INTRODUCTION

Top Abstract Introduction Materials & Methods Results & Discussion References

Replication of the Escherichia coli chromosome is performed by the DNA polymerase III holoenzyme (Pol III HE) (15, 16, 21). Pol III HE is an asymmetric, dimeric complex containing a total of 18 subunits (10 distinct), which are capable of coordinately synthesizing leading and lagging strands. The complex contains two polymerase cores, one for each strand, composed of , , and  subunits (22).

With regard to the fidelity of the replication process, the  and  subunits of the Pol III core are of particular importance. The  subunit (dnaE gene product) is the DNA polymerase, which selects the correct nucleotides during template-directed DNA synthesis (18). The  subunit (dnaQ gene product) performs the 3'5' exonucleolytic proofreading activity, which preferentially removes incorrect bases inserted by the polymerase (27). The function of the third subunit, the  subunit, has yet to be identified. The  subunit of the Pol III HE plays a complex role in DNA replication. Besides its proofreading activity, it stabilizes the core by tightly binding to both the  and  subunits (22, 29). Interestingly, the  and  subunits are each less active individually than when bound together in the Pol III core (19). One may hypothesize that the fidelity of DNA replication depends not only on the intrinsic accuracy of the polymerase and the strength of the 3'5' exonuclease activity but also on the appropriate interactions between the subunits within the Pol III core. Thus, the decreased fidelity of DNA replication observed in E. coli strains carrying mutations within the dnaQ gene could be due either to the defective catalytic properties of the  subunit or to aberrant subunit interactions within the Pol III core.

Several mutators which carry mutations in the dnaQ gene have been isolated, e.g., dnaQ49, mutD5, and dnaQ926 strains. Two of these mutations, mutD5 and dnaQ49, have been extensively studied. mutD5 is a particularly strong mutator allele, leading to mutation rates of up to 10⁵-fold above the wild-type level (3, 4). This is due not only to the proofreading defect but also to the concomitant impairment (by saturation) of postreplicative mismatch repair (25, 26). dnaQ49 strains differ from mutD5 strains in several respects. First, dnaQ49 is a temperature-sensitive mutator, possessing modest mutator activity below 30°C but strongly enhanced activity at 37°C (7, 10, 11). Second, dnaQ49 strains are unable to grow at 44.5°C in salt-free rich medium because of the inhibition of DNA synthesis (10). Third, the dnaQ49 allele is recessive with respect to the wild-type gene, while mutD5 is dominant (3, 20). The dnaQ49 and mutD5 alleles result from different missense mutations within the dnaQ gene (9, 30; see also Table 1). On the basis of genetic data, a model in which the DnaQ49 protein has a reduced ability to bind to the  subunit has been proposed (30). The third allele, dnaQ926, is the strongest known mutator of E. coli (9). It was constructed by site-specific mutagenesis by changing the codons for two conserved amino acid residues in the ExoI motif of the  subunit (Table 1) known to be essential for the catalytic activity of other polymerase-associated proofreading exonucleases (1). When residing on a plasmid, dnaQ926 confers a strong, dominant mutator phenotype, suggesting that the protein, although deficient in exonuclease activity, may still efficiently bind to the  subunit. When dnaQ926 was transferred to the chromosome, replacing the wild-type gene, the cells were essentially inviable. dnaQ926 strains survived well, however, when carrying a dnaE antimutator mutation (6, 8) or a multicopy plasmid containing the E. coli mutL⁺ gene. Thus, the poor viability of dnaQ926 strains was proposed to result from excessively high mutation rates due, as in the case of mutD5 strains, both to the proofreading defect and to the collapse of the mismatch repair system (error catastrophe).

A mechanism coordinating DNA polymerization and DNA excision, relying on structural and functional communication between the different subunits of Pol III HE, may play an important role in maintaining optimal fidelity of DNA replication. We have been particularly interested in elucidating the physiological role of interactions between the  and  subunits. In an attempt to achieve this goal, we used the Saccharomyces cerevisiae two-hybrid system to investigate the in vivo interactions between mutant and wild-type DnaQ protein with the wild-type DnaE and HolE proteins.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results & Discussion References

Bacterial and yeast strains and media. E. coli DH5 [supE44 lacU169 (80lacZM15) hsdR17 recA1 endA1 gyrA96 thi-1 relA1] was the transformation recipient strain for all plasmid constructions. All two-hybrid system experiments were done with S. cerevisiae Y187 (MAT gal4 gal80 his3 trp1-901 ade2-101 ura3-52 leu2-3,112 met URA3::GAL1-lacZ). Strain Y187 was kindly provided by S. Elledge (Baylor College of Medicine, Houston, Tex.). Yeast extract-peptone-dextrose medium and synthetic medium (SMM) were prepared as described previously (23). For drug selection, Luria broth plates were supplemented with ampicillin (100 µg/ml).

Methods. Manipulations and sequencing of DNA were carried out by standard procedures (24). The S. cerevisiae Y187 strain was transformed simultaneously with a pGBT9-derived plasmid (e.g., pGBT9dnaE) and a pGAD424-derived plasmid (e.g., pGAD424-2dnaQ) by the method of Chen et al. (2).

-Galactosidase assay. For quantitative studies, yeast strains were grown at 25 or 28°C to stationary phase in synthetic medium (SMM plus 3% glucose) lacking leucine and tryptophan, diluted 10 times in SMM plus 2% ethanol (lacking leucine and tryptophan), and then incubated at 25 or 28°C for 48 h. The -galactosidase activity was determined as described previously (23).

Construction of GAL4 protein fusion plasmids. Plasmids for the GAL4 two-hybrid fusion assay were prepared by cloning PCR-amplified fragments into pGBT9 (Clontech) or pGBT9-2 (13) (both containing amino acids 1 to 147 of the DNA-binding domain of GAL4) and pGAD424 (Clontech) or pGAD424-2 (13) (both containing amino acids 768 to 881 from the trans-activation domain of GAL4). The dnaQ coding region was obtained from plasmid pIP1 (12), the dnaE coding region was obtained from plasmid pMWE103 (kindly provided by C. McHenry [University of Colorado Health Sciences Center, Denver]), and the holE coding region was obtained from the chromosome of E. coli DH5. In all cases these were obtained via PCR amplification with the following forward and reverse primers, respectively: dnaQ, 5'-ATGAGCACTGCAATTACACGC-3' and 5'-TTTTTAGCGCCTTCACAGG-3'; dnaE, 5'-ATGTCTGAACCACGTTTCGTA-3' and 5'-AATCAAGGAAATTCAGACTCA-3'; and holE, 5'-ATGCTGAAGAATCTGGCTA-3' and 5'-CAGGCGTTATGTAAGAAAG-3'. The PCR conditions and cloning of PCR amplification products were as described previously (13). To confirm the presence of the in-frame junction of the genes with the respective GAL4 domains, recombinant plasmids pGBT9-2dnaQ⁺, pGBT9-2holE⁺, and pGBT9dnaE⁺ were sequenced with the primer GAL4bd (5'-GAAGAGAGTAGTAACAAAGG-3'). The entire DNA sequences of the PCR-derived holE and dnaQ inserts were verified by dideoxy sequencing. The dnaE sequence was verified by sequencing pGBT9dnaE⁺ from the ATG codon (position 796) to the HpaI site (position 1035) and from the SmaI site (position 4022) to the 3' end of the gene (numbering system as in reference 31). Then the region in pGBT9dnaE⁺ between the HpaI site (position 1035) and the SmaI site (position 4022) was removed and replaced by the HpaI-SmaI fragment (2,987 bp) from pMWE103 carrying the wild-type dnaE gene.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials & Methods Results & Discussion References

Interactions within the Pol III core. It has been shown that the heterotrimer can be formed in vitro (22, 29). Structural studies showed the  subunit to bind to the  subunit and the  subunit to bind to the  subunit but not to the  subunit, indicating a linear arrangement of the , , and  subunits in the Pol III core. To evaluate the usefulness of the two-hybrid system for studying interactions within the Pol III core, we have examined interactions between the DnaE ( subunit), DnaQ ( subunit), and HolE ( subunit) fusion proteins. We cloned the complete DNA coding sequences of the dnaE, dnaQ, and holE genes, each from the first ATG codon, into both pGBT9 and pGAD424. All pairwise combinations of pGBT9, pGBT9dnaE⁺, pGBT9-2dnaQ, or pGBT9-2holE⁺ and pGAD424, pGAD424dnaE⁺, pGAD424-2dnaQ⁺, or pGAD424-2holE⁺ were introduced into the yeast reporter strain Y187. After selection, cotransformants were screened for their ability to produce -galactosidase by filter assay (data not shown) and by quantitative measurement of -galactosidase activity (Table 2). The results in Table 2 indicate that the DnaE fusion protein is able to interact specifically with the DnaQ fusion protein. Also, the DnaQ fusion protein binds tightly to the HolE fusion protein. The HolE fusion protein does not interact with DnaE fusion protein. The observed efficiency of (hetero)dimer formation, measured by the activity of the lacZ reporter gene, followed the order DnaQ-HolE  DnaQ-DnaE. This does not necessarily reflect the relative strengths of the  subunit- subunit and  subunit- subunit interactions in E. coli cells, as the two-hybrid system results are additionally determined by the levels of expression of various fusion proteins and by any effects of the GAL4 fusion domains on the binding efficiencies. None of the three tested proteins formed homodimers. We noticed stronger interactions of the DnaE and HolE fusion proteins with the DnaQ fusion protein when dnaQ was cloned into pGBT9 than when cloned into pGAD424. We interpret these differences to reflect the need for appropriate tertiary structures of the respective proteins in order to permit the specific interactions. The presence of the additional 113 amino acids of the GAL4 trans-activation domain fused at the N terminus of DnaQ in the case of pGAD424 may interfere with the optimal folding of the protein and, consequently, affect its interaction with the other proteins. Our results on the specificities and strengths of the  subunit- subunit and  subunit- subunit interactions are in good agreement with those obtained from in vitro experiments using purified subunits (29). The present data as well as previous data demonstrating specific interactions between the UmuC and UmuD' or UmuD and UmuD' proteins (13) indicate that the yeast two-hybrid system can be used successfully for investigating protein-protein interactions of E. coli proteins.

View this table: [in this window] [in a new window]   TABLE 2.   Interaction of DnaQ, HolE, and DnaE fusion proteins: quantitative assay of -galactosidase activity

Interactions of the HolE and DnaE fusion proteins with mutant DnaQ proteins. To investigate the importance of proper DnaE-DnaQ interactions in maintaining the high fidelity of DNA replication, we examined the ability of the DnaE fusion protein to interact with several DnaQ mutant proteins. E. coli strains carrying the dnaQ49, mutD5, and dnaQ926 mutations were originally isolated as strong mutators (3, 9, 10, 20). Amino acid substitutions responsible for the mutator phenotype of each mutant allele (Table 1) have been identified (9, 30). If the proper interaction of DnaE with DnaQ has biological significance in maintaining the high fidelity of DNA replication, one may expect that at least some mutations in the dnaQ gene affect the ability of DnaQ to interact with DnaE.

View this table: [in this window] [in a new window]   TABLE 3.   Interaction of HolE and DnaE fusion proteins with wild-type and mutant DnaQ proteins: quantitative assay of -galactosidase activity