Glycolysis: Preparatory Phase

A glucose molecule arriving in the cytosol is chemically inert. It can sit unreacted for hours. Yet within milliseconds of entering a starving muscle cell, it is committed. It gets trapped, bent, and split in two. The five reactions that accomplish this are hexokinase, phosphoglucose isomerase, phosphofructokinase-1, aldolase, and triose phosphate isomerase. They spend two molecules of ATP before any energy has been recovered. This looks like poor accounting until you see what the investment buys. It buys irreversible commitment of glucose to catabolism. It buys activation of an otherwise unreactive carbon skeleton. And it buys conversion of a six-carbon sugar into a symmetric pair of phosphorylated three-carbon units. These units can then be oxidized in parallel by the payoff phase. This preparatory phase is where the logic of Glycolysis: Overview and Strategy becomes concrete chemistry. Every design principle mentioned there has a specific enzyme mechanism behind it here. These principles include substrate trapping by phosphorylation and use of ATP as a group-transfer reagent. They also include control of flux at essentially irreversible steps. Understanding how hexokinase transfers a phosphoryl group matters. So does understanding why isomerization must precede the aldol cleavage. So does understanding why the cleavage produces two different trioses that must be interconverted. Together, these turn glycolysis from a memorized list into a sequence of chemically necessary moves. The big picture is a carbon-accounting and energy-accounting story running side by side. Carbon: glucose (C6) is phosphorylated, isomerized to a ketose, and phosphorylated again. It is then cleaved at C3–C4 into dihydroxyacetone phosphate and glyceraldehyde-3-phosphate. One glyceraldehyde-3-phosphate forms directly; the other forms via isomerization of the DHAP. Energy: two ATP are consumed, at hexokinase and phosphofructokinase-1. This investment is repaid four-fold in the payoff phase. Control: three of the five reactions are essentially irreversible under cellular conditions. These are hexokinase, PFK-1, and by extension the whole flux. It is exactly these reactions that are allosterically regulated. This pattern holds true for every pathway you will meet later in metabolism. This article works through each of the five reactions mechanistically. It covers the in-line phosphoryl transfer at hexokinase, with its induced-fit closure. It covers the acid–base isomerizations catalyzed by phosphoglucose isomerase and triose phosphate isomerase, via cis-enediol(ate) intermediates. It covers the second ATP-dependent phosphorylation at PFK-1. It also covers the retro-aldol C–C bond cleavage carried out by aldolase. This cleavage proceeds through an imine (Schiff base) or metal-stabilized enolate mechanism. It ends by setting up exactly where the carbon and phosphoryl-bond energy go next. That story continues in Glycolysis: Payoff Phase and ATP Generation.

Key takeaways

Hexokinase phosphorylates glucose using ATP·Mg²⁺, via in-line nucleophilic attack of the C6-OH oxygen on the ATP γ-phosphorus, with induced-fit domain closure excluding water.
Phosphoglucose isomerase converts glucose-6-phosphate (aldose) to fructose-6-phosphate (ketose) through a cis-enediol(ate) intermediate stabilized by active-site acid/base residues.
Phosphofructokinase-1 (PFK-1) performs the second ATP-dependent phosphorylation, is essentially irreversible, and is the primary flux-control point of glycolysis.
Aldolase cleaves fructose-1,6-bisphosphate between C3 and C4 by a retro-aldol mechanism — Schiff-base (Class I, animals) or metal-ion-stabilized enolate (Class II, bacteria/fungi) — giving DHAP and G3P.
Triose phosphate isomerase (TIM) interconverts DHAP and G3P via a cis-enediolate, achieving catalytic perfection (diffusion-limited kcat/KM), so all six original carbons proceed as G3P.
Two ATP are spent in this phase (hexokinase, PFK-1); the phase ends with a net carbon "doubling" of downstream units without yet recovering any energy. Key idea: glucose is spent, twisted, and split — not oxidized — in the preparatory phase; irreversible phosphorylations trap and commit the carbon skeleton before any ATP or NADH is harvested.

Where this lives today

Hexokinase and PFK-1 are not museum pieces — they sit at the center of modern metabolic medicine and drug design. Tumor cells characteristically overexpress hexokinase II and anchor it to mitochondrial voltage-dependent anion channels (VDAC), coupling glycolytic phosphorylation directly to mitochondrially generated ATP and simultaneously suppressing apoptotic signaling; this is part of the molecular basis of the Warburg effect exploited diagnostically in ¹⁸F-fluorodeoxyglucose PET imaging, where a non-metabolizable glucose analog is phosphorylated by hexokinase and trapped (exactly like G6P) inside metabolically hyperactive tumor tissue, lighting it up on a scan. Small-molecule hexokinase II inhibitors (e.g., 2-deoxyglucose derivatives, lonidamine analogs) are actively investigated as anticancer agents specifically because they exploit this trapping chemistry. Liver glucokinase (hexokinase IV) is a direct pharmacological target in type 2 diabetes: glucokinase activators are designed to shift its regulatory conformational equilibrium, boosting hepatic glucose uptake and insulin secretion signaling at physiological glucose concentrations. Fructose-2,6-bisphosphate, the master allosteric activator of PFK-1, is synthesized and degraded by the bifunctional enzyme PFK-2/FBPase-2, whose two activities are reciprocally controlled by glucagon/insulin-driven phosphorylation — a textbook case of hormonal control converging on the exact enzyme highlighted in this pathway, and central to the moment-to-moment balance between glycolysis and gluconeogenesis in liver. Aldolase deficiency (aldolase A, in red blood cells and muscle) causes hereditary hemolytic anemia and exercise intolerance, directly demonstrating in human disease why the retro-aldol step cannot be bypassed. Triose phosphate isomerase deficiency is a severe, often fatal, autosomal recessive disorder, underscoring that this step, though "merely" an isomerization, is non-negotiable for full carbon flux. Industrially, immobilized aldolases are used in stereoselective organic synthesis to build complex sugars and sugar-mimetic pharmaceuticals, directly exploiting the same retro-aldol/aldol chemistry described mechanistically here.

Ask the tutor: Why does the cell tolerate a positive ΔG°′ at the aldolase step instead of evolving a thermodynamically favorable cleavage enzyme directly?

The physical picture — trapping, bending, and cutting a sugar

Glucose entering a cell faces a phosphate group waiting to be attached at C6. This single covalent modification does two separate jobs that are easy to conflate. First, it converts glucose into glucose-6-phosphate (G6P), a species with no significant transporter in the plasma membrane — the negatively charged phosphate cannot cross the bilayer or be recognized by GLUT transporters, so G6P is kinetically trapped inside the cell. Second, phosphorylation destabilizes the molecule relative to further metabolism: the C6–O–PO₃²⁻ ester is a handle that later enzymes (phosphofructokinase, aldolase) can grip and use to enforce ring-opening, isomerization, and eventual cleavage. The overall preparatory-phase transformation is

\text{glucose} + 2\,\text{ATP} \longrightarrow \text{fructose-1,6-bisphosphate} + 2\,\text{ADP} + 2\,\text{H}^+

followed by symmetric cleavage,

\text{fructose-1,6-bisphosphate} \longrightarrow \text{DHAP} + \text{G3P (via TIM: DHAP} \rightleftharpoons \text{G3P)}

so that ultimately two molecules of glyceraldehyde-3-phosphate emerge from one glucose, each carrying three of the original six carbons forward into the payoff phase. Two chemical problems must be solved along the way, and each has its own enzyme. Problem one: an aldose (glucose, with its carbonyl at C1) cannot be cleaved symmetrically at C3–C4 by a retro-aldol reaction, because a retro-aldol cleavage needs a carbonyl positioned so that the resulting alkoxide/enolate is stabilized — that requires a ketose with the carbonyl at C2. Hence phosphoglucose isomerase must convert the aldose G6P into the ketose fructose-6-phosphate (F6P) before cleavage is chemically feasible. Problem two: the retro-aldol cleavage of fructose-1,6-bisphosphate produces two different three-carbon fragments — dihydroxyacetone phosphate (a ketose) and glyceraldehyde-3-phosphate (an aldose) — but every subsequent glycolytic enzyme is built to act only on the aldose. Triose phosphate isomerase (TIM) solves this by rapidly and reversibly converting DHAP to G3P, effectively rescuing the "wrong half" of the split sugar.

Figure

Carbon flow and phosphorylation sites through the preparatory phase

Figure

Active-site geometry: hexokinase induced fit vs. aldolase Schiff-base cleavage

The core ideas — governing relationships

The energetics of each step, and of the phase as a whole, are best organized around standard free energies and the actual free energies operating inside the cell, since these differ enormously for the "irreversible" steps.

\Delta G = \Delta G^{\circ\prime} + RT\ln Q

where: $\Delta G^{\circ\prime}$ is the standard transformed free-energy change (pH 7, 1 M standard states, Mg²⁺ present), $Q$ is the actual mass-action ratio of products to reactants in the cell, $R = 8.314\ \text{J mol}^{-1}\text{K}^{-1}$ , $T$ is absolute temperature.

Reading: a reaction with unfavorable $\Delta G^{\circ\prime}$ can still run forward in vivo if the cell keeps $Q$ far below $K_{eq}$ by continuously removing products — this is exactly the situation for phosphoglucose isomerase and aldolase, whose $\Delta G^{\circ\prime}$ values are near zero or even positive, yet which run forward because downstream reactions consume their products. For the two ATP-coupled phosphorylations (hexokinase, PFK-1), the relevant quantity is the free energy of the coupled reaction,

\Delta G^{\circ\prime}_{\text{coupled}} = \Delta G^{\circ\prime}_{\text{phosphoryl transfer}} + \Delta G^{\circ\prime}_{\text{ATP hydrolysis}}

where: $\Delta G^{\circ\prime}_{\text{ATP hydrolysis}} \approx -30.5\ \text{kJ/mol}$ is used as the "energy currency" contribution, and $\Delta G^{\circ\prime}_{\text{phosphoryl transfer}}$ is the (often unfavorable) free energy of transferring a phosphoryl group to the sugar hydroxyl in the absence of ATP.

Reading: ATP does not "contain" the energy that appears at the sugar — rather, the phosphoanhydride bond of ATP is broken in the same elementary step (in-line transfer) that forms the new sugar phosphoester bond, so the two processes are thermodynamically summed, not sequentially released and recaptured. For the aldol cleavage itself, the key relationship is the Haldane relation connecting kinetics and thermodynamics of a freely reversible enzyme:

K_{eq} = \frac{k_{cat,f}\,K_{M,r}}{k_{cat,r}\,K_{M,f}}

where: $k_{cat,f}, k_{cat,r}$ are forward and reverse turnover numbers, $K_{M,f}, K_{M,r}$ are the corresponding Michaelis constants for forward (cleavage) and reverse (condensation) directions.

Reading: aldolase's $K_{eq} \approx 10^{-4}$ M favors condensation, yet cleavage dominates in vivo because triose phosphate isomerase and downstream glycolytic enzymes keep the triose-phosphate concentrations low, pulling $Q$ below $K_{eq}$ and making $\Delta G$ negative for cleavage in the cell.

Derivation AHexokinase: in-line phosphoryl transfer and the energetics of glucose trapping

Goal: show mechanistically why hexokinase transfers the γ-phosphate of ATP to glucose C6-OH via associative in-line attack, and quantify why this step is effectively irreversible in the cell.

Step 1 — substrate binding and induced fit. Free hexokinase exists in an open conformation with two lobes separated by a cleft. Glucose binding triggers a large conformational change (~30–40° hinge rotation between the two domains, first visualized by Koshland-inspired crystallography of yeast hexokinase) that closes the cleft around the sugar. This "induced fit" excludes bulk water from the active site. Its chemical purpose is critical: without it, ATP's γ-phosphate would be equally exposed to attack by water (giving simple ATP hydrolysis, wasting the phosphoryl group) as to attack by the C6 hydroxyl. Closing the lobes positions the C6-OH oxygen of glucose directly opposite the γ-phosphorus of Mg·ATP and physically occludes water.

Step 2 — deprotonation of the C6 hydroxyl. An active-site base (in yeast hexokinase, an aspartate carboxylate positioned near C6-OH) partially deprotonates the sugar hydroxyl as it approaches phosphorus, increasing its nucleophilicity. The developing alkoxide oxygen lone pair is oriented in-line with the γ-phosphorus–oxygen(bridging)-to-β,γ bond of ATP.

Step 3 — associative in-line attack at phosphorus. The C6 oxygen attacks the γ-phosphorus of ATP from the side opposite the leaving β,γ-bridging oxygen, forming a trigonal-bipyramidal pentacoordinate phosphorane-like transition state with partial bonds to both the incoming glucose oxygen and the leaving ADP oxygen. Mg²⁺ coordinates the β- and γ-phosphates throughout, neutralizing developing negative charge and orienting the triphosphate chain.

Step 4 — collapse to products. The transition state collapses as the O–Pγ bond to ADP fully breaks and the new O–Pγ bond to glucose C6 fully forms, releasing ADP and glucose-6-phosphate, both still Mg²⁺-coordinated. The reaction proceeds with inversion of configuration at phosphorus, the classic signature (established using chiral ¹⁶O/¹⁷O/¹⁸O-phosphorothioate substrates) of a single in-line associative mechanism rather than a two-step phosphoenzyme intermediate.

Step 5 — energetics. The uncoupled phosphoryl-transfer reaction

\text{glucose} + P_i \longrightarrow \text{glucose-6-phosphate} + H_2O

has $\Delta G^{\circ\prime} \approx +13.8\ \text{kJ/mol}$ (unfavorable — glucose-6-phosphate has lower phosphoryl-transfer potential than $P_i$ would suggest, but forming the ester from free sugar and inorganic phosphate is still endergonic because entropy is lost and no anhydride bond is broken). Coupling to ATP hydrolysis,

\text{ATP} + H_2O \longrightarrow \text{ADP} + P_i \qquad \Delta G^{\circ\prime} \approx -30.5\ \text{kJ/mol}

gives the overall hexokinase reaction

\Delta G^{\circ\prime}_{HK} = (+13.8) + (-30.5) = -16.7\ \text{kJ/mol}

Step 6 — why this makes the step irreversible in vivo. Cellular ATP/ADP ratios are typically ~10:1 or higher, and G6P is continuously consumed by phosphoglucose isomerase, keeping $[G6P]$ low. Both effects make $Q \ll K_{eq}$ , so the actual $\Delta G$ in the cell (typically around $-33$ kJ/mol) is even more negative than $\Delta G^{\circ\prime}$ . This large, essentially unidirectional free-energy drop is why hexokinase is one of the three flux-controlling, allosterically regulated steps of glycolysis (inhibited by its own product, G6P, in most isozymes — a classic case of feedback inhibition preventing futile phosphorylation when downstream flux is already saturated).

Go deeper · Derivation B · 3 minAldolase: retro-aldol cleavage of fructose-1,6-bisphosphate and the role of isomerization

Derivation BAldolase: retro-aldol cleavage of fructose-1,6-bisphosphate and the role of isomerization

Goal: derive why glucose must be isomerized to a ketose before cleavage, and show electron-arrow by electron-arrow how aldolase splits the C3–C4 bond of fructose-1,6-bisphosphate.

Step 1 — why an aldose cannot be cleaved directly. A retro-aldol reaction breaks a C–C bond β to a carbonyl, generating an enolate (or its equivalent) on one fragment and a carbonyl or a stabilized carbanion-equivalent on the other. For cleavage between C3 and C4 of a hexose to give two three-carbon units, the carbonyl must be positioned so that C3–C4 bond cleavage places negative charge alpha to a carbonyl. In glucose (aldose, carbonyl at C1), cleaving C3–C4 would leave the developing negative charge at C4 with no adjacent carbonyl to stabilize it as an enolate — chemically this bond is not activated for retro-aldol chemistry. In fructose (ketose, carbonyl at C2), however, C3–C4 cleavage places the new carbanion/alkoxide directly alpha to the C2 carbonyl of the C1–C3 fragment, exactly the stabilization a retro-aldol requires. This is the chemical reason phosphoglucose isomerase (PGI) must convert G6P to F6P before phosphofructokinase and aldolase can act — it is not a side reaction but a structural prerequisite for symmetric cleavage.

Step 2 — the isomerization mechanism (PGI). PGI opens the pyranose ring of G6P (via general acid catalysis by an active-site His/Glu protonating the ring oxygen as it leaves), producing the open-chain aldose. A general base (Glu/Lys) abstracts the C2 proton, generating a cis-enediolate spanning C1–C2 stabilized by an adjacent metal ion or positively charged residue (analogous to the oxyanion stabilization seen in triose phosphate isomerase). Reprotonation occurs at C1 instead of C2, converting the aldose carbonyl (C1) into an alcohol and installing a new carbonyl at C2 — net result: aldose become ketose without any change in phosphorylation state or carbon connectivity. The enzyme then closes the five-membered furanose ring to release fructose-6-phosphate.

Step 3 — second phosphorylation (PFK-1). Phosphofructokinase-1 transfers the γ-phosphate of a second ATP to the C1-OH of fructose-6-phosphate by the same in-line associative mechanism used by hexokinase, generating fructose-1,6-bisphosphate (F1,6BP). This step has $\Delta G^{\circ\prime} \approx -14.2\ \text{kJ/mol}$ and, like hexokinase, is pulled further negative in vivo by low ATP/AMP-driven allosteric activation and product removal; it is the single most important control point of glycolysis, allosterically inhibited by ATP and citrate (signaling ample energy/citric-acid-cycle intermediates) and activated by AMP and fructose-2,6-bisphosphate (signaling energy need).

Step 4 — retro-aldol cleavage: Schiff-base pathway (Class I aldolase, animals/plants). An active-site lysine ε-amino group attacks the C2 carbonyl of F1,6BP, forming a carbinolamine that dehydrates to a protonated Schiff base (iminium ion) at C2. This electron-withdrawing iminium is the functional equivalent of a carbonyl for stabilizing negative charge — it is the reason the enzyme can act on a substrate whose true carbonyl has been temporarily "hidden" as an imine.

Step 5 — C3–C4 bond heterolysis. With the iminium in place, the C3–C4 σ bond breaks heterolytically: the bonding electron pair moves onto C4, generating an alkoxide at C4 (stabilized by an active-site Zn²⁺ or, in Class I enzymes, by hydrogen bonding) as the leaving group of the top fragment; simultaneously the iminium nitrogen's positive charge is neutralized as the electrons flow from the broken C–C bond through C3 into a resonance-stabilized enamine at C2–C3 (Lys–N=C2–C3, an enamine/eneamine tautomer of the iminium). This releases glyceraldehyde-3-phosphate (the C4–C5–C6 fragment, now a free aldose with its original C1 hydroxyl becoming its aldehyde-adjacent carbon numbering as C1 of G3P) while the C1–C2–C3 fragment remains bound to the enzyme as the resonance-stabilized enamine.

Step 6 — protonation and hydrolysis. The enamine is protonated at C3, regenerating the iminium at C2, which is then hydrolyzed (attack of active-site water at the iminium carbon, followed by carbinolamine collapse) to release free dihydroxyacetone phosphate (DHAP) and regenerate the active-site lysine for another catalytic cycle.

Step 7 — Class II aldolases (bacteria, fungi). Instead of a lysine-Schiff base, Class II aldolases use a catalytic Zn²⁺ ion coordinated at the active site to polarize the C2 carbonyl directly and stabilize the developing C3–C4 cleavage enolate as a metal-bound enolate, without any covalent enzyme-substrate intermediate. Mechanistically this is a metal-Lewis-acid-catalyzed retro-aldol rather than an amine-catalyzed one, but the bond broken and the electronic logic (stabilizing negative charge alpha to the carbonyl) are identical.

Step 8 — energetics and the pull of downstream reactions. The aldolase reaction

\text{F1,6BP} \longrightarrow \text{DHAP} + \text{G3P}

has $\Delta G^{\circ\prime} \approx +23.8\ \text{kJ/mol}$ (strongly unfavorable for cleavage under standard 1 M conditions) and $K_{eq} \approx 10^{-4}\ \text{M}$ . In the cell, the concentrations of DHAP and G3P are kept low by rapid isomerization (TIM) and by the payoff-phase enzyme glyceraldehyde-3-phosphate dehydrogenase, which consumes G3P continuously. This keeps the mass-action ratio $Q = [\text{DHAP}][\text{G3P}]/[\text{F1,6BP}]$ well below $K_{eq}$ , making the actual in vivo $\Delta G$ close to zero or modestly negative — illustrating again that near-equilibrium, thermodynamically "uphill" steps can still carry substantial net flux if the cell maintains disequilibrium by consuming products downstream.

Worked examples

Example 1 — Foundational: free energy of hexokinase in resting muscle. Given approximate resting-muscle cytosolic concentrations $[\text{glucose}] = 5.0\ \text{mM}$ , $[\text{ATP}] = 3.0\ \text{mM}$ , $[\text{ADP}] = 0.020\ \text{mM}$ (note the deliberately non-round, asymmetric numbers), $[\text{G6P}] = 0.083\ \text{mM}$ , and taking $\Delta G^{\circ\prime}_{HK} = -16.7\ \text{kJ/mol}$ at $T = 310\ \text{K}$ , compute the actual $\Delta G$ . The reaction quotient is

Q = \frac{[\text{G6P}][\text{ADP}]}{[\text{glucose}][\text{ATP}]} = \frac{(0.083\times10^{-3})(0.020\times10^{-3})}{(5.0\times10^{-3})(3.0\times10^{-3})}

Q = \frac{1.66\times10^{-6}}{1.5\times10^{-5}} = 0.111

\Delta G = \Delta G^{\circ\prime} + RT\ln Q = -16{,}700 + (8.314)(310)\ln(0.111)\ \text{J/mol}

RT\ln Q = 2577\times(-2.198) = -5665\ \text{J/mol}

\Delta G = -16{,}700 - 5665 = -22{,}365\ \text{J/mol} \approx -22.4\ \text{kJ/mol}

Check (limiting case): if $Q \to K_{eq}$ (i.e., the cell allowed the reaction to approach equilibrium, $\Delta G^{\circ\prime} = -RT\ln K_{eq}$ ), then $\Delta G \to 0$ and the flux would stall — this never happens physiologically for hexokinase precisely because product (G6P) is constantly drained by phosphoglucose isomerase and ATP is constantly regenerated, keeping $Q$ far from $K_{eq}$ and $\Delta G$ strongly negative, consistent with hexokinase being a one-way, flux-generating step.

Example 2 — Applied: how much of the "isomerization step" free energy is actually near zero, and what this implies for enzyme design. Phosphoglucose isomerase has $\Delta G^{\circ\prime} \approx +1.7\ \text{kJ/mol}$ (close to thermoneutral) at 298 K. Suppose in a hepatocyte $[\text{F6P}]/[\text{G6P}] = 0.30$ (asymmetric, physiologically reasonable — F6P is the minority species at equilibrium-like ratios). Compute $\Delta G$ and comment on reversibility.

\Delta G = \Delta G^{\circ\prime} + RT\ln\!\left(\frac{[\text{F6P}]}{[\text{G6P}]}\right) = 1700 + (8.314)(298)\ln(0.30)

RT\ln(0.30) = 2478\times(-1.204) = -2983\ \text{J/mol}

\Delta G = 1700 - 2983 = -1283\ \text{J/mol} \approx -1.3\ \text{kJ/mol}

This is small in magnitude (comparable to $RT \approx 2.5$ kJ/mol at 298 K), confirming that PGI operates very close to equilibrium in the cell — consistent with its role as a freely reversible "conduit" step whose net direction is dictated entirely by whichever neighboring irreversible reaction (hexokinase feeding it, or PFK-1 draining it) currently dominates flux, not by any intrinsic thermodynamic bias of its own.

Limiting case: if $[\text{F6P}] = [\text{G6P}]$ (ratio = 1), $\ln(1) = 0$ and $\Delta G = \Delta G^{\circ\prime} = +1.7$ kJ/mol exactly — mildly unfavorable for the forward (G6P→F6P) direction at truly equal concentrations, which is why cells must keep $[\text{F6P}] < [\text{G6P}]$ (as in the 0.30 ratio used above) for net forward flux, achieved by PFK-1 continuously consuming F6P.

Example 3 — Exam/competition level: using the Haldane relation to find a missing kinetic constant for aldolase. Suppose aldolase has been characterized with the following (illustrative, literature-consistent order-of-magnitude) constants: forward (cleavage) $k_{cat,f} = 8.0\ \text{s}^{-1}$ , $K_{M,f}(\text{F1,6BP}) = 8.0\times10^{-6}\ \text{M}$ ; reverse (condensation) $k_{cat,r} = 20\ \text{s}^{-1}$ , and the reaction has $K_{eq} = 1.0\times10^{-4}\ \text{M}$ for the cleavage direction (dimensioned in M because DHAP + G3P → F1,6BP is a second-order combination). Find $K_{M,r}$ (the Michaelis constant for the reverse reaction, effectively for the product of [DHAP][G3P] approach, treated here as a lumped constant for illustration). Rearranging the Haldane relation,

K_{eq} = \frac{k_{cat,f}\,K_{M,r}}{k_{cat,r}\,K_{M,f}} \quad\Longrightarrow\quad K_{M,r} = \frac{K_{eq}\,k_{cat,r}\,K_{M,f}}{k_{cat,f}}

K_{M,r} = \frac{(1.0\times10^{-4})(20)(8.0\times10^{-6})}{8.0} = \frac{1.6\times10^{-8}}{8.0} = 2.0\times10^{-9}\ \text{M}^{?}

(units here are schematic, since the reverse reaction is bimolecular in substrates — the pedagogical point is the algebraic manipulation and the internal consistency check, not a literal reported constant). Rearranged this way, note that $K_{M,r}$ came out numerically tiny, meaning the enzyme has very high apparent affinity for its (bimolecular) products in the reverse, condensation direction — consistent with condensation being the thermodynamically favored direction ( $K_{eq} = 10^{-4}$ M for cleavage means the equilibrium lies far toward F1,6BP), even though cleavage is what dominates flux physiologically because cellular G3P and DHAP are kept low by downstream consumption.

Sanity/limiting check: if $k_{cat,f} = k_{cat,r}$ and $K_{M,f} = K_{M,r}$ (a symmetric, achiral-like idealization), the Haldane relation collapses to $K_{eq} = 1$ , which is manifestly not the real aldolase case — a useful reminder that the strongly asymmetric, non-unity $K_{eq} \approx 10^{-4}$ M observed experimentally requires correspondingly asymmetric kinetic constants between the forward and reverse directions, exactly the kind of "non-special-case" numbers used above.

Limiting cases, scope & common traps

Situation	What happens	Why
$[\text{ATP}]/[\text{ADP}]$ very high (energy-replete cell)	PFK-1 allosterically inhibited by ATP; flux through preparatory phase throttled	ATP binds a regulatory site distinct from the catalytic site, lowering affinity for F6P
$[\text{AMP}]$ rises (energy-depleted cell)	PFK-1 activated; hexokinase indirectly favored as G6P is drained faster	AMP relieves ATP inhibition allosterically; mass-action pulls hexokinase forward
G6P accumulates (e.g., glycogen breakdown surge)	Hexokinase (not glucokinase) product-inhibited	G6P competes at/near the active site region of hexokinase isozymes I–III
Liver glucokinase (hexokinase IV) instead of hexokinase	High $K_M$ (~10 mM), not inhibited by G6P, cooperative kinetics	Liver must respond to blood glucose swings without being switched off by its own product
Fructose-2,6-bisphosphate elevated	PFK-1 strongly activated even with ATP present	F2,6BP is a separate, dedicated allosteric activator made by PFK-2, overriding ATP inhibition
Aldolase substrate concentration extremely low	Reaction can reverse toward condensation	$K_{eq} \approx 10^{-4}$ M favors F1,6BP; low triose-phosphate levels needed to sustain forward cleavage flux
TIM inhibited or absent	Only half of glucose carbon reaches G3P; DHAP accumulates uselessly	TIM operates at the diffusion limit ( $k_{cat}/K_M \approx 10^{8}$ – $10^{9}\ M^{-1}s^{-1}$ ) precisely because failing to rescue DHAP would halve glycolytic ATP yield

The chemist behind it

Origins · 1 minOtto Fritz Meyerhof and how this was discovered

Otto Fritz Meyerhof

1884–1951 · Elucidation of the glycolytic pathway and muscle biochemistry

Meyerhof, a German physician-turned-biochemist, shared the 1922 Nobel Prize in Physiology or Medicine (with Archibald Hill) for relating heat production and lactic acid formation in contracting muscle to a definable chemical pathway. Working through the 1920s–1930s in Kiel and later Heidelberg, Meyerhof's laboratory — together with contributions from Gustav Embden, whose name is jointly attached to the pathway as "Embden–Meyerhof" glycolysis — established that glucose breakdown to lactate proceeds through a defined sequence of phosphorylated intermediates rather than by a single uncontrolled combustion. Meyerhof emigrated to the United States in 1940 to escape Nazi persecution, continuing enzymological work at the University of Pennsylvania. The specific enzymes of the preparatory phase — hexokinase, phosphofructokinase, aldolase — were characterized piecemeal over subsequent decades by Carl and Gerty Cori, Jacob Parnas, and others, but the pathway's overall architecture, including the recognition that a hexose is split into two interconvertible trioses, traces to Meyerhof's synthesis of fermentation and muscle-glycolysis biochemistry into one coherent scheme.

History

- 1897: Eduard Buchner shows cell-free yeast extract ferments sugar, founding biochemical enzymology.

- 1905: Arthur Harden and William Young discover that fermentation requires inorganic phosphate, first hinting at phosphorylated intermediates.

- 1918–1933: Gustav Embden and Otto Meyerhof map the sequence of phosphorylated hexose and triose intermediates in muscle and yeast.

- 1943: Otto Warburg and collaborators crystallize aldolase and characterize its mechanism in skeletal muscle extract.

- 1965: Rapoport, and later kinetic/structural studies through the 1970s–1990s, establish the diffusion-limited catalytic perfection of triose phosphate isomerase.

Why does the cell spend two ATP before glycolysis has produced anything?

The two ATP investments (at hexokinase and PFK-1) are not wasted — they are irreversible commitment steps that trap glucose in the cell and activate its carbon skeleton for cleavage. Without phosphorylation, glucose would freely diffuse back out through GLUT transporters, and without the second phosphorylation, fructose-6-phosphate would lack the geometry needed for the aldolase active site to properly orient and cleave the C3–C4 bond. The investment is recovered four-fold in the payoff phase, when each of the two triose phosphates yields two ATP by substrate-level phosphorylation, for a net gain of two ATP per glucose overall (four generated minus two invested), plus two NADH.

Why is fructose-1,6-bisphosphate cleaved specifically between C3 and C4, and not elsewhere?

The cleavage site is dictated by the position of the carbonyl. Fructose-6-phosphate (and its bisphosphorylated form) carries its ketone at C2. A retro-aldol reaction breaks a C–C bond beta to a carbonyl, generating a resonance-stabilized enolate. Counting from C2, the bond beta to the carbonyl is the C3–C4 bond — cleaving there places the developing negative charge (as an enolate or, mechanistically, a Schiff-base-stabilized enamine) directly conjugated with the C2 carbonyl equivalent. No other bond in the molecule offers this stabilization, so C3–C4 is the unique chemically favorable cleavage site.

What is the difference between Class I and Class II aldolases, and why does it matter?

Class I aldolases, found in animals, plants, and some bacteria/protists, use an active-site lysine to form a covalent Schiff base (iminium) intermediate with the substrate carbonyl, which then stabilizes the retro-aldol transition state without requiring a metal cofactor. Class II aldolases, found in fungi and many bacteria (including many pathogens), instead use a catalytically essential divalent metal ion (typically Zn²⁺) to polarize the carbonyl and stabilize the enolate directly, with no covalent intermediate. Because Class II aldolases are absent from humans, they are actively pursued as selective antifungal and antibacterial drug targets — a direct pharmacological payoff of understanding the mechanistic difference.

Does phosphoglucose isomerase only act on the open-chain form of glucose-6-phosphate?

Yes. Glucose-6-phosphate exists overwhelmingly (>99%) as a cyclic pyranose in solution, but the isomerization chemistry — proton abstraction at C2 forming a cis-enediolate — can only occur on the open-chain aldehyde form, where C1 and C2 are accessible as a genuine carbonyl/alpha-carbon pair. PGI's active site therefore includes catalytic machinery to open the pyranose ring (general acid protonation of the ring oxygen as the C1–O bond breaks), perform the isomerization on the linear intermediate, and then reclose the new furanose ring of fructose-6-phosphate before release. This ring-opening/closing chemistry is mechanistically analogous to what mutarotases and other sugar-isomerizing enzymes must do throughout carbohydrate metabolism.

Why does the cell need triose phosphate isomerase at all — could glycolysis not just process DHAP directly?

In principle a parallel enzymatic pathway for DHAP oxidation could exist, but no such pathway is used in mainline glycolysis; every downstream glycolytic enzyme (starting with glyceraldehyde-3-phosphate dehydrogenase) is built to recognize the aldose G3P specifically. Without TIM, only the G3P half of each cleaved fructose-1,6-bisphosphate could proceed, capping ATP and NADH yield at half its true value and stranding the DHAP half as a metabolic dead end (it would have to be diverted elsewhere, e.g., toward glycerol-3-phosphate for lipid synthesis). TIM's near-diffusion-limited efficiency ensures essentially complete rescue of the DHAP pool, which is why textbooks legitimately describe one glucose as yielding "two" G3P molecules even though only one is produced directly by aldolase.

How do hexokinase and glucokinase differ, and why does the liver use a different isozyme from muscle?

Hexokinase (isozymes I–III), found in muscle, brain, and most tissues, has a low $K_M$ for glucose (~0.1 mM) and is allosterically inhibited by its product G6P, ensuring that these tissues phosphorylate essentially all available glucose immediately and shut off once G6P accumulates (e.g., when downstream pathways are saturated). Glucokinase (hexokinase IV), expressed mainly in liver and pancreatic β-cells, has a much higher $K_M$ (~10 mM), shows sigmoidal (cooperative-like) kinetics with respect to glucose, and is not inhibited by G6P. This design lets the liver's glucose phosphorylation rate track blood glucose concentration over the physiological range (roughly 4–10 mM) rather than saturating and shutting off, appropriate for an organ whose job is buffering systemic blood glucose rather than greedily consuming it for its own energy needs.

Is any carbon or oxygen atom lost during the preparatory phase?

No — the preparatory phase is purely isomerization, phosphorylation, and C–C bond cleavage; there is no oxidation, no CO₂ release, and no net change in oxidation state of any carbon. All six carbons of glucose are conserved and end up distributed between the two three-carbon glyceraldehyde-3-phosphate molecules that proceed into the payoff phase. Oxidation (loss of electrons, generation of NADH) first occurs at the glyceraldehyde-3-phosphate dehydrogenase step, which belongs to the payoff phase, not the preparatory phase.