The Book of PeptidesThe Book of Peptidesa hormone from the other genome
2001-2003 (humanin, the first mitochondrial-derived peptide, reached by three groups from unrelated directions); 2015 (MOTS-c published in Cell Metabolism, GenBank deposit January 2015)
MOTS-c is written into the mitochondrial genome — but no mitochondrion can build it. Read with the organelle's own genetic code, the gene collapses into tandem start and stop codons, so the transcript has to be shipped out and translated by the cell's ordinary ribosomes.
The mitochondrial genome was supposed to be a spent cartridge: 37 genes, 13 of them protein-coding, all building respiratory machinery, nothing left over. Then in 2001 a Keio University group led by Ikuo Nishimoto, screening for protection against Alzheimer's-linked neuronal toxicity, cloned a 24-amino-acid peptide they called humanin — and its open reading frame sat inside the mitochondrial 16S rRNA gene. Two further routes reached it within two years from unrelated directions: a screen for proteins that block Bax, and a hunt for binding partners of insulin-like growth factor binding protein-3, run jointly with Nishimoto's group. That second search belonged to Pinchas Cohen, a pediatric endocrinologist at UCLA who had not been looking for mitochondrial hormones at all.
Cohen moved to USC's Leonard Davis School of Gerontology and, with Changhan Lee, asked the obvious follow-up: if the 16S rRNA gene hides one functional peptide, what about the 12S? An in silico sweep for short open reading frames returned a sixteen-amino-acid candidate, deposited in GenBank as KP715230 on 26 January 2015 and named MOTS-c — mitochondrial open-reading-frame of the twelve S rRNA, type-c. The authors then published the complication most retellings drop: read with the mitochondrial genetic code, the sequence degenerates into tandem start and stop codons, so translation obligatorily occurs in the cytoplasm, by an export mechanism they conceded was poorly understood. The GenBank annotation says the same.
What it does is stranger than a generic metabolic tonic. Unbiased metabolomics pointed the group at the folate-methionine cycle and the purine synthesis wired to it: obstruct that pathway and AICAR accumulates, activating AMPK, the cell's low-fuel sensor. The authors noted the parallel to methotrexate, an antifolate used in cancer and rheumatoid arthritis. In CD-1 mice, 0.5 mg/kg/day intraperitoneally prevented diet-induced obesity, and not by suppressing appetite — caloric intake was identical between groups. Six years later, in Nature Communications, ten days of 15 mg/kg/day put every treated young CD-1 mouse through the final 23 m/min stage of a treadmill ramp, against 16.6% of vehicle animals and 16.6% of those given a third of that dose.
The verdict turns on three specifics. MOTS-c was not the discovery that mitochondria encode hormones: humanin was, in 2001, as the 2015 paper's opening sentence credits. "Encode" also does more work than "make" — read with the mitochondrial genetic code the gene yields tandem start and stop codons, so the peptide is assembled in the cytoplasm, a caveat the paper and its GenBank annotation both state plainly. And the Japanese-longevity variant story inverted: the same network's 27,527-person analysis in 2021 tied that allele to more diabetes in men, not longer life.
MOTS-c is approved by no regulator, anywhere, for anything. The only completed interventional trial tested an analog, CB4211, not the peptide — run by CohBar, Inc., which Cohen co-founded and whose stock the discovery paper disclosed he held; its outcomes exist only as the sponsor's own SEC-filing characterization, never peer-reviewed. A Phase 2a of MOTS-c itself (NCT07505745, 120 adults with prediabetes) was registered in 2026 with no dose disclosed, and a registry entry is not a result. Anti-doping chemists validated a WADA-standard LC/MS plasma assay in 2019, then could not confirm, down to 100 pg/mL, the concentrations a commercial ELISA reported in twenty healthy subjects — an unresolved discrepancy under much of the age-related-decline literature. It sells meanwhile as a research chemical labeled not for human use.
The counterweights arrived in the same paper. In a separate cohort on a 60%-fat diet, MOTS-c improved running overall but not one mouse reached that final stage; treatment begun near 24 months of age improved grip strength, stride length and walking capacity late in life, while the survival curve run alongside did not reach significance. The longevity story turned around too. A 2015 proposal held that m.1382A>C, a variant inside the MOTS-c reading frame almost confined to Northeast Asia, might help explain Japanese longevity; pooling 27,527 people across three cohorts in 2021, an overlapping set of authors tied the C allele to more type 2 diabetes in men, and to a variant peptide with reduced activity.