Convergent evolution in the supercoiling of prokaryotic flagellar filaments

Supercoiling of bacterial and archaeal flagellar filaments is required for motility, as the rotation of a straight or flexible filament will not generate thrust. The question of how a bacterial flagellar filament, composed of thousands of copies of identical subunits, can supercoil was addressed almost 50 years ago with two-state models for polymorphic switching in filaments comprised of 11 protofilaments, with different flagellar waveforms attributed to different mixtures of these two states. However, no direct visualization of mixed populations of protofilaments has ever been made. Archaeal flagellar filaments have no homology to bacterial ones and are instead homologs of bacterial Type IV pili. The bacterial model cannot simply be applied to the archaeal filaments, and no models exist for archaeal flagellar supercoiling. Great advances in cryo-electron microscopy (cryo-EM) make it now possible to directly see the basis for supercoiling, and here we show atomic structures of supercoiled bacterial and archaeal flagellar filaments. For the bacterial flagellar filament, we identified 11 protofilament conformations, rather than two. The archaeal flagellar filament is assembled from 10 protofilaments with 10 subunit conformations, with a seam on the inside of the curve. Our results reveal a process of convergent evolution where analogous functions have been achieved by completely different paths and mechanisms.