Structure-function studies of GDNF and other members of the TGF-beta superfamily

Abstract: The glial cell line-derived neurotrophic factor (GDNF) family is a distant subclass of the TGF-beta superfamily. The GDNF family of ligands, consisting of GDNF, neurturin (NTN), artemin (ART) and persephin (PSP), are potent survival and differentiation factors for neurons in the central and peripheral nervous systems. These ligands signal through a tyrosine kinase receptor, Ret and an accessory receptor subunit, the GDNF family receptor alphas (GFRas). There are four GFRas known to date (GFRalpha1 to 4). GDNF preferentially binds to GFRalpha1, while NTN, ART, and PSP bind to GFRalpha2, 3 and 4, respectively, but with some degree of cross talk. It has been postulated that the ligand first associates with the GFRalpha and that only then is Ret recruited to the complex, becoming autophosphorylated on several cytoplasmic tyrosine residues. Phosphotyrosine residues on active Ret form a platform for the recruitment of multiple adaptor and effector proteins. There are two main isoforms of Ret, Ret9 and Ret51, differing at their C-terminal sequence. We have assessed the ability of different GDNF mutants to bind to GFRalpha1 and induce tyrosine phosphorylation in Ret. Hydrophobic and negatively charged residues in the tips of GDNF fingers 1 and 2, were found to be important for receptor interaction. Unexpectedly some of the mutants that lost their affinity for GFRalpha1 were still able to induce Ret tyrosine phosphorylation. These mutants however, were not able to activate Ret in cells not expressing GFRalpha1, indicating that GFRalpha1 was still required even if GDNF was unable to bind. These results led us to propose a model including two distinct binding sites for GDNF: one formed by GFRalpha1 alone, requires hydrophobic and acidic residues in finger 1 and 2, and another by a pre-associated GFRalpha1/Ret complex, which requires acidic residues in finger 1. Several mutations have been found in the GDNF gene of patients with Hirschsprung disease (HSCR). We have characterized the effects of these mutants on the ability of GDNF to bind to and activate its receptors. Although none of the four mutations analyzed appeared to affect the ability of GDNF to activate Ret, two of them resulted in a significant reduction in the binding affinity of GDNF for GFRalpha1. Indicating that, although none of the GDNF mutations identified so far in HSCR patients are per se likely to result in HSCR, two of these mutations may, in conjunction with other genetic lesions, contribute to the pathogenesis of this disease. We have demonstrated that Ret51 associates more strongly, than Ret9, with the ubiquitin ligase Cbl, leading to increased ubiquitylation and faster turnover of active Ret51. The association of Cbl with Ret is indirect and mediated through Grb2. A constitutive complex of Grb2 and Cbl can be recruited to both receptor isoforms via docking and tyrosine phosphorylation of Shc. However, Ret 51, but not Ret9, can in addition recruit Cbl via direct interaction of the Grb2/Cbl complex with phosphorylated Tyr-1096, unique to the Ret51. Interestingly, this same phosphotyrosine also allows Ret51 to recruit the adaptor protein CrkL, leading to prolonged activation of MAP kinases ERK1 and ERK2 upon activation of Ret51 in neuronal cells. Our results have therfore established distinct signaling mechanisms by Ret51 and Ret9. Taking advantage of the conserved pattern of cysteine residues in the TGF-beta superfamily, we sought to identify new members of this family, using a novel search engine called Motifer. We identified two genes, provisionally named Motifer Derived Factors (MDF451 and MDF628), in the public human genome database. The MDFs can only be found in genomes of primates and not in other species such as rodents. Both genes are expressed in human fetal brain and cerebellum, but so far we have been unable to isolate full-length cDNA for either of the two MDFs. Further analysis of upstream sequence in MDF628 revealed STOP codons in frame with the TGF-beta like reading frame, thereby ruling out the capacity of this gene to encode a TGF-beta family member. We are suggesting that the genes identified may have recently appeared in evolution, in a common ancestor of the primate lineage, perhaps by duplication of a GDNF/TGF-beta-like gene.