Compartmental Distributions of Carbon Dioxide and Ammonia in Rainbow Trout at Rest and Following Exercise, and the Effect of Bicarbonate Infusion

The mechanisms of CO₂ transport in the blood and its excretion across the gills have been extensively studied in various fish (see reviews by Perry, 1986; Randall, 1990), but little is known about CO₂ distribution or the factors determining that distribution in body tissues of fish. Because of a lack of adequate methods for direct measurement, tissue CO₂ content is calculated by assuming that tissue is equal to venous ⁠, based on the theory of non-ionic diffusion of weak acids (Milne et al. 1958). The validity of such calculations, however, has not been tested, especially in cases where large changes of CO₂ content occur in a short period, such as following exhaustive exercise. The first objective of the present study was therefore to evaluate intracellular CO₂ content in the white muscle of rainbow trout at rest and following exhaustive exercise, based on direct total CO₂ and pH measurements in tissue homogenates.

The distribution and excretion of ammonia have been investigated in some fish (see review by Randall, 1990). The excretion of ammonia and carbon dioxide across fish gills is linked (Wright et al. 1989). In brief, CO₂ excretion across the gills acidifies the water layer next to the gill surface as CO₂ is hydrated at a catalysed rate to form HCO₃⁻ and H⁺. NH₃ excretion however, raises boundary layer pH with the formation of NH4⁺. The latter effect is usually masked by the much larger CO₂ excretion. The formation of an acidic boundary layer at the gill as a result of CO₂ excretion will, therefore, enhance ammonia excretion by trapping NH₃ as NH₄⁺. Carbonic anhydrase has been observed on the sarcolemma of skeletal muscle in a variety of vertebrates (Gros and Dodgson, 1988) including fish (Sanyal et al. 1982). The presence of this enzyme on the sarcolemma raises the possibility that similar interactions between carbon dioxide and ammonia may also occur at the tissue level. The role of carbonic anhydrase on the sarcolemma of muscle tissue is probably to facilitate CO₂ diffusion. However, the functional influence of a catalysed CO₂ hydration reaction on ammonia excretion has not been considered. The second objective of the present study was, therefore, to investigate the possible interactions between carbon dioxide and ammonia transfer from white muscle. This was approached by altering plasma CO₂ content using bicarbonate infusion following exhaustive exercise in rainbow trout, and then examining CO₂ and ammonia contents in extra-and intracellular compartments.

Wright et al. (1988) concluded that fish muscle membranes, unlike those of mammals, were permeable to NH₄⁺. This conclusion has been challenged on the grounds that the metabolic cost that must result from the subsequent shortcircuiting of proton transport is too high (Heisler, 1990). If the muscle membrane is very permeable to ammonium ions, then the removal of the acid boundary layer by bicarbonate infusion should enhance ammonia excretion from the muscle. If the muscle membrane is permeable to ammonia rather than to ammonium ions then the removal of the acid boundary layer should retard ammonia excretion from the muscle.

Rainbow trout [Oncorhynchus mykiss (Walbaum)], weighing 180–250 g, were obtained from a local hatchery, and held outdoors at the University of British Columbia, Canada, in dechlorinated Vancouver tap water (11–13°C) for at least 2 weeks before experimentation. The animals were fed with commercial trout pellets, but feeding was suspended 4 days prior to surgery. Under MS-222 anaesthesia (1:10000 in NaHCO₃-buffered fresh water), fish were fitted with dorsal aortic catheters. Following surgery, fish were allowed to recover for at least 48 h in the darkened Plexiglas box of a 3–1 water recirculating system at constant temperature (12°C).

For measurements at rest, blood samples (800 μl) were taken from the dorsal aorta and immediately analyzed for plasma pH (pHa), plasma total CO₂⁠, plasma total ammonia (Ta_amm) and whole-blood lactate concentration ([Lactate]_a). The animals then received a bolus injection through the dorsal aortic cannula of 1 mg per 300 g body mass of d-tubocurarine chloride (a neuromuscular junction blocker, Sigma). When the animal was unable to move (approximately 10s after injection), it was quickly removed from the chamber, and a sample of white muscle was immediately excised from beside the spine starting at the middle of the dorsal fin and cutting 3–4 cm caudally. Samples were immediately freezeclamped and stored in liquid nitrogen prior to analysis. The time between injection and freezing was 20–30s. The animals were then killed by anaesthetic overdose.

For measurements following exhaustive exercise, animals were chased to exhaustion in a 500-1 circular tank (approximately 6min). Immediately after exercise, blood and white muscle samples were taken as described above.

Animals were exercised to exhaustion as described above. Each fish was then held in the darkened chamber of the recirculating system. The animal was immediately infused with 5 ml kg⁻¹ body mass of saline (control group) or 3 mol l⁻¹ NaHCO₃ in saline (experimental group), using a peristaltic pump to defiver the solution into the dorsal aorta over a period of approximately 10 min. At 15 min or 30min post-exercise (about 5 min or 20 min after infusion), samples of blood and white muscles were taken as described above. For measurement of ammonia excretion in the external medium over the experimental period (0–15 min or 0–30min post-exercise), water samples from the recirculating system were taken at 0 and 15 or 30 min following exercise.

Plasma pH (pHa) was determined with a microcapillary pH electrode (Radiometer G279/G2) coupled to a PHM84pH meter. Plasma total CO₂ was measured using a gas chromatography method (Boutilier et al. 1985) on samples obtained anaerobically. Plasma CO₂ tension and bicarbonate concentration ([HCO₃⁻]_a) were calculated using a rearrangement of the Henderson-Hasselbalch equation with values of plasma pK’ and CO₂ solubility coefficients for trout blood at 12°C (Boutilier et al. 1984). Plasma total ammonia (Ta_amm) was assayed based on the L-glutamic dehydrogenase/NAD method (Sigma reagents). Plasma partial pressure of ammonia arid ammonium ion concentration ([NH₄⁺]_a) were calculated from the Henderson-Hasselbalch equation, using pK and αNH₃ values given by Cameron and Heisler (1983). Whole-blood lactate levels ([Lactate]_a) were analyzed using the L-lactate dehydrogenase/NAD method (Sigma, 1982). Water ammonia concentration was measured by a modification of the salicylate-hypochlorite reaction (McDonald and Wood, 1981). The net fluxes of ammonia (in μmol h⁻¹ kg⁻¹) were calculated from changes in their respective concentrations in the water of the recirculating system.

For the measurements of muscle intracellular ammonia and lactate concentrations, about 500 mg of the tissue powder (wet with liquid nitrogen) was transferred to a preweighed vial containing 1ml of ice-cold 0.6 mol l⁻¹ perchloric acid (PCA) and then reweighed. A further 2 ml of PCA was added and the mixture was immediately homogenized twice on ice for 15 s using an Ultra-Turrax homogenizer. The homogenate was then centrifuged for 10 min at 13 000 revs min⁻¹ and 4°C. A known volume of supernatant was immediately neutralized (pH7.0) with medium containing 1.5 mol l⁻¹K₂CO₃ and 0.5 mol l⁻¹ triethamol-amine, and stored in liquid nitrogen prior to analysis. The concentrations of ammonia and lactate in the supernatant were analyzed in the same way as described above. The values of muscle Ti_amm and [Lactate]_i (mmoll⁻¹ICF) were calculated in the same way as that for (see equations 1 and 2).

Mean values±s.E. are reported throughout. Differences between groups were analyzed statistically using unpaired Student’s t-test. P<0.05 was taken as the fiducial limit of significance.

The present studies are the first direct measurement of carbon dioxide content in the intracellular compartment of fish muscle (Fig. 1; Table 1). In resting animals is similar to and ⁠, whereas after exercise is about 0.2 kPa greater than but is similar to the value calculated for venous blood (Fig. 1A; Table 2). The correlation between and or is demonstrated in Fig. 2. however, is much higher in plasma than in muscle during both rest and exercise (Fig. IB). This reflects the fact that carbon dioxide is distributed according to pH between muscle and blood under all conditions tested (Table 1). Calculated values of muscle pH and total carbon dioxide levels, based on the assumption that carbon dioxide was distributed according to pH, were similar to measured values. Membrane potential was not measured but it is unlikely that it was similar to the calculated equilibrium potential for bicarbonate (Table 1). Plasma and muscle increased markedly following exercise but did not change (Fig. 1). There was a negative linear relationship between and pHa (Fig. 3).

Changes in blood and muscle pH and ammonia concentration were similar to those previously reported for exhaustive activity in trout (Tang and Boutilier, 1991; Wright and Wood, 1988) except that resting ammonia levels were significantly lower than those reported by Wright and Wood (1988). Muscle and Ti_amm levels were higher than those in the plasma, and all increased following exercise (Fig. 4). NH₄⁺ was not distributed according to pH but according to membrane potential (Table 3). Calculated values based on the assumption that ammonia is distributed according to pH were different from measured values, whereas the calculated equilibrium potential was similar to that expected for muscle (Hodgkin and Horowicz, 1959), indicating that NH₄⁺ is distributed according to membrane potential.

The data from the saline and bicarbonate infusion experiments cannot be compared directly with the initial data set because the initial set was collected immediately following exercise, whereas data following infusion were collected 15 or 30 min after exercise. Bicarbonate infusion resulted in a rise in plasma pHa, and bicarbonate levels, which were partially corrected after 30 min (Fig. 5). and ⁠, however, were lower 30 min after exercise in bicarbonate-infused animals and muscle and bicarbonate and lactate concentrations were unaffected (Fig. 5; Table 4). Muscle pHi increased significantly 30 min after exercise in bicarbonate-infused animals (Table 4). The CO₂ partial pressure differences between muscle and plasma decreased 15 min after exercise in bicarbonate-infused animals but returned to normal 30 min after infusion (Table 4).

Bicarbonate infusion caused a significant reduction in plasma total ammonia and NH₄⁺ levels at 15 min, while plasma NH₃ partial pressure and lactate levels were elevated at 30 min (Fig. 6). Bicarbonate infusion also raised the muscle total ammonia, NH₃ and NH₄⁺ levels and the NH₃ partial pressure differences between muscle and plasma (Table 4). The ratio of total ammonia in plasma to that in muscle was lowered after bicarbonate infusion (Table 4). Ammonia excretion by the fish was unaffected by bicarbonate infusion. Ammonia excretion was 536±37μmolh⁻¹kg⁻¹ (mean±s.E., N=6) following saline infusion compared with 543±33μmolh⁻¹kg⁻¹ in bicarbonate-infused fish during the first 30min of the post-exercise period.

The data reported in this study are the only direct measurements of muscle carbon dioxide content in fish. Analysis of these results shows that carbon dioxide is distributed between muscle and plasma according to pH during both rest and exercise in rainbow trout. This pattern is similar to that seen in other vertebrates. The equilibrium potential for bicarbonate is different from the probable membrane potential, indicating that the muscle membranes are not very permeable to bicarbonate. Carbon dioxide must therefore leave the muscle as molecular CO₂ rather than as bicarbonate. As expected, is similar to estimates of and is about 0.2 kPa greater than (Fig. 2).

Resting ammonia levels are lower than those reported by Wright and Wood (1988), presumably because of differences in sampling methods. Ammonia levels are elevated by activity and it is difficult to collect a muscle sample without disturbing the fish. Our method minimized this problem and, as a result, we observed much lower muscle ammonia levels and a significantly lower gradient from muscle to blood in resting fish than those reported by Wright and Wood (1988).

Infusion of bicarbonate after exhaustive exercise resulted in an increase in pH and and HCO₃⁻ and lactate levels in the blood, but little change in the muscle, except for a small increase in pHi (Figs 5, 6; Table 4). Elevated wholeblood lactate levels could be caused by reduced recycling by the liver or increased flux from the muscle, possibly as a result of the increased blood pH following bicarbonate infusion. The observation that there was a significant increase in blood lactate concentration but no change in muscle lactate concentration following bicarbonate infusion can be explained by the large difference in the pool size of the two compartments. The blood pool of lactate represents only about 3% of the total pool; hence, changes in blood levels would represent only a 1% reduction in the muscle pool and would not be detectable by our methods.

Bicarbonate infusion initially lowered blood T_amm and NH₄⁺ levels but subsequently raised blood NH₃ levels. It also caused an increase in muscle NH₃ and NH₄⁺ concentrations. This increase in T_amm could be due to an increase in ammonia production or increased ammonia retention in the muscle. The pattern seen during exercise, however, was identical and so it seems unlikely that there were differences in ammonia production. Thus, it seems probable that the increase in muscle T_amm following bicarbonate infusion was due to ammonia retention. Muscle NH₃ concentration increased more than total ammonia concentration, reflecting the small increase in muscle pH following bicarbonate infusion, but this cannot explain ammonia retention in the muscle. We did not measure the pH of extracelluar fluid surrounding the muscle, but we assume that bicarbonate infusion resulted in an elevation in pH of fluid around the muscle as well as in blood. The more alkaline extracellular environment will raise NH₃ levels and reduce NH₃ flux from the muscle; the reverse, however, will be true for NH₄⁺. If NH₄⁺ transfer is the dominant form of ammonia excretion from the muscle, then raising extracellular pH should augment ammonia excretion by reducing extracellular ammonium ion concentrations. Our observations, however, indicate that alkaline conditions cause ammonia retention in the muscle. This indicates that the muscle membrane in fish, like that of mammals, is permeable to NH₃ but not to NH₄⁺. This is at odds with our calculations indicating that fish muscle is permeable to NH₄⁺ (Table 3). Clearly this requires further investigation. It appears, however, that an acid extracellular environment enhances ammonia removal from the muscle. The location of carbonic anhydrase on the muscle surface (Sanyal et al. 1982) ensures that any CO₂ excreted by the muscle will be rapidly hydrated and will acidify the fluid surrounding the muscle. Thus, it seems likely that CO₂ excretion by the muscle will also augment ammonia excretion from that tissue, in a manner similar to that seen in the gills (Wright et al. 1989). Extracellular carbonic anhydrase activity plays a role, therefore, not only in facilitating CO₂ transfer but also in augmenting ammonia transfer.

Muscle ammonia concentrations increased but ammonia excretion remained the same in fish infused with bicarbonate compared with those infused with saline, indicating increased ammonia production or a decreased rate of muscle ammonia utilization during recovery. The breakdown of adenylate appears to be the major source of muscle ammonia (Mommsen and Hochachka, 1988). This occurs during exhaustive exercise and was the same in both groups because bicarbonate or saline infusion was given after exercise. During recovery, ammonia and IMP are used as substrates to restore the adenylate pool and this process is well under way after 30 min of recovery (Mommsen and Hochachka, 1988), when we measured muscle ammonia concentrations. The only other measured change in the muscle following bicarbonate infusion was a small rise in pH. This may have been sufficient to delay restoration of the adenylate pool. Alternatively, adenylate breakdown may have continued after exhaustive exercise and resulted in a greater ammonium production in the less acidotic, bicarbonate-infused fish. The pH sensitivity of the various components of the purine nucleotide cycle in fish is not known.

was elevated in both blood and muscle following exercise, as reported by others (see Perry and Wood, 1989), and this was inversely related to pH (Figs 2, 3). A reduction in blood pH causes a Root shift and decreases the oxygen content of the blood, but this does not occur during the acidosis following exercise because of β-adrenergic activation of Na⁺/H⁺ exchange by elevated levels of circulating catecholamines, resulting in a rise in erythrocytic pH (Primmett et al. 1986). Thus, as blood flows through the gills, the protons produced by haemoglobin oxygenation fuel both bicarbonate dehydration and Na⁺/H⁺ transfer in the red blood cell. In fact, these two mechanisms compete for the available protons as blood passes through the gills. The dumping of protons into the plasma from red blood cells will create a CO₂/HCO₃⁻ disequilibrium, resulting in a rise in as blood flows away from the gills. Na⁺/H⁺ exchange is enhanced at low pH (Nikinmaa et al. 1987), so the extent of proton dumping will be increased by low pH and will result in a larger rise in ⁠. This explains the observed inverse relationship between and pHa (Fig. 3) in trout following exhaustive exercise, when levels of circulating catecholamines are elevated. is further elevated following bicarbonate infusion because of an exacerbated CO₂/HCO₃⁻ disequilibrium.