1 Introduction

Lectinology can be divided in two major parts, studies on lectins and application of lectins as tools. Studies on lectins are carried out to characterize the structure of lectin molecules, their carbohydrate-binding specificity, conformational/functional properties of their carbohydrate-binding ‘pocket’, and their biological roles. The second area pertains to the application of lectins with known specificity as powerful tools to study the function of glycoconjugates, both in solution and on cell surfaces [1–6]. Lectins have been widely used for preparative and analytical purposes in biochemistry, cell biology, immunology and related areas [1–3, 7]. They are used for structural characterization of glycoconjugates of unknown structure, identification of lectin-reactive structures in biological materials (body fluids, cells, tissues), fractionation and purification of glycoconjugates by affinity chromatography, comparative studies etc. The choice of proper lectin for such studies and its quality as a tool depends well-defined lectin specificity. Many lectins are apparently specific for a monosaccharide, but they react with various oligosaccharide chains terminating with this sugar with different affinities, because the type of linkage and underlying sugar residues also have significance. Quantitative differences in reactivity pattern exist between various lectins recognizing the same terminal monosaccharide residue [4, 5]. There are also lectins specific for more complex structures which are not inhibited by any monosaccharide. In the present article we describe the specificity of lectins studied by us, and the methods used to determine the lectin–carbohydrate interaction.

2 Grouping lectins based on recognition of monosaccharides and oligosaccharides

Lectins that can be used as tools to study the glycobiology system are defined as applied lectins. They must be easily obtained, stable, and well characterized in respect of specificity. Before detailed studies for powerful tools, the carbohydrate specificities of applied lectins are classified into six groups according to their specificities to monosaccharides [4, 5]. They are further subgrouped by the affinities to (a) GalNAcα1→O to Serine(Threonine) (Ser(Thr)) of the peptide chain; (b) mammalian disaccharide structural units; (c) trisaccharides; (d) the number and location of LFucα1→linked to Galβ1→3/4GlcNAc sequence; and (e) α2→, 3/6 linkages of sialic acid. These structures are frequently found in soluble glycoproteins and as cell surface glycoconjugates in mammals. A scheme of the classification based on monosaccharide specificity and oligosaccharide structures (described in Table 1) showing the highest affinity is shown as follows.

  1. I.

    GalNAc-specific lectins

    1. 1.

      F/A, GalNAcα1→3GalNAc (Forssman) and GalNAcα1→3Gal (Blood group A determinant disaccharide)—Dolichos biflorus, Helix pomatia and Wisteria floribunda.

    2. 2.

      A, GalNAcα1→3Gal—Soybean agglutinin (SBA), Lima bean and Psophocarpus tetragonolobus.

    3. 3.

      Tn, GalNAcα1→Ser/Thr—Vicia villosa B4 (VVL-B4) [8] and Salvia sclarea [9].

  2. II.

    Gal-specific lectins

    1. 1.

      T, Galβ1→3GalNAc (\({\mathbf{T}}_\alpha \) or \({\mathbf{T}}_\beta \))—Peanut, Bauhinia purpurea alba [10], Abrus precatorius [11], Agaricus bisporus [12], Sclerotium rolfsii [13], Artocarpus integrifolia (jacalin) [14], Artocarpus lakoocha [15], and Maclura pomifera [16], ricin [17] and Morus nigra (Morniga G) [18].

    2. 2.

      I/II, Galβ1→3/4GlcNAcβ1→—Ricinus communis (RCA1) [19], Datura stramonium (Thorn apple), Wheat germ (WGA) [20], Erythrina cristagalli [21], Erythrina corallodendron [22] and ricin [17].

    3. 3.

      B, Galα1→3Gal—Griffonia (Bandeiraea) simplicifolia B4 [23].

    4. 4.

      E, Galα1→4Gal—abrin-a, mistletoe toxic lectin-I (ML-I) [24] and Aplysia gonad [25, 26].

  3. III.

    Man and/or Glc-specific lectins recognizing complex N-linked oligosaccharides—Concanavalin ensiformis (Jack bean), Lens culinaris, Pisum sativum, Hippeastrum hybrid, Narcissus pseudonarcissus and Morniga M [27].

  4. IV.

    GlcNAc, and/or Galβ1→4GlcNAcβ1→linked specific lectins

Chitin oligosaccharide specific agglutinins—WGA [20], G. (Bandeiraea) simplicifolia II (GSA-II), Solanum tuberosum, Ulex europaeus (UEA) and D. stramonium.

  1. V.

    LFuc-specific lectins (subgroups based on the numbers and location of LFucα linkage)

    1. 1.

      Monofucosyl-specific agglutinins (blood group H, Lea and/or Lex)—UEA-I, UEA-II, Pseudomonas aeruginosa [28, 29] and Anguilla anguilla [30].

    2. 2.

      Difucosyl-specific agglutinins (Leb and Lex/y) – GSA-IV and Lotus tetragonolobus.

    3. 3.

      Others requiring further characterization—Salmonella typhimurium and Ulva lactuca.

  2. VI.

    Sialic acid specific lectins (subgroups based on the recognized linkage of SA)

    1. 1.

      SAα2→6Gal(NAc)—Sambucus nigra [31], Trichosanthes japonica and ML-I [24].

    2. 2.

      SAα2→3Gal—Agrocybe cylindracea and Maackia amurensis [32].

    3. 3.

      Others requiring further characterization—Limax flavus [33], Limulus polyphemus [34] and achatinin-H [35].

Profiles of the binding properties of lectins, based on the affinity of decreasing order of mammalian glycotopes (determinants) is probably one of the easy ways to express carbohydrate specificity and should facilitate the selection of lectins as structural probes for studying mammalian glycobiology.

Eleven mammalian structural units, shown in Fig. 1 and Table 1, have been selected to express the binding properties of Gal and GalNAc specific lectins. They are: (1) F, GalNAcα1→3GalNAc; (2) A, GalNAcα1→3Gal; (3) T, Galβ1→3GalNAc; (4) I, Galβ1→3GlcNAc; (5) II, Galβ1→4GlcNAc; (6) B, Galα1→3Gal; (7) E, Galα1→4Gal; (8) L, Galβ1→4Glc; (9) P, GalNAcβ1→3Gal; (10) S, GalNAcβ1→4Gal; and (11) Tn, GalNAcα1→Ser (Thr) of the peptide chain. Thus, the information on carbohydrate specificity of Gal/GalNAc reactive lectins, divided into classes according to their highest affinity for one (or two) of the above disaccharide monomers and/or Tn residue, can be completed with data showing an order of reactivity with other disaccharides (Table 2) [4]. These data demonstrate that lectins reacting most strongly with the same oligosaccharide may show differences in order of reactivity with other structures that reflect subtle differences in lectin fine specificity.

Fig. 1
figure 1figure 1

Mammalian glycoconjugates structural units used to express and classify the carbohydrate specificity of lectins (Adopted and modified from Wang and Wu 2007 [81])

Full size image
Table 1 Carbohydrate structural units in mammalian glycoproteins and glycosphingolipids [4]
Full size table
Table 2 Expression of binding properties of Gal/GalNAc reactive lectins by carbohydrate structural units
Full size table

3 Effect of polyvalence of glycotopes in glycan on lectin–glycoform interaction

During the past two decades, it has been observed that many multi-branched oligosaccharides exhibit a significant increment in lectin binding reactivity as compared to their linear counterparts [36, 37]. Based on the results of previous studies, the concept of glycoside cluster effect can be classified into two groups: (a) the ‘multi-antennary or simple glycoside cluster effect’ as in reaction of galactosides with hepatic lectin [38, 39] and tri-antennary II sequences reactive with a galectin from chicken liver (CG-16) [40], or Tn glycopeptides. The molecular sizes of these ligands are usually less than 1.5 × 104. The second group is (b) the ‘high-density polyvalent or complex glycoside cluster effect’, such as polyvalent Tn in asialo ovine submaxillary mucin, which generates an enhancement in affinity with VVL-B4 by 3.3 × 105 and 4.5 × 103 times over Gal and GalNAc, respectively, and is about 1,000 times more active than monomeric Tn [8]. In Table 3, the much stronger inhibition of RCA1 by a panel of cyst glycoproteins than by disaccharides and galactose is shown. A similar phenomenon was also observed with some animal lectins [41]. However, the polyvalences of glycotopes do not always significantly affect carbohydrate protein binding. For example, the potency in the interaction of P. aeruginosa II lectin with LFucα1 polyvalent glycans is about as strong as or weaker than the incremental increase by carbohydrate specificity of monomers [28]. Therefore, to obtain a comprehensive picture of the carbohydrate specificities of a lectin in order to elucidate its functional roles and biomedical applications, the following information should be provided: (1) monosaccharide specificity (Gal, GalNAc, GlcNAc, Man, LFuc, and sialic acid), (2) reactivities to mammalian disaccharides and Tn structural units (in decreasing order; Table 2), (3) the most active ligand, (4) simple multivalent or cluster effect of carbohydrate structured units such as Tn glycopeptides and multi-antennary glycotopes to inhibit binding, and (5) complex multivalent or cluster effects present in macromolecules with known glycotopes.

Table 3 Inhibitory potential of human ovarian cyst glycoproteins estimated by amount of cyst glycoproteins giving 50% inhibition of binding of RCA1 (2.5 ng/50 μl) to asialo human α1-acid gp (25 ng/50 μl)a [19]
Full size table

The varying reactivities of lectins toward mammalian glycotopes and other sugar units, may reflect the possible existence of different combining sites or subsites in the same molecule, and the differential binding properties of these combining sites (if any) have to be characterized. Furthermore, effects of polyvalence on binding have to be studied. Establishing the relationship among the amino acid sequences of the combining sites of plant lectins and mammalian glycotopes can also be an important direction to be addressed in lectinology.

4 Overview of quantitative methods used to determine lectin–carbohydrate interactions

A similar methodical approach is used for the determination of lectin specificity and application of lectins as tools. It concerns the problem how to measure glycoconjugate–lectin interaction.

In early research, when most lectins were detected by agglutination of erythrocytes (and for this reason were commonly called agglutinins), determination of lectin specificity by semiquantitative hemagglutination-inhibition assay was most frequently used. A more accurate approach was binding of radiolabelled lectins to erythrocytes or other cells [42].

A historical method, which did not require cells, was the quantitative precipitation assay (originally developed for antigen–antibody complexes) introduced in Kabat’s laboratory [43]. Lectins precipitate serially diluted high molecular weight glycoconjugates and precipitation is inhibited by various mono- and oligosaccharides. Application of quantitative precipitin/precipitin-inhibition assays (QPA/QPIA) to study the binding properties of lectin glycotopes has several limitations, especially the fact that relatively large amounts of reagents (lectins, glycoproteins and inhibitors) are required. However, due to the accuracy of these methods, various versions have been used for many years in several laboratories. QPA and QPIA have also been successfully used by us to characterize the binding properties of lectins between the early 1970s and 1990s [44–46], until they were replaced by solid phase assays.

The enzyme-linked immunosorbent assay (ELISA), performed on 96-well microtiter plates, was introduced for determination of antibody activity. In the frequently used version, the antibody bound to the antigen-coated surface of the well was detected using a proper secondary enzyme-linked anti-immunoglobulin antibody. Direct adaptation of this method to lectins was initially based on binding the lectin followed by binding an anti-lectin antibody [47]. However, when several lectins are studied, obtaining an antibody for each lectin is troublesome and this technique was replaced by using labeled lectins, which could be detected by a universal anti-label reagent. One of the reagents for labeling lectins is digoxigenin (DIG) that can be detected with enzyme-linked anti-DIG antibody [48]. This method is relatively popular, because anti-DIG antibodies and DIG-labeled lectins are commercially available. An alternative and very convenient method, in which no antibody is used, is based on the biotin/avidin system. Lectins are biotinylated and detected with enzyme-conjugated avidin (alkaline phosphatase or horseradish peroxidase are most frequently used enzymes). This method was introduced in our laboratories [49] and is described in further part of this article.

In recent years, the microtiter plate methods have lead to micro-array techniques which have been described in rapidly increasing number of publications [50–55]. Microarrays are solid surfaces which can carry a great number of glycans bound (covalently or noncovalently) as small spots. Binding of lectins to immobilized glycans allows rapid and exhaustive analysis of their specificity and is useful for detecting even weak interactions due to multivalent display of ligands on the surface. The various methods described in the literature differ in the ways of derivatization and immobilization of glycans and of determination of lectin binding (frequently fluorescence-based methods were used).

Interaction of lectins with glycoconjugates is also measured by affinity chromatography, using columns of immobilized lectins. The greater the affinity of oligosaccharide or glycopeptide to the lectin, the slower its elution from the column [56]. This method is useful not only for characterization of oligosaccharides, but also for their effective fractionation by using sequential chromatography on various lectin columns [57, 58]. Another approach is frontal affinity chromatography (FAC) [59], in which an excess volume of diluted fluorescently labeled glycan is applied to a lectin column. When the column is saturated with the glycan, it starts to leak and concentration of the glycan in the eluate increases, reaching the concentration equal to that of the applied solution. Comparison of the elution volumes of a ligand not reactive with the lectin and the glycan studied, allows detection of low-affinity interactions and calculation of dissociation constants. Recently, an automated FAC system has been developed, which allows fast analysis of multiple lectins and glycan samples [60].

An excellent method for observation of lectin–carbohydrate interaction ‘in real time’ is surface plasmon resonance (SPR) [61]. In this method one partner of the reaction is immobilized on a sensor chip channel and solution of the second partner flows over the sensor chip (association stage) that is followed by passing a buffer (dissociation stage). This method allows to follow the interaction and to determine the kinetic and thermodynamic parameters of the reaction. The SPR method was initially designed for measuring interaction between macromolecules, but recently techniques allowing to measure interactions of proteins with small molecules (e.g. lectins with glycans) have been developed. However, the SPR method requires an access to rather expensive highly computerized BIAcore or other instrument.

5 Biotin/avidin based microtiter plate enzyme-linked lectinosorbent assay

Efficient noncovalent binding of glycoproteins or other high-molecular weight glycoconjugates to MaxiSorp plates allows determination of binding of biotinylated lectins to the coated wells by using enzyme-conjugated ExtrAvidin (modified avidin from Sigma) and a proper enzyme substrate [49]. Biotinylation of lectins via amino groups is a simple procedure and does not alter their binding capacity [62]. Comparison of lectin binding to different immobilized glycoconjugates (glycoproteins or conjugates of an oligosaccharide with a carrier) or binding of various lectins to one glycoprotein (an example of the latter version is shown in Table 4) can be compared. However, binding assay may give rise to some uncertainties due to possible differences in coating efficiency or orientation of immobilized ligands. This problem can be overcome by binding of a lectin to one selected ligand and inhibition of the binding not only with oligosaccharides and monosugars, but also with high molecular weight glycoconjugates. This allows comparison of interaction of lectins with various monomeric and multivalent glycoforms in solution [8, 12, 18, 49, 63–65].

Table 4 Characterization of binding properties of native and asialo-cyst glycoproteins (HOC 350) with established lectins by ELLSA [64]
Full size table

This method is accurate, sensitive and allows testing of multiple glycoconjugates [64, 65]. In this system, the amount of reagents required is greatly reduced (about 1,000-times) as compared to that needed for precipitation assays. When starting experiments with a new lectin, it is important to select a proper coating concentration of the target glycoform, because frequently with increasing coating concentration binding of the lectin increases to some point, and then strongly decreases [10, 49, 66]. Optimal binding usually occurred when solution used for coating contained the ligand at 0.2–2μg/ml concentration (10–100ng/50μl/well).

The biotin/avidin-based enzyme-linked lectinosorbent assay may be combined with partial deglycosylation of the glycoprotein by Smith degradation, which can be performed in solution or directly on the plate. For example, when Gal/GalNAc-specific lectins were studied, the plate was coated with asialoglycophorin A (asialo-GPA) containing Galβ1→3GalNAcα→ chains (T antigen), which was efficiently transformed on the plate into asialo-agalacto-GPA (Tn antigen) [49]. An ‘artificial’ Tn antigen obtained from GPA has several applications [67], it can replace the natural Tn antigen from rare Tn erythrocytes in studying Tn-specific lectins and antibodies. Smith degradation combined with lectin studies can also be applied to other glycoproteins to obtain information on the structure of glycans.

The biotin/avidin-based plate assay facilitated the study of fine specificity of multiple plant and bacterial lectins, where interaction of each lectin with several tens of high- and low-molecular mass glycoforms was examined [12, 18, 19, 21, 28, 68]. Similar studies were performed on galectins, rat galectin-4 (G4-N) and galectin-5, and chicken liver galectin CG-16 [40, 41, 63, 69, 70]. Generally, galectins recognize glycoforms with a terminal β-galactose residue. However, determination of their fine specificity showed quantitative differences in interaction with various oligosaccharide structures. For example, CG-16 showed preference in reactivity with oligosaccharides terminated with Galβ1→4GlcNAc unit, while G4-N reacted better with Galβ1→3GlcNAc-terminated ones. These subtle differences may have significance for functions of lectins in vivo and the knowledge of lectin specificity has potential for medical applications.

5.1 Methods

5.1.1 Biotinylation of lectins

Lectins were biotinylated via amino groups with biotinamidocaproate-N-hydoxysuccinimide ester (Sigma), which will be henceforth termed a biotin ester. A 0.025% solution was obtained by dissolving 500μg of biotin ester in 50μl methanol and mixing this sample with 1.95ml phosphate-buffered saline (PBS; 0.02M phosphate buffer/0.15M NaCl, pH 7.4). A 0.1% lectin solution in PBS was mixed with double volume of biotin ester solution (weight ratio of lectin to the ester should be 2:1) and after 30min at room temperature the sample was dialyzed for several hours against water and overnight against Tris-buffered saline (TBS; 0.05 M Tris–HCl buffer/0.15M NaCl, pH 7.4). The sample was diluted with TBS and 20% sodium azide was added to a final concentration of 0.02% lectin and 0.1% NaN3. This solution can be stored at 4°C for several months without loss of activity.

5.1.2 The microtiter plate lectin assay

The volumes of all reagents applied to the plate (Nunc, MaxiSorp) were 50μl/well, and all incubations, if not stated otherwise, were performed at room temperature using reagents diluted with TBS containing 0.05% Tween 20 (TBS-T). The plates were coated with a target ligand in 0.05M carbonate buffer, pH 9.6 overnight at 4°C. After washing the plate, biotinylated lectins were incubated in the wells for 30–60min. The plates were washed and the ExtrAvidin–alkaline phosphatase solution (Sigma, diluted 1:10,000) was added. After 1h the plates were washed at least four times and incubated with p-nitrophenyl phosphate (Sigma 104 phosphatase substrate 5-mg tablets) in 0.05M carbonate buffer, pH 9.6, for 30min, or longer, and the absorbance at 405nm was read in a microtiter plate reader (Fig. 2).

Fig. 2
figure 2

Principle and an example (RCA1) of (a) biotin/avidin based microtiter plate enzyme-linked lectinosorbent (ELLSA) and (b, c) inhibition assays as tools for characterizing binding properties of soluble glycoproteins, their inhibitory potential and/or determining carbohydrate specificities of lectins [19, 64]. a The uncharacterized glycoprotein immobilized on a microtiter plate can bind an established biotinylated lectin, for e.g. R. communis agglutinin-1 (RCA1), which can recognize glycotopes (Galβ1→4GlcNAcβ) of unknown glycans. The overall activity of ExtrAvidin–alkaline phosphatase complexed to the biotinylated lectin can be estimated. b The biotinylated (known) lectin reacting with small haptens (M.W. < 1.5 × 103) or clusters (M.W. 1.5 × 103 to 1.5 × 104) in solution can compete with a known coated glycoprotein for binding. Similarly, c if a series of well-characterized glycans are precoated in the wells of a microtiter plate, an uncharacterized glycan (M.W. > 1.5 × 104), containing polyvalent glycotopes, can be used as an inhibitor to block the binding and its inhibitory potential toward known lectins can also be determined

Full size image

For inhibition studies, the serially diluted inhibitor samples were mixed with an equal volume and constant concentration of lectin solution (can be done in a separate microtiter plate). The control lectin sample was diluted twofold with TBS-T. The lectin concentration used for inhibition should be selected from the binding curve of serially diluted lectin to the plate coated with constant ligand concentration, and should be located below the saturating concentration, usually in the upper part of the decreasing arm of the curve. After 30min incubation at room temperature, the samples were tested in the binding assay. The inhibitory activity is expressed as the concentration of inhibitor giving 50% inhibition of the control lectin binding. All experiments were done at least in duplicate and the standard deviation was usually less than 5% of the mean value. The control wells, where addition of biotinylated lectin or coating were omitted, served as blank (absorbance below 0.1).

5.1.3 Smith degradation on ELISA plates

Wells coated with asialoglycoprotein were filled with 50μl of 0.05M sodium periodate in 0.05M acetate buffer, pH 4.5, and left overnight at 4°C (oxidation of sugars having two vicinal free hydroxyl groups). The plates were washed with TBS, and incubated with 50μl of 0.15M NaBH4 in 0.1M borate buffer, pH 8.0, for 2–3h (reduction of aldehyde groups formed during oxidation). After washing, the plates were filled with 50μl of 0.025M sulfuric acid, tightly covered with parafilm and plastic cover, and incubated overnight at 37°C (selective release of oxidized/reduced sugars). After washing with TBS the plates were ready for the binding assay.

6 Lectinoblotting and modifications lectin ligands on the blots

This method allows identification of lectin-binding components in cell extracts or other biological material, which is fractionated by electrophoresis and blotted to nitrocellulose or polyvinylidene fluoride (PVDF) membranes. It is a qualitative method which is used to detect the lectin-binding glycoproteins in mixtures, e.g. to compare the lectin-binding patterns to glycoproteins of different cells, or cells at different stages of differentiation. The use of biotinylated lectins allows detection of reactive glycoproteins with high sensitivity. This method can also be combined with desialylation and Smith degradation of glycoproteins on the blot, thus allowing detection of cryptic lectin-reactive structures. Degradation on the blot is convenient, because modified glycoproteins are detected at their original positions that may further facilitate their identification. We first used this method to detect glycophorin Mi.III hybrid variant, which is relatively frequent in Asian populations [71]. Glycophorins are unique densely O-glycosylated red cell membrane glycoproteins which, after desialylation or desialylation/Smith degradation, can be specifically detected on the blots of erythrocyte membrane lysates with anti-T and anti-Tn lectins, respectively. Later, the method was used to characterize glycans of chicken glycophorins with lectins [72].

Some lectin-reactive glycotopes (e.g. ABH/Lewis) are present on O-glycans and N-glycans. To identify the characteristic of glycans to which the lectin is bound, the β-elimination of O-glycans by mild alkaline treatment of the blots was performed [73]. Nitrocellulose sheets are mechanically destroyed by alkaline treatment, and Immobilon P (PVDF) transfer membranes, which are not affected, were used. Best results were obtained by treating the blots with 0.055 MNaOH for 16h at 40°C. Sodium borohydride used in the β-elimination procedure was omitted because in these experiments the reduction of the released O-glycans was not necessary. The results obtained with erythrocyte membrane glycoproteins (used as a model) showed that peanut lectin detected specifically all glycophorin bands on desialylated blots, and no bands were detected on desialylated and then alkali-treated blots. On the other hand, the reactions of red cell membrane glycoproteins with Phaseolus vulgaris lectin (recognizing N-glycans) and anti-peptide antibodies were not affected. This method was used to show that sialyl-Lewisa receptors in human colon carcinoma cell line CX-1.1 were carried mainly by O-glycans of mucin-type glycoproteins.

The chemical modifications are cheaper and frequently more efficient than enzymatic deglycosylation with the use of glycosidases. However, Smith-degradation may be less conclusive than sequential glycosidase digestion, when applied to a glycoprotein with glycans of unknown structure.

6.1 Methods

6.1.1 Modifications of glycoproteins on blots

To desialylate glycoproteins, blots (nitrocellulose or Immobilon P) were incubated in 0.025M sulfuric acid for 1h at 80°C and washed with water. Smith degradation was performed by successive incubations of the desialylated blots in (1) 0.05M NaIO4 in 0.05M acetate buffer, pH 4.5, overnight at 4°C; (2) 0.15M NaBH4 in 0.1M sodium borate buffer, pH 8.0, for 2–3h at room temperature; (3) 0.025M sulfuric acid for 1h at 80°C. The blots were washed with water between the incubations. For β-elimination, the untreated or desialylated Immobilon P blots were overlaid with 0.055M NaOH, left for 16h at 40°C, and washed with water.

6.1.2 Detection of glycoprotein bands on blots with biotinylated lectins

The blots were blocked with 5% bovine serum albumin (BSA) for 1h at 37°C, or overnight at 4°C, and were overlaid with biotinylated lectin solution (usually 5–10μg/ml) in TBS-T for 1h at room temperature. The blots were washed 5-times for 10min each with 0,15M NaCl and overlaid with ExtrAvidin–horseradish peroxidase conjugate (Sigma; diluted 1:1,000 with TBS-T) for 1h at room temperature. After five to six washes, an enzymatic reaction was developed at room temperature with 4-chloro-1-naphthol (Sigma; 6mg in 2ml methanol, diluted with 10ml TBS and treated with 6μl H2O2). Alternatively, ExtrAvidin–alkaline phosphatase conjugate and its substrate, nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (both from Sigma) can be used.

7 Interactions of lectins with glycosphingolipids

Glycosphingolipids have unique oligosaccharide sequences, but also share several glycotopes with glycoproteins [74]. Insolubility of glycolipids in water requires a different experimental approach to study their interaction with lectins and antibodies [75]. Glycolipids can be tested in a microtiter plate assay, but are applied in organic solvent and solvent-resistant plates must be used. Vinyl Assay Plates (COSTAR) are suitable for this purpose. Glycolipids are dissolved in a minimal volume of chloroform/methanol (2:1) and diluted with methanol (concentration of chloroform should not be higher than 5%). These samples are applied to the plates (20μl/well) and after organic solvents evaporate, the plates are ready to use. A commonly used method for fractionation of glycolipids is thin-layer chromatography (TLC) and detection of glycolipids with lectins on the TLC-plates. This method was very helpful in our studies on unique glycosphingolipids identified in rare polyagglutinable NOR erythrocytes.

Preliminary results suggested that the antigen responsible for polyagglutination of NOR erythrocytes is a neutral glycolipid. Therefore, the pool of neutral glycolipids from NOR erythrocytes was fractionated on a TLC-plate and probed with biotinylated lectins (Fig. 3). Three unusual glycolipid bands, absent in control erythrocytes, were detected, two strongly stained by G. simplicifolia isolectin B4 (GSL-IB4, specific for Galα), and one by SBA (specific for GalNAcα/β). It greatly facilitated the isolation of these glycolipids and their structural studies, which showed them to be a globoside elongated with Galα1-4 (NOR1), Galα1-4GalNAcβ1-3Galα1-4 (NOR2) and GalNAcβ1-3Galα1-4 (NORint), respectively [76, 77].

Fig. 3
figure 3

Detection of human erythrocyte glycosphingolipids with lectins on TLC-plates. The plates containing fractionated glycolipids from blood group A or NOR erythrocytes were stained with orcinol or overlaid with lectins: RCA1, R. communis agglutinin-1 recognizing Galα/β-related ligands; GSL-IB4, G. simplicifolia isolectin B4 specific for Galα-terminating oligosaccharides; SBA, soybean agglutinin specific for GalNAcα/β; HPA, H. pomatia agglutinin specific for GalNAcα. NOR1 and NOR2, NOR-related glycolipids, both terminating with Galα1-4GalNAcβ1-3Galα-

Full size image

During these studies the following observations were made: GSL-IB4 did not react on the plate with Pk glycolipid (Galα1-4Galβ1-4Galβ1-Cer), and SBA did not or very weakly react with globoside (GalNAcβ1-3Galα1-4Galβ1-4Glcβ1-Cer; Fig. 3). These glycolipids contained proper terminal sugars and were present in much greater amount than NOR-related longer glycolipids mentioned above, were, strongly reactive with these lectins. A possible reason is the short oligosaccharide chain in Pk or globoside, and the inhibitory effect of closely located hydrophobic ceramide residue. However, it may not apply to all lectins. For example, Moluccella laevis lectin (anti-Tn) surprisingly bound to the Pk glycolipid on the TLC plate, but not in the microtiter plate assay [78]. Comparison of the activity of lectins recognizing an α or α/β form of the same sugar gave surprising results. R. communis agglutinin 1 (specific for Galα/β related ligands) reacted on the TLC plate with the NOR1 glycolipid only, while GSL-IB4 (recognizing Galα residue) reacted strongly with NOR1 and NOR2 glycolipids (Fig. 3) [79]. Similarly, SBA recognized only the shortest blood group A glycolipid (migrating in the pentaglycosidic glycolipid region), while H. pomatia agglutinin (specific for GalNAcα1-3Gal(NAc)β) reacted strongly not only with the same blood group A glycolipid, but also with its elongated homologs (Fig. 3) [77]. These results indicate that binding of lectins to glycosphingolipids is influenced (in both directions?) by the lengths of their oligosaccharide chains. This binding may also depend on the method used, presumably due to different conformation or orientation of glycosphingolipids on various surfaces. Therefore, the results of lectin binding to glycosphingolipids should be interpreted with caution, and influence of various factors on this binding is currently unknown.

7.1 Method

7.1.1 Binding of lectins to glycosphingolipids on thin-layer plates

Glycolipid samples solubilized in chloroform/methanol (2:1, v/v) were applied to high-performance TLC plates (Kieselgel 60, Merck) and developed with a solvent system selected for glycolipids studied. Erythrocytes neutral glycolipids were fractionated in chloroform/methanol/water (55:45:10, v/v/v). The dried plates were immersed in 0.05% polyisobutylmethacrylate (Aldrich) in hexane for 1 min, dried again, sprayed with TBS and immersed in 5% BSA for 1 h. The plates were overlaid successively with (1) biotinylated lectin solution (5 μg/ml) in TBS-T containing 1% BSA for 1 h; (2) ExtrAvidin–alkaline phosphatase conjugate (Sigma) diluted with TBS-T/BSA for 1 h; (3) a substrate solution (nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate, Sigma) until development of color.

8 Conclusions

Lectins with well defined specificity are excellent tools for several analytical and preparative purposes. They are used to identify and characterize the structure of glycans, when the amount of biological material is not sufficient for instrumental techniques, such as mass spectrometry or nuclear magnetic resonance, or lectins may be supplementary to instrumental research. The methods for determination of lectin–carbohydrate interaction described here do not cover the full spectrum of assays used. They are limited to the most frequently used techniques measuring interaction of lectins with soluble (or solubilized) lectin ligands, with special attention to the methods applied in our laboratories. There are several techniques for detection or determination of lectin-reactive glycoconjugates in cells or tissues, such as fluorescence-activated cell sorter, histochemistry, and others. The choice of the method depends on the purpose and material studied. Use of biotinylated lectins is convenient in many assays. Based on practical and economic considerations, the speed and the accuracy of the assay, the range of binding vision, and the amounts of glycans, lectins and ligands required, the highly sensitive microtiter plate assay by using biotinylated lectins should be one of the best approaches to analyze the specificity of combining sites of lectins and/or to characterize the binding properties (glycotopes) of complex carbohydrates and carbohydrate haptens. This system has been adopted in our labs for over a dozen years and successfully used for determination of specificity of several tens of lectins. As all methods have their own limitations, other methods can be used for advanced purposes, like crystallography for characterization of lectin-binding site, SPR for kinetic studies and other techniques.