Modelling membrane reshaping by staged polymerization of ESCRT-III filaments.

ESCRT-III filaments are composite cytoskeletal polymers that can constrict and cut cell membranes from the inside of the membrane neck. Membrane-bound ESCRT-III filaments undergo a series of dramatic composition and geometry changes in the presence of an ATP-consuming Vps4 enzyme, which causes stepwise changes in the membrane morphology. We set out to understand the physical mechanisms involved in translating the changes in ESCRT-III polymer composition into membrane deformation. We have built a coarse-grained model in which ESCRT-III polymers of different geometries and mechanical properties are allowed to copolymerise and bind to a deformable membrane. By modelling ATP-driven stepwise depolymerisation of specific polymers, we identify mechanical regimes in which changes in filament composition trigger the associated membrane transition from a flat to a buckled state, and then to a tubule state that eventually undergoes scission to release a small cargo-loaded vesicle. We then characterise how the location and kinetics of polymer loss affects the extent of membrane deformation and the efficiency of membrane neck scission. Our results identify the near-minimal mechanical conditions for the operation of shape-shifting composite polymers that sever membrane necks. Author summary: ESCRT-III proteins have the unique ability to cut membrane necks from within, which is needed for a vast number of cell remodelling events including the release of cargo-containing vesicles. ESCRT-III proteins exist in different forms, which can assemble into spiral and helical homopolymers of different curvatures, and they have been suggested to polymerize and depolymerize with each other in a staged manner to deform and cut membranes. We developed a computer model to explore the physical mechanisms behind vesicle budding driven by the staged assembly and disassembly of multiple elastic filaments. We identified rules that determine the outcomes of membrane remodelling, which depend on the relative physical features of the distinct filaments, the dynamics of their disassembly, and on the presence of cargo; thereby providing experimentally testable predictions. Our study provides new physical insights into the ESCRT-III-mediated vesicle budding process, at a single subunit level, and identifies the general design principles of nanomachines built from shapeshifting copolymers, which might also be realized in synthetic systems. [ABSTRACT FROM AUTHOR]