Molecular basis of H2O2 resistance mediated by Streptococcal Dpr. Demonstration of the functional involvement of the putative ferroxidase center by site-directed mutagenesis in Streptococcus suis.

H(2)O(2) is an unavoidable cytotoxic by-product of aerobic life. Dpr, a recently discovered member of the Dps protein family, provides a means for catalase-negative bacteria to tolerate H(2)O(2). Potentially, Dpr could bind free intracellular iron and thus inhibit the Fenton chemistry-catalyzed formation of toxic hydroxyl radicals (H(2)O(2) + Fe(2+) --> (.)OH + (-)OH + Fe(3+)). We explored the in vivo function of Dpr in the catalase- and NADH peroxidase-negative pig and human pathogen Streptococcus suis. We show that: (i) a Dpr allelic exchange knockout mutant was hypersensitive ( approximately 10(6)-fold) to H(2)O(2), (ii) Dpr incorporated iron in vivo, (iii) a putative ferroxidase center was present in Dpr, (iv) single amino acid substitutions D74A or E78A to the putative ferroxidase center abolished the in vivo iron incorporation, and (v) the H(2)O(2) hypersensitive phenotype was complemented by wild-type Dpr or by a membrane-permeating iron chelator, but not by the site-mutated forms of Dpr. These results demonstrate that the putative ferroxidase center of Dpr is functionally active in iron incorporation and that the H(2)O(2) resistance is mediated by Dpr in vivo by its iron binding activity.

Gram-positive bacteria, including the important human pathogens Streptococcus pyogenes and Streptococcus pneumoniae (11), relatively little is known of the molecular mechanisms of its action. The primary amino acid sequence shares similarity with Escherichia coli Dps (DNA-binding protein from starved cells) (12,13), a prototype for a large group of similar oligomeric proteins (14), which is abundantly expressed in starved cells and involved in DNA protection by DNA-Dps biocrystal formation (15,16). Studies on Dps family members also indicate that the protective function of Dpr against H 2 O 2 might be mediated by H 2 O 2 degradation due to a catalase-like activity (14) or by chelation of free intracellular iron (17).
It is known that toxicity of H 2 O 2 is relatively weak (3), although it easily diffuses across biological membranes (18,19) and oxidize thiols (3). However, if reduced transition metal ions, especially iron, are present, H 2 O 2 is nonenzymatically cleaved into highly toxic hydroxyl radicals by Fenton chemistry (H 2 O 2 ϩ Fe 2ϩ 3 ⅐ OH ϩ -OH ϩ Fe 3ϩ ) (20,21). Dps family members share a conserved amino acid motif, similar to mammalian ferritin L-subunit iron nucleation center, in the N-terminal halves of the proteins (14,(22)(23)(24). Because several members, including the streptococcal Dpr (11,25), are known to bind iron (17,22,26,27), the conserved motif is believed to serve a functional role. It has been suggested that the motif catalyzes iron oxidation by its putative ferroxidase activity and also directs the formation of an iron core into the inner cavity of the oligomer (17,22,28,29). However, there is no direct experimental evidence for any of the Dps family members to support the functionality of the putative ferroxidase center in iron incorporation in vivo. Furthermore, the biological significance of iron incorporation is not well established.
Streptococcus suis is an important pig pathogen that causes severe infections such as sepsis and meningitis, and it occasionally causes life-threatening disease also in humans (30). We have previously identified in S. suis a galactose-specific adhesion activity (31)(32)(33). One of the proteins identified as a candidate adhesin displaying binding activity to glycoproteins had a 64% primary amino acid sequence identity with S. mutans Dpr. 1 S. suis seemed an ideal model organism to study the in vivo function of Dpr in H 2 O 2 resistance because it not only lacks catalase but also lacks NADH peroxidase (34). The results of the present study demonstrate that the putative ferroxidase center of Dpr is involved in iron incorporation and that the H 2 O 2 resistance mediated by Dpr depends on its iron incorporation activity in vivo. * This study was supported by the grants from the Sigrid Jusélius Foundation and the Academy of Finland. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AY154459.

Bacterial Strains, Plasmids, and Primers
The bacterial strains, plasmids, and primers used in this study are listed in Tables I and II. S. suis was cultured aerobically or under 6% CO 2 at 37°C with or without agitation in Todd Hewitt Broth medium (DIFCO) supplemented with 0.5% (w/v) yeast extract (Biokar Diagnostics) (THY). 2 E. coli was grown aerobically at 37°C with agitation in Luria-Bertani medium. When needed, media were solidified by 1.5% agar. All bacteria were stored at Ϫ70°C in growth media containing 15% (v/v) glycerol. Antibiotics (Sigma) were used at the following concentrations unless otherwise indicated: (i) E. coli, 100 g/ml ampicillin, 30 g/ml kanamycin, 100 g/ml spectinomycin, and 30 g/ml chloramphenicol; and (ii) S. suis, 20 g/ml ampicillin, 500 g/ml kanamycin, 1000 g/ml spectinomycin, and 10 g/ml chloramphenicol.

DNA Techniques
Genomic DNA of S. suis was isolated as described previously (38). Standard protocols were used for PCR, DNA modification, cloning, E. coli transformation, and Southern and colony hybridization as described by Sambrook and Russel (39). DNA-modifying enzymes were purchased from Promega and Fermentas, and Vent DNA polymerase was purchased from New England Biolabs. DNA molecular mass markers were from Promega. Plasmids amplified in E. coli DH5␣ were isolated using QIAprep Spin Miniprep Kit (Qiagen) as described by the manufacturer. DNA fragments were purified from agarose gels using QIAquick Gel Extraction Kit or from PCR and other enzymatic reactions using QIAquick PCR Purification Kit as described by the manufacturer (Qiagen). DNA sequences were determined by ABI Prism Big Dye Terminator Cycle Sequencing Ready Reaction Kit (PerkinElmer Life Sciences) with AmpliTaq DNA Polymerase FS (Roche Molecular Biochemicals). Sequencing primers were purchased from Interactiva Biotechnologie GmbH. For Southern, Northern, and colony hybridization, radioactive DNA probes were labeled with [␣-32 P]CTP (Amersham Biosciences) using Prime-a-Gene Labeling System (Promega) according to the instructions of the manufacturer. Templates for labeling were generated by PCR using primers DPR-5Ј-IN and DPR-3Ј-IN specific for dpr and 16S-5Ј and 16S-3Ј specific for 16S rRNA genes and gel-isolated before labeling. The hybridized membranes were analyzed with Fujifilm BA-2500 Phosphor Imaging Plate System (Fuji Photo Film Co.) according to the instructions of the manufacturer.

Cloning and Sequence Analysis of the dpr Locus
Southern hybridization was used to identify a 6.5-kb ClaI-EcoRI genomic fragment that seemed to contain the entire dpr from S. suis serotype 2 strain 628 (data not shown). To clone the 6.5-kb fragment, we ligated gel-isolated, ClaI-and EcoRI-digested DNA fragments ranging from 6 to 7 kb into EcoRI-NarI-digested pBR322, electrotransformed the ligation mixture into E. coli DH5␣ cells, and screened the transformants by colony hybridization. One hybridizing clone, designated pDPR6500, was isolated and sequenced. Sequencing data were assembled, and the consensus sequence was edited using the MAC DNAsis software (Hitachi). The web-based program ORF Finder (www.ncbi. nlm.nih.gov/gorf/gorf.html) was used to predict the coding regions. The BLAST software package at the National Center for Biotechnology Information (www.ncbi.nlm.nih.gov) was used to search for protein sequences homologous to the deduced amino acid sequences. Promoter sequence features were searched using WWW Signal Scan (bimas. dcrt.nih.gov/molbio/signal).

Electrotransformation of S. suis
Overnight S. suis cultures were diluted 100-fold into 50 ml of fresh THY supplemented with 30 mM glycin. Cultures were incubated at 37°C with slight agitation to A 600 of ϳ0.2. Cells were harvested by centrifugation (2000 ϫ g, 20 min, 4°C) and washed twice with 10 ml of ice-cold 0.5 M sucrose (2000 ϫ g, 20 min, 4°C) and once with 10 ml of ice-cold 0.5 M sucrose supplemented with 15% (v/v) glycerol (2000 ϫ g, 20 min, 4°C). The cells were resuspended in 50 l of ice-cold 0.5 M sucrose supplemented with 15% glycerol (v/v) and were either used directly or stored at Ϫ70°C. Electrotransformations were done using the Gene Pulser II Electroporation System (Bio-Rad). 1 g of suicide vector or 100 ng of shuttle vector was mixed with 50 l of the electrocompetent cells on ice. The mixtures were transferred into prechilled sterile Gene Pulser cuvettes (interelectrode distance, 0.1 cm; Bio-Rad) and pulsed with a setting of 15 microfarads, 1.8 kV, and 200 ohms. After the electric pulse, the cells were diluted in 1 ml of THY supplemented with 0.3 M sucrose and incubated for 2 h at 37°C under 6% CO 2 . The cells were then plated on THY agar containing the appropriate antibiotics. The cells routinely yielded ϳ10 6 transformants/g shuttle vector.

Isolation of Total RNA and Northern Hybridization
For extraction of total RNA, overnight S. suis cultures were diluted 100-fold into fresh THY and grown at 37°C with vigorous shaking. At various levels of turbidity, 1-ml aliquots were taken from the cultures. The cells were harvested by centrifugation (15,000 ϫ g, 2 min, 4°C), immediately snap-frozen in liquid N 2 , and stored at Ϫ70°C. The cells were later thawed, and total RNA was extracted using the RNeasy Mini kit (Qiagen) according to instructions of the manufacturer. 20 g of RNA was electrophoresed on a 1% agarose gel containing 2.2 M form-

Preparation of Cellular Protein Extracts of S. suis
Overnight S. suis cultures were diluted 100-fold into fresh THY and grown at 37°C with vigorous shaking. At various levels of turbidity, cells were harvested by centrifugation (2000 ϫ g, 20 min, 4°C), washed with P i /NaCl (10 mM sodium phosphate buffer and 0.15 M NaCl (pH 7.4)) (2000 ϫ g, 20 min, 4°C), and resuspended in 750 l of P i /NaCl. The cells were disrupted by sonication five times (20 s each time), with a chilling interval of 1 min between the sonications. EDTA-free Protease Inhibitor Mixture Tablets (Roche Molecular Biochemicals) were added to the sonicates according to instructions of the manufacturer. After unbroken cells and cell debris were removed by centrifugation (15,000 ϫ g, 30 min, 4°C), the cleared lysate was collected and stored at 4°C. Protein concentrations were determined at least in triplicate using the Bio-Rad Protein Assay, based on the Bradford dye binding procedure (40), using bovine serum albumin (BSA) (Sigma) as a standard.

Preparation of Polyclonal Antibodies Specific for Dpr
Expression and purification of recombinant Dpr have been described previously (41). Two rabbits were immunized subcutaneously with 1.0 ml of a mixture of the recombinant Dpr (100 g/ml) and Freund's complete adjuvant (1:1, v/v). Booster injections with the same protein mixture in Freund's incomplete adjuvant were given at 28 and 56 days, and sera were collected 14 days after each immunization. The sera collected 14 days after the last booster had the highest activity against Dpr and were used for subsequent work.

Western Analysis of Dpr Expression
For Western blotting, the proteins were resolved under denaturing conditions in a 12% polyacrylamide gel or under nondenaturing conditions in a 4 -15% gradient Tris-HCl Ready Gel (Bio-Rad) using SDS-PAGE Low Range (Bio-Rad) or equine spleen type I ferritin (Sigma) and BSA as the molecular mass markers, respectively. Proteins were subsequently transferred to a Protran Nitrocellulose Transfer Membrane (Schleicher & Schell) using the 2117 Multiphor II Electrophoresis Unit (LKB Bromma). The membranes were saturated with 3% (w/v) BSA and 0.1% (v/v) Tween 20 in P i /NaCl at room temperature for 1 h. The anti-Dpr polyclonal antibodies were diluted 1:10,000 in 1% (w/v) BSA and 0.05% (v/v) Tween 20 in P i /NaCl, and the membranes were incubated at room temperature for 30 min. After washing with P i /NaCl, the membranes were incubated at room temperature for 30 min in 1% BSA (w/v) and 0.05% (v/v) Tween 20 in P i /NaCl containing a 1:10,000 dilution of peroxidase-conjugated goat anti-rabbit immunoglobulins (DAKO). After washing with P i /NaCl, the ECL chemiluminescence detection kit (Amersham Biosciences) was used to detect the binding according to instructions of the manufacturer.

Inactivation of dpr in S. suis
We generated a genetically stable deletion mutant (D282⌬dpr) of the dpr by adopting the double cross-over method.
Construction of the Suicide Vector pDPR2-A 2.8-kb fragment around the dpr was amplified by PCR using primers TOT-5Ј and TOT-3Ј. The resulting PCR product was digested with BclI and KpnI, and the resulting 2.6-kb fragment was cloned into the KpnI-BamHI site of pID700 to generate pDPR1. The dpr was deleted from the pDPR1 by PCR with primers DEL-5Ј and DEL-3Ј and replaced by a spectinomycin resistance gene (spc) devoid of terminator generated by PCR with primers SPCS and SPCA using pKUN19-Spc as the template. The resulting plasmid, pDPR2, was analyzed by restriction enzyme digestions and PCR to contain the spc in the same direction of transcription as the dpr.
Generation of the Mutant-Because strain 628 has been poorly transformable in our hands, we used D282 cells of same serotype and with identical dpr sequence as strain 628, and colonies with the single cross-over genotype (Chl R , Spc R ) were selected. One Chl R and Spc R colony was chosen, and the double cross-over genotype (Spc R ) was selected as follows. Single cross-over mutant was first grown overnight with spectinomycin (1000 g/ml), allowing the vector to excise out of the genome, leaving the spc in the place of the dpr. Resulting double cross-over mutants were enriched by using the bacteriostatic activity of chloramphenicol (10 g/ml) and at the same time killing the dividing bacteria (still having the vector insertion) by using ampicillin (20 g/ ml). The enriched bacteria were plated on THY with spectinomycin, and the double cross-over genotype (Spc R ) was verified by replica plating. The replacement of dpr by spc was further verified by PCR with primers DPR-5Ј-OUT and DPR-3Ј-OUT annealing upstream and downstream of dpr, respectively, and by Western analysis.

Ectopic Expression of Dpr in D282⌬dpr
The dpr mutation was complemented by introducing wild-type dpr into D282⌬dpr in an E. coli/Streptococcus sp. shuttle vector pLZ12-Km/ Spc. pLZ12-Km/Spc was generated by ligating the spc generated by PCR with primers SPCS and SPCB using pKUN19-Spc as the template to NcoI/EcoRI-digested pLZ12-Km. Two dpr-containing DNA fragments, L4A and P3, were amplified by PCR using primers 5Ј-DCOMP and 3Ј-COMP or 5Ј-UPCOMP and 3Ј-COMP, respectively. These DNA fragments were digested with KpnI and ligated into KpnI-digested and dephosphorylated pLZ12-Km/Spc to generate the plasmid constructs pLZ12-L4A and pLZ12-P3, respectively. In these constructs, dpr transcription is under the control of spc promoter or the possible promoter activity of dpr upstream sequence. The constructs were sequenced to ensure that no mistakes were introduced into dpr during the amplification and introduced into D282⌬dpr, and complementation genotype (Spc R , Kan R ) was selected on agar plates. The expression of Dpr was verified by Western analysis.

Radiometric Determination of Iron Content of Dpr
Overnight S. suis cultures were diluted 100-fold into 25 ml of fresh THY containing 55 FeCl 3 (663.3 MBq/mg; PerkinElmer Life Sciences) as 0.35 MBq/ml and grown to early-stationary phase at 37°C with vigorous shaking. The bacteria were harvested by centrifugation (2000 ϫ g, 20 min, 4°C), washed once with P i /NaCl (2000 ϫ g, 20 min, 4°C), and resuspended in 750 l of P i /NaCl. Cellular protein extracts were prepared as described. For determination of iron content of Dpr, 50 g of cellular protein extracts was resolved under nondenaturing conditions in a 4-15% gradient Tris-HCl Ready Gel (Bio-Rad). After the run, one gel was stained with Coomassie Blue to show the pattern of protein bands for equal loading in each lane, and another gel was placed between two plastic sheets and analyzed for 55 Fe-containing proteins using Fujifilm BA-2500 Phosphor Imaging Plate System (Fuji Photo Film Co.) according to instructions of the manufacturer.

H 2 O 2 Sensitivity Assays
Overnight S. suis cultures were diluted 100-fold into fresh THY and grown to early-stationary phase at 37°C with vigorous shaking. At this point, 1-ml aliquots were taken from the cultures and exposed to H 2 O 2 in concentrations of 0, 1.0, 2.5, and 5.0 mM. Cells were left in contact with H 2 O 2 for 2 h at 37°C without agitation. Immediately after the incubation, cells were diluted in P i /NaCl and plated for viability counts onto THY agar. Colonies were counted after overnight incubation under 6% CO 2 at 37°C. In the iron chelator complementation analysis of the H 2 O 2 hypersensitivity of D282⌬dpr, 1-ml aliquots of early-stationary phase cultures were first incubated with deferoxamine mesylate (DFOM) (Sigma) or diethylene triamine pentaacetic acid (DTPA) (Sigma) at concentrations of 0, 10, 50, and 100 M for 30 min at 37°C without agitation. The cells were then incubated with or without 5.0 mM H 2 O 2 for 2 h at 37°C without agitation. Immediately after the incubation, cells were diluted in P i /NaCl and plated for viability counts onto THY agar. Colonies were counted after overnight incubation under 6% CO 2 at 37°C.

RESULTS
General Features of S. suis dpr Locus-We have previously identified in S. suis a galactose-specific adhesion activity (31)(32)(33) and purified a candidate adhesin (42). Recently, the corresponding gene was cloned, based on peptide sequence data of the purified protein and by genome walking. 1 The gene turned out to encode a protein with 64% identity and 82% similarity to S. mutans Dpr (NCBI Protein Database accession number BAA96472) at the level of primary amino acid sequence. In the present study, we identified by Southern hybridization and subsequently cloned and sequenced a 6.5-kb ClaI-EcoRI frag-ment containing the entire dpr. As shown in Fig. 1A, we identified eight possible ORFs in addition to dpr, some of which were in the same direction of transcription. This indicated that dpr might be transcribed in a multicistronic mRNA. To study this possibility, we first analyzed the upstream region of dpr for sequence features commonly associated with bacterial promoters. As shown in Fig. 2, we found a possible ␦-factor Ϫ10 recognition sequence TATAAT and a Ϫ35 recognition sequence TTGCAA between ORFI and dpr. A putative Shine-Dalgarno box was found a few nucleotides upstream of the initiation codon. Northern hybridization demonstrated that dpr is transcribed monocistronically with an approximate mRNA size of 550 bp (Fig. 1B). Thus, there seems to be a functional promoter between ORFI and dpr possibly including the sequence features found in this study (Fig. 2). However, further work including primer extension analysis is needed to characterize the putative dpr promoter.
Inactivation and Complementation of dpr in S. suis-The sequence information of the 6.5-kb ClaI-EcoRI fragment allowed us to construct a suicide plasmid pDPR2 and subsequently generate a dpr knockout mutant (D282⌬dpr) using allelic replacement (Fig. 3A). PCR analysis with primers DPR-5Ј-OUT and DPR-3Ј-OUT annealing upstream and downstream of dpr, respectively, confirmed the replacement of dpr by spc by homologous recombination (Fig. 3B). This was further verified by Western analysis (Fig. 3C). The dpr mutation was complemented by introducing wild-type dpr into D282⌬dpr in an E. coli/Streptococcus sp. shuttle vector pLZ12-Km/Spc (Fig.  2). In pLZ12-L4A, dpr was inserted into spc containing only its putative Shine-Dalgarno box, leading to a Dpr expression level comparable with that of the wild-type strain D282 at the earlystationary phase (Fig. 3C). In pLZ12-P3, dpr was inserted into spc with its own putative promoter and Shine-Dalgarno box, leading to overexpression of Dpr at the early-stationary phase (Fig. 3C). The overexpression might have resulted from better stability of pLZ12-P3 or from inclusion of the putative dpr promoter. In any case, the complementation strains constructed allowed us to analyze the function of Dpr with two cellular levels of the protein. Both of the constructed plasmids pLZ12-L4A and pLZ12-P3 induced a constant Dpr expression level in early-stationary phase as analyzed by Western blotting of several independent protein extracts.
Role of Dpr in Aerotolerance of S. suis-Dpr is an important aerotolerance factor for S. mutans (11). To study whether Dpr is involved in aerotolerance of S. suis, we analyzed the growth of our strains in solid and liquid THY incubated under normal or 6% CO 2 atmosphere. At the same time, we quantified dpr mRNA and Dpr levels at different time points of the growth phase. We first grew D282 and D282⌬dpr to early-stationary phase in THY at 37°C under 6% CO 2 . The cultures were then diluted 100-fold into fresh THY and incubated at 37°C with vigorous shaking, which causes an extensive aeration of the FIG. 1. General features and organization of the dpr locus. A, genetic and physical map of the cloned 6.5-kb ClaI-EcoRI fragment of S. suis serotype 2 strain 628. Recognition sites for a few commonly used restriction enzymes are indicated. The arrows represent potential ORFs with more than 300 nucleotides complete coding sequence. For possible functions of the gene products, see Table III. B, Northern hybridization analysis of total RNA (20 g) with a dpr-specific DNA probe to reveal the size of dpr mRNA. culture medium. The growth was monitored by measuring A 600 (Fig. 4A). As shown in Fig. 4B, Dpr expression was transcriptionally induced just after the culture entered the actively growing state. The transcriptional activity sharply decreased when the bacterial population reached the stationary phase. The protein levels, on the other hand, reached the maximum in the actively growing state and remained relatively unchanged at the stationary phase (Fig. 4C). Thus, wild-type D282 cells clearly expressed Dpr during the growth period assayed. Yet, as shown in Fig. 4A, there was no significant growth retardation of D282⌬dpr as compared with D282. As analyzed by plating of early-stationary phase cultures, there were no significant differences between the colony forming abilities of D282⌬dpr and D282, regardless of whether the THY plates were incubated aerobically or under 6% CO 2 (data not shown). Taken together, the results indicate that Dpr is not an essential aerotolerance factor for S. suis. (11). Sensitivities of our S. suis strains to H 2 O 2 were tested by exposing early-stationary phase cultures to different concentrations of H 2 O 2 and counting viable cells after plating onto THY. Fig. 5 shows that over the whole range of H 2 O 2 concentrations used, there was only a slight loss of viability of the wild-type strain D282. In contrast, D282⌬dpr was highly sensitive toward H 2 O 2 with a ϳ10 6 -fold reduced viability after exposure to 5.0 mM H 2 O 2 . To examine the possibility that the H 2 O 2 hypersensitivity of D282⌬dpr was not due to its Dpr deficiency but rather was caused by polar effects of the spc insertion, genetic complementation analyses were done. These experiments were important because D282 was used for mutant construction due to poor transformability of 628, which served as the initial source of sequence information (Fig. 1A). Also, ORFV with a predicted gene product of 16.2 kDa ( Fig. 1A; Table III) was in the complementary strand to dpr and was also deleted in D282⌬dpr. Fig. 5 shows that over the whole range of H 2 O 2 concentrations used, both of the complementation strains were resistant to H 2 O 2 . This confirmed that Dpr was responsible for the detected H 2 O 2 resistance in S. suis and ruled out the participation of ORFV and other downstream effects. Overexpression of Dpr in D282P3 (Fig. 3C) did not significantly increase the ability of bacteria to tolerate H 2 O 2 (Fig. 5).

Role of Dpr in H 2 O 2 Resistance of S. suis-Dpr is involved in H 2 O 2 resistance of S. mutans
Iron  (20,21). We analyzed how iron chelators with different membrane permeabilities altered the H 2 O 2 hypersensitivity of D282⌬dpr. The hydrophilic iron chelator DFOM, a siderophore produced by Streptomyces pilosus, has been used in several studies to interfere with the intracellular iron pool of both prokaryotic and eukaryotic cells (43)(44)(45)(46). DTPA, on the other hand, has been used as an extracellular iron chelator (46). The effects of these compounds on the H 2 O 2 hypersensitivity of D282⌬dpr were assayed by exposing early-stationary phase cultures to 5.0 mM H 2 O 2 after preincubating the bacteria with different concentrations of the chelators. As shown in Fig. 6 the H 2 O 2 hypersensitivity of D282⌬dpr was complemented by a preincubation with 100 M DFOM. In contrast, DTPA did not complement the H 2 O 2 hypersensitivity at any of the concentrations studied. The results indicate that the intracellular but not the extracellular iron pool is involved in H 2 O 2 sensitivity of the bacteria and that the absence of functional Dpr can be complemented by a molecule capable of iron chelation.
Site-directed Mutagenesis of the Putative Ferroxidase Center of Dpr-Amino acid sequence alignment of Dps family members with known crystal structures revealed that S. suis Dpr contained a putative ferroxidase center in its N-terminal half (Fig. 7). It has been suggested that this amino acid motif is involved in iron incorporation in a fashion similar to that of classical ferritins (22,27,28,47), but there is no direct experimental evidence to support this proposal in vivo. When we analyzed cellular protein extracts from early-stationary phase cultures grown in the presence of 55 FeCl 3 , Dpr had clearly incorporated iron (Fig. 8). To investigate the possible relationship of the in vivo iron incorporation activity with the putative ferroxidase center, we utilized site-directed mutagenesis. The negatively charged residues Asp-74 and Glu-78 of the putative ferroxidase center (Fig. 7) were independently substituted with Ala. As shown in Fig. 8A, the mutations had no apparent effects on the level of expression or solubility of Dpr. The mutations also seemed to have no effects on the oligomeric stability of Dpr as indicated by similar mobility to wild-type Dpr in nondenaturing polyacrylamide gels (Fig. 8B). In contrast, the mutations independently caused a complete inactivation of the iron incorporation activity in vivo (Fig. 8C). Taken together, S. suis Dpr was able to incorporate iron in vivo, and this process was dependent on Asp-74 and Glu-78, the amino acids conserved in the putative ferroxidase centers of several Dps family members (Fig. 7) (14,22,27,28,47).
H 2 O 2 Sensitivities of D282⌬dpr Expressing Wild-type or Sitemutated Forms of Dpr-The H 2 O 2 sensitivities of D282⌬dpr expressing wild-type or site-mutated forms of Dpr were assayed by exposing early-stationary phase cultures to different concentrations of H 2 O 2 and counting viable cells after plating onto THY. As shown in Fig. 9 bacteria expressing wild-type Dpr were fully resistant to H 2 O 2 . In contrast, bacteria expressing the site-mutated and iron incorporation-negative forms of Dpr were hypersensitive to H 2 O 2 with a comparable phenotype to dpr knockout. Thus, the detected H 2 O 2 resistance of S. suis was not only critically dependent on Dpr but specifically on its iron incorporation activity. Bacteria were grown to early-stationary phase in THY and treated for 2 h with different concentrations of H 2 O 2 at 37°C, and their viabilities were counted after plating to THY. Values are given as means of two independent experiments. Strains D282L4A (E) and D282P3 (‚) are otherwise like D282⌬dpr (f) but express Dpr from the pLZ12-L4A and pLZ12-P3, respectively, at two different cellular levels (Fig. 3C).

DISCUSSION
In this paper, we provide for the first time direct in vivo evidence on how streptococcal Dpr mediates its protective function against H 2 O 2 . H 2 O 2 resistance is a crucial property for streptococci because several species produce it as a part of their metabolism despite the lack of oxidative phosphorylation (4 -7). H 2 O 2 is also encountered as a part of host defenses (8), and under certain conditions, streptococci seem to utilize it as their own virulence factor (4,9,10). Thus, in addition to maintenance of normal cellular physiology, Dpr may also have a role in the pathogenesis of streptococcal infections. S. suis seemed an ideal model organism to study the in vivo function of Dpr because it not only lacks catalase, like other streptococci, but also lacks NADH peroxidase (34). This enzyme is capable, to some extent, of substituting for the absence of catalase (48,49).
It is known that defects in the regulation of intracellular iron homeostasis may lead to enhanced oxidative stress (46,50,51). Iron in its reduced form nonenzymatically cleaves H 2 O 2 into hydroxyl radicals, the most deleterious forms of reactive oxy-gen intermediates, by Fenton chemistry (H 2 O 2 ϩ Fe 2ϩ 3 ⅐ OH ϩ -OH ϩ Fe 3ϩ ) (20,21,52). Based on a primary amino acid sequence comparison, Dpr has been reported to be a member of the Dps protein family (11), in which one of the functional features is iron binding activity (17,22,26,27). The family members seem to be able to oxidize Fe 2ϩ and store it inside the oligomeric protein shell as Fe 3ϩ (17,29,53,54), resembling classical ferritins in this respect (23). Indeed, Bozzi and coworkers (22) identified in L. innocua Flp, a member of the Dps family that is well characterized in vitro (53,54), a homologous region to the iron nucleation center of mammalian ferritin L-subunit (24). They suggested that this region could carry out the initial steps in iron core formation and also serve as a ferroxidase center. Although a homologous region to this amino acid motif can be found in several Dps family members (Fig. 7) (14), no direct experimental evidence is available on its functionality in iron incorporation in vivo. Furthermore, the biological significance of iron incorporation is not well established. In the present study, a putative ferroxidase center was identified in the N-terminal half of S. suis Dpr, and the structure-function relationship was studied by site-directed mutagenesis.
The targets for the point mutations in the putative ferroxidase center were chosen by using the three-dimensional structure of L. innocua Flp (28). Flp shares 44% primary amino acid sequence identity with S. suis Dpr. In Flp crystals, 12 iron atoms have been directly observed occupying the putative ferroxidase centers, which are formed at the interfaces of two adjacent subunits (28). The amino acids protruding into the interface of Flp dodecamer and seemingly involved in the actual iron coordination (His-31, His-43, Asp-47, Asp-58, and Glu-62) (28) are all conserved and similarly spaced in S. suis Dpr as His-47, His-59, Asp-63, Asp-74, and Glu-78, respectively (Fig. 7). The involvement of these amino acids in forming a classical dinuclear ferritin-like ferroxidase center has been modeled for Flp (28). In the initial step, the first ferrous iron, guided into the interior of the dodecamer through hydrophilic channels, would bind to His-31 from the A helix and Asp-58 and Glu-62 from the B helix of an adjacent subunit (Fig. 7) (28). Based on the L. innocua Flp model, we decided to independently substitute two of the homologous negatively charged amino acids, Asp-74 and Glu-78, with Ala. The substitutions had no apparent effects on the expression, solubility, or oligomeric stability of the Dpr dodecamer. In contrast, Asp-74 and Glu-78 were crucial for the Dpr to incorporate iron in vivo. This result represents the first direct  demonstration of the functional involvement of the putative ferroxidase center in iron incorporation in vivo for Dpr or any of the Dps family members.
The dpr knockout of S. suis had a ϳ10 6 -fold reduced viability under exposure to 5.0 mM H 2 O 2 as compared with the wild-type strain. By analyzing the H 2 O 2 sensitivities of dpr knockout expressing wild-type or site-mutated forms of Dpr, we determined the role played by the iron incorporation activity of Dpr in H 2 O 2 resistance. Strikingly, bacteria expressing iron incorporation-negative forms of Dpr were hypersensitive to H 2 O 2 with a phenotype comparable to that of dpr knockout. Thus, Dpr seemed to protect the bacteria against H 2 O 2 by its iron incorporation activity. This was supported by the iron chelator complementation analysis of the H 2 O 2 hypersensitivity of D282⌬dpr. DFOM, an iron chelator interfering with the intracellular iron pool (43)(44)(45)(46), was able to rescue the viability defect of D282⌬dpr. Under the same conditions, an extracellular iron chelator, DTPA (46), did not cause any effect. Yamamoto et al. (25) recently reported data indicating that S. mutans Dpr is capable of inhibiting the action of Fenton chemistry in vitro. Thus, Dpr seems to provide an indirect means for catalase-negative bacteria to tolerate H 2 O 2 by inhibiting the Fe 2ϩ -catalyzed cleavage of H 2 O 2 into more toxic reactive oxygen intermediates.
Mechanisms other than iron incorporation have been suggested for Dps family members to explain the protection against H 2 O 2 , and synergy between them is possible (12, 14 -16). A convincing body of evidence indicates that E. coli Dps protects DNA by a direct association, which leads to highly  (28) or in Pb 2ϩ coordination of the Dps heavy atom derivative (47) are shaded black. Amino acids proposed to be involved in iron coordination in Dlp-1 and Dlp-2 (27) are also shaded black. Flp, the closest Dpr homolog with a solved crystal structure, folds into a four-helix bundle, where amino acids from helixes A and B (boxed) form the intersubunit iron-binding pocket. In the dodecamer, 12 pockets are formed, each serving as a putative ferroxidase center (28). The amino acids Asp-74 and Glu-78, chosen for independent substitution with Ala in S. suis Dpr, are marked with an asterisk.

FIG. 8. Effects of the putative ferroxidase center mutations on Dpr and its in vivo iron incorporation activity.
Western analysis of Dpr expression after resolving 2 g of cellular protein extracts either under denaturing (A) or nondenaturing (B) conditions. Under nondenaturing conditions, equine spleen type I ferritin (Fer) and BSA were used as molecular mass markers. Recombinant Dpr (tDpr) lacks seven amino acids in its N terminus (41), leading to slightly faster mobility of the Dpr oligomer. C, iron ( 55 Fe) incorporation into native and site-mutated forms of Dpr during aerobic growth in THY. 50 g of cellular protein extracts from early-stationary phase cultures was loaded into each well of a 4 -15% gradient polyacrylamide gel, with equine spleen type I ferritin and BSA serving as molecular mass markers, and resolved under nondenaturing conditions. The gel was subsequently analyzed for 55 Fe-containing proteins by autoradiography. ordered DNA-Dps biocrystals (12,15,16,55). However, Dps also contains a putative ferroxidase center in its N-terminal half, and a dual function in H 2 O 2 resistance has been proposed. Dps could directly protect the DNA by biocrystal formation but could also inhibit the action of Fenton chemistry by its iron incorporation activity (17,29,47). Whether Dpr also binds and protects DNA remains contradictory. Yamamoto et al. (25) recently reported data indicating that S. mutans Dpr has no DNA binding activity. In our own studies using hydroxyl radical DNA footprinting assays, we have not detected any DNA binding activity for S. suis Dpr. 3 Also, some other members of the Dps family seem to lack DNA binding activity (26,27). The possibility that Dpr could differ from Dps in lacking DNA binding activity is not unexpected, considering that the primary amino acid sequence identity shared between S. suis Dpr and E. coli Dps is only 25%. It also seems unlikely that Dpr could have a catalase-like activity as reported for the DpsA of Synechococcus sp. (14). S. suis Dpr shares an extremely weak homology to DpsA (nondetectable in BLAST-P data base search) and does not seem to contain heme, which has been linked to the enzymatic activity of DpsA (14). Furthermore, Yamamoto et al. (25) have recently presented data indicating that S. mutans Dpr has no catalase-like activity. Indeed, the results of our present study indicate that it is the iron incorporation activity that is the molecular basis of Dpr-mediated H 2 O 2 resistance.
The Dps family includes molecules from diverse taxonomic lineages (14) with striking similarities, even including a recent member in the archaeon Halobacterium salinarum (56). Proteins fold into four-helix bundles resembling the fold of mammalian ferritins and assemble into large hollow globular complexes (23,27,28,47). Based on our present study, S. suis Dpr protected the bacteria against H 2 O 2 by its iron incorporation activity. Whether this is true for the Dps family members in general is not known. However, the amino acid residues forming the putative ferroxidase centers are among the most conserved primary amino acid sequence features of the family (Fig.  7) (14). Also, all members have proven to be iron-binding proteins, if analyzed for this activity (11,17,22,26,27,56). Thus, it is possible that in diverse bacterial species, Dps-related molecules might serve as functional ferritin-like molecules and protect bacteria against H 2 O 2 as inhibitors of Fenton chemistry. However, it has become evident that, despite sharing the few highly conserved amino acid residues, the putative ferroxidase centers vary considerably (Fig. 7), which might lead to different functional properties. Indeed, Zhao et al. (17) have recently reported data on iron oxidation and incorporation properties of E. coli Dps that differ from those of L. innocua Flp in both the rates of Fe 2ϩ binding and oxidation. Thus, it is obvious that more studies are needed to further elucidate the importance of iron incorporation activity for Dps family members in H 2 O 2 resistance.
The functional divergence of the Dps family is an important area of further investigation. The family members seem to be involved in aerotolerance (11), H 2 O 2 resistance (11,12,57), cold shock adaptation (58), starvation tolerance (12), iron storage (22), neutrophil activation (59,60), carbohydrate binding (61), and adhesion (62). The molecular determinants mediating these activities are largely unknown. Additional in vivo studies together with in vitro work using recombinant site-mutated forms of the molecules in combination with crystal structure information (27,28,41,47) are needed to reveal the molecular mechanisms behind these activities and their relationships to each other. Determining the structural basis of these biological activities may eventually contribute to the development of means to prevent and treat diseases caused by pathogenic bacteria.