GDNF and its receptors : a structure-function analysis of the NCAM-GDNF-GFRalpha1 complex and an examination of potential receptor candidates

Abstract

Neurotrophic factors are small, secreted proteins that are important players in the development and maintenance of the nervous system. Their functions span from neuronal survival and migration to neurite outgrowth and axon guidance, and in some cases even programmed cell death. Glial cell line-derived neurotrophic factor (GDNF) is the founding member of a family of four neurotrophic factors; the GDNF family of ligands. GDNF and its fellow family members Neurturin, Artemin and Persephin signal by forming a tripartite complex with one GPI-anchored receptor (GFRalpha1, -2, -3 or -4, respectively) and one transmembrane receptor, the latter being responsible for transmitting the signal through the plasma membrane into the cell. The transmembrane receptor can be either one of two molecules: The receptor tyrosine kinase Ret or the Neural Cell Adhesion Molecule (NCAM).

The interaction determinants of the Ret-GDNF-GFRalpha1 complex have been thoroughly analyzed, as well as the functional importance of its structural features. The more recently discovered NCAM-GDNF-GFRalpha1 complex, however, is much less studied. In paper I we mapped the GDNF binding domain in NCAM to the third immunoglobulin-like domain by analysing the ligand-binding properties of a large set of deletion mutants. Molecular modeling of the interaction interface revealed four charged contacts, and mutation of the NCAM residues involved in these contacts created a receptor unable to bind GDNF while keeping its cell adhesion properties, thus separating these two functions of NCAM.

In paper II we used a similar approach to map the NCAM-GFRalpha1 binding site. The major GFRalpha1-binding site in NCAM was mapped to the fourth Ig-like domain, and we found that the primary NCAM-binding site resides in the N-terminal domain of GFRalpha1. Deletion of the latter domain did not affect the ability of GFRalpha1 to potentiate GDNF-NCAM binding, implying that GFRalpha1 facilitates signaling by a ligand presentation mechanism rather than by allosterically changing the GDNF-binding properties of NCAM. We also showed that deletion of the NCAM-binding N-terminal domain severely impaired the ability of GFRalpha1 to interfere with NCAM-mediated cell adhesion. This suggests that the strong GFRalpha1-NCAM interactions mediated by the N-terminal domain of GFRalpha1 are required for efficient modulation of NCAMmediated cell adhesion.

Although the GDNF family of ligands appears to be a vertebrate-specific invention, orthologs of the receptors are present in other phyla. In paper III we examined the Drosophila Ret ortholog dRet, and showed that it did not bind GDNF, nor did it display any cell adhesion promoting properties despite its cadherin-like domain structural features. Interestingly, we found that a chimeric mammalian Ret molecule that included the dRet kinase domain was capable of inducing PC12 cell differentiation upon stimulation with GDNF and soluble GFRalpha1, suggesting that this functional feature of the intracellular domain is evolutionarily conserved.

GDNF and GFRalpha1 promote the migration and differentiation of cortical interneurons from the medial ganglionic eminence (MGE), a process in which neither RET nor NCAM appear to be involved, suggesting the possible existence of a third transmembrane receptor. In paper IV we examined two potential receptor candidates: The HGF receptor Met and the Neuregulin-1 receptor ErbB4. None of the two receptors were phosphorylated by GDNF in our biochemical assays, and neither HGF nor Neuregulin had any significant effects of their own on neurite outgrowth in our cultures. Intriguingly, two different Met inhibitors displayed outgrowthpromoting effects on their own. However, neither Met nor ErbB4 inhibitors impaired GDNFinduced neurite outgrowth, showing that these receptors are not involved in GDNF signaling in MGE neurons.