INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Escherichia coli DNA polymerase III (DNA Pol III) is the enzyme responsible for replication of the bacterial chromosome. The DNA Pol III holoenzyme consists of 10 subunits, with polymerase activity residing in the  subunit, the product of the dnaE (polC) gene (reviewed in references 16 and 20). Although much is known about the polymerase active centers of several DNA polymerases and an RNA polymerase from structural and genetic studies (14), relatively little is known about the polymerase active centers of bacterial DNA Pol III holoenzymes. One approach has been to identify conserved aspartate residues, since carboxylate residues are an essential feature of the polymerase active centers of DNA and RNA polymerases (3, 4, 14, 15). Protein sequence alignment of the -subunit sequences from Proteobacteria, Spirochaetales, Cyanobacteria, Aquificales, and Fermicutes revealed conserved aspartate residues, and three of these residues, D401, D403, and D555, in the E. coli  subunit were determined to be essential for polymerase activity by site-directed mutagenesis (22).

Genetic selection and screening procedures have also been used to identify amino acid residues in the  subunits of bacterial DNA Pol III holoenzymes that are important for function. One advantage of genetic strategies is that structural information is not required, and a second advantage is that the mutant DNA polymerases are studied in the cell in the presence of the full complement of DNA replication proteins. Fijalkowska et al. (7) used a genetic screen to identify mutations in the E. coli dnaE (polC) gene that confer an antimutator phenotype, which is increased accuracy of DNA replication. A mutation that decreases replication fidelity and produces a strong mutator phenotype has also been identified (18). Mutations in the Bacillus subtilis polC gene that confer resistance to the dGTP analog, 6-(p-hydroxyphenylhydrazino)uracil, have also been identified (1, 9). We have extended these studies by identifying amino acid residues in the E. coli  subunit that affect the sensitivity of the DNA Pol III holoenzyme to chain-terminating nucleotide analogs. These studies may reveal amino acids that affect the ability of the DNA Pol III holoenzyme to discriminate in the incorporation of nucleotide analogs or in their removal by exonucleolytic proofreading.

The genetic strategy is based on two previous findings about bacterial DNA replication in the presence of chain-terminating nucleotide analogs. First, cytosine and adenine arabinosides (AraC and AraA) (23) and 2',3'-dideoxyadenosine (ddA) (25) are converted in bacteria to the deoxynucleoside triphosphates and then incorporated into DNA by DNA polymerases. Incorporation of chain-terminating nucleotides is not expected to block further replication, however, since the DNA Pol III holoenzyme has 3'-to-5' exonucleolytic proofreading activity, which is reduced only sixfold by the chain terminator ddA compared to deoxyadenosine at the 3' primer terminus (10). Yet ddA hinders DNA replication in vivo and is lethal to the cell (25), which indicates that at least some of the incorporated chain-terminating nucleotides are not removed. The second important observation is that B. subtilis and E. coli polA mutant strains which are deficient in DNA Pol I activity are more sensitive than wild-type strains to the killing activity by AraC and AraA (23). Furthermore, the increased killing activity by chain-terminating nucleotides in polA mutant strains can be suppressed by certain mutations in the dnaE (polC) gene (23). Together, these results indicate a synergism between the DNA Pol III holoenzyme and DNA Pol I for the repair of modified 3' primer termini. While the DNA Pol III holoenzyme is responsible for incorporation of chain-terminating nucleotides during chromosome replication, DNA Pol I appears to assist in their removal.

The increased sensitivity of polA mutant strains to killing by chain-terminating nucleotides was exploited to identify mutations in the dnaE (polC) gene that alter sensitivity to ddA. The starting point was an E. coli polA1(Am) strain (5), which has about 1% of the DNA Pol I activity of wild-type cells (17). The dnaE (polC) gene was mutagenized selectively by the procedure of Hong and Ames (12). Strains with increased sensitivity or resistance to ddA were isolated and characterized. These and additional mutants identified by the genetic selection method will be useful for in vivo and in vitro biochemical studies of nucleotide interactions with the  subunit of the DNA Pol III holoenzyme.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and media. E. coli strains used in this study are described in Table 1. Bacteria were grown in Luria broth supplemented with thiamine (50 µg/ml) and thymidine (10 µg/ml) (TLB). Solid media contained 1.5% Bacto agar. Antibiotics, when required, were added as follows: chloramphenicol (CAM), 10 µg/ml; tetracycline (TET), 15 µg/ml; rifampin, 100 µg/ml; kanamycin, 70 µg/ml; ampicillin, 50 µg/ml.

P1 transduction and localized mutagenesis. P1 transduction was used to construct bacterial strains and to map mutations in the bacterial chromosome and for localized mutagenesis. Ten-milliliter bacterial cultures were grown overnight in TLB; the cells were pelleted and resuspended in 2.5 ml of 10 mM MgSO₄. CaCl₂ was then added to give a final concentration of 5 mM. The recipient cells (0.1 ml) were mixed with 0.05 ml of a fresh P1 vir lysate and incubated at 37°C for 30 min without shaking. Sodium citrate (0.1 ml of a 1 M solution) was added to prevent phage readsorption. TLB (0.5 ml) was then added, and incubation continued for another 60 min. The infected cells were pelleted and resuspended in 0.1 ml of TLB supplemented with 20 mM sodium citrate. The cells were then plated under conditions to select for P1 transductants.

F plasmid conjugation. The polA⁺ gene was introduced into D110 cells by conjugation with the donor strain CJ300, which carries the polA⁺ gene and a gene conferring Cam^r on the F plasmid (13). The recipient D110 cells were made Tet^r by introduction of the transposon zae-502::Tn10(Tet) linked to the wild-type or to a mutant dnaE gene. Fresh, early-log-phase cultures of recipient and donor cells were prepared. The cells were pelleted and suspended in TLB to give 2 × 10⁸ bacteria/ml. The donor and recipient cells were diluted 10-fold in TLB and incubated at 37°C for 60 min without shaking. The titers of the cells on plates containing TET and CAM to select for antibiotic-resistant conjugants were then determined. The conjugants also became more resistant to UV and other types of DNA damage due to the introduction of the polA⁺ gene (11).

Sensitivity to ddA. (i) Plating efficiency method. Titers of fresh cultures were measured to determine the total number of cells, and cultures were also plated on ddA-containing plates to determine the number of drug-resistant cells. ddA (Sigma) was dissolved in dimethyl sulfoxide to give a 100 mM stock solution, which was then diluted further into TLB agar to produce plates with a final concentration of 5 or 20 µM ddA.

(ii) Liquid culture method. Cell killing by ddA in liquid culture was determined by first growing each strain at 37°C in TLB without ddA to early log phase, about 10⁷ cells/ml. The cultures were then divided, and ddA was added to one set at a final concentration of 100 µM. Control cultures contained no ddA. The cultures were incubated at 37°C with shaking. The titers of ddA and control cultures were measured at 1-h intervals for 4 h to determine cell viability.

Mutation frequency determination. DNA replication fidelity for the wild-type and mutant strains was determined by measuring the frequency of forward mutations that confer resistance to the antibiotic rifampin. Ten or more colonies from each strain were grown overnight in TLB with aeration by shaking. The cultures were diluted and plated on TLB plates to determine the total number of cells and on freshly made rifampin-containing TLB plates to determine the number of Rif^r cells. The experiments were repeated at least three times; similar results were obtained each time.

DNA sequencing of the dnaE (polC) gene. PCR and sequencing primers were designed from the published sequence of the dnaE (polC) gene (26). The entire dnaE gene (3,480 bp) was amplified within a 3,891-bp PCR fragment using the forward PCR primer 5'-CTGGGCTGCATATTGCG at position 448 to 464, according to the numbering system of Tomasiewicz and McHenry (26), and the reverse primer 5'-CTTCCAGCTCTGCAATC at position 4339 to 4323.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Reduced DNA Pol I function increases sensitivity to the killing activity of ddA. The bacterial strain D110 (21) carries the polA1(Am) allele (5). This strain is predicted to have increased sensitivity to the killing effects of chain-terminating nucleotides because of reduced DNA Pol I activity (23). Sensitivity of the polA1 mutant strain to ddA was measured by growing the D110 cells in the presence of 100 µM ddA (Fig. 1). Cell survival was also determined for a polA⁺ derivative of D110 in which an F' plasmid bearing the polA⁺ gene was introduced. The polA1(Am) mutant strain was about 50-fold more sensitive to killing by ddA after 1 h of exposure and 800-fold more sensitive after 3 h than the strain with restored DNA Pol I activity (Fig. 1).

View larger version (11K): [in this window] [in a new window]   FIG. 1.   Sensitivity of wild-type and DNA Pol I-deficient strains to killing by ddA. Cultures, initially at about 107 cells/ml, were grown with aeration by shaking at 37°C in TLB medium supplemented with 100 µM ddA. Titers of the cultures were determined for cell viability at 1-h intervals. DNA Pol I-deficient D110 cells () were more sensitive to killing by ddA than D110-polA+ cells with restored DNA Pol I activity ().

Isolation of mutants with increased or decreased sensitivity to the killing activity of ddA. Localized mutagenesis of the dnaE (polC) gene was done by the method of Hong and Ames (12). Phage P1 was grown on strain NR9918 (Table 1). This strain has a transposon carrying a gene encoding CAM resistance 5' to the dnaE (polC) gene and a second transposon carrying a gene encoding TET resistance on the 3' side (7). The P1 lysate was treated with hydroxylamine and used to transduce the polA1(Am) strain, D110, as described in Materials and Methods. About 3,000 CAM- or TET-resistant transductants were screened for ddA resistance by spotting diluted cultures on TLB plates with 50 µM ddA or were screened for increased sensitivity to ddA by spotting cultures on TLB plates containing 5 µM ddA. The ddA phenotypes for the ddA-sensitive S64 and S566 strains and for the resistant R3 strain were 60 to 80% linked by P1 transduction to the transposons zae-502::Tn10(Tet) and zae-502::Tn10d(Cam). These transposons are located at equal distances, approximately 0.5 min to either side of the dnaE (polC) gene (7). High linkage to both transposon markers indicates that mutations conferring resistance or sensitivity to ddA reside within or near the dnaE (polC) gene. The P1 transductions were repeated into the polA⁺ C600 strain, and the ddA-sensitive and -resistant phenotypes were still observed. Mapping to the dnaE gene was further confirmed by complementation with a plasmid carrying the dnaE⁺ gene, plasmid pMWE303, supplied by C. McHenry. Introduction of plasmid pMWE303(dnaE⁺) into the C600 strains carrying the S64, S566, or R3 alleles restored the wild-type level of ddA sensitivity. For another ddA-sensitive strain, the S16 strain, cotransduction of antibiotic resistance and ddA sensitivity was observed only at a very low level, about 1%. One mutation in the dnaE (polC) gene in the S16 strain was identified by sequencing (see below), but this mutation alone did not confer ddA sensitivity.

DNA sequence analysis. The dnaE (polC) genes of the mutant strains were sequenced from PCR fragments. A single nucleotide change in the dnaE (polC) gene of each strain was detected. GC-to-AT transition mutations were expected from the hydroxylamine mutagenesis, and these were observed (Table 2). Mutations in the S64 and S566 strains encoded the L329F and H417Y amino acid substitutions, respectively. These substitutions are fairly conservative, as expected if these residues affect polymerase function, but still permit DNA replication at a level sufficient to sustain viability. A less conservative amino acid substitution, G365S, was detected in the R3 strain. The conservative V977I substitution was detected in the S16 strain, but the failure to observe cotransduction of ddA sensitivity with either of the transposon markers that flank the dnaE (polC) gene indicates that a second mutation, either in combination with the dnaE (polC) mutation or alone, is required to produce ddA sensitivity.

Sensitivity and resistance to ddA. Altered sensitivity to ddA produced by the dnaE (polC) S64, S566, and R3 alleles was examined further by plating cells on TLB plates supplemented with 5 or 20 µM ddA (Table 3). The plating efficiency of the parental polA1(Am) strain (D110), which has only about 1% DNA Pol I activity, was not reduced by 5 µM ddA, but the colonies were small compared to colonies produced in the absence of ddA. Twenty micromolar ddA prevented colony formation of the DNA Pol I-deficient strain. Increased sensitivity to ddA was observed for the polA1(Am) dnaE (polC)-S64 and polA1(Am) dnaE (polC)-S566 strains since colony formation was prevented at 5 µM ddA. Increased resistance was detected for the polA1(Am) dnaE (polC)-R3 strain; no reduction in growth rate was detected at 5 µM ddA, and colony formation was observed on plates containing 20 µM ddA, which restricted growth of the parental polA1(Am) mutant strain.

Fidelity of DNA replication. Altered ability to discriminate in the incorporation of chain-terminating nucleotides or in their removal may also affect incorporation and proofreading of nonmodified nucleotides, which can be detected as a change in the accuracy of DNA replication. Thus, mutant strains with increased sensitivity to ddA may be expected to display a mutator phenotype, while an antimutator phenotype is expected for the ddA-resistant strain. Spontaneous mutation frequencies were determined for the ddA strains by measuring the number of cells in a late-log-phase culture that were resistant to rifampin (Rif^r mutants). Error-prone mutator DNA polymerases produced more Rif^r mutants, while antimutator DNA polymerases produced fewer Rif^r mutants, than the wild-type DNA polymerase. DNA replication accuracy was first measured in the DNA Pol I deficient polA1(Am) background (Table 4). Only small changes in replication fidelity were produced by the S64 and R3 dnaE (polC) alleles: a weak mutator phenotype was observed for the S64 strain, and an antimutator phenotype was observed for the R3 strain. No significant difference in DNA replication fidelity was produced by the S566 allele.

Replication accuracy in the presence of 2AP. 2-Aminopurine (2AP) is a base analogue that is converted to deoxynucleoside triphosphate in the bacterial cell and incorporated into DNA by DNA polymerases. Incorporation of the 2AP deoxynucleotide is mutagenic because incorporation may be opposite template cytosine as well as template thymine. Increased 2AP-induced mutagenesis was detected for the S566 allele, and decreased mutagenesis was detected for the R3 allele (Table 6). Reduced 2AP mutagenesis was also detected in the polA⁺ strain compared to that in the DNA Pol I-deficient polA1(Am) strain (Table 6).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Mutations in the dnaE (polC) gene that confer increased sensitivity or resistance to the killing activity of ddA were identified. Isolation of the mutant strains was assisted by the polA1(Am) mutation, which reduces DNA Pol I activity to just 1% of the wild-type level (17) and, as a consequence, increases sensitivity to the killing activity by ddA (Fig. 1). In previous studies of the DNA Pol III from B. subtilis (23), certain mutations in the dnaE (polC) gene were observed to suppress sensitivity to chain-terminating nucleotides. This finding was confirmed for the E. coli DNA Pol III by selection of the R3 allele, which increased survival in the presence of 20 µM ddA (Table 3). Mutations in the E. coli dnaE (polC) gene that increased sensitivity to killing by ddA, represented by alleles S64 and S566, were also identified (Table 3).

Amino acid substitutions that alter sensitivities to nucleotide analogs may also affect the fidelity of DNA replication if these changes affect discrimination in nucleotide incorporation or exonucleolytic proofreading. A weak but consistent mutator phenotype was produced by the S64 allele (Tables 4 and 5). A weak mutator phenotype was also produced by the S566 allele; this phenotype was most pronounced in the presence of the wild-type level of DNA Pol I (Table 4), in the absence of mismatch repair (Table 5), or in the presence of 2AP (Table 6). An antimutator phenotype was detected for the R3 allele under all of the assay conditions (Tables 4 to 6).

The dnaE (polC) genes of the mutant strains were sequenced in order to identify the mutations responsible for the altered ddA sensitivity and DNA replication fidelity. The L329F and H417Y substitutions, which produce increased sensitivity to ddA, and the G365S substitution, which confers resistance, are all located in conserved sequences in the vicinity of the candidate polymerase active-center residues D401 and D403 (Fig. 2). Two amino acid substitutions that confer the antimutator phenotype, P357L and E395K (6), are also located in this region (Fig. 2).

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Alignment of dnaE  subunit protein sequences. Protein sequence alignments are adapted from Pritchard and McHenry (22). Boldface, conserved residues; asterisks; proposed active-center aspartate residues (22). Amino acid substitutions in E. coli dnaE (polC) mutant strains are indicated on the top line. The L329F, G365S, and H417Y substitutions were identified in this study. The L329F and H417Y substitutions increase sensitivity to ddA and 2AP and decrease DNA replication fidelity (mutator phenotype). The G365S substitution produces resistance to nucleotide analogs and an antimutator phenotype. The P357L and E395K substitutions, identified by Fijalowska and Schaaper (6), were identified on the basis of an antimutator phenotype. B. burgdorferi, Borrelia burgdorferi; A. aeolicus; Aquifex aeolicus; M. tuberculosis, Mycobacterium tuberculosis; M. pneumoniae, Mycoplasma pneumoniae.

Our studies confirm previous observations about the synergism between the DNA Pol III holoenzyme and DNA Pol I during chromosome replication. The ability of DNA Pol I to reduce the block to DNA replication by chain-terminating nucleotides (23) (Fig. 1) and to reduce both spontaneous (2) and 2AP-induced DNA replication errors (Tables 4 and 6) indicates that DNA Pol I has access to nonextendable and mismatched primer termini in certain instances. One possibility that would allow access of DNA Pol I to the primer terminus is if incorporation of a chain-terminating nucleotide or misinsertion of an incorrect nucleotide sometimes results in dissociation of the DNA Pol III holoenzyme. Uncoupling of leading and lagging strand synthesis may ensue. Proofreading by DNA Pol I may then be necessary to reload the DNA Pol III holoenzyme if assembly is hindered by a chain-terminated or mismatched primer terminus. DNA Pol I also has access to the primer terminus after the lagging strand complex dissociates when replication nears a primer for an Okazaki fragment (reviewed in reference 19). If the unbound primer terminus happens to be mismatched or is terminated by a nucleotide analog, proofreading by DNA Pol I could perform the repairs. In the absence of DNA Pol I, the unrepaired primer terminus may be bound by a proofreading-deficient polymerase, such as the umuD'₂C polymerase (24). If the primer terminus is mismatched, extension will introduce a replication error (8). If there is a chain-terminating nucleotide at the primer terminus, association of the umuD'₂C polymerase may produce a stalled replication complex that is incapable of either extension or repair, which would prevent complete DNA replication and ultimately lead to cell death.

In conclusion, the identification of amino acid substitutions in the vicinity of proposed active-center residues that alter sensitivity to nucleotide analogs and alter the fidelity of DNA replication will be useful in characterizing the polymerase active center of the E. coli DNA Pol III holoenzyme. The mutant strains reported here also provide a starting point for the isolation of additional mutations, most importantly second-site mutations that suppress the increased sensitivity to ddA produced by the L329F and H417Y substitutions. The locations of amino acid substitutions encoded by these mutations can be used to map nucleotide-binding interactions in the polymerase active center of the  subunit.