For most of the history of molecular biology, the mitochondrial genome was described as encoding 13 proteins. That number is in the textbooks, and it is not wrong. It is just incomplete. Beginning in 2001, researchers started finding short open reading frames hidden inside mitochondrial ribosomal RNA genes that appear to produce small signalling peptides. Humanin came first, then MOTS-c, then the SHLPs.
The Genome That Was Supposed to Encode 13 Proteins
Human mitochondrial DNA is a small circular molecule of 16,569 base pairs, sequenced in full in 1981. It carries 37 genes: 22 encoding transfer RNAs, two encoding ribosomal RNAs, and 13 encoding proteins. All 13 of those proteins are subunits of the oxidative phosphorylation machinery, the enzyme complexes in the inner mitochondrial membrane that generate ATP.
The picture this produced was tidy. Over evolutionary time the mitochondrion transferred almost all of its ancestral bacterial genome to the nucleus, keeping only what it needed to build its own power plant on site. A 16.5 kilobase genome with no obvious spare room was not where anyone expected to find signalling molecules.
There was a technical obstacle too. The mitochondrial genetic code is not identical to the cytoplasmic one. In human mitochondria, UGA specifies tryptophan rather than a stop, and AGA and AGG function as stop codons rather than arginine. Reading frames that look meaningless under the standard code can be meaningful under the mitochondrial one, and the reverse.
Humanin: The Peptide Nobody Was Looking For
Humanin was not found by scanning the mitochondrial genome. It turned up in a screen designed to answer a completely different question.
In 2001, Hashimoto and colleagues, working in Ikuo Nishimoto's laboratory, ran a functional screen on a cDNA library built from the occipital lobe of a patient with Alzheimer's disease, a brain region relatively spared in that condition. They were hunting for clones that protected neuronal cells from death induced by familial Alzheimer's disease genes and by amyloid beta. The work appeared in the Proceedings of the National Academy of Sciences.
One clone protected the cells. When the group worked out what it encoded, the answer was a 24 amino acid peptide, and the sequence traced back to a short open reading frame sitting inside the mitochondrial 16S ribosomal RNA gene. They named it humanin.
The result was strange on several counts. The coding sequence was nested inside a gene whose job was to make ribosomal RNA, not protein. The product was very short. And because it arrived through a screen for biological activity, the peptide came with a phenotype attached before anyone had a framework to explain why it should exist.
MOTS-c and the Second Reading Frame
Humanin sat as an isolated curiosity for more than a decade. If it was a one-off, it could be filed away as a quirk of one rRNA gene. If it was the first member of a family, the mitochondrial genome needed re-reading.
In 2015, Pinchas Cohen's group at the University of Southern California reported a second one. Working from the premise that other short open reading frames might be hiding in the mitochondrial rRNA genes, they searched the 12S ribosomal RNA gene and found one predicted to encode a 16 amino acid peptide. They called it MOTS-c, for mitochondrial open reading frame of the 12S rRNA type-c, and published in Cell Metabolism.
MOTS-c mattered less as a molecule than as a proof of concept. Humanin was no longer a singular oddity, and there was now a method.
The SHLPs: From Two Molecules to a Class
The Cohen group applied that method systematically. In work published in the journal Aging in 2016, they searched the same 16S rRNA region that had yielded humanin for further short open reading frames and reported six more peptides. They named them small humanin-like peptides, SHLP1 through SHLP6, reported in the range of roughly 20 to 40 amino acids.
With eight candidate molecules coming out of two ribosomal RNA genes, this stopped being a list of exceptions and started being a class. The umbrella term now in use is mitochondrial-derived peptides, often shortened to MDPs.
The label flattens real differences, though. Humanin and MOTS-c have by far the most literature behind them. Several SHLPs remain thinly characterised, and the field is still working through which predicted reading frames are genuinely translated in living cells rather than merely plausible on paper.
What Retrograde Signalling Actually Means
The reason cell biologists find this class interesting is not really the peptides. It is the direction of information flow they imply.
Communication between nucleus and mitochondria runs both ways. Anterograde signalling runs from nucleus to mitochondria: the nucleus encodes most mitochondrial proteins and governs mitochondrial biogenesis, so that direction was never controversial. Retrograde signalling runs the other way, from mitochondria back to the nucleus, changing nuclear gene expression in response to conditions inside the organelle.
Retrograde signalling is not a new idea. It was worked out in budding yeast, where a pathway known as RTG, built around the proteins Rtg1, Rtg2 and Rtg3, alters nuclear transcription when mitochondrial function is compromised. What had been missing was a clean example of the mitochondrial genome itself encoding a diffusible messenger for that conversation.
A 2018 paper in Cell Metabolism reported that MOTS-c translocates to the nucleus under metabolic stress and is associated with regulation of nuclear gene expression there. If that model generalises, the mitochondrial genome is not only a parts list for the respiratory chain.
How This Class Differs From Conventional Peptides
| Conventional signalling peptides | Mitochondrial-derived peptides | |
|---|---|---|
| Encoded by | Nuclear DNA | Mitochondrial DNA |
| Genetic code | Standard code | Mitochondrial code differs at several codons |
| Gene context | Dedicated protein-coding genes | Short reading frames nested inside rRNA genes |
| Route to discovery | Sequence annotation and cloning | Functional screening, then targeted searching |
| Depth of literature | Decades for many | Roughly 25 years at most, uneven across members |
The nested-gene point is what surprised people. Genes inside other genes are not unknown in biology, but finding them in a genome this small and this long assumed to be fully annotated was not an expected result.
Where the Research Currently Stands
This is a young field, and its limits are worth stating plainly.
- Much of the published work sits in cell culture and rodent models. Human data is sparse.
- Measuring these peptides reliably in human blood is technically difficult, and assay differences complicate comparison between studies.
- Whether the concentrations detected are sufficient to do meaningful signalling work in the body is an open question.
- Which predicted reading frames are actually translated, and where, is still being worked out.
- No mitochondrial-derived peptide is an approved therapeutic anywhere.
None of that makes the discovery less interesting. A genome sequenced, annotated and taught for decades turned out to contain reading frames nobody had catalogued, found because someone ran a screen for a biological effect rather than staring harder at the sequence. That is a good story about how molecular biology works. It is not a settled mechanism.
Regulatory Status: What Is and Is Not True
MOTS-c is not FDA-approved for any indication. Neither humanin nor the SHLPs are approved therapeutics. They are research molecules.
On July 23 and 24, 2026, the FDA's Pharmacy Compounding Advisory Committee voted to recommend adding MOTS-c to the 503A Bulks List, against FDA staff's own pre-meeting recommendation not to add it. PCAC recommendations are advisory and non-binding. The FDA must complete notice-and-comment rulemaking, which can take 12 months or more. Nothing has changed yet in what may lawfully be compounded.
For the full sequence, see the TelosRX explainer on the July 2026 FDA PCAC vote, and the FDA's Pharmacy Compounding Advisory Committee page. Related TelosRX reading: a research overview of MOTS-c and background on how peptide safety questions are evaluated.
Frequently Asked Questions
What are mitochondrial-derived peptides?
They are small peptides encoded by short open reading frames located within the mitochondrial genome, inside the 12S and 16S ribosomal RNA genes. The group described so far includes humanin, MOTS-c and the small humanin-like peptides SHLP1 through SHLP6. They are studied as candidate signalling molecules. None is an approved therapeutic.
Why was their discovery surprising?
The human mitochondrial genome is only 16,569 base pairs, was fully sequenced in 1981, and was described for decades as encoding 13 proteins, all subunits of the oxidative phosphorylation machinery. Finding additional reading frames nested inside ribosomal RNA genes ran against a long-standing assumption that the annotation was complete.
What is retrograde signalling?
Retrograde signalling is communication running from the mitochondria to the nucleus, altering nuclear gene expression in response to conditions inside the organelle. The opposite direction is called anterograde signalling. Retrograde pathways were characterised in yeast well before mitochondrial-derived peptides were described, and these peptides are studied as possible messengers in that direction.
How do these peptides differ from conventional peptides?
Conventional signalling peptides are encoded by nuclear DNA in dedicated genes and read using the standard genetic code. Mitochondrial-derived peptides are encoded within mitochondrial DNA, in short reading frames nested inside rRNA genes, and the mitochondrial genetic code differs from the standard code at several codons.
Is MOTS-c approved or newly legal after the July 2026 PCAC vote?
No. MOTS-c is not FDA-approved for any indication. The FDA's Pharmacy Compounding Advisory Committee voted on July 23 and 24, 2026 to recommend adding it to the 503A Bulks List, against FDA staff's own pre-meeting recommendation. That recommendation is advisory and non-binding. The FDA must complete notice-and-comment rulemaking, which can take 12 months or more, and nothing has changed yet in what may lawfully be compounded.
How well established is this field?
It is roughly 25 years old at most, and it is uneven. Much of the published work is preclinical, measuring these peptides accurately in humans is technically hard, and questions about which predicted reading frames are genuinely translated are still open. The discovery is well documented. The biology is not settled.
This article is educational and is not medical advice. Humanin, MOTS-c and the small humanin-like peptides are research molecules. None of them is FDA-approved for any indication, and nothing here describes an outcome any of them produces in humans. The July 2026 PCAC vote on MOTS-c was an advisory, non-binding recommendation; a final FDA decision requires notice-and-comment rulemaking, and no change in legal compounding status has taken effect. Compounded medications are not FDA-approved and are prepared under federal and state compounding regulations. Any prescription requires review by a US-licensed provider, and approval is not guaranteed. TelosRX is LegitScript-certified and operates as an online-first, asynchronous telehealth service.