Effective translation of the second cistron in two Drosophila dicistronic transcripts is determined by the absence of in-frame AUG codons in the first cistron.

The novel dicistronic transcript encoded by the Drosophila melanogaster stoned gene was recognized as being unusual in that the protein encoded by the first open reading frame, stoned-A (STNA), contains no internal methionine residues in a protein of 93 kDa. The dicistronic nature of the stoned locus and the lack of methionine residues in STNA is conserved across dipteran species. A second methionine-free cistron, encoding Snapin, was identified in Drosophila and also found to be dicistronic, the second open reading frame (ORF) encoding a methyltransferase. We have replaced the methyltransferase cistron with green fluorescent protein (GFP) and used this dicistronic construct to show that the GFP cistron is translated in Drosophila S2 cells. The insertion of in-frame AUG codons into the snapin ORF attenuates the translation of GFP, and the level of attenuation correlates with the number of inserted AUGs. Increasing the efficiency of translation-initiation of the Snapin cistron also attenuates the translation of GFP. This indicates that failure to initiate translation at the first AUG allows ribosomes to scan through the Snapin ORF and to initiate translation of the second cistron, unless new AUG codons are inserted. These data are used to interpret the expression of the stoned locus and in particular, to explain the altered stoned protein levels in the stoned-temperature-sensitive mutant allele, which replaces a lysine with a methionine codon early in the first, stonedA, cistron.