Deinococcus radiodurans exhibits extreme resistance to DNA damage and is one of only few bacteria that encode two Dps (DNA protection during starvation) proteins. Dps-1 was shown previously to bind DNA with high affinity and to localize to the D. radiodurans nucleoid. A unique feature of Dps-2 is its predicted signal peptide. In the present paper, we report that Dps-2 assembly into a dodecamer requires the C-terminal extension and, whereas Dps-2 binds DNA with low affinity, it protects against degradation by reactive oxygen species. Consistent with a role for Dps-2 in oxidative stress responses, the Dps-2 promoter is up-regulated by oxidative stress, whereas the Dps-1 promoter is not. Although DAPI (4′,6-diamidino-2-phenylindole) staining of Escherichia coli nucleoids shows that Dps-1 can compact genomic DNA, such nucleoid condensation is absent from cells expressing Dps-2. A fusion of EGFP (enhanced green fluorescent protein) to the Dps-2 signal peptide results in green fluorescence at the perimeter of D. radiodurans cells. The differential response of the Dps-1 and Dps-2 promoters to oxidative stress, the distinct cellular localization of the proteins and the differential ability of Dps-1 and Dps-2 to attenuate hydroxyl radical production suggest distinct functional roles; whereas Dps-1 may function in DNA metabolism, Dps-2 may protect against exogenously derived reactive oxygen species.

INTRODUCTION

Members of the bacterial family Deinococcaceae are among the most radiation-resistant organisms characterized. They form a distinct phylogenetic lineage most closely related to the genus Thermus, and several species have been described of which Deinococcus radiodurans RI was the first to be discovered. Their natural distribution includes environments in which they may encounter extended periods of desiccation [1,2]. D. radiodurans is best known for its remarkable resistance to ionizing radiation, a resistance thought to occur as a consequence of its resistance to desiccation, as both conditions are associated with extensive DNA damage [3]. Precisely how it grows under chronic γ-radiation or how it recovers from acute doses of >10 kGy without induced mutation is not clear, but several contributing factors have been proposed, including a greater efficiency of DNA repair pathways, recombination events facilitated by its ring-like nucleoid or accumulation of Mn(II) [411].

DNA damage may occur as a consequence of irradiation or due to the associated production of ROS (reactive oxygen species) [3]. ROS are also produced at greater levels during starvation, desiccation or other stress-related conditions. The ROS H2O2 can be reduced by transition metals such as Fe2+ to form highly reactive hydroxyl radicals through Fenton chemistry. Free cellular iron can therefore lead to the proliferation of cellular damage through the production of ROS [12,13]. Among several mechanisms that prokaryotes use to combat the consequences of such ROS production, the non-specific DNA-binding protein Dps (DNA protection during starvation) features prominently.

Dps proteins can protect DNA from ROS by both DNA binding and by oxidizing and sequestering cellular iron. Dps proteins form shell-like structures composed of 12 subunits, each of which consist of a four-helix bundle and often containing an N-terminal extension that participates in DNA binding and dodecameric assembly [1423]. Mycobacterium smegmatis Dps1 also has a C-terminal extension that plays a role in DNA binding and assembly [23]; in general, extensions beyond the central four-helix bundle correlate with DNA binding. Dps proteins oxidize Fe2+ to Fe3+ and sequester it in a bioavailable mineralized iron core, thus making it non-reactive to free radical production through Fenton chemistry. That Dps proteins generally prefer H2O2 to molecular oxygen as the oxidant is also important, as this preference allows Dps to detoxify Fe2+ and H2O2 simultaneously [16,20,24,25].

For a bacterial species to encode more than one Dps homologue is not common, but it is seen, for example, in Bacillus anthracis, M. smegmatis and D. radiodurans [2527]. D. radiodurans encodes two proteins: Dps-1 (DR2263) and Dps-2 (DRB0092) [28]. The structure of Dps-1 reveals the expected four-helix bundle subunit, but also a novel metal site within the N-terminal extension [29,30]. For Dps-1, this N-terminal metal site is necessary for oligomeric assembly, and residues preceding the metal site are required for DNA binding [21,31]. Dps-1 is associated with the D. radiodurans nucleoid, but its function is not clear, and it is not essential for nucleoid organization [32]. The crystal structure of Dps-2 was solved in both its apo and iron-loaded form. The Dps-2 subunit also has the four-helix bundle fold, but features a single helical turn C-terminal extension that chelates a metal ion [33]. In the present study, we show that this C-terminal extension is required for oligomeric assembly and report on another unique feature of Dps-2, its signal peptide, which directs expression of EGFP (enhanced green fluorescent protein) to the perimeter of D. radiodurans cells. The distinct functional properties and cellular localization of Dps-1 and Dps-2 suggests different in vivo roles; whereas Dps-1 may participate in organization of genomic DNA, Dps-2 may protect against exogenous ROS.

EXPERIMENTAL

Cloning of Dps-2, mutagenesis and construction of signal peptide–EGFP reporter

The gene encoding full-length D. radiodurans Dps-2 (including the sequence encoding the predicted signal peptide) was PCR-amplified from D. radiodurans genomic DNA. The PCR product was cloned into T7-NT/TOPO (Invitrogen). The construct was digested with NdeI to remove the N-terminal His6 tag and the remaining fragment was religated to yield pTOPO-dps2. Dps-2 truncated for the predicted 30-amino-acid signal peptide was created by whole-plasmid PCR using pTOPO-dps2 as a template. The C-terminally truncated Dps-2 (CLess) was PCR-amplified from D. radiodurans genomic DNA and cloned into the Champion pET100 TOPO vector (Invitrogen). Dps-1 and Dps-met have been characterized previously [21,27].

To determine cellular localization of Dps-2 in D. radiodurans, the promoter (210 bp upstream of the translational start) and signal peptide of Dps-2 were amplified and cloned into the pRAD1 shuttle vector after digestion with AgeI and BamHI to generate pRADpsp. pd1EGFP-N1 (Clontech) was digested with BamHI and EcoRI and the gene encoding EGFP was subcloned into pRADpsp to generate pRADpspEGFP.

All constructs were confirmed by sequencing. Primer sequences are available in the Supplementary Online Data at http://www.BiochemJ.org/bj/447/bj4470381add.htm.

Protein expression and purification

Plasmid encoding Dps-2 truncated for the signal peptide or CLess Dps-2 was transformed into Escherichia coli Rosetta2, and cultures were grown at 30°C in LB (Luria–Bertani) medium with 50 μg/ml ampicillin to a D600 of 0.2 and induced with 1 mM IPTG (isopropyl β-D-thiogalactopyranoside) for 2 h. Cells were pelleted by centrifugation.

Cells were resuspended and incubated for 1 h in lysis buffer [50 mM Tris/HCl (pH 8.0), 0.25 M NaCl, 5 mM sodium EDTA, 5% (v/v) glycerol, 5 mM 2-mercaptoethanol and 0.1 mM PMSF] and subsequently sonicated at 15 s intervals for 5 min. Cellular debris was pelleted by centrifugation at 8000 g for 20 min. For Dps-2 purification, the supernatant was incubated in a 70°C water bath for 30 min and the suspension once again pelleted by centrifugation at 8000 g for 30 min. Supernatants for Dps-2 and CLess Dps-2 were dialysed overnight against HA buffer [50 mM Tris/HCl (pH 7.6), 50 mM KCl, 5% (v/v) glycerol, 1 mM sodium EDTA, 3.5 mM 2-mercaptoethanol and 0.2 mM PMSF] and centrifuged at 4000 g for 15 min. The supernatants were applied to a DEAE-cellulose column equilibrated with HA buffer. Column flowthrough for Dps-2 was concentrated, and the buffer was brought to 150 mM KCl. The sample was then applied to a Sephadex size-exclusion column, and fractions containing Dps-2 were pooled. The DEAE washes for CLess Dps-2 were run on a CM column, the wash was concentrated, and buffer was brought to 150 mM KCl. The sample was then applied to a Sephadex size-exclusion column and fractions containing CLess Dps-2 were pooled. Purity was confirmed by SDS/polyacrylamide gels stained with Coomassie Brilliant Blue. Concentrations were established by comparison with BSA on SDS/polyacrylamide gels stained with Coomassie Brilliant Blue. Dps-1 was purified and characterized as described in [27].

Determination of oligomeric state

Gel filtration was carried out at 4°C. HiLoad 16/60 Superdex 30 prep grade column (bed length 60 cm, inner diameter 16 mm; GE Healthcare) was equilibrated with buffer A, pH 8.0 [50 mM sodium phosphate buffer, 10 mM imidazole and 10% (v/v) glycerol]. The gel-filtration standard (Bio-Rad Laboratories), a mixture of bovine thyroglobulin (670 kDa), bovine γ-globulin (158 kDa), chicken ovalbumin (44 kDa), horse myoglobin (17 kDa) and vitamin B12 (1.35 kDa), was used to calibrate the column. The concentration of protein applied to the column was 5 mg/ml. Dps-2 and CLess Dps-2 were run independently under the same conditions and eluted with a flow rate of 0.5 ml/min.

Sedimentation equilibrium was used to determine oligomeric state. CLess Dps-2 was dialysed overnight at 4°C against AU buffer [20 mM Tris/HCl (pH 8.0), 50 mM NaCl and 5 mM MgCl2]. The reference and solution sectors of an analytical cell with a double-sector centrepiece were loaded with AU buffer on one side and CLess Dps-2 on the other side. The cell was spun at 10000 rev./min at 20°C using a Beckman Optima XL-A analytical ultracentrifuge equipped with an An-60 Ti rotor. The cell was scanned at 5 h intervals at 294 nm with a step size of 0.004 cm until the system reached equilibrium. Equilibrium sedimentation data were analysed using SEDFIT software and fitted to an equation describing a single non-interacting protein species.

Oligomeric assembly was also assessed by cross-linking with glutaraldehyde and by native gel electrophoresis, as described in [21].

Thermal stability

Dps-2 and CLess Dps-2 were diluted to 10 μM in a buffer containing 50 mM Tris/HCl (pH 8.0), 100 mM NaCl and 5× SYPRO Orange (Invitrogen). Fluorescence emission resulting from dye binding to unfolded protein was measured over the temperature range 1–90°C in 1° increments for 45 s using an Applied Biosystems 7500 Real-Time PCR System using the SYBR Green filter. Correction for the total fluorescence yield was made using reactions without protein. The resulting data were exported to Sigma Plot 9, and the sigmoidal part of the curve was fitted to a four parameter sigmoidal equation.

β-Galactosidase assay

For analysis of Dps-1 and Dps-2 promoter activity in D. radiodurans, the promoters were PCR-amplified from D. radiodurans genomic DNA and digested with BglII. The PCR products were cloned into pRADZ1 containing a promoter-less lacZ gene. Both promoters were cloned in both their forward and reverse directions. For Dps-1, the cloned promoter fragment spanned positions −87 to −10 relative to the start codon; for Dps-2, a 130 bp DNA fragment spanning positions −131 to −2 relative to the start codon was used. The HU promoter (DRA0065; 187 bp spanning positions −191 to −5 relative to the start codon) was similarly amplified and cloned into the BglII site of pRADZ1. Integrity of all constructs was confirmed by sequencing. Primer sequences are available in the Supplementary Online Data.

Exponentially growing D. radiodurans was transformed with reporter constructs using the protocol described in [34], except chloramphenicol was used in place of mitomycin. Transformants were grown until cells reached exponential phase, and cultures were then treated with either 10 mM H2O2 or 10 mM Fe(NH4)2(SO4)2 for 60 min. Cells were harvested by centrifuging 500 μl of culture while measuring the D600. The cell pellets were resuspended in 500 μl of Z-buffer (60 mM Na2HPO4, 40 mM NaH2PO4, 10 mM KCl, 1 mM MgSO4 and 50 mM 2-mercaptoethanol, pH 7.0) supplemented with lysozyme (25 μg/ml) and DNase I (50 ng/ml). After incubation for 30 min at 37°C, 20 μl of toluene was added. After another 60 min of incubation at 37°C, 300 μl of 4 mg/ml ONPG (o-nitrophenyl β-D-galactopyranoside) was added to each aliquot. The reaction was stopped by the addition of 500 μl of 1 M Na2CO3. Absorbances were measured at 420 and 550 nm and the Miller unit activity was measured using the equation: {A420−(1.75×A550)/[reaction time (min)×volume×A600]}×1000.

EMSA (electrophoretic mobility-shift assay)

To remove divalent cations, proteins were incubated with 50 mM bipyridyl for 20 min at 4°C. The bipyridyl or metalbipyridyl complex was removed by dialysis against a high-salt buffer [10 mM Tris/HCl (pH 8.0), 500 mM KCl, 5% (v/v) glycerol, 0.5 mM 2-mercaptoethanol and 0.2 mM PMSF] at 4°C for 2 h. The bipyridyl-treated proteins were incubated with 800 nM of MnCl2 for 6 h. DNA binding was assessed in 10 μl reaction mixtures where 100 ng of pGEM5 was incubated with 10 pmol of protein in binding buffer [20 mM Tris/HCl (pH 8.0), 100 mM KCl, 0.1 mM EDTA, 0.1 mM dithiothreitol, 0.05% Brij58, 10 μg/ml BSA and 5% (v/v) glycerol] at room temperature (22°C) for 1 h. Reactions were resolved on 1% agarose gels with 0.5× TBE buffer [22.5 mM Tris/borate (pH 8.3) and 0.5 mM EDTA] and stained with ethidium bromide. EMSA was also performed with 26 bp DNA as described in [27], except that the binding buffer contained 50 mM NaCl as the monovalent salt.

DNA protection

DNA protection from hydroxyl radicals was determined by incubation of 100 ng of pGEM5 and 10 pmol of protein in 20 mM Tris/HCl (pH 7.5), 400 mM KCl for 30 min at room temperature. Then 150 μM Fe(NH4)2(SO4)2 and 10 mM H2O2 were added followed by incubation at room temperature for 10 min. DNA protection from DNase I was ascertained by incubating 10 pmol of Dps-2 and 100 ng of pGEM5 in 20 mM Tris/HCl and 400 mM KCl for 30 min at room temperature. Then 1 unit of DNase I, 0.5 mM MgCl2 and 0.5 mM CaCl2 were added to the reaction mixture, which was incubated for 5 min at room temperature. Reactions were terminated with 2 μl of stop solution (5% SDS and 15% glycerol) and plasmids were resolved on a 1% agarose gel with 0.5× TBE buffer and stained with ethidium bromide. Reactions were performed at least in triplicate.

Dps-2 localization in D. radiodurans and nucleoid condensation

To determine cellular localization of Dps-2, D. radiodurans transformed with plasmid pRADpspEGFP was grown in 10 ml of TGY medium (0.8% Tryptone, 0.4% yeast extract and 0.1% glucose) with 3 μg/ml chloramphenicol at 25°C until cultures reached a D600 of 0.8. Cells were viewed using a Leica DM IRE2 under 100× NA (numerical aperture) 1.4 objective.

E. coli Rosetta2 was transformed with TOPO-dps2, pET5a-dps1, or a Champion pET100/D-TOPO vector containing the Dps-met gene. Cultures were grown in 10 ml of LB medium with 50 μg/ml ampicillin at room temperature until the D600 reached 0.2. Protein expression was induced with 1 mM IPTG for 1 h. Then 2 μl of each cell culture was incubated with 2 μl of 50 mg/ml DAPI (4′,6-diamidino-2-phenylindole) for 5 min. Cells were viewed using a Leica DM IRE2 under a 100× NA 1.4 objective. Nucleoids were visualized using a Leica A4 filter cube.

RESULTS

Sequence analysis

Sequence alignment of D. radiodurans Dps proteins reveals significant homology, particularly with regard to residues involved in iron oxidation (Figure 1A). For Dps-1, ferroxidase activity has been reported previously, and the N-terminal extension preceding the four-helix bundle fold has been shown to be required for DNA binding as well as assembly into a dodecamer [21,27,31]. For Dps-2, sequence analysis predicts the ability to oxidize Fe2+ and that the ferroxidase centre lies at the interface between two monomers. The ligands to iron in the predicted ferroxidase centre of Dps-2 are His70 from one subunit and Asp97 and Glu101 from the other, residues that are completely conserved, as well as a water molecule co-ordinated by His82 (residues marked by asterisks in Figure 1A) [33]. Trp71 is also conserved; this residue has been shown to be instrumental in preventing the release of radical by-products of the ferroxidase centre by capturing free electrons [35]. Analysis of the Dps-2 sequence using SignalP 4.0 (http://www.cbs.dtu.dk/services/SignalP/) predicts that the first 30 residues of Dps-2 constitute a signal peptide that may direct Dps-2 to a non-cytoplasmic cellular localization (residue numbering excludes the signal peptide sequence; predicted cleavage site indicated by an arrow in Figure 1A). No other Dps homologues have been shown to encode signal peptides.

Sequence and assembly of Dps-2

Figure 1
Sequence and assembly of Dps-2

(A) Sequence alignment of D. radiodurans Dps-1 (D.rad-Dps1) and Dps-2 (D.rad-Dps2) with homologues from Helicobacter pylori (H.pyl-HNAP), Listeria innocua (L.inn-Dps), M. smegmatis Dps1 (M.smeg-Dps1) and Dps2 (M.smeg-Dps2), E. coli (E.coli-Dps) and Agrobacterium tumefaciens (A.tum-Dps). Amino acids that encompass the four-helix bundle are identified above the alignment. Ligands to iron at the ferroxidase centre are indicated by asterisks. Black shading represents conserved amino acids and grey shading indicates homology. The predicted signal peptide cleavage site is indicated by a vertical arrow above the alignment; the site was predicted using SignalP 4.0 (http://www.cbs.dtu.dk/services/SignalP/). (B) Dps-2 dodecamer with each monomer represented in individual colours (PDB code 2C6R). (C) Enlarged view of the iron exit pore formed by the C-terminal extensions of three monomers. C-terminal extensions of each monomer are coloured, starting with proline (side chain shown in green). Iron is represented by red spheres.

Figure 1
Sequence and assembly of Dps-2

(A) Sequence alignment of D. radiodurans Dps-1 (D.rad-Dps1) and Dps-2 (D.rad-Dps2) with homologues from Helicobacter pylori (H.pyl-HNAP), Listeria innocua (L.inn-Dps), M. smegmatis Dps1 (M.smeg-Dps1) and Dps2 (M.smeg-Dps2), E. coli (E.coli-Dps) and Agrobacterium tumefaciens (A.tum-Dps). Amino acids that encompass the four-helix bundle are identified above the alignment. Ligands to iron at the ferroxidase centre are indicated by asterisks. Black shading represents conserved amino acids and grey shading indicates homology. The predicted signal peptide cleavage site is indicated by a vertical arrow above the alignment; the site was predicted using SignalP 4.0 (http://www.cbs.dtu.dk/services/SignalP/). (B) Dps-2 dodecamer with each monomer represented in individual colours (PDB code 2C6R). (C) Enlarged view of the iron exit pore formed by the C-terminal extensions of three monomers. C-terminal extensions of each monomer are coloured, starting with proline (side chain shown in green). Iron is represented by red spheres.

The crystal structure of Dps-2 (Figures 1B and 1C) revealed the four-helix bundle subunit conformation as well as a C-terminal extension thus far seen only in Dps1 from M. smegmatis [23,33]; the level of homology between the two C-terminal extensions is minimal, except that both contain positively charged residues. No structural information is available for residues preceding Thr42 (the first residue of helix A of the four-helix bundle), suggesting that the N-terminal extension is flexible or disordered. The presence of both N- and C-terminal extensions beyond the four-helix bundle predicts the ability of Dps-2 to bind DNA.

The C-terminal extension is required for dodecameric assembly

The Dps-2 gene (including the sequence encoding the predicted signal peptide) was expressed in E. coli and found to yield functional Dps-2 protein, as was evident by its DNA-binding and ferroxidase activity in vitro (results not shown). To optimize yields, Dps-2 was expressed without its predicted signal peptide and purified to apparent homogeneity (Supplementary Figure S1B at http://www.BiochemJ.org/bj/447/bj4470381add.htm). This deletion was also used to obtain the structure of Dps-2 [33], and experiments presented below were performed using this preparation of Dps-2.

In dodecameric Dps-2, three adjacent subunits interact via their C-termini to form the iron exit channels (Figure 1C). The C-terminal extensions contain a novel metal-binding site, and they mediate contacts between subunits by interacting with the linker regions connecting helices A and B and helices BC and C of the four-helix bundle from an adjacent dimer. In the light of the structural evidence suggesting an architectural role for the C-terminus, a Dps-2 C-terminal truncation mutant (CLess) was created that also lacks the predicted signal peptide sequence. CLess Dps-2 was purified to apparent homogeneity (Supplementary Figure S1A).

The Dps-2 interfaces were also analysed using the PISA (Protein Interfaces, Surfaces and Assemblies) server [36]. The PISA algorithm uses a given structure to predict the most thermodynamically stable assemblies. The PISA algorithm predicts two protein–protein interfaces that contribute significantly to the stability of the Dps-2 dodecamer, one of which is the dimer interface with a total of 24 hydrogen bonds and salt bridges (Table 1). Interaction between the C-terminal loop and an adjacent subunit at the three-fold Dps-like axis is also predicted to be very significant for stability of dodecameric Dps-2, whereas interactions between subunits at the N-terminal (ferritin-like) axes are predicted to contribute less. Metal binding at various sites is also predicted to be significant for stability of the assembly, including metals co-ordinated at the ferroxidase centre, at the C-terminal loop, and at the three-fold N-terminal ferritin-like iron entry channels. The PISA analysis predicts that dodecameric assembly of CLess Dps-2 would be compromised, and that dodecameric Dps-2 would be significantly more stable than isolated subunits.

Table 1
PISA analysis predicts several interfaces that contribute to assembly of Dps-2

HB+SB represents number of hydrogen bonds and salt bridges. CSS, Complex Formation Significance Score, an indicator of interface relevance to complex formation, with CSS=1.000 implying an interface that is essential to stability of the assembly. In addition to the interfaces listed, contacts to a metal near Fe2+ at the ferroxidase centre are also predicted to contribute to complex stability. Metals at the ferroxidase centre are co-ordinated by two subunits, generating two distinct interfaces. Metal co-ordinated at the C-terminal loop contacts an adjacent subunit, likewise generating two distinct interfaces. *Not a significant contribution. Calculations were done using the PISA server at http://www.ebi.ac.uk/msd-srv/prot_int/pistart.html and PDB code 2C6R, treating all ligands as free particles.

Type of interactionHB+SBCSS
Dimer interface 24 1.000 
C-terminal loop and adjacent subunit 1.000 
Adjacent subunits at N-terminal three-fold axis 0.070* 
Metal at ferroxidase centre/subunit 1 0.282 
Metal at ferroxidase centre/subunit 2 0.254 
Metal at N-terminal ferritin-like three-fold axis 0.538 
Metal near C-terminal Dps-like three-fold axis 0.269 
Metal at C-terminal loop 0.166 
Metal at C-terminal loop/adjacent subunit 0.475 
Type of interactionHB+SBCSS
Dimer interface 24 1.000 
C-terminal loop and adjacent subunit 1.000 
Adjacent subunits at N-terminal three-fold axis 0.070* 
Metal at ferroxidase centre/subunit 1 0.282 
Metal at ferroxidase centre/subunit 2 0.254 
Metal at N-terminal ferritin-like three-fold axis 0.538 
Metal near C-terminal Dps-like three-fold axis 0.269 
Metal at C-terminal loop 0.166 
Metal at C-terminal loop/adjacent subunit 0.475 

Dps-1 exists almost exclusively as a dodecamer in solution [27]. Similarly, gel-filtration chromatography shows that the vast majority of Dps-2 (molecular mass of 23612 Da) elutes as a single peak with an estimated molecular mas of 240 kDa, indicating that it also exists mainly as a dodecamer (Figure 2A). This result was expected on the basis of the conserved dodecameric assembly of Dps proteins analysed previously and the structure of Dps-2. In contrast, CLess Dps-2 elutes at an elution volume corresponding to a monomer. As this result was unexpected on the basis of reports that other Dps proteins form subassemblies corresponding to either dimers or trimers, we confirmed it using analytical ultracentrifugation (Figure 2B). Data were best-fitted to a model describing a single non-associating protein species, yielding a molecular mass average of 21 881 Da for CLess Dps-2, consistent with a monomer. To assess the prediction from PISA analyses that dodecameric Dps-2 would be significantly more stable than an individual subunit, we compared thermal stability of Dps-2 and CLess Dps-2 using SYPRO Orange as a fluorescent reporter of protein unfolding (Figure 2C). Consistent with predictions, the melting temperature of Dps-2 is 68.1±0.0°C, consistent with complete assembly (and comparable with the melting temperature for dodecameric Dps-1 [27]), whereas that for CLess Dps-2 is 41.2±0.2°C.

Oligomeric assembly of Dps-2

Figure 2
Oligomeric assembly of Dps-2

(A) Elution of molecular mass standards (1.4, 17, 44, 158 and 670 kDa) measured by gel-filtration chromatography. Elution of Dps-2 and CLess Dps-2 indicated by arrows. Inset shows native gel electrophoresis of Dps-dn (molecular mass 17 kDa, pI 4.7; lane 1), Dps-2 (lane 2), Dps-2 and CLess Dps-2 mixed together (lane 3), and CLess Dps-2 (molecular mass 23 kDa, pI 5.3; lane 4). (B) Equilibrium sedimentation profile of CLess Dps-2. Lower panel shows the absorbance of 80 μM CLess Dps-2 as a function of the radial cell position. Data were fitted to a model describing a single non-associating species using the SEDFIT software [49]. The upper panel shows the residuals to the fit, which were distributed randomly. (C) Thermal stability of Dps-2 (○) and CLess Dps-2 (●), measured as fluorescence emitted from SYPRO Orange upon binding to denatured protein as a function of temperature.

Figure 2
Oligomeric assembly of Dps-2

(A) Elution of molecular mass standards (1.4, 17, 44, 158 and 670 kDa) measured by gel-filtration chromatography. Elution of Dps-2 and CLess Dps-2 indicated by arrows. Inset shows native gel electrophoresis of Dps-dn (molecular mass 17 kDa, pI 4.7; lane 1), Dps-2 (lane 2), Dps-2 and CLess Dps-2 mixed together (lane 3), and CLess Dps-2 (molecular mass 23 kDa, pI 5.3; lane 4). (B) Equilibrium sedimentation profile of CLess Dps-2. Lower panel shows the absorbance of 80 μM CLess Dps-2 as a function of the radial cell position. Data were fitted to a model describing a single non-associating species using the SEDFIT software [49]. The upper panel shows the residuals to the fit, which were distributed randomly. (C) Thermal stability of Dps-2 (○) and CLess Dps-2 (●), measured as fluorescence emitted from SYPRO Orange upon binding to denatured protein as a function of temperature.

Oligomeric state was also assessed by native gel electrophoresis, in which Dps-2 fails to migrate from the well with no faster migrating species detected (Figure 2A, inset, lane 2). The migration of CLess Dps-2 (Figure 2A, inset, lane 4) was compared with that of Dps-dn, a truncated derivative of Dps-1 shown previously to exist exclusively as a dimer (Figure 2A, inset, lane 1) [27]. Considering the shared four-helix bundle folds and the lower molecular mass of Dps-dn (17364 Da) as well as its lower pI of 4.7 compared with the 5.3 for CLess Dps-2, dimeric Dps-dn should migrate significantly faster than dimeric CLess Dps-2; the faster migration of CLess Dps-2 is therefore consistent with its existence as a monomer. Glutaraldehyde cross-linking also failed to detect any species other than the monomer for CLess Dps-2, whereas cross-linking of Dps-2 was very efficient, yielding no species smaller than a hexamer (results not shown). Glutaraldehyde reacts primarily with lysine to cross-link individual subunits. Inspection of the Dps-2 structure shows that most lysine residues extend from the ends of the four-helix bundle near the three-fold axis formed by three adjoining N-termini and that one lysine residue is located at the dimer interface, predicting cross-linking of dimeric species should they occur in solution. The failure of CLess Dps-2 to self-assemble is unexpected given that Dps proteins generally exist as dimers, trimers or dodecamers. Taken together, these results indicate that Dps-2 exists primarily as a dodecamer in solution, and that the C-terminal extension is required for oligomeric assembly.

Dps-2 would be predicted to exhibit ferroxidase activity. Indeed, the kinetics of Fe2+ oxidation in molecular oxygen shows the rapid production of Fe3+ by Dps-2. In contrast, CLess Dps-2 does not have ferroxidase activity, consistent with formation of the ferroxidase centre at the interface between two subunits, and reinforcing the conclusion that CLess Dps-2 exists as a monomer (Supplementary Figure S2 at http://www.BiochemJ.org/bj/447/bj4470381add.htm).

DNA binding and protection

Dps-1 binds DNA with high affinity (Kd of 0.5 nM for binding to 26 bp DNA [21,27]). Under identical conditions (500 mM NaCl), no complex may be detected with Dps-2, and neither does lowering the ionic strength (to 50 mM NaCl) result in a detectable complex (results not shown). In contrast, Dps-2 does bind plasmid DNA, as evidenced by a shift in mobility on agarose gels (results not shown). We also compared DNA binding by bipyridyl-treated Dps-2 and protein to which Mn2+ was added following bipyridyl treatment (Figure 3A). As reported previously, chelation of divalent metal from dodecameric Dps-1 destroys its ability to bind DNA, whereas addition of Mn2+ results in protein–DNA complexes that do not migrate from the well (Figure 3A, lanes 2 and 3); for Dps-1, this metal-dependence of DNA binding is associated with occupancy of the metal-binding site within the N-terminal extension [21]. Bipyridyl treatment of dodecameric Dps-2 also eliminates DNA binding, whereas Mn2+ addition restores complex formation (Figure 3A, lanes 4 and 5), perhaps due to a requirement for occupancy of the C-terminal metal site (oligomeric state of both Dps-1 and Dps-2 is unaffected by bipyridyl treatment [21], and results not shown). That complexes with dodecameric Dps-2 migrate much faster than those with Dps-1 may reflect both a lower affinity and that individual Dps-1 protomers are more prone to association with multiple DNA sites [31], thus creating large networks of DNA and protein that cannot migrate from the well.

DNA binding and protection by Dps-2

Figure 3
DNA binding and protection by Dps-2

(A) Binding of Dps homologues to plasmid DNA. Lane 1, DNA only; lanes 2 and 3, DNA and 10 pmol of Dps-1; lanes 4 and 5, DNA and 10 pmol of Dps-2. Reaction mixtures in lanes 2 and 4 contained bipyridyl-treated protein, and reaction mixtures in lanes 3 and 5 bipyridyl-treated protein with Mn2+. (B) Protection from hydroxyl-mediated DNA cleavage. Lane 1, DNA; lane 2, DNA degraded by hydroxyl radical; lane 3, DNA and 10 pmol of Dps-2; lane 4, DNA and Dps-2, incubated with H2O2 and Fe2+. (C) Protection from DNase I-mediated DNA cleavage. Reaction mixture in lane 1 contains DNA. Reaction mixtures in lanes 2 and 3 contain DNA incubated with DNase I for 10 and 5 min respectively. Reaction mixtures in lanes 4 and 5 contain DNA and 10 pmol of Dps-2 incubated with DNase I for 10 and 5 min respectively.

Figure 3
DNA binding and protection by Dps-2

(A) Binding of Dps homologues to plasmid DNA. Lane 1, DNA only; lanes 2 and 3, DNA and 10 pmol of Dps-1; lanes 4 and 5, DNA and 10 pmol of Dps-2. Reaction mixtures in lanes 2 and 4 contained bipyridyl-treated protein, and reaction mixtures in lanes 3 and 5 bipyridyl-treated protein with Mn2+. (B) Protection from hydroxyl-mediated DNA cleavage. Lane 1, DNA; lane 2, DNA degraded by hydroxyl radical; lane 3, DNA and 10 pmol of Dps-2; lane 4, DNA and Dps-2, incubated with H2O2 and Fe2+. (C) Protection from DNase I-mediated DNA cleavage. Reaction mixture in lane 1 contains DNA. Reaction mixtures in lanes 2 and 3 contain DNA incubated with DNase I for 10 and 5 min respectively. Reaction mixtures in lanes 4 and 5 contain DNA and 10 pmol of Dps-2 incubated with DNase I for 10 and 5 min respectively.

To ascertain whether Dps-2 affords DNA protection from ROS and DNase I, in vitro DNA protection assays were performed (Figures 3B and 3C). When pGEM5 is incubated with H2O2 and Fe2+ (Figure 3B, lane 2), DNA is degraded. Dps-2 protects DNA from ROS-mediated cleavage, with only a modest increase in nicked DNA species (Figure 3B, lanes 3 and 4). Similarly, incubation of DNA with DNase I resulted in complete DNA degradation (Figure 3C, lanes 2 and 3), whereas Dps-2 can afford some protection against cleavage, with nicked and linear DNA species still detectable after 10 min of incubation with DNase I (Figure 3C, lanes 4 and 5). In comparison, Dps-1 affords efficient protection against DNase I-mediated cleavage, consistent with its higher-affinity binding, but it cannot protect against ROS-mediated degradation, an effect ascribed to its inability to sequester iron stably [29]. Thus only Dps-2 conserves the ability to provide measurable protection against ROS-mediated degradation.

Nucleoid condensation by Dps-1

In E. coli, up-regulation of Dps in stationary phase leads to compaction of its nucleoid and formation of a ring-like structure [3739]. However, knocking out either Dps-1 or Dps-2 has no observable effect on nucleoid morphology in D. radiodurans, although Dps-1 was seen to associate with the nucleoid. Instead, the essential HU protein played a major role in maintaining a compact genome [32]. To determine whether Dps-1 and Dps-2 have the ability to compact genomic DNA, nucleoid condensation was visualized by DAPI staining of E. coli expressing the respective proteins. Dps-met, a mutant of Dps-1 that lacks the flexible N-terminal extension and hence is unable to bind DNA, was used as a control [21]. Overexpression was verified by SDS/PAGE (results not shown).

Cells expressing Dps-met have disperse nucleoids that span the length of the cell (Figure 4A) as seen previously for wild-type E. coli [40,41]. However, ~90% of cells expressing Dps-1 acquire compact spherical nucleoids indicating significant DNA condensation (Figure 4C). Consistent with such genomic DNA condensation, growth of cells expressing Dps-1 is adversely affected (results not shown). In contrast, cultures expressing Dps-2 maintain a disperse nucleoid similar to that seen in the Dps-met control or in cells not expressing any heterologous protein (Figure 4B). The Dps-1-mediated compaction of E. coli nucleoids was examined further by TEM (transmission electron microscopy). E. coli recombinants expressing plasmid-borne Dps-1 look significantly different from E. coli recombinants not expressing Dps-1 or the wild-type cells, consistent with Dps-1-mediated nucleoid condensation; in wild-type cells, the nucleoid is expansive and irregularly shaped, whereas the nucleoid in Dps-1-expressing cells is highly condensed (Supplementary Figure S3 at http://www.BiochemJ.org/bj/447/bj4470381add.htm).

In vivo nucleoid condensation

Figure 4
In vivo nucleoid condensation

E. coli BL21(DE3)pLysS expressing Dps-met (A), Dps-2 (B) and Dps-1 (C) were imaged. Nucleoid condensation is visualized by DAPI staining. Overlay images are a combination of both DIC (differential interference contrast) and DAPI images.

Figure 4
In vivo nucleoid condensation

E. coli BL21(DE3)pLysS expressing Dps-met (A), Dps-2 (B) and Dps-1 (C) were imaged. Nucleoid condensation is visualized by DAPI staining. Overlay images are a combination of both DIC (differential interference contrast) and DAPI images.

In vivo localization of Dps-2

The Dps-2 sequence includes a predicted signal peptide comprising the first 30 residues. We therefore constructed an in-frame fusion of GFP (green fluorescent protein) to the C-terminus of Dps-2. GFP has been suggested as a reporter for localization of proteins in E. coli, as it often fails to fold correctly in the oxidizing environment of the periplasm; overexpression of fusions to GFP that do not fluoresce would therefore be consistent with a periplasmic localization [42]. Consistent with these expectations, we found that the Dps-2–GFP fusion is abundantly expressed in E. coli (Supplementary Figure S4A at http://www.BiochemJ.org/bj/447/bj4470381add.htm), yet very few cells show any green fluorescence (Supplementary Figure S4B). Notably, in the few cells in which expression of the Dps-2–GFP may be detected, fluorescence is seen around the perimeter of the cells, primarily at the poles. No overlap of nucleoid-associated DAPI stain and GFP fluorescence is seen. This suggests that the signal peptide directs the Dps-2–GFP fusion to the periplasm. The absence of fluorescence from most cells suggests that most of the fusion protein fails to fold properly. Even cells in which fluorescence is seen may not express properly assembled protein; we consider it most likely that the bulky GFP impedes dodecameric assembly.

To determine the cellular localization in D. radiodurans, we therefore fused EGFP directly to the Dps-2 signal peptide with the construct under control of the Dps-2 promoter. To alleviate issues with slow folding of GFP resulting in poor fluorescence yield [43], cultures were grown at room temperature to reduce the growth rate. As shown in Figure 5, fluorescence may be seen around the perimeter of the cells, particularly visible at the junctions between cells within each tetrad, with the vast majority of cells expressing the reporter gene. The observed pattern of GFP expression is again consistent with a non-cytoplasmic localization.

In vivo localization of Dps-2

Figure 5
In vivo localization of Dps-2

(A) DIC (differential interference contrast) image of wild-type D. radiodurans (left-hand panel) and cells expressing EGFP fused to the Dps-2 signal peptide under control of the Dps-2 promoter (right-hand panel). (B) GFP fluorescence of wild-type and recombinant cells. (C) Overlay of DIC and green fluorescence. (D) Expanded view of D. radiodurans expressing EGFP fused to the Dps-2 signal peptide.

Figure 5
In vivo localization of Dps-2

(A) DIC (differential interference contrast) image of wild-type D. radiodurans (left-hand panel) and cells expressing EGFP fused to the Dps-2 signal peptide under control of the Dps-2 promoter (right-hand panel). (B) GFP fluorescence of wild-type and recombinant cells. (C) Overlay of DIC and green fluorescence. (D) Expanded view of D. radiodurans expressing EGFP fused to the Dps-2 signal peptide.

Activity of Dps-1 and Dps-2 gene promoters in response to ROS

Dps-1 is unusual in its inability to protect DNA from ROS [27,29], whereas Dps-2 conserves this property (Figure 3B). We therefore wondered whether activity of the respective gene promoters is modulated by ROS. Reporter constructs were used in which lacZ is under control of the respective promoters in either their forward or reverse directions, and β-galactosidase activity was measured in response to H2O2 or Fe2+ (Figure 6). As shown in Figure 6(A), the Dps-1 promoter does not respond to the addition of either H2O2 or Fe2+. In contrast, activity of the Dps-2 promoter is enhanced under both conditions, during both exponential and stationary phase growth (Figure 6C). The lower β-galactosidase activity measured during stationary phase can be attributed to a lower plasmid copy number as the plasmid used for these assays exists in a copy number equivalent to the genome content, which is lower in stationary phase [44,45]. As expected, little β-galactosidase is seen when promoters are in their reverse orientations (Figures 6B and 6D). As Dps-2 is expected to inactivate H2O2 (by using it as an oxidant in ferroxidase reactions), we also monitored the response of the HU promoter as a control; as shown in Figure 6(E), HU promoter activity is increased on addition of H2O2, indicating that H2O2 does enter the cytosol.

β-Galactosidase assay reporting on promoter activity in response to H2O2 or Fe2+

Figure 6
β-Galactosidase assay reporting on promoter activity in response to H2O2 or Fe2+

(A and B) β-Galactosidase activity of reporter construct with Dps-1 promoter in its forward and reverse orientations in untreated (white bars), H2O2-treated (dark grey bars) and iron-treated (light grey bars) cells. (C and D) β-Galactosidase activity with Dps-2 promoter in its forward and reverse orientations in untreated (white bars), H2O2-treated (dark grey bars) and iron-treated (light grey bars) cells. (E) β-Galactosidase activity with HU promoter in untreated (white bars) and H2O2-treated (grey bars) cells. Results are means+S.D. of three replicates.

Figure 6
β-Galactosidase assay reporting on promoter activity in response to H2O2 or Fe2+

(A and B) β-Galactosidase activity of reporter construct with Dps-1 promoter in its forward and reverse orientations in untreated (white bars), H2O2-treated (dark grey bars) and iron-treated (light grey bars) cells. (C and D) β-Galactosidase activity with Dps-2 promoter in its forward and reverse orientations in untreated (white bars), H2O2-treated (dark grey bars) and iron-treated (light grey bars) cells. (E) β-Galactosidase activity with HU promoter in untreated (white bars) and H2O2-treated (grey bars) cells. Results are means+S.D. of three replicates.

DISCUSSION

DNA binding and protection

Dps-2 contains a unique metal-bound C-terminal extension as well as an N-terminal extension that is not visible in the structure, probably due to flexibility [33]. Considering the content of positively charged residues in the C-terminal extension, this segment would be expected to contribute to DNA binding. Consistent with this prediction, Dps-2 binds DNA, albeit with low affinity (Figure 3A); the differential ability to detect complexes of Dps-2 with 26 bp DNA and plasmid DNA may reflect different complex stability in the two gel systems or stabilizing protein–protein interactions that occur on the longer DNA substrates. We also find that bipyridyl-treated dodecameric Dps-2 no longer binds DNA; a similar phenomenon was observed for Dps-1, where it is removal of metal from its N-terminal extension that is responsible for the loss of DNA binding [21,31]. A possible explanation for the failure of bipyridyl-treated Dps-2 to bind DNA is a loss of metal from the C-terminus that leads to a conformation of the C-terminal extension that is incompatible with DNA binding.

Dps-1 cannot protect DNA from ROS, and this unusual property was inferred to arise from release of iron through a novel exit channel [21,29]. That it can condense the E. coli nucleoid raises the possibility that it may instead protect DNA by means of its high-affinity binding. In vivo, Dps-1 is indeed associated with the nucleoid, but its deletion was reported not to cause significant changes in nucleoid morphology as was evident by DAPI staining of cells in exponential or early stationary phase; however, it is conceivable that other nucleoid-associated proteins such as HU may be up-regulated to compensate for the loss of Dps-1 [32].

In contrast, Dps-2 protects DNA from ROS-mediated cleavage. Its main function may therefore be to protect against exogenously derived ROS. Consistent with this interpretation, we find that only the Dps-2 promoter is sensitive to ROS (Figure 6). The inability of H2O2 to up-regulate the Dps-1 promoter was reported previously and suggested to be due to repression by OxyR; we note that the upstream edge of the promoter fragment cloned in front of lacZ (Figure 6) is within a few base pairs of the DNA fragment used to assess binding by OxyR [46]. That study did not reveal up-regulation of the Dps-2 promoter either as measured by qRT-PCR (quantitative real-time PCR), a difference perhaps due to different times of exposure to H2O2.

Dps-2 is non-cytoplasmically localized

During stationary phase, the E. coli nucleoid adopts highly ordered toroidal structures, which have been attributed to the up-regulation of Dps [3840]. This toroidal conformation is not seen on overexpression of E. coli Dps in exponentially growing cells, indicating additional contributing factors present during stationary phase. Likewise, overexpression of D. radiodurans Dps-1 in E. coli does not appear to result in toroidal DNA assemblies (Supplementary Figure S3). Whereas Dps-1-mediated effects on nucleoid organization could be indirect by affecting coupled transcription–translation processes [47], the DNA compaction seen both in vitro and in vivo would be consistent with a role for Dps-1 in nucleoid organization. Such functional roles may be global, provided sufficient cellular concentration, or local, as Dps-1 has been reported to bind DNA co-operatively [21,27,31].

The N-terminal signal peptide of Dps-2 directs the EGFP reporter to a non-cytoplasmic localization in D. radiodurans, with fluorescence particularly visible at the interface between cells within a tetrad. This is intriguing in the light of the reported sequestration of Fe2+ in D. radiodurans [11]; whereas Mn2+, which accumulates to millimolar concentrations is globally distributed, Fe2+ is largely sequestered outside the cytoplasm in a region overlapping the septum between dividing cells. Such distribution of Fe2+ would rationalize both the inability of cytoplasmically localized Dps-1 to protect DNA from OH derived from Fenton chemistry and the non-cytoplasmic localization of Dps-2 indicated in the present study (Figure 5). Only DpsA from Synechococcus sp. has been shown to have a non-cytoplasmic localization, partitioning to the thylakoid membrane where a role in metal transport was proposed [48].

Even localized iron pools have the potential for oxidative damage, and the Fe2+ must be sequestered to prevent damage to membrane proteins and other periplasmic components. D. radiodurans does not encode genes corresponding to the two major iron storage proteins in bacteria, i.e. ferritin or bacterioferritin. The observed localization of Dps-2 makes it a prime candidate for serving this iron-sequestering role as well as to serve as a first line of defence against exogenously derived ROS.

Abbreviations

     
  • DAPI

    4′,6-diamidino-2-phenylindole

  •  
  • Dps

    DNA protection during starvation

  •  
  • EGFP

    enhanced green fluorescent protein

  •  
  • EMSA

    electrophoretic mobility-shift assay

  •  
  • GFP

    green fluorescent protein

  •  
  • IPTG

    isopropyl β-D-thiogalactopyranoside

  •  
  • LB

    Luria–Bertani

  •  
  • NA

    numerical aperture

  •  
  • PISA

    Protein Interfaces, Surfaces and Assemblies

  •  
  • ROS

    reactive oxygen species

  •  
  • TEM

    transmission electron microscopy

AUTHOR CONTRIBUTION

All authors contributed to experimental design and data analysis and to writing the paper. Brian Reon cloned, expressed and purified Dps-2 and CLess Dps-2. He also performed gel-filtration analyses and microscopy in E. coli cells. Khoa Nguyen performed DNA-binding and β-galactosidase assays, analytical ultracentrifugation, thermal stability assays and microscopy in D. radiodurans cells. Gargi Bhattacharyya contributed to construction of lacZ reporter constructs and electron microscopy.

We thank C. Henk for her expert assistance with the electron microscopy, M. Lidstrom for providing plasmid pRADZ1, M.E. Newcomer for the use of her FPLC and C.C. Liu for assistance with analytical ultracentrifugation.

FUNDING

This publication was made possible by a grant from the Howard Hughes Medical Institute through the Undergraduate Biological Sciences Education Program to Louisiana State University and by the National Science Foundation [grant numbers MCB-0744240 and MCB-1051610 to A.G.].

References

References
1
Bauermeister
 
A.
Moeller
 
R.
Reitz
 
G.
Sommer
 
S.
Rettberg
 
P.
 
Effect of relative humidity on Deinococcus radiodurans’ resistance to prolonged desiccation, heat, ionizing, germicidal, and environmentally relevant UV radiation
Microb. Ecol.
2011
, vol. 
61
 (pg. 
715
-
722
)
2
Dartnell
 
L. R.
Hunter
 
S. J.
Lovell
 
K. V.
Coates
 
A. J.
Ward
 
J. M.
 
Low-temperature ionizing radiation resistance of Deinococcus radiodurans and Antarctic Dry Valley bacteria
Astrobiology
2010
, vol. 
10
 (pg. 
717
-
732
)
3
Mattimore
 
V.
Battista
 
J. R.
 
Radioresistance of Deinococcus radiodurans: functions necessary to survive ionizing radiation are also necessary to survive prolonged desiccation
J. Bacteriol.
1996
, vol. 
178
 (pg. 
633
-
637
)
4
Daly
 
M. J.
Minton
 
K. W.
 
An alternative pathway of recombination of chromosomal fragments precedes recA-dependent recombination in the radioresistant bacterium Deinococcus radiodurans
J. Bacteriol.
1996
, vol. 
178
 (pg. 
4461
-
4471
)
5
Makarova
 
K. S.
Aravind
 
L.
Wolf
 
Y. I.
Tatusov
 
R. L.
Minton
 
K. W.
Koonin
 
E. V.
Daly
 
M. J.
 
Genome of the extremely radiation-resistant bacterium Deinococcus radiodurans viewed from the perspective of comparative genomics
Microbiol. Mol. Biol. Rev.
2001
, vol. 
65
 (pg. 
44
-
79
)
6
Liu
 
Y.
Zhou
 
J.
Omelchenko
 
M. V.
Beliaev
 
A. S.
Venkateswaran
 
A.
Stair
 
J.
Wu
 
L.
Thompson
 
D. K.
Xu
 
D.
Rogozin
 
I. B.
, et al 
Transcriptome dynamics of Deinococcus radiodurans recovering from ionizing radiation
Proc. Natl. Acad. Sci. U.S.A.
2003
, vol. 
100
 (pg. 
4191
-
4196
)
7
Levin-Zaidman
 
S.
Englander
 
J.
Shimoni
 
E.
Sharma
 
A. K.
Minton
 
K. W.
Minsky
 
A.
 
Ringlike structure of the Deinococcus radiodurans genome: a key to radioresistance?
Science
2003
, vol. 
299
 (pg. 
254
-
256
)
8
Tanaka
 
M.
Earl
 
A. M.
Howell
 
H. A.
Park
 
M. J.
Eisen
 
J. A.
Peterson
 
S. N.
Battista
 
J. R.
 
Analysis of Deinococcus radiodurans's transcriptional response to ionizing radiation and desiccation reveals novel proteins that contribute to extreme radioresistance
Genetics
2004
, vol. 
168
 (pg. 
21
-
33
)
9
Zimmerman
 
J. M.
Battista
 
J. R.
 
A ring-like nucleoid is not necessary for radioresistance in the Deinococcaceae
BMC Microbiol.
2005
, vol. 
5
 pg. 
17
 
10
Daly
 
M. J.
Gaidamakova
 
E. K.
Matrosova
 
V. Y.
Vasilenko
 
A.
Zhai
 
M.
Venkateswaran
 
A.
Hess
 
M.
Omelchenko
 
M. V.
Kostandarithes
 
H. M.
Makarova
 
K. S.
, et al 
Accumulation of Mn(II) in Deinococcus radiodurans facilitates γ-radiation resistance
Science
2004
, vol. 
306
 (pg. 
1025
-
1028
)
11
Daly
 
M. J.
Gaidamokova
 
E. K.
Matrosova
 
V. Y.
Vasilenko
 
A.
Zhai
 
M.
Leapman
 
R. D.
Lai
 
B.
Ravel
 
B.
Li
 
S. M.
Kemner
 
K. M.
Fredrickson
 
J. K.
 
Protein oxidation implicated as the primary determinant of bacterial radioresistance
PLoS Biol.
2007
, vol. 
5
 pg. 
e92
 
12
Yu
 
T. W.
Anderson
 
D.
 
Reactive oxygen species-induced DNA damage and its modification: a chemical investigation
Mutat. Res.
1997
, vol. 
379
 (pg. 
201
-
210
)
13
Cornelis
 
P.
Wei
 
Q.
Andrews
 
S. C.
Vinckx
 
T.
 
Iron homeostasis and management of oxidative stress response in bacteria
Metallomics
2011
, vol. 
3
 (pg. 
540
-
549
)
14
Martinez
 
A.
Kolter
 
R.
 
Protection of DNA during oxidative stress by the nonspecific DNA-binding protein Dps
J. Bacteriol.
1997
, vol. 
179
 (pg. 
5188
-
5194
)
15
Grant
 
R. A.
Filman
 
D. J.
Finkel
 
S. E.
Kolter
 
R.
Hogle
 
J. M.
 
The crystal structure of Dps, a ferritin homolog that binds and protects DNA
Nat. Struct. Biol.
1998
, vol. 
5
 (pg. 
294
-
303
)
16
Zhao
 
G.
Ceci
 
P.
Ilari
 
A.
Gianqiacomo
 
L.
Laue
 
T. M.
Chiancone
 
E.
Chasteen
 
N. D.
 
Iron and hydrogen peroxide detoxification properties of DNA-binding protein from starved cells: a ferritin-like DNA-binding protein of Escherichia coli
J. Biol. Chem.
2002
, vol. 
277
 (pg. 
27689
-
27696
)
17
Gupta
 
S.
Chatterji
 
D.
 
Bimodal protection of DNA by Mycobacterium smegmatis DNA-binding protein from stationary phase cells
J. Biol. Chem.
2003
, vol. 
278
 (pg. 
5235
-
5241
)
18
Ceci
 
P.
Cellai
 
S.
Falvo
 
E.
Rivetti
 
C.
Rossi
 
G. L.
Chiancone
 
E.
 
DNA condensation and self-aggregation of Escherichia coli Dps are coupled phenomena related to the properties of the N-terminus
Nucleic Acids Res.
2004
, vol. 
32
 (pg. 
5935
-
5944
)
19
Stillman
 
T. J.
Upadhyay
 
M.
Norte
 
V. A.
Sedelnikova
 
S. E.
Carradus
 
M.
Tzokov
 
S.
Bullough
 
P. A.
Shearman
 
C. A.
Gasson
 
M. J.
Williams
 
C. H.
, et al 
The crystal structures of Lactococcus lactis MG1363 Dps proteins reveal the presence of an N-terminal helix that is required for DNA binding
Mol. Microbiol.
2005
, vol. 
57
 (pg. 
1101
-
1112
)
20
Su
 
M.
Cavallo
 
S.
Stefanini
 
S.
Chiancone
 
E.
Chasteen
 
N. D.
 
The so-called Listeria innocua ferritin is a Dps protein: iron incorporation, detoxification, and DNA protection properties
Biochemistry
2005
, vol. 
44
 (pg. 
5572
-
5578
)
21
Bhattacharyya
 
G.
Grove
 
A.
 
The N-terminal extensions of Deinococcus radiodurans Dps-1 mediate DNA major groove interactions as well as assembly of the dodecamer
J. Biol. Chem.
2007
, vol. 
282
 (pg. 
11921
-
11930
)
22
Ceci
 
P.
Mangiarotti
 
L.
Rivetti
 
C.
Chiancone
 
E.
 
The neutrophil-activating Dps protein of Helicobacter pylori, HP-NAP, adopts a mechanism different from Escherichia coli Dps to bind and condense DNA
Nucleic Acids Res.
2007
, vol. 
35
 (pg. 
2247
-
2256
)
23
Roy
 
S.
Saraswathi
 
R.
Gupta
 
S.
Sekar
 
K.
Chatterji
 
D.
Vijayan
 
M.
 
Role of N- and C-terminal tails in DNA binding and assembly in Dps: structural studies of Mycobacterium smegmatis Dps deletion mutants
J. Mol. Biol.
2007
, vol. 
370
 (pg. 
752
-
767
)
24
Ilari
 
A.
Latella
 
M. C.
Ceci
 
P.
Ribacchi
 
F.
Su
 
M.
Gianqiacomo
 
L.
Stefanini
 
S.
Chasteen
 
N. D.
Chiancone
 
E.
 
The unusual intersubunit ferroxidase center of Listeria innocua Dps is required for hydrogen peroxide detoxification but not for iron uptake: a study with site-specific mutants
Biochemistry
2005
, vol. 
44
 (pg. 
5579
-
5587
)
25
Schwartz
 
J. K.
Liu
 
X. S.
Tosha
 
T.
Diebold
 
A.
Theil
 
E. C.
Solomon
 
E. I.
 
CD and MCD spectroscopic studies of the two Dps miniferritin proteins from Bacillus anthracis: role of O2 and H2O2 substrates in reactivity of the diiron catalytic centers
Biochemistry
2010
, vol. 
49
 (pg. 
10516
-
10525
)
26
Ghatak
 
P.
Karmakar
 
K.
Kasetty
 
S.
Chatterji
 
D.
 
Unveiling the role of Dps in the organization of mycobacterial nucleoid
PLoS ONE
2011
, vol. 
6
 pg. 
e16019
 
27
Grove
 
A.
Wilkinson
 
S. P.
 
Differential DNA binding and protection by dimeric and dodecameric forms of the ferritin homolog Dps from Deinococcus radiodurans
J. Mol. Biol.
2005
, vol. 
347
 (pg. 
495
-
508
)
28
White
 
O.
Eisen
 
J. A.
Heidelberg
 
J. F.
Hickey
 
E. K.
Peterson
 
J. D.
Dodson
 
R. J.
Haft
 
D. H.
Gwinn
 
M. L.
Nelson
 
W. C.
Richardson
 
D. L.
, et al 
Genome sequence of the radioresistant bacterium Deinococcus radiodurans R1
Science
1999
, vol. 
286
 (pg. 
1571
-
1577
)
29
Kim
 
S. G.
Bhattacharyya
 
G.
Grove
 
A.
Lee
 
Y. H.
 
Crystal structure of Dps-1, a functionally distinct Dps protein from Deinococcus radiodurans
J. Mol. Biol.
2006
, vol. 
361
 (pg. 
105
-
114
)
30
Romão
 
C. V.
Mitchell
 
E. P.
McSweeney
 
S. M.
 
The crystal structure of Deinococcus radiodurans Dps protein (DR2263) reveals the presence of a novel metal centre in the N terminus
J. Biol. Inorg. Chem.
2007
, vol. 
11
 (pg. 
891
-
902
)
31
Nguyen
 
K. H.
Smith
 
L. T.
Xiao
 
L.
Bhattacharyya
 
G.
Grove
 
A.
 
On the stoichiometry of Deinococcus radiodurans Dps-1 binding to duplex DNA
Proteins
2012
, vol. 
80
 (pg. 
713
-
721
)
32
Nguyen
 
H. H.
de la Tour
 
C. B.
Toueille
 
M.
Vannier
 
F.
Sommer
 
S.
Servant
 
P.
 
The essential histone-like protein HU plays a major role in Deinococcus radiodurans nucleoid compaction
Mol. Microbiol.
2009
, vol. 
73
 (pg. 
240
-
252
)
33
Cuypers
 
M. G.
Mitchell
 
E. P.
Romão
 
C. V.
McSweeney
 
S. M.
 
The crystal structure of the Dps2 from Deinococcus radiodurans an unusual pore profile with a non-specific metal binding site
J. Mol. Biol.
2007
, vol. 
371
 (pg. 
787
-
799
)
34
Earl
 
A. M.
Rankin
 
S. K.
Kim
 
K. P.
Lamendola
 
O. N.
Battista
 
J. R.
 
Genetic evidence that the uvsE gene product of Deinococcus radiodurans R1 is a UV damage endonuclease
J. Bacteriol.
2002
, vol. 
184
 (pg. 
1003
-
1009
)
35
Bellapadrona
 
G.
Ardini
 
M.
Ceci
 
P.
Stefanini
 
S.
Chiancone
 
E.
 
Dps proteins prevent Fenton-mediated oxidative damage by trapping hydroxyl radicals within the protein shell
Free Radical Biol. Med.
2010
, vol. 
48
 (pg. 
292
-
297
)
36
Krissinel
 
E.
Henrick
 
K.
 
Inference of macromolecular assemblies from crystalline state
J. Mol. Biol.
2007
, vol. 
372
 (pg. 
774
-
797
)
37
Azam
 
T. A.
Iwata
 
A.
Nishimura
 
A.
Ueda
 
S.
Ishihama
 
A.
 
Growth phase-dependent variation in protein composition of the Escherichia coli nucleoid
J. Bacteriol.
1999
, vol. 
181
 (pg. 
6361
-
6370
)
38
Frenkiel-Krispin
 
D.
Levin-Zaidman
 
S.
Shimoni
 
E.
Wolf
 
S. G.
Wachtel
 
E. J.
Arad
 
T.
Finkel
 
S. E.
Kolter
 
R.
Minsky
 
A.
 
Regulated phase transitions of bacterial chromatin: a non-enzymatic pathway for generic DNA protection
EMBO J.
2001
, vol. 
20
 (pg. 
1184
-
1191
)
39
Frenkiel-Krispin
 
D.
Ben-Avraham
 
I.
Englander
 
J.
Shimoni
 
E.
Wolf
 
S. G.
Minsky
 
A.
 
Nucleoid restructuring in stationary-state bacteria
Mol. Microbiol.
2004
, vol. 
51
 (pg. 
395
-
405
)
40
Ohniwa
 
R. L.
Morikawa
 
K.
Kim
 
J.
Ohta
 
T.
Ishihama
 
A.
Wada
 
C.
Takeyasu
 
K.
 
Dynamic state of DNA topology is essential for genome condensation in bacteria
EMBO J.
2006
, vol. 
25
 (pg. 
5591
-
5602
)
41
Mukherjee
 
A.
DiMario
 
P. J.
Grove
 
A.
 
Mycobacterium smegmatis histone-like protein Hlp is nucleoid associated
FEMS Microbiol. Lett.
2009
, vol. 
291
 (pg. 
232
-
240
)
42
Feilmeier
 
B. J.
Iseminger
 
G.
Schroeder
 
D.
Webber
 
H.
Phillips
 
G. J.
 
Green fluorescent protein functions as a reporter for protein localization in Escherichia coli
J. Bacteriol.
2000
, vol. 
182
 (pg. 
4068
-
4076
)
43
Andrews
 
B. T.
Gosavi
 
S.
Finke
 
J. M.
Onuchic
 
J. N.
Jennings
 
P. A.
 
The dual-basin landscape in GFP folding
Proc. Natl. Acad. Sci. U.S.A.
2008
, vol. 
105
 (pg. 
12283
-
12288
)
44
Meima
 
R.
Lidstrom
 
M. E.
 
Characterization of the minimal replicon of a cryptic Deinococcus radiodurans SARK plasmid and development of versatile Escherichia coli–D. radiodurans shuttle vectors
Appl. Environ. Microbiol.
2000
, vol. 
66
 (pg. 
3856
-
3867
)
45
Meima
 
R.
Rothfuss
 
H. M.
Gewin
 
L.
Lidstrom
 
M. E.
 
Promoter cloning in the radioresistant bacterium Deinococcus radiodurans
J. Bacteriol.
2001
, vol. 
183
 (pg. 
3169
-
3175
)
46
Chen
 
H.
Xu
 
G.
Zhao
 
Y.
Tian
 
B.
Lu
 
H.
Yu
 
X.
Xu
 
Z.
Ying
 
N.
Hu
 
S.
Hua
 
Y.
 
A novel OxyR sensor and regulator of hydrogen peroxide stress with one cysteine residue in Deinococcus radiodurans
PLoS ONE
2008
, vol. 
3
 pg. 
e1602
 
47
Woldringh
 
C. L.
 
The role of co-transcriptional translation and protein translocation (transertion) in bacterial chromosome segregation
Mol. Microbiol.
2002
, vol. 
45
 (pg. 
17
-
29
)
48
Durham
 
K. A.
Bullerjahn
 
G. S.
 
Immunocytochemical localization of the stress-induced DpsA protein in the cyanobacterium Synechoccus sp. strain PCC 7942
J. Basic Microbiol.
2002
, vol. 
42
 (pg. 
367
-
372
)
49
Schuck
 
P.
 
A model for sedimentation in inhomogeneous media. I. Dynamic density gradients from sedimenting co-solutes
Biophys. Chem.
2004
, vol. 
108
 (pg. 
187
-
200
)

Author notes

1

These authors made an equal contribution to this work.

2

Present address: University of Virginia School of Medicine, Charlottesville, VA 22908, U.S.A.

Supplementary data