Discovery of a pre-mRNA structural scaffold as a contributor to the mammalian splicing code [in vitro]

The specific recognition of splice signals at or near exon-intron junctions is not explained by their weak conservation and instead is postulated to require a multitude of features embedded in the pre-mRNA strand. We explored the possibility of three-dimensional structural scaffold of AdML – a model pre-mRNA substrate – guiding early spliceosomal components to the splice signal sequences. We find that mutations in the non-cognate splice signal sequences impede recruitment of early spliceosomal components due to disruption of the global structure of the pre-mRNA. We further find that the pre-mRNA segments potentially interacting with the early spliceosomal component U1 snRNP are distributed across the intron, that there is a spatial proximity of 5' and 3' splice sites within the pre-mRNA scaffold, and that an interplay exists between the structural scaffold and splicing regulatory elements in recruiting early spliceosomal components. These results suggest that early spliceosomal components can recognize a three-dimensional structural scaffold beyond the short splice signal sequences, and that in our model pre-mRNA, this scaffold is formed across the intron involving the major splice signals. This provides a conceptual basis to analyze the contribution of recognizable three-dimensional structural scaffolds to the splicing code across the mammalian transcriptome. Overall design: Estimation of SHAPE reactivity by SHAPE-MaP of Adenovirus 2 major late transcript IVS1 and its mutants