The state of the union. Neuromuscular junction

The vitronectin receptor associates with clathrin-coated membrane domains via the cytoplasmic domain of its beta5 subunit

Rat myotubes cultured in fetal calf serum adhere to vitronectin-coated substrates through two distinct structures, focal contacts and clathrin-coated membrane domains. We studied the integrins in myotubes to learn how they associate with these two domains. Double label immunofluorescence studies with antibodies specific for clathrin, vinculin and several forms of integrin showed that focal contacts and clathrin-coated membrane domains contain both vitronectin receptors (VnR, containing beta-3 and beta-5integrins) and fibronectin receptors (FnR, containing beta1-integrin). VnR but not FnR associates tightly with the substrate in both domains, as the VnR alone remains attached to the coverslip when the lipid bilayer and other membrane proteins are removed by detergent. Ultrastructural studies confirmed the localization of the beta5 subunit of the VnR at both domains. We used intracellular injection and affinity chromatography to test the possibility that clathrin at coated membrane domains associates with the cytoplasmic sequence of the beta5 subunit of the VnR. Injection of a synthetic peptide containing the NPXY motif from the cytoplasmic domain of the human beta5 subunit, SRARYEMASNPLYRKPIST, depleted clathrin from coated membrane domains without affecting clathrin in perinuclear structures or vinculin at focal contacts. Injection of the homologous beta1 peptide, MNAKWDTGENPIYKSAVITT, also containing an NPXY motif, had no significant effect on any of these structures. Affinity matrices containing the beta5 but not the beta1 peptide selectively retained clathrin from myotube extract, and bound clathrin could be selectively eluted by soluble forms of the beta5 but not the beta1 peptide. Thus, a sequence including the NPXY motif in the integrin beta5 subunit is involved in the specific anchoring of the VnR, but not the FnR, to clathrin-coated membrane.

The zinc finger domain of the 43-kDa receptor-associated protein, rapsyn: role in acetylcholine receptor clustering

We injected rat myotubes with proteins and antibodies to assess the importance of the zinc finger (ZnF) domain of the 43-kDa receptor-associated protein, rapsyn, in clustering acetylcholine receptors (AChR). Injection of rat myotubes with a fusion protein containing the ZnF domain of rapsyn disrupted AChR clusters. Clusters were unaffected by a fusion protein containing a double mutant that does not bind zinc. Similar results were obtained with the purified wild type and mutant ZnF domains. The ZnF of HIV-1 nucleocapsid protein had no effect. AChR clusters were also disrupted in myotubes injected with antibodies to the ZnF domain, followed by injection of anti-antibodies. Injection of antibodies directed against a different rapsyn epitope or against the cytoplasmic domain of the AChR had no effect. In transfection experiments with HEK 293 cells, the ZnF domain failed to associate with membrane aggregates containing full-length rapsyn, AChR, or rapsyn and AChR together. We conclude that the ZnF domain of rapsyn provides a binding site essential for AChR clustering, but that this site is unlikely to be involved in high affinity binding of rapsyn to itself or to AChR.

Formation of postsynaptic-like membranes during differentiation of embryonic stem cells in vitro

To analyze the formation of neuromuscular junctions, mouse pluripotent embryonic stem (ES) cells were differentiated via embryoid bodies into skeletal muscle and neuronal cells. The developmentally controlled expression of skeletal muscle-specific genes coding for myf5, myogenin, myoD and myf6, alpha 1 subunit of the L-type calcium channel, cell adhesion molecule M-cadherin, and neuron-specific genes encoding the 68-, 160-, and 200-kDa neurofilament proteins, synaptic vesicle protein synaptophysin, brain-specific proteoglycan neurocan, and microtubule-associated protein tau was demonstrated by RT-PCR analysis. In addition, genes specifically expressed at neuromuscular junctions, the gamma- and epsilon-subunits of the nicotinic acetylcholine receptor (AChR) and the extracellular matrix protein S-laminin, were found. At the terminal differentiation stage characterized by the formation of multinucleated spontaneously contracting myotubes, the myogenic regulatory gene myf6 and the AChR epsilon-subunit gene, both specifically expressed in mature adult skeletal muscle, were found to be coexpressed. Only the terminally differentiated myotubes showed a clustering of nicotinic acetylcholine receptors (AChR) and a colocalization with agrin and synaptophysin. The formation of AChRs was also demonstrated on a functional level by using the patch clamp technique. Taken together, our results showed that during ES cell differentiation in vitro neuron- and muscle-specific genes are expressed in a developmentally controlled manner, resulting in the formation of postsynaptic-like membranes. Thus, the embryonic stem cell differentiation model will be helpful for studying cellular interactions at neuromuscular junctions by "loss of function" analysis in vitro.

Intracellular calcium regulates agrin-induced acetylcholine receptor clustering

Agrin is an extracellular matrix protein that directs neuromuscular junction formation. Early signal transduction events in agrin-mediated postsynaptic differentiation include activation of a receptor tyrosine kinase and phosphorylation of acetylcholine receptors (AChRs), but later steps in this pathway are unknown. Here, we have investigated the role of intracellular calcium in agrin-induced AChR clustering on cultured myotubes. Clamping intracellular calcium levels by loading with the fast chelator BAPTA inhibited agrin-induced AChR aggregation. In addition, preexisting AChR aggregates dispersed under these conditions, indicating that the maintenance of AChR clusters is similarly dependent on intracellular calcium fluxes. The decrease in AChR clusters in BAPTA-loaded cells was dose-dependent and reversible, and no change in the number or mobility of AChRs was observed. Clamping intracellular calcium did not block agrin-induced tyrosine phosphorylation of the AChR beta-subunit, indicating that intracellular calcium fluxes are likely to act downstream from or parallel to AChR phosphorylation. Finally, the targets of the intracellular calcium are likely to be close to the calcium source, since agrin-induced AChR clustering was unaffected in cells loaded with EGTA, a slower-binding calcium chelator. These findings distinguish a novel step in the signal transduction mechanism of agrin and raise the possibility that the pathways mediating agrin- and activity-driven changes in synaptic architecture could intersect at the level of intracellular calcium fluxes.