Systematic generation and imaging of tandem repeats reveal multivalent base-pairing as a major determinant of RNA aggregation

Abstract

Expansion of short tandem repeats (STRs) in the human genome underlies over fifty genetic disorders. A common pathological feature of repeat RNAs is their propensity to aggregate in cells. While these RNA aggregates have been shown to cause toxicity by sequestering RNA-binding proteins, the molecular mechanism of repeat RNA aggregation remains unclear. Here we devised a generalizable method to efficiently generate long tandem repeat DNAs de novo and applied it to systematically determine the sequence features underlying RNA aggregation. Live-cell imaging of repeat RNAs indicated that aggregation was mainly driven by multivalent RNA-RNA interactions via either canonical or noncanonical base pairs. While multiple short runs of two consecutive base pairs were sufficient, longer runs of consecutive base pairs such as those formed by a neurodegeneration-associated hexanucleotide repeat further enhanced aggregation. In summary, our study provides a unifying model for the molecular basis of repeat RNA aggregation and a generalizable approach for identifying the sequence and structural determinants of repeat RNA properties.