Inhibition of Escherichia coli chromosome replication by rifampicin treatment or during the stringent response is overcome by de novo DnaA protein synthesis

Abstract Initiation of Escherichia coli chromosome replication is controlled by the DnaA initiator protein. Both rifampicin‐mediated inhibition of transcription and ppGpp‐induced changes in global transcription stops replication at the level of initiation. Here, we show that continued DnaA protein synthesis allows for replication initiation both during the rifampicin treatment and during the stringent response when the ppGpp level is high. A reduction in or cessation of de novo DnaA synthesis, therefore, causes the initiation arrest in both cases. In accordance with this, inhibition of translation with chloramphenicol also stops initiations. The initiation arrest caused by rifampicin was faster than that caused by chloramphenicol, despite of the latter inhibiting DnaA accumulation immediately. During chloramphenicol treatment transcription is still ongoing and we suggest that transcriptional events in or near the origin, that is, transcriptional activation, can allow for a few extra initiations when DnaA becomes limiting. We suggest, for both rifampicin treated cells and for cells accumulating ppGpp, that a turn‐off of initiation from oriC requires a stop in de novo DnaA synthesis and that an additional lack of transcriptional activation enhances this process, that is, leads to a faster initiation stop.


| INTRODUC TI ON
In Escherichia coli chromosome replication commences from a single origin of replication, oriC. The oriC region contains an AT-rich region as well as multiple binding sites for the DnaA initiator protein (Leonard and Mechali, 2013). The DnaA protein belongs to the family of AAA + (ATPases associated with diverse cellular activities) proteins. DnaA binds ATP and ADP with similar affinities (Sekimizu et al., 1987), but only the ATP-bound form, DnaA ATP , is active in replication initiation (Skarstad and Katayama, 2013). At initiation DnaA ATP oligomerizes on oriC in a highly ordered manner, which leads to origin unwinding in the AT-rich region, and subsequent loading of two replisomes to direct chromosome replication in a bidirectional manner, until terminated at the opposite replication terminus region, terC.
Initiation of chromosome replication is a highly regulated step in the E. coli cell cycle during which all cellular origins are simultaneously initiated once and only once per cell cycle .
This tight control is ensured mainly by fluctuations in the cellular Rifampicin has been known to inhibit replication initiation for decades (von Meyenburg et al., 1979), while having no effect on replication elongation. This feature is often employed for determination of the number of replication origins per cell by flow cytometry (Skarstad et al., 1986). The exact mechanisms by which rifampicin inhibits replication initiation remain unclear, but two models are frequently proposed. In one suggestion, a consequence of transcription inhibition is that de novo DnaA protein synthesis ceases, and hence initiation is inhibited. Inhibition of protein synthesis by chloramphenicol (Vazquez, 1979) is in agreement with this hypothesis. Another suggestion is that transcriptional activation of oriC is required for replication initiation and this process is inhibited by rifampicin. Specifically, transcriptional activity of promoters located with close proximity to or within oriC is known to alter the DNA structure in such a way that facilitates DNA duplex unwinding, either by creating a stable R-loop, or by introducing negative supercoils in the wake of the migrating RNA polymerase complex, thereby stimulating replication initiation . Transcription units located close to oriC include promoters of the highly conserved mioC and gidA genes. Given their opposite orientations, mioC transcripts read toward oriC introducing initiation-inhibiting positive supercoils, whereas gidA transcripts read away from oriC introducing initiation-activating negative supercoils (Theisen et al., 1993;Bates et al., 1997). Some mioC transcripts will progress completely through oriC (Rokeach and Zyskind, 1986).
Initiation of chromosome replication is also inhibited by accumulation of the secondary messenger guanosine tetra-or penta-phosphate, ppGpp, and pppGpp, respectively (together abbreviated as ppGpp below (Levine et al., 1991, Schreiber et al., 1995, Ferullo and Lovett, 2008). However, as with rifampicin, the exact mechanism by which ppGpp accumulation contributes to this inhibition is not well understood. ppGpp accumulates during stressful conditions, such as amino acid starvation, iron limitation, heat shock, and antibiotic treatment in what is collectively referred to as the "stringent response" (Dalebroux and Swanson, 2012;Hauryliuk et al., 2015).
Two enzymes, RelA and SpoT, control the levels of ppGpp. The main ppGpp synthetase, RelA, synthesizes ppGpp when associating with ribosomes in the ribosomal acceptor site (A-site) to sense uncharged tRNAs during amino acid starvation (Cashel et al., 1996;Potrykus and Cashel, 2008;Hauryliuk et al., 2015). SpoT, moreover, exhibits dual functions and is able to synthesize ppGpp in response to starvation of other nutrients, such as iron, fatty acids, and certain carbon sources, as well as being responsible for hydrolyzing ppGpp and resetting cells after ppGpp induction (Sarubbi et al., 1989). Activation of the stringent response causes significant alterations in global gene expression, which are driven by the direct binding of ppGpp to two sites in the bacterial RNAP located at the interface of the ω and β' subunits (Ross et al., 2013;. Binding of ppGpp to RNAP results in an extensive transcriptional reprograming including downregulation of genes involved in stable RNA (ribosomal RNA; rRNA, and transfer RNA; tRNA, respectively), synthesis, DNA replication, transcription, translation, and various metabolic pathways, while genes related to stress and amino acid biosynthesis are upregulated (Cashel et al., 1996;Haugen et al., 2008;Potrykus and Cashel, 2008;Traxler et al., 2008;Kanjee et al., 2012;Hauryliuk et al., 2015). Overall, these transcriptional changes assist bacteria in entering a physiologically state that enable them to survive the stressful conditions.
The effects of ppGpp on replication initiation is argued to be indirect, that is, a result of changes in general cell physiology and rate of stable RNA synthesis (Hernandez and Bremer, 1993). However, dnaA is among the genes downregulated by ppGpp with the transcriptional activity of both dnaA gene promoters, dnaAp1 and dnaAp2 (Hansen et al., 1982), being reduced in a RelA-dependent manner (Chiaramello and Zyskind, 1990;Zyskind and Smith, 1992). Lack of de novo DnaA synthesis or even degradation (Gross and Konieczny, 2020) during ppGpp accumulation could therefore contribute to, or even be responsible for, the cessation of initiation seen in the presence of ppGpp. A recent study further suggests that a ppGpp-driven reduction in transcripts from promoters located close to oriC, leading to less negative supercoiling of oriC, that is, less transcriptional activation, could explain the negative effect of ppGpp on replication initiation (Kraemer et al., 2019).
Here, we address the role of DnaA limitation in rifampicin-and ppGpp-mediated inhibition of DNA replication initiation. We show that driving dnaA gene transcription continuously from a rifampicin-resistant and ppGpp-insensitive T7 RNAP-dependent promoter, p T7 , enables cells to continue replication initiation in the presence of both rifampicin and RelA-synthesized ppGpp.

| Continuous synthesis of DnaA allows initiation of chromosome replication in the presence of rifampicin
To determine whether rifampicin blocks replication initiation indirectly through a stop in de novo DnaA protein synthesis, we constructed a MG1655 based strain where DnaA is produced even if rifampicin is present. Briefly, this strain carried a chromosomal λDE3 with the rifampicin-resistant bacteriophage T7 RNAP under control of the isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible placUV5 promoter (Tseng et al., 2009), and a plasmid, pA-CYC184-T7-DnaA, containing the dnaA gene under the control of a T7 RNAP-dependent promoter, p T7 (Tabor, 2001) (Figure 1a). Cells were treated with IPTG for 10 min to allow for synthesis of T7 RNAP prior to addition of rifampicin and cephalexin ( Figure 2a). Because rifampicin blocks E. coli RNAP, but not the T7 RNAP, DnaA was the sole protein synthesized during rifampicin treatment and it accumulated approximately 16-fold over 2 hr (Figure 1b). Cephalexin prevented residual cell division during rifampicin treatment.
Wild-type cells treated with rifampicin and cephalexin ceased initiations while replication elongation continued. Consequently they ended up containing integral numbers of fully replicated chromosomes over time ( Figure 2b). Also, the origin-to-terminus (ori/ter) ratio, reflecting the copy number ratio of oriC to terC, approached 1 over time indicating that cells had finished on-going replication and carried a complete set of fully replicated chromosome(s), also defined as run out replication (Figure 2c; black bars). The cellular DNA content increased somewhat as a result of run out replication ( Figure 2d; black bars). No decrease in DnaA protein level was observed in agreement with the DnaA protein being stable (Torheim et al., 2000). Cells that continued to express DnaA in the presence of rifampicin and cephalexin behaved quite different. No clear run out replication phenotype was observed (Figure 2b), the ori/ter ratio increased after drug addition, representing cells with on-going chromosome replication (Figure 2c; blue bars), and the cellular DNA content continued to increase for 120 min, which was the duration of the experiment (Figure 2d; blue bars).
These data show that continuous de novo synthesis of DnaA is sufficient to initiate new rounds of DNA replication in the presence of rifampicin. As rifampicin blocks initiation of global RNA synthesis, including transcription from promoters located in or around oriC, transcriptional activation of the origin is not necessary for initiation to take place under these conditions. This suggests that rifampicin inhibits initiation of DNA replication primarily through preventing de novo DnaA protein synthesis.
F I G U R E 1 DnaA is synthesized from a T7 RNAP-dependent promoter during rifampicin treatment. (a) Schematic overview of the applied molecular system. The bacteriophage T7 RNA polymerase (T7 RNAP) is resistant to rifampicin (Tabor, 2001). IPTG-mediated induction of the lacUV5 promoter leads to expression of the chromosomally encoded T7 RNAP, which induce transcription of dnaA from promoter p T7 . During rifampicin treatment, global RNA synthesis initiation is blocked, whereas the dnaA gene is continuously expressed under the control of p T7 in the presence of IPTG. (b) DnaA protein content determined by Western blot analysis for cells carrying the "empty" pACYC184-T7 plasmid (Wild type; left) or the DnaA expression plasmid, pACYC184-T7-DnaA (T7-dnaA; right), at the indicated times. Cells were grown at 37°C in AB minimal medium supplemented with glucose and casamino acids. At time −10 min 1 mM IPTG was added to induce T7 RNAP synthesis, and at time 0 min rifampicin and cephalexin was added to block initiation of RNA synthesis and cell division, respectively. Samples were adjusted to the same total protein content before loading onto the Western blot. All quantifications, normalized to pretreated wildtype cells (−11 min), include standard deviations and are based on triplicates.

| Continuous synthesis of DnaA allows initiation of chromosome replication in the presence of ppGpp
For decades, initiation of chromosome replication has been known to be inhibited by accumulation of ppGpp during the stringent response (Levine et al., 1991;Schreiber et al., 1995;Ferullo and Lovett, 2008). dnaA is downregulated by ppGpp (Chiaramello and Zyskind, 1990), thus it was investigated whether accumulation of ppGpp inhibits replication initiation by preventing de novo DnaA protein synthesis, or by preventing transcriptional activation of oriC as previously suggested (Kraemer et al., 2019), or both.
To address this question, we used the pACYC184-T7-DnaA expression plasmid described above and constructed a compatible plasmid, pBR322-RelA * , where the active 455 amino acid N-terminal relA fragment (Svitil et al., 1993) is controlled by the IPTG inducible p A1/04-03 promoter (Lanzer and Bujard, 1988). The use of these two plasmids alone or together in the MG1655 λDE3 strain allows for overproduction of DnaA, ppGpp, or both in the same cells in an IPTG dependent manner (Figure 3a). The simultaneous production of ppGpp and DnaA was possible as the dnaA gene is transcribed from the ppGpp-insensitive T7 RNAP-dependent promoter, p T7 (Friesen and Fiil, 1973;Tabor, 2001). This allowed us to study the effect of  (Chiaramello and Zyskind, 1990) and/or from a decrease in DnaA protein translation F I G U R E 2 de novo DnaA synthesis allows replication initiation in the presence of rifampicin. (a) Overview of the timeline of the experiment. (b) Cells of strain ALO6511 carrying the "empty" pACYC184-T7 plasmid (Wild type; upper panel) or the DnaA expression plasmid, pACYC184-T7-DnaA (T7-dnaA; lower panel), were grown at 37°C in AB minimal medium supplemented with glucose and casamino acids. At time −10 min 1 mM IPTG was added to induce T7 RNAP and hence DnaA synthesis, and at time 0 min rifampicin and cephalexin was added to block initiation of RNA synthesis and cell division, respectively. Samples for flow cytometry analysis and quantitative PCR (qPCR) were removed at the indicated timepoints. (c) Origin-to-terminus ratios for cells of strain ALO6511 carrying the "empty" pACYC184-T7 plasmid (Wild type) or the DnaA expression plasmid, pACYC184-T7-DnaA (T7-dnaA) at the indicated time points. (d) Cellular DNA content (DNA/cell) for cells of strain ALO6511 carrying the "empty" pACYC184-T7 plasmid (Wild type) or pACYC184-T7-DnaA (T7-dnaA) at the indicated time points. DNA/cell is normalized to wild-type cells at t = 0 min (the time of rifampicin and cephalexin addition). Standard deviations are based on duplicates.
rate caused by ppGpp-mediated inhibition of promoters for rRNA and tRNA synthesis (Potrykus and Cashel, 2008). We find it unlikely that ppGpp-induced degradation of DnaA (Gross and Konieczny, 2020) affected its accumulation as we observed no decrease in DnaA level when ppGpp was increased in otherwise wild-type cells  (Figure 4e; gray bar). This is in agreement with previous observations (Ferullo and Lovett, 2008;Kraemer et al., 2019). The replication elongation rate did not seem significantly affected within the timeframe of the experiment, as also seen in earlier studies (Levine et al., 1991;Ferullo and Lovett, 2008;Denapoli et al., 2013). Moreover, overproduction of DnaA resulted in an increase in cellular origin content along with an inability to finish replication in the presence of rifampicin ( Figure 4c). Both the ori/ter ratios increased dramatically (Figure 4e; dark blue bars) and the cellular DNA content also increased ( Figure 4f; dark blue bars). This demonstrates that additional DnaA triggers excessive replication initiation as also previously observed (Atlung et al., 1987;Skarstad et al., 1989).
To address additional role(s) of ppGpp on replication initiation ppGpp synthesis was induced in cells containing a mutant RNAP that is non-responsive, that is, does not bind ppGpp due to several amino acid substitutions in RpoC as well as deletion of amino acids 2-5 in RpoZ (referred to as −1−2 (Ross et al., 2016)). This allowed the clarification of whether the effect of ppGpp on replication initiation was mediated through RNAP-driven transcriptional changes only, or whether there is an additional direct effect of ppGpp on replication proteins.
Induction of RelA * synthesis in wild-type cells slowed down growth significantly, most likely due to ppGpp accumulation, whereas growth of the −1−2 strain was largely unaffected by an increase in ppGpp level (Figure 5a,b). Accumulation of ppGpp resulted in cessation of replication initiation leading to complete run out replication over time in wild type (+1+2), but not in −1−2, cells ( Figure 5c). Further, rifampicin and cephalexin treatment of −1−2 cells showed that ppGpp accumulation neither affected the cellular number of replication origins nor the synchrony of initiation to any significant level (Figure 5c; insert). Therefore, the inhibiting effect of ppGpp on DNA replication initiation is mediated indirectly through RNAP-driven reprograming of transcription and not through altering the activity of replication proteins. Thus, a change in one or more ppGpp-regulated gene transcripts works in concert with a lowered DnaA level to limit replication initiation during ppGpp accumulation.

| Chromosome replication initiation is more sensitive to transcriptional than translational inhibition
We proceeded to test whether the efficient turn-off of replication initiation imposed from rifampicin resulted from its effect on protein, that is, DnaA, synthesis only, or whether the inhibitory effect was enhanced by inhibition of additional transcripts as well. The kinetics of replication initiation inhibition of rifampicin was compared to that of chloramphenicol; a bacteriostatic and broad-spectrum antibiotic that acts on the 50S ribosomal subunit to inhibit translation without affecting transcription (Vazquez, 1979).
Previously, rifampicin and chloramphenicol have been shown to shut down transcription and translation, respectively, within 5-10 s after addition to highly permeable AS19 cells (Pato and Von Meyenburg, 1970;Pato et al., 1973).
Rifampicin and cephalexin or chloramphenicol and cephalexin was added to exponentially growing cultures of AS19 cells, respectively, and the residual DNA synthesis was followed over time ( Figure 6). Because cephalexin inhibits cell division, the number of fully replicated chromosomes per cell following run out synthesis will depend on how fast each antibiotic inhibited replication initiation, whereas the time required to finish replication will depend on how the replication elongation rate is affected. Cells treated with rifampicin and cephalexin finished replication approximately 60-90 min after drug addition, and contained mainly two, four, and eight fully replicated chromosomes with an average of 4.1 cellular origins (Figure 6; left). Cells treated with chloramphenicol and cephalexin ended up with primarily four and eight fully replicated chromosomes, and the average number of cellular origins was 5.1, that is, higher than for rifampicin treated cells (Figure 6; right), which F I G U R E 6 Inhibition of transcription inhibits replication initiation faster than inhibition of translation. E. coli AS19 cells were grown exponentially at 37°C in AB minimal medium supplemented with glucose and casamino acids. At time 0 min the culture was split in two. One subculture received rifampicin and cephalexin ("RIF + CEPH," left) and one received chloramphenicol and cephalexin ("CAM + CEPH," right) to inhibit initiation of RNA synthesis and cell division, or protein synthesis and cell division, respectively. Samples for flow cytometry analysis were removed at the indicated times. Samples representing a complete replication run out were removed 4 hr after drug addition F I G U R E 5 The effect of ppGpp on replication initiation is mediated through RNAP-driven transcriptional changes. MG1655 cells with either the wild-type RNAP (+1+2; ALO5045) or the ppGpp-insensitive mutant RNAP (−1−2; ALO6798) carrying either the "control" plasmid with the relAgene expressed downstream the p A1/04-03 promoter, pBR322-RelA -(relA -), or the RelA expression plasmid with the relA * gene expressed downstream the p A1/04-03 promoter, pBR322-RelA * (relA * ), were grown at 37°C in AB minimal medium supplemented with glucose and casamino acids. At time 0 min 1 mM IPTG was added to induce RelA synthesis. (a) Growth curves for +1+2 cells carrying pBR322-RelA -(black) or pBR322-RelA * (gray), and for −1−2 cells carrying pBR322-RelA -(dark blue) or pBR322-RelA * (light blue). Optical density at 450nm (OD 450 ) is depicted as a function of time in a semilogarithmic diagram. Time of addition of 1 mM IPTG is indicated. Generation times before and after IPTG addition are indicated in the table. (b) Quantification of (p)ppGpp accumulation (total amount of ppGpp and pppGpp) in +1+2 and −1−2 cells carrying either pBR322-RelAor pBR322-RelA * at the given times. Autoradiograms were used to calculate and visualize the fractional content of (p)ppGpp as a function of time. Standard deviations are based on three technical replicates. (c) Samples of exponentially growing +1+2 and −1−2 cells carrying either pBR322-RelAor pBR322-RelA * were removed at the given times for direct analysis by flow cytometry (marked by blue). Samples of exponentially growing +1+2 and −1−2 cells carrying either pBR322-RelAor pBR322-RelA * were removed at the given times for treatment with rifampicin and cephalexin prior to flow cytometry analysis (inserts).
indicates that rifampicin inhibits replication initiation in a faster mode than chloramphenicol. Chloramphenicol treated cells took longer (≥120 min) to finish replication most likely due to chromatin condensation (Zimmerman, 2002), hence slowing down the chromosome replication rate. Similar results were obtained with MG1655 (Supporting Information Figure S1).
Accumulation of DnaA ceases immediately in chloramphenicol treated cells due to inhibition of translation and somewhat later in rifampicin treated cells due to the residual capacity for synthesis from the dnaA mRNA present at the time of rifampicin addition. Yet, chloramphenicol treated cells continued initiation for a longer period relative to rifampicin treated cells. As transcription is still on-going in chloramphenicol treated cells, this suggests that the efficient inhibition of chromosome replication initiation seen for rifampicin treated cells is caused via both a stop in de novo protein, that is, DnaA synthesis, and an arrest of transcription. This also explains why overinitiation caused by T7 RNAP-directed DnaA synthesis was lowered in cells treated with rifampicin relative to untreated cells where transcription continues (compare Figures 2 and 4). This indicates that reduced/altered transcription of genes in addition to dnaA contribute to initiation inhibition.

| Replication initiation and de novo DnaA synthesis
In E. coli, initiation of chromosome replication is inhibited during rifampicin treatment (von Meyenburg et al., 1979), which blocks initiation of global RNA synthesis through direct binding to-and inhibition of RNAP. Here, we found that de novo synthesis of DnaA from the rifampicin-insensitive T7 RNAP-dependent p T7 promoter was sufficient to continue replication initiation in the presence of rifampicin. This is in agreement with an earlier report where high levels of DnaA allowed for some additional initiations from oriC in the presence of rifampicin (Pierucci et al., 1987). This underlines the importance of de novo DnaA protein synthesis for initiation and suggests that rifampicin inhibits this process by preventing dnaA gene transcription and hence de novo DnaA protein synthesis.
While accumulation of ppGpp in wild-type cells resulted in initiation inhibition (Levine et al., 1991;Ferullo and Lovett, 2008), this was not the case for cells carrying a mutated RNAP non-responsive to ppGpp (i.e., −1−2; (Ross et al., 2016)). Therefore, ppGpp imposed changes in RNAP-driven transcription of one or several genes are responsible for inhibition of initiation, whereas there is no evidence for a direct ppGpp effect on proteins involved in replication initiation as also reported previously (Kraemer et al., 2019). This is in line with findings that DnaA is not a direct target for ppGpp (Zhang et al., 2018;Wang et al., 2019). de novo DnaA protein synthesis directed from the ppGpp-insensitive T7 RNAP-dependent p T7 promoter was sufficient to sustain continued replication initiation during ppGpp accumulation. Therefore downregulation of dnaA gene transcription (Chiaramello and Zyskind, 1990) and reduced de novo DnaA protein synthesis is responsible for replication initiation inhibition by ppGpp in wild-type cells. We suggest that high levels of ppGpp inhibit replication initiation in a manner similar to rifampicin, that is, mainly by halting de novo DnaA protein synthesis.

| Importance of DnaA activity
Recently it was found that overproduction of DnaA induced overinitiation in E. coli, but subsequent induction of ppGpp synthesis still inhibited replication initiation under these conditions (Kraemer et al., 2019). This would seemingly contradict the present finding that continued de novo DnaA synthesis allows for new rounds of replication initiation during accumulation of ppGpp and during rifampicin treatment. A major difference between the two experimental setups is that in the previous report (Kraemer et al., 2019), DnaA was overproduced prior to ppGpp accumulation, while it is overproduced during ppGpp accumulation in the data presented here. Overproduction of DnaA in wild-type cells induces a burst in initiations (Atlung et al., 1987;Skarstad et al., 1989), which can be explained by de novo DnaA protein being mainly ATP bound and thus active. This is due to ATP being more abundant than ADP within the cell (Petersen and Møller, 2000), and because DnaA binds these nucleotides with similar affinity (Sekimizu et al., 1987). These initiations accelerate processes that convert DnaA ATP to DnaA ADP , and a new steady state is reached where replication initiation takes place at a slightly reduced mass, and without any significant increase in the DnaA ATP -to-DnaA ADP ratio (Flatten et al., 2015); that is, the activity of DnaA for initiation action is unchanged or only moderately increased. Moreover, overproduction of mainly ATP-bound DnaA in RIDA deficient cells leads to massive overinitiation and inviability (Riber and Lobner-Olesen, 2005;Fujimitsu et al., 2008). Based on these observations, we suggest that it is primarily the activity of DnaA, that is, DnaA ATP -to-DnaA ADP ratio, which determines whether or how fast initiations cease in response to rifampicin treatment or an elevated ppGpp level, and that both conditions fail to arrest replication initiation when this ratio is high. In agreement with this, both rifampicin treatment and ppGpp fail to arrest initiations in cells that overproduce a hyperactive DnaA protein that mimics DnaA in its ATP bound form (Simmons et al., 2004, Kraemer et al., 2019. When DnaA is de novo synthesized in the presence of either high ppGpp or rifampicin as presented here, the pool of DnaA ATP is continuously replenished and the cellular DnaA ATP -to-DnaA ADP ratio increased. An increased ATP level during the stringent response and during rifampicin treatment (Lobritz et al., 2015;Varik et al., 2017) could add to the increase. As the level of T7-mediated DnaA overproduction was quite high it may overpower the ability of RIDA and DDAH to reduce the DnaA ATP -to-DnaA ADP ratio, which could even further add to increasing the overall DnaA ATP -to-DnaA ADP ratio.
Consequently, cells with continued de novo DnaA synthesis are able to continue initiation, at least for a while, both during ppGpp accumulation and during rifampicin treatment.
These observations underline the importance of DnaA activity and indicate that initiation inhibition by ppGpp accumulation or rifampicin treatment results, at least in part, from failure to replenish the DnaA ATP pool.

| Transcriptional activation of the origin
Recently, it was suggested that ppGpp inhibits replication initiation mainly by reducing the negative superhelicity of oriC, either by reduced transcription of the mioC and gidA genes adjacent to oriC (Kraemer et al., 2019), or by reducing gyrase and topoisomerase IV expression (Fernandez-Coll et al., 2020). Transcriptional activation is, however, not a requirement for replication initiation in wild-type cells, and both mioC and gidA promoters can be deleted without measurable effects (Lobner-Olesen and Boye, 1992;Bates et al., 1997;Lies et al., 2015).

Moreover, transcriptional activation becomes important for initiation
during suboptimal conditions, such as when the chromosomal oriC is truncated by deletion of DnaA binding site R4 (Bates et al., 1997).
Here we show that accumulation of ppGpp prevents replenishment of the DnaA pool, presumably by reducing dnaA gene transcription. DnaA therefore becomes limiting for initiations. We suggest that transcriptional activation of oriC can assist during these suboptimal conditions and delay the replication initiation arrest. This explains why replacement of the ppGpp regulated gidA promoter The same applies to rifampicin that negatively affects the gidA promoter as well. In rifampicin treated cells, accumulation of DnaA gradually ceases due to the lack of transcription, whereas translation continues as long as intact dnaA mRNA is present. The translational inhibitor chloramphenicol, moreover, arrests DnaA accumulation immediately. Yet, we found rifampicin treatment to stop replication initiation faster than chloramphenicol. This is in line with previous findings that RNA synthesis contributes directly to an early stage of chromosome replication initiation (Lark, 1972;Messer, 1972;Zyskind et al., 1977). As chloramphenicol enables continued transcription, this supports that transcriptional activation of the origin allows for, at least a few, additional initiations during suboptimal, that is, DnaA limiting, conditions, and that cessation of de novo DnaA synthesis along with an inability to transcriptionally activate oriC, leads to a faster replication initiation stop as seen for rifampicin treated cells.
Previously, a direct interaction of DnaA with RNAP at oriC was reported to partially protect RNA polymerase against rifampicin at the gidA promoter (Flatten et al., 2009). The increased DnaA levels reported here could therefore allow for a degree of continued transcriptional activation mediated by gidA transcription in the presence of rifampicin, which in turn could contribute to continued replication initiation in the presence of rifampicin.
In summary, we propose that binding of rifampicin or ppGpp to RNAP leads to a stop/reduction in de novo DnaA protein synthesis. Although this results in a failure to replenish the DnaA ATP pool, which should be sufficient for an arrest in replication initiation, it does not ensure a fast turn-off. The efficacy of both rifampicin and ppGpp in turning off initiations is further enhanced by their negative effect on transcriptional activation of oriC. Therefore, a combination of several mechanisms including reduced de novo DnaA synthesis and lower negative supercoiling of oriC resulting from either reduced gidA (Kraemer et al., 2019) or gyrA/parC (Fernandez-Coll et al., 2020) transcription contributes to efficiently shutting down initiations during ppGpp accumulation or rifampicin treatment.

| Bacterial strains
Bacterial background strains are listed in Table 1.

| Construction of bacterial plasmids
All plasmids are listed in Table 2. Primers are listed in Table 3.
To construct a plasmid expressing DnaA in a rifampicin-and ppGpp-independent manner, the dnaA gene with ribosome binding site (RBS) was amplified as a 1,560 bp PCR fragment from MG1655 using primers DnaA_539_fw_XbaI and DnaA_539_bw_XhoI. This fragment was digested with XbaI and XhoI and ligated into the ex-
Incubation continued for a minimum of 4 hr at 37°C, unless otherwise specified. When replication initiation is inhibited, and ongoing rounds of replication allowed to finish, the number of fully replicated chromosomes per cell (i.e., genome equivalents) represents the number of origins per cell at the time of drug addition. In case of initiation synchrony, the integral number of chromosomes is two, four, or eight (i.e., 2 n ). In cells with asynchronous initiation additional peaks appear representing three, five, six, or seven cellular origins (i.e., ≠2 n ). Flow cytometry was performed as described previously  using an Apogee A10 Instrument (Apogee, Inc.). For each sample, 40,000-100,000 cells were analyzed.

| Quantification of cellular levels of (p)ppGpp
Cultures were grown in low phosphate (0.2 mM K 2 HPO 4 ) MOPS minimal medium supplemented with 0.2% glucose, 0.5% casamino acids and 10 μg/ml thiamine to an OD 450 of 0.150. Cultures were labeled with 32 P (final specific activity 100 μCi/ml; Hartmann Analytic GmbH, Braunschweig, Germany) for 2-3 generations of growth. ppGpp accumulation was induced by the addition of 1 mM IPTG. At the times indicated, samples of 50 μl were removed and mixed with 10 μl of icecold 2 M formic acid. Samples were kept at ice and stored at −20°C.
Prior to ppGpp detection, the frozen samples were centrifuged at 4°C for 45 min at 15,000g to pellet the cell debris, and 5 μl aliquots of the supernatants were added to polyethyleneimine (PEI) cellulose thin layer chromatography plates (Merck, Sigma, MO, US), and resolved with 1.5 M KH 2 PO 4 (pH 3.4) before being air dried and exposed by phosphorimaging (Amersham Typhoon phosphorimager, GE Healthcare, IL, US). Quantification of cellular (p)ppGpp levels was done using the ImageJ software (LOCI, University of Wisconsin, US), and the relative abundance of ppGpp and pppGpp was measured as a percentage value relative to the total amount (ppGpp+pppGp-p+GTP) as previously described (Mechold et al., 2013). as previously reported (Riber et al., 2006). All ori/ter ratios were normalized to the ori/ter ratio of a sample of wild-type cells treated with rifampicin and cephalexin where the ori/ter ratio is one.

ACK N OWLED G M ENTS
We thank Professor Judy Zyskind for kindly providing the pALS13 and pALS14 plasmids, and Professor Richard Gourse for providing the MG1655 RNAP (+1+2) and (−1−2) strains. Also, we would like to thank Dr. Anurag Kumar Sinha for his kind and excellent instructions with the ppGpp measurements. This work was supported by the Center for Bacterial Stress Response and Persistence (BASP) by a grant from the Danish National Research Foundation (DNRF120).

AUTH O R CO NTR I B UTI O N S
LR and ALO have both made major contributions to (a) the conception or design of the study, (b) the acquisition, analysis, or interpretation of the data; and (c) writing of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that supports the findings of this study are available from the corresponding authors upon a reasonable request.

R E FE R E N C E S
Atlung, T., Løbner-Olesen, A. and Hansen, F.G. (1987) Overproduction of DnaA protein stimulates initiation of chromosome and minichromosome replication in Escherichia coli. Molecular and General Genetics, 206, 51-59.