Crystallographic Evidence for Dimerization of Unliganded Tumor Necrosis Factor Receptor

Abstract

Activation of the cell surface receptors for tumor necrosis factor (TNF) is effected by the aggregation of cytoplasmic domains that occurs when the extracellular domains of two or three receptors bind to trimeric TNF alpha or TNF beta. The structure of the type I TNF receptor extracellular domain (sTNF-R1), crystallized in the absence of TNF, has now been determined at 2.25-A resolution. The receptor itself is an elongated molecule comprising four disulfide-rich domains in a nearly linear array. Contrary to expectations, the unliganded domains are found to associate into dimers of two distinct types, in which monomers are related by local two-fold axes of symmetry. In one case, the receptors are antiparallel to each other and associate through an interface that overlaps the TNF binding site. If intact receptors were capable of such an association, their cytoplasmic domains would be separated by over 100 A. This interaction could inhibit signaling in the absence of TNF. Parallel dimers are also observed in which the dimer interface is well separated from the TNF binding site. Associations among TNF-bound parallel dimers could cause receptor clustering. Both dimers bury substantial areas of protein surface and are formed by polar and non-polar interactions.

Homomeric and heteromeric interactions of the extracellular domains of death receptors and death decoy receptors

“…Previously, ligand-independent self-association of DR4-ECD was detected on the cell surface by fluorescence resonance energy transfer (FRET) [25]. In other TNFR superfamilies, the unliganded ECD of TNFR1 was found to crystallize as a dimeric complex [27], and pre-ligand homomeric clustering of TNFR1, TNFR2, CD40, and Fas (CD95) was observed on the cell surface by FRET studies [25,28]. However, no heterologous interactions were observed between TNFR1-TNFR2 and CD40-DR4 on the cell surface [25].…”

“…The homomeric interactions of TNFR1, TNFR2, and Fas occurred through the N-terminal CRD1 domain, which is distinct from the ligand contact regions of CRD2 and CRD3 [25,27,28]. The TNFRs and Fas have a complete CRD1 containing three disulfide bonds, while DRs and DcRs have only a partial CRD1 with one disulfide bond, questioning the feasibility of the interaction through the truncated CRD1.…”

The role of polar interactions in the molecular recognition of CD40L with its receptor CD40

“…The combination of A1 and B2 modules, together with variation in the spacing between them (0 to 2 residues), is used to describe the structures of domains 1 to 3, and the first half of domain 4, of TNFR. The structures of six monomers of TNFR have been determined in different space groups, at different pH, in unliganded forms and in complex with TNF-p (Banner et al, 1993;Naismith et al, 1995. An analysis of the six independent crystallographic structures of TNFR confirmed that the modules rather than the domains are structurally rigid .…”

“…These proteins share little primary sequence similarity in their extracellular domains, apart from the cysteine pattern. The X-ray structure of the TNFR was first determined as part of a complex with TNF-,B (Banner et al, 1993), and subsequently in the unliganded state at pH 7.5 (Naismith et al, 1995). The extracellular region of TNFR is composed of four domains of 44 amino acids.…”

“…An alternative approach has been to use the known X-ray structures of TNFR (Banner et al, 1993;Naismith et al, 1995) to build a homology model of CD40. The first model of CD40 was part of a complex with CD40L, which was built using the structure of the…”