Deinococcus grandis ATCC43672^T and D. radiodurans ATCC13939^T were purchased from the American Type Culture Collection. E. coli strain JM109 was purchased from Takara Bio. Inc. (Shiga, Japan) (Table 1). D. grandis and D. radiodurans were grown in tryptone glucose yeast (TGY) broth containing 0.5% Bacto Tryptone (BD, NJ, United States), 0.1% glucose, 0.3% Bacto Yeast Extract (BD), or on TGY agar supplemented with 1.5% Bacto Agar (BD) at 30°C, unless otherwise mentioned. E. coli was grown in LB broth-Lennox or LB Agar-Lennox (BD) at 37°C. The following antibiotics were added as needed: for D. grandis, chloramphenicol (3 μg/mL), hygromycin B (50 μg/mL), streptomycin (2 μg/mL); for E. coli, chloramphenicol (15 μg/mL), hygromycin B (100 μg/mL), streptomycin (17.5 μg/mL) or spectinomycin (100 μg/mL).

Genomic DNA was extracted from D. grandis and D. radiodurans using a FastDNA SPIN Kit (MP Biomedicals, CA, United States) and a FastPrep 24 Instrument Version 4 (MP Biomedicals). PCR was performed using Tks Gflex DNA polymerase (Takara Bio. Inc.) or AmpliTaq Gold 360 DNA polymerase (Thermo Fisher Scientific, MA, United States). PCR was performed as per manufacturer’s instructions. Oligonucleotide primers used in this study are listed in Supplementary Table 1. PCR products were purified using a Gel/PCR Extraction Kit (Nippon Genetics Co., Ltd., Tokyo, Japan).

A 263-bp DNA fragment containing the promoter for the DNA damage response repressor-encoding ddrO gene was PCR-amplified using D. grandis genomic DNA as a template and treated with KpnI and NcoI. In addition, using the pNL1.1[Nluc] vector (Promega, WI, United States) as a template, a 564-bp DNA fragment of the luciferase gene from deep sea shrimp Oplophorus gracilirostris was PCR-amplified and treated with NcoI and XhoI. These were ligated to the KpnI-XhoI site of pKatAAD2 to construct plasmid pRNKAAD. Fragments upstream and downstream of the DEIGR_500008 gene were PCR-amplified using D. grandis genomic DNA as a template (769 bp and 770 bp, respectively); they were mixed and treated with the restriction enzymes EcoRI, KpnI, PstI, and HindIII. In addition, a PCR-amplified fragment of 3,025 bp containing the Nluc and streptomycin resistance gene (aad), using the plasmid pRNKAAD as a template, was treated with KpnI and PstI. These three treated fragments were ligated to the EcoRI-HindIII site of pUC19 to construct the plasmid pΔ500008LA (Supplementary Figure 1). Plasmids used in this study are listed in Table 1.

To generate D. grandis strain TY1, wild-type strain was incubated at 40°C for 48 h with shaking and then spread on TGY agar. Ten colonies formed after incubation at 30°C for 48 h were randomly selected, inoculated into TGY broth, and incubated at 30°C for 24 h.

To generate D. grandis strain TY3, wild-type strain was transformed with PCR products that were amplified with primers pKat-FP2 and pKat-RP (Supplementary Table 1) using pΔ500008LA as a template to generate strain TY1Nluc. The strain was cultured in TGY broth supplemented with 5 μg/mL rifampicin at 40°C for 48 h with shaking, diluted 100-fold, inoculated in the same medium, and cultured at 40°C for 72 h with shaking. The culture was then spread on TGY agar and incubated at 30°C for 48 h. Luciferase-nonproductive and streptomycin-sensitive colonies were selected from TGY agar, inoculated into TGY broth, and incubated at 30°C for 24 h.

Using the E. coli vector pACYC184 (Chang and Cohen, 1978) as a template, a 933-bp DNA fragment containing the p15A replicon was PCR-amplified and treated with SphI and KpnI. Plasmid pKatAAD2 (Satoh et al., 2006) was also treated with SphI and KpnI to obtain a 981-bp DNA fragment. These DNA fragments were ligated to yield the plasmid pKatAAD2L. A 2,022-bp DNA fragment containing the replicon of plasmid pDEGR-3 of D. grandis was then PCR-amplified, treated with KpnI, and ligated to the KpnI site of pKatAAD2 to construct plasmid pGRC5 (Supplementary Figure 2).

A 2,109-bp DNA fragment of the shuttle vector pZT29 (Satoh et al., 2009), which contained the replicon of D. radiopugnans plasmid pUE30 and that of E. coli vector pUC19 treated with KpnI, was ligated to a 2,492-bp DNA fragment of plasmid pKatHPH4 (Satoh et al., 2006) treated with KpnI, to construct plasmid pZT29H. In this plasmid, the chloramphenicol resistance gene, a marker gene of pZT29, was replaced with a hygromycin resistance gene (Supplementary Figure 3).

Using plasmid pRADN1 (Ohba et al., 2005) as a template, a 3,164-bp DNA fragment containing the pUE10 replicon was PCR-amplified and treated with XhoI and EcoRV. The plasmid pKatCAT5 (Satoh et al., 2009) was treated with XhoI and EcoRV. These fragments were ligated to construct a plasmid pRADN7. Using the E. coli vector pGBM5 as a template, a 1,997-bp fragment containing the replicon was then PCR-amplified and treated with XhoI and HindIII. Plasmid pRADN7 was treated with XhoI and HindIII to obtain a 4,063-bp DNA fragment. These DNA fragments were ligated to construct a plasmid pRADN8 (Supplementary Figure 4).

Escherichia coli cells were transformed with plasmids using an ECM 399 Electroporation System (BTX, MA, United States). D. grandis was transformed with plasmids using the calcium chloride method based on previous studies. Streptomycin was used as a selection marker for the pGRC5-transformants, chloramphenicol for the pRADN8-transformants, and hygromycin B for the pZT29H-transformants.

Plasmids from the E. coli transformants were extracted as per the standard protocol of the FastGene Plasmid Mini Kit (Nippon Genetics). Transformant cultures of D. grandis (20 mL) were collected via centrifugation and resuspended in buffer mP1 (Nippon Genetics) containing 10 mg/mL lysozyme. After incubating the suspension at 37°C for 30 min, plasmid DNA was extracted using the FastGene Plasmid Mini Kit, as per manufacturer’s instructions.

Quantitative PCR (qPCR) was performed to determine the copy number of shuttle vectors in D. grandis transformants essentially as described (Lee et al., 2006). Total DNA was extracted using the SPINeasy DNA Kit for Microbiome (MP Biomedicals). dnaA was selected as a target gene on the D. grandis chromosome. A 101-bp dnaA fragment was PCR-amplified and ligated with TA-cloning vector pGEM-T (Promega) to yield pDgra-dnaA. This plasmid was used to produce standard curves for the dnaA target. pDEGR-3 rep, pUE10 rep, and pUE30 rep were selected as targets for pGRC5, pRADN8, and pZT29H, respectively. Plasmid copy number was calculated as the ratio of quantified rep to dnaA. Oligonucleotide primers for qPCR analysis are listed in Supplementary Table 1. qPCR was carried out using a QuantStudio 1 Real-time PCR System (ThermoFisher Scientific) with PowerTrack SYBR Green Master Mix (ThermoFisher Scientific) according to the manufacturer’s instructions. The final reaction volume was 20 μL. Standard cycling mode was taken as thermal protocol. A tenfold serial dilution series of plasmids containing dnaA or rep was prepared ranging from 1 × 10⁻⁵ to 1 × 10⁻⁹ ng/reaction and standard curves displaying the cycle threshold parameter (Ct) values plotted against the log of the initial DNA concentration were generated. Plasmid copies were calculated based on the amplicon size of the target genes and the weight of 1 bp (1.095 × 10⁻¹² ng). The relative copy numbers of dnaA and rep genes in samples were determined from qPCR analysis using extracted total DNA with reference to the standard curves.

Plasmid copy number of pRAND8 in D. radiodurans was also determined using qPCR. In this case, a 118-bp D. radiodurans dnaA fragment was PCR-amplified and ligated with TA-cloning vector pGEM-T to yield pDrad-dnaA, and this plasmid was used to produce standard curves for the dnaA target.

Using plasmid pZT90 (Satoh et al., 2009) as a template, a 152-bp fragment containing the D. radiodurans groES minimal promoter region was PCR-amplified and treated with KpnI and NcoI. This DNA fragment was inserted into the KpnI-NcoI site of pRNKAAD to construct pGNKAAD (Supplementary Figure 5). Next, three plasmids were constructed to evaluate the plasmid stability in the transformants without selection pressure. (1) A 669-bp fragment (groEp-Nluc cassette) containing the D. radiodurans groES minimal promoter and deep sea shrimp luciferase gene amplified using pGNKAAD as a template was treated with BamHI and PstI and inserted into the SalI-HindIII site of the shuttle vector pGRC5 to yield the luciferase gene expression plasmid pGRC5GN. (2) The PCR-amplified groEp-Nluc cassette (667 bp) using pGNKAAD as a template was treated with SalI and HindIII and inserted into the SalI-HindIII site of shuttle vector pRADN8 to yield pRADN8GN. (3) The PCR-amplified groEp-Nluc cassette (669 bp), using pGNKAAD as a template, was treated with BamHI and XbaI and inserted into the BamHI-XbaI site of shuttle vector pZT29H to yield pZT29HGN (Supplementary Figure 6).

Evaluation of plasmid stability was performed as described previously (Satoh et al., 2009). D. grandis transformants (strain TY3 carrying pGRC5GN, pRADN8, or pZT29HGN) were cultivated for 24 h in TGY broth supplemented with antibiotics. Cultures were diluted 4,096-fold in TGY broth without antibiotic addition and cultivated for 24 h. Following cell division through 12 generations, cultures were again diluted 4,096-fold in TGY broth. This procedure was repeated until cell division reached 48 generations. At each dilution step, ten milliliters of the culture were transferred to flat-bottomed microplates with 5 mm diameter, and luciferase activity was determined using a Nano-Glo Luciferase Assay System (Promega). Chemiluminescence signals were captured using a LumiCube Chemiluminescence Imaging System (Liponics, Tokyo, Japan) with the Digital Photo Professional 4 software (ver. 4.10.40, CANON) for 5 s at the optimal ISO sensitivity and signal intensities were measured using the JustTLC software (ver. 4.6.3, Sweday, Stockholm, Sweden). Relative chemiluminescence intensity was calculated by setting the chemiluminescence intensity at the 0 generation to 1.

A 263-bp promoter fragment for the DNA damage response repressor-encoding ddrO gene was PCR-amplified using D. grandis genomic DNA and treated with XhoI and NcoI. In addition, a 564-bp fragment of the deep sea shrimp luciferase gene was PCR-amplified using the pNL1.1[Nluc] vector (Promega) and treated with NcoI and HindIII. These were ligated to the XhoI-HindIII site of pGRC5 to construct the plasmid pRDR-Nluc. A PCR product containing the D. radiodurans groES minimal promoter and its downstream region containing the firefly luciferase gene luc + (groEp-luc+) cloned into pZTGL93 (Satoh et al., 2009) was treated with EcoRI and SacI, and then inserted into the EcoRI-SacI site of pRDR-Nluc to yield the dual-luciferase gene expression plasmid pRDR-NlucD (Supplementary Figure 7). This plasmid was used to transform the D. grandis strain TY3 carrying plasmid pZT29H.

The culture of D. grandis transformants (strain TY3 carrying pRDR-NlucD and pZT29H; 1 mL) grown in TGY broth supplemented with streptomycin and hygromycin for 18 h was washed with 1 mL of 10 mM sodium phosphate buffer (pH 7.0) and resuspended in 0.4 mL of the same buffer. The suspensions were transferred to a 35-mm diameter polystyrene dish (Model 3,000–035, AGC Techno Glass Co., Ltd., Shizuoka, Japan) and irradiated at room temperature at constant UV irradiation of 4.2 J/m²/s using four UV lamps. To 0.1 mL of the irradiated suspension, a 25-fold volume of TGY broth supplemented with streptomycin and hygromycin was added and incubated at 30°C for 4.5 h. Every 30 min, 10 μL of the culture was withdrawn and luciferase activity was measured using the Nano-Glo Dual-Luciferase Reporter Assay System (Promega), as described in Section 2.8. Relative reporter activity was calculated as the ratio of chemiluminescence from the experimental reporter (Nluc) to that from the control reporter (luc+). Normalized relative reporter activity was calculated via setting the relative reporter activity at 0 min for each dose to 1.

The data of the plasmid copy number estimation in Section 2.7 was reported as an average of four qPCR assays. The data of the relative chemiluminescence intensity in Section 2.8 was presented as mean ± standard error (n = 9). The data of the normalized relative reporter activity in Section 2.9 was presented as mean ± standard error (n = 3).

Various methods have been developed to delete plasmids from microorganisms (Trevors, 1986; Spengler et al., 2006). In this study, we attempted to culture D. grandis at 40°C, the upper permissive temperature limit for D. grandis, as describes in Section 2.4. The presence of the plasmid pDEGR-3 in bacteria was confirmed using agarose gel electrophoresis. The plasmid was detected in all isolates regardless of incubation at 40°C (Supplementary Figure 8).

Figure 1 shows the results of agarose gel electrophoresis for the presence of PCR products derived from the glutamine synthetase gene (glnN) located on the chromosome of D. grandis, DEIGR_400031, DEIGR_400086, and DEIGR_400127 genes located in pDEGR-PL, and the DEIGR_500007 gene located in pDEGR-3. PCR fragments using primer pairs to amplify the glnN gene (Figure 1, lanes 1 and 6) and genes located at pDEGR-3 (Figure 1, lanes 5 and 10) were detected in both bacterial isolates that were grown at 40°C and those at 30°C. In contrast, three PCR fragments using primer pairs that amplified different DNA regions of pDEGR-PL were detected in the control genome extracted from the bacterial isolate cultured at 30°C (Figure 1, lanes 2–4) but not in that cultured at 40°C (Figure 1, lanes 7–9). Upon culturing at 40°C, the bacterial isolate lost the plasmid pDEGR-PL but retained pDEGR-3, leading to its designation as D. grandis strain TY1.

In addition to high-temperature incubation, rifampicin can eliminate plasmids from microorganisms (Johnston and Richmond, 1970; McHugh and Swartz, 1977). In this experiment, the addition of these chemical agents to the culture medium was combined with incubation at high temperature. To facilitate selecting pDEGR-3 deletion strains from the D. grandis TY1 strain, the DEIGR_500008 gene present in pDEGR-3 was replaced with a luciferase gene from deep sea shrimp and a streptomycin resistance gene to generate the D. grandis strain TY1Nluc (Section 2.4). Then, luciferase-nonproductive and streptomycin-sensitive colonies were selected as described in Section 2.4. The results of agarose gel electrophoresis are shown in Supplementary Figure 9. This result indicates that the bacterial isolate lacked pDEGR-3Δ8LA of 8,961 bp that was present in strain TY1Nluc and designated as D. grandis strain TY3.

Figure 2 shows the genetic organization of the D. grandis cryptic plasmid, pDEGR-3. The plasmid contained nine putative open reading frames, of which the protein encoded by DEIGR_500002 (Accession No. GAQ24009) was considered to be the pDEGR −3 replication initiation protein for the following reasons: Figure 3 shows a comparison of this protein with the replication initiation protein of the cryptic plasmid pUE30 from D. radiopugnans (Accession No. BAH03365), and that of cryptic plasmid pUE10 of D. radiodurans strain Sark (Accession No. AAF44040). The amino acid identity between the replication initiation proteins of pDEGR-3 and pUE30 was 31.7%, and that between the replication initiation proteins of pDEGR-3 and pUE10 was 26.3%. The GC% of a 139-bp non-coding region between DEIGR_500009 (methyltransferase) and DEIGR_500001 (hypothetical protein) was 35.5% and more AT-rich than other plasmid regions because the GC% of the entire pDEGR-3 was 64.9%. The AT-rich region is considered a putative replication origin (Figure 2). Therefore, in this study, the DNA region containing the AT-rich region and DEIGR_500002 was PCR-amplified as a replicon of pDEGR-3 and used to construct a shuttle vector.

The E. coli-D. grandis shuttle vector pGRC5 was constructed as follows: First, plasmid pKatAAD2L was constructed via ligating plasmid pKatAAD2, which expressed a streptomycin resistance gene (aad) controlled by the D. radiodurans catalase gene promoter (katp), with the replicon of the E. coli vector pACYC184. A DNA fragment containing the replicon of pDEGR-3 of D. grandis was PCR-amplified and ligated to pKatAAD2L to yield pGRC5. The second E. coli-D. grandis shuttle vector, pZT29H, was constructed via changing the selection marker chloramphenicol resistance gene (cat) in the plasmid pZT29 to a hygromycin resistance gene (hyg). The third E. coli-D. grandis shuttle vector pRADN8 was constructed as follows: First, plasmid pKatCAT5, which could express a chloramphenicol resistance gene (cat) by the D. radiodurans catalase gene promoter (katp), and the replicon (pUE10 rep) of the D. radiodurans Sark strain plasmid pUE10 to yield pRAND7. The plasmid was ligated to E. coli vector pGBM5 replicon to yield pRADN8. Structures of the three shuttle vectors are shown in Figure 4, and their DNA sequence information is shown in Supplementary Figures 10–12. Plasmid profiles in D. grandis transformants carrying the three shuttle vectors are shown in Supplementary Figure 13. pGRC5 was mainly detected as a monomer in D. grandis. The dimer forms of pZT29H and pRADN8 were also detected (Supplementary Figure 13A). Each of the plasmids tested had a single XhoI digestion site (Figure 4); the size of the linear plasmids after XhoI digestion was in good agreement with the estimated size of monomers (Supplementary Figure 13B).

Relative plasmid copy number was estimated using qPCR assay as described in Section 2.7. The results showed that the relative plasmid copy numbers of pGRC5, pZT29H, and pRADN8 in D. grandis transformants were 11, 26, and 5 per genome, respectively (Supplementary Table 2).

In D. grandis, DNA fragments (groEp-Nluc cassette) with the deep sea shrimp luciferase gene placed downstream of D. radiodurans groES minimum promoter were inserted into pGRC5, pZT29H, and pRADN8 to express foreign genes and construct expression plasmids pGRC5GN, pZT29HGN, and pRADN8GN, as described in Section 2.8. The expression plasmid profiles in D. grandis are shown in Supplementary Figure 14. In lane 1, a band was observed at 4,545 bp derived from pGRC5GN, indicating that the plasmid with pDEGR-3 replicon was present only as a monomer in D. grandis strain TY3, similar to the results for pGRC5 shown in Supplementary Figure 13A. In lane 2, bands were observed at 5,250 bp derived from pZT29HGN monomer and 10,500 bp derived from pZT29H dimer. Lane 3 shows a band at 6,683 bp derived from the pRADN8GN monomer and a larger band indicating the plasmid’s multimers. The size of the linear plasmids after XhoI digestion was in good agreement with the estimated size of monomers (Supplementary Figure6, 14B).

To determine whether the shuttle vectors could stably replicate under non-selective conditions, D. grandis transformants carrying pGRC5GN, pZT29HGN, and pRADN8GN were grown in non-selective broth until 48 generations by repeated dilution and incubation, and relative chemiluminescence intensity was determined as described in Section 2.8. As shown in Figure 5, D. grandis transformants carrying pGRC5GN exhibited constant luminescence intensity until 48 generations, whereas the luminescence intensity of D. grandis transformants carrying pZT29HGN or pRADN8GN decreased as the generation progressed. This result was in good agreement with the analysis of plasmid retention in culture under non-selective conditions using agarose gel electrophoresis (Supplementary Figure 14B).

To determine gene expression induction via UV irradiation in D. grandis, a dual-luciferase reporter plasmid, pRDR-NlucD, was constructed as described in Section 2.9. This plasmid is based on shuttle vector pGRC5. It contains the deep sea shrimp luciferase gene (Nluc) with D. grandis ddrO promoter (ddrOp) and firefly luciferase gene (luc+) with the D. radiodurans groES minimal promoter (groEp). This plasmid was introduced into TY3, which already contained pZT29H, and luciferase activity was monitored after UV irradiation. Figure 6 shows the normalized relative reporter activity in the pRDR-NlucR-transformants exposed to different UV-C doses. When irradiated with 20 J/m² of UV-C, the maximum value (approximately 1.5 times the value immediately after irradiation) was observed 1 h post-irradiation incubation; at 50 J/m², (approximately 4.9 times the value immediately after irradiation) 2 h post-irradiation incubation; at 150 J/m², (approximately 9.8 times the value immediately after irradiation) 3.5 h post-irradiation incubation (Figure 6). Thus, using pRDR-NlucD enabled gene expression induction in TY3 via UV irradiation and revealed the irradiation dose dependence of gene expression levels.

We attempted to introduce the newly developed shuttle vector into D. radiodurans. As expected, D. radiodurans transformants were generated with pRADN8. Relative plasmid copy numbers of pRADN8 in D. radiodurans transformants were 15 per genome (Supplementary Table 2). D. radiodurans transformants carrying pRADN8GN exhibited approximately 15.5% of the initial luminescence intensity at the 48 generation (Figure 5). Similar to previous studies, we were unable to generate D. radiodurans transformants using pUE30-based shuttle vector pZT29H (Satoh et al., 2009). Likewise, D. radiodurans transformants could not be generated using pGRC5.

We previously analyzed the draft and whole genome sequences of D. grandis (Satoh et al., 2016; Shibai et al., 2019); both are available in public genome databases (GenBank Assembly ID: GCA_001485435.1 and GCA_009177165.1). Table 2 summarizes results of the two analyses. Four circular and three linear contigs were registered in the draft genome analysis; four circular DNAs were registered in the whole genome sequence analysis: chromosome, pDEGR-1, pDEGR-2, and pDEGR-3. Next-generation sequencing analyses were performed using genomic DNA extracted from the D. grandis type strain. However, whole genome sequence analysis did not detect one of the circular contigs (BCMS01000006) found in the draft genome analysis. Previous studies have not described a specific name for this plasmid; hence, it was referred to as pDEGR-PL. The laboratory stock of the D. grandis type strain used in draft genome sequencing was used in this study.

In this study, we generated two strains: TY1 that lacked pDEGR-PL, and TY3 that lacked both pDEGR-PL and pDEGR-3. Plasmid pDEGR-PL was deleted via cultivation at 40°C, the upper limit of the permissible temperature, whereas pDEGR-3 was stable against heat treatment. Plasmid pDEGR-PL was undetected via whole genome sequencing in a previous study (Shibai et al., 2019). As the largest coding DNA sequence of the undetected circular contig was similar to that of the phage tail protein, this circular DNA was assumed to be a mobile genetic factor and thus lost prior to whole-genome sequencing. In this study, pDEGR-PL was believed to be easily cured from the D. grandis type strain owing to the instability of the plasmid against heat. Thus, pDEGR-PL may have been transferred from other psychrophiles to D. grandis.

When multiple plasmid vectors are used, homologous recombination reactions between them must also be suppressed. To overcome this, we constructed a shuttle plasmid as a replication initiation region of three plasmids from E. coli with different nucleotide sequences and three plasmids from Deinococcus spp. ligated together, thereby avoiding plasmid instability owing to nucleotide sequence homology. The replication initiation regions of the three plasmids from E. coli used in this study are known to coexist without interfering with each other (Bartolome et al., 1991; Manen et al., 1997). However, the compatibility of replication initiation regions of the three plasmids from Deinococcus spp. is unclear. In this study, we experimentally demonstrated that, despite the limited homology of these replication initiation proteins (Figure 3), the plasmids did not interfere with each other and could stably coexist in D. grandis (Supplementary Figure 13).

Each shuttle vector carries a different antibiotic resistance gene as a selection marker, which ensures the plasmid presence in the bacteria in a selective culture medium, extraction and purification of plasmids from the bacteria, and introduction of plasmids into the bacteria via transformation. The antibiotic resistance genes in the three shuttle vectors are regulated by the D. radiodurans catalase gene promoter (katp), and this promoter region (123 bp) is the same sequence in the three shuttle vectors. Because homologous recombination between different shuttle vectors did not occur under the experimental conditions, homologous recombination in D. grandis is thought to require a longer homologous DNA size. In a previous study, D. grandis could not be transformed with the pUE10-based shuttle vector pRADN1 (Satoh et al., 2009); however, in the present study, D. grandis was transformed with pRADN8. This difference is because of the different plasmid DNA sequences, and likely depends on the presence or absence of plasmid DNA cleavage by the restriction enzymes present in D. grandis.

pGRC5, pZT29H, and pRADN8 have pUC19-derived multiple cloning sites upstream and downstream of the antibiotic marker gene, and some of these restriction sites can be used as cloning sites. However, in the process of constructing the shuttle vector, additional recognition sites were created at other sites in the plasmid, making some of the restriction sites in the multiple cloning sites unusable as cloning sites. By appropriately modifying the nucleotide sequence of the newly developed shuttle vectors, the unusable cloning sites can be restored in the future.

Among the three plasmids, the copy number per genome for pRADN8 was 5. The propensity of pRADN8 and its derived plasmids to form multimeric forms in D. grandis may be related to its low copy numbers (Field and Summers, 2011; Crozat et al., 2014). The pUE10-based shuttle vector pRAD1 was shown to be unretained and shed by D. radiodurans under culture conditions with no selection pressure (Meima and Lidstrom, 2000). Plasmid pRADN8 is presumed to be unstably retained by D. grandis under culture conditions without selection pressure. In this study, we determined plasmid stability in D. grandis transformants carrying plasmids with the groEp-Nluc cassette (Figure 5). As expected, pGRC5 showed high stability in D. grandis strain TY3, whereas pRADN8 showed low stability.

pUE30-based pZT29H was retained in D. grandis, mainly as a monomer and partly as a dimer. The mode of presence of this plasmid was similar to that of pZT29 in a previous study (Satoh et al., 2009). The pUE30-based shuttle vector pZTGL93 that contains the D. radiodurans groES minimal promoter and firefly luciferase gene has been shown to be stably retained by D. grandis wild-type strain for 48 generations under culture conditions without selection pressure (Satoh et al., 2009). Unlike this, D. grandis strain TY3 transformants carrying pZT29HGN exhibited approximately 6% of the initial luminescence intensity at the 48 generation (Figure 5). We suggest that plasmids present in wild strain but absent in strain TY3, namely pDEGR-PL or pDEGR-3, may encode genes responsible for the stability of pUE30-based shuttle vectors.

In this study, we constructed pRDR-NlucD, in which the luciferase gene was controlled by the ddrO promoter in D. grandis. We showed that UV irradiation increased luciferase activity in pRDR-NlucD-transformants (Figure 6). The ddrO promoter region contains an operator sequence called the radiation desiccation-responsive motif (RDRM) that binds to the repressor DdrO protein and represses its gene expression. The DdrO protein is cleaved by the protease activity of the PprI protein activated by DNA damage stress, de-repressing the downstream gene expression (Ishino and Narumi, 2015). This DNA damage stress response mechanism by PprI and DdrO is highly conserved in several Deinococcus spp., including D. grandis, for which genome analysis has been completed (Lim et al., 2019). The dual-luciferase reporter plasmid pRDR-NlucD coexisted with another type of shuttle vector, pZT29H. In the future, cloning genes that affect the DdrO/PprI stress response mechanism into pZT29H could provide a tool to elucidate the molecular mechanisms of the DdrO/PprI-dependent stress response in D. grandis.

In addition to the ddrO promoter, promoters that are upstream of the pprA, ddrA, ddrB, and gyrB genes and have RDRM operator sequences in their vicinity can be used as promoters to induce the expression of foreign genes in D. grandis following UV irradiation (Narasimha and Basu, 2021). Their activity in foreign gene expression in D. grandis should be investigated in the future. Promoters of DR1261, rpmB, and dnaK have been reported to allow constitutively high expression of these genes in D. radiodurans (Chen et al., 2019). Therefore, these promoters should be tested in the future to determine the similarity of their effects on D. grandis. It will also be desirable to develop vectors for the expression of recombinant genes in the periplasm of enlarged D. grandis spheroplasts in the future.

Thus this study reports on the successful development of a novel system that stably introduced and retained expression plasmids for D. grandis. From the wild-type strain, two cryptic plasmids were eliminated to generate the strain TY3. Three shuttle vector plasmids, pGRC5, pRADN8, and pZT29H, were shown to coexist in D. grandis. These plasmids are expected to facilitate the removal of toxic materials and produce useful compounds in D. grandis.