In vitro Interactions Between Oxygen and Carbon Dioxide Transport in the Blood of the Sea Lamprey (Petromyzon Marinus)

There is growing evidence indicating that the strategy for gas transport in the blood of lampreys may be very different from that in most other vertebrates (Nikinmaa and Tufts, 1989; Tufts and Boutilier, 1989, 1990; Nikinmaa and Mattsoff, 1992; Tufts et al. 1992). Lampreys are extant members of a phylogenetically primitive group of vertebrates, the agnathans. Thus, a thorough understanding of the unique strategy for blood gas transport in these primitive animals may provide valuable insights into the evolution of respiratory systems in vertebrates.

In most vertebrates, the majority of CO₂ added to the blood by the tissues is transported in the plasma as HCO₃^- to the respiratory organ (Roughton, 1964; Boutilier et al. 1979; Boutilier and Toews, 1981; Perry, 1986; Heming, 1984; Klocke, 1987). In the blood of lampreys, however, CO₂ transport is largely dependent on erythrocyte CO₂ carriage (Tufts and Boutilier, 1989, 1990; Nikinmaa and Mattsoff, 1992; Tufts et al. 1992). Indeed, Tufts et al. (1992) have recently demonstrated that changes in erythrocyte CO₂ concentrations account for as much as 78 % of the total CO₂ difference between arterial and venous blood in the sea lamprey after exercise in vivo. Several factors may contribute to the unique distribution of CO₂ transport observed in the blood of lampreys. Chloride/bicarbonate exchange across the red blood cell membrane is an integral component of the CO₂ transport process in the blood of most vertebrates, but appears to be functionally limited or absent in the blood of agnathans (Nikinmaa and Railo, 1977; Ellory et al. 1977). Sodium-dependent movements of acid-base equivalents may also affect the distribution of CO₂ across the red blood cell membrane in the sea lamprey (Tufts, 1992). Finally, Nikinmaa and Mattsoff (1992) have recently demonstrated that the blood of the river lamprey shows a relatively large Haldane effect. At present, however, there is a paucity of information on the relative importance of these factors under conditions resembling those occurring in arterial and venous blood in vivo.

The unique pattern of carbon dioxide transport may also have critical implications for our understanding of oxygen transport in lampreys. As suggested by Heisler (1986), chloride/bicarbonate exchange across the red blood cell membrane will have an important effect on the red blood cell pH at any given (Tufts and Boutilier, 1990). Red blood cell pH in lampreys is often considerably higher than in other vertebrates and is well maintained during relatively large fluctuations in extracellular pH either in vitro or in vivo (Nikinmaa and Weber, 1985; Mattsoff and Nikinmaa, 1988; Tufts and Boutilier, 1989; Tufts, 1991; Tufts et al. 1992). Previous investigations of the oxygen transport properties of lamprey blood, however, have not considered the importance of erythrocyte pH. Thus, as suggested by Nikinmaa and Mattsoff (1992), the magnitude of the Bohr coefficient in lamprey whole blood has probably been largely underestimated. Moreover, the functional significance of the Bohr effect in lampreys may not be fully appreciated (Lapennas, 1983).

The purpose of the present study is, therefore, to elucidate further the strategy for gas transport in the blood of the sea lamprey Petromyzon marinus. In these experiments, particular emphasis was placed on the interactions between oxygen and carbon dioxide transport in an attempt to account for the unique relationships recently documented in the arterial and venous blood of these primitive vertebrates in vivo (Tufts et al. 1992).

Adult sea lampreys {Petromyzon marinus) weighing approximately 300 g each were collected from the Shelter Valley River, Ontario, during the annual spawning migration (May-June). The animals were transported to the Biology Department at Queen’s University where they were held in circular (1.5mxl.5m) tanks containing aerated dechlorinated Kingston tapwater (pH 7.4, 8–10 °C). At least 2 weeks were allowed for acclimation prior to experimentation. The ionic compositon of the water, in mequiv I^-1, was as follows: Na⁺=2.1, K⁺=0.05, Ca²⁺=2.2, Cl^-=1.3, HCO₃^-=1.5.

The dorsal aorta of anaesthetized lampreys was cannulated with PE50 tubing as described earlier (Tufts, 1991). The animals were permitted to recover in darkened Perspex boxes with flowing aerated fresh water for at least 24 h prior to blood collection. Following the recovery period, blood was drawn from each animal using 1 ml syringes and placed in ice-chilled heparinized glass vessels.

6 ml of blood was placed into each of two heparinized glass tonometry vessels and equilibrated in intermittently rotating tonometer vessels against air containing 0.2, 1.0, 3.0 or 5.0% CO₂ at 10°C (Wöstoff gas-mixing pump). Following equilibration, 1 ml of blood was removed in an air-tight Hamilton syringe for the analysis of total oxygen content ⁠, haematocrit, haemoglobin, plasma pH (pHe) and red cell pH (pHi). At this time, the partial pressure of oxygen of the equilibration gas was decreased and that of nitrogen increased. Equilibration and sampling was performed at six different levels such that oxygen saturation curves were constructed at the following carbon dioxide partial pressures (kPa): 0.20, 0.99, 2.84 and 4.65. Hill analysis of those data values lying between 25 and 90% oxygenation was performed for each oxygen dissociation curve.

In a second series of experiments, pooled blood was divided into two 6 ml portions which were then transferred to heparinized tonometer vessels. The blood was equilibrated for 1 h at 10 °C against 0.2 % CO₂ in either air or nitrogen. Following equilibration, two 1 ml blood samples were removed from each vessel using 1 ml air-tight Hamilton syringes. The first 1 ml blood sample was analyzed for total CO₂ content of whole blood and true plasma ⁠, pHe, pHi, whole-blood pH (pH_wb) and haematocrit (Hct). The second 1 ml blood sample was dispensed into two 0.5 ml Eppendorf tubes and centrifuged. Following centrifugation, the plasma from each of these tubes was discarded and the red cell pellets were saved for analysis of red blood cell water content and sodium concentration respectively. A similar sampling procedure was performed after the blood had been equilibrated at 1.0 (0.96 kPa) and 3.0 % (2.9 kPa) CO₂.

In the final series of experiments, blood was collected from resting lampreys and divided into two 4 ml portions which were then transferred to intermittently rotating tonometers. In each experiment, ouabain (final concentration in blood 10⁻⁴moll^-1) was added to the blood in both tonometer vessels. The blood was then equilibrated at 10 °C against a gas mix of 0.2 % CO₂ balance air. Following this initial equilibration period, a 1 ml sample was taken from each tonometer vessel for the determination of red cell water and sodium ion content. In one of the tonometers, the equilibration gas mixture was then switched to a mix of 0.2 % CO₂ balance nitrogen. The other tonometer served as a control and remained at 0.2 % CO₂ in air. The sampling procedure was then repeated for both vessels after an additional 10, 30 and 60 min of equilibration.

A PHM 73 pH meter and associated micro-pH unit (Radiometer, Copenhagen) thermostatted at 10 °C was used to measure pH. The pHi was determined on freeze-thawed haemolysates according to the method of Zeidler and Kim (1977).

Haematocrit (Hct) was measured in triplicate using a Clay-Adams microhaematocrit centrifuge. The method of Drabkin and Austin (1935) was used to determine haemoglobin concentration using the procedure and reagents supplied by Sigma Chemical Co. (St Louis).

The total oxygen content of whole blood was obtained by the method of Tucker (1967) using an apparatus similar to that described by Hughes et al. (1982).

Red blood cell water content and sodium concentration were determined by first dispensing two 300 μl portions of whole blood into two dried and tared Eppendorf tubes. The water content of the red cells was then obtained by centrifuging (Eppendorf microfuge) the whole blood for 4 min, removing the plasma supernatant, and drying the red cell pellet to a constant weight (2–3 days at 80 °C). For red blood cell sodium analysis, the other red blood cell pellet was agitated to a homogeneous slurry after addition of 300 μl of 8.0 % perchloric acid and then allowed to digest for 12 h. Sodium composition was determined on the supernatant of this acid-extracted slurry using a Corning 410 flame photometer (Ciba Corning Canada). Plasma sodium and water contents were used to correct the red blood cell values for trapped plasma (2.5 %, Houston, 1985). Only such corrected values are reported.

Unless stated otherwise, all data are presented as means ± 1 standard error of the mean (S.E.M.). Statistical analysis by two-tailed paired or unpaired t-tests, where appropriate, was used to determine the significance of differences. In the comparison of values obtained through repetitive sampling, the significance of differences was obtained by first employing an analysis of variance followed by Dunnett’s two-tailed i-test for repeated measures. Lines derived from linear regressions were compared for significant differences in slope by an analysis of covariance. A fiduciary limit of P<0.05 was employed throughout.

Whole blood used in the determination of oxygen transport characteristics had a haematocrit of 16.9±0.7% and a maximal O₂-carrying capacity of 1.01±0.05mol O₂mol^-1 Hb at 100% oxygenation, N=28. Haematocrit (Hct) and haemoglobin content (Hb) of the whole blood were related by the following linear regression equation: Hb (gdl^-l)=0.26×Hct(%) + 0.3l9, P=0.0001, r=0.96, N=28. Sea lamprey Hb exhibited sigmoidal (NH= 1.52–1.89) O₂ dissociation relationships (Table 1). There was no significant difference between the NH at 0.20 kPa and the HH values obtained at the higher tensions (Table 1). An increase in the ambient did result in a rightward shift in the O₂ dissociation curves. Hill analysis of the data indicated a progressive increase (reaching 1,7-fold) in the P₅₀ with increasing CO₂ tension. An additional set of experiments confirmed that full saturation of haemoglobin was reached over the entire range of CO₂ tensions used in these experiments.

The Bohr effect, indicated by the data in Table 1, was determined for both plasma pH (pHe) and red blood cell pH (pHi; Fig. 1). It is evident that haemoglobin oxygen-carriage was more strongly dependent on pHi than on pHe. Thus, this inequality resulted in a significantly lower Bohr coefficient calculated by linear regression analysis of pHe data (ΔlogP₅₀/ΔpH=−0.173±0.029) compared with that obtained when the pHi data are considered (ΔlogP₅₀/ΔpH=−0.631±0.081 ; Fig. 1).

Fig. 2 illustrates that haemoglobin oxygen-saturation has a considerable effect on the pHi of sea lamprey blood, but not on the pHe. Red blood cell pH is inversely related to and significant increases in pHi occur as falls below 80% at all experimental levels. The magnitude of the increase in pHi observed with decreasing % (from 100 to approximately 10–20%) appears to be smaller at 4.65 kPa than at 0.20 kPa. This may be due, however, to the nature of the logarithmic pH scale, since the calculated changes in actual proton concentration are approximately similar.

The correlation between pHi, pHe and blood oxygenation was further examined using air-equilibrated and nitrogen-equilibrated blood (Hct=17.6±l.8) over a range of CO₂ tensions to determine the magnitude of the increase in pHi associated with the Haldane effect (Fig. 3). In these experiments, significantly greater pHi values were obtained in deoxygenated blood over the entire pHe range. The increase in pHi associated with the Haldane effect ranged from 0.453 pH units at 0.2 kPa to 0.169 pH units at 2.9 kPa ⁠.

Deoxygenated blood also had a significantly greater CO₂-carrying capacity than oxygenated blood in the sea lamprey (Fig. 4A). At a of 2.9 kPa, the of deoxygenated blood was 2.95±0.43 mmol 1’¹ higher than that of oxygenated blood. The molar concentration of Hb for the blood used in these experiments can be calculated on the basis of a monomeric relative molecular mass for lamprey Hb of 18 200 (Lenhert et al. 1956) and the previously described relationship between Hct and Hb content of lamprey whole blood. At a of 2.9 kPa, the Haldane coefficient, which can then be calculated from the present data, is 1.1 mmol CO₂ mmol^-1 monomer Hb or 4.4mmol CO₂mmol^-1 tetrameric Hb. This Haldane effect can be attributed to a large increase in the of the sea lamprey red blood cells (Fig. 4B). Indeed, at 2.9 kPa, oxygenated red blood cells contain 21.85±1.96 mmol CO₂l^-1 cell water, or only 55% of the of deoxygenated red blood cells (39.9111.09mmolCO₂I^-1) at the same ⁠. There were no significant differences between the of oxygenated and deoxygenated plasma (Fig. 4C).

Non-bicarbonate buffer relationships for oxygenated and deoxygenated lamprey blood are presented in Fig. 5. Whole-blood bicarbonate concentration is significantly greater in deoxygenated blood over the entire pH range. Again, this is largely attributable to differences between oxygenated and deoxygenated red blood cells (Fig. 5B). Only minor differences are observed in the plasma concentration of bicarbonate under these conditions (Fig. 5C). There was a significant difference in the nonbicarbonate buffer value (β= ΔHCO₃^-/ ΔpH) between oxygenated and deoxygenated blood (Table 2). However, there was no significant effect of oxygenation on the nonbicarbonate buffer values determined for red blood cells or plasma (Table 2).

The present data can be used to predict the actual distribution of CO₂ between plasma and red blood cells in oxygenated and deoxygenated blood (Fig. 6). In these experiments, the haematocrit was somewhat lower than that observed in vivo (Tufts, 1991 ; Tufts et al. 1992). This difference is probably because some animals were used in two ‘separate’ series of experiments and considerable volumes of blood were removed from each animal. Thus, to facilitate comparisons with previous in vivo studies, a resting haematocrit of 27.4%, as observed in vivo (Tufts et al. 1992), was used to predict the distribution of CO₂ between plasma and red blood cells in oxygenated and deoxygenated blood (Fig. 6). For both oxygenated and deoxygenated blood, the importance of the red blood cell in CO₂ carriage increases as the increases. In oxygenated whole blood, the CO₂ attributable to the red blood cells increases from 17 % to 35 % over the experimental range. Similarly, in deoxygenated blood, the increase is from 26 % to as much as 48 % at a of 2.90 kPa.

Sodium-dependent pH regulation has been demonstrated in lamprey red blood cells (Nikinmaa, 1986; Nikinmaa et al. 1986; Tufts, 1992). Thus, in order to characterize further the factor(s) associated with the observed relationships between red blood cell pH, and oxygenation, changes in erythrocyte sodium concentration ([Na⁺]) were examined. It was found that [Na⁺] increased with deoxygenation and/or decreased pH (increased ⁠; Fig. 7). A final series of experiments was therefore carried out in which ouabain (10⁻⁴moll^-1) was added to block the Na⁺/K⁺-ATPase. The purpose of these experiments was to eliminate the potential impact of reduced ATP concentrations on the Na⁺ distribution across the erythrocyte membrane. In the presence of ouabain, red blood cell [Na⁺] became significantly elevated in deoxygenated blood compared to blood that remained under oxygenated conditions (Fig. 8). Thus, the observed increases in red blood cell [Na⁺] under deoxygenated conditions in the previous experiments (Fig. 7) are apparently not attributable to a reduction in Na⁺/K⁺-ATPase activity.

As in other vertebrates, the oxygen transport functions of lamprey blood are largely determined by the intrinsic affinity of haemoglobin for oxygen, the cooperativity of haemoglobin oxygen-binding and the effects of pH and CO₂ on haemoglobin function. To our knowledge, the present experiments are the first to examine these characteristics in lampreys using intact red blood cells collected from quiescent cannulated animals at a temperature (10 °C) typical of their environment. Unlike previous studies, we have also demonstrated the critical importance of intracellular pH measurements in these experiments.

The P50 values presented in Table 1 indicate that the affinity of sea lamprey haemoglobin for oxygen is much lower in the present study than previously reported by Manwell (1963) using intact red blood cells. This is somewhat surprising considering that the previous investigation was carried out at a much higher temperature (25 °C). However, the difference may be related to the fact that the present experiments use blood collected from relatively unstressed sea lampreys. In this regard, it is noteworthy that river lampreys exposed to hypoxic stress have a higher blood oxygen-affinity than that of normoxic animals (Nikinmaa and Weber, 1984). Furthermore, stress responses in many species of fish have a significant impact on blood oxygen-affinity (Nikinmaa and Tufts, 1989; Nikinmaa, 1990).

The oxygen dissociation characteristics of sea lamprey whole blood indicate haemoglobin cooperativity and could be described as sigmoidal (nH=1.52–1.89; Table 1). These nH values are higher than the values (nH=10–1.2) originally determined by Wald and Riggs (1951) for dialyzed sea lamprey haemoglobins. Under conditions of increased concentration, deoxygenation or reduced pH, the predominantly monomeric lamprey haemoglobins will interact in dimer and tetramer aggregations (Riggs, 1972; Hardisty, 1979). Thus, the greater concentrations present in intact erythrocytes in the present study have probably resulted in some aggregation and positive cooperativity (i.e. nH> 1). The present values are also somewhat greater than the value of 1.2 reported by Manwell (1963) for intact erythrocytes of adult P. marinus. However, it should be noted that direct comparisons are difficult since our study was conducted at a temperature normally experienced by sea lampreys (10 °C) whereas the experiments of Manwell (1963) were carried out at 25 °C, quite close to the upper limit of their temperature range. We found no evidence of a relationship between the nH values and pH; the nH values were not significantly different over the range of ⁠. This may be explained by the fact that relatively large changes in and/or pHe of lamprey blood are associated with much smaller changes in pHi of the intact erythrocytes (Tufts and Boutilier, 1989, 1990; Tufts, 1991 ; Fig. 3). The described relationship between nH and pH for lamprey haemoglobin may, therefore, be obscured or perhaps not relevant under more physiological conditions using intact cells. Nevertheless, our results do indicate that haemoglobin cooperativity is present in intact sea lamprey erythrocytes at a temperature (10 °C) compatible with their natural environment.

The present study is the first to determine experimentally the Bohr coefficient of lamprey whole blood using both the extracellular and the erythrocyte pH (Fig. 1). To date, extracellular pH has been used exclusively for such determinations and investigators have reported relatively small Bohr coefficients between −0.14 and −0.34 in lamprey whole blood (Bird et al. 1976; Nikinmaa and Weber, 1984). It has been demonstrated, however, that the erythrocyte pH is largely maintained over a broad range of extracellular pH values in the blood of these animals (Nikinmaa, 1986; Tufts and Boutilier, 1989, 1990; Tufts, 1991). As suggested by Nikinmaa and Mattsoff (1992), consideration of only the extracellular pH would therefore largely underestimate the Bohr coefficients. Indeed, in the present study, the Bohr coefficient obtained with respect to the erythrocyte pH is much larger than the value obtained when the extracellular pH values are considered (Fig. 1). Moreover, the Bohr coefficient based on the erythrocyte pH (− 0.63) is much closer to that obtained for sea lamprey haemoglobin solutions (−0.7; Manwell, 1963). In the river lamprey, Nikinmaa and Mattsoff (1992) have also estimated a Bohr coefficient of −0.93 based on the erythrocyte pH. Thus, in whole blood, it is clear that erythrocyte pH values must be considered in order to obtain a reliable measure of the pH sensitivity of lamprey haemoglobin. These results may provide important insights into gas transport in these primitive vertebrates. According to the analysis of Lapennas (1983), oxygen delivery at the tissues will be optimized in mammals when the magnitude of the Bohr effect is about half of the respiratory quotient (i.e. −0.35 to −0.50). This analysis also indicates that pH compensation and CO₂ transport will be optimized when the Bohr effect approaches the respiratory quotient. In lampreys, therefore, Bohr effects previously determined on intact cells using pHe suggested that the system was oriented towards oxygen delivery. However, the present experiments using pHi to determine the Bohr factor clearly suggest that the system may be oriented more towards CO₂ transport, or the functional equivalent of the Bohr effect, the Haldane effect. Indeed, a large Haldane effect may be essential for CO₂ transport in lampreys since an integral component of CO₂ transport in higher vertebrates, erythrocyte chloride/bicarbonate exchange, appears to be limited or absent in agnathans (Ellory et al. 1977; Nikinmaa and Railo, 1977).

Several in vivo studies have demonstrated that the erythrocyte pH in lampreys is considerably higher than in most other vertebrates and negative transmembrane pH gradients have even been reported under certain conditions (Nikinmaa and Weber, 1984; Mattsoff and Nikinmaa, 1988; Tufts et al. 1992). In the river lamprey, L. fluviatilis, the oxygenation state of haemoglobin has a large impact on the erythrocyte pH (Nikinmaa and Mattsoff, 1992). Our results indicate that this is also the case in the sea lamprey (Fig. 2). In the present study, we have further examined this relationship over a range of levels in an attempt to explain the unique relationship between pHe and pHi often observed in vivo. This analysis confirms that very small or negative transmembrane pH gradients may occur under a number of circumstances in sea lampreys in vivo (Fig. 2). Indeed, the negative transmembrane pH gradients recently observed in vivo by Tufts et al. (1992) would be expected in venous blood after exercise as a result of the elevation in and simultaneous reduction in ⁠. Thus, the unique relationship between pHe and pHi which has been documented in lampreys in vivo may be largely explained by the effects of oxygen and carbon dioxide on the pH gradient in these animals.

The large dependence of red blood cell pH on haemoglobin oxygenation is not surprising considering both the large intracellular Bohr effect (Fig. 2) and the fact that the Bohr and Haldane effects in vertebrate haemoglobin are interrelated (Wyman, 1964; Jensen, 1989). In the sea lamprey, the difference in erythrocyte pH between fully oxygenated and deoxygenated blood is greater than in most vertebrates. At an extracellular pH of 7.9, our data indicate a pHi difference of 0.40 pHunits (Fig. 3). In comparison, the pHi difference at the same extracellular pH is 0.35 pH units in the tench (Jensen, 1986) and about 0.27 pH units in the carp (Albers et al. 1983); two animals generally considered to have large Haldane-mediated increases in red blood cell pH. Nikinmaa and Mattsoff (1992) have recently demonstrated that pHi increases by 0.24 pH units at a pHe of 7.5 in another agnathan, the river lamprey, Lampetra fluviatilis, when the oxygen saturation of the blood is reduced from 100% to 7%. This value approaches the 0.29 pH unit difference obtained in the present experiments for fully oxygenated and deoxygenated sea lamprey blood at the same extracellular pH. Thus, the relatively large dependence of red blood cell pH upon oxygenation in comparison with most other vertebrates may be a common feature of the blood of lampreys.

The importance of the Haldane effect for CO₂ carriage in sea lamprey blood is illustrated in Fig. 4A. Over the range examined, the total CO₂ content of whole blood was enhanced by 15–27 % after deoxygenation. It is noteworthy that this increase is entirely due to an increase in the total CO₂ content of the red blood cells (Fig. 4B). At a of 2.9 kPa, the of deoxygenated red blood cells was approximately double that of oxygenated red blood cells. In comparison, plasma was not significantly different between oxygenated and deoxygenated blood (Fig. 4C). The differences in between oxygenated and deoxygenated lamprey erythrocytes are not a result of a significant difference in nonbicarbonate buffer values (Table 2), but there is considerably more bicarbonate formed in deoxygenated lamprey red blood cells at any given (Fig. 5; Nikinmaa and Mattsoff, 1992). In most vertebrates, a similar increase in red blood cell [HCO₃^-] would be associated with the transfer of bicarbonate to the plasma via the anion exchanger within the red blood cell membrane (Roughton, 1964; Cameron, 1979; Perry, 1986; Nikinmaa, 1990). However, there is growing evidence that this process is functionally limited or absent in lamprey blood (Nikinmaa and Railo, 1987; Tufts and Boutilier, 1989, 1990; Nikinmaa, 1990; Nikinmaa and Mattsoff, 1992; Tufts et al. 1992). The present results also indicate that anion exchange is at least functionally absent in vitro since none of the bicarbonate formed as a result of the Haldane effect is transferred to the plasma.

Assuming an in vivo haematocrit of 27.4 % reported by Tufts et al. (1992), the present data can be used to partition the distribution of CO₂ in fully oxygenated and deoxygenated whole blood. This distribution can then be compared with the pattern of CO₂ transport observed in sea lampreys in vivo. At a of 0.2 kPa, the red blood cells carry about 17 and 26% of the total CO₂ in oxygenated and deoxygenated blood respectively (Fig. 6). These values are quite similar to values of 16 and 22 % determined by Tufts et al. (1992) for arterial and venous blood of resting sea lampreys in vivo. Moreover, these values are substantially greater than the 8 and 10% values determined for arterial and venous erythrocytes in the rainbow trout (Heming, 1984). In vitro, the contribution of the red blood cell to total CO₂ carriage increases markedly to 21 % and 45 % for oxygenated and deoxygenated blood, respectively, when the is increased to 0.96 kPa. Similarly, immediately after exercise in vivo, there is an increase in in both arterial and venous blood, and the CO₂ attributable to the red blood cells increases to 25 and 38 %, respectively, in these compartments (Tufts et al. 1992). The distribution of CO₂ between plasma and red blood cells observed in arterial and venous blood in vivo can therefore be largely explained by the present in vitro experiments.

Carbon dioxide excretion across the gills will be largely determined by (i) the carbon dioxide tensions in arterial and venous blood and (ii) the carbon dioxide dissociation characteristics of plasma and red blood cells under both oxygenated and deoxygenated conditions. Taking these factors into consideration, the present in vitro results indicate that arteriovenous differences in red blood cell should account for a large fraction of the CO₂ excreted across the gills of the sea lamprey. For example, when the CO₂ distribution in deoxygenated blood at an elevated CO₂ tension (2.9 kPa) is compared to that of oxygenated blood at a low CO₂ tension (0.2 kPa), it can be determined that 71 % of the difference in whole-blood is attributable to changes in within the red blood cells (Fig. 6). Again, this correlates well with the recent observations of Tufts et al. (1992) indicating that, unlike other vertebrates, the majority of CO₂ excreted across the gills of sea lampreys in vivo can be attributed to arteriovenous changes in red blood cell ⁠. In vivo, changes in red blood cell account for about 60 % of the arteriovenous difference in lamprey whole blood at rest and almost 80 % after exercise (Tufts et al. 1992).

As suggested by Tufts and Boutiler (1990), the absence of net bicarbonate movements across the membrane of lamprey red blood cells may also have a considerable impact on the red blood cell pH in these animals. In most vertebrates, transfer of bicarbonate across the erythrocyte membrane via the anion exchanger is associated with a decrease in red blood cell pH at any given (Heisler, 1986). In contrast to other vertebrates, however, all the bicarbonate formed by deoxygenated haemoglobin in lampreys appears to be retained within their red blood cells (Nikinmaa and Mattsoff, 1992; Figs 4, 5). Thus, the functional absence of bicarbonate transfer across the red blood cell membrane in vitro may be associated with a relatively larger Haldane-mediated increase in the red blood cell pH compared to that in other vertebrates (Figs 2, 3). The relatively high red blood cell pH observed in lampreys in vivo may also be related to the absence of bicarbonate movements from the red blood cell to the plasma (Nikinmaa and Weber, 1984; Mattsoff and Nikinmaa, 1988; Tufts, 1991; Tufts et al. 1992).

There is a marked elevation in the sodium concentration in sea lamprey red blood cells when the of the equilibration gas is increased (Fig. 7). This is also observed in the presence of ouabain after a step increase in (Tufts, 1992). The increase in intracellular sodium concentration is therefore probably not due to a reduction in the activity of the Na⁺/K⁺-ATPase and may indicate activation of ion transport processes across the red blood cell membrane. The present results suggest that sodium-dependent ion transport processes may also be activated in deoxygenated blood (Figs 7, 8). In contrast, Nikinmaa and Mattsoff (1992) recently observed no changes in sodium levels in deoxygenated blood of the river lamprey. Nevertheless, it is noteworthy that the red blood cell pH of both the river lamprey and the sea lamprey are dependent on the extracellular sodium concentration (Nikinmaa, 1986; Nikinmaa et al. 1986; Tufts, 1992). Moreover, extracellular sodium concentration also influences the distribution of CO₂ across the red blood cell membrane in the sea lamprey (Tufts, 1992). Thus, the activation of ion transport processes may, in fact, significantly contribute to the respiratory properties of lamprey blood and further experiments to clarify their relative importance are clearly warranted.

In conclusion, the present results confirm the view that the strategy for gas transport in agnathans is markedly different from that in more recent vertebrates. When the red blood cell pH is used to determine the Bohr effect, it is apparent that CO₂ transport may take precedence over oxygen delivery in the blood of these animals. The present in vitro results also confirm that the majority of CO₂ added to the blood by the tissues is probably transported within the red blood cell to the gills. This strategy may be necessary for effective CO₂ transport since red blood cell chloride/bicarbonate exchange appears to be functionally limited or absent in agnathans. It is possible, however, that retention of bicarbonate within the red blood cell may also compromise oxygen delivery by limiting red blood cell pH changes as blood passes through the tissues. Thus, oxygen and carbon dioxide transport appear to be largely uncoupled in the blood of these primitive vertebrates. In view of the present results, one can speculate that more effective coupling between oxygen and carbon dioxide transport may have been an important selective pressure in the evolution of rapid chloride/bicarbonate exchange in red blood cells.

This study was supported by an NSERC Operating Grant to B.L.T., an NSERC Postdoctoral Fellowship to R.A.F. and NSERC Summer Awards to N.S. and B.B. The authors would also like to thank the Lamprey Control Center (Department of Fisheries and Oceans) in Sault Ste Marie, Ontario, for their assistance in obtaining the sea lampreys.