**3. Glycan binding assays for galectins**

As galectins play a fundamental role in cell adhesion, cell signalling, inflammation, tumor progression etc. there is an enormous interest in the evaluation of galectin-glycan interactions regulating those functions.

### **3.1 Comparison of different common assays**

Various assay set-ups have been designed to analyse the binding behaviour of different galectins to specific glycan structures. Binding assays can be subdivided regarding the presentation of the different binding partners: 1) the glycan structure is immobilised, 2) the galectin is immobilised and 3) both binding partners are soluble.

The chosen assay format influences the data generated as each assay set-up has its own advantages and disadvantages (Rapoport et al., 2008):

Assays in which one of the binding partners is immobilised raise the problem that the amount of this ligand is not completely known. Moreover it is possible that side interactions with the surface occur or that the conformation and flexibility of the bound partner differ slightly from its soluble parameters. The natural oligomerisation of galectins is blocked after immobilisation. Beside this the presentation of the immobilised binding partner is multivalent which influences the binding (Sörme et al., 2004). This can be useful for the glycan structures, as they are multivalently presented in nature as well, but not for galectins. Examples for studies with immobilised glycans or glycoproteins are glycan arrays, ELISA

Guévremont et al., 2004; Ochieng et al., 1994). The single CRD is mainly described to have an increased affinity for different carbohydrates such as *N*-acetyllactosamine, the glycoprotein asialofetuin or glycans presented on endothelial cells but to have less biological activity as it looses the ability to form oligomers. This reveals the possible regulatory function of galectin-3 cleavage (Dam et al., 2005; Dumic et al., 2006; Ochieng et al., 1998a; Shekhar et al., 2004). In terms of this regulation it is suggested that the single galectin-3-CRD binds with high affinity to glycans on cell surfaces thereby blocking these interaction partners for full-length galectin-3 binding. After this blockage the full-length protein cannot perform its physiological functions anymore. In this way galectin-3 cleavage could act as

The specific properties of galectin-8 are also implied in its structure and the different isoforms arising from it. At least 6 different isoforms are described so far of which some only consist of the N-terminal CRD with an extension and others consist of both CRDs linked by different hinge domains (Bidon et al., 2001; Delgado et al., 2011; Zick et al., 2004). The two galectin-8 CRDs show approximately 35% sequence similarity but reveal different fine specificity for glycan structures. Therefore galectin-8 can act as "hetero-bifunctional crosslinking agent" (Zick et al., 2004). The length and structure of the linker domain has direct influence on the biological function (Levy et al., 2006). Moreover the linker domain regulates susceptibility to protease cleavage. It was for example shown that a long linker can be cleaved by thrombin while shorter linker variants are not substrate for this protease

As galectins play a fundamental role in cell adhesion, cell signalling, inflammation, tumor progression etc. there is an enormous interest in the evaluation of galectin-glycan

Various assay set-ups have been designed to analyse the binding behaviour of different galectins to specific glycan structures. Binding assays can be subdivided regarding the presentation of the different binding partners: 1) the glycan structure is immobilised, 2) the

The chosen assay format influences the data generated as each assay set-up has its own

Assays in which one of the binding partners is immobilised raise the problem that the amount of this ligand is not completely known. Moreover it is possible that side interactions with the surface occur or that the conformation and flexibility of the bound partner differ slightly from its soluble parameters. The natural oligomerisation of galectins is blocked after immobilisation. Beside this the presentation of the immobilised binding partner is multivalent which influences the binding (Sörme et al., 2004). This can be useful for the glycan structures, as they are multivalently presented in nature as well, but not for galectins. Examples for studies with immobilised glycans or glycoproteins are glycan arrays, ELISA

down-regulation of galectin-3 function (John et al., 2003; Shekhar et al., 2004).

**2.3.4 Galectin-8: Several isoforms of a tandem-repeat galectin** 

(Nishi et al., 2006).

**3. Glycan binding assays for galectins** 

**3.1 Comparison of different common assays** 

advantages and disadvantages (Rapoport et al., 2008):

galectin is immobilised and 3) both binding partners are soluble.

interactions regulating those functions.

assays and surface plasmon resonance (Appukuttan, 2002; Blixt et al., 2004; Bohorov et al., 2006; Ideo et al., 2003; Munoz et al., 2010; Song et al., 2009b; Stowell et al., 2008a). For glycan arrays it can be important which linker is used to bind the glycan epitopes to the surface (length, chemical structure). Moreover it is possible to use chemically and/or enzymatically produced ligands as well as glycans from natural compounds like glycopeptides and glycolipids (Blixt et al., 2004; Bohorov et al., 2006). The latter allows the analysis of complex and even unknown glycan structures of different cells (Song et al., 2009a; Song et al., 2009b; Song et al., 2010). Immobilised galectins are for example used in frontal affinity approaches and ELISA assays (Hirabayashi et al., 2002; Sörme et al., 2002).

Variations of binding assays with immobilised partners are assays in which the surface binding is inhibited by a soluble ligand. Such inhibition studies of surface interactions allow a direct read-out of IC50 values and thereby the direct comparison of relative affinities (Sörme et al., 2002). For the calculation of affinity constants assumptions have to be made to simplify calculations which might not be correct for each single interaction measurement. Additional the problems mentioned before still persist (Sörme et al., 2004).

Most assays with one immobilised component as well as some direct interaction assays are based on the read-out of a fluorescence signal or other labels. Therefore either the galectin or the glycan structures have to be labelled. This leads to some additional problems: If the glycan is chemically labelled the linker or label itself can alter the binding affinity with specific effects for different galectins (Sörme et al., 2004). Therefore the affinity constants measured do not exactly fit to the unmodified glycan structures. Moreover the labelling of glycans is time-consuming. The labelling of galectins can also alter the binding specificities. It is in most cases done by random chemical modification of specific functional groups such as amino or thiol functionalities (Carlsson et al., 2007; Patnaik et al., 2006; Rapoport et al., 2008; Salomonsson et al., 2010; Song et al., 2009b; Stowell et al., 2008a; Stowell et al., 2008b). Although this labelling is assumed not to influence binding specificity or inactive galectins are removed after the labelling reaction, binding and oligomerisation still might be slightly affected. Moreover lot-specific aberrations between different labelling reactions occur. Labelled galectins are for example used for glycan arrays and ELISA-type set-ups (Blixt et al., 2004; Carlsson et al., 2007; Rapoport et al., 2008; Salomonsson et al., 2010; Song et al., 2009b; Stowell et al., 2008a) while fluorescence labelled glycans are used in frontal affinity chromatography or fluorescence polarisation (Carlsson et al., 2007; Hirabayashi et al., 2002; Salomonsson et al., 2010; Sörme et al., 2004).

Assays using both binding partners in its soluble form overcome most of the mentioned problems. But although those assays have different advantages the results cannot directly be compared with the natural set-up in which the glycans are immobilised on glycoproteins or glycolipids and thereby multivalently presented. Fluorescence polarisation is one of these methods measuring direct interactions of ligands in solution, but facing negative side effects of glycan labelling. Similarly, titration calorimetry also measures the interaction of two soluble binding partners. As for titration calorimetry no labelling reaction has to be performed this assay set-up might be considered as the one with fewest problems. But needed galectin concentrations for this test are usually (but not always) in high micromolar ranges and therefore above the physiological range. In this concentration range galectins tend to oligomerise, aggregate or precipitate (Ahmad et al., 2004b; Bachhawat-Sikder et al., 2001; Cho & Cummings, 1995; Dam et al., 2005; Sörme et al., 2004). Moreover titration

Galectins: Structures, Binding Properties and Function in Cell Adhesion 11

*N*-glycans, preferred poly-LacNAc

Non-reducing terminal and internal LacNAc, high affinity for repetitive LacNAcunits (grey in upper scheme)

3-O-sulfation 3 O-sulfation 3 O-sulfation and

β4

β4

SO4

Blood group A and B antigens

α2

α3 β4

α

α3 β4

Table 1. Preferred ligands of the single carbohydrate recognition domains of galectin-1, -3 and -8 following Rapoport et al. 2002. Symbols according to the consortium of functional glycomics (Brewer, 2004; Carlsson et al., 2007; Dell, 2002; Hirabayashi et al., 2002; Ideo et al., 2003; Ideo et al., 2011; Leppänen et al., 2005; Patnaik et al., 2006; Rabinovich & Toscano, 2009; Rapoport et al., 2008; Salomonsson et al., 2010; Stowell et al., 2004; Stowell et al., 2008a;

Complex *N*-glycans Increasing affinity with increasing number of antennas

Non-reducing terminal LacNAc type I or II (grey in upper scheme)

β3/4

β3/4

SO4

Yamamoto et al., 2008)

Galectin-1 Galectin-3 Galectin-8 N-CRD Galectin-8 C-CRD

*N*-glycans and glycosphingolipids (e.g. GM3 and GD1a)

Preference for lactose (but also binding to LacNAcunits)

β4

Cer

α3-sialylation

β4

α3 β4

SO4

*N*-glycans, preferred poly-LacNAc

LacNAc type I or II high affinity for repetitive LacNAcunits (grey in upper scheme)

β3/4

Blood group A and B antigens

α3 β4

α2

α3 β4

α2

calorimetry experiments are suitable for comparative studies of different glycans but do not lead to accurate calculation of affinity constants (Ahmad et al., 2004b). Another way to determine the direct interaction of soluble galectins and glycans is the use of hemagglutination assays, but those are limited to multivalent glycans or the inhibition of interactions between galectins and multivalent glycans or erythrocytes (Ahmad et al., 2004a; Ahmad et al., 2004b; Appukuttan, 2002; Giguere et al., 2008).
