We have obtained a 3D reconstruction of intact microtubules, using cryoelectron microscopy and image processing, at a resolution of about 8 A, sufficient to resolve much of the secondary structure. The level of detail in the map allows docking of the tubulin structure previously determined by electron crystallography, with very strong constraints, providing several important insights not previously available through docking tubulin into lower-resolution maps. This work provides an improved picture of the interactions between adjacent protofilaments, which are responsible for microtubule stability, and also suggests that some structural features are different in microtubules from those in the zinc sheets with which the tubulin structure was determined.

No abstract

Three-dimensional reconstructions from electron cryomicrographs of the rotor of the flagellar motor reveal that the symmetry of individual M rings varies from 24-fold to 26-fold while that of the C rings, containing the two motor/switch proteins FliM and FliN, varies from 32-fold to 36-fold, with no apparent correlation between the symmetries of the two rings. Results from other studies provided evidence that, in addition to the transmembrane protein FliF, at least some part of the third motor/switch protein, FliG, contributes to a thickening on the face of the M ring, but there was no evidence as to whether or not any portion of FliG also contributes to the C ring. Of the four morphological features in the cross section of the C ring, the feature closest to the M ring is not present with the rotational symmetry of the rest of the C ring, but instead it has the symmetry of the M ring. We suggest that this inner feature arises from a domain of FliG. We present a hypothetical docking in which the C-terminal motor domain of FliG lies in the C ring, where it can interact intimately with FliM.The bacterial flagellum of Salmonella enterica serovar Typhimurium has a reversible rotary motor powered by the proton gradient across the cell's plasma membrane. The flagellar filament, with its corkscrew shape, converts the motor's torque to thrust. Counterclockwise (CCW) rotation of the filament by the motor pushes the cell through the liquid medium, whereas a brief intervening burst of clockwise (CW) rotation causes the cell to tumble. Following the CW burst, the motor resumes CCW rotation, and the cell swims off in a new direction. Approximately 40 proteins are involved in the regulation, assembly, and operation of the flagellum. At least 24 of them are components of the completed flagellum. Of the 24, only 5, i.e., MotA, MotB, FliG, FliM, and FliN, appear to be involved in torque generation (for reviews, see references 1, 2, and 25). MotA and MotB (8, 37) form a proton channel through the plasma membrane (3, 38); they are assumed to form the stator (7, 18). The remaining three are cytoplasmic proteins that form the switch complex (49) and are assumed to be part of the rotor. FliG appears to be most directly involved in torque generation (14,22), and the C-terminal domain of FliG contains key charged residues that interact with charged residues in MotA (53). The N-terminal portion of FliG interacts with FliF, the transmembrane component of the rotor (20, 31), which functions as a mechanical mount for the rotor and couples the motor to the rod or drive shaft. FliM is involved in switching between CCW and CW rotation (34) and binds phospho-CheY (48), tipping the motor's bias toward CW rotation. FliN appears to play a smaller role in rotation and switching (14, 22) but has an important role in flagellar export (46).The gross structural features of the flagellum are a filamentous axial structure and a set of coaxial rings (Fig. 1). MotA and MotB, the putative stator, form a ring of about 10 studs in the plasma membrane (17, 18)...

A transmission electron micrograph is essentially a projection of the specimen in the direction of view. In order to reconstruct a three-dimensional image of the specimen, it is necessary to be able to combine data from a number of different views. A formal solution of this problem is given in terms of Fourier transforms. Its realization requires data reduction and interpolation. The final solution is given by a least squares approach, which also indicates how many views must be included to give a valid reconstruction of a given particle to a given degree of resolu­tion. Interpolation procedures of varying power are given, to be employed according to the economy with which the available data must be used. An alternative procedure is described for direct reconstruction without the use of Fourier transforms, but it is shown to be in general less practicable than the Fourier approach.