Biochimica et Biophysica Acta 1760 (2006) 603 – 609
http://www.elsevier.com/locate/bba
Review
Design of carbohydrate multiarrays
V.I. Dyukova b , N.V. Shilova a , O.E. Galanina a , A.Yu. Rubina b , N.V. Bovin a,⁎
a
Shemyakin and Ovchinnikov Institute of Bioorganic Chemistry, Russian Academy of Sciences, 16/10 ul. Miklukho-Maklaya, 117997, Moscow, Russian
b
Engelhardt Institute of Molecular Biology, Russian Academy of Sciences, 32 ul. Vavilova, 119991 Moscow, Russia
Received 30 September 2005; received in revised form 8 December 2005; accepted 8 December 2005
Available online 5 January 2006
Abstract
Recently, microarray technology has increasingly been widely applied in glycobiology. This technology has rather evident potential
advantages: unlimited number of carbohydrate ligands coated onto one small sized chip, enormously low consumption of both carbohydrate
ligands and carbohydrate-binding proteins to be tested, etc. Literature data demonstrate that three approaches are used for glycoarray design. The
first one is based on the physical adsorption of glycomolecules on a surface (as in a common ELISA), the second one—on covalent
immobilization, and the third one—on a streptavidin–biotin system. In all of the described methods, carbohydrate ligands were placed on chips as
a 2D monolayer and high sensitivity was achieved due to fluorescent detection. Notably, a tendency of stepping from model chips toward real
multiarrays, where the number of carbohydrate ligands can be up to two hundred, has been observed the last 2 years, this already producing a
number of interesting findings when studying carbohydrate-binding proteins.
In 2005 new construction, 3D glycochip was described, where 150 μm diameter polyacrylamide gel elements serve as microreactors instead of
2D dots. As a result of the 3D placement of a ligand, two orders of magnitude increase of its density is possible, this providing principal signal
improvement during fluorescent detection and increasing method sensitivity. At the same time, carbohydrate consumption is low, i.e., ∼1 pmol per
gel element. Copolymerization chemistry enables the immobilization of several glycomolecule classes to the gel, in particular, aminospacered
oligosaccharides, polyacrylamide conjugates, and even 2-aminopyridine derivatives of oligosaccharides, which are widely used in the structural
analysis of glycoprotein N-chains.
© 2005 Elsevier B.V. All rights reserved.
Keywords: Glycochip; Carbohydrate array; Oligosaccharide
1. 2D arrays
Microarray technology allowing for the simultaneous
multiparameter testing of a large number of samples at a
minimal consumption of the used reagents has been increasingly applied widely recently for the study of biological
systems, particularly in glycobiology. Several reviews have
already been dedicated to glycoarrays and glycobiological
problems that could be solved with their use [1–5]. Multiarray
technology has rather evident advantages: unlimited
(thousands) carbohydrate ligands, which can be placed on a
Abbreviations: MAb, momoclonal antibodies; Atri, GalNAcα1-3(Fucα1-2)
Gal; Btri, Galα1-3(Fucα1-2)Gal; PAA, polyacrylamide; CBP, carbohydratebinding protein; 2AP, 2-aminopyridine; Sug, saccharide residue; ELISA,
enzyme-linked immunosorbent assay; All, allyl,CH 2 =CH-CH 2 -; OS,
oligosaccharide
⁎ Corresponding author. Tel.: +7 495 330 71 38; fax: +7 495 330 55 92.
E-mail address: [email protected] (N.V. Bovin).
0304-4165/$ - see front matter © 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.bbagen.2005.12.005
single chip and a principal consumption decrease of carbohydrate ligands and analytes. The latter fact can be illustrated by
the work [6], where 96-well polystyrene plates were compared
to a chip coated with the same polystyrene demonstrating that
polysaccharide consumption at the assigned antibody concentration was 50 pg and 80 fg, respectively. Another advantage of
multiarrays with a maximally wide glycan set should be noted:
screening expands our tunnel vision and eliminates prejudice
during the selection of potential ligands for particular
carbohydrate-binding protein, thus providing a chance of
serendipity in the search of new ligands. Such a set also
removes the limitations of choice due to economic reasons: few
researchers can afford purchasing one hundred or even several
dozens commercial glycoconjugates for carrying out a single
experiment, whereas several hundred glycans will be presented
on a single chip for a reasonable price.
As the key technology components, arrayers, scanners, chip
platforms (microscope glass slides are usually used) and the
604
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
corresponding surface chemistry were adopted from DNA
arrays at the first stage of the carbohydrate microarray
development, it became necessary to adapt them to glycan
particularities. This related mainly to immobilization methods.
The simplest way of immobilization inherited from the
standard ELISA is based on physical adsorption of carbohydrates, in particular on polystyrene [6], nitrocellulose [7],
polyvinylidene fluoride membrane [8], etc. The first of the cited
platforms, glass slides coated with black polystyrene for
fluorescence detection, is commercially available.
The second method is the chemical (covalent) immobilization of glycoligands on an activated surface or, oppositely,
activated glycoligands on a functionalized surface. A microarray platform based on dextran-coated glass slides (PhotoChips obtained from CSEM, Switzerland) was reported in [9].
The polymer was derivatized with aryl-trifluoromethyl-diazirine groups that upon illumination formed reactive carbenes,
which reacted with any vicinal molecule to form covalent
bonds. In the paper [10] an oligomer of 1,8-diamino-3,6dioxaoctan was coupled to solid support (GlycoChip, Glycominds, Lod, Israel) to obtain an activated surface and after that
p-aminophenyl glycosides were covalently bound to a plastic
surface via a cyanurchloride-activated linker. SH-coated glass
slides (Diometrix Technology, Korea), onto which maleimideconnected carbohydrates were covalently immobilized, were
used as an activated surface in [11]. The same authors
developed a method for chemical microarray fabrication by
way of immobilization of hydrazide-linked substances on
epoxide-coated glass surfaces [12]. Commercially available
aminoreactive NHS-activated slides were used in a printed
glycan array designed by Blixt et al. [13]. It allowed the
covalent attachment of glycans containing a terminal aliphatic
amine by forming an amide bond under aqueous conditions at
room temperature. Somewhat apart from the solid-phase
platforms cited above stands the suspension microsphere
array [14], where the particles were bar coded and detection
was carried out by using flow cytometry; coupling of
glycoligand to carboxylated microspheres was performed with
the use of carbodiimide chemistry.
A number of arrays are based on streptavidin–biotin
immobilization. They were realized in various formats, in
particular, on a 384-well plate with well volume 25 μl [15,16]
and 192-spot slide format [17]. The first of them was designed
to be in maximal proximity to the traditional immunochemical
assay using commercial streptavidin coated black 384-well
plates. A carbohydrate ligand was biotinylated via a short
aliphatic spacer or at the peptide part in the case of
glycopeptides.
Efficiency of the substance coating on the chip largely varies
in the cited literature. Physical immobilization on smooth
materials, such as polystyrene, usually proceeds with several
percent yields, increasing up to 20% only for the molecules
having m.w. of approximately 106 Da [18]. The applied amount
can be decreased up to 10 −13 g as mentioned in [6].
Approximately 10−11 mol of substance was applied to a spot
in the study where physical adsorption on nitrocellulose (i.e., a
porous material) was used [7], but the percentage of actual
immobilized substance was not reported. In [14] approximately
108 coupling sites were present on each microsphere; efficiency
of glycopeptides coupling to microspheres was 72–89% [14].
Streptavidin array on 384-well plates needed approximately
10−11 mol/well [16]; in the case of correct ratio streptavidin/
biotin, the efficiency of attachment is specified only by the
streptavidin amount in a well, thus it is independent from the
carbohydrate ligand. Therefore, it is possible to suppose that
this approach leads to quantitative saccharide immobilization.
The degree of immobilization for the microarray reported in
[13] was approximately 50%, (O. Blixt, personal communication). Glycans were applied in two concentrations, i.e., 6 × 10−14
and 6 × 10−15 mol/spot. As for all the other papers cited above, it
was impossible to obtain information regarding the yields at the
immobilization stage from them.
When comparing coating methods, it is necessary to take into
account not only the chemistry and efficiency of carbohydrate
coatings but also the immobilization consistency between
different classes of glycomolecules. Indeed, if the efficiency
(yield) of the attachment is different for glycopeptides,
polysaccharides, and oligosaccharides (or even oligosaccharides of different size), this will lead to errors during the
interpretation of the further assay results. In this respect, a
streptavidin–biotin system seems to be the most reliable,
whereas physical adsorption—the most risky one, in particular,
neoglycolipids, glycosylamines, and polysaccharides differ
greatly by their ability for adsorption onto nitrocellulose [19].
Chemical immobilization lies in the line; a good solution of the
consistency problem is the amino ligand attachment to the
activated polymer [13,18,20], i.e., the case when the ligand can
be taken in deficiency with respect to the activated surface. In
particular, the printed array [13,21] demonstrated good surface
reproducibility independently of the differences in the structures
of glycomolecules.
Immobilization chemistry is important from another point of
view. Carbohydrate–protein interactions are considerably less
affine compared to other biologically important interactions
studied using chip technology, such as protein–protein one or
oligonucleotide hybridization. This means that the requirements
for a background signal for a carbohydrate chip are,
correspondingly, more strict [22], which in turn very seriously
limits the method (chemistry) of immobilization, first of all, the
spacer (linker) nature. Obviously, such hydrophobic linkers as
long aliphatic, any aromatic ones and also charged or just bulky
ones will sooner or later lead to a non-specific interaction, even
in the case of the use of blocking by the addition of BSA,
Tween, etc. With respect to this, the printed array [13] is
favorable to many others because glycan linkage with a carrier
is formed with the help of short C2 or C3 spacers, whereas a
hydrophilic polymer serves as the carrier. Due to chain
flexibility, the polymer executes two more useful functions
for the assay: firstly, it assures glycan delivery to even deep
pockets of a carbohydrate-binding protein (CBP) and secondly,
provides multivalent interactions independent of the distance
between subunits of a CBP.
Additionally, not only in micro—but also in multiarrays, of
primary importance is representativity of the carbohydrate set,
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
in other words, microchip technology is worth applying only
when dozens or, better, hundreds of saccharides are coated
onto a chip. Most platforms based on glass slides and printed
with the help of robotic arrayers allow coating of up to
10,000 spots. The real number of carbohydrates used in these
arrays was always considerably smaller, reaching two hundred
at best. A set used in [6] included polysaccharides (pectic
polysaccharide, galactan, arabinan, xylogalacturonan, arabinogalactan, proteoglycans) and neoglycoproteins, fourteen
compounds in total. Each substance was coated in fifteen
repetitions for the reliability of quantitative measurement.
Forty eight microbial polysaccharides (Klebsiella, Pneumococcus, Meningococcus, etc.) were used in [8]. Glycoarray
consisting of glycolipids and neoglycolipids included 95
positions, in particular the derivatives of neutral, sialylated
and sulfated oligosaccharides, and oligosaccharides GAG
fragments [7]. Blixt et al., Consortium for Functional
Glycomics [13,21], placed about 200 synthetic glycans on a
chip (in six replicates), which represented the most typical
terminations and core fragments of mammal glycoproteins
and glycolipids. A similar set of glycans (approximately 180
in total) though as biotinylated oligosaccharides was used for
the design of another glycochip of the Consortium [16,21].
Importantly, the results of parallel testing of the same CBP
with the help of the two latter arrays mostly coincided [21].
Concluding this review, it is necessary to mention the recent
study of Xia et al. [23] where 2,6-diaminopyridine (DAP)
derivatives of OS were used for microarray design. The
second amino group in DAP label possess sufficient reaction
ability for both direct printing on NHS-activated slides [13]
and biotinylation followed by the placement of biotinylated
OS onto streptavidin platform [21]. Critical analysis demonstrates that excepting several above-mentioned studies, the
rest that were cited in this paper are rather attempts to
develop model glycochips than to design actual multiarrays.
In this paper, we report the platform principally different
from all of the above described, i.e., a chip where saccharides
605
are immobilized inside 3D drops of porous polymeric gel
(Fig. 1). This platform has already demonstrated its
advantages over 2D analogs during the design of oligonucleotide chips [24,25]. Immobilization of oligosaccharides as ωaminospacered glycosides proceeds due to the formation of a
covalent bond between the amino group and growing polymer
chain during photoinitiated polymerization in the presence of
a cross-linking agent. We hypothesized that due to the larger
amount of the ligand displaced in the volume, the 3D chip
would demonstrate higher sensitivity than the similar 2D
variants.
2. 3D array
The advantages of microarray technology can be realized
only when the ligand set is rather representative, i.e., includes
several hundred substances. At the same time, literature data
analysis (see above) demonstrates that most of the described
arrays are far from this desirable criterion, whereas the most
representative of them contains the substances of one type, e.g.,
synthetic spacered oligosaccharides, neoglycolipids, or polysaccharides. Obviously, a more comprehensive methodology,
which would enable the utilization of different classes of
carbohydrate molecules and different derivatives of them with
equal or similar efficiency, should be sought in order to further
increase the number of ligands. One of the still unused sources
of complex, and thus particularly interesting glycans, is the
analytical chemistry of carbohydrates. In particular, the number
of described and well-characterized 2AP derivatives of Nglycans [26,27] reaches several hundred; many of these
derivatives are commercially available. To expand the range
of carbohydrate ligand forms that could be covalently bound on
a chip, an approach developed earlier for the grafting of
oligonucleotides and proteins [24,25] into 3D polyacrylamide
gel, was applied for carbohydrate immobilization [28]. Three
very different classes of glycomolecules, i.e., (1) OS with
primary amino group, OS-OCH2CH2CH2NH2, (2) allyl-
Fig. 1. Photograph of microchip gel elements in transmitted light. The diameter of the gel drops was 150 μm.
606
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
substituted polyacrylamide conjugates, OS-PAA-All (30 kDa)
[29], and (3) 2AP derivatives of oligosaccharide alditols were
taken for grafting into the chip element. All of the three forms,
OS-OCH2CH2CH2NH2, OS-PAA-All, and OS-2AP are readily
subjected to covalent attachment in the same conditions, in
radical process of gel-forming. Below, we present this new, 3D
technology, more narrowly than 2D approaches.
For hydrogel microchips manufacturing, the gel-forming
monomers, i.e., methacrylamide, methylenebisacrylamide, and
compounds to be immobilized are printed onto hydrophobized
glass and are UV irradiated. Double bond of methylenebisacrylamide readily reacts with the amino group of Sug-sp-NH2
at pH 10.5 giving rise to a Michael addition product.
Methacrylamide is taken as the gel-forming monomer instead
of acrylamide, since it is inactive in the reaction of the
Michael addition, whereas acrylamide enters readily into this
reaction thus blocking the polymerization process. Sug-PAAAll conjugate (5% mol. of allyl groups) is inserted into the
polymeric chain due to a double bond of the allyl groups. In
both versions, the trisaccharides were immobilized in
concentrations from 1.7 to 0.15 μmol/ml. After polymerization, an array of individual three-dimensional ∼1 nl gel drops,
150 μm in diameter and 25 μm in height is formed. Before
the assays, all microchips were checked for quality, i.e.,
microchips with a deviation in positions and radii of gel
elements not more than 5% from the average values were
utilized for the experiments. We compared the gel chip based
on oligosaccharide 3-aminopropyl glycosides with the chip
based on Glyc-PAA-All chemistry (Fig. 2). Chip-immobilized
trisaccharides were tested by the interaction with anti-Atri
MAb. In both cases anti-Atri antibodies interacted only with
trisaccharide A and did not bind trisaccharide B, i.e., there
was not any non-specific binding of anti-carbohydrate MAb.
The signal recorded from the gel elements with aminospacered Àtri was 1.5–2 times higher than that from immobilized
Atri-PAA-All, though equal amounts of Atri were taken per
spot in both cases. As the gel was formed in the same manner
in both cases, i.e., the availability and presentation of the
carbohydrate ligand was expected to be similar, the difference
in the MAb binding can be explained by a lower degree of
immobilization in the second case. Taking into account 50%
yield in case of primary amines immobilization [24], the yield
of allyl derivative grafting seems to be approximately 30%.
This result is explained by the low reaction ability of the allyl
group in the reaction of radical copolymerization. It should be
noted that the degree of OS-PAA-All grafting could be
improved by the optimization of gel formation conditions.
Both variants of carbohydrate ligand presentation on a gel
chip have their own advantages. Immobilization of ωaminospacered ligands on the chip allows the use of already
existent OS libraries [21], whereas OS-PAA-All conjugates
with different content of the carbohydrate ligand enable the
variation of the density of the saccharide in the composition
of the PAA chain at the stage of the conjugate synthesis. The
concentration dependence obtained on 3D microchips and 96well plates (data not shown) as well as the minimal
concentration of antibodies (maximal dilution) that can be
detected using chip and 96-well plate techniques are similar;
the minimal concentration was determined as a concentration
of antibodies at which the signal is twice as high as the
background signal, i.e., the signals when Btri was immobilized. However, as was expected, the difference in the amount
of carbohydrate ligand taken for immobilization is striking. In
the case of the microchips, the amount of saccharide (Fig. 2)
was ∼1 pmol per gel element, whereas for 96-well ELISA,
∼1 nmol of saccharide per well was taken for physisorption
[18].
Reaction ability of the amino group (secondary and
aromatic) in 2AP is very low. Thus, it was impossible to
immobilize 2AP-modified OS by N-acylation reaction (data
not shown). Nevertheless, we hoped that the reaction ability
in the Michael reaction at pH 10.5 would be sufficient. The
possibility of the immobilization of 2AP derivatives was
studied in the example of a melibiose derivative Galα1–
6Glc2AP, whereas presumably a reactive aminoalditol analog,
Galα1–6Glcol, was taken for comparison. As in the case of
trisaccharides A and B the fact of the inclusion and its
efficiency were evaluated indirectly, due to the interaction
with specific antibodies. As was demonstrated with the help
of a standard ELISA, corresponding monospecific human
antibodies [30] were capable of recognizing Galα1–6Glcol in
the form of PAA-conjugate (data not shown). Interaction of
Fig. 2. Fluorescence of gel elements with immobilized Atri-OCH2CH2CH2NH2
( ) and Atri-PAA-All (▴), both 1.7 pmol of carbohydrate per gel element.
Dependence of fluorescence on anti-Atri monoclonal antibodies dilution.
Microchips were incubated with MAb and developed with Cy5-labeled antimouse antibodies. The dotted line indicates the background signal from gel
elements without immobilized trisaccharide.
▪
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
607
Table 1
Comparison of two glycoarray variants with classic ELISA on 96-well plate.
Fig. 3. Comparison of antibody-binding potency of melibiose in two forms,
Galα1–6Glcol and Galα1–6Glc2AP. The 3D-microchip was incubated with
human antibodies affinity purified on Galα1–3Gal-Sepharose and developed
with Cy5-labeled anti-human Ig antibodies.
these antibodies with both disaccharide versions was observed
in the case of glycochip. Antibody interaction with gel
elements containing Galα1–6Glc2AP took place in the
described experiment (Fig. 3) though it was 2–2.5 times
weaker than that of the aminoalditol variant. According to the
interaction with antibodies, the degree of 2AP derivative
attachment proved to be lower but still sufficient for the
reliable reading of a high signal. Optimization of OS-2AP
grafting into the gel has not been performed; this is a subject
for further studies aimed to increase the yield.
3. Bridging glycochip and structure analysis
The amount of carbohydrate ligand in a single gel element
is approximately 1 pmol. Assuming that several gel elements
with the ligand must be placed on each chip and that serial
work needs several chips, the actually necessary substance
amounts to approximately 10 pmol. Routine mapping of
glycoprotein's carbohydrate chains based on chromatographic
separation of OS is performed within pmol–nmol range. This
opens an attractive prospect of using OS obtained after
analytical HPLC for immobilization on a chip followed by
assaying lectins and other carbohydrate-binding proteins.
Such a “link” variant, when OS bears a fluorescent label, is
Parameter of test-system
3D chip a
96-well
plate b
Printed
array c
OS immobilized in the
form of:
Working area, diameter
OS amount used for
immobilization
Immobilization yield
Working dilution of MAb
Volume of MAb per one
chip, μl
OS-sp-NH2
OS-PAA
OS-sp-NH2
150 μm
∼10−12 mol/gel
element
∼50%
1:103
102
5000 μm
∼10−10 mol/
well
2–20% d
1:103
104
200 μm
∼10−14 mol/
spot
∼ 50%
1:102
102
Anti blood group A monoclonal antibody A16 was assayed in all cases, OS is Atri
a
3D gel-chip.
b
ELISA on 96-well plate from [17,18].
c
Printed array from [13].
d
2% yield corresponds to 30 kDa OS-PAA, 20% – to 1000 kDa OS [18].
especially attractive: on the one hand, this makes it easier the
HPLC separation and more sensitive OS detection, on the
other hand, this makes real quantitative dosage of glycans
during chip fabrication. In this case, the experimental
technique could be the following: (i) carbohydrate chains
are cleaved from the protein core, (ii) oligosaccharide pool is
labeled with fluorescent reagent, (iii) HPLC is performed,
possibly in two-dimensional mode, i.e., using consequently
two columns, (iv) after detection, the collected peaks are
concentrated to the optimal concentration, and (v) are placed
into the microarrayer. Such bridging of structure analysis and
glycoarray technique seems to be very rational and
prospective.
4. Ligand density
Array miniaturization accompanied with the decrease of
the substance that is coated to the chip has its negative points.
Indeed, minimization of the substance consumption and,
correspondingly the decrease of carbohydrate quantity on the
chip inevitably leads to the decrease of signal value. This
makes it necessary to increase compensatory the concentrations of CBP used in the array, which, in turn, is unwanted.
For example, the glycochip described by us earlier [17]
Fig. 4. Cartoon illustrating the effect of overdense ligand placement on the chip surface. Left: the ligand interacts readily with CBP at an optimal ligand density, right:
interaction is hindered or impossible in the case of excessive ligand density on the chip surface.
608
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
consumed two orders less of the carbohydrate than common
ELISA but needed two orders higher concentration of
antibodies in the antibody assay for obtaining of the same
result. To increase the assay signal, it is advantageous to
increase, in a reasonable manner, the amount (density) of the
ligand but it is impossible to endlessly increase the ligand
density on the surface. Moreover, excessive density can lead
to a decrease rather than increase of the interaction with CBP
as illustrated by Fig. 4. In this respect, the 3D variant is much
more advantageous allowing more ‘comfortable’ ligand
colonization on the same area (in the given case 150 μm)
thus providing stronger and correspondingly more reliable
signal. An objection could be the assay slowdown due to
impeded CBP diffusion in gel. However, this problem has
already been solved [28,31], the time of the protein contact
with gel is approximately 1 h. Thus, 3D chip-based assay
occupies a niche between common ELISA and 2D printed
array (Table 1) in respect of OS consumption (10−12, 10−10,
10−14 mol per dot correspondingly); requiring 100 times less
OS this assay works with the same concentration of
antibodies as ELISA, consuming, in the same time, 100
times less of these antibodies. At the same time, several
intrinsic limitations of 3D chip should be noted, in particular,
hampered penetration of viruses and bacteria into the gel and
destruction of alkali-labile ligands due to high pH value
during immobilization. The experimental results reported here
demonstrate certain prospects for further work in this
direction.
Acknowledgements
This study was supported by NIH grants NIGMS 1U54
GM62116 and GM60938, Physico-Chemical Biology Program,
RAS. The authors would like to thank I.M. Belayanchikov for
assistance with manuscript preparation. The authors are grateful
to E. Ya. Kreindlin, K.B. Yevseev, D.A. Urasov for their help in
microchip manufacturing, V.E. Barsky for the construction of
fluorescence microscopes, V.V. Petukhov and A.A. Stomakhin
for the construction of a mixing device, A.Yu. Turygin, D.V.
Prokopenko, and R.A. Urasov for the development of software
for the processing of experimental data, and all of the members
of the protein biochips group for the help and useful
discussions.
References
[1] J. Hirabayashi, Oligosaccharide microarrays for glyconomics, Trends
Biotechnol. 21 (4) (2003) 141–143.
[2] C.O. Mellet, J.M.G. Fernandes, Carbohydrate microarrays, ChemBioChem. 3 (2002) 819–822.
[3] T. Feizi, W. F.Fazio, C.-H. Chai, Carbohydrate microarrays—A new set of
technologies at the frontiers of glyconomics, Curr. Opin. Struct. Biol. 13
(2003) 637–645.
[4] D.M. Ratner, E.W. Adams, M.D. Disney, P.H. Seeberg, Tools for
glyconomics: mapping interactions of carbohydrates in biological systems,
ChemBioChem. 5 (2004) 1375–1383.
[5] I. Shin, S. Park, M.-R. Lee, Carbohydrate microarrays: an advanced
technology for functional studies of glycans, Chem. Eur. J. 11 (2005)
2894–2901.
[6] W.G. Willats, S.E. Rasmussen, T. Kristensen, J.D. Mikkelsen, J.P. Knox,
Sugar-coated microarrays: a novel slide surface for the high-throughput
analysis of glycans, Proteomics 2 (2002) 1666–1671.
[7] D. Wang, S. Liu, B.J. Trummer, C. Deng, A. Wang, Carbohydrate
microarrays for the recognition of cross-reactive molecular markers of
microbes and host cells, Nat. Biotechnol. 20 (2002) 275–281.
[8] S. Fukui, T. Feizi, C. Galustian, A.M. Lawson, W. Chai, Oligosaccharide microarrays for high-throughput detection and specificity
assignments of carbohydrate–protein interactions, Nat. Biotechnol. 20
(2002) 1011–1017.
[9] S. Angeloni, J.L. Ridet, N. Kusy, H. Gao, F. Crevoisier, S. Guinchard, S.
Kochhar, H. Sigrist, N. Sprenger, Glycoprofiling with micro-arrays of
glycoconjugates and lectins, Glycobiology 15 (2005) 31–41.
[10] M. Schwarz, L. Spector, A. Gargir, A. Shtevi, M. Gortler, R.T. Alstock, A.
A. Ducler, N. Dotan, A new kind of carbohydrate array, its use for profiling
antiglycan antibodies, and the discovery of a novel human cellulosebinding antibody, Glycobiology 13 (2003) 749–754.
[11] S. Park, I. Shin, Fabrication of carbohydrate chips for studying
protein–carbohydrate interactions, Angew. Chem., Int. Ed. 41 (2002)
3180–3182.
[12] M. Lee, I. Shin, Fabrication of chemical microarrays by efficient
immobilization of hydrazide-linked substances on epoxide-coated glass
surfaces, Angew. Chem., Int. Ed. 44 (2005) 2881–2884.
[13] O. Blixt, S. Head, T. Mondala, C. Scanlan, M.E. Huflejt, R. Alvares, M.C.
Bryant, F. Fazio, D. Calarese, J. Stevens, N. Razi, D.J. Stevens, J.J. Skehel,
I. van Die, D.R Burton, I.A. Wilson, R. Cummings, N. Bovin, C.-H. Wong,
J.C. Paulson, Printed covalent glycan array for ligand profiling of diverse
glycan binding proteins, Proc. Natl. Acad. Sci. U.S.A. 101 (2004)
17033–17038.
[14] K. Yamamoto, S. Ito, F. Yasukakawa, Y. Konami, N. Matsumoto,
Measurement of the carbohydrate-binding specificity of lectins by a
multiplexed bead-based flow cytometric assay, Anal. Biochem. 336 (2005)
28–38.
[15] Y. Guo, H. Feinberg, E. Conroy, D.A. Mitchell, R. Alvarez, O. Blixt, M.E.
Taylor, W.I. Weis, K. Drickamer, Structural basis for distinct ligandbinding and targeting properties of the receptors DC-SIGN and DCSIGNR, Nat. Struct. Mol. Biol. 11 (2004) 591–598.
[16] B.S. Bochner, R.A. Alvares, P. Nehta, N.V. Bovin, O. Blixt, J.R. White,
R.L. Schnaar, Glycan array screening reveals a candidate ligand for
siglec-8, J. Biol. Chem. 280 (2005) 4307–4312.
[17] O.E. Galanina, M. Mecklenburg, N.E. Nifantiev, G.V. Pazynina, N.V.
Bovin, GlycoChip: multiarray for the study of carbohydrate-binding
proteins, Lab. Chip 3 (2003) 267–272.
[18] N.V. Shilova, O.E. Galanina, T.V. Pochechueva, A.A. Chinarev, V.A.
Kadykov, A.B. Tuzikov, N.V. Bovin, High molecular weight neoglycoconjugates for solid phase assays, Glycoconj. J. 22 (2005) 43–51.
[19] T. Feizi, Glyco XVIII, Florence, 2005.
[20] N.V. Bovin, E.Yu. Korchagina, T.V. Zemlyanukhina, N.E. Byramova, O.E.
Galanina, A.E. Zemlyakov, A.E. Ivanov, V.P. Zubov, L.V. Mochalova,
Synthesis of polymeric neoglycoconjugates based on N-substituted
polyacrylamides, Glycoconj. J. 10 (1993) 142–151.
[21] Public site of Consortium for Functional Glycomics: www.
functionalglycomics.org.
[22] N.V. Bovin, Neoglycoconjugates as probes in glycobiology, in: M.P.
Schneider (Ed.), Chemical Probes in Biology, Kluwer, the Netherlands,
2003, pp. 207–225.
[23] B. Xia, Z.S. Kawar, T. Ju, A. Alvarez, G.P. Sachdev, R.D. Cummings,
Versatile fluorescent derivatization of glycans for glycomic analysis, Nat.
Methods 2 (2005) 845–850.
[24] A.Yu. Rubina, S.V. Pankov, E.I. Dementieva, D.N. Penkov, A.V. Butugin,
V.A. Vasiliskov, A.V. Chudinov, A.L. Mikheikin, V.M. Mikhailovich, A.D.
Mirzabekov, Hydrogel drop microchips with immobilized DNA: properties and methods for large-scale production, Anal. Biochem. 325 (2004)
92–106.
[25] A. Yu. Rubina, E.I. Dementieva, A.A. Stomakhin, S.V. Pankov, V.E.
Barsky, S.M. Ivanov, E.V. Konovalova, A.D. Mirzabekov, Hydrogel-based
protein microchips: manufacturing, properties, and applications, BioTechniques 34 (2003) 1008–1022.
V.I. Dyukova et al. / Biochimica et Biophysica Acta 1760 (2006) 603–609
[26] N. Takahashi, Y. Wada, J. Awaya, M. Kurono, N. Tomiya, Twodimensional elution map of GalNAc-containing N-linked oligosaccharides, Anal. Biochem. 208 (1993) 96–109.
[27] N. Takahashi, Three-dimensional mapping of N-linked oligosaccharides
using anion-exchange, hydrophobic and hydrophilic interaction modes of
high-performance liquid chromatography, J. Chromatogr., A 720 (1996)
217–225.
[28] V.I. Dyukova, E.I. Dementieva, D.A. Zubtsov, O.E. Galanina, N.V. Bovin,
A.Yu. Rubina, Hydrogel glycan microarrays, Anal. Biochem. 347 (2005)
94–105.
609
[29] P.D. Rye, N.V. Bovin, Selection of carbohydrate-binding cell phenotypes
using oligosaccharide-coated magnetic particles, Glycobiology 7 (1997)
179–182.
[30] L.S. Khraltsova, M.A. Sablina, T.D. Melikhova, D.H. Joziasse, H. Kaltner,
H.-J. Gabius, N.V. Bovin, An enzyme-linked lectin assay for α1,3galactosyltransferase, Anal. Biochem. 280 (2000) 250–257.
[31] D.A. Zubtsov, S.M. Ivanov, A.Yu. Rubina, E. I. Dementieva, V.R.
Chechetkin, A.S. Zasedatelev, Effect of mixing on reaction–Diffusion
kinetics for protein hydrogel-based microchips, J. Biotechnol. (in press)
(doi:10.1016/j.jbiotec.2005.08.010).