SEM mapping of sequence-specific protein-DNA interactions on long DNA molecules

초록

Direct visualization of protein-binding positions along individual DNA molecules provides direct readouts of binding location, occupancy, and heterogeneity beyond the reach of ensemble assays. However, existing single-molecule imaging methods face a persistent trade-off between spatial resolution, field of view, and throughput. Here, we establish an scanning electron microscopy (SEM)-based approach that combines contrast enhancement with sequence-defined labeling to image extended DNA molecules and resolve protein-bound regions along their contour. We validate this capability across distinct binding regimes, including sequence-defined streptavidin-fluorescent protein labels on biotinylated lambda DNA, mapping of densely bound dCas9 regions on plasmid DNA, and machine-learning-assisted detection of localized dCas9 binding on human genomic DNA (F1 = 0.97). SEM achieves mean positional offsets of 410 +/- 325 bp for nick-translated labels and 116 +/- 63 bp for dCas9-bound regions, similar to 3-fold improved over fluorescence imaging, while supporting large-area surveys of extended DNA molecules across multi-scale magnifications-a capability not accessible by transmission electron microscopy or atomic force microscopy. These results establish SEM as a scalable platform for simultaneous structural visualization and quantitative mapping of sequence-specific DNA-protein interactions along individual DNA molecules.[GRAPHICS]

키워드