Characterization of diurnal urea excretion in the mangrove killifish, Rivulus marmoratus

Many behavioral, biochemical and physiological oscillations result in rhythmic circadian events that are often synchronized to various environmental cues. In many cases, these daily oscillations are more than a passive response to variations in the organism's surroundings; they are in fact controlled by an endogenous and complex timekeeping system. The classical view of the circadian timekeeping model describes a hierarchical arrangement of highly specialized pathways connected to a central pacemaker that accepts input from both photic and non-photic influences(Yanelli and Harrington,2004). Closely coupled biochemical and molecular mechanisms are upregulated in the central pacemaker, triggering neuroendocrine cascades that result in the coordination of various oscillators in the central and peripheral tissues (Jin et al.,1999; Whitmore et al.,2000; Cahill,2002).

Many catabolic processes are extremely complex. They require an internal temporal organization and integration of various physiological and biochemical pathways. The central pacemaker facilitates the synchronization of these functions. Excretion of nitrogenous wastes requires the coordination of digestive pathways, catabolic processes in the liver and transport proteins in various tissues. The majority of teleosts are ammoniotelic, excreting 70-90%of their nitrogenous waste products as ammonia(Wood, 1993). The remaining 10-30% is largely due to the excretion of urea(Kajimura et al., 2004). The mangrove killifish Rivulus marmoratus is no exception to this general rule, although an unusual characteristic of urea excretion(J_urea) is the fact that urea is excreted in a diurnal pattern, whereas ammonia is excreted (J_amm) at a steady rate (Frick and Wright,2002a). The majority of urea is excreted during the day, with the peak of J_urea occurring between the hours of 12:00 h and 18:00 h. Two other teleosts studied to date, the gobiid fish Mugilogobius abei and the marine toadfish Opsanus beta, both ureogenic, have a non-uniform pattern of J_urea. The ammoniotelic R. marmoratus is a small (∼100 mg) self-fertilizing fish that inhabits temporal pools and crab burrows in mangrove habitats. They are also known to survive for weeks in moist leaf litter and detritus, as well as tolerating a range of water conditions including variations in water temperature, salinity and oxygen tensions (Abel et al.,1987; King et al.,1989; Davis et al.,1990; Frick and Wright,2002a; Frick and Wright,2002b). The purpose of our study was to determine whether the diurnal pattern of J_urea in R. marmoratus meets the criteria for a circadian rhythm.

The foremost defining feature of circadian rhythms is a periodicity of approximately 24 h, with only slight deviations from this value. Although biotic and abiotic factors may influence the timing of the phases and length of the period, the rhythm should remain in the absence of these factors. If a diurnal or nocturnal pattern is sustained under constant conditions, the rhythm is defined as free-running. For instance, goldfish Carassius auratus, held under conditions of constant darkness, showed free-running rhythms in locomotor and feeding activity with an average of 25.3±1.8 h and 24.4±1.7 h, respectively(Sánchez-Vázquez et al.,1996). Observations on locomotor activity in zebrafish Danio rerio under similar constant conditions demonstrated a comparable pattern with a period of 25.5 h (Cahill et al.,1998; Hurd et al.,1998). The third characteristic of a circadian rhythm is that they can be entrained to new environmental stimuli, which can change or reset the phase of a circadian clock. The light:dark cycle clearly appears to be the dominant zeitgeber for all organisms; however, other non-photic stimuli such as feeding can have an equally powerful effect on resetting the clock(Yanelli and Harrington,2004). Light-sensitive circadian oscillators have been isolated in many peripheral tissues of the zebrafish, including the heart, kidney and liver (Whitmore et al., 1998; Whitmore et al., 2000). The rhythms in zebrafish are entrained fairly quickly, that is, within 2-3 days of exposure to the new photoperiod. Furthermore, cell lines derived from zebrafish embryos have demonstrated that these light-entrainable circadian oscillators are present quite early in development(Pando et al., 2001).

In the present study we hypothesized that the diurnal J_urea pattern in R. marmoratus follows a circadian rhythm. We predict that J_urea in R. marmoratus has a periodicity of 24 h, which is sustained under constant conditions, and can be entrained to different photoperiods. J_urea and J_amm were measured in fasted fish under normal conditions (12 h:12 h light:dark), continuous darkness (0 h:24 h light:dark), and under an inverted photoperiod. Comparisons were made between fasted and fed individuals under normal light:dark conditions to determine if feeding influenced the diurnal pattern. Furthermore, nitrogen excretion rates were measured in zebrafish Danio rerio, a tropical fish of similar size and metabolic demand to R. marmoratus. In a companion study, we investigated the mechanisms regulating J_urea patterns in R. marmoratus(Rodela and Wright, 2006).

A colony of mangrove killifish Rivulus marmoratus Poey was held at the Hagen Aqualab at the University of Guelph (Guelph, ON, Canada) in brackish water (15‰ seawater) (Frick and Wright, 2002a). Mature fish (0.07-0.25 g) were held in individual translucent containers in 60 ml of 15‰ artificial seawater (pH 8.1)using distilled water and marine salt (Instant Ocean™, Crystal Sea,Baltimore, MA, USA). Water in all containers was changed biweekly. The fish were held in environmental chambers at 25°C on a photoperiod of 12 h:12 h light (L):dark (D). Fish were fed Artermia salina nauplii once a day 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 unless otherwise indicated.

Adult zebrafish Danio rerio Hamilton (DAP International,Etobicoke, ON, Canada) were kept in static aerated freshwater (pH 8.1,25°C, 12 h:12 h L:D) in 2 l plastic holding tanks. Zebrafish (0.27-0.37 g)were fed TetraMin fish flakes (Pet Paradise, Guelph, ON, Canada) every other day except 48 h prior and for the duration of the experiments. Treatment of animals followed approved experimental protocols and guidelines of the University of Guelph Animal Care Committee.

To investigate the diurnal nitrogen excretion rhythms, five series of experiments were conducted.

Series I: nitrogen excretion measurements under 12 h:12 h L:D in fasted R. marmoratus.

Series II: nitrogen excretion measurements under 12 h:12 h L:D in fed R. marmoratus.

Series III: nitrogen excretion measurements under 0 h:24 h L:D in fasted R. marmoratus.

Series IV: nitrogen excretion measurements under 12 h:12 h D:L in fasted R. marmoratus.

Series V: nitrogen excretion measurements under 12 h:12 h L:D in fasted D. rerio.

Initial experiments were conducted on fish under their natural photoperiod(pH 8.1, 25°C, 12 h:12 h L:D) in containers with 30 ml of water. Water samples were collected every 6 h for a 3-day period for measurement of J_amm and J_urea. Water was changed after each sampling interval to avoid the accumulation of urea or ammonia in the water. To eliminate the stress of handling, the fish were kept in a double container. The inner container was made out of mesh in the exact same configuration as the plastic translucent holding container. When a water change was required, 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 double-walled 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.

For Series II-V, the fish were acclimated to the experimental conditions 2 days preceding the start of the experiment. For Series II, fish were fed 5 ml of Artemia nauplii at 12:30 h daily and the photoperiod remained unchanged (12 h:12 h L:D). In Series III, food-deprived fish were exposed to complete darkness for 24 h (0 h:24 h L:D). For the final series of experiments(IV) in R. marmoratus, the fish were deprived of food and excretion rates were measured under an inverse photoperiod (12 h:12 h D:L).

As a comparison to other tropical teleost species, both J_urea and J_amm were measured in food-deprived zebrafish Danio rerio, exposed to a 12 h:12 h L:D photoperiod, also kept at 25°C. Water was sampled and renewed as described above for the killifish experiments. Following the termination of each series of experiments, the fish were blotted dry and weighed for calculation of excretion rates.

Urea levels in both freshwater and seawater samples were quantified by a colorimetric assay (Rahmatullah and Boyde,1980) using an Ultrospec 3300 Pro spectrophotometer(Biochrom, Cambridge, UK). Ammonia content of seawater samples was measured as described elsewhere (Ivancic and Degobbis,1984). Freshwater samples were analyzed for ammonia using a colorimetric assay method previously described(Verdouw et al., 1978). The rates of excretion (J) were calculated using the methodology outlined elsewhere (Wright and Wood,1985).

Control experiments were performed to determine if bacterial contamination from fish sources, water sources or attached to the experimental chamber could have influenced nitrogen excretion rates. Fish were placed in 30 ml of water for a 1 h interval and subsequently removed. Ammonia and urea content were monitored in the water over the following 6 h. Analysis revealed that microbial contamination did not significantly affect nitrogen excretion rates(P>0.05).

defines three circadian parameters: C_o represents the mesor, a rhythm-adjusted median value; C is the amplitude (one half the difference between the maximum and minimum values); Φ is the time at which the curve reaches its highest value or acrophase. Time, which is represented by t_i, is measured as a fraction of 24 h. An F-test was used to determine the statistical significance of the circadian rhythm(P<0.05). This test allows subjective analysis of the hypothesis that the rhythm amplitude differs from zero. Furthermore, comparisons of the maximal absolute urea and ammonia excretion values between experimental series were performed using t-tests with either assumptions for equal or unequal variances (P<0.05).Verification of equal variances was accomplished using an F-test (P<0.05).

Under a normal photoperiod of 12 h:12 h L:D, J_urea in food-deprived R. marmoratus followed a significant circadian rhythm(F_2,9=23.8, P=0.0003). The highest rates of J_urea occurred during the day between the hours of 12:00 h and 18:00 h with an amplitude of 47% of the meson; the acrophase occurred at 14:30±00:39 h (Fig. 1A). Fasted R. marmoratus were predominantly ammoniotelic, excreting 69-89% of their nitrogenous wastes in the form of ammonia. However, in contrast to J_urea, ammonia excretion did not demonstrate any significant circadian activity (F_2,9=0.84, P=0.46; Fig. 1B).

Fed fish demonstrated a significant circadian rhythm in both J_urea (F_2,9=18.3, P=0.0007)and J_amm (F_2,9=11.6, P=0.0032). J_urea in fed fish had peaks and troughs at times identical to the pattern observed in fasted fish(Fig. 2A). The amplitude of the oscillation was marked; it was 72% of the meson and the peak of the rhythm occurred at 15:45±00:18 h. In comparison with fasted fish(Fig. 1A), fed fish(Fig. 2A) excreted approximately 2.3-6.3 times more urea, the variations for the minimum and maximum rates respectively. Diurnal J_amm comprised 57-85%of total nitrogen wastes excreted in fed fish(Fig. 2B). The amplitude of the diurnal J_amm pattern was 88% of the meson with the highest rates of excretion occurring between 12:00 h and 18:00 h and the lowest rates occurring between 06:00 h to 12:00 h, just prior to feeding. The acrophase for the circadian rhythm for J_amm occurred at 15:50±0:12 h. J_amm rates were 1.4-8.3 times larger compared with values obtained from fasted fish.

When food-deprived R. marmoratus were kept in darkness for 24 h (0 h:24 h L:D) the circadian pattern of J_urea persisted(F_2,9=60.7, P=0.00001; Fig. 3). The highest rates of J_urea occurred during 12:00 h and 18:00 h with the amplitude of the oscillation equaling 48% of the meson. The absolute rates of J_urea were significantly (1.6-1.8 times) lower relative to values obtained from fasted fish (12 h:12 h L:D). The acrophase occurred at 14:31±00:33 h. J_amm was constant over time(F_2,9=0.003, P=0.99) and excretion rates were not significantly different than values obtained from fasted fish (12 h:12 h L:D)(data not shown).

A significant circadian rhythm in J_urea(F_2,9=39.6, P=0.00003) was present in fish acclimated to an inverted photoperiod (12 h:12 h D:L). The period of the single cycle remained at 24 h, however, the occurrence of the J_urea peaks and troughs were different from individuals on a normal photoperiod (12:12 h L:D). The highest rates of J_urea occurred between the 00:00 h and 06:00 h, with an amplitude 52% of the meson (Fig. 4). The acrophase occurred at 02:16±00:28 h. The absolute values of J_urea were significantly lower by 1.6 (peak) to 2.3 (trough) times compared to fish on a normal photoperiod. J_amm remained constant over time(F_2,9=0.48, P=0.63) and there was no significant difference between J_amm rates in fish on an inverted photoperiod compared to a normal photoperiod (data not shown).

In fasted D. rerio, there was no significant circadian rhythm in urea and ammonia excretion (F_2,9=2.33, P=0.15, F_2,9=1.28, P=0.33; Fig. 5A,B). Approximately 83%of total nitrogen in D. rerio was excreted in the form of ammonia. J_urea was up to 2.6 times higher in R. marmoratuscompared to D. rerio with the largest difference occurring during the daytime (12:00 h to 18:00 h).

Free-running rhythms, 24 h periodicity and entrainment are all criteria for classification of a circadian rhythm. Our results support the hypothesis that the diurnal J_urea pattern in R. marmoratusfollows a circadian rhythm. Under various photoperiod manipulations and feeding regimes, J_urea in R. marmoratusdemonstrated a significant rhythmicity with a period of approximately 24 h,the acrophase occurring during mid-afternoon of the subjective day. In contrast, no diurnal pattern in J_urea was observed in D. rerio. Under continuous conditions of darkness, the J_urea pattern in R. marmoratus appeared to free-run with an approximate period of 24 h and similar amplitude and acrophase to fish exposed to a normal photoperiod. Furthermore, upon exposure to an inverted photoperiod, the diurnal pattern became entrained to the new conditions (12 h:12 h D:L), with the acrophase of J_ureapattern still occurring during subjective afternoon of the photoperiod.

Natural 24 h light:dark cycles play an important part in circadian rhythms as daily biological patterns often parallel environmental changes. In many cases, this synchrony between the animal and its external milieu gives the individual the ability to anticipate temporal changes in the surroundings. However, a true endogenously generated rhythm persists even in the absence of abiotic cues and continues unperturbed until the animal is presented with a new external cue that acts to reset the central clock(Whitmore et al., 2000; Pando and Sassone-Corsi,2002). Locomotor activity in several fish species continues in a diurnal fashion with a periodicity of 24.4-25.5 h under constant dim or darkened conditions(Sánchez-Vázquez et al.,1996; Cahill et al.,1998; Hurd et al.,1998; Gerkema et al.,2000). Free-running rhythms in darkness have been recorded in electric discharge of the gymnotid electric fish, Eigenmannia virescens (Deng and Tseng,2000). Furthermore, a daily pattern in J_ureain M. abei was maintained in darkness(Kajimura et al., 2002). In R. marmoratus, the peak of J_urea in constant darkness occurred during the time of subjective day of the previous photoperiod (12 h:12 h L:D). Aschoff demonstrated that the free-running rhythm in diurnal animals tends to be slightly longer than 24 h, whereas the period length in nocturnal animals under free-running conditions is somewhat shorter than 24 h (Aschoff, 1981). As measurements for J_urea and J_amm in R. marmoratus were made over 6 h intervals, a precise estimation of period length was not possible.

An intriguing feature of the free-running rhythm (0 h:24 h L:D) in fasted R. marmoratus is that the absolute rates of J_ureawere significantly lower (1.6-1.8 times) than values obtained from fasted fish held under a normal photoperiod (12 h:12 h L:D). Usually, under constant conditions, when some peripheral oscillators are removed from the light-sensitive influence of the master pacemaker the amplitude of the circadian rhythm becomes diminished over time and eventually ceases. For example, circulating thyroxine levels in red drum Sciaenops ocellatus(Leiner and MacKenzie, 2001)and melatonin rhythms in the pineal organ of the ayu Plecoglossus altivelis are reduced over time in constant darkness(Iigo et al., 2004). As well,circadian rhythms of Clock mRNA in the kidney and heart of zebrafish oscillated with a lower amplitude until they disappeared after a period of 2-3 days in darkness (Whitmore et al.,2000). Urea excretion rates in R. marmoratus were measured for three complete cycles after a 2-day acclimation period. It is possible that J_urea is under control of a peripheral oscillator and perhaps over time, a period of several weeks, the rhythm may become dampened and eventually cease under conditions of continuous darkness.

The transition between dark and light appears to be the dominant zeitgeber for all organisms. Lighting cues appear to play an important role in resetting the clock that controls J_urea in R. marmoratus.Like most animals, R. marmoratus can become entrained to new photoregimens within a day or two (Pohl,1978). Measurements of zebrafish Clock gene expression in the heart and kidney revealed inherent and independent light-sensitive circadian oscillators that entrained to altered photoperiods within 2-3 days(Whitmore et al., 1998; Whitmore et al., 2000).

Under the influence of an inverted photoperiod, the amplitude of J_urea in food-deprived R. marmoratus was significantly lower (by twofold) relative to fish exposed to a normal photoperiod (Table 1). There are several possible explanations for this difference in amplitude. Non-photic stimuli such as feeding may act as an important cue regulating and resetting the clock (Yanelli and Harrington,2004). It is possible that while light may have a role in entraining the J_urea pattern in R. marmoratusthere may also be some confounding influences from a prior non-photic zeitgeber that may be resisting the change to the new photoperiod. Stokkan et al. reported that rhythmicity of certain genes in the liver of rats retain a phase relation to previous restricted feeding schedule even during fasting(Stokkan et al., 2001). Furthermore, there is a partial coupling between food-entrained and light-entrained activity in sea bass, Dicentrarchus labrax(Sánchez-Vázquez et al., 1995). When goldfish were held under conflicting zeitgebers (light and feeding cycles), activity rhythms were initially synchronized to the light cycle but gradually feeding time assumed greater importance (Aranada et al., 2001). In R. marmoratus,J_urea may depend on temporal regulation of the rate of food digestion, urea synthesis pathways, and changes in transport mechanisms that may ultimately be under control of light and food entrainable circadian clocks.

Rates of nitrogen excretion in fed fish are typically correlated with dietary protein content and the proportion of nitrogen wastes excreted as urea versus ammonia remains constant (e.g. Wright, 1993; Walsh and Milligan, 1995; Verbeeten et al., 1999). Indeed, the ratio of nitrogen waste products did not vary under different feeding regimes in R. marmoratus. Both J_urea and J_amm demonstrated up to a six- and eightfold increase,respectively, in absolute rates and this is consistent with trends described by other studies (Brett and Zala,1975; Kaushik,1980; Kaushik et al.,1983; Wright,1993; Alsop and Wood,1997).

J_urea and J_amm in zebrafish were steady over time, whereas J_urea in captive (this study)and wild R. marmoratus (Frick and Wright, 2002a) demonstrate a circadian rhythm. This disparity between nitrogen excretion strategies may relate to the ecology and life history of each species. Zebrafish are primarily a freshwater species that prefers to inhabit lotic bodies of water, such as rivers and streams, and they will often remain near the surface of the water column in large groups(Talwar and Jhingran, 1991). R. marmoratus tend to exhibit a more solitary and fossorial lifestyle, often seeking shelter in detritus and leaf litter; however, they also frequently inhabit the burrows of land crabs in estuarine environments(Taylor, 1988; Davis et al., 1990). Limited water flow and elevated ammonia concentrations appear to be major shortcomings for the cryptic burrowing strategy used by R. marmoratus(Frick and Wright, 2002a). A few exceptional species shift towards ureotelism under these conditions(Forster and Goldstein, 1966; Gordon et al., 1978; Ramaswamy and Gopalakrishna Reddy,1983; Saha and Ratha,1990; Wood et al.,1998), but this is not the case for R. marmoratus. Exposure to ammonia concentrations up to 10 mmol l^-1 does not change the proportion of ammonia versus urea excreted in R. marmoratus (Frick and Wright,2002a). An alternative explanation may relate to predation. O. beta, a facultative ureotelic teleost, is a preferred prey item of the grey snapper (Lutjanus griseus). Preliminary data from behavioral assays indicates that the grey snapper has a stronger attraction to ammonia than either urea alone or an ammonia/urea mixture(Barimo and Walsh, 2005). Hence, it is possible that the circadian J_urea pattern in R. marmoratus is correlated with potential predator activity but further ecological data is required to verify this hypothesis.

In conclusion, this study provides strong evidence that the diurnal J_urea pattern in R. marmoratus meets the criteria for a circadian rhythm, namely a 24 h periodicity, a free-running rhythm in constant darkness, and entrainment to a reversed photoperiod. The findings suggest that light cues may have an important role in coordinating the various physiological processes involved in nitrogen metabolism, primarily catabolism of proteins, urea synthesis and urea transport in R. marmoratus. The absence of a diurnal pattern of J_urea in D. reriois consistent with the majority of teleost fishes.

The authors would like to thank Dr David Noakes for the donation of the fish and Dr Don Stevens for his assistance with the statistical analyses. This work was supported by a NSERC Discovery Grant to P.A.W.