Synthesis of a sialic acid containing complex-type n-glycan on a solid support

Article abstract of DOI:10.1002/asia.200800411

A new solid-phase synthesis of N-linked glycans featuring 1) highly ste-reoselective b-mannosylation and microfluidic a-sialylation and 2) efficient glycosylation of the N-phenyltrifluoroacetimidate units on JandaJel resin is reported. Reagent concentration effects by a fluorous solvent are effectively applied, and the use of these methods results in the first synthesis of a sialic acid containing complex-type N-glycan on a solid support.

Full text of DOI:10.1002/asia.200800411

FULL PAPERS
DOI: 10.1002/asia.200800411
Synthesis of a Sialic Acid Containing Complex-Type N-Glycan on a Solid
Support
Katsunori Tanaka, Yohei Fujii, Hiroomi Tokimoto, Yasutaka Mori, Shin-ichi Tanaka,
Guang-ming Bao, Eric R. O. Siwu, Aiko Nakayabu, and Koichi Fukase*[a]
Abstract: A new solid-phase synthesis of N-linked glycans featuring 1) highly ste-
reoselective b-mannosylation and microfluidic a-sialylation and 2) efficient glyco-
sylation of the N-phenyltrifluoroacetimidate units on JandaJel resin is reported.
Reagent concentration effects by a fluorous solvent are effectively applied, and
the use of these methods results in the first synthesis of a sialic acid containing
complex-type N-glycan on a solid support.
Keywords: carbohydrates
proteins · glycosylation · microflui-
dics · solid-phase synthesis
· glyco-
Introduction
supports.[4] Although the nonreducing end sialic acid residue
was not introduced to these oligosaccharides, a small library
of complex-type N-glycans, which consisted of 17 structures,
was elegantly constructed.[4d] Seeberger and co-workers have
also been investigating automated oligosaccharides synthesis
on solid supports;[5] a core pentasaccharide structure of com-
plex-type N-glycans was prepared.[5a]
We have been investigating a variety of solid-phase and
solid-supported methods,[6] such as chemical fishing, catch-
and-release,[6a–c] or separation based on affinity separation
(SAS)[6d–g] in combination with microfluidic glycosylation
technology,[6g,7] and these methods have been successfully
applied to the library synthesis of oligosaccharides. While
we continuously strive to develop efficient solid-phase meth-
ods, our interests in elucidating unknown biological func-
tions of mammalian N-glycans as well as sialic acid contain-
ing oligosaccharides[8] have motivated us to establish a prac-
tical solid-phase synthesis for complex-type N-glycans.
Herein, we report an efficient solid-phase protocol that can
readily deal with the structural diversity to the N-glycans.
The first chemical synthesis of the sialic acid containing
complex-type N-glycan on solid supports was achieved
based on the established method.
Among the various types of oligosaccharide structures, as-
paragine-linked oligosaccharides (N-glycans) are the most
prominent in terms of diversity and complexity.[1] It is be-
coming clear that they are involved in a variety of important
physiological events, such as cell–cell recognition, adhesion,
signal transduction, quality control, and circulatory resi-
dence of proteins.[1] However, isolating large quantities of
structurally pure N-glycans from natural sources is difficult.
Thus, chemical synthesis provides an attractive opportunity
to evaluate their biological functions. Although a number of
new chemical or combined methods employing biological
technology have been actively investigated,[2] an efficient ap-
proach to these complex oligosaccharides has yet to be es-
tablished in terms of 1) selectivity in the glycosyl bond for-
mations, namely, b-mannosylation and a-sialylation, and 2) a
nontedious purification process during each step of glycosy-
lation and deprotection. A solid-supported protocol is the
most attractive and practical approach to the library-direct-
ed synthesis of these diverse structures of natural and non-
natural N-glycans.[3] Schmidt and co-workers recently report-
ed the first library-directed synthesis of N-glycans on solid
[a] Dr. K. Tanaka, Y. Fujii, Dr. H. Tokimoto, Y. Mori, Dr. S.-i. Tanaka,
G.-m. Bao, Dr. E. R. O. Siwu, A. Nakayabu, Prof. Dr. K. Fukase
Department of Chemistry, Graduate School of Science
Osaka University
Results and Discussion
The initial target of our strategy was sialic acid-containing
N-glycan 1[1] (see structure in Scheme 4) with asymmetrical
branching chains, which is difficult to obtain from natural
sources. To prepare target N-glycan 1 as well as other di-
verse structures of this family efficiently, we designed frag-
1-1 Machikaneyama-cho, Toyonaka-shi, Osaka 560-0043 (Japan)
Fax : (+81)6-6850-5419
E-mail: koichi@chem.sci.osaka-u.ac.jp
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/asia.200800411.
574
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 574 – 580

ments a–d (Figure 1) by using ideas similar to Schmidtꢁs
precedent,[4] namely, two challenging glycosyl bond forma-
tions (i.e., b-mannosylation and a-sialylation) were con-
structed in advance in the solution-phase[9] or under micro-
for the sugar extension on the solid phase, while an azido-
chlorobenzyl (AzClBn) group[6a–c] was applied at the C-3 hy-
droxy group of the branching mannose. These temporary
protecting groups allowed N-glycans to be prepared with di-
verse structures at the C-3’’ and C-6’’ hydroxy groups. A tri-
chloroethoxycarbonyl (Troc) group was applied to protect
the nitrogen atom on glucosamine because this protecting
group not only ensured glycosylation selectivity by the
neighboring participation, but also conferred high reactivity
as the glycosyl acceptor.[11] After a sugar extension on solid
supports, alcoholysis of the p-xylyleneglycol ester linkage by
NaOBn provided N-glycans protected with hydrogenolysis-
sensitive groups (benzyl on hydroxy groups, benzyloxycar-
bonyl on the glycosaminyl amino group, and 4-hydroxyme-
thylbenzyl at the anomeric hydroxy group).
Scheme 1 shows the synthesis for fragments a–c. Glucosa-
minyl fragment a was prepared from known 2[9] by Fmoc
protection of the C-4 hydroxy group followed by conversion
of the 1-O-allyl group into an N-phenyltrifluoroacetimidate
in quantitative yield in three steps. Mannosyl fragment b
was synthesized from 4[9] by selective benzylation of the C-3
hydroxy group; 4 was treated with dibutyltin(IV) oxide in
toluene to provide the intermediary cyclic tin acetal, which
Figure 1. Synthetic strategy of N-glycan library on solid supports.
fluidic conditions.[7b] Then, suitably protected mono- and dis-
accharide donors a–d, activated as the corresponding glyco-
syl N-phenyltrifluoroacetimidates,[10] could stereoselectively
be glycosylated on solid supports with the aid of neighboring
group participation (Figure 1). N-Phenyltrifluoroacetimidate
was selected as the leaving group because of its higher sta-
bility than the corresponding trichloroacetimidate, which
makes it suited for the longer storage of fragments a–d. An
O-Fmoc group was used as a temporary protecting group
Abstract in Japanese:
Scheme 1. Synthesis of imidate fragments a–c. Reagents and conditions:
a) FmocCl, pyridine, CH2Cl2, RT, quant. for 3, 88% for 5a, 95% for 10;
b) “Ir” complex, H2, THF, RT, then I2, H2O, RT; c) N-phenyltrifluoroace-
timidoyl chloride, Na2CO3 or K2CO3 acetone, RT, quant. for a, 86% for
b, 81% for 7, 97% for c (2 steps); d) Bu2SnO, toluene, 1208C, 3 h, then,
CsF, TBAI, BnBr or AzClBnBr, RT, 73% for R1 =Bn, 87% for R1 =
AzClBn; e) BnBr, NaH, DMF, RT, 95%; f) Et3SiH, BF3·Et2O, CH2Cl2,
RT, 82% or microfluidic conditions;[12] g) Ac2O, pyridine, RT, 83% for 6,
quant. for 10; h) 1m HCl, MeOH, 608C, 92%.
Chem. Asian J. 2009, 4, 574 – 580
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
575

K. Fukase et al.
FULL PAPERS
was subsequently treated with
benzyl bromide in the pres-
ence of CsF and TBAI. After
the C-2 hydroxy group was
protected with Fmoc group, 5a
was obtained in 64% in two
steps. Reductive opening of
the 4,6-O-benzylidene acetal in
5a by treating with triethylsi-
lane (Et3SiH) in the presence
of BF3·OEt2 selectively gave
the C-6 benzyloxy function,
from which 6 was obtained by
acetylation (68% in two steps,
1 g scale). However, reductive
opening of the acetal on a
large scale produced a signifi-
cant amount of the diol by-
product through the concomi-
tant acid-mediated hydrolysis
of the acetal. These problems
were circumvented under mi-
crofluidic reaction conditions,
which our group previously de-
veloped;[12] 6 was continuously
and reproducibly obtained in
nearly quantitative yields even
on a 10 g scale. Finally, the de-
Scheme 2. Synthesis of imidate fragment d by microfluidic a-sialylation. Reagents and conditions: a) PPh3,
MeCN, RT, quant; b) Ac2O, pyridine, 708C, 87%; c) “Ir” complex, H2, THF, RT, then I2, H2O, RT; d) N-phe-
nyltrifluoroacetimidoyl chloride, K2CO3 acetone, RT, 82% for 2 steps.
protection of the anomeric allyloxy group followed by the
imidate formation provided fragment b in 86% in two steps.
The synthesis of fragment c was achieved according to
our highly b-selective mannosylation method,[9] which uti-
lized the trimethylsilyl tetrakis(pentafluorophenyl)borate,
neuraminic acids, namely, N-acetyl or N-glycolyl groups. As
anticipated, sialylation between 11b and galactosyl acceptor
12[7] in the presence of TMSOTf as an activator and 4 ꢂ
M.S. in propionitrile provided 13b in 90% yield with good
a-selectivity (a/b=9:1). Furthermore, applying the continu-
ous microfluidic sialylation, which was developed by our
group for 11a,[7b] improved both the yield and a-selectivity;
13b was obtained quantitatively with near total a-selectivity
(a/b=20:1). The a and b stereoisomers were easily separat-
ed by chromatography on silica-gel, and the pure a-isomer
was readily converted into the N-acetyl derivative 14 in
95% yield by treatment with PPh3 and subsequent N-acety-
lation. Finally, 14 was converted into imidate d in 82% yield
using the general procedure. Hence, we successfully pre-
pared fragments a–d on a 5–10 g scale.
Prior to the glycosylation trials on the solid supports using
imidates a–d towards the N-glycan synthesis, detailed glyco-
sylation conditions as well as the effects of the co-solvents
were optimized by using N-Troc glucosaminyl fragment a
(Table 1, entries 1–4). Initially, we examined polystyrene
supports (co-polymer of styrene with 1% of divinylben-
zene), which are currently commonplace in solid-phase syn-
thesis. The loading yields of the glucosaminyl moiety were
calculated after Et3N-induced deprotection of the Fmoc
group and subsequent UV/Vis analysis of the resulting di-
benzofulvene (absorbance intensity at 301 nm). When poly-
styrene loaded with a p-xylyleneglycol ester linker (0.5 me-
quivgÀ1)[4,6c] was treated with 2.5 equivalents of fragment a
in the presence of TMSOTf (0.5 equiv) in CH2Cl2 at room
TMSBACHTUNGTRENNUNG a Lewis acid/cation trap activator
(C6F5)4,[9b] as
(Scheme 1). Thus, mannosyl donor 7 with azidochlorobenzyl
group at the C-3 hydroxy group, which was prepared from
5b via common mannosyl fragment 4, was glycosylated with
N-Troc-glucosaminyl acceptor 2 in the presence of 20 mol%
of TMSBACHTUNGTRENNUNG(C6F5)4 8 and 4 ꢂ molecular sieves (4 ꢂ M.S.) in
CH2Cl2 at À788C to provide 9 in 79% with excellent b-se-
lectivity (b/a=97:3). Hydrolysis of the benzylidene acetal
(92%), sequential protection of C-6’ and C-4’ hydroxy
groups by Fmoc and acetyl groups (95% in two steps), and
conversion of 1-O-allyl into the imidate leaving group (97%
in two steps) provided fragment c.
The most critical endeavor was a-sialylation toward frag-
ment d (Scheme 2). We recently discovered the fixed-dipole
moment effects of the C-5 N-phthalimide group in sialyl
donor 11a,[7a] and realized the quantitative and completely
a-selective sialylation under microfluidic conditions.[7b] How-
ever, the selective deprotection of the N-phthalyl group in
the presence of the C-1 methoxycarbonyl in 13a was trou-
blesome on the 1–2 g reaction scale. As an alternative to the
N-phthalyl function, we employed the C-5 azide group in
sialyl donor 11b because this azide group should direct simi-
lar fixed-dipole moment effects,[13] but should be easier to
convert into naturally occurring nitrogen substituents of
576
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 574 – 580

Sialic Acid Containing Complex-Type N-Glycan
Table 1. Glycosylation trials on various solid supports.
Entry
Resin[a]
Solvents
Yield [%][b]
1
2
3
4
5
6
7
8
polystyrene
polystyrene
polystyrene
polystyrene
ArgoPore
ArgoPore
JandaJel
CH2Cl2
13
16
19
23
30
39
100
100
CH2Cl2/ionic liquid 16 (3:1)
CH2Cl2/C4F9OEt (3:1)
CH2Cl2/C4F9OEt (1:1)
CH2Cl2
CH2Cl2/C4F9OEt (1:1)
CH2Cl2
JandaJel
CH2Cl2/C4F9OEt (1:1)
[a] p-Xylyleneglycol linker was introduced to each amino group-loaded
resin in the presence of HOBt and DIC in DMF (loading amount:
0.5 mequivgÀ1). [b] Estimated from UV/Vis spectrum of dibenzofulvene
at 301 nm in CH2Cl2.
temperature for 1 h (Table 1, entry 1), 13% of the xylylene
linker was glycosylated. Interestingly, the addition of a 1/3
volume of ionic liquid 16 relative to CH2Cl2 slightly in-
creased the loading yield (16%, Table 1, entry 2). Further-
more, the addition of a fluorous solvent (C4F9OEt) showed
a similar effect (19%, Table 1, entry 3), and when a 1:1
mixed solvent of C4F9OEt and CH2Cl2 was applied, the effi-
ciency of glycosylation was enhanced twofold (23%) rela-
tive to the reaction performed in neat CH2Cl2 (Table 1,
entry 1 vs. 4). The profound effects of these co-solvents can
be rationalized by the reagent concentration effects on the
solid supports. That is, glycosyl donor a (or the correspond-
ing oxocarbenium ion equivalent) might efficiently be accu-
mulated onto the polystyrene supports because of favorable
hydrophobic interactions in the presence of an ionic liquid
or fluorous solvent additives such that the glycosylation re-
activity may be enhanced. These reagent concentration ef-
fects have been applied to the solid-phase synthesis of pep-
tides and other reactions,[14] but to the best of our knowl-
edge, this is the first observation for glycosylation.
After successfully activating the glycosyl acceptor on the
solid supports towards glycosylation, we investigated the
solid-phase synthesis of core pentasaccharide 17 of complex-
type N-glycans (Scheme 3). The success of each glycosyla-
tion step was evaluated by TLC, ESI-MS, and HPLC after
cleaving the oligosaccharides from the resin by treatment
with sodium methoxide (cleavage products obtained as N-
methoxycarbonyl derivatives). Thus, the C-4 hydroxy group
of the glucosaminyl fragment introduced on the supports
Scheme 3. Solid-phase synthesis of pentasaccharide 17 on polystyrene
supports. Reagents and conditions: a) 15% Et3N, CH2Cl2, RT, 4 h;
b) PBu3, CH2Cl2, 1 h, then DDQ/AcOH/H2O (1:1:1), THF, RT, 2 h;
c) (Bu3Sn)2, microwave (50 W, 20 psi, 2008C), DMF, 40 min, then Ac2O,
THF, 1 h; d) 1m NaOMe, MeOH/THF (1:1), RT, 1 h.
C4F9OEt=1:1, room temperature) to provide the corre-
sponding trisaccharide in 96% yield (Scheme 3). However,
Fmoc-deprotection and subsequent continuous glycosylation
with mannose fragment b under the same conditions result-
ed in only 10% of the desired tetrasaccharide, and most of
the starting trisaccharide was recovered. After optimizing
the conditions, this glycosylation could only be achieved in a
reasonable yield (79%) when using 10 equivalents of the
donor and excess fluorous solvent (twice the volume relative
to CH2Cl2), as well as repeating the procedure (double gly-
cosylation). Furthermore, subsequent glycosylation with
fragment b after deprotection of the AzClBn group on solid
supports (PBu3–H2O, then DDQ–AcOH),[6a–c] also necessi-
tated this rather inefficient glycosylation operation; penta-
saccharide 17 was obtained in 70% yield from the cleavage
of the AzClBn group by radical-mediated N-Troc deprotec-
tion under microwave irradiation on a solid support.[14e]
Disappointed by these glycosylation results on polystyr-
ene supports, we reinvestigated solid supports to achieve ef-
ficient glycosylation even during the late stage of the glyco-
sylation on the solid supports, namely, when the hydroxy
groups are glycosylated as the size of the oligosaccharide ac-
ceptors increases. It was found that the dioxybutane-linked
polystyrene, JandaJel,[15] meets our requirements (Table 1,
entries 5–8). Although the use of macroporous polystyrene,
ArgoPore,[6b,c,14a,b,16] achieved a threefold increase in the re-
(15) was further glycosylated with ManbACTHNUTRGNE(NUG 1-4)GlcNTroc imi-
date c under identical conditions established in Table 1
(donor; 2.5 equiv, TMSOTf; 0.5 equiv, 4 ꢂ M.S., CH2Cl2/
Chem. Asian J. 2009, 4, 574 – 580
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
577

K. Fukase et al.
FULL PAPERS
activity relative to that of the regular polystyrene-based
resin (30% and 39% without and with fluorous solvent ad-
ditive, Table 1, entries 5 and 6, respectively), JandaJel af-
forded quantitative glycosylation without the addition of a
fluorous co-solvent (Table 1, entries 7 and 8). The flexible
dioxybutane-linked polystyrene structure as well as the ac-
cumulation of Lewis acid by coordinating to the oxygen
inside this resin might explain the high glycosylation effi-
ciency, although the detailed reactivity on JandaJel supports
needs further to be evaluated.
Based on the discovery of JandaJel as an effective solid-
phase platform for glycosylation, the assembly of the N-
glycan structures was investigated (Scheme 4). By simply re-
peating the sequence of glycosylation (donor; 2.5 equiv,
TMSOTf; 0.5 equiv, and reaction performed only in CH2Cl2)
and deprotection (15% Et3N in CH2Cl2 for Fmoc, and
PBu3/H2O-DDQ/AcOH for AzClBn) in the glycosylation
order of a-c-b-b, pentasaccharide 18 was obtained in 60%
total yield (from the first glycosylation) after cleavage from
the resin by NaOMe treatment. TLC analysis of the inter-
mediary tri-, tetra-, and pentasaccharides revealed that each
Scheme 4. Solid-phase synthesis of N-glycans 18 and 1 on JandaJel supports. Reagents and conditions: a) donor (2.5 equiv), TMSOTf (0.5 equiv), 4 ꢂ
M.S., CH2Cl2, RT, 1 h; b) 15% Et3N, CH2Cl2, RT, 4 h; c) PBu3, CH2Cl2, 1 h, then DDQ/AcOH/H2O (1:1:1), THF, RT, 2 h; d) donor (2.5 equiv), TMSOTf
(0.5 equiv), 4 ꢂ M.S., CH2Cl2/C4F9OEt (1:1), RT, 1 h; e) 1m NaOMe, MeOH/THF (1:1), RT, 1 h; f) 1m NaOBn, BnOH/THF (1:2), RT, 3 h, then 3m
NaOH, RT, 12 h; g) 20% Pd(OH)2/C, H2, Ac2O/MeOH (1:1), RT, 9 h.
578
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 574 – 580

Sialic Acid Containing Complex-Type N-Glycan
(125 mmol, 2.0 mL) in CH2Cl2 (or with fluorous co-solvent) was added to
a resin (50 mmol of the acceptor hydroxy group on the resin) at room
temperature, and the resulting mixture was shaken at this temperature
for 10 min. TMSOTf (4.5 mL, 25 mmol) was added, and the resulting sus-
pension was shaken at this temperature for 1 h. The solution was re-
moved, and the resin was washed sequentially with CH2Cl2 and MeOH
(each 1.5 mL, 1.5 min, 5 sets), and dried in vacuo.
glycosylation step proceeded in nearly quantitative yield.
Neither a fluorous solvent nor a large excess of glycosyl
donors was required.
Encouraged by these results, we then examined the syn-
thesis of octasaccharide 1, which contains a sialic acid resi-
due, by sequentially glycosylating fragments a-c-b-a-d-b
(Scheme 4). Although glycosylation was quite successful up
to the fourth glycosylation with glycosaminyl imidate a, the
unreacted pentasaccharide acceptor was mainly recovered
Fmoc deprotection: The resin (50 mmol of the protected hydroxy group
on the resin) was suspended in CH2Cl2 (1.0 mL) and 15% Et3N solution
in CH2Cl2 (2.0 mL) was added. The resulting mixture was shaken at
room temperature for 4 h. The solution was removed the resin was
washed with DMF (1.5 mL, 1.5 min, 14 times) and then sequentially with
CH2Cl2 and Et2O (each 1.5 mL, 2.0 min, 10 sets), and dried in vacuo for
5 h to prepare for subsequent glycosylation.
during the fifth glycosylation with NeuaACTHNUTRGNE(NUG 2-6)Gal imidate d.
To circumvent the retarded reactivity of the acceptor hy-
droxy groups on the extended oligosaccharide structures on
the solid supports, the reagent concentration effect shown in
Table 1 was effectively applied. Thus, a mixed solvent of
CH2Cl2/C4F9OEt (1:1) was employed in the fifth and the
final glycosylation with fragments d and b; after treating the
resin with NaOBn, protected octasaccharide 19 was ob-
tained in 27% total yield (calculated from the first introduc-
tion of the fragment a to the resin), after gel filtration fol-
lowed by the removal of the strongly UV/Vis absorbing
polystyrene derivatives by HPLC, produced during the
cleavage from the JandaJel resin. Finally, 19 was treated
with acetic anhydride in the presence of 20% Pd(OH)2 on
carbon under a hydrogen atmosphere to give octasaccharide
Azidochlorobenzyl (AzClBn) deprotection; To the resin (50 mmol of the
protected hydroxy group on the resin) suspended in CH2Cl2 (2.0 mL) was
added tributylphosphine (43.1 mL, 173 mmol) at room temperature, and
the resulting mixture was shaken at this temperature for 1 h. The solution
was filtered off and the resin was washed with CH2Cl2 (each 1.5 mL,
2.0 min, 4 times). To the resin again suspended in THF (2.0 mL) were
added 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ; 39.2 mg,
173 mmol), AcOH (99.2 mL, 1.73 mmol), and H2O (31.2 mL, 1.73 mmol) at
room temperature and the resulting mixture was shaken at this tempera-
ture for 2 h. After the solution was filtered off, the resin was washed se-
quentially with DMF, MeOH, and CH2Cl2 (each 1.5 mL, 2.0 min, 5 sets),
and dried in vacuo for 5 h to prepare for subsequent glycosylation.
Reaction monitoring (cleavage from the resin by NaOMe treatments): A
28% solution of NaOMe in MeOH (500 mL) was added to a suspension
of the resin (50 mmol of the oligosaccharide on the resin) in THF
(1.0 mL) and MeOH (1.0 mL) at room temperature, and the resulting
mixture was shaken at this temperature for 1 h. The resin was washed
with THF (1.5 mL, 2.0 min, 5 times), and the resulting solution was neu-
tralized by ion-exchange resin, dowex H+ to provide the crude products.
The residue was purified by thin layer chromatography on silica gel (6%
MeOH in chloroform) to give the corresponding N-methoxycarbonyl de-
rivatives as the colorless solids. The efficiency of the reaction was evalu-
ated by TLC, HPLC (column; 5C18-AR300 (nacalai tesque) 4.6ꢃ250 mm,
eluent; MeCN in H2O), and ESI- or MALDI-TOF-MS; mono-GlcN m/z
1
1 in 74% yield. The H NMR spectrum showed good match
with an authentic N-glycan sample with asparagine residue,
and hence, the first solid-phase synthesis of the sialic acid-
containing N-glycan was achieved.[17]
Conclusions
calcd for C30H35NNaO8 [M+Na]+: 560.2, found: 560.2; Manb
(1–4)GlcN m/z calcd for C72H80ClN5NaO19 [M+Na]+: 1377.9, found:
1377.4; Mana(1–6)Manb(1–4)GlcNb(1–4)GlcN m/z calcd for
C92H102ClN5NaO24 [M+Na]+: 1720.3, found: 1720.6; Mana
(1–6)[Mana(1–
3)]Manb(1–4)GlcNb
(1–4)GlcN 18 m/z calcd for C105H121N2O29 [M+H]+:
1873.8, found: 1873.9; GlcNb(1–2)Mana(1–6)Manb(1–4)GlcNb(1–
4)GlcN m/z calcd for
114H127ClN6NaO30 [M+Na]+: 2117.8, found:
2117.6; methyl ester of Neua(2–6)Galb(1–4)GlcNb(1–2)Mana(1–
6)Manb(1–4)GlcNb
(1–4)GlcN m/z calcd for C132H157ClN7O43 [M+H]+:
ACHTUNGTRENUN(NG 1–4)GlcNb-
In conclusion, we developed an efficient solid-phase method
of N-glycan synthesis and successfully applied it to the first
chemical synthesis of a complex-type N-glycan with a nonre-
ducing end sialic acid. Our solid-phase protocol was success-
ful because of 1) the highly selective b-mannosylation and
a-sialylation either in the solution phase or under microflui-
dic conditions, which led to the large scale preparation of N-
phenyltrifluoroacetimidate fragments a–d, 2) the nearly
quantitative glycosylation on JandaJel supports, and 3) the
fluorous-solvent-assisted reagent concentration effects on
the solid-phase glycosylation. Because a variety of natural
and nonnatural N-glycans could be easily prepared by glyco-
sylating imidates a–d or their slightly structural variants
(such as fucosylated congener of the glucosamino fragment
a), the present protocol might be applicable to a general N-
glycan synthesis, even in an automated synthesis. A library
synthesis of the mammalian N-glycans and their biofunction-
al studies, including the PET biodistribution,[18] is currently
in progress in our laboratory.
AHCTUNGTRENNUNG
G
R
ACHTUNGTRENNUNG
G
A
ACHTUNGTRENNUNG
N
ACHTUNGTRENNUNG
G
E
U
ACHTUNGTRENNUNG
C
G
G
N
ACHTUNGTRENNUNG
G
ACHTUNGTRENNUNG
2563.0, found: 2562.8.
Protected octasaccharide (19): To a suspension of octasaccharide-loaded
JandaJel resin (theoretical amount of oligosaccharide loaded on the
resin; 30.2 mmol) in dry THF (1.0 mL) was added an 1m NaOBn solution
in BnOH (500 mL, 500 mmol) at room temperature, and the resulting mix-
ture was shaken at this temperature for 3 h. Subsequently, aqueous 3m
NaOH (500 mL, 1.5 mmol) was added and the resulting suspension was
shaken overnight at room temperature. The resin was washed with THF
(1.5 mL, 2.0 min, 5 times), and the resulting solution was neutralized by
ion-exchange resin, dowex H+ (500Wꢃ8). After the excess benzyl alco-
hol was removed by size-partitioning gel filtration through a column
filled with sephadex LH-20 (eluent: MeOH), the protected octasacchar-
ide 19 (a white solid, 38.4 mg) was further separated from the strongly
UV/Vis absorbing polystyrene derivatives by HPLC, produced during the
cleavage from JandaJel resin (23.9 mg, 27% overall yield; column: naca-
lai tesque 5C18-AR300, 4.6ꢃ250 mm; MeCN in H2O (50–100% gradient
over 60 min); retention time of 19: 14.4 min): MALDI-TOF-MS m/z
calcd for C162H185N4O48 [M+H]+: 2955.2, found: 2955.4. Owing to the
very low solubility of 19 in a variety of the NMR solvents, only broad,
but characteristic proton signals of the oligosaccharide with benzyl type-
protection groups could be observed in CD3CN; 1HNMR (600 MHz;
Experimental Section
General procedure for glycosylation, deprotection of Fmoc and azido-
chlorobenzyl (AzClBn) groups on JandaJel, and reaction monitoring
(cleavage from the resin): Glycosylation: A solution of fragments a–d
Chem. Asian J. 2009, 4, 574 – 580
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemasianj.org
579

K. Fukase et al.
FULL PAPERS
CD3CN): d=1.82 (s, 3H), 1.88 (dd, 1H, J=12.0 Hz, 3-HaxNeu), 1.99 (s,
3Hꢃ3), 2.69–2.87 (brdm, including 3-HeqNeu), 3.19–4.11 (m, 22H), 4.16–
5.42 (m, 32H), 5.59–5.97 (m, 6H), 7.07–7.62 (m, 67H), 7.85–8.03 ppm (m,
7H).
[3] a) “Combinatorial and Solid-Phase Methods in Oligosaccharide Syn-
thesis in Glycoscience” in:K. Fukase, Chemistry and Chemical Biol-
ogy, Vol. I–III (Eds.: B. O. Fraser-Reid, K. Tatsuta, J. Thiem),
Springer, New York, 2001; b) “Polymer-Supported and Tag-Assisted
Methods in Oligosaccharide Synthesis in Glycoscience” in:K.
Tanaka, K. Fukase, Chemistry and Chemical Biology, Vol. I–III, 2nd
ed. (Eds.: B. O. Fraser-Reid, K. Tatsuta, J. Thiem), Springer, New
York, 2008.
[4] a) X. Wu, M. Grathwohl, R. R. Schmidt, Angew. Chem. 2002, 114,
4664–4668; Angew. Chem. Int. Ed. 2002, 41, 4489–4493; b) X. Wu,
R. R. Schmidt, J. Org. Chem. 2004, 69, 1853–1857; c) X. Wu, R. R.
Schmidt, Eur. J. Org. Chem. 2004, 2826–2832; d) S. Jonke, K.-G.
Liu, R. R. Schmidt, Chem. Eur. J. 2006, 12, 1274–1290.
[5] a) D. M. Ratner, E. R. Swanson, P. H. Seeberger, Org. Lett. 2003, 5,
4717–4720; b) K. R. Love, P. H. Seeberger, Angew. Chem. 2004,
116, 612–615; Angew. Chem. Int. Ed. 2004, 43, 602–605; c) P. H.
Seeberger, D. B. Werz, Nature 2005, 4, 751–763, and references
therein.
[6] a) K. Egusa, K. Fukase, S. Kusumoto, Synlett 1997, 675–676; b) K.
Egusa, K. Fukase, S. Kusumoto, Synlett 2000, 27–32; c) K. Egusa, S.
Kusumoto, K. Fukase, Synlett 2001, 777–780; d) S.-Q. Zhang, K.
Fukase, S. Kusumoto, Tetrahedron Lett. 1999, 40, 7479–7483; e) S.-
Q. Zhang, K. Fukase, M. Izumi, Y. Fukase, S. Kusumoto, Synlett
2001, 590–596; f) Y. Fukase, S.-Q. Zhang, K. Iseki, M. Oikawa, K.
Fukase, S. Kusumoto, Synlett 2001, 1693–1698; g) K. Fukase, M. Ta-
kashina, Y. Hori, D. Tanaka, K. Tanaka, S. Kusumoto, Synlett 2005,
2342–2346.
[7] a) K. Tanaka, T. Goi, K. Fukase, Synlett 2005, 2958–2962; b) S.
Tanaka, T. Goi, K. Tanaka, K. Fukase, J. Carbohydr. Chem. 2007,
26, 369–394.
[8] a) K. Tanaka, T. Masuyama, K. Hasegawa, T. Tahara, H. Mizuma, Y.
Wada, Y. Watanabe, K. Fukase, Angew. Chem. Int. Ed. 2008, 47,
102–105; b) K. Tanaka, T. Masuyama, K. Minami, Y. Fujii, K. Hase-
gawa, T. Tahara, H. Mizuma, Y. Wada, Y. Watanabe, K. Fukase,
Peptide Science 2007, 91–94.
[9] a) S. Tanaka, M. Takashina, H. Tokimoto, Y. Fujimoto, K. Tanaka,
K. Fukase, Synlett 2005, 2325–2328; b) K. Tanaka, K. Fukase, Elec-
tronic Encyclopedia of Reagents for Organic Synthesis, 2007.
[10] a) B. Yu, H. Tao, Tetrahedron Lett. 2001, 42, 2405–2407; b) M. Adi-
nolfi, G. Barone, A. Iadonisi, M. Schiattarella, Synlett 2002, 269–
270; c) S. Cai, B. Yu, Org. Lett. 2003, 5, 3827–3830.
Octasaccharide (1): To a solution of the protected octasaccharide 19
(2.0 mg, 680 nmol) in MeOH (150 mL) and Ac2O (150 mL) was added
20% Pd(OH)2 on carbon (1.7 mg), and the resulting mixture was stirred
under a hydrogen atmosphere for 9 h at room temperature. After the cat-
alyst was filtered off and washed with 20% MeCN in H2O, the solution
was lypophilized to give 1 (1.1 mg, 74%): MALDI-TOF-MS m/z calcd
for C59H98N4NaO44 [M+Na]+: 1589.5, found: 1590.5, m/z calcd for
C59H96N4NaO43 [MÀH2O+Na]+: 1571.5, found: 1571.9, m/z calcd for
C59H97N4O43 (MÀH2O+H]+: 1549.6, found: 1548.8, m/z calcd for
C59H95N4O42 (MÀ2H2O+H]+: 1531.5, found: 1531.7, and m/z calcd for
C48H79N3NaO35 (M-sialic acid-H2O+Na)+: 1279.4, found: 1279.0;
1H NMR (500 MHz; D2O): d=1.64 (dd, 1H, J=12.0, 12.0 Hz), 1.95 (s,
6H), 1.99 (s, 6H), 2.60 (dd, 1H, J=11.9, 4.6 Hz), 3.40–3.93 (m, 46H),
3.99 (brs, 1H), 4.04 (brs, 1H), 4.18 (brs, 1H), 4.37 (d, 1H, J=7.9 Hz),
4.52 (brs, 1H), 4.70 (brs, 0.5H), 4.86 (s, 1H), 5.03 ppm (s, 1H).
Acknowledgements
We thank Hiroaki Asai, Dr. Michio Sasaoka, Kazuhiro Fukae, and Azusa
Hashimoto (Otsuka Chemical Co., Ltd.) for supplying the authentic
sample of the octasaccharide 1 with asparagine residue. Parts of this
work were financially supported by Grants-in-Aid for Scientific Research
from the Japan Society for the Promotion of Science (Nos. 19681024 and
19651095), Collaborative Development of Innovative Seeds from the
Japan Science and Technology Agency (JST), New Energy and Industrial
Technology Development Organization (NEDO, project ID:
07A01014a), Research Grants from Yamada Science Foundation, as well
as Molecular Imaging Research Program, Grants-in-Aid for Scientific
Research from the Ministry of Education, Culture, Sports, Science and
Technology (MEXT) of Japan.
[1] “Analysis of Glycans, Polysaccharide Functional Properties & Bio-
chemistry of Glycoconjugate Glycans, Carbohydrate-mediated Inter-
actions” in:Comprehensive Glycoscience, From Chemistry to Systems
Biology, Vol. II & III (Eds.: J. P. Kamerling, G.-J. Boons, Y. C. Lee,
A. Suzuki, N. Taniguchi, A. G. J. Voragen), Elsevier, UK, 2007.
[2] a) “Introduction to Glycoscience Synthesis of Carbohydrates” in:-
Comprehensive Glycoscience, From Chemistry to Systems Biology,
Vol. I (Eds.: J. P. Kamerling, G.-J. Boons, Y. C. Lee, A. Suzuki, N.
Taniguchi, A. G. J. Voragen), Elsevier, UK, 2007; also see recent
synthesis of N-glycan core pentasaccharides: b) A. Dan, Y. Ito, T.
Ogawa, Tetrahedron Lett. 1995, 36, 7487; c) P. H. Seeberger, P. F. Ci-
rillo, S. H. Hu, X. Beebe, M. T. Bilodeau, S. J. Danishefsky, Enantio-
mer 1996, 1, 311; d) Z. W. Guo, Y. Nakahara, Y. Nakahara, T.
Ogawa, Bioorg. Med. Chem. 1997, 5, 1917; e) S. J. Danishefsky, S.
Hu, P. F. Cirillo, M. Eckhardt, P. H. Seeberger, Chem. Eur. J. 1997,
3, 1617; f) I. Matsuo, M. Isomura, K. Ajisaka, J. Carbohydr. Chem.
1999, 18, 841; g) M. Takatani, T. Nakama, K. Kubo, S. Manabe, Y.
Nakahara, Y. Ito, Y. Nakahara, Glycoconjugate J. 2000, 17, 361;
h) Z. G. Wang, X. F. Zhang, D. Live, S. J. Danishefsky, Angew.
Chem. 2000, 112, 3798; Angew. Chem. Int. Ed. 2000, 39, 3652;
i) V. Y. Dudkin, J. S. Miller, S. J. Danishefsky, Tetrahedron Lett. 2003,
44, 1791; j) J. S. Miller, V. Y. Dudkin, G. J. Lyon, T. W. Muir, S. J.
Danishefsky, Angew. Chem. 2003, 115, 447; Angew. Chem. Int. Ed.
2003, 42, 431; k) M. R. Pratt, C. R. Bertozzi, J. Am. Chem. Soc.
2003, 125, 6149.
[11] S. Inamura, K. Fukase, S. Kusumoto, Tetrahedron Lett. 2001, 42,
7613–7616.
[12] K. Tanaka, K. Fukase, Synlett 2007, 164–166.
[13] Similar effects of C-5 azide on the a-selective sialylation were re-
ported: C.-S. Yu, K. Niikura, C.-C. Lin, C.-H. Wong, Angew. Chem.
2001, 113, 2984–2987; Angew. Chem. Int. Ed. 2001, 40, 2900–2903.
[14] a) M. Bodanzsky, A. Bodanzsky, Practice of Peptide Synthesis 2nd
ed., Springer, New York, 1994; b) S. A. Kates, F. Albericio, Solid-
Phase Synthesis A Practical Guide, Marcel Dekker, New York, 2000;
c) F. Z. Dorwald, Organic Synthesis on Solid Phase: Supports, Link-
ers, Reactions, 2nd ed., Wiley-VCH, Weinheim, 2002; d) H. Akamat-
su, K. Fukase, S. Kusumoto, Synlett 2004, 1049–1053; e) H. Tokimo-
to, K. Fukase, Tetrahedron Lett. 2005, 46, 6831–6832.
[15] Available from Sigma–Aldrich.
[16] K. Fukase, Y. Nakai, K. Egusa, J. A. Porco, Jr., S. Kusumoto, Synlett
1999, 1074.
[17] Y. Kajihara, Y. Suzuki, N. Yamamoto, K. Sasaki, T. Sakakibara,
L. R. Juneja, Chem. Eur. J. 2004, 10, 971–985.
[18] a) K. Tanaka, K. Fukase, Org. Biomol. Chem. 2008, 6, 815–828;
b) K. Tanaka, K. Fukase, Mini-Rev. Org. Chem. 2008, 5, 153–162.
Received: October 28, 2008
Published online: January 20, 2009
580
www.chemasianj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Asian J. 2009, 4, 574 – 580