We constructed a three-dimensional model of TNFRSF25 (death receptor-3; DR3), a tumor necrosis-receptor family member that is expressed on immune cells and on osteoblasts, to determine whether mutations that are linked to rheumatoid arthritis are likely to have effects on receptor function. Since the crystal structure of DR3 is not known, comparative modeling was used, aligning structural elements of the primary sequences of DR3 with TNFs which have been determined by crystallography, substituting the amino acids of the target protein for those in the known structure, introducing necessary deletions or insertions, followed by energy minimization to yield a putative structure. This approach has been validated by studies of other TNF-family receptors. The results show that the DR3 extracellular domain is comprised of four homologous cysteine-rich domains (CRDs), and that a mutation linked to rheumatoid arthritis is in a region critical for structural integrity of ligand-receptor complexes at the end of CRD3. Specifically, the D158G mutation eliminates two hydrogen bonds normally present in a N/D-T-V/D-C consensus motif typically found flanking the last cysteine of each CRD. This may cause aberrations in either T cell function or in response of bone cells to DR3 ligands, which may contribute to pathology in rheumatoid arthritis. Comparison of RA mutants to mutants in other TNFRSF receptors shows that these occur in homologous positions in CRDs, so that this site is proposed to be a 'hot spot' for mutations in TNFRSF family proteins.