Synthesis and evaluation of 5-thio-L-fucose-containing oligosaccharide

Article abstract of DOI:10.1002/chem.200400831

5-Thio-L-fucose-containing trisaccharide H-type II was synthesized. The 3′,4′-O-isopropylidene-2-azido-2-deoxylactoside derivative, which was prepared from lactose by azidonitration of lactal, was used as a starting material. By regio- and stereo-selectiv

Full text of DOI:10.1002/chem.200400831

Synthesis and Evaluation of 5-Thio-l-Fucose-Containing Oligosaccharide
Masayuki Izumi,*[a, d] Osamu Tsuruta,[a] Yasuhiro Kajihara,[b] Shin Yazawa,[c]
Hideya Yuasa,[a] and Hironobu Hashimoto*[a]
Abstract:
5-Thio-l-fucose-containing
containing H-type II 1. Conformational
analysis of the 5-thio-l-fucose-contain-
ing H-type II and the native H-type II
was carried out through NOESY ex-
periments. The observed NOE values
between N-acetylglucosamine and gal-
actose, and galactose and fucose were
same for these two trisaccharides.
However, NOE values between fucose
and N-acetylglucosamine were signifi-
cantly different. Binding of the 5-thio-
l-fucose-containing H-type II to lectins
and antibodies were in some case
stronger and in some case weaker than
those of the native trisaccharide.
trisaccharide H-type II was synthe-
sized. The 3’,4’-O-isopropylidene-2-
azido-2-deoxylactoside
which was prepared from lactose by
azidonitration of lactal, was used as a
starting material. By regio- and stereo-
selective 5-thio-l-fucosylation of the
6,6’-dibenzoate 5 with 5-thiofucosyl tri-
derivative,
Keywords: biological activity
carbohydrates conformation
analysis · lectins · synthesis design
·
·
chloroacetimidate
deprotection gave the 5-thio-l-fucose-
6 and subsequent
Introduction
regulation and other intercellular communication and signal
transduction events.[1] For the study of such recognition
processes, development of biologically stable oligosacchar-
ide ligand is very important. Since natural oligosaccharides
are easily degraded in the blood stream, biologically stable
oligosaccharides are also important for the development of
carbohydrate-based therapeutics and vaccines. In this con-
text, we have been investigating 5-thiosugar-containing oli-
gosaccharides because glycosidic bond of 5-thiosugars are
known to be resistant to enzymatic degradation by glycosi-
dases.[2–5] However, replacing a sugar residue in an oligosac-
charide ligand with a 5-thiosugar would alter its affinity to
its receptor. One of the known effects, which originates di-
rectly from the sulfur atom, is the increase of the affinity by
a stronger hydrophobic interaction compared with the
oxygen atom. This was observed in the binding of 5-thio-l-
fucose to a-l-fucosidase.[6] On the contrary, it could decrease
affinity by destroying hydrogen bond to the ring oxygen,
which was observed in the binding of the 1,6-linked 5-thio-
d-mannose-containing di- and trisaccharide to concanavalin
A.[7] We selected the H-type II trisaccharide 1[8,9] as our
target (Figure 1) because there are a lot of reports not only
on its synthesis but also on its conformation and biological
activity. This trisaccharide was compared with the native oli-
gosaccharide in terms of conformation and affinity with car-
bohydrate binding proteins to evaluate 5-thiosugar-contain-
ing oligosaccharides as molecular probes for biological and
medicinal studies.
It has recently become apparent that carbohydrates play
crucial roles in important biological recognition processes,
including: bacterial and viral infections, cell adhesion in in-
flammation and metastasis, differentiation, development,
[a] Dr. M. Izumi, Dr. O. Tsuruta, Prof. Dr. H. Yuasa,
Prof. Dr. H. Hashimoto
Department of Life Science
Graduate school of Bioscience and Biotechnology
Tokyo Institute of Technology
4259 Nagatsuta, Midori-ku, Yokohama 226-8501 (Japan)
Fax : (+81)45-924-5704
E-mail: masayuki.izumi@aist.go.jp
hhashimo@bio.titech.ac.jp
[b] Prof. Dr. Y. Kajihara
Graduate school of Integrated Science
Yokohama City University, 22-2 Seto
Kanazawa-ku, Yokohama 236-0027 (Japan)
[c] Dr. S. Yazawa
Japan Immunoresearch Laboratories Co. Ltd. (JIMRO)
351-1 Nishiyokote-cho, Takasaki, Gunma 370-0021 (Japan)
[d] Dr. M. Izumi
Current address:
Research Center of Advanced Bionics
National Institute of Advanced Industrial Science and Technology
(AIST)
1-1-1 Higashi, Tsukuba, Ibaraki 305-8565 (Japan)
Fax : (+81)29-861-4680
3032
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200400831
Chem. Eur. J. 2005, 11, 3032 – 3038

FULL PAPER
puter simulation of the H-type II trisaccharide, Lemieux
and co-workers[22] used HSEA calculations and predicted
the torsion angles to be F’=558, Y’=08 and F=508, Y=
158. Imberty and co-workers[23] restudied 14 histo-blood
group oligosaccharides through a combination of molecular
mechanics (MM3) and conformational search (CICADA)
methods. For H-type II, they reported that 95% of conform-
ers have the torsion angles of F’ ꢁ34–478, Y’ ꢁ17–338 and
F ꢁ41–538, Y ꢁ8–178, and 4% of them have F’ ꢁ16–288,
Y’ ꢁꢀ45 to ꢀ448 and F ꢁ29–348, Y ꢁꢀ65 to ꢀ578.
Figure 1. Target 5-thio-l-fucose-containing H-type II 1.
For the synthesis of 5-thiosugar-containing oligosacchar-
ide, glycoside formation is the most important step. This
step has been investigated extensively by using glycosyl tri-
chloroacetimidate of 5-thio-l-fucose[5,10] and 5-thio-d-glu-
cose[11–13] as the donors. In most cases, a 1,2-cis axial glyco-
side was formed predominantly in the presence of an equa-
torial acyloxy group at the 2-position. Glycosylation of the
native (ring oxygen) sugar under similar conditions usually
leads to a formation of a 1,2-trans equatorial glycoside as a
result of the neighboring group participation. Alternative
access to construct 5-thioglycoside is the use of glycosyl-
transferases, although there are only three examples report-
ed to date. Galactosyltransferase could transfer 5-thio-d-gal-
actose[2] and lactose synthetase could transfer N-acetyl-5-
thio-d-galactosamine from their UDP derivatives.[14] Chemo-
enzymatic synthesis of the 5-thio-l-fucose-containing Lewis
X by using GDP-5-thio-l-fucose and a-1,3-fucosyltransfer-
ase has been reported by our group.[15]
Several biological activities of 5-thiosugar-containing oli-
gosaccharides have been reported to date. Methyl 5’-thioiso-
maltoside binds to glucoamylase, contrary to methyl isomal-
toside which binds to a different subsite.[4] Methyl 5’-thio-
maltoside also binds to glucoamylase.[3] Four disaccharides
having 5-thio-l-fucose glycosidically linked to 3-, 4- or 6-po-
sition of N-acetylglucosamine or 2-position of galactose
were synthesized and tested as inhibitors of a-1,2-fucosi-
dase.[5] It turned out that only the a(1!2)Gal-linked disac-
charide was a competitive inhibitor and the other three did
not have any inhibitory activity. The 5-thio-d-mannose-con-
taining oligosaccharides were tested as concanavalin A li-
gands.[7] Five of the structures had a decreased affinity. Sev-
eral lectins and antibodies are known to recognize H-type
II, though each protein has a different mode of recognition.
For example, key hydroxyl groups of H-type II involved in
the binding with Ulex europaeus agglutinin-I are H2, 3, 4 of
fucose, whereas those with Galactia tenuiflora lectin are H3,
4 of galactose and H3 of N-acetylglucosamine, and those
with Psophocarpus tetragonolobus lectin II are H3, 4 of gal-
actose and H2 of fucose.[24]
Herein we report the chemical synthesis of the 5-thio-l-
fucose-containing H-type II trisaccharide from a disacchar-
ide 2-azido-2-deoxy-lactose derivative. Conformational anal-
ysis of the 5-thio-l-fucose-containing H-type II trisaccharide
was carried out through NMR experiments, and the result
was discussed with the aid of molecular mechanics confor-
mational search. Inhibition study of antibody and lectin
binding to native saccharides by 5-thio-l-fucose-containing
oligosaccharides is also reported.
Several possible effects are known to have influence on
the conformation of 5-thiosugar-containing oligosaccharide.
The conformational preferences and associated geometrical
variations have been rationalized by anomeric effects.[16] The
puckering distortion of the 5-thiopyranose ring is suggested
ꢀ
as the result of longer C S bond length (1.78 ꢁ, 1.42 ꢁ for
[17]
ꢀ
C O) and exact C-S-C bond angle (998, 1148 in C-O-C).
Conformational analysis of the 5-thioglucose-containing dis-
accharide kojibiose[12] and maltose[18] were reported by Pinto
group. Conformations found by grid search supported by nu-
clear Overhauser enhancement (NOE) data coincide well
with their oxygen counterpart. Theoretical study of 5-thio-a-
d-glucopyranosyl-(1!3)-deoxymannojirimycin by Izumi et
al.[19] also suggested good similarity of its three-dimensional
structure with that of its oxygen counterpart. For H-type II
trisaccharide, Barker and co-workers[20] reported the so-
lution conformation using the 13C enriched trisaccharide syn-
Results and Discussion
3
3
thesized by using glycosyltransferases. JC,C and JC,H cou-
plings of interglycosidic bonds were used to estimate their
torsion angles. The F and Y torsion angles of the most
abundant conformer were estimated to be F’ ꢁ558, Y’ ꢁ08
Synthesis of 5-thio-l-fucose-containing H-type II: We have
previously reported the convergent synthesis of 5-thio-l-
fucose-containing H-type II and Lewis X.[8] H-type II and
Lewis X (Lex) trisaccharide both contains N-acetyllactos-
amine. The difference between them is the fucosylated posi-
tion, which is the 2’-OH group in H-type II, whereas it is 3-
OH in Lex. We used O-(3,4,6-tri-O-acetyl-2-O-levulinoyl-b-
d-galactopyranosyl)-(1!4)-1,6-anhydro-2-azido-3-O-(tert-
butyldimethylsilyl)-2-deoxy-b-d-glucopyranose as a common
intermediate for an efficient synthesis of both trisaccharides.
The reducing end was protected in a 1,6-anhydro ring be-
cause it can easily be converted into the glycosyl donor so
that diverse aglycons can be introduced.
(F’=H1’’-C1’’-O2’-C2’,
Y’=C1’’-O2’-C2’-H2’)
in
the
Fuca1!2Gal and F ꢁ608 and Y ꢁ 158 (F = H1’-C1’-O4-
C4, Y = C1’-O4-C4-H4) in the Galb1!4GlcNAc linkage.
However, this approach requires a 13C enriched compound
and is not generally applicable. NOE data are generally
used for a conformational analysis of oligosaccharides. Bush
and co-workers[21] used NOE data and CD data with confor-
mational energy calculations to estimate the torsion angles
to be F’=308, Y’=308 and F=608, Y=ꢀ108. For a com-
Chem. Eur. J. 2005, 11, 3032 – 3038
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3033

M. Izumi, H. Hashimoto et al.
During our first synthesis, Windmꢂller and Schmidt[25] re-
ported the efficient synthesis of lactoneo series oligosacchar-
ides including H-type II and Lex. Their synthetic strategy of
these trisaccharides is based on the regioselective benzoyla-
tion and fucosylation of the 2-azido-2-deoxy-3’,4’-O-isopro-
pylidene-lactose derivative. The lactose derivative was syn-
thesized from disaccharide lactose so that the glycosylation
step was avoided. Since we had some difficulties with the
glycosylation of 1,6-anhydro-2-azido-3-O-(tert-butyldime-
thylsilyl)-2-deoxy-b-d-glucopyranose with 3,4,6-tri-O-acetyl-
2-O-levulinoyl-a-d-galactopyranosyl bromide in our first
synthetic route, we examined their synthetic route for the
syntheses of the 5-thio-fucose-containing H-type II 1.
4’,6’-O-isopropylidene isomer as a byproduct in 18% yield.
The primary hydroxyl groups of 4 were selectively benzoy-
lated[27] with BzCN in DMF at ꢀ408C to give the diol 5 in
71% yield. The glycosylation of the diol 5 with the 5-thiofu-
cosyl trichloroacetimidate 6 in the presence of BF3·OEt2 in
CH2Cl2 at ꢀ208C proceeded a-selectively and regioselec-
tively to 2’-OH group, yielding the H-type II precursor 7.
The acceptor 5 could not be separated from 7 at this point.
Hydrolysis of isopropylidene group with 50% TFA in
CH2Cl2 after acetylation of the mixture of 5 and 7, and sub-
sequent deacylation and acetylation gave the pure octaace-
tate 9. The structure of 9 was confirmed at this stage by
1
comparing its H NMR spectrum with the spectrum of our
The synthesis of the H-type II trisaccharide 1 from the
hexaacetyl-2-azido-2-deoxy-a-lactosyl trichloroacetimidate 2
was showed in Scheme 1. Glycosidation of the a-imidate 2
with 8-methoxycarbonyloctanol gave b-lactoside 3 in 33%
yield. The low yield is due to the difficulty of the separation
of the pure b-glycoside from the complex mixture including
a-glycoside formed in the glycosylation. Deacetylation of 3
with methanolic sodium methoxide followed by treatment in
previously[8] synthesized sample. Then azido group was con-
verted into acetamido group in 95% yield by treatment with
hydrogen sulfide in aqueous pyridine and subsequent acety-
lation. Finally, O-deacetylation with methanolic sodium
methoxide gave the 5-thio-l-fucose-containing H-type II tri-
saccharide 1 in 62% yield after purification by solid-phase
extraction with reverse-phase C18 cartridge.
[26]
refluxing acetone in the presence of FeCl3 gave the 3’,4’-
Conformational analysis of 5-thio-l-fucose-containing H-
type II: The chemical shift of every proton of the 5-thio-l-
O-isopropylidene derivative 4 in 40% yield along with the
fucose-containing H-type II
1
was determined by
HOHAHA and HSQC experiments and are summarized in
Table 1, along with the native H-type II 11. Most of the res-
onances of N-acetylglucosamine and galactose of 1 are in
good accordance with those of its oxygen counterpart; only
the resonance of 5-thio-l-fucosylated site (H2 of Gal) has
been shifted about 0.25–0.3 ppm to low field. The conforma-
tional analysis of the 5-thio-l-fucose-containing H-type II 1
and the H-type II 11 was carried out with NMR experi-
ments. The NOESY experiments were carried out with sev-
eral mixing times, and NOESY spectra and observed NOE
values are depicted in Figure 2. NOEs between H1 of galac-
tose and H4 and H6a,b of N-acetylglucosamine, and H2 of
galactose and H1 of fucose were observed in the both com-
pounds. However, we could not determine which proton of
galactose had an NOE with H1 of 5-thiofucose since H2, 3
and 4 of galactose overlapped. A significant difference was
the NOEs between H5 of fucose and H5, 6a of N-acetylglu-
cosamine, which were observed in 11 but were not observed
in 1. ROESY experiments were also carried out for both
compounds and they gave similar results.
Molecular modeling studies of the methyl glycoside of the
5-thio-l-fucose-containing H-type II (5’’S H-type II, 12) and
the native H-type II 13 (Figure 3) were carried out. Monte
Carlo conformational search was performed on MacroMo-
del ver.5.5 software[28] by using AMBER* force field and
GB/SA water model and NOESY data were included as
constraints. After 15000 Monte Carlo steps, the global mini-
mum structure found for 5’’S H-type II 12 had the intergly-
cosidic torsion angles of F’=49.88, Y’=20.28 (F’=H1’’-
C1’’-O2’-C2’, Y’=C1’’-O2’-C2’-H2’) in the 5SFuca1!2Gal
and F = 50.38, Y = ꢀ5.68 (F = H1’-C1’-O4-C4, Y = C1’-
O4-C4-H4) in the Galb1!4GlcNAc linkage. The global
minimum structure of native H-type II (13) had the torsion
Scheme 1. a) HO(CH2)8CO2Me, BF3·OEt2, (CH2Cl)2/hexane, ꢀ408C
(33%); b) 1. NaOMe, MeOH, 2. acetone, FeCl3 (40%); c) BzCN, Et3N,
DMF, ꢀ408C (71%); d) BF3·OEt2, MS 4 ꢁ, CH2Cl2, ꢀ208C; e) Ac2O,
pyridine; f) 1. 50%TFA, CH2Cl2, 2. NaOMe, MeOH, 3. Ac2O, pyridine
(24% from 5); g) 1. H2S, pyridine, H2O, 2. Ac2O, pyridine (95%);
h) NaOMe, MeOH (62%).
3034
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3032 – 3038

Fucose Derivatives
FULL PAPER
Figure 2. Selected region of NOESY spectrum and observed NOEs of 5’’S-H-type II 1, and native H-type II 11.
angles of F’ = 45.98, Y’ = 20.48 and F = 52.28, Y =
ꢀ2.38, which was in good accordance with those values re-
ported previously. The conformation of the global minimum
of 5’’S H-type II (12) is very similar to that of the native tri-
saccharide 13. Further extensive computational studies are
required to explain the difference of the observed NOE be-
tween 1 and 11. CPK models of the global minimum of 12
and 13 are illustrated in Figure 4. It is noteworthy that the
sulfur atom sits in the upper surface of the trisaccharide,
which consists of the b face of the galactose and the fucose
and the a face of the N-acetylglucosamine. The existence of
the sulfur atom at the surface of the molecule may affect its
binding affinity to lectins and antibodies.
amined as inhibitors along with their oxygen counterparts
(Figure 5). Three fucose-binding lectins, that is aleuria aur-
antia (AAL) isolated from a mushroom,[31] Angiulla anguilla
agglutinin (AAA) isolated from an eel,[32] and Ulex euro-
paeus agglutinin I (UEA-I) isolated from a fern, and two
anti-H monoclonal antibodies (MoAb) against H-type II tri-
saccharide and disaccharide (Fuca1!2Gal), were used as
agglutinins. Minimal concentrations required for an inhibi-
tion of hemagglutination reactions are summarized in
Table 2. No inhibition activities of 5-thio-l-fucose deriva-
tives were observed against AAL. Only weak inhibition ac-
tivities were observed against AAA by 5-thio-l-fucose and
the disaccharide 14 but not by the trisaccharide 1. The disac-
charide 14 and trisaccharide 1 showed good inhibition activi-
ties against UEA-I. The trisaccharide 1 showed very strong
inhibition activity against both monoclonal antibodies com-
pared to its oxygen counterpart. The disaccharide 14
showed weak activity against anti-H disaccharide MoAb but
no activity against anti-H trisaccharide MoAb. No inhibition
activities of 5-thio-l-fucose-containing saccharides against
AAL and AAA are presumably because of the involvement
Biological activities of 5-thio-l-fucose-containing oligosac-
charides: For the preliminary screening of the affinity of the
5-thio-l-fucose-containing saccharides with variety of
fucose-binding proteins, inhibition studies of hemagglutina-
tion reactions were performed. 5-Thio-l-fucose-containing
H-type II 1, allyl O-(5-thio-a-l-fucopyranosyl)-(1!2)-b-d-
galactopyranoside (14)[5] and 5-thio-l-fucose[29,30] were ex-
Chem. Eur. J. 2005, 11, 3032 – 3038
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3035

M. Izumi, H. Hashimoto et al.
Table 1. 1H NMR chemical shifts of 5’’S-H-type II 1 and native H-type II
11.
1 against antibodies also supports this assumption. Strong in-
hibition activity of the trisaccharide 1 against antibodies
compared with its oxygen counterpart might be the result of
additional hydrophobic interaction of the ring sulfur atom at
the surface of the molecule.
1
11
GlcNAc
H1
H2
H3
H4
4.55
3.78
3.71
3.79
3.54
4.06
3.87
2.09
4.54
3.77
3.70
3.81
3.50
4.02
3.84
In conclusion, the 5-thio-l-fucose-containing H-type II tri-
saccharide 1 was synthesized from 2-azido-2-deoxy-lactose
that can be derived from lactose by azidonitration of lactal.
The 5-thio-l-fucosylation of the diol 5 proceeded a-selec-
tively and regioselectively to 2’-OH. The resulting fully func-
tionalized trisaccharide was converted to 1 in good yield.
Conformational analysis of the 5-thio-l-fucose-containing
H-type II and the native H-type II was carried out through
NOESY experiments with the aid of molecular modeling
conformation search. The observed NOEs between N-ace-
tylglucosamine and galactose, and galactose and fucose were
same in those two trisaccharides. However, NOEs between
fucose and N-acetylglucosamine were significantly different.
Comparing the inhibition activity of oligosaccharide–protein
binding by the 5-thio-l-fucose-containing oligosaccharides
with their oxygen counterparts suggested that simply replac-
ing one ring oxygen in oligosaccharide to sulfur resulted in
making a compound with the different inhibition activity
spectrum. Taking into consideration that each protein has
the different way of recognition, it is very difficult to predict
the effect of using 5-thiosugar in oligosaccharide to its bio-
logical activity, as has already been pointed out by Yuasa
et al.[33] Still, stability of 5-thiosugar-containing oligosacchar-
ide to degradation enzymes and the cross-reactivity against
antibodies suggest its potential usefulness in carbohydrate-
based therapeutics and vaccines.
H5
H6a
H6b
NAc
Gal
H1
H2
H3
H4
H5
H6a
H6b
Fuc
H1
H2
H3
H4
H5
H6
4.57
3.96
3.95
3.95
3.75
3.85
3.80
4.58
3.71
3.92
3.94
3.74
3.83
3.77
5.21
4.05
3.83
4.10
3.45
1.27
5.33
3.86
3.84
3.85
4.27
1.29
Figure 3. Structures of 12 and 13 used in Monte Carlo conformational
search.
Experimental Section
General: Melting points are uncorrected. Optical rotations were mea-
sured with a JASCO DIP-4 polarimeter using 0.5 dm cell. NMR spectra
were recorded on JEOL EX-270 or Varian Unity-400 or Bruker AM-500
instruments. 1H NMR spectra recorded in CDCl3 or in D2O were refer-
enced to tetramethylsilane at d=0 ppm or to acetone at d=2.225 ppm,
respectively. 13C NMR spectra recorded in CDCl3 or in D2O were refer-
enced to the central peak of CDCl3 at d=77.0 ppm or to 1,4-dioxane at
d=67.4 ppm, respectively. Column chromatography was performed with
Kieselgel 60 (E. Merck) or Wakogel C-300 (Wako Pure Chem.). TLC
was carried out on plates precoated with silica gel 60 F254 (E. Merck),
with detection by UV light (254 nm) and/or by charring with 5% H2SO4
in MeOH or 1% Ce(SO4)2 and 1.5% (NH4)6Mo7O24·4H2O in 10%
H2SO4. SepPak C18 cartridge was purchased from Waters Inc. Dichloro-
methane was distilled twice from P2O5 and stored over MS4A. Crushed
molecular sieves were activated by heating at 2508C overnight and
cooled in vacuo prior to use.
Figure 4. CPK model of the global minimum of (I) 12 and (II) 13.
of the ring oxygen in recognition by the hydrogen-bonding
network. It is known that key hydroxyl groups for binding
to UEA-I are the 2,3,4-OH groups of fucose and involve-
ment of 3-OH of galactosyl is also suggested.[24] The fact
that UEA-I recognized the 5-thio-l-fucose-containing di-
and trisaccharide 14 and 1, respectively, suggested that these
oligosaccharides possess those key hydroxyl groups in
proper position, which is consistent with the result of confor-
mational analysis. The inhibition activity of the trisaccharide
8-Methoxycarbonyloctyl O-(2,3,4,6-tetra-O-acetyl-b-d-galactopyranosyl)-
(1!4)-3,6-di-O-acetyl-2-azido-deoxy-b-d-glucopyranoside (3): A suspen-
sion of O-(2,3,4,6-tetra-O-acetyl-b-d-galactopyranosyl)-(1!4)-3,6-di-O-
acetyl-2-azido-2-deoxy-a-d-glucopyranosyl
trichloroacetimidate
(2,
1.484 g, 1.94 mmol), HO(CH2)8CO2Me (0.549 g, 2.92 mmol) and crushed
activated MS 4 ꢁ (0.94 g) in (CH2Cl)2/hexane (1:1 v/v, 16.6 mL) was stir-
red for 1 h under Ar atmosphere. The suspension was cooled to ꢀ408C,
and to the suspension was added dropwise a solution of BF3·OEt2
(0.13 mL, 1.06 mmol) in (CH2Cl)2/hexane (1:1 v/v, 7.0 mL). After being
stirred for 30 min at ꢀ408C, the reaction mixture was warmed to room
3036
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3032 – 3038

Fucose Derivatives
FULL PAPER
zation from EtOAc/Et2O gave
5
(0.181 g). The mother liquor was con-
centrated and another silica gel chro-
matography (hexane/EtOAc 6:5) gave
further 5 (93.5 mg, 71%). M.p. 125–
1268C; [a]2D0
=
+43.8 (c=0.97 in
CHCl3); 1H NMR (CDCl3): d=8.17–
7.42 (m, 10H; 2ꢃPh), 4.89–4.84 (m,
2H; H6a,6’a), 4.65 (brs, 1H; OH3),
4.46–4.37 (m, 2H; H6b,6’b), 4.30 (d, 3J-
(H,H)=8.3 Hz, 1H; H1’), 4.26 (d, 3J-
(H,H)=8.3 Hz, 1H; H1), 4.23–4.19
(m, 2H; H4’,5’), 4.13 (dd, 3J(H,H)=
5.6, 6.9 Hz, 1H; H3’), 3.93–3.85 (m,
1H; OCH2), 3.72–3.66 (m, 1H; H2’),
3.66 (s, 3H; OMe), 3.65–3.57 (m, 2H;
H3,5), 3.57–3.49 (m, 1H; OCH2), 3.39
(dd, 3J(H,H)=8.3, 9.7 Hz, 1H; H4),
3.28 (dd, 3J(H,H)=8.3, 9.7 Hz, 1H;
H2), 2.29 (t, 3J(H,H)=7.4 Hz, 2H;
COCH2), 1.63–1.25 (m, 12H; (CH2)6),
1.56, 1.37 ppm (each s, 2ꢃ3H; CMe2);
elemental analysis calcd (%) for
C39H51N3O14: C 59.61, H 6.54, N 5.35;
found: C 59.59, H 6.51, N 5.29.
Figure 5. H-type II derivatives used in hemagglutination inhibition assay.
Table 2. Minimum concentration (mm) required for inhibition of hemagglutination reaction.[a]
1
11
14
15
5SFuc
Fuc
2.5
1.25
AAL
AAA
UEA I
anti-H di
anti-H tri
>10[a]
1.25
0.313
0.313
5.0
>10[a]
n.d.
n.d.
n.d.
2.5
>10[a]
5.0
>10[a]
5.0
0.625
1.25
0.625
0.154
>10[b]
>10[b]
>10[b]
>10[b]
5.0
>10[b]
>10[b]
5-Thio-l-fucose-containing H-type II
1.25
>10[b]
5.0
1:
thio-l-fucopyranosyl trichloroacetimi-
date (77.1 mg, 0.171 mmol),
A mixture of 2,3,4-tri-O-acetyl-5-
[a] See Figure 5 for structures of inhibitors; AAL: Aleuria aurantia lectin; AAA: Anguilla anguilla agglutinin;
UEA I: Ulex europaeus agglutinin I; anti-H di: anti H disaccharide monoclonal antibody; anti-H tri: anti H-
type II trisaccharide monoclonal antibody; n.d. not determined. [b] No inhibition at 10 mm.
6
5
(133 mg, 0.169 mmol) and crushed acti-
vated MS 4 ꢁ (0.53 g) in CH2Cl2
(3.6 mL) was stirred under Ar atmos-
phere for 1 h at room temperature and
cooled to ꢀ208C. A solution of BF3·OEt2 (7.2 mL, 56 mmol) in CH2Cl2
(1.4 mL) was added dropwise to the cooled mixture, which was stirred
for 30 min. The reaction mixture was neutralized with Et3N and filtered
through a Celite pad. The filtrate was washed with saturated aqueous
NaHCO3, dried over MgSO4 and concentrated. The residue was purified
on a column of silica gel (hexane/EtOAc 12:5) to give a mixture of trisac-
charide 7 and 5 (184 mg) as a syrup. 1H NMR (CDCl3): d=5.29 (s, 3J-
(H,H)=2.6 Hz, 1H; H1’’), 1.15 ppm (d, 3J(H,H)=6.9 Hz, 1H; H6’’). The
mixture was treated with Ac2O and pyridine to give 8. To the solution
was added MeOH and concentrated. The residue was treated with 50%
TFA (0.5 mL) in CH2Cl2 (15 mL). The solution was concentrated, and
the residue was treated with methanolic sodium methoxide. The solution
was neutralized with Dowex 50W-X8 (H+), and the resin was filtered off.
The filtrate was concentrated, and the residue was purified by silica gel
chromatography (CHCl3/MeOH 4:1) to give a syrup. The syrup was ace-
tylated with Ac2O and pyridine to give octaacetate 9 (42 mg, 24%) as a
syrup. [a]1D9 = ꢀ108 (c=1.0 in CHCl3); 1H NMR (CDCl3): d=5.48 (brs,
1H; H4’’), 5.30 (brd, 3J(H,H)=3.3 Hz, 1H; H4’), 5.23 (dd, 3J(H,H)=2.3,
10.6 Hz, 1H; H2’’), 5.16 (brd, 3J(H,H)=10.6 Hz, 1H; H3’’), 5.07 (d, 3J-
(H,H)=2.3 Hz, 1H; H1’’), 5.02 (dd, 3J(H,H)=3.3, 10.2 Hz, 1H; H3’),
4.85 (t, 3J(H,H)=9.9 Hz, 1H; H3), 4.48 (brd, 3J(H,H)=12.2 Hz, 1H;
H6a), 4.39 (d, 3J(H,H)=7.9 Hz, 1H; H1’), 4.35 (d, 3J(H,H)=7.9 Hz, 1H;
H1), 4.28 (dd, 3J(H,H)=5.6, 12.2 Hz, 1H; H6b), 4.18 (dd, 3J(H,H)=6.6,
10.9 Hz, 1H; H6’a), 4.12–4.04 (m, 2H; H2’,6’b), 3.92 (m, 1H; OCH2),
3.84 (t, 3J(H,H)=6.6 Hz, 1H; H5’), 3.79 (t, 3J(H,H)=9.2 Hz, 1H; H4),
3.66 (s, 3H; OMe), 3.60–3.54 (m, 2H; H5’’,OCH2), 3.43 (m, 1H; H5),
3.23 (dd, 3J(H,H)=7.9, 9.9 Hz, 1H; H2), 2.30 (t, 3J(H,H)=7.3 Hz, 2H;
COCH2), 2.17, 2.14, 2.07, 2.00, 1.97, 1.96 (each s, 3H, 9H, 5ꢃ3H; 8ꢃAc),
1.64–1.26 (m, 12H; (CH2)6), 1.19 ppm (d, 3J(H,H)=6.9 Hz, 3H; H6’’);
13C NMR (CDCl3): d=170.9, 170.7, 170.5, 170.3, 170.0, 170.0, 169.8,
169.7, 102.2, 100.1, 80.3, 73.7, 73.4, 73.0, 72.6, 71.8, 71.7, 71.3, 70.9, 70.7,
68.5, 67.3, 63.8, 62.3, 61.0, 51.5, 34.2, 34.2, 29.5, 29.1, 29.0, 25.8, 24.9, 20.9,
20.8, 20.7, 20.6, 15.5 ppm.
temperature and filtered through a pad of Celite. The filtrate was washed
with saturated aqueous NaHCO3, and the organic layer was dried over
MgSO4 and concentrated. The residue was purified on a column of silica
gel (hexane/EtOAc 1:1) to give 3 (0.499 g, 33%) as a syrup. [a]2D5 = ꢀ6.4
(c=1.1 in CHCl3); 1H NMR (CDCl3): d=5.35 (dd, 3J(H,H)=0.7, 3.3 Hz,
1H; H4’), 5.09 (dd, 3J(H,H)=7.6, 10.6 Hz, 1H; H2’), 4.97 (t, 3J(H,H)=
9.6 Hz, 1H; H3), 4.94 (dd, 3J(H,H)=3.3, 10.6 Hz, 1H; H3’), 4.46 (dd, 3J-
(H,H)=2.0, 11.9 Hz, 1H; H6a), 4.45 (d, 3J(H,H)=7.6 Hz, 1H; H1’), 4.34
(d, 3J(H,H)=8.1 Hz, 1H; H1), 4.17 (dd, 3J(H,H)=6.6, 11.2 Hz, 1H;
H6’a), 4.14–4.04 (m, 2H; H6b,6’b), 3.93–3.85 (m, 1H; CH2O), 3.87 (t, 3J-
3
(H,H)=6.6 Hz, 1H; H5’), 3.70 (t, J(H,H)=9.6 Hz, 1H; H4), 3.66 (s, 3H;
OMe), 3.58–3.50 (m, 2H; H5, CH2O), 3.38 (dd, 3J(H,H)=7.6, 10.2 Hz,
1H; H2), 2.30 (t, 3J(H,H)=7.4 Hz, 2H; CH2CO), 2.16, 2.12, 2.12, 2.07,
2.03, 1.97 ppm (each s, 6ꢃ3H; 6ꢃAc); elemental analysis calcd (%) for
C34H51N3O18: C 51.71, H 6.51, N 5.32; found: C 51.70, H 6.52, N 4.95.
8-Methoxycarbonyloctyl O-(6-O-benzoyl-3,4-O-isopropylidene-b-d-galac-
topyranosyl)-(1!4)-2-azido-6-O-benzoyl-2-deoxy-b-d-glucopyranoside
(5): A 0.5m methanolic NaOMe solution (0.27 mL) was added to a so-
lution of 3 (0.499 mg, 0.632 mmol) in MeOH (6.2 mL). After being stir-
red overnight at room temperature, the solution was neutralized with
Dowex 50W-X8 (H+). The resin was filtered off and the filtrate was con-
centrated. FeCl3 (30 mg, 0.11 mmol) was added to a solution of the resi-
due in acetone (132 mL). After being stirred under reflux for 30 min, the
solution was cooled and neutralized with 10% aqueous K2CO3 (26 mL),
and acetone was evaporated. The aqueous layer was extracted with
EtOAc (132 mL), and the organic layer was dried over MgSO4 and con-
centrated. The residue was purified on a column of silica gel (CHCl3/
MeOH 15:1) to give 4 (146 mg, 40%) and its 4’,6’-O-isopropylidene
isomer (66.5 mg, 18%) each as a syrup. A solution of BzCN (0.15 mL,
0.13 mmol) and Et3N (0.21 mL) in DMF (1.0 mL) was added under Ar
atmosphere at ꢀ408C to a solution of 4 (0.283 g, 0.490 mmol) in DMF
(3.9 mL). After being stirred for 1 h at ꢀ408C, excess of the reagent was
decomposed by adding MeOH and the solution was concentrated. The
residue was dissolved in EtOAc, and washed with brine. The organic
layer was dried over MgSO4 and concentrated. The residue was purified
on a column of silica gel (hexane/EtOAc 6:5) to give a syrup. Recrystalli-
A solution of 9 (42 mg 41 mmol) in pyridine/water (1:1, 3.5 mL) was satu-
rated with H2S at room temperature. After being stirred overnight at
room temperature, the reaction mixture was concentrated, and the resi-
Chem. Eur. J. 2005, 11, 3032 – 3038
www.chemeurj.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3037

M. Izumi, H. Hashimoto et al.
due was acetylated with Ac2O/pyridine, and purified on a column of
silica gel (hexane/EtOAc 1:2) to give acetamide 10 (40 mg, 95%) as a
syrup. 1H NMR (CDCl3): d=5.69 (d, 3J(H,H)=8.9 Hz, 1H; NH), 5.50
(brs, 1H; H4’’), 5.30 (brd, 3J(H,H)=2.6 Hz, 1H; H4’), 5.22 (m, 2H;
H2’’,3’’), 5.11 (t, 3J(H,H)=8.6 Hz, 1H; H3), 5.09 (d, 3J(H,H)=2.3 Hz,
1H; H1’’), 5.04 (dd, 3J(H,H)=2.6, 10.2 Hz, 1H; H3’), 4.56–4.53 (m, 1H;
H6a), 4.54 (d, 3J(H,H)=7.3 Hz, 1H; H1), 4.44 (d, 3J(H,H)=7.6 Hz, 1H;
H1’), 4.29 (dd, 3J(H,H)=5.3, 11.5 Hz, 1H; H6b), 4.19–4.03 (m, 3H;
H2’,6’a,6’b), 3.99–3.81 (m, 4H; H2, 4, 5’, OCH2), 3.67–3.61 (m, 5H; H5,
5’, OMe), 3.47–3.43 (m, 1H; OCH2), 2.30 (t, 3J(H,H)=7.3 Hz, 2H;
COCH2), 2.17, 2.13, 2.13, 2.09, 2.06, 2.00, 1.97, 1.96 (each s, 6ꢃ3H, 6H,
3H; 9ꢃAc), 1.60–1.26 (m, 12H; (CH2)6), 1.18 ppm (d, 3J(H,H)=6.9 Hz,
3H; H6’’).
[3] S. Mehta, J. S. Andrews, B. D. Johnston, B. Svensson, B. M. Pinto, J.
Am. Chem. Soc. 1995, 117, 9783–9790.
[4] H. Hashimoto, M. Kawanishi, H. Yuasa, Chem. Eur. J. 1996, 2, 556–
560.
[5] M. Izumi, O. Tsuruta, S. Harayama, H. Hashimoto, J. Org. Chem.
1997, 62, 992–998.
[6] H. Yuasa, Y. Nakano, H. Hashimoto, Carbohydr. Lett. 1996, 2, 23–
26.
[7] H. Yuasa, S. Matsuura, H. Hashimoto, Bioorg. Med. Chem. Lett.
1998, 8, 1297–1300.
[8] M. Izumi, O. Tsuruta, H. Hashimoto, S. Yazawa, Tetrahedron Lett.
1996, 37, 1809–1812.
[9] O. Tsuruta, H. Yuasa, H. Hashimoto, S. Kurono, S. Yazawa, Bioorg.
Med. Chem. Lett. 1999, 9, 1019–1022.
[10] H. Hashimoto, M. Izumi, Tetrahedron Lett. 1993, 34, 4949–4952.
[11] S. Mehta, B. M. Pinto, Tetrahedron Lett. 1992, 33, 7675–7678.
[12] S. Mehta, K. L. Jordan, T. Weimar, U. C. Kreis, R. J. Batchelor,
F. W. B. Einstein, B. M. Pinto, Tetrahedron: Asymmetry 1994, 5,
2367–2396.
[13] S. Mehta, J. S. Andrews, B. D. Johnston, B. M. Pinto, J. Am. Chem.
Soc. 1994, 116, 1569–1570.
[14] O. Tsuruta, G. Shinohara, H. Yuasa, H. Hashimoto, Bioorg. Med.
Chem. Lett. 1997, 7, 2523–2526.
To a solution of 10 (40 mg, 38 mmol) in MeOH (3.0 mL) was added a cat-
alytic amount of NaOMe. After being stirred overnight, the solution was
neutralized with Dowex 50W-X8 (H+), and the resin was filtered off. The
filtrate was concentrated, and the residue was dissolved in water and ab-
sorbed to the SepPak C18 cartridge. After wash with water, compound 1
(17.2 mg, 62%) was eluted with MeOH. 1H NMR see Table 1; 13C NMR
(D2O): d=178.7, 175.2, 101.9, 101.5, 84.8, 77.6, 77.2, 76.1, 76.0, 74.8, 74.5,
73.2, 72.1, 71.5, 71.4, 70.0, 61.9, 61.1, 56.1, 52.9, 37.1, 34.5, 29.3, 29.1, 29.0,
25.8, 25.1, 23.1, 16.4 ppm; HR-MALDI-FTMS: m/z: calcd for
C30H53NO16SNa: 738.2983, found 738.2963 [M+Na]+.
[15] O. Tsuruta, H. Yuasa, H. Hashimoto, K. Sujino, A. Otter, H. Li,
M. M. Palcic, J. Org. Chem. 2003, 68, 6400–6406.
[16] B. M. Pinto, R. Y. N. Leung in The anomeric effect and associated
stereoelectronic effects, Vol. 539 (Ed.: G. R. J. Thatcher), American
Chemical Society, Washington, DC, 1993, pp. 126–155.
[17] J. B. Lambert, S. M. Wharry, Carbohydr. Res. 1983, 115, 33–40.
[18] T. Weimar, U. C. Kreis, J. S. Andrews, B. M. Pinto, Carbohydr. Res.
1999, 315, 222–233.
[19] M. Izumi, Y. Suhara, Y. Ichikawa, J. Org. Chem. 1998, 63, 4811–
4816.
[20] P. R. Rosevear, H. A. Nunez, R. Barker, Biochemistry 1982, 21,
1421–1431.
NMR methods for conformational analysis: The NMR experiments were
run on a Bruker DMX-500 instrument. The temperature was maintained
at 313 K. Two-dimensional 1H–13C HMQC, TOCSY, NOESY, and
ROESY spectra were measured by use of pulse programs in the Bruker
standard library (invbtp, mlevprtp, noesyprtp, and roesyprtp, respective-
ly). During acquisition, GARP decoupling was performed toward 13C
(HMQC). TOCSY spectrum was recorded by using MLEV-17 pulse se-
quence with a total spin locking time for 100 ms. The mixing times in
NOESY pulse sequence were varied from 100 to 1000 ms. Spin locking
time in ROESY were varied from 200 to 300 ms and the carrier frequen-
cy was placed at the left side of the spectrum at 6 ppm in order to mini-
mize HOHAHA type magnetization transfer.
[21] B. N. N. Rao, V. K. Dua, C. A. Bush, Biopolymers 1985, 24, 2207–
2229.
Molecular modeling: All calculations were performed on
a Silicon
Graphics O2 workstation using MacroModel ver.5.5 software. Initial
structures of 12 and 13 were built within MacroModel. The Monte Carlo
(MC) approach was used for the global conformation search, and the tor-
sion angles of interglycosidic linkages were randomly modified at each
MC step. MC steps (15000) were carried out for both compounds and re-
sultant geometry was minimized using gradient conjugate steps with the
AMBER* force field and the GB/SA water model with NOE data as
constraints.
[22] H. Thøgersen, R. U. Lemieux, K. Bock, B. Meyer, Can. J. Chem.
1982, 60, 44–57.
[23] A. Imberty, E. Mikros, J. Koca, R. Mollicone, R. Oriol, S. Pꢄrez,
Glycoconjugate J. 1995, 12, 331–349.
[24] M.-H. Du, U. Spohr, R. U. Lemieux, Glycoconjugate J. 1994, 11,
443–461.
[25] R. Windmꢂller, R. R. Schmidt, Tetrahedron Lett. 1994, 35, 7927–
7930.
Inhibition studies of hemagglutination reaction: Lectins (AAA, AAL,
UEA-I) and antibodies (anti-H type II antibody, anti-H disaccharide an-
tibody) were diluted to titer 4 with phosphate-buffered saline (PBS). In-
hibitors (1, 14, 15, 16, 5SFuc, Fuc) were diluted serially with PBS starting
from final concentration of 10 mm. Agglutinins were incubated with in-
hibitors, and then human type O erythrocyte in PBS was added to see
the residual hemagglutination activity. The minimum concentration of in-
hibitors that inhibited the hemagglutination reaction completely was de-
termined.
[26] R. Bommer, W. Kinzy, R. R. Schmidt, Liebigs Ann. Chem. 1991,
425–433.
[27] K. C. Nicolaou, T. Caulfield, H. Kataoka, T. Kumazawa, J. Am.
Chem. Soc. 1988, 110, 7910–7912.
[28] F. Mohamadi, N. G. J. Richards, W. C. Guida, R. Liskamp, M.
Lipton, C. Caufield, G. Chang, T. Hendrickson, W. C. Still, J.
Comput. Chem. 1990, 11, 440–467.
[29] H. Hashimoto, T. Fujimori, H. Yuasa, J. Carbohydr. Chem. 1990, 9,
683–694.
[30] M. Izumi, O. Tsuruta, H. Hashimoto, Carbohydr. Res. 1996, 280,
287–302.
[31] F. Fukumori, N. Takeuchi, T. Hagiwara, H. Ohbayashi, T. Endo, N.
Kochibe, Y. Nagata, A. Kobata, J. Biochem. 1990, 107, 190–196.
[32] S. Baldus, J. Thiele, Y. O. Park, F. G. Hanisch, J. Bara, R. Fischer,
Glycoconjugate J. 1996, 13, 585–590.
Acknowledgements
This work was supported by a Grant-in-Aid for Scientific Research on
Priority Area (A) No. 11121208 from the Ministry of Education, Cul-
ture, Sports, Science and Technology.
[33] H. Yuasa, H. Hashimoto, Rev. Heteroat. Chem. 1999, 19, 35–65.
[1] A. Varki, Glycobiology 1993, 3, 97–130.
[2] H. Yuasa, O. Hindsgaul, M. Palcic, J. Am. Chem. Soc. 1992, 114,
5891–5892.
Received: August 12, 2004
Revised: January 24, 2005
Published online: March 15, 2005
3038
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2005, 11, 3032 – 3038