Mitochondria: The Bacteria That Never Left

Every cell you have is running on a partnership that began about 1.5 billion years ago, when one cell swallowed another and the smaller one never left. What it became is the mitochondrion, and it is still carrying its own set of instructions.

This article looks at how that relationship shows up in the evidence. It covers how the energy is made, and why the ATP figure in most textbooks has been wrong for decades.

How Many, and Where

Mitochondria are usually drawn as bean-shaped, but they are not fixed like that. They fuse and divide continuously, and they move along the cytoskeleton. Boguszewska and colleagues describe them as a dynamic network rather than a bag of separate organelles.

The number per cell varies enormously. It runs from a handful to several thousand, and it tracks demand. Liver and heart muscle carry large numbers because they work continuously. A cell that mostly sits still needs fewer.

Mitochondria in a cell. Their number tracks how much energy the cell spends, which is why liver and heart muscle carry so many. Ryan Kissinger for NIAID, public domain, via Wikimedia Commons

Structurally there are two membranes. The outer one is porous, and porin channels may account for more than a third of its surface area. The inner one is the opposite. It is close to impermeable, and that is the point: it holds the gradient that the energy comes from.

The parts of a mitochondrion. The inner membrane folds inward into cristae, and the matrix between them is where the citric acid cycle runs. OpenStax, CC BY 4.0, via Wikimedia Commons

Folding is not decoration. Every fold increases the membrane area available to hold respiratory chain complexes and ATP synthases. That is why the inner membrane looks like a set of shelves rather than a smooth bag.

How We Know It Was a Bacterium

The strongest support for the endosymbiotic story is not a diagram. It is that mitochondria arrived carrying their own genome, and that the genome still works.

A tomogram of a single mitochondrion. The folded inner membrane, the cristae, is where the respiratory chain sits. Terrence G. Frey, CC BY 3.0, via Wikimedia Commons

A Genome, and a Very Small One

Boguszewska and colleagues report that human mitochondrial DNA carries 37 genes. Thirteen of them encode protein. The rest are ribosomal RNA and transfer RNA, which the mitochondrion needs to read those thirteen.

Now set that against the proteins a mitochondrion actually needs. Boguszewska and colleagues put the figure at about 1,100, and all but thirteen of them are encoded in the nucleus, made in the cytosol, then imported. The cell and the organelle have been sharing the work for so long that only a sliver of the original bacterial genome is left.

That sharing happened by a named process: endosymbiotic gene transfer. Genes for biosynthesis and other core functions moved to the host over time. That is why the bacterial ancestor and the nucleus are hard to separate from each other.

The Ancestor, More Precisely

The symbiont was not a generic bacterium. Boguszewska and colleagues place it among the alpha-proteobacteria, related to the Rickettsiales lineage. The host, they add, was most probably an archaeon, of the Asgard superphylum.

Martin and colleagues reach the same conclusion from a different direction. They argue that current versions of the theory have the host as an archaeon rather than another eukaryote.

That is a real correction to the older textbook version, in which one cell simply engulfed another. The host was not a fully formed eukaryotic cell. It was something simpler that ended up with more complexity afterwards.

The Argument That Held Up

The clean version says an aerobic bacterium was swallowed because oxygen was available and breathing was useful. Martin and colleagues note that more than 20 different versions of endosymbiotic theory have been published. The classical ones struggled with a specific problem.

If the organelle was always an oxygen breather, then the anaerobic versions of it are awkward. Those exist. They are called hydrogenosomes, and they make hydrogen instead of using oxygen. Classic accounts tended to explain them away as not being mitochondria at all.

Giardia intestinalis. It lives without oxygen and has no ordinary mitochondria, which is the strongest evidence that the organelle began as a free-living bacterium. Servier Medical Art, CC BY 2.0, via Wikimedia Commons

A Prediction That Held

The hydrogen hypothesis, as Martin and colleagues set it out, predicts three things. Every eukaryote has mitochondria or has secondarily lost them. The host was an archaeon. And aerobic and anaerobic forms should interleave across the eukaryotic tree, rather than sitting in separate groups.

They report that all three predictions have since been borne out. That is unusual here, and it is worth stating plainly. This is a case where a sharp prediction turned out to be right, not a story that kept being adjusted.

The energetic reason follows the same way. Martin and colleagues argue that only cells that had mitochondria had the bioenergetic means to reach eukaryotic complexity. They note there are no true intermediates between prokaryote and eukaryote, which is what you would expect if the transition was fast and once.

One historical correction is worth making, because it is often told wrongly. Martin and colleagues point out that Altmann is not credited with proposing the symbiotic origin of mitochondria. In his 1890 book he did not mention mitochondria, or their older name chondriosomes, and did not propose endosymbiosis.

How the Energy Is Made

Cellular respiration is usually taught as three stages. Keeping to three is fine, but the labels hide where each one actually happens and what it costs.

Glycolysis, in the Cytosol

One glucose molecule is split into two pyruvate, with a small amount of ATP made along the way. This happens outside the mitochondrion, in the cytosol, and it does not need oxygen.

That last point matters more than it looks. Glycolysis is the reason a cell can keep working when oxygen runs short, which is exactly what happens in exercising muscle.

The Citric Acid Cycle, in the Matrix

Pyruvate is broken down further, releasing carbon dioxide and loading electrons onto carriers. This runs in the matrix, the fluid inside the inner membrane.

Most of the energy is not stored as ATP at this stage. It goes onto carrier molecules, which is why the cycle is often described as producing very little ATP directly.

It is also wrong to say this cycle requires oxygen. It consumes no oxygen itself. It depends on oxygen only indirectly. The electron transport chain has to keep regenerating the oxidised carrier the cycle hands electrons to. Cut off the chain, and the cycle stalls on its own backlog.

The Chain, in the Inner Membrane

The carriers deliver their electrons to the respiratory chain, a series of complexes set into the inner membrane. As electrons move along it, protons get pumped out of the matrix into the intermembrane space.

Water forms at the end, as oxygen takes up the electrons and the protons that have come back through the motor. That is the only step in the whole pathway where molecular oxygen is consumed.

So the sequence is not "oxygen lets the cell make energy". It is "the chain moves electrons, and pumping protons needs somewhere to put them, and oxygen is what lets the used-up carriers start again."

The Motor That Does the Work

The protons now fall back through the inner membrane, and the thing they fall through is a rotary machine.

An ATP synthase. It is a rotary motor: protons flow through it down their gradient and the shaft turns, like a water wheel. Atpsynthesizer, CC BY-SA 3.0, via Wikimedia Commons

One Residue Does the Work

Klusch and colleagues solved the structure of an intact mitochondrial ATP synthase and found something surprisingly simple at the center of it.

The membrane part, called Fo, contains a ring of identical subunits. Each one has a single glutamate residue, and that residue is the proton-binding site. The lumen-facing channel protonates it. As the ring turns, that glutamate reaches the matrix-facing channel and releases the proton.

Meanwhile an arginine sits between the two channels and prevents protons leaking straight through. Klusch and colleagues argue the steep potential across that sub-nanometre gap is what produces net directional rotation. The gradient does the turning, and ATP comes out the other end.

Why the Ring Size Matters

The c-ring is not the same size in everything. Klusch and colleagues report 8 subunits in mammalian mitochondria and 10 in yeasts and in the alga they worked on. Across organisms the range runs from 8 to 15.

Each subunit carries one proton across per turn. A bigger ring means more protons per turn, and that single structural fact is part of why one cell type gets more ATP out of the same glucose than another.

The Number in the Textbooks Is Outdated

Here is the part most likely to surprise you, because the figure has been printed for fifty years.

Amthor traces the history. By the 1970s, standard biochemistry textbooks gave 36, or sometimes 38, ATP per glucose as the total from aerobic respiration. That was already under downward revision before 1980, as the mechanism of oxidative phosphorylation became better understood.

Later estimates kept falling: 29.5 or 31, then 32, then 30 or 32 again. Refining the count for the real structure of ATP synthases moved it further.

Source and year	ATP per hexose
Lehninger and Stryer, 1970s	36 or 38
Hinkle et al., 1991	29.5 or 31
Taiz and Zeiger, 1991	32
Stryer, 1995	30 or 32
Amthor model, 2023, potential	about 27.5

Estimated potential yield of aerobic respiration, from the figures Amthor (2023) reviews.

Amthor's own model, built for plant respiration, comes out near 27.5 ATP per hexose when respiring sucrose. And he is careful to say that is the potential.

Real yields sit below it. Energy-conserving steps in the chain get bypassed even in healthy, unstressed cells. If a quarter of oxygen uptake goes through an alternative oxidase, which Amthor says is a typical fraction, the yield falls about 15 percent below potential.

So there is no single correct number, and Amthor says so directly. A potential ATP yield is unlikely to be a universal biological constant. The transport pathways differ, the shuttles that move molecules across the membrane differ, and so does the ATP synthase.

What you can say is bounded. Glycolysis yields 2 ATP directly, and the cycle yields a small amount more. The overwhelming majority comes from the chain, which is where the oxygen is needed.

Who Figured This Out

The Nobel Foundation records that Albert Szent-Gyorgyi had already identified several important reactions in metabolism. In 1937 Hans Krebs was able to present a complete picture of a cyclical part of it.

Krebs shared the 1953 Nobel Prize in Physiology or Medicine. Half of it went to Fritz Lipmann, recognised for discovering coenzyme A and its importance in intermediary metabolism.

That pairing is the useful part. The cycle that became metabolism's central loop was assembled in the same year, by the same kind of work, as the coenzyme that carries its products into the next step. Neither made sense on its own.

A single cell of Chlamydomonas under the electron microscope. Even in a green alga, most of the organelles doing the work are mitochondria. Dartmouth Electron Microscope Facility, public domain, via Wikimedia Commons

Key Terms in Plain English

These are the words in this article that would send you to a dictionary. Each one is given here the way it is actually used above.

Term	What it means here
ATP	Adenosine triphosphate, the molecule cells use to pay for work.
Matrix	The fluid inside a mitochondrion's inner membrane.
Cristae	Folds of the inner membrane that hold the respiratory chain.
Glycolysis	The splitting of glucose into pyruvate, done in the cytosol.
Citric acid cycle	The cyclic pathway that strips electrons off pyruvate in the matrix.
Electron transport chain	The membrane complexes that move electrons and pump protons.
Oxidative phosphorylation	Making ATP from a proton gradient across a membrane.
ATP synthase	The rotary motor that turns proton flow into ATP.
Proton gradient	The imbalance of hydrogen ions across a membrane that stores energy.
Endosymbiosis	One organism living permanently inside another, to both their benefit.
Hydrogenosome	An anaerobic form of the mitochondrion that produces hydrogen instead of using oxygen.

The words in this article that would otherwise send you to a dictionary.

Key Takeaways

• Mitochondria carry their own DNA: 37 human genes, 13 encoding protein, against roughly 1,100 mitochondrial proteins that have to be made from nuclear DNA.
• The ancestor was an alpha-proteobacterium and the host was an archaeon, not another eukaryote.
• Anaerobic forms exist, and the hydrogen hypothesis predicts them. That prediction has held up.
• The citric acid cycle does not consume oxygen directly. It stalls without oxygen only because the chain cannot regenerate its carriers.
• ATP synthase is a rotary motor. One glutamate residue per ring subunit picks up and releases each proton.
• The 36 ATP figure is obsolete. Current estimates are nearer 27 to 30, and lower still once real-world bypasses are counted.
• There is no universal ATP yield, because the machinery differs between organisms.

Frequently Asked Questions

Do mitochondria produce all of a cell's ATP?

No. Glycolysis happens in the cytosol and makes some. The mitochondrion handles the rest, including the large majority that comes from the chain.

Why are there so many in heart muscle?

Because cardiac muscle does not stop. Boguszewska and colleagues note that cells needing more energy carry more mitochondria, and liver and heart are the standard examples.

If Giardia has no mitochondria, does that disprove endosymbiosis?

No, and Martin and colleagues treat it as supporting evidence. The hydrogen hypothesis predicts that some lineages lost the organelle afterwards, which is what Giardia looks like.

Do all eukaryotes have mitochondria?

On the hydrogen hypothesis, all of them either have them or have lost them. Martin and colleagues report that prediction as borne out.

Where does the 36 ATP figure come from?

It was a reasonable estimate in the 1970s, based on assumptions about stoichiometry that later work revised. Amthor traces the corrections and finds the textbooks still citing the old figure.

Is heat a waste product?

In a plant cell, partly. Amthor notes that respiration can produce substantial useful heat, and that some plants do deliberately generate it.

Key references
• Martin, Garg and Zimorski 2015: endosymbiotic theories
• Amthor 2023: ATP yield of plant respiration
• Boguszewska et al. 2020: mitochondria and bacteria
• Nobel Prize: Hans Krebs, facts
• Klusch et al. 2017: the ATP synthase motor

Images: Wikimedia Commons, with the author and license named in each caption. Figures and dates follow the five sources above.