Deinococcus radiopugnans DY59 was kindly provided by Dr. M. Kim at Seoul Women’s University. This strain was deposited into the Korean Collection for Type Cultures (KCTC33033^T). Deinococcus geothermalis DSM11300^T and D. radiodurans R1 were used as a positive control for the active transposition of ISs. Deinococcus geothermalis and D. radiodurans R1 were cultured on TGY medium containing 1% tryptone, 0.5% yeast extract, and 0.1% glucose with 150 rpm for broth culture at 48 and 30°C, respectively. Strain DY59 is known to grow well in R2A or LB medium. Therefore, the cells were cultured for 3–4 days at 30°C in R2A medium, whereas the cells in the TGY medium required only 2–3 days to grow. All downstream Deinococcus cultures were thus conducted using TGY culture medium.

Deinococcus radiopugnans DY59 cells were typically grown to a maximum optical density (OD₆₀₀) for 2–3 days. When the culture reached an absorbance of 2.0 or 4.0 at OD₆₀₀ as measured by a UV–VIS spectrophotometer, the cells were harvested by centrifugation at 3,000 rpm and resuspended in a 0.9% NaCl solution to an OD₆₀₀ of 2.0 for exposure to H₂O₂.

Next, the cells were treated with final H₂O₂ concentration ranges of 80 and 100 mM/200 and 300 or 0 mM as a control and continuously cultured for 1 h at 30°C with 150 rpm shaking. The samples were then directly diluted to 10⁻⁴–10⁻⁵ with 0.9% NaCl solution and 100 μL of the cell suspensions were spread on TGY agar medium using sterilized glass beads. After discarding the glass beads, the TGY plates were incubated at 30°C for 2–3 days. Non-pigmented colonies were detected and cell culture dilutions were streaked on TGY agar plates for pure cultivation. Non-pigment forming frequency was calculated the ratio of number of non-pigment colony and CFU. We also evaluated whether hypochlorite and gamma irradiation induced the occurrence of IS transposition. Active transposition was detected via PCR amplification of four selected target genes related to carotenoid biosynthesis (Supplementary Table S1).

The D. radiopugnans DY59 genome sequence was obtained from the NCBI database (GenBank accession number NZ_CP010028.1). IS elements were detected from the annotated transposase genes using protein profiling data obtained from the GenBank database. We first extracted sequences from upstream and downstream of a transposase (1 kb region; 3 kb total length). Then, terminal inverted repeat (TIR) sequences and direct repeat (DR) sequences were determined using DNA sequence alignment software such as BLAST, ClustalW, and BioEdit. All detected IS elements were sorted into IS family members. Unfortunately, the IS detection platform ISfinder (https://isfinder.biotoul.fr; Siguier et al., 2006) still does not support the detection and classification of D. radiopugnans DY59 IS elements with calling ID as of September 2022. Therefore, we classified the IS ourselves based on IS structure analysis and nomenclature rules (Supplementary Table S2).

Afterward, 10 ml of cultured cells till to an OD₆₀₀ of 2.0 and 4.0 were harvested, resuspended with 0.9% NaCl, and separately exposed for 1 h to various H₂O₂ concentrations or 100 μg/mL sodium hypochlorite treatment or for in total 5 kGy gamma-irradiation (160 Gy/min). Then, to measure the cell viability using colony forming units, samples were diluted to a 10⁻⁵ ratio and spread on TGY agar plates. After 5 kGy gamma-irradiation, DY59 strain was exhibited 0.5% viability comparing to non-radiation. The viability of hypochlorite treatment was not analytical determined. To measure the minimum inhibitory concentration (MIC) of antibiotics, the strain DY59 cells of 100 μL on OD₆₀₀ = 4.0 were spread on TGY agar plates, and the disk diffusion assay was conducted with different concentrations of 10 μL antibiotic solutions, such as streptomycin, kanamycin, and ampicillin. The MIC values were determined based on the formation of a clear zone.

Strain DY59 cells were harvested at OD₆₀₀ values of 2.0 and 4.0, resuspended to OD₆₀₀ of 2.0 in 5 mL 0.9% NaCl, and then treated with 50–150 mM H₂O₂ for 1 h at 150 rpm. After the H₂O₂ challenge, the samples were centrifuged at 10,000 rpm for 5 min and washed once more with 0.9% NaCl. The supernatants were discarded and the pellets were stored at −20°C. The cell wall was broken using phenol and DNA digestion was performed using DNase I. RNA was extracted using an RNA prep kit for RNA isolation (RNeasy mini purification kit; Qiagen, Germany). After measuring the extracted RNA concentration, the concentration was normalized to 1,000 ng in an 8 μL volume for all samples. cDNA synthesis was performed with a dNTP mixture and 6-mer random primers using the following protocol: 60°C for 5 min, 4°C for 3 min, 30°C for 10 min, 42°C for 60 min, and finally 95°C for 5 min (PrimeScript™ first strand cDNA Synthesis Kit; TaKaRa, Japan). In the step at 4°C for 3 min, 4 μL 5× buffer, 4.5 μL RNase free water, 1 μL RTase, and 0.5 μL RNase inhibitor were added. qRT-PCR analysis was performed using TB Green® Premix Ex Taq™ (TaKaRa, Japan) on a Bio-RAD RT-PCR model CFX96™ Optics Module (Bio-RAD, United States). The expression of catalase (QR90_06310) and three LysR family regulators (QR90_13110, QR90_14595, and QR90_15105) was normalized to that of GAPDH, a gene that is constitutively expressed throughout all growth phases. The related expression levels of both catalase and LysR genes were calculated as described in a previous study (Choo et al., 2020). Differences between samples were determined via the Student’s t-test using the Prism™ software (ver. 8.0). Differences were deemed statistically significant at p < 0.05 (^*) and p < 0.01 (^**). The primer sequences for qRT-PCR were shown in Supplementary Table S3.

Generally, D. radiopugnans DY59 formed reddish-colored colonies after being cultured for 2–3 days on TGY medium at 30°C. The strain also exhibited a stronger viscosity compared to D. geothermalis. Here, we observed different MIC values, which were indicative of different levels of antibiotic resistance. The MIC of streptomycin was less than 25 μg/mL, whereas those of ampicillin, kanamycin, and chloramphenicol were 30, 300, and 60 μg/mlL, respectively (Figure 1A; Supplementary Figure S1). Therefore, we conclude that strain DY59 displays a resistant phenotype to kanamycin and chloramphenicol, and a tolerant phenotype to streptomycin and ampicillin when compared to the typical working concentrations of antibiotics.

Surprisingly, the results of our H₂O₂ viability assays indicated that D. radiopugnans DY59 could form colonies after being exposed to H₂O₂ concentrations of up to 600 mM for 1 h, which is two times higher than the resistance of two well-studied control strains of D. radiodurans and D. geothermalis (300 mM H₂O₂; Figure 1B).

Genes related to carotenoid biosynthesis in genus Deinococcus, which were linked to the phenotypic reddish color of the colonies, were detected via KEGG pathway analysis and genome-wide genomics studies (Tian and Hua, 2010; Lim et al., 2019). We selected four genes that are involved in the carotenoid biosynthesis pathway, which were marked in a simplified pathway schematic (Figure 2). The selected four carotenoid biosynthesis genes of strain DY59 exhibited amino acid sequence similarity with more than 60% to D. geothermalis genes: QR90_03795, a phytoene synthase, is 69.4% similar to Dgeo_0523; QR90_10400, a phytoene dehydrogenase (desaturase), is 78.9% similar to Dgeo_0524; QR90_14380, a carotenoid hydratase, is 58.6% similar to Dgeo_2309; and QR90_14400, a FAD-dependent oxidoreductase, is 74.5% similar to Dgeo_2306 (Supplementary Table S1). Interestingly, some carotenoid biosynthesis-related genes of D. geothermalis were adjacently clustered in the genome, for examples crtB-crtI and cruC-cruD-cruF-crtO. In contrast, each carotenoid biosynthesis gene in D. radiopugnans was separated.

From the NCBI genome sequence data (submitted at December 2014) and protein information of DY59 (Kim et al., 2015), we first selected transposases and their genomic loci and defined full-length IS elements. The genome of strain DY59 contains six defined IS family members with DDE type transposase (Tpase) including IS1, IS4, IS5, IS66, IS630, and IS701, as well as five unclassified Tpases within a total of 36 ISs (Supplementary Table S2; Supplementary Figure S2). In general, Tpase is a major component of IS elements and simplified schematics of the IS structures except for the IS1 family member, which has only a partial fragment, are illustrated in Supplementary Figure S3.

The IS1 family member is an IS element (QR90_RS04110) with a 119 aa-length Tpase. In general, the IS1 family IS element is composed of two transposase-related open reading frames (ORFs) with the same transcriptional direction. Therefore, TIR and DR sequences for this IS region of the IS1 family could not be easily determined because it was only a partial fragment.

The IS4 family members had five copies in the genome. The IS elements including transposases were separated into two subtypes. QR90_RS09690, QR90_04660, and QR90_01215 have identical IS elements with a 320 aa-long transposase, a conserved TIR sequence (CTCTGTACCGGACAACT), and DR sequences of variable sequences with 9 nt (“GCCGTGATC,” “GTCACGCCC,” and “GAAAGCAAT,” respectively). QR90_RS10860 and QR90_04150 possess another identical IS element with an identical transposase with 327 aa-length. Both IS elements have a conserved TIR sequence “CTCGGTAGCTGACAACTTCA” and DR sequences “AGATTGAAC” and “CAGGGTCAG.” Thus, these IS4 family members were divided into the two subtypes ISDrpg2 and ISDrpg3, with their Tpase sequences having a 67.19% amino acid identity.

The IS5 family members have four copies in the genome with two subtypes elements: QR90_RS06350 with a 277 aa-length Tpase and QR90_RS04355, QR90_RS07350, and QR90_06585 with 265 aa-length Tpase sequences. The three 265 amino acid-long transposase genes are highly similar, with similarity rates of 97.7–100%. However, the Tpase of QR90_RS06350 with a 277 amino acid length exhibited a quite low similarity of 12–12.8% to the 265 aa-length Tpase in the IS5 family. Therefore, the IS5 family members were separated into two subtypes: ISDrpg4 and ISDrpg5. The TIR and DR sequences of three IS elements with a 265 aa length Tpase were “AGGCTG” and “TAG,” respectively. The TIR and DR sequences of QR90_RS06350 were “ACCTCCTGCGAAAGTC” and “TAG,” respectively. The structural schemes for the forward and reverse area of the Tpase are identical, with a 54 and 9 nt distance, respectively. One interesting finding is that the loci of the IS5 family IS elements are positioned close to the IS701 family member, except for QR_ RS06350.

The IS66 family IS elements have five copies in the DY59 genome: QR90_07275, QR90_RS09525, QR90_07340, QR90_11595, and QR90_09840, with a 472 aa-long Tpase named ISDrpg6. The IS elements have a conserved terminal inverted repeat (TIR) sequence “GTCTGTGATTAGCGGTCG” and 8 nt variable direct repeat (DR) sequences (“GATGGGGG,” “GGTGCAGG,” “ATGTCGTC,” “GGC GAGAG,” and “TATTTTTG”).

The IS630 family members were divided into the two subtypes ISDrpg7 and ISDrpg8: QR90_RS17010 with a 181 aa-long Tpase, QR90_RS08750, QR90_RS17180, QR90_RS17220, QR90_RS17305, and QR90_RS17410 with a 187 aa-long Tpase. Although QR90_17010 and QR90_RS17410 exhibited different Tpase lengths, the amino acid sequence similarities were 100% identical, whereas the 187 aa-length Tpases exhibited sequence identities ranging from 92.8 to 98.9%. In contrast, the Tpase of QR90_RS08750 exhibited a 64.7–67.9% identity when compared to the 187 aa-long Tpase. Thus, ISDrpg7 included five IS elements except QR90_RS08750 for ISDrpg8. All of the examined IS630 family members exhibited three-nucleotide DR sequences “TGA/TAA/TCA” and their TIR sequence was “TACGGACTCCGATTAA.”

The IS701 family members included the 10 IS elements QR90_RS00720, QR90_04350, QR90_RS06590, QR90_05955, QR90_10425, QR90_RS17070, QR90_RS17080, QR90_RS17170, QR90_RS17235, and QR90_RS17395, with a 432 aa-length. Two distinct subtypes were identified according to Tpase identity: QR90_RS17070, QR90_RS17080, QR90_RS17170, QR90_RS17235, and QR90_RS17395 had a DNA sequence of 100% identity, whereas QR90_RS00720, QR90_04350, QR90_RS06590, QR90_05955, and QR90_10425 had a 99.5% identity. In contrast, the two subtypes exhibited only a 89.12% Tpase identity between each other. All IS701 family members had a unique DR sequence (“nTAG”) and TIR sequence (“CTGTACTTTG GGGATATTCA”). Interestingly, the 3′ end of the TIR sequence of all IS701 family members overlapped into the Tpase ORF.

In this study, we identified five unclassified IS members: QR90_08760 with a 270 aa-long Tpase; QR90_RS08625 with a 477 aa-long Tpase; QR90_08735 with a 434 aa-long Tpase; QR90_RS05950 with a 120 aa-long Tpase; and QR90_04880 with an 89 aa-long Tpase.

Next, the active transposition of IS elements was detected on non-pigment phenotypic selection under oxidative stress conditions induced by H₂O₂ treatment in the present study.

A total of 24 and one non-pigmented colonies were detected after low concentration (80 and 100 mM) and high concentration (200 and 300 mM) H₂O₂ treatment for 1 h, respectively. It seems that generation of non-pigmented mutants is less efficient with higher H₂O₂ concentration.

In the high-concentration treatment, one is a complete non-pigmented strain (w3) and two isolates exhibited a pale reddish color on TGY plates (w1 and w2; Figure 3A). When carotenoid biosynthesis genes were amplified by PCR with the appropriate primers, the complete non-pigmented strain only exhibited an enlarged PCR product from QR90_10400, a phytoene desaturase (Figure 3B). However, other carotenoid biosynthesis genes were not affected. The ISDrpg3 of the IS4 family was integrated at the 275th nucleotide of QR90_10400 (Figure 4). The TIR sequence of this IS element was “CTCGGTAGCTGACAACTTCA” and the DR sequence was “ACCCGCCCC.”

In the low-concentration H₂O₂ treatment, five out of 24 non-pigmented strains exhibited PCR amplicons with QR90_10400 gene enlargement: two strains with an OD₆₀₀ of 2.0 and three strains with an OD₆₀₀ of 4.0 at 80 mM H₂O₂ (Figure 5A). The frequency of IS transposition in gene QR90_10400 exhibited 5.1 × 10⁻⁴ and 1.8 × 10⁻⁴ from OD₆₀₀ of 2.0 and 4.0, respectively (Supplementary Table S4). All five of these IS-integrated mutants exhibited IS4 family transposition (Figure 5B; Supplementary Figure S4). Other carotenoid-related genes QR90_03795, QR90_14380, and QR90_14400 were not affected (Figure 5C). The 19 remaining clones did not exhibit any changes in the size of the PCR products, indicating that there was no IS transposition in the selected four carotenoid biosynthesis genes. However, there might be point mutations in carotenoid-related genes or perhaps IS transposition in non-analyzed carotenoid biosynthesis-related genes.

At an OD₆₀₀ of 2.0, one IS element (w3) was integrated in the forward direction at the 401st nt of phytoene desaturase. The DR sequence of this element was “CTTCTTCGA” and the TIR sequence was “CTCGGTAGCTGACAACTTCA.” Another (w7) integrated in the reverse direction at the 617th nt. The DR sequence of this element was “GTAAACGAG” and the TIR sequence was “CTCGGTA GCTGACAACTTCA” (Figure 5B; Supplementary Figure S4A). Both IS elements were belonged to ISDrpg3 of QR90_04150 or QR90_RS10860 (Supplementary Table S2).

At an OD₆₀₀ of 4.0, one IS element (w5) integrated in the reverse direction at the 587th nt of phytoene desaturase. The DR sequence of this element was “AGGCGCTC” and the TIR sequence was “CTCGGTAGCTGACAACTTCA.” Another (w11) integrated in the reverse direction at the 474th nt; its DR sequence was “GCTCGTAGC” and its TIR sequence was identical to that of w5. Both IS elements were identical to ISDrpg3 (Supplementary Figure S4B). The last IS element (w7) integrated in the forward direction at the 22nd nt with DR sequence “CCAGCAGGC” and TIR sequence “CTCT GTACCGGACAACT” (Figure 5B; Supplementary Figure S4C). This IS element was identical to ISDrpg2 of QR90_01215, QR90_04660, or QR90_09690. Therefore, D. radiopugnans DY59 exhibits active transposition of IS4 family members in H₂O₂ treatment conditions. The PCR detection of five copies of IS4 family members at the location as found in the genome sequence indicates that the active transposition occurred through replicative mode in present (Supplementary Figure S5).

The 100 μM sodium hypochlorite does not affected viability of the strain DY59. 5 kGy irradiated strain DY59 exhibits 99.5% reduction of CFU. The selection of non-pigmented colonies was performed via gamma irradiation exposure and sodium hypochlorite treatment. Unlike the lack of pigment production of D. geothermalis, gamma irradiation of total 5 kGy did not induce non-pigment phenotypes in the D. radiopugnans DY59 wild-type strain (Ye et al., 2022). Moreover, although hypochlorite treatment induced the non-pigment phenotype, the non-pigmented colony did not exhibit IS element integration in the four analyzed carotenoid biosynthesis genes (data not shown). This may be the same explanation for the absence of transposition of IS in phenotypic changes due to point mutations in four carotenoid biosynthetic genes and the decay of other genes related to pigment formation.

To determine the mechanisms underlying the high H₂O₂ resistance of DY59, the expression levels of catalase and three LysR family regulators including possible oxyR were first measured by qRT-PCR analysis. The chromosome of DY59 strain has a single catalase QR90_06310 with 72.17 and 74.17% amino acid sequence similarity to D. radiodurans catalase KatE1 and D. geothermalis KatE, respectively. After exposure to various concentrations of H₂O₂ (50, 100, 150 mM, and an unexposed control) for 1 h and growing the cells to OD₆₀₀ values of 2.0 and 4.0, the relative expression levels of catalase and LysR family members were measured via qRT-PCR by the basal expression level of OD₆₀₀ of 2.0 with unexposed control as 1.0. Unexpectedly, catalase was not dramatically induced at OD₆₀₀ 2.0 or 4.0 in any of the tested H₂O₂ concentrations. Nevertheless, there was a 2-fold increase in catalase expression at OD₆₀₀ 4.0 when the cells were challenged with 50 and 100 mM H₂O₂ (Figure 6A).

LysR1 QR90_13110 exhibited a more than 6-fold induction at an OD₆₀₀ of 4.0 upon exposure to 100 mM H₂O₂ and LysR3 QR90_15105 exhibited a more than 6-fold induction at an OD₆₀₀ of 2.0 in the 100 mM H₂O₂ condition. Our findings thus indicated that 100 mM H₂O₂ controlled the expression of LysR1 and LysR3 at two different growth phases (Figure 6B). However, LysR2 (QR90_14595) was not affected regardless of growth phase and oxidation condition (data not shown). LysR2 was exhibited 70.3 and 75.7% identities of amino acid sequence to the proposed OxyR of D. radiodurans and D. geothermalis, respectively. Therefore, the high H₂O₂ resistance phenotype of the DY59 strain cannot be attributed to changes in the expression of catalase and its three potential LysR family regulator members, as determined by RNA production levels measured by qRT-PCR analysis.

There are two subtypes of IS4 family which were actively transposed to other loci. Tpase of ISDrpg3 such as QR90_04150 and QR90_10860/ISDrpg2 such as QR90_01215, QR90_04660, and QR90_09690 exhibited a more than 5-6-fold induction at an OD₆₀₀ of 4.0 in the 50 mM H₂O₂ condition. The related expression levels were reduced at an OD₆₀₀ of 4.0 in the 100 and 150 mM H₂O₂ condition (Figure 6C). However, both ISDrpg2 and ISDrpg3 Tpases were gradually induced over 2-fold at an OD₆₀₀ of 2.0 in the different H₂O₂ conditions. Thus, IS elements of the IS4 family were induced by the H₂O₂ and actively transposed into other genomic loci with replicated mode.