A facile method for the preparation of gold glyconanoparticles from free oligosaccharides and their applicability in carbohydrate-protein interaction studies

Article abstract of DOI:10.1002/ejoc.200500256

The weak binding affinity of monomeric oligosaccharides with carbohydrate-binding proteins are hampering their use in in-vivo and in-vitro bio-assays. Gold glyconanoparticles (GNPs), prepared from synthetic oligosaccharides, have been used to overcome thi

Full text of DOI:10.1002/ejoc.200500256

FULL PAPER
A Facile Method for the Preparation of Gold Glyconanoparticles from Free
Oligosaccharides and Their Applicability in Carbohydrate-Protein Interaction
Studies
Koen M. Halkes,[a] Adriana Carvalho de Souza,[a] C. Elizabeth P. Maljaars,[a]
Gerrit J. Gerwig,[a] and Johannis P. Kamerling*[a]
Dedicated to Professor András Lipták on the occasion of his 70th birthday
Keywords: Gold glyconanoparticles / Carbohydrates / Surface plasmon resonance / UV/Vis spectroscopy / Transmission
electron microscopy
The weak binding affinity of monomeric oligosaccharides
with carbohydrate-binding proteins are hampering their use
in in-vivo and in-vitro bio-assays. Gold glyconanoparticles
(GNPs), prepared from synthetic oligosaccharides, have been
used to overcome this weak binding affinity. In this paper,
a convenient method for the preparation of GNPs from free
oligosaccharides is presented. The reductive amination of
saccharides with trityl-protected cysteamine, followed by de-
tritylation, afforded cysteamine-extended saccharides that
could be used for the preparation of GNPs under reducing
conditions in water. The robust chemistry and facile purifica-
tion of intermediate and final compounds ensure high yields
and reproducible results and the, subsequent, preparation of
GNPs proceeded smoothly, even with minute quantities
(nanomolar scale) of the cysteamine-extended saccharide.
The described method was used to synthesize a series of
gluco- and manno-oligosaccharide-containing GNPs. The
prepared GNPs were validated in interaction studies with
Con A, using either surface plasmon resonance (SPR), UV/
Vis spectroscopy, or transmission electron microscopy (TEM).
The described method for the preparation of water-soluble
gold glyconanoparticles can be used for the identification of
carbohydrate ligands for novel carbohydrate-binding pro-
teins, and can find application as inhibitors of pathological
interactions.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2005)
The interest in designing multivalent or clustered carbo-
hydrate analogues, that will mimic the natural cell or tissue
surface and bind with high affinity to proteins, has been
growing. Synthetic oligosaccharides, that are elongated with
a suitable spacer, have been conjugated to carrier-proteins,[6]
dendrimers,[7] and polystyrenes.[8] Recently, gold nanopar-
ticles were coated with biologically relevant, synthetic sac-
charides.[9,10] The globular shape and the multivalent dis-
play of oligosaccharides at their surface make gold glyco-
nanoparticles (GNPs) extremely suitable to overcome both
the low binding affinity of monomeric oligosaccharides to
carbohydrate-binding proteins[11,12] and the weak carbo-
hydrate–carbohydrate self-recognition.[13,14] Complete water
solubility, high storage stability, and no or low cytotoxic
activity are some of the advantages associated with the use
of GNPs in in-vitro and in-vivo bio-assays.
Introduction
Complex glycans that are attached to proteins and lipids
coat all eukaryotic cells and are also found in the extracellu-
lar space between cells. They affect, through non-covalent
interactions with other biomolecules, a great variety of bio-
logical processes ranging from development to cell-ad-
hesion, and from signaling to infections by viral and bacte-
rial agents.[1–4] Therefore, complex glycans represent a
promising source for the development of new pharmaceuti-
cal agents. Before being able to tap the vast potential of
complex glycans in drug development, fundamental chal-
lenges to clarify important structure-activity relationships
have to be overcome. One of these challenges is the weak
interaction of a carbohydrate in a 1:1 complex with its com-
plementary biomolecule (Kd Ϸ 10–3 ). To give these inter-
actions biological relevance, Nature often presents carbo-
hydrates multivalently or clustered on cell or tissue sur-
faces.[5]
The synthesis of the thiol-spacer-containing oligosaccha-
rides, necessary for the preparation of the above-mentioned
GNPs, can be difficult and time-consuming. To further the
use of GNPs in the field of glycobiology, simple methods
for the preparation of GNPs from isolated glycans (e.g.
from N-glycoproteins or polysaccharide fragments) has to
[a] Bijvoet Center, Department of Bio-Organic Chemistry, Utrecht
University,
Padualaan 8, 3584 CH Utrecht, The Netherlands
E-mail: J.P.Kamerling@chem.uu.nl
3650
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/ejoc.200500256
Eur. J. Org. Chem. 2005, 3650–3659

Gold Glyconanoparticles in Carbohydrate-Protein Interaction Studies
FULL PAPER
be explored. Robust chemistry for the introduction of a
thiol-containing spacer in free oligosaccharides, facile puri-
fication protocols of the thiol-spacer-containing oligosac-
charides, and the reliable generation of stable and water-
soluble GNPs are a few of the required characteristics for
the method to become successful.
Results and Discussion
Synthesis of Compounds 1–21
To synthesize GNPs from free oligosaccharides, a thiol-
containing spacer has to be introduced at their reducing
In the current paper, such a method for the facile prepa- end. Reductive amination is a well-documented and robust
ration of GNPs on a small scale from free saccharides is chemical method to connect an aglycon to a glycan and
described. The applicability of the approach is demon- was, therefore, used for the introduction of cysteamine into
strated by the generation of a wide range of gluco- and free saccharides. Prior to the reductive amination reaction,
manno-oligosaccharide-containing GNPs and their use in the thiol function in cysteamine was protected with a trityl
SPR, UV/Vis, and TEM based interaction studies with the group to avoid side reactions. The trityl-protecting group
model lectin Concanavalin A.
was chosen as it can be smoothly removed under mild
Figure 1. Synthesized thiol-spacer-containing saccharides.
Eur. J. Org. Chem. 2005, 3650–3659
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3651

J. P. Kamerling et al.
FULL PAPER
acidic conditions and, additionally, its hydrophobic charac- after several cycles of lyophilization and redissolving in
ter facilitates the purification of the reaction products by water. The GNPs Au-13 to Au-17 and Au-19 to Au-21, that
RP-HPLC or in solid-phase extraction procedures. The hy- were derived from oligosaccharides, showed complete col-
drochloric acid salt of cysteamine was treated with trityl loidal stability, even in buffers of elevated salt concentra-
chloride in DMF to afford pure 1 (90% yield).
tions (HEPES buffer: 10 m HEPES pH 7.4, 0.15  NaCl).
1
The free saccharide (glucose, maltose – maltohexaose, The various compounds were characterized by H NMR,
mannose, mannotriose, mannopentaose, or lactose) and 1 transmission electron microscopy (TEM), and monosaccha-
were dissolved in anhydrous, acidified dimethyl sulfoxide, ride composition analysis. The mean diameter, the percen-
and sodium cyanoborohydride was added. After heating at tage of carbohydrate, and the carbohydrate-surface cover-
65 °C for 2 h, followed by degradation of the reducing age of GNPs Au-13 to Au-17 and Au-19 to Au-21 are shown
agent, dimethyl sulfoxide and unreacted saccharide were re- in Table 1. Although the size and percentage of carbo-
moved from the mixture by solid-phase extraction. Remain- hydrate on the GNPs vary to some extent, the carbo-
ing 1 could be easily removed from the desired product by hydrate-surface coverage of the GNPs are in a narrow range
silica gel chromatography for the smaller saccharides (between 9 and 16%) when the thiol-spacer derivatized oli-
(mono- to trisaccharide) or by either RP-HPLC or solu- gosaccharides are treated with tetrachloroauric acid in a 1:1
tion-phase extraction procedures for the larger oligosaccha- molar ratio.
rides. The intermediates 2–11 (Figure 1) were, after charac-
As it is well known that reductive aminations can be eas-
terization by high-resolution matrix-assisted laser desorp- ily performed with nanomolar amounts of starting oligo-
tion ionization time-of-flight (MALDI-TOF) mass spec- saccharides, it was investigated if GNPs could be prepared
trometry, directly used in the next reaction step. The trityl with this amount of cysteamine-extended oligosaccharides,
group was smoothly removed under acidic conditions, in thereby, making the method more attractive for the field of
the presence of a cation scavenger, to yield the pure prod- glycoscience. By replacing magnetic for ultrasonic stirring
ucts 12–21 (Figure 1) in good overall yields (61 to 85%), and using a more dilute NaBH4 solution, it was possible to
after solid-phase extraction. All products were charac- prepare stable GNPs from 17 and tetrachloroauric acid
terized by high-resolution mass spectrometry and 1D and (1:3) on a 95 nmol scale (Au-17a, Dcore = 2.46 1.1 nm).
2D NMR spectroscopy. It should be noted that the isolated
products oxidized, over time, completely into their corre-
sponding disulfide analogues, as seen by a downfield shift Application of GNPs Au-13 to Au-17 and Au-19 to Au-21
of approximately 0.25 ppm for the proton signals of the cys- in Interaction Studies with Concanavalin A
teamine residue.
To validate the usefulness of the prepared GNPs in vari-
ous types of interaction studies, a binding study with the
model lectin Concanavalin A (Con A) was initiated. Con A
Synthesis of Gold Glyconanoparticles
is a well-described lectin that binds α--mannose/α--glu-
A modified literature procedure was used for the prepa- cose-containing glycans and, has long been recognized as
ration of GNPs Au-12 to Au-21.[15] Thus, an aqueous solu- an excellent system to study multivalent effects. The lectin
tion of thiol-spacer-containing saccharide 12 to 21 was is a tetramer above pH 7.0, and a dimer below pH 6.0. The
added to an equimolar amount of aqueous tetrachloroauric carbohydrate-binding site in Con A is an extended groove
acid, and incubated for 2 h with an excess of sodium on the surface of the protein, that can easily accommodate
borohydride. The generated GNPs were purified by centrif- carbohydrates ranging from mono- to pentasaccharides,
ugal filtration. GNPs Au-12 and Au-18, derived from glu- and is best suited to recognize the unique spatial arrange-
cose and mannose, respectively, were not completely stable, ment of the trisaccharide α--Manp-(1Ǟ3)-[α--Manp-
as judged by the minute amounts of precipitation formed (1Ǟ6)]--Man.[18] Extrapolating from literature,[18] GNPs
Table 1. Physical characteristics of gold glyconanoparticles.
Glyconanoparticle
Dcore (nm)
% CHO (carbohydrate)[16]
# gold atoms/
# CHO molecules[a]
% surface coverage
/
surface atoms[17]
Au-13
Au-14a
Au-14
Au-14b
Au-15
Au-16
Au-17
Au-19
Au-20
Au-21
3.38 1.7
1.99 0.9
2.16 1.1
3.12 1.8
2.00 0.9
2.79 1.0
1.79 0.9
3.84 1.6
2.71 1.1
1.97 0.8
10%
11%
19%
10%
12%
21%
28%
9%
1298/482
314/174
309/162
1133/436
314/174
807/348
314/174
2406/752
807/348
314/174
70/15%
32/12%
25/16%
64/15%
77/12%
47/13%
23/13%
53/11%
34/10%
98/9%
16%
9%
[a] Calculated according to nCHO = [(nAu ×MWAu/100 – %CHO)×%CHO]/MWCHO; where nCHO is the number of sugar molecules on GNP,
nAu is the number of gold atoms in GNP.
3652
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 3650–3659

Gold Glyconanoparticles in Carbohydrate-Protein Interaction Studies
FULL PAPER
Au-13 to Au-17, containing α--glucose ligands, can be clas-
sified as weak Con A binders, while the GNPs Au-19 to Au-
20, containing α--mannose ligands, constitute the strong
Con A binders. In this paper, SPR, UV/Vis, and TEM stud-
ies of binding events between the model lectin Con A and
the different GNPs are used as examples of their applica-
bility in interaction studies.
SPR Studies: Con A was immobilized, as a dimer, on
two channels of a CM5 sensor chip at 400 RU and 7,000
RU, respectively. Ethanolamine was immobilized on a third
channel of the CM5 sensor chip, and served as the blank
surface. Solutions of Au-13 to Au-17 and Au-19 to Au-21
(100 µg mL–1) in HEPES buffer, containing 10 mM CaCl2,
were flown over the sensor chip. As expected, Au-21, dis-
playing β--galactose residues on its surface, showed no
binding on either of the Con A channels nor on the blank
surface. On the sensor chip channel with 400 RU of Con
A, only Au-19 and Au-20 showed a strong SPR binding
signal (175 and 625 RU, respectively) while no detectable
binding events with GNPs Au-13 to Au-17 could be ob-
served. On the channel with 7,000 RU Con A, all gluco-
and manno-oligosaccharide-containing GNPs showed
strong SPR signals. An explanation of this finding can be
ascribed to the affinity order for the carbohydrate-binding
site of Con A: manno-oligosaccharides ϾϾ dextran poly-
mers Ͼ gluco-oligosaccharides.[18] As dextran polymers
form the matrix of CM5 sensor chips, competition reactions
between the analytes and matrix polymers for the binding
site of immobilized Con A have to be taken into consider-
ation. As a consequence of the ligand affinity order of Con
A, the mannose-containing GNPs (Au-19 and Au-20) can
efficiently compete for the carbohydrate-binding site of Con
A leading to a SPR signal on both the 400 and 7,000 RU
Con A channel. However, the gluco-oligosaccharide-con-
taining GNPs will not be able to remove dextran from the
carbohydrate-binding site of Con A. In the flow cell with
400 RU Con A, the chance that all the binding sites are
occupied by dextran is high and, therefore, no SPR signal
Figure 2. Biacore sensorgrams of the interactions between various
GNPs (100 µg mL–1) and Con A. a: Curves from top to bottom:
Au-13, Au-16, Au-14, Au-15, Au-17, and Au-21. b: Curves from top
to bottom: Au-19, Au-20, and Au-21. c: Curves from top to bottom:
is observed when GNPs Au-13 to Au-17 are injected. Statis- Au-14b, Au-14, and Au-14a.
tically, in the 7,000 RU Con A channel free carbohydrate-
binding sites are present for binding with Au-13 to Au-17.
In Figure 2 (parts a and b), the binding curves of the is clearly visible that the largest particle gives rise to the
various GNPs, at a 100 µg mL–1 concentration, on the 7,000 highest SPR response.
RU Con A sensor surface are shown. The effect of the avid-
The avidity of the GNPs for Con A can be, qualitatively,
ity on the SPR response is most clearly seen when parts a established from the dissociation phase of the binding
and b of Figure 2 (a: sensorgrams of weak Con A binders; curves. Dissociation can be clearly seen in the Au-13 to Au-
b: sensorgrams of strong Con A binders) are compared. 17 series (Figure 2, part a), some dissociation can be ob-
The strong interaction of the manno-oligosaccharide-pre- served for Au-19, and no or negligible dissociation is seen
senting GNPs ensures that more material is collected on the for Au-20 (Figure 2, part b). Another qualitative indication
sensor surface as compared to the gluco-oligosaccharide- for the avidity of the GNPs for the lectin is the method of
presenting GNPs, thereby, leading to a stronger SPR re- regeneration of the sensor chip surface. Complete regenera-
sponse. Besides the avidity of the GNPs for Con A, also the tion of the Con A surface that has been treated with Au-13
size of the GNPs is reflected in the observed SPR response. to Au-17 is achieved by injecting, for 5 min, 150 m methyl
To demonstrate this effect, GNPs Au-14a (Dcore = 1.99
α--mannopyranoside. Under the same regeneration condi-
0.9 nm), Au-14 (Dcore = 2.16 1.1 nm), and Au-14b (Dcore tions, approximately 30% of Au-19 and none of the Au-
= 3.12 1.8 nm) were prepared from 14 and tetrachloroau- 20 was removed from the sensor surface. More stringent
ric acid, in different molar ratios (see Exp. Sect.). The sen- regeneration conditions are required for the removal of Au-
sorgrams of these GNPs are shown in part c of Figure 2. It 19 and Au-20 from the sensor surface (see Exp. Sect.).
Eur. J. Org. Chem. 2005, 3650–3659
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3653

J. P. Kamerling et al.
FULL PAPER
To establish the minimum GNP concentration that could clearly demonstrating the multivalent binding character be-
still be observed in SPR binding experiments, a dilution tween the GNPs and Con A.
series of Au-15 (Figure 3, part a) and Au-19 (Figure 3, part
UV/Vis Studies: Previously, it has been shown[11] that lac-
b) was injected over the 7,000 RU Con A sensor surface. It tose-containing GNPs (prepared starting from p-ami-
is clear that even at the lowest concentrations tested (2.34 nophenyl lactoside and PEGylated gold nanoparticles) ex-
µg mL–1 and 0.26 µg mL–1 for Au-15 and Au-19, respec- hibit selective aggregation when exposed to Ricinus commu-
tively) a detectable SPR response could be seen (10 to 20 nis agglutinin (RCA), a bivalent lectin specifically recogniz-
RU).
ing terminal β--galactose residues. During the aggregation
of the GNPs with RCA, a broadening and red-shift in the
particle surface plasmon band from 520 nm to a longer
wavelength[19] was observed, leading to a loss of absorption
at 520 nm. Recently, the same phenomenon was described
for mannose-coated gold nanoparticles that were incubated
with tetravalent Con A.[20] The loss of absorption at
520 nm[11] or the increase of absorption at 620 nm[20] could
be followed by UV/Vis spectroscopy, and used in qualitative
and quantitative bio-assays. Although the GNPs Au-13 to
Au-17 and Au-19 to Au-21 are significantly smaller than the
ones used above (8.9 nm[11] or 16 nm[20]), an absorption
band near λ = 520 nm could be observed in all cases and
used for monitoring aggregation events in an UV/Vis-spec-
trometer. Figure 5 shows the representative UV/Vis spectra
of GNPs Au-13 to Au-17 and Au-19 to Au-21, recorded for
6 h after addition of tetravalent Con A. With the exemption
of Au-21 (red line in Figure 5) that, as expected, showed no
specific binding to Con A, all GNPs show the same pattern
in the UV/Vis spectra: An initial increase in absorbance at
520 nm that is followed by a decrease in absorbance. The
initial increase in absorbance is suggested to originate from
the initial formation of small clusters (approximately
10 nm) of Con A-interlinked GNPs that may lead to a
stronger surface plasmon band. Over the time, the cluster
Figure 3. Composite sensorgrams of GNP concentration experi-
ments with Con A. a: Au-15 (concentration from top to bottom:
300, 150, 75, 37.5, 18.75, 9.38, 4.69, 2.34 µg mL–1). b: Au-19 (con-
centration from top to bottom: 66.67, 33.33, 16.67, 8.33, 4.16, 2.08,
1.04, 0.52, 0.26 µg mL–1).
Finally, a competition experiment between Au-17 (100 µg
mL–1; containing 0.028 mg intact maltopentaose) and mal-
topentaose (0.03–1.79 mg) was performed. The composite
sensorgrams shown in Figure 4 illustrate that a large excess
of free maltopentaose had to be used to obtain complete
inhibition of binding between Au-17 and Con A, thereby
Figure 4. Composite sensorgrams of competition experiments be-
tween GNPs Au-17 (100 µg mL–1 stock solution) and maltopen-
taose for the binding to Con A. To 100 µL of the Au-17 stock
solution were added increasing amounts of maltopentaose (from
top to bottom: 0, 0.03, 0.06, 0.12, 0.24, 0.89, and 1.79 mg).
Figure 5. Composite UV/Vis spectra of the aggregation between
different GNPs (50 µg) and Con A (50 µg) in 550 µL HEPES
buffer, containing 10 m CaCl2 (pH 7.4).
3654
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 3650–3659

Gold Glyconanoparticles in Carbohydrate-Protein Interaction Studies
FULL PAPER
size increases with the concomitant broadening and red-
shift in the particle surface plasmon band from 520 nm to
a longer wavelength.[19] This results in a decrease of the
measured absorbance at 520 nm, which is further aug-
mented by the precipitation of the larger aggregates. The
aggregation process is completely reversible, as judged by
the clear, red/brown colored solution that was obtained af-
ter the addition of aqueous 150 mM methyl α--mannopyr-
anoside (50 µL) to the precipitated aggregates. The time-
frame in which this increase and subsequent decrease of
absorbance at 520 nm occurs, depends on the strength of
the interaction between Con A and the ligand that is dis-
played on the surface of the GNP. In the ligand series tested
in this study, both literature[18] and our SPR data have pre-
dicted that ligand 20 would have the highest affinity for
Con A. Accordingly, the decline of the UV/Vis spectrum of
Au-20 (light-green line) is the steepest of the tested com-
pounds.
TEM Studies: The TEM micrographs of the aggregates
formed during the UV/Vis experiments between either Au-
17 or Au-20 and tetravalent Con A are shown in Figure 6.
While the TEM micrographs of the aggregation between
Au-17 and Con A showed aggregates of up to 150 nm
throughout the whole grid (see Figure 6, part b), the aggre-
gates formed between Au-20 and Con A were much larger
(up to 7 µm; see Figure 6, part d).
Conclusions
A method has been established for the preparation of
GNPs from free oligosaccharides. The crucial part of this
method is the two-step reaction sequence for the introduc-
tion of a thiol-spacer in the free oligosaccharides. Via re-
ductive amination, trityl-protected cysteamine was intro-
duced into the glycan. Subsequent removal of the trityl
group rendered thiol-spacer-containing oligosaccharides, in
good overall yields, that can be used in the preparation of
GNPs. The preparation of GNPs from the thiol-spacer-ex-
tended oligosaccharides proceeded smoothly, even with
minute amounts of starting material. The prepared GNPs
show complete colloidal stability, even in buffers with high
salt concentrations. The prepared GNPs have been vali-
dated by their application in SPR-, UV/Vis-, and TEM-
based binding assays with model lectin Con A.
The study of the proteome has identified a range of novel
carbohydrate-binding proteins and, their role in biological
and pathological phenomena has to be established. To
achieve this, it is necessary to identify their carbohydrate
ligands and, subsequently, perform interaction studies with
them. Glycans, released from proteins and lipids or ob-
tained via (partial) hydrolysis of polysaccharides, form a
rich source of potential ligands for these biomolecules.
GNPs, prepared from isolated glycans by using the pre-
sented method, can find application in ligand-identification
processes and interaction studies with carbohydrate-binding
proteins and, ultimately, help to elucidate their role in bio-
logical processes. In addition, the GNPs can find applica-
Figure 6. TEM images of aggregation events of GNPs Au-17 and
Au-20 with Con A. a: Au-17 in water (50 nm). b: Au-17 incubated
overnight with Con A in HEPES buffer (pH 7.4), containing 10 m
CaCl2 (500 nm). c: Au-20 in water (50 nm). d: Au-20 incubated
overnight with Con A in HEPES buffer (pH 7.4), containing 10 m
CaCl2 (500 nm).
Eur. J. Org. Chem. 2005, 3650–3659
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3655

J. P. Kamerling et al.
FULL PAPER
S2-(Trityl)cysteamine (1): Cysteamine hydrochloride (200 mg,
1.73 mmol) was dissolved in dichloromethane-dimethylformamide
(1:1, 10 mL). Trityl chloride (725 mg, 2.60 mmol) was added, and
the solution was stirred for 2 h at room temperature. Then, the
mixture was concentrated in vacuo and subsequently co-concen-
trated with toluene (3 × 15 mL), ethanol (3 × 15 mL), and dichloro-
methane (3 × 15 mL). The crude product was dissolved in dichloro-
methane (50 mL) and washed with saturated aq. NaHCO3 (25 mL).
The organic phase was dried, filtered, and concentrated in vacuo.
The product was purified by flash chromatography (dichlorometh-
ane-ethyl acetate, 9:1 Ǟ dichloromethane/methanol, 4:1) to yield
tion in the bio-medical field as inhibitors of pathological
interactions and in the development of site-specific drug
targeting devices.
Experimental Section
General: All chemicals were of reagent grade, and were used with-
out further purification. The malto-oligosaccharides were pur-
chased from Sigma–Aldrich Chemie BV (Zwijndrecht, The Nether-
lands) and the manno-oligosaccharides were obtained from Dextra
Laboratories Ltd. (Reading, United Kingdom). NaCNBH3,
NaBH4, and tetrachloroauric acid were purchased from Acros Or-
ganics (Geel, Belgium). N-Hydroxysuccinimide was purchased
from Merck (NJ, USA), N-ethyl-NЈ-(dimethylaminopropyl)car-
bodiimide, ethanolamine, and Concanavalin A lectin from Canav-
alia ensiformis (Con A) from Sigma (St. Louis, USA). Cysteamine
hydrochloride was supplied by Fluka (Buchs, Switzerland). C-18
Extract-CleanTM columns were purchased from Alltech (Breda,
The Netherlands).
1
pure 1 (500 mg, 90%). H NMR (300 MHz, CDCl3): δ = 2.33 (t, 2
H; H2NCH2CH2STrt), 2.59 (t, 2 H; H2NCH2CH2STrt), 7.22 and
7.43 (2 m, 9 H and 6 H; 15 CHarom).
General Procedure for the Reductive Amination of Saccharides with
1 (Ǟ 2–11): The free saccharide (1 equiv.) and 1 (2 equiv.) were
dissolved in a 5% anhydrous solution of acetic acid in dimethyl
sulfoxide to a saccharide concentration of approximately 30 m.
Sodium cyanoborohydride (20 equiv.) was added, and the mixture
was heated for 2 h at 65 °C. After acidification with acetic acid and
dilution with water (15 mL), the mixture was loaded on a C-18
Extract-CleanTM column and unreacted saccharide was eluted with
water (15 mL). Subsequent elution with methanol (15 mL) afforded
a mixture of 1 and the S2-(trityl)cysteamine-extended saccharide.
The methanol fraction was concentrated in vacuo. Starting saccha-
rides smaller than a tetrasaccharide were purified by column
chromatography (dichloromethane-methanol, 6:4 Ǟ methanol Ǟ
methanol/water, 1:1), while larger oligosaccharides were purified by
semi-preparative RP-HPLC or solution-phase extraction pro-
cedures. The pure intermediates were characterized by high-resolu-
tion MALDI-TOF MS, and directly used in the next reaction step.
Reactions were monitored by TLC on Silica Gel 60 F254 (Merck);
after examination under UV light, compounds were visualized by
heating with 10% (v/v) methanolic H2SO4, orcinol (2 mg mL–1) in
20% (v/v) methanolic H2SO4, or ninhydrin (1.5 mg mL–1) in 1-bu-
tanol/water/acetic acid (38:1.75:0.25). In the work-up procedures of
reaction mixtures, organic solutions were washed with appropriate
amounts of the indicated aqueous solutions, then dried with
MgSO4, and concentrated under reduced pressure at 30–50 °C on
a water bath. Column chromatography was performed on Silica
Gel 60 (Merck, 0.040–0.063 mm). Purification of the larger S2-(tri-
tyl)cysteamine-extended oligosaccharides was performed by re-
verse-phase HPLC on a Knauer HPLC system with a Polaris C18-
A column (250 × 4.6 mm) and a linear gradient of solvent A (0.1%
trifluoroacetic acid in 10% aq. acetonitrile) and solvent B (0.08%
trifluoroacetic acid in 90% aq. acetonitrile) at a flow rate of
1 mL min–1.
(1-Deoxy-D
-glucityl)-(1ǞN)-S2-(trityl)cysteamine (2): High-resolu-
tion MS data of C27H33NO5S (M = 483.208): M + Na found
506.189, calculated 506.198.
α-D-Glucopyranosyl-(1Ǟ4)-(1-deoxy-D
-glucityl)-(1ǞN)-S2-(trityl)-
1H NMR spectra were recorded at 300 K with a Bruker AC 300
(300 MHz) or a Bruker AMX 500 (500 MHz) spectrometer; δH val-
ues are given in ppm relative to the signal for internal acetone (δH
= 2.22, D2O). Two-dimensional 1H-1H TOCSY (mixing times 7
and 100 ms) and 1H-13C correlated HSQC spectra were recorded
at 300 K with a Bruker AMX 500 spectrometer; δC values are given
in ppm relative to the signal of internal acetone (δC = 30.9, D2O).
Surface plasmon resonance studies were carried out on a Biacore
2000 instrument, using a CM5 sensor chip and Biaevaluation soft-
ware 4.1 (Pharmacia Biosensor AB, Uppsala, Sweden). Spectro-
photometric measurements were performed on a Hewlett–Packard
8452A diode array instrument. All the exact masses of the oligosac-
charides were measured by matrix-assisted laser desorption ioniza-
tion time-of-flight mass spectrometry (MALDI-TOF MS) using a
Voyager-DE Pro (Applied Biosystems) instrument in the reflector
mode at a resolution of 5000 FWHM. 2,5-Dihydroxybenzoic acid
in 50% aq. acetonitrile (10 mg mL–1) was used as a matrix. All
spectra were recorded in the positive ion-mode and a ladder of
malto-oligosaccharides (G3-G13) was added as internal standard.
The electrospray ionization mass spectra of compounds 12 and 18
were measured on a LCQ DecaXP ion-trap mass spectrometer
equipped with an electrospray ion source (Thermo Finningan, San
Jose, CA). The samples were dissolved in the mobile phase, propa-
nol/water (1:1) to a concentration of approximately 10–30 pM µL–1
and introduced via a syringe pump at a flow rate of 5 µL min–1.
The capillary temperature was set to 225 °C and spectra were ac-
quired in the positive ion-mode by scanning over m/z 50–2000 with
a spray voltage of 4.8 kV.
cysteamine (3): High-resolution MS data of C33H43NO10S (M =
645.261): M + Na found 668.253, calculated 668.251.
α-D-Glucopyranosyl-(1Ǟ4)-α-D-glucopyranosyl-(1Ǟ4)-(1-deoxy-D-
glucityl)-(1ǞN)-S2-(trityl)cysteamine (4): High-resolution MS data
of C39H53NO15S (M = 807.313): M + Na found 830.297, calculated
830.303.
α-
D
-Glucopyranosyl-(1Ǟ4)-α-
D-glucopyranosyl-(1Ǟ4)-α-D-glucopy-
ranosyl-(1Ǟ4)-(1-deoxy-
D
-glucityl)-(1ǞN)-S2-(trityl)cysteamine
(5): High-resolution MS data of C45H63NO20S (M = 969.366): M
+ Na found 992.360, calculated 992.356.
α-
D-Glucopyranosyl-(1Ǟ4)-α-
D-glucopyranosyl-(1Ǟ4)-α-D-glucopy-
ranosyl-(1Ǟ4)-α-
D
-glucopyranosyl-(1Ǟ4)-(1-deoxy-
D-glucityl)-
(1ǞN)-S2-(trityl)cysteamine (6): High-resolution MS data of
C51H73NO25S (M = 1131.419): M + Na found 1154.424, calculated
1154.409.
α-
ranosyl-(1Ǟ4)-α-
(1Ǟ4)-(1-deoxy-
-glucityl)-(1ǞN)-S2-(trityl)cysteamine (7): High-
D-Glucopyranosyl-(1Ǟ4)-α-
D-glucopyranosyl-(1Ǟ4)-α-D-glucopy-
D
-glucopyranosyl-(1Ǟ4)-α-
D
-glucopyranosyl-
D
resolution MS data of C57H83NO30S (M = 1293.472): M + Na
found 1316.444, calculated 1316.462.
(1-Deoxy-D
-mannityl)-(1ǞN)-S2-(trityl)cysteamine (8): High-reso-
lution MS data of C27H33NO5S (M = 483.208): M + Na found
506.197, calculated 506.198.
α-
oxy-
D
-Mannopyranosyl-(1Ǟ3)-[(α-
D
D
-mannopyranosyl)-(1Ǟ6)]-(1-de-
-mannityl)-(1ǞN)-S2-(trityl)cysteamine (9): High-resolution
3656
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 3650–3659

Gold Glyconanoparticles in Carbohydrate-Protein Interaction Studies
FULL PAPER
MS data of C39H53NO15S (M = 807.313): M + Na found 830.299,
calculated 830.303.
amorphous powder (4 mg), overall yield 68%. 1H NMR (500 MHz,
D2O, 2D TOCSY): δ = 5.40 (d, J1ЈЈ,2ЈЈ = J1ЈЈЈ,2ЈЈЈ = 3.9 Hz, 2 H;
H1ЈЈ and H1ЈЈЈ), 5.13 (d, J1Ј,2Ј = 3.8 Hz, 1 H; H1Ј), 4.16 (m, 1 H;
H2), 3.88 (m, 2 H; H3 and H4), 3.45 (m, 2 H; DNCH2CH2SD),
3.42 (br. t, J3ЈЈЈ,4ЈЈЈ = J4ЈЈЈ,5ЈЈЈ = 9.5 Hz, 1 H; H4ЈЈЈ), 3.26 (m, 2 H;
H1a and H1b), 3.06 (t, 2 H; DNCH2CH2 SD). 13 C NMR
(125 MHz, D2O, HSQC): δ = 101.1 (C1Ј), 100.4 (C1ЈЈ and C1ЈЈЈ),
82.3 (C4), 72.3 (C3), 70.1 (C4ЈЈЈ), 67.8 (C2), 50.9 (C1), 46.7
(NCH2), 33.2 (CH2S); High-resolution MS data of C26H49NO20S
(M = 727.257): M + Na found 750.239, calculated 750.247.
{α-
mannopyranosyl-(1Ǟ6)}-[(α-
-mannityl)-(1ǞN)-S2-(trityl)cysteamine (10): High-resolution MS
D
-Mannopyranosyl-(1Ǟ6)-[(α-
D-mannopyranosyl)-(1Ǟ3)]-α-D-
D
-mannopyranosyl)-(1Ǟ3)]-(1-deoxy-
D
data of C51H73NO25S (M = 1131.419): M + Na found 1154.397,
calculated 1154.409.
β-D-Galactopyranosyl-(1Ǟ4)-(1-deoxy-D
-glucityl)-(1ǞN)-S2-(tri-
tyl)cysteamine (11): High-resolution MS data of C33H43NO10S (M
= 645.261): M + Na found 668.249, calculated 668.251.
α-
D-Glucopyranosyl-(1Ǟ4)-α-
D-glucopyranosyl-(1Ǟ4)-α-D-glucopy-
General Procedure for the Removal of the Trityl Protecting Group
(Ǟ 12–21): The S2-(trityl)cysteamine-extended saccharide was dis-
solved in a mixture of dichloromethane (1 mL) and H2O (1 mL).
Triisopropylsilane (75 µL) and trifluoroacetic acid (4 mL) were
added, and the two-phase system was vigorously stirred for 15 min,
then co-concentrated with toluene (3 × 20 mL) and ethanol (2 ×
15 mL). The amorphous solid was dissolved in water (5 mL) and
loaded on a C-18 Extract-CleanTM column. The fraction eluted
with water was lyophilized to yield the cysteamine-extended saccha-
ride.
ranosyl-(1Ǟ4)-α-
D
-glucopyranosyl-(1Ǟ4)-(1-deoxy-
D-glucityl)-
(1ǞN)-cysteamine (16): White amorphous powder (6 mg), overall
1
yield 81%. H NMR (500 MHz, D2O, 2D TOCSY): δ = 5.40 (m,
3 H; H1ЈЈ, H1ЈЈЈ, and H1ЈЈЈЈ), 5.14 (d, J1Ј,2Ј = 4.0 Hz, 1 H; H1Ј),
4.20 (m, 1 H; H2), 3.53 (m, 2 H; DNCH2CH2SD), 3.42 (br. t,
J3ЈЈЈЈ,4ЈЈЈЈ = J4ЈЈЈЈ,5ЈЈЈЈ = 9.5 Hz, 1 H; H4ЈЈЈЈ), 3.33 (m, 2 H; H1a and
H1b), 3.09 (t, 2 H; DNCH2CH2SD). 13C NMR (125 MHz, D2O,
HSQC): δ = 101.0 (C1Ј), 100.4 (C1ЈЈ, C1ЈЈЈ, and C1ЈЈЈЈ), 70.0
(C4ЈЈЈЈ), 67.4 (C2), 50.9 (C1), 46.5 (NCH2), 32.5 (CH2S); High-
resolution MS data of C32H59NO25S (M = 889.309): M + Na found
912.291, calculated 912.299.
(1-Deoxy-D-glucityl)-(1ǞN)-cysteamine (12): White powder
(8.0 mg), overall yield 63 %. 1H NMR (500 MHz, D2O, 2D
TOCSY): δ = 4.12 (m, 1 H; H2), 3.83 (dd, J2,3 = 5.1, J3,4 = 2.2 Hz,
1 H; H3), 3.82 and 3.66 (2 m, each 1 H; H6a and H6b), 3.77 (m,
1 H; H5), 3.65 (dd, J4,5 = 4.1 Hz, 1 H; H4), 3.47 (m, 2 H;
DNCH2CH2SD), 3.28 (dd, J1a,2 = 2.9, J1a,1b = 12.2 Hz, 1 H; H1a),
3.19 (dd, J1b,2 = 9.6 Hz, 1 H; H1b), 3.06 (t, 2 H; DNCH2CH2SD).
13C NMR (125 MHz, D2O, HSQC): δ = 71.6 (C5), 71.5 (C4), 71.4
(C3), 69.1 (C2), 63.4 (C6), 50.4 (C1), 46.7 (NCH2), 33.1 (CH2S);
High-resolution MS data of C8H19NO5S (M = 241.098): M + Na
found 264.081, calculated 264.088.
α-
ranosyl-(1Ǟ4)-α-
(1Ǟ4)-(1-deoxy- -glucityl)-(1ǞN)-cysteamine (17): White amorph-
D
-Glucopyranosyl-(1Ǟ4)-α-
D
-glucopyranosyl-(1Ǟ4)-α-
D-glucopy-
D
-glucopyranosyl-(1Ǟ4)-α-
D
-glucopyranosyl-
D
1
ous powder (4 mg), overall yield 76%. H NMR (500 MHz, D2O,
2D TOCSY): δ = 5.39 (m, 4 H; H1ЈЈ, H1ЈЈЈ, H1ЈЈЈЈ, and H1ЈЈЈЈЈ),
5.12 (d, J1Ј,2Ј = 4.0 Hz, 1 H; H1Ј), 4.14 (m, 1 H; H2), 3.42 (br. t,
J3ЈЈЈЈЈ,4ЈЈЈЈЈ = J4ЈЈЈЈЈ,5ЈЈЈЈЈ = 9.5 Hz, 1 H; H4ЈЈЈЈЈ), 3.40 (m, 2 H;
DNCH2CH2SD), 3.20 (m, 2 H; H1a and H1b), 3.04 (t, 2 H;
DNCH2CH2SD). 13C NMR (125 MHz, D2O, HSQC): δ = 101.1
(C1Ј), 100.4 (C1ЈЈ, C1ЈЈЈ, C1ЈЈЈЈ, and C1ЈЈЈЈЈ), 70.0 (C4ЈЈЈЈЈ), 68.1
(C2), 51.1 (C1), 46.8 (NCH2), 33.7 (CH2S); High-resolution MS
data of C38H70NO30S (M = 1051.363): M + Na found 1074.364,
calculated 1074.353.
α-D-Glucopyranosyl-(1Ǟ4)-(1-deoxy-D-glucityl)-(1ǞN)-cysteamine
(13): White amorphous powder (16 mg), overall yield 61 %. 1H
NMR (500 MHz, D2O, 2D TOCSY): δ = 5.12 (d, J1Ј,2Ј = 3.8 Hz,
1 H; H1Ј), 4.17 (m, 1 H; H2), 3.99 (m, 1 H; H5), 3.88–3.87 (m, 2
H; H3 and H4), 3.86 (m, 1 H; H5Ј), 3.85 (dd, J5Ј,6aЈ = 2.2, J6aЈ,6bЈ
= 12.4 Hz, 1 H; H6aЈ), 3.79 (dd, J5,6a = 4.6, J6a,6b = 11.9 Hz, 1 H;
H6a), 3.78 (dd, J5,6b = 7.0 Hz, 1 H; H6b), 3.74 (dd, J2Ј,3Ј = 9.9,
J3Ј,4Ј = 9.5 Hz, 1 H; H3Ј), 3.69 (dd, J5Ј,6bЈ = 5.5 Hz, 1 H; H6bЈ),
3.60 (dd, 1 H; H2Ј), 3.45 (m, 2 H; DNCH2CH2SD), 3.44 (dd, J4Ј,5Ј
= 9.8 Hz, 1 H; H4Ј), 3.26 (d, 2 H; H1a and H1b), 3.07 (t, 2 H;
DNCH2CH2SD). 13C NMR (125 MHz, D2O, HSQC): δ = 101.3
(C1Ј), 82.2 (C4), 73.5 (C3Ј), 73.3 (C5Ј), 73.2 (C5), 72.3 (C3), 72.2
(C2Ј), 70.1 (C4Ј), 67.7 (C2), 63.0 (C6), 61.1 (C6Ј), 51.0 (C1), 46.6
(NCH2), 33.1 (CH2S); High-resolution MS data of C14H29NO10S
(M = 403.151): M + Na found 426.150, calculated 426.141.
(1-Deoxy-D-mannityl)-(1ǞN)-cysteamine (18): Syrup (7.0 mg),
overall yield 74%. 1H NMR (500 MHz, D2O, 2D TOCSY): δ =
4.03 (m, 1 H; H2), 3.85 (dd, J5,6a = 3.4, J6a,6b = 11.8 Hz, 1 H; H6a),
3.67 (dd, J5,6b = 5.7 Hz, 1 H; H6b), 3.52 (m, 2 H; DNCH2CH2SD),
3.49 and 3.19 (2 m, 2 H; H1a and H1b), 3. 08 (t, 2 H;
DNCH2CH2SD). 13C NMR (125 MHz, D2O, HSQC): δ = 67.1
(C2), 63.7 (C6), 51.2 (C1), 46.7 (NCH2), 32.7 (CH2S); High-resolu-
tion MS data of C8H19NO5S (M = 241.098): M + Na found
264.097, calculated 264.088.
α-
D
-Mannopyranosyl-(1Ǟ3)-[(α-
D
-mannopyranosyl)-(1Ǟ6)]-(1-de-
oxy-
D
-mannityl)-(1ǞN)-cysteamine (19): White amorphous powder
(7 mg), overall yield 85%. 1H NMR (500 MHz, D2O, 2D TOCSY):
δ = 5.03 (d, J1ЈЈ,2ЈЈ = 1.9 Hz, 1 H; H1ЈЈ), 4.89 (d, J1Ј,2Ј = 1.9 Hz, 1
H; H1Ј), 4.16 (m, 1 H; H2), 4.04 (dd, J2ЈЈ,3ЈЈ = 3.1 Hz, 1 H; H2ЈЈ),
α-D-Glucopyranosyl-(1Ǟ4)-α-D-glucopyranosyl-(1Ǟ4)-(1-deoxy-D-
glucityl)-(1ǞN)-cysteamine (14): White amorphous powder (9 mg),
overall yield 73%. 1H NMR (500 MHz, D2O, 2D TOCSY): δ =
5.41 (d, J1ЈЈ,2ЈЈ = 3.8 Hz, 1 H; H1ЈЈ), 5.13 (d, J1Ј,2Ј = 3.8 Hz, 1 H;
H1Ј), 4.18 (m, 1 H; H2), 3.89–3.87 (m, 2 H; H3 and H4), 3.64 (dd,
J2Ј,3Ј = 10.1 Hz, 1 H; H2Ј), 3.60 (dd, J2ЈЈ,3ЈЈ = 9.9 Hz, 1 H; H2ЈЈ),
3.48 (m, 2 H; DNCH2CH2SD), 3.42 (br. t, J3ЈЈ,4ЈЈ = J4ЈЈ,5ЈЈ = 9.5 Hz,
1 H; H4ЈЈ), 3.29 (m, 2 H; H1a and H1b), 3.07 (t, 2 H;
DNCH2CH2SD). 13C NMR (125 MHz, D2O, HSQC): δ = 101.1
(C1Ј), 100.5 (C1ЈЈ), 82.3 (C4), 72.5 (C2ЈЈ), 72.4 (C3), 72.0 (C2Ј),
70.1 (C4ЈЈ), 67.7 (C2), 50.8 (C1), 46.6 (NCH2), 33.0 (CH2S); High-
resolution MS data of C20H40NO15S (M = 565.285): M + Na found
588.198, calculated 588.295.
3.99 (dd, J2Ј,3Ј = 3.5 Hz, 1 H; H2Ј), 3.56 (dd, J1a,2 = 2.6, J1a,1b
=
12.9 Hz, 1 H; H1a), 3.53 (m, 2 H; DNCH2CH2SD), 3.23 (dd, J1b,2
= 10.6 Hz, 1 H; H1b), 3.08 (t, 2 H; DNCH2CH2SD). 13C NMR
(125 MHz, D2O, HSQC): δ = 103.3 (C1ЈЈ), 100.7 (C1Ј), 70.9 (C2ЈЈ),
70.7 (C2Ј), 68.1 (C2), 50.4 (C1), 46.6 (NCH2), 32.6 (CH2S); High-
resolution MS data of C20H40NO15S (M = 543.222): M + Na found
566.212, calculated 566.212.
{α-
mannopyranosyl-(1Ǟ6)}-[(α-
-mannityl)-(1ǞN)-cysteamine (20): White amorphous powder
(4 mg), overall yield 69%. 1H NMR (500 MHz, D2O, 2D TOCSY):
δ = 5.13 (d, J1ЈЈЈЈ,2ЈЈЈЈ = 1.2 Hz, 1 H; H1ЈЈЈЈ), 5.04 (d, J1ЈЈЈ,2ЈЈЈ
0.6 Hz, 1 H; H1ЈЈЈ), 4.91 (d, J1ЈЈ,2ЈЈ = 1.7 Hz, 1 H; H1ЈЈ), 4.86 (d,
D
-Mannopyranosyl-(1Ǟ6)-[(α-
D-mannopyranosyl)-(1Ǟ3)]-α-D-
D
-mannopyranosyl)-(1Ǟ3)]-(1-deoxy-
D
α-
D-Glucopyranosyl-(1Ǟ4)-α-D-glucopyranosyl-(1Ǟ4)-α-
D-glucopy-
=
ranosyl-(1Ǟ4)-(1-deoxy-
D-glucityl)-(1ǞN)-cysteamine (15): White
Eur. J. Org. Chem. 2005, 3650–3659
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3657

J. P. Kamerling et al.
J1Ј,2Ј = 1.5 Hz, 1 H; H1Ј), 4.15 (br. t, 1 H; H2Ј), 4.17 (m, 1 H; H2), SPR Binding Assays: CM5 sensor surfaces were equilibrated with
FULL PAPER
4.07 (m, 1 H; H2ЈЈЈЈ), 4.04 (br. t, 1 H; H2ЈЈЈ), 3.99 (dd, J2ЈЈ,3ЈЈ
=
HEPES buffer (10 m HEPES pH 7.4, 0.15  NaCl), then acti-
vated with a 14-min pulse of a 1:1 mixture (v/v) of 0.1  N-hydroxy-
3.5 Hz, 1 H; H2ЈЈ), 3.56 and 3.26 (2 m, each 1 H; H1a and H1b),
3.51 (m, 2 H; DNCH2CH2SD), 3.07 (t, 2 H; DNCH2CH2SD). 13C succinimide and 0.1  N-ethyl-NЈ-(dimethylaminopropyl)carbodii-
NMR (125 MHz, D2O, HSQC): δ = 103.2 (C1ЈЈЈ), 103.1 (C1ЈЈЈЈ), mide, at 25 °C and a flow rate of 5 µL min–1. Ethanolamine hydro-
100.8 (C1Ј), 100.1 (C1ЈЈ), 70.7 (C2ЈЈЈ and C2ЈЈЈЈ), 70.6 (C2ЈЈ), 70.4
(C2Ј), 68.1 (C2), 50.6 (C1), 46.5 (NCH2), 32.6 (CH2S); High-resolu-
tion MS data of C32H59NO25S (M, 889.310): M + Na found
912.311, calculated 912.300.
chloride was immobilized on channel 1 via an injection of 7 min
(1.0 , pH 8.5; Ϸ 300 RU) to measure the level of non-specific
binding and to serve as blank channel for mathematical data treat-
ment. Con A was attached to channel 2 via an injection of 1 min
(50 µg mL–1 in 10 m sodium acetate buffer, pH 4.8; Ϸ 400 RU)
and to channel 3 via an injection of 7 min (100 µg mL–1 in 10 m
sodium acetate buffer, pH 4.8; Ϸ 7.000 RU); remaining N-hydroxy-
succinimide esters were blocked by a 7-min pulse of 1.0  ethanol-
amine hydrochloride, pH 8.5.
β-D-Galactopyranosyl-(1Ǟ4)-(1-deoxy-D-glucityl)-(1ǞN)-cyste-
amine (21): Slightly yellow powder (5 mg), overall yield 67%. 1H
NMR (500 MHz, D2O, 2D TOCSY): δ = 4.53 (d, J1Ј,2Ј = 7.8 Hz,
1 H; H1Ј), 4.23 (m, 1 H; H2), 3.93 (br. d, J3Ј,4Ј = 3.5, J4Ј,5Ј Ͻ 1 Hz,
1 H; H4Ј), 3.80 (dd, J2,3 = 2.0, J3,4 = 5.1 Hz, 1 H; H3), 3.67 (dd,
J2Ј,3Ј = 9.9 Hz, 1 H; H3Ј), 3.55 (dd, 1 H; H2Ј), 3.48 (m, 2 H;
GNPs Au-13 to Au-17 and Au-19 to Au-21 (100 µg mL–1, and di-
DNCH2CH2SD), 3.40 (dd, J1a,2 = 4.6, J1a,1b = 12.7 Hz, 1 H; H1a), lutions thereof) in HEPES buffer, containing 10 m CaCl2, were
3.18 (dd, J1b,2 = 10.1 Hz, 1 H; H1b), 3.07 (t, 2 H; DNCH2CH2SD). flowed across the sensor chip surfaces for 3 min at a flow rate of
13C NMR (125 MHz, D2O, HSQC): δ = 103.6 (C1Ј), 75.9 (C3), 15 µL min–1, and were allowed to dissociate for 5 min. To restore
73.1 (C3Ј), 71.7 (C2Ј), 71.6 (C4Ј), 68.4 (C2), 50.4 (C1), 46.7 the response level to zero after Au-13 to Au-17 injections, 150 m
(NCH2), 32.9 (CH2S); High-resolution MS data of C14H29NO10S
(M = 403.151): M + Na found 426.138, calculated 426.141.
methyl α--mannopyranoside and 3  NaCl were sequentially in-
jected for 5 min. After an Au-19 injection, the sensor chip surface
was regenerated by sequentially injecting 600 m methyl α--man-
nopyranoside and 3  NaCl for 5 min. Multiple (6x), 10 min long
sequential injections of 1.2  methyl α--mannopyranoside con-
taining 50 m EDTA and 1  NaCl were required to regenerate
the sensor chip surface after Au-20 was injected.
General Procedure for the Preparation of GNPs Au-13 to Au-17 and
Au-19 to Au-21: A solution of a cysteamine-extended saccharide
(12 to 21) in water (10 m, 1 equiv.) was added to a solution of
tetrachloroauric acid in water (25 m, 1 equiv.). Then, aq. NaBH4
(1 , 40 equiv.) was slowly added under vigorous stirring, and the
obtained red/brown solution was stirred for 2 h at room tempera-
ture. After neutralization with 30% aq. acetic acid, the solution
was diluted with water (10 mL), loaded on a 30 kDa Nalgene cen-
trifugal filter, and filtered. The residue was several times redissolved
in water and filtered (5 × 5 mL). After lyophilization from water,
the GNPs Au-13 to Au-17 and Au-19 to Au-21 were obtained as
brown, amorphous powders. The GNPs were characterized by
500 MHz 1H NMR spectroscopy in D2O, monosaccharide analysis,
and transmission electron microscopy (TEM).
Spectrophotometric Aggregation Assay: A solution (500 µL) of
GNPs Au-13 to Au-17 and Au-19 to Au-21 (100 µg mL–1) in
HEPES buffer (pH 7.4), containing 10 m CaCl2, and 50 µL of
Con A (1 mg mL–1), dissolved in the same HEPES buffer, were
mixed in a quartz cuvet. Spectrophotometric analysis was per-
formed by measuring, every 5 min for 6 h, the absorbance at
520 nm on a Hewlett–Packard 8452A diode array spectrophotome-
ter.
Transmission Electron Microscopy: Prior to the examination of the
aggregates formed between Au-17 and Con A, and Au-20 and Con
A (see spectrophotometric aggregation assay in experimental sec-
tion), the HEPES buffer, containing 10 m CaCl2, was removed by
centrifugal filtration through a 30 kDa eppendorf centrifugal filter
and replaced by water. Examinations were performed with a Philips
Tecnai12 microscope at 120 kV accelerating voltage. Aliquots
(1 µL) of GNPs Au-17 and Au-20 (0.1 mg mL–1), before and after
aggregation, were placed onto copper grids coated with a carbon
film (QUANTIFOIL on 200 square mesh copper grid, hole shape
R 2/2). The grids were left to dry at room temperature for several
hours. The particle size distribution of the GNPs was automatically
determined from several micrographs of the same sample, using
analySIS® 3.2 (Soft Imaging System GmbH).
Preparation of GNPs Au-14a and Au14b: The procedure is the same
as described above, only 2 equiv. or 0.33 equiv. of 14, as compared
to tetrachloroauric acid, were added for the preparation of GNPs
Au-14a and Au-14b, respectively.
Preparation of GNPs Au-17a: To a solution of 17 (0.1 mg, 95 nmol)
in water (50 µL) was added 12 µL aq. tetrachloroauric acid (25 n,
285 nmol). Then, 120 µL aq. NaBH4 (0.1 , 1.14 µmol) was slowly
added under ultrasonic stirring. The obtained red/brown solution
was sonicated for 5 min, then left for 2 h at room temperature. Af-
ter neutralization with 30% aq. acetic acid, the solution was diluted
with water (1 mL), loaded on a 30 kDa eppendorf centrifugal filter,
and filtered. The residue was several times redissolved in water and
filtered (5 × 1 mL). After lyophilization from water, the GNPs Au-
17a were obtained as a brown, amorphous powder. The GNPs were
characterized by 500 MHz 1H NMR spectroscopy in D2O and
transmission electron microscopy (TEM).
Acknowledgments
The financial support of the Academic Biomedical Center, Utrecht
University (Expertise Center for Carbohydrate Analysis and Syn-
thesis) is highly acknowledged (J. P. K., K. M. H. and A. C. S.). P.
Rohfritsch is kindly acknowledged for measuring the ES-MS spec-
tra. J. P. Meeldijk is kindly acknowledged for his assistance with
the TEM measurements.
Monosaccharide Analysis: After adding internal standard (manni-
tol), the GNP samples were subjected to methanolysis (1  meth-
anolic HCl, 24 h, 85 °C), followed by trimethylsilylation. The tri-
methylsilylated methyl glycosides were analyzed by GLC on an EC-
1 capillary column (30 m × 0.32 mm, Alltech) using a Chrompack
CP 9002 gas chromatograph (temperature program, 140–240 °C at
4 °C min–1). The identification of the monosaccharide derivatives
was confirmed by gas chromatography/mass spectrometry on a Fi-
sons Instruments GC 8060/MD 800 system (Interscience) equipped
with an AT-1 capillary column (30 m × 0.25 mm, Alltech), using
the same temperature program.[16]
[1] N. Perrimon, M. Bernfield, Nature 2000, 404, 725–728.
[2] P. R. Crocker, T. Feizi, Curr. Opin. Struct. Biol. 1996, 6, 679–
691.
[3] Y. C. Lee, R. T. Lee, Acc. Chem. Res. 1995, 28, 321–327.
3658
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2005, 3650–3659

Gold Glyconanoparticles in Carbohydrate-Protein Interaction Studies
FULL PAPER
[4] H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637–674.
[5] Y. C. Lee, FASEB J. 1993, 6, 3193–3200.
[6] B. Benaissa-Trouw, D. J. Lefeber, J. P. Kamerling, J. F. G. Vlieg-
enthart, H. Snippe, K. Kraaijeveld, Inf. Imm. 2001, 69, 4698–
4701.
[7] S. Andre, P. J. Ortega, M. A. Perez, R. Roy, H.-J. Gabius, Gly-
cobiology 1999, 9, 1253–1261.
[13] A. G. Barrientos, J. M. de la Fuente, T. C. Rojas, A. Fernández,
S. Penadés, Chem. Eur. J. 2003, 9, 1909–1921.
[14] A. Carvalho de Souza, K. M. Halkes, J. D. Meeldijk, A. J. Ver-
kleij, J. F. G. Vliegenthart, J. P. Kamerling, ChemBioChem
2005, 6, 828–831.
[15] M. Brust, M. Walker, D. Berthell, D. J. Schiffrin, R. Whyman,
J. Chem. Soc. Chem. Commun. 1994, 801–802.
[8] K. Matsuura, H. Kitakouji, N. Sawada, H. Ishida, M. Kiso,
K. Kitajima, K. Kobayashi, J. Am. Chem. Soc. 2000, 122,
7406–7407.
[9] J. M. de la Fuente, A. G. Barrientos, T. C. Rojas, J. Roja, A.
Fernández, S. Penadés, Angew. Chem. Int. Ed. 2001, 40, 2257–
2261.
[10] A. Carvalho de Souza, K. M. Halkes, J. D. Meeldijk, A. J. Ver-
kleij, J. F. G. Vliegenthart, J. P. Kamerling, Eur. J. Org. Chem.
2004, 4323–4339.
[16] J. P. Kamerling, J. F. G. Vliegenthart, in Clinical Biochemistry –
Principles, Methods, Applications 1. Mass Spectrometry (Ed.:
A. M. Lawson), Walter de Gruyter, Berlin, 1989, 176–263.
[17] M. J. Hostetler, J. E. Wingate, C.-J. Zhong, J. E. Harris, R. W.
Vachet, M. R. Clark, J. D. Londono, S. J. Green, J. J. Stokes,
G. D. Wignall, G. L. Glish, M. D. Porter, N. D. Evans, R. W.
Murray, Langmuir 1998, 14, 17–30.
[18] H. Kaku, I. J. Goldstein, Carbohydr. Res. 1992, 229, 337–346.
[19] G. Mie, Ann. Phys. 1908, 25, 377.
[11] H. Otsuka, Y. Akiyama, Y. Nagasaki, K. Kataoka, J. Am.
Chem. Soc. 2001, 123, 8226–8230.
[12] 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.
[20] D. C. Hone, A. H. Haines, D. A. Russell, Langmuir 2003, 19,
7141–7144.
Received: April 12, 2005
Published Online: July 12, 2005
Eur. J. Org. Chem. 2005, 3650–3659
www.eurjoc.org
© 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3659