Control of Red Blood Cell Metabolism in Rainbow Trout After Exhaustive Exercise

In teleost fish, the red blood cell (RBC) is not only the vehicle that carries oxygen to the tissues but it is also a significant consumer of this oxygen (Hunter and Hunter, 1957; Eddy, 1977). Indeed, it has been shown recently that aerobic metabolism accounts for more than 90 % of resting nucleoside triphosphate (NTP) production in salmonid RBCs (Ferguson and Boutilier, 1988, 1989; Ferguson et al. 1989). These workers demonstrated a close coupling between aerobic metabolism, maintenance of intracellular NTP levels and regulation of intracellular pH (pHi). Thus, in response to stressful situations such as strenuous exercise, the energy demand of the RBC may increase above resting levels in association with adrenergic activation of RBC Na⁺/H⁺ exchange. This mechanism serves to regulate RBC intracellular pH and thereby maintain blood O2 transport capacity in the face of extracellular acidosis (Primmett et al. 1986; Milligan and Wood, 1987). In addition to elevated plasma catecholamine levels, a variety of other variables can potentially affect RBC metabolism after strenuous exercise. These include altered blood pH, and HCO₃⁻ levels (typically a mixed respiratory/ metabolic acidosis occurs), as well as altered plasma concentrations of oxidative substrates, in particular greatly elevated plasma lactate (Wood and Perry, 1985). Pervasive direct actions of acid-base variables and catecholamines on the metabolism of hepatic cells in salmonids have been demonstrated recently (Walsh et al. 1988; Moon et al. 1988; Mommsen et al. 1988; Wright et al. 1989). Therefore, it seems likely that flux through metabolic pathways of RBCs could be subject to similar perturbation during acid-base stress and adrenergic pHi regulation, though nothing is known on this point.

The preceding paper described the basic metabolic pathways present in trout RBCs, developed in vitro methods for assessing their oxidative metabolism under defined acid-base conditions, identified lactate and glucose as important oxidative substrates and characterized the kinetic relationships involved (Walsh et al. 1990). The goal of the present study was to extend this experimental framework to an analysis of RBC metabolism in vivo after strenuous exercise in the rainbow trout Oncorhynchus my kiss. We first assessed the conditions present in the bloodstream (acid-base status, catecholamines, substrate levels) and the oxidative metabolism of the resident RBCs (substrate oxidation rates, total O₂ consumption rates) using blood withdrawn from catheterized fish at rest and at selected times after exhaustion. During the measurements, the RBCs were incubated in vitro under conditions closely duplicating those monitored in vivo. Subsequently, we established the mechanisms of exercise-induced alterations in metabolism by simulating the various altered parameters both singly and in combination. When combined with the kinetic relationships determined in the preceding paper (Walsh et al.1990), this multifactorial modelling approach allowed an assessment of the independent and integrated effects of catecholamines, substrate levels and the various components of acid-base status on RBC metabolism.

Rainbow trout (Oncorhynchus mykiss=Salmo gairdneri; 150–400 g) were obtained and held at 9–12°C as described earlier (Walsh et al. 1990). All fish, whether employed experimentally in the exercise trials or used as a source of blood in the factorial modelling experiments, were fitted with indwelling arterial catheters (Soivio et al. 1972). Prior to sampling or exercise, the fish were allowed to recover for 24–48 h in darkened Perspex chambers served with aerated flowing tapwater at the experimental temperature (10±l°C).

Trout were transferred, without air exposure, to a 5001 tank, and then exercised to exhaustion by vigorous chasing for 6 min (see Turner et al. 1983). At the end of exercise, fish were returned to their individual chambers and sampled at various times during recovery. Arterial blood samples (approximately 2.0ml, 50i.u. heparin ml⁻¹) were drawn anaerobically into ice-cold gas-tight syringes (Hamilton) at rest, at Oh (immediately post-exercise, prior to return to the holding chamber), and at 1, 2, 4 and 12 h of recovery. Fish used for the rest samples were not subsequently exercised in case the removal of blood altered swimming performance. At other times, fish were sampled no more than twice, at intervals no shorter than 2 h, and the blood removed was replaced with an equal volume of non-heparinized Cortland saline (Wolf, 1963).

Blood samples were analyzed for arterial pH (pHa), RBC intracellular pH (pHi), true plasma total CO₂ content ⁠, haemoglobin (Hb), haematocrit (Het), plasma glucose, lactate, epinephrine and norepinephrine levels, and either RBC CO₂ production rates individually from lactate and glucose or total RBC O₂ consumption rate ⁠. Note that values were calculated and, if samples were not at equilibrium, it is possible that the 0 h post-exercise values may be slight overestimates. Not all parameters were measured on all samples. For the metabolic determinations, the samples were placed in 20 ml glass vials in volumes of either 700 μl (for CO₂ production measurements) or 1400 μl (for measurements) and additionally heparinized (125 i.u.). The vials were placed in a shaking water bath at 10±l°C and equilibrated for 15 min to the ⁠, appropriate to that sample time, prior to the start of the rate measurement (see below). The open gassing system described by Walsh et al. (1990) was employed with the following humidified CO₂ mixtures (balance air) to maintain in vivo acid-base status: 0.25 % CO₂ for rest, 4h and 12 h samples; 1.00% for Oh samples, 0.50% for Ih samples; and 0.40% for 2h samples.

Blood was withdrawn from the arterial catheters of resting trout (2–5 ml per fish) as described by Walsh et al. (1990), care being taken to avoid struggling or disorientation of the fish. The blood was initially heparinized at 150 i.u. ml⁻¹ and kept on ice for up to 30 min until use. For each run in a matrix experiment, blood from several different fish was pooled. The pooled blood was very lightly centrifuged (500g for 60s), an appropriate volume of plasma removed, and the RBCs resuspended in the remaining plasma to ensure a representative Hct (25–30%) after addition of other solutions. Each pooled preparation was considered as one sample for statistical purposes. In the CO₂ production experiments, it was possible to run all treatments in the matrix simultaneously on each pool. This was not possible in the O₂ consumption experiments because of equipment constraints. Therefore, a slightly different protocol was used in which a subset of the matrix was examined in each pool, but the appropriate control was always included.

In the matrix experiments that examined CO₂ production from [U-¹⁴C]L-lactate, the pooled blood was placed in 20 ml glass vials in 700 μl volumes, each receiving an additional 125 i.u. of heparin. The appropriate volume of a neutralized sodium lactate stock solution (140 mmol l⁻¹) was added to each vial to create a final nominal concentration of 5 mmol l⁻¹. Experimental factors were added by equilibration with elevated (for respiratory acidosis), and/or appropriate volumes of 140 mmol l⁻¹ HC1 (for metabolic acidosis), 140 mmol l⁻¹ NaHCO₃ (for metabolic alkalosis), and a stock solution of 3×1010⁻⁶moll⁻¹ L-epinephrine bitartrate (Sigma) plus 3×10⁻⁷moll⁻¹ L-norepinephrine bitartrate (Sigma) in 140mmol l⁻¹ NaCl (for elevated catecholamines). The final nominal concentrations of added catecholamines were 150 nmol l⁻¹ epinephrine and 15 nmol l⁻¹ norepinephrine. An appropriate amount of 140 mmol l⁻¹ NaCl was added to each vial so that the total sample volume was always 850 μl. In the matrix experiments that examined total O₂ consumption, the treatments were identical, but the initial samples of blood were 1400 μl, so the volumes of all additions were doubled to yield final sample volumes of 1700 μl.

Once all the additions were complete, the vials were placed in a shaking water bath at 10±l°C, and equilibrated for 30 min to the appropriate before the start of metabolism measurements. The open gassing system described by Walsh et al. (1990) was employed, using humidified mixtures of 0.25 % CO₂ (balance air) for normocapnia and 1.00% CO₂ (balance air) for respiratory acidosis. CO₂ production or O₂ consumption rates were then determined as outlined below. At the end of all runs, measurements were made of blood pH (pHe), true plasma ⁠,Hb, Hct and cell mass, either on dummy vials or on samples recovered from the measurement chamber. RBC pHi was also measured in some experiments.

For O₂ consumption measurements in both the exercise and factorial experiments, samples were prepared as outlined above, then transferred by gas-tight Hamilton syringe to an measurement chamber thermostatted to 10±l°C. In the factorial experiments, 5 mmol l⁻¹ sodium lactate was added to all vials. The rate of O₂ consumption was measured over a 15 min period exactly as described by Walsh et al. (1990).

In the exercise experiments, CO₂ production rates were measured separately from [U-¹⁴C]L-lactate (90–105 mCi mmol⁻¹; ICN) and [U-¹⁴C]o-glucose (310 mCi mmol⁻¹; ICN) using the native substrate levels present in the blood, which were measured in both initial and final samples. In the factorial experiments, CO₂ production rates from [U-¹⁴C]L-lactate were measured using nominal added concentrations of 5 mmol l⁻¹ unlabelled sodium lactate; exact concentrations were measured in both initial and final samples. Methodology was identical to that described by Walsh et al. (1990) and gassing conditions and other experimental considerations have been outlined above. Experimental runs lasted 2h, except in one case where a 30min run was carried out on Oh and 2h postexercise samples to ensure that transient events were not being missed. The results were the same as in the 2h runs.

All analytical methods and calculations were identical to those described by Walsh et al. (1990). All values are presented as means±l S.E.M. (N). The statistical significance (P<0.05) of differences was assessed using Student’s two-tailed t-test, paired or unpaired format as appropriate to the design of the experiments. In the factorial experiments, a model III two-way analysis of variance (ANOVA) together with Student-Newman-Keuls and Duncan’s new multiple-range test was employed (Zar, 1974).

We have chosen to report CO₂ production rates of RBCs from specific substrates in units of /rmolCO₂g⁻¹ cell mass h⁻¹, the standard unit in the fish cell metabolism literature (e.g. French et al. 1981; Mommsen et al. 1988; Perry and Walsh, 1989; Walsh, 1989). In contrast, we have reported RBC O₂ consumption rates in terms of nmol O₂g⁻¹Hb min⁻¹, the standard unit in the fish red cell respiration and acid-base literature (e.g. Ferguson and Boutilier, 1988; Ferguson et al. 1989). A reported advantage of the latter is that it corrects for changes in RBC mass occurring simply because of water flux, i.e. swelling and shrinking. The mean cell haemoglobin concentration (MCHC=Hb/Hct), which is often considered an index of this water flux (e.g. Milligan and Wood, 1987; Ferguson et al. 1989), was monitored in all experiments. While MCHC did change significantly as a result of some treatments, the absolute changes were never greater than 25 %, and would not have affected any of the basic trends in metabolism seen in either data set. At a typical MCHC of 0.25 g Hb ml⁻¹ RBC, the conversion is 1 μmol g⁻¹ cell h⁻¹=67 nmol g⁻¹ Hb min⁻¹.

Strenuous exercise in rainbow trout, using an established protocol (Turner et al. 1983; Milligan and Wood, 1986a,b, 1987), caused typical changes in blood acid-base status and catecholamine and metabolite levels (Figs 1,2,3). Arterial blood displayed a pronounced mixed respiratory/metabolic acidosis, as shown by an immediate elevation of and longer-lasting depression of plasma HCO₃” concentration (Fig. 1A,B,C). The acidosis was greatest at 0-1h post-exercise, and had been corrected by 4h with a slight alkalotic overshoot which persisted at 12 h. RBC pHi, in contrast, was extremely well regulated in the face of this extracellular acidosis. RBC pHi fell only marginally at 1 h (by 0.06 units relative to an extracellular drop of 0.32units), was completely compensated at 2h, and manifested a significant alkalosis at 4 and 12 h (Fig. 1A).

Plasma catecholamine levels were elevated dramatically at 0 h post-exercise, but declined rapidly thereafter (Fig. 2). Epinephrine was the dominant catecholamine both at rest and after exhaustive exercise, approximately 10-fold higher than norepinephrine, with a persistent slight elevation at 1 and 2h. MCHC (not shown) fell significantly from 0.251±0.011 (16) g ml⁻¹ at rest to 0.224±0.007 (15) g ml⁻¹ at Oh, indicating a slight swelling of the RBCs, but there were no other significant changes. Swelling may have been slightly lower than prior studies because the changes in and pHe were not as pronounced in these fish. Plasma lactate levels increased immediately post-exercise, reaching a peak of about 12 mmol l⁻¹ at 1–2 h and declining thereafter; levels remained significantly elevated at 12 h (Fig. 3A). Plasma glucose levels exhibited a biphasic response (Fig. 3B) with a significant immediate rise at 0 h, a return to resting levels at 1 and 2 h, and a second elevation at 4 and 12 h. On both a relative and an absolute basis, these changes in glucose were much smaller than those in lactate.

Metabolic rates of the RBCs were measured directly on samples of this same blood under acid-base conditions closely approximating those in vivo. At rest, CO₂ production from glucose was approximately three times higher than from lactate, as reported by Walsh et al. (1990). However, this situation changed markedly after exercise (Fig. 4). Glucose oxidation was depressed by 30-50%, despite the fact that plasma glucose levels stayed the same or increased (Fig. 3B). The depression became significant at 1 h and roughly coincided with the peak of plasma lactate (Fig. 3A). Glucose oxidation thereafter recovered to control levels at 4 and 12 h. CO₂ production from lactate, in contrast, increased by about 250% immediately after exercise, and reached a peak 450% increase at 2h. By 12 h, lactate oxidation had returned to resting levels. Thus, at 1–2 h post-exercise, lactate oxidation was 3–4 times faster than glucose oxidation. These changes in lactate oxidation closely followed the changes in plasma lactate concentrations (cf. Fig. 3A). However, it is apparent that rates at 0, 1 and 2h post-exercise were elevated substantially beyond the levels predicted by the measured plasma lactate concentrations (Fig. 3A) and the kinetic relationship established for resting blood (cf. Fig. 1 of Walsh et al. 1990, and present Fig. 4). Conversely, at these same times, rates of CO₂ production from glucose were depressed below the levels predicted by the measured plasma glucose concentrations (Fig. 3B) and the comparable kinetic relationship for glucose oxidation by resting blood (cf. Fig. 1 of Walsh et al. 1990, and present Fig. 4).

Despite these large variations in blood chemistry and in rates of glucose and lactate oxidation after exhaustive exercise, rates of total O₂ consumption by RBCs stayed remarkably stable (Fig. 5). The only significant change was a 20 % decline in at 2h, a time when the blood was still acidotic (Fig. 1A), but circulating catecholamines had returned close to resting levels (Fig. 2). On an absolute basis, combined CO₂ production from both glucose and lactate remained a small fraction of total O₂ consumption, though this fraction increased from about 10 % at rest to 20–25% post-exercise. Clearly, other substrates were of greater quantitative importance.

Since the exercise regimen produced a large number of changes in acid-base variables and catecholamine and substrate concentrations, it was difficult to determine which blood variables, if any, were responsible for the observed metabolic changes. The influence of substrate concentration has been examined by Walsh et al. (1990). Therefore, here we assessed the independent and combined effects of the various acid-base components and catecholamines on RBC lactate oxidation and in vitro under simulated Oh post-exercise conditions. Lactate (as sodium lactate) was added to all treatments at 5 mmol l⁻¹, the measured Oh level (Fig. 3A). Metabolic acidosis was simulated by the addition of 5 mmol l⁻¹ HC1, because at Oh, the lactate and ‘metabolic acid’ loads to the blood were the same, as calculated from the data of Fig. 1 (see also Turner et al. 1983; Milligan and Wood, 1986a). Respiratory acidosis was simulated by 1% CO₂ (1kPa), somewhat higher than the mean measured value at 0 h in the present experiments (Fig. 1B), but representative of that recorded in several previous studies (Turner et al. 1983; Milligan and Wood, 1986a). Catecholamine additions (150nmolC¹ epinephrine plus 15 nmol l⁻¹ norepinephrine) were close to measured elevations in vivo at Oh (Fig. 2).

Table 1 summarizes the measured acid-base parameters and MCHC from each treatment in these factorial experiments. Respiratory and metabolic acidoses caused the same depressions in pHe and pHe-pHi, whereas the combined respiratory and metabolic treatment (‘mixed acidosis’) caused larger depressions in both parameters. Respiratory acidosis raised plasma HCO₃⁻ concentration metabolic acidosis lowered it, while the mixed acidosis caused no significant change. Metabolic alkalosis raised pHe and plasma HCO₃^-, the latter to the same level as induced by respiratory acidosis. All acidotic treatments induced a swelling of the RBCs, as shown by a decline in MCHC, which was greatest for the mixed acidosis. Metabolic alkalosis caused a slight rise in MCHC. The addition of catecholamines caused a decrease in pHe-pHi and a decrease in MCHC, indicative of active RBC pHi regulation and swelling, effects that were significant in all treatments except metabolic acidosis. Catecholamines did not significantly affect pHe or plasma HCO₃⁻ in any treatment.

The factorial analysis demonstrated that the rate of lactate oxidation by RBCs was sensitive to blood acid-base status (Fig. 6). Respiratory acidosis (which Lowered pHe and raised plasma [HCO₃^-]) and mixed acidosis (which lowered pHe even more without changing plasma [HCO₃^-]) both caused a significant 40% increase in the rate. However, metabolic acidosis alone (which lowered pHe by the same amount as respiratory acidosis, but also lowered plasma [HCO₃⁻]) had no effect. These results suggest that lactate oxidation is enhanced by increased and/or [HCO₃⁻], while pHe is without influence. However, when the cells were subjected to metabolic alkalosis (which raised pHe and elevated plasma [HCO₃⁻] to the same level as in respiratory acidosis), there was no effect on lactate oxidation. This indicates that the rate enhancement during respiratory and mixed acidoses was due solely to increased and not to changes in either [HCO₃⁻] or pHe.

Elevated plasma catecholamine levels also significantly altered the rate of lactate oxidation, but in an interactive manner with acid-base status (Fig. 6). Under control conditions, catecholamines were without effect. However, under all three types of acidosis (respiratory, metabolic and mixed), the rate was significantly enhanced by typical post-exercise levels of epinephrine and norepinephrine. In each case, the elevation was approximately 60 % of the control rate. Thus, the combination of respiratory (or mixed) acidosis plus catecholamines accounted for about a 100 % increase.

The rate of total O₂ consumption by RBCs was also sensitive to acid-base status and catecholamines, but in a very different pattern from that of lactate oxidation (Fig. 7). Both respiratory and metabolic acidoses caused a 35 % decrease in however, these effects were not additive, because mixed acidosis (where the pHe decrease was greater) also caused a 35 % decrease. In contrast, metabolic alkalosis induced a significant 30% increase above control ⁠. In general, these results indicate that pHe, but not or plasma [HCO₃⁻], was the important acid-base variable affecting RBC ⁠.

The role of catecholamines was again interactive with acid-base status (Fig. 7). Elevated epinephrine and norepinephrine levels did not affect the control ⁠, but significantly enhanced the lower rates seen under all three acidotic conditions, returning them close to the control rate. Thus, catecholamines served to reverse the depression in total aerobic metabolism of RBCs caused by acidosis.

During strenuous exercise and recovery in rainbow trout, there were distinct variations in blood acid-base status, lactate and glucose concentrations and levels of circulating catecholamines (Figs 1,2,3; see also Turner et al. 1983; Primmett et al. 1986; Milligan and Wood, 1986a,b, 1987). The relative oxidation rates of lactate and glucose by RBCs in vivo would be expected to vary under these circumstances, based solely on the observed changes in substrate concentrations |(Fig. 3) and the pathway K_m values for lactate and glucose determined by Walsh et al. (1990). However, observed changes in CO₂ production rates from lactate or glucose following strenuous exercise did not entirely follow these simple kinetic patterns (Fig. 4). Lactate oxidation was elevated above the predicted level, while glucose oxidation was depressed below the predicted level during the 0–2 h postexercise period. Subsequent experiments designed to factor out the potential effects of the many altered blood variables (Fig. 6) demonstrated that: (i) CO₂ production from lactate was stimulated by the respiratory component of the postexercise acidosis (i.e. elevated ⁠); and (ii) elevated plasma catecholamine levels were effective in stimulating lactate oxidation under the acidotic conditions prevailing post-exercise (but not under control conditions). The combination of these two factors with the increased blood lactate levels was sufficient to explain the observed lactate oxidation rates, at least at 0 and 1h post-exercise. The mechanisms of the and catecholamine effects are not known, though direct actions on lactate entry appear unlikely, for this is a passive process, at least in tuna, Katsuwonus pelamis, RBCs (Moon et al. 1987). It is of interest to ask whether these elevated rates of lactate oxidation could serve as a significant route for lactate clearance from the bloodstream after exhaustive exercise. While exact calculations are problematical (see also Milligan and McDonald, 1988), the answer is clearly negative -the observed rates would remove less than 0.5% h⁻¹ of the peak blood lactate load.

The depression in glucose oxidation 0–2 h after exercise (Fig. 4) in the face of unchanged or increased plasma glucose concentrations (Fig. 3B) was at least partially explained by the increase in plasma lactate levels at these times. Walsh et al. (1990) demonstrated that elevated lactate levels had a small inhibitory effect on glucose oxidation at high concentrations of glucose and lactate. Additional factors that were not assessed in the present study could include the post-exercise changes in acid-base variables and catecholamines, though Tse and Young (1990) have reported that catecholamines did not affect glucose uptake by trout RBCs. Furthermore, it should be remembered that our methods utilized uniformly labelled [¹⁴C]glucose and thus measured CO₂ production from both the tricarboxylic acid cycle and the pentose-phosphate shunt (Wood and Katz, 1958; Walsh et al. 1990). It is possible that the marked depression in CO₂ production from glucose included a decrease in pentose-shunt activity, a pathway that produces NADPH and five-carbon sugars for biosynthesis. Clearly the control of glucose metabolism by RBCs in vivo deserves further study.

Total O₂ consumption rates of RBCs in vivo were remarkably stable after exhaustive exercise (Fig. 5), despite the clear evidence of RBC pHi regulation (Fig. 1A) and elevation in plasma catecholamine levels (Fig. 2). In vitro, Ferguson and Boutilier (1988) and Ferguson et al. (1989) have reported that adrenergically induced pHi regulation is accompanied by a marked increase in RBC ⁠,. the presumed metabolic cost of activation of the secondarily active Na⁺/H⁺ exchange. However, these results were obtained with 5 x 10⁻⁴ mol l⁻¹ epinephrine, more than three orders of magnitude greater than the physiological levels of catecholamines measured in vivo after exhaustive exercise in the present study. Interestingly, Ferguson et al. (1989) also reported the results of an in vivo exercise experiment rather similar to that of the present study, though single measurements were made over a 2h period starting at Oh post-exercise. They found no consistent response -in one experimental group, RBC ⁠, increased significantly after exercise, while in another it exhibited a non-significant decrease. It is notable that, in the present study, catecholamines only stimulated O₂ consumption under conditions of low pHe, where it is well known that adrenergically stimulated Na⁺/H⁺ exchange is potentiated.

The present in vivo results (Fig. 5) were entirely consistent with the results of the in vitro factorial experiments (Fig. 7), which showed that: (i) was inhibited by all types of acidosis, an effect attributable mainly to the fall in pHe; and (ii) this inhibition was reversed by catecholamines, though these had no effect under control conditions. Thus, the stability of RBC in the face of postexercise acidosis was explained by the protective effect of elevated levels of circulating catecholamines. This inhibitory effect of acidosis became evident in vivo at 2h (Fig. 5), when the elevation in plasma catecholamine levels had largely disappeared (Fig. 2) but acidosis persisted (Fig. 1A). In vitro, physiological levels of epinephrine and norepinephrine never elevated RBC above control levels, yet Na⁺/H⁺ exchange was clearly activated, as shown by decreases in the pHe-pHi gradient and MCHC (Table 1). These results suggest that the aerobic cost of RBC pHi regulation may be lower than estimated by Ferguson and Boutilier (1988) and Ferguson et al. (1989) when physiological levels of catecholamines are involved. In support of this argument, decreases in RBC NTP stores after exhaustive exercise in vivo (Milligan and Wood, 1987) were also much smaller than those seen with high levels of epinephrine in vitro (Ferguson and Boutilier, 1988; Ferguson et al. 1989). Indeed, the constancy of RBC in vivo after exercise means that the erythrocytes consume an even smaller percentage of whole-animal O₂ uptake than the 2–3 % at rest (Walsh et al. 1990). One note of caution should be added regarding the correlation of our in vivo and in vitro results: we cannot eliminate the possibility that additional red blood cells (with different metabolic capabilities) were recruited during exercise by release from the spleen.

It is now becoming evident that acid-base effects on metabolism are widespread in fish, but their nature appears specific to the individual tissue, pathway and species, and no general conclusions can yet be drawn. Indeed, even within the relatively narrow area of aerobic respiration by trout RBCs, there is a marked dichotomy in the present results. Thus, elevated stimulated lactate oxidation while pHe and HCO₃⁻ had little effect (Fig. 6). This is the first demonstration to our knowledge of a stimulatory effect of on CO₂ production from a single substrate in fish tissue. However, total was depressed by respiratory acidosis, an effect largely correlated with the drop in pHe, with little influence of or [HCO₃⁻] (Fig. 7). By way of contrast, elevated and lowered pHe independently inhibited lactate oxidation and gluconeogenesis in trout hepatocytes, while elevated HCO₃⁻ stimulated these processes (Walsh et al. 1988; Mommsen et al. 1988). However, in toadfish (Opsanus beta) hepatocytes, lowered pHe stimulated gluconeogenesis from lactate (Walsh, 1989), while elevated [HCO₃⁻] stimulated urea synthesis (Walsh et al. 1989); in both cases, other acid-base variables had no effect. Clearly a variety of patterns exists, but almost nothing is known regarding the specific mechanisms involved.

The role of catecholamines in metabolic regulation is becoming more obvious, from both the present study and other recent investigations (Walsh et al. 1988; Mommsen et al. 1988; Moon et al. 1988; Perry et al. 1988; Wright et al. 1989). In general, these substances appear to have little influence under control circumstances, when they would not normally be mobilized anyway. However, at times of acid-base disturbance, they become extremely important in stimulating, protecting or modifying metabolism, thereby fulfilling their basic role in adaptation to stress. This is particularly clear in the trout RBC, where catecholamines have small or negligible effects on pHi, haemoglobin O₂-affinity (Nikinmaa and Tufts, 1989) and oxidative metabolism under normal conditions, but extremely important actions during acidotic or hypoxic stresses. Trout RBCs represent an appealing experimental system for further studies of the interrelationships between acidbase status, catecholamines and metabolic control at the molecular level.

This research was supported by grants from NSERC of Canada (CMW and SFP), NSF and NIH (PJW) and CNRS (ST). We thank Richard Kinkead and Tina Goodison for valuable assistance, and Dr Tom Mommsen for commenting on an earlier draft of this paper.