A novel regulatory pathway recognizes and degrades transcripts with long 3' untranslated regions

Abstract

Quantitative control of gene expression is fundamental to cellular function, and post-transcriptional regulation is a consequential, additional layer of control over protein output. Within mammalian mRNAs, the 3' untranslated region (3'UTR) acts as a hub of regulatory control, typically mediated by regulatory sequence elements within the 3'UTR. We have found that expression from transcripts with long 3'UTRs is strongly repressed compared to those with short 3'UTRs, and this phenomenon appears to be independent of sequence elements within the 3'UTR. The repression increases with 3'UTR length and is substantial; reporters with 2,000 nucleotide long 3'UTRs are repressed >10-fold compared to 400 nucleotide 3'UTRs. Conversely, increasing the length of the coding region has no effect on expression, demonstrating that it is 3'UTR length, not transcript length, that elicits repression. Reporters with different length 3'UTRs show no difference in transcription rates nor translation efficiency but have clear differences in mRNA half-lives and nucleocytoplasmic distribution, indicating repression of long-3'UTR transcripts is mediated by accelerated RNA decay. However, this transcript degradation does not involve the nonsense-mediated decay (NMD) pathway nor other canonical RNA decay pathways. In order to identify trans factors involved in 3'UTR-length-mediated decay, we quantified the differences in proteins bound to reporters with long and short 3'UTRs, and identified over 100 differentially associated factors. Knockdowns of two differentially associating proteins, the mRNA nuclear export factors DDX39B and ZC3H11A, indicate they play a role in repression of mRNAs with long 3'UTRs. This study establishes a novel length-dependent regulatory feature of 3'UTRs, which potentially regulates many genes.