Molecular insights into replication initiation in a multipartite genome harboring bacterium Deinococcus radiodurans

Deinococcus radiodurans harbors a multipartite ploid genome system consisting of two chromosomes and two plasmids present in multiple copies. How these discrete genome elements are maintained and inherited is not well understood. PprA, a pleiotropic protein involved in radioresistance, has been characterized for its roles in DNA repair, genome segregation, and cell division in this bacterium. Here, we show that PprA regulates ploidy of chromosome I and II and inhibits the activity of drDnaA, the initiator protein in D. radiodurans. We found that pprA deletion resulted in an increased genomic content and ploidy of both the chromosomal elements. Expression of PprA in trans rescued the phenotypes of the pprA mutant. To understand the molecular mechanism underlying these phenotypes, we characterized drDnaA and drDnaB. As expected for an initiator protein, recombinant drDnaA showed sequence-specific interactions with the putative oriC sequence in chromosome I (oriCI). Both drDnaA and drDnaB showed ATPase activity, also typical of initiator proteins, but only drDnaB exhibited 5′→3′ dsDNA helicase activity in vitro. drDnaA and drDnaB showed homotypic and heterotypic interactions with each other, which were perturbed by PprA. Interestingly, PprA has inhibited the ATPase activity of drDnaA but showed no effect on the activity of drDnaB. Regulation of chromosome copy number and inhibition of the initiator protein functions by PprA strongly suggest that it plays a role as a checkpoint regulator of the DNA replication initiation in D. radiodurans perhaps through its interaction with the replication initiation machinery.

Deinococcus radiodurans harbors a multipartite ploid genome system consisting of two chromosomes and two plasmids present in multiple copies. How these discrete genome elements are maintained and inherited is not well understood. PprA, a pleiotropic protein involved in radioresistance, has been characterized for its roles in DNA repair, genome segregation, and cell division in this bacterium. Here, we show that PprA regulates ploidy of chromosome I and II and inhibits the activity of drDnaA, the initiator protein in D. radiodurans. We found that pprA deletion resulted in an increased genomic content and ploidy of both the chromosomal elements. Expression of PprA in trans rescued the phenotypes of the pprA mutant. To understand the molecular mechanism underlying these phenotypes, we characterized drDnaA and drDnaB. As expected for an initiator protein, recombinant drDnaA showed sequence-specific interactions with the putative oriC sequence in chromosome I (oriCI). Both drDnaA and drDnaB showed ATPase activity, also typical of initiator proteins, but only drDnaB exhibited 5 0 →3 0 dsDNA helicase activity in vitro. drDnaA and drDnaB showed homotypic and heterotypic interactions with each other, which were perturbed by PprA. Interestingly, PprA has inhibited the ATPase activity of drDnaA but showed no effect on the activity of drDnaB. Regulation of chromosome copy number and inhibition of the initiator protein functions by PprA strongly suggest that it plays a role as a checkpoint regulator of the DNA replication initiation in D. radiodurans perhaps through its interaction with the replication initiation machinery.
The origin of replication in bacterial chromosome (oriC) is a discrete locus that contains AT-rich conserved DNA motifs and a varying number of 9 mer repeats of nonpalindromic sequences called DnaA boxes. These boxes are recognized by a replication initiator protein named DnaA, followed by the assembly of replication initiation complex at oriC (1,2). Mechanisms underlying replication initiation have been characterized in large number of bacteria harboring limited copies of single circular chromosome as inheritable genetic material (3,4). In Escherichia coli, it has been shown that DnaA-ATP oligomer binds to DnaA boxes at oriC and unwinds the adjacent AT-rich region. Subsequently, a hexameric complex of replicative helicase DnaB and its loader DnaC (DnaB 6 -DnaC 6 ) is recruited to the unwound region in oriC resulting in the formation of the prepriming complex (5,6). This provides the site for binding of primase and activation of various events required for the progression of the replication complex that includes DNA polymerase III holoenzyme. DnaB hexameric ring translocates bidirectionally to unwind the parental duplex DNA while the 3 0 end of the primer is extended by DNA polymerase complex (7,8). In E. coli, the initiation of DNA replication at oriC site is tightly regulated and the next round of oriC-mediated DNA replication must have to wait till the oriC of newly replicated daughter chromosomes is fully methylated. Therefore, under normal growth conditions, the number of copies of the primary chromosome per cell is expected to be less than 2 (5). Recently, the bacteria with multiple copies of the multipartite genome system have been reported. Notably, most of them are either parasites to some forms of life or exhibit super tolerance to abiotic stresses (9). The ploidy of chromosomes in these bacteria allowed us to revisit the mechanism of oriC regulation as known in bacteria containing less than two copies of circular chromosome per cell. Mechanisms underlying the regulation of oriC function in multipartite genome harboring bacteria have not been studied in detail. In the case of Vibrio cholerae, which harbors two chromosomes, namely chromosome I (Chr I) and chromosome II (Chr II), the Chr I replicates similar to the E. coli chromosome while Chr II follows a replication mechanism that is akin to low copy number P1 and F plasmids (10)(11)(12)(13)(14)(15).
Deinococcus radiodurans is characterized for its extraordinary resistance to the lethal doses of DNA-damaging agents including radiation and desiccation (16)(17)(18)(19). It harbors a multipartite genome system comprised of two chromosomes (Chr I (2,648,638 bp) and Chr II (412,348 bp)) and a megaplasmid (177,466 bp) and a plasmid (45,704 bp) (20). Interestingly, each genome element is present in multiple copies per cell (21). Chr I of D. radiodurans encodes the putative DnaA (DR_0002) and DnaB (DR_0549) (hereafter named drDnaA and drDnaB, respectively) while chromosome II encodes PprA (DR_A0346), which has been characterized for various functions (22)(23)(24)(25). Recently, the extended structure of PprA has been reported where a possibility of it acting as a protein scaffold has been suggested (26). Here, for the first time, we report the functional characterization of chromosome replication initiation proteins drDnaA and drDnaB in D. radiodurans and demonstrate that PprA plays an important role in regulation of DNA replication. The pprA mutant showed an increased copy number of chromosome I (ChrI) and chromosome II (ChrII), which was complemented by the in trans expression of the wild-type PprA. Interestingly, there was no effect of PprA over expression on copy number of the different genome elements in the wild-type cells. drDnaA was characterized as a sequence-specific origin of replication (oriCI) binding protein and an oriCI responsive ATPase. drDnaB was found to be an ATP-dependent 5 0 →3 0 dsDNA helicase and showed higher affinity for single-stranded DNA (ssDNA) than double-stranded DNA (dsDNA). Interestingly, PprA interacted with drDnaA at a relatively higher affinity than drDnaB and inhibited both homotypic and heterotypic interactions of these proteins. Further, PprA downregulated the ATPase activity of drDnaA but showed no effect on ATPase and helicase activities of drDnaB. These results suggest that drDnaA and drDnaB carry out the necessary functions required for initiation of replication at oriCI in D. radiodurans. Furthermore, the interference imposed by PprA in the physicochemical properties of these replication proteins as well as an increase in the copy numbers of both primary and secondary chromosomes in its absence together suggested the involvement of PprA in regulation of chromosomal replication in this bacterium.

Results
The pprA deletion affects genomic content in D. Radiodurans Earlier, the regulatory role of PprA in cell division and genome maintenance has been demonstrated (23,24,27,28). In this study, the DNA content and the copy number of genome elements in pprA deletion mutant were compared with the wild-type D. radiodurans. The total DNA content in mutant cells (8.41 ± 1.05 fg per cell) was found to be approximately threefold higher than wild-type cells (3.05 ± 0.47 fg per cell) (Fig. 1A). Similarly, the DAPI fluorescence in ΔpprA mutant was approximately twofold higher than that of wild type (Fig. 1B). The cell scan analysis of DAPI stained cells showed that 80% of ΔpprA cells have approximately twofold higher DAPI fluorescence as compared with wild-type cells grown identically (Fig. 1C). When we checked the copy number of each replicon per cell, we found that the average copy number of Chr I and Chr II was 2.5to 3-fold higher in ΔpprA mutant as compared with the wild type. For instance, the average copy number of Chr I was 8 per cell in wild type while it was 18 per cell in the ΔpprA mutant. The average Figure 1. Change in genomic content in pprA deletion mutant of D. radiodurans. D. radiodurans R1 (WT) and its pprA mutant (ΔpprA) were grown to mid-logarithmic phase. The amount of DNA per cell (fg; femtograms) was measured spectrophotometrically (A) and through DAPI staining of nucleoid (B). The changes in DAPI fluorescence in wild type and ΔpprA mutant cells were scanned in 150 cells and relative frequency of cells containing different amounts of DAPI stained nucleoid was estimated (C). Similarly, wild type (WT) and mutant (ΔpprA) populations were expressed with PprA in trans (WT + PprA, ΔpprA + PprA) and the copy number of chromosome I (ChrI), chromosome II (Chr II), megaplasmid (Mp), and small plasmid (Sp) was determined (D). Data presented in A, B, and D are mean ± SD (n = 9). Data given in C are from the wild-type cell population (n = 148) and mutant population (n = 159). Data was analyzed by Student's t-test and p values less than 0.5 and 0.001 were denoted as (*) and (***), respectively. copy number of Chr II also increased from 7 in wild type to 17 per cell in the ΔpprA mutant (Fig. 1D). Further we checked the effect of in trans expression of PprA on the copy number of these genome elements in wild type and ΔpprA mutant. We observed nearly no effect of PprA over expression in wild type, whereas the ΔpprA mutant expressing PprA in trans could restore the wild-type copies of the genome element (Fig. 1D). Interestingly, the copy number of megaplasmid and small plasmid did not change under any of these conditions. These results indicated a possible role of PprA in the maintenance of the chromosome copy number possibly by regulating the replication of DNA in D. radiodurans.

PprA interacts with drDnaA and drDnaB
To obtain the mechanistic insights into the regulatory role of PprA in DNA replication, the physical and functional interaction of PprA with drDnaA or drDnaB proteins was monitored using bacterial two-hybrid system. For this, T18tagged PprA and T25-tagged drDnaA or drDnaB were coexpressed in E. coli BTH101 (cyaA − ). Interaction of two target proteins tagged with T18 and T25 domains of CyaA would reconstitute the active CyaA resulting in induction of transcription from β-galactosidase gene in BTH101. Expression of β-galactosidase as an indication of protein-protein interaction was monitored using spot test. PprA showed interaction with both drDnaA and drDnaB as indicated by a blue color in the spot test ( Fig. 2A). The interaction of PprA with these proteins was further checked using surface plasmon resonance (SPR) using purified recombinant drDnaA or drDnaB (Fig. S1). Results showed a concentration-dependent increase in the SPR signals for both drDnaA and drDnaB (Fig. 2, B and C). The dissociation constant (Kd) for drDnaA was 5.41 × 10 −7 ± 1.8 X 10 −8 [M] while it was 9.71 × 10 −7 ± 1.1 × 10 −8 [M] for drDnaB indicating that PprA interacts with drDnaA with approximately twofold higher affinity than with drDnaB. Interaction of C-terminal deletion (CTD) mutant of drDnaA (DnaAΔCt) with PprA was tested in surrogate E. coli using coimmunoprecipitation. Results showed that the deletion of CTD did not affect drDnaA interaction with PprA ( Fig. 2D) suggesting that drDnaA might interact with PprA through its N-terminal domain and/or the middle region of the protein.
In vivo interaction of PprA with drDnaA and dDnaB was tested using co-immunoprecipitation assays with the cellular protein extracts from D. radiodurans cells transformed with T18tagged drDnaA or drDnaB expressing plasmid. Endogenous PprA was precipitated using anti-PprA antibodies and the PprA regulates replication initiation through DnaA and DnaB presence of T18-tagged drDnaA/drDnaB was probed using immunoblotting with T18 antibodies. Consistent with the above results, both drDnaA and drDnaB had coimmunoprecipitated with PprA in vivo (Fig. 2E). These results together suggested that PprA interacts with both drDnaA and drDnaB of D. radiodurans.
drDnaA is a sequence-specific oriCI binding protein The putative origin of replication of chromosome I (oriCI) in D. radiodurans R1, spanning 1183 to 1903 bp upstream of drdnaA gene (DR_0002), was predicted using the DOriC database (DoriC accession number -ORI10010007) (29). Sequence analysis using WebLogo online tool revealed the presence of consensus sequences of 13 copies of 9-mer DnaAboxes (30) between 1273 and 1772 bp (500 bp) upstream of the drdnaA gene. A 46.2% GC content of oriCI indicated an AT-rich sequence, typical of canonical oriC sequences. The structure of oriCI was found to be different as compared with the E. coli oriC. It is comparatively longer than the E. coli oriC and contains only eight out of 13 perfect E. coli like DnaA boxes (TTATCCACA). Five out of 13 DnaA boxes were imperfect with single nucleotide difference from the E. coli type boxes and were similar to other bacteria such as Cyanothece 51142, Thermus thermophilus, and Bacillus subtilis ( Fig. S2) (31)(32)(33). A web logo of these DnaA boxes was created yielding a consensus sequence of T(A/G)TA(T) TCCACA. These DnaA boxes are distributed randomly on both sense (four DnaA boxes) and antisense (nine DnaA boxes) strands of chromosome I. This suggested that oriCI in chromosome I of D. radiodurans is largely similar to oriC of E.coli with some deviations. Whether these changes have any functional significance, in the context of regulation of replication initiation, needs to be explored.
Binding of recombinant drDnaA (Fig. S1A) to [ 32 P] end labeled oriCI containing 13 repeats of DnaA-boxes was checked by electrophoretic mobility shift assays (EMSA). The recombinant drDnaA showed sequence-specific interaction with oriCI as the binding did not change even in the presence of 50-fold excess of cold nonspecific DNA (Fig. 3, A and B). Interestingly, the binding affinity of drDnaA for oriCI (Kd = 1.68 ± 0.23 μM) was increased by approximately eightfold in presence of ATP (Kd = 0.243 ± 0.02 μM) (Fig. 3, C and D). Further, binding of [ 32 P] end labeled nonspecific DNA with increasing concentrations of drDnaA was displaced by addition of higher molar concentration of cold oriCI DNA confirming the sequence-specific interaction of drDnaA to oriCI ( Fig. 3E). In addition, the binding affinity of drDnaA for oriCI decreased upon reducing the number of DnaA boxes (Fig. S3) in the presence of ATP (Table 1). The effect of histidine tag on DNA-binding activity drDnaA was ruled out as both (his)6 tagged and non-(his)6 tagged drDNA showed nearly similar Kd for dsDNA (Fig. S4). Similar to E. coli DnaA, which shares 44.4% identity and 72% similarity with drDnaA (data not shown), drDnaA showed specific binding even for the oriCI sequence with only a single DnaA box. However, the affinity of drDnaA for a perfect DnaA box (TTATCCACA) was approximately twofold higher (Kd = 2.88 ± 0.28 μM) than the affinity (Kd = 4.21 ± 0.15 μM or 4.36 ± 0.18 μM) for imperfect DnaA boxes (TTTTCCACA or GTATCCACA) ( Table 1). These results suggested that drDnaA binds to oriCI in a sequence-specific manner and this interaction is stimulated by ATP.

C-terminal domain of drDnaA regulates oriCI stimulation of its ATPase activity
Since, drDnaA showed a relatively higher affinity for oriCI in the presence of ATP (Fig. 3C), ATP hydrolysis by drDnaA in the presence and absence of oriCI was tested. Results showed that drDnaA hydrolyzes [ 32 P] αATP into [ 32 P] αADP (Fig. 4A). Interestingly, the ATP hydrolysis by drDnaA is stimulated in the presence of oriCI (Fig. 4, B and C) but not in the presence of nonspecific dsDNA (Fig. S5). The helix-turn-helix motif in domain IV at the C terminal of DnaA is known to confer sequence specificity for DnaA-oriC interaction in many bacteria (34)(35)(36)(37)(38)(39). Therefore, the binding of drDnaA having a deleted C terminal (domain IV) (DnaAΔCt) ( Further, the effect of CTD deletion on rate of ATP hydrolysis was calculated using both drDnaA and DnaAΔCt proteins at 30 nM [ 32 P] αATP. The rate of ATP hydrolysis in drDnaA (0.76 ± 0.04 nM/min) was found to be nearly similar to DnaAΔCt (0.81 ± 0.05 nM/min). These results suggested that the C-terminal domain IV of drDnaA is essential for its binding to oriCI and thus for oriCI-dependent stimulation of ATPase function but seems to have no role in the ATPase activity of drDnaA per se.
drDnaB is an ssDNA-responsive ATPase and an ATPdependent 5 0 →3 0 dsDNA helicase Binding of recombinant drDnaB to the oriCI was checked by EMSA. drDnaB showed nonspecific binding to oriCI as chase with nonspecific dsDNA significantly competed out oriCI bound to drDnaB. The affinity of drDnaB for oriCI (Kd = 8.89 ± 0.92 μM) did not change in the presence of ATP (Kd = 8.23 ± 0.41 μM). Interestingly, drDnaB showed approximately threefold higher affinity to ssDNA substrate (Kd = 3.11 ± 0.11 μM) as compared with dsDNA (oriCI) (Kd = 8.89 ± 0.92 μM). Further, binding of drDnaB to ssDNA increased by nearly twofold in the presence of ATP (Kd = 1.42 ± 0.05 μM) (Fig. 6) and approximately threefold in the presence of ATPγS (1.02 ± 0.06 μM) (Fig. S6). Addition of ATP increased drDnaB affinity for ssDNA but not for dsDNA. These results suggested that drDnaB preferentially binds to ssDNA over dsDNA both in the presence and absence of ATP. Similar to drDnaA, drDnaB could hydrolyze [ 32 P]-αATP to [ 32 P]-αADP (Fig. 7A). The stimulatory effect of ssDNA on ATP hydrolysis was observed albeit at a very low level and only at the lower ATP to protein ratios (Fig. 7, B and C). Similar observations were reported earlier for other DnaB homologs (40,41). These results suggested that drDnaB is an ssDNA-binding ATPase and its binding to ssDNA is affected by ATP hydrolysis in vitro.
DnaBs are characterized as replicative helicases and play an essential role in oriC-mediated DNA replication in other bacteria (42)(43)(44). Therefore, the helicase activity of recombinant drDnaB was checked on dsDNA having either 3 0 or 5 0 overhangs and blunt end as a substrate in the presence and absence of ATP. The subsequent products were monitored on the gel as well as using a Fluorescence Resonance Energy Transfer (FRET)-based assay. Results showed that drDnaB is able to unwind only dsDNA with 5 0 overhang, not with 3 0 overhang (Compare Fig. 8A with Fig. 8B) or blunt end substrate (Fig. S7). This 5 0 → 3 0 dsDNA helicase activity required the hydrolyzable ATP as no DNA unwinding was observed when ATP was substituted with ATPγS ( Fig. 8C). Similar results were obtained in FRET assays with FAM as a donor and BHQ as an acceptor moiety (Fig. 8D). Results showed an increased fluorescence in the presence of drDnaB and ATP, which was lost when ATPγS was substituted for ATP (Fig. 8, E and F). Interestingly, drDnaB showed binding to blunt end dsDNA substrate, even though no helicase activity was observed for the same. These results suggested that drDnaB is an ATP-dependent 5 0 → 3 0 dsDNA helicase, which is consistent with the observed characteristics of known DnaB helicases from other bacteria (42,45,46).

drDnaA and drDnaB show homotypic and heterotypic interactions
The possible interaction between drDnaA and drDnaB proteins was monitored using bacterial two-hybrid (BACTH) system and co-immunoprecipitation. Both N-terminal and C- Recombinant purified drDnaA was incubated with radiolabeled repeat variants of linear oriCI in the presence (+) and absence (−) of ATP and EMSA was carried out, and autoradiograms were developed as shown in Figure 4. Fractions of DNA bound to protein were estimated densitometrically and plotted as a function of protein concentration. The Kd for the curve fitting of individual plots was determined using GraphPad Prism6 software.
PprA regulates replication initiation through DnaA and DnaB terminal fusions of drDnaA and drDnaB with T18 and T25 tags of BACTH were coexpressed in E. coli BTH101 (cyaA − ) in different combinations (Fig. S8). The expression of β-galactosidase upon reconstitution of an active CyaA from its T18 and T25 domains that would indicate that the interactions of target proteins were monitored. The drDnaA and drDnaB proteins tagged with T18 and T25 through their C terminals showed both homotypic and heterotypic interactions in E. coli (Fig. 9A). Some of these interactions were further confirmed by immunoprecipitation and results agreed with the BACTH analysis in E. coli (Fig. 9, B and C). In vivo interaction of these proteins was monitored in D. radiodurans. For this the whole cell extracts from the cells coexpressing drDnaA or drDnaB with T18 tag and/or polyhistidine tag on the respective plasmids were immunoprecipitated using polyhistidine antibodies, and the potential interacting partner was detected using anti-T18 antibodies. Results confirmed both homotypic and heterotypic interactions of these proteins in vivo (Fig. 9D). Further, the DnaAΔCt protein also showed interaction with full-length drDnaA and drDnaB (Fig. 9E) indicating that CTD is not required for the interaction. Since, C-terminal domain of drDnaA is required for oriCI binding but not for both homotypic and heterotypic interaction with drDnaA and drDnaB, the interaction of these proteins seems to be  independent of drDnaA interaction with oriCI. These results suggested that both these proteins interact through their Nterminal domains as reported earlier for the known homologs (1,(47)(48)(49). Further studies would require for the exact mapping of the interaction domains or residues of these proteins.

PprA modulates in vitro function of drDnaA but not drDnaB
The ATP hydrolysis is the key feature of bacterial DnaA and DnaB proteins, and this activity is required for the initiation and progression of oriC mediated replication. The recombinant drDnaA and drDnaB showed ATPase activity in vitro (Figs. 4 and   7). The effect of histidine-tagged PprA (His-PprA) on the ATPase activity of both drDnaA and drDnaB was checked in vitro. The ATPase activity of drDnaA was significantly reduced in the presence of an equimolar concentration of His-PprA irrespective of the order of addition of these proteins in the reaction mixture (Fig. 10, A and B). There was no effect of His-PprA on either ATPase activity (Fig. 10, C and D) or 5 0 →3 0 dsDNA helicase activity (Fig. S9) of drDnaB in vitro. His-PprA did not affect drDnaB function in vitro and also did not hydrolyze ATP into ADP by itself. Therefore, the possibility of ATP degradation or sequestration by His-PprA affecting the ATPase activity of drDnaA was ruled out. The effect of histidine tag on PprA function was further evaluated. In trans expression of His-PprA showed nearly complete complementation to the loss of gamma radioresistance in pprA mutant (Fig. S10A). These results suggested that PprA affects drDnaA functions. Since the ATPase activity of DnaA is required for oriC-dependent replication initiation in bacteria (1,6), the suppression of ATPase activity of drDnaA by PprA may be an important step in oriCI regulation in D. radiodurans.

PprA negatively regulates drDnaA and drDnaB interaction in surrogate E. coli
Since PprA showed interaction with drDnaA or drDnaB with different affinities, the possibility of PprA affecting drDnaA and drDnaB interactions was tested using bacterial two-hybrid system. The E. coli BTH101 cells coexpressing drDnaA and drDnaB in different combination were transformed with PprA expressing plasmid (pSpecpprA) (50) and the expression of β-galactosidase activity was monitored. We observed that the expression of β-galactosidase activity originating from the established interaction of drDnaA and drDnaB ( Fig. 9) was reduced in the presence of PprA as compared with the control with no PprA (Fig. 11, A-C). This effect of PprA was observed on homotypic and heterotypic interactions of both the replication proteins (Fig. 11D). These results suggested that PprA interferes with the oligomerization of drDnaA and drDnaB proteins of D. radiodurans.

Discussion
Molecular basis of regulation of DNA replication initiation has been extensively studied in bacteria with limited copies of the single circular chromosome. A significant proportion of bacteria harbor multiple genome elements, which are split into two or more chromosomes. Though many of the important pathogenic and plant symbiotic bacteria (e.g., Burkholderia, Brucella, Vibrio, Shinorhizobium, etc.) and D. radiodurans harbor multipartite genomes, most of our current knowledge on the DNA replication of multipartite genomes has come from the study on V. Cholerae. In V. cholerae, Chr I replication is initiated by DnaA protein while Chr II replication is initiated by RctB protein that binds and unwinds an array of repeats present in oriCII of Chr II (10,12,13,51). Since Chr I is larger than Chr II in V. cholerae, the temporal regulation of replication initiation could help in synchronization of replication termination of both the chromosomes before the onset of cell division (52)(53)(54).
D. radiodurans harbors a polyploid multipartite genome comprising two chromosomes and two plasmids (20). Molecular mechanism responsible for control of DNA replication initiation in this bacterium has not been reported yet. Homology search for the known replication initiator proteins revealed the presence of dnaA and dnaB genes, encoding for putative replication initiator proteins drDnaA and drDnaB, on the chromosome I of this bacterium. However, there was no homology with the known replication initiation proteins encoded on secondary genome elements (e.g., RctB and RepA) of other bacteria, in the genome of this bacterium (data not shown). PprA (DR_A0346), a pleiotropic protein of D. radiodurans, has been earlier characterized for its role in radioresistance. Here, we have brought forth some evidence to suggest that PprA regulates ploidy of Chromosome I and II and inhibits the activity of DnaA, the initiator protein, by direct interaction in D. radiodurans. In silico sequence analysis identified putative oriC-like sequence in chromosome I (oriCI) comprising two AT-rich clusters (12 bp and 20 bp), two GATC regions, and 13 AT-rich DnaA boxes with consensus sequence similar to that of E. coli (eight perfect DnaA boxes) or Cyanothece 51142, T. thermophilus, and B. subtilis (5 imperfect DnaA boxes) (Fig. S1). The oriC1 was located between drdnaA and drdnaN genes.
Characterization of the purified recombinant proteins using in vitro studies showed that both drDnaA and drDnaB have ATPase activity. Similar to other bacteria (32,(55)(56)(57)(58), drDnaA showed sequence-specific interaction with oriCI, specifically with a higher affinity for the E. coli-like DnaA boxes as well as in the presence of ATP suggesting its probable role as the replication initiator protein. drDnaB binds DNA in a nonspecific manner with a preference of ssDNA over dsDNA. The ATPase activity of drDnaA and drDnaB was stimulated by oriCI and ssDNA, respectively. The DnaBs or its homologs from different bacteria have shown different polarities in dsDNA helicase functions (59, 60). Similar to DnaB helicases from E. coli (42), Helicobacter pylori (46), Mycobacterium tuberculosis (41), and in Bacillus anthracis (60), drDnaB showed ATP dependent 5 0 → 3 0 helicase activity on dsDNA with 5'-overhangs.
The oriC-mediated initiation of DNA replication requires a minimum of DnaA binding at oriC site, the recruitment of (DnaB) 6 -(DnaC) 6 hexameric replicative helicase, DNA unwinding by helicase function, and energy-dependent elongation of replication fork (6,43,61). We observed that drDnaA and drDnaB undergo both homotypic and heterotypic oligomerization and interact with oriCI region and ssDNA, respectively. It is known that energy from the ATP hydrolysis by DnaA is used for melting of the AT-rich region in oriC necessary for the initiation of replication (1,6). The presence of ATPase activity in drDnaA and drDnaB proteins and its stimulation by oriCI and ssDNA, respectively, as well as ATPdependent dsDNA helicase function in drDnaB are therefore some of the typical characteristics known for the similar proteins in other bacteria. Altogether, above findings suggest that drDnaA and drDnaB protein confer the desired function to   demonstrated an increase in the genomic content and ploidy levels in D. radiodurans cells deficient in PprA indicating a relaxed replication initiation control as compared with the wild-type cells (Fig. 1). We next asked if PprA modulates the activity of the initiation proteins. Our results showed that PprA inhibited the homotypic and heterotypic macromolecular interactions between drDnaA and drDnaB proteins. In addition, PprA specifically inhibited the ATPase activity of drDnaA while having no effect on the ATPase and dsDNA helicase activity of drDnaB. These results suggest that PprA might regulate the replication initiation by inhibiting drDnaA function and oligomerization of drDnaA and drDnaB. Nonetheless, a direct effect on the replication initiation in PprA mutant remains to be seen and a conditional knockout of the genes encoding these proteins will confirm these functions in vivo.
Regulation of DNA replication by a cell cycle regulatory protein competing for binding to oriC site was suggested earlier (62); however, absence of the specific interaction of PprA with oriCI has ruled out this possibility (Fig. S10, B and  C). Interestingly, treatment with Hydroxy Urea (HU), a replication inhibitor, improved post irradiation recovery (PIR) of wild-type D. radiodurans (Fig. S11). It may be possible that the delayed PIR in the absence of PprA might be due to the loss of replication initiation control, which is otherwise necessary for rapid PIR in D. radiodurans. Different mechanisms have been reported for the regulation of DNA replication in different bacteria. In E. coli the initiation of DNA replication is regulated by the differential affinity of DnaA and SeqA proteins to methylated and hemimethylated oriC, respectively, and by the regulated inactivation of DnaA by Hda (5). Further, CtrA involvement in the regulation of DNA replication in Caulobacter crescentus (63) and CrtS protein interaction with methylated DNA in the case of V. cholerae (15) have been reported. Involvement of genome segregation proteins such as Soj and Spo0J in case of B. subtilis (64)(65)(66) and ParA and ParB in the regulation of secondary genome copy number in D. radiodurans (67,68) have also been suggested.
In conclusion, we have characterized oriCI, drDnaA, and drDnaB, the components involved in oriC-mediated initiation of bacterial chromosome replication in multipartite genome harboring radioresistant bacterium D. radiodurans. PprA, a protein with pleiotropic functions in this bacterium, was found to regulate genome ploidy and the physicochemical properties of drDnaA and drDnaB necessary for their functions in chromosome replication. Further studies would be required to address if PprA acts as a checkpoint regulator of oriC-mediated DNA replication under normal conditions or during DNA double-strand break (DSB) repair. Also, if the phosphorylation status of PprA (50) has any role in these functions would be worth investigating independently.

Bacterial strains and plasmids
All the bacterial strains, plasmids, and oligonucleotides used in this study are given in Tables S1 and S2, respectively. D. radiodurans R1 (ATCC13939) was grown in TGY (tryptone (1%), glucose (0.1%), and yeast extract (0.5%)) medium at 32 C (69). E. coli strain NovaBlue was used for cloning and maintenance of all the plasmids while E. coli strain BL21(DE3) pLysS was used for the expression of recombinant proteins. E. coli strain BTH 101 was used for Bacterial Two-Hybrid System (BACTH)-based protein-protein interaction studies. Standard protocols for all recombinant techniques were used as described in (70). Molecular biology-grade chemicals and enzymes were procured from Merck Inc and New England Biolabs. Radiolabeled nucleotides were purchased from the Department of Atomic Energy-Board of Radiation and Isotope Technology (DAE-BRIT), India.

Cloning, expression, and purification of proteins
The drDnaA was expressed from pETDnaA as described in (71). The drdnaB (DR_0549) gene was PCR amplified from the genomic DNA of D. radiodurans R1 using pETdnaBF and pETdnaBR primers (Table S2). The PCR product was cloned in pET28a(+) plasmid at BamHI and EcoRI sites to yield pETDnaB plasmid (Table S1). The C-terminal (domain IV) truncation of drDnaA was made using pETDAΔCtR primer along with pETdnaAF (Table S2), and the resulting plasmid was named pETDAΔCt (Table S1). The recombinant plasmids were sequenced, and the correctness of inserts was ascertained. The recombinant proteins were purified by nickel affinity chromatography as described earlier (65). Fractions showing more than 95% purity were pooled and dialyzed in buffer A containing 200 mM NaCl and further purified from HiTrap Heparin HP affinity columns (GE Healthcare Life sciences) using a linear gradient of NaCl. Different fractions were analyzed on SDS-PAGE and fractions containing desired protein were pooled and precipitated with 30% w/v ammonium sulphate at 8 C. The precipitate was dissolved in Rbuffer (20 mM Tris-HCl pH 7.6, 0.1 mM EDTA, 0.5 mM DTT) containing 1 M NaCl. After centrifugation at 16,000 rpm for 30 min, the supernatant containing soluble proteins was processed for gel filtration chromatography.The purified protein was dialyzed in dialysis buffer containing 20 mM Tris-HCl pH 7.6, 200 mM NaCl, 50% glycerol, 1 mM MgCl 2, 0.5 mM DTT, and 1 mM PMSF. Similarly, PprA was expressed from pETp-prA and purified as described in (72) followed by gel filtration as described above. The correct refolding of all purified proteins was confirmed by circular dichroism (CD) spectroscopy in phosphate buffer as described previously (67) (Fig. S1, B, D,  F and H).

Protein-DNA interaction studies
The interaction of recombinant proteins with different forms of DNA was studied using EMSA as described earlier (67,71). In brief, the 500 bp oriCI region (1272-1771) containing 13 repeats of 9 mer of DnaA boxes (T(A/G)TA(T) TCCACA) was PCR amplified using OriIFw and OriIRw primers (Table S1 and Fig. S2) and gel purified. In addition, different fragments of oriCI with 11, 7, 3, and 1 repeat(s) were also either PCR amplified or chemically synthesized and annealed (Fig. S3). DNA substrates were labeled with [γ- 32 Table S1) and further incubated as per experimental requirements. Similarly, oligonucleotide 99F (73) was radiolabeled at 5' end and used as an ssDNA substrate for interaction with drDnaB. The interaction of drDnaB with oriCI (dsDNA) was monitored in the absence and presence of 1 mM ATP as described for drDnaA above. The reaction mixtures were separated on 6 to 8% native PAGE gels, the gels were dried, and autoradiograms were developed on X-ray films. The band intensity of unbound and bound fraction was quantified densitometrically and computed using Image J 2.0 software. The fraction of DNA bound to the protein was plotted against the protein concentrations by using GraphPad Prism6, and the Kd values for the curve fitting of individual plots were determined as described in (67).

DNA helicase activity assay
The unwinding of dsDNA substrates (with 5' overhang as well as 3' overhang) by drDnaB (2 μM) in the presence of different combinations of 1 mM ATP and 2 μM PprA was monitored using FRET approach in 96-well plates as described in (74). In brief, the substrates used were FAM and BHQ labeled complementary oligonucleotides (Table S1). For the preparation of overhang dsDNA, the oligonucleotides of complementary strands producing 5'overhang or 3' overhang were mixed, heated at 95 C for 10 min, and then slowly annealed by switching off the heating block for overnight. The helicase assays were performed in volume of 100 μl in helicase buffer (25 mM Tris pH 7.6, 10 mM NaCl, 25 mM KCl, 1 mM MgCl 2, 100 μM DTT, 100 μg/ml BSA, and 2% glycerol) at 37 C. To avoid reannealing of unwound DNA, excess of DNA trap complementary to the BHQ labeled DNA was added to each reaction mixture. The reaction was excited at 490 nm and emission was recorded at 520 nm. The fluorescence signals were monitored at an interval of 30 s using a microplate reader (Biotek Synergy H1).
Similarly, gel-based DNA helicase activity assay of drDnaB was carried out as described in (74). In brief, the 3 0 overhang substrate was made by annealing the [ 32 P] radiolabeled 3OV99 F (5 0 TTTTTGCGGTTCATATGGAATTCC3 0 ) with 99F (5 0 GGAATTCCAATGAACCGCAAAACCGCAAAAAC CGTACCGA3 0 ) as described in (74). Similarly, a 5 0 overhang substrate was generated by annealing of radiolabeled 5OV99F (5 0 TCGGTACGGTTTTTGCGGTTCATA3 0 ) with 99F oligonucleotides. The annealed substrates were incubated with 2 μM drDnaB with or without 2 μM PprA proteins in the presence and absence of 1 mM ATP or ATP-γS in helicase assay buffer at 37 C. Aliquots were taken at different time intervals and reactions were stopped with stop solution (50 mM KCl, 20% glycerol, 20 mM EDTA, 125 μg Proteinase K, 0.2% SDS) at 37 C for 10 min. The products were separated on 8% denaturing PAGE, dried, and exposed to X-ray films. The autoradiograms were developed and documented.
Bacterial two-hybrid (BACTH) system assay drDnaA and drDnaB interactions were monitored in the absence and presence of PprA using a BACTH, described in (75)(76)(77)(78). In brief, the coding sequence of DR_0002 (drDnaA) was cloned at BamHI and EcoRI sites in pUT18, pUT18C, and pKT25 plasmids to yield pUT18DA, pUT18CDA, and pKTDA, respectively. Similarly, coding sequences of DR_0549 (drDnaB) were cloned at KpnI and EcoRI sites in pUT18, pUT18C, and pKT25 plasmids to yield pUT18DB, pUT18CDB, and pKTDB, respectively. The construction of pKNTDA and pKNTDB plasmids has been described in (68) and pUTpprA in (23). E. coli strain BTH101 (cyaA − ) was cotransformed with these plasmids in different combinations and expression of βgalactosidase was monitored as described earlier (76). E. coli BTH101 harboring only vectors were used as negative control while those harboring pUTEFA and pKNTEFZ (79) were used as a positive control. For monitoring the effect of PprA on drDnaA and drDnaB interactions, the PprA was expressed on pSpecpprA (50), whereas the p11559 vector was used for negative control. Expression of β-galactosidase as an indication of protein-protein interaction was monitored using spot assay and in liquid culture, and β-galactosidase activity was calculated in Miller units as described in (76).

Co-immunoprecipitation assay
Protein-protein interactions in surrogate E. coli (BTH101) and D. radiodurans were monitored by coimmunoprecipitation as described in (76,77). In brief, the total proteins of the recombinant BTH101 cells coexpressing drDnaA and drDnaB proteins on BACTH plasmids (Table S1) were immunoprecipitated using polyclonal antibodies against T25 and counterpart was detected using T18 monoclonal antibodies raised in the mouse. PprA's effect on the interactions among drDnaA and drDnaB was studied by in trans expression of PprA from pSpecpprA plasmid. Signals were detected using anti-mouse secondary antibodies conjugated with alkaline phosphatase in the presence of BCIP/NBT substrates (Roche Biochemical, Mannheim). Similarly, for monitoring interactions among drDnaA, DnaAΔCt, and drDnaB in D. radiodurans by co-immunoprecipitation, the coding sequences of drDnaA, DnaAΔCt, and drDnaB with polyhistidine tag were PCR amplified using pETHisFw and pETHisRw primers and cloned in pRADgro plasmid (80) at ApaI and XbaI sites to yield pRADhisDA, pRADhisDACt, and pRADhisDB, respectively (Table S1). The expressions from these plasmids were confirmed in D. radiodurans using anti-polyhistidine antibodies (Fig. S7, E and F). In addition, the coding sequences of T18-tagged drDnaA and drDnaB were PCR amplified using BTHF(pv) and BTHR(pv) primers from pUT18DA and pUT18DB plasmids and cloned in p11559 plasmid at NdeI and XhoI sites to yield pV18DA and pV18DB, respectively (Table S1). Expression of T18-tagged drDnaA and drDnaB from pV18DA and pV18DB in D. radiodurans was confirmed using anti-T18 antibodies (Fig. S8A). These plasmids were cotransformed in different combinations in D. radiodurans and induced with 5 mM IPTG as required. The cell-free extracts of D. radiodurans expressing drDnaA, DnaAΔCt, and drDnaB in different combinations were prepared and immunoprecipitated using antipolyhistidine antibodies as described earlier (67). The immunoprecipitates were purified using Protein G Immunoprecipitation Kit (Cat. No. IP50, Sigma-Aldrich Inc) and were separated on SDS-PAGE for western blotting using monoclonal antibodies against T18 as described above. To show the interaction of PprA with drDnaA or drDnaB, the wild-type cells of D. radiodurans expressing native PprA were transformed with pV18DA or pV18DB and expression of T18-tagged drDnaA or drDnaB was induced with 5 mM IPTG. The cell-free extracts of these transformants were prepared and immunoprecipitated using Anti-PprA antibodies. The purified immunoprecipitates were processed for western blotting using T18 antibodies as described above.

Surface plasmon resonance
Interaction of drDnaA and drDnaB with PprA was also investigated using surface plasmon resonance (SPR; Autolab Esprit, Netherland) as described in (77). For this, 20 μM PprA was immobilized on a bare gold sensor chip using EDC-NHS chemistry at 20 C as described in the user manual, which results in 200 response units in running buffer [20 mM Tris (pH 7.6)]. The different concentrations (4-20 μM) of recombinant drDnaA or drDnaB incubated with 1 mM ATP and 1 mM MgCl 2 were used in the mobile phase and passed over the PprA-bound sensor chip in one channel. Reaction buffer containing 20 mM Tris (pH 7.6), 1 mM ATP, and 1 mM MgCl 2 was used as buffer control to flow from another channel over immobilized PprA. The response units for each concentration of proteins were recorded and normalized with buffer control. Further, data were processed using the inbuilt Autolab kinetic evaluation software (V5.4) to find dissociation constant and plotted after curve smoothening using GraphPad Prizm6 software.

Thin layer chromatography for ATPase assay
ATPase activity of drDnaA, DnaAΔCt, and drDnaB was measured as the release of [ 32 P] αADP from [ 32 P] αATP using Thin Layer Chromatography (TLC) as described earlier (67).
In brief, different concentrations (0-3 μM) of drDnaA or drDnaB were mixed with 30 nM [ 32 P] αATP in the absence and presence of 0.2 pmol dsDNA or 0.1 pmol ssDNA in 10 μl reaction mixture containing buffer B, respectively. The reaction mixture was incubated at 37 C for 30 min. For monitoring the effect of PprA on ATPase activity, 2 μM PprA was used with 2 μM drDnaA or 2 μM drDnaB as described above. The reaction mixture was stopped at the different time points (0-30 min) with 10 mM EDTA solution and 1 μl of it was spotted on the PEI-Cellulose F + TLC sheet. The spots were airdried, and products were separated on a solid support in a buffer system containing 0.75 M KH 2 PO 4 /H 3 PO 4 (pH 3.5). After separation, the TLC sheets were air-dried and exposed to X-ray film. The autoradiograms were developed. Spot intensities of samples were determined by densitometry using Image J 2.0 software. The percentage ratio of ADP to ATP was calculated and plotted using the GraphPad Prizm6 software.

Fluorescence microscopy
The D. radiodurans R1 (WT) and its ΔpprA mutant cells were grown till the exponential phase, and the equal number of cells were incubated with 2.5 μg 4',6-diamidino-2phenylindole dihydrochloride (DAPI) per 10 8 cells for nucleoid staining. These cells were washed twice in phosphate buffer saline (PBS) and then mounted on slides coated with 0.8% agarose. Cells were imaged in DIC channel (automatic exposure) and DAPI channel (50 milliseconds exposure) under identical conditions for both WT and ΔpprA mutant using an Olympus IX83 inverted fluorescence microscope equipped with an Olympus DP80 CCD monochrome camera. Many cells (150) from both WT and ΔpprA mutant were processed and the fluorescence intensity of DAPI stained nucleoid was measured using the "intensity profile" tool in cellSens1. 16 software installed with the microscope. Mean fluorescence intensity was plotted against sample type using GraphPad Prizm6 software. In addition, relative frequency distribution (%) was also plotted against fluorescence intensity using GraphPad Prizm6 software.

Determination of ploidy in wild type and ΔpprA mutant
The amount of DNA and copy number of all four genome elements such as chromosome I, chromosome II, megaplasmid, and plasmid per cell of wild type and ΔpprA mutant were determined as described earlier (67). In brief, the exponentially growing wild type and ΔpprA cells were adjusted at an optical density of 600 nm and their cell numbers were determined using a Neubauer cell counter. The collected cells were washed in PBS followed by 70% ethanol wash and lysed in a solution containing 10 mM Tris pH 7.6, 1 mM EDTA, 4 mg/ ml lysozyme. The cells were incubated at 37 C and cellular debris was removed by centrifugation at 10,000 rpm for 5 min. The integrity of isolated genomic DNA was confirmed by agarose gel electrophoresis using 0.8% agarose gel. The DNA content was measured at OD260 nm and the genomic copy number was determined using quantitative real-time PCR as described in (81). For the quantification of genome copy numbers, two different genes showing similar PCR efficiency were taken per replicon. For example, the ftsZ and ftsE for chromosome I, DR_A0155 and DR_A0002 for chromosome II, DR_B003 and DR_B0076 for megaplasmid, and DR_C001 and DR_C018 for small plasmid (Table S2). PCR efficiency of each gene was analyzed and was found to be >96% for each (data not shown). We have followed the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) guidelines (82) for qPCR using Roche Light cycler and the cycle threshold (Cp) values were determined. The experiment was performed using three independent biological replicates for each sample. The Cp value for each gene of the respective replicon was compared with a dilution series of a PCR product of known concentration, as a standard (Fig. S12). The copy number of each replicon by both genes per cell was determined using the cell number present at the time of cell lysis. Thus, an average of copy numbers from two genes per replicon in WT and ΔpprA mutant was calculated and represented along with biostatistical analysis.

Cell survival studies
The D. radiodurans cells were grown till early logarithmic phase and divided into two sets. One was treated with 500 mM hydroxyurea for 1 h and other set was kept as control. Both these samples were centrifuged and resuspended into the same volume of fresh TYG medium. Each of these was divided into two aliquots. One aliquot of each set was exposed to 6 kGy gamma radiation and the other was used as SHAM control. These cells were inoculated in normal TYG medium in six replicates and growth was monitored for overnight using Synergy H1 multimode plate reader. Data were analyzed for statistical deviations using GraphPad Prizm6 software and plotted at mean ± SD (n = 6). The wild type, pprA mutant, and mutant complemented in trans with histidine tagged PprA on a plasmid were exposed to 6 kGy gamma radiation and different dilutions were spotted in TYG agar plate with SHAM control.

Data availability
All data are contained within the article in form of main figures, table, and supplementary materials.
Supporting information-This article contains supporting information.