Cell wall synthesis and remodelling dynamics determine division site architecture and cell shape in Escherichia coli

Abstract

The bacterial division apparatus catalyses the synthesis and remodelling of septal peptidoglycan (sPG) to build the cell wall layer that fortifies the daughter cell poles. Understanding of this essential process has been limited by the lack of native three-dimensional views of developing septa. Here, we apply state-of-the-art cryogenic electron tomography (cryo-ET) and fluorescence microscopy to visualize the division site architecture and sPG biogenesis dynamics of the Gram-negative bacterium Escherichia coli. We identify a wedge-like sPG structure that fortifies the ingrowing septum. Experiments with strains defective in sPG biogenesis revealed that the septal architecture and mode of division can be modified to more closely resemble that of other Gram-negative (Caulobacter crescentus) or Gram-positive (Staphylococcus aureus) bacteria, suggesting that a conserved mechanism underlies the formation of different septal morphologies. Finally, analysis of mutants impaired in amidase activation (ΔenvC ΔnlpD) showed that cell wall remodelling affects the placement and stability of the cytokinetic ring. Taken together, our results support a model in which competition between the cell elongation and division machineries determines the shape of cell constrictions and the poles they form. They also highlight how the activity of the division system can be modulated to help generate the diverse array of shapes observed in the bacterial domain.

Fig. 1. In situ cell envelope architecture and dynamics during E. coli cell division.

a, Overview of different stages of cell division. Summed, projected central slices of cryo-electron tomograms visualizing different stages in division of wild-type E. coli are shown. Black arrowhead indicates the side of the division site displayed in (b). b, Top row: NAD-filtered cryo-electron tomograms visualizing the cell wall. Left panels show a 2D slice, right panels show the corresponding slice with segmentations for the observed PG signal in cyan, IM in green and OM in magenta (see Methods). White arrowheads indicate where the PG layer appears to thicken from one to two layers, and black arrowhead indicates the side of the division site shown in the schematic overview below. Bottom row: corresponding labelled summary diagrams. The left two bottom panels correspond to arrowhead-marked top division side rotated 90° to the left. Segmented PG signal is not indicative of specific glycan strand network. c–e, Representative time-lapse series from 3 biological replicates of wild-type E. coli expressing Pal-mCherry and ZipA-sfGFP as OM and IM markers, respectively, imaged at 30 °C on M9 supplemented with 0.2% casamino acids and d-glucose. Fluorescence signals were deconvolved (see Methods). The yellow triangle marks division sites used for line scans of fluorescence intensity (FI) profiles (d) and kymograph analysis of cytokinesis (e). f, Average constriction velocities of the IM and OM were derived from the slopes of the fluorescence signals in kymographs (see Methods). Black line indicates mean. Two-sided unpaired Mann-Whitney test; **P < 0.01; N = 150 division kymographs. g, Instantaneous constriction velocities for ZipA (IM, green) and Pal (OM, magenta) are plotted against normalized cell width. Second order polynomial fits with 95% confidence intervals are shown. Scale bars: a and b, 100 nm; c, 2 µm; e, 200 nm (vertical) and 5 min (horizontal). Source data

Fig. 2. Divisome mutants display altered division site ultrastructure and constriction kinetics in E. coli.

a, Schematic overview of the septal PG loop pathway for the activation of sPG synthesis (see text for details). b, Left: NAD-filtered cryo-electron tomograms of division sites in the indicated division mutants of E. coli shown as in Fig. 1b. Right: summary diagrams of the cell envelope architecture visualized. Black arrowheads indicate the side of the division site represented in the schemes. Segmented PG signal is not indicative of specific glycan strand network. c, Top: representative time-lapse series from 3 biological replicates of indicated E. coli division mutants expressing Pal-mCherry and ZipA-sfGFP as OM and IM markers, respectively, imaged as in Fig. 1. Bottom: kymograph analysis and line scans of fluorescence intensity profiles of cytokinesis, from division sites marked with yellow triangles in the top row. d, Constriction velocities of the IM and OM were determined as in Fig. 1. Black line indicates mean. Data from wild type are replotted from Fig. 1f for comparison. Brown-Forsythe and Welch ANOVA test with Dunnett’s correction for multiple comparisons, significance of differences is tested relative to wild type (wt); **P < 0.01, ***P < 0.001, ****P < 0.0001; NS, not significant (P = 0.09); N = 150 (wt), 48 (ftsN-∆SPOR), 74 (∆envC), 68 (ftsL*) kymographs. e, Instantaneous constriction velocities for IM (top) and OM (bottom) are plotted against normalized cell width. Second order polynomial fits with 95% confidence intervals are shown. See Extended Data Fig. 4b,c for individual instantaneous constriction velocity traces. Data from wild type are replotted from Fig. 1g for comparison. Scale bars: b, 100 nm; c, top row, 2 µm; bottom row kymographs, 200 nm (vertical), 5 min (horizontal). Source data

Fig. 3. Measuring cell wall synthesis and hydrolysis rates during division and elongation in E. coli.

a, Labelling patterns observed for an FDAA pulse-chase experiment. New cell wall material is labelled with HADA (blue), while old material is stained with YADA (yellow). b, Representative images from 3 biological replicates of indicated strains after 2, 4 and 8 min pulses with HADA. Overlay images are provided in Extended Data Fig. 8l,m. c,d, Mean fluorescence intensity was measured at the division site for new (c) and old (d) PG. c, Data were fit to a linear regression to derive sPG synthesis rates. Data points represent median ± 95% confidence intervals. d, Reduction in old (YADA) fluorescence intensity was fit to a one-phase exponential decay curve. e, Mean septal PG hydrolysis rates were derived from decay curves in d. Points represent the average value of the three biological replicates and bars indicate mean + 1 s.d. N = 1,054 (wt), 716 (ftsN-∆SPOR), 819 (∆envC), 880 (ftsL*) cells. f, Side wall incorporation of new cell wall material (HADA fluorescence intensity) was measured after 8 min due to low signal intensities in earlier time points. Black line indicates median, one-way ANOVA with Dunnett’s correction for multiple comparisons, significance of difference is tested relative to wild type; NS, P = 0.06; ***P < 0.001, ****P < 0.0001; N = 103 (wt), 107 (ftsN-∆SPOR), 101 (∆envC), 100 (ftsL*) cells. g, The ratio between sPG and side wall synthesis was calculated by dividing the mean HADA fluorescence intensity after the 8 min pulse. h–j, Labelling patterns observed for the pulse-chase experiment in cells with inhibited division by SulA expression (h). New cell wall material is labelled with Alexa488-labelled MurNAc-alkyne probes (yellow), while old material is stained with HADA (blue). Mean fluorescence intensity was measured along the side wall for both MurNAc-alkyne (i) and HADA (j) and fitted to a quadratic exponential Malthusian exponential growth function (i) or one-phase exponential decay (j). Data points represent median ± 95% confidence intervals. N = 578 (wt), 456 (ftsN-∆SPOR), 427 (∆envC), 501 (ftsL*) cells. k, Representative images from 3 biological replicates of indicated strains after 15, 20 and 30 min pulses with MurNAc-alkyne. Scale bar, 2 µm. Source data

Fig. 4. sPG hydrolysis is required for normal Z-ring placement and condensation in E. coli.

a, Distribution of cell wall material in ∆envC cells was assessed by FDAA staining in 3 biological replicates. Images are sum-projections of a 1 µm spanning z-stack and were deconvolved. White arrowheads indicate double septa. b, Representative time-lapse series from 3 biological replicates of a ∆envC mutant expressing Pal-mCherry and ZipA-sfGFP as OM and IM markers, respectively. An example of double septum formation is shown. c, Examples of membrane blebbing (yellow arrowheads) and polar septa (blue arrowheads) formation are highlighted. d, Formation of double constrictions observed in cryo-electron tomograms of ∆envC ∆nlpD cells. Black arrowheads indicate constriction sites. e, The frequency of double septum formation was quantified from counting the number of Pal-mCherry doublets per cell. No Pal doublets were found in >10,000 cells for wild-type or ftsN-∆SPOR cells in 3 biological replicates (N.A., not applicable). Data are represented as median + 95% confidence interval. f, The distance between Pal doublets was measured manually using the line tool in Fiji. N = 91 (∆envC), 46 (ftsL* ∆envC) Pal doublets measured. g, The frequency of polar septa per cell was measured for the indicated strains. No polar septa were observed in >10,000 wild-type or ftsN-∆SPOR cells. Data are represented as median + 95% confidence interval. h–j, Three-dimensional maximum intensity renderings showing Z-ring condensation based on ZipA-sfGFP localization (h). The degree of Z-ring condensation was quantified from averaged fluorescence intensity projections from summed 3D volumes (i) or from 5 time points (corresponding to 10 min) of a time-lapse series (j) (see Methods). Insets: FWHM of the fluorescence signal, with data represented as boxplots; line represents median, error bars depict minimum–maximum range. Inserts show average fluorescence intensity projection at the septum. Significance was tested against wild type by one-way ANOVA with Dunnett’s correction for multiple comparisons: *P < 0.05. N = 100 (wt, ∆envC, ftsL* ∆envC, ftsN-∆SPOR) Z-rings from 3 biological replicates. Averaged Z-rings are shown and colour-coded according to graphs. Scale bars: a–c, 2 µm; d, 200 nm; h, 2 µm; i and j, 200 nm. Source data

Fig. 5. Competition between the divisome and elongation machinery defines polar cell shape in E. coli.

a, MreB dynamics were followed by SIM-TIRF in indicated strains (see Methods). Time-lapse series were sum projected and overlayed with single-particle tracking results from TrackMate and 3D-SIM Pal-mCherry reference images. The Pal-mCherry signal serves to identify constricting cells. Early division site (yellow arrowheads) displayed Pal foci that were resolvable as two distinct foci, whereas late division sites (blue arrowheads) displayed a continuous Pal signal across the cell, indicative of complete or near-complete cytokinesis. b, Directionally moving MreB tracks were filtered by MSD analysis (see Methods), represented as boxplots (line indicates median; error bars depict minimum–maximum range) and normalized by cell area. Significance in each group was tested against wild type by one-way ANOVA with Dunnett’s correction for multiple comparisons: *P < 0.05, **P < 0.01; NS, P ≥ 0.05. N = 30 (wt, ftsN-∆SPOR, ∆envC, ftsL*) time-lapse series from 3 biological replicates. c, Representative phase-contrast micrographs showing segmented cells in ‘Morphometrics’ for the indicated division mutants. d, Summed, projected central 3D slices through cryo-electron tomograms of indicated strains visualizing cell poles. Black arrowheads indicate 3D-rendered pole. The corresponding 3D-volume renderings show polar curvature determined by shape index (see Methods). e–g, Polar curvature was measured by the two highest points of positive cell outline curvature (f), while constriction curvature was assessed by measuring the opposing contour-matched lowest curvature values at the division site (g) using Morphometrics and normalized to cell width (see Methods). Polar and division site curvatures are negatively correlated (R² = 0.27) (e). Data are represented as mean ± s.d. For f and g, significance was tested against wild type by one-way ANOVA with Dunnett’s correction for multiple comparisons: ***P < 0.001, ****P < 0.0001; NS, P = 0.057. N = 460 (wt), 999 (ftsL*), 292 (ftsN-∆SPOR), 164 (∆envC) cells from 3 biological replicates. Scale bars: a, 1 µm; c, 2 µm; d, summed projection images, 200 nm and 3D renderings, 100 nm. Source data

Fig. 6. Septal PG architecture and divisome activity modulate bacterial morphogenesis in E. coli.

a, Wild-type E. coli divides via a mixed constriction-septation mechanism in which a partial septum with two discernible plates of sPG is formed at later stages of the division process. A wedge structure is observable at the lagging edge of the septum where the dual layers of PG signal of the developing septum meet the single-layered signal of the side wall. Although not clearly resolved in the tomograms, we assume that the two layers of PG signal within the septum are probably connected by additional PG material (drawn as crosshatches). b, A constrictive mode of cell division is observed for the ftsN-∆SPOR mutant, where OM and IM invaginate at similar velocities due to lower sPG synthesis rates. The result is a V-shaped constriction that is similar to that formed by the distantly related Gram-negative bacterium C. crescentus. In contrast, inhibition of sPG hydrolysis causes a temporal separation of IM and OM constriction, leading to septation. These septa as well as the partial septa in wild-type cells are reminiscent of the Gram-positive bacterium S. aureus, which also displays two distinctive plates of sPG within its septa. c, The activities of the two major synthetic cell wall machineries, the Rod complex and the divisome, are anti-correlated probably due to competition for limited substrate (lipid II). The balance of their relative activities determines the shape of the cell division site and the resulting poles they form. Cells with higher Rod complex activity are thinner and form pointier poles, while cells with elevated divisome activity are shorter and wider, with blunt poles.

Extended Data Fig. 1. Cryo-FIB / cryo-ET pipeline utilized in this study.

Schematic cartoons showing the steps in sample preparation for cryo-ET. In brief, bacteria are grown to OD₆₀₀ = 0.3 and applied onto an EM grid for vitrification in liquid ethane. Cryo-EM grids are kept in liquid nitrogen until transfer into the cryo-FIB microscope for milling. An illustration shows the result of milling vitrified bacteria distributed onto the holey carbon film on the mesh EM grid. (a-d) Images taken from the Aquilos Thermo Fisher Scientific graphical user interface during cryo-FIB milling performance. (a) Target bacteria (red circles) are first identified by SEM, e^- beam. Yellow box indicates region visualized in (b) and magenta box indicates the area visualized in (c). (b) Corresponding FIB, ion Ga⁺ beam, view (52° with respect to the e^- beam) of the targeted grid square in (a) (yellow box). Green box indicates region visualized in (d). (c) SEM view of the same region shown in (a) and (b) after platinum deposition and milling. Cyan box indicates obtained lamella shown in (e). (d) FIB view of region shown in (b) (green box) after platinum deposition and milling. Scale bars in (a-d) are indicated on each image. After milling, cryo-EM grids containing bacterial lamellae are transferred into a TEM microscope. (e) Low magnification TEM 2D image of the lamella shown in (c-d), cyan box. Dashed black box indicates target region for cryo-ET acquisition. Dashed white line indicates the tilt axis for cryo-ET data acquisition. (f) 3D slice of the cryo-electron tomogram obtained from 3D reconstruction of aligned cryo-ET tilt series acquired in (e) (dashed black box). Scale bars: e = 1000 nm; f = 200 nm. (g) A representative lamella from a wild-type E. coli cell imaged at indicated imaging conditions. White box highlights region for corresponding high-magnification acquisition. Scale bars = 1 µm (low magnification); 200 nm (high magnification and cryo-electron tomogram).

Extended Data Fig. 2. Distance measurements in cryo-ET data of dividing E. coli cells.

Three dimensional slices visualizing the division site during (a) constriction, (b) septation and (c) cytokinesis. Dashed white line indicates OM-OM distance and white bold line indicates IM-IM distance. (a.i-c.i) Measured distances in nm of IM-IM, OM-OM and the difference between these distances at (a.i) constriction (N = 7 (wt); 10 (ftsN-∆SPOR); 5 (∆envC); 4 (∆nlpD ∆envC); 8 (ftsL*) images), (b.i) septation (N = 5 (wt); 4 (ftsN-∆SPOR); 3 (∆envC); 2 (∆nlpD ∆envC); 3 (ftsL*) images), and (c.i) cytokinesis (N = 5 (wt); 4 (ftsN-∆SPOR); 11 (∆envC); 3 (∆nlpD ∆envC); 4 (ftsL*) images). (a.ii-b.ii) Measured distances in nm for each strain grouped. Scale bars = 200 nm. All data are expressed as mean + SEM. (d-f) Schematic representing the division stages color-coded at where periplasmic width was measured. Thirty euclidean distances were measured per region (see Methods), N values for each region per stage are: (d) at septum (N = 14 (wt); 16 (ftsN-∆SPOR); 7 (∆envC); 6 (∆nlpD ∆envC); 10 (ftsL*)); at curve (N = 15 (wt); 27 (ftsN-∆SPOR); 13 (∆envC); 8 (∆nlpD ∆envC); 19 (ftsL*)); (e) at septum (N = 10 (wt); 8 (ftsN-∆SPOR); 6 (∆envC); 6 (∆nlpD ∆envC); 6 (ftsL*)); at curve (N = 19 (wt); 23 (ftsN-∆SPOR); 12 (∆envC); 6 (∆nlpD ∆envC); 12 (ftsL*)); (f) at septum (N = 4 (wt); 3 (ftsN-∆SPOR); 12 (∆envC); 3 (∆nlpD ∆envC); 6 (ftsL*)); at curve (N = 16 (wt); 12 (ftsN-∆SPOR); 39 (∆envC); 6 (∆nlpD ∆envC); 7 (ftsL*)). All analyzed data points are displayed, bar represents mean + SD. Significance was tested using unpaired t-test with Welch correction for paired groups when data followed gaussian distribution and Mann-Whitney when data did not follow gaussian distribution and tested relative to wild-type. Significance difference among all groups compared was tested using Welch ANOVA and Brown-Forsythe when data followed gaussian distribution and Kruskal-Wallis test when data did not follow gaussian distribution. In (a.ii) difference plot, statistical significance is shown based on F-test statistics. Ns = non-significant, * = p < 0.05, ** = p < 0.01, *** = p < 0.001, **** = p < 0.0001. Source data

Extended Data Fig. 3. Subtomogram averaging, NAD filtering and segmentation of the cell envelope of E. coli.

(a) STA 3D structure of the cell envelope at the septum and side wall are displayed in Chimera using solid and surface rendering. 3D slices of averages are shown. 212 particles contributed to the septum average while 8072 particles from N = 5 tomograms contributed to the side wall average. Blue dots plotted on a tomogram slice represent particles that contributes to ‘side wall’ and yellow dots represent particles contributing to septum average. Scale bar = 40 nm. In the tomogram rendering, 100 pixels blocks in the cartesian axes correspond to 102.6 nm. (b-f) Gallery of corresponding zoom-in summed projected central slices of cryo-electron tomograms visualizing the indicated division mutants. First column shows original image, second column shows filtered image, and third column shows filtered image with segmentation layers indicating IM = green, PG = cyan and OM = magenta. A full cryo-ET gallery can be found in Extended Data Fig. 5. A complete overview of the number of tomograms is reported in Supplementary Tables 2, 3. Scale bars = 100 nm.

Extended Data Fig. 4. Measuring cell envelope constriction from kymograph data.

(a) Schematic representation of workflow for the generation of kymographs using Kymoclear and KymogrphaDirect software. Instantaneous constriction velocity for (b) IM (ZipA-sfGFP) and (c) OM (Pal-mCherry) are plotted against normalized cell width. Bold lines show second order polynomial fits as in Fig. 2e,f. N = 150 cells (wt); N = 48 (ftsN-∆SPOR); N = 81 (∆envC); N = 68 (ftsL*) kymographs. (d) Additional examples of cell envelope constriction kymographs for the corresponding strains in panels b-c. (e) Duration of IM (left) and OM (right) constriction was derived from kymograph measurements. Data are represented as boxplots. The line represents median; error bars depict Min-Max range. The significance of differences were tested relative to wild-type by one-way ANOVA with Dunnett’s correction for multiple comparisons; ns = non-significant (p = 0.99), * = p < 0.05, *** = p < 0.001, **** = p < 0.0001; N = 150 cells (wt); N = 48 (ftsN-∆SPOR); N = 81 (∆envC); N = 68 (ftsL*). Scale bar = 2 µm, in kymographs = 200 nm horizontal and 10 min vertical. Source data

Extended Data Fig. 5. Cell division and polar morphology of E. coli viewed by cryo-ET.

Gallery of summed projected central slices of cryo-electron tomograms visualizing the indicated division mutants. Black arrowhead = division site; green arrowhead = envelope bulging. Dashed white box indicates corresponding zoom-in region show in Extended Data Fig. 3. A complete overview of number of tomograms is reported in Supplementary Tables 2, 3. Scale bars = 200 nm.

Extended Data Fig. 6. Measuring bulk growth rates of E. coli cell division mutants analyzed in this study.

Growth curves were measured in biological triplicates by OD₆₀₀ readings in a 96-well plate reader at 30 °C. Data is represented as mean ± SD. (a) Untagged strains used for cryo-ET and FDAA labeling as well as (b) ∆murQ mutants expressing amgK and murU for MurNAc-alykyne labeling experiments were grown in LB. Cells harboring fluorescent fusion proteins for live-cell imaging of (c) cell envelope constriction or (d) MreB tracking were grown in M9 medium supplemented with 0.2% glucose and casamino acids. Source data

Extended Data Fig. 7. A hyperactivated divisome leads to anisotropic cell envelope constriction.

(a) Orthogonal views of XZ and XY slices of 3D cryo-electron tomograms of the indicated division mutants. Magenta and green arrowheads indicate OM and IM, respectively. 3D volumes are displayed in cartesian 3D grids with axes indicating the dimensions in pixels. For wild-type and ∆envC ∆nlpD 100 pixels = 102.6 nm, and for ftsN-∆SPOR and ftsL* 100 pixels = 110.3 nm. (b) Corresponding 3D surface segmentation renderings of OM (magenta) and IM (green) are shown on the right. (c) Schematic overview of a theoretical kymograph for an isotropic (left) and anisotropic (right) constriction of the cell envelope. Representative examples form 3 biological replicates for wild-type (left) and ftsL* (right) are provided. Scale bars: 200 nm (vertical); 5 min (horizontal). (d) An anisotropy score was calculated by taking the ratio of the constriction velocity from both sides of the cell. Red line (= 1) indicates a perfectly isotropic cell envelope constriction process. Data are represented as mean ± SD, Kruskal-Wallis with Dunn’s correction for multiple comparisons among all values was calculated, exact p values are shown. N = 65 (wt); 24 (ftsL*); 23 (ftsN-∆SPOR); 44 (∆envC) constriction were analyzed (e) Cells were vertically immobilized using small micro pillars imprinted into agarose pads, allowing to image the cell division site along its long axis. Representative example of the cell envelope position in vertically imaged wt and ftsL* cells. Scale bar = 2 µm. Circularity was quantified using Morphometrics. Red line (circularity = 1) indicates a perfect circle. Black line indicates median. Two-way ANOVA with Sidak’s multiple comparison test; * = p < 0.05; *** = p < 0.001. N = 132 (wt); 172 (ftsL*) cells imaged in three biological replicates. Source data

Extended Data Fig. 8. Cell wall synthesis and hydrolysis measurements.

(a) New and old cell wall material were detected with Alexa488 labelled MurNAc-alykyne (yellow) or HADA (blue), respectively. (b) Representative images from labeling. (c) Label incorporation at the division site. Bars show the mean and dots show the average from 9 different images. ftsN-∆SPOR labeling example. (d) Rate of sPG synthesis (line = median). (e) Septal PG hydrolysis. Bar represents median + 95 % confidence interval; points indicate the average of three biological replicates. (f) Side wall labeling (line = median). (g) sPG and side wall synthesis ratio. (h) Representative YADA labeling pattern. Yellow arrow heads indicate polar label accumulation. (i) Average polar YADA fluorescence. Black line indicates mean, one-way ANOVA with Dunnett’s correction for multiple comparison. Significant differences are relative to wild-type; ns = non-significant (p = 0.76), p < 0.00. (j) Summed projected central 3D slices through tomograms of poles. Yellow arrowhead indicates enlarged periplasm. (k) Periplasm thickness in cryo-ET data at the side wall (green) and pole (red) (mean + SD). Thirty Euclidean distances were measured per region. One-way ANOVA (side wall) and Kruskal-Wallis test (pole); significance was tested among all groups within each region; ns = non-significant (p = 0.91), **** = p < 0.0001, N.A. = not applicable. (l) Labeling patterns observed for the pulse-chase. (m) Sum-projection of deconvolved images after HADA pulses as shown in Fig. 3b. (n) Expected labeling patterns for cells expressing sulA. New and old wall is labelled with HADA (blue) and YADA (yellow), respectively. (o) Representative images of indicated strains after HADA pulses. Mean side wall fluorescence intensity for (p) HADA and (q) YADA fit to a (p) Malthusian exponential function or (q) one phase exponential decay. Mean fluorescence intensity ± one SD is shown, N = 360 cells each. (r) Fluorescence intensity (p,q) was measured in original non-deconvolved (raw) SUM projections. Linear regression shows strong positive (R² = 0.87) correlation to deconvolved fluorescence intensity values. N = 1035 values. Scale bars = 2 µm (fluorescence) or 200 nm (cryo-ET). Source data

Extended Data Fig. 9. Z-ring views during constriction in cryo-electron tomograms of E. coli.

Summed projections of 10 slices of XZ and XY views during constriction of indicated strains. Representative examples of all strains are shown. Green arrowheads indicate IM, magenta arrowheads indicate OM, red arrowheads indicate cytoskeletal ring and red asterisks indicate zones of diffuse signal. Note, Z-ring signal is weaker in ∆envC mutants due to issues with Z-ring condensation as shown by fluorescence microscopy (Fig. 4h-j) Scale bar = 100 nm.

Extended Data Fig. 10. The balance between elongation and division affects cell morphology.

(a) MreB-sw-mNeonGreen dynamics were followed by SIM-TRIF microscopy for 3 min at 3 s acquisitions per frame in indicated mutants. Time-lapse series was sum-projected and overlayed over a 3D-SIM Pal-mCherry and brightfield reference image. Larger fields of view are shown as compared to Fig. 5g and are representative from three biological replicates. Bar = 1 µm (b) Slopes of MSD curves (α) were analyzed following log-log fit to log[MSD] versus log[t] using the MATLAB class msdanalyzer. Particles displaced by diffusive motion are characterized by a slope of their log[MSD] = 1, while transported particles have slopes of 2 and constrained particles display slopes < 1. No significant difference (ns = non-significant, p = 0.0693; Kruskal Wallis test with Dunn’s correction for multiple comparisons) in the slopes of MSD curves were found indicating that MreB is displaced at a similar rate and manner in all strains. Box plot error bars displaying Min-Max range of values, blackline represents median. Tracks fit to log[MSD] log[t] with R² ≤ 0.95 (c) Cell length and (d) width was measured from three independent biological replicates for the indicated mutants using Morphometrics. Line represents median. Differences in significances were tested relative to wild-type using Kruskal-Wallis one-way ANOVA with Dunnett’s correction for multiple comparisons; * = p < 0.05, **** = p < 0.0001, ns = non-significant, p = 0.69. (e) MreB dynamics were imaged as in (a). Representative temporal SUM projections of MreB-mNeon trajectories were overlayed to brightfield images. Bar = 2 µm. (f) Directionally moving MreB tracks were filtered by MSD analysis (see Methods) and represented as boxplots (line indicating median; error bars depict Min-Max range) and normalized by cell area. Significance in each group was tested against non-filamented control (-Ara) by two-sided unpaired t-test; ** p < 0.01, ns = non-significant, wt p = 0.053, fstN-∆SPOR = 0.253. (g) Constriction curvature values of wild-type cells are plotted against division site width in 566 cells. Linear regression (R² = 0.135) indicates the negative correlation between cell width at the division site and constriction angle. Source data