Stingrays of the family Potamotrygonidae are the only stenohaline freshwater elasmobranchs. Potomotrygon sp. collected from the ion-poor blackwaters ([Na+], [Cl-] and[Ca2+]=10-30 μmol l-1, pH 6.1) of the Rio Negro,Amazonas, Brazil, were ammoniotelic (91% ammonia-N, 9% urea-N excretion) and exhibited blood chemistry (Na+, Cl-, urea, ammonia and glucose levels and osmolality) typical of freshwater teleosts. Unidirectional Na+ and Cl- influx rates, measured with radiotracers,displayed saturation kinetics. The relationships for Cl- and Na+ had similar Km values (300-500 μmol l-1), but Jmax values for Cl-(approximately 950 μmol kg-1 h-1) were almost twice those for Na+ (approximately 500 μmol kg-1h-1). Cl- efflux rates varied with external concentration, but Na+ efflux rates did not. There were no differences in the kinetic variables (Km, Jmax) for influx between animals acclimated to their native ion-poor blackwater or to ion-rich hard water, but efflux rates for both Na+ and Cl- were lower in the former, yielding much lower balance points (external Na+ or Cl- levels at which influx and efflux were equal). Na+, Cl- and Ca2+ uptake were all strongly inhibited by acute exposure to pH 4.0, but efflux rates and Ca2+ binding to the body surface did not change. Na+ influx was inhibited by amiloride (10-4 mol l-1) and by two of its analogs, phenamil (4×10-5mol l-1) and HMA (4×10-5 mol l-1), with the latter being slightly more potent, while Cl- fluxes were unaffected. Cl- fluxes were insensitive to DIDS(2×10-5 mol l-1 or 10-4 mol l-1) and SITS (10-4 mol l-1), but both influx and efflux rates were strongly inhibited by DPC (10-4 mol l-1) and thiocyanate (10-4 mol l-1). Ammonia excretion was unresponsive to large changes in water Na+concentration, but was elevated by 70% during acute exposure to pH 4.0 and transiently inhibited by approximately 50% by amiloride and its analogues. The strategy of adaptation to ion-poor blackwater appears similar to that of some Rio Negro teleosts (Cichlidae) in which low-affinity transport systems are relatively sensitive to inhibition by low pH but are complemented by low diffusive loss rates. Ionic transport systems in these freshwater elasmobranchs, although superficially similar to those in some freshwater teleosts, may bear more resemblance to their presumed evolutionary precursors in marine elasmobranchs.

Stingrays of the family Potamotrygonidae are widespread throughout river systems of South America that drain into the Atlantic Ocean. While some members of the family Dasyatidae may complete their entire life cycle in freshwater (Compagno and Roberts,1982; Johnson and Snelson,1996), the potamotrygonid stingrays are considered to be the only group of elasmobranchs to have speciated exclusively within freshwaters(Lovejoy, 1996). In contrast to all other freshwater-tolerant elasmobranchs, potamotrygonid rays have lost their ability to live in seawater and are restricted in nature to waters with a salinity of less than 3 ppt (Brooks et al., 1981). Unique features include an inability to retain urea to any great extent when challenged with higher salinity(Thorson et al., 1967; Junqueira et al., 1968; Thorson, 1970; Griffith et al., 1973; Gerst and Thorson, 1977; Bittner and Lang, 1980), the absence of salt excretion by the rectal gland, which is degenerate(Thorson et al., 1978), and modifications of the electroreceptive ampullae of Lorenzini for freshwater operation (Raschi and Mackanos,1989).

The exact route and timing of evolutionary invasion from the ocean remains controversial, either from the Pacific Ocean during the late Cretaceous 90-100 million years ago (Brooks et al.,1981; Brooks, 1995)or from the northern coast of South America (i.e. Caribbean Atlantic) during the early Miocene 15-23 million years ago(Lovejoy, 1996; Lovejoy et al., 1998). Regardless of which scenario proves correct, it is apparent that the Potamotrygonidae have experienced as long a period for evolutionary adaptation to freshwater as many teleosts endemic to the same areas(Lovejoy et al., 1998). It is therefore possible that they have evolved ionoregulatory strategies similar to those of freshwater teleosts and different from those of their euryhaline cousins, stingrays of the family Dasyatidae. Indeed, it is known that the Potamotrygonidae are ammoniotelic rather than ureotelic(Gerst and Thorson, 1977; Barcellos et al., 1997) and regulate their blood salts, osmolality and urea at levels similar to those of teleosts (Thorson et al.,1967; Griffith et al.,1973; Gerst and Thorson,1977; Mangum et al.,1978; Bittner and Lang,1980) and at levels lower than those of freshwater-adapted Dasyatidae (Thorson, 1967; Shuttleworth, 1988; Piermarini and Evans, 1998). However, little is known about the details of the mechanisms involved.

Some species of the Potamotrygonidae are endemic to the most extreme freshwater environment of the continent, the acidic, ion-poor blackwaters of the Rio Negro and its tributaries, so called because of their high content of dissolved organic carbon (DOC) comprising humic, fulvic and other organic acids derived from the breakdown of jungle vegetation. Recently, Gonzalez and colleagues have begun to characterize ionoregulatory mechanisms in various teleost species endemic to these blackwaters(Gonzalez, 1996; Gonzalez et al., 1997, 1998, 2002; Gonzalez and Preest, 1999; Gonzalez and Wilson, 2001). Two basic strategies have emerged in the Rio Negro teleosts. In one strategy,as seen in many members of the Family Characidae, the presence of a high-affinity (i.e. low Km), high-capacity Na+transport system, which is relatively insensitive to inhibition by low pH,compensates for high rates of diffusive loss. In the other, as seen in some members of the Family Cichlidae, the transporter is sensitive to inhibition by low pH, and affinity is much lower (i.e. higher Km),although capacity may still be high. However, the key to adaptation in these species is that diffusive loss rates are much lower. In both strategies,diffusive loss rates are relatively resistant to stimulation by low pH, but show variable patterns of change in response to alterations in external Na+ concentration.

With this background in mind, the objectives of the present study were several-fold. First, using radiotracers to measure unidirectional flux rates,we described the basic concentration-dependent kinetics of both Na+and Cl- uptake and efflux in a Potamotrygon species collected from and tested in its native blackwater. This point is important because Gonzalez et al. (2002)concluded that transport mechanisms of Rio Negro teleosts bred in captivity and/or held in other water qualities for long periods overseas by aquarists may differ from those of native fish in natural blackwater. Second, we examined whether these kinetic relationships changed when the freshwater rays were acclimated to ion-rich hard water. Third, we evaluated the responses of Na+, Cl- and Ca2+ transport and ammonia excretion to acute short-term exposure to low pH (pH 4.0) in natural blackwater. The only previous work on ammonia excretion in Rio Negro teleosts indicated that it generally increases at low pH and does not appear to be linked to Na+ uptake (Wilson,1996; Wilson et al.,1999). Lastly, we used a range of pharmacological inhibitors to characterize the nature of the Na+, Cl- and ammonia transport systems in stingrays acclimated to and tested in blackwater. To date, little research has been done on Cl- or Ca2+transport or on the pharmacological characterisation of any of the transporters in Rio Negro fish.

Experimental animals

Stingrays (70-770 g; mean 270±40 g, N=21) were collected in November-December 1999 by beach seine or throw-net from sandy, shallow regions of the Rio Negro and its tributaries between Barcelos and the Anavilhanas Archipelago. The systematics of Potamotrygon are controversial. The rays were morphologically uniform but heterogeneous in patterning. They were identified by a collector as Potamotrygon thorsoni but by local experts as Potamotrygon aff. hystrix, recognizing that hystrix may include up to three species (including thorsoni)and/or subspecies that hybridize and show considerable regional variability in color patterns (Brooks et al.,1981). We therefore refer to the fish as Potamotrygon sp. Since freshwater rays have protected status in Brazil, animal numbers were limited, and all procedures were designed to be non-lethal. Experiments were planned such that after exposure to any pharmacological treatment (always by addition of drugs to the water), animals were allowed 48 h for recovery prior to further pharmacological tests. In all cases, control measurements were made prior to tests to ensure that baseline conditions had been re-established.

After capture, most (16) of the rays were transported to the Laboratory of Ecophysiology and Molecular Evolution [Instituto Nacional de Pesquisas da Amazonia (INPA)], Manaus, Brazil, where they were held for 10 days in a recirculating, filtered tank (3001) in which the INPA wellwater (which is very soft) had been supplemented with NaCl and limestone chips, a common procedure used to reduce the stress of captivity. This provided us with an opportunity to examine the kinetics of both Na+ and Cl- uptake and efflux (see below) in this species after acclimation to ion-rich hard water(composition in Table 1). After completion of tests at INPA, the fish were transported back to the Anavilhanas Archipelago on board the research vessel Amanai II. The research vessel was moored in an embayment close to the Scientific Base of the Anavilhanas Archipelago of the Instituto Brasiliero do Meio Ambiente e dos Recursos Naturais Renovaveis. During this time, the 3001 holding tank was continually flushed with fresh blackwater (which is very soft) from the Rio Negro (composition in Table 1),i.e. the original water quality from which the fish had been collected. The rays were held in this water for at least 5 days prior to further tests, and all other experiments were performed in this water quality on board the research vessel. During the 14-day period of research at Anavilhanas, an additional five rays were collected locally, from the wild and added to the pool of experimental animals. The fish were offered food (pellets, tubifex worms) but did not appear to eat in captivity. Holding and experimental temperatures were 28-31°C throughout.

Chemicals

The radioisotopes 22Na (as NaCl), 36Cl (as HCl) and 45Ca (as CaCl2) were manufactured by New England Nuclear(Dupont) and supplied by REM (Sao Paulo). Dimethylsulphoxide (DMSO), amiloride hydrochloride, 5-(N,N)-hexamethylene amiloride (HMA),4,4′-diisothiocyanatostilbene-2,2′-disulfonic acid disodium salt(DIDS), 4-acetomido-4′-isothiocyanatostilbene-2,2′-disulfonic acid sodium salt (SITS), sodium thiocyanate, lithium heparin and tricaine methane sulfonate (MS-222) were all obtained from Sigma. Diphenylamine-2-carboxylic acid (DPC: N-phenylanthranilic acid) and phenamil methane sulphonate were obtained from RBI Pharmaceuticals.

General methods

In all experiments, rays were placed in individual polyethylene containers and allowed to settle for at least 12h prior to tests. The containers were fitted with airstones and lids and flushed with water from the holding reservoir or directly from the river at approximately 200 ml min-1. Container volume was 2.61, except for the largest rays (450-770 g), for which 61 chambers were used. At the start of all flux experiments, the inflowing water was stopped, and the volume was set to a known volume, generally 2.01 for fish up to 450 g, and 3.0 or 4.01 as appropriate for the largest rays.

In most experiments, unidirectional Na+ and Cl- flux rates were determined simultaneously by adding both 22Na (a dual gamma and beta emitter) and 36Cl (a beta emitter only; neutralized with KOH) to the external water in a 1:2.5 ratio(Wood, 1988). Differential gamma and scintillation counting (see below) was employed to monitor the disappearance of 22Na and 36Cl counts into the fish over time. Initial tests using `cold displacement' (a rinse with 20 mmol l-1 NaCl, approximately 1000× ambient concentrations, at the end of an experiment) demonstrated that there was no significant adsorption of either radioisotope to the walls of the container or the surface of the fish. Unidirectional influx rates were low in these fish, so flux periods of 2 h (or longer) were routinely employed. Water samples (generally 15 ml) were taken at the beginning and end of each flux period for the measurement of 22Na and 36Cl radioactivity and total concentrations of water Na+ (by flame photometry) and Cl- (by colorimetric assay). Influx rates (JinX, by convention positive) were calculated from the mean external specific activity and the disappearance of counts from the external water (factored by time, volume and fish mass), net flux rates (JnetX) were calculated from the change in total Na+ or Cl-concentration in the water (similarly factored) and unidirectional efflux rates (JoutX, by convention negative) were calculated by difference, as outlined in detail by Wood(1988). This approach makes no assumptions about steady state and allows comparisons of different treatments over time. Experiments were designed such that internal specific activity,calculated from the measured accumulation of counts in the fish (i.e. cumulative disappearance from the water) and an estimated internal Na+ or Cl- pool of 40 mmol kg-1, was never more than 5% of measured external specific activity at the end of an experiment, so there was no need for `backflux correction'. In these same experiments, net ammonia flux rates were determined simultaneously by the appearance of total ammonia (measured colorimetrically) in the external water. Net urea-N excretion was similarly measured in a few experiments.

Measurements of the kinetics of Na+ and Cl-uptake and efflux

Kinetic relationships were determined by starting at the lowest water Na+ and Cl- concentrations and then moving upwards in approximately twofold steps until approximately 2000 μmol l-1was reached. For stingrays acclimated to ion-rich hard water(Table 1), this was achieved by replacing the acclimation water immediately prior to the first test period with INPA wellwater ([Na+]≈20 μmol l-1,[Cl-]≈20 μmol l-1, [Ca2+]≈10 μmol l-1) supplemented 24 h previously with sufficient CaCO3to raise the measured [Ca2+] to 912 μmol l-1,approximately equal to that in the acclimation water. For stingrays acclimated to ion-poor Rio Negro blackwater (Table 1), the starting point was Rio Negro water. Radioisotopes (18.5 kBq of 22Na, 46.3 kBq of 36Cl or proportionately more for the largest rays) were added, allowed to mix for 10 min, and then an initial 15 ml water sample was taken, followed by a final 15 ml water sample after 2h. Thereafter, sufficient quantities of NaCl, 22Na and 36Cl were added (as a single stock solution) to approximately double the concentrations of all in the external bath. The process was repeated until six flux determinations had been made, each at progressively twofold higher external Na+ and Cl- concentrations and radioactivity. This procedure ensured that external specific activity remained approximately constant throughout the kinetics experiment. To achieve comparable resolution at these progressively higher external concentrations,the length of the flux period was gradually increased from 2 h to 2.5 h, 3.25 h, 4.5 h, 7 h and finally 10 h. As both Na+ and Cl-uptake clearly exhibited saturation kinetics, the JinNa and JinCldata for each individual were fitted separately to a Michaelis—Menten model by Eadie—Hofstee regression(Michal, 1985) to yield individual estimates of Km (affinity constant) and Jmax (maximum transport rate). Grand means ± S.E.M. were then calculated for all fish in a treatment group.

Measurement of Ca2+ flux rates

We found that the methodology used for Na+ and Cl-flux determinations would not work for measuring unidirectional Ca2+ fluxes, because Ca2+ influx rates were lower and there was a significant adsorption of 45Ca to the body surface of the stingrays (in all experiments) and to the walls of the container (in some experiments). Therefore, a longer-term measurement with correction for the latter was developed. The method assumed steady-state conditions and therefore allowed only a single measurement of JinCaunder a single treatment condition. However, the method also provided an estimate of the amount of Ca2+ bound to the surface of the fish. At the start of the experiment, 148 kBq of 45Ca was added to the 2.01 of water in each container, allowed to mix for 10 min, and 15 ml water samples(for measurement of 45Ca radioactivity and total Ca2+concentration) were then taken at 0 h, 3 h, 6 h and 9 h. The slope of the regression line of the natural logarithm of total water radioactivity against time yielded the rate constant (k) for turnover, and the product of k and the mean external total Ca2+ pool, factored by mass,yielded an uncorrected estimate of JinCa (see Kirschner, 1970).

After the 9 h sample, the ray was gently transferred to new container containing 1.01 of fresh Rio Negro water. A water sample was taken, and then 10 ml of 1 mol l-1 Ca(NO3)2 was added to raise the ambient Ca2+ concentration to 10 mmol l-1(approximately 1000× ambient concentration) for `cold displacement'. A second water sample was taken, and the increase in 45Ca radioactivity, factored by volume, mean specific activity of the original water and mass, was taken as the amount bound to the surface of the fish(B). Similarly, the water remaining in the original container(approximately 2.01) was spiked with 20 ml of 1 mol l-1Ca(NO3)2, and any increase in 45Ca radioactivity that occurred, factored by volume and mean specific activity of the original water, was taken as the amount bound (C) to the container's surface. Corrected JinCa was then calculated as:
\[\mathrm{corrected}J_{\mathrm{in}}^{\mathrm{Ca}}=\frac{(\mathrm{uncorrected}J_{\mathrm{in}}^{\mathrm{Ca}}{\times}m{\times}9)-(B{\times}m)-C}{m{\times}9},\]
where m is mass and 9 refers to 9 h sample. JnetCa and JoutCa were then calculated as for Na+ and Cl- fluxes. The approach therefore assumes that 45Ca simply equilibrates with preexisting `cold' Ca2+ on surface sites and that these sites do not contribute to net fluxes.

Responses to low pH

The responses to low pH in Rio Negro water were assessed by measuring unidirectional and net Na+ and Cl- flux rates and net ammonia flux rates during a 2 h control period, followed by a 2 h low-pH period and a 2 h recovery period, using each fish as its own control. Water samples (15 ml) were taken at the beginning and end of each period. At 10 min prior to the start of the control period, 37 kBq of 22Na and 92.5 kBq of 36Cl were added to the external water (or proportionately more if the volume was greater than 2.01). A 20-min interval between each period was used to lower the pH to approximately 4.0 with 0.5 mol l-1 HNO3 and then raise it back to approximately 6.1 with 0.5 mol l-1 KOH for recovery. In practice, because pH tended to rise during the low-pH period as a result of ammonia production by the fish, pH was monitored in each container every 30 min and was held between 3.8 and 4.2 by addition of more HNO3. During the control and recovery periods, pH was held between 5.9 and 6.3.

To assess the effects of low pH on unidirectional Ca2+ fluxes,two separate experiments were performed using the Ca2+ methodology outlined above. In the first 9 h experiment (control), pH was held at approximately 6.1. In the second experiment, water pH was lowered to approximately 4.0 approximately 30 min before the start of the experiment and held at this level for the ensuing 9 h by monitoring at 30 min intervals.

Responses to pharmacological treatments

We tested the effects of various pharmacological agents on the unidirectional and net flux rates of Na+ and/or Cl- and the net flux rates of ammonia in Rio Negro water. The basic protocol was a 2 h control period followed by three successive 2 h experimental periods in the presence of the drug, using each fish as its own control. Water samples (15 ml) were taken at the beginning and end of each period. At 10 min prior to the start of the control period, 37 kBq of 22Na and 92.5 kBq of 36Cl were added to the external water (or proportionately more if the volume was greater than 2.01). In tests with amiloride (10-4mol l-1), DIDS (2×10-5 mol l-1),phenamil (4×10-5 mol l-1) and HMA(4×10-5 mol l-1), both Na+ and Cl- exchanges were evaluated. In tests with DIDS (10-4mol l-1), SITS (10-4 mol l-1), DPC(10-4 mol l-1) and thiocyanate (10-4 mol l-1), only Cl- exchanges were evaluated. In all cases,drugs were initially dissolved in DMSO, yielding a final concentration of 0.1%DMSO in the experimental water. The same concentration of DMSO was therefore used during the control periods.

Blood sampling

At the end of the 14-day period on the Amanai II, stingrays acclimated to Rio Negro water were lightly anaesthetized in MS-222. Approximately 1 ml of blood was withdrawn non-terminally from either the caudal arch or the heart using a 23-gauge needle and syringe heparinized with 50 μl of 5000 i.u. ml-1 lithium heparin. Haematocrit and plasma protein (by refractometry) were measured immediately after centrifugation of blood in microcapillary tubes at 500 g for 10 min. The remainder of the plasma was separated by centrifugation at 10 000 g, portioned,and frozen in liquid nitrogen for later analysis of Na+,Cl-, glucose, urea and total ammonia concentrations and osmolality.

Analytical methods

22Na radioactivity was determined by manually counting 2 ml water samples in a Picker Cliniscaler gamma counter fitted with a 10 cm NaI crystal with a well. 22Na plus 36Cl radioactivity was determined by mixing 3.0 ml water samples with 7.5 ml of Ecolite fluor and manually counting on a Triathler portable scintillation counter. 36Cl radioactivity was obtained by subtraction after correcting for differences in efficiency of 22Na counting between the two instruments, determined by counting the same standards on both(Wood, 1988). 45Ca radioactivity was determined directly by scintillation counting. Water and plasma total Na+ and water total Ca2+ concentrations were determined by flame photometry (CELM flame photometer) or atomic absorption spectrophotometry (Perkin Elmer 1100B), water DOC by a Rosemount total carbon analyzer, and water pH using an Orion (model 266) portable meter and electrode. Water total Cl- concentrations were determined by the colorimetric assay of Zall et al.(1956), and water total ammonia concentrations by the salicylate hypochlorite method of Verdouw et al.(1978). In both assays,standards were made up with the same concentrations of DMSO and the appropriate drug, as in the experimental medium. In addition, we ran blanks to correct for the absorbance due to the natural color of blackwater, and confirmed that the slopes of the assays were not altered by this color. Water and plasma urea concentrations were determined by the diacetyl monoxime method of Rahmatullah and Boyde(1980). Plasma Cl-concentration was determined by coulometric titration (Radiometer CMT10),plasma osmolality by vapour pressure osmometry (Wescor 5100C), plasma protein concentration by refractometry (American Optical TS meter; Alexander and Ingram, 1980),plasma total ammonia concentration by the L-glutamate dehydrogenase method(Mondzac et al., 1965; Sigma kit 171-UV) and plasma glucose concentration by the hexokinase method(Bergmeyer, 1985; Sigma kit 17).

Statistical analyses

All data are reported as means ± 1 S.E.M. (N). Relationships were assessed by analysis of variance (ANOVA), and individual means were compared using Student's paired or unpaired two-tailed t-tests, as appropriate, with the Bonferroni correction for multiple comparisons (Nemenyi et al.,1977). A significance level of P≤0.05 was used throughout.

Blood chemistry and nitrogen metabolism

Concentrations for most measured plasma components (ions, glucose, urea and total ammonia and osmolality; Table 2) in these freshwater elasmobranchs held in Rio Negro blackwater were typical of those for freshwater teleosts, although plasma protein concentration was notably low. Nitrogen metabolism was very characteristic of the teleost pattern, with plasma urea concentration being very low relative to other elasmobranchs but approximately fourfold higher than plasma total ammonia concentration (eightfold higher on a per unit N basis). Urea-N excretion comprised only approximately 9% (58±10 μmol kg-1 h-1) and ammonia-N excretion approximately 91%(607±29 μmol kg-1 h-1; N=6) of the total N excretion.

Kinetics of Na+ and Cl- uptake and efflux

Both JinNa and JinCl exhibited clear saturation kinetics as external Na+ and Cl- concentrations were increased(Fig. 1). There were no differences in the kinetic variables (Km, Jmax) for these relationships between animals acclimated to their native ion-poor blackwater of the Rio Negro or to ion-rich hard water(Table 3), so single lines have been fitted to the data of Fig. 1. The relationships for Cl- were significantly different from those for Na+, with 1.7- to twofold higher Jmax values but virtually identical Kmvalues (Table 3). As an additional check that the relationships for Rio Negro water had not changed as a result of the animals' holding and manipulation in the laboratory, uptake kinetics of Na+ and Cl- were also measured in four stingrays freshly collected from the Rio Negro. These yielded values of Km=563±86 μmoll-1 and Jmax=400±73 μmol kg-1 h-1for Na+ and Km=569±98μmoll-1 and Jmax=487±160 μmol kg-1 h-1 for Cl-, not significantly different from the corresponding values (Table 1) for fish acclimated to Rio Negro water in the laboratory. The slightly lower Jmax values in the freshly collected animals probably reflected their 2.5-fold greater mean body mass, and we therefore elected not to pool these data in the kinetic relationships.

To evaluate whether there was any effect of acclimation water chemistry on ammonia excretion or any direct coupling of ammonia excretion to JinNa, the net ammonia flux was measured at the lowest and highest external Na+ concentrations used in establishing the kinetic relationships (Table 4). There were no differences in ammonia excretion associated with acclimation water chemistry, and no differences in ammonia excretion between the lowest and highest water Na+ levels, despite approximately sevenfold differences in JinNa at the two concentrations. The four freshly collected Rio Negro fish exhibited the same patterns (data not shown).

JoutNa did not vary significantly with external Na+ concentration in either acclimation condition and was generally lower than JinNa, so that JnetNa was positive over most of the range tested (Fig. 2). In the lower range of external Na+ levels, JoutNawas lower in the stingrays acclimated to Rio Negro water, with a significant difference at around 400 μmoll-1(Fig. 2). While the absolute differences appear small, the net effect was to shift the balance point(marked by arrows in Fig. 2),where JinNa=JoutNa(and hence JnetNa is zero), from 220μmoll-1 in ion-rich hard water to 75 μmoll-1 in ion-poor Rio Negro water.

JoutCl values were generally greater than JoutNa values at comparable external ion concentrations. In contrast to JoutNa, JoutCl varied significantly with external Cl- concentration under both acclimation conditions, in a manner similar to the influx relationship (Fig. 3). Over most of the range of external Cl- levels,there were no differences between the two treatments, but the relationships diverged around 300 μmoll-1, with a significantly lower value of JoutCl in the fish acclimated to Rio Negro water. As a result, the JinCl=JoutClbalance point (marked by arrows in Fig. 3) shifted from 460 μmol l-1 in ion-rich hard water to 65 μmol l-1 in ion-poor Rio Negro water.

All other experiments were performed with stingrays held in Rio Negro water. Under the conditions of our experiments, blackwater Na+ and Cl- concentrations were generally around 20-40 μmol l-1 (Table 1; i.e. slightly below the balance points), so the animals were in slight negative balance for both Na+ and Cl- during control treatments(Figs 4,5,6,7). JinNa and JinClwere low, approximately 30-80 μmol kg-1 h-1(generally higher for JinCl than for JinNa), while efflux rates were approximately twice these values. DMSO (0.1%) had no effect on any of the flux rates measured.

Responses to low pH

Acute exposure to pH 4.0 induced substantial reductions (approximately 80%)in both JinNa and JinCl (Fig. 4A). However, there was no increase in net ion loss rates because efflux rates were slightly reduced (Fig. 4A). At the same time, net ammonia excretion was increased by approximately 70% (Fig. 4B). When the water pH was returned to 6.1 after 2 h of exposure to low pH, the recovery of JinNa was only partial, and JoutNa was now significantly lower (i.e. Na+ turnover was reduced). In contrast, Cl- fluxes returned to normal levels, and ammonia flux was no longer significantly elevated.

At pH 6.1, JinCa was only approximately 19μmol kg-1 h-1(Table 5) under control conditions, lower than JinNa or JinCl. JoutCawas approximately 3.5 times higher than JinCa,so JnetCa was negative. Exposure to pH 4.0 completely abolished JinCa, while JoutCa and JnetCa did not change significantly. Interestingly, the amount of Ca2+ bound to the surface of the fish(6-7 μmol kg-1) was not significantly altered by exposure to low pH (Table 5).

Responses to pharmacological treatments

Exposure to the cationic transport inhibitor amiloride (10-4 mol l-1) caused a 70% inhibition of JinNa, an effect that was stable throughout all three 2 h periods of the experimental treatment(Fig. 5A). There was no significant effect on JoutNa or JnetNa, or on the unidirectional or net flux rates of Cl-. The non-significant tendency for a large stimulation of JinCl and JoutCl probably reflected the fact that the drug was presented as the hydrochloride salt, which therefore raised the water Cl- concentration by 100 μmol l-1 to around 130μmol l-1. In this range, both JinCl (Fig. 1) and JoutCl(Fig. 3) are very sensitive to small changes in external water Cl- concentration, and the increases were approximately equal to those expected from the kinetic relationships.

Exposure to the anionic transport inhibitor DIDS (2×10-5mol l-1) had no effect on JinCl, JoutCl or JnetCl(Fig. 5B). Unidirectional and net flux rates of Na+ were also unaffected.

To pursue the amiloride effect, two analogues (HMA, phenamil) of reputedly greater specificity were tested at a lower concentration(4×10-5 mol l-1). Responses to the two drugs were very similar, although HMA appeared to be slightly more potent. Exposure to HMA (4×10-5 mol l-1) induced a 90% inhibition of JinNa, which persisted throughout all three 2 h periods of the treatment (Fig. 6A). Phenamil (4×10-5 mol l-1) caused a 70% inhibition of JinNa, which was significant during the first two experimental periods only(Fig. 6B). During the first and third periods of drug exposure, JinNa was significantly lower in the presence of HMA than in the presence of phenamil. Both drugs caused significant increases in JoutNa and more negative values of JnetNa throughout the exposures, whereas there were no significant effects on unidirectional or net flux rates of Cl-, apart from a decrease in JinClat 2-4 h in the presence of HMA.

To pursue the surprising lack of response of Cl- transport to DIDS, both this drug and a related stilbene derivative (SITS) were tested at a higher concentration (10-4 mol l-1). In addition, two other anionic transport blockers (DPC, thiocyanate) were evaluated at this same concentration. To increase the sensitivity for Cl- uptake measurement, only 36Cl was used in these experiments, so that 36Cl counts could be obtained by direct scintillation counting,avoiding the need for differential gamma and scintillation detection to eliminate 22Na counts. Furthermore, DIDS, SITS and thiocyanate were all presented as Na+ salts, thereby raising water Na+concentration significantly, which would have complicated interpretation of Na+ flux rates.

Even at 10-4 mol l-1, both DIDS(Fig. 7A) and SITS(Fig. 7B) were without significant effect on JinCl, JoutCl or JnetCl. However, both DPC(Fig. 7C) and thiocyanate(Fig. 7D) strongly inhibited JinCl, reducing it to almost zero by the third 2 h period. JoutCl was also strongly reduced during exposure to DPC and thiocyanate (significant in the second and third periods), so that Cl- turnover was greatly reduced, and therefore JnetCl became only slightly more negative.

Exposure to amiloride (at 10-4 mol l-1) and its analogues (HMA and phenamil at 4×10-5 mol l-1)exerted similar effects on net ammonia excretion, inhibiting it by approximately 50% during the first 2 h of treatment, with recovery in subsequent periods. This response was significant for amiloride(Fig. 8A) and HMA(Fig. 8B), but just below the level of significance (P≤0.10) for phenamil(Fig. 8C). None of the anionic transport blockers tested had any effect on net ammonia excretion (data not shown).

Strategies of adaptation

Here, we confirm earlier reports of typical `teleostean' blood chemistry(low Na+, Cl- and urea levels, Table 2; Thorson et al., 1967; Griffith et al., 1973; Gerst and Thorson, 1977; Mangum et al., 1978; Bittner and Lang, 1980) in potamotrygonid stingrays, very different from that of dasyatid stingrays(Piermarini and Evans, 1998)and other euryhaline elasmobranchs in freshwater(Shuttleworth, 1988), which have much higher levels of these plasma constituents. More importantly, we provide the first data on ion transport kinetics in the only stenohaline freshwater elasmobranch. To our knowledge, the only previously published information on ion transport in the family Potamotrygonidae appeared in a review article by Pang et al.(1977). These authors stated that, on the basis of preliminary experiments, `the influx rate of Na+ was rather low', the stingrays `are capable of picking up Na+ from surrounding water against a concentration gradient' and that `the uptake system is saturable'. However, no data were reported. The present results demonstrate that this preliminary report was correct and extend the same conclusions about low rates of uptake occurring against large gradients with typical saturation kinetics to Cl- transport(Fig. 1). Indeed, this is the first time that Cl- transport has been characterized in any Amazonian blackwater fish.

At least superficially, the freshwater stingrays appear to be physiologically similar to some of the freshwater teleosts that are also endemic to the dilute blackwaters of the Amazon basin. These teleosts may or may not have shared the same route of invasion(Brooks, 1995; Lovejoy et al., 1998). Gonzalez et al. (2002)recently summarized kinetic variables and acid sensitivity for Na+transport in a number of teleost species collected from the Rio Negro. The data divided into two strategies for adaptation to acidic, ion-poor water: (i)high-capacity, high-affinity (i.e. low Km) Na+transport systems relatively resistant to inhibition by low pH, but linked with high diffusive loss rates; and (ii) low-affinity (i.e. high Km) transport systems relatively sensitive to inhibition by low pH, but complemented by low diffusive loss rates. The freshwater rays represent an extreme version of the second strategy, with Km values for Na+ (and Cl-) of approximately 400-500 μmol l-1(Fig. 1; Table 3), at least double that of any of the teleosts surveyed, Jmax values at the lower end of the range for the teleosts surveyed, normal influx and efflux rates of less than 200 μmol kg-1 h-1 and an 80% inhibition of influx but no increase in efflux or net loss rate upon challenge with pH 4.0(Fig. 4). As pointed out by Gonzalez et al. (2002), this strategy, typical of Amazonian cichlids, is designed to limit net ion loss and thereby minimize physiological disturbance so as to `wait out' exposures to especially acidic and dilute challenges.

Most of the teleosts surveyed by Gonzalez et al.(2002) exhibited JinNa=JoutNabalance points below 43 μmol l-1, whereas the Na+ and Cl- balance points in the stingrays were higher, 65-75 μmol l-1 (Figs 2, 3), approximately twice the normal levels in the blackwater in which they were living(Table 1). Thus, the rays were in negative balance in their native water, but they were not feeding in our studies. We suspect the explanation is that these animals normally obtain significant supplementary ions from the diet, as documented for some teleosts(e.g. Smith et al., 1989). The importance of food-derived salts in successful adaptation to low environmental pH has recently been highlighted for salmonids(D'Cruz and Wood, 1998). A second explanation may be that the stingrays normally live at the bottom of the water column, just above the benthos; ion concentrations in the water column may be slightly higher here through remobilization from the sediments.

A surprising result was the absence of any shift in the Na+ or Cl- influx kinetic curves between acclimation to ion-rich hard water and ion-poor Rio Negro water (Fig. 1; Table 3). The`normal' pattern, documented in many freshwater teleosts and crustaceans (e.g. Shaw, 1959; Maetz, 1974; McDonald and Rogano, 1986; Perry and Laurent, 1989; Potts, 1994), is for a higher affinity (lower Km) and/or greater Jmax upon acclimation to ion-poor water. The explanation may be that modification of ion uptake systems originally designed for function in seawater (see below) has been pushed to a maximum during the evolutionary invasion of freshwater, and there is no capacity in these elasmobranchs for further modification to differentiate between normal freshwater and very dilute freshwater. In this regard, examination of teleost families with a similar distribution to the Potamotrygonidae, and which today have both marine and blackwater representatives (e.g. Scianidae, Engraulidae,Clupeidae, Belonidae, Soleidae; Gery,1969; Chao, 1978; Fink and Fink, 1979), may be instructive.

On the other hand, upon acclimation to ion-poor blackwater, there were subtle but significant reductions in both JoutNa and JoutCl, which substantially shifted the balance points to much lower external Na+ and Cl- concentrations(Figs 2, 3). Efflux rates may consist of two components in these fish (see below), but it appears likely that an attenuation of the simple diffusive component via epithelial tightening was responsible for the overall reductions. This is a classic response to softwater acclimation (e.g. McDonald and Rogano, 1986). Possibly, DOC may also play a role. Since dissolved organic matter is now known to bind to fish gills, especially at low pH(Campbell et al., 1997),Gonzalez et al. (1998, 2002) have speculated that the high DOC content of ion-poor blackwaters(Table 1) may also be important in limiting gill permeability.

Mechanisms of Ca2+ transport and binding

To our knowledge, there are no previous measurements of unidirectional Ca2+ fluxes (Table 5) in any elasmobranch, marine or freshwater, or in any Amazonian blackwater teleosts. In other freshwater teleosts, in which transport rates are similar, Ca2+ uptake is active and is mediated by apical Ca2+ channels and basolateral high-affinity Ca2+-ATPase and/or Na+/Ca2+ exchange mechanisms in gill ionocytes(Flik et al., 1995; Jurss and Bastrop, 1995; Perry, 1997). As for Na+ and Cl- balance, these stingrays were in negative Ca2+ balance when acclimated to their native blackwater(Table 5), again suggesting an important role for dietary uptake or ionic microclimates in benthic waters in normal homeostasis. In two northern hemisphere teleost species, Hobe et al.(1984) reported 65-75%reductions of JinCa upon acute exposure to pH 4.0. In stingrays, JinCa was abolished at pH 4.0 (Table 5), similar to the severe reduction of JinNa and JinCl during acid exposure(Fig. 4). Interestingly, for all three ions, there was no stimulation of the efflux components. Gonzalez and coworkers (Gonzalez and Dunson,1989; Gonzalez et al., 1997, 1998, 2002; Wood et al., 1998; Gonzalez and Preest, 1999; Wilson et al., 1999; Gonzalez and Wilson, 2001)have developed the idea that some blackwater teleosts are able to minimize or prevent low-pH-induced increases in diffusive permeability by having very high affinity Ca2+ binding sites which stabilize external cell junctions. The present measurements showing that surface-bound Ca2+in stingrays are unaffected by exposure to pH 4.0(Table 5) provide direct support for this hypothesis.

Mechanisms of Na+ and Cl- transport

In the following discussion, we assume that ion exchanges occur mainly at the gills, as in other fish. However, given the large surface area and soft texture of the well-perfused skin (which becomes pink with blood flow during hypoxia), we cannot eliminate a significant role for the skin in both ion uptake and efflux, and of course for the renal system in ion efflux (see Mangum et al., 1978). Another important caveat is that the present experiments were performed on whole animals, in which side effects of the pharmacological treatments or water chemistry manipulations may potentially confound interpretation. These might include changes in transepithelial potential(Kirschner, 1997) or alterations in the potency of drugs by trapping in mucus or by interactions with dissolved organic matter in the natural blackwater(Campbell et al., 1997).

The presence of traditional Michaelis—Menten saturation kinetics(Fig. 1) for both JinNa and JinCl(Fig. 1) against large water-to-blood concentration gradients(Table 1) suggests that both occur by carrier-mediated active transport in these freshwater elasmobranchs. They were of different magnitude and responded entirely differently to various pharmacological treatments (Figs 5, 6), suggesting that they are independent, as in most freshwater teleosts (Kirschner, 1970, 1997; Goss et al., 1992; Potts, 1994; Jurss and Bastrop, 1995; Perry, 1997; Evans et al., 1998). JoutCl appeared to exhibit a strong kinetic linkage with JinCl(Fig. 3), a conclusion reinforced by the inhibitory effects of several drugs on both components(Fig. 7C,D). This pattern is reminiscent of the `exchange diffusion' phenomenon seen in many freshwater teleosts and crustaceans (e.g. Shaw,1960; Kerstetter and Kirschner, 1972; de Renzis,1975; Wood et al.,1984; Goss and Wood,1990), in which the same transport mechanism may perform both Cl-/Cl- self-exchange and vectorial transport (e.g. Cl-/base exchange). The phenomenon can be equally well explained by an exchange protein or a selective channel mechanism linked to an electrogenic pump (Potts, 1994). Exchange diffusion has also been seen for Na+ transport where Na+/Na+ and Na+/acid exchange may co-exist(e.g. Shaw, 1959; Wood and Randall, 1973; Goss and Wood, 1990), but this was not apparent in the stingrays. JoutNa and JinNa appeared to be kinetically independent(Fig. 2), and did not co-vary when JinNa was pharmacologically inhibited(Figs 5A, 6A,B). Interestingly, Gonzalez et al. (2002) reported kinetic linkage between JoutNa and JinNa in approximately half the Rio Negro teleost species they surveyed (Cl- fluxes were not measured), but were unable to relate the presence or absence of this exchange diffusion phenomenon to either of the strategies outlined above.

According to theory (Shaw,1959; Kirschner,1970; Wood and Randall,1973; de Renzis,1975; Goss and Wood,1990), the exchange diffusion components of efflux, when present,are always superimposed on the simple diffusive components, and the latter are predominant at low external Na+ and Cl- concentrations. It was probably a reduction in this simple diffusive permeability, rather than a change in the exchange diffusion component, which explained the lower values of JoutNa and JoutCl, and the resulting reductions in Jin=Jout balance points upon acclimation to Rio Negro blackwater (Figs 2, 3).

Amiloride (10-4 mol l-1) inhibited JinNa by 70% in these freshwater stingrays(Fig. 5A). In this regard, the stingray is similar to most freshwater teleosts, in which a 60-95% blockade of Na+ uptake induced by this concentration of amiloride has been widely reported (e.g. Kirschner et al.,1973; Perry and Randall,1981; Perry et al.,1981; Wright and Wood,1985; Wilson et al.,1994; Nelson et al.,1997; Clarke and Potts,1998; Patrick and Wood,1999). Interestingly, in two native Rio Negro characids that demonstrated the high-capacity, high-affinity, low-pH-resistant-strategy of Na+ transport (i.e. opposite to that of the rays), amiloride(10-4 mol l-1) had little effect (13-26% inhibition of JinNa; Gonzalez et al., 1997; Gonzalez and Preest, 1999). However, amiloride at this concentration is a relatively non-selective drug,blocking both Na+ channels and Na+/H+exchangers (Benos, 1982). In an attempt to separate these two possibilities in the stingrays, we employed two drugs of reputedly greater specificity (HMA, phenamil) at lower concentration(4×10-5 mol l-1). In higher vertebrate systems(cell lines), HMA is reported to be over 500 times more potent against Na+/H+ exchange and only 1/30th as potent against Na+ channels as amiloride(Kleyman and Cragoe, 1988). In contrast, phenamil is reported to be 17 times more potent against Na+ channels, and only 1% as potent against Na+/H+ exchange as amiloride(Kleyman and Cragoe, 1988). While HMA was somewhat more effective than phenamil (90% versus 70%inhibition, with greater persistence), our results were not clear-cut because both drugs strongly reduced JinNa(Fig. 6A,B). Amiloride analogue specificities defined in mammalian cell lines may not necessarily apply in intact lower vertebrates. No definitive conclusion as to the importance of Na+/H+ exchange versus Na+channel/H+-ATPase systems in mediating Na+ uptake can be drawn.

Stilbene drugs such as DIDS and SITS are effective blockers of JinCl in some freshwater teleosts(Perry and Randall, 1981; Perry et al., 1981) although they are ineffective in others (Marshall et al., 1997). In the present study, these agents were completely ineffective against Cl- transport in freshwater stingrays, even at concentrations (10-4 mol l-1) above those(2×10-5 mol l-1) usually considered to be specific for blockade of Cl-/HCO3- exchange (Figs 5B, 7A,B). We took precautions to shield these agents from light and to make them up immediately before use, and we also ran one test in water without DOC to ensure that the lack of effect was not due to immobilization of DIDS by dissolved organic molecules in blackwater. Nevertheless, from this result, we cannot eliminate a role for anion exchange in Cl- uptake, because many members of the anion exchange family are insensitive to stilbenes.

Thiocyanate is considered to be a more general blocker (competitive and non-competitive) of many anion exchange systems and has been shown to inhibit Cl- transport in several freshwater teleosts(Epstein et al., 1973; Kerstetter and Kirschner,1974; de Renzis,1975; Perry et al.,1984), even those in which stilbenes are ineffective(Marshall et al., 1997). In the stingrays, at a concentration (10-4 moll-1) only slightly higher than that of Cl- in the water, thiocyanate almost abolished JinCl and also strongly reduced JoutCl (Fig. 7D). These actions are very similar to those reported in the goldfish, which were interpreted as a blockade of both Cl-/HCO3- exchange and Cl-/Cl- exchange diffusion in the very thorough study of de Renzis (1975). Since it is known that thiocyanate is a `sticky' anion that will compete for and block Cl- channels, while DIDS and SITS are generally ineffective against Cl- channels, at least when applied externally(McCarty, 2000), this raised the possibility that a Cl- channel might be involved.

DPC (10-4 moll-1) was employed because it is widely effective in blocking Cl- channels(Gogelein, 1988; Distefano et al., 1985; Chang and Loretz, 1993; McCarty, 2000), including those recently characterized in isolated gill cells of freshwater teleost fish(O'Donnell et al., 2001),although it has never been tested on intact freshwater teleosts. In the stingrays, DPC (10-4 moll-1) essentially duplicated the action of thiocyanate, abolishing JinCl and greatly reducing JoutCl, indicating blockade of both vectorial transport and Cl-/Cl- exchange diffusion. While this would suggest the involvement of a Cl- channel in both processes, the result must be treated with caution because it is unclear where this channel is located. DPC is lipophilic, so even though it was applied in the external medium, it may well have penetrated to the basolateral membrane of gill ionocytes, thereby blocking basolateral Cl- channels, which in turn would disrupt the whole transcellular Cl- transport pathway at the `exit' step rather than the `entry' step. A further caveat is that DPC at this concentration has also been reported to block Cl-/HCO3- exchange in at least one system(Reuss et al., 1987). Whatever the mechanism of Cl- uptake, given a negative intracellular potential, any reasonable estimate of intracellular Cl-concentration in the low millimolar range and an extracellular concentration in the low micromolar range, it would have to be `energized' at the apical membrane, perhaps by a local elevation of intracellular HCO3- concentration coupled to intracellular carbonic anhydrase and/or H+-ATPase or Cl-/HCO3--ATPase systems.

Nitrogen metabolism and ammonia excretion

Urea-N excretion was low (approximately 60 μmol kg-1h-1) and comparable to rates previously reported in other freshwater Potamotrygonidae, while ammonia-N excretion was high (500-600μmol kg-1 h-1) but in the mid-range of previously reported values (200-1000 μmol kg-1 h-1), confirming the ammoniotelic nature of these fish(Goldstein and Forster, 1971; Gerst and Thorson, 1977; Barcellos et al., 1997). The high rate of ammonia-N production may have resulted from metabolism of endogenous protein in these non-feeding animals. Since net ammonia flux was approximately eight- to tenfold higher than normal JinNa in animals held in Rio Negro blackwater,any contribution from Na+/NH4+ exchange was presumably small, and this was in accord with the complete insensitivity of ammonia excretion to approximately sevenfold variations in JinNa in the kinetic experiments(Table 4). Furthermore, the 70%increase in ammonia excretion at pH 4.0, when JinNa was severely reduced, was consistent with a dominant role for simple NH3 diffusion along PNH3 gradients(Wilson, 1996; Wilkie, 1997). It was therefore surprising that amiloride and its analogues (HMA, phenamil) all caused a transitory 50% depression in ammonia excretion(Fig. 8). Indeed, this was larger than the marginal decreases in ammonia efflux (0-30%) normally seen in freshwater teleosts in response to amiloride blockade of JinNa(Kirschner et al., 1973; Wright and Wood, 1985; Wilson et al., 1994; Nelson et al., 1997; Clarke and Potts, 1998; Patrick and Wood, 1999). While originally seen as evidence for Na+/NH4+exchange, the latter observations have more recently been interpreted as an inhibition of `diffusion trapping' of NH3 in gill boundary layer water associated with blockade of H+ extrusion viaNa+/H+ and/or Na+channel/H+-ATPase systems(Wilson, 1996; Wilkie, 1997). Perhaps this diffusion trapping is particularly important in the poorly buffered blackwater, especially if H+ and NH3 excretion sites are`downstream' from alkalizing sites (e.g. HCO3-excretion) on the gills. A related explanation is that blockade of H+ excretion mechanisms by amiloride and its analogs may also reduce internal pH, thereby trapping more ammonia internally as NH4+ and transiently reducing the internal PNH3 and, therefore, the PNH3 gradient for outward NH3diffusion.

Evolutionary perspective and future directions

Based on the limited knowledge of branchial Na+ and Cl- transport in marine elasmobranchs (for a review, see Shuttleworth, 1988), it is possible to put the present results in an evolutionary perspective. In marine elasmobranchs, the rectal gland normally performs the bulk salt excretion, and the gills usually achieve a very small net uptake of Na+ and Cl-. The gills exhibit unidirectional ion flux rates that are only a small fraction of those in marine teleosts and are quantitatively more similar to those of freshwater teleosts. Unidirectional Cl- fluxes are generally greater than unidirectional Na+ fluxes(Bentley et al., 1976). Although the fine details of both fluxes are unknown, the system appears to be designed for pH regulation, normally achieving a small net uptake of Na+ in exchange for `acid', as in freshwater teleosts (Payan and Maetz, 1974; Evans et al.,1979). These characteristics are all similar to those we have identified in these freshwater stingrays and suggest that gill transport mechanisms already present in marine elasmobranchs may have been modified for operation in dilute external solution during the evolutionary invasion of freshwater by the Potamotrygonidae.

In the euryhaline cousins of the Potamotrygonidae, the Dasyatidae,Piermarini and Evans (2000, 2001) have recently identified two types of ionocyte in the gills on the basis of immunocytochemistry, one of which stains for Na+/K+-ATPase and the other for H+-ATPase. After acclimation to freshwater, the expression of both enzymes increased, as measured by western blotting, and the H+ATPase appeared to become localized in the basolateral membrane. Based solely on these data, for no flux measurements or pharmacological characterizations were carried out, Piermarini and Evans(2001) have proposed a model(see their Fig. 7) whereby basolateral Na+/K+-ATPase energizes an apical Na+/H+ exchanger in one cell type, while basolateral H+-ATPase energizes an apical Cl-/HCO3- exchanger in the other. This scheme is certainly unusual compared to most current models for the gills of freshwater teleosts (Goss et al.,1992; Potts, 1994; Jurss and Bastrop, 1995; Perry, 1997; Kirschner, 1997; Evans et al., 1998), but is coherent with our present data on potamotrygonid stingrays. Nevertheless, it must be emphasized that the present pharmacological results cannot conclusively differentiate between Na+/H+ exchange and a Na+ channel/H+-ATPase system on the apical membrane, or between a Cl- channel and Cl-/HCO3- exchange on this membrane. Clearly,the next step is to apply all techniques simultaneously (flux measurements,pharmacology, immunocytochemistry and molecular characterization of transporters) to these two different families of freshwater stingrays, one euryhaline, the other stenohaline.

We thank the captain and crew of the Amanai II for their hospitality and expertise, Maria Lucia Goes de Araujo (Projeto Potamotrygon-ACEPOAM) and the fishermen of INPA for the collection of stingrays, the Director of the Instituto Brasileiro do Meio Ambiente e dos Recursos Naturais Renovaveis for allowing us to use the Scientific Base of the Anavilhanas Archipelago, Dr M. J. O'Donnell for helpful advice, Dr R. Playle for DOC analyses and Gamble Technology (Mississauga, Ontario, Canada) for loan of the Triathler scintillation counter. This work was supported by an NSERC(Canada) research grant to C.M.W., a Royal Society (UK) research grant (RSG 21085) to R.W.W., an NSF (USA) international dissertation enhancement grant to M.L.P., a University Professorship award from the University of San Diego to R.J.G. and a research grant from the Brazilian Research Council to A.L.V. C.M.W. is supported by the Canada Research Chair Program and A.L.V. was the recipient of a research fellowship from the National Research Council of Brazil.

Alexander, J. B. and Ingram, G. A. (
1980
). A comparison of five of the methods commonly used to measure protein concentrations in fish sera.
J. Fish Biol.
16
,
115
-122.
Barcellos, J. F. M., Wood, C. M. and Val, A. L.(
1997
). Ammonia and urea fluxes in Potamotrygon sp., a freshwater stingray of the Amazon. In
The Physiology of Tropical Fish, Symposium Proceedings
(ed. A. L. Val, D. J. Randall and D. MacKinlay), pp.
33
-37. San Francisco, CA: American Fisheries Society.
Benos, D. J. (
1982
). Amiloride: a molecular probe of sodium transport in tissues and cells.
Am. J. Physiol.
242
,
C131
-C145.
Bentley, P. J., Maetz, J. and Payan, P. (
1976
). A study of the unidirectional fluxes of Na+ and Cl-across the gills of the dogfish Scyliorhinus canicula(Chondrichthyes).
J. Exp. Biol.
64
,
629
-637.
Bergmeyer, H. U. (
1985
).
Methods of Enzymatic Analysis
. New York: Academic Press.
Bittner, A. and Lang, S. (
1980
). Some aspects of the osmoregulation of Amazonian freshwater stingrays (Potamotrygon hystrix). I. Serum osmolality, sodium and chloride content, water content, hematocrit, and urea level.
Comp. Biochem. Physiol.
67A
,
9
-13.
Brooks, D. R. (
1995
). Neotropical freshwater stingrays and their parasites: a tale of an ocean and a river long ago.
J. Aquaricult. Aquat. Sci.
7
,
52
-61.
Brooks, D. R., Thorson, T. B. and Mayes, M. A.(
1981
). Fresh-water stingrays (Potamotrygonidae) and their helminth parasites: testing hypotheses of evolution and co-evolution. In
Advances in Cladistics, Proceedings of the First Meeting of the Willi Hennig Society
(ed. V. A. Funk and D. R. Brooks), pp.
147
-175. Bronx, New York: The New York Botanical Garden.
Campbell, P. G. C., Twiss, M. R. and Wilkinson, K. J.(
1997
). Accumulation of natural organic matter on the surfaces of living cells: implications for the interaction of toxic solutes with aquatic biota.
Can. J. Fish. Aquat. Sci.
54
,
2453
-2554.
Chang, W. and Loretz, C. A. (
1993
). DPC blockade of transepithelial chloride absorption and single anion channels in teleost urinary bladder.
Am. J. Physiol.
265
,
R66
-R75.
Chao, L. N. (
1978
). A basis for classifying western Atlantic Scianidae (Teleostei: Perciformes).
NOAA Technical Report NMFS Circ. 415
.
64
pp.
Clarke, A. P. and Potts, W. T. W. (
1998
). Sodium, net acid and ammonia fluxes in freshwater-adapted European flounder(Platichthys flesus L.). Pharmacological inhibition and effects on gill ventilation volume.
J. Zool., Lond.
246
,
427
-432.
Compagno, L. J. V. and Roberts, T. R. (
1982
). Freshwater stingrays (Dasyatidae) of Southeast Asia and New Guinea, with a description of a new species of Himantura and reports of unidentified species.
Environ. Biol. Fish.
7
,
321
-339.
D'Cruz, L. M. and Wood, C. M. (
1998
). The influence of dietary salt and energy on the response to low pH in juvenile rainbow trout.
Physiol. Zool.
71
,
642
-657.
de Renzis, G. (
1975
). The branchial chloride pump in the goldfish Carassius auratus: relationship between Cl-/HCO3- and Cl-/Cl-exchanges and the effect of thiocyanate.
J. Exp. Biol.
63
,
587
-602.
Distefano, A., Wittner, M., Schlatter, E., Lang, H. J., Englert,H. and Greger, R. (
1985
). Diphenylamine-2-carboxylate, a blocker of the Cl- conductive pathway in Cl-transporting epithelia.
Pflügers Arch.
405
,
S95
-S100.
Epstein, F. H., Maetz, J. and de Renzis, G.(
1973
). Active transport of chloride by the teleost gill:inhibition by thiocyanate.
Am. J. Physiol.
224
,
1295
-1299.
Evans, D. H., Kormanik, G. A. and Krasny, E. J.(
1979
). Mechanisms of ammonia and acid excretion by the little skate, Raja erinacea.
J. Exp. Zool.
208
,
431
-437.
Evans, D. H., Piermarini, P. M. and Potts, W. T. W.(
1998
). Ionic transport in the fish gill epithelium.
J. Exp. Zool.
283
,
641
-652.
Flik, G., Verbost, P. M. and Wendelaar Bonga, S. E.(
1995
). Calcium transport processes in fish. In
Cellular and Molecular Approaches to Fish Ionic Regulation
(ed. C. M. Wood and T. J. Shuttleworth), pp.
317
-342. San Diego: Academic Press.
Fink, W. L. and Fink, S. V. (
1979
). Central Amazonia and its fishes.
Comp. Biochem. Physiol.
62A
,
13
-29.
Gerst, J. W. and Thorson, T. B. (
1977
). Effects of saline acclimation on plasma electrolytes, urea excretion, and hepatic urea biosynthesis in a freshwater stingray, Potamotrygon sp. Garman, 1877.
Comp. Biochem. Physiol.
56A
,
87
-93.
Gery, J. (
1969
). The freshwater fishes of South America. In
Biogeography and Ecology in South America
,vol.
2
(ed. E. J. Fitkau), pp.
828
-848. The Hague: Dr W. Junk.
Gogelein, H. (
1988
). Chloride channels in epithelia.
Biochim. Biophys. Acta
947
,
521
-547.
Goldstein, L. and Forster, R. P. (
1971
). Urea biosynthesis and excretion in freshwater and marine elasmobranchs.
Comp. Biochem. Physiol.
39B
,
415
-421.
Gonzalez, R. J. (
1996
). Ion regulation in ion poor waters of low pH. In
Physiology and Biochemistry of the Fishes of the Amazon
(ed. A. L. Val, V. M. Almeida-Val and D. J. Randall), pp.
111
-121. Manaus: INPA.
Gonzalez, R. J., Dalton, V. M. and Patrick, M. L.(
1997
). Ion regulation in ion-poor, acidic water by the blackskirt tetra (Gymnocorymbus ternetzi), a fish native to the Amazon River.
Physiol. Zool.
70
,
428
-435.
Gonzalez, R. J. and Dunson, W. A. (
1989
). Acclimation of sodium regulation to low pH and the role of calcium in the acid-tolerant sunfish Enneacanthus obesus.
Physiol. Zool.
62
,
977
-992.
Gonzalez, R. J. and Preest, M. (
1999
). Ionoregulatory specializations for exceptional tolerance of ion-poor, acidic waters in the neon tetra (Paracheirodon innesi).
Physiol. Biochem. Zool.
72
,
156
-163.
Gonzalez, R. J. and Wilson, R. W. (
2001
). Patterns of ion regulation in acidophilic fish native to the ion-poor acidic Rio Negro.
J. Fish Biol.
58
,
1680
-1690.
Gonzalez, R. J., Wilson, R. W., Wood, C. M., Patrick, M. L. and Val, A. L. (
2002
). Diverse strategies for ion regulation in fish collected from the ion-poor, acidic Rio Negro.
Physiol. Biochem. Zool.
75
,
37
-47.
Gonzalez, R. J., Wood, C. M., Wilson, R. W., Patrick, M. L.,Bergman, H. L., Narahara, A. and Val, A. L. (
1998
). Effects of water pH and calcium concentration on ion balance in fish of the Rio Negro,Amazon.
Physiol. Zool.
71
,
15
-22.
Goss, G. G., Perry, S. F., Wood, C. M. and Laurent, P.(
1992
). Mechanisms of ion and acid—base regulation at the gills of freshwater fish.
J. Exp. Zool.
263
,
143
-159.
Goss, G. G. and Wood, C. M. (
1990
). Na+ and Cl- uptake kinetics, diffusive effluxes and acidic equivalent fluxes across the gills of rainbow trout. I. Responses to environmental hyperoxia.
J. Exp. Biol.
152
,
521
-547.
Griffith, R. W., Pang, P. K. T., Srivastava, A. K. and Pickford,G. E. (
1973
). Serum composition of freshwater stingrays(Potamotrygonidae) adapted to fresh and dilute sea water.
Biol. Bull.
144
,
304
-320.
Hobe, H., Laurent, P. and McMahon, B. R.(
1984
). Whole body calcium flux rates in freshwater teleosts as a function of ambient calcium and pH levels: a comparison between the euryhaline trout, Salmo gairdneri, and stenohaline bullhead, Ictalurus nebulosus.
J. Exp. Biol.
113
,
237
-252.
Johnson, M. R. and Snelson, F. F. (
1996
). Reproductive life history of the Atlantic stingray, Dasyatis sabina,(Pisces, Dasyatidae) in the freshwater St. Johns River, Florida.
Bull. Mar. Sci.
59
,
74
-88.
Junqueira, L. C. U., Hoxter, G. and Zago, D.(
1968
). Observations on the biochemistry of fresh water rays and dolphin blood serum.
Rev. Bras. Pesquisas Med. Biol.
1
,
225
-226.
Jurss, K. and Bastrop, R. (
1995
). The function of mitochondria-rich cells (chloride cells) in teleost gills.
Rev. Fish Biol. Fish.
5
,
235
-255.
Kerstetter, T. H. and Kirschner, L. B. (
1972
). Active chloride transport by the gills of rainbow trout (Salmo gairdneri).
J. Exp. Biol.
56
,
263
-272.
Kerstetter, T. H. and Kirschner, L. B. (
1974
). HCO3--dependent ATPase activity in the gills of rainbow trout (Salmo gairdneri).
Comp. Biochem. Physiol.
48B
,
581
-589.
Kirschner, L. B. (
1970
). The study of NaCl transport in aquatic animals.
Am. Zool.
10
,
365
-376.
Kirschner, L. B. (
1997
). Extrarenal mechanisms in hydromineral metabolism and acid—base regulation in aquatic vertebrates. In
The Handbook of Physiology
(ed. W. H. Dantzler), pp.
577
-622. Bethesda, MD: The American Physiological Society.
Kirschner, L. B., Greenwald, L. and Kerstetter, T. H.(
1973
). Effect of amiloride on sodium transport across body surfaces of freshwater animals.
Am. J. Physiol.
241
,
832
-837.
Kleyman, T. R. and Cragoe, E. J., Jr (
1988
). Amiloride and its analogs as tools in the study of ion transport.
J. Membr. Biol.
105
,
1
-21.
Lovejoy, N. R. (
1996
). Systematics of myliobatoid elasmobranchs: with emphasis on the phylogeny and historical biogeography of neotropical freshwater stingrays (Potamotrygonidae:Rajiformes).
Zool. J. Linn. Soc.
117
,
207
-257.
Lovejoy, N. R., Bermingham, E. and Martin, A. P.(
1998
). Marine incursion into South America.
Nature
396
,
421
-422.
Maetz, J. (
1974
). Aspects of adaptation to hypo-osmotic and hyper-osmotic environments. In
Biochemical and Biophysical Perspectives in Marine Biology
, vol.
1
(ed. D. C. Malins and J. R. Sargent), pp.
1
-167. London: Academic Press.
Mangum, C. P., Haswell, M. S., Johansen, K. and Towle, D. W.(
1978
). Inorganic ions and pH in the body fluids of Amazon animals.
Can. J. Zool.
56
,
907
-916.
Marshall, W. S., Bryson, S. E., Darling, P., Whitten, C.,Patrick, M., Wilkie, M., Wood, C. M. and Buckland-Nicks, J.(
1997
). NaCl transport and ultrastructure of opercular epithelium from a freshwater-adapted euryhaline teleost, Fundulus heteroclitus.
J. Exp. Zool.
277
,
23
-37.
McCarty, N. A. (
2000
). Permeation through the CFTR chloride channel.
J. Exp. Biol.
203
,
1947
-1962.
McDonald, D. G. and Rogano, M. S. (
1986
). Ion regulation by the rainbow trout, Salmo gairdneri, in ion-poor water.
Physiol. Zool.
59
,
318
-331.
Michal, G. (
1985
). Determination of Michaelis constants and inhibitor constants. In
Methods of Enzymatic Analysis
, vol.
1
(ed. H. U. Bergmeyer),pp.
86
-103. Weinheim: Verlag Chemie.
Mondzac, A., Ehrlich, G. E. and Seegmiller, J. E.(
1965
). An enzymatic determination of ammonia in biological fluids.
J. Lab. Clin. Med.
66
,
526
-531.
Nelson, I. D., Potts, W. T. W. and Huddart, H.(
1997
). The use of amiloride to uncouple branchial sodium and proton fluxes in the brown trout, Salmo trutta.
J. Comp. Physiol. B
167
,
123
-128.
Nemenyi, P., Dixon, S. K., White, N. B. and Hedstrom, M. L.(
1977
).
Statistics from Scratch
. San Francisco: Holden-Day.
O'Donnell, M. J., Kelly, S. P., Nurse, C. A. and Wood, C. M.(
2001
). A maxi Cl- channel in cultured pavement cells from the gills of the freshwater rainbow trout Oncorhynchus mykiss.
J. Exp. Biol.
204
,
1783
-1794.
Pang, P. K. T., Griffith, R. W. and Atz, J. W.(
1977
). Osmoregulation in elasmobranchs.
Am. Zool.
17
,
365
-377.
Patrick, M. L. and Wood, C. M. (
1999
). Ion and acid—base regulation in the freshwater mummichog (Fundulus heteroclitus): a departure from the standard model for freshwater teleosts.
Comp. Biochem. Physiol.
122A
,
445
-456.
Payan, P. and Maetz, J. (
1973
). Branchial sodium transport mechanisms in Scyliorhinus canicula: evidence for Na+/NH4+ and Na+/H+exchanges and for a role of carbonic anhydrase.
J. Exp. Biol.
58
,
487
-502.
Perry, S. F. (
1997
). The chloride cell:structure and function in the gills of freshwater fishes.
Annu. Rev. Physiol.
59
,
325
-347.
Perry, S. F., Haswell, M. S., Randall, D. J. and Farrell, A. P. (
1981
). Branchial ionic uptake and acid—base regulation in the rainbow trout, Salmo gairdneri.
J. Exp. Biol.
92
,
289
-303.
Perry, S. F. and Laurent, P. (
1989
). Adaptational responses of rainbow trout to lowered external NaCl concentration: contribution of the branchial chloride cell.
J. Exp. Biol.
147
,
147
-158.
Perry, S. F., Payan, P. and Girard, J. P.(
1984
). Effect of perfusate HCO3- and PCO2 on chloride uptake in perfused gills of rainbow trout (Salmo gairdneri).
Can. J. Fish. Aquat. Sci.
41
,
1768
-1773.
Perry, S. F. and Randall, D. J. (
1981
). Effects of amiloride and SITS on branchial ion fluxes in rainbow trout, Salmo gairdneri.
J. Exp. Zool
.
215
,
225
-228.
Piermarini, P. M. and Evans, D. H. (
1998
). Osmoregulation of the Atlantic stingray (Dasyatis sabina) in freshwater Lake Jesup of the St. Johns River, Florida.
Physiol. Zool.
71
,
553
-560.
Piermarini, P. M. and Evans, D. H. (
2000
). Effects of environmental salinity on Na+/K+-ATPase in the gills and rectal gland of a euryhaline elasmobranch (Dasyatis sabina).
J. Exp. Biol.
203
,
2957
-2966.
Piermarini, P. M. and Evans, D. H. (
2001
). Immunochemical analysis of the vacuolar proton-ATPase B-subunit in the gills of a euryhaline stingray (Dasyatis sabina): effects of salinity and relation to Na+/K+-ATPase.
J. Exp. Biol.
204
,
3251
-3259.
Potts, W. T. W. (
1994
). Kinetics of sodium uptake in freshwater animals: a comparison of ion-exchange and proton pump hypotheses.
Am. J. Physiol.
266
,
R315
-R320.
Rahmatullah, M. and Boyde, T. R. (
1980
). Improvements in the determination of urea using diacetyl monoxime; methods with and without deproteinization.
Clin. Chim. Acta
107
,
3
-9.
Raschi, W. and Mackanos, L. A. (
1989
). The structure of the Ampullae of Lorenzini in Dasyatis garouaensis and its implications on the evolution of the freshwater electroreceptive system.
J. Exp. Zool.
Suppl. 2
,
101
-111.
Reuss, L., Costantin, J. L. and Bazile, J. E.(
1987
). Diphenylamine-2-carboxylate blocks Cl-—HCO3- exchange in Necturusgallbladder epithelium.
Am. J. Physiol.
253
,
C79
-C89.
Shaw, J. (
1959
). The absorption of sodium ions by the crayfish, Astacus pallipes Lereboullet. I. The effect of external and internal sodium concentrations.
J. Exp. Biol.
36
,
126
-144.
Shaw, J. (
1960
). The absorption of chloride ions by the crayfish, Astacus pallipes Lereboullet.
J. Exp. Biol.
37
,
557
-572.
Shuttleworth, T. J. (
1988
). Salt and water balance — extrarenal mechanisms. In
Physiology of Elasmobranch Fishes
(ed. T. J. Shuttleworth), pp.
171
-199. Berlin: Springer-Verlag.
Smith, N. F., Talbot, C. and Eddy, F. B.(
1989
). Dietary salt intake and its relevance to ionic regulation in freshwater salmonids.
J. Fish Biol.
35
,
749
-753.
Thorson, T. B. (
1967
). Osmoregulation in fresh-water elasmobranchs. In
Sharks, Skates, and Rays
(ed. P. W. Gilbert, R. F. Mathewson and D. P. Rall), pp.
265
-270. Baltimore: Johns Hopkins University Press.
Thorson, T. B. (
1970
). Freshwater stingrays, Potamotrygon spp.: failure to concentrate urea when exposed to a saline medium.
Life Sci.
9
,
893
-900.
Thorson, T. B., Cowan, C. M. and Watson, D. E.(
1967
). Potamotrygon spp.: elasmobranchs with low urea content.
Science
158
,
375
-377.
Thorson, T. B., Wooton, R. M. and Georgi, T. A.(
1978
). Rectal gland of freshwater stingrays, Potamotrygon spp. (Chondrichthyes: Potamotrygonidae).
Biol. Bull.
154
,
508
-516.
Verdouw, H., van Echteld, C. J. A. and Dekkers, E. M. J.(
1978
). Ammonia determinations based on indophenol formation with sodium salicylate.
Water Res.
12
,
399
-402.
Wilkie, M. P. (
1997
). Mechanisms of ammonia excretion across fish gills.
Comp. Biochem. Physiol
.
118A
,
39
-50.
Wilson, R. W. (
1996
). Ammonia excretion in fish adapted to an ion-poor environment. In
Physiology and Biochemistry of the Fishes of the Amazon
(ed. A. L. Val, V. M. Almeida-Val and D. J. Randall), pp.
123
-138. Manaus:INPA.
Wilson, R. W., Wood, C. M., Gonzalez, R., Patrick, M. L.,Bergman, H. L., Narahara, A. and Val, A. L. (
1999
). Ion and acid—base balance in three species of Amazonian fish during gradual acidification of extremely soft water.
Physiol. Biochem. Zool.
72
,
277
-285.
Wilson, R. W., Wright, P. M., Munger, S. and Wood, C. M.(
1994
). Ammonia excretion in rainbow trout (Oncorhynchus mykiss) and the importance of gill boundary layer acidification: lack of evidence for Na+/NH4+ exchange.
J. Exp. Biol.
191
,
37
-58.
Wood, C. M. (
1988
). Acid—base and ionic exchanges at gills and kidney after exhaustive exercise in the rainbow trout.
J. Exp. Biol.
146
,
461
-481.
Wood, C. M. and Randall, D. J. (
1973
). Sodium balance in the rainbow trout (Salmo gairdneri) during extended exercise.
J. Comp. Physiol.
82
,
235
-256.
Wood, C. M., Wheatly, M. G. and Hobe, H.(
1984
). The mechanisms of acid—base and ionoregulation in the freshwater rainbow trout during environmental hyperoxia and subsequent normoxia. III. Branchial exchanges.
Respir. Physiol
,
55
,
175
-192.
Wood, C. M., Wilson, R. W., Gonzalez, R. J., Patrick, M. L.,Bergman, H. L., Narahara, A. and Val, A. L. (
1998
). Responses of an Amazonian teleost, the tambaqui (Colossoma macropomum) to low pH in extremely soft water.
Physiol. Zool.
71
,
658
-670.
Wright, P. A. and Wood, C. M. (
1985
). An analysis of branchial ammonia excretion in the freshwater rainbow trout:effects of environmental pH change and sodium uptake blockade.
J. Exp. Biol.
114
,
329
-353.
Zall, D. M., Fisher, M. and Garner, M. Q.(
1956
). Photometric determination of chlorides in water.
Anal. Chem.
28
,
1665
-1668.