Among the multitude of carbohydrate structures found in nature, the ones that seem most likely to bear specific information are the complex oligosaccharides attached to extracellular proteins and to proteins and lipids at the surface of eukaryotic cells. Although our understanding of the roleof the saccharide portionsof such glycoconjugates is still incomplete, results presented in this issue of Cell provide important insight into their functions. To appreciate the significance of these findings, it is useful to reflect on the unique properties of complex oligosaccharides that may make them especially suited for conveying information in particular situations. Multiple Roles ot Giycosyiation The idea that cell surface sugars form an important part of the identity of cells has been given a substantial boost by the recent demonstration that oligosaccharide recognition by the sefectins is an important first step in adhesion of leukocytes to selected portions of the endothelium (Stoolman, 1989). The cell surface oligosaccharides present a high concentration of target ligands, which facilitates the initial interaction between stationary and rapidly moving cells. High affinity of individual saccharide-receptor interactions is not so important as high valency. The initial sugar-mediated binding need only be sufficiently tight to allow time for more specific protein-protein adhesion systems to consolidate the interaction. Although the generality of this type of function for cell surface sugars remains to be demonstrated, we at least have a provocative paradigm. In contrast to the plasma membrane, each serum glycoprotein carries one or at most a few oligosaccharides. The sugar units on these glycoproteins must, therefore, function in a more individualistic manner. One possibility is that the conjugation of saccharides to proteins allows proteins to achieve unique conformations that facilitate their intrinsic activities, such as catalysis, ligand binding, or receptor binding. Indeed, the ability of specific saccharide structures to modify the intrinsic activity of individual serum proteins has now been well established (Parekh, 1991). However, given the subtle variety of function that can be achieved with polypeptidesalone, it is hard to imagine how sugars could add sufficiently to the repertoire to be worth all the trouble it takes to put them on. The vast array of intracellular protein functions achieved without attachment of N-linked oligosaccharides bears this out. While recognizing that attachment of sugars to proteins affects their intrinsic activities, one cannot help but feel that the sugars must be put there to serve some general purpose. The tangential effects on enzymatic or other activities have been exploited to various degrees in specific instances, but what is the general purpose? Recent results Minireview