Endosymbiotic theory proposes that some of the most important structures inside your cells, particularly mitochondria and chloroplasts, were once free-living bacteria that took up permanent residence inside a host cell billions of years ago. What began as a fringe idea in the late 1960s is now one of the most well-supported explanations in biology, backed by converging lines of evidence from genetics, biochemistry, and cell structure. The story is richer and stranger than many people realize, though, because endosymbiosis did not just happen once and stop.
Where the Idea Came From
Scientists had noticed the resemblance between mitochondria and bacteria as far back as the late 1800s, but the idea never gained traction. In 1967, Lynn Margulis (then publishing under the name Lynn Sagan) laid out a detailed, testable version of the hypothesis in a paper called “On the Origin of Mitosing Cells.” She argued that the eukaryotic cell, the complex cell type found in animals, plants, fungi, and protists, was not built by a single lineage gradually gaining complexity. Instead, it was assembled from mergers between different microbes.1PubMed Central. Lynn Margulis and the endosymbiont hypothesis: 50 years later She proposed endosymbiotic origins not only for mitochondria and chloroplasts, but also for structures like the eukaryotic flagellum. Not all of those ideas held up. The flagellum claim, for instance, never found compelling evidence. But the mitochondria and chloroplast parts of her hypothesis proved transformative.2PubMed. On the origin of mitosing cells: A historical appraisal of Lynn Margulis endosymbiotic theory
Margulis herself acknowledged she was not the first to float these ideas, but no one before her had pulled the strands together so thoroughly or generated so many testable predictions. Her 1967 paper was rejected by over a dozen journals before it was published. Within a couple of decades, as molecular biology tools matured, the evidence stacked up so convincingly that the endosymbiotic origin of mitochondria and plastids became textbook biology.3PubMed. Revisiting the theoretical basis of the endosymbiotic origin of plastids in the original context of Lynn Margulis on the origin of mitosing, eukaryotic cells
The Core Claim for Mitochondria
Mitochondria, the structures that generate most of a cell’s usable energy, descend from an alphaproteobacterium that was engulfed by (or invaded) an ancient host cell closely related to a group of archaea called the Asgard archaea.4Current Biology. The Origin and Evolution of Mitochondria That merger likely happened somewhere around two billion years ago, and every mitochondrion in every organism today traces back to that single event. The host cell was probably an archaeon, not a bacterium, which means you are, at the deepest level, an archaeal cell running on bacterial power plants.
Recent genomic analysis has confirmed that the Asgard archaea are the closest known archaeal relatives of eukaryotes, while the mitochondrial contribution from the alphaproteobacterial endosymbiont was largely concentrated in energy metabolism and iron-sulfur cluster assembly.5PubMed Central. Dominant contribution of Asgard archaea to eukaryogenesis That finding makes intuitive sense: the engulfed bacterium’s main value to its host was its ability to use oxygen to produce energy far more efficiently than the host could alone.
The Core Claim for Chloroplasts
Chloroplasts, the structures that carry out photosynthesis in plants and algae, have their own endosymbiotic origin, separate from the mitochondrial one. They descend from an ancient cyanobacterium, a photosynthetic bacterium sometimes loosely called “blue-green algae.”6PubMed Central. Are Cyanobacteria an Ancestor of Chloroplasts or Just One of the Gene Donors for Plants and Algae? This primary endosymbiosis happened in the ancestor of a large group called the Archaeplastida, which includes land plants, green algae, red algae, and a lesser-known lineage called glaucophytes. The closest known living relative of the cyanobacterium that gave rise to plastids appears to be a freshwater species called Gloeomargarita lithophora.7PubMed Central. An Early-Branching Freshwater Cyanobacterium at the Origin of Plastids
So in simplified terms, the recipe for a plant cell involved at least two major endosymbiotic events: first the mitochondrial ancestor was acquired, then a cyanobacterium was incorporated as a chloroplast. The two events happened independently and at different times.
Lines of Evidence That Clinched It
No single piece of evidence proved the theory. Instead, multiple independent lines of evidence all pointed the same direction, which is what made the case so persuasive.
Their Own DNA
Both mitochondria and chloroplasts carry their own small, circular genomes that look far more like bacterial chromosomes than like the linear chromosomes in a cell’s nucleus. These organellar genomes encode some of their own proteins and use a gene-expression system with strong bacterial characteristics. A review of the structural and genomic similarities between human mitochondria and bacteria found that the parallels extend beyond genome shape to include the DNA repair machinery, particularly the base excision repair system that fixes oxidative damage.8PubMed Central. The Similarities between Human Mitochondria and Bacteria in the Context of Structure, Genome, and Base Excision Repair System
Bacterial-Type Ribosomes
Mitochondria and chloroplasts build proteins using ribosomes that resemble bacterial ribosomes rather than the larger ribosomes found in the cell’s own cytoplasm. Early work on the chloroplast ribosomes of the green alga Chlamydomonas showed that antibiotic resistance patterns in those ribosomes matched what researchers saw in bacteria, supporting the idea that at least some chloroplast ribosomal proteins are encoded by genes in chloroplast DNA.9PubMed Central. Localization of five antibiotic resistances at the subunit level in chloroplast ribosomes of Chlamydomonas This detail turns out to have real medical significance, as we will see later.
Double Membranes
Mitochondria and chloroplasts are enclosed by two membranes. The inner membrane corresponds to the original bacterium’s own cell membrane, while the outer membrane is thought to derive from the host cell’s engulfing membrane. The inner membranes of both organelles contain cardiolipin, a distinctive phospholipid that is otherwise found almost exclusively in bacterial membranes and energy-transducing membranes.10PubMed. Cardiolipin, a lipid found in mitochondria, hydrogenosomes and bacteria was not detected in Giardia lamblia You do not find cardiolipin in most other eukaryotic membranes, which is a chemical fingerprint of the organelles’ bacterial past.
Division by Fission
Bacteria reproduce by splitting in two, a process called binary fission. Mitochondria and chloroplasts do the same thing inside the cell rather than being built from scratch by the host. Chloroplasts in particular retain a striking molecular echo of this ancestry: they use the FtsZ protein to organize their division, the same protein that bacteria use to form a constricting ring at the cell’s midpoint during division.11PubMed Central. FtsZ and the division of prokaryotic cells and organelles Plastid division in plants operates through an elaborate system that combines these ancestral bacterial components with newer eukaryotic proteins, but the cyanobacterial imprint remains clearly visible in the positioning and ring-forming machinery.12PubMed. Plant organelle division orchestrated by the mosaic machinery of endosymbiotic relics and eukaryotic host factors
Mitochondrial division is more complex and appears to have replaced FtsZ with dynamin-related proteins in most lineages, but the basic principle of fission-style division rather than de novo construction still holds.13PubMed. Organelle division: from coli to chloroplasts
The Massive Gene Migration to the Nucleus
If mitochondria and chloroplasts were once independent bacteria, you might expect them to still carry large genomes encoding everything they need. They do not. Over hundreds of millions of years, the vast majority of the endosymbiont’s genes have migrated to the host cell’s nucleus. The mitochondrial genome in humans encodes only 37 genes, while a free-living alphaproteobacterium typically carries thousands. The rest have been transferred to the nuclear genome, where they are transcribed, and their protein products are shipped back into the mitochondrion.14PubMed Central. The functional transfer of genes from the mitochondria to the nucleus: the effects of selection, mutation, population size and rate of self-fertilization
This gene transfer is not just ancient history. It is an ongoing process that researchers can track in living organisms. Modeling work has shown that the rate and probability of successful gene transfer depend on factors like mutation rates, population size, and whether the organism self-fertilizes. The practical result is that the endosymbiont becomes increasingly dependent on the host for proteins it can no longer make itself, tightening the partnership into something irreversible.
Secondary and Tertiary Endosymbiosis
The story gets wilder. The primary endosymbiosis that created chloroplasts happened once, but those chloroplast-bearing cells were themselves engulfed by other eukaryotes, creating a kind of Russian-nesting-doll arrangement. This process, called secondary endosymbiosis, is how many algal lineages got their photosynthetic ability. Green algal plastids were picked up by euglenids and chlorarachniophytes, while red algal plastids were taken up in a separate event that gave rise to a huge and diverse group of organisms including diatoms, brown algae, and many dinoflagellates.15PubMed Central. The endosymbiotic origin, diversification and fate of plastids
Some dinoflagellates went even further and replaced their existing plastids by engulfing yet another alga, producing tertiary endosymbioses. In a few lineages, the engulfed alga’s nucleus has not entirely disappeared. These shrunken remnant nuclei, called nucleomorphs, survive inside some cryptophytes and chlorarachniophytes. Researchers have found similar structures in certain dinoflagellates that harbor green algal endosymbionts, with evidence of DNA still present in the nucleomorphs and gene transfer to the host genome actively underway.16PubMed Central. Dinoflagellates with relic endosymbiont nuclei as models for elucidating organellogenesis These organisms are essentially snapshots of organelle evolution in progress.
A Second Independent Origin of Plastids
For a long time, the mainstream view was that primary plastid endosymbiosis (a eukaryote engulfing a cyanobacterium and keeping it permanently) happened exactly once, roughly 1.6 billion years ago. Then researchers took a closer look at a peculiar amoeba called Paulinella chromatophora. Paulinella contains photosynthetic bodies called chromatophores that are derived from a different lineage of cyanobacteria than the one that gave rise to plant chloroplasts, and the endosymbiosis happened far more recently, an estimated 90 to 140 million years ago.17PubMed Central. Paulinella, a model for understanding plastid primary endosymbiosis
Phylogenetic analysis of Paulinella’s photosynthetic endosymbiont confirmed that it is not related to the original Archaeplastida chloroplast lineage, meaning the evolution of a photosynthetic organelle from a cyanobacterium is not a unique, unrepeatable event. It can and did happen more than once, suggesting it may be an ongoing process.18PubMed. A plastid in the making: evidence for a second primary endosymbiosis That changes how biologists think about organelle evolution: it is not a freak accident frozen in deep time, but a transition that natural selection can drive repeatedly given the right conditions.
The Nitroplast, Discovered in 2024
In a finding announced in 2024, researchers showed that an intracellular cyanobacterium called UCYN-A, living inside a marine alga called Braarudosphaera bigelowii, has crossed the line from endosymbiont to genuine organelle. UCYN-A has been integrated into the algal cell’s architecture and division cycle, and the host’s nuclear genome encodes proteins that are imported into UCYN-A, a hallmark of true organelle status.19PubMed. Nitrogen-fixing organelle in a marine alga The researchers proposed calling it a “nitroplast,” a nitrogen-fixing organelle, which would make it the first known organelle with that function.
Complementary work found that the size ratio between UCYN-A and its host is strikingly constant across different sublineages, resembling the stable size relationships between mitochondria or chloroplasts and their host cells rather than the variable ratios typical of looser symbioses.20PubMed. Metabolic trade-offs constrain the cell size ratio in a nitrogen-fixing symbiosis The nitroplast discovery is significant because nitrogen fixation, the ability to convert atmospheric nitrogen gas into biologically usable forms, was previously thought to be restricted to free-living or symbiotic bacteria and archaea. If a eukaryote can acquire this ability through endosymbiosis, it opens the door to questions about engineering nitrogen fixation into crop plants.
Endosymbiosis Caught in the Act
Paulinella and UCYN-A represent different stages along the continuum from loose partnership to full organelle, but there are even earlier stages visible in nature. The freshwater ciliate Paramecium bursaria routinely harbors Chlorella green algae inside its cells.21PubMed Central. Mutualism on the edge: Understanding the Paramecium-Chlorella symbiosis The relationship is mutualistic: the algae photosynthesize and share sugars; the Paramecium provides shelter and nutrients. But it is not obligate in the way mitochondria or chloroplasts are. You can cure a Paramecium of its Chlorella, and both can survive independently, at least under favorable conditions.
Under starvation, though, the advantage of the partnership becomes stark. Experiments showed that symbiotic Paramecium bursaria survived for up to five weeks in nitrogen- and prey-depleted conditions by continuously feeding on a fraction of its algal endosymbionts while the remaining Chlorella cells kept proliferating using host metabolites and light energy. Aposymbiotic Paramecium (those without algae) and free-living Chlorella both died or bleached under the same conditions.22PubMed Central. The closed nutrient recycling system in the Paramecium-Chlorella photosymbiosis contributes to survival under oligotrophic conditions This “cyclical farming” creates a closed nutrient-recycling loop that gives the partnership a major survival edge in poor environments. It is easy to see how, over evolutionary time, this kind of selective advantage could push a loose partnership toward obligate dependency.
What Mitochondria Became in Oxygen-Poor Environments
Not every descendant of the original mitochondrial endosymbiont looks like a textbook mitochondrion. In organisms that live in oxygen-poor or completely anaerobic environments, the mitochondrion has been repurposed. Hydrogenosomes, found in some parasitic protists and certain fungi, generate energy by producing hydrogen gas instead of using the oxygen-dependent pathway. Mitosomes, found in organisms like Giardia, are even more stripped down and appear to function mainly in assembling iron-sulfur clusters rather than producing energy at all. Both share a common ancestor with mitochondria.23PubMed Central. Multiple secondary origins of the anaerobic lifestyle in eukaryotes
Some organisms bridge the gap between these forms. The Anaeramoebae, a group of free-living protists, possess hydrogen-producing organelles that have the most “mitochondrion-like” protein set of any metamonad studied, representing what researchers describe as missing links between typical aerobic mitochondria and the highly reduced anaerobic versions.24Current Biology. Anaeramoebae are deeply-branching eukaryotes that clarify the transition from anaerobic mitochondria to hydrogenosomes
And then there is Monocercomonoides exilis, a gut protist that appears to have lost its mitochondrial organelle entirely. Genome sequencing found no trace of the hallmark mitochondrial proteins, including the iron-sulfur cluster assembly pathway that was considered universally conserved in eukaryotes. The organism replaced that pathway with a bacterial sulfur-mobilization system acquired by horizontal gene transfer.25PubMed. A Eukaryote without a Mitochondrial Organelle This is not a primitive organism that never had mitochondria; phylogenetic placement shows it lost the organelle secondarily.26PubMed Central. The Oxymonad Genome Displays Canonical Eukaryotic Complexity in the Absence of a Mitochondrion It demonstrates that while mitochondria were central to the origin of eukaryotes, they are not absolutely essential to keeping a eukaryotic cell running once alternative systems are in place.
Not Every Organelle Has an Endosymbiotic Origin
It is tempting to assume that if mitochondria and chloroplasts came from engulfed bacteria, maybe other membrane-bound compartments in the cell did too. Peroxisomes, small organelles involved in breaking down fatty acids and detoxifying harmful molecules, were once considered possible candidates for an endosymbiotic origin. But comparative genomic studies have consistently found otherwise. The protein import machinery in mitochondrial and chloroplast membranes contains clear remnants of bacterial transport systems, but peroxisomal membranes show no evidence of proteins with prokaryotic ancestry.27Molecular Biology and Evolution. The Evolutionary Origin of Peroxisomes: An ER-Peroxisome Connection Instead, the peroxisome appears to have originated from the endoplasmic reticulum, an internal membrane system that the host cell built on its own.28PubMed Central. Origin and evolution of the peroxisomal proteome
This distinction matters because it establishes limits on the theory. Endosymbiosis explains the origin of specific organelles, not all compartmentalization in eukaryotic cells. The nucleus itself remains debated; some researchers have proposed endosymbiotic origins for it, but there is no consensus and no evidence as strong as what supports the mitochondrial and plastid cases.
Why Antibiotics Sometimes Harm Your Cells
The bacterial ancestry of mitochondria has a practical consequence that shows up in medicine. Many antibiotics work by targeting bacterial ribosomes or other bacterial machinery. Because mitochondrial ribosomes still closely resemble their bacterial counterparts, some antibiotics can accidentally inhibit mitochondrial function as a side effect. Aminoglycosides (used for serious infections) and oxazolidinones (used for drug-resistant bacteria) are among the classes whose side effects appear to result from direct inhibition of mitochondrial ribosomes, echoing their intended antibacterial action.29PubMed Central. Adverse effects of antimicrobials via predictable or idiosyncratic inhibition of host mitochondrial components
This is not a minor pharmacological footnote. Mitochondrial toxicity is implicated in hearing loss from aminoglycosides, lactic acidosis from certain antiretrovirals, and liver damage from other drug classes. Recognizing that mitochondria are, in a sense, tamed bacteria has given researchers a framework for predicting and understanding these side effects rather than treating each one as an isolated surprise.
Open Questions About the Original Merger
The evidence for endosymbiotic origins of mitochondria and chloroplasts is overwhelming, but significant questions remain about the details of the original events. Researchers disagree about whether the mitochondrial endosymbiont was acquired early in the evolution of eukaryotic cells or late, after the host had already developed some degree of cellular complexity. Some models envision a simple archaeal host swallowing a bacterium as essentially the founding act of the eukaryotic lineage. Others propose that the host was already a sophisticated proto-eukaryote with internal membranes and other features before the mitochondrion arrived.30PubMed Central. Relative timing of mitochondrial endosymbiosis and the “pre-mitochondrial symbioses” hypothesis
The nature of the initial relationship is also unclear. Was the engulfed bacterium a parasite that was eventually domesticated? A prey item that resisted digestion? A mutualist from the start? Modeling work suggests that mitochondria, or respiring symbionts more broadly, expanded the maximum cell volume that cells could achieve, potentially enabling the larger, more complex cells we associate with eukaryotes. But they are not strictly required for all the cell-size and division-time combinations seen in modern eukaryotes, which means the energetic advantage was real but not the only factor at play.31PubMed Central. The role of mitochondrial energetics in the origin and diversification of eukaryotes These debates are active, with new genomic data from Asgard archaea and deep-branching eukaryotic lineages regularly reshaping what seems plausible.