Hepatocellular nuclei require glutathione, glutathione S-transferases (GSTs) and Se-dependent glutathione peroxidase (GPx) for intranuclear protection against damage from electrophiles or products of active oxygen. Data so far available from the literature on nuclei isolated in aqueous systems range from glutathione, GSTs and GPx either being absent altogether to being present in quantities in excess of those in the cytoplasm. This paper describes a small-scale preparation of a nuclear fraction from rat liver by a non-aqueous technique, designed to retain nuclear water-soluble molecules in situ, since low-molecular-mass compounds can diffuse freely into other compartments during aqueous separation. This non-aqueous procedure shows the nucleus to contain glutathione at 8.4 mM and soluble GSTs at 38 micrograms/mg of protein, the enrichment over the homogenate being 1.2-1.4-fold. Se-dependent GPx activity was also present in the nucleus (56 m-units/mg), although with slightly lower activity than in the homogenate (0.7-fold).

The reversible reaction of GSH with two dietary anticarcinogens, benzyl isothiocyanate (BITC) and phenethyl isothiocyanate (PEITC), has been studied in the absence and presence of human glutathione S-transferases (GSTs). The spontaneous reaction at pH 7.4 and 37 degrees C yielded values for k2 of 17.9 and 6.0 M-1.s-1 for GSH conjugation of BITC and PEITC respectively (forward reaction), and k1 values of 6.9 x 10(-4) and 2.4 x 10(-4) s-1 for dissociation of the respective GSH conjugates, BITC-SG and PEITC-SG (reverse reaction). GSTs A1-1, A2-2, M1a-1a and P1-1 catalysed both the forward and reverse reactions with specific activities (mumol/min per mg at 30 microM isothiocyanate or GSH conjugate) ranging from 23.1 for the GSH conjugation of BITC by GST P1-1 to 0.03 for the dissociation of BITC-SG by GST A1-1. When present at similar concentration to substrates (12 microM), GSTs A1-1 and A2-2 but not GST M1a-1a shifted the equilibrium in favour of BITC-SG or PEITC-SG. Kinetic studies confirmed that GST A1-1 interacted selectively with the GSH conjugates in the micromolar range (Km 6.9 microM, Ki 4.3 microM), whereas GST M1a-1a interacted with BITC-SG and PEITC-SG with approx. 5-fold lower affinity. In conclusion, GSTs are true catalysts; at high intracellular concentration they also sequester GSH conjugates, promoting GSH conjugation, whereas trace extracellular GSTs promote dissociation of effluxed organic isothiocyanate-GSH conjugates.

The glutathione transferases (GSTs) form a group of enzymes responsible for a wide range of molecular detoxications. The photoaffinity label S-(2-nitro-4-azidophenyl)glutathione was used to study the hydrophobic region of the active site of the rat liver GST 1-1 and 2-2 isoenzymes (class Alpha) as well as the rat class-Mu GST 3-3. Photoaffinity labelling was carried out using a version of S-(2-nitro-4-azidophenyl)glutathione tritiated in the arylazido ring. The labelling occurred with higher levels of radioisotope incorporation for the Mu than the Alpha families. Taking rat GST 3-3, 1.18 (+/- 0.05) mol of radiolabel from S-(2-nitro-4-azidophenyl)glutathione was incorporated per mol of dimeric enzyme, which could be blocked by the presence of the strong competitive inhibitor, S-tritylglutathione (Ki = 1.4 x 10(-7) M). Radiolabelling of the protein paralleled the loss of enzyme activity. Photoaffinity labelling by tritiated S-(2-nitro-4-azidophenyl)glutathione on a preparative scale (in the presence and absence of S-tritylglutathione) followed by tryptic digestion and purification of the labelled peptides indicated that GST 3-3 was specifically photolabelled; the labelled peptides were sequenced. Similarly, preparative photoaffinity labelling by S-(2-nitro-4-azidophenyl)glutathione of the rat liver 1-1 isoenzyme, the human GST A1-1 and the human-rat chimaeric GST, H1R1/1, was carried out with subsequent sequencing of radiolabelled h.p.l.c.-purified tryptic peptides. The results were interpreted by means of molecular-graphics analysis to locate photoaffinity-labelled peptides using the X-ray-crystallographic co-ordinates of rat GST 3-3 and human GST A1-1. The molecular-graphical analysis indicated that the labelled peptides are located within the immediate vicinity of the region occupied by S-substituted glutathione derivatives bound in the active-site cavity of the GSTs investigated.

In humans, glutathione-dependent conjugation of halomethanes is polymorphic, with 60% of the population classed as conjugators and 40% as non-conjugators. We report the characterization of the genetic polymorphism causing the phenotypic difference. We have isolated a cDNA that encodes a human class Theta GST (GSTT1) and which shares 82% sequence identity with rat class Theta GST5-5. From PCR and Southern blot analyses, it is shown that the GSTT1 gene is absent from 38% of the population. The presence or absence of the GSTT1 gene is coincident with the conjugator (GSST1+) and non-conjugator (GSTT1-) phenotypes respectively. The GSTT1+ phenotype can catalyse the glutathione conjugation of dichloromethane, a metabolic pathway which has been shown to be mutagenic in Salmonella typhimurium mutagenicity tester strains and is believed to be responsible for carcinogenicity of dichloromethane in the mouse. In humans, the enzyme is found in the erythrocyte and this may act as a detoxification sink. Characterization of the GSTT1 polymorphism will thus enable a more accurate assessment of human health risk from synthetic halomethanes and other industrial chemicals.

The nuclear/cytoplasmic distributions of the GSH and N-acetylcysteine conjugates of monochlorobimane (MCB) have been studied in rat hepatocytes in primary culture after microinjection into the cytoplasm. The immediately occurring fluorescence (within 1-2 s) in the nucleus was much greater than that in the cytoplasm. Other water-soluble fluorescent markers such as Cascade Blue, 6-carboxyfluorescein and Lucifer Yellow also exhibited a similar nuclear localization. MCB, which is poorly soluble in water, gave similar results whether exposed to the exterior of the cell or microinjected as a solution in ethanol. It is concluded that the fluorophores studied here diffuse into the nucleus and accumulate there, including MCB-GSH formed intracellularly from MCB and GSH. Therefore MCB may be of limited value in estimating concentrations of GSH in the nucleus and the cytoplasm, as recently suggested by Bellomo, Vairetti, Stivala, Mirabelli, Richelmi and Orrenius [(1992) Proc. Natl. Acad. Sci. U.S.A. 89, 4412-4416].

GSH transferase P1-1 (GSTP1-1) was modified with group-specific reagents. Kinetic experiments demonstrated that inactivation of GSTP1-1 occurred upon reaction of one arginine residue per subunit with diacetyl, one lysine residue per subunit with 2,4,6-trinitrobenzene sulphonate, or one carboxylate group per subunit with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide. All three inactivation reactions were inhibited by compounds known to bind at the GSH site of the enzyme but were unaffected by the electrophile 1-chloro-2,4-dinitrobenzene. N-terminal sequence analysis showed that Arg-13 was modified by diacetyl and that this modification was inhibited by GSH. Arg-11 was not modified. The lysine residue modified by 2,4,6-trinitrobenzene sulphonate and protected by S-octylglutathione was identified as Lys-44 by sequencing of tryptic peptides. The findings are in agreement with the involvement of Arg-13 and Lys-44 in binding of GSH, as determined from the crystal structure [Reinemer, Dirr, Ladenstein, Huber, Lo Bello, Frederici and Parker (1992) J. Mol. Biol. 227, 214-226]. The present data also implicate a single carboxylate in GSH binding, consistent with the involvement of Asp-98 of subunit B determined from the crystallographic study. The GSH-binding determinants of GSTP1-1 are compared using sequence similarity with those of GSTs of Alpha, Mu and Theta classes.

Reaction of human GSH transferase P1-1 (GSTP1-1) with diethylpyrocarbonate (DEPC) at pH 7.0 and 4 degrees C resulted in covalent modification of an equivalent of one histidine and one tyrosine residue per subunit, with loss of activity. Sequence analysis showed that His-71 and Tyr-7 were modified. Reference to the three-dimensional structure of GSTP1-1 [Reinemer, Dirr, Ladenstein, Huber, Lo Bello, Frederici and Parker (1992) J. Mol. Biol. 227, 214-226] shows that the modification of Tyr-7 is most likely to affect enzyme activity. Kinetic analysis of the DEPC modification of Tyr-7 in GSTP1-1 gave a k2 approx. 150 times that of a peptide comprising residues 1-11 of GSTP1-1. The reaction of Tyr-7 of GSTP1-1 with DEPC was poorly inhibited by 1 mM GSH (14%) or 10 microM S-hexylglutathione (18%). DEPC treatment of the enzyme altered the absorbance at 290 nm in second-derivative spectra, suggesting that a significant amount of tyrosinate ion occurs in the enzyme. GSH, however, did not significantly alter the A290. The data provide the first evidence of unusual chemical reactivity of Tyr-7 and are consistent with its proposed role as a proton acceptor during catalysis.

Glutathione S-transferases (GSTs) are a group of enzymes which play an important role in the detoxication of xenobiotics. It is shown that the expression of human glutathione S-transferase P1-1 (GSTP1-1) is suppressed by retinoic acid (RA) as the result of decreased transcription from its gene, GSTP1. Chloramphenicol acetyltransferase (CAT) assays indicate that the effect of RA on the transcription of a GSTP1 promoter-CAT fusion gene is mediated by the region -99 to +72 of GSTP1. A consensus activator protein 1-binding site, located at nucleotide position -59 to -65 of GSTP1, is suggested to be responsible for RA repression. This effect of RA on GSTP1 expression is mediated by the human beta-type RA receptor, hRAR beta, but not the chicken retinoid X receptor, cRXR. The retinoid X receptor does not augment the action of hRAR beta on GSTP1. In addition, it is shown that GSTP1-1 expression is enhanced by insulin as a result of increased transcription of GSTP1. Assay of CAT activity indicates that the effect of insulin on the transcription of GSTP1 is also mediated by the region -99 to +72 of GSTP1. Comparison with sequences of other insulin-responsive genes, suggests that insulin enhancement of GSTP1 expression is effected by an eight-base-pair sequence, ‘CCCGCGTC’, located at +48 to +55 in intron 1 of the gene. These results are discussed in relation to the increased expression of GSTP1-1 in many tumour cells.

The constitutive and inducible cytosolic glutathione S-transferase (EC 2.5.1.18) subunit compositions of parenchymal cells (hepatocytes) and biliary epithelial cells (BEC) from rat liver have been quantitatively analysed using reverse-phase h.p.l.c. Hepatocytes, analysed in the absence of non-parenchymal cells, expressed constitutively the following subunits, in order of their concentration: 3, 4, 2, 1a, 1b, 8, 6 and 10. BEC express constitutively only four of the GST subunits expressed by hepatocytes and these are, in order of their concentration: subunits 2, 7, 4 and 3. Notable differences from hepatocytes are that BEC completely lack the Alpha-class subunits 1a and 1b that are major subunits in hepatocytes, Mu-class subunits make up a very low proportion of the total, and the Pi-class subunit 7 is a major subunit in BEC, whereas it is essentially absent from hepatocytes. For the first time, the effects of the inducing agents phenobarbitone (PB), beta-naphthoflavone (beta-NF) and ethoxyquin (EQ) have been characterized in a comprehensive and quantitative manner in both cell types. PB, beta-NF and EQ increased total GST protein in hepatocytes by approx. 2-fold, 3-fold and 4-fold respectively. Subunits significantly induced in hepatocytes were (in order of fold-induction): by PB, 1b > 8 > 3 > 2 > 4; by beta-NF, 1b > 8 > 2 > 3 > 4; and by EQ, 7 > 1b > 10 > 8 > 3 > 2 > 1a > 4. In BEC, neither PB nor beta-NF had significant effects on the total amount of GST protein, although PB did significantly induce subunit 3 at the expense of other subunits. EQ increased total GST protein nearly 5-fold in BEC, subunits 7 and 3 being induced dramatically above constitutive levels.

The ontogeny of rat liver glutathione S-transferase (EC 2.5.1.18) (GSTs) during foetal and postnatal development was investigated. The GSTs are dimers, the subunits of which belong to three multigene families, Alpha (subunits 1, 2, 8 and 10), Mu (subunits 3, 4, 6, 9 and 11) and Pi (subunit 7) [Mannervik, Alin, Guthenberg, Jennsson, Tahir, Warholm & Jörnvall (1985) Proc. Natl. Acad. Sci. U.S.A. 82, 7202-7206; Kispert, Meyer, Lalor, Coles & Ketterer (1989) Biochem. J. 260, 789-793]. There is considerable structural homology within each gene family, with the result that whereas reverse-phase h.p.l.c. successfully differentiates individual subunits, immunocytochemical and Northern-blotting analyses may only differentiate families. Enzymic activity, h.p.l.c. and Northern blotting indicated that expression of GST increased from very low levels at 12 days of foetal growth to substantial amounts at day 21. At birth, GST concentrations underwent a dramatic decline and remained low until 5-10 days post partum, after which they increased to adult levels. During the period under study, GST subunits underwent differential expression. The Mu family had a lower level of expression than the Alpha family, and, within the Alpha family, subunit 1 was more dominant in the adult than the foetus. Subunit 2 is the major form in the foetus. Most noteworthy were subunits 7 and 10, which were prominent in the foetus, but present at low levels post partum. Immunocytochemical analysis of the 17-day foetal and newborn rat livers showed marked differences in the distribution of GSTs in hepatocytes. In the 17-day foetal liver Pi greater than Alpha greater than Mu whereas in the newborns Alpha greater than Mu much greater than Pi. Erythropoietic cells were not stained for any of the three GST families. Steady-state mRNA concentrations in the foetus correlated with the relative transcription of the Alpha, Mu and Pi class genes. However, in those genes expressed post partum, namely the Alpha and Mu class, low transcriptional activity was associated with high concentrations of mRNA. This suggests that there is a switch from transcriptional control to post-transcriptional control at birth. GST 7-7 appears to be regulated predominantly by transcription throughout the period of liver development under observation.

A rat liver mitochondrial-matrix fraction was prepared and shown to have 1-chloro-2,4-dinitrobenzene(CDNB)-metabolizing glutathione transferase (GST) activity. Further fractionation by sequential gel filtration, isoelectric focusing or chromatofocusing and hydroxyapatite chromatography yielded three GSTs of pI 9.3, 8.9 and 7.5, none of which bound to a GSH-agarose affinity matrix. Most of the activity was associated with the pI-9.3 form, which was selected for further study. Its activity was tested with the following potential substrates in addition to CDNB: 1,2-dichloro-4-nitrobenzene, p-nitrobenzyl chloride, trans-4-phenylbut-3-en-2-one, 1,2-epoxy-3-(p-nitrophenoxy)propane, ethacrynic acid, menaphthyl sulphate, cumene hydroperoxide, linoleic acid hydroperoxide and 4-hydroxynon-2-enal. Appreciable activity was obtained only with CDNB and ethacrynic acid (82 and 26 mumol/min per mg of protein respectively). The apparent Km for GSH, using 1 mM-CDNB, was 1.9 mM. The enzyme is a dimer of subunit Mr 26,500. It has a free N-terminus, which has enabled the first 33 amino acids to be sequenced. This portion of primary structure has a sequence in common with members of the Theta class of GSTs (eg. 36% identity with subunit 12) and also a sequence which might function as a mitochondrial import signal. It is novel and has been named ‘GST 13-13’.

Glutathione transferases (GSTs) of a novel class, which it is proposed to term Theta, were purified from rat and human liver. Two, named GST 5-5 and GST 12-12, were obtained from the rat, and one, named GST theta, was from the human. Unlike other mammalian GSTs they lack activity towards 1-chloro-2,4-dinitrobenzene and are not retained by GSH affinity matrices. Only GST 5-5 retains full activity during purification, and its activities towards the substrates 1,2-epoxy-3-(p-nitrophenoxy)propane, p-nitrobenzyl chloride, p-nitrophenethyl bromide, cumene hydroperoxide, dichloromethane and DNA hydroperoxide are 185, 86, 67, 42, 11 and 0.03 mumol/min per mg of protein respectively. Earlier preparations of GST 5-5 or GST E were probably a mixture of GST 5-5 and GST 12-12, which was largely inactive, and may also have been contaminated by less than 1% with another GSH peroxidase of far greater activity. Partial analysis of primary structure shows that subunits 5, 12 and theta are related to each other, particularly at the N-terminus, where 25 of 27 residues are identical, but have little relationship to the Alpha, Mu and Pi classes of mammalian GSTs. They do, however, show some relatedness to subunit I of Drosophila melanogaster [Toung, Hsieh & Tu (1990) Proc. Natl. Acad. Sci. U.S.A. 87, 31-35] and the dichloromethane dehalogenase of Methylobacterium DM4 [La Roche & Leisinger (1990) J. Bacteriol, 172, 164-171].

Analogues of GSH in which either the gamma-glutamyl or the glycyl moiety is modified were synthesized and tested as both substrates for and inhibitors of glutathione S-transferases (GSTs) 7-7 and 8-8. Acceptor substrates for GST 7-7 were 1-chloro-2,4-dinitrobenzene (CDNB) and ethacrynic acid (ETA) and for GST 8-8 CDNB, ETA and 4-hydroxynon-trans-2-enal (HNE). The relative ability of each combination of enzyme and GSH analogue to catalyse the conjugation of all acceptor substrates was similar with the exception of the combination of GST 7-7 and gamma-L-Glu-L-Cys-L-Asp, which used CDNB but not ETA as acceptor substrate. In general, GST 7-7 was better than GST 8-8 in utilizing these analogues as substrates, and glycyl analogues were better than gamma-glutamyl analogues as both substrates and inhibitors. These results are compared with those obtained earlier with GSH analogues and GST isoenzymes 1-1, 2-2, 3-3 and 4-4 [Adang, Brussee, Meyer, Coles, Ketterer, van der Gen & Mulder (1988) Biochem. J. 255, 721-724] and the implications with respect to the nature of their active sites are discussed.

GSSG selectively elutes two GSH transferases from a mixture of rat GSH transferases bound to a GSH-agarose affinity matrix. One is a form of GSH transferase 1-1 and the other is shown to be GSH transferase 8-8. By using tissues that lack this form of GSH transferase 1-1 (e.g. lung), GSH transferase 8-8 may thus be purified from cytosol in a single step. Quantitative analysis of the tissue distribution of GSH transferase 8-8 was obtained by h.p.l.c.

A labile GSH transferase homodimer termed 11-11 was purified from rat testis by GSH-agarose affinity chromatography followed by anion-exchange f.p.l.c. The enzyme is unstable in the absence of thiol(s) and has relatively low affinity for both 1-chloro-2,4-dinitrobenzene (Km 4.4 mM) and GSH (Km(app.) 4.4mM). Its mobility on SDS/polyacrylamide-gel electrophoresis is slightly less than that of subunits 3 and 4 and its pI is 5.2. Subunit 11 has a blocked N-terminal amino acid residue, but after CNBr cleavage fragments accounting for 113 amino acid residues were sequenced and showed 65% homology with corresponding sequences in subunit 4, indicating that it is a member of the Mu family. GSH transferase 11 is a major isoenzyme in testis, epididymis, prostate and brain and present at lower concentrations in other tissues.

The human glutathione S-transferase pi gene has been isolated from a cosmid library. The gene spans approximately 3 kb, is interrupted by six introns and the region around its 5′ end has the high G + C and CpG content typical of an HTF (HpaII tiny fragment) island. In addition to a TATA box at position -28 relative to the transcription initiation site and two G + C boxes (GGGCGG), the promoter contains a sequence motif matching the phorbol ester- and ras-responsive element from the polyoma virus enhancer.

DNA peroxidized by exposure to ionizing radiation in the presence of oxygen is a substrate for the Se-independent GSH peroxidase activity of several GSH transferases, GSH transferases 5-5, 3-3 and 4-4 being the most active in the rat liver soluble supernatant fraction (500, 35 and 20 nmol/min per mg of protein respectively) and GSH transferases mu and pi the most active, so far found, in the human liver soluble supernatant fraction (80 and 10 nmol/min per mg respectively). Although the GSH transferase content of the rat nucleus was found to be much lower than that of the soluble supernatant, nuclear GSH transferases are likely to be more important in the detoxification of DNA hydroperoxide produced in vivo. Two nuclear fractions were studied, one extracted with 0.075 M-saline/0.025 M-EDTA, pH 8.0, and the other extracted from the residue with 8.5 M-urea. The saline/EDTA fraction contained subunits 1, 2, 3, 4 and a novel subunit, similar but not identical to 5, provisionally referred to as 5*, in the proportions 40:25:5:5:25 respectively. The 8.5 M-urea-extracted fraction contained principally subunit 5* together with a small amount of subunit 6 in the proportion 95:5 respectively. GSH transferase 5*-5* purified from the 8.5 M-urea extract has the highest activity towards DNA hydroperoxide of any GSH transferase so far studied (1.5 mumol/min per mg). A Se-dependent GSH peroxidase fraction from rat liver was also active towards DNA hydroperoxide; however, since this enzyme accounts for only 14% of the GSH peroxidase activity detectable in the nucleus, GSH transferases may be the more important source of this activity. The possible role of GSH transferases, in particular GSH transferase 5*-5*, in DNA repair is discussed.

The stereoselectivity of purified rat GSH transferases towards alpha-bromoisovaleric acid (BI) and its amide derivative alpha-bromoisovalerylurea (BIU) was investigated. GSH transferase 2-2 was the only enzyme to catalyse the conjugation of BI and was selective for the (S)-enantiomer. The conjugation of (R)- and (S)-BIU was catalysed by the isoenzymes 2-2, 3-3 and 4-4. Transferase 1-1 was less active, and no catalytic activity was observed with transferase 7-7. Isoenzymes 1-1 and 2-2 of the Alpha multigene family preferentially catalysed the conjugation of the (S)-enantiomer of BIU (and BI), whereas isoenzymes 3-3 and 4-4 of the Mu multigene family preferred (R)-BIU. The opposite stereoselectivity of conjugation of BI and BIU previously observed in isolated rat hepatocytes and the summation of activities of enzymes known to be present in hepatocytes on the basis of present data are in accord.

A simple method is described for the separation and quantification of the subunits of GSH transferases present in rat tissue extracts. This method, involving GSH-agarose affinity chromatography followed by reverse-phase h.p.l.c., is rapid and sufficiently sensitive to measure 5 micrograms of each subunit in a mixture. Examples are given of its application to extracts of rat kidney, adrenal, testicular interstitial cells and seminiferous tubules. The analysis of seminiferous tubules indicates that the technique may be of value for the identification of novel subunits. Preliminary separations of subunits from human GSH transferases are also described.

Polyadenylated RNA isolated from NN-dimethyl-4-aminoazobenzene-induced rat hepatoma was used to prepare a cDNA library in lambda gt10. Full-length clones complementary to mRNA coding for glutathione transferase subunit 7 were isolated and one of these clones (pGSTr7) was fully characterized. In Northern blot analysis, mRNA hybridizing to 32P-labelled pGSTr7 was found in poly(A)-containing RNA isolated from seven normal rat tissues but not from testis and liver. A similar hybridizing mRNA species was also detected in human placental mRNA. The same probe, used in a Southern blot analysis of genomic DNA, suggests the presence of a multigene family in the rat.

The content of GSH transferase mRNAs in poly(A)-containing RNA isolated from eight rat tissues was examined by immunoprecipitation of cell-free translation products and by Northern blotting. Considerable tissue-specific distribution and heterogeneity of immunoprecipitable GSH transferase subunits 1 and 2 synthesized in vitro was observed. These results were confirmed by Northern blotting using a 32P-labelled subunit 1 cDNA probe. The same probe, used in a Southern blot analysis of genomic DNA, provided confirmation that GSH transferase subunits 1 and 2 comprise a multigenic family in the rat. The results show that the selection of cDNA clones coding for chosen subunits can be made easier by making use of qualitative and quantitative tissue differences in GSH transferase mRNAs.

Lipid peroxidation in vitro in rat liver microsomes (microsomal fractions) initiated by ADP-Fe3+ and NADPH was inhibited by the rat liver soluble supernatant fraction. When this fraction was subjected to frontal-elution chromatography, most, if not all, of its inhibitory activity could be accounted for by the combined effects of two fractions, one containing Se-dependent glutathione (GSH) peroxidase activity and the other the GSH transferases. In the latter fraction, GSH transferases B and AA, but not GSH transferases A and C, possessed inhibitory activity. GSH transferase B replaced the soluble supernatant fraction as an effective inhibitor of lipid peroxidation in vitro. If the microsomes were pretreated with the phospholipase A2 inhibitor p-bromophenacyl bromide, neither the soluble supernatant fraction nor GSH transferase B inhibited lipid peroxidation in vitro. Similarly, if all microsomal enzymes were heat-inactivated and lipid peroxidation was initiated with FeCl3/sodium ascorbate neither the soluble supernatant fraction nor GSH transferase B caused inhibition, but in both cases inhibition could be restored by the addition of porcine pancreatic phospholipase A2 to the incubation. It is concluded that the inhibition of microsomal lipid peroxidation in vitro requires the consecutive action of phospholipase A2, which releases fatty acyl hydroperoxides from peroxidized phospholipids, and GSH peroxidases, which reduce them. The GSH peroxidases involved are the Se-dependent GSH peroxidase and the Se-independent GSH peroxidases GSH transferases B and AA.

Free polyribosomal poly(A)-containing RNA isolated from normal rat liver was used to prepare a complementary DNA plasmid library in the Pst1 site of the plasmid pAT153 . A plasmid pGSTr155 complementary to mRNA coding for a glutathione transferase Ya subunit was selected by differential hybridization in situ and preliminary characterization was performed by hybrid-selected mRNA translation, immunoprecipitation and polyacrylamide-gel electrophoresis of the product synthesized in vitro. The nucleotide sequence of the complementary DNA contained within pGSTr155 was determined and shown to contain a single open reading frame corresponding to the first 129 amino acids of the N-terminus of the Ya subunit and a further 63 nucleotides upstream of the initiating methionine codon.

A theory is presented that deals with the involvement of the intracellular binding proteins ligandin and aminoazodye-binding protein A (otherwise known as Z-protein or fatty-acid-binding protein) on the uptake and intracellular transport and metabolism of their ligands. Equations are derived that combine steady-state diffusional fluxes of small molecules that are (a) free in the aqueous phase of the cell, (b) bound to the two proteins and (c) partitioned into intracellular membranes, for model systems that resemble conditions in the rat hepatocyte. These equations are then combined with expressions for the enzyme-catalysed metabolic reactions undergone by these small molecules to assess the influence of diffusion rats on the overall metabolic rates. It is concluded that ligandin and protein A can enhance the rate of intracellular of their ligands by an order of magnitude or more and that this could make the hepatocyte several times more efficient in metabolizing these ligands. Various ways of testing this theory are discussed.

1. To assess the possible involvement of ligandin and aminoazo-dye-binding protein A in intracellular transport it is necessary to know how their ligands, most of which are molecules with hydrophobic moieties, interact with cellular membranes. To obtain such information we examined the interactions of bromosulphophthalein, oestrone sulphate, haem and bilirubin with aqueous dispersions of egg phosphatidylcholine and egg phosphatidylchone/cholesterol (1:1, molar ratio) by equilibrium dialysis and spectrophotometry. 2. In all four cases, saturation effects were observed. Values of Vmax (v = mol of compound bound/mol of lipid phosphorus) at 25 degrees C were: for bromosulphophthalein, approximately 0.1; for oestrone sulphate, approximately 0.25; for haem, approximately 0.25 (all at pH 7.4); and for bilirubin 0.1–0.2 (at pH 8.2). 3. Limiting values of v/c (c = unbound concentration) as v leads to 0 at 25 degrees C and pH 7.4 are: for bromosulphophthalein, 6.25 × 10(4) litre-mol-1; for oestrone sulphate, 7.8 × 10(2) litre-mol-1; for haem, 4.5 × 10(5) litre-mol-1; and for bilirubin, approximately 1.2 × 10(4) litre-mol-1. For haem the result depends on the assumption that only the monomeric form binds to the lipid. 4. The binding of each compound was decreased by cholesterol; bromosulphophthalein and oestrone sulphate were affected more than haem and bilirubin. 5. Bromosulphophthalein at saturating concentration decreased the limiting values of v/c of the other three compounds by approximately one order of magnitude. 6. By assuming that the interactions with egg phosphatidylcholine resemble those with the phospholipid components of mammalian intracellular membranes the binding data for phosphyatidylcholine, together with data for binding to the intracellular proteins ligandin and aminoazo-dye-binding protein A, enable the subcellular distributions of the four compounds to be estimated. For the rat hepatocyte up to 92, 51, 98 and 47% of the total bromosulphophthalein, oestrone sulphate, haem and bilirubin respectively may be membrane-bound.

Ligandin (glutathione S-transferase B, EC 2.5.1.18)was treated with p-mercuribenzoate, N-(4-dimethylamino-3,5-dinitrophenyl)-maleimide, 5,5,-dithiobis-(2-nitrobenzoic acid), N-ethylmaleimide, iodoacetamide or iodoacetate. Although performic acid oxidation revealed the presence of four cysteines, p-mercuribenzoate and N-(4-dimethylamino-3,5-dinitrophenyl)maleimide, the most effective of the reagents studied, reacted with only three residues. N-Ethylmaleimide and 5,5′-dithiobis-(2-nitrobenzoic acid) each reacted with two cysteines: iodoacetamide reacted with only one cysteine and iodoacetate was essentially unreactive. Modification of three thiol groups decreased both the enzymic and binding activities of ligandin although the number of binding sites was unaffected. Modification of only one or two of the thiol groups had little effect on the ligandin activities. It therefore appears that there is a thiol group in the common hydrophobic-ligand- and substrate-binding site of ligandin. Ligandin was separated into two fractions on CM-cellulose. Both fractions gave the same results with p-mercuribenzoate and iodoacetamide.

Spectrophotometric and equilibrium-dialysis measurements show that ligandin (glutathione S-transferase B, EC 2.5.1.18) binds monomeric porphyrins at a single site with association constants in the range 10(4)-10(6) litre/mol at pH 7.0. Binding affinities are paralleled by the tendencies of the porphyrins to aggregate, increasing in the order: uroporphyrins I and III less than coproporphyrins I and III approximately haematoporphyrin less than protoporphyrin IX. From this it is deduced that the hydrophobic effect is the predominant driving-force for binding. The porphyrins can be displaced from their binding site on ligandin by bromosulphophthalein and oestrone sulphate. In enzyme inhibition studies, 50% inhibition was brought about by 8 micron-haematoporphyrin and by 1 micron-protoporphyrin IX. In the analysis of the haemotoporphyrin-ligandin system the self-association of haematoporphyrin was studied in detail. It was found to be limited to dimerization in the concentration range 0-200 micron at pH 7.0, 25 degrees C and a dimerization constant of 1.9 × 10(5) litre/mol was determined. Coproporphrin III has a dimerization constant of 5.2 × 10(5) litre/mol under the same conditions.

1. The interactions of ferriprotoporphyrin IX with ligandin and aminoazo-dye-binding protein A result in absorption spectra in the Soret region characteristic of the ligand in its monomeric state. 2. Both proteins are able to bind ferrous as well as ferric haem. 3. Ferriprotoporphyrin IX is bound at a single site on both proteins. At pH7.0, I 0.16M, difference-spectrophotometric measurements gave association constants of 10(7) and 4 × 10(6) LITRE/MOL FOR LIGANDIN AND PROTEin A respectively. Under the same conditions fluorescence-quenching experiments gave an association constant of 2 × 10(7) litre/mol for ligandin. 4. Bilirubin, bromosulphophthalein and oesterone sulphate each compete with haem for binding by the two proteins. 5. Ferriprotoporphyrin IX bound to both ligandin and protein A is able to form co-ordination complexes with CN-, but not, to any measurable extent, with either N3- or F-. From these results it is suggested that binding by the two proteins may not involve the haem iron atom. 6. Both haem-protein complexes give rise to measurable extrinsic Cotton effects in the Soret region. 7. The formation and properties of the ligandin- and protein A-haem complexes are compared with those of haem-albumin, haemoglobin, myoglobin and other haemoproteins.

Ligandin and aminoazo-dye-binding protein A both bind bilirubin at a single site. Quantitative studies of the interactions using difference spectrophotometry show that at pH 7.0, protein A binds the tetrapyrrole with an association constant (K) greater than or equal to 2 × 10(7) litre/mol, whereas binding by ligandin is slightly weaker (K = 7 × 10(6) litre/mol) at this pH. The protein-bilirubin complexes give rise to absorption and fluorescence spectra quite different from those of unbound bilirubin and also to large Cotton effects. It appears that on binding to both proteins, the ligand is forced into a rigid twisted configuration in a hydrophobic environment. Ligandin and protein A resemble serum albumin in their interactions with bilirubin.

A protein of S20,W 1.6S and mol.wt. 14000, which binds covalently a metabolite of the aminoazodye carcinogen NN-dimethyl-4-amino-3′-methylazobenzene, was isolated from rat liver cytosol from both carcinogen-treated and normal rats. The protein binds non-covalently palmitoyl-CoA, fatty acids, bilirubin, sex steroids and their sulphates, bile acids and salts, bromosulphophthalein, diethylstilboestrol and 20-methylcholanthrene with a wide range of affinities. The protein is isolated as three components with isoelectric points of 5.0, 5.9 and 7.6 by a method involving isoelectric focusing. All three components have closely similar amino acid analyses, tryptic-peptide ‘maps’ and u.v. spectra. Each single component redistributes into all three on further electrophoresis. However, the three forms differ in their binding characteristics, the form of pI 7.6 having much the highest affinity for compounds bound non-covalently. The protein was identified immunologically in rat liver, small intestine, adipose tissue, skeletal muscle, myocardium and testis. The protein was compared with other hepatic binding-protein preparations of similar molecular weight.