Almost every human pre-mRNA contains introns that must be removed through splicing before translated into proteins. Alternative splicing is a major contributor to proteomic diversity. Moreover, splicing mistakes are involved in at least 15% of human genetic diseases. Therefore, precise recognition of exon-intron boundaries is critical to gene expression. However, loosely conserved splice site sequences generally do not contain sufficient information to determine their usage in splicing. The main objectives of this dissertation were to identify the sequence contexts that suppress 5’ splice sites and elucidate the mechanisms by which they regulate splicing.
To understand how 5’ splice sites are recognized and regulated, we employed a sensitive splice site competition assay coupled with a randomization-selection (functional SELEX) strategy in vitro to identify sequence elements capable of silencing a strong (consensus) 5’ splice site; intronic and exonic elements were identified in parallel. Bioinformatics analysis revealed a striking correlation between the presence of these sequences and negatively regulated 5’ splice sites in the human genome. Surprisingly, however, none of the silencers can inhibit splicing of the strong 5’ splice sites in the absence of the distal competing site; instead, they only caused a subtle kinetic defect, although the defect became more striking when both splicing signals were weakened. Therefore, our results demonstrated that splicing silencers can switch 5’ splice site choice without blocking its usage, which suggested that alternative splicing regulation is very dynamic and can be modulated by subtle changes.
We subsequently focused on the mechanism by which a new class of intronic silencer functions to suppress splicing. We demonstrated that U1 snRNP bound to the 5’ splice site even though the site was not used. Interestingly, the inhibition is caused by a secondary structure looping around the silenced 5’ splice site. Further experiments showed that the suppressed 5’ splice site contacts U1 snRNP differently than a non-suppressed site. This altered interaction attacked the early recognition of the 5’ splice site by U4/U6.U5 tri-snRNP. Therefore, we concluded that secondary structure can silence 5’ splice site by affecting the ability of U1 snRNP to either recruit or interact with U4/U6.U5 tri-snRNP.