OUP user menu

How radiation kills cells: Survival of Deinococcus radiodurans and Shewanella oneidensis under oxidative stress

Debabrota Ghosal , Marina V. Omelchenko , Elena K. Gaidamakova , Vera Y. Matrosova , Alexander Vasilenko , Amudhan Venkateswaran , Min Zhai , Heather M. Kostandarithes , Hassan Brim , Kira S. Makarova , Lawrence P. Wackett , James K. Fredrickson , Michael J. Daly
DOI: http://dx.doi.org/10.1016/j.fmrre.2004.12.007 361-375 First published online: 1 April 2005


We have recently shown that Deinococcus radiodurans and other radiation resistant bacteria accumulate exceptionally high intracellular manganese and low iron levels. In comparison, the dissimilatory metal-reducing bacterium Shewanella oneidensis accumulates Fe but not Mn and is extremely sensitive to radiation. We have proposed that for Fe-rich, Mn-poor cells killed at radiation doses which cause very little DNA damage, cell death might be induced by the release of Fe(II) from proteins during irradiation, leading to additional cellular damage by Fe(II)-dependent oxidative stress. In contrast, Mn(II) ions concentrated in D. radiodurans might serve as antioxidants that reinforce enzymic systems which defend against oxidative stress during recovery. We extend our hypothesis here to include consideration of respiration, tricarboxylic acid cycle activity, peptide transport and metal reduction, which together with Mn(II) transport represent potential new targets to control recovery from radiation injury.

  • Oxidative stress
  • UV
  • Desiccation
  • Ionizing radiation
  • Cytochromes
  • Flavins
  • Nucleoid
  • Manganese
  • Iron
  • Shewanella
  • Deinococcus
  • Kineococcus

1 What makes Shewanella oneidensis so sensitive to ionizing radiation?

The central dogma of radiation biology is that the cytotoxic and mutagenic effects of radiation are principally the result of DNA damage caused during the course of irradiation. This might not always be the case since environmental organisms such as Shewanella oneidensis (MR-1) (ATCC 700550) [1], which encode relatively complex DNA repair systems [2, 3], are killed at radiation doses that cause relatively little DNA damage. The ionizing radiation doses that yield 17% survival of Escherichia coli and Deinococcus radiodurans are higher by factors of 20 and 200, respectively, than those for S. oneidensis [2]. Whereas 90% of S. oneidensis cells do not survive 70 Gy, a dose that induces less than one DNA double strand break (DSB) per genome, 10% of D. radiodurans cells survive 12,000 Gy, a dose that induces 120 DSBs per genome [2], and 10% of E. coli survive 700 Gy, a dose that induces 7 DSBs per genome (Table 1) [2]. Drying cells is also known to cause genomic DSBs [4], and S. oneidensis is killed after exposure to desiccation for only one day whereas similarly treated D. radiodurans can survive for months (Table 1) [2]. When the generation of reactive oxygen species (ROS) (superoxide, hydrogen peroxide and hydroxyl radicals) produced by irradiation or metabolism exceeds the capacity of endogenous scavengers to neutralize them, cells become vulnerable to damage, a condition referred to as oxidative stress (Fig. 1) [5, 6].

View this table:

Relationship between c-type cytochrome number, Mn/Fe levels and resistance

Strain (genomic sequence)Genome size (Mb)Total no. of c-type cytochromesa Intra-cellular Mn/Fe ratiobD 10, IR survival (kGy)bD 10, desiccation survival (days)
cDeinococcus radiodurans3.2870.2410–1230
dDeinococcus geothermalis3.2370.461030
eEscherichia coli4.6460.0070.77
fPseudomonas putida6.1817<0.0010.252
gShewanella oneidensis5.1339<0.0010.071

Overview of the basic metabolic pathways of D. radiodurans and systems that generate and defend against ROS (reactive oxygen species). The pathways are color-coded as follows: Black, catabolism of fructose, lactate and acetate; pyruvate oxidative decarboxylation; TCA cycle; and carotenoid biosynthesis. Brown, reactions of the glyoxylate bypass. Blue, methionine, serine biosynthesis, and electron transfer coupled to ATP synthesis. Purple, cysteine biosynthesis and sulfate reduction. Green, basic ammonia metabolism. For each reaction, the gene name and/or D. radiodurans gene number is shown. Red crosses show the location of predicted disruptions (missing genes) of biosynthetic pathways. Red arrows and boxes show the dominant reactions leading to the generation of ROS. Black boxes include potential defense systems against free radical damage. Black box around Mn(II), accumulated non-enzymic Mn(II)-based scavengers of Embedded Image and related ROS. Blue boxes outside the cell show proteases and substrates: blue substrates in these boxes have predicted specified transporters, orange substrates do not have predicted specified transporters, but some are believed to be transported based on experimental evidence (Table 2). Long dashed blue lines show conventional pathways leading to degradation of amino acids. Long solid blue lines indicate transport based on experimental evidence (Table 2). Inside the cell, blue boxes indicate substrates that support moderate/poor growth, orange ovals indicate substrates that were tested and shown not to support growth (Table 2). Abbreviations: APS, adenylylsulfate; PAPS, phosphoadenylylsulfate; NAD, nicotinamide adenine dinucleotide (oxidized form); NADH, nicotinamide adenine dinucleotide (reduced form); NADP, nicotinamide adenine dinucleotide phosphate (oxidized form); NMN, nicotinamide mononucleotide; FMN, flavin mononucleotide; FAD, flavin adenine dinucleotide (oxidized form); FADH, flavin adenine dinucleotide (reduced form); ATP, adenosine triphosphate; ADP, adenosine diphosphate; PEP, phosphoenolpyruvate; nicotinate-AD, nicotinic acid adenine dinucleotide; nicotinate-NM, nicotinic acid mononucleotide; E1, subunit of phosphotransferase system; Pi, orthophosphate; PPi, pyrophosphate; Q, ubiquinone; SOD, superoxide dismutase; Embedded Image, superoxide radical; Embedded Image, peroxide ion; H2O2, hydrogen peroxide; HO, hydroxyl radical; and Embedded Image, hydrated electrons.

Until recently, there have been no clear physiologic predictors of a cell's ability to recover from radiation or desiccation. In general, most of the resistant bacteria reported have been Gram-positive and the most sensitive have been Gram-negative [7, 8]. However, there are several reported exceptions to this paradigm, the Gram-negative cyanobacterium Chroococcidiopsis is extremely radiation- and desiccation-resistant [9], whereas the Gram-positive Micrococcus luteus (Sarcinia lutea) is sensitive [1012]. We recently reported that the differences in resistance to γ-radiation and desiccation for different bacteria mirror their intracellular Mn/Fe concentration ratios, where very high, moderate and very low Mn/Fe ratios correlate with very high, moderate and very low resistances, respectively [2]. D. radiodurans (Mn/Fe ratio: 0.24) accumulates 150 times more Mn than S. oneidensis (Mn/Fe ratio: 0.0005) and is sensitized to ionizing radiation when Mn(II) is restricted, and S. oneidensis accumulates 3.3 times more Fe than D. radiodurans (Table 1) [2]. In the case of S. oneidensis exposed to doses ⩾70 Gy, Fe(II)-dependent oxidative stress produced during recovery might lead to additional DNA, RNA, lipid and protein damage [2].

2 Competing views of radiation resistance mechanisms

Studies have shown that for a given dose of ionizing radiation, the number of DNA double strand breaks in D. radiodurans compared to S. oneidensis is about the same, and that genomic DNA of D. radiodurans is not endowed with unusual protection from in vivo irradiation [2, 13]. Thus, the key to surviving DNA damage appears to reside in the ability of D. radiodurans to repair DNA, and two competing views of the mechanisms responsible for its extraordinary survival now dominate [2]: (i) there are repair functions encoded among its hypothetical genes, but most novel genes implicated in recovery by transcriptome profiling have had little effect on survival when disrupted [14]; or (ii) a combination of non-enzymic and enzymic antioxidant defenses allow relatively conventional repair pathways to function with greater efficiency than in other bacteria [2]. A third view, that highly condensed ‘ringlike’ nucleoids (RNs) are key to resistance [15] is generally discounted [2, 16]. Specifically, Levin-Zaidman et al. [15] have proposed that each cell of a typical D. radiodurans cluster (tetracoccus, four cells) (Fig. 2A) contains a single copy of its genome partitioned as a ringlike nucleoid (RN), and following irradiation, is involved in a two-stage repair process. Where, in the first stage, DNA fragments arising from DNA double strand breaks are prevented from defusing by a ridged, toroid-shaped nuclear matrix and undergo error-free blunt-end joining; and in the second stage, RNs in adjacent cells fuse and undergo template-specific recombination. The overriding reasons why such a model is unlikely to be correct include: (i) each cell of tetracoccus contains 4–8 copies of its genome [17]; (ii) blunt-end joining of chromosomal DNA double strand breaks in D. radiodurans is not observed [18]; (iii) D. radiodurans cells grown to the logarithmic-phase in defined minimal medium do not display condensed RNs (Fig. 2B), but are equally resistant [2]; (iv) several other highly resistant deinococci do not form condensed RNs [2] (http://www.usuhs.mil/pat/deinococcus/index_20.htm); and (v) the relatively sensitive bacterium E. coli, which accumulates 18 times less intracellular Mn than D. radiodurans [2], displays RN nucleoid morphologies by transmission electron microscopy (TEM) [19, 20] and fluorescent microscopy [21], but cross-sections support sphere/coralline-like structures with central cores of unknown composition [19], not toroids; sections through random planes of highly condensed late stationary-phase D. radiodurans nucleoids typically reveal circular-shaped structures [2, 22] (Fig. 2). For S. oneidensis (Fig. 2C), we have proposed that ROS generated by energy-metabolism after irradiation during recovery might underlie this organism's great sensitivity [2]. One source of electrons which can leak from the substrate-side of the respiratory chain is heme-containing c-type cytochromes, yielding the reduced oxygen-product superoxide anion Embedded Image (Fig. 1) [5]; Embedded Image can be converted to other ROS such as hydrogen peroxide (H2O2) by dismutation, and the extremely reactive hydroxyl radical (HO) produced by Fe(II)-dependent reduction of H2O2 (Fenton-type chemistry) [5]. In this context, we note that S. oneidensis has one of the highest numbers of c-type cytochromes, as predicted from genome sequence, and iron contents reported for any bacteria (Table 1) [1, 2].


Transmission electron microscopy. Unpublished images derived from work reported in Daly et al. [2]. (A) D. radiodurans (ATCC BAA-816), tetracoccus, grown in TGY (rich medium) (stationary-phase); condensed RN morphologies are visible. (B) D. radiodurans diplococci, grown in defined minimal medium (logarithmic-phase); diffuse genomes. (C) S. oneidensis (MR-1) grown in DSM (defined Shewanella medium [2]), stationary-phase. (D–F) Three serial micrographs of the same D. radiodurans cell, grown in TGY (stationary-phase), support that the nucleoid has a sphere/coralline-like structure similar to that reported in E. coli [19]. Scale bars, 0.5 μm. See Appendix for experimental details.

3 Manganese-dependent bacteria

Whereas Mn(II) salts are soluble, Mn(III,IV) oxides are relatively insoluble at circaneutral pH, and both forms are widely distributed in the environment [23]. Manganese-dependent microorganisms such as Deinococcus, Arthrobacter, Bacillus, Streptococcus and cyanobacteria spp. have been implicated in the deposition of manganese oxides in dark manganiferric rock varnish coatings on desert rocks [2426]. Organisms that belong to those groups are known for their radiation and desiccation resistance, and our recent work establishes a link between the role of Mn(II) and an environmental niche they occupy [2]. The existence of high intracellular Mn/Fe concentration ratios in phylogenetically distant, radiation resistant bacteria but not in sensitive cells supports the idea that Mn(II) accumulation (with relatively low Fe) might be a widespread strategy that facilitates survival [2]. Accumulation of high intracellular levels of Mn(II) as a mechanism to fight Embedded Image-related ROS has been reported for a variety of bacteria including Lactobacillus, Borrelia, Neisseria, Deinococcus and Enterococcus spp. [2, 2729].

Fig. 3 shows that D. radiodurans (R1 and 7 b-1) [2, 30] and Deinococcus geothermalis [31] are able to utilize colloidal Mn(IV) oxides for growth in defined minimal medium [2]. These organisms grew equally well when supplemented with Mn(IV) oxides in place of MnCl2, suggesting that they possess the ability to reductively mobilize solid-phase Mn(IV), although abiotic reduction to Mn(II) by growth medium components can not be ruled out. We have previously reported metal reductase activities in these organisms [31, 32], and it is possible that such functions might facilitate the reductive assimilation of environmental sources of Mn(III,IV) oxide. The mechanism by which high levels of intracellular Mn(II) scavenge Embedded Image and related ROS in the absence of superoxide dismutase (SOD) and catalase is not understood [33], nor is it known if Mn(II) acts as bound ions or as non-enzymic catalytic complexes. Reactions that might be relevant to the scavenging ability of intracellular Mn(II) include those which mimic activities of SOD: Embedded Image; oxidases: Embedded Image oxides; and deprotonation: Embedded Image[3335].


Growth with Mn oxides. (a) D. radiodurans R1 (ATCC BAA-816), (b) D. radiodurans 7b-1 [30], and (c) Deinococcus geothermalis (DSM 11300) [62]. No metal, no Mn or Fe added. Mn sources: Colloidal Mn(IV) oxide, product of permanganate oxidation; Pyr, pyrolusite, β-MnO2 (mineral); MnCl2, manganous (Mn(II)) chloride; MnO2, Mn(IV) oxide. See Appendix for experimental details.

Nucleic acids coordinate Fe(II) very well [36]. However, even though Fe levels in D. radiodurans are four times higher than Mn [2], the chromosomes of D. radiodurans preferentially bind Mn(II) [2, 10]. Mn(II) binds predominantly at the N7 position of guanine, possibly leading to preferential binding of Mn(II) over Fe(II) within DNA duplexes [36]. Whereas high intracellular Mn(II) concentrations do not protect DNA during in vivo ionizing radiation [2, 13], Mn(II) ions can mimic as catalase- [35] and SOD-activities [3335, 37, 38], and are not associated with the production of HO by Fenton-type chemistry in vivo [33]. In contrast, “free Fe(II)” meaning Fe(II) that is not incorporated into enzymes or iron-storage proteins catalyzes Fenton-type chemistry in vivo [5, 6], and has been reported to be released from its bound state in cells by oxidative stress and radiation [39, 40]. Since the production of Embedded Image is intimately associated with respiration, the case can be made that Mn-rich, Fe-poor cells are less susceptible to metabolism-induced ROS during recovery than Mn-poor, Fe-rich cells. After exposure to ultraviolet radiation (UV), desiccation or ionizing radiation, non-enzymic Mn(II)-based protection might facilitate survival by preventing sudden increases in metabolism-induced, Fe(II)-dependent ROS at the onset of recovery [2, 41] and before enzymic defense systems are recalibrated [5]. Supporting this model, D. radiodurans cells are able to express very high levels of SOD and catalase [42, 43], but are sensitized to ionizing radiation by limiting Mn(II) [2]. In contrast, S. oneidensis, which does not accumulate Mn(II), encodes a diverse set of protection systems [2] but is extremely sensitive to oxidizing conditions [2, 44].

4 Role of metabolism in recovery of D. radiodurans

Our whole genome transcriptome analysis of D. radiodurans recovering from ionizing radiation (15 kGy) supports the idea that cells limit biosynthetic demands during recovery by importing exogenous peptides and other secondary metabolites [45]. Furthermore, recovering cells suppress the isocitrate (icd) to fumarate (sdhA) steps of the tricarboxylic acid (TCA) cycle while at the same time upregulating the glyoxylate bypass (Fig. 1) [45]. Importing biosynthetic precursors and differentially regulating TCA cycle activity following ionizing radiation could provide biosynthetic intermediates needed for recovery without generating high levels of ROS (oxidative stress). Although wild-type D. radiodurans cannot utilize most exogenously provided TCA cycle intermediates for growth [31, 46], the TCA cycle genes (including genes for its glyoxylate bypass) identified by annotation [3] (Fig. 1) are functionally expressed based on our whole proteome [47] and genome expression analyses [45]. We find that four genes (nadABCD) required for nicotinamide adenine dinucleotide (NAD) biosynthesis are absent (Fig. 1) [3], and consistent with this prediction, D. radiodurans is dependent on exogenous nicotinic acid for growth [31].

A striking informatic (Fig. 1) and experimentally determined feature of wild-type D. radiodurans is its inability to utilize inorganic sulfate (Fig. 1, CysJ is missing), which results in its dependence on an exogenous source of sulfur-containing amino acids for growth [31, 46]. We find that methionine alone in addition to an Embden–Meyerhof–Parnas (EMP) substrate (e.g., fructose) and nicotinic acid will support growth in defined medium (31), however, optimal growth of D. radiodurans in defined minimal medium additionally requires branched-chain and aromatic amino acids at >100 μg/ml (total amino acids) [46]. Other deinococci were similarly tested (Table 2) showing that the indicated strains also are dependent on methionine and NAD, and cannot utilize inorganic sulfate, suggesting that this metabolic configuration contributes to their resistance phenotypes.

View this table:

Utilization of carbon substrates by Deinococci grown in a defined minimal medium

D. radiodurans (ATCC BAA-816)bD. radiodurans/sodAcD.radiodurans (7b-1)dD. radiopugnanseD.murrayifD. proteolyticus
g, h Substrates
Fructose + 7 amino acids++++++++++++++++++
Fructose + Met + NAD+++++++++/−+/−
Fructose (−aa) + NAD
Fructose (−NAD) + Met
Pyruvate + Met + NAD+++++++/−
Acetate + Met + NAD++++++++/−+/−
α-Ketoglutarate + Met + NAD
Succinate + Met + NAD+/−+/−
Fumarate + Met + NAD+/−+/−
Malate + Met + NAD+/−+/−
Oxaloacetate + Met + NAD+/−+/−+++/−
Fructose + ammonium sulfate + NAD
  • aDefined minimal medium [2] was used to test the growth requirements of the indicated strains. The current table is derived from growth assays conducted as part of work reported previously [31].

  • bThe superoxide dismutase A mutant (KKW7004) does not encode sodB, and was reported previously [58].

  • cStrain 7b-1 was isolated in 2000 from an environmental US Department of Energy radioactive waste site [30].

  • d(DSM 12027) [60].

  • e(DSM 11303) [62].

  • f(DSM 20540) [61].

  • gGrowth on substrates compared to TGY, tryptone/glucose/yeast extract (rich) medium [2]: +++, good; ++, moderate; +/−, poor; −, absent.

  • hNotes and abbreviations: Fructose + 7 amino acids, fructose plus Met, His, Cys, Lys, Asp, Trp, and Pro, each @ 50 μg/ml; +Met (100 μg/ml), only methionine added; −aa, no amino acids added; +NAD, nicotinic acid added (1 μg/ml); −NAD, no nicotinic acid added; ammonium sulfate added to 15 mM.

We have also examined TCA cycle regulation in non-irradiated, nutrient-limited D. radiodurans grown in defined minimal medium [2, 46] versus undefined rich medium (TGY) (Table S2 in the Appendix). For both growth conditions, 15 TCA cycle gene expression levels were examined by real time-polymerase chain reaction (RT-PCR). Notably, in defined minimal medium, only one gene (aceA) was statistically significantly upregulated. Six genes displayed no significant changes (gltA, sucB, lpd, sucD, fumC, and aceB DR1155); and eight genes were significantly downregulated (acnA, icd, sucA, sucC, sdhB, sdhA, aceB (DRA0277), and mdh) (Fig. 1). Thus, the response of the TCA cycle in D. radiodurans to nutrient-depletion or radiation is similar, with upregulation of isocitrate lyase (aceA) observed in both conditions (Table S2) [45]. Taken collectively, our results indicate that aceA is a potential new target to control recovery from radiation injury.

Manganese(II) accumulation in D. radiodurans is energy-dependent [2] and the inhibitory effect of nutrient-depletion on recovery of irradiated D. radiodurans is documented [2, 46, 48, 49]. However, a recent report by Harris et al. [50] claims that γ-irradiated wild-type D. radiodurans (R1) is able to maintain viability during prolonged incubation in the absence of nutrients in 10 mM MgSO4. We were not able to reproduce those findings and it is not clear what the conditions could have been that gave rise to those results. As expected in the absence of nutrients, incubation of irradiated (5 kGy) logarithmic-phase D. radiodurans (ATCC 13939) in 10 mM MgSO4 caused a ∼1000-fold decrease in viability after 4 days (Fig. 4). Further, the exponential loss of viability of irradiated ATCC 13939 in MgSO4 is essentially the same as reported for irradiated MgSO4-treated mutant ddrA (DR0423), described as encoding part of a DNA end-protection system [50]. Thus, DdrA does not appear to facilitate cell survival when energy sources are lacking; the ionizing radiation dose yielding 10% CFU (colony forming unit) survival (D10) of ddrA in rich medium is 8–9 kGy [50] compared to ∼12 kGy for the wild-type.


Irradiated D. radiodurans requires nutrient conditions for recovery. Each data point is derived from three biological replicates, and standard deviations are shown. IR, ionizing radiation. See Appendix for experimental details.

5 Criteria for selecting organisms for comparison

Genome-based comparisons could help define how Mn homeostasis and the configuration of metabolic systems involved in energy production influence recovery, and highlight the possible nature of such processes during the evolution of environmentally robust microorganisms. There has been some confusion in defining what constitutes radiation resistance in bacteria. In assessing the resistance of bacterial cultures, failure to correct for cell-grouping (aggregation) has resulted in exaggerated reports of resistance. One approach to evaluating cells for resistance to ionizing radiation involves growth in rich medium, irradiation on ice and recovery on fresh medium, where survival frequencies are determined by dilution of irradiated cultures and colony-counts on nutrient agar plates [2]. When cells in liquid culture grow singly, death of one cell will correlate with the loss of one colony. However, where bacteria cluster in groups of cells (e.g., Fig. 1), each constituent viable cell of a cluster must be killed to eliminate a CFU. For a culture comprised of cells grouped as four or more, the shoulders of survival curves are significantly increased and the effect of cell-grouping on resistance values must be compensated [2]. Furthermore, as the cell-grouping number increases the cell-survival characteristics of an irradiated organism are expected to increase as a result of the effects imposed by oxygen limitation, where cells inside a cluster are shielded from oxygen-effects [51]. One example of an organism that might be substantially less resistant than reported is Kineococcus radiotolerans, which grows in liquid rich medium as very large clumps (>100 cells/CFU) [52]. The growth physiologies (cell-grouping) for the organisms listed in Table 1 have been characterized and the cell-survival frequencies have been adjusted appropriately [2]. Therefore, based on the most resistant and sensitive bacteria yet reported (Table 1), we present genome comparisons between D. radiodurans and S. oneidensis viewed from the perspective of systems that produce and defend against oxidative stress (Table 3).

View this table:

Predicted systems that produce and protect against oxidative stress

a COG #COG/protein annotationD. radiodurans, 3.28 Mb (3182 ORFs)S. oneidensis, 5.1 Mb (4472 ORFs)Comments
Protein groups associated with the production of ROS
Cytochromes (all types) (see also Table 1)1053Binary profiles for all types of cytochromes, flavoproteins, Fe-S domains and NADH/NADPH domains were taken from the b CDD database (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?CMD = search&DB = cdd). D. radiodurans and S. oneidensis were searched for these domains using profiles with default parameters as described previously [53]
Iron–sulfur proteins3253
NADPH- and NADH-dependent enzymes102159
Transport of protein precursors which might reduce biosynthetic demands and suppress the production of ROS
Peptide transporters184The approximate total number of transporters (for ABC transporters, each subunit was calculated separately)
Aminoacid transporters4841The approximate total number of transporters (for ABC transporters, each subunit was calculated separately)
Subtilisin-like serine proteases106See text and Fig. 5[3]
Proteins associated with the production of ROS
COG1053Succinate dehydrogenase/fumarate reductase, flavoprotein subunit: fumarate reductase, FrdA0SO0398FAD-containing enzymes are responsible for the catalytic interconversion of fumarate and succinate; fumarate reductase is used in anaerobic growth; in operon with SO0396 (frdC), SO0397 (frdC) and SO0399 (frdB)
COG1053Succinate dehydrogenase, SdhADR0952SO1928Succinate dehydrogenase is used in aerobic growth. In operon with DR0951, DR0953 & DR0954; and SO1927 (sdhC) & SO1929 (sdhB)
COG1053Fumarate reductase, IfcA0SO1421Flavocytochrome that is induced during anaerobic respiration of Fe(III) by Shewanella frigidimarina; individual (not in cluster)
COG1053Fumarate reductase flavoprotein subunit precursor0SO0970Individual (not in cluster)
COG1053Flavocytochrome c flavin subunit0SO3301In cluster with SO3300 cytC subunit
SO1414In operon with SO1413 4
SO3624In operon with SO3623 4
SO3058In operon with SO3056 4
COG0029Aspartate oxidase, NadB0SO1341Catalyzes the oxidation of l-aspartate to iminoaspartate; monomer; individual (not in cluster)
COG0369Sulfite reductase, alpha subunit (flavoprotein), CysJ0SO3738In operon with SO3736 (cysI), hemoprotein beta-component
Proteins that defend against ROS, participate in repair of damaged molecules (non-DNA) and redox regulation
COG0605Superoxide dismutase, SodA/B1 (DR1279)1 (SO2881)Removes Graphic
COG2032Cu/Zn superoxide dismutase, SodC2 (DR1546, DRA0202_1)0Removes Graphic
COG0753Catalase,KatE/A/B3 (DRA025, DR1998, and related DRA0146)1 (SO1070)Removes H2O2
COG0376Catalase (peroxidase I), KatG02 (SO0725, SO4405)Removes H2O2; activated by SoxR
COG2837Predicted iron-dependent peroxidase1 (DRA0145)1 (SO0740)Removes hydroperoxides
COG0386Glutathione peroxidase02 (SO1563, SO3349)Removes hydroperoxides (may contain redox active selenocysteine residue in the active site); btuE.
COG1858Cytochrome c peroxidase,MauG1 (DRA0301)1 (SO2178)Catalyzes the peroxidative oxidation of azurin and cytochrome c551. Likely provides protection against toxic peroxides.
COG0450Peroxiredoxin, AhpC1 (DR2242)1 (SO0958)Peroxiredoxin family of proteins; might represent a major line of defense against hydroperoxide damage; and removes organic hydroperoxides at the expense of thiols
COG0678Peroxiredoxin01 (SO4640)
COG1225Peroxiredoxin2 (DR0846, DR1209)2 (SO1381, SO1877)
COG2077Peroxiredoxin01 (SO3341)
COG0225Peptide methionine sulfoxide reductase, MsrA1 (DR1849)2 (SO2337, SO2588_2)Repair of oxidized proteins (reduction of protein-bound methionine sulfoxide back to methionine via a thioredoxin-recycling process)
COG0229Conserved domain frequently associated with peptide methionine sulfoxide reductase, MsrB1 (DR1378)1 (SO2588_1)Repair of oxidized proteins (reduction of protein-bound methionine sulfoxide back to methionine via a thioredoxin-recycling process)
COG0583Transcriptional regulator, LisR family (OxyR)2 (DR0615, possible distant relationship)52 (SO1328)DR0615 has similarity to OxyR; oxidized OxyR can activate genes responding to oxidative stress
COG0789Predicted transcriptional regulators (SoxR)7 (DR2519)5 (SoxR-SO1687, more distant than DR)Phylogenetic tree shows that DR2519 is probably an ortholog of SoxR; SoxR activates soxS transcription
COG1304Isopentenyl-diphosphate delta-isomerase2 (DR1087)0Carotenoid biosynthesis
COG1562Phytoene/squalene synthetase1 (DR0862)0Carotenoid biosynthesis
COG1233Phytoene dehydrogenase and related proteins7 (DR0861)0Carotenoid biosynthesis
No COGLycopene cyclase, superfamily: FAD/NAD(P)-binding protein1 (DR0801)0Carotenoid biosynthesis
COG0142Octaprenyl-diphosphate synthase, IspB1 (DR0932)1 (SO3653) 1Carotenoid biosynthesis
Geranylgeranyl pyrophosphate synthase, IspA1 (DR1395)(SO1526)
COG0584Glycerophosphoryl diester phosphodiesterase31Esterase/lipase, class hydrolase, helps maintain integrity of cellular membranes under oxidative stress conditions
COG2267Lysophospholipase24Esterase/lipase, class hydrolase, helps maintain integrity of cellular membranes under oxidative stress conditions
COG2272Carboxylesterase type B20Esterase/lipase, class hydrolase, helps maintain integrity of cellular membranes under oxidative stress conditions
COG2755Lysophospholipase L1 and related esterases01Esterase/lipase, class hydrolase, helps maintain integrity of cellular membranes under oxidative stress conditions
COG0657Esterase/lipase52Esterase/lipase, class hydrolase; acts on ester bonds; carboxylic ester hydrolases; phospholipid degradation; glycerolipid metabolism; helps maintain integrity of cellular membranes under oxidative stress conditions
COG1166Arginine decarboxylase1 (DR0243)1 (SO1870)Polyamine biosynthesis (agmatine); polyamines inhibit the toxic effects of oxygen
COG0010Arginase/agmatinase/formimionoglutamate hydrolase, arginase family1 (DRA0149)0Polyamine biosynthesis (putrescine); polyamines inhibit the toxic effects of oxygen; putrescine is a modulator of oxyR transcriptional activity
COG1982Arginine/lysine/ornithine decarboxylases01 (SO0314)Polyamine biosynthesis (putrescine); polyamines inhibit the toxic effects of oxygen; putrescine is a modulator of oxyR transcriptional activity
COG4262Predicted spermidine synthase with an N-terminal membrane domain01 (SO3763)Polyamine biosyntheses (spermidine); polyamines inhibit the toxic effects of oxygen
COG0241Spermidine synthase1 (DR1052)1(SO0960)Polyamine biosynthesis (spermidine); polyamines inhibit the toxic effects of oxygen
Mn- and Fe-dependent systems
COG1914NRAMP family transporter1 (DR1709)0Mn-homeostasis: transport [2]
COG1121ABC-type Mn/Zn transport systems, ATPase component1 (DR2284)0Mn-homeostasis: transport [2].
COG1108ABC-type Mn/Zn transport systems, permease component1 (DR2283)1 (SO0566)
COG0803ABC-type Mn/Zn transport systems, perplasmic component1(DR2523)1 (SO0565)
COG1321Mn-dependent transcriptional regulator [2]1 (DR2539)0Mn-homeostasis: regulation
COG3486Lysine/ornithine N-monooxygenase0SO3030Fe-homeostasis: siderophore biosynthesis; alcaligin biosynthesis
No COGSiderophore biosynthesis protein0SO3031Fe-homeostasis: siderophore biosynthesis; alcaligin biosynthesis
COG4264Siderophore synthetase component0SO3032Fe-homeostasis: siderophore biosynthesis; alcaligin biosynthesis
COG1528Ferritin-like protein, Ftn01 (SO0139)Fe-homeostasis: iron storage
COG1629Outer membrane receptor proteins, mostly Fe transport012Fe-homeostasis: transport
COG4771Outer membrane receptor for ferrienterochelin and colicins09Fe-homeostasis: transport
COG0810Periplasmic protein TonB, links inner and outer membranes04Fe-homeostasis: transport; TonB-ExbBD complex provides energy for transfer of ferrisiderophores to the periplasm via the outer membrane
COG4772Outer membrane receptor for Fe(III)-dicitrate01 (SO1102)Fe-homeostasis: transport
COG4773Outer membrane receptor for ferric coprogen and ferric-rhodotorulic acid01 (SO3033)Fe-homeostasis: transport
COG0783DNA-binding ferritin-like protein (oxidative damage protectant),Dps2 (DR2263, DRB0092)1 (SO1158)Fe-homeostasis: iron storage. Protects DNA from oxidative damage and induced by OxyR
COG1120ABC-type cobalamin/Fe(III)-siderophores transport systems, ATPase components2 (DR2590, DRB0121)1 (SO1033)Fe-homeostasis: transport
COG4559ABC-type hemin transport system, ATPase component1 (DRB0016)1 (SO3675)Fe-homeostasis: transport
COG0614ABC-type Fe(III)-hydroxamate transport system, periplasmic component51 (SO3709)Fe-homeostasis: transport
COG0609ABC-type Fe(III)-siderophore transport system, permease component42 (SO1034, SO3674)Fe-homeostasis: transport
COG2375Siderophore-interacting protein2 (DRB0017, DRB0124)0Fe-homeostasis: transport
COG1918Fe(II) transport system protein A1 (DR1220)1 (SO1783)Fe-homeostasis: Fe(II) transport system (active under anaerobic conditions)
COG0370Fe(II) transport system protein B1 (DR1219)1 (SO1784)Fe-homeostasis: Fe(II) transport system (active under anaerobic conditions)
COG0735Fe(II)/Zn(II) uptake regulation proteins1 (DR0865)1 (SO1937)Fe-homeostasis: regulation; DR0865 is similar to zur of B. subtilus
  • aCOG, clusters of orthologous groups database resources of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov). The number of proteins identified for each COG are indicated.

  • bConserved Domain Database (CDD) database resources of the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov).

6 Genome comparisons

At least four bacteria reported to be radiation resistant (D. radiodurans, D. geothermalis, Lactobacillus plantarum and Rubrobacter xylanophilus) have been subjected to genome sequencing [3] (http://www.jgi.doe.gov/), and comparative analyses support that their DNA repair systems are encoded by a variety of genes which have functional homologs in other prokaryotic species [2]. However, no shared group of uncharacterized genes has been identified in those sequenced organisms that might comprise an expanded gene core involved in recovery. Further, DNA repair systems identified in D. radiodurans appear less complex and diverse than those reported for S. oneidensis [2]. Viewed in this context, it is possible that D. radiodurans and other radioresistant bacteria use relatively conventional DNA repair and protection systems, but with greater efficiency than other organisms.

Recently, four predicted genes of D. radiodurans (DR0003, DR0070, DR0326 and DR0423) were identified by transcriptional profiling following ionizing radiation or desiccation as the most highly induced [14, 45, 50]. These genes have been disrupted, and the mutants characterized for resistance to ionizing radiation. Remarkably, the radiation resistance of the corresponding mutants (ddrC, ddrB, ddrD and ddrA, respectively) remained very high [14, 45, 50]. In contrast, we have shown that lowering the intracellular Mn concentration of D. radiodurans by a factor of just 3.3 substantially reduced the D10 cell-survival value [2]. Since high intracellular Mn(II) levels serve as antioxidants in several phylogenetically distinct organisms [2729], we have proposed that Mn(II)-accumulation in D. radiodurans facilitates cellular repair processes after irradiation or desiccation by providing protection from metabolism-induced ROS [2]. We extend our hypothesis here to include consideration of how the metabolic configuration of D. radiodurans contributes to the radiation resistance phenotype.

6.1 Systems that produce oxidative stress

The prevailing hypothesis that most metabolically induced ROS are generated from within the respiratory chain has been expanded to include sources outside the respiratory chain [54]. It is known that terminal oxidases of respiring cells yield water, and not Embedded Image or H2O2 [55]; instead, respiratory flavins have been implicated [54, 56]. One source of ROS arises from the autoxidation of respiratory dehydrogenases (Fig. 1), where adventitious transfer of electrons from reduced flavins (FADH2) associated with c-type cytochromes to oxygen yields Embedded Image and/or H2O2+ FAD+ [5, 6]. Other sources might include flavoproteins and cytochromes not associated with the respiratory chain [54]. Therefore, the total intracellular titer of cytochromes and flavins might serve as a marker for the proclivity of cells to generate oxidative stress [56], as well as the ability of cells to survive radiation and other oxidizing conditions. Viewed in this context, the high Fe and low Mn contents of S. oneidensis [2] in combination with its predicted 53 cytochromes and 59 flavoproteins (Table 3) appear to make this organism ripe for Fenton-type chemistry under growth conditions where these proteins are highly expressed. In contrast, the relatively Fe-independent, Mn-rich D. radiodurans [2] encodes only 10 predicted cytochromes and 39 flavoproteins (Table 3); D. radiodurans can grow in the presence of Fe-chelators and in media with less than 0.4 μM Fe [2]. The rates at which different flavins react with oxygen vary over several orders of magnitude, depending on the extent of flavin exposure. Two enzymes with highly exposed flavins and some of the highest in vitro reactivities are fumarate reductase (FrdA) and aspartate oxidase (NadB is an Frd homologue that lacks an Fe–S subunit) [56]. Notably, S. oneidensis encodes several fumarate reductases and NadB, but D. radiodurans does not (Fig. 1) (Table 3).

Irrespective of source, H2O2 is relatively inert and diffusible throughout the cell. However, when H2O2 encounters “free Fe(II)”, Fenton-type chemistry ensues [5, 6]. Therefore, conditions which liberate bound Fe(II) are extremely dangerous to metabolically active Fe-containing cells, not only because of the generation of HO, but because the loss of Fe from iron-dependent enzymes will lead to the failure of the biochemical pathways within which they operate. Typically during irradiation, ∼80% of DNA damage is caused indirectly by irradiation-induced ROS, the remaining ∼20% by direct interaction between γ-photons and DNA [6]. Hydroxyl radicals are the primary product of the radiolysis of water, and in the presence of oxygen can also generate some Embedded Image, and H2O2 by dismutation of Embedded Image[6]. In contrast, the primary ROS generated by metabolism are Embedded Image and H2O2 [5]. Hydroxyl radicals (standard reduction potential 2.31 V) are substantially more oxidizing than Embedded Image (−0.33 V) and H2O2 (−2.84 V) [6]. Since bound Fe(II), particularly from iron–sulfur (2Fe–2S, 4Fe–4S) clusters [5, 56], can be released within cells when they have been oxidatively damaged [57], HO generated by radiolysis might increase the amount of “free Fe(II)” in cells, with the result that irradiated cells transferred to recovery medium become susceptible to sudden increases in Fe(II)-dependent oxidative stress once metabolism resumes. In contrast, cytochromes coordinate Fe extremely tightly, and likely remain functional under oxidizing conditions [5]. Notably, S. oneidensis encodes 65% more Fe–S proteins than D. radiodurans (Table 3).

6.2 Systems that defend against oxidative stress

Wild-type E. coli cultured under hyperbaric oxygen displays severe growth defects, which resemble those observed in SOD mutants grown in minimal medium, and are caused by Embedded Image and related ROS [5]. E. coli cells poisoned by Embedded Image lose their ability to assimilate sulfate, become auxotrophic for sulfurous and other amino acids, require NAD, and non-fermentable substrates (e.g., acetate, α-ketoglutarate, succinate, fumarate and malate) can no longer support growth [5]. Metabolic regulation underlying these phenotypes in E. coli limits TCA cycle-dependent precursor-biosynthesis, energy metabolism and the ROS that ensue [5]. Remarkably, these responses in E. coli lacking SOD mirror the growth deficiencies of D. radiodurans, which appears to have evolved a metabolic configuration that constitutively suppresses the generation of ROS (Fig. 1); wild-type D. radiodurans is highly dependent on methionine and other amino acids for growth, the NAD biosynthetic pathway is disrupted, sulfate assimilation is blocked, and cells are unable to utilize α-ketoglutarate, succinate, fumarate or malate for growth (Fig. 1) (Table 2). Add to this the intracellular accumulation of Mn(II) as an antioxidant [2], and it is perhaps not surprising that the role of SOD has been marginalized in D. radiodurans; D. radiodurans lacking SodA/B (KKW7004 [3, 58]) has similar growth requirements (Table 2) and resistance to ionizing radiation as the wild-type [2, 58]. In this context, we note that D. radiodurans encodes 3 catalases with some of the highest activities reported for any bacteria [42], and this would serve to remove H2O2 generated by non-enzymic Mn(II)-based dismutation of Embedded Image. S. oneidensis also encodes 3 catalases but is sensitive to conditions known to produce H2O2 [2, 44]. It is unlikely that the metabolic configuration of D. radiodurans (ATCC BAA-816) (Fig. 1, Table 2) is the product of mutations accumulated during maintenance under laboratory conditions since its isolation in 1956 [59]. Deinococcus radiopugnans (isolated 1963) [60], Deinococcus proteolyticus (1973) [61], Deinococcus murrayi (1997) [62], and another strain of D. radiodurans (7b-1) isolated from sediment samples collected in 2000 [32] have very similar nutrient requirements (Table 2), with the conclusion that this metabolic configuration is prevalent among Deinococcus species.

Table 3 contrasts the systems in D. radiodurans and S. oneidensis predicted to contribute to the defense against oxidative stress, with the most prominent differences summarized as follows: D. radiodurans encodes the Mn(II) transporter Nramp and a putative Mn-dependent transcriptional regulator (DR2539) [2], but S. oneidensis encodes neither and does not accumulate Mn(II) [2]. The absence of accumulated Mn(II) in the Fe-rich S. oneidensis might make this organism highly susceptible to oxidative stress, and it is noteworthy that S. oneidensis encodes 52 proteins predicted to be related to the LysR family, which includes OxyR, compared to just 2 in D. radiodurans (Table 3); OxyR is a global transcriptional regulator involved in activating genes that fight oxidative stress [6]. Carotenoids have also been implicated in protection from oxidative stress [63], and the red-pigmented D. radiodurans encodes 13 predicted genes involved in carotenoid biosynthesis compared to 2 in S. oneidensis. D. radiodurans also encodes 18 peptide transporters and has numerous expanded protein families for branched-chain amino acid transport, many of which are upregulated during recovery from radiation [45]. In comparison, S. oneidensis encodes 4 peptide transporters and displays no auxotrophies for amino acids, sulfur or NAD [1, 2]. Although there is currently no evidence to support that proteases might facilitate recovery from ionizing radiation, we note that at least 10 complete open reading frames with considerable sequence similarity to secreted subtilisin-like proteases of B. subtilis have been identified in D. radiodurans [3], which might relieve its amino acid auxotrophies. D. radiodurans growth was tested on protease-indicator plates containing skimmed milk, and showed luxuriant growth and generated large halo-shaped areas of clearing around colonies engaged in protein hydrolysis (Fig. 5A and D); S. oneidensis is predicted to encode 6 subtilisin-like proteases, and showed less protease secretion (Fig. 5C). Although, S. oneidensis encodes a similar assortment of catalases and peroxidases as D. radiodurans [2], S. oneidensis is exceptionally sensitive to conditions that produce H2O2, which include UV and ionizing radiation, and desiccation [2, 44].


Protease secretion assay on an indicator plate containing skimmed milk. (A) D. radiodurans (7b-1 [30]). (B) E. coli (strain K-12) (MG1655). (C) S. oneidensis (MR-1) (ATCC 700550). (D) D. radiodurans (ATCC BAA-816). See Appendix for experimental details.

7 Conclusion

In the laboratory, the Fenton reaction is readily demonstrated by adding ferrous chloride to H2O2, which causes intense oxygen-effervescence; in comparison, manganous chloride does not. Given the structural similarities between Fe(II) and Mn(II) atoms, no great difference in reactivity with H2O2 is predicted, yet Mn(II) does not actively participate in Fenton-type chemistry [33]. We have reported a strong correlation between the intracellular Mn/Fe concentration ratio and the resistance to ionizing radiation and desiccation for a variety of phylogenetically distinct bacteria, and have proposed that Mn(II) ions accumulated in bacteria serve as antioxidants that reinforce enzymic systems which defend against oxidative stress during recovery [2]. For Fe-rich, Mn-poor cells such as S. oneidensis, death at low doses of ionizing radiation might not be caused by DNA damage inflicted during irradiation but instead by the release of Fe(II) and consequent toxic byproducts of energy-metabolism after irradiation. In addition, oxidative damage to DNA in S. oneidensis during recovery might induce lytic prophages, as proposed following recovery from UV radiation (200–290 nm) [44]. Conversely, D. radiodurans irradiated to high doses might be able to prevent bursts of oxidative stress at the onset of recovery with accumulated Mn(II), allowing repair systems and enzymic ROS-defenses to be expressed with little interference from metabolism-induced ROS. Other recent reports also attribute the release of Fe(II) and Fenton-type chemistry to radiation toxicity. Reelfs et al., (2004) [40] have reported the immediate release of labile “free Fe(II)” in skin cells as a result of oxidative damage during UV irradiation; Mn-based SOD-mimetic compounds [33, 35, 37, 38] and Fe-chelators [40, 64] have been shown to be effective in promoting radiorecovery in animals [6567]; and it has been proposed that the radioresistance of the human osteosarcoma cell line HS-Os-1 is related to the strong scavenging of the cells for ROS following irradiation [68].

This view that the release of bound Fe(II) in irradiated cells is a primary cause of death could be tested further in sensitive versus resistant organisms by examining (i) the relationship between radiation dose and the size of the intracellular “free Fe(II)” pool, (ii) intracellular Mn/Fe localization, (iii) how genes that participate in Mn/Fe homeostasis, metal reduction, metabolism (e.g., aceA) and phage-induction affect resistance, (iv) the production of ROS and resistance of S. oneidensis under aerobic versus strictly anaerobic conditions, (v) the extent to which organisms can be made more resistant by increasing their intracellular Mn/Fe concentration ratios, and (vi) the effectiveness of Fe chelators, Mn-based catalase- and SOD-mimetic compounds, and other antioxidants (e.g., vitamins E and C) to control recovery from radiation injury.

Appendix A Supplementary data

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.fmrre.2004.12.007.


This research was supported by the Office of Science (Biological and Environmental Research), US Department of Energy (DOE) Grant Nos. DE-FG02-01ER63220 and DE-FG02-04ER63918. Pacific Northwest National Laboratory is operated for the DOE by Battelle Memorial Institute under Contract DE-AC06-76RLO 1830.


  • Edited by Michael Y. Galperin.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
View Abstract