AMPK

AMPK is a three-part enzyme, one catalytic subunit and two regulatory ones, and its job is to notice when a cell is running short of energy and do something about it. Everything it does downstream follows one rule: nothing expensive gets built while the gauge reads low.

The crucial thing to understand about AMPK is that it does not measure how much ATP a cell has. It measures a ratio. Cells hold ATP in a narrow band, so absolute ATP is a poor signal. When ATP consumption outruns production, ADP accumulates, and an enzyme called adenylate kinase converts two ADP into one ATP plus one AMP. Because of the arithmetic of that reaction, a small proportional fall in ATP produces a much larger proportional rise in AMP. AMP is therefore a high gain amplifier of energy stress, which is why evolution built the sensor around it.

The honest headline for a fasting page is proportionality. Exercise activates AMPK far more strongly and far more acutely than fasting does, and the human muscle biopsy data are consistent with that.

The regulatory gamma subunit contains domains that bind adenine nucleotides competitively. AMP binding does three things at once: it promotes phosphorylation of the activation loop threonine, Thr172, on the catalytic subunit; it protects that phosphate from being removed by phosphatases; and it allosterically activates the already-phosphorylated kinase. ADP contributes to the first two. ATP competes and switches the sensor off. That is what makes it a ratio sensor rather than an AMP counter.

Thr172 is phosphorylated by an upstream kinase, principally LKB1, a tumour suppressor that works in a complex with two accessory proteins, STRAD and MO25. LKB1 is not itself regulated by energy status. It is constitutively active. What is regulated is AMPK's susceptibility to being phosphorylated by it and its resistance to being dephosphorylated. A second upstream kinase, CaMKK2, phosphorylates the same site in response to rising intracellular calcium, which is how muscle contraction activates AMPK independently of the energy ratio.

What AMPK switches on: fatty acid oxidation, by phosphorylating and inhibiting acetyl-CoA carboxylase, which lowers malonyl-CoA, which opens the CPT1 gate into the mitochondrion; glucose uptake, by moving GLUT4 to the muscle membrane without insulin; mitochondrial biogenesis, by directly phosphorylating the coactivator PGC-1 alpha; and autophagy, by phosphorylating ULK1 at serine 555 and other sites.

What it switches off: mTORC1, by two routes, phosphorylating TSC2 and phosphorylating raptor directly; fatty acid and cholesterol synthesis; glycogen synthase; and protein synthesis.

Note where PGC-1 alpha sits. AMPK phosphorylates it and SIRT1 deacetylates it. The energy sensor and the redox sensor converge on the same protein by two different chemistries.

A rising AMP to ATP ratio. In practice that means muscle contraction, hypoxia, and any intervention that impairs ATP production or raises its consumption. Glucose availability is also sensed at the lysosome by a non-canonical route involving fructose-1,6-bisphosphate, aldolase and an Axin and LKB1 arrangement, and glycogen is sensed by a carbohydrate binding module on the beta subunit.

AMPK is the node where the popular framing and the human measurement diverge most sharply, and the divergence is informative rather than disappointing. Fasting is routinely described as a strong AMPK activator on the basis of rodent work. In human skeletal muscle, across the durations anyone has biopsied, that activation has not been observed, and one careful study found the markers drifting the other way in the first 8 hours.

The reason is proportionality. Exercise moves the AMP to ATP ratio hard by consuming ATP directly, and it releases calcium that activates the same kinase through a second route. Fasting does neither. If the outcome of interest is AMPK activation in muscle, exercise is the substantially stronger stimulus and the human data say so.

What this does not establish is that AMPK is unactivated during a human fast. Liver is where the energetic transition is largest and where the gluconeogenic and ketogenic programmes actually run, and nobody biopsies a healthy volunteer's liver. That is an unmeasured tissue, not a measured null.