The Strange Biology of Mitochondria: How Two Lifeforms Became One Two Billion Years Ago
The standard middle-school summary of mitochondria — "the powerhouse of the cell" — captures one functional fact and misses everything that makes mitochondria the strangest object in biology. Every cell in every plant, animal, fungus, and protist contains hundreds to thousands of organelles that began as free-living bacteria, were engulfed by an ancestor cell roughly two billion years ago, and have been inherited as semi-autonomous passengers ever since. They retain their own DNA, replicate by binary fission on their own schedule, encode some of their own proteins, and are inherited almost exclusively from the mother. Every multicellular eukaryote on Earth is, structurally, a chimera of two lifeforms.
This post covers the endosymbiotic theory, the evidence that won the half-century argument, what mitochondria actually do at the biochemistry level, the strange genetics of inheritance and replication, and the open questions that the standard story does not yet answer.
The endosymbiotic theory
Lynn Margulis published the modern formulation of endosymbiotic theory in 1967 under her then-name Lynn Sagan, in a paper titled "On the Origin of Mitosing Cells." The proposal was that mitochondria are descended from free-living alpha-proteobacteria, that chloroplasts in plant cells are descended from free-living cyanobacteria, and that the eukaryotic cell is the product of an ancient endosymbiosis in which one prokaryotic cell engulfed another and the two became permanently integrated.
The reaction from the cell biology establishment was scathing. The paper was rejected by fifteen journals before publication. The theory was treated as fringe speculation through the 1970s. The shift came with molecular evidence in the 1980s and 1990s: mitochondrial DNA was sequenced and shown to be more similar to alpha-proteobacterial DNA than to nuclear DNA; the inner mitochondrial membrane was shown to share lipid composition with bacterial membranes; mitochondrial ribosomes were shown to be more similar to bacterial ribosomes than to eukaryotic cytoplasmic ribosomes; and the mitochondrial protein-import machinery was shown to derive from bacterial outer-membrane translocases.
By the 2000s the theory was textbook orthodoxy. The 2014 closure of the matter came with the discovery of intermediate forms — bacterial endosymbionts in modern protists that are partway through the process of becoming organelles, with reduced genomes, dependence on host metabolism, and inability to survive outside the host. The intermediate forms are alive in the present and demonstrate the mechanism the theory predicted.
What mitochondria actually do
The schoolroom phrase "powerhouse of the cell" is correct but uninformative. The actual work is the synthesis of ATP via oxidative phosphorylation: the electron transport chain on the inner mitochondrial membrane uses the energy from electron transfer to pump protons across the membrane against their gradient, and ATP synthase uses the proton gradient to drive the chemical synthesis of ATP from ADP and inorganic phosphate. Peter Mitchell's chemiosmotic theory, proposed in 1961 and accepted with the 1978 Nobel, established that energy storage in cells works by proton gradients across membranes rather than direct chemical coupling.
The efficiency of mitochondrial ATP synthesis — about 38 ATP per glucose molecule under aerobic conditions, versus 2 ATP per glucose for the glycolysis-only pathway available to most prokaryotes — is the reason eukaryotic life exists at the size and complexity it does. Without the energy density of oxidative phosphorylation, cells could not afford the metabolic cost of being large enough to be eukaryotic. Nick Lane's argument in The Vital Question is that the endosymbiosis was not an incremental improvement but a step change that enabled cellular complexity that prokaryotes have never matched in the four billion years since.
The strange genetics
Mitochondria retain their own DNA — about 16,500 base pairs in humans, encoding 37 genes — descended from the original bacterial endosymbiont's genome. The vast majority of the original genome (about 1,500 of the original genes) has migrated to the nuclear genome over evolutionary time. The reasons for the partial migration are debated: hypotheses include reduction of risky cytoplasmic DNA replication, regulatory advantages of nuclear coding, and selection against transferring genes whose products must be made on-site for membrane-insertion reasons.
The remaining 37 genes are inherited exclusively through the maternal line in almost all animals. Mitochondria from sperm cells are tagged with ubiquitin during fertilization and degraded by the egg's autophagy machinery. The mechanism is so efficient that paternal mitochondrial DNA is essentially absent in offspring, and the mitochondrial genome forms a strictly clonal lineage descended from one's mother, maternal grandmother, and so on backward into prehistory. This is the foundation of mitochondrial DNA phylogeography, which has reconstructed the female line of human ancestry through Mitochondrial Eve, the most recent common female ancestor of all living humans, who lived approximately 150,000 to 200,000 years ago in Africa.
The 2018 discovery of paternal mitochondrial DNA inheritance in three unrelated human families with mitochondrial disease — the Luo et al PNAS paper — corrected the textbook claim of strict maternal inheritance to "almost always strict maternal inheritance." The mechanism that suppresses paternal inheritance can fail in rare cases, and when it does, the offspring inherit a mixture of maternal and paternal mitochondrial DNA. The molecular basis of the failure is not yet understood.
The replication strangeness
Mitochondria do not replicate in synchrony with the cell cycle. Each mitochondrion replicates on its own schedule, by binary fission, when the cell needs more mitochondria. Mitochondrial fission and fusion are dynamic — the mitochondrial population in a cell is constantly merging and splitting, exchanging contents, and being culled when individual mitochondria become damaged. The cell maintains an active surveillance system, mitophagy, that degrades dysfunctional mitochondria specifically. The discovery of the PINK1/Parkin pathway that marks damaged mitochondria for destruction earned the 2013 Lasker Award for Pickrell and Youle.
The total number of mitochondria varies by cell type — a few hundred per skin cell, several thousand per liver cell, hundreds of thousands per oocyte. The variation tracks the energy demand of the cell type. Cells that are heavily aerobic have many mitochondria; cells that mostly use glycolysis have few. The regulatory mechanism that adjusts mitochondrial population to metabolic demand involves PGC-1α and a network of transcription factors that respond to ATP-to-ADP ratios.
The open questions
The major open questions are: which lineage of bacteria was the original endosymbiont, and which lineage of cells engulfed it? The bacterial side is reasonably well constrained — alpha-proteobacteria, with the closest modern relatives being Rickettsia and Anaplasma — but the host cell side is contested. The Asgard archaea hypothesis, supported by the discovery of Lokiarchaeota in 2015, proposes that the host was an archaeon with eukaryote-like features rather than a true bacterium. The 2020 PNAS paper by Imachi et al cultured the first Asgard archaeon, Prometheoarchaeum syntrophicum, and found it forms physical interactions with bacterial partners that look structurally analogous to the proposed pre-mitochondrial relationship.
The other open question is why endosymbiosis happened only once. Chloroplasts are the result of a separate endosymbiosis with a cyanobacterium, but the mitochondrial endosymbiosis appears to have happened exactly once in the entire history of life on Earth. Every eukaryote has the same mitochondrial ancestor. If endosymbiosis is the natural mechanism for evolutionary innovation, why did it not happen multiple times independently? The leading hypothesis is that the conditions that made it possible — the rise of atmospheric oxygen, the availability of aerobic bacteria, the existence of a host that could survive engulfment without immediate digestion — were a narrow historical window that closed once eukaryotes had occupied the niche.
The deepest implication is that the eukaryotic lifestyle is not a continuous evolutionary improvement on the prokaryotic lifestyle but a one-time fusion event whose probability per unit time across all of Earth's history may be very small. If true, the universe may contain many planets with prokaryotic life and very few with eukaryotic life, and the great filter on the path to complex life may be the endosymbiotic event itself rather than anything that came before or after it. Every cell in your body is, in this sense, evidence of a cosmic accident that may not be repeated.