A Polymerization-Associated Structural Switch in FtsZ That Enables Treadmilling of Model Filaments

Abstract

Bacterial cell division in many organisms involves a constricting cytokinetic ring that is orchestrated by the tubulin-like protein FtsZ. FtsZ forms dynamic filaments close to the membrane at the site of division that have recently been shown to treadmill around the division ring, guiding septal wall synthesis. Here, using X-ray crystallography of Staphylococcus aureus FtsZ (SaFtsZ), we reveal how an FtsZ can adopt two functionally distinct conformations, open and closed. The open form is found in SaFtsZ filaments formed in crystals and also in soluble filaments of Escherichia coli FtsZ as deduced by electron cryomicroscopy. The closed form is found within several crystal forms of two nonpolymerizing SaFtsZ mutants and corresponds to many previous FtsZ structures from other organisms. We argue that FtsZ's conformational switch is polymerization-associated, driven by the formation of the longitudinal intersubunit interfaces along the filament. We show that such a switch provides explanations for both how treadmilling may occur within a single-stranded filament and why filament assembly is cooperative.IMPORTANCE The FtsZ protein is a key molecule during bacterial cell division. FtsZ forms filaments that organize cell membrane constriction, as well as remodeling of the cell wall, to divide cells. FtsZ functions through nucleotide-driven filament dynamics that are poorly understood at the molecular level. In particular, mechanisms for cooperative assembly (nonlinear dependency on concentration) and treadmilling (preferential growth at one filament end and loss at the other) have remained elusive. Here, we show that most likely all FtsZ proteins have two distinct conformations, a "closed" form in monomeric FtsZ and an "open" form in filaments. The conformational switch that occurs upon polymerization explains cooperativity and, in concert with polymerization-dependent nucleotide hydrolysis, efficient treadmilling of FtsZ polymers.

FIG 1

FL T66W and F138A mutant SaFtsZ proteins have compromised polymerization and GTPase activities. (A) Polymerization of FtsZ proteins at 10 μM was assayed by sedimentation in the presence of GTP and GMPCPP (CPP) with and without the FtsZ inhibitor PC190723 (PC). Pelleted (P) and soluble (S) protein samples were subjected to SDS-PAGE in the same gel lane with a delay. The percentage of pelleted protein was estimated from integration of band intensities. (B) GTPase activity of FtsZ proteins at 10 and 20 μM in the presence of GTP or GMPCPP. (C) Polymerization of FtsZ proteins in the presence of GTP and GMPCPP with and without the FtsZ inhibitor PC190723 (PC) was assessed by negative-stain electron microscopy. All images are at the same magnification (scale bar, 200 nm). WT, wild type.

FIG 2

Nucleotide-bound SaFtsZ crystal structures group into two conformations, open and closed. (A) The five SaFtsZ (TR to residues 12 to 316) structures determined here and PDB ID 3VOA are shown in cartoon representations with nucleotides as sticks colored by element. The 2-methyl-2,4-pentanediol (MPD) molecule in 1FOf is shown as green sticks. The structures are colored according to conformation. Closed structures are red, with the N-terminal GTP-binding domain in light red and the C-terminal GTPase activation domain in dark red, and the central helix, H-7, is highlighted is yellow. Open structures are blue, with the NTD in light blue and the CTD in dark blue, and the central helix H-7 is highlighted in orange. All structures shown in the same orientation, after alignment with the NTD of 3VOA (residues 13 to 165). The 4FCs domain-swapped pseudomonomer is formed of two polypeptides. Note the different positions of the CTD in the two sets of structures. The position of the PC190723 binding pocket is indicated on the 3VOA molecule. (B) Superposition of the six structures in panel A shown in Cα ribbon representation after alignment as for panel A, with the same color scheme. Nucleotides are shown as sticks. Side chains of residues F138 and T66 of the wild-type structure are shown as spheres, and noncarbon atoms are colored by element (left). The same view as in panel A (right) with molecules rotated 90° as indicated. The axis of interdomain rotation is indicated by the circled dot and the curved arrows. (C, D) Census of available nucleotide-bound SaFtsZ structures. (C) Bar graph indicating that DynDom, model-free assessment of dynamic protein domains, essentially produces two groups when SaFtsZ structures are compared to PDB ID 3VOA (no interdomain rotation or an ~27° shift around the axis in panel B, right). (D) Table with information about nucleotide-bound SaFtsZ structures. The horizontal line separates the structures determined here (top) from previous structures in the PDB. PC, PC190723; reso, resolution.

FIG 3

All FtsZ structures can be grouped into two conformations, open and closed. (A) All previous FtsZ structures were obtained from the PDB as listed in panel C. Chain A was extracted from each downloaded structure and the five structures determined here and aligned with the NTD (residues 12 to 176) of 3VOA by using the PyMOL align command, which matches residues via sequence and then minimizes the RMSD with five cycles of outlier rejection, except for PDB ID 1W5F and our structure 4FCs, which are both domain swapped. In these cases, a pseudomonomer was generated for each. Also, S. aureus apo structures (PDB IDs 3VO9 and 3VPA), which have a very different conformation (21), were excluded. N- and C-terminal extensions were removed, and the aligned structures are shown in ribbon representations from the same view as in Fig. 2B. Closed structures are red, except for closed S. aureus structures, which are in white, and open structures are blue. The structural conservation of FtsZ proteins is clear from the quality of alignment at the NTD (the outlier-excluded RMSDs, and the number of Cα used is shown in the last two columns of panel C). The two groups of structures can be distinguished because of the relative motion of the CTD—the open blue structures are separated from the closed white and red ones. (B) The discrete distinction between the two groups is made clearer by zooming into the CTD as indicated. (C) Cα RMSDs were calculated for all structures versus all structures, by using the PyMOL align command with zero cycles of outlier rejection (i.e., all residues matched via sequence are included in the RMSD calculation). The RMSD for each pair of structures is indicated with a linear three-color gradient as indicated below the matrix. Within each species, sets of highly similar structures are found (blue squares on the diagonal filling the black lines), with the exception of S. aureus, where the two conformations, open and closed, align poorly. The S. aureus closed structures are more similar to FtsZ proteins from other species than they are to open S. aureus structures, indicating that all existing non-S. aureus FtsZ structures are in similar, closed, conformations.

FIG 4

Atomic details of the SaFtsZ regions allowing the conformational switch. Cartoon representation of structures 1FOf (open, blue) and 5FCm (closed, red) are shown superposed after alignment on the NTD. Nucleotides and labeled residues are shown as sticks. Noncarbon atoms are colored by element, except in panel A. The viewpoint is indicated in small cartoons with coloring as in Fig. 2. (A) Top view of FtsZ NTD. Helices are numbered. Note the very minimal rearrangements in this region after both a conformational switch and nucleotide hydrolysis. (B) View of the top of H-7 and into the nucleotide binding pocket. Cartoons are semitransparent. The inset is on same scale and shows the molecule rotated 90° as indicated. The shift of R191 Cα is 3.6 Å. Note the rearrangement of individual side chains between conformations. (C, D) Identical views of the three-way interaction among the NTD, the CTD, and H7 at the top of H-7. 1FOf is semitransparent in panel C with no side chains shown, in panel D 5FCm is semitransparent. Identical side chains are shown in both. The three-way interaction is different in each conformation, but solvent is excluded from the hydrophobic core in both cases.

FIG 5

The closed conformation corresponds to the monomeric state of FtsZ, and the open conformation corresponds to the filament, including in E. coli. (A, B) FtsZ pairs were extracted from crystal lattices as described in the text. Structures are shown in cartoon representations, and each chain is rainbow colored blue to red, from the N terminus to the C terminus. Nucleotide atoms are colored by element. In each case, the view is from the same orientation after the lower molecule is aligned with the NTD of the lower subunit from 1FOf. 4FCs is shown with one chain colored and the other white to highlight the domain swap. EcFtsZ filament cryoEM density is also shown in panel A at a threshold of 7.5 σ (middle) and also with a 1FOf filament fitted into it (right). Open structures can be arranged to have 44-Å repeats by using favorable tubulin-like interfaces. Closed structures (B) have smaller, incomplete, intersubunit interfaces within crystals and cannot be sensibly arranged to produce straight filaments with a 44-Å repeat. (C) Typical micrograph of frozen-hydrated EcFtsZ GMPCPP filaments. Curved, straight, single, and double/bundled filaments are shown. (D) Representative EcFtsz filament 2D class produced by RELION. (E, F) FtsZ structures as indicated were fitted into the EcFtsZ cryoEM density with the CHIMERA volume viewer fitting tool. The flexible T3 loop region is indicated. (E) A 1FOf 5-mer fits very well into the density, as does a 1FOf monomer, with both fits extremely similar. RMSD is for the middle subunit in rigidly fitted 5-mer and monomer molecules fitted into the middle subunit density. (F) Closed structures do not fit well into the electron density, certainly not so that a repeating filament can be constructed. Some regions of poor fit are indicated by arrowheads. 2TCm was fitted by using only the NTD, which fit into the same position as the open-structure NTDs. EM, electron microscopic.

FIG 6

FtsZ’s polymerization-associated conformational switch allows treadmilling of single-stranded filaments. Black arrows indicate rates roughly in proportion to their width, and similarly colored arrows in panel A indicate rates that are exactly equivalent. See the text for a discussion of the limitations and assumptions of these simplified models, particularly regarding implied orientation of molecules. (A) An idealized rigid (lacking a conformational switch), tubulin-like, filament-forming protein, for which addition or loss of a given nucleotide:monomer complex is isodesmic. This filament cannot do robust treadmilling, as breakage is the same as minus end subunit loss, and it cannot couple structural and kinetic polarity. (B) A single-stranded version of panel A with a polymerization-associated conformational switch (between blue and red forms) able to treadmill robustly and with coupled kinetic and structural polarities. The conformational switch allows filament breakage and subunit loss from ends to be different and for the stereochemistry of subunit addition at either end to be different, meaning that addition will take place at different rates in a manner defined by structural polarity.

FIG 7

A polymerization-associated conformational switch generates asymmetry between filament end interfaces. (Middle) Three molecules from the open-form 1FOf crystal filament, slightly separated for clarity, are shown as Cα ribbons. The middle subunit is rainbow colored blue to red from the N terminus to the C terminus, and the top and bottom subunits are gray (right, left). The middle subunit is replaced with a closed-form 3FCm molecule aligned with the middle subunit NTD (right) or CTD (left), as indicated by the arrows. The different pairs of approaching surfaces are labeled B/T_m/f for bottom/top monomer/filament. These modeled closed and open interfaces do not represent the transition state of subunit addition at either end of a filament (or even any position on the reaction pathway), but they illustrate the fact that the conformational switch will necessarily lead to stereochemically different reaction pathways at each end that allow the two ends to have different rates of subunit addition, linking structural and kinetic polarity.