Display of azido glycoside on a sensor chip

Article abstract of DOI:10.1246/cl.2004.580

Using 12-azidododecyl β-mannoside, we successfully immobilized azido glycoside onto a sensor chip by either the Staudinger reaction or reduction of azido group followed by condensation reaction. Specific binding of Concanavalin A to the sensor chip proved

Full text of DOI:10.1246/cl.2004.580

5
80
Chemistry Letters Vol.33, No.5 (2004)
Display of Azido Glycoside on a Sensor Chip
ꢀ
y
y
yy
Toshinori Sato, Sinya Fujita, Maria Carmelita Z. Kasuya, Kenichi Hatanaka, and Tatsuya Yamagata
Department of Biosciences and Informatics, Keio University, 3-14-1 Hiyoshi, Yokohama 223-8522
y
Institute of Industrial Science, University of Tokyo, 4-6-1 Komaba, Meguro-ku 153-8505
Shenyang Pharmaceutical University, P. O. Box 29, 103 WenHua Road, Shenyang 110016, P. R. China
yy
(Received February 16, 2004; CL-040172)
8
Using 12-azidododecyl ꢀ-mannoside, we successfully im-
reprecipitation (chloroform/n-hexane). Deacetylation was car-
mobilized azido glycoside onto a sensor chip by either the Stau-
dinger reaction or reduction of azido group followed by conden-
sation reaction. Specific binding of Concanavalin A to the sensor
chip proved immobilization of the glycoside by either method.
ried out in dry methanol in the presence of NaOMe. The total
yield from mannose was 29%. The molecular mass determined
by MALDI–TOF–MS (Autoflex, Bulker) was 413.06 (exact
þ
mass as [M + Na ] = 412.47).
Immobilization of the Man-C12-N3 was carried out on the
sensor chip of a surface plasmon resonance-based IAsys9
(Thermo Labsystems) as shown in Figure 1.
It is known that oligosaccharides are important molecules
that are correlated with various cell functions and diseases. To
accelerate the research on the function of carbohydrate, the de-
velopment of saccharide microarray to comprehensively detect
carbohydrate recognition has long been sought for. For the de-
velopment of a saccharide microarray, a saccharide library con-
sisting of a variety of oligosaccharides is required to immobilize
individual oligosaccharides on the microplate. For the construc-
tion of a saccharide library, we have developed biocombinatorial
1
synthesis of oligosaccharides by animal cells in culture. From a
single building block chemically synthesized, a large number of
oligosaccharides can be constructed by various kinds of cells.
This is possible since animal cells have a potential to produce
particular glycoconjugates depending on their origin. Dodecyl
ꢀ
-lactoside is one of the building blocks which can be glycosy-
1
lated after incorporation into cells, and such a building block is
also called as saccharide primer. In order to utilize the oligosac-
charide on a sensor chip, oligosaccharides in saccharide library
must possess a functional group amenable to covalent reaction.
It was recently found that 12-azidododecyl ꢀ-lactoside, a sac-
charide primer having azido group, was as well glycosylated
2
by cells as dodecyl ꢀ-lactoside.
Several approaches on saccharide arrays have been reported.
Fukui et al. presented the neoglycolipids on nitrocellulose mem-
3
brane. Houseman and Mrksich prepared carbohydrate array by
the Diels–Alder mediated immobilization of carbohydrate–cy-
Figure 1. Immobilization of Man-C12-N3 (method 1) and Man-C12-
NH2 (Method 2) on sensor chip.
4
clopentadiene conjugates. Park and Shin reported the attach-
ment of maleimide-linked carbohydrates to a thiol group-coated
Method 1: 2-Diphenylphosphonylterephthalic acid 1-methyl
10
5
glass slide. Furthermore, Faizio et al. described the attachment
of oligosaccharides by 1,3-dipolar cycloadditions between
ester was employed as linker to immobilize azido glycoside. 1-
Methyl hydrogen 2-iodoterephthalate was synthesized from 1-
methyl hydrogen 2-aminoterephthalate through the treatment
6
azides and alkynes to microtiter plate.
1
1
We synthesized 12-azidododecyl ꢀ-mannoside (Man-C12-
N3), and immobilized it to sensor chip by two methods, the Stau-
dinger reaction (Method 1) and conversion of azides to amines
followed by condensation reaction (Method 2).
with NaNO2 followed by KI. The linker was synthesized by
coupling 1-methyl hydrogen 2-iodoterephthalate and diphenyl-
phosphine in the presence of palladium acetate and triethylamine
ꢁ
12
in acetonitrile under a nitrogen atmosphere for 5 h at 85 C.
Man-C12-N3 was synthesized as follows. 1-Bromo-1-de-
oxy-2,3,4,6-tetra-O-acetyl-ꢁ-D-mannopyranose was prepared
from mannose through acetylation and subsequent treatment
The linker was first conjugated with amino silane cuvette (Si–
C6H12–NH2, sensor chip) in 200 mM acetate buffer (pH 4) con-
taining 400 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodii-
mide (EDC) and 100 mM N-hydroxysuccinimide (NHS). There-
after, the linker was coupled with 20 mM Man-C12-N3 dissolved
in methanol. The amount of immobilized linker was estimated to
with HBr–HOAc. 12-Azidododecanol was synthesized from
7
1,12-dodecandiol. The O-acetyl mannosyl bromide was reacted
with 12-azidododecanol in CH2Cl2 in the presence of Ag2CO3
and molecular sieves 4A for 21 h. The product was purified by
silica gel chromatography (ethyl acetate:n-hexane = 1:1) and
2
be 0.1 ng/mm from response due to mass increase on the sur-
face of sensor chip, while the amount of immobilized Man-
Copyright ꢀ 2004 The Chemical Society of Japan

Chemistry Letters Vol.33, No.5 (2004)
581
2
Table 1. Kinetic parameters (k1, kꢂ1, Kd, and response) for immobi-
lized mannose
C12-N3 was below the detection limit (0.14 pg/mm ).
Method 2: Azido group of Man-C12-N3 was reduced to ami-
no group in the presence of triphenylphosphine in DMF/H20 for
k1
k
ꢂ1
ꢂ1
Kd
/mM
Responsea
/arc sec
ꢁ
ꢂ1 ꢂ1
6
h at 50 C. The production of 12-aminododecyl ꢁ-D-manno-
pyranoside (Man-C12-NH2) was confirmed by molecular mass
[M + H ] = 364.20) and ninhydrin test. Man-C12-NH2 was
immobilized on carboxylate cuvette in 5 mM maleic acid buffer
pH 6) containing 400-mM EDC and 100-mM NHS. Unreacted
carboxyl groups were blocked with ethanolamine. When the
concentrations of Man-C12-NH2 in the reaction solution were
5, 7.5, and 10 mM, the immobilized amounts were 0.14, 1.09,
and 2.35 pmol/mm , respectively.
/M
s
/s
Method 1
þ
(
ConA
BSA
2516
459
0.021 8.3
0.037 80.6
28
4
(
Method 2
Man = 0.14b
ConA
BSA
ConA
BSA
885
514
2825
531
3564
1567
0.043 48.6
0.020 38.9
38
19
17
1
10
2
b
Man = 1.09
0.025
0.034 64.0
0.023 6.5
0.024 15.3
8.8
2
Man = 2.35b ConA
BSA
a
Response after washing with buffer. [ConA] = 10 mM.
The density of the immobilized mannose (pmol/mm ).
b
2
sponds to 13% displacement of carboxyl group on the sensor
chip with mannose.
In conclusion, we developed two technologies to immobi-
lize azido glycoside on sensor chip. These technologies will lead
to the preparation of saccharide microarray using a strategy that
involves immobilization of the saccharide library obtained by
the saccharide primer method. Method 1 was very favorable
for us to immobilize azide-bearing saccharide library. We are
preparing mictotiter plates displaying azide-bearing oligosac-
charide produced by animal cells.
Figure 2. Binding of ConA (M1) and BSA (M2) to mannose immo-
bilized by method 1. L1 represents the binding of ConA to linker with-
out mannose. [ConA] = [BSA] = 10 mM.
This work was partly supported by Special Coordination of
Funds for Promoting Science and Technology from the Ministry
of Education, Culture, Sports, Science and Technology, the
Japanese Government (T. S.)
Whether or not mannose was successfully fixed and dis-
played on the sensor chip by Methods 1 and 2 was confirmed
by analysis with an IAsys. Figure 2 showed the typical sensor-
grams for the binding of mannose–specific lectin, Concanavalin
A (ConA), to mannose displayed by Method 1. Though the
amount of mannose immobilized on a sensor chip was below de-
tection limit of instrument, a specific binding of ConA (curve
M1 in Figure 2) was observed. The binding of ConA was inhib-
ited by the addition of 100 mM mannose, and no significant
binding of ConA to the linker surface without mannose was ob-
served. The bindings of ConA were measured at the ConA con-
centrations of 1, 3, 5, and 10 mM, and the kinetic parameters
References and Notes
1
a) Y. Miura, T. Arai, and T. Yamagata, Carbohydr. Res., 289, 193 (1996).
b) H. Nakajima, Y. Miura, and T. Yamagata, J. Biochem., 124, 148 (1998).
M. C. Z. Kasuya, L. X. Wang, Y. C. Lee, M. Mitsuki, H. Nakajima, Y.
Miura, T. Sato, K. Hatanaka, S. Yamagata, and T. Yamagata, Carbohydr.
Res., 329, 55 (2000).
2
3
S. Fukui, T. Feizi, C. Galustian, A. M. Lawson, and W. Chai, Nat. Biotech-
nol., 20, 1011 (2002).
4
5
6
B. T. Houseman and M. Mrksich, Chem. Biol., 9, 443 (2002).
S. Park and I. Shin, Angew. Chem., Int. Ed., 41, 3180 (2002).
F. Faizo, M. C. Bryan, O. Blixt, J. C. Paulson, and C.-H. Wong, J. Am.
Chem. Soc., 124, 14397 (2002).
(
binding rate k1, dissociation rate kꢂ1, dissociation constant
1
7
H NMR (500 MHz, CDCl3); ꢂ 3.75 (2H, t, HOCH2–), 3.23 (2H, m,
Kd, and response) were determined by FASTfit attached to IA-
sys. Those results were summarized in Table 1. The Kd value
for ConA-binding was 8.3 mM, and 10 times lower than that
for BSA-binding. The mannose-immobilized sensor chip pre-
pared by Method 2 also showed specific binding of ConA. The
specificity of ConA-binding depends on the density of immobi-
N3CH2–), 1.58 (5H, m, HOCH2CH2–, N3CH2CH2–, –OH), 1.38–1.28
(16H, m, –CH2–). C NMR (125 MHz, CDCl3) ꢂ 62.73 (s, 1C, –CH2OH),
13
5
1.31 (s, 1C, –CH2N3).
þ
8
MS (MALDI-TOF): m=z ¼ 581:30 [M + Na ]. 1H NMR (300 MHz,
CDCl3) ꢂ 1.27 (16H, s, –CH2–), 1.57 (4H, m, –CH2–), 1.75 (3H, s,
–CH ), 2.05 (3H, s, –CH ), 2.08 (3H, s, –CH ), 2.12 (3H, s, –CH ),
3
3
3
3
3.26 (2H, t, –CH2N3), 3.47 (2H, m, O–CH2–), 3.68 (1H, m, H5), 4.14
0
2
(1H, dd, H6), 4.24 (1H, dd, H6 ), 4.59 (1H, dd, H2), 5.15 (1H, dd, H3),
lized mannose. For the mannose density of 0.14 pmol/mm , no
5
.30 (1H, t, H4), 5.47 (1H, d, J ¼ 2:44 Hz, H1).
significant differences in the Kd values between ConA and
BSA were observed, though response due to the binding amount
of ConA was higher than that of BSA. For the mannose density
9
N. Athanassopoulou, R. J. Davies, P. R. Edwards, D. Yeung, and C. H.
Maule, Biochem. Soc. Trans., 27, 340 (1998).
10 a) E. Saxon, J. I. Armstrong, and C. R. Bertozzi, Org. Lett., 2, 2141 (2000).
b) E. Saxon, S. J. Luchansky, H. C. Hang, C. Yu, S. C. Lee, and C. R.
Bertozzi, J. Am. Chem. Soc., 124, 14893 (2002).
2
of 1.09 pmol/mm , significant differences in Kd values between
ConA and BSA were observed. These values well agreed with
þ
1
11
MS(FAB): m=z ¼ 306 [M + H ]. H NMR (300 MHz, CDCl3); ꢂ 3.67
those obtained by Method 1. For the mannose density of
2
(1H, s, –CH ), 7.83 (1H, d, J ¼ 8:05 Hz, Ar–H), 8.11 (1H, d,
3
J ¼ 8:63 Hz, Ar–H), 8.68 (1H, s, Ar–H).
2.35 pmol/mm , the specific binding of ConA was observed,
though the differences in Kd value between ConA and BSA
þ
1
1
2
MS (MALDI-TOF): m=z ¼ 365:31 [M + H ]. H NMR (300 MHz,
DMSO-d6) ꢂ 3.41 (3H, s, –CH3), 7.20 (4H, m, Ar–H), 7.40 (6H, m, Ar–
was not so large. From these results, it is considered that the op-
2
31
H), 7.52 (1H, d, J ¼ 3:78 Hz, Ar–H), 8.02 (2H, m, Ar–H). P NMR
timum density of mannose is about 1 pmol/mm , which corre-
(300 MHz, DMSO-d6) ꢂ ꢂ5:72 (P:), 28.13 (P=O).
Published on the web (Advance View) April 19, 2004; DOI 10.1246/cl.2004.580