How Did Mitochondria and Chloroplasts Develop?

Mitochondria and chloroplasts began as free-living bacteria that were taken up by ancient host cells and, over hundreds of millions of years, became permanent residents. This process, called endosymbiosis, is one of the most consequential events in the history of life. A bacterium related to modern alphaproteobacteria gave rise to mitochondria, and a cyanobacterium gave rise to chloroplasts. The details of how these partnerships formed, stabilized, and reshaped the cells that host them are richer and stranger than a simple story of one cell swallowing another.

The Idea That Reshaped Cell Biology

The notion that organelles might have once been independent organisms floated around in biology for decades before anyone took it seriously. In 1967, Lynn Margulis (then publishing as Lynn Sagan) laid out a detailed, evidence-rich case in her paper “On the Origin of Mitosing Cells.” She argued that mitochondria descended from bacteria, that chloroplasts descended from cyanobacteria, and that even the cell’s internal movement apparatus might have symbiotic origins. Not all of her proposals held up, but her promotion of endosymbiosis as a driving force in the evolution of complex cells fundamentally changed how biologists thought about where eukaryotic cells came from.¹ As later historians of science noted, Margulis was not the first to discuss symbiosis as a factor in organelle origins, but no one before her had done it with the same depth or provided as many testable predictions.²

The Host Cell and Its Bacterial Partners

For decades, the identity of the host cell that engulfed the ancestor of mitochondria was murky. Recent breakthroughs in environmental DNA sampling have changed that. Researchers have reconstructed genomes of a group called the Asgard Archaea from deep-sea sediment and other extreme environments. These organisms sit on the tree of life as the closest known relatives of eukaryotes, suggesting that the host cell that gave rise to all complex life was an archaeon, not a bacterium.³ In 2020, Japanese researchers managed to culture a living Asgard archaeon for the first time, a painstaking effort that took over a decade. This organism offers a living window into what the ancestral host might have looked like roughly two billion years ago.⁴

On the mitochondrial side, the ancestor was a member of the alphaproteobacteria, the large group that today includes soil bacteria, plant symbionts, and ocean-dwelling microbes. For years, the leading candidate for the closest living relative was the order Rickettsiales, a group of obligate intracellular parasites. But network-based analyses of key energy-producing proteins suggest that mitochondria are actually more closely related to a broader group of alphaproteobacteria outside Rickettsiales, possibly sharing a common ancestor with orders like Rhizobiales and Rhodospirillales.⁵ The exact branching point remains debated, but the bacterial lineage is clear.

Chloroplasts trace back to cyanobacteria, the photosynthetic microbes responsible for oxygenating Earth’s atmosphere billions of years ago. Recent phylogenetic work points to a deep-branching freshwater cyanobacterium called Gloeomargarita lithophora as the closest cultured relative of the original chloroplast ancestor. Three independent datasets, covering plastid-encoded proteins, ribosomal RNA genes, and genes transferred to the host nucleus, all converge on this result.⁶

What Made the Partnership Stick

One of the hardest questions about endosymbiosis is why a host cell would tolerate an internal bacterium instead of digesting it. One influential idea, the hydrogen hypothesis, offers an ecological answer. It proposes that the original partnership was based on chemistry: an archaeon that depended on hydrogen for its metabolism teamed up with a bacterium that produced hydrogen as a waste product of fermentation. The bacterium could also use oxygen when available, making it metabolically flexible. This model neatly explains why modern mitochondria share ancestry with hydrogenosomes, the hydrogen-producing organelles found in some anaerobic organisms.⁷

A second piece of the puzzle involves energy exchange. Modern-day obligate intracellular bacteria and plastids use a specific type of transporter, an ATP/ADP translocator, to shuttle energy currency between host and symbiont. Structural work on these transporters shows they exist in both plastids and the endosymbionts of certain protists, suggesting they played a key role in cementing the metabolic dependence between host and organelle.⁸ Researchers have even recreated a rough version of this arrangement in the lab by engineering E. coli to express an ATP/ADP translocator and then placing these bacteria inside yeast cells that couldn’t produce their own energy. The yeast-bacterium chimera could grow on carbon sources that the yeast alone could not use.⁹ That experiment does not prove how endosymbiosis originally happened, but it demonstrates that the metabolic logic is sound.

The host cell also had to avoid digesting the newcomer. Many non-photosynthetic protists and animals today can host algal endosymbionts by adapting their internal digestion pathways, recognizing specific partners and housing them within membrane-bound compartments without breaking them down.¹⁰ Ancient host cells likely evolved similar tricks.

Moving Genes to the Nucleus

Once a bacterium is living inside a host cell permanently, something remarkable happens to its genome: most of its genes migrate to the host’s nucleus. Modern mitochondria encode only about 13 proteins in humans (other species vary), down from the thousands a free-living bacterium would carry. Chloroplasts retain somewhat more, usually around 60 to 200 protein-coding genes, but this is still a fraction of what a cyanobacterium needs to survive on its own.

The mechanism behind this transfer involves a process where organelle DNA gets inserted into the nuclear genome during the repair of broken chromosomes. Over evolutionary time, this has resulted in massive relocation of genes from organelle to nucleus. Many of the proteins encoded by those relocated genes are still sent back to the organelle after being made, targeted by special signal sequences. Others have been repurposed to work elsewhere in the cell. The leftover bits of transferred DNA that do not end up coding for anything useful are called NUMTs (from mitochondria) and NUPTs (from plastids), and they litter the nuclear genomes of most eukaryotes. Occasionally, one of these non-coding fragments gets recruited as part of a new gene, adding novel function by a mechanism that has no equivalent in the rest of genetics.¹¹

Why Organelles Still Have Their Own DNA

If genes keep migrating to the nucleus, why haven’t mitochondria and chloroplasts lost all of theirs? The most widely supported explanation is the CoRR hypothesis: co-location for redox regulation of gene expression. The genes that remain in the organelle are overwhelmingly those encoding core components of the energy-producing machinery. These proteins need to be made on-site so their production can respond instantly to changes in the organelle’s energy state. If the cell had to send a signal to the nucleus, wait for transcription, ship the protein back, and then assemble it, the delay could be catastrophic for energy production.¹²

Computational work across many eukaryotic species confirms that the genes most likely to be retained in organelle DNA are those whose protein products sit at the structural core of energy-producing complexes. These proteins tend to be highly hydrophobic (difficult to transport across membranes) and occupy central positions in protein assemblies, meaning their quantities directly control how fast the whole complex gets built.¹³ Cross-species analysis has further validated that hydrophobicity and binding energy centrality are universal predictors of which genes stay in organelle DNA, regardless of the organism.¹⁴

Building New Machinery From Scratch

The transition from free-living bacterium to organelle required new infrastructure. Most mitochondrial proteins are now made in the cell’s cytoplasm and must be imported. The molecular machines that carry out this import show traces of bacterial ancestry in their structure, hinting that some bacterial membrane components were repurposed during the transition.¹⁵ But the core protein translocases that drive mitochondrial import have no clear bacterial counterparts, suggesting they were created from scratch as the organelle became dependent on the host for most of its proteins.¹⁶ The same general pattern holds for the way organelles divide. Free-living bacteria split using a protein called FtsZ, and some chloroplasts still use it. But mitochondrial division in plants relies on dynamin-like proteins, a eukaryotic invention, that pinch the organelle at constriction sites.¹⁷ The host gradually took control of functions the bacterium once managed on its own.

When Chloroplasts Changed Hands

The original chloroplast acquisition, called primary endosymbiosis, happened once. It gave rise to the lineage that includes green algae, red algae, and land plants. But photosynthesis spread far beyond that group through secondary and tertiary endosymbiosis, events where a non-photosynthetic eukaryote swallowed a eukaryote that already had chloroplasts. Green algal plastids were picked up by euglenids and chlorarachniophytes. Red algal plastids were taken up in what appears to have been a single event, producing a sprawling group of organisms called chromalveolates that includes diatoms, brown algae, and the parasites that cause malaria.¹⁸ In some cases, plastids were subsequently lost or replaced. Certain dinoflagellates have swapped their red-derived plastids for ones stolen from other chromalveolates or even from green algae. The result is a tangled history of gain, loss, and replacement that makes tracing chloroplast ancestry in some lineages genuinely difficult.¹⁹

Despite all this complexity, the majority of algal biodiversity on Earth today traces its photosynthetic ability to these eukaryote-eukaryote events rather than to primary endosymbiosis. The organisms most people picture when they think of photosynthesis, green plants and green algae, represent only one branch of a much larger photosynthetic family tree.

Paulinella and Endosymbiosis Caught Mid-Process

Almost all plastids on Earth descend from a single endosymbiotic event that happened roughly 1.6 billion years ago. But there is one stunning exception. The amoeba genus Paulinella independently acquired its own photosynthetic compartment, called a chromatophore, from a cyanobacterium only about 90 to 140 million years ago.²⁰ Because this event is so much more recent, Paulinella offers researchers a window into an earlier stage of organelle integration. Its chromatophore still retains far more genes than a typical chloroplast, but it has already lost genes essential for independent survival, and some of its functions now depend on proteins encoded in the host nucleus.²¹ Comparing genome evolution across multiple Paulinella species reveals the same patterns of gene loss and host dependence that shaped the ancient Archaeplastida plastid, just on a vastly compressed timescale.²²

When Mitochondria Lost Their Oxygen Habit

Not every descendant of the original mitochondrial endosymbiont looks like a textbook mitochondrion. In organisms that live in oxygen-poor environments, the organelle has been remodeled, sometimes beyond recognition. Hydrogenosomes, found in certain anaerobic protists, produce energy and release hydrogen gas instead of consuming oxygen. For years they were thought to be unrelated to mitochondria, but molecular evidence now firmly places them as modified mitochondria that lost their oxygen-dependent metabolism.²³ Even more stripped-down versions, called mitosomes, exist in some parasites and retain almost no metabolic function, serving mainly in iron-sulfur cluster assembly. The spectrum from full aerobic mitochondria to mitosomes demonstrates that the organelle’s evolution did not stop after the initial partnership. It has continued to be reshaped by the metabolic needs of whatever organism houses it.

Timing the Events

Pinning absolute dates to events that happened over a billion years ago is inherently uncertain, but cross-calibrated molecular clock analyses provide rough estimates. One such study, using duplicated energy-producing proteins as molecular markers, placed the mitochondrial endosymbiosis at roughly 1.2 billion years ago and the primary plastid endosymbiosis at about 900 million years ago.²⁴ These dates are younger than some other estimates and remain debated, but the relative order is broadly accepted: mitochondria came first, and chloroplasts followed later in a lineage that already had mitochondria. No known eukaryote has chloroplasts but lacks mitochondria (or their remnants), which is consistent with this sequence.

Why You Inherit Mitochondria From Only One Parent

In most animals and plants, mitochondrial DNA passes exclusively from mother to offspring. This uniparental inheritance is not a random quirk. Mixing mitochondria from two parents creates a situation called heteroplasmy, where a cell contains two genetically distinct populations of mitochondria. Because these populations can compete with each other for replication regardless of whether they are good for the host, heteroplasmy opens the door to selfish mitochondrial mutations that replicate fast but function poorly.²⁵ Modeling work shows that uniparental inheritance, once it arises, reliably outcompetes biparental inheritance in a population, independent of mutation rate or the number of mitochondria per cell.²⁶ In practical terms, this means that the ancient conflict between host and organelle genome left a lasting mark on how genetic material is transmitted in nearly all complex life.

Stolen Chloroplasts in the Modern Ocean

Endosymbiosis is not only an ancient phenomenon. Some animals alive today run a version of it in miniature. Sacoglossan sea slugs, sometimes called “solar-powered sea slugs,” feed on algae and retain the chloroplasts from their food in their own gut cells. The slug Plakobranchus ocellatus, for instance, keeps stolen chloroplasts functional for several months, potentially supplementing its nutrition through photosynthesis.²⁷ This phenomenon, called kleptoplasty, is not true endosymbiosis because the chloroplasts cannot be passed to offspring and eventually degrade. But it shows that the basic biological trick of capturing and exploiting another organism’s photosynthetic machinery is not locked away in deep evolutionary history. It keeps happening, in various forms, wherever the ecological incentive exists.