RT Journal
A1 Litterman, Adam J.
A1 Kageyama, Robin
A1 Le Tonqueze, Olivier
A1 Zhao, Wenxue
A1 Gagnon, John D.
A1 Goodarzi, Hani
A1 Erle, David J.
A1 Ansel, K. Mark
T1 A massively parallel 3′ UTR reporter assay reveals relationships between nucleotide content, sequence conservation, and mRNA destabilization
JF Genome Research 
JO Genome Research 
YR 2019 
FD June 01 
VO 29 
IS 6 
SP 896 
OP 906 
DO 10.1101/gr.242552.118 
UL http://genome.cshlp.org/content/29/6/896.abstract 
AB Compared to coding sequences, untranslated regions of the transcriptome are not well conserved, and functional annotation of these sequences is challenging. Global relationships between nucleotide composition of 3′ UTR sequences and their sequence conservation have been appreciated since mammalian genomes were first sequenced, but the functional relevance of these patterns remain unknown. We systematically measured the effect on gene expression of the sequences of more than 25,000 RNA-binding protein (RBP) binding sites in primary mouse T cells using a massively parallel reporter assay. GC-rich sequences were destabilizing of reporter mRNAs and come from more rapidly evolving regions of the genome. These sequences were more likely to be folded in vivo and contain a number of structural motifs that reduced accumulation of a heterologous reporter protein. Comparison of full-length 3′ UTR sequences across vertebrate phylogeny revealed that strictly conserved 3′ UTRs were GC-poor and enriched in genes associated with organismal development. In contrast, rapidly evolving 3′ UTRs tended to be GC-rich and derived from genes involved in metabolism and immune responses. Cell-essential genes had lower GC content in their 3′ UTRs, suggesting a connection between unstructured mRNA noncoding sequences and optimal protein production. By reducing gene expression, GC-rich RBP-occupied sequences act as a rapidly evolving substrate for gene regulatory interactions.