Original Contribution

Characterization and structural analysis of humanselenium-dependent glutathione peroxidase 4 mutantexpressed in Escherichia coli

Yang Yu a,1, Jian Song b,1, Xiao Guo a, Shuan Wang a, Xiao Yang a,Long Chen a, Jingyan Wei a,c,n

a College of Pharmaceutical Science, Jilin University, Changchun 130021, Chinab College of Electronic Science and Engineering, Changchun 130000, Chinac State Key Laboratory of Theoretical and Computational Chemistry, Institute of Theoretical Chemistry, Jilin University, Changchun 130000, China

a r t i c l e i n f o

Article history:Received 4 August 2013Received in revised form20 March 2014Accepted 21 March 2014Available online 28 March 2014

Keywords:AntioxidantsCys auxotrophic strainGlutathione peroxidase 4MutantsOverexpressionSelenoenzymesFree radicals

a b s t r a c t

Glutathione peroxidase 4 (GPx4) is a monomeric selenium-dependent glutathione peroxidase highlyexpressed in mammalian cells, which can reduce phospholipid hydroperoxides. However, it has beendifficult to express recombinant mammalian GPx4 in Escherichia coli because of the differences in theselenocysteine (Sec) incorporation machinery between eukaryotes and prokaryotes. In this study, anE. coli BL21(DE3)cys auxotrophic strain was used to express GPx4 mutants. We found that untargetedsubstitution of Cys-2, Cys-37, Cys-75, Cys-107, and Cys-148 with Sec led to loss of activity, suggesting thatmutation of any of these Cys residues in GPx4 could result in a structural change. Additionally, we foundthat the catalytic activity of GPx4 mutants increased as the number of noncatalytic Sec residuesdecreased, indicating that the negative effects could be mitigated by replacing these Cys residues withSer residues. A GPx4 mutant with all Cys residues converted to Ser exhibited a “Ping–Pong” mechanismand structure similar to that of native GPx4, indicating that it could act as a substitute for GPx4, whenheterologously expressing the protein in E. coli. This research provides an important foundation forbiosynthesis of selenium-dependent GPx4 mutants in E. coli.

& 2014 Elsevier Inc. All rights reserved.

Glutathione peroxidase 4 (GPx4; EC 1.11.1.12), also known asphospholipid hydroperoxide glutathione peroxidase, is a selenium-dependent GPx that was first purified from pig liver in 1982 by Ursiniet al. [1]. Human GPx4 consists of three different isoforms encodedby one gene, including a cytosolic isoform (c-GPx4), mitochondrialisoform, and nuclear isoform [2–4]. In addition to reducing smallperoxides like H2O2, GPx4 is also capable of reducing membrane-bound hydroperoxides and lipids, such as phospholipids and choles-terol hydroperoxides [5–9]. Over the course of the past few decades,it has become increasingly clear that GPx4 is far more than anantioxidant enzyme. Beyond its antioxidant activity, GPx4 plays anessential role in cellular differentiation during embryonic developmentand is also thought to be involved in spermatogenesis and a variety ofother cellular processes [10–12].

GPx4 is a monomeric enzyme containing a selenocysteine (Sec)at the active site [5]. The catalytically active Sec forms a “catalytictetrad” with Trp, Gln, and Asn, which is conserved in the entire

selenium-dependent GPx family [13]. The proposed catalytic cycleof GPx initiates with the oxidation of the catalytically activeselenol of GPx (E-SeH) by hydroperoxides, producing the corre-sponding selenenic acid form (E-SeOH). The E-SeOH intermediatereacts with glutathione (GSH) to generate a selenenyl sulfideintermediate, which then reacts with another GSH to regenerateE-SeH [14–16]. The E-SeOH intermediate is extremely unstableand thus far has not been clarified in the selenoproteins. Because,unlike other tetrameric GPx proteins, GPx4 lacks a surface-exposed loop domain, it could react with a large variety of proteinthiols when GSH becomes limited [17–20].

Sec is encoded by the codon UGA, which is usually decoded as astop codon. Additionally, because the mechanism of Sec incorporationdiffers from prokaryotes to eukaryotes, it is difficult to heterologouslyexpress mammalian GPx in Escherichia coli. Although a variety ofmimetics have been constructed, mass preparation of selenium-containing GPx in E. coli is still an effective way to analyze its biologicalfunction. One such method involves mischarging of tRNACys with Secin an E. coli BL21(DE3)cys auxotrophic strain [21–23].

In this study, a series of mutants in human c-GPx4 was pre-pared in E. coli BL21(DE3)cys. Various Cys residues in GPx4 weremutated to Ser to evaluate the impact that untargeted Sec has on

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/freeradbiomed

Free Radical Biology and Medicine

http://dx.doi.org/10.1016/j.freeradbiomed.2014.03.0320891-5849/& 2014 Elsevier Inc. All rights reserved.

n Corresponding author at: Jilin University, College of Pharmaceutical Science,1266 Fujin Road, Changchun 130021, China. Fax: þ86 431 85619252.

E-mail address: jingyanweijluedu@163.com (J. Wei).1 These authors contributed equally to this work.

Free Radical Biology and Medicine 71 (2014) 332–338

the structure and catalytic activity of this protein. This workresulted in the construction of a new GPx4 mutant with significantactivity.

Materials and methods

Bacterial strains, plasmids, and growth conditions

The bacterial strains and plasmids used in this research arelisted in Table 1. In site-directed mutagenesis and subcloningexperiments, E. coli cells carrying the mutant plasmids weregrown in Luria–Burtani medium at 37 1C. The synthetic medium“modified M9 minimal medium” (MM medium) was employed asdescribed previously for production of selenoprotein [23,24].MM medium was supplemented with 250 μg/ml L-methionine.

Construction of plasmids

The sequence of the human GPx4 gene was codon-optimizedfor E. coli expression, and ACA sequences in the GPx4 gene werereplaced while retaining its amino acid sequence. The codons forCys-10 (TGC) and Cys-66 (TGC) were changed to the codon for Ser(AGC) to avoid aggregation of the selenoenzyme. The codon-optimized GPx4 (C10/66S) gene was synthesized and cloned intothe NdeI/HindIII sites of pUC57 plasmid by Sangon Biotech Co. Ltd.(Shanghai, China), yielding plasmid pUC57-GPx4 (C10/66S).

The expression plasmid pCGPx4 (C10/66S) encoding humanGPx4 (C10/66S) was constructed by subcloning the NdeI/HindIIIfragment from pUC57-GPx4 (C10/66S) into the NdeI/HindIII sites ofthe pCold I vector (TaKaRa). Plasmids were prepared using theMiniBEST plasmid purification kit (TaKaRa). Mutations were gen-erated using the QuikChange site-directed mutagenesis kit (Stra-tagene). The plasmid pCGPx4 (C10/66S) was used as the templatefor the site-directed mutagenesis reactions and the mutagenicprimers used in this study are listed in Table 2. Mutagenesis PCRcontained 50 ng of template plasmid, 125 ng of each mutagenicprimer, 0.2 mM dNTP, 2.5 U of PfuTurbo DNA polymerase, and 1�reaction buffer in a volume of 50 μl. The PCR program usedconsisted of 30 s at 95 1C followed by 18 cycles of 30 s at 95 1C,1 min at 55 1C, and 14 min at 68 1C. One microliter of DpnIrestriction enzyme (10 U/μl) was added directly to the amplifica-tion reaction to digest the parental DNA template and select forthe synthesized DNA containing mutations. All mutated constructswere verified by DNA sequencing.

Overexpression and purification of GPx4 mutants

An E. coli BL21(DE3)cys auxotrophic strain was used in thisstudy to produce selenoenzyme via tRNACys misloading. pCGPx4and six mutant plasmids were prepared using the MiniBESTplasmid purification kit (TaKaRa). The BL21(DE3)cys auxotrophicstrain was cotransformed with pCGPx4 (or the other six mutantplasmids) and pMazF, which contains themazF gene. Production ofselenoprotein was carried out according to the previouslydescribed protocol with a slight modification [25]. A single colonyof E. coli BL21(DE3)cys carrying plasmids pMazF and pCGPx4(C10/66S) (or the other six mutant plasmids) was grown overnightat 37 1C in 10 ml of MM medium supplemented with ampicillin(100 mg/ml), kanamycin (25 μg/ml), chloramphenicol (25 mg/ml),and Cys (50 mg/ml) on a shaker. The overnight culture wasadded to 200 ml of MM medium supplemented with ampicillin(100 mg/ml), kanamycin (25 μg/ml), chloramphenicol (25 mg/ml),and Cys (50 mg/ml) and incubated at 37 1C for several hours. Whenthe optical density of the culture at 600 nm reached 0.5, theculture was chilled in an ice-water bath for 5 min and incubated at15 1C for 45 min. Cells were harvested by centrifugation (5000 gfor 5 min at 15 1C) and washed twice with 0.9% NaCl to remove Cysadded to the MM medium. The pellet was resuspended in 5 mlof MM medium supplemented with ampicillin (100 mg/ml),kanamycin (25 μg/ml), chloramphenicol (25 mg/ml), isopropyl-β-D-1-thiogalactopyranoside (1 mM), and Sec (600 μM). The culturewas incubated at 15 1C for 16 more hours on a shaker to produceselenoprotein, and the cells were collected by centrifugation(5000 g for 5 min at 15 1C). Purification of recombinant proteinsfrom E. coli was carried out using the immobilized metal-affinity

Table 1Strains and plasmids used in this study.

Strain/plasmid Relevant genotype or description Source or reference

StrainDH5α F� Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17 (rK� , mKþ)

phoA supE44 λ� thi-1gyrA96 relA1Invitrogen

BL21(DE3)cys BL21(DE3) selB::kan cysE51 [22,23]Plasmid

pUC57-GPx4 (C10/66S) pUC57 with GPx4 (C10/66S) fragment in NdeI/HindIII Sangon Biotech (Shanghai)pMazF MazF (mRNA interferase) expression plasmid, Cmr TakarapCold I Cloning and expression vector, Ampr TakarapCGPx4 (C10/66S) pCold I with human GPx4 (C10/66S) fragment in NdeI/HindIII This studypCGPx4 (C10/66/148S) pCold I with human GPx4 (C10/66/148S) fragment in NdeI/HindIII This studypCGPx4 (C10/66/107/148S) pCold I with human GPx4 (C10/66/107/148S) fragment in NdeI/HindIII This studypCGPx4 (C10/37/66/107/148S) pCold I with human GPx4 (C10/37/66/107/148S) fragment in NdeI/HindIII This studypCGPx4 (C10/37/66/75/107/148S) pCold I with human GPx4 (C10/37/66/75/107/148S) fragment in NdeI/HindIII This studypCGPx4 (C2/10/37/66/107/148S) pCold I with human GPx4 (C2/10/37/66/107/148S) fragment in NdeI/HindIII This studypCGPx4 (C2/10/37/66/75/107/148S) pCold I with human GPx4 (C2/10/37/66/75/107/148S) fragment in NdeI/HindIII This study

Table 2Primers used in this study.

Sequence 50 - 30

C2S-F GAAGGTAGGCATATGTCCGCGAGCCGCGATGATTGC2S-R CAATCATCGCGGCTCGCGGACATATGCCTACCTTCC37S-F GATAAATATCGCGGCTTTGTGAGCATCGTGACCAACGTGGCC37S-R GCCACGTTGGTCACGATGCTCACAAAGCCGCGATATTTATCC75S-F CATCCTGGCCTTTCCGTCCAACCAGTTTGGCAAGCC75S-R GCTTGCCAAACTGGTTGGACGGAAAGGCCAGGATGC107S-F GATATGTTTAGCAAAATCAGCGTGAACGGCGATGATGC107S-R CATCATCGCCGTTCACGCTGATTTTGCTAAACATATCC148S-F CTGATCGATAAAAACGGCTCCGTGGTGAAACGCTATGGCC148S-R GCCATAGCGTTTCACCACGGAGCCGTTTTTATCGATCAG

Bold font indicates the substitutions that resulted in mutations.

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338 333

chromatography purification system with Ni–NTA resin. Fractionscontaining recombinant proteins were dialyzed against 50 mMphosphate-buffered saline (PBS; pH 7.4) and confirmed by SDS–PAGE and Western blotting. Western blotting was performed asdescribed previously with small modifications [26]. Samples wereseparated on an SDS–PAGE gel and electrotransferred to nitrocel-lulose membranes. The membrane was blocked with 5% nonfatdry milk in Tween–Tris-buffered saline (TTBS; 100 mM Tris,pH 7.5, 150 mM NaCl, 0.2% Tween 20) at 37 1C for 1 h, followedby incubation with mouse monoclonal anti-His6 IgG (Sigma)at a dilution of 1:3000 at 37 1C for 1 h. The membrane waswashed with TTBS (3�5 min) and then incubated with horse-radish peroxidase-conjugated goat anti-mouse IgG at a dilution of1:3000 at 37 1C for 1 h. After three washes with TTBS, proteinbands were visualized by DAB staining. Total protein content wasestimated by the Bradford method using bovine serum albumin asa standard.

Mass spectrometric analysis

Matrix-assisted laser desorption time-of-flight mass spectro-metry (MALDI–TOF MS) analysis was performed using the ABSCIEX TOF/TOF 5800 system (AB SCIEX). The purified proteinsamples were dialyzed against water and prepared using a con-ventional dried-droplet protocol with sinapinic acid as the matrixto determine the molecular weight of intact selenoprotein. Forpeptide mass fingerprinting, proteins were digested with trypsinin a ratio of 1 μg of protease per 50 μg of protein at 37 1C for 15 h.The reaction was stopped with 1 μl of 10% (v/v) trifluoroacetic acidand tryptic fragments were analyzed by MALDI–TOF MS using α-cyano-4-hydroxycinnamic acid as the matrix.

Assay of enzyme activity

GPx activities of various GPx4 mutants were determined usinga previously described method [27]. PBS (50 mM, pH7.4), EDTA(1 mM), GSH (1–2 mM), NADPH (0.25 mM), glutathione reductase(1 U), and samples (2 μM) were mixed in a cuvette at 37 1C. Thereagent mixture was incubated for 3 min at 37 1C and the reactionwas initiated by addition of 60–500 μM peroxides, at a totalvolume of 0.5 ml. GPx activity was determined by measuring thedecrease in NADPH absorption at 340 nm per minute. Activityunits (U) are defined as the amount of enzyme necessary tooxidize 1 μmol of NADPH per minute at 37 1C and specific activityis expressed in U/mg. The specific activity of GPx4 mutant wasmeasured using 1-palmitoyl-2-(13-hydroperoxy-cis-9, trans-11-octadecadienoyl)-L-3-phosphatidylcholine (PLPC–OOH) as sub-strate, which was prepared by enzymatic hydroperoxidationof 1-palmitoyl-2-linoleoyl-L-3-phosphatidylcholine by lipoxidase,precisely as described [28].

Determination of optimum temperature and pH

The optimum temperature and pH for GPx activity of Se-GPx4(C2/10/37/66/75/107/148S) were determined by performing enzy-matic assays at various temperatures (15 to 50 1C) and pH levels(4.8 to 12). GPx activities were measured using H2O2 as anoxidizing substrate with the same method as described above.

Steady-state kinetics of Se-GPx4 (C2/10/37/66/75/107/148S)

Steady-state kinetics were carried out following the methoddescribed above [29]. Steady-state parameters for Se-GPx4 (C2/10/37/66/75/107/148S) were determined using the same assay withthe concentration of one substrate varied while the other waskept fixed. GPx activities were measured using PLPC–OOH as an

oxidizing substrate with the same method as described above at37 1C and pH 7.4. Kinetic data values were calculated from theDalziel coefficients as described previously [5,30,31].

Molecular modeling

Molecular modeling was performed with the Insight II package,version 2000 (Accelrys, San Diego, CA, USA) [32,33]. The X-raycrystal structure of the GPx4 mutant (Protein Data Bank ID: 2OBI)was obtained and used for the starting coordinates for calcula-tions. The three-dimensional (3D) structure of Se-GPx4 (C2/10/37/66/75/107/148S) was built and the model with the best value ofdiscrete optimized potential energy statistical potential wasselected [34]. The resulting model was refined by MD simulationsand analyzed with Profile-3D and Procheck following the methodas described previously [24,35,36]. The electrostatic potential wascalculated using the DelPhi module.

Results

Overexpression and purification of GPx4 mutants

To investigate the influence of multiple Sec residues onstructure and activity, Cys residues of GPx4 were mutated to Serresidues. Plasmids containing various mutations in the GPx4 geneare listed in Table 1. Sequencing analysis confirmed that the Cyscodons of GPx4 were changed to Ser codons successfully and thatno other mutations were introduced during mutagenesis. Thepurified selenoprotein showed a single band with a molecularmass of �20,000 Da, as determined by SDS–PAGE and Westernblotting (Fig. 1), indicating that GPx4 mutants were successfullyexpressed and purified from E. coli BL21(DE3)cys. The yield ofrecombinant selenoprotein was approximately 3.9 μg/ml of culture.

Fig. 1. (A) SDS–PAGE and (B) Western blotting analysis of various purified GPx4mutants. Lane M, marker; lane 1, Se-GPx4 (C10/66S); lane 2, Se-GPx4 (C10/66/148S); lane 3, Se-GPx4 (C10/66/107/148S); lane 4, Se-GPx4 (C10/37/66/107/148S);lane 5, Se-GPx4 (C10/37/66/75/107/148S); lane 6, Se-GPx4 (C2/10/37/66/107/148S);lane 7, Se-GPx4 (C2/10/37/66/75/107/148S).

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338334

Mass spectrometric analysis

MALDI–TOF MS analysis of intact Se-GPx4 (C10/66S) andSe-GPx4 (C2/10/37/66/75/107/148S) indicated a mass increase of378 Da revealing that Sec residues were introduced into Se-GPx4(C10/66S) successfully (Fig. 2A and B). Furthermore, MALDI–TOFMS analysis of Se-GPx4 (C2/10/37/66/75/107/148S) was performedafter digestion with trypsin and a peak at mass/charge 1592.28 ofthe peptide Gly-34–Lys-48 was detected. We observed a massincrease of 34.56 Da compared with the predicted mass/charge of1557.72 of the peptide, indicating that two oxygen atoms werebound to the selenium atom in Sec-46 (Fig. 2C). Thus, we foundthat the average substitution ratio of Sec for Cys residues wasapproximately 80%, with the selenoproteins at least partially inseleninic acid (RSeO2H) form.

Assay of enzyme activity

Fig. 3 illustrates that the catalytic activity of GPx4 mutantsincreases as the number of noncatalytic Sec residues decreases. Themutant Se-GPx4 (C2/10/37/66/75/107/148S) with all noncatalytic Secresidues replaced by Ser residues showed the highest activity, whichwas about 15-fold more efficient than Se-GPx4 (C10/66S) in reducingH2O2. To make a comparison of activities between Se-GPx4 (C2/10/37/66/75/107/148S) and human native GPx4, specific activity was thendetermined under the same conditions as described previously (2 mMGSH, 60 μM peroxide, pH 7.4, at 37 1C) [29]. The activity of this mutantwith H2O2 and PLPC–OOH as an oxidizing substrate was 14.0 and26.4 U/mg of protein, respectively. This activity was on the same orderof magnitude as that of human native GPx4 (34.8 and 64.8 U/mg, withH2O2 and PLPC–OOH, respectively).

Determination of optimum temperature and pH

High values of activities of Se-GPx4 (C2/10/37/66/75/107/148S)were observed at temperatures of 20–37 1C, with the maximum

activity at 20 1C. Activity of the mutant decreased at temperaturesabove 20 1C and it was almost completely inactivated at 50 1C(Fig. 4A). Activity of Se-GPx4 (C2/10/37/66/75/107/148S) wasmeasured in the pH range of 4.8 to 12. The optimum pH wasdetermined to be 9 and the relative activity was almost inactive atpH levels lower than 7 or higher than 11 (Fig. 4B).

Steady-state kinetics of Se-GPx4 (C2/10/37/66/75/107/148S)

Using double-reciprocal plots of [E0]/ν0 versus the reciprocalconcentration of PLPC–OOH at various fixed concentrations of GSHgave a set of parallel lines (Fig. 5A); we found that Se-GPx4 (C2/10/37/66/75/107/148S) follows a “Ping–Pong” mechanism as described

Fig. 2. MALDI-TOF MS analysis of the purified recombinants (A) Se-GPx4 (C2/10/37/66/75/107/148S) and (B) Se-GPx4 (C10/66S). (C) MALDI–TOF mass spectrum of trypticfragments of Se-GPx4 (C2/10/37/66/75/107/148S).

Fig. 3. Comparison of GPx activities of various GPx4 mutants. The activities weremeasured using H2O2 as oxidizing substrate. The concentrations of GSH and H2O2

were 1 and 0.5 mM, respectively.

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338 335

by a Dalziel equation (Eq. (1)):

½E�0=ν0 ¼Φ0þðΦ1=½PLPC–OOH�ÞþðΦ2=½GSH�Þ: ð1ÞIn the secondary plot (Fig. 5B), the apparent maximum velo-

cities for infinite PLPC–OOH concentrations were replotted againstthe reciprocal concentrations of GSH. The intercept approximates

to 0, indicating that the coefficient Φ0 approximates 0. Accord-ingly, Vmax and Km values of the selenoenzyme are infinite. Φ1 isthe reciprocal rate constant k0 þ1 for the net forward reaction ofreduced enzyme with PLPC–OOH. k0 þ1 is defined as kþ1 � k�1,and may be regarded as kþ1, because the partial reaction shouldbe irreversible. Φ2 is the reciprocal k0 þ2 for regeneration of thereduced enzyme by GSH. The rate constants kþ1 and k0 þ2 of theGPx4 mutant were 1.6�105 and 2.3�103 M�1 s�1, respectively,which were 2 and 1 order of magnitude lower than those ofhuman native GPx4.

Molecular modeling

The Profile-3D score of Se-GPx4 (C2/10/37/66/75/107/148S) wasdetermined to be 67.5, compared with the expected high score of 71.1and the expected low score of 32.0. Procheck Ramachandran plotanalysis showed 84.4% of residues in the core region, 15.6% in theallowed region, and no residues in the disallowed region. Theseresults suggested that the conformation of Se-GPx4 (C2/10/37/66/75/107/148S) was reliable. We found the catalytic tetrad consisting ofSec-46, Gln-81, Trp-136, and Asn-137 to be detectable in the GPx4mutant (Fig. 6A). The Se atom in Sec-46 lies in an exposed andpositively charged area (Fig. 6B), which might facilitate the interac-tion between Sec and the broad range of substrates.

Discussion

GPx catalyzes the reduction of hydroperoxides to their respec-tive alcohols and therefore protects cells and tissues from oxida-tive damage. Because the incorporation of Sec into proteinsrequires a special mechanism, very limited work has been donein the area of heterologous expression and purification ofselenium-dependent GPx with high activity from E. coli over thepast few decades. Although selenoproteins could be obtained byusing E. coli BL21(DE3)cys, all Cys residues in the protein would beinevitably replaced by Sec residues.

Results in Fig. 3 support our finding that the replacement of Cysby Sec residues has adverse effects on the reaction catalyzed byGPx4. There are seven Cys residues (Cys-2, Cys-10, Cys-37, Cys-66,Cys-75, Cys-107, and Cys-148) in native human GPx4. Because Cys-10 and Cys-66 had been considered to be related to proteinaggregation [17], the codons for these two Cys residues werechanged to Ser codons in the synthetic GPx4 gene. The remainingCys residues in various GPx4 mutants were substituted by Secowing to the mischarging of tRNACys with Sec in E. coli BL21(DE3)cys. Sec has a structure similar to that of Cys, except that in Sec, a

Fig. 4. Effects of temperature and pH on activity of Se-GPx4 (C2/10/37/66/75/107/148S). (A) Plot of GPx activity versus temperature. The activities were measured at thetemperature range of 15–50 1C. (B) Plot of GPx activity versus pH. The activities were measured in the pH range of 4.8–12. The enzyme activity was assumed as 100% at 37 1Cand at pH 7.4. The concentrations of GSH and H2O2 were 1 and 0.5 mM, respectively.

Fig. 5. Double-reciprocal plots for the reduction of PLPC–OOH by GSH catalyzed bySe-GPx4 (C2/10/37/66/75/107/148S). (A) [E]0/ν0 versus 1/[PLPC–OOH] (μM�1) at[GSH] of 0.5, 1.0, 2.0, and 3.0 mM. The slope corresponds to the reciprocal rateconstant kþ1. (B) Secondary plot of the apparent maximum velocities, calculatedfor each GSH concentration data set, versus the reciprocal GSH concentrations.The slope corresponds to the reciprocal rate constant k0 þ2.

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338336

selenium atom takes the place of the sulfur atom. However, it hasbeen reported in crystallization studies of bovine GPx1 and humanGPx3 that Sec in selenoproteins isolated from natural sources mayundergo air oxidation and exist as a seleninic acid (E-SeO2H) afterexposure to excess peroxides [37,38]. The same conclusion wassupported by the results of MALDI–TOF MS analysis in this work. Aseleninic acid group could be reduced back to selenol, and Sec atthe active center could participate in the catalytic reaction [39]. Toensure the regeneration of selenol, thiol is required as a reductant.Although a structure of a thiol–GPx4 intermediate has not beensolved to date, it is not surprising that similar reactions would notoccur in the case of the other Sec, as there is not another thiolbinding site nearby. Therefore, all Sec residues except Sec-46would be easily oxidized to seleninic acid by hydroperoxide inthe reaction mixture. Because of the presence of the two oxygenatoms bonded to the selenium atom, the structural differencesbetween Sec and Cys could not be ignored. These differenceswould eventually lead to structural changes of the protein andwould therefore affect the catalytic efficiency.

Considering the fact that, among the 20 common amino acids,Ser has the properties most similar to Cys, replacement of Cys bySer is predicted to cause less structural change than replacementby Sec. Therefore, it may be feasible to maintain the active-sitegeometry if Cys were replaced by Ser. To test this hypothesis, Cysresidues in native human GPx4 were mutated to Ser and variouscorresponding GPx4 mutants were prepared. We found that GPx4activity did increase as the number of noncatalytic Sec residuesdecreased. Accordingly, the mutant with all noncatalytic Secresidues converted to Ser showed the highest catalytic activity(Fig. 3). This mutant retained about 40% of the activity of humannative GPx4 isolated from human placental cytosol [29]. This ratiowas higher than those of other tetrameric GPx mutants with allCys converted to Ser, including a GPx1 mutant (about 28%) [26]and a GPx3 mutant (about 10%, data not shown). Because humanGPx4 is a single polypeptide chain protein and because no lipidposttranslational modifications, such as glycosylation, weredetected [40,41], a reduction in catalytic efficiency might beattributed to the substitution of Ser for Cys. Despite this relativelylow activity, steady-state kinetics results indicate that Se-GPx4(C2/10/37/66/75/107/148S) follows a Ping–Pong mechanism with-out apparent saturation, which is similar to native GPx4. Addi-tionally, a similar structure between this mutant and native GPx4was calculated (Fig. 6). These results suggest that the negativeeffects caused by replacement of Cys with Sec residues can bemitigated by substituting Ser for these Sec residues.

So far, direct structural information on human GPx is availableonly for GPx3 [38]. This is largely due to the lack of appropriatemethods available for obtaining sufficient protein with Sec in theactive center. In this work, large quantities of highly pureselenium-containing GPx4 mutants were obtained easily from anE. coli BL21(DE3)cys auxotrophic strain. The mutant and nativeGPx4 exhibited similar properties with respect to catalytic activity.Thus, it would be very helpful if detailed structural information onthis GPx4 mutant was obtained.

In conclusion, we have successfully purified a novel humanGPx4 mutant with high activity from E. coli BL21(DE3)cys byrational mutation, which may act as a substitute for native GPx4.By studying the relationship between activities of various GPx4mutants, we confirmed that substitution of noncatalytic Secresidues for Cys could cause a reduction in the catalytic activityof GPx4. The negative effects of untargeted introduction of Sec canbe minimized by replacing these Sec residues with Ser residues.This research is of great significance for heterologous expression ofmammalian selenium-dependent GPx4 mutants in E. coli and laysthe foundation for clarifying the catalytic mechanism of GPx4.

Acknowledgments

The authors thank Professor Marie-Paule Strub and AugustBock for providing the E. coli Cys auxotrophic strain, BL21(DE3)cys.This work is supported by the National Natural Science Foundationof China (Nos. 30970633 and 31270851).

References

[1] Ursini, F.; Maiorino, M.; Valente, M.; Ferri, L.; Gregolin, C. Purification from pigliver of a protein which protects liposomes and biomembranes from perox-idative degradation and exhibits glutathione peroxidase activity on phospha-tidylcholine hydroperoxides. Biochim. Biophys. Acta 710:197–211; 1982.

[2] Brigelius-Flohé, R.; Aumann, K. -D.; Blocker, H.; Gross, G.; Kiess, M.; Kloppel, K.;Maiorino, M.; Roveri, A.; Schuckelt, R.; Usani, F. Phospholipid–hydroperoxideglutathione peroxidase: genomic DNA, cDNA, and deduced amino acid sequence.J. Biol. Chem. 269:7342–7348; 1994.

[3] Arai, M.; Imai, H.; Sumi, D.; Imanaka, T.; Takano, T.; Chiba, N.; Nakagawa, Y.Import into mitochondria of phospholipid hydroperoxide glutathione perox-idase requires a leader sequence. Biochem. Biophys. Res. Commun. 227:433–-439; 1996.

[4] Pfeifer, H.; Conrad, M.; Roethlein, D.; Kyriakopoulos, A.; Brielmeier, M.;Bornkamm, G. W.; Behne, D. Identification of a specific sperm nucleiselenoenzyme necessary for protamine thiol cross-linking during spermmaturation. FASEB J. 15:1236–1238; 2001.

[5] Ursini, F.; Maiorino, M.; Gregolin, C. The selenoenzyme phospholipid hydro-peroxide glutathione peroxidase. Biochim. Biophys. Acta 839:62–70; 1985.

Fig. 6. 3D structure of Se-GPx4 (C2/10/37/66/75/107/148S). (A) Stereo view of the active site. (B) Electrostatic potential surface representation of the GPx4 mutant. The circlerepresents the location of the catalytic tetrad. Red denotes negative electrostatic potential and blue denotes positive electrostatic potential. (For interpretation of thereferences to color in this figure legend, the reader is referred to the web version of this article.)

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338 337

[6] Maiorino, M.; Roveri, A.; Ursini, F.; Gregolin, C. Enzymatic determination ofmembrane lipid peroxidation. J. Free Radic. Biol. Med. 1:203–207; 1985.

[7] Sattler, W.; Maiorino, M.; Stocker, R. Reduction of HDL- and LDL-associatedcholesterylester and phospholipid hydroperoxides by phospholipid hydroper-oxide glutathione peroxidase and ebselen (PZ-51). Arch. Biochem. Biophys.309:214–221; 1994.

[8] Maiorino, M.; Thomas, J. P.; Girotti, A. W.; Ursini, F. Reactivity of phospholipidhydroperoxide glutathione peroxidase with membrane and lipoprotein lipidhydroperoxides. Free Radic. Res. 12:131–135; 1991.

[9] Bao, Y.; Jemth, P.; Mannervik, B.; Williamson, G. Reduction of thyminehydroperoxide by phospholipid hydroperoxide glutathione peroxidase andglutathione transferases. FEBS Lett. 410:210–212; 1997.

[10] Imai, H.; Hirao, F.; Sakamoto, T.; Sekine, K.; Mizukura, Y.; Saito, M.; Kitamoto, T.;Hayasaka, M.; Hanaoka, K.; Nakagawa, Y. Early embryonic lethality caused bytargeted disruption of the mouse PHGPx gene. Biochem. Biophys. Res. Commun.305:278–286; 2003.

[11] Conrad, M.; Moreno, S.; Sinowatz, F.; Ursini, F.; Kolle, S.; Roveri, A.; Brielmeier, M.;Wurst, W.; Maiorino, M.; Bornkamm, G. The nuclear form of phospholipidhydroperoxide glutathione peroxidase is a protein thiol peroxidase contributingto sperm chromatin stability. Mol. Cell. Biol. 25:7637–7644; 2005.

[12] Conrad, M.; Schneider, M.; Seiler, A.; Bornkamm, G. W. Physiological role ofphospholipid hydroperoxide glutathione peroxidase in mammals. Biol. Chem.388:1019–1025; 2007.

[13] Tosatto, S. C.; Bosello, V.; Fogolari, F.; Mauri, P.; Roveri, A.; Toppo, S.; Flohé, L.;Ursini, F.; Maiorino, M. The catalytic site of glutathione peroxidases. Antioxid.Redox Signaling 10:1515–1526; 2008.

[14] Dalziel, K. Initial steady state velocities in the evaluation of enzyme–coen-zyme–substrate reaction mechanisms. Acta Chem. Scand. 11:7; 1957.

[15] Flohé, L.; Loschen, G.; Günzler, W. A.; Eichele, E. Glutathione peroxidase. V.The kinetic mechanism. Hoppe-Seyler’s Z. Physiol. Chem. 353:987–1000; 1972.

[16] Gunzler, W. A.; Vergin, H.; Muller, I.; Flohe, L. Glutathione peroxidase. VI.The reaction of glutathione peroxidase with various hydroperoxides. Hoppe-Seyler's Z. Physiol. Chem. 353:1001–1004; 1972.

[17] Scheerer, P.; Borchert, A.; Krauss, N.; Wessner, H.; Gerth, C.; Hohne, W.; Kuhn, H.Structural basis for catalytic activity and enzyme polymerization of phospholipidhydroperoxide glutathione peroxidase-4 (GPx4). Biochemistry 46:9041–9049;2007.

[18] Mauri, P.; Benazzi, L.; Flohé, L.; Maiorino, M.; Pietta, P. G.; Pilawa, S.; Roveri, A.;Ursini, F. Versatility of selenium catalysis in PHGPx unraveled by LC/ESI–MS/MS. Biol. Chem. 384:575–588; 2003.

[19] Godeas, C.; Tramer, F.; Micali, F.; Roveri, A.; Maiorino, M.; Nisii, C.; Sandri, G.;Panfili, E. Phospholipid hydroperoxide glutathione peroxidase (PHGPx) in rattestis nuclei is bound to chromatin. Biochem. Mol. Med. 59:118–124; 1996.

[20] Maiorino, M.; Roveri, A.; Benazzi, L.; Bosello, V.; Mauri, P.; Toppo, S.; Tosatto, S. C.;Ursini, F. Functional interaction of phospholipid hydroperoxide glutathioneperoxidase with sperm mitochondrion-associated cysteine-rich protein disclosesthe adjacent cysteine motif as a new substrate of the selenoperoxidase. J. Biol.Chem. 280:38395–38402; 2005.

[21] Mueller, S.; Senn, H.; Gsell, B.; Vetter, W.; Baron, C.; Boeck, A. The formation ofdiselenide bridges in proteins by incorporation of selenocysteine residues:biosynthesis and characterization of (Se) 2-thioredoxin. Biochemistry33:3404–3412; 1994.

[22] Sanchez, J. -F.; Hoh, F.; Strub, M. -P.; Aumelas, A.; Dumas, C. Structure of thecathelicidin motif of protegrin-3 precursor: structural insights into the

activation mechanism of an antimicrobial protein. Structure 10:1363–1370;2002.

[23] Strub, M. -P.; Hoh, F.; Sanchez, J. -F.; Strub, J. M.; Bock, A.; Aumelas, A.; Dumas, C.Selenomethionine and selenocysteine double labeling strategy for crystallo-graphic phasing. Structure 11:1359–1367; 2003.

[24] Yu, Y.; Song, J.; Song, Y.; Guo, X.; Han, Y.; Wei, J. Characterization of catalyticactivity and structure of selenocysteine-containing hGSTZ1c-1c based on site-directed mutagenesis and computational analysis. IUBMB Life 65:163–170;2013.

[25] Suzuki, M.; Mao, L.; Inouye, M. Single protein production (SPP) system inEscherichia coli. Nat. Protoc. 2:1802–1810; 2007.

[26] Guo, X.; Song, J.; Yu, Y.; Wei, J. Can recombinant human glutathioneperoxidase 1 with high activity be efficiently produced in Escherichia coli?Antioxid. Redox Signaling 20:1524–1530; 2014.

[27] Wilson, S. R.; Zucker, P. A.; Huang, R. R. C.; Spector, A. Development ofsynthetic compounds with glutathione peroxidase activity. J. Am. Chem. Soc.111:5936–5939; 1989.

[28] Maiorino, M.; Gregolin, C.; Ursini, F. Phospholipid hydroperoxide glutathioneperoxidase. Methods Enzymol. 186:448–457; 1990.

[29] Takebe, G.; Yarimizu, J.; Saito, Y.; Hayashi, T.; Nakamura, H.; Yodoi, J.;Nagasawa, S.; Takahashi, K. A comparative study on the hydroperoxide andthiol specificity of the glutathione peroxidase family and selenoprotein P.J. Biol. Chem. 277:41254–41258; 2002.

[30] Maiorino, M.; Aumann, K. -D.; Brigelius-Flohé, R.; Doria, D.; van den Heuvel, J.;McCarthy, J.; Roveri, A.; Ursini, F.; Flohé, L. Probing the presumed catalytictriad of selenium-containing peroxidases by mutational analysis of phospho-lipid hydroperoxide glutathione peroxidase (PHGPx). Biol. chem. Hoppe-Seyler376:651–660; 1995.

[31] Toppo, S.; Flohé, L.; Ursini, F.; Vanin, S.; Maiorino, M. Catalytic mechanismsand specificities of glutathione peroxidases: variations of a basic scheme.Biochim. Biophys. Acta 1790:1486–1500; 2009.

[32] Insight II, version 98.0. Accelrys, Inc., San Diego; 1998.[33] Discover 3 User Guide. Accelrys, Inc., San Diego; 1999.[34] Shen, M. y.; Sali, A. Statistical potential for assessment and prediction of

protein structures. Protein Sci. 15:2507–2524; 2006.[35] Profile-3D User Guide. Accelrys, Inc., San Diego; 1999.[36] Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thornton, J. M. PROCHECK: a

program to check the stereochemical quality of protein structures. J. Appl.Crystallogr 26:283–291; 1993.

[37] Epp, O.; Ladenstein, R.; Wendel, A. The refined structure of the selenoenzymeglutathione peroxidase at 0.2‐nm resolution. Eur. J. Biochem. 133:51–69; 1983.

[38] Ren, B.; Huang, W.; Åkesson, B.; Ladenstein, R. The crystal structure of seleno-glutathione peroxidase from human plasma at 2.9 Å resolution. J. Mol. Biol.268:869–885; 1997.

[39] Briviba, K.; Kissner, R.; Koppenol, W. H.; Sies, H. Kinetic study of the reactionof glutathione peroxidase with peroxynitrite. Chem. Res. Toxicol. 11:1398–-1401; 1998.

[40] Roveri, A.; Maiorino, M.; Nisii, C.; Ursini, F. Purification and characterization ofphospholipid hydroperoxide glutathione peroxidase from rat testis mitochon-drial membranes. Biochim. Biophys. Acta 1208:211–221; 1994.

[41] Morandat, S.; Bortolato, M.; Nicol, F.; Arthur, J. R.; Chauvet, J. -P.; Roux, B.Adsorption of a phospholipid–hydroperoxide glutathione peroxidase intophospholipid monolayers at the air–water interface. Colloids Surf., B 35:99–-105; 2004.

Y. Yu et al. / Free Radical Biology and Medicine 71 (2014) 332–338338