Accumulation of ammonia in the body and NH₃ volatilization from alkaline regions of the body surface during ammonia loading and exposure to air in the weather loach Misgurnus anguillicaudatus

The weather loach Misgurnus anguillicaudatus is a small freshwater fish that lives in rice fields. It is capable of using its intestine as an additional respiratory organ (McMahon and Burggren, 1987) to breathe air. It can gulp air, pass it through the digestive tract and release it from the vent. This enables it to live in oxygen-poor waters or to bury itself in mud during long-lasting droughts in the summer. When there is a lack of water, it will bury and position itself in the soft mud such that the mouth has access to air through a small opening, the fish gulping air occasionally and respiring through its intestine. By doing so, this loach can survive for more than a month until the drought is over.

Chew et al. (2001) reported that M. anguillicaudatus was capable of partial amino acid catabolism leading to a slight accumulation of alanine in the first 24 h of aerial exposure. Thus, it is possible that this loach is able to use certain amino acids as an energy source to sustain muscular activities on land, as has been suggested for the mudskipper Periophthalmodon schlosseri (Ip et al., 1993, 2001a,b). After 24 h of aerial exposure, glutamine accumulated in the muscle and liver of M. anguillicaudatus (Chew et al., 2001). Simultaneously, ammonia accumulated to very high levels in all the tissues examined. The present study was undertaken to determine whether M. anguillicaudatus could excrete NH₃ by volatilization during air exposure. Attempts were made to evaluate the roles of the skin and the digestive tract in NH₃ volatilization. The effects of placing M. anguillicaudatus in concentrated ammonia solutions on ammonia volatilization from the gut and on tissue ammonia levels were also determined.

Mature Misgurnus anguillicaudatus (8–15 g) were obtained from the Yuen Long wet-market in Hong Kong and transferred to Singapore by air. They were kept in dechlorinated tap water at 25°C in the laboratory. No attempt was made to separate the sexes. Fish were acclimated to laboratory conditions for at least 48 h before experimentation. During this period, half the water was changed daily and the fish were fed with commercial fish food. Fish were fasted for 24 h prior to experiments and were not fed during the experimental period.

Fish were exposed to 30 mmol l^–1 NH₄Cl in 20 vols (w/v) of water, adjusted to pH 7.2 using NaOH, at 25°C for various periods (12, 24 or 48 h). Those collected at the beginning of the experiment and those exposed to dechlorinated tap water (pH 7.2) for 48 h served as controls.

Fish exposed to terrestrial conditions were placed in individual plastic boxes (20.5 cm×14.5 cm×6.0 cm) containing a thin film of dechlorinated tap water (50 ml). The boxes were, in turn, placed in a fibreglass humidity chamber for the entire period of aerial exposure.

Fish were kept individually in conical flasks containing 100 ml water or NH₄Cl (10 mmol l^–1 or 20 mmol l^–1) solution. The flasks were tightly sealed such that they were both airtight and watertight, except for the air inlet and outlet. Air was bubbled through the water in the flask at a constant rate of 20 l h^–1. The flask was connected to two test tubes. The first test tube served as a water trap. The second, containing 3 ml of freshly prepared 0.5 mmol l^–1 H₂SO₄, served to trap any ammonia gas present in the air stream, which was bubbled through the acid. At the end of each 24 h period, the acid in the trap was collected and analysed for ammonia as described above. Blank control experiments were also run in which animals were not present in the flasks to account for volatilization of ammonia from the bathing solutions. In one experiment, fish were placed in small net cages before being submerged in the flask to determine whether preventing the fish from gulping air at the water surface would affect ammonia volatilization.

Experiments in which fish were air-exposed were performed at 20, 25 and 30°C. Fish were kept initially as described above in conical flasks containing 100 ml of dechlorinated water. After 24 h, the acid in the trap was analysed for ammonia content. The water in the flask was collected for ammonia assays, and the flask was rinsed. The same fish was then immediately exposed to terrestrial conditions for 24 h in the same apparatus, but with only 1 ml of dechlorinated water present in the flask. At the end of the second 24 h period, the acid in the trap was collected for ammonia assay and totally replenished. The small volume of water inside the flask was collected, and the flask was rinsed and refilled with 1 ml of fresh water. Ammonia production under terrestrial conditions was then monitored for 72 h. To evaluate the effects of a larger volume of water on NH₃ volatilization, an experiment was performed at 25°C with 3 ml, instead of 1 ml, of water in the flask.

Specimens were anaesthetised in 0.02 % 3-aminobenzoic acid ethyl ester (MS222; Sigma, USA). The tail was cut, and arterial blood was collected in heparinised capillary tubes. Blood variables (pH, P_CO2 and P_O2) were determined using an IL 1306 blood gas analyzer.

To collect muscle and liver tissue, specimens were killed by a blow to the head. The lateral muscle and liver were excised and immediately freeze-clamped in liquid nitrogen with pre-cooled aluminium tongs (Faupel et al., 1972). The entire procedure was carried out within 1 min. Samples were stored at –80°C until analysis.

The frozen samples were powdered in liquid nitrogen, weighed and homogenized three times in 5 vols of 6 % trichloroacetic acid (TCA) using an Ika-werk Staufen Ultra-Turrax homogenizer (Janke & Kunkel, Germany) at a maximum speed of 24 000 revs min^–1 for 20 s each, with 10 s intervals. The homogenate was centrifuged at 10 000 g at 4°C for 15 min in a Beckman J2-21M/E refrigerated centrifuge (Beckman Instruments Inc., USA).

A second group of specimens was used for the collection of blood samples. The tail of the fish was cut, and blood exuding from the severed caudal artery was collected in heparinized (sodium heparin) micro-haematocrit capillary tubes. The blood samples were centrifuged at 5000 g at 4°C for 5 min. The resulting plasma was collected and deproteinised immediately by adding an equal volume of 6 % TCA, followed by centrifugation at 10 000 g for 10 min. The supernatant fluid was kept at –80°C until analysis.

For ammonia and urea analyses, the deproteinized samples were neutralised to pH 5.5–6.0 with 2 mol l^–1 KHCO₃. Ammonia content was determined by the method of Kun and Kearney (1974). The change in absorbance at 340 nm (25°C) was monitored using a Shimadzu UV-160A spectrophotometer. Freshly prepared ammonium chloride (Merck) was used as a standard.

The urea content in 0.2 ml of deproteinised sample was analysed colorimetrically according to the method of Jow et al. (1999). The difference in absorbance of the sample in the presence and absence of urease was used as an estimate of the urea content of the sample.

To measure free fatty acid (FAA) levels, the deproteinised samples were diluted with an equal volume of 0.2 mol l^–1 lithium citrate buffer (Pierce, USA), and the pH was adjusted to 2.2 with 4 mol l^–1 lithium hydroxide. FAA levels were measured using a Shimadzu LC-6A amino acid analysis system with a Shim-pack ISC-07/S1504 Li type column. Analytical-grade FAA standards purchased from Sigma Chemical Co. served as references. Results were expressed as μmol g^–1 wet mass tissue or μmol ml^–1 plasma.

Specimens were exposed to 200 ml of 30 mmol l^–1 NH₄Cl solution (pH 7.2) at 25°C for 48 h. The external medium was changed daily. At 24 h and 48 h, a water sample was collected and analysed for urea content. Preliminary studies revealed that urea excretion rate was linear for at least 24 h. Urea content was analysed colorimetrically as described above. The rate of urea excretion was expressed as μmol day^–1 g^–1 fish.

The pH of water samples and the surface of the skin of M. anguillicaudatus was monitored using a PHR-146 micro combination pH electrode (Lazar Research Laboratories) and an Orion 720A pH meter. Preliminary results indicated that the pH over different regions of the body differed by less than 0.1 unit. Subsequently, it was decided to report the pH obtained from the side of the fish just above the pectoral fins. The specimen was then killed, and the digestive tract was dissected out. The digestive tract was divided into three regions: anterior, middle and posterior. Each region was cut open longitudinally, and the pH of the mucosal surface was measured using the micro combination pH electrode.

Results are presented as means ± s.e.m. Student’s t-test and one-way analysis of variance (ANOVA) followed by Student–Neuman–Keul’s multiple range test were used to compare differences between means where applicable. Differences at P<0.05 were regarded as statistically significant.

Exposure to 30 mmol l^–1 NH₄Cl at pH 7.2 did not affect blood pH, P_O2 or P_CO2, but aerial exposure resulted in a significant increase in blood pH (Table 1). Ammonia built up progressively in the muscle and liver (Fig. 1A) following ammonia exposure. Ammonia content in the muscle and liver was 18.9 and 17.5 μmol g^–1, respectively, after a 48 h exposure. In comparison, the plasma ammonia level rose abruptly in the first 12 h of ammonia exposure and reached 4.2 μmol ml^–1 by 48 h (Fig. 1A). There were no increases in urea level in the muscle, liver or plasma (Fig. 1B). The urea excretion rate of the specimen kept under control condition was 0.75±0.12 μmol day^–1 g^–1 (N=5), which was unaffected by exposure to 30 mmol l^–1 NH₄Cl.

The total free amino acid (TFAA) concentrations in the muscle (Table 2), liver (Table 3) and plasma (Table 4) were unaffected by ammonia exposure. However, there were minor changes in the levels of some free amino acids in these tissues. In the muscle, there was a transient increase in alanine level from 1.3 to 5.2 μmol g^–1 after 12 h of ammonia exposure, but it decreased thereafter. This was accompanied by transient increases in the levels of some essential amino acids. Ammonia exposure did not affect glutamine levels in the muscle, liver or plasma, but led to an increase in the glutamate level in the liver and a decrease in the glutamate level in the plasma.

With 20 mmol l^–1 NH₄Cl in the external medium, the amount of ammonia recovered from the acid trap in the presence of M. anguillicaudatus after a 24 h period was significantly higher than that of the blank without a fish (Fig. 2A). However, this difference disappeared when the fish was prevented from accessing air (data not shown). With a fish in the medium, the pH recorded in the presence of 20 mmol l^–1 NH₄Cl after 24 h was significantly lower than that of the blank without fish (Fig. 2B). For fish exposed to 30 mmol l^–1 NH₄Cl, the mucosal surface pH of the anterior portion of the digestive tract was significantly higher (more alkaline) than that of the control (Table 5); however, the pH values of the middle and posterior regions of the intestine were unaffected by ammonia exposure (Table 5). Air exposure caused a larger increase in pH of the anterior intestine than ammonia exposure in water. The pH of the skin surface during aerial exposure was significantly higher than that of submerged fish (Table 5).

Aerial exposure led to a marked decrease in the rate of ammonia excretion (Fig. 3). At the same time, there was a progressive increase in the amount of NH₃ volatilized (Fig. 4). The pH of the film of water underlying the fish in air increased from an initial value of 7.2 to between 7.9 and 8.1. Increasing the environmental temperature to 30°C increased the rate of NH₃ volatilization by M. anguillicaudatus exposed to terrestrial conditions (Table 6).

M. anguillicaudatus inhabits muddy swamps and ponds and rice fields subject to periodic drying and has to solve problems of ammonia toxicity under two situations. First, to survive drought, M. anguillicaudatus burrows into the mud and breathes air through a hole in the surface of the mud. In the absence of water, ammonia excretion may be impeded, leading to the accumulation of ammonia in its tissues. Second, many farmers add fertilizers to the rice paddies, which may release ammonium salts into the environment and expose the loach to high levels of environmental ammonia.

It is well known that the primary route of ammonia excretion in fish is through the gills (Wood, 1993; Wilkie, 1997). The principal mechanism of ammonia excretion across the gill epithelia is the non-ionic diffusion of NH₃ down a favourable transbranchial gradient of NH₃ from blood to water (Randall and Wright, 1987). An increase in external NH₃ concentration, if large enough, will result in the reversal of the blood-to-water NH₃ gradient, leading to net NH₃ influx, an inhibition of excretion of endogenous ammonia and a rise in the total plasma ammonia level in the body (Wood, 1993; Wilkie, 1997).

Although the ammonia level in the plasma of M. anguillicaudatus in the submerged condition falls within the range of those of other facultative air-breathers, it was unexpectedly high in ammonia-exposed (4.2 μmol ml^–1) and air-exposed (5.09 μmol ml^–1) individuals (Chew et al., 2001) after 48 h of exposure. The ammonia levels in the plasma of Amphipnous euchia, Clarias batrachus, Heteropneustes fossilis, Anabas testudineus and Channa punctatus range from 0.43 to 0.89 μmol ml^–1 (Saha and Ratha, 1989). The plasma of the tilapia Oreochromis nilotica (in ammonia-loading conditions), the climbing gourami Anabas scandens and the snakehead Channa gachua (in terrestrial conditions) contained 1.0, 1.6 and 1.6 μmol ml^–1 ammonia, respectively (Randall et al., 1989; Ramaswamy and Reddy, 1983). To our knowledge, no other fish have been reported to exhibit such a high ammonia concentration in the plasma under ammonia-loading or terrestrial conditions.

Tissue ammonia levels increased markedly in M. anguillicaudatus exposed to NH₄Cl in water or in air. The levels in the muscle and liver reached 18.9 μmol g^–1 and 17.5 μmol g^–1, respectively, following NH₄Cl exposure, and 11.08 and 14.52 μmol g^–1, respectively, after 48 h of aerial exposure (Chew et al., 2001). In the mudskipper Periophthalmodon schlosseri, 3.46 and 1.64 μmol g^–1 ammonia were accumulated in the liver and muscle, respectively, after 24 h of aerial exposure (Ip et al., 1993) and 7.24 μmol g^–1 and 14.8 μmol g^–1, respectively, after 6 days of exposure to 100 mmol l^–1 NH₄Cl (Randall et al., 1999). In the portunid crab Necora puber, the muscle ammonia content was 10.42 μmol g^–1 after 12 h of aerial exposure (Durand and Regnault, 1998). Clearly, M. anguillicaudatus can tolerate much higher levels of ammonia in its tissues and organs than can other animals, and future research on these ammonia-tolerant processes may increase our understanding of the cellular mechanisms of ammonia toxicity. For most vertebrates, such high levels of ammonia would disturb intracellular pH and membrane integrity (Campbell, 1991), leading to convulsions and death. Working on mammals, Marcaida et al. (1992) proposed that ammonia toxicity was due to the activation of N-methyl-d-aspartate (NMDA)-type glutamate receptors, leading to elevated intracellular Ca²⁺ concentrations and cell death. Perhaps M. anguillicaudatus has modified NMDA receptors.

Levels of free amino acids (FAAs) were variable in both the ammonia-exposed and control groups. The fish were not fed during the experiment, and this may explain the significant increases in levels of a number of FAAs observed after 48 h of fasting (Table 2). This somewhat obscures the FAA response to ammonia exposure. However, it is clear that there was a transient increase in alanine level in the muscle after 12 h of exposure to 30 mmol l^–1 NH₄Cl. M. anguillicaudatus might have resorted to partial amino acid catabolism at the onset of exposure to NH₄Cl to slow down the rate of internal ammonia production. Ip et al. (2001a,b) proposed partial amino acid catabolism as a strategy for fish to reduce the rate of production of endogenous ammonia during aerial exposure. The amino groups of certain amino acids are transferred directly or indirectly to pyruvate to form alanine. The resulting carbon chain enters the Krebs cycle and is partially oxidised to malate, which replenishes pyruvate through the reaction catalysed by malic enzyme. In this way, ATP can be derived from amino acids without producing ammonia.

Glutamine has been found to play a role in ammonia detoxification in fish in response to high environmental ammonia concentrations (Arillo et al., 1981; Dabrowska and Wlasow, 1986; Mommsen and Walsh, 1992; Peng et al., 1998). In this regard, M. anguillicaudatus is atypical, for although glutamine accumulated in its muscle after 48 h of aerial exposure, accumulation of glutamine did not occur in muscle and liver when specimens were exposed to ammonia. Hence, even if the mechanisms to detoxify endogenous ammonia were present, they were not utilised when the fish were confronted with an influx of exogenous ammonia.

The levels of urea in various tissues and the urea excretion rate of M. anguillicaudatus exposed to ammonia remained unchanged during ammonia loading. The activities of glutamine synthetase and ornithine–urea cycle enzymes were too low to render them effective in detoxifying ammonia (Chew et al., 2001). The production of urea as a strategy to detoxify ammonia in response to aerial exposure or ammonia loading does not appear to be a common phenomenon in teleosts (Ip et al., 2001a). This is probably because glutamine and urea production are energy-expensive processes, involving 4 and 5 moles of ATP (or equivalent), respectively, per mole of product formed (Ip et al., 2001a). It would appear that M. anguillicaudatus simply tolerated high tissue ammonia levels. In addition, it lowered the pH of the medium during ammonia exposure. This would reduce the NH₃ concentration in the external medium and decrease the rate of ammonia influx.

Aerial exposure resulted in a marked reduction in the rate of ammonia excretion. Presumably, ammonia is excreted across the gills into the water, and this is inhibited during air exposure. A small amount of ammonia is volatilized into the air when the animal is submerged (see Fig. 4). The amount excreted by volatilization increases when ammonia is added to the water and during aerial exposure. NH₃ volatilization has also been observed in the mangrove killifish Rivulus marmoratus in response to prolonged air exposure (Frick and Wright, 2002). M. anguillicaudatus can breathe air using its intestine (McMahon and Burggren, 1987), gulping air with its mouth and blowing gas out of the vent. Our results indicate that M. anguillicaudatus was capable of excreting some ammonia in gaseous form via the gut during ammonia-loading. A small quantity of ammonia was volatilized from the NH₄Cl solution in the absence of a fish. This is understandable because a small fraction (0.6 %) of ammonia was present as NH₃ even at pH 7, and it would be carried over by the air current into the acid trap. The fraction of NH₃ present in solution decreases with decreasing pH. Since there was a significant reduction in the pH of the external medium in the presence of a fish, the amount of ammonia volatilized should theoretically be smaller if it was due only to volatilization from the water. However, that was not the case: volatilization increased in the presence of the fish. Furthermore, M. anguillicaudatus stopped volatilizing NH₃ when it was prevented from accessing the water surface to gulp air. Hence, it can be concluded that the excess NH₃ volatilized represented ammonia excreted from the fish into the air passing through its gut and blown out of the vent. If it had been excreted across the skin into the water, it would have appeared even when the fish was forcibly prevented from breathing air.

In M. anguillicaudatus, the anterior portion of the digestive tract became more alkaline in ammonia-loaded and air-exposed fish. Alkalization would raise the fraction of NH₃ in the ammonia excreted into that region of the digestive tract, and this would augment NH₃ volatilization. When M. anguillicaudatus was exposed to terrestrial conditions, the surface of the skin also became significantly more alkaline, and the pH of the water underneath the fish was significantly higher than that of the submerged control. Thus, the skin of M. anguillicaudatus may also be a site of NH₃ volatilization during aerial exposure. Similarly, Wieser (1972b) suggested that, in isopods, ammonia diffuses into the film of water retained between the pleopods and is volatilized as a result of alkalization of the solution. Some terrestrial crabs have also been reported to volatilize NH₃ by alkalization of urine (De Vries and Wolcott, 1993; Varley and Greenaway, 1994). Although the skin and digestive tract may both be involved in NH₃ volatilization when the fish is moving on land, excretion of ammonia through the skin would be ineffective when the fish burrows into the mud. It seems probable that, when the fish is surrounded by mud, the digestive tract might become the only avenue for NH₃ volatilization.

For certain isopods (Wieser, 1972a) and for M. anguillicaudatus exposed to terrestrial conditions, the rate of NH₃ volatilization increases with temperature. Although the amount of ammonia excreted into the water remained the same at different temperatures, the increase in the rate of NH₃ volatilization led to increases in total ammonia excretion at higher temperatures (Table 6). The observed effects are probably related to the direct effect of temperature on the process of volatilization.

The progressive increase in the amount of NH₃ volatilized by M. anguillicaudatus over a 3-day period was apparently related to the increased internal ammonia levels. Besides building up ammonia to a very high level in the plasma during aerial exposure, the blood pH simultaneously became more alkaline. This would lead to a high level of NH₃ in the blood. The high P_NH3 gradient would facilitate the efflux of ammonia onto the body surface. Thus, the increase in ammonia volatilization seen during periods of elevated ammonia level in the water and aerial exposure are related to the alkalization of the gut and blood and the increase in total ammonia concentrations in the body, all raising NH₃ levels and, therefore, ammonia volatilization. This is in agreement with the observation of Wright and O’Donnell (1993) on the isopod Porcellio scaber that periodic volatilization of NH₃ was correlated with a large increase in ammonia levels in the haemolymph and pleon fluid.

The loach Misgurnus anguillicaudatus is unusual in that it tolerates very high ammonia levels in its tissues, but in turn these high levels, together with alkaline body surfaces, permit a significant ammonia excretion by volatilization during aerial exposure.