Domain organization of the subunit of the Salmonella typhimurium flagellar hook.

The deduced amino acid sequences of the family of axial proteins of the bacterial flagellum possess N and C-terminal heptad repeats of hydrophobic amino acid residues, which suggests that these proteins all fold to form bundles of alpha-helices (e.g. coiled coils). There is evidence that flagellin, which is one of the axial proteins, has an axially oriented bundle of alpha-helices that gives rise to the inner, rod-shaped domains seen in electron density maps. We present evidence that a second member of the family, the hook subunit, also has such an axially oriented, rod-shaped domain. In three-dimensional reconstructions from electron micrographs of the helical hook of Salmonella typhimurium, the rod-shaped domain has a diameter of 18 A, which is that expected for a coiled coil. The corresponding domain in the flagellin subunit of the filament, however, is larger, having a diameter of 24 A suggesting a bundle of three or more alpha-helices. In addition to the rod-shaped domain, the hook has two other domains. At a radius of 55 A is the middle spheroidal domain about 25 A in diameter and at a radius of 75 A is the outer ellipsoidal domain about 20 A by 30 A by 40 A. The flagellin subunit also has a middle and an outer domain although they appear different from those of the hook. This is no doubt a result of the lack of any sequence similarity of the hook and flagellin subunits, apart from the N and C-terminal heptad repeats. Along the hook axis, there is a 25 A wide channel, which presumably serves in the export of hook and flagellin subunits in the assembly of the filament. There is a comparably sized channel in the filaments as deduced from electron micrographs. Thus, electron microscopy consistently finds a small channel, whereas in X-ray diffraction studies of the filament, the channel size appeared to be about 60 A. At a diameter of 60 A, the channel could pass the flagellin or hook subunit in its completely folded state, but if the channel is only 25 A in diameter, the subunit would have to be at least partially unfolded in order to pass through the channel.