Investigating the role of supercoiling in DNA damage detection using all-atom implicit solvent molecular dynamics simulations

DNA damage is not uniformly distributed across the genome, and crystal structures show that repair proteins interact with flipped-out bases or bent DNA. When supercoiled, DNA often forms plectonemes: plait-like structures of interwound superhelices with sharply bent apices. To accommodate the bend, flipped bases, kinks, and bubbles form. Because plectonemes protrude from the dense genome, it has been hypothesised that damage may localise to the plectoneme tips to improve accessibility to repair proteins and provide bent or flipped structures for binding. Here, we use DNA minicircles to study the interplay between DNA sequence and structure in positioning plectonemes under both positive and negative supercoiling. We present all-atom, implicit-solvent molecular dynamics simulations of minicircles of two sizes and multiple sequences. The simulations' ability to capture sequence-dependent effects was validated experimentally using a novel smFRET assay and supported by data from the literature. Our results confirm implicit solvent molecular dynamics to be an accurate and accessible method to study the effects driving plectoneme pinning. We introduced all twelve mismatches into sequences with different plectoneme density landscapes, enabling us to examine both damage-driven pinning and the competition between sequence and damage for the tip. We found that the mismatches had different propensities for pinning the plectoneme, which further depended on the surrounding sequence context and direction of supercoiling. Multiple consecutive mismatches were introduced to study the relative influences of sequence versus bubble formation and elucidate a novel sequence-dependent threshold below which the bubble sequence dominates over bubble size in positioning the plectoneme. Overall, the conformational response of damaged DNA under torsional strain may explain why some regions of the genome are repaired more efficiently than others and provide one solution to the protein-DNA search problem.