DNA gyrase activity regulates DnaA-dependent replication initiation in Bacillus subtilis

SUMMARY

In bacteria, initiation of DNA replication requires the DnaA protein. Regulation of DnaA association and activity at the origin of replication, oriC, is the predominant mechanism of replication initiation control. One key feature known to be generally important for replication is DNA topology. Although there have been some suggestions that topology may impact replication initiation, whether this mechanism regulates DnaA-mediated replication initiation is unclear. We found that the essential topoisomerase, DNA gyrase, is required for both proper binding of DnaA to oriC as well as control of initiation frequency in Bacillus subtilis. Furthermore, we found that the regulatory activity of gyrase in initiation is specific to DnaA and oriC. Cells initiating replication from a DnaA-independent origin, oriN, are largely resistant to gyrase inhibition by novobiocin, even at concentrations that compromise survival by up to four orders of magnitude in oriC cells. Furthermore, inhibition of gyrase does not impact initiation frequency in oriN cells. Additionally, deletion or overexpression of the DnaA regulator, YabA, significantly modulates sensitivity to gyrase inhibition, but only in oriC and not oriN cells. We propose that gyrase is a negative regulator of DnaA-dependent replication initiation from oriC, and that this regulatory mechanism is required for cell survival.

ABBREVIATED SUMMARY

DNA gyrase controls replication initiation by inhibiting DnaA binding and activity at the origin of replication, oriC. Inhibition of gyrase increases replication initiation frequency and DnaA association with oriC, and is harmful to cell survival if replication initiates from oriC. We propose a model where modulation of DNA topology by gyrase regulates replication initiation at an early step during orisome assembly.

INTRODUCTION

DNA replication is an essential process in all organisms. In bacteria, replication initiates from a single origin of replication, oriC, and proceeds bi-directionally until the replication forks reach the terminus, ter (1). Replication initiation from oriC depends on ordered binding of the replication initiation protein, DnaA, to specific 9-mer consensus sequences (2–4). Oligomerization and cooperative binding of DnaA leads to the melting of the origin at the DNA unwinding element (DUE), and subsequent replisome assembly. Regulation of replication is important for proper cell proliferation and generally occurs at the initiation step through modulation of DnaA binding and activity (5–7).

The regulatory mechanisms for DnaA association and function with oriC in Bacillus subtilis and other Gram-positive bacteria are generally different than in Gram-negatives. For example, key regulators of initiation in B. subtilis include YabA and SirA, which are not found in Escherichia coli (6, 8–10). Additionally, the mechanism of DnaA regulation is different in B. subtilis compared to E. coli: YabA disrupts oligomerization and cooperative binding of DnaA to the origin and this type of regulatory mechanism has not been reported for the key E. coli initiation regulators such as Hda, SeqA, or Dam (6). YabA has also been proposed to sequester DnaA at the replication forks (11).

One critical factor in replication initiation and progression is DNA topology (7, 12, 13). Much of what is known regarding the role of supercoiling in bacterial replication comes from in vitro studies that use plasmid-based systems and proteins purified from E. coli (13–18). These studies have established that a negatively supercoiled DNA template is required for replication initiation by DnaA from oriC in these reconstituted systems (13–18). Furthermore, in vitro, supercoiling can enhance association of Helicobacter pylori and E. coli DnaA to certain consensus sites at oriC (19, 20), and Aquifex aeolicus ATP-DnaA oligomerization has been shown to induce positive DNA supercoils (21). In agreement with this, transcription-induced negative supercoiling near oriC activates replication initiation in E. coli (22, 23). However, though these studies suggest a role for topology in DnaA association and replisome assembly at the origin for these particular Gram-negative bacteria, whether changes in DNA superhelicity regulate replication timing or frequency in vivo or for Gram-positives is unclear.

In vitro, the type II topoisomerase DNA gyrase has been utilized to obtain the necessary supercoiling status for replication reactions to initiate and progress (13–18). Other processes affect DNA topology as well – including DNA replication and transcription (24). Without gyrase, which regulates DNA topology by introducing negative supercoils and relieving excess positive supercoils ahead of replication forks (25, 26), replication cannot proceed (27, 28). Furthermore, in vitro studies suggest that negative supercoiling at the origin promotes replication initiation and increases DnaA binding (19, 20), although gyrase is not required for open complex formation or helicase loading (29). If these in vitro-based models are correct, then upon gyrase inhibition, the rate of replication initiation as well as DnaA binding to the origin should decrease in vivo. These predictions have not been thoroughly tested in living cells. Moreover, the few existing in vivo studies of gyrase contradict the predictions from in vitro work. For example, recent work from E. coli suggest that gyrase promotes ATP hydrolysis by DnaA, at the DnaA sequestration locus, datA, which negatively regulates initiation (30). This is in contrast to in vitro work, which suggests that gyrase promotes DnaA-dependent initiation. Therefore, though various studies have suggested that gyrase may influence replication initiation, the mechanism and potential role of gyrase as a regulator of DnaA binding or activity at oriC in vivo remain to be determined.

We found that gyrase is an essential, negative regulator of replication initiation in vivo, in B. subtilis. Our data indicate that gyrase activity decreases DnaA association with oriC and inhibits replication initiation. The regulatory function of gyrase is specific to DnaA and oriC: replication initiation from an ectopic, DnaA-independent origin, oriN is unaffected by gyrase activity. Furthermore, gyrase inhibition is significantly more detrimental to cell survival when replication initiates from oriC compared to oriN. Lastly, over-expression of the DnaA negative regulator YabA promotes survival of gyrase inhibition. These results suggest that the essentiality of gyrase stems at least partially from a regulatory activity in oriC and DnaA-dependent replication initiation.

RESULTS

Inhibition of type II topoisomerases leads to over-initiation

To test if and how DNA topology impacts replication initiation in vivo, we measured the impact of novobiocin on initiation dynamics. Novobiocin is a useful tool for understanding the importance of DNA topology for essential processes, such as DNA replication and transcription (31–34). Through binding to the GyrB subunit of gyrase at the ATP binding site, novobiocin blocks ATP hydrolysis, and enzymatic activity (33, 35). Novobiocin also binds to the ParE subunit of topoisomerase IV (Topo IV), although, the K_i is several orders of magnitude higher than for gyrase (36). Our understanding of these drug-enzyme interactions are based largely on studies performed in E. coli. However novobiocin demonstrates a high affinity for B. subtilis gyrase as well, suggesting that the preferential effect of novobiocin on B. subtilis gyrase is likely similar to that of E. coli (37).

To assess the role of DNA topology on replication, we determined the marker frequency pattern along the genome for wild-type cells grown in the presence and absence of novobiocin using whole genome sequencing and quantitative PCR (Fig. 1A–B). Using whole genome marker frequency analyses through new generation sequencing, we found that novobiocin increases DNA copy number near oriC two-fold, which gradually decreases to levels exhibited by the untreated control samples (Fig. 1A–B), at around 36 and 340 degrees along the chromosome (Fig. 1A). Consistent with our marker frequency analysis, we also observed this reduction in copy-number at 45 and 315 (indicated as “−45” in the figure) degrees by qPCR (Fig. 1B). These data suggest that changes in DNA topology impact replication initiation in vivo.

Figure 1. Inhibition of type II topoisomerases by novobiocin increases oriC-dependent replication initiation.

A) Marker frequency analysis as measured by deep sequencing for oriC cells with and without 0.75 μg/mL novobiocin treatment for 40 minutes. The x-axis indicates chromosomal location, and the y-axis represents the abundance of reads relative to the total number of reads in the sequencing library. B) Marker frequency as measured by quantitative PCR at 0, 22.5, 45, 135, 225 (−135), 315 (−45), and 337.5 (−22.5) degrees along the chromosome for oriC cells with and without 0.75 μg/mL novobiocin treatment for 40 minutes. Data shown are averages for at least 9 biological replicates, examined on 3 different days. C) Cartoon map of the oriN ΔoriC-S mutant strain, showing the deletion in the region between dnaA and dnaN. D) Origin-to-terminus ratios for oriC and oriN cells with 0, 0.50, and 0.75 μg/mL novobiocin, and with 1 μg/mL MMC are plotted. Data shown are averages from 6–12 biological replicates. Error bars represent standard error of the mean. Statistical significance was calculated using t-test (**p<0.01). E) Marker frequency as measured by quantitative PCR at 0, 22.5, 45, 135, 225 (−135), 315 (−45), and 337.5 (−22.5) degrees along the chromosome for oriN cells with and without 0.75 μg/mL novobiocin treatment. Data shown are averages for at least 9 biological replicates. F) The ratio of total DNA amplified from the origin and terminus for cultures treated with 0.50 and 0.75 μg/mL novobiocin divided by the amount of DNA amplified from the same regions for cultures without treatment are shown. Absolute levels of DNA used are derived from Cq values from qPCR (raw data). Ratios of DNA amplified for oriC and oriN cells are plotted. Data shown are averages from 6–9 biological replicates.

Over-initiation of novobiocin treated cells is specific to oriC

To determine if the impact of novobiocin on replication initiation is specific to cells that undergo DnaA-dependent initiation at oriC, we measured the effect of novobiocin on origin-to-terminus ratios for oriC cells and oriN ΔoriC-S cells that undergo DnaA-independent and unregulated replication initiation from a heterologous origin, oriN (38). oriN is the origin of replication used by pLS32, a plasmid present in Bacillus natto (a.k.a hay or grass Bacillus) (39, 40). Initiation from oriN depends on the initiator protein RepN, which is also expressed in the strains we used in our experiments (Fig. 1C). Both oriN and repN are integrated at spoIIIJ on the chromosome (38). Importantly, the oriN strain lacks the DUE region found at oriC and thus its replication initiation depends on oriN and RepN activity (40, 41) (Fig. 1C).

We measured the ratio of origins to termini upon exposure of cells to increasing concentrations of novobiocin by amplification of oriC and ter using qPCR. Increased replication initiation generally leads to an increased ratio of origin to terminus DNA. For oriC cells, we found that novobiocin treatment increases origin-to-terminus ratios in a dose-dependent manner (Fig. 1D) indicating that cells are over-initiating. In contrast, we did not observe any change in origin-to-terminus ratios in the presence of novobiocin for the oriN cells (Fig. 1D). The oriC novobiocin phenotype remained the same under slow growth conditions, when cultures were grown in minimal glucose media (Fig. S1). Analysis of DNA copy number at several different locations along the genome of the oriN cells showed results consistent with this observation: novobiocin did not increase the DNA copy number at any of the loci analyzed around the chromosome (Fig. 1E).

The changes in origin-to-terminus ratios we observed were due to changes in the copy number at oriC and not ter: 1) the genome-wide marker frequency analysis showed a specific increase in the copy number of origin-proximal DNA (rather than the terminus), and 2) the ratio of total oriC DNA, but not ter DNA, from the novobiocin treated cells compared to untreated cells was over 1 (Fig. 1F). Furthermore, the increased ratio of origin-proximal DNA was only detected in oriC and not oriN cells (Fig. 1F). These results altogether indicate that topoisomerase inhibition by novobiocin impacts replication initiation specifically from oriC.

The effect of novobiocin on replication initiation is not due to a general block in replication elongation

The consensus in the field is that topoisomerase inhibition primarily effects replication elongation. However, the oriC specificity of the phenotypes we observed following novobiocin treatment suggests that either gyrase or Topo IV is primarily impacting replication initiation. If inhibition of topoisomerases was primarily effecting replication elongation, then oriN cells would also display similar phenotypes to those observed in oriC cells given that the elongation complexes (and the chromosome) are the same in both strains. It is formally possible that topoisomerase inhibition does lead to changes in replication elongation and that these changes indirectly modulate initiation from oriC. In this scenario, either gyrase or Topo IV would not be primarily acting at the origin, and, other elongation inhibitors would also have an oriC-specific effect (there would be over-initiation from oriC but not oriN upon elongation block through any inhibitor of replication).

To further clarify whether the oriC-specific effects of topoisomerase inhibition is due to a replication elongation block, we measured origin-to-terminus ratios for oriC and oriN cells in the presence and absence of the replication elongation inhibitor, Mitomycin C (MMC) (Fig. 1D). We found that cells treated with MMC over-initiate, but this over-initiation phenotype is not specific to oriC (Fig. 1D). In both oriC and oriN cells, origin-to-terminus ratios increased upon treatment with 1 μg/mL MMC: a concentration at which replication is inhibited (42). Although this effect was more pronounced for oriC cells, the increase in oriN cells was statistically significant (P-value<0.01). Together, these results are consistent with our observations that replication elongation is largely unaltered in oriN cells upon topoisomerase inhibition. The results of the oriN and MMC experiments together strongly suggest that arrest of replication elongation is unlikely to be responsible for the over-initiation phenotypes we see in oriC cells upon inhibition of topoisomerases through novobiocin treatment.

Additional experiments were performed to test if novobiocin-induced over-initiation could be leading to replication fork collapse. Microscopy using oriC+ recA-gfp cells grown in LB and in LB supplemented with novobiocin yielded no obvious cell morphology defects (i.e. filamentous cells) (Fig. S2A). The total number of RecA-GFP foci and the total number of cells (DAPI stained nucleoids) were quantified for these two conditions in order to calculate the percentage of cells with RecA foci. We observed no difference in the percentage of cells with RecA-GFP foci, which was roughly 10% under both conditions (Fig. S2B). Together, these results provide evidence against novobiocin inducing a DNA damage response, which would be expected under conditions of replication fork collisions or collapse.

The over-initiation phenotype of novobiocin treated cells is specifically due to gyrase inhibition

Novobiocin inhibits both Topo IV and gyrase, therefore it was unclear if the effects we observed were due to the activity of one or the other (or both) enzymes at the origin. To identify which topoisomerase was responsible for the observed effects of novobiocin inhibition, we plated wild-type B. subtilis cells on 4 μg/mL novobiocin, and isolated a novobiocin resistant mutant of gyrase that contained a single mutation in the gyrB gene, converting Arginine at the 138 position to Leucine. The R138L mutant carries an amino acid change in the ATP-binding domain of gyrase. This mutation is in the same domain and the amino acid change is analogous to previously identified novobiocin resistant gyrase mutants in other bacteria (43, 44). Mutations that lead to novobiocin resistance can sometimes arise in the parE gene, which codes for one of the two subunits of Topo IV (36, 43). We confirmed that there was no mutation in the parE gene in the gyrB (R138L) mutant by sequencing.

We quantified the survival of wild-type and gyrB mutants with increasing concentrations of novobiocin. As expected, the novobiocin resistant mutant can grow on novobiocin, at concentrations which are lethal for wild-type cells. The gyrB mutant grew normally on LB supplemented with 0.45, 0.55, 0.65, 0.75, and 0.85 μg/mL novobiocin, whereas the wild-type strain displayed 1–3-logs (or more) of killing when grown on these concentrations (Fig. 2A).

Figure 2. The over-initiation phenotype of oriC+ cells after topoisomerase inhibition is due to effects on gyrase.

A) Plating efficiency of wild-type and gyrB (R138L) mutant cells on LB or LB supplemented with 0.45, 0.55, 0.65, 0.75, and 0.85 μg/mL novobiocin were calculated. Data shown are averages from 12–24 biological replicates. Error bars represent standard error of the mean. B) Origin-to-terminus ratios of wild-type and gyrB (R138L) cells upon treatment with 0.75 μg/mL novobiocin are shown. Data shown are averages from 6 biological replicates. Error bars represent standard error of the mean.

We then measured origin-to-terminus ratios for wild-type and gyrB mutant cells, in the presence and absence of novobiocin (Fig. 2B). As before, wild-type cells displayed a 2-fold increase in origin-to-terminus ratios; however, the gyrase mutant did not show an increase in origin-to-terminus ratios (Fig. 2B). This indicates that the impact of novobiocin on initiation frequency is through inhibition of gyrase, and is unlikely to be related to inhibition of Topo IV.

Gyrase inhibits DnaA association with oriC

The effect of gyrase inhibition on replication initiation, but not elongation, led us to investigate its impact on DnaA association at oriC. oriN cells do not use DnaA to initiate replication, therefore, one explanation for how gyrase might modulate initiation in an oriC-specific manner is through modulation of DnaA binding or activity. To test this, we performed Chromatin Immunoprecipitations (ChIPs) of DnaA for oriC, oriN, and gyrB (R138L) mutant cells. We measured DnaA association at two different loci within oriC: upstream of dnaA, (PdnaA), and at the DNA unwinding element (DUE) using anti-DnaA rabbit polyclonal antiserum. We normalized the signal for the various loci in oriC to a previously established control locus (9, 45, 46), yhaX, with and without novobiocin treatment (Fig. 3).

Figure 3. Inhibition of gyrase activity increases DnaA association with oriC.

DnaA enrichment was measured at the DUE and PdnaA by ChIP-qPCR (relative to the control locus yhaX). The X’s indicate the absence of the DUE in the oriN strain. DnaA ChIPs were performed from WT, oriN, and gyrB (R138L) cells, grown in the presence or absence of 0.75 μg/mL novobiocin. Data shown are averages from at least 6 biological replicates. Error bars represent standard error of the mean.

In oriC cells, DnaA enrichment at PdnaA and DUE increased roughly two-fold upon treatment with novobiocin (Fig. 3). Interestingly, although the overall levels of DnaA found at oriC were higher in the oriN strain, novobiocin treatment did not lead to any further increase in DnaA enrichment at PdnaA in the oriN background (Fig. 3). DnaA association with the DUE region could not be measured, as the DUE is deleted in this strain background. DnaA ChIPs were also performed in the novobiocin resistant gyrB mutant. Novobiocin treatment did not change DnaA binding patterns at the two loci tested in this strain (Fig. 3). These results suggest that gyrase activity modulates DnaA association with oriC, which is consistent with the over-initiation phenotypes observed in oriC cells.

Given the impact of gyrase inhibition on initiation frequency and DnaA association at oriC, we wanted to test if inhibition of gyrase activity indirectly impacted DnaA through its known regulators. For this, we measured expression levels of several known DnaA regulators with and without novobiocin. RNA levels were quantified for yabA, sirA, spo0A, and dnaA, and normalized to dnaK for oriC+ cells grown in the presence and absence of novobiocin (Fig. S3). We did not detect any changes in expression profiles for these DnaA regulators after novobiocin treatment (Fig. S3).

Initiation from oriN decreases sensitivity to gyrase inhibition

Genome-wide marker frequency, origin-to-terminus ratios, and DnaA ChIP analyses all showed that gyrase inhibition alters DnaA-dependent initiation from oriC, but not RepN-dependent initiation from oriN. Given the importance of well-timed replication initiation and gyrase activity for DNA replication, we were curious if oriC cells are more susceptible to gyrase inhibition than cells initiating replication from oriN. To test this, we measured the survival efficiency of oriC and oriN cells upon exposure to novobiocin. Interestingly, we found that strains initiating from oriN are significantly less sensitive to gyrase inhibition, showing minimal survival defects when plated on concentrations of novobiocin that reduce survival of wild-type cells by up to 3-logs (Fig. 4A). In order to ensure that secondary mutations are not responsible for the increase in novobiocin survival observed among the oriN strains, we sequenced the gyrB and parE genes, where mutations that lead to resistance usually arise in response to novobiocin treatment. No mutations were found in these genes in the oriN strains.

Figure 4. The survival defect observed upon topoisomerase inhibition is oriC-dependent.

Plating efficiencies of oriC and oriN cells on increasing concentrations of novobiocin are presented. A) Colony forming units per mL on LB and LB supplemented with 0.45, 0.55, 0.65, 0.75, and 0.85 μg/mL novobiocin were quantified. Data shown are averages from at least 10 biological replicates. Error bars represent standard error of the mean. B) Plating efficiencies of oriC and oriN cells on LB supplemented with 20 and 30 ng/mL of mitomycin C (MMC) were quantified. Data shown are averages from 4 biological replicates. Error bars represent standard error of the mean.

We confirmed that the decrease in colony forming units on LB supplemented with novobiocin is due to cell death, rather than growth inhibition. To do this, wild-type exponential phase cells were plated on novobiocin. After 24–48 hours of incubation at 30°C, colony forming units were quantified and replica plating was performed to transfer cells to LB plates, which were incubated overnight. Colonies did form in the same places as observed on the antibiotic plates, however, there was no increase in colony forming units on LB plates compared to LB supplemented with novobiocin (data not shown).

To determine if the resistance of oriN cells to novobiocin is related to replication elongation, we performed plating efficiency experiments on MMC. We found that, as expected, oriC cells are sensitive to MMC treatment (Fig. 4B). Importantly, however, unlike what we observed with novobiocin, oriN cells did not show any resistance beyond that of oriC cells to MMC. In fact, oriN cells were 1–2 logs more sensitive to various concentrations of MMC, as compared to oriC cells (Fig. 4B). In the context of our novobiocin experiments, these data again confirm that the oriC-specific growth defects observed upon gyrase inhibition are unlikely to be related to a block in replication elongation.

YabA-DnaA interactions at oriC can counteract gyrase inhibition

YabA is a negative regulator of DnaA that lowers rates of replication initiation through disrupting oligomerization and cooperative binding of DnaA to oriC (46, 47) and tethering DnaA to the replisome (11). Deletion of yabA leads to over-initiation of replication, whereas over-expression of yabA inhibits this process (48). If gyrase has an essential role in regulating DnaA-dependent initiation dynamics, then known regulators of DnaA may be important for modulating the impact of gyrase inhibition on replication initiation and survival. To test this model, using yabA deletion and over-expression mutants, we investigated the impact of YabA on modulating novobiocin susceptibility. In addition, we wanted to determine whether, specifically, the interaction of YabA with DnaA is important for novobiocin susceptibility. For this, we constructed a strain that harbors a mutant of YabA (yabA-aim) containing a previously characterized single point mutation that disrupts interactions between YabA and DnaA (yabA-N85D) (10). The mutant allele of yabA was placed under an IPTG-inducible promoter, in a yabA deletion background. Both yabA-aim expression and yabA over-expression were confirmed by measuring RNA levels (Fig. 5A).

Figure 5. Regulation of initiation by YabA increases the ability of cells to survive gyrase inhibition.

A) RNA levels for yabA (normalized to dnaK) for oriC yabA+, oriC yabA over-expression (P_{spank (hy)} - yabA) with inducer (1 mM IPTG), and oriC P_{spank (hy)} – yabA-aim with inducer (1 mM IPTG) are presented. Data shown are averages from 3 biological replicates. B) Colony forming units per mL of exponentially growing oriC yabA+, oriC ΔyabA, oriC yabA-aim, oriC yabA over-expression, oriN yabA+, and oriN ΔyabA cells plated on LB and LB supplemented with 0.55 μg/mL novobiocin were quantified. Data shown are averages from at least 6 biological replicates. Statistical significance was calculated using t-test (*p<0.05, **p<0.01).

We found that in the absence of YabA, cells are significantly more sensitive to novobiocin. Survival of yabA deletion mutants was 2–3 logs lower than wild-type (Fig. 5B). Conversely, YabA over-expression increased survival efficiency on novobiocin by 1–2 logs (Fig. 5B). These data are consistent with the findings presented above, and suggest that sensitivity to gyrase inhibition is strongly influenced by changes in regulation of DnaA association and activity at oriC.

We found that initiation mutants with deletions in yabA are highly sensitive to novobiocin (Fig. 5B). To further confirm that this is due to the activity of these proteins at oriC, we constructed yabA deletions in the oriN background and quantified survival of these mutants on novobiocin. In contrast to cells initiating replication from oriC, we did not observe any increase in novobiocin sensitivity in yabA deletion mutants initiating replication from oriN (Fig. 5B). The plating efficiency for these mutants remained high, with no detectable decrease in survival, as is seen with cells that initiate replication from oriN (Fig. 5B).

Experiments using the yabA-aim mutants, where YabA-DnaA interactions are disrupted, provided results consistent with the overall hypothesis we have developed based on the data presented above. We grew yabA-aim mutants on increasing concentrations of novobiocin with 1mM IPTG (to induce expression of yabA-aim) and found that loss of YabA-DnaA interactions (yabA-aim) increases novobiocin sensitivity by up to 3 logs (Fig. 5B). This suggests that it is specifically the interaction of YabA with DnaA that is important for modulation of novobiocin survival and not indirect effects found in strains lacking YabA.

DISCUSSION

In vitro work has demonstrated the importance of gyrase and negative supercoiling for melting of the DUE (18, 49). Both our work and previous in vitro studies are in agreement that gyrase activity is important for replication initiation. However, our findings strongly suggest that gyrase is important for replication initiation at an earlier step than DUE melting in vivo. Our results point to a model where gyrase negatively regulates DnaA association with oriC, and decreases replication initiation frequency in B. subtilis.

A previous study reported that novobiocin treatment actually decreases initiation in B. subtilis (50). Ogasawara and colleagues reported that only a limited region near the origin of replication is replicated in the presence of novobiocin (50). This result is not surprising as replication and transcription can both be completely inhibited at high concentrations of novobiocin (32, 34). However, based on the marker frequency patterns presented, had Ogasawara and colleagues measured origin-to-terminus ratios, they would have likely seen an increase in initiation frequency similar to what we found. Nevertheless, at lower concentrations of novobiocin such as those used in our study, the impact of gyrase on initiation is clearly detectable.

There are various ways in which gyrase might regulate DnaA-dependent replication initiation. One possibility is that inhibition of gyrase indirectly affects DnaA binding to the origin through changes in DNA topology at the DnaA consensus binding sites, which are adjacent to the transcriptionally active dnaA gene. The origin of replication must be negatively supercoiled for initiation to proceed. However, given that gyrase introduces negative supercoils into DNA, our data suggest that negative supercoiling beyond what is necessary for DUE melting is actually inhibitory to DnaA association and/or function at oriC. Another possible model for how topology might impact initiation dynamics is through indirect effects of superhelical torsion on replication elongation, which could be communicated to the origin specifically through DnaA regulation. The Grossman group previously reported that in elongation-arrested B. subtilis cells, DnaA association at the origin of replication and copy number of origin-proximal genes increases (44). This is consistent with our findings using MMC. However, in the context of gyrase inhibition, we find that elongation block alone cannot explain the observed effects on DnaA and oriC firing. Cells initiating replication from oriN continue replication elongation unaltered upon gyrase inhibition. Furthermore, unlike with novobiocin inhibition of gyrase, the effects of blocking replication elongation through MMC is not specific to oriC. These results argue against the model that gyrase inhibition targets DnaA-dependent initiation through arrest of replication elongation.

Given the global role of gyrase on chromosome topology, we cannot rule out the possibility that supercoiling effects on DnaA are indirect. However, we did not detect changes in mRNA levels of several regulators of DnaA, including: YabA, SirA, Spo0A, and DnaA upon gyrase inhibition. These results argue against an indirect effect, but do not rule it out. It is possible that the impact of gyrase inhibition on replication initiation is independent of both origin topology and DNA replication. For example, recent work by Magnan and colleagues showed that chromosome tethering at sites as far as 1 Megabases away from the origin of replication can alter global DNA topology and inhibit replication initiation (51). Furthermore, inhibition of gyrase may affect replication initiation indirectly. Further experimentation is required to dissect these models.

Interestingly, it appears that dis-regulation of replication initiation can be lethal for cells if the topological status of the chromosome is compromised. Our studies suggest that over-initiation becomes lethal if positive supercoiling is not resolved around the chromosome. In the case of other regulators such as YabA, over-initiation is not lethal, perhaps because gyrase is able to resolve topological problems that arise away from the origin, in front of converging replication forks.

We propose a model where the topological status of the origin is used as a regulatory mechanism for DnaA binding at an early step prior to the melting of the origin. This model can explain previous observations from studies in E. coli, where changes in DNA supercoiling or transcriptional activity at promoters near oriC were shown to impact replication initiation dynamics (22, 23, 52). Transcriptional activation of gidA, which is adjacent to oriC in E. coli, can alter topology at the origin by introducing negative supercoiling within the region (22, 23). Furthermore, decreased supercoiling was shown to lead to asynchronous replication initiation (53, 54). Additionally, initiation phenotypes observed in the temperature sensitive alleles of dnaA could be suppressed by mutations in topA (Topoisomerase I) (55). All of these observations point to a possible regulatory role for topology in DnaA binding and activity during replication initiation. However, to our knowledge, a role for gyrase in DnaA association and initiation from oriC has not been demonstrated in vivo prior to our study. This role may or may not be direct, and the mechanism DnaA regulation by gyrase should be further investigated.. Additionally, this work establishes a role for topology in DnaA activity at the origin in B. subtilis. Since DnaA, gyrase, and topological constraints are ubiquitously found across bacterial species, we anticipate that regulation of DnaA association or activity by gyrase at oriC is not specific to B. subtilis. Topology-mediated DnaA binding may therefore be one of the few common mechanisms that regulate replication initiation across both Gram-negative and Gram-positive organisms.

EXPERIMENTAL PROCEDURES

Bacterial Culture Conditions

Cells were grown overnight at 37°C on Luria-Bertani (LB) agar plates supplemented with 10 μg/mL chloramphenicol, 100 μg/mL spectinomycin, 5 μg/mL kanamycin, or 500 ng/mL erythromycin and 12.5 μg/mL lincomycin, when appropriate. Cultures were started from single colonies and grown at 30°C or 37°C with aeration (260 rotations per minute), in LB broth, supplemented with 5 μg/mL chloramphenicol, 50 μg/mL spectinomycin, 2.5 μg/mL kanamycin, or 250 ng/mL erythromycin and 6.25 μg/mL lincomycin, when appropriate.

Sporulating cells were prepared by growing cells in 2 mL LB media at 37°C with aeration (260 rotations per minute). Mid-exponential phase cultures (OD 0.3–0.5) were pelleted, LB supernatant was removed, and cells were back diluted (to OD 0.05) into 10 mL of sporulation media, using a previously described recipe (56). Cells were grown for 18.5 hours, harvested in methanol (1:1 ratio), and pelleted for genomic DNA extraction.

Plating Efficiencies

Cultures were started from single colonies and grown at 37°C with aeration (260 rotations per minute) in LB media. Cultures were grown to OD 0.3, and serial dilutions were plated on LB alone and LB supplemented with varying concentrations of novobiocin. Plates were incubated at 30°C. Total viable cells were quantified after 24–72 hours of incubation. Novobiocin concentrations used include: 0.45, 0.55, 0.65, 0.75, and 0.85 μg/mL. Mitomycin C concentrations used include: 20 and 30 ng/mL.

Origin-To-Terminus Ratios and Marker Frequency Analysis

Cultures were started from single colonies and grown at 37°C with aeration (260 rotations per minute) in LB media. Cultures were grown to exponential phase (optical density 0.3–0.5), set back to OD 0.05 in LB media and grown at 30°C. Cultures were grown to OD 0.2, divided into 5mL cultures with no treatment, 0.50 μg/mL novobiocin, 0.75 μg/mL novobiocin, and 1 μg/mL Mitomycin C. After 40 minutes of growth at 30°C, cells were harvested in methanol (1:1 ratio) and pelleted. Genomic DNA was isolated by phenol-chloroform extractions. The copy number of the origin and terminus were quantified by quantitative PCR (qPCR). QPCR was done using SSoAdvanced SYBR Green master mix and CFX96 Touch Real-Time PCR system (Bio-Rad).

Primers used to amplify the origin and terminus were the same as previously described (57). Quantification of oriC was done using primers: HM1583 (5′ – GATCAATCGGGGAAAGTGTG – 3′) and HM1584 (5′ – GTAGGGCCTGTGGATTTGTG – 3′). Quantification of the origin in strains that initiate replication at oriN was done using primers: HM2510 (5′ – GAATTCCTTCAGGCCATTGA – 3′) and HM2511 (5′ – GATTTCTGGCGAATTGGAAG – 3′). Quantification of the terminus region was done using primers: HM1585 (5′ – TCCATATCCTCGCTCCTACG – 3′) and HM1586 (5′ – ATTCTGCTGATGTGCAATGG – 3′). Origin-to-terminus ratios were determined by dividing the number of sequence reads (as indicated by the Cq values measured through qPCRs) from the origin by the number of sequence reads quantified at the terminus. Ratios were normalized to the origin-to-terminus ratio of sporulating B. subtilis cells, which were quantified using primers HM1583 and HM1584.

For marker frequency analysis, quantification of DNA at six additional chromosomal positions was measured: 22.5°, 45°, 135°, 225°, 315°, 337.5°. Primers HM4853 (5′ – CCTTATCGTTCGGTATCGTC – 3′) and HM4854 (5′ – GCTTTGCAATGCGC TTG – 3′) amplify DNA at position 22.5°. HM3011 (5′ – GTCGAGATGGTGGCGATCG – 3′) and HM3012 (5′ –CGCGGCATGTCTCTGAGTAC – 3′) amplify DNA at position 45°. Primers HM3015 (5′ – GATACGGCATTACAGCATGC – 3′) and HM3016 (5′ – GCTTTTAATCAGAATGAGCTGTCC – 3′) amplify DNA at position 135°. Primers HM3017 (5′ – GTGCTCACTGAAGACGATCTTCCC – 3′) and HM3018 (5′ – CATCTTCTTGAAGGGTTCCGAC – 3′) amplify DNA at position 225°. Primers HM3021 (5′ – CGTTTGTAAGAGGGGCGCACC – 3′) and HM3022 (5′ – GGTGATTGCGTCATGATCCGTACC – 3′) amplify DNA at position 315°. Primers HM4855 (5′ – GCTGTCAAGTGGACATGTC – 3′) and HM4856 (5′ – GTATTCGCGGTGTGAAAACCTTG – 3′) amplify DNA at position 337.5°. Marker frequency ratios were determined by dividing the number of sequence reads quantified at each site (as indicated by the Cq values measured through qPCRs) by the number of sequence reads quantified at the terminus (using primers HM1585 and HM1586). These values were normalized to ratios measured for sporulating B. subtilis cells.

Chromatin Immunoprecipitations

Cultures were started from single colonies and grown at 37°C with aeration (260 rotations per minute) in LB media. Cultures were grown to exponential phase (optical density 0.3–0.5), set back to OD 0.05 in 25 mL LB media and grown at 30°C. After reaching OD 0.2, cultures were grown for 40 more minutes at 30°C with no treatment or 0.75 μg/mL novobiocin. Cultures were processed for ChIPs as described (58). ChIPs were performed using 2 μl anti-DnaA rabbit polyclonal antiserum (59). Quantitative PCR was done using SSoAdvanced SYBR green master mix and Bio-Rad CFX96 Touch Real-Time PCR system. DnaA association at the DNA unwinding element (DUE) was normalized to DnaA association at yhaX, a control locus that does not have increased DnaA association. Enrichment of the DUE was detected using primers HM459 (5′-GGGAAAGTGTGAATAACTTTTCG -3′) and 460 (5′ - GTAGGGCCTGTGGATTTGTG -3′). Enrichment upstream of dnaA (referred to in the text as PdnaA) was done using primers HM457 (5′ - ACCATTGCAAGCTCTCGTTT - 3′) and 458 (5′ - CCACACTTTGTGGATAAAGAGGA - 3′). Enrichment of yhaX was detected using primers HM192 (5′-CCGTCTGACCCGATCTTTTA-3′) and HM193 (5′-GTCATGCTGAATGTCGTGCT-3′).

DNA-Sequencing

Cultures were prepared as described under the “origin-to-terminus ratios” section. Whole genomic DNA was sonicated using a Covaris ultrasonicator and sequenced on an Illumina Next-Seq yielding approximately 15M reads. Reads were mapped to the B. subtilis strain JH642 (GenBank: CP007800.1) genome using Bowtie 2 (60). The resulting sam file was processed using SAMtools, view, and sort functions. PCR and optical duplicates were then removed using Picard (61). The resulting .sam file was processed by SAMtools mpileup functions to produce wiggle plots. To account for differences in read depth between samples, the signal at each base position was normalized to total signal for the genome. The resulting total-normalized wiggle files were then visualized in MochiView.

Quantification of RNA levels

RNA levels were quantified by growing cultures from single colonies at 37°C with aeration (260 rotations per minute) in LB media. Cultures were grown to exponential phase (optical density 0.3–0.5), set back to OD 0.05 in LB media (with and without 1mM IPTG), and grown to OD=0.3. Cultures were mixed with ice-cold Methanol (1:1 ratio). Cells were pelleted and RNA was extracted using the Thermo Scientific GeneJET RNA Purification kit. DNase was added to 1μg RNA samples for 30 minutes at 37°C, and inactivated by adding 1μL EDTA to the reaction and incubating at 65°C for 10 minutes. Reverse Transcriptase PCR was performed to make cDNA. Samples were diluted 1:5 and qPCR was performed using the housekeeping gene dnaK as a control locus. QPCR was done using SSoAdvanced SYBR Green master mix and CFX96 Touch Real-Time PCR system (Bio-Rad). Primers to amplify dnaK were HM770 (5′ - TCTCCAGCTGTGATAAACGGTA – 3′) and HM771 (5′ - AAAACGGCATTGATTTGTCA – 3′). Primers to amplify yabA were HM2967 (5′-GCAGCTGGGGGATTTGAAGC -3′) and HM2968 (5′-GTCCAGCCGTTTGCGCAAATG -3′). Primers to amplify sirA were HM2995 (5′-GCCAATCACTATTTCGGCCGGG -3′) and HM2996 (5′-GCTGCTTTTCAAGGCTTGTCC -3′). Primers to amplify spo0A were HM4849 (5′-TAACGGACAGGAATGCCTGT -3′) and HM4850 (5′ACGGCCTTTTTCGTGACATC -3′). Primers to amplify dnaA were HM4654 (5′-CCT GTGGAACCAA GCCCTTGCTC -3′) and HM4655 (5′-GGCAAATTCATTGGGAGCCGTG -3′).

Replica Plating

Wild-type exponential phase cells were plated on novobiocin. After 24–48 hours of incubation at 30°C, colony forming units were quantified. Using sterile velvet cloths, cells were transferred to LB plates and incubated overnight. The number and location of colonies was recorded and compared to what was observed on the antibiotic plates.

Microscopy

Cultures were started from single colonies and grown at 37°C with aeration (260 rotations per minute) in LB media. Cultures were grown to exponential phase (optical density 0.3–0.5), set back to OD 0.05 in LB media and grown at 30°C. Cultures were grown to OD 0.2, divided into 5 mL cultures with no treatment, 0.50 μg/mL novobiocin. Cells were then fixed with 4% formaldehyde (vol/vol), stained with DAPI (4-,6-diamidino-2-phenylindole), and transferred to 1% agarose pads for visualization by microscopy. Images were taken using a Leica inverted microscope with CCD camera fitted with a 60× oil objective. DAPI fluorescence was used to quantify total cells and GFP fluorescence was used to quantify total RecA-GFP foci. DAPI-stained nucleoids and RecA-GFP foci were counted using ImageJ. At least 2,000 cells from 3 biological replicates were counted per condition.

Strain Construction

New strains were confirmed by PCR and sequencing. HM89 was constructed using site-directed mutagenesis by a previously described method (62). Primers HM20 (5′ – CGAAAAGAGGATTGTCCTTTCTGTCTGTCATTC – 3′) and HM21 (5′ –GAATGACAGACAGAAAGGACAATCCTCTTTTCG – 3′) were used to make the point mutation. This point mutation has been made and characterized by Noirot-Gros and colleagues (63). The gyrase mutant (HM3387) was isolated by plating a large saturated culture of wild type B. subtilis onto LB supplemented with 4 μg/mL novobiocin. The mutation in gyrB, R138L, was identified by amplification of gyrB by PCR, followed by sequencing. The parE gene was also amplified and sequenced. As expected, no mutations in this gene were found.

Table 1.

Strains used in this study.

Strain	Relevant Genotype	Reference
HM1 (JH642)	trpC2, pheA1 (Wild Type)	Perego et al., 1988
HM2 (AIG109)	trpC2, pheA1 yabA::cat	Goranov et al., 2009
HM3 (AIG80)	trpC2, pheA1, amyE::[ Pspank(hy)- yabA specR ]	Goranov et al., 2009
HM77 (MMB170)	trp+ pheA1 122 spoIIIJ::[ oriN repN kan ] oriC-S	Goranov et al., 2009
HM78 (AIG185)	trp+ pheA1 122 spoIIIJ::[ oriN repN kan ] oriC-S yabA::cat	Goranov et al., 2009
HM89	yabA::cat amyE::[ Pspank(hy)-yabA(N85D) specR ]	This study
HM313 (LAS40)	recA-mgfp-mut2(A206K)-spec	Simmons et al., 2007
HM3387	trpC2, pheA1, gyrB(R138L)	This study