INTRODUCTION

The initiation of replication of a plasmid that is able to be maintained in a diverse group of bacteria (broad host range) is considerably less well understood than the initiation of replication of other prokaryotic replicons. It is of particular interest to determine the nature of the interactions between a broad host plasmid origin of replication, the plasmid-encoded replication initiation protein and the host specific replication proteins that are responsible for initiating replication. The broad host range plasmid RK2 requires for its replication in Escherichia coli an origin of replication (oriV) (Fig. 1) and a plasmid encoded initiation protein (TrfA) (1-3) which binds as a monomer to 17-bp¹ direct repeats (iterons) at this origin (4, 5). In addition, RK2 replication in E. coli requires the host specified proteins DnaA, DnaB, DnaC, DNA gyrase, DnaG primase, DNA polymerase III holoenzyme, and SSB (6, 7). It has been shown recently that the E. coli DnaA protein binds to four DnaA consensus sequences that potentially can form a cruciform structure within the RK2 origin (8). In this same study it was found that the TrfA initiation protein in the presence of HU produces an opening of a set of 13-mers located in the A + T-rich region of the RK2 oriV. The DnaA protein enhances and/or stabilizes this open complex formation, but cannot on its own form an open complex.

Initiation of replication at the E. coli chromosome origin (oriC) has been studied extensively and the formation of several distinct nucleoprotein complexes has been described (9). The initial step involves binding of the DnaA protein to DnaA boxes localized within oriC (10, 11). This binding results in destabilization of the duplex DNA at the A + T-rich region and open complex formation (12-14). The DnaB helicase (15) in the form of a DnaB-DnaC complex is specifically loaded at the open region of the origin (16-18). Cross-linking, enzyme-linked immunosorbent assay, and monoclonal antibody interference studies have shown that a physical interaction between the DnaA protein and the helicase is required for loading (19). An E. coli prepriming complex consisting of oriC DNA and the DnaA, DnaB, DnaC, and HU proteins can be isolated in vitro (19, 20). Two stages of prepriming complex formation at oriC have been described (9). During the first stage (prepriming complex I) helicase is loaded but is not active. It has been proposed that the activation of prepriming complex I is the result of the repositioning of the helicase leading to the formation of prepriming complex II. At this stage helicase can unwind template DNA thus allowing the priming reaction to occur.

This study utilizes an in vitro replication system reconstituted from purified components to address the mechanism of helicase loading during the initiation of plasmid RK2 replication in E. coli and the role of the DnaA and TrfA proteins in this process. We show that not unlike that observed with oriC (19), a specific interaction between DnaA and DnaB is required for helicase delivery to the initiation complex. The TrfA protein activates the DnaB helicase for template unwinding presumably by forming a specific nucleoprotein structure and an open complex at the A + T-rich region of the origin. These results are discussed in the context of the broad host range replication properties of the RK2 plasmid.

MATERIALS AND METHODS

Purified proteins were used for the various assays. DnaB (21), DnaC (22), and the copy-up mutant TrfA-33 254D/267L (23) proteins were purified as described previously. Preparations of DnaA protein and histidine tagged version of the copy-up mutant TrfA protein designated His6-TrfA 254D/267L were provided by Dr. Aresa Toukdarian (University California, San Diego) and Dr. Alessandra Blasina (Scripps Research Institute). DNA polymerase III holoenzyme was kindly provided by Dr. Michael O'Donnell (Cornell University Medical College). The anti-DnaA monoclonal antibody M7, the anti-DnaC antibody and pBSoriC plasmid containing oriC DNA fragment (24) were kindly provided by Dr. Jon Kaguni (Michigan State University). Polyclonal anti-DnaA antibody was provided by Dr. Walter Messer (Max Plank Institute, Berlin). Anti-DnaB antibody was provided by Dr. Jaroslaw Marszalek (University of Gdansk). pTJS42 is a mini-replicon of plasmid RK2 and contains the five iteron minimal oriV (25). Commercially available proteins and chemicals used in this study were: DNA gyrase, DnaG primase, HU, and SSB from Enzyco, Inc.; bovine serum albumin (fraction V), creatine phosphate, creatine kinase, and rNTPs from Sigma; dNTPs and Sepharose CL-4B from Pharmacia; [methyl-³H]dTTP from ICN Radiochemicals and goat anti-rabbit IgG from Bio-Rad.

The RK2 oriV DNA replication reaction was established with purified components similar to those required for in vitro replication of oriC (26, 27). The in vitro replication mixture (25 µl) contained: 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; 2 mM ATP; 50 µM of each dNTP; [methyl-³H]TTP (150 cpm/pmol); 500 µM (each) CTP, GTP, and UTP; 8 mM creatine phosphate; 20 µg/ml creatine kinase; 230 ng of SSB; 120 ng of DNA gyrase; 1600 ng of DnaB; 100 ng of DnaG primase; 55 ng of DNA polymerase III core subunit; 55 ng of  subunit; 15 ng of  subunit; 10 ng of  complex; 600 ng of DnaA; 800 ng of DnaC; 300 ng of RK2 oriV (pTJS42); and 400 ng of His6-TrfA 254D/267L. Reactions were assembled on ice and then incubated at 32 °C for 30 min. Reactions were stopped by placing on ice and adding 1 ml of 0.1 M sodium pyrophosphate in 10% trichloroacetic acid. Total nucleotide incorporation (picomoles) was measured by liquid scintillation counting after filtration onto Whatman GF/C glass fiber filters. The reaction mixture for oriC plasmid replication was assembled as described for RK2 oriV DNA replication except that the TrfA protein and RK2 oriV plasmid DNA were omitted and the reaction was supplemented with 200 ng of pBSoriC DNA. Reaction mixtures were incubated for 30 min at 32 °C.

Column gel filtration was used to isolate RK2 prepriming complexes. The reaction mixture (total volume 100 µl: 40 mM Hepes/KOH pH 8.0, 40 mM potassium glutamate, 10 mM magnesium acetate, 50 µg/ml bovine serum albumin, 4% sucrose, 4 mM dithiothreitol, and 2 mM ATP) contained the amount of proteins equivalent to four standard in vitro replication reactions except that SSB, DNA gyrase, primase, and DNA polymerase III holoenzyme, CTP, GTP, UTP, dNTP's, and the ATP regeneration system were omitted. The TrfA-33 254D/267L protein was used instead of His6-TrfA 254D/267L and HU protein (100 ng) was added. The reaction mixtures were incubated for 20 min at 32 °C. After incubation the reactions ware run through a Sepharose CL-4B column (0.5 × 12 cm), equilibrated at room temperature with the incubation buffer and 0.01% Brij 58. Fractions (80 µl) were collected and a portion of each (40 µl) was analyzed by SDS-polyacrylamide gel electrophoresis, followed by a semi-dry protein transfer and immunoblot with rabbit antisera specific against DnaA, TrfA, DnaB, and DnaC proteins. Bound rabbit antibody was detected by a colorimetric reaction with an alkaline phosphatase conjugate goat anti-rabbit IgG.

The replication reaction assay was similar to the RK2 in vitro replication reaction reconstituted with purified components described above. A 20-µl portion of each fraction collected from the Sepharose CL-4B column was supplemented by the addition of a mixture (10 µl) containing all other replication components in a standard replication buffer. After 30 min incubation at 32 °C, reactions were stopped by placing on ice followed by the addition of 0.1 M sodium pyrophosphate and 10% trichloroacetic acid. The total nucleotide incorporation (picomoles) was measured as described above.

The reaction mixture is identical to the RK2 in vitro replication system reconstituted with purified components except that the DnaG primase and DNA Pol III holoenzyme components were omitted. The reaction mixtures were incubated for 30 min at 32 °C and the reactions were stopped by the addition of EDTA and SDS at a final concentration of 10 mM and 2%, respectively, followed by 2 min incubation at 65 °C. Sucrose and bromphenol blue were then added to the reaction mixture to a final concentration of 10 and 0.05%, respectively. The mixture was then analyzed on a 1% agarose gel in TBE buffer (0.09 M Tris borate, 0.002 M EDTA). The samples were electrophoresed at 25 V for 20 h and the gel was stained with ethidium bromide solution.

RESULTS

The RK2 in vitro replication system described in this study was established with highly purified proteins. As for the RK2 in vitro replication system using E. coli crude extract (6, 7), the replication mixture contained an ATP regeneration system, rNTPs, dNTPs, and MgOAC. The RK2 oriV replication template was the supercoiled DNA form of plasmid pTJS42 which contains a 393-bp RK2 minimal origin (Fig. 1). The replication reaction demonstrated a stringent dependence on oriV containing DNA and the copy-up mutant His6-TrfA 254D/267L protein (Table I). The largely dimeric His6-TrfA protein was unable to support replication in the purified system (data not shown). The monomer form of TrFA is the active form for binding to the iterons at the RK2 origin (5) and since the His6-TrfA 254D/267L protein is largely in the form of a monomer,² this protein was used for these and also previous studies (8).

[View Larger Version of this Image (12K GIF file)]

Table I. Protein requirements for RK2 oriV replication in the reconstituted system The purified TrfA protein designated His6-TrfA 254D/267L is the double copy-up mutant form of the TrfA-33 protein. Component omitted DNA synthesis pmol % None 295 100 oriV DNA 2 <1 TrfA 5 2 DnaA 2 <1 DnaB 3 1 DnaC 6 2 DnaG primase 17 6 DNA gyrase 40 13 DNA pol III 4 1 SSB 8 3

While there is a strong dependence on DnaB, DnaC, DNA gyrase, DnaG primase, SSB, and polymerase III holoenzyme in the reconstituted system (Table I), deleting the  subunit of polymerase III holoenzyme only slightly lowered in vitro replication activity (data not shown). The HU protein was found to be dispensable for RK2 replication in the purified system although a previous study showed it was essential for open complex formation in the absence of the DnaA protein (8).

Replication of intact RK2 plasmid or RK2 mini-replicons in E. coli is DnaA dependent (6, 7, 28). In the reconstituted RK2 replication system there is a stringent requirement for the E. coli DnaA protein (Table I). It is of interest that the amount of DnaA protein required for E. coli oriC and RK2 oriV replication in vitro differs (Fig. 2). In comparison to oriC, RK2 oriV replication in vitro requires at least 5-fold less DnaA protein for maximum DNA synthesis (with molar ratios approximately 1:75 for oriC:DnaA and 1:15 for oriV:DnaA). This may indicate differences in affinity for the DnaA boxes in the two origins, or differences in the nucleoprotein structures of oriV-DnaA and oriC-DnaA. Unlike with oriC (29, 30) an excess of DnaA protein at the highest concentrations used did not result in an inhibition of oriV replication (Fig. 2). The kinetics of the replication reactions with supercoiled pTJS42 or pBSoriC DNA as templates were not found to be significantly different (data not shown).

[View Larger Version of this Image (18K GIF file)]

The initiation of replication of E. coli oriC and bacteriophage  DNA requires that the helicase be delivered to and loaded on ssDNA to convert the open complex into an activated prepriming complex (18, 31). During the initiation of replication of oriC DNA in vitro, DnaA, DnaB, and DnaC proteins interact and associate with oriC to form a prepriming complex (32, 33). It has been shown that a specific interaction between DnaB and DnaA proteins is critical for the formation of this complex and helicase delivery at oriC (19). To investigate the mechanism of helicase delivery to oriV during the initiation of RK2 replication, prepriming complexes were formed and then isolated by gel filtration using a Sepharose CL-4B column. When the DnaB helicase was incubated with supercoiled oriV plasmid DNA (pTJS42) and the DnaC, DnaA, and TrfA proteins, all four proteins were detected along with pTJS42 DNA in the column void volume (Fig. 3A). When oriV template DNA was incubated only with DnaA, DnaB, and DnaC proteins, and not with the TrfA protein, again DnaA, DnaB, and DnaC were found in the void volume with pTJS42 DNA. Although under the conditions of this experiment these three proteins together do not open pTJS42 at the origin in the absence of the TrfA protein (8), it is clear that DnaB and DnaC form a prepriming complex with the DnaA protein and oriV DNA (Fig. 3B). When both the SSB and HU proteins were added to the preincubation mixture containing oriV and the DnaA, DnaB, and DnaC proteins, the DnaB helicase and DnaC proteins were once again observed in the void volume along with the DnaA protein (data not shown). In contrast, when the experiment was carried out in the absence of the DnaA protein but in the presence of the TrfA protein and under conditions that promote origin opening by the TrfA protein (i.e. the presence of HU), we did not observe DnaB helicase and DnaC in the void volume, indicating that the TrfA protein by itself cannot deliver the helicase to the RK2 origin (Fig. 3C). Finally, incubation of only DnaB and DnaC with the supercoiled pTJS42 DNA did not result in the presence of DnaB or DnaC in the void volume (Fig. 3D).

[View Larger Version of this Image (39K GIF file)]

We also tested the Sepharose CL-4B void fractions for replication activity. The complex of oriV, TrfA, DnaA, DnaB, and DnaC found in the void volume (Fig. 3A) was active under complete in vitro replication assay conditions, i.e. when gyrase, DnaG primase, SSB, Pol III holoenzyme, rNTPs, and dNTPs were added (Fig. 4). The recovery of replication activity was estimated to be about 80% of the activity of an equivalent amount of incubation mixture that was not subjected to Sepharose CL-4B chromatography. The in vitro replication activity found for the prepriming complex formed from oriV and the DnaA, DnaB, and DnaC proteins without the TrfA protein (Fig. 4) was approximately at a background level, i.e. similar to that obtained when both the TrfA and DnaA proteins were not present during the formation of the complex.³ The relatively high background level observed for the complex formed in the absence of the TrfA protein was probably due to degradation of a portion of the template during the course of the experiment and, consequently, nonspecific nucleotide incorporation. The subsequent addition of TrfA to the complex formed in the absence of TrfA restored replication activity to a level approaching that seen with the complex formed from all four proteins (Fig. 4). These results indicate that the TrfA protein is still required for replication even after the helicase is delivered to the RK2 origin and that the formation of a prepriming complex that includes oriV, DnaA, DnaB, and DnaC can precede the formation of a TrfA-dependent open complex.

[View Larger Version of this Image (21K GIF file)]

We determined the kinetics of the oriV in vitro replication reaction using different preincubation conditions. Preincubation of TrfA, DnaA, DnaB, DnaC, and gyrase proteins with pTJS42 in the presence of HU and SSB resulted in an approximately 2-min advance in the time for significant nucleotide incorporation (Fig. 5). Presumably under these conditions, a prepriming complex consisting of oriV, TrfA, DnaA, DnaB, and DnaC is formed. By comparison, a 1-min time difference was observed when either the TrfA or the DnaA protein was omitted during the preincubation period (Fig. 5). These results may indicate that the DnaA and TrfA proteins act independently in the formation of a prepriming complex during the preincubation period.

[View Larger Version of this Image (20K GIF file)]

Several template unwinding assays have been developed for measuring helicase unwinding activity (17, 18, 34, 35). In this study the agarose TBE-electrophoresis method was used to determine unwinding of the supercoiled pTJS42 DNA template (35). The formation of an electrophoretically distinct form of DNA, designated FI^*, has been shown to be produced as a result of the unwinding activity of a helicase in the presence of gyrase and SSB proteins (18, 34). This covalently closed circular DNA exhibits more rapid electrophoretic mobility and, as shown by electron microscopy (18), is extensively single-stranded. Using RK2 in vitro replication conditions including ATP, TrfA, DnaA, DnaB, DnaC, gyrase, HU, and SSB proteins, we observed unwinding of a substantial fraction of the oriV supercoiled DNA template molecules as determined by the appearance of the FI^* form of pTJS42 DNA (Fig. 6). Not unlike that found for the initiation of oriC and  DNA replication, the formation of extensively unwound DNA is gyrase dependent (18, 34). In comparison to unreacted supercoiled pTJS42 DNA, and not unexpectedly, we observed that this extensively single-stranded reaction product was very sensitive to nuclease P1 digestion (data not shown). Omission of TrfA or DnaA protein resulted in a failure to produce the fast migrating FI^* form (Fig. 6). Thus, DnaB helicase activity on the oriV template requires both the DnaA and TrfA proteins. These results suggest that helicase remains inactive in the prepriming complex consisting of oriV, DnaA, DnaC, and DnaB and that activation requires the TrfA protein.

[View Larger Version of this Image (56K GIF file)]

A specific interaction between the DnaA and DnaB proteins has been shown to be required for helicase loading at oriC (19). Interference of this protein-protein interaction using monoclonal antibody M7 results in an inhibition of in vitro oriC replication (19). We examined the effect of the M7 antibody on RK2 replication in vitro. OriV replication was inhibited by approximately 50% with 50 ng and 90% with 200 ng of M7 antibody (Fig. 7). These results indicate a similar requirement for a specific DnaA-DnaB interaction for the formation of an RK2 prepriming complex as was found for the E. coli replication origin.

[View Larger Version of this Image (19K GIF file)]

Altering the order of addition of the M7 monoclonal antibody and the DnaA, DnaB, DnaC, and TrfA proteins during the assay for helicase activity allowed us to obtain information as to what stage the antibody was inhibitory. When M7 antibody was added to the incubation mixture with DnaA but before DnaB, DnaC, and TrfA were added, we observed an approximately 75% inhibition of the formation of the FI^* form (Fig. 8). This is consistent with the M7 antibody blocking the DnaA-DnaB interaction and, therefore, delivery of the helicase to the template. In contrast, the addition of M7 after incubation of DnaA, DnaB, DnaC, and TrfA had little or no inhibitory effect on helicase activity (Fig. 8). Surprisingly, when the order of addition was changed and the TrfA protein was added at the last step together with M7 antibody, we also observed reduced helicase activity (Fig. 8). This result may indicate instability of the prepriming complex involving the DnaA, DnaB, and DnaC proteins in which case the M7 monoclonal antibody can continue to compete with the DnaB protein for binding to DnaA even if added after the addition of both the DnaA and DnaB proteins to the supercoiled plasmid DNA template. If this is the case, then when the TrfA protein is present along with the DNA, DnaB, and DnaC proteins, the DnaA-DnaB interaction is stabilized or the DnaB protein is repositioned on the template so that it no longer is associate with the DnaA protein.

[View Larger Version of this Image (62K GIF file)]

DISCUSSION

We have demonstrated that plasmid RK2 can be replicated in vitro using purified proteins if the plasmid-specific initiation protein, TrfA, is added. This reconstituted replication system, similar to the system described previously for oriC (26, 27), thus, can support the replication of supercoiled RK2 oriV and E. coli oriC DNA. In the case of RK2 oriV replication, we found that a much lower concentration of the DnaA protein is required for maximum replication of oriV when compared with oriC (Fig. 2). We also found that the excess of DnaA protein at the highest concentrations used in our experiments inhibited oriC replication but did not result in an inhibition of oriV replication (Fig. 2). The inhibition of oriC in vitro replication by high concentrations of DnaA protein was observed previously (29, 30). These observations may reflect differences in the structure of DnaA-oriV and DnaA-oriC nucleoprotein complexes. The wild-type RK2 replication initiation protein was unable to support in vitro replication in this reconstituted system, however, the largely monomeric mutant TrfA 254D/267L protein was active with the reconstituted purified system (Table I). It has been shown previously that the monomer form of the TrfA protein is the active form for binding to the iterons at RK2 origin (5). The TrfA 254D/267L protein is largely in the form of a monomer as a result of two specific mutations in the trfA gene.⁴ The double mutant can be considered an activated form of the RK2 initiation protein. In a crude extract system the wild-type TrfA is active (7) presumably because it is converted to the monomer form by chaperones. In fact, recently it has been found that the E. coli ClpX chaperone activates wild-type TrfA by conversion of the dimer form of the protein to the monomer form.⁵

Several DnaA binding consensus sequences have been identified within the RK2 origin region (36). Recently, DNase I footprinting experiments have shown that the DnaA protein binds to four DnaA boxes that are located upstream of the TrfA-binding sites (8). These four DnaA-binding sites are arranged as two pairs in an inverted orientation with respect to each other. Unlike the E. coli oriC and the Pl plasmid origins of replication, binding of the DnaA protein to these sites does not result in the formation of an open complex at oriV (8). The present study, however, demonstrates that the DnaA protein has an indispensable role in the delivery and loading of the DnaB helicase at the RK2 origin. This is supported by the finding that an oriV-DnaA-DnaB-DnaC prepriming complex (prepriming complex I) can be isolated by gel filtration after incubation of the replication proteins with a supercoiled RK2 template (Fig. 3B). The presence or absence of the TrfA protein does not appear to influence recruitment of the helicase by the DnaA protein. It was found further that, not unlike the findings for the ABC priming reaction or oriC in vitro replication (19), oriV replication is inhibited by the M7 anti-DnaA monoclonal antibody (Fig. 7). Using enzyme-linked immunosorbent assay and column fractionation, the studies with oriC showed that the M7 antibody specifically interferes with the interaction between DnaA and DnaB but does not interfere with the binding of the DnaA protein to DNA containing a DnaA box consensus sequence (19). The observation that helicase activity at the RK2 origin can be inhibited by M7 when added prior to the addition of the DnaB-DnaC complex provides additional support for the role of the DnaA protein in the delivery and loading of the DnaB helicase at the RK2 origin. Since the DnaA protein cannot produce an opening in oriV (8) but can recruit the helicase, it is likely that the helicase is delivered to oriV by protein-protein (DnaA-DnaB) and protein-dsDNA (DnaA-oriV) interactions.

The gel filtration results show that an oriV DnaA-DnaB-DnaC prepriming complex requires the addition of the TrfA protein for initiating replication activity (Fig. 3). An investigation of the kinetics of the replication reactions showed that preformation of this prepriming complex in the absence of TrfA resulted in a significantly shorter lag period for nucleotide incorporation upon addition of the TrfA protein when compared with reactions assembled on ice de novo. Using the FI^* assay for determining helicase activity, it is clear that activation of the DnaB helicase and consequently unwinding of oriV (formation of form FI^*) is dependent on the TrfA protein. As shown for the DnaA protein acting either at oriC (10, 12, 13) or the replication origin of the Pl plasmid (37) and the O protein acting at the  origin (38), TrfA by virtue of its binding to the 17-bp iterons at the RK2 origin destabilizes a discrete segment of the A + T-rich region and forms an open complex (8). It has been shown that the DnaB helicase in a complex with DnaC or P can be loaded onto ssDNA formed as a result of the destabilization of the E. coli and bacteriophage  origins, respectively (34, 39, 40). Recently it was demonstrated that both P and DnaC proteins contain a cryptic ssDNA binding activity which is mobilized when each forms a complex with the DnaB helicase (41). Interaction of P or DnaC with ssDNA may precede the transfer of helicase onto DNA. Interestingly O protein enhances interaction of the P-DnaB complex with ssDNA (41). Our results indicate that TrfA destabilization of the RK2 origin itself does not bring about loading of DnaB helicase but the presence of TrfA is necessary for activation of the helicase that is already delivered to the dsDNA by the DnaA protein (Fig. 9). Using both enzyme-linked immunosorbent assay and immunoprecipitation techniques no evidence was found for the formation of a specific DnaB-TrfA complex⁶ as has been shown for the R6K replication protein  and the DnaB protein (42).

[View Larger Version of this Image (11K GIF file)]

Our experiments showed that the DnaC protein was present in prepriming complexes formed with or without TrfA protein. Although DnaC is required for helicase recruitment it has an inhibitory effect on oriC replication in vitro (16, 43, 44). Six DnaC monomers (22) form a stable complex with the DnaB hexamer in the presence of ATP (45) but during in vitro replication with ssDNA template (ABC priming reaction) the DnaC protein is released and, therefore, is no longer present in the nucleoprotein complex (45, 46). In contrast, and as seen with double-stranded oriC DNA (19), in our studies the DnaC protein was found within the prepriming complex even after activation of the helicase. During RK2 replication initiation the TrfA dependent change in origin structure possibly repositions DnaC in such a way that it not longer inhibits the DnaB helicase. Alternatively the detection of DnaC protein in the void fractions together with oriC DNA (19) or oriV DNA (this work) may be an artifact caused by the presence of an inactive nucleoprotein complexes containing DnaC that is formed during incubation in vitro.

RK2 is a self-transmissible plasmid of the Inc-P1 incompatibility group which replicates and is stable maintained in a wide range of Gram-negative bacteria (47). This property is unlike that of most naturally occurring plasmid elements of Gram-negative bacteria which generally display a narrow host-range in that they are stable maintained in their natural host but fail to replicate in distantly related bacteria. Is there a unique aspect to the initiation of RK2 replication that may contribute to the ability of this plasmid to replicate in a widely diverse set of host cell backgrounds? While the present study was carried out with a minimal replication origin of RK2 that contains five instead of the eight iterons present in the complete origin region, some conclusions can be drawn. Our results show that the DnaA protein is the key factor that directs the helicase to the RK2 replication origin and this occurs in the absence of the TrfA protein. dnaA genes are ubiquitous among bacteria, and DnaA amino acid sequences are highly conserved (9, 48). Furthermore, on the basis of sequence similarities the DnaA protein has been divided into domains with varying degrees of homology (9). Very well conserved regions correspond to the ATP and DNA-binding domains of the protein (9). It has been shown that the DnaA protein of Pseudomonas putida can bind in vivo to the E. coli DnaA-box consensus sequence found in E. coli (49). These observations suggest that many if not all of the bacterial DnaA proteins of Gram-negative bacteria are capable of binding to one or more of the DNA binding consensus sequences localized within the RK2 origin and are able to deliver the specific host DnaB helicase to the RK2 origin. Evidence has been obtained that the essentiality of each of the DnaA boxes at the origin for DnaA binding may be host dependent (50). In contrast to a very conserved DNA-binding domain (9, 51), the region of the DnaA protein which is the target of the M7 monoclonal antibody and, therefore, which presumably interacts with the DnaB helicase appears to show only a low level of conservation (35, 51). This suggests that the DnaB-DnaA interaction required for helicase delivery to the oriC region of different bacteria or the RK2 oriV is host specific. However, once the DnaB helicase is delivered to oriV, the helicase activation step, triggered by the formation of a TrfA-DNA nucleoprotein structure, should be host nonspecific. On the basis of these observations, replication of RK2 in different hosts may depend on a host nonspecific interaction of the DnaA protein with one or more of the DnaA boxes at the RK2 ori region, a specific interaction between the host DnaA and DnaB proteins and a host nonspecific activation of the helicase by TrfA induced conformational changes at the RK2 origin. The specific structural organization of the RK2 origin and the unique properties of the plasmid-encoded TrfA protein are, therefore, likely to be key factors accounting for the broad host range replication properties of this plasmid.