Haemoglobin H⁺ equilibria in lamprey (Lampetra fluviatilis) and hagfish (Myxine glutinosa)

Reversible binding of hydrogen ions (H⁺) to haemoglobin (Hb) is of central importance for blood CO₂ transport. Carbon dioxide produced in the tissues diffuses into the red blood cells where carbonic anhydrase catalyses the rapid hydration of CO₂ to carbonic acid, which quickly dissociates to H⁺ and HCO₃⁻. For the blood to have a high CO₂-carrying capacity, the reaction products must be rapidly removed from the red blood cell cytosol, allowing the equilibrium reactions inside the red blood cell to proceed further towards the formation of H⁺ and HCO₃⁻. In general, H⁺ is effectively removed through reversible binding to Hb, whereas HCO₃⁻ is moved to the plasma in exchange for Cl⁻via the anion exchanger in the erythrocyte membrane.

Vertebrates, however, exploit different strategies of blood CO₂ transport as a result of differences in the functional properties of their haemoglobins and differences in their red blood cell HCO₃⁻/Cl⁻ exchange capacity. In elasmobranchs, H⁺ binding to Hb is ensured by high buffer values at a constant Hb conformation, whereas oxygenation-linked H⁺ binding (fixed-acid Haldane effect) to Hb is very low. Many teleosts, in contrast, have low buffer values (because of low numbers of titratable histidines and α-NH₂ groups) and large Haldane effects (Jensen, 1989). The special Hb/H⁺ equilibria in teleosts result in a large increase in red blood cell pHi upon deoxygenation and enhance red blood cell HCO₃⁻ formation, which may compensate for rate limitations in HCO₃⁻/Cl⁻ exchange (Jensen, 1989). This view is supported to some extent by species differences in red blood cell anion-transport capacity (Jensen and Brahm, 1995).

Agnathans are particular interesting in this regard, since the lamprey (Nikinmaa and Railo, 1987) and hagfish (Ellory et al., 1987) are practically devoid of red blood cell anion exchange. Rather than transporting most CO₂ in the blood as plasma HCO₃⁻, lampreys carry CO₂ as free HCO₃⁻ inside the red blood cells (Tufts and Boutilier, 1989). In the virtual absence of HCO₃⁻/Cl⁻ exchange, agnathans apparently rely exclusively on H⁺ removal to ensure an appropriate blood CO₂-carrying capacity. This makes it pertinent to study the H⁺-binding properties of agnathan Hbs. Agnathans are the most ancient vertebrates, and their Hbs differ markedly from those of other vertebrates. Lamprey and hagfish Hbs are monomeric when oxygenated and aggregate into oligomers upon deoxygenation and when pH is reduced (Riggs, 1972; Nikinmaa et al., 1995; Fago and Weber, 1995). The amino acid sequences of hagfish and lamprey Hbs, however, show large differences, suggesting evolutionary divergence between these two groups of agnathans some 400–500 million years ago (Liljeqvist et al., 1982). Recent analysis of the protein-coding genes of mitochondrial DNA supports an early separation of hagfish and lampreys during evolution and even suggests that the hagfish (Myxiniformes) are a sister group of the vertebrates (Rasmussen et al., 1998). The large differences in primary structure between hagfish and lampreys Hbs are likely to cause major differences in Hb/H⁺ equilibria in the two groups. The aim of the present study was to evaluate the H⁺-binding properties of lamprey and hagfish haemoglobins to determine how these properties have evolved in the virtual absence of red blood cell anion exchange.

River lampreys (Lampetra fluviatilis) were kept in fresh water in 400 l tanks at Odense University, Denmark, and Atlantic hagfish (Myxine glutinosa) were held in sea water with a salinity of 32 ‰ at Kristineberg Marine Biological Station, Sweden. Blood was sampled from the caudal vessel (lamprey) or sinus (hagfish) of animals anaesthetised with MS 222 (3-aminobenzoic acid ethyl ester, Sigma). The red blood cells were separated from the plasma by centrifugation and subsequently washed three times in physiological saline. Following the final wash and centrifugation, the packed red blood cells were frozen in liquid nitrogen. The hagfish red blood cells were transported to Odense while frozen in liquid N₂. Distilled water was added upon thawing, and cell debris was removed by centrifugation. The haemolysate was stripped from cell solutes by passing it three times through a mixed-bed ion-exchange column (Amberlite MB1, BHD). The iso-ionic Hb solution was brought to a KCl concentration of 0.1 mol l⁻¹. This KCl concentration is commonly used in Hb titration studies and is close to that inside the red blood cells of many vertebrates, including the lamprey (Nikinmaa, 1990) and even the hagfish (Fincham et al., 1990). The total concentration of Hb was measured spectrophotometrically both as oxyhaemoglobin (using a millimolar extinction coefficient of 16.1 at 576 nm) and after converting a subsample of the Hb to cyanmethaemoglobin (using a millimolar extinction coefficient of 11 at 540 nm). The Hb concentrations determined using these two methods typically differed by less than 5 %, reflecting a minimal presence of methaemoglobin.

Hydrogen ion titration curves were recorded using a computer-controlled Radiometer (Copenhagen, Denmark) TitraLab 90 titration system. A sample of Hb solution (9 ml) was transferred to the titration chamber, which was thermostatted at 15 °C and sealed with a lid. The Hb solution was magnetically stirred, and humidified pure O₂ was supplied to the chamber through a gas inlet to oxygenate the Hb fully. The pH was measured using a Radiometer GK2401C combined pH electrode. After 45 min of equilibration, a stable pH was recorded, and this pH was considered to be the iso-ionic pH. Microlitre samples of freshly prepared 0.1 mol l⁻¹ NaOH were then added until a pH of approximately 9 had been attained. After an additional 5 min of equilibration, titration with 0.1 mol l⁻¹ HCl was started. The titration was continued until pH 5 was reached. A new 9 ml sample from the same Hb stock solution was then transferred to the titration chamber. This sample was equilibrated with pure N₂ to deoxygenate the Hb and then subjected to the same titration procedure as the oxygenated Hb. The number of titrations on oxygenated and deoxygenated Hb was nine for lamprey Hb and four for hagfish Hb. The monomeric Hb concentration in the Hb solutions varied between 0.5 and 1.0 mmol l⁻¹ for Lampetra fluviatilis and between 0.2 and 0.8 mmol l⁻¹ for Myxine glutinosa. No distinct dependency of hydrogen ion equilibria on Hb concentration was observed within these ranges of Hb concentration.

Representative hydrogen ion titration curves of oxygenated and deoxygenated Hbs from lamprey (Lampetra fluviatilis) and hagfish (Myxine glutinosa) have been drawn to the same scale in Fig. 1. The curves depict how net proton charge (Z_H) changes as a function of pH. Proton charge is given on a monomer basis (i.e. mol H⁺ mol⁻¹ monomer), since both lamprey and hagfish Hbs are monomers when oxygenated, whereas they aggregate to oligomers upon deoxygenation.

Hb/H⁺ equilibria varied considerably between lamprey and hagfish (Fig. 1). The slopes of the titration curves give the intrinsic Hb-specific buffer values at a constant degree of oxygenation (i.e. when pKa values of titratable groups tend to be fixed). The direct titration curves were fitted individually to polynomial equations (Fig. 1), from which the buffer values (−dZ_H/dpH, mol H⁺ mol⁻¹ monomer pH unit⁻¹) were obtained by differentiation (Fig. 2). The buffer values varied with pH, degree of oxygenation and species (Fig. 2). Buffer values were much lower in lamprey than in hagfish for both oxygenated and deoxygenated Hb. Oxygenated lamprey Hb even showed a global minimum at approximately pH 7.8, which is the normal pH in oxygenated lamprey red blood cells (Nikinmaa, 1993). Deoxygenated lamprey Hb showed a local maximum close to this pH, but the absolute buffer value remained low. In hagfish Hb, buffer values at physiological pH (approximately pH 7.2 in oxygenated red blood cells; Tufts et al., 1998) were relatively high and comparable with values measured in mammalian Hb (e.g. pig; see Jensen, 1989).

In both lamprey and hagfish Hb, buffer values increased when pH approached 9 (where titration of basic groups such as the ε-amino group of lysine and the guanidyl group of arginine begins) or when pH values decreased below approximately 6 (where amino acid residues containing carboxyl groups are titrated). In the intermediate ‘neutral’ and physiologically relevant pH range, it is mainly the imidazole group of histidine residues and α-amino groups that are titrated (Tanford, 1962). The pKa values of acid, neutral and basic groups in proteins are often clearly separated, and titration curves therefore contain inflection points that can be used to estimate the number of titratable groups (Tanford, 1962). In differential titration curves (−dpH/dZ_Hversus Z_H), the inflection points become separated into two peaks, and the distance between these peaks gives the number of titratable neutral groups (De Bruin and van Os, 1968). This method was used to estimate the number of titratable neutral groups in deoxygenated Hb from the two agnathan species. In lamprey Hb, the differential titration curve showed two clear peaks, which were separated by approximately 1 Z_H charge unit (Fig. 3). There was, however, slight variation among experiments, and in some instances the distance between the peaks was approximately 1.5 Z_H charge unit. In hagfish Hb, one peak (corresponding to the distinction between neutral and basic groups) was well defined, whereas the other (separating neutral and acid groups) was less apparent (Fig. 3). The data suggested that some 4–5 neutral groups were titrated per monomer in hagfish Hb (Fig. 3).

The fixed-acid Haldane effect (moles of H⁺ taken up per mole of monomer upon deoxygenation at constant pH) is given by the vertical distance between the H⁺ titration curves for oxygenated and deoxygenated Hb (Fig. 1). This oxygenation-linked H⁺ binding was significantly (P<0.0001) greater in lamprey than in hagfish Hb. In the examples shown (Fig. 1), the maximal H⁺ uptake upon deoxygenation was 0.84 mol H⁺ mol⁻¹ monomer for lamprey Hb and 0.38 mol H⁺ mol⁻¹ monomer for hagfish Hb. Mean values were 0.92±0.15 mol H⁺ mol⁻¹ monomer for lamprey and 0.35±0.11 mol H⁺ mol⁻¹ monomer (means ± S.D.; N=9 for lamprey and 4 for hagfish).

The data presented here show that the early evolutionary divergence between lampreys and hagfish is accompanied by large differences in the H⁺ equilibria of their Hbs. Lamprey Hb is characterised by very low buffer values and a large fixed-acid Haldane effect, whereas hagfish Hb has large buffer values and a moderate Haldane effect. The distinct difference between lamprey and hagfish Hbs to some extent resembles the difference in H⁺ equilibria between teleost and elasmobranch Hbs (Jensen, 1989). However, the buffer values of lamprey Hb are even lower than those of teleosts, while the buffer values of hagfish are lower than those of elasmobranchs but as large as those of mammalian Hbs. Furthermore, the fixed-acid Haldane effect in hagfish Hb is larger than that of elasmobranchs.

The large difference in buffer values is a consequence of the different amino acid sequences of the Hbs. The amino acid sequence is known for the major Hb components in both lamprey (Zelenik et al., 1979; Hombrados et al., 1983) and hagfish (Liljeqvist et al., 1982). Even though multiple Hb components are present in lamprey and hagfish blood, the primary structures of the main components provide a good reference for interpreting the average Hb/H⁺ equilibria of the natural Hb mixture (Fig. 1).

Monomeric lamprey Hb contains only two histidine residues (Zelenik et al., 1979; Hombrados et al., 1983) and thus has an even lower histidine content than that of most teleost Hbs, which have some 2–6 histidine residues per chain (see Jensen, 1989). Furthermore, the two histidine residues in lamprey Hb are the proximal histidine and the distal histidine (i.e. the haem-linked histidines), which do not normally exchange protons (Perutz, 1990): the proximal histidine residue does not ionize, and the distal histidine residue typically has a pKa of approximately 5.5 (but see below). The very small number of histidines explains the very low buffer values of lamprey Hb in the physiological pH range. Differential titration curves of deoxygenated lamprey Hb suggest that the number of titratable ‘neutral’ groups is between 1 (Fig. 3) and 1.5 (see Result section). One titratable group could be the solitary N-terminal α-amino group. It has been suggested that the N terminus is formylated in one of the main lamprey Hb components (Zelenik et al., 1979), but Hombrados et al. (1983) have pointed to an unblocked N-terminal proline residue as a possible contributor to the Bohr effect. It has also been suggested that the distal histidine is involved in the Bohr effect. Perutz (1990) proposed that, in the oxygenated monomeric form of lamprey Hb, the distal histidine residue has an internal position and a pKa value of approximately 5.5, whereas in the dimeric form it has an external position and a pKa value of approximately 8. If the distal histidine residue is titratable at physiological pH in deoxygenated lamprey Hb, and some Hb components have titratable α-amino groups, this would account for the 1–1.5 titratable neutral groups.

The main Hb component in hagfish contains four histidine residues. One of these is the proximal histidine residue, whereas the distal histidine residue has been substituted by glutamine (Liljeqvist et al., 1982). This leaves three histidine residues, plus the α-amino group of the N-terminal proline, potentially available for titration in each monomer. Four titratable groups in the main Hb component compare well with the value of 4–5 titratable ‘neutral’ groups in the composite Hb suggested by the differential titration curve (Fig. 3), especially when it is taken into account that some of the minor Hb components contain more than four (perhaps up to eight) histidine residues (Paléus and Liljeqvist, 1972). The larger number of titratable ‘neutral’ groups present in hagfish (4–5 per monomer) than in lamprey (1–1.5 per monomer) Hb explains why hagfish Hb has larger buffer values than lamprey Hb (Fig. 2). For comparison, carp Hb (Jensen, 1989) and anodal eel Hb (Brauner and Weber, 1998) have 2.25 titratable ‘neutral’ groups per monomer and human Hb has 6 (Janssen et al., 1970).

The Haldane effect is equivalent to the Bohr effect. In lamprey Hb, it originates in the change from monomer to oligomer upon deoxygenation rather than in conformational changes within a tetramer as in other vertebrates (Riggs, 1972; Nikinmaa et al., 1995). In the present study, the maximal H⁺ uptake upon deoxygenation occurred at pH 6.85, and it was 0.92±0.15 mol H⁺ mol⁻¹ monomer (mean ± S.D., N=9). This compares well with the value of 1 H⁺ per monomer deduced from oxygen equilibria (i.e. the Bohr factor) in intact red blood cells from Lampetra fluviatilis (Nikinmaa, 1993). The amino acid sequences of the main Hb component in Petromyzon marinus and Lampetra fluviatilis only differ at three positions (Hombrados et al., 1983). The difference involves two substitutions of amino acid residues with uncharged polar side chains (Thr→Ser) and one substitution of amino acid residues with nonpolar side groups (Leu→Met), none of which should influence H⁺ titration behaviour. A great similarity in H⁺ binding properties may thus be expected in these two lamprey species. In Petromyzon marinus, the Bohr factor in Hb solutions suggests that 0.7 H⁺ per monomer is taken up upon deoxygenation (Riggs, 1972), and in red blood cells the value is 0.63 H⁺ per monomer (Ferguson et al., 1992). These values for oxygenation-linked H⁺ binding are slightly lower than in Lampetra fluviatilis, which may be due to different H⁺ binding properties of the minor Hbs in the two lamprey species.

Myxine glutinosa possesses a significant fixed-acid Haldane effect (Fig. 1), but the magnitude of this effect is lower than in lampreys. The H⁺ uptake upon deoxygenation was estimated to be some 0.35 H⁺ per monomer in the present study. This compares with a numerical value for the Bohr factor (ΔlogP₅₀/ΔpH) of 0.31 in Myxine glutinosa Hb in the absence of Cl⁻ (Fago and Weber, 1995). In the presence of Cl⁻, however, the Bohr effect was reduced (Fago and Weber, 1995). The present estimate for the oxygenation-linked H⁺ uptake thus appears to be slightly larger than that deduced from measurements of the pH-dependency of oxygen affinity.

The H⁺ equilibria in lamprey Hb are ideally suited to providing a high CO₂-transporting capacity inside the red blood cells. The large oxygenation-linked exchange of H⁺ between Hb and the red blood cell cytosol and the low buffer values of the oxygenated and deoxygenated Hb will produce large changes in red blood cell pHi when the degree of blood oxygenation changes. Intracellular pH in lamprey red blood cells increases by 0.3–0.4 units upon deoxygenation (Ferguson et al., 1992; Nikinmaa and Mattsoff, 1992; Nikinmaa, 1993), which is much larger than pH fluctuations seen in other vertebrates, except in teleosts, which also have large Haldane effects and low buffer values (Jensen, 1986). In lampreys, Na⁺/H⁺ exchange across the red blood cell membrane keeps the intraerythrocytic pH high, and the further increase in pHi upon deoxygenation ensures a high bicarbonate concentration inside the red blood cells in venous blood (e.g. Nikinmaa and Mattsoff, 1992). Thus, CO₂ can be effectively transported in spite of the absence of HCO₃⁻/Cl⁻ exchange across the red blood cell membrane.

The Hb/H⁺ equilibria of hagfish appear to be less effective in ensuring blood CO₂ transport in the face of a limited membrane HCO ⁻ permeability. H⁺ produced when the CO₂ concentration in the blood increases will be effectively removed because the Hb buffer values are high. The intracellular pH can, however, be expected to decrease, because the amount of CO₂ added (0.7–1 mol of CO₂ per mol of O₂ consumed, as dictated by the respiratory quotient) surpasses the magnitude of the fixed-acid Haldane effect. Furthermore, intracellular pH is lower in hagfish than in lamprey red blood cells, so that the red blood cell bicarbonate concentration in hagfish is lower than that in the lamprey. The red blood cell [HCO₃⁻] accordingly changes less with level of oxygenation in hagfish (Tufts et al., 1998) than in lamprey (Nikinmaa and Mattsoff, 1992). Against this background, one may expect that addition of CO₂ in tissue capillaries could produce a relatively high venous and low pH in hagfish.

There is, however, an alternative possibility to be considered. It has recently been suggested that an oxygenation-linked binding of bicarbonate might occur in hagfish Hb (Fago et al., 1999). This functional property has so far been reported only for crocodile Hbs (Bauer et al., 1981), where it gives rise to a unique strategy for blood CO₂ transport in which most of the CO₂ taken up into the blood in the tissue capillaries is carried within the red blood cells as Hb-bound HCO₃⁻ until CO₂ is excreted in the lungs (Jensen et al., 1998). Binding of HCO₃⁻ to Hb provides a mechanism for removing free HCO₃⁻ from the red blood cell cytosol that is alternative to shifting HCO₃⁻ to the plasma. Given the virtual absence of HCO₃⁻/Cl⁻ exchange and that Hb/H⁺ equilibria and red cell pH do not favour high concentrations of free HCO₃⁻ inside the red blood cells, an oxygenation-linked binding of HCO₃⁻ to the Hb would indeed be useful in hagfish. The potential role of oxygenation-linked bicarbonate binding and its quantitative importance for blood CO₂ transport must, however, await future studies.

The study was supported by the Danish Natural Science Research Council (Centre for Respiratory Adaptation). Ms Annie Bach is thanked for skilled technical assistance with the experiments.