The marble goby Oxyeleotris marmoratus activates hepatic glutamine synthetase and detoxifies ammonia to glutamine during air exposure

The catabolism of amino acids, proteins, nucleotides and other related substances results in the formation of end-products that must be excreted by the animal (Mommsen and Walsh, 1991). Fishes excrete mainly ammonia, with some rare examples that also excrete urea and trimethylamine oxide as nitrogenous end-products (Goldstein and Forster, 1970; Campbell and Anderson, 1991; Anderson, 1995).

The marble goby Oxyeleotris marmoratus is a facultative air-breather capable of surviving under terrestrial conditions for up to 7 days (Choo, 1974). In its natural environment, O. marmoratus may encounter an absence of water during voluntary emergence or habitat desiccation (Choo, 1974). On land, O. marmoratus may face the problem of nitrogenous end-products accumulating to toxic levels in the body. This is due to the absence of a water current to allow the excretion of ammonia through the branchial epithelial surfaces. Thus, attempts were made in this study to determine the amount of ammonia excreted by O. marmoratus under submerged and terrestrial conditions.

The accumulation of ammonia to lethal concentrations in the body may be partially prevented by the conversion of some ammonia to urea. This strategy is adopted by certain species of fish exposed to air (Gordon et al., 1969, 1978; Ip et al., 1993), to high levels of ambient ammonia (Olson and Fromm, 1971; Saha and Ratha, 1987; Walsh et al., 1990), to confinement and/or crowding (Walsh et al., 1994; Walsh and Milligan, 1995) or to alkaline environments (Randall et al., 1989; Wood et al. 1989; Walsh et al., 1993). Urea production is metabolically expensive, but urea is less toxic than ammonia. Thus, the possibility of O. marmoratus producing more urea during air exposure was explored in the present study. Attempts were made to verify the presence of all the components of the urea cycle in the liver of this fish.

Ammonia detoxification may be achieved by other means. Ip et al. (1993) suggested that partial amino acid catabolism, leading to the formation of alanine, may serve to lower the amount of ammonia produced by the mudskipper Periophthalmodon schlosseri during air exposure. In addition, glutamine has been found to accumulate in the brain of various teleosts exposed to toxic levels of environmental ammonia (Levi et al., 1974; Arillo et al., 1981; Dabrowska and Wlasow, 1986; Peng et al., 1998). Therefore, we also examined the free amino acid levels and the activities of glutamate dehydrogenase and glutamine synthetase in various tissues of O. marmoratus. It is hypothesized that O. marmoratus must have adopted one or more of these strategies to allow it to maintain low levels of ammonia in its tissues during air exposure.

Oxyeleotris marmoratus Bleeker (500–800 g body mass) were purchased locally from fish farms at Jalan Kayu, Singapore. Groups of no more than 20 fish were kept submerged in fibreglass tanks (72 cm×55 cm×56 cm; length×width×height) containing 130 l aerated dechlorinated tap water at 25 °C. No attempt was made to separate the sexes. The fish were allowed at least 48 h of acclimation prior to use in experiments. During this period, half the water was changed daily and the fish were not fed. The ammonia concentration in the water was always below 0.02 mmol l⁻¹.

Fish submerged in aerated dechlorinated tap water at 25 °C were regarded as controls. Fish exposed to terrestrial conditions at 25 °C were placed in individual plastic tanks (28 cm×18 cm×14 cm; length×width×height) containing a thin layer of tap water (100 ml). These plastic tanks were kept within a moisture-saturated fibreglass tank for the entire period (6 h, 24 h, 48 h or 72 h) of air exposure.

Fish submerged in aerated dechlorinated tap water were anaesthetized with 3-aminobenzoic acid ethyl ester (MS 222) (Sigma Chemical Co.) neutralized to pH 7.0 at a final concentration (w/v) of 0.1 % for 10 min; fish exposed to terrestrial conditions were anaesthetized by exposing the fish to diethylether vapour for 10 min. They were then individually killed with a blow on the head. The lateral muscle and liver were excised and immediately freeze-clamped in liquid N₂ using pre-cooled aluminium tongs (Faupel et al., 1972). The whole procedure was carried out within 1 min. Samples were then stored at −80 °C until analyzed.

The frozen sample was weighed, powdered under liquid N₂ and homogenized three times in 5 vols of 5 % trichloroacetic acid using an Ultra-Turrax homogenizer at maximum speed (24 000 revs min⁻¹) for 20 s each time with 10 s beween bursts. The homogenate was then centrifuged at 10 000 g for 10 min in a Beckman J2-21M/E refrigerated centrifuge. The supernatant obtained was used for the analysis of ammonia, urea and free amino acid (FAA) contents.

Plasma samples were collected from another group of anaesthetized fish. The tail of the anaesthetized fish was cut, and the blood exuding from the severed caudal artery was collected in heparinized (Na⁺) capillary tubes and centrifuged at 5000 g for 5 min at 4 °C. The plasma obtained was immediately deproteinized by adding an equal volume of 6 % trichloroacetic acid, followed by centrifugation at 10 000 g for 10 min. The supernatant fluid was kept at −80 °C until analyzed.

The deproteinized sample was neutralized with 2 mol l⁻¹ KHCO₃, and the ammonia content was determined according to the method of Kun and Kearney (1974). The change in absorbance at 25 °C and 340 nm was monitored using a Shimadzu UV-160A spectrophotometer. Freshly prepared NH₄Cl solution was used as the standard.

Urea content in 0.2 ml of deproteinized sample was assayed using a Sigma urea assay kit (procedure 535). A blank assay was also performed after 0.2 ml of the sample had been incubated with 0.2 ml of 20 mmol l⁻¹ imidazole buffer (pH 7.2) containing 2 units of urease for 15 min at 30 °C. Urea obtained from Sigma Chemical Co. (USA) was used as the standard.

For analysis of FAA content, the deproteinized sample was diluted with an equal volume of 0.2 mol l⁻¹ lithium citrate buffer (pH 2.2) and adjusted to pH 2.2 with 4 mol l⁻¹ LiOH. Determination of FAA levels was performed using a Shimadzu LC-6A amino acid analysis system with a Shim-pack ISC-07/S1504 Li column. Analytical grade amino acid standard purchased from Sigma Chemical Co. (USA) served as a reference.

Ammonia, urea and FAA contents were expressed as μmol g⁻¹ wet mass for muscle and liver and as μmol ml⁻¹ for plasma.

Specimens were submerged individually in plastic aquarium tanks containing 4 l of aerated dechlorinated tap water at 25 °C. After 24 h, a small volume of water was sampled for ammonia and urea analyses.

Subsequently, the same individuals were exposed to terrestrial conditions in plastic tanks containing a thin layer (100 ml) of tap water. After 24 h, the fish were sprayed thoroughly with de-ionized water. The water collected was analyzed for ammonia and urea. The fish were then transferred to other tanks, and the experiment was repeated twice. After 72 h of air exposure, the fish were submerged again to study the rates of ammonia and urea excretion during recovery.

Ammonia was determined colorimetrically according to the method of Anderson and Little (1986). Freshly prepared NH₄Cl solution was used as the standard. Urea was analyzed using a Sigma urea assay kit (Procedure 535) as described above. Ammonia and urea excretion rates were expressed as μmol day⁻¹ g⁻¹ wet mass of the fish.

Freshly excised liver samples were homogenized in 10 vols of mitochondrial extraction buffer (285 mmol l⁻¹ sucrose, 3 mmol l⁻¹ Tris–HCl and 3 mmol l⁻¹ EDTA, pH 7.2) using a Glas-Col tissue grinder (Terre Haute, USA) using one stroke at low speed. The homogenized sample was centrifuged at 600 g for 15 min. The resulting supernatant fluid was centrifuged three times at 10 000 g for 15 min. Between each spin, the supernatant fluid was discarded and replaced with fresh mitochondrial extraction buffer. After the last spin, the loose pellet was mixed with 1 ml of buffer containing 50 mmol l⁻¹ Hepes (pH 7.6), 50 mmol l⁻¹ KCl, 0.5 mmol l⁻¹ EDTA and 1 mmol l⁻¹ DL-dithiothreitol (DTT). This suspension was sonicated three times for 20 s each time with 10 s between bursts. The mitochondrial fraction was obtained by passing the suspension through a 10 ml Econo-Pac 10DG desalting column (Bio-Rad Laboratories Inc., CA, USA) equilibrated with the same buffer.

To obtain the cytosolic fraction, the freshly excised liver was homogenized with 5 vols of mitochondrial extraction buffer (see above) and centrifuged at 10 000 g for 15 min. The resulting supernatant was collected and passed through a 10 ml Econo-Pac 10DG desalting column equilibrated with Hepes buffer. Samples were immediately assayed for carbamoyl phosphate synthetase (CPS) II activity. CPS activity was detected using the method described by Korte et al. (1997). Preliminary studies showed that CPS activities in both the cytosolic and mitochondrial samples were low in the presence of NH₄⁺, but high when glutamine was present. The standard reaction mixture contained 50 mmol l⁻¹ Hepes (pH 7.6), 50 mmol l⁻¹ KCl, 5 mmol l⁻¹ EDTA, 20 mmol l⁻¹ ATP, 25 mmol l⁻¹ MgCl₂, 5 mmol l⁻¹ NaHCO₃, 22.5 mmol l⁻¹ phosphoenolpyruvate (PEP), 2 units of pyruvate kinase (PK), 74 MBq ml⁻¹ [¹⁴C]NaHCO₃, 20 mmol l⁻¹ glutamine and 0.1 ml of the sample in the presence or absence of 1.67 mmol l⁻¹N-acetylglutamate (AGA) or 1.67 mmol l⁻¹ AGA + 1.67 mmol l⁻¹ UTP. Radioactivity was measured using a Wallac 1414 liquid scintillation counter. CPS activity was expressed as μmol [¹⁴C]urea min⁻¹ g⁻¹ wet liver mass.

The activity of ornithine transcarbamoylase was determined according to Xiong and Anderson (1989). Ornithine transcarbamoylase activity was expressed as μmol citrulline min⁻¹ g⁻¹ wet liver mass.

Argininosuccinate synthetase + lyase activities were determined together assuming that both were present by measuring [¹⁴C]fumarate formation from [¹⁴C]aspartate following the method of Cao et al. (1991). Radioactivity was measured using a Wallac 1414 liquid scintillation counter. Argininosuccinate synthetase + lyase activity was expressed as μmol [¹⁴C]fumarate min⁻¹ g⁻¹ wet liver mass.

Arginase activity was assayed according to Casey and Anderson (1982). The urea formed was determined colorimetrically as described above. Arginase activity was expressed as μmol urea min⁻¹ g⁻¹ wet liver mass.

The frozen liver and muscle samples were homogenized three times in 5 vols of ice-cold buffer containing 50 mmol l⁻¹ imidazole (pH 7.4), 50 mmol l⁻¹ NaF, 3 mmol l⁻¹ EGTA, 3 mmol l⁻¹ EDTA and 0.5 mmol l⁻¹ phenylmethylsulphonyl-fluoride (PMSF) at maximum speed (24 000 revs min⁻¹) for 20 s each with 10 s between bursts. The resulting homogenate was centrifuged at 10 000 g for 15 min at 4 °C. The supernatant fluid was passed through a 10 ml Econo-Pac 10DG desalting column equilibrated with the buffer without PMSF. The resulting fluid was used for enzymatic analysis.

Enzyme activities were recorded at 25 °C and 340 nm unless otherwise stated. All chemicals and coupling enzymes were obtained from Sigma Chemical Company (MO, USA). Enzyme activities are expressed in μmol NADH min⁻¹ g⁻¹ wet tissue mass unless otherwise stated.

GDH activities were assayed in the aminating and deaminating directions according to the method of Peng et al. (1994). GDH activity in the deaminating direction was recorded at 492 nm and is expressed as μmol formazan min⁻¹ g⁻¹ wet tissue mass. ALT activity in the direction of alanine degradation was determined following the method of Peng et al. (1994). Glutamine synthetase activity was measured according to Shankar and Anderson (1985). Glutamine synthetase activity was measured at 500 nm and is expressed as μmol γ-glutamylhydroxymate min⁻¹ g⁻¹ wet tissue mass.

Results are presented as means ± S.E.M. One-way analysis of variance (ANOVA) followed by Student–Newman–Keuls multiple-range test was used to compare differences between means where applicable. Differences with P<0.05 were regarded as statistically significant.

Air exposure had no significant effect on the ammonia levels in the muscle of O. marmoratus (Table 1). In the liver and plasma, the ammonia level rose transiently to a significantly higher value after air exposure, but returned to the control (submerged) value thereafter. The rate of ammonia excretion by O. marmoratus during air exposure was significantly lower than that when the fish was submerged (Fig. 1). When the fish was resubmerged, this rate returned to the initial control value.

The urea content of the muscle of O. marmoratus rose significantly after 6 h of air exposure but returned to the control (submerged) level thereafter (Table 2). Air exposure did not affect the concentration of urea in the plasma. However, the urea content of some of the liver samples collected after 24 h, 48 h and 72 h of air exposure became undetectable. Air exposure also significantly lowered the rate of urea excretion by O. marmoratus (Fig. 2). Upon resubmergence, this rate returned to the initial control value.

Using glutamine as a substrate, CPS activity was detected in the hepatic mitochondrial and cytosolic fractions. CPS activity in the mitochondrial fraction was activated by AGA but this stimulated activity was unaffected by UTP (Table 3). In contrast, CPS activity in the cytosol was unaffected by AGA but was inhibited by UTP. Air exposure had no effect on the activity of CPS in either fraction. Ornithine transcarbamoylase, argininosuccinate synthetase + lyase and arginase activities were detected in the liver of O. marmoratus (Table 4). These activities were unaffected by air exposure, except for arginase from the cytosolic fraction, which exhibited significantly lower activity.

After 72 h of air exposure, the glutamine level in the muscle of O. marmoratus increased threefold (Table 5). Small but insignificant changes were detected for aspartate, phenylalanine and threonine (Table 5). The total free amino acid (TFAA) content of the muscle remained unchanged during air exposure. In the liver, significant increases in the contents of glutamine, histidine, isoleucine, leucine, serine, threonine and valine were observed after 24 h of air exposure (results not shown). However, by 72 h, these values had either returned to control values (glutamine, histidine, leucine and threonine) or fallen to levels lower than that of the control (isoleucine, serine and valine) (Table 5). In the blood, few changes in FAA content were detected and TFAA remained unchanged during air exposure (Table 5).

A time course study of the glutamine level in the muscle (Fig. 3) showed that glutamine rose to peak levels by 48 h of air exposure and levelled off thereafter. In contrast, the glutamine level of the liver reached its peak at approximately 24 h, when it was more than five times the control value (Fig. 3). Subsequently, at 48 h, it returned to the control level. There was no significant difference between GDH (amination) activities from the muscle of O. marmoratus exposed to submerged and terrestrial conditions (Table 6). No GDH activity in the deamination direction was detected in the muscle of O. marmoratus. There was a threefold increase in GDH (amination) activity in the liver of fish exposed to terrestrial conditions, while deamination levels remained unchanged. Hence, air exposure led to an increase in the amination:deamination ratio of GDH in the liver of O. marmoratus (Table 6).

ALT activities from both muscle and liver were unaffected by air exposure (Table 7). There was a significant increase in the hepatic glutamine synthetase activity of O. marmoratus after 48 h and 72 h of air exposure (Table 7; Fig. 4). The activity of hepatic glutamine synthetase remained low when O. marmoratus was exposed to terrestrial conditions for up to approximately 24 h, after which the activity increased approximately thirtyfold (Fig. 4). In contrast, no such effect was observed for muscle glutamine synthetase (Table 7).

In contrast to the amphibious mudskippers (Periophthalmus cantonensis and Boleophthalmus pectinirostris, Morii, 1979; Boleophthalmus boddaerti and Periophthalmodon schlosseri, Ip et al., 1993), the ammonia level in the muscle of O. marmoratus never increased during air exposure. In the liver and plasma of O. marmoratus, ammonia content rose only transiently after air exposure to 1.74 μmol g⁻¹ and 0.18 μmol ml⁻¹, respectively. These are very small increases compared with the values of 3.46 μmol g⁻¹ and 0.92 μmol ml⁻¹, respectively, in air-exposed P. schlosseri (Ip et al., 1993). Since the ammonia excretion rate in O. marmoratus exposed to air was one-fifth of that of the submerged controls, and since the ammonia excretion rate did not increase above the control value upon resubmersion, this fish must be capable of reducing its rates of protein and/or amino acid catabolism, or detoxifying ammonia to other less toxic products, or both, during air exposure.

Buttle et al. (1996) reported that the African catfish Clarias gariepinus excreted more ammonia during resubmersion only if the period of air exposure was short (<5 min). Between 5 and 180 min of air exposure, C. gariepinus resorted to decreasing the rate of ammonia production and converting ammonia to urea. The period of air exposure imposed on C. gariepinus was short, possibly because it was unable to survive on land like O. marmoratus for 72 h or more.

O. marmoratus excreted a large proportion of its nitrogenous waste as urea in the control submerged condition. The ratio of ammonia:urea excreted was approximately 5.1, which is lower than values (7.6–33.4) for the sockeye salmon Oncorhynchus nerka (Brett and Zala, 1975), for Blennius pholis (Davenport and Sayer, 1986) and for five other facultative air-breathers, Amphipnous cuchia, Clarias batrachus, Heteropneustes fossilis, Anabas testudineus and Channa punctatus (Saha and Ratha, 1989). Although O. marmoratus did not detoxify ammonia to urea during air exposure, all the five enzymes involved in the ornithine–urea cycle were present in the liver of O. marmoratus. CPS III, which is responsible for urea synthesis in the majority of fish species (Mommsen and Walsh, 1991), was detected only at low levels (approximately 5 nmol min⁻¹ g⁻¹) in this fish. CPS II, which does not require AGA for activity and is inhibited by UTP (Campbell and Anderson, 1991), was found in the cytosol.

Air exposure exhibited no significant effect on the activities of CPS III, ornithine transcarbamoylase and argininosuccinate synthetase + lyase from the liver of O. marmoratus. However, arginase activity in the cytosol was significantly lower at 72 h of air exposure. Incidentally, the rate of urea excretion in O. marmoratus exposed to the terrestrial condition decreased to one-fiftieth of the submerged value, leading to a higher ammonia:urea excretion ratio of 20. After all, it is metabolically expensive to produce urea. Four moles of ATP are required per mole of urea formed from two moles of ammonia (Mommsen and Walsh, 1991).

Superficially, it would appear that O. marmoratus suppressed both ammonia and urea production during air exposure. Taking the submerged ammonia excretion rate to be 0.09 μmol ammonia day⁻¹ g⁻¹ wet mass, and assuming that the rate dropped to one-fifth of this value during air exposure, approximately 0.072 μmol ammonia day⁻¹ g⁻¹ wet mass would theoretically have been accumulated in the tissues of this fish. Since ammonia levels in the muscle exposed to the submerged and terrestrial conditions were comparable, a 500 g fish would have reduced ammonia production by approximately 108 μmol in 72 h. The decrease in urea excretion for the entire 72 h period of air exposure was approximately 5.3×10⁻² μmol g⁻¹. Since urea content in the muscle was not affected by air exposure, the fish had decreased overall urea production by the same amount (5.3×10⁻² μmol g⁻¹) during this period. Again, a 500 g fish would have produced 26.5 μmol less urea. Two ammonia molecules are utilized during the formation of urea; hence, the fish would have reduced production by approximately 53 μmol of ammonia-equivalent during 72 h of air exposure. A 500 g O. marmoratus would have decreased its total ammonia production by 161 μmol (108+53 μmol) during this period.

However, the glutamine content of the muscle of O. marmoratus exposed to air increased threefold in 72 h. Glutamine may be produced from glutamate and NH₄⁺ catalyzed by glutamine synthetase in the muscle and/or liver. Glutamate may, in turn, be produced from α-ketoglutarate (αKG) and NH₄⁺ catalyzed by GDH, or from αKG and other amino acids catalyzed by various transaminases. In other words, the formation of one glutamine molecule allows the uptake of two ammonia molecules (Campbell, 1973). The hepatic glutamine content peaked at 24 h of air exposure, while that in the muscle peaked only after 48 h. This suggests that the glutamine formed in the liver was perhaps later shuttled to the muscle, which acted as a reservoir for glutamine accumulation. Consequently, the steady-state glutamine level in the plasma was maintained.

Given that the glutamine level rose by approximately 0.7 μmol g⁻¹ in the muscle after 72 h of air exposure, and that approximately 60 % of a 500 g O. marmoratus is muscle, approximately 210 μmol of glutamine would have been accumulated. Since two NH₄⁺ are needed to produce one glutamine molecule, 420 μmol of ammonia would have been detoxified in this manner. This would be more than enough to account for the apparent decrease of 161 μmol of ammonia produced during the entire period. Therefore, it was unlikely that amino acid catabolism had been reduced; instead, the ammonia produced was channelled towards the formation of glutamine. Protein or amino acid catabolism in O. marmoratus may, in fact, have been increased during air exposure because glutamine accumulated to levels far in excess of that needed to detoxify the ammonia produced during this period.

To the best of our knowledge, glutamine has been found to play a role in ammonia detoxification in fish only in response to high environmental ammonia concentrations (Levi et al., 1974; Arillo et al., 1981, Dabrowska and Wlasow, 1986; Mommsen and Walsh, 1992; Peng et al., 1998). Thus, O. marmoratus may represent the first known case in which a fish is able to detoxify internally generated ammonia in this way.

The increase in hepatic GDH (amination) activity in O. marmoratus exposed to the terrestrial condition did not lead to any increase in glutamate or alanine content, as in P. schlosseri (Ip et al., 1993), in its tissues. Under the same conditions, hepatic glutamine synthetase activity rose from 0.021 μmol min⁻¹ g⁻¹ to 0.62 μmol min⁻¹ g⁻¹. This was an increase of approximately thirtyfold. Hence, the glutamate produced must be channelled into the formation of glutamine. There was a good correlation between the peak of hepatic glutamine synthetase activity and the peak of glutamine accumulation in the muscle of the fish exposed to terrestrial condition.

In other cases where environmental ammonia is detoxified to glutamine (Levi et al., 1974; Arillo et al., 1981; Peng et al., 1998), neither cerebral nor hepatic glutamine synthetase activity was found to be significantly different from control values. This suggests that pre-existing glutamine synthetase is sufficient to account for the increased glutamine formation (Walsh et al., 1993; Wright et al., 1993). Although Peng et al. (1998) observed some increases in the specific activities of cerebral glutamine synthetase in two species of mudskipper in response to ammonia loading, those from the liver remained unchanged. In the gulf toadfish Opsanus beta kept under confinement/crowding conditions, hepatic glutamine synthetase activity increased several-fold, leading to an induction of ureotely (Walsh et al., 1994; Walsh and Milligan, 1995).

Thus, O. marmoratus appears to be the first known teleost that responds to air exposure by activating hepatic glutamine synthetase and detoxifying internally produced ammonia to glutamine.

The authors wish to thank Dr D. Randall for his helpful comments on this research project.