Abstract
Selection of a translation initiation site is thought to be determined by relative proximity to the 5' end and sequence context of a potential initiator codon. These guidelines seem insufficient to explain translation of the Drosophila E74A mRNA, whose 5' untranslated region is exceptionally long (1.8 kb) and contains many AUG triplets preceding the long open reading frame. In an effort to understand how the appropriate initiator codon is chosen, we have undertaken a study of E74A translation in transfected Drosophila cells. The results show that translation of the E74A protein utilizes at least three initiator codons: two minor forms of the protein are initiated at a CUG and an AUG, while the most abundant form initiates at a CUG. This main initiator CUG appears to be in a good context; however, it lies downstream of 17 AUG and 24 other CUG codons, several of which are also in good contexts. Unexpected results were obtained from sequence perturbations upstream and downstream of the main CUG initiator. Creating an AUG with a good context 72 bases 5' to the main CUG has only a modest inhibitory effect on initiation frequency at that CUG. Replacing sequences 44 bases 3' to the main CUG has an inhibitory effect on its use as an initiator as well as on the CUG 72 bases further upstream. These results indicate that factors other than context and relative proximity to the 5' end must be involved in initiator codon selection and may include elements such as secondary or tertiary structure of the RNA.
