Glycosaminoglycan-dependent and -independent inhibition of neurite outgrowth by agrin.

Agrin is a proteoglycan that can inhibit neurite outgrowth from multiple neuronal types when present as a substrate. Agrin's neurite inhibitory activity is confined to the N-terminal segment of the protein (agrin N150), which contains heparan sulfate (HS) and chondroitin sulfate (CS) side chains. We have examined the activities of various purified recombinant agrin fragments and their glycosaminoglycan (GAG) side chains in neurite outgrowth inhibition. Inhibitory activity was tested using dissociated chick ciliary ganglion neurons or dorsal root ganglion explants growing on laminin or N-cadherin. Initial experiments demonstrated that agrin N150 lacking GAG chains inhibited neurite outgrowth. Both halves of N150, each containing HS and/or CS side chains, could also inhibit neurite growth. Experiments using agrin fragments in which the GAG acceptor residues were mutated, or using agrin fragments purified from cells deficient in GAG synthesis, demonstrated that inhibition by the N-terminal portion of N150 requires GAGs, but that inhibition from the C-terminal part of N150 does not. Thus, the core protein or other types of glycosylation are important for inhibition from the more C-terminal region. Our results suggest that there are two distinct mechanisms for neurite outgrowth inhibition by agrin, one that is GAG-dependent and one that is GAG-independent.

A neuronal inhibitory domain in the N-terminal half of agrin.

Agrin is required for appropriate pre- and postsynaptic differentiation of neuromuscular junctions. While agrin's ability to orchestrate postsynaptic differentiation is well documented, more recent experiments have suggested that agrin is also a "stop signal" for the presynaptic neuron, and that agrin has actions on neurons in the CNS. To elucidate the neuronal activities of agrin and to define the receptor(s) responsible for these functions, we have examined adhesions of neurons and their neurite-outgrowth responses to purified agrin in vitro. We find that both full-length agrin and the C-terminal 95 kDa of agrin (agrin c95), which is sufficient to induce postsynaptic differentiation, are adhesive for chick ciliary ganglion (CG) and forebrain neurons. Consistent with previous findings, our results show that N-CAM binds to full-length agrin, and suggest that alpha-dystroglycan is a neuronal receptor for agrin c95. In neurite outgrowth assays, full-length agrin inhibited both laminin- and N-cadherin-induced neurite growth from CG neurons. The N-terminal 150 kDa fragment of agrin, but not agrin c95, inhibited neurite outgrowth, indicating that domains in the N-terminal portion of agrin are sufficient for this function. Adhesion assays using protein-coated beads and agrin-expressing cells revealed differential interactions of agrin with members of the immunoglobulin superfamily of cell adhesion molecules. However, none of these, including N-CAM, appeared to be critical for neuronal adhesion. In summary, our results suggest that the N-terminal half of agrin is involved in agrin's ability to inhibit neurite outgrowth. Our results further suggest that neither alpha-dystroglycan nor N-CAM, two known binding proteins for agrin, mediate this effect.