Immunodeficiency Virus Type 1 mRNA 3' End Formation

At least two mechanisms have been implicated in regulating poly(A) site use in human immunodeficiency virus type 1 (BIVy1): inhibition ofbasal signals within 500 nucleotides (nt) ofthe cap site, leading to specific suppression of the 5 ' poly(A) site, and stimulation of basal signals by long terminal repeat U3 sequences, leading to specific activation of the 3 ' poly(A) site. We determined the relative contributions of these mechanisms in a HeLa cell transcription/processing reaction and by transient transfection analysis. In vitro, the efficiency of basal signals is equivalent close to (270 nt) and far from (1,080 nt) the promoter and is stimulated at least 30-fold in both positions by upstream U3 sequences. In vivo, U3 sequences also enhance processing at both positions. There are two additional effects when the poly(A) site is close to the cap site: at least a 15-fold reduction in total RNA levels and a 5-fold decrease in relative levels of RNA processed at the HIV-1 site in constructs containing U3. Both effects are overcome by insertion of upstream splicing signals in an orientation-dependent manner. Splicing appears to influence poly(A)+ RNA levels by two distinct mechanisms: stabilizing nuclear transcripts and directly stimulating 3 ' end formation. It is proposed that upstream elements play major roles in regulating poly(A) site choice and in controlling the subsequent fate ofpolyadenylated RNA. The impact ofthese findings on mechanisms of mRNA biogenesis in the HIV-1 provirus is discussed. During infection, the RNA genome of the retrovirus i