pal-1 is a gene required for the formation of rays by the V6 cell in males. As we have described previously,
pal-1 effectively prevents the T cell from inducing V6 to generate alae and, thereby, allows V6 to generate rays. In the absence of the T cell (following laser ablation),
pal-1 is not required for V6 ray formation. Previously we presented evidence that
pal-1 positively activates mab- 5. Supporting this model we have found that
pal-1 is suppressed by the
mab-5 gain of function allele
e1751 (formerly
lin-21). In
unc-79(
e1061) 91) 51); 90) males, only 5/100 sides had fewer than the wild- type number of rays in the fan; and in those cases, it appeared that V6 generated a few rays but not the wild-type number (V6 makes no rays at all on 85% of the sides in
pal-1(
e2091); 90)). The
e1751 lesion is a duplication of a region of DNA containing
mab-5. As Salser and Kenyon describe in an abstract in this gazette, it seems likely that the
e1751 phenotype is the result of inappropriate transcription of
mab-5. The fact that
pal1+ activity is not required in the presence of the
e1751 mutation suggests that
pal-1 turns on mab- 5 at the level of transcription. Whether or not it does so directly is unknown. At last report we had narrowed the search for
pal-1 to a 14 kb clone that carried the homeobox
ceh-3. Now we have a 5.5 kb,
ceh-3 containing clone, which rescues the
pal-1 phenotype completely. We have found 2 independent cDNA clones which are each 1.3 kb (from Chris Martin's miracle library). One is just a few bases longer than the other. The 5.5 kb clone that rescues
pal-1 contains about 1 kb upstream of the 5' end of the cDNAs and about 400 basepairs downstream of the 3' end of the cDNAs. For this reason we believe that
ceh-3 is
pal-1 (although we have not yet ruled out the alternative that there is a second gene on this clone). These cDNAs encode a ~250aa protein that contains the
ceh-3 homeodomain. As Burglin and Ruvkin showed, this homeodomain is homologous to the caudal gene of Drosophila. Both cDNAs have part of what appears to be a trans-spliced leader sequence (the longer has 10 bases homologous to the splice-leader), so we believe these cDNAs are virtually full length. We have found no other conserved domains of any known significance, but we have found one curious homology to the
mab-5 gene. In the
mab-5 cloning paper, Costa et al. pointed out a serine rich region followed by a poly-alanine stretch. Interestingly,
pal-1 has a very similar region, which includes an identical stretch of 9 amino acids SAAAAAAAN (the San Box). Such poly-amino acid stretches are common in homeodomain proteins, but their function is unknown. While this homology may be insignificant, the fact that both
pal-1 and
mab-5 act in V6 to promote ray formation makes this homology mildly but not profoundly tantalizing. Finally, to determine if
pal-1 acts directly on
mab-5 we are looking for
pal-1 binding sites in the vicinity of
mab-5. (In principle, if
pal-1 binds we have made a
pal-1/glutatione-S-transferase fusion protein (glu-pal) and expressed it in E. coli. Our results are very preliminary; but we have been able to purify this fusion protein, and we have seen sequence specific binding (as judged by band shifts) to an oligo-nucleotide probe known to be bound in vitro by engrailed and several other homeodomain proteins (E. Martin-Blanco and T Kornberg, personal communication). We hope to be able to use the glu-pal fusion to identify authentic
pal-1 binding sites near
mab-5.