Effect of burst swimming and adrenaline infusion on O₂ consumption and CO₂ excretion in rainbow trout, Salmo gairdneri

In fish, the rate of carbon dioxide excretion is largely dependent on the catalysed dehydration of plasma bicarbonate by erythrocytic carbonic anhydrase (Randall & Daxboeck, 1984). As in mammals, plasma bicarbonate enters fish erythrocytes via a chloride/bicarbonate exchange mechanism on the erythrocyte membrane (Cameron, 1978; Obaid, Critz & Crandall, 1979).

Wood & Perry (1985) have reported that adrenaline inhibits bicarbonate entry into rainbow trout erythrocytes in vitro. In the intact animal, this inhibition would result in the retention of plasma bicarbonate during branchial blood transit and, therefore, a reduction of carbon dioxide excretion. Indeed, Perry (1986) proposes that adrenergic control of erythrocytic chloride/bicarbonate exchange may be important for acid-base regulation in fish by increasing internal bicarbonate stores. Circulating catecholamine levels are elevated in fish following acid-base disturbances (Boutilier, Iwama & Randall, 1986; Perry, 1986; Primmett, Randall, Mazeaud & Boutilier, 1986). At present, however, there is little evidence that these hormones modulate carbon dioxide excretion under these conditions in the intact animal. Van den Thillart et al. (1983) have reported low respiratory exchange ratios (carbon dioxide retention) in exercising coho salmon (Oncorhynchus kisutch) which would indicate a reduction in erythrocytic chloride/bicarbonate exchange. The effect of their experimental protocol on circulating catecholamine levels in these fish, however, is unknown.

The purpose of the present experiment was to determine if catecholamines modulate carbon dioxide excretion and, therefore, the respiratory exchange ratio in the rainbow trout in vivo. Burst swimming is known to cause a large increase in circulating catecholamines and an acid-base disturbance in the rainbow trout, Salmo gairdneri (Primmett et al. 1986). We have, thus, examined the effect of (1) burst swimming and (2) adrenaline infusion on carbon dioxide excretion and the respiratory’ exchange ratio in these animals.

Rainbow trout (Salmo gairdneri), weighing between 285 and 700 g, from the Sun Valley Trout Farm (Mission, BC) were maintained in large outdoor tanks. Water temperature varied from 10 to 15°C. The fish were regularly fed and in good health. The experimental protocol consisted of two separate series of experiments, in one of which fish were induced to burst swim and in the other adrenaline was infused into resting fish.

The first series of experiments was performed at 15±0·5°C in a Brett-type swimming respirometer (Brett, 1964) with a total water volume of 37·51. Oxygen consumption was determined over20-min periods from the decline in oxygen tension of the recirculating water in the closed respirometer, and the amount of pure oxygen injected into the system, as described by van den Thillart et al. (1983). Oxygen tension of the water was measured continuously by recirculating a small fraction of the water through an O₂ electrode mounted in a cuvette (Radiometer, E-5046 and D-616), as described by Steffensen, Johansen & Bushnell (1984). The O₂ electrode was connected to a Radiometer PHM71 acid—base analyser and a chart recorder.

CO₂ excretion from the fish was calculated from the difference in total CO₂ content of the water, determined every’ 20 min by analysing water samples using a Carle Series 111 analytical gas chromatograph with a Poropak Q column. Total CO₂ was determined as described by Boutilier, Iwama, Heming & Randall (1985) with thé following modifications. A water sample of 2·0 ml was injected into a glass syringe (10ml) containing pure nitrogen, and acidified with 50μl of 10moll⁻¹ HC1. After 2 min of shaking, at least 6 ml of the gas in the syringe was injected into the gas chromatograph loop (volume = 1 ·0 ml) to ensure complete wash-out of the loop. The CO₂ concentration was calculated by integrating the signal from the gas chromatograph with an HP 3497 data acquisition/control unit and an HP9135A computer.

The water pH in the swimming respirometer was kept constant at 6·80 by means of a pH electrode (Canlab, GK2401C) permanently mounted in the system, connected to a Radiometer PHM71 acid—base analyser and a comparator. The recorder output was connected to a circuit controlling a Harvard Linear displacement pump injecting 0·25 mol C⁻¹ NaOH, as described by van den Thillart et al. (1983). When the pH of the water fell below 6-80, the pump was activated and NaOH injected until the pH was re-established.

The fish were acclimated in the respirometer for 24h before the experiment. During this period, the respirometer was continuously flushed with thermostatically controlled, aerated water of pH 6·80. The experiment was started by closing the respirometer. Control O₂ consumption and CO₂ excretion were measured for 20-min periods for 80 min at the acclimation swimming speed (40 cms⁻¹). The swimming speed was then increased to 80-85 cms⁻¹ for 10 min, causing the fish to burst swim to ‘exhaustion’. Burst swimming causes an increase in catecholamines in trout (Primmett et al. 1986). After 10min, the speed was returned to 40cms^-1 and the same parameters were measured during the following 80 min, when the experiment was terminated.

In the second series of experiments, the fish were anaesthetized with MS-222 and cannulated in the dorsal aorta with PE 50 polyethylene tubing, as described by Soivio & Oikari (1976). The cannula was used to sample blood for determination of red blood cell (RBC) pH (pH,), according to the freeze-thaw method described by Zeidler & Kim (1977) with a Radiometer PHM71 acid-base analyser and a micro-pH unit. In addition, the cannula was used to infuse 0·25 ml of 10⁻⁴mol I⁻¹ adrenaline in the dorsal aorta. This large dose ensured that the β-adrenoceptors were stimulated and saturated. The increased RBC pH_i indicated that the bolus acted on the β-receptors of the RBC. The adrenaline solution was prepared no more than 10 min before the infusion. The fish was acclimated for at least 24h in the respirometer. Control measurements were performed for 10-min periods for 70 min. Adrenaline was then injected and the measurements were continued for the following 60 min. O₂ consumption, CO₂ excretion and erythrocyte pH (pH_i) were measured in resting fish before and after infusion of adrenaline. The fish were housed in a flow-through respirometer with a volume of 3·01. The respirometer was constructed with a recirculating circuit to ensure adequate mixing. Water temperature was 10-0 ± 0-5°C.

In both series of experiments the ambient oxygen tension was always kept above 14·66 kPa.

Statistical analysis was based on Student’s t-test. All values in the text and tables are given as mean ± standard deviation.

Results of oxygen consumption, CO₂ excretion and respiratory exchange ratio (RE) for six fish swimming at 40cms⁻¹ before and after burst swimming are summarized in Table 1. Control oxygen consumption at a swimming speed of 40cms⁻¹ ranged from 118·3 to 135·6μmolkg⁻¹ min⁻¹ (mean ± S.D. = 126·8 ± 29·2μmol kg⁻¹ min⁻¹). CO₂ excretion varied from 86·7 to 102·1 μmol kg⁻¹ min ¹ (94-2 ± 21-3 μmol kg^-1 min^-1). The calculated respiratory exchange ratio varied from 0·68 to 0·80 with a mean of 0·74 ± 0·08.

After burst swimming for 10 min, oxygen consumption increased significantly, by 71%, to 217·3 ± 32·1 μmol kg⁻¹ min⁻¹ during the first 20 min of recovery. CO₂ excretion increased by 104% to 192·1 ± 49·9μmolkg⁻¹ min⁻¹; thus RE increased significantly, by 17 %, to 0·87 ±0·11.

20–40 min after burst exercise, O₂ consumption and CO₂ excretion were still significantly elevated as compared to the control values, 32% (167·1 ±26·9μmol kg⁻¹min⁻¹) and 50% (141·6 ± 23· μmol kg⁻¹ min⁻¹), respectively. RE was not significantly different from the control value.

During the following 20min, oxygen consumption was 152·8 ± 25 ·7μmol kg⁻¹ min⁻¹ and CO₂ excretion was 118·4 ± 34·5μmol kg⁻¹ min⁻¹, or 21% and 26% higher than control, respectively. RE was not significantly different from the control value.

During the last period (60—80 min) none of the measured parameters was significantly different from control values.

The effects of infusing adrenaline into nine resting rainbow trout are summarized in Table 2. Oxygen consumption during the 70-min control period was 47·9 ± 8·0μmol kg⁻¹ min⁻¹, CO₂ excretion was 35·2 ± 6·7//mol kg⁻¹ min⁻¹, and RE, consequently, was 0·74 ±0-09. Intracellular pH measured prior to adrenaline infusion was 7·391 ± 0·017. Adrenaline had no significant effect on either O₂ corísumption, CO₂ excretion or RE during the following 60 min. 10 min after injection of adrenaline, pH_i had increased to 7·447 and remained elevated (7·432) for 60 min after adrenaline infusion.

Several types of stress have been shown to cause an increase in the levels of circulating catecholamines in fish (Nakano & Tomlinson, 1968; Mazeaud & Mazeaud, 1981; Boutilier et al. 1986; Perry, 1986). Recently, Primmett et al. (1986) have documented 25-to 35-fold increases in circulating adrenaline and noradrenaline levels in the rainbow trout following a burst swim. It was therefore assumed that the burst swim in the present experiment would cause a similar increase in the levels of circulating catecholamines.

It has been reported that catecholamines affect CO₂ transport by inhibition of bicarbonate flux through the erythrocyte (Wood & Perry, 1985). This inhibition of erythrocytic chloride/bicarbonate exchange could explain the low RE values reported by van den Thillart et al. (1983) in the coho salmon (Oncorhynchus kisutch). In coho salmo (exercised in sea water at pH7·0) the RE was found to be 0·21. Van den Thillart et al. (1983) also measured and after burst swimming in normal sea water, but only as mean rates for a 6-h period ‘since most rates did not change very much during each run’. They determined RE to be 0-64 (JV=4). It appears, however, that in their representative fish (their fig. 3), the bicarbonate excretion during the first hour after burst swimming was only 1/15 that of the following 3h. Accordingly, RE must have been approximately 0-1 during the first hour. The results of the present experiments, in contrast, provide no evidence that increased circulating catecholamines lead to a reduction in carbon dioxide excretion in the whole animal.

Infusion of adrenaline into resting rainbow trout in vivo caused no significant changes in either CO₂ excretion or O₂ consumption and, consequently, the RE value did not change significantly from its control value of 0·74. In addition, burst swimming, which is associated with an increase in blood catecholamines, caused an increase in both O₂ consumption (71%) and carbon dioxide excretion (104%) during the recovery period. Thus, in contrast to the ‘CO₂ retention’ proposal of Wood & Perry (1985), there was, in fact, a significant increase in the respiratory exchange ratio to 0·87. The increased RE following the burst swim is likely to result, as in other animals, from the titration of the blood bicarbonate pool by protons entering the blood from the exercising tissues.

It is possible that the absence of an effect of catecholamines on CO₂ excretion in our study could have been due to seasonal variation in inhibition of erythrocytic chloride/bicarbonate exchange by catecholamines caused by seasonal variation in μ-receptor activity. It has been documented that the gills and heart (Peyraud-Waitzenegger, Barthelemy & Peyraud, 1980) of eels may lose β-adrenergic sensitivity during the winter. Similarly, Nikinmaa & Jensen (1986) have suggested that this may also be the case for the erythrocytes of the rainbow trout. However, the fact that an elevation in catecholamine levels caused a significant increase in the erythrocyte pH, as shown in other studies (Nikinmaa, 1983; Nikinmaa & Huestis, 1984; Heming el al. 1986), indicates that the β-adrenergic receptors were still functional in the trout used in this study. Thus, the absence of an effect of catecholamines on CO₂ excretion in the present study cannot be attributed to reduced activity of erythrocytic β-adrenergic receptors.

There is also a discrepancy between the present experiment and those of van den Thillart et al. (1983) concerning oxygen uptake following burst swimming. These authors reported no change in oxygen consumption during the approximately 4-h recovery period, whereas we found O₂ consumption had initially increased, but decreased to control values within 80min. Brett (1964) reported an oxygen debt replacement up to 5 h after fatigue in yearling sockeye salmon (Oncorhynchus nerka). Similarly, Stevens & Randall (1967) and Steffensen et al. (1984) found that the O₂ consumption of rainbow trout decreased to control values within 0·5–4 h after strenuous exercise. Why van den Thillart el al. (1983) found no such increase in O₂ consumption (or oxygen debt) after burst swimming is not clear: an oxygen debt after strenuous exercise can be expected, since lactate is removed metabolically. Holeton, Neumann & Heisler (1983) found that strenuous exercise resulted in a severe lactacidosis, which was corrected within 4h by a transient net transfer of H⁺ to the environmental water. The lactate was removed metabolically within 6–8h. This curious lack of an oxygen debt after burst swimming in the study of van den Thillart et al. (1983) indicates that their surprisingly low RE values may be due to technical limitations.

In summary, the present experiments provide no evidence to support the view that increased catecholamine levels in fish cause a reduction in carbon dioxide excretion (Wood & Perry, 1985; Perry, 1986). In addition, carbon dioxide excretion is proportional to haematocrit and inhibition by the anion exchange blocker 4-acetamido-4’-isothiocyanatosilbene-2,2’ disulphonic acid (SITS) in the blood-perfused trout preparation (Perry, Davie, Daxboeck & Randall, 1982). The present experiment also provides no evidence that erythrocytic chloride/bicarbonate exchange is inhibited in the rainbow trout, in vivo by elevated catecholamines and contrasts with the results obtained with in vitro preparations (Wood & Perry, 1985).

JFS was supported by a NATO Science Fellowship, the Danish National Science Research Council and the Carlsberg Foundation, Denmark. BLT and DJR were supported by the National Science and Engineering Research Council of Canada. Assistance from D. Mense and L. E. Fidler is gratefully acknowledged.