Cooperativity and Allostery in Oxygen Binding

A single myoglobin molecule binds oxygen the way a lock accepts a key. One site, one ligand, no drama. Its binding curve is hyperbolic, looking exactly like the Langmuir isotherm your general chemistry course promised. Hemoglobin is different. It is built from four myoglobin-like subunits packed into a tetrahedron. It refuses to behave so simply. Its oxygen-binding curve is S-shaped — sigmoidal. This means the tetramer somehow "remembers" that one subunit has already bound oxygen. It uses that memory to make the next binding event easier. This is cooperativity, and it is not a minor curiosity. It is the molecular feature that lets you unload roughly twice as much oxygen in resting muscle as a non-cooperative pigment ever could. This happens over the same pressure drop between lungs and tissue. The puzzle deepens once you ask how one heme tells the others it has bound O₂. It sits thirty angstroms from its neighbor, with no direct electronic contact. The answer is allostery — literally "other shape." It is a conformational relay that runs through the protein's quaternary architecture. It does not run through space or through bonds between the ligands themselves. Hemoglobin exists in two interconvertible quaternary states. These are historically labeled T (tense, low affinity) and R (relaxed, high affinity). The entire cooperative machine is a population-shifting device. Oxygen binding does not chemically alter distant hemes. Instead, it biases the equilibrium between T and R. Layered on top of this T⇌R switch is a second regulatory trick, the Bohr effect. Protons and CO₂ shift the oxygen affinity of hemoglobin. This couples gas transport to tissue metabolism and blood pH. A hard-working muscle is acidic and CO₂-rich. Hemoglobin passing through it releases oxygen more readily, precisely where it is needed most. This is allostery recruited for a second job. It provides chemical feedback between two different diprotic/gaseous species and a third, O₂. All of this is routed through the same quaternary conformational switch. This article works through the thermodynamic language of cooperativity. It covers the Hill equation and the concerted MWC and sequential KNF models. It also examines the structural chemistry of the T→R transition. This includes the heme "doming" mechanism, the Perutz salt bridges, and the α₁β₂ interface. Finally, it explores the physiological logic of the Bohr effect. Each of these is connected to the crystal-field and coordination chemistry established for isolated heme–O₂ binding.

Key takeaways

Myoglobin binds O₂ hyperbolically (Langmuir isotherm); hemoglobin binds sigmoidally because binding at one subunit raises affinity at the others.
Cooperativity is quantified by the Hill coefficient $n_H$ ; hemoglobin has $n_H \approx 2.8$ –3.0, far above the mechanistic maximum of 4 for a perfectly synchronized tetramer only in the sense of apparent steepness, and far above the $n_H = 1$ of an independent-site protein.
Two quaternary states, T (tense, low-affinity, deoxy-stabilized) and R (relaxed, high-affinity, oxy-stabilized), are the physical substrate of allostery; ligand binding shifts the T⇌R equilibrium, not the local heme chemistry alone.
The trigger is mechanical: Fe(II) moves into the porphyrin plane on O₂ binding, pulling the proximal histidine and its helix, which propagates through the α₁β₂ interface to break/reform Perutz salt bridges.
The Monod–Wyman–Changeux (MWC) concerted model and the Koshland–Némethy–Filmer (KNF) sequential model are the two limiting descriptions; real hemoglobin sits closest to MWC but with measurable sequential character.
The Bohr effect (H⁺/CO₂ lowering O₂ affinity) and 2,3-bisphosphoglycerate (2,3-BPG) both stabilize the T state, linking O₂ delivery to tissue pH and altitude/fetal physiology.
Key idea: cooperativity in hemoglobin is not a property of any single heme — it is an emergent, allosteric consequence of a two-state quaternary equilibrium that ligand binding tips, subunit by subunit.

Where this lives today

The T/R allosteric framework established for hemoglobin is now the template for understanding an enormous class of regulated proteins beyond gas transport. Allosteric drug design explicitly targets the T⇌R (or analogous inactive/active) equilibrium in enzymes and receptors — allosteric activators and inhibitors of kinases, GPCRs, and metabolic enzymes are designed to bind at sites remote from the catalytic center and shift a conformational equilibrium exactly as O₂ shifts hemoglobin's, rather than competing directly at the active site. Drugs like voxelotor, used clinically to treat sickle cell disease, work by covalently stabilizing the R state of hemoglobin, raising O₂ affinity and reducing the population of deoxy-HbS available to polymerize — a direct pharmacological exploitation of the MWC two-state switch described here. In altitude physiology and clinical medicine, the 2,3-BPG-driven right-shift of the oxygen curve explains high-altitude acclimatization (elevated red-cell BPG lowers Hb affinity, favoring O₂ unloading in tissue despite lower atmospheric $p_{O_2}$ ), banked-blood storage lesions (BPG depletion in stored blood transiently increases O₂ affinity, impairing immediate post-transfusion oxygen delivery), and the diagnostic use of oxygen dissociation curve shifts in anemia and chronic obstructive pulmonary disease. Fetal-maternal oxygen transfer depends on HbF's structurally reduced BPG affinity (a γ-chain substitution removes a key BPG-binding histidine), giving fetal blood intrinsically higher O₂ affinity than maternal blood across the placental interface — a beautiful evolutionary tuning of the same allosteric machinery. In biophysical instrumentation, single-molecule and time-resolved X-ray free-electron laser studies now directly watch the heme-doming-to-quaternary-switch cascade on picosecond-to-microsecond timescales after laser photolysis of CO-hemoglobin, turning Perutz's static crystallographic mechanism into a genuine molecular movie. The same T/R conceptual vocabulary — a concerted, ligand-driven population shift between two rigid quaternary states — underlies allosteric enzymes throughout metabolism (aspartate transcarbamoylase, phosphofructokinase, glycogen phosphorylase), making hemoglobin the pedagogical Rosetta Stone for Allosteric Regulation of Enzymes generally, while the underlying Fe–O₂ coordination chemistry connects forward to Dioxygen Activation and Reversible O2 Binding and backward to the crystal-field description in Oxygen Transport by Hemoglobin and Myoglobin.

Ask the tutor: How would the oxygen-binding curve and Hill coefficient change for a hemoglobin mutant that cannot form the Lys40α1–His146β2 Perutz salt bridge?

The physical picture

Myoglobin is monomeric, and its oxygen-binding equilibrium,

\text{Mb} + \text{O}_2 \rightleftharpoons \text{MbO}_2,

is a one-site Langmuir isotherm. Fractional saturation $Y$ depends on the partial pressure of oxygen $p$ through a rectangular hyperbola, rising steeply at low $p$ and saturating smoothly — exactly the shape you get from any simple 1:1 binding equilibrium with a single dissociation constant. Hemoglobin's curve looks different: at low $p_{O_2}$ it is nearly flat (hemoglobin resists picking up the first oxygen), then it turns upward steeply through the physiological range, then flattens again near saturation. This S-shape is the fingerprint of positive cooperativity — the affinity for the second, third, and fourth oxygen is progressively higher than for the first. The structural root of this behavior is hemoglobin's quaternary architecture: two α and two β subunits, arranged with two-fold symmetry, held together mostly through two chemically distinct interfaces. The α₁β₁ (and α₂β₂) interface is large, hydrophobic, and barely moves during the T→R transition — it is the "packing" interface. The α₁β₂ (and α₂β₁) interface is smaller, polar, and does nearly all the work: a roughly 15° rotation of one αβ dimer relative to the other converts T to R, sliding this interface over itself like a ratchet between two discrete, stereochemically locked positions. There is essentially no stable intermediate quaternary structure — the switch is binary at the level of quaternary packing, even though the population of T versus R tetramers shifts continuously with the number of ligands bound. At the heme itself, the trigger for this switch is a small but consequential geometric event described already for isolated heme–O₂ binding: five-coordinate deoxy-Fe(II) is high-spin ( $d^6$ , $t_{2g}^4 e_g^2$ in an approximately $C_{4v}$ ligand field) and too large to fit in the porphyrin plane, so it sits domed roughly 0.4–0.6 Å out of the N₄ plane, toward the proximal histidine (His F8). Binding O₂ converts iron to low-spin, six-coordinate Fe(II) ( $t_{2g}^6$ ), which is smaller and moves into the plane. This in-plane motion drags the proximal histidine, and with it helix F, by about 0.6 Å. In myoglobin this local shift has nowhere to go — it just happens. In hemoglobin, helix F is mechanically coupled through the FG corner to the α₁β₂ interface, so the same 0.6 Å heme-doming motion is amplified into the large-scale rigid-body rotation that flips the whole tetramer from T to R.

Figure

Sigmoidal vs. hyperbolic O₂-binding curves for hemoglobin and myoglobin

Figure

T→R quaternary switch: rigid-body rotation at the α₁β₂ interface

The core ideas

The empirical description of cooperative binding begins with the Hill equation. If $n$ ligand molecules bound in an idealized, perfectly concerted fashion, saturation would follow

Y = \frac{p^{\,n_H}}{K_{n_H} + p^{\,n_H}}

where:

$Y$ is fractional saturation (0 to 1),
$p$ is the partial pressure (or concentration) of O₂,
$K_{n_H}$ is an empirical constant with units of pressure $^{n_H}$ , and
$n_H$ , the Hill coefficient, measures the steepness/cooperativity of the transition — $n_H = 1$ for independent sites, $n_H \to n$ (the number of sites) for infinitely cooperative all-or-none binding.

Rearranged into the linear Hill plot,

\log\!\left(\frac{Y}{1-Y}\right) = n_H \log p - \log K_{n_H},

where the slope of $\log[Y/(1-Y)]$ versus $\log p$ is $n_H$ , read directly off experimental oxygen-binding data. For hemoglobin, $n_H \approx 2.8$ –3.0 (never the maximal value 4, because binding is cooperative but not perfectly all-or-none); for myoglobin $n_H = 1.0$ exactly, since there is only one site and nothing to cooperate with. The mechanistic content behind this empirical steepness is supplied by two limiting structural models.

The Monod–Wyman–Changeux (MWC), or concerted, model posits that the whole tetramer exists in only two accessible quaternary states, T and R, in a pre-existing equilibrium even in the absence of ligand:

L = \frac{[T_0]}{[R_0]}

where:

$L$ is the allosteric equilibrium constant between unliganded T and unliganded R, and
for deoxyhemoglobin $L \approx 10^{5}$ – $10^{6}$ , meaning essentially all unliganded tetramers sit in T.

Each subunit within a given quaternary state binds O₂ with one intrinsic dissociation constant regardless of how many neighbors are already occupied — $K_T$ for a subunit in a T-state tetramer, $K_R$ for a subunit in an R-state tetramer, with

c = \frac{K_R}{K_T} \ll 1

where $c$ is the ratio of R-state to T-state intrinsic dissociation constants (R binds far more tightly, so $c$ is small, typically $\sim 0.01$ – $0.001$ ). Crucially, ligands cannot bind "part-way" between T and R — the whole tetramer flips as a rigid unit — so cooperativity here is entirely a population-shift phenomenon: each O₂ bound preferentially stabilizes R over T (by mass action, since R binds tighter), which shifts the T⇌R equilibrium toward R, which raises the apparent affinity for the next ligand even though no individual subunit's intrinsic chemistry changed.

The Koshland–Némethy–Filmer (KNF), or sequential, model instead allows each subunit to change conformation individually upon binding its own ligand, and propagates that change to neighbors via subunit–subunit interaction energies. There is no obligatory two-state symmetric constraint; mixed-conformation tetramers (some subunits "tense," some "relaxed," within one physical tetramer) are allowed and are populated at intermediate saturation. Cooperativity here arises from sequential, positive interaction energies transmitted through the interfaces — each successive conformational change makes the next one thermodynamically easier because of favorable subunit–subunit coupling, not because of a symmetry-imposed all-or-none switch. Real hemoglobin data (oxygen equilibrium curves, X-ray structures of partially liganded intermediates, and NMR studies of quaternary state populations) show that the concerted MWC model captures the bulk of the cooperative free energy — deoxy- and oxyhemoglobin crystallize overwhelmingly as pure T and pure R, with no continuum of quaternary structures — but doubly and triply liganded intermediates show measurable tertiary (within-subunit) relaxation before the full quaternary switch fires, a sequential-model correction on top of the concerted skeleton. Hemoglobin is therefore best described as predominantly MWC with a KNF-like tertiary "pre-equilibration."

Derivation AThe Hill coefficient from the MWC two-state model

Goal. Show how a purely mass-action, two-state (T/R) equilibrium generates an S-shaped binding curve and connect the model parameters $L$ and $c$ to the observed Hill coefficient $n_H$ .

Step 1 — define the species. Let $T_i$ and $R_i$ denote tetramers in the T or R quaternary state with $i$ O₂ bound ( $i = 0,1,2,3,4$ ). Each T-state site binds with dissociation constant $K_T$ ; each R-state site binds with $K_R = cK_T$ . Using $p$ for the oxygen partial pressure and defining reduced pressure $\alpha \equiv p/K_R$ , the equilibrium populations (binomial in the number of ways to place $i$ ligands on 4 sites) are

[T_i] \propto \binom{4}{i}(\alpha c)^i, \qquad [R_i] \propto L^{-1}\binom{4}{i}\alpha^i,

with an overall normalization dropped for clarity and $L$ folded in so that unliganded $T_0$ carries the statistical weight $L$ relative to unliganded $R_0 \equiv 1$ .

Step 2 — total ligand and total protein. Sum over all species to get total tetramer concentration and total bound-ligand count:

[\text{tetramer}]_{\text{tot}} \propto L(1+\alpha c)^4 + (1+\alpha)^4,

[\text{O}_2]_{\text{bound}} \propto 4L\alpha c(1+\alpha c)^3 + 4\alpha(1+\alpha)^3,

obtained by differentiating the binomial sums with respect to $\alpha$ (equivalently, summing $i \times$ each term). Fractional saturation is the ratio, divided by 4 sites per tetramer:

Y = \frac{L\alpha c(1+\alpha c)^3 + \alpha(1+\alpha)^3}{L(1+\alpha c)^4 + (1+\alpha)^4}.

Step 3 — the two limiting regimes. When $c \to 0$ (R-state sites bind essentially infinitely tighter than T-state sites, the physically relevant limit for hemoglobin where $c \sim 10^{-2}$ – $10^{-3}$ ), the $T$ -branch terms with any power of $\alpha c$ beyond the constant vanish rapidly, and at low $L$ -weighted occupancy the expression simplifies to a form dominated by the crossover between the all-T branch (nearly flat, since $K_T$ is large) and the all-R branch (steep, since $K_R$ is small). The curve is forced to look flat at low $p$ (tetramers sit in T, resisting ligand) and steep once enough ligand has accumulated to tip the population to R — this crossover is the sigmoid.

Step 4 — extracting $n_H$ at half-saturation. The Hill coefficient is formally the log-log slope at $Y = 1/2$ :

n_H = \left.\frac{d\ln[Y/(1-Y)]}{d\ln p}\right|_{Y=1/2}.

For the MWC expression above, this slope can be shown to satisfy the bound

1 \le n_H \le 4,

with $n_H \to 1$ as $L \to 0$ or $c \to 1$ (T and R become indistinguishable — no cooperativity), and $n_H \to 4$ only in the unphysical limit $L \to \infty$ together with $c \to 0$ with an infinitely sharp switch (perfect all-or-none behavior). For hemoglobin's actual parameters, $L \approx 3\times10^5$ – $10^6$ and $c \approx 0.003$ – $0.01$ , numerically evaluating $n_H$ from the slope formula gives $n_H \approx 2.8$ – $3.0$ , matching experiment.

Reading. The sigmoid is not built from any single subunit changing its chemistry mid-titration; it is the shape of a population redistributing between two fixed-affinity states as mass action tips the balance. $L$ sets how reluctant the switch is to fire with no ligand aboard (how "tense" T really is), and $c$ sets how much reward there is for firing (how much tighter R really binds) — together they set how steep the observed curve becomes.

Go deeper · Derivation B · 2 minLinkage between O₂ binding and proton release — the Bohr effect from a thermodynamic cycle

Derivation BLinkage between O₂ binding and proton release — the Bohr effect from a thermodynamic cycle

Goal. Derive the quantitative statement of the Bohr effect — that O₂ binding and H⁺ binding to hemoglobin are thermodynamically linked, reciprocal processes — from a simple four-corner linkage (box) diagram, and connect it to the number of Bohr protons released per O₂ bound.

Step 1 — set up the linkage cycle. Consider hemoglobin able to exist as HbH $_n^{4+}$ (protonated, deoxy) or Hb(O₂) $_4$ (deoxygenated proton lost). Write the square cycle

\text{HbH}_n \xrightleftharpoons[K_{O_2}^{\text{deoxy}}]{} \text{Hb(O}_2)_4\text{H}_n \qquad \text{(top and bottom binding O}_2\text{)}

\text{HbH}_n \xrightleftharpoons[K_a^{\text{deoxy}}]{} \text{Hb} + n\text{H}^+ \qquad \text{(left and right releasing H}^+\text{)}

with four species at the corners: deoxy-protonated, deoxy-deprotonated, oxy-protonated, oxy-deprotonated. Because free energy is a state function, the cycle must close:

\Delta G_{O_2}^{\text{(protonated path)}} + \Delta G_{H^+}^{\text{(oxy path)}} = \Delta G_{H^+}^{\text{(deoxy path)}} + \Delta G_{O_2}^{\text{(deprotonated path)}}.

Step 2 — express this as an equality of equilibrium constants. Each leg of the square is an equilibrium constant; closure of the thermodynamic cycle requires the product of constants going one way around equal the product going the other way:

K_{O_2}^{\text{deoxy}} \cdot K_a^{\text{oxy}} = K_a^{\text{deoxy}} \cdot K_{O_2}^{\text{oxy}}

where $K_{O_2}^{\text{deoxy}}$ and $K_{O_2}^{\text{oxy}}$ are O₂-binding constants for the protonated and deprotonated forms respectively, and $K_a^{\text{deoxy}}, K_a^{\text{oxy}}$ are the corresponding acid dissociation constants for the Bohr-effect ionizable groups (chiefly the C-terminal His146β imidazole and the N-terminal α-amino groups).

Step 3 — rearrange to link O₂ affinity directly to pH. Since the Bohr groups are more acidic (lower $pK_a$ , i.e., larger $K_a$ ) in the oxy conformation than in the deoxy conformation — a direct experimental fact, because forming R breaks the salt bridges that were stabilizing the protonated (deoxy) form of His146β — we have $K_a^{\text{oxy}} > K_a^{\text{deoxy}}$ . The cycle-closure relation then forces

K_{O_2}^{\text{oxy-favored path}} > K_{O_2}^{\text{deoxy-favored path}}

in the sense that O₂ affinity measured at low pH (more protonated) must be lower than affinity measured at high pH, because the equality above requires $K_{O_2}^{\text{deoxy}}/K_{O_2}^{\text{oxy}} = K_a^{\text{deoxy}}/K_a^{\text{oxy}} < 1$ : the linkage forces $K_{O_2}$ measured under high-\[H⁺\] conditions to be smaller (weaker O₂ binding) than under low-\[H⁺\] conditions. This is the Bohr effect stated as a linkage relation, not an assumption.

Step 4 — quantify the number of Bohr protons. Differentiating the linkage relation with respect to $\ln p_{O_2}$ at fixed pH (a Wyman linkage relation) gives

\left(\frac{\partial \ln K_a}{\partial \ln p_{O_2}}\right)_{pH} = \left(\frac{\partial \bar{n}_{H^+}}{\partial \bar{n}_{O_2}}\right)_{pH},

i.e. the number of protons released per O₂ bound, $\Delta H^+/\Delta O_2$ , is exactly a cross-derivative in the linkage free energy — proton release and oxygen uptake are the same underlying conformational event, viewed from two different conjugate variables. Experimentally this ratio is about $-0.4$ to $-0.6$ mol H⁺ released per mol O₂ bound near physiological pH (negative sign: protons are released as O₂ is taken up), consistent with roughly 2 protons released per fully oxygenated tetramer (4 O₂ bound) under near-physiological conditions.

Reading. The Bohr effect is not a separate chemical event bolted onto O₂ binding — it is the same T→R quaternary transition, viewed through its effect on ionizable side chains whose pKa's differ between T and R. Any perturbation that stabilizes T (protons, CO₂ as carbamate, 2,3-BPG, chloride) must, by the same linkage logic, lower O₂ affinity, because all these effectors act through the identical two-state switch.

Worked examples

Example 1 — Foundational: reading Hill coefficients off two data points. Suppose an oxygen-binding curve for a hypothetical hemoglobin variant gives $Y = 0.20$ at $p_{O_2} = 15$ torr and $Y = 0.80$ at $p_{O_2} = 42$ torr. Estimate $n_H$ using the two-point Hill-plot slope,

n_H = \frac{\log\!\big[(Y_2/(1-Y_2))\big] - \log\!\big[(Y_1/(1-Y_1))\big]}{\log p_2 - \log p_1}.

Compute the two logit terms:

\log\!\left(\frac{0.80}{0.20}\right) = \log(4.00) = 0.602,

\log\!\left(\frac{0.20}{0.80}\right) = \log(0.250) = -0.602.

Difference: $0.602 - (-0.602) = 1.204$ . Now the pressure term:

\log(42) - \log(15) = 1.623 - 1.176 = 0.447.

n_H = \frac{1.204}{0.447} = 2.69.

This value, close to but below hemoglobin's textbook 2.8–3.0, is exactly the kind of number a real variant or a partially cooperative mutant might show — a useful sanity check that this variant is substantially, but not maximally, cooperative (the ceiling being $n_H = 4$ for four sites). As a limiting-case check: if the same two saturations had instead been reached over a ten-fold narrower pressure range (say $p_1 = 24$ , $p_2 = 27$ torr, $\log$ ratio $= \log(27/24) = 0.051$ ), the same numerator would give $n_H = 1.204/0.051 \approx 23.6$ — an unphysically high value, flagging that real data never compress that tightly; this illustrates why $n_H$ is bounded by the number of sites in any physically sound model, and any apparent $n_H$ above 4 signals an experimental artifact (e.g., aggregation) rather than genuine higher-order cooperativity.

Example 2 — Applied: P₅₀ shift from the Bohr effect during exercise. Resting arterial blood has pH ≈ 7.40 and hemoglobin $P_{50} \approx 26$ torr. During intense exercise, lactic acid and CO₂ production drop local tissue pH to about 7.20 in working muscle capillaries. Using the empirical Bohr coefficient

\phi \equiv \frac{\Delta \log P_{50}}{\Delta \text{pH}} \approx -0.48

(typical human value near physiological pH), estimate the new $P_{50}$ at pH 7.20. Step 1: $\Delta\text{pH} = 7.20 - 7.40 = -0.20$ . Step 2: $\Delta \log P_{50} = \phi \times \Delta\text{pH} = (-0.48)(-0.20) = 0.096$ . Step 3: convert to a ratio,

\frac{P_{50}(\text{pH }7.20)}{P_{50}(\text{pH }7.40)} = 10^{0.096} = 1.248.

Step 4: $P_{50}(\text{pH } 7.20) = 26 \times 1.248 = 32.4$ torr. So a drop of just 0.2 pH units — entirely physiological, generated by working muscle — raises hemoglobin's $P_{50}$ from 26 to about 32 torr, meaning hemoglobin releases oxygen more readily right in the tissue that is consuming it fastest. As a limiting case: if pH were held rigorously constant ( $\Delta\text{pH} = 0$ ), the formula correctly returns $P_{50}$ unchanged (ratio $=10^0=1$ ), confirming the Bohr shift vanishes exactly when there is no pH perturbation, as it must.

Example 3 — Exam/competition level: extracting $L$ and $c$ from two experimental Hill-plot asymptotes. The Hill plot ( $\log[Y/(1-Y)]$ vs. $\log p$ ) for hemoglobin has slope 1 at both very low and very high saturation (the curve is only steep — slope $\approx n_H$ — near the midpoint), because at the extremes essentially all tetramers sit in a single quaternary state and each subunit behaves as an independent binder within that state. The low-saturation asymptote extrapolates (as $Y\to 0$ ) to intercept $\log p^\ast_{\text{low}}$ satisfying $p^\ast_{\text{low}} = K_T$ (T-state intrinsic dissociation pressure), while the high-saturation asymptote (as $Y \to 1$ ) extrapolates to $p^\ast_{\text{high}} = K_R$ (R-state intrinsic dissociation pressure). Suppose these two asymptotes, extrapolated from real oxygen equilibrium curve data, give

K_T = 180\ \text{torr}, \qquad K_R = 0.65\ \text{torr}.

(a) Compute the non-cooperative parameter $c$ :

c = \frac{K_R}{K_T} = \frac{0.65}{180} = 3.61\times 10^{-3}.

(b) The allosteric constant $L$ is separately obtained from the position of the inflection point (where the Hill plot crosses slope $n_H$ at $Y = 1/2$ ); suppose the fitted midpoint pressure is $p_{1/2} = 27$ torr. A standard MWC relation locates the midpoint approximately where the two branch populations are comparable, giving (to leading order, using the two-state symmetry model at $Y = 1/2$ )

L \approx \left(\frac{p_{1/2}}{K_T}\right)^{-4}\cdot\left(\frac{1-c}{c}\right)\quad\text{(order-of-magnitude estimator)}.

Plugging in numbers: $p_{1/2}/K_T = 27/180 = 0.150$ , so $(p_{1/2}/K_T)^{-4} = (0.150)^{-4} = 1975$ . And $(1-c)/c = (1-0.00361)/0.00361 = 275.9$ . Multiplying,

L \approx 1975 \times 275.9 \approx 5.45\times 10^{5},

consistent (order of magnitude) with the textbook range $L \sim 10^5$ – $10^6$ obtained from independent crystallographic and kinetic estimates. As a limiting-case check: if T and R bound identically ( $K_T = K_R$ , so $c = 1$ ), the estimator collapses to $L \propto (1-c)/c \to 0$ regardless of $L$ 's "true" thermodynamic value — correctly signaling that $L$ becomes experimentally unobservable once the two states are chemically indistinguishable, exactly as it must, since cooperativity requires $c \ll 1$ to be measurable at all.

Limiting cases, scope & common traps

Regime / limit	Behavior	Physical reason
Single-subunit protein (myoglobin)	$n_H = 1$ , hyperbolic curve	No quaternary partner exists; nothing to allosterically couple to
$c \to 1$ ( $K_T = K_R$ )	$n_H \to 1$ , sigmoid flattens to hyperbola	T and R states chemically indistinguishable; population shift has no effect on affinity
$L \to 0$ (deoxy-Hb already mostly R-like)	Cooperativity collapses, curve shifts to high affinity, low $n_H$	No T-state reservoir to switch out of; e.g., some mutant/fetal hemoglobins
$L \to \infty$ , $c\to 0$	$n_H \to 4$ (ideal maximum for a tetramer)	Perfectly all-or-none switching; never fully reached physiologically
High CO₂ / low pH (Bohr)	$P_{50}$ increases, curve shifts right	T state stabilized by protonated Bohr groups and carbamate formation
High 2,3-BPG (e.g., high altitude, anemia)	$P_{50}$ increases	BPG binds the central cavity of T state specifically, stabilizing T
Fetal hemoglobin (HbF, γ instead of β)	Lower BPG affinity, higher O₂ affinity than maternal HbA	γ-subunit lacks a key BPG-binding His, weakening BPG's T-stabilization
CO or NO bound instead of O₂	Extremely high affinity, strong left shift, still cooperative via same T/R switch	CO/NO stabilize R even more strongly than O₂ (much smaller $c$ )
Isolated α or β chains (no tetramer)	$n_H = 1$	No quaternary interface exists to transmit the allosteric signal

The chemist behind it

The modern structural understanding of hemoglobin allostery is inseparable from Max Ferdinand Perutz, who spent over two decades solving the X-ray crystal structures of deoxy- and oxyhemoglobin and, in doing so, gave molecular substance to what had previously been a purely thermodynamic curiosity (the sigmoidal binding curve, known since the 1900s work of Christian Bohr, August Krogh, and Karl Hasselbalch). Perutz identified the specific salt bridges — Lys40α1–His146β2, Asp94β1–His146β1, and others — that lock the deoxy tetramer in the T conformation, and he traced the heme-doming-to-helix-F-to-interface mechanical pathway that couples local ligand binding to global quaternary change. His 1970 "stereochemical mechanism" paper in Nature remains one of the foundational syntheses connecting coordination chemistry (spin-state change at Fe) to macromolecular allostery.

Origins · 1 minMax Ferdinand Perutz and how this was discovered

Max Ferdinand Perutz

1914–2002 · Cambridge, Laboratory of Molecular Biology

Austrian-born, educated in Vienna and Cambridge; began X-ray work on hemoglobin in 1937 under J. D. Bernal and continued through internment as an "enemy alien" during WWII, wartime work on the Habbakuk ice-ship project, and eventually the founding of the MRC Laboratory of Molecular Biology. Shared the 1962 Nobel Prize in Chemistry with John Kendrew for determining the structures of globular proteins (hemoglobin and myoglobin respectively) — the first atomic-resolution protein structures ever solved.

History

- 1904 — Christian Bohr, Karl Hasselbalch, and August Krogh report that CO₂ shifts hemoglobin's oxygen affinity (the Bohr effect) and that the binding curve is sigmoidal.

- 1937–1959 — Max Perutz begins X-ray diffraction studies of hemoglobin crystals; solves the low-resolution and then atomic structures of deoxy- and oxyhemoglobin.

- 1965 — Jacques Monod, Jeffries Wyman, and Jean-Pierre Changeux publish the concerted (MWC) two-state allosteric model.

- 1966 — Daniel Koshland, George Némethy, and David Filmer publish the sequential (KNF) induced-fit allosteric model.

- 1970 — Perutz publishes the stereochemical mechanism of cooperativity, linking heme spin-state/doming changes to the T⇌R quaternary switch via the Perutz salt bridges.

Why is a sigmoidal binding curve physiologically better than a hyperbolic one for oxygen transport?

A sigmoidal curve is steep exactly in the physiological pressure window between lung capillary $p_{O_2}$ (~100 torr) and resting-tissue $p_{O_2}$ (~20–40 torr), so a modest pressure drop unloads a large fraction of bound oxygen. A hyperbolic curve with the same $P_{50}$ would already be near saturation at both ends of that range, unloading much less oxygen per pressure drop. Cooperativity effectively "compresses" the useful, oxygen-releasing part of the curve into the exact pressure range the body operates over, roughly doubling the oxygen delivered per unit of hemoglobin compared to a non-cooperative pigment with the same overall affinity.

Is the Hill coefficient a real mechanistic quantity or just a curve-fitting parameter?

It is fundamentally an empirical, phenomenological descriptor — the Hill equation assumes literal simultaneous binding of $n_H$ ligands, which is not what physically happens in hemoglobin (ligands bind one at a time, sequentially). $n_H$ should be read as a steepness index bounded between 1 (no cooperativity) and the number of binding sites (perfect all-or-none cooperativity), not as a literal ligand stoichiometry. Mechanistic models like MWC and KNF explain why $n_H$ takes the value it does in terms of real molecular parameters ( $L$ , $c$ , interaction energies), and the Hill coefficient can be derived as an output of those models rather than assumed as an input.

Does every subunit have to bind O₂ before the T→R switch happens?

No — the switch can and does happen at intermediate ligation states, which is exactly why cooperativity is graded rather than perfectly all-or-none ( $n_H \approx 2.8$ –3, not 4). Crystal structures of singly and doubly liganded hemoglobin intermediates show that some tetramers with only 1–2 O₂ bound have already flipped to the R quaternary structure, while others with the same ligation number remain in T — consistent with the MWC picture that ligand binding shifts a pre-existing T⇌R equilibrium rather than triggering a deterministic switch at a fixed ligation number.

How does carbon monoxide poisoning connect to this same allosteric machinery?

CO binds heme iron with an intrinsic affinity roughly 200–250 times that of O₂ at the isolated heme, and it also strongly stabilizes the R quaternary state, just as O₂ does but more so. This has two compounding lethal effects: CO occupies binding sites (reducing oxygen-carrying capacity directly), and by locking hemoglobin toward R, it raises the apparent O₂ affinity of any remaining unliganded hemes in the same tetramer — shifting the oxygen curve further left and making it harder for tissues to extract what little oxygen is still being carried. The cooperative machinery that normally aids oxygen delivery works against you once CO is aboard.

Why does fetal hemoglobin (HbF) have higher oxygen affinity than adult hemoglobin (HbA)?

HbF replaces the two β subunits with γ subunits, which differ at a key residue (His143β becomes Ser143γ) in the 2,3-BPG binding pocket in the central cavity of the T state. This substitution removes a positive charge that HbA uses to grip BPG tightly, so HbF binds BPG much more weakly. Weaker BPG stabilization of the T state shifts HbF's effective T⇌R equilibrium toward R even at physiological BPG concentrations, raising its intrinsic O₂ affinity relative to maternal HbA — allowing the fetus to extract oxygen from maternal blood across the placenta despite both proteins sharing the identical heme chemistry.

Are the MWC and KNF models mutually exclusive, and which one is "correct" for hemoglobin?

They are limiting cases of a general framework, not mutually exclusive true/false hypotheses. Structural and kinetic data show hemoglobin's quaternary structure is strongly bimodal — essentially only pure T or pure R tetramers are populated, with no stable intermediate quaternary state — which favors MWC's concerted picture over strict KNF sequential switching. However, tertiary (within-subunit, not quaternary) structural relaxation at partially liganded states shows sequential character not captured by pure MWC. The accurate modern picture layers KNF-like tertiary relaxation on top of an MWC-like quaternary switch.

Why doesn't myoglobin show any Bohr effect worth mentioning?

The Bohr effect requires a linkage between an ionizable group's pKa and a conformational change (T⇌R) that itself controls O₂ affinity — but myoglobin has no quaternary structure and no T/R switch to link to. Myoglobin does show a very small intrinsic pH dependence of O₂ affinity from local heme-pocket protonation effects, but it lacks the interfacial salt bridges (like His146β's) whose pKa shifts constitute the bulk of hemoglobin's Bohr effect, so the physiologically significant, large-magnitude Bohr effect is a cooperativity-linked, tetramer-exclusive phenomenon.