Orthogonal glycosylation reactions on solid phase and synthesis of a library consisting of a complete set of fucosyl galactose isomers

Article abstract of DOI:10.1002/anie.200600433

A dream team: Orthogonal glycosylation and solid-phase synthesis were shown to be an excellent combination for polysaccharide synthesis (see scheme). The method was applied to generate a complete library of structural isomers with an L-fucosyl-D-galactose

Full text of DOI:10.1002/anie.200600433

Angewandte
Chemie
Table 1: Anomeric configurations and linkage positions of librarycom-
pounds 1–16.
Oligosaccharide Synthesis
DOI: 10.1002/anie.200600433
Fuc
Anomeric
configuration
Gal
Anomeric
configuration
Compound
Linkage
position
Orthogonal Glycosylation Reactions on Solid
Phase and Synthesis of a Library Consisting of a
Complete Set of Fucosyl Galactose Isomers**
a
b
a
b
2
2
2
2
a
a
b
b
1
2
3[a]
4
Osamu Kanie,* Isao Ohtsuka, Takuro Ako,
Shusaku Daikoku, Yoshimi Kanie, and Rumiko Kato
a
b
a
b
3
3
3
3
a
a
b
b
5
6
7
8
The development of efficient methods for accessing synthetic
oligosaccharides with diverse structures is essential to our
understanding of the involvement of such molecules in
biological phenomena.[1] Recent efforts directed toward the
synthesis of complex glycoconjugates have made it possible to
control the stereo- and regiospecificity of glycosylation
reactions in solution- and solid-phase organic synthesis.[2–10]
Current syntheses, however, can not be applied directly to the
combinatorial chemistry of carbohydrates, for which the
synthetic protocols must be extremely simple.[11] Various
approaches to the creation of an oligosaccharide library have
been described,[11–14] however, there are no reports of
successful multiple glycosylation reactions that occur at
every hydroxy-group position of a monosaccharide acceptor
in a position-specific manner. We describe herein the syn-
thesis of a complete disaccharide library with an l-fucosyl-d-
galactose (Fuc-Gal) sequence with three parameters, namely,
the anomeric configuration of the Fuc and Gal residues and
the linkage position (Table 1). The library was synthesized on
the basis of an orthogonal glycosylation strategy. The
potential of this approach for solid-phase oligosaccharide
synthesis is also illustrated with a focus on overall efficiency.
The particular Fuc-Gal sequence used was chosen because
aFuc(1!2)bGal is an O-blood-type epitope and also relates
to the Leb antigen, which has been reported to act as a ligand
of Helicobacter pylori. Thus, a library may provide important
information for the potential treatment of ulcers.
a
b
a
b
4
4
4
4
a
a
b
b
9
10
11
12
a
b
a
b
6
6
6
6
a
a
b
b
13
14
15
16
[a] Synthesis was reported in reference [21].
orthogonal glycosylation strategy[16] with solid-phase chemis-
try.[17–19] In the orthogonal glycosylation method, reaction
starts from the nonreducing terminus. In this way, the
formation of the deletion sequence is suppressed without
the need for capping reactions, as has been described for
glycal chemistry, because the by-products can not serve as
glycosylating agents in the next cycle.[5] At the final stage of
glycosylation, a hydrophobic tag was introduced to facilitate
rapid isolation of the final product(s) after the deprotection
reactions.[20–22]
In our strategy, the position of the interglycosidic linkage
is fixed and the anomeric configuration is randomized to
avoid formation of a highly complex mixture that would make
isolation extremely difficult.[11] The glycosyl fluorides 17–20
were synthesized from the corresponding phenylthioglyco-
sides of galactopyranose[23] by protection of a hydroxy group
with a chloroacetyl group (ClAc), conversion of the phenyl-
thio group into a fluoride substituent by treatment with N,N-
diethylaminosulfur trifluoride, and removal of the ClAc
group.[24] The synthesis of a disaccharide library was carried
out with 17–20, which contain orthogonal leaving groups, and
a fucosyl donor in the form of the phenylthioglycoside 21
(Scheme 1). The phenylthio group was activated chemo-
selectively over glycosyl fluoride with dimethyl(methylthio)-
sulfonium trifluoromethanesulfonate (DMTST)[25,26] to yield
disaccharides in generally good yields (TLC analysis). Indi-
vidual anomers in the disaccharide mixtures 22–25 were not
isolated at this stage, to avoid the unnecessary loss of some
anomers. The disaccharide fractions were collected as mix-
tures by gel-permeation chromatography. The mixtures were
then used as the glycosyl donors and treated with n-octanol in
the presence of hafnocene bis(trifluoromethanesulfonate)
[Cp2Hf(OTf)2].[27–30] An n-octyl group was introduced at the
reducing-end anomeric position to facilitate the removal of
In planning the synthesis of an oligosaccharide library, we
focused on the efficiency of the overall process. It was
necessary to minimize the number of 1) types of protecting
groups used and 2) reaction steps in the synthesis. To fulfill
the first criterion, a benzyl group was chosen as the sole
protecting group, with the exception of protection, when
needed, at the connecting position to the resin.[15] To satisfy
the second criterion, we chose to combine the use of an
[*] Dr. O. Kanie, Dr. I. Ohtsuka, T. Ako, S. Daikoku, Dr. Y. Kanie, R. Kato
Mitsubishi Kagaku Institute of Life Sciences (MITILS)
11 Minamiooya, Machida-shi, Tokyo 194-8511 (Japan)
Fax: (+81)42-724-6317
E-mail: kokanee@libra.ls.m-kagaku.co.jp
[**] Financial support provided byMitsubishi Chemical Co. (MCC) and
the KeyTechnologyResearch Promotion Program of the New Energy
and Industrial Development Organization (NEDO), Ministryof
Economy, Trade, and Industry (METI) of Japan is gratefully
acknowledged.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 3851 –3854
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3851

Communications
group were collected at this stage (compounds with 2-, 3-, 4-,
and 6-linked Fuc-Gal were isolated in 58, 49, 65, and 68%
yield, respectively, over three steps). Finally, anomeric
mixtures were separated by HPLC with a C18 column to
give the individual anomers 1–16. Chemical shifts and
coupling constants for the anomeric positions, retention
times, and compound ratios are summarized in Table 2.[24,31]
Such a library might be used to obtain information
regarding the specificities of various enzymes, especially in
connection with methods of structural elucidation.[32]
A
preliminary assay of a library of 16 compounds with an
a(1!2)-“specific” fucosidase from Corynebacterium[33,34]
revealed that several compounds: 1 (0.5), 3 (4.0), 5 (1.0), 7
(1.2) and 15 (0.7), were accepted as substrates. (The numbers
in parentheses indicate the hydrolysis rate relative to that of
aFuc(1!2)bGal(1!4)Glc.) Compound 1 with an a-Gal
moiety was found unexpectedly to undergo very slow
hydrolysis, as well as a(1!3)-linked 5 and 7, and a(1!6)-
linked 15. Although no information regarding the structure of
this particular enzyme is available, it is possible that the 2-OH
group of Gal in 5 and 7 may occupy the same space as the 3-
OH group in 3 and the reference compound. The presence of
these groups may not be essential, as PNP-fucoside (p-
nitrophenyl a-fucoside) is a better substrate, but they do not
Scheme 1. a) DMTST, MS-AW300, C2H4Cl2/CH3CN (1:1), 08C, 4 h;
b) n-octanol, [Cp2Hf(OTf)2], MS-AW300, C2H4Cl2/CH3CN (1:1), ꢀ208C,
4 h; c) H2, Pd/C, MeOH. Bn=benzyl.
hydrophilic by-products after complete deprotection. This
approach is considered particularly useful in solid-phase
synthesis.[20] The octyl glycosides 26–29 were subjected to
hydrogenation and purified roughly by using a C18 cartridge
column (Sep-Pak). Ideally, only compounds with an octyl
Table 2: Properties of compounds 1–16.
[b]
Fuc
Gal
tR
Ratio of
[min] anomers[c]
d [ppm][a] J [Hz][a]
d [ppm][a] J [Hz][a]
1
2
5.05
4.45
5.26
4.67
3.6
7.9
3.6
7.8
5.02
5.08
4.62
4.54
3.5
3.7
7.9
7.3
73.3 12.0
46.4 1.0
93.3 37.0
49.0 21.0
3[21]
4
5
6
7
8
5.13
4.67
5.15
4.49
4.0
7.8
3.9
7.8
4.94
4.97
4.43
4.42
3.7
3.8
7.4
8.0
87.4 1.0
140
1.0
74.2 1.7
123
1.6
9
10
11
12
5.12
4.33
5.14
4.31
4.0
7.8
3.9
7.7
4.94
4.95
4.41
4.41
3.1
3.2
7.9
7.9
62.4 1.0
49.8 3.1
54.4 7.8
36.9 4.2
13
14
15
16
4.94
4.38
4.93
4.39
3.3
7.9
3.9
7.9
4.90
4.91
4.38
4.39
3.6
3.7
7.9
7.9
39.7 1.0
78.7 1.9
46.9 6.4
82.7 15.0
[a] NMR spectra (500 MHz) were recorded in D2O at 298 K (DOH: d=
4.768 ppm). [b] Retention time was measured byLC–MS; for conditions
and chromatograms, see the Supporting Information. [c] Ratios of
anomers were estimated from the LC–MS spectra on the basis of the
assumption that in each case both anomers were ionized similarly. The Scheme 2. a) DMTST, C2H4Cl2/CH3CN (1:1), ꢀ30!08C (2.58Chꢀ1);
ratio of the major isomer to the minor isomer of the same sequence is b) [Cp2Hf(OTf)2], C2H4Cl2/CH3CN (1:1), ꢀ15!08C (1.258Chꢀ1);
given.
c) NaOMe/MeOH/C2H4Cl2; d) H2, Pd(OH)2/C, MeOH/EtOAc (1:1).
3852
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3851 –3854

Angewandte
Chemie
interfere with binding or catal-
ysis. Interference by aglycons
in a-linked Gal moieties with
the surface of the protein may
be responsible for the differ-
ence in the observed activity.
Having confirmed the eli-
gibility of the orthogonal-gly-
cosylation protocol in a library
synthesis, we next investigated
a combined method with solid-
phase chemistry. Studies with
potential orthogonal glycosyl
donors attached to a solid sup-
port have been reported, but
only single glycosylation reac-
tions were carried out.[17–19] We
conducted solid-phase reac-
tions in which the glycosyl
donor was linked to a bead
through a succinyl ester (see
30; Scheme 2).[24] Glycosyla-
tion with the acceptors 17 and
20, which contain a secondary
and a primary hydroxy group,
respectively,
proceeded
smoothly to give the resin-
bound disaccharides 31 and
32, each of which consists of a
mixture of anomers.[35] The
excellent performance of the
reactions was confirmed by the
formation of disaccharides
(determined by ESI MS) upon
the treatment of a few beads
with methanolic NaOMe (Fig-
ure 1 A).[36] The disaccharide
fluoride 32 thus obtained was
subjected to a further glycosy-
lation reaction with 33, which
has a hydroxy group at C6 and
a phenylthio group at C1, in
Figure 1. Mass spectra of glycosylation products. Spectra of the A) disaccharide, B) trisaccharide, C) octyl trisacchar-
ide, and D) deprotected octyl trisaccharide 36 after purification byusing a Sep-Pak C18 cartridge column.
the presence of [Cp2Hf(OTf)2], to give 34 (Figure 1B).
Compound 34 underwent glycosylation with n-octanol readily
with DMTST as a promoter. (For MS analysis of the
trisaccharide, see Figure 1C.) After hydrogenolysis, the
anomeric mixture 36 was obtained in 33% yield (five steps)
by using a Sep-Pak column, which removed all accumulated
by-products (Figure 1D). The incorporation of a hydrophobic
tag in the final step on the solid support is important, as then
only the compounds with that tag are isolated. The yield of
the glycosylation reaction in this solid-phase synthesis (gly-
cosylation at the primary hydroxy group) was 72%, which is
comparable to that of the solution reaction (76%).[37]
In summary, we synthesized a library consisting of a
complete set of structural isomers with Fuc-Gal sequences as
octyl glycosides. Furthermore, in an assay of the library
compounds, unexpected substrate specificity of fucosidase
was discovered. The successful extension of an orthogonal-
glycosylation protocol to solid-phase synthesis and the
demonstration of a result comparable to that observed for
the solution reaction suggest the potential usefulness of the
present method in future combinatorial oligosaccharide syn-
thesis.
Received: February 1, 2006
Published online: April 28, 2006
Keywords: chemoselectivity· compound libraries ·
.
glycosylation · oligosaccharides · solid-phase synthesis
[1] A. Varki, Glycobiology 1993, 3, 97 –130.
[2] H. Paulsen, Angew. Chem. 1982, 94, 184 –201; Angew. Chem. Int.
Ed. Engl. 1982, 21, 155 –175.
[3] R. R. Schmidt, Angew. Chem. 1986, 98, 213 –236; Angew. Chem.
Int. Ed. Engl. 1986, 25, 212 –235.
Angew. Chem. Int. Ed. 2006, 45, 3851 –3854
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3853

Communications
[4] G.-J. Boons, Tetrahedron 1996, 52, 1095 –1121.
[5] S. J. Danishefsky, K. F. McClure, J. T. Randolph, R. B. Ruggeri,
Science 1993, 260, 1307 –1309.
[6] O. J. Plante, E. R. Palmacci, P. H. Seeberger, Science 2001, 291,
1523 –1527.
[7] Solid Support Oligosaccharide Synthesis and Combinatorial
Carbohydrate Libraries (Ed.: P. H. Seeberger), Wiley-Inter-
science, New York, 2001, and references therein.
[8] T. Kanemitsu, O. Kanie, Comb. Chem. High Throughput
Screening 2002, 5, 339 –360.
and Sep-Pak purification. Experimental details and the struc-
tural assignments of all anomers will be reported elsewhere.
[37] The calculation was based on the assumption of a quantitative
yield for the hydrogenolysis and 90% yield for the cleavage
reaction. Calculations were based on the data obtained for O6-
linked saccharides.
[9] Y. Ito, S. Manabe, Curr. Opin. Chem. Biol. 1998, 2, 701 –708.
[10] L. Yan, C. M. Taylor, R. Goodnow, Jr., D. Kahne, J. Am. Chem.
Soc. 1994, 116, 6953 –6954.
[11] O. Kanie, F. Barresi, Y. Ding, J. Labbe, A. Otter, L. S. Forsberg,
B. Ernst, O. Hindsgaul, Angew. Chem. 1995, 107, 2912 –2915;
Angew. Chem. Int. Ed. Engl. 1995, 34, 2720 –2722.
[12] R. Liang, L. Yan, J. Loebach, M. Ge, Y. Uozumi, K. Sekanina, N.
Horan, J. Gildersleeve, C. Thompson, A. Smith, K. Biswas, W. C.
Still, D. Kahne, Science 1996, 274, 1520 –1522.
[13] G.-J. Boons, B. Heskamp, F. Hout, Angew. Chem. 1996, 108,
3053 –3056; Angew. Chem. Int. Ed. Engl. 1996, 35, 2845 –2847.
[14] M. Izumi, Y. Ichikawa, Tetrahedron Lett. 1998, 39, 2079 –2082.
[15] T. Kanemitsu, C.-H. Wong, O. Kanie, J. Am. Chem. Soc. 2002,
124, 3591 –3599.
[16] O. Kanie, Y. Ito, T. Ogawa, J. Am. Chem. Soc. 1994, 116, 12073 –
12074.
[17] L. O. Kononov, Y. Ito, T. Ogawa, Tetrahedron Lett. 1997, 38,
1599 –1602.
[18] T. Doi, M. Sugiki, H. Yamada, T. Takahashi, J. A. Porco, Jr.,
Tetrahedron Lett. 1999, 40, 2141 –2144.
[19] J. Ferguson, C. Marzabadi, Tetrahedron Lett. 2003, 44, 3573 –
3577.
[20] M. M. Palcic, L. D. Heerze, M. Pierce, O. Hindsgaul, Glyco-
conjugate J. 1988, 5, 49 –63.
[21] T. Lowary, O. Hindsgaul, Carbohydr. Res. 1993, 249, 163 –195.
[22] U. J. Nilsson, E. J.-L. Fournier, O. Hindsgaul, Bioorg. Med.
Chem. 1998, 6, 1563 –1575.
[23] K. Suzuki, I. Ohtsuka, T. Kanemitsu, T. Ako, O. Kanie, J.
Carbohydr. Chem. 2005, 24, 219 –236.
[24] I. Ohtsuka, T. Ako, R. Kato, S. Daikoku, S. Koroghi, T.
Kanemitsu, O. Kanie, Carbohydr. Res. 2006, in press.
[25] M. Ravenscroft, R. M. G. Roberts, J. G. Tillett, J. Chem. Soc.
Perkin Trans. 2 1982, 1569 –1572.
[26] P. Fügedi, P. J. Garegg, Carbohydr. Res. 1986, 149, C9 –C12.
[27] K. Suzuki, H. Maeta, T. Matsumoto, G. Tsuchihashi, Tetrahedron
Lett. 1988, 29, 3571 –3574.
[28] K. Suzuki, H. Maeta, T. Matsumoto, Tetrahedron Lett. 1989, 30,
4853 –4856.
[29] K. C. Nicolaou, T. J. Caulfield, H. Kataoka, N. A. Stylianides, J.
Am. Chem. Soc. 1990, 112, 3693 –3695.
[30] Y. Ito, S. Nunomura, S. Shibayama, T. Ogawa, J. Org. Chem.
1992, 57, 1821 –1831.
[31] HRMS values obtained for compounds 1–16 were within the
generally tolerated range.
[32] C. J. Edge, T. W. Rademacher, M. R. Wormald, R. B. Parekh,
T. D. Butters, D. R. Wing, R. A. Dwek, Proc. Natl. Acad. Sci.
USA 1992, 89, 6338 –6342.
[33] A. Sano, T. Sakai, K. Hayakawa, I. Kato, Glycoconjugate J. 1991,
8, 271.
[34] K. Yamashita, K. Totani, Y. Iwaki, M. Kuroki, Y. Matsuoka, T.
Endo, A. Kobata, J. Biol. Chem. 1989, 264, 17873 –17881.
[35] Solid-phase reactions were carried out in the absence of
molecular sieves to avoid problems such as clogging and/or
resin breakdown.
[36] The yield of the solid-phase reaction was not determined at each
step. Instead, the overall yield was calculated after deprotection
3854
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3851 –3854