Prokaryotes, the single-celled organisms that include bacteria and archaea, have a cytoskeleton, and discovering that fact overturned decades of biological dogma. Until the early 1990s, textbooks taught that internal scaffolding proteins like tubulin and actin belonged exclusively to eukaryotic cells. Then three independent labs showed that a well-known bacterial cell-division protein called FtsZ was structurally related to tubulin, and the field cracked open. Since then, researchers have identified bacterial and archaeal versions of actin, intermediate filaments, and even microtubules, revealing that the cytoskeleton is far older and more widespread than anyone expected.
How the Prokaryotic Cytoskeleton Was Found
The story starts with FtsZ, a protein that had already been known for years as essential for bacterial cell division. In 1992, researchers recognized that FtsZ carries a “tubulin signature sequence” found in all forms of eukaryotic tubulin. That was the first hard evidence that prokaryotes possessed anything resembling a cytoskeletal protein.1PubMed Central. The discovery of the prokaryotic cytoskeleton: 25th anniversary Within a few more years, MreB was identified as the bacterial counterpart of actin. Together, these two discoveries established that the cytoskeleton originated in prokaryotes, not in the more complex eukaryotic cells that came later.2Nature Reviews Molecular Cell Biology. The prokaryotic cytoskeleton
What makes the story interesting is how well the structural similarities were hidden. FtsZ and tubulin share very little obvious amino-acid sequence in some regions, yet their three-dimensional protein folds are strikingly alike. When crystal structures of both proteins were solved, the shapes of even the parts with no detectable sequence match turned out to be virtually identical.3PubMed. Atomic structures of tubulin and FtsZ The same pattern holds for MreB and actin. In each case, the three-dimensional architecture is conserved even when the underlying code has drifted far apart over billions of years of evolution.
FtsZ and the Ring That Splits a Cell in Two
FtsZ is present in nearly all bacteria and most archaea, and its primary job is orchestrating cell division. When a bacterium is ready to divide, FtsZ molecules assemble into a ring-shaped polymer at the future division site. This structure, called the Z ring, acts as a scaffold that recruits dozens of other proteins into a larger complex known as the divisome.4PubMed Central. At the Heart of Bacterial Cytokinesis: The Z Ring The Z ring also contains FtsA and ZipA, two anchor proteins that tether FtsZ filaments to the inner membrane.5PubMed Central. Condensation of FtsZ filaments can drive bacterial cell division
One of the more elegant discoveries in recent years is that FtsZ filaments treadmill around the division ring. Treadmilling means subunits are added at one end and removed at the other, so the filament appears to glide in a circle without the whole structure physically moving. This circumferential motion turns out to be functionally important: the treadmilling filaments drag along the enzymes that build new cell-wall material. As the enzymes follow the filaments around, they lay down progressively smaller concentric hoops of peptidoglycan, which gradually pinch the cell in half.6PubMed Central. Treadmilling by FtsZ filaments drives peptidoglycan synthesis and bacterial cell division Changing the speed of FtsZ treadmilling directly changes how fast new cell wall is built and how quickly the cell divides.
There is a subtlety worth noting here. FtsZ on its own generates only a few piconewtons of force when its curved filaments bend against a membrane. That is enough to pinch a membrane vesicle in the lab, but in a living cell the internal pressure (turgor) is much higher.7PubMed Central. Turgor Pressure and Possible Constriction Mechanisms in Bacterial Division The real constriction force comes from the coordinated synthesis of new peptidoglycan. FtsZ organizes the process; the cell wall machinery does the heavy mechanical lifting.
MreB and the Shape of a Bacterium
If FtsZ governs when and where a cell divides, MreB governs what shape it takes while growing. MreB is the bacterial actin homolog, and it plays a central role in maintaining cell shape, directing the placement of new cell wall material, and even helping coordinate chromosome segregation.8PubMed Central. MreB: unraveling the molecular mechanisms of bacterial shape, division, and environmental adaptation Bacteria that naturally take a rod shape, like Escherichia coli or Bacillus subtilis, rely on MreB. Remove it, and those cells swell into spheres and eventually die.
Early imaging studies suggested MreB formed long helical cables running the length of the cell. Better microscopy later revised that picture. MreB actually assembles into short, discrete patches that travel around the circumference of the cell, perpendicular to the long axis.9PubMed. Processive movement of MreB-associated cell wall biosynthetic complexes in bacteria Each patch carries along a set of cell-wall-building enzymes, and as the patches move they insert radial hoops of new peptidoglycan that elongate the cell while preserving its rod shape.10PubMed Central. Coupled, circumferential motions of the cell wall synthesis machinery and MreB filaments in B. subtilis
Here is a genuinely surprising twist: the motion of MreB patches is not driven by MreB’s own polymerization. When researchers blocked cell-wall synthesis with antibiotics, the patches stopped moving, even though MreB itself was still capable of forming polymers.11PubMed Central. The bacterial actin MreB rotates, and rotation depends on cell-wall assembly The cell-wall machinery outside the membrane seems to act as a kind of motor that pushes or pulls MreB along. This is the reverse of what anyone initially assumed: the cytoskeleton organizes the wall builders, but the wall builders provide the force to move the cytoskeleton. It is a deeply coupled system.
When FtsZ and MreB Work Together
Cell division and cell elongation are not independent processes, and neither are the cytoskeletal proteins that drive them. In E. coli, FtsZ and MreB interact directly, and that physical contact turns out to be essential. When the interaction is disrupted, the Z ring cannot contract properly, and the cell fails to build the dividing wall (the septum). What appears to happen is that MreB hands off cell-wall-building enzymes from the elongation system to the division system as the cell transitions from growing longer to splitting in two.12PubMed Central. Direct interaction of FtsZ and MreB is required for septum synthesis and cell division in Escherichia coli This means bacterial cell division is coupled to cell elongation through a direct handshake between the two main cytoskeletal proteins. Even in multicellular cyanobacteria, MreB forms filamentous structures that interact with FtsZ, suggesting this partnership is widespread.13PubMed Central. The role of the cytoskeletal proteins MreB and FtsZ in multicellular cyanobacteria
Crescentin and Intermediate Filament Relatives
Eukaryotic cells have a third major cytoskeletal component beyond tubulin and actin: intermediate filaments, the rope-like proteins that provide tensile strength to tissues like skin and muscle. Bacteria have their own version. In 2003, researchers studying Caulobacter crescentus, a crescent-shaped freshwater bacterium, discovered a protein they named crescentin. It has the hallmark features of intermediate filament proteins, including the ability to self-assemble into filaments without any energy input or helper molecules.14PubMed. The bacterial cytoskeleton: an intermediate filament-like function in cell shape
Crescentin lines one side of the cell interior, creating an asymmetry that bends the cell into its characteristic curved shape. Delete the gene for crescentin, and the cells become perfectly straight rods. Put crescentin into E. coli, a normally straight bacterium, and those cells start curving too.15PubMed Central. Bacterial cell curvature through mechanical control of cell growth The mechanism works by causing one side of the cell to grow more slowly than the other, bending the whole structure like a bimetallic strip in a thermostat.
Crescentin is not the only non-actin, non-tubulin cytoskeletal protein in bacteria. Bactofilins, discovered more recently, form a distinct class of filament-forming proteins that are widespread across bacterial species. Structural studies show they fold into a triangular solenoid shape quite unlike anything in eukaryotes, suggesting they are a genuinely prokaryotic invention rather than a shared ancestor with eukaryotic proteins.16PubMed Central. Structure of the Bacterial Cytoskeleton Protein Bactofilin by NMR Chemical Shifts and Sequence Variation
Pushing Plasmids Apart
One of the more visually striking prokaryotic cytoskeletal systems has nothing to do with cell shape or division. It is a miniature spindle apparatus encoded by plasmids to ensure their own inheritance. The best-studied version is the ParMRC system from the R1 plasmid in E. coli. It consists of just three parts: ParM (an actin-like protein), ParR (an adaptor protein that binds DNA), and parC (a short DNA sequence that acts like a centromere).17PubMed. The ParMRC system: molecular mechanisms of plasmid segregation by actin-like filaments
When the plasmid replicates, ParM filaments grow between the two copies, forming a bipolar spindle of at least two antiparallel filaments that physically push the plasmid copies toward opposite ends of the cell.18PubMed Central. A bipolar spindle of antiparallel ParM filaments drives bacterial plasmid segregation The resemblance to a eukaryotic mitotic spindle is hard to miss, yet the system arose independently in a completely different cellular context. ParM filaments are dynamic, searching through the cell space for plasmid copies to attach to. Once both ends of a filament are stabilized by ParR-parC complexes, the filament elongates and drives the two plasmids apart.19Current Biology. Chromosome Segregation: Pushing Plasmids Apart It is an astonishingly simple machine: three components accomplish reliable DNA segregation with no motor proteins, no kinetochore complexes, and no nuclear envelope.
Chromosomal DNA segregation in bacteria uses a different system. The ParABS system, found on many bacterial chromosomes, relies on a gradient of the ATPase ParA across the cell to move newly replicated chromosome origins toward opposite poles. This mechanism is less obviously cytoskeletal than ParM’s actin-like filaments, but it achieves a similar outcome through a different biophysical strategy.
Magnetosome Chains and Organelle Organization
Some bacteria build membrane-bound organelles called magnetosomes, tiny crystals of iron minerals enclosed in lipid vesicles that function like compass needles, allowing the cells to navigate along Earth’s magnetic field lines. For this navigation to work, the magnetosomes need to be lined up in a chain. That alignment is maintained by MamK, an actin-like cytoskeletal protein specific to magnetotactic bacteria.20PubMed Central. Interplay between two bacterial actin homologs, MamK and MamK-Like, is required for the alignment of magnetosome organelles in Magnetospirillum magneticum AMB-1 MamK polymerizes into filaments that run alongside the magnetosome chain and hold the vesicles in register. The system requires cooperation with a second actin homolog called MamK-Like, showing that even relatively simple organisms can deploy multiple cytoskeletal proteins for specialized tasks.
The Archaeal Cytoskeleton
Archaea, the other major domain of prokaryotic life, have their own cytoskeletal toolkit, and in some respects it bridges the gap between bacteria and eukaryotes more visibly than anything in the bacterial world. Many archaea use FtsZ for cell division, just as bacteria do. But a subset of archaea, particularly the Crenarchaeota, have discarded FtsZ entirely and instead divide using a system based on ESCRT-III-like proteins, the same protein family that eukaryotic cells use for membrane remodeling and the final step of cell division.21PubMed Central. Molecular structure of the ESCRT-III-based archaeal CdvAB cell division machinery Physical modeling and live-cell imaging suggest these archaea use a division mechanism that is distinct from both the FtsZ-based constriction of most bacteria and the actomyosin ring of animal cells.22PubMed Central. Physical mechanisms of ESCRT-III-driven cell division
On the actin side, crenarchaea possess crenactin, an actin homolog that is functional at temperatures up to 90°C. Despite very low sequence similarity to eukaryotic actin, the three-dimensional structure is remarkably close. Crenactin even shares some structural features with eukaryotic actin that are absent from bacterial MreB, hinting that the eukaryotic actin lineage may trace back through an archaeal ancestor rather than directly from bacteria.23PubMed. Structure of crenactin, an archaeal actin homologue active at 90°C
Microtubules Before Eukaryotes Existed
Perhaps the most striking recent finding in the field came in 2025, when researchers reported actual microtubules in Asgard archaea, the group of organisms widely considered the closest living relatives of eukaryotes. These Asgard tubulins form heterodimers resembling eukaryotic alpha/beta-tubulin pairs and assemble into bona fide microtubules with five protofilaments, compared to the thirteen protofilaments in most eukaryotic microtubules. A variant form, incorporating a paralog that cannot bind a nucleotide, produces seven-protofilament microtubules with non-standard properties. Cryo-tomography of the archaeon Candidatus Lokiarchaeum ossiferum revealed tubular structures inside cells, and expansion microscopy confirmed these were tubulin-based assemblies.24PubMed. Microtubules in Asgard archaea
This discovery suggests that microtubules are not a eukaryotic invention at all. They appear to have a pre-eukaryotic origin in the archaeal lineage that eventually gave rise to all complex life. The finding reshapes how biologists think about the transition from simple to complex cells: the machinery for building internal scaffolding may already have been in place before the first true eukaryote appeared.
Viruses That Build Their Own Spindles
Cytoskeletal proteins are not limited to cellular organisms. Certain large bacteriophages, viruses that infect bacteria, encode their own tubulin-like protein called PhuZ. When one of these phages infects a host cell, PhuZ assembles into a dynamic spindle-like array that positions the viral DNA at the center of the cell. Correct centering is required for optimal virus reproduction: if PhuZ filaments lose their dynamic behavior, the DNA drifts off-center and the phage produces fewer progeny.25PubMed Central. A Phage Tubulin Assembles Dynamic Filaments by a Novel Mechanism to Center Viral DNA within the Host Cell
The phenomenon is conserved across several large Pseudomonas phages, including φKZ and φPA3. These phages go even further: they build a proteinaceous shell around their replicating DNA that resembles a nucleus, and PhuZ forms a bipolar spindle that displays dynamic instability, growing and shrinking in a way reminiscent of eukaryotic microtubules, to keep that shell centered in the host cell.26PubMed Central. The phage nucleus and tubulin spindle are conserved among large Pseudomonas phages A virus constructing a nucleus-like compartment and a spindle apparatus inside a bacterial cell is one of the stranger outcomes of evolution’s tinkering with cytoskeletal parts.
Why Bacterial Cytoskeletons Matter for Antibiotics
Because FtsZ is found in virtually all bacteria but has no counterpart in human cells, it is an attractive target for new antibiotics. The logic is straightforward: a drug that jams FtsZ would prevent bacteria from dividing without affecting human cellular processes. The protein’s high conservation across bacterial species also means a single FtsZ-targeting drug could theoretically work against a wide range of pathogens.27PubMed Central. FtsZ as a novel target for antibiotics development: Promises and challenges
In practice, developing FtsZ inhibitors has been challenging. The protein sits inside the bacterial cell, so any drug has to cross the outer membrane, and in Gram-negative bacteria that means penetrating two membranes. Several compounds have shown potent activity in lab assays but poor performance in animal models because they cannot reach their target efficiently. MreB has also attracted interest as a drug target, given its essential role in maintaining cell shape and viability in rod-shaped pathogens.8PubMed Central. MreB: unraveling the molecular mechanisms of bacterial shape, division, and environmental adaptation The structural differences between MreB and human actin, while real, are narrower than those between FtsZ and tubulin, which complicates selectivity. Still, as antibiotic resistance grows, the cytoskeleton remains one of the more promising frontiers for new drug development.
Seeing the Invisible Scaffold
Much of what we know about the prokaryotic cytoskeleton has been shaped by advances in imaging technology. Standard light microscopy cannot resolve structures at the scale of individual cytoskeletal filaments inside a bacterium, which is only about a micrometer wide. Super-resolution fluorescence microscopy transformed the field by allowing researchers to track single molecules of FtsZ and MreB in living cells, revealing behaviors like treadmilling and circumferential patch movement that were invisible with older methods. Cryo-electron tomography has pushed even further, producing three-dimensional views of cytoskeletal filaments in their native cellular context without the distortions introduced by chemical fixation or staining.28PubMed Central. Three-dimensional organization of the cytoskeleton: A cryo-electron tomography perspective The Asgard archaeal microtubule discovery, for example, relied on cryo-tomography to catch tubular structures inside cells that are notoriously difficult to grow in the lab.24PubMed. Microtubules in Asgard archaea As these imaging tools continue to improve in resolution and throughput, the list of known prokaryotic cytoskeletal elements will almost certainly keep growing.