Leucine and Muscle Protein Synthesis

Leucine is a signal, and your cells have a receptor for it#

Of the twenty amino acids in food, leucine is the one that does a second job. Besides being a brick, it is a message: cells detect free leucine and read it as evidence that protein has arrived, then switch on mTORC1, the master regulator that licenses muscle protein synthesis. This is not a metaphor that got standardized. There is a specific protein that physically holds a leucine molecule and releases it, and its binding affinity matches the leucine concentration at which the pathway turns on.

The practical conclusion, though, is the opposite of what the supplement aisle drew from it. Flipping the switch is necessary — block the pathway pharmacologically and a protein meal stops building muscle at all. It is also nowhere near sufficient: pooled across randomized trials, supplementing leucine on its own changes lean mass by essentially zero. The gap between those two facts is the interesting part of this subject, and it is explained by a timer nobody put in the marketing. Your daily total still comes from how much protein per day; this is what happens inside each dose of it.

The sensor is a protein that lets go when leucine arrives#

For decades "leucine signals mTORC1" was a black box. The box was opened by identifying the molecule doing the detecting.

Sestrin2 sits inside cells bound to a complex called GATOR2, inhibiting it. GATOR2, when free, ultimately permits mTORC1 activation — so as long as Sestrin2 is clamped onto it, the pathway stays down. Leucine binds Sestrin2 and breaks that grip. Crucially, the affinity is not arbitrary: leucine binds Sestrin2 with a dissociation constant of 20 ± 5 µM, and half-maximal mTORC1 activation occurs at 20–40 µM leucine — the sensor is tuned to the concentration range at which the switch actually moves. And it is specific: leucine, but not arginine, disrupts the Sestrin2–GATOR2 interaction, with arginine binding neither Sestrin2 nor the control protein tested¹.

That is what "leucine is a signal" means in physical terms: a molecule shaped to hold leucine, at a strength calibrated to the concentrations a meal produces, whose only job on binding is to stop inhibiting something.

The story is not fully closed. The same authors note that leucyl-tRNA synthetase — a different cytosolic protein, one that normally loads leucine onto transfer RNA — binds leucine with a reported affinity of 45 µM, in the same neighbourhood². They do not present it as a rival explanation, and that review comes from the laboratory that identified Sestrin2, which is worth knowing when reading a field's assessment of its own discovery. Cellular leucine sensing has more than one leucine-binding protein in it, and the accounting is still being done.

Block the switch and the meal stops working#

A sensor in a cell line is a long way from a human leg. The experiment that closed that distance ran the pathway backwards: rather than adding leucine and watching muscle respond, it added amino acids and blocked the pathway.

Eight young adults completed two randomized trials, ingesting 10 g of essential amino acids, once alone and once after being given rapamycin — a direct mTORC1 inhibitor — with vastus lateralis biopsies before and after. In the control trial, muscle protein synthesis rose about 60%, alongside phosphorylation of mTOR, S6K1 and 4E-BP1. With rapamycin on board, the increase in muscle protein synthesis was completely blocked, and the signaling proteins were blocked or attenuated — while amino acid availability stayed comparable between trials³.

Same amino acids in the blood, no mTORC1, no response. That is as close to a demonstration of necessity as human physiology gets: the amino acids are not building muscle by mass action, they are building it because a signal was received and acted on.

The switch has an off-timer, and the timer ignores your bloodstream#

Here is the finding that reorganizes everything else, and it is twenty-five years old.

Researchers infused mixed amino acids intravenously at a constant rate — a square-wave increase in availability, held steady — and tracked quadriceps muscle protein synthesis for six hours. Basal synthesis ran 0.076 ± 0.008 %/h. Nothing detectable happened in the first 30 minutes. Then the rate climbed steeply to a peak at about 2 hours of roughly 2.8 times basal. And then, between 120 and 360 minutes, it fell markedly from that peak, becoming not significantly different from baseline — while the infusion continued and amino acids stayed elevated the entire time⁴.

Muscle protein synthesis responds rapidly to amino acids and is then shut down despite their continued availability. The ceiling is a timer, not a shortage.

This is the phenomenon usually nicknamed "muscle full," and it dissolves a lot of confusion. The muscle does not stop responding because it ran out of raw material — the raw material was still arriving by catheter. It stops because the response is self-limiting. Which means the anabolic effect of a meal is a pulse with a shape: a latency, a rise, a peak, and a refractory decline. You cannot lengthen the pulse by keeping leucine high, because leucine being high is not what ends it. This is also why the per-meal ceiling and the per-day ceiling behave so differently, a distinction worked through in the protein per meal limit.

Signal is not substrate: what leucine alone actually does#

Now put the pieces together and a prediction falls out. If leucine's role is to flip a switch that then times out on its own, then giving someone leucine without giving them the other amino acids should produce a real signal and no building material — a lit switch in an empty workshop.

That is what the long trials find. Pooling 17 randomized controlled trials in 1,418 older adults, leucine supplementation given in isolation had no effect on total lean mass: a weighted mean difference of 0.03 kg (95% CI −0.51 to 0.57, P = 0.917). Handgrip strength, no effect (WMD 1.23 kg, CI −0.58 to 3.03). Leg press, no effect (WMD −1.35 kg, CI −7.46 to 4.77). Leucine given combined with other nutrients including vitamin D did move some outcomes — handgrip strength by 2.17 kg (CI 0.24 to 4.10, P = 0.027) and gait speed by 0.03 m/s (CI 0.01 to 0.05)⁵.

So the acute literature and the chronic literature are not in conflict, and the thing that reconciles them is nameable. Leucine genuinely triggers synthesis; that finding replicates. But a trigger fired into a bloodstream short of the other eight essential amino acids produces a pulse that has nothing to build with, and the pulse ends on schedule regardless. Isolated leucine is the one intervention that maximizes signal and minimizes substrate, which is exactly why it is the one that shows nothing.

The corollary matters more than the myth-busting: the leucine content of a protein is a reasonable tiebreaker between sources of similar dose, which is why it predicts differences between powders — see whey, casein, or plant protein — and a poor substitute for dose itself.

What this looks like on a plate#

The popular translation of all this is a per-meal leucine target, usually quoted around 2.5 to 3 grams. Set that number against what real servings contain and it is tighter than it sounds. These are USDA reference values in the household measures USDA itself uses⁶:

Food (USDA measure)	Leucine
Swiss cheese, 1 cup diced	3.91 g
Yellowtail, cooked, half fillet	3.52 g
Black beans, raw, 1 cup	3.35 g
Top sirloin, lean, roasted, 3 oz	2.57 g
Bison chuck, lean, braised, 3 oz	2.46 g
Top round steak, grilled, 3 oz	2.39 g
Peanuts, dry-roasted, 1 cup	2.24 g

Read the top of that list skeptically. A cup of diced Swiss cheese and a cup of raw black beans are the portions that win a leucine ranking, and neither is a thing anyone eats. The rows worth attending to are the 3-ounce ones — a palm-sized piece of cooked meat, which lands at 2.4 to 2.6 g. In other words, an ordinary serving of a good protein food sits right on the threshold rather than comfortably past it, which is the real argument for a decent-sized portion at each meal instead of a supplement.

Where this genuinely changes behaviour is at the ends of the age range, because the dose needed to clear the threshold rises as muscle becomes harder to stimulate — the evidence for that shift, and what it means for how an older adult should build breakfast, is in protein for older adults. For everyone else the practical conclusion is deflating in the useful way: hit the daily total, put a real serving of protein in each main meal, and the switch takes care of itself. The dose-response evidence behind that total is in how much protein to build muscle.

FAQ#

How much leucine do you need in a meal?#

The commonly quoted figure is about 2.5 to 3 grams per meal, and ordinary servings sit close to that line rather than far above it: USDA puts a 3-ounce grilled top round steak at 2.39 g of leucine and a 3-ounce roasted top sirloin at 2.57 g⁶. Rather than counting leucine, the reliable move is a full protein serving at each main meal — the amino acid arrives with the substrate it needs, which isolated leucine does not.

Do leucine supplements build muscle?#

On the pooled evidence, no. Across 17 randomized controlled trials in 1,418 older adults, isolated leucine supplementation changed total lean mass by 0.03 kg (95% CI −0.51 to 0.57, P = 0.917), with no effect on handgrip strength or leg press⁵. Leucine is the signal that starts muscle protein synthesis, not the material it is built from — firing the signal without supplying the other essential amino acids gives you a response with nothing to construct.

Why does muscle protein synthesis fall while amino acids are still high?#

Because the response is self-limiting rather than supply-limited. Under a constant intravenous amino acid infusion, muscle protein synthesis showed no change for 30 minutes, peaked at about 2 hours at roughly 2.8 times basal, then fell between 120 and 360 minutes to a rate not significantly different from baseline — all while amino acids stayed elevated⁴. Keeping leucine high does not extend the pulse, because leucine running out was never what ended it.

Sources#

This article was researched and drafted with AI assistance and reviewed for accuracy by the BurnWeek team. It is general information, not medical advice. How we research and correct our articles →