Agrin is a synapse-organizing protein likely to mediate nerve-induced aggregation of acetylcholine receptors and other postsynaptic components at the neuromuscular junction. We used in situ hybridization and polymerase chain reaction (PCR) to define the localization of agrin mRNA and its alternatively spliced forms in the chick embryo nervous system. Agrin cRNA probes intensely labeled motor neurons, dorsal root ganglia, cerebellar Purkinje neurons, and retinal ganglion cells. Neuronal layers in optic tectum and ventricular regions were also labeled. Analysis by PCR showed that all parts of the nervous system at embryonic day 10 contained three major forms of agrin mRNA. Our results raise the possibility that agrin isoforms play a role in synapse formation or other aspects of neuronal development in the central nervous system.

Antibodies against chicken and Torpedo agrin were used for immunofluorescent staining in order to assess the spatial distribution and temporal appearance of agrin-like molecules at newly formed synaptic contacts in cultures of embryonic Xenopus nerve and muscle cells. The antibodies stained Xenopus neuromuscular junctions and removed ACh receptor (AChR)-aggregating activity from extracts of Xenopus brain. Immunofluorescence was observed at almost all nerve- induced AChR aggregates, even at microaggregates in cocultures as young as 7.5 hr and at nerve-muscle contacts less than 2 hr old. Microdeposits of immunofluorescence extended as far distally as, or farther than, the microaggregates of AChRs along young nerve-muscle contacts. They also occurred along portions of growing neurites that were not in contact with muscle. By contrast, immunofluorescence was rarely observed at the nonsynaptic aggregates of AChRs that form on noninnervated muscle cells. These results raise the possibility that neuronally derived microaggregates of agrin-like molecules may be primary sites of nerve-induced clustering of AChRs, and they indicate that these molecules are present at embryonic nerve-muscle synapses from the very onset of AChR aggregation. The cellular origin of the agrin-like molecules at synapses was examined in cross-species cocultures in which the neurons and muscle cells were obtained from embryos of Xenopus laevis and Rana pipiens. Immunofluorescent staining with anti-agrin antibodies reactive at both Rana and Xenopus neuromuscular junctions revealed immunofluorescence at AChR aggregates along nerve-muscle contacts involving both cross-species combinations. Immunofluorescent staining with an anti-agrin antibody reactive at Rana but not at Xenopus neuromuscular junctions was positive only at cross- species nerve-muscle contacts involving Rana neurons. These results provide the first demonstration that embryonic neurons supply agrin- like molecules to the synapses they form with embryonic muscle cells.

Recent experiments have indicated that neural agrin is deposited at newly forming nerve-muscle synapses and has a primary synaptogenic role there. As a step toward assessing how the spatial arrangement of new synaptic sites is regulated, we compared the pattern of agrin deposition by Xenopus neurites on culture substrate and on muscle cells. The neurons were grown on a substrate that bound their externalized agrin so tightly that it remained bound even when the neurites retracted spontaneously or were eliminated experimentally. By contrast, the neural cell adhesion molecule, NCAM, was not left behind on the substrate when the neurites were eliminated. Agrin, visualized by immunofluorescent staining, was deposited on the culture substrate in a continuous fashion along virtually the entire neuritic arbor of many spinal cord (SC) neurites. The pattern of agrin deposition by the same neurites changed from continuous to discontinuous when the neurites contacted muscle cells, and it became continuous again when the neurites returned to the culture substrate. The sites of agrin deposition on muscle cells were also sites of accumulation of ACh receptors (AChRs). Dorsal root ganglion (DRG) neurons and some SC neurons did not deposit agrin along their neuritic outgrowth, either on the culture substrate or on the muscle cells, and did not induce AChR accumulation at sites of contact with muscle cells. Besides adding to the evidence in support of agrin's synaptogenic role, the findings indicate that muscle cells significantly influence how neural agrin and synaptic sites become distributed along paths of neurite-muscle contact.

NGF is essential for the development and maintenance of sympathetic and certain sensory neurons. The NGF receptors on the surface of sympathetic ganglion cells from chick embryos were characterized; they consist of high-affinity receptors with a dissociation constant of about 10(-11) M, and low-affinity receptors with a dissociation constant of about 10(-9) M. There are more than 10 times as many low-affinity as high-affinity receptors per cell. The heterogeneity of NGF binding is not due to negatively cooperative interactions among the receptors. The high- and low-affinity components of NGF binding defined at steady state correspond to slowly and rapidly dissociating components of bound NGF seen in kinetic experiments. In addition, a very slowly dissociating component of bound NGF was observed; this component was a small fraction of binding at low concentrations of NGF but increased to 20-60% of bound NGF at the highest NGF concentrations examined. This very slowly dissociating component of bound NGF accounts for several peculiarities in the binding data not accounted for by steady-state binding of NGF to its high- and low-affinity receptors. Developmental studies showed that both high- and low-affinity NGF receptors were present on chick embryo sympathetic ganglion cells from 6.5 to 20 d in ovo. No significant differences in the numbers or affinities of the receptors were seen with cells from ganglia at 9, 11, or 15 d of development. Cultured non-neuronal cells from sympathetic ganglia had only low-affinity NGF receptors.