Abstract

The complexity of cellular DNA is a consequence of its innate flexibility, compaction in the nucleus, and manipulation of its structure by DNA processing enzymes, resulting in a vast conformational landscape. Within this landscape DNA adopts intricate structures, conformations and topologies and is frequently maintained under superhelical stress, in an under or over- wound state. The effect of superhelical stress on the structure of DNA is challenging to quantify, because of the length scale at which this occurs, 100x less than the wavelength of light. We develop new microscopy and image analysis tools to determine how the structure of individual DNA molecules varies under superhelical stress, and how this affects its interactions with DNA -binding proteins and therapeutic agents. High-resolution atomic force microscopy (AFM) is unique in its ability to visualise DNA structure and interactions in liquid with sub-molecular resolution without the need for labelling or averaging, enabling routine visualisation of highly flexible and dynamic molecules, such as DNA , with sub-molecular resolution [1]. To quantify the structural and conformational variability of these molecules, we have developed TopoStats, a high-throughput, open-source Python package which enables us to measure the physical properties of DNA molecules from AFM images, from contour length, through curvature, writhe and even twist as they ‘explore’ their complex conformational space [2]. We combine these measurements with atomistic molecular dynamics simulations to demonstrate that DNA under superhelical stress is far richer in structure than can be observed in short linear sequences, containing kinks and defects at the atomistic level [3]. We build on this work to determine how DNA supercoiling affects the interactions of DNA binding proteins. For example, we determine the structure of the protein NDP52 and show that it binds specifically and with high affinity to double-stranded DNA changing its local conformation [4].  [1] Pyne, A et al. Small. 12, 1053 (2014)[2] Beton, JG et al. Methods 193, 68-79 (2021)[3] Pyne, ALB , Noy A et al. Nature Communications. 12, 1053 (2021)[4] Dos Santos, A et al. Nature Communications 14, 2855 (2023)