Nonsense mutant mRNAs are generally less stable than WT mRNAs. The seven smg genes are required for this "nonsense-mediated" mRNA decay. In smg mutants, otherwise unstable nonsense mRNAs accumulate to approximately WT levels. We have previously reported the cloning of
smg-5 (Anders and Anderson, 1995 Worm Meeting Abstracts) and
smg-7 (Cali and Anderson, 1995 Worm Meeting Abstracts).
smg-5 encodes a 549 amino acid protein that is not similar to any protein sequence in the public databases.
smg-7 is predicted to encode a 53 kDa protein with two notable features: (i.) three repeats of a conserved 34 amino acid domain known as a tetratricopeptide repeat (TPR) and (ii.) a negatively charged COOH-terminus. TPRs are predicted to form amphipathic helices and are believed to mediate homo- and heteromeric protein:protein interactions. Here we describe the characterization of SMG-5- and SMG-7-specific antisera and provide evidence that these proteins associate in vivo. Polyclonal antisera to both SMG-5 and SMG-7 were generated in rabbits using bacterially expressed fusion proteins as immunogen. Both antisera specifically detect 0.1 ng of the appropriate fusion protein. Antisera raised to SMG-5 fusion protein detect a 56 kDa protein in N2 but not
smg-5(-) extracts by immunoprecipitation followed by Western blotting. Antisera raised to SMG-7 fusion protein detect 54 and 57 kDa proteins on Western blots of total worm protein. Several results indicate the 54 kDa protein is encoded by
smg-7. First, its apparent molecular weight is quite close to that predicted for SMG-7. Second, the abundance of the 54 kDa protein is reduced (but not eliminated) in
smg-7(
r1131::Tc1) animals.
smg-7(
r1131::Tc1) animals accumulate approximately 60% of WT levels of WT-sized
smg-7 mRNA. Thus, it is not surprising that this mutation does not eliminate SMG-7 synthesis. We are currently investigating whether the 57 kDa protein is also encoded by
smg-7 (perhaps by an alternatively spliced
smg-7 mRNA), and are considering the possibility this protein represents a post-translationally modified form of SMG-7. The presence of a protein:protein interaction domain in SMG-7 suggests it might associate with other proteins, perhaps including the products of some or all of the other smg genes. Therefore, we have tested whether SMG-5 and SMG-7 associate in vivo as judged by co-immunoprecipitation. Both SMG-5 and SMG-7 co-precipitate from extracts when either antisera is used. The proteins are detected by Western blot analysis of immunoprecipitates. Coprecipitation of SMG-7 with SMG-5-specific antisera requires WT SMG-5, as SMG-7 does not coprecipitate from
smg-5(-) extracts (although SMG-7 is present in
smg-5(-) extracts). Only the 54 kDa (not the 57 kDa) polypeptide identified by anti-SMG-7 antisera co-precipitates with SMG-5. We conclude that SMG-5 and SMG-7 physically interact as components of a multi-protein complex. Work is in progress to determine if the two proteins interact directly or indirectly, and if the integrity of the complex is maintained in other smg mutants.