Activation of distant replication origins in vivo by DNA looping as revealed by a novel mutant form of an initiator protein defective in cooperativity at a distance.

We have isolated mutants of the pi initiator protein of the plasmid R6K that are defective in DNA looping in vitro but retain their normal DNA binding affinity for the primary binding sites (iterons) at the gamma origin/enhancer. One such looping defective mutant called R6 was determined to be a proline to leucine change at position 46 near the N terminus of the pi protein. Using a set of genetic assays that discriminate between the activation of the gamma origin/enhancer from those of the distantly located alpha and beta origins, we show that the looping defective initiator protein fails to activate the alpha and beta origins but derepresses initiation from the normally silent gamma origin in vivo. The results conclusively prove that DNA looping is required to activate distant replication origins located at distances of up to 3 kb from the replication enhancer.     

Roles of a 106-bp origin enhancer and Escherichia coli DnaA protein in replication of plasmid R6K.

A dnaA 'null' strain could not support replication of intact plasmid R6K or derivatives containing combinations of its three replication origins (alpha, gamma, beta). DnaA binds in vitro to sites in two functionally distinct segments of the central gamma origin. The 277-bp core segment is common to all three origins and contains DnaA box 2, which cannot be deleted without preventing replication. Immediately to the left of the core lies the 106-bp origin enhancer, which contains DnaA box 1. When the origin enhancer is deleted, the core alone can still initiate replication if levels of wt pi protein are decreased or if copy-up pi mutant proteins are provided in trans. DnaA does not effect expression of R6K replication initiator protein pi, although several DnaA boxes were identified in the coding segment of the pir gene, which encodes pi. Together these data suggest that a single DnaA box, 2, is sufficient for initiation from the gamma origin and might be sufficient for initiation from the gamma origin and might be sufficient and required for the activity of the alpha and beta origins as well. Implications of the DnaA protein binding to two domains of the gamma origin and the role of the 106-bp origin enhancer in replication are discussed.     

Enhancer-origin interaction in plasmid R6K involves a DNA loop mediated by initiator protein

Initiation of DNA replication from ori beta of plasmid R6K requires the presence of the ori gamma sequence in cis. We demonstrate that binding of initiator protein to the seven strong, tandem binding sites in gamma increases binding of the protein at the very weak binding site present in ori beta by cooperativity at a distance. The gamma-beta interaction via the initiator results in a DNA loop, as revealed by the novel technique of cyclization enhancement and as confirmed by exonuclease III protection, electron microscopy, and chemical footprinting. The protein-mediated gamma-beta interaction in vitro suggests that the cooperative interaction of gamma-bound protein with the beta sequence by DNA looping is an early step in the initiation of DNA replication at the beta origin of R6K.     

Binding of DnaA protein to a replication enhancer counteracts the inhibition of plasmid R6K gamma origin replication mediated by elevated levels of R6K pi protein.

Replication of the gamma origin of Escherichia coli plasmid R6K requires pi protein, encoded by the R6K pir gene, and many host factors, including DnaA protein. Pi has dual roles, activating replication at low levels and inhibiting replication at high levels. The inhibitory function of pi is counteracted by integration host factor and a specific sequence of the origin called the enhancer. This 106-bp DNA segment contains a binding site for DnaA protein (DnaA box 1). In this study, we mutated this site to determine if it was required for the enhancer's function. Using gamma origin derivative plasmids with the DnaA box 1 altered or deleted, we show that this site is necessary to protect the origin against levels of wild-type pi protein that would otherwise inhibit replication. To show that the base substitutions in DnaA box 1 weakened the binding of DnaA, we developed a new application of the agarose gel retardation assay. This quick and easy assay has broad applicability, as shown in binding studies with DNA fragments carrying a different segment of the R6K origin, the chromosomal origin (oriC), or the pUC origin. The gel retardation assay suggests a stoichiometry of DnaA binding different from that deduced from other assays.     

Conformational changes in a replication origin induced by an initiator protein

The replication initiator protein of the plasmid R6K binds to seven contiguous 22 bp direct repeats that form an indispensable part of the three replication origins alpha, beta, and gamma. Binding of the initiator to the direct repeats induced a marked bending of the region of gamma replication origin. Binding of the initiator also promoted unwinding of the origin DNA by at least two turns. Distamycin appeared to antagonize the binding of the initiator to the seven 22 bp direct repeats. At the appropriate DNA and protein concentrations the initiator enhanced topoisomerase-induced catenation of the origin containing supercoiled DNA but not of DNA lacking the origin sequence. Thus, the initiator protein caused significant changes in the secondary and tertiary structures of the replication origin.     

Bending of the origin of replication of E. coli by binding of IHF at a specific site

In studies of DNA replication in Escherichia coli, an important question concerns the role of the initiator protein DnaA. This protein is known to bind to a specific 9-bp sequence in the origin of replication, but it is not understood how it can recognize another, relatively distant, 13-bp sequence that has no homology to the binding site but is where the DnaA protein serves its catalytic function in the initiation of DNA replication. This effect of DnaA might be achieved by bending of DNA in this region. I have searched for putative binding sites for integration host factor (IHF), a protein known to bend DNA. Here I report the finding of an IHF binding site in the E. coli origin and present direct evidence that IHF binds and causes DNA bending in this region. On the basis of these results I propose a model wherein formation of a higher-order nucleoprotein structure would facilitate the action of DnaA protein in the initiation events.     

Interaction of the initiator protein of an IncB plasmid with its origin of DNA replication

The replication initiator protein RepA of the IncB plasmid pMU720 was purified and used in DNase I protection assays in vitro. RepA protected a 68-bp region of the origin of replication of pMU720. This region, which lies immediately downstream of the DnaA box, contains four copies of the sequence motif 5'AANCNGCAA3'. Mutational analyses identified this sequence as the binding site specifically recognized by RepA (the RepA box). Binding of RepA to the RepA boxes was ordered and sequential, with the box closest to the DnaA binding site (box 1) occupied first and the most distant boxes (boxes 3 and 4) occupied last. However, only boxes 1, 2, and 4 were essential for origin activity, with box 3 playing a lesser role. Changing the spacing between box 1 and the other three boxes affected binding of RepA in vitro and origin activity in vivo, indicating that the RepA molecules bound to ori(B) interact with one another.     

Interaction of initiator proteins with the origin of replication of an IncL/M plasmid

The origin of replication of the IncL/M plasmid pMU604 was analyzed to identify sequences important for binding of initiator proteins and origin activity. A thrice repeated sequence motif 5'-NANCYGCAA-3' was identified as the binding site (RepA box) of the initiator protein, RepA. All three copies of the RepA box were required for in vivo activity and binding of RepA to these boxes appeared to be cooperative. A DnaA R box (box 1), located immediately upstream of the RepA boxes, was not required for recruitment of DnaA during initiation of replication by RepA of pMU604 unless a DnaA R box located at the distal end of the origin (box 3) had been inactivated. However, DnaA R box 1 was important for recruitment of DnaA to the origin of replication of pMU604 when the initiator RepA was that from a distantly related plasmid, pMU720. A mutation which scrambled DnaA R boxes 1 and 3 and one which scrambled DnaA R boxes 1, 3 and 4 had much more deleterious effects on initiation by RepA of pMU720 than on initiation by RepA of pMU604. Neither Rep protein could initiate replication from the origin of pMU604 in the absence of DnaA, suggesting that the difference between them might lie in the mechanism of recruitment of DnaA to this origin. DnaA protein enhanced the binding and origin unwinding activities of RepA of pMU604, but appeared unable to bind to a linear DNA fragment bearing the origin of replication of pMU604 in the absence of other proteins.     

A DNA segment conferring stable maintenance on R6K gamma-origin core replicons.

The plasmid R6K gamma origin consists of two adjacent modules, the enhancer and the core, and requires R6K initiator protein pi for replication. While the core alone can replicate at a low level of wild-type pi protein, we show here that host cells do not stably maintain core plasmids. The presence of the enhancer segment confers stable inheritance on core plasmids without a significant change in average plasmid copy number. Deletions and site-directed mutagenesis indicated that the stability of core plasmids is not mediated by binding sites or consensus sequences in the enhancer for DnaA, pi protein, gyrase, Fis, or Dcm methylase. Proper segregation of core plasmids requires only the R6K stb or stability-related region, which includes the 20-bp segment of the 100-bp enhancer adjacent to the core. The use of the pi 116 mutant protein, which increases plasmid copy number fourfold, does not stabilize core plasmids lacking the enhancer. We also show that at an elevated level of wild-type pi, the gamma-origin plasmid is unstable, even in the presence of the enhancer. We discuss the differences and similarities between the R6K stability system and those found in other plasmids.     

Origin Remodeling and Opening in Bacteria Rely on Distinct Assembly States of the DnaA Initiator

The initiation of DNA replication requires the melting of chromosomal origins to provide a template for replisomal polymerases. In bacteria, the DnaA initiator plays a key role in this process, forming a large nucleoprotein complex that opens DNA through a complex and poorly understood mechanism. Using structure-guided mutagenesis, biochemical, and genetic approaches, we establish an unexpected link between the duplex DNA-binding domain of DnaA and the ability of the protein to both self-assemble and engage single-stranded DNA in an ATP-dependent manner. Intersubunit cross-talk between this domain and the DnaA ATPase region regulates this link and is required for both origin unwinding in vitro and initiator function in vivo. These findings indicate that DnaA utilizes at least two different oligomeric conformations for engaging single- and double-stranded DNA, and that these states play distinct roles in controlling the progression of initiation.     

Conformational changes induced by integration host factor at origin gamma of R6K and copy number control.

We have investigated the role of integration host factor (IHF) in the replication of plasmid R6K by studying the maintainance of the plasmid in a strain of Escherichia coli that lacks both subunits of IHF and in an isogenic wild type strain and found that all three origins, alpha, beta, and gamma, were functional in the absence of IHF; however, loss of IHF reduced the copy number of those replicons initiating solely from ori gamma by 5-fold. Concomitant loss of direct repeats within the origin that bind the R6K replication initiator protein, Pi, resulted in a further reduction in copy number. Using gel mobility shift analysis, we showed that IHF bound specifically only to one site within the A/T rich region of the minimal origin adjacent to the Pi binding sites. The origin region possessed no intrinsic DNA curvature although IHF induced a strong bend upon binding. Combination footprinting with different orders of addition of Pi and IHF suggested that there was no cooperativity between the two proteins with regard to DNA binding. Hydroxyl-radical footprinting revealed hypersensitive asymmetric periodic cleavage sites within the origin region in the presence of IHF that extended over 200 base pairs and a localized perturbation of cleavage chemistry. The presence of periodic cleavages was dependent upon the presence of the wild type R6K origin sequence and was not observed when the IHF binding site was positioned adjacent to a heterologous sequence. We observed that the conformational changes induced by IHF upon binding to the R6K origin were negatively correlated with the observed decrease in copy number, and therefore, origin conformation altered by protein-DNA interaction may play an important role in the regulation of replication initiation.     

A comprehensive set of DnaA-box mutations in the replication origin, oriC, of Escherichia coli

We probed the complex between the replication origin, oriC , and the initiator protein DnaA using different types of mutations in the five binding sites for DnaA, DnaA boxes R1–R4 and M: (i) point mutations in individual DnaA boxes and combinations of them; (ii) replacement of the DnaA boxes by a scrambled 9 bp non-box motif; (iii) positional exchange; and (iv) inversion of the DnaA boxes. For each of the five DnaA boxes we found at least one type of mutation that resulted in a phenotype. This demonstrates that all DnaA boxes in oriC have a function in the initiation process. Most mutants with point mutations retained some origin activity, and the in vitro DnaA-binding capacity of these origins correlated well with their replication proficiency. Inversion or scrambling of DnaA boxes R1 or M inactivated oriC -dependent replication of joint replicons or minichromosomes under all conditions, demonstrating the importance of these sites. In contrast, mutants with inverted or scrambled DnaA boxes R2 or R4 could not replicate in wild-type hosts but gave transformants in host strains with deleted or compromised chromosomal oriC at elevated DnaA concentrations. We conclude that these origins require more DnaA per origin for initiation than does wild-type oriC . Mutants in DnaA box R3 behaved essentially like wild-type oriC , except for those in which the low-affinity box R3 was replaced by the high-affinity box R1. Apparently, initiation is possible without DnaA binding to box R3, but high-affinity DnaA binding to DnaA box R3 upsets the regulation. Taken together, these results demonstrate that there are finely tuned DnaA binding requirements for each of the individual DnaA boxes for optimal build-up of the initiation complex and replication initiation in vivo     

Mechanistic studies of initiator-initiator interaction and replication initiation.

Unlike the chromosome of Escherichia coli that needs only one replication initiator protein (origin recognition protein) called DnaA, many plasmid replicons require dual initiators: host-encoded DnaA and a plasmid-encoded origin recognition protein, which is believed to be the major determinant of replication control. Hitherto, the relative mechanistic roles of dual initiators in DNA replication were unclear. Here, we present the first evidence that DnaA communicates with the plasmid-encoded pi initiator of R6K and contacts the latter at a specific N-terminal region. Without this specific contact, productive unwinding of plasmid ori gamma and replication is abrogated. The results also show that DnaA performs different roles in host and plasmid replication as revealed by the finding that the ATP-activated form of DnaA, while indispensable for oriC replication, was not required for R6K replication. We have analyzed the accessory role of the DNA bending protein, integration host factor (IHF), in promoting initiator-origin interaction and have found that IHF significantly enhances the binding of DnaA to its cognate site. Collectively, the results further advance our understanding of replication initiation.     

Structural properties of the beta origin of replication of plasmid R6K

The beta origin of replication of plasmid R6K, one of three active R6K origins of replication, requires most or all of a 1962-base pair (bp) sequence for activity. The nucleotide sequence of a portion of this functional beta origin was determined in an earlier study (Stalker, D., Kolter, R., and Helinski, D. (1982) J. Mol. Biol. 161, 33-43). In this work, the sequence of the remaining portion of this 1964-bp segment was obtained. In addition to its activity as an origin of replication, this sequence also contains sufficient information for autonomous replication in Escherichia coli. A 277-bp region containing seven 22-bp direct repeats is present at one end of the beta origin segment (Stalker, D., Kolter, R., and Helinski, D. (1979) Proc. Natl. Acad. Sci. U. S. A. 76, 1150-1154) while the other end contains a 140-bp sequence that includes a relaxation complex site. The 277-bp direct repeat region is required for activity of the beta origin. The start of the beta origin of replication as mapped by electron microscopy (Crosa, J. (1980) J. Biol. Chem. 255, 11075-11077) lies approximately 1000 bp away from the 277-bp region. The pi structural gene, which makes up most of the sequence between the direct repeats and the beta origin, is required in cis for beta origin activity. The pi protein also is required for beta origin activity but can be provided in trans. The nucleotide sequence just beyond the pi structural gene and within or near the start of beta origin of replication contains an open reading frame for a 151-amino acid protein. Deletions ranging from 94 bp to 1590 bp were obtained within the 1964-bp beta origin region. In every case, the deletion results in loss of origin activity even when the deleted sequence plus adjacent regions are provided in trans. These observations suggest a requirement for a specific secondary structure over an extensive region for beta origin activity.     

Role of RepA and DnaA proteins in the opening of the origin of DNA replication of an IncB plasmid

The replication initiator protein RepA of the IncB plasmid pMU720 was shown to induce localized unwinding of its cognate origin of replication in vitro. DnaA, the initiator protein of Escherichia coli, was unable to induce localized unwinding of this origin of replication on its own but enhanced the opening generated by RepA. The opened region lies immediately downstream of the last of the three binding sites for RepA (RepA boxes) and covers one turn of DNA helix. A 6-mer sequence, 5'-TCTTAA-3', which lies within the opened region, was essential for the localized unwinding of the origin in vitro and origin activity in vivo. In addition, efficient unwinding of the origin of replication of pMU720 in vitro required the native positioning of the binding sites for the initiator proteins. Interestingly, binding of RepA to RepA box 1, which is essential for origin activity, was not required for the localized opening of the origin in vitro.     

Deletion analysis of the mini-P1 plasmid origin of replication and the role of Escherichia coli DnaA protein

The mini-P1 plasmid origin of replication is contained on a 246 base pair (bp) piece of DNA. At one end there are five 19-bp binding sites for the P1 initiator protein, RepA, and near the other end there are two 9-bp DnaA protein-binding sites. To further define the limits of the origin, we cloned the origin region in M13 and constructed deletions of either end. We sequenced the DNA and tested the replicative form I DNA of the deletion phages for their ability to support RepA-dependent DNA replication in an in vitro system. The origin that is functional in vitro could be reduced to 202 bp. It includes three intact and one incomplete RepA-binding sites at one end and the two DnaA-binding sites at the other end. When the two naturally occurring DnaA-binding sites were replaced with one or two synthetic sites, only the construction containing two sites was active in vitro. We found that the minimal origin that is functional in vivo contains all of the five RepA and the two DnaA-binding sites. Mini-P1 plasmid replication both in vivo and in vitro requires two initiator proteins, the Escherichia coli DnaA protein and the P1 RepA protein. We have found that the ADP form of DnaA is as active as the ATP form of the protein in the in vitro replication of mini-P1. In contrast, only the ATP form is active for in vitro replication of plasmids carrying the E. coli origin (Bramhill, D., and Kornberg, A. (1988) Cell 52, 743-755).     

DNA replication in Escherichia coli is initiated by membrane detachment of oriC. A model

An adequate model for the initiation of chromosome replication in Escherichia coli should explain why the introduction of multiple copies of the chromosomal origin of replication, oriC, does not perturb cells seriously and why such multiple origins are replicated synchronously; it should explain why the key initiator protein, DnaA, is activated in vitro by binding specifically to acidic phospholipids and why the Dam methyltransferase is essential for the correct timing of initiation; it should explain why phospholipid synthesis and fluidity are necessary for initiation. In the detachment model, presented here, cyclical changes in the phospholipid composition of the cytoplasmic membrane activate initiator proteins such as DnaA protein and cause origins to detach; this detachment allows torsional stresses to open 13mer sequences in oriC; DnaA assists in the serial opening of these sequences and guides the entry of the helicase to form a pre-priming complex and trigger initiation; the greater affinity of hemi-methylated origin for membrane is re-interpreted as a mechanism for preventing re-initiation.     

Integration host factor of Escherichia coli reverses the inhibition of R6K plasmid replication by pi initiator protein.

Integration host factor (IHF) protein is the only host-encoded protein known to bind and to affect replication of the gamma origin of Escherichia coli plasmid R6K. We examined the ability of R6K origins to replicate in cells lacking either of the two subunits of IHF. As shown previously, the gamma origin cannot replicate in IHF-deficient cells. However, this inability to replicate was relieved under the following conditions: underproduction of the wild-type pi replication protein of R6K or production of normal levels of mutant pi proteins which exhibit relaxed replication control. The copy number of plasmids containing the primary R6K origins (alpha and beta) is substantially reduced in IHF-deficient bacteria. Furthermore, replication of these plasmids is completely inhibited if the IHF-deficient strains contain a helper plasmid producing additional wild-type pi protein. IHF protein has previously been shown to bind to two sites within the gamma origin. These sites flank a central repeat segment which binds pi protein. We propose a model in which IHF binding to its sites reduces the replication inhibitor activity of pi protein at all three R6K origins.     

The integration host factor of Escherichia coli binds to multiple sites at plasmid R6K gamma origin and is essential for replication.

Examination of the effect of the himA and himD mutants of E. coli on the maintenance of plasmid R6K has revealed that the gamma origin-containing replicons cannot be established in any of the mutants deficient in the production of E. coli Integration Host Factor (IHF). Contrary, the R6K derivatives containing other origins of the plasmid (alpha and/or beta) replicate in a host lacking functional IHF protein. We show that IHF protein binds specifically to a segment of the replication region which is essential for the activity of all three R6K origins. Mapping the IHF binding sequence with neocarzinostatin showed that the protein protects three segments of the origin: two strong binding sites reside within an AT-rich block, while the third, considerably weaker site is separated from the other two by a cluster of the seven 22 bp direct repeats. These seven repeats have been shown previously to bind the R6K-encoded initiator protein pi. We also demonstrate that the establishment of pi-origin complexes prior to IHF addition prevents the binding of the IHF protein to the gamma origin. The binding sequences of IHF and pi proteins do not overlap, therefore, we propose that the binding of pi protein alters the structure of the DNA and thereby prevents the subsequent binding of IHF protein.