Background Previous studies have demonstrated that nearly all proteins are glycosylated, and that the oligosaccharide structure of the protein has an important structural, functional and regulatory role.

One of the main mechanisms by which oligosaccharide portions of glycoprotein interact on a molecular level is through binding to specific receptors called lectins. As the level of knowledge about the structure and function of glycosylated proteins

is expanding, it is increasingly important to be able to analyse the properties of lectins.

It is known that lectins are involved in numerous vital physiological processes. For example, selectins mediate lymphocyte adhesion to inflamed endothelium, and mannose-binding lectin functions as a part of the host defence system.

Lectin activity is being discovered in a growing number of previously known proteins and there are still whole families of lectins, for example, galectins, whose role remains obscure even after extensive research.

Examples of previously discovered lectins and their respective typical ligands and functions include: Lectin Typical Ligands Examples of functions Group Protein sorting in. the Calnexin GlclMang endoplasmic reticulum Endoplasmic reticulum- M-type M-type Man8 associated degradation of lectins glycoproteins L-type Protein sorting in the Various lectins endoplasmic reticulum P-type Man 6-phospate Protein sorting post Golgi lectins Cell adhesion (selectins) C-type Glycoprotein clearance Various lectins Innate immunity (Collectins)

Lectin Typical Ligands Examples of functions Group Glycan crosslinking in the Galectins ß-Galactosides extracellular matrix 1-type Sialic acid Cell adhesion (Siglecs) Lectins (Glc = Glucose; Man = Mannose) A major problem in the study of lectins is the lack of adequate methods to analyse their activity.

Current methods are based on either affinity chromatography, or the use of radiolabelled glycoconjugates."Glycoconjugate"is a generic term for biological macromolecules containing a carbohydrate moiety. Examples of such glycoconjugates comprise, amongst others, glycolipids, glycoproteins and proteoglycans.

With a few rare exceptions, most methods to study lectins require the lectin to be purified before its properties can be analysed satisfactorily.

When used to isolate lectins, glycans need to be linked to an appropriate tag because their binding to lectins is difficult to detect. Biotin (vitamin H) is a widely used non-radioactive tag, whose main advantage is the ability to bind proteins based on avidin. In addition, the multimeric nature of avidin-biotin interactions results in clustering of the reporter enzyme molecules and amplified sensitivity. However, there is a significant drawback to using biotin. Biotin is an endogenous prosthetic group for ubiquitously expressed

carboxylase enzymes, for example pyruvate- carboxylase. Since endogenous biotin/enzyme complexes cannot be differentiated from biotin that is used as a label, all methods that use biotin- containing conjugates are prone to give false- positive results when investigating complex biological mixtures. This is clearly demonstrated in Fig. 1, where endogenous biotin in rat tissue homogenates is visualised using a streptavidin- biotinylated alkaline phosphatase complex. Several bands can be seen, and these could create a false positive result in any assay system that uses a biotin label.

Digoxin is a cardiac glycoside, originally derived from the foxglove (Digitalis). The structure of digoxin is shown in Fig. 2. Digoxin has been used as a drug for centuries to treat circulatory disorders. As digoxin can be toxic if taken at too high a dosage, highly specific and strongly-binding monoclonal and polyclonal antibodies have been developed as an antidote to treat digoxin overdose.

These antibodies can be used to detect both digoxin and digoxigenin (a deglycosylated form of digoxin), when they are labelled with an appropriate enzyme or other molecule. The digoxin and detector molecule complex forms an effective label for polysaccharides, which eliminates the problem of false positive results.

Digoxin was previously considered unsuitable for use as a label as the high acid sensitivity of the

glycosidic bonds between the digitoxoses, and the alkaline sensitivity of the lactone ring, has impeded chemical derivatization of this molecule and therefore its use as a label.

STATEMENT OF THE INVENTION According to the present invention there is provided a digoxin-labelled glycan which comprises a digoxin moiety having a glycosidic chain attached to a steroid moiety and a glycan moiety.

Preferably the glycan moiety is attached to said digoxin moiety through an amino group.

It is further preferred that the digoxin is covalently bound to the glycan.

It is further preferred that the glycan is bound to the extremity of the glycosidic chain of the digoxin moiety which is at the opposite end to said steroid moiety and advantageously at position 4 of the last digitoxose ring (as shown in figures 3 and 4).

The term digoxin moiety is used to encompass other molecules which achieve the same function as the specific molecule identified. For example, the term includes analogues or mimetics of the named compounds which contain substitutions and/or structural variations. Modifications to the structure of digoxigenin or digoxin, for example by adding one or more alkyl groups, metal ions or

other prosthetic groups or substituents, would be obvious to the person skilled in the art and such equivalents would fall within the scope of the present invention. Exemplary functional equivalents of digoxin would include"digoxin-like" commpounds having a steroid structure and a glycosidic chain linked thereto.

The term glycan is used to encompass not only oligosaccharide structures but also monosaccharides and sugar based polymers like polysaccharides.

According to another embodiment of the present invention there is provided a method of labelling a glycan with a digoxin moiety, said method comprising the steps of: 1. reacting digoxin with CNBr to form activated digoxin; 2. reacting the glycan to be labelled with the activated digoxin under suitable conditions to form a digoxin-labelled glycan; and optionally 3. separating the digoxin-labelled glycan from the reaction mixture.

Preferably the glycan is reduced prior to reaction with activated digoxin. More preferably the glycan is reductively aminated.

Reductive amination of the glycan can conveniently be achieved by reacting the glycan with hydrazine, a diamine or a dihydrazide. Preferably hydrazine is used because unreacted hydrazine can be

conveniently removed by repeated evaporation from toluene. Preferably substantially all the glycan to be labelled is reductively aminated.

Preferably the hydrazine, diamine or dihydrazide is provided in a large excess relative to the glycan.

For example the molar amount of hydrazine may be from 2 to 100 fold (preferably 2 to 20 fold) to that of glycan.

Preferably the reaction between the activated digoxin and the reduced glycan proceeds substantially to completion, i. e. virtually all reduced glycan is bound to digoxin. It is further preferred that, during this reaction, repeated additions of activated digoxin are provided. This permits maintaining levels of activated digoxin in the reaction mixture. According to a further embodiment of the present invention there is provided the use of a digoxin-labelled glycan as a probe to study the properties of a macromolecule.

Preferably the macromolecule is a biomolecule such as a lectin or other glycan-binding species.

According to yet a further embodiment of the present invention there is provided a method of analysing a glycan-binding macromolecule in a mixture of undetermined or partially determined biomolecules (like a natural extract), said method comprising:

1. exposing said mixture to a digoxin-labelled glycan under suitable condition to form a digoxin-labelled glycan/macromolecule complex; 2. removing unbound digoxin-labelled glycan; and 3. determining the presence and/or amount of the digoxin-labelled glycan/macromolecule complex.

Suitably in a preliminary step the mixture of biomolecules are first separated using, for example, ELISA, dot-blotting or histological assays. Electrolysis could also be considered for separation but the fact that the proteins have to be denatured would not make electrolysis a particularly preferred mode of preparation.

Preferably the biomolecules are blotted onto a suitable membrane, such as Immobilon PVDF membrane or nitrocellulose, prior to exposure to the digoxin-labelled glycan.

Suitably the glycan-binding macromolecule is a lectin.

The digoxin-labelled glycan of the present invention can be used in any type of assay in which a glycan probe may generally be used. These include but are not restricted to, blotting (e. g onto PVDF membrane), multiwell (e. g. ELISA), slot blotting or histological assays. In addition digoxin-glycans may be used to study glycan-binding molecules on the membranes of animal cells, plant cells and micro-organisms or on the surface of any tissue.

Preferably the presence and/or amount of digoxin- labelled glycan is visualised by exposing the mixture to a digoxin specific probe containing a chromogenic, chemoluminescent or radioactive marker. Preferably an anti-digoxin antibody linked to alkaline phosphatase is used.

According to an alternative embodiment of the present invention there is provided a method of analysing a glycan-binding macromolecule in a mixture of biomolecules, said method comprising: 1. adding digoxin-labelled glycan to a mixture of biomolecules under suitable conditions to form digoxin-labelled glycan/glycan-binding macromolecule complexes; 2. exposing the complexes thus formed to an immobilised anti-digoxin antibody; and 3. removing unbound digoxin-labelled proteins.

An example of the present invention will now be described, by way of example only, with reference to the accompanying drawings in which: Fig. 1: Endogeneous biotin in rat tissue: shows the results of a Western blot of various homogenates (10 Rg protein per lane) of rat tissues (1-cortex, 2-thymus, 3-serum, 4-spleen and 5-liver) which were separated by reducing SDS-PAGE, blotted onto PVDF membrane and developed with a streptavidin- biotinylated alkaline phosphatase complex (panel A) or an anti-digoxin antibody labelled with alkaline phosphatase (panel B).

Fig. 2 shows the structure of digoxin.

Fig. 3 shows the activation of digoxin by CNBr and its subsequent reaction with the hydrazone derivative of a glycan.

Fig. 4 shows an example of a glycan digoxin- labelled glycan.

Fig. 5 shows the results of the determination of GS-I lectin activity using a digoxin-labelled glycan probe.

The following example describes a method to synthesise digoxin-labelled glycans by activating digoxin with cyanogen bromide (CNBr) and reacting the activated digoxin with the hydrazone derivative of glycans. The reaction sequence is summarised in Fig. 3.

The digoxin-labelled tetra-antennary glycan structure shown in Fig. 4 is an example of this invention. Several oligosaccharides have been labelled in this way. An application of the binding of a digoxin/glycan probe to lectin is demonstrated by the results in Fig. 5 using Griffonia simplicipholia lectin I (GS-I). This lectin binds to terminal structures of the type (Gala (1-4) Gal) as shown in Fig. 4.

Activation of digoxin using CNBr

Digoxin (Aldrich Chem. Co. , Milwaukee, WI) (100 gmol, 78.1 mg) was dissolved in 3 ml 33% tetrahydrofuran, 66% 2 M potassium phosphate buffer, pH 12, to form a biphasic mixture containing 33 mM digoxin. CNBr (Aldrich Chem. Co., Milwaukee, WI) (10-fold excess) was added to this mixture as a 5 M solution in tetrahydrofuran. The reaction mixture was stirred at 22°C for 30-60 minutes. Formation of the product was monitored by thin-layer chromatography (TLC) on 0.2 mm Silica Gel 60 F254 precoated on aluminum sheets (Merck, Darmstadt, Germany), which was developed in chloroform: methanol: water (80: 20: 10, v: v: v).

Digoxin was detected by"charring"after spraying with 15% H2SO4 in 80% ethanol. "Activated"digoxin appeared as the major product with Rf slightly lower than the digoxin. The estimated product yield was usually 40-60%. If CNBr was excluded from the reaction mixture, TLC analysis revealed no change in digoxin for 60 min, suggesting that the lactone ring was stable under the alkaline conditions.

The reaction mixture was evaporated under reduced pressure, and the dried powder was redissolved in 20 ml mixture of chloroform and 1 M NaCl (1: 1, v/v). After vigorous shaking, the phases were separated and the water phase was extracted with additional 10 ml chloroform. Virtually all the digoxin derivative was found in the combined chloroform phases. The combined chloroform phases were briefly washed with 5 ml water to remove any

residual water-soluble material, and dried under reduced pressure. Any remaining unreacted digoxin did not interfere with the reaction of the activated digoxin with the glycan. If the activated digoxin/unreacted digoxin was stored dry, the ratio of"activated-digoxin"to digoxin remained constant for several weeks at room temperature.

Reductive amination of glycans A pool of glycans (approximately 1 jlrnol) were dissolved in 10 Rl 100 mM triethylamine/CO2 buffer pH 8.5, containing 10% hydrazine, 10% pyridine borane and incubated for 24 hours at 42°C to reductively aminate the reducing end of the glycans. A large surplus of hydrazine was used to prevent the formation of dimers. The unreacted pyridine borane complex was removed by two extractions with 0.5 ml of diethyl ether. Free borate was removed by evaporating twice from 20 Fl 10 mM acetic acid; and hydrazine was removed by evaporating ten times from toluene. All evaporations were performed under reduced pressure at 37 °C.

Labelling of glycans with digoxin The glycan hydrazone derivatives were dissolved in 50 p. l 0.1 M triethylamine adjusted to pH 8.5 (pH lowered by adding solid carbon dioxide) and rapidly mixed with the CNBr-activated digoxin (5-to 10-

fold molar excess) dissolved in 50 pl tetrahydrofuran. After overnight incubation at room temperature, the reaction was stopped by evaporation under reduced pressure. The product was dissolved in 100 Rl water and all the non- reacted digoxin was removed by two extractions with 500 Zl chloroform.

Notes on the digoxin labelling procedure 1. It is highly preferable to convert all the glycan to the hydrazone derivative since any remaining free glycan would act as a competitive inhibitor for the lectin binding.

2. Any diamine or dihidrazide can be used instead of hydrazine, but hydrazine is particularly convenient since it can be nearly quantitatively removed by repeated evaporation from toluene.

3. The glycan/digoxin reaction can be performed in a number of suitable buffers (for example phosphate, carbonate, or triethylamine) plus tetrahydrofuran or acetonitrile.

4. The glycan/digoxin reaction should be allowed to proceed as close to completion as possible since any remaining free glycan will inhibit subsequent labelled glycan binding.

5. Activated digoxin decomposes quite quickly at elevated pH, thus repeated addition of"fresh" activated digoxin is the method of choice when

adding a large surplus of digoxin at the beginning of the glycan/digoxin reaction.

6. Unreacted or hydrolysed digoxin can be removed by extraction with toluene or chloroform.

Probing with a digoxin-labelled glycan A microtiter plate was coated with purified pigeon ovalbumin (10 Rg/ml in 50 mM Tris/HCl pH 7. 4, 10O mM NaCl ; TBS) for 4 hours at 37°C, followed by blocking with 3% (w/v) bovine serum albumin in TBS for 6 hours at 37°C and then three 10 minute washes with TBS. Different concentrations (0.1-10 Fg/ml) of GS-I in TBS were added to the wells and incubated for 1 hour at 37°C, followed by three 10 minute washes with TBS. Digoxin labelled glycans (0.2 nmol/well) were added to the wells and incubated for 4 hours at 37°C and the wells were washed with three 10 minute washes of TBS. Glycan binding was visualised by treating the washed wells with an anti-digoxin antibody conjugated with alkaline phosphatase (diluted 1: 10000) (Sigma) for 2 hours at 37°C and using p-nitrophenyl phosphate (3 mM) as a substrate. Absorbance was measured by a plate reader at 405 nm. The results are shown in Fig. 5; a linear response curve was obtained for the binding of the digoxin-probe to the lectin over a wide range of lectin concentrations.

Digoxin-labelled glycans can be used like antibodies in any multiwell, slot blotting or

histology assay; however, they have the advantage over antibodies that they can be used for activity measurements as well as detection. As well as investigating soluble lectins/carbohydrate binding proteins, digoxin labelled glycans can also be used for screening these molecules on the membranes/coats of animal cells, plant cells and micro-organisms.