Date Published: June 01, 2018
Publisher: International Union of Crystallography
Author(s): Florian M. Rossmann, Morgan Beeby.
The use of high-throughput in situ electron cryotomography and subtomogram averaging to study the the architecture and diversity of the bacterial flagellar motor is reviewed. Together with phenotypic analysis, this information can be used to better understand the evolution of molecular machines.
Understanding how molecular machines evolve is important for reasons ranging from antibiotic design to synthetic biology. The bacterial flagellar motor is an ideal model system for probing the principles of molecular evolution. The flagellar motor powers the rotation of bacterial flagella, which are helical proteinaceous filaments extending from the bacterial cell body that act as propellers for bacterial propulsion through liquid medium or swarming across surfaces (Jarrell & McBride, 2008 ▸). The flagellar motor is widespread, enabling cross-species comparison, and well characterized in terms of its components and function (Ohnishi et al., 1997 ▸; Imada et al., 2016 ▸; Khan et al., 1992 ▸; Thomas et al., 2006 ▸), opening the way for deeper questions on molecular evolution.
ECT is a technique that spans scales from structural biology to cell biology (Fig. 1 ▸). Cryo-EM involves the vitrification of a specimen by rapid freezing, preventing the formation of damaging ice crystals. This results in a sample suspended in a thin layer of vitreous ice that immobilizes the sample in a hydrated, close-to-native state that is stable for imaging in an electron cryo-microscope (Dubochet, 2012 ▸). Unlike conventional transmission electron microscopy, which requires chemical fixation and staining of the specimen, contrast in cryo-EM is derived from induced phase contrast of biological material in the microscope. Two major approaches are applied to obtain the molecular structure of large proteins or protein complexes using cryo-EM: SPA and ECT. SPA involves imaging many thousands of identical purified particles that are randomly oriented in vitreous ice. Given sufficient different orientations, it is possible to reconstruct a three-dimensional structure from these many two-dimensional images, in the process averaging out noise. Recent improvements in the developments of direct electron detectors have enabled full realization of this potential: the so-called ‘resolution revolution’ (Grigorieff, 2013 ▸; Ruskin et al., 2013 ▸; Kühlbrandt, 2014 ▸). SPA can now provide very high resolution structures of protein complexes (2–4 Å).
One of the most amenable systems to tomography, which has yielded considerable biological insights, is the bacterial flagellar motor (Fig. 3 ▸). The flagellar motor is a molecular rotary motor centred around a core cytoplasmic stator–rotor interaction that drives the rotation of a helical extracellular propeller through torque transmitted across the periplasm by an axial driveshaft (Chevance & Hughes, 2008 ▸). The stator component is a ring of inner membrane-embedded motor-protein ion channels immobilized by binding to the periplasmic peptidoglycan; ion flux drives interaction with the cytoplasmic rotor component called the C-ring. Torque applied to the C-ring is transmitted across the periplasm via an inner membrane-embedded MS-ring, which is connected to an axial driveshaft: the rod. To traverse the peptidoglycan layer and outer membrane, the rod passes through the P-ring and the L-ring, respectively, which act as bushings, to connect to an extracellular universal joint, called the hook, which finally transmits torque to the multimicrometre-long helical propeller: the flagellar filament. All axial structures are assembled by an integral flagellar type 3 secretion system (T3SS), with inner membrane components housed within the MS-ring together with a cytoplasmic ATPase. Until recently, much of what was known about the flagellar motor was derived from biochemical (Altegoer & Bange, 2015 ▸), genetic (Chevance & Hughes, 2008 ▸) and structural studies of purified components (Thomas et al., 2006 ▸), preventing mechanistic insights into the whole, assembled molecular machine, and furthermore the majority of studies focused exclusively on the motor from the model enteric bacteria Salmonella enterica and Escherichia coli, preventing comparative insights.
Although these advances in in vivo structure determination allowed the determination of the architecture of intact flagellar motors, the specific locations of proteins remained inferences from previous knowledge, leaving it difficult to decipher the locations of proteins within the in situ architecture.
The ability to locate specific proteins in a subtomogram average allowed a deeper investigation of the function of additional structures in diverse flagellar motors composed of accessory proteins. Towards understanding motor evolution, a recent study probed the selective benefits of motor diversity, finding that additional motor structures serve as a scaffold to assemble larger motors that output higher torque. Torque is a measurement of rotary force, and it follows that higher torques will enable propulsion through more viscous media that would otherwise immobilize motors that produce lower torque. Motor torque varies substantially between different bacteria and correlates with their swim speed and ability to propel themselves in highly viscous media such as gastrointestinal mucus. Three different bacteria that produce different torques were compared using electron cryotomography in an effort to rationalize different torque outputs: Salmonella with ∼1300 pN nm torque output, Vibrio with ∼2000 pN nm and C. jejuni with ∼3600 pN nm. Using the selective deletion strategy, ECT of V. fischeri and C. jejuni not only verified the location of the stator ring in the motor structure but also enabled determination of the number of stator complexes in the stator ring. Strikingly, the number of stator complexes, and their radius from the axis of rotation, differed in these higher torque-generating species from the ∼11 stator complexes in Salmonella positioned ∼20 nm from the axis of rotation (Reid et al., 2006 ▸; Leake et al., 2006 ▸): V. fischeri had 13 stator complexes located at a radius of 21.5 nm from the rod, while C. jejuni had 17 stator complexes located 26.5 nm from the rod, and the C-rings were also correspondingly wider in both. Indeed, the number and the location of the stator complexes, combined with previously measured stator-complex force exertion, was sufficient to accurately quantitatively predict the torque outputs of structurally diverse bacteria (Beeby et al., 2016 ▸).
Subsequent studies have sought to understand how these high-torque motors evolved. Naively, these motors are ‘irreducibly complex’ in that they are nonfunctional upon the deletion of individual components. Indeed, many of these components were first identified by screening for nonmotile motors resulting from mutations in genes that were not encoded in organisms with simpler motors.
Another intriguing aspect of flagellar evolution that ECT has provided insights into is the degeneration of an ancestral motor to form the hypodermic syringe-esque ‘injectisome’ complex used by many pathogens. Injectisomes, also referred to as type III secretion systems, are used by diverse pathogens to inject virulence factors into host cells to hijack their physiology. Phylogenetic studies indicate that injectisomes are degenerate flagella that have lost their stator complexes and have adapted their flagellar filament to become a short, rigid, hollow needle for virulence-factor delivery (Abby & Rocha, 2012 ▸; Fig. 5 ▸). ECT studies of injectisomes requires considerable sample optimization, as many pathogens (for example Salmonella, E. coli and Yersinia species) are too thick for high-resolution imaging. Successful studies have employed minicell systems (Hu et al., 2015 ▸; Kawamoto et al., 2013 ▸) or selected thin bacteria that are more amenable to imaging (Nans et al., 2015 ▸).
Future prospects for understanding bacterial flagellar motors are significant as the capabilities of ECT continue to mature. The deliverables from ECT are fairly straightforward: better data, and more of it. The impacts of these deliverables will be major advances in understanding flagellar assembly, mechanism and evolution.