Hemoglobin & Oxygen Transport
A resting adult consumes about 250 mL of oxygen every minute. Yet only 3 mL of O₂ physically dissolve in each liter of plasma. If blood relied on dissolved gas alone, the heart would need to pump roughly 80 liters per minute. This is just to meet resting metabolic demand. Instead, cardiac output sits near 5 liters per minute. This feat is made possible by a single protein packed into red blood cells: hemoglobin. This article explains how hemoglobin's four-chain architecture works. It lets hemoglobin load oxygen almost completely in the lungs. It also lets hemoglobin unload a large fraction of it in tissues. All of this happens without any energy expenditure. The central puzzle is a chemical one. How can one molecule bind O₂ tightly enough to saturate in the lungs? Yet it must also release oxygen easily enough to supply exercising muscle. The answer is cooperativity. This property emerges purely from the geometry of protein subunits shifting shape as they bind ligand. Hemoglobin is not a passive sponge. It is an allosteric machine whose oxygen affinity changes continuously with how much oxygen it is already carrying. It also responds to local pH, CO₂, and temperature. This behavior is not an incidental biochemical curiosity. It is the physiological linchpin of the entire circulatory-respiratory system. Every downstream topic depends on the molecular mechanics described here. This includes the shape of the oxygen-hemoglobin dissociation curve. It also includes the transport of carbon dioxide and buffering of blood pH. Understanding hemoglobin also explains several diseases. These include sickle cell anemia, carbon monoxide poisoning, and high-altitude adaptation. It also explains clinical tools like pulse oximetry and blood gas analysis. All of these rest on this single protein's behavior. This article builds the mechanism from the atom up. It covers heme iron, the globin fold, and the T-to-R conformational switch. It also covers the environmental modulators (Bohr effect, 2,3-BPG, temperature). It then walks through the classical experiments. These include X-ray crystallography by Perutz. They also include the kinetic and structural work that established the two-state allosteric model. Finally, it applies the concept to worked physiological problems.
- Hemoglobin is a tetramer (2α + 2β chains), each subunit holding one heme group with a central Fe²⁺ that reversibly binds one O₂ molecule.
- Oxygen binding is cooperative: binding O₂ at one subunit increases the O₂ affinity of the remaining subunits, producing a sigmoidal dissociation curve.
- This cooperativity arises from a quaternary conformational shift between a low-affinity Tense (T) state and a high-affinity Relaxed (R) state.
- Affinity is tuned by pH, CO₂, temperature, and 2,3-bisphosphoglycerate (2,3-BPG) — the Bohr effect shifts the curve right in metabolically active tissue.
- Each gram of hemoglobin carries about 1.34 mL of O₂ when fully saturated; total blood O₂ capacity is roughly 200 mL O₂ per liter of blood.
- Perutz's X-ray crystal structures of deoxy- and oxyhemoglobin directly visualized the T→R conformational switch, confirming the allosteric model proposed from binding-curve kinetics.
Hemoglobin's four subunits communicate through a conformational switch, so oxygen binding is self-reinforcing in the lungs and self-releasing in the tissues — turning a simple iron-oxygen bond into a finely tuned delivery system.
The core idea
Hemoglobin (Hb) is the oxygen-carrying metalloprotein packed at roughly 33 g per 100 mL inside red blood cells (Blood Composition & Hematopoiesis covers how erythrocytes are made and maintained to house it). Structurally, adult hemoglobin (HbA) is a tetramer of four polypeptide globin chains — two α chains (141 amino acids each) and two β chains (146 amino acids each) — arranged as a pair of αβ dimers. Each chain folds into a compact globular shape, the globin fold, built almost entirely of α-helices, and each chain cradles one non-protein prosthetic group called heme. Heme is a porphyrin ring with a single iron atom held at its center by four nitrogen atoms from the ring itself. The iron has two additional coordination positions perpendicular to the ring plane: one binds a histidine residue of the globin chain (the "proximal histidine"), anchoring heme into the protein; the other, facing outward, is the binding site for O₂. Critically, the iron must remain in the ferrous (Fe²⁺) state to bind oxygen reversibly — oxidation to ferric (Fe³⁺) produces methemoglobin, which cannot carry O₂. The defining physiological property of hemoglobin is cooperativity: the four heme sites do not bind oxygen independently. Binding of O₂ at one subunit triggers a small shift in that subunit's iron and helix positions, which propagates through subunit-subunit contacts to make the remaining three subunits bind O₂ more readily. This produces a sigmoid (S-shaped) oxygen-hemoglobin dissociation curve rather than the hyperbolic curve seen with monomeric oxygen carriers like myoglobin, and it is this sigmoid shape that lets hemoglobin load efficiently at high O₂ tension and unload efficiently at low O₂ tension.
Mechanism
Step 1 — Iron-oxygen coordination. In the lung capillaries, alveolar O₂ partial pressure (PO₂) is about 100 mmHg. O₂ diffuses across the alveolar-capillary membrane into plasma and then into red blood cells, where it coordinates to the sixth position of a heme iron. This is not an oxidation-reduction reaction in the net sense — iron stays Fe²⁺ — but the Fe²⁺ shifts slightly toward the plane of the porphyrin ring upon binding O₂, because unliganded Fe²⁺ sits slightly outside the ring plane (domed) due to its ionic radius, while liganded Fe²⁺ is smaller and pulled in-plane.
Step 2 — Transmission through the proximal histidine. Because the iron is covalently tethered to the proximal histidine (His F8), the iron's movement into the ring plane drags that histidine, and the attached helix (the F-helix), along with it. This small ripple (on the order of 0.02–0.06 nm at the iron, amplified through the helix) is the mechanical trigger for everything that follows.
Step 3 — Quaternary conformational switch (T → R). Deoxyhemoglobin is locked in the Tense (T) state, stabilized by a network of ionic bonds (salt bridges) between the C-termini of the four chains and by weaker binding of O₂. As successive subunits bind oxygen, the helix movements destabilize these inter-subunit salt bridges. The whole tetramer eventually flips to the Relaxed (R) state, in which the two αβ dimers rotate about 15° relative to each other, salt bridges break, and every remaining heme site's affinity for O₂ rises roughly 300-fold compared with the T-state affinity.
Step 4 — Progressive, not simultaneous, binding. Binding of the first O₂ is difficult (T-state, low affinity); this is why the dissociation curve has a flat lower-left region. Each subsequent O₂ is progressively easier to bind as the equilibrium shifts toward R-state character, producing the steep middle portion of the curve. The fourth O₂ binds roughly 100–300 times more readily than the first — this ratio, the Hill coefficient (n ≈ 2.8–3.0 for human HbA, versus a theoretical maximum of 4.0 for perfect cooperativity), is the standard quantitative index of cooperativity.
Step 5 — Unloading in tissues. In systemic capillaries, tissue PO₂ is low (often 20–40 mmHg at rest, lower during exercise). Hemoglobin, now in the R state, begins losing O₂; as each O₂ leaves, the tetramer becomes progressively more T-like, and its remaining affinity drops — so the last O₂ molecules leave more easily than a non-cooperative carrier would release them. This self-accelerating unloading is exactly what tissues need during high metabolic demand.
Step 6 — Allosteric modulators fine-tune the curve.
| Modulator | Effect on curve | Physiological logic |
|---|---|---|
| ↑ H⁺ (↓pH) — Bohr effect | Shifts right (↓affinity) | Active tissue is acidic (CO₂ → carbonic acid); promotes O₂ unloading exactly where needed |
| ↑ CO₂ | Shifts right | CO₂ forms carbamino compounds with globin N-termini, stabilizing T state |
| ↑ Temperature | Shifts right | Exercising/febrile tissue is warmer; favors unloading |
| ↑ 2,3-bisphosphoglycerate (2,3-BPG) | Shifts right | Produced by RBC glycolysis; binds a pocket between the β chains, stabilizing T state; rises in chronic hypoxia, anemia, high altitude |
| Fetal hemoglobin (HbF, α₂γ₂) | Left relative to HbA | γ chains bind 2,3-BPG weakly, giving HbF higher O₂ affinity to pull O₂ from maternal blood |
The rightward shift produced by falling pH, rising CO₂, and rising temperature — all hallmarks of metabolically active tissue — is called the Bohr effect, named for Christian Bohr, who first described it in 1904.
The evidence
Perutz's crystallography (1959–1970s): seeing the T and R states directly. Max Perutz spent over two decades solving the X-ray crystal structure of hemoglobin, first achieving a low-resolution structure in 1959 and progressively refining it. By comparing crystals of deoxyhemoglobin with crystals of oxyhemoglobin (or with hemoglobin bound to CO as an O₂ analogue), Perutz's group directly visualized the quaternary rearrangement: the two αβ dimers rotate relative to each other by about 15°, inter-subunit salt bridges present in deoxyhemoglobin are absent in the oxy form, and the iron atom moves measurably relative to the porphyrin plane upon ligand binding. This structural evidence converted the T/R allosteric model — until then inferred purely from binding kinetics — into a directly observed molecular mechanism, work for which Perutz shared the 1962 Nobel Prize in Chemistry with John Kendrew (who solved myoglobin's structure).
The binding-curve evidence that demanded an allosteric model. Before any structure existed, physiologists already knew hemoglobin's O₂-binding curve was sigmoidal, not hyperbolic like myoglobin's. Archibald Hill's 1910 analysis of oxygen-binding data produced the empirical Hill equation, and fitting Hb data yielded a Hill coefficient (n) around 2.8–3.0 — far above the n = 1 expected for independent, non-interacting binding sites. This number was itself indirect evidence of cooperation between subunits, since a Hill coefficient greater than 1 is mathematically impossible unless binding at one site influences binding at others. The Monod-Wyman-Changeux (MWC) two-state allosteric model (1965) then formalized this: it proposed that the whole tetramer exists in equilibrium between two conformations (T, low affinity, and R, high affinity) and that ligand binding shifts the equilibrium toward R — a model whose predictions matched both the kinetic Hill-coefficient data and, later, Perutz's crystal structures.
Sickle cell hemoglobin as confirmatory (and clinically consequential) evidence of structure-function coupling. Linus Pauling's 1949 study of sickle cell hemoglobin (HbS) — using electrophoresis to show that HbS migrates differently from normal HbA — established hemoglobin as the first "molecular disease," directly linking an altered protein to a specific clinical phenotype. Vernon Ingram's 1956–1957 peptide fingerprinting then pinpointed the defect to a single amino acid substitution (glutamate → valine at position 6 of the β chain). This exposed hydrophobic valine promotes polymerization of deoxy-HbS into rigid fibers, distorting red cells into the sickle shape — powerful confirmation that hemoglobin's tertiary and quaternary structure, not just its heme chemistry, is what governs its physiological (and pathological) behavior.
Worked examples
Example 1 — Calculating total blood oxygen-carrying capacity. A patient has a hemoglobin concentration of 15 g/dL and arterial saturation (SaO₂) of 98%, PaO₂ 95 mmHg. Known constants: each gram of fully saturated hemoglobin carries 1.34 mL O₂ (the Hüfner constant, empirically close to the theoretical 1.39 mL/g); dissolved O₂ contributes 0.003 mL O₂ per dL of blood per mmHg PO₂. Step 1 — Hb-bound O₂: ``` O2(bound) = Hb (g/dL) × 1.34 (mL O2/g) × SaO2 = 15 × 1.34 × 0.98 = 19.7 mL O2 / dL blood ``` Step 2 — Dissolved O₂: ``` O2(dissolved) = 0.003 × PaO2 = 0.003 × 95 = 0.285 mL O2 / dL blood ``` Step 3 — Total arterial O₂ content: ``` CaO2 = 19.7 + 0.285 ≈ 20.0 mL O2 / dL blood ``` This means >98% of transported oxygen rides on hemoglobin, and dissolved plasma O₂ is nearly negligible — the number that motivated the overview's opening comparison. Now suppose this patient becomes severely anemic, with Hb dropping to 7 g/dL but SaO₂ unchanged at 98% (lungs still oxygenate normally). Recalculating: bound O₂ = 7 × 1.34 × 0.98 = 9.2 mL/dL; total ≈ 9.5 mL/dL — roughly half the oxygen-carrying capacity, despite completely normal lung function and normal PaO₂. This illustrates why anemia causes tissue hypoxia even when the respiratory system is functioning perfectly: the bottleneck is hemoglobin quantity, not oxygenation efficiency.
Example 2 — Reasoning through a right-shifted curve at altitude. A mountaineer ascends to 4,300 m, where atmospheric pressure drops to about 460 mmHg (versus 760 mmHg at sea level) and alveolar PO₂ falls to roughly 50–55 mmHg (versus ~100 mmHg at sea level). Immediate concern: at PO₂ 50 mmHg, sea-level-adapted hemoglobin would only be about 80–85% saturated (reading off the steep-to-plateau transition region of the curve), a meaningful drop from the 97–98% typical at sea level. Physiological compensation over days: chronic hypoxia stimulates red blood cell 2,3-BPG production, shifting the dissociation curve rightward — lowering hemoglobin's O₂ affinity. This seems paradoxical (why reduce affinity when oxygen is already scarce?), but the logic resolves once loading and unloading are considered separately: at the flat, high-PO₂ portion of the curve (lungs), a modest rightward shift barely reduces saturation, because the curve is nearly flat there. But at the tissue level, where PO₂ is much lower and the curve is steep, the same rightward shift substantially increases O₂ unloading for a given drop in PO₂. Net effect: altitude acclimatization sacrifices a small amount of pulmonary loading efficiency to gain a larger amount of peripheral unloading efficiency — a trade that only makes sense because the curve's shape differs at the lung end versus the tissue end.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "Hemoglobin's iron gets oxidized when it binds O₂." | Iron stays Fe²⁺ (ferrous) throughout normal O₂ binding and release; oxidation to Fe³⁺ (methemoglobin) is a separate, pathological/enzymatically-reversed process that cannot carry O₂. |
| "All four O₂ molecules bind with equal ease." | Binding is sequential and cooperative — the first O₂ binds with low affinity (T state), the fourth with much higher affinity (R state); this asymmetry creates the sigmoid curve. |
| "The Bohr effect means CO₂/H⁺ bind at the heme site." | CO₂ and H⁺ act at allosteric sites away from heme (carbamino groups at N-termini, protonatable side chains), stabilizing the T conformation rather than competing for the O₂ pocket. |
| "Fetal hemoglobin binds O₂ better because its heme is different." | HbF (α₂γ₂) has the same heme chemistry; its higher O₂ affinity comes from γ chains binding 2,3-BPG more weakly than β chains, shifting the T/R equilibrium toward R. |
| "Dissolved O₂ in plasma is a major transport route." | Dissolved O₂ contributes under 2% of total blood O₂ content at normal PaO₂; hemoglobin-bound O₂ dominates almost entirely. |
| "More hemoglobin always means better oxygen delivery." | Beyond a point, elevated hematocrit (e.g., polycythemia) raises blood viscosity, impairing flow and can reduce net O₂ delivery despite higher carrying capacity per liter. |
The scientist behind it
Origins · 1 minMax Ferdinand Perutz and how this was discovered
Austrian-born chemist who fled Nazi-annexed Austria in 1936 to work at Cambridge; interned as an "enemy alien" in Canada during WWII before returning to protein crystallography. Spent over 20 years wrestling X-rays into a hemoglobin structure, ultimately revealing the T-to-R quaternary shift that explains cooperative oxygen binding at atomic resolution.
Perutz began X-ray diffraction studies of hemoglobin crystals in the 1930s under J.D. Bernal, but the technical problem — solving the "phase problem" for a protein with thousands of atoms — seemed nearly intractable for decades. His breakthrough came through the method of isomorphous replacement: soaking hemoglobin crystals in heavy-metal salts (like mercury compounds) that bind at specific sites without disrupting the crystal lattice, then comparing diffraction patterns with and without the heavy atoms to extract phase information computationally. This technique, refined through the 1950s, finally yielded a low-resolution (5.5 Å) structure of hemoglobin in 1959 and progressively higher-resolution structures through the 1960s that resolved the T-state and R-state conformations distinctly. Perutz continued refining hemoglobin structures and their functional interpretation for the rest of his career, also founding and directing the Laboratory of Molecular Biology at Cambridge, which became one of the most Nobel-productive research institutions in history.
Origins · 1 minhow this was discovered
Where this matters today
Clinical diagnostics. Pulse oximetry, used continuously in hospitals and during anesthesia, exploits the different light-absorption spectra of oxygenated versus deoxygenated hemoglobin to estimate SaO₂ noninvasively — a direct application of hemoglobin's conformation-dependent optical properties. Arterial blood gas analysis quantifies PaO₂, SaO₂, and calculated CaO₂ to distinguish hypoxemia (low PaO₂) from anemic hypoxia (low Hb) from histotoxic hypoxia (e.g., cyanide poisoning, where O₂ delivery is normal but cells cannot use it).
Hemoglobinopathy management. Understanding the E6V mutation in sickle cell disease has driven modern gene-editing therapies (CRISPR-based reactivation of fetal hemoglobin production to dilute HbS polymerization) that received regulatory approval in the 2020s. Thalassemias, caused by reduced α- or β-globin chain production, are diagnosed and monitored using the same structural logic — imbalance between chain types disrupts normal tetramer assembly.
Carbon monoxide poisoning and antidotes. CO binds hemoglobin's heme iron with roughly 200–250 times the affinity of O₂, competitively occupying oxygen sites and additionally shifting remaining subunits toward higher O₂ affinity (a leftward curve shift) — worsening tissue O₂ unloading even at sites still carrying O₂. Treatment with high-flow or hyperbaric oxygen works by mass-action competition, displacing CO from heme.
High-altitude physiology and athletic training. Altitude acclimatization protocols for athletes and mountaineers are designed around the 2,3-BPG-mediated rightward curve shift described above; blood doping bans in competitive sport exist precisely because artificially raised hemoglobin/hematocrit alters O₂-carrying capacity.
Fetal-maternal medicine. HbF's higher O₂ affinity relative to maternal HbA is what allows placental oxygen transfer to the fetus down a favorable gradient — clinically relevant in understanding fetal hemoglobin persistence disorders and neonatal oxygen physiology, and conceptually a bridge to Oxygen-Hemoglobin Dissociation Curve and Carbon Dioxide Transport, where these same allosteric principles govern gas exchange in both directions.
Ask the tutor: How does the shape of the dissociation curve change quantitatively at each step of gas exchange from alveolus to mitochondrion?
Why doesn't myoglobin show cooperative oxygen binding like hemoglobin does?
Myoglobin is a single-chain, single-heme protein found in muscle, so it has no other subunits to communicate with — cooperativity requires at least two interacting binding sites capable of shifting a shared quaternary conformation. Myoglobin's oxygen-binding curve is therefore a simple hyperbola, reflecting independent, non-cooperative binding governed by straightforward mass-action equilibrium. This actually suits myoglobin's biological role: it needs high affinity even at low PO₂ to extract oxygen from blood and store it within muscle for release during intense contraction, rather than the tunable, environment-sensitive delivery role hemoglobin plays in blood.
What exactly is the Bohr effect, and why does it make physiological sense?
The Bohr effect is the rightward shift of hemoglobin's oxygen dissociation curve caused by increased CO₂ and decreased pH, first described by Christian Bohr in 1904. Mechanistically, H⁺ ions and CO₂ (via carbamino compound formation) stabilize the low-affinity T conformation, lowering hemoglobin's overall O₂ affinity. Physiologically this is elegant: metabolically active tissue generates CO₂ and lactic/carbonic acid, locally acidifying its capillary blood exactly where more oxygen unloading is needed — the chemistry of local metabolism directly promotes local oxygen delivery without requiring any nervous or hormonal signal.
How does carbon monoxide poisoning kill, mechanistically, given that CO binds the same site as O₂?
Carbon monoxide binds heme iron with roughly 200–250 times greater affinity than O₂, so even low ambient CO concentrations progressively occupy hemoglobin's binding sites, directly reducing the number of sites available for O₂. Additionally, when CO occupies some subunits, it stabilizes the R (high-affinity) conformation in the whole tetramer, which paradoxically increases the remaining subunits' grip on any O₂ they do carry — shifting the dissociation curve left and impairing oxygen release in tissues. The combined effect (reduced O₂ capacity plus impaired unloading of what O₂ remains) causes tissue hypoxia disproportionate to the fraction of hemoglobin actually occupied by CO.
Does the Hill coefficient of ~2.8-3.0 mean hemoglobin only has three cooperative binding events?
Not exactly — the Hill coefficient is an empirical index of the steepness of the binding curve, not a literal count of interacting sites. A perfectly, maximally cooperative four-site protein would show a theoretical Hill coefficient approaching 4 (all sites behaving as if bound simultaneously), while a value of 1 indicates complete independence. Hemoglobin's real value near 2.8–3.0 indicates strong but imperfect cooperativity — consistent with the two-state MWC model, in which some intermediate liganded states still have appreciable populations rather than the tetramer flipping instantaneously between only fully-T and fully-R states.
Why does fetal hemoglobin have higher oxygen affinity than adult hemoglobin?
Fetal hemoglobin (HbF) is composed of two α chains and two γ chains (α₂γ₂) rather than adult HbA's two α and two β chains. The γ chains have a different amino acid sequence at the 2,3-BPG binding pocket, which lies between the two β chains in HbA; this altered pocket binds 2,3-BPG much more weakly in HbF. Since 2,3-BPG normally stabilizes the low-affinity T state, its weaker binding in HbF leaves the tetramer biased more toward the R state, raising overall O₂ affinity. This lets fetal blood extract oxygen from maternal blood across the placenta, where maternal hemoglobin is simultaneously releasing it.
How is total blood oxygen content actually calculated, and why does hemoglobin dominate the number?
Total arterial oxygen content (CaO₂) is calculated as (Hb concentration in g/dL × 1.34 mL O₂/g × fractional saturation) plus (0.003 × PaO₂ in mmHg) for dissolved oxygen. With typical values (Hb 15 g/dL, SaO₂ 98%, PaO₂ 95 mmHg), the hemoglobin-bound term contributes about 19.7 mL O₂/dL while dissolved oxygen contributes only about 0.29 mL O₂/dL — under 2% of the total. This dominance exists because oxygen's solubility in aqueous plasma is intrinsically very low; hemoglobin functions precisely to circumvent that solubility limit by providing millions of high-affinity binding sites per red cell.
What would happen to oxygen delivery if hemoglobin bound oxygen non-cooperatively, with the same overall affinity?
A non-cooperative (hyperbolic) hemoglobin with an affinity tuned to load well at lung PO₂ (~100 mmHg) would also bind O₂ too tightly to release much of it at typical tissue PO₂ (~40 mmHg), since a hyperbolic curve is already near-saturated well before typical arterial pressures and stays relatively flat down to much lower pressures than hemoglobin's actual curve. Conversely, an affinity tuned to unload well in tissue would fail to saturate fully in the lungs. Cooperativity is what allows a single molecule to satisfy both requirements simultaneously: steep transition through the physiological PO₂ range between lung and tissue, high saturation at the lung end and substantial release at the tissue end.