HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 275, Issue 24, 18454-18461, June 16, 2000

This Article Abstract Full Text (PDF) All Versions of this Article: 275/24/18454    most recent M000552200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Caspi, R. Articles by Konieczny, I. Search for Related Content PubMed PubMed Citation Articles by Caspi, R. Articles by Konieczny, I. Social Bookmarking                      What's this?

Interactions of DnaA Proteins from Distantly Related Bacteria with the Replication Origin of the Broad Host Range Plasmid RK2^*

Ron Caspi, Donald R. Helinski^§, Marcin Pacek^¶, and Igor Konieczny^¶

From the  Department of Biology, University of California, San Diego, La Jolla, California 92093-0322 and the ^¶ Department of Molecular and Cellular Biology, Faculty of Biotechnology, University of Gdansk, Gdansk, Poland

Received for publication, January 24, 2000, and in revised form, March 28, 2000

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Replication initiation of the broad host range plasmid RK2 requires binding of the host-encoded DnaA protein to specific sequences (DnaA boxes) at its replication origin (oriV). In contrast to a chromosomal replication origin, which functionally interacts only with the native DnaA protein of the organism, the ability of RK2 to replicate in a wide range of Gram-negative bacterial hosts requires the interaction of oriV with many different DnaA proteins. In this study we compared the interactions of oriV with five different DnaA proteins. DNase I footprint, gel mobility shift, and surface plasmon resonance analyses showed that the DnaA proteins from Escherichia coli, Pseudomonas putida, and Pseudomonas aeruginosa bind to the DnaA boxes at oriV and are capable of inducing open complex formation, the first step in the replication initiation process. However, DnaA proteins from two Gram-positive bacteria, Bacillus subtilis and Streptomyces lividans, while capable of specifically interacting with the DnaA box sequences at oriV, do not bind stably and fail to induce open complex formation. These results suggest that the inability of the DnaA protein of a host bacterium to form a stable and functional complex with the DnaA boxes at oriV is a limiting step for plasmid host range.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

DnaA proteins are the universal initiators of replication from the chromosomal replication origin (oriC) in bacteria. They are also essential for initiation of replication of a large number of plasmids. The DnaA protein recognizes and binds specifically to asymmetric 9-base consensus sequences (5'-TTWTNCACA or a close match), named DnaA boxes, which are present in all bacterial chromosomal replication origins that have been studied, as well as in replication origins of DnaA-dependent plasmids (1-3). In Escherichia coli, as well as in Bacillus subtilis, the binding of DnaA proteins to the DnaA boxes at the replication origin promotes destabilization of nearby AT-rich sequences, resulting in unwinding of the DNA double helix, and the formation of an open complex (4, 5). The formation of this open complex, the first step in replication initiation, is followed by the delivery of the DnaB helicase into the open region, and helicase-mediated unwinding of the duplex DNA. The mechanism by which DnaA binding results in unwinding of the AT-rich region is not clearly understood. It has been shown that the DnaA protein of E. coli can recognize and bind to the DnaA boxes at the replication origin of B. subtilis, and, conversely, the DnaA protein from B. subtilis binds to DnaA boxes at the E. coli oriC (5, 6). However, DnaA-mediated unwinding of oriC is species-specific, since neither protein can unwind the heterologous oriC DNA (5). The study of DnaA-dependent broad host range plasmids offers a good opportunity to analyze the nature of DnaA-DNA interactions since DnaA proteins from many different bacterial hosts must bind functionally to DnaA boxes at the plasmid origin of replication to initiate replication of the plasmid.

Plasmid RK2 is a 60-kilobase plasmid that replicates and is stably maintained in a wide range of Gram-negative bacteria (7). A minimal functional origin derived from this plasmid, which retains this broad host range replication activity, consists of the following elements: four DnaA boxes, five repeats of highly conserved 17-mer sequences (iterons), and an AT-rich region containing four 13-mer sequences. Replication from this origin (oriV) requires a plasmid encoded replication protein, TrfA, which binds specifically to the iterons (8, 9). The DnaA boxes in oriV are not identical. Boxes A2, A3, and A4 have one mismatch each from the consensus sequence, while box A1 has two mismatches. It has been shown recently that box A4 plays a particularly important role at least in E. coli in that it directs the binding of the DnaA protein to the other three boxes (10).

Previous work has suggested a model for the initiation of replication of plasmid RK2 in E. coli (11, 12). According to this model, there are at least three steps in the initiation process. In the first step, the host encoded DnaA protein binds to the DnaA boxes, and the plasmid encoded TrfA protein independently binds to the iterons. Binding of the TrfA protein to the iterons by itself results in the formation of a partially open complex, but in the presence of the DnaA protein the TrfA protein forms a fully open complex involving all four 13-mers at the RK2 origin. In the second step of replication initiation, the host encoded helicase, DnaB, in complex with the host DnaC protein, binds to the plasmid-bound DnaA proteins. In the last step of this model, the bound DnaB is delivered to the open DNA region, where it is loaded on the single-stranded DNA and is active in unwinding the plasmid DNA (12). The E. coli DnaA protein, in addition to specifically interacting with the DnaB protein (13), has also been shown to interact with the replication initiation proteins encoded by plasmids R6K and pSC101 as well as with the E. coli DnaG (primase) protein and the tau subunit of DNA polymerase III (14, 15).

A major difference between replication initiation in plasmid RK2 (as well as most of the DnaA-dependent plasmids studied so far) and that of chromosomal origins, is the fact that in most cases this plasmid requires, in addition to the binding of DnaA to DnaA boxes in most bacterial hosts, the binding of a plasmid encoded replication initiation protein to the origin for the formation of an open complex (16). Unlike for E. coli and B. subtilis oriC, binding of DnaA to oriV by itself does not result in opening (melting) of the AT-rich region (11).

Nevertheless, a critical step in the initiation of replication of plasmid RK2 is the successful interaction between the host encoded DnaA protein and the DnaA boxes at the plasmid origin of replication. In the case of plasmid RK2, this is a particularly demanding requirement, considering the ability of the plasmid to replicate in a wide range of Gram-negative bacteria (7) and, therefore, interact at its replication origin with many different DnaA proteins.

To determine the nature of the interactions between different DnaA proteins and the replication origin of plasmid RK2, we obtained DnaA proteins from five organisms, three of which are Gram-negative and capable of maintaining plasmid RK2, and the other two Gram-positive bacteria which are incapable of stably maintaining the plasmid. These proteins were tested for their ability to form functional nucleoprotein complexes at the plasmid replication origin. For this purpose, the dnaA gene of Pseudomonas aeruginosa was isolated and along with the available dnaA genes of E. coli and Pseudomonas putida, polyhistidine-tagged versions of the corresponding DnaA proteins were constructed to facilitate their purification. These three DnaA proteins and the DnaA proteins of the Gram-positive bacteria B. subtilis and Streptomyces lividans were compared for their ability to bind to the four DnaA boxes of plasmid RK2 to form an open complex in the presence of the TrfA protein at the origin region, and to activate the E. coli DnaB helicase.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Bacterial Strains, Media, DNA, Proteins, and Reagents-- The strains and plasmids used in this assay are described in Table I. E. coli strain JP313 was kindly provided by Dr. Joe Pogliano. E. coli strain TC3482 and plasmid pHI206 containing the dnaA gene of P. putida were kindly provided by Dr. Tove Atlung. All bacteria were grown on LB media. Antibiotics concentrations were: 250 mg/liter penicillin, 100 mg/liter ampicillin, 100 mg/liter kanamycin, and 10 mg/liter chloramphenicol. Plasmid pGL5 was provided by Dr. Kjell Anderson. Plasmid pdnaA116 was provided by Dr. Walter Messer. Wild type E. coli DnaA protein was purified by Dr. Alessandra Blasina. C-terminal His₆-tagged DnaA proteins from E. coli, P. putida, and P. aeruginosa were purified as described below. His₆-tagged DnaA protein from S. lividans was kindly provided by Dr. J. Zakrzewska-Czerwinska. B. subtilis DnaA protein was provided by Dr. Walter Messer. The His₆-TrfA/G254D/S267L protein was purified as described (17). The TrfA/G254D/S267L protein, which contains two plasmid copy-up mutations, initiates replication with kinetics similar to those of the largely dimeric wild type protein, but, unlike the wild type, the mutant protein is present largely in the form of a monomer, which is the active form of TrfA for binding to the iterons at the origin (18).

The construction and purification of the C-terminal His₆-tagged E. coli DnaB protein will be described elsewhere. DnaC protein was purified as described (19). Commercially available proteins were HU, SSB, and DNA gyrase from Enzyco, Inc., creatine kinase and bovine serum albumin (fraction V) from Sigma, and DNA restriction and modification enzymes from various companies.

Isolation of dnaA Genes and Construction of His₆-tagged Proteins-- DNA cloning was performed using established laboratory methods. E. coli strain XL1-Blue was used throughout as host. The dnaA gene of plasmid pdnaA116 was amplified by PCR¹ from this plasmid using the primers ECdnaAHis5' (5'-TCACTTTCGCTTTGGCAGCAG-3') and ECdnaAHis3' (5'-GCTCTAGATCAGTGGTGGTGGTGGTGGTGCGATGACAATGTTCTGATTAA-3').

The P. putida dnaA gene in plasmid pHI206 (20) was amplified from the plasmid by PCR, using the primers PPdnaAHis5' (5'-TCAGTGGAACTTTGGCAGCAG-3') and PPdnaAHis3' (5'-GCTCTAGATCAGTGGTGGTGGTGGTGGTGGGTCGTCAGCGTCCGCAGCAG-3').

The P. aeruginosa dnaA gene was identified by BLAST searches against the unfinished P. aeruginosa genome sequencing project. PCR primers were designed based on the putative reading frame, and the gene was amplified by PCR from chromosomal DNA using the primers PAdnaAHis5' (5'-TCCGTGGAACTTTGGCAGCAG-3') and PAdnaAHis3' (5'-GCTCTAGATCAGTGGTGGTGGTGGTGGTGGGTTGTCAGGGTACGCAGCAG-3'). In all cases the 5' primers started at the 4th nucleotide of the gene, and the 3' primers added codons for 6 histidine residues immediately upstream of the stop codon, which was followed by an XbaI restriction site.

The PCR products of the dnaA genes of E. coli, P. putida, and P. aeruginosa were cleaved with the restriction endonuclease XbaI and cloned into the vector pBAD24 which was prepared as described (21), resulting in the plasmids pGK1, pGK2, and pRC57, respectively. A 928-bp PvuII-BbsI fragment from the PCR-amplified E. coli gene in plasmid pGK1 was replaced with the same fragment isolated from pdnaA116, resulting in plasmid pGK3. A 892-bp AatII-BglII fragment from the PCR-amplified P. putida gene in plasmid pGK2 was replaced with the same fragment isolated from pHI206, resulting in plasmid pGK4. All PCR-amplified DNA that was not replaced by wild type DNA was sequenced to ensure that no errors were introduced by the PCR procedure. The nucleotide sequence for the P. aeruginosa dnaA gene has been deposited in the GenBank data base under GenBank accession no. AF229442.

Purification of His₆-tagged DnaA Proteins-- His₆-tagged DnaA proteins from E. coli, P. putida, and P. aeruginosa were over-expressed in E. coli strain JP313 by induction at 37 °C with 0.2% L-arabinose for 3 h, and purified on Ni²⁺-nitrilotriacetic acid-agarose (Qiagen, Inc.) according to the manufacturer's instructions with the following modifications; 500 ml of induced cells were spun down, washed, and resuspended in 3 ml of sonication buffer (50 mM phosphate buffer, pH 8.0, 0.3 M NaCl, 2 mM imidazole, 0.1% Triton X-100). Following sonication and centrifugation, the supernatant was applied to a column containing 1 ml of the Ni²⁺-nitrilotriacetic acid resin, washed twice with buffer containing 50 mM phosphate buffer, pH 8.0, 0.3 M NaCl, 20 mM imidazole, and 0.1% Triton X-100, and once with storage buffer (0.5 M potassium glutamate, 45 mM HEPES/KOH, pH 7.6, 10 mM MgAc, 0.5 mM EDTA, 20% sucrose, 0.1% Triton X-100). The protein was eluted with 0.5 ml of storage buffer containing 250 mM imidazole, and dialyzed against 2 liters of storage buffer.

DNase I Footprinting Assay-- The DNA template in the footprinting assay was a 429-bp EcoRI-HincII fragment isolated from plasmid pSP6 (9). The top strand of this DNA fragment was labeled with [-³²P]dATP by filling-in the recessed end with DNA polymerase I large (Klenow) fragment (Promega). Unincorporated nucleotides were removed using a nucleotide removal kit (Qiagen, Inc.).

DnaA proteins at increasing concentrations were allowed to bind to 5 ng of template DNA at 32 °C for 20 min in the following buffer: 20 mM Tris-HCl, pH 7.6, 80 µg/ml bovine serum albumin, 4% sucrose, 4 mM dithiothreitol, 32 mM HEPES/KOH, 8 mM magnesium acetate, 2 mM ATP, 20 µg/ml poly(dI-dC), 0.08% Nonidet P-40, in a total volume of 50 µl. DNase I footprinting was performed using Promega RQ1 RNase-free DNase, following the procedure recommended by the manufacturer. The samples were loaded on a 8% denaturing polyacrylamide gel.

Gel Mobility Shift and Strand Opening Assays-- The DNA template for the gel mobility shift assay was the same as that used for DNase I footprinting. The fragment was labeled either with [-³²P]dATP by filling-in (as described for footprinting) or with [-³²P]ATP using T4 polynucleotide kinase (Life Technologies, Inc.) according to the manufacturer instructions. Template and proteins were incubated as described for DNase I footprinting, with the exception of using only 0.1 ng of DNA template and a total volume of 20 µl. The samples were loaded on a 5% acrylamide gel with 2 µl of 25% Ficoll solution and no loading dye. Strand opening assays based on permanganate (KMnO₄) footprinting and primer extension using PCR followed by electrophoresis were performed as described previously (22).

Helicase Unwinding Assay-- Helicase unwinding assays were performed as described previously (22) using 300 ng of supercoiled plasmid pKD19L1 and the following concentrations of proteins: DnaA, 375 ng; DnaB, 2000 ng; DnaC, 250 ng; TrfAG254D/S267L, 560 ng; HU, 5 ng; gyrase, 120 ng; and SSB, 230 ng. It was found that HU protein is essential for the strand opening and helicase unwinding assays only when using polyhistidine-tagged versions of DnaA, possibly because previously purified non-His-tagged DnaA proteins contain some HU protein as impurity.

Computer Analysis-- DNA and protein sequence analysis and alignments were performed with the programs VectorNTI (Informax, Inc.), and ClustalX (23). The phylogenetic tree was drawn with the program TreeView (24). Sequence searches were performed over the Internet using BLAST and gapped-BLAST (25).

Surface Plasmon Resonance (SPR)-- To generate the DNA molecule that was used as a substrate for SPR experiments, a 64-bp 5'-biotinylated oligonucleotide (5'-AACGCCTGATTTTACGCGAGTTTCCCACAGATGATGTGGACAAGCCTGGGGATAAGTGCCCTGC -3') was hybridized with a non-biotinylated complementary oligonucleotide. The oligonucleotides were preheated to 95 °C for 10 min and then cooled to 25 °C at a rate of 0.2 °C/min. The resulting double-stranded 64-bp linear DNA fragment contained the four DnaA boxes from oriV, exactly as they appear in the plasmid. The DNA was then immobilized on the streptavidin matrix-coated Sensor Chip S.A. (Pharmacia Biosensor AB) by biotin covalent linkage, following the manufacturer's instructions, by passing the DNA fragment (10 pmol/µl in 300 mM sodium chloride, 1 mM EDTA, and 10 mM Tris pH 7.5) over the chip for 2 min, followed by washing with SPR buffer (100 mM potassium acetate, 10 mM magnesium acetate, 25 mM HEPES, pH 7.6, and 0.005% P20). The final change in response units for immobilization was 67 response units. For the purpose of control experiments, a second DNA fragment was prepared in the same way, in which the DnaA boxes were scrambled, so that there was no specific binding of DnaA proteins to this fragment.

SPR analysis was performed by injecting 10 µl of 500, 1000, and 2000 nM solutions of the five DnaA proteins from E. coli, P. putida, P. aeruginosa, B. subtilis, and S. lividans, respectively, in binding buffer (40 mM HEPES/KOH, pH 8.0, 25 mM Tris/HCl, pH 7.4, 80 µg/ml bovine serum albumin, 4% sucrose, 4 mM dithiothreitol, 11 mM magnesium acetate, and 2 mM ATP) for 2 min at 25 °C. Protein injections were followed by SPR buffer for 3 min for analysis of the K_d value. Following the completion of each protein binding analysis, the surface of the chip was regenerated by injection of 10 µl of 0.05% SDS, which releases all bound protein without affecting the binding capacity of the immobilized DNA. All proteins were tested for binding to both DNA fragments, and the data for binding to the control fragment (which contained scrambled DnaA boxes) were subtracted from the data for binding to the fragment containing the DnaA boxes prior to calculations of binding coefficients and sensograms, to correct for nonspecific binding. All apparent K_a and K_d values were determined through nonlinear curve fitting using the Pharmacia Biosensor kinetics software (BIA evaluation 2.1) assuming the simplest case: A + B  AB.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Sequence Comparison of the Different DnaA Proteins-- Three of the five dnaA genes encoding the proteins used in this study (from E. coli, P. putida, and B. subtilis) have been sequenced and published (26-28). The S. lividans dnaA gene has not been sequenced, but the dnaA gene from Streptomyces coelicolor, which is closely related to S. lividans, has been sequenced (29), and was used for the purpose of sequence alignment. The dnaA gene of P. aeruginosa has not been published, but BLASTN searches against the unfinished P. aeruginosa genome project resulted in a match to a DNA stretch (within contig 54) with 85% identity to the P. putida dnaA gene. The DNA sequence from contig 54 was analyzed, and an open reading frame of the expected size was found. PCR primers designed to amplify this gene were used to amplify it from chromosomal DNA extracted from P. aeruginosa, the amplified PCR product was sequenced, and the translated sequence was used for sequence alignment.

As can be seen in Figs. 1 and 2, the five genes show a relatively high degree of similarity, ranging from 47.1% to 83.7%. The most conserved regions are found in domain III of the protein, which includes the nucleotide binding region (P-loop), and domain IV, which contains the DNA binding site (for a recent review discussing the functional and structural domains of the DnaA protein, see Ref. 30). The S. coelicolor DnaA protein has a much longer domain II, which appears to be not essential in E. coli, but may play an important role in Streptomyces. A phylogenetic analysis (Fig. 2) grouped the DnaA proteins from B. subtilis and S. coelicolor in the same group, but long branch lengths indicate that they are not very close to each other, and the similarity between them (61.4% identity) is comparable to the similarity between the E. coli and the Pseudomonas proteins (64.2% and 63.8% identity with P. putida and P. aeruginosa, respectively) (Fig. 2).

View larger version (85K): [in this window] [in a new window]   Fig. 1.   Amino acid sequence comparison of DnaA proteins. The alignment was performed by the ClustalX program (23). Domains of the E. coli DnaA protein that have been implicated in interactions with other proteins (30) are indicated by arrows. Residues that are conserved in all five proteins are indicated by an asterisk.

View larger version (10K): [in this window] [in a new window]   Fig. 2.   A, relationship among DnaA proteins based on amino acid sequences, as determined by the Neighbor Joining method (44). Alignment was performed by the ClustalX program, and the phylogenetic tree was drawn by the TreeView program. Branch lengths represent evolutionary distance between proteins. B, amino acid identities between the DnaA proteins. The amino acid sequences of the DnaA proteins were aligned as described in Fig. 1, and the percent identity of amino acid between each pair of organisms is shown.

Activity in Vivo of Polyhistidine-tagged Proteins-- Genes encoding C-terminal His₆-tagged fusions of the DnaA proteins from E. coli, P. putida, and P. aeruginosa were constructed to facilitate purification of the proteins. The activity of the His₆-tagged E. coli DnaA protein was tested in vivo using a dnaA(null) mutant of E. coli. This mutant is viable because it possesses a second mutation in the rnh locus, which allows for non-DnaA-dependent chromosome replication initiation, through several sites in the chromosome collectively called oriK (31). However, such mutants are not capable of supporting the replication of DnaA-dependent plasmids. Thus, it is possible to test the activity of a DnaA protein by introducing the dnaA gene into such a dnaA(null) mutant and testing for the ability of the strain to support the replication of a DnaA-dependent plasmid. E. coli TC3482 (dnaA::cat rnh-373) was transformed with either the expression vector pBAD24 (as a negative control), or with plasmid pGK3, a construct based on pBAD24 that expresses the His₆-tagged version of the E. coli DnaA protein under the control of a P_BAD promoter. Since these plasmids contain a colE1 replicon, they can replicate in the absence of DnaA protein. Cells harboring these plasmids were transformed with either pKT230, a derivative of plasmid RSF1010 that does not require DnaA for replication, or with plasmid pGL5, a derivative of plasmid RK2 that requires DnaA for replication (both pKT230 and pGL5 carry kanamycin resistance), plated on media containing kanamycin, and colonies were counted. The results are summarized in Table II.

While transformation of the dnaA(null) strain with plasmid pGL5 by itself resulted in no colonies, transformation of the strain producing the DnaA-His₆ protein with this plasmid resulted in a similar number of colonies to that of the control plasmid pKT230, indicating that the protein is functional in vivo.

The identical assay using the P. putida DnaA-His₆ protein (cloned into plasmid pGK4) instead of the E. coli DnaA-His₆ did not yield transformants containing the RK2-derived plasmid pGL5, indicating the P. putida DnaA protein cannot replace the E. coli DnaA protein for the replication of plasmid RK2 in an E. coli strain. A similar result has been reported earlier, for the inability of the P. putida DnaA protein to support chromosomal replication in a dnaA ts mutant of E. coli (20). We were not able to test the in vivo function of the Pseudomonas proteins in supporting the replication of plasmid RK2 since dnaA(null) Pseudomonas mutants are not available. However, we were able to verify in vitro activity of these proteins by performing the various in vitro assays described below.

Footprints of Different DnaA Proteins at oriV-- DNase I footprinting was carried out to determine whether each of the proteins recognizes the different DnaA boxes at oriV, and to localize and characterize this interaction. Using a 429-bp fragment of oriV that contains the minimal origin of RK2, we compared the extent and nature of DNase I protection conferred by binding of DnaA proteins to the DNA. The results of this assay (Fig. 3) clearly demonstrated that all five proteins, including the two proteins from Gram-positive bacteria, recognize and bind to the DnaA boxes of oriV. The protection conferred by the E. coli protein, as has been documented previously for the non-His-tagged DnaA protein (11), is strongest at boxes 3 and 4, intermediate at box 2, and weak at box 1. The pattern of the two Pseudomonas proteins was indistinguishable from each other, and similar to that of the E. coli protein with small differences in the protection at box 3. The protection pattern of the proteins from the two Gram-positive organisms, however, was strikingly different. In both cases the pattern of protection extended strongly to all four boxes. In addition, the binding of either of these two proteins resulted in new or stronger hypersensitive bands, suggesting that these proteins form a significantly different nucleoprotein structure when compared with the proteins from the Gram-negative bacteria.

View larger version (110K): [in this window] [in a new window]   Fig. 3.   DNase I footprinting of DnaA proteins from five different bacteria bound to the RK2 minimal origin (top strand is shown). The reactions were carried out as described under "Materials and Methods." In each reaction the top strand of a 429-bp EcoRI-HincII fragment from plasmid pSP6 was labeled with [-32P]dATP and used at a concentration of 5 ng. Incubation was carried out with either 50 or 150 ng of DnaA protein for 20 min, prior to the addition of DNase I. Arrows indicate location and orientation of DnaA boxes. Lane 1, no DnaA control; lane 2, E. coli DnaA, 50 ng; lane 3, E. coli DnaA, 150 ng; lane 4, P. putida DnaA, 50 ng; lane 5, P. putida DnaA, 150 ng; lane 6, P. aeruginosa DnaA, 50 ng; lane 7, P. aeruginosa DnaA, 150 ng; lane 8, B. subtilis DnaA, 50 ng; lane 9, B. subtilis DnaA, 150 ng; lane 10, S. lividans DnaA, 50 ng; lane 11, S. lividans DnaA, 150 ng.

Comparison of Binding Activities of DnaA Proteins by Gel Mobility Shift and SPR Analyses-- At a concentration range of 10-40 ng of protein/2 fmol of DNA template, all three DnaA proteins from Gram-negative bacteria produced a gel mobility shift of the oriV fragment, while discrete bands of retarded DNA did not occur with the proteins from Gram-positive bacteria (Fig. 4). Even at a concentration as high as 250 ng of protein/reaction, the Bacillus protein did not retard the DNA template, while the Streptomyces protein apparently formed either DNA-protein aggregates, resulting in high molecular weight complexes, or diffuse complexes due to weak nonspecific binding to the DNA. In addition, a fundamental difference was observed between the DnaA protein from E. coli and those from Pseudomonas species. While the E. coli DnaA protein binds to oriV cooperatively, with preferential binding at low protein concentrations to two DnaA boxes rather than a single one (10), DnaA proteins from both Pseudomonas species apparently bind to the DnaA boxes non-cooperatively (Fig. 4), with stepwise binding to the DnaA boxes as the concentration of these proteins is increased.

View larger version (77K): [in this window] [in a new window]   Fig. 4.   Retardation of a DNA fragment containing oriV by bound DnaA proteins. Gel electrophoretic mobility shift generated by the binding of increasing amounts of DnaA protein to a 429-bp EcoRI-HincII fragment from plasmid pSP6, containing the minimal origin of plasmid RK2. The labeled fragment was used at a concentration of 0.1 ng/reaction. DnaA proteins at different concentrations were allowed to bind to the DNA for 20 min, and then loaded on a polyacrylamide gel. The arrow indicates the position of free DNA with no bound protein. Lane 1 has no DnaA protein added. Increasing amounts of 10, 20, and 40 ng of E. coli DnaA protein (lanes 2-4), P. putida DnaA protein (lanes 5-7), and P. aeruginosa DnaA protein (lanes 8-10), respectively, and of 10, 20, 40, and 250 ng of B. subtilis DnaA protein (lanes 11-14) and S. lividans DnaA protein (lanes 14-18), respectively, were used for incubation with the labeled fragment.

The binding properties of the DnaA proteins to the DnaA boxes at oriV were also examined using SPR. A 64-bp linear fragment of double strand DNA containing the exact sequences of the four DnaA boxes present at oriV was constructed by hybridization of two complementary synthetic oligonucleotides. The DNA fragment was immobilized on a Biosensor chip surface, and DnaA proteins were allowed to interact with it. The change in surface plasmon resonance, which is proportional to the change in mass concentration on the chip surface layer, was used to calculate the amount of protein bound to the DNA. Following binding, the chip surface was washed with buffer, and the dissociation of the protein from the DNA was monitored and used to create a sensogram. The sensograms showed that DnaA proteins from E. coli and Pseudomonas have similar real time kinetics binding to the DNA fragment, while DnaA proteins from B. subtilis and S. lividans have a lower response (Fig. 5). The sensograms were used to calculate apparent constants for association (K_a) and dissociation (K_d) of the DnaA protein with the DNA fragment. Since the proteins can potentially bind to any or all four sites within this fragment, these constants can be interpreted only as apparent constants, representing the affinity of the protein to the whole DNA fragment, rather than constants for a specific protein binding site. As shown in Fig. 5, the DnaA proteins from Gram-negative bacteria have similar values for both constants, with the E. coli protein having somewhat higher affinity than the Pseudomonas proteins. The B. subtilis and the S. lividans DnaA proteins have K_a values somewhat lower than the other proteins, while the averaged K_d values are higher by approximately 100-fold, suggesting that while the proteins exhibit similar association kinetics to the proteins from Gram-negative bacteria, the binding is unstable, and the proteins dissociate from the DNA rapidly.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Binding and dissociation kinetics of DnaA proteins from five different bacteria with a DNA fragment containing the four DnaA boxes of oriV. A, surface plasmon resonance sensograms showing the binding and dissociation kinetics of five different DnaA proteins to a 64-bp double-stranded DNA fragment containing the four DnaA boxes from oriV, immobilized on a Biosensor Chip SA surface. Data are presented as the observed change in response units divided by the molecular weight of the protein. a, E. coli; b, P. aeruginosa; c, P. putida; d, B. subtilis; e, S. lividans (all at 500 nM concentrations). B, apparent binding and dissociation constants of various DnaA proteins to the 64-bp double-stranded DNA fragment containing the four DnaA boxes of oriV, as measured by surface plasmon resonance.

Differential Activities of the Various DnaA Proteins in Strand Opening and Helicase Activation-- It has been shown that the binding of TrfA protein to the iteron sequences at oriV, in the presence of the DNA-bending protein HU, results in a partial local strand melting in the AT-rich region, particularly at two (L, M1) of the four 13-mers present within this region (11). The concurrent binding of the E. coli DnaA protein to the DnaA boxes of this region enhances or stabilizes the open complex formed by TrfA, and extends the opening fully over the four 13-mers. We performed this assay with each of the five DnaA proteins, in the presence of the TrfA protein, and the E. coli HU protein.

The DnaA proteins from the Gram-negative bacteria, but not from the Gram-positive bacteria, were able to enhance the opening by TrfA, and produce an open structure that encompassed all four 13-mers (Fig. 6). The Pseudomonas DnaA proteins were equal to or more active than the E. coli DnaA protein in forming an open complex at oriV. However, the DnaA proteins from either B. subtilis or S. lividans failed to enhance open complex formation by TrfA.

View larger version (66K): [in this window] [in a new window]   Fig. 6.   TrfA-dependent opening of oriV is enhanced by DnaA proteins from Gram-negative bacteria but not by DnaA proteins from Gram-positive bacteria. DnaA proteins from five different bacteria, respectively, were bound to supercoiled DNA of plasmid pKD19L1, containing the origin of plasmid RK2, in the presence of TrfA and HU proteins, as described under "Materials and Methods." Following incubation the reaction mixtures were oxidized by KMnO4, primer extended, and electrophoresed. Panel A, lane 1 has no DnaA protein added; increasing DnaA protein concentrations were 125, 250, 500, and 1000 ng of E. coli DnaA (lanes 2-5), P. putida DnaA (lanes 6-9), and P. aeruginosa DnaA (lanes 10-13), respectively. Panel B, lane 1 has no DnaA protein added; increasing DnaA protein concentrations were 125, 250, 500, and 1000 ng of E. coli DnaA (lanes 2-5), B. subtilis DnaA (lanes 6-9), and S. lividans DnaA (lanes 10-13), respectively.

A critical step following open complex formation in the initiation of replication is the formation of a pre-priming complex. In the case of plasmid RK2, this step involves the recruitment of the DnaB helicase (which in E. coli is present as a DnaB/DnaC complex) by the DnaA protein bound to the DnaA boxes. The DnaB/DnaC complex is then delivered, in the presence of the TrfA protein, to the open complex, where the helicase unwinds the DNA template further.

Helicase activity in vitro on supercoiled DNA, in the presence of ATP, gyrase, and SSB proteins, results in the formation of an extensively unwound form of the supercoiled DNA, designated FI*, which can be detected by agarose gel electrophoresis (32).

Using this assay, each of the DnaA proteins, bound to the DnaA boxes at oriV, was tested for its ability to form an active pre-priming complex consisting of the E. coli DnaB and DnaC proteins, as well as the plasmid encoded TrfA protein. The results, shown in Fig. 7, clearly show that under these conditions only the E. coli DnaA protein supports the formation of the F1* form of the RK2 oriV plasmid.

View larger version (94K): [in this window] [in a new window]   Fig. 7.   Only the E. coli DnaA protein can form an active pre-priming complex with the E. coli helicase. Helicase mediated unwinding of the oriV template was performed as described under "Materials and Methods." All reactions included E. coli DnaB and DnaC proteins, as well as E. coli SSB protein, HU protein, gyrase, and creatine kinase, except for lane 2, where DnaC was omitted. The only factor that was changed in different reactions was the source of the DnaA protein. F III, open circular DNA; F II, linear DNA; F I, covalently closed circular DNA; F I*, extensively unwound covalently closed circular DNA, generated by helicase activity. Lane 1, E. coli DnaA; lane 2, E. coli DnaA, no DnaC present; lane 3, P. putida DnaA; lane 4, P. aeruginosa DnaA; lane 5, B. subtilis DnaA; lane 6, S. lividans DnaA.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The mechanism by which binding of the DnaA protein to DnaA boxes at bacterial replication origins promotes unwinding of the AT-rich region is not understood, but some insight to this process can be gained by studying the interaction of DnaA proteins with the DnaA box sequences of different organisms. The results of the present study with the broad host range plasmid RK2 and published work on bacterial replication origins clearly show that dnaA protein recognition of DnaA box sequences per se is not sufficient for productive interaction. When the sources of the DnaA protein and the target DNA are closely related, the specific binding activities of the proteins are very similar and they are able to substitute for each other. For example, the DnaA proteins from the enteric bacteria Salmonella typhimurium and Serratia marcescens are so similar to the E. coli protein that they can complement an E. coli dnaA ts mutation (33). However, when the phylogenetic distance between two bacteria is greater, the DnaA proteins fail to form an active nucleoprotein complex with the heterologous DNA, despite the fact that they retain the ability to recognize the heterologous DnaA box sequences (which are highly conserved), and to bind to them (5). This implies that the DnaA protein from closely related bacteria not only exhibit similar binding activities and form similar nucleoprotein structures but they also can interact with heterologous initiation proteins such as DnaB in the initiation of replication.

DnaA protein from P. putida, which phylogenetically is not very distant from E. coli, was shown to bind to E. coli DnaA boxes, but could not complement a dnaA(null) mutation in vivo (20). It has also been shown that DnaA protein from E. coli can recognize and bind to the DnaA boxes at the replication origin of B. subtilis, and that the DnaA protein from B. subtilis binds to DnaA boxes at E. coli oriC (5, 6); however, neither protein could form an open complex with the heterologous oriC DNA (5). In this context, the ability of a broad host range plasmid to interact with DnaA proteins from distantly related bacteria, and to form with each of them a functional open complex, is quite remarkable, and makes broad host range plasmids valuable systems for comparing functional and non-functional interactions of DnaA proteins with DnaA boxes at the origin of replication.

The present study examines for the first time the nature of molecular interactions between the origin region of a broad host range plasmid and DnaA proteins from distantly related bacteria. While a considerable amount of knowledge exists about the replication initiation of plasmid RK2 in E. coli, little is known about this process in other bacteria. In a recent study different deletion mutations of DnaA boxes within oriV were constructed and analyzed for the ability of the plasmid to replicate in four Gram-negative bacteria (34). It was shown that the first two DnaA boxes (boxes A1 and A2) were dispensable for plasmid replication in E. coli, P. putida, P. aeruginosa, and Azotobacter vinelandii. Deletion of the first three boxes (A1-A3) resulted in severe impairment of replication in E. coli and P. putida, but not in P. aeruginosa and A. vinelandii, and deletion of all four boxes prevented replication in E. coli, P. putida, and A. vinelandii, and reduced the level of stable maintenance of the plasmid in P. aeruginosa. The observation that alterations of the DnaA binding region of oriV results in modification of host range properties further supports the notion that DnaA binding plays a crucial role in RK2 replication initiation in various organisms.

In order to be able to analyze the interactions of the various DnaA proteins with oriV, five different assays were used to resolve several distinct steps in the initiation process: DNase I footprinting, gel mobility shift, and SPR analyses examine the formation of the initial nucleoprotein complex involving the binding of the DnaA protein to the DnaA boxes at the RK2 origin region. DNase I footprinting allows for measurements of both weak and strong binding to the target DNA while the gel mobility assay provides some measure of the stability of the nucleoprotein complex in addition to binding specificity. Indeed, the results from each of the three assays indicated differences in the binding properties of all five proteins. Presumably, under the conditions of the gel mobility shift assay, the DnaA proteins from Gram-positive bacteria did not bind stably enough to the DnaA boxes to produce distinct bands of retardation. This qualitative observation was confirmed quantitatively by SPR analysis, which revealed that, although the K_a of all proteins was similar, the K_d for B. subtilis and S. lividans was much higher, indicating that the binding of these proteins to the DnaA boxes at oriV is very weak.

These results by themselves do not disallow the binding of the Bacillus and Streptomyces proteins in vivo. However, the different pattern of binding of these proteins, as evidenced by changes in the hypersensitive bands generated in the footprinting assay, as well as their inability to form an open complex in the presence of the TrfA protein, as shown by the strand opening assay, clearly indicate that these proteins are unable to form a functional nucleoprotein complex at oriV. This lack of functional binding is likely an important factor in the inability of the RK2 plasmid to replicate in Bacillus and Streptomyces.

An interesting feature of DnaA protein binding to oriV is cooperativity. It has been shown before that the E. coli DnaA protein binds to oriV in a cooperative way, with box A4 acting as an "organizer" for the formation of the DnaA-oriV nucleoprotein complex (10). This cooperative binding of the DnaA protein, with the exception of plasmid pSC101, is unique to RK2. In the case of pSC101, the cooperativity requires interactions with the replication protein, RepA, and the DNA-bending protein IHF (35). In other plasmids, as well as in oriC, the E. coli DnaA protein binds without observed cooperativity. There is a certain similarity between the E. coli and P. putida DnaA proteins in their requirements for the oriV DnaA boxes in vivo (34); plasmids containing a deletion of DnaA boxes A1-A3 had only marginal replication activity in P. putida or E. coli, and plasmids that were missing all four boxes were not able to replicate at all. In contrast, deletion of the four DnaA boxes at oriV had a less severe affect on replication in P. aeruginosa (34). The possibility has been raised that an alternative mechanism of initiation of replication that is not absolutely dependent on the four DnaA boxes can be used by the plasmid in P. aeruginosa (34). Despite the in vivo similarities in DnaA box requirements for RK2 replication in P. putida and E. coli, their respective DnaA proteins differ in their binding activity with respect to cooperativity.

The formation of a full open complex at oriV by all three DnaA proteins from Gram-negative bacteria indicates that not only do the three proteins bind specifically and strongly enough to the DnaA boxes but these proteins form a nucleoprotein complex along with the plasmid TrfA initiation protein that destabilizes all four 13-mers in the AT-rich region. To date there is no evidence for a specific interaction between the DnaA protein and the TrfA protein, as in the case of the plasmids R6K (14) and pSC101 (15) replication initiation proteins. If a specific interaction is required between TrfA and DnaA for the formation of an open complex, then the three DnaA proteins from the Gram-negative bacteria must share a similar site for interacting with TrfA.

While the failure of the DnaA proteins from the Gram-positive bacteria to form an open complex and activate the E. coli helicase at the RK2 origin is likely due to the inability of these proteins to form a stable and functional complex at this origin, the inability of the Pseudomonas DnaA proteins, which form a functional nucleoprotein complex, to activate the DnaB helicase at oriV as measured by F1* formation is most likely due to their being unable to interact with the E. coli helicase. The DnaA and DnaB proteins of E. coli have been shown to specifically interact with each other (13). Various functional domains of DnaB have been identified, and although comparative analysis indicates that most of the major domains in the E. coli, P. putida, and P. aeruginosa DnaB proteins are similar in sequence, the domain of the DnaB helicase responsible for specific interaction with the DnaA protein is dissimilar in sequence between the three DnaB proteins.² This dissimilarity in the DnaA binding domain of the helicases likely accounts for the failure of the Pseudomonas DnaA proteins to interact with, and along with the TrfA protein, activate the helicase and generate the F1* replication intermediate.

A feature of plasmid RK2, which was suggested in the past as a potential adaptation for broad host range specificity, is the fact that the gene encoding the replication protein, trfA, has two in frame translational start sites, resulting in a 44-kDa protein and a truncated 33-kDa protein. It has been shown that, while in E. coli there is not much difference between the two forms, in P. aeruginosa there is a strong preference for the larger protein (36-38). With the DnaA protein from P. aeruginosa at hand, we had the ability to compare the activity of the two forms of the TrfA proteins in the presence of the P. aeruginosa DnaA protein. Strand opening assays were conducted with these two forms of TrfA (data not shown), and no differences in activity were found. It is possible that the difference between the two proteins is significant in a later step, e.g. in the loading and activation of the helicase. This possibility will be tested when the DnaB helicases from the two Pseudomonas species, which are in the process of being isolated, are available and tested in the F1* assay. The availability of the DnaB proteins from Gram-negative bacteria distantly related to E. coli will in general allow additional experimentation in vitro involving heterologous combinations of proteins that should substantially extend our understanding of the unique properties of the RK2 TrfA protein and replication origin that account for the broad host range replication properties of this plasmid.

	ACKNOWLEDGEMENTS

We thank Gavin Klinger for laboratory assistance. We are grateful to Dr. A. Toukdarian for discussions and suggestions during the course of this study, and for critical reading and assistance in preparation of this manuscript.

	FOOTNOTES

^* This work was supported in part by National Institutes of Health Research Grant AI-07194.The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submitted to the GenBank^TM/EMBL Data Bank with accession number(s) AF229442.

^§ To whom correspondence should be addressed: Dept. of Biology, University of California, San Diego, 9500 Gilman Dr., La Jolla, CA 92093-0322. Tel.: 858-534-3638; Fax: 858-534-0559; E-mail: dhelinski@ucsd.edu.

Supported by MEN/NIH Research Grant 98-349 from the United States-Polish Maria Sklodowska Curie Fund II and by Polish State Committee for Scientific Research Grant 6P04A01115.

Published, JBC Papers in Press, April 3, 2000, DOI 10.1074/jbc.M000552200

² R. Caspi, unpublished observations.

	ABBREVIATIONS

The abbreviations used are: PCR, polymerase chain reaction; contig, group of overlapping clones; SPR, surface plasmon resonance; bp, base pair(s).

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

CiteULike

Complore

Connotea

Del.icio.us

Digg

Reddit

Technorati

This article has been cited by other articles:

				M. Rajewska, L. Kowalczyk, G. Konopa, and I. Konieczny Specific mutations within the AT-rich region of a plasmid replication origin affect either origin opening or helicase loading PNAS, August 12, 2008; 105(32): 11134 - 11139. [Abstract] [Full Text] [PDF]

				T. C. Jarvis, A. A. Beaudry, J. M. Bullard, N. Janjic, and C. S. McHenry Reconstitution of a Minimal DNA Replicase from Pseudomonas aeruginosa and Stimulation by Non-cognate Auxiliary Factors J. Biol. Chem., March 4, 2005; 280(9): 7890 - 7900. [Abstract] [Full Text] [PDF]

				T. Betteridge, J. Yang, A. J. Pittard, and J. Praszkier Role of RepA and DnaA Proteins in the Opening of the Origin of DNA Replication of an IncB Plasmid J. Bacteriol., June 15, 2004; 186(12): 3785 - 3793. [Abstract] [Full Text] [PDF]

				Z. Zhong, D. Helinski, and A. Toukdarian A Specific Region in the N Terminus of a Replication Initiation Protein of Plasmid RK2 Is Required for Recruitment of Pseudomonas aeruginosa DnaB Helicase to the Plasmid Origin J. Biol. Chem., November 14, 2003; 278(46): 45305 - 45310. [Abstract] [Full Text] [PDF]

				Y. Jiang, M. Pacek, D. R. Helinski, I. Konieczny, and A. Toukdarian A multifunctional plasmid-encoded replication initiation protein both recruits and positions an active helicase at the replication origin PNAS, July 22, 2003; 100(15): 8692 - 8697. [Abstract] [Full Text] [PDF]

				I. Konieczny and K. Liberek Cooperative Action of Escherichia coli ClpB Protein and DnaK Chaperone in the Activation of a Replication Initiation Protein J. Biol. Chem., May 17, 2002; 277(21): 18483 - 18488. [Abstract] [Full Text] [PDF]