The seven C. elegans smg genes are necessary for identifying and degrading mRNAs that contain premature nonsense codons. In smg(+) genetic backgrounds, nonsense mutation-containing mRNAs are less stable than their WT counterparts. Mutations in any of the seven smg genes eliminate the system of nonsense-mediated mRNA decay, which cause nonsense-mutant mRNAs to accumulate to near WT levels (Pulak and Anderson. 1993. Genes and Dev. 7: 1885). Nonsense-mediated mRNA decay is a non-essential system and smg(-) animals have a near WT phenotype. We are working on the cloning of several smg genes to gain a better understanding of this system. We have taken advantage of the well characterized genetic and physical maps of C. elegans in the cloning of
smg-1. After three-factor mapping, a transformation rescue assay was used to identify
smg-1-containing clones. The
smg-1 rescuing region is contained within a 13 kb genomic clone. Southern blot analysis of ten
smg-1 alleles isolated in a transposon-active background identifies insertions in three alleles (
r904,
r910, and
r913). All three insertions map to the 13 kb rescuing region. Clones from this region identify a single transcript of 7.8 kb in N2. This transcript is larger in
smg-1(
r913) and absent in
smg-1(
r904) mutants. Thus, this 7.8 kb transcript is likely
smg-1 and
smg-1(
r904) is a null. The phenotype of
smg-1(
r904) is indistinguishable from other viable smg alleles. Like
smg-2 (Carr, B. Meeting Abstracts, 1993),
smg-1 is a non-essential gene, confirming the
smg-1 null phenotype suggested by genetic analysis. The large size of the putative
smg-1 mRNA is consistent with the fact that
smg-1 alleles are the most commonly isolated alleles in general screens for smg(-) mutations. Partial cDNAs from oligo-d(T)- and randomly primed libraries have been identified and assembled into a contig comprising most of the
smg-1 transcript. Sequence analysis of genomic and cDNA clones demonstrates that the C-terminal region of SMG-1 is approximately 30% identical over 300 AA to the kinase domain of several phosphatidylinositol-3-kinases, including the recently cloned human gene ATM. This homology appears to be limited to the kinase domain; no significant homologies to other regions of SMG-1 have been identified. We are continuing to sequence
smg-1 and are preparing polyclonal antibodies to investigate the SMG-1 protein and its subcellular localization.