Helical insertion of peptidoglycan produces chiral ordering of the bacterial cell wall

Abstract

The regulation of cell shape is a common challenge faced by organisms across all biological kingdoms. In nearly all bacteria, cell shape is determined by the architecture of the peptidoglycan cell wall, a macromolecule consisting of glycan strands crosslinked by peptides. In addition to shape, cell growth must also maintain the wall structural integrity to prevent lysis due to large turgor pressures. Robustness can be accomplished by establishing a globally ordered cell-wall network, although how a bacterium generates and maintains peptidoglycan order on the micron scale using nanometer-sized proteins remains a mystery. Here, we demonstrate that left-handed chirality of the MreB cytoskeleton in the rod-shaped bacterium Escherichia coli gives rise to a global, right-handed chiral ordering of the cell wall. Local, MreB-guided insertion of material into the peptidoglycan network naturally orders the glycan strands and causes cells to twist left-handedly during elongational growth. Through comparison with the right-handed twisting of Bacillus subtilis cells, our work supports a common mechanism linking helical insertion and chiral cell-wall ordering in rod-shaped bacteria. These physical principles of cell growth link the molecular structure of the bacterial cytoskeleton, mechanisms of wall synthesis, and the coordination of cell-wall architecture.

Fig. 1.

MreB/C/D form left-handed helices in E. coli. (A) A representative z-series of YFP-MreB fluorescence (Left to Right, then Top to Bottom). (B) Schematic illustration of the z-series cylindrical unwrapping algorithm and radon analysis. (C–E) Gallery of unwrapped fluorescence distributions of MreB (C), MreC (D), and MreD (E). A left-handed helix is a straight line directed up and to the right. Scale bar represents 1 μm. (F–H) Distributions of the helix angles from radon transform analysis for MreB (F, N = 36 cells), MreC (G, N = 24 cells), and MreD (H, N = 42 cells). A negative angle corresponds to a left-handed helix. Magenta dotted lines indicate the mean values. Strain: (C, F), MC1000/pLE7 (Plac∷yfp-mreB); (D, G), MC1000/pVP1 (Plac∷yfp-mreC); (E, H), MC1000/pVP2 (Plac∷yfp-mreD).

Fig. 2.

E. coli cells twist in a left-handed direction during elongation. (A) Schematic illustration of the different helix angles used in this paper. (B) A cell expressing cytoplasmic mCherry (red) with two Dragon Green-labeled beads (yellow) bound to its top surface imaged at two instances in time. (i) DIC microscopy, (ii) fluorescence microscopy, and (iii) computational extraction of the cell outline and bead centers all indicate a left-handed twist (arrow). Scale bars represent 2 μm. (C, D) Circumferential and longitudinal distances between the two beads are linearly related for both cephalexin-treated (C) and untreated (D) cells. (Inset) DIC and fluorescence images of a representative cell, scale bar represents 1 μm. Line fits (solid lines) and 95% confidence intervals of the fits (dotted lines) are shown. The growth angle in (D), deduced from the slope of the fit line, is -6.4 ± 0.7° (95% confidence interval). (B–D) Strain, MC4100 ara^R/pSW1 (P_BAD ∷ mCherry). (E) Circumferential and longitudinal distances between two fluorescently labeled flagellar hooks are also linearly related. (Inset) Fluorescence and analyzed images of a representative cell expressing cytoplasmic EGFP (scale bar represents 5 μm). The growth angle for this cell is -5.2 ± 0.5° (95% confidence interval). Strain, MTB9/pWR20 (EGFP). (F) Mean and standard error of the growth angle of various E. coli strains. N = 10 for YS34. N = 5 for MC1000, MC4100 ara^R/pSW1, MC4100 ara^R/pSW1—no cephalexin, and WA220. N = 3 for MTB9/pWR20. The YS34 and MC1000 cells are labeled with membrane dye FM4-64. (G) Ensemble distribution of the growth angles in F. The magenta line indicates the ensemble average.

Fig. 3.

The growth twist of E. coli depends on the MreB cytoskeleton. (A) DIC images of an FtsZ-depleted cell grown at room temperature with A22. Scale bar represents 5 μm. (B) The circumferential and longitudinal distance between the beads are linearly related during a phase of twisting elongation until twisting ceases and circumferential distance remains constant as the cell elongates during straight elongation. (A, B) Strain, WX7 (ΔftsZ)/λGL100(Plac∷ftsZ). (C) The growth angle (mean and standard error) of E. coli cells with and without A22 treatment. A22 treatment, which causes rapid depolymerization of MreB, results in a slow decrease in the growth angle over a timescale of one doubling time. YS34: without A22 (N = 10). A22 (short): cells were incubated with A22 for 3 min before observation, and imaged for 45 min at room temperature in the presence of A22 (N = 12). A22 (long): cells were incubated with A22 for 30 min at 37 °C before observation, and then imaged at room temperature in the presence of A22 (N = 10). WA220 (parental strain for A22-resistant strain WA221): without A22 (N = 5) and with A22 (long) (N = 6). WA221 (A22-resistant strain): without A22 (N = 5) and with A22 (long) (N = 5). Double asterisks (**) indicate significant difference from zero growth angle in t test, with p < 0.01.

Fig. 4.

Modeling of cell elongation demonstrates that helical peptidoglycan insertion gives rise to growth twist. (A) After glycan-strand insertion initiates near the MreB helix (yellow), synthesis and crosslinking of new strands (blue) proceeds along the direction defined by the existing glycan strands (green), which are initially oriented roughly circumferentially. Peptide crosslinks are in red. (B) Cell wall grown in silico to twice its initial length. Teal spheres represent beads bound to a single vertex of the peptidoglycan network. (C) Circumferential versus longitudinal distance between pairs of beads (blue curves) bound to vertices along the cell walls in B, with the average overall bead pairs shown in red. Twisting remains left-handed in a simulation with a small percentage of right-handed MreB helical segments (maroon), but handedness is abolished with equal levels of left-handed and right-handed segments (gray). (D) The growth angle decreases linearly with the MreB helix angle. Light blue region indicates one standard deviation above and below the average growth angle for pairs of beads on a single cell. Experimental measurements of MreB helix angle and growth angle are shown in black.

Fig. 5.

Modeling reveals chiral self-organization of peptidoglycan during helical insertion. (A) Average orientation of new, old, and all glycan strands during simulations of growth with MreB. Strands rapidly adopt a right-handed chirality opposite the direction of MreB, as illustrated for early, middle, and late stages of growth in B–D, respectively. (B–D) Cell wall at three stages of growth. Rectangles are zoomed-in versions of the regions marked on the corresponding cell walls. (E) Average orientation of glycan strands during simulations in which helical insertion has been replaced by uniform insertion at t = 0. (F) Circumferential versus longitudinal distances between representative pairs of beads during simulation in E.

Fig. 6.

Twist during osmotic shock reveals the chiral organization of peptidoglycan. (A) Illustration of the effect of osmotic shock on a chiral peptidoglycan network with right-handed glycan strands. Arrows indicate the glycan orientation, and teal and purple spheres correspond to virtual beads labeling the same vertices in all schematics. A decrease in turgor pressure increases the helix angle of the glycan strands, leading to a left-handed twist. (B) DIC and fluorescence images of an FM4-64-labeled E. coli cell before and after an osmotic up-shock. The beads twist in a left-handed direction (arrow), similar to the growth twist in Fig. 2. Scale bar represents 2 μm. (C) Histogram of stiffness angles for wild-type cells without A22 treatment. (D) The stiffness angle (mean and standard error) of E. coli YS34 cells with and without A22 treatment. Wild-type: without A22 (N = 6). A22 (short): cells were incubated with A22 for 3 min before shock (N = 7). A22 (long): cells were incubated with A22 for 1.5 h at 37 °C before shock (N = 7). Double asterisks (**) indicate significant difference from wild-type in t test, with p < 0.01.

Fig. 7.

Simulations reproduce left-handed osmotic twist. (A) Simulations of a peptidoglycan network with a steady-state average glycan angle after varying the turgor pressure from 0.9 to 1.1 atm. Circumferential and longitudinal displacement for the cell-wall vertices of the cell wall relative to P = 0.9 atm. The majority of the vertices display a left-handed osmotic twist. (B) Stiffness angle measurements from simulations in which the cell wall in A was further elongated with a random insertion pattern to mimic growth in the presence of A22, after which the turgor pressure was varied from 0.9 to 1.1 atm. (C) Circumferential and longitudinal displacement for the vertices of the cell wall in B relative to P = 0.9 atm. The average osmotic twist is nearly eliminated by nonhelical growth.

Fig. 8.

B. subtilis cells twist in a right-handed fashion during growth. (A) Filamentous B. subtilis cell labeled with beads whose rotation during growth indicates a right-handed twist (arrow). The cell membrane is labeled with FM4-64. (B) Circumferential and longitudinal distances between the two beads are linearly related. The ensemble growth angle is 4.3 ± 0.3° (N = 11). Line fits (solid lines) and 95% confidence intervals (dotted lines) are shown. Strain: wild-type B. subtilis 168. Scale bar represents 2 μm.

Fig. P1.

Helical insertion of material into the bacterial cell wall (green) during elongational growth, guided by the protein MreB (yellow), leads to an emergent chiral order in the cell-wall network and twisting of the cell that can be visualized using surface-bound beads (red).