Survival in Nuclear Waste, Extreme Resistance, and Potential Applications Gleaned from the Genome Sequence of Kineococcus radiotolerans SRS30216

Kineococcus radiotolerans SRS30216 was isolated from a high-level radioactive environment at the Savannah River Site (SRS) and exhibits γ-radiation resistance approaching that of Deinococcus radiodurans. The genome was sequenced by the U.S. Department of Energy's Joint Genome Institute which suggested the existence of three replicons, a 4.76 Mb linear chromosome, a 0.18 Mb linear plasmid, and a 12.92 Kb circular plasmid. Southern hybridization confirmed that the chromosome is linear. The K. radiotolerans genome sequence was examined to learn about the physiology of the organism with regard to ionizing radiation resistance, the potential for bioremediation of nuclear waste, and the dimorphic life cycle. K. radiotolerans may have a unique genetic toolbox for radiation protection as it lacks many of the genes known to confer radiation resistance in D. radiodurans. Additionally, genes involved in the detoxification of reactive oxygen species and the excision repair pathway are overrepresented. K. radiotolerans appears to lack degradation pathways for pervasive soil and groundwater pollutants. However, it can respire on two organic acids found in SRS high-level nuclear waste, formate and oxalate, which promote the survival of cells during prolonged periods of starvation. The dimorphic life cycle involves the production of motile zoospores. The flagellar biosynthesis genes are located on a motility island, though its regulation could not be fully discerned. These results highlight the remarkable ability of K radiotolerans to withstand environmental extremes and suggest that in situ bioremediation of organic complexants from high level radioactive waste may be feasible.


Introduction
High-level radioactive waste (HLW) is an anthropogenic disturbance to which few organisms are resistant. During the Cold War, Pu 239 production for national defense began by irradiating uranium or other target elements in a nuclear reactor. At Hanford, WA and the Savannah River Site (SRS), SC, irradiated fuel and targets were reprocessed to reclaim approximately 99% of the U 235 and Pu 239 isotopes. All remaining radionuclides, fission products, fuel components, and nonradioactive chemicals used during reclamation make up the HLW, which currently resides at over 100 different sites across the contiguous U.S. and exceeds 1 billion curies. The majority of this lasting Cold War legacy is located at Hanford (roughly 65 million gallons) and SRS (roughly 35 million gallons).
The SRS waste contains Fe, Al, Si, Ca, F, K, alkali cations, organic solvents, radionuclides, and other fission products. Organic constituents include complexants used during separations, radiolysis products from degradation of complexants and solvents, and waste tank decontamination reagents. One of the preferred decontamination reagents is oxalic acid, which can create problems for storage and final disposal due to its solubility properties as a sodium salt. Additional reagents in use at SRS and Hanford include glycolic acid, citric acid, and formic acid. Removal of organic constituents directly in HLW tanks could greatly improve processing efficiency of HLW.
In 2001, the Committee on Long-Term Research Needs for Radioactive High-Level Waste at Department of Energy Sites recommended the investigation of radioactive waste to identify promising new radiation resistant microorganisms that might be used to degrade some of the organic constituents in the HLW. Radiation dosage is measured in gray (Gy); 1 Gy causes the first signs of radiation sickness in humans. Half of all people exposed to 4.5 Gy die, and doses of 8 Gy or more are invariably fatal to humans. Bacteria are the most extreme examples of radiation resistant organisms. The paradigm is Deinococcus radiodurans, which was isolated from canned meat that received 4,000 Gy of ionizingradiation [1], but many other radiation resistant bacteria have also been identified.
Kineococcus radiotolerans SRS30216 was isolated from HLW within a shielded cell work area at the SRS [2]. K. radiotolerans is an orange-pigmented, aerobic bacterium belonging to the Actinobacteria phylum that is capable of withstanding relatively high concentrations of metals and alkali cations, as well as exposure to extreme doses of ionizing radiation. K. radiotolerans exhibits radiation resistance approaching that of D. radiodurans [2]. D. radiodurans and K. radiotolerans belong to different phyla and it remains unknown whether both organisms attain radiation resistance through a common set of gene products. Because K. radiotolerans survived a HLW environment, it is expected to possess potent cellular defense and repair mechanisms for radiation exposure, osmotic stress and chemical toxicity. The occurrence of all these features in a naturally occurring bacterium may have direct applications for the bioremediation of nuclear waste.
In this work the genome sequence of K. radiotolerans SRS30216 was examined in three different contexts. First, the presence of genes known to confer radiation resistance in D. radiodurans was examined in the K. radiotolerans genome. Second, the capacity for bioremediation was assessed by comparative genomics as well as growth and respiration studies. Finally, the dimorphic life cycle of the organism, in particular the production of motile zoospores, was examined by identifying genes involved in flagellar motility and chemotaxis.

Results and Discussion
Phylogeny K. radiotolerans belongs to the suborder Frankineae in the order Actinomycetales and the phylum Actinobacteria. Complete genome sequences are available for only two other members of the Frankineae suborder, Frankia alni and Acidothermus cellulolyticus. F. alni is distinguished by its ability to fix nitrogen in symbiosis with alder (Alnus spp.) and myrtle (Myrica spp.), two pioneer plants in temperate regions. The K. radiotolerans genome shows no potential for nitrogen fixation. A. cellulolyticus is a thermotolerant organism isolated from the hot spring in Yellowstone National Park that degrades cellulose. While the K. radiotolerans genome may encode proteins with the potential for degrading complex carbohydrates like cellulose (Krad4622, 3823), cellobiose (Krad0408, 2526, 2530, 2531, 2539, 3436, 3480, 3961), glycogen, and starch (Krad1294, 1298), growth on these substrates has not been demonstrated experimentally.
More distant relatives in the same phylum include Streptomyces and Mycobacterium, whose physiology has been extensively examined, and which serve as reference points for comparative genomics.

Genome size and organization
The K. radiotolerans SRS30216 genome was sequenced by the US DOE Joint Genome Institute with the discovery of three replicons. The bulk of the DNA is contained on a 4,761,183 bp linear chromosome (CP000750). In addition there is a 182,572 bp linear plasmid (pKRAD01; CP000751) and a 12,917 bp circular plasmid (pKRAD02; CP000752). The three contigs were derived from at least 20 reads and average 74.2 mol % G+C. The genome is predicted to contain 4,715 genes.
Linear chromosomes are rare among prokaryotes. Aside from one atypical Agrobacterium isolate, the two principle examples are Streptomyces species [3] and Borrelia burgdorferi [4]. In both Streptomyces and Borrelia bidirectional replication occurs from a single origin of replication (oriC) located near the middle of the replicon. A plot of the GC skew = (C2G)/(C+G) along the chromosome sometimes inverts at the replication origin [5]. The shift in GC skew is thought to be due to accumulated mutational biases during leading vs lagging strand synthesis. There is no obvious GC skew at the center of the chromosome (Figure 1, compare green vs magenta peaks in inner circle). There is however a remarkable GC skew at the telomers but it seems unlikely that replication proceeds from the telomers toward the center based on the Borrelia and Streptomyces systems. A second method used to locate oriC is the presence of the dnaA gene and DnaA binding sites. dnaA (Krad0001) is near the center of the chromosome, as with Streptomyces and Borrelia. Putative, imperfect blocks of DnaA binding sites, observed using DoriC [6], are widely scattered in the center of the chromosome with the closest being about 94 Kb from the dnaA gene.
The topology of the chromosome was investigated by Southern hybridization using an end-specific probe homologous with Krad2223. Three restriction enzymes were predicted to generate single fragments with this probe. If the chromosome is linear then products with sizes predicted from the DNA sequence should be visible after hybridization. If the chromosome is circular then the hybridization products with the probe should be the sum of the sizes of the predicted restriction fragments from each end plus any missing sequence. Restriction fragments homologous with the Krad2223 probe were similar in size to those predicted by the DNA sequence (Fig. 2). These hybridization results confirm the linear topology predicted by the inability of JGI to close the DNA sequence.
The ends of linear DNA replicons have special features that preserve their integrity. B. burgdorferi linear replicons contain covalently closed hairpin ends [4]. ResT, telomere resolvase, hydrolyzes a phosphodiester bond on each DNA strand then joins the opposite strands to form a covalently closed telomere. The process is reversed during chromosome replication and ResT is essential for B. burgdorferi growth. A ResT homolog is not encoded by the K. radiotolerans genome. Streptomyces coelicolor replicon ends are composed of single-stranded sequences that can anneal to form a noncovalent circular molecule [3]. Streptomyces telomeres bind a family of conserved terminal binding proteins that have no orthologs in the K. radiotolerans genome. Thus, unique mechanisms must protect the K. radiotolerans telomers.

Ionizing radiation resistance
Gamma radiation is one of the most energetic forms of electromagnetic radiation. Gamma rays penetrate tissues and cells, causing direct damage to DNA (namely double strand breaks), proteins, and membranes. Gamma radiation also induces indirect cellular damage through the ionization of water with formation of free radical species, primarily NOH. Oxygen free radicals are extremely reactive, compounding cellular and DNA damage. DNA damage blocks transcription and replication, and if not correctly repaired, could introduce detrimental mutations or cause cell death. Relatively few DNA double strand breaks (DSB) are lethal for most bacteria. Escherichia coli cells succumb to around 10 DSB and Shewanella oneidensis cells die after 1 DSB (based on calculations of 0.0114 DSBs/Gy/Genome; Daly et al., 2004).
Radioresistance has been partially characterized for K. radiotolerans ( Figure 3). Acutely irradiated, exponentially grown cultures have a broad shoulder of death, which contrasts with the exponential death of E. coli. This shoulder is due in part to efficient repair systems and in part to the multicellular nature of the organism. In rich medium K. radiotolerans grows in cubical packets that form by alternating cell division planes (Phillips et al. 2002). Thus, the colony forming unit method used to estimate survivorship likely overestimates culture viability. Nevertheless, these results may portray a more ecologically relevant context as cell clustering is common for certain species or developmental stages of Actinobacteria (eg., Frankia, Geodermatophilus, Actinoplanes, Micrococcus) or other extremophiles (Kocuria, Deinococcus, Chroococcidiopsis). Remarkably though, Kineococcus can withstand the damaging effects of 20 kGy of c-radiation (theoretically generating more than 200 DSB/genome; Daly et al., 2004) and cell division resumes within 4 days ( Figure 4).
Several other Actinobacteria species exhibit extraordinary radiation resistance including Rubrobacter radiotolerans, Rubrobacter xylanophilus, and Kocuria rosea [7]. While more ionizing radiation resistant bacterial species are found in the Actinobacteria than any other phylum, radiation resistance is not a widespread trait in this phylum and may have multiple evolutionary origins. The mechanisms that render this trait remain poorly understood. Radiation resistant bacteria suffer severe damage from c-radiation [8], which implies that molecular repair processes function with high efficiency.
Analysis of genes common to four ionizing radiation resistant bacteria with fully sequenced genomes indicated that DNA repair From the outer to the inner concentric circle: circles 1 and 2, predicted protein coding sequences (CDS) indicated by blue arrows on the forward (outer wheel) and reverse (inner wheel) strands. Red indicates tRNA genes and purple indicates rRNA genes; circle 3, GC content showing deviation from average (74.2%); circle 4, GC skew (+ is green and 2 is purple); circle 5, genomic position in kb beginning with Krad2223 and proceeding clockwise. The chromosome map was generated using CGview [56]. doi:10.1371/journal.pone.0003878.g001 must have played a major role in evolutionary adaptation to ionizing radiation [9]. D. radiodurans has served as the paradigm for radiation resistant organisms and both genetic and biochemical approaches are converging to reveal a complex network of repair and protection processes [7,10]. Many D. radiodurans genes required for ionizing radiation resistance have been identified. In table 1 they are ordered according to the c-radiation sensitivity of a D. radiodurans strain lacking that gene based on published kill curves. While the roles of only a few of these genes are known with certainty, K. radiotolerans apparently employs a different genetic toolbox from that of D. radiodurans. Absent from K. radiotolerans are homologs of the D. radiodurans pprA, ddrA, ddrB, ddrC, ddrD genes [11].
Mutations in recA render D. radiodurans as sensitive to ionizing radiation as E. coli illustrating the importance of recombinational repair in repairing DSB (table 1). D. radiodurans cells exposed to 10 kGy c-radiation accumulate about 100 DSB per genome that are repaired over the course of several hours. D. radiodurans DSB are repaired by homologous recombination in the extended synthesisdependent strand annealing (ESDSA) process [12]. ESDSA repair is carried out by RecA and PolA, enzymes found in the K. radiotolerans genome (table 2). Based on their genomic analysis, recA and polA exhibit signs of coevolution in ionizing radiation-resistant bacteria [9].
Genes involved in carotenoid biogenesis have been shown to confer a modest level of radiation resistance [13][14][15] by scavenging electrons from reactive oxygen species (table 1). K. radiotolerans produces carotenoids [2] and the carotenoid biosynthetic pathway is similar to that found in other organisms. In addition to the two genes listed in table I encoding phytoene synthetase and phytoene desaturase, K. radiotolerans produces polyprenyl synthase (Krad3227), lycopene cyclase (crtY, Krad0091), and neurosporene dehydrogenase (crtD, Krad3225). A hydroxlase gene (crtZ) was not discovered suggesting that the K. radiotolerans carotenoids are not hydroxylated.
Higher eukaryotes repair DSB using nonhomologous end joining (NHEJ) ( [16,17] Ionizing radiation also causes many other types of DNA damage. The K. radiotolerans DNA replication, recombination and repair gene set is overlapping with [9], but generally different from those of D. radiodurans [18] and E. coli [19] which may be expected since these organisms belong to different phyla (table 2). As one illustration of this difference, K. radiotolerans but not D. radiodurans contains RecB and RecC which are involved in recombinational repair in E. coli (reviewed in [20]). DNA replication, repair and recombination systems in K. radiotolerans are more similar to those of Mycobacterium tuberculosis [21], possibly a reflection of the phylogenetic proximity of the two organisms. In addition, M. tuberculosis is at least 10-fod more resistant to ionizing radiation than E. coli [22]. Because M. tuberculosis is an actively studied pathogen its genome is wellannotated. Because it has coevolved the majority of its DNA replication, repair, and recombination mechanisms with K. radiotolerans we chose it as a reference organism, together with E. coli and .
Both K. radiotolerans and M. tuberculosis lack the classical bacterial mismatch repair genes MutS, MutH, MutL, RecJ, ExoVIII, ExoI and Mug (table 2). A compensating factor may be production of DNA polymerases with increased fidelity and proofreading efficiency as K. radiotolerans contains four exonuclease (e; DnaQ) subunits, three polymerase (a; DnaE) subunits and two b (DnaN) subunits of the replicative DNA polymerase III. Mismatch repair in K. radiotolerans may also be handled by proteins unrelated to PCR was used to generate probes for the first end in the presumptive linear chromosome homologous with a portion of the first gene Krad2223. The genome was digested with one of three restriction enzymes predicted from the DNA sequence to generate a hybridization product with each probe. Lane 1, NcoI; Lane 2, BamHI; Lane 3, BglII. Sizes of molecular markers in base pairs are given in the right hand column. The predicted sizes of the homologous restriction fragments from the DNA sequence are NcoI, 865 bp, BamHI, 3692 bp, and BglII, 1619 bp with the Krad2223 probe. The predicted sizes of the restriction fragments from the other end of the chromosome are NcoI, 775 bp, BamHI, 1363 bp, and BglII, 4018 bp. The observed sizes were calculated using bacteriophage lambda DNA standards digested with HindIII. In all cases the observed sizes were approximately equal to the sizes predicted from a linear chromosome rather than the sum of the sizes from both ends suggesting a linear topology. doi:10.1371/journal.pone.0003878.g002 Figure 3. Resistance of K. radiotolerans to acute c-radiation exposure. E. coli was used as a reference strain. Prior to irradiation, both strains were grown to exponential phase in TGY and LB, respectively. Colony forming units determinations were conducted in triplicate. doi:10.1371/journal.pone.0003878.g003 classical bacterial mismatch repair proteins. Some base excision repair genes that may take on such roles are also uniquely overrepresented. There are three Fpg and four Nei base excisionases in K. radiotolerans, compared with one Fpg and no Nei homologs in D. radiodurans. As an additional example of divergence with D. radiodurans, 3-methyladenine DNA glycosylase I (Tag) is present in K. radiotolerans, but absent from D. radiodurans.
The nucleotide excision repair pathway is also overrepresented in K. radiotolerans by three uvrA orthologs and by five genes encoding UvrD-like helicases. In addition to these helicases, the K. radiotolerans genome, similarly to M. tuberculosis, contains an ERCC3 (XPB)-like superfamily II helicase, whose eukaryotic homolog performs essential functions in nucleotide excision repair and transcription [23]. The K. radiotolerans and M. tuberculosis XPB helicases have been recently demonstrated to be functional in vitro (Biswas and Tsodikov, unpublished).
Five homologs of histone-like proteins (IHF or HupB-like) may package DNA in order to protect it from damage, aid recombinational repair, or maintain multiple chromosomes. The direct reversal pathway used to repair pyrimidine dimers includes two copies of the phrB photolyase gene and an additional, splB photolyase gene, which is absent in E. coli, D. radiodurans and M. tuberculosis. Other differences from M. tuberculosis include the presence in K. radiotolerans of PolA, Topo IV, HolIII, YejH, which are absent in M. tuberculosis. The translesion DNA repair system UmuC/UmuD, present in E. coli but not D. radiodurans or M. tuberculosis, is also a part of the K. radiotolerans genome.
Similar to mycobacteria, the K. radiotolerans replication gene set lacks a well-defined homolog of the helicase loader (DnaC), but contains the other main replication genes (DnaA, DnaB, DnaG). K. radiotolerans and M. tuberculosis both contain a eukaryotic-like DNA primase gene.
In summary, many of the genes known to confer radiation resistance in D. radiodurans are missing from K. radiotolerans suggesting novel components to the repair and protection toolbox. Two pathways are involved in DSB repair in other organisms, ESDSA repair mediated by RecA and PolA and NHEJ repair mediated by Ku and LigD. RecA and PolA are present and may be aided by the presence of RecB and RecC. K. radiotolerans lacks the Ku protein making the presence of the NHEJ pathway unlikely. Base excision and nucleotide excision repair pathway genes are over represented in K. radiotolerans relative to other bacteria.

Reactive oxygen species detoxification
Most organisms have both RecA and PolA without exhibiting extreme radiation resistance. Remarkably, D. radiodurans cells rendered ionizing radiation-sensitive by a polA mutation are fully complemented by expression of the polA gene from ionizing radiation-sensitive E. coli [24]. Repair proteins, either native or cloned, may function better after irradiation in D. radiodurans cells due to protection from protein oxidation [25]. The genetic components and molecular mechanisms of protein repair/ protection remain unknown but appear to be correlated with a Mn/Fe ratio in the range of 0.12-0.37 [26]. The Mn/Fe ratio in K. radiotolerans is 0.09, slightly lower than Deinococcus but much higher than that of radiation sensitive organisms [27].    D. radiodurans is adept at detoxifying reactive oxygen species (ROS) during radiation exposure when many free radicals are generated from hydrolysis of water [13][14][15]. Like Deinococcus, K. radiotolerans produces carotenoids as one level of protection. K. radiotolerans also possesses an impressive suite of genes involved in ROS detoxification (table 3) comparable to those found in bacterial pathogens of mammals such as Neisseria gonorrhoeae [28]. Pathogenic bacteria are routinely exposed to oxidative stress by the host as part of an innate immune response. These results suggest a thorough ROS detoxification network.

Bioremediation potential
Because K. radiotolerans was isolated from a HLW facility, the genome was examined for gene products that could be used to generate carbon or energy from organic compounds in the tanks. Orthologs belonging to known degradation pathways were not detected for U.S. Departments of Energy (DOE) and Defense (DOD) selected priority pollutants for including benzene, toluene, ethylbenzene, and xylenes (the BTEX compounds), chlorinated hydrocarbons, polynuclear aromatic hydrocarbons, polychlorinated biphenyls, ketones, alkanes, phenols, phthalates, explosives, and pesticides. Several homologues for atrazine degradation were noted (AtzABZ, Krad1947, 4281, 0533, respectively), though the pathway seems incomplete.
The SRS HLW also contains low molecular weight organic complexants which interfere with downstream chemical stabilization and processing. Commonly used complexants and decontamination reagents at the SRS include oxalate, glycolate, citrate, and formate. Growth using each organic acid as a sole carbon source was evaluated for K. radiotolerans in comparison with glucose as a positive control and 0.1% yeast extract as a negative control. Respiration was examined using O 2 consumption and CO 2 evolution. Preliminary studies revealed that citrate and glycolate were unsuitable growth substrates (data not shown). Respiration rates and biomass yields on glucose were higher than those of other carbon sources (table 4). Biomass yields with 5 mM oxalate and 5 mM formate were generally consistent with the 0.1% yeast extract control though higher concentrations were toxic. Similarly, O 2 consumption and CO 2 production were both better at lower oxalate and formate concentrations. While biomass production was similar between formate, oxalate, and YE, the respiration rates were higher for formate and oxalate.  While none of the carbon sources stimulated growth, formate and oxalate promote survival and quicken recovery of K. radiotolerans following prolonged starvation (table 5). In these experiments, chloramphenicol was used to accelerate starvation and to uncouple growth from survival. At intervals of 3, 7 or 14 days cells were resuspended in fresh TGY medium and respiration measured. After three days of incubation, substrate specific differences were already apparent. Once starvation was lifted, the highest respiration rates were recorded for cells starved in the presence of oxalate and formate. While all of the cultures recovered, the formate-starved culture was first to enter exponential growth, 4 hours ahead of all other treatments (data not shown). Remarkably, after 7 days of starvation, the highest rate of respiration was measured from the oxalate treated culture, which recovered and began exponential growth 60 hours after starvation was relieved ( Figure 5). The yeast extract treated culture also recovered following more than 100 hours of post-starvation recovery. None of the cultures responded after 14 days of treatment.
E. coli contains three formate dehydrogenases. Formate dehydrogenase H (FdnF is coupled to a hydrogenase and used for hydrogen production by the formate hydrogen lyase reaction. K. radiotolerans appears to have this enzyme (Krad1331). Formate dehydrogenase N (FdnGHI) is induced when cells are grown anaerobically with nitrate.   K. radiotolerans has one of the two (Krad1601, Krad1602, Krad1603), but it is difficult to tell based on the amino acid sequence whether it is an N-type or an O-type. The K. radiotolerans genome encodes glycolysis, pentose phosphate, tricarboxylic acid (TCA) cycle and glyoxylate bypass pathways (table 6). The pentose phosphate cycle [29] and the glyoxylate bypass [30] may both be involved in radiation resistance and post-irradiation recovery as in D. radiodurans. Both pathways are important routes for the production of biosynthetic metabolites while minimizing the production of damaging oxygen radicals. K. radiotolerans exhibits aerobic respiration on 5-and 6carbon sugars (eg., fructose, amylose, maltose) and sugar alcohols (eg., glycerol, mannitol, sorbitol) (data not shown). A partial gluconeogenesis pathway was identified (missing pyruvate carboxylase, pcb, and glucose-6-phosphatase, glp). A partial Entner-Doudoroff pathway was also found, lacking glucose dehydrogenase, gdh, and phosphogluconate dehydratase, edd. Carbohydrate metabolism is supported by many ABC-type sugar transporters. Electron acceptors other than O 2 increase habitat diversity by enabling growth when oxygen tension is low. Krad1328 has strong identity with Arthrobacter aurescens TC1 nitrate reductase. Nitrate respiration may be a general feature of Arthrobacter species as Arthrobacter globiformis and Arthrobacter nicotianae have been shown to utilize NO 2 2 as a terminal electron acceptor [31]. Arthrobacter spp. are routinely recovered from subsurface environments contaminated by high level radioactive waste [32,33]. Krad1328 is not related to the Kuenenia stuttgartiensis enzyme that participates in anammox, anaerobic ammonium oxidation during which nitrite and ammonium are converted directly into dinitrogen gas.

Formate dehydrogenase O (FdoGHI) is expressed aerobically and
The K. radiotolerans genome possesses a large number of molybdopterin oxidoreductases. Krad4344 has striking identity to the trimethylamine-N-oxide (TMAO) reductase TorA from E.coli K12, which is a terminal electron acceptor that is specific for N-oxides. TorA is connected to the quinone pool via a membranebound penta-haem cytochrome (TorC) [34]. DorCA in Rhodobacter is very similar to TorCA with the difference that DorA catalyses reduction of both dimethylsulfoxide and TMAO. K. radiotolerans lacks a homolog to TorC or DorC and reduction of TMAO has not been experimentally examined. There is no evidence for the type of fumarate reductase that reduces fumarate to succinate during anaerobic growth or for proteins homologous those from Shewanella and Geobacter involved in metal reduction.
In summary, K. radiotolerans exhibits moderate nutritional versatility with genes for growth on a variety of carbon sources and at least two terminal electron acceptors. This organism appears to utilize two of the commonly used complexants, oxalate and formate. These results suggest that K. radiotolerans should be investigated for use in scrubbing complexants from HLW tanks. The ability of formate and oxalate to stimulate respiration and prolong viability might provide certain clues into the metabolic ecology of K. radiotolerans. Kineococcus species are commonly found on weathered rock, stone monuments, paintings, desert crust or varnish along with cyanobacteria and fungi which together form biotic crusts, lichens, or endolithic communities ( 1981). Several studies have demonstrated heterotrophic mineralization of oxalate to CO 2 without assimilation but the mechanism remains unknown.

Life cycle
The production of motile zoospores is widespread but patchy among members of the Actinobacteria phylum. All three Kineococcus species produce zoospores during part of their life cycle, but zoospore formation has not been extensively examined in any species [2,35,36]. Zoospores are produced at the tips of substrate hyphae and in clusters on sporophores in the closely related genus Kineosporia [37]. However, K. radiotolerans cells do not produce hyphae or sporangioles. Cells grown in rich PYTG medium produce large clusters of nonmotile cells that alternate division planes to form cubes of cells connected by a thick extracellular matrix [2]. During stationary phase spherical, dispersed zoospores, approximately 1 mm in diameter, are produced. In PTYG medium, the spores lack flagella. Motile spores were observed only after exposure to 10% sandy loam extract or with certain carbon sources [2]. Thus the production of zoospores and the development of flagella on those spores are subject to environmental cues that remain to be elucidated.
Movement of Kineosporia SR11 zoospores up chemical gradients has been observed for a variety of inorganic compounds [38]. Remarkable speeds of up to 160 mm s 21 are achieved in Kineosporia SR11 zoospores [38]. Kineospora SR11 is attracted to K + , Mg +2 , and Ca +2 salts of phosphate, sulfate and halides, but not other combinations of these ions, suggesting that chemotaxis may be dependent on specific cations and anions pairs. The K. radiotolerans genome contains complete pathways for flagellar biogenesis and chemotaxis suggesting that zoospore dispersal is designed to colonize new niches for growth.
The bacterial flagellum is a complex nanomachine that requires dozens of gene products for its assembly and function. Flagellar gene expression is temporally regulated to produce proteins as they are needed for flagellar assembly. Where it has been examined, temporal control of flagellar gene expression is achieved through a transcriptional hierarchy initiated by a master regulator [39,40]. Flagellar genes in K. radiotolerans are clustered within a motility island located between Krad1621 and Krad1673 (table 7). Genes within this region encode components of the basal body, hook, filament, flagellar protein export apparatus, flagellar chaperones, flagellar motor and motor switch. None of the genes within the motility island encode proteins that share homology with known master regulators of flagellar biogenesis though some of the genes encode proteins involved in regulating expression of flagellar biogenesis or motility. One such regulatory protein is a FliA (s 28 ) homolog encoded by Krad1625. FliA is an alternative s factor that is required for transcription of flagellar genes in many bacteria whose products are needed late in flagellar assembly [39,40]. Activity of FliA is often regulated by anti-sigma factor FlgM [39][40][41]. Upon completion of the hook-basal body complex FlgM is secreted from the cytoplasm via the flagellar protein export apparatus, which allows expression of the FliA-dependent flagellar genes. Bacteria that possess a fliA ortholog often also possess an ortholog of flgM [42], suggesting that negative control of FliA activity by FlgM is a general feature of flagellar gene regulation. K. radiotolerans does not possess a flgM ortholog, nor do the other flagellated Actinobacteria sequenced to date (Norcardioides sp. JS614, Acidothermus cellulolyticus and Leifsonia xyli). It is possible that K. radiotolerans possesses a protein that is functionally equivalent to FlgM but does not share homology with it. Alternatively, regulation of FliA may occur through a novel mechanism in K. radiotolerans and related Actinobacteria.
A second regulatory gene within the K. radiotolerans motility island is csrA, which encodes the carbon storage regulator. The csrA gene is also associated with flagellum biosynthetic genes in Norcardioides sp. JS614, A. cellulolyticus and L. xyli. CsrA is a global regulator that binds mRNA to regulate gene expression either positively or negatively [43,44]. CsrA-mediated control of flagellar biogenesis and/or motility has been demonstrated in some Gramnegative bacteria [45][46][47]. CsrA stimulates expression of E. coli flhDC (encodes the master regulator for flagellar biosynthesis) by binding the flhDC operon leader transcript and stabilizing the mRNA [47]. CsrA was recently shown to repress expression of the Bacillus subtilis flagellin gene, hag, by preventing ribosome binding to the transcript [48]. RNA footprinting assays identified two CsrA-binding sites in the hag transcript, one of which (BS2) overlaps the Shine-Dalgarno sequence. The corresponding region of the predicted transcript of the K. radiotolerans flagellin gene (Krad1626) matches well to the BS2 site of B. subtilis hag (9 of 12 bases match). Thus, CsrA may regulate expression of the K. radiotolerans flagellin gene as it does B. subtilis hag.
Developmental control of flagellar genes has been extensively examined during the dimorphic life cycle of Caulobacter crescentus, where a non-flagellated, stalked mother cell undergoes asymmetric binary fission to produce a flagellated swarmer cell [49][50][51]. The swarmer cell eventually differentiates into a mother cell, losing its flagellum and elaborating a stalk. Several genes involved in developmental regulation of flagellar biogenesis in C. crescentus have been identified [49][50][51] including cckA, a histidine kinase in an operon of export genes that activates the master regulator, CtrA [52]. Krad1670, a predicted histidine kinase, appears to be the last gene of an operon that includes fliA, fliR and flhB, genes encoding components of the flagellar protein export apparatus. Given this conservation in synteny K. radiotolerans Krad1670 may play a role in developmental regulation of flagellar biogenesis similar to that of cckA in C. crescentus.
Signal transduction in K. radiotolerans has been examined using the microbial signal transduction database MiST [53]. K. radiotolerans exhibits regulatory complexity comparable to that of Streptomyces spp. with regard to two-component systems. K. radiotolerans is unusual in having 116 di-guanylate cyclase domains. In general these protein domains, which are used to synthesize the second messenger cyclic di-GMP, are over represented in members of the Frankineae suborder of the Actinobacteria but even still K. radiotolerans exhibits large over representation. These results suggests extensive transcriptional regulation by cyclic di-GMP, possibly through the use of riboswitches [54].

Conclusions
Exposure to extreme levels of ionizing radiation does not occur naturally on this planet. While natural nuclear reactors may have existed on Earth about 2 billion years ago, for example in the Oklo region of Gabon, the extreme radiation resistance observed today is more likely to be an evolutionary response to desiccation, which challenges cell viability in a manner similar to ionizing radiation [55]. Like many other radiation resistant bacteria K. radiotolerans exhibits resistance to desiccation [2]. Nevertheless, such organisms may have tremendous potential in remediating metabolizable organic constituents of HLW. Removal of organic constituents directly in HLW tanks could greatly improve processing efficiency of HLW. K. radiotolerans appears to utilize two of the commonly used complexants at the SRS, oxalate and formate, and should be investigated for use in scrubbing complexants from HLW tanks.

Culture conditions and chemicals
Kineococcus radiotolerans SRS30216 (BAA-149) was obtained from the American Type Culture Collection (ATCC; Manassas, VA, USA). Cultures were maintained on TGY plating medium (1.0% tryptone, 0.1% glucose, 0.5% yeast extract, 1.4% Difco agar). Except where noted all experiments were conducted in semidefined growth medium (6.1 g/L Tris (pH 7.0), 0.1% YE) amended with different carbon sources to a final concentration of 5 mM unless otherwise mentioned. All cultures were grown at 28uC and liquid cultures were shaken at 150 rpm. appear to be linear. PCR and blunt end adapter PCR were used unsuccessfully to attempt to close the gaps.

Southern blotting and PCR
A culture of Kineococcus radiotolerans was grown to mid log phase in 100 ml of PTYG (0.5% (w/v) yeast extract, 0.5% (w/v) tryptone, 0.5% (w/v) peptone, 0.006% (w/v) MgSO 4 , 0.0006% (w/v) CaCl 2 ). The cells were collected by centrifugation at 10,000 6 g for five minutes, resuspended in 9.5 ml TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0), and frozen at 220uC for approximately one hour. The cells were quickly thawed at 37uC, mixed with 0.5 ml of 10% SDS and 50 ml of 20 mg/ml proteinase K, and incubated at 37uC for one hour. Then 1.8 ml of 5 M NaCl was added along with 1.8 ml of 1 g CTAB (Hexadecyltrimethylammonium Bromide) and 10 ml of 0.7 M NaCl. The mixture was incubated for 20 minutes at 65uC, extracted with an equal volume of chloroform/isoamyl alcohol (24:1), and centrifuged for 10 minutes at 10,000 6 g at room temperature. The aqueous upper phase was transferred to a fresh tube and the DNA precipitated with 0.6 volume of isopropanol then washed with 70% ethanol. The pellet was resuspended in 2 ml TE buffer and incubated at 37uC with 40 mg/ml DNase free RNase for at least 1 hr.
The sample was extracted with an equal volume of Tris buffered phenol (pH 7.9) and the upper aqueous phase was transferred to a fresh tube, where it was extracted with an equal volume of 1:1 phenol and chloroform and then an equal volume of 24:1 chloroform and isoamyl alcohol. The DNA was precipitated with 0.5 M ammonium acetate and 60% ethanol at 220uC for 20 minutes and then pelleted. The pellet was washed with 70% ethanol and allowed to air dry. The pellet was resuspended in 0.5 ml TE buffer and the DNA concentration was about 1 mg/ml.
Probes were created for each theoretical end of the K. radiotolerans chromosome using PCR and digoxygenin tnucleotides (Roche Biochemicals). Primers Krad2223F 59CGGCAGATACCGGGTTTGATGTTC39 and Krad2223R 59CGATGTCGGTGCGGCTGTAGC39 generated a DNA fragment that was homologous with Krad2223. The PCR reaction contained 0.1 ng of genomic DNA, 0.5 mM of each primer solution. The genomic DNA was initially denatured 2 minutes at 95uC. In subsequent cycles the DNA was denatured for 30 seconds at 95uC, the primers were annealed at 60uC for 30 seconds, and the DNA extended at 72uC for 90 seconds. After 30 cycles the final extension was at 72uC for 3 minutes. The PCR products were separated by agarose gel electrophoresis to determine the success of the PCR reaction.
Genomic K. radiotolerans DNA was digested with restriction enzymes, BamHI, BglII or NcoI as specified by the manufacturer. The restriction fragments were separated by electrophoresis in 0.8% agarose. The gel was submerged in denaturing solution (0.5 M NaOH, 1.5 M NaCl) for 30 minutes and neutralized (0.5 M Tris-HCl, 1.5 M NaCl, pH 7.5) for 30 minutes. The contents of the gel were transferred to a nylon membrane (Roche Biochemicals) using 206 SSC (0.3 M sodium citrate, 3 M NaCl, pH 7.0). The membrane was exposed in a UV crosslinker. Hybridizations were performed in roller bottles with 20 ml hybridization solution and 15 ml probe at 60uC overnight as recommended by the manufacturer (Roche Biochemicals). CPSD was added to the membrane, incubated at 25uC for 10 minutes and 37uC for 15 minutes, then exposed to X-ray film for 1-10 minutes.

Starvation Experiments
K. radiotolerans was grown in TGY medium to mid-exponential phase (,26 hr) at 28uC and 150 rpm. Biomass was harvested by centrifugation, washed in 16 PBS (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 ), and split evenly into different tubes containing 50 mM Tris HCl (pH 7.0), 0.01% yeast extract, chloramphenicol (100 mg/ml), and an individual substrate at a final concentration of 5 mM. Cultures were incubated at 28uC. After 3, 7, and 14 days of imposed starvation, the biomass was harvested, suspended in fresh TGY medium (50 ml), and the ability of each culture to recover and resume growth was compared across substrates.

Respirometry
Metabolic rates were measured using the Micro-Oxymax multichannel respirometer (Columbus Instruments; Columbus, OH, USA). Respirometry bottles (250 ml capacity) were filled with 50 ml of semi-defined growth medium and headspace gases were analyzed every 2 hr. Bottles were incubated at 28uC and 100 rpm. The Micro-Oxymax respirometer measures the volume of gas consumed (O 2 ) or produced (CO 2 ). Conversions were made to moles of gas using the ideal gas law. End point biomass determinations were made in 125 ml shake flasks using 25 ml of growth medium. Cells were filtered onto 0.22 mm GS filters (Millipore) and dried to a constant weight at 65uC.