Mutational Analysis of Residues in PriA and PriC Affecting Their Ability To Interact with SSB in Escherichia coli K-12

Escherichia coli PriA and PriC recruit the replication machinery to a collapsed replication fork after it is repaired and needs to be restarted. In vitro studies suggest that the C terminus of SSB interacts with certain residues in PriA and PriC to recruit those proteins to the repaired fork, where they help remodel it for restart. Here, we placed those mutations on the chromosome and tested the effect of mutating these residues in vivo. The priC mutation completely abolished function. The priA mutations had no effect by themselves. They did, however, display modest phenotypes in a priB-null strain. These phenotypes were partially suppressed by SSB overproduction. These studies give us further insight into the reactions needed for replication restart.

ABSTRACT

Escherichia coli PriA and PriC recognize abandoned replication forks and direct reloading of the DnaB replicative helicase onto the lagging-strand template coated with single-stranded DNA-binding protein (SSB). Both PriA and PriC have been shown by biochemical and structural studies to physically interact with the C terminus of SSB. In vitro, these interactions trigger remodeling of the SSB on ssDNA. priA341(R697A) and priC351(R155A) negated the SSB remodeling reaction in vitro. Plasmid-carried priC351(R155A) did not complement priC303::kan, and priA341(R697A) has not yet been tested for complementation. Here, we further studied the SSB-binding pockets of PriA and PriC by placing priA341(R697A), priA344(R697E), priA345(Q701E), and priC351(R155A) on the chromosome and characterizing the mutant strains. All three priA mutants behaved like the wild type. In a ΔpriB strain, the mutations caused modest increases in SOS expression, cell size, and defects in nucleoid partitioning (Par⁻). Overproduction of SSB partially suppressed these phenotypes for priA341(R697A) and priA344(R697E). The priC351(R155A) mutant behaved as expected: there was no phenotype in a single mutant, and there were severe growth defects when this mutation was combined with ΔpriB. Analysis of the priBC mutant revealed two populations of cells: those with wild-type phenotypes and those that were extremely filamentous and Par⁻ and had high SOS expression. We conclude that in vivo, priC351(R155A) identified an essential residue and function for PriC, that PriA R697 and Q701 are important only in the absence of PriB, and that this region of the protein may have a complicated relationship with SSB.

IMPORTANCE Escherichia coli PriA and PriC recruit the replication machinery to a collapsed replication fork after it is repaired and needs to be restarted. In vitro studies suggest that the C terminus of SSB interacts with certain residues in PriA and PriC to recruit those proteins to the repaired fork, where they help remodel it for restart. Here, we placed those mutations on the chromosome and tested the effect of mutating these residues in vivo. The priC mutation completely abolished function. The priA mutations had no effect by themselves. They did, however, display modest phenotypes in a priB-null strain. These phenotypes were partially suppressed by SSB overproduction. These studies give us further insight into the reactions needed for replication restart.

INTRODUCTION

DNA replication is a demanding task, made so by the sheer amount of DNA to be copied, millions to billions of base pairs per cell, and the fact that cellular metabolism places many obstacles in its way: transcription complexes and different types of DNA damage often cause replication fork collapse. The need for near-perfect copies of the genome to be passed to future generations prompted evolution of efficient ways of repairing broken replication forks and then restarting them (1–3). A key molecule in these reactions and in most DNA transactions (DNA replication, repair, and recombination) is the single-stranded-DNA-binding protein (SSB). It protects ssDNA from the chemical insults cellular metabolism may hurl at it while the DNA is being processed. Studies have shown that it is able to recruit many proteins to the DNA and that these partner proteins often help to remodel the SSB and/or ssDNA so that the DNA transaction can move forward (4, 5).

Escherichia coli SSB is a homotetramer. Each monomer has three parts: an N-terminal DNA binding core containing an oligonucleotide/oligosaccharide binding (OB) fold, an intrinsically disordered linker region and the C-terminal acidic tip (6–8). SSB can bind ssDNA in at least two major modes. The first is the SSB₆₅ mode, where the ssDNA is wrapped around all four SSB monomers, and the second is the SSB₃₅ mode, where only two of the four SSB monomers on average interact with the ssDNA (9, 10; reviewed in references 4 and 5). Biochemical and biophysical studies suggest that the SSB₆₅ mode is favored in DNA repair and recombination (11, 12), whereas the SSB₃₅ mode is favored in DNA replication (13, 14).

The last 9 residues of the SSB C terminus form a structure often referred to as the acidic tip. In addition to acidic residues, it also has hydrophobic amino acid residues. It is evolutionarily conserved among SSBs from different members of the Bacteria (15). In E. coli, sequence changes to this tip affect cell viability (16–18). It is proposed that SSB uses this structure to recruit more than a dozen proteins to the DNA (4). Structural studies reveal that several of these partner proteins have SSB-binding pockets: PriA (19), PriC (20), RecO (21), RecQ (22), RNase HI (23), the polymerase III (Pol III) χ subunit (24), and exonuclease I (ExoI) (15). The pockets of these proteins have amino acid residues that envelop the SSB F177 residue (a critical residue in the acidic tip). Mutations of the amino acid residues in these pockets cause perturbations in the interactions of these binding partners with SSB in vitro (15, 19–24).

Two large-scale proteomic screens and subsequent work identified at least 14 genome maintenance proteins that interact with SSB (25, 26; reviewed in references 4 and 5). Two of these SSB-interacting partners are the replication restart proteins (RRP) PriA and PriC. PriA and PriC are necessary to restart a repaired replication fork away from the origin of DNA replication in a structure-specific and cell cycle-independent manner. PriA, PriB, PriC, and DnaT combine to form three biochemically and genetically distinct pathways of replication restart called PriA-PriB, PriA-PriC, and PriC (27). A key feature of these pathways is the ability to recognize a repaired fork and remodel it to liberate a region of ssDNA so that the DnaC protein can load the DnaB replicative helicase onto the lagging-strand template at the point of the fork in the proper orientation.

PriA-SSB interactions has been extensively studied (28, 29). Crystallization of the Klebsiella pneumoniae PriA with a peptide comprising the acidic tip of SSB revealed that the tip is bound in an evolutionarily conserved pocket in PriA (19). The SSB-binding pocket of PriA is located at the interface of the 3′-OH DNA binding domain, the C-terminal domain, and the helicase core (Fig. 1A). It has been proposed that R697 and Q701 residues within the PriA pocket interact with the SSB tip (Fig. 1A). E. coli PriA R697 was analyzed in vitro for its importance in the PriA interaction with SSB using a single-molecule fluorescence resonance energy transfer (smFRET) assay that distinguishes between SSB binding modes. Bhattacharyya and colleagues demonstrated that PriA stabilized the SSB₃₅ binding mode, whereas a PriA341(R697A) mutant failed to alter SSB-DNA complexes (19). Taken together, these data support the idea that PriA R697 is important for the ability of PriA to be recruited to the ssDNA and remodel SSB.

FIG 1.

(A) Domain organization of Klebsiella pneumoniae PriA helicase as determined by X-ray crystallography (PDB code 4NL4). PriA architecture includes an N-terminal DNA-binding domain and a C-terminal helicase domain. The DNA-binding domain is divided into a 3′-DNA-binding domain (red) and a winged helix (orange). The helicase domain is composed of a helicase lobe 1 (blue) and lobe 2 (green). Two elements within lobe 2 include a Cys-rich region (purple) and a C-terminal domain (yellow). Residues involved in SSB binding are Arg-697 (magenta) and Gln-701 (cyan). (B) NMR structure of a full-length PriC protein from Cronobacter sakazakii (PDB code 2NCJ). Residues stabilizing its interaction with SSB are Arg-121 (green) and Arg-151 (yellow), which correspond to residues Arg-121 and Arg-155 of E. coli PriC.

PriC was shown to be an SSB-interacting partner through a dual tandem affinity purification-based proteomic screen and subsequent yeast two-hybrid assay (30). PriC R121 and R155 are two arginine residues of 25 tested that were found to be required for PriC’s interaction with the SSB C terminus using the yeast two-hybrid assay and isothermal titration calorimetry (30). A nuclear magnetic resonance (NMR) structure of Cronobacter sakazakii PriC sharing 41% sequence identity with E. coli PriC revealed that PriC R121 and R155 were on adjacent α-helices, on a single face of the protein. This is consistent with the idea that these two basic side chains are in close proximity and mediate an interaction with SSB (20) (Fig. 1B). Both E. coli priC(R121A) and priC351(R155A) mutations expressed from a plasmid did not complement a priC303::kan chromosomal null mutation (30). Using an smFRET assay as described above for the PriA-SSB interaction, PriC was shown to be recruited to the ssDNA substrate and stabilized the SSB₃₅ binding mode, whereas the PriC351(R155A) mutant protein failed to alter the SSB-DNA complexes (30). Together, these results demonstrate that PriC R155 is required for PriC binding to SSB and is important for PriC function both in vitro and in vivo.

In this work, we further investigated whether PriA and PriC residues identified as being important in the smFRET assay in vitro for interactions with SSB are essential for PriA and PriC function in vivo. To do this, we placed mutated priA and priC genes on the chromosome at their endogenous loci under regulation from their native promoters. Similar to priC303::kan, priC351(R155A) caused priA null-like phenotypes in ΔpriB strains. In contrast, three single priA mutations in the SSB-binding pocket, priA341(R697A), priA344(R697E), and priA345(Q701E), were not sufficient to produce any noticeable phenotype in vivo in an otherwise wild-type cell. In ΔpriB strains, however, all three mutations caused modest negative changes in several phenotypes tested. SSB overproduction could partially suppress these negative phenotypes for two of the three priA mutations. The data are consistent with the PriC R155 residue being important for SSB interactions and remodeling in vivo. It appears, however, that PriA has abilities similar to those of PriC to interact with and remodel SSB in vitro. While it is not clear yet exactly how PriA modulates SSB to efficiently help the restart process, it is clear that the role of PriA R697 and Q701 in SSB interactions and remodeling is more complicated than that of PriC R155, as they have a greater effect in the PriA-PriC pathway than in the PriA-PriB pathway.

RESULTS

It was previously shown that defects in replication restart mutants can be characterized by changes in cell viability, UV resistance, recombination ability, basal levels of SOS expression, cellular filamentation in the presence of sulB103 (see below), and defects in chromosome partitioning (reviewed in references 3 and 27). To facilitate this type of phenotypic analysis, we used phase-contrast and fluorescence microscopy on strains that have two types of reporter genes: a hupA::mCherry translational fusion gene that allows direct visualization of nucleoids without cell fixation or staining to assign a phenotype for nucleoids partitioning (24) and a sulAp-gfp transcriptional reporter gene inserted at the λ attachment site to measure SOS expression (31, 32). Neither reporter construct has any negative effect on the cell. All strains also contained a sulB103 mutation. This is an ftsZ allele that is not sensitive to the inhibitory effects of the SOS cell division inhibitor SulA (33). Thus, cellular filamentation (cell area) is an SOS-independent phenotype for these cells.

Single priA mutations within the SSB-binding pocket do not perturb cellular phenotypes.

The mutations priA341(R697A), priA344(R697E), and priA345(Q701E) in the SSB binding pocket were placed on the chromosome at the native locus for priA to facilitate the testing of cellular phenotypes. This was done using recombineering techniques in a “scarless” fashion (see Materials and Methods). The three priA mutations were chosen, as follows. priA341(R697A) is the mutation evaluated in the smFRET assay in vitro. priA344(R697E) is a charge reversal mutation of that same residue. It is predicted that the SSB acidic tip would not bind at all due to charge repulsion (21–24) on the acid tip. Finally, priA345(Q701E) mutates another residue indicated by the crystallographic studies to have an interaction with the SSB acidic tip (19).

priA-null mutants have severe phenotypes. They grow poorly (34, 35) and are Rec⁻ (36, 37), sensitive to UV irradiation (UV^s) (34), and sensitive to rich media (38). They also have high basal levels of SOS expression (35) and defects in nucleoid partitioning (31; reviewed in reference 27) and are filamentous even in the presence of a sulB103 mutation (31). If recruitment by SSB and its remodeling are essential for PriA function in replication restart, then it is predicted that all of these mutants would have phenotypes like that of a priA-null mutant. Table 1 and Fig. 2 show that all three mutant strains behave like the priA⁺ strain with regard to the phenotypes of average cell area, basal levels of SOS expression, and chromosome partitioning. The three mutant strains grew well in both minimal and rich media, were resistant to UV irradiation, and were recombination proficient (qualitative patch tests [data not shown]). We conclude that these three priA mutations within the SSB-binding pocket do not perturb PriA function in an otherwise wild-type cell.

TABLE 1.

Phenotypic characterization of three priA mutations within the SSB-binding pocket

Strain	Mutation	Avg cell area (pixels)a,b	SOS expression (RFI)b,c	Partitioning abilityd	Cell count
SS6321	+	572 ± 6	2.23 ± 0.04	++	758
SS11228	priA341(R697A)	548 ± 5	2.45 ± 0.07	++	1,103
SS11245	priA344(R697E)	644 ± 7	2.50 ± 0.08	++	1,005
SS11244	priA345(Q701E)	567 ± 5	2.26 ± 0.04	++	1,274
SS7064	priA2::kan	2,157 ± 161	13.80 ± 0.80	−	342

^aCells were grown as described in Materials and Methods. One pixel is equal to 0.0625 μm according to our microscopic parameters. The standard errors of the means for the entire cell population are shown.

^bStatistical analysis was done using Student's t test. The average cell areas of the priA341 (SS11228) and the priA344 (SS11245) strains are significantly different from that of the wild type (SS6321) (P < 0.005 and P < 0.001, respectively). The average cell area of the priA345 (SS11244) mutant is not significantly different from that of the wild type (P = 0.6). The differences in level of SOS expression between the priA341 and the priA344 mutants and the wild type are significant (P < 0.01 and P < 0.005, respectively). The level of SOS expression of the priA345 mutant is not significantly different from that of the wild type (P = 0.6).

^cLevels of SOS expression were monitored using a sulAp-gfp transcriptional reporter inserted into the λ attachment site (32). SOS expression was quantified as the average relative fluorescence intensity (RFI) of pixels for the entire population of cells normalized to the average fluorescence intensity of the background of the images. For each strain, the data were obtained from at least nine different fields taken on three different days. The standard errors of the means for the entire cell population are shown.

^dPartitioning ability was scored visually by assessing shape of nucleoids across the cell population using a hupA::mCherry translational fusion reporter (24). Examples of cells that partition their nucleoids similarly to the wild type and those that do not are shown in Fig. 2.

FIG 2.

The mutations R697A, R697E, and Q701E in the SSB-binding pocket of priA do not cause any significant changes in the cell phenotype. The log-phase cells were grown on minimal medium in slabs of 2% low-melting-point agarose at 37°C for 3 to 4 h before imaging. All strains have a sulAp-gfp transcriptional reporter for the SOS response and a hupA::mCherry translational fusion reporter showing the position of the nucleoids. The three images in each panel are a phase-contrast micrograph and images of nucleoids (red) and SOS (green). Different rows show results for different strains and genotypes: SS6321 (WT), SS11228 (priA341), SS11245 (priA344), SS11244 (priA345), and SS7064 (priA2::kan). Total magnification, ×1,000.

priA341, priA344, and priA345 alleles create modest phenotypes in a priB-null strain.

As mentioned above, there are three pathways of replication restart. The PriA-PriB and PriA-PriC pathways are efficient and redundant under most conditions (reviewed in reference 27). The PriC pathway, however, is not very efficient, as seen by the many phenotypes associated with a priA2::kan mutant. It was shown previously that some priA mutations [e.g., priA300(K230R) and priA301(C479Y)] have no or few phenotypes in an otherwise wild-type cell but do have priA2::kan null mutant phenotypes if they are combined with a ΔpriB mutation (39). The simplest explanation for this is that these priA mutations remove functions that are required only for the PriA-PriC pathway (and not the PriA-PriB pathway). For this reason, we tested whether the addition of ΔpriB302 would lead to measurable phenotypes for the priA341(R697A), priA344(R697E), and priA345(Q701E) mutant strains. This was done by introducing ΔpriB302 with an appropriately placed Tn10 marker into strains with the priA mutations. Tetracycline-resistant transductants were selected and then screened by PCR for the presence of the deletion. Table 2 shows that in the ΔpriB302 strain, all three priA mutants had decreased abilities to perform PriA functions. priA344(R697E) showed the largest defect in partitioning ability and the largest increases in average cell area (4-fold) and basal levels of SOS expression (3-fold) of the three mutants. The increases in average cell area and basal levels of SOS expression were 54% and 77%, respectively, for priA341(R697A) and 29% and 45%, respectively, for priA345(Q701E). In a similar way, the ΔpriC307 mutation was introduced into the three priA mutant strains to test if any of the priA mutations affected only the PriA-PriB pathway. Table 2 shows also that combination of any of the three priA alleles with the ΔpriC307 mutation led to no change in the phenotype relative to the wild type. We conclude that priA341(R697A), priA344(R697E), and priA345(Q701E) affect some activity of PriA necessary in the PriA-PriC pathway but not in the PriA-PriB pathway.

TABLE 2.

Effects of priA mutations on cell area and level of SOS expression in the ΔpriB and ΔpriC strains

Strain	priA	priB	priC	Avg cell area (pixels)a ,b	SOS (RFI)b,c	Partitioning abilityd	Cell count
SS6321	+	+	+	572 ± 6	2.23 ± 0.04	++	758
SS9116	+	ΔpriB302	+	584 ± 7	2.42 ± 0.09	++	777
SS11233	+	+	ΔpriC307	575 ± 6	2.13 ± 0.04	++	1,023
SS11228	priA341	+	+	548 ± 5	2.45 ± 0.07	++	1,103
SS11230	priA341	ΔpriB302	+	846 ± 31	4.33 ± 0.21	+	490
SS11232	priA341	+	ΔpriC307	573 ± 5	2.38 ± 0.04	++	1,246
SS11245	priA344	+	+	644 ± 7	2.50 ± 0.08	++	1,005
SS11255	priA344	ΔpriB302	+	2,418 ± 139	7.81 ± 0.34	−	500
SS11253	priA344	+	ΔpriC307	700 ± 25	2.30 ± 0.03	++	691
SS11244	priA345	+	+	567 ± 5	2.26 ± 0.04	++	1,274
SS11251	priA345	ΔpriB302	+	730 ± 22	3.28 ± 0.10	+	715
SS11249	priA345	+	ΔpriC307	535 ± 5	1.95 ± 0.01	++	1,270

^aCells were grown as described in Materials and Methods. One pixel is equal to 0.0625 μm according to our microscopic parameters. The standard errors of the means for the entire cell population are shown.

^bThe average cell areas and levels of SOS expression of the priA341 ΔpriB302 (SS11230), priA344 ΔpriB302 (SS11255) and priA345 ΔpriB302 (SS11251) double mutants are significantly different (P < 0.001) from those of the priA single mutants (SS11228, SS11245, and SS11244, respectively) and the ΔpriB302 single mutant (SS9116) (P < 0.001) by Student's t test.

^cLevels of SOS expression were monitored using a sulAp-gfp transcriptional reporter inserted into the λ attachment site (32). SOS expression was quantified as the average relative fluorescence intensity (RFI) of pixels for the entire population of cells normalized to the average fluorescence intensity of the background of the images. For each strain, the data were obtained from at least nine different fields taken on three different days. The standard errors of the means for the entire cell population are shown.

^dPartitioning ability was scored visually by assessing the shape of nucleoids across the cell population using a hupA::mCherry translational fusion reporter (24). Examples of cells that partition their nucleoids similarly to the wild type and those that do not are shown in Fig. 2.

SSB overproduction partially suppresses some phenotypes associated with priA341, priA344, and priA345 mutations in the ΔpriB302 strain.

To better understand the nature of the defect in the priA priB strains presented in Table 2, we wanted to test if we could create conditions that would suppress the phenotypes of the priA priB mutants. Two lines of thinking suggested that overproduction of SSB might be able to do this. The first is that PriB and SSB share sequence similarity and an oligonucleotide/oligosaccharide ssDNA-binding fold (40, 41) and that phylogenetic data suggest that a priB gene arose as a result of a an ssb gene duplication event (42). The second is that it is possible that the priA mutations merely decreased the ability of SSB to bind PriA instead of eliminating its ability to bind completely. Therefore, increasing the concentration of SSB might push the equilibrium in favor of binding. In general, however, SSB overproduction could have negative effects on the cell (43, 44).

To overproduce SSB, we introduced a strong constitutive promoter and an improved ribosome binding site (RBS) upstream of the ssb gene as previously described for radA and recN (45, 46). Briefly, the strong promoter and optimized RBS were linked to a flippable chloramphenicol resistance gene and inserted just upstream of the ssb gene (see Materials and Methods). This construct was called ssbop. The priA341 ΔpriB302 ssbop, priA344 ΔpriB302 ssbop, and priA345 ΔpriB302 ssbop triple mutants were constructed in a two-step process. First, ssbop was transduced into the priA341, priA344, and priA345 single mutant strains by selecting for chloramphenicol resistance, and then the ΔpriB302 mutation was introduced with tetracycline resistance. Table 3 shows that ssbop positively impacted two priA priB double mutants and negatively impacted the third. The largest positive effect was seen with the priA344 ΔpriB302 double mutant. Here, all of the phenotypes improved: the defect of nucleoid partitioning was less apparent, the cells had a smaller average cell area (3-fold) and the levels of SOS expression were 2-fold less. Overproduction of SSB rescued only the high SOS expression of the priA341 ΔpriB302 double mutant (a change in relative fluorescence intensity [RFI] from 4.33 to 2.42). There was no significant effect on the average cell area. ssbop failed to suppress any of the phenotypes of the priA345 ΔpriB302 mutation. In fact, the presence of ssbop increased the SOS levels by 30%. Table 3 also shows that for the phenotypes tested here, ssbop had no negative impact on the cell even though overproduction of SSB from a plasmid has been shown to negatively impact cellular function related to SOS induction and recombination (43, 44). We conclude that overproduction of SSB is able to suppress some of the negative phenotypes of the priA341 ΔpriB302 and priA344 ΔpriB302 double mutants but not of the priA345 ΔpriB302 mutant.

TABLE 3.

Effect of SSB overproduction on the priA mutants in the ΔpriB302 background

Strain	priA	priB	ssb	Avg cell area (pixels)a ,b	SOS (RFI)b ,c	Partitioning abilityd	Cell count
SS6321	+	+	+	572 ± 6	2.23 ± 0.04	++	758
SS11234	+	+	ssbop	576 ± 5	2.18 ± 0.02	++	1,381
SS9116	+	ΔpriB302	+	584 ± 7	2.42 ± 0.09	++	777
SS11236	+	ΔpriB302	ssbop	658 ± 7	2.24 ± 0.02	++	983
SS11228	priA341	+	+	548 ± 5	2.45 ± 0.07	++	1,103
SS11235	priA341	+	ssbop	557 ± 5	2.08 ± 0.01	++	1,328
SS11230	priA341	ΔpriB302	+	846 ± 31	4.33 ± 0.21	+	490
SS11237	priA341	ΔpriB302	ssbop	940 ± 24	2.42 ± 0.05	+	1,358
SS11245	priA344	+	+	644 ± 7	2.50 ± 0.08	++	1,005
SS13467	priA344	+	ssbop	595 ± 5	2.73 ± 0.04	++	1,298
SS11255	priA344	ΔpriB302	+	2,418 ± 139	7.81 ± 0.34	−	500
SS13468	priA344	ΔpriB302	ssbop	902 ± 21	4.49 ± 0.10	+	1,389
SS11244	priA345	+	+	567 ± 5	2.26 ± 0.04	++	1,274
SS11256	priA345	+	ssbop	521 ± 4	2.48 ± 0.06	++	1,648
SS11251	priA345	ΔpriB302	+	730 ± 22	3.28 ± 0.10	+	715
SS11257	priA345	ΔpriB302	ssbop	719 ± 17	4.68 ± 0.18	+	1,312

^aCells were grown as described in Materials and Methods. One pixel is equal to 0.0625 μm according to our microscopic parameters. The standard errors of the means for the entire cell population are shown.

^bStatistical analysis was done using Student's t test. The average cell area of the priA341 ΔpriB302 ssbop (SS11237) triple mutant is not significantly different from that of the priA341 ΔpriB302 (SS11230) double mutant (P = 0.03). The levels of SOS expression between SS11237 and SS11230 differ significantly (P < 0.001). The differences in the average cell area and level of SOS expression between the priA344 ΔpriB302 ssbop (SS13468) triple mutant and the priA344 ΔpriB302 (SS11255) double mutant are significant (P < 0.001). The average cell area of the priA345 ΔpriB302 ssbop triple mutant (SS11257) is not significantly different (P = 0.7) from the average cell area of the priA345 ΔpriB302 double mutant (SS11251). The difference in levels of SOS expression between SS11257 and SS11251 is significant (P < 0.001).

^cLevels of SOS expression were monitored using a sulAp-gfp transcriptional reporter inserted into the λ attachment site (32). SOS expression was quantified as the average relative fluorescence intensity (RFI) of pixels for the entire population of cells normalized to the average fluorescence intensity of the background of the images. For each strain, the data were obtained from at least nine different fields taken on three different days. The standard errors of the means for the entire cell population are shown.

^dPartitioning ability was scored visually by assessing shape of nucleoids across the cell population using a hupA::mCherry translational fusion reporter (24). Examples of cells that partition their nucleoids similarly to the wild type and those that do not are shown in Fig. 2.

A strain with priC351(R155A) on the chromosome acts like a priC-null mutant in vivo.

priC351(R155A) was not able to complement priC303::kan in a plasmid-based assay in vivo (30). We wanted to measure priC351’s phenotypes when it was expressed from its native locus on the chromosome. To do this, we first placed this mutation on the chromosome using recombineering techniques as described in Materials and Methods. We then characterized the phenotypes of priC351(R155A) by standard phase-contrast and fluorescence microscopy. Table 4 shows that the single priC351(R155A) mutant behaved as the wild type for the phenotypes tested.

TABLE 4.

Phenotypic characterization of the priC351 mutation within the SSB-binding pocket

Strain	priA	priB	priC	Avg cell area (pixels)a ,b	SOS (RFI)b ,c	Cell area < 300		Cell area > 300		Partitioning abilityd	Cell count
						SOS (RFI)	% population	SOS (RFI)	% population
SS6321	+	+	+	184 ± 1	3.1 ± 0.1	2.9	98	9.3	2	++	6,489
SS12997	+	+	priC351	197 ± 1	4.4 ± 0.1	4.0	96	12.1	4	++	2,829
SS9114	+	+	priC303::kan	193 ± 1	2.8 ± 0.1	2.7	96	4.6	4	++	2,265
SS9116	+	ΔpriB302	+	187 ± 2	4.7 ± 0.2	4.3	96	15.8	4	++	1,654
SS13407	+	ΔpriB302	priC351	806 ± 39	22.7 ± 0.7	17.2	56	29.8	44	−	1,404
SS13420	+	ΔpriB302	priC303::kan	538 ± 34	13.6 ± 0.6	10.3	64	19.4	36	−	1,347
SS7064	priA2::kan	+	+	675 ± 24	24.5 ± 1.0	16.8	46	31.0	54	−	988

^aCells were grown as described in Materials and Methods. One pixel is equal to 0.1073 μm according to our microscopic parameters. The standard errors of the means for the entire cell population are shown.

^bStatistical analysis was done using Student's t test. The average cell area of the priC351 (SS12997) was not statistically significantly different from the average cell area of the priC303::kan (SS9114) mutant (P = 0.03). The difference in levels of the SOS expression between SS12997 and SS9114 is significant (P < 0.001). The average cell area and level of SOS expression of the ΔpriB302 priC351 (SS13407), ΔpriB302 priC303::kan (SS13420), and priA2::kan (SS7064) mutants are significantly different (P < 0.001) from those of the wild type (SS6321).

^cLevels of SOS expression were monitored using a sulAp-gfp transcriptional reporter inserted into the λ attachment site (32). SOS expression was quantified as average relative fluorescence intensity (RFI) of pixels for the entire population of cells normalized to the average fluorescence intensity of the background of the images. For each strain, the data were obtained from at least nine different fields taken on three different days. The standard errors of the means for the entire cell population are shown.

^dPartitioning ability was scored visually by assessing shape of nucleoids across the cell population using a hupA::mCherry translational fusion reporter (24). Examples of cells that partition their nucleoids similarly to the wild type and those that do not are shown in Fig. 2.

The above result was expected, as priB and priC have redundant functions (47). Initial characterization of the ΔpriB302 priC303::kan double mutant showed that it grew very poorly, even more poorly than a priA2::kan mutant (47). The small colonies that did grow often acquired suppressor mutations (47). Here, we report how we could construct and grow these very poorly growing mutants reproducibly so that we could further characterize them. To do this, we transduced the ΔpriB302 mutation into the priC mutant, incubating the transductant on minimal medium plates at 37°C for 3 to 4 days. Very small transductants were formed. These transductants were then subjected to 2 rounds of colony purification (incubated for 48 h each on minimal medium). By purifying the small colonies each time, we were able to consistently get small colonies without too many large ones. Individual small colonies were then used to make patches on minimal medium plates that could then be used to reproducibly inoculate and grow cultures. Since these strains grew very slowly and there was a high probability of suppressors arising in the population, we tested our cultures at the end of all experiments by streaking cells on minimal medium plates to be sure that only the small-colony phenotype was seen and no suppressors (large colonies) arose in that culture.

Table 4 shows the results of the microscopic characterization of the priBC double mutants. These cells were imaged at a magnification of ×600 instead of ×1,000, the magnification used for the strains characterized in Tables 1 to 3. Hence, the average cell area and levels of SOS expression are not directly comparable between tables. Microscopic analysis of cultures of ΔpriB302 priC351(R155A) and ΔpriB302 priC303::kan cells grown in minimal medium showed that the populations of cells were composed of both wild-type-looking and filamentous cells (Fig. 3). Table 4 shows that in general, ΔpriB302 priC351(R155A) and ΔpriB302 priC303::kan mutants were similar to the priA2::kan mutants. They showed 3- to 4-fold increases in average cell area and 3- to 5-fold increases in SOS expression, with the ΔpriB302 priC351 mutant showing greater increases than the ΔpriB302 priC303::kan mutant for both phenotypes. Both strains showed the partitioning defect. Binning of the data into two groups based on cell area (less than or equal to 300 pixels in cell area with cells that are greater than 300 pixels) showed that between 36 and 44% of the cells were filamentous. The level of SOS expression in the filamentous cells with areas greater than 300 pixels were about 2-fold higher on average than the level in cells with areas less than or equal to 300 pixels (Fig. 3 and Table 4). Both groups had higher values than the wild type or either single mutant in both categories. The data support the previous findings that priC351(R155A) does not complement a priC303::kan mutation and show that this mutation tends to produce slightly more detrimental phenotypes than the priC303::kan mutation. It is also seen that the priBC double mutants have phenotypes similar to that of a priA null mutant, with two distinct populations of cells (normal-length and highly filamentous cells), high basal levels of SOS expression, and defects in nucleoid partitioning.

FIG 3.

Phenotypes of the ΔpriB302 priC351 and ΔpriB302 priC303::kan double mutants. (Left) The mutation R155A in the SSB-binding pocket of priC does not have any noticeable phenotype by itself compared to the wild-type strain. In contrast, both ΔpriB302 priC351 and ΔpriB302 priC303::kan double mutants show many filamentous cells that are highly induced for SOS and have poorly partitioned nucleoids. Overnight cultures were grown on minimal medium in slabs of 2% low-melting-point agarose at 37°C for 3 to 4 h before imaging. All strains have a sulAp-gfp transcriptional reporter for the SOS response and a hupA::mCherry translational fusion reporter showing the position of nucleoids. The images are overlays of a phase-contrast image and images showing nucleoids (red) and SOS (green). Bars, 10 μm. Total magnification, ×600. (Middle) Binning of the cell area from 100 pixels (lowest) to 14,000 pixels (highest), with 50-pixel increments. The population of cells representing less than 1% was combined in one column (15th through 548th increments). (Right) Distribution of individual cells in the entire population by cell area and RFI (SOS). Different rows show results for different strains and genotypes: SS6321 (WT), SS12997 (priC351), SS9114 (priC303::kan), SS9116 (ΔpriB302), SS13407 (ΔpriB302 priC351), SS13420 (ΔpriB302 priC303::kan) and SS7064 (priA2::kan).

DISCUSSION

In vitro, the replication restart proteins PriA and PriC have been shown to interact with SSB through the C-terminal acidic tip (19, 20) and remodel the conformation of the SSB on ssDNA from an SSB₆₅ mode to a SSB₃₅ mode (9, 10; reviewed in references 4 and 5). This work included the identification of certain residues in PriA and PriC that were required for the in vitro reactions (19, 20). It has been hypothesized that the in vivo importance of these reactions is that a region of ssDNA is liberated so that DnaC can load DnaB onto the lagging strand template during replication restart. Only the PriC mutants were shown not to be able to complement a chromosomal null mutation expressed from a plasmid (20). At the time of the previous publications, the identified residues in the PriA’s SSB-binding pocket had not been tested for complementation in vivo (19). Thus, this work continued those studies and tested the importance of the PriA and PriC residues in vivo by first placing them on the chromosome and then characterizing them using standard assays for replication restart mutants. As expected, the mutations in PriC that destroyed the ability to interact with SSB and remodel the SSB on ssDNA in vitro effectively result in a null mutant phenotype in vivo. Unexpectedly, however, we found that the three mutations tested in the PriA SSB-binding pocket did not cause a null mutant phenotype as single mutations in an otherwise wild-type cell. This was unexpected because they had the same effects as mutations in the SSB-binding site of PriC on the assay in vitro and it was assumed, supported by experiments and rationale mentioned above, that the SSB-PriA interactions would be important for PriA’s function in vivo.

The removal of PriB unmasked a requirement for the SSB-binding pocket in the PriA-PriC pathway for the three priA mutant strains. This too was unexpected, because PriC, as well as PriA, can interact with SSB and remodel it, as shown in vitro. This prompts the questions of why the PriA SSB-binding pocket is required only in the PriA-PriC pathway (and not the PriA-PriB pathway) and why the PriC SSB-binding pocket is not sufficient in vivo in the priB mutant. While the answers to these questions are not clear, one possibility is that the interaction of PriA with SSB in replication restart is required before PriC is loaded. Another is that the SSB tetramer forms a bridge between the PriA and PriC proteins via two of the acidic tips.

Another unexpected observation was that overproduction of SSB was able to suppress some of the phenotypes of some of the priA single mutants in the ΔpriB strain. The fact that more suppression was seen with some mutations than others suggests that the mode of suppression is not a new general indirect pathway (like dnaC809 suppression of priA-null mutations [37]) but is instead some sort of direct interaction. It is possible that the mutations decrease the ability of PriA to bind SSB to different degrees, and then the increased concentration of SSB is able to overcome this decreased interaction to various degrees by mass action. An alternative explanation for the SSB suppression in the absence of PriB could be that PriB and SSB have similar structures and that increased amounts of SSB are somehow able to partially substitute for PriB in this situation.

This is the first study in which priBC double mutants have been characterized in more depth other than showing a severe qualitative growth defect relative to either of the two single mutations or the priA2::kan mutation (47). The nucleoid partitioning ability, the two distinct subpopulations of cells, and the high basal levels of SOS expression are highly reminiscent of priA-null mutants (31). It should be remembered, however, that a distinguishing characteristic of the priBC mutants, versus the priA-null mutants, is that they take a much longer time for colony formation.

Given the current model of E. coli having three pathways and the assumption that replication restart is essential for cell viability, one must rationalize the marginal viability of priBC mutants. One simple idea is suggested by the two populations seen in the microscopic images of priBC mutants that contain both wild-type-looking cells and filamentous cells with Par⁻ nucleoids. This suggests that when a replication fork breaks and is repaired, it cannot simply be restarted because all three pathways of restart are inactivated. What then happens to this repaired fork is critical for the phenotype seen. It is possible that the newly synthesized arms of the replicating chromosome are degraded so that a single circular chromosome remains and it can begin replication again at oriC (48–50). This cell would likely be similar in size to a normal cell with a single nucleoid and might show higher-than-normal levels of SOS expression (these types of cells are seen [Fig. 3]). Alternatively, the abandoned repaired fork is likely still a substrate for the DNA replication and recombination machineries. Its further processing does not, however, yield a whole chromosome that can be segregated to a daughter cell. Instead, it produces a contiguous mass of DNA along the length of the filamentous cell. This process might be similar to late DNA replication of bacteriophage T4, where recombination primes DNA replication to make branched structures (51). This model is supported by the observation that the Par⁻ phenotype in a priA2::kan mutant is largely suppressed by removing the recombination machinery (31). These priA recA cells have reasonably partitioned nucleoids but still filament and grow poorly. They can, however, restart forks because of the presence of the PriC pathway. The marginal viability of priBC mutants is then seen as the low probability that the chromosome can be replicated without suffering an insult that requires restart. While models predict that this would be infrequent, it could happen just often enough to explain the very slow growth of the priBC colony or culture when the most favorable of conditions are used.

Finally, it is worth emphasizing that the ability to control and modulate SSB binding and ssDNA during replication restart must be extremely important, as the cell has evolved three separate mechanisms to do this for the three different pathways of replication restart. For the PriA-PriB, PriA-PriC, and PriC pathways, these are the PriA-PriB interaction presumably through the cysteine-rich region (52) (remembering that SSB and PriB both bind ssDNA and share structure and evolutionary history [see above]), the PriA-SSB interaction through the SSB binding pocket, and PriC-SSB interaction, respectively.

MATERIALS AND METHODS

Strains, media, and other reagents.

All bacterial strains are derivatives of Escherichia coli K-12 and are characterized in Table S1 in the supplemental material. The strains were generated using either linear transformation or P1 transduction according to previously described protocols (53, 54). Transformants and transductants were selected on 2% agar plates containing either Luria broth or 56/2 minimal medium supplemented with 0.2% glucose, 0.001% thiamine, 0.02% arginine, 0.005% histidine, 0.02% proline, and appropriate antibiotics (53). Ampicillin was used at 50 μg/ml, anhydrotetracycline hydrochloride at 5 μg/ml, chloramphenicol at 25 μg/ml, kanamycin at 50 μg/ml, and tetracycline at 10 μg/ml. The cells were purified on the same type of medium on which they were selected and grown at 30°C or 37°C. l-Arabinose was used for induction of the λ Red expression plasmid pKD46 at a final concentration of 0.5% (wt/vol). All plasmids and oligonucleotide primers used in this work are described in Tables S1 and S2, respectively.

Mutagenesis of the priA gene.

Mutations in the priA gene, priA341(R697A), priA344(R697E), and priA345(Q701E), were introduced via scarless site-directed mutagenesis using a method described by Blank and colleagues (55). The method involves the combination of the Red recombinase system of phage λ and counterselection with the help of I-SceI endonuclease introducing double-strand breaks in DNA of the unsuccessful recombinants. Oligonucleotide primers prSJS1511 and prSJS1512 (Table S2) were used to amplify an I-SceI recognition site and a chloramphenicol resistance cassette between two Flp recombination target (FRT) sites (I-SceI FRTcat) from pWRG100 template plasmid. These primers contained 40-bp extensions at the 5′ ends homologous to regions immediately upstream and downstream of a codon of the priA gene encoding the R697 residue. After amplification, a 1,080-bp PCR product was integrated within the priA gene on the chromosome by linear transformation into SS11301 strain using standard recombineering methods (54). Transformants were selected for chloramphenicol resistance on rich medium at 37°C. The proper insertion of the cassette was verified by PCR with prSJS1511 and prSJS1512. The mutant cells were cured of pKD46 and tested for ampicillin sensitivity. The resulting strain was designated SS11225. This strain was then chemically transformed with pWRG99 and selected for ampicillin resistance on rich medium at 30°C to give SS11226.

Pairs of 70-mer or 80-mer DNA oligonucleotides were designed for site-directed mutagenesis of the priA gene: prSJS1498 and prSJS1513 for priA341, prSJS1560 and prSJS1561 for priA344, and prSJS1558 and prSJS1559 for priA345. These oligonucleotides contain mutated priA alleles in the central part flanked by the homologous sequences used for integration of the I-SceI FRTcat cassette. In addition, the priA341 and priA345 alleles contain a novel BsiWI restriction site introduced at V698 to facilitate initial screening of mutants via restriction fragment length polymorphism analysis. For each pair, oligonucleotides were hybridized to form DNA duplexes. These duplexes were then transformed into electrocompetent SS11226 cells and integrated within the priA gene to replace the I-SceI FRTcat cassette using λ Red recombinase genes expressed from pWRG99 plasmid (55). Transformants were selected for ampicillin resistance on rich medium containing 5 μg/ml of anhydrotetracycline hydrochloride at 30°C and screened for chloramphenicol sensitivity. Successful recombinants were verified by PCR of genomic DNA with prSJS1235 and prSJS1330 followed by digestion with BsiWI (if applicable) and finally by DNA sequencing with prSJS1235, prSJS1237, and prSJS1330. The strains harboring the priA341, priA344, and priA345 alleles were designated SS11227, SS11241, and SS11242, respectively. The cells were cured of pWRG99 by P1 transduction of the mutations into SS11217 to create SS11228 (priA341), SS11245 (priA344), and SS11244 (priA345).

Mutagenesis of the priC gene.

priC351(R155A) was introduced in the priC gene using a method similar to the one applied for an insertion of the I-SceI FRTcat cassette into the priA gene. Oligonucleotide primers prSJS1638 and prSJS594 were used to amplify a mutated priC allele and a kanamycin resistance cassette from SS4395. prSJS1638 was designed with homology to the priC gene and contains a priC351 allele in its second half and a novel Sau96I site to facilitate screening. A 1,537-bp PCR fragment was used to transform electrocompetent SS11301 cells using standard recombineering methods (54). Transformants were selected for kanamycin resistance on minimal medium at 37°C. A construct with priC351(R155A) was verified by PCR analysis with prSJS626 and prSJS625. The kanamycin resistance marker was removed from the chromosome by Flp-mediated recombination using plasmid pCP20 (56) to give strain SS12677. The presence of the priC351 mutation was confirmed by DNA sequencing with prSJS718 and prSJS594. ybaM⁺ was then reintroduced on the chromosome next to priC351 using a flippable chloramphenicol resistance cassette and a method similar to the one described above. Briefly, a ybaM⁺ FRTcat fragment (1,218 bp) was generated in a crossover PCR with a ybaM gene amplified from strain JC13509 with prSJS1694 and prSJS1696 and the chloramphenicol resistance cassette amplified from the (cat) recNop construct (45) with prSJS1695 and prSJS1697. This fragment was transformed into electrocompetent SS12977 cells and placed downstream of the priC351(R155A) mutation using standard recombineering methods (54). The chloramphenicol resistance marker was removed using Flp-mediated recombination as previously described (56) to give strain SS12997. The construct was confirmed by DNA sequencing with prSJS718 and prSJS608.

Microscopy and image analysis.

Cells were grown overnight in 56/2 minimal medium. Then, 175 μl of the culture was added to 3 ml of 56/2 minimal medium and grown for 3 h to early log phase. Three microliters of the log-phase or overnight culture was loaded onto a 2% agarose slab prepared from 56/2 minimal medium and low-melting-point agarose. A coverslip was mounted on top, and the slides were incubated for 3 to 4 h at 37°C. Cells were visualized using a Nikon E600 microscope equipped with automated filter wheels, shutters, a CoolLED light source, and an ORCA-ER camera. Phase-contrast and fluorescent images were taken for at least 9 different fields of view (3 fields on 3 different days). These images were analyzed with the following software: I-Vision (BioVision Technologies, Inc.), OpenLabs 5.5.1 (Improvision, Inc.), Oufti (57), SuperSegger (58), and MatLab R2016a and R2019a (MathWorks, Inc.). Individual cells were outlined using Oufti or SuperSegger. Strains were analyzed for number of cells, cell area, and intensity of the fluorescent signal using specially written Matlab programs. Statistical analysis of data was performed using Student's t test. A cutoff P value of <0.001 was used to determine significance. P values are reported in the footnotes for each table. The cell area is given as a total number of pixels in a cell, where one pixel is equal to 0.0625 μm at a ×1,000 total magnification and 0.1073 μm at a ×600 total magnification.