Identification and Characterization of a Novel Allele of Escherichia coli dnaB Helicase That Compromises the Stability of Plasmid P1

Reagents, media, and buffers.The suppliers for reagents were as follows: restriction enzymes and enzymes for DNA manipulations, New England Biolabs and Stratagene; ethane methylsulfonate (EMS) and antibiotics, Sigma; 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal), Roche; and radioisotopes, Amersham Biosciences. Bacterial cells were grown in Luria-Bertani (LB) medium or M9 medium (33). Antibiotics, when present, were used at the following concentrations: ampicillin, 100 μg/ml; chloramphenicol, 25 μg/ml; kanamycin, 25 μg/ml; and rifampin, 100 μg/ml. M9Xgal plates were M9 agar plates supplemented with glucose (0.2%), Casamino Acids (0.2%), tryptophan (50 μg/ml), adenine (70 μg/ml), thymine (40 μg/ml) and X-Gal (40 μg/ml). MinA buffer is 0.09 M potassium phosphate buffer (pH 7) with 1 g of (NH₄)₂SO₄ per liter and 0.5 g of sodium citrate per liter (33).

Bacterial strains, bacteriophages, and plasmids. E. coli K-12 strains and genotypes are listed in Table 1. E. coli N99 was used as the background strain for all genetic studies, and DH5 was used for plasmid construction. P1rev6 and P1c17 were the transducing phages used for strain construction (52). λNK1316 and λNK1323 encoded miniTn10kan and miniTn10tet, respectively (26). λlacI was constructed by cloning the lacI^q gene from pMC9 (8) (the EcoRI fragment with HindIII synthetic linkers) into the HindIII site of the i²¹ phage λD69 (34).

To construct E. coli N99 lacI::tet, a Gal⁺ derivative of N99 was first randomly mutagenized by miniTn10tet insertion by using λNK1323 infection as described previously (26). lacI mutants were isolated based on their ability to grow on phenyl-β-d-galactoside as the sole carbon source. Disruption of the lacI gene by miniTn10 was confirmed by Southern hybridization. The lacI::tet allele was introduced into N99 by transduction with P1rev6 to create strain BF226. The lacI::cam allele was constructed by linear transformation (62) by using a fragment of pMC9 in which the lacI coding sequence was interrupted at its MluI site by a chloramphenicol resistance cassette. The lacI::cam allele was then also transduced into N99, creating strain BF826.

Alleles of dnaB were linked to miniTn10kan as follows. A library of miniTn10kan insertions in the E. coli chromosome was created by infecting λNK1316 (26) into N99 and pooling all Kan^r progeny. A P1rev6 lysate was prepared with the library and used to transduce N99 I1-18 cells, selecting for Kan^r progeny that grew at 42°C (dnaB⁺ miniTn10kan). This miniTn10 insertion, which was over 90% linked to the dnaB gene, was then linked to all dnaB alleles used in this study (except dnaB252 [Table 1]) and used to move them into N99 and/or N99 lacI::tet. The dnaB252 mutation was already linked to Tn10(tet) in strain RS162 and was moved into N99 and N99 lacI::cam.

N99 recA56 dnaB277 miniTn10kan was constructed by first transducing an srl::Tn10 marker from an srl::Tn10 derivative of MC4100 (50) into N99 dnaB277 miniTn10kan and selecting for Tet^r colonies. Next, the recA56 allele was transduced from MC4100 recA56 srl⁺ into N99 dnaB277 miniTn10kan with P1c17, selecting for Srl⁺ transductants. One-half of the Srl⁺ transductants (srl is 50% linked to recA) were confirmed to be recA by UV sensitivity screening.

Individual plasmids used in this study are listed in Table 2. pBEF224 was constructed by replacing the P1 KpnI fragment in pLG49 (which contains rep and par [20]) with the fragment from λ-mini-P1parS_md2 (19). pBEF231 was constructed by replacing the Cam^r cassette of pBEF224 with a 1.2-kb Kan^r cassette from pUC4-K. pAJM4 was constructed by replacing the kan gene in pALA318kan with lacI^q from pMC9 (as BamHI fragments) and then replacing the cam region (a PstI fragment) with the 1.2-kb kan cassette from pUC4K. pDnaB2 was constructed as follows. A 1,676-bp fragment that included dnaB (with its promoter) was amplified from N99 cells by using primers 5′-GCGCGACGTCCCACCGTGCTTGTGAAATTC-3′ and 5′-CGCGCCCGAGACCAGTTCACGAAGACGTTG-3′ and was inserted into pBR322 at its EcoRV site to produce pDnaB. The former primer contained an AatII restriction site; cleavage of pDnaB with AatII and subsequent religation deleted the pBR322 tet promoter and created pDnaB2. pDnaB277 was made in the same way as pDnaB2, except that the amplified DNA product from N99 I1-18 was used. To construct pBEF261, we used PCR primers 5′-CGCGAGATCTGCCATAGTGAATGGAGTTAC-3′ and 5′-CGCGAGATCTGGCTGAATCCCGCGTCTTTAAAATT-3′ to amplify the entire pDnaB2 plasmid except the region in dnaB between the 1st and 158th codons. Each oligonucleotide also contained a BglII site, and both oligonucleotides were used to prime DNA synthesis with Pfu Turbo DNA polymerase. The PCR products were digested with BglII and religated to form pBEF261. Finally, a plasmid library of E. coli genes was constructed by insertion of a partial Sau3AI digest of E. coli DH5Δlac DNA into the BglII site of pST52 (D. Xu and B. Funnell, unpublished results).

E. coli mutagenesis and isolation of mutations that compromise P1 stability. E. coli N99 lacI::tet(pBEF224) cells were mutagenized with EMS essentially as described by Miller (33). Briefly, cells were grown until the A₆₀₀ was approximately 0.6 in LB medium with chloramphenicol, collected by centrifugation, washed twice in MinA buffer, and resuspended in 0.75 volume of MinA buffer. The mixture was then treated with 7.5 μl of EMS per 0.5 ml of culture for 45 min at 37°C. Cells were again collected by centrifugation, washed twice in MinA buffer, and resuspended in an equal volume of MinA buffer. The culture was titrated to measure viability and diluted 20-fold into LB medium for overnight growth at 30°C. Cells were then plated on M9Xgal plates to screen for plasmid stability mutants and on plates containing LB medium with rifampin to select for rifampin resistance. This treatment with EMS resulted in approximately 80% viable cells and increased the frequency of Rif^r mutations about 1,000-fold (from approximately 10⁻⁷ to 10⁻⁴) (data not shown).

All screening steps were performed at 30°C, and wild-type controls were included for comparison at each step. First, blue colonies on M9Xgal plates were transferred with toothpicks to plates containing LB medium with chloramphenicol so that only cells with miniP1 plasmids were sampled. The resulting colonies were divided into pools (usually 200 colonies per pool), and the pools were transformed by pBEF231 to Kan^r. This step effectively replaced pBEF224 with pBEF231 (due to incompatibility) so that mutations in the miniP1 plasmid pBEF224 were eliminated. The cultures (with pBEF231) were again plated on M9Xgal plates and screened for blue colonies. This time, individual blue colonies were picked and tested several times on M9Xgal plates following plasmid enrichment on plates containing LB medium with kanamycin. Colonies that were reproducibly blue were retained as described above. At least 1,000 colonies were screened for each pool, but mutants isolated from the same pool were considered to be siblings unless later phenotypes (extent of plasmid stability or growth phenotypes, for example) distinguished them. Plasmid-free segregants were isolated and lysogenized by λlacI. Lysogens that were blue on M9Xgal plates (i.e., the blue color was not due to miniP1 plasmid stability) were discarded. Finally, cells were again transformed with pBEF231 to confirm that the phenotype was due to a host mutation.

Plasmid stability tests.A culture of E. coli cells containing either P1 or F plasmid was grown overnight in LB medium at 30°C from a single colony under selective conditions. This culture was diluted 10⁴-fold into 5 ml of fresh LB medium without antibiotics, grown overnight at 30°C (about 15 generations), and then diluted and plated onto LB medium plates. The resulting colonies were transferred with toothpicks onto plates containing LB medium and LB medium with antibiotics to determine plasmid retention. A minimum of 50 colonies were tested for each strain, and each assay was repeated at least twice.

Measurement of DNA and protein synthesis.The rate of DNA synthesis was measured as described by Khidhir et al. (22). Culture aliquots (500 μl) were taken from log-phase cultures in M9-glucose medium and pulse-labeled for 2 min with [³H]thymidine (6 μCi/ml; specific activity, ∼70 Ci/mmol) at various times before and after culture cells were shifted from 30 to 42°C. For protein synthesis, cells were similarly pulse-labeled with [³H]leucine (6 μCi/ml) for 2 min. Ice-cold 15% trichloroacetic acid was added to stop synthesis and precipitate macromolecules. The samples were filtered through glass fiber filters, and the amounts of incorporated ³H-labeled thymidine or leucine were measured by scintillation counting.

Isolation of E. coli mutations that compromise P1 plasmid stability.We sought to identify new E. coli mutations that compromised P1 stability, and we wanted to be able to differentiate between mutations with effects on replication and mutations with effects on partition. Due to observations that components of the replication machinery are specifically localized in bacterial cells (28), we also considered the possibility that certain mutations might affect both processes. We designed a screening procedure for bacterial mutations that destabilize miniP1 plasmids with weakened Par systems. The rationale was that we would be more likely to uncover mutations that reduced plasmid stability (a synthetic lethal approach). The Par system was weakened due to the presence of a specific parS mutation, parS_md2 (19), which reduced ParB affinity for the site so that it was no longer sufficient for partition in the absence of IHF. The mutation had only a minor effect on plasmid stability in wild-type E. coli, but it drastically reduced partition in cells lacking IHF (19). Thus, another benefit of the approach is that we expected to find IHF mutants in this screening analysis as a positive control. In this analysis, the miniP1 plasmid pBEF224 contained the E. coli lacI repressor gene in addition to the parS_md2 mutation, and plasmids were propagated in E. coli that was lacZ⁺ but lacI. On agar plates containing X-Gal (but with no selection for miniP1), cells in which P1 was stable produced white or very light blue colonies, while cells in which P1 was unstable produced darker blue colonies as the plasmid was lost during the growth of the colony. The intensity of the blue color approximately correlated with the number of cells in the colony containing miniP1 (colonies with no miniP1 were intensely blue, colonies with some miniP1 were less blue, etc.), so that we could qualitatively estimate the retention of miniP1 plasmids in each colony.

N99 lacI::tet cells containing pBEF224 were mutagenized with EMS and then screened for mutations that reproducibly produced blue colonies as described above (see Materials and Methods). The screening procedure and all initial tests were performed at 30°C so that we could recover conditional (temperature-sensitive) mutations in essential genes. Cells that produced blue colonies were retransformed by pBEF231, a Kan^r version of pBEF224, to eliminate plasmid mutations. Candidate mutants were then lysogenized by a λ phage carrying lacI to eliminate mutations that produced a blue phenotype independently of plasmid stability (blue λlacI lysogens were eliminated). Using this approach, we screened approximately 65,000 mutagenized cells and identified nine mutants that reduced the stability of miniP1.

One mutant, designated N99 I1-18, was extremely temperature sensitive for growth (see below), while the other eight mutants were not. In this study, we focused on the I1-18 mutation, as it represented a change in a gene that is essential for E. coli. However, we also noted that one of the other mutations, designated I1-20, was most likely in ihfA, based on two major criteria. First, it could not be lysogenized by the λlacI phage (IHF is required for λ integration [23]). Second, the plasmid stability defect was complemented by introduction (by mating) of F′506 or F′500, whose chromosomal inserts overlap for approximately five min of the E. coli map, including the region encoding the ihfA gene (37, 40) (data not shown).

We next characterized the stability of a higher-copy-number derivative of miniP1 in the color assay. This plasmid, pAJM4 (Table 2), which contained the E. coli lacI gene but not P1 par, was reasonably stable because its copy number was about 4 (derived from pALA318 [42]). N99 lacI::tet(pAJM4) cells were white on M9Xgal plates. N99 I1-20(pAJM4) cells were also white, which was consistent with the properties of an IHF mutant (IHF is involved in partition but not replication [17]). However, N99 I1-18(pAJM4) cells were blue, indicating that the I1-18 mutation did not depend on the P1 partition system for its plasmid stability phenotype and implying that I1-18 influenced the replication of miniP1.

Identification of dnaB277.We used an E. coli K-12 plasmid library to identify candidate genes altered by the I1-18 conditional mutation. Following transformation of N99 I1-18 by this library, we recovered two different transformants that allowed N99 I1-18 to grow at 42°C. These isolates (with pBEF231) also formed white colonies on X-Gal plates, indicating that miniP1 stability was restored in the cells (Fig. 2). The library plasmids from each isolate, designated p3d and p4c, were characterized by restriction mapping, and the ends of the chromosomal inserts were sequenced. Although the lengths of chromosomal DNA in p3d and p4c differed, the DNA sequences and restriction maps indicated that the two plasmids possessed an overlapping and contiguous region of E. coli DNA between 91.8 and 91.9 min of the chromosome (Fig. 2). One end of each of the inserts in p3d and p4c was identical and was located within the dnaB replication helicase gene (at position 473 of the 1,413-bp dnaB gene). Both plasmids also carried alr, whose product (alanine racemase) converts l-alanine to d-alanine for subsequent catabolism (30), and all (p3d) or a fraction (p4c) of tyrB, whose product (tyrosine aminotransferase) is involved in the biosynthesis of tyrosine and phenylalanine (15, 27). The p3d insert was about 1 kb longer than that of p4c and included the 5′ end of aphA, which encodes diadenosine tetraphosphatase, a class B acid phosphatase (54). Although the coding sequence of dnaB was not complete, dnaB was the only essential gene represented in the chromosomal inserts, and so we considered it to be the most likely gene mutated in N99 I1-18. To test this hypothesis, we cloned the dnaB gene with its natural promoter from wild-type strain N99 into the pBR322 vector (see Materials and Methods). The resulting plasmid, pDnaB2, was able to restore viability to N99 I1-18 at 42°C, indicating that dnaB⁺ was sufficient to suppress the temperature-sensitive-lethal defect. We similarly cloned and sequenced the dnaB gene from N99 I1-18. Sequencing revealed a G-to-A transition at bp 830 in the coding strand of dnaB, corresponding to a change from serine to leucine at amino acid position 277 (Fig. 1). The allele was designated dnaB277 and encoded DnaB S277L. Finally, pDnaB277 expressing the dnaB277 allele could not restore viability to N99 I1-18 at 42°C, further supporting the hypothesis that the temperature-sensitive-lethal mutation in N99 I1-18 was in dnaB.

• Open in new tab
• Download powerpoint

FIG. 2.

Library plasmids that suppress the I1-18 mutation. As shown at the top, the inserts in plasmids p3d and p4c were identified by sequencing 500 to 700 bp at each end from the vector sequence, followed by BLAST, and were confirmed by restriction mapping. Note that the dnaB gene is incomplete and lacks the first 472 bp of coding sequence. The properties of the N99 I1-18(p3d) and N99 I1-18(p4c) isolates are indicated at the bottom. Single colonies of cells were streaked onto preheated LB medium plates at either 30 or 42°C and scored for viability after about 16 h. Single colonies of cells transformed by the pBEF231 miniP1 lacI^q plasmid (Table 2) were scored for blue color (loss of miniP1) after 8 h of growth on M9Xgal plates at 30°C.

To confirm these results and to characterize dnaB277 further, we moved the I1-18 mutation into a clean, unmutagenized N99 background by linking the Ts allele to miniTn10kan (see Materials and Methods), and we repeated viability and plasmid stability tests. We also compared it to a variety of other dnaB(Ts) alleles that we obtained from the E. coli Genetic Stock Center (Table 1). Most alleles were linked to the same miniTn10kan used for I1-18 and were transferred into N99 lacI::tet (Table 1). One allele, dnaB252, was already linked to Tn10(tet), which was used to transfer the allele into N99 and N99 lacI::cam.

The temperature sensitivities conferred by the various dnaB alleles varied slightly. The dnaB252 and dnaB277 alleles were the most severe alleles and inhibited colony formation at temperatures above 34°C, but all the dnaB mutants were incapable of growth at temperatures above 36°C (data not shown). N99 lacI::tet dnaB277 was transformed by p4c and pST52 (its vector) at 30°C, but only p4c transformants grew when they were streaked out at 42°C, confirming the ability of p4c to suppress the temperature-sensitive-lethal defect in N99 I1-18. The suppression was not complete, however; when measured by plating, the ratio of colonies that grew at 42°C to colonies that grew at 30°C was 0.06. This ratio for pST52 transformants was <10⁻⁶. In contrast, suppression by pDnaB2 was complete (the ratio was 1.0 [see below]). We constructed a recA56 derivative of N99 dnaB277 (Table 1), tested the viability of p4c transformants at 42°C, and found that the viability dropped to <10⁻⁵ in the absence of RecA function. We concluded that the p4c plasmid suppressed the dnaB277 allele by marker rescue.

Interestingly, we observed that although the viability of wild-type cells was not significantly affected by p4c transformation (the 42°C/30°C viability ratio was 0.9), the plasmid did have a low cell growth rate so that N99(p4c) colonies were smaller than N99(pST52) colonies. This observation implied that p4c produced a dnaB product that was slightly deleterious to growth in the presence of wild-type DnaB. We created a similarly truncated dnaB gene in the pDnaB2 context that was deleted in frame between the dnaB start codon and bp 473 within dnaB but was expressed from the dnaB promoter signals, producing pBEF261 (see Materials and Methods). We found that this plasmid could transform wild-type E. coli at 37°C but not at 30°C (reduced by at least 1,000-fold). The 37°C transformants were cold sensitive and unable to form colonies at 30°C. We could not recover pBEF261 transformants in any of the dnaB mutant strain set at either temperature. These observations indicate that this fragment of DnaB is deleterious to cell growth in a wild-type background.

We measured the ability of the pDnaB2 plasmids to suppress dnaB277 in the clean genetic background and confirmed that pDnaB2, but not pDnaB277, could completely suppress the temperature-sensitive viability defects (Table 3). We also tested the suppression of the other dnaB alleles by pDnaB2. This plasmid was able to efficiently suppress some alleles (dnaB6, dnaB8) but not all alleles (Table 3). We observed an incomplete ability of wild-type dnaB⁺ in pDnaB2 to suppress the viability defects of the dnaB22, dnaB70, dnaB107, and dnaB252 alleles. These observations suggest that in some contexts the mutant allele interferes with wild-type DnaB function, at least when it is expressed from a multicopy plasmid.

dnaB277 is a fast-stop replication allele.Conditional replication mutations are often divided into fast-stop mutations, which immediately inhibit the propagation of the replication fork, and slow-stop alleles, which prevent replication initiation. Since helicase function is required for replication elongation, most dnaB alleles, such as dnaB8, are fast-stop alleles. However, one allele originally misclassified as a dnaA allele is the exception (63). The dnaB252 mutation is a slow-stop mutation; it does not influence helicase activity (45), and DNA propagation can be suppressed by overexpression of the DnaC replication initiation protein, suggesting that the dnaB252 mutation may potentially interfere with initiation protein interactions (47). We measured the rates of DNA synthesis before and after a shift from a permissive temperature (30°C) to a nonpermissive temperature (42°C) (Fig. 3A). We compared dnaB277 to both dnaB252 (slow stop) and dnaB8 (fast stop). First, we observed that the overall rate of DNA synthesis by the dnaB277 mutant at 30°C was less than one-half that by the wild type or by mutants with other dnaB alleles for cultures at similar absorbance values (Fig. 3). However, the doubling time of dnaB277 cells (162 min) was much longer than the doubling times of dnaB8 cells (85 min) and dnaB252 cells (106 min), implying that the dnaB277 mutation was slowing the replication forks. Next, for each mutant we compared the rates of DNA synthesis before and after a shift to 42°C. Following the temperature shift, the dnaB8 mutant immediately ceased DNA synthesis, as expected, while the dnaB252 derivative continued DNA synthesis for 30 min following the shift to 42°C before a decline in replication began. The dnaB277 mutant exhibited a replication fast-stop profile, although the cells did not cease replication quite as rapidly as the dnaB8 mutant cells and exhibited >80% thymidine incorporation after 5 min at 42°C, compared to <3% for dnaB8.

• Open in new tab
• Download powerpoint

FIG. 3.

DNA synthesis profile of dnaB(Ts) fast-stop, slow-stop, and dnaB277 alleles. (A) DNA synthesis measured by [³H]thymidine incorporation into N99 dnaB⁺ (•), dnaB8 (▪), dnaB252 (♦), and dnaB277 (▴) before and after a shift from 30 to 42°C in rich media. Time zero was the time of temperature shift from 30 to 42°C. At −30 min, the average counts were as follows: for dnaB⁺ (A₆₀₀, ∼0.21), 43,492 cpm; for dnaB8 (A₆₀₀, ∼0.2), 45,239 cpm; for dnaB252 (A₆₀₀, ∼0.25), 59,459 cpm; and for dnaB277 (A₆₀₀, ∼0.29), 23,060 cpm. The plotted values are averages based on duplicate samples. (B) DNA synthesis and protein synthesis were measured in tandem by measuring incorporation of [³H]thymidine and [³H]leucine into untreated N99 dnaB277 cells or chloramphenicol (Cam)-treated (125 μg/ml) cells shifted from 30 to 42°C and held at this temperature for 30 min before they were shifted back to 30°C. [³H]leucine incorporation into untreated (▿) and chloramphenicol-treated (▾) cells and [³H]thymidine incorporation into untreated (○) and chloramphenicol-treated (•) cells were determined. Time zero was the point at which cells were shifted from 30 to 42°C. The time of chloramphenicol addition is indicated by the arrow.

Next, we asked whether the function of heat-labile DnaB S277L could be restored if a culture was shifted back to a permissive temperature. We monitored replication and protein synthesis in N99 dnaB277 cells that were returned to 30°C following 30 min of incubation at 42°C (Fig. 3B). Prior to the shift back to 30°C, the 42°C cultures were divided, and one of the cultures was treated with chloramphenicol to prevent de novo protein synthesis. We monitored [³H]thymidine incorporation in parallel with [³H]leucine incorporation as measures of DNA and protein synthesis, respectively. Chloramphenicol reduced protein synthesis to less than 3% of the synthesis in untreated cells within 5 min. Following the shift back to the permissive temperature, the cells treated with chloramphenicol exhibited a small but reproducible increase in DNA synthesis. Therefore, the activity of temperature-labile DnaB S277L can be partially restored by shifting the protein back to a permissive temperature, although the activity resumes slowly and recovery is incomplete.

Influence of dnaB alleles on P1 and F plasmid stability.We assayed the stability of unit-copy miniP1 (pBEF224) and miniF (pMF3) plasmids in the dnaB mutant set by measuring plasmid retention after growth for approximately 15 generations in the absence of selection (Table 4). The dnaB277 allele compromised P1 stability but not F stability at 30°C, while none of the other dnaB alleles assayed reduced the stability of either plasmid. The effect of the dnaB277 allele was even more dramatic at 34°C (8% P1 retention, compared to >85% P1 retention for all other dnaB derivatives [data not shown]). These results were a quantitative measure of the instability of miniP1 that we observed in the initial genetic screening with N99 I1-18. We also measured the stability of the higher-copy-number but Δpar miniP1 derivative pALA318Kan (which had the same origin and copy number as pAJM4 [Table 2]). This miniP1 derivative was also destabilized by dnaB277 but not by the other dnaB(Ts) alleles.

Since a plasmid expressing dnaB⁺ (pDnaB2) could fully complement the dnaB277 allele and restore viability at 42°C, we asked whether it could restore P1 stability in the cells in the type of plasmid retention assay described above (Table 5). We observed that pDnaB2 could restore P1 stability to nearly wild-type levels in N99 dnaB277. This finding confirmed that the dnaB277 allele is responsible for both the TS-lethal and P1 instability phenotypes seen in this strain. Interestingly, pDnaB2 and (to a lesser extent) pDnaB277 destabilized miniP1 in wild-type cells at 30°C. These results support a role for DnaB in P1 plasmid stability and further suggest that overexpression of dnaB⁺ somehow compromises plasmid maintenance.

Finally, because the miniP1 plasmid pBEF224 contains a weakened parS site (parS_md2), we assayed the effect of dnaB277 on the parS⁺ miniP1 plasmid pLG49 (which had the same replicon and copy number as pBEF224 [Table 2]). We found that 37% ± 7% of N99 dnaB277 cells retained pLG49 after 15 generations of growth in nonselective rich medium at 30°C, compared to 98% ± 1% for wild-type N99 cells. These results indicate that destabilization of miniP1 by the dnaB277 mutation is similar for pLG49 and pBEF224 and thus is not dependent on the specific parS site present in the plasmid. In addition, the difference between P1 stability and F stability is not due to the weakened par system of pBEF224 and pBEF231.

Next, we measured and compared the copy numbers of a miniP1 plasmid (pLG49) and a miniF plasmid (pBEF215) in dnaB277 mutants and wild-type cells by Southern hybridization (Fig. 4). These versions of miniP1 and miniF plasmids each contain the E. coli lacI gene (Table 2), and lacI sequences were used as a probe to measure the ratio of plasmid DNA to chromosomal DNA in wild-type and dnaB277 cells. Within the error of the experiment, the copy numbers of miniP1 and miniF relative to chromosomal DNA were slightly reduced by the dnaB277 mutation. This result indicates that this dnaB allele influenced replication of both P1 and F plasmids. In addition, overall copy number differences do not simply explain the differential effects of dnaB277 on miniP1 and miniF (see Discussion).

• Open in new tab
• Download powerpoint

FIG. 4.

Determination of plasmid (miniP1) copy number in N99 dnaB(Ts) mutants. DNA was isolated, cut with different restriction enzymes, and analyzed by Southern hybridization as described by Funnell and Gagnier (20). (A) Lanes 1 and 2, N99 dnaB⁺ (BF890); lanes 3 and 4, dnaB⁺(pLG49); lanes 5 and 6, dnaB8 (pLG49) (BF891); lanes 7 and 8, dnaB277 (pLG49) (BF892). (B) Lanes 1 and 2, N99 dnaB⁺; lanes 3 and 4, dnaB⁺ (pBEF215); lanes 5 and 6, dnaB8 (pBEF215); lanes 7 and 8, dnaB277 (pBEF215). Total genomic DNA was digested with EcoRI and EcoRV. The amount of DNA in lanes 2, 4, 6, and 8 was twice the amount in lanes 1, 3, 5, and 7. The lacI probe was a 942-bp EcoRI-HincII fragment from pMC9 (8). The arrows indicate the positions of the chromosomal 1.5-kb (lac C) and plasmid 0.85-kb(lac P) bands hybridizing to the lacI probe. The plasmid-to-chromosome (P/C) ratio was calculated for each lane by dividing the intensity of the lac P band by the intensity of the lac C band. The copy number results for pLG49 were averaged and corrected for the percentage of cells that lost pLG49 (dnaB⁺, <1%; dnaB8, <1%; dnaB277, 12%).

Plasmid localization.Since components of the replication machinery are known to be specifically localized in bacterial cells, we considered the possibility that dnaB277 could affect P1 plasmid localization. In addition, genetic evidence has suggested a role for DnaB in the partition of plasmid pSC101 (32). We examined P1 positioning by immunofluorescence of ParB, which forms large foci that colocalize with parS and thus with P1 (14). In wild-type cells these foci colocalize with bacterial nucleoids (14) (Fig. 5 and Table 6). In contrast, in the absence of a complete partition system (no ParA), the ParB foci are usually located between and outside the space occupied by the nucleoids. We reasoned that if DnaB were influencing partition as well as replication, ParB foci may also be mislocalized. We observed a 10- to 15-fold-greater number of filamentous N99 dnaB277 cells than of wild-type cells at 30°C, although the majority of mutant cells exhibited a healthy morphology at this temperature (Table 6). However, shifting dnaB277 cells to 42°C resulted in extensive filamentation by 60 min (data not shown). We noted that a greater number of dnaB277 cells had lost the P1 plasmid at 30°C and that the number of foci per plasmid-containing cell was lower than the number seen in wild-type cells (74% of the wild-type value). However, we observed no irregular positioning of foci in the N99 dnaB277 strain at 30°C (Fig. 5D and E), indicating that P1 localization, as measured by its ability to colocalize with the bacterial nucleoid, is not significantly affected by the dnaB277 allele.

• Open in new tab
• Download powerpoint

FIG. 5.

Visualization of P1 ParB by immunofluorescence. E. coli cells were collected and fixed by treatment with paraformaldehyde plus glutaraldehyde (7, 14). The fixed cells were stained with affinity-purified rabbit anti-ParB and then with Cy3-conjugated goat anti-rabbit immunoglobulin G and 0.01% 4′,6-diamidino-2-phenylindole (DAPI) prior to mounting. E. coli N99 dnaB277 (strain BEF892) (Table 1) cells containing the Par⁺ pLG44 plasmid (ParA⁺ ParB⁺parS⁺) were examined and compared to N99 wild-type cells containing pLG44 or Par⁻ pBEF246 (parS⁺) in the presence of only ParB (pBEF251). (A, D, and G) Log-phase N99(pLG44) cells. (B, E, and H) Log-phase N99 dnaB277(pLG44) cells. (C, F, and I) Log-phase N99(pBEF246/pBEF251) cells. Panels A to C are Nomarski images of cells. Panels D to F show ParB visualized by Cy3-labeled secondary antibodies. Panels G to I show DAPI staining of cells to visualize nucleoids. In the absence of ParA, the pBEF246 plasmid was unstable and resulted in a large number of cells that had lost the plasmid (I).

Effects of dnaB277 on E. coli.The dnaB277 mutation reduces the rate of total DNA replication and exhibits a fast-stop DNA synthesis phenotype (Fig. 3), which indicates a defect during elongation of DNA replication. The DnaB S277L mutation identified in this study is 15 residues C terminal to the conserved aspartate of the Walker B ATP binding motif (Fig. 1). The mutation is within fragment II (defined by partial proteolysis) and the β domain (defined by deletion analysis) of the DnaB polypeptide. The proximity of the ATP binding site and the observation that dnaB277 is a fast-stop replication mutation (Fig. 3A) may suggest that the change affects helicase activity, perhaps by affecting ATP binding or hydrolysis. Alternatively, other interactions at the replication fork may be compromised so that the fork disassembles at nonpermissive temperatures. DnaB interacts with a variety of chromosomal replication proteins (for example, DnaA and DnaC) during initiation of replication and the τ subunit of DNA polymerase III holoenzyme to coordinate DNA replication on the leading and lagging strands (24, 53).

Effects of dnaB277 on plasmid maintenance.Based on our observations that the dnaB277 allele reduced the plasmid copy number (Fig. 4) and destabilized P1 plasmids that do not depend on par for their stability (pALA318kan [Table 4]), we inferred that the destabilization of P1 is primarily due to a reduction in plasmid replication. The phenotypes provide direct genetic evidence that DnaB is required for plasmid replication, which is consistent with its requirement for replication in vitro (60). However, although the dnaB277 allele confers similar reductions in P1 and F copy numbers, it is interesting that F remains relatively stable in the cells and P1 does not. One possibility is that the sop system of F may be more effective than P1 par in plasmid distribution and thus can suppress the slight reduction in plasmid copy number. Alternatively, the dnaB277 allele may differentially influence the copy number distribution of F and P1 in the population, without affecting the average copy number. A wider distribution of P1 could result in a greater number of cells that lose the plasmid. Finally, we cannot formally rule out the possibility that dnaB277 also interferes with P1 partition either directly or indirectly (for example, by delaying the separation of plasmids due to delayed replication).

Hyperrecombination mutants with plasmid stability defects have also been identified (2), since an increase in the number of plasmid dimers and multimers reduces the number of partitionable units. Recently, DnaB has also been shown to drive branch migration at Holliday junctions and thus may be involved in the resolution of recombination intermediates (21). However, we think that it is unlikely that dnaB277 exerts its effect via recombination for two reasons. We saw no difference in the pattern of plasmid monomers versus multimers when plasmids were isolated from dnaB⁺ and dnaB277 cells, and miniP1 plasmids were unstable in the Rec⁻dnaB277 derivative (unpublished results). In addition, it is difficult to explain why miniP1 would be more affected than miniF in this regard.

The β domain also contains the regions that interact with pSC101 RepA (amino acids 208 to 261) (13) and R6K π (amino acids 151 to 189) (44). In light of these observations, it is possible that the specific defect in P1 plasmid maintenance involves a weakened interaction between DnaB S277L and the P1 replication initiator, RepA. Interestingly, we also observed that high-copy expression of dnaB⁺ from pDnaB2 in a dnaB⁺ host destabilized miniP1 plasmids (Table 5). We speculated that this destabilization could result from an excess of DnaB titrating RepA away from other initiation factors.

Interactions of mutant and wild-type dnaB alleles.We were able to suppress the dnaB277 allele with a multicopy dnaB⁺ plasmid that restored viability and P1 stability to wild-type levels. These results imply that the dnaB277 allele is recessive, although we have not confirmed this using single-copy dnaB⁺ in a complementation test. Alleles dnaB252, dnaB22, dnaB70, and dnaB107 were only partially complemented by the dnaB⁺ plasmid at 42°C. These alleles may compete for protein interactions, with wild-type DnaB diluting out available interactions, or may have a negative dominant influence, generating heteromultimers with wild-type DnaB at their nonpermissive temperatures that reduce DnaB function. Alternatively, the promoter present on the pDnaB2 plasmid may influence dnaB gene regulation by diluting out transcriptional stimulators or repressors.

We also found that wild-type cells carrying multicopy plasmids that encoded only the βγ domains of DnaB (p4c and pBEF261) interfered with cell growth. pBEF261 conferred a cold-sensitive phenotype to cells. We concluded that the βγ fragment interferes with wild-type DnaB function, possibly by competing with DnaB for a protein interaction. This effect may be particularly detrimental at lower temperatures, at which the rate of replication is lower.

We designed a genetic screening procedure to isolate host mutations that compromise the stability of plasmid P1 with the goal of improving our understanding of the mechanisms of P1 stability. We used a destabilized miniP1 in order to design a synthetic lethal screening procedure, with which we could isolate mutations that might not have been identified by using a wild-type P1 system. We think that the isolation of an ihfA mutation validates this approach, and we suggest that the method could be extended to identify other factors that play accessory roles in plasmid maintenance in this and other plasmid systems. It should of course yield mutations in genes that are essential, such as the identification here of the dnaB277 allele. We continue to investigate the involvement of DnaB in P1 plasmid stability, and we are extending this screening analysis to search for other factors that play roles in plasmid maintenance.