Azido glycoside primer: A versatile building block for the biocombinatorial synthesis of glycosphingolipid analogues

Article abstract of DOI:10.1016/S0008-6215(00)00238-X

A lactoside primer, 12-azidododecyl β-lactoside, was synthesized via the Koenigs-Knorr method by glycosylation of 1,12-dodecyldiol with perbenzoylated lactosyl bromide. The presence of the 2-O-acyl substituent in the donor gave the β-lactoside, and an excess of acceptor ensured monoglycosylation of the diol. Mesylation of the ω-hydroxyl group in the aglycon, followed by displacement of the mesylate with azide and subsequent O-debenzoylation gave the desired ω-azidododecyl β-lactoside. The azido glycoside primer was examined in mouse B16 melanoma cells for its feasibility as a building block for oligosaccharide biosynthesis. Uptake of the azido glycoside primer by B16 cells resulted in the sialylation of the galactose residue of the primer to give a glycosylated product having the same glycan as in ganglioside GM3. After 24 h incubation of B16 cells with the primers, the amount of sialylated ω-azidododecyl β-lactoside primer was 75% of the amount of sialylated n-dodecyl β-lactoside. However, after 48 h incubation, both primers gave equal amounts of the sialylated products. Interestingly, the remaining azido glycoside primer after 48 h incubation was 5.6-fold greater than that of the alkyl primer, indicating degradation of the alkyl primer to a larger extent than the ω-azido glycoside primer. The facile chemical synthesis and the efficient uptake in cells make the azido glycoside primer a versatile building block for the biocombinatorial synthesis of glycolipid oligosaccharides.

Full text of DOI:10.1016/S0008-6215(00)00238-X

www.elsevier.nl/locate/carres
Carbohydrate Research 329 (2000) 755–763
Azido glycoside primer: a versatile building block for
the biocombinatorial synthesis of glycosphingolipid
analogues
Maria Carmelita Z. Kasuya a,b, Lai X. Wang c, Yuan C. Lee c, Motoaki Mitsuki a,
Hideki Nakajima a, Yoshiaki Miura a, Toshinori Sato a, Kenichi Hatanaka a,
Sadako Yamagata a,b, Tatsuya Yamagata a,b,
*
aDepartment of Biomolecular Engineering, Tokyo Institute of Technology, 4259 Nagatsuta-cho, Midori-ku,
Yokohama 226-8501, Japan
bDi6ision of Glycobiology and Glycotechnology, Japan Institute of Leather Research, 1-1 Senju, Midori-cho,
Adachi-ku, Tokyo 120-8601, Japan
cDepartment of Biology, The Johns Hopkins Uni6ersity, Baltimore, MA 21218, USA
Received 28 March 2000; accepted 8 August 2000
Abstract
A lactoside primer, 12-azidododecyl b-lactoside, was synthesized via the Koenigs–Knorr method by glycosylation
of 1,12-dodecyldiol with perbenzoylated lactosyl bromide. The presence of the 2-O-acyl substituent in the donor gave
the b-lactoside, and an excess of acceptor ensured monoglycosylation of the diol. Mesylation of the v-hydroxyl group
in the aglycon, followed by displacement of the mesylate with azide and subsequent O-debenzoylation gave the
desired v-azidododecyl b-lactoside. The azido glycoside primer was examined in mouse B16 melanoma cells for its
feasibility as a building block for oligosaccharide biosynthesis. Uptake of the azido glycoside primer by B16 cells
resulted in the sialylation of the galactose residue of the primer to give a glycosylated product having the same glycan
as in ganglioside GM3. After 24 h incubation of B16 cells with the primers, the amount of sialylated v-azidododecyl
b-lactoside primer was 75% of the amount of sialylated n-dodecyl b-lactoside. However, after 48 h incubation, both
primers gave equal amounts of the sialylated products. Interestingly, the remaining azido glycoside primer after 48
h incubation was 5.6-fold greater than that of the alkyl primer, indicating degradation of the alkyl primer to a larger
extent than the v-azido glycoside primer. The facile chemical synthesis and the efficient uptake in cells make the
azido glycoside primer a versatile building block for the biocombinatorial synthesis of glycolipid oligosaccharides.
© 2000 Elsevier Science Ltd. All rights reserved.
Keywords: Lactosyl ceramide analogue; Oligosaccharide library; Oligosaccharides, biocombinatorial synthesis; Azido glycoside
primer; Saccharide primer, glycosylation
1. Introduction
Glycopolymers are an important class of
bioactive macromolecules that have found a
* Corresponding author. Tel.: +81-3-38885111; fax: +81-
3-38709631.
E-mail address: tcyamagata@pop16.odn.ne.jp (T. Yama-
gata).
variety of biomedical applications in tumor
diagnosis, human vaccines, detection and
trapping of viruses and toxins, and targeted
0008-6215/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(00)00238-X

756
M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
shown to occur at the terminal residue of the
lactoside, proceeded with the involvement of
innate glycosyltransferases in the cell.
drug delivery systems [1]. Increasing attention
has been paid to synthetic polymers, carrying
various pendant mono- and oligosaccharides
as biomaterials useful in applications utilizing
an oligosaccharide-recognizing functional
group. For example, various types of poly-
mers that incorporate certain sialic acid
residues have been synthesized in an attempt
to inhibit the influenza virus [2]. Moreover,
lactose-containing polystyrene and polypep-
tides are useful for culture substrates and as
artificial antigens [3].
The usual synthetic strategy for the con-
struction of glycopolymers involves polymer-
ization or co-polymerization of a biologically
important oligosaccharide unit [4]. However,
chemical synthesis of oligosaccharides remains
a laborious task requiring intricate methods of
protection, glycosylation, and deprotection
steps. We embarked on a biocombinatorial
synthesis to prepare glycopolymers without
these shortcomings.
Biocombinatorial synthesis has the potential
to become an important tool for generating
libraries of oligosaccharides and glycopoly-
mers with desirable biological activities. The
strategy involves the preparation of oligosac-
charides or their glycosides by combining
chemical synthetic methods with cellular
biosynthetic processes. From a single building
block, a large number of oligosaccharides can
be constructed by using a variety of cells [5].
The ideal building blocks are synthetically ac-
cessible glycoside-based primers that act as
substrates for additional glycosylation after
incorporation into the cells. Moreover, such
glycosides must possess functional groups
amenable to further manipulation to enable
preparation of glycopolymers.
Whereas incorporation of the lactoside
primers into B16 cells afforded an a-(2“3)
sialyl lactoside derivative with the same
oligosaccharide structure as that of GM3, the
main ganglioside synthesized by B16 cells [5],
incubation of the same lactoside primers with
PC12 cells gave a-Gal-(1“4)- and a-Gal-(1“
3)-a-Gal-(1“4) lactosides. N-Acylaminoethyl
lactosides were also taken up and glycosylated
by B16 cells to give the GM3 analogue like-
wise [6].
As part of our work on the biocombinato-
rial synthesis for the construction of an
oligosaccharide library, we report in this pa-
per the chemical synthesis of another building
block, 12-azidododecyl b-lactoside primer (2),
and its uptake by B16 cells for the synthesis of
glycosphingolipid analogues. The efficiency of
12-azidododecyl b-lactoside (2) to prime
oligosaccharide synthesis was also compared
with that of the n-dodecyl b-lactoside primer
(1). The introduction of an azido functional
group in the aglycon allows its facile conver-
sion into an amino group for specific incorpo-
ration into desired glycopolymers for future
work.
In our work in biocombinatorial synthesis
to construct an oligosaccharide library, the
building blocks used were amphipathic gly-
cosides, a hydrophilic lactosyl moiety with
hydrophobic aglycons of various alkyl chain
lengths (e.g., a Cn alkyl lactoside and a Cn
amide lactoside, where n=C8, C12, C14, C16).
Incubation of amphipathic lactosides, appro-
priately referred to as primers, with different
cell types primed the synthesis of oligosacchar-
ides of different structures depending on the
cells used. Direct glycosylation, which was
Employing biocombinatorial synthesis of
oligosaccharides, it is possible to construct a
glycolipid-specific saccharide library using dif-
ferent glycoside primers in combination with a
wide variety of cells. Using primers of diverse
structures, it may be further possible to obtain
a variety of N-linked glycans, O-glycans, and
glycosaminoglycan–saccharides, leading to
construction of oligosaccharide libraries that
will pave the way for better understanding of

M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
757
Scheme 1. Reagents and conditions: (a) HO(CH2)12OH, AgOTf, THF, rt, 4 h; (b) MsCl, pyridine, rt, 4 h; (c) NaN3, DMF, 60 °C,
4 h then 12 h at rt; (d) MeOH–NaOMe, rt, overnight.
the biological function of glycoconjugates in
biological systems.
azide in DMF) afforded compound 8 in 64%
yield in two steps. Usual O-deacylation with
NaOMe–MeOH and purification by silica gel
chromatography afforded the 12-azidododecyl
b-lactoside (2).
Effect of amphipathic lactoside primers on
cell growth.—Mouse melanoma cells (B16)
were cultured in the absence or in the presence
of 50 mM azido or alkyl lactoside primer. The
growth of cells cultured for 2 days in the
presence of the primers was similar to that of
the control (Fig. 1). The primers did not ex-
press toxic activity toward the cells at this
concentration.
2. Results
Chemical synthesis of the azido lactoside
primer.—The chemical synthesis of 12-azi-
dododecyl b-lactoside (2) is shown in Scheme
1. The glycosyl donor, perbenzoyllactosyl bro-
mide (5), was readily prepared in two steps
from lactose following a known procedure [7].
Glycosylation of 1,12-dodecyldiol using silver
triflate as catalyst in dichloromethane (under
Koenigs–Knorr conditions) yielded a partial
analog of ceramide glycoside. Although there
are two possible sites for glycosylation in do-
decyldiol, the major product was 6. The excess
glycosyl acceptor (3.4 equivalents) resulted in
mostly the monoglycosylation product, and
the presence of the 2-O-benzoyl group led to
the desired b-lactoside via neighboring-group
participation. Purification by silica gel chro-
matography afforded the v-hydroxyldodecyl
b-benzoyl lactoside (6). The b-anomeric
configuration of the glycosylation product was
confirmed by its NMR spectrum. The rela-
tively large coupling constant for J1,2 (7.9 Hz)
indicates a b configuration of newly formed
glycosidic linkage.
Fig. 1. Effect of amphipathic lactoside primers on cell growth.
B16 cells (4.0×106, 100 mm dish) were incubated for 48 h in
serum-free 1:1 DMEM–F12 supplemented with transferrin
and insulin (TI–DF) in the absence (control) or presence of
50 mM lactoside primer.
Activation of the hydroxyl group of com-
pound 6 by reaction with methanesulfonyl
chloride (MsCl), followed by displacement of
the resulting mesylate with azide (with sodium

758
M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
Fig. 2. Glycosylated lactosides produced by adding amphipathic lactoside primers to cells. B16 cells (3.0×106, 100 mm dish) were
incubated for 24 h in serum-free 1:1 DMEM–F12 supplemented with transferrin and insulin (TI–DF) in the presence of 50 mM
alkyl 1 or azido glycoside 2 primer. The cells were washed with PBS (−) and harvested. Lipids were extracted, purified, separated
by HPTLC, and analyzed as described in the Experimental Section. The left and right panels show the glycosylated products
accumulated in the cell fraction and culture medium fraction, respectively; Lane 1, control (TI–DF only); Lane 2, alkyl primer
1; Lane 3, azido primer 2.
tic GM3 confirmed that the glycosylated
product is an a-(2“3)-sialylated lactoside —
an analogue of GM3, which is the glycosphin-
golipid predominantly expressed on the cell
surface of B16 melanoma cells (Scheme 2)
[5,8]. To confirm whether reduction of the
azide to an amine proceeded during incuba-
tion with cells, the HPTLC plate was sprayed
with ninhydrin reagent. The absence of the
characteristic stain confirmed that the azido
function remained intact even after glycosyla-
tion. Thus, the azido glycoside primer entered
the glycosphingolipid (GSL) biosynthetic
pathway and functioned as an acceptor for the
Glycosylation of the azido lactoside
primer.—To examine the versatility of the
12-azidododecyl b-lactoside (2) as a building
block for the synthesis of glycosphingolipid-
type oligosaccharide, B16 cells expressing high
levels of GM3 were fed with 50 mM of the
primer. After incubation for 24 h, culture
media were collected and the lipids were
purified using SepPak C18 column. Lipids
from the cell homogenates were extracted with
chloroform–methanol and then with chloro-
form–2-propanol–water. Analysis by HPTLC
of the lipid extracts from the cell homogenates
and the culture media indicated a band corre-
sponding to a putative glycosylated lactoside
primer (Fig. 2, Lane 3).
Identification of the glycosylated lactoside.—
To elucidate the structure of the glycan unit
modified in cells, the silica gel zone corre-
sponding to the putative glycosylated lactoside
was scraped from the HPTLC plate and ex-
tracted. Analysis of the MALDI–TOF mass
spectrum (Fig. 3) revealed a peak at m/z 842.9
corresponding to a monosialylated lactoside
primer. To establish whether the sialylated
lactoside is a-(2“3)- or a-(2“6)-linked, the
purified product was hydrolyzed with endogly-
cosylceramidase [6]. HPTLC analysis of the
released product alongside those from sialyl
n-dodecyl lactoside standards [5] and authen-
Fig. 3. MALDI-TOF MS of the glycosylated product ob-
tained from the incubation of cells with 12-azidododecyl
b-lactoside (2) for 24 h.

M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
759
Scheme 2. Biocombinatorial synthesis of GM3 analogue after uptake of 12-azidododecyl b-lactoside primer (2) into B16 cells.
for both azido and alkyl glycoside primers
were about the same (9.0 and 8.4%, respec-
tively) (Fig. 5).
GM3 synthase, a-(2“3)-sialyl transferase, re-
sulting in direct glycosylation occurring at the
terminal galactosyl residue of the lactose moi-
ety subsequent to uptake by cells.
Comparison of priming ability between the
alkyl glycoside primer and azido glycoside
primer.—In our previous report, the ability of
different amphipathic lactosides with various
alkyl aglycons [5] to prime the synthesis of
glycolipid-type oligosaccharides was exam-
ined. Among the lactosides tested, n-dodecyl
lactoside (the alkyl glycoside primer) was
found to be the best acceptor for sialylation in
B16 melanoma cells.
To determine whether the priming ability of
the azido glycoside primer behaves similarly
as the dodecyl glycoside primer, the primers
were incubated for 24 and 48 h with B16
melanoma cells. The lipid fractions were ex-
tracted from the respective cell and medium
fractions, and analyzed by HPTLC. As shown
in Fig. 2, the glycosylated primers were mostly
secreted into the culture medium.
Fig. 4. Sialylated lactosides from the cell and medium frac-
tions. B16 cells (3.0×106, 100 mm dish) were incubated for
24 h in serum-free 1:1 DMEM–F12 supplemented with trans-
ferrin and insulin (TI–DF) in the presence of 50 mM alkyl 1
or azido glycoside primer 2 for 24 h. The amounts of sialy-
lated primer (nmol) from the cell and medium fractions were
calculated. Each value represents the average of three deter-
minations and the experiment was repeated five times.
After incubation of the lactoside primers for
24 h, HPTLC of the culture medium fraction
showed clearly the preference for the uptake
of the alkyl glycoside primer over the azido
glycoside primer as judged from the band
corresponding to the respective glycosylated
product (Fig. 2). Quantification of the sialy-
lated primers showed that the amount of sialy-
lated azido glycoside primer was 75% of the
amount of sialylated alkyl glycoside primer.
As shown in Fig. 4, the amounts of glycosy-
lated product using the azido glycoside primer
2 and the alkyl glycoside primer 1 were 10.3
nmol (4.1%) and 14.0 nmol (5.6%) of the
administered primers, respectively. However,
after 48 h incubation, the amounts of glycosy-
lation products accumulated in the cultures
Fig. 5. Sialylated lactoside production with time. B16 cells
(3.0×106, 100 mm dish) were incubated for 24 h in serum-
free 1:1 DMEM–F12 supplemented with transferrin and in-
sulin (TI–DF) in the presence of 50 mM alkyl 1 or azido
glycoside primer 2. The amounts of sialylated products were
estimated at 24 and 48 h.

760
M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
tosides used have hydrophobic aglycons that
mimic the ceramide group of LacCer and help
the lactoside reach the Golgi apparatus. It is
evident that the lactosides are taken up by the
cells, conveyed to the Golgi apparatus, and
taken through the GSL biosynthetic pathway
of glycosylation. The sialylated lactosides were
mostly secreted into the culture medium.
Glycosylation was cellular GSL-specific
[10], and the efficiency of glycosylation is de-
pendent on the aglycon structure. In B16 cells,
amphipathic lactosides synthesized GM3
oligosaccharide, whereas in PC12 cells, lac-
tosides produced Gb3 and a-Gal-(1“3)-Gb3
[5,11]. Thus, the carbohydrate structure of the
glycosylation product depends on the glyco-
lipid the cells synthesize. Primers with longer
hydrophobic chains hampered the release of
glycosylated products from cells. Lactosides
with a dodecyl aglycon were glycosylated 2–3-
fold greater than hexadecyl lactosides, and 1.4
times more than the aminoethyl lactoside with
the dodecanoyl chain. The presence of an
N-acyl chain suppressed the secretion of gly-
cosylated products. Octyl lactosides did not
function as glycosylation acceptors under the
conditions used in the study.
Fig. 6. Amount of remaining lactoside primer after incubation
in B16 cells for 48 h. B16 cells (3.0×106, 100 mm dish) were
incubated in serum-free 1:1 DMEM–F12 supplemented with
transferrin and insulin (TI–DF) in the presence of 50 mM
alkyl 1 or azido glycoside primer 2. The amounts of remain-
ing primers 1 or 2 were estimated after 48 h.
Interestingly, although the amounts of sialy-
lated primers were almost the same for both
primers after 48 h incubation, the remaining
azido glycoside primer recovered from the cul-
ture medium was 78% of the input while that
of the alkyl primer was only 14%, as shown in
Fig. 6.
3. Discussion
Among many lactosides tested, the dodecyl
b-lactoside (1) appears to be the best acceptor
for glycosylation in vivo [5,6]. However, for
the preparation of glycopolymers, the alkyl
lactoside has limited utility because the alkyl
aglycon is not readily modified further. Thus,
for the preparation of biocombinatorial glyco-
polymers, a different lactoside primer with a
readily modifiable functional aglycon that is
amenable to cellular uptake, glycosylation,
and secretion was deemed necessary.
In this report, we synthesized 12-azidodode-
cyl b-lactoside (2), and examined its feasibility
for the biocombinatorial synthesis of gly-
cosphingolipid analogues. After internaliza-
tion and glycosylation by the cells, the azido
function in the aglycon can be selectively and
easily reduced into an amino group, enabling
it for further manipulation such as incorpora-
tion into glycopolymers.
In the study of oligosaccharide structure
and cellular functions, one of the critical steps
is the facile synthesis of the desired oligosac-
charide analogues. We have described in this
paper the synthesis of GM3 oligosaccharide
by combining chemical synthesis and cellular
biosynthesis. This is one of the segments in
our ongoing work of building an oligosacchar-
ide library by biocombinatorial synthesis, a
novel technique using different glycoside
primers and subsequent incorporation in a
variety of cells. The glycosylated products
thus obtained can be used for glycopolymer or
other modulations by treatment with endogly-
cosylceramidase [9] or by further modification
of the aglycon unit.
Lactosylceramide (LacCer) is a common in-
termediate in glycosphingolipid synthesis, and
the lactose structure in glycoconjugates is
unique to glycosphingolipids [10]. Recently,
we reported a series of artificial glycolipids
consisting of a lactosyl moiety that functions
effectively as a GSL primer subsequent to
their incorporation into cells [5,6]. The lac-
The synthesis of the 12-azidododecyl gly-
coside primer (2) was achieved under the
Koenigs–Knorr conditions using 1,12-dode-
cyldiol and per-O-benzoylated lactosyl bro-

M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
761
scope of this study, our data suggest that
phenol red should be used with caution as a
pH indicator in cell cultures, especially when
glycosylation is under study.
mide. Through mesylation, displacement of
the mesylate with azide, and subsequent O-
debenzoylation, the 12-azidododecyl b-lac-
toside (2) was obtained in a moderate yield.
This synthetic scheme should be applicable to
the preparation of other azido alkyl glycosides
using diols with various chain lengths.
Although a more detailed pathway of the
primers after uptake by cells needs to be es-
tablished, the azido glycoside primer 2 is more
favorable as an acceptor in oligosaccharide
synthesis over the alkyl glycoside primer 1
because of the greater extent of glycosylation
and the higher efficiency in the uptake of
primer. The azido group of the glycosylated
products can be reduced and conjugated for
the synthesis of various glycoconjugates, with
which the saccharide functions can be ana-
lyzed widely. Development of methods for the
large-scale preparation of GM3 oligosacchar-
ide from the azido primer 2 using cells are
underway in order to produce chemical quan-
tities with homogeneity required for applica-
tion to polymer or antigen synthesis. The
simple chemical synthetic route and efficient
uptake by cells, coupled with the potential for
the facile preparation of glycopolymers, make
the azido glycoside primer a versatile building
block for the biocombinatorial synthesis of
glycosphingolipid analogues. Since no inher-
ent damage to B16 melanoma cells has been
brought about by the administration of the
azide primer, these results open new avenues
leading to further use of the azido group to
different primers for the synthesis of other
types of glycans.
Both the alkyl (1) and azido glycoside (2)
primers gave the GM3 oligosaccharide moiety
subsequent to the uptake by B16 cells. Incuba-
tion of the lactoside primers with the cells for
24 h showed that the amount of sialylated
product from the azido glycoside primer was
75% of that from the alkyl glycoside primer.
However, during the 48 h incubation period,
the amount of glycosylated product in both
primers was almost the same. The amount of
recovered azido glycoside primer after 48 h
incubation was 5.6-fold greater than the re-
covered alkyl glycoside primer. This indicates
that after incorporation into cells, the alkyl
glycoside primer and/or its sialylated form
possibly underwent greater degradation than
the azido glycoside primer. Thus, the azido
glycoside primer functions more efficiently as
an acceptor in oligosaccharide synthesis in the
B16 cells than the alkyl glycoside primer over-
all. These results provide fundamental infor-
mation on the effects of aglycon function-
alities and the factors important for the effi-
cient biocombinatorial synthesis of oligosac-
charides using lactoside primers.
Phenol red is a commonly used indicator to
monitor pH changes during cell culture. To
determine if phenol red in the culture medium
interferes in the glycosylation of lactoside
primers, the cells were cultured in DMEM,
with or without phenol red, prior to the 24-h
incubation with the primers. Treatment with
the alkyl glycoside primer gave almost the
same amount of sialylation product with or
without phenol red in the culture medium.
However, the amount of glycosylated prod-
ucts, 5.0 and 8.8 nmol, from the azido gly-
coside primer with or without phenol red,
respectively, may represent a significant differ-
ence. These results revealed that phenol red, a
commonly used indicator for pH changes in
the culture medium, interferes with intracellu-
lar glycosylations. Although the intrinsic ef-
fect of phenol red in cells goes beyond the
4. Experimental
Synthesis of 12-azidododecyl i-lactoside
primer (2)
General methods.—1H NMR spectra were
recorded at 200 MHz with a JEOL spectrome-
ter in CDCl3 using Me4Si as the internal refer-
ence. All reactions were monitored by thin
layer chromatography (TLC) on Silica Gel 60
F-254 (E. Merck), with detection by UV light
or by visualizing with spraying with anisalde-
hyde–H2SO4 and heating. Column chro-
matography was performed on Silica Gel 60
(70–230 mesh, E. Merck, Darmstadt). The
MALDI mass spectrum was recorded on a
KOMPACT MALDI 1 mass spectrometer

762
M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
mixture was diluted with CHCl3, and washed
with aq NaHCO3 and water. The organic
layer was dried with anhyd Na2SO4 and
filtered, and the filtrate was concentrated in
vacuo. Purification by column chromatogra-
phy with 1:1 toluene–EtOAc as eluent af-
(Kratos, UK) with a 2,5-dihydroxybenzoic
acid (DHB) matrix. The n-dodecyl lactoside
primer was synthesized from lactose as de-
scribed [5]. The per-O-benzoyl lactosyl bro-
mide (5) was prepared from lactose through
per-O-benzoylation and subsequent treatment
with HBr–HOAc [9].
1
forded compound 8 in 64% yield. H NMR
(CDCl3): 7.124–8.036 (m, 35 H, 7 Bz), 5.807
(t, 1 H, J=9.5 Hz, H-3), 5.756–5.695 (m, 2
H, H-2%,4%), 5.462 (dd, 1 H, J=7.9, 9.3 Hz,
H-2), 5.367 (dd, 1 H, J=3.4, 10.3 Hz, H-3%),
4.870 (d, 1 H, J=7.9 Hz, H-1%), 4.683 (d, 1 H,
J=7.9 Hz, H-1), 4.619–4.480 (m, 2 H, H-
6a,b), 4.259 (t, H, J=9.5 Hz, H-4), 3.925–
3.710 (m, 4 H, H-5,5%,6%a,b), 3.700–3.400 (m, 2
H, OCH2ꢀ), 3.260 (t, 2 H, J=6.9 Hz,
CH2N3), 1.621–1.092 [m, 20 H, ꢀ(CH2)10ꢀ].
Anal. Calcd for C73H73N3O18 (1280.37): C,
68.48; H, 5.75; N, 3.28. Found: C, 68.69; H,
5.90; N, 3.05.
12-Hydroxydodecyl 2,3,4,6-tetra-O-benzoyl-
i -
D
- galactopyranosyl - (1“4) - 2,3,6 - tri - O-
-glucopyranoside (6).—To 30 g
benzoyl-i-
D
,
(26.5 mmol) BzLacBr (5) and 4 A molecular
sieves in THF (500 mL) were added 8.0 g of
AgOTf (31.1 mmol, 1.2 equiv) and 18.0 g
(88.9 mmol, 3.4 equiv) of 1,12-dodecanediol,
and the resulting mixture was stirred at rt. The
reaction was monitored by TLC. After the
reaction was over (4 h), CHCl3 was added,
and the mixture was filtered. The filtrate was
neutralized with aq NaHCO3, washed with
water, dried with anhyd Na2SO4 and concen-
trated in vacuo. Silica gel column chromatog-
raphy of the residue with 1:1 toluene–EtOAc
as eluent, afforded compound 6 in 72% yield.
1H NMR (CDCl3): 8.037–7.125 (m, 35 H, 7
Bz), 5.816 (t, 1 H, J=9.5 Hz, H-3), 5.757–
5.696 (m, 2 H, H-2%,4%), 5.463 (dd, 1 H, J=
7.9, 9.8 Hz, H-2), 5.371 (dd, 1 H, J=3.4, 10.3
Hz, H-3%), 4.870 (d, 1 H, J=7.9 Hz, H-1%),
4.687 (d, 1 H, J=7.9 Hz, H-1), 4.618–4.473
(m, 2 H, H-6a,b), 4.256 (t, 1 H, J=9.5 Hz,
H-4), 3.915–3.400 (m, 8 H, H-5,5%,6%a,b,
ꢀCH2OH, and OCH2ꢀ), 1.621–1.090 [m, 20 H,
ꢀ(CH2)10ꢀ]. Anal. Calcd for C73H74O19
(1255.36): C, 69.84; H, 5.94. Found: C, 70.13;
H, 6.10.
12-Azidododecyl i-lactoside (2).—Com-
pound 8 (5 g, 3.48 mmol) was dissolved in dry
MeOH (50 mL), and to the solution was
added 5% NaOMe in MeOH. The mixture
was stirred overnight at rt and then neutral-
ized with cation-exchange resin (Dowex 50W-
X2, H+ form). After filtration, the filtrate was
evaporated, and the residue was treated with
toluene and evaporated repeatedly to give a
white solid. Recrystallization with EtOH af-
forded the 12-azidododecyl b-lactoside primer
(2) in 80% yield. 1H NMR (Me2SO-d6 contain-
ing 5% D2O): 4.202 (d, 1 H, J=7.9 Hz, H-1%),
4.141 (d, 1 H, J=7.9 Hz, H-1), 3.183 (t, 2 H,
J=6.8 Hz, CH2N3), 1.470–1.130 [m, 20 H,
ꢀ(CH2)10ꢀ]. Anal. Calcd for C24H45N3O11·0.5
H2O (560.63): C, 51.41; H, 8.11; N, 7.49.
Found: C, 51.70; H, 8.40; N, 7.50.
12-Azidododecyl 2,3,4,6-tetra-O-benzoyl-i-
-galactopyranosyl-(1“4)-2,3,6-tri-O-ben-
D
zoyl-i- -glucopyranoside (8).—To a solution
D
of 6 (10 g, 7.97 mmol, in pyridine (250 mL)
was added MsCl (1.82 g, 15.9 mmol, 2 equiv),
and the solution was stirred for 4 h at rt. The
mixture was filtered and evaporated to give a
syrup, which was diluted with CHCl3 and
neutralized with aq NaHCO3. The organic
layer was dried with anhyd Na2SO4 and con-
centrated in vacuo to afford the mesylated
compound 7, which was used immediately for
the preparation of 8. To a solution of 7 in
DMF (500 mL) was added NaN3 (1.36 g, 20.9
mmol), and the solution was stirred at 60 °C
for 4 h and then at rt for 12 h. The reaction
Incubation of B16 melanoma cells with lac-
toside primer
General methods.—B16 cells were obtained
from Riken Cell Bank (Tsukuba, Japan). Dul-
becco Modified Eagles’s Medium (DMEM),
DMEM and Ham F12 (1:1), and fetal bovine
serum (FBS) were from GIBCO. Insulin and
transferrin were from Wako Pure Chemical,
Tokyo. SepPak C18 was from Waters. HPTLC
(Silica Gel 60) and preparative TLC plates
were from E. Merck, Darmstadt, Germany.
The lactoside primers were dissolved in
Me2SO to an initial concentration of 50 mM.

M.C.Z. Kasuya et al. / Carbohydrate Research 329 (2000) 755–763
763
Cell culture. Mouse B16 melanoma cells
were cultured in 1:1 DMEM–F12 supple-
mented with 10% fetal bovine serum (FBS).
Cells were detached through application of
0.25% trypsin–EDTA, passaged every 3 days
and maintained in humidified atmosphere of
5% CO2 air at 37 °C.
Research Activities for Innovative Biosciences
(Pro BRAIN) (T.Y.), a Grant-in-Aid for Sci-
entific Research on Priority Areas (no.
5274106) to T.Y. from the Ministry of Educa-
tion, Science and Culture of Japan, and by a
fellowship from Pro BRAIN to M.C.K.
Incubation of cells with lactoside primer. In-
ocula of 3.0×106 cells were seeded into 100-
mm culture dishes containing 7 mL of
medium and incubated for 48 h. This was
followed by washing with TI–DF without
phenol red (1:1 DMEM–Hams F12 contain-
ing 30 nM SeO2, 5 mg/mL transferrin, and 50
mg/mL insulin) to remove the serum, and cells
were incubated with either 50 mM of the alkyl
primer 1 or 50 mM of the azido primer 2 for
24 and 48 h at 37 °C. After incubation, culture
media were collected and cells were washed
with PBS (−), harvested with 0.25% EDTA
in PBS (−), and centrifuged at 1000 rpm for
10 min.
The lipids were extracted from the cell pellet
with 2:1 CHCl3–MeOH, then with 7:11:2
CHCl3–2-propanol–water, in a sonicated
bath. Lipids from the culture media were
purified using a SepPak C18 column. Lipids
from the cell homogenate and culture medium
fractions were analyzed by HPTLC with 5:4:1
CHCl3–MeOH–0.2% aq KCl as developing
solvent. HPTLC plates were sprayed with re-
sorcinol, and with orcinol–H2SO4 reagent,
and heated to detect the separated glycolipids
which were later quantified using BIO-PRO-
FIL BIO Image Analyzer.
References
[1] (a) S.J. Danishefsky, M.T. Bilodeau, Angew. Chem., Int.
Ed. Engl., 35 (1996) 1380–1419. (b) A. Tsuchida, S.
Akimoto, T. Usui, K. Kobayashii, Biochemistry, 123
(1998) 715–721.
[2] (a) H. Kamitakahara, T. Suzuki, N. Nishigori, Y. Suzuki,
O. Kanie, C.H. Wong, Angew. Chem., Int. Ed. Engl, 37
(1998) 1524–1528. (b) S. Hakomori, Ann. Re6. Biochem.,
50 (1981) 733–764. (c) G. Sigal, M. Mammen, G. Dah-
mann, G. Whitesides, J. Am. Chem. Soc., 118 (1996)
3789–3800. (d) S.-K. Choi, M. Mammen, G. Whitesides,
J. Am. Chem. Soc., 119 (1997) 4103–4111.
[3] (a) K. Hatanaka, Y. Ito, A. Maruyama, Y. Watanabe, T.
Akaike, Macromolecules, 26 (1993) 1483–1486. (b) R.
Roy, F.D. Tropper, A. Romanowska, J. Chem. Soc.,
Chem. Commun., (1992) 1611–1613.
[4] K. Kobayashi, A. Tsuchida, Macromolecules, 30 (1997)
2016–2020.
[5] (a) Y. Miura, T. Arai, T. Yamagata, Carbohydr. Res.,
289 (1996) 193–199. (b) H. Nakajima, Y. Miura, T.
Yamagata, J. Biochem., 124 (1998) 148–156.
[6] Y. Miura, T. Yamagata, Biochem. Biophys. Res. Com-
mun., 241 (1997) 698–703.
[7] H. Paulsen, Angew. Chem., Int. Ed. Engl., 21 (1982)
155–173.
[8] (a) M. Ito, H. Komori, J. Biol. Chem., 271 (1996) 12655–
12660. (b) H. Komori, S. Ichikawa, Y. Hirabayashii, M.
Ito, J. Biol. Chem., 274 (1999) 8981–8987. (c) S.
Ichikawa, N. Nakajo, H. Sakiyama, Y. Hirabayashii,
Proc. Natl. Acad. Sci. USA, 91 (1994) 2710–2717.
[9] Y. Miura, T. Arai, A. Ohtake, M. Ito, K. Yamamoto, T.
Yamagata, Glycobiology, (1999) 957–960.
[10] (a) S. Hakomori, Sci. Am., 254 (1986) 44–53. (b) K.
Nicolau, T. Caulfield, H. Kataoka, Carbohydr. Res., 202
(1990) 177–191. (c) H.H. Freeze, D. Sampath, A. Varki,
J. Biol. Chem., 268 (1993) 1618–1627. (d) J.M. Las-
saletta, K. Carlson, P.J. Garegg, R.R. Schmidt, J. Org.
Chem., 61 (1996) 6873–6880.
Acknowledgements
This work was supported in part by funds
from the Program for Promotion of Basic
[11] M. Shimamura, T. Hayase, M. Ito, M.L. Rasilo, T.
Yamagata, J. Biol. Chem., 263 (1988) 12124–12128.
.