Overlapping oriC and centromere-like functions in secondary genome replicons determine their maintenance independent of chromosome I in Deinococcus radiodurans

The Deinococcus radiodurans multipartite genome system (MGS) consists of chromosome I (ChrI) and secondary genome elements; Chr II and megaplasmid (MP). The sequences upstream to parAB operons in Chr II (cisII) and MP (cisMP) helped an E. coli plasmid maintenance in D. radiodurans and showed sequence specific interactions with DnaA and ParBs. The cells devoid of cisII (ΔcisII) or cisMP (ΔcisMP) showed reduced γ radiation resistance and copy number of Chr II and MP. Fluorescent Reporter-Operator System (FROS) developed for ChrI, ChrII and MP in ΔcisII or ΔcisMP mutants showed no change in wild type pattern of Chr I localization. However, the relative copy numbers of Chr II and MP had reduced while anucleate cells had increased in mutants. These results suggested that cisII and cisMP elements contain both ori and centromere-like functions, and like other MGS bacteria, the Chr I and secondary genome are maintained independently in D. radiodurans.


Introduction
Genome duplication and segregation are highly coordinated macromolecular events critical for growth of any organism 1 . The chromosomal replication starts at oriC a well conserved region comprised of the non-palindromic repeats of 9 mer as DnaA boxes and 13 mer AT rich repeats [2][3][4][5] . In the best studied, single circular chromosome harbouring, E. coli model, the DnaA boxes are bound by ATP-complexed DnaA protein following the recruitment of replication initiation complex comprised of DnaB helicase with the help of DnaC, primase and DNA polymerase III 6,7 . Segregation of duplicated replicons is also linked with the number of copies per cell. Various models of active genome segregation by tripartite genome segregation (TGS) system have been suggested for single circular chromosome and limited copy large plasmids 8 . In TGS, the nucleation of adapter protein like ParB on centromeric sequences is a limiting step for faithful segregation of duplicated chromosome 9 . The distribution of high copy number plasmids in dividing population is believed to be a random passive process 10 . It has been observed that the initiation of DNA replication and segregation of duplicated genome occurs concurrently and therefore believed to be interdependently regulated in bacteria 11 . In majority of the cases where DNA replication and segregation have been studied, the centromeric sequences are found close to ori. For instance, B. subtilis and XbaI in case of pNOKcisII and ApaI and EcoRI in case of pNOKcisMP and release of desired inserts was verified on agarose gel.

Protein purification
The recombinant DnaA of D. radiodurans (hereafter referred as DnaA unless other mentioned) expressing on pETDnaA in E. coli BL21 (DE3) pLysS was purified by Ni-NTA affinity chromatography, as described earlier 36 . In brief, E. coli cells were treated with 0.5 mM isopropyl-β-D-thio-galacto-pyranoside (IPTG) and cell pellet was suspended in buffer A (20 mM Tris-HCl, pH 7.6, 300 mM NaCl and 10% glycerol), 0.5 mg/ml lysozyme, 1 mM PMSF, 1mM MgCl 2 , 0.05 % NP-40, 0.05 % TritonX-100, protease inhibitor cocktail), and incubated on ice for 1h. The cell lysate made by sonication was centrifuged at 11,000 rpm for 30 min at 4 • C and the clear supernatant was purified through Ni-NTA column chromatography (GE Healthcare) following kit protocols. The column was thoroughly washed with buffer A containing 50 mM imidazole. Recombinant protein was eluted in steps using 100 mM, 200 mM, 250 mM and 300 mM imidazole in buffer A and analyzed on 10% SDS-PAGE. Fractions with >95% pure protein were pooled and re-purified from HiTrap Heparin HP affinity columns (GE Healthcare Life sciences) using a linear gradient of 0 -300 mM NaCl, followed by precipitation with 30% w/v ammonium sulphate at 8˚C and molecular exclusion chromatography using standard protocol. The recombinant ParBs encoded on chromosome II (ParB2) and megaplasmid (ParB3) in D. radiodurans were also purified from recombinant E. coli using Ni-NTA affinity chromatography followed by molecular exclusion chromatography as described earlier 37 . Protein concentration was determined by taking OD at 280 nm in Nano drop (Synergy H1 Hybrid Multi-Mode Reader Biotek) and mass extinction co-efficient of the proteins.

DNA-protein interaction studies
DNA binding activity of DnaA, ParB2 and ParB3 was monitored using electrophoretic mobility shift assay (EMSA) with different version of cis elements as DNA substrates 36 . In brief, full length cisII and cisMP and their repeat derivatives (Fig. S2) were PCR amplified using sequence-specific primers (Table S1). The PCR products were gel purified and labeled with [γ-32 P] ATP using T4 polynucleotide kinase. Two repeats and one repeat of both the cis elements were chemically synthesised, radiolabelled with [γ-32 P] ATP using T4 polynucleotide kinase, and annealed with unlabelled complimentary strands (Table S1).  (Table S1) and a 10-fold higher concentration of same cis sequence. Mixtures were separated on 6-10 % native PAGE gel as required, dried, and autoradiogram was developed on X-ray films. The band intensity of bound and unbound fraction was determined by using Image J 2.0 software. The fraction of DNA probe bound to the protein was plotted as a function of the protein concentration by using GraphPad Prism 5.
The Kd for the curve fitting of individual plots was determined as described before 36,38 .

Construction of deletion mutants and cell survival studies
The constructs for generating knockout of both the cis elements were made in pNOKOUT plasmid as described earlier 34 . In brief, ~500 bps upstream and ~500 bps downstream fragments to the cisII and cisMP were PCR amplified with the primers (Table S1) and cloned at KpnI & EcoRI, and BamHI & XbaI sites in pNOKOUT plasmid to yielded pNOKCII and pNOKCM plasmids, respectively (Table S2). These plasmids were linearized with XmnI and transformed into D. radiodurans cells. The homozygous replacement of cisII and cisMP by nptII was achieved through several rounds of sub-culturing and ascertained by diagnostic PCR using flanking primers (Table S1) of the target as described earlier (Maurya et al., 2018). The confirmed mutant cells were maintained in the absence of selection pressure but ensured for antibiotic resistance before the experiments.
The survival of Δ cisII and Δ cisMP deletion mutants was monitored under normal conditions and in response to different doses of gamma radiation as described earlier 39 . In brief, the cells were grown in the absence and presence of antibiotics and exposed to different doses of

Estimation of genome copy number
The genome copy number of each replicon in D. radiodurans was determined in exponentially growing cells. The density and cell numbers of independent sample were normalized for a fix OD 600 nm and by cfu.  (Table S1). The experiments were carried out with three independent biologic replicates for each sample. The replicon copy number is quantified by comparing the results with a standard plot. The copy number of each replicon as determined by considering the amplification of all genes was calculated using the cell number present at the time of cell lysis. Average of copy number reflected from all three genes per replicon was analysed using statistical analysis. The copy numbers of pNOKcisII and pNOKcisMP in D. radiodurans were determined using similar approach as described above except that the genes used were nptII (kanamycin resistance gene) and bla (ampicillin resistance gene) from plasmid backbone (Table S1).

Construction of tetO insertion derivatives of genome elements in D. radiodurans.
We used tetO operator cassette (an array of 240 repeats) from pLAU44 42  radiodurans and listed in Table S2.

Construction of TetR-GFP expression plasmid in D. radiodurans.
The reporter gene tetR was PCR amplified from pLAU53 42 using gene specific primers (Table S1) and cloned at SacI and SalI sites in pDSW208 43 to give pDTRGFP plasmid.
Further, we PCR amplified tetR-gfp fusion fragment from pDTRGFP plasmid using TetRscIApIFw and GFPXbaIRw primers and cloned in pRADgro 39 at ApaI-XbaI sites to give pRADTRGFP plasmid (Table S2). The constitutive expression of TetR-GFP from their respective plasmids in D. radiodurans was confirmed by western blotting and fluorescence microscopy.
We transformed pRADTRGFP plasmid into D. radiodurans R1 and its Δ cisII and Δ cisMP mutants (Table S2)  presented as scattered plot. One representative image belonging to more than 90 % population is presented separately for DIC, DAPI, Nile Red and GFP as well as merged.
Experiments were repeated independently to ensure the reproducibility and significance of these data.

cisII and cisMP elements confer ori function in D. radiodurans
Since, cisII and cisMP sequences contain typical DnaA boxes and are located upstream to parAB operons, a possibility of these elements acting as origin of replication and centromerelike elements in D. radiodurans was investigated. We used pNOKOUT a colE1 based E. coli plasmid for stability assay in D. radiodurans. For that, these elements were separately cloned in pNOKOUT and the resultant plasmids pNOKCisII and pNOKCisMP were transformed into D. radiodurans. Unlike pNOKOUT vector, the transformants harbouring pNOKCisII or pNOKCisMP were able to grow under kanamycin (6 µg/ml) selection both on TGY plates and in broth, which was found to be nearly similar to wild type grown without antibiotics ( Fig. 2A, C). In order to rule out that the expression of plasmid born antibiotic resistance was not due to integration of these plasmids into chromosome, the maintenance of these plasmids was checked in recombinant negative Δ recA mutant of D. radiodurans. The results were same as wild type and the Δ recA transformants of pNOKCisII or pNOKCisMP plasmids could grow in the presence of required dose of antibiotic (Fig. 2B). Further, the independent multiplication of these plasmids was confirmed independently by isolating them from kanamycin resistant recombinant D. radiodurans and restriction analysis. Results showed the release of expected size insert from recombinant derivatives of pNOKOUT plasmid (Fig.   2D). When we measured the copy number by qPCR, it was 6.86 ± 0.71 for pNOKCisII and 12.15 ± 0.92 for pNOKCisMP, which was found to be very close to in vivo copy number of Chr II and MP, respectively in D. radiodurans. These results together provided the clear evidence that pNOKCisII and pNOKCisMP plasmids were independently maintained in D.
radiodurans because of the presence of cisII and cisMP elements, as the vector without them could not survive, suggesting the 'ori' nature of cisII and cisMP elements for cognate replicons in this bacterium.

cisII and cisMP elements provide sequence specific interaction to DnaA
To understand the molecular mechanisms of cisII and cisMP roles as ori, the recombinant deinococcal replication initiator protein DnaA was purified (Fig. S1), and its DNA binding activity was checked with radiolabelled cisII and cisMP elements and their repeat derivatives ( Fig. S2). Results showed that DnaA establish the sequence specific interaction with both cisII and cisMP albeit with different affinity (Fig. 3) and its affinity varies with the number of these repeats. For instance, DnaA affinity to cisII was higher than cisMP. The affinity of DnaA to both the cis elements had reduced with the reduction of repeats' number (Fig. 3, Table 1). As expected, the DnaA bound to either of cis elements remained unaffected even in the presence of a 100-fold higher molar concentration of non-specific DNA, which could be titrated out completely with 10-fold molar excess of specific DNA (Fig. 3). The formation of DnaA complex with cisII and cisMP containing several directs repeats clearly suggests the 'ori' nature of both the cis elements in D. radiodurans.

cisII and cisMP elements contain sites for specific interaction of ParB proteins
Earlier, we had shown the sequence specific interaction of ParB2 and ParB3 with both cisII and cisMP albeit with different affinity 36 . Since, ParBs function in genome segregation as dimer, we wanted to check the minimum number of repeats required for their sequence specific interaction with cognate ParBs. The radiolabelled cisII and cisMP and their repeat variants were incubated with purified recombinant ParB2 and ParB3 proteins (Fig. S1) and products were resolved on native PAGE. The ParB2 and ParB3 showed higher affinity with full length cisII (Kd = 0.40 ± 0.007 µM) and cisMP (0.59 ± 0.05 µM), respectively (Fig. 4 A, E). The affinity of ParBs to cognate cis elements were nearly same upto 3 repeats and reduced significantly when the number of repeats become less than 3 (Fig. 4C, G; Table 1).
These proteins could rarely bind when number of repeats were less than 2 ( Fig. 4 D, H). The ParBs binding to cognate cis elements was sequence specific as the ParB-cis nucleoprotein complex was not affected in the presence of a 100-fold higher molar concentration of nonspecific DNA (Fig 4 A-C, E-G). Further, it was noticed that the full length cisII and cisMP bound ParBs migrated under electrophoresis with the peculiar patterns that normally seen when a tertiary or quaternary structure is introduced either naked DNA or by binding with proteins. Since, reduction in number of repeats has abolished this pattern, a possibility of fulllength cis elements producing structures upon protein binding cannot be ruled out.
Nonetheless, these findings suggested that both cisII and cisMP elements have specific sites for ParB nucleation, which might function as centromere -like sequence during segregation of cognate replicons.

cisII and cisMP determine cognate replicons copy number and their segregation
Since both cisII and cisMP have 'ori' signatures, their roles in the regulation of genome copy number and their segregation pattern were investigated. For that, the deletion mutants of cisII (ΔcisII) and cisMP (ΔcisMP) elements were made in D. radiodurans (Fig. S3) (Fig. 5 A). Deletion of cisMP has not affected the copy number of Chr I and Chr II when compared with wild type.
Such a reduction in ploidy of secondary genome elements after deletion of their cognate cis elements could be due to either defective replication and/or segregation of these replicons.
The possibility of ∆ cisII and ∆ cisMP mutants have defects in segregation was checked indirectly by monitoring the frequency of cells conferring antibiotic resistance (integrated in Chr II and MP) when grown in the absence and presence of selection pressure. In general, the growth pattern of cis mutants was different from wild type. Further, the growth of both the mutants had reduced when grew in the presence of antibiotics as compared to the absence of antibiotics (Fig. 5 B). Since, antibiotic is being expressed in place of cis elements in 3.6. cisII and cisMP mutants showed reduction in γ -radiation resistance.
Since, the deletion of cisII and cisMP in D. radiodurans has drastically reduced the copy number of respective replicons, a possibility of this if has affected the extraordinary resistance to DNA damage in this bacterium was hypothesized and tested. We monitored the survival of ∆ cisII and ∆ cisMP mutants under γ -radiation exposure and found that ∆ cisII mutant was in general more sensitive to gamma radiation than ∆ cisMP mutant. Further, the mutants that were maintained without antibiotics showed nearly 2-fold higher sensitivity to gamma radiation as compared to those grown with antibiotics ( Fig 5C). Since, the deletion of these cis elements had caused reduction in cognate replicon copy number, the loss of gamma radiation resistance could be due to either reduction in copy number or complete loss of secondary genome elements in these mutants. Earlier, it was shown that the reduction of copy number of Chr II and MP in △ parA2△parA3 double mutants has affected gamma radiation response in D. radiodurans 36 . These results suggested that deletion of cisII or cisMP from genome can make these cells defective in DNA replication and /or segregation of secondary genome elements like Chr II and MP, which might have contributed to the loss radioresistance in this bacterium.

cisII and cisMP deletions affected the maintenance of cognate genome elements
The multipartite genome system, ploidy and packaging of complete genetic materials into a doughnut shaped toroidal nucleiod are the remarkable cytogenetic feature in D. radiodurans [50][51][52]  repeats of tetO cassette (Table S1) in wild type and cis mutants background (Fig. S4) and TetR-GFP was expressed in trans ( Fig S5). These cells were grown exponentially and stained with Nile red and DAPI, and imaged microscopically in DIC, FITC, TRITC and DAPI channels. Interestingly, we could not find any change in dynamics and localisation of Chr I in Δ cisII and Δ cisMP mutant when compared with wild type (Fig. 6) (Fig. 7, Fig 8). Since, qPCR study has shown highest number (

Discussion
The radiodurans. The possibility of such type organization of 'ori' and centromere-like region is helpful in temporal regulation of replication and segregation in this bacterium. Functionally, the conversion of a non-replicative plasmid into replicative one by these cis elements (Fig. 2), reduction in copy number of cognate replicon without these elements (Fig. 5), and increased frequency of Chr II and MP anucleate cells in deletion mutant of cognate cis element (Fig. 7, Likewise, if these cis elements are believed to contain centromere-like sequences, then their deletion should show segregation defect in at least respective replicons. Our results supported the hypothesis and the mutants maintained without antibiotic selection had a heterogenous population of cells with and without cognate replicon as seen where nearly half of the population was sensitive to antibiotics ( Fig 5B). Presumably, these were the ones that survived with chromosome I and without secondary genome elements due to segregation defects. Additionally, 2-3 cells in tetrad were devoid of FROS foci in respective genome elements (Fig 7 & Fig 8). When these cis mutants were checked for gamma radiation response, the mutants maintained without selection pressure showed a greater sensitivity to gamma radiation as compared to those maintained under selection pressure. This further supported that the cells maintained without selection pressure were heterogenous with respect to secondary genome elements and those without secondary genome would have become sensitive to gamma radiation. Reduction in ploidy of secondary genome elements in D.
radiodurans was previously found to be associated with reduction in growth and sensitivity for abiotic stresses including γ -radiation 36 . These findings clearly suggest a regulated crosstalk between DNA replication and segregation in bacteria that have evolved a complex     Table 1.  Table 2.