Structural determinants of glutathione transferases with azathioprine activity identified by DNA shuffling of alpha class members

A library of alpha class glutathione transferases (GSTs), composed of chimeric enzymes derived from human (A1-1, A2-2 and A3-3), bovine (A1-1) and rat (A2-2 and A3-3) cDNA sequences was constructed by the method of DNA shuffling. The GST variants were screened in bacterial lysates for activity with the immunosuppressive agent azathioprine, a prodrug that is transformed into its active form, 6-mercaptopurine, by reaction with the tripeptide glutathione catalyzed by GSTs. Important structural determinants for activity with azathioprine were recognized by means of primary structure analysis and activities of purified enzymes chosen from the screening. The amino acid sequences could be divided into 23 exchangeable segments on the basis of the primary structures of 45 chosen clones. Segments 2, 20, 21, and 22 were identified as primary determinants of the azathioprine activity representing two of the regions forming the substrate-binding H-site. Segments 21 and 22 are situated in the C-terminal helix characterizing alpha class GSTs, which is instrumental in their catalytic function. The study demonstrates the power of DNA shuffling in identifying segments of primary structure that are important for catalytic activity with a targeted substrate. GSTs in combination with azathioprine have potential as selectable markers for use in gene therapy. Knowledge of activity-determining segments in the structure is valuable in the protein engineering of glutathione transferase for enhanced or suppressed activity.     

Cytosolic glutathione transferases from rat liver. Primary structure of class alpha glutathione transferase 8-8 and characterization of low-abundance class Mu glutathione transferases.

Six GSH transferases with neutral/acidic isoelectric points were purified from the cytosol fraction of rat liver. Four transferases are class Mu enzymes related to the previously characterized GSH transferases 3-3, 4-4 and 6-6, as judged by structural and enzymic properties. Two additional GSH transferases are distinguished by high specific activities with 4-hydroxyalk-2-enals, toxic products of lipid peroxidation. The most abundant of these two enzymes, GSH transferase 8-8, a class Alpha enzyme, has earlier been identified in rat lung and kidney. The amino acid sequence of subunit 8 was determined and showed a typical class Alpha GSH transferase structure including an N-acetylated N-terminal methionine residue.     

DNA改组技术在水蛭素实验进化中的应用

蛋白质的改造是生物工程的重大研究课题.由于结构和功能预测的不精确性,而使按照三维结构信息进行定位诱变往往达不到预期的目的.近年来,另一条改造蛋白质的途径有较大的发展,即在实验室条件下模拟生物分子的自然进化,通过变异和靶功能的选择来获得改进性能的蛋白质[1],此过程称为生物分子实验定向进化.DNA改组(DNAshuffling)是一种改造基因和蛋白质的有效实验进化技术[2].它是在体外进行基因随机片段的重组,从而增加基因的多样性,促使有利变异与不利变异分离,通过选择使有利变异得到优化组合[3].DNA改组包含3个步骤:基因的随机片段化,自身引发PCR和重组合PCR.经过DNA改组的突变体库有可能选择到性能更优的突变体.为进行亲和淘选,需将突变体展示在噬菌体的表面[4].     

Importance of a hypervariable active-site residue in human Mu class glutathione transferases catalyzing the bioactivation of chemotherapeutic thiopurine prodrugs

Glutathione transferases (GSTs) catalyze the bioactivation of the thiopurine prodrugs azathioprine, cis-6-(2-acetylvinylthio)purine (cAVTP) and trans-6-(2-acetylvinylthio)guanine (tAVTG), thereby releasing the antimetabolites 6-mercaptopurine and 6-thioguanine. In the GST Mu class, GST M1-1 has the highest catalytic efficiency, whereas GST M2-2 and other enzymes are less active. In the evolution of Mu class GSTs, residue 210 appears hypervariable and has particular functional significance. We demonstrate that the catalytic activity of GST M1-1 with cAVTP or tAVTG is successively diminished when wild-type Ser-210 is mutated into Ala followed by Thr. Conversely, mutating wild-type Thr-210 in GST M2-2 into Ala and Ser enhanced the corresponding activities. Comparisons were also made with GST M2-2 distinguished by Gly or Pro in position 210, as well as wild-type GSTs M4-4 and M5-5. The results suggest that the hydroxyl group of Ser in position 210 stabilizes the transition state of the GST-catalyzed reaction. The low activity of GSTs containing Thr in position 210 is probably due to steric hindrance caused by the beta-methyl group of the side chain. The ratios of the different catalytic efficiencies were translated into differences in the Gibbs free energies of transition state stabilization. The effects of the mutations were qualitatively parallel for the alternative substrates, but vary significantly in magnitude. From the evolutionary perspective the data show that a point mutation can alternatively enhance or attenuate the activity with a particular substrate and illustrate the functional plasticity of GSTs.     

Activation of rat glutathione transferases in class mu by active oxygen species

The activities of rat glutathione transferases (GSTs) 3-3, 3-4, 4-4 in Class mu towards 1-chloro-2,4-dinitrobenzene (CDNB) but not 1,2-dichloro-4-nitrobenzene were increased up to 5-fold during preincubation with 0.4 mM xanthine and xanthine oxidase in 50 mM potassium phosphate, pH 7.8, containing 0.1 mM EDTA. The activated GST 3-4, purified by S-hexylglutathione affinity chromatography after the treatment, had a higher specific activity (130 units/mg) than that of the nontreated (35 units/mg), the Km and Vmax values for glutathione or CDNB also were increased. Other rat GSTs in Class alpha and pi were inactivated by the same treatment. In the presence of superoxide dismutase, the activation of GST 3-4 did not occur.     

Combinatorial chemical reengineering of the alpha class glutathione transferases

Previously, we discovered that human glutathione transferases (hGSTs) from the alpha class can be rapidly and quantitatively modified on a single tyrosine residue (Y9) using thioesters of glutathione (GS-thioesters) as acylating reagents. The current work was aimed at exploring the potential of this site-directed acylation using a combinatorial approach, and for this purpose a panel of 17 GS-thioesters were synthesized in parallel and used in screening experiments with the isoforms hGSTs A1-1, A2-2, A3-3, and A4-4. Through analytical HPLC and MALDI-MS experiments, we found that between 70 and 80% of the reagents are accepted and this is thus a very versatile reaction. The range of ligands that can be used to covalently reprogram these proteins is now expanded to include functionalities such as fluorescent groups, a photochemical probe, and an aldehyde as a handle for further chemical derivatization. This site-specific modification reaction thus allows us to create novel functional proteins with a great variety of artificial chemical groups in order to, for example, specifically tag GSTs in biological samples or create novel enzymatic function using appropriate GS-thioesters.     

Detoxification of DNA hydroperoxide by glutathione transferases and the purification and characterization of glutathione transferases of the rat liver nucleus.

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.     

Binding and glutathione conjugation of porphyrinogens by plant glutathione transferases

Overexpression in Escherichia coli of a tau (U) class glutathione transferase (GST) from maize (Zea mays L.), termed ZmGSTU1, caused a reduction in heme levels and an accumulation of porphyrin precursors. This disruption was highly specific, with the expression of the closely related ZmGSTU2 or other maize GSTs having little effect. Expression in E. coli of a series of chimeric ZmGSTU1/ZmGSTU2 proteins identified domains responsible for disrupting porphyrin metabolism. In addition to known heme precursors, expression of ZmGSTU1 led to the accumulation of a novel glutathione conjugate of harderoporphyrin(ogen) (2,7,12,18-tetramethyl-3-vinylporphyrin-8,13,17-tripropionic acid). Using the related protoporphyrinogen as a substrate, conjugation could be shown to occur on one vinyl group and was actively catalyzed by the ZmGSTU. In plant transgenesis studies, the ZmGSTUs did not perturb porphyrin metabolism when expressed in the cytosol of Arabidopsis or tobacco. However, expression of a ZmGSTU1-ZmGSTU2 chimera in the chloroplasts of tobacco resulted in the accumulation of the harderoporphyrin(ogen)-glutathione conjugate observed in the expression studies in bacteria. Our results show that the well known ability of GSTs to act as ligand binding (ligandin) proteins of porphyrins in vitro results in highly specific interactions with porphyrinogen intermediates, which can be demonstrated in both plants and bacteria in vivo.     

Evolution of a cytokine using DNA family shuffling.

DNA shuffling of a family of over 20 human interferon-alpha (Hu-IFN-alpha) genes was used to derive variants with increased antiviral and antiproliferation activities in murine cells. A clone with 135,000-fold improved specific activity over Hu-IFN-alpha2a was obtained in the first cycle of shuffling. After a second cycle of selective shuffling, the most active clone was improved 285,000-fold relative to Hu-IFN-alpha2a and 185-fold relative to Hu-IFN-alpha1. Remarkably, the three most active clones were more active than the native murine IFN-alphas. These chimeras are derived from up to five parental genes but contained no random point mutations. These results demonstrate that diverse cytokine gene families can be used as starting material to rapidly evolve cytokines that are more active, or have superior selectivity profiles, than native cytokine genes.     

Programmed delivery of novel functional groups to the alpha class glutathione transferases

Here we describe a new route to site- and class-specific protein modification that will allow us to create novel functional proteins with artificial chemical groups. Glutathione transferases from the alpha but not the mu, pi, omega, or theta classes can be rapidly and site-specifically acylated with thioesters of glutathione (GS-thioesters) that are similar to compounds that have been demonstrated to occur in vivo. The human isoforms A1-1, A2-2, A3-3, and A4-4 from the alpha class all react with the reagent at a conserved tyrosine residue (Y9) that is crucial in catalysis of detoxication reactions. The yield of modified protein is virtually quantitative in less than 30 min under optimized conditions. The acylated product is stable for more than 24 h at pH 7 and 25 degrees C. The modification is reversible in the presence of excess glutathione, but the labeled protein can be protected by adding S-methylglutathione. The stability of the ester with respect to added glutathione depends on the acyl moiety. The reaction can also take place in Escherichia coli lysates doped with alpha class glutathione transferases. A control substance that lacks the peptidyl backbone required for binding to the glutathione transferases acylates surface-exposed lysines. There is some acyl group specificity since one out of the three different GS-thioesters that we tried was not able to acylate Y9.     

Purification and characterization of glutathione transferases with an activity toward nitroglycerin from human aorta and heart. Multiplicity of the human class Mu forms

Although recent studies suggest involvement of glutathione transferase (GST) of blood vessels in vasodilation by nitroglycerin, GST forms in blood vessels remain to be studied. In this study, three GST forms (pI values 8.3, 6.6, and 4.8) were purified from human aorta and four (pI values 6.0, 5.6, 5.3, and 4.6) from the heart by affinity chromatography followed by chromatofocusing. The major form of both aorta (pI 4.8) and heart (pI 4.6) was identified as GST-pi, and the other five forms were immunologically related to GST-mu, suggesting that the five belong to the Mu class. Among nine human GST forms, including three in the Alpha class purified from the liver, GST-mu, aorta pI 8.3 form, and GST-I (a form of the Alpha class, corresponding to GST-epsilon (B1B1)) showed high activities toward nitroglycerin, 1.08, 0.85, and 0.78 units/mg protein, respectively. GST-pi did not exhibit the activity. The Km values of the aorta form (pI 8.3) for glutathione (GSH) and nitroglycerin were calculated as 0.12 and 1.1 mM, respectively. The Km values of GST-mu and GST-I for GSH were 0.29 and 0.09 mM, and those for nitroglycerin were 2.5 and 0.3 mM, respectively. The activity of the pI 8.3 form as well as GST-mu toward nitroglycerin was inhibited by bromosulfophthalein, which is known to inhibit the relaxation of rabbit aorta induced by nitroglycerin, at the lower concentration (IC50, 2 microM) than was GST-I (IC50, 32 microM). Two-dimensional gel electrophoresis and N-terminal amino acid sequence analysis revealed that five forms in the Mu class are homo- or heterodimers of five different subunits named M1 (pI 7.0/Mr 27,000), M2 (6.6/27,000), M3 (6.0/27,000), N1 (6.5/26,500), and N2 (5.9/26,500). The subunit structures of the five forms are as follows: pI 8.3 form, M1M2; 6.6 form, M2N1; 6.0 form, M3M3; 5.6 form, M3N2; and 5.3 form, N2N2. M3 and N2 seem to correspond to the subunits of GST-mu, and -4 (Board, P. G., Suzuki, T., and Shaw, D. C. (1988) Biochim. Biophys. Acta 953, 214-217), respectively. These subunits except N1 are different from each other at two or three positions in the first 20 residues of N-terminal amino acid sequence. These results indicate the presence of five different subunits in the human Mu class and also suggest that GST-M1M2 and -M2N1 found in the aorta are involved in the expression of the pharmacologic effect of nitroglycerin.     

Complementary substrate specificities of class I and class II collagenases from Clostridium histolyticum

The substrate specificities of three class I (beta, gamma, and eta) and three class II (sigma, epsilon, and zeta) collagenases from Clostridium histolyticum have been investigated by quantitating the kcat/KM values for the hydrolysis of 53 synthetic peptides with collagen-like sequences covering the P3 through P3 subsites of the substrate. For both classes of collagenases, there is a strong preference for Gly in subsites P1' and P3. All six enzymes also prefer substrates that contain Pro and Ala in subsites P2 and P2' and Hyp, Ala, or Arg in subsite P3'. This agrees well with the occupancies of these sites by these residues in type I collagen. However, peptides with Glu in subsites P2 or P2' are not good substrates, even though Glu occurs frequently in these positions in collagen. Conversely, all six enzymes prefer aromatic amino acids in subsite P1, even though such residues do not occur in this position in type I collagen. In general, the class II enzymes have a broader specificity than the class I enzymes. However, they are much less active toward sequences containing Hyp in subsites P1 and P3'. Thus, the two classes of collagenases have similar but complementary sequence specificities. This accounts for the ability of the two classes of enzymes to synergistically digest collagen.     

Development of innovative pediocin PA-1 by DNA shuffling among class IIa bacteriocins

Pediocin PA-1 is a member of the class IIa bacteriocins, which show antimicrobial effects against lactic acid bacteria. To develop an improved version of pediocin PA-1, reciprocal chimeras between pediocin PA-1 and enterocin A, another class IIa bacteriocin, were constructed. Chimera EP, which consisted of the C-terminal half of pediocin PA-1 fused to the N-terminal half of enterocin A, showed increased activity against a strain of Leuconostoc lactis isolated from a sour-spoiled dairy product. To develop an even more effective version of this chimera, a DNA-shuffling library was constructed, wherein four specific regions within the N-terminal half of pediocin PA-1 were shuffled with the corresponding sequences from 10 other class IIa bacteriocins. Activity screening indicated that 63 out of 280 shuffled mutants had antimicrobial activity. A colony overlay activity assay showed that one of the mutants (designated B1) produced a >7.8-mm growth inhibition circle on L. lactis, whereas the parent pediocin PA-1 did not produce any circle. Furthermore, the active shuffled mutants showed increased activity against various species of Lactobacillus, Pediococcus, and Carnobacterium. Sequence analysis revealed that the active mutants had novel N-terminal sequences; in active mutant B1, for example, the parental pediocin PA-1 sequence (KYYGNGVTCGKHSC) was changed to TKYYGNGVSCTKSGC. These new and improved DNA-shuffled bacteriocins could prove useful as food additives for inhibiting sour spoilage of dairy products.     

Proposed reductive metabolism of artemisinin by glutathione transferases in vitro

Artemisinin is a sesquiterpene lactone containing an endoperoxide bridge. It is a promising new antimalarial and is particularly useful against the drug resistant strains of Plasmodium falciparum. It has unique antimalarial properties since it acts through the generation of free radicals that alkylate parasite proteins. Since the antimalarial action of the drug is antagonised by glutathione and ascorbate and has unusual pharmacokinetic properties in humans, we have investigated if the drug is broken down by a typical reductive reaction in the presence of glutathione transferases. Cytosolic glutathione transferases (GSTs) detoxify electrophilic xenobiotics by catalysing the formation of glutathione (GSH) conjugates and exhibit glutathione peroxidase activity towards hydroperoxides. Artemisinin was incubated with glutathione, NADPH and glutathione reductase and GSTs in a coupled assay system analogous to the standard assay scheme with cumene hydroperoxide as a substrate of GSTs. Artemisinin was shown to stimulate NADPH oxidation in cytosols from rat liver, kidney, intestines and in affinity purified preparations of GSTs from rat liver. Using human recombinant GSTs hetelorogously expressed in Escherichia coli, artemisinin was similarly shown to stimulate NADPH oxidation with the highest activity observed with GST M1-1. Using recombinant GSTs the activity of GSTs with artemisinin was at least two fold higher than the reaction with CDNB. Considering these results, it is possible that GSTs may contribute to the metabolism of artemisinin in the presence of NADPH and GSSG-reductase We propose a model, based on the known reactions of GSTs and sesquiterpenes, in which (1) artemisinin reacts with GSH resulting in oxidised glutathione; (2) the oxidised glutathione is then converted to reduced glutathione via glutathione reductase; and (3) the latter reaction may then result in the depletion of NADPH via GSSG-reductase. The ability of artemisinin to react with GSH in the presence of GST may be responsible for the NADPH utilisation observed in vitro and suggests that cytosolic GSTs are likely to be contributing to metabolism of artemisinin and related drugs in vivo.     

Induction of anionic glutathione transferases in rat liver by propylthiouracil

The treatment of rats with propylthiouracil (PTU) resulted in an induction of not only cationic but also anionic glutathione (GSH) transferases. DE-52 column chromatography of the anionic GSH transferases divided into five main peaks (I-V). Peaks I-IV had a high activity toward 1-chloro-2,4-dinitrobenzene (CDNB). Peak V, which is a new form of the enzyme, showed high activity to ethacrynic acid. PTU induced peaks I and V, whereas phenobarbital increased the activity of only peak I.     

Identification, characterization, and crystal structure of the Omega class glutathione transferases

A new class of glutathione transferases has been discovered by analysis of the expressed sequence tag data base and sequence alignment. Glutathione S-transferases (GSTs) of the new class, named Omega, exist in several mammalian species and Caenorhabditis elegans. In humans, GSTO 1-1 is expressed in most tissues and exhibits glutathione-dependent thiol transferase and dehydroascorbate reductase activities characteristic of the glutaredoxins. The structure of GSTO 1-1 has been determined at 2.0-A resolution and has a characteristic GST fold (Protein Data Bank entry code ). The Omega class GSTs exhibit an unusual N-terminal extension that abuts the C terminus to form a novel structural unit. Unlike other mammalian GSTs, GSTO 1-1 appears to have an active site cysteine that can form a disulfide bond with glutathione.     

Heme transport and detoxification in nematodes: subproteomics evidence of differential role of glutathione transferases

In contrast to their mammalian hosts, parasitic nematodes are heme auxotrophs and require pathways for the uptake and transport of exogenous heme for incorporation into hemoproteins. Phase II detoxification Nu-class glutathione transferase (GST) proteins have a proposed role as heme-binding ligandins in parasitic nematodes. The genome-verified free-living nematode Caenorhabditis elegans also cannot synthesize heme and is an ideal functional genomics model to delineate the role of individual nematode GSTs in heme trafficking and heme detoxification. In this study, C. elegans was exposed to externally controlled heme concentrations ranging from 20-fold suboptimal growth levels to 10-fold supra-optimal growth levels to mimic fluctuations in blood- and tissue-feeding parasitic cousins from the same nematode group. A new heme-responsive GST (GST-19) was identified by subproteomics approaches. Functional characterization of this and two other C. elegans GSTs revealed that they all have high affinity for heme compounds similar to mammalian soluble heme carrier proteins such as HBP23 ( K d approximately 10 (-8) M). In the genomics-predicted absence of orthologous mammalian soluble heme-binding proteins in nematodes, we propose that Nu-class GSTs are candidates in the cellular processing of heme compounds. Toxic heme binding may be coupled to enzymatic protection from its breakdown as several GSTs possess glutathione peroxidase activity.     

Structural evidence for conformational changes of Delta class glutathione transferases after ligand binding

We report four new crystal structures for Delta class glutathione transferases from insects. We compare these new structures as well as several previously reported structures to determine that structural transitions can be observed with ligand binding. These transitions occurred in the regions around the active site entrance, including alpha helix 2, C-terminus of alpha helix 4 including the loop to helix 5 and the C-terminus of helix 8. These structural movements have been reported or postulated to occur for several other glutathione transferase classes; however, this is the first report showing structural evidence of all these movements occurring, in this case in Delta class glutathione transferases. These fluctuations also can be observed occurring within a single structure as there is ligand bound in only one subunit and each subunit is undergoing different conformational transitions. The structural comparisons show reorganizations occur both pre- and post-GSH ligand binding communicated through the subunit interface of the quaternary assembly. Movements of these positions would allow 'breathing' of the active site for substrate entrance, topological rearrangement for varying substrate specificity and final product release.     

Equilibrium folding of dimeric class mu glutathione transferases involves a stable monomeric intermediate

The conformational stabilities of two homodimeric class mu glutathione transferases (GSTM1-1 and GSTM2-2) were studied by urea- and guanidinium chloride-induced denaturation. Unfolding is reversible and structural changes were followed with far-ultraviolet circular dichroism, tryptophan fluorescence, enzyme activity, chemical cross-linking, and size-exclusion chromatography. Disruption of secondary structure occurs as a monophasic transition and is independent of protein concentration. Changes in tertiary structure occur as two transitions; the first is protein concentration dependent, while the second is weakly dependent (GSTM1-1) or independent (GSTM2-2). The second transition corresponds with the secondary structure transition. Loss in catalytic activity occurs as two transitions for GSTM1-1 and as one transition for GSTM2-2. These transitions are dependent upon protein concentration. The first deactivation transition coincides with the first tertiary structure transition. Dimer dissociation occurs prior to disruption of secondary structure. The data suggest that the equilibrium unfolding/refolding of the class mu glutathione transferases M1-1 and M2-2 proceed via a three-state process: N(2) <--> 2I <--> 2U. Although GSTM1-1 and GSTM2-2 are homologous (78% identity/94% homology), their N(2) tertiary structures are not identical. Dissociation of the GSTM1-1 dimer to structured monomers (I) occurs at lower denaturant concentrations than for GSTM2-2. The monomeric intermediate for GSTM1-1 is, however, more stable than the intermediate for GSTM2-2. The intermediates are catalytically inactive and display nativelike secondary structure. Guanidinium chloride-induced denaturation yields monomeric intermediates, which have a more loosely packed tertiary structure displaying enhanced solvent exposure of its tryptophans and enhanced ANS binding. The three-state model for the class mu enzymes is in contrast to the equilibrium two-state models previously proposed for representatives of classes alpha/pi/Sj26 GSTs. Class mu subunits appear to be intrinsically more stable than those of the other GST classes.