Molecular cues are proposed to guide axon growth cones along particular pathways in developing animals. "Guidance molecules" are either secreted or presented on the surface of pre-existing substrates. This model for understanding axon pathfinding assumes that growth cones have receptor molecules that allow them to distinguish among various molecular cues. Therefore, the pathway that a particular growth cone follows may depend on the specific spectrum of hypothetical receptor molecules that it expresses. This notion has been demonstrated for
unc-5: growth cones of neurons that express the UNC-5 receptor migrate dorsally in response to the UNC-6 secreted signal. However, this begs the question of how a neuron determines which receptor molecules to express. Put another way, a neuron has to "know" which pathway its axon should follow in order to express the appropriate receptors. The
fax-1 (defective fasciculation of axons) gene may provide this function for neurons that extend axons along the left bundle of the ventral nerve cord (VNCL). Mutations in
fax-1 lead to defects in axon pathfinding by the HSNL, AVKR and PVQL axons of the VNCL. The axons extend to their normal length, indicating that they are not defective for extension. Moreover, each of the VNCL neurons is normal in every way we have examined other than axon pathfinding, suggesting
fax-1 functions in axon pathfinding, but not other aspects of neuronal identity. We cloned
fax-1 and determined that it encodes a predicted member of the nuclear hormone receptor superfamily of zinc-finger transcription factors. The FAX-1 DNA-binding domain is most similar to a group of nuclear receptors from several species that are related to the Drosophila tailless gene. We have defined the nature of the molecular lesions of the two existing
fax-1 alleles:
gm83 is a nonsense mutation in the second zinc-finger, and
gm27 is a deletion that removes the entire
fax-1 gene and flanking genes. Therefore,
fax-1 is likely to function as a transcription factor and may regulate genes that function in specifying which pathway cues VNCL growth cones follow. A requirement of this hypothesis is that
fax-1 should be expressed in the very same VNCL neurons that exhibit axon guidance defects in its absence. We have created
fax-1::GFP fusion transgenes as a first step in addressing this issue. We observe
fax-1::GFP expression in all stages, from mid-embryogenesis through adults, in a subset of the nervous system (and the non-neuronal dtc's). Approximately 15 different neurons or neuron pairs express
fax-1::GFP, including AVKR. None of the
fax-1::GFP reporters are expressed to detectable levels in postembryonic HSNL and PVQL neurons. Embryonic expression of the
fax-1::GFP reporters is somewhat broader and currently under investigation. Thus,
fax-1 may function in at least AVKR to direct the transcription of genes that mediate axon pathfinding. For a description of
fax-1 function outside of the VNCL, see the abstract by Carmean et al.