Abstract
Circular RNAs (circRNAs) are an abundant class of RNAs of largely unknown function. Eukaryotic circRNAs are intrinsically stable RNA molecules that are largely resistant to degradation by known RNA decay mechanisms and are produced co-transcriptionally by the spliceosome. Further, circRNAs accumulate with age in neuronal tissues, at least in nematodes, fruit flies, mice, and humans though it is unclear whether circRNAs have a direct role in aging. The main aim of this thesis is to understand the connection between circRNAs and aging. Briefly, we took two different approaches. In the first study, we profiled circRNA expression in fly heads across the lifespan. In the second study, we screened for functional circRNAs with special emphasis on aging-related functions. From the first study, we concluded that circRNAs could be used for markers of biological age. First, we aged wild-type male and female flies for 0, 10, 20, 30, 40, or 50 days old and sequenced total RNA-seq libraries generated from fly heads. We found that globally, circRNA levels increase linearly with chronological age, unlike mRNAs. Moreover, our data identify several circRNAs that strongly correlate with age strongly suggesting circRNAs are better age markers than mRNAs. We identified a few predominant classes of circRNA expression patterns: circRNAs that continuously increase with age, those that plateau at middle life, groups of circRNAs that show constant levels with age, and a group of circRNAs that decreases their abundance with age. The majority of circRNAs accumulate with age and among these, we found three subgroups that increase at different rates. Importantly, the mRNAs generated from the hosting genes, as well as general transcription from the locus (assessed by quantification of intronic sequences), remain largely constant. Moreover, although we observed age-dependent changes in splicing efficiency and alternative splicing, they can't account for the changes in circRNAs as the fly ages. Together, our data strongly suggest that stability is the main driver of circRNA accumulation. The intrinsic stability of circRNAs on the nervous system could potentially allow the use of circRNAs markers for the stress history of the animal. To understand if circRNAs have roles in development, behavior, and/or aging, we generated a collection of genetically encoded shRNAs to systematically knockdown (KD) the top 100 most abundant circRNAs in the fly head. KD of five circRNAs leads to lethality during developmental stages. KD of 21 circRNAs provoked defects in behavior as measured by a larval locomotion assay. To determine if circRNAs have a role in aging, we performed a survival assay with our collection. Interestingly, we found 7 circRNA KD that extended the lifespan and 17 that shortened the lifespan. Together, this shows that circRNAs have diverse roles in development and aging. We proceeded to study two circRNA, circKlg and circSide-III that consistently showed increased median and maximum lifespan upon repeat of our survival assay. We generated additional shRNAs to validate the effect of circKlg and circSide-III KD on the healthspan and lifespan assays. We confirmed the KD of these circRNAs by RT-qPCR and did not observe an effect on the levels of Klg or Side-III mRNA. Further, we confirmed that the extension of median lifespan did translate into improved health by 30 days old with a climbing assay. To understand the mechanism by which KD of these circRNAs improves the healthspan and lifespan, we performed 3’-RNAseq from heads of control and circRNA KD flies that were 3, 15, and 30 days old. Principle component analysis (PCA) revealed that KD of circKlg and circSide-III lead to a shift in the transcriptome to resemble that of younger flies. Indeed, we found that differentially expressed genes (DEGs) that change with age in the control are largely unchanged in circKlg and circSide-III KD samples at 30 vs 15 days. In addition, when comparing the control and circRNA KDs we found DEGs that have been previously associated with longevity. Among these DEGs are mthl8, Glut, tobi, and smp-30, which are known for their roles in insulin signaling, immunity, and metabolism. For further study, we generated circRNA overexpression (OE) strains and developed strategies to remove circRNA-producing exons using CRISPR/Cas9. Together, this dissertation revealed the main cause behind the age-dependent accumulation of circRNAs in the brain and uncovered new connections between circRNAs and aging.