Article Figures & Data
Figures
- FIG 1
The coarse-grained model. The membrane (yellow) was modeled as a single layer of beads, originally forming a cylinder. Each FtsZ filament (cyan) was modeled as a chain of beads in the filament sliding model (A) or as a chain of cubes in the filament bending model (B). The filament was connected to the membrane via a set of linkers (orange), each composed of two springs (shown as rods) joined at a bead that represents an FtsA or a ZipA protein. The cell wall (pink) was modeled as a grid of beads, originally forming a cylinder surrounding the membrane. Note that the same colors are used for all following figures unless stated otherwise.
- FIG 2
Simulation results of the filament sliding model. Italic font indicates simulation times. (A) Axial view of the initial system in which the FtsZ filaments overlapped to form complete rings. (B) Implementing a Lennard-Jones potential interaction between the filaments resulted in formation of a bundle of rings. (C and D) Axial (C) and side (D) views of the filament bundle in panel B. (E) Implementing depolymerization with a rate above the critical rate quickly resulted in loss of ring integrity. (F) With a depolymerization rate slower than rc, deep constriction occurred and rings broke much more slowly. (G) Removing depolymerization and implementing treadmilling also resulted in a deep constriction. (H) Side view of panel G. (I and J) Time series of the constriction rate with the simulation conditions were the same as those in panels F and G, respectively. The average was calculated over four simulations. Error bars indicate standard deviation.
- FIG 3
Simulation results of the filament bending model. Italic font indicates simulation times. (A) The initial system with FtsZ filaments in the GTP state running circumferentially. (B) After the filaments were switched to the GDP state, they did not constrict the membrane but, instead, rolled to bend on the plane of the membrane. (C) Implementing treadmilling did not prevent rolling, but filaments treadmilled in circles on the plane of the membrane. (D) Aligning membrane beads (yellow) connected to the same filament to the circumferential direction did not prevent filament rolling but only stretched the filament circles into a more elliptical shape. Note that, except for beads connected to filaments, the rest of the membrane (visualized as a surface) is shown with low opacity. (E) The presence of rigid linkers that prevented rolling allowed filaments to exert force on the membrane. Treadmilling resulted in uniform membrane constriction. (F) Time series of the constriction rate with the simulation conditions the same as in panel E. The average was calculated over four simulations. Error bars indicate standard deviation.
- FIG 4
(A and B) Traditional bending (A) versus reverse bending (B). (Top) Schematic of how filament bending has traditionally been thought to occur versus bending in the opposite direction to the membrane (“reverse” bending). In both cases, the C termini face the membrane. Arrows indicate the directions of forces exerted on the membrane by the filament. (Middle) Simulation results after 100 s of systems in which the membrane initially had a diameter of 500 nm and the preferred curvature of the filament was 1/250 nm−1 (twice that of the initial membrane). Traditional bending did not result in constriction, but reverse bending did. (Bottom) Side views.
- FIG 5
Single-filament constriction. (A) Axial view of the initial single-filament system. (B) Side view of the initial system. (C) Side view of the system after 100 s of simulation time showing that even a single FtsZ filament can constrict the membrane through reverse bending and treadmilling. (D) Cell wall constriction rate averaged over four simulations. Error bars indicate standard deviation.
Tables
- TABLE 1
Summary of key parameters and simulation results of each model
Supplemental material
Files in this Data Supplement:
- Supplemental file 1 -
Fig. S1 to S17
PDF, 1377K
- Supplemental file 1 -
