Chemoenzymatic synthesis of 2-azidoethyl-ganglio-oligosaccharides GD3, GT3, GM2, GD2, GT2, GM1, and GD1a

Article abstract of DOI:10.1016/j.carres.2005.06.008

We have synthesized several ganglio-oligosaccharide structures using glycosyltransferases from Campylobacter jejuni. The enzymes, α-(2→3/8)-sialyltransferase (Cst-II), β-(1→4)-N- acetylgalactosaminyltransferase (CgtA), and β-(1→3)- galactosyltransferase (CgtB), were produced in large-scale fermentation from Escherichia coli and further characterized based on their acceptor specificities. 2-Azidoethyl-glycosides corresponding to the oligosaccharides of GD3 (α-D-Neup5Ac-(2→8)-α-D-Neup5Ac-(2→3)-β-D-Galp- (1→4)-β-D-Glcp-), GT3 (α-D-Neup5Ac-(2→8)-α-D-Neup5Ac- (2→8)-α-D-Neup5Ac-(2→3)-β-D-Galp-(1→4)-β-D-Glcp-) , GM2 (β-D-GalpNAc-(1→4)-[α-D-Neup5Ac-(2→3)]-β-D-Galp- (1→4)-β-D-Glcp-), GD2 (β-D-GalpNAc-(1→4)-[α-D-Neup5Ac- (2→8)-α-D-Neup5Ac-(2→3)]-β-D-Galp-(1→4) -β-D-Glcp-), GT2 (β-D-GalpNAc-(1→4)-[α-D-Neup5Ac-(2→8) -α-D-Neup5Ac-(2→8)-α-D-Neup5Ac-(2→3)]-β-D-Galp- (1→4)-β-D-Glcp-), and GM1 (β-D-Galp-(1→3)-β-D-GalpNAc- (1→4)-[α-D-Neup5Ac-(2→3)]-β-D-Galp-(1→4) -β-D-Glcp-) were synthesized in high yields (gram-scale). In addition, a mammalian α-(2→3)-sialyltransferase (ST3Gal I) was used to sialylate GM1 and generate GD1a (α-D-Neup5Ac-(2→3)-β-D-Galp-(1→3)- β-D-GalpNAc-(1→4)-[α-D-Neup5Ac-(2→3)]-β-D-Galp- (1→4)-β-D-Glcp-) oligosaccharide. We also cloned and expressed a rat UDP-N-acetylglucosamine-4′epimerase (GalNAcE) in E. coli AD202 cells for cost saving in situ conversion of less expensive UDP-GlcNAc to UDP-GalNAc.

Full text of DOI:10.1016/j.carres.2005.06.008

Carbohydrate
RESEARCH
Carbohydrate Research 340 (2005) 1963–1972
Chemoenzymatic synthesis of 2-azidoethyl-ganglio-oligosaccharides
GD3, GT3, GM2, GD2, GT2, GM1, and GD1a
Ola Blixt,a,* Daniela Vasiliu,a Kirk Allin,a Nathan Jacobsen,a
a
a
a
b
´
Dawn Warnock, Nahid Razi, James C. Paulson, Stephane Bernatchez,
Michel Gilbertb and Warren Wakarchukb
aCarbohydrate Synthesis and Protein Expression Core D, Consortium for Functional Glycomics, The Scripps Research Institute,
Department of Molecular Biology, CB-248A, 10550 North Torrey Pines Road, 92037 La Jolla, United States
bInstitute for Biological Sciences, National Research Council of Canada, Ottawa, Ontario, Canada K1A 0R6
Received 24 March 2005; received in revised form 10 June 2005; accepted 13 June 2005
Available online 11 July 2005
Abstract—We have synthesized several ganglio-oligosaccharide structures using glycosyltransferases from Campylobacter jejuni. The
enzymes, a-(2!3/8)-sialyltransferase (Cst-II), b-(1!4)-N-acetylgalactosaminyltransferase (CgtA), and b-(1!3)-galactosyltransfer-
ase (CgtB), were produced in large-scale fermentation from Escherichia coli and further characterized based on their acceptor spec-
ificities. 2-Azidoethyl-glycosides corresponding to the oligosaccharides of GD3 (a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-b-D-
Galp-(1!4)-b-D-Glcp-), GT3 (a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)-b-D-Glcp-), GM2
(b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!3)]-b-D-Galp-(1!4)-b-D-Glcp-), GD2 (b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!8)-a-D-
Neup5Ac-(2!3)]-b-D-Galp-(1!4)-b-D-Glcp-), GT2 (b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!8)-a-D-Neu-
p5Ac-(2!3)]-b-D-Galp-(1!4)-b-D-Glcp-), and GM1 (b-D-Galp-(1!3)-b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!3)]-b-D-Galp-
(1!4)-b-D-Glcp-) were synthesized in high yields (gram-scale). In addition, a mammalian a-(2!3)-sialyltransferase (ST3Gal I)
was used to sialylate GM1 and generate GD1a (a-D-Neup5Ac-(2!3)-b-D-Galp-(1!3)-b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-
(2!3)]-b-D-Galp-(1!4)-b-D-Glcp-) oligosaccharide. We also cloned and expressed a rat UDP-N-acetylglucosamine-40epimerase
(GalNAcE) in E. coli AD202 cells for cost saving in situ conversion of less expensive UDP-GlcNAc to UDP-GalNAc.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Keywords: Ganglioside; Oligosaccharide; Chemoenzymatic; Synthesis; Glycosyltransferase
1. Introduction
apoptosis. They are also well-known tumor associated
antigens and active immunization using gangliosides
may suppress melanoma growth.1–6
Gangliosides are glycolipids that comprise a structurally
diverse set of sialylated molecules that are found in most
cells but are particularly abundant in neuronal tissues.
They have been found to act as receptors for growth fac-
tors, toxins, viruses, and to facilitate the attachment of
human melanoma and neuroblastoma cells. Specific
gangliosides are also present in early stages of human
neural development and affect major cellular processes
including proliferation, differentiation, survival, and
Despite the importance of the sialylated ganglioside
structures, methods for their efficient preparation have
been limited. The introduction of the sialic acid to a
glycolipid core structure has shown to be a daunting
task and complicated engineering with well executed
synthetic strategies is needed. Several elegant chemical
approaches have been reported over decades by various
groups7–10 as well as chemoenzymatic synthesis of
GM3.11–14 A recent report also demonstrated the in vivo
production of the glycan chain of GM1 and GM2 oligo-
saccharides using metabolically engineered bacteria.15,16
Until now, the synthesis of N-acetylneuraminic acid
*
Corresponding author. Tel.: +1 858 784 7538; fax: +1 858 784
7542; e-mail: olablixt@scripps.edu
0008-6215/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2005.06.008

1964
O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
(Neup5Ac) containing ganglio-oligosaccharides and
particularly the difficult formation of a a-(2!8)-linkage
between two Neup5Ac molecules has been the most
tedious.
assayed under similar conditions to give 67 U/L using
a-(2!3)-sialyllactose (16) and uridine-50diphospho-
N-acetylgalactosamine (UDP-GalNAc) as acceptor
and donor substrates, respectively. The CgtB enzyme
was assayed using GM2 (24) as the acceptor and
uridine-50diphosphogalactose (UDP-Gal) as the donor
substrate to obtain an activity of 17 U/L of the expres-
sion media. All enzymes were stable as frozen cell lysate
at ꢀ20 ꢁC for at least 6 months of storage.
Due to recent advances in molecular biology and char-
acterizations of various pathogens, several bacterial gly-
cosyltransferases have been identified that accomplish
transformations analogous to mammalian enzymes.
Because they can be produced in simple bacterial expres-
sion systems, they are increasingly being explored as po-
tential biochemical catalysts for glycosidic linkage
formations in the preparation of synthetically defined
carbohydrate structures.17–21 Several glycosyltransferase
genes from Campylobacter jejuni have been identified as
coding genes responsible for the expression of various
ganglioside related lipooligosaccharide (LOS) synthe-
sized by this pathogenic bacterium.22 Among these
genes, Cst-II, coding for a bifunctional a-(2!3/8)-sialyl-
transferase (a3/8SiaT), CgtA, coding for a b-(1!4)-N-
acetylgalactosaminyltransferase (b4GalNAcT), and CgtB,
coding for a b-(1!3)-galactosyltransferase (b3GalT),
have been cloned into Escherichia coli expression vectors
and expressed as active glycosyltransferases.23,24 We
have now further explored the large-scale production
of these enzymes and developed efficient synthetic
procedures for ganglio-oligosaccharides GD3 (a-D-
Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)-b-
D-Glcp-), GT3 (a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-
(2!8)-a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)-b-D-Glcp-),
GM2 (b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!3)]-b-
D-Galp-(1!4)-b-D-Glcp-), GD2 (b-D-GalpNAc-(1!4)-
[a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)]-b-D-Galp-(1!
4)-b-D-Glcp-), GT2 (b-D-GalpNAc-(1!4)-[a-D-Neu-
p5Ac-(2!8)-a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)]-b-
D-Galp-(1!4)-b-D-Glcp-), GM1 (b-D-Galp-(1!3)-b-D-
GalpNAc-(1!4)-[a-D-Neup5Ac-(2!3)]-b-D-Galp-(1!4)-
b-D-Glcp-), and GD1a (a-D-Neup5Ac-(2!3)-b-D-Galp-
(1!3)-b-D-GalpNAc-(1!4)-[a-D-Neup5Ac-(2!3)]-b-D-
Galp-(1!4)-b-D-Glcp-).
Multi-gram-scale synthesis requires stoichiometric
additives of generally expensive sugar nucleotide donor
substrates. Since current prices of UDP-GalNAc are
20 times that of UDP-GlcNAc, we produced a rat
UDP-N-acetylglucosamine-40epimerase (GalNAcE) that
converts less expensive UDP-GlcNAc to UDP-GalNAc
in situ. The rat epimerase gene, kindly provided by Karl
Johnson, Neose Technologies, was cloned into vector
pCWori+ and expressed into E. coli strain AD20225 with
the standard procedure reported previously.18 The cell
lysates were used directly in the enzymatic reaction mix-
tures and no endogenous activity was seen in control
reactions using cell lysate lacking the GalNAcE-produc-
ing plasmid (data not shown). Approximately 496 U/L
media was produced as determined using the assay con-
ditions described below. The GalNAcE enzyme also
catalyzes the epimerization of UDP-Glc to UDP-Gal
(174 U/L media). Therefore, the UDP-GalNAc epimer-
ase worked exemplary well in combination with the
CgtA and CgtB in preparative synthesis (see below).
2.2. a3/8SiaT and b4GalNAcT acceptor specificities
Bacterial glycosyltransferases typically have more re-
laxed substrate specificities than mammalian enzymes,
which can be used to advantage for various synthetic
applications.20 To further explore the utility of the Cst-
II and CgtA enzymes in preparative synthesis of various
ganglio-oligosaccharides, we performed a detailed oligo-
saccharide acceptor substrate specificity study (Table 1).
The Cst-II a-(2!3)-activity was measured using galacto-
sides at high acceptor substrate concentrations (5 mM)
and short incubation times (30 min) to minimize any
a-(2!8)-sialylation on formed a-D-Neup5Ac-(2!3)-
products. The mono-sialylated assay products were iso-
lated using Dowex-ion-exchange column as described
previously.17 Our data showed that neither terminal N-
acetylglucosaminide nor N-acetylgalactosaminide could
serve as acceptors. A spacer-derivatized galactoside (2)
showed almost 2-fold higher activity than free galactose
(1), suggesting the importance of beta configuration at
the anomeric center of the terminal galactose moiety.
Similar assumption can be made for lactose (7) and lac-
toside (8). No activity was observed with the detected
thiogalactoside (3), N-acetylgalactosamine (4 and 5),
or N-acetylglucosamine (6). As expected, lactosides (7
and 8) had the highest specificity of the compounds
2. Results and discussion
2.1. Large-scale production of recombinant enzymes
The genes encoding cst-II (a3/8SiaT), cgtA, (b4Gal-
NAcT) and cgtB (b3Gal) were cloned into the expres-
sion vector pCWori+ and expressed in E. coli strain
AD202 cell line.22,25 Large-scale fermentation (100 L)
gave high enzymatic activities from crude cell lysates
after passing cells through a microfluidizer and removal
of cell debris. The Cst-II was assayed based on a-(2!3)-
activity using lactose (7) as acceptor substrate and cyti-
dine-50monophospho-N-acetylneuraminic acid (CMP-
Neu5Ac) as donor substrate to give an enzyme activity
of 50 U/L of the expression media. The CgtA was

O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
1965
Table 1. Relative acceptor substrate specificities of Cst-II and CgtA
No. Acceptor
Relative specific activity (%)
Cst-II
CgtA
a-(2!3)-
a-(2!8)-
Dowex assays (Method A)
1
D-Gal
21
2
b-D-Galp-spa
38
<1
0
3
b-D-GalpSiPrb
D-GalpNAc
4
5
a-D-GalpNAc-sp
D-GlcpNAc
0
6
0
7
b-D-Galp-(1!4)-b-D-Glcp
b-D-Galp-(1!4)-b-D-Glcp-sp
b-D-Galp-(1!4)-b-D-GlcpSCrc
b-D-Galp-(1!4)-b-D-GlcpNAc
b-D-Galp-(1!4)-b-D-GlcpNAc-sp
b-D-Galp-(1!4)-b-D-GlcpNAc-(1!3)-b-D-Lacd-sp
[b-D-Galp-(1!4)-b-D-GlcpNAc-(1!3)]2-sp
[b-D-Galp-(1!4)-b-D-GlcpNAc-(1!3)]3-sp
100
123
8
0
8
9
10
11
12
13
14
64
88
27
10
5
Reversed phase (C18) assays (Method B)
15
16
17
18
19
20
21
22
23
24
25
a-D-Neup5Ac-(2!3)-b-D-GalpRe
63
100
199
116
43
95
100
84
8
a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)GlcpR
a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)GlcpNAcR
a-D-Neup5Ac-(2!6)-b-D-Galp-(1!4)GlcpR
a-D-Neup5Ac-(2!6)-b-D-Galp-(1!4)GlcpNAcR
a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-b-D-Lac-R
[a-D-Neup5Ac-(2!8)]2-a-D-Neup5Ac-(2!3)-b-D-Lac-R
a-D-Neup5Gc-(2!3)-b-D-Galp-(1!4)GlcpNAcR
b-D-Galp-(1!4)-b-D-GlcpNAc-(1!3)-b-D-Lac-R
a-D-Neup5Ac-(2!3)-[b-D-GalpNAc-(1!4)]-b-D-Lac-R
a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-[b-D-GalpNAc-(1!4)]-b-D-Lac-R
7
12
37
33
98
0
5
104
27
10
12
Relative specific activity for each compound is calculated as a percentage of the reference compound (7) according to the formula:
(value) = (acceptor/reference acceptor) · 100. Neutral acceptors were assayed by Method A and charged acceptors with Method B (see
Experimental).
a sp, OCH2CH2N3.
b SiPr, thio-iso-propyl.
c pSCr, para thio-cresyl.
d Lac, lactose.
e R, OCH2CH2NH(Fmoc).
tested. N-Acetyllactosamines (LacNAc) (10 and 11)
demonstrated excellent activity, but the extended
dimeric and trimeric LacNAc (13 and 14) showed an
approximate decrease of 9- and 18-fold, respectively,
compared to monomeric LacNAc (11).
ger oligosaccharides GD3 (20) and GT3 (21) showed a
significant reduced activity (10–20-fold), compared to
a-(2!3)-sialyllactose (16). A similar trend was also ob-
served with the a-(2!3)-activity of the extended di- and
trimeric-LacNAc structures above. Additionally, the N-
glycolylneuraminic acid (Neu5Gc) oligosaccharide (22)
was also a very good substrate and enabled the synthesis
of various N-glycolylneuraminic acid structures com-
monly found in the murine species.
We also investigated the specificity of the b4GalNAcT
(cgtA) with Fmoc acceptor substrates (15–25). The
CgtA enzyme did not show activity on non-sialylated
substrates as demonstrated with lactose (7)24 and lacto-
N-neotetraose (23) in contrast to the mammalian GM2
synthase that was reported capable in generating asi-
alo-GM2 structures.27 Sialic acid a-(2!3)-linked to ter-
minal galactosides (15–17) showed 10-fold higher
activity compared to a-(2!6)-sialosides (18, 19) and
approximately one third of the activity on the GD3
The Cst-II a-(2!8)-activity was measured with differ-
ent sialosides (15–25) tagged with a hydrophobic 9-flu-
orenyl methoxycarbonyl (Fmoc) functional group for
convenient solid-phase reversed phase C-18 SepPak
extraction assays.26 We found that products containing
up to four sialic acid units could bind to the C18 column
in our study (confirmed in parallel size exclusion chro-
matography assays, data not shown). The a-(2!8)-
activity study showed that both a-(2!3)-sialosides
(15–17) and a-(2!6)-sialosides (18, 19) were excellent
substrates for Cst-II. To confirm the a-D-Neup5Ac-
(2!8)-a-D-Neup5Ac-(2!6)-product formation, 19 was
synthesized in 100 mg quantities and verified by 1D
and 2D-NMR-spectroscopy (data not shown). The lar-

1966
O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
and GT3 oligosaccharides, compared to 16. We
speculate that 16 could be the preferred substrate for
the enzyme in C. jejuni LOS biosynthesis. The CgtA
glycosyltransferase could also utilize the sugar nucleo-
tide donor UDP-Gal to the same extent as UDP-
GalNAc (data not shown), and thus increase the
synthetic diversity of oligosaccharide derivatives.16
onstrates extensive a-(2!8) multi-sialylation activities.
Therefore, the synthesis of GD3 (27), GT3 (28), and
tetra-sialyllactose (29) required carefully controlled
conditions using restricted amounts of CMP-Neu5Ac.
2-Azidoethyllactoside (8) was elongated with Cst-II and
2.5 equiv of CMP-Neu5Ac17 to produce 2.4 g (34%) of
GD3 and 0.39 g (6%) of GT3 (in Scheme 1). GM3 (26)
was completely consumed and converted to GD3 and
GT3 using this molar ratio of CMP-Neu5Ac donor.
By increasing the molar ratio of CMP-Neu5Ac to
4 equiv, the reaction was driven to >80% conversion
of the newly formed GD3 to generate larger amounts
of GT3 (1.0 g, 20%). Smaller amounts of the tetra-sialic
2.3. Synthesis of GD3- (27), GT3- (28), and
poly-sialylated-oligosaccharides
To synthesize ganglio-oligosaccharide structures for
conjugation, the reducing end on lactose was equipped
with the 2-azidoethyl-functionality prior to enzymatic
modifications. These derivatives can be used as is in
competitive inhibition experiments, conjugated to vari-
ous supports or converted to other functional groups.
With excess amounts of CMP-Neu5Ac, the Cst-II dem-
acid
a-D-Neup5Ac-(2!8)-a-D-Neup5Ac-(2!8)-a-D-
Neup5Ac-(2!8)-a-D-Neup5Ac-(2!3)-b-D-Galp-(1!4)-
b-D-Glcp-(29) (0.25 g, 4%) were also isolated and traces
of higher sialylated fractions were also detected but not
purified. Products were subject to ion-exchange and size-
Neu5Ac
CTP
CMP-Neu5Ac-
synthetase
OH OH
OH
CMP-Neu5Ac (2.5-4 eqv.)
O
O
O
HO
HO
O
N3
OH
OH
a(2-3/8)-sialyltransferase
8
OH
OH
OH
O
HO
O
AcHN
O
O
O
O
HO
N3
OH
OH
HO
OH
OH
COOH
26, consumed in situ
OH
OH
OH
HO
OH
O
O
AcHN
HO
O
O
O
O
HO
OH
HO
OH
N3
OH
AcHN
O
O
COOH
27, 2.4g (34% at 2.5 eqv donor)
HO
OH
COOH
OH
OH
OH
HO
O
O
AcHN
O
O
O
O
HO
HO
N3
OH
OH
OH
HO
AcHN
HO
O
COOH
O
HO
OH
AcHN
O
28, 0.39 g (6% at 2.5 eqv donor)
COOH
O
1.0 g (20% at 4 eqv donor)
HO
OH
OH
COOH
OH
OH
OH
HO
OH
O
O
AcHN
HO
O
O
O
O
HO
HO
N3
OH
OH
AcHN
O
COOH
O
HO
OH
HO
AcHN
HO
O
COOH
29, 0.25g (4% at 4 eqv donor)
O
HO
OH
OH
O
AcHN
HO
COOH
O
OH
COOH
Scheme 1.

O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
1967
exclusion chromatography to give pure material (>90%)
per NMR and MS analysis.
2.5. Conclusion
This work demonstrates that efficient synthesis of com-
plex oligosaccharides representing the glycan chains of
gangliosides can be achieved using microbial glyco-
syltransferases. Detailed acceptor specificities of Cst-II
and CgtA revealed a relative broad substrate flexibility
that enabled various enzymatic strategies. Large-scale
fermentation in E. coli generated up to 5000 units of
active enzymes that were incorporated into gram-scale
synthesis of GD3, GT3, GM2, GD2, GT2, and GM1
ganglioside oligosaccharides. Expensive UDP-GalNAc
was generated in situ from less expensive UDP-GlcNAc
by the E. coli produced rat UDP-GalNAc-40-epimerase.
Furthermore, spacered GM1 was a substrate for a mam-
malian ST3Gal-I to generate GD1a, a structure that may
have an important therapeutic value in neural disease.
We are currently extending the ganglioside oligosaccha-
ride series with the GD1b and GD1c using the approach
described above. All structures have been linked to a con-
venient reducible azido-functional group that can be con-
jugated to proteins, solid-supports,30 lipids, or other
functional groups,31 and they are available free of charge
from CFG (http://www.functionalglycomics.org).
2.4. Synthesis of GM2- (30), GD2- (31), GT2- (32),
GM1- (33), and GD1a- (34) oligosaccharides
To extend the series of ganglioside oligosaccharide
structures, GM3 (26),18 GD3 (27), and GT3 (28) were
further branched by using the one-pot mixture of CgtA
(b4GalNAcT), UDP-GlcNAc, and GalNAcE to effi-
ciently generate the GM2 (30), GD2 (31), and GT2
(32) derivatives in high yields (typical 80–90%) (Scheme
2). We have further elongated GM2 (30) to GM1 (33) by
using the CgtB (b3GalT), UDP-glucose, and GalNAcE
in high yields (0.5 g, 86%). Recently, chemical synthesis
of similar aminoethyl glycoside was reported, in which
total amounts of 38 mg, in a yield of 30%, were pro-
duced.7 Starting from lactoside (8), we enzymatically
generated gram-scale amounts of GM1 in >75% total
yield within a week.
GD1a is one of the several structures identified as
target epitopes for immunological responses in Guil-
24
lian–Barre syndrome. We were also able to generate
´
spacered GD1a (34) in a simple procedure using a a-
(2!3)-sialyltransferase (ST3Gal-I). The ST3Gal-I is an
enzyme known for its substrate specificity on O-glycans,
where it creates sialylated T-antigen but it was also re-
ported to have activity for ganglioside structures such
as GM1 and GD1b.28 Here, we used a recombinant por-
cine ST3Gal-I expressed in a baculovirus system29 for
sialylation of 2-azidoethyl GM1 (33) to synthesize
GD1a (34) in high yield (Scheme 2).
3. Experimental
3.1. General methods
Concentration of solvents was performed under reduced
pressure at <40 ꢁC bath temperature. NMR spectra
OH
β4GalNAcT
GalNAcE
UDP-GlcNAc (1.8 eqv)
OH
HO
O
O
OH
OH
NHAc
R
O
O
26, 27 or 28
O
O
HO
O
N3
OH
OH
0.5-0.6 g (80-85 %)
GM2 30 R = α-D-Neup5Ac
GD2 31 R = α-D-Neup5Ac-(2-8)α-D-Neup5Ac
GT2 32 R = α-D-Neup5Ac-(2-8)-α-D-Neup5Ac-(2-8)α-D-Neup5Ac
1. β3GalT
OH
OH
O
GalNAcE
OH
OH
UDP-GlcNAc (1.8 eqv)
2. pST3Gal-I
CMP-Neu5Ac (2 eqv.)
O
O
O
O
HO
O
OH
OH
R
NHAc
OH
O
O
AcHN
O
30
O
HO
O
N3
OH
OH
OH
OH
HO
COOH
GM1 33 R = H (0.5 g, 86 %)
GD1a 34 R = α-D-Neup5Ac (56 mg, 88%)
Scheme 2.

1968
O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
were recorded at 25 ꢁC using a Bruker DRX-500 or 600
spectrometer. The following reference signals were used:
acetone 2.225 ppm (1H in D2O), acetone 29.7 ppm (13C
in D2O). ESI-TOF high-accuracy MS spectra were re-
corded with an LC MSD TOF (Agilent Technologies)
using dihydroxybenzoic acid as matrix. Thin-layer
chromatography was performed on Kieselgel 60 F254
Fertigplatten (Merck, Darmstadt, Germany). After
development with appropriate eluants (ethyl acetate–
MeOH–acetic acid–water by volume 10—3:3:3:2), spots
were visualized by UV light and/or by dipping in 5%
sulfuric acid in ethanol, followed by charring. Water
was purified from a NanoPure Infinity Ultrapure water
system (Barnstead/Thermolyne, Dubuque, Iowa, USA)
and was degassed by vacuum treatment before use.
Compounds 10–12 were synthesized by chemoenzymatic
methods as described previously18 and compounds 2
and 9 were prepared as reported.32 Compounds 13
and 14 were synthesized by consecutive reaction of
b-(1!3)-N-acetylglucoaminyltransferase33 and b-(1!4)-
galactosyltransferase18 and will be described elsewhere.
Radioactive labeled sugar nucleotides were from NEN
Life Science Products Inc. (Boston, MA). The radioac-
tive nucleotide-sugars were diluted with unlabeled
UDP-sugars (Sigma) to the desired specific radioactiv-
ity. All other oligosaccharide substrates were purchased
from Sigma Chemical Co. (St. Louis, MO).
with NdeI and SalI and cloned in pCWori+ giving the
constructs pCJL-30 (CgtA), pCJL-20 (CgtB), and
pCST-68 (Cst-II). The cst-II gene in construct CST-68
also includes a point mutation (Ile53Ser) that enhances
the production of sialyltransferase activity.23
3.4. Production of enzymes; Cst-II, CgtA, CgtB, and
GalNAcE in E. coli
All the desired enzymes were overexpressed in E. coli
AD202 and produced in a 100 L Fermentor (Braun
Fermentor). Bacteria were cultured in 2xYT/Ampicillin
(150 lg/mL) and induced with iso-propyl-thiogalactopy-
ranoside (IPTG, 1 mM) with the conditions optimal for
each enzyme. The constructs expressing the bacterial
enzymes from C. jejuni (CgtA, CgtB, and Cst-II) demon-
strated a slow growth and required incubation up to
24 h at 37 ꢁC for CgtA, CgtB and 25 ꢁC for Cst-II to
reach the optimal OD of 0.3–0.6 at A600 for induction
with IPTG. After induction, temperature decreased to
20 ꢁC for CgtA and CgtB, whereas Cst-II continued to
grow at 25 ꢁC to reach the maximum growth of bacteria.
The rat GalNAcE was incubated at 37 ꢁC for 4 h to reach
the optimal OD = 0.3–0.6 at A600 for induction. The
incubation continued at 37 ꢁC to complete the growth.
Cells were harvested by spinning at 5000g for 30 min.
The cell paste was weighed and re-suspended in the
appropriate buffer (see assay conditions) in a 1 g/mL
ratio of cell weight to the buffer. The cell paste was ali-
quoted to the desired volume and stored at ꢀ20 ꢁC. To
release the enzyme, cells were lysed with a microfluidizer
procedure by adding 1/1 volume of HEPES buffer to the
resuspended cell paste. The released enzymes were used
as a crude suspension in the assays.
3.2. Construction of UDP-GalNAc-epimerase expression
plasmid
Plasmid pACGP67B-GalNAcE was amplified in E. coli
DH5a and the GalNAcE gene was subcloned into the
expression plasmid pCWori+ using the restriction sites
NdeI and EcoRI of the vector. The 50 primer (gal-
NacE-5p) was designed with 30 nucleotides containing
the NdeI restriction site. The start codon, ATG, is in-
cluded within the NdeI restriction site (bold italics) (50-
GTGAGACATATGGAGGAGAAGGTGCTCGTC-30).
The 30 primer (gal NAcE-3p) was a 30 mer containing
the EcoRI restriction site (bold) followed by the com-
plement (TCA) of the stop codon, TGA, (bold italics);
50-GAACCCCGCGTACCGACTCTTAAGTTGATC-30.
The resulting vector, pCWori-GalNAcE, was first trans-
formed into E. coli DH5a (cloning host) and subse-
quently to the bacterial cells AD20225 for expression.
3.5. Enzyme activity assays
The assays were generally conducted in a final volume of
100 lL at 37 ꢁC for 30 min with the desired enzyme
(10 lL of crude mixture), MnCl2 (100 mM), bovine ser-
um albumin (1%), and the appropriate acceptor/donor
substrate for each enzyme as follows: Cst-II assays were
performed with lactose (100 mM) and [C14]-CMP sialic
acid (20 nmol, 2500 specific activity) in HEPES buffer
(50 mM, pH 7.5). CgtA assays were carried out with
sialyllactose (5 mM) and [H3]-UDPGalNAc (50 nmol,
1000 specific activity) in HEPES (50 mM, pH 7.0). CgtB
assays were performed with GM2 (20 mM) and [H3]-
UDPGal (50 nmol, 1000 specific activity) in MES
(20 mM, pH 6.0). The enzyme activity of the GalNAcE
was indirectly detected by GM2 formation using both
epimerase and CgtA in the presence of UDP-GlcNAc
(150 mM) as the donor and sialyllactose (5 mM) as an
acceptor substrate. The CgtA and CgtB reactions were
terminated by adding the ice-cold phosphate buffer
(1 mL, 10 mM) to the reaction and were analyzed on
3.3. Construction of CgtA, CgtB, and Cst-II expression
plasmids
The cgtA gene was amplified from C. jejuni ATCC 43456
(GenBank accession # AF401528), the cgtB gene was
amplified from C. jejuni NCTC 11168 (GenBank acces-
sion # AL139077, Cj1139c), and the cst-II gene was
amplified from C. jejuni OH4384 (GenBank accession
# AF130984). All three PCR products were digested

O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
1969
phosphate-Dowex-ion exchange column (method A).
The Cst-II reaction was terminated with 1 mL of sodium
formate (30 mM) and further analyzed on formate-
Dowex-ion exchange.
The supernatant was diluted to 50 mL with water and
loaded onto a SepPak cartridge (10 g). The column was
washed with water and products were eluted with a gra-
dient of MeOH (0–50%). Appropriate fractions were col-
lected and lyophilized to dryness. Compounds were
identified by MS and typical yields of 50–80% were ob-
tained with a purity >90% according to TLC.
3.6. Acceptor substrate specificity assays
3.6.1. Method A—neutral acceptors. Enzyme activity
for different neutral acceptors was assayed at 25 ꢁC for
30 min in a final volume of 100 lL containing HEPES
(50 mM, pH 7.5), [H3]-UDP-GalNAc (2 mM, 1.7 mCi/
mol) (0.5 mM), or [C14]-CMP-Neu5Ac acceptor
(5 mM) final specific activity (0.70 Ci/mol) and diluted
enzyme. Reactions were terminated by the addition of
ice cooled EDTA (10 mM, 0.9 mL). In all cases, reac-
tions were limited to transfer of less than 10% of the
sugar nucleotide donor. The reaction mixtures were then
passed through Pasteur pipet columns of Dowex resin
(1-·8, 200–400 mesh, formate form, 0.5 · 5 cm, 1 mL).
The columns were washed with sodium formate
(20 mM, pH 7.0, 1.3 mL) and the effluents were directly
collected in scintillation vials. The radioactivity was
assayed as described above.
3.8. Synthesis of 2-azidoethyl O-b-D-galactopyranosyl-
(1!4)-b-D-glucopyranoside (8)
O-(2,3,4,6-Tetra-O-acetyl-b-D-galactopyranosyl)-(1!4)1,
2,4,6-tetra-O-acetyl-b-D-glucopyranose (90 g, 130 mmol)
and 2-azidoethanol34, (23 g, 260 mmol, 2 equiv) were
dissolved in dry CH2Cl2 (100 mL) containing molecular
˚
sieves (4 A) and cooled on ice. After 30 min stirring,
boron trifluoride etherate (23 g, 160 mmol, 2 equiv)
was added and the reaction was kept under a nitrogen
atmosphere for 12 h while it was slowly reaching room
temperature. The mixture was washed with water (1 L)
and sodium hydrogen carbonate (1 M, 2 · 1 L). The
organic layer was dried with magnesium sulfate, filtered,
and evaporated to dryness. The crude mixture was
dissolved in dry CH2Cl2 (100 mL), cooled in ice for
30 min, and then hydrazine acetate (6.0 g, 65 mol,
0.5 equiv) was added to hydrolyze un-reacted per-acetyl-
ated lactose. After 30 min, the mixture was washed with
sodium hydrogen carbonate (1 M, 2 · 1 L). The organic
layer was dried with magnesium sulfate, filtered, and
evaporated to dryness. The crude mixture was loaded
onto a silica gel column (10 · 40 cm) packed in toluene
and the product was eluted with toluene–ethyl acetate
(4:1 by volume, 1 vol), toluene–ethyl acetate (3:1 by vol-
ume, 1 vol), and toluene–ethyl acetate (2:1 by volume,
1 vol). Appropriate fractions containing 8 were collected
and evaporated to dryness to give 38 g of 2-azidoethyl
O-(2,3,4,6-tetra-O-acetyl-b-D-galactopyranosyl)-(1!4)-
2,3,6-tri-O-acetyl-b-D-glucopyranoside. The clear syrup
was dissolved in dry MeOH (200 mL) and de-acetylated
by adding sodium methoxide (20 mL, 0.5 M) at room
temperature overnight. The mixture was neutralized
with MeOH-washed Dowex-50-H+, filtered, and evapo-
rated to approximately 100 mL. Diethyl ether (100 mL)
was added to initiate crystallization. After filtration and
3.6.2. Method B—sialylated hydrophobic acceptors.
The assays were conducted in a final volume of 100 lL
at 25 ꢁC for 30 min containing the acceptor (5 mM)
and [H3]-UDP-GalNAc (2 mM, 1.7 mCi/mol) (0.5 mM)
or [14C]-CMP-Neu5Ac in sodium HEPES buffer (50
mM, pH 7.5), MnCl2 (10 mM), and bovine serum albu-
min (1%). Reactions were initiated by addition of the
appropriate crude glycosyltransferase diluted in buffer
containing bovine serum albumin (1%). Reactions were
terminated by addition of ice-cooled EDTA buffer
(10 mM, 0.9 mL). In all cases, reactions were limited to
transfer of less than 10% of the sugar nucleotide donor.
The mixture was applied to Sep-Pak C-18 reverse-phase
cartridges (500 mg) (Isolute, International Sorbent Tech-
nology, Mid-Glamorgan, UK), which were washed with
5 mL of water. Radioactive products were eluted with
MeOH (2 mL) and the eluate was evaporated to dryness
and radioactivity was counted as described above.
3.7. General procedure for Fmoc-derivatization of
2-azidoethyl glycosides (15–25)
drying, white crystals of 8 were obtained (25 g, 63 mmol,
20
47%). Mp 145–147 ꢁC; ½aꢁ +36.8 (c 0.025, water).
D
2-Azidoethyl glycoside (50 lmol) and Pd/C (10%, 50 mg)
were suspended in water (5 mL) and ethanol (5 mL) and
the mixture was vigorously stirred under a hydrogen
atmosphere for 1 h at room temperature. When TLC
(ethyl acetate–acetic acid–MeOH–water, 6:3:3:2 by vol-
ume) indicated complete reduction of azide to amine,
sodium hydrogen carbonate was added (1 M, 2 mL)
followed by 9-fluorenyl methylcarbonylchloroformate
(2 equiv). After 30 min reaction at room temperature,
the mixture was centrifuged to remove solid particles.
Selected 1H NMR (500 MHz, D2O), d (ppm): 4.53
(d, 1H, J = 8 Hz, Glc H-1), 4.45 (d, 1H, J = 8 Hz, Gal
H-1), 4.06–4.04 (m, 1H, OCH2CH2N3), 3.98 (dd, 1H,
Gal H-4), 3.92 (d, 1H, Glc H-4), 3.36–3.33 (m, 1H,
OCH2CH2N3). Selected13C NMR (600 MHz, D2O), d
(ppm): 102.57, 101.81, 77.95, 75.00, 74.46, 73.98,
2-Azidoethanol is potentially explosive and adequate precaution
should be taken when handling this reagent.

1970
O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
72.41, 72.16, 70.60, 68.19, 60.68, 59.69, 50.17. ESI-TOF
high-accuracy MS m/z calculated for (M+H), 412.1567;
found, 412.1533.
50% GT3 oligosaccharide, 20% GD3 oligosaccharide,
and <10% lactoside as visualized by TLC), the mixture
was centrifuged and loaded onto a column of Sephadex
G15 (5 · 170 cm) equilibrated, packed, and eluted with
water. Appropriate fractions, containing all products
and nucleotide sugar, were collected (1.2 L) and loaded
onto a Dowex 1 · 8-400 formate ion exchange resin
(30 · 2.5 cm). Compound 8, GM3, oligosaccharide,
and free Neu5Ac were in the void fractions, whereas
GD3 oligosaccharide, GT3 oligosaccharide, and nucleo-
tides absorbed to the resin. The column was washed
with water (500 mL) followed by elution with a gradient
of aqueous sodium formate (20–80 mM, pH 7.5, 2.5 L).
Fractions containing each product were concentrated by
lyophilizing and further purified by size exclusion chro-
matography (Sephadex G15, 5 · 170 cm, packed and
eluted with 5% aqueous n-BuOH) to give 28 (1.0 g,
20%) and tetra-sialo-lactoside 29 (0.25 g, 4%). GT3 oli-
3.9. Synthesis of GD3 oligosaccharide (27)
Compound 8 (2 g, 5 mmol) and crude CMP-Neu5Ac18
(25 g, 25 mmol, 46% by weight) were dissolved in cacodyl-
ate buffer (200 mM, 50 mL, pH 7.5) containing MnCl2
(40 mM). The Cst-II enzyme (24 U) was added and the
reaction was slowly stirred at room temperature for
48 h. When 95% of the starting material was converted
to products (10% GM3 oligosaccharide, 60% GD3 oligo-
saccharide, and 25% GT3 oligosaccharide approximately
as visualized by TLC), the mixture was centrifuged and
loaded onto a column of Sephadex G15 (5 · 170 cm)
equilibrated and eluted with water. Appropriate frac-
tions, containing all products and nucleotide sugar, were
collected (1.2 L) and loaded onto a Dowex 1 · 8-400 for-
mate ion exchange resin (30 · 2.5 cm). Compound 8,
GM3, and free Neu5Ac were in the void fractions whereas
GD3 oligosaccharide, GT3 oligosaccharide, and nucleo-
tides absorbed to the resin. The column was washed with
water (500 mL) followed by a gradient of aqueous sodium
formate (20–80 mM, pH 7.5, 2.5 L). Fractions containing
each product were concentrated by lyophilizing and each
oligosaccharide was further purified by size exclusion
chromatography (Sephadex G15, 5 · 170 cm and packed
and eluted with 5% aqueous n-BuOH) to give 27 (2.42 g,
34%) and 28 (0.39 g, 6%). GD3 oligosaccharide 27
1
gosaccharide 28 selected H NMR (600 MHz, D2O), d
(ppm): 4.53 (d, 2H, J = 8 Hz, Glc H-1, Gal H-1), 4.07–
4.05 (2m, 2H, OCH2CH2N3), 3.37–3.34 (m, 1H,
OCH2CH2N3), 2.77 (dd, 2H, J = 4 Hz, NeuAc H-3eq),
2.69 (dd, 1H, J = 4 Hz, NeuAc H-3eq), 2.07 (s, 6H,
NHCOCH3), 2.03 (s, 3H, NHCOCH3), 1.75 (t, 2H,
J = 12 Hz, NeuAc H-3ax), 1.70 (t, 1H, J = 12 Hz,
NeuAc H-3ax). Selected 13C NMR (600 MHz, D2O):
174.48, 102.12, 101.65, 77.89, 77.38, 74.99, 74.71,
74.33, 73.72, 73.22, 73.00, 72.27, 72.10, 71.20, 69.74,
68.82, 68.75, 68.55, 68.03, 68.03, 67.98, 67.88, 67.68,
67.62, 67.51, 67.45, 67.05, 66.93, 66.46, 62.03, 60.89,
60.57, 60.31, 59.47, 51.81, 51.21, 51.13, 49.99, 40.50,
39.89, 39.66, 39.14, 38.71, 21.91, 21.63, 21.50. ESI-
TOF high-accuracy MS m/z calculated for (M+Na),
1351.3888; found, 1351.3858.
1
selected H NMR (600 MHz, D2O), d (ppm): 4.53 (d,
2H, J = 8 Hz, Glc H-1, Gal H-1), 4.06–4.04 (2m, 2H,
OCH2CH2N3), 3.35–3.33 (m, 1H, OCH2CH2N3), 2.78
(dd, 1H, J = 4 Hz, NeuAc H-3eq), 2.68 (dd, 1H,
J = 4 Hz, NeuAc H-3eq), 2.07 (s, 3H, NHCOCH3), 2.03
(s, 3H, NHCOCH3), 1.75 (t, 2H, J = 12 Hz, NeuAc
H-3ax). Selected 13C NMR (600 MHz, D2O), d (ppm):
102.32, 101.85, 77.92, 77.55, 75.13, 74.89, 74.51, 73.89,
73.71, 72.45, 72.32, 74.45, 68.95, 68.19, 67.60, 67.11,
62.20, 62.10, 61.29, 61.26, 60.79, 59.61, 51.92, 51.40,
50.18, 40.19, 39.40, 21.99, 21.70. ESI-TOF high-accuracy
MS m/z calculated for (M+H), 1038.3114; found,
1038.3098.
1
Tetra-sialyllactoside 29 selected H NMR (600 MHz,
D2O), d (ppm): 4.53 (d, 1H, J = 8 Hz, Glc H-1), 4.51
(d, 1H, J = 8 Hz, Gal H-1), 4.07–4.05 (2m, 2H,
OCH2CH2N3), 3.36–3.33 (m, 1H, OCH2CH2N3), 2.76
(dd, 1H, J = 4 Hz, NeuAc H-3eq), 2.68 (dd, 3H,
J = 4 Hz, NeuAc H-3eq), 2.08 (s, 6H, NHCOCH3),
2.07 (s, 3H, NHCOCH3), 2.03 (s, 3H, NHCOCH3),
1.74 (q, 4H, J = 12 Hz, NeuAc H-3ax).
13C NMR (600 MHz, D2O), d (ppm): 174.65, 174.58,
102.30, 101.83, 78.14, 77.55, 77.41, 75.17, 74.89, 74.52,
73.89, 73.42, 73.16, 72.86, 72.45, 72.27, 71.40, 69.24,
69.09, 68.94, 68.74, 68.55, 68.22, 68.17, 67.92, 67.87,
67.68, 67.12, 62.24, 61.17, 61.01, 60.86, 60.76, 59.64,
52.09, 52.02, 51.99, 51.39, 50.175, 40.08, 39.72, 39.30,
22.11, 21.99, 21.70. ESI-TOF high-accuracy MS m/z cal-
culated for (M+Na), 1620.5015; found, 1620.5019.
3.10. Synthesis of GT3 oligosaccharide (28) and tetra-
sialyllactoside (29)
Compound 8 (1.5 g, 3.6 mmol) and crude CMP-
Neu5Ac18 (20 g, 46% by weight, 14.6 mmol) were dis-
solved in cacodylate buffer (200 mM, 50 mL, pH 7.5)
containing MnCl2 (40 mM), then Cst-II (20 U) was
added, and the reaction was slowly stirred at room tem-
perature for 48 h. After 4 h, 50% GD3 oligosaccharide
and 10% GT3 oligosaccharide were formed and 30%
of 8 remained. When <90% of the starting material
was converted to products (15% tetra-sialyllactoside,
3.11. General procedure for the synthesis of GM2 (30),
GD2 (31), and GT2 (32) oligosaccharides
Sialylated acceptors 26–28 (0.70 mmol) and UDP-Gal-
NAc (1.1 mmol, 1.8 equiv) were dissolved in Tris

O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
1971
(100 mM, 10 mL, pH 7.5) containing MnCl2 (40 mM),
then NAD+ (0.3 mM), CgtA (30 U), and GalNAcE
(800 U) were added, and the reaction was kept at
37 ꢁC for 24 h. When >95% of the starting material
was converted to product, the mixture was centrifuged
racy MS m/z calculated for (M+H), 1554.4682; found,
1554.4675.
3.12. Synthesis of GM1 oligosaccharide (33)
and loaded onto
a
column of Sephadex G15
Compound 30 (500 mg, 0.552 mmol) and UDP-Glc
(404 mg, 0.662 mmol) were dissolved in cacodylate buf-
fer (200 mM, 1 mL, pH 7.5) containing MnCl2 (40 mM),
then CgtB (3U) and GalNAcE (30U) were added, and
the reaction was slowly stirred at room temperature
for 48 h. More than 95% of the starting material was
converted to product and the mixture was centrifuged
(2.5 · 170 cm) equilibrated and eluted with 5% n-BuOH
in water. Appropriate fractions, containing the product,
were collected by lyophilization and loaded onto a
Sephadex G15 (1.6 · 170 cm) column and eluted with
water to give 30, 31, or 32 (580–650 mg, 80–85%).
GM2 oligosaccharide 30, Selected 1H NMR
(600 MHz, D2O), d (ppm): 4.74 (d, 1H, J = 8 Hz, Gal-
NAc H-1), 4.54 (d, 1H, J = 8 Hz, Glc H-1), 4.53 (d,
1H, J = 8 Hz, Gal H-1), 4.06–4.04 (2m, 2H,
OCH2CH2N3), 3.36–3.33 (m, 1H, OCH2CH2N3), 2.66
(dd, 1H, J = 4 Hz, NeuAc H-3eq), 2.03 (s, 3H, NH-
COCH3), 2.017 (s, 3H, NHCOCH3 of GalNAc), 1.92
(t, 1H, J = 12 Hz, NeuAc H-3ax). Selected 13C NMR
(600 MHz, D2O) d (ppm): 102.43, 102.26, 101.83,
78.20, 76.84, 74.57, 74.38, 74.12, 74.01, 73.68, 72.74,
72.36, 71.95, 70.94, 69.69, 68.38, 68.21, 67.67, 67.44,
74.57, 74.12, 62.50, 60.84, 60.22, 59.74, 56.10, 52.00,
51.25, 50.19, 36.58, 22.27, 21.71. ESI-TOF high-accu-
racy MS m/z calculated for (M+H), 928.3134; found,
929.3174.
GD2 oligosaccharide 31, selected 1H NMR
(600 MHz, D2O), d (ppm): 4.70 (d, 1H, GalNAc H-1),
4.53 (d, 1H, J = 8 Hz, Glc H-1), 4.51(d, 1H, J = 8 Hz,
Gal H-1), 4.06–4.04 (m, 1H, OCH2CH2N3), 3.35–3.33
(m, 1H, OCH2CH2N3), 2.77 (dd, 1H, J = 4 Hz, NeuAc
H-3eq), 2.68 (dd, 1H, J = 4 Hz, NeuAc H-3eq), 2.07
(s, 3H, NHCOCH3), 2.04 (s, 3H, NHCOCH3), 2.03 (s,
3H, NHCOCH3), 1.75 (m, 2H, J = 12 Hz, NeuAc H-
3ax). Selected 13C NMR (600 MHz, D2O) d (ppm):
102.40, 101.83, 81.20, 77.99, 77.87, 75.63, 74.56, 74.18,
73.91, 73.39, 72.72, 72.38, 72.31, 71.42, 71.04, 70.52,
69.37, 68.92, 68.87, 68.17, 67.78, 67.37, 61.13, 60.60,
60.30, 59.70, 59.64, 59.57, 52.02, 51.47, 51.31, 50.18,
49.81, 40.19, 40.15, 40.06, 38.89, 22.24, 22.01, 21.70.
ESI-TOF high-accuracy MS m/z calculated for
(M+H), 1241.3908; found, 1241.3892.
and loaded onto
a
column of Sephadex G15
(5 · 170 cm) equilibrated and eluted with water. The
fractions containing product were concentrated by
lyophilizing to give pure 33 (508 mg, 0.475 mmol,
86%). Selected 1H NMR (600 MHz, D2O), d (ppm):
4.78 (d, 1H, J = 8 Hz, GalNAc H-1), 4.54 (d, 2H,
J = 8 Hz, Gal H-1), 4.53 (d, 1H, J = 8 Hz, Glc H-1),
4.06–4.04 (m, 2H, OCH2CH2N3), 3.36–3.33 (m, 1H,
OCH2CH2N3), 2.66 (dd, 1H, J = 4 Hz, NeuAc H-3eq),
2.03 (s, 3H, NHCOCH3), 2.017 (s, 3H, NHCOCH3 of
GalNAc), 1.92 (t, 1H, J = 12 Hz, NeuAc H-3ax).
Selected 13C NMR (600 MHz, D2O) d (ppm): 174.63,
174.39, 173.75, 104.37, 102.20, 101.78, 101.27, 79.94,
78.15, 76.76, 74.50, 74.40, 73.97, 73.68, 72.70, 72.30,
72.10, 71.90, 70.29, 69.64, 68.34, 68.20, 68.16, 67.62,
67.52, 62.43, 60.72, 60.55, 60.22, 59.68, 51.21, 50.81,
50.15, 36.52, 29.85, 22.20, 21.66. ESI-TOF high-accu-
racy MS m/z calculated for (M+H), 1090.3663; found,
1090.3672.
3.13. Synthesis of GD1a oligosaccharide (34)
Compound 33 (50 mg, 46.8 lmol) and crude CMP-
Neu5Ac18 (108 mg, 93.6 lmol, 55% by weight) were dis-
solved in Tris buffer (100 mM, 2 mL, pH 7.5) containing
MnCl2 (20 mM), then ST3Gal I (2U) was added, and
the reaction was slowly stirred at room temperature
for 12 h. The conversion to product seemed complete
by TLC. The mixture was centrifuged and loaded onto
a column of Sephadex G15 (5 · 170 cm) equilibrated
and eluted with water. Fractions containing product
GT2 oligosaccharide 32, selected 1H NMR
(600 MHz, D2O), d (ppm): 4.71 (d, 1H, J = 8 Hz, Gal-
NAc H-1), 4.53 (d, 2H, J = 8 Hz, Glc H-1), 4.50 (d,
1H, J = 8 Hz, Gal H-1), 2.77 (dd, 2H, J = 4 Hz, NeuAc
H-3eq), 2.68 (dd, 2H, J = 4 Hz, NeuAc H-3eq), 2.05 (m,
12H, NHCOCH3), 1.74 (m, 3H, J = 12 Hz, NeuAc H-
3ax). Selected 13C NMR (600 MHz, D2O) d (ppm):
174.64, 174.59, 174.46, 102.41, 101.81, 77.92, 77.34,
76.82, 75.74, 74.78, 74.50, 74.42, 74.36, 74.14, 74.100,
73.97, 73.89, 73.66, 73.45, 72.99, 72.71, 72.47, 72.34,
72.30, 71.93, 71.77, 71.70, 71.40, 70.90, 70.50, 69.91,
60.81, 60.48, 60.26, 68.16, 68.04, 67.63, 67.41, 67.36,
67.19, 66.54, 62.21, 62.12, 60.58, 51.97, 51.32, 50.17,
40.67, 38.86, 22.21, 21.82, 21.70. ESI-TOF high-accu-
1
were lyophilized to give 34 (56 mg, 88%). Selected H
d (ppm): 4.61 (d, 1H,
NMR (600 MHz, D2O)
J = 8 Hz, GalNAc H-1), 4.54 (d, 2H, J = 8 Hz, Gal
H-1), 4.52 (d, 1H, J = 8 Hz, Glc H-1), 2.75 (dd, 1H,
J = 4 Hz, NeuAc H-3eq), 2.68 (dd, 1H, J = 4 Hz, NeuAc
H-3eq), 2.03 (d, 6H, NHCOCH3), 2.00 (s, 3H, NH-
COCH3 of GalNAc), 1.91 (t, 1H, J = 12 Hz, NeuAc
H-3ax), 1.80 (t, 1H, J = 12 Hz, NeuAc H-3ax). Selected
13C NMR (600 MHz, D2O) d (ppm): 174.65, 174.35,
173.70, 104.15, 102.25, 101.81, 101.08, 99.38, 80.14,
78.17, 76.51, 75.13, 74.45, 74.29, 73.99, 73.73, 72.70,
72.44, 72.34, 71.48, 69.62, 68.78, 68.36, 68.20, 68.15,
67.78, 67.69, 67.41, 67.06, 62.47, 62.15, 60.78, 60.60,

1972
O. Blixt et al. / Carbohydrate Research 340 (2005) 1963–1972
16. Antoine, T.; Priem, B.; Heyraud, A.; Greffe, L.; Gilbert,
M.; Wakarchuk, W. W.; Lam, J. S.; Samain, E. Chembio-
chem 2003, 4, 406–412.
17. Blixt, O.; Allin, K.; Pereira, L.; Datta, A.; Paulson, J. C. J.
Am. Chem. Soc. 2002, 124, 5739–5746.
60.25, 59.71, 51.25, 50.70, 50.19, 39.31, 36.86, 22.26,
21.71. ESI-TOF high-accuracy MS m/z calcd for
(M+H), 1403.4358; found, 1403.4382.
18. Blixt, O.; Brown, J.; Schur, M.; Wakarchuk, W.; Paulson,
J. C. J. Org. Chem. 2001, 66, 2442–2448.
19. Fang, J.; Chen, X.; Zhang, W.; Janczuk, A.; Wang, P. G.
Carbohydr. Res. 2000, 329, 873–878.
Acknowledgments
This work was funded by NIGMS and Consortium for
Functional Glycomics GM62116 and GM60938. We
thank NEOSE Technologies for providing the plasmid
pACGP67B-GalNAcE and Tokyo Research Laborato-
ries, Kyowa Hakko Kogyo Co. Ltd. for providing
UDP-GlcNAc.
20. Johnson, K. F. Glycoconj. J. 1999, 16, 141–146.
21. Endo, T.; Koizumi, S.; Tabata, K.; Ozaki, A. Appl.
Microbiol. Biotechnol. 1999, 53, 257–261.
22. Gilbert, M.; Brisson, J.-R.; Karwaski, M.-F.; Mich-
niewicz, J.; Cunningham, A.-M.; Wu, Y.; Young, N. M.;
Wakarchuk, W. W. J. Biol. Chem. 2000, 275, 3896–3906.
23. Chiu, C. P.; Watts, A. G.; Lairson, L. L.; Gilbert, M.;
Lim, D.; Wakarchuk, W. W.; Withers, S. G.; Strynadka,
N. C. Nat. Struct. Mol. Biol. 2004, 11, 163–170.
24. Gilbert, M.; Karwaski, M. F.; Bernatchez, S.; Young, N.
M.; Taboada, E.; Michniewicz, J.; Cunningham, A. M.;
Wakarchuk, W. W. J. Biol. Chem. 2002, 277, 327–337.
25. Akiyama, Y.; Ito, K. Biochem. Biophys. Res. Commun.
1990, 167, 711–715.
References
1. Kannagi, R.; Izawa, M.; Koike, T.; Miyazaki, K.;
Kimura, N. Cancer Sci. 2004, 95, 377–384.
2. Gagnon, M.; Saragovi, H. U. Exp. Opin. Ther. Pat. 2002,
12, 1215–1223.
26. Palcic, M. M.; Heerze, L. D.; Pierce, M.; Hindsgaul, O.
Glycoconj. J. 1988, 5, 49–63.
3. Svennerholm, L. Adv. Gen. 2001, 44, 33–41.
4. Schnaar, R. L. Carbohydr. Chem. Biol. 2000, 4, 1013–
1027.
27. Yamashiro, S.; Haraguchi, M.; Furukawa, K.; Takamiya,
K.; Yamamoto, A.; Nagata, Y.; Lloyd, K. O.; Shiku, H. J.
Biol. Chem. 1995, 270, 6149–6155.
5. Ravindranath, M. H.; Gonzales, A. M.; Nishimoto, K.;
Tam, W.-Y.; Soh, D.; Morton, D. L. Ind. J. Exp. Biol.
2000, 38, 301–312.
28. Lee, Y. C.; Kojima, N.; Wada, E.; Kurosawa, N.;
Nakaoka, T.; Hamamoto, T.; Tsuji, S. J. Biol. Chem.
1994, 269, 10028–10033.
6. Hakomori, S.-I.; Zhang, Y. Chem. Biol. 1997, 4, 97–104.
7. Cheshev, P. E.; Khatuntseva, E. A.; Tsvetkov Iu, E.;
Shashkov, A. S.; NifantÕev, N. E. Bioorg. Khim. 2004, 30,
68–79.
8. Ishida, H.; Kiso, M. Trends Glycosci. Glyc. 2001, 13, 57–
64.
9. Castro-Palomino, J. C.; Simon, B.; Speer, O.; Leist, M.;
Schmidt, R. R. Eur. J. Chem. 2001, 7, 2178–2184.
10. Hasegawa, A.; Kiso, M. Carbohydrates 1992, 243–266.
11. Duclos, R. I. Carbohydr. Res. 2000, 328, 489–507.
12. Zehavi, U.; Tuchinsky, A. Glycoconj. J. 1998, 15, 657–
662.
13. Ito, Y.; Paulson, J. C. J. Am. Chem. Soc. 1993, 115, 1603–
1605.
14. Liu, K. K. C.; Danishefsky, S. J. J. Am. Chem. Soc. 1993,
115, 4933–4934.
15. Antoine, T.; Heyraud, A.; Bosso, C.; Samain, E. Angew.
Chem., Int. Ed. 2005, 44, 1350–1352.
29. Williams, M. A.; Kitagawa, H.; Datta, A. K.; Paulson, J.
C.; Jamieson, J. C. Glycoconj. J. 1995, 12, 755–761.
30. Blixt, O.; Head, S.; Mondala, T.; Scanlan, C.; Huflejt, M.
E.; Alvarez, R.; Bryan, M. C.; Fazio, F.; Calarese, D.;
Stevens, J.; Razi, N.; Stevens, D. J.; Skehel, J. J.; van Die,
I.; Burton, D. R.; Wilson, I. A.; Cummings, R.; Bovin, N.;
Wong, C. H.; Paulson, J. C. Proc. Natl. Acad. Sci. 2004,
101, 17033–17038.
31. Guo, Y.; Feinberg, H.; Conroy, E.; Mitchell, D. A.;
Alvarez, R.; Blixt, O.; Taylor, M. E.; Weis, W. I.;
Drickamer, K. Nat. Struct. Mol. Biol. 2004, 11, 591–598.
32. Fazio, F.; Bryan, M. C.; Blixt, O.; Paulson, J. C.; Wong,
C.-H. J. Am. Chem. Soc. 2002, 124, 14397–14402.
33. Blixt, O.; van Die, I.; Norberg, T.; van den Eijnden, D. H.
Glycobiology 1999, 9, 1061–1071.
34. Chernyak, A. Y.; Sharma, G. V. M.; Kononov, L. O.;
Krishna, P. R.; Levinsky, A. B.; Kochetkov, N. K.; Rao,
A. V. R. Carbohydr. Res. 1992, 223, 303–309.