The initiation of chromosomal replication occurs only once during the cell cycle in both prokaryotes and eukaryotes. Initiation of chromosome replication is the first and tightly controlled step of a DNA synthesis. Bacterial chromosome replication is initiated at a single origin, oriC , by the initiator protein DnaA, which specifically interacts with 9-bp nonpalindromic sequences (DnaA boxes) at oriC . In Escherichia coli , a model organism used to study the mechanism of DNA replication and its regulation, the control of initiation relies on a reduction of the availability and/or activity of the two key elements, DnaA and the oriC region. This review summarizes recent research into the regulatory mechanisms of the initiation of chromosomal replication in bacteria, with emphasis on organisms other than E. coli .
The events involved in the initiation of chromosomal replication are similar in Eubacteria, eukaryotes, and Archea: replication starts with the binding of specific initiator protein(s) to DNA sites, termed origins, and results in the localized unwinding of the DNA duplex and the establishment of replication forks. In eukaryotes, chromosomes contain multiple start sites for DNA synthesis, and the initiator origin recognition complex (ORC) is a six-subunit heteromultimer that binds to the origin region. In contrast, bacteria replicate their chromosome(s) from a single replication origin ( oriC ), and the initiation of chromosome replication is mediated by a single initiator protein, DnaA, which specifically interacts with 9-bp nonpalindromic sequences (DnaA boxes) at oriC ( Messer et al. , 2002 ; Kaguni et al. , 2006 ). This process has been particularly well characterized in Escherichia coli . Twenty to 30 DnaA monomers interact with 11 binding sites. Three of these sites are high-affinity binding sites; the others [two 9-mer DnaA boxes, three 6-mer DnaA–ATP boxes ( Weigel et al. , 1997 ; Speck & Messer, 2001 ) and three 9-mer I sites ( Leonard & Grimwade, 2005 )] require oligomerization of DnaA. Erzberger (2002 , 2006) proposed a model of DnaA oligomerization at oriC based on the recently resolved crystal structure of the major part of Aquifex aeolicus DnaA (domains III and IV). In this model, DnaA monomers bound to DnaA boxes together with DnaA monomers oligomerize into right-handed filament; a newly created nucleoprotein complex is stabilized by specific protein–ATP interactions of adjacent DnaA monomers. Additional stability may be provided by domain I, which is responsible for self-oligomerization ( Weigel et al. , 1999 ; Felczak et al. , 2005 ). Wrapping of the oriC region around the DnaA filament promotes a local unwinding of an AT-rich region that leads to the formation of the open complex. Because there are flexible links between the respective DnaA domains responsible for oligomerization, ATP binding, and DNA binding, it is easy to imagine how the origins of other bacteria with different numbers and orientations of DnaA boxes adjust to their cognate DnaA proteins ( Zawilak-Pawlik et al. , 2005 ). As a result of DnaA–DnaB and DnaC–ssDNA interactions, the DnaB/DnaC helicase complex is loaded into the unwound origin region ( Konieczny et al. , 2003 ), and then DnaB loads DnaG ( Lu et al. , 1996 ; Tougu & Marians, 1996 ) and DNA polymerase III ( Kim et al. , 1996 ). The overall architecture of the eukaryotic ORC is very similar to that of DnaA oligomers ( Clarey et al. , 2006 ). This and the importance of ATP binding for ORC–origin interaction exemplify the global similarity of these basic processes. By analogy to the DNA-wrapping activity of DnaA, ORC together with Cdc6 prepares origins for helicase loading through mechanisms related to the established pathway of bacteria ( Clarey et al. , 2006 ).
Replication initiation has to occur at the correct time in the cell cycle, and any one origin must initiate once and only once per cell cycle ( Boye et al. , 2000 ; Messer, 2002 ; Kaguni et al. , 2006 ). In Eubacteria, eukaryotes, and, very likely, Archaea as well, replication is controlled at the initiation stage ( Maaloe & Kjeldgaard, 1966 ). Various mechanisms are involved in the regulation of this process. In Eubacteria, the control of initiation relies on a reduction of the availability and/or activity of both the DnaA protein and the oriC region at the various steps after initiation, for example before unwinding and/or immediately after the establishment of replication forks.
Regulatory mechanisms in E. coli – a short overview
The initiation of replication is controlled by affecting the assembly of the orisomes (protein– oriC complexes). Recent extensive studies have shown that the E. coli orisome structure is dynamic, changing in stages as it progresses through the cell cycle of this bacterium ( Cassler et al. , 1995 ; Leonard & Grimwade, 2005 ; Schaeffer et al. , 2005 ). In E. coli , several orisome components have been identified, including the histone-like DNA-binding proteins IHF and Fis, and other oriC -binding elements such as HU, Dpi, IciA, Cnu, Hha, Rob, SeqA, and ArcA ( Kim et al. , 2005 ). Orisome assembly is regulated by a dynamic interplay among these proteins; for example, Fis and IHF directly modulate the interaction of DnaA–ATP with its weaker binding sites, while HU modulates the binding of IHF to oriC and presumably enhances the ability of DnaA to unwind the origin ( Hwang & Kornberg, 1992 ; Ryan et al. , 2004 ). In contrast to IHF, IciA inhibits the unwinding of oriC . However, the contribution of IciA and other proteins, Dpi, Cnu, Hha, and Rob, to orisome assembly or disassembly during the cell cycle remains to be elucidated. The assembly of orisomes could also be affected by proteins that directly interact with DnaA. Recently, DiaA, a novel DnaA-binding protein crucial to ensuring the timely initiation of replication, was identified in E. coli ( Ishida et al. , 2004 ).
Three key negative regulation mechanisms preventing reinitiation from the newly replicated origins have been described in E. coli : (1) inhibition of DnaA activity, (2) titration of the free form of DnaA, and (3) sequestration of the oriC . Inactivation of DnaA protein occurs by conversion of the active initiator DnaA–ATP form to inactive DnaA–ADP, which is stimulated by the replisome elements, namely the DnaN sliding clamp of DNA polymerase III and Hda protein, and is called RIDA (regulatory inactivation of DnaA) ( Katayama et al. , 1998 ; Kato & Katayama, 2001 ; Gon et al. , 2006 ; Riber et al. , 2006 ). A second mechanism preventing reinitiation involves titration of DnaA protein by a cluster of high-affinity DnaA boxes (named datA , D na At itration), which reduces the level of DnaA shortly after this region is duplicated. The datA region is able to bind over 300 DnaA molecules ( Kitagawa et al. , 1998 ). In the third regulatory mechanism, the newly replicated and therefore hemimethylated oriC regions are sequestered by the binding of SeqA protein. SeqA recognizes GATC sequences overrepresented within oriC and prefers binding to hemimethylated over binding to fully or unmethylated oriC .
These three mechanisms of E. coli have been discussed in a number of excellent reviews ( Boye et al. , 2000 ; Katayama et al. , 2001 ; Messer et al. , 2002 ; Margolin & Bernander 2004 ; Camara et al. , 2005 ; Cunningham & Berger, 2005 ; Kato et al. 2005 ; Lobner-Olesen et al. , 2005 ; Kaguni et al. , 2006 ). Because much of what is know about the regulation of the initiation of bacterial chromosomal replication comes from studies of E. coli , this review focuses mainly on regulatory mechanisms in organisms other than E. coli ( Fig. 1 ).
Escherichia coli -like mechanisms in other organisms
The inactivation of DnaA–ATP by ATP hydrolysis is likely to take place in all bacteria possessing a dnaA gene. Besides being established for E. coli DnaA, ATPase activity has also been demonstrated for other bacterial DnaA proteins, including Bacillus subtilis ( Fukuoka et al. , 1990 ) , Helicobacter pylori (A. Zawilak-Pawlik unpublished results), Mycobacterium tuberculosis ( Yamamoto et al. , 2002 ; Madiraju et al. , 2006 ), Streptomyces coelicolor ( Majka et al. , 1997 ), Thermus thermophilus ( Schaper et al. , 2000 ), and Thermotoga maritima ( Ozaki et al. , 2006 ). So far, all sequenced dnaA genes encode an AAA+ATPase motif responsible for the binding and hydrolysis of ATP. Furthermore, this motif is also present in proteins that initiate chromosome replication in eukaryotes (three ORC proteins, namely Orc1p, Orc4p and Orc5p, and Cdc6) ( Speck et al. , 2005 ) and Archaea (Cdc6/Orc1) ( Robinson & Bell, 2005 ). All these proteins are replication-active in the ATP-bound form ( Lee and Bell 2000 ; Davey et al. , 2002 ). Thus, ATP binding and hydrolysis acts as a universal switch that regulates the initiation of replication in three domains of life. Little is known about the mechanism(s) responsible for the inactivation of the ATP-bound form of initiators. It should be noted that orthologues of Hda are present only in certain gammaproteobacterial genomes, suggesting the presence of different regulation systems in the replication initiation process of Eubacteria.
Titration of the DnaA protein by a cluster of high-affinity DnaA boxes localized outside oriC also appears to be involved in the regulation of chromosome replication in other bacteria. The S. coelicolor chromosome contains a cluster of high-affinity DnaA boxes in the vicinity of the oriC region ( Smulczyk-Krawczyszyn et al. , 2006 ). Deletion of the cluster caused more frequent chromosome replication and led to earlier colony maturation. In contrast, delivery of high-affinity DnaA boxes caused slow colony growth, presumably because of a reduction in the frequency of replication initiation. In silico analysis of bacterial chromosomes revealed that many of them contain at least two clusters of DnaA boxes in the vicinity of the oriC region ( Mackiewicz et al. , 2004 ). Thus the presence of additional clusters of DnaA boxes in chromosomes other than those of E. coli or S. coelicolor suggests that such control may be a common mechanism for many bacteria.
Up until now, sequestration of the oriC region has seemed to be a mechanism exclusively characteristic of E. coli and other enterobacteria. However, some recent data suggest that the initiation of replication of Agrobacterium tumefaciens ( Kahng & Shapiro, 2001 ), Brucella abortus , Caulobacter crescentus ( Stephens et al. , 1995 ), Rhizobium meliloti , and Rickettsia prowazekii , and probably of other organisms from α-subdivision bacteria, might involve an analogous system; these organisms possess a cell cycle-regulated CcrM DNA methyltransferase that recognizes the GANTC sequence ( Stephens et al. , 1996 ). Interestingly, in some bacteria that do not have the sequestration mechanism, such as B. subtilis and Streptomyces , minichromosomes are unstable, and only low copy numbers occur; in contrast to E. coli , minichromosomes of these organisms compete with chromosomes ( Zakrzewska-Czerwińska & Schrempf, 1992 ; Moriya et al. , 1999 ; Paulsson & Chattoraj, 2006 ; Smulczyk-Krawczyszyn et al. , 2006 ). From studies on these bacteria, it is becoming clear that their chromosomal replication control shares some similarities with that of low-copy-number plasmids, such as miniP1. In both cases, DnaA boxes or plasmid iterons (binding sites for initiator protein) serve as incompatibility elements ( Mukhopadhyay & Chattoraj, 2000 , Ogura et al. , 2001 , Park et al. , 2001 ). Perhaps handcuffing (origin pairing causing steric hindrance) may also apply to B. subtilis and Streptomyces replication control.
As in E. coli , the assembly of orisomes in other bacteria is affected by proteins that interact with oriC and/or DnaA. In C. crescentus , IHF and CtrA (see below) bind the oriC region specifically. Bacillus subtilis lacks genes encoding proteins homologous to E. coli accessory proteins, IHF, Hha, Fis, SeqA, Dam, but several novel proteins modulating orisome assembly have been identified in B. subtilis , including Spo0A, an oriC -binding protein (see below), DnaD and DnaB, which participate in loading the DnaC–DnaI helicase complex (equivalent to E. coli DnaBC) at oriC ( Bruand et al. , 2005 ; Zhang et al. , 2005 , 2006 ; Carneiro et al. , 2006 ), and YabA, interacting with DnaA and DnaN ( Noirot-Gros et al. , 2002 , 2006 ; Hayashi et al. , 2005 ). However, their exact biological function in B. subtilis orisome assembly is not yet fully understood. Recently, in H. pylori a novel, essential architectural component of the orisome, HobA, was found (A. Zawilak-Pawlik unpublished results).
In E. coli , DnaA protein, besides being the initiator, is also a transcription factor regulating the expression of genes that are involved in replication (e.g. mioC , nrd ), including its own gene; binding of the DnaA protein to DnaA boxes located within promoter regions influences gene expression ( Messer & Weigel, 1997 ). The DnaA protein also regulates gene expression in B. subtilis ( Goranov et al. , 2005 ), C. crescentus ( Hottes et al. , 2005 ), and, presumably, in other organisms whose promoter regions contain DnaA-binding motifs.
A global network relevant to regulating replication is likely to be more intricate in organisms that undergo a complex life cycle or in those that have to adapt to highly fluctuating environmental conditions. Under unfavourable conditions, the growth rate should be reduced and/or the bacteria should undergo morphological changes. In these organisms the decision ‘to replicate or not to replicate’ has to be precisely controlled at a number of levels. Bacteria, much like eukaryotic cells, coordinate cell division with DNA replication. Little is known about how the replication machinery coordinates its action with other cellular processes in variable environmental conditions. Adaptation in bacterial cells is often achieved through two-component signal-transduction systems, which consist of a sensor kinase and a response regulator. So far, only a few bacterial two-component signal-transduction systems involved in the regulation of replication initiation have been described. One of them is E. coli Arc (anoxic redox control), a two-component signal-transduction system that participates in regulating chromosomal initiation under anaerobic growth conditions ( Iuchi & Weiner, 1996 ). The Arc system regulates the expression of numerous operons in response to respiratory growth conditions. It consists of the ArcB, a transmembrane sensor kinase, and its cognate response regulator, ArcA. Anaerobic conditions that induce the Arc two-component signal-transduction system lead to a reduction in the growth rate. Lee (2001) demonstrated in vitro that ArcA∼P specifically binds the left part of the E. coli oriC region and prevents the formation of the open complex. Thus, oxygen depletion stress promotes the conversion of ArcA to ArcA∼P and its binding to oriC , which consequently reduces the frequency of chromosomal initiation to sustain a slow growth rate in adverse environmental conditions. Similarly, PhoB protein, a transcriptional regulator of the PhoB–PhoR two-component system regulating phosphate uptake, affects E. coli chromosome initiation. PhoB∼P, under reduced phosphate availability, activates iciA transcription ( Han et al. , 1999 ). IciA is a negative regulator of oriC unwinding in vitro ; hence PhoB∼P might reduce the initiation frequency during phosphate starvation.
Switching off replication by preventing a new round of replication must also occur at certain stages of the life cycles of bacteria that undergo cellular differentiation, for example the formation of morphologically different cells ( C. crescentus ) and/or the production of endospores or exospores ( B. subtilis or S. coelicolor ). An interesting example of a response regulator of the two-component system involved in the regulation of the cell cycle, including inhibition of chromosome replication, is CtrA (cell-cycle transcription regulator) in C. crescentus ( McAdams & Shapiro, 2003 ; McGrath et al. , 2004 ). This is a free-living bacterium in an aquatic environment that divides asymmetrically, generating two distinct cell types at each cell division: a stalked cell competent for DNA replication and a swarmer cell that is unable to initiate DNA replication until it differentiates into a stalked cell later in the cell cycle ( Crosson et al. , 2004 ; Brazhnik & Tyson, 2006 ; Holtzendorff et al. , 2006 ; Jensen et al. , 2006 ). In swarmer cells, CtrA∼P binds specifically to five sites located within the oriC region, preventing the formation of the replisome ( Quon et al. , 1998 ; Siam & Marczynski, 2000 ; Wortinger et al. , 2000 ; Marczynski & Shapiro, 2002 ). At the swarmer-to-stalked cell transition, CtrA is temporally degraded by the ClpXP protease, which releases the origin for replication initiation ( Jenal & Fuchs, 1998 ). Shortly after replication initiation, the proteolysis of CtrA is stopped and a positive transcriptional feedback loop results in the accumulation of new CtrA protein ( Domain et al. , 1999 ; Hung & Shapiro, 2002 ), thus preventing premature reinitiation of DNA replication ( Quon et al. , 1998 ). In C. crescentus , DnaA protein is also selectively targeted for proteolysis, but DnaA proteolysis uses a different mechanism from that of CtrA ( Gorbatyuk & Marczynski, 2005 ). Unlike the case for E. coli DnaA, the degradation of C. crescenus DnaA depends on cell-cycle- and nutrition-specific signalling; it takes place preferentially in swarmer cells. In C. crescentus , both proteins, DnaA and CtrA, regulate the transcriptions of multiple genes: DnaA controls the expression of genes encoding several replisome components, and CtrA controls the expression of many genes involved in flagella biogenesis and cell division. In addition to CtrA, a second master regulatory protein, GcrA, is involved in the cell-cycle regulation of C. crescentus ( Holtzendorff et al. , 2006 ). GcrA is present predominantly in stalked cells. GcrA also affects the expression of many genes: it inhibits dnaA expression and activates genes encoding components of the segregation machinery and the ctrA gene (p1 promoter) ( Holtzendorff et al. , 2004 ). Its expression is inhibited by CtrA and activated by DnaA protein ( Collier et al. , 2006 ).
At certain stages of life, usually in response to nutritional stress, bacteria can form dormant, nonreproductive bodies (e.g. spores). The formation of such temporarily inactive cells has to be preceded by the completion of a final replication round and by the prevention of a new round of replication at the initiation step. During sporulation of B. subtilis , a single cell divides asymmetrically (in contrast to vegetative growth) near one pole, producing a small endospore and a large mother cell that participates in the maturation of the spore and finally lyses to release it. In B. subtilis , Spo0A is a key transcriptional regulator controlling the entrance into sporulation. Spo0A belongs to a superfamily of phosphorylation-activated signal-transduction proteins that mediate adaptive responses to environmental or metabolic signals ( Baldus et al. , 1994 ; Burkholder et al. , 2001 ) and activate the transcription of crucial genes for the sporulation process. It has recently been demonstrated that Spo0A is involved in the regulation of replication frequency. In addition to being required for the onset of sporulation, Spo0A is a transcriptional activator/repressor that influences the expression of over 500 genes. Spo0A binds to so-called ‘0A-boxes’, which have been found not only within the promoter regions, but also within the B. subtilis oriC region, suggesting a novel function for the protein. Indeed, Castilla-Liorente (2006) showed that binding of Spo0A protein to ‘0A-boxes’, which overlap with functional DnaA-binding sites of the oriC region, prevents open complex formation.
Recent advances in bacterial cell biology have revealed that the nucleoid is a highly organized structure that undergoes dynamic changes through the cell cycle, for example during segregation ( Errington et al. , 2005 ). Chromosomal regions are organized into highly ordered structures that are placed in discrete spatial locations at specific times. It has been suggested that, in B. subtilis , Soj and Spo0J proteins (ParA and ParB homologues) required for chromosome segregation may be the negative regulator of replication initiation. In vivo , Spo0J binds to eight parS sites distributed around oriC ( Lee et al. , 2003 ; Murray et al. , 2006 ). The formation of a massive nucleoprotein complex that compacts the oriC region presumably prevents reinitiation from the newly replicated origins by reducing origin accessibility for the initiator protein DnaA ( Lee & Grossman, 2006 ). Indeed, deletion of soj and spo0J causes overinitiation of replication in B. subtilis ( Lee & Grossman, 2006 ).
Similarly, in Streptomyces , ParB, besides being the protein involved in segregation, may also regulate the initiation of replication. Streptomyces , which are known for their ability to produce many valuable antibiotics, are among the most striking examples of multicellular bacteria. Their hyphae grow by tip extension, forming a branched vegetative mycelium that consists of hyphal compartments containing multiple chromosomes; thus cross-wall formation is uncoupled from chromosome replication and segregation. During further growth, Streptomyces colonies form an aerial mycelium that develops into long chains of uninucleoidal exospores. At this stage, intensive replication in rapidly growing compartments is subsequently switched off as the chromosomes become condensed and segregated into spores. As in B. subtilis , it was shown that ParB is engaged, particularly during sporulation, in the formation of large nucleoprotein complexes encompassing oriC regions (twenty parS sequences are clustered around the origin) ( Jakimowicz et al. , 2002 , 2005 ). It has been postulated that compaction of the oriC region by ParB may prevent further rounds of replication.
Intracellular pathogens and endosymbionts
Little is known about the signalling pathways that link the facultative intracellular cell cycle of pathogens with the host environment. It has recently been shown that in M. tuberculosis the initiation of chromosome replication is regulated by the signal-transduction system MtrA–MtrB, which is activated by specific host–pathogen interactions. The M. tuberculosis MtrA response regulator affects chromosome replication in a phosphorylation-dependent manner by inducing M. tuberculosis dnaA expression. The dnaA promoter is a MtrA target, as confirmed by immunoprecipitation experiments using anti-MtrA antibodies ( Fol et al. , 2006 ). Elevating the intracellular levels of MtrA has no effect on bacterial growth in broth, while it prevents the proliferation of M. tuberculosis in macrophages and in mice lungs and spleens. In human macrophage cell lines, the transcript levels of dnaA were significantly increased ( c . 40-fold) in an mtrA overexpression strain relative to the wild type. Furthermore, the same phenotypes were observed when dnaA expression was artificially induced ( dnaA was under the control of an inducible promoter), which suggests that the overexpression of DnaA inhibits the proliferation of M. tuberculosis ( Hoskisson & Hutchings, 2006 ). Fol (2006) proposed that the proliferation of M. tuberculosis in vivo depends, in part, on the optimal ratio of phosphorylated to nonphosphorylated MtrA response regulator.
Obligatory intracellular parasites and endosymbionts existing continuously within the host or extremophiles living under extreme conditions (e.g. high temperature) do not experience the extreme environmental fluctuations encountered by free-living bacteria ( Moran et al. , 2002 ). Sequence analysis of their chromosomes has demonstrated that they have lost many genes and regulatory elements, including those involved in the regulation of replication initiation in free-living organisms.
In contrast to the case for pathogens, the proliferation of obligatory endosymbionts has to be beneficial for the host. Thus these organisms needed to adapt their replication in a balanced way so that their growth rates were coordinated with the development of their hosts. The development of a stable symbiosis with cytosolic bacteria might have required a more direct control of DNA replication of the symbionts by the host ( Gil et al. , 2003 ). In extreme cases, some obligatory insect endosymbionts, for example Wigglesworthia glossinidia (endosymbiont of the tsetse fly) ( Akman et al. , 2002 ), Blochmannia floridanus (carpenter ant) ( Gil et al. , 2003 ), Blochmannia pennsylvanicus (black carpenter ant) ( Degnan et al. , 2005 ), and Baumannia cicadellinicola (glassy-winged sharpshooter) ( Wu et al. , 2006 ), lost the dnaA gene ( Foster et al. , 2005 ). The lack of DnaA protein might allow the host to protect itself from overreplication of the bacterium in its cytosol ( Akman et al. , 2002 ). It cannot be excluded that other alternative replication initiation pathways based on priA , priC or recA genes ( Gil et al. , 2003 ; Wu et al. , 2006 ) – commonly used in E. coli to re-establish replication at damage sites – might be used in these endosymbionts. Although Blochmannia also lacks priA, priC and recA genes, it contains recBCD genes, which may play some role in the initiation of replication ( Wu et al. , 2006 ). Buchnera is an interesting endocellular bacterial symbiont of aphids, which, in contrast to Wiggelsworthia and Blochmannia , retains the dnaA gene to initiate replication (in B. aphidicola , only two DnaA boxes were identified within its putative oriC region; Mackiewicz et al. , 2004 ). However, it should be noted that Buchnera resides in vacuole-like organelles, while Wiggelsworthia and Blochmannia directly contact the host cytoplasm.
Organisms that possess more than one chromosome
Not all bacteria have a single chromosome: some bacteria have multiple chromosomes; for example, Vibrio cholerae possesses two chromosomes, while Burkholderia cepacia has three chromosomes. The control of replication in bacteria with multiple chromosomes is much less well understood than that in organisms with a single chromosome. Among the organisms possessing more than one chromosome, the initiation of replication and the mechanism(s) regulating this process have been studied so far only in V. cholerae . This bacterium, the causative agent of cholera, has two differently sized circular chromosomes, namely chromosome I (chrI) and chromosome II (chrII), of 2.96 and 1.07 Mbp, respectively ( Heidelberg et al. , 2000 ). Vibrio cholerae lives primarily as a free organism in aquatic environments and associates with the host only during short outbreaks. Egan & Waldor (2003) suggested that a bipartite genomic arrangement may provide an evolutionary advantage by facilitating a faster replication time or by allowing chromosome-specific replication control in certain environments. However, so far, none of these hypotheses has been experimentally proven. Although the two chromosomes replicate synchronously, they exhibit distinct replication requirements ( Egan & Waldor, 2003 ; Egan et al. , 2004 , 2005 ; Duigou et al. , 2006 ). The structure of oriCI resembles that of E. coli oriC , whereas oriCII shares some features with certain plasmid replicons. The functional oriCII requires internal 12-bp repeats and two hypothetical genes that flank the origin. One of these genes encodes the protein RctB that specifically binds oriCII . DnaA and RctB independently control replication initiation of the two chromosomes chrI and chrII, respectively. Overproduction of DnaA or RctB protein promoted exclusively overinitiation of chromosome I or chromosome II, respectively. The distinct replication requirements of the two origins may minimize competition between chromosomes, ensuring the maintenance of the divided genome, but, on the other hand, raise the question of how the regulation of replication initiation of two chromosomes is coordinated. Although oriCII has very little sequence similarity to oriCI , it contains a single DnaA box and an overrepresentation of GATC sequences, targets for the Dam methyltransferase. Although DnaA protein does not initiate replication of the chromosome II, its minimum concentration is required for the initiation of chromosome II replication. A similar situation is observed in the case of the replication of certain plasmids, including RK2, where DnaA is known to promote their replication, although they encode their own initiators, such as TrfA ( Duigou et al. , 2006 ). The overrepresentation of methylation motifs in both origins, oriCI and oriCII , of V. cholerae and the presence of seqA and dam genes suggest that these factors may mediate the coordination of replication of two chromosomes in a similar manner to sequestration of the E. coli origin. Additional, so far unknown, elements probably coordinate the activity of both initiators, DnaA and RctB.
Streptomyces and Rhizobium belong to the intriguing group of bacteria whose cells are multinucleoidal at certain stage(s) of the cell cycle. Thus these bacteria do not obey the once-and-only-once doctrine of DNA replication at particular life stages. This characteristic feature raises interesting questions on how the replication of multiple chromosomes within a single cell/compartment is regulated. Very little is known about the synchronization and regulation of chromosome replication in the multinucleoidal compartments of Streptomyces . Recent data show that chromosome replication appears to be asynchronous within a single compartment; only selected chromosomes undergo replication at one time ( Ruban-Osmialowska et al. , 2006 ). Rhizobium endosymbiotic soil bacteria can incite root nodule formation in certain leguminous plants. After infection of plant cells, the free-living bacteria are converted into nitrogen-fixing bacteroids. The transformation includes an elongation of the cells and repeated chromosome replication (endoreduplication) without cytokinesis, leading to the multinucleate cells. It has been demonstrated that plant factors present in the nodules trigger endoreduplication ( Mergaert et al. , 2006 ).
So far, Rhizobium and M. tuberculosis are the only two examples for host cells–plant cells and human macrophages, respectively – capable of triggering bacterial replication frequency.
Much of what we know about the molecular mechanisms of the regulation of bacterial chromosome replication comes from studies of E. coli . So far, not many factors involved in the regulatory network of the replication of organisms other than E. coli have been identified. However, those identified have demonstrated that the diversity in the mechanisms involved in the regulation of replication initiation results from the different life cycles and lifestyles of the microorganisms. Comparative analysis of bacterial genomes together with expression profiles of their genes under different conditions and the application of high-throughput two-hybrid systems ( E. coli or yeast) for searching interacting proteins will allow us to identify new regulatory mechanisms in the near future. The elements involved in the novel mechanisms may provide valuable drug and vaccine targets against bacterial pathogens.
This work was supported by the Ministry of Science and Higher Education (grant 2P04A 054 29). D.J. acknowledges support from the Marie Curie Reintegration Grant MERG-6-CT-2005-014851. A. Z.-P. was supported by a Marie Curie Intra-European Fellowship within the 6th European Community Framework Programme. J.Z.-C., A.Z.-P. and D.J. acknowledge support from the Scientific and Technological International Cooperation Joint Project Polonium.