Genetics and Regulatory Impact of Alternative Polyadenylation in Human B-Lymphoblastoid Cells

Abstract

Gene expression varies widely between individuals of a population, and regulatory change can underlie phenotypes of evolutionary and biomedical relevance. A key question in the field is how DNA sequence variants impact gene expression, with most mechanistic studies to date focused on the effects of genetic change on regulatory regions upstream of protein-coding sequence. By contrast, the role of RNA 3′-end processing in regulatory variation remains largely unknown, owing in part to the challenge of identifying functional elements in 3′ untranslated regions. In this work, we conducted a genomic survey of transcript ends in lymphoblastoid cells from genetically distinct human individuals. Our analysis mapped the cis-regulatory architecture of 3′ gene ends, finding that transcript end positions did not fall randomly in untranslated regions, but rather preferentially flanked the locations of 3′ regulatory elements, including miRNA sites. The usage of these transcript length forms and motifs varied across human individuals, and polymorphisms in polyadenylation signals and other 3′ motifs were significant predictors of expression levels of the genes in which they lay. Independent single-gene experiments confirmed the effects of polyadenylation variants on steady-state expression of their respective genes, and validated the regulatory function of 3′ cis-regulatory sequence elements that mediated expression of these distinct RNA length forms. Focusing on the immune regulator IRF5, we established the effect of natural variation in RNA 3′-end processing on regulatory response to antigen stimulation. Our results underscore the importance of two mechanisms at play in the genetics of 3′-end variation: the usage of distinct 3′-end processing signals and the effects of 3′ sequence elements that determine transcript fate. Our findings suggest that the strategy of integrating observed 3′-end positions with inferred 3′ regulatory motifs will prove to be a critical tool in continued efforts to interpret human genome variation.

Tandem repeat variation in human and great ape populations and its impact on gene expression divergence

“…Moreover, an association with expression divergence held also for repeats in other genic regions. The strongest of them was evident for 3 ′ UTRs, consistent with their known role in gene regulation (Yoon et al 2012). In addition, repeats in first introns were associated with greater expression divergence than repeats in other introns.…”

The roles of RNA processing in translating genotype to phenotype

“…Loss or gain of miRNA-mediated regulation of mRNAs has been proposed to link changes in APA with altered gene output 63 . Consistent with this proposal, regulatory cis -acting elements, including miRNA-binding sites, are enriched between proximal and distal polyadenylation sites 45 . In a GWAS using HapMap data across multiple populations, SNVs that shortened the mRNA 3′ UTR and caused a loss of miRNA-binding target sites were associated with increased gene expression 63 .…”

“…A notable example involves susceptibility to facioscapulohumeral muscular dystrophy (FSHD), in which individual differences in the polyadenylation hexanucleotide sequence of the last exon of double homeobox 4 ( DUX4 ) are determinative for disease (BOX 2). Analysis of allele-specific expression in LCLs from six individuals revealed that variants within the hexanucleotide are significantly more likely to have an effect on gene expression than variants in the rest of the 3′ UTR, and that single-nucleotide substitutions are sufficient to change the use of distal versus proximal poly adenylation sites 45 . A detailed analysis of blood samples from 94 individuals to find SNVs in linkage disequilibrium with a change in polyadenylation site usage identified 5 top candidates, all of which were in the AATAAA motif.…”

“…This SNV shifts polyadenylation from the proximal to the distal site and thereby produces a comparatively unstable IRF5 transcript with a longer 3′ UTR that contains a destabilizing AU-rich element (ARE). This results in reduced levels of IRF5 protein, the consequences of which are proposed to be relevant to increased SLE risk 45,47,48 . In another lupus-associated example, the rs6598 SNV in the proximal polyadenylation hexanucleotide site of GTPase IMAP family member 5 ( GIMAP5 ) marks the risk haplotype for SLE and shifts polyadenylation to alternative sites 49 .…”

Ubiquitously transcribed genes use alternative polyadenylation to achieve tissue-specific expression

“…Across different tissue samples, we observed that about twothirds of multi-UTR genes show significant differences in usage of the first pA sites and thus can potentially escape post-transcriptional regulation. There is a higher frequency of noncanonical hexamer signals at the proximal pA sites of multi-UTR genes than in single-UTR genes, and these noncanonical signals were shown to be weaker pA sites (Wickens and Stephenson 1984;Yoon et al 2012). However, the 39 UTRs of pAM genes are more highly conserved, their distal UTR regions are enriched for miRNA-binding sites, and the regions surrounding their proximal pA signals are more conserved relative to NpAM genes.…”