Hydrogen Ion Equilibria in Fish Haemoglobins

An important functional property of haemoglobins is their reversible hydrogen ion exchange between the protein and the solvent. Haemoglobin (Hb) serves acid-base homeostasis by being the predominant nonbicarbonate buffer in blood. The Hb-H⁺ equilibria guarantees effective blood CO₂ transport, via the binding and release of H⁺ when red cells cross capillaries in tissues and respiratory surfaces, respectively. The overall H⁺ binding properties are determined by the pK of individual titratable groups, and provide Hb with a constellation of charges that are important for molecular structure and binding of heterotropic ligands, and which control red cell pH via the Donnan-like distribution of H⁺ across the red cell membrane.

In contrast to the large amount of information available on the O₂ transporting properties of vertebrate Hbs, relatively little is known about interspecific differences in H⁺ binding properties and their molecular origin. The intrinsic H⁺ binding properties of Hbs consist of buffering at a constant degree of oxygenation (i.e. with pK values fixed) and oxygenation-linked H⁺ binding (i.e. due to oxygenation-dependent pK changes - the Haldane effect). The purpose of this study was to compare H⁺ equilibria of teleost Hbs with that of elasmobranch and mammalian Hbs, and to seek an explanation for observed differences within the primary structure of the Hbs. The study shows that the H⁺ equilibria of teleost Hbs differ significantly from those in other vertebrate groups, and this correlates well with structure.

Red cells were separated from freshly drawn blood of carp (Cyprinus carpio), rainbow trout (Salmo gairdneri), spiny dogfish (Squalus acanthias) and pig (Sus scrofa domesticus) by centrifugation, and washed twice in cold 0·9% NaCl. The red cells were haemolysed by addition of weakly buffered (5 mmol I⁻¹ Na-Hepes, pH7·8) distilled water and subsequent freezing, and the cell debris was removed by centrifugation. The haemolysates were dialysed against distilled water at 5°C, and passed repeatedly through a mixed-bed ion-exchange column (Amberlite MB1, IRA 400 and IR120, BDH). KC1 was added to the resulting isoionic Hb solutions to a concentration of 0·1 mol I⁻¹, and they were then stored frozen (–24°C for a maximum of 2 days).

In each experiment a uniform Hb solution was divided into four 3 ml portions, which were equilibrated (40 min) at 15 °C with either humidified oxygen (oxyHb) or humidified pure nitrogen (deoxyHb) in Eschweiler tonometers (Kiel, FRG). Starting from the isoionic pH (zero net proton charge), two 3 ml Hb portions were used for titration with HC1, and two 3 ml portions for titration with NaOH, in the oxygenated and deoxygenated states, respectively. Freshly prepared and analysed 0·1 mol I⁻¹ HC1 or 0·1 mol I⁻¹ NaOH solutions were added from a Hamilton syringe fitted with a titration adaptor. pH was measured with Radiometer (Copenhagen) combined pH electrodes (GK2321C) connected to PHM71 acid-base analysers and displayed on Rec80 recorders. The tetrameric haemoglobin concentrations were measured spectrophotometrically after bubbling with humidified oxygen. Three or four separate experimental runs were performed for each type of haemoglobin, and gave reproducible results for each species.

The H⁺ titration curves of oxygenated and deoxygenated haemoglobins from two teleosts (carp and rainbow trout), one elasmobranch (spiny dogfish) and one mammal (pig) are drawn to the same scale in Fig. 1 with zero net proton charge (Tanford, 1962) as the reference point. The slope of the individual curves gives the buffer values (mol H⁺ mol tetramer⁻¹ pH unit⁻¹) when the oxygenation degree, and thus the protein conformation and pK values of titratable amino acid residues, tends to be fixed. These buffer values vary with pH and oxygenation degree as well as with species (Fig. 2A). The vertical distance (ΔZ_H) between the titration curves for oxygenated and deoxygenated Hbs (Fig. 1) gives the fixed-acid Haldane effect (mol H⁺ taken up per mol tetramer upon deoxygenation at constant pH), which results from pK changes of specific titratable groups associated with the conformation change upon deoxygenation. The magnitude of this oxygenation-linked H⁺ binding varies with pH and with species (Fig. 2B).

The buffer values of the Hbs from carp and trout were low in comparison with the buffer values of dogfish and pig Hbs at physiological pH values (Figs 1, 2A). This suggests that low intrinsic buffer values are a general feature of teleost Hbs, as it also applies to other species (tuna, Brunori, 1966; eel, Breepoel et al. 1980; tench, Jensen & Weber, 1985). Above pH 7, the buffer values of the Hb from an elasmobranch, the dogfish, in contrast, were very high, exceeding those of pig Hb (Fig. 2A). The H⁺ equilibria of pig Hb (Fig. 1) resemble those of human Hb (Rollema et al. 1975) and horse Hb (Janssen et al. 1972), and may represent the general trend in mammalian Hbs.

Low buffer values in teleost Hbs mean that the number of titratable groups must be reduced in comparison with other vertebrate Hbs. The type of groups titratect can be divided into three classes (Tanford, 1962; De Bruin & van Os, 1968). At acid pH (i.e. below about pH 6) the groups titrated are carboxyl groups of aspartic acid and glutamic acid residues, and at alkaline pH (i.e. above about pH9) it is the ∈-amino group of lysine and the guanidyl group of arginine, plus phenolic and sulphhydryl groups that are titrated. In the intermediate ‘neutral’ pH range, H⁺ buffering by Hb mainly concerns the imidazole group of histidine residues and the terminal α-amino groups (i.e. groups with pK values close to physiological pH). Thus it is tempting to seek an explanation for low buffer values in teleost Hbs in a reduced number of these neutral groups within the Hbs. An analysis of the available amino acid sequences of Hb chains confirms this hypothesis (Table 1). Teleost Hbs are characterized by a much lower histidine content (2–6 His per chain) than that seen in other phylogenetic groups of fish (dipnoans and elasmobranchs) and in amphibians, reptiles, birds and mammals (typically 8–12 His per chain) (Table 1). Cyclostomes seem to be the only other group of vertebrates that share with teleosts the property of a low haemoglobin His content. Hagfish and river lamprey have 4 and 2 His residues in their monomeric Hb, respectively (Kleinschmidt & Sgouros, 1987). The correlation of buffer values with the haemoglobin His content is confirmed by the parallel rise in buffer values (Fig. 2A) and His content (Table 1) when comparing carp and trout with pig and dogfish. In addition to a low histidine content, the acetylation of the N-terminal amino acids in α-chains of teleost Hbs (Table 1) contributes to lower buffer values of teleost Hbs. Acetylation (i.e. R-NH-CO-CH₃ in contrast to R-NH₂) excludes proton exchange at physiological pH, so that tetrameric teleost Hbs contain only two titratable α-amino groups (i.e. those of the β-chains) in comparison with four titratable α-amino groups in most other vertebrate Hbs.

The reduced His content in teleost Hbs can partly be explained by the different allosteric modifiers used among vertebrate Hbs. In mammalian Hbs, position H21β and often also position NA2β are occupied by histidine residues, which are involved in 2,3-diphosphoglycerate (DPG) binding. In teleosts these histidines are substituted by Arg and Glu or Asp residues, respectively, which makes the phosphate binding site in the cleft between the two β-chains stereochemically complementary to ATP and GTP (Perutz, 1983; Gronenborn et al. 1984). Such amino acid substitutions will, however, not always necessitate a reduction in the total His content, as can be seen in crocodilian haemoglobins. In these Hbs, His H21β and His NA2β have also been replaced (by Ala and Pro or Ser, respectively), which partly explains their phosphate insensitivity and their use of bicarbonate as allosteric modifier (Perutz, 1983). The β-chain histidine content of crocodilian Hbs is, however, very high (12–14 His per chain), as is the α-chain His content (11–12) (Leclercq et al. 1981).

Not all the histidine residues included in Table 1 are titratable. Some will be buried in the interior of the molecule, or otherwise be unavailable for hydrogen ion exchange. In differential titration curves (i.e. -dpH/dZ_Hvs Z_H), the inflection points of ordinary titration curves are localized as two peaks, the distance between which gives an accurate measure of the number of imidazole and α-amino groups titrated (De Bruin & van Os, 1968; Janssen et al. 1970,1972). This procedure also applies to naturally occurring mixtures of Hb components (Janssen et al. 1972). The differential titration curves of carp Hb suggest that nine neutral groups are titratable per tetramer (Fig. 3). With two of these being the α-amino groups of the β-chains, this leaves seven titratable histidine residues out of 18 histidines per α₁β₂ tetramer (Table 1). This compares with 20 and 22 titratable histidines out of 38 histidines in human and horse Hbs, respectively (Janssen et al. 1970, 1972). Thus, low buffer values seem to be a unique and fundamental feature of teleost Hbs, and correlate well with both a reduced total His content and a reduced titratable His content.

Low intrinsic buffer values seems to be maladaptive with respect to the role of Hb in blood H⁺ buffering and CO₂ transport, but can be compensated by the possession of large Haldane effects. Indeed, the data reveal an inverse relationship between the magnitudes of buffer values and the magnitudes of fixed-acid Haldane effects (Fig. 2A,B). Thus, two fundamentally different strategies to safeguard the Hb-H⁺ exchange essential to blood CO₂ transport are evident. The elasmobranch dogfish has very large buffer values, but only a minor, and almost insignificant, Haldane effect, whereas the carp has an exceptionally large Haldane effect, but small buffer values. Some variability in the magnitude of Haldane effects is present among teleosts. In trout Hb it is smaller than in carp Hb. Judged from oxygen equilibria (i.e. the Bohr effect, which is a phenomenon equivalent to the Haldane effect) within teleosts (e.g. Jensen, 1989), the Bohr-Haldane effects of carp and trout haemoglobins may encompass a typical range of values found within teleosts. The present data on carp and trout may accordingly also span a typical range of H⁺ equilibria encountered by teleosts.

The maximum number of H⁺ taken up upon deoxygenation at constant pH is 3·8 per tetramer in carp Hb at pH6·6, whereas it is 2·7 in trout (at pH6·95), 1·6 in pig (at pH7·4) and 0·75 in dogfish (at pH7·5) under the present experimental conditions (Fig. 2B). The relatively low pH at which the maximal Haldane H⁺ uptake occurs in carp (pH 6·6 in comparison with a typical red cell pH of above 7), does not reduce the physiological significance of the large Haldane effect. In vivo, the presence of organic phosphates (ATP and GTP) within the red cells will actually increase the magnitude of the fixed-acid Haldane effect and shift the pH of its maximum towards a higher pH (Jensen & Weber, 1985). This safeguards a very high H⁺ uptake when red cells cross tissue capillaries. Also, the physiologically insignificant specific CO₂ effect on O₂ affinity in teleost Hbs (Weber & Jensen,1988) suggests that the amount of oxylabile carbamino formation is small. This tends to keep the Haldane effect high, by avoiding the release of H⁺ from the preferential reaction of CO₂ with amino groups in deoxyHb. That low buffer values and high Haldane effects appear to be restricted to teleost Hbs means that the pronounced influence of Hb O₂-saturation on red cell pH in teleosts (Jensen, 1986) is also a special feature among vertebrates.

The Haldane effect results from an increase in pK of specific amino acid residues when their microenvironment changes as result of the change in protein conformation upon deoxygenation. The groups contributing are, however, only partly resolved for fish Hbs. In human Hb in the absence of organic phosphates Bohr-Haldane protons bind mainly to His HC3β, Vai NAlα and Lys EF6β (Perutz, 1983; Riggs, 1988). In teleost Hbs, His HC3β is also a major Bohr-Haldane group, and it may contribute more than in mammalian Hbs (some 50% of the large effect in carp; Parkhurst et al. 1983). Val NA1α (the N-terminal amino acid), however, is acetylated in teleosts (Ac-Ser or Ac-Thr), and will not contribute to the Bohr effect. Thus, a large fraction of the Bohr-Haldane effect remains to be accounted for structurally.

Why do many teleosts Hbs have high Haldane effects and low buffer values to safeguard the H⁺ uptake necessary for CO₂ transport? One possibility is that a high Haldane effect allows a large H⁺ uptake without producing a pH decrease which, according to the mass law equation for CO₂ hydration, gives a larger HCO₃⁻ formation when CO₂ is added in tissues than when the Haldane effect is low but buffer values are high (since here pH decreases). This could compensate for a relatively slow HCO₃⁻/C1⁻ exchange across the red cell membrane (Jensen,1989) Thus, even though the Haldane effect is very important for CO₂ transport in mammals (Wieth et al. 1982), its role must be even greater in many teleosts.