The informational suppressor genes
smg-1 through
smg-7 play a role in identifying and degrading mRNAs that contain nonsense codons. In smg(+) genetic backgrounds,
unc-54 nonsense and frameshift mutation-containing mRNAs are unstable. Mutations in any of the seven smg genes cause these mutant mRNAs to accumulate to near wild-type levels (see Pulak and Anderson. 1993. Genes and Dev. 7: 1885; Cali and Anderson, Abstract this meeting). We used three factor mapping and a microinjection assay to identify two overlapping cosmids, F56F2 and C14A12, that contain
smg-1 (O'Connor, 1993 Meeting Abstract). Subsequent subcloning and microinjection experiments localized the
smg-1 rescuing activity to a 10 kb region. Southern analysis of ten
smg-1 alleles isolated in a
mut-2(
r459) background identifies insertions in three alleles (
r904,
r910, and
r913). All three insertions map to the 10 kb rescuing region. A 4 kb subclone from the middle of the rescuing region identifies a single transcript of 7.8 kb on an N2 northern blot. The size of this transcript is altered in
smg-1(
r904) and
smg-1(
r913) mutants. Thus, it seems likely that this 7.8 kb transcript is
smg-1. The large size of the putative
smg-1 mRNA is consistent with the large size of the
smg-1 mutational target. Partial cDNAs of this transcript have been identified and the isolation of full length clones is underway. Sequence analysis of genomic and cDNA clones predicts that SMG-1 protein contains a kinase domain. The C-terminal region of SMG-1 has approximately 50% identity over 400 AA to the kinase domain of several phosphatidylinositole 3-kinases. This homology appears to be limited to the kinase domain; no significant homologies to other regions of SMG-1 have been identified. Sequence analysis is continuing.