Transcription of a cloned rainbow trout protamine gene is accurately initiated following transfection into HeLa cells but the majority of the transcripts fail to polyadenylate at the correct site.

The expression of a cloned trout protamine gene transfected into mammalian cells in culture has been studied. This small intronless gene has a consensus TATA-box, a classical AATAAA sequence and the cap and polyadenylation sites are separated by only 228 base pairs (Gregory et al., ref 10). When 1kb of cloned trout genomic DNA containing this sequence was introduced into HeLa cells, S1-mapping showed that transcripts of the protamine gene were accurately initiated at the in vivo cap site but were not polyadenylated at the authentic 3'-site. Replacement of the 3'-end of the protamine transcription unit with a fragment of SV40 containing the small-t intron and early polyadenylation site resulted in only a modest increase in transcript levels over the wild-type gene in HeLa cells. However, transcripts of a fusion gene in which the 5'-end of the protamine gene was replaced by the SV40 early promoter were present at extremely low levels in transfected COS cells. The data are discussed in the context of the involvement of RNA processing events in the stabilisation of eukaryotic gene transcripts.