The BOHR effect before Perutz


  • Maurizio Brunori

    Corresponding author
    1. Department of Biochemical Sciences, Sapienza-University of Rome, P.le A. Moro, 5-00185 Roma, Italy
    • Department of Biochemical Sciences, Sapienza-University of Rome, P.le A. Moro, 5-00185 Roma, Italy
    Search for more papers by this author
    • Tel.: +39-06-4450291; Fax: +39-06-49910717.


Before the outbreak of World War II, Jeffries Wyman postulated that the Bohr effect in hemoglobin demanded the oxygen linked dissociation of the imidazole of two histidines of the polypeptide. This proposal emerged from a rigorous analysis of the acid–base titration curves of oxy- and deoxy-hemoglobin, at a time when the information on the chemistry and structure of the protein was essentially nil. The magnetochemical properties of hemoglobin led Linus Pauling to hypothesize that the (so called) Bohr histidines were coordinated to the heme iron in the fifth and sixth positions; and Wyman shared this opinion. However, this structural hypothesis was abandoned in 1951 when J. Wyman and D. W. Allen proposed the pK shift of the oxygen linked histidines to be the result of “…a change of configuration of the hemoglobin molecule as a whole accompanying oxygenation.” This shift in paradigm, that was published well before the 3D structure of hemoglobin was solved by M.F. Perutz, paved the way to the concept of allostery. After 1960 the availability of the crystallographic structure opened new horizons to the interpretation of the allosteric properties of hemoglobin.

In 1937, one-year after his arrival in Cambridge (UK), Max Perutz took the first X-ray diffraction pictures of hemoglobin (Hb) crystals that he obtained from Gilbert Adair, the man who discovered by osmotic pressure that Hb is a tetramer (MW = ∼67,000) and published the equation to describe the binding of four molecules of O2 to a tetrameric protein. More or less at the same time Jeffries Wyman published, in the Journal of Biological Chemistry, a pioneering paper [1] presenting an investigation on Hb aimed at identifying the chemical groups responsible for the Bohr effect. At the time very little was known about the chemistry of the protein: the presence of the α- and β-chains was unknown not to mention their amino acid sequences, and even their amino acid composition was preliminary. Nevertheless by a rigorous analysis of simple measurements of the O2 binding curve and the acid–base titration of Hb, Wyman came to the conclusion that the Bohr effect implied a change in pK of the imidazole group of two histidine residues, coupled to binding of O2 (or CO) at the heme iron. It is interesting to outline the analysis involved.

Evidence for the dependence of O2 affinity of Hb on CO2 was published in 1904 by Bohr et al. [2]. These authors discovered that increasing the partial pressure of CO2 decreases the overall affinity of Hb for O2. It was later shown that this phenomenon is due to acidification of the medium following hydration of CO2; and thus the linkage between pH and O2 affinity in Hb has been called the Bohr effect [3], which has been ever since widely employed to define the pH dependence of enzymatic activity.

Given that, upon binding of O2 to deoxy Hb some dissociable group becomes more acidic, the reaction must be associated with the release of protons. Therefore, in unbuffered solution a change in pH should be detected upon O2 binding. When Hb is the only buffer, the measured ΔpH may be expressed as ΔH+ (i.e. the number of protons released per O2 bound to each heme) if the acid–base titration of the protein is known. Figure 1 reproduces the data by German and Wyman [1] obtained as the difference between the acid–base titrations of deoxy- and oxy-Hb from horse. It may be seen that, upon oxygenation, there is a release of protons around physiological pH called the alkaline Bohr effect; which turns into an uptake of protons below pH ∼ 6, called the acid or reverse Bohr effect. Using these experimental data, the acid dissociation constants in deoxy- and oxy-Hb were calculated (legend to Fig. 1). In addition, the data in Fig. 1 could be used to predict the Bohr effect using the linkage equation:

equation image

which correlates (i) the dependence of the charge of the protein (X) on O2 saturation (Y) with (ii) the O2 affinity (log p1/2) as a function of pH [4]. The agreement was quite satisfactory.

Figure 1.

Difference in charge between deoxy-Hb and oxy-Hb. Smooth curve calculated for the case in which the pK of one histidine changes upon oxygenation from 7.81 to 6.80, and that of the other histidine from 5.25 to 5.75. The maximum number of protons released per heme around physiological pH amounted to ∼0.5–0.6/O2 bound (from Ref. 4).

Since several possible amino acid side chains (imidazole, α-amino, sulphydryl, …) may be compatible with the pK's (see Fig. 1), Wyman [5] decided to determine the enthalpy change of the O2 linked groups. This was achieved by measuring at different temperatures (i) the acid–base titration curves of oxy and deoxy Hb, and (ii) the O2 equilibrium curve on the assumption that temperature would shift the binding isotherm along the O2 pressure axis without changing its shape. It may be recalled that the overall enthalpy change associated to O2 binding at constant pH includes the heat contribution due to dissociation of the ligand-linked protons; in so far as the shape of the saturation curve is temperature invariant (which at the time seemed to be verified), the same number of protons would be released by the oxygenation of each of the four hemes in the tetramer, over the whole pH range explored. The enthalpy change estimated for both O2 linked ionizable groups was +6,500 cal/heme, compatible with the value characteristic of the imino group of histidine.

The results outlined above suggested a structural hypothesis in which all the four hemes are bound to structurally identical local conformations of the globin, and the irons are coordinated to histidines. Coryell and Pauling [6] published their structural interpretation of the acidity of what they called the heme-linked acid groups; and for the alkaline Bohr effect they postulated that “...the imino group of a histidine residue whose 3-nitrogen atom is strongly coordinated by either an essentially ionic or an essentially covalent bond with the iron atom.” This hypothesis was consistent with and in fact based on the demonstration [7] that oxygenation of Hb is associated to a large change in magnetic moment, from paramagnetic deoxyHb to diamagnetic oxy (or CO) Hb. An identical change in magnetic moment was to be demonstrated by Taylor [8] for myoglobin, the monomer with higher O2 affinity and no Bohr effect.

The outbreak of World War II overturned everybody's life, including those of our heroes. Max Perutz, being an Austrian Jew and thus potentially sympathetic to the Nazis (sic!), was arrested in May 1940, and sent to a prisoner of war camp, first in England and than in Canada. He was eventually allowed back to Cambridge in January 1941, where he was asked to work on a war project called Habakkuc [9]. Wyman was also engaged in a research project on smoke screens and sonar for the US Navy [10], and eventually was at sea on a destroyer where life was generally quiet (so he told me!) but for a really “exciting” encounter with a German U-Boat! About Pauling, I have no info….

When WWII finally ended, they all went back to Hb, but it took several years to arrive at the conclusion that the Pauling/Wyman interpretation of the heme-linked acidic groups was incorrect. In 1948 at the Barcroft Memorial Conference in Cambridge (UK), essentially the same data and concepts elaborated and published before the war were presented. The volume of the Proceedings [11] was given to me years later by my Professor Alessandro Rossi Fanelli, a biochemist at the University of Pavia and the only Italian invited to the Cambridge meeting. Reading in that book the paper by Max Perutz demonstrates how difficult it was to obtain sensible information by X- ray diffraction, before the advent of isomorphous replacement; the only interesting result was that oxy-Hb has an axis of symmetry that is lost upon deoxygenation, in accord with the decrease in entropy detected when O2 or CO are bound. Perutz et al. [12] published in 1960 the low resolution structure of horse metHb that, however, had essentially no information about the molecular mechanism of functional control. This paper appeared in Nature side-by-side with the (higher) resolution structure of sperm whale myoglobin by Kendrew et al. [13], who showed indeed that the heme iron is at a short distance from two histidines, the so-called proximal His(F8), directly coordinated to the Fe, and the distal His(E7), which faces the ligand binding site. Yet myoglobin has no Bohr effect [3].

A decade later, a detailed structural hypothesis for the allosteric behavior of Hb and the ionizable groups responsible for the Bohr effect was published by Perutz [14] in a famous paper called “Stereochemistry of cooperative effects in haemoglobin.” based on a comparison of the structures of met and deoxy-Hb. Here he analyzed the structural basis of the functional peculiarities of Hb, and presented his hypothesis for the residues involved in the alkaline Bohr effect (nowadays in every Biochemistry textbook). A salt bridge between the C-terminal His146(β) and Asp94(β) in deoxyHb significantly increases the pK of the imidazole, that becomes “normal” in oxyHb, in which the C-terminus is disordered. This accounts for one-half of the Bohr effect. Perutz also proposed that at least one quarter of the Bohr effect comes from Val1(α), whose normal pK in oxy-Hb is raised in the deoxy form by interaction with the carboxylate of the C-terminal Arg141(α). The residual contribution was more uncertain but he ventured to invoke His122(α), which in deoxy-Hb is kept close to Asp126(α). Overall, this detailed structural biology view was not inconsistent with Wyman's hypothesis given that α-amino terminal and imidazole have approximately the same enthalpy change. Actually, the crucial role of the C-terminal residues of the α- and β-chains had been demonstrated in 1961 by Antonini et al. [15] who digested human oxy-Hb with carboxypeptidases A or B, proteolytic enzymes just discovered that removed Arg-Tyr- from the α-chain and His-Tyr- from the β-chain; the modified tetramers had high affinity, no cooperativity and no Bohr effect.

In spite of bitter criticisms and conflicting results, Perutz' hypothesis is considered essentially correct [16, 17]. Nevertheless, it should be remarked that the most interesting conceptual breakthrough, which highlighted the crucial role of a ligand-linked global conformational change of the whole tetramer mediating heme–heme interactions and perturbing the pKs of the Bohr groups, had already been proposed much earlier. In 1951 Wyman and Allen [18] published a theoretical paper in the Journal of Polymer Science in which they argued for a change of paradigm. I would guess that only a dozen people have carefully read that paper, largely written by Wyman while in Japan as scientific Attaché to the U.S. Embassy (Jan–Nov, 1950). One reads that: “.....the reason why certain acid groups are affected by oxygenation is simply the alteration of their position and environment which results from the change of configuration of the hemoglobin molecule as a whole accompanying oxygenation.” This was consistent with the remarkable experiment by Haurowitz [19], a distant relative to Perutz, who demonstrated that deoxy-Hb crystals crack upon oxygenation, revealing a striking difference in shape and surface structure between oxy- and deoxy-Hb.

It is interesting to follow the reasoning whereby between 1948 and 1951 Wyman changed his mind, rejecting the direct iron-histidine linkage as the mechanism for the Bohr effect and proposing the new view that possibly paved the way to allostery. An obvious objection to the role of a direct linkage between iron and histidine was the fact that oxygenation of myoglobin, which has no Bohr effect, is nevertheless associated to a drop in magnetic moment from paramagnetic to diamagnetic, just like Hb [see Refs. 7 and 8]. In addition, the so-called redox Bohr effect (extensively analyzed by Wyman and by Pauling) is the same as the O2 Bohr effect [4], although ferric Hb, the product of oxidation of deoxy, is also paramagnetic. But possibly even more convincing was the result obtained by Riggs [20], working at Harvard for his PhD, that the two Hbs from the bullfrog are characterized by the same Hill coefficient (n = 2.8) yet the adult has full Bohr effect while the tadpole Hb has none; Wyman and Allen [18] wrote: “ is to be expected that different Hbs, differing in the structure of the polypeptide chains, should have different Bohr effects.” The objection they highlighted and discussed was the puzzling observation that the shape of the O2 binding curve was essentially pH independent, contrary to expectation based on linkage. This inconsistency was tentatively by-passed by an explanation based on a “rectangular model” for cooperativity, with fairly strong interactions within each dimer and much weaker ones between dimers [4], which proved incorrect many years later. In fact much more accurate data by Imai [21] have shown that the O2 binding curve of human Hb is pH dependent, but mostly at the extremes of the saturation curve.

The paper by Wyman and Allen [18] paved the way to the concept that a global conformational change of the whole protein is the crucial event in functional regulation, be it heme–heme interactions or the Bohr effect. Their visionary perspective went quite a bit further as they wrote: “......if we are prepared to accept hemoglobin as an enzyme, its behavior might give us a hint as to the kind of process to be looked for in enzymes more generally,” and “A protein, with its enormous complexity of possible configurations and corresponding richness of entropy effects, would on this basis be uniquely fitted to play the role of an enzyme.” This paper probably had a minuscule number of citations. Nevertheless, I would guess that Jacques Monod knew of it when, in 1964, decided to recruit Wyman as an author of the famous Monod–Wyman–Changeux theory of allostery [22, 23]. My path trough science was enlightened by the intelligent people I had the fortune to collaborate with; this memoir is a small tribute to their shining vision.