Magnesium: The Mineral That Makes ATP Work

Textbooks tell you ATP is the energy currency. That is only half the story. Bare ATP is biologically useless. Its three phosphate groups carry a dense negative charge, they shove each other apart, and an enzyme cannot grip the thing. Magnesium fixes that. A single Mg²⁺ ion clamps the β- and γ-phosphates, neutralises the charge, and the calmed complex — Mg-ATP — is the actual substrate of essentially every kinase and ATPase in your body. Press Play and watch bare ATP get rejected at an enzyme's active site, then watch magnesium dock and the enzyme fire.

Try this: watch Normal for ten seconds, then press Deficiency — the common, subclinical kind. Notice what the serum magnesium readout does. It stays inside the normal range. Meanwhile the cells empty, ATP loses its clamp, the magnesium plug pops out of the NMDA channel and the muscle fibre stops letting go. That gap between a normal blood test and a depleted body is the whole point of this page.

Speed Diagram is illustrative — not to scale.

Live magnesium readout

Serum Mg — the standard blood test

0.86 mmol/L

2.09 mg/dL · reference band 0.75–0.95 mmol/L (1.8–2.3 mg/dL)

Inside the cells — RBC Mg

2.25 mmol/L

typical lab band 1.65–2.65 mmol/L — ranges differ between labs

ATP in the Mg-ATP form

91%

the rest is bare ATP — charged, flailing, and useless to the enzyme

Enzyme throughput

100 % of normal

0reactions fired

100%Na⁺/K⁺ pump rate

NMDA channel block

93 % plugged

the Mg²⁺ cork. Lose it and glutamate gets a free hand.

Muscle relaxation time

45 ms (model)

Calcium clears between twitches. The fibre lets go.

Magnesium is where it should be: cells full, ATP clamped, channels corked, muscle relaxing.

What’s happening

Bare ATP drifts toward the enzyme, its three phosphates repelling each other…

phosphate (ATP) adenine Mg²⁺ Ca²⁺ K⁺ glutamate

What is real and what is model: the reference bands (serum Mg 0.75–0.95 mmol/L = 1.8–2.3 mg/dL; RBC Mg roughly 1.65–2.65 mmol/L), the body distribution (~50–60% bone, ~40% intracellular, ~1% serum), the fact that most cellular ATP is Mg-complexed, the 600+ Mg-dependent enzymes, the voltage-dependent Mg²⁺ block of the NMDA receptor and SERCA’s dependence on Mg-ATP are all real, textbook biochemistry. The numbers that move — throughput %, block %, relaxation time in ms, reactions fired — are an illustrative model tuned to be directionally right, not a measurement of your body. One honest caveat: the NMDA plug is relieved mainly by depolarisation, and falling magnesium lowers the threshold for that rather than yanking the cork out at rest. The animation exaggerates the pop-out to make the mechanism visible.

The Science in Plain Language

“But my magnesium test came back normal.”

It probably did. That is the problem. About 1% of your body’s magnesium is in your serum. Roughly 50–60% is locked in bone and about 40% is inside your cells. Serum magnesium is not a fuel gauge — it is a thermostat. Your body defends that 1% relentlessly, pulling magnesium out of bone and muscle to keep the blood number steady, because the blood number controls your heart rhythm and it is not allowed to drift.

So a normal serum magnesium tells you that the defence is working. It does not tell you the tank is full. You can be genuinely depleted with a perfectly normal serum magnesium — and by the time serum does fall, you are usually far down the road. Press Deficiency on the animation and watch it happen: the serum needle sits inside the green band, the verdict chip says NORMAL, and the cells behind it are emptying.

The chemistry: why bare ATP is a broken tool

ATP is adenosine triphosphate — a tail of three phosphate groups hanging off an adenosine. At the pH inside your cells, that tail carries about four negative charges, crammed into a molecule a fraction of a nanometre long. Like charges repel. The tail is under permanent internal strain, it flexes and squirms, and its electric field screams at every other negative charge nearby.

Picture handing someone three strong magnets taped in a row, all facing the same way. They shove each other apart, the bundle writhes, and nobody can hold it steady enough to do anything useful with it. Now clip a heavy opposite-pole magnet across the last two. The squirming stops. The bundle sits still in the hand.

Mg²⁺ is that clip. It binds across the β- and γ-phosphates, its double positive charge cancelling the repulsion and locking the tail into a defined shape. The enzyme’s active site is not built to hold ATP — it is built to hold Mg-ATP. The magnesium is part of the key, not part of the lock. Kinases do not just tolerate magnesium; they require it, positioning the metal ion to steer the transferred phosphate onto its target and stabilising the transition state as the bond breaks.

Because the affinity is high and free magnesium inside a cell sits around 0.5–1.0 mmol/L, the great majority of cellular ATP — well over 90% — is walking around with a magnesium ion clamped to it at any given moment. When a biochemist writes “ATP” in a reaction, what they almost always mean is Mg-ATP. The textbook line “ATP is the energy currency” quietly omits the fact that the currency is only legal tender with magnesium attached.

Why “600 enzymes” is not marketing

Supplement labels love the number, which makes it easy to dismiss. It happens to be true. Magnesium is a required cofactor for more than 600 enzymes and an activator of roughly 200 more. That is not because magnesium is magic. It is because so much of metabolism runs on phosphate chemistry, and every reaction that hands off a phosphate from ATP needs the magnesium clamp.

Look at the five enzymes on the animation’s enzyme floor and the scale becomes obvious:

Hexokinase takes glucose and phosphorylates it — the very first step of glycolysis. No Mg-ATP, no glucose metabolism.
Na⁺/K⁺-ATPase, the pump that maintains the electrical gradient across every cell membrane you own, and which consumes a large fraction of your resting energy budget. It hydrolyses Mg-ATP.
DNA polymerase holds two magnesium ions in its catalytic centre; they are what makes the chemistry of copying and repairing DNA possible at all.
Adenylate cyclase converts ATP into cyclic AMP — the second messenger through which adrenaline, glucagon and many other hormones actually speak to a cell.
Glutathione synthase builds glutathione, your master intracellular antioxidant. Low magnesium means low glutathione synthesis.

Magnesium is also the second most abundant cation inside your cells, after potassium. It stabilises ribosomes, holds the phosphate backbone of DNA and RNA together, and is required for the enzymes that copy them. This is why a deficiency does not produce one clean symptom. It produces a diffuse, hard-to-pin-down malaise across a dozen systems — which is exactly what patients describe, and exactly why they are so often told nothing is wrong.

The potassium link: you cannot fix low potassium without fixing magnesium

This is the single most clinically useful fact on this page, and it is standard hospital medicine.

Potassium leaves your cells and enters the urine through a channel in the kidney called ROMK. Under normal conditions, magnesium inside the cell physically plugs that channel from behind, limiting how much potassium can escape. When intracellular magnesium falls, the plug goes with it — and ROMK starts leaking potassium into the urine continuously.

The result is refractory hypokalaemia: you give potassium, the level barely moves; you give more, it still barely moves, because it is running straight out through an unplugged channel. The mechanism was worked out and described clearly by Huang and Kuo (Journal of the American Society of Nephrology, 2007), and the practical rule that falls out of it is the one every hospital doctor learns: if the potassium will not come up, check — and replace — the magnesium. Correct the magnesium and the potassium often corrects itself.

If you have ever been told your potassium is stubbornly low and nobody mentioned magnesium, this is the conversation to have.

The calcium link: magnesium is calcium’s brake

Calcium is the go signal. It makes muscles contract, nerves fire, platelets clump, and cells commit to action. Magnesium is the stop signal. Chemically they are close enough that magnesium can sit in many of calcium’s seats without doing calcium’s job — it competes at channels, at binding sites on the contractile machinery, and at the pores that let calcium in. Magnesium is, in the most literal sense, a natural calcium antagonist.

This is not a wellness metaphor. It is a licensed, life-saving drug. Intravenous magnesium sulfate is the standard of care for preventing and treating the seizures of pre-eclampsia and eclampsia. The Magpie Trial (Lancet, 2002) randomised over 10,000 women with pre-eclampsia and found that magnesium sulphate more than halved the risk of eclampsia. Earlier trials had already shown it beats both diazepam and phenytoin for stopping eclamptic seizures. Magnesium is also first-line for the dangerous arrhythmia torsades de pointes. When the emergency is “a cell is firing when it should not be,” magnesium is what the crash trolley reaches for.

The two panels at the bottom of the animation are the same principle in two tissues. In the neuron, a magnesium ion sits inside the pore of the NMDA receptor, corking it. Glutamate can bind, but nothing gets through until the membrane depolarises enough to spit the magnesium out — a genuine, well-characterised, voltage-dependent block. It is one of the most elegant switches in neuroscience: the channel only opens if the cell is already excited and glutamate is present. Lower the magnesium and the switch gets sloppy — less depolarisation is needed to open it, calcium enters more freely, and neurons become jumpy. The neurological face of magnesium deficiency — anxiety, insomnia, hyperreflexia, migraine, eyelid twitching — is what a nervous system without enough brake looks like.

In the muscle fibre, contraction is easy and relaxation is the expensive part. To let go, the fibre has to sweep calcium back into its storage tank using SERCA — a pump which is an ATPase, which means it runs on Mg-ATP. Short of magnesium, SERCA slows, calcium lingers around the filaments, and the fibre cannot fully release. Stimulate it again before it has cleared and the twitches fuse: sustained, involuntary contraction. That is cramp, and at the extreme, tetany.

An honest caveat, because you deserve one: the mechanism above is solid, but that does not mean magnesium pills reliably cure ordinary leg cramps. A Cochrane review of magnesium for skeletal muscle cramps concluded it is unlikely to provide clinically meaningful cramp prevention in older adults, with the evidence in pregnancy-related cramps conflicting and low quality. Both things are true at once: magnesium is essential for muscle relaxation, and most people whose calves seize up at 3 a.m. are not deficient. Fix a deficiency if you have one; do not expect magnesium to be a cramp cure if you do not.

What quietly drains it

Magnesium deficiency is rarely dramatic. It is usually the slow arithmetic of a diet that is short and a life that leaks.

Proton-pump inhibitors (omeprazole, esomeprazole, lansoprazole and relatives). Long-term use impairs the active, channel-mediated absorption of magnesium in the gut. This is not a fringe claim: the FDA issued a formal safety communication in 2011 about low magnesium with prolonged PPI use. If you have been on a PPI for years, this is worth knowing about — and worth asking whether you still need it.
Loop diuretics (furosemide) and thiazides. They act on exactly the parts of the kidney tubule where magnesium is reabsorbed, and they push it into the urine.
Alcohol. It drives magnesium out through the kidneys, and heavy drinkers are frequently and substantially depleted.
Poorly controlled diabetes. High blood glucose spills into the urine and drags water with it — an osmotic diuresis — and magnesium goes along for the ride. Low magnesium is common in type 2 diabetes, and because magnesium is needed for insulin signalling, the deficiency and the insulin resistance feed each other.
Chronic diarrhoea, coeliac disease, Crohn’s, bowel resection — anything that shortens or inflames the absorptive surface.
Plain ordinary eating. Refining grains strips most of the magnesium out; a diet built on white flour, white rice and ultra-processed food is a low-magnesium diet by construction. Population surveys consistently find a large share of adults taking in less than the recommended 310–320 mg/day for women and 400–420 mg/day for men.

So what should you test instead?

Here is the honest answer, including the uncomfortable part: there is no perfect test, and no good one is routine. Anyone who tells you otherwise is selling something.

Serum magnesium — cheap, universal, and largely useless for detecting depletion. It is genuinely useful for detecting danger (a serum magnesium that is actually low is a serious finding). It is not a screening test for whether your tissues are full.
RBC (red blood cell) magnesium — measures magnesium inside a cell rather than in the fluid around it, which is at least the right compartment. Better, and orderable from most labs, but red cells are long-lived and are not muscle or brain, so it is an imperfect proxy, and reference ranges vary between laboratories.
Ionised magnesium — measures the biologically active, unbound fraction rather than the total. Conceptually the most sensible blood measurement; in practice it is not widely available and not standardised.
The magnesium loading (retention) test — give a measured magnesium load, then collect urine for 24 hours and see how much comes back out. A depleted body hangs on to it; a replete body dumps it. This is the closest thing to a gold standard, it is used in research, and almost no one does it in clinic because it takes a day and a urine collection.

What that adds up to, in practice: the diagnosis is often clinical. The pattern — a magnesium-poor diet, a drug or a condition that wastes it, plus the constellation of twitching, cramps, poor sleep, anxiety, palpitations, and a potassium that will not stay up — is more informative than the blood test, and a careful, bounded trial of repletion is often more informative than either.

Food first — the real numbers

Magnesium sits at the centre of the chlorophyll molecule, the way iron sits at the centre of haem. Green leaves are literally built around it. Seeds and nuts store it. The numbers below are approximate USDA values, per typical serving:

Pumpkin seeds — about 150 mg per ounce (28 g). The most magnesium-dense everyday food most people can buy.
Spinach, cooked — about 155 mg per cup. (Raw spinach is mostly water — a cup of raw leaves is a fraction of that.)
Almonds — about 80 mg per ounce, roughly 23 nuts.
Black beans, cooked — about 60 mg per half cup.
Dark chocolate (70–85% cocoa) — about 65 mg per ounce. The craving is not entirely a character flaw.
Cashews — about 75 mg per ounce. Avocado — about 45 mg per cup.

Bear in mind you only absorb roughly 30–40% of the magnesium you eat — the fraction rises when you are depleted and falls when you are replete, which is your gut quietly doing the right thing. A handful of pumpkin seeds and a bowl of greens is not a gimmick; it is a serious dose.

The forms, compared honestly

Every magnesium supplement is a magnesium ion attached to something else. That something else determines how much you absorb, what it costs, and how urgently you will need a bathroom.

Magnesium glycinate (bisglycinate) — magnesium bound to the amino acid glycine. Well absorbed and the gentlest on the bowel, which makes it the sensible default if you are actually trying to replete rather than to move your bowels. Glycine itself is mildly calming, which is probably part of why this form has the reputation it does for sleep.
Magnesium citrate — well absorbed, cheap, widely available. It is also mildly laxative at higher doses, which is a feature if you are constipated and a nuisance if you are not.
Magnesium oxide — the one in most cheap tablets and in a great deal of “high strength” marketing, because it is 60% magnesium by weight and therefore looks impressive on the label. It is poorly absorbed. In a published comparison of commercial preparations, oxide came out the worst absorbed of the forms tested, while chloride, lactate and aspartate did substantially better. Most of an oxide dose stays in your gut — which is precisely why magnesium oxide is sold as a laxative and an antacid. If your goal is to refill your cells, oxide is close to the worst thing you can buy, no matter what the milligram number on the front says.
Magnesium L-threonate — marketed hard on the claim that it uniquely raises magnesium in the brain. Be careful here. That claim originates largely in rodent work (Slutsky and colleagues, Neuron, 2010) plus a small number of small human trials, often industry-linked. The idea is biologically plausible and it may well turn out to be right — but “plausible, with limited evidence” is not the same as “proven,” and threonate is several times the price. It is a reasonable thing to try. It is not an established fact you should pay a premium for on faith.
Magnesium sulfate (Epsom salts) — the injectable form is a genuine emergency drug. The bath is pleasant, but the evidence that meaningful magnesium crosses intact skin is thin. Enjoy the bath; do not count on it as a dose.

The diarrhoea ceiling — and the one group who must not do this

Magnesium has a built-in safety brake that almost no other supplement has. Any magnesium your gut does not absorb stays in the bowel, where it is osmotically active — it pulls water in behind it. Take too much and you get loose stools, then diarrhoea. That is not a side effect to be pushed through; it is a signal, and it arrives long before your blood level goes anywhere. It is why the tolerable upper limit for supplemental magnesium is set at 350 mg/day — that limit is about the bowel, not about toxicity, and it does not apply to magnesium from food. Press Too much on the animation and notice what does not happen: with working kidneys, serum magnesium barely twitches. The gut and the kidney handle it. Your bowel is the dose-limiter, and it is a reasonably honest one.

The exception is serious. The kidney is magnesium’s only meaningful exit. In advanced chronic kidney disease — and especially in anyone on dialysis — that exit is closed, and magnesium taken by mouth can accumulate to genuinely dangerous levels. Rising magnesium suppresses neuromuscular transmission: deep tendon reflexes fade first, then breathing and the heart rhythm are threatened. There are documented deaths from magnesium-containing laxatives and antacids in people with impaired kidneys. If you have advanced kidney disease, do not load magnesium without your nephrologist — and check whether the antacid or laxative you already take contains it.

For everyone else with working kidneys: eat the seeds and the greens, correct the drains if you have them, treat a normal serum magnesium as reassurance about your heart rather than a verdict on your cells, and if you supplement, pick a form designed to be absorbed rather than one designed to look impressive on a label.

↑ Back to the animation