Scientists are broadly united on the answer: mitochondria descended from free-living bacteria that took up residence inside another cell roughly two billion years ago. This idea, called endosymbiotic theory, holds that an ancient bacterium from the alphaproteobacteria group entered into a partnership with an archaeal host cell, and over vast stretches of time the bacterium lost its independence, becoming the energy-producing organelle found in nearly every complex cell on Earth today. The story of how that partnership formed, who the partners were, and what it meant for the evolution of all complex life is richer and stranger than the textbook summary suggests.
How the Idea Developed
The notion that parts of our cells might once have been independent organisms is not new. In 1905, the Russian biologist Constantin Mereschkowsky proposed that chloroplasts, the photosynthetic structures in plant cells, descended from cyanobacteria-like organisms that had been swallowed by a host cell. By 1927, Ivan Wallin had extended the same logic to mitochondria, arguing they evolved from once free-living bacteria.1PubMed. Endosymbiosis, cell evolution, and speciation These proposals were largely ignored or ridiculed for decades. The intellectual climate of early twentieth-century biology wasn’t ready for the idea that major cellular components had foreign origins.
That changed in 1967, when Lynn Margulis (publishing under her then name Lynn Sagan) wrote “On the Origin of Mitosing Cells” in the Journal of Theoretical Biology. Margulis didn’t just revive the old idea; she built a far more detailed case, arguing that both mitochondria and plastids descended from bacterial ancestors that had been engulfed by a host cell. She even proposed that the eukaryotic flagellum originated from an endosymbiotic spirochete-like organism, a claim that has not held up nearly as well as her arguments about mitochondria and chloroplasts.2PubMed Central. Lynn Margulis and the endosymbiont hypothesis: 50 years later Margulis’s paper faced enormous resistance at first, but the accumulation of molecular evidence over the following decades proved her core thesis right. Today, the bacterial origin of mitochondria is one of the most firmly established facts in cell biology.
Identifying the Bacterial Ancestor
Saying mitochondria came from “bacteria” is true but vague. Scientists have spent decades trying to pin down exactly which bacterial lineage gave rise to them. The consensus points squarely at the alphaproteobacteria, a large and diverse class of bacteria that today includes organisms living in soil, water, and inside other cells. When researchers compare the genes still carried by mitochondrial DNA to those of living bacteria, the best matches consistently fall within this group.3PubMed. A genome phylogeny for mitochondria among alpha-proteobacteria and a predominantly eubacterial ancestry of yeast nuclear genes
Narrowing the identification further has proven frustrating. The event happened so long ago, and so much genetic change has occurred since, that the phylogenetic signal is faint. Early analyses placed mitochondria near the Rickettsiales, an order of alphaproteobacteria that includes obligate intracellular parasites. More recent work using metagenome-assembled genomes collected from marine environments has repositioned the mitochondrial ancestor closer to the root of most alphaproteobacteria, rather than nested within any one modern order.4PubMed Central. Origin and evolution of mitochondrial inner membrane composition One 2022 analysis using multiple independent methods suggested that the closest living bacterial relatives of the proto-mitochondrion may be found in the Iodidimonadales, an order sitting in the middle of alphaproteobacterial diversity.5bioRxiv. On the bacterial ancestry of mitochondria: new insights with triangulated approaches The honest summary is that we know the mitochondrial ancestor was an alphaproteobacterium, but exactly where it sat on that family tree remains actively debated.
Who Was the Host Cell?
If the bacterium that became the mitochondrion is one partner, the other partner is the host cell that took it in. For years, researchers assumed the host must have already been a fairly complex cell, perhaps something resembling a simple modern eukaryote. The discovery of Asgard archaea, starting around 2015, upended that picture.
Asgard archaea are a group of microorganisms whose genomes contain a surprising number of genes previously thought to be exclusive to eukaryotes, the complex cells that make up animals, plants, and fungi. Phylogenomic analyses consistently place eukaryotes within the Asgard group, making these archaea the closest known relatives of the host cell that acquired the mitochondrial ancestor.6PubMed. Asgard archaea illuminate the origin of eukaryotic cellular complexity The implication is striking: the host cell wasn’t already a complex eukaryote. It was an archaeon, a prokaryotic organism without a nucleus or the elaborate internal membrane systems we see in eukaryotic cells today.
A major 2025 study reinforced this view, finding that Asgard archaea contributed the dominant share of genes to most conserved eukaryotic functional systems. The authors proposed a model in which key features of eukaryotic cell organization evolved within the Asgard lineage before the mitochondrial endosymbiosis occurred, followed by the capture of the alphaproteobacterial partner and the gradual acquisition of additional bacterial genes over time.7PubMed Central. Dominant contribution of Asgard archaea to eukaryogenesis The last eukaryotic common ancestor, the cell from which all modern eukaryotes descend, already possessed mitochondria. Whatever evolutionary path led to the first eukaryote, mitochondria were there from the start.
Why Did They Merge?
Knowing who the partners were doesn’t explain why an archaeon and a bacterium would enter into such an intimate relationship. Several competing hypotheses try to answer this, and the field has not settled on a single winner.
The hydrogen hypothesis, proposed by William Martin and Miklós Müller in 1998, remains one of the most influential. It suggests the archaeal host was a strict autotroph that depended on molecular hydrogen for its metabolism. The bacterial partner, which could both respire oxygen and produce hydrogen as a waste product of anaerobic metabolism, provided a steady supply of the gas the host needed. Over time, the host became so dependent on this hydrogen source that the two organisms fused into one.8PubMed. The hydrogen hypothesis for the first eukaryote This model is appealing because it provides a concrete selective pressure, a metabolic dependency, rather than relying on one cell randomly swallowing another and getting lucky.
Other models have been proposed, and recent reviews describe a shift toward symbiogenetic frameworks broadly, meaning models built on metabolic give-and-take between archaea and bacteria. These include variations on the hydrogen hypothesis, the “Reverse Flow” model, and the “Entangle-Engulf-Endogenize” model, each proposing different metabolic currencies and physical mechanisms for how the two partners became inseparable.9PubMed. The Syntrophy hypothesis for the origin of eukaryotes revisited What all of these models share is a rejection of the older “accidental meal” story, where a predatory cell simply ate a bacterium and failed to digest it. The metabolic partnership view has largely replaced it.
Did Complexity Come Before or After Mitochondria?
One of the most persistent debates in this field asks a chicken-and-egg question: did the host cell already have the machinery to engulf other cells (phagocytosis) before it acquired mitochondria, or did mitochondria come first and enable the evolution of that complexity?
In what are called autogenous or “complexity first” models, an ancestor gradually evolved eukaryotic features like an internal membrane system and the ability to engulf particles. Eventually it swallowed a bacterium that became the mitochondrion. In “mitochondria first” models, the host was a relatively simple archaeon. The acquisition of the bacterial endosymbiont provided the energy and the raw genetic material needed to evolve all the complexity that followed, including internal membranes potentially derived from the endosymbiont’s own membrane vesicles.10PubMed Central. The Physiology of Phagocytosis in the Context of Mitochondrial Origin Both camps can point to supporting evidence, and some researchers have proposed intermediate models that don’t require phagocytosis at all, suggesting instead that the two organisms could have grown together in a shared extracellular matrix before one physically enclosed the other.11PubMed Central. Division of labour in a matrix, rather than phagocytosis or endosymbiosis, as a route for the origin of eukaryotic cells
This debate matters because it shapes our understanding of what the host cell looked like. If complexity came first, the host was already eukaryote-like. If mitochondria came first, the host was a prokaryote, and virtually everything we associate with complex cells evolved afterward. The Asgard archaea findings nudge the story in an interesting direction: the host had some eukaryotic-like genes but was structurally still a prokaryote. Reality may not fit neatly into either camp.
The Energy Windfall
Whatever the initial reason the partnership formed, one of its most consequential outcomes was energetic. A landmark 2010 paper argued that the endosymbiosis that gave rise to mitochondria allowed a roughly 200,000-fold expansion in the number of genes a cell could support. By moving energy production onto internal membranes housed in many copies of the organelle, the host cell broke free of the energetic constraints that keep prokaryotic genomes small.12PubMed. The energetics of genome complexity This enormous leap in genomic capacity, the argument goes, was the prerequisite for everything from multicellularity to brains.
More recent modeling has complicated this picture somewhat. A 2022 analysis found that the energetic advantage of having mitochondria is less dramatic than originally claimed for smaller genomes. At genome sizes comparable to a typical bacterium, the advantage is less than about ten percent. The advantage grows substantially only at larger genome sizes, where eukaryotes with mitochondria can maintain vastly more DNA than any prokaryote could.13PubMed Central. The role of mitochondrial energetics in the origin and diversification of eukaryotes In practical terms, this suggests mitochondria didn’t instantly supercharge their hosts. The energy boost became transformative only as genomes expanded, creating a kind of feedback loop: mitochondria enabled larger genomes, and larger genomes enabled more complex organisms that benefited even more from mitochondrial energy.
How Mitochondria Lost Their Independence
If mitochondria were once free-living bacteria, they have surrendered nearly all of their autonomy. A free-living alphaproteobacterium carries thousands of genes. Human mitochondrial DNA contains just 37. Most of the genes the original endosymbiont carried have been transferred to the host cell’s nucleus over evolutionary time, a process called endosymbiotic gene transfer. The proteins encoded by those nuclear genes are manufactured in the cell’s cytoplasm and then shipped back into the mitochondrion, which can no longer make them on its own.
Mathematical modeling suggests that the rate of this gene transfer depends on several factors, including how fiercely mitochondria within the same cell compete with one another, the probability of inheriting organelles from both parents rather than just one, and the overall population size of the organism. Intense competition among mitochondria within a cell tends to slow gene transfer, because a mitochondrion that loses an essential gene to the nucleus becomes dependent on that nuclear copy, putting it at a disadvantage against fellow mitochondria that still carry their own.14PubMed Central. Rate of gene transfer from mitochondria to nucleus: effects of cytoplasmic inheritance system and intensity of intracellular competition
To receive all those proteins back, mitochondria had to evolve an elaborate import system. The protein transport machinery that moves nuclear-encoded proteins into mitochondria was already mostly in place in the last eukaryotic common ancestor, meaning it evolved early in the endosymbiotic relationship. Some components of this import system trace their ancestry back to bacterial outer membrane proteins, while others are novel eukaryotic inventions.15PubMed Central. Origin and Evolutionary Alteration of the Mitochondrial Import System in Eukaryotic Lineages The result is that mitochondria are locked into a permanent partnership. They retain just enough genetic independence to carry a small set of their own genes, but they depend entirely on the host cell for the vast majority of the proteins they need to function.
Fingerprints of the Bacterial Past
Even after billions of years of co-evolution, mitochondria carry unmistakable signs of their bacterial origins. They have their own ribosomes, the molecular machines that build proteins, and these ribosomes are distinctly bacterial in character. An ancestral mitochondrial ribosome of alphaproteobacterial descent has been reconstructed from comparative genomic data, and it has more than doubled its protein content in most eukaryotic lineages since the endosymbiosis, accumulating new components while retaining its bacterial core.16PubMed Central. Reconstructing the evolution of the mitochondrial ribosomal proteome
Mitochondria also still divide by fission, somewhat like bacteria, though the machinery has been substantially reworked. In primitive eukaryotes like certain red algae, the bacterial cell-division protein FtsZ still operates on the inner mitochondrial membrane, while a eukaryotic protein called dynamin handles the outer membrane.17PubMed. Triple immunofluorescent labeling of FtsZ, dynamin, and EF-Tu reveals a loose association between the inner and outer membrane mitochondrial division machinery in the red alga Cyanidioschyzon merolae In more recently evolved organisms like flowering plants, the bacterial FtsZ protein has been lost entirely, and dynamin-like proteins have taken over the whole job.18PubMed Central. A dynamin-like protein (ADL2b), rather than FtsZ, is involved in Arabidopsis mitochondrial division You can see the transition from bacterial to eukaryotic control playing out across the tree of life.
When the Immune System Sees the Stranger Within
One of the more medically consequential echoes of mitochondria’s bacterial past involves the immune system. Because mitochondria retain bacterial molecular signatures, when cells are damaged and mitochondrial contents leak into the bloodstream, the body’s innate immune system can mistake them for invading bacteria. Mitochondrial DNA, like bacterial DNA, carries distinctive chemical modifications. Mitochondrial membrane lipids like cardiolipin resemble bacterial cell-wall components. When these molecules escape the organelle, they act as damage-associated molecular patterns and can trigger inflammatory responses comparable to those provoked by a genuine bacterial infection.19PubMed Central. Circulating mitochondrial DAMPs cause inflammatory responses to injury
This phenomenon has been implicated in the excessive inflammation that follows major trauma, burns, and surgery. The immune system evolved to recognize bacterial molecules as threats, and mitochondria still look bacterial enough to set off that alarm.20PubMed. Innate immunity and tolerance toward mitochondria Researchers are actively investigating the role of these mitochondrial danger signals in a range of diseases, from autoimmune conditions to neurodegenerative disorders, though much remains unexplored.21PubMed. Immunomodulatory role of mitochondrial DAMPs: a missing link in pathology?
Can Cells Survive Without Mitochondria?
Given how central mitochondria are to eukaryotic life, a natural question is whether any eukaryote has managed to get rid of them. For a long time, the assumption was no. But a handful of organisms have pushed back against that idea in surprising ways.
Many organisms living in low-oxygen environments have drastically reduced their mitochondria. In some anaerobic parasites, the organelle has lost the ability to produce energy via the standard oxygen-dependent pathway and has been repurposed. These reduced versions go by names like hydrogenosomes, which produce hydrogen instead of consuming oxygen, and mitosomes, which are so stripped down they appear to do little more than assemble iron-sulfur clusters essential for certain enzymes.22PubMed. Highly divergent mitochondrion-related organelles in anaerobic parasitic protozoa These are still recognizably derived from the original mitochondrial endosymbiont, just radically downsized.
The most extreme case is the protist Monocercomonoides exilis, which appears to have lost its mitochondria entirely. Extensive analysis of its genome found no trace of mitochondrial genes or a mitochondrial organelle. The organism compensates by using a bacterial-type system for assembling iron-sulfur clusters, a function that is typically the last one retained by even the most reduced mitochondrial remnants. Apart from the absence of mitochondria and peroxisomes, the organism has otherwise standard eukaryotic complexity, including a full cytoskeleton and endomembrane system.23PubMed Central. The Oxymonad Genome Displays Canonical Eukaryotic Complexity in the Absence of a Mitochondrion It lives inside the guts of chinchillas, a low-oxygen environment where aerobic respiration isn’t needed. This discovery demonstrates that while mitochondria were present at the very origin of eukaryotic cells, they are not, in the strictest sense, absolutely required once alternative metabolic arrangements are in place.
Why You Inherit Mitochondria from Only One Parent
In the vast majority of sexually reproducing species, mitochondria pass from mother to offspring only. Sperm contribute essentially none. This pattern of strict uniparental inheritance probably evolved as a defense mechanism. If mitochondria were inherited from both parents, “selfish” mutations in mitochondrial DNA that benefited their own replication at the expense of the host cell could spread through a population much more rapidly. Having only one parent contribute mitochondria keeps competition between genetically different mitochondrial lineages to a minimum, limiting the damage selfish mutations can do.24PubMed. Evolutionary origin and consequences of uniparental mitochondrial inheritance
This inheritance pattern has an interesting evolutionary side effect. Because mitochondrial DNA passes only through females, natural selection can only “see” mitochondrial mutations through their effects on female fitness. A mutation that harms males but has no effect on females will not be selected against. Evolutionary theory predicts this should lead to an accumulation of mitochondrial mutations that impair male fertility, and experimental evidence in multiple species is consistent with that prediction.25PubMed. Maternal inheritance of mitochondria: implications for male fertility? This so-called “mother’s curse” is a direct consequence of the way we inherited our ancient bacterial tenants.
A Modern Endosymbiosis in Progress
One reason the mitochondrial origin story is so compelling is that we can watch a similar process unfolding in real time, or at least in geological slow motion. The amoeba Paulinella chromatophora acquired a photosynthetic cyanobacterium as an endosymbiont relatively recently in evolutionary terms. This internal cyanobacterium, called a chromatophore, has already lost much of its genome and depends on the host for survival. The host cell has even begun importing proteins into the chromatophore, a step that closely parallels what happened with mitochondria billions of years ago.26PubMed Central. Trafficking of protein into the recently established photosynthetic organelles of Paulinella chromatophora Studying Paulinella gives researchers a living snapshot of the early stages of organelle formation, the transition from independent symbiont to fully integrated cellular component. It suggests that the process by which mitochondria became organelles, while extraordinary, is not a one-off miracle. Given the right conditions, cells can repeat this trick.