Letter | Published:

The bacterial DnaA-trio replication origin element specifies single-stranded DNA initiator binding

Nature volume 534, pages 412416 (16 June 2016) | Download Citation

  • An Erratum to this article was published on 29 June 2016


DNA replication is tightly controlled to ensure accurate inheritance of genetic information. In all organisms, initiator proteins possessing AAA+ (ATPases associated with various cellular activities) domains bind replication origins to license new rounds of DNA synthesis1. In bacteria the master initiator protein, DnaA, is highly conserved and has two crucial DNA binding activities2. DnaA monomers recognize the replication origin (oriC) by binding double-stranded DNA sequences (DnaA-boxes); subsequently, DnaA filaments assemble and promote duplex unwinding by engaging and stretching a single DNA strand3,4,5. While the specificity for duplex DnaA-boxes by DnaA has been appreciated for over 30 years, the sequence specificity for single-strand DNA binding has remained unknown. Here we identify a new indispensable bacterial replication origin element composed of a repeating trinucleotide motif that we term the DnaA-trio. We show that the function of the DnaA-trio is to stabilize DnaA filaments on a single DNA strand, thus providing essential precision to this binding mechanism. Bioinformatic analysis detects DnaA-trios in replication origins throughout the bacterial kingdom, indicating that this element is part of the core oriC structure. The discovery and characterization of the novel DnaA-trio extends our fundamental understanding of bacterial DNA replication initiation, and because of the conserved structure of AAA+ initiator proteins these findings raise the possibility of specific recognition motifs within replication origins of higher organisms.


The master bacterial DNA replication initiator, DnaA, is a highly conserved multifunctional protein that utilizes distinct domains to achieve its two key DNA binding activities. DnaA recognizes double-stranded (ds)DNA using a helix–turn–helix motif (domain IV), whereas an ATP-dependent DnaA filament interacts with a single DNA strand using residues within the initiator specific motif (ISM; an α-helical insertion that distinguishes the family of replication initiators) of the AAA+ domain (domain III) (Extended Data Fig. 1a–d)4,5. In contrast to DnaA, bacterial replication origins are diverse; they contain variable numbers of DnaA-boxes and seemingly lack a common architecture6,7. Therefore, the sequence information within oriC that directs DnaA filament assembly onto a single DNA strand is unknown.

To investigate how DnaA filament formation could be localized to the DNA replication origin of Bacillus subtilis, we began by characterizing site-directed mutants of the DNA unwinding region in vivo (Fig. 1a and Extended Data Fig. 1e). To enable identification of essential sequences without selecting for suppressor mutations, we generated a strain in which DNA replication could initiate from a plasmid origin (oriN) integrated into the chromosome (Fig. 1b and Supplementary Information). Activity of oriN requires its cognate initiator protein, RepN; both of these factors act independently of oriC/DnaA8. Expression of repN was placed under the control of a tightly regulated inducible promoter, thus permitting both the introduction of mutations into oriC and their subsequent analysis after removal of the inducer to shut off oriN activity (Fig. 1c and Extended Data Fig. 2).

Figure 1: Genetic analysis of the oriC DNA unwinding element reveals a critical region required for initiation activity.
Figure 1

a, B. subtilis oriC unwinding region. DnaA-box colouring indicates conservation (consensus 5′-TTATCCACA-3′ in black). b, The oriC-independent strain used for constructing replication origin mutations. c, Growth of an oriC deletion mutant is dependent upon oriN activity. d, Deletions extending beyond the AT-cluster into the initially unwound region inhibit cell growth. e, Sequences between the GC-rich and AT-rich clusters are essential for origin function. f, Sequences proximal to DnaA-boxes are most important for origin function. Marker frequency analysis was used to measure the rate of DNA replication initiation (mean and s.d. of three technical replicates). g, Open complex formation by DnaA requires the native sequence between the GC- and AT-clusters. DNA duplex unwinding was probed by KMnO4 and detected by primer extension.

At the B. subtilis replication origin, DNA unwinding by DnaA is detected downstream of DnaA-box elements and includes a sequence of 27 continuous A:T base pairs that is thought to facilitate DNA duplex opening (Fig. 1a)9. Surprisingly, we were able to delete the entire AT-rich sequence (Δ27) without abolishing origin activity, although the mutant strain did display a slow growth phenotype indicating that the AT-cluster is required for efficient origin function (Fig. 1d). Interestingly, further deletions extending three or six base pairs (Δ30, Δ33) severely impaired oriC-dependent initiation (Fig. 1d), and a deletion series targeting the sequence between the GC-rich and AT-rich clusters confirmed that this region alone was essential for origin function (Fig. 1e). Scrambling this entire region also inhibited cell growth (t1–t6Scr), demonstrating that the specific sequence is required, rather than the spacing between the flanking elements (Fig. 1f). To explore this region in more detail, sequences were scrambled three base pairs at a time by exchanging each triplet for its complement. Phenotypic and marker frequency analyses revealed that disruption of sequences closest to the GC-cluster (t1Scr and t2Scr) caused the greatest defect in DNA replication initiation, indicating that the region proximal to the DnaA-boxes is most important for origin activity (Fig. 1f). Although mutagenesis of neither t4 nor t5 alone produced a detectable effect on DNA replication initiation under the conditions tested, they may become important when origin firing is suboptimal as was observed in the AT-cluster deletion mutant (Fig. 1d).

To determine whether this essential DNA sequence between the GC- and AT-clusters has a role in DNA melting per se, an open complex formation assay was performed. DnaA was incubated with oriC plasmids containing either the wild-type or scrambled sequence (t1–t6Scr), potassium permanganate was added to oxidize distorted bases within the DNA, and base modification was detected by primer extension. Scrambling the sequence inhibited open complex formation, indicating that this region is necessary for DnaA-dependent unwinding (Fig. 1g).

DnaA monomers are thought to bind DnaA-boxes before ATP-dependent filament formation10. Using the strain capable of oriC-independent initiation, the seven DnaA-box sequences were individually scrambled to abolish DnaA binding11. Culturing these strains in the absence of oriN activity revealed that mutation of DnaA-box6 severely inhibited growth, and mutation of DnaA-box7 resulted in a marked growth defect, while mutation of the remaining DnaA-boxes had no observable effect (Fig. 2a). Marker frequency analysis confirmed that mutation of DnaA-box7 markedly impaired origin activity, whereas mutation of the remaining DnaA-boxes resulted in only modest decreases in initiation frequency (Fig. 2a). These results indicate that DnaA-boxes proximal to the essential unwinding region are most critical for origin activity.

Figure 2: DnaA filaments are loaded from DnaA-boxes onto a specific single-strand sequence within the initially unwound region.
Figure 2

a, DnaA-boxes proximal to the unwinding region are most important for origin function. Marker frequency analysis was used to measure the rate of DNA replication initiation (mean and s.d. of three technical replicates). b, Sequence of the origin region used for constructing DNA scaffolds in c. c, DnaA filament formation using cysteine-specific crosslinking on DNA scaffolds. DnaA complexes were resolved by SDS–polyacrylamide gel electrophoresis (SDS–PAGE) and detected by western blot analysis. d, Crystal structure showing ssDNA (dA12) bound to the DnaA filament through the AAA+ domain (Protein Data Bank (PDB) accession number 3R8F). e, Sequence of the origin region used for constructing DNA scaffolds in f. f, DnaA filament formation on tailed substrates is arrested by a poly(A) tract. Long oligomers highlighted within the dotted box are shown above with increased contrast.

To directly test whether these DnaA-boxes promote DnaA filament assembly at the essential unwinding region we used a previously described DnaA filament formation assay12. Here two cysteine residues are introduced within the AAA+ domain such that the protein remains functional and when the DnaA filament assembles the cysteine residues from interacting protomers come into close proximity. DNA scaffolds were assembled using oligonucleotides, and the cysteine-specific crosslinker bis(maleimido)ethane (BMOE; 8 Å spacer arm) was used to capture the oligomeric species formed on each substrate.

Incubation of DnaA with duplex substrates containing DnaA-box6, DnaA-box7 and the GC-rich region produced a dimeric species (Fig. 2b, c), whereas incubation of DnaA with a longer duplex substrate containing the unwinding region produced a set of larger oligomeric complexes. We wondered whether the larger species were being formed on the duplex DNA or on a single DNA strand. To test these models scaffolds containing single-stranded (ss)DNA tails were assembled. DnaA filaments readily formed on substrates containing a 5′-tail but were absent when the corresponding 3′-tail was provided (Fig. 2c and Extended Data Fig. 3). Formation of DnaA oligomers on the 5′-tailed substrate was dependent upon both ATP and the ssDNA binding residue Ile190 located within the ISM of the AAA+ domain4, indicating that the assay was capturing DnaA filament formation on ssDNA (Fig. 2c, d and Extended Data Fig. 1d).

Critically, DnaA oligomer formation on the 5′-tailed substrate was specific. DnaA filament assembly was abolished when the DnaA-box sequences within the duplex region were scrambled and it was notably reduced when the single-stranded region was replaced with its complementary sequence (Fig. 2c). Taken together, these results suggest that DnaA filaments are loaded from duplex DnaA-boxes onto ssDNA bearing a 5′-tail. This model is consistent both with biochemical experiments showing that Escherichia coli DnaA preferentially interacts with the corresponding single-strand of its DNA unwinding element and with single molecule studies showing that Aquifex aeolicus DnaA filaments form with 3′→5′ polarity11,13,14.

DnaA oligomer size was proportional to the length of the 5′-tail up to the formation of a heptamer, after which further DNA extension did not promote longer filaments (Fig. 2e, f). We noted that this limit corresponded to a poly(A) tract in the DNA sequence and wondered whether this sequence inhibited DnaA filament formation. When the poly(A) tract was replaced by sequences from the beginning of the DNA unwinding region, DnaA oligomer length increased beyond a heptamer (Fig. 2e, f). This result suggests that the origin unwinding region is designed to limit DnaA filament formation to a precise position within oriC.

To identify a possible single-strand binding motif recognized by DnaA, individual base pairs within the essential unwinding region were inverted and origin activity was analysed in vivo. Marker frequency analysis revealed that altering either of two A:T base pairs, which were spaced three nucleotides apart from each other, resulted in the most significant loss of origin activity; in contrast the surrounding mutations had only modest effects (Fig. 3a). Re-examination of the unwinding region shows that A:T base pairs are spaced at three nucleotide intervals throughout this sequence (Fig. 1a). This observation is strikingly congruent with the mechanism proposed for binding of the DnaA filament to ssDNA, where each protomer engages a set of three nucleotides (Fig. 3b)4.

Figure 3: Analysis of the key origin unwinding region provides evidence for functional trinucleotide repeats.
Figure 3

a, Mutagenesis identifies A:T base pairs spaced three nucleotides apart are most critical for origin activity. Marker frequency analysis was used to measure the rate of DNA replication initiation (mean and s.d. of three technical replicates). b, Crystal structure showing the interaction of DnaA with sets of three nucleotides (PDB accession number 3R8F). Residues for A. aeolicus indicated above; B. subtilis below. c, In vivo deletion analysis of the unwinding region. Isogenic deletions indicated in black. Marker frequency analysis was used to measure the rate of DNA replication initiation (mean and s.d. of three technical replicates). d, Growth of mutants used in c. e, In vitro deletion analysis of tailed substrates.

We hypothesized that an array of triplet nucleotide motifs recognized by DnaA are present within the unwinding region and that the motifs proximal to the DnaA-boxes are most important for origin activity. To test this model in vivo we created a set of nested deletions that removed either one or three base pairs (Extended Data Fig. 4). All of the single base-pair deletions significantly lowered the replication initiation frequency and several considerably inhibited cell growth (especially the same A:T base pairs noted above), whereas triplet deletions encompassing the single deletions had little or no effect (Fig. 3c, d). These results are consistent with the model that single base-pair deletions act both by disrupting a specific trinucleotide motif and by shifting the register of downstream trinucleotide motifs relative to the DnaA filament start point at the DnaA-boxes.

To test the model that the ssDnaA binding motif is indeed a repeating trinucleotide, DnaA filament formation was analysed in vitro using tailed substrates that contained either single or triplet base deletions (Fig. 3e). Whereas deletion of one base produced shorter oligomers, deletion of three bases restored formation of full-length complexes. Taken together with the in vivo deletions, these results indicate that DnaA filaments bind to ssDNA by recognizing a specific trinucleotide motif found within the unwinding region. We have termed this trinucleotide motif the ‘DnaA-trio’.

To define the precise sequence of the DnaA-trio, DnaA filament formation was observed using a series of DNA scaffolds in which the 5′-tails were extended by increments of one nucleotide. We observed that additional oligomeric species appeared after the following sequences were added: 3′-GAT-5′, 3′-AAT-5′ and 3′-GAA-5′, suggesting that these triplets represent individual DnaA-trio motifs (Fig. 4a).

Figure 4: Identification of the DnaA-trio motif.
Figure 4

a, Varying the length of 5′-tailed substrates identifies the likely DnaA-trio sequence. Lane 2 shows DnaA filament formation on a duplex DNA scaffold (DnaA-box6, DnaA-box7, GC-rich cluster). Letters indicate the nucleotide sequentially added to the 5′-tail. b, Targeted mutagenesis of the proposed DnaA-trio motif. c, Bioinformatic analysis identifies DnaA-trio motifs adjacent to a DnaA-box throughout the bacterial kingdom. Underlined sequences indicate experimentally determined DnaA-dependent unwinding sites. d, DnaA-trios sequence logo (WebLogo28). e, Schematic of DnaA filament formation from double-stranded DnaA-boxes (triangles) onto a single strand containing the DnaA-trios.

However, it was surprising that a longer oligomer was not formed after addition of the first 3′-GAT-5′ motif proximal to the GC-cluster, since mutagenesis of this sequence in vivo resulted in strong phenotypes (Fig. 3a, c, d). In structures of the archaeal initiator Orc1 bound to a replication origin the protein was observed to make two contacts with the DNA, one through its carboxy (C)-terminal DNA binding domain (analogous to DnaA domain IV) and another through its AAA+ motif15,16. We wondered whether DnaA might similarly be capable of contacting both a DnaA-box and the first DnaA-trio, thereby accounting for the absence of a DnaA trimer. Importantly, BMOE crosslinking of cysteines in the AAA+ domain would not detect this activity as the assay captures DnaA oligomers formed on either dsDNA or ssDNA12.

To test this hypothesis, we used the amine-specific crosslinker bis(sulfosuccinimidyl)suberate (BS3) which, in contrast to BMOE, only captures DnaA oligomers formed on a single DNA strand (Extended Data Figs 3 and 5). Crosslinking by BS3 reveals a DnaA dimer forming in the presence of the first 3′-GAT-5′, indicating that DnaA does recognize this sequence (Extended Data Fig. 5). Taken together with the BMOE crosslinking showing that a DnaA dimer is formed on the dsDNA scaffold containing just DnaA-boxes 6 and 7 and the GC-cluster, the data suggest that the DnaA protein initially bound at DnaA-box7 undergoes a conformational change (detected by BS3) to engage the first DnaA-trio motif following the GC-cluster. Several lines of evidence support the notion that DnaA adopts distinct conformations when it engages either dsDNA or ssDNA17,18.

To support the assignment of the DnaA-trio, we performed a targeted mutagenesis of the proposed sequence. The results indicate that each of the positions (3′-GAT-5′) appears important for DnaA filament formation, specifically the nucleotides at positions 1 and 2, and the deoxyribose group at position 3 (Fig. 4b and Extended Data Fig. 6). Interestingly, in the crystal structure of DnaA bound to a ssDNA substrate, the protein makes no base-specific contacts4. These observations suggest either that the sequence of the DnaA-trios is important for an intermediate step in DNA duplex recognition and melting before full engagement of the product single-strand, or that the specific base sequence promotes the DNA backbone to adopt a favourable geometry for DnaA binding.

Using this information, we first searched for DnaA-trios within other well-characterized origin unwinding elements (Fig. 4c, underlined)19,20,21. In these cases a set of at least three DnaA-trios could be identified. These DnaA-trios were located proximal to a DnaA-box that shared the same orientation as B. subtilis DnaA-box7, and the regions between the DnaA-box and the DnaA-trios were GC-rich. Using these additional criteria we next interrogated predicted bacterial DNA replication origins (DoriC22) for similar patterns. Figure 4c shows that similar elements can be identified within putative oriC regions throughout the bacterial kingdom. A sequence logo of the DnaA-trios indicates that the preferred motif is 3′-G/AAT-5′ (Fig. 4d), with the central adenine being most highly conserved. We also observed that in most cases a pair of tandem DnaA-boxes preceded the GC-cluster (Extended Data Table 1).

We propose that the DnaA-trio constitutes a new element within bacterial replication origins. Our findings indicate that DnaA-trios play an essential role during DNA replication initiation by providing specificity for DnaA filament formation on a single DNA strand, thereby promoting DNA duplex unwinding. Our analysis also indicates that the arrangement of tandem DnaA-boxes in close proximity to DnaA-trios is a widespread strategy used to direct DnaA filament growth onto the unwinding region, with a single DnaA protein probably binding dsDNA via domain IV before engaging a DnaA-trio via its AAA+ motif (Fig. 4e and Extended Data Fig. 5). Together our data are consistent with the two-step DnaA assembly model for DNA melting18.

We note that the configuration between DnaA-boxes and the DnaA-trios is not strictly required for DnaA to be loaded onto the single DNA strand in vitro. Scaffolds containing either a single DnaA-box or containing DnaA-boxes in reverse orientation are competent to promote DnaA filament formation from the duplex DNA onto the 5′-tail, although in the latter situation DnaA filament formation was reduced suggesting that DnaA-box orientation is important (Extended Data Fig. 7). Furthermore, loading does not require the flexibly tethered domains I/II of DnaA, consistent with previous observations suggesting that domains III/IV can adopt multiple conformations (Extended Data Fig. 7)12,17,18. These results suggest that some plasticity can be accommodated between duplex and single-strand DNA binding elements, which is in agreement with recent reanalyses of essential DnaA-boxes in E. coli and might also explain the location of atypical origin unwinding sites23,24,25,26.

Analysis of replication initiator proteins from both bacteria and archaea shows that the ISM within AAA+ domains is used for DNA binding, and the recent structure of the Drosophila origin recognition complex (ORC) suggests that this is also probably the case for eukaryotes, supporting the model that DNA binding by the ISM is a universal feature of replication initiators4,15,16,27. We find here, for the first time, that the interaction of the B. subtilis replication initiator ISM with the origin involves recognition of a specific DNA sequence. We speculate that motifs analogous to the DnaA-trio might be present in replication origins of higher organisms and recognized by the ISM of ORC proteins. These sites need not be trinucleotides, nor would they necessarily share the same spacing observed for the DnaA-trios as they would need to accommodate the arrangement of AAA + interactions within the respective heterohexameric ORC17,27. The discovery of ISM binding motifs in higher organisms would greatly facilitate origin identification, an elusive problem precluding the understanding of DNA replication control in eukaryotes.


No statistical methods were used to predetermine sample size. The experiments were not randomized. The investigators were not blinded to allocation during experiments and outcome assessment.

Media and chemicals

Nutrient agar (Oxoid) was used for routine selection and maintenance of both B. subtilis and E. coli strains. For experiments in B. subtilis cells were grown using Luria-Bertani medium. Supplements were added as required: chloramphenicol (5 μg ml−1), erythromycin (1 μg ml−1), kanamycin (5 μg ml−1), spectinomycin (50 μg ml−1). Unless otherwise stated, all chemicals and reagents were obtained from Sigma-Aldrich.

Phenotype analysis of oriC mutants using the inducible oriC-independent strain

Strains were grown for 18–72 h at 37 °C on nutrient agar plates either with or without IPTG (1 mM). All experiments were independently performed at least twice and representative data are shown.

Marker frequency analysis

Genomic DNA was harvested from cells during the exponential growth phase and the relative amount of DNA from the replication origin (ori) and terminus (ter) was determined by qPCR. Strains were grown in Luria-Bertani medium to an absorbance, A600 nm, of 0.3–0.5 whereupon sodium azide (0.5%) was added to prevent further metabolism. Chromosomal DNA was isolated using a DNeasy Blood and Tissue Kit (Qiagen). The DNA replication origin (oriC) region was amplified using primers 5′-GAATTCCTTCAGGCCATTGA-3′ and 5′-GATTTCTGGCGAATTGGAAG-3′; the region adjacent to oriN was amplified using primers 5′-CTTTCTGCCGCAAAGGATTA-3′ and 5′-CCTCTTCATAGCCGTTTTGC-3′; the DNA replication terminus (ter) region was amplified using primers 5′-TCCATATCCTCGCTCCTACG-3′ and 5′-ATTCTGCTGATGTGCAATGG-3′. Either Rotor-Gene SYBR Green (Qiagen) or GoTaq (Promega) qPCR mix was used for PCR reactions. qPCR was performed in a Rotor-Gene Q Instrument (Qiagen). By use of crossing points (CT) and PCR efficiency a relative quantification analysis (ΔΔCT) was performed using Rotor-Gene Software version 2.0.2 (Qiagen) to determine the origin:terminus (ori:ter) ratio of each sample. These results were normalized to the ori:ter ratio of a DNA sample from B. subtilis spores, which only contain one chromosome and thus have an ori:ter ratio of 1. Error bars indicate the standard deviation of three technical replicates. All experiments were independently performed at least twice and representative data are shown.

Protein expression

BL21 (DE3)-pLysS cells were transformed with the appropriate expression construct (Supplementary Table 2) and selected on nutrient agar plates containing 100 ng μl−1 of ampicillin and 34 ng μl−1 of chloramphenicol. A single transformant colony was used to inoculate an overnight starter culture grown at 37 °C, 180 rpm, in Luria-Bertani medium supplemented with 100 ng μl−1 of ampicillin and 34 ng μl−1 of chloramphenicol. The following morning a 1/100 dilution of overnight culture was used to inoculate 1,200 ml of Luria-Bertani medium supplemented with 100 ng μl−1 of ampicillin and grown at 37 °C, 180 rpm, to A600 nm = 0.5. Cells were induced with 1 mM IPTG and cultured for a further 3 h at 30 °C. Cells were pelleted at 3,000g, 4 °C for 10 min before resuspension in 45 ml of resuspension buffer (25 mM HEPES-KOH (pH 7.6); 500 mM potassium glutamate; 10 mM magnesium acetate; 20% sucrose; 30 mM imidazole; 1 × cOmplete EDTA-free protease inhibitor tablet (Roche). The cell suspension was then flash-frozen in liquid nitrogen.

Protein purification

DnaA (WT, WT-CC and I190A-CC) was purified as follows. A frozen 50 ml BL21 cell pellet suspension was thawed on ice with 32 mg of lysozyme and gentle agitation for 1 h then disrupted by sonication at 20 W for 5 min in 2 s pulses. Cell debris was pelleted by centrifugation at 31,000g, 4 °C for 45 min and the supernatant further clarified by filtration (0.45 μm). All subsequent steps were performed at 4 °C unless otherwise stated. The clarified lysate was applied at 1 ml min−1 to a 1 ml HisTrap HP column (GE), which had previously been equilibrated with Ni binding buffer (25 mM HEPES-KOH (pH 7.6); 250 mM potassium glutamate; 10 mM magnesium acetate; 20% sucrose; 30 mM imidazole). The loaded column was washed with a 10 ml one-step gradient of 10% Ni elution buffer (25 mM HEPES-KOH (pH 7.6); 250 mM potassium glutamate; 10 mM magnesium acetate; 20% sucrose; 30 mM imidazole). Specifically bound proteins were eluted using a 7.5 ml one-step gradient of 100% Ni elution buffer and the entire fraction collected and diluted into 42.5 ml of Q binding buffer (30 mM Tris-HCl (pH 7.6); 100 mM potassium glutamate; 10 mM magnesium acetate; 1 mM DTT; 20% sucrose). The diluted fraction was then applied at 1 ml min−1 to a 1 ml HiTrap Q HP column (GE), which had previously been equilibrated with Q binding buffer. The loaded HiTrap Q HP column was washed with 10 ml of Q binding buffer then eluted using a linear 10 ml gradient of 0–100% Q elution buffer (30 mM Tris-HCl (pH 7.6); 1 M potassium glutamate; 10 mM magnesium acetate; 1 mM DTT; 20% sucrose) with 1 ml fractions collected. The peak 3 × 1 ml fractions, based on ultraviolet absorbance, were pooled and dialysed into 1 Ll of FactorXa cleavage buffer (25 mM HEPES-KOH (pH 7.6); 250 mM potassium glutamate; 20% sucrose; 5 mM CaCl2), using 3.5k MWCO SnakeSkin dialysis tubing (Life Technologies) at 4 °C overnight. The dialysed protein was diluted to 5 ml total volume in FactorXa cleavage buffer and incubated at 23 °C for 6 h with 80 μg of FactorXa protease (NEB). The sample was applied at 1 ml min−1 to a 1 ml HisTrap HP column (GE), which had previously been equilibrated with Factor Xa cleavage buffer. The Factor Xa-cleaved fraction was eluted in 7.5 ml of Ni binding buffer. The eluted fraction was diluted into 42.5 ml of Q binding buffer and purified on a 1 ml HiTrap Q HP column as previously described. Peak fraction(s) were pooled and dialysed into 1 L of final dialysis buffer (40 mM HEPES-KOH (pH 7.6); 250 mM potassium glutamate; 1 mM DTT; 20% sucrose; 20% PEG300), using 3.5k MWCO SnakeSkin dialysis tubing (Life Technologies) at 4 °C overnight before aliquoting, flash-freezing in liquid nitrogen and storage at −80 °C. Removal of the amino (N)-terminal His-tag, after incubation with FactorXa, was confirmed by anti-pentaHis (Qiagen) western blotting.

C-terminally His-tagged DnaA (WT-CC and Δ(domainI-II)-CC) purification was performed as for the tag-free variants, except that the protein was dialysed into final dialysis buffer after the first HiTrap Q HP column purification before aliquoting, flash-freezing and storing.

HBsu purification was performed exactly as for DnaA, except that the HiTrap Q HP column was substituted for a 1 ml HiTrap Heparin HP column (GE) and the composition of buffers was modified accordingly. Ni binding buffer (25 mM Tris-HCl (pH 8.0); 400 mM NaCl; 30 mM imidazole). Ni elution buffer (25 mM Tris-HCl (pH 8.0); 400 mM NaCl; 500 mM imidazole). Heparin binding buffer (25 mM Tris-HCl (pH 8.0); 100 mM NaCl; 1 mM EDTA). Heparin elution buffer (25 mM Tris-HCl (pH 8.0); 2 M NaCl; 1 mM EDTA). Factor Xa cleavage buffer (25 mM Tris-HCl (pH 8.0); 100 mM NaCl; 2 mM CaCl2; 20% sucrose). Final dialysis buffer (25 mM Tris-HCl (pH 8.0); 400 mM NaCl; 2 mM CaCl2; 20% sucrose; 20% PEG300). Peak fractions were determined by SDS–PAGE and Coomassie staining owing to the absence of tryptophan, tyrosine and cysteine residues.

Open complex formation assays

KMnO4 footprinting assays were essentially performed as described in ref. 9, except for the following changes. DnaA was not pre-incubated with ATP. The unwinding buffer contained 2 mM ATP, rather than 5 mM, and 500 ng of plasmid pTR541 (wild type) or pTR542 (t1-t6scr) was used per 75-μl-scale reaction. DnaA was added to final concentrations of 0, 100, 250, 500 and 1,000 nM. Assembled reactions were incubated at 37 °C for 10 min. KMnO4 treatment was then performed at 37 °C for 10 min. Six microlitres of β-mercaptoethanol was used to quench reactions; however, EDTA was omitted. KMnO4-treated DNA was immediately purified using a Qiagen PCR clean-up kit, eluting in 20 μl of EB buffer. KMnO4-treated templates were not linearized before primer extension. Primer extensions were performed on a 20 μl scale using 0.1 U μl−1 of Vent exo- DNA polymerase (NEB) in 1 × manufacturer’s reaction buffer supplemented with 4 mM MgSO4, 200 μM each dNTP, 200 nM Cy5-labelled oligonucleotide (5′-Cy5-AGCTTCAGCAGCATGTAAAAG-3′) and 4 μl of PCR-purified template DNA per reaction. Reactions were subjected to thermocycling using a 3Prime thermal cycler (Techne) with 1 min initial denaturation at 98 °C, followed by 35 cycles of (10 s at 98 °C; 30 s at 55 °C; 30 s at 72 °C). Reactions were quenched by addition of an equal volume of stop buffer (95% formamide; 10 mM EDTA; 10 mM NaOH; 0.01% Orange-G) and products subjected to denaturing PAGE (6% acrylamide:bisacrylamide (19:1); 8 M urea in 1 × TBE). Resolved products were visualized using a Typhoon Trio Variable Mode Imager (GE Healthcare). The DnaA-trio marker was generated by primer extension performed under the same conditions as described for KMnO4-treated substrates, but using a PCR product as template generated with a primer corresponding to the end of the first DnaA-trio (5′-TAGGGCCTGTGGATTTGTG-3′). All experiments were independently performed at least twice and representative data are shown.

Filament assembly assays (BMOE)

DNA scaffolds were prepared by mixing each oligonucleotide (50 nM final concentrations) in 10 mM HEPES-KOH (pH 7.6), 100 mM NaCl and 1 mM EDTA. Mixed oligonucleotides were heated to 98 °C for 5 min in a heat-block and slowly cooled to room-temperature in the heat-block before use. Filament formation was promoted by mixing DnaA-CC proteins (WT, I190A, ΔdomainI-II) (200 nM final concentration) with DNA scaffold (15 nM) on a 20 μl scale in 30 mM HEPES-KOH (pH 7.0), 100 mM potassium glutamate, 100 mM NaCl, 10 mM magnesium acetate, 25% glycerol, 0.01% Tween-20 and 2 mM nucleotide (ADP or ATP). Reactions were incubated at 37 °C for 5–12 min before addition of 4 mM BMOE (ThermoFisher Scientific). Reactions were incubated at 37 °C for 5–12 min before quenching by addition of 60 mM cysteine. Reactions were incubated once more at 37 °C for 10–12 min before fixing in NuPAGE LDS sample buffer (ThermoFisher Scientific) at 98 °C for 5 min. Complexes were resolved by running 500 fmol of cross-linked DnaA from each reaction on a NuPAGE Novex 3–8% Tris-acetate gel (ThermoFisher Scientific) then transferred to Hybond 0.45 μm PVDF membrane (Amersham) in 0.5 × NuPAGE Tris-acetate SDS running buffer with 20% MeOH at 35 mA, 4 °C overnight using wet transfer apparatus (Biorad). Complexes were visualized by western blotting using a polyclonal anti-DnaA antibody (Eurogentec). NB: all filament assembly assays were performed using tag-free proteins with the exception of that shown in Extended Data Fig. 7, in which C-terminally His-tagged proteins (ΔdomainI-II-CC and wild-type-CC) were used. All experiments were independently performed at least twice and representative data are shown.

Filament assembly assays (BS3)

Filament assembly assays using bis(sulfosuccinimidyl)suberate (BS3) were performed as described for BMOE, except a tag-free fully wild-type recombinant DnaA protein was used for Extended Data Fig. 3. Tag-free ‘CC’ variants of wild-type and I190A DnaA were used for Extended Data Fig. 5a, b. Crosslinking was performed using BS3 (15 mM final concentration) in place of BMOE and quenching performed by addition of Tris-HCl (pH 7.6) (30 mM final concentration). All experiments were independently performed at least twice and representative data are shown.


To visualize GFP-DnaN, starter cultures were grown overnight in defined minimal medium base (Spizizen minimal salts supplemented with Fe-NH4-citrate (1 μg ml−1), MgSO4 (6 mM), CaCl2 (100 μM), MnSO4 (130 μM), ZnCl2 (1 μM), thiamine (2 μM)) supplemented with casein hydrolysate (200 μg ml−1) and glycerol (0.5%) with IPTG (1 mM) at 37 °C, diluted 1:100 into fresh medium with IPTG (1 mM) and allowed to grow at 37 °C for several generations until they reached A600 nm = 0.3. Cells were collected by centrifugation, washed to remove IPTG, and resuspended into fresh medium at A600 nm = 0.1 and allowed to grow until A600 nm = 0.6. Cells were mounted on 1.5% agar pads (0.5 × growth media) and a 0.13–0.17 mm glass coverslip (VWR) was placed on top. Microscopy was performed on an inverted epifluorescence microscope (Nikon Ti) fitted with a Plan-Apochromat objective (Nikon DM 100 × /1.40 Oil Ph3). Light was transmitted from a 300 W xenon arc-lamp through a liquid light guide (Sutter Instruments) and images were collected using a CoolSnap HQ2 cooled CCD (charge-coupled device) camera (Photometrics). All filters were Modified Magnetron ET Sets from Chroma and details are available upon request. Digital images were acquired and analysed using METAMORPH software (version 6.2r6). All experiments were independently performed at least twice and representative data are shown.


Strains are listed in Supplementary Table 1. The genotype of all origin mutants was confirmed by DNA sequencing.


All oligonucleotides were purchased from Eurogentec. Oligonucleotides used for plasmid construction are listed in Extended Data Table 2. Oligonucleotides used to construct DNA scaffolds are listed in Extended Data Table 3.


Plasmids are listed in the Supplementary Table 2 (sequences are available upon request). DH5α (F Φ80lacZΔM15 Δ(lacZYA-argF) U169 recA1 endA1 hsdR17(rk, mk+) phoA supE44 thi-1 gyrA96 relA1 λ)29 was used for plasmid construction, except where noted. Descriptions, where necessary, are provided below.

pHM327 derivatives were generated by quickchange mutagenesis using the oligonucleotides listed in Extended Data Table 2. After sequencing to confirm mutated regions, sequences were subcloned using BglII/FspAI.

pHM446 (bla aprEkan lacI Pspac-MCS ′aprE) is a derivative of pAPNC213 (ref. 30) with a kanamycin resistance cassette replacing the spectinomycin resistance cassette (gift from H. Strahl).

pHM453 (bla rpnArpmH erm ΔincAB Pspac-dnaA′) was created in multiple steps. First, pJS1 was generated by ligation with a HindIII-BamHI PCR product containing 5′ end of dnaA and pMUTIN4 (ref. 31) cut with HindIII-BamHI (gift from J. Errington). Second, pHM396 was generated by digestion of pJS1 with PvuII (to remove lacZ and lacI) and ligation of the vector backbone. Finally, pHM453 was generated by ligation of an AatII PCR product containing rpmH and the 5′ end of rpnA (oHM319 + oHM320 and 168CA genomic DNA as template) with pHM396 cut with AatII.

pHM492 (bla aprEkan lacI Pspac-repN(oriN) ′aprE) was generated by ligation of an EcoRI–XhoI PCR product containing repN(oriN) (oHM313 + oHM315 and MMB208 (ref. 32) genomic DNA as template) with pHM446 cut with EcoRI–SalI.

pHM560 (bla rpnArpmH erm ΔincA PdnaA ΔincB dnaA′) was generated by ligation of an EcoRV–HindIII PCR product containing the dnaA promoter (oHM510 + oHM511 and 168CA genomic DNA as template) with pHM453 cut with EcoRV–HindIII.

pTR72, pTR73, pTR74, pTR102, pTR168 were generated by quickchange mutagenesis using the oligonucleotides listed in Extended Data Table 2.

pTR208 was generated by two-fragment PCR. oTR384/oTR385 and oTR386/oTR387 were used to amplify products using pTR74 and B. subtilis 168CA genomic DNA as templates, respectively. An equal volume of each PCR product was mixed, heated to 98 °C and allowed to cool to room temperature before DpnI digestion and transformation.

pTR229 (bla PT7(his6-link-Xa-dnaA)) was generated by subcloning a HindIII–XhoI fragment of dnaA from pHM239 into the pTR74 backbone.

pTR541 and pTR542 were generated by two-fragment PCR. oTR537 and oTR538 were used to amplify the plasmid backbone of pSG1301. oTR535 and oTR536 were used to amplify incC with B. subtilis 168CA genomic DNA and pTR84 used as the templates for pTR541 and pTR542, respectively. An equal volume of each PCR product was mixed, heated to 98 °C and allowed to cool to room temperature before DpnI digestion and transformation into EH3827 (asnB32 relA1 spoT1 thi-1 fuc-1 lysA ilv-192 zia::pKN500 ΔdnaA mad-1)33. DNA sequencing confirmed the construction of each origin including flanking sequences (>400 base pairs upstream and downstream).


  1. 1.

    & A structural framework for replication origin opening by AAA + initiation factors. Curr. Opin. Struct. Biol. 23, 144–153 (2013)

  2. 2.

    The bacterial replication initiator DnaA. DnaA and oriC, the bacterial mode to initiate DNA replication. FEMS Microbiol. Rev. 26, 355–374 (2002)

  3. 3.

    , & The dnaA protein complex with the E. coli chromosomal replication origin (oriC) and other DNA sites. Cell 38, 889–900 (1984)

  4. 4.

    , & DNA stretching by bacterial initiators promotes replication origin opening. Nature 478, 209–213 (2011)

  5. 5.

    et al. Structural basis of replication origin recognition by the DnaA protein. Nucleic Acids Res. 31, 2077–2086 (2003)

  6. 6.

    , , , & Where does bacterial replication start? Rules for predicting the oriC region. Nucleic Acids Res. 32, 3781–3791 (2004)

  7. 7.

    , , & oriC-encoded instructions for the initiation of bacterial chromosome replication. Front. Microbiol. 5, 735 (2015)

  8. 8.

    et al. Suppression of initiation defects of chromosome replication in Bacillus subtilis dnaA and oriC-deleted mutants by integration of a plasmid replicon into the chromosomes. J. Bacteriol. 179, 2494–2502 (1997)

  9. 9.

    , , & Complexes at the replication origin of Bacillus subtilis with homologous and heterologous DnaA protein. J. Mol. Biol. 274, 365–380 (1997)

  10. 10.

    & Regulation of DnaA assembly and activity: taking directions from the genome. Annu. Rev. Microbiol. 65, 19–35 (2011)

  11. 11.

    & Mechanism of origin unwinding: sequential binding of DnaA to double- and single-stranded DNA. EMBO J. 20, 1469–1476 (2001)

  12. 12.

    , & Soj/ParA stalls DNA replication by inhibiting helix formation of the initiator protein DnaA. EMBO J. 31, 1542–1555 (2012)

  13. 13.

    , , & Bacterial initiators form dynamic filaments on single-stranded DNA monomer by monomer. Nucleic Acids Res. 43, 396–405 (2015)

  14. 14.

    et al. A common mechanism for the ATP-DnaA-dependent formation of open complexes at the replication origin. J. Biol. Chem. 283, 8351–8362 (2008)

  15. 15.

    , , & Structural basis of DNA replication origin recognition by an ORC protein. Science 317, 1213–1216 (2007)

  16. 16.

    , , & Replication origin recognition and deformation by a heterodimeric archaeal Orc1 complex. Science 317, 1210–1213 (2007)

  17. 17.

    , & Structural basis for ATP-dependent DnaA assembly and replication-origin remodeling. Nature Struct. Mol. Biol. 13, 676–683 (2006)

  18. 18.

    et al. Origin remodeling and opening in bacteria rely on distinct assembly states of the DnaA initiator. J. Biol. Chem. 285, 28229–28239 (2010)

  19. 19.

    & DnaA proteins of Escherichia coli and Bacillus subtilis: coordinate actions with single-stranded DNA-binding protein and interspecies inhibition during open complex formation at the replication origins. Gene 228, 123–132 (1999)

  20. 20.

    , , , & Helicobacter pylori oriC – the first bipartite origin of chromosome replication in Gram-negative bacteria. Nucleic Acids Res. 40, 9647–9660 (2012)

  21. 21.

    , , & The DnaA homolog of the hyperthermophilic eubacterium Thermotoga maritima forms an open complex with a minimal 149-bp origin region in an ATP-dependent manner. Genes Cells 11, 425–438 (2006)

  22. 22.

    , & DoriC 5.0: an updated database of oriC regions in both bacterial and archaeal genomes. Nucleic Acids Res. 41, D90–D93 (2013)

  23. 23.

    , & In-vitro helix opening of M. tuberculosis oriC by DnaA occurs at precise location and is inhibited by IciA like protein. PLoS ONE 4, e4139 (2009)

  24. 24.

    et al. Mechanism for the TtDnaA-Tt-oriC cooperative interaction at high temperature and duplex opening at an unusual AT-rich region in Thermoanaerobacter tengcongensis. Nucleic Acids Res. 35, 3087–3099 (2007)

  25. 25.

    et al. Building the bacterial orisome: high-affinity DnaA recognition plays a role in setting the conformation of oriC DNA. Mol. Microbiol. 91, 1148–1163 (2014)

  26. 26.

    , , & The Arg fingers of key DnaA protomers are oriented inward within the replication origin oriC and stimulate DnaA subcomplexes in the initiation complex. J. Biol. Chem. 290, 20295–20312 (2015)

  27. 27.

    , & Crystal structure of the eukaryotic origin recognition complex. Nature 519, 321–326 (2015)

  28. 28.

    , , & WebLogo: a sequence logo generator. Genome Res. 14, 1188–1190 (2004)

  29. 29.

    , & E. coli host strains significantly affect the quality of small scale plasmid DNA preparations used for sequencing. Nucleic Acids Res. 21, 1677–1678 (1993)

  30. 30.

    et al. Six GTP-binding proteins of the Era/Obg family are essential for cell growth in Bacillus subtilis. Microbiology 148, 3539–3552 (2002)

  31. 31.

    , & A vector for systematic gene inactivation in Bacillus subtilis. Microbiology 144, 3097–3104 (1998)

  32. 32.

    & Subcellular positioning of the origin region of the Bacillus subtilis chromosome is independent of sequences within oriC, the site of replication initiation, and the replication initiator DnaA. Mol. Microbiol. 63, 150–165 (2007)

  33. 33.

    & Host participation in plasmid maintenance: dependence upon dnaA of replicons derived from P1 and F. Proc. Natl Acad. Sci. USA 83, 4423–4427 (1986)

  34. 34.

    , , , & Cloning of an autonomously replicating sequence (ars) from the Bacillus subtilis chromosome. Mol. Microbiol. 6, 309–315 (1992)

  35. 35.

    , , & Replication origin of the Escherichia coli K-12 chromosome: the size and structure of the minimum DNA segment carrying the information for autonomous replication. Mol. Gen. Genet. 178, 9–20 (1980)

  36. 36.

    & Conserved gene arrangement in the origin region of the Streptomyces coelicolor chromosome. J. Bacteriol. 174, 3220–3226 (1992)

  37. 37.

    et al. Light-dependent and asynchronous replication of cyanobacterial multi-copy chromosomes. Mol. Microbiol. 83, 856–865 (2012)

Download references


We thank J. Errington and W. Vollmer for reviewing the manuscript. We thank G. Scholefield for preliminary data, A. Koh for research assistance and I. Selmes for technical assistance. Research support was provided to H.M. by a Royal Society University Research Fellowship and a Biotechnology and Biological Sciences Research Council Research Grant (BB/K017527/1), and to O.H. by an Iraqi Ministry of Higher Education and Scientific Research Studentship.

Author information


  1. Centre for Bacterial Cell Biology, Institute for Cell and Molecular Biosciences, Newcastle University, Newcastle upon Tyne NE2 4AX, UK

    • Tomas T. Richardson
    • , Omar Harran
    •  & Heath Murray


  1. Search for Tomas T. Richardson in:

  2. Search for Omar Harran in:

  3. Search for Heath Murray in:


H.M. and T.T.R. conceived and designed experiments; H.M., T.T.R. and O.H. constructed plasmids and strains; H.M. and O.H. performed growth and marker frequency analysis experiments; H.M. performed microscopy experiments; T.T.R. purified proteins, performed the open complex assay, and performed the DnaA filament formation assays; H.M. and T.T.R. interpreted results and wrote the paper.

Competing interests

The authors declare no competing financial interests.

Corresponding author

Correspondence to Heath Murray.

Extended data

Supplementary information

PDF files

  1. 1.

    Supplementary Information

    This file contains Supplementary Text, Supplementary Tables 1-2 and additional references.

About this article

Publication history






Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.