Four That Act in Concert: Allostery, Hemoglobin, and the Logic of Cooperative Binding

The Hook: Why a single oxygen molecule changes the rules

Imagine you want to fill four identical seats on a bus, but the seats behave strangely. As long as all of them are empty, sitting down is almost forbiddingly difficult. Yet the moment the first passenger takes a seat, the other three suddenly become invitingly comfortable, and the next three passengers sit down almost in the same instant. And when the bus empties again, the effect reverses: as soon as the first person stands up, the others practically leap to their feet.

This peculiar behavior — "hard to start, then it rushes along, and just as abruptly when letting go" — is not a curiosity. It is the fundamental mode of motion of the molecule keeping you alive this very second: hemoglobin, the red oxygen carrier in your blood cells. And it is the most vivid embodiment of one of the most important and widespread control principles in all of biochemistry — allostery.

The concrete problem nature solved here is this. In the lungs, the oxygen partial pressure is high, and there hemoglobin should take up as much oxygen as possible. In the working tissues — in the muscle climbing stairs, in the brain concentrating — the oxygen pressure is low, and there the same molecule should release its cargo as completely as possible. A simple carrier that bound oxygen strictly according to the concentration on hand could not manage this balancing act: it would load only moderately in the lungs and barely unload in the tissue. Hemoglobin does both at once, because its four binding sites do not work independently but talk to one another.

This article tells how that molecular conversation works — and how figuring it out became one of the loveliest intellectual stories in molecular biology. We move from the unusual shape of the oxygen-binding curve, through two rival models from the 1960s, through Max Perutz's atom-by-atom picture of the molecule's two states, all the way to medicine: to sickle cell disease, to CRISPR therapy, and to an allosteric drug that was pulled from the market in 2024, leaving an uncomfortable lesson behind.

Part 1: The problem — why simple binding is not enough

The oxygen-binding curve: sigmoidal, not hyperbolic

If you measure what fraction of a protein's binding sites are occupied at a given oxygen pressure (its saturation, plotted against the oxygen partial pressure), you would naively expect a simple saturation curve: a steep rise, then a leveling off toward saturation — mathematically, a hyperbola. This is exactly how myoglobin behaves, the oxygen-binding protein of muscle, which consists of a single subunit.

Hemoglobin, by contrast, shows a completely different shape: a sigmoidal (S-shaped) curve. At low oxygen pressure the binding is at first sluggish and reluctant (the flat beginning); then comes a steep, almost sudden rise in saturation; and finally the curve levels off toward saturation. This S-shape is the graphical signature of positive cooperativity: binding the first oxygen molecule makes it easier to bind the next.

The physiological payoff lies precisely in the steep middle portion of the curve. In humans, this steep region falls roughly in the pressure window between the lungs and the tissues. That means a relatively small drop in oxygen pressure — the transition from the lung environment to the tissue environment — produces a large change in saturation and thus a large quantity of oxygen released. A molecule with a hyperbolic curve (like myoglobin) would release only a fraction of that over the same pressure change. Cooperativity is therefore not a biochemical end in itself, but a clever amplification mechanism: it makes the carrier sensitive precisely in the pressure range where it matters.

Myoglobin as a contrast

The comparison with myoglobin is instructive, because it shows that cooperativity cannot reside in the individual binding site. The actual chemical binding site — the heme group, with a central iron atom to which oxygen docks — is essentially the same in myoglobin and in every hemoglobin subunit. Myoglobin binds oxygen very tightly (it is a store, not a delivery service) and does so without any cooperativity, because it has only one subunit and can "talk" to no one.

Hemoglobin is a tetramer: four subunits, which in the adult human are two α-chains and two β-chains, each with its own heme group and its own iron atom. Cooperativity is therefore an emergent property of the assembly — it arises only because four binding sites are clamped together in a shared protein scaffold and communicate their states to one another. It is precisely this action-at-a-distance between spatially separated sites of a single molecule that lies at the heart of allostery.

Part 2: The language of allostery

The term allostery comes from the Greek words allos (other) and stereos (solid, spatial): an event at one location on the molecule changes the behavior at another, spatially separated location. Before we turn to the models, it is worth sharpening the vocabulary, because it structures the whole discussion.

Homotropic and heterotropic

We distinguish two kinds of effects. Homotropic effects are interactions between molecules of the same type: oxygen influencing the binding of further oxygen. Hemoglobin's cooperativity toward its own ligand is a homotropic effect. Heterotropic effects, by contrast, are interactions with a different molecule that binds at a separate site and modulates the ligand affinity. For hemoglobin, protons (acid), carbon dioxide, and the molecule 2,3-bisphosphoglycerate are such heterotropic effectors — more on these below.

Positive and negative cooperativity

Positive cooperativity means the first binding event raises the affinity for the next (hemoglobin for oxygen). Negative cooperativity is the opposite case: the first binding event lowers the affinity for the following one — this also exists, for instance in certain enzymes and receptors. This distinction will matter shortly, because it is one of the touchstones against which the two great models of allostery can be measured.

The decisive conceptual insight: cooperativity presupposes that there are channels of communication through the protein. The question of the 1960s was not whether subunits influence one another, but how the message travels from one binding site to the next. Two research groups gave two elegant, initially competing answers.

Part 3: The two great models

The concerted model (Monod, Wyman, Changeux 1965)

The first model was put forward in 1965 by Jacques Monod, Jeffries Wyman, and Jean-Pierre Changeux and is known as the MWC model or the concerted model. Its basic assumption has an almost puristic elegance and rests on a strong symmetry rule.

The MWC model postulates that the entire tetramer can adopt only two possible conformational states: a T state (from tense) with low ligand affinity, and an R state (from relaxed) with high affinity. The crucial condition is the symmetry constraint: all four subunits always switch together — either all four are in the T state or all four are in the R state. There are no mixed intermediate forms in which some subunits have already flipped and others have not. The switching event is "concerted"; it happens en bloc.

How does this generate cooperativity? Even in the complete absence of ligand, both states exist in a chemical equilibrium described by an equilibrium constant L₀; in the unloaded state, the equilibrium lies far on the side of the low-affinity T state (hence the sluggish start of the curve). Oxygen binds preferentially to the high-affinity R state. Each bound oxygen molecule therefore "pulls" on the equilibrium and shifts it toward R. Because all four subunits flip together into the high-affinity R state, the still-empty binding sites abruptly become high-affinity as well when the switch occurs — and the remaining oxygen molecules bind far more easily. From a simple state equilibrium plus a preference, the S-shaped curve emerges.

The strength of the MWC model: it is mathematically economical, explains positive cooperativity elegantly, and at the same time provides a natural account of heterotropic effectors. A molecule that binds preferentially to the T state shifts the equilibrium toward T and thereby lowers affinity (a negative effector); a molecule that prefers the R state does the reverse. Its decisive weakness: in its pure form, the concerted model cannot describe negative cooperativity — the symmetry rule simply does not permit it.

The sequential model (Koshland, Némethy, Filmer 1966)

A year later, in 1966, Daniel Koshland, George Némethy, and David Filmer formulated an alternative, the KNF model or sequential model. It builds on Koshland's older idea of induced fit and starts from a quite different premise.

In the sequential model there is no strict symmetry constraint. Each subunit can change its state individually. When a ligand binds to a subunit, that subunit changes its conformation by induced fit, and this change influences the neighboring subunits through the contact interfaces and alters their affinity — step by step, sequentially. Intermediate forms with mixed conformations are explicitly allowed here.

The great advantage: because the subunits respond individually, the KNF model can describe both positive and negative cooperativity. If the conformational change triggered by the first binding event makes the neighbors more receptive, positive cooperativity results; if it makes the neighbors more reluctant, negative cooperativity results. The sequential model thus covers a broader range of observed behavior than the concerted one — and it is mechanistically more intuitive, because it lies closer to the picture of a ligand actively reshaping its binding site.

Who is right?

The honest answer is: both, a little — and hemoglobin lies closer to the concerted model, though it is more complicated than any pure scheme. For hemoglobin itself, the MWC picture proved remarkably robust, because the molecule really does exist in two clearly distinguishable quaternary structures corresponding to the T (deoxy) and R (oxy) states. The fundamental two-state idea captures the quaternary structure well.

At the same time, later high-precision measurements — above all the thermodynamic work of Gary Ackers and colleagues on the binding energies of the individual intermediates — showed that reality is richer: there are subtleties at the level of tertiary structure (the shape of individual subunits) within a given quaternary state, and the distribution of the cooperative free energy across the various binding steps does not fit the purest two-state picture exactly. Modern, often statistical-mechanical, models no longer treat MWC and KNF as adversaries but as limiting cases of a more general framework in which both concerted quaternary transitions and local tertiary adjustments have their place. I am of the opinion that it is precisely this outcome — two brilliant, seemingly irreconcilable models that ultimately turn out to be complementary approximations of a finer truth — that makes the intellectual history so instructive.

Part 4: The structural proof — Perutz and stereochemistry

Models are one thing; looking at the molecule itself is another. Here Max Perutz enters, who devoted much of his research life to hemoglobin. Together with John Kendrew (who solved the structure of myoglobin), he received the Nobel Prize in Chemistry in 1962 for elucidating the structures of globular proteins by X-ray crystallography. In 1970 Perutz presented his famous stereochemical mechanism, which explained the cooperative effect at atomic resolution for the first time.

T and R states, salt bridges, and the iron out of the plane

Perutz compared the crystal structures of the deoxygenated and the oxygenated forms and interpreted the two states structurally. The T state (deoxy-hemoglobin) is held together by a network of salt bridges between the subunits — ionic bonds, among others at the C-termini of the chains — that "brace" the structure and lock it into a geometry of low oxygen affinity. The R state (oxy-hemoglobin) has shed this bracing; the salt bridges are broken, and the binding sites are high-affinity.

The triggering mechanism is astonishingly subtle and begins with a single atom. In the deoxy state, the iron atom sits slightly outside the plane of the flat heme group, pulled a little out of it. When oxygen binds, the iron slips into the heme plane — a tiny movement of only a few hundredths of a nanometer. But because the iron is firmly coupled through an amino acid (the proximal histidine) to a helix of the protein (the F-helix), this minute movement drags the whole helix along. This local shift propagates across the contact interfaces between the subunits, breaks salt bridges, and finally triggers the large quaternary switch from T to R. In this way a chemical bond at one site is translated mechanically into a change of affinity at the others — Perutz thereby made the abstract "communication" of the subunits visible as a concrete chain of atomic movements.

The Bohr effect and 2,3-BPG: the heterotropic masters

The real physiological finesse shows itself in the heterotropic effectors, and Perutz's salt-bridge picture explains them mechanistically.

The Bohr effect (after Christian Bohr, 1904) describes the fact that hemoglobin binds oxygen less well at low pH (that is, higher acidity) and at high carbon dioxide content — the binding curve shifts to the right. This is physiologically exactly right: in working tissues, metabolism produces plenty of CO₂ and acid (lactic acid, carbonic acid). The Bohr effect ensures that hemoglobin releases its cargo more readily there, where oxygen is most urgently needed. Mechanistically, protons and CO₂ bind preferentially to the T state (protons to certain amino acid groups, CO₂ as a carbamate at the N-termini) and stabilize its salt bridges — they shift the equilibrium toward the low-affinity T form.

2,3-bisphosphoglycerate (2,3-BPG) is the second great heterotropic regulator, a small, strongly negatively charged molecule from the metabolism of red blood cells. It fits precisely into a central cavity of the tetramer that is open only in the T state, and anchors itself there via positive charges on the β-chains. In doing so, 2,3-BPG stabilizes the T state and lowers oxygen affinity. This sounds counterproductive, but it is ingenious: without 2,3-BPG, hemoglobin would bind oxygen so tightly that it would hardly release it in the tissue. 2,3-BPG "tunes" the affinity to a physiologically useful value. During adaptation to high altitude, the 2,3-BPG level rises, which improves oxygen release in the tissue — a fast, reversible fine-adjustment of the carrier to thinner air.

A particularly beautiful demonstration of the importance of 2,3-BPG is fetal hemoglobin (HbF). Instead of β-chains, the fetus has so-called γ-chains (composition α₂γ₂). These γ-chains bind 2,3-BPG considerably more weakly, which is why fetal hemoglobin has a higher oxygen affinity than the mother's. It is precisely this affinity gradient that allows the fetus to effectively "take" oxygen from the maternal circulation across the placenta. Here, then, nature regulates oxygen affinity not through the heme or the iron, but purely through the allosteric receptiveness to an effector.

Part 5: The mathematics — the Hill equation

As early as 1910, long before any structural picture, the physiologist Archibald Vivian Hill attempted to capture the sigmoidal curve mathematically. His eponymous Hill equation describes saturation as a function of ligand concentration with an exponent, the Hill coefficient (n_H). If you plot the data suitably (as a so-called Hill plot on logarithmic axes), the coefficient emerges from the slope.

The interpretation is memorable:

• n_H = 1 means no cooperativity (independent binding sites) — this is how myoglobin behaves.
• n_H > 1 signals positive cooperativity — hemoglobin lies at about 2.8 to 3.0.
• n_H < 1 indicates negative cooperativity.

A common misconception deserves clarification: hemoglobin has four binding sites, but its Hill coefficient is not 4. A value of 4 would be reachable only under "infinitely strong" cooperativity — that is, if the four oxygen molecules bound strictly all-or-nothing together. The real value of around 2.8–3.0 shows that cooperativity is strong but not perfect: the Hill coefficient is a measure of cooperativity, not a count of binding sites. It represents a lower bound on the number of interacting sites. Historically, Hill's approach was a remarkable anticipation: he described a cooperative phenomenon quantitatively half a century before anyone could see the underlying conformational states.

Part 6: When it goes wrong — and when we intervene

Allostery is not a laboratory phenomenon but medicine. Nowhere is this clearer than in the diseases of hemoglobin and the attempts to intervene in them allosterically.

Sickle cell disease: one amino acid with enormous consequences

Sickle cell disease is perhaps the most famous molecular disease in history — Linus Pauling coined the term for it in 1949. Its cause is a single point mutation: at position 6 of the β-chain, the amino acid glutamate is replaced by valine (β6 Glu→Val). The mutant hemoglobin is called HbS.

The effect of this tiny change is devastating, and it is itself allosteric in the broader sense: the substituted valine creates a sticky hydrophobic patch on the molecule's surface, but one that is exposed only in the deoxy state (T form). As soon as HbS releases its oxygen, these sticky patches lock into one another, and the molecules polymerize into long, rigid fibers. These fibers distort red blood cells into their eponymous sickle shape, make them stiff and fragile, and trigger vascular occlusions, pain crises, and organ damage. The key: it is the deoxygenated HbS that polymerizes. Anyone wanting to combat sickling can therefore either lower the number of HbS molecules — or prevent them from entering the polymerization-prone T state.

Fetal hemoglobin as a therapeutic lever: the CRISPR connection

This is exactly where a circle closes back to fetal hemoglobin. HbF does not polymerize and moreover "disrupts" the fiber formation of HbS. People who, for genetic reasons, still produce a great deal of HbF in adulthood have significantly milder courses of the disease. This observation became the target of one of the most spectacular new therapies: Casgevy (exagamglogene autotemcel), the first approved CRISPR-based gene therapy. Using gene editing, it switches off a switch (the repressor BCL11A) that normally silences the γ-chains, and thus fetal hemoglobin, after birth. As a result, the red blood cells produce HbF again — and sickling is curbed. I have described the mechanics of this editing separately in the vault (see cross-references).

Voxelotor: an allosteric drug — and a lesson

The most direct attack on allostery is a drug that shifts the state equilibrium itself. Voxelotor (trade name Oxbryta) is exactly that: an allosteric modulator that binds to the N-terminus of the α-chain and stabilizes hemoglobin in the high-affinity R state. The logic is compelling and follows directly from the MWC picture: if you push the equilibrium toward R (the oxy form), there is less deoxy-HbS — and less deoxy-HbS means less polymerization and less sickling. Voxelotor received accelerated approval from the U.S. FDA in 2019 for patients aged twelve and older.

But the story took a sobering turn. On September 25, 2024, the manufacturer Pfizer voluntarily withdrew voxelotor worldwide and discontinued all ongoing clinical trials. The reason was an imbalance that had emerged in the data concerning vaso-occlusive crises and deaths: in a study of children at elevated risk of stroke, there were more deaths in the voxelotor group than in the comparison group, and the regulatory authorities concluded that the overall benefit no longer outweighed the risk.

The case is instructive for two reasons. First, biochemically: voxelotor reliably improved the laboratory value — it raised the hemoglobin level and lowered markers of blood-cell destruction, because fewer cells sickled. But a better surrogate measure is not the same as better patient well-being. Possibly the stronger oxygen binding in the R state buys the prettier lab value at the price of hemoglobin releasing its oxygen less readily in the tissue — a direct echo of the allosteric logic underlying this whole article. Second, methodologically: the case became a much-cited warning against approvals that rely on new surrogate endpoints instead of hard clinical outcomes. I am of the opinion that it is precisely this connection — a molecule whose clean allosteric mechanism at first convinced everyone, and whose clinical balance sheet then failed to add up — that shows how much humility the leap from molecular structure to therapy demands.

Part 7: Allostery beyond hemoglobin

Hemoglobin is the teacher, but not the special case. Allostery is a universal control principle of biochemistry.

Enzymes, receptors, signaling pathways

The classic case of metabolic regulation is feedback inhibition: the end product of a multi-step synthetic pathway binds to an enzyme at the start of the chain — not at its active site, but at a separate, allosteric site — and thereby throttles its own production. The molecule regulates its own manufacture, entirely without a central authority. A textbook example is aspartate transcarbamoylase (ATCase), on which enzymologists have played out the MWC logic almost paradigmatically.

A large part of modern pharmacology also rests on allostery. Many drugs act not at a receptor's actual binding site but as allosteric modulators at separate sites — amplifying or dampening the effect of the natural messenger rather than replacing it. This holds for important drug classes used in anxiety disorders, epilepsy, and many other indications. The appeal of allosteric drugs lies in their ability to regulate more subtly than a crude on/off switch at the active site.

A lesson for systems thinkers

Beyond biochemistry, allostery contains an idea that carries far: a local event can trigger a global change of state if a system exists in a few clearly distinguishable overall states and an equilibrium between them is sensitively tuned. This is the signature of many switched systems — from cellular signal processing, through phase transitions in physics, to tipping points in organizations and markets. Hemoglobin's S-curve is thus a relative of every other S-curve in which a threshold turns a sluggish initial phase into a steep, self-reinforcing flip.

The Central Takeaway

Allostery in hemoglobin condenses several great ideas into a single molecule:

1. Function emerges from cooperation. Four identical building blocks accomplish together something none could do alone: a switching, context-sensitive oxygen binding. The S-shaped curve is not a chemical detail but the graphical signature of this coordination, and the reason you can breathe equally well at altitude and in a sprint.
2. Structure is mechanism. Perutz showed that "communication" between distant parts of a molecule is nothing mystical, but a concrete chain of atomic movements — set off by an iron atom that slips a few hundredths of a nanometer into a plane.
3. Models are tools, not dogmas. The concerted (MWC) and sequential (KNF) models seemed to be adversaries; today they are understood as complementary limiting cases of a finer reality. And the voxelotor case warns that even a convincing mechanism cannot replace the clinical test.

A concrete call to action: In a system you know well — a software architecture, a team, a process — look for an "allosteric site": a point where a small, targeted intervention does not act directly on the core activity, but shifts the equilibrium of the whole and thereby produces a disproportionate effect. Such leverage points are often more effective and gentler than a frontal intervention at the "active site." Learning to recognize them is thinking the way nature did when it built hemoglobin.

A question to reflect on: Hemoglobin owes its entire usefulness to the fact that it does not bind maximally tightly, but adjusts its affinity depending on context — load firmly, release readily. Where in your own work do you confuse "hold on as tightly as possible" with "function optimally" — and where would the true value lie instead in the ability to let go deliberately, depending on the situation?

Cross-References in the Vault

• The Molecular Turbine: ATP Synthase and the Engine of Life – ATP synthase, too, translates a tiny mechanical movement into chemical function; both molecules show how conformational changes are the true language of molecular machines.
• The Shape of Life: Levinthal's Paradox, Chaperones, and How Proteins Find Their Form – Allostery is inconceivable without the protein's precise three-dimensional fold; that article shows how a protein even finds the shape that can then switch between T and R.
• The Programmable Scissors: CRISPR and the Rewriting of Life – The CRISPR therapy Casgevy mentioned in the sickle-cell section is explained there in detail; a contrast between a genetic intervention (CRISPR) and an allosteric one (voxelotor) in the same disease.

Sources and Further Reading

• Wikipedia – Monod–Wyman–Changeux model (concerted model, symmetry rule, L₀): https://en.wikipedia.org/wiki/Monod%E2%80%93Wyman%E2%80%93Changeux_model
• ScienceDirect Topics – Monod-Wyman-Changeux Model (overview, T/R equilibrium): https://www.sciencedirect.com/topics/biochemistry-genetics-and-molecular-biology/monod-wyman-changeux-model
• NCBI PMC – Structural Basis of Sequential and Concerted Cooperativity (MWC vs. KNF, structural basis): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9687781/
• NCBI PMC – A retrospective on statistical mechanical models for hemoglobin allostery (modern, unifying view): https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9830738/
• Annual Review of Biophysics – The Stereochemical Mechanism of the Cooperative Effects in Hemoglobin Revisited (Perutz, T/R, salt bridges): https://www.annualreviews.org/doi/pdf/10.1146/annurev.biophys.27.1.1
• Biophysical Journal – Half a Century of Hemoglobin's Allostery (overview, open questions): https://www.cell.com/biophysj/fulltext/S0006-3495(15)00608-6
• Pfizer – Pfizer Voluntarily Withdraws All Lots of OXBRYTA (voxelotor) (withdrawal, September 25, 2024): https://www.pfizer.com/news/press-release/press-release-detail/pfizer-voluntarily-withdraws-all-lots-sickle-cell-disease
• FDA – FDA is alerting patients and health care professionals about the voluntary withdrawal of Oxbryta: https://www.fda.gov/drugs/drug-alerts-and-statements/fda-alerting-patients-and-health-care-professionals-about-voluntary-withdrawal-oxbryta-market-due
• American Journal of Hematology – FDA Approval Based on Novel Surrogate Endpoints: Lessons From the Voluntary Withdrawal of Voxelotor: https://onlinelibrary.wiley.com/doi/10.1002/ajh.27635

Created as part of the daily learning workflow. Interessensgebiet: Biochemistry. Estimated reading time: ~30 minutes.

← All articles