Coordination of Polyploid Chromosome Replication with Cell Size and Growth in a Cyanobacterium.

Polyploidy has evolved many times across the kingdom of life. The relationship between cell growth and chromosome replication in bacteria has been studied extensively in monoploid model organisms such as Escherichia coli but not in polyploid organisms. Our study of the polyploid cyanobacterium Synechococcus elongatus demonstrates that replicating chromosome number is restricted and regulated by DnaA to maintain a relatively stable gene copy number/cell volume ratio during cell growth. In addition, our results suggest that polyploidy confers resistance to UV, which damages DNA. This compensatory polyploidy is likely necessitated by photosynthesis, which requires sunlight and generates damaging reactive oxygen species, and may also explain how polyploid bacteria can adapt to extreme environments with high risk of DNA damage.

C hromosomal ploidy has diversified during evolution in all three domains of life, bacteria, archaea, and eukaryotes. Most eukaryotic organisms are diploid, but some plants are triploid or polyploid. Likewise, ploidy level varies in bacteria. Many bacteria adopted as models for cell biological studies, such as Escherichia coli and Bacillus subtilis, possess a single circular chromosome (genome) per cell (monoploid) during slow growth and become mero-origoploid during fast growth. In the case of Caulobacter crescentus, cells are monoploid irrespective of growth rate (1,2). Alternatively, some bacteria maintain multiple chromosome copies per cell (polyploid) irrespective of growth rate as observed in cyanobacteria (3,4), Deinococcus (5,6), Thermus thermophilus (7), and symbiotic or parasitic bacteria such as Buchnera (8), Neisseria (9), Borrelia (10), and Epulopiscium (11). Chloroplasts and mitochondria, which evolved in eukaryotic cells from bacterial endosymbionts, are also polyploid. In addition, polyploidy has been reported in several lineages of archaea (12)(13)(14)(15)(16).
Although mechanisms that regulate polyploidy are not known, a positive association between ploidy level and cell size has been observed in a certain species of polyploid prokaryotes and eukaryotes. In polyploid prokaryotes such as the Grampositive bacteria Epulopiscium spp. (17), the cyanobacterium Synechococcus elongatus (18), and the halophilic archaeon Halobacterium salinarum (13), larger cells possess more copies of individual chromosomes than smaller cells. Similarly, artificial polyploidy in the yeast Saccharomyces cerevisiae was associated with greater cell size (19). In plant epidermal cells as well, a positive association of ploidy with cell size has been observed (20). It has been assumed that cytoplasmic cell volume changes in a manner depending on ploidy level which regulates rate of ribosome biogenesis, gene expression level, or other factors in yeasts and plants (21,22). However, the studies did not rule out the opposite possibility that cell volume regulates the ploidy level. Thus, it is unclear whether there is any causal relationship between ploidy level and cell size in polyploid organisms and, if so, whether cell growth increases ploidy level or increased ploidy level leads to increased cell size (21,22). Further, the biological significance of this relationship remains obscure. Monoploidy is apparently advantageous for proliferation compared to polyploidy due to the lower energetic costs and reduced risks associated with DNA replication during cellular proliferation. Nevertheless, polyploidy has evolved many times in both prokaryotes and eukaryotes, implying that polyploidy confers certain survival advantages in specific environments.
Precise chromosomal DNA replication is essential for inheritance of advantageous genetic traits during proliferation. Molecular studies in model bacteria such as E. coli have shown that chromosome replication is tightly controlled mainly at the initiation stage of DNA replication. The initiator protein DnaA first binds to the replication origin (oriC) and then recruits components of the replisome (23,24). DnaA AAAϩ ATPase is highly conserved among bacteria. It binds both ATP and ADP, but only ATP-bound DnaA is capable of forming an oligomeric structure at the oriC region and initiating DNA replication. In E. coli, DnaA level is constant throughout cell cycle progression, whereas the ratio of ATP-DnaA to the total DnaA pool peaks just before chromosome replication in the cell cycle (25,26). A similar regulatory mechanism for DnaA activity during the cell cycle has been observed in other bacteria (23,27,28). In contrast, DnaA activity in C. crescentus is regulated by a change in the total DnaA level (29,30). It is believed that regulation of DnaA activity ensures chromosome replication once per cell cycle in monoploid bacteria, although the specific mode of regulation (i.e., DnaA level or ratio of ATP-DnaA to total DnaA) has diverged among species (31). In contrast to these monoploid bacteria, regulatory mechanisms for DNA replication are still poorly understood in polyploid bacteria.
Cyanobacteria are photosynthetic, and it has been reported that many species are polyploid (32)(33)(34)(35)(36). Recent studies in Synechococcus elongatus PCC 7942 have begun to reveal the regulatory mechanism of polyploid DNA replication in cyanobacteria. In S. elongatus, DNA replication is initiated at the oriC region and replication requires DnaA as in monoploid bacteria (37,38). However, all multicopy chromosomes (four to six) are not replicated simultaneously; rather, only one or two chromosomes are replicated at any stage of the cell cycle (37)(38)(39)(40). In addition, chromosomal copy number per cell changes and exhibits a positive and linear relation with cell size (18,39,40). Furthermore, despite the increase in cell volume, protein concentration remains constant (18); thus, it has been suggested that the increase in chromosomal copy number with growth allows the cell to maintain individual mRNA and protein concentrations (18). However, it is still unclear how replication of multiple chromosome copies is regulated so that only a few are replicated at once and how cell size and copy number are coordinated. In addition, it is not known whether all chromosomal copies contribute mRNA and protein production to maintain a constant protein concentration per cell.
Here we show that initiation of DNA replication and the number of replicating chromosomes are regulated by DnaA and that the rate of cell growth determines the number of replicating chromosomes by modulating DnaA level and activity in S. elongatus. In addition, it is shown that increasing the ploidy level also increases cellular resistance to UV irradiation, suggesting that possessing multicopy chromosomes allows the organism to cope with DNA damage.

RESULTS
Only a few chromosomal copies are replicated at once while genes are transcribed from all chromosomal copies in cyanobacteria. Previous reports suggested that the number of replicating chromosomes is restricted to one (or in some case two) in S. elongatus (37)(38)(39)(40). Zheng and O'Shea recently reported that chromosomal copy number and protein concentration per cell volume are maintained constant independently of cell size and growth rate (18) in S. elongatus. These observations imply that genes are transcribed from all chromosome copies to maintain a constant concentration of transcripts and proteins; however, the transcriptional activity at chromosomes was not examined.
When S. elongatus cells were fixed and stained with a DNA-specific fluorescent dye, SYBR Green I, multiple (3 to 6) copies of chromosomes were observed aligned along the long axis of the cell (Fig. 1A). To examine transcriptional activities of multicopy chromosomes, we visualized replicating chromosomes and transcribing chromosomes in live S. elongatus cells. In addition, to assess the generality of the results, the same assays were applied to Synechocystis sp. strain PCC 6803 (here called Synechocystis), a model cyanobacterium phylogenetically distant from S. elongatus (41,42). To this end, we expressed GFP-tagged single-strand binding protein (SSB; see Fig. S1 and S2 in the supplemental material), which is known to localize at replication forks (40,43), and RNA polymerase beta subunit (RpoC2; Fig. S3 and S4). Consistent with previous reports (37,39,40), the majority of S. elongatus cells exhibited only one SSB focus (replication fork) per cell during growth under illumination (Fig. 1B). In contrast, RNA polymerase was observed on all chromosome copies (Fig. 1C).
Synechocystis possesses more chromosome copies (10 to 20) than S. elongatus (3 to 6) (3,36). When genomic DNA of Synechocystis was stained with SYBR Green I, many small foci were observed in each cell (Fig. 1D). In contrast, SSB-GFP was detected at only one focus per cell (Fig. 1E). In a similar manner, only one chromosome per cell (or two in the case of dividing cells) incorporated 5-bromo-2=-deoxyuridine (BrdU), an analog of thymidine incorporated into newly synthesized DNA, during a 1-h observation period (Fig. 1F). In contrast, RNA polymerase localized on all chromosome copies (Fig. 1G).
These results indicate that the number of replicating chromosomes is restricted to one or two while all chromosome copies are used as the templates for transcription both in S. elongatus and in Synechocystis.
DnaA binding to the oriC region depends on photosynthesis. To address how replicating chromosome number is regulated, we examined the relationships of DnaA protein level and activity with chromosomal replication and chromosomal copy number in S. elongatus. It was previously shown that genomic DNA replication absolutely depends on photosynthetic activity in S. elongatus, so replication ceases under darkness (44). To examine whether DNA replication correlates with DnaA activity, we first compared DnaA protein level and oriC-binding activity between light and dark condi-tions. Cells were grown in an inorganic medium for 5 h under illumination and then kept under light or transferred to dark conditions ( Fig. 2A). Immunoblot analysis using anti-DnaA antibody showed that DnaA protein level was constant for 3 h under both light and dark conditions ( Fig. 2B and Fig. S5). In contrast, chromatin immunoprecipitation using the anti-DnaA antibody and subsequent quantitative PCR (ChIP-qPCR) analysis of the oriC region showed that the affinity of DnaA to oriC decreased by approximately two-thirds after 30 min under darkness (Fig. 2C). When an inhibitor of photosynthetic electron flow, 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) or 2,5dibromo-3-methyl-6-isopropyl-p-benzoquinone (DBMIB), was added to the culture under light, oriC binding of DnaA was almost completely abolished although DnaA level remained constant ( Fig. S5 and Fig. S6A, B, and C). These results suggest that DNA replication activity is associated with the oriC-binding activity of DnaA rather than DnaA protein level. Further, oriC binding and DNA replication appear to depend on photosynthesis, which was confirmed in subsequent studies.
ATP-dependent binding of DnaA to oriC in S. elongatus. In many model bacterial species, binding of DnaA to the oriC region is controlled by the ATP/ADP ratio, as only the ATP-bound form is capable of forming the oligomeric structure at the oriC region required for initiation of DNA replication (45). To assess the mechanism of light/ photosynthesis-dependent activation of DnaA activity in S. elongatus, we examined the effect of changing the proportion of ATP-bound DnaA on oriC-binding activity under constant illumination.
The arginine residue at position 334 of E. coli DnaA is essential for ATP hydrolysis, and the amino acid substitution R334H inactivates this intrinsic ATPase activity, which leads to accumulation of DnaA-ATP (46). This arginine is conserved in DnaA of many organisms (47), including cyanobacteria (Fig. 2D). In order to examine the effect of ATP binding to DnaA on oriC affinity in S. elongatus, we expressed HA-DnaA (DnaA WT ) or HA-DnaA R328H (DnaA R328H ) (which corresponds to R334H of E. coli DnaA) under the control of the dnaA promoter from the genomic neutral site (NS I) on an endogenous dnaA-knockout background ( Fig. 2E and Fig. S7). Binding to oriC was then examined by chromatin immunoprecipitation (ChIP). For these studies, HA tag was fused to the N terminus of DnaA WT or DnaA R328H so that the protein could be precipitated with an anti-HA antibody instead of anti-DnaA because the affinity of the polyclonal anti-DnaA antibody would likely differ between DnaA WT and DnaA R328H .
Both DnaA WT and DnaA R328H proteins were expressed in respective transformants at similar levels, and the expression levels remained constant for 3 h after cells were  (44). The samples were quantified by qPCR using primers specific to the oriC region and the Syf1294 gene (1294), which is farthest from oriC in the circular chromosome. The value indicated is the ratio of percent recovery (oriC/1294) normalized to the value of DnaA WT at hour 0 (defined as 1.0).
Coordination of Chromosome Number with Cell Size ® transferred to darkness (Fig. 2F). Similarly to endogenous DnaA (Fig. 2C), DnaA WT bound to the oriC region in cells cultured under illumination and the affinity was reduced when the cells were transferred to darkness (Fig. 2G). The R328H mutation elevated DnaA activity due to an approximately 6-fold increase in oriC affinity compared to DnaA WT under illumination (Fig. 2G). When the DnaA R328H cells were transferred to darkness, the affinity of DnaA R328H to oriC decreased but the level was still higher than DnaA WT (Fig. 2G). These results suggest that the ATP-bound form of DnaA possesses higher affinity for the oriC site than the ADP-bound form as shown in other model bacteria. In addition, the light/photosynthesis-dependent activation of DnaA is, at least partly, independent of DnaA intrinsic ATPase activity because the affinity of constitutive DnaA R328H for oriC decreased when the cells were transferred to darkness or treated with DCMU ( Fig. 2G and Fig. S6E). This assumption is based on the following observations in E. coli. The R334H mutation completely abolished the intrinsic ATPase activity of DnaA in vitro (46). However, ATP bound to the mutated DnaA was gradually hydrolyzed by the addition of crude extract of E. coli in vitro (46). In addition, ATP bound to DnaA was slowly hydrolyzed by ATPases other than DnaA in vivo (23).
To examine the effect of DnaA R328H expression and light/photosynthesis on replication fork components other than DnaA, we compared affinity of DnaB to oriC between DnaA WT and DnaA R328H transformants by ChIP-qPCR analysis. DnaB is a DNA helicase recruited to the oriC region in a DnaA-and DnaC-dependent manner in bacteria. Similar to DnaA, the affinity of DnaB to oriC was higher in DnaA R328H cells than in DnaA WT cells under both light and dark conditions (Fig. 2G). In addition, the affinity of DnaB to oriC decreased in both DnaA R328H and DnaA WT cells after transfer to darkness. These results suggest that the increased affinity of DnaA to oriC by light/ photosynthesis or by an increase in ATP-DnaA by the R328H mutation promoted the formation of replication forks.
The oriC-binding activity of DnaA R328H was 6-hold higher than that of DnaA WT under illumination, although the expression level of DnaA R328H protein was comparable to that of DnaA WT (Fig. 2G). Thus, it is suggested that DnaA activity is limited, probably by moderation of ATP hydrolysis, even under illumination to restrict the number of replicating chromosomes as shown below.
Simultaneous replication of multiple chromosome copies by increasing ATP-DnaA. As mentioned above, only one copy of multiple chromosomes is replicated in wild-type S. elongatus (37,39,40). This finding led us to examine the effect of artificially elevating DnaA activity by expressing DnaA R328H on the number of replicating chromosome copies. To this end, we expressed GFP-tagged SSB to visualize replicating chromosomes in DnaA WT and DnaA R328H cells. Consistent with previous reports (37,39,40) and the results shown in Fig. 1, the majority of DnaA WT cells exhibited only one SSB focus per cell during growth under illumination ( Fig. 3A to C). In contrast, the majority of DnaA R328H cells exhibited two or more SSB foci under the same growth conditions ( Fig. 3B and C). When DnaA WT and DnaA R328H cells were labeled with BrdU, one or two BrdU foci were detected in DnaA WT cells (Fig. 3B) whereas two or more foci were detected in DnaA R328H cells 6 h after inoculation under illumination (Fig. 3B). These results indicate that elevation of DnaA activity by enhancing the ratio of ATP-DnaA to total DnaA results in simultaneous replication of multiple chromosome copies.
In order to compare frequency of DNA replication initiation between DnaA WT and DnaA R328H cells on a genome-wide basis, relative copy numbers of chromosomal regions in these strains were analyzed by next-generation sequencing. When model bacteria possessing a single chromosome (genome) per cell, such as E. coli, are rapidly grown in nutrient-rich medium, DNA is replicated in a multifork mode. In this case, the oriC/terC ratio is more than 2, thus yielding a V-shaped profile in the depth of sequencing reads (lowest at terC and highest at oriC) (37). In the case of cyanobacteria possessing multiple chromosome copies, of which only one copy is being replicated, the ratio of terC and oriC to total DNA is almost 1.0, thus yielding an almost linear profile in the depth of sequencing reads (37) (Fig. S8A, DnaA WT ) throughout the genome. The sequencing profile of DnaA R328H cells 6 h after inoculation under illumination exhibited a slight V-shape, in contrast to the sequencing profile of DnaA WT cells (highest at oriC region; Fig. S8A). However, the ratio of oriC to terC was less than 2 (Fig. S8A). These results suggest that some chromosomes are replicated simultaneously in DnaA R328H cells under illumination but that the number is still restricted, most likely by ATPase proteins other than DnaA. Nonetheless, this restriction was compromised to a certain extent when the intrinsic ATPase activity was abolished by the R328H mutation.
Relationship between cell growth and DNA replication in S. elongatus. Chromosome copy number exhibits a positive, linear correlation with cell volume in S. elongatus (18,39,40). However, it is not known whether increased chromosomal copy number leads to increased cell volume or if cell growth leads to an increased number of chromosomal copies. To address whether there is such a causal relationship between chromosomal copy number and cell size, we examined the relationships among growth rate, cell volume, and frequency of chromosomal replication.
First, we investigated the effect of changing the chromosome copy number on cell growth and cell volume by comparing DnaA R328H and DnaA WT cells. When DnaA R328H and DnaA WT were cultured under illumination, DnaA R328H cells possessed more chromosome copies than DnaA WT cells (Fig. 3D). Thus, an increase in the number of replicating chromosome by expression of DnaA R328H led to an increase in chromosomal Coordination of Chromosome Number with Cell Size ® copy number per cell. However, DnaA R328H and DnaA WT cells exhibited similar growth rates (Fig. 3A) and cell volumes (Fig. 3E). Thus, the cell volume occupied by one chromosome in DnaA R328H cells was smaller than that in DnaA WT cells ( Fig. 3F and G). Similar results were also observed by measuring cell volume and DNA level by flow cytometry (Fig. S8B). These results suggest that initiation frequency and number of chromosome copies have little impact on cellular growth rate and volume.
Next, to examine the effect of cell growth on chromosomal replication and chromosome copy number, we cultured S. elongatus under different light intensities (Fig. 4A) (Fig. 4A). As light intensity and growth rate increased, the number of SSB foci per cell also increased (average of 0.26 foci per cell at 5 E m Ϫ2 s Ϫ1 , 0.92 at  (Fig. 4B). On the other hand, cells growing under different light intensities exhibited similar chromosome copy numbers and cell volumes and thus similar positive correlations between cell volume and chromosomal copy number (Fig. 4C to G). Comparable results were also observed by measuring cell volume and DNA level by flow cytometry (Fig. S9A). Changing the light intensity also influenced DnaA level. Specifically, DnaA level decreased when light intensity was reduced to 5 E m Ϫ2 s Ϫ1 (hour 6) from 70 E m Ϫ2 s Ϫ1 (hour 0) (Fig. 4H and Fig. S5) and increased when the light intensity was raised to 250 E m Ϫ2 s Ϫ1 (hour 6) from 70 E m Ϫ2 s Ϫ1 (hour 0) (Fig. 4H and Fig. S5). In addition, the oriC-binding activity of DnaA was lower at 5 E m Ϫ2 s Ϫ1 than at 70 or 250 E m Ϫ2 s Ϫ1 (Fig. 4I; note that in the ChIP-qPCR analysis, sample input DnaA depends on the cellular DnaA level). These results suggest that an increase or decrease in cellular growth rate leads to a corresponding increase or decrease in chromosomal replication activity to maintain a constant chromosomal copy number per unit cell volume at a constant temperature.
When the cells were transferred from light to dark or photosynthesis was inhibited with DCMU or DBMIB, cellular growth and chromosome replication ceased whereas DnaA level remained constant ( Fig. 2 and Fig. S5 and S6). In contrast, when the light intensity was reduced and cellular growth slowed down, DnaA level decreased ( Fig. 4 and Fig. S5). Thus, DnaA level changes only when cells are growing and the growth rate changes to match chromosome replication rate, but does not change when cellular growth and chromosome replication cease.
We also examined the relationship between cell growth and chromosomal replication and copy number at different temperatures. As temperature increased, cell growth accelerated (0.01 h Ϫ1 at 20°C, 0.03 h Ϫ1 at 30°C, and 0.05 h Ϫ1 at 40°C; Fig. 5A). In contrast to results at constant temperature (30°C) and variable light intensity (Fig. 4), the number of SSB foci per cell increased at lower temperature ( Fig. 5B) even though cell growth was slower (Fig. 5A). Consistent with these observations, microscopic analysis and flow cytometry showed that chromosome copy number per cell was higher at lower temperature 6 h after temperature reduction (Fig. 5C, D, and E) while cell volume per chromosome was reduced at lower temperature ( Fig. 5F and G). Similar results were also observed by measuring cell volume and DNA level by flow cytometry (Fig. S9B). The level and oriC-binding activity of DnaA were higher at lower temperature ( Fig. 5H and I; note that in the ChIP-qPCR assay, sample input DnaA level depends on cellular DnaA level). Although cell volume per chromosome depended on temperature, cell volume and chromosomal copy number exhibited a linear correlation at all temperatures examined. These results suggest that the number of replicating chromosomes (initiation frequency) and chromosomal copy number are regulated by a certain temperature-affected factor.
Multicopy chromosomes confer greater UV resistance to S. elongatus. Photoautotrophic organisms require light for growth. However, UV in sunlight damages DNA directly and the photosystems generate reactive oxygen species (ROS) that also damage DNA (48). Thus, photosynthetic organisms are exposed to higher risk of DNA damage during cell growth than many other heterotrophic organisms. Polyploid bacteria often inhabit extreme environments such those with high temperatures (e.g., Thermus thermophilus) and possess resistance to high UV (e.g., Deinococcus radiodurans). In a similar manner, polyploid chloroplasts and mitochondrial DNA are also exposed to oxidative stresses that result from photosynthesis and respiration (49,50). Thus, one plausible advantage of multiple chromosome copies is to compensate for a damaged copy using the information from the undamaged replicate.
In order to assess this possibility, we investigated the relationship between ploidy level and cellular resistance to UV in S. elongatus. Populations with different ploidy levels were generated by cultivation of DnaA WT cells and DnaA R328H cells at 20, 30, or 40°C ( Fig. 3 and 5). After cultivation in liquid culture under light for 6 h, cells were spotted onto agar medium. After UV irradiation, cells were cultured on agar medium at 30°C to allow surviving single cells to produce colonies ( Fig. 6A and B). Cell viability just Coordination of Chromosome Number with Cell Size ® after UV irradiation was quantified based on the number of single colonies (Fig. 6C and E). As temperature increased and ploidy level decreased (Fig. 5), susceptibility to UV increased in the wild type (Fig. 6E). Ploidy was higher in DnaA R328H than DnaA WT cells at respective temperatures, and resistance to UV was also greater in DnaA R328H cells than in DnaA WT cells at respective temperatures ( Fig. 6C and D). Thus, cells of higher ploidy were more resistant to UV irradiation, supporting the hypothesis that polyploidy benefits organisms under higher risk of DNA damage.

DISCUSSION
DnaA-dependent regulation of multicopy chromosomes in cyanobacteria. Our results suggest that DNA replication in S. elongatus is absolutely dependent on photosynthesis through regulation of DnaA activity ( Fig. 2;   Ohbayashi et al. ® mental material) and that the number of replicating chromosomes is restricted so that chromosome replication frequency is matched to cellular growth rate. Through this coordination, chromosome copy number increases and decreases according to cell size ( Fig. 4 and 5), thereby maintaining protein and metabolite concentrations during cell growth.
Supporting this conclusion, inhibition of DnaA intrinsic ATPase activity by the R328H mutation increased the number of replicating chromosomes during cell growth (Fig. 3). However, DNA replication was still limited to a certain chromosome copy number in DnaA R328H cells ( Fig. 3; Fig. S8). In addition, the oriC-binding activity of DnaA R328H decreased when photosynthesis and cell growth were blocked by transfer to darkness or chemical inhibition of photosynthesis, although DnaA R328H cells still exhibited higher oriC-binding activity than DnaA WT cells ( Fig. 2; Fig. S6). These results suggest that in addition to the intrinsic ATPase activity of DnaA, other, as-yet-unknown factors also regulate DnaA activity in S. elongatus. There are several steps during initiation in which regulatory systems have been found to control bacterial DNA replication (for a review, see reference 23). For example, oriC binding of ATP-DnaA is inhibited by SeqA in E. coli, Spo0A in B. subtilis, and CtrA in C. crescentus, all of which occupy the oriC region.
In addition, there are also likely DnaA-independent mechanisms that regulate the number of replicating chromosomes in cyanobacteria. In this study, we showed that the number of replicating chromosome is also restricted to one in Synechocystis cells during growth (Fig. 1). As in S. elongatus, chromosome replication in Synechocystis depends on Coordination of Chromosome Number with Cell Size ® photosynthesis (44). However, even when DnaA protein is depleted, Synechocystis grows in a manner similar to the wild type, and it is likely that initiation and regulation of chromosome replication in Synechocystis are governed mainly by unknown DnaAindependent mechanisms (38) that also function in S. elongatus.
Relationship between cell growth and initiation of chromosome replication. In this study, we showed that DnaA level, DnaA activity, and number of replicating chromosomes changed depending on cellular growth rate ( Fig. 4 and 5). In addition, chromosome copy number per unit cell volume was nearly constant during cell growth at a constant temperature ( Fig. 4 and 5). On the other hand, increasing the number of replicating chromosomes and chromosome number per cell by inhibition of DnaA ATPase activity did not affect cell growth (Fig. 3). These results suggest that the number of replicating chromosomes is regulated by cell growth through regulation of DnaA level and activity in S. elongatus.
Regulation of chromosome replication depending on cellular growth has also been demonstrated in monoploid bacteria such as E. coli, in which chromosome replication is initiated when the cell grows to a certain fixed volume (51,52). However, in both monoploid and polyploid bacteria, it is still unclear how a cell senses its volume to regulate chromosome replication. When cell volume increases, concentrations of RNAs transcribed from chromosomes and proteins translated from these mRNAs are predicted to decrease until additional chromosome copies are generated by replication (Fig. 7). Therefore, chromosomal replication is likely triggered by a reduction in the concentration of an as-yet-unknown factor or factors by cell growth. For example, chromosomal replication may be initiated when the concentration of a replication repressor decreases below a certain threshold. This notion is supported by the observation that DnaA level and activity, number of replicating chromosomes, and chromosome number per cell volume were lower at higher temperature in S. elongatus (Fig. 5). Increased temperature accelerates chemical reactions, including enzymatic reactions; thus, transcriptional and translational activities increase with temperature. At higher temperature, cells require lower levels of template chromosomal DNA and thus a smaller number of chromosomal copies to sustain protein and metabolite levels than at lower temperature.
Biological significance of polyploidy in bacteria. Polyploidy is widespread in bacteria, archaea, and eukaryotes, but the advantages conferred by multicopy genomes are unclear. Results of this study suggest that polyploid bacteria are able to change chromosomal copy number per cell during cell growth to maintain nearly constant gene copy number per unit cell volume (Fig. 7) and that multiple copies of chromosomes in bacteria act as backup genetic information to compensate for damage to the other chromosomal copies. In accord with the first notion, compared to monoploid organisms or polyploid organisms exhibiting simultaneous replication of chromosomes, in which gene number per unit cell volume changes up to 2-fold (53), stepwise replication of chromosomal copies in polyploids such as S. elongatus reduces changes in the gene number/cell volume ratio during cell growth and division. Thus, the stepwise replication of multiple chromosomal copies enables a cell to keep nearly constant chromosome number per unit cell volume, which coordinates chromosomal number with cell size (Fig. 7) and probably contributes to maintaining a constant mRNA and protein concentration during cell growth as observed in S. elongatus (18). Consistent with chromosome copies as backup genetic information, some known polyploid eukaryotes inhabit extreme environments (7,(12)(13)(14)(15)(16) that can damage DNA (48). In addition, cyanobacterial growth depends on sunlight, which contains UV, and photosynthesis generates ROS, and both UV radiation and ROS damage DNA (48,54). In this study, S. elongatus cells of higher ploidy exhibited greater UV resistance. Thus, despite the higher metabolic cost of polyploidy than of monoploidy, possessing multiple copies of a chromosome that replicates asynchronously probably benefits the organism by compensating for a damaged copy.

MATERIALS AND METHODS
Strains and culture conditions. Strains used in this study were Synechococcus elongatus PCC 7942 (WT) and several transformants (described below) as well as Synechocystis sp. PCC 6803 (GT-I strain) (55) and several transformants (described below) including a thymidine kinase (TK) strain (38) expressing thymidine kinase for the BrdU assay and a strain expressing DnaB-FLAG under the control of the dnaB promoter (44). All cyanobacterial strains were cultured at 30°C unless otherwise indicated in BG-11 liquid medium with air bubbling and illumination of 70 E m Ϫ2 s Ϫ1 unless otherwise indicated. DCMU or DBMIB was added to the culture at a final concentration of 5 M to inhibit photosynthetic electron flow.
ChIP-qPCR analyses. Combined chromatin immunoprecipitation and qualitative PCR (ChIP-qPCR) analyses were performed according to the method of Hanaoka and Tanaka (56) with minor modifications. Cells were fixed with 1% formaldehyde for 15 min at room temperature. The cross-link reaction was stopped by the addition of glycine (final concentration of 125 mM) and incubation at room temperature for 5 min. Cells were harvested by centrifugation, washed twice with ice-cold Tris-buffered saline (TBS) (20 mM Tris-HCl, pH 7.4, 150 mM NaCl), and stored at Ϫ80°C. Fixed cells were broken using Beads Crusher (Taitec) with glass beads (Ͻ106 m; Sigma Aldrich) in TBS at 4°C, and the genome was fragmented to approximately 500 bp using a Covaris sonication system (MS Instrument Inc.). After centrifugation for 15 min to precipitate the insoluble fraction, the supernatant containing genomic DNA was subjected to immunoprecipitation using an anti-HA antibody (clone 16B12; BioLegend), anti-SyfDnaA antibody (38), or anti-FLAG M2 antibody (F1804; Sigma) at a dilution of 1:250. Precipitated DNA was quantified by qPCR using primer sets oriC-F and oriC-R for the oriC region and 1294-F and 1294-R for the genomic region farthest from oriC in the circular genome. Primer sequences are listed in Text S1 in the supplemental material.
Immunofluorescence microscopy. Cells were fixed with 1% paraformaldehyde and 10% dimethyl sulfoxide dissolved in methanol at Ϫ30°C for 5 min. After washing twice with phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.76 mM KH 2 PO 4 , pH 7.4), fixed cells were treated with 0.05% Triton X-100 in PBS for 15 min at room temperature. After washing twice with PBS, the fixed cells were further permeabilized with 0.2 mg ml Ϫ1 lysozyme (dissolved in 25 mM Tris-HCl, pH 7.5, 10 mM EDTA) at 37°C for 30 min. After two additional washes with PBS, genomic DNA was digested in situ with 4 M HCl at 37°C for 1 h. After washing twice with PBS, cells were blocked with Blocking One (Toyobo) at room temperature for 30 min and then were reacted with the anti-BrdU antibody (clone BMC9318; Roche) diluted in Blocking One (1:20) at 37°C overnight. After washing twice with PBS, cells were reacted with Alexa Fluor 488 goat anti-mouse antibody (Invitrogen) diluted in Blocking One (1:200). After two final washes with PBS, cells were observed by fluorescence microscopy.
DNA staining for flow cytometry and observation of nucleoids by fluorescence microscopy. Cells were harvested by centrifugation, fixed with 1% glutaraldehyde for 10 min, and washed with PBS. Fixed cells were stained with SYBR Green I (Invitrogen) (final concentration was 1:1,000) to count chromosomal copy number by fluorescence microscopy or with 10 M Sytox Green for flow cytometry analysis. After staining, cells were subjected to flow cytometry (BD Accuri C6) as previously described (37).
UV irradiation and evaluation of cell viability by spotting assay. Mid-log-phase cultures of S. elongatus DnaA WT or DnaA R328H at 30°C were inoculated into fresh inorganic medium (BG-11) and cultured under 70-E m Ϫ2 s Ϫ1 illumination at 20, 30, or 40°C. A portion of the culture was harvested 6 or 24 h after the inoculation and then adjusted to an OD 750 of 0.2 for preparation of a serial dilution series Coordination of Chromosome Number with Cell Size ® (10 Ϫ1 to 10 Ϫ5 ) using fresh medium. The diluted cells were spotted on BG-11 agar plates and then irradiated with 0 to 600 J m Ϫ2 UV-C (254 nm) under a UV lamp (UVP; Upland). The plates were further incubated at 30°C under 70-E m Ϫ2 s Ϫ1 illumination for 7 days, and cell survival rate was calculated by counting the colonies.