We constructed a hybrid replication origin that consists of the main part of oriC from Escherichia coli, the DnaA box region and the AT‐rich region from Bacillus subtilis oriC. The AT‐rich region could be unwound by E. coli DnaA protein, and the DnaB helicase was loaded into the single‐stranded bubble. The results show that species specificity, i.e. which DnaA protein can do the unwinding, resides within the DnaA box region of oriC.
Bacterial initiator protein DnaA binds to iterated binding sites in replication origins, forming specialized nucleoprotein structures (Echols, 1990; Kornberg and Baker, 1992). The net result is the local unwinding of an AT‐rich region, which in the case of Escherichia coli consists of an AT cluster and three 13mer repeats (Bramhill and Kornberg, 1988). This is the crucial reaction in the initiation of replication. The unwound region is the entry site for the replicative helicase, which in E. coli is loaded as a double‐hexamer of the helicase DnaB and the loader protein DnaC (Fang et al., 1999). DnaA and interaction between DnaA and DnaB are required for the loading reaction (Marszalek et al., 1996; Seitz et al., 2000). DnaB helicase extends the initial bubble allowing primase to enter the complex (Fang et al., 1999). The resulting primer is extended by the DNA polymerase III holoenzyme (Kornberg and Baker, 1992).
We have shown recently that the mechanism of the unwinding reaction relies on a cooperative binding of the ATP‐complexed form of DnaA protein to special binding sites in the AT‐rich region of E. coli oriC, 6mer ATP–DnaA boxes, using the adjacent 9mer DnaA box as an anchor. ATP–DnaA then binds to the exposed single‐stranded ATP–DnaA boxes, thereby stabilizing the single‐stranded region (Messer et al., 2001; Speck and Messer, 2001).
The region corresponding to the E. coli AT‐rich 13mers in Bacillus subtilis oriC is a 27mer containing exclusively A and T residues, and AT‐rich sequences close to it. The two origins could only be unwound by the homologous DnaA protein (Krause et al., 1997). The regions of oriC that are unwound by DnaA protein in E. coli and B. subtilis are thus AT rich, but otherwise quite different in sequence. Alignment of both regions using the adjacent DnaA box reveals, however, that the spatial pattern of unwinding is very similar. In both cases the first unpaired nucleotide, measured by its sensitivity to oxidation by KMnO4, is 14 bp from the border of the DnaA box, R1 in the case of E. coli. The following 28 bp are unwound without SSB, and 16–18 additional base pairs in the presence of SSB, both in E. coli and B. subtilis oriC (Figure 1) (Krause and Messer, 1999). In addition, the ATP–DnaA boxes, which mediate cooperative binding of ATP–DnaA to the AT‐rich region, are present at corresponding positions in both origins (Figure 1) (Speck and Messer, 2001).
This similarity prompted us to construct a hybrid replication origin that contains the AT‐rich region from B. subtilis and the main part of oriC from E. coli, to the right of DnaA box R1 (Figure 1), and to analyze which initiation reactions can be observed with such an origin.
Structure of pOCBS
For the construction of such an E. coli–B. subtilis hybrid origin we used plasmid pOC180. It contains the oriC region from E. coli and the replication origin of pBR322 (Skarstad et al., 2000). The AT‐rich region from this plasmid, encompassing the AT cluster and the three 13mers (E. coli oriC coordinates 6–79) (Buhk and Messer, 1983), was replaced by the PCR‐amplified AT‐rich region from B. subtilis, positions 2216–2064 in Moriya et al. (1985) (Figure 1). The fragment from B. subtilis oriC is larger than the E. coli fragment in order to include the complete region of helical instability (DNA unwinding element) (Krause et al., 1997).
The hybrid origin is unwound by E. coli DnaA protein
Unwinding of the AT‐rich region was monitored using the reactivity of unpaired T residues to KMnO4 as described previously (Krause et al., 1997; Krause and Messer, 1999). Supercoiled pOCBS was incubated with E. coli protein HU and different concentrations of purified E. coli DnaA protein, and subjected to oxidation by KMnO4. The positions of modified bases were determined by primer extension using 5′‐32P‐end‐labelled primer A (see Methods), electrophoresis in a sequencing gel and detection in a PhosphorImager. A sequencing reaction with the same primer and pOCBS as template was run in parallel. The plasmid with the E. coli oriC, pOC180, served as a control.
With E. coli DnaA protein modified bases were observed with the hybrid origin from 14 to 41 bases upstream of the DnaA box, with signals of lower intensity up to 51 bases (Figure 2). This region of unwinding is identical to the unwound region observed with a plasmid containing a bona fide B. subtilis replication origin using B. subtilis DnaA protein (Krause et al., 1997). The hybrid origin thus shows a normal unwinding pattern. With pOCBS we did not observe any modified bases using purified DnaA protein from B. subtilis (data not shown).
DnaA can load helicase to the unwound hybrid origin
The next step after the unwinding reaction is the loading of helicase. Therefore we tested whether helicase can be loaded to the hybrid origin. Loading of DnaB helicase, as a DnaB–DnaC double‐hexamer, and the subsequent action of helicase can be monitored using the FI* test. DnaB helicase action on a negatively supercoiled plasmid in the presence of DNA gyrase changes the linking number of the plasmid. Such plasmids can be detected as a faster moving band (FI*) in agarose gels (Baker et al., 1987; Konieczny and Helinski, 1997).
Plasmids pOCBS and pOC180 were incubated with purified proteins from E. coli: DnaA, DnaBC, SSB, HU and DNA gyrase. The reaction was stopped with EDTA–SDS and subjected to agarose electrophoresis. With the complete mixture, a FI* band was observed with both pOC180 and pOCBS (Figure 3, lanes 5 and 11). This shows that helicase can be loaded to the hybrid origin. Omission of DnaA, DnaBC, SSB or DNA gyrase abolished the reaction (Figure 3, lanes 2–4). Without protein HU, there was only a small amount of a FI* band with pOC180 (Figure 3, lane 6), which was further reduced by increasing the DnaA concentration or adding IHF protein (data not shown). With pOCBS, nearly as much FI* band was observed with and without HU (Figure 3, lane 12). The results show that the hybrid origin cannot only be unwound by DnaA protein, but DnaA‐mediated loading of helicase is also possible.
The hybrid origin does not replicate in vivo
The ability of the hybrid replication origin to replicate in vivo was tested by transformation of a polA1 strain. No transformants were found in multiple experiments, whereas the same plasmid preparation resulted in >1000 transformants per plate with a polA+ strain. pOC180 as a control gave comparable numbers of transformants (>1000 per plate) with both strains. The results show that DnaA‐mediated unwinding and helicase loading are not sufficient to allow replication in vivo.
We show here that a hybrid replication origin, consisting of the DnaA box region from E. coli oriC and the AT‐rich region from B. subtilis oriC, can be unwound in vitro by E. coli DnaA protein. The unwound bases, measured by KMnO4 footprinting, cover an identical part in the AT‐rich region of the hybrid as in B. subtilis oriC. In addition, helicase, E. coli DnaBC protein can be loaded onto such a bubble substrate, mediated by DnaA protein (Seitz et al., 2000). DnaB helicase then extends the bubble (Fang et al., 1999). Since this is a pre‐requisite for the formation of the FI* structure, we can also conclude that this step occurs in the hybrid origin.
It has been shown previously that the unwinding reaction is species specific. Unwinding of E. coli oriC requires E. coli DnaA protein, likewise B. subtilis oriC is only unwound by B. subtilis DnaA (Krause et al., 1997). The observation that the hybrid origin is unwound by E. coli DnaA protein, and only by E. coli DnaA, suggests that the species specificity resides in the DnaA box part of the origin, probably in the spatial arrangement of DnaA boxes.
Although it is possible to replace the AT‐rich region of E. coli oriC by the corresponding region of another species, it cannot be completely random, and AT richness is not sufficient for its function (Speck and Messer, 2001). We have shown that several 6mer consensus sequences, ATP–DnaA boxes that are different from the 9mer consensus sequences of regular DnaA boxes, mediate the cooperative binding of the ATP‐complexed form of DnaA protein. Such ATP–DnaA boxes are present in the AT‐rich regions of both E. coli and B. subtilis origins in the same spatial arrangement (Figure 1). Binding of six or more ATP–DnaA protein molecules destabilizes the helix. ATP–DnaA protein then binds with high affinity to the exposed single‐stranded ATP–DnaA boxes (Messer et al., 2001; Speck and Messer, 2001), preparing them for the loading of helicase.
It is not clear why the hybrid origin is unable to replicate in vivo. One specific requirement for replication initiation in vivo is transcriptional activation (Lark, 1972; Messer, 1972). Crucial for transcriptional activation is the gidA promoter at the left boundary of oriC, and the spacing between gidA and oriC (Asai et al., 1990). The different spacing between gidA and the AT‐rich region in the hybrid origin is likely to be one reason for its inactivity in vivo.
Unwinding at the replication origin is the central reaction in the initiation of most origins. Most cellular origins, prokaryotic and eukaryotic, have AT‐rich regions that are unwound with the help of an initiator protein, DnaA in the case of bacteria. The analysis of the hybrid B. subtilis/E. coli replication origin allows us to define that the DnaA box region of oriC contains the species‐specific elements of this reaction.
Bacterial strains and plasmids.
Escherichia coli WM1963 [XL1‐blue] (endA1, gyrA46, hsdR17, lac, recA1, relA1, supE44, thi; F'lac: lacIQ, lacZΔM15, Tn10, proAB+) (Sambrook et al., 1989) and SCS1 [F−, endA1, hsdR17(rK−, mK+), supE44, thi‐1, recA1, gyrA96, relA1; Stratagene, La Jolla, CA] were used for cloning and in the polA tests as control. WM1933 (wild‐type except for polA1) was the strain used for the polA tests.
Plasmid pOC170 (Roth and Messer, 1995) contains the oriC region (coordinates −176 to +1497; coordinates refer to Buhk and Messer, 1983), the replication origin of pBR322 and the bla gene from pT7‐7. Plasmid pOC180 is pOC170 with 24 bp of the multicloning region deleted. Plasmid DNA was purified using Qiagen purification kits (Qiagen, Hilden, Germany).
Restriction enzymes, DNA polymerases and T4 DNA ligase were from New England Biolabs (Beverly, MA). Thermostable polymerase for PCR was from Promega (Madison, WI), DNA gyrase from Gibco BRL and E. coli SSB from Promega. DnaA protein was overproduced from expression vector pdnaA116 and purified as described previously (Schaper and Messer, 1995; Krause et al., 1997). DnaB and DnaC were simultaneously overproduced from plasmid pPS562 and the DnaB–DnaC complex separated from DnaB and DnaC by chromatography on a Mono Q‐column and gel filtration as described previously (San Martin et al., 1995).
In vitro KMnO4 footprinting was performed as described previously (Krause et al., 1997) with minor modifications. All analysis contained 1 μg oriC plasmid, pOC180 or pOCBS, and as assisting protein 100 ng of E. coli HU protein in 75 μl. Purified DnaA protein was added on ice. Open complex formation and subsequent KMnO4 modification was allowed for 2 min each as described previously (Krause et al., 1997). Primer extension was performed using a 5′‐32P‐labelled oligonucleotide complementary to the top strand of oriC (primer A, oriC coordinates 142–119: 5′‐CATTCACAGTTAATGATCCTTTCC) as primer. The samples were run on sequencing gels and analyzed by a PhosphorImager (Molecular Dynamics).
Reactions (25 μl) contained 40 mM HEPES–KOH pH 7.6, 11 mM Mg++‐acetate, 2 mM ATP, 1 mM DTT, 0.1 pmol plasmid DNA, 23 pmol SSB protein (440 ng), 2.1 pmol HU protein (40 ng) and 10 pmol DnaA (540 ng). After incubation for 5 min at 4°C and 5 min at 32°C, 0.8 pmol DnaB*C complex (410 ng) and 0.65 pmol DNA gyrase (125 ng) were added and incubation continued for an additional 30 min at 32°C. The reactions were stopped by adding 5 μl stop mix (10 mM EDTA, 2.5% SDS) followed by incubation at 65°C for 2 min. Twenty microlitres of the reaction was loaded onto a 0.8% agarose gel in 0.5× TBE buffer with 0.1 μg/ml ethidium bromide. The samples were electrophoresed for 20–24 h at a constant 1 V/cm, then stained with Vistra GreenTM and scanned.
This work was in part supported by grant Me659/6‐1 from the Deutsche Forschungsgemeinschaft.
- Copyright © 2001 European Molecular Biology Organization