Engineering Glutathione Transferase to a Novel Glutathione Peroxidase Mimic With High Catalytic Efficiency INCORPORATION OF SELENOCYSTEINE INTO A GLUTATHIONE-BINDING SCAFFOLD USING AN AUXOTROPHIC EXPRESSION SYSTEM*

Glutathione peroxidase (GPx, EC 1.11.1.9) protects cells against oxidative damage by catalyzing the reduction of hydroperoxides with glutathione (GSH). Several attempts have been made to imitate its function for me-chanical study and for its pharmacological development as an antioxidant. By replacing the active site serine 9 with a cysteine and then substituting it with selenocysteine in a cysteine auxotrophic system, catalytically essential residue selenocysteine was bioincorporated into GSH-specific binding scaffold, and thus, glutathione S transferase (GST, EC 2.5.1.18) from Lucilia cuprina was converted into a selenium-containing enzyme, seleno-LuGST1-1, by genetic engineering. Taking advantage of the important structure similarities between seleno-LuGST1-1 Assay of Enzyme The activities of enzymes were meas- ured using a UV-visible spectrophotometer (Shimadzu UV-3100). The GPx activities of enzymes were measured according to Wilson’s method (25). The reaction was carried out at 37 °C in 500 (cid:3) l of solution containing 50 m M , pH 7.0, potassium phosphate buffer, 1 m M GSH, 1 unit of GSH reductase, and 0.5–2.5 (cid:3) g of enzyme. The mixture was preincubated for 7 min, and 0.25 m M NADPH solution was added. After the mixture was incubated for 3 min at 37 °C, the reaction was initiated by addition of 0.5 m M hydrogen peroxide. The activity was determined by the decrease of NADPH absorption at 340 nm. Background absorption was run without enzyme and was subtracted. The activity unit of enzyme is defined as the amount of enzyme that catalyzes the turnover of 1 (cid:3) mol of NADPH per min. The specific activity is expressed in (cid:3) mol (cid:1) min (cid:6) 1 (cid:1) (cid:3) mol (cid:6) 1 of enzyme. The GST activities of wild type LuGST1-1 and its various mutants were measured as described by Habig (26). The reaction was carried out at 30 °C in 1 ml of solution containing 100 m M , pH 6.5, sodium phos- phate buffer, 1 m M GSH, and 0.5–5 (cid:3) g of enzyme. After preincubation for 3 min, 1 m M 1-chloro-2,4-dinitrobenzene was added and then the absorbance was recorded at 340 nm for 3 min. Background absorption was run without enzyme and was subtracted. The activity unit of enzyme is defined as the amount of enzyme that catalyzes the turnover of 1 (cid:3) mol of 1-chloro-2,4-dinitrobenzene per min. The specific activity is expressed in (cid:3) mol (cid:1) min (cid:6) 1 (cid:1) (cid:3) mol (cid:6) 1 of enzyme. Determination of Optimal pH and Temperature for Seleno-LuGST1-1 Catalysis— The initial rates were measured using 1 m M GSH and 0.5 m M hydrogen peroxide. The pH value of the buffer was changed from 6.0 to 10.0 to determine the initial rates of the reaction to obtain the optimal pH condition for seleno-LuGST1-1-catalyzed reaction. Simi- larly, a catalytic reaction was carried out at different temperatures from 20 °C to 45 °C to determine the optimal temperature for the seleno-LuGST1-1-catalyzed reduction of hydroperoxide. Steady-state Kinetics of Seleno-LuGST1-1— The assay of kinetics of seleno-LuGST1-1 for the reduction of H 2 O 2 by GSH was similar to that of selenium-containing catalytic antibody, Se-4A4 (5). The initial rates were measured by observing the decrease of NADPH absorption at 340 nm at several concentrations of one substrate while the concentration of the second substrate was kept constant. All kinetic experiments were performed in a total volume of 0.5 ml containing 50 m M potassium phosphate buffer (pH 7.0), 1 m M EDTA, 1 unit of GSH reductase, 0.25 m M NADPH, and varying concentrations of GSH, H 2 O 2 , and seleno- LuGST1-1. After the enzyme was preincubated with GSH, NADPH, and GSH reductase, the reaction was then initiated by the addition of H 2 O 2 . Subtraction of the nonenzymatic background absorption gave the rate of the enzyme-catalyzed reaction.

Glutathione peroxidase (GPx, 1 EC.1.11.1.9) is a selenoenzyme that functions to catalyze the reduction of hydroperoxides using glutathione (GSH) as a reducing substrate (1). It therefore plays an important role in the organismal antioxidant defense mechanism protecting cells from oxidative stress, which is responsible for many diseases such as reperfusion injury, brain ischemia, tumor, cataract, inflammation, and physiological aging. Extensive investigations of the structure and the catalytic mechanism of GPx reveal that a selenocysteine (Sec) residue is in the enzyme active site and thus participates in GPx activity. In the catalytic cycle, the selenol in the reduced Sec residue is oxidized by hydroperoxide to produce the selenenic acid, which is further converted to the selenenyl sulfide by the nucleophilic attack of GSH. Reaction of selenenyl sulfide with second equivalent of GSH regenerates the reduced selenol (2).
Because the Sec is encoded by a stop codon UGA, it is difficult to prepare GPx with traditional recombinant DNA technology. To imitate GPx properties, Sec has been incorporated into some natural enzymes by chemical modification or genetic engineering in an auxotrophic expression system (3,4). The first semisynthetic GPx-like selenoenzyme is selenosubtilisin, which was generated by chemical conversion of the catalytically active residue serine (Ser-221) of subtilisin to Sec. Alternatively, the biosynthetic substitution of the catalytically essential residue cysteine (Cys-149) of phosphorylating glyceraldehyde-3-phosphate dehydrogenase by Sec led to selenoglyceraldehyde-3-phosphate dehydrogenase, which displayed the GPx-like properties. However, the mechanism that these two selenoenzymes applied to catalyze the reduction of hydroperoxide involves use of aryl thiols instead of GSH as reducing substrate due to the lack of a GSH-specific binding site in their active sites. To enhance the GPx activities, the GSH binding site has been successfully introduced into GPx mimics using monoclonal antibody (5) and bioimprinting techniques (6), and the as-synthesized selenoantibody and bioimprinted selenoprotein containing the GSH binding site exhibited high GPx activities. Besides the efficient binding of substrates, however, some other factors, such as the correct geometric conformation and the microenvironment of the active site, are also dominant for the enhancement of enzyme catalysis. Thus, it is challenging to generate novel models that are more suitable for GPx catalysis. Intensive studies of protein structures have revealed that the evolution of proteins for novel functions is largely based on the redesign of existing protein frameworks in nature (7,8). This principle can be exploited in generation of novel GPx-like biocatalyst by redesigning existing protein scaffolds that are similar to that of natural GPx.
Glutathione S-transferases (GSTs, EC 2.5.1.18) are a family of multifunctional proteins capable of detoxifying endogenous and xenobiotic electrophiles by addition of GSH to the electro-philes (9 -11). Owing to their modular features, their stability, ease of purification, and the wealth of accumulated knowledge of their structure-activity relationship, GSTs have been used as ideal candidates to redesign novel binding properties and catalytic functions (9,12,13). GSTs are divided into different classes according to the sequence similarity, but all of their structures consist of two domains, a conserved GSH binding site at the N terminus and a hypervariable hydrophobic xenobiotic substrates binding site at the C terminus (9 -11). GSTs and GPxs both belong to the thioredoxin superfamily (also including thioredoxin, glutaredoxin, and disulfide-bond formation facilitator) classified by the common glutathione binding domain-adopted thioredoxin fold (9,14). Their active site residues (Tyr or Ser in GST and Sec in GPx) that interact with the thiol group of the substrate glutathione hold the similar positions in their protein structures (14). Moreover, several GSTs display GPx-like catalytic properties toward organic hydroperoxides, although they do not catalyze the reduction of hydrogen peroxide (15). Accordingly, GSTs appear to offer model protein scaffolds to confer GPx properties by engineering Sec residue into the GSH binding sites. Recently, we have successfully converted the rat -class glutathione transferase T2-2 into a selenoenzyme by chemically modifying the active site Ser to Sec (16). This novel selenium-containing enzyme displayed dramatically high GPx activities for catalyzing the reduction of hydrogen peroxide by GSH. However, because the chemical modification is incapable of specifically targeting amino acid residues in the active site, other hydroxyl groups in the protein are inevitably converted into selenols, which would hamper the further structure-function studies of this important selenoenzyme. In view of this, genetic engineering should provide a better and more suitable alternative to incorporate Sec into the defined GST binding site. Here, we report the conversion of Australian sheep blowfly (Lucilia cuprina) glutathione transferase (LuGST1-1) to selenoenzyme (seleno-LuGST1-1) by means of genetic engineering. The serine 9 in the active site of the LuGST1-1 was mutated to cysteine and then biosynthetically substituted to selenocysteine in an auxotrophic expression system. The as-generated novel selenium-dependent enzyme exhibited high catalytic activity toward the reduction of H 2 O 2 by GSH, which is in the same order of magnitude compared with natural GPx. For the first time, a selenium-containing enzyme with such remarkable GPx activity was generated by genetic engineering in bacteria.

MATERIALS AND METHODS
Bacterial Strains, Plasmids, and Media-Strains and plasmids used in this study are listed in Table I. The media employed in cultivation and expression experiments were two types of modifications of M9 minimal medium as described previously (17).
DNA Manipulations-All restriction enzymes and nucleic acid-modifying enzymes were obtained from TAKARA. Plasmid isolation, DNA restriction endonuclease analysis, ligation, and transformation were performed as described previously (18). The Qiagen minipreps DNA purification system and Biotech DNA clean-up system were used to prepare plasmid DNA for restriction enzyme digestion and recover DNA fragment from low melting agarose gels, respectively.
Overexpression and Purification of Wild Type LuGST1-1 and the Mutants GST(S9C) and GST(S9C C86/200S)-Wild type LuGST1-1 was expressed and purified by affinity chromatography on glutathione-Sepharose 4B as described previously (19). Plasmids pSM4 and pSM3 were transformed into BL21(DE3), and the strains were grown in the presence of isopropyl thio-␤-D-galactopyranoside to induce expression of mutant GST(S9C) and GST(S9C C86/200S), respectively (21). With regard to the purification, the two mutants were passed through DE52 ion-exchange columns and then applied to glutathione-Sepharose 4B columns as described previously (22) with a slight modification.
Overexpression and Purification of Seleno-LuGST1-1-Plasmid pSM3 were transformed into strain BL21cysE51. Overexpression of seleno-LuGST1-1 in the presence of selenocysteine was performed as already described for (Se) 2 -thioredoxin (17). And the protein was also purified by affinity chromatography on glutathione-Sepharose 4B as described previously (19) with a slight modification. After elution from the affinity column with 5 mM GSH in 50 mM Tris/HCl, pH 9.6, the enzyme fraction was dialyzed and applied to the Sephadex G-25 column. The enzyme fraction was collected, dialyzed again, and lyophilized.
Determination of Protein Concentration-Protein concentration was determined by a Bio-Rad protein assay using bovine serum albumin as a standard following the method of Bradford (23). The electrophoresis (SDS-PAGE) was carried out on a Hoefer Mighty Small II electrophoresis unit according to the protocols from the manufacturer (40).
Electrospray Mass Spectrometry Analysis-All MALDI-TOF mass spectra were acquired on a Voyager DE-STR Biospectrometry workstation (PerSeptive Biosystems, Framingham, MA) using a nitrogen laser (337 nm). The protein samples were purified by high-performance liquid chromatography and prepared using a conventional dried droplet protocol in which sinapin acid was used as the matrix. The sinapin acid matrix was prepared as a saturated, aqueous solution that contained 60% acetonitrile and 0.3% trifluoroacetic acid. An aliquot sample of 1 l was mixed with 20 l of sinapin acid matrix before depositing 1.2 l of the sample matrix mixture on the MALDI sample stage.

This study
Assay of Enzyme Activities-The activities of enzymes were measured using a UV-visible spectrophotometer (Shimadzu UV-3100). The GPx activities of enzymes were measured according to Wilson's method (25). The reaction was carried out at 37°C in 500 l of solution containing 50 mM, pH 7.0, potassium phosphate buffer, 1 mM GSH, 1 unit of GSH reductase, and 0.5-2.5 g of enzyme. The mixture was preincubated for 7 min, and 0.25 mM NADPH solution was added. After the mixture was incubated for 3 min at 37°C, the reaction was initiated by addition of 0.5 mM hydrogen peroxide. The activity was determined by the decrease of NADPH absorption at 340 nm. Background absorption was run without enzyme and was subtracted. The activity unit of enzyme is defined as the amount of enzyme that catalyzes the turnover of 1 mol of NADPH per min. The specific activity is expressed in mol⅐min Ϫ1 ⅐mol Ϫ1 of enzyme.
The GST activities of wild type LuGST1-1 and its various mutants were measured as described by Habig (26). The reaction was carried out at 30°C in 1 ml of solution containing 100 mM, pH 6.5, sodium phosphate buffer, 1 mM GSH, and 0.5-5 g of enzyme. After preincubation for 3 min, 1 mM 1-chloro-2,4-dinitrobenzene was added and then the absorbance was recorded at 340 nm for 3 min. Background absorption was run without enzyme and was subtracted. The activity unit of enzyme is defined as the amount of enzyme that catalyzes the turnover of 1 mol of 1-chloro-2,4-dinitrobenzene per min. The specific activity is expressed in mol⅐min Ϫ1 ⅐mol Ϫ1 of enzyme.
Determination of Optimal pH and Temperature for Seleno-LuGST1-1 Catalysis-The initial rates were measured using 1 mM GSH and 0.5 mM hydrogen peroxide. The pH value of the buffer was changed from 6.0 to 10.0 to determine the initial rates of the reaction to obtain the optimal pH condition for seleno-LuGST1-1-catalyzed reaction. Similarly, a catalytic reaction was carried out at different temperatures from 20°C to 45°C to determine the optimal temperature for the seleno-LuGST1-1-catalyzed reduction of hydroperoxide.
Steady-state Kinetics of Seleno-LuGST1-1-The assay of kinetics of seleno-LuGST1-1 for the reduction of H 2 O 2 by GSH was similar to that of selenium-containing catalytic antibody, Se-4A4 (5). The initial rates were measured by observing the decrease of NADPH absorption at 340 nm at several concentrations of one substrate while the concentration of the second substrate was kept constant. All kinetic experiments were performed in a total volume of 0.5 ml containing 50 mM potassium phosphate buffer (pH 7.0), 1 mM EDTA, 1 unit of GSH reductase, 0.25 mM NADPH, and varying concentrations of GSH, H 2 O 2 , and seleno-LuGST1-1. After the enzyme was preincubated with GSH, NADPH, and GSH reductase, the reaction was then initiated by the addition of H 2 O 2 . Subtraction of the nonenzymatic background absorption gave the rate of the enzyme-catalyzed reaction.

RESULTS AND DISCUSSION
Production and Isolation of Seleno-LuGST1-1-The method used was based on the assumption that an efficient charging of tRNA Cys occurred with selenocysteine when cysteine was omitted (17). Thus, Cys could be substituted efficiently by Sec in a cysteine auxotrophic strain when Cys was omitted in the growth medium. In the case of LuGST1-1 from the Australian sheep blowfly (L. cuprina), the active residue Ser-9 was mutated to Cys and then converted to Sec during expression in the auxotrophic strain. However, because there are two other cysteines in this protein, i.e. the Cys-86 and the Cys-200, interpretation of kinetic data would be hampered if all the three cysteines were substituted by Sec. Therefore, before introducing selenocysteine at position 9, the other two cysteines were changed into serines. Since their positions are apart from the active site and they don't form a disulfide bridge, the substitutions were expected to have no effect on protein folding and catalytic activity.
Analysis of the results of SDS-PAGE (Fig. 1) indicated that seleno-LuGST1-1 was about 13% of all the proteins expressed in the strain. Inclusion bodies and soluble form of seleno-LuGST1-1 were produced simultaneously in the expression. The yield of soluble seleno-LuGST1-1 purified from the soluble protein was about 1.13 mg/liter of culture.
Biochemical Characterization of the Purified Protein-The purified selenoenzyme sample was characterized by MALDI-TOF mass spectrometry. The MALDI-TOF mass spectrometry results revealed that there were two enzymes components in the sample. The mass of the main component was 38,811 Da, which was predicted for the ESeO 2 H form. The minor peak (38,651 Da) was identified as the mutant (S9C C86/200S). Based on the relative peak intensities (curves not shown), it was estimated that the protein sample was composed of 70% seleno-LuGST1-1 and 30% GST (S9C C86/200S) mutants. Similar yields were also reported for (Se) 2 -thioredoxin (17) and selenoglyceraldehyde-3-phosphate dehydrogenase (4).
Binding Experiments-For seleno-LuGST1-1, the dissociation constant of GSH at pH 6.5 was determined to be 104 M, which was approximately the same as that of wild type LuGST1-1, 107 M. This result suggests that substitution of the hydroxyl group of Ser-9 has no effect on the affinity toward GSH, which has also been proved by the mutant GST S9C (24).
Peroxidase and Transferase Activities of Seleno-LuGST1-1-The GPx and GST activities of seleno-LuGST1-1 and other catalysts are listed in Table II. The glutathione transferase activity of seleno-LuGST1-1 (23.5 mol⅐min Ϫ1 ⅐mol Ϫ1 ) was one order of magnitude lower than that of wild type LuGST1-1 (355.3 mol⅐min Ϫ1 ⅐mol Ϫ1 ) when 1-chloro-2,4-dinitrobenzene was used as the substrate, and the two GST mutants (S9C and S9C C86/200S) displayed no GST activities under the same conditions due to the removal of the active hydroxyl group of Ser-9, which could form a hydrogen-bond with the thiol group of GSH to stabilize the ionized GSH and therefore enhance the nucleophilicity of GSH (24).
The GPx activities of the wild type LuGST1-1, GST mutants (S9C and S9C C86/200S), and seleno-LuGST1-1 were estimated by a coupled enzyme system under the same conditions. LuGST1-1 and its two mutants exhibited no evident GPx activity, whereas the purified selenoenzyme displayed a remarkably activity of 2957 mol⅐min Ϫ1 ⅐mol Ϫ1 . Because the cysteinecontaining mutant (S9C C86/200S, 30% in the purified selenoenzyme) did not catalyze the reduction of hydroperoxide by GSH, all the activity of the purified selenoenzyme should be attributed to the seleno-LuGST1-1 (70% in the purified selenoenzyme). Accordingly, the high GPx activity of the purified selenoenzyme should result from the conversion of Cys-9 to Sec in the auxotrophic strain instead of the three mutations (S9C, C86S, and C200S). Seleno-LuGST1-1 was found to be much more efficient than most of other GPx mimics for the reduction of H 2 O 2 by GSH. For instance, it was at least 500-fold and 2900-fold more efficient than the first semisynthetic selenoenzyme, selenosubtilisin and the well known model compound ebselen (27), which exhibited GPx activities of only 4.6 mol⅐min Ϫ1 ⅐mol Ϫ1 and 0.99 mol⅐min Ϫ1 ⅐mol Ϫ1 , respectively (as shown in Table II). Its activity was even comparable with those of some natural GPxs, such as rabbit liver GPx (28), bovine liver GPx (29), human hepatoma HepG 2 cell giGPx (29), and human plasma pGPx (30) whose activities were in the order of 10 2 -10 3 mol⅐min Ϫ1 ⅐mol Ϫ1 .
The Optimal pH and Temperature for Seleno-LuGST1-1-catalyzed Reduction of Hydrogen Peroxide by GSH-The GPx activity of seleno-LuGST1-1 was examined over the pH range from 6.0 to 10.0 and the temperature range from 20.0 to 45.0°C. As shown in Fig. 2, pH and temperature profiles of seleno-LuGST1-1 activity were similar to those exhibited by natural GPx (30). Seleno-LuGST1-1 had an optimal pH of 9.0, which was close to that of natural GPx (8.8) (31), and a temperature optimum of 41.2°C, which was lower than that of natural GPx (50°C) (31). The activities of seleno-LuGST1-1 at 37°C and at pH 7.0 were only 30 and 13% of their maximums: 22,746 (at 41.2°C) and 9,857 mol⅐min Ϫ1 ⅐mol Ϫ1 (at pH 9.0), respectively.
Steady-state Kinetics and Catalytic Mechanism of Seleno-LuGST1-1-The initial rates for H 2 O 2 reduction catalyzed by seleno-LuGST1-1 were determined as a function of substrate concentration at 37.0°C and pH 7.0 when the concentration of one substrate was varied and that of the other was fixed. Michaelis-Menten kinetics was observed under all conditions investigated. Double reciprocal plots of the initial rate versus substrate concentration revealed the characteristic parallel lines (Fig. 3) of a ping-pong mechanism, in analogy with those of natural GPx (2). Treated with excess iodoacetate in the presence of GSH, the enzyme was found to completely lose GPx activity, suggesting the presentation of the enzyme-bound selenol in the catalytic cycle (2,32). The further characterizations of the intermediates and the molecular mechanism need to be fully established.
The kinetic parameters for the reactions between GSH and hydroperoxide catalyzed by seleno-LuGST1-1 were detailed in Table III. These values were deduced from fitting the experimental data to a ping-pong kinetic scheme. The relevant steady-state rate equation is shown as Equation 1.
(Eq. 1) The first-order rate constant k cat H 2 O 2 and the apparent Michaelis constant K mH 2 O 2 at 1 mM GSH were determined to be 2100 min Ϫ1 and 2.1 ϫ 10 Ϫ4 M, respectively. At 0.5 mM H 2 O 2 , K mGSH was 1.7 ϫ 10 Ϫ4 M. The apparent second-order rate constant k cat /K mH 2 O 2 and k cat /K mGSH provide measures of the rates of reactions between free enzyme and the relative substrates (GSH and hydrogen peroxide, respectively). k cat /K mGSH and k cat /K mH 2 O 2 of seleno-LuGST1-1 were both ϳ10 7 M Ϫ1 ⅐min Ϫ1 . k cat /K mGSH of seleno-LuGST1-1 was in the same order of magnitude as that of natural GPx (33,35), indicating  2. Effects of the catalytic conditions on the GPX activity for the reduction of H 2 O 2 by GSH catalyzed by seleno-LuGST1-1. A, relative GPX activity versus pH. B, relative GPX activity versus temperature. The activity was determined when the concentrations of GSH and H 2 O 2 were 1 mM and 0.5 mM, respectively. The GPX activity was converted to the relative value, and the activity detected at 37°C and pH7.0 was defined to be 100%. that they had similar affinities to GSH. Although k cat /K mH 2 O 2 of seleno-LuGST1-1 was still one order of magnitude lower than that of natural GPx (10 8 M Ϫ1 ⅐min Ϫ1 ) (35), it was much higher than those of most GPx mimics (for example, 4.5 ϫ 10 3 M Ϫ1 ⅐min Ϫ1 for selenium-containing catalytic antibody Se-4A4 (5)).
The remarkably high GPx activity of seleno-LuGST1-1 could be ascribed to its special enzyme scaffold. Studies on crystal structures of LuGST1-1 and natural GPx (9,11,14) reveal that there are some important structure similarities between these two enzymes. Both of the two enzymes include specific GSH binding sites (14) and thus could both strongly bind to their common specific substrate GSH, although their residues involved in the binding are different. As has been proved by binding experiments, the specific affinity of seleno-LuGST1-1 toward GSH was hardly weakened after the conversion of serine 9 to Sec-9 in its active site. Because the natural GPx couples the reduction of hydroperoxides with the oxidation of GSH in vivo, the ability to bind this thiol substrate is essential for GPx activity. This has been suggested by the catalytic efficiency difference among several GPx types. Cellular GPx contains a GSH binding site consisting of one lysine and four arginine residues, extracellular GPx lacks three of the arginine residues, and phospholipids hydroperoxide GPx lacks all of the five residues. As a result, cellular GPx, which has the strongest binding to GSH, exhibits the highest GPx activity (34). The favor for special thiol substrate of the first semisynthetic selenoenzyme, selenosubtilisin, which decomposes hydroperoxides with aryl thiols as reductant instead of GSH, is also due to the deficiency of the specific GSH binding site (35). Therefore, the specific GSH binding site of seleno-LuGST1-1 and its high GSH affinity should have contributed greatly to its significantly high GPx activity. Moreover, GPx and GST adopt a common thioredoxin fold, even though they share low sequence identity (9,11,14). Their active site residues locate in relatively the same positions of their thioredoxin folds. Furthermore, on alignment of their structures, their substrate GSH is apparently bound at the same accessible point by an interaction with the conserved hydroxyl moiety of active residue Ser in LuGST and with the selenol group of the catalytic Sec in GPx, respectively. Therefore, the enzyme scaffold of seleno-LuGST1-1 seems to provide its catalytic Sec with an optimal geometric conformation even comparable with that of naturally occurring GPx. In addition, the secondary structure adjacent to the active site in this enzyme model also seems to have a considerable contribution to enzyme catalysis. In the thioredoxin fold, the active site residue (Ser in LuGST1-1 or Sec in GPx) is positioned at the N-terminal end of helix ␣ 1 of a ␤␣␤ structure (11,36). In the ␣ helix, the alignment of the peptide dipoles parallel to the helix axis gives rise to a macrodipole with considerable strength. For the active residue point, the effect of the dipole is equivalent to the effect of half a positive unit charge (37). In natural GPx, the electric field due to the dipole moment of helix ␣ 1 could stabilize the active site selenolate and enhance its nucleophilic reactivity (36). We could imagine that the same effect may exist in the similar enzyme scaffold of the engineering seleno-LuGST1-1. Therefore, it is not surprising that this engineered selenoenzyme has high values for its apparent second-order rate constants comparable with those of naturally occurring GPx.
However, because naturally occurring GPx have evolved a perfect structure for the decomposition of hydroperoxide with near optimal efficiency, differences between the active site structures of LuGST and natural GPx may unavoidably cause the lower catalytic activity of seleno-LuGST1-1. The active site of natural GPx is found in a flat depression on the molecular surface, and no prominent clefts or crevices are visible at the active center in the crystals (36). Exposure of the catalytically active Sec residues at the molecular surface permits easy access of the substrate and thus results in the high reaction rate. The active site of LuGST1-1, however, is located in a deeper (20 Å) V-shaped cleft (11), and the active Ser-9, which is substituted by Sec in our system, is located at the base of this pocket and therefore is presumably less accessible to hydroperoxide. This may be partly responsible for the lower k cat /K mH 2 O 2 value of seleno-LuGST1-1 than that of natural GPx. In addition, all of the GPx family contains a catalytic triad in the active site consisting of selenocysteine, glutamine, and tryptophan residues (38) for the perfect balance between stabilizing a high energy species and lowering its energy. Deletion of two of them in the seleno-LuGST1-1 model may also cause a decrease of its efficiency to some extend. However, this also promises seleno-LuGST1-1 high potential to be redesigned on the basis of modeling studies and to enhance its reactivity and specificity by rational mutation.
GST and GPx have been proposed to have a "glutathionebinding protein" ancestor in previous studies on the basis of the similarities in their overall structures and the positioning of their important active-site residues despite their functional differences and low sequence identity (9,14). Endowing GST with efficient GPx activity by the conversion of the key Ser in the active site to Sec provides a new proof for the previous assumption on their evolution and suggests that their activesite residues would have evolved separately from their common thioredoxin-like ancestor to accommodate different functions along with evolution. This result leads further credence to the principles that the evolution of new catalytic activities involves the incorporation of new catalytic groups within an active site and the reservation of those groups necessary to catalyze the partial reaction common to all of them (7,8). In addition, it supports the notion that the dominant factor in the evolution of new enzymatic activities is chemistry in nature rather than binding specificity. This could also be deduced from some other enzyme superfamilies within which their members share a common structural scaffold but catalyze different overall reactions (8).
In all organisms investigated to date, Sec is encoded by a UGA opal codon, usually a stop codon. However, the presence of a downstream mRNA stem-loop structure, designated as the Sec insertion sequence, precludes termination of the polypeptide biosynthesis and promotes Sec incorporation into the nascent protein (39). Consequently, expression of a eukaryotic selenoprotein in transformed bacteria would require the presence of a bacterial-type Sec insertion sequence inside the open reading frame of the recombinant protein, which, however, could result in alteration of its amino acid sequence and may affect its function. So far, the only success of heterologous expression of genes coding for eukaryotic selenoproteins in bacteria is the expression of selenium-containing citrus phospholipid hydroperoxide glutathione peroxidase, but functional enzyme could not be purified to homogeneity (39). Chemical modification is an efficient method to incorporate Sec into protein. The successful chemical incorporation of functional Sec has been demonstrated by the first semisynthetic selenoenzyme, selenosubtilisin (3), and other high efficiency GPx-like biocatalysts, such as selenium-containing catalytic antibody (5) and imprinted proteins (6). But non-site-directed substitution hampers further structure-function characterization of as-generated proteins. In contrast, the method of using the auxotrophic expression system to incorporate Sec into protein could circumvent this problem efficiently. Mischarging of tRNA Cys with Sec in the cysteine auxotrophic strain when Cys was omitted from the growth medium allows the efficient biosynthetic substitution of cysteine residues by selenocysteine. 70 -80% substitution ratios and high level yields were achieved for all the three proteins that have been successfully expressed using this method: (Se) 2 -thioredoxin, selenoglyceraldehyde-3phosphate dehydrogenase, and seleno-LuGST1-1. Because of less structural modifications, the efficient substitution, and high product yields, this system is of high potential for performing mutational analysis that is unconceivable in animal cell system. It is also an excellent strategy either to yield and characterize other novel selenoproteins or to extend our understanding of the mechanisms and the evolution of selenoenzymes.
In conclusion, we have successfully engineered LuGST1-1 into seleno-LuGST1-1 by substitution of the catalytic residue Ser-9 by Sec using the auxotrophic expression system. Because of the important structure similarities in the specific GSH binding site and geometric conformation for the active Sec in thioredoxin fold between seleno-LuGST1-1 and naturally occurring GPx, the engineered selenoenzyme exhibited the high efficiency. Its GPx activity was in the same order of magnitude compared with that of naturally occurring GPx. This study provided a proof that both GST and GPx are evolved from a common "glutathione-binding protein" ancestor. We expect that the seleno-GST would offer a more suitable enzymatic model for a further understanding of the relationships between structure and function of GPx. In addition, this novel engineered selenoenzyme is also a prospective catalyst as an excellent antioxidant especially useful for both industrial and medical applications.