The cytoskeleton builds and supports the complex architecture of neurons. It orchestrates the specification, growth, and compartmentation of the axon: axon initial segment, axonal shaft, presynapses. The cytoskeleton must then maintain this intricate architecture for the whole life of its host, but also drive its adaptation to new network demands and changing physiological conditions. Microtubules are readily visible inside axon shafts by electron microscopy, whereas axonal actin study has long been focused on dynamic structures of the axon such as growth cones. Super-resolution microscopy and live-cell imaging have recently revealed new actin-based structures in mature axons: rings, hotspots and trails. This has caused renewed interest for axonal actin, with efforts underway to understand the precise organization and cellular functions of these assemblies. Actin is also present in presynapses, where its arrangement is still poorly defined, and its functions vigorously debated. Here we review the organization of axonal actin, focusing on recent advances and current questions in this rejuvenated field.

1-Deoxynojirimycin (DNM) is a saccharide decoy that inhibits cellular ␣-glucosidase I-II activity. Treatment by DNM of human immunodeficiency virus (HIV)-infected lymphocyte cultures inhibits virus spread. The functional properties of the membraneassociated Env glycoprotein (Env) modified in the presence of DNM remain unclear because previous reports on this subject have essentially used recombinant soluble Envs whose properties differ notably from those of Env anchored on the surface of the virus. To model virus-associated Env synthesized in the presence of DNM, native Env was expressed at the surface of mammalian cells treated with DNM. As expected, its glycosylation pattern was altered in the presence of the inhibitor. Env was found able to bind CD4, whereas its ability to induce membrane fusion was abolished. The immunoreactivity of regions involved in interactions of Env with CXCR4 (V1, V2, C2, and V3) was modified and Env displayed altered interaction with this coreceptor. These results are consistent with the inhibition by DNM of virus entry at the Env/coreceptor interaction step. Finally, preliminary data indicate that suboptimal concentrations of DNM and natural or synthetic CXCR4 ligands used in combination potently inhibit the Env-mediated membrane fusion process. Altogether, our results suggest that DNM and its analogs deserve further investigation as anti-HIV agents in combination with experimental compounds targeting CXCR4 to inhibit each partner of this crucial step of HIV entry.