Skip to main content

Regulation of DNA replication in Escherichia coli by DnaA

Final Report Summary - DNAA (Regulation of DNA replication in Escherichia coli by DnaA)

The gamma-proteobacterium Escherichia coli K-12 have for more than 40 years been the model organism of choice to study bacterial chromosome replication. The biochemistry of replication initiation from the unique origin of replication, governed by an accumulation of active ATP bound DnaA initiator protein is now well characterised, yet the mechanisms that set the frequency of initiation in vivo remain poorly understood. This is important because deviations from using the origin once and only once per cell cycle lead to aberrant cell growth and/or inviability. Such defects in higher eukaryotes are often associated with the onset of cancer. In prokaryote, the hope is to eventually be able to trigger specific events leading to failures in the DNA replication system in order to limit growth or kill bacterial pathogen. Initiation of replication is a highly regulated step in the E. coli cell cycle, and the frequency of initiation is controlled by cell cycle dependent variation in the availability of active initiator protein, DnaAATP. DnaA protein binds ATP and ADP with equal affinity. Because the ATP/ADP ratio is high in living cells, newly synthesised DnaA protein will mainly be in the ATP bound form. Initiation occurs when free DnaA ATP has accumulated above a certain threshold level. It induces formation of a DnaA-DNA nucleoprotein complex on oriC that promotes DNA strand opening and subsequent recruitment of the replication machinery. Following initiation, the new and hemimethylated origins are bound (sequestered) by the SeqA protein to prevent immediate re-initiation. During sequestration the level of free DnaA ATP is lowered by shutting down de novo synthesis of DnaA while titrating existing protein to binding sites outside oriC that are generated by replication. The activity of DnaA is reduced by the regulatory inactivation of DnaA (RIDA) mechanism which converts active DnaA ATP to inactive DnaA ADP by hydrolysis. RIDA involves a complex of two proteins: Hda and the DNA-loaded beta clamp of the DNA Pol. III holoenzyme. Together they stimulate the ATPase activity of DnaA to promote conversion of DnaA ATP to the inactive DnaA ADP. Mutations in hda result in accumulation of DnaA ATP, excessive initiation and inviability. Consequently, hda mutant cells carry compensatory mutations called hsm (hda suppressor mutation). We determined the nature of eight such hsm mutations by whole genome sequencing. These mutations are important to identify factors involved in negative regulation of the initiation frequency. One of our objectives was to characterise these hsm mutants. We discovered that the excessive initiation resulting from the loss of hda could be suppressed by several mechanisms. The first was by direct alteration of DnaA function: two hsm mutations were located in the dnaA gene. Importantly one of these mutations is located in a DnaA domain whose function was unknown and thought to be dispensable. The second mechanism consists of upregulation of pathways that normally repress initiation: this is the case for the hsm-1 mutation, which results in overproduction of SeqA. However, our results point to a more complex picture in which SeqA could act indirectly through factors yet to be found. A third mechanism involves rearrangement of the chromosome itself as is seen in two mutants (hsm-7 and hsm-8) where large chromosomal fragment are reoriented or duplicated. We speculate that interfering with the so called domain architecture of the chromosome would slow down the DNA replication and therefore limit the detrimental effects of aberrant number of replication forks. The forth group of suppressors appeared unrelated at first because the mutations are located in genes never associated with DNA replication before: stpA, iscU and frE. However, upon inspection of their transcriptional profile we discovered that they possess at least one thing in common: a deregulation of genes involved in dNTP biosynthesis. And this was the aim of our second project objective. It is generally assumed that a coupling exists between biosynthesis of DNA precursors (dNTP's) and the rate of DNA synthesis, which in bacteria is set by the initiation frequency. Although experimental evidence is scarce such a link may exist in bacteria and could be fulfilled by the nucleotide bound status of the DnaA protein, through expression of nucleotide biosynthetic genes. It was reported that overproduction of ribonucleotide reductase (RNR) suppresses overinitiation in Hda deficient cells. A tempting hypothesis is that these dNTPs determine the activity of DnaA and hence the initiation frequency from oriC. Cell cycle dependent variation in dNTP content would thus contribute to maintaining once-per-cell cycle replication initiation. To test this idea, we created strains where RNR expression was perturbed. Specifically, we removed binding sites for DnaA (DnaA is thought to contribute directly to the regulation of RNR expression) or a regulatory region involved in the cell cycle controlled expression from the nrdAB promoter. Cell cycle parameters of these strains such as the DNA content and mass were severely affected. Moreover, the effect of the mutations depended on medium composition. These results further support the hypothesis that the synthesis of dNTPs regulates the initiation of DNA replication but it also implies a new interconnection between media composition, dNTP pool size and DnaA function. Transcription profiling of hsm mutants also points in this direction. One mutation lies in the frE gene that is an activator of RNR. Our data indicates that key genes involved in the biosynthesis of purine and pyrimidine are down-regulated in this mutant.

The stpA and iscU hsm mutations results in down regulation of a transcriptional repressor of all ribonucleotide reductases in E. coli. Taken together, these results suggest that an imbalance in the proportions of various NTP's or dNTPs that negatively affect replication initiation from oriC. The last and most ambitious objective of our project was to determine the subcellular localisation of DnaA. Since attempts to localise DnaA using conventional GFP-tagging techniques had failed before our project began, we set out to use and develop a brand new technology that would enable us to visualise DnaA in living cells. The cornerstone of our technique lies in the ability to place a very small fluorescent tag virtually anywhere along a chosen protein sequence, thereby minimising the risk of affecting protein function. A fluorescent amino acid (CouAA) is encoded into a given sequence by use of an orthogonal tRNA / aminoacyl-tRNA synthetase pair that transfers the CouAA to a growing polypeptide chain at the amber codons (UAG), provided this is present present in the coding sequence. As a premise to demonstrate that this system can be effectively used as a means to visualise the in vivo subcellular location of a DnaA CouAA-labelled protein, we chose to label FtsZ and GroEL, which have been extensively studied both in vitro and in vivo. We could not only confirm previously known subcellular localisation pattern FtsZ and GroEL, but we could also show that these prototypes were fully functional both in vivo and in vitro. We created and expressed a DnaA protein in which the CouAA was encoded at amino acid number u8, named DnaA8CouAA.

We observed that DnaA8CouAA was localised uniformly throughout the cell. We confirmed that we specifically labelled the DnaA protein and that it was still capable of initiate replication. At the time, we started the project, two independent groups published the localisation of DnaA using GFP. While both group presented evidences that DnaA is generally distributed throughout the cell, they also presented significant differences. One presented evidences that DnaA is bound to chromosomal DNA and covers the entire nucleoid, while the other group suggests that DnaA forms a helical structure.

We believe that one plausible explanation for the different observation is the nature and placement of the fluorescent emitter. The GFP construct published by the two groups consist of a chimera where GFP replaces the second domain of DnaA. We and other recently demonstrate that the removal of this very domain can profoundly affect DnaA function. It is therefore possible that our observation reflect a situation closest to normal. Nevertheless, the project ended up by the development and validation of a new avant-garde technology to visualise intracellular protein location.