Haemoglobin function in intact lamprey erythrocytes: interactions with membrane function in the regulation of gas transport and acid–base balance

Haemoglobin function in lampreys has mainly been studied using (dilute) solutions (e.g. Wald and Riggs, 1951; Briehl, 1963; Antonini et al. 1964; Andersen and Gibson, 1971; Dohi et al. 1973; see also reviews by Riggs, 1972; Perutz, 1990). On the basis of these studies, it is clear that lamprey haemoglobins in solution show cooperativity as well as distinct Bohr and Haldane effects. Furthermore, the oxygen affinity of lamprey haemoglobin depends on its concentration in the solution. This finding, and the observation that the apparent molecular mass of deoxygenated haemoglobin is higher than that of oxygenated haemoglobin (Briehl, 1963; Andersen, 1971; Dohi et al. 1973), indicates that, in lampreys, the cooperativity of oxygen binding and the effect of protons on the haemoglobin oxygen-affinity are caused by reversible association/dissociation reactions of monomers/oligomers (for reviews, see Riggs, 1972; Perutz, 1990). This contrasts with the situation in all other vertebrate haemoglobins, in which these effects are the result of conformational changes of the tetrameric molecule.

The behaviour of lamprey haemoglobin in solution can be explained by the following model (Perutz, 1990). The dimeric/oligomeric form of the molecule has a low oxygen affinity owing to constraints resulting from interactions between its subunits. In the dimer, binding of oxygen weakens these interactions, the molecule dissociates to its subunits and, consequently, the oxygen affinity of the other subunit is increased. It is likely that the distal histidine (E7) is involved in this allosteric transition (Perutz, 1990): in the monomeric form, it would be in an internal position within the haem pocket (where it could form a hydrogen bond with oxygen), and in the dimeric form it would be in an external position (where it could form a hydrogen bond with an aspartate or glutamate of the neighbouring molecule). The Bohr effect would, in this case, be accounted for by the differences in the pK values of the histidine in its internal position (5.5–6) and in its external position hydrogen-bonded to a carboxylate (approximately 8). Until recently, it has been difficult to interpret the oxygen binding data of whole blood on the basis of the behaviour of haemoglobin in solution. For example, the oxygen affinity of lamprey blood was much higher than expected on the basis of the behaviour of haemoglobin in solution (Riggs, 1972). Also, the numerical value of the Bohr factor estimated for whole blood was much smaller than expected on the basis of data from haemoglobin solutions (Bird et al. 1976; Nikinmaa and Weber, 1984).

The apparent disagreement between oxygen binding data for haemoglobin in solution and for haemoglobin within intact erythrocytes was resolved when it was observed that the properties of the lamprey erythrocyte membrane were markedly different from those of the erythrocyte membrane of other vertebrates. The permeability of lamprey erythrocytes to chloride and to acid or base equivalents is very low (Ohnishi and Asai, 1985; Nikinmaa and Railo, 1987; Tufts and Boutilier, 1989, 1990). As a consequence of the minute passive fluxes of acid equivalents across the erythrocyte membrane, intracellular pH is maintained at a high value mainly by sodium/proton exchange across the red cell membrane (Nikinmaa, 1986; Nikinmaa et al. 1986, 1993; Tufts and Boutilier, 1990; Tufts, 1992; Virkki and Nikinmaa, 1994).

In this review, we examine how these unique membrane properties affect haemoglobin function within intact lamprey erythrocytes and how the properties of haemoglobin and of the cell membrane affect not only oxygen transport but also carbon dioxide transport and acid–base regulation.

The resting pH of lamprey Lampetra fluviatilis blood at 10 °C is 7.7–7.9 (Mattsoff and Nikinmaa, 1988). Since plasma pH decreases by 0.011 units with each 1 °C increase in temperature (Nikinmaa and Weber, 1984), at 18 °C the values would be approximately 7.6–7.8. At an extracellular pH of 7.67, well within the range of resting plasma pH values at 18 °C, the P₅₀ value for lamprey erythrocytes is 2.9 kPa (Nikinmaa, 1993a). The resting plasma pH value is similar to, and the P₅₀ value similar to or higher than, the values measured in rainbow trout (plasma pH at 19 °C is 7.67 and the P₅₀ value is 3.4 kPa; Nikinmaa, 1981; plasma pH at 18 °C is 7.65 and the in vivo P₅₀ value is 4.9 kPa in normoxia and 3.6 kPa in hypoxia; Nikinmaa and Soivio, 1982). In carp, Cyprinus carpio, however, the P₅₀ value at an extracellular pH of 7.7 is clearly smaller, i.e. 1.1 kPa (a value calculated from the data of Salama and Nikinmaa, 1988) than in lamprey. Thus, lamprey erythrocytes show oxygen affinities of the range expected for an active animal.

When erythrocyte pH is related to the oxygen affinity of haemoglobin within intact erythrocytes (Fig. 1), it becomes obvious that, to have a blood oxygen-affinity that is similar to or higher than that of rainbow trout at rest, Lampetra fluviatilis must have a much higher intraerythrocytic pH (pHi). Indeed, at 10 °C, the intraerythrocytic pH of Lampetra fluviatilis, at a resting extracellular pH of 7.78, was 7.96 (Mattsoff and Nikinmaa, 1988). At a similar temperature, and at an extracellular pH of 7.8, the intraerythrocytic pH of rainbow trout was 7.45 (Perry et al. 1987).

When the lampreys Lampetra fluviatilis and Petromyzon marinus are compared, it is apparent that the intrinsic haemoglobin oxygen-affinity of Petromyzon marinus is higher than that of Lampetra fluviatilis, since the P₅₀ value of Petromyzon marinus haemoglobin in dilute solutions is lower than that of Lampetra fluviatilis haemoglobin at similar pH values (Antonini et al. 1964). This difference is also seen in the whole blood; the P₅₀ value of Petromyzon marinus haemoglobin at a pHi of 7.64 at 10 °C was 3.1 kPa (Ferguson et al. 1992), i.e. similar to that of Lampetra fluviatilis haemoglobin at a pHi of 7.67 and at 18 °C. Taking into account the exothermic reaction of oxygen binding and the dependence of intracellular pH on temperature (Nikinmaa and Weber, 1984), this result indicates clearly the higher oxygen affinity of Petromyzon marinus than of Lampetra fluviatilis haemoglobin in intact erythrocytes at a given pH. Since the intraerythrocytic pH of Petromyzon marinus is approximately 0.2 units lower than that of Lampetra fluviatilis (see Tufts, 1991; Ferguson et al. 1992; Mattsoff and Nikinmaa, 1988), it is likely that the resting in vivo whole-blood oxygen-affinities of the two species are similar.

Both lamprey species maintain their intraerythrocytic pH using a sodium-dependent acid extrusion mechanism, most likely sodium/proton exchange (Nikinmaa, 1986; Nikinmaa et al. 1986; Tufts and Boutilier, 1990; Tufts, 1992). As stated in the Introduction, we have suggested that the maintenance of intracellular pH by sodium/proton exchange is possible because the permeability of the lamprey erythrocyte membrane to acid equivalents is low. This suggestion can be experimentally verified using, for example, tributyltin chloride, a hydroxyl ion/chloride exchanger. Tufts and Boutilier (1990) showed that treatment of lamprey erythrocytes with tripropyltin chloride caused a reduction in the steady-state intraerythrocytic pH of Petromyzon marinus. The effect of organotin compounds is immediate (Fig. 2; L. V. Virkki and M. Nikinmaa, unpublished data) showing that the effect is indeed caused by the ion-exchange action of the compound and not by its metabolic effects. On the basis of the data in Figs 1 and 2, a reduction of intracellular pH to the values observed in tributyltin-chloride-treated cells would increase the P₅₀ value of Lampetra fluviatilis haemoglobin to about 10 kPa at 18 °C. Such a value would seriously impair oxygen loading at the gills, even in normoxic conditions.

In conclusion, the haemoglobin oxygen-affinity within intact erythrocytes of lampreys is similar to that in other active, aerobic fish. The maintenance of this oxygen affinity requires that the intraerythrocytic pH be maintained at a higher level than in teleosts. The maintenance of a high intraerythrocytic pH is possible because the passive permeability of red cell membrane to acid equivalents is low (Nikinmaa and Railo, 1987; Tufts and Boutilier, 1989, 1990) and because the cells possess a sodium/proton exchanger which influences the red cell pH even in resting conditions (Nikinmaa, 1986; Tufts, 1992); the exchanger is even further activated by acidification (Virkki and Nikinmaa, 1994).

Bird et al. (1976) found that the Hill plots of oxygen binding by Lampetra flluviatilis blood were curvilinear. Whereas at low oxygen saturations, the calculated n values were close to 1, they increased with increasing oxygen saturation and the highest values obtained were above 2. Subsequent studies using intact erythrocytes over an intracellular pH range of 8–6.7 (Nikinmaa, 1993a) showed that the cooperativity was also markedly affected by pH. For example, at a pHi of 6.7, the Hill plot was linear and the n value was close to 1. The higher the intracellular pH, the more curvilinear the Hill plot became (Fig. 3). These findings are compatible with the suggestion that deoxygenated lamprey haemoglobin is in the low-affinity oligomeric form, which is further stabilized by low pH values, such that at the lowest pH values examined no dissociation occurs. As the pH increases, the interactions between the individual chains of the oligomers are weakened, and oxygenation causes dissociation of the low-affinity oligomers to the high-affinity monomers. Consequently, an oxygenation-dependent increase in Hill’s n value is observed. The higher the pH, the lower the oxygen tension required for the dissociation of oligomers and for the consequent increase in Hill’s n value.

Early studies on the effects of pH on the oxygen binding properties of lamprey blood suggested that the Bohr factor would be small, with values between -0.1 and -0.3 (Bird et al. 1976; Nikinmaa and Weber, 1984). However, in those studies, the haemoglobin oxygen-affinity was given as a function of extracellular pH, since at that time the unique regulation of intracellular pH in lamprey erythrocytes was not understood. Later studies (Ferguson et al. 1992; Nikinmaa, 1993a), in which the change in oxygen affinity was related to intracellular pH, showed that the Bohr factor was quite large. In Petromyzon marinus at 10 °C and at a pHi of 7.3–7.6, the Bohr factor (Δlog P₅₀/ΔpHi) was -0.63 (Ferguson et al. 1992), and in Lampetra fluviatilis at 18 °C and at a pHi of 6.7–8.0, the Bohr factor was -1.03 (Nikinmaa, 1993a; Fig. 1). For Lampetra fluviatilis, the value of -1.03 therefore indicates the release of one proton per monomeric haemoglobin. A similar value has been obtained on the basis of the nonbicarbonate buffering capacity of the erythrocyte, the intraerythrocytic haemoglobin concentration and the measured intracellular pH changes upon oxygenation (Nikinmaa, 1993a). Notably, conclusions based on studies of haemoglobin solutions indicate that the Bohr effect of lamprey haemoglobin is due to the release of one proton per haemoglobin chain upon oxygenation (Andersen, 1971; Riggs, 1972; Perutz, 1990).

As is apparent from the concept of linked functions (e.g. Wyman, 1964), a large Bohr effect (the effect of protons on oxygen binding and dissociation) implies that oxygenation has a large effect on proton uptake and release, i.e. the Haldane effect is large. Consequently, deoxygenation increases the red cell pH of Lampetra fluviatilis by up to 0.33 units and the red cell pH of Petromyzon marinus by up to 0.45 units (Nikinmaa and Mattsoff, 1992; Ferguson et al. 1992).

Lapennas (1983) has discussed the significance of the Bohr (and Haldane) factor in mammals. In his analysis, he used the ‘physiological’ oxygen equilibrium curve, i.e. the relationship between oxygen saturation and oxygen tension that pertains in vivo, where blood pH is not constant but changes as oxygen is exchanged for carbon dioxide. The analysis took into account the changes in blood pH caused by the liberation of carbon dioxide into the blood and the uptake of protons upon deoxygenation of haemoglobin (Haldane effect), and the effect of protons on haemoglobin oxygen-affinity (Bohr effect). Lapennas (1983) concluded that Bohr factors between -0.35 and -0.5, i.e. half of the respiratory quotient, give maximal benefit in terms of oxygen delivery, since they produce the greatest possible rightward shift of the physiological oxygen equilibrium curve in working tissues (i.e. the net effect of the decrease in the pH of capillary blood and the pH-induced decrease in haemoglobin oxygen-affinity is maximal). However, large Bohr factors (i.e. those approaching the respiratory quotient) minimize arterio-venous pH changes and facilitate carbon dioxide transport. When applying this analysis to lamprey blood, and taking into account that intraerythrocytic pH must be used, Ferguson et al. (1992) concluded that the large Bohr/Haldane effect is probably oriented towards carbon dioxide transport. Indeed, the Haldane effect within intact lamprey erythrocytes is so large that the intraerythrocytic pH in venous blood is higher than that in arterial blood (Tufts et al. 1992).

The role played by the large Haldane effect of haemoglobin in carbon dioxide transport by lamprey blood becomes obvious when the mechanism of carbon dioxide transport by lamprey blood is examined. Tufts and Boutilier (1989, 1990) and Nikinmaa (1990) first presented the model for carbon dioxide transport by lamprey blood (Fig. 4). Owing to the negligible permeability of the lamprey red blood cell membrane to bicarbonate and other acid–base equivalents (Nikinmaa and Railo, 1987; Tufts and Boutilier, 1989; Nikinmaa and Mattsoff, 1992; Nikinmaa, 1993b; Cameron and Tufts, 1994), the carbonic anhydrase contained within the erythrocyte (Nikinmaa et al. 1986; Henry et al. 1993) is inaccessible to plasma bicarbonate. As a consequence, only a small proportion of plasma bicarbonate can be converted to carbon dioxide in the uncatalysed dehydration reaction during the passage of blood through the gills, because the half-time of the uncatalysed reaction is approximately 1 min at 15 °C (Heming, 1984) and the residence time of blood in the gills is at the most a few seconds (Bhargava et al. 1992). Thus, most of the carbon dioxide excreted from the blood in the gills must come from the erythrocytic bicarbonate stores, and the efficiency of carbon dioxide transport by lamprey blood depends critically on how much of the total carbon dioxide the erythrocytes can load in the tissues and unload in the gills. The major factors influencing the loading and unloading of carbon dioxide within the erythrocytes are (1) the changes in carbon dioxide tension in the circulation and (2) the carbon dioxide dissociation curves of oxygenated and deoxygenated erythrocytes. On the basis of the carbon dioxide dissociation curves for deoxygenated and oxygenated Lampetra fluviatilis erythrocytes (Nikinmaa and Mattsoff, 1992; Fig. 5), more than 80 % of the maximal decrease in the total carbon dioxide content of the erythrocytes (in vitro conditions resembling the situation in gills) can be attributed to the oxygenation-dependent decrease in the total carbon dioxide content (at a constant carbon dioxide tension) and less than 20 % to the decrease in carbon dioxide tension. Ferguson et al. (1992) also observed a marked difference in the total carbon dioxide content of oxygenated and deoxygenated erythrocytes of Petromyzon marinus. Furthermore, they showed that oxygenation did not affect the plasma carbon dioxide content. The oxygenation-dependent decrease in the intraerythrocytic carbon dioxide content is fully accounted for by the oxygenation-dependent decrease in intraerythrocytic pH caused by the Haldane effect of haemoglobin (Nikinmaa and Mattsoff, 1992). Thus, on the basis of in vitro data, the Bohr/Haldane effect appears to play a critical role in carbon dioxide transport by lamprey blood.

Tufts et al. (1992) investigated the role of erythrocytes in carbon dioxide transport in vivo. According to their results, 62 % of the excreted carbon dioxide originates from the erythrocytes in resting Petromyzon marinus. After exercise, the contribution of erythrocytes is increased to 78 % of the total carbon dioxide content difference between venous and arterial blood. It is possible that these values underestimate the role of erythrocytes in carbon dioxide excretion because of the inherent time delays in the measurements of total carbon dioxide content (of true plasma in postbranchial blood). Nevertheless, the results indicate the major role played by erythrocytes in carbon dioxide excretion in lampreys in vivo. In comparison, only 15 % of the excreted carbon dioxide originates from the erythrocytes in rainbow trout (Heming, 1984).

In conclusion, the large Bohr effect of lamprey haemoglobin within intact erythrocytes plays an important role in facilitating carbon dioxide transport. As a consequence of the high venous intraerythrocytic pH (0.15–0.2 units higher than arterial intraerythrocytic pH) in Petromyzon marinus (Tufts et al. 1992), carbon dioxide is efficiently loaded into erythrocytic stores, mainly as bicarbonate. As a consequence of the decrease in carbon dioxide tension of blood at the gills, and especially the marked reduction in intraerythrocytic pH due to the release of protons upon oxygenation of haemoglobin, carbon dioxide is efficiently unloaded in the gills. Thus, the measured decrease in total carbon dioxide content during the passage of blood through the gills in the lamprey (Tufts et al. 1992) is similar to that observed for active teleosts such as rainbow trout (e.g. Nikinmaa and Jensen, 1986), despite the fact that, in lampreys, plasma bicarbonate plays only a limited role in carbon dioxide excretion owing to the slow permeation of bicarbonate through the membrane.

In dilute solutions, the oxygen affinity of lamprey haemoglobin is markedly concentration-dependent (Wald and Riggs, 1951; Briehl, 1963; Antonini et al. 1964; Andersen and Gibson, 1971; Dohi et al. 1973; Nikinmaa and Weber, 1993). An increase in the concentration decreases haemoglobin oxygen-affinity by increasing the proportion of low-affinity oligomers. Since haemoglobin within intact erythrocytes shows both cooperativity and a large Bohr effect, it is probable that association/dissociation reactions of haemoglobin also occur within the erythrocyte. If this were the case, then the haemoglobin oxygen-affinity should respond to variations in red cell volume (which affect intracellular haemoglobin concentration). This possibility could be tested experimentally by shrinking the cells osmotically and then measuring oxygen equilibrium curves, because the volume of osmotically shrunk Lampetra fluviatilis erythrocytes does not change after the initial decrease (Virkki and Nikinmaa, 1994). We have recently carried out these experiments (Airaksinen and Nikinmaa, 1995). As Table 1 indicates, there is a clear decrease in the haemoglobin oxygen-affinity (an increase in P₅₀) within intact erythrocytes when the mean cellular haemoglobin concentration increases. Although this decrease occurs, there is a tendency for the shrunk erythrocytes to have a higher intracellular pH than that observed in erythrocytes of normal volume (Virkki and Nikinmaa, 1994).

Thus, concentration changes occurring within intact erythrocytes appear to be adequate to cause changes in the aggregation state of the haemoglobin molecule and, as a consequence, in the haemoglobin oxygen-affinity. It is not yet clear whether such volume changes are part of the normal physiological regulation of haemoglobin function. However, data from hypoxia-acclimated Lampetra fluviatilis indicate that the mean cellular haemoglobin concentration is lower after a week’s acclimation to hypoxia than during normoxia (Nikinmaa and Weber, 1984).

The general vertebrate pattern of buffering extracellular acid loads involves the Jacobs–Stewart cycle: protons entering the plasma react with bicarbonate ions to form carbon dioxide (and water). The carbon dioxide formed diffuses into the erythrocytes and is hydrated to bicarbonate and protons. The protons are taken up by intracellular buffers, notably haemoglobin, and bicarbonate is able to leave the cell via rapid anion exchange, thus replenishing the extracellular buffer store. Although the erythrocytic buffering capacity is substantial in lampreys (35–70 μmol pH unit^–1 g^–1; Tufts and Boutilier, 1989, 1990; Nikinmaa and Mattsoff, 1992), this scheme is not possible in lampreys because of the slow movement of bicarbonate across the red cell membrane. Consequently, extracellular acid loads can only be buffered by plasma buffers, and the nonbicarbonate buffering capacity of whole blood (or true plasma, i.e. plasma that is in contact with erythrocytes during the determination of the buffering capacity) is very low, i.e. 3 μmol pH unit^–1 g^–1 in Petromyzon marinus (Tufts and Boutilier, 1990) and 2.5 μmol pH unit^–1 g^–1 in Lampetra fluviatilis (Mattsoff and Nikinmaa, 1988). When the erythrocyte membrane is made permeable to acid equivalents using tripropyltin chloride, the nonbicarbonate buffering capacity of true plasma is significantly increased in Petromyzon marinus (Tufts and Boutilier, 1990), confirming that intraerythrocytic buffers do not take part in buffering extracellular acid loads in normal conditions.

Extracellular acid loads can thus cause marked fluctuations in the extracellular pH of lampreys. For example, exposure of Lampetra fluviatilis to pH 4 for 24 h caused a 0.9 unit decrease in the plasma pH (Mattsoff and Nikinmaa, 1988); a similar exposure in rainbow trout caused only a 0.1–0.3 unit decrease (Ultsch et al. 1981; Booth et al. 1982).

Haemoglobin function within lamprey erythrocytes offers a unique solution to gas transport among vertebrates. A schematic representation of factors controlling haemoglobin function is given in Fig. 6. Both the cooperativity of oxygen binding and the effect of protons on haemoglobin oxygen-affinity are due to dissociation/association reactions of the haemoglobin molecules. In intact erythrocytes, dissociation/association reactions can occur, since the intracellular pH can be maintained at a relatively high value by sodium/proton exchange. This is possible because the permeability of the red cell membrane to acid and base equivalents is very low.

Owing to the low bicarbonate permeability of the red cell membrane, only a minor proportion of plasma bicarbonate can be dehydrated to carbon dioxide during the residence time of blood in the gills. This potential limitation on carbon dioxide excretion is overcome, however, by the high intracellular pH and the marked oxygenation-linked pH changes, which greatly increase the effectiveness of carbon dioxide excretion by the erythrocytes.

Although both oxygen and carbon dioxide transport are as effective in lamprey blood as in that of active teleosts, there is one significant disadvantage caused by the relative impermeability of the lamprey erythrocyte membrane to acid equivalents: the buffering of extracellular acid loads becomes extremely ineffective. As a consequence, extracellular acid loads cause marked fluctuations in lamprey plasma pH. Thus, the major advantage gained by the evolution of rapid anion exchange appears not to be related to gas transport but to extracellular buffering.

The authors’ research has been supported by the Academy of Finland. We thank Dr Bruce L. Tufts for valuable comments on the manuscript.