The structure of F-actin. Results of global searches using data from electron microscopy and X-ray crystallography

The structure of F-actin was investigated by fitting the crystallographically determined actin monomer structure to F-actin electron microscopy data sets obtained by a variety of methods. A reciprocal-space global search procedure was applied to non-equatorial reciprocal-space amplitudes and phases of the microscopy data to locate minima. Fits were performed over a range of cross-sectional radii-of-gyration encompassing values obtained from X-ray solution scattering measurements. Five data sets from four laboratories were investigated: one from frozen-hydrated single filaments, three from negatively stained single filaments, and one from negatively stained single-layer paracrystals. In this last case the paracrystal data were straightened to improve resolution. The best fits nearly unanimously favored a monomer orientation having long-pitch connectivity that was close to that obtained by fitting X-ray fiber diffraction patterns. In certain cases where the resolving power was low, competitive fits were obtained with a quite different orientation, one having only protomer connectivity along the genetic helix. Using a running 10-residue deletion from the monomer in the Holmes-type orientation, subtle differences between the monomer structure and the protomer structure in F-actin could be detected. In particular, differences in the "DNase loop" (residues 41 to 50) and the "hydrophobic loop" (residues 264 to 273) were seen in single-filament data. In addition, a perturbation of the structure was seen in a region near residues 81 to 90. A rearrangement within the protomer structure reported in the literature did not produce fits as good as those obtained when using the undistorted monomer crystallographic structure without 10-residue deletions. These results, taken as a whole, provide strong support for a structure of Mg(2+)-ADP F-actin similar to that originally suggested by Holmes et al. but with alterations of the hydrophobic loop, the DNase loop and in the region near residues 81 to 90. These latter two regions have been proposed as secondary binding sites for myosin heads. The available evidence from electron microscopy and from other sources suggests that residues 40 to 49 are disordered in Mg(2+)-ADP F-actin.