The Absence of Rapid Chloride/Bicarbonate Exchange in Lamprey Erythrocytes: Implications for CO₂ Transport and Ion Distributions Between Plasma and Erythrocytes in the Blood of Petromyzon Marinus

In mammals, as blood passes through the tissues, CO₂ diffuses into the erythrocyte where it combines with water to form bicarbonate and a proton in the presence of erythrocytic carbonic anhydrase. The bicarbonate formed by the CO₂ hydration reaction is rapidly exchanged for plasma chloride and the bulk of CO₂ transported in the blood is in the form of plasma bicarbonate (Roughton, 1964). At the lung, the bicarbonate re-enters the erythrocyte where it combines with a proton to form water and CO₂, which is excreted. Rapid chloride/bicarbonate exchange is an integral part of this CO₂ transport system. In this system, the chloride/bicarbonate exchange and the acid-base reactions within the erythrocyte also link the movements of protons, bicarbonate and chloride, and these ions are distributed in a Donnan equilibrium between the plasma and the erythrocyte (Hladky & Rink, 1977). Rapid, passive anion exchange has also been demonstrated in nucleated erythrocytes (Cameron, 1978; Obaid et al. 1979; Romano & Passow, 1984). Thus, the CO₂ transport properties of the blood of the lower vertebrates are basically similar to those of mammalian blood (Boutilier et al. 1979; Boutilier & Toews, 1981; Randall & Daxboeck, 1984). As in mammals, the presence of rapid anion exchange on the erythrocyte membrane also contributes to the similar distribution ratio for protons, bicarbonate and chloride ions in the blood of these animals (McDonald et al. 1980; Albers & Goetz, 1985; Heming et al. 1986; Tufts & Randall, 1988).

Recent evidence suggests that chloride/bicarbonate exchange may be absent in Lampetra fluviatilis erythrocytes (Nikinmaa & Railo, 1987). In addition, Ellory et al. (1987) have demonstrated that chloride/bicarbonate exchange activity is extremely low in erythrocytes of the hagfish, Eptatretus stouti. Clearly, the low activity or absence of this exchanger would have a large influence on the CO₂ transport properties and distributions of ions in the blood of these animals. The purpose of these experiments was, therefore, to determine whether the CO₂ transport properties and distributions of ions in the blood of the lamprey Petromyzon marinus support the view that the anion exchanger is virtually absent in agnathan erythrocytes.

Adult lampreys, Petromyzon marinus Linnaeus, were collected during their spawning migration from a fishladder on the Lahave River in Nova Scotia, Canada. The animals were transported to the Aquatron facility at Dalhousie University where they were held in freshwater aquaria at 10°C for at least 1 week prior to the experiment.

The blood used in these experiments was obtained from anaesthetized lampreys via caudal vessel puncture or from resting lampreys via a dorsal aortic cannula which had been surgically implanted at least 24 h prior to the experiment.

In the first series of experiments, the collected blood was equilibrated in an intermittently rotating glass tonometer at 10 °C with a humidified gas mixture of 0·2%, 1% or 3% CO₂ in air (Wösthoff gas-mixing pumps, Bochum, FRG). Following a 30min equilibration period, a 1ml sample was removed from the tonometer using a 1ml Hamilton gas-tight syringe. A 50μl sample of this blood was immediately dispensed into another Hamilton syringe for the analysis of total CO₂ content. Another 100 μl portion of the original sample was used for duplicate hematocrit measurements. Following the determination of the hematocrit, 50 μl of the plasma fraction from the hematocrit tubes was analyzed for total CO₂ content. The remainder of the original sample was divided equally between two 0·5 ml Eppendorf tubes and centrifuged. After centrifugation, the plasma pH of the first tube was measured and the remaining plasma from this tube was discarded. The cell pellet was frozen in liquid nitrogen for later analysis of erythrocyte pH. The plasma from the second tube was removed, frozen in liquid nitrogen and stored for subsequent ion analyses. The erythrocyte pellet from this tube was saved for the determination of both erythrocyte water content and ion concentrations. This equilibration and sampling procedure was repeated at each of the specified gas concentrations.

Another series of experiments examined the effects of the protonophore 2,4-dinitrophenol (2,4-DNP) on the distribution of ions across the lamprey erythrocyte membrane. In these experiments, the collected blood was divided equally between two tonometers. The blood from the first tonometer was incubated in the presence of 10⁻⁴moll^-1 2,4-DNP. An equivalent volume (50μl) of saline was added to the second tonometer which served as a control. The equilibration and sampling procedure were similar to the first series of experiments. In this series, however, the total CO₂ content determinations were omitted and the humidified gas mixtures also included 0·6% and 2% CO₂ in air.

CO₂ dissociation curves of whole blood, true plasma (plasma equilibrated with erythrocytes) and erythrocytes are compared in Fig. 1. At a of 0·2 kPa (=1·5 mmHg), values for the three compartments were similar. At elevated values, however, the erythrocytes had the highest followed by whole blood and true plasma. At the highest experimental (3·01 kPa), the of the erythrocytes reached 13·4 ± 1·0mmoll^-1 whereas the of whole blood at this was 8·1 ± 0·5 mmol l^-1 and that of true plasma was 5·7 ± 0·5 mmol l^-1. The erythrocytes also had the highest buffering capacity of the three compartments examined (Fig. 2). The erythrocyte buffer value (ΔHCO₃^-/ ΔpH) was —37·0mequivl^-1pHunit^-1, whereas the buffer values of whole blood and true plasma were only –3·3 and –0·1 mequivl^-1 pH unit^-1, respectively.

The consequences of the large buffering capacity of the erythrocyte and the low buffering capacity of the plasma were demonstrated when erythrocyte pH was plotted against plasma pH in lamprey blood (Fig. 3). The slope of the regression line generated for this relationship was only 0·27 and, therefore, a plasma pH change of 0·368 units was required to cause an erythrocyte pH change of 0·1 unit. The relationship between pHe and pHi was not significantly different when lamprey blood was equilibrated in the presence of the protonophore 2,4-DNP. Thus, the protons appeared to be passively distributed across the erythrocyte membrane in lamprey blood in the present experiments. It should also be noted that the relationship between pHe and pHi for blood obtained via caudal vessel puncture was not significantly different from that for blood obtained via the dorsal aortic cannula in resting animals.

The distribution ratios of sodium, potassium, chloride and protons across the erythrocyte membrane in lamprey blood are plotted against pHe in Fig. 4. Of these, only the distribution ratio for protons (rH⁺) changed according to the extracellular pH. It is noteworthy that the distribution ratio for chloride was not only independent of pHe, but was far removed from rH⁺. Indeed, the overall mean rCl^- in the present experiments was only 0·04. The low rCl^- values are a reflection of the extremely low erythrocyte concentrations of chloride measured in these experiments (Table 1). It is unlikely that these values represent an error in analysis since simultaneous analysis of trout blood (N = 6) resulted in an erythrocyte chloride concentration of 33·6mequivl^-1 and an rCl^- of 0·25. These trout values are similar to results obtained by other researchers (Ferguson & Boutilier, 1988; Heming et al. 1986). It should also be pointed out that any error caused by trapped plasma in the erythrocyte pellet would result in an overestimate of the erythrocyte chloride concentration. Thus, the actual erythrocyte chloride concentration in Petromyzon marinus may be slightly lower than the value of 4·3 ± 0·3mequiv l^-1 obtained in these experiments.

The erythrocyte water content is plotted against extracellular pH in Fig. 5-Water content increased as pHe decreased, but the magnitude of the increase was only 0·07 %/0·l pH unit. Thus, the erythrocyte water content varied less than 1 % over the entire pH range studied.

Nikinmaa & Railo (1987) have demonstrated that the apparent chloride permeability of Lampetra fluviatilis erythrocytes is similar to the apparent chloride permeability of lipid bilayers at the same temperature. In the same study, these authors found that chloride movements across the erythrocyte membrane were not affected by the anion exchange inhibitor 4,4-diisothiocyanostilbene 2,2-disul-phonic acid (DIDS). Ellory et al. (1987) have also demonstrated extremely low chloride/bicarbonate exchange activity in erythrocytes from the hagfish, Epta-tretus stouti. In the present study, the CO₂ transport properties of the blood from another agnathan species, Petromyzon marinus, clearly support the growing view that erythrocytic chloride/bicarbonate exchange is virtually absent in agnathan erythrocytes.

The CO₂ transport properties of mammalian blood have been extensively described by Roughton (1964). Similar models of CO₂ transport have been described in the blood of other vertebrates (Boutilier et al. 1979; Boutilier & Toews, 1981; Randall & Daxboeck, 1984). In these systems, the majority of the bicarbonate formed by the CO₂ hydration reaction within the erythrocyte is exchanged for plasma chloride via the anion exchange mechanism on the erythrocyte membrane. Thus, at any given the total CO₂ content of true plasma is greater than that of either whole blood or erythrocytes (Roughton, 1964; Boutilier et al. 1979; Boutilier & Toews, 1981). In the present study, the total CO₂ content of true plasma was lower than that of either whole blood or erythrocytes (Fig. 1). Indeed, in Petromyzon marinus blood, the erythrocytes had the highest content over the majority of the range studied. Carbamino formation could contribute somewhat to the high erythrocyte levels. These results suggest, however, that the bicarbonate formed within the erythrocytes of Petromyzon marinus is not transferred to the plasma, but remains within the erythrocyte.

According to Nikinmaa & Railo (1987), erythrocyte buffering is effectively isolated from the extracellular compartment in Lampetra fluviatilis since practically no extracellular pH recovery occurred in an unbuffered erythrocyte suspension after acidification or alkalinization. The nonbicarbonate buffer lines of true plasma, whole blood and erythrocytes for Petromyzon marinus provide further evidence that erythrocytic bicarbonate does not have access to the plasma in agnathan blood (Fig. 2). The slope of the nonbicarbonate buffer line for true plasma in this study is only —0·1mequivl^-1pHunit^-1, whereas that of whole blood is –3·3mequivl^-1pHunit^-1. In other species, the slope of the true plasma buffer line is invariably higher than that of whole blood (Woodbury, 1974; Boutilier et al. 1979; Boutilier & Toews, 1981; Heisler, 1986). This relationship occurs because bicarbonate formed within the erythrocyte (after nonbicarbonate buffering of protons) is transferred via the anion exchanger to the plasma to reestablish the Donnan equilibrium for bicarbonate (Woodbury, 1974; Heisler, 1986). Only the nonbicarbonate buffer lines of separated plasma (i.e. plasma that has not been equilibrated in the presence of erythrocytes) for other species approach the slope of the true plasma buffer line found for Petromyzon marinus blood. The explanation for this result can only be that the bicarbonate formed within the erythrocyte does not reach the plasma. This would also explain the unusually high nonbicarbonate buffer value for the erythrocytes (—37·0mequiv l^-1 pH unit^-1). In trout erythrocytes, the nonbicarbonate buffer value at a similar temperature is only 16·5mequivl^-1pHunit^-1 (calculated from Heming et al. 1986). The measured nonbicarbonate buffer values of Petromyzon true plasma and erythrocytes do not resemble the ‘apparent’ buffer values commonly measured for these compartments in other vertebrates. Our buffer values do, however, resemble the actual nonbicarbonate buffer values which would be expected for these compartments prior to the bicarbonate redistribution process (Heisler, 1986).

The erythrocyte pH in Lampetra fluviatilis is actively maintained by a sodium/ proton exchange mechanism (Nikinmaa, 1986; Nikinmaa et al. 1986). In the present experiments, however, the pH of Petromyzon erythrocytes was not affected by the protonophore 2,4-DNP (Fig. 3). These results suggest that protons are passively distributed across the erythrocyte membrane in Petromyzon. Thus, in Petromyzon the relatively low slope of the pHi versus pHe regression line must be due solely to the high buffering capacity of the erythrocytes.

In nucleated erythrocytes, the chloride/bicarbonate exchanger normally increases the permeability of the erythrocyte membrane to chloride ions and links the movements of chloride, bicarbonate and protons (Hladky & Rink, 1977; Fortes, 1977). Thus, in erythrocytes with a rapid anion exchange mechanism, the distribution ratios for protons and chloride ions are similar (Albers & Goetz, 1985; Heming et al. 1986; Tufts & Randall, 1988). In Petromyzon blood, however, the distribution ratios for protons and chloride ions are strikingly different from each other (Fig. 4). Such differences suggest that either protons or chloride ions are not in Donnan equilibrium. Our data indicate that protons are in electrochemical equilibrium in Petromyzon erythrocytes since the protonophore 2,4-DNP did not alter the relationship between pHe and pHi (Fig. 3). Chloride ions are, therefore, probably not distributed according to a Donnan equilibrium in Petromyzon erythrocytes. This result is not entirely explained by the absence of the anion exchanger since passive conductive pathways for chloride should bring this ion into electrochemical equilibrium. Ellory et al. (1987) have suggested that chloride ion movements may occur across the membrane of Eptatretus stouti erythrocytes via an unidentified transporter which is not DIDS-sensitive. A similar transport mechanism may be keeping chloride out of electrochemical equilibrium in Petromyzon erythrocytes and this possibility clearly warrants further investigation.

In erythrocytes with an anion exchanger, there is an increase in erythrocyte chloride concentration, and therefore rCl^-, with decreasing pH, and water is drawn into the cell osmotically (Hladky & Rink, 1977; Heming et al. 1986; Nikinmaa et al. 1987; Tufts & Randall, 1988). In Petromyzon rCl^- did not change significantly and the water content of erythrocytes varied less than 1 % over the pH range studied (Figs 4, 5). A similar situation occurs when trout erythrocytes have been incubated in the presence of the anion exchange blocker DIDS (Nikinmaa et al. 1987). Thus, the absence of rapid chloride movements across the erythrocyte membrane also results in a rather constant erythrocyte volume.

The extremely low rCl^- in Petromyzon blood results from the unusually low levels of chloride within the erythrocyte (Table 1). It is unlikely that these levels represent a measurement error since the distribution ratios for sodium and potassium in Petromyzon blood (Fig. 4) and simultaneously determined chloride distribution ratios for Salmo gairdneri blood (see Results section) gave values similar to those in the literature. According to Hladky & Rink (1977), bicarbonate movements via the anion exchanger will be greatly influenced by the erythrocyte chloride concentration. Thus, it follows that vertebrate erythrocytes contain ample levels of chloride to meet the requirements of the anion exchanger during the excretion of CO₂. In this regard, it is interesting that the chloride concentrations in Petromyzon erythrocytes are extremely low, since the results of this study suggest that the anion exchanger is absent in Petromyzon erythrocytes.

Clearly, the transport and excretion of CO₂ in agnathans does not conform to any of the models described for other vertebrates since these models all involve rapid chloride/bicarbonate exchange across the erythrocyte membrane. In agnathan blood, CO₂ still diffuses across the erythrocyte membrane and is hydrated to form bicarbonate and a proton. Nikinmaa et al. (1986) have demonstrated that carbonic anhydrase is present in Lampetra fluviatilis erythrocytes. As in other vertebrates, therefore, this reaction proceeds at the catalysed rate. In the absence of rapid chloride/bicarbonate exchange, however, the resulting bicarbonate is transported within the erythrocyte to the gas exchange organ rather than exported for carriage in the plasma. At the gas exchange organ, oxygenation of hemoglobin presumably liberates the protons that combine with bicarbonate to evolve CO₂. These events are summarized in Fig. 6.

The lamprey is an extant member of a phylogenetically primitive group of vertebrates. These results, therefore, invite speculation about the evolution of gas transport by the vertebrate erythrocyte. In its early stages, the erythrocyte may have evolved to carry both oxygen and carbon dioxide. The importance of the plasma in CO₂ transport and, therefore, of elevated levels of plasma bicarbonate must only have arisen after the evolution of erythrocytic chloride/bicarbonate exchange.

Financial support for this study was provided by an NSERC operating grant to RGB and an NSERC Postdoctoral Fellowship to BLT. We would also like to thank the Department of Fisheries and Oceans and Eric and Elmer Jefferson for their help in obtaining the lampreys.