High-throughput carbohydrate microarray analysis of 24 lectins

Article abstract of DOI:10.1002/anie.200600591

Hip hip array: Lectins are involved in a wide range of biological processes and are used frequently as tools for basic and applied research. Through the use of a carbohydrate microarray, high-throughput analysis of lectin specificity was carried out on 24

Full text of DOI:10.1002/anie.200600591

Angewandte
Chemie
anticancer agent.[3] Therefore, a fundamental understanding
of carbohydrate–protein interactions and comprehensive
information on lectin specificity is critical. Unfortunately,
evaluation of lectin specificity is not trivial. One common
method involves measuring the binding of lectins to cells,
tissues, and glycoproteins. This approach frequently uncovers
interesting and useful binding properties, however, cells,
tissues, and glycoproteins display complexmitxures of
carbohydrate epitopes. Therefore, it is exceedingly difficult
to determine the specific carbohydrate structures being
recognized by the lectin. An alternative approach involves
measuring binding to structurally defined carbohydrate
epitopes through techniques such as isothermal calorimetry
(ITC), mono- and oligosaccharide inhibition studies, enzyme-
linked lectin assays (ELLA), and surface plasmon resonance
assays (SPR). Unfortunately, these methods can be labor
intensive, require large amounts of carbohydrates, and/or be
difficult to perform in a high-throughput fashion. Moreover,
these studies have typically been limited to the small number
of carbohydrate epitopes that were readily accessible.
Although lectin specificity has been studied often, much
more comprehensive information is still needed.
Carbohydrate microarrays are an emerging technology
for the high-throughput evaluation of carbohydrate-macro-
molecule interactions.[4–27] Analogous to DNA and protein
arrays, carbohydrate microarrays contain numerous carbohy-
drate epitopes immobilized on a solid support in a mini-
aturized fashion. The microarray format allows one to rapidly
evaluate many potential interactions with a minimal amount
of sample. Our group has recently developed a carbohydrate
microarray and a highly sensitive assay to detect binding.[21]
To maximize throughput, each slide contained 16 wells with
an entire array printed in each well (see Figure 1a). To
illustrate the capabilities of our microarray and study lectin
sepcificity, 24 lectins were evaluated at eight different
concentrations by using the microarray. As one of the largest
and most-comprehensive lectin studies ever reported, the
results should be a useful resource for scientists conducting
basic and applied research with lectins. Interestingly, micro-
array analysis revealed unexpected ligands for many of the
lectins.
Carbohydrate Microarrays
DOI: 10.1002/anie.200600591
High-Throughput Carbohydrate Microarray
Analysis of 24 Lectins**
Joseph C. Manimala, Timothy A. Roach, Zhitao Li, and
Jeffrey C. Gildersleeve*
Lectins, non-immunoglobulin proteins that bind carbohy-
drates, play a central role in a wide range of biological
processes such as cell–cell recognition, viral and bacterial
pathogenesis, and inflammation.[1,2] Moreover, they are used
extensively as research tools, diagnostics, and therapeutics.
For example, mistletoe lectin is in clinical trials as an
Our approach for fabricating arrays involves printing
carbohydrate–bovine serum albumin (BSA)/human serum
albumin (HSA) conjugates and glycoproteins on epoxide-
functionalized glass microscope slides. This strategy permits
immobilization of both structurally discrete synthetic carbo-
hydrates as well as natural carbohydrates presented on
glycoproteins. Our first array contained 29 components. To
increase the diversity, 23 additional glycans were chemically
synthesized and 21 were purchased (see the Supporting
Information). The array then contained 73 different compo-
nents: 4 controls, 54 BSA/HSA conjugates, and 15 glycopro-
teins (see Figure 1c).
[*] Dr. J. C. Manimala, Dr. T. A. Roach, Dr. Z. Li, Dr. J. C. Gildersleeve
Laboratory of Medicinal Chemistry
Center for Cancer Research, NCI-Frederick
376 Boyles St., 376/109, Frederick, MD 21702 (USA)
Fax: (+1)301-846-6033
E-mail: gildersleevej@ncifcrf.gov
Lectins were evaluated by using our previously reported
assay. Briefly, slides were incubated with biotinylated lectins
in serial dilutions, washed, and then incubated with strepta-
vidin–horseradish peroxidase (HRP). Finally, wells were
incubated with a Cy3-labeled tyramide substrate. The fluo-
rescence intensities of each spot were then measured by using
[**] We thank Jack Simpson (Protein Chemistry Laboratory, SAIC/NCI-
Frederick) for MALDI-MS analysis of BSA conjugates. This research
was supported by the Intramural Research Program of the NIH,
NCI.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 3607 –3610
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3607

Communications
Figure 1. The carbohydrate microarray: a) the 16-well slide format; b) typical assay results for a single well; c) components of the array in the
order that they were printed (black=controls, blue=BSA/HSA conjugates, green=glycoproteins).
a standard DNA microarray scanner (for an image of a typical
well, see Figure 1b). At any single concentration of lectin,
some spots may be saturated while others are barely
distinguishable from the background signal. By evaluating
binding over a broad range of lectin concentrations, one can
determine the relative affinities for many ligands. Relative
binding has been expressed in terms of the detection limit
(DL), which is the lowest concentration of lectin that
produced a signal five-times higher than the background for
a given epitope. Better ligands have lower detection limits.
Results for each of the 24 lectins are summarized in
Table 1.[28] Several findings are of special interest. First, many
lectins that bound epitopes would not have been predicted
based on their designated specificity.[29] For example, BPL is
known as a GalNAc/Galb1–3GalNAc-binding lectin and
RCA120 is considered a Gal/GalNAc-binding lectin. Inter-
estingly, both lectins bound tightly to rhamnose (DL = 80 pm
for BPL and 21 pm for RCA120). Furthermore, soluble
rhamnose competitively inhibited the binding of these lectins
to lactose (see the Supporting Information). From a chemical
perspective, recognition of both galactose and rhamnose is
remarkable given the significant structural differences (i.e.
stereochemistry of the hydroxyl groups at the 2- and 4-
positions, oxygenation compared with deoxygenation at the 6-
position, and overall d compared with l stereochemistry; see
Scheme 1). Recognition of rhamnose may also be important
biologically as rhamnose is found in a variety of plant and
bacterial carbohydrates. ACL, a Galb1-3GalNAc binding
lectin, was found to bind Manb1-4Manb1-4Manb1-4Man
(DL = 29 nm),
a plant oligosaccharide. Again, Galb1-
3GalNAc and Manb1-4Man contained substantial structural
differences (see Scheme 1). As a third example, a lectin that
binds sialic acid, SNA, demonstrated good binding to some
glycans that lack sialic acid such as blood group B (DL =
3608 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3607 –3610

Angewandte
Chemie
Table 1: Lectin specificity profiles.
Lectin (specificity)[a]
Epitopes bound by the lectin (detection limits)[b]
ACL (Galb1-3GalNAc) a-Gn, Di-Lex (1 nm); Lex, SLex, Gn (4 nm); aOSM, PSA, HSP90, AFP (14 nm); Manb4, aBSM (29 nm)
BPL (TF, GalNAc)
a-Gn, aOSM, aBSM, LeX, Di-LeX (40 pm); GalNAc-b, Rha-b, Tn, GalNAca1-6Galb, Adi, LNT (80 pm); TFdi, Lac, Lec, LSTb, GA1
(164 pm); FABP (320 pm); Gala1-3Gal, GA2di, KLH, OSM (1 nm); GalNAc-a (3 nm); BG-A, BG-B, Lea, BSM, deAcBSM, Gn,
PSA, HSP90, Tgl, AFP, CEA (10 nm)
Con A (Mana, Glca)
ManT (51 pm), Tgl (102 pM), Man-a (200 pm); KLH, CEA (410 pm); Glc-a (820 pm); BSM, AFP, aBSM, deAcBSM, OSM
(3 nm); Maltose, Glc4 (5 nm); oxKLH, aOSM, PSA (10 nm); GlcNAca 1-4Galb, Mana1-6Mana (12 nm)
BG-A (892 pm); Tn , GalNAca1-6Galb, Adi (4 pm); BG-B (7 nm)
aGn (9 nm); Lac (11 nm); LacNAc (22 nm); aBSM, aOSM (44 nm); Gal-a, Tgl (88 nm)
BG-B (55 pm); Gal-a (110 pm); Gb3 (440 pm); Gala1-3Gal (880 pm); aOSM (4 nm)
GlcNAca1-4Galb (110 pm); a-Gn, aBSM (440 pm); aOSM (880 pm); GlcNAc-b (9 pm); Ley, LeX, Di-LeX, BSM, DeAcBSM
(18 nm)
DBL (GalNAca)
ECL (LacNAc)
GS-I (Gal/GalNAca)
GS-II (GlcNAc)
HAA (GalNAca)
HPL (GalNAca)
Tn, GalNAca1-6Galb, Adi, BG-A (7.9 nm); BG-B, aGn, GlcNAca1-4Galb (16 nm); GalNAc-a, OSM, aOSM (32 nm)
aBSM, aOSM (13 pm); Tn, GalNAca1-6Galb, Adi, GlcNAca1-4Gal, BG-A, OSM, aGn (26 pm); GalNAc-a (53 pm); BG-B
(110 pm); BSM (210 pm); DeAcBSM (790 pm); Gn (2 nm); GalNAc-b (3 nm); GlcNAc-b (6 nm)
BG-B (25 nm); BG-A (100 nm)
LBA (BG-A)
LTL (Fuc, Lex)
MPL (GalNAca)
Fuc-a (170 pm); Ley, Lex (340 pm); Di-Lex, CEA, Fuc-b (680 pm)
aBSM, aOSM, aGn (780 pm); Tn (6 nm); BG-B, GalNAc-a, Gal-a, GalNAca1-6Galb, OSM (25 nm)
PNA (Galb1-3GalNAc) a-Gn (11 pm); GA1 (1 nm); TFdi (2 nm); Lac (5 nm); BG-B (6 nm); aBSM (11 nm)
PTL (Gal/GalNAc) BG-A, BG-B (1 nm); Adi (6 nm); Tn, GalNAca1-6Galb, aOSM (11 nm); aBSM (29 nm); GalNAc-a (57 nm)
RCA120 (Gal/GalNAc) Rha-b, a-Gn (21 pm); Lac, Tgl (330 pm); LacNAc, BG-B, LSTc, aBSM, OSM, aOSM, Gn, FABP (1 nm); Gal-a, GalNAca1-
6Galb, Gala1-3Gal, TFdi, Galb1-3GlcNAcb, Gb3, LNT, LSTb, BSM, CEA (2 nm); Gala1-4Galb, BG-H1, KLH, deAcBSM, PSA,
AFP (4 nm); HSA, (5 nm); GalNAc-b, GA1 (10 nm)
SBA (GalNAc)
aBSM, aOSM (33 pm); Tn, Gb3 (65 pm); GalNAca1-6Galb, Adi (130 pm); GalNAc-a (260 pm); OSM (520 pm); GalNAc-b
(1 nm); aGn (4 nm); GA2di (5 nm); BG-A, BG-B, deAcBSM (10 nm); Gal-a, BSM (21 nm); Gn (42 nm)
AFP (420 pm); LSTc (830 pm); BG-B, aOSM, Tgl (3.3 nm); TFdi, GM3, 6’SLac, BSM, aBSM, deAcBSM, OSM, Gn, PSA, CEA,
FBP (6.7 nm); HSA, GalNAc-a, Man3, aGn (13 nm)
SNA (Siaa2-6Gal)
SSA (Tn)
UEA-I (Fuca)
VAA (Galb)
aBSM, aOSM (100 nm)
Ley (6 pm); Fuc-a, Fuc-b (590 pm); BG-B, aGn (6 nm); BG-H1, BSM (12 nm)
aOSM, a-Gn (17 mm)
VFA (Mana, Glca)
VVL-B4 (Tn)
Tgl (200 nm)
aOSM (8 pm); aBSM (15 pm); Tn (60 pm); GalNAca1-6Galb, Adi, OSM (240 pm); GalNAc-a, DeAcBSM, a-Gn (480 pm);
GalNAc-b, GA2di, BSM (719 pm)
WFA (GalNAca/b)
aOSM (11 pm); aBSM (23 pm); Tn (46 pm); GalNAca1-6Galb, Adi (92 pm); GalNAc-b, aGn (180 pm); GalNAc-a, GA2di
(730 pm); Lac, Lec, LNT, LSTb (2 nm); OSM, Gn, Tgl (3 nm) BSM, deAcBSM (29 nm)
GalNAc-a, GalNAc-b, GlcNAc-b, Adi, GlcNAca1-4Gal (36 pm); BG-A, aBSM (140 pm); 3’SLacNAc, aOSM, Gn, (290 pm); Tn,
GalNAca1-6Gal (580 pm); SLea,OSM, a-Gn (1 nm); SLex, GM3, 6’Slac, LSTa, oxKLH, BSM, deAcBSM (5 nm); PSA, Tgl, AFP,
CEA (17 nm); LacNAc (35 nm)
WGA (GlcNAc)
[a] Nominal specificities listed by the commercial suppliers. [b] The detection limit (DL) is the lowest concentration of lectin that produced a signal 5-
times higher than the background. Abbreviations (highest concentration tested): ACL=Amaranthus caudatus lectin (29 nm), BPL=Bauhinia purpurea
lectin (10 nm), Con A=concanavalin A (12 nm), DBL=Dolichos biflorus agglutinin (7 nm), ECL=Erythrina cristagalli lectin (88 nm), GS-I=Griffonia
simplicifolia I (4 nm), GS-II=Griffonia simplicifolia II (18 nm), HAA=Helix aspersa agglutinin (130 nm), HPL=Helix pomatia lectin (13 nm), LBA=Lima
bean agglutinin or Phaseolus lunatus (100 nm), LTL=Lotus tetragonolobus lectin (7 nm), MPL=Maclura pomifer lectin (50 nm), PNA=peanut
agglutinin (46 nm), PTL=Psophocarpus tetragonolobus lectin (57 nm), RCA120=Ricinus communis agglutinin I (10 nm), SBA=soybean agglutinin
(42 nm), SSA=Salvia sclarea agglutinin (400 nm), SNL=Sambucus nigra lectin or eldeberry lectin (13 nm), UEA-I=Ulex europaeus agglutinin I
(12 nm), VAA=Viscum album agglutinin or mistletoe lectin (17 mm), VFA=Vicia faba agglutinin (29 nm), VVL-B4=Vicia villosa lectin B4 (14 nm),
WFL=Wisteria floribunda lectin (29 nm), WGA=wheat germ agglutinin (140 nm).
3.3 nm) and the T disaccharide (Galb1-3GalNAc, DL =
6.7 nm). Interestingly, binding to LSTc, a carbohydrate with
a sialic acid at the nonreducing end, could be inhibited with
the T disaccharide but not with blood group B. Perhaps one of
the most unexpected results was that SNA and RCA120
bound to HSA (non-glycosylated protein used as a control in
our array).[30] HSA competitively inhibited carbohydrate
binding. These results highlight how easily proteins, cells,
and tissues could be mischaracterized when using lectins.
Several other features of lectin binding should also be
mentioned. First, lectins frequently exhibit binding require-
Scheme 1. Structures of selected carbohydrates.
Angew. Chem. Int. Ed. 2006, 45, 3607 –3610
ments beyond simple mono/disaccharide specificity. For
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 3609

Communications
[14] D. M. Ratner, E. W. Adams, J. Su, B. R. OꢀKeefe, M. Mrksich,
P. H. Seeberger, ChemBioChem 2004, 5, 379.
[15] S. Angeloni, J. L. Ridet, N. Kusy, H. Gao, F. Crevoisier, S.
Guinchard, S. Kochhar, H. Sigrist, N. Sprenger, Glycobiology
2005, 15, 31.
example, RCA120 showed little or no binding within the
concentration range tested to a number of Gal and GalNAc
terminal carbohydrates such as GalNAca1-4Galb, Gal-
NAca1-3Galb, Galb1-6Mana, GalNAcb1-4Gal, BG-A, Lea,
and Lex. Although this aspect of lectin recognition is known,
detailed information on secondary binding requirements is
not available for most lectins. Carbohydrate microarrays
provide rapid access to comprehensive binding profiles.
Second, lectin specificity can be difficult to predict based on
binding (or lack of binding) to closely related structures. For
example, the GalNAc binding lectin SBA showed little or no
binding to the Gala monosaccharide and the Gala1-3Gal
disaccharide. However, SBA bound very well to Gb3 (DL =
65 pm, Gala1-4Galb1-4Glc). Therefore, it is critical to eval-
uate binding to a wide range of carbohydrates.
In summary, a carbohydrate microarray was used to
evaluate over 1700 potential lectin–ligand interactions over a
range of lectin concentrations. The format allowed for very
rapid analysis with a minimal amount of expensive and
difficult to obtain carbohydrates. For comparison, an ELLA
assay carried out in 96-well plates would have required at
least 150 plates and 100-fold larger quantities of each BSA
conjugate and protein. The unexpected binding properties
uncovered in this study and the extensive binding information
obtained highlight the utility of carbohydrate microarrays for
rapid evaluation of carbohydrate–protein interactions.
[16] M. B. Biskup, J. U. Muller, R. Weingart, R. R. Schmidt, Chem-
BioChem 2005, 6, 1007.
[17] B. S. Bochner, R. A. Alvarez, P. Mehta, N. V. Bovin, O. Blixt,
J. R. White, R. L. Schnaar, J. Biol. Chem. 2005, 280, 4307.
[18] D. A. Calarese, H. K. Lee, C. Y. Huang, M. D. Best, R. D.
Astronomo, R. L. Stanfield, H. Katinger, D. R. Burton, C. H.
Wong, I. A. Wilson, Proc. Natl. Acad. Sci. USA 2005, 102, 13372.
[19] V. I. Dyukova, E. I. Dementieva, D. A. Zubtsov, O. E. Galanina,
N. V. Bovin, A. Y. Rubina, Anal. Biochem. 2005, 347, 94.
[20] K. S. Ko, F. A. Jaipuri, N. L. Pohl, J. Am. Chem. Soc. 2005, 127,
13162.
[21] J. Manimala, Z. Li, A. Jain, S. VedBrat, J. C. Gildersleeve,
ChemBioChem 2005, 6, 2229.
[22] I. Shin, S. Park, M. R. Lee, Chem. Eur. J. 2005, 11, 2894.
[23] S. J. van Vliet, E. van Liempt, E. Saeland, C. A. Aarnoudse, B.
Appelmelk, T. Irimura, T. B. H. Geijtenbeek, O. Blixt, R.
Alvarez, I. van Die, Y. van Kooyk, Int. Immunol. 2005, 17, 661.
[24] M. Vila-Perello, R. G. Gallego, D. Andreu, ChemBioChem 2005,
6, 1831.
[25] C. Y. Huang, D. A. Thayer, A. Y. Chang, M. D. Best, J.
Hoffmann, S. Head, C. H. Wong, Proc. Natl. Acad. Sci. USA
2006, 103, 15.
[26] J. C. Kehr, Y. Zilliges, A. Springer, M. D. Disney, D. D. Ratner,
C. Bouchier, P. H. Seeberger, N. T. de Marsac, E. Dittmann, Mol.
Microbiol. 2006, 59, 893.
[27] M. M. Ngundi, C. R. Taitt, S. A. McMurry, D. Kahne, F. S. Ligler,
Biosens. Bioelectron. 2006, 21, 1195.
Received: February 14, 2006
Published online: April 26, 2006
[28] Different lectins contain different levels of biotin incorporation.
As the biotin/lectin ratio will affect the detection limit, the
results for different lectins should be compared qualitatively
rather than quantitatively.
[29] Some epitopes contain a thioglycoside linker at the reducing end
of the sugar. Although the linker can potentially affect
recognition, inhibition was also observed with sugar lactols
(see the Supporting Information).
[30] Although possible, several lines of evidence suggest that the
observed binding is not due to a minor glycoprotein impurity.
First, the manufacturer states that the HSA is non-glycosylated,
greater than 99% pure, and essentially free of globulins. Second,
binding to HSA was not observed with other lectins with similar
specificity. Recognition of proteins/peptides by carbohydrate-
binding proteins has been reported previously; for examples, see
A. Pashov, M. Perry, M. Dyar, M. Chow, T. Kieber-Emmons,
Curr. Top. Med. Chem. 2005, 5, 1171 and K. R. Oldenburg, D.
Loganathan, I. J. Goldstein, P. G. Schultz, M. A. Gallop, Proc.
Natl. Acad. Sci. USA 1992, 89, 5393.
Keywords: carbohydrates · glycoconjugates · glycosides ·
.
lectins · microarrays
[1] H. Lis, N. Sharon, Chem. Rev. 1998, 98, 637.
[2] N. Sharon, H. Lis, Glycobiology 2004, 14, 53R.
[3] P. Schoffski, S. Riggert, P. Fumoleau, M. Campone, O. Bolte, S.
Marreaud, D. Lacombe, B. Baron, M. Herold, H. Zwierzina, K.
Wilhelm-Ogunbiyi, H. Lentzen, C. Twelves, Ann. Oncol. 2004,
15, 1816.
[4] B. T. Houseman, M. Mrksich, Chem. Biol. 2002, 9, 443.
[5] E. W. Adams, J. Ueberfeld, D. M. Ratner, B. R. OꢀKeefe, D. R.
Walt, P. H. Seeberger, Angew. Chem. 2003, 115, 5475; Angew.
Chem. Int. Ed. 2003, 42, 5317.
[6] M. Schwarz, L. Spector, A. Gargir, A. Shtevi, M. Gortler, R. T.
Altstock, A. A. Dukler, N. Dotan, Glycobiology 2003, 13, 749.
[7] D. N. Wang, Proteomics 2003, 3, 2167.
[8] O. Blixt, S. Head, T. Mondala, C. Scanlan, M. E. Huflejt, R.
Alvarez, 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. Cummings, N. Bovin, C. H. Wong, J. C. Paulson, Proc. Natl.
Acad. Sci. USA 2004, 101, 17033.
[9] 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.
[10] M. D. Disney, P. H. Seeberger, Chem. Biol. 2004, 11, 1701.
[11] T. Feizi, W. G. Chai, Nat. Rev. Mol. Cell Biol. 2004, 5, 582.
[12] C. Galustian, C. G. Park, W. G. Chai, M. Kiso, S. A. Bruening,
Y. S. Kang, R. M. Steinman, T. Feizi, Int. Immunol. 2004, 16, 853.
[13] L. Nimrichter, A. Gargir, M. Gortler, R. T. Altstock, A. Shtevi,
O. Weisshaus, E. Fire, N. Dotan, R. L. Schnaar, Glycobiology
2004, 14, 197.
3610
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 3607 –3610