Cellular and molecular mechanisms of neurite targeting in the zebrafish visual system

Histologically discrete, parallel layers occur frequently in the nervous system. In many cases, each lamina is a target for innervation by a subset of neurons. We are interested in how neurites select their target laminae. Young zebrafish larvae develop two laminated neuropils in the visual system: the inner plexiform layer (IPL) of the retina and the synaptic region of the optic tectum. Using cell type-specific markers, we have characterized the IPL and tectal neuropil in detail, identifying the complement of neurites that compose each lamina. Using these maps, we investigated of the role of activity in IPL sublamination, and completed a forward genetic screen to identify molecules regulating lamination in the retina and tectum. In mammals, retinal activity is important for the sublamination of ganglion cell (GC) dendrites. Using pharmacological tools and the brudas/NSF mutant, we show that the zebrafish IPL develops in an activity-independent manner, at least until 7 dpf. Hard-wiring mechanisms may be conserved across these species, but other, activity-dependent mechanisms are not. Our screen uncovered five informative mutants. We exploited the moonraker (mra) and notorious (noto) mutants to explore the importance of cell-cell interactions in IPL development. Transplantation studies with mra demonstrate that a subset of amacrine cells (ACs) rely on cell non-autonomous cues to sublaminate. The mra locus encodes DEAD-box protein 19, an RNA helicase not previously implicated in neurite targeting. In noto, GC, AC, and bipolar cell neurites appear to branch outside their target sublaminae. These IPL defects are autonomous to GCs; GC dendrites are able to instruct the sublamination of partner neurites. In the tectal neuropil, noto GC axons and some tectal dendrites are highly disorganized. While wildtype axons project to a single tectal lamina, noto GCs can form arbors in two or three. We explored the developmental relationship between GC axons tectal dendrites by characterizing the tectum of lakritz (lak) mutants, which lack GCs. Our observations uncovered only slight defects in lak mutant tectal dendritogenesis. The grossly visible aspects of lamination in the tectum are largely GC-independent.