Phenotypic plasticity in cell elongation among closely related bacterial species

Abstract

Cell elongation in bacteria has been studied over many decades, in part because its underlying mechanisms are targets of numerous antibiotics. While multiple elongation modes have been described, little is known about how these strategies vary across species and in response to evolutionary and environmental influences. Here, we use fluorescent D-amino acids to track the spatiotemporal dynamics of bacterial cell elongation, revealing unsuspected diversity of elongation modes among closely related species of the family Caulobacteraceae. We identify species-specific combinations of dispersed, midcell and polar elongation that can be either unidirectional or bidirectional. Using genetic, cell biology, and phylogenetic approaches, we demonstrate that evolution of unidirectional-midcell elongation is accompanied by changes in the localization of the peptidoglycan synthase PBP2. Our findings reveal high phenotypic plasticity in elongation mechanisms, with implications for our understanding of bacterial growth and evolution.

Fig. 1. Diversity of growth modes in rod-shaped bacteria and members of the Caulobacteraceae family.

a Schematic illustrating different growth modes observed among rod-shaped bacteria. Green lines represent lateral or dispersed elongation, solid green denotes different modes of localized cell elongation, and solid blue marks cell division. b Dimorphic cell cycles and growth modes in C. crescentus, A. excentricus, and A. biprosthecum. C. crescentus swarmer cells elongate by dispersed cell elongation (green lines) before bidirectional midcell elongation (green), followed by cell division (blue). A. excentricus swarmer cells also undergo dispersed elongation (green lines). However, this is followed by an uncharacterized, unidirectional midcell elongation towards the new pole (green) prior to bidirectional cell division (blue). A. biprosthecum cells elongate through yet another novel mode of elongation, a combination of polar and unidirectional midcell elongation. c Schematic depicting the pulse-chase experiment using FDAA. Whole-cell PG was labeled with FDAA (green) over two generations, followed by washes with PYE to remove free FDAA from the medium. Subsequent growth in the absence of FDAA was observed using time-lapse microscopy. During the chase period, the loss of the FDAA signal corresponds to new PG synthesis/turnover. The magenta dashed line indicates the position of the ZapA fluorescent protein fusion, as a marker of the future division site. d Pulse-chase experiments using FDAAs in C. crescentus and A. excentricus cells carrying fluorescent fusions of ZapA. Images were taken every 5 min during the chase period. Kymographs show the loss of FDAA fluorescence as the cells grow. In C. crescentus (left), the FDAA BADA was used in combination with ZapA-mCherry. In A. excentricus (right), the FDAA TADA was used in combination with ZapA-sfGFP. Kymographs present the FDAA signal in green and the ZapA fluorescent fusion signal in magenta. White stars indicate cells starting division. See Supplementary Fig.1 for additional kymographs from each species, n  =  3 biological replicates. Scale bars: 2 µm (cell length). e Schematic of FDAA signal loss in C. crescentus and A. excentricus cells. The magenta dashed line indicates the position of ZapA as a marker of the future division site. Green shading represents the old PG labeled with FDAA.

Fig. 2. Sequential FDAA labeling reveals that C. crescentus grows bidirectionally from the midcell, while A. excentricus grows unidirectionally towards the new pole.

a Schematic depicting the dual short-pulse experiment: sequential, dual short FDAA pulses show sites of active PG synthesis. Cells were first labeled with one FDAA (HADA in C. crescentus or TADA in A. excentricus, magenta) for 5% of their generation time, washed with PYE to remove the free FDAA, allowed to grow for 15 % of their generation time, and then labeled with a second FDAA (green) for 5% of their generation time, washed again, and imaged with microscopy. b–d Sequential FDAA labeling in C. crescentus (b) and A. excentricus c,d. Top: Representative images are shown (FDAA 1, FDAA 2, and merge, n  =  3 biological replicates). Scale bars: 2 µm. Bottom: population-level demographs showing the localization of the fluorescence intensities of FDAA 1 and FDAA 2, and graphs showing the fluorescence intensity of the FDAAs against their relative positions along the cell length for 50 cells. In (d), the old pole in A. excentricus cells is additionally labeled with fluorescent WGA (cyan). In all demographs, cells are arranged by length, with 50% of maximum fluorescence intensities shown. Demographs in Panels (b,c) are oriented with the maximum fluorescence intensity of the second FDAA to the right, and in Panel d with the WGA-labeled old pole (cyan) to the left. In (b,c, and d), the white brackets show the 50 cells selected to plot the fluorescence intensities of the two FDAAs. The lines represent the mean values, with error bars showing the standard error of the mean (SEM). See Supplementary Fig. 2 for additional demographs. Source data are provided as a Source Data file. e Schematic illustrating PG synthesis during A. excentricus cell cycle. Swarmer cells undergo dispersed cell elongation (green dots). Stalked cells elongate unidirectionally from the midcell towards the new pole, with the first FDAA signal (magenta) located on the new pole side of the second FDAA signal (green). Predivisional cells exhibit bidirectional growth at the midcell, with magenta signals on both sides of the second FDAA signal (green). The old pole is indicated by the holdfast (cyan).

Fig. 3. Distinct PBP2 localization in A. excentricus and C. crescentus and its correlation with active PG synthesis.

a Subcellular localization of mCherry-PBP2 in A. excentricus. Top: Representative images are shown (n  =  3 biological replicates). Left to right: Phase, mCherry-PBP2 fluorescence, and merged images. Scale bar: 2 µm. A heatmap of mCherry-PBP2 foci at the population level is displayed. In the heatmap, cells were oriented using the old pole labeled with WGA (cyan), with the white line indicating the midcell. Bottom: A demograph showing the localization of mCherry-PBP2 fluorescence at the population level, with cells arranged by length and oriented with the old pole towards the right. b Short-pulse FDAA (BADA) labeling of the A. excentricus mCherry-PBP2 strain. Top: A schematic depicting the experiment. Cells were labeled with 250 μM BADA for 5% of their generation time, fixed with 70% (v/v) ethanol, and imaged. Middle: Representative phase, fluorescence (mCherry-PBP2 and BADA), and merged images are shown (n  =  3 biological replicates). Scale bar: 2 µm. Bottom: Population-level demographs showing the localization of the fluorescence intensities of mCherry-PBP2, BADA, and their overlays. See Supplementary Fig. 4a for heatmaps and density maps of mCherry-PBP2 and BADA at the population level. c Short-pulse FDAA (TADA) labeling of the C. crescentus GFP-PBP2 strain. Top: A schematic depicting the experiment. Cells were labeled with 250 μM TADA for 5% of their generation time, fixed with 70% (v/v) ethanol, and imaged. Middle: Representative phase, fluorescence (GFP-PBP2 and TADA), and merged images are shown (n  =  3 biological replicates). Scale bar: 2 µm. Bottom: Population-level demographs showing the localization of the fluorescence intensities of GFP-PBP2, TADA, and their overlays. See Supplementary Fig. 4b, c for additional fluorescence images and population-level demographs showing the differences in localization of PBP2 in A. excentricus vs. C. crescentus.

Fig. 4. PBP2 is responsible for unidirectional PG synthesis at the midcell in A. excentricus.

a Schematic depicting pulse-chase experiments in A. excentricus ZapA-sfGFP cells with (right) or without (left) mecillinam treatment. Whole-cell PG was labeled with TADA (magenta) over two generations, followed by washing and growth with or without mecillinam (50 µg ml⁻¹) over one generation before imaging. Representative phase, fluorescence (TADA and TADA overlaid with ZapA-sfGFP), and merged images with WGA fluorescence are shown (n  =  3 biological replicates). See Supplementary Fig. 5c for population-level demographs. Scale bar: 2 µm. b ShapePlots of A. excentricus cells after an FDAA pulse-chase experiment, with mecillinam. Top: Shape plots showing bulge localization. Bottom: ShapePlots show ZapA-sfGFP and FDAA signal loss. Each ShapePlot is divided longitudinally (black line) with FDAA signal loss on the left and overlaid with ZapA-sfGFP on the right. ShapePlots show four categories of cells binned by cell length (left) and the entire population of 160 cells (right) oriented using WGA-labeled holdfast (old pole at the bottom). Horizontal white lines represent the midcell. c Top: Schematic of a dual short-pulse experiment with or without mecillinam (50 µg ml⁻¹). A. excentricus cells were grown over 120 min with or without mecillinam, then labeled with TADA (magenta), washed, allowed to grow, and labeled with BADA (green). Cells were washed again and imaged with microscopy. Middle: Representative phase and merged images from both treatment conditions (n  =  3 biological replicates). Merged images show phase contrast overlaid with fluorescence signals from the two FDAAs and WGA-labeled holdfast (cyan). Scale bar: 2 µm. Look-up tables (LUTs) were adjusted for each condition to have a visible FDAA signal. Bottom: Population-level demographs showing the fluorescence intensities of both FDAA signals, with and without mecillinam. Cells were arranged by length, with the old pole to the left. 50% of the maximum fluorescence intensities are shown. d Dot plots graphs showing quantification of fluorescence intensities for each FDAA pulse, with or without mecillinam (n = 686 cells control and n = 514 mecillinam treated-cells). **** P < 0.0001, unpaired two-tailed t-test with Welch’s correction. Error bars show the SEM. Source data are provided as a Source Data file.

Fig. 5. Diversity of cell elongation modes in members of the Caulobacteraceae family.

a Schematic depicting the pulse-chase experiment using FDAAs. Whole-cell PG was labeled with TADA for two generations, washed to remove excess FDAA, followed by a chase period of 75 min before imaging. b Representative merge images are shown (n  =  3 biological replicates). Demograph analysis of pulse-chase experiments in WT cells of C. crescentus, C. henricii, P. conjunctum, B. diminuta, A. excentricus, and A. biprosthecum. Cells are arranged by length, with each cell oriented so that the pole with the maximum fluorescence intensity is to the left. White stars indicate signal loss at one pole. Red arrows highlight unidirectional midcell elongation in A. excentricus, A. biprosthecum, and P. conjunctum, as observed by greater FDAA signal loss towards the right of the demograph. Scale bar: 2 µm. c Phylogenetic tree of representative species from the Alphaproteobacteria, which includes the family Caulobacteraceae. Taxon label colors correspond to different modes of cell elongation: dispersed cell elongation (green), unidirectional midcell elongation (red), polar elongation (blue), budding (violet), polar and unidirectional midcell elongation (orange and red), bidirectional midcell elongation (magenta), binary fission (black), and unknown (gray). Species for which the cell elongation mode has been studied using FDAAs, TRSE, or other methods are highlighted in bold. The node where pbp2 and mreB (and the associated mreCD and rodA genes) are predicted to have been lost within the Rhizobiales is highlighted in blue. The tree, based on a concatenation conserved protein-coding gene sequences, is fully supported, with posterior probabilities of 1 for all clades. See the Methods section for details on phylogenetic reconstruction and refer to Supplementary Table S4 for genome IDs. d This schematic, derived from a pruned version of the phylogenetic tree in Fig. 5a to highlight the species of interest, illustrates the identified modes of cell elongation and the possible number of transitions. The transitions are depicted assuming two scenarios: unidirectional cell elongation as the ancestral state (left, red) or bidirectional midcell elongation as the ancestral state (right, magenta). Colored lines and arrows (red or magenta) indicate where the transitions might have occurred.

Fig. 6. FDAA labeling experiments demonstrate polar and unidirectional midcell elongation in A. biprosthecum and unidirectional midcell elongation in R. capsulatus.

a Schematic depicting the FDAA pulse-chase experiments in A. biprosthecum and R. capsulatus. Whole-cell PG was labeled with 500 µm TADA (green) over two generations, followed by washes to remove free FDAA from the medium. Subsequent growth in the absence of the FDAA was followed by time-lapse microscopy. b,f Kymographs of the pulse chase experiments showing the loss of FDAA fluorescence during the chase period in (b) A. biprosthecum and (f) R. capsulatus. Images were taken every 5 min (n  =  3 biological replicates). Scale bars: 2 µm. See Supplementary Fig. 6 and 7 for additional kymographs. c Schematic depicting the dual short-pulse experiment. Cells were first labeled with TADA for 5% of their generation time, washed with fresh media, allowed to grow without FDAA, and then labeled with BADA for 5% of their generation time. Cells were then washed again and imaged with microscopy. d,g Representative images and demographs showing the fluorescence intensity of both FDAA signals in (d) A. biprosthecum and (g) R. capsulatus (n  =  3 biological replicates). In the demographs, cells were arranged by length, with the old pole (labeled with WGA) to the left in A. biprosthecum, and with the maximum fluorescence intensity of the first FDAA to the right in R. capsulatus. 50% of the maximum fluorescence intensities are shown. The white brackets indicate the 50 cells selected to show the fluorescence profiles of the two FDAAs in (e and f). Scale bars: 2 µm. See Supplementary Figs. 6 and 7 for additional demographs. e, h Fluorescence intensity profiles of FDAA 1 (TADA, magenta line) and FDAA 2 (BADA, green line) in (e) A. biprosthecum cells and (h) R. capsulatus cells, plotted from n  =  50 cells for both species. Points were selected along the medial axis of each cell, and the normalized signal was plotted relative to position along the cell length. The lines represent the mean values, with error bars showing the standard error of the mean (SEM). Source data are provided as a Source Data file.