Growth curves measurements were performed as previously described (Dai et al., 2020). WT and mazEF mutant strains were grown to early exponential phase, and then 100 μL cultures were resuspended in 50 mL fresh prepared TGY broth at 30°C. The optical density at 600 nm (OD₆₀₀) was measured at every 4 h for a total incubation period of 40 h. Experiments were repeated three times.

Total RNA of both wild type and mazEF mutant strains were extracted as described (Dai et al., 2020) and then cDNA was amplified using SYBR Premix Ex Taq^TM GC (TaKaRa). RT-qPCR assays were performed using the Mx3005^TMP Real-time detection system (Agilent, United States) and groEL was used to normalize RNA input. Experiments were repeated three times. RT-qPCR primers are listed in Supplementary Table S1.

The mazEF (DR_0416-DR_0417) mutant strain was constructed as previously described (Supplementary Figures S1a,b) (Li et al., 2017). Compared with WT, cells devoid of mazEF gene showed similar cell growth under normal conditions (Supplementary Figure S1c). The transcriptional levels of two adjacent genes, DR_0415 and DR_0418, in both wild-type (WT) and mazEF mutant strains exhibited no significant difference under normal conditions (Supplementary Figure S1d), indicating that disruption of mazEF did not affect the functions of bordering genes. The E. coli strains used in this study were grown in lysogeny broth (LB) (1% tryptone, 0.5% yeast extract, and 1% sodium chloride) or LB agar at 37°C supplemented with 40 μg/mL kanamycin if necessary. All D. radiodurans strains were grown at 30°C in TGY broth (0.5% tryptone, 0.1% glucose, and 0.3% yeast extract) or on TGY agar, containing 30 μg/mL kanamycin or 10 μg/mL streptomycin if needed. The bacterial strains and plasmids used in this study are listed in Supplementary Table S1.

Wild-type and mazEF mutant cells were incubated to an OD₆₀₀ of 0.8, centrifuged, and resuspended in PBS buffer (pH 7.0). Each suspension was divided in half: one half was treated with 15 μg/mL MMC for 30 min, and the other half was incubated without MMC under the same conditions. Total RNA of two biological replicates from each group was isolated using the TRIzol method according to the manufacturer’s protocol (Ambion, Foster City, CA, United States). Four micrograms of RNA per sample was used as the input material for library preparation, and then, a Ribo-Zero^TM Magnetic Kit (Epibio, MRZB12424) was used to remove rRNA. RNA concentration and RNA integrity were assessed using an RNA Nano 6000 Assay Kit and a Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, United States). Index codes were added to attribute sequences to each sample. The resultant libraries were sequenced on an Illumina HiSeq platform using the sequencing by synthesis (SBS) method and paired-end reads were generated. Raw reads were submitted to the NCBI Gene Expression Omnibus (GEO) under accession number GSE97563.

The Kolmogorov–Smirnov test was used for the randomness test to ensure that the position distribution of the read length was distributed evenly. Differential expression analysis was performed using the differentially expressed genes (DEG) method (1.18.0), which provides statistical routines for determining differential expression in digital gene expression data using a negative binomial distribution model. Genes with adjusted p-values <0.05 were considered significantly differentially expressed.

Gene Ontology and KEGG pathway analyses were used to identify the enriched molecular functions and associated biological pathways of DEGs. The enrichment analyses were carried out through an online bioinformatics tool named The Database for Annotation, Visualization, and Integrated Discovery (DAVID2) (Huang et al., 2009). Only terms with a p < 0.05 were considered to be significant. Subcellular localization was analyzed using CELLO¹ v.2.5.

MazF-dr was cloned into the expression vector pET28a+. E. coli BL21 (λDE3) pLysS cells transformed with expression vectors were cultured in LB with 40 μg/mL kanamycin. Protein expression and purification were performed according to a previously described method (Li et al., 2017).

The MazF-dr protein containing an N-terminal 6× His-tag was concentrated to 5 mg/mL, and crystals were grown by the sitting-drop vapor diffusion method, using the Index^TM-HR-144 kit (Hampton Research, Aliso Viejo, CA, United States) for initial screening at 293 K. MazF-dr crystals were optimized and obtained in the reservoir solution containing 0.2 M ammonium sulfate, 0.1 M Bis-Tris (pH 5.5), and 25% w/v polyethylene glycol 3350. Cryocooling was done by soaking the crystals in the reservoir solution containing 10, 20, and 30% glycerol for 3 min and flash freezing in liquid nitrogen. Diffraction intensities were recorded on beamline BL17U at the Shanghai Synchrotron Radiation Facility (Shanghai, China) and were integrated and scaled using the XDS suite (Kabsch, 2010). The structure was determined by molecular replacement using a published MazF structure (PDB ID: 6KYS) (Chen R. et al., 2020) as the search model. Structures were refined using PHENIX (Liebschner et al., 2019) and interspersed with manual model building using COOT (Emsley et al., 2010). Later stages of refinement utilized TLS group anisotropic B-factor refinement. MazF-dr forms a dimer in the crystallographic asymmetric unit. The refined structure includes 110 amino acids of MazF-dr (residues 4–113 from each protomer). All the residues were in the most favorable (97%) and allowed regions (3%) of the Ramachandran plot. All structural figures were rendered in PyMOL².

Deinococcus radiodurans strain was cultured in 500 mL of TGY broth at 30°C until the OD₆₀₀ reached 0.8 and then divided into three parts, two of which were treated with 5 or 15 μg/mL MMC for 30 min, and the other part was untreated as a control group. Cells were then collected by centrifugation, washed twice with PBS, and resuspended in 3 mL of PBS for sonication. After centrifugation, the supernatant was subjected to SDS-PAGE using a 12% acrylamide resolving gel and electrotransferred onto a nitrocellulose membrane (0.22 μm, Pall Biotech, Shanghai, China). The membrane was blocked with 5% non-fat milk in 1× TBST and probed with anti-MazE-dr polyclonal antibody (homemade in the Animal Center of Zhejiang University of Traditional Chinese Medicine) overnight at 4°C, followed by HRP goat anti-rabbit antibody (1:10,000 dilution, Servicebio, Wuhan, China). The membrane was washed with TBST and developed using ECL reagents (Pierce, MA, United States) and imaged by exposure to light-sensitive film. The expression level of GroEL detected by a rabbit anti-GroEL polyclonal antibody was served as an internal control (Sigma, United States).

Inductively coupled plasma mass spectrometry (ICP-MS) was performed as previously described (Dai et al., 2020). Cells were cultured in 500 mL of TGY broth until the OD₆₀₀ reached 0.8, centrifuged and resuspended in TGY broth. Each suspension was divided in half: half was treated with 15 μg/mL MMC for 30 min, and the other half was incubated without MMC for the same duration. After harvesting, cells were washed using PBS containing 1 mM EDTA three times and then rinsed three times with PBS without EDTA. After drying for 12 h with a freeze-dryer (FD-1A-50, BIOCOOL, China), the cells were treated with 5 mL of Ultrex II nitric acid (Fluka AG., Buchs, Switzerland) and 1 mL of H₂O₂ at 100°C for 2 h. The concentrations of metal ions were measured by ICP-MS (ELAN DRC-e, PerkinElmer, United States).

Reactive oxygen species (ROS) levels were measured using 2′,7′-dichlorofluorescein diacetate (DCFH-DA) as a molecular probe. A 4-mL sample of cell cultures (OD₆₀₀ = 0.8) was washed three times with PBS, resuspended in PBS containing 0.1 μm DCFH-DA, and incubated at 37°C for 30 min. After incubation, the cells were washed three times with PBS and resuspended in 4 mL of TGY. Each 1-mL sample was then treated with 10, 20, and 30 μg/mL MMC for 30 min, respectively, and a 1-mL sample without treatment was used as a control. For H₂O₂ treatment, each 1-mL sample was treated with 20, 40, and 80 mM H₂O₂ for 30 min, respectively, and a 1-mL sample without treatment was used as a control. DCFH-DA is hydrolyzed to DCFH by esterase and then oxidized by intracellular ROS to DCF, which produces fluorescence that can be measured using a fluorescence spectrophotometer (SpectraMax M5) at an excitation wavelength of 488 nm and an emission wavelength of 525 nm.

Protein carbonylation was measured using the 2,4-dinitrophenyl hydrazine (DNPH) spectrophotometric method. Cell culture (OD₆₀₀ = 0.8) was treated with 15 μg/mL MMC for 30 min, harvested, and resuspended in PBS. Cells were lysed using sonication, and the protein concentration in the cell-free extract was determined by the Bradford method. The cell extract was then incubated with 10 mM DNPH in 2 M HCl for 2 h in the dark. After precipitation by 10% trichloroacetic acid at 4°C, the precipitated proteins were washed three times with 50% ethyl acetate in ethanol. After evaporation, the protein sediments were dissolved in 6 M guanidine hydrochloride and centrifuged. The absorbance of the supernatant was determined at 370 nm against a protein control that had been processed in parallel but with 2 M HCl instead of DNPH. The protein carbonyl content was quantified in mmol/μg protein.

Sequence alignment revealed that MazF-dr has a similar overall domain arrangement as other bacterial MazF proteins, sharing 23% and 46% amino acid identity with the Bacillus subtilis and E. coli MazF proteins, respectively (Figure 1A). Crystal structure of full-length MazF-dr was determined at 1.3 Å resolution, which is the highest resolution currently known for MazF structures. MazF-dr crystallizes in space group P4₃2₁2 with a twofold symmetric dimer molecule in the crystallographic asymmetric unit. These two protomers could be well superimposed on each other with a root mean square deviation (rmsd) of 0.121 Å over 82 pairs of Cα atoms. The crystal data, together with the data collection and refinement statistics, are summarized in Table 1.

Each protomer of MazF-dr is composed of three α-helices and seven β-strands, which form the typical MazF/CcdB fold frequently observed in published MazF structures (Figure 1B). As a typical MazF protein, MazF-dr could be aligned well with MazF homologs, including MazF from E. coli (PDB ID: 5CK9) (Zorzini et al., 2016), B. subtilis (PDB ID: 1NE8) (Gogos et al., 2003), and Staphylococcus aureus (PDB ID: 4MZM) (Zorzini et al., 2014) (Supplementary Figure S2). The rmsd of MazF-dr among these structures ranged from 0.469 to 1.347 Å over 58–73 Cα atoms, indicating the conserved overall conformation of MazF-dr. We performed the structure similarity search via DALI server using the entire MazF-dr structure as the query, and the highest score was for the MazF-mt1 protein from Mycobacterium tuberculosis (PDB ID: 6KYS, Supplementary Figure S2) (Chen R. et al., 2020). Similarly, the dimeric interface of MazF-dr calculated by the PISA server occupies a total buried surface area of 1,262 Ų (21% of the total solvent-accessible surface area), suggesting the tight dimer formation of MazF-dr (Figure 1C). Two interlocked loops (β1–β2 loop) located at the dimeric interface between two protomers are fully ordered in the MazF-dr structure, in contrast to the high flexibility of this loop in its E. coli counterpart protein (Figure 1C and Supplementary Figure S2) (Kamada et al., 2003). Moreover, two catalytic residues, namely, Arg29 and Thr52 are well organized, suggesting the conserved ribonuclease activity of MazF-dr (Figure 1B).

The protein concentration of MazE in vivo is crucial to the regulation of the MazEF system under environmental stresses. Under normal conditions MazE could self-regulate the expression level of its operon, thus inhibiting the ribonuclease activity of MazF. Previously, we showed that the MazEF-dr system induced the cell death of D. radiodurans in response to a sublethal dose of MMC (Li et al., 2017). To confirm the degradation of MazE-dr under DNA damage, we checked the expression levels of MazE-dr in vivo under MMC treatments (Supplementary Figure S3). While strong immunoblot signals were detected under normal growth conditions, cells treated with 5 or 15 μg/mL MMC displayed decreased MazE-dr signals, implying the activation of the MazEF-dr system upon DNA damage.

To further investigate the possible regulatory role of MazEF-dr-mediated DNA damage response, the transcriptomes of WT and mazEF mutant strains were analyzed in the presence or absence of MMC treatments using RNA sequencing. Four paired-end libraries were generated, yielding a total of 15.7, 15.7, 15.9, and 19.4 M clean reads from R (WT strain without MMC treatment), M (mazEF mutant strain without MMC treatment), TR (WT strain with MMC treatment), and TM (mazEF mutant strain with MMC treatment) with 150 bp as the average sequence length, which corresponded to 83.37–88.25% of the mapped genome of D. radiodurans (Supplementary Table S2). To explore the potential genes regulated by MazEF-dr upon MMC treatment, we compared the transcriptomes of TR vs R, TM vs M, and R vs M, respectively. Numbers of genes that were differentially expressed more than one-fold (log2) (p < 0.05) from each comparison group are summarized in Figure 2A. A total of 203 genes were downregulated and 801 genes were upregulated in the mazEF mutant strain compared with the WT strain under normal conditions (R vs M). In the presence of MMC treatment, a total of 1,029 DEGs were identified in the WT strain (TR vs R), including 961 downregulated genes and 68 upregulated genes. For the mazEF mutant strain (TM vs M), a total of 420 DEGs were identified under MMC treatment, with 354 downregulated genes and 66 upregulated genes. Given that potential genes regulated by the MazEF-dr system are expected to be cleaved by MazF-dr upon MMC treatment, we focused on the downregulated genes in TR vs R but not on those in R vs M or TM vs M to avoid overinterpretation of our data. For example, the transcriptional level of DR_A0320 (a putative urea transporter) was downregulated 2.25-fold (log2) in TR vs R but not significantly downregulated in TM vs M [0.13-fold (log2)] or R vs M [2.81-fold (log2)], suggesting that this gene was potentially regulated by the MazEF-dr system upon MMC treatment. As a result, a total of 650 downregulated genes in TR vs R were chosen as the potential genes regulated by the MazEF-dr system in response to MMC treatment in D. radiodurans (Figure 2B and Supplementary Data Sheet 1).

We next performed functional annotation analysis using the GO and KEGG databases. In the biological process category, these genes were involved in multiple biological processes including transmembrane transport, RNA processing, cell division, cell wall organization, regulation of transcription, and carbohydrate metabolism (Figure 2C). The results of KEGG pathway analysis revealed that these genes were mostly enriched in ATP-binding cassette (ABC) transporters participating in iron transportation, osmotic tolerance, and cell division (Table 2). These results suggested that the MazEF-dr system regulates gene expression upon MMC treatment in a variety of cellular processes. Moreover, approximately 30% of the proteins encoded by these genes were located in the cell membrane (Figure 2D), which may explain the observed markers of apoptotic cell death, e.g., membrane blebbing and loss of membrane integrity, in D. radiodurans upon MMC treatments (Li et al., 2017).

Among all the 650 genes potentially regulated by the MazEF-dr system upon MMC treatment, we found that two sets of enriched genes in the KEGG pathway analysis, and these genes were involved in the iron complex transport system (Table 2). Three of these genes (DR_B0122, DR_B0123, and DR_B0125) encode the key components of the Fec system (FecBCD), which is important for ferrous iron transport. Two genes (DR_B0014 and DR_B0016) are involved in the transportation of the ferric-heme complex. Given that these genes were downregulated in the WT under MMC treatment, we next measured the intracellular concentrations of both Fe and Mn, using ICP-MS analyses (Figure 3A). Despite the similar intracellular concentration of Mn, the mazEF mutant strain had a 29.5% lower intracellular concentration of Fe than the WT strain in the absence of MMC. Upon MMC treatments, the intracellular concentrations of Fe and Mn in the WT strain (wild type + MMC in Figure 3A) increased by 70.1% and 40.9%, respectively. However, such elevation in metal ion concentrations was severely compromised in the mazEF mutant strain under MMC treatment (mazEF + MMC in Figure 3A), indicating that the MazEF-dr system plays a role in the increased intracellular concentrations of Mn and Fe upon MMC treatment.

Excessive iron can cause damage to cellular components, e.g., proteins, through ROS formation via the Fenton reaction or Haber-Weiss reaction, which may further lead to cell death (Wardman and Candeias, 1996). Given the elevated intracellular Fe level in the WT strain upon MMC treatment, the total intracellular ROS levels were measured to further investigate the function of the MazEF-dr system in D. radiodurans (Figure 3B). The intracellular ROS levels in the WT strain increased with increasing concentrations of MMC, showing an MMC dose-dependent ROS accumulation profile (Figure 3B). In contrast, the ROS levels of the mazEF mutant strain were much lower than those of the WT strain, with compromised ROS accumulation upon MMC treatment. Similar results were observed when cells were treated with various concentrations of H₂O₂; much less ROS accumulation was observed in the mazEF mutant strain than in the WT strain (Figure 3C). These results were consistent with the distinct profile of Fe concentration in the WT and mazEF mutant strains in response to MMC treatments (Figure 3A). To further confirm the accumulation of ROS in vivo, protein carbonylation, the widely used marker of severe protein oxidation, in both the WT and mazEF mutant strains upon MMC treatment was measured (Figure 3D). Consistent with the increased ROS levels, high level of protein carbonylation (3.5-fold increase) was observed for the WT strain under MMC treatment. However, the mazEF mutant cells showed a significantly reduced level of protein carbonylation under MMC treatment. These results indicated that the MazEF-mediated cell death in D. radiodurans could be possibly caused by an increase in ROS accumulation upon DNA damage.