Intracellular ion concentrations and cation-dependent remodelling of bacterial MreB assemblies

Here, we measured the concentrations of several ions in cultivated Gram-negative and Gram-positive bacteria, and analyzed their effects on polymer formation by the actin homologue MreB. We measured potassium, sodium, chloride, calcium and magnesium ion concentrations in Leptospira interrogans, Bacillus subtilis and Escherichia coli. Intracellular ionic strength contributed from these ions varied within the 130–273 mM range. The intracellular sodium ion concentration range was between 122 and 296 mM and the potassium ion concentration range was 5 and 38 mM. However, the levels were significantly influenced by extracellular ion levels. L. interrogans, Rickettsia rickettsii and E. coli MreBs were heterologously expressed and purified from E. coli using a novel filtration method to prepare MreB polymers. The structures and stability of Alexa-488 labeled MreB polymers, under varying ionic strength conditions, were investigated by confocal microscopy and MreB polymerization rates were assessed by measuring light scattering. MreB polymerization was fastest in the presence of monovalent cations in the 200–300 mM range. MreB filaments showed high stability in this concentration range and formed large assemblies of tape-like bundles that transformed to extensive sheets at higher ionic strengths. Changing the calcium concentration from 0.2 to 0 mM and then to 2 mM initialized rapid remodelling of MreB polymers.


Results intracellular ion concentrations in bacteria.
Ion-selective electrode/colorimetry is able to measure concentrations of free ions, while flame photometry measures the total concentrations of ions, including ions bound to proteins and DNA. To compare these two methods, we measured the intracellular concentration of Na + , K + and Ca 2+ in samples extracted from Escherichia coli using both ion-selective electrode/colorimetry and flame photometry. The two methods detected similar concentrations of the three ions ( Fig. 1A) (with no statistically significant difference, p > 0.05), indicating that the ion extraction method used here was effective in releasing the majority of bound ions. Ion-selective electrode/colorimetry is able to measure more types of ions than flame photometry, therefore in subsequent experiments only ion-selective electrode/colorimetry was used to measure ion concentrations (including Na + , K + , Ca 2+ , Mg 2+ and Cl − ). Our aim was to understand whether a relationship exists between the intracellular ionic conditions and the cell shapes of two Gram-negative species, Leptospira interrogans (Li), which is a Spirochaete with a helical shape, and Escherichia coli (Ec) with rodshaped cell morphology. For comparison, we examined a Gram-positive species Bacillus subtilis (Bs), which also Scientific RepoRtS | (2020) 10:12002 | https://doi.org/10.1038/s41598-020-68960-w www.nature.com/scientificreports/ has a rod-shaped cell. The ion concentrations determined in all three species are presented in Fig. 1B. Sodium ion concentrations were high in all three species: 296 ± 12 mM in Li, 219 ± 42 mM in Ec and 122 ± 11 mM in Bs, and show statistically significant differences. By contrast, potassium ion concentrations were relatively low: 38 ± 10 mM and 27 ± 10 mM in Ec and Bs, respectively, and significantly lower 5 ± 1 mM in Li. The calcium ion concentration was 1 ± 0.08 mM (significantly higher than for Ec and Bs) and the magnesium concentration was 0.38 ± 0.2 mM in Li (statistically significantly lower than for the other two species). Interestingly, in Ec and Bs the calcium ion concentrations were sub-millimolar, while the magnesium ion concentration was between 1 and 2 mM. Chloride ions are known to be the major anion composition of bacterial cells. Here, we measured the chloride ion concentrations to be 164 ± 0.7, 207 ± 41 and 106 ± 13 mM in Li, Ec and Bs, respectively (there is a statistically significant difference only between Li and Ec). Na + , K + , Cl − , Ca 2+ , and Mg 2+ are the major contributors to intracellular ionic strength [36][37][38] . Calculation (Eq. 1) of the intracellular ionic strengths in the differences species, based on these ions, gave values of 235 mM (Li), 237 mM (Ec) and 130 mM (Bs). Despite of the different levels between species, the ratios between various ions in a single species were similar between the three species.

Reorganization of MreB polymers by changes in calcium levels.
Li-MreB shows monovalent cation dependent sensitivity to calcium ion concentrations in the range of 0.1-2 mM. In the absence of KCl and MgCl 2 , a fivefold increase in light scattering was observed on the addition of 2 mM CaCl 2 to Li-MreB monomers (50 µM). However, in the presence of 300 mM KCl and 2 mM MgCl 2 the signal from the polymers was unchanged on the calcium addition (dashed line in Fig. 3A). Similar effects were observed for Rr-MreB (Fig. 3B).
To further investigate these effects, we carried out time-lapse microscopy imaging to explore any structural changes of fluorescently labeled Li-MreB polymers in response to calcium treatment. The initial large assemblies (gray, Fig. 3C), appeared to shorten and become less flexible after 5 min (cyan, Fig. 3C) and 10 min (green, Fig. 3C) on the addition of 2 mM CaCl 2 (Movie 1). White arrows in Fig. 3C indicate the shortening of the MreB assembly, by the contraction of the whole structure. At the start, the superstructure moves relatively freely. After calcium treatment, the thermal motion of the superstructure slowed and stopped. Both characteristics suggest a calcium-induced structural change in the MreB assembly. However, after 6 mM EDTA treatment in 300 mM KCl, in the absence of divalent cations, the Li-MreB and Rr-MreB polymers precipitated ( Fig. 3D and E). The addition of 6 mM EGTA, rather than EDTA, in the absence of calcium ions, led to the Li-MreB superstructures remodelling into a single ribbon-like sheet ( Fig. 3F) (Supp. Movie 2), while Rr-MreB formed irregular bundled assemblies (Fig. 3G). These data show that calcium influences the formation of MreB polymer superstructures.
To better understand the role of calcium, we carried out experiments in which the responses of polymers were monitored by light scattering on fast changes in calcium ion concentrations. In the presence of 6 mM EGTA, where free calcium was eliminated from the solution, the addition of 2 mM Ca 2+ led to a jump in light scattering (black line in Fig. 4A), which subsequently decreased slowly, indicating a moderate size change of the Li-MreB (50 µM) polymers. Similarly, after extended incubation with 2 mM CaCl 2 followed by a sudden switch to EGTA (6 mM) a fast increase in light scattering was observed (orange line in Fig. 4A) followed by a gradual decrease, indicating that the Li-MreB polymers went through a relatively quick supramolecular reorganization. Likewise, a sudden change of calcium level, by consecutive application of EGTA then calcium, had a similar effect on Rr-MreB sheets indicated by a rapid increase in light scattering (Fig. 4B). To explore how calcium affects MreB polymers in the presence of EGTA, we followed the change with fluorescence microscopy. Alexa488 labeled Li-MreB assemblies in the presence of 2 mM CaCl 2 were dissociated and precipitated after the addition of 6 mM EGTA (Fig. 4C), explaining the quick change of light scattering after this treatment (Fig. 4A, orange line). Surprisingly, in the presence of 6 mM EGTA, Li-MreB polymers formed novel web-like structures after 2 mM CaCl 2 addition (Fig. 4D) as was also the case for Rr-MreB (Fig. 4E) and Ec-MreB (Fig. 4F) polymers. Possibly, magnesium can replace calcium in binding to the MreB polymers to stabilize the assemblies, since the elimination of calcium by EGTA in the presence of magnesium did not affect the MreB assemblies. Subsequent addition of calcium, in excess of the local EGTA concentration, reorganized the MreB filaments. Microscopy images of whole cell lysate from E. coli overexpressing Li-MreB or Rr-MreB ( Fig. 4G and H, respectively), in which all cysteines of the entire sample were labeled with Alexa488-maleimide, showed extensive web-like structures after treatment with EGTA (6 mM) followed by CaCl 2 (2 mM excess). This indicates that changes in intracellular calcium levels may play a role in intracellular polymer reorganization. the persistence length of MreB under various salt conditions. Since ionic conditions have substantial effects on MreB polymerization and superstructure reorganization, we sought to characterize the flexibility of MreB polymers under various salt conditions. We analyzed the microscope images with Easyworm software to estimate the persistence lengths of the MreB structures. Ionic strength was effective in changing the polymerization properties of Rr-MreB (Supp. 2B, C, D), therefore the persistence length was examined as the function of the KCl concentration (Fig. 5A). The buffers contained different amounts of KCl (50-300 mM) and 2 mM MgCl 2 , 0.1 mM CaCl 2 . Persistence lengths were ionic strength dependent and fell in the range of 3.9-35.5 µm (images not shown). Interestingly, the optimal polymerization conditions of Rr-MreB (200 mM KCl) resulted the longest persistence length and the lowest flexibility of polymers. However, the persistence lengths of Li-MreB and Rr-MreB polymers in the presence of 300 mM KCl were similar (14.6 µm), indicating similar flexibilities. For reference, in the case of actin, the persistence length is 12 µm in the activated state of the thin filaments, in the presence of calcium and binding of the tropomyosin-troponin complex 40 . The persistence length of ADPbound state of the actin filaments is 9 ± 0.5 µm, while the filaments become much stiffer in the absence of calcium (20 ± 1 µm). The persistence lengths estimated here for MreB superstructures varied within a broad range (3.9-35.5 µm), which is similar to the range observed for actin filaments. Interestingly, in the presence of 100 mM KCl the persistence length of Rr-MreB (25.73 ± 9.3 µm) polymers was similar to the values of actin thin filaments in the absence of calcium (20 ± 1 µm) 40 . These

Discussion
Here, we used two methods, ion-selective electrodes/colorimetry and flame photometry, to measure the concentrations of Na + , K + and Ca 2+ cell extracts, and used ion-selective electrodes/colorimetry to determine the Mg 2+ and Cl − concentrations. These methods reveal the average intracellular ion conditions and do not report inhomogeneous distributions in different areas of a cell. Nevertheless, the agreement between the Na + , K + and Ca 2+ levels, measured by the two techniques, suggests that these average values are accurate, and that in general the levels of sodium ions exceed potassium ions in bacteria. Bacterial cells can exchange monovalent cations with the media to adjust intracellular ionic strength during osmoadaptation 41 . The types of broth used here in growing bacterial cultures were major source of sodium ions. The conditions for culturing Leptospira interrogans (Korthof), Escherichia coli (Luria-Bertani) and Bacillus subtilis (Mueller-Hinton) contain 54, 85 and 342 mM sodium ions, respectively, which show a different relationship to the measured intracellular sodium ion concentrations, Leptospira interrogans (296 mM), Escherichia coli (219 mM) and Bacillus subtilis (122 mM). We measured almost identical intracellular ion concentration-derived ionic strengths in Leptospira interrogans (235 mM) and Escherichia coli (237 mM), which mainly resulted from sodium and chloride ions. The intracellular ionic strength in Bacillus subtilis was significantly lower (130 mM), suggesting that non-ion osmolytes may have a significant role in maintaining the osmotic balance. The in vitro polymerization and formation of MreB superstructures is highly dependent on the presence of cations. Millimolar magnesium and hundred millimolars potassium or sodium are necessary for efficient MreB polymerization. Addition of calcium caused the stacking into ribbon-like structures and large assemblies, and we hypothesize that calcium binding may change the strain in filaments. Subsequent calcium depletion, via EGTA treatment, reordered the polymers into extensive sheets in the presence of magnesium, and further treatment with calcium led to fissured monolayer sheets and the dissociation of filaments into web-like structures (Fig. 6). Structural studies have shown that pairs of MreB protofilaments associate together in an antiparallel manner, while molecular dynamics simulations suggest the possibility of curvature in the protofilaments 42,43 . The in vitro assemblies observed here are likely to be formed by the non-polarized filaments associating side by side. In vivo the assemblies will also be stabilized by membrane binding 43 . Molecular dynamics simulations have shown that the twist of MreB double protofilaments can be reduced by membrane binding. The moderated dynamics of MreB filaments resulted in shorter filaments, and possibly provides tuning to their flexibilty and length 44 . The untwisted antiparallel structure of pairs of protofilaments can allow bending of the MreB filaments, which may stabilize the curvature of a membrane 44 . High calcium concentrations 45 , near a membrane, may induce longer persistence lengths in MreB polymers. Thus, we hypothesize that local calcium concentration changes may lead to the reshaping of membranes. Through calcium-ion induced stiffness, MreB filaments may contribute to The persistence length of Alexa488-labeled Li-MreB assemblies (red columns) changed significantly in the presence of 2 mM calcium, and in the absence of calcium by the treatment with 6 mM EGTA, and Rr-MreB assemblies (blue columns) showed a non-significant change in the presence of 2 mM calcium, but a significant change in the presence of 6 mM EGTA. Both assemblies show non-significant differences to the pretreated samples on subsequent addition of 2 mM calcium after EGTA treatment. The persistence length of Alexa488-labeled Ec-MreB assemblies (green columns) did not show a large difference on the addition of 6 mM EGTA followed by 2 mM calcium. Error bars refer to mean ± SD of five independent measurements. Double asterisks indicate statistically significant difference relative to the initial conditions. The analysis was based on a two-sample t-test, p < 0.005. Scientific RepoRtS | (2020) 10:12002 | https://doi.org/10.1038/s41598-020-68960-w www.nature.com/scientificreports/ specifying membrane curvature at points of membrane-MreB interaction (Supp. 4). Presumably, the final cell shape will be a product of the MreB properties and the ionic milieu. Calcium concentration-induced changes of MreB structure may also participate in membrane remodelling during cell division or osmotic adaptation. Rr-MreB polymerization is more sensitive to monovalent cations than Li-MreB, which suggests that the shape and stability of cytoskeletal systems will vary between organisms under similar intracellular conditions. Since, the high salt conditions present in bacteria fluctuate in response to osmotic shock, the influence on the stability of MreB scaffold and its ability to reattach bacterial membranes to cell wall is likely to be affected by divalent cations. www.nature.com/scientificreports/ Varying ionic conditions do not in general change cell shape, however they are known to modify membrane stiffness and enzyme activity, which effects the stability of bacterial envelope 46 . However, the effect of calcium levels appears to be more sensitive. It is likely that calcium fluxes may have a role in regulating and remodelling the in vivo MreB cytoskeleton, which in turn may influence the mobility and localization of links between the membrane and cell wall during the cell division.
One of the limitations of our study is that due to the limited number of species analyzed, it is not clear whether the differences seen are species specific or common features of several relative strains. But we highlighted the relevance of intracellular ion conditions influence on MreB polymer formation.
Cells were harvested by centrifugation (4500 × g for 10 min), then the extracellular medium was removed by washing. Five times the pellet volume of distilled water was added, incubated for 2 min at RT, and subsequently inverted gently ten times and centrifuged again (4500 × g for 10 min). The pellet volume was measured in a scaled 2 mL Eppendorf tube. Cells killed and the extracellular water was evaporated by heat treatment (100 °C for 20 min). The cell pellets were resuspended in five times the pellet volume of water, frozen and subsequently boiled to generate a homogenous cell lysate. The slurry was centrifuged for 10 min at 100,000 × g. Ion concentrations were measured from the supernatants. Unbound Na + , K + , Cl − (ion-selective electrode) Ca 2+ and Mg 2+ (colorimetry) levels were analyzed using a COBAS INTEGRA 400 plus analyzer (Roche Diagnostics, GmbH, Mannheim, Germany) following the manufacturer's instructions. Total (protein-bound and free) Na + , K + and Ca www.nature.com/scientificreports/ MreB polymers were not but monomers were clarified by ultracentrifugation (100,000 × g, at 4 °C for 30 min) and the concentrations of protein and fluorophore in the supernatant were determined using spectrophotometry (ε Alexa488 = 73,000 M −1 cm −1 , ε Alexa568 = 88,000 M −1 cm −1 , ε LiMreB = 10,555 M −1 cm −1 , ε RrMreB = 15,025 M −1 cm −1 , ε EcMreB = 7575 M −1 cm −1 ). The ratio of labeling was calculated as the concentration ratio of the probe to the protein and was found to be approximately 0.1 in cases of MreB polymers and more than 0.5 in case of monomers.
in vitro polymerization and polymer remodelling assays of MreB. The polymerization kinetics of MreB was investigated using light scattering assays with a Perkin Elmer LS-50 spectrofluorimeter. The excitation and emission monochromators were set to wavelength 400 nm and the excitation and emission slits to 2.5 nm.
The samples (2 mL) were stirred continuously with a magnetic stirrer during the polymerization process. Polymerization was initiated by adding monovalent and/or divalent cations, as indicated. The same set up was used to study the time dependent and ionic concentration dependent morphology changes of MreB polymers. The settings for the method based on our protocol and experience 39 . fluorescence microscopy. In these experiments 15 μL of fluorescent labeled MreB was dropped on slides, incubated in the presence of a buffer containing 300 mM KCl, 2 mM MgCl 2 , 0.1 mM CaCl 2 , unless otherwise indicated, then covered by coverslips. The morphologies of the labeled MreB assemblies were analyzed using a Leica TCS SP confocal scanning microscope system (Leica Microsystems GmbH Germany) equipped with a 10-63X objective lens. Image acquisition was carried out a fluorescent-probe specific wavelengths, Alexa488 Data analysis. Data are presented as means ± standard deviations (SD) throughout. Comparisons were performed using Students T-test and statistically significant differences between groups were defined as p values < 0.05 or < 0.005 and are indicated in the legends of figures.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.