# The Pyruvate Dehydrogenase Complex: The One-Way Gate to the Cycle

> The pyruvate dehydrogenase complex is the one-way gate between glycolysis and the TCA cycle: it converts pyruvate to acetyl-CoA, releasing one CO2 and banking one NADH, and once a cell walks through it there is no walking back. One physical machine, three enzymes and five coenzymes, hands the carbon down an assembly line on a swinging arm without ever letting it escape into solution. Because the step is irreversible and sits at a metabolic crossroads, it is also one of the most heavily regulated reactions in the cell.

Course: Biochemistry (Central Metabolism). Level: university elective. About 31 minutes.
Lesson page (link this one when citing): https://stemhop.com/lessons/pyruvate-dehydrogenase/

Builds on: Glycolysis: The Investment-and-Payoff Pathway · Metabolism: Basic Concepts and Design · Basic Concepts of Enzyme Action

## What you'll learn

- Write and explain the net PDC reaction (pyruvate + CoA + NAD+ -> acetyl-CoA + CO2 + NADH + H+), and state why it is irreversible in animals and what that forces gluconeogenesis to do.
- Walk the four steps across the three enzymes and five coenzymes (TPP, lipoamide, CoA, FAD, NAD+), following the carbons and the electrons separately: name the step that releases CO2 (and why it removes no electrons) and identify E3 as the ONLY NADH-producing step.
- Describe substrate channeling and the lipoamide swinging arm, and explain why a single ~10-MDa complex is more efficient than three free enzymes.
- Predict how the two regulatory layers (product inhibition and covalent PDK/PDP phosphorylation, with the Ca2+ muscle axis) turn PDC on or off, and connect failures to beriberi, arsenic/mercury poisoning, and the Warburg effect.
- Explain why hydrolyzing acetyl-CoA releases about 31 kJ/mol (a thioester is far less resonance-stabilized than an ordinary oxygen ester), and use the energy ledger to show that the oxidation pays for both the NADH and the thioester, with enough left over, together with the loss of CO2, to make the gate one-way.

**Eat more sugar than you burn and your body stores the surplus as fat. Burn that fat later, and almost none of it can come back as sugar. Why does the road run only one way?**

The answer starts with one reaction. The pyruvate dehydrogenase complex (PDC) turns pyruvate, the three-carbon end of glycolysis, into acetyl-CoA, releasing a carbon as CO₂ and banking one NADH. Picture an airport: glycolysis in the cytosol is Terminal 1, the citric acid cycle in the mitochondrial matrix is Terminal 2, and PDC is a security gate between them that opens one way. Fat breaks down mostly to acetyl-CoA, which is already on the far side of that gate.

The net reaction is worth memorizing exactly:

> **Pyruvate + CoA + NAD⁺ → Acetyl-CoA + CO₂ + NADH + H⁺**

It is an **oxidative decarboxylation**: pyruvate (3 carbons) loses its carboxyl carbon as CO₂ and is oxidized to a 2-carbon acetyl group carried by coenzyme A. The two halves of that name happen at different moments. It runs in the mitochondrial matrix, which pyruvate enters through the mitochondrial pyruvate carrier.

*Figure (AI-generated illustration).* Where the gate sits. Pyruvate crosses the inner mitochondrial membrane through the mitochondrial pyruvate carrier; in the matrix the pyruvate dehydrogenase complex makes acetyl-CoA, CO2 and NADH. The reaction is irreversible in animals, and with no glyoxylate cycle, acetyl-CoA cannot yield net glucose; glycerol and propionyl-CoA from odd-chain fatty acids still can. Image: two panels show pyruvate crossing the inner mitochondrial membrane through the mitochondrial pyruvate carrier into the matrix, where the pyruvate dehydrogenase complex makes acetyl-CoA, CO2 and NADH; an inset shows fatty-acid-derived acetyl-CoA with the route back to pyruvate crossed out.

> **The one idea to carry through this lesson.** Follow two things through the machine: the carbons and the electrons. Pyruvate’s carboxyl carbon leaves as CO₂ at E1 *before* any oxidation happens. The two acetyl carbons that stay and the two electrons removed from them then ride the same swinging arm, and they get off at different stops: the carbons at E2 (onto CoA), the electrons at E3 (through FAD to NAD⁺).

This two-track picture explains the net reaction, the product brakes and the poisons. The phosphorylation switch at E1 is a separate layer on top.

*About the film at the top of the page:* the “three reactions” it names are steps 1 to 3 below, and its “step four” is our step 4. One correction to its wording: Thiamine pyrophosphate (TPP) grabs pyruvate first, and the CO₂ leaves from that grip.

## Following one pyruvate: two tracks on one arm

Keep the two tallies, **carbons** and **electrons**, as you go. **Oxidation** just means losing electrons; for a carbon, that shows up as trading a bond to hydrogen for a bond to oxygen (or, here, sulfur). Chemists rank a carbon’s oxidation level by those bonds: alcohol, then aldehyde, then acid, each with one more.

**TPP, the grabber on E1.** Thiamine pyrophosphate (made from vitamin B1) has an unusual carbon between a nitrogen and a sulfur in a small ring. It can drop its hydrogen, become negatively charged and grab pyruvate’s middle carbon. The ring then soaks up electrons, which lets the carboxyl group break away as CO₂.

**Lipoamide, the arm.** Lipoic acid is a short chain ending in a ring that contains two sulfur atoms bonded to each other (an S–S bond, a disulfide). Bolted to a lysine on E2 (by an amide bond), it becomes lipoamide, riding on a floppy lipoyl domain that jiggles at random among the three enzymes, all within reach. Give the S–S two electrons (with two protons) and it breaks into two –SH groups. So the arm can carry two things at once: an acetyl group on one sulfur, and two electrons in the broken S–S.

**Predict.** Step 1 removes pyruvate’s carboxyl carbon as CO₂. Does that step remove any electrons from the molecule?

> **Check your prediction.** No. CO₂ leaves its bonding electron pair behind (TPP’s ring soaks it up), so no electrons leave the molecule overall, and the two carbons left behind sit at the oxidation level of an aldehyde (acetaldehyde, in effect). Yeast’s pyruvate decarboxylase uses the same TPP chemistry but releases acetaldehyde for fermentation, with no NAD⁺ involved. The carbon is lost for good (it is gone as gas), but the oxidation, the main energy source, is still to come.

Now the walk. Three steps build the product, and one more resets the machine.

| Step | Where | Carbon track | Electron track |
| --- | --- | --- | --- |
| **1 · Decarboxylation** | E1, with TPP | Pyruvate’s carboxyl carbon (C1) leaves as **CO₂**. C2 and C3 stay on TPP as a two-carbon (hydroxyethyl) group. | Nothing removed yet. |
| **2 · Oxidative transfer** | E1, handing to the arm | The two carbons are oxidized to an **acetyl group** and hooked onto one sulfur of the arm (a thioester). TPP is free again. | **Two electrons removed**, and they break the arm’s S–S. The arm now carries both tracks. |
| **3 · Acetyl transfer** | E2, with CoA | The arm swings to E2, which moves the acetyl group onto CoA. **Acetyl-CoA leaves: the carbon track is finished.** | The two electrons are still on the arm, as two –SH groups. |
| **4 · Regeneration** | E3, with FAD, then NAD⁺ | — | The arm swings to E3. The electrons pass through E3’s bound FAD (FADH₂ for an instant) to NAD⁺, making **NADH**. The S–S is restored. **The electron track is finished.** |

**Predict.** The oxidation happens at E1, in step 2. So why is the NADH not made there?

> **Check your prediction.** Because E1 has no NAD⁺ site. The electrons it removes go onto the nearest acceptor, the arm’s disulfide, and they ride the arm until it reaches E3, the only enzyme that can hand them to NAD⁺. That is the exam trap in one line: **the oxidation is at E1, but the NADH is made at E3.** The FADH₂ along the way never leaves E3, which is why the net reaction lists NADH, not FADH₂.

Read down each column and the net reaction writes itself: three carbons in, one CO₂ and one acetyl-CoA out; two electrons out of the carbons, into one NADH. Step 4 makes no product carbon, but it has to happen: an arm still holding electrons cannot accept the next acetyl group.

*Figure (AI-generated illustration).* Two tracks on one arm (schematic). The carboxyl carbon leaves as CO2 at E1 before any electrons are removed. The two acetyl carbons and the two electrons taken from them then ride the same lipoamide arm and get off at different stops: the carbons at E2, onto CoA; the electrons at E3, through FAD to NAD+. The oxidation is at E1, but the NADH is made at E3. Image: a four-frame strip of the pyruvate dehydrogenase complex. Frame 1, at E1, one of pyruvate's three carbons leaves as CO2 and no electrons are removed. Frame 2, still at E1, the two remaining carbons and a pair of electrons both load onto the swinging lipoamide arm, opening its clasp. Frame 3, at E2, the two carbons leave on CoA as acetyl-CoA while the electrons stay on the arm. Frame 4, at E3, the electrons pass through FAD to make NADH and the arm's clasp closes again.

*Interactive on the lesson page:* Pyruvate dehydrogenase complex · entry plus four steps tap a step to reveal it Running total 0 NADH0 CO₂ regulated step 

Card 0 is the trip into the matrix; cards 1 to 4 are the four steps.

Count the coenzymes you just used: **TPP** (step 1), **lipoamide** (steps 2 to 4), **CoA** (step 3), **FAD** and **NAD⁺** (step 4). That is the mnemonic **“Tender Loving Care For Nancy,”** in the order the pyruvate meets them. Four come from vitamins you must eat: thiamine B1 (TPP, at E1), pantothenate B5 (CoA, at E2), riboflavin B2 (FAD, at E3) and niacin B3 (NAD⁺, at E3). Lipoic acid is the exception; the body makes it, so it is not a true vitamin.

No step uses molecular O₂. Pyruvate dehydrogenase complex (PDC) needs oxygen only indirectly: the electron transport chain must turn NADH back into NAD⁺, or step 4 runs out of acceptor and the relay stops.

### One machine, so nothing floats free

PDC is one of the largest enzyme assemblies in the cell, larger than a ribosome. E2 forms the core, and E1 and E3 dock onto it, close enough for the arm to reach every active site. So the intermediates never diffuse into solution, where they could drift away or be lost to side reactions. That is **substrate channeling**.

*Figure (AI-generated illustration).* One machine, three enzymes (schematic). E1 and E3 are drawn on the E2 core; in mammals E3 attaches through E3BP. The E2-tethered lipoamide arm takes the acetyl group from E1 to CoA at E2, then visits E3 for reoxidation. TPP works at E1, CoA at E2, FAD and NAD+ at E3. Block numbers are illustrative. Image: a round E2 core with E1 and E3 enzyme blocks drawn around it, a flexible lipoamide arm tethered to the core reaching toward the other enzymes, and labelled tokens for TPP, CoA, FAD and NAD+ on the enzymes where they work. Schematic; in mammals E3 attaches through E3BP.

## Where the energy goes

In animals this reaction is **irreversible**: no enzyme runs it backward, and acetyl-CoA cannot become pyruvate again. **The oxidation is the main energy source: it pays for the NADH and for the thioester, and what is left over, together with the loss of CO₂, makes the gate one-way.** To see it, keep the free-energy books (standard values, pH 7).

| Piece of the reaction | ΔG°′ (kJ/mol) |
| --- | --- |
| The oxidation: the two carbons go from aldehyde level to acid level, and the electrons end on NAD⁺ | about −50 |
| Storing the acetyl group as a thioester, instead of letting it go as free acetate | about +31 (a cost) |
| Losing CO₂ (found by difference, so approximate) | about −14 |
| **Overall** | **about −33** |

Building the thioester spends most of the oxidation’s drop, about 31 of the 50. What is left (about −19), plus about −14 from the CO₂ loss, gives about −33 kJ/mol: an equilibrium constant (K = e−ΔG°′/RT) of about 4 × 10⁵ at body temperature, so far toward the products that in practice the reaction runs only forward.

The +31 line is the same number you get back when you hydrolyze acetyl-CoA: about −31 kJ/mol, roughly ATP’s −30.5. Hydrolyzing an ordinary oxygen ester gives only about −21. This is not energy stored in the bond; breaking a bond always costs energy. Hydrolysis releases energy because the products, acetate and CoA-SH, are much more stable than the thioester, which gets little resonance stabilization: sulfur is larger than oxygen, so its orbitals overlap the carbonyl carbon poorly, whereas an oxygen ester’s lone pairs stabilize its C=O. CoA is the handle, not the valuable part. So acetyl-CoA is a wound spring, and the oxidation wound it. That stored potential lets acetyl-CoA drive the first step of the citric acid cycle and seed fatty-acid synthesis. Acetyl-CoA has two main fates: complete oxidation for ATP, or storage as fat. The liver can also turn it into ketone bodies, and it is the starting point for cholesterol.

**Predict.** Your body burns fat to acetyl-CoA. Can that acetyl-CoA make new glucose?

> **Check your prediction.** Not a net amount. It is on the far side of the gate and can never become pyruvate, and animals lack the glyoxylate cycle, the separate route plants and bacteria use to make new oxaloacetate (and so glucose) from acetyl-CoA. Gluconeogenesis must start on the near side: from pyruvate (made from lactate or amino acids such as alanine), with pyruvate carboxylase and PEPCK carrying it back to PEP, or from other entry points. That answers the opening puzzle. Two parts of a fat are the exceptions: its glycerol, and the propionyl-CoA made at the end of odd-chain fatty-acid breakdown.

## Two layers of regulation

A gate this important is heavily controlled, and pyruvate dehydrogenase complex (PDC) reads the cell’s **energy charge** (how much ATP it has relative to ADP and AMP) with the same logic as phosphofructokinase-1 (PFK-1) from the glycolysis chapter. The first layer is **product inhibition**.

**Predict.** Acetyl-CoA and NADH, the two products, both slow the complex, each by competing with the substrate it resembles. Which enzyme does each one slow?

> **Check your prediction.** Acetyl-CoA competes with CoA at E2, backing up the carbon track. NADH competes with NAD⁺ at E3, backing up the electron track. Either way the relay slows: “we have plenty, slow down.”

The second layer is **covalent**, and it acts at E1. A dedicated kinase, **PDK** (pyruvate dehydrogenase kinase), phosphorylates a serine on E1 and switches the complex OFF. A phosphatase, **PDP**, removes that phosphate and switches it back ON. Both are bound to the complex itself. Note the direction: here phosphorylation **inactivates**, the same convention as glycogen synthase. Much of the energy-charge sensing at this gate runs through PDK: ATP is its substrate and ADP competes with it, so PDK reads the ATP/ADP ratio.

| The cell signals | What happens | E1 | The gate |
| --- | --- | --- | --- |
| Plenty: high ATP/ADP ratio, high NADH, high acetyl-CoA | PDK is stimulated | phosphorylated | closed |
| Need: ADP and pyruvate; in working muscle, Ca²⁺ | ADP and pyruvate inhibit PDK, so PDP wins; Ca²⁺ activates PDP | dephosphorylated | open |

The plenty signals reinforce the product brake.

Try it below: pick a state (fed, sprinting, fasting) or set the effectors, and watch which way PDK and PDP flip the gate.

*Interactive on the lesson page:* Switching the PDH complex on and off

The cleanest illustration is muscle. A contracting muscle fills with **Ca²⁺**; some is taken up into the mitochondrial matrix, where it switches on the phosphatase (the PDP1 isoform). PDC switches on, and acetyl-CoA surges into the citric acid cycle exactly when the muscle needs ATP. The calcium that triggers contraction also unlocks the fuel.

## When the gate fails

Each failure below breaks one of the two tracks, and a stall anywhere stops the relay.

*Figure (AI-generated illustration).* Where the brakes and poisons hit (schematic). Each product slows the stop where it leaves: acetyl-CoA competes with CoA at E2 on the carbon track, NADH with NAD+ at E3 on the electron track. Without thiamine the carbon track never starts; arsenite locks the reduced arm so neither track can reset. The PDK/PDP phosphate switch on E1 (inset) is a separate layer on top. Image: the pyruvate dehydrogenase complex drawn as E1, E2 and E3 in a row, crossed by an upper carbon track ending in acetyl-CoA at E2 and a lower electron track ending in NADH at E3, both sharing one swinging arm. Markers show thiamine lack stopping the carbon track at E1, acetyl-CoA braking E2, NADH braking E3, and arsenite clamping the arm. A separate inset shows PDK adding a phosphate to E1 and PDP removing it.

**Beriberi** is B1 deficiency: no thiamine, no thiamine pyrophosphate (TPP), so the carbon track stops before it starts. Pyruvate and lactate pile up, and the nervous and cardiovascular systems suffer. TPP-dependent α-ketoglutarate dehydrogenase and transketolase fail too, and in the brain they may matter as much as pyruvate dehydrogenase complex (PDC). This is why a malnourished or alcohol-dependent patient should get thiamine before or with a glucose drip: in severe deficiency the sugar load can worsen lactic acidosis, and it has been reported to precipitate Wernicke’s encephalopathy. (Glucose for dangerously low blood sugar is not delayed to wait for thiamine.) Thiamine deficiency is also one part of refeeding syndrome, which is defined by falling blood phosphate, potassium and magnesium after feeding restarts.

**Predict.** Arsenite grips the arm’s two –SH groups, the state the arm is in after step 3. Which track stops, and why does the other stop too?

> **Check your prediction.** The electron track stops: the locked arm cannot hand its electrons to E3, so no NADH is made and the S–S is never restored. Then the carbon track stops too, because an arm that cannot be reset cannot accept the next acetyl group. The whole complex freezes. Arsenite binds closely spaced pairs of thiols; in lipoamide these are the sulfurs on carbons 6 and 8.

Mercury (Hg²⁺) also binds thiols such as lipoamide’s, though its toxicity involves many other targets.

**E3 deficiency** stops the electron track in at least three complexes at once, because PDC, α-ketoglutarate dehydrogenase and the branched-chain α-keto acid dehydrogenase share one E3 (it also serves the glycine cleavage system). Patients show lactic acidosis, often with raised branched-chain amino acids and α-ketoglutarate.

The **Warburg effect** in cancer is partly PDC regulation hijacked: in many tumors HIF-1α (a low-oxygen transcription factor that many tumors keep switched on) drives up PDK, E1 is held largely phosphorylated and off, pyruvate is shunted to lactate even when oxygen is abundant, and the tumor gets fast ATP plus glycolytic intermediates for biosynthesis. Other tumors keep substantial PDC activity and mitochondrial oxidation.

## Next: the same relay, again

With acetyl-CoA in hand, we are at the door of Terminal 2, where you will meet this machine again: α-ketoglutarate dehydrogenase runs the same E1 → arm → E2 → E3 relay with the same E3; its own E1 and E2, the group carried (succinyl instead of acetyl) and the build of its core differ. The phosphorylation switch is not shared: α-ketoglutarate dehydrogenase is not controlled by phosphorylation. The next chapter runs that acetyl group through the citric acid cycle, where each turn squeezes out CO₂ and GTP plus the NADH and FADH₂ that feed the electron transport chain.

## Going deeper

### The energy ledger in numbers

The −50 kJ/mol comes from the reduction potentials of the acetate/acetaldehyde couple (−0.581 V) and NAD⁺ (−0.320 V). The +31.4 thioester cost is 62% of that. The overall −33.4 kJ/mol agrees, within a few kJ, with a cross-check from the acetyl-CoA + CO₂/pyruvate couple (E°′ about −0.48 to −0.50 V, giving −31 to −35 kJ/mol). The −14 for losing CO₂ is found by difference and mixes tables, so it is approximate; the conclusion that the drop comes mainly from the oxidation, with the CO₂ loss contributing the rest, does not hang on it.

### Size and docking of the complex

Pyruvate dehydrogenase complex (PDC) is roughly 10 megadaltons in eukaryotes, against about 4 for a eukaryotic ribosome. E2 forms a core of 60 subunits arranged as a pentagonal dodecahedron; in mammals some of those positions are taken by a related protein, E3BP. Some bacteria, such as *E. coli*, build a 24-subunit cube instead, which is the film’s “24 or 60.” E1 docks onto subunit-binding domains of E2, and E3 onto those of E3BP (in bacteria, E2 binds both). The lipoamide arm, carried on a mobile lipoyl domain, is about 14 Ångströms long.

### Hyperglycemia and superoxide

This concerns the electron transport chain, not PDC. In hyperglycemia, one proposed model holds that excess fuel overloads the chain so that more electrons leak out as superoxide. Oxidative stress is thought to contribute to diabetic neuropathy, but its sources are still debated and other pathways are involved too.

---

From STEMhop (https://stemhop.com), free lessons by Christopher Deppmann. Lesson text © the authors. Practice questions: https://stemhop.com/lessons/pyruvate-dehydrogenase/
