Skip to content
NatureHQ

Ecologyphenomenon

Endosymbiosis

Your mitochondria have their own genome, divide on their own schedule, and are sensitive to antibiotics. Five separate clues, one answer.

One organism living inside another — and, twice in deep history, staying. The mitochondria in your cells descend from free-living bacteria. They are not bacteria now: most of their genes are in your nucleus.

The claim sounds like a flourish and is one of the better-supported historical statements in biology, because no single observation carries it. Mitochondria have their own DNA, in a small circular genome of the kind bacteria have. They divide by splitting, on their own schedule, rather than being manufactured by the cell. They are wrapped in two membranes, as an engulfed cell would be. Their protein-making machinery is bacterial in type, which is why certain antibiotics that target bacteria also affect mitochondria. And when their genes are placed on a tree of life, they sit among bacteria rather than among their host’s. Each of those has an innocent explanation on its own; together they do not. What the popular version gets wrong is the tense. "Mitochondria used to be bacteria" is true about ancestry and misleading about the present. Over roughly two billion years the great majority of the ancestral genes have moved to the host cell’s nucleus, so a mitochondrion cannot make most of its own parts and cannot live outside the cell at all. It has been reduced, rebuilt and integrated to the point where calling it a captive bacterium describes a distant ancestor rather than the thing itself. Some eukaryotes have taken the reduction further still, keeping organelles that no longer respire. And this is not only ancient history. Endosymbiosis is going on now, at every stage of establishment: aphids carrying bacteria they cannot live without and which cannot live without them, corals carrying photosynthetic cells inside their own, sea slugs retaining chloroplasts from the algae they eat. The old events are the extreme end of a process still visibly underway.

Early coverage · 48% complete · reviewed 2026-09-04

What this page covers

The ancient events concern all eukaryotes. Living endosymbioses are everywhere — in insects, corals, molluscs and protists — and several are recent enough to show the process partway through.

Often confused with: The idea that mitochondria are bacteria now, rather than descended from them; A single event, when plastid endosymbiosis happened repeatedly; Something that only happened long ago, when it is observable in living organisms

Quick facts

Ancestry, not identity
Descended from bacteria; most of their genes now sit in the nucleus
Five independent clues
Own DNA, own division, double membranes, bacterial ribosomes, phylogeny
Still happening
Aphids, corals and sea slugs show the process at different stages
The idea was rejected first
Proposed early in the twentieth century and dismissed for decades

Five clues that only agree for one reason

Each has an innocent explanation. Together they do not.

No single observation established endosymbiosis. Own DNA, own division, double membranes, bacterial-type ribosomes and phylogenetic placement each point the same way for separate reasons.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

The endosymbiotic origin of mitochondria and plastids is supported by independent lines: retention of an organellar genome, division by fission independent of nuclear division, envelope membrane structure, prokaryote-type translation machinery and antibiotic sensitivity, and phylogenetic placement of organellar genes among bacteria.

Who this applies to
Mitochondria and plastids across eukaryotes.
Studied in
Eukaryota
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

Convergence of independent evidence types is what makes this among the better-supported historical claims in biology.

How far it can be extended

Each line of evidence has been examined across eukaryotic diversity.

Caveats

  • Independent lines can share a hidden assumption; here they rest on different kinds of data, which is what makes the convergence informative.
  • Some of the 1967 synthesis’s other proposals were not supported, and the case for organelles does not carry them.

Still unanswered

  • How many times plastid endosymbiosis occurred, and the details of secondary and tertiary events.

Last reviewed 2026-09-04

The evidence (2 studies)
The lines of evidence, and what each would otherwise require
ObservationWhat endosymbiosis explainsThe alternative
Own circular genomeA remnant of the bacterium’s ownA cell keeping DNA in two places for no reason
Divides by fission, independentlyBacterial division retainedA manufactured structure that assembles itself instead
Two envelope membranesThe bacterium’s, plus the host’s engulfing oneA doubled membrane arising de novo
Bacterial-type ribosomesInherited translation machineryConvergence on bacterial machinery inside a eukaryote
Genes group with bacteriaBecause they came from bacteriaExtensive independent similarity by chance

Diagram

Five clues that agree for five separate reasons

How a claim about the deep past gets settled without a witness.

Five clues that agree for five separate reasonsOwn circular genomeDivides by fission, on its own scheduleTwo envelope membranesBacterial-type ribosomesGenes group among bacteriaBacterialancestryEach clue has an innocent explanation on its own. All five holding of one structuredoes not — which is how a claim about the deep past gets settled.
The same explanation in words

Five observations are listed and all converge on one conclusion: own circular genome; divides by fission on its own schedule; two envelope membranes; bacterial-type ribosomes; and genes that group phylogenetically among bacteria. Arrows from each lead to a single box reading bacterial ancestry. A note beneath records that each clue has an innocent explanation on its own, and that all five holding of the same structure does not — which is how a claim about the deep past gets settled.

The right-hand column is the point. Any one row could be argued away; the alternatives are not absurd. What is implausible is all five being separately true of the same structure — and this is a good general model of how historical claims in biology get settled, since nobody was there to watch.

Descended from is not the same as being

The popular phrasing is right about the past and wrong about the present.

Mitochondria descend from free-living bacteria taken up by an ancestral cell. They are not bacteria now: most of their genes moved to the host nucleus, and none can live independently.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

Mitochondria are of bacterial ancestry, phylogenetically placed among the alphaproteobacteria. They retain a small remnant genome, with the large majority of ancestral genes transferred to the host nucleus, and are incapable of independent replication outside the cell.

Who this applies to
All eukaryotes, including lineages whose mitochondria are highly reduced.
Studied in
Eukaryota, Alphaproteobacteria
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

Several independent lines converge: remnant genome, division, membrane structure, antibiotic sensitivity, and phylogenetic placement.

How far it can be extended

Established across eukaryotic diversity by phylogenomic evidence.

Caveats

  • The precise bacterial sister lineage is actively debated, and the nature of the host cell is much less constrained.
  • "Used to be bacteria" is right about ancestry and misleading about the present, which is the whole content of this claim.

Still unanswered

  • What the host cell was at the time of the event, which the bacterial evidence does not constrain.

Last reviewed 2026-09-04

The evidence (2 studies)

The scale of the change is easy to understate. A free-living bacterium of the relevant kind carries a few thousand genes. A human mitochondrion carries thirty-seven. The rest either disappeared or moved into the host nucleus, where they remain — so most mitochondrial proteins are made in the cell and imported. Whatever a mitochondrion is now, it is not a bacterium sheltering inside a cell; it is a component that cannot exist independently and that the cell cannot exist without.

Related

  • Symbiosis

    The category, and what the word actually means

  • Lichen

    Partners that built a shared body without merging

  • Fungi

    A eukaryote lineage that kept the mitochondria and lost the plastid

It is still going on

Living examples at every stage of establishment.

The ancient events are the far end of a process visible in living organisms. Aphids carry bacteria inside specialised cells, passed from mother to offspring, which supply amino acids the insect cannot make and which have lost so many genes that they can no longer live outside their host — a reduction of the same kind as the mitochondrial one, caught midway. Corals hold photosynthetic cells inside their own cells and can lose and reacquire them. Some sea slugs eat algae and keep the chloroplasts working inside their tissues for a while, which is endosymbiosis so temporary it barely qualifies, and shows how the process can begin.

Reading those alongside the ancient events makes the deep history less mysterious. Nothing extraordinary is required — an organism taken up and not digested, doing something useful, retained. What is rare is not the event but its permanence.

Claims about this, checked

Things people have heard, and what the evidence actually supports.

The research behind this page

5 studies, newest first. Each one has a page explaining what it found and what it could not show.

2017Current Biology

The origin and diversification of mitochondria

Mitochondria descend from a bacterial lineage, retain only a small remnant genome, and have transferred most ancestral genes to the host nucleus.

2017Nature

Global warming and recurrent mass bleaching of corals

Bleaching severity tracked accumulated heat exposure.

2006Proceedings of the Royal Society B: Biological Sciences

The role of zooxanthellae in the thermal tolerance of corals: a nugget of hope for coral reefs in an era of climate change

Colonies hosting a more heat-tolerant symbiont type withstood higher temperatures than colonies of the same coral species hosting a less tolerant type, and symbiont composition shifted following transplantation.

1977BioScience

Reef corals: mutualistic symbioses adapted to nutrient-poor environments

Symbionts within coral tissue supply a large share of the animal’s energy from photosynthesis, while the animal supplies nitrogen and carbon dioxide and a lit position near the surface.

1967Journal of Theoretical Biology

On the origin of mitosing cells

The features of mitochondria and chloroplasts are consistent with descent from engulfed bacteria rather than with origin from within the host cell.

Where to go from here

Each of these follows from something on this page — a relationship in the evidence, a claim people ask about, or the next mechanism along.

How complete this page is, and what it is still missing

NatureHQ publishes its own gaps. This page is at 48% completeness against what we would call a finished subject, and was last reviewed on 2026-09-04. It carries 4 claims and answers 2 mapped search questions.

  • no research from the last few years is attached — check for newer work
  • more experiments could be explained in plain English
  • The nature of the host cell at the time of mitochondrial acquisition is barely covered, and is far less settled than the bacterial side.
  • Secondary and tertiary plastid endosymbioses are mentioned only in passing.
  • Insect endosymbionts are used as an example rather than covered in their own right.