Tumor necrosis factor receptor superfamily.

Tumor necrosis factor (TNF) is a powerful cytokine which is involved in the immune and pro-inflammatory response. The TNF receptors (TNF-R1 and TNF-R2) are the sole mediators of TNF signaling. The receptors consist of a disulfide rich domain which recognizes TNF, a transmembrane helix, and a cytoplasmic domain. Signaling occurs when a TNF trimer binds two or three receptors in an extracellular complex which permits aggregation and activation of the cytoplasmic domains. The complex is then endocytosed where it dissociates at low pH. We have now determined the structure of the soluble extracellular domain of TNF-R1 in two crystal forms at pH 3.7 in addition to our earlier report of one form at pH 7.5. One low pH form diffracts to 1.85 A and the entire polypeptide sequence has now been traced for this protein. The C-terminal 20 residues of the protein which were disordered in all previous structures show a different topology and disulfide connectivity to that seen in the remainder of the structure. In all crystal forms, the uncomplexed soluble extracellular domain of the type I TNF-R (sTNF-R1) exists as a dimer. At low pH the dimer buries a large amount of solvent accessible surface (2,900 A2), over 800 A2 greater than the area buried by TNF complexation. This dimer at low pH is different than both dimers observed in our previous pH 7.5 structure of unliganded sTNF-R1. We suggest that the low pH dimer forms during endocytosis and as the dimer completely buries the TNF interaction surface, the dimer would break up the receptor TNF complex. We have identified two distinct structural modules in sTNF-R1, a type A and a type B module. We suggest that these modules are the unit of structural conservation rather than the 6 cysteine subdomain. Although the orientation of these modules with respect to each other is sensitive to crystal packing, complexation, and pH, the modules themselves are structurally well conserved between and within the known sTNF-R1 structures. This modular approach will allow us to build accurate models for all members of the TNF-R superfamily.