Christoph Weigel

ATP– and ADP–DnaA protein, a molecular switch in gene regulation

DnaA protein functions by binding to asymmetric 9mer DNA sites, the DnaA boxes. ATP–DnaA and ADP–DnaA bind to 9mer DnaA boxes with equal affinity, but only ATP–DnaA protein binds in addition to an as yet unknown 6mer site, the ATP–DnaA box AGATCT, or a close match to it. ATP–DnaA protein binding to ATP–DnaA boxes is restricted to sites located in close proximity to DnaA boxes, suggesting that protein–protein interaction is required for its stabilization. We show that ATP–DnaA represses dnaA transcription much more efficiently than ADP–DnaA. DnaA is thus a regulatory molecule that, depending on the adenosine nucleotide bound, can bind to different sequences and thereby fulfill distinct functions.

DnaA initiator—also a transcription factor

The replication‐initiator protein DnaA is ubiquitous in the eubacterial world. It binds to an asymmetric 9u2003bp consensus DNA sequence, the DnaA box. Besides its primary function as an initiator, it acts as a transcription factor that represses or activates several genes, or terminates transcription, depending on the location and arrangement of DnaA boxes.

The interaction domains of the DnaA and DnaB replication proteins of Escherichia coli.

The initiation of chromosome replication in Escherichia coli requires the recruitment of the replicative helicase DnaB from the DnaBC complex to the unwound region within the replication origin oriC, supported by the oriC‐bound initiator protein DnaA. We defined physical contacts between DnaA and DnaB that involve residues 24–86 and 130–148 of DnaA and residues 154–210 and 1–156 of DnaB respectively. We propose that contacts between DnaA and DnaB occur via two interaction sites on each of the proteins. Interaction domain 24–86 of DnaA overlaps with its N‐terminal homo‐oligomerization domain (residues 1–86). Interaction domain 154–210 of DnaB overlaps or is contiguous with the domains known to interact with plasmid initiator proteins. Loading of the DnaBC helicase in vivo can only be performed by DnaA derivatives containing (in addition to residues 24–86 and the DNA‐binding domain 4) a structurally intact domain 3. Nucleotide binding by domain 3 is, however, not required. The parts of DnaA required for replication of pSC101 were clearly different from those used for helicase loading. Domains 1 and 4 of DnaA, but not domain 3, were found to be involved in the maintenance of plasmid pSC101.

The N-terminus promotes oligomerization of the Escherichia coli initiator protein DnaA

Initiation of chromosome replication in Escherichia coli is governed by the interaction of the initiator protein DnaA with the replication origin oriC. Here we present evidence that homo‐oligomerization of DnaA via its N‐terminus (amino acid residues 1–86) is also essential for initiation. Results from solid‐phase protein‐binding assays indicate that residues 1–86 (or 1–77) of DnaA are necessary and sufficient for self interaction. Using a ‘one‐hybrid‐system’ we found that the DnaA N‐terminus can functionally replace the dimerization domain of coliphage lambda cI repressor: a λcI‐DnaA chimeric protein inhibits λ plasmid replication as efficiently as λcI repressor. DnaA derivatives with deletions in the N‐terminus are incapable of supporting chromosome replication from oriC, and, conversely, overexpression of the DnaA N‐terminus inhibits initiation in vivo. Together, these results indicate that (i) oligomerization of DnaA N‐termini is essential for protein function during initiation, and (ii) oligomerization does not require intramolecular cross‐talk with the nucleotide‐binding domain III or the DNA‐binding domain IV. We propose that E. coli DnaA is composed of largely independent domains — or modules — each contributing a partial, though essential, function to the proper functioning of the ‘holoprotein’.

DnaA protein binding to individual DnaA boxes in the Escherichia coli replication origin, oriC.

The formation of nucleoprotein complexes between the Escherichia coli initiator protein DnaA and the replication origin oriC was analysed in vitro by band‐shift assays and electron microscopy. DnaA protein binds equally well to linear and supercoiled oriC substrates as revealed by analysis of the binding preference to individual DnaA boxes (9‐mer repeats) in oriC, and by a competition band‐shift assay. DnaA box R4 (oriC positions 260–268) binds DnaA preferentially and in the oriC context with higher affinity than expected from its binding constant. This effect depends on oriC positions 249 to 274, is enhanced by the wild‐type sequence in the DnaA box R3 region, but is not dependent on Dam methylation or the curved DNA segment to the right of oriC. DnaA binds randomly to the DnaA boxes R1, M, R2 and R3 in oriC with no apparent cooperativity: the binding preference of DnaA to these sites was not altered for templates with mutated DnaA box R4. In the oriC context, DnaA box R1 binds DnaA with lower affinity than expected from its binding constant, i.e. the affinity is reduced to approximately that of DnaA box R2. Higher protein concentrations were required to observe binding to DnaA box M, making this low‐affinity site a novel candidate for a regulatory DnaA box.

Functional domains of DnaA proteins

Functional domains of the initiator protein DnaA of Escherichia coli have been defined. Domain 1, amino acids 1-86, is involved in oligomerization and in interaction with DnaB. Domain 2, aa 87-134, constitutes a flexible loop. Domain 3, aa 135-373, contains the binding site for ATP or ADP, the ATPase function, a second interaction site with DnaB, and is required for local DNA unwinding. Domain 4 is required and sufficient for specific binding to DNA. We show that there are three different types of cooperative interactions during the DNA binding of DnaA proteins from E. coli, Streptomyces lividans, and Thermus thermophilus: i) binding to distant binding sites; ii) binding to closely spaced binding sites; and iii) binding to non-canonical binding sites.

Bacterial replication initiator DnaA. Rules for DnaA binding and rolesof DnaA in origin unwinding and helicase loading

We review the processes leading to the structural modifications required for the initiation of replication in Escherichia coli, the conversion of the initial complex to the open complex, loading of helicase, and the assembly of two replication forks. Rules for the binding of DnaA to its binding sites are derived, and the properties of ATP-DnaA are described. We provide new data on cooperative interaction and dimerization of DnaA proteins of E. coli, Streptomyces and Thermus thermophilus, and on the stoichiometry of DnaA-oriC complexes of E. coli.

Analysis of the DNA‐binding domain of Escherichia coli DnaA protein

The DNA‐binding domain of the Escherichia coli DnaA protein is represented by the 94 C‐terminal amino acids (domain 4, aa 374–467). The isolated DNA‐binding domain acts as a functional repressor in vivo, as monitored with a mioC::lacZ translational fusion integrated into the chromosome of the indicator strain. In order to identify residues required for specific DNA binding, site‐directed and random PCR mutagenesis were performed, using the mioC::lacZ construct for selection. Mutations defective in DNA binding were found all over the DNA‐binding domain with some clustering in the basic loop region, within presumptive helix B and in a highly conserved region at the N‐terminus of presumptive helix C. Surface plasmon resonance (SPR) analysis revealed different binding classes of mutant proteins. No or severely reduced binding activity was demonstrated for amino acid substitutions at positions R399, R407, Q408, H434, T435, T436 and A440. Altered binding specificity was found for mutations in a 12 residue region close to the N‐terminus of helix C. The defects of the classical temperature sensitive mutants dnaA204, dnaA205 and dnaA211 result from instability of the proteins at higher temperatures. dnaX suppressors dnaA71 and dnaA721 map to the region close to helix C and bind DNA non‐specifically.

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 9u2003bp 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

The sequence requirements for a functional Escherichia coli replication origin are different for the chromosome and a minichromosome.

We have developed a simple three‐step method for transferring oriC mutations from plasmids to the Escherichia coli chromosome. Ten oriC mutations were used to replace the wild‐type chromosomal origin of a recBCsbcB host by recombination. The mutations were subsequently transferred to a wild‐type host by transduction. oriC mutants with a mutated DnaA box R1 were not obtained, suggesting that R1 is essential for chromosomal origin function. The other mutant strains showed the same growth rates, DNA contents and cell mass as wild‐type cells. Mutations in the left half of oriC, in DnaA boxes M, R2 or R3 or in the Fis or IHF binding sites caused moderate asynchrony of the initiation of chromosome replication, as measured by flow cytometry. In mutants with a scrambled DnaA box R4 or with a modified distance between DnaA boxes R3 and R4, initiations were severely asynchronous. Except for oriC14 and oriC21, mutated oriCs could not, or could only poorly, support minichromosome replication, whereas most of them supported chromosome replication, showing that the classical definition of a minimal oriC is not valid for chromosome replication. We present evidence that the functionality of certain mutated oriCs is far better on the chromosome than on a minichromosome.