U-rich elements drive pervasive cryptic splicing in 3’ UTR massively parallel reporter assays

Abstract

ABSTRACTNon-coding RNA sequences play essential roles in orchestrating gene expression. However, the sequence codes and mechanisms underpinning post-transcriptional regulation remain incompletely understood. Here, we revisit the finding from a prior massively parallel reporter assay (MPRA) that AU-rich (U-rich) elements in 3’ untranslated regions (3’ UTRs) can drive upregulation or downregulation of mRNA expression depending on 3’ UTR context. We unexpectedly discover that this variable regulation arises from widespread cryptic splicing, predominately from an unannotated splice donor in the coding sequence of GFP to diverse acceptor sites in reporter 3’ UTRs. Splicing is activated by U-rich sequences, which function as potent position-dependent regulators of 5’ and 3’ splice site choice and overall splicing efficiency. Splicing has diverse impacts on reporter expression, causing both increases and decreases in reporter expression via multiple mechanisms. We further provide evidence that cryptic splicing impacts between 10 to 50% of measurements made by other published 3’ UTR MPRAs. Overall, our work emphasizes U-rich sequences as principal drivers of splicing and provides strategies to minimize cryptic splicing artifacts in reporter assays.