nob-1 IIIR, identified in screens for genes that control posterior embryonic patterning, is defined by three alleles:
ct223 and
ct351 are severe loss-of-function mutations that result in the embryonic lethal Nob (no back end) phenotype, and
ct230 is a weak loss-of-function mutation that results in viable animals with mildly deformed tails. We recently reported the cloning and characterization of two
nob-1 cDNAs (kindly provided by Yuji Kohara) [WBG 15(1):29, 1997; 15 (2):28, 1998]. The smaller of the two cDNAs, which contains a 1.35 Kb insert, can encode a protein containing a homeodomain similar to those of the posterior paralog group of Hox genes, including Abd-B and the vertebrate Hox 9-13 genes. The second,
nob-1B, a longer clone of 2.2 Kb, appears to be alternatively spliced: it contains an additional 25 amino acids from an extended second exon, and the third intron, between the codons for amino acids 50 and 51 in the homeodomain, is not removed. Therefore, this transcript should not encode a functional homeodomain. To understand the molecular basis for the observed phenotypes, we are using PCR and newly generated genomic sequence (thanks to Alan Coulson and the C. elegans G.S.C.) to amplify and sequence the
nob-1 region for each mutant allele. So far, we have determined that the viable
ct230 allele, is a 192 bp deletion that removes the last 3 bp of exon 2 (or the last 78 base pairs of extended exon 2 in the larger cDNA) and 189 bp of the second intron. Barring a cryptic splicing event, this mutation should result in truncated messages that terminate before the homeobox.
ct223 and
ct351 also appear to be deletions, although their precise endpoints are not yet known.
ct223 is a deletion from the predicted translational start site to at least 1.5 Kb upstream, which could eliminate essential transcriptional regulatory elements resulting in lack of all
nob-1 transcripts. In contrast, the
ct351 deletion begins in the second intron and extends at least 4.0 Kb in the 3' direction past the predicted end of
nob-1. Interestingly, in addition to the
nob-1 homeobox,
ct351 also deletes a second homeobox that resides approximately 2.5 Kb 3' of
nob-1. The latter sequence is predicted to encode another homeodomain of the posterior paralog group with 88% sequence similarity to Drosophila Abd-B. This is in comparison to the homeodomains of
nob-1 and
egl-5, which are approximately 72% similar to Abd-B. We do not yet know if the second homeobox is part of a new gene (Genefinder fails to predict one), but preliminary RNA blot analysis shows that this homeobox is transcribed. We predict that
ct351 should result in loss not only of
nob-1 transcripts, but also of transcripts containing the second homeobox. Since
ct223 and
ct351 result in similar loss-of-function phenotypes, and
ct223 appears to be a deletion at the 5' end of
nob-1, we hypothesize that these two homeobox genes share some 5' regulatory sequences, if not 5' coding sequence, as well. Thus, the viable mutation may only affect one of the homeodomains, while the lethal mutations affect both. Our current experiments are directed towards determining the precise nature of the lesions in
ct223 and
ct351, and how each mutation affects transcript production. We also plan to isolate full-length transcripts containing the second Abd-B-like homeodomain to determine their structure, and to begin using RNAi to explore the roles of each transcript in posterior embryonic patterning. We are also searching for more homeobox genes in this region, so far without success; however, sequencing in this region is still unfinished.