Conformation of the glycoprotein glycans of the N-acetyl-lactosaminic type (complex type).

hibit flexibility with respect to the rigid core. Surprisingly, similar studies on the a ( 1-3) arm of ‘bisected’ sugars from rabbit IgG (see below) and sialylated oligosaccharides isolated from bovine prothrombin (D. L. Fernandes, unpublished work) which contain a GalB(1-3)GlcNAc linkage in the outer chains show similar NOE effects when compared to the ‘complex’ oligosaccharide from human serotransferrin or non-bisected oligosaccharide obtained from rabbit IgG. This suggests that in biantennary ‘complex’ structures the a ( 1-3) antenna as well as the mannosyl-chitobiosyl core may exist in a well-defined conformation. Further n.m.r. studies indicate that the a(l-6) arm of the un-bisected carbohydrate enjoys considerably greater flexibility than that of the bisected structure. Although no NOE’s have been detected between the a(1-6) arms and the mannosylchitobiosyl cores of the carbohydrates examined a selective line broadening of the anomeric resonance of the Man(4‘) residue occurs in the presence of a bisected GlcNAc (S. W. Homans, unpublished work). Our preliminary interpretation of this broadening is that either the Man(4’) anomeric proton is in chemical exchange or is experiencing an alteration in the dipolar interaction as a result of an increased number of interacting protons; the most likely sources of such perturbation would be protons of the mannosyl-chitobiosyl core and of the bisecting GlcNAc. The structural significance of this is presently under investigation. The rabbit species is unique in that IgG sub-classes have not been detected. Examination of the Fc-derived oligosaccharide chains should therefore reflect structural requirements for CH2-C,2 cross-domain architecture. It was therefore surprising to find the presence of four mannosyl-chitobiosyl cores on rabbit IgG and extensive outer-chain heterogeneity. In addition, despite the degree of carbohydrate heterogeneity, single crystals of rabbit Fc have carbohydrate profiles identical with those of pooled Fc fragments. This ‘lack’ of selective crystallization of immunoglobulin molecules having identical oligosaccharide chains suggests that the specific oligosaccharide sequence has little role in maintaining a unique protein conformation. Immunoglobulins derived from species (human, rat, mouse and bovine) which have multiple sub-classes also have outer-chain and core variations. Experiments are currently in progress to determine if any of these structures are sub-class restricted and whether or not rabbit utilizes carbohydrate diversity to generate functionally-distinct sub-classes. Recent X-ray crystallographic data on human Fc fragments suggest that the a( l -6) arms are involved in carbohydrateprotein interactions while a(1-3) arms are involved in carbohydrate-carbohydrate interactions (Deisenhofer, 198 1). Our NOE data expand upon this result and suggest that the mannosyl-chitobiosyl core and the a(1-3) arm exist in a well-defined continuous unit. In our limited series of a oligosaccharide structures examined the relative orientation of the a(1-3) arm is invariant. Conversely, the a(l-6) arm varies from relative freedom of internal interactions to semi-rigid structures. The a( l -6) arm may therefore play a crucial role in maintaining unique structural variants which serve as recognition units in glycoproteins. Finally, the presence of core variants and outer-chain microheterogeneity in rabbic Fc-derived oligosaccharides would appear to preclude any form of specific interaction of the sugar on one C,2 domain with either the contiguous protein or the adjacent carbohydrate on the opposite C,2 domain. However, based upon our analysis of the molar ratios of oligosaccharide species present on Fc, n.m.r. of oligosaccharides in free solution, model building and available X-ray crystallographic data on both rabbit and human Fc carbohydrates, it is our hypothesis that the Fc unit has dissimilar oligosaccharides specifically paired such that the structural requirements of the proteincarbohydrate and carbohydrate-carbohydrate interaction can be maintained. Experiments are currently in progress to determine the validity of this hypothesis.