ABSTRACT
For the first time in a marine teleost (the long-horned sculpin; Myoxocephalus octodecimspinosus), the maintenance of blood pH, , [HCO3−] and the net movements of NH4+, HCO3− and H+ between the fish and the water have been studied during exposure to ammonia stress induced either by infusion (NH4C1 or NH4HCO3; 5 mmol kg−1) or by external application (NH4C1; approx. 1 mmol I−1).
Following NH4C1 infusion, a rapid decrease in blood pH (0·36units) and [HCO3−] (2·38 mmol I−1) was observed, and within 1h about 40% of the ammonia load had been excreted to the water. Analysis of NH4+ and HCO3− transfers revealed that the total ammonia (TAmm) efflux was due to a loss of NH3 and NH4+ in approximately equal proportions when an outwardly directed NH3 diffusion gradient was established.
Infusion of NH4HCO3 induced only small changes in plasma pH, and the rate of net HCO3− excretion was some 90% higher than that of NH4+ over 20 h. These data indicate a predominance of NH3 as the form of ammonia lost. In both infusion experiments, a presumed intracellular buffering of a majority of the ammonia load was noted.
High external TAmm induced an initial uptake of NH4+, but after 4 h of exposure ammonia efflux resumed even though NH3 diffusion gradients were negligible. Thus, in this seawater teleost, a role for the excretion of ammonia in the form of NH4+ is also likely.
INTRODUCTION
Teleost fish excrete nitrogenous waste products in the form of ammonia (NH3 + NH4+). Although it is well accepted that the gills are the major site of this excretion (Smith, 1929; Evans, 1982; see review by Kormanik & Cameron, 1981; Evans & Cameron, 1986), the mechanism of ammonia transfer across the gill epithelium and the form(s) of the transferred ammonia are not as clear. In the freshwater trout (Salmo gairdneri), ammonia loss under control conditions can be accounted for principally by simple diffusive loss of NH3 (Cameron & Heisler, 1983). It has been theorized that branchial transfers such as NH4+/Na+ exchange are only necessary when the animal is exposed to an abnormally high ambient [ammonia] (Cameron & Heisler, 1983). Recently, Cameron (1986) has hypothesized that ammonia excretion in the freshwater catfish (Ictalurus punctatus) can be maintained in the face of reversed NH3 and NH4+ concentration gradients by activating an NH4+/H+ exchange system. Using the isolated perfused head preparation, roles for the ionic diffusion of NH4+ as well as NH4+/Na+ exchange have also been postulated for two seawater teleosts (Opsanus beta and Myoxocephalus octodecimspinosus; Goldstein et al. 1982; Claiborne et al. 1982).
The ratio of NH3 to NH4+ diffusion across the gills is dependent on pH, pK’, temperature and the ionic strength of both the blood and the external environment, as well as the relative gill permeabilities of NH3 and NH4+ (Cameron & Heisler, 1983; Boutilier et al. 1984). Cameron (1986) has calculated that the ratio of NH3 to NH4+ transferred may range from 32:1 to 1:9, depending on the relative permeabilities and blood/water parameters utilized. Though it is likely that diffusive loss of NH3 can account for the majority of ammonia transfer in freshwater teleosts under normal conditions, the sparse data for seawater varieties indicate a role for some of the additional pathways described above. For example, since it is generally thought that the gill ion permeability of seawater-adapted teleosts is more than 10 times that of freshwater fish (Evans, 1984), ionic diffusion of NH4+ could be much more significant in seawater animals. To gain further insight into the relative roles of these transfers in a marine species, we monitored internal acid-base regulation and gill ammonia excretion in the long-horned sculpin (Myoxocephalus octodecimspinosus Mitchell) following exposure to infused ammonia loads (NH4C1 or NH4HCO3) or a reversed external ammonia gradient.
MATERIALS AND METHODS
Long-horned sculpin (Myoxocephalus octodecimspinosus), mass = 165·1 ± 10·4 g (N = 16) (mean ± S.E.) were caught by local fishermen in Frenchman’s Bay, Maine near the Mount Desert Island Biological Laboratory. The animals were maintained in large wooden or fibreglass tanks and supplied with running sea water (13–15°C). Before use, specimens were held for 2–6 days without feeding. To facilitate blood sampling in these relatively small animals, the fish were cannulated in a manner similar to that described for the measurement of ventral aortic blood pressure (Claiborne & Evans, 1981). Each fish was anaesthetized (MS-222,1:10 000), placed in a moist tray, and periodically ventilated with aerated sea water during the 5–10 min operative procedure. The cut tip of a 23 gauge needle, connected to a short length of heparinized, Ringer-filled cannula (PE-50) was inserted into the afferent artery of the third branchial arch and secured in place with a suture. In some animals, an additional Ringer-filled cannula (PE-50) was inserted through the skin and peritoneal musculature into the peritoneal cavity, and secured with a suture through the skin. The animals were then placed in darkened Plexiglas boxes (1·3–2·01) and allowed to recover for 20–48h. During this period, fresh running sea water was directed through the experimental chamber. Several hours prior to the start of each experimental series, the running sea water was disconnected so that control net ion and ammonia fluxes (see below) could be measured. Duplicate control blood samples were also drawn during this period.
Each blood sample (0·2–0·4ml) was analysed for pH (I.L. model 213 or Radiometer ‘gun’ electrode with an Orion 701A pH meter), total CO2 concentration, (Capnicon-Con II; Cameron Instruments Inc.) and plasma total ammonia concentration (TAmm; Sigma kit no. 170-UV). Plasma and [HCO3−] were calculated from and pH using values for CO2 solubility and pK’ derived from Boutilier et al. (1984). Water samples (20 ml) were collected periodically and analysed for TAmm using the phenolhypochlorite method (Solorzano, 1969). The net titratable base was determined by volumetric titration of a portion of the sample to a pH of 3·800 or 3·700 with 0·1 mol I−1 HC1 using a syringe micrometer burette (model SB2, Micro Metric Instrument Co.) according to the methods of Cameron & Kormanik (1982). This method was typically repeatable within ±1 μl of acid (<1%) in a 10ml water sample, thus the overall resolution was approx. 25 μmol in a 1·31 chamber. The ΔHCO3− was then calculated as the difference between the titratable base at the beginning and end of each time interval. Mucus, proteins and other buffers excreted by the animal would cause an overestimation of ΔHCO3−. In some cases, seawater pH was also recorded before acid was added.
Ammonia infusion series
After all control measurements had been made, animals were infused intraperitoneally with either NH4C1 or NH4HCO3 (5 mmol kg−1, a 3–5 ml bolus of a 200 mmol I−1 stock solution infused over about 2 min). After a 5-min equilibration, time zero blood and water samples were collected, and additional samples were obtained at 1, 2, 4, 8 and 20 h post-infusion. At 4 and 8h of the experimental period, the water within the box was flushed with sea water to limit the accumulation of external ammonia.
High external ammonia series
Following the control period, the ammonia concentration within the experimental chamber was increased to approx. 1 mmol I−1 by the addition of NH4C1 (approx. 6 ml of 200 mmol I−1 stock). After a 6-h exposure, the high ammonia bath was replaced by normal sea water for an additional 15 h. Blood was drawn at 0·5, 1,2,4 and 6 h of the high external ammonia (HEA) exposure, and at 1 and 15 h of the recovery period. External bath samples were also collected during all perturbations and the water within the box was flushed at 4 h of the recovery period to maintain a relatively low external [TAmm].
Calculations
ΔHCO3− and ANH4+ (mmol kg−1) were calculated for all time periods by multiplying the measured concentration of each ion by the volume of the experimental bath, and adjusting for volume changes due to sampling and the mass of the animal (TAmm is effectively [NH4+] in sea water since the pK’ of the NH3/NH4+ equilibrium is about 9·6; Cameron & Heisler, 1983). The total amount of H+ transferred between the sculpin and the water (ΔH+) is therefore the difference between ΔNH4+ and ΔHCO3− (for details see Claiborne & Heisler, 1984, 1986; Heisler, 1984). ‘Net Δ’ values are the differences between the experimental and control rate of transfers for each time period. Data analysis was performed on a microcomputer (Franklin 1200 or Apple IIe) and Student’s t-test (one-or two-tailed) was applied where appropriate.
RESULTS
Control measurements
Pooled control values for acid-base parameters and ion transfers are listed in Table 1. The calculated ΔH+ during control conditions was due to a measured HCO3− uptake (20%) and a more significant TAmm excretion (80%).
NH4 Cl infusion
As shown in Fig. 1A, intraperitoneal infusion of NH4C1 induced an increase in plasma TAmm, which rose to approx. 5400 μmol I−1 immediately following the infusion and then returned to levels near control values within 8h (TAmm control, 234 ± 29μmoll−1; TAmm 8h post-infusion, 314 ± 70μmoll−1; mean ± S.E., N= 5). Fig. 1B–D depicts the concurrent changes in blood acid-base status following the ammonia infusion. Plasma pH decreased from 7·80 ± 0·01 to 7·44 ± 0·06 (P < 0·01, N = 5) and then regained control levels at hour 4. Plasma rose by 74% from 2·3 to 4·0 mmHg (1 mmHg = 133-3 Pa) and remained significantly above pre-infusion values until hour 8, while [HCO3−] was depressed by 2·0mmoll−1 (approx. 32%) after the infusion but had returned to normal at the 4-h sample.
The effects of the infusion on transfers of NH4+, HCO3− and H+ between the fish and the external bath are shown in Fig. 2A–C, respectively. In these and all following ion transfer figures, control lines represent mean ± S.E. of the preliminary control flux period extrapolated over the length of the subsequent experimental period(s). NH4C1 infusion induced a large increase in ammonia efflux from the animal (Fig. 2A). Over the first hour after infusion, TAmm loss increased by about sevenfold from a control rate of 0·29 ± 0·06 mmol kg−1 h−1 to 2·15 ± 0·08 mmol kg−1 h−1 (P<0·001, N = 4). The rate of ammonia transfer remained above the control rate until the final sampling period (hours 8–20). The ‘net’ loss of ammonia (the experimental rate minus the extrapolated control rate) was equal to 3·57 ± 0·25 mmol kg−1 at hour 4 and totalled 4·61 ± 0·54 mmol kg−1 at hour 20. Apparent HCO3− movements between the animal and the water (Fig. 2B) also changed following the infusion: the control uptake rate of 0·11 ± 0·02 mmol kg−1 h−1 was altered to a net efflux of 1·06 ± 0·28 mmol kg−1 h−1 during the first hour after NH4C1 application. HCO3− efflux was maintained up to and during hour 20 (0·14 ± 0·08 mmol kg−1 h−1, hours 1–20). Net HCO3− loss was 1·39 ± 0·38 mmol kg−1 at hour 4 and 5·82 ± 1·64 mmol kg−1 after 20 h (P < 0·05, N = 4). The calculated H+ transfer (see Materials and methods) from fish to environment (Fig. 2C) increased significantly over the first 4h after infusion (control, 0·40 ± 0·08 mmol kg−1 h−1; at hour 4, 0·90 ±0·05 mmol kg−1, P<0·02, N=4), and then decreased 60% below the control ΔH+ by the end of the experiment (0·15 ± 0·03 mmol kg− 1h−1, P<0·05). Net ΔH+ reached a maximum of 2·18 ± 0·48 mmol kg−1 at hour 8, but was not significantly different from 0 by hour 20.
NH4 HCO3 infusion
Infusion of NH4HCO3 (5 mmol kg−1) caused a rapid and significant increase in both plasma TAmm and [HCO3−] (Fig. 3A,D). Plasma TAmm increased from a control value of 0·24 ± 0·07 to 4·62 ± 0·38 mmol I−1 (mean ± S.E., N = 5) immediately after the infusion but had returned to control levels by hour 2 after infusion. Plasma [HCO3−] reached a maximum of 11·25 ±0·45 mmol I−1 after NH4HCO3 application, and then returned to values not different from the pre-infusion control of 4·69 ± 0·26 mmol I−1 by the hour 4 sample. Fig. 3B shows the effect of the infusion on blood pH. The pH increased slightly but significantly at hour 2 (control, 7·78 ± 0·02; hour 2, 7·81 ± 0·01, P<0·05, N= 5), dropped below control at hour 8 (7·73 ±0·01), and then approached pre-infusion levels by the end of the experiment. Plasma (Fig. 3C) increased from 1·80 ±0·13 to 4·69 ± 0·20 mmHg immediately after infusion. This parameter then slowly decreased to control levels by hour 4.
The rate of NH4+ excretion (Fig. 4A) increased by 12-fold during the first hour following the NH4HCO3 load (control, 0·30 ± 0·06 mmol kg−1 h−1; hours 0–1, 3·55 ± 0·43 mmol kg−1 h−1, N = 5) but regained control rates by hour 2. By hour 2, the net ΔNH4+ was 3·76 mmol kg−1 and reached a maximum of 4·42 mmol kg−1 at hour 8. ΔHCO3− also increased rapidly during the first hour post-infusion (Fig. 4B), increasing from a control rate of 0·01 ± 0·04 mmol kg−1h−1 to 3·60 ± 0·35 mmol kg−1 h−1. Though much reduced from the initial hour, HCO3− loss continued through hours 4–8 (0·33 ± 0·06 mmol kg−1 h−1, P < 0·01), resulting in a net ΔHCO3− of 6·29 mmol kg−1 at the end of this period and 7·11 mmol kg−1 after 20 h. ΔH+ transfers between the animal and the water are shown in Fig. 4C. Though variable, ΔH+ remained near zero throughout the experiment. When compared with a control excretion rate of 0·29 ± 0·06 mmol kg−1 h−1, the low experimental values resulted in a significant net H+ uptake of 3·40 mmol kg−1 (P<0·05) over the 20 h post-infusion period.
High external ammonia concentration
High external ammonia concentration (HEA) induced several effects on acid-base balance and ion transfers in the sculpin (N = 6). Plasma pH and appeared to increase slightly, but did not vary significantly from the control measurements of 7·78 ± 0·02 and 5·29 ± 0·55 mmol I−1, respectively (Fig. 5B,C). After 30min in HEA, plasma TAmrn (Fig. 5A) had increased by nearly threefold from a control of 230 ± 30 to 690 ± 103 μmoll−1. TAmm reached a maximum of 839 ± 157 μmol I−1 at hour 6 (but remained below that of the external bath: 1175 ± 43 μmol I−1), and then decreased to 552 μmol I−11 h after normal sea water was reinstated.
A control ammonia efflux (0·271 ± 0·042 mmol kg−1 h−1) was reversed to an influx as a ΔNH4+ of –1·216 ± 0·390 mmol kg−1 was observed during the first hour of HEA (Fig. 6A). This uptake resulted in a calculated net NH4+ gain of approx. 2·0 mmol kg−1 after 4h. From hours 4 to 6, ΔNH4+ returned to an efflux once again (0·442 ± 0·088 mmol kg−1 h−1, P<0·01). During the first 3h of the recovery period, a large net NH4+ loss was observed which equalled the ammonia gained in the preceding period. The rate of NH4+ efflux remained significantly higher than the control (approx. 30%) until the end of the experiment. ΔHCO3− (Fig. 6B) was reversed from a control uptake of 0·10 ±0·05 mmol kg−1 h−1 to an excretion of 0·22 ±0·08 mmol kg−1 h−1 at hour 2 of HEA. During the recovery period, ΔHCO3− remained near zero until the last sampling interval when an efflux of 0·24 ± 0·06mmol kg−1 h−1 was observed. A large but variable net HCO3− loss totalling 4·66 ± 1·14 mmol kg−1 (P<0·01, N =6) was observed during the 15-h recovery period. The measured NH4+ and HCO3− transfers resulted in a negative net ΔH+ (a net base loss; Fig. 6C) of 2·74 ± 0·58 mmol kg−1 from hours 1 to 4 of HEA. ΔH+ then resumed rates similar to or slightly higher than the control AH+ measurement (0·37 ± 0·05 mmol kg−1 h−1) from hours 4 to 10, such that during this interval a net H+ efflux of 1·00 mmol kg−1 occurred. During the last sampling period of the recovery, ΔH+ again decreased to 0·12 ± 0·04mmolkg−1h−1.
To study the gradients driving the movement of NH3 across the gills (Fig. 7), we utilized the measured plasma and water pH and TAmm values as well as appropriate solubility and pK’ constants derived for trout plasma and sea water (Cameron & Heisler, 1983). Under control conditions, a positive (from fish to water) NH3 diffusion gradient of about 60nmHg was measured. Within 1h after exposure to the HEA, however, plasma was not significantly different from that of the surrounding sea water. When the external bath was flushed with normal sea water, plasma decreased (but remained above the control average), and a positive diffusion gradient was established once more. Sea water [NH3] decreased during the first 2h of HE A as ammonia entered the fish, and then significantly increased again during hours 2–6.
DISCUSSION
As seen in Table 1, acid-base parameters for the sculpin are similar to those for other teleosts (Heisler, 1984). Ion transfer rates also appear to be ‘normal’ for a seawater species (Wood et al. 1977; Turner et al. 1983). Although we have no data on the branchial versus renal partitioning of the observed ion movements, it is likely that a large portion of the net transfers were transbranchial (Kormanik & Cameron, 1981; Heisler, 1984). It should be noted that owing to the relatively small size and the particular vascular structure of these animals, we have found it very difficult to cannulate the dorsal aorta using the usual methods (e.g. Soivio et al. 1972; Claiborne & Heisler, 1986) but have utilized a method which we previously developed for this species (Claiborne & Evans, 1981). In contrast to dorsal aortic cannulation, it is likely that mixed venous blood was drawn from the animal when utilizing the present method. Possible disadvantages of our procedure may include disruption of proper water flow through the branchial cavity and ischaemia of the cannulated gill arch. However, the lack of any undue struggling or perturbations in the acid-base status of the animal during control conditions demonstrates that this procedure is acceptable for this species. In about 80% of the cannulations, afferent blood flow did not appear to be impeded (upon visual examination of the gill arch at the termination of an experiment) and blood acid-base balance was maintained even by animals in which gill flow was blocked. In several fish returned to large holding tanks, the cannula remained patent for up to 2 weeks and periodic sampling of blood pH, and indicated that these parameters remained normal over the length of the experiment (J. B. Claiborne & D. H. Evans, unpublished observations).
NH4 Cl infusion
In many aspects, the present data confirm those presented by Cameron & Kormanik (1982) and Cameron & Heisler (1983) for two freshwater-adapted teleosts. We found that infusion of NH4C1 induced a rapid decrease in plasma pH and [HCO3−] and an increase in . The responses are similar to those expected following a metabolic acid load or infusion (Holeton et al. 1983; Turner et al. 1983; Cameron & Kormanik, 1982). The above authors ascribe these effects to the rapid loss of the infused ammonia as NH3 (into the water), with a more gradual equimolar excretion or intracellular buffering of the remaining H+. An alternative explanation for the observed extracellular acidosis could be that the infused ammonia enters metabolic pathways and is thus buffered intracellularly as NH3 (for example with glutamate; L. Goldstein, personal communication) thus again liberating equimolar amounts of H+. The NH3 may then be slowly shuttled back into the blood where it could be extruded along with the remaining protons (as NH3 + H+ and/or NH4+).
By examining the transfers of NH4+, HCO3− and net H+ between the fish and the water, one can attempt to quantify the relative amounts of acid-base exchanges occurring across the gills following the ammonium infusion. As seen in Fig. 2, during the first hour after infusion, a net NH4+ excretion of 2·1 mmol kg−1 in combination with a ΔHCO3− of approx. 1·0 mmol kg−1 resulted in a net H+ loss of approx. 1·1 mmol kg−1. Since any NH3 leaving the animal would create an equivalent amount of HCO3− in the water (the majority of NH3 lost to the sea water would be immediately protonated to NH4+), it appears that, during this first hour, about 48% of the total ammonia excretion was due to the loss of NH3, and the remainder was via the combined loss of NH3 and H+, the direct transfer of NH4+ along with Cl−, or the exchange of NH4+ with external Na+ (see review by Evans & Cameron, 1986). Unfortunately, the transfer of NH3 + H+ is indistinguishable from the movement of NH4+ using present analytical methods. When the control rate of ammonia excretion is taken into account, the calculated net NH4+ efflux is 1·9 mmol kg−1. Thus, about 40% of the infused ammonia had appeared in the water within 1 h. At the same time, a net H+ loss of 0·7 mmol kg−1 amounted to about 15% of the infused load, again an indication of the direct loss of NH3.
Using 1·0 mmol kg−1 as an estimate of the NH3 leaving the fish during the first hour after infusion (and thus the proton load remaining within the animal), an extracellular space estimate of 20% (Cameron, 1980) and a blood buffer value of – 12mequivpHunit−1 (Cameron & Heisler, 1983), if all the proton load were to remain in the extracellular fluids during this period, the plasma pH would have decreased by approximately 0·4 units; in fact, the pH had increased slightly after 1h (Fig. 1B). This is clear, if indirect, evidence that the H+ liberated from the NH4CI was removed from the extracellular space and probably buffered intracellularly. As calculated by Heisler (1980) and Cameron & Kormanik (1982), the relatively large volume and buffering capacity of the intracellular space could absorb this excess H+ without a drastic effect on intracellular pH.
From hours 1 to 4, a shift in the mode of ammonia and H+ transfer is apparent. During this period, the net H+ loss was driven almost completely by the NH4+ (or NH3 + H+) efflux (the net ΔHCO3” was near 0). 1·5 mmol kg−1 (cumulative total, 40%) of H+ was cleared from the animal during this period. Paralleling these transfers was a recovery of plasma pH that was complete by hour 4.
Over the subsequent 16 h of the experiment, a cumulative NH4+ loss of 5·7mmolkg−1 and a ΔHCO3− of 2·8mmol kg−1 indicate that about 47% of the total ammonia excretion was again due to NH3 while the remainder must have resulted from the transfer of NH4+ (or NH3 + H+). The net NH4+ lost after 20 h was 4·61 ± 0·54 mmol kg−1, or 92% of the infused load. In contrast, from hours 4 to 20, net ΔH+ became negative. Owing to this reversal, the total H+ excreted over the entire 20 h was not different from that expected from control fish. In other words, the proton load induced by the infusion of NH4C1 and any non-ionic elimination of NH3 had not been excreted by the animal after 20 h. Paradoxically, 40% of the infusion had been lost by hour 4, but the animals developed a net uptake of H+ (or an efflux of HCO3−) during the next 16 h (Fig. 2C). Cameron & Kormanik (1982) found that after an infusion of NH4C1 the freshwater catfish (Ictalurus punctatus) excreted only a portion of the total ammonia (65%) and H+ (48%) load. They theorized that the remainder of these components was either excreted slowly by the animal or retained in the intracellular compartment. The plasma pH of the sculpin had returned to normal within 4h post-infusion and remained at this level for the next 16 h. Since plasma pH could be maintained even while the animal was not excreting the infused H+, it is again likely that the net H+ gain was sequestered intracellularly. Though we have no data past hour 20, it is possible that the sculpin also excrete the sequestered H+ load over the next few days. If the reversal in H+ transfer is a maladaptive response, perhaps it is due to the artificially high concentrations of internal ammonia to which the animals were exposed following the infusion (about 23 times control). It remains to be seen whether an ammonia ‘overload’ alters the metabolic homeostasis of amino acid catabolism and synthesis of NH3 and H+ within the gill epithelium (Goldstein et al. 1964; Cameron & Heisler, 1983), and thereby indirectly affects intra- and extracellular H+ and HCO3− balance.
Although the modes of ammonia excretion in these fish appear to include the transfer of both NH4+ (or NH3 + H+) and NH3, the data indicate that the ratio of each of these components in the total ammonia lost may vary at different stages of the experiment. During the first hour post-infusion, and again after hour 4, about 50% of ammonia excretion was due to the excretion of NH3, whereas from hours 1 to 4, NH4+ or NH3 + H+ transfer accounted for the entire ammonia lost. The question arises as to why these animals would switch from a passive non-ionic diffusion of NH3 to a presumably carrier-mediated transport of NH4+ or H+ during hours 1–4. A probable answer becomes clear when one considers the ammonia gradients between the fish and the water through the course of the experiment. The observed alterations in the mode of ammonia, transfer may have been due to a build-up of external ammonia in the environmental water following the infusion. The bath was flushed with fresh sea water at hour 4 (and again at hour 8; see Materials and methods), but by hour 4 the external was approx. 95 nmHg and the plasma was approx. 125 nmHg [calculated using solubility and pK’ derived for sea water and trout plasma by Cameron & Heisler, 1983, observed plasma pH values (this study), and a mean water pH of 7·5, J. B. Claiborne & D. H. Evans, unpublished]. Thus, only a small net gradient (30 nmHg) was available to drive diffusion of NH3 during this period. In contrast, the gradient for NH3 efflux immediately following the NH4C1 infusion was approx. 640nmHg (20 versus 660nmHg, water and plasma, respectively). Subsequent to the seawater change at hour 4, the of the water decreased to approx. 9 nmHg, thus creating a larger gradient for the outward diffusion of NH3 once more. That ammonia excretion could continue (Fig. 2A) when the gradient was relatively small, is an indication that these animals can utilize some active form of NH4+ excretion when necessary. This finding led us to perform a separate series of experiments designed to assess further these transfers during high external ammonia exposure (see below).
NH4 HCO3 infusion
The infusion of NH4HCO3 is different from that of NH4C1 (described above), in that NH4HCO3 dissociation will result in the fish receiving equal amounts of an acid (NH4+) and a base (HCO3−). If the animal were to lose the infused ammonia as NH4+, any HCO3− not concurrently excreted should cause an internal base excess and concomitant pH increase. In contrast, if ammonia were transferred as NH3, the remaining H+ would be buffered by the additional HCO3− (leading to an increase in plasma due to the dehydration of HCO3− to CO2 and H2O), and no pH change should be apparent. The present data indicate that both processes may have occurred during the hours subsequent to the infusion. In the first 2h, NH4+ and HCO3− excretion rates were similar (ΔH+ ≈0). Plasma TAmm rapidly increased and then returned to control levels. At the same time, extracellular [HCO3−] was elevated but subsequently did not decrease to the same extent. Both plasma and pH increased significantly during this interval. Since plasma [HCO3−] remained elevated and blood pH increased, one could hypothesize that a portion of the excreted ammonia was lost in the form of NH4+ without a parallel loss of HCO3−. Given the maximum observed pH increase (0·071 units; the difference between hours 0 and 2), an estimate for extracellular space of 20%, and a blood buffer value of – 12mequivpH−1 (see previous section), only about 4% of the observed NH4+ transfer could have been voided in this manner. Thus, the pH changes following the infusion were much less than one would expect if a significant proportion of the ammonia excretion were due to NH4+ transport without the concurrent loss of HCO3−. Likewise, had all ammonia lost during the first few minutes post-infusion been due to NH4+ transfer, the resulting plasma [HCO3−] of 11·25 mmol I−1 (and no change in ) would have led to a metabolic alkalosis with serosal pH values of approx. 8·12 (calculated from using solubility and pK’ values derived from Boutilier et al. 1984). That the pH did not increase during the initial period, concurrent with the observed rapid elevation in plasma , is evidence for a diffusive NH3 transfer component.
As shown in the previous section, extracellular-intracellular transfers of the infused loads can play a role in the regulation of the observed exchanges. For example, the sum of the net excreted ammonia and that measured in the extracellular compartment after 1 h can only account for approx. 3·5 mmol (70%) of the infused load. The remainder was presumably sequestered intracellularly. In contrast, beginning at hour 1, net HCO3− transferred out of the animal plus the extracellular HCO3− increase was equal to or greater than the amount infused. Upon infusion of equal amounts of acid and base, the net H+ lost by the fish might be expected to remain near 0 (no different from control rates). Interestingly, net ΔH+ for the sculpin did not remain constant and, over the 20-h experiment, these animals exhibited a net H+ gain (or a net HCO3− loss) of 3·6 mmol kg−1 due mainly to a 42% ‘overshoot’ in HCO3− loss (approx. 7·1 mmolkg−1 over the 20-h experiment). The maintenance of the pre-infusion plasma [HCO3−] vis-à-vis the net efflux of HCO3− is evidence for a contribution of the intracellular HCO3− pool to the observed net loss of this ion, and could again be due to the ‘ammonia stress’ imposed on the animal (see above).
High external ammonia concentration
When exposed to high external ammonia concentration (HEA) the plasma TAmm of the sculpin increased threefold during the first 30 min of the exposure. This increase was due to a large net ammonia influx (approx 2·0 mmol kg−1; Fig. 6A) which lasted for the first 4h. Importantly, alterations in the rate of ammonia transfer accounted for about 90% of the calculated ΔH+ during the first few hours of HEA. This result, in combination with the negligible effects of the HEA on plasma pH, may suggest that a large portion of the TAmm which entered the animal was in the form of NH4+. Had NH3 been the predominant form taken up by the animal, a reduction in ΔHCO3− and a blood alkalosis might be predicted (Cameron & Heisler, 1983; Cameron, 1986). These results, though in contrast to those described for freshwater trout and catfish, are not unexpected since the ionic permeability of the gills in marine species is thought to be relatively high (Evans, 1979), and these animals are capable of a more rapid recovery from acid-base disturbances (Toews et al. 1983; Claiborne & Heisler, 1984, 1986). Indeed, evidence for a role of ionic NH4+ diffusion across the gills of the sculpin has been described previously (Goldstein et al. 1982). Interestingly, NH4+ transfer changed to an efflux from hours 4 to 6 of HEA and remained so for the entire recovery period such that net ΔNH4+ returned to zero. Stated another way, the net ammonia gained during the first part of the HEA exposure was subsequently excreted again. Furthermore, the initial loss of ammonia was accomplished by the fish when the external [NH4C1] was still elevated to approx. 1·2 mmol I−1. As can be seen in Fig. 7, this excretion was also measured at a time when the diffusion gradient for NH3 between the fish and the water was nil. That a NH4+ efflux could be maintained during a period when NH3 gradients between the animal and the water were effectively zero and NH4+ gradients were negative (from the water into the animal) suggests that the sculpin is capable of actively extruding ammonia (perhaps via Na+/NH4+ or NH4+/H+ exchange; see review by Evans & Cameron, 1986; Cameron, 1986) under these extreme conditions. However, sculpin (unlike the freshwater trout and catfish; Cameron & Heisler, 1983; Cameron, 1986, respectively) cannot maintain plasma levels below that of an elevated external environment. When the high-ammonia water was flushed with normal sea water, plasma TAmm remained above the control level until hour 20. It is likely that ammonia taken up during the HEA was shuttled back out of intracellular compartments and, finally, into the water.
In conclusion, the maintenance of acid-base balance and the movements of NH4+ and HCO3− have been studied in a marine teleost during exposure to either infused or externally induced ammonia stress. The present results demonstrate that the sculpin is permeable to both NH3 and NH4+. Blood and ion-exchange data following ammonia infusion indicate that at least 50% of the infused load is excreted in the form of NH3. In contrast, when external TAmm levels are elevated, ammonia appears to enter the animal as NH4+. A role for the excretion of ammonia in the form of NH4+ is also apparent (either via ionic diffusion of NH4+ or perhaps by some form of cation exchange), especially when outwardly directed diffusive NH3 gradients are negligible. After each infusion, a majority of the ammonia and associated acid or base load is sequestered intracellularly and then presumably released to the extracellular fluid over time.
ACKNOWLEDGEMENTS
The authors would like to thank Drs James Cameron, Leon Goldstein and Gregg Kormanik as well as Ms Julie Walton for helpful discussions concerning this work. This research was supported by NSF DCM 86-02905 and Georgia Southern College Faculty Research Grants to JBC, and NSF PCM 83-02621 to DHE.