SR proteins mediate the coupling of Rous sarcoma virus RNA splicing and polyadenylation control

Rous sarcoma virus (RSV) requires large amounts of unspliced RNA for viral replication. Splicing and polyadenylation are coupled in the cells that RSV infects, which raises the question of how viral RNA is efficiently polyadenylated in the absence of splicing. Optimal RSV polyadenylation requires a far-upstream splicing control element, the negative regulator of splicing (NRS). This doctoral dissertation explores the link between NRS-mediated splicing inhibition and efficient polyadenylation. As a beginning to understanding the role of the NRS in RSV polyadenylation, the viral polyadenylation signal was characterized in vitro. The RSV substrate showed little or no polyadenylation in vitro, indicating that the polyadenylation signal is suboptimal. Polyadenylation sites often have identifiable upstream and downstream elements (USEs and DSEs) in close proximity to the conserved AAUAAA signal, however, the USEs and DSEs in RSV deviate from those found in efficiently used sites. Analysis of chimeric substrates composed of the USEs and DSEs of the well-characterized SV40 late polyadenylation signal in combination with elements from RSV indicated that USEs and DSEs from RSV are suboptimal but functional. Further analysis showed that the inactivity of the RSV polyadenylation site was at least in part due to poor CstF64 binding, consistent with poor polyadenylation factor binding sites as the basis for inefficient use of the RSV polyadenylation site and pointing to the importance of additional elements within RSV RNA in promoting 3' end formation.;Previous work has indicated that the NRS contributes to efficient viral polyadenylation. The NRS binds several factors involved in splicing (e.g., SR proteins and U1 snRNP) and is proposed to function as a pseudo-5' splice site that sequesters 3' splice sites in a non-productive splicing complex. The hypothesis was that this nonproductive splicing complex stimulates polyadenylation in the absence of splicing. In vitro, however, the NRS alone activated RSV polyadenylation. The polyadenylation stimulatory effect did not require the snRNP binding sites or a downstream 3' splice site, but SR proteins were required. Consistent with this, high-affinity binding sites for specific SR proteins were able to stimulate RSV polyadenylation both in vitro and in vivo. The high-affinity sites improved polyadenylation in proviral clones only when the NRS-3'splice site complex could form, but deletions that positioned the SR protein-binding sites closer to the polyadenylation site eliminated this requirement. These results suggest a previously undescribed role for SR proteins in RSV polyadenylation and a more general role for SR proteins in polyadenylation of cellular mRNAs.;The second part of this dissertation examined four intronic mutations and one exonic mutation in the pituitary homeobox transcription factor 2 (PITX2) gene that are associated with Axenfeld-Rieger syndrome (ARS). PITX2 isoform C minigenes were used to address the hypothesis that the mutations affect RNA splicing. The mutations analyzed included a G→T mutation within the AG 3' splice site junction of exon 4 (IVS4-1G→T), a G→C mutation at position +5 of the 5' splice site of exon 4 (IVS4+5G→C), an A→G substitution at position -11 of the 3' splice site of exon 5 (IVS5-11A→G), a C→→A mutation at position +62 of the 3' splice site of exon 4 (IVS4-62C→A), and a conservative mutation within exon 5 (IVS5+460C→T). While no splicing defects were seen with the IVS4-62C→A and IVS5+460C→T mutations, IVS4+5G→C showed 71% retention of the intron between exons 4 and 5, and poorly expressed protein. Wild-type protein levels were proportionally expressed from correctly spliced mRNA. IVS4-1G→T shifted splicing to a new AG and resulted in a severely truncated, poorly expressed protein. Similarly, IVS5-11A→G shifted splicing to a newly created upstream AG and resulted in generation of truncated protein. This is the first evidence to suggest aberrant RNA splicing as the mechanism underlying ARS in some patients, and the data suggest that the magnitude of the splicing defect may contribute to the variability of ARS disease phenotypes.