A strategy for the one-pot synthesis of sialylated oligosaccharides

Article abstract of DOI:10.1139/v02-131

A new strategy has been developed for the synthesis of branched sialylated oligosaccharides using one-pot technology. Sialyl donors are in general too weak in reactivity to be used as the first glycosyl donors in the one-pot synthesis. When sialic acid is linked to a different sugar such as galactose, the reactivity is, however, significantly enhanced and can be tuned to enable the one-pot synthesis. A combination of NIS-TfOH-AgOTf was used for activation of the thioglycosides to improve the glycosylation yield when a hindered acceptor was used, as illustrated in the one-pot assembly of sialylated hexasaccharide.

Full text of DOI:10.1139/v02-131

1051
A strategy for the one-pot synthesis of sialylated
oligosaccharides
Zhiyuan Zhang, Kenichi Niikura, Xue-Fei Huang, and Chi-Huey Wong
Abstract: A new strategy has been developed for the synthesis of branched sialylated oligosaccharides using one-pot
technology. Sialyl donors are in general too weak in reactivity to be used as the first glycosyl donors in the one-pot
synthesis. When sialic acid is linked to a different sugar such as galactose, the reactivity is, however, significantly en-
hanced and can be tuned to enable the one-pot synthesis. A combination of NIS–TfOH–AgOTf was used for activation
of the thioglycosides to improve the glycosylation yield when a hindered acceptor was used, as illustrated in the one-
pot assembly of sialylated hexasaccharide.
Key words: one-pot synthesis, sialyl oligosaccharides, new activation.
Résumé : Faisant appel à une technologie monotope, on a développé une nouvelle stratégie pour réaliser la synthèse
d’oligosaccharides sialylés ramifiés. La réactivité des donneurs sialyles est généralement trop faible pour les utiliser
comme premiers donneurs de glycosyle dans une synthèse monotope. Lorsque l’acide sialique est lié à divers sucres,
comme le galactose, la réactivité est toutefois considérablement augmentée et elle peut être ajustée pour utilisation dans
une synthèse monotope. Lorsque des accepteurs stériquement empêchés ont été utilisés, on a fait appel à une
combinaison de NIS–TfOH–AgOTf pour l’activation des thioglycosides afin d’augmenter le rendement de la
glycosylation; cette technique est illustrée par la synthèse monotope d’un hexasaccharide sialylé.
Mots clés : synthèse monotope, saccharides sialylés, nouvelle activation.
[Traduit par la Rédaction] Zhang et al. 1054
Introduction
of building blocks, with the most reactive one being added
first. However, sialic acid thioglycosides (even a per-O-
benzylated thioglycoside) are much less reactive and less in-
fluenced by the protecting group than the other
thioglycosides (5). We have found that the major effect on
the anomeric reactivity of sialic acid is from the carboxyl,
not from the side-chain protecting groups. The reactivity dif-
ference between per-O-benzylated and per-O-acetylated
sialosides is less than 100, compared to ~1 × 105 for the cor-
responding thioglycosides of the other sugars (5). Reduction
of the carboxyl group does increase the reactivity by >1 ×
104, but the ꢁ-selectivity is completely diminished and gives
mainly the undesirable ꢀ-glycoside product (5). To tackle
this problem, we decided to use sialylated disaccharides as
building blocks in the one-pot synthesis, as the reactivity of
a disaccharide or trisaccharide glycosyl donor is mainly de-
termined by the reducing end unit (4d). Here we report the
representative one-pot synthesis of a protected sialyl Lewis
X hexasaccharide (1) to illustrate this strategy.
Sialylated oligosaccharides are attractive targets for syn-
thesis as most cancer cells carry sialylated oligosaccharides
as unique markers on their surface (1). However, synthesis
of sialylated oligosaccharides remains a major problem (2),
mainly due to the low yield and low stereoselectivity in the
sialylation step. Strategically, the difficult sialylation should
be conducted at the early stage to maximize the overall yield
and stereoselectivity. Sialic acid is, however, often found at
the nonreducing end of oligosaccharides, and most methods
developed for oligosaccharide synthesis start from the reduc-
ing end to the nonreducing end. The strategy based on
Danishefsky et al.’s (3) glycal assembly, via epoxide inter-
mediates, starts from the nonreducing end, but it is not appli-
cable to sialic acid, as undesirable ꢀ-selectivity occurs.
Alternatively, fragment coupling using sialyl oligosaccha-
rides as glycosylation reagents may minimize the problems.
Another approach is based on the one-pot strategy (4),
where an oligosaccharide is assembled rapidly from the
nonreducing end to the reducing end by sequential addition
Experimental and results and discussion
Received 4 December 2001. Published on the NRC Research
Press Web site at http://cjc.nrc.ca on 20 August 2002.
As shown in Fig. 1, the target molecule could be assem-
bled from the sialylated disaccharide 2, the thiofucoside 4,
the 2-N-Phth glucoside 3, and the lactoside 5. This strategy
is based on the following analysis: when a 2-O-acylated
glycosyl donor couples with a 2-N-phthalimido protected
3,4-diol glucosyl acceptor (i.e., 2 + 3), it is expected to give
a 1,4-ꢀ linked product (5). This would allow the one-pot
strategy to proceed through a fucosylation with 4 after the
Dedicated to the memory of Professor Raymond U. Lemieux.
Z. Zhang, K. Niikura, X.-F. Huang, and C.-H. Wong.1
Department of Chemistry and the Skaggs Institute for
Chemical Biology, The Scripps Research Institute, 10550
North Torrey Pines Road, La Jolla, CA 92037, U.S.A.
1Corresponding author (e-mail: wong@scripps.edu).
Can. J. Chem. 80: 1051–1054 (2002)
DOI: 10.1139/V02-131
© 2002 NRC Canada

1052
Can. J. Chem. Vol. 80, 2002
Fig. 1. Retro-synthetic analysis for the one-pot assembly of protected sialyl Lewis X hexasaccharide.
HO
OH
OH
OH
O
HO
NHAc
O
OH
O
OH
O
O
HO
O
O
O
HO
O
1
O
OR
HO
O C
2
HO
OH
HO
O
HO
OBn
O
HO
HO
OH
OBn
O
N
H
HO
BnO
c
O
OBn
O
4
A
HO
OBn
STol
BnO
O
OMe
OBn
OBn
Ph
5
PhthN
O
O
HO
HO
O
STol
STol
OLev
O
MeO C
2
O
3
ONBz
O
AcO
AcO
2
OAc
AcO
N
H
c
A
first step, followed by coupling with the final acceptor
lactoside 5.
Encouraged by this result, we then conducted the assem-
bly of the four components 2–5 in a one-pot procedure. Our
experience suggests that exclusively dry conditions are the
key to a successful glycosylation step, and these conditions
are even more important in the one-pot procedure as more
reagents and substrates are added to the reaction flask. The
relative amounts of building blocks used is also one of the
important factors. Normally, the use of less equivalents of
the acceptor at each step can have many advantages. First, it
increases the selectivity; for example, if the concentration
ratio of the donor and the acceptor in the beginning is 1.5,
the ratio would roughly be ~2.0 in the middle of the reac-
tion, and much higher near the end. Second, the excess
amount of donor can still be activated, even though all of the
acceptor is consumed and trapped by the released
succinimide to form the glycosyl succinimide. When the last
acceptor is an O-linked glycoside, using an excess of the ac-
ceptor will increase the yield. Through several attempts at
the synthesis of the SLe X hexasaccharide, we found that the
optimal ratio of building blocks (2:3:4:5) is 1.3:1:1.5:2.
We first tried the one-pot assembly using NIS–TfOH as
the promoter. TLC analysis showed a good result in the first
step (coupling 2 with 3); however, almost no reaction was
observed in the second step, probably due to the steric hin-
drance at 3-OH after the 4-OH position was glycosylated to
the disaccharide 2. In addition, most of the thiofucoside 4
was converted to its succinimide derivative.
To reduce the amount of the succinimide, we have devel-
oped a new procedure (Fig. 4) using NIS–TfOH–AgOTf. In
this reaction the side product TolS-I is generated, as shown
in Scheme 1. In the presence of AgOTf, it reacts with TolS-I
to generate another more powerful promoter, TolS-OTf (8),
which can further activate the donor. In this manner,
0.5 equiv of NIS, together with 0.5 equiv of AgOTf, activate
1 equiv of the donor. As an extended application of this
reaction, we used 1.5 equiv of compound 2 and 1.5 equiv of
NIS in the first step (Fig. 4). The resultant mixture ideally
is composed of about 1 equiv of trisaccharide product,
0.5 equiv of succinimide derivative of compound 2, and
We expect that the 2,3-linked disaccharide 2 would have a
similar reactivity as the methylphenyl 2,3-di-O-benzoyl-4,6-
O-benzylidene-1-thio-ꢀ-D-galactopyranoside (RRV = 285),
and would be at least 10 times more reactive than 6-O-
nitrobenzoyl-2-N-phthalimido ꢀ-glucoside 3. Compound 2
was synthesized from phosphite 6 (7) and thiogalactoside 7
followed by acylation. The desired ꢁ-product 8 was isolated
in 21% yield, and the lactonized product 9 was the major
side product (8% yield). Protection of the 2-OH group using
levulinic acid gave the desired product 2 in 86% yield.
To reduce the reactivity of the 2-N-Phth glucoside, a
nitrobenzoyl group was selected as the 6-O protecting group.
Monobenzoylation of the 2-N-phthalimido glucoside 3 was
first tried with p-NO2BzCl–DMAP–pyridine in THF. This
reaction mainly gave two products, the 3,6-di-p-nitro-
benzoylated compound 10 (35% yield), and 6-O-NBz 3
(53% yield), which were difficult to purify. However, pure
compound 3 was easily obtained in 83% yield when we
treated compound 10 with (Bu2Sn)2O followed by p-nitrobenzoyl
chloride.
The last component, lactoside acceptor 5, was prepared
(in 74% overall yield) from the free lactoside 11 (Fig. 2)
through three steps and a single purification. An improved
procedure, using DMF as the cosolvent together with ꢁ,ꢁ-
dimethoxypropane under H2SO4 treatment, gave a good
yield of isopropylidene lactoside. The crude product was
subjected to further benzylation and TFA treatment to give
the diol 5 after final purification.
Before we started the assembly, we measured the reactivi-
ties of compound 2 and compound 3, using the HPLC
method developed in this group. Compound 2 (RRV = 1308)
was found to be 23 times more reactive than compound 3
(RRV = 57). To make sure that the glycosylation and the
regio- and stereoselectivities would work as predicted during
the one-pot synthesis, we tried a one-pot synthesis (Fig. 3)
of the protected Lewis X trisaccharide 14, and the product
was isolated in more than 85% yield.
© 2002 NRC Canada

Zhang et al.
1053
Fig. 3. One-pot synthesis of a protected Lewis X trisaccharide.
Fig. 2. Preparation of building blocks. (a) TMSOTf, MS AW-
300, CH2CI2ꢂCH3CN; (b) Levilunic acid, EDC, DMAP, CH2CI2;
(c) (Bu3Sn)2O, o-NBzCl, 82%; (d) CMA, ꢁ,ꢁ-dimethoxy
isopropane; (e) NaH, BnCl, DMF; ( f ) TFAꢂCH2Cl2, overall
82% yield.
Ph
OBn
O
OAc
AcO
AcHN
AcO
P
O
O
OBn
O
+
O
CO2Me
HO
STol
OAc
OH
6
7
a
Ph
O
OAc
CO2Me
AcO
AcHN
AcO
O
Fig. 4. One-pot synthesis of sialyl Lewis X hexasaccharide.
O
8 R = H, 21 %
2 R = Lev
O
b
O
STol
STol
RO
OAc
O
O
OAc
AcO
AcHN
AcO
O
O
O
9
O
OAc
Ph
OH
ONBz
O
c
O
HO
HO
HO
HO
STol
STol
PhthN
PhthN
10
3
OH
HO
HO
OH
O
O
O
HO
O
11
OMe
OH
Scheme 1. Proposed mechanisms of NIS–TfOH and AgOTf acti-
OH
vation of thioglycosides.
d-f
O
OBn
HO
HO
O
N
O
OR
HOR
OBn
O
O
O
O
BnO
OBn
O
5
OMe
O
I N
O
OBn
O
O
1.5 equiv amount of ToS-I. Therefore, addition of AgOTf
after adding thiofucoside 4 in the second step would gener-
ate an adequate amount of TolS-OTf (8) to activate the
trisaccharide and to give the tetrasaccharide. In the last
step we used NIS for the coupling, as the lactoside diol 5
(2.5 equiv) is much more reactive than succinimide, and less
hindered. By using this one-pot procedure, we were able to
isolate the final hexasaccharide 15 in 42% yield after flash
chromatography.2
STol
OTf
O
TfOH
H N
+
TolS-I
AgOTf
AgI
O
hυ
TolS-STol + I2
TolS-OTf
In summary, we have successfully developed a one-pot
regioselective glycosylation. We have also developed a new
method using NIS–TfOH–AgOTf for activation of
thioglycosides. Further application of the one-pot strategy in
synthesis of sialyl oligosaccharide using
a sialylated
disaccharide and other building blocks in combination with
2 Supplementary material may be purchased from the Depository of Unpublished Data, Document Delivery, CISTI, National Research Coun-
cil Canada, Ottawa, ON K1A 0S2, Canada (http://www.nrc.ca/cisti/irm/unpub_e.shtml).
© 2002 NRC Canada

1054
Can. J. Chem. Vol. 80, 2002
3. S.J. Danishefsky, K.F. McClure, J.T. Randolph, and R.B.
Ruggeri. Science (Washington, D.C.), 260, 4 (1993).
4. (a) S. Raghavan and D. Kahne. J. Am. Chem. Soc. 115, 1580
(1993); (b) H. Yamada, T. Harada, and T. Takahashi. J. Am.
Chem. Soc. 116, 7919 (1994); (c) S.V. Ley and H.W.M.
Priepke. Angew. Chem. Int. Ed. Engl. 33, 2292 (1994); (d) Z.
Zhang, I.R. Ollmann, X.-S. Ye, R. Wischnat, T. Baasov, and
C.-H. Wong. J. Am. Chem. Soc. 121, 734 (1999); (e) F.
Burkhart, Z. Zhang, S. Wacowich-Sgarbi, and C.-H. Wong.
Angew. Chem. Int. Ed. Engl. 40, 1274 (2001).
the syntheses of more complicated and highly branched nat-
ural oligosaccharides is in progress.
Acknowledgements
This research was supported by the U.S. Department of
Health and Human Services National Institutes of Health
(NIH) and the Skaggs Chemical Biology Program.
5. X.-S. Ye, X.-F. Huang, and C. H. Wong. Chem. Commun. 11,
974 (2001).
6. U. Ellervik and G. Magnusson. J. Org. Chem. 63, 9314 (1998).
7. M.M. Sim, H. Kondo, and C.H. Wong. J. Am. Chem. Soc.
115, 2260 (1993).
References
1. (a) T. Feizi. Nature (London), 314, 53 (1985); (b) P.O. Livingston.
Curr. Opin. Immunol. 4, 624 (1992); (c) S. Hakomori and
Y. Zhang. Chem. Biol. 4, 97 (1997).
2. G.-J. Boons and A.V. Demchenko. Chem. Rev. 100, 4539
(2000).
8. V. Martichonok and G.M. Whitesides. J. Org. Chem. 61, 1702
(1996).
© 2002 NRC Canada