The binding of initiator methionyl-tRNA to ribosomes is catalyzed in eukaryotic organisms by the heterotrimeric factor eIF2, whereas in prokaryotes a single polypeptide factor IF2 performs the same function. We have identified and characterized IF2 homologs in archaea, the yeast Saccharomyces cerevisiae and humans. Previous studies demonstrated that the yeast IF2 homolog, encoded by the FUN12 gene, is a general translation initiation factor. Biochemical assays demonstrated that the human IF2 protein possesses ribosome-dependent GTPase activity and promotes the ribosomal subunit joining step of protein synthesis. In recognition of this activity the eukaryotic IF2 homologs have been renamed eIF5B. Mutation of conserved residues in the eIF5B GTP-binding domain revealed a critical role for GTP-binding and hydrolysis by eIF5B for translation initiation. Using an eIF5B mutant that utilizes XTP in place of GTP, we have demonstrated that at least two nucleotide (GTP) hydrolysis events are required for eukaryotic translation initiation. Consistent with a role in subunit joining, yeast strains lacking eIF5B show increased levels of leaking scanning. Finally, yeast two-hybrid, in vitro protein binding assays and co-immunoprecipitation experiments revealed that yeast eIF5B directly interacts with the translation factor eIF1A (a homolog of the prokaryotic factor IF1). In addition, overexpression of eIF1A specifically exacerbated the growth defect of strains lacking, or expressing truncated forms of, eIF5B. This physical and functional interaction between the two evolutionarily conserved translation initiation factors may facilitate methionyl-tRNA binding to the ribosomal P site.A second research interest is phosphorylation of the translation initiation factor eIF2. The mammalian kinases PKR, HRI, and PERK and the yeast kinase GCN2 specifically phosphorylate serine-51 on the alpha subunit of eIF2 to regulate translation during stress conditions. Mutational analysis of yeast eIF2alpha identified amino acid substitutions at residues 49 and 50 as well as in a conserved sequence motif around 30 residues C-terminal of the Ser-51 phosphorylation site that impair translational regulation. Biochemical studies revealed that a subset of the mutations in eIF2alpha blocked phosphorylation by the GCN2 and PKR kinases both in vivo and in vitro. These results demonstrate that kinase recognition of eIF2alpha utilizes residues both nearby and, surprisingly, remote from the phosphorylation site. We demonstrated that the vaccinia virus K3L protein and the swine pox virus C8L protein are pseudosubstrate inhibitors of PKR, and can suppress PKR toxicity in yeast. This inhibition of PKR by K3L and C8L was dependent on residues conserved among eIF2alpha, K3L and C8L. Fourteen independent mutations in the carboxyl-terminal half of the PKR kinase domain rendered the kinase resistant to K3L inhibition, and these mutations are predicted to alter contacts between the kinase and substrate. Finally, experiments in yeast and mammalian cells demonstrated the importance of dimerization for PKR activation in vivo. Whereas an isolated PKR kinase domain was inactive in vivo, fusion of the kinase domain to heterologous dimerization domains was found to restore activity.
Showing the most recent 10 out of 38 publications
