Prostaglandin H synthase: distinct binding sites for cyclooxygenase and peroxidase substrates

Prostaglandin H synthase has two distinct catalytic activities: a cyclooxygenase activity that forms prostaglandin G2 from arachidonic acid; and a peroxidase activity that reduces prostaglandin G2 to prostaglandin H2. Lipid hydroperoxides, such as prostaglandin G2, also initiate the cyclooxygenase reaction, probably via peroxidase reaction cycle enzyme intermediates. The relation between the binding sites for lipid substrates of the two activities was investigated with an analysis of the effects of arachidonic and docosahexaenoic acids on the reaction kinetics of the peroxidase activity, and their effects on the ability of a lipid hydroperoxide to initiate the cyclooxygenase reaction. The cyclooxygenase activity of pure ovine synthase was found to have an apparent Km value for arachidonate of 5.3 microM and a Ki value (competetive inhibitor) for docosahexaenoate of 5.2 microM. When present at 20 microM neither fatty acid had a significant effect on the apparent Km value of the peroxidase for 15-hydroperoxyeicosatetraenoic acid: the values were 7.6 microM in the absence of docosahexaenoic acid and 5.9 microM in its presence, and (using aspirin-treated synthase) 13.7 microM in the absence of arachidonic acid and 15.7 microM in its presence. Over a range of 1 to 110 microM the level of arachidonate had no significant effect on the initiation of the cyclooxygenase reaction by 15-hydroperoxyeicosatetraenoic acid. The inability of either arachidonic acid or docosahexaenoic acid to interfere with the interaction between the peroxidase and lipid hydroperoxides indicates that the cyclooxygenase and peroxidase activities of prostaglandin H synthase have distinct binding sites for their lipid substrates.     

Prostaglandin H synthase and hydroperoxides: peroxidase reaction and inactivation kinetics

The reaction kinetics of the peroxidase activity of prostaglandin H synthase have been examined with 15-hydroperoxyeicosatetraenoic acid and hydrogen peroxide as substrates and tetramethylphenylenediamine as cosubstrate. The apparent Km and Vmax values for both hydroperoxides were found to increase linearly with the cosubstrate concentration. The overall reaction kinetics could be interpreted in terms of an initial reaction of the synthase with hydroperoxide to form an intermediate equivalent to horseradish peroxidase Compound I, followed by reduction of this intermediate by cosubstrate to regenerate the resting enzyme. The rate constants estimated for the generation of synthase Compound I were 7.1 X 10(7) M-1 s-1 with the lipid hydroperoxide and 9.1 X 10(4) M-1 s-1 with hydrogen peroxide. The rate constants estimated for the rate-determining step in the regeneration of resting enzyme by cosubstrate were 9.2 X 10(6) M-1 s-1 in the case of the reaction with lipid hydroperoxide and 3.5 X 10(6) M-1 s-1 in the case of reaction with hydrogen peroxide. The intrinsic affinities of the synthase peroxidase for substrate (Ks) were estimated to be on the order of 10(-8) M for lipid hydroperoxide and 10(-5) M for hydrogen peroxide. These affinities are quite similar to the reported affinities of the synthase for these hydroperoxides as activators of the cyclooxygenase. The peroxidase activity was found to be progressively inactivated during the peroxidase reaction. The rate of inactivation of the peroxidase was increased by increases in hydroperoxide level, and decreased by increases in peroxidase cosubstrate. The inactivation of the peroxidase appeared to occur by a hydroperoxide-dependent process, originating from synthase Compound I or Compound II.     

The role of selenium peroxidases in the protection against oxidative damage of membranes

The present review deals with the chemical properties of selenium in relation to its antioxidant properties and its reactivity in biological systems. The interaction of selenite with thiols and glutathione and the reactivity of selenocompounds with hydroperoxides are described. After a short survey on distribution, metabolism and organification of selenium, the role of this element as a component of the two seleno-dependent glutathione peroxidases is described. The main features of glutathione peroxidase and phospholipid hydroperoxide glutathione peroxidase are also reviewed. Both enzymes reduce different hydroperoxides to the corresponding alcohols and the major difference is the reduction of lipid hydroperoxides in membrane matrix catalyzed only by the phospholipid hydroperoxide glutathione peroxidase. However, in spite of the different specificity for the peroxidic substrates, the kinetic mechanism of both glutathione peroxidase and phospholipid hydroperoxide glutathione peroxidase seems identical and proceeds through a tert-uni ping pong mechanism. In the reaction cycle, indeed, as supported by the kinetic data, the oxidation of the ionized selenol by the hydroperoxide yields a selenenic acid that in turn is reduced back by two reactions with reduced glutathione. Special emphasis has been given to the role of selenium-dependent glutathione peroxidases in the prevention of membrane lipid peroxidation. While glutathione peroxidase is able to reduce hydrogen peroxide and other hydroperoxides possibly present in the soluble compartment of the cell, this enzyme fails to inhibit microsomal lipid peroxidation induced by NADPH or ascorbate and iron complexes. On the other hand, phospholipid hydroperoxide glutathione peroxidase, by reducing the phospholipid hydroperoxides in the membranes, actively prevents lipid peroxidation, provided a normal content of vitamin E is present in the membranes. In fact, by preventing the free radical generation from lipid hydroperoxides, phospholipid hydroperoxide glutathione peroxidase decreases the vitamin E requirement necessary to inhibit lipid peroxidation. Finally, the possible regulatory role of the selenoperoxidases on the arachidonic acid cascade enzymes (cyclooxygenase and lipoxygenase) is discussed.     

Enzymatic reduction of phospholipid and cholesterol hydroperoxides in artificial bilayers and lipoproteins

Lipid hydroperoxides (LOOHs) in various lipid assemblies are shown to be efficiently reduced and deactivated by phospholipid hydroperoxide glutathione peroxidase (PHGPX), the second selenoperoxidase to be identified and characterized. Coupled spectrophotometric analyses in the presence of NADPH, glutathione (GSH), glutathione reductase and Triton X-100 indicated that photochemically generated LOOHs in small unilamellar liposomes are substrates for PHGPX, but not for the classical glutathione peroxidase (GPX). PHGPX was found to be reactive with cholesterol hydroperoxides as well as phospholipid hydroperoxides. Kinetic iodometric analyses during GSH/PHGPX treatment of photoperoxidized liposomes indicated a rapid decay of total LOOH to a residual level of 35-40%; addition of Triton X-100 allowed the reaction to go to completion. The non-reactive LOOHs in intact liposomes were shown to be inaccessible groups on the inner membrane face. In the presence of iron and ascorbate, photoperoxidized liposomes underwent a burst of thiobarbituric acid-detectable lipid peroxidation which could be inhibited by prior GSH/PHGPX treatment, but not by GSH/GPX treatment. Additional experiments indicated that hydroperoxides of phosphatidylcholine, cholesterol and cholesteryl esters in low-density lipoprotein are also good substrates for PHGPX. An important role of PHGPX in cellular detoxification of a wide variety of LOOHs in membranes and internalized lipoproteins is suggested from these findings.     

The selenoenzyme phospholipid hydroperoxide glutathione peroxidase

The reduction of membrane-bound hydroperoxides is a major factor acting against lipid peroxidation in living systems. This paper presents the characterization of the previously described 'peroxidation-inhibiting protein' as a 'phospholipid hydroperoxide glutathione peroxidase'. The enzyme is a monomer of 23 kDa (SDS-polyacrylamide gel electrophoresis). It contains one gatom Se/22 000 g protein. Se is in the selenol form, as indicated by the inactivation experiments in the presence of iodoacetate under reducing conditions. The glutathione peroxidase activity is essentially the same on different phospholipids enzymatically hydroperoxidized by the use of soybean lipoxidase (EC 1.13.11.12) in the presence of deoxycholate. The kinetic data are compatible with a tert-uni ping-pong mechanism, as in the case of the 'classical' glutathione peroxidase (EC 1.11.1.9). The second-order rate constants (K1) for the reaction of the enzyme with the hydroperoxide substrates indicate that, while H2O2 is reduced faster by the glutathione peroxidase, linoleic acid hydroperoxide is reduced faster by the present enzyme. Moreover, the phospholipid hydroperoxides are reduced only by the latter. The dramatic stimulation exerted by Triton X-100 on the reduction of the phospholipid hydroperoxides suggests that this enzyme has an 'interfacial' character. The similarity of amino acid composition, Se content and kinetic mechanism, relative to the difference in substrate specificity, indicates that the two enzymes 'classical' glutathione peroxidase and phospholipid hydroperoxide glutathione peroxidase are in some way related. The latter is apparently specialized for lipophylic, interfacial substrates.     

α-Tocopherol enhances the peroxidase activity of hemoglobin on phospholipid hydroperoxide

Summary

We have used direct separation of phospholipid hydroperoxide and phospholipid hydroxide by high performance liquid chromatography to examine the phospholipid hydroperoxide peroxidase activity of hemoglobin (Hb) in the presence of hydrogen donors. Hb exhibits phospholipid hydroperoxide peroxidase activity and rapidly breaks down phospholipid hydroperoxide to thiobarbituric acid-reactive substances. However, in the presence of α-tocopherol, some phospholipid hydroperoxide is converted to phospholipid hydroxide, which is more stable than the hydroperoxide and is much less reactive with thiobarbituric acid. Other electron donors such as glutathione and ascorbate are less effective than α-tocopherol. Free cysteine also shows some ability to reduce phospholipid hydroperoxides to corresponding hydroxides, but cys-93β of Hb did not participate in the reaction, as shown by N-ethylmaleimide modification. Hemin alone catalysed the reaction, in the absence of protein. The results therefore show that Hb catalyses an apparent phospholipid hydroperoxide α-tocopherol peroxidase reaction due to bound hemin, and that the reduction depends on the ability of hydrogen donors to react with the intermediate phospholipid alkoxyl radical and does not involve reduction by deprotonated sulfhydryl groups.     

Lipid peroxidation and antioxidant status in blood of patients with uterine myoma, endometrial polypus, hyperplastic and malignant endometrium

Oxidative stress is considered to be involved in pathogenesis of many disorders of the female genital tract. In this study, we explored the lipid peroxidation levels and antioxidant enzyme activities in women diagnosed with different forms of uterine diseases in order to evaluate the extent of oxidative stress in blood of such patients. Blood samples of healthy subjects and gynecological patients were collected and subjected to assays for superoxide dismutase, catalase, glutathione peroxidase, glutathione reductase and lipid hydroperoxides. The results show that alterations of measured parameters vary with the enzyme type and diagnosis. However, both reduction in antioxidants and elevation of lipid peroxidation were observed in general. Lipid hydroperoxides level was negatively correlated to superoxide dismutase and glutathione peroxidase activities, as well as positively correlated to catalase activity. In addition, the lipid hydroperoxides/ glutathione peroxidase ratio was found to be increased, according to the type of uterine disease. The obtained results show that perturbation of antioxidant status is more pronounced in blood of patients with premalignant (hyperplastic) and malignant (adenocarcinoma) lesions, compared to those with benign uterine changes such as polypus and myoma.     

The role of human glutathione S-transferases hGSTA1-1 and hGSTA2-2 in protection against oxidative stress.

In order to elucidate the protective role of glutathione S-transferases (GSTs) against oxidative stress, we have investigated the kinetic properties of the human alpha-class GSTs, hGSTA1-1 and hGSTA2-2, toward physiologically relevant hydroperoxides and have studied the role of these enzymes in glutathione (GSH)-dependent reduction of these hydroperoxides in human liver. We have cloned hGSTA1-1 and hGSTA2-2 from a human lung cDNA library and expressed both in Escherichia coli. Both isozymes had remarkably high peroxidase activity toward fatty acid hydroperoxides, phospholipid hydroperoxides, and cumene hydroperoxide. In general, the activity of hGSTA2-2 was higher than that of hGSTA1-1 toward these substrates. For example, the catalytic efficiency (kcat/Km) of hGSTA1-1 for phosphatidylcholine (PC) hydroperoxide and phosphatidylethanolamine (PE) hydroperoxide was found to be 181.3 and 199.6 s-1 mM-1, respectively, while the catalytic efficiency of hGSTA2-2 for PC-hydroperoxide and PE-hydroperoxide was 317.5 and 353 s-1 mM-1, respectively. Immunotitration studies with human liver extracts showed that the antibodies against human alpha-class GSTs immunoprecipitated about 55 and 75% of glutathione peroxidase (GPx) activity of human liver toward PC-hydroperoxide and cumene hydroperoxide, respectively. GPx activity was not immunoprecipitated by the same antibodies from human erythrocyte hemolysates. These results show that the alpha-class GSTs contribute a major portion of GPx activity toward lipid hydroperoxides in human liver. Our results also suggest that GSTs may be involved in the reduction of 5-hydroperoxyeicosatetraenoic acid, an important intermediate in the 5-lipoxygenase pathway.     

Mitochondrial defects and oxidative damage in patients with peripheral arterial disease

Abnormal mitochondrial function is present in patients with peripheral arterial disease and may contribute to its clinical manifestations. However, the specific biochemical mitochondrial defects and their association with increased oxidative stress have not been fully characterized. Gastrocnemius muscle was obtained from peripheral arterial disease patients (n = 25) and age-matched controls (n = 16) and mitochondrial parameters were measured. Complexes I through IV of the electron transport chain were individually evaluated to assess for isolated defects. Muscle was also evaluated for protein and lipid oxidative changes by measuring the levels of protein carbonyls, lipid hydroperoxides, and total 4-hydroxy-2-nonenal binding and for the activities of the antioxidant enzymes superoxide dismutase, catalase, and glutathione peroxidase. Mitochondrial electron transport chain complexes I, III, and IV in arterial disease patients demonstrated significant reductions in enzymatic activities and mitochondrial respiration compared to controls. Oxidative stress biomarker analysis demonstrated significantly increased levels of protein carbonyls, lipid hydroperoxides, and 4-hydroxy-2-nonenal compared to control muscle. Antioxidant enzyme activities were altered, with a significant decrease in superoxide dismutase activity and significant increases in catalase and glutathione peroxidase. Peripheral arterial disease is associated with abnormal mitochondrial function and evidence of significant oxidative stress.     

Enzymes utilizing hydroperoxides and O ./2 in the myocardium of rats of different ages

Superoxide dismutase activity of myocardial cytosol increases with rat age (1, 3, 12, 24 months) whereas catalase activity is unchanged. Glutathione peroxidase activity with H2O2, cumene and t-butyl hydroperoxides as substrates increases more than 1.5-fold over the period from 1 to 12 months. The high activity of these enzymes is maintained in 24-month-old rats.     

The role of selenium in the regulation of lipid peroxidation in biological membranes and glutathione peroxidase activity

The antioxidative effect of selenium cannot be exclusively due to the functioning of the selenium-dependent glutathione peroxidase mechanism of utilization of various hydroperoxides. This hypothesis is based on the following experimental evidence. Selenium ions are readily incorporated into animal organs and tissues immediately after injection (2 hours) as well as into cell organelles and cytosol where they inhibit lipid peroxidation. The activity of glutathione peroxidase (EC 1.1.1.19) in rat liver and guinea pig cytosol is thereby unchanged but increases drastically after 12 hours reaching a maximum an the 3rd-4th day. The effectiveness of lipid peroxidation inhibition does not increase under these conditions. Although the glutathione peroxidase activity is absent in the nuclei and microsomes, exogenous selenium inhibits lipid peroxidation in these organelles. The activity of the rat liver cytosolic enzyme markedly exceeds that of its guinea pig counterpart. However, lipid peroxidation in guinea pig liver occurs less intensively than that in rat liver cytosol.     

Purification and physicochemical characterization of a recombinant phospholipid hydroperoxide glutathione peroxidase from Oryza sativa

Phospholipid hydroperoxide glutathione peroxidase (PHGPx) is an unique antioxidant enzyme that directly reduces lipid hydroperoxides in biomembranes. In the present work, the entire encoding region for Oryza sativa PHGPx was expressed in Escherichia coli M15, and the purified fusion protein showed a single band with 21.0 kD and pI = 8.5 on SDS- and IFE-PAGE, respectively. Judging from CD and fluorescence spectroscopy, this protein is considered to have a well-ordered structure with 12.2% alpha-helix, 30.7% beta-sheet, 18.5% gamma-turn, and 38.5% random coil. The optimum pH and temperature of the enzyme activity were pH 9.3 and 27 degrees C. The enzyme exhibited the highest affinity and catalytical efficiency to phospholipid hydroperoxide employing GSH or Trx as electron donor. Moreover, the protein displayed higher GSH-dependent activity towards t-Butyl-OOH and H(2)O(2). These results show that OsPHGPx is an enzyme with broad specificity for hydroperoxide substrates and yielded significant insight into the physicochemical properties and the dynamics of OsPHGPx.     

Inhibition of arachidonate metabolism in human epidermoid carcinoma a431 cells overexpressing phospholipid hydroperoxide glutathione peroxidase

Phospholipid hydroperoxide glutathione peroxidase (PHGPx), a selenium-dependent glutathione peroxidase, can interact with lipophilic substrates, including phospholipid hydroperoxides, fatty acid hydroperoxides and cholesterol hydroperoxides, and can reduce them to hydroxide compounds. It also seems to be a major regulator of lipid oxygenation in human epidermoid carcinoma A431 cells. In order to study the functional role of PHGPx in the regulation of 12-lipoxygenase and cyclooxygenase, cDNA of PHGPx was inserted into pcDNA3.1/His, and a plasmid designated as S4 with the His-tag sequence inserted between PHGPx and its 3'-untranslated region was constructed. A number of stable transfectants of A431 cells that could express the tag-PHGPx were generated using plasmid S4. Using an intact cell assay system, the metabolism of arachidonic acid to prostaglandin E(2) significantly decreased in stable transfectants of overexpressing PHGPx compared to that in a vector control cell line. If the intact cell assay was carried out in the presence of 13-hydroperoxyoctadecadienoic acid as a stimulator of lipid peroxidation, formation of 12-hydroxyeicosatetraenoic acid from arachidonic acid also significantly decreased in stable transfectants of overexpressing PHGPx compared to that in a vector control cell line, indicating that PHGPx could downregulate the 12-lipoxygenase activity in cells. These results support the hypothesis that PHGPx plays a pivotal role in the regulation of arachidonate metabolism in A431 cells.     

Effects of selenium-deficient diets on the production of prostaglandins and other oxygenated metabolites of arachidonic acid and linoleic acid by rat and rabbit aortae

Selenium is an essential component of glutathione peroxidase, which reduces free and esterified hydroperoxides of polyunsaturated fatty acids. Adequate glutathione peroxidase activity could be important for the maintenance of prostacyclin synthesis by blood vessels, since hydroperoxides can inhibit the formation of this substance. We have investigated the effects of dietary selenium deficiency on glutathione peroxidase activity and the synthesis of 6-oxoprostaglandin F1 alpha and monohydroxy and trihydroxy metabolites of polyunsaturated fatty acids by aorta. The latter products can be formed either by the actions of cyclooxygenase or lipoxygenase or by lipid peroxidation. Aortic glutathione peroxidase activity was reduced by over 80% by feeding rats a selenium-deficient diet for 4 weeks, and to undetectable levels after 6 weeks. There were no appreciable differences in the levels of free and esterified oxygenated metabolites of linoleic acid or arachidonic acid between the control and treated groups after 4 weeks. However, after 6 weeks, there were modest, but statistically significant reductions in the formation of 6-oxoprostaglandin F1 alpha and monohydroxy products formed by cyclooxygenase. On the other hand, the amounts of esterified 18:2 metabolites appeared to be higher in aortae from animals on the selenium-deficient diet, although only the increase in esterified 9-hydroxy-10,12-octadecadienoic acid was statistically significant. These results suggest that selenium deficiency can affect the formation of prostacyclin and other oxygenated metabolites of polyunsaturated fatty acids by aorta, possibly by increasing lipid peroxidation. However, the differences between control and selenium-deficient rats after 6 weeks were not very dramatic, in spite of the fact that glutathione peroxidase activity was undetectable. It would therefore appear that additional mechanisms are also involved in controlling the levels of lipid hydroperoxides in aorta.     

Different effects of Triton X-100, deoxycholate, and fatty acids on the kinetics of glutathione peroxidase and phospholipid hydroperoxide glutathione peroxidase

The effects of Triton X-100, deoxycholate, and fatty acids were studied on the two steps of the ping-pong reaction catalyzed by Se-dependent glutathione peroxidases. The study was carried out by analyzing the single progression curves where the specific glutathione oxidation was monitored using glutathione reductase and NADPH. While the "classic" glutathione peroxidase was inhibited only by Triton, the newly discovered "phospholipid hydroperoxide glutathione peroxidase" was inhibited by deoxycholate and by unsaturated fatty acids. The kinetic analysis showed that in the case of glutathione peroxidase only the interaction of the lipophilic peroxidic substrate was hampered by Triton, indicating that the enzyme is not active at the interface. Phospholipid hydroperoxide glutathione peroxidase activity measured with linoleic acid hydroperoxide as substrate, on the other hand, was not stimulated by the Triton concentrations which have been shown to stimulate the activity on phospholipid hydroperoxides. Furthermore a slight inhibition was apparent at high Triton concentrations and the effect could be attributed to a surface dilution of the substrate. Deoxycholate and unsaturated fatty acids were not inhibitory on glutathione peroxidase but inhibited both steps of the peroxidic reaction of phospholipid hydroperoxide glutathione peroxidase, in the presence of either amphiphilic or hydrophilic substrates. This inhibition pattern suggests an interaction of anionic detergents with the active site of this enzyme. These results are in agreement with the different roles played by these peroxidases in the control of lipid peroxide concentrations in the cells. While glutathione peroxidase reduces the peroxides in the water phase (mainly hydrogen peroxide), the new peroxidase reduces the amphyphilic peroxides, possibly at the water-lipid interface.     

Microsomal-catalyzed hydroperoxide-dependent C-oxidation of amines

Organic hydroperoxides are capable of supporting the C-oxidation of several different amines in the presence of hepatic microsomes. Evidence is presented that indicates that microsomal cytochrome P-450 acts as the catalyst. Removal of the NADPH-cytochrome $\text{c}$ oxidoreductase or essential phospholipid from microsomes does not significantly affect the peroxidase activity. Of the amine substrates C-oxidized by organic hydroperoxides in the presence of microsomes, only aminopyrine and dimethylaniline are rapidly oxidized by hydroperoxides in the presence of catalase. The catalase-mediated reaction can also be distinguished from the microsomal-catalyzed reaction by the use of differential inhibitors.     

Iodometric measurement of lipid hydroperoxides in human plasma

Many assay techniques have been used to measure lipid hydroperoxides in plasma, including absorbance of conjugated dienes and reactivity with thiobarbituric acid. Because these measurements are not specific for lipid hydroperoxides, we modified an exisiting iodometric method to correct for interfering phenomena and to provide a more specific measurement of the lipid hydroperoxide content of plasma. To ensure reproducible extraction of hydroperoxides from the many possible forms in plasma, the plasma was treated to hydrolyze enzymatically cholesterol ester, triglycerides, and phospholipids, and the nonesterified fatty acid peroxides were then extracted with ethyl acetate. Extracted lipids were reacted with potassium iodide in acetic acid and methylene chloride, and the resulting triiodide ion (I3-) was measured spectrophotometrically. Correction for nonoxidizing chromophores was made after back-titration of the triiodide ion to iodide with sodium thiosulfate and other non-peroxide oxidants were estimated by their resistance to reduction with glutathione peroxidase. Recovery of added hydroperoxide standards provided routine validations of the procedure's efficiency. The method indicated that insignificant amounts of hydroperoxide may be in the less polar lipids, but the total amount of lipid hydroperoxide esterfied in the plasma lipids of apparently healthy humans may be as much as 4.0 +/- 1.7 microM.     

Damage to erythrocytes caused by 2,3,7,8-tetrachloro-dibenzo-p-dioxin (in vitro)

The effects of the exposure of human erythrocytes to different concentrations of 2,3,7,8-tetrachlorodibenzo-p-dioxin were studied. Particular attention was paid to lipid peroxidation, haemoglobin oxidation, and changes in the activity of catalase and glutathione peroxidase. Human erythrocytes at a 5% haematocrit were incubated with 2,3,7,8-TCDD at concentrations of 0.2 ppm to 1.6 ppm (ng-microg/ml erythrocytes) for 1 hour. The results obtained show that 2,3,7,8-TCDD induces the generation of lipid peroxides and the oxidation of Hb, and decreases the activity of catalase and glutathione peroxidase. This supports the thesis that TCDD causes oxidative stress in erythrocytes.     

Cytoprotection against lipid hydroperoxides correlates with increased glutathione peroxidase activities, but not selenium uptake from different selenocompounds

Cells cultivated under standard conditions were highly deficient in tocopherol, selenium, and glutathione peroxidase (GPx) activities. We investigated whether and to what extent the addition of different selenocompounds to growth media would alter biochemical, physiological, and pathophysiological parameters of cultured liver cells. Cellular uptake of selenium, GPx activities, and cytoprotection were measured and compared in human hepatoma cells (HepG2). Selenite and selenocystine were Se donors of high bioavailability (i.e., with these culture supplements, the increased Se uptake, induction of GPx isoenzymes, and protection of treated cells from lipid hydroperoxides were well correlated). In contrast, selenium from selenomethionine was incorporated into cellular proteins but had no effect on GPx activities or cytoprotection. The data show that not all selenium donors provide selenium, which is bioactivated to act as antioxidant. Thus, cellular selenium content, in general, did not correlate with cytoprotective activity of this trace element. However, cellular GPx activities at different times, with different concentrations, and with different Se donors always correlated with protection from lipid hydroperoxides and may, thus, represent a more reliable parameter to define adequate Se supply.