sulfated chondroitin can induce a strong allergic-type re- polysaccharides was planned through a modular approach
sponse,[10] as demonstrated by hundreds of serious adverse (Scheme 1). After a preliminary esterification of GlcA carboxylic
events, including 149 deaths,[11] which occurred after the con- acid and protection of GalNAc 4,6-diol as benzylidene to con-
tamination of some heparin lots with per-O-sulfated chondroi- vert CS-0 into polysaccharide 1,[16] five reactions were applied
tin.[12] Furthermore, together with their anticoagulant activity, in two different sequences to access fCSs and fCs with Fuc
natural fCS polysaccharides exhibit some adverse effects, such branches placed exclusively on either the GlcA or GalNAc
as platelet aggregation, factor XII activation, and bleeding.[13] units. Indeed, in synthetic sequence A, glycosylation of poly-
As low-molecular-mass fCS polysaccharides do not show these saccharide acceptor 1, followed by i) capping of the unreacted
adverse effects, many efforts have recently been devoted to hydroxyls by acetylation, ii) oxidative cleavage of benzyl (Bn)
the development of mild and selective depolymerization meth- and benzylidene protecting groups,[22] iii) sulfation of the re-
ods aimed at shortening the fCS polymer chain without cleav- leased hydroxyls, and iv) global deprotection by ester hydroly-
ing the sulfate groups as well as the labile Fuc branches.[2b,14] sis, can furnish polysaccharides with Fuc branches exclusively
To avoid the use of animal sourced molecules, a total synthetic on the GlcA units. Alternatively (synthetic sequence B), by per-
approach of some fCS di- and trisaccharide fragments has also forming the acetylation and benzylidene ring oxidative cleav-
been accomplished.[15] To the best of our knowledge, their anti- age before the glycosylation reaction and then the Bn cleav-
coagulant activity has not yet been reported, nonetheless the age, sulfation, and ester hydrolysis steps, fCSs and fCs with Fuc
obtainment of higher fCS oligosaccharides by selective depoly- branches exclusively on the GalNAc units could be obtained. A
merization of the natural polysaccharide has very recently second level of differentiation of the target polysaccharides
pointed to the fact that nonasaccharide is the minimum struc- concerned the sulfation pattern of the Fuc branches. This
tural unit responsible for anticoagulant activity.[14b] An alterna- could be achieved by using a set of Fuc donors with a different
tive approach to non-animal sourced fCS species has very re- patterns of protecting groups. Indeed, Bn groups were oxida-
cently been proposed by us, by accomplishing the chemical tively cleaved before the sulfation step, whereas the benzoyl
fucosylation of a microbial sourced chondroitin polysaccharide. (Bz) ones were deprotected only at the end of the semi-syn-
Subsequent sulfation of the fucosylated polysaccharide afford- theses. Thus, a per-O-benzylated Fuc donor could be used to
ed for the first time a semi-synthetic fCS, displaying per-O-sul- obtain fCS polysaccharides with per-O-sulfated Fuc branches,
fated Fuc branches randomly linked at GlcA and GalNAc units whereas 2,3-di-O-Bn-4-O-Bz-Fuc and 2,4-di-O-Bn-3-O-Bz-Fuc
and with additional sulfate groups at either the O-4 or O-6 po- donors could allow the access to fCSs with 2,3-di-O-sulfated or
sitions of GalNAc.[16] A preliminary anticoagulant assay revealed 2,4-di-O-sulfated Fuc branches, respectively. It is worth noting
an activity profile similar to low-molecular-mass fCS polysac- that the 2,4-di-O-sulfation pattern of Fuc branches was chosen
charides obtained from natural sources, thus encouraging us because it is one of the most commonly found in natural
to further work on this topic. With the aim of performing fCSs[4] and it has also been identified to give the strongest anti-
a wide structure–activity relationship (SAR) investigation of fCS coagulant activity.[4b,23] However, some results published during
polysaccharides that would be not restricted to the natural the development of this work pointed to the independence of
structures found up to now, we have now prepared a small li- the anticoagulant and antithrombotic activities from the differ-
brary of six semi-synthetic fCSs with different linkage positions ent sulfation patterns of Fuc branches in natural fCSs.[3] The
and/or sulfation patterns of the Fuc branches, and two unsul- other two chosen Fuc sulfation patterns were selected be-
fated fucosylated chondroitins (fCs).
cause, to the best of our knowledge, they have been never
found in natural fCSs, and therefore will us allow to enlarge
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
2
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 1. Modular approach for the transformation of unsulfated chondroitin into a library of fCSs (see Scheme 2 for the structures of Fuc donors 2, 3, and
4).
the SAR investigations and also serve as additional standards
for structural characterization of fCSs from novel natural sour-
ces. A third level of differentiation could be also achieved
among the target polysaccharides by avoiding the sulfation
step in the semi-synthetic sequences, thus opening access to
non-natural fCs.
The key step of the semi-synthetic strategies depicted in
Scheme 1 was the glycosylation reaction that connected the
Fuc branches with the linear chondroitin backbone. Before per-
forming it, Fuc donors and polysaccharide acceptors were pre-
Scheme 2. Synthesis of Fuc donors 2, 3, and 4. (a) See ref. [25]; (b) see
ref. [26]; (c) i) TBDMSCl, imidazole, DMF, rt, 3 h, 61%; ii) BnBr, NaH, DMF, rt,
1 h, 93%; iii) TBAF, THF, rt, 2 h, 94%; (d) BzCl, 1:1 v/v pyridine/CH2Cl2, rt, 3 h,
80%.
pared. For the donors, per-O-benzylated Fuc derivatives with
both a N-phenyl-trifluoroacetimidate[24] and an ethyl thioglyco-
side[25] leaving group were shown to be efficient in the
random fucosylation of a polysaccharide,[16] nonetheless thio-
glycosides were preferred owing to their longer shelf life and
shorter sequence of steps for their preparation. Therefore, thio-
glycosides 2, 3, and 4 were selected as Fuc donors (Scheme 2).
The former two were synthesized from ethyl 1-thio-b-l-fuco-
pyranoside 5[25] according to known procedures,[25,26] whereas
4 was unprecedented. It was thus obtained by regioselective
protection of triol 5 in three steps (3-O-silylation, benzylation,
and de-O-silylation)[27] to give 6 (53% overall yield), which was
then benzoylated to afford 4 in 80% yield.
The polysaccharide acceptors for Fuc branching on GlcA or
GalNAc were 1 and 8, respectively. The former was already
known[16] and served also for the synthesis of the latter in two
steps (Scheme 3). First, the 2,3-diol on the GlcA units was pro-
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
3
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
fucosylation reactions. Therefore, 9i–iii were directly subjected
to further steps to transform them into target polysaccharides
fC-i and fCS-i–iii. In particular, they were first treated with
Ac2O and Et3N in CH3CN in the presence of a catalytic amount
of DMAP to protect the (possible) unreacted hydroxyls on the
GlcA units (!10i–iii). Then, the oxidative cleavage of both
benzylidene and Bn protecting groups with NaBrO3 and
Na2S2O3 in a H2O/ethyl acetate mixture[22] furnished derivatives
11 i–iii with two or three free hydroxyls on the Fuc branches as
well as at either 4-O- or 6-O-position of the GalNAc units. Sulfa-
tion of the free alcohol moieties with the SO3·pyridine complex
in DMF at 508C and subsequent hydrolysis of Ac, Bz, and
methyl ester groups under alkaline aqueous conditions afford-
ed fCS-i–iii. By skipping the sulfation reaction on 11 i, unsulfat-
ed fucosylated chondroitin fC-i was obtained.
A slightly different sequence of steps allowed the transfor-
mation of polysaccharide acceptor 8 into fC-ii and fCS-iv–vi.
Fucosylation of 8 with thioglycosides 2–4 was conducted
under almost the same conditions indicated above for the gly-
cosylation of polysaccharide 1. Only the reaction temperature
was changed, by increasing it to room temperature, because
no acid-cleavable benzylidene groups had to be preserved.
The obtained derivatives 12i–iii were then subjected to Bn oxi-
dative cleavage (!13i–iii) followed by sulfation of the depro-
tected hydroxyls on Fuc and final ester hydrolysis to afford
fCS-iv–vi. By skipping the sulfation reaction on 13i, fC-ii was
obtained.
1H NMR spectra of semi-synthetic fC-i,ii and fCS-i–vi polysac-
charides (see Figure 2 and the Supporting Information) allowed
the determination of their degree of fucosylation (DF) and the
a/b stereochemical ratio of the Fuc glycosidic bonds. DF was
evaluated as the ratio between the H-6 Fuc methyl (d=1.21–
1.35 ppm) and the acetyl signal (d=2.01–2.10 ppm) integra-
tions. No significant variations among the fC and fCS polysac-
charides was observed, as the DF value ranged from 0.34 to
0.51 (Table 1). Interestingly, no differences could be detected
between polysaccharides branched at the GlcA and GalNAc
positions (fC-i, fCS-i–iii and fC-ii, fCS-iv–vi, respectively), al-
Scheme 3. Synthesis of 8. (a) Ac2O, Et3N, DMAP, CH3CN, rt, overnight
(DS=2.00); (b) NaBrO3, Na2S2O4, H2O/ethyl acetate, rt, overnight, 63% over
two steps (DS=0.97).
tected by acetylation with Ac2O and Et3N in CH3CN in the pres-
ence of a catalytic amount of 4-dimethylaminopyridine (DMAP)
to afford derivative 7 with a quantitative degree of substitution
(DS=2.00), as calculated by integration of the H-2 and H-3 car-
binolic signals (d=4.65 and 5.04 ppm, respectively) with re-
spect to benzylidene methine one (d=5.44 ppm) in the
1H NMR spectrum (see also the Supporting Information). The
benzylidene ring of derivative 7 was then cleaved under oxida-
tive conditions (NaBrO3 and Na2S2O3 in a H2O/ethyl acetate
mixture)[21] to afford polysaccharide acceptor 8 in 63% yield
(over two steps from 1)[28] with either 4-O- or 6-O-benzoylated
GalNAc units randomly distributed along the polymer chain.
The almost quantitative DS (0.97) of the benzylidene cleavage
1
reaction was evaluated by H NMR integration of the ortho-H
Bz aromatic signals (d=7.86 and 7.95 ppm) with respect to the
Ac ones (d=1.72–1.98 ppm). The 4-O-benzoylated-/6-O-ben-
zoylated-GalNAc ratio was estimated to be 1.3 by HSQC DEPT
2D NMR integration of the 4-O-Bz-GalNAc and 6-O-BzGalNAc 6-
O-methylene signals (dH/C =3.29/59.3 and 4.21–4.34/62.9 ppm,
respectively; see also the Supporting Information), assuming
1
that the signals to be compared displayed similar JCH coupling
constants and that a difference of around 5–8 Hz from the ex-
perimental set value did not cause a substantial variation in
the integrated peak volumes.[29]
With fucosyl donors 2–4 and polysaccharide acceptors 1 and
8 in hand, the reaction conditions for glycosylations were set:
N-iodosuccinimide (NIS)/TMSOTf was chosen as the typical thi-
oglycoside activator system and 5:3 v/v CH2Cl2/DMF as solvent
mixture, because DMF was shown to act as an a-stereodirect-
ing modulator in glycosylations with 2-O-benzylated thioglyco-
side donors,[30] including the first 1,2-cis-glycosylation of a poly-
saccharide acceptor.[16] The reaction temperature was set at
ꢀ208C to make the benzylidene acid cleavage rate slow
enough to avoid Fuc branching also on the GalNAc units. The
glycosylation of 1 with thioglycosides 2–4 was run under ho-
mogeneous conditions and gave derivatives 9i–iii, respectively
(Scheme 4). The complexity of their 1H NMR spectra and the
coprecipitation of part of the Fuc byproducts deriving from 2–
4 with the polysaccharides impeded any DS evaluation of the
1
Figure 2. Magnified region of H and HSQC DEPT NMR spectra of fC-
i (600 MHz, D2O, 298 K).
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
4
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
Scheme 4. Synthesis of fSC-i–vi and fC-i,ii. (a) 2 (!9-i/12-i) or 3 (!9-ii/12-ii) or 4 (!9-iii/12-iii), NIS, TMSOTf, 5:3 v/v CH2Cl2/DMF, AW-300 4 ꢁ-MS, ꢀ208C (for
1) or rt (for 8), 4 h; (b) Ac2O, Et3N, DMAP, CH3CN, rt, overnight; (c) NaBrO3, Na2S2O4, H2O/ethyl acetate, rt, overnight; (d) i) SO3·pyridine, DMF, 508C, overnight
(skipped for fC-i,ii: see text); ii) NaOH, H2O, rt, 6 h.
Table 1. Yield and structural data of fC-i,ii and fCS-i–vi.
[f]
[g]
Yield[a] [%]
DF[b]
a/b ratio[c]
Branching site ratio[d]
Degree of sulfation[e]
Mw [kDa]
Mw/Mn
fC-i
33
39
41
41
0.34
0.38
0.51
0.43
5.2
3.5
3.6
5.5[j]
0.9[h]
0.9[i]
–
–
7.3
6.9
10.9
9.4
1.35
1.35
1.50
1.49
fC-ii
fCS-i
fCS-ii
0.9[h]
0.6[h,j]
GalNAc O-4=0.32; GalNAc O-6=0.63
GalNAc O-4=0.31; GalNAc O-6=0.62
GlcA O-2=0.06: GlcA O-3=0.07
GalNAc O-4=0.28; GalNAc O-6=0.56
GlcA O-2=0.13: GlcA O-3=0.19
GalNAc O-4=0.25; GalNAc O-6=0.33
GalNAc O-4=0.25; GalNAc O-6=0.36
GalNAc O-4=0.14; GalNAc O-6=0.39
fCS-iii
17
0.44
2.8
1.1[h]
8.3
1.58
fCS-iv
fCS-v
fCS-vi
47
38
40
0.45
0.43
0.50
4.3
4.0
1.4
1.3[i]
2.1[i]
2.4[i]
9.4
8.7
9.9
1.56
1.45
1.44
[a] Overall mass yield determined with respect to starting glycosyl acceptor (1 or 8). [b] Determined by 1H NMR integration of Fuc methyl and GalNAc
acetyl signals. [c] Estimated by 1H NMR integration according to Eq. (1), except where indicated differently. [d] Estimated by 1H NMR integration, except
where indicated differently. [e] Estimated by HSQC DEPT integration (see Figures 4 and 5 and the Supporting Information). [f] Mw =weight-averaged molec-
ular mass. [g] Mn =number-averaged molecular mass. [h] [a-Fuc-(1!3)-GlcA]/[a-Fuc(1!2)-GlcA] branching ratio. [i] [a-Fuc-(1!6)-GalNAc]/[a-Fuc(1!4)-
GalNAc] branching ratio. [j] Determined by HSQC DEPT integration, owing to signals overlapping.
though fucosylation reactions were conducted at different
temperatures (ꢀ208C and room temperature, respectively).
The a/b ratio of Fuc branching could not be determined by in-
tegration of the a- and b-Fuc anomeric signals, because the
latter could be not unambiguously detected in NMR spectra
owing to low intensity and overlap with the signals of the a-
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
5
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
Fuc[31] or GlcA and GalNAc units. Therefore, an indirect estima-
tion of the a/b ratio was done by evaluating the b-Fuc branch-
es amount as the difference between the H-6 and H-1 a-Fuc
signal integrations—as the former counts for both a- and b-
linked Fuc units—according to Equation (1):
for the assignment of Fuc branching sites was repeated for fC-
ii as well as for all the semi-synthesized fCSs (see Figure 3 and
the Supporting Information). Then, the integration of the
1
anomeric a-Fuc signals in the H NMR spectra allowed an eval-
uation of the branching site regiochemistry. Polysaccharides
obtained by fucosylation of acceptor 1 (fC-i and fCS-i–iii)
showed no preference for the GlcA O-2 or O-3 position, with
the branching site ratio close to 1 in all cases (Table 1). A slight
preference for fucosylation at O-6 with respect to O-4 of
GalNAc units was observed for polysaccharides obtained by
glycosylation of acceptor 8 (possessing a 6-OH/4-OH ratio of
1.3: see above), but limited only to fCS-v,vi.
IðCH ꢀ 1aFucÞ
a=b ¼
ð1Þ
IðCHꢀ6FucÞ ꢀ IðCH ꢀ 1aFucÞ
3
The a/b ratio was found to be rather variable, but with the
a-stereochemistry clearly predominant in all cases, as expected
from the a-stereodirecting effect of DMF in fucosylations.[16,30]
It is worth noting that no evident increase in a/b ratio could
be observed for polysaccharides deriving from fucosylations
conducted with donors 3 and 4 (to give fCS-ii–iii and fCS-v–vi,
respectively), in spite of the Bz groups at position O-4 and O-3,
respectively, which are believed to enhance the a-stereoselec-
tivity of fucosylation by remote participation.[32] Conversely, the
lowest a/b ratios were observed for fCS-iii and fCS-vi, both de-
riving from fucosylations with 3-O-benzoylated donor 4.
In the case of fCS polysaccharides, NMR analysis also allowed
us to confirm the Fuc sulfation patterns expected from the po-
sition of orthogonally cleavable (Bn) and permanent (Bz) pro-
tecting groups on fucosyl donors 2, 3, and 4 used in the glyco-
sylations. For example, the cross peaks related to anomeric a-
Fuc signals in the COSY, TOCSY, and NOESY spectra confirmed
the differences in the sulfation pattern of Fuc units a-linked to
the O-4 or O-6 atoms of GalNAc residues in fCS-iv–vi. Indeed,
a-Fuc branches of fCS-iv (obtained from fucosylation reaction
with perbenzylated thioglycoside 2) displayed H-2, H-3, and H-
4 signals all downfield shifted for O-sulfation (d=4.52, 4.70–
4.74, and 4.95 ppm, respectively), whereas only the H-2 and H-
3 signals showed a similar shift (d=4.53, 4.67–4.69 ppm, re-
spectively; d=4.22 for H-4) in fCS-v (deriving from 2,3-di-O-
benzylated donor 3), as well as H-2 and H-4 ones in fCS-vi (d=
4.41, 4.69 ppm, respectively; d=4.07–4.20 for H-3), which was
obtained by glycosylation with 2,4-di-O-benzylated thioglyco-
side 4 (see Figure 3). The same NMR analysis was applied to
confirm the differences in the a-Fuc sulfation patterns of fCS-
i–iii (see the Supporting Information). 2D NMR analysis also al-
lowed us to define the sulfation pattern in residues other than
the Fuc branches. For example, the HSQC DEPT spectrum of
A 2D NMR (COSY, TOCSY, NOESY, HSQC DEPT, HSQC TOCSY)
analysis of the semi-synthetic polysaccharides was then con-
ducted to investigate the regiochemistry of the Fuc branches
and, limited to fCS products, the sulfate group distribution.
The HSQC DEPT spectrum of fC-i (Figure 2) showed at the
1
highest H chemical shift values the presence of two densities
(dH/C =5.38/99.5 and 5.25/100.4 ppm), which could be associat-
ed with the anomeric CH of two a-linked Fuc units. The analy-
sis of the cross peaks related to these signals in the NOESY
spectrum showed a correlation of the more downfield shifted
signal (dH =5.38 ppm) with a density at dH =3.70 ppm, attribut-
able to GlcA H-3 by means of the other 2D NMR spectra and
literature data.[16,33] Similarly, the other H-1 a-Fuc signal (dH =
5.25 ppm) could be assigned to a branching unit linked at O-2
(dH/C =3.50/78.6 ppm) of some GlcA units. The same approach
fCS-iv (Figure 4) displayed signals for both sulfated (dH/C
=
Figure 3. Magnified region of TOCSY (black) and NOESY (grey) NMR spectra (600 MHz, D2O, 298 K) of fCS polysaccharides derived from acceptor 8.
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
6
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
1
Figure 4. Magnified region of H and HSQC DEPT NMR spectra (600 MHz, D2O, 298 K) of fCS-iv (densities enclosed in the dotted circles were integrated for
GalNAc O-4 and O-6 sulfation degree estimation).
4.22/68.8 ppm) and non-sulfated (fucosylated or not: dH/C
=
i–iii) display a cumulative value of sulfation degree on GalNAc
units that is very close to 1. This is indicative of an A,C sulfa-
tion pattern—as expected from the non-regioselective oxida-
tive opening of the benzylidene ring before the sulfation
step[18e] (Scheme 4, 10-i–iii!11-i–iii)—with a very low amount
of unsulfated GalNAc units along the polysaccharide chain.
Lower cumulative values for sulfation degree on the GalNAc
units were obtained for fCS-iv–vi (Table 1). Indeed, in these
cases, sulfate groups could be inserted only on those GalNAc
residues that were not fucosylated in the glycosylation step, as
ascertained by values close to 1 for the sum of DF and GalNAc
O-6 and O-4 sulfation degrees.[35]
3.90/69.4 or 3.71–3.85/62.5 ppm, respectively) CH2 at position
6, as well as sulfated (dH/C =4.70/76.1 and 4.74/77.7 ppm) and
non-sulfated (fucosylated or not: dH/C =4.29/73.8 or 4.17/
68.8 ppm, respectively) CH at position 4 of the GalNAc
units[18e,34] (see also the Supporting Information). This was in
agreement with the non-regioselective distribution of free hy-
droxyls at either position 4 or 6 of the GalNAc units on poly-
saccharide acceptor 8 together with their non-quantitative fu-
cosylation (8!12-i, Scheme 4) before the sulfation step (13-
i!fCS-iv), as already discussed above (DF=0.45, Table 1). An
estimation of sulfation degree at positions O-4 and O-6 was
possible by HSQC DEPT density integration. Owing to the par-
tial overlap of the densities related to the CH2 at position 6 of
GalNAc6S and CH at position 4 of unsulfated and non-fucosy-
lated GalNAc, the integrations were made with respect to the
CH-2 density of GalNAc unit (dH/C =4.01/52.2 ppm). In particu-
lar, the relative integration of GalNAc4S CH-4 densities with re-
spect to the CH-2 one gave a 0.25 degree of sulfation at posi-
tion O-4. A value of 0.33 was estimated for position O-6, by in-
tegrating the densities related to unsulfated GalNAc units (6-O-
fucosylated or not) with respect to the CH-2 one and then ap-
plying Equation (2):
A careful investigation of the 2D NMR spectra of the semi-
synthetic library allowed us to detect a small amount of sulfate
groups also on the GlcA units in the case of fCS-ii and fCS-iii
polysaccharides, as demonstrated by the presence of charac-
teristic signals for O-2 and O-3 sulfation at dH/C =4.11/80.5[29a]
and 4.35/82.5 ppm,[18d] respectively. This could be ascribed to
the non-quantitative protection of GlcA hydroxyls in the acety-
lation step (9-ii,iii!10-ii,iii) that left some free alcohol moiet-
ies ready to be sulfated at the end of the semi-synthesis
(Scheme 4). Similarly to the GalNAc case, the GlcA sulfation de-
grees at O-2 and O-3 positions (Table 1) could be estimated by
integration of these densities with respect to the GalNAc CH-2
one in the HSQC DEPT spectrum (Figure 5 and the Supporting
Information). A comprehensive list of NMR assignment data for
all the semi-synthesized polysaccharides is reported in the Sup-
porting Information. These could be useful as standard data
for researchers dealing with the structural characterization of
novel natural fCSs, which could show a Fuc sulfation pattern
and/or branching site different from those already reported in
the literature.
O-6 degree of sulfation :
I½ðCH2 ꢀ 6GalNAc ꢀ ð6 ꢀ OHÞꢁ þ I½CH2 ꢀ 6GalNAcð6 ꢀ FucÞꢁ
1 ꢀ
2 ꢂ IðCH ꢀ 2GalNAcÞ
ð2Þ
A similar approach was repeated for the other fCS polysac-
charides of the semi-synthetic library. Data for 4- and 6-O-sulfa-
tion at GalNAc sites are reported in Table 1. It is worth noting
that fCS polysaccharides with Fuc branches on GlcA units (fCS-
The structural characterization of the semi-synthetic fC and
fCS polysaccharides was completed with the determination of
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
7
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
1
Figure 5. Magnified region of H and HSQC DEPT NMR spectra (600 MHz, D2O, 298 K) of fCS-iii (densities enclosed in the dotted circles were integrated for
GalNAc O-4 and O-6 as well as GlcA O-2 and O-3 sulfation degree estimation).
their molecular mass by high-performance size exclusion chro-
matography combined with a triple detector array (HP-SEC-
TDA).[36] All polysaccharides showed a weight-averaged molec-
ular mass (Mw) that was much lower (7–10 kDa, Table 1) with
respect to natural fCSs (Mw ꢃ55–65 kDa).[14d] This is presumably
due to acid-mediated reactions (carboxylic acid esterification[16]
and glycosylation) in the semi-synthetic sequences, which
shorten the polysaccharide chain of the starting microbial
chondroitin (Mw =38 kDa), although the polydispersity (Mw/Mn)
was found to be only poorly increased (1.35–1.58 for the semi-
synthetic polysaccharides; 1.34 for starting chondroitin). The fC
species showed slightly lower Mw values with respect to fCS
polysaccharides, owing to the absence of sulfate groups on
the polymer chain. Interestingly, fCS derivatives with a Mw of
8–12 kDa are known to retain the anticoagulant activity of the
longer, native polysaccharides, while minimizing the undesired
effects, such as platelet aggregation and bleeding exhibited by
the latter.[14]
cies obtained by partial depolymerization of natural polysac-
charides.[5,14d,33] A behavior similar to natural species was found
also for the reduction of HC-II mediated factor IIa activity.[14d]
Indeed, data were very close to heparin for fCSs with Fuc
branches on the GlcA units (fCS-i–iii), regardless of their sulfa-
tion pattern (Figure 6), whereas two of the three fCSs with Fuc
branches on the GalNAc units (fCS-v–vi), as well as unsulfated
polysaccharides (fC-i–ii), displayed a much reduced anticoagu-
lant activity. These preliminary data encourage a much more
detailed screening of the anticoagulant and antithrombotic ef-
fects of the semi-synthesized polysaccharides. This work is cur-
rently underway, and the results will be published in due
course.
Two very preliminary assays of the anticoagulant activity of
the semi-synthetic polysaccharides were made by evaluating
their AT-dependent activity against factor Xa and HC-II-mediat-
ed anti-factor IIa activity. Approximately 600–1000-fold lower
values with respect to heparin were found for anti-Xa activity
(Table 2), as already reported for low-molecular-mass fCS spe-
Figure 6. HC-II-mediated anti-factor IIa activity.
Table 2. AT-dependent activity against factor Xa.
fC-i
fC-ii
fCS-i
fCS-ii
fCS-iii
Activity [IUmgꢀ1
Activity [IUmgꢀ1
]
]
0.20
0.25
0.18
0.27
n.d.[a]
Conclusion
A library of six fCS polysaccharides with different Fuc sulfation
patterns and/or branch positions, and two fCs with different
Fuc linkage sites, was produced from microbial sourced chon-
droitin through semi-synthetic strategies based on the modu-
fCS-iv
fCS-v
0.33
fCS-vi
0.23
heparin
198
0.26
[a] Not determined.
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
8
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
tographies were performed with a Bio-Gel P2 column (0.75ꢂ
67.5 cm, Bio-Rad) by using 50 mm ammonium bicarbonate as
a buffer at a flow rate of 0.2 mLminꢀ1. The column eluates were
monitored continuously with a Knauer K-2310 refractive index re-
fractometer. Freeze-dryings were performed with a 5 Pascal Lio 5P
4 K freeze-dryer. Optical rotations were measured with a JASCO P-
1010 polarimeter. Elemental analyses were performed with a Carlo
Erba 1108 instrument. Positive MALDI-MS spectra were recorded
with a Applied Biosystem Voyager DE-PRO MALDI-TOF mass spec-
trometer in positive mode: compounds were dissolved in CH3CN at
a concentration of 0.1 mgmLꢀ1 and one microliter of these solu-
tions was mixed with one microliter of a 20 mgmLꢀ1 solution of
2,5-dihydroxybenzoic acid in 7:3 v/v CH3CN/H2O. NMR spectra were
recorded with a Bruker DRX-400 (1H: 400 MHz, 13C: 100 MHz) in-
strument or with a Bruker DRX-600 (1H: 600 MHz, 13C: 150 MHz) in-
strument equipped with a cryo probe, in D2O (acetone as internal
standard, 1H: (CH3)2CO at d=2.22 ppm; 13C: (CH3)2CO at d=
31.5 ppm) or [D6]DMSO (1H: CHD2SOCD3 at d=2.49 ppm; 13C:
CD3SOCD3 at d=39.5 ppm) or CDCl3 (1H: CHCl3 at d=7.26 ppm;
13C: CDCl3 at d=77.0 ppm). Gradient-selected COSY, phase-sensi-
tive NOESY, and TOCSY experiments were performed by using
spectral widths of 6000 Hz in both dimensions, and by using data
sets of 4096ꢂ256 points. Quadrature indirect dimensions were
achieved through the States-TPPI (time proportional phase incre-
mentation) method; spectra were processed by applying an un-
shifted Qsine function to both dimensions and the data matrix was
zero-filled by a factor of 2 before Fourier transformation. TOCSY
and NOESY mixing times were set to 120 and 200 ms, respectively.
lar use of five different reactions: glycosylation, acetylation, se-
lective cleavage of orthogonal protecting groups (Bn and ben-
zylidene), sulfation, and global deprotection. To the best of our
knowledge, this is the first example of the obtainment of a li-
brary of chemically modified polysaccharides by using a O-gly-
cosylation as the key step, a reaction that is very well known
in the field of oligosaccharide synthesis[37] but is almost unex-
plored for polysaccharides,[38] especially in the case of branch-
ing at the secondary, less reactive hydroxyl positions. The
structures of the semi-synthetic fCS and fC polysaccharides
were fully characterized by 2D NMR spectroscopy, confirming
in all cases the fucosylation and sulfation patterns expected
from the designed reaction sequences. A single, well-defined
sulfation pattern was found for Fuc branches in all the semi-
synthesized fCS polysaccharides, contrary to the natural ones
studied up to now, which display two or even more different
sulfation patterns within the same polymer chain. The ob-
tained NMR data could be useful for structural characterization
purposes, as they can be employed as standards for compari-
son with data obtained from novel natural fCSs showing a Fuc
sulfation pattern and/or branching site different from the
known ones. More importantly, this library of semi-synthetic
polysaccharides will allow a wider and more accurate SAR
study of fCSs with respect to those performed up to now that
have been limited to only the natural, heterogeneous struc-
tures found in sea cucumbers. By HP-SEC-TDA analysis, a lower
molecular mass was found for all the obtained polysaccharides
with respect to the natural ones. This is encouraging for avoid-
ing the adverse effects displayed by the latter, while at the
same retaining the anticoagulant activity. To confirm this,
a careful screening of the biological activities of these semi-
synthetic fCS and fC polysaccharides has been launched. The
results, which will be published in due course, will drive further
chemical efforts towards the obtainment of fCS polysaccha-
rides with an even more well-defined structure in terms of Fuc
site branching. In particular, the development of further regio-
selective protecting steps for the differentiation of the hydrox-
yl groups in GlcA 2,3- and GalNAc 4,6-diols will be pursued, as
well as the polysaccharide branching with more complex gly-
cosyl donors such as Fuc disaccharide ones. The final aim will
be the full development of an effective, safe, non-animal
sourced antithrombotic drug candidate.
1
HSQC DEPT experiments were measured in the H-detected mode
by single quantum coherence with proton decoupling in the 13C
domain, by using data sets of 2048ꢂ256 points and typically 32 in-
crements. As for HSQC TOCSY, data sets of 2048ꢂ256 points were
used, with 100 increments, the mixing time was set to 100 ms and
spectra were transformed as indicated for homonuclear spectra,
data matrix were doubled, and linear prediction applied the new
points by using 30 coefficients, Qsine functions were used as
window functions, and the values selected depended by the best
resolution obtained. A ViscotekTM instrument (Malvern) was used to
determine molecular mass data.
Ethyl 3-O-benzoyl-2,4-di-O-benzyl-1-thio-b-l-fucopyranoside
(4)
A solution of triol 5[25] (1.966 g, 9.452 mmol) in DMF (16 mL) was
cooled to 08C and treated with imidazole (1.608 g, 23.63 mmol)
and
tert-butyldimethylsilyl
chloride
(TBDMSCl;
1.952 g,
12.95 mmol). After 10 min of stirring at 08C, the temperature was
raised to room temperature and stirring was continued for an addi-
tional 3 h. The solution was then diluted with ethyl acetate
(100 mL) and washed with water (100 mL). The organic phase was
collected, dried over anhydrous Na2SO4, filtered, and concentrated.
The obtained residue was purified by column chromatography
(10:1 to 6:1 v/v n-hexane/ethyl acetate) to afford pure ethyl 3-O-
tert-butyldimethylsilyl-1-thio-b-l-fucopyranoside (1.844 g, 61%),
Experimental Section
General methods
Commercial grade reagents and solvents were used without fur-
ther purification, except where indicated differently. The term
“pure water” refers to water purified by a Millipore Milli-Q Gradient
system. Centrifugations were performed with an Eppendorf Centri-
fuge 5804R instrument at 48C (4600 g, 10 min). Dialyses were con-
ducted with Spectra/Por 3.5 kDa cut-off membranes at 48C. Analyt-
ical thin layer chromatographies (TLCs) were performed on alumi-
num plates precoated with Merck Silica Gel 60 F254 as the adsorb-
ent. The plates were developed with 10% H2SO4 ethanolic solution
and then heating to 1308C. Flash column chromatographies were
performed with Kieselgel 60 (63–200 mesh). Size-exclusion chroma-
1
which showed H and 13C NMR data fully in line with those already
reported for its d-enantiomer.[27] A solution of the product (1.796 g,
5.580 mmol) in DMF (16 mL) was cooled to 08C and treated with
benzyl bromide (1.66 mL, 13.4 mmol) and then NaH (60% disper-
sion in mineral oil, 321 mg, 13.4 mmol). After 10 min of stirring at
08C, the temperature was raised to room temperature and stirring
was continued for another 1h. The reaction was then quenched by
careful addition of water (50 mL) at 08C and in small aliquots, and
then of CH2Cl2 (100 mL). The organic layer was collected, dried
over anhydrous Na2SO4, filtered, and concentrated to give a residue
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
9
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
that was subjected to column chromatography (12:1 to 8:1 v/v n-
hexane/ethyl acetate) to afford pure ethyl 2,4-di-O-benzyl-3-O-tert-
butyldimethylsilyl-1-thio-b-l-fucopyranoside (2.605 g, 93%), which
showed 1H and 13C NMR data fully in line with those already report-
ed for its d-enantiomer.[27] The product (2.570 g, 5.120 mmol) was
then dissolved in THF (37 mL) and treated with a 1.0m solution of
tetra-n-butylammonium fluoride (TBAF) in THF (12.8 mL). After 2 h
of stirring at room temperature, the solution was concentrated
and the obtained residue was subjected to column chromatogra-
phy (10:1 to 6:1 v/v n-hexane/ethyl acetate) to afford pure ethyl
2,4-di-O-benzyl-1-thio-b-l-fucopyranoside 6 (1.894 g, 94%), which
showed 1H and 13C NMR data fully in line with those already report-
ed for its d-enantiomer.[27] A solution of 6 (840 mg, 2.17 mmol) in
1:1 v/v CH2Cl2/pyridine (18.7 mL) was treated with benzoyl chloride
(377 mL, 3.25 mmol). After 3 h of stirring at room temperature, the
reaction was quenched by adding CH3OH (2.0 mL). The solution
was diluted with CH2Cl2 (40 mL) and washed with 0.1m HCl
(40 mL). The organic layer was dried over anhydrous Na2SO4, fil-
tered, and concentrated to give a residue that was subjected to
column chromatography (12:1 to 8:1 v/v n-hexane/ethyl acetate)
to afford pure 4 (857 mg, 80%) as a white powder. Rf =0.6 (3:1 v/v
hexane/ethyl acetate); [a]D25 =ꢀ89.9 (c=1.0 in CH2Cl2); 1H NMR
(400 MHz, CDCl3): d=8.04–7.18 (m, 15H, H-Ar), 5.25 (dd, J=9.7,
3.1 Hz, 1H, H-3), 4.90 (d, J=10.6 Hz, 1H, OCHHPh), 4.72 (d, J=
11.7 Hz, 1H, OCHHPh), 4.69 (d, J=11.7 Hz, 1H, OCHHPh), 4.59 (d,
J=10.6 Hz, 1H, OCHHPh), 4.55 (d, J=9.6 Hz, 1H, H-1), 4.01 (t, J=
9.7 Hz, 1H, H-2), 3.89 (d, J=3.1 Hz, 1H, H-4), 3.74 (quartet, J=
6.4 Hz, 1H, H-5), 2.81 (m, 2H, SCH2CH3), 1.35 (t, J=7.4 Hz, 3H,
SCH2CH3), 1.29 ppm (d, J=6.4 Hz, 3H, H-6); 13C NMR (100 MHz,
CDCl3): d=165.9 (CO), 137.9–127.6 (C-Ar), 85.0 (C-1), 78.0, 77.3,
76.4, 75.5, 75.2, 74.3 (C-2, C-3, C-4, C-5, 2OCH2Ph), 24.9 (SCH2CH3),
16.9 (C-6), 14.9 ppm (SCH2CH3); MALDI-TOF MS for C29H32O5S (m/z):
Mr (calcd): 492.20; Mr (found): 514.13 (M+Na)+; elemental analysis
calcd (%) for C29H32O5S: C 70.70, H 6.55, S 6.51; found: C 70.45, H
6.68, S 6.41.
and then DMF (3.2 mL) and CH2Cl2 (5.2 mL), which were freshly
dried over 4 ꢁ molecular sieves, were added to the mixture under
an argon atmosphere. The mixture was stirred at ꢀ208C (or at
room temperature for 8) for 10 min, and then treated with NIS
(180 mg, 0.799 mmol) and a 0.45m solution of TMSOTf in freshly
dried CH2Cl2 (460 mL, 0.207 mmol). After 4 h of stirring under an
argon atmosphere at ꢀ208C (or room temperature for 8), the mo-
lecular sieves were removed by decantation. A few drops of trie-
thylamine and then 1:1 v/v hexane/ethyl acetate (40 mL) were
added, and the mixture stored at ꢀ288C overnight. The obtained
yellowish precipitate was collected by centrifugation and then
dried under vacuum overnight to afford a mixture containing de-
rivative 9i (99.5 mg).
Example of acetylation reaction
A suspension of derivative 9i (96.1 mg) in CH3CN (2.1 mL) was
treated with triethylamine (160 mL), acetic anhydride (710 mL), and
finally with DMAP (4.8 mg). After overnight stirring at room tem-
perature, the resulting brownish solution was concentrated to give
a residue that was coevaporated several times with toluene. A
brownish oil (165 mg) was obtained. It was not further purified,
but used in the following synthetic step as it was.
Example of cleavage of benzyl (and benzylidene) protecting
groups
Derivative 10i (163 mg) was treated with ethyl acetate (6.7 mL)
and the resulting suspension was treated with a 0.27m solution of
NaBrO3 in pure water (6.7 mL). A 0.24m solution of Na2S2O4 in pure
water (6.3 mL) was added portionwise over a period of 10 min. The
triphasic mixture was vigorously stirred at room temperature over-
night under visible-light irradiation. The yellowish solid was then
collected by centrifugation. It was dried under vacuum overnight
to afford derivative 11 i (37.5 mg).
Chondroitin derivative (7)
Example of sulfation and deprotection reactions
Polysaccharide 1[16] (202 mg, 0.419 mmol) was treated with CH3CN
(4.4 mL) and then with Et3N (1.2 mL), acetic anhydride (2.8 mL), and
DMAP (18.8 mg, 0.168 mmol). After overnight stirring at room tem-
perature, a clear solution was obtained. Diisopropyl ether (15 mL)
was then added to give a yellowish precipitate that was collected
by centrifugation and then dried under vacuum overnight to
afford derivative 7 (227 mg), which was contaminated mostly by
Et3N and DMAP.
A suspension of derivative 11 i (26.9 mg) in DMF (1.7 mL), which
was freshly dried over 4 ꢁ molecular sieves, was treated with
a 1.17m solution of pyridine/sulfur trioxide complex in freshly
dried DMF (0.6 mL). After overnight stirring at 508C, the obtained
clear solution was cooled to room temperature and then treated
with a saturated NaCl solution in acetone (20 mL). The obtained
yellowish precipitate was collected by centrifugation and then dis-
solved in pure water (8.0 mL). A 15% w/v NaOH solution was then
added to adjust the pH to 13. The solution was stirred at room
temperature for 6 h, then it was neutralized by treatment with 1m
HCl. Dialysis and subsequent freeze-drying yielded a slightly yellow
solid, which was further purified by filtration through a Sep-pak C-
18 cartridge and then by size-exclusion chromatography. Freeze-
drying of the fractions afforded fCS-i (14.5 mg) as a white waxy
solid.
Chondroitin derivative (8)
Derivative 7 (96.5 mg) was suspended in ethyl acetate (1.9 mL) and
then treated with a 0.27m solution of NaBrO3 in pure water
(1.9 mL). A 0.24m solution of Na2S2O4 in pure water (1.9 mL) was
added portionwise over a period of 10 min. The mixture was vigo-
rously stirred at room temperature overnight under visible-light ir-
radiation. The yellowish solid was collected by centrifugation and
then dried under vacuum overnight to give pure polysaccharide 8
(66.9 mg, 63% over two steps from 1, DSacetylation =2.0,
DSdebenzylidenation =0.97).
Determination of molecular mass
Hydrodynamic characterization of fC and fCS samples was per-
formed by using the SEC-TDA equipment by Viscotek (Malvern,
Italy). The chromatographic system consists of two modules: (i) a
GPCmax VE 2001 integrated system composed of a specific pump
for gel permeation chromatography, an in-line solvent degasser,
and an autosampler; and (ii) a TDA305 module (triple detector
array) that includes a column oven and a triple detector. The latter
is the key element of the GPC-Viscotek as it is equipped with a re-
Example of glycosylation reaction
A mixture of polysaccharide acceptor 1 (51.5 mg, 0.107 mmol) and
fucosyl donor 2 (256 mg, 0.535 mmol) was coevaporated three
times with dry toluene (2.5 mL each). AW-300 4 ꢁ molecular sieves
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
10
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
[3] G. R. C. Santos, B. F. Glauser, L. A. Parreiras, E. Vilanova, P. A. S. Mour¼o,
Glycobiology 2015, 25, 1043–1052.
[4] a) P. Myron, S. Siddiquee, S. Al Azad, Carbohydr. Polym. 2014, 112, 173–
178; b) S. Chen, C. Xue, L. Yin, Q. Tang, G. Yu, W. Chai, Carbohydr. Polym.
2011, 83, 688–696.
[5] P. A. S. Mour¼o, M. S. Pereira, M. S. Pav¼o, B. Mulloy, D. M. Tollefsen,
M. C. Mowinckel, U. Abildgaard, J. Biol. Chem. 1996, 271, 23973–23984.
[6] B. F. Glauser, M. S. Pereira, R. Q. Monteiro, P. A. S. Mour¼o, Thromb. Hae-
mostasis 2008, 100, 420–428.
[7] R. J. Fonseca, P. A. S. Mour¼o, Thromb. Haemostasis 2006, 96, 822–829.
[8] P. A. S. Mour¼o, Mar. Drugs 2015, 13, 2770–2784.
fractive index (RI) detector, a four-bridge viscosimeter (VIS), and
a light scattering (LS) detector. The latter consists of a right-angle
light scattering (RALS) detector and a low-angle light scattering
(LALS) detector that performed measurements of the scattered
light at 78 with respect to the incident beam with an optimal
signal-to-noise ratio. The OmniSEC software program was used for
the acquisition and analysis of the Viscotek data. Two in series TSK-
GEL GMPWXL columns (Tosoh Bioscience, Cat. No. 8–08025, hy-
droxylated polymethacrylate base material, 100–1000 ꢁ pore size,
13 mm mean particle size, 7.8ꢂ30.0 cm), preceded by a TSK-GEL
guard column GMPWXL (Tosoh Bioscience, Cat. No. 08033, 12 mm
mean particle size, 6.0ꢂ4.0 cm), were used. The samples were ana-
lyzed at concentrations ranging from 0.3 to 4 gLꢀ1 to have
a column load for each analysis (injection volumeꢂsample concen-
trationꢂintrinsic viscosity) of approximately 0.2 dL and, at the
same time, appreciable LALS and VIS signals when analyzing low-
molecular-mass species (see equations below). An isocratic elution
with 0.1m NaNO3 aqueous solution (pH 7.0) at a flow rate of
0.6 mLminꢀ1 was carried out. Analyses were performed at 408C
with a running time of 60 min. The molecular mass and size distri-
bution, polydispersity, hydrodynamic radius, and intrinsic viscosity
were derived. For the detected signals, the following equations
apply: RI signal=K1 ꢂdn/dcꢂC; VIS signal=K2 ꢂ[h]ꢂC; LALS
signal=K3 ꢂMw ꢂ(dn/dc)2 ꢂC; where [h] is the intrinsic viscosity
(dLgꢀ1); C is the mass concentration (mgmLꢀ1); dn/dc is the refrac-
tive index increment (mLgꢀ1); and K1, K2, and K3 are instrumental
constants.[39] By solving the equations above, the system allowed
the simultaneous determination of sample concentration, molecu-
lar mass, and intrinsic viscosity.[36a]
[9] N. Mackman, Nature 2008, 451, 914–918.
[10] A. Greinacher, J. Michels, M. Schꢄfer, V. Keifel, C. Mueller-Eckhardt, Br. J.
Haematol. 1992, 81, 252–254.
[11] T. K. Kishimoto, K. Viswanathan, T. Ganguly, S. Elankumaran, S. Smith, K.
Pelzer, J. C. Lansing, N. Sriranganathan, G. Zhao, Z. Galcheva-Gargova, A.
Al-Hakim, G. S. Bailey, B. Fraser, S. Roy, T. Rogers-Cotrone, L. F. Buhse, M.
Whary, J. Fox, M. Nasr, G. J. Dal Pan, Z. Shriver, R. S. Langer, G. Venkatara-
man, K. F. Austen, J. Woodcock, R. Sasisekharan, New Engl. J. Med. 2008,
358, 2457–2467.
[12] M. Guerrini, D. Beccati, Z. Shriver, A. Naggi, K. Viswanathan, A. Bisio, I.
Capila, J. C. Lansing, S. Guglieri, B. Fraser, A. Al-Hakim, N. S. Gunay, Z.
Zhang, L. Robinson, L. Buhse, M. Nasr, J. Woodcock, R. Langer, G. Venka-
taraman, R. J. Linhardt, B. Casu, G. Torri, R. Sasisekharan, Nat. Biotechnol.
2008, 26, 669–675.
[13] R. J. Fonseca, S. Oliveira, V. H. Pomin, A. S. Mecawi, I. G. Araujo, P. A. S.
Mour¼o, Thromb. Haemostasis 2010, 103, 994–1004.
[14] a) X. Liu, J. Hao, X. Shan, X. Zhang, X. Zhao, Q. Li, X. Wang, C. Cai, G. Li,
G. Yu, Carbohydr. Polym. 2016, 152, 343–350; b) L. Zhao, M. Wu, C.
Xiao, L. Yang, L. Zhou, N. Gao, J. Chen, J. Chen, J. Liu, H. Qin, J. Zhao,
Proc. Natl. Acad. Sci. USA 2015, 112, 8284–8289; c) N. Gao, F. Lu, C. Xiao,
L. Yang, J. Chen, K. Zhou, D. Wen, Z. Li, M. Wu, J. Jiang, G. Liu, J. Zhao,
Carbohydr. Polym. 2015, 127, 427–437; d) M. Wu, D. Wen, N. Gao, C.
Xiao, L. Yang, L. Xu, W. Lian, W. Peng, J. Jiang, J. Zhao, Eur. J. Med.
Chem. 2015, 92, 257–269; e) L. Zhao, S. Lai, R. Huang, M. Wu, N. Gao, L.
Xu, H. Qin, W. Peng, J. Zhao, Carbohydr. Polym. 2013, 98, 1514–1523.
[15] a) N. E. Ustyuzhanina, P. A. Fomitskaya, A. G. Gerbst, A. S. Dmitrenok,
N. E. Nifantiev, Mar. Drugs 2015, 13, 770–787; b) J.-I. Tamura, H. Tanaka,
A. Nakamura, N. Takeda, Tetrahedron Lett. 2013, 54, 3940–3943.
[16] A. Laezza, A. Iadonisi, C. De Castro, M. De Rosa, C. Schiraldi, M. Parrilli, E.
Bedini, Biomacromolecules 2015, 16, 2237–2245.
Anticoagulant activity
AT-dependent antifactor Xa activity was evaluated by a kinetic col-
orimetric method by using a commercial kit (Stachrom Heparin,
Stago). Heparin was used as the standard for building a calibration
curve (r2 =0.976) with 0.1 to 0.5 IUmgꢀ1 active solutions. The
sample (100 mL) was added to purified bovine AT (100 mL) diluted
in Tris EDTA buffer (pH 8.4), and the mixture was incubated with
factor Xa (200 mL) for 2 min at 378C. Then, factor Xa chromogenic
substrate (200 mL) was added. The absorbance was read at 405 nm
against a blank and was inversely proportional to the concentra-
tion of the polysaccharides. The anti HC-IIa assay was performed
by using a sandwich enzyme immunoassay (USCN, Life Science
Inc.) according to the manufacturer’s protocol (http://www.nlbio-
chemex.com/e90284ga.html).
[17] a) P. L. DeAngelis, Appl. Microbiol. Biotechnol. 2012, 94, 295–305; b) C.
Schiraldi, D. Cimini, M. De Rosa, Appl. Microbiol. Biotechnol. 2010, 87,
1209–1220.
[18] a) E. Bedini, A. Laezza, A. Iadonisi, Eur. J. Org. Chem. 2016, 3018–3042;
b) A. Laezza, C. De Castro, M. Parrilli, E. Bedini, Carbohydr. Polym. 2014,
112, 546–555; c) E. Bedini, M. Parrilli, Carbohydr. Res. 2012, 347, 75–85;
d) E. Bedini, C. De Castro, M. De Rosa, A. Di Nola, O. F. Restaino, C. Schir-
aldi, M. Parrilli, Chem. Eur. J. 2012, 18, 2123–2130; e) E. Bedini, C. De Ca-
stro, M. De Rosa, A. Di Nola, A. Iadonisi, O. F. Restaino, C. Schiraldi, M.
Parrilli, Angew. Chem. Int. Ed. 2011, 50, 6160–6163; Angew. Chem. 2011,
123, 6284–6287.
[19] a) C. Schiraldi, A. Alfano, D. Cimini, M. De Rosa, A. Panariello, O. F. Restai-
no, M. De Rosa, Biotechnol. Prog. 2012, 22, 1012–1018; b) D. Cimini,
O. F. Restaino, A. Catapano, M. De Rosa, C. Schiraldi, Appl. Microbiol. Bio-
technol. 2010, 87, 1779–1787.
[20] O. F. Restaino, D. Cimini, M. De Rosa, C. De Castro, M. Parrilli, C. Schiraldi,
Electrophoresis 2009, 30, 3877–3883.
[21] A. Stellavato, V. Tirino, F. de Novellis, A. Della Vecchia, F. Cinquegrani, M.
De Rosa, G. Papaccio, C. Schiraldi, J. Cell. Biochem. 2016, 117, 2158–
2169.
Acknowledgments
Regione Campania (Progetto di Ricerca POR-BIP Biotechnologi-
cal Industrial Processes) and Ministero Italiano dell’Istruzione
dell’Universitꢃ e della Ricerca (MIUR, PON03PE_00060_7) are
acknowledged for financial support.
[22] M. Adinolfi, G. Barone, L. Guariniello, A. Iadonisi, Tetrahedron Lett. 1999,
40, 8439–8441.
[23] S. Chen, G. Li, N. Wu, X. Guo, N. Liao, X. Ye, D. Liu, C. Xue, W. Chai, Bio-
chim. Biophys. Acta Gen. Subj. 2013, 1830, 3054–3066.
Keywords: carbohydrates · glycosylation · polysaccharides ·
semi-synthesis · sulfation
[24] a) E. Bedini, A. Carabellese, D. Comegna, C. De Castro, M. Parrilli, Tetrahe-
dron 2006, 62, 8474–8483; b) D. Comegna, E. Bedini, A. Di Nola, A. Ia-
donisi, M. Parrilli, Carbohydr. Res. 2007, 342, 1021–1029.
[25] H. Lonn, Carbohydr. Res. 1985, 139, 105–113.
[26] R. Fujiwara, S. Horito, Carbohydr. Res. 2011, 346, 2098–2103.
[27] D. Comegna, E. Bedini, M. Parrilli, Tetrahedron 2008, 64, 3381–3391.
[1] V. H. Pomin, Mar. Drugs 2014, 12, 232–254.
[2] a) N. E. Ustyuzhanina, M. I. Bilan, A. S. Dmitrenok, E. A. Tsvetkova, A. S.
Shashkov, V. A. Stonik, N. E. Nifantiev, A. I. Usov, Carbohydr. Polym. 2016,
153, 399–405; b) J. Yang, Y. Wang, T. Jiang, L. Lv, B. Zhang, Z. Lv, Int. J.
Biol. Macromol. 2015, 72, 699–705.
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
11
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ

Full Paper
[28] Molecular masses of disaccharide repeating units of 1 and 8 were used
for molar yield calculations.
[29] a) V. Gargiulo, R. Lanzetta, M. Parrilli, C. De Castro, Glycobiology 2009,
19, 1485–1491; b) M. Guerrini, A. Naggi, S. Guglieri, R. Santarsiero, G.
Torri, Anal. Biochem. 2005, 337, 35–47.
[36] a) A. La Gatta, M. De Rosa, I. Marzaioli, T. Busico, C. Schiraldi, Anal. Bio-
chem. 2010, 404, 21–29; b) S. Bertini, A. Bisio, G. Torri, D. Benzi, M. Ter-
bojevich, Biomacromolecules 2005, 6, 168–173.
[37] X. Zhu, R. R. Schmidt, Angew. Chem. Int. Ed. 2009, 48, 1900–1934;
Angew. Chem. 2009, 121, 1932–1967.
[30] S.-R. Lu, Y.-H. Lai, J.-H. Chen, C.-Y. Liu, K.-K. T. Mong, Angew. Chem. Int.
Ed. 2011, 50, 7315–7320; Angew. Chem. 2011, 123, 7453–7458.
[31] B. Mulloy, P. A. S. Mour¼o, E. Gray, J. Biotechnol. 2000, 77, 123–135.
[32] A. G. Gerbst, N. E. Ustyuzhanina, A. A. Grachev, E. A. Khatuntseva, D. E.
Tsvetkov, D. M. Whitfield, A. Berces, N. E. Nifantiev, J. Carbohydr. Chem.
2001, 20, 821–831.
[38] a) M. Ishimaru, M. Nagatsuka, A. Masubuchi, J. Okazaki, K. Kurita, Polym.
Bull. 2014, 71, 301–313; b) K. Kurita, Y. Matsumura, H. Takahara, K.
Hatta, M. Shimojoh, Biomacromolecules 2011, 12, 2267–2274; c) K.
Kurita, H. Akao, J. Yang, M. Shimojoh, Biomacromolecules 2003, 4,
1264–1268; d) K. Kurita, K. Shimada, Y. Nishiyama, M. Shimojoh, S. Nishi-
mura, Macromolecules 1998, 31, 4764–4769.
[33] C. Panagos, D. Thomson, C. Moss, A. D. Hughes, M. S. Kelly, Y. Liu, W.
Chai, R. Venkatasamy, D. Spina, C. P. Page, J. Hogwood, R. J. Woods, B.
Mulloy, C. Bavington, D. Uhrin, J. Biol. Chem. 2014, 289, 28284–28298.
[34] A. Mucci, L. Schenetti, N. Volpi, Carbohydr. Polym. 2000, 41, 37–45.
[35] Values slightly higher than 1, which are not possible in theory, might
be due to the sums of numbers obtained through different integration
methods (1H NMR for DF, and HSQC DEPT 2D NMR for GalNAc O-6 and
O-4 sulfation degrees).
[39] W. K. Hartmann, N. Saptharishi, X. Y. Yang, G. Mitra, G. Soman, Anal. Bio-
chem. 2004, 325, 227–239.
Received: July 26, 2016
Published online on && &&, 0000
&
&
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
12
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Paper
FULL PAPER
A small library of fucosylated chondroi-
tin sulfate polysaccharides was obtained
from microbial sourced unsulfated
chondroitin by a semi-synthetic modular
approach based on the different combi-
nation of only five reactions. An almost
unprecedented polysaccharide branch-
ing by O-glycosylation was the key step.
Eight polysaccharides with different
fucose and/or sulfation patterns were
obtained and carefully characterized by
2D NMR spectroscopy.
&
Polysaccharide Chemistry
A. Laezza, A. Iadonisi, A. V. A. Pirozzi,
P. Diana, M. De Rosa, C. Schiraldi,
M. Parrilli, E. Bedini*
&& – &&
A Modular Approach to a Library of
Semi-Synthetic Fucosylated
Chondroitin Sulfate Polysaccharides
with Different Sulfation and
Fucosylation Patterns
Chem. Eur. J. 2016, 22, 1 – 13
www.chemeurj.org
13
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
These are not the final page numbers! ÞÞ