An in vitro and in vivo Study of the Distribution of Ammonia Between Plasma and Red Cells of Rainbow Trout (Salmo Gairdneri)

In the in vitro experiments blood was collected from the dorsal aortic catheter (for technique, see Soivio, Westman & Nyholm, 1972) of donor fish, pooled, and then divided between tonometers (4ml per tonometer). Each tonometer received either a 0·2 % CO₂ (control) or a 1·0 % CO₂ (hypercapnia) humidified gas mixture in air and was shaken for 90min in a 9°C water bath before measurements were taken. Blood was analysed for whole blood pH (pH_e), red cell pH (pH_i), whole blood and plasma ammonia concentrations (T_amm), plasma and red cell water content, and haematocrit (Hct). In the in vivo experiment, dorsal aortic cannulated fish were placed in individual, low volume (21) flow-through chambers to recover for 48 h. In the 30 min prior to sampling, inflow water was turned off and fish chambers were aerated with either 100% air (water pH = 7·0, control) or switched to 1 % CO₂ in air (water pH = 5·6, hypercapnia). A 2ml blood sample was withdrawn from each fish at the end of 30 min and analysed for pH_c, pH_i, whole blood and plasma T_amm, plasma and red cell water content, and Hct.

Slight quantitative differences between in vitro and in vivo data are shown in Table 1, but the overall results and conclusions are the same whether blood was held in tonometers (in vitro) or in live animals (in vivo) prior to analysis. Red cell ammonia levels are consistently higher than plasma levels, resulting in ammonia concentration ratios (plasma-to-red cell) of between 0·3 and 0·4 (Table 1). Control red cell pH, predicted from the plasma-to-red cell ammonia distribution was not significantly different from measured pH_i and there was no difference between calculated plasma and red cell NH₃ tensions in the control experiment (in vitro and in vivo, Table 2). The same was not true for the hypercapnia experiment (in vitro and in vivo), where predicted red cell pH, was significantly different from measured pH₁ and calculated red cell was greater than plasma (Table 2). Our calculations of levels assume an equilibrium between NH₃ and NH₄⁺ in each compartment. Thus, when there is an NH₃ gradient from red cell to plasma there will also be an electrochemical gradient for NH₄⁺. In hypercapnia, therefore, there is a net diffusion gradient for both NH₃ and NH₄⁺ out of the red cell.

Ammonia gradients during hypercapnia may develop between intra- and extra-cellular compartments because of high rates of ammonia production. We tested this possibility in trout blood in vitro, by following whole blood T_amm levels over time during hypercapnic exposure, and found that ammonia levels did not change. Thus, intracellular ammoniagenesis is not a factor in the development of ⁠. gradients during hypercapnia. Ammonia accumulation in the red cell, therefore, can only be maintained by the active uptake of NH₄⁺ in the face of NH₃ diffusion out of the red cell down the ⁠, gradient and NH₄⁺ electrochemical gradient. Secondary active transport of NH₄⁺ is linked to the energetically favourable movement of Na⁺ in many cells (see Maetz & Garcia-Romeu, 1964; Kinsella & Aronson, 1981; Wright & Wood, 1985). The trout red cell membrane Na⁺/H⁺ exchange mechanism is known to be active during an acidosis (Nikinmaa, Steffensen, Tufts & Randall, 1987), but if NH₄⁺ can replace H⁺ in exchange for Na⁺, then red cell ammonia stores would be depleted during hypercapnia. Instead, we observed an accumulation of ammonia during hypercapnia, therefore NH₄⁺ substitution for H⁺ in Na⁺/H⁺ exchange cannot be involved. The ability of NH₄⁺ to replace K⁺ in Na⁺,K⁺-ATPase is well established in many tissues, including red cell membranes (Post & Jolly, 1957; Sorensen, 1981). We tested the possibility that NH₄⁺ was replacing K⁺ in the Na⁺,K⁺-ATPase by adding the specific Na⁺,K⁺-ATPase inhibitor, ouabain, to hypercapnic blood, in vitro (Tables 3,4). The addition of ouabain did not alter the distribution of ammonia between red cells and plasma (Table 3) and red cell-to-plasma gradients were not abolished (Table 4). This implies that even if Na⁺,K⁺-ATPase plays a role in ammonia accumulation within the red cell during hypercapnia, it cannot be a major one.

It is possible that changes in pH and water content, which will lead to changes in ammonia distribution, may have caused the development of ammonia gradients between red cell and plasma during hypercapnia. Whole blood pH remained stable after 30 min of hypercapnia in vitro. Red cell water content increased significantly between control (in vitro, 65·1 ±0·4%; in vivo, 65·7 ±0·3%) and hypercapnia (in vitro, 68·4 ± 0·3 %; in vivo, 68·9 ± 0·3 %) experiments. It seems likely that the water content of red cells was stable following 90 min of exposure to hypercapnia in vitro. Thus it appears that non-steady states for pH and water content cannot account for the red cell-to-plasma gradients in vitro during hypercapnia. It also seems to be an unlikely explanation of the in vivo results because of the similarity of the in vitro and in vivo data.

We conclude that ammonia is passively distributed according to the plasma-to-red cell H⁺ distribution in blood at resting pH values, but not during hypercapnia. Ammonia accumulation during hypercapnia cannot be accounted for by red cell ammoniagenesis or NH₄⁺ substitution for K⁺ in the Na⁺,K⁺-ATPase, but must be due to some other active NH₄⁺ uptake process. Whether ammonia is passively distributed between plasma and other intracellular compartments in fish is not known. This question is interesting in light of the fact that the distribution of H ⁺ across intracellular compartments, other than red cell membranes (Lassen, 1977; Heming et al. 1986), is not passive (see Roos & Boron, 1981). Thus, if NH₄⁺ is able to move across tissue membranes, then one would predict that the distribution of ammonia would follow the membrane potential and not the H⁺ distribution. However, if tissue membranes are essentially impermeable to NH₄⁺ then one would expect the distribution of ammonia to follow the H⁺ distribution, as do other weak acids and bases with impermeant ion forms (see Randall & Wright, 1987).

The authors wish to express their appreciation to Dr G. K. Iwama for helpful discussions during this project. This work has been funded by an EPA grant (CR 812 397-01-0) and an NSERC operating grant to DJR.