Ammonia, Na⁺, K⁺ and Cl⁻ Levels in Rainbow Trout Yolk-Sac Fry in Response to External Ammonia

Ammonia is the major nitrogenous excretory product of teleost fish that is lost from the body across the gills to the aqueous environment. Ammonia is a weak base (pK value about 9.64 at 10°C) and at physiological pH values (pH7–8) less than 2.5% of the total ammonia exists as the unionised form, NH₃. Unionised ammonia accounts for the major part of metabolic ammonia excretion by diffusion down its concentration gradient (Cameron and Heisler, 1983; Wright and Wood, 1985). The ionised form, NH₄⁺, may be excreted in exchange for Na⁺ and possible mechanisms have been discussed on numerous occasions (e.g. Maetz, 1973; McDonald et al. 1989).

The chemistry of ammonia in fresh water has been extensively reviewed and toxicity to fish is usually expressed as NH₃ concentration (EIFAC, 1970; Alabaster and Lloyd, 1982; Erikson, 1985; WHO, 1986). Ammonia toxicity to rainbow trout is apparently not influenced by the nature of the ammonium salt (Thurston and Russo, 1983), although chloride is marginally more toxic than sulphate to channel catfish (Sheehan and Lewis, 1986). Ammonia has severe and often lethal effects on sensitive species such as salmonids at concentrations as low as 12 μmol l⁻¹ NH₃ (EIFAC, 1970; Alabaster and Lloyd, 1982). Physiological effects include disturbances of ionic balance and acid–base balance (Maetz, 1973; Cameron and Heisler, 1983; Cameron, 1986; Twitchen and Eddy, 1993). However, in considering toxic effects, the temperature, pH and total ammonia content of the water should be known; e.g. an increase in pH by 1 unit will increase unionised ammonia by a factor of ten. The competitive inhibition of Na⁺ influx by NH₄⁺ at the gills may also be important (Maetz and Garcia-Romeu, 1964; Maetz, 1973; Twitchen and Eddy, 1993).

When considering the physiological effects of external ammonia, it is important to know its rate of entry into the body, how much can be accumulated and the rate of loss once the exposure has ended. There is very little information of this type apart from some values for changes in blood plasma concentrations (Cameron and Heisler, 1983; Cameron, 1986; Wilson and Taylor, 1992), and there are few data for tissue ammonia levels in fish (e.g. Randall and Wright, 1987; Dobson and Hochachka, 1987).

This paper describes experiments on the response of late yolk-sac fry to a pulse of ammonia, explores the rate of ammonia build up in the body, together with its effects on ionic balance, and examines post-exposure effects.

Rainbow trout [Oncorhynchus mykiss (Walbaum)] late alevins (triploids), similar to development stage 13–14 described for salmon alevins, Salmo salar L. (Pelluet, 1944), were obtained from a local hatchery (Cloan Hatcheries, Perthshire). Brown trout (Salmo trutta L.) late alevins at a similar developmental stage were obtained from Loch Leven Fisheries, Fife. Stocks were maintained in running Dundee aquarium water at 6.5–8°C. Towards the end of the experimental period, the fish were almost at the start-feed fry stage, but were not fed. The experimental temperature was 9–11°C. Mean values for water quality were (in mmol l⁻¹): Na⁺, 0.19; K⁺, 0.02; Ca²⁺, 0.24; Mg²⁺, 0.07; Cl⁻, 0.3; free CO₂, 0.02. Alkalinity as CaCO₃ was 20.5mg l⁻¹; total hardness as CaCO₃, 31.2mg l⁻¹ and non-bicarbonate hardness as CaCO₃, 10.6mg l⁻¹; pH was 8.2.

Between 12 and 18 rainbow trout alevins were placed in each of six darkened containers with 1l of constantly aerated aquarium water and were allowed to equilibrate at 10°C and pH7. After 24h, the water was replaced and ammonium chloride was added to give nominal concentrations of 0 (control), 7.2 (3.2), 14.4 (6.3), 21.7 (9.5), 28.9 (12.7) and 36.2 (15.8) μmoll⁻¹ NH₃ (mmol l⁻¹ ammonia), calculated according to the method of Cameron and Heisler (1983). The pH value was adjusted to and maintained at pH7 with dilute sulphuric acid. After 24h, eight alevins were removed, gently blotted to remove surface moisture, and the wet mass of each individual was determined. To avoid stress, alevins were weighed only at the end of each experiment (see below). Each alevin was individually homogenised in 1ml of ice-cold deionised water with a few strokes of a glass homogeniser, a procedure lasting for 10–20s. After centrifugation at 13000revsmin⁻¹ for 2.5min, the supernatant was withdrawn, immediately frozen to -30°C and stored until required for ammonia and Cl⁻ analysis. Since Na⁺ may be released from glass homogenisers, this procedure was unsuitable for Na⁺ analysis and an alternative method was selected. The remaining alevins (in most cases at least five, see Fig. 1) from each ammonia exposure were removed and surface water was blotted as before. They were individually wet weighed, each placed in a plastic tube, killed by compressing the head with forceps and dissolved in 0.5ml of concentrated nitric acid for subsequent Na⁺ and K⁺ analysis. This method is unsuitable for chloride determination, since HCl is lost during the acid digestion (Conway, 1957).

In a further experiment to examine loading and unloading of body ammonia, at least 300 rainbow trout alevins in 2.5l water were exposed to 36.2 μmol l⁻¹ NH₃, pH7, a level designed to provoke a maximal sublethal effect in 24h. Sixteen individuals were removed at 2, 4, 8 and 24h for analysis. Then fresh ammonia-free water, adjusted to pH7 as before, was substituted and further batches of 16 individuals were removed after 2, 4, 8, 24, 48, 72, 120 and 168h, together with water samples for ionic analysis. Water was changed every other day. A control experiment with 100 alevins was run concurrently. Alevins were individually weighed and processed for ammonia and ionic analysis, as described previously.

To investigate the effects of ammonia exposure on wet mass and percentage water content, 50 brown trout late alevins were placed in 1l of aquarium water containing 0 or 36.2 μmol l⁻¹ NH₃, pH7 and 10°C, as described above. Groups of eight alevins were removed at intervals, wet mass determined for each individual as described above, and then they were dried at 65°C for 48h before being reweighed (see Table 2). This involved wet weighing the alevin only once, since this procedure was very stressful. It was observed in a similar experiment that when individuals were wet weighed more than twice there was 100% mortality of both unexposed and exposed groups.

Ammonia analysis of homogenates was by the Boehringer Mannheim UV method, where ammonia content of the sample is equivalent to the amount of NADH oxidised by glutamate dehydrogenase. Known amounts of ammonia were added to some of the samples, resulting in recoveries in excess of 75%. The presence of glutamate dehydrogenase (or other ammonia-metabolising enzymes) in tissue samples may result in an underestimation of tissue ammonia content, although this would be minimal since tissue samples were immediately frozen following preparation. Reanalysis of samples which had been frozen for several weeks showed no loss of ammonia. This method of tissue preparation was preferred since it offered few opportunities for ammonia release from proteins, a possible cause of overestimation of tissue ammonia. As all rainbow trout alevins were processed in the same way, differences between control and ammonia-exposed alevins are directly comparable. Ammonia content of the water was determined by the indophenol method (Solorzano, 1969). Na⁺ and K⁺ concentrations were determined using an EEL 100 flame photometer, and chloride levels in the homogenate were determined with a Jenway chloride meter.

Ammonia refers to total ammonia, NH₃ refers to unionised ammonia and NH₄⁺ to ionised ammonia. In the Discussion, terms referring to ion fluxes are: unidirectional influx, influx; unidirectional efflux, efflux. The difference between unidirectional fluxes represents the net flux: net gain, uptake; net loss, loss.

The body ammonia concentration of rainbow trout alevins increased from about 0.6 μmol g⁻¹ for unexposed fish to nearly 4 μmolg⁻¹ after 24h of exposure to 36.2 μmol l⁻¹ NH₃ (15.8mmol l⁻¹ ammonia) (Fig. 1D). As well as gaining ammonia, the fish also lost body ions, about 10 μmol g⁻¹ alevin for Na⁺ and Cl⁻, and over double that value for K⁺ at 36.2 μmol l⁻¹ NH₃. Lower ammonia exposure levels produced lesser ammonia burdens and lower ionic losses (Fig. 1). There was no significant difference in mass of rainbow trout alevins following ammonia exposure, suggesting that changes in body ion content were not due to tissue dilution and mass increase (Table 1).

There was no significant difference in wet mass or percentage water content between unexposed and exposed late brown trout alevins (Table 2), confirming that mass increase through increased water content during ammonia exposure of rainbow trout alevins is extremely unlikely (Tables 1 and 3). Subsequent calculations are therefore based on this assumption.

Upon exposure to 36.2μmoll⁻¹ NH₃, rainbow trout alevins showed a rapid build up of body ammonia that was almost complete by 10h and reached a value of about 6 μmol g⁻¹ alevin by 24h (Fig. 2D). Subsequent exposure to ammonia-free water resulted in a decrease in body ammonia, that was again almost complete in 10h, to a value of 2–3 μmol g⁻¹, which is higher than the initial level of around 1 μmol g⁻¹ but comparable to that of the control group at this time (Fig. 2). Ammonia exposure did not result in any significant mass differences between exposed and corresponding unexposed alevins (Table 3).

During ammonia exposure, 4–5h elapsed before body Na⁺ levels began to fall, and by 24h some 4–5 μmol g⁻¹ had been lost. Following return to ammonia-free water, there was again a delay of some 4–5h before body Na⁺ values began to recover, reaching normal levels 8h after exposure (Fig. 2). Body Cl⁻ levels fell by 8μmol g⁻¹ after 24h of ammonia exposure, but did not show the delay noted for Na⁺. Cl⁻ recovery was complete within 10h but was subsequently unstable compared with the control. Body K⁺ values fell by about 14 μmol g⁻¹ after 24h of ammonia exposure, but recovery to normal values of about 70 μmol g⁻¹ was slow and incomplete even after a week in ammonia-free water. Over the 8-day experiment, Na⁺, K⁺ and ammonia values for unexposed rainbow trout alevins showed a steady increase, while Cl⁻ values decreased (Fig. 2).

Net fluxes of Na⁺, K⁺, ammonia and Cl⁻ were calculated for each group of rainbow trout alevins from the change in body ionic content during the ammonia exposure (Tables 4 and 5) and the results parallel the trends shown in Figs 1 and 2.

Following exposure of rainbow trout alevins to ammonia for 24h, the water Na⁺ and K⁺ values increased (Table 6), reflecting net Na⁺ and K⁺ loss from the alevins (Fig. 2 and Table 5). The net fluxes calculated from changes in water content (assuming 300 alevins of average mass 100mg) were about 1.11 μmol g⁻¹ h⁻¹ for K⁺ and 0.49 μmol g⁻¹ h⁻¹ for Na⁺, which are greater than those estimated from changes in body ionic content (Table 5). This suggests that net ion flux rates based on body ionic content changes (Table 5) may be an underestimate for the exposed population, since moribund alevins which were likely to have lost greater amounts of body ions were not selected for analysis. The experimental protocol precluded an estimate of net fluxes for individuals from water ionic content changes. Because of high background levels (15.8mmol l⁻¹), significant changes in external Cl⁻ level were undetectable. Throughout the experiment (Fig. 2), except during ammonia exposure, water Na⁺, K⁺ and Cl⁻ values were within 10% of those given in Materials and methods and total ammonia levels were less than 0.01mmol l⁻¹.

No mortalities were observed in any of the experiments, although 24h of exposure to

21.7 μmol l⁻¹ NH₃ and above resulted in hyperexcitability, darker coloration and morbidity of rainbow trout alevins. One week after exposure, the rainbow trout alevins appeared to be almost fully recovered. Amongst the controls, no abnormalities or mortalities were observed.

The body ammonia levels of about 0.6 μmol g⁻¹ for unexposed rainbow trout late alevins is an average value for all tissues. Body ammonia levels appear to be a function of development since, at the end of the 8-day experiment, they had increased to around 2 μmol g⁻¹ (Fig. 2). These values are higher than blood plasma values of about 0.3mmol l⁻¹ (Cameron and Heisler, 1983; Randall and Wright, 1987), but similar to tissue values of about 1–2mmol l⁻¹ for white muscle of unexercised rainbow trout (Dobson and Hochachka, 1987; Wright and Wood, 1988) and estimated whole-body ammonia values of about 0.6mmol l⁻¹ for a 1kg fish (Randall and Wright, 1987). Values of about 6mmol l⁻¹ in lemon sole muscle were noted by Wright et al. (1988), suggesting differences between freshwater and marine fish.

Body Na⁺, K⁺ and Cl⁻ values for unexposed rainbow trout alevins were in the range reported for freshwater and seawater salmonids by Talbot et al. (1982, 1986) and Rombough and Garside (1984). Over the 8-day experimental period, control alevins showed an increase in values for Na⁺ and K⁺, but a decrease in the value for Cl⁻. Such changes for cations have been previously noted (Runn and Sohtell, 1982; Rombough and Garside, 1984; McWilliams and Shephard, 1991). These changes in body ions probably relate to growth, development and metabolism.

Exposure of rainbow trout alevins to 36.2 μmol l⁻¹ NH₃ (15.8mmol l⁻¹ ammonia) for 24h increased the body ammonia levels to about 4 μmol g⁻¹ (Fig. 1). The response was not linear with lower ammonia levels, e.g. 14.4 μmol l⁻¹ NH₃ produced nearly the same increase as the higher levels (Fig. 1). The reason for this is unknown, but a threshold value for body ammonia build up may be involved.

The diffusion gradients for both ionised and unionised species of ammonia are from water to body fluids, but it is not possible to conclude whether ammonia loading is through uptake from the water or through the build up of metabolic ammonia, or a combination of both. Diffusive uptake of NH₃ down its partial pressure gradient is believed to be the primary route for ammonia entry into the fish (WHO, 1986; Randall and Wright, 1987; Wilson and Taylor, 1992). This would result in intracellular alkalisation through titration of protons (Avella and Bornancin, 1989; Wilson and Taylor, 1992). Although ammonia loading may stimulate Na⁺/NH₄⁺ exchange (Cameron and Heisler, 1983; Wright and Wood, 1985), intracellular alkalisation might reduce proton excretion and thus Na⁺ influx via the Na⁺ channel, which is the primary route for Na⁺ uptake (Avella and Bornancin, 1989; Lin and Randall, 1991), thus contributing to further Na⁺ loss. External NH₄⁺ may also contribute to ammonia loading and Na⁺ imbalance, since it is a competitive inhibitor of Na⁺ uptake (Twitchen and Eddy, 1993).

Ammonia exposure of both rainbow trout and catfish resulted in increased blood plasma ammonia levels, which reached a steady state despite the net inward ammonia gradient, explained by export of NH₄⁺ in exchange for Na⁺ or H⁺ (Cameron and Heisler, 1983; Cameron, 1986; Wilson and Taylor, 1992). However, in exposed alevins, ammonia loading of body tissues as well as blood plasma might be expected, since in many membrane processes NH₄⁺ is known to displace cellular K⁺ (Binstock and Lecar, 1969), and ammonia distribution between tissues has been shown to be a function of membrane potential (Wood et al. 1989; Tang et al. 1992). It seems likely that the body tissues of ammonia-exposed alevins also reach a steady state (Fig. 2).

After exposure, NH₃ would rapidly diffuse down its partial pressure gradient from fish to water, resulting in intracellular acidification. Stimulation of proton excretion may be expected and hence Na⁺ influx via Na⁺ channels (Avella and Bornancin, 1989; Lin and Randall, 1991), so facilitating recovery of Na⁺ balance (Fig. 2). However, intracellular acidification may reduce Cl⁻/HCO₃⁻ exchange contributing to the accelerated chloride loss immediately after exposure (Fig. 2).

After 24h of ammonia exposure, rainbow trout alevins showed a net cationic loss of 18 μmolg⁻¹ (4 μmol g⁻¹ Na⁺ + 14 μmolg⁻¹ K⁺, Fig. 2), which exceeded the measured anionic loss (8 μmol g⁻¹ Cl⁻) and cation gain (4 μmolg⁻¹ NH₄⁺); i.e. a net anionic loss (or cationic gain) of 6 μmol g⁻¹ would be required for electroneutrality. A similar trend was observed in Fig. 1, although the differences were greater, principally on account of greater K⁺ losses. A possible explanation is a gain of hydrogen ions via H⁺/NH₄⁺ exchange (Wilson and Taylor, 1992), though the H⁺ concentration of the external medium was low and the proton pump in gills is directed outwards (Avella and Bornancin, 1989). Another possibility that may partly account for the deficit is loss of bicarbonate in exchange for external Cl⁻ (Maetz and Garcia-Romeu, 1964; McDonald et al. 1989), i.e. Cl⁻ influx should be operating at its maximum rate since the external Cl⁻ concentration was at a saturating level of 15.8 mmoll⁻¹ (Williams and Eddy, 1986) and cellular alkalisation, together with increased hydration rates of metabolic CO₂ to HCO₃⁻, could provide an additional source of HCO₃⁻. Immediately after exposure, Cl⁻ loss accelerated prior to recovery, possibly because exposure to water of relatively low Cl⁻ content (0.3mmol l⁻¹) would significantly decrease Cl⁻ influx but might leave Cl⁻ efflux unaltered. The Cl⁻ loss during exposure might have been greater had another ammonium salt been used.

After 24h of ammonia exposure, body K⁺ values of rainbow trout alevins had decreased by some 14 μmol g⁻¹, but recovery to normal values of about 70 μmol g⁻¹ was slow and incomplete even after a week in ammonia-free water (Fig. 2). These trends are reflected in Table 5, which shows a high rate of K⁺ loss during ammonia exposure with low rates of uptake during recovery. This could be accounted for by the low levels of K⁺ in the water and its relatively slow uptake from the water compared with uptake of Na⁺ and Cl⁻ (Eddy, 1985). Since alevins were not feeding, dietary K⁺ was unavailable. The build up of K⁺ in the closed system during ammonia exposure may have reduced further K⁺ loss (Table 6).

The mechanisms of K⁺ loss from alevins during ammonia exposure are unclear, but may be connected with NH₄⁺ entry to cells from the extracellular fluid via Na⁺/K⁺(NH₄⁺)-ATPase and subsequent deprotonation of NH₄⁺ (Evans et al. 1989; Lin and Randall, 1991), which may induce loss of cellular K⁺. Also exercise caused increased muscle ammonia in lemon sole (Wright et al. 1988) and rainbow trout (Dobson and Hochachka, 1987) as well as loss of muscle cell K⁺ to the plasma in rainbow trout (Nielsen and Lykkeboe, 1992). Thus, during ammonia exposure, the primary source of K⁺ loss from alevins would appear to be from the intracellular compartment to the blood plasma, then to the exterior via body surface epithelia or paracellular routes.

Increased plasma ammonia levels occurred within an hour in freshwater rainbow trout exposed to 21.6 μmol l⁻¹ NH₃ (995 μmol l⁻¹ total ammonia, pH7.85, 15°C), but there were no changes in plasma Na⁺, K⁺, Cl⁻, Ca²⁺ or Mg²⁺ levels (Wilson and Taylor, 1992). However, in our experiments at a comparable NH₃ level (Fig. 1), the total ammonia concentration in the water was about ten times higher because of lower pH and temperature. Therefore, the disturbed ionic regulatory patterns observed in our experiments may be due to both NH₃ and NH₄⁺. Meade (1985), Erickson (1985) and WHO (1986) reviewed several studies where NH₄⁺ toxicity has been implicated. Further evidence for the role of NH₄⁺ in ammonia toxicity was demonstrated by Twitchen and Eddy (1993), who showed that, at approximately 30 μmol l⁻¹ NH₃, juvenile rainbow trout lost 19% of their body Na⁺ in 24h at pH7 compared with only 5% at pH8, because the NH₄⁺ levels were ten times higher at pH7.

Lloyd and Orr (1969) observed no mass changes in juvenile rainbow trout exposed to ammonia, suggesting that the increased urine flow was balanced by increased branchial water permeability. Net fluxes of Na⁺ (Tables 4 and 5), calculated from changes in body content, were similar to the values for rainbow trout alevins obtained from radio tracer studies in which unidirectional Na⁺ influx and efflux where determined (Twitchen and Eddy, 1993). Such comparisons indicate an additional effect, i.e. a net loss of body Na⁺, K⁺ and Cl⁻, some of which may be urinary and the remainder from the body surface, again without mass change.

These findings have important implications for alevins exposed to intermittent or cycling ammonia levels. Body ammonia levels rise and fall reasonably rapidly in response to external ammonia, but Na⁺ levels show a delay of about 4–5h before changes are detectable upon exposure and immediately after exposure, although both Na⁺ and ammonia levels recover simultaneously (Fig. 2). Far more significant is the fall in body K⁺ values, which decreased by more than 20%, then showed a very slow recovery which was incomplete even after a week. A second ammonia exposure during the recovery period could be extremely serious for alevins.

This project was financed by The Department of The Environment, London.