Polymorphism of FtsZ filaments on lipid surfaces: role of monomer orientation

How Does the Spatial Confinement of FtsZ to a Membrane Surface Affect Its Polymerization Properties and Function?

FtsZ is the cytoskeletal protein that organizes the formation of the septal ring and orchestrates bacterial cell division. Its association to the membrane is essential for its function. In this mini-review I will address the question of how this association can interfere with the structure and dynamic properties of the filaments and argue that its dynamics could also remodel the underlying lipid membrane through its activity. Thus, lipid rearrangement might need to be considered when trying to understand FtsZ’s function. This new element could help understand how FtsZ assembly coordinates positioning and recruitment of the proteins forming the septal ring inside the cell with the activity of the machinery involved in peptidoglycan synthesis located in the periplasmic space.

Protein Recruitment through Indirect Mechanochemical Interactions

Some of the key proteins essential for important cellular processes are capable of recruiting other proteins from the cytosol to phospholipid membranes. The physical basis for this cooperativity of binding is, surprisingly, still unclear. Here, we suggest a general feedback mechanism that explains cooperativity through mechanochemical coupling mediated by the mechanical properties of phospholipid membranes. Our theory predicts that protein recruitment, and therefore also protein pattern formation, involves membrane deformation and is strongly affected by membrane composition.

Surface Orientation and Binding Strength Modulate Shape of FtsZ on Lipid Surfaces

We have used a simple model system to test the prediction that surface attachment strength of filaments presenting a torsion would affect their shape and properties. FtsZ from E. coli containing one cysteine in position 2 was covalently attached to a lipid bilayer containing maleimide lipids either in their head group (to simulate tight attachment) or at the end of a polyethylene glycol molecule attached to the head group (to simulate loose binding). We found that filaments tightly attached grew straight, growing from both ends, until they formed a two-dimensional lattice. Further monomer additions to their sides generated a dense layer of oriented filaments that fully covered the lipid membrane. After this point the surface became unstable and the bilayer detached from the surface. Filaments with a loose binding were initially curved and later evolved into straight thicker bundles that destabilized the membrane after reaching a certain surface density. Previously described theoretical models of FtsZ filament assembly on surfaces that include lateral interactions, spontaneous curvature, torsion, anchoring to the membrane, relative geometry of the surface and the filament ‘living-polymer’ condition in the presence of guanosine triphosphate (GTP) can offer some clues about the driving forces inducing these filament rearrangements.

Bacterial cell division: modeling FtsZ assembly and force generation from single filament experimental data

The bacterial cytoskeletal protein FtsZ binds and hydrolyzes GTP, self-aggregates into dynamic filaments and guides the assembly of the septal ring on the inner side of the membrane at midcell. This ring constricts the cell during division and is present in most bacteria. Despite exhaustive studies undertaken in the last 25 years after its discovery, we do not yet know the mechanism by which this GTP-dependent self-aggregating protein exerts force on the underlying membrane. This paper reviews recent experiments and theoretical models proposed to explain FtsZ filament dynamic assembly and force generation. It highlights how recent observations of single filaments on reconstituted model systems and computational modeling are contributing to develop new multiscale models that stress the importance of previously overlooked elements as monomer internal flexibility, filament twist and flexible anchoring to the cell membrane. These elements contribute to understand the rich behavior of these GTP consuming dynamic filaments on surfaces. The aim of this review is 2-fold: (1) to summarize recent multiscale models and their implications to understand the molecular mechanism of FtsZ assembly and force generation and (2) to update theoreticians with recent experimental results.

Gain-of-function variants of FtsA form diverse oligomeric structures on lipids and enhance FtsZ protofilament bundling

Escherichia coli requires FtsZ, FtsA and ZipA proteins for early stages of cell division, the latter two tethering FtsZ polymers to the cytoplasmic membrane. Hypermorphic mutants of FtsA such as FtsA* (R286W) map to the FtsA self-interaction interface and can bypass the need for ZipA. Purified FtsA forms closed minirings on lipid monolayers that antagonize bundling of FtsZ protofilaments, whereas FtsA* forms smaller oligomeric arcs that enable bundling. Here, we examined three additional FtsA*-like mutant proteins for their ability to form oligomers on lipid monolayers and bundle FtsZ. Surprisingly, all three formed distinct structures ranging from mostly arcs (T249M), a mixture of minirings, arcs and straight filaments (Y139D) or short straight double filaments (G50E). All three could form filament sheets at higher concentrations with added ATP. Despite forming these diverse structures, all three mutant proteins acted like FtsA* to enable FtsZ protofilament bundling on lipid monolayers. Synthesis of the FtsA*-like proteins in vivo suppressed the toxic effects of a bundling-defective FtsZ, exacerbated effects of a hyper-bundled FtsZ, and rescued some thermosensitive cell division alleles. Together, the data suggest that conversion of FtsA minirings into any type of non-miniring oligomer can promote progression of cytokinesis through FtsZ bundling and other mechanisms.

Escherichia coli ZipA Organizes FtsZ Polymers into Dynamic Ring-Like Protofilament Structures

ZipA is an essential cell division protein in Escherichia coli. Together with FtsA, ZipA tethers dynamic polymers of FtsZ to the cytoplasmic membrane, and these polymers are required to guide synthesis of the cell division septum. This dynamic behavior of FtsZ has been reconstituted on planar lipid surfaces in vitro, visible as GTP-dependent chiral vortices several hundred nanometers in diameter, when anchored by FtsA or when fused to an artificial membrane binding domain. However, these dynamics largely vanish when ZipA is used to tether FtsZ polymers to lipids at high surface densities. This, along with some in vitro studies in solution, has led to the prevailing notion that ZipA reduces FtsZ dynamics by enhancing bundling of FtsZ filaments. Here, we show that this is not the case. When lower, more physiological levels of the soluble, cytoplasmic domain of ZipA (sZipA) were attached to lipids, FtsZ assembled into highly dynamic vortices similar to those assembled with FtsA or other membrane anchors. Notably, at either high or low surface densities, ZipA did not stimulate lateral interactions between FtsZ protofilaments. We also used E. coli mutants that are either deficient or proficient in FtsZ bundling to provide evidence that ZipA does not directly promote bundling of FtsZ filaments in vivo. Together, our results suggest that ZipA does not dampen FtsZ dynamics as previously thought, and instead may act as a passive membrane attachment for FtsZ filaments as they treadmill.

Bacterial cells use a membrane-attached ring of proteins to mark and guide formation of a division septum at midcell that forms a wall separating the two daughter cells and allows cells to divide. The key protein in this ring is FtsZ, a homolog of tubulin that forms dynamic polymers. Here, we use electron microscopy and confocal fluorescence imaging to show that one of the proteins required to attach FtsZ polymers to the membrane during E. coli cell division, ZipA, can promote dynamic swirls of FtsZ on a lipid surface in vitro. Importantly, these swirls are observed only when ZipA is present at low, physiologically relevant surface densities. Although ZipA has been thought to enhance bundling of FtsZ polymers, we find little evidence for bundling in vitro. In addition, we present several lines of in vivo evidence indicating that ZipA does not act to directly bundle FtsZ polymers.

Unite to divide: Oligomerization of tubulin and actin homologs regulates initiation of bacterial cell division

To generate two cells from one, bacteria such as Escherichia coli use a complex of membrane-embedded proteins called the divisome that synthesize the division septum. The initial stage of cytokinesis requires a tubulin homolog, FtsZ, which forms polymers that treadmill around the cell circumference. The attachment of these polymers to the cytoplasmic membrane requires an actin homolog, FtsA, which also forms dynamic polymers that directly bind to FtsZ. Recent evidence indicates that FtsA and FtsZ regulate each other’s oligomeric state in E. coli to control the progression of cytokinesis, including the recruitment of septum synthesis proteins. In this review, we focus on recent advances in our understanding of protein-protein association between FtsZ and FtsA in the initial stages of divisome function, mainly in the well-characterized E. coli system.

Collective effects of torsion in FtsZ filaments

Recent evidence points to the presence of torsion in FtsZ bonds. In addition, experiments with FtsZ mutants on surfaces resulted in new aggregates that cannot be explained by older models for FtsZ dynamics. We use an interaction model for FtsZ derived from molecular dynamics simulations and expand a fine-grained lattice model used to describe FtsZ aggregates on a surface. This new model includes different anchoring angles for the monomers and allows bond twist, two ingredients that oppose each other resulting in a more dynamic and interesting system. We study the role and importance of these conflicting elements and how the aggregates are characterized by the different interaction parameters.

Modeling the interplay between protein and lipid aggregation in supported membranes

We present a theoretical model that deals with the complex interplay between lipid segregation and the self-aggregation of lipid-attached proteins. The model, in contrast to previous ones that consider proteins only as passive elements affecting the lipid distribution, describes the system including three terms: the dynamic interactions between protein monomers, the interactions between lipid components, and a mixed term considering protein-lipid interactions. It is used to explain experimental results performed on a well-defined system in which a self-aggregating soluble bacterial cytoskeletal protein polymerizes on a lipid bilayer containing two lipid components. All the elements considered in a previously described protein model, including torsion of the monomers within the filament, are needed to account for the observed filament shapes. The model also points out that lipid segregation can affect the length and curvature of the filaments and that the dynamic behavior of the lipids and proteins can have different time scales, giving rise to memory effects. This simple model that considers a dynamic protein assembly on a fluid and active lipid surface can be easily extended to other biologically relevant situations in which the interplay between protein and lipid aggregation is needed to fully describe the system.

Torsion and curvature of FtsZ filaments

FtsZ filaments participate in bacterial cell division, but it is still not clear how their dynamic polymerization and shape exert force on the underlying membrane. We present a theoretical description of individual filaments that incorporates information from molecular dynamic simulations. The structure of the crystallized Methanococcus jannaschii FtsZ dimer was used to model a FtsZ pentamer that showed a curvature and a twist. The estimated bending and torsion angles between monomers and their fluctuations were included in the theoretical description. The MD data also permitted positioning the curvature with respect to the protein coordinates and allowed us to explore the effect of the relative orientation of the preferred curvature with respect to the surface plane. We find that maximum tension is attained when filaments are firmly attached and oriented with their curvature perpendicular to the surface and that the twist serves as a valve to release or to tighten the tension exerted by the curved filaments on the membrane. The theoretical model also shows that the presence of torsion can explain the shape distribution of short filaments observed by Atomic Force Microscopy in previously published experiments. New experiments with FtsZ covalently attached to lipid membranes show that the filament on-plane curvature depends on lipid head charge, confirming the predicted monomer orientation effects. This new model underlines the fact that the combination of the three elements, filament curvature, twist and the strength and orientation of its surface attachment, can modulate the force exerted on the membrane during cell division.