Regulation of Hepatic Gluconeogenesis and Glycogenolysis By Catecholamines in Rainbow Trout During Environmental Hypoxia

When trout encounter environmental hypoxia they respond initially by lowering metabolic rate, then by increasing anaerobic metabolism (Boutilier et al. 1988).

Some tissues are glucose-dependent (e.g. nervous tissue), but have little reserve capacity and, for this reason, it is important that plasma glucose levels are regulated. Moreover, rapid increases in glucose availability are often required to meet the increased metabolic demands of tissues. It is not surprising, therefore, that catecholamine levels, well known as mediators of rapid physiological change, increase markedly in arterial blood during exposure to low environmental oxygen (Butler et al. 1979; Tetens & Christensen, 1987; Fievet et al. 1987; Boutilier et al. 1988). The purpose of this study was to determine whether catecholamines play a role in the regulation of glucose metabolism in vivo during short-term moderate hypoxia.

Catecholamines exert hyperglycemic effects in fish (e.g. Thorpe & Ince, 1974; deRoos & deRoos, 1978; Morata et al. 1982; Ottolenghi et al. 1984, 1985, 1986; Janssens & Lowrey, 1987; Mommsen et al. 1988), which are due, in part, to activation of glycogenolysis (Nakano & Tomlinson, 1967; Birnbaum et al. 1976; Rush & Umminger, 1978; Ristori & Laurent, 1985; Mommsen et al. 1988). Epinephrine also activates gluconeogenesis, but its relative importance is not clear. Mommsen et al. (1988) reported that glycogenolysis accounted for 97 % of glucose production in trout hepatocytes with or without epinephrine. Ottolenghi et al. (1985), however, concluded that in isolated, perfused catfish livers, glycogenolysis dominated at high doses of epinephrine, but gluconeogenesis was the dominant source of glucose at low doses.

The role of catecholamines in glucose metabolism has often been studied in vitro, but less frequently in intact fish. One of the difficulties with using hepatocytes in vitro is that freshly isolated cells are in negative glycogen balance and rates are unlikely to reflect those found in situ (Mommsen et al. 1988). In this study we use intact fish to eliminate this in vitro preparation artifact.

A second problem addressed by this study relates to methodological differences in the measurement of glycogen phosphorylase (GPase) activity. Wosilait & Sutherland (1956) first separated the two forms of GPase (b, inactive; a, active) by measuring activities in the absence and in the presence of AMP, respectively. Unless caffeine is present, however, the activity of rat liver GPase b is significantly stimulated by AMP (Stalmans & Hers, 1975). This may explain the consistently high liver resting percent GPase a [= a/(a+b)] values reported in the literature (90–95%, Ottolenghi et al. 1985, 1986; 83%, Perry et al. 1988; 72%, Mommsen et al. 1988). Such high percent GPase a values suggest that GPase is almost completely activated at rest, a situation which affords little possibility of regulatory control. In the present study, we rectified this technical problem by applying the criteria of Stalmans & Hers (1975) to GPase activities.

We have analyzed the individual effects of epinephrine, norepinephrine and hypoxia (with/without β-adrenoreceptor blockade) on tissue glucose and glycogen levels, and on the kinetic behavior of two key enzymes involved at critical points in glucose metabolism, glycogen phosphorylase and pyruvate kinase. PK was measured because of its key role in regulating metabolic flux through glycolysis and gluconeogenesis (Pilkis et al. 1988).

Rainbow trout (Salmo gairdneri) of both sexes, weighing 150-250 g, were obtained from Thistle Springs Trout Farm (Ashton, Ontario) and maintained as reported previously (Perry & Vermette, 1987) at seasonal temperatures (15–18°C). Preliminary PK kinetic properties were determined on resting intact fish, held in individual, opaque, Perspex chambers. Fish were killed by a sharp blow to the head, the liver was immediately removed, and frozen between aluminum blocks cooled in liquid N₂. Tissue samples were stored at −70 °C until analysis.

In all other experiments, trout were anesthetized, surgically fitted with a dorsal aortic (DA) catheter (Soivio et al. 1972), and left to recover for 48 h in individual Perspex chambers. Three series of experiments were performed (series I, II and III) and in each series there were four separate groups of fish (Table 1). Experiments were performed over the summer season when ambient water temperature increased from 15 °C to a peak of 18°C. Fish in series I were held at 15 °C, series II at 18°C and series III at 16°C.

In each series, fish were pre-injected with either heparinized, physiological saline (Wolf, 1963) or propranolol, which was followed 1 h later by the experimental treatment of saline, epinephrine or norepinephrine injection, or exposure to external hypoxia (Table 1). Propranolol was dissolved in heparinized saline immediately before use, adjusted to pH 7·8, and injected over a 30 s period into the DA catheter. The volume injected was 0·2 ml at a dose of 2 mg kg⁻¹ body mass and was followed by an infusion of 0·3 ml of saline to flush the catheter. The calculated circulating level of propranolol was approximately 5×10⁻⁵moll⁻¹, about three orders of magnitude greater than peak catecholamine levels (see Fig. 2). Epinephrine and norepinephrine solutions (bitartrate salts; pH 7·8) were prepared immediately prior to infusion. The volume injected was 0-2 ml at an appropriate dose to achieve a circulating epinephrine/norepinephrine concentration of 5×10⁻⁸moll⁻¹, followed by a 0·3 ml saline bolus to flush the catheter.

In series I and II, an initial blood sample (0·8–1·0 ml) was collected 30 min after the pre-injection and a final blood sample (0·8–l·0ml) was collected 15 min after the experimental injection (Table 1). The initial blood sample was taken for control measurements, and compared to blood samples taken 15 min after the experimental injection.

In series III, the degree and duration of hypoxia were chosen to induce a marked increase in circulating catecholamines within a minimum time (Boutilier et al. 1988). External hypoxia was achieved by gassing the inflowing water with 100% N₂ in a counter-current gas-exchange column. By carefully adjusting the water and gas flow rates through the column, the oxygen tension of the inflowing water was kept constant for a 30min experimental period.

In all series, fish were killed by a sharp blow to the head immediately after the final blood sample had been withdrawn through the DA catheter. The liver was excised and frozen between aluminum blocks cooled in liquid N₂. White muscle samples (0·5–1·0 g) were then excised and quickly frozen in the same manner. Tissue samples were stored at −70°C until analysed (<24h for glucose and glycogen, <4 days for liver enzyme analyses).

Blood was analysed for pH (pHe), red cell pH (pHi), plasma glucose and catecholamine levels. In series III, additional measurements were made of whole blood oxygen tension content and plasma lactate ([lactate]). pHe, and were determined as described by Vermette & Perry (1988). Plasma was separated by centrifugation (2min at 13000g), acidified (0·6 moll⁻¹ HClO₄), and stored (−20°C) for later determination of plasma glucose levels. Plasma samples (100–200 μl) were also stored for subsequent determination of catechol amines and [lactate] (series III only) at −70 °C. Plasma glucose and lactate levels were measured enzymatically according to Bergmeyer (1974). Epinephrine and norepinephrine levels were measured by HPLC and electronic detection, as described in detail by Woodward (1982).

Liver and muscle glucose and glycogen concentrations were determined in the following manner. Frozen tissue samples were weighed, deproteinized in iced HC1O₄ (0·6moll⁻¹, 1:4 dilution), and homogenized. A sample of neutralized supernatant was hydrolyzed with amyloglucosidase (in acetate buffer, pH4·6–4·8) and glucose was determined as above. Glycogen (in μmol glucosyl units g⁻¹ tissue) was determined as the difference in glucose concentrations in the hydrolyzed and nonhydrolyzed samples.

The activities of GPase and PK were determined spectrophotometrically by recording the change in absorbance of NAD(P) at 340nm (E_mu = 6·22). Liver tissue was sonicated (Koutes MicroUltrasonic cell disrupter) in a phosphorylation-déphosphorylation ‘stopping buffer’ (5:1 w/v; 50 mmol 1⁻¹ imidazole, 15 mmol 1⁻¹β-mercaptoethanol, 100 mmol 1⁻¹ KF, 5 mmol 1⁻¹ EDTA, 5 mmol 1⁻¹ EGTA, pH7·5, at room temperature; after Stalmans & Hers, 1975), layered onto a 3 ml Sephadex (G-25) column equilibrated with stopping buffer, and centrifuged at low speeds according to Christopherson (1983) to remove small metabolites. PK activity was determined in 50 mmol 1⁻¹ imidazole buffer (pH 7·5) containing 10mmoll⁻¹ MgCl₂, 2·5 mmol 1⁻¹ ADP, 100mmoll⁻¹ KC1, 0·12 mmol 1⁻¹ NADH, excess lactate dehydrogenase (10 units) and either 0·2, 0·4 or 5·0 mmol 1⁻¹ phosphoenolpyruvate (PEP, omitted for control). GPase activity was determined in 50 mmol 1⁻¹ phosphate buffer (pH 7·0) containing 0·4 mmol 1⁻¹ NADP⁺, 15 mmol 1⁻¹ MgSO₄, 0·2 mmol 1⁻¹ glucose-l,6-bisphosphate, 2 mmol 1⁻¹ AMP, excess glucose-6-phosphate dehydrogenase and phosphoglycerate mutase (each l unit), 2 mg ml⁻¹ dialyzed glycogen (omitted for control) and either 0·25mmoll⁻¹ EDTA or 10mmoll⁻¹ EDTA and 10mmoll⁻¹ caffeine. Caffeine inactivates the inactive or b form of GPase and therefore provides a direct measure of the active or a form of the enzyme. GPase a activity is defined as that in the presence of caffeine and total GPase as activity without caffeine (Stalmans & Hers, 1975).

Cyclic AMP was measured in liver tissue using a cyclic-[³H]AMP kit (Amersham Corp.). Tissues were homogenized in 0·7% perchloric acid containing EDTA, neutralized and analyzed as directed.

Results are given as means ± 1 standard error (S.E.) of the mean. Initial and final blood sample values were compared using a Student ‘s paired r-test (P < 0·05). Comparisons between groups of fish were performed using a Student ‘s unpaired t-test (P<0·05).

The substrate-dependence of pyruvate kinase is shown in Fig. 1A. The substrate concentration for half-maximal saturation (S_0·5) for the enzyme with respect to phosphenolpyruvate (PEP) was 0·60 mmol 1⁻¹ (Fig. IB) and V_m at 15°C was 16·1 μmolg⁻¹ min⁻¹ (Fig. 1A). The Hill constant (n_H), calculated from the slope of the fine in Fig. 1B, was 1·80. An n_H value greater than 1 indicates positive operativity in the binding between ligand and regulatory enzyme, L-Alanine mmol 1⁻¹) did not significantly modify the kinetic properties of PK with respect to PEP (S_0·5 = 0·66 mmol 1⁻¹, V_m = 15·7 μmolg⁻¹ min⁻¹, n_H = l ‘84; Fig-1A,B); this effect was independent of alanine concentration or the presence of the activator fructose-l,6-bisphosphate (F-1,6-BP) (Fig. 1D). In contrast, F-1,6-BP (0-lmmoll⁻¹) had a strong activating effect, decreasing S_0·5 from 0·53 to 0·07 mmol 1⁻¹ (Fig. 1A,B). This stimulation of PK by F-1,6-BP occurs at low activator concentrations (Fig. 1C), with a K_a value of 0·60μmoll⁻¹. F-1,6-BP decreased the n_H value from 1·80 to 0·99, indicating an absence of cooperative between ligand (PEP) and enzyme.

It is not possible to determine the S_0·5 for GPase with respect to glycogen because of the presence of unknown, but substantial, levels of endogenous tissue glycogen.

It is clear that plasma catecholamine levels increased in response to epinephrine and norepinephrine injections and hypoxia (Fig. 2). Because of the large variability in the magnitude of the responses, however, the changes were generally not significant. Control catecholamine levels were approximately 1 nmoll⁻¹, except for fish pre-treated with propranolol, where initial levels were two-to sevenfold higher (Fig. 2A, 2B). With epinephrine and norepinephrine injections, the final plasma concentration was 60 nmol I⁻¹ in both cases. Hypoxia elevated epinephrine (22 nmol I⁻¹, Fig. 2A) and, to a smaller extent, norepinephrine levels (7 nmol 1⁻¹, Fig. 2B). Pre-treatment with propranolol prior to hypoxia resulted in a considerably larger total catecholamine release to the plasma (140nmol 1⁻¹, Fig. 2C).

Epinephrine elevated plasma glucose levels (46%) relative to initial values (Table 2A). Liver glucose also increased compared with control values, but this was not significant. There were no changes in muscle glucose levels and, although muscle glycogen levels decreased, high variability masked any significant trends. Pre-injection with propranolol eliminated the increase in glucose following epinephrine treatment, implying that the epinephrine effect was mediated by β-receptors. Propranolol alone had no significant effect on glucose, but liver and muscle glycogen levels were well above values in the other treatments. pHe and pHi were not altered by any of the experimental treatments (Table 2B).

Epinephrine significantly modified PK activity (Table 3A). At low PEP concentrations (0·2 mmol 1⁻¹), activity was reduced two-to fivefold compared with the control and the other treatments. The ratio of activity rates measured at low (0·2 mmol 1⁻¹) and saturating (5 mmol 1⁻¹) concentrations of PEP provides an indication of the phosphorylation state of PK: that is, a decrease in the ratio corresponds to an increase in PK phosphorylation. The ratio of activities (Table 3A), was significantly lower in fish treated with epinephrine compared with values for saline and both propranolol treatments; as maximal activities (5 mmol 1⁻¹ PEP) remained unchanged, this decrease indicates an increase in the S_0·5 (PEP) value. Propranolol alone increased the ratio and eliminated the inhibitory effect of epinephrine on PK.

In preliminary experiments, PK activities were estimated in the presence of μmoll⁻¹ L-alanine and 0·1 mmol 1⁻¹ F-1,6-BP under all experimental treatments. F-1,6-BP had a strong stimulatory effect on PK activity, increasing the 0·2/5·0PEP ratio; however, this effect was independent of experimental treatment. Alanine did not significantly modify PK activity in any of the experimental manipulations. These data are not shown since they do not add any new information to that presented in Fig. 1.

GPase values in epinephrine-treated fish were not significantly different from control values. The active form of GPase (percent GPase a), total GPase (a+b) and GPase a activities, however, were significantly higher in epinephrine-treated fish than in the propranolol control and propranolol + epinephrine experiments (Table 3B). Propranolol alone significantly reduced GPase activity (total and a) compared with the saline control value.

The results of the norepinephrine experiments (Tables 4, 5) were similar to those of series I for epinephrine. The major difference was that at the same circulating concentration as epinephrine (60 nmol 1⁻¹, Fig. 2B), the effects of norepinephrine on the measured variables were smaller. Following norepinephrine injection, plasma glucose level increased significantly (16 %), but this increase was only half of that observed with epinephrine (Table 2). Pre-injection with propranolol eliminated the increase in glucose associated with norepinephrine treatment, again suggesting mediation by β-receptors. pHe and pHi were not altered by any of the experimental treatments (Table 4B).

PK activity at 0·2 mmol 1⁻¹ PEP was reduced by norepinephrine compared to the other treatments, but this effect was not significant. Norepinephrine, however, did significantly depress the 0·2/5·0 PEP ratio compared with the saline control value.

Norepinephrine increased total GPase, GPase a and percent GPase a activities (Table 5A), but these trends were not significant.

Plasma epinephrine and norepinephrine were both elevated after 30 min of hypoxia, although epinephrine levels were three times those of norepinephrine Fig. 2). Plasma glucose levels were constant during hypoxia. The combination of propranolol and hypoxia, however, caused a significant reduction of plasma glucose (20%) compared to pre-treatment levels. There were no significant changes in liver or muscle glucose and glycogen levels. Red cell pHi was significantly elevated during hypoxia, which indicates β-adrenoreceptor stimulation (Tetens & Christensen, 1987). Blood and were severely depressed during hypoxia. Pre-treatment with propranolol significantly reduced initial blood oxygen levels, and hypoxia compounded this effect by lowering and values below those seen with hypoxia alone (Table 6C). Hypoxia resulted in a three-to fourfold increase in plasma [lactate], indicating an activation of anaerobic metabolic pathways.

PK activity remained-constant during hypoxia (Table7). Pre-treatment with propranolol plus hypoxia, however, caused a significant increase in PK activity at 0·2mmoll⁻¹ PEP, and an increase in the 0·2/5·0PEP ratio (Table 7). Propranolol alone decreased maximal PK activity (at 5·0 mmol 1⁻¹ PEP), although a similar change at 0·2 mmol 1⁻¹ PEP meant that there was no change in the 0·2/5·0PEP ratio.

Hypoxia increased total GPase activity by two-to threefold over saline and propranolol control treatments. GPase a activity was 5–7 times greater during hypoxia compared to values for propranolol treatment with or without hypoxia. Percent GPase a activity was significantly higher during hypoxia compared with values during treatments with propranolol + hypoxia (Table 7).

Liver cyclic AMP levels increased slightly with epinephrine, norepinephrine and hypoxia, but none of these treatments resulted in significant changes compared with control levels (Table 8).

Pyruvate kinase in teleost liver tissue was initially reported to be nonregulatory (Somero & Hochachka, 1968; Guderley et al. 1978; Moon & Hulbert, 1980; Guderley & Cardenas, 1980). The results of this study, however, and those of four recent investigations (Petersen et al. 1988; Rahman & Storey, 1988; Sand, 1988; Moon et al. 1989) indicate the contrary. An explanation for this discrepancy is that differences in the physiological state of the fish and/or the preparation of the liver homogenate result in variable states of phosphorylation of the enzyme. The sensitivity of PK to allosteric modulators is dependent on whether the enzyme is phosphorylated (inactive state) or dephosphorylated (active state) (Pilkis et al. 1988). Furthermore, the procedure for isolating hepatocytes in rats, which is similar to that in trout, results in the progressive dephosphorylation of PK (Riou et al. 1985). The substrate concentration required for half-maximal saturation (S_{0 ·5}) dependent on the phosphorylation state, and this may explain why our in vivo S_0·5value (0·6mmoll⁻¹ PEP) is an order of magnitude greater than the measured by Petersen et al. (1988) in trout hepatocytes. Hence, kinetic parameter for PK measured in this study, using freeze-clamped liver tissue and a phosphorylation-dephosphorylation stopping buffer, may be closer to in situ values than those measured on isolated hepatocytes or in many of the previous studies made on this enzyme in fish.

Alanine had virtually no effect on PK activity in the present study, but is a strong inhibitor of mammalian liver and invertebrate hepatopancreas PK (see Munday et al. 1980). Moon et al. (1989) have also shown an effect of alanine on rainbow trout liver PK; the reason for this discrepancy is not understood. In contrast, relatively low concentrations of F-1,6-BP activated PK, lowered the S_0·5 value for PEP and modified the shape of the PEP saturation curve to a hyperbola. The sensitivity of trout PK to F-1,6-BP activation has been reported previously in rainbow trout (Moon et al. 1989), goldfish (Rahman & Storey, 1988) and flounder (Sand, 1988).

The net rate of glycogen breakdown depends largely on the relative activity of the active form of GPase or percent GPase a. Glycogenolysis is stimulated during stress in many vertebrates (Hems & Whitton, 1980) and, therefore, one would expect activity of the a form to be low in resting animals and increase with stress. Published liver percent GPase a values for resting fish are high, ranging from 60 to 90% (Ottolenghi et al. 1985, 1986; Perry et al. 1988; Mommsen et al. 1988; Moon et al. 1989), suggesting near maximal rates of glycogen breakdown under restine conditions. These high percent GPase a values may possibly be an artifact of the assay technique. Stalmans & Hers (1975) reported that AMP activated both the a and b forms of GPase and only when caffeine was added to liver homogenates was the b form completely inhibited. Their technique provided a more accurate estimate of percent GPase a. This study is one of the first to use the criteria of Stalman & Hers (1975), and reports percent GPase a values of between 10 and 50 %, depending upon conditions. These values, to our knowledge, are the lowest reported for any teleost fish, including a recent study on rainbow trout (Moon et al. 1989), or for mammalian liver (e.g. 70–85 % ; Stalmans & Hers, 1975; Hems & Whitton, 1980).

It should be noted that total GPase (a+b) was not constant between different experimental treatments (Tables 3B, 5B, 7B). Although this observation has been reported by others (Ottolenghi et al. 1986, 1988; Mommsen et al. 1988; Perry et al. 1988), we do not have an explanation for these results.

The results of this study support the hypothesis that catecholamines, released during hypoxia, have an important role in regulating the availability of glucose. Injections of epinephrine or norepinephrine into resting fish increased plasma glucose levels, and this hyperglycemia was blocked by pre-treatment with the β-adrenoreceptor antagonist propranolol. During hypoxia, plasma catecholamine is increased but there was no change in plasma glucose levels. We suggest that reason we did not observe catecholamine-mediated hyperglycemia was because there was an increase in tissue glucose utilization, as a result of inhibition of aerobic, and activation of anaerobic, metabolism (Dunn & Hochachka, 1987; Boutilier et al. 1988). Plasma lactate levels increased fourfold during hypoxia, supporting an enhancement in anaerobic metabolism. Finally, plasma glucose levels were significantly depressed during hypoxia after blockade of β-adrenoreceptors. This indicates an essential role for β-adrenoreceptor-mediated mobilization of glucose during hypoxia to meet metabolic requirements.

Catecholamines regulate glucose metabolism, in part by modifying the activity of two key regulatory enzymes: PK, a regulatory enzyme in glycolysis and gluconeogenesis, and GPase, the enzyme directly responsible for glycogen breakdown. Injections of epinephrine, and to a lesser extent norepinephrine, increased the percent GPase a (statistically significant compared with propranolol-treated fish) and decreased PK activity in the liver, indicating an activation of glycogenolysis and an inhibition of glycolysis and/or stimulation of gluconeogenesis, respectively. The catecholamine effects on GPase and PK were blocked completely by propranolol, identifying β-receptors as the mediators of the trout hepatic catecholamine response. Blockade of β-adrenoreceptors caused substantial reductions of percent GPase a and total GPase activities, while PK activity was markedly stimulated. Thus, it is apparent that the maintenance of blood glucose levels during hypoxia was due to adrenergic activation of GPase and inhibition of PK. In turn, these enzymatic changes accelerated hepatic glycogenolysis and gluconeogenesis while inhibiting glycolysis. The unusually high activity of PK (assayed at low PEP concentrations) and low activity of GPase during hypoxia after β-adrenoreceptor blockade suggest that internal hypoxia is specifically affecting the phosphorylation state, that is dephosphorylating these two enzymes. Thus, it would appear that the mobilization of catecholamines during hypoxia serves to offset the deleterious effects of hypoxemia on PK and GPase.

The hyperglycemic effect of catecholamines and involvement of β-adrenoreceptors have been demonstrated in vitro in trout (Mommsen et al. 1988) and in catfish (Brighenti et al. 1987c; Janssens & Lowrey, 1987). The present study is the first to show β-adrenoreceptor-mediated catecholamine effects on glucose metabolism in the intact trout with administration of physiologically relevant epinephrine or norepinephrine levels. In addition, we have shown that liver percent GPase a is twofold greater in epinephrine-treated fish than in propranolol-or propranolol + epinephrine-treated fish, whereas previous hepatocyte studies have failed to demonstrate a marked change (Ottolenghi et al. 1985, 1986; Mommsen et al. 1988). Epinephrine inhibits PK activity in trout hepatocytes (Mommsen et al. 1988), although the magnitude of the effect was considerably less than we report in the present study. Mommsen and co-workers found that the 0·1/5·0 PEP ratio decreased by 20%, compared to our results in which the 0·2/5·0PEP ratio was reduced by 60%. As discussed above, the hepatocyte isolation procedure may have altered the phosphorylation state of PK, resulting in a reduced sensitivity to catecholamines. Norepinephrine mimicked the effects for epinephrine but was less potent (Tables 4, 5). Similar results were obtained by Brighenti et al. (1987a,b) in catfish hepatocytes. In trout hepatocytes, however, both hormones induced similar changes in glucose production and lactate gluconeogenesis (Mommsen et al. 1988).

In many animals, the immediate response to oxygen deprivation is the breakdown of liver glycogen and subsequent mobilization of glucose via the blood to other tissues for anaerobic utilization. In the longer term, as glycogen levels decline, liver gluconeogenesis must be activated to supply glucose required for anaerobic metabolism or, alternatively, overall metabolism may be depressed. Our results did not show a significant drop in liver or muscle glycogen stores following 30 min of hypoxia (Table 6A), but the increases in liver total GPase and percent GPase a activity (Table 7B) suggest that glycogen catabolism may have been enhanced. Liver glycogen levels are extremely variable in fish and without initial glycogen contents it is difficult to evaluate the glycogenolytic effect of hypoxia. In fish hepatocytes and perfused livers, glycogen level decreases in the presence of catecholamines (Janssens et al. 1983; Ottolenghi et al. 1985, 1986; Brighenti et al. 1987a; Janssens & Lowery, 1987). Dunn & Hochachka (1987) found that trout exposed to water with a of 4 kPa for 3 h showed no significant fall in liver glycogen. Thus, it may be that 30 min of moderate hypoxia is not of sufficient intensity and duration to cause a discernible depletion of glycogen stores, despite elevated catecholamine levels.

Catecholamine release during hypoxia is dependent on the intensity and duration of the hypoxic stress (Tetens & Christensen, 1987; Boutilier et al. 1988). Plasma epinephrine levels in trout exposed to hypoxia (water for 30 min, Fig. 2A) were two-to 10-fold greater than that reported by Fievet et al. (1987) and Tetens & Christensen (1987) for trout in water of for 20–30 min. Norepinephrine levels (Fig. 2B) were lower than in the other two studies (Fievet et al. 1987; Tetens & Christensen, 1987) and, therefore, it may be more instructive to look at the total catecholamine levels ([epinephrine + norepinephrine]; Fig. 2C). Total plasma catecholamine concentration was 29nmoll⁻¹ during hypoxia (Fig. 2C), a value very similar to the total level in the Fievet et al. (1987) study and 2·5-fold higher than levels reported by Tetens & Christensen (1987). Control or initial catecholamine levels were approximately l nmoll⁻¹, which is representative of true resting values (Mazeaud & Mazeaud, 1985). Stress levels of epinephrine and norepinephrine may be as high as 200–400 nmol 1⁻¹ and, therefore, the measured levels of either catecholamine (60 nmol 1⁻¹) after injection in the present study were well within the physiological range, although higher than those achieved during hypoxia. An interesting observation is that propranolol alone stimulated catecholamine release and, during hypoxia, resulted in a higher plasma catecholamine concentration (Fig. 2). Tetens & Christensen (1987) also noted a catecholamine stimulation with propranolol, although they did not comment on its significance. The fact that pre-treatment with propranolol reduced blood and (Table 6C) suggests that fish were hypoxic before the experimental treatment (epinephrine, norepinephrine or hypoxia). This might plain the initial small rise in both epinephrine and norepinephrine in fish treated with propranolol. The marked increase in catecholamines with both propranolol and hypoxia, compared to hypoxia alone, is presumably related to the much greater hypoxic stress ⁠,Table 6C) in these fish.

Resting or baseline catecholamine levels (1 nmoll⁻¹) may be important in the regulation of glucose metabolism. When propranolol was administered to resting fish in series I, total GPase and GPase a activities were reduced by three-to fivefold compared to control values (Table 3B). (The same trends were observed in series II, but the differences were not statistically significant.) There are two possibilities; catecholamines tonically regulate glycogenolysis in the liver, or pre-mortem handling induced a catecholamine surge that resulted in an immediate β-adrenoreceptor response (Moon et al. 1988). In this study, the procedure of killing the fish, excising the liver and freezing the tissue in liquid N₂ took less than 10s. It is possible, however, that some catecholamines were released within this short time. To differentiate between tonic adrenergic control of GPase or pre-mortem stress, further studies on intact fish are necessary where the potential problem of stress-related handling is reduced or eliminated.

Binding of catecholamines to the β-adrenoreceptor activates adenylate cyclase and increases cellular cyclic AMP levels (see Levitzki, 1988). In the present study, cyclic AMP levels did increase slightly with catecholamines and hypoxia, although the changes were not significant. In contrast, cyclic AMP levels increased rapidly with both epinephrine or norepinephrine treatment in fish hepatocyte studies (Janssens & Lowrey, 1987; Brighenti et al. 1987a). There are several possible explanations for the discrepancy between in vivo and in vitro results. First, stimulation of cellular cyclic AMP levels is dose-dependent. A plasma catechol amine level in vivo, however, is probably not physiologically equivalent to the same in vitro dose, because of the more complex structure and circulation of the whole organ versus isolated hepatocytes. Second, the sampling time of our in vivo study may not have corresponded with the maximum surge of intracellular cyclic AMP levels. Based on the ubiquitous nature of cyclic AMP as the β-adrenoceptor intracellular messenger in higher vertebrates (Levitzki, 1988), teleosts (Janssens & Lowrey, 1987; Brighenti et al. 1987a) and amphibians (Janssens et al. 1986), it is probable that the low (but physiological) catecholamine dose used in our study, the inherent variability of in vivo data and transient surges have masked the role of cyclic AMP as the intracellular messenger in trout liver tissue.

We thank Dr Tom Mommsen for his helpful discussions during the initial stages of this study. This work was supported by NSERC of Canada operating and equipment grants to TWM and SFP. PAW was the recipient of a Killam Postdoctoral Fellowship.