Effects of Water pH and Calcium Concentration on Ion Balance in Fish of the Rio Negro, Amazon

We examined the effects of acute low‐pH exposure on ion balance (Na+, Cl−, K+) in several species of fish captured from the Rio Negro, a dilute, acidic tributary of the Amazon. At pH 5.5 (untreated Rio Negro water), the four Rio Negro species tested (piranha preta, Serrasalmus rhombeus; piranha branca, Serrasalmus cf. holandi; aracu, Leporinus fasciatus; and pacu, Myleus sp.) were at or near ion balance; upon exposure to pH 3.5, while Na+ and Cl− loss rates became significant, they were relatively mild. In comparison, tambaqui (Colossoma macropo‐mum), which were obtained from aquaculture and held and tested under the same conditions as the other fish, had loss rates seven times higher than all the Rio Negro species. At pH 3.0, rates of Na+ and Cl− loss for the Rio Negro fish increased three‐ to fivefold but were again much less than those observed in tambaqui. Raising water Ca2+ concentration from 10 mmol L−1 to 100 mmol L−1 during exposure to the same low pH's had no effect on rates of ion loss in the three species tested (piranha preta, piranha branca, aracu), which suggests that either they have such a high branchial affinity for Ca2+ that all sites are saturated at 10 mmol L−1 and additional Ca2+ had no effect, or that Ca2+ may not be involved in regulation of branchial ion permeability. For a final Rio Negro species, the cardinal tetra (Paracheirodon axelrodi), we monitored body Na+ concentration during 5 d of exposure to pH 6.0, 4.0, or 3.5. These pH's had no effect on body Na+ concentration. These data together suggest that exceptional acid tolerance is a general characteristic of fish that inhabit the dilute acidic Rio Negro and raise questions about the role of Ca2+ in regulation of branchial ion permeability in these fish.

of the paracellular tight junctions, which is dependent on Ca 2/ testing. They were then transported up the river to the experimental site on board the Amanai II. They were kept in fiberglass binding to the membrane-bound junctional proteins (Hunn 1985;Madara 1988). When ambient Ca 2/ levels and/or pH are tanks filled with Rio Negro water during transport up the river and at the experimental site for a total of about 48 h before low, Ca 2/ is leached from tight junctions and branchial ion permeability and ion loss rates rise. Together, the inhibition the start of experiments. The fish were not fed during the time they were held. Cardinal tetras (n Å 44, mean wet mass Å 0.157 of ion uptake and stimulation of efflux results in a net loss of ions, and if the rate of loss is too high or the total amount { 0.008 g) that had been collected from the upper reaches of the Rio Negro were obtained from a commercial dealer in lost too great (around 50%), then serious, potentially fatal internal ionic and coupled fluid balance disturbances result Manaus and were held in Manaus groundwater for a few days before the start of the test. (Milligan and Wood 1982).
Despite these many challenges to ion regulation, the dilute, acidic waters of the Rio Negro still support an incredible diver-Experimental Protocol sity of fish. Recent estimates indicate that over 1,000 different species inhabit the Rio Negro (Val and Almeida-Val 1995). To Net Na / flux, net Cl 0 flux, and net K / flux were measured in unmodified water from the river (pH 5.5) as a control and better understand how these fish are able to maintain ion balance (Na / , Cl 0 , K / ) and inhabit the waters of the Rio Negro, during serial 1-h exposures to river water acidified to pH 3.5 and 3.0. In a second series, on a separate group of fish, the same we examined the ionoregulatory ability during acute low-pH exposure of four species collected directly from the Rio Negro protocol was used except that the water Ca 2/ concentration was raised to 100 mmol L 01 (from 10 mmol L 01 ) after the initial in the region of the Anavilhanas archipelago: piranha preta (Serrasalmus rhombeus), piranha branca (Serrasalmus cf. ho-control flux in unmodified water. At the end of the flux measurements, the water Ca 2/ concentrations were confirmed with landi), pacu (Myleus sp.), and aracu (Leporinus fasciatus), and, for comparison, a fifth species obtained from aquaculture in the flame photometer, and in all cases the actual concentration was within 3 -5 mmol L 01 of the nominal 100 mmol L 01 concen-Manaus (tambaqui, Colossoma macropomum). These species were chosen because they were of good size, collected in ade-tration.
To make the measurements, six fish of each species were quate number, and represented some of the diversity of the species-rich Rio Negro. In addition, we examined the role that placed into individual 3.5-L chambers (water in each chamber was aerated) connected to a 100-L recirculating system filled water Ca 2/ concentration played in the severity of the ion disturbances experienced during low-pH exposure by these with river water and allowed to recover overnight. Flow rate into each chamber was approximately 500 mL min 01 . At the species. Finally, we examined the cardinal tetra (Paracheirodon axelrodi), which was collected from the upper reaches of the beginning of a measurement period, flow was stopped to all containers, and a 20-mL water sample was removed. One hour Rio Negro, for low-pH tolerance by exposing it to pH 4.0 or 3.5 for 5 d and measuring body Na / concentration. later another 20-mL water sample was removed and water flow was restored. The pH was lowered to 3.5 with concentrated H 2 SO 4 , and after a 30-min exposure another measurement period was started. After the completion of the second mea-Material and Methods surement period, the water flow was restored, the pH was Experimental Animals lowered to 3.0, and after 30 min the final measurements were made. At the conclusion of the third flux period the fish were Piranha preta (n Å 9, mean wet mass { SE Å 231.1 { 32.1 g), piranha branca (n Å 11, mean wet mass Å 176.3 { 16.5 removed, weighed, and released to the river. The second series (on a new group of fish) was performed exactly as the first, g), pacu (n Å 6, mean wet mass Å 205.2 { 21.6 g), and aracu (n Å 10, mean wet mass Å 236.3 { 14.8 g) were collected by except that as the pH was lowered to 3.5, the water Ca 2/ concentration was raised to 100 mmol L 01 by addition of seine net from the Rio Negro in the Anavilhanas archipelago and held on board the research vessel Amanai II. The fish were CaSO 4 . During the low-pH exposures, water pH was monitored continuously with an Orion model 250A pH meter and allowed to recover overnight in large fiberglass tanks supplied on a flow-through basis with water pumped directly from the adjusted as needed. The pH of the bathwater was kept within 0.05 units of the nominal pH. At the end of the flux periods, river. The river water was analyzed with a flame photometer and found to have the following ion concentrations (in mmol the pH of the water in each test chamber was checked and found to have risen less than 0.05 units in virtually all cases. L 01 ): Na / , 52; K / , 27; Ca 2/ , 10, and Cl 0 , 55 (pH Å 5.5, temperature Å 30ЊC). Tambaqui (n Å 6, mean wet mass Å 160.8 The water samples were analyzed for Na / , K / , and Ca 2/ concentrations with a flame photometer. Water Cl 0 concentra-{ 9.9 g) were supplied by the National Institute for Amazon Research aquaculture station in Manaus and held in Manaus tion was determined using a colorimetric assay (Zall et al. 1956). Net ion fluxes (J Ion net ) were calculated from the changes groundwater (in mmol L 01 : Na / , 15; K / , 9; Ca 2/ , 10; and Cl 0 , 16; pH Å 6.0, temperature Å 30ЊC) for several weeks before in the ion concentration of the bathwater over the 1-h period 9g11$$ja18 12-23-97 08:03:00 pza UC: PHYS ZOO for all ions (Figs. 1 -5), which indicates that they had recovered from any stress induced by placement into the test chambers. using the following equation: Upon exposure to pH 3.5, tambaqui obtained from aquaculwhere [Ion] 1 and [Ion] 2 are the bath ion concentrations at the ture experienced rates of Na / and Cl 0 loss 12 -13 times greater beginning and end of the flux period, respectively, V is the than at pH 5.5; rates of K / loss were five times greater (Fig. bath volume in liters, M is the mass of the fish in grams, and 1). In sharp contrast, at pH 3.5, pacu, piranha preta, piranha t is the duration of the flux period in hours.
branca, and aracu from the Rio Negro all experienced only In Manaus, the cardinal tetras were divided among three 7mildly elevated net losses of Na / and Cl 0 , averaging about 300 L containers filled with aerated Manaus groundwater. After a nmol g 01 h 01 (Figs. 2 -5; drawn to the same scale as tambaqui day to recover from the transfer, two fish were removed from to facilitate comparisons). Further, while net K / flux was negaeach of the three containers, weighed, placed in individual tive in all four Rio Negro species at pH 3.5, it was unchanged beakers, and dried in an oven at 90ЊC. While the pH of one relative to pH 5.5 measurements. For the three species tested container was maintained at 6.0, the pH of one of the re-(piranha preta, piranha branca, and aracu), raising the water maining two was lowered to 4.0 and the other to 3.5 with Ca 2/ concentration from 10 to 100 mmol L 01 with exposure dilute H 2 SO 4 . After 1 and 5 d, five to seven fish were removed to pH 3.5 had no significant effect on any ion loss rates (Figs. from each container and processed as before. When the expo-3 -5). sure was completed, the dried fish were dissolved in concen-When the pH was dropped to pH 3.0, rates of Na / and Cl 0 trated analytical grade HNO 3 , and the resulting liquid was loss for tambaqui rose 78% and 48%, respectively, relative to diluted and analyzed for whole body Na / concentration with rates at pH 5.5 (Fig. 1). In the Rio Negro species, ion loss rates the flame photometer.
rose as well (Figs. 2 -5), but they were still much lower than rates for tambaqui. However, some differences among the four Statistical Analyses other species became apparent. Piranha preta experienced the smallest net loss of Na / and Cl 0 (net Na / and Cl 0 fluxes were All data are reported as means { 1 SE. Means were compared three and two times greater than at pH 3.5, respectively). The using paired t-tests or ANOVA (overall P°0.05) with multiple other three Rio Negro species all had net Na / and Cl 0 loss comparisons (Scheffé test) if the ANOVA proved significant.
rates about 60% -70% higher than piranha preta. Along with the greatly stimulated Na / and Cl 0 losses at pH 3.0, all four Results species from the Rio Negro experienced increased rates of K / At the beginning of each of the two series of exposures for loss. In these cases, the magnitude of net K / flux rose by about 50%. As at pH 3.5, increased water Ca 2/ concentration had each species, net ion fluxes were measured in untreated Rio 9g11$$ja18 12-23-97 08:03:00 pza UC: PHYS ZOO so it is unclear whether the greater sensitivity in the present Asterisks indicate significant differences from corresponding fluxes at pH 5.5. study was due to the aquacultural origin of the stock tested or reflected a general characteristic of the species.
At pH 3.0, although all species experienced significant ion little effect on rates of Na / and Cl 0 loss at pH 3.0. Only aracu disturbances, the results still indicate great acid tolerance in exhibited a significant reduction in ion losses when water Ca 2/ the Rio Negro species. On exposure to pH 3.0, the Rio Negro concentration was raised to 100 mmol L 01 ; rates of Na / and fish experienced a three-to fivefold increase in net Na / and Cl 0 loss dropped by 50% (Fig. 5). Interestingly, when the water Cl 0 losses relative to fluxes at pH 3.5. Further, the doubling Ca 2/ concentration was raised, piranha preta appeared to lose of K / loss rates in each species indicated that they were experieven more K / (Fig. 3).
encing an internal osmotic imbalance, since K / is lost primarily Cardinal tetras that were held at pH 4.0 or 3.5 for 5 d from intracellular pools. However, the rates of Na / and Cl 0 appeared to be unaffected by the exposure (Fig. 6). Body Na / loss in these species were only one-half to one-fourth as great concentration of fish held at the two low pH's did not change as that observed in tambaqui and were also less than those significantly over the 5 d relative to those held at pH 6.0. observed in other species at less severe pH's. For example, pacu at pH 3.0 (the species from the Rio Negro with the highest Discussion rates of ion loss at that pH) lost Na / at only one-third the rate of common shiners (Notropis cornutus) and two-thirds the Our flux measurements clearly show that the five species from rate of rainbow trout (Oncorhynchus mykiss) at pH 4.0 (Freda the Rio Negro tested here were exceptionally tolerant of low and McDonald 1988). pH. All four large species experienced only mild disruptions The fifth Rio Negro species examined, the cardinal tetra, of Na / and Cl 0 balance during the first hour of exposure was also very tolerant of low pH. During 5 d of exposure to (when the disturbance is usually greatest) to pH 3.5 and no pH 4.0 or 3.5, they did not experience any drop of body Na / disruption of K / balance, an indicator of serious internal osconcentration, which indicates that they were able to maintain motic disturbances (McDonald et al. 1980;Milligan and Wood ion balance at those pH's. Similarly, the acid-tolerant banded 1982; Audet and Wood 1988). The magnitude of the ion distursunfish does not experience a depression of body Na / concenbances observed in the Rio Negro fish at pH 3.5 was similar tration at pH 4.0 (Gonzalez and Dunson 1987). However, while to that observed in the banded sunfish (Enneacanthus obesus), sunfish can survive extended periods at pH 3.5, they lose about an acid-tolerant species native to North America that can sur-30% of their body Na / concentration during the first 2 wk of vive indefinitely at this pH Dunson 1987, 1989). In contrast, tambaqui, which were cultured in Manaus, experi-exposure (Gonzalez and Dunson 1987). It would seem, then, 9g11$$ja18 12-23-97 08:03:00 pza UC: PHYS ZOO Figure 3. Effects of water pH and Ca 2/ concentration on net Na / Negro species tested here, in contrast to previous findings, flux (J Na net ), net Cl 0 flux (J Cl net ), and net K / flux (J K net ) of piranha Ca 2/ may not be involved to any great extent in regulation of preta (Serrasalmus rhombeus) from the Rio Negro. Fluxes were branchial permeability at low pH. In our tests, raising the Ca 2/ measured during the first hour of exposure to each low pH. Values concentration 10-fold had little or no effect on rates of ion are means { SE. The y-axis is drawn with the same scale as Figure 1 to facilitate comparisons. Asterisks indicate significant loss in the three species tested. This insensitivity of ion losses differences from corresponding fluxes at pH 5.5. to water Ca 2/ concentration suggests the possibility of a novel mechanism for the control of paracellular tight junction permeability, one that does not involve Ca 2/ . that the cardinal tetra's ability to maintain ion balance at low Of course, given the way our experiments were performed, pH is superior to that of the banded sunfish.
several alternative explanations cannot be ruled out. For in-It should not be surprising that these fish exhibit such a stance, it is possible that an actual reduction in diffusive efflux high degree of tolerance to low pH given the extremely ionin water with increased Ca 2/ concentration was masked by a poor nature of their native waters and the extremely low pH simultaneous drop in active uptake. Since we measured only found in some locations. The question is, How do they do it?
net Na / and Cl 0 fluxes and not unidirectional movement (i.e., Numerous studies on several North American species have radioisotope fluxes; legal restrictions prevented the use of rashown that the key to tolerance of dilute waters of low pH is dioisotopes in this project) of Na / and Cl 0 across the gills, we the ability to avoid increased branchial ion permeability. Elecannot determine the response of the unidirectional fluxes to vated gill permeability, which, evidence suggests, is caused by added Ca 2/ . However, given the overall elevated rates of ion displacement by H / of Ca 2/ from paracellular tight junction loss, such masking seems unlikely. proteins (see review by Wood [1988]), leads to increased diffu-Another possibility is that the Rio Negro fish possess such sive ion losses. This claim is bolstered by experiments in which a high branchial affinity for Ca 2/ that tight junction binding addition of Ca 2/ to test water causes a reduction or even an sites are saturated even in 10 mmol L 01 water. This seems elimination of elevated rates of ion losses at low pH (McDonald reasonable, since these fish inhabit waters with typical Ca 2/ et al. 1980; McDonald and Wood 1981;McDonald 1983). Exconcentrations that are 10 mmol L 01 or less. If this is the tending this concept, it has been proposed that the basis of case, then raising water Ca 2/ concentration would have no tolerance to low pH is an increased branchial affinity for Ca 2/ , additional effect on branchial permeability or ion balance. A which resists displacement at low pH (Hunn 1985;Gonzalez similar conclusion was reached with the acid-tolerant banded and Dunson 1987Dunson , 1989McDonald et al. 1991). Indeed, several sunfish and yellow perch. For the sunfish, raising water Ca 2/ studies have indicated a very high branchial affinity for Ca 2/ concentration from 0 to 50 mmol L 01 at pH 3.25 causes Na / in fish that are tolerant of low pH (McWilliams 1982; losses to drop by more than 50%, but additional Ca 2/ produces and McDonald 1988;Gonzalez and Dunson 1989), including no further reductions (Gonzalez and Dunson 1989). Na / losses a study of an Amazonian species (Gonzalez et al. 1997).
of yellow perch ( Perca flavescens ) were found to be insensitive While our results support the concept that control of ion efflux is key for survival, they also indicate that for the Rio to water Ca 2/ concentration at a variety of pH's, although the 9g11$$ja18 12-23-97 08:03:00 pza UC: PHYS ZOO and Na / uptake ceases. However, if Rio Negro fish can survive extended periods (perhaps their whole life) at very low pH's (°4.5), it seems likely that they can actively transport salts at lowest concentration tested, 50 mmol L 01 , may have been too these pH's. A recent study provides some support for this high to detect any effect (Freda and McDonald 1988).
notion. The blackskirt tetra experiences partial inhibition of One final, interesting alternative is that the organic com-Na / uptake at pH 4.0, but at pH 4.5, when loss rates are pounds that give the Rio Negro its tea color (and its name) stimulated, they exhibit a rapid stimulation of Na / uptake somehow play a role in eliminating the effects of water Ca 2/ (Gonzalez et al. 1997). Interestingly, support also comes from concentration. Furch (1984) found an average total carbon studies of some North American and European fish. Yellow concentration in the Rio Negro of 10.5 mg L 01 . Given this perch and brown trout (Salmo trutta), for instance, are able to large quantity of organic compounds in the river, it is possible take up salts at pH 4.0 (McWilliams 1982; Freda and McDonald that they could interact directly with the branchial tight junc-1988), which is well below the proposed theoretical limit. It tions and influence permeability. Dissolved organic comseems we must begin to reevaluate our model to make it consispounds have been shown to bind to North American amphibtent with these results. ian egg membranes and prevent hatching (Karns 1983; Finally, the exceptional acid tolerance of all five species et al. 1989). It also seems possible that the organic compounds tested here suggests that this tolerance is a general charactercould effectively bind to Ca 2/ in the water, rendering it unavailistic of fishes of the Rio Negro (Dunson et al. 1977). Given able to the fish (Freda et al. 1989), and promote the developthat it is estimated that over 1,000 species from many differment of a new mechanism to limit branchial permeability. ent families inhabit the Rio Negro, there may be a variety Further studies are needed to identify the possible involvement of different patterns of ion regulation in these fish. Further of dissolved organic substances in ion regulation of Rio Negro studies of these species are likely to yield novel mechanisms fish.
of ion regulation. While our results raise questions concerning the role of water Ca 2/ in regulating branchial ion permeability in Rio Negro fish, they also prompt questions about other aspects of Acknowledgments models of ion regulation that have been developed in recent years from studies of North American and European species. This work was supported by a National Science and Engineering Research Council research grant to C.M.W. R.W.W. For example, Lin and Randall (1995) argue, on the basis of their examination of proton pumps in trout, that Na / uptake was supported by a Royal Society Research Grant. A.L.V. was 9g11$$ja18 12-23-97 08:03:00 pza UC: PHYS ZOO Chico Buarque and Jorge Ben for their excellent technical assistance.