Oxygen, while being an obligate fuel for aerobic life, has been shown to be toxic through its deleterious reactive species, which can cause oxidative stress and lead ultimately to cell and organism death. In marine organisms, reactive oxygen species (ROS), such as the superoxide anion and hydrogen peroxide, are generated within respiring cells and tissues and also by photochemical processes in sea water. Considering both the reduced metabolic rate of nektonic organisms thriving in the deep sea and the physico-chemical conditions of this dark, poorly oxygenated environment, the meso- and bathypelagic waters of the oceans might be considered as refuges against oxidative dangers. This hypothesis prompted us to investigate the activities of the three essential enzymes (superoxide dismutase, SOD; catalase, CAT; glutathione peroxidase, GPX) constitutive of the antioxidative arsenal of cells in the tissues of 16 species of meso- and bathypelagic fishes occurring between the surface and a depth of 1300 m. While enzymatic activities were detected in all tissues from all species, the levels of SOD and GPX decreased in parallel with the exponential reduction in the metabolic activity as estimated by citrate synthase activity. In contrast, CAT was affected neither by the metabolic activity nor by the depth of occurrence of the fishes. High levels of metabolic and antioxidative enzymes were detected in the light organs of bioluminescent species. The adjustment of the activity of SOD and GPX to the decreased metabolic activity associated with deep-sea living suggests that these antioxidative defense mechanisms are used primarily against metabolically produced ROS, whereas the maintenance of CAT activity throughout all depths could be indicative of another role. The possible reasons for the occurrence of such a reduced antioxidative arsenal in deep-sea species are discussed.

It has long been known that oxygen is not distributed uniformly in the oceans (Sewell and Fage, 1948). Most regions of the world’s oceans are characterized by high oxygen concentrations in the euphotic zone and by a decrease to lower oxygen levels in the deep sea. In some areas, such as the Pacific Ocean, oxygen levels reach a minimum value (the oxygen minimum layer, OML) corresponding to only a few per cent of that near the surface, followed by an increase in oxygen levels at greater depths (Wyrtki, 1962). Since oxygen is essential for aerobic life, problems associated with the existence of such a gradient focused on the adaptations of organisms able to survive at the low oxygen concentrations occurring near the OML (Childress and Siebel, 1998). However, in addition to its importance as an electron acceptor in cellular energetic processes, oxygen is also a problem for living organisms because it can be toxic. Indeed, its electronic structure favors univalent reductions that lead to the formation of highly reactive intermediates, such as the superoxide anion (O2•−), hydrogen peroxide (H2O2) and the hydroxyl radical (OH).

These intermediates, called reactive oxygen species (ROS), can oxidize most cellular constituents, such as DNA, proteins and lipids, and markedly affect the physiology of the cell, leading to the initiation of cancer or cellular death (Cadenas, 1995). Most major pathologies in humans and animals are known to involve ROS at an early stage of the disease (Chen et al., 1995), and the involvement of ROS in the aging process is well documented in many species (Pacifici and Davies, 1991; Harman, 1994). Since oxygen was released into the atmosphere by photosynthetic processes, its toxicity has posed a major threat to life (Hassan and Schiavone, 1991), and all extant organisms are endowed with a complex arsenal of antioxidative defense mechanisms. Besides some small diffusible molecules, such as vitamins C and E, which are able to quench ROS in cells and tissues, specialized enzymes are aimed at neutralizing ROS. The three most important enzymes are the superoxide dismutases (SODs) that converts O2•− into H2O2, which can in turn be detoxified into water and oxygen by either catalase (CAT) or peroxidases. A major peroxidase is glutathione peroxidase (GPX), which utilizes glutathione (GSH) as an electron donor (Ahmad, 1995).

Accordingly, in addition to aerobic life in a hypoxic environment such as the OML being a real physiological challenge, it can also be seen as beneficial because it relieves organisms from a major danger. From that point of view, the deep sea appears to be a somewhat protected place to live. Marine organisms face two threatening sources of ROS: one is exogenous and is linked to the presence of ROS in sea water; the other is endogenous and is associated with metabolic processes in the organisms. The mitochondrial respiratory chain, which continuously generates O2•− in every cell at a rate of 1–5 % of the total oxygen consumed, is the main producer of ROS (Boveris and Cadenas, 1982). Work by Petasne and Zika (1987) and Van Baalen and Marler (1966) has revealed that superoxide anions and hydrogen peroxide are formed in sea water exposed to sunlight. Measurements in the Gulf of Mexico indicated maximal levels of 2×10−8 mol l−1 and 10−7 to 2×10−7 mol l−1 for O2•− and H2O2, respectively (Zika et al., 1985). Because these ROS are produced by photochemical reactions, they are restricted to near the euphotic zone and their concentration decreases rapidly with increasing depth. As a consequence, levels of ROS were undetectable at depths below 100 m.

From these data, one could expect deep-sea organisms to face a reduced threat from ROS. First, the absence of exogenously produced ROS eliminates dangers associated with them. Second, the general decrease in the metabolic rate with increasing depth of occurrence of fishes, crustaceans and cephalopods (Childress, 1995; Childress and Siebel, 1998) should be accompanied by a decrease in the amounts of ROS produced in their tissues. Thus, it is possible that these reduced dangers linked to oxygen toxicity could have led deep-sea organisms to reduce their biochemical defenses against oxidative damage. To test this hypothesis, we investigated the activity of the three most important antioxidative enzymes in a variety of tissues of meso- and bathypelagic fishes occurring between a depth of 1300 m and the surface.

Meso- and bathypelagic fishes were collected in the San Clemente Basin (32°N, 119°W; 2000 m maximal depth), northeastern Pacific, during a 15-day cruise aboard the R/V Point Sur. Animals were captured with an opening/closing Mother Tucker trawl (10 m2 mouth), equipped with a 30 l thermally insulated cod end for reducing mechanical damage and heat shock to the animals during ascent (Childress et al., 1978). Once on board, the animals were immediately transferred to seawater tanks (5 °C), where they were identified. Selected individuals were immediately frozen in liquid nitrogen and then stored in a freezer at −20 °C. Others were rapidly dissected on ice, and the resulting tissue samples were also frozen in liquid nitrogen before storage at −20 °C. After the cruise, samples were stored at −80 °C until processed. Taking into account the number of individuals captured by species and their minimum depth of occurrence (MDO), 16 species were selected to cover as far as possible the range between the surface and a depth of 1300 m (Table 1). Minimum depth of occurrence is defined as the minimum depth below which 90 % of the population of the species concerned can be found (Childress and Nygaard, 1973).

Table 1.

Minimum depth of occurrence (MDO) and number of individuals of fish species used in this study

Minimum depth of occurrence (MDO) and number of individuals of fish species used in this study
Minimum depth of occurrence (MDO) and number of individuals of fish species used in this study

Sample preparation

Eleven tissues and organs were selected: eyes, brain, gills, heart, liver, stomach, kidney, white muscle, red muscle, skin and skin with photophores (Aksnes and Njaa, 1981; Filho et al., 1993). After the dissection, which was carried out on ice, similar tissues of each species were pooled, weighed and homogenized for 20 s at 9500 revsmin−1 with an electronic homogenizer (Polytron) in 5 volumes of phosphate-buffered saline (PBS) containing 1 % Triton X-100. Supernatants were obtained by a 30 min centrifugation at 4000 revsmin−1 (2880 g) at 4 °C (Minifuge GL, Heraeus). Part of the supernatant was stored at −80 °C, while the rest was dialysed (Medicell International Ltd; 12–14 kDa) overnight at 4 °C against the PBS-Triton buffer. Some samples were assayed directly for CAT while others were stored at −80 °C.

Assay for catalase (EC 1.11.1.6)

The chemiluminescent method of Maral et al. (1977) was adapted for 96-well microplates to process numerous samples and replicates simultaneously. The assay for catalase (CAT) was performed on fresh samples immediately after dialysis. The measurements were carried out at 25 °C on a PC-controlled microplate luminometer (Berthold) equipped with two 50 μl injectors. A first injection of 10−6 mol l−1 H2O2 solution initiated its reduction by the CAT contained in the diluted sample (50 μl), which was added to 100 μl of 100 mmol l−1 phosphate buffer (pH 7.8) containing 0.6 mmol l−1 EDTA. After a 30 min incubation, the injection of 50 μl of a solution containing 20 mmol l−1 luminol and 11.6 units ml−1 horseradish peroxidase produces an emission of light whose intensity was proportional to the H2O2 remaining in the mixture. CAT activity in the sample was estimated with purified bovine liver CAT solubilized in the homogenization buffer as standard and is expressed in international units.

Assay for superoxide dismutase (EC 1.15.1.1)

A new microplate assay method was developed to detect the very low activities in many samples. The method is based on the chemiluminescence of an imidazolopyrazinone molecule (CLZm) when this reacts with O2•− (Nakano, 1990). Exploiting this chemiluminescent property, total superoxide dismutase (SOD) activity was estimated by the inhibition of light production in conditions where the enzymatically generated O2•− is the rate-limiting factor of the luminescence. In these circumstances, SOD competes with CLZm for O2•−, and the luminescence will become weaker as the activity of SOD in the sample increases. The injection of 50 μl of hypoxanthine solution (34 mg ml−1) into a mixture of 25 μl of diluted sample, 165 μl of 100 mmol l−1 phosphate buffer (pH 7.8) containing 0.6 mmol l−1 EDTA and 0.01 % bovine serum albumin (BSA), 5 μl of 0.8 μmol l−1 CLZm and 5 μl of xanthine oxidase (27.56 munits ml−1) initiates the emission of light. The assay was carried out at 25 °C, and SOD activity was calculated from a standard curve established with Cu,Zn-SOD purified from bovine erythrocytes in PBS-Triton buffer. The activity of SOD is expressed in international units.

Assay for glutathione peroxidase (EC 1.11.1.9)

Glutathione peroxidase (GPX) activity was measured according to the spectrophotometric method of Paglia and Valentine (1967). Diluted samples (65 μl) were added to 1 ml of 50 mmol l−1 Tris-HCl buffer (pH 7.6) containing mmol l−1 EDTA, 0.14 mmol l−1 NADPH, 1 mmol l−1 GSH and 1 unit ml−1 glutathione reductase (EC 1.6.4.2). The consumption of NADPH was monitored at 340 nm (25 °C) in a spectrophotometer (Beckman DU-8) after the addition of mmol l−1t-butyl hydroperoxide. GPX activity was estimated from a standard curve constructed with GPX purified from bovine erythrocytes solubilized in PBS-Triton buffer and is expressed in international units.

Metabolic activity

Metabolic activity was estimated from the activity of citrate synthase (CS; EC 4.1.3.7), a mitochondrial enzyme that is a good indicator of the metabolic activity of the tissues (Childress and Somero, 1979). Samples (10 μl) were added to 995 μl of 100 mmol l−1 Tris-HCl (pH 8.0) buffer containing 1 mmol l−1 5,5′-dithio-bis(2-nitrobenzoic acid) (DTNB) and 0.5 mmol l−1 acetyl-CoA. The consumption of acetyl-CoA was measured at 25 °C from the increase in absorbance at 420 nm after the addition of 5 μl of 5 mmol l−1 oxaloacetate (Shepherd and Garland, 1969). The activity of CS was estimated from a standard curve established with CS from chicken heart and is expressed in international units.

Protein measurements were performed according to the method of Lowry et al. (1951) with bovine serum albumin as standard. Oxygen partial pressures in the sea water at different depths of the area studied were obtained from Childress (1995).

Reagents

All reagents were purchased from Sigma (St Louis, USA), except for Triton X-100, GSH and glutathione reductase, which were obtained from Boehringer (Mannheim, Germany). Methyl-coelenterazine [CLZm; 3,7-dihydro-2-methyl-6-(p-hydroxyphenyl)-8-benzylimidazo(1,2-a]pyrazin-3-one)] was synthesized as described by Inoue et al. (1975).

Statistical analyses

The enzymatic activities are presented as means of quadruplicate measurements with the standard error of the mean and are expressed relative to both the protein content and to the wet mass of the tissue.

To compare metabolic activity levels among different fishes and tissues (Childress and Somero, 1979) with various protein and water contents (Childress and Nygaard, 1973), the data were presented as a function of wet mass because this standardization procedure seems to be the most appropriate to analyze these results as a function of depth. This standardization procedure indicates the metabolic activity per unit of living tissue. However, for the activities of antioxidative enzymes, standardization to protein content seems to be more appropriate because the antioxidative status of the cell is linked to cellular targets (lipids, proteins, DNA) and not to the size, mass or water content of the animal. This standardization procedure indicates priorities for the organism in the synthesis of antioxidative proteins within the context of their global protein synthesis. Accordingly, expressing activity with respect to protein content appears to be most appropriate for comparing enzyme activities.

Simple linear regressions were carried out on log-transformed data, as for previous studies on these fishes (Childress and Siebel, 1998). The figures are, however, semilogarithmic plots. Confidence intervals on the regression coefficients are at the 95 % level; P values are from F-tests.

Protein content and metabolic activity of fish tissues

Our measurements confirmed the previously reported exponential decrease in the protein content (Table 2) of the white muscle with increasing minimum depth of occurrence (Childress and Nygaard, 1973). Interestingly, similar relationships were found in the eyes, liver, stomach and skin (Table 2). The previously reported (Childress and Somero, 1979) exponential decrease in the metabolic activity in the white muscle, as revealed by CS activity, with increasing minimum depth of occurrence was also confirmed (Table 2) and was also demonstrated in the eyes (b=−0.64; Fig. 1) and the stomach (b=−0.62). Although CS activity was detected in all the tissue of all species, some activities tended to be very low when expressed as a function of protein content or wet mass. Among species, the level of CS in the liver of the deepest-living species, Eurypharynx pelicanoides (minimum depth of occurrence 1300 m), was only 2.5 % of that assayed in the myctophid Symbolophorus californiensis, a species reaching the surface. Compared with all other tissues, the skin associated with light organs, when present, showed a significantly higher CS activity (P<0.05). The size of the fish could be correlated neither with the minimum depth of occurrence nor with CS activity.

Table 2.

Stastically significant relationships between biochemical variables and the minimum depth of occurrence in tissues of meso- and bathypelagic fishes collected off California

Stastically significant relationships between biochemical variables and the minimum depth of occurrence in tissues of meso- and bathypelagic fishes collected off California
Stastically significant relationships between biochemical variables and the minimum depth of occurrence in tissues of meso- and bathypelagic fishes collected off California
Fig. 1.

Exponential relationship between the citrate synthase (CS) activity (munits g−1 wet mass) of the eyes and the minimum depth of occurrence (MDO) (m) of meso- and bathypelagic fishes: CS activity=16 841MDO−0.64. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Fig. 1.

Exponential relationship between the citrate synthase (CS) activity (munits g−1 wet mass) of the eyes and the minimum depth of occurrence (MDO) (m) of meso- and bathypelagic fishes: CS activity=16 841MDO−0.64. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Superoxide dismutase

Although SOD activity was detected in all tissues from all species, large differences were observed among tissues and species. The highest mean SOD activity was detected in the photophores (416.40 units mg−1 protein), while the lowest levels were found in the eyes (16.46 units mg−1 protein). Variations of up to three orders of magnitude were found in a single organ. For example, the SOD level in the liver of the deepest-living species, Eurypharynx pelicanoides (minimum depth of occurrence 1300 m), was 1.02 units mg−1 protein, but it was 175.72 units mg−1 protein in Argyropelecus affinis (minimum depth of occurrence 200 m). Although there appeared to be a trend for SOD activity (units g−1 wet mass) to decrease with increasing depth of occurrence, attempts to link SOD activity to either the minimum depth of occurrence or the ambient oxygen concentration did not yield significant relationships. In contrast, the activity of SOD was tightly correlated to CS activity in most tissues and organs (Fig. 2). The relationship is best described as an exponential increase in SOD activity (units mg−1 protein) with increasing CS activity (munits mg−1 protein) following the equation SOD activity =a(CS activity)b, with b ranging from a minimum 0.51 in the stomach to more than 1 in the muscles and gills (Table 2). Although high CS activity was always associated with high levels of SOD in the tissues, large variations in SOD activity were observed at low CS levels (<25 munits mg−1 protein), suggesting that other factors could be important in determining SOD activity when the metabolic rate is low. As an example, it can be seen that SOD levels in the liver of Scopelengys tristis and Parvilux ingens are higher than would be expected from their CS activity.

Fig. 2.

Exponential relationship between superoxide dismutase (SOD) (units mg−1 protein) and citrate synthase (CS) activities (munits mg−1 protein) in the liver of meso- and bathypelagic fishes: SOD activity=2.29(CS activity)0.70. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Fig. 2.

Exponential relationship between superoxide dismutase (SOD) (units mg−1 protein) and citrate synthase (CS) activities (munits mg−1 protein) in the liver of meso- and bathypelagic fishes: SOD activity=2.29(CS activity)0.70. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Glutathione peroxidase

The activity of GPX shows a very similar pattern to that of SOD, i.e. large variations among tissues and species and a dependence on the metabolic activity as measured by CS activity. The highest activities were again associated with the photophores (P<0.05 compared with all other tissues), with levels (9308.87 munits mg−1 protein) far higher than in the liver, which ranked second (541.62 munits mg−1 protein). Very low levels were measured in the eyes (68.18 munits mg−1 protein) and the red muscle (71.45 munits mg−1 protein). Variations in activity of three orders of magnitude were observed in most tissues. In the liver, the GPX activity measured in Eurypharynx pelicanoides (minimum depth of occurrence 1300 m) reached 39.3 munits mg−1 protein, while the level in Symbolophorus californiensis (minimum depth of occurrence 10 m) was 1522.76 munits mg−1 protein. The relationship between GPX activity (munits mg−1 protein) and CS activity (munits mg−1 protein) was very similar to that described between SOD and CS, with GPX activity=a(CS activity)b and with b ranging from a minimum of 0.57 in the stomach to 0.97 in the white muscle and brain (Fig. 3; Table 2). Here again, high CS activity was associated with high levels of the antioxidative enzyme while large variations in GPX activity were found at low CS levels (<25 munits mg−1 protein), suggesting that other factors also affect GPX activity when metabolic rate is low.

Fig. 3.

Exponential relationship between glutathione peroxidase (GPX) (munits mg−1 protein) and citrate synthase (CS) activities (munits mg−1 protein) in the liver of meso- and bathypelagic fishes: GPX activity=6.62(CS activity)0.92. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Fig. 3.

Exponential relationship between glutathione peroxidase (GPX) (munits mg−1 protein) and citrate synthase (CS) activities (munits mg−1 protein) in the liver of meso- and bathypelagic fishes: GPX activity=6.62(CS activity)0.92. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

The similar evolution of SOD with metabolic activity (units mg−1 protein) and GPX activity (munits mg−1 protein) is further illustrated by the existence of a significant relationship, GPX activity=a(SOD activity)b, linking the activities of the two antioxidative enzymes in most tissues (Table 2; Fig. 4). The value of the exponent b, which characterizes the relationships, is close to 1 in most tissues (eyes, brain, liver, stomach, kidney, white muscle and skin), indicating a linear relationship. In the gills, where b is only 0.43, it appears that the activity of SOD increases faster than that of GPX (Table 2).

Fig. 4.

Exponential relationship between glutathione peroxidase (GPX) (munits mg−1 protein) and superoxide dismutase (SOD) activities (units mg−1 protein) in the liver of meso- and bathypelagic fishes: GPX activity=7.69(SOD activity)0.95. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Fig. 4.

Exponential relationship between glutathione peroxidase (GPX) (munits mg−1 protein) and superoxide dismutase (SOD) activities (units mg−1 protein) in the liver of meso- and bathypelagic fishes: GPX activity=7.69(SOD activity)0.95. Each letter represents one species (see Table 1). Values are means ± S.E.M. (N=4).

Catalase

Contrary to what was observed for SOD and GPX, no relationship between CAT and CS activity could be found in any tissue whether data were expressed as a function of the wet mass or of the protein content. Although large variations occurred among tissues, the activity of CAT in a single tissue varied little among species. In most tissues, the highest activity of CAT was found in the liver, representing 34–99 % of the total CAT activity of the fish. The lowest activity was found in the white muscle, where its content varied from 0.01 to 0.64 % of the total CAT activity. It is interesting to note that skin containing photophores has a higher CAT activity than skin lacking these luminous organs; this relationship was independent of the minimum depth of occurrence of the fish.

The decline in metabolic activity of pelagic fishes and crustaceans with increasing minimum depth of occurrence has been reported in many studies (for a review, see Childress, 1995), and has also recently been observed in pelagic cephalopods (Siebel et al., 1997). Many variables could be implicated in such a decline: reduced residual light, temperature, available nutrients, the low-oxygen environment or increased pressure, among others. Childress (1995) reviewed these variables and refuted their possible implications in the evolution of metabolic rates. According to Childress (1995), the lower metabolic rates of deep-sea fishes, crustaceans and cephalopods are functionally adaptive for life in the deep sea and the oxygen minimum layer and are not specifically evolved adaptations to the low food availability or low O2 levels in these habitats. Several lines of evidence, including the observation that no decrease in metabolic activity occurs in the brain of deep-sea fishes although such a decrease can be measured in white muscles, have led to the hypothesis that a decrease in visual prey/predator interactions was the key factor in the evolution of lower metabolic rates at depth, through a reduction in the locomotor abilities needed for foraging. This view is consistent with observations on chaetognaths (Thuesen and Childress, 1993) and medusae (Thuesen and Childress, 1994), which show no decrease in metabolic activities with increasing depth of occurrence.

Our results confirm previous data indicating that the metabolic activity of the brain does not decrease with increasing depth of occurrence (Childress and Somero, 1979). Interestingly, the present work indicates that the decline in metabolic activity occurs not only in the muscles of fish but also in their eyes, a result that could support the above model. However, the reduction in ocular CS activity seems to take place at a faster rate (b=−0.64) than the reduction in the overall metabolic rate when data from all tissues are combined (b=−0.43), a result that might be related to the rapid decline in ambient light levels with increasing depth. Such a decline would certainly decrease the operational cost of the visual system. Also, the depth-related decline in CS activity in the stomach may indicate that the metabolic needs of tissues involved in the digestion of prey could have adjusted to those of the energy-consuming locomotory system. This result suggests that selection could have favored low-energy-consuming digestion processes acting over longer periods in a context of lower densities of prey with a lower nutritional value.

Whatever the exact reasons for these declines in metabolic activity in pelagic organisms could be, the present data demonstrate that these reduced metabolic requirements are associated with lower activities of two essential antioxidative enzymes, superoxide dismutase and glutathione peroxidase. The decline in the activities of both enzymes cannot be ascribed to some experimental artifact associated with an excessive assay temperature favoring enzymes from species thriving in warmer conditions, i.e. near the surface. Because of the dependency of the reaction between SOD and O2•− on diffusion rate, SOD is little affected by temperature. Issels et al. (1986) demonstrated that SOD activity increased by a factor of only 1.6 when the temperature was raised from 20 to 60 °C. In any case, experimental procedures limited the exposure of the enzymes to 25 °C to a maximum of 15 min. Controls indicated that the exposure of these enzymes extracted from the deepest-living species to 25 °C for 15 min did not affect their activity. Also, the possible interference of the assay temperature with the determination of CS activity seems very unlikely because the results are very similar to those reported in other studies in which this assay was carried out at a lower temperature (Childress and Somero, 1979; Childress, 1995).

If one considers that the reduced CS activity with increasing depth could be accomplished by reduced numbers of mitochondria rather than reduced enzymatic efficiency or reduced quantities of enzymes within mitochondria, the possibility arises that the reduced activities of SOD and GPX could simply reflect the attenuation of mitochondrial densities. However, although both antioxidant enzymes are partly localized within mitochondria, their reduced activities with decreasing CS activity cannot be explained by their intracellular distribution within cells. Indeed, mitochondrial Mn-based SOD activity of vertebrates accounts for only 10–20 % of the total SOD activity, while the remainder represents Cu,Zn-SOD located in the cytosol (McCord and Fridovich, 1988). In the case of GPX, several forms coexist within cells, but only 10 % of the total activity is located in the mitochondrial matrix (Meister, 1995). If one assumes that the reduction in CS activity with increasing depth corresponds exclusively to a decreasing mitochondrial density within cells, then this could account only for a maximum reduction of 10–20 % in both SOD and GPX activity. Our results indicate a far more drastic reduction in activity of both enzymes. To support this, differential assays of Mn-SOD and Cu,Zn-SOD were performed according to the method of Nakano (1990) on liver homogenates from the fishes described in this study. These revealed that Cu,Zn-SOD represents 75.91±5.74 % of the total SOD activity in the liver of surface-living fish, 92.13±9.39 % in fish living between 250 and 500 m, and 84.23±8.96 % (means ± S.E.M., N=4) in deeper living species. These data are consistent with results obtained for other marine fish. For example, Cu,Zn-SOD in the liver of saithe and mackerel represents 83 % and 93 % of the total SOD content, respectively (Aksnes and Njaa, 1981).

The correlation between the decline in both SOD and GPX activity and oxidative metabolism indicates that levels of both antioxidant enzymes were adjusted to the endogenous source of ROS. Since superoxide is known to be produced continuously in all aerobic cells during the transfer of electrons in mitochondria, a lower rate of oxygen consumption would be expected to reduce the rate of O2•− production and, thus, lower requirements for the corresponding detoxification systems. Similarly, the decreased production of superoxide together with a reduction in the lipid content should reduce the risks of formation of H2O2 and lipid peroxides, thus diminishing the requirement to maintain high GPX activities in the tissues.

The analysis of the possible role of exogenous sources of ROS and the influence of the oxygen content of the sea water will require further studies of populations thriving in different geographical areas at different depth and oxygen concentrations.

It is interesting to note that three of the deepest-living species studied, Scopelengys tristis (650 m), Parvilux ingens (700 m) and Bajacalifornia burragei (1000 m), while showing a very low CS activity, contained higher levels of SOD and GPX than could be expected from the relationships described (Figs 2, 3). This suggests that other sources of ROS might be present in these deep-living species. It is well known that many pollutants present in aquatic systems, such as organochlorines, exert toxic effects related to oxidative stress. In response to increasing levels of organic xenobiotics, contaminated animals show an induction of mono-oxygenases (Stohs, 1995), such as cytochrome-P450-associated ethoxyresorufin-O-deethylase (EROD), and higher activities of antioxidative enzymes (Rodriguez-Ariza et al., 1993; Bainy et al., 1996). Kramer et al. (1984) showed that the amount of unmetabolized 4,4′-dichlorophenyltrichloroethane (4,4′-DTT) and other pollutants in deep-sea benthic fishes is far higher than that found in surface-living fish. Some deep-sea fish appear to contain approximately 50 times more polychlorinated biphenyls (PCB) than surface-living fish. In addition, very high activities of EROD and other xenobiotic-metabolizing enzymes have been detected in deep-sea fish (Förlin et al., 1995). Thus, it is possible that the high antioxidative activities found in some pelagic deep-sea species could be linked to the oxidative stress due to pollutants contaminating the deep-sea biota.

Considering the above statements, it is very surprising that the activity of CAT, which catabolizes the production of oxygen and water from H2O2, remained unchanged while metabolic rates declined with increasing depth of occurrence. In epipelagic fishes, close positive correlations were found between the CAT and SOD activity of most tissues (Filho et al., 1993), and CAT activity would have been expected to decrease in the same manner as does SOD activity. Several hypotheses can be proposed for this unexpected observation.

First, it could indicate that the dangers associated with H2O2 do not decrease with increasing depth of occurrence and, while the endogenous sources undergo a rapid decline, the activity of CAT could be essentially devoted to the elimination of exogenous H2O2 diffusing into the fish tissue. It is remarkable that the activity of CAT does not vary with the depth, while that of GPX shows a rapid decline to very low levels, although both enzymes are involved in the detoxification of peroxides. Since GPX can decompose H2O2 and lipidic peroxides while CAT is devoted solely to the catabolism of H2O2, it is surprising that the reduced enzyme activity is that detoxifying the wider range of ROS. The possible reason for this apparent preference for CAT might be related to the limited resources available in the deep sea. Indeed, CAT requires neither cofactors nor energy for its activity, while GPX consumes glutathione, which is oxidized upon reaction with GPX and must then be recycled by a NADPH-consuming enzyme (glutathione reductase). In an environment characterized by a decline in the resources available, favoring the least energy-consuming enzyme could be an efficient strategy, in particular when the total amount of lipid targets is decreased. In a preliminary series of experiments, we assayed H2O2 in water samples taken nightly at depths of 250, 700 and 1200 m. Although the H2O2 content of the sea water increased from 250 to 700 m, levels remained much lower (3×10−11 to 2.8×10−12 mol l−1) than those reported for surface waters (Van Baalen and Marler, 1966) and could, therefore, probably not account for the high CAT activities in deep-living species.

A second reason could lie in the possible exposure of the fish to important occasional oxidative stress events. The respective shares of the two enzymes in the catabolism of H2O2 could explain the maintenance of CAT activity. It has been shown in mammalian cells that, although CAT decomposes H2O2 very efficiently, the enzyme is active only at rather high H2O2 concentrations (Chance et al., 1979). Thus, CAT normally plays a relatively minor role in the catabolism of H2O2 at low rates of H2O2 generation (Jones et al., 1981), but becomes indispensable when the rate of H2O2 production is enhanced, such as during oxidative stress. Godin and Garnett (1992) demonstrated that a low GPX activity could be compensated by a high CAT activity, the activities of these two enzymes being negatively correlated in various mammalian species. So, it is possible that GPX would maintain normal cell function and adjust to the metabolic activity of the cell, while CAT would be part of a stress response mechanism regardless of the low metabolic rate or the oxygen concentration in the surrounding water.

One more interesting finding is that the light organs show significantly higher activities of the three antioxidative enzymes and of CS, while their protein content is similar to that of skin lacking photophores. It is possible that these high levels of antioxidative defense in the light organs are devoted to catabolizing ROS generated by the high metabolic activity that also characterizes this tissue. However, the observation that CAT level is also higher than in skin lacking photophores suggests that this antioxidative arsenal could be related to the bioluminescent reactions taking place in these organs. These reactions may involve radical intermediates and ROS, which could be potentially harmful to the cells. Also, the luciferin in these fish, coelenterazine or some derivatives, is very sensitive to oxygen and ROS (Rees et al., 1998), and an elevated antioxidative status could therefore help preserve the dietary-acquired luciferin against spontaneous autoxidation. The high metabolic activity of the light organs may be related to the occurrence of energy-consuming inhibitory mechanisms maintaining luminous cells in a resting state, as described previously in the bioluminescent fish Porichthys notatus (Rees and Baguet, 1987, 1989).

This research was supported by the UCL Scientific Development Fund and the FRIA (National Scientific Research Fund) (B.J.J.) and by NSF grant OCE 9415543 to J.J.C.. J.-F.R. is senior researcher associate of the Belgian ‘Fonds National de la Recherche Scientifique’.

Ahmad
,
S.
(
1995
).
Antioxidant mechanisms of enzymes and proteins
.
In Oxidative Stress and Antioxidant Defenses in Biology
(ed.
S.
Ahmad
), pp.
238
272
.
New York
:
Chapman & Hall
.
Aksnes
,
A.
and
Njaa
,
L. R.
(
1981
).
Catalase, glutathione peroxidase and superoxide dismutase in different fish species
.
Comp. Biochem. Physiol
.
69B
,
893
896
.
Bainy
,
A. C. D.
,
Saito
,
E.
,
Carvalho
,
P. S. M.
and
Junqueira
,
V. B. C.
(
1996
).
Oxidative stress in gill, erythrocytes, liver and kidney of Nile tilapia (Oreochromis niloticus
).
Aquat. Toxicol
.
34
,
151
162
.
Boveris
,
A.
and
Cadenas
,
E.
(
1982
).
Production of superoxide radicals and hydrogen peroxide in mitochondria
.
In Superoxide Dismutase
, vol.
2
(ed.
L. W.
Oberley
), pp.
15
30
.
Boca Raton, FL
:
CRC Press
.
Cadenas
,
E.
(
1995
).
Mechanisms of oxygen activation and reactive oxygen species detoxification
.
In Oxidative Stress and Antioxidant Defenses in Biology
(ed.
S.
Ahmad
), pp.
1
60
.
New York
:
Chapman & Hall
.
Chance
,
B.
,
Sies
,
H.
and
Boveris
,
A.
(
1979
).
Hydroperoxide metabolism in mammalian organs
.
Physiol. Rev
.
59
,
527
605
.
Chen
,
Y.
,
Miles
,
A. M.
and
Grisham
,
M. B.
(
1995
).
Pathophysiology and reactive oxygen metabolites
.
In Oxidative Stress and Antioxidant Defenses in Biology
(ed.
S.
Ahmad
), pp.
62
95
.
New York
:
Chapman & Hall
.
Childress
,
J. J.
(
1995
).
Are there physiological and biochemical adaptations of metabolism in deep-sea animals?
Trends Ecol. Evol
.
10
,
30
36
.
Childress
,
J. J.
,
Barnes
,
A. T.
,
Quetin
,
L. B.
and
Robison
,
B. H.
(
1978
).
Thermally protecting cod ends for the recovery of living deep-sea animals
.
Deep-Sea Res
.
25
,
419
422
.
Childress
,
J. J.
and
Nygaard
,
M. H.
(
1973
).
The chemical composition of midwater fishes as a function of depth of occurrence off Southern California
.
Deep-Sea Res
.
20
,
1093
1109
.
Childress
,
J. J.
and
Siebel
,
B. A.
(
1998
).
Life at stable oxygen levels: adaptations of animals to oceanic oxygen minimum layers
.
J. Exp. Biol
.
201
,
1223
1232
.
Childress
,
J. J.
and
Somero
,
G. N.
(
1979
).
Depth-related enzymatic activities in muscle, brain and heart of deep-living pelagic marine teleosts
.
Mar. Biol
.
52
,
273
283
.
Filho
,
D. W.
,
Giulivi
,
C.
and
Boveris
,
A.
(
1993
).
Antioxidative defences in marine fish: Teleosts
.
Comp. Biochem. Physiol
.
106C
,
409
413
.
Förlin
,
L.
,
Lemaire
,
P.
and
Livingstone
,
D. R.
(
1995
).
Comparative studies of hepatic xenobiotic metabolizing and antioxidant enzymes in different fish species
.
Mar. Env. Res
.
39
,
201
204
.
Godin
,
D. V.
and
Garnett
,
M. E.
(
1992
).
Species-related variation in tissue antioxidant status. I. Differences in antioxidant enzymes profiles
.
Comp. Biochem. Physiol
.
103B
,
737
742
.
Harman
,
D.
(
1994
).
Free-radical theory of aging. Increasing the functional life span
.
Ann. N.Y. Acad. Sci
.
717
,
1
15
.
Hassan
,
H. M.
and
Schiavone
,
J. R.
(
1991
).
The role of oxygen free radicals in biology and evolution
.
In Metazoan Life without Oxygen
(ed.
C.
Bryant
), pp.
17
37
.
London
:
Chapman & Hall
.
Inoue
,
S.
,
Sugiura
,
S.
,
Kakoi
,
H.
,
Hashizume
,
K.
,
Goto
,
T.
and
Iio
,
H.
(
1975
).
Squid bioluminescence. II. Isolation from Watasenia scintillans and synthesis of 2-(p-hydroxybenzyl)-6-(p-hydroxyphenyl)-3,7-dihydroimidazo(1,2-a)pyrazin-3-one
.
Chem. Lett
.
197
,
141
144
.
Issels
,
R. D.
,
Fink
,
R. M.
and
Lengfelder
,
E.
(
1986
).
Effects of hyperthermic conditions on the reactivity of oxygen radicals
.
Free Radical Res. Commun
.
2
,
7
18
.
Jones
,
D. P.
,
Eklow
,
L.
,
Thor
,
H.
and
Orrenius
,
S.
(
1981
).
Metabolism of hydrogen peroxide in isolated hepatocytes: Relative contributions of catalase and glutathione peroxidase in decomposition of endogenously generated H2O2
.
Arch. Biochem. Biophys
.
210
,
505
516
.
Kramer
,
W.
,
Buchert
,
H.
,
Reuter
,
U.
,
Biscoito
,
M.
,
Maul
,
D. G.
,
Le Grand
,
G.
and
Ballschmiter
,
K. H.
(
1984
).
Global baseline pollution studies. IX. C6–C14 organochlorine compounds in surface-water and deep-sea fish from the Eastern North Atlantic
.
Chemosphere
13
,
1255
1267
.
Lowry
,
O. H.
,
Rosebrough
,
N. L.
,
Farr
,
A. L.
and
Randall
,
R. I.
(
1951
).
Protein measurement with Folin phenol reagent
.
J. Biol. Chem
.
193
,
265
275
.
Maral
,
J.
,
Puget
,
K.
and
Michelson
,
A. M.
(
1977
).
Comparative study of superoxide dismutase, catalase and glutathione peroxidase levels in erythrocytes of different animals
.
Biochem. Biophys. Res. Commun
.
77
,
1525
1535
.
Mccord
,
J. M.
and
Fridovich
,
I.
(
1988
).
Superoxide dismutase: the first twenty years (1968–1988
).
Free Radical Biol. Med
.
5
,
363
369
.
Meister
,
A.
(
1995
).
Mitochondrial changes associated with glutathione deficiency
.
Biochim. Biophys. Acta
1271
,
35
42
.
Nakano
,
M.
(
1990
).
Assay for superoxide dismutase based on chemiluminescence of luciferin analog
.
Meth. Enzymol
.
186
,
227
232
.
Pacifici
,
R. E.
and
Davies
,
K. J. A.
(
1991
).
Protein, lipid and DNA repair systems in oxidative stress: the free-radical theory of aging revisited
.
Gerontology
37
,
166
180
.
Paglia
,
D. E.
and
Valentine
,
W. N.
(
1967
).
Studies on the quantitative and qualitative characterization of erythrocyte glutathione peroxidase
.
J. Lab. Clin. Med
.
70
,
158
168
.
Petasne
,
R. G.
and
Zika
,
R. G.
(
1987
).
Fate of superoxide in coastal sea water
.
Nature
325
,
516
518
.
Rees
,
J. F.
and
Baguet
,
F.
(
1987
).
Metabolic control of spontaneous glowing isolated photophores of Porichthys
.
J. Exp. Biol
.
135
,
289
299
.
Rees
,
J. F.
and
Baguet
,
F.
(
1989
).
Metabolic control of luminescence in the luminous organs of the teleost Porichthys: effects of iodoacetic acid and potassium cyanide
.
J. Exp. Biol
.
143
,
347
357
.
Rees
,
J. F.
,
De Wergifosse
,
B.
,
Noiset
,
O.
,
Dubuisson
,
M.
,
Janssens
,
B.
and
Thompson
,
E. M.
(
1998
).
The origin of marine bioluminescence: turning oxygen defence mechanisms into deep-sea communication tools
.
J. Exp. Biol
.
201
,
1211
1221
.
Rodriguez-Ariza
,
A.
,
Peinado
,
J.
,
Pueyo
,
C.
and
Lopez-Barea
,
J.
(
1993
).
Biochemical indicators of oxidative stress in fish from polluted littoral areas
.
Can. J. Fish. Aquat. Sci
.
50
,
2568
2573
.
Sewell
,
R. B. S.
and
Fage
,
L.
(
1948
).
Minimum oxygen layer in the ocean
.
Nature
162
,
949
951
.
Shepherd
,
D.
and
Garland
,
P. B.
(
1969
).
Citrate synthase assay
.
Meth. Enzymol
.
13
,
11
16
.
Siebel
,
B. A.
,
Thuesen
,
E. V.
,
Childress
,
J. J.
and
Gorodezky
,
L. A.
(
1997
).
Decline in pelagic cephalopod metabolism with habitat depth reflects differences in locomotory efficiency
.
Biol. Bull
.
192
,
262
278
.
Stohs
,
S. J.
(
1995
).
Synthetic pro-oxidants: drugs, pesticides and other environmental pollutants
.
In Oxidative Stress and Antioxidant Defenses in Biology
(ed.
S.
Ahmad
), pp.
117
180
.
New York
:
Chapman & Hall
.
Thuesen
,
E. V.
and
Childress
,
J. J.
(
1993
).
Enzymatic activities and metabolic rates of pelagic chaetognaths – lack of depth-related declines
.
Limnol. Oceanogr
.
38
,
935
948
.
Thuesen
,
E. V.
and
Childress
,
J. J.
(
1994
).
Oxygen-consumption rates and metabolic enzyme activities of oceanic California medusae in relation to body size and habitat depth
.
Biol. Bull
.
187
,
84
98
.
Van Baalen
,
C.
and
Marler
,
J. E.
(
1966
).
Occurrence of hydrogen peroxide in sea water
.
Nature
211
,
951
.
Wyrtki
,
K.
(
1962
).
The oxygen minima in relation to ocean circulation
.
Deep-Sea Res
.
9
,
11
23
.
Zika
,
R. G.
,
Moffett
,
J. W.
,
Petasne
,
R. G.
,
Cooper
,
W. J.
and
Satlzman
,
E. S.
(
1985
).
Spatial and temporal variations of hydrogen peroxide in Gulf of Mexico waters
.
Geochim. Cosmochim. Acta
49
,
1173
1184
.