Cell polarity control by an unconventional G-protein complex in bacteria

In Myxococcus xanthus, cell polarity and motility are regulated by the small GTPase MglA and its associated regulators, MglB, a potential GTPase Activating Protein and the RomRX complex, a potential Guanine nucleotide Exchange Factor. However, recent results have questioned the exact function of these regulators. Using a new type of fluorescent nucleotides for fluorescence-based anisotropy, we first demonstrate that RomRX does not function as a GEF but as an effector that binds strictly to MglA GTP. Secondly, based on an enzymology model, we found that MglB functions as a weak catalyst, unlike typical GAPs, and requires additional factors to dissociate from MglA. We demonstrate that RomRX, and specifically RomR, effectively competes with MglA by sequestering MglB. This low affinity interaction is still permissive for GAP activity but sufficient to partition these proteins in vivo as indicated by a biomimetic assay in oil emulsion droplets reconstituting these interactions in a cell-sized environment. These results suggest a new model for cell polarity where RomRX exerts a dual function, targeting MglA GTP to the pole via its effector function and preventing its accumulation at the lagging pole by destabilizing the MglA-MglB complex.

Cell polarity/ GTPase/ MglA/ Myxococxxus xanthus /Roadblock proteins .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 3

Cell polarity, the asymmetric organization of cellular components, is a fundamental characteristic of many living organisms. It is essential for effective cell motility and a myriad of other directional functions, such as epithelial transport, neuronal signaling, and body axis specification in embryos (Allam et al, 2018; Mayor & Etienne-Manneville, 2016) . A polarized cell establishes a distinct axis with a specific front, known as the polarity site, enabling it to maintain functionally distinct domains. This ability relies on intricate interactions between various protein complexes, including small GTPases, polarity proteins, and cytoskeletal components, that regulate multiple signaling pathways. In most eukaryotic cells, small GTPases of the Ras superfamily act as molecular switches in many signaling pathways and play a central role in cell polarization processes. For example, the Rho-family GTPases, including Cdc42 (one of the most highly conserved GTPases) and its relatives (Rac in animals, Rop in plants), are crucial for controlling polarization. Polarization signals act through Cdc42 regulators to accumulate Cdc42-GTP at the future cell front. Cdc42-GTP then organizes the cytoskeleton via various effectors, shaping the polarized morphology of the cell (Chiou et al, 2017; Iden & Collard, 2008). Understanding the molecular details of polarity is essential for understanding various diseases, as disruptions in polarity mechanisms are often linked to cancer, neurodegenerative disorders, and immune system deficiencies (Piroli et al, 2019; Houston, 2017). The discovery that bacterial cells are also polarized dates back to the first electron micrograph showing flagella at the cell poles of bacteria. Since then, efforts to study polarity-based mechanisms on model organisms including B. subtilis, C. crescentus, and E. coli have offered fundamental insights into the mechanisms underlying the spatiotemporal organization of bacterial cells (Kirkpatrick & Viollier, 2011; Laloux & Jacobs-Wagner, 2014; Rowlett & Margolin, 2013; Wettmann & Kruse, 2018; MacCready et al , 2018; Mauriello, 2019; Schumacher & Søgaard-Andersen, 2017) . Similar to eukaryotes, bacterial polarity can be stably maintained over time, dictating the assembly and activity of specific cellular organelles such as flagella, pili and stalks. Alternatively, polarity can be a dynamic process involving the active diffusion and accumulation of polarized protein to precise locations as seen in the front-rear polarity of moving Myxococcus xanthus or Pseudomonas aeruginosa cells (Herrou & Mignot, 2020; Carreira et al, 2020; Kühn et al, 2023). M. xanthus, a Gram-negative delta proteobacterium, glides on solid surfaces along their long axes with well- defined front-rear polarity. Directed cell motility is critical for its complex life cycle, which includes multicellular development and cooperative predation (Dinet & Mignot, 2023; Herrou & Mignot, 2020; Mercier & Mignot, 2016). M. xanthus cells move using two genetically distinct motility systems. The gliding (A, adventurous) motility system enables single-cell movement at the colony edges. At the molecular level, A-motility is powered by the Agl-Glt motor, which propels individual cells by adhering to specific points along the cell surface known as bacterial focal adhesions (Mignot et al, 2007; Faure et al, 2016). M. xanthus cells can also move in groups using a twitching-like (S, social) motility that involves type IV pili, which extend and retract to pull groups of cells together, facilitating collective movement across surfaces (Mercier et al, 2020). Remarkably, M. xanthus cells control their movement by reversing direction, achieved through the redirection of their motility apparatus to the opposite cell pole (Herrou & Mignot, 2020; Zhang et al, 2012a). These events termed “reversals” correspond at the cellular level, to a switching of the leading and lagging cell poles. The establishment and maintenance of polarity reversals in M. xanthus involves a finely tuned interplay between signaling proteins and regulatory complexes, with the small Ras-like GTPase MglA playing a central role. Similar to other well-characterized eukaryotic small Ras-like GTPases, MglA binds to GTP (MglA GTP) and localizes at the leading pole where it interacts with effector proteins to activate each of the motility systems spatially, SgmX for the S-motility system (Mercier et al, 2020; Bautista et al, 2023), GltJ, AglZ and MreB for the A-motility system (Attia et al, 2024; Treuner-Lange et al, 2015). Inactivation of MglA occurs upon GTP hydrolysis because the GDP-bound state (MglA GDP) cannot interact with effector proteins (Dinet & Mignot, 2023; Zhang et al , 2010 ). In the cell, the MglA GTP/GDP cycle is tightly regulated by MglA .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 4 interacting partners, notably the MglB protein and the complex formed by the RomR and RomX proteins (RomRX). Together, these four proteins form the core regulatory components of the polarity control network (Leonardy et al, 2010; Zhang et al, 2010, 2012b; Keilberg et al, 2012). The exact design of the regulatory interactions by which these proteins establish the polarity axis remains a topic of debate and is the subject of this article. The current model suggests that the functions of MglB and RomRX are spatially separated, with MglB acting at the lagging cell pole and the RomRX complex at the leading cell pole (Herrou & Mignot, 2020; Szadkowski et al, 2019; Carreira et al, 2022). This is believed to generate a polarity axis via the following mechanism: (i) At the lagging cell pole, MglB, a so-called roadblock domain protein is recruited by another roadblock domain protein MglC (Carreira et al, 2023). MglB interacts with MglA and activates GTP hydrolysis activity allosterically (see below, MglB is hence defined as a GTPase-Activating Protein (GAP) for MglA). In vitro, MglB is sufficient to activate hydrolysis, but in the cell this process is further assisted by a cognate activator, RomY (Szadkowski et al, 2022). In vivo, the action of MglB precludes accumulation of MglA at the lagging pole. (ii) Conversely at the leading pole, RomRX is proposed to exert a dual function: first, promoting the insertion of GTP over GDP in MglA and thus acting as a guanine nucleotide exchange factor (GEF), and second, acting as a primary determinant for the polar targeting of MglAGTP (Szadkowski et al, 2019). This model is attractive and inspired by other polarity regulations, for example in yeast (Chiou et al, 2017). There are however a number of reported inconsistencies that question its validity: - First, the mechanism of MglA polar localization is not well understood. While it is clear that an interaction between RomRX and MglA GTP is required for localization at the leading cell pole, these proteins only colocalize during a short window of time after a reversal. Once the cell starts moving in the new direction, the RomRX complex relocalizes gradually at the lagging cell pole but MglA GTP remains bound to the leading pole (Guzzo et al, 2018; Szadkowski et al, 2019; Herrou & Mignot, 2020) . This clearly suggests that other factors maintain MglA at the leading cell pole once its localization is established. There is also the conundrum that antagonistic GAP and GEF systems co-localize at the lagging pole once polarity is established, which could only be explained by additional layers of regulations. - Second, the GEF activity of RomRX remains questionable because conventional GEFs generally operate by first forming a complex with the GDP-bound GTPase, leading to nucleotide dissociation, and then allowing GTP-binding to displace the GEF (Bos et al, 2007; Cherfils & Zeghouf, 2013) . In the RomRX system, RomR binds to RomX, which in turn interacts with MglAGTP. This ternary complex has a higher affinity for GTP, but as it cannot form in the presence of MglA GDP (Szadkowski et al, 2019) . However, GEFs generally bind to both the GTP and GDP-bound states of their substrates with equal affinity, favoring the detachment of nucleotides irrespective of their specificity, the directionality of the exchange being dependent on the relative concentrations of the nucleotides (Bos et al, 2007; Hennig et al, 2015; Cherfils & Zeghouf, 2013) . Therefore, the strict interaction between RomX interacts and MglA GTP is inconsistent with a GEF activity and requires a re-examination of the function of RomX and RomR in vitro. - Third, although all studies converge to show that MglB activates GTP hydrolysis, its function as a canonical GAP is also questionable. First, MglB binds to MglA in a 2:1 stoichiometry, a unique configuration for GTPase/GAP complexes. This binding positions a critical arginine of MglA (Arg53) into the GTP binding site, remodeling it into a catalytic site (Baranwal et al, 2019; Miertzschke et al, 2011). While this explains how MglB promotes GTP hydrolysis, there are also a number of unexpected additional regulations that arise from the asymmetry of the MglB dimer : i) the binding of additional regulators on one of the MglB .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 5 monomers, i.e. the GltJ ZnR domain (Attia et al, 2024), and perhaps RomY, which activates the GAP activity (Szadkowski et al , 2022) ; and ii), the association of MglA with a C-terminal helix of one of the MglB monomers (Baranwal et al, 2019; Chakraborty et al, 2024). This feature is especially intriguing because the Ct-helix prevents the dissociation of MglA GDP following hydrolysis (Baranwal et al, 2019). This property is not canonical for GAPs, which are known to dissociate from the cognate GDP-bound GTPase (Cherfils & Zeghouf, 2013). In fact this interaction is very specific of the interaction with MglA, because MglB can also act as a GAP for SofG, another small GTPase in M. xanthus , but in this case it dissociates from SofG GTP (Kanade et al, 2021). In vivo, the MglB Ct-helix is critical for the unipolar localization of MglB because in its absence, MglA, RomR and MglB all co-localize at both cell poles (Baranwal et al , 2019) . The molecular mechanism for this localization pattern is not clear but it shows that regulating the interaction between MglA and MglB is essential. In summary, the Ct helix prevents the MglA-MglB complex from dissociating, which is the main basis of the current polarity model. The observations described above call for in-depth quantitative analyses of the interactions between MglA and its regulators MglB and RomRX. In this study, we have developed a new fluorescence anisotropy assay to study these interactions. By combining these tools with an enzymatic assay that measures GTP hydrolysis, we elaborated a new enzymatic model of the MglB-dependent MglA GTP hydrolysis pathway. The model revealed the need of an additional partitioning system, which we identified as the RomRX complex. Rather than functioning as a GEF, it either interacts with MglA-GTP, acting as an effector, or prevents the formation of an MglA-MglB complex through a competitive interaction with MglB. The results suggest a model in which RomRX functions as a partitioning factor, promoting the localization of MglA at the leading cell pole immediately after reversals while preventing its localization at the lagging cell poles by interacting with MglB. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 6

Analysis of MglA nucleotide-binding specificities by fluorescence anisotropy Previous work analyzed MglA nucleotide exchange in vitro using non-hydrolysable mant-GTP analogues (m- GMPPNP or m -GppNHp) or radioactive nucleotides (Baranwal et al, 2019; Chakraborty & Gayathri, 2024; Chakraborty et al, 2024; Zhang et al, 2010) of low sensitivity. The affinity of GTP for MglA could not be measured because no signal could be detected, and the affinity for GDP was not accurate because of weak and unstable signals in nucleotide exchange assays. Their binding to MglA was detected in the presence of MglA’s interacting partners; however, their binding was unstable, as the fluorescence curve showed a decrease at steady state, hindering accurate measurement of nucleotide exchange dynamics (Baranwal et al, 2019; Chakraborty & Gayathri, 2024; Chakraborty et al , 2024) . We hypothesized that more sensitive fluorescent GTP/GDP derivatives, structurally compatible with MglA, could be used as alternative tools to study the dynamics of MglA nucleotide exchange both in the absence and presence of regulator proteins. Nucleotide analogues conjugated to bright dyes come in various chemistries, with fluorophores covalently bound to the base, ribose, or phosphate group of the nucleotide (Bagshaw, 2001) . Because the structure of MglA bound to GTP suggests that modification of the GTP ribose should not interfere with its interaction with MglA, we selected the 2’/3' -O-(2-Aminoethyl-carbamoyl)-Guanosine-5'-tri/diphosphate GTP/GDP analogues, labeled with ATTO-488 and referred to as GTP488 or GDP488 in the remainder of the study. In these molecules, the ATTO-488 dye is attached to GTP/GDP via a long and flexible linker that allows various orientations relative to MglA and positions the dye away from the nucleotide-binding pocket (Appendix Fig. S1). The low molecular weight of these fluorescent nucleotides (MW < 1.5 kDa) relative to MglA (MW ≈ 22 kDa) and the long fluorescence lifetime of ATTO-488 (τfl = 4.1 ns) allow their binding to be monitored over time using fluorescence anisotropy (Colombo et al , 2021) . Anisotropy measurements are not sensitive to photobleaching, which allows for the recording of signals over long periods of time. Purified MglA is found to have a GDP in its active site pocket (Baranwal et al , 2019) . Addition of GTP 488 or GDP 488 to MglA (MglAGDP) leads to an increase of fluorescence anisotropy signals to plateau values corresponding to steady- state (Fig. 1A). At initial time points, the rate limiting step of the reaction is the dissociation of GDP from MglA, which is similar in both conditions (k off GDP). The higher plateau value in the presence of GDP 488 indicates a higher affinity of GDP 488 compared to GTP 488. Because MglA does not exist in a nucleotide-free state, it is only possible to determine the relative affinity of GTP 488 or GDP488 for MglA. To this aim, we first incubated MglA GDP with GTP488 or GDP 488 and allowed the reaction to reach equilibrium and subsequently added an excess of the corresponding unlabeled nucleotide (100 µM) and monitored the decrease in fluorescence anisotropy (Fig. 1B and C) . The corresponding decay reaction is a measure of the dissociation rate of the fluorescent nucleotide k off. Fitting the data to exponential binding equations (see methods) allowed us to measure k off GTP488 = 2.8×10 -3±5.6×10-4s-1 and k off GDP488= 6.0×10 -4±8.7×10-5s-1, confirming the higher affinity of MglA to GDP 488 than GTP 488. These measurements validate our methodology; moreover, the technique is very sensitive because we could record binding at GTP 488 concentrations as low as 10 nM (Appendix Fig. S2A) . In comparison nucleotide exchange with m-GTP could not be recorded at 100 nM (Baranwal et al, 2019; Chakraborty & Gayathri, 2024; Chakraborty et al, 2024). RomRX does not function as a Guanine nucleotide Exchange Factor (GEF) in vitro Next, we investigated the effect of RomR X in this experimental system. Addition of RomX led to higher values of anisotropy at steady -state in the presence of GTP 488, but not in the presence of GDP 488. This result indicates an increase in the molecular weight of MglA GTP488 specifically, due to it s binding to RomX. In contrast, RomR did not interact directly with either form of MglA (Fig. 2A-B). When RomR and RomX were added together with MglA GTP488, a ternary complex formed as evidenced by an additional increase in the fluorescence anisotropy sign al (Fig. 2A). The binding mode of the RomRX complex is thus clearly different .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 7 from that expected for a GEF and more similar to the binding of an MglA effector. To test this directly, we compared RomX binding to that of a known MglA effector, the Ct Tetratr icopeptide Repeat triad of the SgmX Type -IV pilus activator (SgmX -Ct) (Mercier et al , 2020) . Similarly, SgmX-Ct binds only to MglAGTP488 but not to MglAGDP488 (Fig. 2C-D). We could determine that MglA GTP488 binds RomX with a K D = 1.7 ± 0.3 µM, which was slightly decreased to K D = 1.0 ± 0.4 µM upon addition of RomR (Fig. 2E-F). MglAGTP488 binds SgmX-Ct with a slightly lower K D = 0.6 ± 0.3 µM, but in the same order of magnitude (Fig. 2G). Thus, MglAGTP488 binds RomX and SgmX-Ct by similar modes. Small G-protein effectors tend to stabilize the GTP-bound state of the cognate G-protein, which might explain why RomRX was proposed to facilitate nucleotide exchange (Szadkowski et al, 2019). If this was correct, this effect should also be observed with SgmX-Ct. We tested that possibility by conducting chase experiments, preloading MglA with GDP 488 or GTP 488 and then adding an excess of unlabelled GTP to monitor rates of nucleotide release. We did not detect any effect of RomX, RomR or RomRX on the rate of GDP 488 to GTP exchange (Fig. 2H) but we observed that binding to RomX stabilizes the binding of GTP 488 on MglA (Fig. 2I). In the presence of RomX, we measured a k off GTP488 = 8.4×10-4 ± 3.3×10-4s-1, compared to 2.8×10-3±5.6×10- 4s-1 in its absence (Fig. 2I and Table 1) . This effect is even more pronounced in the presence of RomRX (koff GTP488 = 2.4×10 -4 ± 5.5×10 -5 s-1). Most importantly, SgmX-Ct also stabilized MglA GTP488 and to similar levels (koff GTP488 = 2.7×10-4 ± 2.4 ×10 -5s-1) (Appendix Fig. S2B and Table 1) . Thus, we conclude that RomX and the RomRX complex are bona fide MglA effectors like SgmX-Ct and not an MglA GEF as originally proposed. MglB is an atypical MglA GTPase-activating protein We next used the anisotropy assay to explore the interaction between MglA and MglB. Addition of GTP 488 to a mix of MglA GDP and MglB led to higher anisotropy levels than observed in the absence of MglB, demonstrating that an MglA GTP488-MglB complex is formed (Fig. 3A) . However, after reaching a maximal value, the anisotropy signal decreased with slow kinetics. This decrease must corresp ond to the hydrolysis of the GTP 488 nucleotide bound to MglA into GDP 488 driven by the GAP activity of MglB, because it was not observed when an MglA mutant (MglA Q82L) that is resistant to hydrolysis was co -incubated with MglB (Appendix Fig. S2C and D ). The persistently high anisotropy signal also suggests that MglA GDP488 remains attached to MglB. To prove this, we incubated MglA GDP with MglB and added GDP 488. We observed a sharp increase in the anisotropy signal indicating that MglB also interacts with MglA GDP488 (Fig. 3A). The signals observed with MglAGTP488 and MglAGDP488 converged over time, confirming an MglB-induced GTP hydrolysis (Fig. 3A inset) . Thus, consistent with findings by Baranwal et al., MglB behaves as a non-typical GAP, activating GTP hydrolysis but not dissociating from MglA GDP. This result was further confirmed with an MglB version lacking the Ct helix (MglB ΔCt), which bound MglA GTP488 but not MglA GDP488 (Appendix Fig. S2E and F). MglB has also been proposed to facilitate GTP exchange upon binding to MglA (Baranwal et al , 2019; Chakraborty et al, 2024), which might again be due to an effector-type interaction. To verify this result, we performed chase experiments with unlabeled nucleotides in the presence of MglB or MglB ΔCt (Fig. 3B and C). Consistent with previous observations, we found that MglA exhibits enhanced affinity for GTP 488 than for GDP488 in the presence of MglB (k off GTP488 = 9.0×10 -4 ± 1.0×10 -4 s -1 and k off GDP488 = 1.2×10 -3 ± 2.6×10 -4 s -1) (Table 2). We also used isothermal titration calorimetry (ITC) to measure MglA affinities for MglA GTPγS (a non-hydrolysable form of MglA GTP) and MglAGDP in the presence of MglB. We observed as expected a lower dissociation constant K D = 310 ± 180 nM for MglA GTPγS than for MglA GDP, where K D = 751 ± 111 nM (Fig. 3D and E). Furthermore, the Ct helix of MglB stabilizes the MglA GTP-MglB complex, favoring GTP binding (Fig. 3B and Table 2). This stabilization is further evidenced by the lower K D of 150 ± 100 nM for the interaction between MglBΔCt and MglAGTPγS (Fig. 3F). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 8 A new enzymatic model of MglA and MglB interaction The results presented above highlight the function of RomRX as an MglA effector and confirm that MglB does not dissociate from MglA GDP, which deeply questions the prevailing model. Similarly, MglB cannot be called a GEF (Baranwal et al, 2019; Chakraborty et al, 2024) because it does not dissociate from MglA GTP to which it binds with high affinity. Thus, we developed an enzymatic model based on the current knowledge of the system to elucidate the exact properties of the MglA-MglB complex. In this model, MglA binds with GTP or GDP, forming a binary complex. Each of these complexes can also bind to MglB, forming MglAGTP- MglB or MglA GDP-MglB ternary complexes. Each ternary complex can spontaneously exchange nucleotides, but also hydrolyse GTP leading to the release of inorganic phosphate (Pi) without complex dissociation (Fig. 4A). The differential equations describing the model are consistent with a rapid equilibrium random fixation mechanism (Ivanetich & Goold, 1989) (Extra Methods ). The term “rapid equilibrium” indicates that the substrates and products binding and release occur much faster than the catalytic conversion step. The system involves two reactants (i.e. the substrate; GTP and the activator; MglB) and two products (i.e. P i and GDP). However, in our case GDP remains in the nucleotide pocket until the second round of fixation. This model predicts that varying [GTP] or [MglB] results in linear regressions in double reciprocal Lineweaver-Burk (LW) plots (1/ν vs. 1/[MglB] or 1/[GTP]) consistent with Henri-Michaelis-Menten kinetics, intersecting at the same value left of the 1/ν axis. Seconda ry plots of the y intercepts and slopes from the primary lineweaver- burk plots, plotted as function of 1/[GTP] or 1/[MglB], should also be linear. To test this model experimentally, we measured GTP hydrolysis by determining Pi release kinetics using an established enzyme- coupled assay (Webb, 1992) and constructed LW plots and their secondary plots by varying [MglB] or [GTP] (Fig. 4B and EV1A). The obtained experimental curves aligned remarkably well with the rapid equilibrium random fixation model for the MglA-MglB complex, allowing determination of the system constants (Fig. 4B, Table 3). To further understand the function of the MglB Ct helix, we also constructed a model analyzing the behavior of the system when this helix is absent, making the interaction of MglB with MglA specific to its GTP -bound form (Appendix Fig. S2E and F ). In this scenario, the system becomes coherent with a rapid equilibrium ordered fixation mechanism (Posner et al, 1992), where the MglA-MglB ΔCt ternary complex forms only upon interaction with MglA GTP. This leads to the dissociation of MglA GDP from MglB following hydrolysis (Fig. 4C). The differential equations for this model are shown in Extra Methods and the LW plots in Fig. 4D and EV1B. The LW plots differ from the random model when 1/v is plotted against 1/[MglB ΔCt] at various fixed GTP concentrations, intersecting on the vertical axis (Fig. 4D) . This indicates no velocity dependence on MglBΔCt concentration at saturating GTP, similar to previous observation (Baranwal et al , 2019) and characteristic of a rapid equilibrium ordered mechanism with GTP binding before MglBΔCt. The kinetic parameters from the two enzymatic models are remarkably consistent with the affinity values measured by ITC as one falls within the u ncertainty range of the other. (compare affinities calculated for MglAGTP and MglB K’ B = 370 ± 70 nM vs 310 ± 180 nM by ITC, and for MglAGDP and MglB, KB = 630 ± 240 nM vs 751 ± 111 nM by ITC). In addition, the model is also consistent with a stabilizing effect of MglB on MglA-GTP (compare KS [GDP] = 2.45 ± 1.43 µM in the presence of MglB to K’ S [GDP] = 4.1 ± 1.6 µM in its absence), following the same trend as in the anisotropy exper iments. The measurements also confirm that the deletion of the MglB Ct helix increases the affinity to MglA -GTP (compare K ’ B = 370 ± 70 nM for MglB ΔCt versus K’ B = 130 ± 100 nM for MglB, which is very similar to K D = 150 ± 100 nM, measured by ITC). The measured k cat values were similar for MglB ΔCt (2 ×10 -3 ± 5.7 ×10 -5s–1) compared to MglB (2.67×10-3 ± 1.1 ×10-4 s–1), which are both very low compared to canonical GAPs. For example, p120 GAP stimulates the GTPase reaction of Ras with a k cat ~19 s -1 under saturating conditions (Wittinghofer et al , 1997) . This is .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 9 consistent with measurements by Galicia et al. (Galicia et al , 2019) and suggests that MglB may require interacting proteins to fully activate or enhance its GAP activity (ie RomY, see discussion). Taken together, our results demonstrate that MglA -RomRX and MglA -MglB interactions do not correspond to canonical GEF and GAP intera ctions. The enzymology model above should therefore be taken into consideration in any new hypothesis attempting to explain their function in cell polarization (See below and Discussion) RomR acts on MglB to inhibit the formation of the MglA-MglB complex In vivo, MglB is required to exclude MglA from the lagging pole, which was previously assumed to be linked to its GTPase activity. The observation that MglA, whether bound to GTP or GDP, forms a strong complex with MglB that does not dissociate upon GTP hy drolysis raises the question of the involvement of additional regulators. We first tested how effectors such as SgmX and RomX compete with formation of the MglA - MglB complex. We found that even at high concentrations (up to 30 µM), these proteins were not able to dissociate the complex (Fig. 5A and Appendix Fig. S2G-J). Thus, effector competition may not be sufficient to dissociate MglA -MglB. Given that RomR localizes at the lagging pole and directly interacts with MglB (a lthough there are conflicting evidence for this interaction, which has recently been challenged by the same group (Keilberg et al, 2012; Carreira et al, 2023)), we tested whether the addition of RomR affects the MglA GTP-MglB complex. Addition of 30 µM RomR led to an abrupt decrease in fluorescence anisotropy signal (Fig. 5B) indicating that RomR triggers the dissociation of the complex. Adding 30µM RomR also disrupted the MglA GDP-MglB complex (Fig. 5C). Since RomR does not interact with MglA, it likely competes with MglA for binding to MglB, thereby dissociating the complex and reducing the available MglB for MglA interaction. There is conflicting data regarding a potential direct interaction between RomR and MglB, as some pull-down experiments detect this interaction while others do not (Keilberg et al , 2012; Carreira et al , 2023) . We revisited this potential direct interaction using ITC and found a low affinity interaction with a K D of 31 ± 15 µM (Fig. 5D). Microscale thermophoresis also confirmed these interactions, giving a K D value in the same order of magnitude (~ 59 µM, Appendix Fig. S3A). Previous studies (Szadkowski et al , 2019) have shown that RomR does not affect the GTPase activity of MglA when added at equal concentration to MglB. However, due to the low affinity interaction between RomR and MglB, significant changes in activity should not be expected unless RomR is present at high concentrations. To test this prediction, we conducted GTPase assays in the presence of varying concentrations of MglB and a fixed concentration of RomR (20 µM) (Fig. 5E and Appendix Fig. S3B and C ). Michaelis- Menten and Lineweaver-Burk plots show that RomR significantly reduced the reaction rate, mirroring the effect observed at lower effective concentrations of MglB (an approximately twofold decrease in enzyme efficiency 𝒌𝒄𝒂𝒕 𝒂𝒑𝒑/𝑲𝒎 𝒂𝒑𝒑) (Fig. 5E and Table 4). This shows that when the concentration of RomR is close to the dissociation constant of the MglB-RomR complex, it effectively competes with the formation of the MglA- MglB complex. In vitro reconstitution of MglA exclusion by the RomR -MglB complex in a cell -like configuration The results above sugg est that RomR modulates the interaction between MglA and MglB and thus controls the turnover rate of the MglA -MglB complex in vivo . However, because the affinity between MglB and RomR is low, this effect is only expected if there is sufficient accumulation of RomR to compete with the formation of the high affinity MglA -MglB complex. Such accumulation is possible in vivo because most of the RomR pool relocalizes with RomX to the lagging pole ( Fig. 6A and B ). To test this possibility in .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 10 conditions that mimic the cellular context closely, we formed individual spherical droplets of 10 to 20 µm - diameter that reproduce the concave shape of the bacterial poles, although the radii of curvature are larger and constant around the droplets (Fig. 7A and Appendix Fig. S4A ).The lipids stabilizing the droplet interface were selected to be similar in composition to the inner membrane of M. xanthus. We fluorescently labeled MglA, MglB, and RomR. MglA was either labeled directly ( Appendix Fig. S4B ) or its localization was tracked using bound fluorescent nucleotides as a label (Fig. 7B-C and Appendix Fig. S4C). As shown by the anisotropy measurements, fluorescent nucleotides remain stably bound to MglA for at least 20 minutes, justifying colocalization of fluorescent nucleotides and MglA for the duration of these experiments. MglA (in its GTP or GDP states) and RomR localized diffusely in these compartments at any concentration (Fig. 7B-D and Appendix Fig. S4B-C and S5A). In contrast, MglB localized in large domains formed at the lipid/water interface (Fig. 7E and Appendix Fig. S5B ). Control experiments in the absence of lipids did not reveal any MglB binding ( Appendix Fig. S6 ). This observation is consistent with previous work showing that MglB binds weakly to liposomes (Galicia et al , 2019) . The characteristic localization of MglB in the droplet provided a good opportunity to characterize protein-protein interactions in this cell-like configuration using two-color imaging. Mixing red -labeled MglB with green -labeled MglA GTP488 or MglA GDP488 did not modify the localization of MglB but led to the recruitment of MglA at the lipid/water interface with MglB (Fig. 7F-G). This observation demonstrates that MglB and MglA interact in the droplet assay and that this interaction does not interfere with MglB lipid-binding. Mixing red-labeled MglB with RomR -GFP also led to the recruitment of RomR -GFP at the droplet interface, indicating that this interaction is also detectable in this system (Fig. 7H). Importantly, mixing red-labeled MglB with unlabeled R omR and MglAGTP488 led to diffuse localisation of MglA (Fig. 7I). This is due to the formation of the MglB -RomR complex because RomR -GFP was recruited to the lipid interface together with MglB when all three proteins were mixed and MglA was not labeled (Fig. 7J). Thus, the RomR-MglB complex effectively competes with the MglA-MglB complex in oil droplets.

The polarity axis formed by MglA and MglB has long been considered to be formed via an exclusion mechanism linked to the activation of GTP hydro lysis at the lagging cell pole. This mechanism was proposed based on strong evidence: (i) in vitro studies on MglA and MglB homologs purified from T. thermophilus (MglBtt) showed that MglB tt only binds to MglA GTPtt and (ii), in vivo, an MglA Q82A/L that cannot hydrolyze GTP, even in the presence of MglB, localizes symmetrically in M. xanthus cells (Treuner-Lange et al, 2015). Surprisingly, when Baranwal et al. (Baranwal et al, 2019) elucidated the structure of the M. xanthus MglA- MglB complex, they found that MglB remains associated to MglA GDP after hydrolysis, which is due to the presence of a Ct-helix, a motif that is not present in MglB tt. This finding raised profound questions about how cell polarity is established via MglA and MglB and prompted this study. We re-examined how MglA interacts with MglB and how this affects the MglA hydrolysis cycle. We discovered that the MglA system is not controlled by a canonical GAP/GEF system as originally proposed, but is centrally regulated by the RomRX complex, which acts both as an MglA effector and a regulator of the MglA-MglB interaction. We discuss below how these interactions could lead to the spatial partitioning of MglA in the Myxococcus cell. Cell polarity is dictated by a complex interaction network involving MglA, RomRX and MglA effectors at the leading pole, and MglA, MglB, RomRX, RomY and MglC at the lagging cell pole. Carreira et al. (Carreira et al, 2023) proposed that cell polarity could be established if interactions between RomR and MglB could be blocked at the leading pole and conversely formed at the lagging pole. Combined with the recent discovery of MglC as a localization factor of MglB, this study provides molecular roots to the model of Carreira et al. and suggests the following scenario: i) At the leading cell pole, RomRX functions as an MglA effector, recruiting MglA GTP. The high concentration of MglA GTP prevents formation of the MglB -MglC complex and thus excludes MglB from the .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 11 leading cell pole. The interaction of MglA with RomRX is transient and most critical immediately after the polarity switch, as RomRX relocalizes to the lagging pole within a few minutes. During this time, MglA GTP associates with polar effectors such as SgmX for the Type -IV pili and the GltJ ZnR -GYF motif for the Agl - Glt comp lex. In these complexes, MglA adopts a third state insensitive to the action of MglB, further reinforcing polar localization (Galicia et al, 2019) (Fig. 8). ii) At the lagging pole, the spatial localization of GAP activity is essential to establish the polarity axis and the disassembly of the Agl -Glt complex (Attia et al, 2024). At this pole, MglB is recruited by MglC, which is favored by the low concentration of MglA (Carreira et al, 2023). However, since MglB does not dissociate from MglAGDP, a dissociation factor is needed to ensure MglA release following nucleotide hydrolysis. Based on direct in vitro evidence, we propose that the RomRX complex is the dissociation factor, acting between reversals when its concentration is high at the lagging pole (Fig. 8). This hypothesis is strongly supported by the co-localization of MglA and MglB at both cell poles observed in a romR mutant (Zhang et al, 2012b). Thus, functioning as a polar effector, RomRX first ensures that MglA GTP localizes to the leading pole, while blocking access to the lagging pole by forming a MglB -RomRX complex. Consequently, the common view that the Mgl system is a prokaryotic version of a clas sical small G -protein system, regulated by spatially localized GAP (MglB) and GEF (RomRX), is obsolete. So how can this system function without a bone fide GEF and GAP system? Considering all available data, a view emerges that distinct signaling hubs, one formed by MglA (which is insensitive to the action of MglB (Galicia et al, 2019)), the other formed by MglB, segregate at opposite cell poles. The lagging pole system is formed by numerous interactions, between RomR and MglC, MglC and MglB (Carreira et al, 2023) and now MglB and RomR. This system is thus the site of numerous regulations: as discussed above it can be destabilized by RomRX and there are likely a number of additional regulators acting on it. There are currently two examples of such regulation: (i) RomY which activates the GAP activity and may be critical to accelerate the reaction in vivo because the kcat of MglB alone is exceedingly low. (ii) Any regulation that modulates interaction of the MglB Ct -helix with MglA converts MglB into a bona fide GAP (Chakraborty et al , 2024) . Such regulation occurs in the Agl-Glt complex where the GltJ ZnR domain recruits and converts it into a GAP via a direct action on this helix (Attia et al, 2024). Since MglB does not dissociate from MglA, it can n either be called a GAP nor a GEF. Remarkably, the intervention of a dissociation factor could direct the reaction towards either of these functions, depending on how fast the dissociation occurs and whether it competes with or allows GTP hydrolysis (which again is a slow step). Since RomR dissociates both GTP and GDP -bound MglA from MglB it could help direct either reaction depending on the context. In the cell it is clearly the GAP that dominates (possibly due to the presence of RomY, (Szadkowski et al, 2022) at the lagging pole between reversals. The notable asymmetry of the complex formed between MglA and the MglB dimer, along with recent findings that the Ct-helix of the proximal MglB protomer is sufficient to stabilize MglA in its GDP- and GTP-bound states (Chakraborty et al, 2024), suggests that the second MglB protomer provides an important binding interface for additional regulators (Szadkowski et al, 2022). In combination with RomR, these regulators could thus functionalize the complex as a GAP or a GEF. This emerging picture makes it very difficult to propose a precise reversal mechanism at this stage. We propose that when they become phosphorylated, FrzX and FrzZ each act on the polar signaling complexes to release MglAGTP and allow its relocalization at the lagging pole. The molecular targets of these regulators are yet to be discovered and could involve yet unidentified polar proteins. Nevertheless, this work makes apparent that modulating the partitioning function of Rom RX might be a key. Indeed, it can be imagined that promoting the formation of MglA -MglB complexes by alleviating the inhibition of RomRX could trigger a domino effect leading to the release of MglB from MglC and its relocalization at the opposite pole. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 12 Recently, structural homologs of MglA and MglB have been found in Asgard Archaea , demonstrating an ancient origin of these systems (Tran et al., 2024). Remarkably, Thor-Rab, the MglA (and Rab) homolog, also contains an arginine residue positioned similarly to the MglA Arg53, suggesting that interaction with the archaeal MglB counterpart, Thor -RB, might lead to a Thor complex functioning with similar enzymatic properties as MglAB. Thus, in all three domains of life, Roadblock domain -containing proteins (as well as the structurally related Longins) may form dimeric platforms that often directly interact with small GTPases (Koonin & Aravind, 2000; Levine et al, 2013; Liu et al, 2021). In eukaryotic cells, these proteins primarily function as GEFs, but occasionally also as GAPs, through allosteric mechanisms or by incorporating additional regulatory components (Levine et al., 2013; Jansen & Hurley, 2023; Cabrera et al., 2014; Miertzschke et al., 2011). One notable example is the Ragulator complex in the mTOR pathway, where multiple Roadblock domain-containing proteins act as a scaffold for Rag GTPases, again creating a multi- layered regulatory hub crucial for mTOR signaling. As discussed above, the MglB system might also be viewed as such and all these systems could be modular signaling units. In conclusion, studying the MglA- MglB system may reveal universal features of these ubiquitous regulatory modules.

Methods and protocols Protein expression, purification and labeling - Expression and purification of MglA-His6, MglAQ82L-His6, MglB-His6, MglBΔCt-His6, SgmX-Cter and RomX-His6 The gene sequences of mglA, mglAQ82L, mglB, mglBΔCt, sgmX-Cter and romX from M. xanthus were inserted in a pET28a plasmid for expression of C-terminal hexa-histidine tagged proteins. Bacterial strains and primers used in this study are listed in Appendix Tables S1 and S2. Plasmid sequences were verified by Sanger sequencing (Eurofins GATC‐Bio tech, Germany). All constructs were transformed in the BL21(DE3) E. coli strain. Cells were grown in Luria-Bertani (LB) medium at 32°C and induced at OD 600nm 0.5-0.8 for 3 hours by the addition 0.5 mM IPTG (isopropyl-h-d-thiogalactopyranoside). Cells were harvested by centrifugation and pellets were resuspended in buffer A (10 mM Tris-HCl, pH = 7.4, 50 mM NaCl, 5 mM MgCl2 , 10 mM Imidazole, pH = 7.4) supplemen ted with cOmplete™ EDTA -free protease inhibitor cocktail (Roche) and 20 µg/ml DnaseI (Merck). Cells were lysed with a French press at 1 kbar and lysates were centrifuged at 18,000 rpm for 20 minutes at 4 °C. Supernatants were loaded on 2 ml HisTrap beads columns (GE Healthcare) functionalized with Nickel and equilibrated with Buffer A. After 10 mn of incubation, the resin was washed with 5-column volume of Buffer A and 5-column volume of Buffer B (Buffer A containing 75 mM Imidazole). The protein was then eluted with Buffer C (Buffer A containing 200 mM Imidazole). Fractions containing the protein were pooled, concentrated, dialysed overnight in Buffer D (Buffer A without imidazole) and stored at -80 °C in Buffer D. Protein purity was analyzed by SDS-PAGE, and protein concentrations were quantified using NanoDrop™ (Thermo Scientific™). The following extinction coefficients, calculated based on the amino acid sequence of the constructed proteins using ProtParam from ExPASy, were used: 16,515 M -1·cm-1 for MglA-Hi s6 and MglAQ82L-His 6, 2,980 M -1·cm-1 for MglB-His 6 and MglBΔCt-His6, 11,920 M-1·cm-1 for SgmX-Cter-His6, and 8,480 M-1·cm-1 for RomX-His6. - Expression and purification of RomR-His6 and RomRGFP-His6 Synthetic coding sequences of RomR and RomRGFP were obtained from Twist Bioscience, cloned in a modified pET28 vector for expression of C -terminal hexa-histidine tagged proteins . Constructs were transformed in the LEMO21(DE3) E. coli strain. Cells were grown overnight in LB medium at 37°C and transferred into NZY auto-induction LB medium (NZYtech) for 24 hours at 18 °C for overexpression. Cells were harvested by centrifugation and pellets were resuspended in Buffer E (50 mM tris -HCl, pH = 8.0, 300 mM NaCl, 5 mM MgCl 2, 10 mM imidazole, pH = 8.0) supplemented with cOmplete™ EDTA -free protease inhibitor cocktail (Roche) and 20 µg/ml DnaseI. Cells were lysed with a French press at 1 kbar and lysates .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 13 were centrifuged at 18,000 rpm for 20 minutes at 4°C. Supernatants were loaded on a 2 ml HisTrap beads column (GE Healthcare) functionalized with Nickel and equilibrated with a Buffer E. After 10 mn of incubation, the resin was washed with 5 -column volume of Buffer E and 5 -column volume of Buffer F (Buffer A containing 75 mM Imidazole). The protein w as then eluted with Buffer G (Buffer A containing 200 mM Imidazole). Fractions containing the protein were pooled, concentrated, dialyzed overnight in Buffer H (Buffer A without imidazole) for ITC and MST experiments or dialyzed in Buffer D for all other e xperiments and stored at -80°C. Protein purity was analyzed by SDS -PAGE, and protein concentrations were quantified using NanoDrop™ (Thermo Scientific ™). The following extinction coefficients were used: 6990 mol -1·cm-1 for RomR-His6 and; 26025 mol-1·cm-1 for RomRGFP-His6. - Labeling of MglA and MglB Purified MglAGDP and MglB were labeled overnight at 4°C under gentle agitation in the presence of a 5-fold excess of Alexa Fluor™ 488 C5 -maleimide (ThermoFisher scientific) or Janelia Fluor® 646 Maleimide (TOCRIS) in Buffer E supplemented with 2 mM TCEP. Unbound fluorophore was removed by passing the reaction mixture through a pre -equilibrated protein desaltin g column (Zeba Spin Desalting columns from ThermoFisher scientific). Labeled proteins were analyzed on SDS -PAGE gel and imaged using ChemiDoc imaging system (Bio-Rad) to quantify the fraction of labeling. Labeled proteins were stored at -80°C. Fluorescence Anisotropy experiments - Fluorescent nucleotides analogs Fluorescent nucleotides used in this study were purchased from Jena Bioscience (https://www.jenabioscience.com/), specifically EDA-GTP-ATTO-488 (ref. NU-820-488) and EDA-GDP- ATTO-488 (ref. NU-840-488). When needed, fluorescent nucleotide analogs were diluted in 20 mM HEPES pH 7.5 (Colombo et al, 2021). - Fluorescence anisotropy The binding of fluorescent nucleotides analogs to proteins was monitored by measuring changes in anisotropy using excitation and emission wavelengths of 504 nm and 521 nm, respectively. These experiments were performed using a Safas Xenius XC spectrofluorometer (Safas Monaco) controlled by SP2000 software, version 7.8.13.0 (Colombo et al, 2021). - Nucleotides exchange assays For kinetic experiments, MglA GDP, either alone or in combination with MglB, MglB ΔCt, RomR, RomX, RomRX or SgmX-Ct, was mixed in a volume of 150 µl in Buffer D (supplemented with 2mM TCEP) at the indicated concentrations. Nucleotide exchange was initiated by adding 0.1 µM of the fluorescent nucleotide analog, and the increase in fluorescence anisotropy was recorded over time until equilibrium was reached. The main reaction controlling the rate in this condition is the dissociation of GDP from MglA (k off GDP). Subsequently, exchange kinetics were monitored by competing the bound fluorescent nucleotides with an excess of unlabeled nucleotides (100 µM GTP or GDP), resulting in a decrease in fluorescence anisotropy due to the release of the fluorescence nucleotide from MglA. Here, the primary reaction rate is controlled by the dissociation of the fluorescent nucleotide from MglA (koff N1*). The reactions involved are described by the following equations: 𝑃 + 𝑁 ⇔ 𝑃𝑁 𝑃 +𝑁∗ ⇔ 𝑃 𝑁 ∗ Where P represents MglA GDP or the MglA GDP-based protein complex, N is the unlabeled nucleotide, N * is the labeled nucleotide and PN represents the protein-nucleotide complex. These equations implies four rate constants:kon N, koff N, kon N* and koff N*. The binding rate constants (k on) cannot be determined by this assay, as they are too fast for anisotropy measurements. However, the dissociation rates corresponding to the slowest reactions can be mathematically described by the following differential equation: .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 14 𝑑[𝑃𝑁 𝑜𝑟 𝑃𝑁∗ ] 𝑑𝑡 = − 𝑘𝑜𝑓𝑓. [𝑃𝑁 𝑜𝑟 𝑃𝑁∗ ] This equation can be resolved by fitting an exponential decay function to estimate the koff value. 𝑟(𝑡) = 𝑟𝑚𝑖𝑛 + (𝑟0 − 𝑟𝑚𝑖𝑛)𝑒−𝑘𝑜𝑓𝑓 𝑁1∗.𝑡 Here, r(t) represents the fluorescence anisotropy at time (t), r min is the minimum anisotropy value for the fluorescent nucleotide alone, r 0 is the initial fluorescence anisotropy before adding the excess unlabeled nucleotide and koff N* is the dissociation rate constants of the fluorescent nucleotide. - Steady-state experiment for binding constant determination For binding affinity experiments, changes in the fluorescence anisotropy values of EDA-GTP-ATTO-488 (0.1 µM) were measured in the presence of MglA GDP (2 µM) and increasing concentrations of RomX, RomRX (where RomR and RomX were pre-mixed in equal concentrations), or SgmX-Ct. Each measurement was taken after 30 min when steady state was reached. Data were fitted using Igor Pro (version 9.0.5.1) to the binding equation to estimate KD: 𝑟 = 𝑟𝑚𝑎𝑥 ∗ [𝑃] 𝐾𝐷 + [𝑃] + 𝑟𝑚𝑖𝑛 Where r represents the fluorescence anisotropy readout, r max represents the maximum anisotropy value when all the fluorescent nucleotide is bound, [P] is the protein concentration, K D is the binding constant and r min is the minimum anisotropy value corresponding to the fluorescent nucleotide alone in solution. Results were plotted using Igor Pro. Each data point represents the mean of at least three independent measurements, with error bars indicating the standard deviation. - Protein competition experiment MglAGDP (2 µM) was incubated with MglB (4 µM). The reaction was initiated by adding GTP 488 or GDP 488 (0.1 µM) and fluorescence anisotropy was monitored until steady state. Subsequently, RomR, RomX, RomRX or SgmX-ct was added at the indicated concentrations and measurements were continuously taken. GTPase assay Rates of GTP hydrolysis were measured using the EnzChek Phosphate Assay Kit (E -6646, Thermo Fisher Scientific) based on the method originally described by Webb (Webb, 1992). The assay was performed using a constant concentration of MglA GDP (2 μM) and varying concentration of MglB, MglB ΔCt or RomR-RomX complex, in the presence of 2 mM TCEP, 0.2 mM MESG, PNP (1 U/ml), and a range of GTP concentration as indicated. The reaction was started by the addition of MglA, and phosphate release was recorded by the change in absorbance at 360 nm using 96-well flat-bottom plates (Greiner Bio-One) and a SPARK multimode microplate reader (Tecan) every 30 seconds for at least 30 min. Experiments were repeated a minimum of three times. An R script was used to perform multiple nonlinear regressions on the experimental data to accurately determine the kinetic constants (K B, K’ B, Ks[GDP] and K’ s[GDP]) by fitting the data to the random or ordered model, and to calculate the coefficient of determination (R²) to assess the goodness of fit. Isothermal Titration calorimetry Isothermal titration calorimetry (ITC) was employed to determine the thermodynamic parameters and binding constants of the binding interactions between a) MglA GTPλS or GDP and MglB; b) MglA GTPλS and MglBΔCt and c) RomR or the RomR-RomX complex and MglB . Case a-b: Experiments were conducted using a MicroCal PEAQ-ITC microcalorimeter (Malvern, UK) at 25°C. Each titration consisted of 19 injections, starting with an initial injection of 0.4 μl followed by 18 injections of 2 μl each. Injections were carried out over 4 seconds with 150 -second intervals between injections. The experiments were carried out in Buffer E Supplemented with 5 mM dithiothréitol (DTT). Purified MglA GDP “macromolecule” was placed in the sample cell, while the ligand MglB or MglB ΔCt was .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 15 loaded into the syringe. Both the cell and the syringe contained 100 μM of GTPλS or GDP. Control experiments involved titrating the ITC buffer into MglA GDP alone to account for heat changes due to dilution. This fitted offset “a constant control heat” was subtracted from the experimental data to accurately calculate the dissociation constant. Data were analyzed using the MicroCal PEAQ -ITC Analysis software, the binding isotherm was fit to a one set of sites model. Case c: Experiments were performed using a MicroCal PEAQ-ITC microcalorimeter (Malvern, UK) at 12.5°C. Each titration consisted of 13 injections, with the first injection of 0.4 μl and the subsequent 12 injections of 3 μl each. Injections lasted for 4 seconds with 150 -second intervals between each injection. The experiments were conducted in Buffer J supplement ed with 5 mM DTT. RomR or the RomR-RomX complex (pre-mixed at equal concentrations) was placed in the sample cell, while MglB was loaded into the syringe. A second series of MglB injections with the same parameters was performed, and the resulting ITC data files were concatenated for analysis as a single data file. Control experiments involved titrating the ITC buffer into RomR alone to account for heat changes due to dilution. This fitted offset “a constant control heat” was subtracted from the experimental data to accurately calculate the dissociation constant. Data analysis was performed using the MicroCal PEAQ-ITC Analysis software. Microscale Thermophoresis Microscale thermophoresis (MST) experiments were conducted using a MonolithX MM-307 instrument (NanoTemper Technologies). MglB was labeled with the Monolith NT Protein Labeling Kit - Red NHS and used at a final concentration of 40 nM. The RomR-RomX complex was titrated in 1:1 serial dilution, ranging from 136 µM to 4 nM. Buffer J supplemented with 5mM DTT and 0.05% tween was used for this experiment. All experiments were performed in standard Monolith Series capillaries (cat# MO-Ko22) with 100 % excitation LED power and medium IR-laser power at 24°C. The spectral shift ratio 670nm/650nm was measured as a function of ligand concentration. Each experiment was performed in duplicate to ensure reproducibility. The dissociation constant (K D) for the interaction was determined by fitting the thermophoresis data using the MO Affinity Analysis software, assuming a 1:1 binding stoichiometry. Labeled MglB 40nM was bound to RomR with increasing concentrations as shown by MST (Appendix Fig.S5A). The MST Fnorm values at 650nM with an on time measure at 2.5 seconds in a dose response manner and a signal to noise ratio of 10.3 was used to calculate the binding constant of MglB/RomR . Buffer PBS + 0.05% Tween. Fluorescence microscopy. For each experiment, 1 ml of a CYE-grown culture with an OD of 0.5 –1 was centrifuged 5 min at 3000 rpm and resuspended to an OD of 2 in TPM CaCl2 buffer (10 mM Tris-HCl, pH 7.6, 8 mM MgSO4, 1 mM KH2PO4, 1mM CaCl2). Then 2µl of the cell suspension was placed on a glass slide, covered with a 1.5% agar pad, and incubated for 20 min at 32°C before imaging. Time-lapse experiments were performed using an automated inverted epifluorescence Ti2E microscope with Perfect Focus (Nikon), a 100x objective, and a back-illuminated Kinetix Scientific CMOS (sCMOS) camera (Teledyne). All fluorescence images were acquired with minimal exposure time to reduce bleaching and phototoxicity. For time-lapse movies, images were captured every 2 seconds over a 15-minute period. Water-in-oil lipid stabilized emulsion The lipids used for this experiment included L- α-phosphatidylethanolamine ( E. coli PE), L-α - phosphatidylglycerol (Egg, Chicken) (sodium salt) (EGG PG), L-α -phosphatidylserine (Brain, Porcine) (sodium salt) (Brain PS), 1',3'-bis[1,2-dioleoyl-sn -glycero-3-phospho]-glycerol (sodium salt) (cardiolipin 18:1) and L-α -lysophosphatidylcholine (Egg, Chicken) (EGG LysoPC), 1,2-dipalmitoyl-sn -glycero-3- phosphoethanolamine-N-(lissamine rhodamine B sulfonyl) (ammonium salt)(PE*). All lipids were purchased from Merck. A lipid solution was prepared in a glass vial by mixing 76 mol% E. coli PE (or 70 mol% E.coli PE + 6 mol% PE* for fluorescent lipids), 4.9 mol% EGG PG, 9.3 mol% cardiolipin 18:1, 6.5mol% PS and 3 mol% Egg .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 16 Lyso PC in chloroform. This composition was chosen to mimic the inner membrane of M. xanthus. The mixture was dried overnight in a vacuum desiccator and then rehydrated in Buffer E to a final concentration of 2mg/ml (≈ 0.83 mM). The resulting suspension was sonicated usi ng a tip sonicator with refrigeration, followed by centrifugation at 1000 rpm for 10 mn to remove titanium particles. The suspensions were stored at 4 °C under nitrogen and used within a week. To prepare water-in-oil emulsions, the lipid suspension was dil uted to a final concentration of 3 µM with the proteins of interest in Buffer E. The oil phase was prepared by dissolving 1 mg/m l cithrol (CRODA) surfactant in Squalene oil (Mreck). To form the emulsion, the water phase was added to the oil phase at a typical volume ratio of 1:20 in a microcentrifuge tube. The mixture was gently mixed by pipetting up and down, avoiding vigorous agitation to maintain emulsion stability. Brief vortex mixing at low speed was used for additional mixing. The successful formation of the droplets was indicated by the uniform cloudiness of the sample. The emulsion was placed in a glass microscope chamber , with droplets positioned between a microscope coverslip and glass slide spaced with double -sided adhesive tape (Appendix Fig.S4A). Droplets and fluorescent proteins were imaged using a Zeiss Axio Observer Z1 microscope equipped with a 100x/1.4NA Oil Ph3 Plan-Apochromat objective and a Hamamatsu ORCA-Flash 4.0LT camera. Images were acquired with Zen 2.3 blue edition.

We acknowledge Jessica Colombo for her help in setting the lipid droplet system. This work was funded by the Centre National de la Recherche Scientifique (CNRS), Aix-Marseille University, the European Research Council to TM (JAWS-Advanced Grant ERC-2019-ADG: 885145). CD was funded by a CENTURI postdoctoral fellowship and by an ERC advanced grant to TM. Author contributions CD, AM, JH and TM designed research. CD performed the experiments. CD and DB-K performed and analyzed ITC experiments, SL performed and analyzed the in vivo data. RV and FD did the first purification of RomR and provided the protocol. CS-K and CD analyzed kinetic experiments and wrote the enzymology model. JH assisted with the initial setup of experiments and contributed valuable insights for interpreting the results. CD, AM and TM analyzed the data, interpreted the results, and wrote the manuscript. TM acquired funding. Disclosure and competing interest statement The authors declare that they have non-conflict of interest Data Availability Section All data porting this study are included in the main text, EV figures and Appendix. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 17

Allam AH, Charnley M & Russell SM (2018) Context-Specific Mechanisms of Cell Polarity Regulation. Journal of Molecular Biology 430: 3457–3471 Attia B, My L, Castaing JP, Dinet C, Le Guenno H, Schmidt V, Espinosa L, Anantharaman V, Aravind L, Sebban-Kreuzer C, et al (2024) A molecular switch controls assembly of bacterial focal adhesions. Science Advances 10: eadn2789 Bagshaw C (2001) ATP analogues at a glance. J Cell Sci 114: 459–460 Baranwal J, Lhospice S, Kanade M, Chakraborty S, Gade PR, Harne S, Herrou J, Mignot T & Gayathri P (2019) Allosteric regulation of a prokaryotic small Ras-like GTPase contributes to cell polarity oscillations in bacterial motility. PLoS Biol 17: e3000459 Bautista S, Schmidt V, Guiseppi A, Mauriello EMF, Attia B, Elantak L, Mignot T & Mercier R (2023) FrzS acts as a polar beacon to recruit SgmX, a central activator of type IV pili during Myxococcus xanthus motility. The EMBO Journal 42: e111661 Bos JL, Rehmann H & Wittinghofer A (2007) GEFs and GAPs: Critical Elements in the Control of Small G Proteins. Cell 129: 865–877 Carreira LAM, Szadkowski D, Lometto S, Hochberg GKA & Søgaard-Andersen L (2023) Molecular basis and design principles of switchable front-rear polarity and directional migration in Myxococcus xanthus. Nat Commun 14: 4056 Carreira LAM, Szadkowski D, Müller F & Søgaard-Andersen L (2022) Spatiotemporal regulation of switching front–rear cell polarity. Current Opinion in Cell Biology 76: 102076 Carreira LAM, Tostevin F, Gerland U & Søgaard-Andersen L (2020) Protein-protein interaction network controlling establishment and maintenance of switchable cell polarity. PLOS Genetics 16: e1008877 Chakraborty S & Gayathri P (2024) RomX, a novel prokaryotic regulator, links the response receiver domain of RomR with GTP-bound MglA for establishing Myxococcus xanthus polarity. doi:10.1101/2024.02.20.581209 [PREPRINT] Chakraborty S, Kanade M & Gayathri P (2024) Mechanism of GTPase activation of a prokaryotic small Ras-like GTPase MglA by an asymmetrically interacting MglB dimer. Journal of Biological Chemistry 300: 107197 Cherfils J & Zeghouf M (2013) Regulation of Small GTPases by GEFs, GAPs, and GDIs. Physiological Reviews 93: 269–309 Chiou J-G, Balasubramanian MK & Lew DJ (2017) Cell Polarity in Yeast. Annu Rev Cell Dev Biol 33: 77–101 Colombo J, Antkowiak A, Kogan K, Kotila T, Elliott J, Guillotin A, Lappalainen P & Michelot A (2021) A functional family of fluorescent nucleotide analogues to investigate actin dynamics and energetics. Nat Commun 12: 548 Dinet C & Mignot T (2023) Unorthodox regulation of the MglA Ras-like GTPase controlling polarity in Myxococcus xanthus. FEBS Lett 597: 850–864 Faure LM, Fiche J-B, Espinosa L, Ducret A, Anantharaman V, Luciano J, Lhospice S, Islam ST, Tréguier J, Sotes M, et al (2016) The mechanism of force transmission at bacterial focal adhesion complexes. Nature 539: 530–535 Galicia C, Lhospice S, Varela PF, Trapani S, Zhang W, Navaza J, Herrou J, Mignot T & Cherfils J (2019) MglA functions as a three-state GTPase to control movement reversals of Myxococcus xanthus. Nat Commun 10: 5300 Goryachev AB & Leda M (2019) Autoactivation of small GTPases by the GEF–effector positive .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 18 feedback modules. F1000Res 8: F1000 Faculty Rev-1676 Guzzo M, Murray SM, Martineau E, Lhospice S, Baronian G, My L, Zhang Y, Espinosa L, Vincentelli R, Bratton BP, et al (2018) A gated relaxation oscillator mediated by FrzX controls morphogenetic movements in Myxococcus xanthus. Nat Microbiol 3: 948–959 Hennig A, Markwart R, Esparza-Franco MA, Ladds G & Rubio I (2015) Ras activation revisited: role of GEF and GAP systems. Biological Chemistry 396: 831–848 Herrou J & Mignot T (2020) Dynamic polarity control by a tunable protein oscillator in bacteria. Current Opinion in Cell Biology 62: 54–60 Houston DW (2017) Cell Polarity in Development and Disease Iden S & Collard JG (2008) Crosstalk between small GTPases and polarity proteins in cell polarization. Nat Rev Mol Cell Biol 9: 846–859 Ivanetich KM & Goold RD (1989) A rapid equilibrium random sequential bi-bi mechanism for human placental glutathione S-transferase. Biochimica et Biophysica Acta (BBA) - Protein Structure and Molecular Enzymology 998: 7–13 Kanade M, Singh NB, Lagad S, Baranwal J & Gayathri P (2021) Dual specificity of a prokaryotic GTPase-activating protein (GAP) to two small Ras-like GTPases in Myxococcus xanthus. FEBS J 288: 1565–1585 Keilberg D, Wuichet K, Drescher F & Søgaard-Andersen L (2012) A Response Regulator Interfaces between the Frz Chemosensory System and the MglA/MglB GTPase/GAP Module to Regulate Polarity in Myxococcus xanthus. PLOS Genetics 8: e1002951 Kirkpatrick CL & Viollier PH (2011) Poles Apart: Prokaryotic Polar Organelles and Their Spatial Regulation. Cold Spring Harb Perspect Biol 3: a006809 Koonin EV & Aravind L (2000) Dynein light chains of the Roadblock/LC7 group belong to an ancient protein superfamily implicated in NTPase regulation. Curr Biol 10: R774-776 Kühn MJ, Macmillan H, Talà L, Inclan Y, Patino R, Pierrat X, Al‐Mayyah Z, Engel JN & Persat A (2023) Two antagonistic response regulators control Pseudomonas aeruginosa polarization during mechanotaxis. The EMBO Journal 42: e112165 Laloux G & Jacobs-Wagner C (2014) How do bacteria localize proteins to the cell pole? J Cell Sci 127: 11–19 Leonardy S, Miertzschke M, Bulyha I, Sperling E, Wittinghofer A & Søgaard‐Andersen L (2010) Regulation of dynamic polarity switching in bacteria by a Ras‐like G‐protein and its cognate GAP. The EMBO Journal 29: 2276–2289 Levine TP, Daniels RD, Wong LH, Gatta AT, Gerondopoulos A & Barr FA (2013) Discovery of new Longin and Roadblock domains that form platforms for small GTPases in Ragulator and TRAPP- II. Small GTPases 4: 62–69 Liu Y, Makarova KS, Huang W-C, Wolf YI, Nikolskaya AN, Zhang X, Cai M, Zhang C-J, Xu W, Luo Z, et al (2021) Expanded diversity of Asgard archaea and their relationships with eukaryotes. Nature 593: 553–557 MacCready JS, Hakim P, Young EJ, Hu L, Liu J, Osteryoung KW, Vecchiarelli AG & Ducat DC (2018) Protein gradients on the nucleoid position the carbon-fixing organelles of cyanobacteria. Elife 7: e39723 Mauriello E (2019) How bacteria arrange their organelles. Elife 8: e43777 Mayor R & Etienne-Manneville S (2016) The front and rear of collective cell migration. Nat Rev Mol Cell Biol 17: 97–109 Mercier R, Bautista S, Delannoy M, Gibert M, Guiseppi A, Herrou J, Mauriello EMF & Mignot T (2020) The polar Ras-like GTPase MglA activates type IV pilus via SgmX to enable twitching motility in Myxococcus xanthus. Proceedings of the National Academy of Sciences 117: 28366–28373 Mercier R & Mignot T (2016) Regulations governing the multicellular lifestyle of Myxococcus xanthus. Current Opinion in Microbiology 34: 104–110 Miao W, Eichelberger L, Baker L & Marshall MS (1996) p120 Ras GTPase-activating Protein Interacts with Ras-GTP through Specific Conserved Residues. Journal of Biological Chemistry 271: .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 19 15322–15329 Miertzschke M, Koerner C, Vetter IR, Keilberg D, Hot E, Leonardy S, Søgaard‐Andersen L & Wittinghofer A (2011) Structural analysis of the Ras‐like G protein MglA and its cognate GAP MglB and implications for bacterial polarity. The EMBO Journal 30: 4185–4197 Mignot T, Shaevitz JW, Hartzell PL & Zusman DR (2007) Evidence that focal adhesion complexes power bacterial gliding motility. Science 315: 853–856 Piroli ME, Blanchette JO & Jabbarzadeh E (2019) Polarity as a physiological modulator of cell function. Front Biosci (Landmark Ed) 24: 451–462 Posner I, Engel M & Levitzki A (1992) Kinetic model of the epidermal growth factor (EGF) receptor tyrosine kinase and a possible mechanism of its activation by EGF. Journal of Biological Chemistry 267: 20638–20647 Rowlett VW & Margolin W (2013) The bacterial Min system. Curr Biol 23: R553-556 Schumacher D & Søgaard-Andersen L (2017) Regulation of Cell Polarity in Motility and Cell Division in Myxococcus xanthus. Annu Rev Microbiol 71: 61–78 Szadkowski D, Carreira LAM & Søgaard-Andersen L (2022) A bipartite, low-affinity roadblock domain- containing GAP complex regulates bacterial front-rear polarity. PLoS Genet 18: e1010384 Szadkowski D, Harms A, Carreira LAM, Wigbers M, Potapova A, Wuichet K, Keilberg D, Gerland U & Søgaard-Andersen L (2019) Spatial control of the GTPase MglA by localized RomR–RomX GEF and MglB GAP activities enables Myxococcus xanthus motility. Nat Microbiol 4: 1344–1355 Treuner-Lange A, Macia E, Guzzo M, Hot E, Faure LM, Jakobczak B, Espinosa L, Alcor D, Ducret A, Keilberg D, et al (2015) The small G-protein MglA connects to the MreB actin cytoskeleton at bacterial focal adhesions. J Cell Biol 210: 243–256 Webb MR (1992) A continuous spectrophotometric assay for inorganic phosphate and for measuring phosphate release kinetics in biological systems. Proceedings of the National Academy of Sciences 89: 4884–4887 Wettmann L & Kruse K (2018) The Min-protein oscillations in Escherichia coli: an example of self- organized cellular protein waves. Philos Trans R Soc Lond B Biol Sci 373: 20170111 Wittinghofer A, Scheffzek K & Ahmadian MR (1997) The interaction of Ras with GTPase-activating proteins. FEBS Letters 410: 63–67 Zhang Y, Ducret A, Shaevitz J & Mignot T (2012a) From individual cell motility to collective behaviors: insights from a prokaryote, Myxococcus xanthus. FEMS Microbiol Rev 36: 149–164 Zhang Y, Franco M, Ducret A & Mignot T (2010) A Bacterial Ras-Like Small GTP-Binding Protein and Its Cognate GAP Establish a Dynamic Spatial Polarity Axis to Control Directed Motility. PLoS Biol 8: e1000430 Zhang Y, Guzzo M, Ducret A, Li Y-Z & Mignot T (2012b) A Dynamic Response Regulator Protein Modulates G-Protein–Dependent Polarity in the Bacterium Myxococcus xanthus. PLOS Genetics 8: e1002872 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 20 Figure legends Figure 1: Measurement of MglA nucleotide exchange rates by fluorescence anisotropy A. Binding kinetics of GTP 488 (0.1 µM; dark green) or GDP 488 (0.1 µM; light green) in the presence of MglAGDP (1 µM). B. I- Binding kinetics of GTP 488 (0.1 µM; dark green) or GDP 488 (0.1 µM; light green) in the presence of MglAGDP (1 µM). II- After 600s, competition of GTP 488 or GDP 488 by addition of an excess of unlabeled GTP (100 µM) (represented by the dotted black line labeled “+ excess GTP”). C. I- Binding kinetics of GTP 488 (0.1 µM; dark green) or GDP 488 (0.1 µM; light green) in the presence of MglAGDP (1 µM). II- After 600s, competition of GTP 488 or GDP 488 by addition of an excess of unlabeled GDP (100 µM) (represented by the dotted black line labeled “+ excess GDP”). The shaded regions around each curve represent the standard deviations of the fluorescence anisotropy measurement at each time point, calculated from at least three independent experimental replicates. Figure 2: RomRX complex acts as an MglA effector. A. Binding kinetics of GTP 488 (0.1 µM) to MglA GDP (1 µM) in the absence (green) or in the presence of RomX (1µM; yellow), RomR (1µM; purple) or RomRX (1µM; black). B. Binding kinetics of GDP 488 (0.1 µM) to MglA GDP (1 µM) in the absence (green) or in the presence of RomX (1µM; yellow), RomR (1µM; purple) or RomRX (1µM; black). C. Binding kinetics of GTP 488 (0.1 µM) to MglA GDP (1 µM) in the absence (green) or in the presence of SgmX-Ct (1µM; blue). D. Binding kineti cs of GDP 488 (0.1 µM) to MglA GDP (1 µM) in the absence (green) or in the presence of SgmX-Ct (1µM; blue). E-G. Fluorescence anisotropy measurements for RomX , RomRX or SgmX -Ct respectively binding to MglAGTP488. Binding was estimated by measuring the chan ge in fluorescence anisotropy of MglA GTP488 upon titration with increasing concentrations of RomX, RomRX or SgmX -Ct. Each reading corresponds to the value of fluorescence anisotropy (30 min after the start of the experiment, at steady state). Each point represents the mean value from at least three independent measurements, and the error bars indicate standard deviation. H. I- Binding kinetics of GDP 488 (0.1 µM) to MglA GDP (1 µM) in the absence (green) or in the presence of RomX (1µM; yellow), RomR (1µM; purple) or RomRX (1 µM, black). II- After 600s, competition of GDP488 by addition of an excess of unlabeled GTP (100 µM) (represented by the dotted black line labeled “+ excess GTP”). I. I- Binding kinetics of GTP 488 (0.1 µM) to MglA GDP (1 µM) in the abse nce (green) or in the presence of RomX (1µM; yellow), RomR (1µM; purple) or RomRX (1 µM, black). II- After 600s, competition of GTP488 by addition of an excess of unlabeled GTP (100 µM) (represented by the dotted black line labeled “+ excess GTP”). The shaded regions around each curve in panels A -D, H and I represent the standard deviations of the fluorescence anisotropy measurements at each time point, calculated from at least three independent experimental replicates. Figure 3: Characterization of MglA interactions with MglB and its mutants A. Binding kinetics of GTP 488 (0.1 µM; dark red) or GDP 488 (0.1 µM; light red) to MglA GDP (1 µM) in the presence of MglB (2 µM). The inset corresponds to a zoomed view of the dashed rectangle in the graph. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 21 B. I- Binding kinetics of GTP 488 (0.1 µM; dark red) or GDP 488 (0.1 µM; light red) to MglA GDP (1 µM) in the presence of MglB (2 µM) or MglB ΔCt (blue; 2 µM). II- Competition of GTP 488 or GDP 488 by addition of an excess of unlabeled GTP (100 µM) at 600s (represented by the dotted black line labeled “+ excess GTP”). C. I- Binding kinetics of GTP 488 (0.1 µM; dark red) or GDP 488 (0.1µM; light red) to MglA GDP (1 µM) in the presence of MglB (2 µM). II- Competition of GTP 488 or GDP488 by addition of an excess of un labeled GDP (100 µM) at 600s (represented by the dotted black line labeled “+ excess GDP”). The shaded regions around each curve in panels A -C represent the standard deviations of the fluorescence anisotropy measurements at each time point, calculated fro m at least three independent experimental replicates. Raw (Top panel) and integrated heat (bottom panel) plots for the titration of: D. 200 µM MglB into 20 µM MglA and 100 µM GTPγS E. 492 µM MglB into 20 µM MglA and 100 µM GDP F. 1.1mM MglB ΔCt into 50 µM MglA and 100 µM GTPγS. The binding isotherms are a representation of replicate experiments. Figure 4: Kinetic analysis of MglA’s interaction with MglB or MglBΔCt in the presence of GTP. A. Schematic representation of the rapid equilibrium ran dom fixation model describing MglA and MglB kinetic interaction. B. Top: Plot of 1/v as a function of 1/[MglB] at varying GTP concentrations (2, 5, 10 and 20µM). The data were modeled using the equation (i). All lines converge to a single point correspond ing to -1/KB. Bottom: Following equation (ii), the y -intercept (corresponding to the plot of 1/v in the function of 1/[MglB] at a single GTP concentration) was plotted in the function of 1/[GTP] at four different GTP concentrations 5, 10,20 and 50 µM. The Y-intercept corresponds to 1/Vm and the x-intercept corresponds to -1/Ks[GDP]. C. Schematic representation of the rapid equilibrium ordered fixation model describing MglA and MglB ΔCt kinetic interaction D. Top: Plot of 1/v as a function of 1/[MglBΔCt] at varying GTP concentrations (2, 5, 10 and 20µM). The data were modeled using the equation (i). All lines converge to a single point corresponding to 1/V m. Bottom: Following equation (ii), the slope (corresponding to the plot of 1/v in the function of 1/[ MglBΔCt] at a single GTP concentration) was plotted in the function of 1/[GTP] at three different GTP concentrations 2, 5 and 10 µM. The Y-intercept corresponds to -KB ’/Vm and the x-intercept corresponds to -1/Ks ’[GDP]. Figure 5: Impact of the RomR on the MglA-MglB interaction A. Binding kinetics of GTP488 (dark green;0.1 µM) or GDP488 (light green; 0.1 µM) to MglA GDP (1 µM). Binding kinetics of GTP488 (0.1 µM) to MglA GDP in the presence of MglB (2µM; dark red) or binding kinetics of GTP488 (0.1 µM) to MglA GDP (1 µM) in presence of MglB (2µM; dark red) followed by the addition of RomX (30 µM; orange) or SgmX-Ct (30 µM; dark blue) at 600s. B. Binding kinetics of GTP488 (da rk green;0.1 µM) or GDP488 (light green; 0.1 µM) to MglA GDP (1 µM). Binding kinetics of GTP488 (0.1 µM) to MglA GDP in the presence of MglB (2µM; dark red) or binding kinetics of GTP488 (0.1 µM) to MglA GDP (1 µM) in presence of MglB (2µM; dark red) followed by the addition of RomR (30 µM; purple) at 600s. C. Binding kinetics of GTP488 (dark green;0.1 µM) or GDP488 (light green; 0.1 µM) to MglA GDP (1 µM). Binding kinetics of GDP488 (0.1 µM) to MglA GDP in the presence of MglB (2µM; light red) or binding kinetics of GDP 488 (0.1 µM) to MglA GDP (1 µM) in presence of MglB (2µM; light red) followed by the addition of RomR (30 µM; purple) at 600s. D. Isothermal titration calorimetry (ITC) profile of MglB -RomR interaction. Raw (left panel) and integrated heat (right panel) plots for the titration of 1.3 mM MglB into 70 µM RomR. The fit yielded a K D of 31 ± 15 µM. E. Initial rate data for Pi release at different concentration of GTP in presence of MglA GDP (2 µM) + MglB (4 µM) (empty circle), MglA GDP (2 µM) + MglB (2 µ M) (empty triangle) or MglA GDP (2 µM) + MglB (4 µM) + RomR (20 µM) (purple filled square). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 22 Figure 6: In vivo localization and dynamics of RomR and RomX. A. Time-lapse fluorescence microscopy images of pole-to-pole dynamics of RomR-mcherry (red) and RomX- GFP (green) (20s intervale). The merged image highlights colocalization of both proteins at polar regions. Arrows point to reversal events. The scale bar represents 6 μm. B. Quantitative analysis of fluorescence intensity dynamics over time, comparing the l ocalization of RomR - mCherry and RomX-GFP from Pole 1 to Pole 2. The difference in intensities of RomR -mCherry and RomX- GFP between the two poles was normalized to the range defined by the minimum and maximum intensity. Shaded regions represent the standard deviation across measurements from 3 independent cells. The 0 on the y-axis indicates symmetrical localization between the two poles. The gray line indicates the time of reversal. Figure 7: Epi -fluorescence imaging of MglA, MglB, and RomR localization in water-in-oil emulsion droplets stabilized by lipids A. Schematic representation of the formation of the lipid droplets; surfactants are in orange, lipids are in black and proteins are in green. B. Epi-fluorescence image of lipid droplets containing MglA (1 µM) labeled with GTP488 (0.2 µM) C. Epi-fluorescence image of lipid droplets containing MglA (1 µM) labeled with GDP488 (0.2 µM) D. Epi-fluorescence image of lipid droplets containing RomR-GFP (10 µM). E. Epi-fluorescence image of lipid droplets containing red-labeled MglB (2 µM). F. Epi-fluorescence images of lipid droplets containing MglA (1 µM) labeled with GTP 488 (0.2 µM) in presence of red-labeled MglB (2 µM). G. Epi-fluorescence images of lipid droplets containing MglA (1 µM) labeled with GDP 488 (0.2 µM) in presence of red-labeled MglB (2 µM). H. Epi-fluorescence images of lipid droplets containing red-labeled MglB (10 µM) in presence of RomR-GFP (25 µM). I.Epi-fluorescence images of lipid droplets containing red -labeled MglB (10 µM) in presenc e of RomR-GFP (25 µM) and unlabeled MglA (5 µM) J. Epi-fluorescence images of lipid droplets containing red -labeled MglB (10 µM) in presence of unlabeled RomR (25 µM) and MglA (5 µM) labeled with GTP488 (0.2 µM). Rhodamine was used to label the lipids. Figure 8: Role of the RomRX complex in establishing cell polarity in M. xanthus. Upper Panel: At the leading pole, the RomRX complex regulates polarity by interacting with MglA GTP through RomX. This interaction supports the recruitment of MglA GTP to the pole where it can associate with polar motility specific effectors present in the A - and S- motility complex (such as SgmX for the Type -IV pili and the GltJ ZnR - GYF motif for the Agl -Glt complex). At the lagging pole, the low concentration of MglA facilitate s the recruitment of MglB by MglC. Middle panel: RomRX detaches from the leading and it relocalizes to the lagging pole possibly via several interactions (ie MglC, MglB and unknown). As it accumulates at high concentrations, RomRX competes with the formati on of the MglA -MglB complex interacting with MglB. Lower Panel: However, as observed in vitro, this competition does not entirely block the formation of the MglA-MglB complex due to its low affinity. Given that formation of the MglA -MglB complex and GTP hydrolysis occur rapidly, and that the dissociation of the MglA -MglB complex is the rate -limiting step, RomRX mostly acts as a dissociation factor following hydrolysis.Additionally, RomY plays a crucial role in orienting MglB function toward a GAP in vivo, e nhancing its GAP activity and thus accelerating the reaction. This regulation is essential to drive GTP hydrolysis efficiently, as the kcat of MglB alone is low. Tables and their legends koff GTP488 (s-1) .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 23 MglAGDP 2.8×10-3±5.6×10-4 MglAGDP+ SgmX-Ct 2.7×10-4± 2.4×10-5 MglAGDP+ RomX 8.4×10-4±3.3×10-4 MglAGDP+ RomR n.r. MglAGDP+ RomRX 2.4×10-4 ± 5.5×10-5 Table 1: Dissociation rates constant (K off) corresponding to the graph in Fig .2I. The errors represent the standard deviations calculated from at least three independent measurements. koff GTP488 (s-1) koff GDP488 (s-1) MglAGDP 2.8×10-3 ± 5.6×10-4 6.0×10-4 ± 8.7×10-5 MglAGDP + MglB 9.0×10-4 ± 1.0×10-4 1.2×10-3 ± 2.6×10-4 MglAGDP + MglBΔCt 5.8×10-4 ± 9.4×10-5 n.r. Table 2: The dissociation rates constant (k off) corresponding to the graph in Fig. 3B and C. The errors represent the standard deviations calculated from at least three independent measurements. MglA + MglB MglA + MglBΔCt KB 0.63 ± 0.24 µM n.r. K’ S [GDP] 4.1 ± 1.6 µM 2.58 ± 0.36 µM K’ B 0.37 ± 0.07 µM 0.13 ± 0.06 µM KS [GDP] 2.45 ± 1.43 µM n.r. kcat 2.67×10-3 ± 1.1 ×10-4 s–1 2 ×10-3 ± 5.7 ×10-5s–1 Table 3: Rapid equilibrium random/ordered mechanism. Kinetic parameters obtained from solving the graphs in Fig. 4B and D and Fig. EV1A and B 𝒌𝒄𝒂𝒕 𝒂𝒑𝒑(min-1) 𝑲𝒎 𝒂𝒑𝒑 (µM) 𝒌𝒄𝒂𝒕 𝒂𝒑𝒑/𝑲𝒎 𝒂𝒑𝒑(M-1s-1) MglA (2µM) + MglB (4µM) 0.16 ± 0.011 1.02 ± 0.014 2.60 ×103 MglA (2µM) + MglB (2µM) 0.12 ± 0.001 0.98 ± 0.22 2.04 ×103 MglA (2µM) + MglB (4µM) + RomR (20µM) 0.11 ± 0.002 1.00 ± 0.023 1.83 ×103 Table 4: Kinetic parameters table. Initial rate data were fitted to the Michaelis -Menten equation. The assays were performed using constant concentration of proteins and a range of GTP concentration (0.5 to 100 µM). .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 24 Expanded view figure legends Figure EV1: Kinetic analysis of MglA toward GTP and MglB or MglBΔCt. A. Top: Plot of 1/v as a function of 1/[GTP] at varying MglB concentrations (0.25, 0.5, 2 and 4µM). The data were modeled using the equation (i). All lines converge to a single point corresponding to -1/Ks ’ [GDP], highlighting the substrate -dependent modulation of enzymatic activity. Bottom: Following equation (ii), the y-intercept (corresponding to the plot of 1/v in the function of 1/[GTP] at a single MglB concentration) was plotted in the function of 1/[M glB] at three different MglB concentrations 0.25, 0.5 and 4 µM. Y -intercept corresponds to 1/Vm and the x-intercept corresponds to -1/KB ’. B.Top: Plot of 1/v as a function of 1/[GTP] at varying MglBΔCt concentrations (0.5, 10, 15 and 20µM). The data were modeled using the equation (i). All lines converge to a single point corresponding to -1/Ks ’ [GDP]. Bottom: Following equation (ii), the y -intercept (corresponding to the plot of 1/v in the function of 1/[GTP] at a single MglBΔCtconcentration) was plotted i n the function of 1/[ MglBΔCt] at four different MglBΔCt concentrations 0.5, 2.5, 5 and 10 µM. Y -intercept corresponds to 1/Vm and the x -intercept corresponds to - 1/KB ’. .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint 25 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 1 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 2 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 3 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint A CMglAGDP + GTP 𝐾𝑠′ [GDP] MglB 𝐾𝐵 ′ kcat + MglAGTPMglB MglAGDPMglB + Pi MglAGTP MglB + 𝐾𝐵 MglAGDPMglB + GTP 𝐾𝑠 [GDP] MglAGDP + GTP MglBΔCt 𝐾𝐵 ′ kcat + MglAGTPMglBΔCt MglAGDP + MglBΔCt + Pi MglAGTP 𝐾𝑠′ [GDP] i- 1/v= 1 𝑀𝑔𝑙𝐵 𝐾𝐵𝐾𝑆 𝐺𝐷𝑃 𝑉𝑚 𝐺𝑇𝑃 + 𝐾𝐵 ′ 𝑉𝑚 + 1 𝑉𝑚 1 + 𝐾𝑆 [𝐺𝐷𝑃] 𝐺𝑇𝑃 i- 1/v= 1 𝑀𝑔𝑙𝐵𝛥𝐶𝑡 𝐾𝐵 ′ 𝐾𝑆 ′ 𝐺𝐷𝑃 𝑉𝑚 𝐺𝑇𝑃 + 𝐾𝐵 ′ 𝑉𝑚 + 1 𝑉𝑚 -6 -4 -2 0 2 4 6 8 -0,4 -0,3 -0,2 -0,1 0 0,1 0,2 0,3 0,4 1/[MglBΔCt] [GTP] = 2µM [GTP] = 5µM [GTP] = 10µM [GTP] = 20µM1/Vm 1/v [GTP] = Cts ; varying [MglB] ii. Y intercept = 𝟏 𝑽𝒎 𝟏 + 𝑲𝑺 [𝑮𝑫𝑷] 𝑮𝑻𝑷 0 1 2 3 4 5 6 7 8 -1,5 -1 -0,5 0 0,5 1 1 Vm Y intercept 1/[GTP] ii- Slope = 𝑲𝑩 ′ 𝑲𝑺 ′ 𝑮𝑫𝑷 𝑽𝒎 𝑮𝑻𝑷 + 𝑲𝑩 ′ 𝑽𝒎 [GTP] = Cts ; varying [MglBΔCt] Slope 1/[GTP] -1 -0,5 0 0,5 1 1,5 2 -2,5 -2 -1,5 -1 -0,5 0 0,5 1 −KB ′ Vm −1 KS ′ [GDP] B D −1 KS [GDP] -10 -5 0 5 10 15 20 -5 -3 -1 1 3 5 − 1 𝐾𝐵 [GTP] = 2µM [GTP] = 5µM [GTP] = 10µM [GTP] = 20µM 1/v 1/[MglB] Figure 4 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 5 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 6 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Figure 7 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint Lagging pole Leading pole Motility direction MglAGDP Lagging pole Leading pole Motility direction MglB Effectors MglAGTP RomRX MglA-MglB dissociation MglB MglAGTP MglAGDP Hydrolysis Pi RomRX MglC MglC RomY RomY Figure 8 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint EV1 .CC-BY-NC-ND 4.0 International licenseavailable under a (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in perpetuity. It is made The copyright holder for this preprintthis version posted November 1, 2024. ; https://doi.org/10.1101/2024.10.31.621274doi: bioRxiv preprint