Rapid-throughput competitive colorimetric assay based on monosaccharide-capped gold nanoparticles for detecting lectin-protein interactions

Article abstract of DOI:10.1002/cplu.201200014

Identification of protein binding partners is one of key challenges in proteomics. A rapid-throughput competitive colorimetric assay is presented that uses gold nanoparticles capped by sugars such as mannopyranoside (Man-GNPs), N-acetylglucosamine (GlcNAc-GNPs), glucose (Glc-GNPs), or N-acetylgalactosamine (GalNAc-GNPs). The assay expediently detects protein-protein interactions in solution, particularly protein-lectin interactions. The competitive assays were conducted in microtiter plates; ten proteins (two glycoproteins and eight lectins) and three sugar-binding lectins (concanavalin A (ConA), wheat germ agglutinin (WGA), and Ricinus communis agglutinin (RCA₁₂₀)) were combinatorially arranged in a 30-well plate, and constant concentrations of the monosaccharide-capped GNPs (sugar-GNPs) were added to each well. If interactions occurred between the proteins, the sugar-GNPs retained their burgundy color. If no interactions occurred between the proteins, the sugar-GNPs were agglomerated by the corresponding binding lectin, thus producing a blue color. Several new binding pairs were identified for the first time by using this assay, and the binding constants and stoichiometric ratios were determined on the basis of the wavelength shifts. The results were further verified by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and fluorescence resonance energy transfer (FRET) spectroscopy. The assay is very sensitive, requiring only nanomolar protein concentrations.

Full text of DOI:10.1002/cplu.201200014

DOI: 10.1002/cplu.201200014
Rapid-Throughput Competitive Colorimetric Assay Based
on Monosaccharide-Capped Gold Nanoparticles for
Detecting Lectin–Protein Interactions
Charng-Sheng Tsai and Chao-Tsen Chen*[a]
Identification of protein binding partners is one of key chal-
lenges in proteomics. A rapid-throughput competitive colori-
metric assay is presented that uses gold nanoparticles capped
by sugars such as mannopyranoside (Man–GNPs), N-acetylglu-
cosamine (GlcNAc–GNPs), glucose (Glc–GNPs), or N-acetylgalac-
tosamine (GalNAc–GNPs). The assay expediently detects pro-
tein–protein interactions in solution, particularly protein–lectin
interactions. The competitive assays were conducted in micro-
titer plates; ten proteins (two glycoproteins and eight lectins)
and three sugar-binding lectins (concanavalin A (ConA), wheat
germ agglutinin (WGA), and Ricinus communis agglutinin
(RCA120)) were combinatorially arranged in a 30-well plate, and
constant concentrations of the monosaccharide-capped GNPs
(sugar–GNPs) were added to each well. If interactions occurred
between the proteins, the sugar–GNPs retained their burgundy
color. If no interactions occurred between the proteins, the
sugar–GNPs were agglomerated by the corresponding binding
lectin, thus producing a blue color. Several new binding pairs
were identified for the first time by using this assay, and the
binding constants and stoichiometric ratios were determined
on the basis of the wavelength shifts. The results were further
verified by sodium dodecyl sulfate polyacrylamide gel electro-
phoresis (SDS-PAGE) and fluorescence resonance energy trans-
fer (FRET) spectroscopy. The assay is very sensitive, requiring
only nanomolar protein concentrations.
Introduction
Protein–protein interactions play key roles in structural and
functional organization of living cells, such as signal transduc-
tion, cell–cell recognition, regulation of intracellular transport
pathways, gene expression, and enzyme activities.[1] However,
many pairs of interacting proteins remain unknown. Identifying
and mapping such pairs would further our understanding of
which protein components are involved and how the protein
network operates, and thus advance the exploration of proteo-
mics.
tive primary sensory neurones.[25] However, many carbohydrate
functions remain unknown. Discovering the related functions
would provide valuable information on biological processes
and facilitate the development of potent biomedical agents.
Gold nanoparticles can be modified with various molecular
receptors and biomolecules, including DNA or proteins that
can be used to control particle aggregation through simple
host–guest interactions or complementary hybridization, thus
producing distinct colors.[26,27] Gold nanoparticles capped with
various monosaccharides (sugar–GNPs) are particularly interest-
ing because they can be regarded as glycoform mimicries of
cellular proteoglycans, glycoproteins, or glycolipids and used
for purification or detection, and for assaying glyco-binding
proteins such as lectins.[28–31] In light of these studies and moti-
vated by a desire to examine glycoprotein functions and relat-
ed issues, in our previous studies we devised a gold-nanoparti-
cle-based competitive colorimetric assay based on mannopyra-
noside-modified gold nanoparticles (Man–GNPs) for detecting
possible concanavalin A (ConA)–protein interactions in solu-
tion.[32,33] The assay allows the direct visualization and quantifi-
cation of interactions without protein labeling or the use of
specialized instruments. The approach uses Man–GNPs (Fig-
Although a variety of techniques have been developed for
investigating protein–protein interactions, including protein
microarrays,[2,3] two-hybrid analysis,[4] immunoassays,[5] mass
spectrometric analysis by affinity capture,[6] proximity biotinyla-
tion,[7] fluorescence spectroscopy, chemical cross-linking of in-
teracting proteins, and calorimetry,[8] they rely heavily on speci-
alized instruments or highly specific expertise to perform ex-
periments in vivo or in vitro. A facile, rapid, label-free, and sen-
sitive tool for identifying this type of interaction, especially at
a rapid throughput and possibly extending to a high-through-
put fashion, remains a daunting challenge in proteomics re-
search.[9,10]
Specific interactions between carbohydrates and proteins
have been found in a wide variety of important biological and
pathological processes,[11–15] such as cell–cell communication,
cell adhesion,[16] fertilization,[17] differentiation,[18] development
of the mammalian nervous system,[19,20] inflammation, tumor
cell metastasis,[21] and infection of host cells by bacteria[22] and
viruses.[23,24] There is increasing evidence to suggest that some
lectins such as BSI-B4 may be selectively expressed by nocicep-
[a] Dr. C.-S. Tsai, Prof. Dr. C.-T. Chen
Department of Chemistry
National Taiwan University
No. 1, Sec. 4, Roosevelt Road, Taipei, 10617 (Taiwan (R.O.C))
E-mail: chenct@ntu.edu.tw
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cplu.201200014.
314
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2012, 77, 314 – 322

Protein Interactions
acetylglucosamine (GlcNAc) and N-acetylgalactosamine
(GalNAc) are ubiquitous in glycoproteins, proteoglycans, and
glycolipids and explains their pivotal roles in biological func-
tions.[42] Moreover, O-GlcNAc, the modification of serine (Ser)
and threonine (Thr) residues of nuclear and cytoplasmic pro-
teins with O-linked a-N-acetylglucosamine, is one of a growing
number of posttranslational modifications of proteins thought
to modulate the function or activity of proteins in cells.[7, 43–46]
The GalNAc that is a-linked to Thr or Ser represents the core
unit, known as the Tn antigen, in mucin-type glycoproteins. In
eukaryotes, mucins are the most prevalent type of O-glycans,
constituting a polydisperse group of glycoproteins and proteo-
glycans involved in inflammation and cellular recognition.[47]
Thus, in addition to Man–GNPs, glucopyranoside-modified
Figure 1. Schematic illustration of colorimetric detection of protein–protein
interactions.
GNPs
(Glc–GNPs),
N-acetylglucosamine-modified
GNPs
ure 1A) to capture ConA, which binds specifically a-d-manno-
syl and a-d-glucosyl residues at the terminal position of rami-
fied structures from B-glycans. Concanavalin A may interact
with diverse receptors such as rhodopsin, lipoproteins,[34] im-
munoglobulins, and carcinoembryonic antigen through multi-
valent interactions, thus red-shifting the surface plasmon reso-
nance (SPR) absorption of the colloidal particles and changing
the color from burgundy to blue (Figure 1B). However, in the
presence of protein Y, which may be capable of interacting
with ConA, sugar-modified GNPs remain dispersed, so the bur-
gundy color is retained (Figure 1C). Among the seven proteins
we tested for interactions with ConA, four (thyroglobulin,[35]
bandeiraea simplicifolia lectin I (BS-I), soybean agglutinin (SBA),
and Maackia amurensis leukoagglutinin (MAL)) were found to
have very dramatic effects on the absorption spectrum of
Man–GNP–ConA. Moreover, we determined for the first time
that BS-I, SBA, and MAL interact with ConA. Because the extent
of agglutination is associated with the binding affinity of pro-
teins X and Y, the magnitude of the absorption wavelength
change could be used to discriminate among different pro-
teins.
(GlcNAc–GNPs), and N-acetylgalactosamine-modified GNPs
(GalNAc–GNPs) were fabricated to demonstrate the general ap-
plicability of the assay.
The carbohydrate derivatives (Man-SH, Glc-SH, GlcNAc-SH,
and GalNAc-SH) used for capping gold nanoparticles share the
same structural features (Figure 2). They all consist of a cognate
substrate that enables specific lectin recognition, a spacer that
provides the necessary flexibility, a long alkyl chain with a mer-
capto group that promotes the formation of a self-assembled
monolayer, and an immobilization unit that attaches to the
surface of gold nanoparticles.
Figure 2. Structural features of carbohydrate derivatives (Man-SH, Glc-SH,
GlcNAc-SH, and GalNAc-SH) used for capping gold nanoparticles.
Results and Discussion
Herein, we extend this rather facile, sensitive detection
methodology by altering the sugar entity and binding lectins
to investigate other biologically important sugar–lectin interac-
tions and possible lectin–protein interactions. The reasons for
choosing lectins for investigation are manifold.[36–39] First, lec-
tins play important roles in biological and medical research,
such as immunology, cell biology, cell structure, and cancer re-
search.[21] Lectin-based antiadhesion drugs in pill form may be
used to treat infection, inflammation, heart disease, and even
cancer.[40,41] Gaining a deeper understanding of the lectins in
terms of their functions and characteristics will shed lights on
drug developments and reveal unknown functions. Second,
lectins are highly specific for their sugar moieties and often oli-
gomerize into homodimers, homotrimers, and homotetramers,
which increases their avidity for multivalent ligands. These
structural features render the lectins well adapted to the com-
petitive colorimetric assay and its possible development into
a high-throughput format for expedient detection of possible
lectin–protein interactions in solution.
Synthesis of carbohydrate derivatives (Man-SH, Glc-SH,
GlcNAc-SH, and GalNAc-SH)
To elucidate the role of the electronic nature of the spacer in
the interaction between sugar–GNPs and lectins, two different
synthetic routes were designed to attain carbohydrate deriva-
tives having hydrophilic or hydrophobic spacers for capping
the gold nanoparticles, as depicted in Scheme 1. The Man-SH
and GlcNAc-SH carbohydrate derivatives having an ethylene
glycol moiety to impart the required hydrophilic characteristics
were synthesized using a previously reported method.[32] The
synthesis of GlcNAc-SH is described here to illustrate the pro-
cess. N-acetylglucosamine (GlcNAc) was first activated using
acetyl chloride[48] and subsequently treated with 8-azido-3,6-
dioxa-1-octyl-ol in the presence of silver triflate and 1,1,3,3-tet-
ramethylurea and provided glycosylated peracetate 1. Under
Staudinger conditions, the azido group was reduced to the
corresponding amine and coupled with undec-10-enoic acid,
which was obtained by Jones oxidization of commercially
available w-undecylenyl alcohol, in the presence of DCC to
yield compound 2. The terminal olefin was then converted
The vast majority of glycoproteins have different terminal
sugars and sometime multiple terminal sugars. For example, N-
ChemPlusChem 2012, 77, 314 – 322
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chempluschem.org
315

C.-T. Chen and C.-S. Tsai
Fabrication of sugar–GNPs
(Man–GNPs, Glc–GNPs, GlcNAc–
GNPs, and GalNAc–GNPs)
The Frens method was chosen
to prepare solutions of gold
nanoparticles with diameter of
32 nm that display the desired
optical properties (i.e. a large
plasmonic band) and a uniform
particle distribution.[52] The gen-
eral procedure for preparing
sugar–GNPs is as follows. To the
freshly prepared solution of gold
nanoparticles (32 nm; 100 mL)
was added thiol-derivatized
sugar pyranoside (100, 150, or
200 mL, 1.0 mm). Self-assembly of
the pyranosides was facilitated
by leaving the solution for
a period of 15 hours with gentle
stirring. The resulting colloidal
solution was centrifuged at
7000 rpm with an Eppen-
dorf 5810R centrifuge for 15 mi-
nutes at 108C to precipitate the
capped GNPs. After removal of
the supernatant containing the
unchanged thiol-derivatized pyr-
anosides, the precipitated sugar-
capped GNPs were resuspended
in an equal volume of 0.01%
aqueous sodium citrate solution
to obtain an approximate con-
centration of 0.29 nm and then
Scheme 1. Synthesis of GlcNAc-SH, GalNAc-SH, and Glc-SH. AIBN=2,2’-azobisisobutyronitrile, DCC=1,3-dicyclo-
hexylcarbodiimide, DMAP=4-dimethylaminopyridine, HOBt=1-hydroxybenzotriazole, Tf=trifluoromethanesulfon-
yl, THF=tetrahydrofuran, TMU=N,N,N’,N’-tetramethylurea.
into thioacetate 3 through radical-type addition with thioacetic
acid in the presence of AIBN under photolytic conditions.[49]
Ester and thioester groups were simultaneously saponified by
stirring with sodium methoxide in dry methanol at room tem-
perature to afford the b form of GlcNAc-SH in 80% yield. A hy-
drophobic spacer was incorporated through an alternative
route. Lewis acid BF3·Et2O was used as a catalyst; heating the
commercially available N-acetylgalactosamine (GalNAc) at 708C
for 2 hours and subsequent acetylation gave the desired a-
anomer 4.[50] Subsequent photochemical addition of cystamine
hydrochloride to 4 provided 5 in the a form in 85% yield.[51]
Coupling 5 with undec-10-enoic acid in the presence of DCC
and HOBt yielded 6, which was converted into GalNAc-SH
using radical addition of thioacetic acid and subsequent sapo-
nification described in the final two steps of GlcNAc-SH prepa-
ration. The synthetic route to Glc-SH is very similar to that of
GalNAc-SH synthesis. The only difference is in the amide cou-
pling step. The activated 11-mercaptoundecanoic acid N-hy-
droxysuccinimide ester was used instead of undec-10-enoic
acid for the coupling, and the acetate groups were then sapo-
nified by stirring with sodium carbonate in dry methanol at
room temperature to afford Glc-SH.
sonicated to ensure that they were well suspended in the solu-
tion. The average numbers of glycopyranosides per gold nano-
particle were estimated by elemental analysis; they are 9000
(Man-SH), 8000 (GlcNAc-SH), 7500 (GalNAc-SH), and 5000 (Glc-
SH).[53]
Lectin specificity of sugar–GNPs
Concanavalin A, which can recognize glycoproteins with either
a-Man or a-Glc as a terminal sugar, is known to facilitate the
agglomeration of Man–GNPs through multivalent ligand–pro-
tein interactions, thus producing a blue solution.[32] However,
the proteins capable of agglomerating Glc–GNPs, GlcNAc–
GNPs, and GalNAc–GNPs await assessment. Accordingly, Glc–
GNPs, GlcNAc–GNPs, and GalNAc–GNPs maintained at a con-
centration of 0.29 nm in 0.01% aqueous sodium citrate solu-
tion were titrated with various concentrations of their binding
proteins, ConA, WGA, and RCA120, respectively. The results re-
vealed that 50 mgmLÀ1 of WGA was required to achieve the
maximum absorption shift, whereas only 10 mgmLÀ1 of ConA
and RCA120 were needed (see the Supporting Information, Fig-
ure S1a,c,e,g). We reasoned that differences in the number of
316
www.chempluschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2012, 77, 314 – 322

Protein Interactions
binding sites are responsible for the discrepancy. ConA and
RCA120, which have four sugar-binding sites, could effectively
promote the agglomerations of Glc–GNPs and GalNAc–GNPs.
WGA, which exists as a dimer, not a tetramer like ConA and
RCA120, has only two GlcNAc binding sites to interact with
GlcNAc–GNPs, and thus requires more WGA to reach the same
degree of agglomeration. Note that the absorption maximum
and titration profiles of Man–GNPs and Glc–GNPs are very simi-
lar in the presence of ConA, thus indicating that the electronic
nature of the spacer plays a minor role in lectin recognition.
Presumably, the rather shallow substrate binding pocket of lec-
tins accounted for this observation.
The specificity of these sugar–GNPs was further exploited by
incubation with ten proteins (thyroglobulin, ovalbumin, and
eight lectins).[54] The respective absorption spectra were re-
corded after 30 minutes of incubation (Figure S1b,d,f,h) at
room temperature. The changes in absorption wavelength
(Dl) for sugar–GNPs relative to 528 nm are shown in Figure 3.
They range from 14–60 nm depending on the binding strength
and number of interaction sites. The color changes resulting
from agglomeration are quite discernible and can be observed
by the naked eye. The corresponding photo images of the sol-
utions are shown in the inset of Figure 3. ConA was found to
agglomerate Man–GNPs and Glc–GNPs, causing a large red-
shift in the SPR band; RCA120 also agglomerated Man–GNPs
and Glc–GNPs but with smaller red-shifts, thus indicating that
its binding strength to these GNPs is weaker than that of
ConA. The rest of the proteins had no significant effect on the
SPR band. For GlcNAc–GNPs, a red-shifted SPR band was ob-
served in the presence of WGA, RCA120, and ConA. The GlcNAc
binding lectin, WGA, has a large effect on the SPR band of
GalNAc–GNPs, whereas ConA has a moderate effect, thus re-
flecting the difference in binding strength and moderate lectin
specificity. Neither BS-I, which has a-Gal and a-GalNAc as cog-
nate substrates, nor BS-II, which has GlcNAc as a cognate sub-
strate, exhibit significant effects on the SPR bands of GalNAc–
GNPs and GlcNAc–GNPs. This observation could be attributed
to the fact that lectin–substrate recognition sometime requires
synergetic binding of multiple terminal sugars or to the partici-
pation of the penultimate sugar residue.[36]
Figure 3. Absorption wavelength change profiles of sugar–GNPs (Man–
GNPs, Glc–GNPs, GlcNAc–GNPs, and GalNAc–GNPs) in the presence of two
glycoproteins and eight lectins. Inset: Corresponding photo images of the
same series of proteins interacting with three sugar–GNPs (Man–GNPs,
GlcNAc–GNPs, and GalNAc–GNPs) placed in the microtiter plate. Blue indi-
cates that the tested protein interacts with glycopyranosides protruding
from the surface of the GNPs through multivalency to facilitate the agglom-
eration of GNPs. Burgundy indicates that no interaction occurs between the
glycopyranosides and the tested protein.
ConA/Man–GNPs exhibit a burgundy color, thus showing that
ConA interacts with those proteins that coexist in the solu-
tions. Notably, when comparable amounts of ConA and RCA120
were added to Man—GNPs, precipitation was observed within
5 minutes. This phenomenon can be ascribed to oversaturation
of the Man–GNPs with ConA and RCA120, which are both capa-
ble of agglomerating Man–GNPs. The second row reveals that
only thyroglobulin and ovalbumin interact with WGA, respec-
tively, because the solutions remain burgundy. The color blue
was observed in the presence of other putative proteins, thus
indicating that GlcNAc–GNPs were effectively agglomerated by
WGA through the binding of GlcNAc protruding from the sur-
face of the GNPs to WGA. For RCA120, a burgundy color ap-
peared only in the presence of thyroglobulin, thus indicating
that thyroglobulin interacts strongly with RCA120. In the pres-
ence of other proteins such as SBA, BS-I, BS-II, and ConA,[56] the
solution was blue with a burgundy tint, thus reflecting incom-
plete intervention by the cross-linked GalNAc–GNPs mediated
by RCA120. Thus, we speculate that these proteins exhibit mod-
The combination of a visually detectable color change and
reasonably specific response to certain lectins makes these
sugar–GNPs efficient, selective colorimetric sensor for their
binding lectin under the solution phase conditions employed.
Because Man–GNPs and Glc–
erate binding with RCA120.
The observations described above suggest that investigation
of many more lectin–protein interactions by this assay is highly
feasible simply by replacing the currently used pyranosides
GNPs show very similar absorp-
tion profiles in the presence of
various proteins as well as the
same lectin specificity, only
Man–GNPs were used for further
study.
As the first row of Figure 4
shows, the solutions containing
Figure 4. Photo image of high-throughput competitive colorimetric assay: Columns correspond to ten proteins
(two glycoproteins and eight lectins); rows correspond to three lectins (ConA, WGA, and RCA120) that were added
to each microtiter well, followed by the addition of the sugar–GNPs shown on the left. Burgundy indicates that
the added proteins interacted with the lectin, preventing lectin-mediated agglomeration. Blue indicates that no
interaction occurred between the lectin and the protein.
thyroglobulin/ConA/Man–GNPs,
ovalbumin/ConA/Man–GNPs,[55]
BS-I/ConA/Man–GNPs,
SBA/
ConA/Man–GNPs, and MAL/
ChemPlusChem 2012, 77, 314 – 322
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chempluschem.org
317

C.-T. Chen and C.-S. Tsai
with other sugar entities complementary to the lectins of inter-
est. Moreover, this methodology could be applied to the mea-
surement of various forms of biomolecular interactions and
the screening of drugs that could disrupt protein–protein in-
teractions.
Application of SDS-PAGE and FRET to further confirm the
observed lectin–protein interactions
Figure 6. a) Overlaid absorption spectra of Man–GNPs (0.29 nm) in the pres-
ence of the protein mixture. b) 12% SDS-PAGE used to separate ConA and
BS-I from a protein mixture. Lane 1: protein markers; Lane 2: protein mixture
marker; Lane 3: ConA and BS-I separated from the protein mixture.
To further confirm the results observed in the competitive col-
orimetric assay studies, we next examined the interactions
using SDS-PAGE. Only the Man–GNP/ConA complex was used
here. ConA (15 mg) was first mixed with a comparable amount
of tested proteins for 5 minutes and subsequently incubated
with Man–GNPs at room temperature for 30 minutes. After
centrifugation, protein-bound nanoparticles were isolated and
analyzed by SDS-PAGE on NuPAGE 4–12% Bis-Tris Gel. The pro-
tein bands were visualized by staining with Coomassie blue
and is shown in Figure 5. The gold nanoparticles would not
move and remained at the starting line. More than one band
appeared in Lanes 3, 4, 5, 6, and 10, whereas only one band,
corresponding to ConA, was observed in Lanes 2, 7, 8, and 9.
The results are consistent with the observations in the HTCC
assay. The speculation regarding moderate interactions be-
tween ConA and RCA120 in the HTCC assay is justified by the
and the UV/Vis spectrum of the mixture was identical to that
of Man–GNPs (Figure 6a), thus indicating that the Man–GNPs
did not cross-link with each other even in the presence of its
binding protein, ConA. After centrifugation, the precipitated
debris was analyzed by SDS-PAGE (Figure 6b). Only two dark
bands corresponding to ConA and BS-I were clearly visible in
the gel after it was stained with Coomassie blue. The findings
suggest that either the ConA/BS-I complex is relatively tight
and dense enough to sink at the centrifugation speed used
here, or the binary complex is associated with GNPs through
Man–ConA interactions. The results confirmed that the meth-
odology can effectively detect and separate an interacting pair
of proteins from the rest of the mixture.
observation of two bands, corresponding to ConA and RCA120
respectively, in Lane 10.
,
The FRET efficiency is extremely sensitive to the separation
distance between the energy donor and the acceptor in the
range of 10–100 ꢁ.[57] Thus, it is regarded as a “spectroscopic
ruler” as well as a powerful tool for studying the structural dy-
namic properties of biomolecular interactions and ligand–re-
ceptor interactions.[58–60] To address the dynamic aspects of
To further demonstrate that the competitive colorimetric
methodology can be applied to detecting a particular protein–
protein interacting pair in a protein mixture, a mixture contain-
ing BS-I, trypsin inhibitor, RNase A, bovine serum albumin
(BSA), and ConA was added to Man–GNPs (0.29 nm) and incu-
bated at room temperature for 30 minutes. The solution re-
mained burgundy colored throughout the incubation period,
lectin–proteins interactions,
a
FRET spectrofluorometric
method was employed to exploit one of the interacting pairs
(ConA and BS-I) observed in both the SDS-PAGE studies and
HTCC assay. A fluorescein (Fl)/tetramethylrhodamine (TMRRD)
fluorophore pair with a Fçrster distance (Ro) of 55 ꢁ was
chosen to signal whether interactions occur between ConA
and BS-I in solution. With Fl-labeled ConA as an energy donor
and TMRRD-labeled BS-I as an acceptor, an increase in TMRRD
fluorescence intensity at the expense of Fl intensity was antici-
pated if ConA and BS-I interaction occurred.
A series of fluorescence emission spectra of a fixed concen-
tration of Fl-labeled ConA were acquired by excitation at
490 nm of fluorescein in the presence of various concentra-
tions of TMRRD-labeled BS-I. As shown in Figure 7, as the con-
centration of TMRRD-labeled BS-I increased, there is a signifi-
cant decrease in the fluorescence emission of Fl-labeled ConA
at 520 nm and an increase in the fluorescence emission at
575 nm, therefore strongly indicating that interactions be-
tween ConA and BS-I drove two fluorophores close together,
thus resulting in the observed energy transfer. The rather small
fluorescence increase of TMRRD-labeled BS-I observed here
could be attributed to the interplay of two factors; for exam-
ple, the relative orientation of the donor/acceptor fluorophores
may render the electric dipole transitions ineffective,[61] and
TMRRD’s tendency to aggregate in aqueous solution may
Figure 5. SDS-PAGE on NuPAGE 4–12% Bis-Tris Gel was used to verify the ex-
istence of interactions between ConA and various putative proteins. Other
than Lane 1, each lane contains Man–GNPs, but they would not move in the
gel. Thus, the lane label indicates only the protein names and omits the
GNPs for clarity. Lane 1: protein markers; Lane 2: ConA alone; Lane 3: ConA
+ovalbumin; Lane 4: ConA+BS-I; Lane 5: ConA+SBA; Lane 6: ConA+
MAL; Lane 7: ConA+ECL; Lane 8: ConA+WGA; Lane 9: ConA+BS-II;
Lane 10: ConA+RCA120. ConA monomer=25 kDa; ovalbumin=45 kDa;
BS-I=28.5 kDa, SBA monomer=27.5 kDa; MAL monomer=32.5 kDa
(a,b form); RCA120 monomer=30 kDa. Inset: Corresponding solution photo
images of each lane containing GNPs. Blue indicates that only one protein is
present. Burgundy indicates that two proteins are present in the solution
and they interact.
318
www.chempluschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2012, 77, 314 – 322

Protein Interactions
magnitude is the consequence of synergetic dipoint interac-
tion versus one-point binding to ConA. The sensitivity of this
methodology rivals or exceeds that of SPR and quartz crystal
microbalance techniques, which require the immobilization of
interacting proteins, thus limiting the possible interacting sur-
faces. In addition, its operational simplicity, generality, and use
of near-native conditions make it an attractive method for
quantitative characterization of the binding functions of lectins
and other proteins.
Conclusion
In summary, four sugar–GNPs (Man–GNPs, Glc–GNPs, GlcNAc–
GNPs, and GalNAc–GNPs) were successfully fabricated and ex-
hibited reasonable specificity toward certain lectins regardless
of the electronic nature of the molecular spacer used for link-
ing the substrate and the gold nanoparticles. The color
changes upon recognition enable the use of these sugar–GNPs
as efficient, selective colorimetric sensors for their binding lec-
tins. Moreover, the use of sugar–GNP/binding lectin ensembles
for high-throughput identification of lectin–protein interactions
was demonstrated with no need to label the proteins. Several
new binding pairs were identified for the first time by using
this colorimetric assay, and the binding constants and stoichio-
metric ratios were quantified on the basis of the wavelength
shifts. Its operational simplicity and use of near-native condi-
tions make it an attractive technique, and its sensitivity meets
or exceeds that of surface plasmon resonance and quartz crys-
tal microbalance techniques. The methodology developed
here can be easily adapted to various applications, ranging
from screening larger pools of lectins and a broad range of
proteins in proteomic analysis by simply substituting the sugar
with specific alternatives, to developing molecular libraries
containing potential drugs for disrupting lectin-mediated cell–
cell adhesion.
Figure 7. Fluorescence profile of fluorescein-labeled ConA (0.98 nm) titrated
with various concentrations of rhodamine (TMRRD)-labeled BS-I (0, 4.4, 8.8,
17.5, 35.1, 52.6, 70.2, 87.7, 131.0, 175.4 nm) in PBS buffer (pH 7.4, 0.1 mm
CaCl2, 0.1 mm MnCl2) at 258C for 5 min. lex =490 nm.
reduce the fluorescence output. Nonetheless, this FRET spec-
trofluorometric method is effective for evaluating the dynamic
aspects of lectin–protein interactions in solution and should be
applicable to other identified binding partners for further ap-
plications.
This competitive colorimetric methodology also enables
quantitative estimation of protein–protein interaction strength.
Figure 8 shows the absorption wavelength changes (Dl) rela-
tive to 528 nm as a function of various concentrations of BS-I
and ovalbumin in the presence of a constant concentration of
Man–GNP and ConA (98 nm). A color progression from blue to
purple to burgundy with increasing BS-I or ovalbumin concen-
tration was observed because disruption of the cross-linked
GNPs can be induced by the formation of a ConA/BS-I or
ConA/ovalbumin binding pair. The extent of particle dissipa-
tion is proportional to the BS-I or ovalbumin concentration.
Saturation is reached once the concentration of BS-I or ovalbu-
min is about 49 nm or 100 nm, respectively, thus indicating
that the stoichiometry for the binding of BS-I to ConA is 1:2,
and that of ovalbumin to ConA is 1:1. The binding constant of
ConA/BS-I is calculated to be 1.5x1015 by a nonlinear regression
curve fitting program,[62] and that of ConA/ovalbumin is calcu-
lated to be 2.6x108 based on 1:1 binding,[63] thus suggesting
that the binding affinity of BS-I toward ConA is much stronger
than that of ovalbumin. The large difference in the binding
Experimental Section
General methods
All reagents and starting materials were obtained from commercial
suppliers (Acros, Aldrich, Sigma, and Merck) and were used without
further purification. The IR spectra were recorded with a Nicolet
550 series II spectrometer. The 1H and 13C NMR spectra were re-
corded with a Varian Mercury Plus 400 or Bruker Avance 400 spec-
trometer. The proton and carbon chemical shifts are given in ppm
using CDCl3 as the internal standard (1H: d=7.24 ppm and 13C: d=
77.0 ppm). High-resolution mass spectra were recorded with
a JEOL-102A mass spectrometer. Analytical TLC (silica gel, 60F-54,
Merck) and spots were visualized under UV light and/or staining
with phosphomolybdic acid in ethanol. Flash column chromatogra-
phy was performed with Kiesegel 60 (230–400 mesh) silica gel
(Merck). Melting points are reported without correction. Fl-labeled
ConA and TMRRD-labeled BS-I were purchased from Sigma–Al-
drich.
2-(2-(2-Azidoethoxy)ethoxy)ethyl 2-acetamido-3,4,6-tri-O-acetyl-
2-deoxy-b-d-glucopyranoside (1): AgOTf (3.40 g, 13.3 mmol) and
tetramethylurea (0.98 mL, 8.3 mmol) were added to a solution of 2-
acetamido-3,4,6-tri-O-acetyl-2-deoxy-b-d-glucoyl chloride (1.46 g,
Figure 8. Absorption wavelength changes (Dl) relative to 528 nm plotted as
a function of BS-I concentrations (0, 9, 22, 31, 44, 66, 88, 132 nm) and oval-
bumin concentrations (0, 33, 56, 78, 111, 167, 222, 278 nm).
ChemPlusChem 2012, 77, 314 – 322
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chempluschem.org
319

C.-T. Chen and C.-S. Tsai
10.0 mmol) in CH2Cl2 (20 mL). Subsequently, 8-azido-3,6-dioxa-1-oc-
tylol (5.87 g, 8.3 mmol) in CH2Cl2 (10 mL) was added dropwise to
the reaction mixture at RT and was stirred for 20 h. The reaction so-
lution was filtered and concentrated under reduced pressure. The
residue was redissolved in CH2Cl2 (100 mL) and washed with water
(ꢂ2) and brine. The combined organic layers were dried over anhy-
drous MgSO4, filtered, and concentrated under reduced pressure.
Compound 1 (3.81 g, 90%) was obtained as an oil after flash
column chromatography on silica gel using CH2Cl2/MeOH 95:5
(m, 12H); 13C NMR (100 MHz, CDCl3, 258C) d=195.7, 173.3, 170.9,
169.0, 101.5, 73.3, 71.7, 71.6, 71.3, 70.3, 70.1, 69.8, 68.6, 68.4, 62.1,
54.0, 38.9, 36.6, 30.7, 29.5, 29.4, 29.4, 29.3, 29.1, 29.1, 28.8, 25.8,
23.1, 20.9, 20.8, 20.7; IR (neat): 3370, 2925, 2866, 1752, 1653, 1546,
1381, 1241, 1129, 1049. cmÀ1; HRMS (FAB) calcd for C33H57N2O13S
[M+1]+: 721.3581; found: 721.3585.
2-(2-(2-(11-Mercaptoundecanamido)ethoxy)ethoxy)ethoxy
2-acetamido-2-deoxy-b-d-glucopyranoside (GlcNAc-SH): NaOMe
(225.0 mg, 4.16 mmol) was added to a solution of 3 (500.0 mg,
0.69 mmol) in CH3OH (15 mL) and stirred at RT for 3 h. The mixture
was then neutralized using Dowex 50ꢂ8–400 resin and filtered
through a pad of Celite. The collected filtrate was concentrated
under reduced pressure and GlcNAc-SH (250.0 mg, 80%) was ob-
tained as an oil after column chromatography on silica gel using
MeOH/CH2Cl2 1:9 eluent. Rf =0.33 (CH2Cl2/MeOH 8:2); 1H NMR
(400 MHz, D2O, 258C) d=4.56 (d, J=8.4 Hz, 1H), 4.00 (m, 1H), 3.91
(d, J=11.7 Hz, 1H), 3.82–3.52 (m, 12H), 3.48–3.32 (m, 4H), 2.71 (bs,
2H), 2.24 (bs, 2H), 2.05 (s, 2H), 1.76–1.52 (m, 4H), 1.48–1.20 (m,
12H); 13C NMR (100 MHz, D2O, 258C) d=175.4, 174.0, 101.3, 76.3,
74.3, 70.3, 70.2, 70.0, 69.4, 69.2, 61.2, 55.9, 39.3, 39.3, 36.4, 30.5,
30.2, 30.1, 30.0, 29.3, 26.5, 23.0; IR (neat): 3310, 2919, 2859,
1659 cmÀ1; HRMS (FAB) calcd for C25H49N2O9S [M+1]+: 553.3159;
found: 553.3157.
1
eluent. Rf =0.33 (CH2Cl2/MeOH 9:1); H NMR (400 MHz, CDCl3, 258C)
d=6.36 (d, J=9.2 Hz, 1H), 5.12–5.00 (m, 2H), 4.72 (d, J=8.4 Hz,
1H), 4.20 (dd, J=4.8, 12.5 Hz, 1H), 4.09–4.02 (m, 2H), 3.84–3.57 (m,
11H), 3.47 (t, J=4.8 Hz, 2H), 2.00 (s, 3H), 1.96 (s, 3H), 1.96 (s, 3H),
1.92 (s, 3H); 13C NMR (100 MHz, CDCl3, 258C) d=170.6, 170.5,
170.2, 169.1, 101.8, 73.2, 71.6, 71.6, 70.5, 70.3, 69.7, 68.6, 68.5, 62.1,
53.9, 50.4, 23.1, 20.8, 20.8, 20.7; IR (neat): 2110, 1752, 1653, 1546,
1374, 1248, 1116, 1049 cmÀ1; HRMS (FAB) calcd for C20H33N4O11 [M+
1]+: 505.2146; found: 505.2147.
2-(2-(2-(Undec-10-enamido)ethoxy)ethoxy)ethoxy 2-acetamido-
3,4,6-tri-O-acetyl-2-deoxy-b-d-glucopyranoside (2): PPh3 (1.22 g,
3.9 mmol) and H2O (0.10 mL, 5.3 mmol) were added to a solution
of 1 (1.79 g, 3.6 mmol) in THF (30 mL). The mixture was stirred at
RT for 9 h. Then THF was then removed under reduced pressure
first followed by the addition of CH2Cl2 (30 mL), undec-10-enoic
acid (0.65 g, 5.3 mmol), HOBt (0.53 g, 2.5 mmol), DMAP (catalytic
amount), triethylamine (1.0 mL), and DCC (1.10 g, 5.32 mmol). The
reaction mixture was stirred at RT for 10 h before the reaction mix-
ture was filtered through celite and the filtrate was collected and
concentrated under reduced pressure. The residue was taken up in
EtOAc (100 mL) and washed with 5% citric acid (ꢂ3), 10% NaHCO3
(ꢂ3), and brine. The combined organic layers were dried over an-
hydrous MgSO4, filtered, and concentrated under reduced pressure.
Compound 2 (1.44 g, 63%) was obtained as an oil after flash
column chromatography on silica gel using CH2Cl2/MeOH 95:5
Allyl 2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-a-d-galactopyrano-
side (4): BF3·Et2O complex (0.25 mL, 2 mmol) was added to a solu-
tion of N-acetyl-d-galactosamine (442 mg, 2 mmol) in allyl alcohol
(8 mL) at 08C. The mixture was stirred for 2 h at 708C. After the so-
lution was cooled to RT, the solvent was evaporated under reduced
pressure and then pyridine (10 mL), acetic anhydride (0.75 mL,
8.0 mmol), and DMAP (catalytic amount) were added. The mixture
was then stirred at RT for 10 h. After that, pyridine was removed
under reduced pressure and the residue was taken up in EtOAc
(60 mL) and washed with 5% citric acid (ꢂ3), 10% NaHCO3 (ꢂ3),
and brine. The combined organic layers were dried over anhydrous
MgSO4, filtered, and concentrated under reduced pressure. Com-
pound 4 (380.0 mg, 50%) was obtained as an oil after flash column
chromatography on silica gel using CH2Cl2/EtOAc 8:2 eluent. Rf =
0.33 (CH2Cl2/EtOAc 7:3); 1H NMR (400 MHz, CDCl3, 258C) d=5.86–
5.75 (m, 2H), 5.29 (dd, J=3.1, 0.8 Hz, 1H), 5.26–5.13 (m, 2H), 5.11
(dd, J=11.4, 3.3 Hz, 1H), 4.86 (d, J=3.6 Hz, 1H), 4.50 (m, 1H), 4.14–
3.88 (m, 5H), 2.08 (s, 3H), 1.97 (s, 3H), 1.91 (s, 3H), 1.89 (s, 3H);
13C NMR (100 MHz, CDCl3, 258C) d=170.8, 170.3, 170.2, 170.0,
133.0, 118.2, 96.6, 68.6, 68.2, 67.2, 66.6, 61.8, 47.6, 23.1, 20.6, 20.5,
1
eluent. Rf =0.30 (CH2Cl2/MeOH 9:1); H NMR (400 MHz, CDCl3, 258C)
d=6.66 (d, J=9.2 Hz, 1H), 6.37 (t, J=5.5 Hz, 1H), 5.67 (m, 1H),
5.08–4.79 (m, 4H), 4.67 (d, J=8.4 Hz, 1H), 4.16 (dd, J=4.7, 12.1 Hz,
1H), 3.94 (q, J=9.2 Hz, 1H), 3.80 (m, 1H), 3.72 (m, 1H), 3.63–3.43
(m, 9H), 3.38 (q, J=5.5 Hz, 2H), 2.09 (t, J=7.6 Hz, 2H), 1.99 (s, 3H),
1.98–1.94 (m, 8H), 1.90 (s, 3H), 1.60–1.49 (m, 2H), 1.30–1.19 (m,
10H); 13C NMR (100 MHz, CDCl3, 258C) d=173.2, 170.6, 170.1,
168.9, 138.7, 113.8, 110.4, 73.2, 71.5, 71.1, 70.2, 70.0, 69.7, 68.5, 68.4,
62.0, 53.9, 38.8, 36.5, 33.6, 29.2, 29.0, 28.8, 25.7, 23.0, 20.7, 20.6; IR
(neat): 3303, 2932, 2859, 1752, 1646, 1560, 1374, 1235, 1135,
1043 cmÀ1; HRMS (FAB) calcd for C31H52N2O12 [M+1]+: 645.3599;
found: 645.3597.
20.5; IR (neat): 3383, 1752, 1659, 1553, 1374, 1235, 1049 cmÀ1
;
HRMS (FAB) calcd for C17H26NO9 [M+1]+: 388.1608; found:
388.1610.
(10,22-Dioxo-3,6-dioxa-21-thia-9-azatricosyl)oxy
2-acetamido-
3,4,6-tri-O-acetyl-2-deoxy-b-d-glucopyranoside (3): Thiolacetic
acid (0.59 mL, 7.76 mmol) and azobis(isobutylnitrile) (51.0 mg,
0.31 mmol) were added to a solution of 2 (1.00 g, 1.55 mmol) in
CH3OH (30 mL). The reaction mixture was irradiated in a photo-
chemical reactor (Rayonet reactor lamp, Southern New England Ul-
traviolet Co., model RPR-100, 365 nm) for 6 h under nitrogen. The
mixture was concentrated under reduced pressure and compound
3 (1.00 g, 90%) was obtained as a yellowish oil after a flash column
chromatography on silica gel using CH2Cl2/MeOH 95:5 eluent. Rf =
0.34 (CH2Cl2/MeOH 9:1). 1H NMR (400 MHz, CDCl3, 258C) d=6.57
(bs, 1H), 6.35 (bs, 1H), 5.08–5.01 (m, 2H), 4.73 (d, J=8.4 Hz, 1H),
4.20 (dd, J=3.3, 12.5 Hz, 1H), 4.07 (d, J=12.5 Hz, 1H), 3.98 (q, J=
8.4 Hz, 1H), 3.83 (m, 1H), 3.78 (m, 1H), 3.70–3.50 ( m, 9H), 3.40 (m,
2H), 2.79 (t, J=6.9 Hz, 2H), 2.30–2.22 (m, 3H), 2.13 (t, J=6.6 Hz,
2H), 2.06–1.94 (m, 11H), 1.90 (s, 3H), 1.62–1.42 (m, 4H), 1.34–1.16
3-((2-Aminoethyl)thio)propoxy 2-acetamido-3,4,6-tri-O-acetyl-2-
deoxy-a-d-galactopyranoside (5): The procedure is similar to the
one used to prepare 3. Compound 5 (401.0 mg, 67%) was ob-
tained as an oil after flash column chromatography on silica gel
1
using CH2Cl2/EtOAc 8:2 eluent. Rf =0.35 (CH2Cl2/EtOAc 7:3); H NMR
(400 MHz, CDCl3, 258C) d=6.18 (d, J=9.3 Hz, 1H), 5.25 (d, J=
3.1 Hz, 1H), 5.04 (dd, J=11.4, 3.6 Hz, 1H), 4.77 (d, J=3.6 Hz, 1H),
4.44 (m, 1H), 4.10–3.92 (m, 1H), 3.70 (m, 1H), 3.52 (m, 1H), 2.78 (t,
J=6.8 Hz, 2H), 2.56–2.46 (m, 4H), 2.01 (s, 3H), 1.93 (s, 3H), 1.87 (s,
3H), 1.83 (s, 3H), 1.82–1.74 (m, 2H), 1.67 (s, 2H); 13C NMR
(100 MHz, CDCl3, 258C) d=169.9, 169.4, 169.4, 169.2, 97.3, 68.1,
67.1, 66.6, 66.5, 61.8, 47.6, 41.0, 36.1, 29.2, 28.7, 23.3, 20.9, 20.9,
20.9; IR (neat): 3416, 1745, 1646, 1533, 1149 cmÀ1; HRMS (FAB)
calcd for C19H32N2O9S [M+1]+: 465.1907; found: 465.1905.
320
www.chempluschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2012, 77, 314 – 322

Protein Interactions
3-((2-(Undec-10-enamido)ethyl)thio)propoxy 2-acetamido-3,4,6-
tri-O-acetyl-2-deoxy-a-d-galactopyranoside (6): Undec-10-enoic
acid (130.0 mg, 0.71 mmol), HOBt (60.0 mg, 0.44 mmol), DMAP (cat-
alytic amount), TEA (1.0 mL) and DCC (183.0 mg, 0.89 mmol) were
added to a solution of 5 (205.7 mg, 0.44 mmol) in CH2Cl2 (10 mL).
The reaction mixture was stirred at RT for 10 h and then filtered
through a pad of Celite. The collected filtrate was concentrated
under reduced pressure and the residue was dissolved in EtOAc
(100 mL) and subsequently washing with 5% citric acid (ꢂ3), 10%
NaHCO3 (ꢂ3), and brine. The combined organic layers were dried
over anhydrous MgSO4, filtered, and concentrated under reduced
pressure. Compound 6 (245.0 mg, 88%) was obtained as an oil
after flash column chromatography on silica gel using CH2Cl2/
on silica gel using n-hexane/EtOAc 6:4 eluent. Rf =0.43 (n-hexane/
EtOAc 1:1); H NMR (400 MHz, CDCl3, 258C) d=5.88–5.78 (m, 1H),
1
5.45 (t, J=10.2 Hz, 1H), 5.26 (dd, J=17.2, 1.5 Hz, 1H), 5.17 (dd, J=
10.2, 1.5 Hz, 1H), 5.06–4.99 (m, 2H), 4.83 (dd, J=10.2, 3.7 Hz, 1H),
4.20 (dd, J=12.2, 4.4 Hz, 1H), 4.18–4.11 (m, 1H), 4.05–3.96 (m, 3H),
2.05 (s, 3H), 2.03 (s, 3H), 1.97 (s, 3H), 1.95 (s, 3H); 13C NMR
(100 MHz, CDCl3, 258C) d=170.5, 170.0, 170.0, 169.5, 133.0, 118.1,
94.8, 70.7, 70.1, 68.7, 68.5, 67.3, 61.8, 20.6, 20.6, 20.5, 20.5; MS
(FAB): m/z 389.1 [M+1]+.
3-((2-Aminoethyl)thio)propoxy 2,3,4,6-tetra-O-acetyl-2-a-d-glu-
copyranoside (9): 2-Aminoethanethiol hydrochloride (1.46 g,
12.9 mmol) was added to a solution of 8 (1.0 g, 2.6 mmol) in
CH3OH (20 mL). The reaction mixture was irradiated in a photo-
chemical reactor for 5 h under nitrogen and then concentrated
under reduced pressure. Compound 9 (1.15 g, 96%) was obtained
as an oil after flash column chromatography on silica gel using
CH2Cl2/MeOH 9:1 eluent. Rf =0.4 (CH2Cl2/MeOH 17:3); 1H NMR
(400 MHz, CDCl3, 258C) d=5.56 (bs, 2H), 5.39 (t, J=10.1 Hz, 1H),
5.03–4.97 (m, 2H), 4.80 (dd, J=10.1, 3.7 Hz, 1H), 4.20 (dd, J=12.4,
4.4 Hz, 1H), 4.06 (dd, J=12.4, 2.2 Hz, 1H), 3.98–3.94 (m, 1H), 3.77–
3.73 (m, 1H), 3.50–3.47 (m, 1H), 3.08 (t, J=6.6 Hz, 2H), 2.79 (t, J=
6.6 Hz, 2H), 2.61 (t, J=7.0 Hz, 2H), 2.06 (s, 3H), 2.02 (s, 3H), 1.98 (s,
3H), 1.95 (s, 3H), 1.88–1.84 (m, 2H); 13C NMR (100 MHz, CDCl3,
258C) d=170.7, 170.2, 170.2, 169.6, 95.8, 70.8, 70.1, 68.5, 67.3, 66.6,
61.9, 39.6, 30.8, 28.9, 28.1, 20.8, 20.7, 20.7, 20.6; IR (KBr): 3395,
3290, 1752, 1230 cmÀ1; HRMS (FAB) calcd for C19H32NO10S [M+1]+:
466.1747; found: 466.1751.
1
MeOH 98:2 eluent. Rf =0.30 (CH2Cl2/MeOH 95:5); H NMR (400 MHz,
CDCl3, 258C) d=6.45 (d, J=9.2 Hz, 1H), 6.00 (bs, 1H), 5.71 (m, 1H),
5.31 (d, J=2.7 Hz, 1H), 5.12 (dd, J=11.4, 3.2 Hz, 1H), 4.95–4.82 (m,
3H), 4.51 (m, 1H), 4.13–3.97 (m, 3H), 2.65–2.53 (m, 4H), 2.15–2.03
(m, 5H), 2.02–1.75 (m, 13H), 1.60–1.47 (m, 2H), 1.33–1.13 (m, 10H);
13C NMR (100 MHz, CDCl3, 258C) d=173.3, 170.6, 170.4, 170.3,
170.3, 139.0, 114.0, 97.5, 68.1, 67.2, 66.5, 66.3, 61.8, 47.6, 38.5, 36.5,
33.6, 31.3, 29.2, 29.2, 29.2, 29.1, 28.9, 28.7, 28.0, 25.5, 23.0, 20.6,
20.6; IR (neat): 3323, 2919, 2853, 1759, 1646, 1560, 1367, 1241,
1135, 1056 cmÀ1; HRMS (FAB) calcd for C30H51N2O10S [M+1]+:
631.3264; found: 631.3267.
3-((2-(11-(Acetylthio)undecanamido)ethyl)thio)propoxy 2-acet-
amido-3,4,6-tri-O-acetyl-2-deoxy-a-d-galactopyranoside (7): Thi-
olacetic acid (0.40 mL, 5.10 mmol) and azobis(isobutylnitrile;
46 mg, 0.28 mmol) were added to a solution of 6 (550.0 mg,
0.87 mmol) in CH3OH (20 mL). The reaction mixture was irradiated
in a photochemical reactor for 6 h under nitrogen and then con-
centrated under reduced pressure. Compound 7 (499.0 mg, 81%)
was obtained as a yellowish oil after flash column chromatography
on silica gel using CH2Cl2/MeOH 98:2 eluent. Rf =0.33 (CH2Cl2/
3-((2-(11-Mercaptoundecanamido)ethyl)thio)propoxy
2,3,4,6-
tetra-O-acetyl-2-a-d-glucopyranoside (10): 11-Mercaptoundecano-
ic acid N-hydroxysuccinimide ester (271.0 mg, 0.86 mmol) was
added to a solution of 9 (200.0 mg, 0.43 mmol) in CH2Cl2 (10 mL)
and stirred at RT for 10 h. The mixture was then concentrated
under reduced pressure. Compound 10 (190.0 mg, 68%) was ob-
tained as an oil after flash column chromatography on silica gel
1
MeOH 95:5); H NMR (400 MHz, CDCl3, 258C) d=6.62 (d, J=9.2 Hz,
1H), 6.23 (t, J=5.5 Hz, 1H), 5.28 (d, J=1.9 Hz, 1H), 5.05 (dd, J=
11.4, 2.9 Hz, 1H), 4.81 (d, J=3.7 Hz, 1H), 2.80–2.70 (m, 2H), 2.62–
2.48 (m, 4H), 2.19 (s, 3H), 2.13–2.01 (m, 5H), 1.94 (s, 3H), 1.89 (s,
3H), 1.83 (s, 3H), 1.60–1.39 (m, 4H), 1.30–1.08 (m, 12H); 13C NMR
(100 MHz, CDCl3, 258C) d=195.0, 172.5, 169.7, 169.6, 169.5, 97.2,
68.1, 67.2, 66.4, 66.3, 61.9, 47.7, 38.7, 36.6, 31.5, 30.8, 29.6, 29.5,
29.4, 29.4, 29.2, 29.1, 28.9, 28.2, 25.8, 23.2, 20.9, 20.9; IR (neat):
1
using CH2Cl2/EtOAc 5:1 eluent. Rf =0.25 (CH2Cl2/EtOAc 4:1); H NMR
(400 MHz, CDCl3, 258C) d=5.98 (bs, 1H), 5.41 (t, J=9.9 Hz, 1H),
5.04–4.96 (m, 2H), 4.82 (dd, J=3.6, 10.4 Hz, 1H), 4.21 (dd, J=4.7,
12.5 Hz, 1H), 4.06 (dd, J=2.6, 12.6 Hz, 1H), 3.97 (m, 1H), 3.76 (m,
1H), 3.52–3.35 (m, 3H), 2.65–2.54 (m, 4H), 2.51–2.43 (m, 3H), 2.28
(t, J=7.8 Hz, 1H), 2.15 (t, J=7.3 Hz, 2H), 2.05 (s, 3H), 2.04 (s, 3H),
1.99 (s, 3H), 1.97 (s, 3H), 1.90–1.81 (m, 2H), 1.62–11.52 (m, 4H),
1.36–1.20 (m, 12H); 13C NMR (100 MHz, CDCl3, 258C) d=173.4,
170.6, 170.1, 170.1, 169.5, 95.7, 70.7, 70.0, 68.4, 67.2, 66.4, 61.8,
38.3, 36.6, 33.9, 31.5, 29.3, 29.3, 29.2, 29.2, 29.2, 29.2, 29.1, 28.9,
28.9, 28.2, 27.9, 25.6, 24.5, 20.7, 20.6, 20.5; IR (neat): 1752, 1659,
1540, 1447, 1381, 1228, 1049 cmÀ1; MS (FAB): m/z 666.4 [M+1]+.
3323, 2919, 2853, 1759, 1646, 1560, 1367, 1241, 1135, 1056 cmÀ1
;
HRMS (FAB) calcd for C32H55N2O11S2 [M+1]+: 707.3147; found:
707.3142.
2-(3-((2-(11-Mercaptoundecanamido)ethyl)thio)propoxy 2-acet-
amido-2-deoxy-a-d-galactopyranoside (GalNAc-SH): The proce-
dure is similar to the one used to prepare GlcNAc-SH. GalNAc-SH
(56.0 mg, 65%) was obtained as an oil after flash column chroma-
tography on silica gel using CH2Cl2/MeOH 95:5 eluent. Rf =0.33
(CH2Cl2/MeOH 9:1); 1H NMR (400 MHz, CD3OD, 258C) d=4.72 (d,
J=3.7 Hz, 1H), 4.14 (dd, J=11.0, 3.7 Hz, 1H), 3.77 (d, J=3.1 Hz,
1H), 3.75–3.55 (m, 4H), 3.37 (m, 1H), 3.27–3.16 (m, 3H), 2.58–2.48
(m, 4H), 2.38 (t, J=7.2 Hz, 1H), 2.07 (t, J=7.4 Hz, 2H), 1.88 (s, 3H),
1.80–1.70 (m, 2H),1.54–1.44 (m, 4H), 1.32–1.15 (m, 12H); 13C NMR
(400 MHz, CD3OD, 258C) d=176.6, 174.1, 90.1, 72.8, 70.6, 69.9,
67.7, 63.1, 51.9, 40.3, 37.4, 35.5, 32.3, 30.9, 30.8, 30.8, 30.7, 30.6,
30.5, 29.7, 29.5, 27.3, 25.3, 23.0. IR (neat): 3310, 2932, 2846, 1633,
1547, 1129, 1056 cmÀ1; HRMS (FAB) calcd for C24H47N2O7S2 [M+1]+:
539.2825; found: 539.2829.
2-(3-((2-(11-Mercaptoundecanamido)ethyl)thio)propoxy-a-d-glu-
copyranoside (Glc-SH): The procedure for saponification is similar
to the one used to prepare GlcNAc-SH. The product Glc-SH
(58.0 mg, 65%) was obtained as an oil after flash column chroma-
tography on silica gel using CHCl3/MeOH 95: 5 eluent. Rf =0.30
1
(CHCl3/MeOH 9:1); H NMR (400 MHz, CD3OD, 258C) d=4.80 (d, J=
3.6 Hz, 1H), 3.89–3.79 (m, 2H), 3.63–3.51 (m, 4H), 3.44–3.27 (m,
4H), 2.74–2.62 (m, 4H), 2.51 (t, J=6.8 Hz, 2H), 2.21 (t, J=7.3 Hz,
2H), 2.00–1.84 (m, 2H), 1.68–1.56 (m, 4H), 1.45–1.28 (m, 13H);
13C NMR (400 MHz, CD3OD, 258C) d=175.5, 99.9, 75.0, 73.5, 71.7,
67.4, 62.6, 40.3, 37.3, 35.4, 32.3, 30.9, 29.6, 27.3, 25.3; IR (neat):
3409, 2110, 1633, 1016 cmÀ1; MS (FAB): m/z 498.3 [M+1]+:
Allyl 2,3,4,6-tetra-O-acetyl-2-a-d-glucopyranoside (8): The proce-
dure is similar to the one used to prepare 4. Compound 8 (6.24 g,
58%) was obtained as an oil after flash column chromatography
ChemPlusChem 2012, 77, 314 – 322
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chempluschem.org
321

C.-T. Chen and C.-S. Tsai
[24] M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998, 110,
2908–2953; Angew. Chem. Int. Ed. 1998, 37, 2754–2794.
[25] M. B. Gerke, M. B. Plenderleith, Anatomical Record 2002, 268, 105–114.
[26] R. Elghanian, J. J. Storhoff, R. C. Mucic, R. L. Letsinger, C. A. Mirkin, Sci-
ence 1997, 277, 1078–1081.
Lectin specificity of sugar–GNPs
Individually tested protein (1 mgmLÀ1; 15 mL) was added to sugar–
GNPs (1.5 mL), and then the absorption spectrum was recorded, re-
spectively, after 30 min of incubation time.
[27] J. Liu, Y. Lu, Angew. Chem. 2006, 118, 96–100; Angew. Chem. Int. Ed.
2006, 45, 90–94.
[28] H. Otsuka, Y. Akiyama, Y. Nagasaki, K. Kataoka, J. Am. Chem. Soc. 2001,
123, 8226–8230.
Rapid throughput competitive colorimetric assay
ConA (10 mL), WGA (10 mL), or RCA120 (10 mL) were first mixed with
the putative proteins (10 mL) for 5 min at 48C followed by the addi-
tion of sugar–GNPs (200 mL). After 30 min incubation, the photo-
graph was recorded.
[29] D. C. Hone, A. H. Haines, D. Russell, Langmuir 2003, 19, 7141–7144.
[30] J. M. de La Fuente, A. G. Barrientos, T. C. Rojas, J. Rojo, J. CaÇada, A.
Fernꢃndez, S. Penadꢄs, Angew. Chem. 2001, 113, 2317–2321; Angew.
Chem. Int. Ed. 2001, 40, 2257–2261.
[31] C.-C. Lin, Y.-C. Yeh, C.-Y. Yang, C.-L. Chen, G.-F. Chen, C.-C. Chen, Y.-C.
Wu, J. Am. Chem. Soc. 2002, 124, 3508–3509.
[32] C.-S. Tsai, T.-B. Yu, C.-T. Chen, Chem. Commun. 2005, 4273–4275.
[33] C.-C. Huang, Y.-L. Hung, Y.-C. Shiang, T.-Y. Lin, Y.-S. Lin, C.-T. Chen, H.-T.
Chang, Chem. Asian J. 2010, 5, 334–341.
[34] J. A. Harmony, E. H. Cordes, J. Biol. Chem. 1975, 250, 8614–8617.
[35] J. T. Dulaney, Anal. Biochem. 1979, 99, 254–267.
[36] H. Rꢅdiger, H.-J. Gabius, Glycoconjugate J. 2001, 18, 589–613.
[37] M. Ambrosi, N. R. Cameron, B. J. Davis, Org. Biomol. Chem. 2005, 3,
1593–1608.
[38] A. Cambi, C. G. Figdor, Curr. Opin. Immunol. 2005, 17, 345–351.
[39] J. R. Wright, Nat. Rev. Immunol. 2005, 5, 58–68.
Detection of protein–protein interaction with SDS-PAGE
The assay solution was transferred to vials and centrifuged. The
sunk protein–bound nanoparticles were isolated and washed with
deionized water several times to remove the residual protein solu-
tion. Some of the particles were loaded to 12% SDS-PAGE and sep-
arated by gel electrophoresis and the gel was subsequently stained
with Coomassie Blue.
[40] S. Sueyoshi, T. Tsuji, T. Osawa, Biol. Chem. Hoppe-Seyler. 1985, 366, 213–
221.
Acknowledgements
[41] L.-G. Yu, D. J. Fernig, J. A. Smith, J. D. Milton, J. M. Rhodes, Can.
Res.1993, 53, 4627–4632.
[42] A. Varki, R. Cummings, J. Esko, H. Freeze, G. Hart, J. Marth in Essentials
of Glycobiology, Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, 1999.
We thank the National Science Council of the Republic of China,
Taiwan for financial support.
[43] C. Torres, G. W. Hart, J. Biol. Chem. 1984, 259, 3308–3317.
[44] G. D. Holt, G. W. Hart, J. Biol. Chem. 1986, 261, 8049–8057.
[45] F. I. Comer, G. W. Hart, J. Biol. Chem. 2000, 275, 29179–29182.
[46] L. Wells, K. Vosseller, G. W. Hart, Science 2001, 291, 2376–2378.
[47] R. P. Matthew, C. R. Bertozzi, Chem. Soc. Rev. 2005, 34, 58–68.
[48] T. V. RajanBabu, T. A. Ayers, G. A. Halliday, K. K. You, J. C. Calabrese, J.
Org. Chem. 1997, 62, 6012–6028.
Keywords: colorimetric assay
nanoparticles · immobilization · protein–protein interactions
·
glycosides
·
gold
·
[1] E. M. Phizicky, S. Fields, Microbiol. Rev. 1995, 59, 94–123.
[2] Y. C. Cao, R. Jin, J.-M. C. Nam, S. Thaxton, C. A. Mirkin, J. Am. Chem. Soc.
2003, 125, 14676–14677, and references therein.
[3] J. R. Newman, A. E. Keating, Science 2003, 300, 2097–2101.
[4] T. Ito, T. Chiba, R. Ozawa, M. Yoshida, M. Hattori, Y. Sakaki, Proc. Natl.
Acad. Sci. USA 2001, 98, 4569–4574.
[49] M. N. Yousaf, M. Mrksich, J. Am. Chem. Soc. 1999, 121, 4286–4287.
[50] K. Higashi, S. Miyoshi, S. Nakabayashi, H. Yamada, Y. Ito, Chem. Pharm.
Bull. 1992, 40, 2300–2303.
[5] J. R. Birch, E. S. Lennox in Monoclonal Antibodies: Principles and Applica-
tions, Wiley, New York, 1995.
[51] J. J. La Clair, J. Am. Chem. Soc. 1997, 119, 7676–7684.
[52] G. Frens, Nature 1973, 241, 20–22.
[6] W.-S. Yeo, D.-H. Min, R. W. Hsieh, G. L. Greene, M. Mrksich, Angew. Chem.
2005, 117, 5616–5619; Angew. Chem. Int. Ed. 2005, 44, 5480–5483.
[7] M. Fernꢃndez-Suꢃrez, T. S. Chen, A. Y. Ting, J. Am. Chem. Soc. 2008, 130,
9251–9253.
[8] J. H. Lakey, E. M. Raggett, Curr. Opin. Struct. Biol. 1998, 8, 119–123.
[9] I. Remy, S. W. Michnick, Proc. Natl. Acad. Sci. USA 1999, 96, 5394–5399.
[10] T.-W. Lin, P.-J. Hsieh, C.-L. Lin, Y.-Y. Fang, J.-X. Yang, C.-C. Tsai, P.-L.
Chiang, C.-Y. Pan, Y.-T. Chen, Proc. Natl. Acad. Sci. USA 2010, 107, 1047–
1052.
[53] J. M. Abad, S. F. L. Mertens, M. Pita, V. M. Fernꢃndez, D. J. Schiffrin, J. Am.
Chem. Soc. 2005, 127, 5689–5694.
[54] The eight lectins are: Bandeiraea simplicifolia lectin I (BS-I), soybean ag-
glutinin (SBA), Maackia amurensis leukoagglutinin (MAL), erythrina cris-
tagalli lectin (ECL), wheat germ agglutinin (WGA), bandeiraea simplicifo-
lia lectin II (BS-II), Concanavalin A (ConA), and Ricinus communis Aggluti-
nin (RCA120). Two other proteins are thyroglobulin and ovalbumin.
[55] S. Vincenzi, G. Zoccatelli, F. Perbellini, C. Rizzi, R. Chignola, A. Curioni,
D. B. Angelo, A. D. B. Peruffo, J. Agric. Food Chem. 2002, 50, 6266–6270.
[56] T. T.-S. Lin, S. S.-L. Li, Eur. J. Biochem. 1980, 105, 453–459.
[57] K. E. Sapsford, L. Berti, I. L. Medintz, Angew. Chem. 2006, 118, 4676–
4704; Angew. Chem. Int. Ed. 2006, 45, 4562–4589.
[11] J. Roth, Chem. Rev. 2002, 102, 285–304.
[12] G. E. Ritchie, B. E. Moffatt, R. B. Sim, B. P. Morgan, R. A. Dwek, P. M. Rudd,
Chem. Rev. 2002, 102, 305–320.
[13] S. Park, M.-R. Lee, I. Shin, Chem. Soc. Rev. 2008, 37, 1579–1591.
[14] J. A. Prescher, C. R. Bertozzi, Cell 2006, 126, 851–854.
[15] A. Varki, Nature 2007, 446, 1023–1029.
[58] R. M. Clegg, Methods Enzymol. 1992, 211, 353–388.
[59] M. Lorenz, A. Hillisch, S. Diekmann, Rev. Mol. Biotechnol. 2002, 82, 197–
209.
[16] C. R. Bertozzi, L. L. Kiessling, Science 2001, 291, 2357–2364.
[17] N. E. Zachara, G. W. Hart, Chem. Rev. 2002, 102, 431–438.
[18] C. Armengol, A. Carretero, V. Nacher, J. Ruberte, M. Navarro, J. Anat.
2007, 210, 98–111.
[19] M. B. Plenderleith, L. L. Wright, P. J. Snow, Dev. Brain Res. 1992, 68, 103–
109.
[60] D. M. J. Lilley, T. J. Wilson, Curr. Opin. Chem. Biol. 2000, 4, 507–517.
[61] P. R. Selvin, J. E. Hearst, Proc. Natl. Acad. Sci. USA 1994, 91, 10024–
10028.
[62] K. A. Connors in Binding Constants: The Measurement of Molecular Com-
plex Stability, Wiley-Interscience, New York, 1987.
[63] J. Bourson, J. Pouget, B. Valeur, J. Phys. Chem. 1993, 97, 4552–4557.
[20] G. W. Knott, P. D. Kitchener, N. R. Saunders, J. Comp. Neurol. 1999, 414,
423–436.
[21] R. P. Ecever, Glycoconjugate J. 1997, 14, 585–591.
[22] K. A. Karlsson, Curr. Opin. Struct. Biol. 1995, 5, 622–635.
[23] H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637–674.
Received: January 21, 2012
Published online on February 28, 2012
322
www.chempluschem.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemPlusChem 2012, 77, 314 – 322