We have identified the
unc-3 gene through deficiency end-point mapping using fluorescence in situ hybridization (FISH) followed by identification of RFLPs associated with particular
unc-3 alleles in the defined region. Two alleles have been especially useful.
mn419, isolated in a TR679 mutator background, is associated with a Tc1 insertion; reversion to wild-type locomotion coincides with excision of the Tc1 element.
cn4146 (kindly provided by R. Hosono; WBG 10:1) is a rearrangement in the
unc-3 region most easily explained as an inversion accompanied by a 3-kb deletion that includes most of the
unc-3 coding region, and is the best candidate for an
unc-3 null mutation. The major 2.3-kb transcript from this region is most abundant in embryos and L1 larvae, and is predicted to encode a product homologous to the transcription factors EBF/OLF-1 in rodents (Hagman et al., Genes Dev. 7:760; Wang and Reed, Nature 364:121) and collier in Drosophila (Crozatier et al., Curr. Biol. 6:707). This group of very similar transcription factors has a 200 amino acid DNA-binding domain as well as a predicted helix-loop-helix domain; UNC-3 shares 80% identity in the DNA binding domain and 70% identity in the HLH region with the other family members. The major
unc-3 transcript is derived from an 11-kb genomic sequence (excluding regulatory regions) that extends into a gap in cosmid coverage of the physical map. We isolated a l genomic clone extending into the gap that rescues
unc-3. We engineered a derivative of this phage clone with an in-frame GFP positioned in the loop of the predicted HLH domain; this construct rescues the
unc-3 locomotion defect in 5/5 integrated lines. GFP can be detected in embryos by approximately the 400-cell stage and is localized in nuclei. At the time of hatching, expressing cells can be identified as the DA and DB motor neurons, the ASI amphid neuron, and several other more weakly-expressing cells in the head. By the L2 stage the newly-emerged excitatory motor neuron classes VA, VB, and AS also express GFP, as well as a few other unidentified cells in the tail. Expression in all five excitatory motor neuron classes continues into adulthood. The DD and VD inhibitory motor neurons and the VC motor neurons do not appear to express GFP. The observed GFP expression pattern corresponds well with the phenotype of
unc-3 mutants. EM serial section reconstruction of
unc-3 mutants showed a highly disorganized ventral nerve cord (J. White, The Worm Book), accounting for the locomotory defect in
unc-3 animals. Therefore it appears
unc-3 is required for proper development of excitatory motor neurons in the ventral nerve cord. In addition,
unc-3 mutants have a propensity towards dauer formation in the presence of food, and display a synthetic dauer constitutive phenotype in combination with several other mutations, including
unc-31 (Bargmann et al., CSHSQB 60:529). Bargmann and Horvitz observed (WBG 11:1) that laser ablation of ASI in an
unc-31 background resulted in constitutive dauer formation (Daf-c); therefore it appears that mutation in
unc-3 results in an ASI defect that contributes to the synthetic Daf-c phenotype in
unc-31;
unc-3 animals. We have begun to examine defects in the ASI amphid neuron in a collection of seven
unc-3 alleles by assaying dauer formation and FITC dye-filling. All
unc-3 alleles tested have a synthetic Daf-c phenotype with
unc-31 at 25o; however, the penetrance varies, allowing us to classify the alleles as strong or weak. In wild-type animals ASI is stained by the fluorescent dye FITC, taken up from the environment. Examination of FITC staining in
unc-3 mutants indicates that their ASI neurons are defective in dye-filling and that the penetrance of this phenotype also varies with different alleles. Interestingly, although our
unc-3::GFP construct rescues locomotion very well in a
cn4146 background, it does not rescue the ASI dye-filling defect in the same background, suggesting that the functions of
unc-3 in motor neurons and in ASI may be separable.