SummaryNeurofilament (NF) subunits translocate within axons as short NFs, non-filamentous punctate structures ('puncta') and diffuse material that might comprise individual subunits and/or oligomers. Transport of NFs into and along axons is mediated by the microtubule (MT) motor proteins kinesin and dynein. Despite being characterized as a retrograde motor, dynein nevertheless participates in anterograde NF transport through associating with long MTs or the actin cortex through its cargo domain; relatively shorter MTs associated with the motor domain are then propelled in an anterograde direction, along with any linked NFs. Here, we show that inhibition of dynein function, through dynamitin overexpression or intracellular delivery of anti-dynein antibody, selectively reduced delivery of GFPtagged short NFs into the axonal hillock, with a corresponding increase in the delivery of puncta, suggesting that dynein selectively delivered short NFs into axonal neurites. Nocodazole-mediated depletion of short MTs had the same effect. By contrast, intracellular delivery of anti-kinesin antibody inhibited anterograde transport of short NFs and puncta to an equal extent. These findings suggest that anterograde axonal transport of linear NFs is more dependent upon association with translocating MTs (which are themselves translocated by dynein) than is transport of NF puncta or oligomers.

Neurofilaments (NFs) associate with each other and with other cytoskeletal elements to form a lattice that supports the mature axon. Phosphorylation contributes to formation of this structure by fostering cation-dependent interactions among NF sidearms. By inducing NF bundling, phosphorylation impedes their axonal transport. To examine the impact of the known NF kinase cdk5 on these phenomena, transfected cells with constructs expressing GFP-tagged NF-H sidearms (lacking the rod domain to preclude assembly) with and without site-directed mutagenesis of 7 cdk5 consensus sites, and monitored the impact on NF transport and association with the axonal NF bundle. These mutations did not alter transport but pseudo-phosphorylated mutants displayed a greater association with axonal NF bundles. By contrast, these same mutations in full-length NF-H altered NF transport as well as bundling. Since isolated sidearms cannot assemble, they can only interact with NFs via a single sidearm-sidearm interaction, while assembled NFs can form multiple such interactions. These finding suggest that individual sidearm-sidearm interactions are dynamic and do not persist long enough to slow NF transport, and that bundle formation and maintenance depends upon both the long half-life of NF polymers and the establishment of multiple phosphorylation-dependent sidearm-mediated interactions among NFs.

Neurofilaments (NFs) associate with each other and with other cytoskeletal elements to form a lattice that supports the mature axon. Phosphorylation contributes to formation of this stationary population of NFs by fostering cation-dependent interactions among NF sidearms. Association of NFs with the stationary phase indirectly competes with NF axonal transport by withdrawing NFs from kinesin-dependent motility along microtubules. We therefore hypothesized that inhibition of anterograde NF transport may increase incorporation into the stationary phase. To test this hypothesis, we treated differentiated NB2a/d1 cells expressing GFP-tagged NF subunits with monastrol, a specific inhibitor of kinesin-5. Monastrol significantly inhibited anterograde axonal transport of NF-H but not NF-M, and increased the incorporation of newly-transported NF subunits into axonal NF bundles. These findings support the notion that NF transport and bundling exert opposing forces on axonal NF dynamics, and that inhibition of anterograde transport of NFs can increase their incorporation into the stationary phase.