The majority of higher eucaryotic mRNAs are processed at their 3'-ends by polyadenylation. Most mammalian polyadenylation elements consist of an AAUAAA consensus hexamer and a loosely defined U or GU-rich element between 5 and 70 nucleotides downstream of the AAUAAA. The simian virus 40 (SV40) late polyadenylation signal has these elements; however, we have found that it functions efficiently only in the presence of specific sequences between 16 and 46 nucleotides upstream of the AAUAAA. Hence the SV40 late signal represents a second, more complex class of polyadenylation signals which require an element both upstream and downstream of the AAUAAA. Additionally, the downstream element of the SV40 late signal appears to be more complex than most, since two different downstream elements have been defined in different experimental systems (14,77,78,90). Although the SV40 late signal is a widely used model for both in vitro and in vivo polyadenylation experiments, the complexity of elements and their interrelationship in the polyadenylation process has not been examined. A complex polyadenylation signal utilizing both upstream and downstream elements is not unique to the SV40 late genes, similar polyadenylation signals have been defined in several viral systems (10,17,18,73). We have partially mapped an upstream element in the human immunodeficiency virus (HIV-1) long terminal repeat which imparts efficiency to its polyadenylation signal and to HIV-SV40 chimeric polyadenylation signals. Given this complexity of polyadenylation signals, the primary goals of this proposal are to define the nature of the elements of the SV40 late and HIV polyadenylation signals, their spatial and functional relationships to one another, their involvement in factor binding, their involvement in the biochemistry of cleavage and polyadenylation, their possible role in transcriptional termination and RNA stability and their role in the control of gene expression. Ample evidence suggests that modulation of polyadenylation may be a significant level of gene expression control. Previous data (41) and our preliminary results suggest that the SV40 late polyadenylation signal is specifically activated, or more efficiently utilized, in actively growing cells. Such control of late RNA polyadenylation may be an important aspect of the temporal nature of SV40 gene expression. In this regard, SV40 T antigen, a strong activator of cell growth, may directly or indirectly affect the polyadenylation process. We propose to characterize the growth regulation of the SV40 late polyadenylation signal and to use defined mutations in the elements of the signal to determine which element mediates the control. It is hoped that this will lead to the identification of the cellular factor(s) involved in the process. Similar analysis of the HIV polyadenylation signal will indicate whether alterations in the efficiency of polyadenylation play a role in HIV gene expression control or latency. We will test the efficiency of the HIV signal under conditions known to activate HIV gene expression and in the presence
