Isothiocyanates as substrates for human glutathione transferases: structure-activity studies.

The catalytic properties of four human glutathione transferases (GSTs), A1-1, M1-1, M4-4 and P1-1, were examined with 14 isothiocyanate (R-NCS) substrates. The compounds include aliphatic and aromatic homologues, some of which are natural constituents of human food, namely sulphoraphane [1-isothiocyanato-4-(methylsulphinyl)butane], erucin [1-isothiocyanato-4-(methylthio)butane], erysolin [1-isothiocyanato-4-(methylsulphonyl)butane], benzyl-NCS, phenethyl-NCS and allyl-NCS. All isothiocyanates investigated were substrates for the four GSTs. The enzymes promote addition of the thiol group of GSH to the electrophilic central carbon of the isothiocyanate group to form dithiocarbamates [R-NH-C(=S)-SG] which have high UV absorption at 274 nm. Molar absorption coefficients and non-enzymic rate constants as well as standardized enzyme assay conditions for all compounds were established. Of the four isoenzymes investigated, GSTs M1-1 and P1-1 were generally the most efficient catalysts, whereas GST M4-4 was the least efficient. Isothiocyanates are among the GST substrates that are most rapidly conjugated. On the basis of rate-enhancement data and binding energies, the isothiocyanates were compared with 4-hydroxyalkenals, another class of natural GST substrates previously subjected to systematic kinetic analysis. The incremental transition-state stabilization attributable to an increased number of methylene groups in homologous alkyl isothiocyanates is similar to that previously noted for homologous 4-hydroxyalkenals.

High-level bacterial expression of human glutathione transferase P1-1 encoded by semisynthetic DNA.

A cDNA clone, lambda GTHP1del, encoding glutathione transferase (GST) P1-1, was isolated from a human K562 erythroleukemia cell line cDNA library. The coding sequence was lacking the codons for the N-terminal 34 amino acids. A DNA segment was designed in order to obtain the missing portion and a structure representing the entire protein. The synthetic DNA sequence was constructed to achieve efficient base pairing with Escherichia coli 16S ribosomal RNA, avoidance of internal secondary structure, and optimal codon usage for high-level protein expression in accord with the known preferences in E. coli. The truncated GST P1-1 cDNA sequence and the synthetic segment were ligated into a plasmid to give an inducible expression system. Among the resulting clones a limited number was selected by immunodetection for highest yield of GST P1-1. Maximal expression was obtained from a spontaneously mutated sequence with altered as well as deleted bases as compared to the original construct. This clone, pKXHP1, allowed heterologous expression in E. coli in yields of > 200 mg enzyme per liter culture medium. The physicochemical and catalytic properties of the recombinant protein were indistinguishable from those of the enzyme purified from human placenta.

Reversible conjugation of isothiocyanates with glutathione catalyzed by human glutathione transferases.

Rates were determined for the nonezymatic (second order rate constants) and enzyme-catalyzed conjugations with glutathione of four isothiocyanates that are present in edible plants (allyl-, benzyl-, phenethyl-isothiocyanates, and sulforaphane). Of four cloned human glutathione transferases studied, GSTP1-1 and GSTM1-1 were the most efficient catalysts. GSTA1-1 was less efficient, and GSTM2-2 was the least efficient. Conjugation of benzyl-NCS is the most rapid and that of sulforaphane [CH3S(O)(CH2)4-NCS] is the slowest. The large enzymatic rate enhancements and the abundance of the enzymes suggest that the glutathione transferases play important roles in the metabolic disposition of isothiocyanates in humans. Enzymatic cleavage of the GSH conjugates of isothiocyanates (dithiocarbamates) is catalyzed by glutathione transferases. The importance of these reverse reactions is probably limited because they are slow and inhibited by high intracellular concentrations of glutathione.

Contribution of five amino acid residues in the glutathione-binding site to the function of human glutathione transferase P1-1.

Five amino acids in proximity to GSH bound in the active-site cavity of human Class Pi glutathione transferase (GST) P1-1 were mutated by oligonucleotide-directed site-specific mutagenesis. The following mutations gave catalytically active mutant proteins with the proper dimeric structure: Arg14----Ala, Lys45----Ala, Gln52----Ala, Gln65----His and Asp99----Asn. The mutation Gln65----Ala was also made, but the protein was not characterized because of its poor catalytic activity. Residues Arg14, Lys45, Gln52 and Gln65 all contribute to binding of glutathione, and the substitutions caused an approx. 10-fold decrease in affinity, corresponding to 5 kJ/mol, except for Arg14, for which the effect was larger. In addition, Arg14 appears to have an important structure role, since the Arg14----Ala mutant demonstrated a significantly lower stability as compared with the wild-type and the other mutant enzymes. Asp99 primarily contributes to catalysis rather than to binding. The kcat./Km-versus-pH profile for the Asp99----Asn mutant is shifted by 0.5 pH unit in the alkaline direction, and it is proposed that Asp99 may participate in proton transfer in the catalytic mechanism. The possibility of redesigning the substrate specificity for GSTs was shown by the fact that the mutant Lys45----Ala displayed a higher catalytic efficiency with GSH monoethyl ester than with its natural substrate, GSH.

Participation of the phenolic hydroxyl group of Tyr-8 in the catalytic mechanism of human glutathione transferase P1-1.

The coding region of cDNA corresponding to human class Pi glutathione transferase P1-1 was amplified by the PCR, subcloned into an expression vector, pKHP1, expressed in Escherichia coli, and characterized. The physicochemical and catalytic properties of recombinant glutathione transferase P1-1 were indistinguishable from those of the enzyme previously isolated from human placenta. The active-site residue Tyr-8 of the wild-type enzyme was converted into Phe by means of oligonucleotide-directed mutagenesis. The mutant enzyme Y8F displayed a 300-fold decrease in specific activity, ascribable mainly to a lowered k(cat.) (or V) value. Kinetic parameters reflecting binding affinity, S0.5 (substrate concn. giving 1/2V) and I50 (concn. of inhibitor giving 50% remaining activity), were only moderately elevated in the mutant enzyme. These results indicate that Tyr-8 contributes primarily to catalysis as such, rather than to binding of the substrates. The dependence of k(cat.)/Km on pH shows an optimum at pH 7.0, defined by acidic and basic ionic dissociation constants with pKa1 = 6.7 and pKa2 = 7.3 respectively. The mutant enzyme Y8F does not display the basic limb of the k(cat.)/Km versus pH profile, but shows a monotonic increase of k(cat.)/Km with an apparent pKa1 of 7.1. The results indicate that the phenolic hydroxyl group of Tyr-8 in un-ionized form, but not the phenolate of Tyr-8, contributes to catalysis by glutathione transferase P1-1.