Lys 10

Fig. B.5.1. Crystal structure of a-l-arabinose bound to the binding site of the periplasmic ABP of Escherichia coli (PDB entry 1ABE [5]). Dashed lines represent direct and indirect hydrogen bonds.

Gln 11

Thr 208

h2o whether the protein framework provides binding sites in deep clefts, as in most enzymes and bacterial periplasmic carbohydrate binding proteins, or in relatively shallow depressions in the protein surface, as is typical for lectins [1-4], the molecular basis of carbohydrate binding involves a complex array of non-covalent interactions such as hydrogen bonds, hydrophobic interactions, and, occasionally, an essential contribution of divalent cations to protein structure fixing and/or direct coordination to the substrate.

Hydrogen Bonding

As shown in Figure B.5.1, the protein usually exploits co-operative hydrogen bonding in which a single sugar hydroxyl group acts simultaneously as a hydrogen-bond donor and acceptor to distinguish between different carbohydrate epitopes.

One acidic amino acid side-chain is usually used as a hydrogen-bond acceptor for one or two sugar OH groups. The most common hydrogen bonding scheme is:

Particularly effective in this context are pairs of vicinal sugar OH groups or one OH and the ring oxygen atom interacting with two functional groups in a single amino acid side-chain or with consecutive main-chain amide groups, especially when the vicinal OH groups have an either equatorial/ equatorial or equatorial/axial configuration. Examples include the interaction of Asn 232 with OH-3 and OH-4 or Arg 151 with OH-4 and the ring oxygen atom in the complex of a-L-arabinose with the periplasmic ABP of Escherichia coli [5] (Figure B.5.1). Also water was observed to mediate (indirect) hydrogen bonds between amino acid residues and the saccharide OH groups. In these circumstances the water molecules act as fixed structural elements, equivalent to hydrogen-bonding groups of the protein, and can therefore be regarded as a part of the binding-site architecture.

Non-polar Interactions

Although carbohydrates have many polar functional groups, most biologically relevant saccharides still have significant non-polar patches on their surface that can interact with complementary hydrophobic amino acid residues. This is particularly true for ¿S-D-galactose or related monosaccharide units which have a continuous non-polar surface from C-3 to C-6 and thus this face of the sugar is almost always observed to pack against aromatic side-chains from phenylalanine, histidine, or, most commonly, tryptophan. Hydrophobic interactions are believed to provide a significant contribution to the overall binding energy because apolar patches on the sugar and the aliphatic and aromatic side-chains of the protein are removed from the bulk solvent thus expelling unfavorably bound water molecules.

Divalent Cations

Mammalian C-type lectins are unique among structurally characterized lectins in that most require a calcium(II) ion to form direct coordinative bonds with the sugar ligands which have been found to be the primary determinants of affinity. Figure B.5.2 shows such an example [6] in which the full non-covalent binding potential of two vicinal OH groups is used. One lone pair of electrons from each OH group forms a coordinative bond with Ca2+, the other lone pair accepts a hydrogen bond from a side-chain amino group, and the proton is donated to an acidic oxygen in a hydrogen bond. Calcium and other ions, for example manganese(II) ions, are, however, required for the activity of many families of carbohydrate-binding proteins, even if they do not interact directly with the ligands. In these instances the most common function of these ions is structural, in that the metal ion coordination shell orients important protein functional groups for optimum ligand binding.


The binding affinity of lectins is still surprisingly low - dissociation constants are usually approximately millimolar. Multivalency, i.e. the simulta-

Fig. B.5.2. Crystal structure of an a-d-mannose unit bound to the C-type lectin rat mannose-binding protein A (MBP-A, PDB entry 2MSB [6]). Dashed lines indicate hydrogen and co-ordinative bonds.

neous association of several ligands of one biological unit (macromolecule, cell surface,...) with several receptors of another biological unit, provides a means of achieving high-affinity interactions applied by nature [7, 8]. Extension of carbohydrate binding sites to include direct or water-mediated contacts with multiple sugar units in oligosaccharide ligands ("subsite multivalence'' [9]), however, results only in an increase in affinity to the micromolar range at best. High-affinity carbohydrate-protein interactions in the nanomolar range can be achieved by oligomerization of several lec-tin polypeptides each containing similar or identical simple binding sites (''subunit multivalence'' [9]) or by clustering of several lectins on cell surfaces and interaction of these architectures with multiple carbohydrate epit-opes presented in an appropriate manner on lipid or protein carriers [10].

Clustered binding sites are important because the free energy of binding of a multivalent ligand to multiple sites of an oligomeric lectin can be as large (or even larger) as the sum of the free energies of the individual binding interactions. This means that ideally the individual dissociation constants could potentially be multiplied to achieve high affinity for the complex ligand without the need to involve any additional molecular interactions beyond those seen with individual sugars in the single binding sites. Even if the requirements for ideal additivity are not often met in reality, because of additional geometric constraints, the affinity enhancement is still substantial. The strategy of employing structurally defined multivalent interactions is also very efficient in achieving maximum recognition of certain carbohydrate-coated surfaces - for example cell surfaces - while minimizing competitive binding to smaller saccharides or sugars attached to soluble glycoproteins at the same time. Structural studies revealed that many lectins that bind surfaces achieve the required planar array of sugar-binding sites by arrangement of polypeptide subunits in oligomers with cyclic symmetry, with all binding-sites located on one end of the oligomer.


1 S. H. Barondes, Trends Biochem. Sci. 1988, 13, 480-482.

1997, 243, 543-576; H.-J. Gabius, ed., Animal Lectins, Biochim. Biophys. Acta 2002, 1572, 163-434.

4 D. C. Kilpatrick, Handbook of Animal Lectins: Properties and Biomedical Applications, Wiley, Chichester, 2000.

5 F. A. Quiocho, N. K. Vyas, Nature 1984, 310, 381-386.

6 W. I. Weis, K. Drickamer, W. A. Hendrickson, Nature 1992, 360, 127-134.

7 L. L. Kiessling, T. Young, K. H. Mortell, in B. Fraser-Reid, K.

Tatsuta, J. Tmem, eds., Glycoscience: Chemistry and Chemical Biology, Vol. II, Springer, Heidelberg, 2001, pp. 18171861.

8 M. Mammen, S.-K. Choi, G. M. Whitesides, Angew. Chem. 1998, 110, 2908-2953; Angew. Chem. Int. Ed. 1998, 37, 2754-2794; J. J. Lundquist, E. J. Toone, Chem. Rev. 2002, 102, 555-578; C. F. Brewer, M. C. Miceli, L. G. Baum, Curr. Opin. Struct. Biol. 2002, 12, 616-623.

9 J. M. Rini, Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 551-577.

10 P. R. Crocker, T. Feizi, Curr. Opin. Struct. Biol. 1996, 6, 679691.

0 0

Post a comment