Multiple independent lines of evidence support the endosymbiotic theory, making it one of the best-substantiated explanations in evolutionary biology. DNA sequencing, phylogenetic analysis, structural similarities between organelles and free-living bacteria, shared sensitivity to antibiotics, and even laboratory experiments that recreate early steps of the process all point to the same conclusion: mitochondria descended from a bacterium that took up residence inside another cell, and chloroplasts descended from a photosynthetic cyanobacterium captured in a similar way. The story, though, is richer and stranger than that two-sentence summary suggests.
How the Idea Took Shape
The notion that organelles were once independent organisms dates back to the late nineteenth century, but it languished for decades as a fringe idea. In 1967, Lynn Margulis (then publishing as Lynn Sagan) revived and formalized the hypothesis in a landmark paper, “On the Origin of Mitosing Cells,” proposing that mitochondria and plastids both originated from bacterial ancestors engulfed by a host cell. She also proposed that the eukaryotic flagellum came from a spirochete-like endosymbiont, a claim that has not held up. But her core predictions about mitochondria and chloroplasts have been confirmed by evidence from phylogenetics, biochemistry, and cell biology accumulated over the following half-century.1PubMed. Serial Endosymbiosis Theory: From biology to astronomy and back to the origin of life Margulis was not the first person to suggest that symbiosis played a role in organelle origins, but no one before her had assembled the idea so comprehensively or laid out testable hypotheses with such specificity.2PubMed. On the origin of mitosing cells: A historical appraisal of Lynn Margulis endosymbiotic theory
The DNA Inside Organelles
The single most persuasive piece of evidence is that mitochondria and chloroplasts carry their own DNA, and that DNA is unmistakably bacterial. Mitochondrial DNA is typically a small, circular molecule encoding a handful of genes for proteins, ribosomal RNAs, and transfer RNAs. The circular structure echoes a bacterial chromosome rather than the linear chromosomes found in a eukaryotic nucleus. More importantly, when researchers compare the sequences of mitochondrial genes to genes in free-living bacteria, the match is clear: mitochondrial sequences cluster with those of alphaproteobacteria. Analysis using comprehensive phylogenetic methods and broad taxon sampling places mitochondria robustly within the alphaproteobacterial group, sharing an ancient common ancestor with the Rickettsiales order and some currently unclassified marine lineages.3PubMed. Phylogenetic analyses with systematic taxon sampling show that mitochondria branch within Alphaproteobacteria Network-based analyses have independently reinforced this Alphaproteobacteria-sister placement.4PubMed Central. On the origin of mitochondria: a multilayer network approach
Chloroplast DNA tells a parallel story. The genes inside chloroplasts are clearly cyanobacterial in origin, and phylogenomic analyses using the most comprehensive datasets available strongly support the conclusion that plastids evolved from deep-branching cyanobacteria. The closest known cultured relative of primary plastids is a freshwater cyanobacterium called Gloeomargarita lithophora.5PubMed Central. An Early-Branching Freshwater Cyanobacterium at the Origin of Plastids The phylogenetic relationship between chloroplast genomes and cyanobacterial genomes remains one of the foundational pieces of evidence for endosymbiotic origin.6PubMed Central. Are Cyanobacteria an Ancestor of Chloroplasts or Just One of the Gene Donors for Plants and Algae?
Ribosomes, Antibiotics, and Bacterial Machinery
Mitochondria and chloroplasts do not just have bacterial DNA; they use it in a distinctly bacterial way. Both organelles have their own ribosomes, the molecular machines that translate genetic instructions into proteins. These organellar ribosomes resemble bacterial ribosomes far more than the ribosomes floating in the surrounding cytoplasm of the same cell. One of the most striking demonstrations of this comes from antibiotic sensitivity. Chloramphenicol, a drug that blocks protein synthesis in bacteria by binding to the ribosomal peptidyl transfer center, also inhibits mitochondrial translation in a similar manner. Cryo-electron microscopy of the human mitoribosome bound by linezolid (another antibiotic in the same functional class) shows extensive similarity to the way the drug inhibits bacterial ribosomes.7PubMed Central. Context-specific inhibition of mitochondrial ribosomes by phenicol and oxazolidinone antibiotics Tetracycline and thiostrepton, antibiotics with different mechanisms, likewise inhibit both E. coli and mitochondrial protein synthesis to a similar degree.8PubMed. Antibiotic susceptibility of mammalian mitochondrial translation
If mitochondria had arisen through some purely internal rearrangement of a single cell’s membranes, there would be no reason for their protein-making machinery to be vulnerable to drugs designed to target bacteria. The fact that it is vulnerable, and vulnerable in the same sequence-specific way, is a powerful independent argument for bacterial ancestry.
How Organelles Divide
Mitochondria and chloroplasts do not arise from scratch. They reproduce by splitting in two, much the way bacteria do. In bacteria, cell division depends on a protein called FtsZ, which assembles into a ring at the cell’s midpoint and constricts to pinch the cell apart. Chloroplasts use the same protein. FtsZ self-assembles into a membrane-associated ring structure early in the chloroplast division process, directly paralleling its role in bacterial binary fission.9PubMed Central. FtsZ and the division of prokaryotic cells and organelles The fact that chloroplasts retained this ancestral division machinery, rather than evolving an entirely new system, speaks to a shared evolutionary past with free-living bacteria.10PubMed. Organelle division: from coli to chloroplasts
Mitochondria, interestingly, have largely replaced FtsZ with a different protein called dynamin for their division, though some primitive protists still retain mitochondrial FtsZ. This kind of evolutionary tinkering is expected: over more than a billion years inside a host cell, organelles have had plenty of time to swap out ancestral parts for host-derived substitutes, all while keeping the basic “divide by fission” strategy.
Double Membranes and Internal Structure
Both mitochondria and chloroplasts are enclosed by two membranes. The standard explanation is straightforward: the inner membrane is the original membrane of the engulfed bacterium, while the outer membrane is derived from the engulfing cell’s membrane that wrapped around it during the initial uptake event. Bacteria that were never engulfed have only a single plasma membrane (plus, in Gram-negative species, an outer membrane of their own, which complicates the picture but does not undermine the basic logic).
Beyond the double membrane itself, the internal architecture of mitochondria echoes bacterial structures. Mitochondrial cristae, the folds of inner membrane where energy production occurs, resemble the intracytoplasmic membranes found in certain bacteria that similarly expand their membrane surface area for bioenergetic reactions. Morphological similarities between cristae and these bacterial intracytoplasmic membranes have been proposed as an additional line of evidence for the prokaryotic origins of mitochondria.11PubMed Central. Origin and evolution of mitochondrial inner membrane composition
Genes That Moved to the Nucleus
One of Margulis’s specific predictions was that eukaryotic genomes would turn out to be chimeric, a patchwork of genes from different ancestral sources. That prediction has been confirmed in detail. Over evolutionary time, the vast majority of genes originally carried by the engulfed bacterium have been transferred to the host cell’s nuclear genome, a process called endosymbiotic gene transfer. The result is that modern mitochondria encode only a tiny fraction of the proteins they need; the rest are made in the cytoplasm from nuclear genes and then imported back into the organelle through specialized protein import machinery.12PubMed Central. Tracing the evolutionary pathway: on the origin of mitochondria and eukaryogenesis
The same pattern holds for chloroplasts. Nuclear genomes of plants and algae are riddled with genes of cyanobacterial origin that clearly came from the ancestral plastid. Morphological, biochemical, and geochemical evidence, combined with phylogenomic analyses, have demonstrated the validity of the prediction that eukaryotic genomes are mosaics of genes from the host and its endosymbionts.2PubMed. On the origin of mitosing cells: A historical appraisal of Lynn Margulis endosymbiotic theory This gene transfer is not merely a curiosity. It means that the endosymbiont became permanently dependent on the host for most of its proteins, while the host became dependent on the endosymbiont for energy metabolism or photosynthesis. Neither partner can survive without the other.
Who Was the Host Cell?
For decades, the identity of the cell that swallowed the ancestral mitochondrion was a mystery. Genomic discoveries over the past ten years have pointed toward a group of archaea called the Asgardarchaeota as the closest living relatives of the eukaryotic host lineage. Phylogenetic analyses using best-fitting substitution models robustly support the placement of various Asgard archaeal lineages in positions consistent with a deep relationship to eukaryotes.13PubMed Central. Phylogenomic Analyses Reveal that Panguiarchaeum Is a Clade of Genome-Reduced Asgard Archaea Within the Njordarchaeia This fits neatly with models proposing that the endosymbiotic event involved an alphaproteobacterium engulfed by a host cell originating from the Asgardarchaeota superphylum.11PubMed Central. Origin and evolution of mitochondrial inner membrane composition
The question of what drove this partnership remains actively debated. One influential class of models proposes that the initial relationship was metabolic: the two organisms benefited from exchanging small molecules like hydrogen or organic acids. The Hydrogen Hypothesis and the related Syntrophy Hypothesis suggest that eukaryotes arose through metabolic symbiosis between bacteria and methanogenic archaea, mediated by interspecies hydrogen transfer, and that mitochondria may have initially functioned anaerobically rather than as the oxygen-consuming powerhouses we know today.14PubMed. Metabolic symbiosis at the origin of eukaryotes These symbiogenetic models, based on metabolic interactions between archaea and bacteria, have gained increasing support.15PubMed. The Syntrophy hypothesis for the origin of eukaryotes revisited
Secondary and Tertiary Endosymbiosis
The original capture of a cyanobacterium to form a chloroplast happened once, roughly 1.6 billion years ago, giving rise to the lineage that includes green algae, red algae, and land plants. But endosymbiosis did not stop there. In a process called secondary endosymbiosis, entire algae were themselves engulfed by other eukaryotes, producing “complex” plastids wrapped in additional membranes. Green algal plastids were transferred this way to euglenids and chlorarachniophytes. Red algal plastids were similarly transferred, apparently in a single event, to a diverse group that includes diatoms, brown algae, and dinoflagellates.16PubMed Central. The endosymbiotic origin, diversification and fate of plastids Some dinoflagellates went even further, replacing their existing plastids through tertiary endosymbiosis with other algal lineages.17PubMed. A “green” phosphoribulokinase in complex algae with red plastids: evidence for a single secondary endosymbiosis leading to haptophytes, cryptophytes, heterokonts, and dinoflagellates
These layered endosymbioses leave physical calling cards. A plastid surrounded by three or four membranes, rather than two, reveals its history of being captured inside a cell that was itself already inside another cell. In some cases, a vestigial nucleus from the engulfed alga, called a nucleomorph, persists between the membranes. These Russian-nesting-doll arrangements would be inexplicable without endosymbiosis, and they provide a vivid, visible kind of evidence that is independent of DNA sequences.
It Happened More Than Once
A powerful argument for endosymbiosis as a real biological process, rather than an unrepeatable freak event, comes from the fact that it has happened independently more than once. The best-documented example involves the amoeba Paulinella chromatophora, which acquired its own photosynthetic organelle (called a chromatophore) through a completely separate primary endosymbiosis event roughly 90 to 140 million years ago, well over a billion years after the original chloroplast-forming event in the ancestor of plants and algae.18PubMed Central. Paulinella, a model for understanding plastid primary endosymbiosis Paulinella‘s chromatophore has already transferred many of its genes to the host nucleus, mirroring on a compressed timescale the same gene-transfer process that shaped mitochondria and chloroplasts over much longer periods. Studying Paulinella is like having a time machine: it lets researchers watch the early stages of organelle integration in a lineage where the process is far less advanced than in plants.
Hydrogenosomes and Mitosomes
Some anaerobic eukaryotes lack recognizable mitochondria. For a long time, these organisms were thought to represent lineages that diverged before the mitochondrial symbiosis ever took place. That interpretation turned out to be wrong. The organelles in these organisms, called hydrogenosomes (when they produce hydrogen) or mitosomes (when they are even more reduced), are biochemically modified mitochondria.19PubMed. Anaerobic eukaryote evolution: hydrogenosomes as biochemically modified mitochondria? The discovery of DNA inside the hydrogenosomes of anaerobic ciliates put the matter beyond doubt: the sequences placed these organelles firmly within the mitochondrial family tree, even though their biochemistry had been radically overhauled from oxidative phosphorylation to hydrogen-evolving fermentation.20PubMed. The missing link between hydrogenosomes and mitochondria
The implication is sweeping. It now appears that every known eukaryote either has mitochondria or has an organelle that descended from them.21PubMed Central. Mitochondria and hydrogenosomes are two forms of the same fundamental organelle No eukaryotic lineage has been convincingly shown to have never had a mitochondrial endosymbiont. That universality is hard to explain without invoking a single, early endosymbiotic event at the very root of the eukaryotic tree.
Laboratory Recreation of Endosymbiosis
Perhaps the most conceptually satisfying kind of evidence is experimental: can you make endosymbiosis happen in the lab? Researchers have begun to answer that question. In one set of experiments, cyanobacteria were engineered and introduced into yeast cells that had a defective mitochondrial gene. The cyanobacterial endosymbionts partially rescued the growth of the host yeast under selective conditions, demonstrating that a photosynthetic bacterium living inside a eukaryotic cell can provide a metabolic benefit to its host. The experiment also showed that exogenous translocases (molecular channels that shuttle energy molecules across membranes) were needed for the endosymbiont to support the host’s energy needs, consistent with models of how early endosymbionts may have functioned.22Nature Communications. Engineering artificial photosynthetic life-forms through endosymbiosis These experiments do not prove that endosymbiosis happened in exactly this way billions of years ago, but they demonstrate that the basic process is physically and biochemically plausible, not just a theoretical narrative.
The Competing Idea and Why It Lost
Endosymbiosis is not the only hypothesis ever proposed for the origin of organelles. Autogenous models suggest that the nucleus and cytoplasm arose through internal evolutionary changes within a single prokaryotic lineage, essentially by pinching off portions of the cell membrane to create internal compartments, rather than by engulfing another organism.23PubMed. A comparison of autogenous theories for the origin of eukaryotic cells These models are not crazy, and they may still contribute to explaining certain features of the eukaryotic cell, such as the nuclear envelope itself, which does not show clear endosymbiotic ancestry. But for mitochondria and chloroplasts specifically, the autogenous models cannot account for why these organelles carry bacterial DNA, use bacterial ribosomes, respond to bacterial antibiotics, and nest phylogenetically within known bacterial groups. The evidence is simply too convergent from too many independent directions.
Practical Echoes in Medicine
The bacterial ancestry of organelles is not just an academic curiosity. It has medical implications. The malaria parasite Plasmodium contains a relict plastid called the apicoplast, which it inherited from an ancient algal endosymbiont. Because the apicoplast retains bacterial-like biochemical pathways not found in human cells, it is considered an excellent drug target. Several antimalarial compounds work by disrupting apicoplast function, essentially exploiting the organelle’s endosymbiotic heritage.24PubMed Central. Validation of Putative Apicoplast-Targeting Drugs Using a Chemical Supplementation Assay in Cultured Human Malaria Parasites
The flip side of this coin is less welcome. Because mitochondrial ribosomes resemble bacterial ribosomes, antibiotics that target bacterial protein synthesis can cause side effects by also interfering with mitochondrial translation. Aminoglycosides, for example, are associated with hearing loss and kidney damage in part because of their effects on mitochondrial protein production in vulnerable tissues. The endosymbiotic theory predicted this kind of collateral damage long before the molecular details were worked out.
Living Parallels in Modern Endosymbionts
Endosymbiosis is not just something that happened once or twice in the deep past. Obligate bacterial endosymbionts living inside insects today undergo the same kinds of evolutionary changes that transformed the ancestral mitochondrion. These modern endosymbionts show accelerated molecular evolution, severe shifts in DNA base composition, and massive gene loss, mirroring the trajectory that mitochondrial and plastid genomes took over billions of years.25PubMed Central. Intrahost mutational dynamics parallel long-term genome evolution in endosymbionts Watching gene loss and genome shrinkage happen in real time in insect endosymbionts gives researchers a living model for the process that whittled mitochondrial genomes down from the thousands of genes in a free-living alphaproteobacterium to the mere 13 protein-coding genes in human mitochondrial DNA. The parallel is not just conceptual: the same mutational dynamics observed within individual insect hosts over short timescales match the long-term evolutionary patterns seen in ancient organelle genomes.