Versatility of Approximating Single-Particle Electron Microscopy Density Maps Using Pseudoatoms and Approximation-Accuracy Control

Figure 3

Exploring actual conformational changes by image analysis using normal modes of Gaussian-based (pseudoatomic) representation of EM density maps (analysis with HEMNMA). (a) EM density maps of three conformations of a DNA polymerase Pol α-B subunit complex of the eukaryotic primosome showing bending-unbending motion of the complex and the percentage of images assigned to each density map using ML3D, a method that assumes a small number of different coexisting conformations in the specimen, from [59]. (b) Mapping of images used in the analysis in [59] (producing the density maps shown in (a)) onto a low-dimensional distance space based on a flexible 2D-to-3D fitting between the images and a reference density map (the one with the highest percentage of assigned images in [59]) using normal modes of the pseudoatomic representation of the reference density map (in this space, images are represented with points and the distances between the points correspond to the differences between the corresponding conformations). (c) Displacement of the reference pseudoatomic structure along the trajectories identified in the densest regions of the distance space shown in (b), which indicates a bending-unbending motion, detected also by ML3D, and changes in the length of the flexible linker between the two lobes that could not be detected with ML3D (from left to right: the displacement along the yellow, red, and green trajectory shown in (b); the displacement is shown by providing three frames of an animation represented by red, yellow, and blue isosurfaces of the density maps into which the reference pseudoatomic structure was converted during the displacement; the arrows indicate the motions visualized by the frames). Reproduced with permission from [59] (a) and from [30] (b and c).

(a)

(b)

(c)