How many ATP are produced in cellular respiration?
Stage-by-stage ATP accounting, and why older textbooks say 36 to 38
Matic Broz, PhD7 min
Complete aerobic respiration of one glucose molecule yields about 30 to 32 ATP in human cells under modern accounting. Four of those ATP are made directly, in glycolysis and the citric acid cycle. The other 26 to 28 come from oxidative phosphorylation, and whether the total is 30 or 32 depends on which shuttle carries electrons from the two NADH of glycolysis into the mitochondrion.
Older textbooks give 36 or 38 ATP. They assumed that each NADH makes 3 ATP and each FADH₂ makes 2, whole-number ratios that measurements of mitochondrial coupling have since revised to about 2.5 and 1.5. Both sets of totals are theoretical yields for complete oxidation, not a count of ATP measured from every glucose molecule.
How much ATP does each stage of cellular respiration produce?
Glycolysis splits glucose into two pyruvate molecules in the cytosol and nets 2 ATP. Pyruvate oxidation in the mitochondrial matrix turns both pyruvates into acetyl-CoA, releasing 2 carbon dioxide and making 2 NADH but no ATP. Two turns of the citric acid cycle release the remaining 4 carbon dioxide and make 2 GTP or ATP, 6 NADH and 2 FADH₂.[1]
The 10 NADH and 2 FADH₂ carry most of glucose's energy to the inner mitochondrial membrane, where oxidative phosphorylation converts it into ATP. Molecular Biology of the Cell puts the complete yield at about 30 ATP per glucose, compared with 2 from glycolysis alone.[2]
| Stage | Location | How ATP forms | Modern accounting | Older accounting |
|---|---|---|---|---|
| Glycolysis | Cytosol | Substrate-level phosphorylation | 2 | 2 |
| Pyruvate oxidation | Mitochondrial matrix | None directly | 0 | 0 |
| Citric acid cycle | Mitochondrial matrix | Substrate-level phosphorylation (GTP or ATP) | 2 | 2 |
| Oxidative phosphorylation | Inner mitochondrial membrane | ATP synthase | 26 or 28 | 32 or 34 |
| Total | 30 or 32 | 36 or 38 |
Glycolysis and the Krebs cycle together contribute only four ATP. Oxidative phosphorylation supplies the rest, about 87% of the total in either accounting, which makes it by far the most productive stage. Balanced as a whole, the pathway turns one glucose and six oxygen molecules into six carbon dioxide and six water molecules.[1]
How many ATP does each NADH and FADH₂ produce?
Under modern accounting, each NADH formed in the mitochondrial matrix supports about 2.5 ATP and each FADH₂ about 1.5 ATP.[2] Applied to the 12 electron carriers from one glucose, those values give the 26 to 28 ATP of oxidative phosphorylation.
| Carrier | Made in | Per glucose | ATP each, modern | ATP total, modern | ATP total, older |
|---|---|---|---|---|---|
| NADH | Glycolysis (cytosol) | 2 | 1.5 or 2.5 | 3 or 5 | 4 or 6 |
| NADH | Pyruvate oxidation | 2 | 2.5 | 5 | 6 |
| NADH | Citric acid cycle | 6 | 2.5 | 15 | 18 |
| FADH₂ | Citric acid cycle | 2 | 1.5 | 3 | 4 |
| Total | 12 | 26 or 28 | 32 or 34 |
These values are P/O ratios: ATP made per pair of electrons passed to an oxygen atom. Hinkle reviewed the main measurements made since 1937 and found that values of about 2.5 for NADH-linked substrates and 1.5 for succinate, which feeds electrons in at the FADH₂ level, fit most reports once their systematic errors are explained.[3]
The fractions follow from proton counting. Electrons from NADH enter the respiratory chain at complex I and drive about 10 protons across the inner membrane. Electrons from FADH₂ enter further along, at ubiquinone, and drive only about 6. ATP synthase spends those protons to make ATP, and moving phosphate and ADP into the matrix and ATP out costs more protons still.[4]
Structural work has since pinned down the price. Watt and colleagues found that the rotor ring of bovine ATP synthase has eight c subunits, so one full turn uses 8 protons to make 3 ATP, or 2.7 protons per ATP. Because c-subunit sequences are nearly identical across vertebrates, the same cost probably applies to human mitochondria.[5]
Mookerjee and colleagues added one proton per ATP for transport and recalculated the ceilings: P/O ratios of 2.73 for NADH-linked substrates and 1.64 for succinate. For a whole glucose molecule, that allows at most 33.45 ATP with the malate-aspartate shuttle or 31.81 with the glycerol 3-phosphate shuttle.[4]
The textbook 30 to 32 therefore sits just below the structural maximum. Living cells fall further below it, because some protons leak back across the membrane without passing through ATP synthase.[4]
Why do some sources say 30 ATP and others 32?
The two-ATP difference comes from the two NADH made in glycolysis. They form in the cytosol, but the inner mitochondrial membrane is impermeable to NADH and NAD⁺, so cells pass the electrons into the mitochondrion through one of two shuttles instead.[6]
The malate-aspartate shuttle carries the electrons in as malate, which is oxidized in the matrix to regenerate NADH. That NADH enters at complex I, so each glycolytic NADH keeps a value of about 2.5 ATP, or 5 for the pair, and the total is 32. The glycerol 3-phosphate shuttle hands the electrons to ubiquinone instead, bypassing complex I. Each NADH is then worth about 1.5 ATP, 3 for the pair, and the total drops to 30.[4]
Which shuttle dominates depends on the cell. The mitochondrial enzyme of the glycerol 3-phosphate shuttle is a tissue-specific component that standard textbook schemes of the respiratory chain leave out.[7] In the brain, the malate-aspartate shuttle is considered the main route, particularly in neurons.[6]
Two references on NCBI Bookshelf show how the choice changes the headline number. Molecular Biology of the Cell credits glycolytic NADH with 1.5 ATP and arrives at about 30 ATP per glucose.[2] StatPearls gives approximately 32, the figure that follows when both NADH keep their full value.[8] Neither is wrong; they describe the two shuttle routes.
Why do some sources say cellular respiration produces 36 or 38 ATP?
The 36 and 38 ATP totals come from older whole-number P/O ratios of 3 for NADH and 2 for FADH₂. Applied to the carriers in Table 2, they give 32 or 34 ATP from oxidative phosphorylation and 36 or 38 in total, with 38 assuming the malate-aspartate route and 36 the glycerol 3-phosphate route.
The integers were long read from early experiments as exact. Hinkle traced the apparent contradictions between studies to systematic measurement errors and showed that the fractional values follow from how many protons each step moves.[3] A 36 or 38 in a recent source usually signals older accounting rather than a different biology.
The yield is not universal across life, though. Fungi, bacteria and chloroplasts have ATP synthase rotor rings of 10 to 15 c subunits and need 3.3 to 5 protons per ATP, so the same electron flow buys less ATP than in animal mitochondria.[5]
What the yield means for a cell
Without oxygen, the respiratory chain stops and a human cell keeps only the 2 ATP of glycolysis, regenerating NAD⁺ through fermentation. Full aerobic oxidation returns about 15 times as much ATP from the same glucose, and Alberts and colleagues doubt that complex multicellular animals could run on glycolysis alone.[2]
Glucose is also not the richest fuel. Fatty acids enter the same citric acid cycle and respiratory chain as acetyl-CoA, NADH and FADH₂, and beta oxidation of a fatty acid yields several times more ATP per molecule than glucose under the same 2.5 and 1.5 accounting.[2]
