Degradation of mRNAs in budding yeast by novel decay pathways involving the REX exonucleases

We searched through gene expression databases for long 3’ UTR mRNAs that are targeted for degradation by the cytoplasmic mRNA degradation pathway, nonsense-mediated mRNA decay (NMD) (Sayani et al., 2008). This pathway, known for degrading mRNAs containing premature translation termination codons, also targets mRNAs with long 3’UTRs. We focused on the RTR1 locus which produces mRNAs with two prominent 3’UTR isoforms; the longer of which is targeted by the NMD system. We found from an RNA-seq data set that the RTR1 locus also produces a long non-coding RNA, known as a XUT, which is overlapping and antisense to the 3’UTR (van Dijk et al., 2011). We hypothesized that the XUT may form double-stranded RNA segments with the target mRNA 3’UTR and thus may stabilize the target mRNA. Our pursuit of this hypothesis was spurred by the result that early termination of the 3’UTR antisense XUT, which eliminates overlap between the RTR1 3’UTR and the antisense XUT, decreased the overall RTR1 mRNA steady-state abundance. We pursued a model for the regulation of target mRNAs by 3’UTR antisense XUTs whereby the XUT may compete with an RNA binding protein (RBP) that binds the 3’UTR and facilitates degradation of the mRNA. Ultimately, we invalidated this model by showing that the early termination of the antisense XUT does not result in a change in the half-life of RTR1 mRNAs compared to the wildtype strain upon transcription shut-off, and thus, the mutation must affect the mRNA abundance through changing the rate of transcription. In testing this model, however, we found an RBP site in the 3’UTR of RTR1 revealed by previously published genome-wide PAR-CLIP analysis of all RBP sites in the yeast genome (Freeberg et al., 2013). Deletion of this 3’UTR cis element led to an overall increase in abundance and stability of the RTR1 mRNAs. While deletion of known and well-characterized RBPs did not show an increase in RTR1 mRNA abundance, deletion of the RTR1 gene resulted in an increase in the abundance of an mRNA containing the RTR1 3’UTR that was also epistatic to the deletion of the cis element. This result led to the undertaking of a new study into the role of Rtr1p in mRNA decay that culminated in the manuscript presented in the second chapter of this dissertation. A key finding in this work was that mRNAs containing the Rtr1p cis element are targeted to a novel degradation pathway involving the 5’-3’ RNA helicase, Dhh1p, and REX exonucleases. Because degradation of mRNAs constitutes a previously unrecognized role for the REX exonucleases, we followed up on this study by analyzing the impact of deleting the REX exonucleases on the transcriptome by RNA-seq. We present this RNA-seq analysis in Chapter 3 and further explore other cellular functions and mechanisms of Rex2p and Rex3p by performing affinity purification and mass-spectrometry sequencing of the interacting proteins. Our results show that Rex2p and Rex3p interact with the histone acetylase complex, SAGA, and this interaction may be important in the regulation of ribosomal protein genes (RPGs). Our transcriptome analysis of the REX mutants also reveals a possible role in the quality control of splicing. Overall, this work defines new substrates for the REX exonucleases and establishes a new function for this family of exonucleases in the regulation and quality control of gene expression via mRNA degradation