Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Id, Biochemical Functions Klaus Schwarz (1965), in whose laboratory the essentiality of selenium in animals was discovered (Schwarz and Foltz, 1957), postulated that sele- nium functioned as an essential cofactor at specific sites of intermediary metabolism. Currently, the known biochemical functions of selenium are as a component of the enzyme glutathione peroxidase, found in animals, and of several bacterial enzymes. The selenium deficiency signs observed in animals can be partially explained by a lack of glutathione peroxidase (GSH-Px) (Hoekstra, 1975), but this does not eliminate the possibility of other roles for selenium in animals (see "Other Functions of Selenium," page 491. For example, Burk and Gregory (1982) have recently reported a selenium-binding protein of unknown function in rat liver and plasma that has properties quite distinct from those of GSH-Px. In contrast to animals, microorganisms generally grow and reproduce well in the absence of sele- nium, a lack of certain selenium-containing enzymes being the only sign of selenium deficiency in bacteria. In 1954, Pinsent reported that selenium (and molybdenum) were neces- sary for the appearance of formate dehydrogenase activity in E. coli. At about the same time that Rotruck et al. (1973) reported that GSH-Px in the rat was a selenoenzyme, the microbial enzymes formate dehydrogenase (An- dreesen and Ljungdahl, 1973) and protein A of glycine reductase (Turner and Stadtman, 1973) were shown to be selenoenzymes. These two enzymes are inhibited by iodoacetamide (Turner and Stadtman, 1973; Enoch and Lester, 1975), and the form of selenium in reduced glycine reductase and formate dehydrogenase has been reported to be selenocysteine (Cone et al., 40
Biochemical Functions 41 1976, 1977; Jones et al., 19791. Two other bacterial enzyme activities, nico- tinic acid hydroxylase (Imhoff and Andreesen, 1979) and xanthine dehydro- genase (Wagner and Andreesen, 1979) have been reported recently to re- quire the presence of selenium. A fifth possible selenoprotein, thiolase, isolated from two bacterial species grown in the presence of selenium, has been reported to contain selenium (Hartmanis, 19801. In animals, GSH-Px is presently the only known selenoenzyme, and thus knowledge of the chem- istry and biochemistry of GSH-Px is an important part of our current under- standing of the biochemical function of selenium in animals. NATURE AND PROPERTIES OF GLUTATHIONE PEROXIDASE Glutathione peroxidase (glutathione: H2 O2 oxidoreductase, E. C. 1.1 1.1 .9) was discovered by Mills (1957), who found that this enzyme in the presence of reduced glutathione would protect erythrocytes against H2O2-induced and ascorbate-induced hemoglobin oxidation and hemolysis. The addition of glucose to the incubation medium was shown to protect the erythrocyte at least in part by maintaining the concentration of reduced glutathione (Mills and Randall, 1958; Cohen and Hochstein, 19631. Vitamin E, added to the diet or to the incubation medium, protected erythrocytes against hemolysis (Dam, 1957), but dietary selenium was usually reported not to prevent hemolysis (Christensen et al., 1958; Gitler et al., 1958~. Rotruck et al. (1971, 1972a) integrated these facts and demonstrated that dietary sele- nium protected erythrocytes from ascorbic acid-induced hemolysis only if glucose was included in the incubation medium. The ability of vitamin E to prevent hemolysis was not affected by glucose, and glucose had no protective effect if the erythrocytes were from selenium-deficient rats. Glutathione levels were higher in erythrocytes from selenium-deficient rats than from selenium-adequate rats, suggesting that the defect in selenium deficiency was not in the maintenance of glutathione levels, but rather in its utilization. Rotruck et al. (1972b, 1973) then focused on GSH-Px and discovered that it was a selenoenzyme. Glutathione peroxidase has been purified from the tissues of cattle, humans, swine, sheep, and rats, and shown to be an approximately 80,000- dalton enzyme consisting of four apparently identical subunits (Ganther et al., 1976~. Determination of GSH-Px molecular weight by sedimentation- equilibrium indicates that the molecular weight of GSH-Px differs from species to species: 76,000 + 1,000 for rat liver (Nakamura et al., 1974), 83,800 + 1,200 for bovine erythrocytes (Flohe et al., 1971a), 95,000 + 3,000 for human erythrocytes (Awasthi et al. ,1975~. The size of GSH-Px can vary also from tissue to tissue in the same species (Awasthi et al. ,1975,1979;
42 SELENIUM IN NUTRITION Sunde et al., 1978~. Oh et al. (1974) and Fiche et al. (1973) independently demonstrated that GSH-Px from ovine and bovine erythrocytes contained 4 g-atoms of selenium per mole of GSH-Px. This value has been confirmed by Nakamura et al. (1974) and Awasthi et al. (1975) for rat liver and human erythrocytes. Glutathione peroxidase contains no heme or flavin, in contrast to other peroxidases (Flohe et al., 1971c), and neutron activation analysis has indicated that no metals other than selenium are present in GSH-Px (Flohe et al., 1973~. Following the demonstration that the selenium in protein A of reduced microbial glycine reductase was selenocysteine (Cone et al., 1976, 1977), Forstrom et al. (1978) and Wendel et al. (1978) reported that the selenium in reduced GSH-Px is present also as selenocysteine. These workers alkylated the reduced form of purified GSH-Px with iodoacetate or chloroacetate and then hydrolyzed the enzyme in hydrochloric acid. The alkylated derivatives were separated from the other amino acid residues by standard amino acid analysis procedures. The selenium-containing peak co-chromatographed with authentic carboxymethylselenocysteine, indicating that the selenium in the reduced form of GSH-Px is present as selenocysteine. Sequential Edman degradation was used to demonstrate that the selenocysteine is incorporated into the peptide backbone of GSH-Px (Zakowski et al., 1978~. The amino acid composition of rat liver GSH-Px was determined by Nakamura et al. (1974~. Each subunit was reported to contain two cysteine and three methionine residues out of a total of 153 amino acids. The amino acid composition of bovine erythrocyte GSH-Px is similar; it contains 178 amino acid residues, and a tentative amino acid sequence has been pub- lished (Ladenstein et al., 1979~. Bovine erythrocyte GSH-Px has been crystallized, and the three dimen o signal structure at 2.8 A resolution has been determined (Ladenstein and Wendel, 1976; Ladenstein et al., 19791. The subunits are nearly spherical, o with a radius of 18.7 A; the subunits are identical or at least very similar, with one Se atom per subunit. The GSH-Px tetramer, with dimensions of o 90.4 X 109.5 X 58.6 A, appears to be an almost flat, planar arrangement of two dimers. Each active site consists of regions from both subunits of a dimer. The selenium atoms are located on the surface of the enzyme; the selenium atoms in a dimer are 21 A apart, suggesting that only one sele- nium atom is present per active site. The number of active sites per tetramer has not been established nor has whether all four selenium atoms are active catalytically. The selenium atom appears as a "protrusion of the main chain density," and Ladenstein et al. (1979) concluded that a selenocysteine or selenocysteine derivative satisfactorily fits this density. Thiols other than glutathione are poor substrates for GSH-Px (Mills, 1959~. Flohe et al. (1971b) studied the donor substrate specificity in detail
Biochemical Functions 43 and found no other substrate with more than 30 percent of the activity of glutathione. In contrast to catalase, GSH-Px will destroy a variety of or- ganic hydroperoxides at rates similar to those of H202 destruction (Little and O'Brien, 1968~. Various lipid hydroperoxides, steroid hydroperoxides, thymine hydroperoxide, nucleic acid hydroperoxides, prostaglandin hydro- peroxides, and presumably vitamin K hydroperoxide have been shown to be acceptor substrates for GSH-Px (Christophersen, 1969; Gunzler et al., 1972; Little, 1972; Nugteren and Hazelhof, 1973; Larson and Suttie, 1978~. On the other hand, cholesterol-25- and cholesterol-7ar-hydroperoxide have been reported to be poor substrates (Little, 1972) and the enzyme cannot reduce fatty acid peroxides esterified in phospholipids (McCay et al. ,1981). Because several peroxide substrates elicit similar maximal velocities, a peroxide-enzyme complex is most likely not formed during catalysis. Flohe and co-workers (Flohe et al., 1972; Gunzler et al., 1972; Flohe and Gunzler, 1974) have studied extensively the kinetics of GSH-Px. They concluded that the reaction mechanism is a ter uni ping-pony mechanism: the first step is an oxidation of the enzyme by the peroxide substrate followed by release of the corresponding alcohol. This is followed by two successive additions of reduced glutathione (GSH) to the enzyme, and then release of oxidized glutathione (GSSG). This mechanism takes into ac- count the inability to saturate the enzyme with either substrate and the similar Vmax for most peroxide substrates. Formation of ternary or quater- nary complexes is not predicted during catalysis. The mechanism is best described as a series of three bimolecular steps. Ganther et al. (1976) proposed that the selenium in GSH-Px cycles between a selenol (E-SeH) and a selenenic acid (E-SeOH), or between a selenenic acid and a seleninic acid (E-SeOOH). X-ray photoelectron spec- troscopy (Wendel et al., 1975; Chin et al., 1977) and x-ray crystallographic studies (Ladenstein et al., 1979) have provided some evidence for different oxo-derivatives of selenium during catalysis, but have not conclusively identified these several oxidation states. Purified GSH-Px not exposed to GSH is i rreversibly inhibited by cyanide; reduction of the enzyme with GSH, dithiothreitol, or dithionite prevents this inhibition by cyanide (Prohaska et al., 1977b). However, reduced GSH-Px is irreversibly inhibited by iodoacetate (Flohe and Gunzler, 1974~. Flohe and Gunzler (1974) were unable to detect bound GSH when excess GSH was used, but Kraus et al. (1980) have demon- strated that GSH can bind to a form of GSH-Px different from the fully reduced and fully oxidized forms of the enzyme. These results suggest that GSH binds to an intermediate form of GSH-Px by a selenyl-sulfide (E-Se- SG) or seleninyl-sulfide (E-Se(O)SG) linkage. Thus chemical reactivity can be used to identify the various oxidation states of GSH-Px that
44 SELENIUM IN NUTRITION seemingly correspond to the intermediate forms of the enzyme proposed by the kinetic data of Flohe et al. (1972), and to those of the proposed molecular mechanism of Ganther et al. (1976~. Glutathione peroxidase has been assayed either by direct measurement of the disappearance of GSH (Mills, 1959; Flohe, 1971; Hafeman et al., 1974) or by the use of an excess of glutathione reductase (NAD(P)H:GSSG oxidoreductase EC 1.6.4.2), which couples the utilization of GSH to the disappearance of NADPH (Paglia and Valentine, 1967~. Because the apparent Km for each substrate is dependent on the concentration of the other substrate, the assay is usually conducted with a fixed initial level of GSH and a concentration of peroxide far above the apparent Km for the peroxide. Sodium aside, and not cyanide, should be used to inhibit cata- lase, because of cyanide inhibition of GSH-Px (Prohaska et al., 1977b). Because H2O2 reacts faster spontaneously with GSH than do organic hydroperoxides, causing a higher background, several workers have sug- gested that cumene hydroperoxide or t-butyl hydroperoxide be used for as- saying GSH-Px. However, with the discovery that the GSH-S-tranferases have GSH-Px activity with organic hydroperoxide substrates but not H2O2 (Prohaska and Ganther 1977), these substrates are not recommended for the determination of Se-dependent GSH-Px activity unless the absence of GSH-S-transferase activity has been confirmed chromatographically. Another pitfall encountered in GSH-Px activity determination is that hemoglobin will catalyze the oxidation of GSH by H2O2. Paglia and Valentine (1967) suggested that the conversion of methemoglobin to cyanomethemoglobin would eliminate this problem, but Flohe and Brand (1970) reported that this treatment did not completely eliminate hemoglo- bin-induced peroxidase activity when H202 was used as a substrate. Gunzler et al. (1974) found that non-GSH-Px catalysis of GSH oxidation was completely eliminated by cyanomethemoglobin formation if one of the organic hydroperoxide substrates was used and the cyanomethemoglobin conversion was performed on freshly drawn blood. Provided that GSH-S- transferase activity is not present, this method should be adequate for the determination of GSH-Px in low activity samples, such as human blood. However, some investigators have found that this technique does not compensate satisfactorily for peroxidase activity associated with human hemoglobin (Burk et al., 1981; Butler et al., 1982~. GLUTATHIONE PEROXIDASE ACTIVITY IN ANIMALS The distribution of GSH-PX activity in animal tissues has been reviewed thoroughly by Ganther et al. (1976~. In the selenium-adequate rat the highest GSH-PX activity is found in the liver and erythrocytes; moderate
Biochemical Functions 45 activity in the heart, kidney, lung, and adrenal glands; and low activity in the brain, testis, and lens (Lawrence et al., 19741. In the selenium-ade- quate chick, GSH-Px activity is high in the liver and moderately low in erythrocytes (Omaye and Tappel, 19741. In the selenium-adequate lamb, GSH-Px activity is very high in erythrocytes and low in the liver (Oh et al., 1976a,b). The guinea pig has been reported to have very low levels of liver GSH-Px (Lawrence and Burk, 1978~. These reports demonstrate that the tissue distribution of GSH-Px varies from species to species. Approximately 60 percent of the GSH-Px activity in rat liver is cytosolic and 30 percent mitochondrial (Green and O'Brien, 1970; Flohe and Schle- gel, 1971~. At least 60 percent of rat liver mitochondrial Se is associated with GSH-Px (Levander et al., 19741. Between 75 and 100 percent of ovine erythrocyte selenium is accounted for by selenium incorporated into GSH- Px (Oh et al., 1974), but as little as 10 percent of the selenium in livers from selenium-adequate sheep was reported to be accounted for by GSH- Px (Suede et al., 1978~. These differences may reflect the presence of other specific selenoproteins, or they may simply indicate that in tissue with low GSH-Px activity, a large proportion of the selenium is incorporated into or bound nonspecifically to proteins in general. Selenium supplementation has been shown to increase tissue GSH-Px activity above that of unsupplemented animals in rats (Chow and Tappel, 1974; Hafeman et al., 1974; Lawrence et al., 1974; Reddy and Tappel, 1974; Smith et al., 1974;), chicks (Noguchi et al., 1973; Omaye and Tap- pel, 1974; Cantor et al., 1975b), sheep (Godwin et al., 1975; Oh et al., 1976a,b; Whanger et al., 1977, 1978b), cattle (Anderson et al., 1978), mice (Revis et al., 1979), horses (Caple et al., 1978), swine (Sivertsen et al., 1977; Hakkarainen et al., 1978), Japanese quail (Kling and Soares, 1978) deer (Brady et al., 1978b), and salmon (Poston et al., 19761. In humans, strong correlations between blood selenium and GSH-Px activity have been observed in individuals consuming low levels of selenium (Thomson et al., 1977b; McKenzie et al., 1978), but Schmidt and Heller (1976) and Schrauzer and White (1978) have reported a lack of correlation between blood GSH-Px activity and blood levels of selenium in individuals consum- ing adequate or high levels of selenium. This is most likely due to the non- specific incorporation of selenomethionine into blood proteins, as de- scribed below. In addition, blood selenium and GSH-Px were not corre- lated in blood of pregnant women (Butler et al., 1980, 1982), indicating other factors may be involved. Of the various body components studied, liver and plasma GSH-Px ac- tivities decrease most rapidly during selenium depletion and increase most rapidly during selenium repletion (Noguchi et al., 1973; Chow and Tappel, 1974; Hafeman et al., 1974; Lawrence et al., 1974; Omaye and Tappel,
46 S ELENIUM IN NUTRITI O N 1974; Reddy and Tappel, 1974; Smith et al., 1974; Oh et al., 1976a,b). Tissue GSH-Px activity has been reported by several workers to also in- crease with the log of the dietary selenium concentration (Omaye and Tap- pel, 1974; Smith et al., 1974), yet other workers (Hafeman et al., 1974; Oh et al., 1976a,b) have reported that tissue GSH-Px activity plateaus at ap- proximately 0.1 ppm dietary selenium for all tissues except erythrocytes and pancreas. This leveling off of GSH-Px activity, but not selenium con- tent, with increasing selenium supplementation suggests that tissue GSH- Px activity may be a better indicator of effective selenium status than is tissue selenium content. Direct comparison of the biopotency of various selenium compounds for GSH-Px synthesis has shown that, in selenium-deficient chicks fed diets supplemented with less than 0.1 ppm selenium, selenite was twice as effec- tive as selenomethionine in increasing plasma, liver, and heart GSH-Px (Noguchi et al., 1973), but that with higher levels of selenium supplemen- tation or a longer supplementation period, selenite and selenomethionine had equal biopotency (Omaye and Tappel, 1974; Cantor et al., 1975b). Pierce and Tappel (1977) found that a single large oral dose of selenite or selenomethionine (300 ,ug Se/90 g rat) given to selenium-deficient rats re- sulted in similar increases in liver, kidney, small intestine, and stomach GSH-Px activity 48 hours after selenium administration. However, Sunde et al. (1981) found that with dietary selenium supplementation to sele- nium-deficient rats the increase in liver, plasma, and heart GSH-Px due to selenite supplementation was not affected by the level of dietary methionine, but that suboptimal dietary methionine impaired the biopo- tency of selenomethionine below 0.5 ppm selenium. Additional dietary methionine resulted in a selenomethionine biopotency equivalent to sel- enite biopotency. Selenomethionine degradation has been thought to follow the usual methionine degradation pathways (Schwarz, 1965~. McConnell and cow- orkers (McConnell and Cho, 1965, 1967; Hoffman et al., 1970; McConnell and Hoffman, 1972a,b) have shown that selenomethionine can readily re- place methionine in intestinal uptake, acylation of methionine-tRNA, and incorporation into general body proteins. Clearly, selenomethionine fol- lows the metabolic pathways of intact methionine, and thus when methio- nine is limiting, selenomethionine will be incorporated into general body proteins in place of methionine, where it will be unavailable for GSH-Px synthesis until these proteins turn over. This can explain the observations of Sunde et al. (1981) that selenomethionine biopotency is impaired when dietary methionine is limiting. Hawkes et al. (1979) have indicated that selenocysteine may be incorpo- rated directly into the peptide backbone of GSH-Px via a specific seleno
Biochemical Functions 47 cysteine-tRNA, but Sunde and Hoekstra (1980) have shown with isotope dilution experiments that selenite and selenide are more readily metabo- lized than is selenocysteine to a form of selenium that can be incorporated into GSH-Px. These workers have suggested that a form of selenium other than free selenocysteine is incorporated posttranslationally into an amino acid residue (such as serine) already incorporated in the pre-GSH-Px protein. Germain and Arneson (1977) demonstrated that selenate did not pro- vide selenium for GSH-Px synthesis in mouse neuroblastoma cells, whereas selenite very effectively increased GSH-Px activity, indicating that selenate is not the form used for GSH-Px synthesis, and that selenate re- duction to selenite is impaired in these cells. Selenocystine and selenite had equivalent biopotency in this in vitro system (Germain and Arneson, 1979~. White and Hoekstra (1979) demonstrated that selenium from sel- enite was much more readily incorporated into GSH-Px than was selenium from selenomethionine. Because animals are unable to synthesize sele- nomethionine from inorganic selenium (Cummins and Martin, 1967; Jen- kins, 1968) and are able to synthesize selenocysteine from inorganic sele- nium only to a minimal extent (Olson and Palmer, 1976), these results further indicate that another inorganic form of selenium serves as the im- mediate precursor used for selenium incorporation into GSH-Px. Further experiments will be necessary to establish the pathways of selenium metab- olism leading to the immediate selenium precursor and the mechanism of selenium incorporation into GSH-Px. G~uTATH~oNE- S-TRANsFERAsE In 1976 it was discovered that selenium-deficient rat liver had a non-sele- nium-dependent GSH-Px activity (Lawrence and Burk, 1976; Prohaska and Ganther, 1976~. This activity is catalyzed by one or more of the GSH- S-transferases (EC 2.5.1.18) (Prohaska and Ganther, 1977~. These en- zymes have no catalytic ability to destroy H2O2 but will reduce cumene hydroperoxide and t-butyl hydroperoxide. This peroxidase activity shows zero-order kinetics with respect to GSH, in contrast to GSH-Px kinetics (Prohaska and Ganther, 1977; Lawrence et al., 1978~. Sephadex G-150 chromatography separates these two enzymes; the molecular weight of the rat liver GSH-S-transferases is 39,000. After chromatography to tissue preparations of rat erythrocytes, skin, skeletal muscle, spleen, heart, lung, thymus, and intestine, Lawrence and Burk (1978) reported no detectable peroxidase activity due to the presence of GSH-S-transferase. Liver and adrenals had the highest activity. Approximately 90 percent of the total activity determined with cumene hydroperoxide was accounted for by
48 SELENIUM IN NUTRITION OSH-~-transferase in the testis of selenium-adequate rats, but only 35 per- cent of the total in liver, kidney, and adrenal. GSH-S-transferase ac- counted for 43 percent of the total activity determined with cumene hydro- peroxide in hamster liver; 70 to 85 percent of the activity in livers from pigs, sheep, chickens and humans; and 100 percent of the activity in guinea pig liver. This does not demonstrate, however, that guinea pigs do not have selenium-dependent GSH-Px or a selenium requirement, because preliminary data indicate that GSH-Px can account for at least 15 percent of the total activity in guinea pig liver and 100 percent in erythrocytes (Suede and Hoekstra, unpublished). FUNCTION OF GLUTATHIONE PEROXIDASE The discovery that hydroperoxides were substrates for GSH-Px (Little and O'Brien, 1968) provided an important clue to the biochemical function of GSH-Px, and thus of selenium. The erythrocyte possesses both catalase and GSH-Px activity. From kinetic data and GSH levels present in the erythrocyte, Flohe et al. (1972) hay" calculated that the rate of H2O2 re- duction per heme or per selenium, respectively, is nearly identical for these two enzymes in the erythrocyte. Catalase would therefore seem to be far more important than GSH-Px for H2O2 destruction because of the higher concentrations of catalase in the red cell. However, GSH-Px-deficient eryth- rocytes are susceptible to hemolysis when exposed to oxidizing agents, indi- cating that the ability to reduce hydroperoxides is of critical importance in the erythrocyte. Except in degenerate cells like the mammalian erythrocyte, catalase and GSH-Px are often localized in distinct compartments (catalase in the peroxisomes and GSH-Px in the cytosol and mitochondrial matrix space), such that there is little direct overlap in the competition for H2O2 (Plohe et al., 19761.~In human and guinea pig leukocytes, catalase activity is low and GSH-Px activity is high (Higgins et al., 1978), further demonstrating that these two protective enzymes are generally not in direct competition for H2o2. Decreases in tissue GSH-Px activity and the development of selenium- deficiency signs in animals are well correlated. In weanling rats fed a sele- nium-deficient diet, liver GSH-Px activity falls to undetectable levels at about the time liver necrosis develops (Hafeman et al., 19741. In chicks, depressed plasma GSH-Px activity and the development of exudative diathesis are well correlated (Noguchi et al., 1973; Cantor et al., 1975b). These diseases are prevented either by dietary selenium or vitamin E, sug- gesting that selenium and vitamin E have overlapping roles in the protec- tion of cells.
Biochemical Functions 49 Liver perfusion experiments have helped to substantiate the role of GSH-Px in protecting the liver against peroxidation. Isolated perfused rat liver was shown to destroy H2O2 or organic hydroperoxides added to the perfusate and to release GSSG into the perfusate (Sies et al., 1972; Sies and Summer, 1975~. Burk et al. (1978) demonstrated that GSSG was not released when livers from selenium-deficient rats were perfused with H2O2, indicating that GSH-Px was the catalytic source of the GSSG. Un- der rather unphysiological conditions (i.e., perfusion of liver with high lev- els of organic hydroperoxide), GSH-S-transferase destroyed the organic hydroperoxides added to the perfusate and released GSSG into the perfus- ate. This indicates that the peroxidase activity of GSH-S-transferase may have some physiological significance. Chance et al. (1978) have shown that exposure of perfused liver to hyperbaric oxygen will cause the release of GSSG into the perfusate; GSSG release is especially increased in the vita- min E-deficient liver. These perfusion experiments demonstrate that GSH-Px, vitamin E, and possibly GSH-S-transferase can function in the cell to protect against peroxidation. OTHER FUNCTIONS OF SELENIUM Selenium may have other biochemical functions in higher animals that are not a result of the ability of GSH-Px to serve as a biological antioxidant. It is unlikely that GSH-Px acts directly on fatty acid hydroperoxides in lipid membranes, but it may function by catalyzing the destruction of cytosolic hydrogen peroxide (McCay et al., 19811. A mammalian selenium-binding protein clearly different from GSH-Px was reported to be present in sele- nium-adequate lambs but absent in lambs suffering from nutritional mus- cular dystrophy (Pedersen et al., 19721. This 10,000-dalton protein has proved difficult to purify and characterize (Whanger et al., 1973; Black et al., 1978) and so the claim that it is a selenoprotein must remain tenta- tive until the protein is characterized and the selenium stoichiometry determined. A selenium-binding protein of 15,000 to 20,000 daltons has been ob- served in bovine and rat spermatozoa. Calvin (1978) reported a 17,000- dalton selenium-binding protein located in the midpiece of rat sperm. Pal- lini and Bacci (1979) found a 20,000-dalton selenium-binding protein in bovine sperm mitochondria, and McConnell et al. (1979b) have reported a 15,000-dalton selenium-binding protein from rat testis cytosol. Spermato- zoa from selenium-deficient rats have been reported to show decreased mobility and increased midpiece breakage (Brown and Burk, 1973; Wu et al., 1973, 19791. Thus, spermatozoa may possess a specific selenoprotein that serves as a mitochondrial structural protein or as an enzyme, although
50 SELENIUM IN NUTRITION the possibility that this protein is a GSH-Px subunit or its derivative has not been eliminated. It would seem unlikely that only one or two selenoproteins arising from selenite administration are present in animals, yet GSH-Px appears to be the only substantial 75Se-labeled peak present in chromatograms of cyto- sol of liver and most other tissues, except at early time points, following (75 Se~selenite administration. Several other 75Se-binding proteins have been observed in various tissues, but have not been characterized (Chen et al., 1975; Prohaska et al., 1977a; Herrman, 1977; Gasiewicz and Smith. 19781. Additional selenoproteins may be present in the particulate frac- tions as well as the soluble fractions of animal tissues. Burk et al. (1974) reported that selenium-deficient rats accumulate three times as much injected 75Se in the microsomal fraction of rat liver as do rats supplemented with 0.5 ppm dietary selenium. Burk and Correia (1978) further demonstrated that with selenium deficiency, the rate of heme degra- dation is increased following phenobarbital injection, suggesting that sele- nium may have a role in the regulation of heme catabolism. These reports suggest that additional biochemical roles for selenium as a component of membrane-bound proteins may be discovered. Hoffman and McConnell (1974) and Chen and Stadtman (1980) have reported that selenium can be incorporated into the purine and pyrimidine bases of RNA in bacteria grown in presence of selenite. Hawkes et al. (1979) have reported 75Se-labeled, acylated tRNAs isolated from rat liver, which may contain selenium in the RNA portion of the molecule. Thus, selenium may have a specific role in rare purine or pyrimidine bases. Selenium deficiency, in combination with vitamin E deficiency, has been shown to decrease the ability of ducks to resist infection (Yarrington et al., 19731. The microbicidal activity, but not phagocytizing ability, of neutro- phils from selenium-deficient rats and cattle is also impaired (Serfass and Ganther, 1975, 1976; Boyne and Arthur, 19781. Neutrophil GSH-Px activ- ity was depressed in selenium-deficient animals, so the effect of selenium deficiency on the immune response may or may not involve GSH-Px. Glutathione peroxidase may have a specific role in prostaglandin syn- thesis (Nugteren and Hazelhof, 1973; Van Dorp, 1975~; but prostaglandin endoperoxide synthetases also have peroxidase activity, so GSH-Px may not be essential for prostaglandin metabolism (Christ-Hazelhof and Nug- teren, 19781. Bryant and Bailey (1980) have noted altered metabolism of arachidonic acid via the lipoxygenase pathway in platelets from selenium- deficient rats and have suggested that this may be the first example of a specific function for selenium as a required component in the normal me- tabolism of an essential fatty acid. Burk et al. (1980) recently reported that the toxicity of diquat, a herbi
Biochemical Functions 51 cide similar to paraquat, was reduced, but not eliminated, in selenium- deficient rats by injection of a physiological dose of selenite 6 to 10 hours before diquat administration. These workers suggested that this protec- tion was due to a biochemical function of selenium other than that of GSH- Px because tissue GSH-Px activities were not significantly increased 10 hours after the selenium injection. This report offers strong evidence for an additional role of selenium in animals, although the effect could possibly be mediated by localized increases in GSH-Px activity not detectable in whole tissue homogenates. NUTRITIONAL AND METABOLIC INTERRELATIONSHIPS With the first demonstration of the essentiality of selenium that selenium was the integral part of Factor 3 and prevented liver necrosis in rats it was clear that the biochemistry of selenium is interrelated with other nutri- tional factors. Of these other factors, the biochemical function of vitamin E seems most complementary with that of selenium. As discussed more completely in other parts of this report, selenium and vitamin E deficien- cies in animals cause degenerative lesions; the nature and tissue location depends on the species and the status of other nutritional factors. For ex- ample, in rats a combined deficiency of vitamin E and selenium results in liver necrosis (Schwarz and Foltz, 19571. Vitamin E deficiency alone causes fetal death and resorption (Evans and Bishop, 1922), and selenium deficiency alone results in poor growth and failure to reproduce in rat pups born to selenium-deficient dams and raised on a selenium-deficient diet (McCoy and Weswig, 19691. Chicks develop muscular dystrophy, enceph- alomalacia, exudative diathesis, or pancreatic degeneration, depending on the presence or absence of vitamin E, selenium, sulfur amino acids, and excess dietary unsaturated fatty acids (Scott, 19781. The effects of vitamin E and of selenium deficiency have been postulated to result from the destruction of cellular membranes or of critical cellular proteins and thus of cellular integrity. Addition of polyunsaturated fatty acids to the diet tends to exacerbate these deficiency defects, whereas syn- thetic antioxidants in many cases will alleviate the signs of vitamin E and selenium deficiency. Tappel (1962) suggested a basis for these observations by postulating that the biochemical role of vitamin E is as a lipid antioxi- dant. Selenium was also classified as an antioxidant because of its ability to prevent a number of vitamin E deficiency diseases. Green and Bunyan (1969) attacked this hypothesis because they felt there was no direct evi- dence showing that lipid peroxidation occurred in vivo, that vitamin E was chemically involved in the protection against lipid peroxidation, or that selenium had antioxidant properties in vivo.
52 SELENIUM IN NUTRITION In the 1970s, however, several new discoveries strongly suggested that activated-oxygen and free-radical attack of cellular components occurs, and that generation of hydroperoxides occurs in vivo. As discussed above, GSH-Px was shown to contain selenium, demonstrating that selenium does have antioxidant properties in animals. Superoxide dismutase was identified as a scavenger of superoxide (McCord and Fridovich, 19691; to- gether, superoxide dismutase and GSH-Px may prevent the reaction of su- peroxide with hydrogen peroxide to form hydroxy radical (Fridovich, 1975~. Ethane and pentane evolution in the breath of rats due to the perox- idative breakdown of unsaturated fatty acids also demonstrated that lipid peroxidation does occur in vivo and that such lipid peroxidation is mini- mized by vitamin E and dietary selenium (Dillard et al., 1977; Hafeman and Hoekstra, 1977a,b). Thus selenium, as a component of GSH-Px, and vita- min E both can serve as biological antioxidants; these findings do not elimi- nate the possibility of other biochemical roles for either of these nutrients. Hydrogen peroxide, hydroperoxides, superoxide, various radicals in- cluding hydroxy radical, and possibly singlet oxygen are formed as prod- ucts of necessary reactions in cells or by further chemical reaction of these products, and so protective systems have evolved to contain and ultimately destroy these reactive species before they damage cells. These protective systems are compartmentalized and thus complement one another. As a lipid-soluble antioxidant, vitamin E scavenges free radicals and possibly singlet oxygen (McCay et al., 1978) before they can attack cellular and intracellular membranes. Glutathione peroxidase destroys H2O2 and hy- droperoxides in the cytosol and mitochondrial matrix space. Catalase de- grades H202 in the peroxisome. Superoxide dismutase detoxifies superox- ide in the cytosol and mitochondria before superoxide can react with H2O2 to form hydroxy radical. In deficiency diseases that can be prevented by either vitamin E or sele- nium, such as liver necrosis in the rat, the cells seemingly have vitamin E and GSH-Px organized in a serial arrangement; the origin of the prooxi- dant species is presumably in the soluble portion of the cell, but the molec- ular target is in the membrane. If GSH-Px does not destroy the peroxides, then vitamin E can still protect the membrane. Although lipid hydroperox- ides are excellent substrates for GSH-Px, this soluble enzyme has been re- ported not to reduce lipid hydroperoxides within membranes to the corre- sponding alcohols in vitro. Instead, GSH-Px may protect the cell by de- stroying hydrogen peroxide and thus preventing the formation of hydroxy radical (McCay et al., 1976~. In deficiency diseases that can be prevented by dietary selenium but not by vitamin E, such as pancreatic degeneration in the chick (Cantor et al., 1975a), aqueous prooxidant species are seemingly formed that can then
Biochemical Functions 53 damage critical soluble proteins without initial destruction of the mem- brane. Analogous to the relationship between selenium and vitamin E defi- ciency and pancreatic degeneration, the toxicity of paraquat and nitro- furantoin is reported to be increased by selenium deficiency but not by vitamin E deficiency in the chick (Combs and Peterson, 1979, 1980~. These water-soluble herbicides most likely catalyze the production of superoxide, which could lead to hydroxy radical formation if GSH-Px, as well as super- oxide dismutase, is not present. Apparently in the chick the target of the toxic species, presumably hydroxy radical, is in the cytosol and thus is not protected by lipid-soluble vitamin E. In the rat, however, both selenium and vitamin E deficiencies are reported to increase paraquat toxicity (Bus et al., 1975), suggesting a difference in the relative roles of vitamin E and GSH-Px in these two species. In deficiency diseases that can be prevented by vitamin E but not by selenium, such as encephalomalacia in the chick, presumably both the ori- gin and the target of the activated species lie within the hydrophobic re- gions of the cell. The different abilities of various antioxidants to replace vitamin E and prevent deficiency diseases has been related to their differ- ent abilities to quench these species and to their differential solubilities in hydrophobic regions of tissues such as membrane and adipose tissue. Nutritional muscular dystrophy in lambs can often be prevented by sele- nium supplementation alone (Mush et al., 1958), although dietary unsatu- rated fatty acid supplementation will produce a nutritional muscular dys- trophy that is not prevented by selenium (Blaxter, 1962; Whanger et al., 19771. A high-polyunsaturated-fat diet also accelerates the development of encephalomalacia in vitamin E-deficient chicks. Apparently the increased unsaturated fatty acid content in the diet causes an increased susceptibility of the membrane to peroxidation that cannot be prevented by soluble GSH-Px. The addition of sulfur amino acids to the diet delays the onset of liver necrosis in rats fed selenium- and vitamin E-deficient diets. This effect is not due solely to contamination of the sulfur amino acids by selenium (Schwarz, 1965~. The torula yeast-based diets commonly used in such ex- periments are limiting in the sulfur amino acids; the addition of methionine to the diet has been shown to increase the hepatic GSH concentration (Seligson and Rotruck, 1979~. A higher GSH level increases the apparent Vmax of GSH-Px (Flohe et al., 1972) and thus may explain the ability of methionine to delaythe onset of liver necrosis. Alternatively, supplemental sulfur amino acids may improve the availability of selenomethionine present in the torula yeast (Suede et al., 19811. Schwarz (1965) reported that the addition of various methyl donors such as choline or betaine to the diet exacerbates the development of liver necrosis and may be due to a decreased degradation of
54 SELENIUM IN NUTRITION selenomethionine as well as of methionine. These results suggest that ade- quate dietary methionine, or dietary sulfur in ruminants, is important for maximal selenium biopotency in livestock because a major form of selenium in most feedstuffs of plant origin is selenomethionine (Allaway et al., 1967; Olson et al., 1970~. Additionally, selenomethionine present in general body proteins may be an important source of selenium as these proteins turn over. Tissue GSH-Px activity can be induced or repressed by various factors. Increased tissue GSH-Px activity is observed under conditions of increased oxidant stress, such as exposure to ozone or the inclusion of autooxidized lipids in the diet (Chow and Tappel, 1974; Reddy and Tappel, 19741. Revis et al. (1979) reported increased GSH-Px activity in the muscles from ge- netically dystrophic mice as compared to control mice, suggesting either GSH-Px induction or a selective advantage of higher GSH-Px in these ge- netically dystrophic mice. Excess dietary vitamin E has been shown to de- crease tissue GSH-Px in rats (Yang and Desai, 1978), further suggesting that a relative decrease in oxidant stress will lower the need for GSH-Px. The deficiency of several required nutrients not previously mentioned- iron deficiency in humans and rabbits (Macdougall, 1972; Rodvien et al., 1974), riboflavin deficiency in pigs (Brady et al., 1979), vitamin B6 defi- ciency in rats (Yasumoto et al., 1979), and copper deficiency in rats (Jenkinson et al., 1980) have been reported to decrease tissue GSH-Px activities, but the mechanisms are not understood. Several toxic agents have been shown to decrease tissue GSH-Px activ- ity. Silver acetate decreased liver GSH-Px activity in rats supplemented with 0.5 ppm selenite (Wagner et al., 1975), and 0.2 percent tri-o-cresyl phosphate reduced liver and erythrocyte GSH-Px activity by altering sele- nium metabolism (Swanson, 1975~. Silver may precipitate selenide and thus make it unavailable for GSH-Px synthesis; tri-o-cresyl phosphate de- creases tissue selenium levels, apparently by increasing selenium excre- tion. Doxorubicin, an anticancer drug that is very cardiotoxic, has been reported to reduce heart GSH-Px activity within 4 hours after injection into rabbits. It is unclear whether this reduction is due to specific antagonism of selenium or GSH-Px metabolism, or to general toxicity to the heart (Re- vis and Marusic, 1978; Doroshaw et al., 19801. Selenium has been shown to reduce the toxicity of cadmium, inorganic and methyl mercury, thallium, and silver. Selenium apparently decreases the rate of excretion of these toxic substances and changes the distribution of these elements within the body (Parizek et al., 19741. Recently Ga- siewicz and Smith (1978) identified a specific protein in plasma that binds both cadmium and selenium; either element alone will not result in the formation of this relatively stable complex, suggesting that this complex is
Biochemical Functions 55 Substrates , Cell-Damaging _ , Detoxified + O2 Products Products O2 Oxidases + ~ RH2 O2 + O ~_ , ~_ Cata I ase l SO D + Unsaturated Fatty Acids Li pid Peroxides I Vitamin E SOD = Superoxide Dismutase GSH-Px = Glutathione Peroxidase GSH-T = Glutathione S-Transferase GSH GSH-Px GSH-T GS H-Px - , GSH HOH _ HOH Sulfur | Amino Beta Acids Oxidation FIGURE 5 Interrelationships of selenium, vitamin E, and sulfur amino acids. In oxidative metabolism, highly reactive forms of oxygen are produced, such as superoxide ion, hydroxy radical, and hydrogen peroxide (upper box). Catalase and glutathione peroxidase (a sele- noenzyme) decompose hydrogen peroxide. Unsaturated fatty acids react with oxygen to form lipid hydroperoxides. This process is stimulated by free radicals (curved arrows) and inhib- ited by radical scavengers such as vitamin E. Reduction of organic peroxides by glutathione is catalyzed by glutathione peroxidase and also by various GSH transferases (sometimes called GSH peroxidase II). Glutathione peroxidase helps prevent the formation of lipid peroxides (by destroying hydrogen peroxide) and may also help eliminate those peroxides that are formed. (Prepared by H. E. Ganther, University of Wisconsin, Madison.) a possible biochemical mechanism for the decreased toxicity and meta- bolic changes observed within the body when selenium and cadmium are administered concurrently. Ganther et al. (1972) demonstrated that sele- nium also protects against methylmercury toxicity and suggested that the presence of selenium may lessen the toxicity of the mercury in tuna. The observation that vitamin E and the synthetic lipid soluble antioxidant N,N'-diphenyl-p-phenylene-diamine (DPPD) protected against methyl- mercury poisoning in rats (Welsh, 1979) led to the suggestion that methyl- mercury may exert part of its toxic action by generating free radicals (Gan
56 SELENIUM IN NUTRITION ther, 1980~. Reports that lead (Bell et al., 1978) and copper (Godwin et al., 1978) are more toxic to selenium-deficient animals and that copper pre- treatment will decrease the toxicity of selenium (Stowe and Brady, 1978; Stowe, 1980) suggest that the metabolism of several other elements is inter- related with that of selenium. Characterization of the biochemical interac- tions of these elements awaits further experimentation.
