Genes of the Pax-6 family encode paired domain and homeodomain containing transcriptional regulators that play important roles in eye and brain development in many species. The C. elegans Pax-6 locus encodes multiple homeodomain containing proteins both with and without a paired domain.
vab-3 mutations affect the paired domain of this locus1. The
mab-18 mutation affects alternative N-terminal domains that do not encode a paired domain2.
vab-3 and
mab-18 mutations fully complement one another, suggesting that the paired domain and non-paired domain isoforms of Pax-6 have distinct functions in development. We have found that the C. elegans Pax-6 locus is more complex transcriptionally than previously suspected. Probes specific to the VAB-3 paired domain detect two transcripts of 3.6 and 2.6 kb on Northern blots that are reduced in
vab-3 mutants, suggesting that these transcripts correspond to
vab-3. Probes from genomic regions that contain no
vab-3 exons recognize non-paired domain encoding transcripts of 2.6, 2.3 and 2.0 kb. The 2.0 kb band may correspond to the three
mab-18 transcripts predicted from RT-PCR analysis2. The different transcripts from the Pax-6 locus are differentially expressed during development. The origins and functions of the 3.6 kb (paired domain encoding) and the 2.6 and 2.3 kb (non-paired domain encoding) transcripts are under investigation. Mutations affecting the domains common to both paired domain and non-paired domain isoforms should cause both Mab-18 and Vab-3 phenotypes. Consistent with this, the mutation
e1796, which has missense alterations in the homeodomain, causes a weak Vab-3-like phenotype and a strong Mab-18 phenotype2. We have recently found that the two mutations
e1022 and
e1178, which cause strong Vab-3-like phenotypes, also cause weak Mab-18 phenotypes and have lesions in exons common to
vab-3 and
mab-18.
e1022 is a splice site mutation in a homeodomain encoding exon and
e1178 (which is ICR191 induced) results in a frameshift in the PST-rich C-terminal domain thought to be important for transcriptional activation by Pax-6 proteins.
e1022,
e1178 and
e1796 mutants display H lineage defects distinct from those of other
vab-3 mutants. In previously described
vab-3 mutants, H0 adopts the fate of its posterior neighbor H1, suggesting that
vab-3 functions to specify the H0 fate. In
e1022 and
e1178 mutants both H0 and H1 appear to adopt H2-like fates, based on limited lineage analysis and the presence of ectopic deirid cuticular substructures in
e1022 and
e1178 dauers. In wild-type dauers a distinctive funnel- or horn-shaped structure is found in the cuticle above the deirid and postdeirid3. These structures appear to be generated by the deirid and postdeirid socket cells and thus serve as markers for the deirid socket (H2.aa) fate. Many
e1022 and
e1178 dauers have two ectopic cuticular substructures in positions consistent with their being made by H0.aa and H1.aa, while
e1796 dauers have one ectopic cuticular substructure, consistent with our lineage data indicating that in
e1796 mutants H1 is transformed to H2 and H0 executes an abnormal proliferative lineage. Thus, these mutations disrupt functions required to specify both H0 and H1 fates. Because the
e1022,
e1178 and
e1796 mutations cause phenotypes different from the other
vab-3 alleles we have classified them as Class III mutations in the Pax-6 locus, as distinct from Class I (the eight other
vab-3 mutations) and Class II (
mab-18(
bx23)). The strong Class III mutations
e1022 and
e1178 cause stronger defects in head morphogenesis than strong Class I mutations such as
e648, which is a nonsense mutation in the paired domain. These data suggest that the non-paired domain containing isoforms affected by Class III mutations play roles in head morphogenesis and the specification of the H1 fate. Since
mab-18 mutants have normal head development, such non-paired domain isoforms must be unaffected by the
bx23 deletion, or alternatively the isoforms affected in
mab-18 could function redundantly with paired domain isoforms in head patterning and morphogenesis. 1. Chisholm & Horvitz, Nature 377:52. 2. Zhang & Emmons, Nature 377:55. 3. Albert & Riddle, J. Comp. Neurol. 219:461; W. Fixsen, Ph.D thesis, MIT.