The G-rich Repeats in and Loci Are Hotspots for Local Unpairing of DNA

Pathological mutations involving noncoding microsatellite repeats are typically located near promoters in CpG islands and are coupled with extensive repeat instability when sufficiently long. What causes these regions to be prone to repeat instability is not fully understood. There is a general consensus that instability results from the induction of unusual structures in the DNA by the repeats as a consequence of mispairing between complementary strands. In addition, there is some evidence that repeat instability is mediated by RNA transcription through the formation of three-stranded nucleic structures composed of persistent DNA:RNA hybrids, concomitant with single-strand DNA displacements (R-loops). Using human embryonic stem cells with wild-type and repeat expanded alleles in the (CGGs) and (GGGGCCs) genes, we show that these loci constitute preferential sites (hotspots) for DNA unpairing. When R-loops are formed, DNA unpairing is more extensive, and is coupled with the interruptions of double-strand structures by the nontranscribing (G-rich) DNA strand. These interruptions are likely to reflect unusual structures in the DNA that drive repeat instability when the G-rich repeats considerably expand. Further, we demonstrate that when the CGGs in are hyper-methylated and transcriptionally inactive, local DNA unpairing is abolished. Our study thus takes one more step toward the identification of dynamic, unconventional DNA structures across the G-rich repeats at and disease-associated loci.