bacterial scavengase p20 is structurally and functionally related to peroxiredoxins

BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS 233, 848–852 (1997)ARTICLE NO. RC976564

Bacterial Scavengase p20 Is Structurally and FunctionallyRelated to Peroxiredoxins

Yuan Zhou,1 Xiao-Yu Wan,2 Hai-Lin Wang, Zi-Ying Yan, Yun-De Hou, and Dong-Yan Jin3

National Key Laboratory of Molecular Biology and Genetic Engineering, Chinese Academy of Preventive Medicine,100 Yingxin Street, Beijing 100052, People’s Republic of China

Received March 2, 1997

large family. It has been shown that AhpC and TSAScavengase p20 was recently identified as a novel fam- possess thiol peroxidase activity which is involved with

ily of bacterial antioxidant enzymes possessing thiore- their antioxidant properties to protect glutamine syn-doxin-linked thiol peroxidase activity. In this study, the thetase (GS) from inactivation by metal-catalysed oxi-Escherichia coli gene coding for scavengase p20 was iso- dation (3-5). Thioredoxin was identified as the immedi-lated from three different strains and the nucleotide se- ate bioactive hydrogen donor and the conserved N-ter-quence was determined. Multiple alignment of amino minal cysteine as the active center in the catalysis ofacid sequence revealed that a previously unidentified peroxide reduction by AhpC or TSA (4, 6-9).Cys-61 is most conserved among all bacterial p20 scaven- Besides all these known bacterial antioxidants, agases and corresponds to the active site in the well-char-

novel thioredoxin-linked thiol peroxidase p20 was re-acterized peroxiredoxins. Phylogenetic analysis furthercently identified from the periplasmic space of E. colisupported that scavengase p20 is a novel subfamily of(5). Further studies have revealed that this enzymeperoxiredoxins. Site-directed mutagenesis studies dem-represents a novel group of biologically important anti-onstrated that Cys-61 is indispensable for the antioxi-oxidant enzymes widely distributed in most bacteriadant activities of scavengase p20. Taken together, ourincluding Haemophilus influenzae, Streptococcus spp.,findings strongly suggest that the p20 scavengases areVibrio cholarae and Helibacter pylori (10,11). Althoughstructurally and functionally related to peroxiredoxins.this enzyme possesses GS protection and thioredoxin-q 1997 Academic Pressdependent thiol peroxidase activities, no significant ho-mology to any known peroxidase was found in two inde-pendent studies (5, 10). This enzyme was originallydesignated as thiol peroxidase (5). In order to distin-Aerobic organisms must combat with reactive oxygenguish this structurally distinct antioxidant enzymespecies (ROS) generated intrinsically in metabolism orfrom other known thiol peroxidases such as AhpC andexternally by the environment. The protection fromTSA, we proposed to rename it as ‘‘scavengase’’ (10).ROS toxicity is largely accomplished by antioxidant en-Sequence alignment of multiple bacterial p20 scaven-zymes which decompose peroxides or superoxide anion.gases has identified Cys-95 as one of the most con-These ubiquitous and conserved enzyme systems areserved residues (10, 11) and mutagenesis directed toessential to all living beings from bacteria to human.this site has revealed its essential role (11).Well-documented antioxidant enzymes include su-

In the present study, we extended our previous findingsperoxide dismutase (SOD), catalase and glutathione by sequence analysis, biochemical characterization andperoxidase (1). Moreover, a rapidly expanding family site-directed mutagenesis of E. coli scavengase p20. Sur-of abundant antioxidant enzymes called peroxiredoxin prisingly, we found that Cys-61, which was unidentifiedwas also discovered (2). Bacterial alkyl hydroperoxide in the original sequence reported by another group (5),reductase C22 (AhpC) and mammalian thiol-specific was the most conserved residue among all bacterial p20antioxidant (TSA) are two prototype enzymes in this scavengases. We also showed that scavengase p20 shared

distant but significant homology with peroxiredoxins andthat Cys-61, which is highly conserved in all scavengases1 Present address: Department of Biology, Georgetown University,and peroxiredoxins, is necessary for function.Washington DC 20057.

2 Present address: Johns Hopkins University Oncology Center,MATERIALS AND METHODSBaltimore, MD 21231.

3 Corresponding author. Fax: (10) 63541221. E-mail: yanzy@ccs. Bacterial strains. E. coli strains DH5a, HB101 was purchasedfrom Life Technologies. E. coli K-12 RR1 was originally obtainedcapm.ac.cn.

0006-291X/97 $25.00Copyright q 1997 by Academic PressAll rights of reproduction in any form reserved.

848

AID BBRC 6564 / 6929$$1061 04-10-97 14:51:14 bbrcgs AP: BBRC

Vol. 233, No. 3, 1997 BIOCHEMICAL AND BIOPHYSICAL RESEARCH COMMUNICATIONS

FIG. 1. Nucleotide and deduced amino acid sequence of E. coli scavengase p20 gene. Numbering starts from the translation initiationsite. The nucleotide and the amino acid residue different from the previously reported sequence (GenBank U33213) was double- and single-underlined, respectively. This sequence was deposited in GenBank under accession number U93212.

from American Type Culture Collection (ATCC). Genotypes of these Antioxidant assays. GS protection assay was carried out at 37 7Cessentially as described (10). The reaction mixture (100ml) containsstrains have been described elsewhere (12). An engineered E. coli

DH5a strain with a defective scavengase gene was constructed by 5mg E. coli GS (Sigma), 10mM FeCl3, 10mM dithiothreitol (DTT),5mg purified scavengase p20 or p20/S61 and 100mM Hepes (pH 7.0).homologous recombination. The construction of a similar strain has

been previously described (11). Aliquots (15ml) of reaction were removed at certain time points andassayed for GS activity.DNA cloning. E. coli genomic DNA was isolated by lysis with 1%

Indirect peroxidase assay was performed as described (10). TheSDS, deproteinization with 5M sodium perchloride and extractionreaction mixture (0.5ml) contains E. coli thioredoxin reductasewith chloroform. Scavengase gene was amplified by PCR using stan-(20mg), E. coli thioredoxin (20mg), H2O2 (5mM), purified scavengasedard procedures (12). Primers with the following sequence was usedp20 or p20/S61 (20mg) and NADPH (0.25mM). NADPH oxidationin PCR amplification: 5*-GCGGAATTCATGTCACAAACCGTTCAT-coupled to H2O2 reduction was monitored as absorbance units atTTC-3 * (sense), and 5*-AAAACTGCAGTTATGCTTTCAGTACAG-340nm (AU340). Peroxidase activity was expressed as AU340/min.CC-3 * (antisense). The amplified DNA was cloned via EcoRI and

In vivo alkyl hydroperoxide resistance assay was performed in E.PstI restriction sites into plasmid pUC19 and then subcloned intocoli DH5a transformants. Resistance to cumene hydroperoxide waspKK223-3 (Pharmacia) via the same sites.examined after 2 hour induction by 5mM IPTG and 15min treatment

DNA sequencing and sequence analysis. Double-stranded tem- with 0.2mM cumene peroxide. Percent survival was assayed by plat-plate was sequenced on both strands by the dideoxy method using ing of diluted liquid culture on selective medium.Sequenase 2.0 (United States Biochemical) as per manufacturer’sprotocol. Nucleotide and peptide sequence was analyzed with thehelp of the Wisconsin software package (Version 8.1, Genetic Com- RESULTS AND DISCUSSIONputer Group, Inc) (13) and the PHYLIP package (14). Multiple align-ment of protein sequence was generated with a progressive pairwise Molecular cloning and sequencing of E. coli scaven-algorithm (15). Phylogenetic analysis was based on a matrix of evolu-

gase gene. In an attempt to extend our previous stud-tionary distances and the phylogeny was reconstructed with theies on bacterial p20 scavengases, we obtained a molecu-neighbor-joining method.lar clone of E. coli scavengase gene by PCR cloningOligonucleotide-directed mutagenesis. Site-directed mutagenesisfrom from E. coli DH5a genomic DNA. When we se-was performed with a Clontech transformer kit based on an estab-

lished protocol (16). Sequence of the mutagenic oligonucleotide is 5*- quenced the complete coding region of this clone, weACTGATGCGGCGGAAACACCGGTAT-3 *. found surprisingly that our gene differs from the pre-

Protein expression and purification. Wild type and mutant sca- viously reported sequence (5; GenBank accession num-vengase genes were subcloned into prokaryotic expression vector ber U33213) in one position, namely nucleotide 181pKK223-3 (Pharmacia) and were overexpressed in E. coli upon induc- (double-underlined in Fig. 1; numbering starts fromtion by 5mM IPTG. Recombinant scavengase was purified as de-

the translation initiation site).scribed (10). The purity of the final product is more than 95% asexamined by SDS-PAGE (see Fig.4 below). In order to rule out the possibility of PCR error, we

849



FIG. 3. Distance matrix tree for scavengases and peroxiredoxins.See legend to Fig. 2 for titles and GenBank accession numbers ofthe sequences being analyzed.

Sequence comparison and phylogenetic analysis ofbacterial scavengases p20: homology to peroxiredoxins.Sequence alignment of scavengases with representa-tive peroxiredoxins was generated by use of a progres-sive pairwise algorithm (15). As shown in Fig. 2,bacterial p20 scavengases shared distant but signifi-cant sequence homology with all different types of per-oxiredoxins including the two prototype enzymes AhpCand TSA, and the poorly characterized bacterioferritincomigratory protein (BCP). This homology distributesthroughout the full-length sequence of scavengase p20.It is noteworthy that Cys-61 is one of the three mostconserved residues (highlighted by ‘‘#’’ in Fig. 2) sharedby all scavengases and peroxiredoxins, while the con-

FIG. 2. Sequence alignment of p20 scavengases and representa-tive peroxiredoxins. Residues similar in 7 out of 9 sequences arehighlighted by asterisks (*). Residues identical in all sequences aremarked by ‘‘#’’. Listed below are titles and GenBank accession num-bers of the sequences being compared: ecp20, E. coli scavengase p20,U93212; mtp20, Mycobacterium tuberculosis scavengase p20,Z70692; spp20, S. pneumoniae scavengase p20, U40786; hpp20, H.pylori scavengase p20, X72618; rntsa, R. norvegicus (rat) TSA,U06099; sctsa, Saccharomyces cerevisiae (budding yeast) TSA,L14640; stahpc, Salmonella typhimurium AhpC, J05478; hibcp; H.influenzae BCP, U32711; mtbcp, M. tuberculosis BCP, Z70692.

cloned the scavengase gene from two more E. colistrains HB101 and K-12 RR1. For each of the threescavengase genes, at least two independent clones weresequenced and were shown to be identical (Fig. 1).Thus, residue G-181 is authentic to the E. coli scaven-gase gene. Accordingly, residue 61 of E. coli scavengaseshould be cysteine (underlined in Fig. 1) rather than

FIG. 4. SDS-PAGE (15%) analysis of preparations of wild typetyrosine. This merits further investigation since cys- E. coli scavengase p20 and the Ser-61 mutant (p20/S61). Shown onteine residues are crucial for peroxide reduction cata- the right are positions of the molecular weight markers. Lane 1

contains 5mg wild type scavengase p20, lane 2 contains 5mg p20/S61.lyzed by many known thiol peroxidases (4-9).

850



FIG. 5. Antioxidant activities of wild type (p20) and mutant (p20/S61) E. coli scavengase. (A) GS protection activity. 1mM EDTA was addedto control 1 (ctrl 1) to chelate the catalyst Fe3/. Control 2 (ctrl 2) contains no scavengase p20. Results are representative of triplicate determinations.The standard deviation of each point is below 15%. (B) Thioredoxin-linked peroxidase activity. No scavengase p20 was added to the control.Results are representative of triplicate experiments. Error bars indicate the standard error of the mean. (C) In vivo alkyl hydroperoxide resistanceactivity. E. coli DH5a was transformed individually with empty vector (pKK223-3), plasmid expressing wild type p20 and plasmid expressingp20/S61. Resistance to cumene hydroperoxide was measured as percent survival after treatment. No scavengase p20 was added to the control.Results are representative of triplicate experiments. Error bars indicate the standard error of the mean.

servation of Cys-95 in bacterial scavengases does not dant properties similar to those described for peroxir-extend to peroxiredoxins. Furthermore, the conserved edoxins AhpC and TSA (5, 10). E. coli scavengase p20Cys-61 in scavengase p20 corresponds to Cys-46 in has been further characterized biochemically and ge-AphC and Cys-47 in TSA, which have been character- netically and shown to be actively involved in theized as the redox center residues reacting primarily oxidative stress response in this bacterium (11). Thewith hydroperoxides (3,8,9). structural basis for the functional similarity between

To better define the relatedness of bacterial scaven- scavengase and peroxiredoxin is not established.gases to peroxiredoxins, we performed a phylogenetic However, the sequence homology and the phyloge-analysis based on their amino acid sequences. In the netic relatedness described here are suggestive of aphylogenetic tree based on evolutionary distances (Fig. link between structure and function. Therefore, it3), the group of bacterial p20 scavengases (ecp20, would be of interest to see whether the conservedmtp20, hpp20, and spp20) clusters first with bacterial Cys-61 in scavengase is important to its antioxidantBCPs (hibcp and mtbcp) and the then with eukaryotic functions.TSAs (rntsa and sctsa). Finally, these clusters join with

Cys-61 is essential for antioxidant activities of sca-the bacterial AhpC (staphc). This strongly suggestsvengase p20. To access the role of Cys-61, we con-that bacterial scavengase p20 is a novel subfamily ofstructed a mutant strain of E. coli DH5a in whichperoxiredoxins.the endogenous scavengase gene was deleted. A simi-It has been shown that bacterial scavengase p20

possesses thiol peroxidase activity and other antioxi- lar tpx null strain was described elsewhere (11). Site-

851



specific mutagenesis directed to Cys-61 was also per- lecular mechanisms underlying its catalytic activi-ties. In the apparent absence of interchain disulfides,formed and a gene engineered to encode Ser-61 mu-

tant of scavengase p20 (abbreviated as p20/S61 here- it would be of particularly great interest to test thepossible formation and roles of intrachain disulfideafter) was generated.

The wild type scavengase and the mutant p20/S61 bonds in this protein.were overexpressed in parallel in the mutant E. colistrain defective for endogenous scavengase. Recombi- ACKNOWLEDGMENTSnant scavengase proteins were obtained and purified

The work was supported in part by Chinese Academy of Preventiveto homogeneity as shown by SDS-PAGE (Fig. 4). Sub-Medicine.sequently, the purified proteins were assayed for GS

protection and hydroxide reduction activities. As weREFERENCESexpected, the wild type scavengase p20 was fully com-

petent in both assays whereas the p20/S61 mutant 1. Sies, H. (1993) Eur. J. Biochem. 215, 213–219.was completely dead (Fig. 5, A and B). This result 2. Chae, H. Z., Robison, K., Poole, L. B., Church, G., Storz, G., andwas further confirmed in an in vivo system in which Rhee, S. G. (1994) Proc. Natl. Acad. Sci. USA 91, 7017–7021.functional wild-type scavengase protein accounts for 3. Netto, L. E. S., Chae, H. Z., Kang, S.-W., Rhee, S. G., and Stadt-

man, E. R. (1996) J. Biol. Chem. 271, 15315–15321.the increased resistance of E. coli stains to cumene4. Poole, L. B., and Ellis, H. R. (1996) Biochemistry 35, 56–64.hydroperoxide (Fig. 5C). Collectively, these data5. Cha, M.-K., Kim, H. K., and Kim, I.-H. (1995) J. Biol. Chem.strongly support that the p20/S61 mutant was defec-

270, 28635–28641.tive for antioxidant activities and that Cys-61 is es-6. Chae, H. Z., Uhm, T. B., and Rhee, S. G. (1994) Proc. Natl. Acad.sential for the proper function of bacterial scaven-

Sci. USA 91, 7022–7026.gase p20.7. Kwon, S. J., Park, J. W., Choi, W. K., Kim, I.-H., and Kim, K.Previous studies have suggested an essential role

(1994) Biochem. Biophys. Res. Commun. 201, 8–15.of Cys-95 for antioxidant function of E. coli scaven- 8. Poole, L. B. (1996) Biochemistry 35, 65–75.gase p20 (11). However, Cys-95 is unique to bacterial 9. Chae, H. Z., Chung, S. J., and Rhee, S. G. (1994) J. Biol. Chem.scavengases as shown in Fig.2. Taken together with 269, 27670–27678.data presented here, this may suggest that scaven- 10. Wan, X.-Y., Zhou, Y., Yan, Z.-Y., Wang, H.-L., Hou, Y.-D., andgase would act in a manner slightly different from Jin, D.-Y. (1997) FEBS Lett. in press.

11. Cha, M.-K., Kim, H. K., and Kim, I.-H. (1995) J. Biol. Chem.that of AhpC and TSA. We also note that in contrast270, 28635–28641.to AhpC and TSA that form homodimers, scavengase

12. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecularp20 did not self-associate in a non-reducing gel (5).Cloning: A Laboratory Manual, Cold Spring Harbor LaboratoryFor AhpC and TSA, the two interchain disulfide brid-Press, Cold Spring Harbor, NY.

ges of the dimeric protein, which necessarily involve 13. Devereux, J., Haeberli, P., and Smithies, O. (1984) Nucl. Acids.both N-terminal and C-terminal cysteines (Cys-46 Res. 12, 387–395.and Cys-165 for AhpC, Cys-47 and Cys-180 for TSA), 14. Felsenstein, J. (1996) Methods Enzymol. 266, 418–427.have been shown as the redox-active center in the 15. Feng, D.-F., and Doolittle, R. F. (1987) J. Mol. Evol. 25, 351–catalysis of peroxide reduction (3, 8). For scavengase 360.

16. Zhu, L. (1995) Methods Mol. Biol. 57, 13–29.p20, further studies are required to elucidate the mo-

852


bacterial scavengase p20 is structurally and functionally related to peroxiredoxins

Documents