Structural insight into Helicobacter pylori DNA replication initiation

Abstract

While increasing knowledge is accumulating about the molecular mechanisms allowing the human pathogen Helicobacter pylori to survive and to subvert host defenses, much less is known about fundamental aspects of its biology, including DNA replication. We have studied the initiation step of chromosome replication of H. pylori and particularly the interaction between the initiator protein DnaA and its recently identified regulator HobA. This work has recently culminated in the determination of the crystal structure of the domains I and II of DnaA (DnaA^I−II) in complex with HobA. By combining the structure with a variety of biochemical experiments we show that a tetramer of HobA can accommodate up to four DnaA molecules organized in a particular conformation within the complex. Mutations of the HobA interface that impaired the binding with DnaA were designed and proved to be lethal once introduced into H. pylori. These features suggest that HobA provides a molecular scaffold onto which regular oligomers of DnaA can assemble. The HobA-promoted oligomerization of DnaA could have a determinant role in the formation of the open complex. We propose a speculative model of HobA-dependent DnaA oligomerization leading to DNA unwinding. More generally, the parallel we draw with Escherichia coli DnaA and DiaA (HobA-like E. coli protein) will direct new studies that will contribute to the understanding of bacterial DNA replication.

Introduction

Bacterial DNA replication has been extensively studied over the past fifty years. Many of the general principles regarding this process come from studies of the Escherichia coli replication system. In E. coli, replication consists of three steps: initiation, elongation and termination, which progression is coordinated to other cell cycle events. Initiation requires the DnaA protein, a member of the AAA⁺ family of ATPases, that directly binds to specific DNA sequences (DnaA boxes) on the origin of replication oriC to form the orisome complex (reviewed in ref. 1 and 2). With the aid of other proteins, DnaA forms the initial nucleoprotein complex and unwinds double stranded DNA at a specific AT rich region called DUE (DNA unwinding element). ssDUE-bound DnaA serves as a docking platform to load the hexameric DnaB helicase delivered as a DnaB/DnaC complex. The helicase loader DnaC also belongs to AAA⁺ class of ATPases, is closely related to DnaA and assists the loading of DnaB onto the unwound region in ATP-dependent manner. DnaB recruitment occurs via protein-protein interactions between DnaA and DnaB,^3,4 and most likely also between DnaA and DnaC.⁵ Replication proteins are subsequently loaded (e.g., DnaG primase, DNA III polymerase holoenzyme) to finally assemble an active replisome and allow for replication to start.

DnaA consists of four domains (DnaA^I–IV) playing distinct biochemical and architectural functions (reviewed in ref. 6) that rely on protein oligomerization. In E. coli, DnaA oligomerization is mediated by at least two of these domains: domain I and domain III. On the one hand, oligomerization of E. coli DnaA can occur via self interaction between DnaA^I domains,^7,8 that seems to play a key role in the assembly of the orisome. The flexible linker DnaA^II is not involved in this process. On the other hand, structural studies have shown that Aquifex aeolicus DnaA^III–IV bound to a non-hydrolysable ATP analog form a right-hand super-helical structure.⁹ In this multimeric structure, ATP analog molecules are bound by two adjacent DnaA^III subunits which form an helical filament in the crystal. In that helix, DnaA domains IV (DnaA^IV), containing the DNA binding domain, are located at the external side. Thus, it was proposed that during orisome formation, the DNA could wrap around the helix which would unwind DUE of the oriC.^1,9 The formation of the E. coli orisome is also facilitated by several accessory factors such as IHF, HU and DiaA.^10–12 In particular, DiaA, was found to bind to DnaA^I and regulate the timing of replication since DiaA-depleted cells replicate asynchronously.^13,14 Additional studies revealed that DiaA was specifically binding to ATP-DnaA, hereby stimulating its interaction with weak DnaA boxes (I-sites, τ-sites, M, R2 and R3) and promoted DUE unwinding in vitro and in vivo.¹² DnaA^I also interacts with DnaB, suggesting that this domain is also involved in loading the helicase on the DUE.^3,4

The core principle of the replication fork assembly on oriC seems to be conserved in Helicobacter pylori. The proteins fundamental for the initiation process (DnaA, DnaB) are encoded by H. pylori chromosome, but no sequence homolog has been identified for many E. coli replication-associated proteins including DnaC. Conversely, a high throughput yeast two-hybrid study (Y2H) identified three proteins of unknown functions capable of binding DnaA: HP1230, HP0944 and HP1423.¹⁵ The association of HP1230 with DNA replication was established when it was found to be absolutely required for initiation of H. pylori chromosome replication hence bacterial survival and formed a complex with DnaA in vivo. HP1230 interacted with DnaA-oriC, enhanced orisome formation in vitro and probably also rearranged its structure.^16,17 Accordingly, the protein was named HobA, for Helicobacter orisome binding protein A.¹⁶ The crystal structure of HobA showed that it was in fact a structural homolog of E. coli DiaA.¹⁸ In addition both proteins bind a similar region of DnaA, the domains I–II, form tetramers and contain a SIS domain, which function remains to be elucidated.^12,18 At this time we hypothesized that the two proteins might have similar roles and could reflect a general mechanism that regulates the replication initiation.

In a recent study (Natrajan et al. PNAS 2009) that we summarize here we have solved the crystal structure of the complex between HobA and DnaA^I–II. From the structure, a number of biochemical and in vivo experiments were designed that altogether shed light onto the molecular determinants of the complex formation and show that this complex is essential for the bacterial survival.

Architecture of the DnaA^I–II/HobA Complex

To gain insight into HobA role in DNA replication we have solved the crystal structure of DnaA^I–II domains in complex with HobA.¹⁹ The structure of the HobA/DnaA^I–II complex contains a tetramer of HobA and four molecules of DnaA^I–II located at opposite poles of the HobA tetramers. (Fig. 1A). DnaA^I–II molecules are in fact positioned at the dimer-dimer interface of the tetramer, with each of the DnaAI-II bound to two adjacent monomers of HobA suggesting that the HobA dimer-dimer interface is needed for DnaA interaction. The stoichiometry observed in the crystal was thus four DnaA^I–II for four HobA. The stoichiometry of the complex in solution was determined by combining analytical gel filtration and Isothermal Titration Calorimetry (ITC) experiments. Gel filtration experiments showed that a protein complex formed in the mixture of HobA and DnaA^I–II eluted earlier than HobA suggesting formation of HobA/DnaA^I–II complex of higher molecular weight than the HobA tetramer. Moreover, ITC experiments indicated that, in solution, the molar ratio of the complex was one to one. Therefore the complex in solution was in agreement with the one observed in the crystal structure and consists of 4 molecules of DnaA^I–II bound to one tetramer of HobA (Fig. 1).

Figure 1.

Structure of the HobA/DnaA^I–II complex. (A) Picture depicting two views of the biological assembly of the HobA (two dimers colored in pale and bright orange) and DnaA represented in ribbon. The surface of HobA is also represented with the binding interface of DnaA colored in blue. (B) Detailed view of the binding interface with critical residues represented as ball and stick. Residues of DnaA are colored in blue. Residues of HobA are colored according to the dimer they belong to (pale or bright orange). Residues of HobA that have been mutated in this study are colored in magenta. (C) Table summarizing the mutagenesis study. Interactions of HobA and HobA mutants with DnaA domains I–II (DnaA^I–II) and/or full length DnaA (flDnaA) were tested in pull-down assay, cross-linking experiments and yeast two-hybrid (Y2H).1 In the cross-linking experiments HobA and HobA mutants oligomerization were tested.² The ability of HobA mutants to interact with wt-HobA was tested in the Y2H assays.

Binding Mode of DnaA^I–II onto HobA

In the HobA4(DnaA^I–II)₄ complex, each of the DnaA^I–II molecules interacts with two adjacent subunits of HobA (Fig. 1A). Briefly, the interaction between each DnaA^I–II and HobA is mediated principally by the binding of helices α2 and α3 of DnaA^I to two grooves, located on each side of the α3-β2 loop of HobA (named ERP loop because it includes residues Glu 76, Arg 77 and P78) and involves both electrostatic and hydrophobic interactions (Fig. 1B). In the first binding cavity, α2 from DnaA^I–II sits in a small cavity in HobA formed by the ERP loop and the C-terminus of α6. Residues from α2 of DnaA^I–II bind predominantly by polar contacts with the major interactions being between DnaA^I–II Y29 and HobA E76 and Y175. In the second cavity, residues from α3 from DnaA^I–II sit in a hydrophobic cleft delimited by loops α2-β1 and the ERP loop from one HobA subunit and helix α4 from an adjacent subunit (Fig. 1B). In addition to hydrophobic interactions, binding of α3 is secured by hydrogen bonds between K61 (DnaA^I–II) and E76 from the HobA ERP loop. DnaA^I–II V53 side chain inserts into in a small hydrophobic cavity formed by L170, L80, L174 and L45 of HobA (Fig. 1B). Several water-mediated hydrogen bonds and hydrophobic interactions secure the positioning of α3 from DnaA^I–II in the cavity.

Mutational Analysis of the DnaA^I–II/HobA Interaction

To probe the interface identified in the crystal structure, we generated specific HobA mutants to disrupt DnaA/HobA interaction. Single mutants L80R, A101E, L174A, Y175E and a triple mutant of the ERP loop (E76A/R77A/P78A, named ERP mutant hereafter) were engineered (Fig. 1C). The purified untagged mutant proteins were mixed with His₆-DnaA^I–II or full-length His₆-DnaA and the complexes formation was monitored by pull-down using a metal affinity resin. Wild-type HobA (wt-HobA), A101E and L174A mutants were able to interact with both His₆-DnaA^I–II and His₆-DnaA. In contrast, HobA ERP and L80R mutants were completely defective. Y175E mutant interacted weakly with His₆-DnaA^I–II and not at all with His₆-DnaA (full-length DnaA protein). We also used cross-linking experiments with His₆-DnaA^I–II and HobA mutants, which confirmed that HobA/DnaA^I–II interaction was abolished for ERP, L80R and Y175E mutants (Fig. 1C). Complex formation was maintained with L174A and A101E, although with less efficiency than for wt-HobA (not shown). Interestingly, all the HobA mutants retained their capacity to form tetramers except L80R and ERP, suggesting that HobA tetramer formation might also be important for interaction with DnaA (Fig. 1C).

We then used Y2H to test the HobA mutants for their ability to interact with DnaA. Again, ERP, L80R and Y175E mutants lost their ability to interact with DnaA but L174A and A101E mutants did not, in agreement with the results observed in our pull down assays and cross-linking experiments (Fig. 1C). We also observed that the L80R mutant interaction with wt-HobA was altered, confirming the results obtained in the cross-linking experiments. As an alternative method to site-directed mutagenesis, a recently developed random mutagenesis strategy²⁰ was also performed in parallel on hobA gene to identify HobA mutations that affected DnaA binding in Y2H. Mutations were identified and all pointed towards residues belonging to the interface, thereby confirming the validity of the complex and also of this approach to identify functional residues.

Effects of Introduction of DnaA Interaction-Defective Mutants of HobA in Helicobacter pylori

To evaluate the impact of a deficient DnaA-HobA interaction in vivo, wt-hobA was replaced by hobA mutants in H. pylori strain 26695 cells. HobA mutants A101E and L174A were introduced in H. pylori and the growth curves of bacteria carrying these mutations were similar to those carrying wt-HobA. In the case of the ERP, L80R and Y175E mutants, the aph-3 cassette was introduced into H. pylori chromosome, but the sequence of hobA was that of the wild-type. We hypothesized that alternative recombination events occurred in H. pylori to circumvent the lethal changes introduced by the mutations. Such rearrangements have been described previously and are known to happen when lethal changes are introduced in this bacteria.^21,22 Our hypothesis is also supported by the fact that HobA mutations A101E and L174A, which had no effect on DnaA binding in vitro, were successfully introduced in H. pylori and had also no effect on bacterial growth. We therefore concluded that the interaction of HobA and DnaA is required for H. pylori viability and that the bacteria did not introduce lethal mutations into hobA to overcome a deficient HobA/DnaA interaction.

Oligomerization of H. pylori DnaA^I–II

Oligomerization of E. coli DnaA is a major step in the assembly of an active orisome. Two mechanisms of oligomerization are known: domain I can self-interact^7,8 and ATP-bound domain III assemble to form high order oligomers.¹ In the case of H. pylori no self-interaction was found between DnaA molecules using Y2H.¹⁵ In the crystal structure of the HobA/DnaA complex, few contacts were found between domains I of H. pylori DnaAs. By studying the oligomerization state of H. pylori DnaA^I–II in solution, we found that the protein did not form oligomers neither in gel filtration nor in cross-linking experiments. Consistent with this finding, a yeast three-hybrid (3HB) assay, where both AD- and BD-DnaA fusions were co-expressed in the diploid cells in the presence or in the absence of HobA, showed that HobA triggered an interaction phenotype, while no self-interaction was detected for DnaA in the absence of HobA. Therefore, and in contrast to its E. coli homolog, H. pylori DnaA^I–II does not form oligomers. Interestingly, although DnaA^I adopts a KH fold similar to E. coli DnaA^I NMR structure,^3,23 the residues identified as important for E. coli DnaA self interaction are not conserved in H. pylori DnaA, which provides a structural explanation for the latter protein inability to oligomerise via domain I.

A Role for HobA in DnaA Oligomerization at the Replication Fork?

Although the molecular details of the interaction between HobA and DnaA are now clarified, the exact function of the complex is still unknown. Based on the existing data, a possible role for HobA would be to generate a replication-competent DnaA oligomer at the oriC region. A model of a possible mechanism is depicted in Figure 2. It presents (A) the prediction of a complex of full-length HpDnaA^I–IV on HobA and (B) the participation of the complex to the orisome formation. In DnaA/HobA complex four DnaA^I–IV molecules are bound to one tetramer of HobA via domains I. Interestingly, the resulting position of domains III–IV of each DnaA molecule could facilitate self-interaction and also bring two by two the DNA binding domains of DnaAs (domain IV). Oligomeration via domains III of DnaAs can also occur independently during orisome formation, since the AAA⁺ features of E. coli DnaA seem to be conserved in H. pylori DnaA. The binding of the HobA/DnaA complex to the closely spaced twin boxes 2–3 and 4–5, may help to establish interaction between domains III of DnaA protomers. The DnaA binding to boxes 2–3 and 4–5 (high and moderate affinity, respectively) may somehow facilitate the interaction between DnaA and box 1 (weak affinity) and enhance DnaA oligomerization on oriC.²⁴ Further changes in DNA topology caused by DnaA oligomerization via additional HobA/DnaA complexes incorporation may trigger DNA unwinding and allow further initiation steps to progress.

Figure 2.

Model of DnaA oligomerization on HobA and a speculative model of orisome formation. (A) Ribbon diagram of a speculative model of a full-length DnaA/HobA complex. The model shows the HobA (orange)/DnaA domains I (blue) crystal structure. The crystal structure of DnaA^I was completed by manual building of a linker region for residues of domain II (yellow) followed by an homology model of HpDnaA domains III (green) and IV (red) based on Aquifex aeolicus DnaA crystal structure (pdb code 1L8Q). (B) Schematic representation of a model of HobA/DnaA complex and its assembly onto the oriC region of H. pylori chromosome. The complexes HobA/DnaA can bind to the twin DnaA boxes 5/4 (moderate affinity) and 3/2 (high affinity) and also to box 1 (low affinity). The resulting binding of the DNA and the proximity of adjacent DnaA molecules result in the formation of a higher DnaA/HobA oligomers that can promote DNA unwinding.

DiaA and HobA, Similar Structures, Similar Functions?

This HobA-mediated oligomerization of DnaA seems to have its counterpart in E. coli. We first showed that HobA and DiaA were structural homolog and bind to the same domains of DnaA.¹⁸ More recently, a study by Keyamura and colleagues²⁵ used NMR to demonstrate that the interaction between DiaA and E. coli DnaA occurred at an interface that corresponds remarkably well to the HobA/DnaA interface we identified. Moreover, this group showed that the DiaA complex was important to promote ATP dependent DnaA oligomerization.¹² Therefore HobA and DiaA represent a similar structural scaffold used by DnaA to generate a specific functional complex at the oriC region.

It is worth noting that HobA and DiaA, although being structural homolog share less than 8% sequence identity and might have different functions in their respective bacteria. For instance, HobA is essential to H. pylori survival while a diaA-disruption mutant shows asynchronous cell division but it is not lethal for E. coli. Moreover, a sequence homolog of DiaA exists in H. pylori genome (HP0857)¹² and raises the interesting possibility of the co-existence of these two regulators in a single H. pylori cell, which would have to compete for the same binding site on DnaA. Finally, DnaA-bound DiaA was found to inhibit the loading of the helicase DnaB, and act as a licensing factor similar to eukaryotic proteins.²⁵ Little is known about these steps in H. pylori and no interaction between H. pylori DnaA and DnaB were detected in Y2H.¹⁵ H. pylori DnaB helicase also presents some unique properties, that suggest that unknown molecular interplay exists between HobA, DnaA, DnaB and possibly other proteins that might contribute to DNA replication initiation or its regulation. We anticipate that studies comparing the molecular properties of DiaA, HobA and their interactions with DnaA, DnaB on the oriC region will help to unravel the molecular steps of bacterial DNA replication initiation.