Escherichia coli DinB inhibits replication fork progression without significantly inducing the SOS response

The SOS response is readily triggered by replication fork stalling caused by DNA damage or a dysfunctional replicative apparatus in Escherichia coli cells. E. coli dinB encodes DinB DNA polymerase and its expression is upregulated during the SOS response. DinB catalyzes translesion DNA synthesis in place of a replicative DNA polymerase III that is stalled at a DNA lesion. We showed previously that DNA replication was suppressed without exogenous DNA damage in cells overproducing DinB. In this report, we confirm that this was due to a dosedependent inhibition of ongoing replication forks by DinB. Interestingly, the DinB-overproducing cells did not significantly induce the SOS response even though DNA replication was perturbed. RecA protein is activated by forming a nucleoprotein filament with single-stranded DNA, which leads to the onset of the SOS response. In the DinB-overproducing cells, RecA was not activated to induce the SOS response. However, the SOS response was observed after heat-inducible activation in strain recA441 (encoding a temperature-sensitive RecA) and after replication blockage in strain dnaE486 (encoding a temperature-sensitive catalytic subunit of the replicative DNA polymerase III) at a non-permissive temperature when DinB was overproduced in these cells. Furthermore, since catalytically inactive DinB could avoid the SOS response to a DinB-promoted fork block, it is unlikely that overproduced DinB takes control of primer extension and thus limits single-stranded DNA. These observations suggest that DinB possesses a feature that suppresses DNA replication but does not abolish the cell’s capacity to induce the SOS response. We conclude that DinB impedes replication fork progression in a way that does not activate RecA, in contrast to obstructive DNA lesions and dysfunctional replication machinery.


INTRODUCTION
DNA is damaged by various environmental factors such as endogenous reactive oxygen species and exogenous UV radiation.A replicative DNA polymerase thus often encounters various kinds of obstructive DNA lesions, at which progression of the replication fork is blocked.Escherichia coli cells induce the SOS response to cope with replication fork inhibition that is caused not only by obstructive DNA damage but also by dysfunctional replication machinery in temperature-sensitive dna mutants at high temperatures (Schuster et al., 1973;Friedberg et al., 2006).While there are few regions of single-stranded DNA (ssDNA) in undamaged cells, ssDNA regions accumulate at stalled replication forks (Higuchi et al., 2003;McInerney and O'Donnell, 2007).In the current model of the SOS response, generation of ssDNA regions is the SOS-inducing signal, since it turns on the RecA activator of the SOS response (Sassanfar and Roberts, 1990).RecA binds to ssDNA to form a nucleoprotein filament that activates the co-protease activity of RecA, leading to autocleavage of the transcriptional repressor LexA (Lusetti and Cox, 2002;Butala et al., 2009).This cleavage results in the derepression of more than 40 SOS genes to promote the cell's survival under replication stress (Courcelle et al., 2001).The SOS genes function in damage-tolerance pathways including cell division block, DNA damage repair, replication restart, recombination and translesion synthesis (TLS) (Goodman, 2002;Friedberg et al., 2006;Yang and Woodgate, 2007;Bichara et al., 2011).
The replicative polymerase of E. coli, DNA polymerase (Pol) III, catalyzes DNA chain elongation with high processivity and high fidelity (McHenry, 2003;Maki, 2004;O'Donnell, 2006).When Pol III stalls at a DNA lesion, SOS induction causes upregulation of the genes encoding three DNA polymerases that are specialized for TLS: Pol II, DinB (Pol IV) and Pol V (Courcelle et al., 2001).These TLS polymerases copy the lesion sites in place of the stalled Pol III, and therefore underpin replication fork movement over the lesions (Goodman, 2002;Nohmi, 2006;Fuchs and Fujii, 2007); DinB advances stalled replication forks over chemically induced N 2 -deoxyguanine adducts (Napolitano et al., 2000;Shen et al., 2002;Jarosz et al., 2006).Together with Pol V, DinB belongs to the Y-family DNA polymerases, whose members are characterized by low-processivity DNA synthesis, absence of a proofreading exonuclease activity, and spacious active sites that can accommodate damaged DNA (Ohmori et al., 2001;Jarosz et al., 2007).The Y-family DNA polymerases are present in all the major phylogenetic groups, with DinB homologues such as archaean Dpo IV and eukaryotic Pol κ being the most widely conserved (Gerlach et al., 1999;Ogi et al., 1999;Boudsocq et al., 2001).
The number of DinB molecules per E. coli cell under non-SOS conditions is about 250 (Kim et al., 2001), far more than the 2-8 replication forks in an exponentially growing cell (Skarstad et al., 1986).In SOS-induced cells, enhanced expression of the dinB gene causes a 10fold increase in the number of DinB molecules (Kim et al., 2001).Beyond its established TLS function at replication forks, these large amounts of DinB have led to proposals for additional biological roles that include error-free processing of endogenously formed lesions such as alkylation (Bjedov et al., 2007) and glycation (Yuan et al., 2008), adaptive mutagenesis (McKenzie et al., 2001;Slechta et al., 2003;Tompkins et al., 2003), error-prone double-strand break repair (Ponder et al., 2005) and longterm survival during stationary phase (Yeiser et al., 2002).Recently, based on the suppression of replication fork progression by ectopic dinB overexpression, it has also been proposed that DinB plays a role as a molecular brake in eukaryotic checkpoint-like regulation of replication fork progression in E. coli (Uchida et al., 2008;Indiani et al., 2009;Langston et al., 2009).This fork-brake activity would provide extra time for a DNA lesion to be repaired before the fork encounters another lesion ahead.Biochemical analysis revealed that DinB inhibits a moving Pol III that is rapidly polymerizing nucleotides on the template, and takes over DNA synthesis from Pol III in a dose-dependent manner in vitro (Furukohri et al., 2008;Indiani et al., 2009;Wagner et al., 2009).These findings strongly suggest that upregulation of dinB impedes the progression of replication forks by inhibiting Pol III.Furthermore, amounts of DinB corresponding to those in SOS-induced cells can generate a slow-moving replication machinery with the replicative DnaB helicase in vitro (Indiani et al., 2009).
We have previously reported that replication fork progression was impeded when DinB was rapidly overproduced, upon arabinose addition, from the dinB gene under the control of the P BAD promoter (P BAD -dinB) on a multi-copy plasmid (Uchida et al., 2008).On the other hand, in E. coli cells that accumulate DinB more moderately from a single P BAD -dinB on the chromosome, DNA synthesis slowed gradually (Uchida et al., 2008).In the present report, we found that this slow DNA synthesis eventually stopped after more than one generation time.This slow-stop phenotype during DNA synthesis was further investigated, since it is often caused by inhibition of the initiation of DNA replication (Sevastopoulos et al., 1977;Kornberg and Baker, 1992).Microarray analysis revealed that the phenotype in dinB-overexpressing cells was accounted for by elongation at a slow rate but not by additional interruption during the initiation steps of DNA replication, indicating that DinB antagonizes solely replication fork progression in a dose-dependent manner.To gain a better insight into how DinB inhibits DNA replication forks, we also examined the SOS response, which is ordinarily induced by replication fork block.Remarkably, the dinB-overexpressing cells elicited an extremely weak SOS response.The insignificant SOS response was not attributable either to blockage of RecA activation at the onset of SOS induction or to limitation of ssDNA due to low-processivity DNA synthesis by DinB at inhibited replication forks.These results suggest that the fork-inhibition characteristic of DinB operates through a mechanism distinct from that operating in response to UV irradiation and Pol III dysfunction, both of which readily induce SOS.As we have observed in this study, inhibition of replication fork progression may not inevitably cause a strong induction of the SOS response in E. coli.

MATERIALS AND METHODS
Media and chemicals LB medium was prepared as described (Sambrook and Russel, 2001).E salt is 1 x Vogel and Bonner's synthetic E medium (Vogel and Bonner, 1956), and ECA growth medium is E salt with supplements as described (Horiuchi et al., 1978).[2-14 C]-thymine (52 mCi mmol -1 ) was purchased from American Radiolabeled Chemicals, USA.Rabbit anti-LexA and mouse anti-RopD antibodies were purchased from Bio Academia, Japan and NeoClone Biotechnology International, USA, respectively.Rabbit anti-DinB antibody was obtained from Dr. Takehiko Nohmi (National Institute of Health Sciences, Japan).
Bacterial strains and plasmids All E. coli K12 strains and plasmids used are listed in supplementary Tables S1 and S2, respectively.Replacement of the chromosomal dinB gene by the kanamycin resistance gene (kan) was performed by P1(vir)-mediated transduction (Miller, 1972) with the Keio collection JW0221 as donor (Datsenko and Wanner, 2000;Baba et al., 2006).The kan gene flanked by flippase recognition target (FRT) was then eliminated using the flippase helper plasmid pCP20 (Datsenko and Wanner, 2000).The temperaturesensitive dnaE486 mutation of TS1502 (Shibata et al., 2005) was cotransduced with the tetracycline resistance gene of Tn10 into recipient strains at 30°C.Temperaturesensitive colonies were selected among tetracyclineresistant transductants, yielding MK7146 from MK7136 and MK7149 from MK7140.Likewise, the temperaturesensitive recA441 of RM112 (Maul and Sutton, 2005) was transduced with Tn10 into the recipient MK7140, and MK7176 was obtained.
Cell growth Overnight cultures of cells were prepared in LB medium, and unless otherwise noted the cells were grown to log phase in ECA medium as described previously (Uchida et al., 2008).Measurement of colonyforming ability (CFU ml -1 ) and labeling experiments with [ 14 C]-thymine were carried out as described previously (Uchida et al., 2008).Cells carrying either pSK1002 or pBAD-HisA derivatives (Supplementary Table S2) were grown in medium containing 50 μg ml -1 ampicillin.In addition, 0.2% glucose was routinely added to the medium to repress expression from the P BAD promoter.Exponentially growing cells were diluted to an OD 600 of 0.1 and treated appropriately for each experiment (designated as time zero).L(+)-arabinose was added to a final concentration of 0.2% (w/v) to induce gene expression from the P BAD promoter.Cell cultures were irradiated using a UV Stratalinker 1800 (Stratagene, USA).For temperaturesensitive strains, the cells were grown exponentially at 25 and 30°C for recA441 and dnaE486, respectively, and diluted with pre-heated medium (at 55°C) to raise the temperature of the culture to 42°C.
After appropriate treatments, aliquots of the cells were removed at various time points and immediately chilled on ice.To measure the colony-forming ability of MK7010 and MK6956, each cell suspension was diluted appropriately and spread on LB agar plates.When arabinose had been added to cell cultures, the cells were pelleted at 8,000 x g for 5 min at 4°C and suspended in the same volume of E salt to remove the arabinose.For MK6956 cells, the LB plates contained 0.2% (w/v) glucose to suppress P BAD -dinB expression.The plates were incubated for 14-18 h at 37°C and the colonies formed were scored to determine CFU ml -1 .Microarray analysis Four independent cultures of MK6956 cells were grown in ECA medium at 37°C.Ara-binose (final concentration, 0.2%) was added to each culture to induce P BAD -dinB, and the cultures ware incubated for 2 h at 37°C.After sodium azide was added to a final concentration of 0.2% (w/v), the cultures were pooled and chilled on ice.Three independent cultures of the cells were also grown to stationary phase in the absence of arabinose and pooled.Cells were collected by centrifugation at 8,000 x g for 15 min at 4°C.Genomic DNA from each pooled preparation was purified and hybridized separately to a GeneChip E. coli Antisense Genome Array (Affymetrix, USA) as described previously (Uchida et al., 2008).Probe hybridization signal values of P<0.01 (the threshold of statistical significance of each detection) were divided by those of a reference stationary-phase culture.Each resultant ratio was normalized further by dividing it by the average gene dosage of 128 contiguous ORFs (covering about 76 kb, with uxaB as locus 1 and tus as locus 128) that flank the TerB and TerC sites.
Measurement of the SOS response Induction of the SOS response in cells carrying a lacZ reporter plasmid, pSK1002 (Oda et al., 1985), was measured by determining the specific activity of β-galactosidase (Miller units) of the UmuD-LacZ fusion protein as described (Miller, 1972).The SOS response was also monitored by autocleavage of LexA, which can be examined by detecting LexA repressor in the absence of protein synthesis.To prevent new LexA protein synthesis, the cell samples were further incubated with 50 μg ml -1 of chloramphenicol (Cm) for the times indicated in the figure legends and were subjected to Western blot analysis with anti-LexA antibodies as described below.

Western blot analysis
The cells were collected by centrifugation at 4°C, suspended in 1x SDS-polyacrylamide sample buffer (Sambrook and Russel, 2001) and heated at 99°C for 3 min.Cells equivalent to 1 ml of a suspension at OD 600 = 0.05 were harvested by centrifugation, and total cellular proteins from the cells were loaded in each lane and separated by SDS-PAGE (Sambrook and Russel, 2001).The resolved proteins were transferred to nitrocellulose membranes (Schleicher & Schuell, Germany) and probed with appropriate antibodies as described (Harlow and Lane, 1988).Anti-DinB rabbit antiserum was mixed with cell extracts of a ΔdinB strain to remove non-specific antibodies (Kim et al., 2001).Immunoblots were developed with ECL reagents (GE Healthcare, USA), and visualized by LAS-4000 Mini luminescence image analyzer (Fuji Film, Japan) or by exposure to X-ray film; the film was then scanned with a CCD scanner (Epson, Japan).The σ 70 subunit (RopD) of RNA polymerase served as a loading control for the extracts in the Western blot analysis.In quantitative Western blotting, the linear range for the DinB signals from SOS-constitutive SMR7623 (lexA51(Def)) cells was established by serial dilution (Pennington and Rosenberg, 2007), and relative amounts of cellular DinB were determined by comparison with the SOS-induced level.

Moderate overproduction of DinB slows down and then stops progression of DNA replication forks
Overproduction of DinB in E. coli cells carrying the multi-copy P BAD -dinB plasmid pDB10 completely inhibited cellular DNA synthesis immediately after arabinose was added to the medium, at which time the amount of DinB was 15-fold higher than that in cells undergoing the SOS response (Uchida et al., 2008).This response resembles the "quick-stop" phenotype observed in E. coli expressing conditionally lethal dna genes that render the cells defective in DNA chain elongation (Sevastopoulos et al., 1977;Kornberg and Baker, 1992), and thus indicates that replication fork progression is rapidly inhibited by overproduction of DinB.To reduce the rate of production of DinB after arabinose addition, we constructed E. coli strains carrying a single copy of the P BAD -dinB gene inserted ectopically into the chromosome.The DNA synthesis rate in one such strain, MK6980, was reduced to about half of that in the wild type strain when the intracellular DinB level was 2-to 8-fold higher than that in the SOS-induced cells (Uchida et al., 2008).We repeated this experiment with an extended time course.Whereas protein synthesis was not affected (data not shown), DNA synthesis slowed down and had terminated almost completely within 60 min after arabinose addition (Fig. 1A), which is beyond the doubling time of DNA in MK6980 in the absence of arabinose (about 30 min; data not shown).This pattern of more gradual inhibition of DNA replication resembles the "slow-stop" phenotype, which is generally caused by dysfunction during the initiation step of DNA replication (Sevastopoulos et al., 1977;Kornberg and Baker, 1992).Thus, a DinB level lower than that seen in the plasmid-based overexpression experiment appeared to be interfering with DNA replication initiation.
To examine this possibility, we measured the relative gene dosage of almost every open reading frame (ORF) in MK6956 (chromosomal P BAD -dinB) cells collected 120 min after dinB induction, using genomic DNA extracted from stationary-phase cells as a reference.The resulting profile is displayed in Fig. 1B as smoothed moving median values for the relative abundance of each ORF (Simmons et al., 2004;Camara et al., 2005).The profile shows a maximum at oriC and a minimum at the terminus, exactly the same as the genomic profiles both of normally growing exponential-phase cells and of "quick-stop" DinBoverproducing cells carrying the pDB10 plasmid (Uchida et al., 2008).The gene dosage around oriC (about 0.7) was similar to that in log-phase cells (Uchida et al., 2008).Together with the result shown in Fig. 1A, the slow-stop nature of DinB-overproducing MK6980 cells is probably attributable to a decelerated replication fork movement at random genomic positions, the so-called "slow-rate" phenotype (Sevastopoulos et al., 1977), until 60 min after arabinose addition.Therefore, it seems likely that DinB slows down replication fork progression in a dose-dependent manner, comparable to its negative effects on Pol III observed in vitro (Furukohri et al., 2008).

Inhibition of replication fork progression by DinB does not significantly induce the SOS response
When DNA replication is inhibited by either DNA lesions or dysfunction of the replicative apparatus, ssDNA accumulates at the blocked replication fork (Higuchi et al., 2003;McInerney and O'Donnell, 2007), leading to the induction of expression of various SOS genes in E. coli (Courcelle et al., 2001).To determine whether SOS induction is provoked by the inhibition of replication fork progression by DinB protein, we measured the specific activity of β-galactosidase in cells carrying plasmid pSK1002, which contains the umuD::lacZ fusion gene and serves as a reporter for SOS induction (Oda et al., 1985).When MK7010 (dinB + ) wild type cells carrying pSK1002 were irradiated with UV, which impedes DNA replication, umuD::lacZ expression reached a highly induced level, comparable to that observed with the lexA51 (Def) strain, in which the SOS response is constitutively induced (Fig. 2).This result confirmed that SOS induction could be detected in our strain with this experimental system.Interestingly, the SOS response was not detected in MK6956 cells carrying pSK1002 until at least 30 min after replication was inhibited by arabinose addition.Even when the inhibition of replication fork progression by DinB was prolonged to 120 min, the  level of β-galactosidase activity increased only slightly, to about one tenth of the maximum level observed in cells having a fully induced SOS response.As shown in Fig. 1A, overexpression of the chromosomal P BAD -dinB blocked DNA synthesis almost completely 60 min after arabinose addition.In the absence of arabinose, strain MK6956 carrying pSK1002 was competent for the UV-induced SOS response, comparable to that observed in wild type MK7010 carrying pSK1002 (supplementary Fig. S1A).Therefore, while the overexpression of DinB protein readily inhibits replication fork progression, it induces the SOS response only very weakly.
Any enzymatic reporter assay is sensitive to perturbations of transcription and protein synthesis.Although the overproduction of DinB does not significantly inhibit overall protein synthesis (Uchida et al., 2008), DinB directly interacts with NusA, a modulator of RNA polymerase (Cohen et al., 2009), and may block the transcription of umuD through NusA-dependent transcriptional modulation.To rule out the possibility that SOS was induced but was not detected with the reporter gene, we examined one of the initial steps in the process of SOS induction, namely the autocleavage of LexA protein facilitated by activated RecA.By Western blotting with anti-LexA antibodies, we analyzed how much intact LexA repressor remained in cells (Fig. 3).Cell samples collected at each time point except time zero were incubated at 37°C for 5 min in the presence of chloramphenicol (Cm), which blocks protein synthesis.Under this condition, a constant amount of LexA was observed in exponentially growing MK7010 (dinB + ) cells (data not shown).When MK7010 cells were subjected to UV irradiation, they immediately stopped growing, as judged by measur-ing colony-forming units (CFU) (Fig. 3A).In the UVirradiated cells, LexA disappeared by autocleavage shortly after irradiation and reappeared at later time points (Fig. 3B).LexA was resynthesized when the cells had recovered from DNA damage by repair (data not shown).As with UV irradiation of MK7010, growth of MK6956 (chromosomal P BAD -dinB) cells stopped immediately after arabinose addition.In the DinB-overproducing cells, the amount of LexA was largely unchanged until 30 min after arabinose addition and slightly reduced at 60 min (Fig. 3C).From these data, we concluded that LexA autocleavage does not occur in DinB-overexpressing cells, which rules out the possibility that the negligible SOS induction was due to inhibition of reporter gene expression by excess DinB.
Dysfunction of the replicative apparatus strongly induces the SOS response in E. coli (Schuster et al., 1973).The α subunit of Pol III contains the DNA polymerase activity and is encoded by dnaE (Maki et al., 1985).We examined LexA autocleavage in temperaturesensitive dnaE486 mutant cells, in which DNA synthesis is blocked immediately after the temperature shift from 30°C to 42.5°C (Wechsler and Gross, 1971).Cm was added to the cells at the same time as the temperature shift, and the amount of LexA repressor was determined in cells collected every 3 min (Fig. 4).In Fig. 4, all strains lack the endogenous dinB gene so that DinB is produced only from P BAD -dinB.Consistent with the immediate stall of DNA replication, most of the LexA protein was degraded within 3 min in the dnaE486 cells (Fig. 4, lanes 5 to 8).In contrast, LexA level in dnaE + wild type cells remained unchanged at 3 min after the temperature shift, and then decreased slightly until 9 min (Fig. After arabinose (final 0.2%) was added to exponentially growing cells in ECA medium at 30°C, incubation was continued at 30°C for 180 min except for MK7149 (dnaE486, P BAD -dinB) (150 min).At time zero, the OD 600 was adjusted to 0.1, the temperature was shifted to 42°C and Cm was added (final 50 μg ml -1 ).Aliquots of the cultures were taken at the times indicated above each image and chilled on ice.The cell samples were analyzed by Western blotting with anti-RpoD (top panels), anti-DinB (middle panels) and anti-LexA (bottom panels) antibodies.Samples are MK7147 (dnaE + ) in lanes 1 to 4, MK7146 (dnaE486) in lanes 5 to 8, MK7150 (dnaE + , P BAD -dinB) in lanes 9 to 12 and MK7149 (dnaE486, P BAD -dinB) in lanes 13 to 16. Negligible SOS response to fork arrest by DinB 4, lanes 1 to 4).Strain MK7150 (dnaE + , P BAD -dinB) was cultured at 30°C in the presence of arabinose until the amount of DinB reached approximately 5-fold that in lexA51 (Def) cells (Supplementary Fig. S2B), and the temperature was then shifted to 42°C with concurrent addition of Cm.This amount of DinB slows down DNA synthesis in vivo (Uchida et al., 2008).Under these conditions, the fate of LexA in MK7150 cells showed a pattern similar to that observed in the dnaE + wild type cells, although the amount of LexA at the 0-min time point was slightly lower in MK7150 cells (Fig. 4, lanes 9 to 12).All the data described above suggest that the mechanism of replication fork inhibition caused by excess DinB is distinct from that caused by DNA damage and dysfunction of the replicative apparatus, in which RecA is activated and LexA is autocleaved at the onset of the normal SOS response.
RecA is proficient for activation in the presence of excess DinB DinB physically interacts with RecA and inhibits autocleavage of UmuD 2 , another cellular substrate for the co-protease of RecA (Godoy et al., 2007).Therefore, it seemed possible that overproduced DinB might inhibit the activation of RecA by interacting directly with RecA.To assess this possibility, we examined whether DinB-overproducing cells remain able to induce SOS.However, it was inappropriate to do so using well-known SOS inducers that block replication, such as UV irradiation and DNA-damaging chemicals.This is because the induction of SOS by DNA damage requires blockage of active DNA replication in recA + cells (Sassanfar and Roberts, 1990;Friedberg et al., 2006), but replication fork progression is already suppressed in cells overproducing DinB protein.To overcome this problem, we took advantage of the recA441 mutant strain, which shows constitutive expression of the RecA-mediated SOS response at 42°C but not at 30°C (Kirby et al., 1967).ssDNA is normally formed as an intermediate in replication, recombination and repair, and is immediately coated by ssDNA binding protein (SSB) with high affinity in E. coli cells (Shereda et al., 2008).It has been proposed that the RecA441 protein by itself displaces SSB from the limited ssDNA normally present in undamaged cells to form nucleoprotein filaments, and thereby becomes activated for LexA cleavage independently of any DNA damage when cells are grown at 42°C (Lavery and Kowalczykowski, 1988).We introduced the recA441 mutation into MK6956 (P BAD -dinB) and examined the SOS response at 42°C with DinB overproduction.It should be noted that the endogenous dinB was deleted to avoid additional DinB overproduction caused by thermal induction of LexA autocleavage in the recA441 cells.
The resulting strain MK7176 (recA441, P BAD -dinB) and MK7175 (recA + , P BAD -dinB) were pre-cultured in LB medium at 25°C to completely suppress RecA activation, and the growth temperature was then raised to 42°C (time zero in Fig. 5).MK7176 cells were also incubated at 25°C for 90 min in the presence of arabinose prior to the temperature shift.After arabinose addition, DinB accumulated in the MK7176 cells to about 19-fold more Fig. 5. Thermal induction of SOS by the recA441 mutant in the presence of excess DinB.MK7175 (recA + , P BAD -dinB) and MK7176 (recA441, P BAD -dinB) cells were grown exponentially in LB medium in the absence of arabinose at 25°C.The temperature was changed to 42°C at time zero, and incubation was continued; MK7175 (lanes 1-4) and MK7176 (lanes 5-8).In MK7176 (lanes 9-12), after dinB overexpression had been induced by arabinose (final 0.2%) at 25°C for 90 min, the OD 600 was adjusted to 0.1 and the incubation temperature of the culture was simultaneously shifted to 42°C (designated time zero).Aliquots of the cultures were taken at the times indicated, and the cells at each time point except for time zero were incubated at 42°C with Cm (50 μg ml -1 ) for 5 min.Total cell proteins were analyzed by Western blotting as in Fig. 4. than the level found in SOS-induced cells, 2500 molecules/cell (Supplementary Fig. S2C) (Kim et al., 2001).Thus, DinB was in about 7-fold molar excess over RecA (7200 molecules/cell; Sassanfar and Roberts, 1990) at time zero (lane 9 of Fig. 5), an amount sufficient to immediately abolish replication fork progression (Uchida et al., 2008).Cells were taken at the indicated times and incubated for 5 min in the presence of Cm.Unlike Fig. 4, the amount of DinB was not reduced during this 5-min incubation of the 15-min sample because of higher overproduction (Supplementary Fig. S2).In the 15-min sample, LexA disappeared after the temperature shift, irrespective of DinB overproduction (Fig. 5, lanes 6 and 10).In contrast, LexA remained intact after the temperature shift for recA + cells (Fig. 5, lanes 1 to 4), indicating that its disappearance in MK7176 at 42°C (Fig. 5, lanes 5 to 12) was not caused by the heat shock response.Consequently, we concluded that DinB overproduction has effect on the activation of RecA441.
Even when ssDNA is accumulated, the formation of RecA nucleoprotein filaments is inhibited by SSB tracts that cluster rapidly on ssDNA (Shereda et al., 2008).This inhibition is bypassed in RecA441 at the nonpermissive temperature, as described above.Unlike RecA441, wild type RecA requires RecF, RecO and RecR proteins that relieve the SSB barrier and facilitate RecA filament formation (Shereda et al., 2008).To examine if DinB overproduction negatively affects this RecFOR reaction, strain MK7149 (dnaE486, P BAD -dinB, recA + ) was constructed.The recA-proficient MK7149 requires RecFOR to induce the SOS response by replication block at the nonpermissive temperature.When MK7149 cells were grown in the presence of arabinose at 30°C and accumulated DinB to a level that slows replication, about 5-fold higher than that in SOS-induced cells (Supplementary Fig. S2A) (Uchida et al., 2008), the growth temperature was shifted to 42°C and Cm was added.Different incubation times at 30°C were used for MK7150 (Fig. 4, lanes 9-12) and MK7149 (Fig. 4, lanes 13-16) to obtain the same level of DinB overproduction.The amount of LexA in MK7149 (Fig. 4, lanes 13 to 16) decreased 3 min after the temperature shift and showed a similar pattern of change to that in dnaE486 (Fig. 4, lanes 5 to 8).Thus, thermal inactivation of Pol III at 42°C was capable of inducing the SOS response in the presence of overproduced DinB.This implies that the excess amount of DinB neither interferes with the SOS response caused by malfunction of the replicative apparatus nor inhibits RecFOR-dependent RecA-filament assembly on SSBcoated ssDNA.From the above findings, it is unlikely that excess DinB suppresses SOS induction by inhibiting RecA activation.Rather, a replication fork inhibited by excess DinB does not generate a strong SOS-inducing signal, perhaps because the replication fork remains intact.

DNA polymerase activity of DinB is dispensable for avoidance of the SOS response in dinB-overexpressing cells
The excess amount of DinB at a physiological level directly inhibits DNA synthesis by Pol III holoenzyme in vitro (Furukohri et al., 2008;Indiani et al., 2009).We previously showed that DinB dislocates Pol III from the primer terminus, where Pol III in association with the sliding β clamp is elongating the DNA chain, and then takes over the primer terminus and β clamp to continue DNA chain elongation (Furukohri et al., 2008).When this polymerase switching occurs, the speed of chain elongation drops from 1000 to 3-5 nucleotides per second (Wagner et al., 2000;Furukohri et al., 2008).Considering these in vitro findings, we reasoned that the replication fork might continue to proceed very slowly in cells overproducing DinB protein, despite appearing to stop in [ 14 C]-thymine incorporation experiments (Fig. 1) (Uchida et al., 2008), and that DinB overproduction would not constitute an SOS-inducing signal if the DinB catalyzed a slow movement of the replication fork.To determine whether the very slowly moving replication fork catalyzed by DinB suppresses the SOS response, a catalytically defective mutant DinB (DinB-D8A) was overproduced in MK7136 (ΔdinB) cells carrying pDB11 plasmid (MK7136/pDB11) (Uchida et al., 2008), and the level of LexA in the cells was analyzed (Fig. 6).DinB and DinB-D8A were synthesized equally in MK7136/pDB10 and MK7136/pDB11, respectively, in the presence of arabinose (Fig. 6) (Uchida et al., 2008).[ 14 C]-thymine incorporation into both types of cells stopped 15 min after arabinose addition (Uchida et al., 2008).Under these Fig. 6.The SOS response in cells overexpressing catalytically defective DinB-D8A.MK7136 (ΔdinB) cells carrying pBAD-HisA vector, pDB10 (P BAD -dinB + ) or pDB11 (P BAD -dinB-D8A) were grown exponentially at 37°C.After arabinose was added to 0.2% at time zero (OD 600 = 0.1), aliquots of the cultures were taken at the times indicated.The cell samples (except for lanes 1, 4 and 7) were incubated with Cm (50 μg ml -1 ) at 37°C for 5 min.Total cell proteins were probed by Western blotting with anti-RpoD, anti-DinB and anti-LexA antibodies.Samples are MK7136/pBAD-HisA (lanes 1 to 3), MK7136/pDB10 (lanes 4 to 6) and MK7136/pDB11 (lanes 7 to 9).Negligible SOS response to fork arrest by DinB circumstances, the LexA levels were maintained, with only a transient slight decrease to comparable levels in both MK7136/pDB10 and MK7136/pDB11 around 30 min after arabinose addition (Fig. 6, lanes 4 to 9).On the other hand, a strong SOS response was detected when both types of cells were irradiated with UV in the absence of arabinose (Supplementary Fig. S1B), confirming their competency for SOS induction.These data indicate that DNA polymerase activity does not contribute to the failure of DinB to induce the normal SOS response when it is overproduced.Therefore, it is unlikely that the very slow movement the replication fork under excess DinB keeps the DNA in a double-stranded configuration and prohibits ssDNA from forming around the inhibited replication fork.DinB may inhibit replication fork movement in a novel way so as to prevent the accumulation of ssDNA.

DISCUSSION
The level of DinB protein in E. coli cells rises when the SOS response is induced by replication inhibition such as the blocking of replication forks by DNA damage (Kim et al., 2001;Friedberg et al., 2006).Like other proteins induced by the SOS response, DinB functions to cope with blocked replication.As a TLS DNA polymerase, DinB advances the stalled replication fork over lesions in place of the replicative Pol III (Napolitano et al., 2000;Shen et al., 2002;Jarosz et al., 2006).In contrast to this role in fork rescue, levels of DinB close to those in the SOSinduced cell antagonize the moving replication fork (Uchida et al., 2008;Indiani et al., 2009;Wagner et al., 2009).In this report, we show that DinB can slowly shut down replication fork progression without significant induction of the SOS response, in contrast to DNA damage and dysfunction of the replication apparatus, both of which quickly abolish replication fork progression and lead to strong induction of the SOS.

DinB decelerates replication fork progression in a
dose-dependent manner DNA replication fork progression is inhibited quickly and completely when DinB protein is induced rapidly and to a level 15-fold higher than that in SOS-induced cells (Uchida et al., 2008).When DinB was induced more slowly and to the lower level of 2-to 8-fold higher than that in the SOS response, replication fork progression slowed down gradually (Fig. 1) (Uchida et al., 2008).We therefore concluded that DinB modulates replication fork speed in a dose-dependent manner in E. coli cells.How might this occur?One possible explanation for the dose-dependent brake on the replication fork in vivo is a polymerase switch between DinB and the replicative Pol III.A similar dose-dependent effect of DinB was observed in experiments in which DinB replaced Pol III at the primer terminus on a single-stranded template DNA in vitro (Furukohri et al., 2008).As the level of DinB increases during overproduction, polymerase switching may also occur more frequently in vivo, through both reduced primer occupancy of highly processive Pol III and low processive DNA replication by DinB in place of Pol III, resulting in a perceptible slowdown of the replication fork.This possibility is supported by results showing that the inhibitory effects of various DinB mutants on the replication fork in vivo were almost identical with those on Pol III in vitro (Furukohri et al., 2008;Uchida et al., 2008).The catalytically dead DinB may inhibit the replication fork only by reducing the likelihood of Pol III acting there (Fig. 6).The other possibility is that a slowdown in the unwinding rate of the replicative DnaB DNA helicase might lead to a reduced rate of replication fork progression (Langston et al., 2009).DinB can form the slow replisome with DnaB helicase, and the speed of this alternative replisome depends on the concentration of DinB in vitro (Indiani et al., 2009).In the slow replisome, DnaB helicase may lose the polymerase-helicase coupling, mediated by the τ subunit of Pol III, that is necessary for rapid helicase translocation (Kim et al., 1996;Yuzhakov et al., 1996;Dallmann et al., 2000).The τ subunit may fail to establish fully active DnaB due either to the release of Pol III from replication forks by DinB (Furukohri et al., 2008;Indiani et al., 2009) or to an inactive conformation of Pol III remaining tethered to the β clamp, in accordance with the tool-belt model of TLS (Bunting et al., 2003;Pagès and Fuchs, 2002;Indiani et al., 2005).Recently, it has been reported that an accessory DNA helicase, Rep, interacts with DinB (Sladewski et al., 2011).Rep helicase functions in DNA replication (Atkinson et al., 2011) and may play a role to coordinate DNA unwinding and polymerization for the slow DinB replisome.DnaB may also interact with DinB in a similar manner to Rep, resulting in inhibition of the DNA unwinding activity of DnaB in the slow DinB replisome.

DinB controls replication fork movement without fork breakdown
Interference with replication fork progression by DNA damage and temperature-sensitive dna mutants at nonpermissive temperatures can lead to various structures of collapsed replication forks, including a broken DNA end.In the current model for the SOS response in E. coli, an intracellular signal for SOS induction is generated from segments of ssDNA in the aberrant DNA structures (Sassanfar and Roberts, 1990;Lusetti and Cox, 2002;Friedberg et al., 2006).RecA proteins assemble on ssDNA and form nucleoprotein filaments that stimulate autocleavage of LexA and subsequent induction of the SOS genes.We showed that dinB overexpression induced only slightly β-galactosidase expressed from a LexA-regulated promoter (Fig. 2).In addition, by measuring the in vivo level of LexA repressor using anti-LexA antibodies, we showed that dinB overexpression did not lead to rapid LexA autocleavage (Figs. 3  and 4).Furthermore, the increased level of DinB did not inhibit the activation of RecA (Fig. 5).Therefore, even though replication fork progression was blocked, the SOS response was neither induced nor inhibited in cells overproducing DinB.
From these results, we conclude that ssDNA segments, which would activate RecA protein, are probably not present around the blocked replication forks in DinBoverproducing cells.This model implies that DinB is capable of halting replication fork movement without causing a breakdown of the replication fork that would lead to ssDNA accumulation.The alternative explanation of the results is that ssDNA may be present but unavailable for RecA activation.DinB may inhibit RecA activation on SSB-coated ssDNA because it can bind ssDNA (Grúz et al., 2001;Furukohri et al., 2008).However, the affinity of DinB for ssDNA is approximately 2400-fold lower than that of SSB (Grúz et al., 2001).Even the highest amount of overproduced DinB in this report (about 80 μM; Fig. 5) would thus not effectively compete with SSB (1.7-3.4 μM; Bobst et al., 1985) in terms of ssDNA binding.Moreover, although DinB also interacts with SSB (Furukohri et al., 2012), we have not been able to detect replacement of SSB on SSB-coated ssDNA by DinB (AF, MTA and HM, unpublished results).Thus, it is highly unlikely that DinB makes ssDNA inaccessible to RecA by masking it to suppress SOS induction.
In Bacillus subtilis, the highly conserved RecA and LexA proteins control the induction of the protective SOS response following replication stress.RecA is activated and promotes autocleavage of DinR, the B. subtilis LexA protein, in the presence of ssDNA and a nucleoside triphosphate, both of which are required for nucleoprotein filament formation in vitro (Miller et al., 1996).Replication blockage by UV damage generates activated RecA in B. subtilis as it does in E. coli (Lovett et al., 1994).However, when an array of repressor proteins bound to DNA blocks DNA replication in B. subtilis, the SOS response is not significantly induced despite a robust generation of RecA filament during the replication block (Bernard et al., 2010).Cellular responses to replication fork inhibition are complex and may not invariably result in a strong induction of the SOS response, as we observed here in dinB-overexpressing cells.Interestingly, the human ortholog of DinB, Pol κ, is induced by polycyclic aromatic hydrocarbons that generate N 2 -dG adducts (Ogi et al., 2001), and a moderate ectopic overproduction of the enzyme slows down replication fork progression without activating the ATR-dependent checkpoint, the response that protects eukaryotic cells from replication stress (Pillaire et al., 2007).This ability of bacterial DinB to retard fork progression without causing extensive replica-tion stress thus seems to be highly conserved.Since an elevated level of Pol κ induces genomic instability in hamster cells (Bavoux et al., 2005), it will be of interest to investigate if overproduction of E. coli DinB also results in genomic instability.Such genomic instability might account for the slight degradation of LexA that we observed upon prolonged replication inhibition by DinB (the 60-min time point in Fig. 3).
What is the physiological relevance of DinB's capacity to inhibit the replication fork without significantly inducing the SOS response?To maintain the integrity of the genome, eukaryotes have evolved checkpoint mechanisms which halt cell cycle progression under cellular stress.Among these, the replication checkpoint functions in Sphase and provokes cellular reactions to control overall DNA replication when the replication fork(s) of one or more replicons are impeded by DNA lesions (Ben-Yehoyada et al., 2007;Branzei and Foiani, 2010).This DNA damage response includes a reduction of bulk replication on a damaged template by slowdown of unperturbed replication fork progression by unknown mechanisms (Willis and Rhind, 2009), thereby providing time for the cell to repair the damaged template before unperturbed replication forks encounter the lesions.Based on the low velocity of DNA replication caused by upregulation of E. coli dinB, we and others have proposed DinB as a candidate factor to slow down replication fork progression when cells activate the bacterial damage response (Uchida et al., 2008;Indiani et al., 2009).If further SOS induction signals were generated by the action of DinB, a catastrophic replication inhibition would occur as a result of sustained activation of dinB expression during chronic SOS induction.The negligible SOS induction we now observe in cells overproducing DinB would provide a chance for the cells to restore normal DNA replication after DinB acts at a stalled replication fork for TLS, and at other ongoing replication forks for slowing progression downstream of the initial SOS-inducing signal.In addition, degradation of DinB would also be crucial for the recovery from replication inhibition by upregulation of dinB after DNA damage is repaired.The biological significance of this checkpoint-like model for DinB should be further addressed in future experiments.

Fig. 1 .
Fig. 1.Effects of DinB overproduction on DNA replication.(A) DNA synthesis after induction of dinB expression.Two cultures of MK6980 (P BAD -dinB, thyA) cells were grown at 37°C to log phase in ECA medium containing [ 14 C]-thymine.To one of them, arabinose was added to 0.2% at time zero (closed circles), while the other was incubated in the absence of arabinose (open circles).Aliquots were withdrawn at the times indicated, and the amount of [ 14 C]-thymine incorporated into acid-insoluble material was measured in a liquid scintillation counter.Thymine incorporation was normalized to that at the 0-min time point.(B) Microarray analysis of chromosomal gene content.Sample and reference genomic DNA was prepared from MK6956 (P BAD -dinB) at 2 h after addition of arabinose to 0.2% and at stationary phase in the absence of arabinose, respectively.Gene dosage of the sample relative to the reference was normalized by the average gene dosage of the terminus region and is expressed on the y-axis on a log 2 scale.The x-axis represents the position of each ORF on the chromosome in kb, with the thrA gene at 0 kb.The replication origin oriC and the terminus TerB sites are marked as the positions of the adjacent genes mioC (closed triangle) and tus (open triangle), respectively.Each gray dot marks the relative gene dosage for an individual ORF.The black dots are smoothed values from a moving median with a window of 101 loci.

Fig. 2 .
Fig. 2. Kinetics of SOS induction during replication inhibition by DinB overproduction.All cells carried the pSK1002 (umuD::lacZ) plasmid and were grown exponentially.At time zero, arabinose (to 0.2%) was added to MK6956 (closed circles) to induce dinB expression, and UV (100 J m -2 ) was applied to MK7010 (closed triangles).Cell samples were taken at the time points indicated on the x-axis.The β-galactosidase activities of the samples were measured and are expressed as Miller units on the y-axis.Open and closed squares represent the SOS-defective SMR7467 (lexA3) and SOS-constitutive SMR7623 (lexA51) control strains, respectively.

Fig. 3 .
Fig. 3. Cell growth and the SOS response during replication inhibition by UV and excess DinB.(A) Cell viability.At time zero, arabinose (to 0.2%) was added to MK6956 (closed circles), and MK7010 cells were UV-irradiated (30 J m -2 , closed triangles), at 37°C.MK6956 and MK7010 were also grown in the absence of arabinose (open circles) and UV irradiation (open triangles), respectively.CFU ml -1 at each time point, relative to that at time zero, is represented on the y-axis.(B, C) LexA amounts in cells.Cell samples were taken at each time point and incubated with 50 μg ml -1 Cm at 37°C for 5 min, except for time zero (lane 1), and cellular LexA proteins were detected in (B) UV-irradiated MK7010 and (C) arabinose-treated MK6956.Total cell proteins at the 0-, 10-, 20-, 30-and 60-min time points (lanes 1 to 5) were separated by SDS-PAGE and analyzed by Western blotting with anti-RpoD monoclonal antibody (upper panels) and anti-LexA serum (lower panels).

Fig. 4 .
Fig.4.The SOS response during replication inhibition by Pol III dysfunction and DinB overproduction.After arabinose (final 0.2%) was added to exponentially growing cells in ECA medium at 30°C, incubation was continued at 30°C for 180 min except for MK7149 (dnaE486, P BAD -dinB) (150 min).At time zero, the OD 600 was adjusted to 0.1, the temperature was shifted to 42°C and Cm was added (final 50 μg ml -1 ).Aliquots of the cultures were taken at the times indicated above each image and chilled on ice.The cell samples were analyzed by Western blotting with anti-RpoD (top panels), anti-DinB (middle panels) and anti-LexA (bottom panels) antibodies.Samples are MK7147 (dnaE + ) in lanes 1 to 4, MK7146 (dnaE486) in lanes 5 to 8, MK7150 (dnaE + , P BAD -dinB) in lanes 9 to 12 and MK7149 (dnaE486, P BAD -dinB) in lanes 13 to 16.