Metabolic and neuroendocrine effects on diurnal urea excretion in the mangrove killifish Rivulus marmoratus

The rate of nitrogenous waste excretion in teleosts is largely dependent on the rate of synthesis in the liver and the rate of transport across body surfaces, which is influenced by concentration gradients and membrane permeability. The majority of fish synthesize and excrete ammonia and urea continuously even when deprived of food(Smith and Thorpe, 1976; Kaushik and de Oliva-Teles,1985; Alsop and Wood,1997). In fed individuals, concentrations of nitrogen end products in plasma increase within hours after a meal, followed by a washout period typified by increased rates of excretion(Brett and Zala, 1975; Kaushik and Gomes, 1988; Kaushik and de Oliva-Teles,1985; Wicks and Randall,2002). In contrast to most teleosts, nitrogen excretion in the ammoniotelic Rivulus marmoratus demonstrates a most unusual pattern. Urea (J_urea) but not ammonia excretion(J_amm) in R. marmoratus undergoes a diurnal rhythmicity, with the highest rates of urea excretion occurring during mid-afternoon (Frick and Wright,2002a; Rodela and Wright,2006). Under the influence of circadian clocks, the phase of the J_urea cycle was entrained fairly quickly to photic zeitgebers and possibly feeding cues.

One of the aims of the present study was to determine if the diurnal J_urea pattern in R. marmoratus was related to diurnal changes in metabolism. The ammoniotelic gobiid fish Mugilogobius abei also has a distinct daily pattern of J_urea under light:dark cycles (Kajimura et al.,2002). The authors postulated that the rate of urea synthesis may vary in a similar diurnal pattern; however, rates of oxygen consumption were not measured. An alternative possibility is that the rhythmic pattern of J_urea in R. marmoratus is under the control of mechanisms affecting urea permeability across cell membranes.

Urea permeability across lipid membranes is significantly lower than that of ammonia (Collander, 1937; Galluci et al., 1971) and recent studies have shown the presence of urea transporters (UT) on the branchial epithelium of many fish species(Smith and Wright, 1999; Walsh et al., 2000; Walsh et al., 2001; Mistry et al., 2001; McDonald et al., 2002; McDonald et al., 2004). Pulsatile J_urea patterns in marine toadfish Opsanus beta are attributed to periodic activation of a facilitated branchial UT,rather than to changes in urea production pathways(Wood et al., 1997; Wood et al., 2003). Recent studies have revealed that the glucocorticoid hormone, cortisol and the monoamine, 5-hydroxytryptamine (5-HT; serotonin) may regulate pulse size and frequency, respectively (Wood et al.,1997; Wood et al.,1998; Wood et al.,2003; McDonald et al.,2004).

Similar control mechanisms may be regulating J_urea in R. marmoratus. Both 5-HT and cortisol are appealing candidates for neuroendocrine mediators of diurnal urea excretion as these hormones undergo daily rhythms under the influence of a light:dark photoperiod in other species. For instance, circulating cortisol levels reach peak concentrations at or near the onset of light, immediately prior to the initiation of daily locomotor activity in several fishes(Peter et al., 1978; White and Fletcher, 1984; Fivizzani et al., 1984). Daily fluctuations in brain 5-HT levels have been documented in Anguilla anguilla (van Veen et al.,1982) and O. mykiss(Zuanreiter et al., 1998).

The mangrove killifish R. marmoratus are remarkably hardy fish that tolerate a range of environmental extremes(Abel et al., 1987; King et al., 1989; Frick and Wright, 2002a; Frick and Wright, 2002b). In this study we hypothesized that the daily J_urea pattern in R. marmoratus is due to diurnal changes in metabolic rate, which may in turn alter blood to water urea gradients and ultimately, excretion rates. Tissue concentrations of urea and ammonia, urea excretion rates, as well as whole animal oxygen consumption rates were measured over time. A second hypothesis, that the rate of urea excretion is influenced by neuroendocrine messengers was also tested. We first tested the prediction that RU-486, a cortisol receptor antagonist, will cause a dose-dependent increase in urea,but not ammonia, excretion rates. Second, exposure to ketanserin, a 5-HT₂ receptor antagonist, will cause a dose-dependent decrease in urea excretion rates, but not ammonia excretion rates. Furthermore, we predicted that upon exposure to RU-486 or ketanserin, the amplitude of the diurnal J_urea pattern would be increased and decreased,respectively, based on previous studies in the toadfish(McDonald et al., 2004; McDonald and Walsh, 2004).

Mangrove killifish Rivulus marmoratus Poey were held at the Hagen Aqualab at the University of Guelph (Guelph, ON, Canada)(Frick and Wright, 2002a). The fish were held in an environmental chamber at 25°C on a photoperiod of 12 h:12 h light (L):dark (D). Mature fish (0.07-0.25 g) were kept in individual translucent containers in 60 ml of 15‰ artificial seawater (pH 8.1),which was changed biweekly. Artificial seawater was made from distilled water and marine salt (Instant Ocean™, Crystal Sea, Baltimore, MA, USA). Fish were fed Artermia salina nauplii every other day. To eliminate effects of recent feeding on nitrogen metabolism and excretion, fish were deprived of food 48 h prior to the initiation of an experiment.

Four series of experiments were conducted:

Series I: measurements of tissue urea and ammonia levels over 1.5 days.

Series II: quantification of whole animal oxygen consumption rates over a 3 day period.

Series III: detailed examination of hourly oxygen consumption rates over a 6 h period.

Series IV: hourly nitrogen excretion rates in fasted R. marmoratusover a 6 h period.

In order to determine if changes in the diurnal J_ureapattern were due to metabolic changes (i.e. changes in nitrogen production and storage levels), tissue urea and ammonia concentrations were measured. Fish were killed by flash freezing in liquid nitrogen every 3 h during a 1.5-day period. Samples were stored at -80°C for up to 2 months prior to analysis of urea and ammonia content.

Oxygen consumption rates were measured in R. marmoratus using closed respirometry techniques. A single fish was placed in approximately 60 ml of brackish water in the respiratory chamber. Fish were acclimated to the respirometry chambers 2 days prior to the start of the experiment (25°C). Additional blank chambers (not containing a fish) were run simultaneously to determine the rate of oxygen consumed by the oxygen probe. The background values were subtracted from the oxygen consumption values of the fish, but were minimal (<10.28% of average fish value). At the onset of the measurements, the chamber was sealed for 1 h, during which changes in oxygen levels were measured in mg l^-1. During the measurement, oxygen levels did not fall below 70% saturation in order to ensure that the values obtained were a true representation of normoxic metabolic status in the animal. At the end of the sampling period, the chambers were renewed with fresh aerated water (15‰) for a period of 5 h to allow maximum oxygen saturation levels before the start of the next sampling period. In Series II,experiments involved measuring the change in water oxygen for 1 h intervals every 5 h for 3 days. For Series III, experiments entailed measuring oxygen consumption for every hour between 12:00 h and 18:00 h. At the end of each series of experiments, the fish were removed from their chambers, blotted dry and weighed. Values are presented as μmol O₂ g^-1h^-1.

In order to determine correspondence between J_urea and oxygen consumption, J_urea was measured for each 1 h interval between 12:00 h and 18:00 h. Fish were held in individual containers with 30 ml of 15‰ artificial seawater (pH 8.1, 25°C, 12 h:12 h L:D)and water samples were collected every hour for a 1.5 day period. The water was changed after each sampling interval to prevent accumulation of nitrogenous wastes. Elimination of handling stress was accomplished through the use of double-walled containers, with an inner mesh container made in the exact same configuration as the plastic translucent holding container. Consequently, during water changes, the inner mesh container with the fish was removed and placed in a new outer chamber containing fresh brackish water. Fish were acclimated to the containers 2 days prior to the start of the experiment. Water samples were frozen at -20°C for up to 1 month and later analyzed for urea and ammonia content.

Three series of experiment were conducted:

Series V: the dose effects of ketanserin and RU-486 on nitrogen excretion.

Series VI: the effects of a single application of ketanserin and RU-486 on nitrogen excretion.

Series VII: the effects of ketanserin and RU-486 on diurnal J_urea.

Pharmacological agents were used to determine if the diurnal pattern of J_urea is controlled, in part, by either cortisol or 5-HT. Specifically, RU-486, a cortisol receptor inhibitor, and ketanserin, a 5-HT₂ receptor antagonist, were used. Initial experiments involved testing the effect of each drug on J_urea. Concentrations were chosen based on similar studies on other fish species(Brustein et al., 2003; Dasmahapatra and Lee, 1993; Sathiyaa and Vijayan, 2003). For ketanserin, three different treatment groups of R. marmoratuswere exposed to concentrations of 10, 30 or 50 μmol l^-1. Initially, identical concentrations of RU-486 and ketanserin were used;however, there was a 80% or higher mortality rate associated with these concentrations. Subsequently lower concentrations of RU-486 consisting of 1, 2 or 5 μmol l^-1 were used. Both receptor antagonists required an ethanol vehicle (0.0003% w/v), consequently control fish were exposed to identical concentrations of the vehicle in the absence of the antagonist. Final water concentrations of ethanol in did not exceed 0.83%. Each drug (and ethanol control) was added to 30 ml of water (15‰). Initial water samples were collected following the addition of the drugs and final water samples were taken 6 h later, and handled as described for Series IV. Due to the small size of the fish (<250 mg), it was not possible to measure the plasma concentration of 5-HT and cortisol.

Experiments involved following nitrogen excretion rates over time during and after a 6 h exposure to 30 μmol l^-1 ketanserin or 1 μmol l^-1 RU-486. Initial experiments involved measuring both J_urea and J_amm over 6 h intervals under control conditions for a 24 h period.

As previously described, the control treatment contained ethanol (final concentration less than 0.83%). A single dose of either 30 μmol l^-1 ketanserin or 1 μmol l^-1 RU-486 was administered at 12:00 h. Following a 6 h exposure to the receptor antagonists (12:00 h-18:00 h), freshwater (15‰) with ethanol (0.83%) was replaced in each chamber every 6 h period for the remainder of the experiment in order to follow the recovery phase. Water samples were taken every 6 h and were treated as described in Series IV.

The final series of experiments involved measuring the diurnal excretion pattern when the fish were either exposed to ketanserin or RU-486. Control excretion rates were measured for the first 30 h and following this period the fish was exposed to 30 μmol l^-1 ketanserin or 1 μmol l^-1 RU-486 every 6 h for the next 54 h. Water samples from each experiment were frozen at -20°C for up to 1 month and later analyzed for urea and ammonia content.

Extracts were prepared by grinding the intact fish to a fine powder with a mortar and pestle in liquid nitrogen(Wright et al., 1995). Urea concentrations were analyzed as described previously(Rahmatullah and Boyde, 1980)and values were expressed as mmol N l^-1. Samples were analyzed for ammonia content using a modified published method(Kun and Kearney, 1974) for use with a SpectraMax 190 micoplate reader (Molecular Devices Corp.,Sunnyvale, CA, USA). The efficiency of the assay was determined by spiking samples with known amounts of ammonia and determining the percent recovery(98.5±1.3% recovery rate). Values for ammonia concentration in the tissues are expressed in mmol N l^-1.

Seawater urea concentrations were measured with a colorimetric assay(Rahmatullah and Boyde, 1980)using an Ultrospec 3300 Pro spectrophotometer (Biochrom, Cambridge,UK). Ammonia content of water samples were quantified by assay methods described elsewhere (Ivancic and Degobbis,1984). The rates of excretion (J) were calculated as described previously (Wright and Wood,1985).

Control experiments were carried out to determine if bacterial contamination from various sources such as the individual fish, the water supply or attached to the experimental chamber could have affected nitrogen excretion rates. Fish were placed in 30 ml of water for a 1 h interval and subsequently removed. Ammonia and urea concentrations were monitored in the water during the next 6 h. Analysis revealed that microbial contamination did not significantly affect nitrogen excretion rates from R. marmoratus(P>0.05).

Preliminary experiments were performed to determine if the hormone receptor antagonists used in this study had an effect on the colour development of either the ammonia or urea assays. There was no significant effect of either RU-486 or ketanserin on the ammonia assay, neither was the urea assay affected by RU-486. However, the colour development of the urea assay in the presence of ketanserin was slightly less intense (10%) relative to samples without ketanserin. Due to the fact that this change was relatively small and underestimated the influence of ketanserin (but did not alter the trends in the data), the values were not corrected for this effect.

Oxygen consumption was measured by closed respirometry with eight double-walled glass chambers connected in series to a water bath to control the temperature inside the chambers. Changes in oxygen levels were measured using an automatic temperature compensated dissolved oxygen sensor with built-in thermistor and amplifier (Vernier Software and Technology, Beaverton,OR, USA) and measurements were recorded in mg l^-1. Measurements of water oxygen levels were quantified using the LoggerPro Software (Vernier Software and Technology). Values for oxygen consumption are expressed in(μmol O₂ min^-1 g^-1).

The data are presented as means ± standard error of the mean(s.e.m.). For Series I, III and IV, a one-way analysis of variance (ANOVA) was used to determine if tissue levels and oxygen consumption rates varied over time (P<0.05). A Tukey test (SPSS, SPSS Inc., Chicago, IL, USA)was used to test for differences between time intervals(P<0.05).

For Series II, analysis of oxygen consumption values was carried out using the single cosinor approach (Halberg et al., 1972; Nelson et al.,1979; Rodela and Wright,2006). In this model a cosine function is fitted to the data by least squares regression, defining in the process several parameters of the circadian rhythm: its rhythm-adjusted mean value or mesor; its amplitude,which is one half of the difference between the highest and the lowest values;and the time at which the waveform reaches its peak value or acrophase. Recognition of circadian activity was accomplished by testing a null hypothesis of zero amplitude with an F-test (P<0.05).

For Series V, analysis of the dose-dependence data for both ketanserin and RU-486 was done with a one-way ANOVA (P<0.05) followed by a Dunnett's test to locate statistical differences (P<0.05). For Series VI, a paired t-test was used to test for significance differences between control and treatment values for each individual time period (P<0.05). Due to confounding effects of time and treatment on diurnal nitrogen excretion, a general linear model contrast (SAS, SPSS Inc.) was used to determine if control values during the diurnal antagonist exposure were significantly different from treatment values.

In fasted R.marmoratus kept on their normal photoperiod of 12 h:12 h L:D, whole body tissue urea levels followed a diurnal pattern over time with a definitive a peak and trough over a 1.5 day period(Fig. 1). The period of the cycle was approximately 24 h, with urea accumulating in the tissues overnight to reach maximum concentrations at 12:00 h. Over the next 6 h, tissue urea levels were at their lowest values, 3.5-fold lower that at 12:00 h. In contrast to urea, whole body ammonia concentrations remained constant over time (data not shown), ranging from 6.0 to 6.8 mmol l^-1. The highest levels of urea accumulation in the tissues at 12:00 h did not differ significantly from whole body ammonia values at this time. However, following the decline in tissue urea levels (18:00 h), ammonia concentrations were approximately 3.7-fold higher relative to urea.

Oxygen consumption in R. marmoratus followed a significant circadian rhythm (F_2,9=16.4, P=0.0007). The highest rates of oxygen uptake occurred during the day between 11:30 h and 12:30 h, with an amplitude of 32% of the meson, the acrophase occurred at 13:48±00:34 h (Fig. 2). R. marmoratus consume significantly less oxygen between 17:30 h and 18:30 h. Comparisons between the maximum and minimum rates of oxygen consumption yield a 1.9-fold difference between the peaks and troughs.

A closer examination of oxygen consumption on an hourly basis between 12:00 h and 18:00 h revealed that the definitive peak of oxygen uptake was 2 h in duration and occurred between 12:00 h and 14:00 h(Fig. 3). Oxygen consumption decreased over the afternoon until rates reach their lowest values at 18:00 h.

When J_urea rates were examined over the same 1 h intervals between 12:00 h and 18:00 h, the pattern was similar to the hourly oxygen consumption measurements (Fig. 3). J_urea was significantly higher between 13:00 h and 14:00 h relative to the subsequent hourly intervals. By 17:00-18:00 h, approximately 72% less urea was excreted compared to values obtained between the hours of 13:00 h and 14:00 h. Over the same 6 h intervals, there was no significant difference in ammonia excretion rates(data not shown).

Both receptor antagonists required the use of an ethanol vehicle to facilitate solubility; however, ethanol concentrations in the water did not significantly affect either urea or ammonia excretion rates (t-test, P<0.05). Application of the ketanserin, a 5-HT₂antagonist, caused a dose-dependent inhibition of J_urea(data not shown). A concentration of 50 μmol l^-1 ketanserin caused a 61% inhibition of J_urea. Excretion of ammonia remained unaffected by the application of ketanserin (data not shown). Exposure to RU-486, a cortisol receptor antagonist, also produced a dose-dependent decrease in J_urea (data not shown). Exposure to 5 μmol l^-1 RU-486 resulted in a 65% inhibition of J_urea (data not shown). Ammonia excretion remained unaffected by RU-486 (data not shown).

Under control conditions a diurnal cycle in J_urea was present (data not shown). The highest rates of urea excretion occurred between 12:00 h and 18:00 h, and were 78% higher than rates between 00:00 h and 06:00 h. Exposure to 30 μmol l^-1 of ketanserin between 12:00 h-18:00 h resulted in a 40% inhibition of J_urea relative to control values (Fig. 4A). During the first 6 h of recovery (18:00 h-0:00 h), urea excretion continued to be depressed (50%) in post-treatment fish relative to control values. In the subsequent 6 h interval (0:00 h-6:00 h), however, J_ureareturned to control levels. Ammonia excretion rates remained unaffected by the application of ketanserin (data not shown). RU-486 (1 μmol l^-1)exposure resulted in a 49% inhibition of J_urea rates(Fig. 4B). Following recovery in fresh brackish water, J_urea returned to control values during the next 6 h interval. Ammonia excretion did not change in response to RU-486 (data not shown).

A strong diurnal cycle in urea excretion was present in fish exposed to the ethanol control prior to repeated applications of 30 μmol l^-1ketanserin (Fig. 5). Application of ketanserin did not affect the properties (i.e. period, timing of peaks and troughs, degree of change between peaks and troughs) of the diurnal cycle, however, the absolute values on the rates of J_urea were affected(Fig. 5). Exposure to ketanserin resulted in a 39-72% inhibition in J_urea rates. Ammonia excretion remained constant over time (data not shown). For control fish, J_amm constituted 72-85% of J_nitrogen whereas in the ketanserin-treated fish, J_amm constituted 81-95% of J_nitrogen.

Exposure to repeated doses of 1 μmol l^-1 RU-486 resulted in a 30-42% inhibition in the absolute rate of J_urea(Fig. 6). Similar to ketanserin-treated fish, the period and the timing of the peaks and troughs did not change in RU-486-treated fish; only the absolute values of the rates were affected. Ammonia excretion remained steady for the duration of the experiment (data not shown). The majority of wastes excreted during the experiment were in the form of ammonia.

Urea, but not ammonia, excretion in R. marmoratus undergoes a diurnal rhythmicity with the highest rates of urea excretion occurring during mid-afternoon between 12:00 h and 18:00 h(Rodela and Wright, 2006). We first hypothesized that the diurnal J_urea pattern in R. marmoratus is due to metabolic changes. In particular, fluctuating tissue urea concentrations alter the plasma-to-water urea gradient and these variations are reflected in the pattern of J_urea. Measurements of oxygen consumption in fasted R. marmoratusdemonstrated a diurnal pattern with the highest rates occurring during midday. Diurnal fluctuations in urea, but not ammonia tissue concentrations paralleled rhythms observed in oxygen uptake and urea excretion. The peak of each variable occurred between 12:00 h and 14:00 h. Taken together, these results suggest that urea excretion rates in R. marmoratus may be under metabolic influence.

In R. marmoratus, urea is derived from the routine turnover of uric acid or arginine hydrolysis, rather than the ornithine-urea cycle(Frick and Wright, 2002b). One possibility to explain the diurnal urea excretion pattern in fasted R. marmoratus, is that urea is produced periodically. The highest concentrations of urea occurred at 12:00 h (∼6 mmol l^-1) and lowest levels were observed at 18:00 h (∼1.6 mmol l^-1). This is a profound change (∼ 3.8-fold) in whole body urea levels in a relatively short period of time. Although plasma urea concentrations change by approximately sevenfold over 2 h in gulf toadfish (O. beta) due to their peculiar pattern of pulsatile urea excretion(Wood et al., 1997), this is not surprising given that the extracellular fluid volume is approximately 5%of the mass of the fish. Whether or not similar dramatic changes occur in the intracellular compartment in O. beta is unknown, to our knowledge. Smaller diurnal changes in body urea (∼ twofold) content were observed in the ureogenic gobiid M. abei, which also has a diurnal urea excretion pattern (Kajimura et al.,2002). The activity of several hepatic enzymes related to urea synthesis in M. abei did not show any diurnal changes. Although it is unlikely that enzymes would be rapidly up- or downregulated within a few hours of each day, temporal changes in substrate concentrations may influence the rates of urea synthesis and ultimately affect urea excretion rates. Hence, the results generally support the `metabolic' hypothesis stated above, but upon closer examination of the data, it is clear that other factors are involved in regulating the diurnal pattern of urea excretion in R. marmoratus.

Measurements of oxygen consumption from fasted R. marmoratusrevealed a 24 h diel pattern with approximately a twofold increase in oxygen consumption rates during midday (12:00 h) compared to values obtained at 18:00 h. Clear diurnal patterns in oxygen consumption have been observed in other fish species (Brett and Zala,1975; Ghosh et al.,1986; Ghosh et al.,1990) but the peaks in oxygen consumption are usually indicative of active periods. The oxygen uptake rates obtained in the present study potentially reflect daily changes in routine and active metabolic rate(Rao, 1968; Sims et al., 1993). Hence, the daily variation in urea excretion may not be due to diurnal changes in the rates of all metabolic processes. This idea is supported by the lack of diurnal variation in ammonia content and excretion rates.

There is indirect evidence in M. abei that diurnal fluctuations in urea excretion and urea content are correlated with changes in the permeability of membranes to urea(Kajimura et al., 2002). Hence, the diurnal changes in J_urea in R. marmoratus may be partly explained by diurnal changes in gill urea permeability. Our second hypothesis stated that regulation of J_urea in R. marmoratus is influenced by 5-HT and cortisol, that is, changes in the levels of these hormones results in a change affecting rates of urea excretion. Indeed, ketanserin inhibited J_urea, but not J_amm. The 5-HT antagonist, therefore, probably did not alter protein catabolism but was specific to J_urea mechanisms.

Of the seven different families of receptors, the 5-HT₂ receptor subtype family is the lone category of 5-HT receptors that display sensitivity to ketanserin (Barnes and Sharp,1999). Evidence from numerous studies on teleosts have revealed that ketanserin blocks 5-HT-mediated locomotor activity in zebrafish(Brustein et al., 2003) and 5-HT-mediated gonadotropin release in goldfish(Somoza et al., 1988). Furthermore, 5-HT-induced increases in plasma urea concentrations and excretion rates in O. beta were abolished, in a dose-dependent fashion, by ketanserin (McDonald and Walsh, 2004). These findings indicate that the effects of ketanserin specifically target 5-HT receptors in teleost fish, but whether these physiological responses are specifically mediated by the 5-HT₂ subtype in teleosts remains to be ascertained.

Repeated exposure to ketanserin for 2 consecutive days revealed that the diurnal J_urea pattern still remained under these conditions; however the absolute rates of excretion were significantly diminished. In O. beta, injections of a 5-HT₂ receptor agonist elicited substantial urea pulses of comparable size and duration to natural pulses, whereas a 5-HT₂ receptor antagonist inhibited pulsatile urea excretion implicating 5-HT₂ receptors in the rapid activation of the toadfish UT (McDonald and Walsh, 2004). Based on the ketanserin experiments in the present study, it is possible that 5-HT may have a role in regulating putative urea transport proteins in R. marmoratus; however, further experiments are required to explore this idea.

Ketanserin may be affecting other neuroendocrine cascades that ultimately influence urea excretion pathways. In rats, brain serotonergic activity increases corticosterone release (Fuller,1996). In rainbow trout, 5-HT agonists elevate plasma cortisol levels in a dose-dependent manner, suggesting that this hormone has an important role in cortisol mobilization(Winberg et al., 1997). It is possible that fluctuations in 5-HT activity in the brain in R. marmoratus affect circulating cortisol levels specifically through the HPI axis, which in turn influences various aspects of urea metabolism and excretion (see below).

Exposure to RU-486 resulted in an inhibition of urea, but not ammonia excretion in R. marmoratus. As with ketanserin, the inhibition of urea excretion following exposure to a solitary dose of RU-486 was reversible,therefore confirming that the drug was entering the fish and influencing J_urea pathways. J_urea rates returned to control levels during the next 6 h period postexposure. Similar to ketanserin, repeated application of RU-486 did not alter the timing of the diurnal cycle; however, the absolute rates of urea excretion were significantly depressed. Therefore, blocking cortisol appears to significantly decrease urea excretion in R. marmoratus, in contrast to other studies pertaining to urea transport regulation in fish and mammals by glucocorticoids (Marsh and Knepper,1992; Wood et al.,1997; Wood et al.,2001; Peng et al.,2002; McDonald et al.,2004).

Inhibiting cortisol receptor activity by RU-486 may alter the rate of urea synthesis in R. marmoratus, targeting enzymes in the arginolytic and/or uricolytic pathways. Exposure to exogenous cortisol triggers a twofold increase in arginase activity accompanied by an increase in allantoicase activity in the ammoniotelic sea raven, Hemipterus americanus(Vijayan et al., 1996). An increase in arginase activity would enhance arginine catabolism to form urea,whereas an increase in the uricolytic enzyme, allantoicase, would enhance the synthesis of urea from uric acid. In vivo exposure to cortisol in the rainbow trout resulted in an increase in circulating urea concentrations accompanied by an increase in branchial and renal excretion rates due to an altered blood-to-water urea gradient(McDonald and Wood, 2004). Thus, cortisol may specifically target urea metabolism in R. marmoratus. Furthermore, coupled with the evidence that 5-HT may also increase cortisol mobilization in fish (see above), it is possible that these two hormones act in tandem to increase arginolysis and uricolysis and therefore affect plasma urea concentrations and excretion rates.

Regulation of urea synthesis and excretion by hormones such as 5-HT and cortisol may have an adaptive significance in relation to predation risk in R. marmoratus. Many studies have implicated the involvement of cortisol in modifying behavioural processes and social rank in fish (e.g. Pottinger and Pickering, 1992; Gregory and Wood, 1999; DiBattista et al., 2005). Recent studies have shown that predation or threat of predation significantly raised cortisol levels or serotonergic activity in a number of fish species(Winberg et al., 1993; Woodley and Peterson, 2003). It is possible that predation risk or perceived predation risk by R. marmoratus is greatest between 12:00 h and 14:00 h, and elevated hormone levels, in turn, alter nitrogen excretion rates. Such a strategy may help to evade predators. Preliminary findings from Barimo and Walsh(Barimo and Walsh, 2005) have shown that O. beta may be excreting pulses of urea as a form of chemical crypsis to avoid predator detection by the grey snapper (Lutjanus griseus). There is very little ecological information, however, on R. marmoratus in the wild.

To conclude, there is a strong synchrony between the diurnal patterns of oxygen consumption, whole body urea content and urea excretion rates in R. marmoratus. These results support the hypothesis that the diurnal urea excretion pattern is under metabolic control, but several factors suggested a more complex regulatory process. Cortisol and 5-HT receptor antagonists inhibited the absolute rates of J_urea but did not influence the timing of the J_urea pattern. We propose that 5-HT may regulate the rate of J_urea by affecting the permeability of the gill to urea via a putative gill urea transport protein and/or indirectly altering urea synthesis mediated by other neuroendocrine agents. We further propose that changes in cortisol levels may alter the rate of urea synthesis, which in turn alters the blood-to-water urea gradient, ultimately influencing the rate of urea excretion in R. marmoratus.

The authors would like to thank Dr David Noakes for the donation of the fish. Furthermore, the authors would like to extend their gratitude to Dr Don Stevens for his help with the statistical analyses and Dr Nick Bernier for his guidance in the experimental design. This work was supported by a NSERC Discovery Grant to P.A.W.