Free-Energy Diagrams and the Transition State

Every rationally designed enzyme inhibitor that behaves as a "transition-state analog" is a direct engineering application of this free-energy-diagram logic. Immucillin-H and related compounds mimic the ribooxocarbenium-ion-like transition state of purine nucleoside phosphorylase so closely that they bind roughly $10^{6}$-fold more tightly than the substrate itself, and one such compound (forodesine) is used clinically against T-cell leukemias, exploiting exactly the same "fit the TS, not the substrate" principle Pauling proposed in 1946. HIV protease inhibitors bearing a hydroxyethylene or statine core mimic the tetrahedral gem-diol transition state of the aspartyl-protease mechanism, again binding to a TS-shaped pocket rather than to a cleavable substrate. In protein engineering and directed evolution, computational enzyme design (e.g., Rosetta-based active-site design for Kemp elimination or retro-aldol reactions) works by first calculating the quantum-mechanical structure of the transition state, then scaffolding hydrogen-bond donors and electrostatic groups around that calculated TS geometry — a direct computational implementation of "stabilize the peak, not the valley." Kinetic isotope effect (KIE) measurements, used throughout mechanistic enzymology to probe which bond is breaking at the rate-limiting TS, are interpreted entirely within this free-energy-diagram framework: an unusually large primary KIE signals that C–H bond cleavage occurs at (or very near) the rate-determining transition state itself, while its absence signals that the rate-limiting step lies elsewhere on the diagram (e.g., a conformational change or product release). Even antibody catalysis ("abzymes," raised against stable transition-state-mimicking haptens) is a direct experimental test of Pauling's hypothesis: immunizing animals with a stable TS analog reliably yields antibodies with modest but genuine catalytic activity for the corresponding reaction, confirming that TS-shape complementarity alone, without any of an enzyme's other evolved features, is sufficient to produce real catalysis.

Ask the tutor: How does the shape and depth of the ES and EP valleys on a free-energy diagram determine whether substrate binding or product release becomes the rate-limiting step for an enzyme's overall turnover number?

A free-energy diagram is a plot of Gibbs free energy $G$ on the vertical axis against a "reaction coordinate" on the horizontal axis — an abstract measure of progress from reactants (or enzyme-bound substrate, ES) toward products (EP, then E + P). It is not a plot against time; the horizontal axis has no units of seconds, only "how far along the bond-breaking/bond-making pathway" the system has traveled. Reactants sit in a valley (a local energy minimum, a stable, isolable species). Products sit in another valley, lower or higher depending on whether the reaction is exergonic or endergonic. Between them sits a peak: the transition state, structurally a species with partially broken old bonds and partially formed new bonds, existing for roughly $10^{-13}$ seconds — the time of a single bond vibration — never isolable, never observable by ordinary spectroscopy. The height of that peak above the reactant valley is the activation free energy, $\Delta G^{\ddagger}$. This single number controls the rate constant through the Eyring equation. Crucially, $\Delta G^{\ddagger}$ is completely independent of $\Delta G^{\circ\prime}$, the free-energy difference between the reactant and product valleys, which controls only the equilibrium constant $K_{eq}$. A reaction can be enormously favorable thermodynamically ($\Delta G^{\circ\prime}$ very negative) yet proceed at a negligible rate because $\Delta G^{\ddagger}$ is huge — this is exactly the situation for glucose oxidation by O$_2$ at room temperature, thermodynamically explosive, kinetically inert without a catalyst or a match. Enzymes act entirely on the peak, never on the valleys' relative depths (they do not change $K_{eq}$). The mechanism by which an enzyme lowers $\Delta G^{\ddagger}$ is binding energy: the enzyme active site is a three-dimensional surface of hydrogen-bond donors/acceptors, charged residues, and hydrophobic pockets shaped not to fit the substrate's ground-state geometry perfectly, but to fit the transition-state geometry — with its altered bond lengths, partial charges, and often altered hybridization (sp$^3$→sp$^2$ or vice versa) — even better. Because binding is itself a free-energy-lowering process, and because the enzyme releases much of that binding free energy specifically when the substrate distorts toward the TS geometry, the ES→ETS journey costs less free energy than the uncatalyzed S→TS journey. The uncatalyzed and catalyzed diagrams share the same start and end valley depths (same $\Delta G^{\circ\prime}$ for the overall chemistry) but differ dramatically in peak height and in the presence of extra, shallower intermediate valleys (ES, EP) that the enzyme mechanism introduces.

The second essential picture is what the transition state itself looks like structurally, using a concrete example: the serine protease acylation step. Before the TS, the peptide carbonyl carbon is sp$^2$, planar, with the C=O intact. At the TS, the serine oxygen's lone pair is attacking the carbonyl carbon while the amide nitrogen's leaving-group bond is beginning to break; the carbon is partially sp$^3$, and a negative charge is developing on the oxygen — the classic tetrahedral transition state. This transient tetrahedral geometry with its concentrated negative charge is precisely what the enzyme's "oxyanion hole" (backbone NH groups of Gly193 and Ser195 in chymotrypsin) is built to stabilize via hydrogen bonds that are geometrically too far from the ground-state carbonyl to donate full-strength hydrogen bonds, but perfectly positioned once the oxygen pyramidalizes and the negative charge localizes.

The central relationship is the Eyring equation from transition-state theory, connecting the macroscopic rate constant $k$ to the microscopic free-energy barrier:

where:

• $k$ = observed first-order (or pseudo-first-order) rate constant, s$^{-1}$
• $k_B$ = Boltzmann constant, $1.381\times10^{-23}$ J/K
• $h$ = Planck constant, $6.626\times10^{-34}$ J·s
• $T$ = absolute temperature, K
• $\Delta G^{\ddagger}$ = Gibbs free energy of activation (TS minus reactant valley), J/mol
• $R$ = gas constant, $8.314$ J/(mol·K)

Reading: the prefactor $k_BT/h$ has units of s$^{-1}$ and at 298 K equals $6.21\times10^{12}\,\text{s}^{-1}$ — it represents the frequency with which a species poised at the TS proceeds "downhill" to products (essentially a bond-vibration frequency). The exponential is the equilibrium fraction of reactant molecules that have, at any instant, enough free energy to reach the TS (a Boltzmann factor applied to a free energy, so it implicitly contains both an enthalpic barrier and an entropic cost of organizing the TS geometry). The activation free energy splits into enthalpy and entropy pieces exactly as any free energy does:

where:

• $\Delta H^{\ddagger}$ = activation enthalpy — largely bond stretching/breaking energy at the TS
• $\Delta S^{\ddagger}$ = activation entropy — the entropic cost of achieving the correct TS geometry and orientation (often strongly negative for bimolecular reactions, since two freely tumbling molecules must adopt one precise relative orientation)

The second core relationship links the catalyzed and uncatalyzed barriers to the rate acceleration, the single most important experimentally quoted enzyme number:

where:

• $k_{cat}$ = catalyzed (enzymatic) first-order rate constant
• $k_{uncat}$ = uncatalyzed rate constant for the same chemical step
• $\Delta G^{\ddagger}_{uncat}-\Delta G^{\ddagger}_{cat}$ = the barrier-height reduction achieved by the enzyme, in energy units

Reading: because this is an exponential relationship, modest-looking energy differences produce enormous rate effects. A reduction of just $RT\ln(10) \approx 1.36$ kcal/mol at 25 °C multiplies the rate by exactly 10. Enzymes achieving $10^{10}$-to-$10^{17}$-fold accelerations are therefore only lowering barriers by roughly 14–24 kcal/mol — a modest amount of binding energy, on the same order as three or four hydrogen bonds, provided it is delivered with precision at exactly the right geometry.

Example 1 — Foundational: computing $\Delta G^{\ddagger}$ from a measured rate constant. Chymotrypsin hydrolyzes a synthetic peptide substrate with $k_{cat} = 4.6\times10^{2}\,\text{s}^{-1}$ at 37 °C (310 K). Find $\Delta G^{\ddagger}$. Rearranging the Eyring equation:

Compute the prefactor at 310 K:

Compute the dimensionless ratio inside the log:

Take the natural log:

Multiply by $-RT$ (using $R=1.987\times10^{-3}$ kcal/mol·K):

So a modest-sounding rate of a few hundred turnovers per second corresponds to a barrier of about 14.4 kcal/mol — check the order of magnitude: enzyme barriers for efficient catalysts typically fall in the 11–16 kcal/mol range at physiological temperature, and 14.4 kcal/mol sits squarely inside that band, confirming the arithmetic is sane.

Example 2 — Applied: rate acceleration by orotidine 5′-monophosphate decarboxylase (ODCase), one of the most proficient enzymes known. The uncatalyzed decarboxylation of OMP has a half-life of about 78 million years in water at 25 °C, corresponding to $k_{uncat} \approx 2.8\times10^{-16}\,\text{s}^{-1}$ (from $k = \ln 2 / t_{1/2}$, with $t_{1/2}=78\times10^6\ \text{yr} = 2.46\times10^{15}$ s). The enzyme achieves $k_{cat} = 39\ \text{s}^{-1}$. Compute the barrier-height reduction $\Delta G^{\ddagger}_{uncat}-\Delta G^{\ddagger}_{cat}$ at 298 K. First form the rate ratio:

Take the natural log:

Apply the derived formula with $R=1.987\times10^{-3}$ kcal/(mol·K), $T=298$ K:

ODCase therefore lowers the activation barrier by roughly 23.4 kcal/mol — a huge number by enzymology standards, yet still less than five typical hydrogen bonds' worth of binding energy (each worth roughly 1–5 kcal/mol), which is exactly the point: modest, precisely targeted binding interactions with the TS, not exotic chemistry, explain even the most spectacular known rate accelerations. As a sanity check against Example 1's derivation, note $\Delta G^{\ddagger}_{cat}$ itself can be recovered from $k_{cat}=39\ \text{s}^{-1}$ using the same method as Example 1, giving roughly 15.2 kcal/mol — consistent with the enzymatic barrier again landing in the expected 11–16 kcal/mol window.

Example 3 — Exam/competition level: dissecting binding energy across a two-hump diagram with an unstable intermediate. An enzyme-catalyzed reaction proceeds through the sequence E + S ⇌ ES → EI → EP → E + P, where EI is a covalent intermediate (e.g., an acyl-enzyme). Suppose the free-energy diagram (relative to free E + S set at 0 kcal/mol) has the following values, at 298 K:

• ES valley: $-3.0$ kcal/mol (binding energy released)
• TS1 (ES→EI): $+9.5$ kcal/mol
• EI valley: $+1.0$ kcal/mol
• TS2 (EI→EP): $+12.0$ kcal/mol
• EP valley: $-8.0$ kcal/mol
• E + P (final): $-9.5$ kcal/mol

(a) Identify the rate-limiting step and compute its effective $\Delta G^{\ddagger}$, measured from the lowest point preceding it (the rule for multi-hump diagrams — the barrier that matters kinetically is measured from the deepest accessible valley below it, not necessarily the adjacent one). Both peaks must be checked against every valley that precedes them. TS1 is only 12.5 kcal/mol above ES ($9.5-(-3.0)$). TS2, however, must be checked against both EI (11.0 kcal/mol above it: $12.0-1.0$) and, if EI re-equilibrates back through ES, against the deeper ES valley: $12.0-(-3.0) = 15.0$ kcal/mol. Since ES is a lower, more populated valley than EI and the system can flow back to it before committing to TS2, the operationally rate-limiting barrier — the one that determines overall $k_{cat}$ — is:

(b) Compute $k_{cat}$ at 298 K from this effective barrier.

With $k_BT/h = 6.21\times10^{12}\,\text{s}^{-1}$ at 298 K, and the exponent:

(c) Compute the overall equilibrium constant for S → P, and confirm it depends only on the valley floors (0 for S, $-9.5$ for P), not on either barrier height.

Notice this $K_{eq}$ used none of TS1, TS2, ES, or EI — precisely the independence proven in the second derivation above. As a limiting-case check: if instead the enzyme mutated to raise TS2 by 3 kcal/mol (to 15.0 kcal/mol) while leaving all valleys fixed, $\Delta G^{\ddagger}_{\text{eff}}$ becomes 18.0 kcal/mol, $k_{cat}$ falls by a factor of $e^{3/0.5922}=e^{5.07}\approx159$-fold to about 0.39 s$^{-1}$, while $K_{eq}$ remains exactly $9.24\times10^{6}$ — a clean illustration that mutating a catalytic residue can devastate turnover without at all changing the reaction's thermodynamic favorability.

Henry Eyring (1901–1981), a Mexican-born American chemist trained under Michael Polanyi, built transition-state theory in 1935 by combining statistical mechanics with the emerging picture of potential-energy surfaces from quantum chemistry. Eyring's insight was to treat the transition state as a quasi-equilibrium species — not truly stable, but populated according to Boltzmann statistics just like any other species in a rapid pre-equilibrium — and multiply that population by a universal "frequency of crossing," $k_BT/h$, a term whose surprising universality (independent of the specific reaction, depending only on temperature) was itself the theory's most striking prediction. Eyring's collaboration with Polanyi produced some of the first calculated reaction-rate profiles for simple gas-phase reactions like H + H$_2$ → H$_2$ + H, work that translated directly, decades later, into the language enzymologists now use for every biological reaction: bind, distort, stabilize the TS, release. The application of transition-state theory specifically to enzyme catalysis is largely credited to Linus Pauling, who in a short but hugely influential 1946 paper argued that enzymes work by being "structurally complementary" to the transition state rather than the substrate — a claim that predated any structural biology capable of testing it directly, and that was only vindicated decades later once X-ray crystal structures of transition-state analogs bound to enzymes (and kinetic isotope effect studies) became available.