Iron is an essential nutrient for all organisms and consequently, the ability to bind transferrin and sequester iron from his source constitutes a distinct advantage to a blood-borne bacterial pathogen. Levels of free iron are strictly limited in human serum, largely through the action of the iron-binding protein transferrin. The acquisition of trasferrin-iron is coincident with pathogenicity among Neisseria species and a limited number of other pathogens of human and veterinary signi"cance. In Neisseria meningitidis, transferrin binding relies on two co-expressed, outer membrane proteins distinct in aspects of both structure and function. These proteins are independently and simultaneously capable of binding human transferrin and both are required for the optimal uptake of iron from this source. It has been established that transferrin-binding proteins (designated TbpA and TbpB) form a discrete, speci"c complex which may be composed of a transmembrane species (composed of the TbpA dimer) associated with a single surface-exposed lipoprotein (TbpB). This more exposed protein is capable of selectively binding iron-saturated transferrin and the receptor complex has ligand-binding properties which are distinct from either of its components. Previous in vivo analyses of N. gonorrhoeae, which utilizes a closely related transferrin-iron uptake system, indicated that this receptor exists in several conformations in#uenced in part by the presence (or absence) of transferrin.
Here we propose a dynamic model of the meningococcal transferrin receptor which is fully consistent with the current data concerning this subject. We suggest that TbpB serves as the initial binding site for iron-saturated transferrin and brings this ligand close to the associated transmembrane dimer, enabling additional binding events and orientating transferrin over the dual TbpA pores. The antagonistic association of these receptor proteins with a single ligand molecule may also induce conformational change in transferrin, thereby favouring the release of iron. As, in vivo, transferrin may have iron in one or both lobes, this dynamic molecular arrangement would enable iron uptake from either iron-binding site. In addition, the predicted molecular dimensions of the putative TbpA dimer and hTf are fully consistent with these proposals. Given the diverse data used in the formulation of this model and the consistent characteristics of transferrin binding among several signi"cant Gram-negative pathogens, we speculate that such receptor}ligand interactions may be, at least in part, conserved between species. Consequently, this model may be applicable to bacteria other than N. meningitidis.