UC Irvine

The process of pre-mRNA splicing is conserved from yeast to humans. Alternative splicing drives transcriptomic and proteomic diversity in higher eukaryotes through the creation of multiple unique mRNA transcripts from one gene encoded pre-mRNA. Alternative splicing is under combinatorial control meaning that many cis-acting sequence elements and trans-acting protein factors affect splice site usage. A thorough understanding of the factors that drive alternative splicing is essential in order to understand normal and aberrant splicing in the cell. Splicing regulatory proteins often drive tissue specific alternative splicing. The two most well characterized classes of splicing regulatory proteins are the SR proteins and hnRNPs. These two classes of proteins were shown to have a position-dependent activity on splicing efficiency when bound upstream or downstream of a regulated 5' splice site. To test the hypothesis that splicing regulators affect splicing efficiency by altering the structure of U1 snRNA, a pulldown approach was designed to isolate native U1 snRNPs bound to activated or repressed E-complexes. SHAPE chemistry and psoralen crosslinking were used to show that the core of U1 snRNA does not undergo structural rearrangements, however the 5' end of U1 snRNA is more base paired to the 5' splice site when splicing activators are bound. This study helped to elucidate the mechanism of position-dependent splicing activation and repression. Splice site strength and intron-exon architecture are the primary drivers of the innate ability of an exon to undergo alternative splicing. Previous work had shown that the proximity of the 5' and 3' splice site dictate splice site usage under intron definition. It was hypothesized that the proximity of the 5' and 3' splice sites across the exon influence splice site usage under exon definition. A data base of alternative splicing events comprised of exons with two competing 5' splice sites and one constant 3' splice site was created. Filtering of data to account for differences in MaxEnt scores (computational scores that predict 5' and 3' splice site usage) showed that a genome-wide exon centric proximity rule exists. Observations from the computational studies were confirmed with in vivo minigene analysis. This thesis expands upon the molecular mechanisms of splice site selection in the context of combinatorial alternative splicing. The findings presented herein will aid in the understanding of normal and aberrant alternative splicing programs.