A comparison of cardiac glutathione S-transferases from wild and domestic animals

1. Cardiac glutathione S-transferases from wild animals; hyena, red fox, porcupine, coypu and mountain gazelle were purified and compared with the enzymes from domestic animals; cow, camel, goat and sheep. 2. By using 1-chloro-2,4-dinitrobenzene as a substrate, domestic hearts expressed higher glutathione conjugating activity than wild animals hearts. 3. In all the studied hearts, the bulk of the activity was associated with near neutral and acidic glutathione S-transferase isozymes with pI values ranging from 4 to 7.4. 4. The enzymes from domestic animals displayed homodimeric structure of 25,000 mol. wt subunit while of the wild animals both hyena and coypu displayed homodimers of 26,500 mol. wt subunit and the rest exhibited heterodimers of 25,000 and 28,000 mol. wt subunits.     

Glutathione S-transferases in human prostate

A number of human prostatic tissue biopsies have been analyzed for glutathione S-transferase activity, using 1-chloro-2,4-dinitrobenzene (CDNB) as a substrate. Samples from nine patients (age range 61-90) with benign prostatic hypertrophy who had received no prior chemotherapy had a mean glutathione S-transferase activity of 137 +/- 44 nmol/min per mg with a range of 97-237. A qualitative comparison of the glutathione S-transferase of normal prostate and benign prostatic hypertrophy samples was carried out. Approximately 260-fold purification was achieved using glutathione-Sepharose affinity chromatography, with glutathione S-transferase accounting for approximately 0.19-0.33% of the total protein. Substrate specificity determinations suggested similar, but not identical, glutathione S-transferase subunits in normal prostate and benign prostatic hypertrophy. One- and two-dimensional electrophoresis (isoelectric focusing and 12.5% SDS-polyacrylamide gel electrophoresis) identified at least seven stained polypeptides in the purified glutathione S-transferase preparations. These ranged in Mr from approximately 24,000 to 28,500 and in pI from near neutral to basic. Western blot analysis using polyclonal antibodies raised against rat liver glutathione S-transferase suggested crossreactivity with five of the human isoenzymes in both normal prostate and benign prostatic hypertrophy. One of the glutathione S-transferases, present in both normal prostate and benign prostatic hypertrophy, had an Mr of approx. 24,000 and a near-neutral pI and crossreacted immunologically with a polyclonal antibody raised against human placental glutathione S-transferase (Yf, subunit 7 or pi). These data suggest that four glutathione S-transferases are expressed in human prostate, with subunits from each of the major classes alpha, mu and pi. These are characterized as Ya, Yb, Yb' and Yf (analogous alternative nomenclature subunits 1, 3, 4 and 7).     

Rat lung glutathione S-transferases. Evidence for two distinct types of 22000-Mr subunits.

Two immunologically distinct types of 22000-Mr subunits are present in rat lung glutathione S-transferases. One of these subunits is probably similar to Ya subunits of rat liver glutathione S-transferases, whereas the other subunit Ya' is immunologically distinct. Glutathione S-transferase II (pI7.2) of rat lung is a heterodimer (YaYa') of these subunits, and glutathione S-transferase VI (pI4.8) of rat lung is a homodimer of Ya' subunits. On hybridization in vitro of the subunits of glutathione S-transferase II of rat lung three active dimers having pI values 9.4, 7.2 and 4.8 are obtained. Immunological properties and substrate specificities indicate that the hybridized enzymes having pI7.2 and 4.8 correspond to glutathione S-transferases II and VI of rat lung respectively.     

Cholic acid binding by glutathione S-transferases from rat liver cytosol.

Cholic acid-binding activity in cytosol from rat livers appears to be mainly associated with enzymes having glutathione S-transferase activity; at least four of the enzymes in this group can bind the bile acid. Examination of the subunit compositions of different glutathione S-transferases indicated that cholic acid binding and the ability to conjugate reduced glutathione with 1,2-dichloro-4-nitrobenzene may be ascribed to different subunits.     

Identification of a basic hybrid glutathione S-transferase from human liver. Glutathione S-transferase delta is composed of two distinct subunits (B1 and B2).

The purification of a hybrid glutathione S-transferase (B1 B2) from human liver is described. This enzyme has an isoelectric point of 8.75 and the B1 and B2 subunits are distinguishable immunologically and are ionically distinct. Hybridization experiments demonstrated that B1 B1 and B2 B2 could be resolved by CM-cellulose chromatography and have pI values of 8.9 and 8.4 respectively. Transferase B1 B2, and the two homodimers from which it is formed, are electrophoretically and immunochemically distinct from the neutral enzyme (transferase mu) and two acidic enzymes (transferases rho and lambda). Sodium dodecyl sulphate/polyacrylamide-gel electrophoresis demonstrated that B1 and B2 both have an Mr of 26 000, whereas, in contrast, transferase mu comprises subunits of Mr 27 000 and transferases rho and lambda both comprise subunits of Mr 24 500. Antisera raised against B1 or B2 monomers did not cross-react with the neutral or acidic glutathione S-transferases. The identity of transferase B1 B2 with glutathione S-transferase delta prepared by the method of Kamisaka, Habig, Ketley, Arias & Jakoby [(1975) Eur. J. Biochem. 60, 153-161] has been demonstrated, as well as its relationship to other previously described transferases.     

The development of arylsulphatase in the small intestine of the rat

1. Arylsulphatase activity was measured in stomach, proximal and distal third of small intestine, colon, liver and kidney of foetal and neonatal Sprague-Dawley rats and Swiss mice, with nitrocatechol sulphate as substrate. 2. The specific activity in the distal small intestine, but not in the stomach, proximal small intestine or colon, increased about fourfold between 5 and 16 days after birth in both conventional and germ-free rats. 3. No comparable increase occurred in the distal small intestine of the mouse. 4. The specific activity of acid phosphatase in the distal small intestine of the rat rose only slightly when the arylsulphatase activity increased. 5. The pH optimum and Michaelis constant of arylsulphatase activity of the distal small intestine were similar for 1-day-old, 9-day-old and adult rats. 6. When extracts of distal small intestine of 1-day-old and 9-day-old rats were incubated together, the arylsulphatase activities were additive.     

The influence of diet on the regional distribution of glutathione S-transferase activity in channel catfish intestine

There is evidence that glutathione conjugates are the major metabolites formed following systemic uptake of carcinogenic contaminants from the intestine. The effect of commercial diet versus a semi-purified diet on the distribution of glutathione S-transferase (GST) activity was examined in proximal, medial, and distal sections of catfish intestine. The bulk of GST activity with 1-chloro-2,4-dinitrobenzene, ethacrynic acid, and 3H-benzo[a]pyrene-4,5-oxide, and the percent cytosolic protein cross-reacting with anti-catfish GST-pi were in the more proximal segments and dropped off distally in the two diet groups. However, the total GST-pi cross-reacting protein in the proximal section was significantly higher in fish fed a chow diet. Western blot analysis revealed pi-class GST to be expressed principally in the proximal intestine. Cytosol samples cross-reacted with antibodies to human GST-alpha, -mu, and -pi, but not -theta, classes. Alpha-like GST isoforms of MW 26,200 and 24,600, absent in sections from fish fed a purified diet, were differentially expressed only in the distal section of chow-fed fish. These results indicate that diet significantly elicits regional differences in GST protein levels, that components of the commercial chow affect GST protein expression in the distal intestine, and that maintenance diet should be taken into consideration during dietary exposure studies.     

Nonhistone protein BA is a glutathione S-transferase localized to interchromatinic regions of the cell nucleus

A DNA-binding nonhistone protein, protein BA, was previously demonstrated to co-localize with U-snRNPs within discrete nuclear domains (Bennett, F. C., and L. C. Yeoman, 1985, Exp. Cell Res., 157:379-386). To further define the association of protein BA and U-snRNPs within these discrete nuclear domains, cells were fractionated in situ and the localization of the antigens determined by double-labeled immunofluorescence. Protein BA was extracted from the nucleus with the 2.0 M NaCl soluble chromatin fraction, while U-snRNPs were only partially extracted from the 2.0 M NaCl-resistant nuclear structures. U-snRNPs were extracted from the residual nuclear material by combined DNase I/RNase A digestions. Using an indirect immunoperoxidase technique and electron microscopy, protein BA was localized to interchromatinic regions of the cell nucleus. Protein BA was noted to share a number of chemical and physical properties with a family of cytoplasmic enzymes, the glutathione S-transferases. Comparison of the published amino acid composition of protein BA and glutathione S-transferases showed marked similarities. Nonhistone protein BA isolated from saline-EDTA nuclear extracts exhibited glutathione S-transferase activity with a variety of substrates. Substrate specificity and subunit analysis by SDS polyacrylamide gel electrophoresis revealed that it was a mixture of several glutathione S-transferase isoenzymes. Protein BA isolated from rat liver chromatin was shown by immunoblotting and peptide mapping techniques to be two glutathione S-transferase isoenzymes composed of the Yb and Yb' subunits. Glutathione S-transferase Yb subunits were demonstrated to be both nuclear and cytoplasmic proteins by indirect immunolocalization on rat liver cryosections. The identification of protein BA as glutathione S-transferase suggests that this family of multifunctional enzymes may play an important role in those nuclear domains containing U-snRNPs.     

Ribosome-inactivating proteins from the seeds of Saponaria officinalis L. (soapwort), of Agrostemma githago L. (corn cockle) and of Asparagus officinalis L. (asparagus), and from the latex of Hura crepitans L. (sandbox tree).

A hitherto unknown cytosolic glutathione S-transferase from rat liver was discovered and a method developed for its purification to apparent homogeneity. This enzyme had several properties that distinguished it from other glutathione S-transferases, and it was named glutathione S-transferase X. The purification procedure involved DEAE-cellulose chromatography, (NH4)2SO4 precipitation, affinity chromatography on Sepharose 4B to which glutathione was coupled and CM-cellulose chromatography, and allowed the isolation of glutathione S-transferases X, A, B and C in relatively large quantities suitable for the investigation of the toxicological role of these enzymes. Like glutathione S-transferase M, but unlike glutathione S-transferases AA, A, B, C, D and E, glutathione S-transferase X was retained on DEAE-cellulose. The end product, which was purified from rat liver 20 000 g supernatant about 50-fold, as determined with 1-chloro-2,4-dinitrobenzene as substrate and about 90-fold with the 1,2-dichloro-4-nitrobenzene as substrate, was judged to be homogeneous by several criteria, including sodium dodecyl sulphate/polyacrylamide-gel electrophoresis, isoelectric focusing and immunoelectrophoresis. Results from sodium dodecyl sulphate/polyacrylamide-gel electrophoresis and gel filtration indicated that transferase X was a dimer with Mr about 45 000 composed of subunits with Mr 23 500. The isoelectric point of glutathione S-transferase X was 6.9, which is different from those of most of the other glutathione S-transferases (AA, A, B and C). The amino acid composition of transferase X was similar to that of transferase C. Immunoelectrophoresis of glutathione S-transferases A, C and X and precipitation of various combinations of these antigens by antisera raised against glutathione S-transferase X or C revealed that the glutathione S-transferases A, C and X have different electrophoretic mobilities, and indicated that transferase X is immunologically similar to transferase C, less similar to transferase A and not cross-reactive to transferases B and E. In contrast with transferases B and AA, glutathione S-transferase X did not bind cholic acid, which, together with the determination of the Mr, shows that it does not possess subunits Ya or Yc. Glutathione S-transferase X did not catalyse the reaction of menaphthyl sulphate with glutathione, and was in this respect dissimilar to glutathione S-transferase M; however, it conjugated 1,2-dichloro-4-nitrobenzene very rapidly, in contrast with transferases AA, B, D and E, which were nearly inactive towards that substrate.(ABSTRACT TRUNCATED AT 400 WORDS)     

The elusive roles of bacterial glutathione<Emphasis Type="Italic"> S-</Emphasis>transferases: new lessons from genomes

Glutathione S-transferases constitute a large family of enzymes which catalyze the addition of glutathione to endogenous or xenobiotic, often toxic electrophilic chemicals. Eukaryotic glutathione S-transferases usually promote the inactivation, degradation or excretion of a wide range of compounds by formation of the corresponding glutathione conjugates. In bacteria, by contrast, the few glutathione S-transferases for which substrates are known, such as dichloromethane dehalogenase, 1,2-dichloroepoxyethane epoxidase and tetrachlorohydroquinone reductase, are catabolic enzymes with an essential role for growth on recalcitrant chemicals. Glutathione S-transferase genes have also been found in bacterial operons and gene clusters involved in the degradation of aromatic compounds. Information from bacterial genome sequencing projects now suggests that glutathione S-transferases are present in large numbers in proteobacteria. In particular, the genomes of three Pseudomonas species each include at least ten different glutathione S-transferase genes. Several of the corresponding proteins define new classes of the glutathione S-transferase family and may also have novel functions that remain to be elucidated.     

Irreversible inhibition of hepatic glutathione S-transferase by ciprofibrate, a peroxisome proliferator

Ciprofibrate (2-[4-(2,2-dichlorocyclopropyl) phenoxy]2-methyl propionic acid) which is a hypolipidemic agent and has been shown to cause peroxisome proliferation, non-competitively inhibits glutathione S-transferase activity of rat liver, both in vivo and in vitro. Among all the glutathione S-transferases of rat liver, ligandin is maximally inhibited by ciprofibrate. Studies with the purified glutathione S-transferases of rat liver indicate that the affinities of different subunits of liver enzymes for ciprofibrate are in the order Ya greater than Yb, Yb' greater than Yc.     

Species difference in intestinal drug metabolising enzymes in mouse, rat and hamster and their inducibility by masheri, a pyrolysed tobacco product

Activities of several drug metabolising enzymes in the small intestine were investigated in Swiss mice, Sprague Dawley rats and Syrian Golden Hamsters fed 10% masheri, a pyrolysed tobacco product, in diet, for 20 months. The basal levels of enzymes in proximal (PI), medium (MI) and distal (DI) parts of the intestine in the three species were similar. However, the levels of cytochrome P-450, benzo(a) pyrene hydroxylase (B(a)OH) and glutathione S-transferase (GST) were highest in hamsters followed by rat and mice. Upon treatment with masheri, significant induction of cytochrome P-450 and B(a)PH was observed in PI and DI of all the three species. However, GSH and GST was depleted upon masheri treatment in all the three species again only in proximal and distal parts of the intestine. Thus increase in activating enzymes together with depletion in GSH-GST system upon exposure could be an important factor in the susceptibility of the small intestine to hazardous xenobiotic exposure.     

Purification and characterization of the individual glutathione S-transferases from sheep liver

The glutathione S-transferases (EC 2.5.1.18) have been purified to electrophoretic homogeneity from 105,000g supernatant of sheep liver homogenate by employing a combination of gel filtration on Sephadex G-150 and affinity chromatography on S-hexylglutathione-linked Sepharose-6B columns. Approximately 70% of the original glutathione S-transferase activity toward 1-chloro-2,4-dinitrobenzene and glutathione peroxidase activity toward cumene hydroperoxide could be recovered by this purification method. Of particular importance in developing this procedure was the fact that the enzyme preparation obtained after affinity column chromatography represented all the isozymes of sheep liver glutathione S-transferases. Further purification by CM-cellulose and DEAE-cellulose column chromatography resolved the glutathione S-transferases into seven distinct cationic isozymes designated C-1, C-2, C-3, C-4, C-5, C-6, and C-7 and five overlapping anionic transferases designated A-1, A-2, A-3, A-4, and A-5, respectively, in the order of their elution from the ion-exchange columns. The sodium dodecyl sulfate SDS-gel electrophoretic data on subunit composition revealed that cationic enzymes are composed of two subunits with an identical Mr of 24,000 whereas a predominant subunit with Mr of 26,000 was observed in all anionic isozyme peaks except A-1. Cationic isozymes accounted for approximately 98% of the total peroxidase activity associated with the glutathione S-transferase whereas only A-1 of the anionic isozymes displayed some peroxidase activity. Isozyme C-4 was found to be the most abundant glutathione S-transferase in the sheep liver. Characterization of the individual transferases by their specificity toward a number of selected substrates, subunit composition, and isoelectric points showed some similarities to those patterns for human liver glutathione S-transferases.     

Hepatic glutathione S-transferases in rainbow trout and their interaction with 2,4-dichlorophenoxyacetic acid and 1,4-benzoquinone

Glutathione S-transferase in the cytosol of rainbow trout liver was partially purified by affinity chromatography on a column with glutathione coupled to epoxy-activated Sepharose 6B, which retained 94% of the total activity. Chromatofocussing on a Polybuffer exchanger 118 column separated the glutathione S-transferase into six major cationic isoenzymes (K1-K6), and some minor fractions. SDS-polyacrylamide slab gel electrophoresis showed K1-K3 to be heterodimers with subunits of Mr 25,000 and 26,500, and K4-K6 to be homodimers with subunits of Mr 25,000. The glutathione S-transferase isoenzymes were partially characterized by different biochemical parameters. The hepatic rainbow trout glutathione S-transferases were inhibited by the organic water pollutants, 1,4-benzoquinone and 2,4-dichlorophenoxyacetic acid. The same kinetic inhibition patterns were observed with these inhibitors as for rat liver glutathione S-transferases. It is concluded that rainbow trout glutathione S-transferases can play a key role in the detoxication of organic micropollutants in the aquatic environment.     

Isolation and characterization of the multiple glutathione S-transferases from human liver. Evidence for unique heme-binding sites

Thirteen forms of glutathione S-transferase were isolated from human liver in high yields by glutathione-affinity chromatography and chromatofocusing. Apparent isoelectric points ranged from 4.9 to 8.9 and included neutral forms. All 13 forms appeared to be identical immunochemically in a quantitative enzyme-linked immunosorbent assay. These forms were immunochemically distinct from the major acidic glutathione S-transferase found in placenta and erythrocyte and were immunochemically distinct from two forms of higher molecular weight glutathione S-transferase found in some but not all liver samples. The 13 forms exhibited similar activities with 1-chloro-2,4-dinitro-benzene as substrate, specific activities of 33-94 mumol/min/mg. Likewise, these forms all exhibited glutathione peroxidase activity with cumene hydroperoxide, specific activities of 1.5-8.3 mumol/min/mg. All 13 forms bound bilirubin with subsequent conformational changes leading to states devoid of transferase activity, a process prevented by the presence of foreign proteins. As hematin-binding proteins, however, these multiple transferases exhibited a very broad range of binding extending from nonbinding to high-affinity binding (KD approximately 10(-8) M). Hematin binding was noncompetitive with transferase activity and did not involve the bilirubin-binding site, suggesting the existence of unique heme-binding sites on these proteins. The two forms of the immunochemically distinct glutathione S-transferases transferases found in some liver samples also exhibited both transferase and peroxidase activities. In addition, they also have separate sites for binding bilirubin and hematin.     

Multiple glutathione S-transferase isoforms are present on male germ cell plasma membrane.

Phase II detoxification enzymes, the glutathione S-transferases (GSTs) of 24 kDa are known to be cytosolic enzymes. This study shows that multiple GST isoforms that are 24 kDa in size are present on the extracellular side of the plasma membrane of rat male germ cells. The GST activity of male germ cell plasma membranes is several folds higher than somatic cell plasma membrane GST activity. Isoform composition of the germ cell plasma membrane and the cytosolic pool differ, GSTM5 and GSTPi being absent on the plasma membranes. The molecular masses of the common isoforms are comparable between the two pools and both pools show GST and glutathione peroxidase activity.     

Expression of an enzymatically active Yb3 glutathione S-transferase in Escherichia coli and identification of its natural form in rat brain

Glutathione S-transferases containing Yb3 subunits are relatively uncommon forms that are expressed in a tissue-specific manner and have not been identified unequivocally or characterized. A cDNA clone containing the entire coding sequence of Yb3 glutathione S-transferase mRNA was incorporated into a pIN-III expression vector used to transform Escherichia coli. A fusion Yb3-protein containing 14 additional amino acid residues at its N terminus was purified to homogeneity. Recombinant Yb3 was enzymatically active with both 1-chloro-2,4-dinitrobenzene and 1,2-dichloro-4-nitrobenzene as substrates but lacked glutathione peroxidase activity. Substrate specificity patterns of recombinant Yb3 were more limited than those of glutathione S-transferase isoenzymes containing Yb1- or Yb2-type subunits. Peptides corresponding to unique amino acid sequences of Yb3 as well as a peptide from a region of homology with Yb1 and Yb2 subunits were synthesized. These synthetic peptides were used to raise antibodies specific to Yb3 and others that cross-reacted with all Yb forms. Immunoblotting was utilized to identify the natural counterpart of recombinant Yb3 among rat glutathione transferases. Brain and testis glutathione S-transferases were rich in Yb3 subunits, but very little was found in liver or kidney. Physical properties, substrate specificities, and binding patterns of the recombinant protein paralleled properties of the natural isoenzyme isolated from brain.     

Immunolocalization of antioxidant enzymes in adult hamster kidney

Summary Immunoperoxidase and immunogold techniques were used to localize the following antioxidant enzyme systems in the adult hamster kidney at the light and ultrastructural levels: superoxide dismutases, catalases, peroxidases and glutathione S-transferases. Each cell type in the kidney showed specific patterns of labelling of these enzymes. For example, proximal and distal tubular and transitional epithelial cells showed significant staining for all of these enzymes, while glomerular cells and cells of the thin loop of Henle did not show significant staining at the light microscope level. In addition, high levels of glutathione peroxidase were found in smooth muscle cells of renal arteries. At the ultrastructural level, each enzyme was found in a specific subcellular location. Manganese superoxide dismutase was found in mitochondria, catalase was localized in peroxisomes, while copper, zinc superoxide dismutase and glutathione S-transferase (liver and placental forms) were found in both the nucleus and cytoplasm. Glutathione peroxidase was found to have a broad intracellular distribution, with localization in mitochondria, peroxisomes, nucleus, and cytoplasm. Microvilli of tubular cells were labelled by antibodies to catalase, copper, zinc superoxide dismutase, glutathione peroxidase, and glutathione S-transferases. Cell types that were negative by light microscopy immunoperoxidase studies showed definite labelling with immunogold post-embedding ultrastructural techniques (glomerular cells and cells of the loop of Henle), demonstrating the greater sensitivity of the latter technique. These observations demonstrate that there are large variations in the levels of antioxidant enzymes in different cell types, and that even within a distinct cell type, the levels of these enzymes vary in different subcellular locations. Our results demonstrate for the first time the overall antioxidant enzyme status of individual kidney cell types, thereby explaining why different cell types have differing susceptibilities to oxidant stress. Possible physiological and pathological consequences of these findings are discussed.     

Protein and peptide degradation in the intestine of the common brushtail possum (Trichosurus vulpecula)

A protein (bovine serum albumin: BSA) and a peptide (luteinizing hormone releasing hormone: LHRH) were used to evaluate proteolytic activity in the intestine of common brushtail possums (Marsupiala, Trichosurus vulpecula). Luminal and mucosal extracts were isolated from the duodenum, jejunum, ileum, caecum, proximal colon and distal colon, their protein content assessed and specific activities in metabolising LHRH and BSA determined in vitro. The degradation of LHRH by luminal extracts was compared with that by the pancreatic enzymes, chymotrypsin, trypsin, and elastase. The protein concentration (microg x mg-1) of mucosal extract in the duodenum was higher ( P<0.05) than in the proximal colon, but that of luminal extracts did not differ significantly between regions. Proteolytic activity of luminal extracts was greater ( P<0.01) in the jejunum and ileum than in the hindgut. In the small intestine, proteolytic activity of luminal enzymes far exceeded that of mucosal enzymes ( P<0.05). All three pancreatic enzymes hydrolysed LHRH, but chymotrypsin had the greatest activity. This study has demonstrated that, in possums, proteolysis occurs primarily in the small intestine through luminal enzymes, with chymotrypsin playing a major role. The possum hindgut contributes little to the metabolism of peptides and proteins, identifying it as a potential site to target for their absorption following oral delivery.     

Tissue-specific expression of the rat glutathione S-transferases

Tissue-specific patterns of rat glutathione S-transferase expression have been demonstrated by in vitro translation of purified poly(A) RNAs and by protein purification. Poly(A) RNAs from six rat tissues including heart, kidney, liver, lung, spleen, and testis were used to program in vitro translation with the rabbit reticulocyte lysate system and [35S]methionine. The glutathione S-transferase subunits synthesized in vitro were purified from the translation products by affinity chromatography on S-hexylglutathione-linked Sepharose 6B columns. The affinity bound fractions were analyzed by Na dodecyl SO4-polyacrylamide gel electrophoresis and fluorography. A subunit of Mr = 22,000 detected in the in vitro translation products of poly(A) RNAs from heart, kidney, lung, spleen, and testis is missing from the translation products of liver poly(A) RNAs. This Mr = 22,000 subunit is present only in the anionic glutathione S-transferase fraction purified from rat heart, kidney, lung, spleen, and testis. Purified anionic glutathione S-transferase from rat liver does not contain this subunit. The relative specific activities toward a dozen different substrates also demonstrate the nonidentity between liver and kidney anionic glutathione S-transferases. In addition, among the glutathione S-transferase subunits expressed in the liver, some of them could not be detected in the other tissues investigated. Our results indicate that tissue-specific expression of rat glutathione S-transferases may occur pretranslationally.