CO₂ and H⁺ Excretion By Swimming Coho Salmon, Oncorhynchus Kisutch

Burst swimming activity in fish results in the production of large amounts of lactic acid which, at body pH of between 7 ·0 and 8 ·0, is almost completely dissociated and causes a marked fall in blood pH. The gill epithelia of teleost fish have been shown to be very permeable to hydrogen ions (McWilliams & Potts, 1978). Therefore, a change of internal or external pH will alter the net transport of protons across the gills. It is possible that an important component of acid-base regulation in fish following burst swimming is the diffusive loss of protons from fish across the gill epithelium. Similarly, a reduction of external pH by addition of acid to the water would reduce or even reverse hydrogen ion excretion and therefore impair the acid-base balance of the fish.

In this study we investigated the impact of various pH differences across the gills on H⁺ and CO₂ excretion by coho salmon swimming in sea water for prolonged periods. The pH difference between blood and water across the gill was altered by reducing seawater pH or by reducing blood pH by burst swimming. In addition, bicarbonate concentration difference between blood and water was reduced by lower ing seawater bicarbonate levels.

Experiments were carried out on seawater coho salmon, Oncorhynchus kisutch, in a non-reproductive stage, weighing between 0 ·5 and 1 ·0 kg, obtained from the Pacific Biological Station in Nanaimo, B. C., Canada. The fish were kept in fibreglass hold ing tanks of about 2m diameter and supplied with flowing sea water (10 –12°C) with a maximum water current of 0 ·3 ms⁻¹. The fish were fed daily with Moore-Clarke trout pellets and were treated for furunculosis and fibriosis before transportation and then afterwards with furazolidon to prevent skin infections.

Experiments were carried out in a large volume (1201) tunnel respirometer at 13 °C (Brett, 1964) flushed with local Bamfield sea water (salinity 32 ± l%o). The res pirometer was that used by Brett (1964), transported to Bamfield Marine Station and modified in the following ways. Pure oxygen was injected into the recirculating water of the closed respirometer to replace that used by the swimming fish and this per mitted prolonged closure of the system without causing hypoxia. Oxygen was injected using a 50 ml calibrated glass syringe kept at room temperature and the injected volume measured at atmospheric pressure. The injected O2 dissolved in the water and was completely mixed within 30 min. Oxygen was injected at intervals such that water O2 levels never fell below 110 mmHg The O₂ content of the water was deter mined from the measured with a Radiometer PHM71 blood gas analyser and thermostatted electrode and the Bunsen solubility coefficient for O₂ at the tem perature and salinity of the sea water (Dejours, 1975). Oxygen uptake rate was deter mined from the change in oxygen content of the water, the total volume of water in the respirometer and the amount of oxygen injected into the system. Water sampled from the respirometer was replaced from a small head tank 30 cm above the res pirometer.

Prolonged closure of the respirometer was required to create measurable changes in the CO₂ content of the water. In preliminary experiments, prolonged closure was shown to be associated with reductions in water pH. In subsequent experiments water pH was kept constant by injecting 0 ·25N-NaOH (100% pure, CO₂-free) by means of a Harvard Linear displacement pump at rates between 0 ·5 and 2 ·5 ml min⁻¹. The pump was operated by a small electrical comparator connected to the output of a Radiometer pH meter (PHM65) and electrode, which were used to monitor water pH. If water pH fell below a predetermined setpoint the pump was activated and NaOH injected until the pH of the water was re-established at the setpoint level. The maximum variation in pH was never greater than 0 ·02 units.

Total carbon dioxide of sea water was determined using the method of Cameron (1971) with a Radiometer PHM71 acid-base analyser and associated CO₂ electrode. The bicarbonate concentration of sea water was measured by microtitra tion; 25 ml water samples were acidified to pH3 ·5 –3 ·8 with 0 ·1 N-HC1 added via a Gilmont microburette. CO₂ was then expelled by bubbling with pure N₂ for about 4 min. The samples were returned to the initial pH by titrating with 0 ·02 N-NaOH via Another Gilmont microburette. The difference between the amount of acid and base added represented the amount of bicarbonate in solution. The method was reproducible to 0 ·0005 mequivl⁻¹ for concentrations of between 0 ·1 and 2 ·0mm and gave 100% recovery of pure NaHCO₃.

In some experiments, the dorsal aortae of fish were cannulated (Smith, 1978) and the indwelling PE50 polyethylene cannula was used to sample blood while the fish was swimming. Blood and pH were measured using a Radiometer PHM71 blood gas analyser and the appropriate electrodes thermostatted at the temperatures of the fish and calibrated with gases supplied by Wösthoff gas mixing pumps or pH standard solutions. Erythrocyte pH was measured using the quick freezing method of Zeidler & Kim (1977). Blood total CO₂ was measured using the method of Cameron (1971).

The critical velocity of each cannulated fish (Ucn_t) was determined after it was anaesthetized lightly in MS222 in sea water (50 mg 1⁻¹), placed in the open res pirometer, continuously flushed with sea water, and allowed to recover overnight. The water velocity, and therefore the swimming speed of the fish, was increased using increments of 0 ·5 body lengths per second at 1-h intervals until the fish was exhausted (see Hoar & Randall, 1978, for details of calculation of critical velocity). Cannulated fish were allowed 2 days recovery from the operation before being placed in the respirometer. The Ucrit of these fish was measured with the dorsal aortic cannula trailing in the respirometer.

Following an overnight rest, each cannulated fish was swum at 80% Ucnt and various blood parameters measured. Each fish was allowed 2 h rest and was then swum to exhaustion (burst swimming) at 120% U_crit-Blood parameters were measured again 30 min after fatigue.

A further series of experiments were carried out on uncannulated fish. After deter mining the U_crit, each fish was subjected to four swimming periods of 6 h each, with an overnight rest between each swimming period. The respirometer was closed during each swimming period and changes in O₂, CO₂ and HCO₃⁻ in the sea water were determined. Oxygen was injected into the respirometer and seawater pH held con stant by NaOH injection. The fish were subjected to different seawater conditions during each swimming period. Seawater pH was lowered by adding HC1 to the respirometer. Seawater bicarbonate levels were reduced by first acidifying 4501 of sea water in a header tank with HC1 to pH 3 ·5. The sea water was then aerated until the total CO₂ concentration was below 0 ·1 mm. The seawater pH was then elevated to 8 ·0 with NaOH and used to flush the respirometer with three times its own volume, giving a final total CO₂ level of 0 ·2 mw in the respirometer.

Fish were subjected to the following conditions:

1. Fish were swum for 6h at 80% Ucrit in sea water at pH 7 ·95. The pH of Bamfield sea water is around 8 ·1, so pH was lowered initially by injecting a small volume of 0 ·10N-HC1.

2. Fish were swum for 6 h at 80% Ucrit in low bicarbonate (0 ·2 HIM) sea water. The fish was forced to swim in the closed respirometer and seawater pH was held ai 7 ·95.

3. Fish were swum for 6 h at 50% Ucrit in normal sea water acidified to pH 7 ·1 by adding HC1, and maintained at this pH by the pH stat device. The initial reduction in seawater pH caused a rise in seawater to 4 ·4 mmHg. Thus fish were subjected to hypercapnia as well as reduced seawater pH during this experimental period.

4. Fish were swum at 50% Ucrit following a burst swim in normal sea water at pH 7-95. The protocol was to swim the fish at 50% Ucrit for 1 h after closing the respirometer, then burst swim the fish (at 120% Ucrit) to fatigue before returning the swimming speed to 50% Ucnt for 6h.

Cannulated coho salmon swimming at 80% Ucrit for 1 h showed little or no change in either total CO₂, haematocrit or plasma and erythrocytic pH. There was a non-significant increase in haematocrit and a decrease in plasma pH and blood total CO2 after 10min exercise (Table 1). Burst swimming also had little effect on pleasured blood variables (Table 2) but was associated with a non-significant increase in and decrease in plasma pH.

We found that each fish had to be swum for several hours in the closed respirometer before we could accurately measure changes in seawater CO₂ levels. Repetitive blood sampling over this prolonged period eventually impaired the swimming ability of the fish. As a result we decided to continue our experiments on uncannulated fish, varying only seawater pH and bicarbonate levels. In preliminary experiments, in which uncannulated fish were swum for prolonged periods in the closed respirometer, there was a sharp drop in water pH (Fig. 1). In these experiments pH changes appeared to affect CO₂ excretion, so subsequent experiments were carried out at a constant pH (Fig. 2). The mean accumulated oxygen uptake, CO₂ excretion and net hydrogen ion transfer are illustrated in Fig.,3 and Table 3. Most rates did not change very much during each run (Fig. 3A-D), so mean rates were calculated for each 6-h period (Table 3). The data for conditions 2, 3 and 4 are presented as a fraction of the condition 1 value in Table 3 for ease of comparison. The oxygen uptake of coho salmon swimming at 80% Ucrit in low bicarbonate sea water was reduced, compared with that of fish in normal sea water (condition 1). Oxygen uptake was further reduced in conditions 3 (low pH sea water) and 4 (following burst activity), as these fish were only capable of swimming at a reduced speed of 50% Ucrit

Coho salmon exposed to low pH sea water had a high rate of HCO3⁻ excretion over the first 30 min of the swimming period and this was associated with a low rate of H⁺ transfer (Fig. 3C). After burst activity the reverse was observed, namely a high H⁺ transfer but depressed HCO₃⁻ excretion (Fig. 3D).

The amount of NaOH injected into the respirometer is equal to the H⁺ excreted by the fish, if we regard anything that reduces seawater pH as contributing to H⁺ Excretion by the fish ⁠. CO₂ excreted into the water will reduce pH and cause an increase in seawater bicarbonate. CO₂ production by the fish was derived from the amount of CO₂ neutralized, i.e. the seawater HCO₃⁻ accumulation coupled to NaOH addition, plus the increase in molecular CO₂ calculated using the Henderson-Hasselbalch equation and appropriate values of pK’ and CO₂ solubility in sea water (Dejours, 1975; Strickland & Parsons, 1968).

Table 4 shows the calculated respiratory exchange ratios (RE) based on our estimates of CO2 excretion and oxygen uptake (⁠ Lipids are the major fuel for prolonged swimming in coho salmon (Krueger et al. 1968) and as a result one would predict a respiratory quotient of 0 ·7. Thus, if there were no changes in body CO₂ and O₂ content one would expect a respiratory exchange ratio of 0 ·7 in these experiments on swimming coho salmon. The ratio, however, was low, less than 0 ·7 in all cases. The ratio of to however, was only less than 0 ·7 when seawater pH was reduced to 7 ·1 (Table 4).

If CO₂ is the only substance excreted by the fish into the sea water that affects pH, then one would expect to equal The ratio, however, is less than unity in low pH and low bicarbonate sea water and greater than unity in normal sea water and following a burst swim (Table 4).

The fish gill epithelium is very permeable to CO₂, permeable to hydrogen ions (McWilliams & Potts, 1978), but probably not very permeable to bicarbonate ions (Perry, Davie, Daxboeck & Randall, 1982). Swimming coho salmon excrete CO₂ which reacts with water and reduces pH and elevates the HCO₃⁻ content of the sea water in the respirometer. Most of the CO₂ must enter the water as molecular CO₂ but there may be a separate and occasionally large flux of protons. Thus seawater pH is reduced by the addition of both CO₂ and protons to the water. Fish excrete ammonia and ammonium ions into sea water; at the pH of sea water, excreted ammonia will combine with protons to produce ammonium ions. Thus our values of hydrogen ions excreted, as measured by the amount of NaOH injected to maintain seawater pH at a given level, are probably underestimated because we have not allowed for ammonia excretion. Ammonia production is probably low, however, because the swimming coho salmon were fasting while in the respirometer. If ammonia excretion was similar to that of fasting sockeye salmon (Brett & Zala, 1975), we estimate that ammonia excretion results in an underestimate of hydrogen ion excretion of about 6-8%. Ammonia excretion will have no significant effect on our estimate of CO2 excretion.

The measured respiratory exchange ratio for conditions 1, 2, 3 and 4 was less than the expected 0 ·7 for fat metabolism. Oxygen uptake is a reflection of oxygen utilization by the tissues because oxygen stores are small in fish. Thus the low RE value, of less than 0 ·7, is caused either by a reduction in CO₂ production by the tissues or by an increase in the CO₂ stores in the fish. It is possible that some CO₂ produced by fat metabolism could be utilized by other metabolic pathways, but this seems unlikely because pathways that consume CO₂ are anabolic and probably are not operative during exercise. Thus, the most probable explanation for the low RE values is that some CO₂ is retained in the body of the fish during swimming. The extent of CO₂ retention has been calculated assuming a respiratory quotient for fat metabolism of 0 ·7 and comparing this with the measured RE value for swimming coho salmon (Table 5). In all cases there was CO₂ retention in the body.

If we assume that ammonia excretion has a negligible effect, then we can calculate the acid-base imbalance of the fish from the difference between acid and bicarbonate transfer into sea water. The accumulated imbalance over 6 h is presented in Table 5. As can be seen in Fig. 3, the imbalance was built up in a linear fashion for conditions 1 and 2, while for 3 and 4 the imbalance was mainly due to initial changes.

The initial concentration differences across the blood-water barrier were used to calculate the driving forces for passive diffusion of H⁺ and HCO₃⁻ ions (Table 6).

As can be seen in Table 6 the passive H⁺ ion flux is reversed in condition 3 (pH 7 ·1), directed inwardly. This is consistent with the highly suppressed acid production of the fish (Fig. 3C). In some cases the seawater pH did not change at all for over 1 h, although the [HCCh⁻] increased markedly. Maybe here the H⁺ influx equalled the CO₂ excretion which would result in [HCO₃ ] increase without pH change. The initial H⁺ driving force in condition 4 is enlarged, indicating a higher H⁺ loss by passive diffusion. The excess acid transfer must be due to the anaerobic lactate production during burst swimming. In condition 2, the low [HCO₃⁻] in the sea water results in a high driving force for HCCh⁻ ions and will therefore cause a higher HCO₃⁻ loss by passive diffusion. These initial conditions corroborate with the obser ved HCO₃⁻ loss (Fig. 3B and Table 5).

The very low RE of 0 ·21 in condition 3, with even more reduced hydrogen ion rcretion = 0 ·17), was maintained for the duration of the experiment, that for 6-h in sea water pH 7 ·1 (Table 4). If one assumes a tissue RQ of 0 ·7, then the hydrogen ion load on the buffering capacity of the body is large. A 600 g coho will retain 16 mequiv of acid over the 6-h period, or 27 mequiv kg⁻¹. The buffering capac ity of steelhead trout muscle was found to be 60 mequiv pH unit⁻¹ (G. Somero, unpublished communication). Heisler (1980) calculated the whole body buffer capac ity for dogfish to be 38 mequiv pH unit⁻¹. Assuming similar buffering capacities in coho, then body pH should drop 0 ·7 –1 ·0 pH unit during the course of the experiment. Salmonids can tolerate such changes (Hillaby & Randall, 1979) but a fall in body pH probably impairs many body functions. We observed that, in sea water of pH 7 ·1 or following a burst swim, the maximum prolonged swimming speed of coho salmon was reduced and could only be maintained for 6 h at 50% Ucrit of that in normal sea water. The reduction in body pH may limit aerobic activity either by reducing the blood oxygen capacity via the Root effect or may impair metabolism directly by the inhibit ory effect of low pH on a variety of enzyme activities. Either or both actions may have impaired the ability of the fish to maintain high levels of prolonged swimming in conditions 3 and 4.

If ammonia excretion were to account for the apparent reduction in acid excretion in conditions 2 and 3, then it can be calculated, assuming an RQ of 0 ·7, that the fish would have to excrete 4 ·6 and 8 ·6 mmol NH3 per 600 g fish over the 6-h experimental period, respectively. The body stores of ammonia are small, less than 0 ·25 mmol, so this NH3 would have to be produced metabolically. There are several possible sources: (i) hydrolysis of amides (e.g. glutamine, asparagine and lysine), (ii) purine nucleotide cycle activity, (iii) degradation of nucleic acids, and finally (iv) protein catabolism followed by amino-acid oxidation. Amino-acid oxidation is probably most important from a quantitative point of view, the NH3/ CO₂ ratio being 0 ·28. The RQ for protein oxidation, however, is 0 ·97 (Thillart & Kesbeke, 1978) and ammonia excretion will neutralize only some 30% of the acidity resulting from CO₂ excretion. In fact, the ratio expected from lipids and proteins will be about the same, that is 0 ·7. Thus, even if the swimming coho salmon produced some ammonia, it is very unlikely that ammonia excretion can account for the very low rates of apparent acid excretion observed in condition 3, where values were reduced to 0 ·17.

It is possible that swimming fish subjected to acid waters are able to activate metabolic pathways that consume acids to ameliorate the effects of acid accumulation within the body. Gluconeogenesis, for example, will alkalinize tissues, as will every pathway that converts acids into fats and sugar. Although these mechanisms are important during rest, it is questionable whether they are activated in swimming fish. It seems more probable that the regulatory enzymes controlling the rates of these proton consuming pathways will be inhibited by a reduction in the energy charge of the cell as the adenylate pool is decreased when the fish swims.

What is apparent from these studies is that CO₂ and hydrogen ion production by the fish need not be matched and that changing internal and external conditions can have a marked and separate effect on hydrogen ion and carbon dioxide excretion. Also, CO₂ is not necessarily excreted in a constant ratio with O₂ uptake according to oxidative metabolism, but appears to be influenced by the acid-base condition of the fish.

This study was carried out at the Bamfield Marine Station and supported by a NSERC strategic grant to DJR. Technical assistance of Christine Milliken is grated fully acknowledged. GvdT was a recipient of a NATO-science fellowship granted by the Netherlands Organization for the Advancement of Pure Research (Z.W.O.). LH-R was a Chinese Cultural Exchange Scholar.