About 5 years ago, John White described ventral cord defects in unc- 4
(e120): most VA motor neurons assume the pattern of synaptic input normally reserved for the VB's (White et al., WBG 9,
p81, 1985). These changes account for the inability of
unc-4 worms to back-up which is normally mediated by the VA's. We have previously described a transposon-tagging strategy that identified Tc1 insertions within 0. 15 mu of
unc-4 (WBG 11, p 41, 1989). Burglin et al (Nature 341, 239- 243, 1989) independently discovered a homeobox motif,
ceh-4, in a 1.05 kb EcoRI fragment from the nearby cosmid C07E2. We subsequently probed Southerns with C07E2 and identified a 1.8 kb EcoRI fragment that is either altered or deleted in five different
unc-4 alleles (
e887,
e2308,
e2314,
e2320, wd-1). We have now sequenced the 1.8 kb EcoRI genomic fragment and have detected a short ORF (110 bp) that is highly homologous (~60% identity) to the paired (prd) family of homeodomains first described in Drosophila. The 110 bp exon appears to encode amino acids 11-46 of the 60 aa consensus homeodomain and is in phase with
ceh-4 which encodes amino acids 47-60 (see below). Therefore,
unc-4 is likely to regulate synaptic specificity indirectly by controlling transcription of other genes which in turn define the pattern of synaptic input to the VA's. These downstream genes may act during the terminal stages of VA differentiation since VA morphology and process placement in the ventral cord are essentially normal in
unc-4(
e120). We are currently searching for
unc-4 suppressors. The presumptive
unc-4 transcript exhibits some unusual properties. Northerns of N2 RNA probed with either the 1.8 kb or 1.05 kb fragments detect a large (>7.5 kb) transcript that is absent from
e887 and
e2308 RNA. The
unc-4 transcript does not stick to poly dT which may explain our failure to detect a cDNA clone in the Arhinger embryo and L1 cDNA libraries. The putative
unc-4 transcript is much more abundant in embryos than in larvae. This is surprising since VA neuroblasts are born in the late L1 and ventral cord wiring occurs immediately before the L1/L2 molt. We have performed a temperature shift experiment with the temperature sensitive
unc-4 allele,
e2322 (from M. Shen), and shown that the TSP does not begin until the end of the L1 stage. In other words, VA cell fate is not dependent upon embryonic expression of
unc-4 but is governed by
unc-4 expression in the late L1. A similar pattern has been observed for
lin-10 in which the TSP is restricted to the L2 although
lin-10 mRNA is abundant in embryos (S. Kim and H. Horvitz, Genes and Dev. 4, 357-371, 1990). These data can be interpreted in a number of different ways:
unc-4 may have an embryonic function that has yet to be identified;
unc-4 expression may be subject to post-transcriptional control;
unc-4 action may depend upon another gene product that is selectively expressed during the TSP. Take your pick. Perhaps
unc-4 specific antibodies will shed light on this question. In any event, it is clear that we have a lot of sequencing to do. Our current strategy is to use an internal priming site to generate cDNA clones. [See Figure 1]