Mitochondria are organelles that live inside eukaryotic cells, but they descended from a free-living prokaryote, an ancient bacterium that was absorbed by a host cell roughly two billion years ago. That dual identity makes the question surprisingly tricky: mitochondria carry unmistakable bacterial fingerprints in their DNA, their ribosomes, and the way they divide, yet they are so deeply woven into eukaryotic cell biology that they cannot survive on their own. The honest answer is that they are neither fully one nor the other, and the tension between those two identities turns out to have real consequences, from how your cells manage energy to why certain antibiotics can make you feel terrible.
How a Bacterium Became an Organelle
The story begins with endosymbiosis. Somewhere around 1.5 to 2 billion years ago, a host cell that was related to a group of archaea called the Asgard archaea took in a bacterium from the proteobacterial lineage. Instead of being digested, that bacterium survived, kept dividing, and eventually became a permanent resident. Over vast stretches of evolutionary time, the relationship tightened: the bacterium surrendered most of its genes to the host’s nucleus, lost the ability to live independently, and became what we now call the mitochondrion.
Lynn Margulis, then a 29-year-old biologist named Lynn Sagan, laid out this endosymbiotic hypothesis in a landmark 1967 paper. Her proposal was considered radical at the time, but molecular evidence accumulated over the following decades confirmed the core idea: mitochondria and chloroplasts both trace back to bacterial ancestors that were absorbed by early eukaryotic cells.1PubMed Central. Lynn Margulis and the endosymbiont hypothesis: 50 years later Comparative genomics has since confirmed that the closest bacterial relatives of mitochondria are alphaproteobacteria, while plastids trace to cyanobacteria.2PubMed. Prokaryotic symbiotic consortia and the origin of nucleated cells: A critical review of Lynn Margulis hypothesis
Recent research points to Asgard archaea as the closest living relatives of the host cell that took in the mitochondrial ancestor. A 2025 analysis found that Asgard archaea contributed the majority of genes underlying key eukaryotic systems and pathways, with the capture of the alphaproteobacterial endosymbiont coming after many features of eukaryotic cell organization had already begun to evolve in the Asgard lineage.3Nature. Dominant contribution of Asgard archaea to eukaryogenesis Earlier work on Lokiarchaeota, a phylum within the Asgard group, had already revealed that these archaea possess many genes previously thought to be unique to eukaryotes, supporting a scenario in which an archaeal host merged with a bacterial endosymbiont to produce the first eukaryotic cell.4Nature. Asgard archaea illuminate the origin of eukaryotic cellular complexity
How the initial relationship worked is still debated. Some researchers have proposed that the two organisms started out in a mutually beneficial metabolic partnership, a syntrophy, where each partner supplied something the other needed. Others point out that there is little direct evidence for that scenario and argue instead that the relationship may have begun as parasitism or predation, with the bacterium exploiting the host or the host engulfing the bacterium as prey.5PubMed Central. Endosymbiosis before eukaryotes: mitochondrial establishment in protoeukaryotes Either way, the endosymbiont eventually became indispensable.
The Prokaryotic Traits Mitochondria Still Carry
If you could shrink down and inspect a mitochondrion up close, you would find several features that scream “bacterium.” These are relics of the organelle’s free-living past, and they have persisted for nearly two billion years despite the enormous evolutionary overhaul that transformed the endosymbiont into an organelle.
The most famous holdover is mitochondrial DNA. Human mitochondrial DNA is a small, circular molecule of about 16,569 base pairs encoding 37 genes: two ribosomal RNAs, 22 transfer RNAs, and 13 proteins.6Biochimica et Biophysica Acta (BBA) – Bioenergetics. The mitochondrial genome: structure, transcription, translation and replication That circular chromosome is organized into compact structures called nucleoids, echoing how bacteria package their DNA. This is tiny compared to the nuclear genome, but its very existence sets mitochondria apart from every other organelle except chloroplasts. No purely eukaryotic structure has its own genome.
Mitochondria also have their own ribosomes, and these ribosomes are more similar to bacterial ribosomes than to the ribosomes floating in the eukaryotic cytoplasm. This similarity is not just a curiosity. Antibiotics designed to target bacterial ribosomes can also interfere with mitochondrial ribosomes, a fact with medical implications we will return to later. Research on resistance mutations in bacteria has shown that those mutations can predict whether mitochondrial ribosomes in eukaryotic cells will also be sensitive to a given drug.7PubMed Central. Structural basis for selectivity and toxicity of ribosomal antibiotics
The way mitochondria reproduce is bacterial in flavor as well. They do not arise from scratch; they divide from pre-existing mitochondria through a process that can be separated into division of the internal DNA and division of the matrix compartment itself.8Biochimica et Biophysica Acta (BBA) – Molecular Cell Research. Structure, function and evolution of the mitochondrial division apparatus Researchers have described the machinery responsible for this as a “mitochondrial divisome,” conceptually and operationally similar to the bacterial cell-division machinery.9PubMed Central. Function and regulation of the divisome for mitochondrial fission Even some of the lipids in the mitochondrial membrane trace back to bacterial origins. Cardiolipin, a lipid critical for mitochondrial stability and function, is also a hallmark of bacterial membranes.10PubMed Central. An essential bacterial-type cardiolipin synthase mediates cardiolipin formation in a eukaryote
How Mitochondria Became Eukaryotic Team Players
Despite all those bacterial holdovers, modern mitochondria are profoundly integrated into eukaryotic cell life in ways that no independent bacterium could be. The transition from endosymbiont to organelle involved hundreds of new genes, the evolution of protein import machinery, and the merging of metabolism and reproduction between the two partners.11Current Biology. The Origin and Evolution of Mitochondria
The single biggest change was gene transfer. Over time, the vast majority of genes that once resided in the endosymbiont’s genome were shipped to the host cell’s nucleus, a process called endosymbiotic gene transfer.12PubMed Central. Endosymbiotic gene transfer from prokaryotic pangenomes: Inherited chimerism in eukaryotes This means that most of the proteins a mitochondrion needs are now encoded by nuclear DNA, manufactured by ribosomes in the cytoplasm, and then shipped back into the mitochondrion. The organelle kept just a handful of its original genes, enough to make 13 proteins in humans.
Getting those nuclear-encoded proteins back into the mitochondrion requires sophisticated import machinery that has no equivalent in free-living bacteria. The main gateway is a molecular machine called the TOM complex, which sits in the outer mitochondrial membrane and acts as an entry gate for the bulk of incoming proteins.13PubMed Central. A Biochemical and Structural Understanding of TOM Complex Interactions and Implications for Human Health and Disease Once through that outer gate, proteins are sorted to their final destinations by a pair of inner-membrane machines, TIM23 and TIM22, each handling a different class of cargo.14PubMed. Protein translocation into mitochondria: the role of TIM complexes This TOM/TIM import system is a eukaryotic innovation. No bacterium runs anything like it, because no free-living bacterium needs to import proteins manufactured by a host.
Mitochondria also took on roles that no free-living bacterium performs. One of the most striking is their involvement in programmed cell death, or apoptosis. When a eukaryotic cell receives a signal to self-destruct, mitochondria release cytochrome c, a protein that normally serves as an electron carrier in the energy-producing respiratory chain, into the cytoplasm. Once free in the cytoplasm, cytochrome c triggers a cascade of enzymes called caspases that dismantle the cell from the inside.15PubMed Central. Mitochondrial cytochrome c release in apoptosis occurs upstream of DEVD-specific caspase activation and independently of mitochondrial transmembrane depolarization This dual function, powering respiration when things are normal and triggering cell death when things go wrong, means cytochrome c serves as a kind of life-or-death switch.16Biochimica et Biophysica Acta (BBA) – Bioenergetics. Mitochondrial control of apoptosis: the role of cytochrome c Apoptosis is a strictly eukaryotic process, and mitochondria sit at its center.
Building the energy-producing complexes inside mitochondria requires coordination between two genomes working in two different compartments of the cell. The large protein complexes that carry out oxidative phosphorylation contain subunits encoded by both nuclear and mitochondrial DNA, meaning the cell has to balance gene expression across both genomes to assemble these complexes correctly.17Biochemical Society Transactions. Is mitochondrial gene expression coordinated or stochastic? How cells pull off that coordination is still not fully understood, but it underscores how deeply mitochondrial biology has become entangled with eukaryotic gene regulation.
The Ongoing Debate About the Exact Ancestor
Researchers agree that mitochondria came from a proteobacterium. They agree the host was related to Asgard archaea. Beyond that, the phylogenetic details get contentious. The question of exactly where on the bacterial family tree the mitochondrial ancestor sits has generated conflicting answers depending on the methods and data sets used.
One line of analysis placed mitochondria squarely within the order Rickettsiales, as a sister group to well-known intracellular bacteria in the Rickettsiaceae and Anaplasmataceae families.18Scientific Reports. An integrated phylogenomic approach toward pinpointing the origin of mitochondria That conclusion made intuitive sense because many rickettsial bacteria are already obligate intracellular parasites, so the idea of one of them becoming an endosymbiont felt natural. But a 2018 study using methods specifically designed to correct for biases in sequence analysis reached a very different conclusion: mitochondria branched off before the diversification of all currently known alphaproteobacterial lineages, meaning they do not belong within Rickettsiales or any other recognized alphaproteobacterial group.19Nature. Deep mitochondrial origin outside the sampled alphaproteobacteria
A 2020 study tried to break the tie using broader taxonomic sampling. Its results placed mitochondria back within Alphaproteobacteria, sharing an ancient common ancestor with Rickettsiales and some unclassified marine lineages, but not nested within Rickettsiales itself.20Nature Ecology & Evolution. Phylogenetic analyses with systematic taxon sampling show that mitochondria branch within Alphaproteobacteria The disagreement is not a failure of science; it reflects how difficult it is to reconstruct evolutionary relationships that are nearly two billion years old. Gene sequences change so much over that timescale that the statistical methods used to build family trees can be misled by convergent patterns in the data. Each study improves the analysis, but the community has not yet converged on a single answer.
When Mitochondria Are Not Really Mitochondria
In textbook depictions, mitochondria are bean-shaped organelles with neat internal folds, humming along producing energy via oxygen-dependent respiration. But in the broader world of eukaryotic life, the picture is far messier. Many single-celled eukaryotes that live in low-oxygen or oxygen-free environments have modified their mitochondria so extensively that the organelles barely resemble the textbook version. These variants are collectively called mitochondrion-related organelles, or MROs.21PubMed Central. Diversity and origins of anaerobic metabolism in mitochondria and related organelles
Some MROs, called hydrogenosomes, produce hydrogen gas instead of running the standard oxygen-dependent respiratory chain. Others, called mitosomes, have lost almost all metabolic function and appear to serve mainly in assembling iron-sulfur clusters, a biochemical task so essential that the cell cannot give it up entirely. Within the group Metamonada, for example, mitochondria exist in these various reduced forms.22PubMed Central. Retortamonads from vertebrate hosts share features of anaerobic metabolism and pre-adaptations to parasitism with diplomonads
And then there is the most extreme case. The oxymonad Monocercomonoides exilis appears to have lost its mitochondrion entirely. When researchers sequenced its genome, they found no genes for any hallmark mitochondrial proteins. Phylogenetic context indicates this organism is not primitively amitochondrial; it had a mitochondrion and lost it, making it the first known eukaryote to have completely dispensed with the organelle.23PubMed. A Eukaryote without a Mitochondrial Organelle Follow-up work showed that M. exilis still retains proteins in its genome that would be capable of being imported into a mitochondrion-related organelle if one were present, a ghostly molecular memory of the organelle it once had.24Open Biology. A mitochondrion-free eukaryote contains proteins capable of import into an exogenous mitochondrion-related organelle
Similar detective work on the human parasite Entamoeba histolytica found genes for proteins that in other eukaryotes reside in the mitochondrion, including a chaperone protein whose evolutionary tree clearly places it within the mitochondrial lineage. This suggests that E. histolytica once had a mitochondrion and still harbors a relic of it.25PubMed. Direct evidence for secondary loss of mitochondria in Entamoeba histolytica These cases reinforce the idea that mitochondria are nearly universal among eukaryotes, but not quite. A handful of lineages have found ways to survive without them, or at least without recognizable versions of them.
What Counts as an Organelle Versus an Endosymbiont
If mitochondria started as bacteria living inside another cell, you might wonder at what point they stopped being endosymbionts and became organelles. This is not just a semantic question; it matters for how biologists classify newly discovered symbioses in nature. Researchers have proposed several criteria for drawing the line. These generally include genetic integration, meaning the transfer of genes from the endosymbiont to the host nucleus; cellular integration, meaning the endosymbiont’s cell cycle is synchronized with the host’s; and metabolic integration, meaning neither partner can survive without the other.26PubMed Central. What’s in a name? How organelles of endosymbiotic origin can be distinguished from endosymbionts A further test has been proposed: whether the two partners have become so coupled that independent speciation, one lineage splitting without the other, is no longer possible.
Mitochondria pass all of these tests. They have transferred the bulk of their genes to the nucleus. They divide in step with the cell. They are metabolically essential in most eukaryotes, and no mitochondrial lineage has ever diverged independently from its host lineage. By every proposed criterion, mitochondria are organelles, not merely bacteria living inside cells. But the boundary is a gradient, not a switch, and biologists continue to debate where exactly along that gradient newly discovered endosymbioses (like the nitrogen-fixing organelle recently identified in certain algae) should be placed.
Why the Bacterial Heritage Shows Up at the Pharmacy
The prokaryotic ancestry of mitochondria has a surprisingly direct impact on medicine. Many antibiotics work by targeting the ribosomes of disease-causing bacteria, disrupting their ability to make proteins. Because mitochondrial ribosomes still resemble their bacterial ancestors more than they resemble the ribosomes of the eukaryotic cytoplasm, some of these drugs hit mitochondrial protein synthesis as collateral damage.
Laboratory studies have shown that certain commonly used antibiotics, including tetracyclines and thiostrepton, inhibit mitochondrial protein synthesis with a potency similar to their effect on E. coli.27PubMed. Antibiotic susceptibility of mammalian mitochondrial translation More recent ribosome-profiling experiments found that chloramphenicol and linezolid stall mitochondrial ribosomes in patterns consistent with how those same drugs stall bacterial ribosomes, with a pronounced tendency to pause when certain amino acids occupy a particular position in the growing protein chain.28Nucleic Acids Research. Context-specific inhibition of mitochondrial ribosomes by phenicol and oxazolidinone antibiotics
For most healthy people, this mitochondrial side effect is mild enough to go unnoticed during a standard course of antibiotics. But for patients who already have defects in mitochondrial translation, the additional hit from an antibiotic can be disproportionately harmful. Research on fibroblasts from patients with mitochondrial translation defects found that exposure to translation-targeted antibiotics adversely affected cell growth, suggesting that these individuals may be more vulnerable to side effects from certain drugs.29PubMed. Antibiotic effects on mitochondrial translation and in patients with mitochondrial translational defects
Growing awareness of these effects has fed into broader conversations about antibiotic stewardship. The fact that antibiotics can damage mammalian mitochondria has been cited as one more reason, alongside the headline concern of antibiotic resistance, to avoid using these drugs when they are not genuinely needed.30PubMed Central. Mitochondria and Antibiotics: For Good or for Evil? Structural studies of ribosomal drug binding sites are also being used to guide the design of future antibiotics that distinguish more cleanly between bacterial and mitochondrial ribosomes, potentially reducing off-target effects on human cells.27PubMed. Antibiotic susceptibility of mammalian mitochondrial translation
In a sense, the pharmacological vulnerability of mitochondria is a two-billion-year-old inheritance. Every time a drug designed to kill bacteria also knocks the wind out of your mitochondria, it is because your cells still carry a little piece of a prokaryote inside them, one that never quite finished the transition to fully eukaryotic life.