Mitochondria, Part 1: The Cell Inside the Cell
Written By Kevin Wallace.
What they are, where they came from, and what actually changes with age.
1. Where the curiosity started
When I first learned how a cell is built, I thought I was learning a parts list. Nucleus, ribosomes, membranes, mitochondria, each with a job. That list was only the doorway. The curiosity stayed through veterinary school, through years as an internist, and into retirement on the farm.
The more I understood what each organelle does, the more I wanted to know how it fails, and what actually keeps it working. No organelle rewards that question more than the mitochondrion. This article is the groundwork: what mitochondria are, where they came from, what they do, how many we carry, and what the human evidence says about age. The articles that follow take the levers one at a time.
2. A bacterium that moved in and stayed
Mitochondria began as free-living bacteria. The partnership is not in doubt. The family tree still is.
For years the standard teaching was that mitochondria sit inside the Alphaproteobacteria, the same broad group that includes the rickettsias. A 2021 molecular-clock study, using eukaryotic fossils to date that bacterial tree, put the origin of the Alphaproteobacteria at about 1.9 billion years ago and the last common ancestor of all mitochondria at about 1.55 billion years ago. Those are estimates, with wide uncertainty, not a stopwatch reading. (1)
Newer gene-rich trees have pushed the closest living relatives around. One 2018 analysis placed the mitochondrial ancestor outside the alphaproteobacteria that have been sampled. A 2022 reanalysis, built to handle the compositional bias that misleads these deep trees, favored mitochondria as a sister group to the known Alphaproteobacteria rather than nested inside them. (2) The honest line is this: mitochondria came from a proteobacterium related to the Alphaproteobacteria, more than a billion years ago. Which living bacterial group is the nearest cousin is still being argued.
The evidence is still sitting in your own cells. Mitochondria carry their own small circle of DNA, separate from the chromosomes in the nucleus. In people it is 16,569 base pairs and holds 37 genes: 13 protein-coding genes for parts of the energy machinery, 22 transfer RNAs, and 2 ribosomal RNAs. (3) Over the eons most of the original bacterial genes moved to the nucleus. A mitochondrion today cannot build itself. It depends on proteins shipped in from the rest of the cell.
Mitochondrial DNA also follows its own inheritance rule. Both egg and sperm contain mitochondria, but the sperm’s are almost always destroyed after fertilization. Nearly all of us carry our mother’s mitochondrial DNA, down the maternal line. (4) Rare paternal leakage has been reported. It is the exception, not a second rule.
One more term worth having before the later articles. A cell can hold a mixture of normal and altered mitochondrial DNA. That mixture is called heteroplasmy. Disease and aging questions often turn on the fraction, not on a simple yes or no.
3. Far more than a power plant
The famous job is real. Food is broken down to electron carriers. Those electrons are passed along a chain of protein complexes in the inner membrane. The energy of that fall is used to pump protons out. Protons flowing back in through ATP synthase turn a rotor and rebuild ATP from ADP. We store very little ATP, so it is recycled constantly.
A teaching estimate says a person turns over roughly their own body weight in ATP each day. It comes from oxygen use, not from weighing ATP. Assume a few hundred liters of oxygen a day, and the arithmetic lands near 50 to 75 kilograms of ATP recycled. (5) It is an order-of-magnitude picture. A rest day is lower. A hard work day is higher. The point that survives the arithmetic is the recycling, not the exact number of kilograms.
Calling mitochondria the powerhouse still undersells them.
They help decide when a cell should die. In the intrinsic pathway of apoptosis, cytochrome c released from the mitochondrion is one of the signals that starts the self-destruct program.
They take up and release calcium, and so shape calcium signals inside the cell.
They start steroid hormones, but only in the tissues that make them. Cortisol, the sex steroids, and aldosterone begin as cholesterol converted to pregnenolone on the inner mitochondrial membrane. A muscle mitochondrion is not running that reaction.
They build heme and iron-sulfur clusters. Heme carries oxygen in hemoglobin. Iron-sulfur clusters are essential parts of many enzymes, including parts of the energy chain itself. The first and last steps of heme synthesis are mitochondrial; the middle steps are not.
They can make heat. In brown fat, the protein UCP1 lets protons leak back without making ATP, so the energy leaves as heat. Active brown fat is present in many adults, more readily found in younger, leaner people and in the cold. It is not a furnace everyone can count on. (6)
They send signals. Small, controlled amounts of oxidants such as hydrogen peroxide help drive responses to exercise, immunity, and oxygen level. The same molecules, in excess, damage proteins, lipids, and DNA. Signal and injury are a dose, not two different substances.
They raise an alarm. When mitochondrial DNA escapes into the cell, the immune system can read it as a danger signal, a reminder of the organelle’s bacterial past.
4. How many, and where
Mitochondria are not spread evenly. Cells that work hardest carry the most. Cells with no need carry none. Counting them is harder than a textbook bean implies. In a living cell they fuse and split, so “one mitochondrion” is often a moment in a network, not a fixed object. Volume and DNA copy number are the cleaner measurements.
| Cell or tissue | What was measured | Source, and the limit |
|---|---|---|
| Red blood cell | No mitochondria, and no mitochondrial DNA | Mammals. The DNA is lost as the cell matures. |
| Sperm | About 5 copies of mitochondrial DNA; roughly 50 to 75 mitochondria in the midpiece | Human. The copy-number figure is from a small clinical series, not a census. (7) |
| Liver cell | On the order of 1,000 to 2,000 mitochondrial profiles; about 17 to 20 percent of cell volume | Rat and mouse stereology. Profiles are not the same as separate organelles. (8) |
| Heart muscle | About 32 percent of cell volume in the mouse; about 23 percent in human heart muscle | Not half the cell. Smaller mammals run higher. (9) |
| Egg cell | Mean 193,000 mitochondrial DNA copies; range 20,000 to 598,000 | Human, but these were unfertilized eggs after IVF, not a random sample of healthy eggs. (10) |
The brain makes the demand obvious another way. It is about 2 percent of body weight and accounts for about 20 percent of the oxygen the body uses at rest. (11) That is why the organs that suffer first when energy runs short are the hungriest ones: heart, brain, and muscle.
5. A living network, not a box of batteries
Textbook drawings show mitochondria as separate beans. In a living cell they behave more like a network that is constantly remodeled. Four processes keep it usable.
Fusion. Mitochondria join, share contents, and dilute damage.
Fission. They split, which lets the cell isolate a bad section.
Mitophagy. Damaged pieces are tagged and broken down for recycling. One well-studied route runs through the proteins PINK1 and Parkin. It is not the only route. Other receptors matter in hypoxia, development, and ordinary turnover, and how dominant PINK1-Parkin is in resting human muscle is still unsettled.
Biogenesis. The cell builds new mitochondrial machinery when demand rises. Exercise is the best-documented human trigger.
A protein called PGC-1 alpha is often described as the master switch for building new mitochondria. That holds well in cells and in animal studies. In human muscle the picture is less tidy. A 2018 review found that the nuclear and mitochondrial genes involved do not always switch on together after exercise, the way the simple model predicts. (12) The practical point stands either way. Training reliably builds mitochondrial respiratory capacity in people. The exact wiring is still being worked out.
Most of what helps or harms mitochondria, as the later articles will show, works by pushing one of these four levers.
6. What changes with age, and how much is reversible
Mitochondrial dysfunction is one of twelve hallmarks of aging in the 2023 framework published in Cell. (13) That paper is a synthesis, not a new measurement. In people, the measurable picture looks like this.
Less energy capacity in muscle. In 146 healthy adults aged 18 to 89, muscle mitochondrial DNA, mitochondrial transcripts, and ATP production all declined with age, tracking lower fitness and poorer glucose handling. ATP production fell on the order of 8 percent per decade. Oxidative damage to DNA was higher in older muscle. (14)
A possible drop in NAD, but not a settled one. NAD carries electrons in energy metabolism. It is often said to fall steadily with age. A 2021 review of the human data called that a confounded consensus: modest drops in some brain and liver studies, little or none in muscle or blood in others, and many of the studies small. (15) Later clinical reviews still describe the human tissue evidence as sparse.
Here is the hopeful part, and the limit on it. Studies like the 2005 muscle survey compare older and younger people, but they cannot fully separate age from inactivity, because most of us move less as we get older. When researchers put people on a training program, the gap narrows fast.
In a 2017 Mayo Clinic study in Cell Metabolism, 12 weeks of high-intensity interval cycling raised maximal mitochondrial respiration 49 percent in adults aged 18 to 30, and 69 percent in adults aged 65 to 80. (16) The older group improved more, not less. Resistance training in the same study built strength and muscle but did not raise mitochondrial respiration. The groups were small, on the order of a dozen people in each age-and-training cell. “Reversed aging” is a headline the paper does not earn. What it does earn is this: a large share of the age gap in muscle mitochondrial respiration is trainable, in both directions.
Some changes are stickier. Deleted mitochondrial DNA accumulates in older muscle and is not wiped out by a 12-week program. Use it or lose it is the right working rule. It is not a promise that every mark of age comes off.
7. Where this series goes next
This article was the groundwork. The ones I have planned next, each held to human studies and honest about thin evidence, are these.
What damages mitochondria, and what actually helps.
Supplements, trial by trial.
The research frontier: what is promising, what is marketing, and what is still a mouse.
Food, movement, and sleep are not missing. They sit inside the first of those, and any one of them can become its own article if the evidence deserves the room.
The curiosity that started with a diagram of a cell has never run out. I hope some of it rubs off.
8. Where we land
Mitochondria are former bacteria, still carrying a scrap of their own DNA, mostly from your mother. They make ATP, and they also handle calcium, the start of steroid synthesis, heme, heat, cell death, and danger signals. They are a network, not a stack of batteries. Muscle mitochondrial respiration declines with age and inactivity, and a few months of hard interval work can take a large part of that gap back. The exact bacterial cousin, the NAD story, and the master-switch story are less settled than the popular versions. That is the ground this series will stand on.
9. Sources
Wang and Luo, Dating Alphaproteobacteria evolution with eukaryotic fossils, Nature Communications, 2021. https://www.nature.com/articles/s41467-021-23645-4
Martijn et al., Nature, 2018, https://www.nature.com/articles/s41586-018-0059-5 ; Muñoz-Gómez et al., Nature Ecology & Evolution, 2022, https://pubmed.ncbi.nlm.nih.gov/35027725/
MedlinePlus Genetics, Mitochondrial DNA, NIH. https://medlineplus.gov/genetics/chromosome/mitochondrial-dna/
NIGMS Biomedical Beat, The Maternal Magic of Mitochondria, 2020. https://biobeat.nigms.nih.gov/2020/05/the-maternal-magic-of-mitochondria
Buono and Kolkhorst, Estimating ATP resynthesis during a marathon run, Advances in Physiology Education, 2001. https://journals.physiology.org/doi/full/10.1152/advances.2001.25.2.70
Cypess et al., Identification and importance of brown adipose tissue in adult humans, New England Journal of Medicine, 2009. https://www.nejm.org/doi/full/10.1056/NEJMoa0810780
Gabriel et al., Journal of Assisted Reproduction and Genetics, 2012, via Shokolenko and Alexeyev, Biochimica et Biophysica Acta, 2015. Sperm midpiece count: Hecht et al. and later reviews in Human Reproduction Update, 2026.
Weibel et al., Journal of Cell Biology, 1969; BioNumbers ID 105783 and 100929.
Schaper, Meis, and Stammler, Circulation Research, 1985. https://pubmed.ncbi.nlm.nih.gov/3882260/ ; Hoppeler et al., Respiration Physiology, 1984.
Reynier et al., Molecular Human Reproduction, 2001. https://pubmed.ncbi.nlm.nih.gov/11331664/
Raichle and Gusnard, Appraising the brain’s energy budget, PNAS, 2002. https://www.pnas.org/doi/10.1073/pnas.172399499
Islam, Edgett, and Gurd, Metabolism, 2018. https://pubmed.ncbi.nlm.nih.gov/29126696/
López-Otín, Blasco, Partridge, Serrano, and Kroemer, Hallmarks of aging: An expanding universe, Cell, 2023. https://doi.org/10.1016/j.cell.2022.11.001
Short et al., PNAS, 2005. https://www.pnas.org/doi/10.1073/pnas.0501559102
Peluso, Damgaard, Mori, and Treebak, Age-dependent decline of NAD+ — universal truth or confounded consensus? Nutrients, 2021. https://www.mdpi.com/2072-6643/14/1/101
Robinson et al., Cell Metabolism, 2017. https://www.cell.com/cell-metabolism/fulltext/S1550-4131(17)30099-2
