Construction of N-glycan microarrays by using modular synthesis and on-chip nanoscale enzymatic glycosylation

Article abstract of DOI:10.1002/chem.201001295

An effective chemoenzymatic strategy is reported that has allowed the construction, for the first time, of a focused microarray of synthetic N-glycans. Based on modular approaches, a variety of N-glycan core structures have been chemically synthesized and covalently immobilized on a glass surface. The printed structures were then enzymatically diversified by the action of three different glycosyltransferases in nanodroplets placed on top of individual spots of the microarray by a printing robot. Conversion was followed by lectin binding specific for the terminal sugars. This enzymatic extension of surface-bound ligands in nanodroplets reduces the amount of precious glycosyltransferases needed by seven orders of magnitude relative to reactions carried out in the solution phase. Moreover, only those ligands that have been shown to be substrates to a specific glycosyltransferase can be individually chosen for elongation on the array. The methodology described here, combining focused modular synthesis and nanoscale on-chip enzymatic elongation, could open the way for the much needed rapid construction of large synthetic glycan arrays. Go nano! Arrays of immobilised carbohydrates can be enzymatically processed with the help of glycosyl transferases in nanodroplets placed on top of individual spots. This methodology opens the way for the stereospecific on-chip construction of glycan arrays with very high ligand densities by using immobilised synthetic glycan scaffolds and minute amounts of recombinant glycosyltransferases (see figure).

Full text of DOI:10.1002/chem.201001295

FULL PAPER
DOI: 10.1002/chem.201001295
Construction of N-Glycan Microarrays by Using Modular Synthesis and
On-Chip Nanoscale Enzymatic Glycosylation
Sonia Serna,[a] Juan Etxebarria,[a] Nerea Ruiz,[a] Manuel Martin-Lomas,[a, b] and
Niels-Christian Reichardt*[a]
Abstract: An effective chemoenzymat-
ic strategy is reported that has allowed
the construction, for the first time, of a
focused microarray of synthetic N-gly-
cans. Based on modular approaches, a
variety of N-glycan core structures
have been chemically synthesized and
covalently immobilized on a glass sur-
face. The printed structures were then
enzymatically diversified by the action
of three different glycosyltransferases
in nanodroplets placed on top of indi-
vidual spots of the microarray by a
printing robot. Conversion was fol-
lowed by lectin binding specific for the
terminal sugars. This enzymatic exten-
sion of surface-bound ligands in nano-
droplets reduces the amount of pre-
cious glycosyltransferases needed by
seven orders of magnitude relative to
reactions carried out in the solution
phase. Moreover, only those ligands
that have been shown to be substrates
to a specific glycosyltransferase can be
individually chosen for elongation on
the array. The methodology described
here, combining focused modular syn-
thesis and nanoscale on-chip enzymatic
elongation, could open the way for the
much needed rapid construction of
large synthetic glycan arrays.
Keywords: glycosyltransferases
microarrays nanodroplets
Glycans · oligosaccharides
·
·
·
Introduction
tionation of glycan mixtures into single compounds and the
unambiguous identification by mass spectrometric methods
are the major difficulties associated with this approach.
A concise study of N-glycan binding would ideally require
a structurally diverse collection of synthetic oligosaccharides
with systematic variations in degree and type of branching,
terminal sugars and core modifications like fucosylation or
xylosylation. Specific applications of focused N-glycan
arrays could include the screening of the binding specifici-
ties of mammalian lectins involved in signalling,[3b,c] protein
trafficking[5] and viral entry,[6] viral envelope proteins,[7]
fungal,[8] and invertebrate lectins and glycan processing en-
zymes.[9]
In spite of extensive work on N-glycan synthesis during
the last twenty years, the construction of a library of syn-
thetic N-glycans constitutes still a serious challenge. Never-
theless, a wealth of effective methodologies and strategies
has been reported and a great deal of valuable know-how
exists for the chemo-enzymatic synthesis of N-glycan struc-
tures. The pioneering contributions of Ogawa[10] and Paul-
sen[11] and the further developments into modular strategies
by Unverzagt,[12] Danishefsky[13] and Ito[14] among others
provide a solid basis for the chemical synthesis of a library
encompassing a diversity of N-glycan core structures. Effec-
tive enzymatic processing of these core structures to in-
crease the structural diversity of the library could be per-
Microarrays of printed glycans have recently been estab-
lished as powerful tools for the high-throughput screening of
protein–carbohydrate interactions.[1] However,
a serious
handicap for an extensive use of arrays in glycomics studies
continues to be the limited access to pure and well-charac-
terized ligands. Recently, efforts have been made to include
N-glycans, a particularly important class of carbohydrates
accounting for about 90% of protein glycosylation,[2] into
current array formats.[3]
Most ligands on these N-glycan arrays have been isolated
from complex biological matrices in microgram amounts
and are consequently far less well characterized than ligands
on purely synthetic arrays.[4] The efficient tagging of the re-
leased natural glycans with a chromophoric linker, the frac-
[a] Dr. S. Serna, Dr. J. Etxebarria, Dr. N. Ruiz,
Prof. Dr. M. Martin-Lomas, Dr. N.-C. Reichardt
Biofunctional Nanomaterials Unit, CICbiomaGUNE
Paseo Miramon 182, 20009 San Sebastian (Spain)
Fax : (+34)9-430-053-14
E-mail: nreichardt@cicbiomagune.es
[b] Prof. Dr. M. Martin-Lomas
CIBER-BBN, San Sebastian (Spain)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201001295.
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
13163

formed if experimental conditions could be found to carry
out the enzymatic reactions with minimum amounts of ex-
pensive glycosyltransferases and nucleotide donors, simulta-
neously improving the isolation of the final products.
Based on the on-chip chemical synthesis of oligonucleo-
tides and peptides,[15] which paved the way for the wide-
spread use of microarrays in genomics and peptidomics,
Mrksich[16] recently reported the construction of an array of
21 disaccharides by on-chip chemical synthesis. The direct
assembling of the carbohydrates on the biochip presents ob-
vious advantages but the reported method lacks control of
the stereochemistry of glycosidic bond formation and af-
fords a low density of arrays relative to DNA chips.
The use of glycosyltransferases for the synthesis of oligo-
saccharides on solid supports was investigated over 15 years
ago[17] by the group of Wong and later by Blixt and Nor-
berg.[18] Other examples of enzyme catalysis on solid surfa-
ces that have recently been reviewed[19] include the study of
substrate specificities[20] and enzyme activity,[21] the on-chip
modification of mono- and disaccharides[22] on surface-
bound ligands or the extension of glycans on nanoparti-
cles.[23] In these studies, the entire array was subjected to the
action of glycosyltransferases with no possibility of selecting
individual spots. This obviously excludes the acceptor com-
pounds from subsequent studies in the same array. We now
have found that immobilized carbohydrates on high-density
arrays with spot diameters of 100–200 micrometers can be
enzymatically processed with glycosyl transferases in nano-
droplets placed on top of individual spots (Figure 1).
This methodology could open the way for the enzymatic
on-chip construction of glycan arrays with very high ligand
densities by using immobilized synthetic glycan scaffolds
and recombinant glycosyltransferases that install glycosidic
bonds with complete stereose-
Figure 1. Schematic illustration of the nanodroplet method for spotwise
elongation of printed arrays: a) droplet placement on individual spots,
b) incubation and washing step, c) spotwise incubation with second glyco-
syltransferase, and d) finished high-density array.
In this paper, we report on an effective chemoenzymatic
strategy that has allowed the construction, for the first time,
of a focused microarray of synthetic N-glycans. Based on
previously reported modular approaches, a variety of N-
glycan structures have been chemically synthesized and co-
valently immobilized on a glass surface. The synthetic print-
ed structures were then processed with glycosyl transferases
first in well format and then in nanodroplets placed on top
of the surface-bound oligosaccharides by the printing robot.
By using this methodology, 26 new synthetic oligosacchar-
ides have been prepared on the chip from only seven initial
synthetic structures by the action of three enzymes within a
few days. The printed high-mannose structures, which are
not substrates for the enzymes employed, increase the struc-
tural diversity of the array to a total of 39 structures
(Figure 2). To the best of our knowledge, this is the first ap-
plication of recombinant enzymes in the synthesis of large
surface-bound oligosaccharides and the first example of an
enzymatic synthesis of core fucosylated N-glycans.
lectivity. The on-chip elongation
of ligands in nanodroplets
promises the classical advantag-
es of a solid-phase approach:
Easy product clean-up by
simply washing the slide and
higher yields than solution-
phase methods achieved by cy-
cling and the use of excess re-
agents. In addition, the amount
of glycosyl transferases and
donors needed is reduced by
over seven orders of magnitude
(25000000-fold). And as hun-
dreds of reactions can be car-
ried out in parallel, the synthet-
ic throughput is far higher than
for previously reported meth-
ods. For reactions that do not
go to completion, however, the
purity can be lower than for
compounds synthesized and pu-
rified in solution.
Figure 2. N-Glycan structures included on the microarray after enzymatic extension with GalT, SialT and
FucT.
13164
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
Results and Discussion
and deprotection steps. The tert-butyldimethylsilyl ether
(TBS) in 3 was removed with tetrabutylammonium fluoride
(TBAF) and the resulting hemiacetal activated as N-phenyl-
trifluoroacetimidate 4. Coupling 4 with 5[27] under trimethyl-
silyl trifluoromethanesulfonate (TMSOTf) catalysis pro-
duced the trisaccharidic scaffold 6 in 89% yield. The ortho-
gonality of protecting groups was demonstrated by treating
Analysis of the canonical N-glycan biosynthesis structures
shows a limited number of variations for the 3’’- and the 6’’-
arms of the conserved trisaccharide core.[24] With only 10
structurally different glycosyl donors, including a glucosa-
mine donor for the bisecting GlcNAc, all natural N-glycan
core structures of the complex and hybrid type with termi-
nal GlcNAc would be accessible by assembly around a tri-
saccharidic scaffold. Subsequent enzymatic modifications for
terminal galactose and neuraminic acids or core modifica-
tions, such as fucosylation or xylosylation, would increase
substantially the number of structures accessible. As a proof
of concept, N-glycans were synthesized by modular assem-
bly of a trisaccharide scaffold and a panel of glycosyl donors
for substitutions in 3-OH and 6-OH of the central b-manno-
side. The overall strategy is exemplified for a hybrid-type
oligosaccharide in Figure 3.
6
first with 2,3-dichloro-5,6-dicyano-1,4-benzoquinone
(DDQ), which cleaved the naphthylmethyl group (Nap)
while leaving the acetal untouched (7). Treatment of 6
under transacetalisation conditions cleaved the benzylidene
acetal while leaving the Nap group unaltered (8). From 6,
high-mannose, complex and hybrid-type N-glycans differing
in the terminal sugar, type of branching and number of an-
tennae were synthesized by using a choice of glycosyl
donors (Scheme 2).
Donors 12, 13, 15 and 18 were synthesized by a conver-
gent route based on the simultaneous selective reductive
ring opening of dibenzylidene
mannose thioglycosides 10-exo
and 10-endo. The resulting
products were either used di-
rectly as glycosyl donors or
transformed into trichloro- or
N-phenyl-trifluoroacetimidates,
which allowed us to choose the
best reactivity match between
donor and acceptor for each
glycosylation. Donors 11[28] and
Figure 3. Proposed chemoenzymatic synthesis of N-glycans as ligands in microarrays exemplified for bisecting
hybrid-type structures.
14[12a] are known compounds,
whereas 16 and 17 were pre-
pared by following published
Chemical synthesis: The key trisaccharidic scaffold 6 was
prepared in multigram amounts and in high yield as shown
in Scheme 1. Applying methodology developed by Crich[25]
and Kahne[26] the synthetically challenging b-mannopyrano-
sidic linkage was formed with good stereoselectivity and in
excellent yield by the condensation of nearly equimolar
procedures for analogous compounds.[29] From imidate 13,
glycoside 19 was obtained and then deprotected to provide
20 for surface immobilization. The symmetric pentasacchar-
ide 23, an early intermediate in N-glycan biosynthesis and
thus a conserved motif in all N-glycans was prepared by gly-
cosylation of 7 with imidate 11 at a low temperature to give
tetrasaccharide 21 in 80% yield. After acid-catalysed trans-
amounts of sulfoxide
1 and glucosamine acceptor 2
(Scheme 1). A 2-naphthylmethyl ether protection in 3’ pro-
ACHTUNGTRENNUNG
vided stability under acidic conditions in later glycosylation
Scheme 1. a) Tf2O, TTBP, 4 ꢁ MS in CH2Cl2, ꢀ608C; b) 2, ꢀ60 to ꢀ208C, 92%, (a/b 1:5); c) TBAF, AcOH, THF, 08C. d) CF3C
ACHTUNGTERN(UNGN NPh)Cl, K2CO3, ace-
tone, RT, 78% (2 steps); e) TMSOTf, CH2Cl2, 08C to RT, 89%; f) DDQ, CH2Cl2, MeOH, H2O, RT, 71 %; g) pTsOH, EtSH, CH2Cl2, RT, 87%;
h) NH2CH2CH2NH2, nBuOH, MW (1208C); i) Ac2O, pyridine, 08C to RT; j) Na, NH3, THF, ꢀ788C, 52% (3 steps). TTBP: 2,4,6-tri-tert-butylpyrimidine.
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13165

N.-C. Reichardt et al.
highly efficient for the glycosyl-
AHCTUNGTERGaNNUN tion of the 3-arm with a series
of linear and branched struc-
tures as will be shown for fur-
ther examples later in the text.
Hydrolysis of the benzylidene
acetal in 28 afforded 29, which
was treated with trichloroaceti-
midate 14 to produce the com-
plex protected N-glycan struc-
ture 30.
The potential of this chemical
approach was further shown by
making the non-canonical struc-
tures 33 and 35. The 3,6-
branched trisaccharide donors
12, 15 and 18 provide a usual
Scheme 2. Panel of glycosyl donors employed in the assembly of N-glycan structures: a) HOACHTUNRGTNEG(UN CH2)5NHCbz,
TMSOTf, CH2Cl2, 08C to RT, 91%; b) NH2CH2CH2NH2, nBuOH, MW (1208C); c) Ac2O, pyridine, 08C to
RT; d) Na, NH3, THF, ꢀ788C, 72% (3 steps).
(Scheme 3). Condensation of 22 with imidate 12 gave hepta-
saccharide 24, an intermediate in GlcNAc-initiated branch-
ing in the Golgi apparatus[30] (Scheme 3), in a modest yield
of 25%.
Scheme 4. Synthesis of bi- and triantennary complex-type N-glycans 31
and 32: a) TMSOTf, CH2Cl2, RT, 70%; b) EtSH, BF3·OEt2, CH2Cl2,
69%; c) NH2 (CH2)2NH2, nBuOH, MW (1208C); d) Ac2O, pyridine, 08C
to RT; e) Na, NH3, THF, ꢀ788C, 66% for 31, 58% for 32 (3 steps); f) 14,
TMSOTf, CH2Cl2, 08C to RT, 24%.
substitution pattern of the 6-arm in high mannose-type gly-
cans. By using imidate 15, this motif was first introduced on
the 3-arm to form the non-natural homologue 33 in 76%
yield with complete a-selectivity. After acid-catalysed thioly-
sis of the benzylidene acetal, imidate 12 was employed for
the selective glycosylation of OH-6, giving rise to nonasac-
charide 35 (Scheme 5).
Next the assembly of N-glycans of the complex and
hybrid-type was approached. Condensation of 16 with trisac-
charide 7 gave pentasaccharide 38 in 78% yield (Scheme 6).
This compound was scaled up and served as an intermediate
for the synthesis of 40, 41 and 42. Trichloroacetimidate 14
had been reported to be an efficient glycosylating agent in
the construction of similar structures.[10,12a] However, when
39 was treated with 14 under BF3·OEt2 catalysis heptasac-
charide 40 was only isolated in a modest 29% yield
Scheme 3. Synthesis of high-mannose-type N-glycans 25, 26 and 27:
a) TMSOTf, CH2Cl2, ꢀ408C, 80%; b) EtSH, BF3·OEt2, CH2Cl2, 87%;
c) NH2CH2CH2NH2, nBuOH, MW (1208C); d) Ac2O, pyridine, 08C to
RT; e) Na, NH3, THF, ꢀ788C, 56% for 25, 56% for 26, 47% for 27
(3 steps); f) 11, TMSOTf, CH2Cl2, ꢀ408C, 41%; g) 12, TMSOTf, CH2Cl2,
08C to RT, 25%.
The synthesis of triantennary complex structure 30 started
with the condensation of N-phentyltrifluoracetimidate 13
with trisaccharide 7 to give hexasaccharide 28 in 70% yield
(Scheme 4). N-Phentyltrifluoracetimidate donors proved
13166
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
39 was also reacted with imidate 17 affording the complex-
type triantennary structure 41. Again HPLC analysis of the
crude suggested a conversion of more than 50%, the octa-
saccharide, however, could only be isolated in 27% yield.
Last in this series, the hybrid tetraantennary structure 42
was assembled by reacting pentasaccharide 39 with thiogly-
coside 18 since trichloroacetimidate 16 was found to be far
too reactive for this condensation. Activation of 18 was car-
ried out with the stable radical cation tris(4-bromophenyl)-
AHCTUNGTREGaNNNU mmoniumyl hexachloroantimonate in a one-electron trans-
fer reaction as introduced by Sinay[32] and recently used by
Danishefsky.[13c] This activation protocol proved sufficiently
mild to cleanly conjugate the hindered acceptor 39 with the
reactive donor 18 (Scheme 6).
All described compounds were deprotected in a three-
step sequence to produce synthetic ligands for direct spot-
ting on glass slides in good overall yields and with high
purity. The removal of phthalimide protecting groups[12c] was
carried out in n-butyl alcohol and with excess ethylenedi-
Scheme 5. Synthesis of unnatural high-mannose-type N-glycans 36 and
37: a) 15, TMSOTf, CH2Cl2, RT, 76%; b) BF3·OEt2, EtSH, CH2Cl2, 76%;
c) NH2CH2CH2NH2, nBuOH, MW (1208C); d) Ac2O, pyridine, 08C to
RT; e) Na, NH3, THF, ꢀ788C; 39% for 36, 60% for 37 (3 steps); f) 12,
TMSOTf, CH2Cl2, 08C to RT, 33%.
AHCTUNGTREGaNNNU mine by applying microwave heating to the pressure tight
vials. After reacetylation of the free amines and purification,
the intermediate neo-glycoconjugates were obtained in
yields ranging from 61–81% for the two-step procedure.
The remaining benzyl, acetate, and azide protecting groups
were simultaneously removed by Birch reduction producing
the deprotected ligands 9, 20, 25, 26, 27, 31, 36, 37, 43, 44, 45
and 46 in good yields and with high purity.
(Scheme 6). This low yield may be partially attributed to
loss of material during purification as LCMS showed a con-
version of over 50%. The biantennary structure 40 is the
smallest complete structure of complex N-glycans frequently
found in serum glycoproteins and on the cell surface[31] Diol
Enzymatic extension in well format: After complete charac-
terization, the synthetic ligands were printed on NHS-acti-
Scheme 6. Synthesis of bi- and triantennary complex and triantennary hybrid-type N-glycans 44, 45 and 46: a) TMSOTf, CH2Cl2, RT, 78%; b) pTsOH,
CH3CN, 80%; c) NH2CH2CH2NH2, nBuOH, MW (1208C); d) Ac2O, pyridine, 08C to RT; e) Na, NH3, THF, ꢀ788C, 36% for 43, 17% for 44, 35% for
45, 59% for 46 (3 steps); f) 14, BF3·OEt2, CH2Cl2, ꢀ408C, 29%; g) 17, TMSOTf, CH2Cl2, 08C to RT, 27%; h) 18, (BrC6H4)3NSbCl6, CH3CN, 158C–RT,
51%.
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13167

N.-C. Reichardt et al.
vated glass slides and covalently immobilized by amide-
bond formation by following standard procedures by a non-
contact robotic printer. To find generally applicable condi-
tions for the enzymatic processing of the surface-bound
structures, the galactosylation and sialylation of some select-
ed immobilized substrates was first investigated. Elongation
of immobilized N-acetylglucos-
Once the experimental conditions for the enzymatic proc-
essing had been established, a panel of thirteen synthetic li-
gands (see the Supporting Information, Figure S17) was
printed by using 50 mm spotting concentrations and repli-
cates of five in a total of 14 subarrays (see the Supporting
Information, Figure S4). All ligands with terminal glucosa-
amine (GlcNAc), branched tri-
saccharide 20 and hexasacchar-
ide 31 with b-1,4-galactosyl
transferase (b-1,4-GalT) was
studied by function of time of
incubation, ligand surface den-
sity and enzyme concentration.
Probing the arrays with cya-
nine-3-tagged[33] Ricinus com-
munis agglutinin (Cy3-RCA), a
terminal
galactose/galactosa-
mine specific lectin, at different
time points showed saturation
of fluorescence intensities after
less than 4.5 h of incubation for
the GlcNAc derivative, which
suggests an end-point in the en-
zymatic glycosylation. By using
the same conditions, galactosyl-
ACHTUNGTRENNUNGation of the branched trisac-
charide 20 and hexasaccharide
31 was achieved only after 20 h
of incubation. Enzymatic galac-
tosylation of surface-bound car-
bohydrates has been previously
studied for simple monosac-
Figure 4. Enzymatic on-chip galactosylation with bovine b-1,4-galactosyl transferase. Images show eight spots
of a total of 20 replicates for each ligand, all printed at 50 mm. a) Initial array probed with Cy5-BSL II for ter-
minal GlcNAc recognition. b) Control incubation of initial array with Cy3-RCA specific for terminal galactose
residues. c) Array after galactosylation probed with Cy5-BSL II for terminal GlcNAc recognition. d) Array
after galactosylation probed with Cy3-RCA for terminal galactose recognition. e) Fluorescence intensities for
incubation with Cy5-BSL II recognizing terminal GlcNAc residues before (blue columns) and after galactosy-
lation (grey columns). f) Fluorescence intensity for unmodified (grey columns) and enzymatically galactosylat-
ed ligands (green columns) after incubation with Cy3-RCA recognizing terminal GlcNAc residues (histograms
show average values for 20 spots).
charides,[21,34]
disaccharides[16]
and glycopeptides,[22a] but, to
the best of our knowledge, not
on larger structures.
The optimal conditions for
the enzymatic sialylation with
recombinant a-2,6-sialyl transferase (a-2,6-SialT) were ex-
plored in a similar manner on arrays of printed hexasacchar-
ide 31. Cyanine-5-tagged Sambucus nigra agglutinin (Cy5-
SNA), specific for terminal N-acetylneuraminic acid
(Neu5Ac) a-2,6-galactose/galactosamine structures,[35] clear-
ly showed successful sialylation after incubation overnight
but binding of Cy3-RCA to the remaining galactose residues
was only lost completely after a second cycle of enzymatic
extension.
In contrast to enzymatic galactosylation, very few authors
have reported the use of sialyltransferases for the enzymatic
extension of immobilized sugars. By starting from surface-
bound GlcNAc, Shin reported the enzymatic synthesis of
the sialyl LewisX tetrasaccharide by using three recombi-
nant enzymes, including an a-2,3-SialT,[34a] whereas Blixt[20]
determined the substrate specificities of a series of recombi-
nant sialyltransferases on the Consortium of Functional Gly-
comics glycan array.[3a]
mine were printed in quadruplicate rows. Biofunctionality of
the printed ligands was tested with the tagged lectins cya-
nine-3-tagged Canavalia ensiformis agglutinin (Cy3-ConA),
cyanine-5-tagged Wheat germ agglutinin (Cy5-WGA), which
binds strongly to the chitobiose core of N-glycans and inter-
acts specifically with the Man bACTHNUGTRNENUG(1,4)GlcNAc bACHUTNGTREN(NUNG 1,4)GlcNAc
trisaccharide[36] (see the Supporting Information, Figure S5)
and cyanine-5 tagged Griffonia (Bandeiraea) simplicifolia
(Cy5-BSL-II), the latter showing specific affinity for termi-
nal b-linked N-acetylglucosamine-containing oligosacchar-
ides (Figure 4a). Individual subarrays were incubated with a
solution containing b-1,4-GalT and UDP-Gal for 48 h. After
enzymatic elongation, tagged Cy3-RCA[37] was very strongly
bound to the surface carbohydrates confirming the presence
of new galactose residues (Figure 4b,d), whereas binding of
Cy5-BSL-II (Figure 4c,e) to residual GlcNAc was no longer
visible. A second subarray that had been subjected to enzy-
matic galactosylation but not incubation with the tagged lec-
13168
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
fucosyltransferase from Arabi-
dopsis thaliana expressed in
Pichia pastoris, which specifical-
ly glycosylates the 3-OH of the
reducing end GlcNAc residue.
The a-1,3-core fucosylated N-
glycans are immunogenic and
allergenic epitopes specific to
plants and invertebrates.[40]
Only four of the thirteen immo-
bilized synthetic structures were
core fucosylated after overnight
incubation with a-1,3-FucT,
highlighting the tight substrate
specificities of glycosyltransfer-
ases. Successful fucosylation
was determined by probing the
array with the fucose binding
Cy3-tagged lectin from Aleuria
aurantia (Cy3-AAL), which rec-
ognizes
tures[39a] (Figure 6a,b).
Compounds 31 and 32 with
fucosylated
struc-
AHCTUNGTRENNUNG
Figure 5. Enzymatic on-chip sialylation with recombinant human a-2,6-sialyl transferase. Images show eight
spots of 20 replicates for each ligand printed at 50 mm. a) Galactosylated array probed with Cy3-RCA for ter-
minal Gal recognition. b) Control incubation of initial array with Cy5-SNA specific for terminal Neu5Ac resi-
dues. c) Array after sialylation probed with Cy-3-RCA for terminal Gal recognition. d) Array after sialylation
probed with Cy5-SNA for terminal Neu5Ac recognition. e) Fluorescence intensities for galactosylated (green
columns) and enzymatically sialylated ligands (grey columns) after incubation with Cy3-RCA recognizing ter-
minal Gal residues (histograms show average values for 20 spots). f) Fluorescence intensities for galactosylated
(grey columns) and sialylated ligands (red columns) binding to Cy5-SNA, which recognizes terminal Neu5Ac
residues (histograms show average values for 20 spots).
two terminal GlcNAc-residues
on the 3-arm were not sub-
strates for the enzyme. In con-
trast to the enzymatic galacto-
sylation and sialylations, we
could not follow the fucosyla-
tion progress due to lack of a
lectin recognizing specifically
the non-fucosylated starting
tins was then treated with a solution containing a-2,6-SialT
and CMP-Neu5Ac donor for 48 h at 378C. Incubation of the
array with Cy5-SNA[35] showed clearly the success of the on-
surface sialylation procedure (Figure 5d,f).
compounds. To maximize yields for the enzymatic extension,
however, two cycles of 16 h each were applied.
To further increase the number of ligands on the slide to
a total of 39 ligands, fucosylated wells were treated sequen-
tially with GalT and SialT and incubated with Cy3-RCA and
Cy5-SNA recognizing galactosylation and sialylation, respec-
tively (Figure 6c–f, for control experiments see the Support-
ing Information, Figures S8–S10). To our knowledge, this is
the first example of an enzymatic synthesis of core fucosyl-
After sialylation, binding of Cy3-RCA to carbohydrates
was strongly reduced, as shown by fluorescence intensities
below 10% of initial values for the majority of compounds
tested (Figure 5c,e). Stronger residual fluorescence was ob-
served for the sialylated triantennary octasaccharides 59 and
60. Previous studies with a related sialyltransferase from
bovine colostrum have shown a strong preference for galac-
tose residues on the 3-arm and slower sialylation observed
for residues on the 6-arm under the conditions studied.[38]
This could help to explain the residual binding of tagged
RCA after sialylation observed for the triantennary struc-
tures, particularly compound 59. On the other hand, part of
the residual fluorescence intensity observed, might be attrib-
uted to weak binding of Cy3-RCA to internal galactose resi-
dues.[39] Research is underway in our lab to investigate with
the help of mass spectrometry the influence of surface den-
sity, number of antennae and enzyme concentration on the
enzymatic extension, and, in particular, enzymatic sialylation
of immobilized multiantennary N-glycans.
AHCTUNGTREGaNNUN ted N-glycans.
Enzymatic extension in nanodroplets: Next, the enzymatic
processing on individual spots was investigated. By transfer-
ring the enzymatic reaction from a microlitre-based well
format into nanodroplets placed on top of the printed gly-
cans, the amount of glycosyl transferases and donors needed
could be decreased by up to 100000-fold. To avoid rapid
evaporation of droplets, glycerol was added in varying
amounts to the printing solutions. We found that addition of
30% of glycerol to the buffer was sufficient for the forma-
tion of well-defined droplets without compromising the en-
zymatic activity. Droplet volumes of enzyme spotting solu-
tions were optimized to assure complete coverage of the un-
derlying sugar array with the enzyme cocktail (Figure 7).
Sugars containing terminal GlcNAc, were printed in quadru-
As a third enzyme for probing the enzymatic elongation
of surface-bound N-glycans, we chose a recombinant a-1,3-
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13169

N.-C. Reichardt et al.
Figure 6. Enzymatic on-chip fucosylation with a-1,3-fucosyl transferase, b-1,4-galactosyl transferase and a-2,6-sialyl transferase. a) Array after fucosyla-
tion probed with Cy3-AAL for fucose recognition. b) Analysis of the fluorescence values of the fucosylated ligands incubated with 50 nm solution of
Cy3-AAL. c) Array after galactosylation probed with Cy3-RCA for galactose recognition. d) Analysis of the fluorescence values of the galactosylated li-
gands incubated with 50 nm solution of Cy3-RCA. e) Array after sialylation probed with Cy5-SNA for terminal Neu5Ac recognition. f) Analysis of the
fluorescence values of the sialylated ligands incubated with a 50 nm solution of Cy5-SNA (histograms show an average of 20 spots for every ligand, eight
spots for every ligand shown in the images).
all ligands including octasaccharide 32 (see the Supporting
Information, Figure S16).
We expected the sialylation in nanodroplets to be more
difficult due to the reported lower in vitro activity of the a-
2,6-SialT relative to the bovine b-1,4-GalT,[12e] The first re-
sults with HEPES buffer led to disappointing results, even
after extended reaction times of 72 h and increasing enzyme
or nucleotide donor concentrations. Changing the buffer to
Figure 7. a) Fluorescence scan showing spotsize of immobilized glycans.
b) Micrograph showing size of nanodroplets. c) Overlay of a and b dem-
onstrating complete coverage of spots with enzyme-containing nanodrop-
lets and absence of cross-contamination between vicinal spots.
a cacodylate system resulted in a far higher incorporation of
sialic acid residues in nearly all compounds, as demonstrated
by probing the arrays with Cy3-RCA and Cy5-SNA (Fig-
ure 8b).
plicate rows to be able to accommodate both extended and
non-extended compounds in the same subarray and a GalT-
and UDP-Gal-containing solution was placed on top of se-
lected individual spots by the robotic printer. After 48 h of
incubation, all droplets had maintained their size and posi-
tion, due to the low volatility of glycerol drops on glass
slides. The subsequent incubation of subarrays with a mix-
ture of Cy5-BSL-II and Cy3-RCA showed loss of binding to
residual GlcNAc and intense binding to newly incorporated
galactose residues for all treated structures (Figure 8a),
except for octasaccharide 32, which showed 17% of residual
fluorescence for binding to Cy5-BSL-II. In an attempt to
drive reactions to completion for all ligands, we incubated
an array containing octasaccharides with twice the enzyme
concentration. After an incubation of 48 h, we observed
complete loss of BSL-II binding to any residual GlcNAc for
Octasaccharides 52 and 53 were sialylated in around 80%
yield as suggested by residual fluorescence values, very simi-
lar to the results obtained in the sialylation in well-formats.
Incubation with tagged ECA (Erythrina cristagalli aggluti-
nin; Figure 8b, lower row), which is reported not to tolerate
6-O-sialyation of galactose,[39b] showed some residual bind-
ing as well, confirming incomplete sialylation of the trian-
tennary structures under the tested conditions. Finally, we
proceeded with the spotwise fucosylation of ligands, which
had been shown to be substrates for the a-1,3-FucT in previ-
ous experiments. Enzymatic extension in nanodroplets re-
sulted in a clearly visible core fucosylation of the tested sub-
strates 43, 44, 45 and 46 giving rise to compounds 61, 63, 64
and 62 as seen by probing with Cy3-AAL. Additional incu-
bation cycles with fucosyltransferase did not increase the
fluorescence intensities (Figure 8c).
13170
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
Figure 8. Spotwise enzymatic glycosylation by glycosyltransferases: a) Spot by spot galactosylation of glycan array with GalT. Upper row: Cy3-RCA
showing saturated signals for terminal galactose. Lower row: Cy5-BSL-II to visualize residual GlcNAc. b) Spotwise sialylation of glycan array with a-2,6-
SialT. Upper row: Cy5-SNA showing saturated signals for terminal sialic acid. Middle row: Array probed with Cy3-RCA to visualize residual galactose.
Lower row: Array probed with biotinylated ECA/Cy3-Neutravidine to visualize residual galactose. c) Spot by spot fucosylation with a-1,3-FucT probed
with Cy3-AAL.
Conclusion
40 well-defined N-glycans in a few days. Start and endpoints
of enzymatic reactions were defined where possible by incu-
bating untreated and processed subarrays with fluorescently
tagged lectins that show an affinity for a specific terminal
sugar, a validation method that has been used by others be-
fore.[20,21,34a]
Microarrays focused on specific classes of carbohydrates are
more easily accomplished than more heterogeneous glycan
arrays: The synthetic effort is minimized by taking advant-
age of modular strategies, whereas structural similarity of li-
gands permits the general application of class-specific trans-
ferases exhibiting relatively tight substrate specificities. Ac-
cording to this working hypothesis, we have prepared the
first microarray of fully synthetic N-glycan structures by
using a chemoenzymatic approach. A modular chemical syn-
thesis based on a versatile trisaccharidic scaffold has afford-
ed a series of natural and non-natural high-mannose, com-
plex and hybrid-type N-glycans functionalized with an
anomeric spacer. Overall short synthetic routes and general-
ly good yields combined with a straightforward deprotection
protocol have afforded pure compounds on a scale suited
for microarray preparation. The low yields encountered for
some selective glycosylations are most probably due to mis-
matched reactivities of donor and acceptor and to the diffi-
cult purification of compound mixtures of similar polarity.
After immobilization on glass slides, these synthetic core
structures were extended by enzymatic reactions first in
wells and then, for the first time, in nanodroplets placed on
top of individual spots. This has permitted us to reduce the
amount of precious glycosyltransferases by several orders of
magnitude relative to reactions in solution and in well-for-
mats. Three recombinant glycosyltransferases have been em-
ployed to extend an array of 13 synthetic ligands to nearly
In any case, we are well aware that a general use of
tagged lectin pairs for monitoring conversion of enzymatic
reactions is limited and we are currently working on the
direct analysis of spot compositions by mass spectrometry
techniques to confirm the high conversion rates we observed
by probing the arrays with lectins. As glycosyltransferases
have narrow substrate specificities, some ligands might not
be completely glycosylated, even after repeated cycles when
subjecting the entire array to the enzymatic extension condi-
tions.[20–21] To avoid compound mixtures coming from incom-
plete reactions. the usefulness of a glycosyltransferase would
have to be established for every acceptor individually and
ideally confirmed by a mass spectrometric readout of spot
compositions. For this reason, we propose the enzymatic ex-
tension in well format as a first screen for potential accept-
ors of a particular glycosyltransferase to test its synthetic
usefulness. For the fabrication of arrays, the enzymatic ex-
tension in nanodroplets could then be used to produce pure
ligands on the chip, choosing only spots occupied by validat-
ed acceptors of the enzyme used. This feature of the nano-
droplet approach allows the preparation of arrays of higher
diversity and density than by classical enzymatic on-chip
elongation.
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13171

N.-C. Reichardt et al.
(642 mg, 78% over two steps). Rf =0.16 (hexane/EtOAc 3:1); 1H NMR
(500 MHz, CDCl3 at 508C): d=7.93–7.67 (m, 8H; Ar), 7.62–6.89 (m,
23H; Ar), 6.74 (d, J=7.3 Hz, 1H; Ar), 6.35 (s, 1H; H-1), 5.61 (s, 1H;
CHPh), 5.02–4.94 (m, 4H; CHaHa’Ar, CHbHb’Ar), 4.85 (d, J=12.6 Hz,
1H; CHbHb’Ar), 4.67–4.62 (m, 2H; CHcHc’Ar, H-1’), 4.52–4.37 (m, 4H;
CHaHa’Ar, CHcHc’Ar, H-2, H-3), 4.28–4.16 (m, 3H; H-4, H-4’, H-6a’),
3.88 (d, J=2.4 Hz, 1H; H-2), 3.75–3.41 (m, 5H; H-3’, H-6b’, H-5, H-6a,
H-6b), 3.24 ppm (td, J=4.8, 9.7 Hz, 1H; H-5’); 13C NMR (126 MHz,
CDCl3 at 508C): d=167.47 (CO), 143.06, 138.64, 137.78, 137.57, 136.06
(qC, Ar), 133.82 (CH, Ar), 133.36, 133.00, 131.56 (qC, Ar), 128.77,
128.54, 128.45, 128.23, 128.14, 128.10, 127.99, 127.94, 127.81, 127.76,
127.66, 127.63, 127.53, 126.93, 126.16, 126.04, 125.78, 125.44, 124.36,
123.32, 119.40 (CH, Ar), 115.93 (q, J=284.5 Hz; CF3), 101.63, 101.50 (C-
1’, CHPh), 93.66 (C-1), 78.78, 78.50, 78.44 (C-3’, C-4, C-4’), 77.18 (C-2’),
76.78 (C-3), 75.64 (C-5), 75.14, 74.64, 73.53, 72.58 (CH2Ar), 68.55, 68.05
(C-6’, C-6), 67.44 (C-5’), 54.89 ppm (C-2); HRMS (ESI): m/z: calcd for
C67H59F3N2O12Na: 1163.3917 [M+Na]+; found: 1163.3882.
The methodology here described could open the way for
the rapid production of large synthetic glycan arrays. How-
ever, alternative methods for spot analysis have to be devel-
oped to confirm enzymatic conversions. Work in process in-
cludes the investigation of these analytical methods and the
preparation of a large and representative N-glycan array by
increasing both the number of synthetic core structures and
available glycosyl transferases.
Experimental Section
General methods: General methods and spectroscopic data for the syn-
thesis of the oligosaccharides, commercial sources and detailed informa-
tion for the microarray experiments can be found in the Supporting In-
formation.
5-Azidopentyl 2-O-Benzyl-4,6-O-benzylidene-3-O-(2-naphthylmethyl)-b-
d-mannopyranosyl-(1!4)-3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-
glucopyranosyl-(1!4)-3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-gluco-
tert-Butyldimethylsilyl
methyl)- b-d-mannopyranosyl-(1!4)-3,6-di-O-benzyl-2-deoxy-2-phthal-
imido-b-d-glucopyranoside (3): Tf2O (200 mL, 1.18 mmol) was added to a
stirred solution of donor (714 mg, 1.18 mmol), TTBP (539 mg,
2-O-benzyl-4,6-O-benzylidene-3-O-(2-naphthyl-
pyranoside
(6):
Trimethylsilyltrifluoromethanesulfonate
(4.2 mL,
0.023 mmol) was added to a mixture of donor 4 (631 mg, 0.55 mmol), ac-
ceptor 5 (277 mg, 0.46 mmol) and activated molecular sieves 4 ꢁ in dry
CH2Cl2 (0.9 mL). The reaction mixture was stirred at room temperature
for 40 min and then quenched by the addition of Et3N (100 mL), diluted
with CH2Cl2, filtered over Celite and concentrated. The crude product
was purified by flash chromatography (hexane/EtOAc 3:2) to give the
title compound as a white foam (636 mg, 89%). Rf =0.27 (hexane/EtOAc
3:2); [a]2D0 =+9.1 (c=0.12 in CHCl3); 1H NMR (500 MHz, CDCl3): d=
7.92–7.67 (m, 12H; Ar), 7.53–7.15 (m, 23H; Ar), 7.00–6.96 (m, 4H; Ar),
6.90–6.89 (m, 3H; Ar), 6.80–6.79 (m, 3H; Ar), 5.55 (s, 1H; CHPh), 5.30
(d, J=8.1 Hz, 1H; H-1), 4.98–4.83 (m, 6H; H-1’, CH2Ar, CHaHa’Ar,
CHbHb’Ar, CHcHc’Ar), 4.76 (d, J=12.8 Hz, 1H; CHaHa’Ar), 4.63–4.41 (m,
6H; H-1’’, CHbHb’Ar, CHcHc’Ar, CH2Ar, CHdHd’Ar), 4.36–4.05 (m, 9H;
CHdHd’Ar, H-2, H-2’, H-3, H-3’, H-4, H-4’, H-4’’, H-6a’’), 3.81 (d, J=
ACHTUNGTRENNUNG
1
2.18 mmol) and 4 ꢁ activated molecular sieves in dry CH2Cl2 (10 mL) at
ꢀ608C under argon atmosphere. After five minutes at ꢀ608C, a solution
of acceptor 2 (550 mg, 0.91 mmol) in dry CH2Cl2 (3 mL) by syringe. The
reaction mixture was allowed to warm to ꢀ208C over 1h and was then
quenched by the addition of aqueous saturated NaHCO3 solution
(75 mL), filtered over a plug of Celite and extracted with CH2Cl2 (3ꢂ
75 mL). The combined organic layers were washed with brine, dried over
anhydrous MgSO4, filtered and concentrated. The crude product was pu-
rified by medium-pressure flash chromatography (toluene/EtOAc 100:0
to 96:4) to yield the b-anomer as a white foam (760 mg, 71%). Rf =0.32
(toluene/EtOAc 19:1); [a]2D0 =+4.0 (c=0.21 in CHCl3); 1H NMR
(500 MHz, CDCl3): d=7.90–7.70 (m, 8H; Ar), 7.60–7.24 (m, 18H; Ar),
7.05–7.04 (m, 2H; Ar), 6.97–6.91 (m, 3H; Ar), 5.64 (s, 1H; CHPh), 5.44
(d, J=8.1 Hz, 1H; H-1), 5.02–4.94 (m, 4H; CH2Ar, CHaHa’Ar,
CHbHb’Ar), 4.84 (d, J=12.7 Hz, 1H; CHbHb’Ar), 4.72–4.62 (m, 2H;
CHcHc’Ar, H-1’), 4.58 (d, J=12.4 Hz, 1H; CHaHa’Ar), 4.46 (d, J=
12.1 Hz, 1H; CHcHc’Ar), 4.39 (dd, J=8.6, 10.8 Hz, 1H; H-3), 4.29 (dd,
J=4.7, 10.4 Hz, 1H; H-6a’), 4.25–4.19 (m, 2H; H-4’, H-2), 4.13 (t, J=
9.2 Hz, 1H; H-4), 3.92 (d, J=2.6 Hz, 1H; H-2’), 3.73–3.55 (m, 5H; H-6a,
H-6b, H-6b’, H-3’, H-5), 3.29 (td, J=4.9, 9.7 Hz; 1H; H-5’), 0.76 (s, 9H;
tBu), 0.13 (s, 3H; CH3), ꢀ0.00 ppm (s, 3H; CH3); 13C NMR (126 MHz,
CDCl3): d=168.20, 167.50 (CO), 138.77, 138.54, 137.71, 137.58, 135.88
(qC, Ar), 133.59 (CH, Ar), 133.17, 132.81, 131.52 (qC, Ar), 128.74,
128.32, 128.19, 128.09, 127.93, 127.77, 127.69, 127.54, 127.46, 126.80,
126.05, 125.98, 125.92, 125.72, 125.33, 123.05, 122.91 (CH, Ar), 101.87 (C-
1’, JC1ꢀH1 =158.7 Hz), 101.31 (CHPh), 93.36 (C-1), 79.46 (C-4), 78.58 (C-
3’), 78.17 (C-4’), 76.86 (C-2’), 76.67 (C-3), 74.94 (CH2Ar), 74.65 (C-5),
74.32, 73.35, 72.30 (CH2Ar), 68.60, 68.48 (C-6, C-6’), 67.26 (C-5’), 57.79
(C-2), 25.24 (CH3, tBu), 17.45 (qC, tBu), ꢀ4.31 (CH3), ꢀ5.54 ppm (CH3);
HRMS: m/z: calcd for C65H69NO12SiNa: 1106.4487 [M+Na]+; found:
1106.4537.
2.2 Hz, 1H; H-2’’), 3.70 (dt, J=6.2, 9.9 Hz, 1H; OCHHACTHNUTRGNE(UNG CH2)4N3), 3.63–
3.55 (m, 3H; H-6a, H-6a’, H-6b’’), 3.50 (dd, J=2.7, 9.9 Hz, 1H; H-3’’), 3.46
(dd, J=3.6, 11.1 Hz, 1H; H-6b’), 3.40 (dd, J=1.9, 10.9 Hz, 1H; H-6b),
3.34 (dd, J=2.4, 9.5 Hz, 1H; H-5’), 3.28 (dt, J=6.6, 9.8 Hz, 1H; OCHH-
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
1.18–1.02 ppm (m, 2H; CH2); 13C NMR (126 MHz, CDCl3): d=168.40,
167.49 (CO), 138.80, 138.57, 138.42, 137.69, 137.60, 135.91 (qC, Ar),
133.92, 133.72, 133.55 (CH, Ar), 133.21, 132.86, 131.71, 131.59, 131.40
(qC, Ar), 128.77, 128.38, 128.24, 128.14, 128.11, 127.98, 127.81, 127.72,
127.65, 127.60, 127.55, 127.23, 126.83, 126.78, 126.08, 126.03, 126.00,
125.79, 125.40, 123.55, 123.04 (CH, Ar), 101.73 (C-1’’), 101.34 (CHPh),
98.04 (C-1’), 97.03 (C-1), 79.19, 78.56, 76.96, 76.93, 76.71, 75.83 (C-2, C-2’,
C-3, C-3’, C-4, C-4’, C-2’’), 78.18 (C-3’’), 75.07, 74.60 (CH2Ar), 74.48 (C-5,
C-5’), 74.31, 73.11, 72.59, 72.35 (CH2Ar), 68.80 (OCH
2ACHTUNGTNERN(UNG CH2)4N3), 68.47
(C-6’’), 68.18 (C-6’), 67.85 (C-6), 67.23 (C-5’’), 56.51 (C-2), 55.66 (C-2’),
51.00 (OACTHNUGRTNE(UNG CH2)4CH2N3), 28.59, 28.19, 22.92 ppm (CH2 linker); HRMS
(ESI): m/z: calcd for C92H89N5O18Na: 1574.6100 [M+Na]+; found:
1574.6102.
5-Azidopentyl 2-O-benzyl-4,6-O-benzylidene-b-d-mannopyranosyl-(1!
4)-3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-glucopyranosyl-(1!4)-3,6-
2-O-Benzyl-4,6-O-benzylidene-3-O-(2-naphthylmethyl)-b-d-mannopyra-
nosyl-(1!4)-3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-glucopyranosyl
N-phenyl trifluoroacetimidate (4): TBAF (1m in THF, 1.2 mL,
1.16 mmol) and acetic acid (66 mL, 1.16 mmol) were added to a solution
of compound 3 (785 mg, 0.72 mmol) in dry THF (3.6 mL) at 08C and the
resulting mixture was stirred for 2 h at 08C. The reaction was quenched
with saturated NaHCO3 solution (75 mL) and extracted with EtOAc (3ꢂ
50 mL). The combined organic layers were washed with brine, dried over
anhydrous MgSO4, filtered and concentrated. The crude product was dis-
di-O-benzyl-2-deoxy-2-phthalimido-b-d-glucopyranoside
(7):
2,3-Di-
chloro-5,6-dicyano-p-benzoquinone (153 mg, 0.68 mmol) was added in
one portion to a stirred solution of compound 6 (350 mg, 0.22 mmol) in
wet CH2Cl2/MeOH (4:1, 1.1 mL). The resulting mixture was stirred for
1 h, diluted with CH2Cl2 (50 mL), washed with saturated NaHCO3 and
extracted with CH2Cl2 (2ꢂ100 mL). The combined organic layers were
washed with brine, dried over anhydrous MgSO4, filtered and concentrat-
ed. The crude product was purified by medium-pressure flash chromatog-
raphy (hexane/EtOAc 9:1 to 1:4) to give the title compound (220 mg,
71%). Rf =0.22 (hexane/EtOAc 3:2); [a]2D0 =+3.5 (c=0.062 in CHCl3);
1H NMR (500 MHz, CDCl3): d=7.90–7.69 (m, 8H; Ar), 7.51–7.24 (m,
20H; Ar), 7.06–6.99 (m, 2H; Ar), 6.97–6.95 (m, 2H; Ar), 6.92–6.88 (m,
3H; Ar), 6.82–6.81 (m, 3H; Ar), 5.45 (s, 1H; CHPh), 5.31 (d, J=7.6 Hz,
solved in acetone (7.2 mL) and then ClCACTHNUTRGNEG(NU NPh)CF3 (299 mg, 1.44 mmol)
and K2CO3 (199 mg, 1.44 mmol) were added and the mixture stirred at
room temperature overnight. The reaction mixture was diluted with
EtOAc, filtered over Celite and the solvent was evaporated. The crude
product was purified by medium-pressure flash chromatography (hexane/
EtOAc 94:6 to 60;40) to give the title compound as a white foam
13172
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
1H; H-1), 5.02 (d, J=11.6 Hz, 1H; CHaHa’Ar), 4.97 (d, J=7.8 Hz, 1H;
H-1’), 4.94–4.88 (m, 2H; CHbHb’Ar, CHcHc’Ar), 4.71–4.69 (m, 2H; H1’’,
CHaHa’Ar), 4.63 (d, J=12.0 Hz, 1H; CHdHd’Ar), 4.57–4.54 (m, 3H;
CHcHc’Ar, CH2Ar), 4.48 (d, J=12.1 Hz, 1H; CHdHd’Ar), 4.42 (d, J=
12.2 Hz, 1H; CHbHb’Ar), 4.32–4.09 (m, 7H; H-2, H-2’, H-3, H-3’, H-4, H-
4’, H-6a’’), 3.75–3.66 (m, 4H; H-2’’, H-4’’, H-6a, OCHHACHTUNGRTNEUNG(CH2)4N3), 3.62–
3.57 (m, 2H; H-3’’, H-6a’), 3.54–3.44 (m, 3H; H-6b’’, H-6b, H-6b’), 3.35
(dd, J=3.2, 9.9 Hz, 1H; H-5’), 3.28 (dt, J=6.6, 10.0 Hz, 1H; OCHH-
dried over anhydrous MgSO4, filtered and evaporated. The crude product
was purified by flash column chromatography (hexane/EtOAc 1:1) to
afford the title compound as a transparent oil (77 mg, 80%). 1H NMR
(500 MHz, CDCl3): d=7.90–7.63 (m, 12H; Ar), 7.52 (m, 2H; Ar), 7.40–
7.21 (m, 14H; Ar), 7.01 (m, 2H; Ar), 6.91–6.85 (m, 3H; Ar), 6.81–6.79
(m, 4H; Ar), 5.62 (dd, J=9.4, 10.6 Hz, 1H; H-3E), 5.28 (d, J=8.3 Hz,
1H; H-1A), 5.20 (s, 1H; H-1D), 5.12 (d, J=8.3 Hz, 1H; H-1E), 5.08 (dd,
J=3.0, 9.8 Hz, 1H; H-4D), 5.61–4.98 (m, 4H; H-4E, CHaHa’Ar, CH2’Ar),
4.95 (d, J=7.6 Hz, 1H; H-1B), 4.89 (m, 2H; CHbHb’Ar, CHcHc’Ar), 4.69
(d, J=12.4 Hz, 1H; CHaHa’Ar), 4.62–4.47 (m, 4H; CH2’Ar, CHbHb’Ar, H-
1C), 4.34–4.26 (m, 2H; H-2E, CHcHc’Ar), 4.23–4.11 (m, 8H; H-2A, H-2D,
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
1.46–1.26 (m, 4H; 2ꢂCH2), 1.15–1.03 ppm (m, 2H; CH2); 13C NMR
(126 MHz, CDCl3): d=168.45, 167.50 (CO), 138.72, 138.60, 138.45,
138.14, 137.66, 137.21 (qC, Ar), 133.97, 133.76, 133.60 (CH, Ar), 131.72,
131.62, 131.39 (qC, Ar), 129.01, 128.55, 128.47, 128.17, 128.14, 127.99,
127.95, 127.90, 127.85, 127.77, 127.51, 127.24, 126.92, 126.83, 126.22,
123.59, 123.07 (CH, Ar), 101.88 (CHPh, C-1’’), 98.07 (C-1’), 97.02 (C-1),
79.32, 79.11, 78.92, 77.04, 76.75, 75.81 (C-2’’, C-4’’, C-3, C-3’, C-4, C-4’),
75.75, 74.61 (CH2Ar), 74.55, 74.49 (C-5, C-5’), 74.32, 73.36, 72.59
H-2B, H-3 A, H-3B, H-3D, H-4B, H-6aE), 4.05 (t, J=9.9 Hz, 1H; H-4A), 3.99
C
(t, J=9.9 Hz, 1H; H-4C), 3.92 (m, 1H; H-6bE), 3.81–3.67 (m, 5H; H-6a
,
H-6aD, H-2C, H-6aA, OCHH
6bD, H-6bA), 3.45–3.41 (m, 2H; H-3C, H-6bB), 3.34 (m, 2H; H-5B, H-5D),
3.26 (m, 2H; H-5A, OCHH(CH2)4N3), 3.08 (m, 1H; H-5C), 2.93–2.85 (3H,
m; H-5E, O
(CH2)4CH2N3), 2.11, 2.06, 2.03, 2.00, 1.88, 1.86 (s, 18H; 6ꢂ
(CH2)4N3), 3.61–3.52 (m, 4H; H-6bC, H-6aB, H-
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
AHCTUNGTRENNUNG
OCH3), 1.40–1.27 (m, 4H; 2ꢂCH2), 1.12–1.06 ppm (m, 2H; CH2);
13C NMR (126 MHz, CDCl3): d=170.7, 170.6, 170.1, 170.0, 169.3, 169.2,
168.4, 167.8, 167.7, 167.6, 138.7, 138.5, 137.7, 134.2, 134.1, 133.8, 133.6,
132.5, 131.7, 131.6, 131.5, 131.4, 130.9, 128.8, 128.6, 128.5, 128.4, 128.2,
128.1, 128.0, 127.9, 127.8, 127.6, 127.3, 127.2, 127.0, 126.9, 126.6, 123.6,
123.5, 123.1, 123.0, 100.8 (C-1C, JC1ꢀH1 =158.5 Hz), 98.1, 97.0, 96.4 (C-1A,
C-1B , C-1E), 95.7 (C-1D, JC1ꢀH1 =170.2 Hz), 81.3, 78.6, 76.0, 75.5, 74.5,
74.5, 74.4, 74.3, 73.4, 72.7, 71.3, 70.2, 69.5, 68.9, 68.7, 68.6, 68.2, 67.5, 65.8,
65.5, 62.9, 62.5, 61.6, 60.4, 56.5, 55.7, 54.2, 51.0, 30.3, 28.9, 28.6, 28.3, 23.7,
23.0, 20.8, 20.7, 20.6, 20.5 ppm; HRMS: m/z: calcd for C106H112N6O35Na:
2051.7065 [M+Na]+; found: 2051.5115.
(CH2Ar), 70.88 (C-3’’), 68.84 (OCH
2ACTHNUGTREN(NUNG CH2)4N3), 68.42 (C-6’’) , 68.22 (C-
6’), 67.74 (C-6), 66.80 (C-5’’), 56.48 (C-2), 55.69 (C-2’), 51.03 (O-
ACHTUNGTRENNUNG(CH2)4CH2N3), 28.61, 28.21, 22.94 ppm (CH2 linker); HRMS (ESI): m/z:
calcd for C81H81N5O18Na: 1434.5475 [M+Na]+; found: 1434.5571.
5-Azidopentyl O-[3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-b-d-glucopyr-
anosyl-(1!2)-3,4,6-tri-O-acetyl-b-d-mannopyranosyl-(1!3)-O-(2-O-
benzyl-4,6-O-benzylidene-b-d-mannopyranosyl)-(1!4)-O-(3,6-di-O-
benzyl-2-deoxy-2-phthalimido-b-d-glucopyranosyl)-(1!4)-O-3,6-di-O-
benzyl-2-deoxy-2-phthalimido-b-d-glucopyranoside (38): Trimethylsilyltri-
fluoromethanesulfonate (1.7 mL, 0.01 mmol) was added to a stirred mix-
ture of donor 16 (100 mg, 0.11 mmol), acceptor 7 (135 mg, 0.10 mmol)
and activated molecular sieves 4 ꢁ in dry CH2Cl2 (2 mL) and the reaction
mixture was stirred at room temperature for 40 min. The reaction was
quenched by adding Et3N (100 mL), diluted with CH2Cl2, filtered over
Celite and evaporated. The crude product was purified by flash column
chromatography (hexane/EtOAc 1:1) to give the title compound as clear
oil (159 mg, 78%). [a]D20 =ꢀ19.9 (c=1.0 in CH2Cl2); 1H NMR (500 MHz,
CDCl3): d=7.90–7.86 (m, 3H; Ar), 7.75–7.69 (m, 9H; Ar), 7.56–7.44 (m,
3H; Ar), 7.43–7.22 (m, 16H; Ar), 7.01 (m, 1H; Ar), 6.92–6.89 (m, 4H;
Ar), 6.82–6.80 (m, 2H; Ar), 5.47 (dd, J=9.0, 11.0 Hz, 1H; H-3E), 5.43 (s,
1H; CHPh), 5.30 (d, J=7.8 Hz, 1H; H-1A), 5.05 (dd, J=3.4, 10.4 Hz,
1H; H-4D), 5.01–4.84 (m, 6H; H-1D, H-1B, CHaHa’Ar, CHbHb’Ar, H-4E,
H-1E), 4.79 (s, 2H; CH2Ar), 4.66 (d, J=12.8 Hz, 1H; CHcHc’Ar), 4.55–
4.49 (m, 3H; CH2Ar, CHaHa’Ar), 4.44 (s, 1H; H-1C), 4.39 (d, J=12.8 Hz,
1H; CHbHb’Ar), 4.33 (d, J=12.8 Hz, 1H; CHcHc’Ar), 4.29–4.08 (m, 9H;
H-2A, H-2B, H-2E, H-3A, H-3B, H-3D, H-4A, H-4B, H-6bC), 3.96 (dd, J=4.1,
12.2 Hz, 1H; H-6aE), 3.92 (t, J=10.1 Hz,1H; H-4C), 3.69–3.66 (m, 3H;
5-Azidopentyl O-[3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-b-d-glucopyr-
anosyl-(1!2)-3,4,6-tri-O-acetyl-a-d-mannopyranosyl-(1!3)]-O-{(2-O-
acetyl-3,4,6-tri-O-benzyl-a-d-mannopyranosyl)-(1!3)-O-(2-O-acetyl-
3,4,6-tri-O-benzyl-a-d-mannopyranosyl)-(1!6)-3,6-di-O-benzyl-a-d-man-
nopyranoside-(1!6)}-O-(2-O-benzyl-b-d-mannopyranosyl)-(1!4)-O-
(3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-glucopyranosyl)-(1!4)-O-
3,6-di-O-benzyl-2-deoxy-2-phthalimido-b-d-glucopyranoside (42): A mix-
ture of donor 18 (79.5 mg, 56.7 mmol) with activated 4 ꢁ molecular sieves
in dry CH3CN (0.4 mL) was stirred at room temperature for 20 min. The
resulting mixture was cooled to 158C, tris(4-bromophenyl)ammoniumyl
hexachloroantimonate (TBAP; 88.5 mg, 108.4 mmol) was added followed
by a solution of acceptor 39 (58 mg, 28.3 mmol) in CH3CN (0.3 mL). The
reaction mixture was stirred at 158C for 40 min and then a second por-
tion of TBAP (21 mg, 26.4 mmol) was added. The reaction mixture was
allowed to warm to room temperature and was then stirred overnight.
The solvents were evaporated and the crude product was purified by
flash column chromatography (hexane/EtOAc 1:1 to 1:2) to give the title
compound as a white solid (48 mg, 51%). Rf =0.11 (hexane/EtOAc 1:1);
[a]2D0 =+17.8 (c=1.21 in CHCl3); 1H NMR (500 MHz, CDCl3): d=7.89–
6.59 (m, 137H), 5.72 (dd, J=10.7, 9.1 Hz, 1H), 5.54–5.45 (m, 2H), 5.36
(d, J=8.5 Hz, 1H), 5.25 (d, J=7.8 Hz, 1H), 5.20–5.05 (m, 4H), 5.03–4.77
(m, 9H), 4.76–4.35 (m, 18H), 4.33–4.07 (m, 8H), 4.06–3.78 (m, 13H),
3.76–3.36 (m, 14H), 3.35–3.20 (m, 4H), 3.10 (dt, J=9.4, 3.4 Hz, 1H),
2.97–2.81 (m, 2H), 2.10 (s, 3H), 2.08 (s, 3H), 2.07 (s, 3H), 2.04 (s, 3H),
2.02 (s, 3H), 1.98 (s, 3H), 1.92 (s, 3H), 1.87 (s, 3H), 1.42–1.26 (m, 4H),
1.13–1.03 ppm (m, 2H); 13C NMR (126 MHz, CDCl3): d=170.69, 170.39,
170.34, 170.21, 170.07, 170.04, 169.34, 169.18, 168.22, 167.39, 138.73,
138.63, 138.52, 138.41, 138.25, 138.09, 137.91, 137.74, 134.20, 133.58,
131.40, 129.73, 128.98, 128.86, 128.59, 128.36, 128.30, 128.26, 128.21,
128.14, 128.09, 128.01, 127.76, 127.65, 127.58, 127.52, 127.49, 127.39,
OCHH
3C, H-2D, H-5D, H-6aB, H-6bB, H-6bC), 3.57–3.33 (m, 1H; H-5B), 3.27–3.24
(2H, m; H-5A, OCHH(CH2)4N3), 2.98 (m, 1H; H-5C), 2.88 (m, 2H; O-
(CH2)4CH2N3), 2.07, 2.06, 2.02, 2.01, 1.86 (s, 18H; 6ꢂOCH3), 1.40–1.22
ACHTUNGTRENNUNG
(CH2)4N3, H-6bE, H-6aA), 3.57–3.40 (m, 9H; H-2C, H-6aD, H-6bD, H-
ACHTUNGTRENNUNG
ACHTUNGTRENNUNG
(m, 4H; 2ꢂCH2), 1.10–1.05 ppm (m, 2H; CH2); 13C NMR (126 MHz,
CDCl3): d=170.5, 170.4, 170.3, 170.1, 169.2, 169.1, 168.5, 167.9, 167.8,
167.7, 167.5, 138.8, 138.7, 138.5, 138.1, 137.7, 137.6, 134.1, 134.0, 133.8,
133.6, 131.7, 131.6, 131.5, 131.4, 131.3, 130.2, 129.8, 129.0, 128.9, 128.7,
128.6, 128.5, 128.4, 128.2, 128.0, 127.9, 127.8, 127.7, 127.5, 127.4, 127.3,
127.2, 127.0, 126.9, 126.8, 123.6, 123.5, 123.1, 102.4, 100.9 (C-1C, JC1ꢀH1
=
158.0 Hz), 98.2 (C-1D, JC1ꢀH1 =176.1 Hz), 98.1, 97.2, 95.5, 78.5, 78.1, 77.8,
76.1, 75.0, 74.6, 74.5, 73.5, 72.7, 70.9, 70.3, 69.2, 68.9, 68.5, 68.4, 68.2, 67.5,
66.6, 66.0, 62.9, 61.0, 56.5, 55.7, 54.1, 51.0, 28.6, 28.2, 23.0, 20.7, 20.6, 20.5,
20.4 ppm; HRMS: m/z: calcd for C113H116N6O35Na: 2140.7412 [M+Na]+;
found: 2140.7908.
127.31, 127.25, 127.06, 126.84, 123.47, 123.10, 101.60 (C-1, JC1ꢀH1
158.3 Hz), 99.54 (C-1, JC1ꢀH1 =170.9 Hz), 98.31 (C-1, JC1ꢀH1 =171.9 Hz),
98.08 (C-1, C1ꢀH1 =172.4 Hz), 98.02 (C-1), 97.09 (2ꢂC-1, JC1ꢀH1
=
5-Azidopentyl O-[3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-b-d-glucopyr-
anosyl-(1!2)-3,4,6-tri-O-acetyl-a-d-mannopyranosyl]-(1!3)-O-(2-O-
Benzyl-b-d-mannopyranosyl)-(1!4)-O-(3,6-di-O-benzyl-2-deoxy-2-
J
=
170.1 Hz), 96.71 (C-1), 79.51, 78.19, 78.07, 77.52, 76.01, 75.04, 74.89,
74.73, 74.61, 74.52, 74.45, 74.37, 74.19, 74.11, 74.05, 73.43, 73.30, 73.25,
73.17, 72.61, 72.18, 72.01, 71.74, 71.53, 71.45, 71.11, 70.55, 69.72, 68.83,
68.74, 68.65, 68.30, 68.15, 66.17, 65.86, 62.68, 61.64, 56.49, 55.70, 54.38,
51.06, 28.64, 28.23, 22.96, 21.08, 21.00, 20.75, 20.67, 20.58, 20.47,
20.43 ppm; HRMS (ESI): m/z: calcd for C184H192N6O52Na2: 1683.6292
[M+2Na]+2; found: 1683.6140.
phthal
G
phthalACHTUNGTRENNUNGimido-b-d-glucopyranoside (39): p-Toluenesulfonic acid monohy-
drate (23 mg, 0.12 mmol) was added to a stirred solution of compound 38
(100 mg, 47.2 mmol) in CH3CN (2 mL) at room temperature. After 3 h,
the reaction mixture was neutralized with pyridine and evaporated. The
residue was diluted with CH2Cl2, washed with 1m HCl and KHCO3,
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13173

N.-C. Reichardt et al.
5-Aminopentyl
mannopyranosyl]-(1!3),
mannopyranosyl]
[2-acetamido-2-deoxy-b-d-glucopyranosyl-(1!2)-a-d-
Enzymatic on-chip fucosylation with core a-1,3-fucosyl transferase. (well-
format): A solution (150 mL) containing a-1,3-FucT-core type recombi-
nant from Arabidopsis thaliana (3 mg), GDP-Fuc (500 mm), MnCl2
(30 mm) and alkaline phosphatase (20 mU) in MES buffer (50 mm,
pH 6.5) was added into each well and the slide was placed into a humid
chamber at 308C for 16 h. The slide was washed with the above-men-
tioned protocol previous to examination. To examine successful fucosyla-
tion, one well after incubation with Fuc-T was treated with a 100 nm solu-
tion labelled Cy3-AAL (Aleuria aurantia lectin, core GlcNAc a-1,3-Fuc
recognition) in PBS containing 0.5% Tween-20 for 1 h in an incubation
chamber at 258C. The slide was washed with PBS and water and ana-
lyzed.
G
ACHTUNGTRENNUNG
(1!6)-b-d-mannopyranosyl-(1!4)-2-acetamido-2-
deoxy-b-d-glucopyranosyl-(1!4)-2-acetamido-2-deoxy-b-d-glucopyrano-
side (46): A mixture of compound 42 (48 mg, 14.4 mmol) and ethylenedi-
ACHTUNGTRENNUNGamine (1.2 mL) in nBuOH (4.8 mL) was heated at 1208C under micro-
wave irradiation for 3 cycles of 30 min each. The reaction mixture was
concentrated and co-evaporated with toluene (3ꢂ) and EtOH (3ꢂ). The
crude product was dissolved in anhydrous pyridine (5 mL), cooled to 08C
and Ac2O (2.5 mL) was added dropwise. The reaction was warmed to
room temperature and stirred overnight. The mixture was concentrated
and co-evaporated with toluene (3ꢂ). The crude product was purified by
flash column chromatography (SiO2, EtOAc) to give the title compound
as a transparent oil (35 mg, 78% over two steps). Sodium (12 mg) was
added to liquid ammonia (15 mL) at ꢀ788C and the resulting blue mix-
ture was stirred for 5 min. A solution of the protected product (35 mg,
11.3 mmol) in THF (2 mL) was added dropwise and the resulting mixture
was stirred at ꢀ788C until the blue colour disappeared (10 min). The re-
action was quenched by adding solid NH4Cl until a neutral pH was
reached and was then evaporated to dryness. The crude product was puri-
fied by size-exclusion column chromatography (Biorad P2 gel, aqueous
40 mm NH4HCO3). The fraction was lyophilized to give the title com-
Spotwise elongation with b-1,4-Galactosyltransferase: Nanoliter volumes
(2.0–5.0 nL) from a solution (100 mL) containing b-1,4-GalT (4 mU),
UDP-Gal (800 mm), MnCl2 (5 mm), BSA (0.02%) and alkaline phospha-
tase (160 mU) in HEPES buffer (50 mm, pH 7.4, 30% glycerol) were spot-
ted onto selected ligands of the glycan array by a programmed piezo-
electric printing robot. With this buffer the average diameter of the mi-
crodroplets measured by brightfield microscopy was 300 mm (3.3 nL
volume). The slide was incubated (48 h, 378C, 75% humidity) then soni-
cated in a SDS solution (35 mm in PBS buffer, 408C) to remove reagents
and proteins that could interfere in the fluorescence assay. After washing
with PBST (10 min ꢂ1), PBS (10 min ꢂ2) and water (10 min ꢂ2) the
slide was dried under a stream of argon and was ready for use. The suc-
cessful galactosylation was confirmed by probing one subarray of the
slide with a solution of Cy3-RCA and Cy5-BSL-II (both at 50 nm in PBS,
0.1% Tween-20) for 1 h in an incubation chamber at 258C. The slide was
washed with PBST, PBS and water, dried under a stream of argon and
analyzed in a fluorescence scanner.
1
pound as a white solid (13 mg, 75%). H NMR (500 MHz, D2O): d=5.05
(brs, 1H; H-1), 5.02 (brs, 1H; H-1), 4.84 (d, J=1.1 Hz, 1H; H-1), 4.80
(d, J=1.3 Hz, 1H; H-1), 4.71 (brs, 1H; H-1), 4.52 (dd, J=4.2, 3.5 Hz,
1H; H-1), 4.49 (d, J=8.4 Hz, 1H; H-1), 4.42 (d, J=7.6 Hz, 1H; H-1),
4.18 (s, 1H), 4.13 (s, 1H), 4.08 (s, 1H), 4.00 (s, 1H), 3.95–3.36 (m, 41H),
2.91 (t, J=7.7 Hz, 2H; CH2 linker), 2.00 (s, 3H; NHCOCH3), 1.98 (s,
3H; NHCOCH3), 1.96 (s, 3H; NHCOCH3), 1.66–1.43 (m, 4H; 2ꢂCH3
linker), 1.38–1.29 (m, 2H; CH3 linker); 13C NMR from HSQC experi-
ment (500 MHz, D2O): d=102.21 (C-1, JC1-H1 =171.5 Hz), 101.33 (C-1),
100.98 (C-1), 100.24 (C-1, JC1-H1 =160.0 Hz), 99.66 (C-1, JC1-H1 =171.5 Hz),
99.52 (C-1), 99.47(C-1, JC1-H1 =170.2 Hz), 99.18 (C-1, JC1-H1 =171.9 Hz),
80.32, 79.61, 79.50, 79.24, 79.22, 78.55, 76.28, 75.74, 74.49, 74.35, 73.29,
73.19, 72.47, 71.78, 70.58, 70.06, 69.97, 69.82, 69.80, 69.30, 67.18, 66.61,
65.75, 65.55, 65.41, 65.16, 65.13, 64.97, 61.65, 60.83, 59.96, 54.97, 54.73,
39.20, 27.86, 26.31, 22.17, 21.95 ppm; HRMS (ESI): m/z: calcd for
C59H102N4O41: 1522.610 [M+H]+; found: 1523.622.
Spotwise elongation with a-2,6-sialyltransferase: Nanoliter volumes (2.0–
5.0 nL) from a solution (100 mL) of human a-2,6-sialyltransferase (4 mU),
CMP-Neu5Ac (1 mm), MnCl2 (5 mm), BSA (0.02%) and alkaline phos-
phatase (100 mU) in cacodylate·HCl buffer (50 mm, pH 6.0, 30% glycerol)
were spotted onto selected spots of the glycan array. After incubation
(75% humidity, 24 h, 378C), the slide was sonicated with a SDS solution
(35 mm in PBS buffer) to reduce background fluorescence, washed
(10 min with PBST; 2ꢂ10 min with PBS and 2ꢂ10 min with water) and
dried under a stream of argon. A second cycle was performed to assure
complete reaction. Sialylation was tested by treating one subarray of the
slide in an incubation chamber with a solution of Sambucus nigra lectin
and Ricinus comunis lectin (Cy5-SNA and Cy3-RCA both at 100 nm in
PBS, 0.1% Tween-20) for 1 h at 258C. The slide was washed with PBST,
PBS and water, dried under a stream of argon and analyzed in a fluores-
cence scanner.
Enzymatic on-chip galactosylation with b-1,4-galactosyl transferase.
(well-format): The microarrays were separated into wells containing
single subarrays by a 16-well gasket (Fast Frame incubation chambers
from Whatman). A solution (100 mL) containing bovine b-1,4-galactosyl-
transferase (5 mU), UDP-Gal (800 mm), MnCl2 (5 mm), BSA (0.02%)
and alkaline phosphatase (16 mU) in HEPES buffer (50 mm, pH 7.4) was
added to each well and the slide was incubated at 378C for 48 h and with
70% humidity. The slide was then sonicated in SDS solution (35 mm in
PBS buffer, 408C) to reduce background fluorescence, washed (10 min
with PBST; 2ꢂ10 min with PBS and 2ꢂ10 min with water) and dried
under a stream of argon. Conversion of the acceptors to the galactosylat-
ed species was confirmed by incubation of one subarray with labelled Ri-
cinus communis agglutinin (Cy3-RCA, 50 nm in PBS, 0.5% Tween-20)
and another with a solution of tagged Griffonia (Bandeiraea) simplicifo-
lia lectin (Cy5-BSL-II, 50 nm in PBS, 0.5% Tween-20) for 1 h at 258C.
The slide was washed with PBST, PBS and water, dried and analyzed in a
fluorescence scanner.
Additionally a second sialylated subarray was incubated with biotinylated
ECA (100 nm in HEPES 20 mm. pH 7.7) under the same conditions as
described above, washed as before with PBST, PBS, water and incubated
with tagged Cy3-neutravidine (30 nm PBS, 0.5% Tween) for 1 h, and then
washed again as before, dried and analyzed with a fluorescence scanner.
Spotwise elongation with a-1,3-fucosyltransferase: For core fucosylation
of N-glycans. nanoliter volumes (2.0–5.0 nL) from a solution (100 mL)
containing a a-1,3-fucosyltransferase (3 mg) from Arabidopsis thaliana,
GDP-Fuc (1 mm), MnCl2 (30 mm) and alkaline phosphatase (100 mU) in
MES buffer (50 mm, pH 6.5, 30% glycerol), were spotted onto selected
spots of the glycan array by a printing robot. After incubation (75% hu-
midity, 24 h, 308C), the slide was sonicated with a solution containing
SDS (35 mm in PBS buffer, 408C) to reduce background fluorescence,
washed (10 min with PBST; 2ꢂ10 min with PBS and 2ꢂ10 min with
water) and dried under a stream of argon. To examine fucosylation, one
subarray of the slide was incubated with a solution of labelled Aleuria
aurantia lectin (Cy3-AAL, 100 nm in PBS, 0.1% Tween-20) at 258C. The
slide was washed with PBST, PBS and water, dried under a stream of
argon and analyzed in a fluorescence scanner.
Enzymatic on-chip sialylation with a-2,6-sialyl transferase. (well-format):
For sialylation of terminal LacNAc functions, a solution (100 mL) con-
taining human a-2,6-Sialyltransferase (2 mU), CMP-Neu5Ac (200 mm),
MnCl2 (5 mm), BSA (0.02%) and alkaline phosphatase (20 mU) in
HEPES buffer (50 mm, pH 6.8) was added into individual wells and the
slide incubated at 378C for 48 h and with 75% humidity. The slide was
washed as described in the galactosylation protocol. Sialylation was con-
firmed by incubation of one subarray with labelled Sambucus nigra ag-
glutinin (Cy5-SNA, 50 nm in PBS, 0.5% Tween-20) and another with a
solution of tagged Ricinus communis agglutinin (Cy3-RCA, 50 nm in
PBS, 0.5% Tween-20) for 1 h at 258C. The slide was washed with PBST,
PBS and water, dried and analyzed in a fluorescence scanner.
13174
www.chemeurj.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2010, 16, 13163 – 13175

Construction of N-Glycan Microarrays
FULL PAPER
Chem. Int. Ed. 2004, 43, 2557–2561; c) V. Y. Dudkin, J. S. Miller,
A. S. Dudkina, C. Antczak, D. A. Scheinberg, S. J. Danishefsky, J.
Am. Chem. Soc. 2008, 130, 13598–13607.
Acknowledgements
We are grateful to Prof. I. B. H. Wilson, BOKU, Vienna for providing
the clones and assisting in the expression of a recombinant a-1,3-fucosyl-
transferase from Arabidopsis thaliana. We thank D. Padrꢃ for help with
NMR spectroscopic experiments and C. Mꢄrquez and J. Calvo (MDRe-
nal Lifesciences) for acquisition of mass spectra. Funding from Ministerio
de Ciencia e Innovaciꢃn, CTQ2008–04444/BQA grant, Government of
the Basque Country, Etortek grant, and the European Union, RTN-Eu-
roglycoarrays grant are gratefully acknowledged.
[14] a) Y. Ohnishi, H. Ando, T. Kawai, Y. Nakahara, Y. Ito, Carbohydr.
Res. 2000, 328, 263–276;b) I. Matsuo, T. Kashiwagi, K. Totani, Y.
Ito, Tetrahedron Lett. 2005, 46, 4197–4200; c) I. Matsuo, Y. Ito,
Trends Glycosci. Glycotechnol. 2005, 17, 85–95.
[15] a) R. Frank, Tetrahedron 1992, 48, 9217–9232;b) S. P. A. Fodor, J. L.
Read, M. C. Pirrung, L. Stryer, A. T. Lu, D. Solas, Science 1991, 251,
767–773.
[16] L. Ban, M. Mrksich, Angew. Chem. 2008, 120, 3444–3447; Angew.
Chem. Int. Ed. 2008, 47, 3396–3399.
[17] M. Schuster, P. Wang, J. C. Paulson, C. H. Wong, J. Am. Chem. Soc.
1994, 116, 1135–1136.
[18] O. Blixt, T. Norberg, J. Carbohydr. Chem. 1997, 16, 143–154.
[19] N. Laurent, R. Haddoub, S. L. Flitsch, Trends Biotechnol. 2008, 26,
328–337.
[20] O. Blixt, K. Allin, O. Bohorov, X. F. Liu, H. Andersson-Sand, J.
Hoffmann, N. Razi, Glycoconjugate J. 2008, 25, 59–68.
[21] S. Park, I. Shin, Org. Lett. 2007, 9, 1675–1678.
[1] a) S. Park, M. R. Lee, I. Shin, Chem. Commun. 2008, 4389–4399;
b) T. Horlacher, P. H. Seeberger, Chem. Soc. Rev. 2008, 37, 1414–
1422; c) Y. Liu, A. S. Palma, T. Feizi, Biol. Chem. 2009, 390, 647–
656; d) N. Laurent, J. Voglmeir, S. L. Flitsch, Chem. Commun. 2008,
4400–4412; e) P. H. Liang, C. Y. Wu, W. A. Greenberg, C. H. Wong,
Curr. Opin. Chem. Biol. 2008, 12, 86–92.
[2] R. Apweiler, H. Hermjakob, N. Sharon, Biochim. Biophys. Acta,
Gen. Subj. 1999, 1473, 4–8.
[22] a) N. Laurent, J. Voglmeir, A. Wright, J. Blackburn, N. T. Pham,
S. C. C. Wong, S. J. Gaskell, S. L. Flitsch, ChemBioChem 2008, 9,
883–887; b) M. C. Bryan, F. Fazio, H. K. Lee, C. Y. Huang, A.
Chang, M. D. Best, D. A. Calarese, C. Blixt, J. C. Paulson, D.
Burton, I. A. Wilson, C. H. Wong, J. Am. Chem. Soc. 2004, 126,
8640–8641.
[23] N. Nagahori, S. I. Nishimura, Chem. Eur. J. 2006, 12, 6478–6485.
[24] H. Schachter, in Comprehensive Glycoscience (Ed.: J. P. Kamerling),
Elsevier, Oxford, 2007, pp. 11–32.
[25] a) D. Crich, P. Jayalath, T. K. Hutton, J. Org. Chem. 2006, 71, 3064–
3070; b) D. Crich, S. Sun, J. Am. Chem. Soc. 1998, 120, 435–436.
[26] S. H. Kim, D. Augeri, D. Yang, D. Kahne, J. Am. Chem. Soc. 1994,
116, 1766–1775.
[27] S. Serna, B. Kardak, N. C. Reichardt, M. Martin-Lomas, Tetrahe-
dron: Asymmetry 2009, 20, 851–856.
[28] F. Yamazaki, S. Sato, T. Nukada, Y. Ito, T. Ogawa, Carbohydr. Res.
1990, 201, 31–50.
[29] a) B. G. Davis, A. J. Fairbanks, N.-C. Reichardt, V. U. Angardipar-
ambil (Glycoform Ltd., GB), WO2007/091040A2, 2007; b) B. Yu,
H. C. Tao, Tetrahedron Lett. 2001, 42, 2405–2407.
[30] R. Kornfeld, S. Kornfeld, Annu. Rev. Biochem. 1985, 54, 631–664.
[31] C. Unverzagt, Angew. Chem. 1994, 106, 1170–1173; Angew. Chem.
Int. Ed. Engl. 1994, 33, 1102–1104.
[32] A. Marra, J. M. Mallet, C. Amatore, P. Sinay, Synlett 1990, 572–574.
[33] Cy3 and Cy5 are used as descriptors for the scanner channels oper-
ated in the imaging of microarray slides. See the Supporting Infor-
mation for the cyanine-analogue fluorophores used for tagging the
lectins.
[34] a) S. Park, M. R. Lee, S. J. Pyo, I. Shin, J. Am. Chem. Soc. 2004, 126,
4812–4819; b) J. Su, M. Mrksich, Angew. Chem. 2002, 114, 4909–
4912; Angew. Chem. Int. Ed. 2002, 41, 4715–4718.
[35] N. Shibuya, I. J. Goldstein, W. F. Broekaert, M. Nsimba-Lubaki, B.
Peeters, W. J. Peumans, J. Biol. Chem. 1987, 262, 1596–1601.
[36] K. Yamamoto, T. Tsuji, I. Matsumoto, T. Osawa, Biochemistry 1981,
20, 5894–5899.
[37] S. K. Podder, A. Surolia, Bachhawa. Bk, Eur. J. Biochem. 1974, 44,
151–160.
[38] D. H. Joziasse, W. E. Schiphorst, D. H. van den Eijnden, J. A. van
Kuik, H. van Halbeek, J. F. Vliegenthart, J. Biol. Chem. 1985, 260,
714–719.
[39] a) T. Iskratsch, A. Braun, K. Paschinger, I. B. H. Wilson, Anal. Bio-
chem. 2009, 386, 133–146; b) Y. Itakura, S. Nakamura-Tsuruta, J.
Kominami, N. Sharon, K.-I. Kasai, J. Hirabayashi, J. Biochem. 2007,
142, 459–469.
[40] C. Jin, B. Hantusch, W. Hemmer, J. Stadimann, F. Altmann, J. Aller-
gy Clin. Immunol. 2008, 121, 185–190.
[3] a) O. Blixt, S. Head, T. Mondala, C. Scanlan, M. E. Huflejt, R. Al-
varez, M. C. Bryan, F. Fazio, D. Calarese, J. Stevens, N. Razi, D. J.
Stevens, J. J. Skehel, I. van Die, D. R. Burton, I. A. Wilson, R. Cum-
mings, N. Bovin, C. H. Wong, J. C. Paulson, Proc. Natl. Acad. Sci.
USA 2004, 101, 17033–17038; b) T. Schallus, C. Jaeckh, K. Feher,
A. S. Palma, Y. Liu, J. C. Simpson, M. Mackeen, G. Stier, T. J.
Gibson, T. Feizi, T. Pieler, C. Muhle-Goll, Mol. Biol. Cell 2008, 19,
3404–3414; c) X. Z. Song, B. Y. Xia, S. R. Stowell, Y. Lasanajak,
D. F. Smith, R. D. Cummings, Chem. Biol. 2009, 16, 36–47; d) A. R.
de Boer, C. H. Hokke, A. M. Deelder, M. Wuhrer, Anal. Chem.
2007, 79, 8107–8113; e) X. Song, Y. Lasanajak, C. Rivera-Marrero,
A. Luyai, M. Willard, D. F. Smith, R. D. Cummings, Anal. Biochem.
2009, 395, 151–160.
[4] a) C. C. Wang, Y. L. Huang, C. T. Ren, C. W. Lin, J. T. Hung, J. C.
Yu, A. L. Yu, C. Y. Wu, C. H. Wong, Proc. Natl. Acad. Sci. USA
2008, 105, 11661–11666; b) F. Kamena, M. Tamborrini, X. Liu, Y.-U.
Kwon, F. Thompson, G. Pluschke, P. H. Seeberger, Nat. Chem. Biol.
2008, 4, 238–240; c) S. E. Tully, M. Rawat, L. C. Hsieh-Wilson, J.
Am. Chem. Soc. 2006, 128, 7740–7741.
[5] A. Helenius, M. Aebi, Annu. Rev. Biochem. 2004, 73, 1019–1049.
[6] E. W. Adams, D. M. Ratner, H. R. Bokesch, J. B. McMahon, B. R.
OꢅKeefe, P. H. Seeberger, Chem. Biol. 2004, 11, 875–881.
[7] a) J. Stevens, O. Blixt, T. M. Tumpey, J. K. Taubenberger, J. C. Paul-
son, I. A. Wilson, Science 2006, 312, 404–410; b) J. Stevens, O. Blixt,
L. Glaser, J. K. Taubenberger, P. Palese, J. C. Paulson, I. A. Wilson,
J. Mol. Biol. 2006, 355, 1143–1155.
[8] M. A. Wꢆlti, P. J. Walser, S. Thore, A. Grꢇnler, M. Bednar, M. Kꢇn-
zler, M. Aebi, J. Mol. Biol. 2008, 379, 146–159.
[9] K. Paschinger, E. Staudacher, U. Stemmer, G. Fabini, I. B. H.
Wilson, Glycobiology 2004, 15, 463–474.
[10] T. Ogawa, M. Sugimoto, T. Kitajima, K. K. Sadozai, T. Nukada, Tet-
rahedron Lett. 1986, 27, 5739–5742.
[11] a) H. Paulsen, M. Heume, H. Nurnberger, Carbohydr. Res. 1990,
200, 127–166; b) H. Paulsen, M. Heume, Z. Gyorgydeak, R.
Lebuhn, Carbohydr. Res. 1985, 144, 57–70; c) H. Paulsen, Chem.
Soc. Rev. 1984, 13, 15–45.
[12] a) C. Unverzagt, S. Eller, S. Mezzato, R. Schuberth, Chem. Eur. J.
2008, 14, 1304–1311; b) S. Eller, R. Schuberth, G. Gundel, J. Seifert,
C. Unverzagt, Angew. Chem. 2007, 119, 4251–4253; Angew. Chem.
Int. Ed. 2007, 46, 4173–4175; c) C. Unverzagt, Angew. Chem. 1997,
109, 2078–2081; Angew. Chem. Int. Ed. Engl. 1997, 36, 1989–1992;
d) H. Weiss, C. Unverzagt, Angew. Chem. 2003, 115, 4389–4392;
Angew. Chem. Int. Ed. 2003, 42, 4261–4263; e) C. Unverzagt, Car-
bohydr. Res. 1997, 305, 423–431.
[13] a) B. Wu, Z. H. Hua, J. D. Warren, K. Ranganathan, Q. Wan, G.
Chen, Z. P. Tan, J. H. Chen, A. Endo, S. J. Danishefsky, Tetrahedron
Lett. 2006, 47, 5577–5579; b) M. Mandal, V. Y. Dudkin, X. Geng,
S. J. Danishefsky, Angew. Chem. 2004, 116, 2611–2615; Angew.
Received: May 13, 2010
Published online: September 28, 2010
Chem. Eur. J. 2010, 16, 13163 – 13175
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
13175