Novel DNA-protection Protein (Dps) Originated From Deinococcus Geothermalis

We characterized two interrelated proteins from the Deinococcus geothermalis strain with regard to the role of DNA-binding and protection, such as a novel Dps, Dgeo_0257, and a Dps orthologous protein, Dgeo_0281. Despite the lack of conserved amino acid sequence for ferroxidase activity, Dgeo_0257 exhibited high DNA-binding anity and formed a dodecameric conformation. In contrast, Dgeo_0281 protein showed less effective DNA-binding and was mainly observed monomeric or dimeric forms. Electrophoretic mobility shift assay showed that both puried proteins are non-specically bound to DNA to protect against DNA degradation and the properties of Dps responding to specic metal ions. In the presence of ferrous and ferric ions, Dgeo_0257 or Dgeo_0281 protein does not readily bind to DNA. However, in Zn and Mn ions are present, both proteins had more stable DNA-binding. From the Dps gene disrupted mutant test, each showed a higher sensitivity to oxidative stress than the wild-type strain. In addition, the expression level of each gene was correlated with the presence of each other. In this study, we characterized and conrmed the functional roles of Dgeo_0257 for sensing metal ions and binding to DNAs along with the Dps orthologous Dgeo_0281. Based on various observational evidence, we propose that Dgeo_0257 is a novel Dps protein with no ferroxidase activity. Experiments were performed to compare the expression levels of both dps genes dgeo_0281 (DgDps1) and dgeo_0257 (DgDps3) in the wild-type strain and proceed with Δdgeo_0257 and Δdgeo_0281 mutants. Control was based on the expression level of dgeo_0257 at O.D 600 2.0 of the wild-type strain. In the early stages of growth of wild-type D. geothermalis, DgDps1 (dgeo_0281) is predominant with over 150-fold, and as the growth phase stage progresses, the expression level of DgDps 1 is drastically decreased. Even if hydrogen peroxide is presented, it did not affect the expression of DgDps1 during the growth phases (Fig. 6A). The expression level of DgDps3 (dgeo_0257) is gradually increased when the cells reached the stationary phase with 2.46-fold. Expression of DgDps3 reached maximal level after treatment with 50 mM H 2 O 2 with 11.63-fold at O.D 600 4.0 (Fig. 6B). Therefore, it is expected that the DgDps1 plays a signicant role in the early exponential growth phase without oxidative stress. And DgDps3 plays preventing DNA damage under oxidative stress, especially the late exponential growth phase and growth phase-dependent slightly induction without oxidative stress in D. geothermalis. The Δdgeo_0257 mutant showed a half maximum expression level of DgDps1 with 60.69-fold at the beginning of growth and gradually became extinct on the growth progress without oxidative stress. However, the expression level of DgDps1was increased by two times at both O.D 600 4.0 and 8.0 under oxidative stress


Introduction
Deinococcus species are extremophiles that can survive under various harsh conditions, such as oxidative stress, desiccation, and extreme radioactive environments beyond normal bacterial environmental adaptation zone (Ferreira et al., 1997;Makarova et al., 2007). Deinococcus geothermalis is one of more than 90 species of the intense radiation-resistant Genus Deinococcus including mesophile Deinococcus radiodurans, which can optimally grow at 48°C and survive under acute and chronic exposures of ionizing radiation of in total 10 kGy of 60 Gy/hour, 1 kJ/m 2 of ultraviolet light, and years in desiccation (Daly, 2009; https://lpsn.dsmz.de/). D. geothermalis can reduce Fe (III)-nitrilotriacetic acid, U (VI), Cr (VI), and curtail Hg (II) in the presence of radiation and high temperature and utilized in the bioremediation (Brim et  Dps has been described as a DNA-binding protein that regulate and protects the DNA of starved Escherichia coli (Almirón et al., 1992). It plays an essential role in multifaceted protection by nonspeci c DNA binding, iron sequestration, and ferroxidase activity (Calhoun & Kwon, 2010). These properties appear to correlate with the structure and function of iron storage proteins of ferritin (FtnA) and bacterioferritin (Bfr). E. coli Dps has a shell-like structure and a spherical hollow cavity in the center. This hollow cavity of Dps acts as an iron storage compartment, and iron sequestration is essential in iron detoxi cation and homeostasis. As in Ftns, the internal cavity of the Dps protein is also negatively charged amino acid residues (Asp and Glu), which have a foundation for storing metal ions by electromagnetic force. Expression of Dps is dominantly induced both during starvation and in response to oxidative stress during the exponential phase (Almirón et al., 1992;Altuvia et al., 1994;Lomovskaya et al., 1994). The dps gene can be induced during OxyR-dependent growth by processing exponentially growing cells from E. coli with low concentrations of H 2 O 2 (Altuvia et al, 1994).
The N-terminal region of DrDps1 has a metal-binding site suggested to be important for DNA-binding Reon et al., 2012;Santos et al., 2017). The rst 30 amino acid residues of DrDps2 are expected to be signal peptides and were located in the cell periphery region based on intracellular iron localization and protected against exogenously derived ROS in D. radiodurans (Reon et al., 2012;Santos et al., 2015). Application of H 2 O 2 during the exponential cell growth phase results in changes in the oligomeric formation of DrDps1, which appears to be related to manganese homeostasis and detoxi cation of ROS. On the other hand, H 2 O 2 causes the accumulation of DrDps2 as a dodecameric cage form in stationary phases, which could be considered a 'typical Dps' response (Santos et al., 2015(Santos et al., , 2019. We focused on a novel Dps protein, Dgeo_0257 (DgDps3) in D. geothermalis. Both Dps proteins Dgeo_0257 and Dgeo_0281(DgDps1) were characterized using conformation analysis, DNA-binding capacity, and metal ions effects. Besides, results of the qRT-PCR analysis showed that each Dps gene was expressed dependently on growth phases using oxidative stress and gene expression levels. Thus, the ability of Dps to perform in a regulatory capacity remains, to some extent, speculative. Dgeo_0257 is proposed as one of the Dps DNA-binding proteins in D. geothermalis and its roles probably differ somewhat from those of DgDps1.

Bacterial strains and culture conditions
The strain D. geothermalis KACC 12208 derived from DSM 11300 T was obtained from the Korean Agricultural Culture Collection (KACC, http://www.genebank.go.kr). Dps gene disrupted mutants Δdgeo_0257 and Δdgeo_0281 were constructed in the previous study using the homologous recombination procedure to determine cell viability, growth pattern, and RNA-Seq analysis (Table S1). E. coli DH5α was routinely used for cloning and protein puri cation procedures. E. coli strains containing cloned recombinant plasmids were cultivated in LB broth medium containing 100 ppm of ampicillin at 37•C. For protein expression at the late-log phase (O.D 600 0.8) of E. coli, where the lac promoter is induced by 1 mM IPTG of end concentration and continuously cultured for 4 hours. After 4-hour induction, the culture was centrifuged at 5,000 rpm for 20 min and supernatants were discarded. Stored the pellets at -70 o C for use in further experiments. D. geothermalis was cultured at 48•C in TGY medium contained 1% tryptone, 0.5% yeast extract, and 0.1% glucose at 48°C. E. coli were cultured at 37°C in Luria-Bertani (LB) medium (MB Cell, Kisan Bio. Korea) contained 1% tryptone, 0.5% yeast extract, and 1% NaCl.

Construction of expression vector for Dgeo_Dps genes
The dgeo_0257 and dgeo_0281 genes were constructed and ampli ed using the PCR products encompassing both genes with different restriction enzyme sequences, double digested with NcoI and BglII for dgeo_0257, and NcoI and BamHI for dgeo_0281 and ligated into the pCS19 vector for the carboxyl-terminal His-tag version (Table S1). Then the recombinant plasmids were transformed into E. coli DH5α competent cells using the CaCl 2 chemical transformation method. The transformants were selected on a 100 µg/ml ampicillin-containing plate. Correct construction was detected using colony PCR with primer sets of target genes and nally decided by DNA sequence analysis (Macrogen, Korea).
Puri cation of a putative Dps protein (Dgeo_0257) and Dgeo_Dps1 (Dgeo_0281) Cell pellets stored in a -70 o C refrigerator were resuspended with 5 ml lysis buffer (300 mM NaCl, 50 mM NaH 2 PO 4 , pH8.0, and 10 mM Imidazole) per pellet obtained from 250 ml. Cell membranes were broken using a French Press (with 7,000 psi; Vision Biotech, Korea) or Sonicator (with Max 30% with 20 sec On and 40 sec Off time for three cycles, followed by Max 21%; Model CV334 of Sonics & Materials Inc., USA).
Following the breakage of cells, cells were centrifuged at 5,000 rpm for 20 min, and supernatants were collected for further testing. Nickel-nitrilotriacetic acid (Ni-NTA) resin column chromatography was used to purify over-expressed C-terminal His-tagged target proteins. After a Ni-NTA resin column equilibrated with lysis buffer, the sample solution was loaded and pass-through was collected. Next, 10-20 ml washing buffer (300 mM NaCl, 50 mM NaH 2 PO 4 , pH 8.0, 20 mM Imidazole) was poured. Finally, 10 ml elution buffer (300 mM NaCl, 50 mM NaH 2 PO 4 , pH 8.0, 250 mM Imidazole) was poured into resin to recover the target proteins. The Bradford assay was used to con rm protein production. SDS-PAGE electrophoresis with 12% separation gel was applied to determine the target protein pro les.
Electrophoresis was performed in 1x Tank buffer (25 mM Tris, pH 8.3, 192 mM glycine), followed by placement in staining buffer with 0.1% Coomassie Blue R250 in 10% acetic acid, 50% methanol, and 40% distilled water (DW), for at least 3 hours (usually left overnight). Destaining (10% acetic acid, 50% methanol, and 40% DW) was performed to con rm protein bands of interest visually. Protein dialysis was performed using a cellulose tubular membrane (CelluSep H1, Membrane Filtration Products Inc., USA) with 1x dialysis solution (25 mM Tris, pH 7.5, 150 mM NaCl). The membrane was placed in 1L dialysis solution and rotated to submerge it under water for 1 hour. The process was repeated three times, followed by protein concentration using the manufacturer's guidelines of VivaSpin (7,000 rpm for 15-20 mins, Vivaspin-6 of Sartorius, Germany).

Determination of protein conformation by gel ltration
Size-exclusion chromatography (SEC) was performed using the BioLogic HR Chromatography system (Bio-RAD, USA) with a HiLoad 16/600 Superdex 75 prep grade column (GE Healthcare Life Sciences, USA). The SEC column was lled with a running buffer solution (25 mM Tris-HCl, pH 7.5, 150 mM NaCl) to separate the molecules throughout the resin charged in the column, following the buffer to ow at a rate of 2.0 ml/min for 2 hours. The protein samples were loaded, rst protein size markers such as ferritin (440 kDa), conalbumin (75 kDa), ovalbumin (43 kDa), and anhydrase from erythrocytes, and puri ed proteins, with a ow rate of 0.5 ml/min for 1 min. The buffer was then drained at a ow rate of 2.0 ml/minute for 1 hour and dispersed in the fraction collector. Aliquots of critical parts that exceed the preset threshold values were collected and used in the following experiments.
Electrophoretic mobility shift assay (EMSA) Three putative promoters and the same number of open reading frame (ORF) regions were ampli ed from D. geothermalis genomic DNA by PCR using a forward and reverse primer set (Table S2). For the EMSA, DNA probes with ca 150-350 nucleotides length were puri ed using the AccuPrep® PCR puri cation kit (Bioneer, Korea). Various amounts of DNA and Dgeo_0257 protein were mixed in 5 x EMSA binding buffer Analysis of qRT-PCR according to growth phases in wildtype, Δdgeo_0257, and Δdgeo_0281 After overnight incubation, pre-cultured both wild-type and dps gene disrupted mutants, then used new 5 ml TGY broth with differential inoculation sizes as learned from the delayed growth patterns of mutant strains (approximately 1 hour). After 3 hours, measurement was performed at O.D 600 , adjusted to O.D 600 0.06, and started the main-culture in 50 ml TGY. When samples reached O.D 600 at 2.0, 4.0, and 8.0, centrifuged as 5 ml, 2.5 ml, and 1.25 ml each to match cell number, then discarded supernatants and resuspended 5 ml 0.9% NaCl, then treated 50 mM H 2 O 2 for 30 min. After the H 2 O 2 challenge, samples were centrifuged at 10,000 rpm for 5 minutes and washed once more with 0.9% NaCl. Supernatants were discarded and stored pellets at -20°C. Prepared samples without H 2 O 2 for the control group.
The cell wall was broken using phenol to determine the gene expression level. DNA digestion was performed using DNase I, and RNA was extracted using an RNA prep kit for RNA cleanup (RNeasy mini puri cation kit, Qiagen, Germany). After measuring the extracted RNA concentration, the concentration was made constant of 1,000 ng in 8 µl volume. Then cDNA synthesis was performed with dNTP mixture and 6-mer random primers using followed protocol: 60°C for 5 minutes, 4°C for 3 minutes, 30°C for 10 minutes, 42°C for 60 minutes, and nally 95°C for 5 minutes (PrimeScript™ 1st strand cDNA Synthesis Kit, TaKaRa, Japan). In the step of 3 minutes at 4°C, 4 µl 5x buffer, 4.5 µl RNase free water, 1 µl RTase, and 0.5 µl RNase inhibitor were added. Each primer prepared to determine the expression levels of dgeo_0257 and dgeo_0281 was added to the created cDNA. And then, qRT-PCR analysis was performed using TB Green® Premix Ex Taq™ (TaKaRa, Japan) on the Bio-Rad RT-PCR model CFX96™ Optics Module (Bio-Rad, USA). The expression level of GAPDH, a constitutively expressed gene on all growth phases, was normalized. The related expression levels of both dps genes were calculated from the previous study (Choo et al., 2020). The statistical analysis was done by probability t-test for differences between the samples in Prism™ program and it was considered signi cant at p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****).

Results
Biochemical properties of a putative DgDps D. geothermalis has 222 aa-long DrDps1 orthologous gene dgeo_0281 with 76.3% amino acid sequence identity, but no gene is equivalent to DrDps2. However, the putative Dps protein Dgeo_0257 of 200 aalength is a hypothetic protein of D. radiodurans with 65.5% amino acid sequence identity when 40 aa is deleted in the N-terminal region of the orthologous protein DR0582 ( Fig. 1 and Table 1). Dgeo_0257 has shown over 90% identity to match a DNA-binding protein with 194 aa-long from D. apachensis, D. metallilatus, D. aerius, and D. phoenicis genomes. D. marmoris contains two Dgeo_0257 like proteins: 205 and 210 aa-long with 71 % and 69 % identity, respectively (data not shown). However, Dgeo_0257 has not yet been characterized and revealed only 11.5% amino acid sequence identity with Dgeo_0281 as DgDps1. Therefore, gene disruption and protein puri cation experiments should be conducted to determine functional roles to con rm that Dgeo_0257 is a functional DNA-protecting Dps protein without a ferroxidase center. Puri cation and conformation of Dgeo_0257 and Dgeo_0281 proteins The Dgeo_0257 and Dgeo_0281 proteins were overexpressed in E. coli by attaching C-terminal histidine tags to elucidate the functional role of each protein. After the rst stage of protein puri cation using a Ni-NTA column, analysis of the protein pro le of the dgeo_0257 or dgeo_0281 gene with 6x His-tag was performed using SDS-PAGE. Both proteins were well induced by IPTG triggering at a concentration of 1 mM. The biochemical molecular weights of the monomers of Dgeo_0257 and Dgeo_0281 are approximately 23 kDa and 26 kDa. However, Dgeo_0281 was over 30 kDa in SDS-PAGE gel. The Dgeo_0257 and Dgeo_0281 proteins were puri ed at ca 90% purity per liter of culture in the rst stage of one-step puri cation at 5 mg and 20 mg, respectively (Fig. S1).
Various sizes of protein forms were collected and con rmed using gel ltration. We expected that Dgeo_0257, a 23 kDa monomer with His-tag, would form a 12-mer or 24-mer polymer or a sphere of 240 kDa or 480 kDa like Dps1 and Dps 2 of D. radiodurans. The gel ltration analysis results showed that the intact protein size of Dgeo_0257 was ca 480 kDa representing the form of icositetrahedron compared to the dimer and monomer of Dgeo_0281 as DgDps1. The conformation did not affect the elution of the Dgeo_0281 protein in the presence of iron (Fig. S2). However, the elution fraction of Dgeo_0257 moved to the posterior region when 1 mM of ferrous iron was present. Adversely, the elution fraction moved to the anterior part when 1 mM of ferric iron was present. Therefore, Dgeo_0257 protein conformation is changed by the iron redox state from the dodecameric form when the iron is absent.

DNA-binding and protection properties of DgDps proteins
In general, bacterial Dps proteins have non-speci c DNA-binding activity (Almirón et al., 1992; Calhoun & Kwon, 2010). To con rm this phenomenon, we prepared three intergenic regions with long-distance divergent promoters and three ORF regions in Dgeo_2112 as a hypothetic protein with a 250-350 bp-long DNA fragment, and followed by EMSA detection with 6% native gel (Table S2). Both Dgeo_0257 and Dgeo_0281 showed distinct DNA-binding a nity with DNA fragments. The DNA-protein complex was in the super-shift position for Dgeo_0257. However, the DNA-protein complex was in the lower position for Dgeo_0281. We identi ed the unique DNA binding behavior of two Dps putative proteins. Dgeo_0257 showed DNA binding activity over 5-fold higher than Dgeo_0281 (Fig. 2). Furthermore, Dgeo_0257 protein required 10 µM protein concentration for a complete shift-up migration, but Dgeo_0281 required more than 60 µM protein concentration for full DNA binding. Interestingly, the DNA-binding behavior of Dgeo_0257 and Dgeo_0281 is different between the promoter region and the ORF region. Treatment with ferrous iron resulted in the complete release of the promoter region DNA from the protein-DNA complexes. However, the ORF region DNA was still bound to Dgeo_0257 (partially) and Dgeo_0281 (wholly) with 0.2 mM and 1.0 mM ferrous ions, respectively (Fig. 2).
To examine the DNA protection activity of Dps, we found that DNA disappears in the presence of an iron ion of 0.5 mM. This phenomenon is because unprotected naked DNA is damaged and degraded due to oxidative stress through Fenton reactions by ferrous ions or ferrous ions and hydrogen peroxide mixtures. However, the DNAs were well protected from oxidative stress with the addition and increase of Dgeo_0257 and Dgeo_0281 proteins (Fig. 3). Furthermore, DNA-protein complexes were located in the lower position in the ORF DNA region where ferrous ions existed. Therefore, these two Dps proteins of D. geothermalis are DNA-binding proteins with non-speci c DNA sequence targets with distinct patterns together with ferrous ions in both promoter and ORF regions and can be considered to have DNA protection against oxidative stress reactions.
Metal ion sensing on protein behavior of a novel Dps (DgDps3) and DgDps1 Various metal ions were added to the Dps and DNA complex formation reactions to determine the metal ion effect on the Dps-DNA binding activity. As a result of increased ferric ion concentrations, the DNAprotein complex did not maintain a complex form, and free DNAs were then released from Dps-DNA complexes at more than 0.5 mM and 1 mM Fe 3+ , respectively (Fig. 4A). Ferric ions also showed similar DNA releasing activity to ferrous ions at more than 2-3-fold higher concentrations (Fig. S3A). Free DNA was degraded by a high concentration of ferrous ions (Lane 8-10 in Fig. S3A). Interestingly, as the concentration of manganese and zinc ions was increased, the two Dps proteins were much better bound to DNA, which stabilized the DNA binding even more (Fig. 4BC). Surprisingly, in the presence of more than 2 mM Mn, a novel Dgeo_0281 protein-DNA complex was observed (Fig. 4B). The reaction to manganese is similar to that reported by previous research groups (Nguyen & Grove, 2012; Reon et al., 2012). However, the zinc effect on DNA-binding stabilization is a nding in this study. Lead, copper, and cobalt did not affect DNA-binding ability (Fig. S3BCD). However, DNA-binding activity varied with chromate and cesium depending on metal ion concentration. At a particular concentration, Dgeo_0257 and Dgeo_0281 protein showed a uctuated pattern depending on cesium and chromate concentrations, respectively (Fig.  S3EF). Although the amino acid sequence equivalence of the novel Dps protein Dgeo_0257 (DgDps3) and DgDps1 (Dgeo_0281) was very low, both proteins have the same behavior in DNA-binding capabilities by metal ion sensing with differential conformation and DNA-binding a nity.

Physiological properties of Dps disrupted mutants
In our previous study, disrupted mutant strains of putative Dps dgeo_0257 and dgeo_0281 (Δdgeo_0257 and Δdgeo_0281) were developed by homologous recombination and kanamycin selection (Lee et al., 2019). In the absence of oxidative stress, two mutant strains showed an almost identical growth pattern with 1 hour delayed to the wild-type strain (Fig. S4A). In the late exponential growth of O.D 600 1.47, each strain showed different growth inhibition by treatment with 100 mM H 2 O 2 (Fig. 5A). The viability test was performed to con rm the oxidative stress response among the WT, Δdgeo_0257, and Δdgeo_0281 mutants by treatment with H 2 O 2 concentrations of 80, 100, and 120 mM, respectively. The Δdgeo_0281 mutant strain showed a slight decrease in viability at 80 mM, and cells did not grow above 100 mM H 2 O 2 .
The Δdgeo_0257 mutant strain endured well to 100 mM of slightly higher H 2 O 2 concentrations but did not grow at 120 mM (Fig. 5B). The WT and two Dps gene disrupted mutants were not grown at O.D 600 0.5 with 100 mM H 2 O 2 treatment (Fig. S4B). Thus, both dps genes participated in anti-oxidation responses.

Expression levels of two Dps genes in a growth phasedependent manner
Experiments were performed to compare the expression levels of both dps genes dgeo_0281 (DgDps1) and dgeo_0257 (DgDps3) in the wild-type strain and proceed with Δdgeo_0257 and Δdgeo_0281 mutants. Control was based on the expression level of dgeo_0257 at O.D 600 2.0 of the wild-type strain. In the early stages of growth of wild-type D. geothermalis, DgDps1 (dgeo_0281) is predominant with over 150-fold, and as the growth phase stage progresses, the expression level of DgDps 1 is drastically decreased. Even if hydrogen peroxide is presented, it did not affect the expression of DgDps1 during the growth phases (Fig. 6A). ). However, in Chip-seq analysis, the Dps protein of E. coli was reported to combine genomic DNA in a non-random manner (Antipov et al., 2017). DgDps1 (Dgeo_0281) and Dgeo_0257 also showed different DNA-binding speci city in the selected promoter and ORF regions (Fig. 2,4). The self-assembly characteristics and DNA-binding properties of Dps in solution are extensively agglomerated according to co-crystallization and general defense mechanism (Ceci et al., 2004  VCO139 protein was strongly induced by exposure to hydrogen peroxide in a Gram-negative bacterium Vibrio cholerae (Xia et al., 2017). This protein encodes a Dps homologous protein and has ROS resistance. OxyR regulates Dps during the exponential phase and by RpoS during the stationary phase. This Dps protein can be expressed relatively without H 2 O 2 . However, Dps is dramatically induced in the exponential growth phase but less induced in the stationary growth phase under H 2 O 2 . The involvement of Dps in resistance to various environmental stresses has also been reported. For example, exposure of the Δdps strain to high iron levels during starvation resulted in a less viable Δdps strain compared with wild-type. These data suggest that Dps is vital for the survival of bacteria in starved cells due to its reaction to ROS and iron toxicity tolerance. As with V. cholerae, Dps is related to colonization but does not involve toxic gene expression (Xia et al., 2017).
Interestingly, Campylobacter jejuni Dps has unusually high-temperature tolerance and activates DNA binding by ferrous ions. However, this DNA interaction has been inhibited by NaCl and Mg 2+ (Huergo et al., 2013;Sanchuki et al., 2015). In the hyperthermophilic Crenarchaeon S. solfataricus, Dps protein was directly induced by H 2 O 2 treatment and ferrous ion depletion (Wiedenheft et al., 2005).
Dps protein forms a dimer or dodecamer structure and binds to DNA (Grove & Wilkinson, 2005). In DrDps1 of D. radiodurans, oligomeric states change depending on the growth phase and presence of DNA. This suggests that DrDps1 oligomeric forms and functional roles may adapt to environmental changes (Santos et al., 2015). On the other hand, the DrDps2 oligomeric form maintains a dodecameric structure without affecting the surrounding environment. DrDps2 is thought to be more selective about iron depending on the cell environment and involved in intracellular metal storage (Reon et al., 2012;Santos et al., 2015).
In this study, the protein form of Dgeo_0281, known as DgDps1, the homologous protein of DrDps1, was not identi ed in the dodecameric form. The maximum combining in gel ltration and EMSA test was the dimer structure. The DNA binding a nity was Kd 30 µM, half the maximum of the non-speci c DNA sequence. Interestingly, the novel Dps protein (Dgeo_0257, called DgDps3) showed higher DNA-a nity than DgDps1 (Dgeo_0281) in just 10 µM. EMSA showed that the DgDps3-DNA complex was located near the upper pocket, unlike the DgDps1 found below. In the presence of metal ions together with DNA and protein, EMSA showed that those two DgDps proteins carried out the identical electrophoretic shift with DNA. Therefore, further analysis of protein structure conversion by metal ion and DNA-binding is needed.
Ferritin is a 24-mer protein with an inner diameter of 8 nm and stores 24,000 Fe atoms in vivo (Lawson et al., 1991;Harrison & Arosio, 1996;Yokoyama et al., 2012). However, the inner diameter of 5 nm Dps protein provides storage space for less than 500 iron atoms (Tonello et al., 1999;Ramsay et al., 2006). Thus, DNA-binding Dps proteins can store iron in a bioavailable form and protect cells against oxidative stress. The oxidative protection is achieved by binding Fe 2+ ions and preventing the Fenton reactioncatalyzed formation of toxic hydroxyl radicals or binding DNA to block them from oxidative radicals (Imlay et al., 2013). Despite knowledge of the basic mechanisms of iron entry and oxidation, many essential questions regarding iron core formation and the iron release mechanism in response to cellular needs still need to be elucidated (Papinutto et al., 2002 was absent, substitutional induced at early exponential growth phase (Fig. 6). Therefore, the novel DgDps3 (Dgeo_0257) is proposed as one of the Dps DNA-binding proteins in D. geothermalis and will play a different role than DgDps1 (Dgeo_0281). Further investigation into DgDps3 is required to explain this protein´s function, structure, and network regulation through redox-sensing regulators in D. geothermalis.

Declarations Con ict of interest
The authors declare that they have no con ict of interest.
Author contributions MKB, ES, and SJL designed the experiments, analyzed the data, and wrote the manuscript. The ferroxidase active center is missed in DgDps3 (C). Red words indicated the predicted DNA-binding domains (HTH) by prabi (https://npsa-prabi.ibcp.fr/).   A, Comparison of growth pattern among wild-type, Δdgeo_0257, and Δdgeo_0281 mutants on TGY medium and treatment of 100 mM H2O2 at O.D600 1.47. Optical density was measured at 600 nm wavelength. B, Viability test by hydrogen peroxide treatment of 80, 100, and 120 mM among wild-type, Δdgeo_0257, and Δdgeo_0281 mutants.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.