Chemoselective capture of glycans for analysis on gold nanoparticles: Carbohydrate oxime tautomers provide functional recognition by proteins

Article abstract of DOI:10.1002/chem.200801521

Nanoparticles functionalized with glycans are emerging as powerful solid-phase chemical tools for the study of protein-carbohydrate interactions using nanoscale properties for detection of binding events. Methods or reagents that enable the assembly of glyconanoparticles from unprotected glycans in two consecutive chemoselective steps with meaningful display of the glycan are highly desirable. Here, we describe a novel bifunctional reagent that 1) couples to glycans by oxime formation in solution, 2) aids in purification through a lipophilic trityl tag, and 3) after deprotection then couples to gold nanoparticles through a thiol. NMR studies revealed that these oximes exist as both the open-chain and N-glycosyl oxy-amine tautomers. Glycan-linker conjugates were coupled through displacement of ligands from preformed, citrate-stabilized gold nano-particles. Recognition of these glycans by proteins was studied with a lectin, concanavalin A (ConA), in an aggregation assay and with a processing enzyme and glucoamylase (GA). We demonstrate that the presence of the N-glycosyl oxy-amines clearly enables functional recognition in sharp contrast to the corresponding reduced oxy-amines. This concept is then realized in a novel reagent, which should facilitate nano-glycobiology by enabling the operationally simple capture of glycans and their biologically meaningful display.

Full text of DOI:10.1002/chem.200801521

FULL PAPER
DOI: 10.1002/chem.200801521
Chemoselective Capture of Glycans for Analysis on Gold Nanoparticles:
Carbohydrate Oxime Tautomers Provide Functional Recognition by Proteins
Mikkel B. Thygesen, Jørgen Sauer, and Knud J. Jensen*[a]
Abstract: Nanoparticles functionalized
with glycans are emerging as powerful
solid-phase chemical tools for the study
of protein–carbohydrate interactions
using nanoscale properties for detec-
tion of binding events. Methods or re-
agents that enable the assembly of gly-
conanoparticles from unprotected gly-
cans in two consecutive chemoselective
steps with meaningful display of the
glycan are highly desirable. Here, we
describe a novel bifunctional reagent
that 1) couples to glycans by oxime for-
mation in solution, 2) aids in purifica-
tion through a lipophilic trityl tag, and
3) after deprotection then couples to
particles. Recognition of these glycans
by proteins was studied with a lectin,
concanavalin A (ConA), in an aggrega-
tion assay and with a processing enzyme
and glucoamylase (GA). We demon-
strate that the presence of the N-glyco-
syl oxy-amines clearly enables func-
tional recognition in sharp contrast to
the corresponding reduced oxy-amines.
This concept is then realized in a novel
reagent, which should facilitate nano-
glycobiology by enabling the opera-
tionally simple capture of glycans and
their biologically meaningful display.
gold nanoparticles through
a thiol.
NMR studies revealed that these
oximes exist as both the open-chain
and N-glycosyl oxy-amine tautomers.
Glycan-linker conjugates were coupled
through displacement of ligands from
preformed, citrate-stabilized gold nano-
Keywords: glycoconjugates · gold ·
molecular
recognition
·
nanotechnology · proteins
Introduction
example, reductive amination, hydrazone and acyl hydra-
zone formation, and thiazolidine formation.[13] With the
recent advances in the preparation of complex oligosaccha-
rides of biological relevance through chemical and chemo-
enzymatic methods,[14] researchers have increasing access to
glycans in low quantities related to, for example, studies on
signaling, pathogenic and viral infections, fertility, immunity,
and cancer.[15] The development of tools for the study of
these interactions, providing functional recognition of carbo-
hydrates by proteins and other biomolecules, constitutes an
important basis for glycobiology and glycomics, that is, the
high-throughput study of the biological roles of carbohy-
drates in living organisms.
A central solid-phase chemical tool is represented by gly-
conanoparticles, which have recently attracted attention as
probes for the study of carbohydrate–protein and carbohy-
drate–carbohydrate interactions.[4,16] These metal and semi-
conductor nanoparticles covered with monolayers of carbo-
hydrates offer some unique advantages over other multiva-
lent carbohydrate-functionalized structures (e.g., dendrim-
ers, polymers, liposomes) in the combined properties of the
clustered multivalent presentation of the carbohydrates on
the surface, the possibility for facile tracing by means of
electron microscopy, light scattering, or fluorescence emis-
sion, the relatively small size that allows in vivo bio-imaging,
Glycans are central to many fundamental biological process-
es, such as, cell–cell recognition, detection, and evasion, rais-
ing of immune responses, cell attachment, cell fate, develop-
ment, and morphogenesis.[1–3] Solid-phase chemical methods
are gaining importance as a tool enabling the study of these
processes.[4]
Chemoselective coupling of unprotected carbohydrates to,
for example, polymeric supports,[5–7] microarrays,[8–10] pep-
tides and proteins,[11] and to other surfaces[12] through oxime
formation has been widely exploited in recent years as a
means to capture glycans from natural sources and to obvi-
ate laborious glycosylation and protecting-group chemistry.
The reaction exploits the unique availability of an aldehyde
moiety in the reducing-end of glycans, which also allows, for
[a] M. B. Thygesen, Dr. J. Sauer, Prof. K. J. Jensen
Department of Natural Sciences, Faculty of Life Sciences
Centre for Carbohydrate Recognition and Signalling
University of Copenhagen, Thorvaldsensvej 40
1871 Frederiksberg (Denmark)
Fax : (+45)353-32-398
E-mail: kjj@life.ku.dk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801521.
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1649

and the high degree of solubility and stability in buffered
aqueous media. The three-dimensional presentation on the
spherical nanoparticles results in a flexible globular shape
that can be thought of as a model for the glycocalyx of cell
surfaces.[17] Of particular importance are the optical proper-
ties of gold nanoparticles (AuNPs), notably the changes in
surface-plasmon resonance upon association, which have
been exploited in applications of gold glyconanoparticles
(glyco-AuNPs) for the in vitro detection of, for example, car-
bohydrate–protein interactions. The rapid developments in
this field have to a great extent been fueled by the introduc-
tion of the Brust–Schiffrin two-phase protocol for the prepa-
ration of functionalized AuNPs,[18] in which ligand-capped
AuNPs are produced in a one-step approach. Alternative
methods for AuNP preparation have been available for sev-
eral decades, for example, the methods of Turkevich[19] and
Frens,[20] in which a two-step procedure is typically used,
that is, AuNP preparation followed by the introduction of
the desired ligands. The highly polar ligand shell of glyco-
AuNPs represents a challenge to AuNP preparation by the
Brust–Schiffrin protocol due to the incompatibility of the
polar shell with organic solvents. In typical modifications of
the procedure, the AuNPs are prepared in polar organic sol-
vents such as methanol in a homogenous phase.[21] A consid-
erable limitation of these one-step procedures for compara-
tive (parallel) studies is the high degree of dependency of
particle growth on the nature of the thiol ligands.[17,22,23] This
may be circumvented by use of two-step procedures, in
which the initial preparation of the AuNP cores is followed
by introduction of carbohydrate ligands, hereby allowing the
use of single batches of AuNPs for multiple ligands.[24] Both
types of procedure, however, require access to thiol-func-
tionalized glycans.
drate oximes appears to be recognized. Shin and co-workers
have reported interactions of cyclic tautomer forms of car-
bohydrate acylhydrazones with proteins.[8] Additionally, N-
methyl substituted aminooxy derivatives have been pur-
sued.[9,28]
Carbohydrate oximes and oxime ethers are known to
exist in several ring-chain tautomeric forms,[29] of which the
open-chain E and Z forms generally predominate over the
pyranose b and a forms,[30–32] as shown for d-glucose in
Scheme 1. Additionally, furanose b and a forms may consti-
tute minor tautomers (not shown). Isomerization of the
open-chain oximes is mediated by the closed forms. Howev-
er, the distribution of tautomer forms is dependent on the
monosaccharide unit.
Scheme 1. Ring-chain tautomers of glucose oxime ethers. Adapted from
ref. [30].
In this paper, we describe the use of carbohydrate oxime
formation for the attachment of glycans to gold nanoparti-
cles. To our knowledge, this is the first example of such use.
Moreover, we study the consequences of tautomerism of
carbohydrate oxime ethers during interaction with proteins
by comparing the functional binding of oxime ring-chain
tautomers with analogues “locked” in an open-chain config-
uration by reduction of the oximes to hydroxylamines. For
this purpose, we chose the hydrolytic enzyme glucoamylase
(GA) from Aspergillus niger and the lectin concanavalin A
(ConA) from Concanavalia ensiformis.
Attachment of carbohydrates to AuNPs through thiol-
containing linkers has been performed typically by extensive
protecting-group and glycosylation chemistry in solution.
This approach for the preparation of glyconanoparticles was
introduced by Penadꢁs and co-workers[17,21] to present glu-
cose, lactose, maltose, and the Lewisx trisaccharide on
AuNPs. A similar approach has been taken by the groups of
Russell,[24] Lin,[25] Barchi,[26] and others.
Reductive amination chemistry for glyco-AuNP prepara-
tion was introduced by Kamerling and co-workers for the
thiol-functionalization of unprotected glycans.[22] In this pro-
cedure, the classical method of conversion of the reducing-
end saccharide unit into an open-chain secondary amine by
reduction was used.[27] However, the lack of specific binding
between proteins and glycans attached by means of reduc-
tive amination was highlighted recently by Feizi and co-
workers.[10] Here, lectin and antibody affinity for short gly-
cans, for example, Lea and Lex trisaccharides, was shown to
be dependent on the presence of an unmodified (i.e., pyra-
nose) core-monosaccharide unit. Glycans attached through
oxime formation could be recognized, whereas reductive
amination resulted in complete loss of affinity, demonstrat-
ing that at least for some carbohydrate–protein interactions,
the ring-closed form (i.e., the N-glycoside) of the carbohy-
Results and Discussion
The linker moiety between glycan and nanoparticle is of
crucial importance for interactions of glyco-AuNPs with pro-
teins. The length and nature of the linker between the re-
ducing end of the saccharide and the mercapto group (for
attachment to the gold surface by formation of a covalent
ꢀ
Au S bond) affects the ligand density, but may also play a
role in the organization of the monolayer and on the solubil-
ity and bio-compatibility of the resulting nanoparticles. The
use of well-defined poly(ethylene glycol) (PEG) polymers
and monodisperse oligo(ethylene glycol)s (OEGs) has
emerged as a means to improve biological and pharmaceuti-
cal properties of, for example, proteins[33] and drug mole-
cules[34] and these are also widely used as linkers for surfaces
including nanoparticles. Among the reported effects of
1650
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660

Glyconanoparticles
FULL PAPER
PEGs and OEGs are, for example, prevention of degrada-
tion by hydrolytic and proteolytic enzymes, shielding of anti-
genic or immunogenic epitopes, improvement of aqueous
solubility, reduction of non-specific binding, and prolonga-
tion of systemic circulation.[33,34]
Alkanethiolate-linked OEGs, that is,
RACHTGNUTER(NNUG CH2CH2O)n-
ACHTUNGTRENNUNG(CH2)mSH with n typically around 4 and m around 11, intro-
duced by Whitesides and co-workers,[35] have been used on
gold nanoparticles to produce aqueously soluble particles
with increased non-specific protein resistance.[36,37] Alterna-
tively, the aliphatic chain may be omitted with similiar re-
sults by the use of thiol-terminated OEGs, that is,
Scheme 2. Preparation of heterobifunctional aminooxy linker 3 from tet-
ra(ethylene glycol). a) 1 equiv PhthN-OH, PS-PPh3, diisopropyl diazodi-
carboxylate (DIAD), 16 h; b) TrSH, PS-PPh3, DIAD, 08C!RT, 16 h;
c) hydrazine hydrate, MeCN, RT, 2 h.
RCH2CH2ACHTUNGTRENNUNG(OCH2CH2)nSH with n typically around 4, which
are more easily obtained.[38–41] The differences in the assem-
bly and ordering of monolayers between these two types of
OEG ligands are not clearly understood, however, the OEG
section of the monolayers is generally believed to contain
more random orientations of individual chains than the
close packing of alkanethiolate monolayers.[39,42]
opening and formation of mixed hydroxamic/carboxylic
esters (data not shown). Following the isolation of 1, a
second Mitsunobu reaction was performed with triphenyl-
methanethiol to obtain the orthogonally protected heterobi-
functional linker 2 in 78% yield. The Mitsunobu reaction
with triphenylmethanethiol, reported initially by Jin et al.[44]
in 2005, is a highly convenient method for the “thiolation”
of OEGs that obviates the use of volatile sulfur compounds
typically used in alkylation protocols. Purification of 1 and 2
was simplified by the use of polymer-supported triphenyl-
phosphine that could be removed by simple filtration.
Removal of the labile Phth-protecting group from hetero-
bifunctional linker 2 with hydrazine hydrate at room tem-
perature provided cleanly the aminooxy linker 3, which was
coupled to glucose, maltose, and maltotriose by treatment
with unprotected glycans under mildly acidic conditions.
The resulting carbohydrate oxime ethers 4–6 were isolated
in good yields (74–88%, Scheme 3). A key feature of the
design of linker 3 is that upon formation of glycan-linker
conjugates, it allows easy detection and isolation on silica
gel or by reverse-phase chromatography (HPLC or solid-
phase extraction, SPE) due to the presence of the lipophilic
and UV-active trityl group.
NMR analyses of compounds 4–6 in [D4]methanol al-
lowed the identification of (E)- and (Z)-oxime tautomers as
well as the cyclic ring-chain tautomers as minor constituents.
HSQC data showed that the cyclic ring-chain tautomers had
JH1,H2 coupling constants in the range 9.1–9.2 Hz, thus indi-
cating that they were in the b-pyranose configuration. The
estimated (E)-oxime/(Z)-oxime/b-pyranose ring-chain tauto-
mer ratio of 60:20:20 was identical for compounds 4–6. Di-
agnostic NMR signals for glucose oxime 4 tautomers are
provided in Table 1. These assignments were nearly identical
for maltose and maltotriose oximes 5 and 6 (see Experimen-
tal Section).
The optimal length of OEG linkers for bioanalytical
assays is a compromise between several factors, such as the
flexibility and accessibility of the ligands, the solubility and
stability of the glyco-AuNPs, and a maximum assay re-
sponse, for example, in surface-plasmon-resonance measure-
ments by using UV-visible light. Short linkers may provide
an increased assay sensitivity[24] but may also lead to poor
solubility.[22] Additionally, Li and Peng have shown that the
photochemical stability of AuNPs is dependent on the
length of the linker;[43] long alkanethiolate linkers provide
improved stability over shorter linkers. In light of these con-
siderations, we chose to work with OEG linkers consisting
of thiol-terminated tetra(ethylene glycol).
We have designed and synthesized a novel OEG linker
that allows two sequential highly chemoselective reactions:
1) attachment of a glycan to the linker via carbohydrate
oxime formation, and 2) anchoring of the linker to the parti-
cle surface. The linker incorporates a removable trityl
moiety to aid in isolation of the glycan-linker conjugates.
This approach provides easy-to-handle yet efficient coupling
of biomolecules to nanoparticles to generate functionalized
nanoparticles for bioanalytical assays.
Preparation of OEG glycoconjugates: Mitsunobu chemistry
was used to introduce N-hydroxyphthalimide (PhthN-OH)
on tetra(ethylene glycol), see Scheme 2. The de-symmetriza-
tion of the OEG was performed with stoichiometric
amounts of the nucleophile, thus the theoretical yield would
be 50%. Examination of the reaction mixture indicated
complete consumption of N-hydroxyphthalimide and a 1/
bis-1 ratio of 2:1 as expected, however, the isolated yield of
1 was 37% (74% when corrected for statistical distribution).
Recovery of the bis-substituted product (bis-1) provided a
yield of 18%, indicating that product was lost during the pu-
rification step, possibly by covalent attachment to silica; we
note that, contrary to N-alkylphthalimides, the N-alkoxyph-
thalimide moiety is highly reactive and that addition of alco-
hols such as methanol to solutions of 1 result in partial ring
An investigation of the possible pH dependency of the
ring-chain tautomeric equilibrium was conducted with a 13C-
labeled analogue of 4, prepared from d-[1-13C]glucose and 3.
In [D4]methanol, this compound (16) had an (E)-oxime/(Z)-
oxime/b-pyranose ring-chain tautomer ratio of 60:20:20. The
pH of the solution ([D4]methanol/H2O 4:6) was controlled
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1651

K. J. Jensen et al.
of a proton from Glu179 to the
scissile glycosidic bond and a
nucleophilic attack by water on
C-1 in the transition-state oxo-
carbenium ion, catalyzed by
Glu400.[52–57] The two catalytic
residues protrude in the bottom
of a funnel-shaped passage cre-
ated by six highly conserved re-
gions that constitute the GA-
characteristic
(a/a)6-barrel
domain.[52,54,55,58] These a-helical
segments form the major part
of an active-site pocket that is
believed to contain seven sub-
sites that can accommodate one
glycosyl residue each.[59–61] The
Scheme 3. Preparation of OEG glycoconjugates 4–6 by oxime coupling and subsequent reduction to oxy-
amines 7–9. Removal of the trityl groups furnished thiol glycoconjugates 10–15 for coupling with gold nano-
particles. Carbohydrate oximes are depicted as the b-pyranose tautomer for the sake of clarity (Scheme 1).
a) Glucose/maltose/maltotriose (1.1 equiv), MeCN, H2O, HOAc, 16 h; b) NaBH3CN (2 equiv), HOAc, 2 h; hydrolytic event occurs be-
c) TFA, triethylsilane (TES), CH2Cl2, 5 min.
tween subsites ꢀ1 and 1 and re-
quires the substrate to bind in a
productive mode in these two
Table 1. Diagnostic NMR signals of tautomers of glucose oxime 4.
terminal sites, thereby exposing the glycosidic linkage to the
catalytic residues. The affinities of subsites ꢀ1 and 1 esti-
mated from pre-steady-state association constants indicates
that the initial binding of substrate is controlled by a very
high affinity in subsite ꢀ1 with subsite 1 having much lower
affinity.[62] A combination of site-directed mutagenesis, dif-
ferent dissacharide substrates, and inhibitors have been used
for mapping the affinity of these two subsites and their role
in hydrolysis.[52,54,63–66] One of these inhibitors/substrates is
maltitol (4-O-a-d-glucopyranosyl-d-glucitol), the reduced
form of maltose, that is hydrolyzed only slowly by GA de-
spite the fact that the a-1,4-glycosidic bond is retained after
reduction of maltose. A possible explanation for this has
been suggested by Gꢂnther and Heymann in a study of sub-
sites in pig intestinal glucoamylase–maltase complex[67]
based on the reaction mechanism originally proposed for
glycosidases.[68] According to this model, the reaction rate
depends on the optimal orientation of the glycosidic oxygen
atom relative to the active-site acidic residue. The orienta-
tion of this atom is a function of the glycosidic bond and the
ꢃaglyconꢄ residue. Thus, for optimal hydrolysis, a glycosidic
a-1,4 configuration and an ꢄaglyconꢄ glucopyranosyl ring
structure is necessary. In substrates with other geometric
configurations, for example, open-chain, the initial protona-
tion of the glycosidic oxygen by the catalytic acid is less fa-
vorable.
Tautomer[a]
H1[b]/C1 [ppm]
H2/C2 [ppm]
H3/C3 [ppm]
(E)-oxime
(Z)-oxime
b-pyranose
7.44 (6.7)/151.8
6.78 (6.0)/154.2
4.11 (9.2)/92.9
4.32/72.2
4.94/68.6
3.20/79.2
3.86/72.6
3.90/74.9
N.A.
[a] See Scheme 1. [b] Values in parenthesis denote JH1,H2 coupling con-
stants in Hz.
by acetate buffering at pH 7, 6, 5, 4, 3, and 2, and the rela-
tive ratio of ring-chain tautomers was monitored by record-
ing the anomeric signals in 13C NMR spectroscopy (see Sup-
porting Information). We found that the tautomer ratio was
essentially identical down to pH 2. A related study on car-
bohydrate acetylhydrazones has indicated a similar reduced
basicity of the anomeric nitrogen.[45]
The carbohydrate oximes 4–6 were reduced by treatment
with sodium cyanoborohydride employing acetic acid as the
solvent. The resulting oxy-amines 7–9 were isolated by ion-
exchange chromatography (SPE cartridges) in high yields
(85–92%, Scheme 3). NMR spectroscopy confirmed the
conversion of all ring-chain tautomers of 4–6 into single spe-
cies for 7–9.
Functional recognition of oxime tautomers of OEG glyco-
conjugates by glucoamylase: An assay based on glucoamy-
lase (GA; 1,4-a-d-glucan glucohydrolase, EC 3.2.1.3) from
Aspergillus niger was selected for assessment of the func-
tional recognition of OEG glycoconjugates 4–9 by hydrolyt-
ic enzymes. GA catalyzes the hydrolysis of a-1,4 and to a
lesser extent a-1,6 glucosidic linkages from the non-reducing
end of starch and related glycans to release d-glucose.[46–49]
The active site of GA contains two glutamic acids posi-
tioned at positions 179 and 400 of the protein sequence
acting as a general acid catalyst and base, respectively.[50,51]
The hydrolysis of substrate is accomplished by the donation
We hypothesized that the ring-chain tautomeric equilibri-
um of 4–6 would result in functional recognition of the re-
ducing-end residue, which would not be possible with the
open-chain configuration in 6–9. Indeed, incubation with
GA in 50 mm acetate buffer pH 4.5 at 428C for 20 h fol-
lowed by HPLC analysis showed full conversion of both 5
and 6 into 4, that is, one glucose unit was cleaved from the
maltose oxime and two glucose units were cleaved from the
maltotriose oxime. For carbohydrate oxy-amines 6–9, how-
ever, 9 was converted into 8 as expected, that is, one glucose
1652
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660

Glyconanoparticles
FULL PAPER
unit was cleaved from the maltotriose oxy-amine, but there
was only minor conversion of 8 and 9 to 7, that is, the mal-
tose oxy-amine was not a good substrate for GA (Figure 1).
This result is in agreement with the different substrate spe-
cificities of GA towards maltose and maltitol, as described
above. This leads to the very important conclusion that the
carbohydrate oxime allows functional recognition by gluco-
amylase of ring-closed (N-glycosyl oxy-amine) tautomers.
nmol-substrate scale (Table 2). These results clearly demon-
strate that for carbohydrate oximes 4–6 the maltose and
maltotriose derivatives are hydrolyzed to the glucose stage,
whereas for carbohydrate oxy-amines 7–9 (locked open-
chain) only the maltotriose derivative is a substrate for GA
and that hydrolysis stops at the maltose stage.
As a control experiment, we wanted to study whether the
enzymatic cleavage of 5 to 4 by GA could occur by means
of an unexpected oxime cleavage in acetate buffer. Specifi-
cally, we wanted to study, and exclude, the possibility that
hydrolysis of 5 to 4 proceeded by means of 1) oxime hydrol-
ysis, followed by 2) cleavage of maltose in solution, and fi-
nally by 3) reformation of the oxime between glucose and
linker 3. Two exchange studies were performed in 50 mm
acetate buffer pH 4.5 at 428C for 20 h in which 1) 5 was in-
cubated with an equimolar amount of glucose in the absence
of GA, and 2) 4 was incubated with an equimolar amount of
[1-13C]glucose in the presence of GA. Analysis by mass
spectrometry showed the absence of the putative exchange
product in both cases. These results strongly demonstrated
that under these conditions, the oxime linkage remained
stable and that the glycan of oxime 5 was recognized by
GA, most likely as the ring-closed, N-glycosidic tautomer.
Preparation of glyconanoparticles: OEG glycoconjugates 4–
9 were deprotected by treatment with trifluoroacetic acid
under anhydrous conditions and in the presence of triethyl-
silane as a scavenger to furnish the thiols 10–15 in quantita-
tive yields (Scheme 3). Variable amounts (0–40%) of the
corresponding disulfides were observed by NMR spectrosco-
py.
Citrate-capped gold nanoparticles were prepared accord-
ing to a modification[69] of the Turkevich protocol[19] to pro-
vide spherical AuNPs with an average diameter of 12 nm
and a concentration of 18 nm, as determined by transmission
electron microscopy (TEM) and UV-visible spectroscopy.[70]
The glycan-linker conjugates 10–15 were anchored to the
AuNPs through displacement of the citrate by stirring in
aqueous media for a minimum of 16 h (Scheme 4). Gold gly-
conanoparticles (10-AuNP to 15-AuNP) were purified by
repetitive (5ꢅ) centrifugal filtration (50 kDa cut-off) of
aqueous washings, and redissolved to final concentrations of
0.50 mm in 10 mm phosphate buffer pH 7.0. A slight red-shift
of ꢁ4 nm was observed in the UV-visible spectrum upon
coating with thiols 10–15. All glyconanoparticles had a sur-
face-plasmon-resonance peak maximum at 523 nm in aque-
ous media, and solutions were stable for at least six months,
as witnessed by the lack of changes in their UV-visible spec-
tra during this period. Analysis by TEM confirmed that
there were no changes in AuNP diameter upon treatment
with the thiol-containing glycoconjugates (Supporting Infor-
mation).
Figure 1. HPLC chromatograms (215 nm) of OEG glycoconjugates
before (top three rows) and after (bottom three rows) incubation with
glucoamylase (+GA) in 50 mm sodium acetate buffer pH 4.5 at 428C for
20 h. a) Oximes 4–6; b) Oxy-amines 7–9.
Next, we applied a quantitative colorimetric assay based
on the coupling of the GA assay with a glucose oxidase/per-
oxidase assay. If performed in acetate buffer pH 4.5 at 428C
for 20 h, incubation of the glucose oxime 4 showed a small
signal, indicative of partial hydrolysis (<10%) of the oxime
bond under these assay conditions. Accordingly, we changed
the assay conditions to 10 mm phosphate buffer pH 7.0 at
428C for 20 h, which resulted in a drastic reduction of non-
enzymatic oxime hydrolysis of 4–6, however, we note that
pH 7 is suboptimal for glucoamylase.[52] The glucoamylase–
glucose oxidase/peroxidase assay was conducted on a 10-
Glyconanoparticles 10-AuNP to 15-AuNP were freely
soluble in 10 mm phosphate buffer at pH 7.0, however, rapid
aggregation of 13-AuNP, 14-AuNP, and 15-AuNP, which
contained the reduced oximes, was observed in 50 mm ace-
tate buffer at pH 4.5. On the other hand, under these condi-
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1653

K. J. Jensen et al.
Table 2. Quantitative measurements of release of glucose from free and captured glycans (4–9) by glucoamylase.[a]
Glycan
Glucose
Maltose
Maltotriose
unmodified oxime (4) oxy-amine (7) unmodified oxime (5) oxy-amine (8) unmodified oxime (6) oxy-amine (9)
released glucose [nmol]
9.14ꢂ0.6 0.17ꢂ0.08 0.16ꢂ0.07
19.73ꢂ0.2
8.00ꢂ0.3
0.72ꢂ0.2
30.11ꢂ1.5 17.99ꢂ0.4 8.73ꢂ0.4
[a] Hydrolysis was performed in 10 mm sodium phosphate pH 7.0 at 428C using 10 nmol of each compound. The exact points of hydrolysis are illustrated
on the graphics by dashed lines. Carbohydrate oximes are depicted as the b-pyranose tautomer for the sake of clarity (Scheme 1).
impact on the binding to carbohydrate-binding proteins. For
this purpose we chose the lectin concanavalin A (ConA)
from Canavalia ensiformis. This lectin has an affinity for ter-
minal a-d-glucopyranose residues as well as for a-d-manno-
pyranose residues, and interactions of this lectin with glyco-
nanoparticles have been studied by several groups.[22,24,25]
Comparison of relative binding affinities of methyl a-gluco-
side, maltose, maltitol, and maltotriose in the literature[72,73]
show that these glycans bind to ConA with identical affinity
in the low-millimolar range, consistent with a single-residue
binding site,[74] whereas the affinity of b-glucosides is low-
ered by an order of magnitude.
Glyconanoparticles (5 nm) containing maltotriose as the
Scheme 4. Preparation of glyconanoparticles 10-AuNP to 15-AuNP. Car-
oxime and oxy-amine (12-AuNP and 15-AuNP, respective-
bohydrate oximes are depicted as the b-pyranose tautomer for the sake
ly), as well as glucose as the oxime and oxy-amine (10-
of clarity (Scheme 1).
AuNP and 13-AuNP, respectively) were subjected to a
ConA concentration of 5 mm, and lectin-mediated glycona-
noparticle aggregation was observed as a shift in the sur-
tions 10-AuNP, 11-AuNP, and 12-AuNP, containing carbohy-
drate oximes, remained freely soluble. We attribute this
effect to protonation of the oxy-amines in 13-AuNP, 14-
AuNP, and 15-AuNP, the pKa values of which are expected
to be in the 4–5 range (based on the pKa of N,O-dimethyl
hydroxylamine of 4.75[71]). We speculate that electrostatic in-
teractions between protonated oxy-amines and negative
charges associated with the gold surface led to the observed
aggregation under acidic conditions.
face-plasmon band that could be measured spectrophoto-
metrically (Figure 2). Our results clearly show that both
maltotriose-containing AuNPs underwent extensive lectin-
mediated aggregation, but more importantly, that unlike the
lectin-induced aggregation with the glucose oxime AuNPs
(10-AuNP), the glucose oxy-amine AuNPs (13-AuNP) dis-
played a complete lack of surface-plasmon band shift upon
addition of the lectin. We rationalize this striking difference
in lectin-induced aggregation as a consequence of the differ-
ent glucose structures presented to the lectin by 10-AuNP
and 13-AuNP: the former, which includes a ring-closed, N-
glycosidic tautomer, could be recognized by ConA, whereas
the latter, with its acyclic, open-chain form, could not. This
result also shows that the specificity of the lectin binding
Biological recognition of oxime tautomers on glyco-AuNPs
by the lectin concanavalin A: With the novel glyconanopar-
ticles at hand, we sought to establish whether the ring-chain
tautomerism of the carbohydrate oximes would have an
1654
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660

Glyconanoparticles
FULL PAPER
plays the dominant role and that unspecific protein interac-
tions are minimal.
release of 2 mol glucose/mol maltotriose oxime. For the oxy-
amine AuNPs, however, the ligand density was approximate-
ly three times higher, reaching around 1000 mol ligand/mol
AuNPs for the maltotriose oxy-amine AuNPs (15-AuNP).
We hypothesize that this difference in ligand density be-
tween oxime glyco-AuNPs and oxy-amine glyco-AuNPs
could be due to a steric demand of the dynamic ring-chain
tautomerism of glycoconjugates 10–12 relative to the open-
chain glycoconjugates 13–15.
At lower ConA concentration (1 mm), the glucose oxime
AuNPs (10-AuNP, ring-form predominantly b-pyranoside)
displayed less of a shift of the surface-plasmon band than
seen for 11-AuNP and 12-AuNP (Supporting Information),
consistent with the higher binding of a-glucosides, as de-
scribed above. Control experiments with bovine serum albu-
min (BSA) at 5 mm showed complete absence of a surface-
plasmon band shift with all glyconanoparticles (10-AuNP
through 15-AuNP) throughout the 48 h period (see Support-
ing Information for 10-AuNP), confirming the insignificance
of non-specific interactions and the specificity of the glyco-
nanoparticle–lectin recognitions.
It seems from the results in Figure 2 that the onset of ag-
gregation was slower for the maltotriose oxy-amine AuNPs
(15-AuNP) than for the corresponding oxime AuNPs (12-
AuNP). This effect, given the identical nanoparticle size of
the two samples and supposing identical binding affinities of
the individual maltotriose conjugates for ConA, could be re-
lated to a difference in ligand density of these glyco-AuNPs.
Figure 3. Measurements of release of glucose by glucoamylase after thio-
lytic release of OEG glycoconjugates from glyco-AuNPs (white bars) and
by glucoamylase on glyco-AuNPs (grey bars). The experiment was per-
formed in 10 mm sodium phosphate pH 7.0 at 428C using 50 pmol of
each glyco-AuNP. As expected, 14-AuNP showed no response after thio-
lytic release (Table 2).
Functional recognition of glyco-AuNPs by glucoamylase: Fi-
nally, we applied the coupled assay based on glucoamylase
(GA) and glucose oxidase/peroxidase, as described above,
for estimating the accessibility of the glycan ligands at the
surface. Taking into account the topology of the barrel-
shaped active site, we reasoned that an ability of GA to re-
lease glucose from the surface of the glyconanoparticles
would be indicative of a high degree of accessibility. At the
outset of this study, we noted that Penadꢁs and co-workers
previously examined the enzymatic cleavage of galactose
from lactose-AuNPs by a b-galactosidase and found only
minute activity of the enzyme (<3% cleaved).[21]
Figure 2. Lectin-induced nanoparticle aggregation as monitored by a shift
*
in surface-plasmon band maximum for glyco-AuNPs 10-AuNP ( ), 12-
~
~
*
AuNP ( ), 13-AuNP ( ), and 15-AuNP ( ) upon addition of 5-mm con-
canavalin A.
Maltose- and maltotriose-containing glyconanoparticles
11-AuNP, 12-AuNP, 14-AuNP, and 15-AuNP were subjected
to the GA-assay conditions (Figure 3). The results showed
that the oxime glyco-AuNPs (11-AuNP and 12-AuNP) al-
lowed a certain level of functional recognition by the
enzyme, leading to partial hydrolysis of the glyco-AuNPs
under these conditions. Conversely, the maltotriose oxy-
amine AuNPs (15-AuNP) displayed complete stability to-
wards the enzyme. We ascribe this observation to the in-
creased ligand density observed with these glyco-AuNPs, as
measured in the thiolytic-release assay above. These results
corroborate the increased stability of glycan ligands on gly-
conanoparticles towards enzymatic degradation and empha-
size the ligand density as a pivotal factor for stability.
Determination of ligand density of glyco-AuNPs by gluco-
amylase and thiolytic release: To estimate the ligand density
of the prepared glyconanoparticles, we performed a series of
thiolytic-release (ligand-displacement) experiments on the
glyconanoparticles. Glyconanoparticles were exposed to
excess 2-mercaptoethanol at 378C for 2 h to completely
detach all carbohydrate ligands from the surface of the
AuNPs.[75] After separation from AuNP residues by centrifu-
gal filtration, the samples were concentrated under high
vacuum to remove 2-mercaptoethanol and then subjected to
the coupled glucoamylase–glucose oxidase/peroxidase assay
in 10 mm phosphate buffer pH 7.0 (Figure 3). These results
showed that for maltose and maltotriose oxime AuNPs (11-
AuNP and 12-AuNP, respectively) the ligand density was
around 300 mol ligand/mol AuNPs, when corrected for the
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1655

K. J. Jensen et al.
Synthesis of OEG linkers
N-(2-{2-[2-(2-Hydroxyethoxy)ethoxy]ethoxy}ethoxy)-phthalimide
Conclusion
(1):
Tetra-(ethylene glycol) (0,97 g, 5.0 mmol) was dissolved in dry tetrahy-
drofuran (80 mL) under argon, followed by addition of PS-PPh3 (5.07 g,
7.5 mmol, 1.48 mmolgꢀ1) and N-hydroxyphthalimide (0.82 g, 5.0 mmol).
The highly chemoselective reaction of glycans through their
reducing ends with aminooxy groups, via oxime formation,
has become an important method in glycobiology. However,
it is most often considered that this leaves the monosacchar-
ide at the reducing end in a non-native open-chain form.
Here, we describe, firstly, a novel bifunctional linker for an-
choring glycans by oxime formation to nanoparticles and,
secondly, that the oxime-tautomer forms include ring-closed
forms that are correctly recognized by proteins.
Under stirring,
a solution of diisopropylazodicarboxylate (1.52 g,
7.5 mmol) in dry tetrahydrofuran (20 mL) was added dropwise over
30 min. The mixture was stirred for 4 h and then left unstirred for 16 h at
RT. The resin was filtered off on a pad of Celite, and the filtrate was con-
centrated by rotary evaporation. The residue was purified by filtration on
a silica plug by washing extensively with ethyl acetate to provide the de-
sired mono-substituted product as a clear oil (622 mg, 37%). 1H NMR
(300 MHz, [D]chloroform): d=7.88–7.80 (m, 2H; Phth ortho), 7.78–7.71
(m, 2H; Phth meta), 4.41–4.36 (m, 2H; CH2ON), 3.89–3.84 (m, 2H;
OCH2CH2ON), 3.74–3.66 (m, 4H), 3.66–3.57 (m, 8H), 2.47 ppm (t, J=
6.2 Hz, OH); 13C NMR (75 MHz, [D]chloroform): d=163.4 (C=O), 134.4
(Phth meta), 128.9 (Phth ipso), 123.5 (Phth ortho), 77.2 (CH2ON), 72.4
(OCH2CH2OH), 70.8, 70.6, 70.4, 70.3, 69.2 (OCH2CH2ON), 61.7 ppm
(CH2OH); HRMS (ES): m/z: calcd for C16H22NO7: 340.1396 [M+H]+;
found: 340.1378.
Our design of a novel heterobifunctional OEG linker in-
corporates the following features: 1) an aminooxy group for
the chemoselective capture of reducing glycans under mild
conditions, 2) a trityl group that can be used as a handle for
isolation of the otherwise highly polar glycoconjugates, 3) a
mercapto group that upon deprotection provides chemose-
lective coupling onto the nanoparticle surface, and 4) a tet-
ra(ethylene glycol) chain that provides aqueous solubility to
the nanoparticles and prevents unspecific binding with pro-
teins. Using this linker reagent, we obtained glyconanoparti-
cles that were soluble under a wide range of conditions.
In light of the wide use of the oxime formation for attach-
ment of glycans to surfaces, the consequences of the ring-
chain tautomerism of carbohydrate oxime ethers during in-
teraction with proteins is of considerable interest. Our re-
sults with two model proteins, glucoamylase from Aspergil-
lus niger and concanavalin A from Concanavalia ensiformis,
show that, in both cases, functional recognition of the carbo-
hydrate oximes occurs with the cyclic tautomeric forms, and
that the equilibria exist when the carbohydrate oxime ethers
are attached to gold-nanoparticle surfaces. This was corro-
borated by extensive comparison with the corresponding re-
duced, open-chain forms. We believe this to be an important
finding as it allows efficient anchoring of glycans through
the reducing end while maintaining a correct presentation of
the glycan.
Recovery of the bis-substituted product provided 429 mg (18%) of N,N’-
(2,2’-{2,2’-oxybisACTHNUTRGENNGU[ethoxy]}bisACHTUGNTREN{NUGN ethoxy})diphthalimide (bis-1) as a clear oil.
1H NMR (300 MHz, [D]chloroform): d=7.86–7.78 (m, 4H; Phth ortho),
7.77–7.70 (m, 4H; Phth meta), 4.38–4.33 (m, 4H; 2ꢅCH2ON), 3.86–3.81
(m, 4H; 2ꢅOCH2CH2ON), 3.62–3.55 (m, 4H; 2ꢅCH2OCH2CH2ON),
3.53–3.46 ppm (m, 4H; 2ꢅOCH2CH2OCH2CH2ON); 13C NMR (75 MHz,
[D]chloroform): d=163.4 (C=O), 134.4 (Phth meta), 128.9 (Phth ipso),
123.4 (Phth ortho), 77.2 (CH2ON), 70.7 (OCH2CH2OCH2CH2ON), 70.4
(OCH2CH2OCH2CH2ON), 69.2 ppm (OCH2CH2ON); HRMS (ES): m/z:
calcd for C24H25N2O9S: 485.1560 [M+H]+; found: 485.1572.
N-(2-{2-[2-(2-Tritylsulfanylethoxy)ethoxy]ethoxy}ethoxy)phthalimide (2):
Diisopropylazodicarboxylate (0.607 g, 3.0 mmol) was added to a stirred
suspension of polymer-supported triphenylphosphine (2.03 g, 3.0 mmol,
1.48 mmolgꢀ1) in dry tetrahydrofuran (25 mL) at 08C under argon. The
mixture was stirred for 30 min at 08C. Then, a solution of 1 (0.509 g,
1.5 mmol) and triphenylmethanethiol (0.829 g, 3.0 mmol) in dry tetrahy-
drofuran (20 mL) was added dropwise at 08C. The suspension was left to
reach RT and stirred for an additional 4 h at RT. The resin was filtered
off on a pad of Celite, and the filtrate was concentrated by rotary evapo-
ration. The residue was purified by vacuum liquid chromatography
(VLC) (diethylether/heptane 1:1!1:0) to provide the title compound as
a clear oil (699 mg, 78%). 1H NMR (300 MHz, [D]chloroform): d=7.86–
7.79 (m, 2H; o-ArH), 7.76–7.69 (m, 2H; m-ArH), 7.44–7.38 (m, 6H; Tr),
7.31–7.16 (m, 9H; Tr), 4.38–4.33 (m, 2H; CH2ON), 3.87–3.81 (m, 2H;
OCH2CH2ON), 3.66–3.60 (m, 2H), 3.57–3.51 (m, 2H), 3.51–3.45 (m, 2H),
3.43–3.36 (m, 2H), 3.28 (t, J=6.7 Hz, 2H; OCH2CH2STr), 2.41 ppm (t,
J=6.7 Hz, 2H; CH2STr); 13C NMR (75 MHz, [D]chloroform): d=163.3
(C=O), 144.7 (Tr ipso), 134.3 (Phth meta), 129.5 (Tr), 129.0 (Phth ipso),
127.8 (Tr), 126.6 (Tr para), 123.4 (Phth ortho), 77.1 (CH2ON), 70.7, 70.4,
70.3, 70.0, 69.5, 69.2 (OCH2CH2STr), 66.5 (Tr quaternary), 31.6 ppm
(CH2STr); HRMS (ES): m/z: calcd for C35H36NO6: 620.2083 [M!Na]+;
found: 620.2089.
Experimental Section
Materials and analysis: All chemicals, as well as enzymes and lectins,
were purchased from Sigma–Aldrich Denmark, except for polymer-sup-
ported triphenylphosphine (PL-TPP resin, PS-PPh3), which was a gift
from Varian (Polymer Laboratories). MilliQ water was used for all aque-
ous preparations. All solvent ratios are v/v. 1H, 13C, attached proton test
(APT), HSQC, and COSY NMR spectra were recorded by using a
Bruker Avance 300 spectrometer with a BBO probe. The chemical shifts
are referenced to the residual solvent signal. Assignments were aided by
COSY, HSQC, and APT experiments. Mass determination (high-resolu-
tion MS, HRMS) was performed by using a Micromass LCT instrument
with an ESI probe. UV/Vis spectroscopy was performed by using a Jasco
V-650 spectrophotometer. Analytical HPLC was performed by using a
Dionex Ultimate 3000 system with Chromeleon 6.80SP3 software; a 2.0-
mLminꢀ1 linear gradient flow of MeCN-H2O (0.1% trifluoroacetic acid
(TFA)) 1:20!1:0 over 6.5 min was used for separations on a C-18
(Daiso, 10 mm, 200 ꢆ, 250ꢅ8 mm OD) column from FEF Chemicals.
Transmission electron microscopy (TEM) was recorded by using a Philips
CM20 instrument at 200 keV.
O-(2-{2-[2-(2-Tritylsulfanylethoxy)ethoxy]ethoxy}ethyl)hydroxylamine (3):
Compound 2 (0.598 g, 1.0 mmol) was dissolved in acetonitrile (20 mL)
and hydrazine hydrate (230 mL, 5.0 mmol) was added. Stirring for 2 h at
RT yielded a white suspension that was concentrated by rotary evapora-
tion. The residue was suspended in dichloromethane and filtered through
a plug of Celite. Evaporation to dryness of the filtrate yielded the title
compound as
a
clear oil (0.454 g, 97%). 1H NMR (300 MHz,
[D]chloroform): d=7.44–7.38 (m, 6H; Tr), 7.31–7.16 (m, 9H; Tr), 5.48
(brs, 2H; ONH2), 3.84–3.79 (m, 2H; CH2ON), 3.69–3.64 (m, 2H;
OCH2CH2ON), 3.63 (s, 4H), 3.60–3.55 (m, 2H), 3.48–3.43 (m, 2H), 3.31
(t, J=6.9 Hz, 2H; OCH2CH2STr), 2.43 ppm (t, J=6.9 Hz, 2H; CH2STr);
13C NMR (75 MHz, [D]chloroform): d=144.8 (Tr ipso), 129.6 (Tr), 127.8
(Tr), 126.6 (Tr para), 74.8 (CH2ON), 70.6, 70.5, 70.4, 70.1, 69.7, 69.6, 66.6
(Tr quarternary), 31.6 ppm (CH2STr); HRMS (ES): m/z: calcd for
C27H34NO4S: 468.2209 [M+H]+; found: 468.2199.
1656
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660

Glyconanoparticles
FULL PAPER
General procedure for oxime formation of 3 with glucose, maltose, and
maltotriose: The aminooxy linker 3 (150 mg, 0.32 mmol) was dissolved in
acetonitrile (5 mL) and either d-glucose, d-maltose monohydrate, or d-
maltotriose (0.35 mmol) was added. Water (2–5 mL) was added to pro-
vide clear solutions, followed by addition of glacial acetic acid (150 mL).
The reaction mixtures were stirred at RT for 16 h and then concentrated
by rotary evaporation. The residues were purified by VLC (methanol/di-
chloromethane 1:20!1:8) to provide the glucose oxime 4 (100 mg, 74%),
(2 mL) and sodium cyanoborohydride (13 mg, 0.20 mmol) was added.
The mixtures were stirred for 2 h at RT and then concentrated by rotary
evaporation. The residues were redissolved in methanol and purified on
cation-exchange cartridges (Varian Bond Elut SCX) to provide the glu-
cose-derived oxy-amine 7 (58 mg, 92%), the maltose-derived oxy-amine
8 (67 mg, 85%), and the maltotriose-derived oxy-amine 9 (82 mg, 86%)
as thick, clear oils.
1-Deoxy-1-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethoxyamino)-d-
glucitol (7): 1H NMR (300 MHz, [D4]methanol): d=7.43–7.37 (m, 6H;
Tr), 7.32–7.18 (m, 9H; Tr), 3.99 (ddd, J=8.0, 4.1, 3.9 Hz, 1H; H-2), 3.83–
3.75 (m, 4H; CH2ON, H-3, H-6), 3.70 (dd, J=8.6, 3.1 Hz, 1H; H-4),
3.66–3.58 (m, 8H; 3ꢅCH2 OEG, H-5, H-6), 3.57–3.52 (m, 2H; CH2
OEG), 3.45–3.40 (m, 2H; CH2 OEG), 3.26 (t, J=6.7 Hz, 2H;
OCH2CH2STr), 3.12 (dd, J=13.3, 3.9 Hz, 1H; H-1), 2.92 (dd, J=13.3,
8.0 Hz, 1H; H-1), 2.39 ppm (t, J=6.7 Hz, 2H; CH2STr); 13C NMR
(75 MHz, APT, [D4]methanol): d=146.3 (Tr ipso), 130.8 (Tr), 128.9 (Tr),
127.8 (Tr), 73.7 (C-5), 73.6 (CH2ON), 73.0 (C-4), 72.1 (C-3), 71.5 (CH2
OEG), 71.5 (CH2 OEG), 71.4 (C-2), 71.4 (CH2 OEG), 71.2 (CH2 OEG),
70.6 (CH2 OEG), 70.6 (CH2 OEG), 67.8 (Tr quarternary), 64.8 (C-6),
55.2 (C-1), 32.8 ppm (CH2STr); HRMS (ES): m/z: calcd for
C33H45NO9NaS: 654.2713 [M+Na]+; found: 654.2664.
the maltose oxime
(270 mg, 88%), respectively, as thick, clear oils.
d-Glucose O-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethyl)oxime
5 (214 mg, 86%), and the maltotriose oxime 6
(4): 1H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyranose
tautomeric ratio 60:20:20): d=7.44 (d, J=6.7 Hz, 0.6H; (E)-oxime H-1),
7.42–7.36 (m, 6H; Tr), 7.33–7.17 (m, 9H; Tr), 6.78 (d, J=6.0 Hz, 0.2H;
(Z)-oxime H-1), 4.94 (t, J=6.0 Hz, 0.2H (partly overlapped by H2O
signal); (Z)-oxime H-2), 4.32 (t, J=6.7 Hz, 0.6H; (E)-oxime H-2), 4.21–
4.12 (m, 2.2H; CH2ON+b-pyranose H-1 (J=9.2 Hz from HSQC)), 3.94–
3.85 (m, 0.8H; (E)-oxime+(Z)-oxime H-3), 3.84–3.18 (m, 16.4H),
2.39 ppm (t, J=6.7 Hz, 2H; CH2STr); 13C NMR (75 MHz, APT,
[D4]methanol): d=154.2 ((Z)-oxime C-1), 151.8 ((E)-oxime C-1), 146.3
(Tr ipso), 130.8 (Tr), 129.0 (Tr), 127.8 (Tr), 92.9 (b-pyranose C-1), 79.4,
79.1 (b-pyranose H-2), 74.9 (CH2ON), 74.4 (CH2ON), 74.1 (CH2ON),
73.2, 72.9, 72.7, 72.6, 72.3, 72.2, 72.0, 71.6 (CH2 OEG), 71.6 (CH2 OEG),
1-Deoxy-1-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethoxyamino)-d-
maltitol (8): 1H NMR (300 MHz, [D4]methanol): d=7.43–7.37 (m, 6H;
Tr), 7.32–7.18 (m, 9H; Tr), 5.03 (d, J=4.0 Hz, 1H; H-1’), 4.00 (ddd, J=
8.0, 4.4, 2.0 Hz, 1H; H-2), 3.94–3.59 (m, 17H), 3.58–3.52 (m, 2H; CH2
OEG), 3.49–3.40 (m, 3H; CH2 OEG, C-4’), 3.34–3.28 (m, 1H; overlapped
by [D4]methanol peak), 3.27 (t, J=6.7 Hz, 2H; OCH2CH2STr), 3.05 (dd,
J=13.5, 4.4 Hz, 1H; H-1), 3.00 (dd, J=13.5, 8.0 Hz, 1H; H-1), 2.39 ppm
(t, J=6.7 Hz, 2H; CH2STr); 13C NMR (75 MHz, APT, [D4]methanol):
d=146.3 (Tr ipso), 130.8 (Tr), 128.9 (Tr), 127.8 (Tr), 102.9 (C-1’), 85.5 (C-
5’), 75.1, 74.5, 74.2, 73.8, 73.7 (CH2ON), 73.4, 71.6 (CH2 OEG), 71.5, 71.5
(CH2 OEG), 71.4 (CH2 OEG), 71.2 (CH2 OEG), 70.6 (CH2 OEG), 70.6
(CH2 OEG), 69.2, 67.8 (Tr quarternary), 64.0 (C-6), 62.6 (C-6), 55.8 (C-
1), 32.8 ppm (CH2STr); HRMS (ES): m/z: calcd for C39H55NO14NaS:
816.3241 [M+Na]+; found: 814.3232.
71.4ACHTUNGTRENNUNG(CH2 OEG), 71.2 (CH2 OEG), 70.6 (CH2 OEG), 70.6 (CH2 OEG),
68.5, 67.8 (Tr quarternary), 64.9 (C-6), 63.0 (C-6), 32.8 ppm (CH2STr);
HRMS (ES): m/z: calcd for C33H43NO9NaS: 652.2556 [M+Na]+; found:
652.2552.
d-Maltose
O-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethyl)oxime
(5): 1H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyranose
tautomeric ratio 60:20:20): d=7.49 (d, J=6.7 Hz, 0.6H; (E)-oxime H-1),
7.43–7.36 (m, 6H; Tr), 7.32–7.18 (m, 9H; Tr), 6.83 (d, J=5.4 Hz, 0.2H;
(Z)-oxime H-1), 5.16 (d, J=3.8 Hz, 0.2H; b-pyranose H-1’), 5.05 (d, J=
3.9 Hz, 0.8H; (E)-oxime+(Z)-oxime H-1’), 4.91 (dd, J=5.4, 2.5 Hz, 0.2H
(partly overlapped by H2O signal); (Z)-oxime H-2), 4.43 (dd, J=6.7,
4.3 Hz, 0.6H; (E)-oxime H-2), 4.20–4.11 (m, 2.2H; CH2ON+b-pyranose
H-1 (J=9.1 Hz from HSQC)), 4.08–3.22 (m, 23.2H), 2.39 ppm (t, J=
6.7 Hz, 2H; CH2STr); 13C NMR (75 MHz, APT, [D4]methanol): d=154.8
((Z)-oxime C-1), 152.1 ((E)-oxime C-1), 146.2 (Tr ipso), 130.8 (Tr), 128.9
(Tr), 127.8 (Tr), 102.9 (C-1’), 102.7 (C-1’), 102.4 (C-1’), 92.7 (b-pyranose
C-1), 84.5, 83.0, 81.2, 78.7, 77.9, 75.1, 75.0, 74.8 (CH2ON), 74.7, 74.5, 74.4,
74.1 (CH2ON), 74.1, 73.9, 73.8, 73.1, 71.5 (CH2 OEG), 71.5 (CH2 OEG),
71.5 (CH2 OEG), 71.4 (CH2 OEG), 71.2 (CH2 OEG), 71.1, 70.9, 70.6
(CH2 OEG), 70.6 (CH2 OEG), 70.4 70.6 (CH2 OEG), 67.7 (Tr quarterna-
ry), 67.0, 64.0 (C-6), 63.9 (C-6), 62.7 (C-6), 62.6 (C-6), 62.4 (C-6),
32.8 ppm (CH2STr); HRMS (ES): m/z: calcd for C39H53NO14NaS:
814.3084 [M+Na]+; found: 814.3090.
1-Deoxy-1-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethoxyamino)-d-
1
maltotritol (9): H NMR (300 MHz, [D4]methanol): d=7.43–7.37 (m, 6H;
Tr), 7.32–7.18 (m, 9H; Tr), 5.18 (d, J=3.8 Hz, 1H; H-1’’), 5.04 (d, J=
3.9 Hz, 1H; H-1’), 4.00 (ddd, J=7.8, 4.4, 1.9 Hz, 1H; H-2), 3.96–3.24 (m,
31H), 3.07 (dd, J=13.5, 4.4 Hz, 1H; H-1), 2.98 (dd, J=13.5, 7.8 Hz, 1H;
H-1), 2.39 ppm (t, J=6.7 Hz, 2H; CH2STr); 13C NMR (75 MHz, APT,
[D4]methanol): d=146.2 (Tr ipso), 130.8 (Tr), 128.9 (Tr), 127.8 (Tr), 102.8
(C-1’’), 102.7 (C-1’), 85.2, 81.3, 75.1, 74.9, 74.7, 74.2, 74.2, 73.7 (CH2ON),
73.4, 73.3, 73.0, 71.6 (CH2 OEG), 71.5, 71.5 ((CH2 OEG), 71.4 (CH2
OEG), 71.2 (CH2 OEG), 70.7 (CH2 OEG), 70.6 (CH2 OEG), 69.2, 67.8
(Tr quarternary), 64.0 (C-6), 62.7 (C-6), 62.1 (C-6), 55.8 (C-1), 32.8 ppm
(CH2STr); HRMS (ES): m/z: calcd for C45H65NO19NaS: 978.3769
[M+Na]+; found: 978.3701.
d-Maltotriose O-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]ethoxy}ethyl)ox-
1
ime (6): H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyra-
nose tautomeric ratio 60:20:20): d=7.49 (d, J=6.7 Hz, 0.6H; (E)-oxime
H-1), 7.43–7.36 (m, 6H; Tr), 7.32–7.17 (m, 9H; Tr), 6.84 (d, J=5.4 Hz,
0.2H; (Z)-oxime H-1), 5.18 (d, J=3.8 Hz, 1H; H-1’’), 5.16 (d, J=3.8 Hz,
0.2H; b-pyranose H-1’), 5.08 (d, J=3.9 Hz, 0.8H; (E)-oxime+(Z)-oxime
H-1’), 4.91 (dd, J=5.3, 2.4 Hz, 0.2H (partly overlapped by H2O signal);
(Z)-oxime H-2), 4.43 (dd, J=6.6, 4.3 Hz, 0.6H; (E)-oxime H-2), 4.20–
4.12 (m, 2.2H; CH2ON+b-pyranose H-1 (J=9.2 Hz from HSQC)), 4.08–
3.22 (m, 29.2H), 2.38 ppm (t, J=6.7 Hz, 2H; CH2STr); 13C NMR
(75 MHz, APT, [D4]methanol): d=154.8 ((Z)-oxime C-1), 152.1 ((E)-
oxime C-1), 146.2 (Tr ipso), 130.7 (Tr), 128.9 (Tr), 127.8 (Tr), 102.8 (C-
1’’), 102.7 (C-1’’), 102.4 (C-1’), 102.2 (C-1’), 92.7 (b-pyranose C-1), 84.2,
82.9, 81.2, 78.7, 77.8, 75.0, 74.9, 74.8, 74.7, 74.4 (CH2ON), 74.3, 74.2
(CH2ON), 74.1, 74.1, 73.9, 73.7, 73.4, 73.3, 73.0, 71.6 (CH2 OEG), 71.5
(CH2 OEG), 71.4, 71.3 (CH2 OEG), 71.1 (CH2 OEG), 71.0, 70.9, 70.6
(CH2 OEG), 70.6 (CH2 OEG), 70.3, 70.2, 67.7 (Tr quarternary), 67.1, 64.0
(C-6), 63.9 (C-6), 62.7 (C-6), 62.4 (C-6), 62.0 (C-6), 32.8 ppm (CH2STr);
HRMS (ES): m/z: calcd for C45H63NO19NaS: 976.3613 [M+Na]+; found:
976.3559.
General procedure for removal of the trityl group from compounds 4–9:
Compounds 4, 5, 6, 7, 8, or 9 (0.050 mmol) were dissolved/suspended in
dichloromethane (1 mL) under argon. Triethylsilane (0.5 mL) was added,
followed by a dropwise addition of trifluoroacetic acid (1 mL) under stir-
ring. A transient yellow color was observed during the addition, however,
all reaction mixtures were clear solutions after the addition. The reaction
mixtures were stirred at RT for 5 min and then concentrated by rotary
evaporation. The residues were purified by redissolution in methanol/di-
chloromethane 1:20–1:10 and filtration through
a silica plug. Once
washed, the plug was thoroughly dried and the products were then eluted
by methanol. Evaporation to dryness provided the desired thiols 10–15 as
thick, clear oils in quantitative yields, partially as disulfides (0–40%).
d-Glucose O-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethyl)oxime (10):
1H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyranose tau-
tomeric ratio 60:20:20, 25% disulfide): d=7.45 (d, J=6.9 Hz, 0.6H; (E)-
oxime H-1), 6.79 (d, J=6.1 Hz, 0.2H; (Z)-oxime H-1), 4.94 (dd, J=6.1,
5.4 Hz, 0.2H (partly overlapped by H2O signal); (Z)-oxime H-2), 4.32
(dd, J=6.9, 6.3 Hz, 0.6H; (E)-oxime H-2), 4.23–4.15 (m, 2.2H; CH2ON+
b-pyranose H-1), 3.92 (dd, J=5.4, 2.1 Hz, 0.2H; (Z)-oxime H3), 3.88 (dd,
J=6.3, 1.6 Hz, 0.8H; (E)-oxime H-3), 3.84–3.00 (m, 16.4H), 2.92 (t, J=
General procedure for reduction of carbohydrate oximes 4–6: The carbo-
hydrate oximes 4, 5, or 6 (0.10 mmol) were dissolved in glacial acetic acid
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1657

K. J. Jensen et al.
6.4 Hz, 0.5H; CH2S disulfide), 2.66 ppm (t, J=6.5 Hz, 1.5H; CH2SH);
13C NMR (75 MHz, APT, [D4]methanol, 25% disulfide): d=154.2 ((Z)-
oxime C-1), 151.8 ((E)-oxime C-1), 92.8 (b-pyranose C-1), 86.9
(OCH2CH2S disulfide), 79.4, 79.1, 74.9 (CH2ON), 74.4 (CH2ON), 74.2
(CH2ON), 74.1 (OCH2CH2SH) 73.2, 72.9, 72.7, 72.6, 72.3, 72.2, 72.0, 92.7,
71.6 (CH2 OEG), 71.5 (CH2 OEG), 71.3, CH2 OEG), 71.1 (CH2 OEG),
70.6 (CH2 OEG), 70.5 (CH2 OEG), 70.5 (CH2 OEG), 68.5, 64.9 (C-6),
63.0 (C-6), 39.5 (CH2S disulfide), 24.7 ppm (CH2SH); HRMS (ES): m/z:
calcd for C14H30NO9S: 388.1641 [M+H]+; found: 388.1666.
39.4 ((CH2S disulfide), 24.6 ppm (CH2SH); HRMS (ES): m/z: calcd for
C20H42NO14S: 552.2326 [M+H]+; found: 552.2367.
1-Deoxy-1-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethoxyamino)-d-
maltotritol (15): 1H NMR (300 MHz, [D4]methanol, 10% disulfide): d=
5.17 (d, J=3.8 Hz, 1H; H-1’’), 5.07 (d, J=3.9 Hz, 1H; H-1’), 4.02 (m,
1H; H-2), 3.99–3.23 (m, 31H), 3.08 (dd, J=13.5, 4.5 Hz, 1H; H-1), 2.99
(dd, J=13.5, 7.8 Hz, 1H; H-1), 2.93 (t, J=6.4 Hz, 0.2H; CH2S disulfide),
2.67 ppm (t, J=6.5 Hz, 1.8H; CH2SH); 13C NMR (75 MHz, APT,
[D4]methanol): d=102.9 (C-1’’), 102.5 (C-1’), 84.8, 81.3, 75.1, 74.8, 74.8,
74.3, 74.2, 74.1 (OCH2CH2SH), 73.7 (CH2ON), 73.3, 73.1, 72.8, 71.5 (CH2
OEG), 71.4 (CH2 OEG), 71.4 (CH2 OEG) 71.3, 71.1 (CH2 OEG), 70.7
(CH2 OEG), 70.4, 69.7, 64.0 (C-6), 62.7 (C-6), 62.0 (C-6), 55.7 (C-1), 39.4
(CH2S disulfide), 24.6 ppm (CH2SH); HRMS (ES): m/z: calcd for
C26H52NO19S: 714.2854 [M+H]+; found: 714.2916.
NMR study: d-[1-13C]Glucose O-(2-{2-[2-(2-tritylsulfanylethoxy)ethoxy]-
ethoxy}-ethyl)oxime (16): The aminooxy linker 3 (23 mg, 0.050 mmol)
was dissolved in acetonitrile (2 mL) and d-[1-13C]glucose (14 mg,
0.075 mmol) was added. Water (1 mL) was added, followed by glacial
acetic acid (50 mL). The reaction mixtures were stirred at RT for 16 h
and then concentrated by rotary evaporation. The residue was purified
by VLC (methanol/dichloromethane 1:20!1:6) to provide the
[1-13C]glucose oxime 16 (22 mg, 70% as a clear, thick oil. 1H NMR
(300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyranose tautomeric
ratio 60:20:20): d=7.44 (dd, J=6.9, 166.4 Hz, 0.6H; (E)-oxime H-1),
7.42–7.36 (m, 6H; Tr), 7.33–7.17 (m, 9H; Tr), 6.78 (dd, J=6.1, 178.2 Hz,
0.2H; (Z)-oxime H-1), 4.94 (dt, J=4.3, 6.1 Hz, 0.2H (partly overlapped
by H2O signal); (Z)-oxime H-2), 4.32 (dt, J=4.7, 6.9 Hz, 0.6H; (E)-
oxime H-2), 4.21–4.12 (m, 2.2H; CH2ON+b-pyranose H-1 (J=9.0,
152.9 Hz from H;H-COSY)), 3.94–3.85 (m, 0.8H; (E)-oxime+(Z)-oxime
H-3), 3.84–3.18 (m, 16.4H), 2.39 ppm (t, J=6.7 Hz, 2H; CH2STr);
13C NMR (75 MHz, APT, [D4]methanol): d=154.2 ((Z)-oxime 13C-1),
151.8 ((E)-oxime 13C-1), 146.3 (Tr ipso), 130.8 (Tr), 129.0 (Tr), 127.8 (Tr),
95.6 (minor), 92.9 (b-pyranose 13C-1), 90.7 (minor), 79.4, 79.1, 74.9
(CH2ON), 74.4 (CH2ON), 74.1 (CH2ON), 73.2, 72.9, 72.7, 72.6, 72.3, 72.2,
d-Maltose O-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethyl)oxime (11):
1H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyranose tau-
tomeric ratio 60:20:20, 40% disulfide): d=7.50 (d, J=6.5 Hz, 0.6H; (E)-
oxime H-1), 6.84 (d, J=5.4 Hz, 0.2H; (Z)-oxime H-1), 5.16 (d, J=3.6 Hz,
0.2H; b-pyranose H-1’), 5.05 (d, J=3.8 Hz, 0.8H; (E)-oxime+(Z)-oxime
H-1’), 4.91 (m, 0.2H (partly overlapped by H2O signal); (Z)-oxime H-2),
4.44 (dd, J=6.5, 4.2 Hz, 0.6H; (E)-oxime H-2), 4.24–4.15 (m, 2.2H;
CH2ON+b-pyranose H-1), 4.08–3.00 (m, 23.2H), 2.92 (t, J=6.4 Hz,
0.8H; CH2S disulfide), 2.66 ppm (t, J=6.4 Hz, 1.2H; CH2STr); 13C NMR
(75 MHz, APT, [D4]methanol, 40% disulfide): d=154.8 ((Z)-oxime C-1),
152.1 ((E)-oxime C-1), 146.2 (Tr ipso), 130.8 (Tr), 128.9 (Tr), 127.8 (Tr),
102.9 (C-1’), 102.7 (C-1’), 102.4 (C-1’), 92.7 (b-pyranose C-1), 86.9
(OCH2CH2S disulfide), 84.5, 83.0, 81.2, 78.7, 77.9, 75.1, 75.0, 74.8
(CH2ON), 74.7, 74.5, 74.4, 74.1 (CH2ON), 74.1, 73.9, 73.8, 73.1, 71.6 (CH2
OEG), 71.5 (CH2 OEG), 71.4 (CH2 OEG), 71.3 (CH2 OEG), 71.1 (CH2
OEG), 70.9, 70.7 (CH2 OEG), 70.6 (CH2 OEG), 70.5 (CH2 OEG), 70.3
(CH2 OEG), 67.0, 64.0 (C-6), 63.9 (C-6), 62.7 (C-6), 62.6 (C-6), 62.4 (C-
6), 39.5 (CH2S disulfide), 24.6 ppm (CH2SH); HRMS (ES): m/z: calcd for
C20H40NO14S: 550.2170 [M+H]+; found: 550.2194.
d-Maltotriose
O-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethyl)oxime
(12): 1H NMR (300 MHz, [D4]methanol, (E)-oxime/(Z)-oxime/b-pyra-
nose tautomeric ratio 60:20:20, 30% disulfide): d=7.50 (d, J=6.3 Hz,
0.6H; (E)-oxime H-1), 6.86 (d, J=5.3 Hz, 0.2H; (Z)-oxime H-1), 5.19–
5.15 (m, 1.2H; H-1’’, b-pyranose H-1’), 5.13–5.08 (m, 0.8H; (E)-oxime+
(Z)-oxime H-1’), 4.94–4.90 (m, 0.2H (partly overlapped by H2O signal);
(Z)-oxime H-2), 4.46 (dd, J=6.2, 3.9 Hz, 0.6H; (E)-oxime H-2), 4.24–
4.16 (m, 2.2H; CH2ON+b-pyranose H-1), 4.14–2.98 (m, 29.2H), 2.92 (t,
J=6.4 Hz, 0.6H; CH2S disulfide), 2.67 ppm (t, J=6.4 Hz, 1.4H;
CH2STr); 13C NMR (75 MHz, APT, [D4]methanol): d=154.8 ((Z)-oxime
C-1), 152.1 ((E)-oxime C-1), 102.8 (C-1’’), 102.7 (C-1’’), 102.4 (C-1’),
102.2 (C-1’), 92.7 (b-pyranose C-1), 84.2, 82.9, 81.2, 78.7, 77.8, 75.0, 74.9,
74.8, 74.7, 74.4 (CH2ON), 74.3, 74.2 (CH2ON), 74.1, 74.1, 73.9, 73.7, 73.4,
73.2, 71.5 (CH2 OEG), 71.4 (CH2 OEG), 71.3 (CH2 OEG), 71.2 (CH2
OEG), 71.2, 71.1 (CH2 OEG), 71.0 (CH2 OEG), 70.7 (CH2 OEG), 70.6
(CH2 OEG), 70.4 (CH2 OEG), 67.7, 64.0 (C-6), 63.9 (C-6), 62.7 (C-6),
62.4 (C-6), 62.0 (C-6), 39.4 (CH2S disulfide), 24.6 ppm (CH2SH); HRMS
(ES): m/z: calcd for C26H50NO19S: 712.2698 [M+H]+; found: 712.2717.
72.0, 71.6 (CH2 OEG), 71.6 (CH2 OEG), 71.4ACTHUNGTRNEUNG(CH2 OEG), 71.2 (CH2
OEG), 70.6 (CH2 OEG), 70.6 (CH2 OEG), 68.5, 67.8 (Tr quarternary),
64.9 (C-6), 63.0 (C-6), 32.8 ppm (CH2STr); HRMS (ES): m/z: calcd for
C32H43NO9S13C: 653.2589 [M+Na]+; found: 653.2670.
Preparation of 12-nm diameter citrate-stabilized gold nanoparticles:
Spherical citrate-stabilized gold nanoparticles were prepared by the Tur-
kevich protocol[19] according to a recently published procedure.[69] Briefly,
a
warm (ꢁ50–608C) solution of sodium citrate dihydrate (228 mg,
0.776 mmol) in helium-degassed MilliQ water (20 mL) was added to a
stirred solution of hydrogen tetrachloroaurate(III) trihydrate (79 mg,
AHCTUNGTRENNUNG
0.200 mmol) in helium-degassed MilliQ water (200 mL) at reflux. The re-
sulting dark-red solution was stirred at reflux for 30 min, where after it
was cooled to RT and filtered through a 0.25 mm syringe filter.
1-Deoxy-1-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethoxyamino)-d-glu-
citol (13): 1H NMR (300 MHz, [D4]methanol, 0% disulfide): d=4.00
(ddd, J=8.0, 4.1, 3.9 Hz, 1H; H-2), 3.86–3.79 (m, 3H; CH2ON and H-3),
3.78–3.61 (m, 14H), 3.61 (t, J=6.5 Hz, 2H; OCH2CH2SH), 3.12 (dd, J=
13.3, 3.9 Hz, 1H; H-1), 3.11 (dd, J=13.3, 8.0 Hz, 1H; H-1), 2.66 ppm (t,
J=6.7 Hz, 2H; CH2SH); 13C NMR (75 MHz, APT, [D4]methanol, 0% di-
sulfide): d=74.1 (OCH2CH2SH), 74.1, 73.7 (CH2ON), 73.0, 71.7, 71.6,
71.5 (CH2 OEG), 71.5 (CH2 OEG), 71.4 (CH2 OEG), 71.1 (CH2 OEG),
70.7 (CH2 OEG), 64.7 (C-6), 55.3 (C-1), 24.6 ppm (CH2SH); HRMS
(ES): m/z: calcd for C14H32NO9S: 390.1798 [M+H]+; found: 390.1804.
Ligand exchange and purification of gold glyconanoparticles: Solutions
of 10, 11, 12, 13, 14, or 15 (0.050 mmol) in methanol (5 mL) were added
to stirred solutions of citrate-stabilized gold nanoparticles (50 mL). A
slight color change towards purple was observable. The solutions were
stirred for a minimum of 16 h and the gold glyconanoparticles were puri-
fied in 10-mL batches by centrifugal filtration (Millipore Amicon Centri-
plus 50 kDa) with five consecutive filtration and redissolution (2 mL
MilliQ water, 10ꢅ dilution) steps. After the final filtration, the gold gly-
conanoparticles were redissolved in 10 mm phosphate buffer pH 7.0
(400 mL) to provide final nanoparticle concentrations of 0.50 mm of 10-
AuNP, 11-AuNP, 12-AuNP, 13-AuNP, 14-AuNP, and 15-AuNP.
1-Deoxy-1-(2-{2-[2-(2-mercaptoethoxy)ethoxy]ethoxy}ethoxyamino)-d-
1
maltitol (14): H NMR (300 MHz, [D4]methanol, 15% disulfide): d=5.07
(d, J=4.0 Hz, 1H; H-1’), 4.03 (ddd, J=7.8, 4.6, 1.4 Hz, 1H; H-2), 4.00–
3.62 (m, 21H), 3.61 (t, J=6.5 Hz, 2H; OCH2CH2SH), 3.50 (dd, J=9.7,
4.0 Hz, 1H; C-4’), 3.34–3.27 (m, 1H; overlapped by [D4]methanol peak),
3.08 (dd, J=13.5, 4.6 Hz, 1H; H-1), 2.99 (dd, J=13.5, 7.8 Hz, 1H; H-1),
2.92 (t, J=6.5 Hz, 0.3H; CH2S disulfide), 2.67 ppm (t, J=6.5 Hz, 1.7H;
CH2SH); 13C NMR (75 MHz, APT, [D4]methanol): d=102.6 (C-1’), 84.7
(C-5’), 74.9, 74.7, 74.3, 74.1 (OCH2CH2SH), 73.7, 73.7 (CH2ON), 72.6,
71.5 (CH2 OEG), 71.4 (CH2 OEG), 71.4 (CH2 OEG), 71.3, 71.1 (CH2
OEG), 70.7 (CH2 OEG), 70.4, 69.8, 63.9 (C-6), 62.6 (C-6), 55.8 (C-1),
Glucoamylase assays: Glucoamylase (GA)-catalyzed hydrolysis in solu-
tion (50 mL in 1.5-mL centrifuge tubes) was performed using 1 mgmLꢀ1
GA in 10 mm sodium phosphate pH 7.0 at 428C with 10 nmol of either
the free saccharides d-glucose, d-maltose monohydrate, and d-malto-
triose or compounds 4, 5, 6, 7, 8, and 9, respectively, as substrate. Re-
leased glucose was measured after 20 h by using the glucose oxidase–per-
oxidase assay kit (Sigma–Aldrich), with reaction volumes adapted for use
in microplates and microplate reader (PowerWavex, Bio-Tek Instru-
ments), measuring the absorbance at 540 nm.
1658
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660

Glyconanoparticles
FULL PAPER
Adaptation of the assay for the use on glyconanoparticles (50 pmol parti-
cles) required inclusion of a centrifugal filtration step (Amicon Ultra-
free-MC 30 kDa cut-off from Millipore) immediately after hydrolysis to
remove gold nanoparticles that would otherwise interfere with absorption
measurements at 540 nm.
[18] M. Brust, M. Walker, D. Bethell, D. J. Schiffrin, R. Whyman, J.
Chem. Soc. Chem. Commun. 1994, 801–802.
[19] J. Turkevich, P. C. Stevenson, J. Hillier, J. Phys. Chem. 1953, 57,
670–673.
[20] G. Frens, Nature Phys. Sci. 1973, 241, 20–22.
[21] J. M. de la Fuente, A. G. Barrientos, T. C. Rojas, J. Rojo, J. Canada,
A. Fernandez, S. Penades, Angew. Chem. 2001, 113, 2317–2321;
Angew. Chem. Int. Ed. 2001, 40, 2257–2261.
[22] K. M. Halkes, A. C. de Souza, C. E. P. Maljaars, G. J. Gerwig, J. P.
Kamerling, Eur. J. Org. Chem. 2005, 3650–3659.
[23] A. Sundgren, J. J. Barchi, Jr., Carbohydr. Res. 2008, 343, 1594–1604.
[24] D. C. Hone, A. H. Haines, D. A. Russell, Langmuir 2003, 19, 7141–
7144.
[25] C. C. Lin, Y. C. Yeh, C. Y. Yang, G. F. Chen, Y. C. Chen, Y. C. Wu,
C. C. Chen, Chem. Commun. 2003, 2920–2921.
Thiolytic quantification of ligands on AuNPs: Glyconanoparticles (50 mL,
0.5 mm) were treated with 2-mercaptoethanol (50 mL, 1.0m in water) at
378C for 2 h. The resulting dark suspensions were filtered through micro-
filtration cartridges (Amicon Ultrafree-MC 30 kDa cut-off from Milli-
pore) to remove gold nanoparticles, and the filtrates were concentrated
by freeze drying. Remaining 2-mercaptoethanol was removed under high
vacuum at 808C for 4 h. The completely odorless residues were redis-
solved in 10 mm phosphate buffer pH 7.0 (50 mL). GA-catalyzed hydroly-
sis was performed on the samples with measurement of released glucose
by using the glucose oxidase–peroxidase assay kit, as described above.
[26] S. A. Svarovsky, Z. Szekely, J. J. Barchi, Tetrahedron: Asymmetry
2005, 16, 587–598.
[27] R. J. Baues, G. R. Gray, J. Biol. Chem. 1977, 252, 57–60.
[28] F. Peri, P. Dumy, M. Mutter, Tetrahedron 1998, 54, 12269–12278.
[29] R. E. Valters, W. Flitsch, Ring-Chain Tautomerism, Plenum Press,
New York, 1985.
Lectin assay: Aqueous dispersions of glyconanoparticles (5 nm) were
treated with a final concentration of 5.0 mm of concanavalin A from Con-
canavalia ensiformis by addition of a 1.25 mgmLꢀ1 solution in water. Ag-
gregation was monitored by UV/Vis spectroscopy by measurement of the
shifts in surface-plasmon peak maximum over a period of 48 h.
[30] P. Finch, Z. Merchant, J. Chem. Soc. Perkin Trans. 1 1975, 1682–
1686.
[31] Y. Hatanaka, U. Kempin, P. Jong-Jip, J. Org. Chem. 2000, 65, 5639–
5643.
[32] S. L. Ramsay, C. Freeman, P. B. Grace, J. W. Redmond, J. K. Mac-
Leod, Carbohydr. Res. 2001, 333, 59–71.
[33] C. Delgado, G. E. Francis, D. Fisher, Crit. Rev. Ther. Drug Carrier
Syst. 1992, 9, 249–304.
[34] R. B. Greenwald, Y. H. Choe, J. McGuire, C. D. Conover, Adv. Drug
Delivery Rev. 2003, 55, 217–250.
[35] C. Pale-Grosdemange, E. S. Simon, K. L. Prime, G. M. Whitesides, J.
Am. Chem. Soc. 1991, 113, 12–20.
[36] A. G. Kanaras, F. S. Kamounah, K. Schaumburg, C. J. Kiely, M.
Brust, Chem. Commun. 2002, 2294–2295.
Acknowledgements
The authors kindly thank Inger Jensen at the Nano-Science Center at the
University of Copenhagen for transmission electron microscopy (TEM)
measurements on gold glyconanoparticles. We also thank Dr. Andrew
Coffey from Varian Inc. (Polymer Laboratories) for the polymer-support-
ed triphenylphosphine (PL-TPP resin) and Stig Jacobsen for HRMS de-
terminations. FNU is gratefully acknowledged for a grant (K.J.J.). Finally,
Dr. Emiliano Clꢇ is gratefully acknowledged for help with preparation of
this manuscript.
[37] F. Zhang, M. W. A. Skoda, R. M. J. Jacobs, S. Zorn, R. A. Martin,
C. M. Martin, G. F. Clark, G. Goerigk, F. Schreiber, J. Phys. Chem.
A 2007, 111, 12229–12237.
[38] E. E. Foos, A. W. Snow, M. E. Twigg, M. G. Ancona, Chem. Mater.
2002, 14, 2401–2408.
[39] W. P. Wuelfing, S. M. Gross, D. T. Miles, R. W. Murray, J. Am.
Chem. Soc. 1998, 120, 12696–12697.
[40] M. Zheng, Z. G. Li, X. Y. Huang, Langmuir 2004, 20, 4226–4235.
[41] M. Zheng, F. Davidson, X. Y. Huang, J. Am. Chem. Soc. 2003, 125,
7790–7791.
[42] A. Badia, L. Cuccia, L. Demers, F. Morin, R. B. Lennox, J. Am.
Chem. Soc. 1997, 119, 2682–2692.
[43] J. J. Li, X. G. Peng, J. Nanosci. Nanotechnol. 2004, 4, 565–568.
[44] S. Jin, C. V. Miduturu, D. C. McKinney, S. K. Silverman, J. Org.
Chem. 2005, 70, 4284–4299.
[45] B. Bendiak, Carbohydr. Res. 1997, 304, 85–90.
[46] K. Hiromi, Z. I. Hamauzu, K. Takahash, S. Ono, J. Biochem. 1966,
59, 411–418.
[47] K. Hiromi, K. Takahash, Z. I. Hamauzu, S. Ono, J. Biochem. 1966,
59, 469–475.
[48] P. J. Reilly, Starch/Staerke 1999, 51, 269–274.
[49] J. Sauer, B. W. Sigurskjold, U. Christensen, T. P. Frandsen, E. Mir-
gorodskaya, M. Harrison, P. Roepstorff, B. Svensson, Biochim. Bio-
phys. Acta Protein Struct. Mol. Enzymol. 2000, 1543, 275–293.
[50] E. Boel, M. T. Hansen, I. Hjort, I. Hoegh, N. P. Fiil, EMBO J. 1984,
3, 1581–1585.
[51] J. H. Nunberg, J. H. Meade, G. Cole, F. C. Lawyer, P. McCabe, V.
Schweickart, R. Tal, V. P. Wittman, J. E. Flatgaard, M. A. Innis, Mol.
Cell. Biol. 1984, 4, 2306–2315.
[1] A. Varki, Glycobiology 1993, 3, 97–130.
[2] T. Feizi, Glycoconjugate J. 2000, 17, 553–565.
[3] K. M. Koeller, C. H. Wong, Nat. Biotechnol. 2000, 18, 835–841.
[4] For
a review, see: K. Larsen, M. B. Thygesen, F. Guillaumie,
W. G. T. Willats, K. J. Jensen, Carbohydr. Res. 2006, 341, 1209–1234.
[5] F. Guillaumie, J. D. Sterling, K. J. Jensen, O. R. T. Thomas, D.
Mohnen, Carbohydr. Res. 2003, 338, 1951–1960.
[6] F. Guillaumie, S. F. L. Justesen, K. E. Mutenda, P. Roepstorff, K. J.
Jensen, O. R. T. Thomas, Carbohydr. Res. 2006, 341, 118–129.
[7] S. I. Nishimura, K. Niikura, M. Kurogochi, T. Matsushita, M.
Fumoto, H. Hinou, R. Kamitani, H. Nakagawa, K. Deguchi, N.
Miura, K. Monde, H. Kondo, Angew. Chem. 2005, 117, 93–98;
Angew. Chem. Int. Ed. 2005, 44, 91–96.
[8] M. Lee, I. Shin, Org. Lett. 2005, 7, 4269–4272.
[9] O. Bohorov, H. Andersson-Sand, J. Hoffmann, O. Blixt, Glycobiolo-
gy 2006, 16, 21C–27C.
[10] Y. Liu, T. Feizi, M. A. Campanero-Rhodes, R. A. Childs, Y. Zhang,
B. Mulloy, P. G. Evans, H. M. I. Osborn, D. Otto, P. R. Crocker, W.
Chai, Chem. Biol. 2007, 14, 847–859.
[11] S. E. Cervigni, P. Dumy, M. Mutter, Angew. Chem. 1996, 108, 1325–
1328; Angew. Chem. Int. Ed. Engl. 1996, 35, 1230–1232.
[12] A. Lohse, R. Martins, M. R. Jorgensen, O. Hindsgaul, Angew. Chem.
2006, 118, 4273–4278; Angew. Chem. Int. Ed. 2006, 45, 4167–4172.
[13] F. Guillaumie, O. R. T. Thomas, K. J. Jensen, Bioconjugate Chem.
2002, 13, 285–294.
[14] S. Hanson, M. Best, M. C. Bryan, C. H. Wong, Trends Biochem. Sci.
2004, 29, 656–663.
[15] O. Blixt, K. Allin, L. Pereira, A. Datta, J. C. Paulson, J. Am. Chem.
Soc. 2002, 124, 5739–5746.
[52] T. P. Frandsen, C. Dupont, J. Lehmbeck, B. Stoffer, M. R. Sierks,
R. B. Honzatko, B. Svensson, Biochemistry 1994, 33, 13808–13816.
[53] A. E. Aleshin, B. Stoffer, L. M. Firsov, B. Svensson, R. B. Honzatko,
Biochemistry 1996, 35, 8319–8328.
[16] J. M. De la Fuente, S. Penades, Biochim. Biophys. Acta Gen. Subj.
2006, 1760, 636–651.
[17] A. G. Barrientos, J. M. de la Fuente, T. C. Rojas, A. Fernandez, S.
Penades, Chem. Eur. J. 2003, 9, 1909–1921.
Chem. Eur. J. 2009, 15, 1649 – 1660
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1659

K. J. Jensen et al.
[54] M. R. Sierks, C. Ford, P. J. Reilly, B. Svensson, Protein Eng. 1990, 3,
193–198.
[67] S. Gꢂnther, H. Heymann, Arch. Biochem. Biophys. 1998, 354, 111–
116.
[55] E. M. S. Harris, A. E. Aleshin, L. M. Firsov, R. B. Honzatko, Bio-
chemistry 1993, 32, 1618–1626.
[56] A. E. Aleshin, L. M. Firsov, R. B. Honzatko, J. Biol. Chem. 1994,
269, 15631–15639.
[68] A. Cogoli, G. Semenza, J. Biol. Chem. 1975, 250, 7802–7809.
[69] R. Levy, N. T. K. Thanh, R. C. Doty, I. Hussain, R. J. Nichols, D. J.
Schiffrin, M. Brust, D. G. Fernig, J. Am. Chem. Soc. 2004, 126,
10076–10084.
[57] A. E. Aleshin, C. Hoffman, L. M. Firsov, R. B. Honzatko, J. Mol.
Biol. 1994, 238, 575–591.
[70] W. Haiss, N. T. K. Thanh, J. Aveyard, D. G. Fernig, Anal. Chem.
2007, 79, 4215–4221.
[58] B. Svensson, A. J. Clarke, I. Svendsen, H. Moller, Eur. J. Biochem.
1990, 188, 29–38.
[71] T. C. Bissot, R. W. Parry, D. H. Campbell, J. Am. Chem. Soc. 1957,
79, 796–800.
[59] K. Hiromi, M. Ohnishi, A. Tanaka, Mol. Cell. Biochem. 1983, 51,
79–95.
[72] I. J. Goldstein, C. E. Hollerman, E. E. Smith, Biochemistry 1965, 4,
876–883.
[60] M. M. Meagher, Z. L. Nikolov, P. J. Reilly, Biotechnol. Bioeng. 1989,
34, 681–688.
[73] I. J. Goldstein, R. N. Iyer, E. E. Smith, L. L. So, Biochemistry 1967,
6, 2373–2377.
[61] M. R. Sierks, C. Ford, P. J. Reilly, B. Svensson, Protein Eng. 1989, 2,
621–625.
[74] C. F. Brewer, R. D. Brown, Biochemistry 1979, 18, 2555–2562.
[75] L. M. Demers, C. A. Mirkin, R. C. Mucic, R. A. Reynolds, R. L. Let-
singer, R. Elghanian, G. Viswanadham, Anal. Chem. 2000, 72, 5535–
5541.
[62] S. Natarajan, M. R. Sierks, Biochemistry 1996, 35, 3050–3058.
[63] M. R. Sierks, B. Svensson, Protein Eng. 1994, 7, 1479–1484.
[64] M. R. Sierks, B. Svensson, Protein Eng. 1992, 5, 185–188.
[65] T. P. Frandsen, T. Christensen, B. Stoffer, J. Lehmbeck, C. Dupont,
R. B. Honzatko, B. Svensson, Biochemistry 1995, 34, 10162–10169.
[66] L. M. Firsov, K. N. Neustroev, A. E. Aleshin, C. M. Metzler, D. E.
Metzler, R. D. Scott, B. Stoffer, T. Christensen, B. Svensson, Eur. J.
Biochem. 1994, 223, 293–302.
Received: July 25, 2008
Revised: October 14, 2008
Published online: December 29, 2008
1660
www.chemeurj.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 1649 – 1660