Trityl-derivatized carbohydrates immobilized on a polystyrene microplate

Article abstract of DOI:10.1016/j.carres.2008.08.021

Carbohydrate biosensors, including carbohydrate arrays, are attracting increased attention for the comprehensive and high-throughput investigation of protein-carbohydrate interactions. Here, we describe an effective approach to fabricating a robust microp

Full text of DOI:10.1016/j.carres.2008.08.021

Carbohydrate Research 343 (2008) 2932–2938
Contents lists available at ScienceDirect
Carbohydrate Research
journal homepage: www.elsevier.com/locate/carres
Trityl-derivatized carbohydrates immobilized on a polystyrene microplate
Lan Zou a,b, Hei-Leung Pang a, Pak-Ho Chan a, Zhi-Shu Huang b, Lian-Quan Gu b, Kwok-Yin Wong a,
*
a Department of Applied Biology and Chemical Technology, Central Laboratory of the Institute of Molecular Technology for Drug Discovery and Synthesis,
The Hong Kong Polytechnic University, Hung Hom, Kowloon, Hong Kong, PR China
b School of Pharmaceutical Science, Sun Yat-Sen University, Guangzhou 510275, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 14 May 2008
Received in revised form 14 August 2008
Accepted 19 August 2008
Available online 28 August 2008
Carbohydrate biosensors, including carbohydrate arrays, are attracting increased attention for the com-
prehensive and high-throughput investigation of protein–carbohydrate interactions. Here, we describe
an effective approach to fabricating a robust microplate-based carbohydrate array capable of probing
protein binding and screening for inhibitors in a high-throughout manner. This approach involves the
derivatization of carbohydrates with a trityl group through an alkyl linker and the immobilization of
the trityl-derivatized carbohydrates (mannose and maltose) onto microplates noncovalently to construct
carbohydrate arrays. The trityl carbohydrate derivative has very good immobilization efficiency for poly-
styrene microplates and strong resistance to aqueous washing. The carbohydrate arrays can probe the
interactions with the lectin Concanavalin A and screen this protein for the well-known inhibitors methyl
Keywords:
Carbohydrate array
Microplate
Noncovalent
Lectin
a-D-mannopyranoside and methyl a-D-glucopyranoside in a high-throughput manner. The method
described in this paper represents a convenient way of fabricating robust noncovalent carbohydrate
arrays on microplates and offers a convenient platform for high-throughput drug screening.
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
well as microplate-based arrays.18–24 Among these methods, the
noncovalent type is attractive in the sense that this approach
Protein–carbohydrate interactions are biologically important
processes that govern many critical cellular functions in living
organisms, such as cell recognition, cell communication and cell
adhesion.1–3 These interactions play critical roles in various human
disorders such as HIV infections and cancers, and therefore exten-
sive efforts have been directed to the development of therapeutic
reagents (e.g., antibodies and small ligands) capable of inhibiting
such interactions. The wide variety of natural or synthetic com-
pounds,4,5 however, imposes a big hurdle on the discovery of com-
pounds that have potential for treating disease. Therefore, it is
highly desirable to develop a high-throughout tool for drug candi-
date screening. In this regard, carbohydrate arrays (glycoarrays),
prepared by attachment of carbohydrates to a solid surface in a
spatially discrete pattern, have received considerable attention
because of their ability to screen a large pool of samples in small
quantities in a high-throughput manner6–8 (e.g., using a 96-well
microplate).
avoids the covalent modification of chemically inert solid supports
(e.g., polystyrene microplates), which is often challenging and
complicated.15,17,25 Noncovalent carbohydrate immobilization can
be achieved by means of hydrophobic interactions. Carbohydrate
arrays of this type are particularly promising because polystyrene
microplates are the most commonly used samplers and compatible
with most microplate readers provided by commercial companies.
Moreover, the microtiter plate approach can be coupled with ELISA
assays to probe the carbohydrate- binding interactions.15 In this
category, microplate-based carbohydrate arrays fabricated from
lipid-bearing carbohydrates for screening for carbohydrate–protein
interactions were first reported by the Wong group.26 In this work,
they developed a number of aliphatic alkyl chains as noncovalent
linkers to attach carbohydrates to microtiter plates and further im-
proved this strategy to synthesize and immobilize sugars to micro-
titer plates in situ so as to avoid the difficult handling of
glycolipids.19 Despite this encouraging result, the alkyl-derivatized
carbohydrates are usually easily detached from polystyrene micro-
plates in the presence of aqueous solutions, presumably due to
their weaker hydrophobic forces relative to solvation forces.
Besides, new microarrays have been fabricated by immobilizing
fluorous-tagged sugars on fluorous-derivatized glass slides. The
noncovalent interactions in the fluorous-based arrays are strong
enough to resist detergents used in binding assays.21 The fluorous
derivatives are, however, rather costly, thus limiting their routine
In recent years, a number of carbohydrate arrays have been
developed by attaching carbohydrates to solid supports such as
glass slides, gold surface and polystyrene microplates through
either covalent or noncovalent immobilization strategies.7–17
Examples include nitrocellulose- and gold-coated glass arrays as
* Corresponding author. Tel.: +852 3400 8686; fax: +852 2364 9932.
E-mail address: bckywong@polyu.edu.hk (K.-Y. Wong).
0008-6215/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.carres.2008.08.021

L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
2933
use in carbohydrate studies. As such, it is highly desirable to devel-
op a strong and cost-effective noncovalent immobilizing agent for
construction of microplate-based carbohydrate arrays.
subsequently allowed to couple with mannose. The desired targets
were purified by column chromatography on silica gel in synthet-
ically useful yields. The anomeric configurations of the glycosylam-
ines were determined by measuring the coupling constants and
chemical shifts of the respective products. Analysis of these data
and those reported in the literature30 indicated that the glycosyl-
It is generally believed that p–p interactions can occur between
phenyl rings.27,28 This interesting phenomenon inspired us to de-
vise a noncovalent aromatic immobilizing agent for constructing
carbohydrate arrays on phenyl-rich polystyrene microplates. Trityl
linkers represent an excellent class of noncovalent immobilizing
agents for this purpose, because the trityl group contains three
amines were of the a-configuration and in the pyranose form.
2.2. Fabrication of trityl-derivatized carbohydrate arrays
phenyl rings that allow
p–p interactions. Thus, trityl-derivatized
carbohydrates are expected to bind strongly to polystyrene micro-
plates. This carbohydrate array can serve as a robust and high-
throughput drug screening tool for searching for promising drug
candidates.
The mannose derivatives were then noncovalently immobilized
on polystyrene microplates to fabricate trityl-derivatized carbo-
hydrate arrays (Fig. 1). To investigate whether 4a–f can bind to
polystyrene microplates, phenol–sulfuric acid assay was per-
formed. This colorimetric assay can determine carbohydrate con-
centrations by reacting phenol with carbohydrates to give
chromogenic products, which absorb strongly at ꢀ490 nm.32 The
absorbance at 490 nm (A490) of 4a–f immobilized on the micro-
plate before and after washing with deionized water was deter-
mined. All the mannose derivatives exhibit strong visible
absorbance at 490 nm (A490 >0.8) compared to the phenol–sulfuric
acid solution (A490 = ꢀ0.1) before washing with deionized water.
The A490 of each mannose derivative remains virtually similar
(Table 1) after washing with deionized water, indicating that the
mannose derivatives are able to bind to the microplate.
Here, we report the fabrication of a new and robust microplate-
based carbohydrate array from trityl-derivatized carbohydrates. In
this approach, the carbohydrate (e.g., mannose) is conjugated with
a trityl group, which is responsible for attaching to polystyrene
surface through an alkyl chain as a spacer. The alkyl linker can
enhance carbohydrate accessibility for protein binding. The trityl-
derivatized mannose performs good immobilization efficiency
against polystyrene microplates and strong resistance against
aqueous washing. The mannose-immobilized carbohydrate array
can probe the interactions with the lectin Con A (Concanavalin
A)29 and can also be used to screen Con A against inhibitors in a
high-throughput manner. In addition, the compatibility of the tri-
tyl immobilizing agent with the disaccharide maltose was also
studied.
2.3. Lectin-binding assay
The abilities of the trityl-derivatized carbohydrate arrays with
4a–f to bind to concanavalin A (Con A) were studied by fluores-
cence spectroscopy. Briefly, the wells immobilized with trityl-
derivatized carbohydrates were first washed with deionized water
and then treated with phosphate buffer containing bovine serum
albumin (to reduce nonspecific binding).33 The wells were then
incubated with fluorescein–Con A, followed by extensive washing
with phosphate buffer. In all cases, the mannose derivatives gave
a stronger fluorescence signal compared to the control (the trityl
2. Results
2.1. Synthesis of trityl-derivatized carbohydrates
A
series of trityl-derivatized mannoses (4a–f) were pre-
pared30,31 (Scheme 1). The preparation of the mannose derivatives
requires only a simple two-step synthetic approach. The trityl
group was first conjugated with a diaminoalkyl chain, which was
HO
HO
H2N
NH2
OH
OH
mannose
H
N
H
N
H
N
n
O
Cl
NH2
methanol
n
n
CH2Cl2
50 oC, 1-2 h
r.t., 24 h
1
3a (n = 2)
3b (n = 4)
3c (n = 6)
3d (n = 8)
4a (n = 2)
4b (n = 4)
4c (n = 6)
4d (n = 8)
2a (n = 2)
2b (n = 4)
2c (n = 6)
2d (n = 8)
2e (n = 10)
2f (n = 12)
H2N
NH2
3e (n = 10)
4e (n = 10)
4f (n = 12)
n
3f (n = 12)
OH
O
HO
HO
OH
O
OH
maltose
O
3f
HO
methanol
H
N
12
OH
HN
50 oC, 1-2 h
5
Scheme 1. Synthetic procedures for trityl-derivatized carbohydrates (4a–f and 5).

2934
L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
Figure 1. Fabrication of a trityl-derivatized carbohydrate array. A 40-
l
L aliquot of the stock solution of each sample (100 mM) was pipetted into a 96-well polystyrene
microplate, which was subsequently dried by evaporation.
25000
Table 1
Retention of various trityl-derivatized carbohydrates (4a–f) subjected to three
washing cycles with deionized water
20000
15000
10000
5000
0
Trityl-derivatized
carbohydrates
Retentiona
(%)
HO
HO
OH
OH
O
4a (n = 2)
4b (n = 4)
4c (n = 6)
4d (n = 8)
4e (n = 10)
4f (n = 12)
4a
4b
4c
4d
4e
4f
70 11
90 11
88 18
99 12
94 19
91 16
HN
HN
control
4a
4b
4c
4d
4e
4f
Trityl-derivatized carbohydrates
Trityl-derivatized mannose
a
The retention of (4a–f) was calculated by dividing the A490 of washed trityl-
derivatized carbohydrates by the A490 of unwashed trityl-derivatized carbohy-
drates. Values are given as mean SD.
Figure 2. Fluorescence signals of fluorescein–Con
immobilized with various trityl-derivatized mannoses. Compounds 4a–f and the
A with carbohydrate arrays
trityl group (control) were immobilized onto microplates in similar quantity
(4 lmol). The carbohydrate array was first washed with deionized water and then
treated with BSA in phosphate buffer for 1 h. The microplate was then incubated
with fluorescein–Con A for 1 h, followed by extensive washing with phosphate
buffer and fluorescence measurements with a microplate reader. The error bars
were obtained by five fluorescence measurements (n = 5).
group) after binding to fluorescein–Con A, indicating that these tri-
tyl-derivatized carbohydrates (4a–f) are able to bind to Con A
(Fig. 2). Moreover, the fluorescence intensity of the complexed
fluorescein–Con A increased with the length of the alkyl linker
(Fig. 2). Because 4f gave the strongest fluorescence signal, this
mannose derivative was used for further studies.
In order to optimize the sensing function of the 4f-glycoarray,
different concentrations of 4f (0–200 mM) were immobilized on
a microplate, and their response was examined by the fluores-
cein–Con A assay. As shown in Figure 3, the fluorescence signal
of complexed fluorescein–Con A increases as a function of 4f con-
centration in the concentration range of 0–100 mM. Further
increasing the concentration of 4f, however, resulted in a fluores-
cence plateau, and even a slight decrease of fluorescence signal ap-
peared at a concentration of 160 mM. Because the carbohydrate
array constructed from 100 mM 4f gave a maximum response, this
concentration was chosen to fabricate trityl-derivatized carbohy-
drate arrays.
similar (Fig. S15, see Supplementary data), indicating that 4f can
be completely immobilized on a microplate within 2 h.
As many protein–carbohydrate interactions take place in an
aqueous medium, the robustness of the 4f-glycoarray in this envi-
ronment was investigated. The resistance of 4f against aqueous
washing was studied by the fluorescein–Con A assay. The fluores-
cence signal of the 4f-glycoarray complexed with fluorescein–Con
A showed no significant decrease compared to the unwashed sam-
ple with that after six washing cycles (Fig. S16, see Supplementary
data). However, in the case of trityl-free mannose subjected to sim-
ilar washing steps, no fluorescein–Con A was found to bind to the
microplate, implying that this sugar is unable to bind to polysty-
rene microplates without the trityl agent (data not shown). No
fluorescence signal was detected when the 4f-glycoarray was
washed with ethanol, indicating that 4f can be easily removed
from microplates by this organic solvent (data not shown). These
observations indicate that trityl-derivatized mannose can bind
strongly to polystyrene microplates and is very resistant against
aqueous washing rather than organic solvent.
We then studied the immobilization efficiency of 4f on polysty-
rene microplates. This mannose derivative was immobilized on a
polystyrene microplate for different time intervals (2–28 h), and
the relative amount of immobilized 4f was examined by the fluo-
rescein–Con A assay as described above. In all cases, the fluores-
cence signals of complexed fluorescein–Con A were virtually

L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
2935
The working range for the 4f-glycoarray was then studied by
the fluorescein–Con A assay. Figure 4 shows the fluorescence
response of the carbohydrate array complexed with different
concentrations of fluorescein–Con A. The fluorescence response
28000
24000
20000
16000
12000
8000
4000
0
of the carbohydrate array shows good linearity up to 30
fluorescein–Con A and reaches a plateau in the concentration range
of 40–50 g/mL of fluorescein–Con A, presumably due to the satu-
lg/mL
l
rated binding of Con A to the carbohydrate array. The detection
limit of the 4f-glycoarray for Con A was determined to be
0.12
2.4. Inhibitor screening
Methyl -mannopyranoside and methyl a-D-glucopyranoside
lg/mL.
a
-D
were used in the competitive binding studies of mannose with
fluorescein–Con A. The BSA-treated 4f-glycoarray was incubated
with fluorescein–Con A in the presence of different concentrations
of the inhibitors. In both cases, the fluorescence signal of fluores-
cein–Con A decreases with the increasing concentration of inhibi-
0
40
80
120
160
200
Concentration of 4f used in immobilization (mM)
Figure 3. Plot of the fluorescence signals of fluorescein–Con A with the 4f-glyco-
array against different concentrations of 4f. Different concentrations of 4f
(0–200 mM) were immobilized on a microplate, and the binding to fluorescein–
Con A was examined by fluorescence measurements. The error bars were obtained
by five fluorescence measurements (n = 5).
tors (Fig. 5). The IC50 values for methyl
a-D-mannopyranoside
and methyl -glucopyranoside were determined to be 2.7 and
a-D
5.7 mM, respectively.
2.5. Carbohydrate compatibility
70000
60000
50000
40000
30000
20000
10000
0
In order to investigate the compatibility of the trityl immobiliz-
ing agent with oligosaccharides, we constructed a carbohydrate
array with maltose (consisting of two glucose units) using the
same approach and conditions and studied its sensing function
by the fluorescein–Con A assay. The phenol–sulfuric acid array
showed that the trityl-derivatized maltose (5) can bind to poly-
styrene microplates, as revealed by the strong A490 of 5 (washed
extensively with deionized water) compared to that of the
phenol–sulfuric acid solution (Fig. S17, see Supplementary data).
Fluorescence measurements indicated that the 5-glycoarray can
also sense the binding to Con A (Fig. S18, see Supplementary data);
the fluorescence response of the carbohydrate array increases lin-
early in the concentration range of 0–12 lg/mL of fluorescein–Con
A. Beyond this concentration range, the fluorescence response of
the carbohydrate array levels off to give a plateau, presumably
arising from the saturated binding of fluorescein–Con A to 5. The
0
10
20
30
40
50
Concentration of FITC-ConA (μg/mL)
lower saturation point of 5 (12
(40 g/mL) is likely to be due to the strong binding of maltose to
Con A.34
The effects of methyl
-glucopyranoside on the Con A/5 binding were investigated
lg/mL) compared to that of 4f
Figure 4. Plot of the fluorescence signals of fluorescein–Con A with the 4f-glyco-
array against different concentrations of fluorescein–Con A. The glycoarray was
washed with deionized water (3 cycles) and then treated with BSA in phosphate
buffer. The microplate was incubated with different concentrations of fluorescein–
Con A (0–50 lg/mL), and the fluorescence signal was measured. The error bars were
obtained by five fluorescence measurements (n = 5).
l
a-D-mannopyranoside and methyl
a-D
by the inhibitor screening assay described above. No significant
30000
25000
20000
15000
30000
25000
20000
15000
IC50=5.73 mM
IC50=2.71 mM
10000
10000
-5
-4
-3
-2
-1
0
1
2
3
-5
-4
-3
-2
-1
0
1
2
3
Log [methyl α-D-glucopyranoside]
Log [methyl α-D-mannopyranoside]
(mM)
(mM)
Figure 5. Determination of the IC50 values of (a) methyl
4f-glycoarray. Fluorescein-Con A (10 g/mL, in phosphate buffer) was first incubated with different concentrations of the inhibitors methyl
-mannopyranoside (10 nM–400 mM) for 1 h. A 40- L portion of each solution was then pipetted into the BSA-treated carbohydrate arrays, followed by incubation for
another 1 h. The microplate was then washed with the same buffer for three cycles (5 min each), and the fluorescence signal of each well was measured.
a-
D-mannopyranoside and (b) methyl
a
-
D-glucopyranoside against the Con A/mannose binding using the
l
a-D
-glucopyranoside and methyl
a-
D
l

2936
L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
inhibition was observed when the concentration of the inhibitors
was increased to 400 mM (i.e., IC50 >400 mM). This observation is
presumedly to be due to the strong interactions of Con A with
maltose.
washing (ꢀ50–100% loss).26 These observations reveal that the tri-
tyl immobilizing agent binds more strongly to microplates than the
alkyl type. This interesting phenomenon is likely to arise from the
strong
p
–p
interactions between the phenyl rings of the trityl
interactions appear to
group and polystyrene. The strong
p–p
allow the trityl-derivatized mannose (4f) to be efficienly immobi-
3. Discussion
lized on the microplate within 2 h.
The trityl-derivatized carbohydrate array devised in this study
can serve as a high-throughput drug screening tool. The 4f-glyco-
Carbohydrate-mediated biological events such as protein–car-
bohydrate interactions are important in many living organisms.
Many human disorders such as cancers and viral infections involve
these biological processes, and therefore the development of new
therapeutic agents for inhibiting protein–carbohydrate interac-
tions has attracted much attention in recent years. In response to
this growing research field, carbohydrate arrays are also under
active development because of their potential application in
high-throughput drug screening. In this regard, noncovalent carbo-
hydrate arrays are an attractive choice because they can be easily
constructed compared to the covalent ones, which usually require
complicated chemical modifications on the solid supports. In the
former, robustness is particularly important, and therefore it is
necessary to use a strong immobilizing agent to firmly anchor
the carbohydrate of interest.
In this study, we developed a new trityl immobilizing agent for
the construction of microplate-based carbohydrate arrays. The tri-
tyl immobilizing agent has a very high propensity to bind to poly-
styrene microplates, as revealed by the phenol–sulfuric acid assay;
the A490 values of 4a–f recorded before and after washing are sim-
ilar, with ꢀ70% and 90% retention for 4a and 4b–f, respectively (Ta-
ble 1). The lower retention for 4a, which has a shorter alkyl spacer,
is presumably due to their lower resistance to aqueous solution.
Previous studies have shown that the linker attached to the so-
lid support confers access to carbohydrate–binding proteins/cells
on the carbohydrate, and the length of the linker affects the bind-
ing of proteins to carbohydrate arrays.8,17,35,36 In our case, we
found that the fluorescence signal of the control is much lower
than those of 4a–f (Fig. 2). Moreover, extending the length of the
alkyl linker can enhance the binding to Con A, as revealed by the
increasing fluorescence signal of fluorescein–Con A with 4a–f
(Fig. 2). This observation can be attributed to the fact that increas-
ing the length of the alkyl linker is likely to lead to the greater
accessibility of the trityl-derivatized mannose to fluorescein–Con
A, which facilitates the binding between these two species. This
observation was also obtained in a previous study.36 We attempted
to further study the effects of alkyl linkers with more than 12 car-
bons. However, these alkyl linkers are not commercially available,
and therefore their effects were not studied.
array can screen Con A against methyl
methyl -glucopyranoside and distinguish the inhibitory activi-
ties of these two inhibitors. The IC50 values determined by the car-
bohydrate array indicate that methyl -mannopyranoside is
more effective in inhibiting the mannose–Con A binding than
methyl -glucopyranoside (Fig. 5). This observation is in good
agreement with previous findings that methyl -mannopyran-
a-D-mannopyranoside and
a-D
a-D
a-D
a-D
oside is a stronger inhibitor of Con A binding than methyl
a-D-
glucopyranoside.37
The trityl immobilizing agent is also compatible with oligosac-
charides. As similar to the case of the mannose derivatives, the
disaccharide maltose can also be immobilized on polystyrene
microplates after modified with the trityl immobilizing agent
(Fig. S17) and used to probe the binding to Con A (Fig. S18). These
observations highlight the high applicability of the trityl immobi-
lizing agent in the construction of carbohydrate arrays with differ-
ent carbohydrates.
In summary, we have developed an effective approach to
fabricating a robust microplate-based carbohydrate array. The
key element in this noncovalent carbohydrate array is the trityl
immobilizing agent, which anchors firmly the carbohydrate of
interest on polystyrene microplates. The trityl-derivatized carbo-
hydrate array is very robust, apparently due to the strong
p–p
interactions between the phenyl rings of the trityl group and
polystyrene. This carbohydrate array can be used to probe pro-
tein–carbohydrate interactions and to screen for inhibitors in a
high-throughput manner. The trityl immobilizing agent is com-
patible with oligosaccharides and therefore has high applicability
in the fabrication of various carbohydrate arrays. Moreover, the
trityl immobilizing agent is more resistant to aqueous washing
than the lipid-type immobilizing agent26 and less expensive than
the fluorous linker.21 With these advantageous properties, the tri-
tyl-derivatized carbohydrate array should find its applications in
a variety of areas, ranging from glycomic research to clinical
diagnosis.
4. Experiment
It has been reported that increasing the density of carbohy-
drates immobilized on the surface can facilitate protein–carbohy-
drate binding.22 Excessive carbohydrate density, however, may
lead to a weakening rather than enhancing effect on the binding
to carbohydrate-binding proteins.16 In our case, the fluorescence
response reached the maximum level with 100 mM 4f, implying
that the well of the microplate can be fully coated at this carbohy-
drate concentration (Fig. 3). With 4f concentration larger than
160 mM, a slight decrease in fluorescence intensity appears, indi-
cating that the mannose–Con A binding becomes less favourable
under this condition.
4.1. Materials and instrumentation
NMR spectra were recorded with a Bruker DPX 400 MHz spec-
trometer. The ESIMSs were measured with
spectrometer. Trityl chloride, 1,2-ethylenediamine, 1,4-diamin-
obutane, 1,6-diaminohexane, 1,8-diaminooctane, 1,10-diaminode-
cane and 1,12-diaminododecane were purchased from
a VG-PLATFORM
Sigma–Aldrich.
Sigma–Aldrich.
D-Mannose and maltose were obtained from
Bovine serum albumin (BSA), methyl
a-D-glucopyranoside
The fluorescein–Con A assays showed that the trityl-derivatized
carbohydrate array is very robust. For example, the 4f-glycoarray
was found to bind to fluorescein–Con A to a similar extent before
and after six cycles of aqueous washing. This observation implies
that the trityl-derivatized mannose can bind strongly to polysty-
rene microplates and is highly resistant to aqueous washing. In
contrast, most alkyl-derivatized carbohydrates were significantly
lost from the polystyrene microplates after five cycles of aqueous
(99%), methyl -mannopyranoside (99%) and polyethylene glycol
a-D
sorbitan monolaurate (Tween 20) were obtained from Sigma–Al-
drich. Fluorescein-labelled Concanavalin A (FITC-Con A, 104 kDa)
was purchased from Vector Laboratories, Inc. Microtiter plates
with a hydrophobic surface (no. 3631, 96 wells, flat clear bottom
black polystyrene, nontreated) were obtained from Corning. All
luminescence measurements were conducted on a BMG Labtech
POLARstar microplate reader equipped with a microcomputer.

L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
2937
The wavelengths of the band pass filters for excitation and emis-
sion were 490 and 520 nm, respectively.
146.73, 128.94, 128.82, 128.08, 126.41 (CHAr’s), 87.77 (C1), 78.33
(C3), 75.20 (C2), 71.94 (C5), 70.84 (C4), 68.06 (C(Ph)3), 62.15
(C6), 45.32, 43.77 (2 ꢁ NHCH2), 30.48, 29.55, 29.48, 27.36 (CH2).
(+) ESIMS: m/z 549.4, [M+H]+; HRESIMS: m/z [M+H]+ calcd for
C33H45N2O5, 549.3328; found, 549.3324.
4.2. Synthesis of N-(N0-tritylaminoalkyl)-
a-D-glycosylamines
(4a–f and 5)
A solution of trityl chloride (1.39 g, 5 mmol, dissolved in 50 mL
of CH2Cl2) was added dropwise to alkyldiamine (2a–f, 20 mmol)
dissolved in 15 mL of CH2Cl2 at room temperature, followed by
stirring at room temperature for 24 h.31 The mixture was then di-
luted with 50 mL of CHCl3 and washed with 5% aq Na2CO3 and
satd brine. The organic phase was dried with MgSO4, and the sol-
vent was evaporated. The concentrated fraction was purified by
column chromatography (silica gel) using 8:2:0.1 CH2Cl2–
MeOH–NH4OH as the eluent to give N-tritylalkyldiamine (3a–f,
yield: 60–70%).
4.2.5. N-(N0-Tritylaminodecyl)-
a-D-mannopyranosylamine (4e)
1H NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 8 Hz, 6H, CHAr’s),
7.25 (t, J ꢀ 8 Hz, 6H, CHAr’s), 7.14 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.87 (s,
1H, H1), 3.60–3.64 (m, 2H, H6), 3.49 (m, 2H, H2, H3), 3.22–3.32
(m, 2H, H4, H5), 2.79–2.93 (m, 2H, NHCH2), 1.91 (m, 2H, NHCH2),
1.19–1.41 (m, 16H, 8 ꢁ CH2). 13C NMR (DMSO-d6, 400 MHz): d
146.73, 128.82, 128.07, 126.41 (CHAr’s), 87.76 (C1), 78.33 (C3),
75.18 (C2), 71.93 (C5), 70.84 (C(Ph)3), 68.05 (C4), 62.14 (C6),
44.20, 43.76 (2 ꢁ NHCH2), 30.48, 29.52, 29.45, 27.35 (CH2). (+)
ESIMS: m/z 577.5, [M+H]+; HRESIMS: m/z [M+H]+ calcd for
C35H49N2O5, 577.3641; found, 577.3632.
N-Tritylalkyldiamine (3a–f, 10 mmol) was dissolved in 10 mL of
anhyd MeOH and then mixed with D-mannose (1.80 g, 10 mmol) at
50 °C with stirring for 1–2 h.30 The solvent was then removed by
evaporation, and the concentrated fraction was purified by column
chromatography (silica gel) using 9:1 CH2Cl2–MeOH as the
4.2.6. N-(N0-Tritylaminododecyl)-
a-D-mannopyranosylamine
(4f)
1H NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 8 Hz, 6H, CHAr’s),
7.25 (t, J ꢀ 7.6 Hz, 6H, CHAr’s), 7.14 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.87
(s, 1H, H1), 3.59–3.64 (m, 2H, H6), 3.49 (m, 2H, H2, H3), 3.22–
3.24 (m, 2H, H4, H5), 2.80–2.93 (m, 2H, NHCH2), 1.92 (t, J ꢀ 4.8 Hz,
2H, NHCH2), 1.19–1.41 (m, 20H, 10 ꢁ CH2). 13C NMR (DMSO-d6,
400 MHz): d 146.71, 128.80, 128.05, 126.40 (CHAr’s), 87.78 (C1),
78.33 (C3), 75.21 (C2), 73.61 (C5), 71.09 (C4), 70.83 (C(Ph)3),
62.00 (C6), 43.74 (NHCH2), 30.48, 29.54, 29.49, 29.06, 27.51,
27.35 (CH2). (+) ESIMS: m/z 605.5, [M+H]+; HRESIMS: m/z [M+H]+
calcd for C37H53N2O5, 605.3954; found, 605.3961.
eluent to give N-(N0-tritylaminoalkyl)-
ines (4a–f, yield: 50–60%).
a-D-manno-pyranosylam-
4.2.1. N-(N0-Tritylaminoethyl)-
a-D-mannopyranosylamine (4a)
1H NMR (DMSO-d6, 400 MHz): d 7.37 (d, J ꢀ 7.6 Hz, 6H, CHAr’s),
7.26 (t, J ꢀ 7.6 Hz, 6H, CHAr’s), 7.15 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.66 (s,
1H, H1), 3.57–3.60 (m, 2H, H6), 3.49 (m, 2H, H2, H3), 3.13–3.24 (m,
2H, H4, H5), 2.84–2.97 (m, 2H, NHCH2), 1.91 (m, 2H, NHCH2). 13C
NMR (CDCl3, 400 MHz): d 146.09, 145.97, 128.71, 128.63, 127.93,
127.80, 126.28 (CHAr’s), 87.42 (C1), 76.87 (C3), 74.79 (C2), 72.01
(C5), 70.90 (C4), 70.83 (C(Ph)3), 60.87 (C6), 45.96, 43.21
(NHCH2CH2). (+) ESIMS: m/z 465.4, [M+H]+; HRESIMS: m/z
[M+H]+ calcd for C27H33N2O5, 465.2389; found, 465.2399.
4.2.7. N-(N0-Tritylaminododecyl)-
a-D-maltosylamine (5)
N-Trityldodecyldiamine (3f, 10 mmol) was dissolved in 10 mL
of anhyd MeOH and then mixed with maltose (3.60 g, 10 mmol)
at 50 °C with stirring for 1–2 h.30 The solvent was then evaporated,
and the concentrated fraction was purified by column chromato-
graphy (silica gel) using 9:1 CH2Cl2–MeOH as the eluent to give
4.2.2. N-(N0-Tritylaminobutyl)-
a-D-mannopyranosylamine (4b)
1H NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 8 Hz, 6H, CHAr’s),
7.25 (t, J ꢀ 8 Hz, 6H, CHAr’s), 7.14 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.85 (s,
1H, H1), 3.59–3.63 (m, 2H, H6), 3.46–3.52 (m, 2H, H2, H3), 3.21–
3.26 (m, 2H, H4, H5), 2.82–2.93 (m, 2H, NHCH2), 1.92 (m, 2H,
NHCH2), 1.34–1.44 (m, 4H, CH2CH2). 13C NMR (DMSO-d6,
N-(N0-tritylaminododecyl)- -maltosylamine 5 (yield: 70%). 1H
a-D
NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 7.6 Hz, 6H, CHAr’s), 7.22
(t, J ꢀ 7.2 Hz, 6H, CHAr’s), 7.12 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 5.0 (d,
J ꢀ 3.6 Hz, 1H, CH), 4.31 (d, J ꢀ 3.2 Hz, 1H, CH), 3.37–3.69 (m,
10H, CH), 3.24–3.29 (m, 2H, CH), 2.93–3.08 (m, 2H, NHCH2),
1.92–1.94 (m, 2H, NHCH2), 1.35–1.39 (m, 4H, 2 ꢁ CH2), 1.16–1.18
(m, 16H, 8 ꢁ CH2). 13C NMR (DMSO-d6, 400 MHz): d 146.69,
128.79, 128.04, 126.41 (CHAr’s), 101.41, 91.28, 80.73, 77.63, 76.44,
73.88, 73.51, 70.84, 70.41, 61.33 (CH), 49.09, 46.05, 43.72 (NHCH2),
30.49, 29.58, 29.47, 29.06, 27.34, 26.90 (CH2). (+) ESIMS: m/z 767.2,
[M+H]+; HRESIMS: m/z [M+H]+ calcd for C43H63N2O10, 767.4483;
found, 767.4490.
400 MHz):
d 146.23, 128.66, 128.36, 128.30, 127.79, 126.23
(CHAr’s), 87.85 (C1), 76.77 (C3), 74.95 (C2), 72.07 (C5), 70.96 (C4),
70.88 (C(Ph)3), 66,24 (C6), 45.87, 43.53 (NHCH2), 28.60, 27.89
(CH2CH2). (+) ESIMS: m/z 493.6, [M+H]+; HRESIMS: m/z [M+H]+
calcd for C29H37N2O5, 493.2702; found, 493.2699.
4.2.3. N-(N0-Tritylaminohexyl)-
a-D-mannopyranosylamine (4c)
1H NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 8 Hz, 6H, CHAr’s),
7.25 (t, J ꢀ 7.6 Hz, 6H, CHAr’s), 7.14 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.86
(s, 1H, H1), 3.59–3.65 (m, 2H, H6), 3.42–3.49 (m, 2H, H2, H3),
3.21–3.26 (m, 2H, H4, H5), 2.83–2.91 (m, 2H, NHCH2), 1.91 (m,
2H, NHCH2), 1.19–1.43 (m, 8H, 4 ꢁ CH2). 13C NMR (DMSO-d6,
4.3. Fabrication of trityl-derivatized carbohydrate arrays
A series of 100 mM stock solutions of 4a–f and 5 (in MeOH)
were prepared. A 40-lL portion of each sample was pipetted into
400 MHz):
d
146.31, 146.25, 128.65, 128.33, 127.75, 127.51,
126.17 (CHAr’s), 87.77 (C1), 74.94 (C3), 72.00 (C2), 70.87 (C5),
66.26 (C4), 61.15 (C(Ph)3), 61.07 (C6), 45.86, 43.61 (2 ꢁ NHCH2),
30.98, 30.07, 27.48, 27.35 (4 ꢁ CH2). (+) ESIMS: m/z 521.4,
[M+H]+; HRESIMS: m/z [M+H]+ calcd for C31H41N2O5, 521.3015;
found, 521.3022.
a 96-well polystyrene microplate, which was subsequently dried
by evaporation. Each well was then washed three times with
250
4.4. Phenol–sulfuric acid assay
A 40- L portion of trityl-derivatized carbohydrates (100 mM
lL of deionized water.
4.2.4. N-(N0-Tritylaminooctyl)-
a
-D
-mannopyranosylamine (4d)
l
1H NMR (DMSO-d6, 400 MHz): d 7.36 (d, J ꢀ 8 Hz, 6H, CHAr’s),
7.25 (t, J ꢀ 8 Hz, 6H, CHAr’s), 7.14 (t, J ꢀ 7.2 Hz, 3H, CHAr’s), 3.87 (s,
1H, H1), 3.60–3.64 (m, 2H, H6), 3.49 (m, 2H, H2, H3), 3.22–3.24
(m, 2H, H4, H5), 2.81–2.93 (m, 2H, NHCH2), 1.91 (m, 2H, NHCH2);
1.19–1.41 (m, 12H, 6 ꢁ CH2). 13C NMR (DMSO-d6, 400 MHz): d
4a–f and 5 in MeOH) was pipetted into a 96-well microplate, and
the plate was incubated at room temperature to allow the solvent
to evaporate. The wells were then washed with deionized water (3
cycles). The phenol–sulfuric acid assay was performed according to
a literature method.32 Briefly, a 5% phenol solution (14
lL) was

2938
L. Zou et al. / Carbohydrate Research 343 (2008) 2932–2938
added to each well, followed by the addition of concentrated H2SO4
(70 L). The mixture was incubated for 30 min at room tempera-
Supplementary data
l
ture, and the absorbance at 490 nm (A490) measured to determine
the amount of mannose immobilized on the microtiter plate. For
comparison, similar visible absorption measurements were also
performed with the phenol–sulfuric acid solution (without man-
nose) and the trityl-derivatized carbohydrate controls (4a–f and
5), which were prepared as described above except that no wash-
ing was carried out. The amount of immobilized 4a–f and 5 was
estimated from the ratio of the A490 of immobilized 4a–f and 5
(subjected to 3 washing cycles with deionized water) to the A490
of the corresponding control (unwashed).
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.carres.2008.08.021.
References
1. Bertozzi, C. R.; Kiessling, L. L. Science 2001, 291, 2357–2364.
2. Haseley, S. R. Anal. Chim. Acta 2002, 457, 39–45.
3. Lis, H.; Sharon, N. Curr. Opin. Struct. Biol. 1991, 1, 741–749.
4. Sharon, N. Adv. Exp. Med. Biol. 1996, 408, 1–8.
5. Flitsch, S. L.; Ulijn, R. V. Nature 2003, 421, 219–220.
6. Fukui, S.; Feizi, T.; Galustian, C.; Lawson, A. M.; Chai, W. Nat. Biotechnol. 2002,
20, 1011–1017.
7. Jelinek, R.; Kolusheva, S. Chem. Rev. 2004, 104, 5987–6015.
8. Love, K. R.; Seeberger, P. H. Angew. Chem., Int. Ed. 2002, 41, 3583–3586.
9. Culf, A. S.; Cuperlovic-Culf, M.; Ouellette, R. J. Omics 2006, 10, 289–310.
10. de Paz, J. L.; Seeberger, P. H. QSAR Comb. Sci. 2006, 25, 1027–1032.
11. Feizi, T.; Fazio, F.; Chai, W.-g.; Wong, C.-H. Curr. Opin. Struct. Biol. 2003, 13, 637–
645.
12. Mellet, C. O.; Fernandez, J. M. G. ChemBioChem 2002, 3, 819–822.
13. Shin, I.; Park, S.; Lee, M.-r. Chem. Eur. J. 2005, 11, 2894–2901.
14. Wang, D.-N. Proteomics 2003, 3, 2167–2175.
15. Disney, M. D.; Seeberger, P. H. Drug Dis. Today: Targets 2004, 3, 151–158.
16. Dyukova, V. I.; Shilova, N. V.; Galanina, O. E.; Rubina, A. Y.; Bovin, N. V. Biochim.
Biophys. Acta 2006, 1760, 603–609.
17. Monzo, A.; Guttman, A. QSAR Comb. Sci. 2006, 25, 1033–1038.
18. Bryan, M.; Fazio, F.; Lee, H.; Huang, C.; Chang, A.; Best, M.; Calarese, D.; Blixt, C.;
Paulson, J.; Burton, D.; Wilson, I.; Wong, C.-H. J. Am. Chem. Soc. 2004, 126,
8640–8641.
4.5. Con A binding studies
Trityl-derivatized carbohydrates (4a–f and 5) were immobilized
on 96-well microplates as described above and subsequently incu-
bated with 3% BSA in PBST buffer (phosphate-buffered saline con-
taining 0.2% Tween 20, 0.1 M, pH 6.8) for 1 h. BSA is widely used as
a blocking agent to inhibit nonspecific binding by hydrophobic
interactions in many ELISA, immunoblotting and immunohisto-
chemical studies.38,39 The BSA-treated wells were then incubated
with FITC-Con A in PBST buffer for 1 h, and subsequently washed
with the same buffer three times (5 min each). The binding of
the Con A/mannose complex was monitored by fluorescence mea-
surements using a microplate reader. For comparison, similar
immobilization and fluorescence assays were performed with the
trityl group (control).
19. Fazio, F.; Bryan, M. C.; Blixt, O.; Paulson, J. C.; Wong, C. H. J. Am. Chem. Soc. 2002,
124, 14397–14402.
20. Houseman, B.; Mrksich, M. Chem. Biol. 2002, 9, 443–454.
21. Ko, K.-S.; Jaipuri, F. A.; Pohl, N. L. J. Am. Chem. Soc. 2005, 127, 13162–13163.
22. Park, S.; Shin, I. Angew. Chem., Int. Ed. 2002, 41, 3180–3182.
23. Ratner, D. M.; Adams, E. W.; Su, J.; O’Keefe, B. R.; Mrksich, M.; Seeberger, P. H.
ChemBioChem 2004, 5, 379–382.
24. Wang, D.; Liu, S.; Trummer, B. J.; Deng, C.; Wang, A. Nat. Biotechnol. 2002, 20,
275–281.
25. Larsen, K.; Thygesen, M. B.; Guillaumie, F.; Willats, W. G. T.; Jensen, K. J.
Carbohydr. Res. 2006, 341, 1209–1234.
26. Bryan, M.; Plettenburg, O.; Sears, P.; Rabuka, D.; Wacowich-Sgarbi, S.; Wong, C.
Chem. Biol. 2002, 9, 713–720.
27. Muller-Dethlefs, K.; Hobza, P. Chem. Rev. 2000, 100, 143–167.
28. Waters, M. L. Curr. Opin. Chem. Biol. 2002, 6, 736–741.
29. Concanavalin A as a Tool; Bittiger, H., Schnebli, H. P., Eds.; John Wiley & Sons:
New York, 1976.
4.6. Inhibitor screening
Compounds 4f and 5 (40
well microplates as described above, followed by incubation with
3% BSA in PBST buffer. FITC-Con A (10 g/mL, in PBST buffer) was
first mixed with different concentrations of the inhibitors methyl
-glucopyranoside and methyl -mannopyranoside (10 nM–
400 mM), and each mixture was incubated for 1 h. A 40- L portion
lL, 100 mM) were immobilized on 96-
l
a-
D
a-D
l
of each solution was then pipetted into the BSA-treated carbo-
hydrate arrays, followed by incubation for another 1 h. The wells
were then washed with PBST buffer for three cycles (5 min each),
and the fluorescence signal of each well was measured by a micro-
plate reader.
30. Hayes, W.; Osborn, H. M. I.; Osborne, S. D.; Rastall, R. A.; Romagnoli, B.
Tetrahedron 2003, 59, 7983–7996.
31. Matsoukas, J. M.; Hondrelis, J.; Agelis, G.; Barlos, K.; Gatos, D.; Ganter, R.;
Moore, D.; Moore, G. J. J. Med. Chem. 1994, 37, 2958–2969.
32. Saha, S. K.; Brewer, C. F. Carbohydr. Res. 1994, 254, 157–167.
33. Delgado, A. D. S.; Leonard, M.; Dellacherie, E. Langmuir 2001, 17, 4386–4391.
34. Al-Arhabi, M.; Mrestani, Y.; Richter, H.; Neubert, R. H. H. J. Pharm. Biomed. Anal.
2002, 29, 555–560.
35. Chaki, N.; Vijayamohanan, K. Biosens. Bioelectron. 2002, 17, 1–12.
36. Park, S.; Lee, M.-R.; Pyo, S.-J.; Shin, I. J. Am. Chem. Soc. 2004, 126, 4812–4819.
37. Clegg, R. M.; Loontiens, F. G.; Vanlandschoot, A.; Jovin, T. M. Biochemistry 1981,
20, 4687–4692.
Acknowledgements
We acknowledge support from the Hong Kong Polytechnic
University, Sun Yat-Sen University and the Area of Excellence Fund
of the Hong Kong University Grants Committee (AoE/P-10/01).
38. Ratner, D. M.; Seeberger, P. H. Curr. Pharm. Des. 2007, 13, 173–183.
39. Zhou, X.; Zhou, J. Biosens. Bioelectron. 2006, 21, 1451–1458.