Repeat-associated non-AUG (RAN) Translation in GC-Rich Repeats of C9ALS/FTD and FXTAS

Abstract

Repetitive elements comprise over half of the human genome. Expansion of a subtype within these repetitive elements known as short tandem repeats (STRs) causes over fifty human disorders. C9orf72-associated amyotrophic lateral sclerosis and frontotemporal dementia (C9ALS/FTD) and Fragile X-associated tremor/ataxia syndrome (FXTAS) represent two instances where an STR expansion results in neurodegenerative disease. C9ALS/FTD results from a GGGGCC (G4C2) hexanucleotide repeat expansion within the first intron of C9orf72 while FXTAS is caused by CGG repeat expansion in the 5’UTR of FMR1. These repeat-containing RNAs elicit toxicity at least in part by triggering repeat-associated non-AUG (RAN) translation. RAN translation is a non-canonical process where repetitive regions within mRNAs initiate translation in the absence of a canonical AUG start codon. RAN translation can happen in multiple reading frames and can occur in transcriptional contexts (5’UTRs, introns, 3’UTRs, and antisense RNAs) that do not typically support mRNA translation. The protein products generated by RAN translation with these repeat expansions and are thought to contribute to disease pathogenesis. However, the mechanisms by which RAN translation occurs remain unclear. The goal of this thesis is to better define how RAN translation works and identify selective modifiers of RAN translation. To accomplish this goal, I have taken three approaches. First, I analyzed DHX36 which directly interacts with repeat mRNA and influences its translational efficiency. G4C2 and CGG repeats form stable RNA secondary structures such as G-quadruplex and/or hairpin structures. DEAH-Box Helicase 36 (DHX36) plays an active role in RNA and DNA G-quadruplex resolution. Therefore, I evaluated whether altering the expression of DHX36 impacts repeat transcription and RAN translation. These studies revealed that DHX36 depletion suppresses RAN translation from reporter constructs in a repeat length-dependent manner. Second, performing a candidate-based target screen, I identified factors from the mRNA and protein surveillance pathways that affect the efficiency of translation through GC-rich repeat sequences. I identified three key factors - NEMF, LTN1, and ANKZF1 - from the ribosome-associated quality control (RQC) pathway that impact the generation of RAN products. Depletion of these RQC factors increases the accumulation of RAN proteins while overexpression of RQC factors decreases RAN protein accumulation. To assess this further, I used a dual tagging system to detect partially generated products from G4C2 and CGG repeats. These findings suggest that RQC factors play a critical role in both suppressing RAN translation and in maintaining protein and translational homeostasis in the face of toxic repeats. Third, we completed and analyzed a genome-wide high-throughput small-interfering RNA (siRNA) screen for selective RAN translational modifiers. From gene ontology analysis, we surprisingly found that multiple siRNAs target genes encoding 20S proteasome subunits act as selective suppressors of RAN translation while a separate group of siRNAs targeting genes from the eIF5A hypusination pathway act as selective enhancers of RAN translation. These two are interesting novel pathways emerging as areas for future studies. Taken together, these studies provide insight into the factors and pathways required for RAN translation at two GC-rich nucleotide repeat expansions. These results suggest that repeat RNA structure, translational elongation, and protein homeostasis play active roles in regulating RAN translation. By coupling these newly identified pathways to assays of repeat-associated toxicity in mammalian cell-based reporter systems, Drosophila models of disease, and patient-derived human neurons will reveal novel targets for future therapeutic development.