Leucine, Mitochondria and Cell Energy

Table of Contents

1. 1. Who Did the Study and How
2. 2. Why Only Leucine
3. 3. The GCN2 Route, Not mTORC1
4. 4. SEL1L, the Protein-Disposal Factor
5. 5. More Mitochondrial Proteins, More Respiration
6. 6. The Worm Fertility Experiments
7. 7. The Lung Cancer Cell Experiments
8. 8. Press Release Versus Paper
9. 9. Earlier Research in Context
10. 10. Limits and What Remains Unknown
11. Key Research Papers
12. Connections

1. Who Did the Study and How

The paper is Li Q, Weiss K, Niwa F, Riemer J and Hoppe T, “Leucine inhibits degradation of outer mitochondrial membrane proteins to adapt mitochondrial respiration,” Nature Cell Biology 2025, volume 27, issue 11, pages 1889–1901. It comes from the laboratory of Thorsten Hoppe at the CECAD research centre of the University of Cologne, with Jan Riemer's group, also in Cologne. Funding listed in the paper is public and academic: the German Research Foundation (Deutsche Forschungsgemeinschaft, through a cluster of excellence and several collaborative research programmes) and an Alexander von Humboldt postdoctoral fellowship for the first author. No industry funding is listed. The full text is open access.

The models

• The roundworm C. elegans, a millimetre-long nematode widely used to study ageing, metabolism and protein quality control because its genes can be switched off one at a time with RNA interference (RNAi).
• HEK293 cells, a human cell line derived from embryonic kidney tissue and grown in laboratory dishes.
• Three human non-small-cell lung cancer lines, named H2030, H1437 and H1666.

There were no human participants, no animals other than worms, and no diet or feeding arm of any kind.

A new way to watch protein breakdown

Mitochondria are wrapped in two membranes. The outer one (the outer mitochondrial membrane, or OMM) holds the gateways that let proteins, fats and small molecules in and out. Proteins in that outer membrane are turned over constantly: old or surplus ones are tagged with a small protein called ubiquitin and fed into the proteasome, the cell's protein shredder. This tagging-and-shredding system is called the ubiquitin–proteasome system.

To measure how fast that turnover happens, the team built a reporter they named mitoUFD. It is a green fluorescent protein (GFP) fused to ubiquitin, anchored to the outer membrane by the membrane-spanning segment of a worm protein called FIS-1. Because it already carries the ubiquitin “dispose of me” signal, it is broken down quickly; when the disposal machinery slows, the reporter builds up and the green glow gets brighter. A matching construct without ubiquitin, called mitoGFP, served as a control, as did a version sitting in the cytosol (the fluid inside the cell, outside the mitochondria).

The doses

The leucine exposures were short and concentrated: 20 mM or 50 mM leucine added to the worms' medium for 3 hours, and 1–3 mM added to cell-culture medium for 3 hours. These are laboratory concentrations chosen to produce a clear signal in a dish or on an agar plate. They are not dietary intakes, and the paper does not translate them into amounts of food.

The screens

Before focusing on leucine, the authors ran two RNAi screens in worms. The first switched off 38 mitochondrial genes. The second switched off 135 genes involved in amino-acid metabolism. Knocking down 67 of those 135 genes raised mitoUFD levels, and 16 of them raised it by more than 20%. That pattern pointed to amino-acid metabolism as a regulator of outer-membrane protein breakdown, and the team then tested individual amino acids.

Back to Table of Contents

2. Why Only Leucine

Leucine belongs to a trio called the branched-chain amino acids (BCAAs), together with isoleucine and valine. The three are chemically similar and are often studied, and sold, together. In this paper they did not behave alike.

• In worms, 3 hours of leucine raised mitoUFD levels. Isoleucine and valine did not.
• Leucine did not change the cytosolic control reporter, so the effect was specific to proteins at the mitochondria rather than a general slowing of protein breakdown throughout the cell.
• Leucine did not change mitoGFP, the same outer-membrane construct without the ubiquitin tag. The authors read this as evidence that leucine slows degradation of outer-membrane proteins, rather than simply raising the production or level of everything anchored there.

The finding that leucine stands apart from its two close relatives echoes a long line of earlier work in which leucine, more than isoleucine or valine, drives the mTORC1 growth signal. What is new here is that the leucine-specific effect on mitochondria turned out to run through an entirely different pathway.

Back to Table of Contents

3. The GCN2 Route, Not mTORC1

Cells have two main systems for sensing amino acids. One is mTORC1, which is activated when amino acids are plentiful and switches on growth and protein building; this is the pathway usually invoked when leucine is linked to muscle. The other is GCN2, a kinase best known for responding to amino-acid shortage by sensing uncharged transfer RNAs and slowing general protein production.

The authors tested both, in worms and in HEK293 cells:

• GCN2 was required. Knocking down gcn-2, the worm version of the gene, abolished leucine's effect on mitoUFD.
• mTORC1 was not. Knocking down let-363, the worm version of mTOR, did not block the effect. The paper states that “the regulation of OMM protein degradation is independent of mTORC1 signalling.”

In plain terms, the study describes a second leucine-sensing route in the cell. The familiar one (leucine to mTORC1 to protein synthesis) is about building new proteins. This one (leucine to GCN2 to slower protein disposal at the mitochondria) is about keeping existing mitochondrial proteins around for longer. The two can both be switched on by leucine, but in these experiments the second did not depend on the first.

Back to Table of Contents

4. SEL1L, the Protein-Disposal Factor

Slower breakdown means some part of the disposal machinery is being turned down. To find which part, the team measured the ubiquitin-tagging enzymes (E3 ligases) and their helper proteins in worm mitochondrial fractions after leucine exposure.

• One factor fell. SEL-1, the worm counterpart of the human protein SEL1L, was the only E3 ligase or cofactor that was significantly reduced in mitochondrial fractions after leucine.
• Only the mitochondrial pool fell. SEL1L is best known for its job at the endoplasmic reticulum (ER), the cell's protein-folding factory, where it partners with the ligase HRD1 to clear misfolded proteins — a process called ER-associated degradation, or ERAD. In HEK293 cells, the amount of SEL1L in the ER (microsomal) fraction did not change. The drop was confined to the mitochondria-associated pool.
• Removing it copied leucine. Knocking down sel-1, or sel-11 (the worm version of HRD1), stabilised mitoUFD and increased mitochondrial respiration — the same two effects leucine produced.

Put together, the proposed chain is: leucine → GCN2 → less SEL1L at the mitochondria → fewer outer-membrane proteins tagged for destruction → more of those proteins present → more respiration.

SEL1L is a quality-control protein, and that cuts both ways. In the university's press release (not in the paper), the first author is quoted making this point: SEL1L also prevents damaged proteins from building up, so turning up energy output by suppressing it indiscriminately could backfire. The paper itself does not test long-term suppression or measure damaged-protein build-up over time.

Back to Table of Contents

5. More Mitochondrial Proteins, More Respiration

A selective expansion of the mitochondrial proteome

Using proteomics (measuring thousands of proteins at once), the authors found that leucine raised the levels of mitochondrial and outer-membrane proteins but did not raise the total protein content of the cell. The rise persisted when cytosolic protein production was blocked with the drug cycloheximide, which fits the idea that the extra proteins came from slower disposal rather than faster manufacture.

Nine worm outer-membrane proteins rose consistently: DNJ-9, FKB-6, GOP-3, MTCH-1, PGAM5, TOMM-40, VDAC-1, KMO-1 and FZO-1. Among them are parts of the protein-import gate (TOMM-40), a pore for small molecules (VDAC-1) and a membrane-fusion protein (FZO-1). In HEK293 cells, seven human counterparts of these proteins also rose, including components of the import machinery: TOMM40, SAMM50, MTCH1/2 and DNAJC11.

Oxygen consumption

Mitochondrial respiration was measured as oxygen consumption on a Seahorse analyser, the standard laboratory instrument for this. Three hours of leucine increased respiration in both worms and HEK293 cells. In HEK293 cells, the authors also tried other essential amino acids and report that they “resulted in a slight increase … Leu had the highest impact.”

What abolished the boost

The extra respiration disappeared under any of these conditions:

• overexpression of cdc-48.1, a worm version of the protein-extraction machine that pulls tagged proteins out of membranes for the proteasome;
• overexpression of gcn-2;
• knockdown of tomm-40, the central pore of the mitochondrial protein-import gate;
• blocking mitochondrial protein import with the experimental compound MitoBlock-12.

The authors' interpretation is that keeping more import machinery in the outer membrane lets mitochondria take in more of the proteins they need to run respiration. The respiration effect was measured over hours; the paper does not report whether it persists over days or weeks.

Back to Table of Contents

6. The Worm Fertility Experiments

To ask what happens when leucine breakdown itself is faulty, the team built a worm carrying a change in the gene bcat-1, the first enzyme in BCAA breakdown. The change, called E279K, models a human mutation (E264K in the gene BCAT2) that causes raised levels of branched-chain amino acids in the blood.

• Worms with two copies of the change did not survive.
• Worms with one copy (written E279K/+) had higher mitoUFD levels, consistent with slower outer-membrane protein breakdown. They “produced a normal number of eggs, with no differences in brood size or hatch rate.”
• Brood size fell only when gcn-2 was also knocked down in these mutant worms — that is, only with a second genetic hit.
• Separately, overexpressing gcn-2 in ordinary (wild-type) worms reduced egg laying.

The paper did not measure respiration in the mutant worms. The fertility findings therefore show an interaction between BCAA breakdown and GCN2 in worm reproduction, rather than a direct effect of the leucine-breakdown defect on fertility by itself.

Back to Table of Contents

7. The Lung Cancer Cell Experiments

Some cancers carry changes in BCAA-handling genes, so the authors looked at three human non-small-cell lung cancer cell lines.

• H2030, which carries a loss-of-function mutation in BCAT2, and H1666 had higher BCAA levels than H1437.
• The two higher-BCAA lines also showed lower ubiquitin tagging of outer-membrane proteins — the same direction as the leucine effect in worms and HEK293 cells.
• Blocking mitochondrial protein import with MitoBlock-12 strongly inhibited the growth of H1437 cells, but “only mildly affected” H2030 and H1666.

These are three cell lines grown in dishes. No tumours, animals with tumours or patients were studied, and growth was compared only under a single experimental drug. The result fits the proposed mechanism — cells with more BCAA and more stable import machinery coped better when import was partly blocked — but it does not show how these cancers behave in the body.

Back to Table of Contents

8. Press Release Versus Paper

The press release summarises the core mechanism accurately: leucine lowers SEL1L, so fewer mitochondrial proteins are broken down. Several other statements go further than the paper.

• Fertility. The release says that in worms, problems with leucine breakdown disrupted mitochondrial function and were linked to fertility. In the paper, the disease-model worms (bcat-1 E279K/+) had normal brood size and hatch rate. Fertility fell only with an added gcn-2 knockdown, or when gcn-2 was overexpressed in normal worms, and respiration was not measured in the mutant.
• Cancer survival. The release says mutations affecting leucine metabolism could help cancer cells survive. The paper shows that two higher-BCAA lung cancer lines were less affected by one mitochondrial-import inhibitor, in three cell lines in culture. It does not show better survival in general, or any effect in tumours or people.
• Foods. The release names leucine-rich foods — dairy, meat, beans and lentils — alongside the phrase “powers up your cells,” which implies that eating these foods has the effect described. The paper has no dietary arm. It used 20–50 mM leucine on worms and 1–3 mM on cultured cells for 3 hours, and nothing in it shows that eating leucine changes mitochondria in people.
• Muscle. The headline frames the work as going beyond muscle building, but no muscle tissue was used (worms, kidney-derived cells and lung cancer cells), and the paper states that the mechanism is independent of mTORC1, the pathway that usually links leucine to muscle.
• Timing. The release appeared about eleven months after the paper was published on 31 October 2025, so the finding dates from 2025 rather than from the week of the release.

Back to Table of Contents

9. Earlier Research in Context

Leucine and mTORC1

For more than two decades, leucine's best-documented role beyond being a protein building block has been as a signal to mTORC1. A 2012 review by Dodd and Tee in the American Journal of Physiology – Endocrinology and Metabolism, titled “Leucine and mTORC1: a complex relationship,” summarised that link, and the Cologne paper cites it as background. In 2016, Wolfson and colleagues reported in Science that the protein Sestrin2 is a leucine sensor for the mTORC1 pathway. The 2025 study adds a leucine effect that does not need mTORC1 at all. The site's Leucine and mTOR Activation page covers the mTORC1 side in depth.

Leucine and mitochondria over days

Leucine had been connected to mitochondria before, through a slower route. A 2014 study by Liang, Curry, Brown and Zemel in the Journal of Nutrition and Metabolism reported that leucine modulates mitochondrial biogenesis and SIRT1–AMPK signalling in C2C12 myotubes, a mouse muscle cell line. The Cologne authors describe that work as 48 hours of leucine raising mitochondrial biogenesis (the building of new mitochondria) and respiration in mouse muscle cells through SIRT1–AMPK. The new paper describes a much faster, three-hour effect that works by keeping existing outer-membrane proteins from being broken down, rather than by building new mitochondria.

SEL1L–HRD1 and mitochondria

The idea that the ER's quality-control pair SEL1L–HRD1 reaches mitochondria is not entirely new. A 2020 Science paper by Zhou, Torres, Sha and colleagues, titled “Endoplasmic reticulum-associated degradation regulates mitochondrial dynamics in brown adipocytes,” was earlier evidence that this machinery affects mitochondria. The 2025 study adds a nutrient signal, leucine, that dials the mitochondrial pool of SEL1L down.

BCAA metabolism and worm reproduction

In 2024, Lesnik, Kaletsky and colleagues reported in Nature Metabolism that enhanced branched-chain amino acid metabolism improves age-related reproduction in C. elegans. That work had already tied BCAA handling to worm fertility; the Cologne study adds GCN2 and outer-membrane protein turnover to that picture.

Back to Table of Contents

10. Limits and What Remains Unknown

Limits the authors state

• How SEL1L gets there. SEL-1 “lacks an obvious mitochondrial targeting sequence,” the authors write. Further work is needed to learn how it reaches the outer mitochondrial membrane and how leucine specifically regulates the mitochondria-associated pool while leaving the ER pool untouched.
• Cautious conclusion. The paper says only that the findings “may link diet-induced changes in mitochondrial proteostasis and metabolism to diseases such as cancer and infertility.” The word “may” is the authors' own.

Limits built into the design

• Models. Roundworms and cultured human cell lines only. No mammals, no human tissue taken from people, no clinical measurements.
• Doses. Leucine at 20–50 mM (worms) and 1–3 mM (cells) for 3 hours. Whether ordinary blood leucine levels after a meal produce the same effect in human tissues is untested.
• Time scale. Effects were measured over hours. Days, weeks or a lifetime of different leucine intakes were not studied.
• Tissue. No muscle, heart, brain or liver tissue was examined, even though these are the organs most dependent on mitochondrial energy.
• Trade-offs. Because SEL1L also clears faulty proteins, a sustained reduction could carry costs the three-hour experiments cannot reveal.

Open questions

• Does the leucine–GCN2–SEL1L route operate in mammalian tissues, and at what leucine concentrations?
• How does it interact with the mTORC1 route, which leucine switches on at the same time?
• Does the observed resistance of higher-BCAA lung cancer cells to an import blocker hold in animal tumour models?
• Does a long-term change in outer-membrane protein turnover help or harm mitochondrial quality as cells age?

At present, the study is a mechanism found in worms and cell lines. It identifies a new pathway worth testing further; it does not, on its own, describe what eating leucine does to human energy levels.

Back to Table of Contents

Key Research Papers

1. Li Q, Weiss K, Niwa F, Riemer J, Hoppe T (2025). Leucine inhibits degradation of outer mitochondrial membrane proteins to adapt mitochondrial respiration. Nature Cell Biology 27(11):1889–1901 — PubMed PMID: 41174002
2. Dodd KM, Tee AR (2012). Leucine and mTORC1: a complex relationship. American Journal of Physiology – Endocrinology and Metabolism 302(11):E1329–E1342 — PubMed PMID: 22354780
3. Wolfson RL, Chantranupong L, Saxton RA, Shen K, Scaria SM, Cantor JR, et al. (2016). Sestrin2 is a leucine sensor for the mTORC1 pathway. Science 351(6268):43–48 — PubMed PMID: 26449471
4. Liang C, Curry BJ, Brown PL, Zemel MB (2014). Leucine modulates mitochondrial biogenesis and SIRT1-AMPK signaling in C2C12 myotubes. Journal of Nutrition and Metabolism 2014:239750 — PubMed PMID: 25400942
5. Zhou Z, Torres M, Sha H, Halbrook CJ, Van den Bergh F, Reinert RB, et al. (2020). Endoplasmic reticulum-associated degradation regulates mitochondrial dynamics in brown adipocytes. Science 368(6486):54–60 — PubMed PMID: 32193362
6. Lesnik C, Kaletsky R, Ashraf JM, Sohrabi S, Cota V, Sengupta T, et al. (2024). Enhanced branched-chain amino acid metabolism improves age-related reproduction in C. elegans. Nature Metabolism 6(4):724–740 — PubMed PMID: 38418585

PubMed Topic Searches

1. PubMed: Leucine and outer mitochondrial membrane protein degradation
2. PubMed: GCN2 amino-acid sensing and mitochondria
3. PubMed: SEL1L–HRD1 and mitochondria
4. PubMed: Branched-chain amino acids and mitochondrial respiration
5. PubMed: BCAT2 and lung cancer

Back to Table of Contents

Connections

• All Amino Acids
• Leucine Benefits Hub
• Leucine Overview
• Leucine and mTOR Activation
• Sarcopenia Prevention
• Muscle Protein Synthesis
• Leucine Food Sources
• News, October 11, 2026
• Mitochondria: How Your Cells Make Energy
• Protein Folding and Chaperones
• Valine and Energy Metabolism
• Isoleucine
• Valine

Back to Table of Contents