How Did Eukaryotes Evolve From Prokaryotes?

Eukaryotic cells arose from a merger between at least two prokaryotic partners: an archaeal host cell and a bacterium that became the mitochondrion. This was not a single dramatic moment but a drawn-out evolutionary transition spanning hundreds of millions of years, during which the host and its internal passenger co-evolved a suite of features we now consider hallmarks of complex life, from the nucleus and internal membranes to sexual reproduction. The details of how, when, and in what order these features appeared remain some of the most actively debated questions in biology.

The Host Was an Archaeon, Not a Bacterium

For decades, textbooks presented three domains of life: Bacteria, Archaea, and Eukarya, implying that eukaryotes branched off as a separate lineage very early. Newer genomic analyses tell a different story. When researchers account carefully for how well their evolutionary models fit the data, eukaryotes consistently appear to have originated from within the Archaea rather than alongside them. This means the tree of life has two main trunks, not three: Bacteria on one side, and Archaea (with eukaryotes nested inside) on the other.¹

The closest archaeal relatives of eukaryotes appear to be a group called the Asgard archaea, with a subgroup called the Heimdallarchaeota currently the best candidate for the lineage most closely related to the eukaryotic nuclear ancestor.¹ Earlier analyses had pointed to a broader grouping of archaea known as the TACK superphylum as the eukaryotic sister group.² The Asgard archaea, discovered only in the last decade, sit within or near TACK and have been the breakthrough that sharpened the picture. Their genomes carry genes for proteins previously considered uniquely eukaryotic, including components involved in membrane trafficking and cytoskeletal functions. This suggests that the archaeal ancestor of eukaryotes was already more complex than a typical prokaryote.

The Mitochondrial Partnership

The other essential partner in eukaryotic origins was a bacterium from the alphaproteobacteria group that eventually became the mitochondrion. How that bacterium ended up inside the archaeal host is the central puzzle. The most straightforward scenario, sometimes called the hydrogen hypothesis, proposes that the relationship began as a metabolic partnership: the bacterium produced something the archaeon needed (hydrogen or small organic molecules), and physical proximity gradually became internalization.³ Over time, the bacterium lost the ability to live independently, transferring many of its genes to the host’s genome and becoming an organelle.

Identifying which modern alphaproteobacterium is the closest living relative of mitochondria has proved harder than expected. For years, the Rickettsiales (a group of obligate intracellular parasites) were the leading candidates, largely because they also live inside other cells. But analyses of the mitochondrial ATP-generating machinery suggest a more complicated picture. Comparisons of key subunits from the mitochondrial energy complex show a closer relationship to other alphaproteobacterial orders, including Rhodospirillales and Rhodobacterales, rather than to Rickettsiales.⁴ The hunt for mitochondria’s closest cousin is still open, and the answer likely varies depending on which genes you examine, because the evolutionary history of the mitochondrial ancestor involved a great deal of gene swapping with other bacteria.

Did Complexity Come Before or After Mitochondria?

This is one of the sharpest divides in the field. All living eukaryotes (or their ancestors) had mitochondria, and the last common ancestor of all eukaryotes was already a complex cell. The question is whether the host cell had to become complex first in order to engulf the future mitochondrion, or whether acquiring the mitochondrion is what triggered the evolution of cellular complexity.

The “complexity first” camp argues that the host evolved a primitive ability to engulf other cells (phagocytosis) before mitochondria arrived. In this view, the mitochondrion was essentially a prey item that escaped digestion and stuck around. The “mitochondria first” camp counters that the host was still fundamentally a simple archaeon at the time. It could not have been a phagocyte because phagocytosis requires an elaborate internal membrane system and cytoskeleton that are themselves products of the merger.⁵ Under this scenario, the massive energy budget provided by mitochondrial membranes is what allowed the host to evolve the complex internal architecture of a modern eukaryotic cell.

One compromise scenario imagines an archaeal ancestor that had no rigid cell wall (similar to the modern archaeon Thermoplasma) but did have a simple actin-based cytoskeleton allowing it to form membrane protrusions. These protrusions could have occasionally, accidentally enclosed bacteria, one of which became the mitochondrion. The endosymbiont then triggered the development of a full endomembrane system and, eventually, true phagocytosis.⁶ This picture fits with what we know about Asgard archaea possessing genes for rudimentary cytoskeletal and membrane-shaping proteins.

Where Did the Nucleus Come From?

The nucleus, the membrane-enclosed compartment housing the cell’s DNA, is the defining feature of eukaryotes. You might assume its origin is well understood given how fundamental it is, but there are at least three competing families of ideas, and none has won decisively.

One proposal flips the standard assumption about what happened during the merger. Instead of imagining a bacterium being swallowed by a host cell and the nucleus forming inside the cytoplasm, the “inside-out” model proposes that the ancestral archaeal cell (equivalent to the modern nucleus) extruded membrane blebs beyond its own cell wall, gradually building up an outer compartment that became the cytoplasm.⁷ In this model, the nuclear envelope is not something that was built around the DNA later; the nucleus is the original cell, and everything else grew outward from it.

A second idea focuses on a functional pressure. When the proto-mitochondrion transferred many of its genes into the host genome, it also transferred self-splicing DNA elements called introns. These introns had to be cut out of the messenger RNA before that RNA could be translated into protein. The problem is that splicing is slow and translation is fast. Without a physical barrier, the cell’s protein-making machinery would start reading RNA that still contained junk sequences, producing broken proteins. In this view, the nuclear envelope evolved as a quality-control barrier, keeping the splicing compartment separate from the translation compartment so that only fully processed RNA would reach the protein factories.⁸

A third and more provocative hypothesis proposes that the nucleus descended from a giant virus. Large DNA viruses today build “viral factories” inside host cells, walled-off compartments where the virus replicates its genome and assembles new particles. The viral eukaryogenesis hypothesis suggests that an ancient giant virus infected the archaeal ancestor of eukaryotes and its viral factory eventually became the nucleus.⁹ Some support comes from the observation that the molecular capping machinery eukaryotes use to mark their RNA for export from the nucleus is found in giant viruses of the Mimiviridae but is absent from the closest archaeal relatives of eukaryotes, like Lokiarchaeota.¹⁰ This remains a minority view, but it has not been ruled out.

Building the Internal Machinery

Eukaryotic cells are packed with internal structures that prokaryotes lack or possess only in rudimentary forms: a dynamic cytoskeleton of protein filaments, an endomembrane system of interconnected compartments, and a sophisticated protein-trafficking network. Understanding where all this came from requires looking at both partners in the merger.

The cytoskeleton is a good example. Before 1992, tubulin and actin, the workhorses of the eukaryotic cytoskeleton, were thought to be unique to eukaryotes. Then researchers discovered that bacteria had their own versions. A protein called FtsZ, already known to play a role in bacterial cell division, turned out to carry a signature sequence shared by all forms of tubulin. Around the same time, bacterial relatives of actin were identified. Crystal structures have since confirmed these relationships.¹¹ Prokaryotic filament systems turn out to be structurally and dynamically complex, prefiguring the self-organizing properties of the eukaryotic cytoskeleton.¹² The eukaryotic versions appear to be elaborations of bacterial precursors that had already evolved a range of functions.

The membrane-trafficking system, the network of vesicles and compartments that shuttles proteins around eukaryotic cells, has a partially archaeal pedigree. Asgard archaea like Lokiarchaeota carry genes for small signaling proteins that in eukaryotes regulate vesicle traffic. Phylogenetic analyses show that the eukaryotic versions of these proteins branch from within the Lokiarchaeal versions, consistent with an archaeal origin followed by gene duplication and diversification in the eukaryotic lineage.¹³ Early versions of protein-sorting complexes are also detectable in prokaryotes, providing insight into the earliest stages of how internal membrane compartments formed.¹⁴

The Membrane Lipid Problem

One of the strangest aspects of the prokaryote-to-eukaryote transition is the membrane switch. Archaea build their cell membranes from a fundamentally different chemistry than bacteria and eukaryotes do. Archaeal membranes use isoprenoid chains linked to a mirror-image version of the glycerol backbone, while bacterial and eukaryotic membranes use fatty acid chains on the opposite-handed glycerol. This is the “lipid divide,” and it has long puzzled researchers because eukaryotes have an archaeal host lineage yet bacterial-style membranes.

A leading hypothesis suggests that the last universal common ancestor of all life had a “mixed membrane” containing both types of lipid, and the two lineages later specialized. This scenario gained real-world support when researchers showed that modern bacteria in the FCB group superphylum carry a functional pathway for making archaeal-type ether-linked lipids alongside their normal bacterial fatty acid membranes. When the archaeal pathway genes were expressed in lab-grown E. coli, the bacteria produced a genuine mixed membrane.¹⁵ This demonstrates that mixed membranes are chemically stable and can persist over long evolutionary timescales, making the transition from an archaeal-style membrane to a bacterial-style one during eukaryogenesis more plausible than it once seemed.

Rough Timeline

Pinning dates on these events is notoriously difficult. Molecular clock analyses suggest that the archaeal lineage leading to the eukaryotic nucleus diverged from other Asgard archaea roughly 2.7 to 2.2 billion years ago, and the alphaproteobacterial lineage that became mitochondria split from its free-living relatives around 2.6 to 2.1 billion years ago.¹⁶ The last eukaryotic common ancestor, the population from which all modern eukaryotes descend, is estimated to have lived somewhere between about 2.4 and 1.8 billion years ago, depending on which calibrations and models are used.¹⁶

The fossil and geochemical record broadly supports this timing but adds some wrinkles. Chemical fossils called protosteroids, early intermediates in the sterol-making pathway that eukaryotes use today, have been found in rocks dating from at least 1.64 billion years ago to around 800 million years ago. These compounds point to a “protosterol biota” of ancient organisms, likely including deep-branching stem-group eukaryotes, that were widespread in aquatic environments during the mid-Proterozoic.¹⁷ The gap between the molecular clock dates for LECA and the first clear fossil evidence of recognizable eukaryotic groups (algae, for example) is large enough to suggest there was a long “hidden” period of early eukaryotic evolution.

Chloroplasts Came Later

The story of mitochondria is about the origin of all eukaryotes, but the origin of photosynthetic eukaryotes required a second, separate endosymbiosis. A eukaryote that already had mitochondria engulfed a cyanobacterium, and that cyanobacterium eventually became the chloroplast. Genomic comparisons between modern cyanobacteria and the plant Arabidopsis thaliana leave no doubt about this relationship.¹⁸

This primary plastid endosymbiosis appears to have happened only once, in the common ancestor of the supergroup Archaeplastida, which today includes green algae, land plants, red algae, and glaucophyte algae.¹⁹ After that, some lineages acquired photosynthesis secondhand by swallowing algae that already had chloroplasts, a process called secondary endosymbiosis, which accounts for the photosynthetic ability of organisms like brown algae and some single-celled predators. Molecular clocks estimate the primary plastid diverged from free-living cyanobacteria roughly 2.1 to 1.7 billion years ago.¹⁶

A Living Model of Endosymbiosis in Action

One of the most compelling pieces of evidence that endosymbiosis can and does happen comes from the amoeba Paulinella. This single-celled organism independently acquired its own photosynthetic organelle, termed a chromatophore, from a cyanobacterium. Unlike the ancient origin of the standard chloroplast over 1.5 billion years ago, the Paulinella endosymbiosis happened relatively recently, around 90 to 140 million years ago.²⁰ Having two independent examples of plastid formation on vastly different timescales gives researchers a way to study the general principles of how a free-living bacterium becomes a permanent organelle, including gene transfer to the host nucleus and the loss of the endosymbiont’s genetic independence.

Not Everything Inside a Eukaryote Came From Endosymbiosis

The success of endosymbiotic theory for mitochondria and chloroplasts naturally led researchers to ask whether other eukaryotic organelles had similar origins. Peroxisomes, compartments involved in breaking down fatty acids and detoxifying reactive molecules, were an obvious candidate. They divide by fission like mitochondria and chloroplasts, and they import their proteins after those proteins are made, which is also reminiscent of organelles with endosymbiotic origins.

But the case fell apart under genomic scrutiny. Unlike mitochondria and chloroplasts, whose membranes contain transporter proteins descended from their bacterial ancestors, the peroxisomal membrane is built from proteins that are purely eukaryotic in origin, with no bacterial or cyanobacterial membrane components detectable.²¹ The most conserved proteins involved in building and maintaining peroxisomes turn out to be related to proteins from the endoplasmic reticulum, the cell’s internal membrane factory. Cells that have been experimentally stripped of peroxisomes can regenerate them from the endoplasmic reticulum when the right gene is reintroduced.²² Peroxisomes, it seems, are a homegrown eukaryotic invention, not a captured outsider. The lesson here is that endosymbiosis explains a lot of eukaryotic evolution but not all of it. Some compartments were built from scratch using the cell’s own expanding repertoire of membranes and proteins.

Sex as a Product of the Merger

Sexual reproduction, meaning the shuffling and recombining of genetic material through meiosis, is found across virtually all eukaryotic lineages. Its near-universality strongly suggests that the last eukaryotic common ancestor was already capable of meiotic sex.²³ Why would a brand-new type of cell evolve such a costly and elaborate process so early?

One hypothesis ties sex directly back to the mitochondrial endosymbiosis. The newly acquired proto-mitochondrion generated reactive oxygen species as a byproduct of its energy metabolism. These reactive molecules damaged the host’s DNA at a higher rate than the host had experienced before. Meiotic sex, with its built-in DNA repair step during the pairing and crossing over of chromosomes, could have evolved as a response to this elevated mutation pressure. The raw components were available: archaea already had DNA repair and recombination enzymes that could be repurposed. In this view, sex was one of the earliest consequences of the endosymbiotic merger, not a late addition.²⁴

This connection between mitochondria and sex is speculative but internally consistent: the same event that gave eukaryotes their energy advantage also created a problem (DNA damage) and simultaneously provided the selective pressure to solve it (recombinational repair through meiosis). It is a tidy illustration of how the different threads of eukaryogenesis may be more interconnected than they first appear.