Studies of the functions of carbohydrates — which are increasingly being implicated in diseases such as cancer — have traditionally lagged behind those for nucleic acids and proteins, primarily owing to the lack of general methods for synthesizing and analysing these structurally complex molecules. Recent advances in methods for synthesizing complex carbohydrates have had a significant impact, and now, with the first report of a chemically defined carbohydrate array, the exciting possibility of chips that present arrays of many different carbohydrates for functional studies seems considerably closer to becoming reality.

Development of a carbohydrate chip for characterizing protein–carbohydrate and carbohydrate–carbohydrate interactions and enzyme activities poses several key challenges. To provide both excellent selectivity and quantitative analysis, the chip must prevent non-specific interactions, and the immobilized carbohydrates must be presented in a regular and homogeneous environment to ensure that they have uniform activity towards soluble proteins and enzymes. Furthermore, because many protein–carbohydrate and carbohydrate–carbohydrate interactions are polyvalent in nature, the density of carbohydrate ligands must be controlled. Finally, to be broadly applicable, the chip should be compatible with several common detection methods, such as fluorescence imaging and surface plasmon resonance (SPR) spectroscopy.

Houseman and Mrksich used the Diels–Alder reaction to link carbohydrate ligands to a gold-based monolayer, prepared from a controlled mixture of two components — one that can be linked to the carbohydrate ligands (1%), and one that prevents non-specific association of proteins (99%). Because the Diels–Alder reaction is rapid, selective and quantitative, it can be used to ensure that the carbohydrates are presented homogeneously at a uniform density.

To test the characteristics of their monolayers, the authors first used SPR, a powerful technique for monitoring biomolecular associations. A monolayer presenting a single carbohydrate ligand was completely inert to non-specific adsorption of fibrinogen, a common “sticky” protein, and the ligand could specifically associate with the lectin concanavalin A, a carbohydrate-binding protein. Next, they prepared an array that presented ten carbohydrate ligands, and showed — this time using fluorescence imaging — that it could be used to selectively identify binding interactions between the ligands and several different lectins. Further experiments showed that such interactions could be analysed quantitatively. Finally, the authors showed that the activity of a carbohydrate-modifying enzyme towards its natural substrate bound to an array could be characterized.

So, such arrays fulfil the key requirements for the development of a carbohydrate chip, and importantly, are compatible with recently reported methods for automatic synthesis of complex carbohydrates, suggesting that chips constructed by this chemical approach could find broad application in glycobiology, from research to drug discovery and diagnostics. Moreover, the surface chemistry involved could also prove valuable in the preparation of peptide and protein chips.