Complete suppression of protein synthesis during anoxia with no post-anoxia protein synthesis debt in the red-eared slider turtle Trachemys scripta elegans

There is considerable interest in the mechanisms that allow some vertebrate species to survive long-term exposure to anoxia that would rapidly kill less tolerant species (e.g. Hochachka et al., 1997; Lutz and Nilsson, 1997a). The survival of these vertebrate facultative anaerobes under anoxic conditions is achieved by a large downregulation of ATP-consuming processes coupled to decreased ATP production via anaerobic pathways, allowing cellular ATP concentrations and cell function to be maintained (Hochachka, 1986). In contrast, in anoxia-intolerant animals, cellular ATP concentrations decrease rapidly during anoxia, leading to a loss of ion gradients, a rise in intracellular Ca²⁺ concentrations and the triggering of multiple cell-damaging processes leading ultimately to death (Hochachka, 1986).

As part of the general metabolic downregulation in anoxia-tolerant animals, one would expect to see a decrease in the rate of protein synthesis because this is an energetically expensive process accounting for a large proportion of cellular energy consumption (Muller et al., 1986; Land et al., 1993). The crucian carp, Carassius carassius (L.), an accomplished facultative anaerobe, has previously been shown to demonstrate a tissue-specific reduction in protein synthesis in the liver, heart, red muscle and white muscle, but not in the brain, when exposed to anoxia (Smith et al., 1996). Downregulation of protein synthesis in the liver and muscle of C. carassius during anoxia has been suggested to be an energy-conservation strategy (Smith et al., 1996). In vivo fractional protein synthesis rates have not previously been measured in any other anoxia-tolerant vertebrate.

Trachemys (=Pseudemys=Chrysemys) scripta elegans (Wied), the red-eared slider turtle, and Chrysemys picta bellii (Gray), the western painted turtle, are both highly tolerant of anoxia and capable of surviving up to 5 days of anoxia at 16–18°C and 4–5 months at 3°C (Robin et al., 1964; Ultsch and Jackson, 1982). The behavioural responses to anoxia shown by C. carassius, C. picta bellii and T. scripta elegans are very different: the carp retains some degree of activity, albeit reduced, while the turtles display a comatose-like state (Nilsson et al., 1993; Lutz and Nilsson, 1997b), suggesting that differences in physiological responses may also exist. Previous work has shown that hepatocytes from C. picta bellii exhibit a 92 % reduction in the rate of protein synthesis after 12 h of anoxia (Land et al., 1993). However, in T. scripta elegans, in vivo incorporation rates of ³⁵S-radiolabelled methionine into brain, heart, liver and muscle proteins did not decrease during anoxia (Brooks and Storey, 1993). This is perhaps surprising considering the large energetic cost of protein synthesis previously demonstrated in turtle hepatocytes (Land et al., 1993) and the fact that this species shows an 80–85 % reduction in heat production under anoxia (Jackson, 1968). The continual synthesis of new proteins and turnover of existing proteins in an animal is thought to be important for several reasons, including metabolic regulation and adaptation, mobilisation of amino acids and the elimination of non-functional or damaged polypeptides (Hawkins, 1991). If essential protein synthesis is suppressed during exposure to anoxia, a ‘protein synthesis debt’ may accumulate which could require repayment during the recovery period. The aims of the present work were to investigate (i) whether an anoxia-induced reduction in the rate of protein synthesis is partly responsible for the reduced energy demand and (ii) whether, during recovery, increases in protein synthesis rates above normoxic values occur to repay any ‘protein synthesis debt’ (Garlick et al., 1980). The flooding-dose methodology was used to measure in vivo tissue protein synthesis rates in groups of T. scripta elegans after 1, 3 or 6 h of normoxia or anoxia and after 0.5, 1 or 3 h of recovery from anoxia. Previous work has shown that protein synthesis rates increased to 160 % of control values after 2 h of recovery from anoxia in turtle hepatocytes (Land et al., 1993). To ensure that the criteria necessary for successful use of the flooding-dose method to measure protein synthesis were met, the initial experiment was run as a time course of intracellular free-pool specific radioactivity stability and protein radiolabelling linearity.

Juvenile Trachemys scripta elegans (Wied) were obtained from Lemberger (Oshkosh, Wisconsin, USA) in October 1995 and June 1997. Experiment 1 was carried out in October 1995, when the effects of anoxia on protein synthesis were examined in animals with a mean body mass of 115.41±6.87 g (N=36). Experiment 2 was carried out in June 1997, when the effects of recovery from anoxia on protein synthesis were examined in animals with a mean body mass of 151.01±8.98 g (N=30) (means ± s.e.m.). All animals were maintained and tissue samples collected at Florida Atlantic University with the approval of the Institutional Animal Care and Usage Committee. The turtles were housed in open plastic tanks (six per tank) containing 1–2 cm of fresh water and exposed to a L:D cycle of 12 h:12 h using fluorescent tubes (Vita-Lite) at an air temperature of 24°C. Food [Reptile T.E.N. (Wardley) floating food sticks; 38 % crude protein, 4 % lipid, 3 % crude fibre] was provided on three mornings ad libitum during the first week after the animals arrived in the laboratory. The animals were then starved for 2 weeks prior to the protein synthesis measurements to ensure that all animals were in a similar nitrogen balance.

In vivo tissue protein synthesis rates were measured using a modification of the flooding-dose method (Garlick et al., 1980; Houlihan et al., 1986). The animals were weighed to the nearest 0.1 g after surface drying with tissue paper. Each animal received an intra-peritoneal flooding-dose injection of [³H]phenylalanine (1 ml 100 g^–1 body mass of 135 mmol l^–1l-[2,6-³H]phenylalanine at 3.6 MBq ml^–1 (=100 μCi ml^–1; Amersham International)) (Smith et al., 1996). Injections were administered into the peritoneum by inserting the needle just anterior to one of the hind legs. After injection, groups of three animals were placed in three sealed plastic containers connected in series and continuously flushed with a positive flow of air, for normoxia-exposed animals, or nitrogen, for anoxia-exposed animals. All protein synthesis measurements were carried out at 23°C. Animals were exposed to anoxia or normoxia for 1, 3 or 6 h before being quickly removed from the plastic containers to administer an intra-peritoneal terminal injection of pentobarbitol in ethanol (60 mg kg^–1). Terminal anaesthetic injections were administered just anterior to one of the forelimbs. After injection, the turtles were placed back into the plastic containers until they became completely relaxed, usually within 2–3 min, and were then killed by decapitation. The heart, liver, intestine (pharynx to rectum), brain, head retractor muscle and lungs were dissected from the animals on ice, weighed to the nearest milligram and frozen in liquid nitrogen prior to storage at –70°C. The total time taken for dissection of each animal was approximately 10 min, and this time was not included in the protein synthesis calculations because, although some protein synthesis will occur during this period, rates are likely to be considerably lower than in the intact animal.

Protein synthesis was measured in six animals exposed to 3 h of normoxia and six animals exposed to 3 h of anoxia, as described for experiment 1. A further 18 animals were exposed to 3 h of anoxia before being removed from the anoxia chambers and returned to normoxia. Of these, 12 animals were injected with [³H]phenylalanine upon removal from the anoxia chambers. Six of these 12 animals were killed after 0.5 h, and their tissues were collected; the other six were killed after 1 h, and their tissues were collected. The remaining six animals were allowed to recover from the anoxia exposure for 2 h before they were injected with [³H]phenylalanine; after a further 1 h, these animals were killed and their tissues collected. The protocols used were identical to those in experiment 1. All protein synthesis measurements were carried out at 23°C.

Twelve of the animals used to measure protein synthesis were further dissected after tissue sample collection to remove all the remaining tissue from the carapace and plastron. The combined mass of the plastron and carapace was measured together with the combined mass of all the remaining tissue, which included the limbs, head, skin and musculature.

Sub-samples of the sampled tissues were weighed and homogenised (Tissue Tearor, Biospec Products Inc., USA) in 1.4 ml of 0.2 mol l^–1 perchloric acid (PCA) before centrifugation (3300 g, 10 min, 4°C; Eppendorf Minifuge, fixed rotor) to allow separation of the supernatant, which contained the intracellular free pool, from the precipitated protein, RNA and DNA (Houlihan et al., 1995). The methods used for the following tissue analysis have been described previously (Houlihan et al., 1995). Briefly, NaOH-soluble protein in the pellet was measured (Lowry et al., 1951) using bovine serum albumin as the standard. RNA was measured by comparing the sample concentrations with known RNA (Type IV, calf liver, Sigma) standard concentrations determined spectrophotomically (at 665 nm). The protein pellet was subsequently washed twice in 2 ml of 0.2 mol l^–1 PCA before being hydrolysed in 6 mol l^–1 HCl for 18 h.

Phenylalanine concentrations in the intracellular free pool, the hydrolysed protein pellet and the injection solution were measured using a fluorometric assay after the enzymatic conversion of the phenylalanine to β-phenylethylamine (PEA) (Houlihan et al., 1995). Known phenylalanine standards were also enzymatically converted to PEA to assess the conversion efficiency. The specific radioactivities of the intracellular free pools, protein pellets and injection solution were measured by scintillation counting (³H counting efficiency 45 %; Hionic Fluor scintillation fluid, Packard 1600TR, Liquid Scintillation Analyzer). Intracellular free pool, protein and injection-solution radioactivities were expressed as disintegrations per minute per nmole phenylalanine (disints min^–1 nmol^–1). Tissue fractional rates of protein synthesis were calculated using the following equation (Garlick et al., 1980):

where k_s is the percentage of protein mass synthesised per day, S_b is the specific radioactivity of protein-incorporated radiolabel (disints min^–1 nmol^–1 phenylalanine), S_a is the specific radioactivity of the intracellular free-pool (disints min^–1 nmol^–1 phenylalanine), t is time (min) and 1440 is the number of minutes in a day.

The absolute rates of protein synthesis (A_s) were calculated using the following equation:

where A_s is expressed as mg protein synthesised organ^–1 day^–1, protein concentration as mg g^–1 fresh mass and organ mass as g. After the sampled organs had been removed from the turtles, the remaining tissue, excluding the shell and plastron, i.e. the non-shell tissue, was considered as consisting primarily of muscle. Absolute protein synthesis rates for the non-shell tissue were therefore calculated using k_s and protein concentrations previously calculated for the head retractor muscle.

The RNA-to-protein ratio was expressed as μg RNA mg^–1 protein.

All data are expressed as means ±1 standard error of the mean (s.e.m.). Intracellular free-pool specific radioactivities, protein synthesis rates and RNA-to-protein ratios were compared within a single tissue using analysis of variance (ANOVA) to compare the effects of treatment and time. Pooled tissue intracellular free-pool specific radioactivities and injection-solution radioactivity were compared using single factorial ANOVA. In experiment 1, the linearity of radiolabel incorporation was tested by fitting (SigmaPlot 2001, Version 7.0, SPSS Inc., 233 South Wacker Drive, 11th Floor, Chicago, IL 60606-6307, USA) linear, second- and third-order regression equations and comparing the ‘goodness of fit’ by statistical comparison of residual mean squares (Sokal and Rohlf, 1995). If no significant relationship was found for a treatment between time and radiolabelling of the tissue, a two-factorial ANOVA was used to examine whether significant differences existed between time points for both treatments within a tissue. Tukey’s family error rate pairwise comparison test was used to distinguish between significantly different treatments post ANOVA.

Successful application of the flooding-dose methodology requires that the intracellular free-pool specific radioactivities are elevated and stable over the course of the protein synthesis measurement and that the rate of protein radiolabelling is linear. Each of these criteria is considered in turn for experiments 1 and 2.

Intracellular free-pool specific radioactivities increased rapidly after the flooding-dose injection in all tissues of both normoxia- and anoxia-exposed animals (Fig. 1). For all the tissues examined, there were no significant differences in phenylalanine free-pool specific radioactivities with time after injection or treatment. All the tissues (mean specific radioactivities of the pooled intracellular free-pools from the normoxic and anoxic 1, 3 and 6 h tissues were as follows: intestine 1506.2±53.8 disints min^–1 nmol^–1 phenylalanine, N=36; heart 1171.3±75.7 disints min^–1 nmol^–1, N=36; liver 1615.7± 39.6 disints min^–1 nmol^–1, N=36; brain 996.7±87.2 disints min^–1 nmol^–1, N=36; muscle 919.4±33.2 disints min^–1 nmol^–1, N=33; and lung 1152.7±79.0 disints min^–1 nmol^–1, N=36;) had significantly lower specific radioactivities than the injection solution (mean 2382.5±56.5 disints min^–1 nmol^–1 phenylalanine, N=23). Significant differences existed among the phenylalanine free-pool specific radioactivities of the tissues, with intestine=liver>brain=heart=muscle=lung. The mean free-pool phenylalanine concentrations increased 4.5-fold after injection of the flooding dose.

In the intestine, muscle and lung, there were no significant differences between the free-pool specific radioactivities for any of the treatments. In the brain, heart and liver, there were significant differences between treatments (Table 1). In the brain and heart, intracellular free-pool specific radioactivities were significantly lower at 0.5 and 1 h than at 3 h during recovery, while in the brain the free-pool specific radioactivity at 0.5 h was also lower than the normoxia control (Table 1). The free-pool specific radioactivity of the liver recovery group at 1 h was also significantly lower than that of the 3 h recovery group. However, the significant differences between the intracellular free-pool specific radioactivities of the aforementioned treatments will have no effect on subsequent calculations because the protein synthesis rates were calculated relative to the intracellular free-pool specific radioactivity. There were significant differences among the mean free- pool specific radioactivities of the tissues, with intestine=liver=muscle=lung>brain=heart. Free-pool phenylalanine concentrations showed a mean increase of 3.6-fold after the flooding-dose injection.

Regression analysis revealed a significant linear increase in protein radiolabelling with time in normoxia in all tissues (Fig. 2). No significant improvement on the residual mean squares of the linear model was made by fitting second- or third-order regressions (Sokal and Rohlf, 1995). In all cases, regression intercepts were not significantly different from zero. However, during anoxia, there was no significant increase in radiolabel incorporation after the first hour (Fig. 2). Single factorial ANOVAs revealed no significant difference between anoxic protein-incorporated specific radioactivity at 1, 3 or 6 h in any of the tissues. Comparisons of anoxic and normoxic protein-incorporated specific radioactivities at 1 h with Student’s t-test or Wilcoxon’s signed rank test revealed no significant differences.

There was no significant difference in tissue protein synthesis rates calculated at 1, 3 or 6 h in any of the individual tissues from normoxia-exposed animals. Therefore, within each tissue, the normoxic protein synthesis rates for the three time points were pooled to provide a single normoxic protein synthesis rate. Liver, brain and intestine protein synthesis rates were significantly higher than those in muscle (Table 2). After 1 h of anoxia, protein synthesis rates were not significantly different from those of control normoxic animals. However, after exposure to 3 or 6 h of anoxia, there was no significant increase in protein-incorporated radioactivity, suggesting that protein synthesis rates approached zero in all tissues.

Turtles exposed to 3 h of anoxia showed a large reduction in radiolabel incorporation in all tissues; protein synthesis rates cannot be calculated for these data because radiolabel incorporation is likely to have ceased after approximately 1 h (see above). Although in no tissues were recovery protein synthesis rates significantly different from those of control animals, several non-significant trends in protein synthesis were apparent during recovery (Fig. 3). In the intestine and liver, protein synthesis rates during recovery showed a definite increasing trend and might have exceeded control values after a longer recovery period than 3 h. In the brain and heart recovery group at 0.5 h, protein synthesis rates were higher than control values, albeit not significantly, before decreasing rapidly by 1 h of recovery. In contrast, in the muscle and lungs, recovery protein synthesis rates tended to be higher than those of controls at 0.5 h, decreased at 1.0 h and then increased again at 3 h. In the intestine and liver, there were significant differences between the protein synthesis rates measured after 1 and 3 h recovery from anoxia.

The combined mass of all the sampled tissues and the non-shell tissue account for approximately 43 % of the total mass of the animal; the remaining 57 % of the body mass consisting of the shell, plastron and body fluids lost during the dissection (Table 3).

The intestine and liver accounted for a large proportion of the daily protein synthesis in normoxic conditions (Table 3). Although the remaining non-shell tissue accounted for approximately 30 % of the total mass of the animal, the low estimated fractional protein synthesis rate meant that the absolute amount of protein synthesised was small. The contribution of the brain, heart and lung to the total amount of protein synthesised by the animal per day was very low because of the small protein masses of these organs.

The only tissue that showed a significant change in the RNA-to-protein ratio over the course of the experiment was the heart, in which the anoxic 3 h (8.09±0.97 μg mg^–1, N=6) value was significantly lower than the normoxic 1 h value (14.97±1.72 μg mg^–1, N=6) and the pooled normoxic heart RNA-to-protein ratio value. The pooled mean RNA-to-protein ratio in the other tissues ranged between 7.50 and 37.88 μg mg^–1 (Table 4).

There were no significant differences in the RNA-to-protein ratios of any tissue for any of the treatments.

Measurement of protein synthesis using the flooding-dose technique requires that several criteria are met (Houlihan et al., 1995). The first is that intracellular free-pool specific radioactivities rise rapidly after injection and remain elevated and stable over the course of the protein synthesis measurement period. In experiment 1, the tissue intracellular free-pool specific radioactivities were elevated and stable with no significant differences in any tissue with either treatment or time (Fig. 1). The elevation of the tissue phenylalanine concentrations to 4.5 times the baseline phenylalanine concentrations suggests that the tissues are flooded with phenylalanine.

The second criterion is that there is linear incorporation of the radiolabelled amino acid into the protein pool after injection. This criterion is met in the normoxic tissues, which all showed a significant linear increase in bound radioactivity with time (Fig. 2). The absence of a significant difference in the intercepts of any of the regression equations from zero suggests that protein radiolabelling started rapidly after the flooding-dose injection. However, in the anoxia-exposed animals, protein synthesis rates cannot be calculated after 1 h because there was no significant increase in protein-incorporated radioactivity, suggesting that protein synthesis either ceased or decreased below levels measurable using the flooding-dose technique. The criteria of the flooding-dose technique were therefore fully met by the experiment 1 time course study. A further time course study was not carried out in experiment 2 because the flooding-dose protocol was identical to that in experiment 1.

In vivo fractional protein synthesis rates during anoxia-induced metabolic downregulation have not previously been reported in turtles, although several authors have reported in vivo radiolabelled amino acid incorporation rates, in vitro hepatocyte protein synthesis rates or isolated heart protein synthesis rates (Brooks and Storey, 1993; Land et al., 1993; Bailey and Driedzic, 1995, 1996, 1997). In the present work, protein synthesis rates were lower than, but not significantly different from, those of control animals after 1 h of anoxia. The rates decreased to below measurable levels after 1–3 h of anoxia. It is likely that a large proportion of the protein synthesis measured over the first hour of anoxia was sustained by existing tissue, plasma and lung oxygen stores because turtle internal oxygen stores have been shown to require 30–60 min of exposure to anoxia for depletion at 24°C (Caligiuri et al., 1981). The only previous study of in vivo radiolabel incorporation in turtles (Brooks and Storey, 1993) used a single injection of a small amount of methionine. That study showed no significant difference in radiolabelling rates of proteins in normoxia- or anoxia-exposed turtles and suggested that protein synthesis may not be downregulated during anoxia (Brooks and Storey, 1993). However, in the same study (Brooks and Storey, 1993), incorporation of radiolabel into proteins ceased in all animals, even in normoxic control animals, after 5 h, suggesting that the intracellular free pools, which were not measured, were not stable and therefore that these results should be treated with caution.

After 1 h of anoxia, there was no significant decrease in the protein synthesis rate of isolated T. scripta elegans hearts (Bailey and Driedzic, 1995), although subsequent work by the same authors showed a decrease in both ventricular and mitochondrial protein synthesis rates after 2 or 3 h of anoxia (Bailey and Driedzic, 1996). These results are similar to those of the present study in that there was no significant decrease in protein synthesis rates after 1 h of anoxia, but a significant decrease occurred after longer exposures. Caution needs to be exercised in interpreting radiolabelling of protein during anoxic exposures when using the flooding-dose methodology. If, during the initial stages of anoxia, the rate of protein radiolabelling remains high in the continuing presence of oxygen, but at later time points, when oxygen is depleted, protein-incorporated radioactivity does not increase significantly, protein synthesis rates should not be calculated for the later time points. If no increase in protein-incorporated radioactivity occurs with time, then calculation of protein synthesis rates will result in the protein-phenylalanine specific radioactivity for the early time points being divided by increasingly larger time intervals. This would produce a gradual decrease in protein synthesis rates with continuing anoxia when, in fact, no detectable protein synthesis has occurred after the first time point. A significant linear incorporation of radiolabelled protein over time under anoxia would need to be demonstrated using regression analysis for maintenance of protein synthesis in anoxia to be convincingly demonstrated (Houlihan et al., 1995).

Isolated hepatocytes from C. picta bellii exhibited a 92 % reduction in protein synthesis rates after 12 h of anoxia (Land et al., 1993), which is indistinguishable from the reduction in protein synthesis rates obtained using the cytosolic protein synthesis inhibitor cycloheximide. The same hepatocyte model has been used to show that, although the vast majority of protein synthesis ceases under anoxia in turtles, a small number of specific proteins are still expressed (Land and Hochachka, 1995).

In C. carassius, protein synthesis rates were significantly reduced in the heart and liver after 48 h or 1 week of anoxic exposure and after 48 h in red and white muscle, whereas protein synthesis rates in the brain were not significantly reduced during even 1 week of anoxia (Smith et al., 1996). The behavioural responses of C. carassius and T. scripta elegans to anoxia are quite different: the crucian carp remains relatively active under anoxia, reducing its locomotor activity by only 50 % (Nilsson et al., 1993), whereas the turtle becomes comatose and unresponsive to stimuli (Lutz and Nilsson, 1997b). The crucian carp may need to sustain brain protein synthesis to remain active under anoxia, while inactivity in T. scripta elegans may allow further small energy savings by ‘switching off’ the bulk of brain protein synthesis. It has also been shown that while T. scripta elegans uses ‘channel arrest’ to minimise brain energy use under anoxia (Pek and Lutz, 1997) the carp does not, possibly because it needs to retain a higher level of brain function to remain active (Johansson and Nilsson, 1995).

The present work demonstrated no significant ‘overshoot’ in tissue protein synthesis rates above values for normoxic controls on recovery from 3 h of exposure to anoxia, although several non-significant trends in protein synthesis rates during recovery were apparent. Protein synthesis rates of isolated perfused hearts from T. scripta elegans exposed to 2 h of anoxia prior to recovery for 1 h also showed no increase in protein synthesis above normoxic control levels (Bailey and Driedzic, 1997). However, in isolated C. picta bellii hepatocytes, protein synthesis rates increased to 160 % of normoxic control levels 1 h after recovery from 12 h of anoxia, but were not significantly different from those of control animals 2 h after exposure to anoxia (Land et al., 1993). The protein synthesis ‘overshoot’ in turtle hepatocytes may be due to the longer exposure to anoxia or to differences between the in vivo and in vitro protocols (Land et al., 1993). In the heart, liver and brain of C. carassius, there was no significant difference between normoxic protein synthesis rates and protein synthesis rates measured after 24 h of recovery from 48 h of anoxia exposure (Smith et al., 1996). It would appear, therefore, that at least for short periods of anoxia there is no ‘protein synthesis debt’ to replace when aerobic metabolism restarts.

In the present work, there was a significant decrease in the heart RNA-to-protein ratio after 6 h of anoxic exposure. Although the RNA-to-protein ratio tended to decrease in the liver and brain under anoxia, these reductions were not significant. Other workers have also reported a decrease in RNA-to-protein ratios in heart, ventricle and atria after anoxia exposures ranging from 1 to 3 h (Bailey and Driedzic, 1995, 1996). However, in turtle hepatocytes, there was no significant change in RNA-to-protein ratio after 12 h of anoxia (Land et al., 1993). After 16 h of anoxia or recovery, there was no change in total RNA concentrations of the liver, kidney, heart and red and white skeletal muscle of T. scripta elegans (Douglas et al., 1994), although a significant increase (30 %) in mRNA [poly(A)⁺ RNA] concentration in white muscle under anoxia has been reported (Douglas et al., 1994). Any change in mRNA concentration is unlikely to be detectable in the current study because mRNA represents only 1.3–2.4 % of total RNA (Douglas et al., 1994). However, when C. carassius was exposed to anoxia for periods of 48 h and 1 week, the RNA-to-protein ratio in the brain was significantly lower than normoxic control levels at 48 h and 1 week, the RNA-to-protein ratio in the heart was significantly lower at 48 h but not at 1 week and the RNA-to-protein ratio in the liver was significantly higher at 48 h but significantly lower after 1 week (Smith et al., 1996). These results suggest that, during exposure to longer periods of anoxia, changes in the RNA-to-protein ratio may occur. Since the cost of RNA synthesis has been shown to account for approximately 10 % of total ATP consumption in some cells, one would expect to see a downregulation of RNA synthesis under anoxia concurrent with the reduction in protein synthesis as an energy-saving strategy (Muller et al., 1986).

In conclusion, on exposure to anoxia, T. scripta elegans shows a gradual decrease in protein synthesis over the first hour followed by an almost complete cessation of protein synthesis after 1–6 h of anoxia. The decrease in protein synthesis rates during short-term anoxia does not appear to be controlled by tissue RNA concentrations, except possibly in the heart.

The authors are grateful to Jennica Lowell and Nicole Blackson, who helped with animal husbandry, tissue harvesting and sample preparation at Florida Atlantic University, and Professor Lloyd Peck for useful comments on the manuscript. This research was carried out when K. Fraser was in receipt of a Natural Environment Research Council PhD studentship.