Title: Editor's evaluation: On the role of nucleotides and lipids in the polymerization of the actin homolog MreB from a Gram-positive bacterium
Abstract: Full text Figures and data Side by side Abstract Editor's evaluation Introduction Results Discussion Materials and methods Data availability References Decision letter Author response Article and author information Metrics Abstract In vivo, bacterial actin MreB assembles into dynamic membrane-associated filamentous structures that exhibit circumferential motion around the cell. Current knowledge of MreB biochemical and polymerization properties in vitro remains limited and is mostly based on MreB proteins from Gram-negative species. In this study, we report the first observation of organized protofilaments by electron microscopy and the first 3D-structure of MreB from a Gram-positive bacterium. We show that Geobacillus stearothermophilus MreB forms straight pairs of protofilaments on lipid surfaces in the presence of ATP or GTP, but not in the presence of ADP, GDP or non-hydrolysable ATP analogs. We demonstrate that membrane anchoring is mediated by two spatially close short hydrophobic sequences while electrostatic interactions also contribute to lipid binding, and show that the population of membrane-bound protofilament doublets is in steady-state. In solution, protofilament doublets were not detected in any condition tested. Instead, MreB formed large sheets regardless of the bound nucleotide, albeit at a higher critical concentration. Altogether, our results indicate that both lipids and ATP are facilitators of MreB polymerization, and are consistent with a dual effect of ATP hydrolysis, in promoting both membrane binding and filaments assembly/disassembly. Editor's evaluation This important study makes the case that the assembly of MreB from Geobacillus, a Gram-positive organism differs substantially from MreB from the Gram-negative model organism, Escherichia coli. They make the compelling case that Geobacillus MreB assembly requires both interactions with membrane lipids and nucleotide binding: nucleotide hydrolysis is required for interaction with the membrane and interaction with lipids triggers polymerization. Altogether, these data make the strong case that MreB assembly dynamics can vary in significant, and organism-specific ways. https://doi.org/10.7554/eLife.84505.sa0 Decision letter Reviews on Sciety eLife's review process Introduction Cytoskeletal proteins are known to polymerize into filaments that play critical roles in various aspects of cell physiology, including cell shape, mechanical strength and motion, cytokinesis, chromosome partitioning and intracellular transport. Prokaryotic cells contain homologs of the main eukaryotic cytoskeletal proteins, namely actin, tubulin and intermediate filaments (Cabeen and Jacobs-Wagner, 2010; Lin and Thanbichler, 2013; Shaevitz and Gitai, 2010), which were identified decades after their eukaryotic counterparts. In 2001, MreB proteins of the Gram-positive model bacterium Bacillus subtilis were found to form filamentous structures underneath the cytoplasmic membrane and to play a key role in the determination and maintenance of rod-shape (Carballido-Lopez, 2017; Jones et al., 2001). Soon after, the three-dimensional structure of one of the two MreB isoforms from the Gram-negative thermophilic bacterium Thermotoga maritima (MreBTm) was solved (van den Ent et al., 2001), confirming its structural homology with actin (Bork et al., 1992). Besides, MreBTm in solution was shown to assemble into filaments similar to filamentous actin (F-actin; van den Ent et al., 2001). Research in the field of eukaryotic actin historically focused on elucidating structure-function relationships from in vitro studies. The availability of large amounts of soluble actin purified from several cell types since the 1940s enabled decades of mechanistic studies on actin polymerization (Pollard, 2016). In contrast, MreB from mesophilic bacteria proved particularly difficult to purify in a soluble form, thwarting efforts to perform in vitro assays. Instead, research on MreB primarily focused on cellular studies, driven by the advent of fluorescent microscopy in bacterial cell biology. Over the past two decades, the subcellular localization and dynamics of MreB have been described in several Gram-negative and Gram-positive species (Billaudeau et al., 2017; Billaudeau et al., 2019; Dion et al., 2019; Harris et al., 2014; Hussain et al., 2018; Olshausen et al., 2013; Oswald et al., 2016; Ouzounov et al., 2016; Renner et al., 2013; Schirner et al., 2015). In vivo, MreB proteins form discrete, nanometer-long, membrane-associated polymeric assemblies along the cell cylinder that move processively around the rod circumference together with proteins of the cell wall (CW) elongation machinery (Domínguez-Escobar et al., 2011; Garner et al., 2011; van Teeffelen et al., 2011), forming the so-called Rod complex. The motility of the Rod complex is driven by CW synthesis (Domínguez-Escobar et al., 2011; Garner et al., 2011; van Teeffelen et al., 2011) and MreB assemblies self-align circumferentially, along their direction of motion (Billaudeau et al., 2019; Hussain et al., 2018). Recently, it was proposed that the specific intrinsic curvature of MreB polymers increases their affinity for the greatest concave (negative) membrane curvature within the cell, that is the inner surface of the rod circumference, accounting for their orientation (Hussain et al., 2018). The current model is that self-aligned MreB filaments restrict the diffusion of CW biosynthetic proteins in the membrane and orient their motion to insert new peptidoglycan strands in radial hoops perpendicular to the long axis of the cell, promoting the cylindrical expansion of rod-shaped cells (Domínguez-Escobar et al., 2011; Garner et al., 2011; Hussain et al., 2018). However, many questions remain to be answered. What prompts the assembly of MreB on the inner leaflet of the cytoplasmic membrane? What is the architecture of the membrane-associated MreB polymeric assemblies and how is it controlled? How is their distribution along the cell cylinder regulated? What is the length of individual MreB filaments within these assemblies and how is it controlled? Are the filaments stable? Do they exhibit turnover like actin filaments? In vivo, the length of MreB filamentous assemblies can be affected by the intracellular concentration of the protein (Billaudeau et al., 2019; Salje et al., 2011), but seems to have little impact on MreB function (Billaudeau et al., 2019). No turnover of MreB assemblies was detected in vivo, at least relative to their motion around the cell circumference (Domínguez-Escobar et al., 2011; van Teeffelen et al., 2011). Therefore, MreB polymers are believed to be quite stable despite their motion in the cell. To elucidate in detail the molecular mechanisms underlying the functions of MreB, it remains necessary to understand their biochemical and polymerization properties. The majority of biochemical and structural studies on MreB proteins originally focused on the highly soluble Gram-negative MreBTm (Bean and Amann, 2008; Esue et al., 2005; Esue et al., 2006; Popp et al., 2010b; van den Ent et al., 2001; van den Ent et al., 2010). The tendency to aggregation upon purification hampered most in vitro studies of MreBs from other species. Over the last decade, purification and polymerization assays were nevertheless reported for MreBs from several Gram-negative bacteria, from the Gram-positive B. subtilis (MreBBs) and from wall-less Chlamydophila and Spiroplasma species (Dersch et al., 2020; Gaballah et al., 2011; Harne et al., 2020; Maeda et al., 2012; Mayer and Amann, 2009; Nurse and Marians, 2013; Pande et al., 2022; Salje et al., 2011; Takahashi et al., 2022; van den Ent et al., 2014). Direct binding to the cell membrane was shown for MreB from the Gram-negative Escherichia coli (MreBEc) and T. maritima (Salje et al., 2011) and, more recently, for MreB from Spiroplasma citri (MreB5Sc; Harne et al., 2020). The N-terminal amphipathic helix of MreBEc was found to be necessary for membrane binding and also to cause the full-length purified protein to aggregate (Salje et al., 2011). Although this N-terminal amphipathic helix is dispensable for polymerization, it is required for proper function of MreBEc in vivo (Salje et al., 2011). MreBTm and MreB5Sc are devoid of such an N-terminal amphipathic helix, but instead possess a small hydrophobic loop that protrudes from the monomeric globular structure and mediates membrane binding (Pande et al., 2022; Salje et al., 2011). Additionally, an acidic C-terminal tail was shown to mediate a charge-based interaction of MreB5Sc with the membrane (Pande et al., 2022). Altogether, in vitro work on MreBs from Gram-negative bacteria has shown that MreB polymerizes into straight double filaments in the presence of nucleotides, both in solution and on lipid membrane surfaces (Harne et al., 2020; Salje et al., 2011; Takahashi et al., 2022; van den Ent et al., 2014; van den Ent et al., 2010), and that filaments can assemble into larger sheets by lateral interactions (Esue et al., 2005; Esue et al., 2006; Harne et al., 2020; Nurse and Marians, 2013; Popp et al., 2010b; van den Ent et al., 2001; van den Ent et al., 2014). Furthermore, work on Caulobacter crescentus MreB (MreBCc) and MreBEc indicated an antiparallel arrangement of the straight pairs of protofilaments (van den Ent et al., 2014), in sharp contrast to the helical parallel pairs of protofilaments (double helix) characteristic of F-actin (Pollard, 1990). While the parallel arrangement of a protofilament doublet generates polarity and allows for the characteristic treadmilling of F-actin (Stoddard et al., 2017), the antiparallel arrangement in MreB protofilaments suggests a bidirectional polymerization/depolymerization mechanism (van den Ent et al., 2014). The directionality and the kinetics of MreB polymerization, as well as the role of nucleotides in this process remain to be shown. ATPase activity has been reported in solution for MreBTm, MreBEc, MreBBs and Spiroplasma MreBs (Esue et al., 2005; Esue et al., 2006; Mayer and Amann, 2009; Nurse and Marians, 2013; Pande et al., 2022; Popp et al., 2010b; Takahashi et al., 2022). However, the need for nucleotide binding and hydrolysis in polymerization remains unclear. Early reports suggested a strict dependency on hydrolysable nucleotides (ATP or GTP) for polymerization of MreBTm (Esue et al., 2006; van den Ent et al., 2001), and later for MreBEc (Nurse and Marians, 2013), while others claimed that polymerization occurred similarly in the presence of ADP (Bean and Amann, 2008; Gaballah et al., 2011; Mayer and Amann, 2009; Pande et al., 2022; Popp et al., 2010b; Takahashi et al., 2022), of the non-hydrolysable ATP analogue AMP-PNP (adenylyl-imidodiphosphate) (Bean and Amann, 2008; Mayer and Amann, 2009; Pande et al., 2022; Salje et al., 2011; Takahashi et al., 2022), or even in the absence of nucleotide (Mayer and Amann, 2009). No electron microscopy (EM) images of protofilaments or atomic views of MreB from a Gram-positive bacterium have been reported to date. The two in vitro studies so far reported on B. subtilis MreB investigated polymerization by light scattering and sedimentation assays (Mayer and Amann, 2009) and by fluorescence microscopy (Dersch et al., 2020), respectively, but provided no evidence that MreBBs polymerized into protofilaments in the conditions tested. In Gram-positive bacteria, MreB proteins presumably have no N-terminal amphipathic helix (Salje et al., 2011), and the genome usually encodes several MreB isoforms (in contrast to Gram-negative that usually get by with a single mreB paralog), that may be related to their thicker and more complex CW structure (Chastanet and Carballido-Lopez, 2012). Inter- and intra-species differences in MreBs may exist at the structural or biochemical level, leading to differences in molecular interactions or biological functions. In this study, we aimed to decipher fundamental structural and biochemical properties of MreB from a Gram-positive bacterium. We successfully purified a soluble form of MreB from the Gram-positive thermophilic Geobacillus stearothermophilus (MreBGs) and elucidated its crystal structure, confirming the classical actin/MreB fold and the presence of the small hydrophobic loop shown to mediate membrane binding in MreBTm and MreB5Sc (Pande et al., 2022; Salje et al., 2011). Polymerization assays showed that MreBGs forms straight pairs of protofilaments in the presence of lipids and nucleotide triphosphate (either ATP or GTP), and that these are dynamic. We also show that the interaction with lipids is mediated by electrostatic interactions (Pande et al., 2022) and by two spatially close hydrophobic motifs in the MreBGs monomers that comprise the small hydrophobic loop and the N-terminal end. Free in solution, MreBGs assembled into large sheets regardless of the bound nucleotide, albeit at a higher MreBGs concentration than the one required for polymerization into pairs of protofilaments on a lipid surface. Taken together, our results show a key role for ATP as facilitator of MreB polymerization on the membrane, and suggest that ATP hydrolysis promotes both MreB membrane binding and filament assembly/disassembly. Results Crystal structure of G. stearothermophilus MreB To solve the structure of a MreB protein from a Gram-positive bacterium but overcome the notorious aggregation issues of MreB from mesophilic bacteria, we cloned and purified MreB from the thermophilic G. stearothermophilus (MreBGs). We chose G. stearothermophilus because of its proximity to the Bacillus genus and because of the highly conserved sequence of MreBGs compared to MreB from the model Gram-positive bacterium B. subtilis. MreBBs is more closely related to MreBGs (85.6% identity and 92.6% similarity) than to MreB of Gram-negative for which biochemical or structural data are available (either the thermophilic T. maritima with 55.8% identity, or the mesophilic C. crescentus, 56.9% identity and E. coli, 55.2% identity) (Figure 1—figure supplement 1). MreBGs was purified to homogeneity following a two-step procedure (see Materials and methods). The protein could be purified in a soluble form (Figure 1—figure supplement 2) that remained functional for polymerization at concentrations below 13.4 µM (0.5 mg/mL). When stored frozen at higher concentrations or when conserved overnight at 4 °C, MreBGs rapidly aggregated (Figure 1—figure supplement 2A) and could not be recovered in a monomeric state, consistent with the known tendency of MreB proteins to aggregate. The purified MreBGs protein was crystallized and the structure of its apo form was solved at 1.8 Å resolution (Protein Data Bank Identifier (PDB ID) 7ZPT). The crystals belong to the monoclinic P21 space group and contain one molecule of MreBGs per asymmetric unit (Supplementary file 1). Monomers of apo MreBGs display the canonical fold of actin-like proteins, characterized by four subdomains IA, IIA, IB and IIB (Figure 1A). One of the most similar structures to apo MreBGs is the apo form of MreBTm (PDB ID 1JCF; van den Ent et al., 2001), with a root mean square deviation (rmsd) of 1.92 Å over 305 superimposed Cα atoms and a Z-score of 16.0. Superimposition of the two proteins (Figure 1A) revealed that MreBGs is in a slightly more open conformation than MreBTm, mainly due to a movement of domain IB, which is the less conserved within the actin superfamily of proteins. Loop β6-α2, which connects subdomains IA and IB and closes the nucleotide-binding pocket, is partially disordered in apo MreBGs. In domain IA, the hydrophobic loop α2-β7, which has been shown to be involved in MreBTm membrane binding (Salje et al., 2011) and is 2 residues longer in MreBGs (Figure 1—figure supplement 1), displays a distinct conformation, packed on the N-terminal extremity of the polypeptide chain. Figure 1 with 2 supplements see all Download asset Open asset Crystal structure of the apo protofilament of MreB from G. stearothermophilus. (A) Crystal structure of apo MreBGs (PDB ID 7ZPT), colored by subdomains, superimposed on the crystal structure of apo MreBTm (PDB ID 1JCF), in beige. The sequence similarity between the two proteins is 55.8%. Subdomain IA (blue) of MreBGs is formed by residues 1–32, 66–145 and 315–347; subdomain IB (yellow) by residues 33–65; IIA (red) by residues 146–181 and 246–314 and IIB (green) by residues 182–245. Superimposition of the two forms highlights the distinct positions of loops β6-α2 and α2-β7 as well as the movement of domain IB (two-headed arrow) resulting in slightly distinct subunit interaction modes as shown in panel C. (B) Protofilament structure of apo MreBGs. Three subunits of the protofilament formed upon crystal packing are displayed as cartoon and colored by subdomains. The subunit repeat distance is indicated. (C) Close view of the MreBGs intra-protofilament interface. The two subunits are colored by subdomains as in panel A, and shown as cartoons. Residues involved in putative salt bridges (gray dashed lines) are displayed as sticks colored by atom type (N in blue and O in red) and labeled. (D) Close view of the MreBTm intra-protofilament interface (PDB ID 1JCF). The two subunits are colored in beige as in panel A, and shown as cartoons. Residues involved in putative salt bridges (gray dashed lines) are displayed as sticks colored by atom type (N in blue and O in red) and labeled. Crystal packing analysis revealed straight protofilaments characterized by a subunit repeat distance of 51 Å (Figure 1B), similar to that observed in crystal structures of other actin homologs (Harne et al., 2020; Pande et al., 2022; Roeben et al., 2006; van den Ent et al., 2014). However, because of the open conformation of MreBGs (Figure 1A), domain IB interacts with domain IA (Figure 1C) and not with domain IIA as observed for example for MreBTm (Figure 1D; van den Ent et al., 2001). MreBGs polymerizes into straight pairs of protofilaments in the presence of lipids We next investigated the polymerization of MreBGs by EM of negatively stained samples. In vivo, MreBBs forms membrane-associated nanofilaments (Billaudeau et al., 2019; Hussain et al., 2018; Jones et al., 2001), and MreB filaments from Gram-negative bacteria have been shown to have an intrinsic affinity for membranes (Garenne et al., 2020; Maeda et al., 2012; Salje et al., 2011; van den Ent et al., 2014). We hypothesized that the presence of lipids might be important for the assembly of MreBGs polymers, and thus performed polymerization reactions in the presence and in the absence of lipids. In the presence of ATP, MreBGs readily formed pairs of protofilaments on a monolayer of total E. coli lipid extract, while these were virtually not observed in the absence of lipids (Figure 2A). Using a semi-quantitative workflow analysis of TEM grids (Figure 2—figure supplement 1; Materials and methods), we found that in the presence of both ATP and lipids, MreBGs formed a lawn of double protofilaments in 100% of fields, while only 4% and 8% of the fields contained polymers (often at very low density) in the absence of either ATP or lipids, respectively (Figure 2A). To test if polymers had formed in solution but failed to bind to the hydrophobic EM grids, we instead used glow-discharged hydrophilic grids, which are commonly used to adsorb soluble proteins. Again, MreBGs filaments were not significantly detected in solution (Figure 2—figure supplement 2A). In the presence of lipids, the frequency of polymers was drastically reduced on the glow-discharged grids, consistent with impaired adhesion of a lipid monolayer to a hydrophilic surface (Figure 2—figure supplement 2A). We next hypothesized that the critical concentration for polymerization might be higher in the absence of lipids and thus raised the concentration of MreBGs in the reaction (Figure 2—figure supplement 2B). Again, virtually no pairs of protofilaments were detected in solution in these conditions. The only polymeric structures observed in the bulk solution, albeit very infrequently and only at high MreB concentration, were some large multilayered sheets forming ribbon-like structures among aggregates (Figure 2—figure supplement 2B and C). Taken together, these observations indicated that MreBGs polymerization is strongly enhanced by both ATP and lipids. Figure 2 with 5 supplements see all Download asset Open asset MreBGs forms double protofilaments in the presence of ATP and lipids. (A) Polymerization of MreBGs into pairs of protofilaments depends on the presence of lipids and ATP. MreBGs was set to polymerize in standard conditions in the presence or absence of ATP and lipid total extract from E. coli. Polymer formation is expressed as percent of fields containing high (black) or low (grey) density of polymers (see Figure 2—figure supplement 1 for details of the quantification method of MreB polymers on TEM grids). Values are average of two independent experiments. Error bars are standard deviations. Inset shows an example of a field of dual protofilaments on a negative stained TEM image. Scale bar, 50 nm. (B) Polymer formation as a function of MreBGs concentration. MreBGs was set to polymerize in standard conditions at a concentration ranging from 0.27 to 1.34 µM (0.01–0.05 mg/mL). Values are the average of two independent experiments. Error bars are standard deviations. (C, D) MreBGs polymers assemble into sheets. (C) EM image of MreBGs set to polymerize in standard conditions. Scale bar, 50 nm. Fourier transform (D) was obtained from the area indicated by a white box in (C) and revealed a longitudinal subunit repeat of the filaments of 54 Å and a lateral spacing of ~37 Å (arrowheads). (E) (Left) 2D averaging of images of negatively stained dual protofilaments of MreBGs from 1 554 individual particles. Scale bar, 3 nm. Two copies of the atomic structure of the protofilaments found in the MreBGs crystals shown to scale (Middle, for illustration the two protofilaments are displayed arbitrarily in an antiparallel conformation but could fit in a parallel conformation as well) and docked into the 2D averaged EM image (Right). (F, G) MreBGs polymers assemble on lipid bilayers and distort liposomes as shown by cryo-electron microscopy (cryo-EM). Cryo-EM micrographs of liposomes (0.37 mg/mL) made from E. coli lipid total extracts incubated with purified MreBGs (1.34 µM; 0.05 mg/mL) in the presence of ATP (2 mM), and low (F, 100 mM) or high (G, 500 mM) concentration of KCl. Arrowheads point to MreB accumulations. Scale bars, 50 nm. (H). Cryo-EM micrographs showing the cross-section of the membrane of liposomes in the absence (Left) and in the presence (Right) of ATP-bound MreBGs at 500 mM KCl. Scale bars, 50 nm. On a lipid monolayer, polymers were observed at a concentration of MreB above 0.55 µM (0.02 mg/mL), for a theoretical critical concentration of ~0.45 µM (Figure 2B), which is very similar to the critical concentration reported for ATP-MreBTm (Bean and Amann, 2008). The simplest and most abundant assemblies are paired protofilaments (Figure 2A, Figure 2—figure supplements 1 and 3), as previously observed for MreBTm and MreBCc assembled on lipid monolayers (Salje et al., 2011; van den Ent et al., 2014), and for Spiroplasma MreBs in solution (Pande et al., 2022; Takahashi et al., 2022). Pairs of MreBGs protofilaments are generally straight, and single protofilaments were never observed. Paired protofilaments of different lengths, ranging from below 50 nm up to several micrometers, as well as partial lateral association into two-dimensional sheets of dual protofilaments are often observed on the same EM grid (Figure 2A and C, and Figure 2—figure supplement 3). Importantly, pairs of filaments and sheets always lay flat, indicating that they are oriented relative to the membrane surface. The diffraction patterns of the sheets showed a longitudinal repeat of 54 Å and a lateral spacing of ~37 Å (Figure 2C and D). 2D averaging of negatively stained EM images of 1 554 individual pairs of filaments (Figure 2E and Figure 2—figure supplement 4) confirmed a longitudinal subunit repeat of 54 Å and refined the lateral subunit repeat to 31 Å, and could accommodate well two scaled protofilaments found in the MreBGs crystals (Figure 2E). However, it is not possible to derive the orientation of the two protofilaments (i.e. parallel or antiparallel) from the EM density obtained by 2D averaging. Cations modulate distortion of liposomes by MreBGs filaments MreBGs filaments also formed on lipid bilayers as observed by cryo-electron microscopy (cryo-EM). To this end, we prepared large unilamellar vesicles (LUVs) from E. coli lipid total extract, and incubated them with MreBGs and ATP. LUVs alone were spherical (Figure 2—figure supplement 5A), but vesicles decorated with MreBGs filaments appeared strongly deformed, confirming that MreBGs was bound to the membrane. At 100 mM KCl (our standard polymerization condition), LUVs displayed inward bending (negatively curved areas) where MreBGs filaments accumulated (arrows in Figure 2F and Figure 2—figure supplement 5B), as previously reported for MreBTm and MreBCc (Salje et al., 2011; van den Ent et al., 2014). 100 mM KCl is a salt concentration commonly used in polymerization studies of actin and actin-like proteins, including MreB (Deng et al., 2016; Garner et al., 2004; Polka et al., 2009; Rivera et al., 2011). Yet, while cytoplasmic K+ concentrations are around 50–250 mM in multicellular eukaryotic cells (Rodríguez-Navarro, 2000; Schmidt-Nielsen, 1975), they reach 200–300 mM in yeast (Ariño et al., 2010) and vary greatly depending on the osmolality of the medium in bacteria (Cayley et al., 1991; Epstein and Schultz, 1965; Rhoads et al., 1976). In B. subtilis, the basal intracellular concentration of KCl fluctuates between 100 mM and 800 mM (Eisenstadt, 1972; Whatmore et al., 1990). We therefore tested how higher salt concentrations affect the properties of MreB polymers. At 500 mM KCl, MreBGs readily polymerized into straight pairs of filaments as well, which also distorted liposomes but did not induce negative curvature. Instead, they faceted and tubulated the liposomes (Figure 2G and Figure 2—figure supplement 5C), suggesting that high salt concentration increases the stiffness of MreB filaments and/or of the membrane. Specific binding of cations at discrete sites along the filament has been shown to stiffen actin filaments, determining their bending rigidity (Kang et al., 2012). Our result suggests that physiological salt concentrations may also play a fundamental role in the mechanical properties of MreB filaments. MreBGs largely coated the liposomes and displayed a regular pattern along the cross-section of tubulated vesicles (Figure 2G–H). This view is compatible with longitudinal sections of 2D-sheets of straight filaments aligned in parallel to the longitudinal axis of the cylinder, as previously suggested for the arrangement of MreBTm in rigid lipid tubes (van den Ent et al., 2014). ATP or GTP drive efficient formation of double filaments on a membrane surface The role of nucleotide binding and hydrolysis in MreB polymerization remains unclear. In actin, ATP binding or hydrolysis are not required for polymerization (De La Cruz et al., 2000; Kasai et al., 1965). ATP hydrolysis only occurs subsequent to the polymerization reaction, destabilizing the filaments upon release of the γ-phosphate (Korn, 1982; Korn et al., 1987). In contrast, MreBTm was reported to require either ATP or GTP to polymerize (Esue et al., 2006; Nurse and Marians, 2013; van den Ent et al., 2001). MreB from E. coli, C. crescentus, S. citri, and Leptospira interrogans also formed polymers in the presence of ATP, but the requirement of ATP for polymerization was not established (Barkó et al., 2016; Harne et al., 2020; Maeda et al., 2012; Nurse and Marians, 2013; Salje et al., 2011; van den Ent et al., 2014). However, MreB filaments or sheets of filaments were also observed in the presence of ADP (Gaballah et al., 2011; Pande et al., 2022; Popp et al., 2010b; Takahashi et al., 2022) or AMP-PNP (Pande et al., 2022; Salje et al., 2011; Takahashi et al., 2022). These observations indicated that ATP binding and hydrolysis is not strictly required for filament formation in vitro. An analysis of nucleotide-bound crystal structures of MreBCc also suggested that ATP binding may trigger the transition to the double-protofilament conformation (Pande et al., 2022). Furthermore, liposome binding studies of MreB5Sc pointed to an allosteric effect of ATP binding and hydrolysis for effective polymerization and membrane binding (Pande et al., 2022). We then wondered about the specificity of MreBGs toward nucleotides and their role in polymerization on a lipid membrane. MreBGs formed straight pairs of protofilaments and sheets in the presence of either ATP or GTP, as shown by negative stain EM (Figure 3A). Noteworthy, the average length of double filaments displayed an approximately twofold increase in the presence of GTP compared to ATP (Figure 3—figure supplement 1A). The significance of this observation is unclear at present but it may reflect differential affinity, dissociation rate or hydrolytic activity of the two nucleotide triphosphates (NTPs). Next, we asked whether formation of pairs of filaments required nucleotide hydrolysis and tested if nucleotides diphosphate or non-hydrolysable ATP analogues would also support polymer assembly. Virtually no double filaments were observed when ATP/GTP was replaced by ADP, GDP, AMP-PNP, or ApCpp (5′‐adenylyl methylenediphosphate), either in the presence or in the absence of lipids, in our standard polymerization conditions (Figure 3A and Figure 3—figure supplement 1B). However, differential affinity of MreBGs