Tissue-like structures formed by a bacterium

Abstract

Bacteria generally form only simple multicellular structures lacking the stable cell-cell connections characteristic of eukaryotic tissues. However, when the antibiotic moenomycin modifies peptidoglycan cell wall synthesis, rod-shaped cells of the Gram-negative bacterium Myxococcus xanthus become spherical, fuse their outer membranes, and assemble into stable, honeycomb-like lattices resembling eukaryotic tissues. These findings raise the intriguing possibility that some tissue-like organization could have evolved from stress-induced responses in bacterial ancestors.

The shift from single cells to multicellular tissues marks one of the most pivotal milestones in evolution. Multicellularity has arisen independently on at least 25 occasions across diverse evolutionary lineages, with particularly prominent examples in plants, animals, and fungi¹. Bacteria can also form multicellular structures, with myxobacteria standing out for their well-documented multicellular behavior. The model organism Myxococcus xanthus migrates in large, coordinated swarms across surfaces—a dynamic form of biofilm. Under starvation, these swarms give rise to fruiting bodies containing millions of spores encased in protective polysaccharide shells². However, bacterial multicellular structures differ fundamentally from eukaryotic tissues due to their lack of stable cell-cell connections. While relatively persistent assemblies like biofilms and fruiting bodies are maintained by extracellular matrices, they lack the direct intercellular contacts characteristic of true tissues^3,4. Conversely, direct connections—such as those formed through inner membrane (IM) nanotubes, outer membrane (OM) fusion, or various secretion systems—are typically transient and insufficient to support tissue-like architecture^5–8.

When PG synthesis in the rod-shaped M. xanthus (Fig. 1a)is stressed by the antibiotic moenomycin, the cells do not lyse; instead, they convert into spheres (Fig. 1b)⁹. Unlike L-forms that lack the peptidoglycan (PG) cell wall, often depend on osmo-protectants for survival¹⁰, and generate heterogenous, bubble-like progenies¹¹, these spheres can grow in low-osmolarity media and are uniform in size⁹. Notably, they continued to synthesize PG, as demonstrated by the incorporation of TAMRA 3-amino-D-alanine (TADA), a fluorescent D-amino acid, by the enzymes that assemble the PG matrix (Fig. 2)¹².

Fig. 1. M. xanthus cells form tissue-like structures under moenomycin stress.

(a) Wild-type vegetative cells are rod-shaped. (b, c) After moenomycin treatment (4 μg/ml, 4 h), cells transform into spheres and form honeycomb-like lattices of different sizes through visible cell-cell connections. (d) Cells overexpressing Pal lysed frequently and did not form tissue-like structures after moenomycin treatment. Scale bars, 5 μm.

Fig. 2. Cells in the tissue-like structures connect to each other through extensive OM fusion.

Cells share their OMs but not cytoplasms. The cytoplasm, PG, and OM were labeled with calcein, TADA, and WGA, respectively. BF, bright field. Arrows point to the intracellular OM connections. Scale bars, 5 μm. The cartoon inset depicts the cell sections imaged using HILO illumination.

Remarkably, these spheres organized into lattices of varying sizes on agar surfaces, displaying a hexagonal cellular arrangement with bridge-like connections linking adjacent cells (Fig. 1, 2). These lattices, reminiscent of eukaryotic tissues, exhibited sustained growth over time on agar surfaces. To determine whether individual cells within the lattices retained their integrity, we treated a liquid culture of wild-type M. xanthus with moenomycin and split it into two samples: one incubated with calcein and the other with TADA. Calcein is a small, nonfluorescent, hydrophobic molecule that readily crosses IMs. It serves as an effective probe for IM integrity, as cytoplasmic esterases convert it into a fluorescent, hydrophilic product that becomes trapped inside the cell, accumulating in the cytoplasm⁶. After 1 h of incubation, samples were washed, mixed at a 1:1 ratio and spotted on an agar surface.

Each sample incorporated its respective dye in liquid culture, and upon mixing, they quickly assembled into honeycomb-like lattices on the agar surface. We used the highly inclined and laminated optical sheet (HILO) illumination^13,14 to image half of the cell surface that is close to the coverslips. After one hour of incubation, no calcein leakage was observed, indicating that the cells remained viable and preserved IM integrity. Importantly, TADA-labeled cells showed no fluorescence from calcein, suggesting that cytoplasmic contents were not exchanged between individual cells (Fig. 2).

M. xanthus is a typical Gram-negative bacterium characterized by the presence of an OM surrounding its PG layer. To assess whether cells could establish connections via their OMs, we conducted a similar tissue-forming experiment using cells labeled with different markers. In this setup, calcein was replaced with Alexa Fluor 488-conjugated wheat germ agglutinin (WGA), which cannot penetrate the OM and therefore selectively stains OM¹⁵. In contrast to the lack of cytoplasmic exchange, TADA-labeled cells rapidly acquired fluorescence from WGA, resulting in its nearly uniform distribution in the tissue-like lattices (Fig. 2). Consistent with our recent discovery that the components in M. xanthus OM diffuse freely¹⁶, this result indicate that individual cells share OMs in the tissue-like structures. Strikingly, the “bridges” between cells were also stained by WGA (Fig. 2), indicating that the tissue-like structures are organized by stable OM channels.

In untreated swarms, M. xanthus cells can transiently fuse their OMs in a limited scale¹⁷. But how does moenomycin—an antibiotic that modulates PG synthesis¹⁸—promote the formation of a stable and extensive OM network? Given that Gram-negative bacteria typically anchor their OMs to the PG layer to coordinate cell wall elongation and division, we hypothesize that moenomycin-induced alterations to PG structure weaken these tethering points, thereby enabling extensive OM fusion. To test this hypothesis, we sought to reinforce the PG-OM connection by overexpressing Pal—a protein that binds peptidoglycan noncovalently and links it to outer membrane components of the Tol-Pal system¹⁹—using a vanillate-inducible promoter activated with 100 μM sodium vanillate. Following moenomycin treatment, cells still adopted a spherical morphology but were unable to form tissue-like structures (Fig. 1d), indicating that reinforced PG-OM connections inhibit tissue formation. Notably, widespread cell lysis was observed (Fig. 1d). Given that soil, its natural habitat, is abundant in antimicrobials produced by M. xanthus itself and other organisms like actinobacteria, forming tissue-like structures is likely a survival mechanism.

It is often assumed that tissue formation arose through gain-of-function events during evolution. However, our findings suggest a counterintuitive possibility: the emergence of primitive tissue-like structures may have resulted from loss-of-function events in bacterial cell wall assembly. Bacteria have long been implicated in key evolutionary transitions toward multicellularity, with the origin of animal tissues being particularly significant due to its relevance to human biology. For example, Salpingoeca rosetta, a unicellular eukaryote closely related to animals, forms rosette-shaped colonies in response to specific prey bacteria²⁰. Myxobacteria— the apex predatory microbes with complex behaviors—exhibit foraging strategies reminiscent of early animals² and uniquely produce plasmalogens, lipids otherwise exclusive to animals²¹. Phylogenetic evidence also suggests that ancient myxobacteria contributed to the development of the first eukaryotic cytoplasm²². These insights raise the intriguing possibility that animal tissues may have evolved from stress-induced responses in bacterial ancestors.

Materials and methods

Vegetative M. xanthus cells were grown in liquid CYE medium (10 mM MOPS pH 7.6, 1% (w/v) Bacto^™ casitone (BD Biosciences), 0.5% yeast extract and 8 mM MgSO4) at 32 °C, in 125-ml flasks with vigorous shaking, or on CYE plates that contains 1.5% agar. We used strain DZ2 as the wild-type M. xanthus strain. Pal Overexpression strain was constructed by electroporating DZ2 cells with 4 μg of plasmid DNA. The pal gene, driven by a vanillate-inducible promoter, was inserted into the Mx4 phage attachment site as a merodiploid on the M. xanthus chromosome. Transformed cells were plated on CYE plates supplemented with 10 mg/ml tetracycline and expression was induced by 100 μM sodium vanillate. For microscopy imaging, cultures were grown in liquid CYE to OD600 ~1 and supplemented with 10 μg/mL WGA Alexa Fluor-488 conjugate and 75 μM TADA for one hour. Cells were spun down at 6,000x g for 3 min and the pellet washed three times with CYE. 5 μl of cells were spotted on agar (1.5%) pads and imaged using a Andor iXon Ultra 897^™ EMCCD camera (effective pixel size 160 nm) on an inverted Nikon Eclipse-Ti^™ microscope with a 100× 1.49 NA TIRF objective. Fluorescence of TADA and Alexa Fluor 488-conjugated WGA was excited by the 561-nm and 488-nm lasers, respectively.