The evidence discussed thus far suggests a scheme in which nodes form during remyelination as the low density of axonal internodal Na+ channels cluster adjacent to the tips of myelinating glia, move laterally as the glial processes elongate, and form nodes of Ranvier by fusing with a neighboring cluster (Dugandzija-Novakovic et al., 1995; Vabnick et al., 1996; Lambert et al., 1997; Ching et al., 1999). We now advance the hypothesis that Na+ channels are clustered by exclusion from regions of contact with myelinating glial cells (Kazarinova-Noyes and Shrager, 2002) and that this could occur if the rate of lateral diffusion is increased by this association. Evidence suggests a molecular mechanism that might account for such a scheme. Na+ channels at the node of Ranvier are part of a large complex of transmembrane, extracellular matrix, and intracellular proteins (Kazarinova-Noyes and Shrager, 2002; Poliak and Peles, 2003; Salzer, 2003). Among these are members of the L1-type cell adhesion family, including NrCAM and the 186 kDa form of neurofascin (NF186), whose association with ankyrinG, and hence with the cytoskeleton, is controlled by phosphoryla-tion of a specific intracellular tyrosine in a FIGQY domain (Garver et al., 1997; Jenkins et al., 2001). There is evidence that Na+ channel P subunits are also linked to ankyrinG by a phosphorylation-dependent mechanism (Malhotra et al., 2002). Na+ channels in excitable cell membranes are thought to be normally highly immobile (Stuhmer and Almers, 1982; see also Custer et al., 2003), presumably through the ankyrinG link to the cytoskeleton. We therefore consider a mechanism of clustering in which glial contact activates an axonal tyrosine kinase that phosphorylates one or more key membrane components, breaking the link to ankyrinG, and thereby increasing the effective coefficient of lateral diffusion. As a Na+ channel diffuses beyond the axoglial contact zone, it enters a region in which it may be dephosphorylated by the combined absence of the activated kinase and presence of a tyrosine phosphatase. (It should be noted that there is evidence that receptor protein tyrosine phosphatase P associates with Na+ channels [Ratcliffe et al., 2000)]. This sharp change in enzymatic environment may also be mediated by Schwann cell microvilli that are arrayed close to the nodal axolemma. These structures are rich in the ezrin-radixin-moesin (ERM) protein family, appear very early in node formation, and may contain receptors for axonal proteins, including NF186 (Melendez-Vasquez et al., 2001; Scherer et al., 2001). Channels would then cluster at the edge of the myelinating glial cell as their link to ankyrinG is restored. Although qualitatively plausible, this model requires a quantitative test to judge its applicability. There are several key parameters: the diffusion coefficient of Na+ channels in axons, rates of tyrosine phosphorylation and dephosphorylation of membrane proteins, the Na+ channel densities at the node and internode, and the rate of growth of glial processes. Values for all of these have been estimated from published measurements, the last coming from the experiment shown in Fig. 4. Because it is thus possible to construct a quantitative model in which virtually all critical parameters are under tight constraints, we felt that this exercise might serve as a plausibility test to see if the resulting clustering matches can be seen experimentally.
Was this article helpful?