Where does bacterial replication start? Rules for predicting the oriC
http://www.100md.com
《核酸研究医学期刊》
Department of Genomics, Institute of Genetics and Microbiology, University of Wrocaw, Przybyszewskiego 63/77, 51-148 Wrocaw, Poland, 1 Ludwik Hirszfeld Institute of Immunology and Experimental Therapy, Weigla 12, 53-114 Wrocaw, Poland and 2 Institute of Physics, University of Zielona Góra, Szafrana 4A, 65-516 Zielona Góra, Poland
* To whom correspondence should be addressed. Tel: +48 71 337 11 72; Fax: +48 71 337 13 82; Email: zakrzew@iitd.pan.wroc.pl
ABSTRACT
Three methods, based on DNA asymmetry, the distribution of DnaA boxes and dnaA gene location, were applied to identify the putative replication origins in 120 chromosomes. The chromosomes were classified according to the agreement of these methods and the applicability of these methods was evaluated. DNA asymmetry is the most universal method of putative oriC identification in bacterial chromosomes, but it should be applied together with other methods to achieve better prediction. The three methods identify the same region as a putative origin in all Bacilli and Clostridia, many Actinobacteria and Proteobacteria. The organization of clusters of DnaA boxes was analysed in detail. For 76 chromosomes, a DNA fragment containing multiple DnaA boxes was identified as a putative origin region. Most bacterial chromosomes exhibit an overrepresentation of DnaA boxes; many of them contain at least two clusters of DnaA boxes in the vicinity of the oriC region. The additional clusters of DnaA boxes are probably involved in controlling replication initiation. Surprisingly, the characteristic features of the initiation of replication, i.e. a cluster of DnaA boxes, a dnaA gene and a switch in asymmetry, were not found in some of the analysed chromosomes, particularly those of obligatory intracellular parasites or endosymbionts. This is presumably connected with many mechanisms disturbing DNA asymmetry, translocation or disappearance of the dnaA gene and decay of the Escherichia coli perfect DnaA box pattern.
INTRODUCTION
As replication is controlled at the initiation stage, the events that occur at the replication origin play a central role in the cell cycle. The initiation of DNA replication is a complex process involving several regulated steps: (i) binding the initiator protein(s) to sites located within the origin region (ori); (ii) local unwinding of the ori region; and (iii) loading the DNA helicase and other proteins required to form replication forks (1,2). Chromosomal replication starts from one (Bacteria and Archaea, e.g. Pyroccocus abyssi) (3), two (Archaea, e.g. Sulfolobus solfataricus) (4) or multiple replication origins (Eukaryota) . In bacteria, chromosomal replication initiated at the oriC region proceeds bi-directionally until the replication forks reach the termination site, terC (in the case of circular chromosomes) or chromosomal ends (in the case of linear chromosomes) (6). The initiation of bacterial chromosome replication is mediated by the initiator DnaA protein, which interacts with repetitive non-palindromic nonamer sequences, the DnaA boxes, located within the oriC region. Among bacteria, the initiation of replication is best understood in Escherichia coli (7). The binding of 10 to 20 DnaA protein molecules to five DnaA boxes promotes unwinding within the AT-rich region of oriC.
The sequences of bacterial oriC regions are conserved only among closely related organisms. Bacterial replication origins vary in sizes (from 200 up to 1000 bp), but (nearly) all contain several DnaA boxes and an AT-rich region. A cluster of four or more DnaA boxes is an indication of a functional chromosomal origin. However, only one and two DnaA boxes, respectively, were identified within the functional oriC regions of Caulobacter crescentus (8) and Coxiella burnetti (9). The oriC region is always located within the intergenic region, frequently within the rnpA–rmpH–dnaA–dnaN–recF–gyrB–rnpA genes cluster, usually next to the dnaA gene.
During the last few years, the genomes of approximately 150 bacteria have been completely sequenced, and there are approximately 500 ongoing bacterial genome projects. Thus, within a decade, the genomes of nearly every significant bacterial species will have been sequenced. Since the events that occur at the replication origin are central to the process regulating DNA replication and the cell cycle, it is important to define the rules required for a proper oriC localization within analysed genomes.
Initially, the oriC regions were identified by direct measurement of the replication time of various genetic markers, by construction of minichromosomes (plasmids containing oriC instead of their own replication origin) or by two-dimensional gel electrophoresis. So far, however, only 11 oriC regions have been shown in vivo to be functional. Now that long stretches or the complete sequences of bacterial chromosomes have become available it is possible to identify a potential oriC region in silico. Lobry (10,11) was the first to find bias (asymmetry) in nucleotide composition in some bacterial chromosomes. He noticed that the asymmetry changes its polarity at the origin and the terminus of chromosome replication, where DNA strands change their character from leading to lagging and vice versa. This asymmetry is usually measured as the normalized difference in the content of complementary nucleotides . This observation was then confirmed by other researchers for many bacterial chromosomes . There are many reasons for this asymmetry, but the main one seems to be an asymmetric mutational pressure (different nucleotide substitution patterns) associated with the replication of the leading and lagging strands . Since DNA strands change their mode of synthesis from lagging to leading in the oriC region, it is possible to identify the region in silico by analysis of chromosome asymmetry, using, for instance, the computer program Oriloc created by Frank and Lobry (23). This method is commonly used to identify the putative oriC region within chromosomes, particularly before experimental analyses (24–26). However, the switch in asymmetry does not always exactly correspond to the position of a functional replication origin; e.g. the in silico predicted origin in Helicobacter pylori 26695 was located 2500 bp away form the functional one (26). Therefore, in order to locate the replication origin precisely, GC skew analysis should be accomplished by additional analyses to locate the elements involved in the initiation of chromosomal replication: cluster(s) of DnaA boxes and the dnaA gene. Searching for a cluster of DnaA boxes requires first of all a proper definition of the DnaA box. The perfect sequence for the DnaA box of a model organism, E.coli, is 5'-TTATCCACA-3'. However, the preferred DnaA box sequence for other bacteria differs slightly from that of E.coli; it depends on the G + C content of the analysed chromosome and/or binding specificity of a given DnaA protein (for details see Results and Discussion).
The aim of this study is to evaluate the universality of three different methods of oriC identification (DNA asymmetry, DnaA box distribution and dnaA gene location) in their application to the analysis of completely sequenced bacterial chromosomes. We have tried to group the analysed genomes according to the concordance of these methods and to analyse the problem from the phylogenetic point of view.
MATERIALS AND METHODS
The analyses were performed on 120 complete sequences of bacterial chromosomes representing different taxonomic groups (see Supplementary Table 1). The chromosomal sequences were downloaded from the Genbank (ftp://www.ncbi.nlm.nih.gov). The assumed location of the dnaA gene on the chromosome was that of the database annotations. DNA asymmetry was analysed by DNA walks describing the asymmetry between differently replicating DNA strands . This method differs from that originally proposed (10,11,13), as it cumulates local deviations of a parameter of asymmetry ( or ) from the average value specific to the whole chromosome. This method eliminates the global compositional trend of the whole chromosome and smoothes random fluctuations. By this method the single origin and terminus of replication (if they are present) are easily recognized as the two main extrema of the DNA walk, i.e. where asymmetry changes its sign (Figure 1; Supplementary Figure 6).
Figure 1. DNA asymmetry, DnaA box distribution and dnaA gene location on bacterial chromosomes. x-axis—chromosome coordinates in base pairs, in (A–E and H) beginning and ending in the presumed terminus of replication (according to the DNA asymmetry). The chromosomal coordinates in (F and G) are as in the databases. The grey line and the right y-axis show the DNA walk describing DNA asymmetry (asymmetry for in (F); asymmetry for for the other chromosomes). Black peaks with diamonds and the left y-axis show the DnaA box distribution expressed by the b value. The short vertical line indicates the dnaA gene location.
In order to identify the potential cluster(s) of DnaA boxes we looked for all possible 28 nonamers which differed by no more than one position from the perfect DnaA box: TTATCCACA. The density of the box distribution on the chromosome was presented on charts by b = 1/d, calculated for each box, where d is a sum of the distances of the given box to its two adjacent boxes. The b value corresponds to the density of boxes in the given region of the chromosome, with a high value of b indicating a cluster of three boxes.
To visualize the correlation of the three methods of oriC identification (dnaA gene location, DNA asymmetry and DnaA box distribution), in the charts presenting these analyses (Figures 1–3; Supplementary Figure 6), the x-coordinate begins and ends in the presumed (according to DNA asymmetry) terminus of replication. Then the presumed oriC region is usually located near the middle of the chart. This was done for all circular chromosomes showing two distinct global extrema of DNA asymmetry. The charts for linear chromosomes and chromosomes without distinct global extrema of DNA asymmetry have the original sequence coordinates on the x-axis. In the charts, the DNA walk is shown only for the one parameter of asymmetry ( or ) which exhibited the strongest and most distinct DNA asymmetry.
Figure 2. DNA asymmetry in , DnaA box distribution, and dnaA gene location on the H.pylori J99 chromosome. (A) Distribution of DnaA boxes which differ in more than one position from the perfect DnaA box TTATCCACA; (B) distribution of DnaA boxes experimentally proved to be functional in this genome according to Zawilak et al. (26). The x-axis describing the coordinate position on the chromosome in base pairs begins and ends in the presumed (according to DNA asymmetry) terminus of replication. See Figure 1 for further explanation.
Figure 3. DNA asymmetry in , DnaA box' distribution and dnaA gene location in the oriC region. Example of single cluster of DnaA boxes (A) and a few separated clusters in the region (B). See Figure 1 for further explanation.
Supplementary material
Table 1 (general characteristics of the analysed chromosomes); Table 2 (locations of boxes and their patterns in the presumed oriC region); and Figure 6 (DNA asymmetry, DnaA box distribution and dnaA gene location on 120 bacterial chromosomes).
RESULTS AND DISCUSSION
General remarks
In order to elucidate the rules for predicting the origin of bacterial chromosome replication, we performed a comprehensive in silico analysis of 112 complete genome sequences of bacteria representing different taxonomic groups (Supplementary Table 1). We analysed the DNA asymmetry and the presence of elements involved in the initiation of replication (DnaA boxes and dnaA gene). Most (101) of the analysed bacteria possess a single circular chromosome, while three have a linear chromosome. Eight organisms contain two chromosomes (either with both circular or with one circular and the other linear).
A distinct DNA asymmetry between the leading and lagging strands is present in 93% of the analysed chromosomes (Supplementary Table 1). Such asymmetry is not present in the chromosomes of Aquifex aeolicus, Bifidobacterium longum, Deinococcus radiodurans (chromosome II), Gloeobacter violaceus, Nostoc sp., Phytoplasma asteris, Streptomyces avermitilis, Synechocystis sp. and Wolbachia pipientis. In most of the chromosomes stronger asymmetry was observed for the parameter than for .
dnaA genes have been found in all bacterial genomes studied so far, with the exception of the Wigglesworthia glossinidia and Blochmannia floridanus genomes (which are reduced in size). In addition, the chromosomes of both endosymbionts do not contain any clusters of DnaA boxes. The lack of typical bacterial elements involved in the initiation of replication may reflect the extreme dependency of W.glossinidia and B.floridanus on the host genome functions. These bacteria are freely located in the cytosol of bacteriocytes. Bacteria in the cytosol might be a potential danger for the host cell and more difficult to control. The development of a stable symbiosis with cytosolic bacteria probably requires more direct control of DNA replication of the symbionts by the host, involving the loss of the dnaA gene (27).
If a genome consists of two chromosomes, one of them lacks the dnaA gene (Supplementary Table 1). Two dnaA genes are present in chromosomes of the following organisms: Chlamydia muridarum, Chlamydia trachomatis, Chlamydia pneumoniae, Chlamydia caviae and Pirellula sp. However, it is not yet known whether only one of the dnaA genes or both are transcriptionally active. In Chlamydia it cannot be excluded that two dnaA genes are differentially transcribed during the organism's unusual intracellular developmental life cycle. In Brucella, the additional dnaA gene present on the larger chromosome was probably transferred from the smaller chromosome (via interchromosomal exchange).
As expected, DnaA boxes (differing in at the most one position from the perfect E.coli DnaA box TTATCCACA) were found in all chromosomes. Interestingly, the DnaA boxes were overrepresented (compared with their random occurrence expected from the base composition of a given chromosome) in 70% of the analysed chromosomes. In this group of chromosomes, the overrepresentation was statistically significant (with p < 0.05 in 72% cases). We assume that a potential cluster of DnaA boxes consists of a set of three boxes with an average distance between them of <100 bp (the distance 100 bp corresponds to a b value of 0.005 on the left y-axis in Figures 1–3). Such clusters (located anywhere on a chromosome) were present in 83% of chromosomes (85% of species).
Evaluation of the three methods in oriC identification
We assumed that the methods will identify the same region as the putative origin of replication if the distances between the location of the cluster of DnaA boxes and/or the dnaA gene and/or the position of the extreme of the DNA walk are <1% of the chromosome's length. Some doubtful cases (e.g. when the distance was >1% and <7% of chromosome length) were assigned by using parentheses and described by a suitable notation (see Supplementary Table 1). The assumed 1% limit corresponds to the distance between the dnaA gene and the experimentally identified oriC region in E.coli. In most cases, the discrepancy between the three methods results from fluctuation of DNA asymmetry in the extreme in the local scale (see below for potential causes of the fluctuations).
Taking these assumptions into account, we have classified the chromosomes into five distinct groups, depending on the co-localization of the putative oriC indicated by the three methods (the abbreviations mean: a—method of DNA asymmetry; b—DnaA box distribution; d—dnaA gene location):
abd: all three methods identify the same region as oriC—67 chromosomes (Figure 1A);
ab: the extreme of the DNA walk (asymmetry) is adjacent to the cluster of DnaA boxes—15 chromosomes (Figure 1B);
ad: the extreme of the DNA walk (asymmetry) corresponds to the location of the dnaA gene—18 chromosomes (Figure 1C);
bd: the dnaA gene is in the vicinity of the cluster of DnaA boxes—3 chromosomes (Figure 1D);
O: each method identifies a different region—17 chromosomes (Figure 1E).
The close vicinity of these three features suggests that the analysed region contains a functional origin. So far, only 11 of the putative oriC regions have been shown in vivo to be functional. Eight of them belong to the first group, abd; their in silico predicted oriC regions co-localized the functional origins.
The three methods applied to the identification of oriC agree rather well. In 55% of cases all methods indicate the same region. The most universal method seems to be DNA asymmetry, which was in agreement with at least one of the other methods in 83% of cases. There are only three cases (bd) where the extreme of DNA asymmetry is distant from both the cluster of DnaA boxes and the dnaA gene. In some cases there is imperfect agreement between the location of the cluster of DnaA boxes and/or the dnaA gene and DNA asymmetry, which performs some fluctuations near the global extreme. The asymmetry is probably disturbed by other phenomena (such as transcription, chromosome rearrangements, oligomers and gene distributions . In some cases, the observed shift in the extreme of asymmetry from the cluster of DnaA boxes and the dnaA gene may reflect a very recent translocation of the origin region as well. In this case, replication-associated mutational pressure probably had not enough time to erase the old compositional bias and introduce the new one. The smooth global extremes observed in the chromosomes of Mesorhizobium loti (Figure 1D) and Nitrosomonas europaea may suggest such a possibility.
In the chromosomes grouped in ab and ab(d), the dnaA gene was probably transferred to the other region of the chromosome. This may indicate that a close proximity of the dnaA gene to the origin region is not necessary for its proper functioning (its product can act in trans). Furthermore, it has been reported that mutants of Synechocystis sp. with inactivated dnaA gene were viable (28). It cannot be excluded that replication of this chromosome is initiated by another protein. However, in most cases the dnaA gene is in the vicinity of the oriC region. The reason for the close proximity of both elements is not clear. It has been noted that genomic rearrangements rarely occur in the close vicinity of the origins of replication (5). It is possible that the juxtaposition ensures that the DnaA protein is able to associate with the origin as soon as it is synthesized. Alternatively, the proximity to the origin minimizes the possibility of disruption of the linkage between the well-cooperating dnaA gene–oriC region system.
In some cases , a cluster of DnaA boxes could not be identified in the entire chromosome. However, it must be noted that the preferred DnaA box sequences of certain bacteria may differ slightly from the E.coli perfect DnaA box (TTATCCACA), e.g. in high-G + C organisms such as Micrococcus, Mycobacterium or Streptomyces (GC > 70%), the third position is substituted by G or C. In addition, our own recent experiments have demonstrated that particular DnaA proteins may exhibit different specificity towards DNA than that from E.coli, e.g. the affinity of the DnaA protein from H.pylori to the TCATTCACA sequence (two mismatches) is higher than to the E.coli perfect box. Therefore, the possibility of a different consensus sequence for the DnaA box (and/or binding specificity of DnaA protein) should be considered when the search of a putative DnaA protein binding site is performed. For example, our first search for DnaA box motifs whose sequences differ by no more than one base from the E.coli perfect box failed to find a single cluster of DnaA boxes in the entire H.pylori chromosome (Figure 2A). However, when we looked for DnaA box motifs with two to three mismatches, we found a cluster of DnaA boxes upstream of the dnaA gene (Figure 2B). This region, designated as the H.pylori putative origin of replication, is the only place on the chromosome that contains a cluster of DnaA boxes. The H.pylori DnaA exhibits a high affinity to this region (29). There are also many other examples of origin regions that contain ‘species-specific’ DnaA boxes, e.g. Borrelia burgdorferi, TTTTAAACA, TTATCAAAA, TTTTAAAAA (24); Thermotoga maritima, AAACCTACCACC (30); Caulobacter crescentus, TGATCCACA (8); and Prochlorococcus marinus, TTCCACA (31).
The chromosomes that show asymmetry with more than two distinct extrema are supposed to be replicated from more than one oriC. However, we found in some of them a correlation between the main local extreme of asymmetry and the location of the cluster of DnaA boxes and/or the dnaA gene: Aquifex aeolicus, Bifidobacterium longum, Gloeobacter violaceus, Nostoc sp. (Figure 1F) and Synechocystis sp. The region may be suspected to be the main (the strongest or earlier activated) origin of replication. However, there are many phenomena that influence the course of DNA asymmetry which may disturb the asymmetry introduced by replication-associated mutational pressure . Only experimental studies can definitely solve this problem. Interestingly, in the linear chromosome of S.avermitilis there are two distinct minima suspected of being the origin region (Figure 1G). One of these strongly correlates with the location of distinct clusters of DnaA boxes and the dnaA gene. A few DnaA boxes were also found to be close to the second minimum, but they were rather scattered. The arrangement of the DnaA boxes in close vicinity of the dnaA gene is the same as within the functional oriC region of Streptomyces coelicolor. Interestingly, despite selective pressures or the advantage of central positioning of the origin (to finish replication synchronously), the putative oriC of S.avermitilis is shifted 800 000 bp (10% of the chromosome's length) from the centre. It has been proposed that linear Streptomyces chromosomes resulted from recombination between a linear plasmid and an ancestral circular chromosome (32). Probably, in the case of S.avermitilis, linearization events resulted in the formation of chromosomal arms with different lengths (the oriC is asymmetrically positioned) and therefore the second minimum is observed (see Figure 1G).
Lack of any agreement between the three methods was usually observed in the second chromosomes (the smaller ones). In chromosome II of Brucella suis the extreme of DNA asymmetry correlates with the location of repC, a typical plasmid replication gene. It has been suggested that the chromosome of B.suis is derived from a megaplasmid that was captured by ancestral Brucella (33). Recent experimental studies on the initiation of replication of two Vibrio cholerae chromosomes revealed distinct replication requirements for the two chromosomes. The functional origins, oriCIVC and oriCIIVC, of V.cholerae chromosomes are unrelated; the structure of oriCIVC resembles that of E.coli, whereas oriCIIVC shares some features with certain plasmid replicons (34). The distinct replication requirements of the two chromosomes may minimize competition and thereby help ensure the maintenance of the divided genome. The situation is similar to the observed incompatibility of two plasmids which require identical replication initiation factors.
In organisms which contain a single chromosome and belong to group O, the initiation of chromosome replication is probably also mediated by another protein(s).
Distribution, organization and composition of DnaA boxes
We have found in numerous analysed chromosomes (Supplementary Table 1) that the clusters of DnaA boxes are not distributed randomly in a chromosome, but are densely packed within a relatively short distance around the putative oriC region (Figure 1H). The distance between the distal cluster of DnaA boxes and the oriC varies from 27 to 800 kb (this being up to 23% of chromosome length). The DnaA boxes, most probably involved in initiation of replication, were identified in 76 chromosomes. Detailed data (locations of these boxes and their sequence patterns) are included in the Supplementary Material (Supplementary Table 2), whereas in Table 1 we present the general organization of the clusters (number of boxes in a given cluster, mean distance between boxes in the cluster and distances between clusters). 41 chromosomes contain at least two separated clusters of DnaA boxes near the putative oriC region. In most cases, the distance between the clusters of DnaA boxes is only a few kilobase pairs (up to 7 kb). Longer distances are in Neisseria (12–13 kb) and Actinobacteria (mycobacteria, 15–150 kb; S.coelicolor and Corynebacterium efficiens, 30 kb; Corynebacterium diphtheriae, 25 kb).
Table 1. Organization of the clusters of DnaA boxes in the putative oriC regions
There are a few possible explanations for the occurrence of many clusters of DnaA boxes. One explanation assumes that more than one cluster participates in the initiation of replication. In several chromosomes which have at least two clusters of DnaA boxes in the putative origin region, we observed two local extrema and a plateau between them in the DNA walk describing asymmetry, e.g. Staphylococcus epidermidis (Figure 3, see for comparison the V-shaped asymmetry of the Haemophilus ducreyi chromosome). This may suggest that replication starts from one or both sites. In this case, the region corresponding to the plateau would be replicated sometimes as the leading strand and sometimes as the lagging strand. However, it has been experimentally demonstrated that the single Bacillus subtilis oriC region is composed of two DnaA clusters (separated by the dnaA gene, 1500 bp), which are both required for a functional origin. The unwinding of DNA occurs at the 27 bp AT-rich cluster in the vicinity of dnaN gene. In Mycobacterium (Mycobacterium avium and Mycobacterium tuberculosis), the dnaA gene is also flanked by two clusters of DnaA boxes. However, only one of them acts as oriC. Interestingly, in M.tuberculosis the 3' region of the dnaA gene exhibits oriC activity, while in M.avium the 5' region functions as oriC (35).
The second explanation assumes that the clusters of DnaA boxes (which are not involved in the initiation of replication) may contribute to the negative control of replication by reducing the level of free DnaA protein. Since DnaA proteins interact cooperatively with closely spaced DnaA boxes, the clusters of boxes can bind more molecules than single boxes dispersed randomly along the chromosome. In E.coli there are five regions of the chromosome which bind DnaA protein (36). DnaA protein exhibits particularly high affinity towards one of them, datA, which is located close to the oriC. The datA region has been suggested as playing a pivotal role in controlling replication initiation; it can bind eight times more DnaA protein molecules than the oriC region (37,38). Thus, shortly after initiation of replication, the datA region is duplicated and therefore is able to reduce significantly free DnaA below the level that provokes over-initiation. The presence of many clusters of DnaA boxes near the oriC region in chromosomes other than that of E.coli suggests that such replication control may be a common mechanism for bacteria.
Table 2 presents the general (mean) composition of DnaA boxes, counted only for boxes found in the most probable oriC regions in 76 chromosomes. The most variable are the third and the fifth positions in the consensus. This is in agreement with the recently established crystal structure of the E.coli DnaA binding domain in complex with the perfect DnaA box, which showed that the nucleotide at the fifth position does not play an essential role in the interaction with the DnaA protein. We observed statistically significant correlation (r = 0.62) between G + C content of the DnaA boxes and G + C content counted for the given genome (Figure 4). Assuming that global genomic G + C reflects, to some extent, the composition generated by mutational pressure (39), one can assume that composition of the DnaA boxes may be moderated by mutational pressure and that some changes may be neutral. Actually, a sequence comparison of DnaA boxes derived from the origins of organisms with high G + C content revealed that G is frequently present instead of A at the third position of the DnaA box. Bases in the second, fourth, seventh, eighth and ninth positions are highly conserved. Extensive binding studies on E.coli revealed that these positions are of particular importance (40,7).
Table 2. The composition of DnaA boxes derived from the putative origins identified in 76 chromosomes
Figure 4. Correlation between G + C content of DnaA boxes (derived from the putative origins identified in 76 chromosomes) and genomic G + C content.
Phylogenetic approach
We compared the phylogenetic (taxonomic) status of the analysed organisms with their classification based on DNA asymmetry, DnaA box distribution, and dnaA gene location (see Supplementary Table 1 and Figure 5). Strains belonging to the same species or even the same genus do not exhibit significant differences in DNA asymmetry, distribution of DnaA boxes and dnaA gene location on the chromosome. There are a few exceptions: (i) among Mycoplasma genomes, only Mycoplasma pulmonis shows full agreement of the three methods (it belongs to group abd); (ii) the genome of Bordetella pertussis shows more rough asymmetry (with some local extrema) than other Bordetella genomes; (iii) Haemophilus ducreyi and Haemophilus influenzae were assigned to different groups, ad and O, respectively; (iv) Streptomyces genomes differ in DNA asymmetry; and (v) Treponema denticola shows some shift of the dnaA gene compared with Treponema pallidum.
Figure 5. The 16S rRNA phylogenetic tree for 112 analysed bacteria and their classification based on concordance of methods of oriC identification: DNA asymmetry, DnaA box distribution and the dnaA gene location. The tree was built with the minimum evolution method (ME) by the MEGA 2.1 program (42) assuming the Tamura–Nei model of nucleotide substitutions (43). Abbreviations of the bacterial names are listed in Supplementary Table 1. Archaeoglobus fulgidus (ar f) was chosen as representative of the Archaea.
We observed a strong connection between the phylogenetic relations among the analysed bacteria and the classification based on the three parameters: DNA asymmetry, DnaA box distribution and dnaA gene location. The phylogenetic relationships of the organisms were evaluated by an analysis of the 16S rRNA gene sequences (see phylogenetic tree, Figure 5). All analysed organisms belonging to two subgroups of the Firmicutes (G + C-low Gram positive), Bacilli and Clostridia, and many organisms of the Actinobacteria (G + C-high Gram positive) belong to the abd group. Full correlation of these methods is not observed in other Firmicutes, in most Mollicutes whose genomes underwent considerable reduction because of a parasitic style of life. In all the analysed genomes of Proteobacteria, the extreme of DNA asymmetry co-localizes only with the dnaA gene (the group ad). However, it cannot be excluded that the ‘specific’ DnaA boxes for this lineage have evolved and, therefore, in some cases it is difficult to find a cluster of DnaA boxes. The DnaA boxes found within the H.pylori oriC region support this hypothesis. In many chromosomes of organisms belonging to Proteobacteria, a translocation of the dnaA gene from the putative origin of replication to another place on the chromosome has been observed. Some of these bacteria are obligatory intracellular parasites or symbionts. The ? Proteobacteria genomes exhibit the highest diversity. Nevertheless, some of them show full (or almost full) agreement of the three analysed parameters. Most of the Proteobacteria genomes belong to the abd group. Within this lineage we observed three kinds of exceptions: (i) a distinct cluster of DnaA boxes in the putative origin region is not present in Coxiella burnetii (obligatory intracellular parasite); (ii) the dnaA gene is translocated from the origin region in Pasteurella multocida and H.ducreyi; and (iii) lack of correlation of the three parameters is observed in Haemophilus influenzae and two endosymbionts, W.glossinidia and B.floridanus (neither contains the dnaA gene).
In Chlamydiales (obligatory intracellular parasites), Pirellula and hyperthermophiles (Aquifex aeolicus and Thermotoga maritima), lack of correlation of the three parameters has also been observed. No universal rules can be determined for Spirochaetales, the Bacteroidetes/Chlorobi group and Cyanobacteria.
The results suggest that the state of abd (full agreement of the three methods) is probably the ancestral state for Firmicutes, Actinobacteria and Proteobacteria and was then modified in some lineages belonging to these groups, particularly in Proteobacteria and Mollicutes (usually because of some disturbed DNA asymmetry, translocation or disappearance of the dnaA gene and decay of the pattern or clusters of the E.coli perfect DnaA boxes). In contrast to large genomes, the small genomes of obligatory intracellular parasites or endosymbionts and extremophiles lost many regulatory elements (41), including an extra cluster(s) of DnaA boxes that in other organisms regulates the level of free DnaA protein (titration mechanism). Even the loss of the dnaA gene was observed. This aspect may also be partially attributable to a lack of need; living continuously within the host or under extreme conditions (e.g. high temperature) eliminates the environmental fluctuations encountered by free-living bacteria (41).
CONCLUSIONS
The results presented allow us to draw some major conclusions:
The most universal method of putative oriC identification in bacterial chromosomes is DNA asymmetry, although in some cases it is necessary to apply all three methods.
Many bacterial chromosomes show an overrepresentation of DnaA boxes and the presence of more than one cluster of boxes near the oriC region. The boxes may be involved in the titration of free DnaA protein and play a role in the regulation of chromosome replication.
The E.coli DnaA-like boxes are present in most bacterial chromosomes. However, they have not been found in some lineages and groups, e.g. Proteobacteria. The sequence of DnaA boxes in these organisms was probably modified by mutational pressure.
Signals indicating the oriC region are ambiguous or disappear in bacteria which have become obligatory intracellular parasites or endosymbionts with significantly reduced genomes.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
The work was supported by grant number 1016/IGM/04 and by the Ministry of Scientific Research and Information Research (The State Committee for Scientific Research, grant 3 P04A 079 22).
REFERENCES
Kornberg,A. and Baker,T. ( (1992) ) The DNA Replication, 2nd edn. Freeman W.H. and Co., New York, NY.
Baker,T.A. and Bell,S.P. ( (1998) ) Polymerases and the replisome: machines within machines. Cell, , 92, , 295–305.
Matsunaga,F., Forterre,P., Ishino,Y. and Myllykallio,H. ( (2001) ) In vivo interactions of archaeal Cdc6/Orc1 and minichromosome maintenance proteins with the replication origin. Proc. Natl Acad. Sci. USA, , 98, , 11152–11157.
Robinson,N.P., Dionne,I., Lundgren,M., Marsh,V.L., Bernander,R. and Bell,S.D. ( (2004) ) Identification of two origins of replication in the single chromosome of the archaeon Sulfolobus solfataricus. Cell, , 116, , 25–38.
Kelman,L.M. and Kelman,Z. ( (2003) ) Archaea: an archetype for replication initiation studies? Mol. Microbiol., , 48, , 605–615.
Messer,W. and Weigel,C. ( (1996) ) Initiation of chromosome replication. In Neidhardt,F.C. (ed.), Escherichia coli and Salmonella Cellular and Molecular Biology, 2nd edn. ASM Press Washington, DC, pp. 1579–1601.
Messer,W. ( (2002) ) The bacterial replication initiator DnaA. DnaA and oriC, the bacterial mode to initiate DNA replication. FEMS Microbiol. Rev., , 26, , 355–374.
Marczyski,G.T. and Shapiro,L. ( (2002) ) Control of chromosome replication in Caulobacter crescentus. Annu. Rev. Microbiol., , 56, , 625–656.
Suhan,M., Chen,S.Y., Thompson,H.A., Hoover,T.A., Hill,A. and Williams,J.C. ( (1994) ) Cloning and characterization of an autonomous replication sequence from Coxiella burnetii. J. Bacteriol., , 176, , 5233–5243.
Lobry,J.R. ( (1996) ) Asymmetric substitution patterns in the two DNA strands of bacteria. Mol. Biol. Evol., , 13, , 660–665.
Lobry,J.R. ( (1996) ) A simple vectorial representation of DNA sequence for the detection of replication origins in bacteria. Biochimie, , 78, , 323–326.
Freeman,J.M., Plasterer,T.N., Smith,T.F. and Mohr,S.C. ( (1998) ) Patterns of genome organization in bacteria. Science, , 279, , 1827.
Grigoriev,A. ( (1998) ) Analysing genomes with cumulative skew diagrams. Nucleic Acids Res., , 26, , 2286–2290.
McLean,M.J., Wolfe,K.H. and Devine,K.M. ( (1998) ) Base composition skews, replication orientation, and gene orientation in 12 prokaryote genomes. J. Mol. Evol., , 47, , 691–696.
Salzberg,S.L., Salzberg,A.J., Kerlavage,A.R. and Tomb,J-F. ( (1998) ) Skewed oligomers and origins of replication. Gene, , 217, , 57–67.
Mackiewicz,P., Gierlik,A., Kowalczuk,M., Dudek,M.R. and Cebrat,S. ( (1999) ) How does replication-associated mutational pressure influence amino acid composition of proteins? Genome Res., , 9, , 409–416.
Mackiewicz,P., Gierlik,A., Kowalczuk,M., Dudek,M.R. and Cebrat,S. ( (1999) ) Asymmetry of nucleotide composition of prokaryotic chromosomes. J. Appl. Genet., , 40, , 1–14.
Rocha,E.P., Danchin,A. and Viari,A. ( (1999) ) Universal replication biases in bacteria. Mol. Microbiol., , 32, , 11–16.
Mrazek,J. and Karlin,S. ( (1998) ) Strand compositional asymmetry in bacterial and large viral genomes. Proc. Natl Acad. Sci. USA, , 95, , 3720–3725.
Frank,A.C. and Lobry,J.R. ( (1999) ) Asymmetric substitution patterns: a review of possible underlying mutational or selective mechanisms. Gene, , 238, , 65–77.
Tillier,E.R. and Collins,R.A. ( (2000) ) The contributions of replication orientation, gene direction, and signal sequences to base-composition asymmetries in bacterial genomes. J. Mol. Evol., , 50, , 249–257.
Kowalczuk,M., Mackiewicz,P., Mackiewicz,D., Nowicka,A., Dudkiewicz,M., Dudek,M.R. and Cebrat,S. ( (2001) ) DNA asymmetry and the replicational mutational pressure. J. Appl. Genet., , 42, , 553–577.
Frank,A.C. and Lobry,J.R. ( (2000) ) Oriloc: prediction of replication boundaries in unannotated bacterial chromosomes. Bioinformatics, , 16, , 560–561.
Picardeau,M., Lobry,J.R. and Hinnebusch,B.J. ( (1999) ) Physical mapping of an origin of bidirectional replication at the centre of the Borrelia burgdorferi linear chromosome. Mol. Microbiol., , 32, , 437–445.
Picardeau,M., Lobry,J.R. and Hinnebusch,B.J. ( (2000) ) Analyzing DNA strand compositional asymmetry to identify candidate replication origins of Borrelia burgdorferi linear and circular plasmids. Genome Res., , 10, , 1594–1604.
Zawilak,A., Cebrat,S., Mackiewicz,P., Król-Hulewicz,A., Jakimowicz,D., Messer,W., Gociniak,G. and Zakrzewska-Czerwinska,J. ( (2001) ) Identification of a putative chromosomal replication origin from Helicobacter pylori and its interaction with the initiator protein DnaA. Nucleic Acids Res., , 29, , 2251–2259.
Gil,R., Silva,F.J., Zientz,E., Delmotte,F., Gonzalez-Candelas,F., Latorre,A., Rausell,C., Kamerbeek,J., Gadau,J., Holldobler,B. et al. ( (2003) ) The genome sequence of Blochmannia floridanus: comparative analysis of reduced genomes. Proc. Natl Acad. Sci. USA, , 100, , 9388–9393.
Richter,S., Hagemann,M. and Messer,W. ( (1998) ) Transcriptional analysis and mutation of a dnaA-like gene in Synechocystis sp. strain PCC 6803. J. Bacteriol., , 180, , 4946–4949.
Zawilak,A., Durrant,M.C., Jakimowicz,P., Backert,S. and Zakrzewska-Czerwinska,J. ( (2003) ) DNA binding specificity of the replication initiator protein, DnaA from Helicobacter pylori. J. Mol. Biol., , 334, , 933–947.
Lopez,P., Forterre,P., le Guyader,H. and Philippe,H. ( (2000) ) Origin of replication of Thermotoga maritima. Trends Genet., , 16, , 59–60.
Richter,D., Hess,W.R., Krause,M. and Messer,W. ( (1998) ) Unique organization of the dnaA region from Prochlorococcus marinus CCMP1375, a marine cyanobacterium. Mol. Gen. Genet., , 257, , 534–541.
Chen,C.W., Huang,Ch.-H., Lee,H.-H., Tsai,H.-H. and Kirby,R. ( (2002) ) Once the circle has been broken: dynamics and evolution of Streptomyces chromosomes. Trends in Genet., , 18, , 522–529.
Paulsen,I.T., Seshadri,R., Nelson,K.E., Eisen,J.A., Heidelberg,J.F., Read,T.D., Dodson,R.J., Umayam,L., Brinkac,L.M. and Beanan,M.J. ( (2002) ) The Brucella suis genome reveals fundamental similarities between animal and plant pathogens and symbionts. Proc. Natl Acad. Sci. USA, , 99, , 13148–13153.
Egan,E.S. and Waldor,M.K. ( (2003) ) Distinct replication requirements for the two Vibrio cholerae chromosomes. Cell, , 114, , 521–530.
Madiraju,M.V., Qin,M.H., Yamamoto,K., Atkinson,M.A. and Rajagopalan,M. ( (1999) ) The dnaA gene region of Mycobacterium avium and the autonomous replication activities of its 5' and 3' flanking regions. Microbiology, , 145, , 2913–2921.
Roth,A. and Messer,W. ( (1998) ) High-affinity binding sites for the initiator protein DnaA on the chromosome of Escherichia coli. Mol. Microbiol., , 28, , 395–401.
Boye,E., L?bner-Olesen,A. and Skarstad,K. ( (2000) ) Limiting DNA replication to once and only once. EMBO Rep., , 1, , 479–483.
Ogawa,T., Yamada,Y., Kuroda,T., Kishi,T. and Moriya,S. ( (2002) ) The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli. Mol. Microbiol., , 44, , 1367–1375.
Muto,A. and Osawa,S. ( (1987) ) The guanine and cytosine content of genomic DNA and bacterial evolution. Proc. Natl Acad. Sci. USA, , 84, , 16616–16619.
Schaefer,Ch. and Messer,W. ( (1991) ) DnaA protein/DNA interaction. Modulation of the recognition sequence. Mol. Gen. Genet., , 226, , 34–40.
Moran,N.A. ( (2002) ) Microbial minimalism: genome reduction in bacterial pathogens. Cell, , 108, , 583–586.
Kumar,S., Tamura,K. and Nei,M. ( (1993) ) MEGA: Molecular Evolutionary Genetics Analysis. Pennsylvania State University, University Park, PA.
Tamura,K. and Nei,M. ( (1993) ) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol., , 10, , 512–526.(Pawe Mackiewicz, Jolanta Zakrzewska-Czer)
* To whom correspondence should be addressed. Tel: +48 71 337 11 72; Fax: +48 71 337 13 82; Email: zakrzew@iitd.pan.wroc.pl
ABSTRACT
Three methods, based on DNA asymmetry, the distribution of DnaA boxes and dnaA gene location, were applied to identify the putative replication origins in 120 chromosomes. The chromosomes were classified according to the agreement of these methods and the applicability of these methods was evaluated. DNA asymmetry is the most universal method of putative oriC identification in bacterial chromosomes, but it should be applied together with other methods to achieve better prediction. The three methods identify the same region as a putative origin in all Bacilli and Clostridia, many Actinobacteria and Proteobacteria. The organization of clusters of DnaA boxes was analysed in detail. For 76 chromosomes, a DNA fragment containing multiple DnaA boxes was identified as a putative origin region. Most bacterial chromosomes exhibit an overrepresentation of DnaA boxes; many of them contain at least two clusters of DnaA boxes in the vicinity of the oriC region. The additional clusters of DnaA boxes are probably involved in controlling replication initiation. Surprisingly, the characteristic features of the initiation of replication, i.e. a cluster of DnaA boxes, a dnaA gene and a switch in asymmetry, were not found in some of the analysed chromosomes, particularly those of obligatory intracellular parasites or endosymbionts. This is presumably connected with many mechanisms disturbing DNA asymmetry, translocation or disappearance of the dnaA gene and decay of the Escherichia coli perfect DnaA box pattern.
INTRODUCTION
As replication is controlled at the initiation stage, the events that occur at the replication origin play a central role in the cell cycle. The initiation of DNA replication is a complex process involving several regulated steps: (i) binding the initiator protein(s) to sites located within the origin region (ori); (ii) local unwinding of the ori region; and (iii) loading the DNA helicase and other proteins required to form replication forks (1,2). Chromosomal replication starts from one (Bacteria and Archaea, e.g. Pyroccocus abyssi) (3), two (Archaea, e.g. Sulfolobus solfataricus) (4) or multiple replication origins (Eukaryota) . In bacteria, chromosomal replication initiated at the oriC region proceeds bi-directionally until the replication forks reach the termination site, terC (in the case of circular chromosomes) or chromosomal ends (in the case of linear chromosomes) (6). The initiation of bacterial chromosome replication is mediated by the initiator DnaA protein, which interacts with repetitive non-palindromic nonamer sequences, the DnaA boxes, located within the oriC region. Among bacteria, the initiation of replication is best understood in Escherichia coli (7). The binding of 10 to 20 DnaA protein molecules to five DnaA boxes promotes unwinding within the AT-rich region of oriC.
The sequences of bacterial oriC regions are conserved only among closely related organisms. Bacterial replication origins vary in sizes (from 200 up to 1000 bp), but (nearly) all contain several DnaA boxes and an AT-rich region. A cluster of four or more DnaA boxes is an indication of a functional chromosomal origin. However, only one and two DnaA boxes, respectively, were identified within the functional oriC regions of Caulobacter crescentus (8) and Coxiella burnetti (9). The oriC region is always located within the intergenic region, frequently within the rnpA–rmpH–dnaA–dnaN–recF–gyrB–rnpA genes cluster, usually next to the dnaA gene.
During the last few years, the genomes of approximately 150 bacteria have been completely sequenced, and there are approximately 500 ongoing bacterial genome projects. Thus, within a decade, the genomes of nearly every significant bacterial species will have been sequenced. Since the events that occur at the replication origin are central to the process regulating DNA replication and the cell cycle, it is important to define the rules required for a proper oriC localization within analysed genomes.
Initially, the oriC regions were identified by direct measurement of the replication time of various genetic markers, by construction of minichromosomes (plasmids containing oriC instead of their own replication origin) or by two-dimensional gel electrophoresis. So far, however, only 11 oriC regions have been shown in vivo to be functional. Now that long stretches or the complete sequences of bacterial chromosomes have become available it is possible to identify a potential oriC region in silico. Lobry (10,11) was the first to find bias (asymmetry) in nucleotide composition in some bacterial chromosomes. He noticed that the asymmetry changes its polarity at the origin and the terminus of chromosome replication, where DNA strands change their character from leading to lagging and vice versa. This asymmetry is usually measured as the normalized difference in the content of complementary nucleotides . This observation was then confirmed by other researchers for many bacterial chromosomes . There are many reasons for this asymmetry, but the main one seems to be an asymmetric mutational pressure (different nucleotide substitution patterns) associated with the replication of the leading and lagging strands . Since DNA strands change their mode of synthesis from lagging to leading in the oriC region, it is possible to identify the region in silico by analysis of chromosome asymmetry, using, for instance, the computer program Oriloc created by Frank and Lobry (23). This method is commonly used to identify the putative oriC region within chromosomes, particularly before experimental analyses (24–26). However, the switch in asymmetry does not always exactly correspond to the position of a functional replication origin; e.g. the in silico predicted origin in Helicobacter pylori 26695 was located 2500 bp away form the functional one (26). Therefore, in order to locate the replication origin precisely, GC skew analysis should be accomplished by additional analyses to locate the elements involved in the initiation of chromosomal replication: cluster(s) of DnaA boxes and the dnaA gene. Searching for a cluster of DnaA boxes requires first of all a proper definition of the DnaA box. The perfect sequence for the DnaA box of a model organism, E.coli, is 5'-TTATCCACA-3'. However, the preferred DnaA box sequence for other bacteria differs slightly from that of E.coli; it depends on the G + C content of the analysed chromosome and/or binding specificity of a given DnaA protein (for details see Results and Discussion).
The aim of this study is to evaluate the universality of three different methods of oriC identification (DNA asymmetry, DnaA box distribution and dnaA gene location) in their application to the analysis of completely sequenced bacterial chromosomes. We have tried to group the analysed genomes according to the concordance of these methods and to analyse the problem from the phylogenetic point of view.
MATERIALS AND METHODS
The analyses were performed on 120 complete sequences of bacterial chromosomes representing different taxonomic groups (see Supplementary Table 1). The chromosomal sequences were downloaded from the Genbank (ftp://www.ncbi.nlm.nih.gov). The assumed location of the dnaA gene on the chromosome was that of the database annotations. DNA asymmetry was analysed by DNA walks describing the asymmetry between differently replicating DNA strands . This method differs from that originally proposed (10,11,13), as it cumulates local deviations of a parameter of asymmetry ( or ) from the average value specific to the whole chromosome. This method eliminates the global compositional trend of the whole chromosome and smoothes random fluctuations. By this method the single origin and terminus of replication (if they are present) are easily recognized as the two main extrema of the DNA walk, i.e. where asymmetry changes its sign (Figure 1; Supplementary Figure 6).
Figure 1. DNA asymmetry, DnaA box distribution and dnaA gene location on bacterial chromosomes. x-axis—chromosome coordinates in base pairs, in (A–E and H) beginning and ending in the presumed terminus of replication (according to the DNA asymmetry). The chromosomal coordinates in (F and G) are as in the databases. The grey line and the right y-axis show the DNA walk describing DNA asymmetry (asymmetry for in (F); asymmetry for for the other chromosomes). Black peaks with diamonds and the left y-axis show the DnaA box distribution expressed by the b value. The short vertical line indicates the dnaA gene location.
In order to identify the potential cluster(s) of DnaA boxes we looked for all possible 28 nonamers which differed by no more than one position from the perfect DnaA box: TTATCCACA. The density of the box distribution on the chromosome was presented on charts by b = 1/d, calculated for each box, where d is a sum of the distances of the given box to its two adjacent boxes. The b value corresponds to the density of boxes in the given region of the chromosome, with a high value of b indicating a cluster of three boxes.
To visualize the correlation of the three methods of oriC identification (dnaA gene location, DNA asymmetry and DnaA box distribution), in the charts presenting these analyses (Figures 1–3; Supplementary Figure 6), the x-coordinate begins and ends in the presumed (according to DNA asymmetry) terminus of replication. Then the presumed oriC region is usually located near the middle of the chart. This was done for all circular chromosomes showing two distinct global extrema of DNA asymmetry. The charts for linear chromosomes and chromosomes without distinct global extrema of DNA asymmetry have the original sequence coordinates on the x-axis. In the charts, the DNA walk is shown only for the one parameter of asymmetry ( or ) which exhibited the strongest and most distinct DNA asymmetry.
Figure 2. DNA asymmetry in , DnaA box distribution, and dnaA gene location on the H.pylori J99 chromosome. (A) Distribution of DnaA boxes which differ in more than one position from the perfect DnaA box TTATCCACA; (B) distribution of DnaA boxes experimentally proved to be functional in this genome according to Zawilak et al. (26). The x-axis describing the coordinate position on the chromosome in base pairs begins and ends in the presumed (according to DNA asymmetry) terminus of replication. See Figure 1 for further explanation.
Figure 3. DNA asymmetry in , DnaA box' distribution and dnaA gene location in the oriC region. Example of single cluster of DnaA boxes (A) and a few separated clusters in the region (B). See Figure 1 for further explanation.
Supplementary material
Table 1 (general characteristics of the analysed chromosomes); Table 2 (locations of boxes and their patterns in the presumed oriC region); and Figure 6 (DNA asymmetry, DnaA box distribution and dnaA gene location on 120 bacterial chromosomes).
RESULTS AND DISCUSSION
General remarks
In order to elucidate the rules for predicting the origin of bacterial chromosome replication, we performed a comprehensive in silico analysis of 112 complete genome sequences of bacteria representing different taxonomic groups (Supplementary Table 1). We analysed the DNA asymmetry and the presence of elements involved in the initiation of replication (DnaA boxes and dnaA gene). Most (101) of the analysed bacteria possess a single circular chromosome, while three have a linear chromosome. Eight organisms contain two chromosomes (either with both circular or with one circular and the other linear).
A distinct DNA asymmetry between the leading and lagging strands is present in 93% of the analysed chromosomes (Supplementary Table 1). Such asymmetry is not present in the chromosomes of Aquifex aeolicus, Bifidobacterium longum, Deinococcus radiodurans (chromosome II), Gloeobacter violaceus, Nostoc sp., Phytoplasma asteris, Streptomyces avermitilis, Synechocystis sp. and Wolbachia pipientis. In most of the chromosomes stronger asymmetry was observed for the parameter than for .
dnaA genes have been found in all bacterial genomes studied so far, with the exception of the Wigglesworthia glossinidia and Blochmannia floridanus genomes (which are reduced in size). In addition, the chromosomes of both endosymbionts do not contain any clusters of DnaA boxes. The lack of typical bacterial elements involved in the initiation of replication may reflect the extreme dependency of W.glossinidia and B.floridanus on the host genome functions. These bacteria are freely located in the cytosol of bacteriocytes. Bacteria in the cytosol might be a potential danger for the host cell and more difficult to control. The development of a stable symbiosis with cytosolic bacteria probably requires more direct control of DNA replication of the symbionts by the host, involving the loss of the dnaA gene (27).
If a genome consists of two chromosomes, one of them lacks the dnaA gene (Supplementary Table 1). Two dnaA genes are present in chromosomes of the following organisms: Chlamydia muridarum, Chlamydia trachomatis, Chlamydia pneumoniae, Chlamydia caviae and Pirellula sp. However, it is not yet known whether only one of the dnaA genes or both are transcriptionally active. In Chlamydia it cannot be excluded that two dnaA genes are differentially transcribed during the organism's unusual intracellular developmental life cycle. In Brucella, the additional dnaA gene present on the larger chromosome was probably transferred from the smaller chromosome (via interchromosomal exchange).
As expected, DnaA boxes (differing in at the most one position from the perfect E.coli DnaA box TTATCCACA) were found in all chromosomes. Interestingly, the DnaA boxes were overrepresented (compared with their random occurrence expected from the base composition of a given chromosome) in 70% of the analysed chromosomes. In this group of chromosomes, the overrepresentation was statistically significant (with p < 0.05 in 72% cases). We assume that a potential cluster of DnaA boxes consists of a set of three boxes with an average distance between them of <100 bp (the distance 100 bp corresponds to a b value of 0.005 on the left y-axis in Figures 1–3). Such clusters (located anywhere on a chromosome) were present in 83% of chromosomes (85% of species).
Evaluation of the three methods in oriC identification
We assumed that the methods will identify the same region as the putative origin of replication if the distances between the location of the cluster of DnaA boxes and/or the dnaA gene and/or the position of the extreme of the DNA walk are <1% of the chromosome's length. Some doubtful cases (e.g. when the distance was >1% and <7% of chromosome length) were assigned by using parentheses and described by a suitable notation (see Supplementary Table 1). The assumed 1% limit corresponds to the distance between the dnaA gene and the experimentally identified oriC region in E.coli. In most cases, the discrepancy between the three methods results from fluctuation of DNA asymmetry in the extreme in the local scale (see below for potential causes of the fluctuations).
Taking these assumptions into account, we have classified the chromosomes into five distinct groups, depending on the co-localization of the putative oriC indicated by the three methods (the abbreviations mean: a—method of DNA asymmetry; b—DnaA box distribution; d—dnaA gene location):
abd: all three methods identify the same region as oriC—67 chromosomes (Figure 1A);
ab: the extreme of the DNA walk (asymmetry) is adjacent to the cluster of DnaA boxes—15 chromosomes (Figure 1B);
ad: the extreme of the DNA walk (asymmetry) corresponds to the location of the dnaA gene—18 chromosomes (Figure 1C);
bd: the dnaA gene is in the vicinity of the cluster of DnaA boxes—3 chromosomes (Figure 1D);
O: each method identifies a different region—17 chromosomes (Figure 1E).
The close vicinity of these three features suggests that the analysed region contains a functional origin. So far, only 11 of the putative oriC regions have been shown in vivo to be functional. Eight of them belong to the first group, abd; their in silico predicted oriC regions co-localized the functional origins.
The three methods applied to the identification of oriC agree rather well. In 55% of cases all methods indicate the same region. The most universal method seems to be DNA asymmetry, which was in agreement with at least one of the other methods in 83% of cases. There are only three cases (bd) where the extreme of DNA asymmetry is distant from both the cluster of DnaA boxes and the dnaA gene. In some cases there is imperfect agreement between the location of the cluster of DnaA boxes and/or the dnaA gene and DNA asymmetry, which performs some fluctuations near the global extreme. The asymmetry is probably disturbed by other phenomena (such as transcription, chromosome rearrangements, oligomers and gene distributions . In some cases, the observed shift in the extreme of asymmetry from the cluster of DnaA boxes and the dnaA gene may reflect a very recent translocation of the origin region as well. In this case, replication-associated mutational pressure probably had not enough time to erase the old compositional bias and introduce the new one. The smooth global extremes observed in the chromosomes of Mesorhizobium loti (Figure 1D) and Nitrosomonas europaea may suggest such a possibility.
In the chromosomes grouped in ab and ab(d), the dnaA gene was probably transferred to the other region of the chromosome. This may indicate that a close proximity of the dnaA gene to the origin region is not necessary for its proper functioning (its product can act in trans). Furthermore, it has been reported that mutants of Synechocystis sp. with inactivated dnaA gene were viable (28). It cannot be excluded that replication of this chromosome is initiated by another protein. However, in most cases the dnaA gene is in the vicinity of the oriC region. The reason for the close proximity of both elements is not clear. It has been noted that genomic rearrangements rarely occur in the close vicinity of the origins of replication (5). It is possible that the juxtaposition ensures that the DnaA protein is able to associate with the origin as soon as it is synthesized. Alternatively, the proximity to the origin minimizes the possibility of disruption of the linkage between the well-cooperating dnaA gene–oriC region system.
In some cases , a cluster of DnaA boxes could not be identified in the entire chromosome. However, it must be noted that the preferred DnaA box sequences of certain bacteria may differ slightly from the E.coli perfect DnaA box (TTATCCACA), e.g. in high-G + C organisms such as Micrococcus, Mycobacterium or Streptomyces (GC > 70%), the third position is substituted by G or C. In addition, our own recent experiments have demonstrated that particular DnaA proteins may exhibit different specificity towards DNA than that from E.coli, e.g. the affinity of the DnaA protein from H.pylori to the TCATTCACA sequence (two mismatches) is higher than to the E.coli perfect box. Therefore, the possibility of a different consensus sequence for the DnaA box (and/or binding specificity of DnaA protein) should be considered when the search of a putative DnaA protein binding site is performed. For example, our first search for DnaA box motifs whose sequences differ by no more than one base from the E.coli perfect box failed to find a single cluster of DnaA boxes in the entire H.pylori chromosome (Figure 2A). However, when we looked for DnaA box motifs with two to three mismatches, we found a cluster of DnaA boxes upstream of the dnaA gene (Figure 2B). This region, designated as the H.pylori putative origin of replication, is the only place on the chromosome that contains a cluster of DnaA boxes. The H.pylori DnaA exhibits a high affinity to this region (29). There are also many other examples of origin regions that contain ‘species-specific’ DnaA boxes, e.g. Borrelia burgdorferi, TTTTAAACA, TTATCAAAA, TTTTAAAAA (24); Thermotoga maritima, AAACCTACCACC (30); Caulobacter crescentus, TGATCCACA (8); and Prochlorococcus marinus, TTCCACA (31).
The chromosomes that show asymmetry with more than two distinct extrema are supposed to be replicated from more than one oriC. However, we found in some of them a correlation between the main local extreme of asymmetry and the location of the cluster of DnaA boxes and/or the dnaA gene: Aquifex aeolicus, Bifidobacterium longum, Gloeobacter violaceus, Nostoc sp. (Figure 1F) and Synechocystis sp. The region may be suspected to be the main (the strongest or earlier activated) origin of replication. However, there are many phenomena that influence the course of DNA asymmetry which may disturb the asymmetry introduced by replication-associated mutational pressure . Only experimental studies can definitely solve this problem. Interestingly, in the linear chromosome of S.avermitilis there are two distinct minima suspected of being the origin region (Figure 1G). One of these strongly correlates with the location of distinct clusters of DnaA boxes and the dnaA gene. A few DnaA boxes were also found to be close to the second minimum, but they were rather scattered. The arrangement of the DnaA boxes in close vicinity of the dnaA gene is the same as within the functional oriC region of Streptomyces coelicolor. Interestingly, despite selective pressures or the advantage of central positioning of the origin (to finish replication synchronously), the putative oriC of S.avermitilis is shifted 800 000 bp (10% of the chromosome's length) from the centre. It has been proposed that linear Streptomyces chromosomes resulted from recombination between a linear plasmid and an ancestral circular chromosome (32). Probably, in the case of S.avermitilis, linearization events resulted in the formation of chromosomal arms with different lengths (the oriC is asymmetrically positioned) and therefore the second minimum is observed (see Figure 1G).
Lack of any agreement between the three methods was usually observed in the second chromosomes (the smaller ones). In chromosome II of Brucella suis the extreme of DNA asymmetry correlates with the location of repC, a typical plasmid replication gene. It has been suggested that the chromosome of B.suis is derived from a megaplasmid that was captured by ancestral Brucella (33). Recent experimental studies on the initiation of replication of two Vibrio cholerae chromosomes revealed distinct replication requirements for the two chromosomes. The functional origins, oriCIVC and oriCIIVC, of V.cholerae chromosomes are unrelated; the structure of oriCIVC resembles that of E.coli, whereas oriCIIVC shares some features with certain plasmid replicons (34). The distinct replication requirements of the two chromosomes may minimize competition and thereby help ensure the maintenance of the divided genome. The situation is similar to the observed incompatibility of two plasmids which require identical replication initiation factors.
In organisms which contain a single chromosome and belong to group O, the initiation of chromosome replication is probably also mediated by another protein(s).
Distribution, organization and composition of DnaA boxes
We have found in numerous analysed chromosomes (Supplementary Table 1) that the clusters of DnaA boxes are not distributed randomly in a chromosome, but are densely packed within a relatively short distance around the putative oriC region (Figure 1H). The distance between the distal cluster of DnaA boxes and the oriC varies from 27 to 800 kb (this being up to 23% of chromosome length). The DnaA boxes, most probably involved in initiation of replication, were identified in 76 chromosomes. Detailed data (locations of these boxes and their sequence patterns) are included in the Supplementary Material (Supplementary Table 2), whereas in Table 1 we present the general organization of the clusters (number of boxes in a given cluster, mean distance between boxes in the cluster and distances between clusters). 41 chromosomes contain at least two separated clusters of DnaA boxes near the putative oriC region. In most cases, the distance between the clusters of DnaA boxes is only a few kilobase pairs (up to 7 kb). Longer distances are in Neisseria (12–13 kb) and Actinobacteria (mycobacteria, 15–150 kb; S.coelicolor and Corynebacterium efficiens, 30 kb; Corynebacterium diphtheriae, 25 kb).
Table 1. Organization of the clusters of DnaA boxes in the putative oriC regions
There are a few possible explanations for the occurrence of many clusters of DnaA boxes. One explanation assumes that more than one cluster participates in the initiation of replication. In several chromosomes which have at least two clusters of DnaA boxes in the putative origin region, we observed two local extrema and a plateau between them in the DNA walk describing asymmetry, e.g. Staphylococcus epidermidis (Figure 3, see for comparison the V-shaped asymmetry of the Haemophilus ducreyi chromosome). This may suggest that replication starts from one or both sites. In this case, the region corresponding to the plateau would be replicated sometimes as the leading strand and sometimes as the lagging strand. However, it has been experimentally demonstrated that the single Bacillus subtilis oriC region is composed of two DnaA clusters (separated by the dnaA gene, 1500 bp), which are both required for a functional origin. The unwinding of DNA occurs at the 27 bp AT-rich cluster in the vicinity of dnaN gene. In Mycobacterium (Mycobacterium avium and Mycobacterium tuberculosis), the dnaA gene is also flanked by two clusters of DnaA boxes. However, only one of them acts as oriC. Interestingly, in M.tuberculosis the 3' region of the dnaA gene exhibits oriC activity, while in M.avium the 5' region functions as oriC (35).
The second explanation assumes that the clusters of DnaA boxes (which are not involved in the initiation of replication) may contribute to the negative control of replication by reducing the level of free DnaA protein. Since DnaA proteins interact cooperatively with closely spaced DnaA boxes, the clusters of boxes can bind more molecules than single boxes dispersed randomly along the chromosome. In E.coli there are five regions of the chromosome which bind DnaA protein (36). DnaA protein exhibits particularly high affinity towards one of them, datA, which is located close to the oriC. The datA region has been suggested as playing a pivotal role in controlling replication initiation; it can bind eight times more DnaA protein molecules than the oriC region (37,38). Thus, shortly after initiation of replication, the datA region is duplicated and therefore is able to reduce significantly free DnaA below the level that provokes over-initiation. The presence of many clusters of DnaA boxes near the oriC region in chromosomes other than that of E.coli suggests that such replication control may be a common mechanism for bacteria.
Table 2 presents the general (mean) composition of DnaA boxes, counted only for boxes found in the most probable oriC regions in 76 chromosomes. The most variable are the third and the fifth positions in the consensus. This is in agreement with the recently established crystal structure of the E.coli DnaA binding domain in complex with the perfect DnaA box, which showed that the nucleotide at the fifth position does not play an essential role in the interaction with the DnaA protein. We observed statistically significant correlation (r = 0.62) between G + C content of the DnaA boxes and G + C content counted for the given genome (Figure 4). Assuming that global genomic G + C reflects, to some extent, the composition generated by mutational pressure (39), one can assume that composition of the DnaA boxes may be moderated by mutational pressure and that some changes may be neutral. Actually, a sequence comparison of DnaA boxes derived from the origins of organisms with high G + C content revealed that G is frequently present instead of A at the third position of the DnaA box. Bases in the second, fourth, seventh, eighth and ninth positions are highly conserved. Extensive binding studies on E.coli revealed that these positions are of particular importance (40,7).
Table 2. The composition of DnaA boxes derived from the putative origins identified in 76 chromosomes
Figure 4. Correlation between G + C content of DnaA boxes (derived from the putative origins identified in 76 chromosomes) and genomic G + C content.
Phylogenetic approach
We compared the phylogenetic (taxonomic) status of the analysed organisms with their classification based on DNA asymmetry, DnaA box distribution, and dnaA gene location (see Supplementary Table 1 and Figure 5). Strains belonging to the same species or even the same genus do not exhibit significant differences in DNA asymmetry, distribution of DnaA boxes and dnaA gene location on the chromosome. There are a few exceptions: (i) among Mycoplasma genomes, only Mycoplasma pulmonis shows full agreement of the three methods (it belongs to group abd); (ii) the genome of Bordetella pertussis shows more rough asymmetry (with some local extrema) than other Bordetella genomes; (iii) Haemophilus ducreyi and Haemophilus influenzae were assigned to different groups, ad and O, respectively; (iv) Streptomyces genomes differ in DNA asymmetry; and (v) Treponema denticola shows some shift of the dnaA gene compared with Treponema pallidum.
Figure 5. The 16S rRNA phylogenetic tree for 112 analysed bacteria and their classification based on concordance of methods of oriC identification: DNA asymmetry, DnaA box distribution and the dnaA gene location. The tree was built with the minimum evolution method (ME) by the MEGA 2.1 program (42) assuming the Tamura–Nei model of nucleotide substitutions (43). Abbreviations of the bacterial names are listed in Supplementary Table 1. Archaeoglobus fulgidus (ar f) was chosen as representative of the Archaea.
We observed a strong connection between the phylogenetic relations among the analysed bacteria and the classification based on the three parameters: DNA asymmetry, DnaA box distribution and dnaA gene location. The phylogenetic relationships of the organisms were evaluated by an analysis of the 16S rRNA gene sequences (see phylogenetic tree, Figure 5). All analysed organisms belonging to two subgroups of the Firmicutes (G + C-low Gram positive), Bacilli and Clostridia, and many organisms of the Actinobacteria (G + C-high Gram positive) belong to the abd group. Full correlation of these methods is not observed in other Firmicutes, in most Mollicutes whose genomes underwent considerable reduction because of a parasitic style of life. In all the analysed genomes of Proteobacteria, the extreme of DNA asymmetry co-localizes only with the dnaA gene (the group ad). However, it cannot be excluded that the ‘specific’ DnaA boxes for this lineage have evolved and, therefore, in some cases it is difficult to find a cluster of DnaA boxes. The DnaA boxes found within the H.pylori oriC region support this hypothesis. In many chromosomes of organisms belonging to Proteobacteria, a translocation of the dnaA gene from the putative origin of replication to another place on the chromosome has been observed. Some of these bacteria are obligatory intracellular parasites or symbionts. The ? Proteobacteria genomes exhibit the highest diversity. Nevertheless, some of them show full (or almost full) agreement of the three analysed parameters. Most of the Proteobacteria genomes belong to the abd group. Within this lineage we observed three kinds of exceptions: (i) a distinct cluster of DnaA boxes in the putative origin region is not present in Coxiella burnetii (obligatory intracellular parasite); (ii) the dnaA gene is translocated from the origin region in Pasteurella multocida and H.ducreyi; and (iii) lack of correlation of the three parameters is observed in Haemophilus influenzae and two endosymbionts, W.glossinidia and B.floridanus (neither contains the dnaA gene).
In Chlamydiales (obligatory intracellular parasites), Pirellula and hyperthermophiles (Aquifex aeolicus and Thermotoga maritima), lack of correlation of the three parameters has also been observed. No universal rules can be determined for Spirochaetales, the Bacteroidetes/Chlorobi group and Cyanobacteria.
The results suggest that the state of abd (full agreement of the three methods) is probably the ancestral state for Firmicutes, Actinobacteria and Proteobacteria and was then modified in some lineages belonging to these groups, particularly in Proteobacteria and Mollicutes (usually because of some disturbed DNA asymmetry, translocation or disappearance of the dnaA gene and decay of the pattern or clusters of the E.coli perfect DnaA boxes). In contrast to large genomes, the small genomes of obligatory intracellular parasites or endosymbionts and extremophiles lost many regulatory elements (41), including an extra cluster(s) of DnaA boxes that in other organisms regulates the level of free DnaA protein (titration mechanism). Even the loss of the dnaA gene was observed. This aspect may also be partially attributable to a lack of need; living continuously within the host or under extreme conditions (e.g. high temperature) eliminates the environmental fluctuations encountered by free-living bacteria (41).
CONCLUSIONS
The results presented allow us to draw some major conclusions:
The most universal method of putative oriC identification in bacterial chromosomes is DNA asymmetry, although in some cases it is necessary to apply all three methods.
Many bacterial chromosomes show an overrepresentation of DnaA boxes and the presence of more than one cluster of boxes near the oriC region. The boxes may be involved in the titration of free DnaA protein and play a role in the regulation of chromosome replication.
The E.coli DnaA-like boxes are present in most bacterial chromosomes. However, they have not been found in some lineages and groups, e.g. Proteobacteria. The sequence of DnaA boxes in these organisms was probably modified by mutational pressure.
Signals indicating the oriC region are ambiguous or disappear in bacteria which have become obligatory intracellular parasites or endosymbionts with significantly reduced genomes.
SUPPLEMENTARY MATERIAL
ACKNOWLEDGEMENTS
The work was supported by grant number 1016/IGM/04 and by the Ministry of Scientific Research and Information Research (The State Committee for Scientific Research, grant 3 P04A 079 22).
REFERENCES
Kornberg,A. and Baker,T. ( (1992) ) The DNA Replication, 2nd edn. Freeman W.H. and Co., New York, NY.
Baker,T.A. and Bell,S.P. ( (1998) ) Polymerases and the replisome: machines within machines. Cell, , 92, , 295–305.
Matsunaga,F., Forterre,P., Ishino,Y. and Myllykallio,H. ( (2001) ) In vivo interactions of archaeal Cdc6/Orc1 and minichromosome maintenance proteins with the replication origin. Proc. Natl Acad. Sci. USA, , 98, , 11152–11157.
Robinson,N.P., Dionne,I., Lundgren,M., Marsh,V.L., Bernander,R. and Bell,S.D. ( (2004) ) Identification of two origins of replication in the single chromosome of the archaeon Sulfolobus solfataricus. Cell, , 116, , 25–38.
Kelman,L.M. and Kelman,Z. ( (2003) ) Archaea: an archetype for replication initiation studies? Mol. Microbiol., , 48, , 605–615.
Messer,W. and Weigel,C. ( (1996) ) Initiation of chromosome replication. In Neidhardt,F.C. (ed.), Escherichia coli and Salmonella Cellular and Molecular Biology, 2nd edn. ASM Press Washington, DC, pp. 1579–1601.
Messer,W. ( (2002) ) The bacterial replication initiator DnaA. DnaA and oriC, the bacterial mode to initiate DNA replication. FEMS Microbiol. Rev., , 26, , 355–374.
Marczyski,G.T. and Shapiro,L. ( (2002) ) Control of chromosome replication in Caulobacter crescentus. Annu. Rev. Microbiol., , 56, , 625–656.
Suhan,M., Chen,S.Y., Thompson,H.A., Hoover,T.A., Hill,A. and Williams,J.C. ( (1994) ) Cloning and characterization of an autonomous replication sequence from Coxiella burnetii. J. Bacteriol., , 176, , 5233–5243.
Lobry,J.R. ( (1996) ) Asymmetric substitution patterns in the two DNA strands of bacteria. Mol. Biol. Evol., , 13, , 660–665.
Lobry,J.R. ( (1996) ) A simple vectorial representation of DNA sequence for the detection of replication origins in bacteria. Biochimie, , 78, , 323–326.
Freeman,J.M., Plasterer,T.N., Smith,T.F. and Mohr,S.C. ( (1998) ) Patterns of genome organization in bacteria. Science, , 279, , 1827.
Grigoriev,A. ( (1998) ) Analysing genomes with cumulative skew diagrams. Nucleic Acids Res., , 26, , 2286–2290.
McLean,M.J., Wolfe,K.H. and Devine,K.M. ( (1998) ) Base composition skews, replication orientation, and gene orientation in 12 prokaryote genomes. J. Mol. Evol., , 47, , 691–696.
Salzberg,S.L., Salzberg,A.J., Kerlavage,A.R. and Tomb,J-F. ( (1998) ) Skewed oligomers and origins of replication. Gene, , 217, , 57–67.
Mackiewicz,P., Gierlik,A., Kowalczuk,M., Dudek,M.R. and Cebrat,S. ( (1999) ) How does replication-associated mutational pressure influence amino acid composition of proteins? Genome Res., , 9, , 409–416.
Mackiewicz,P., Gierlik,A., Kowalczuk,M., Dudek,M.R. and Cebrat,S. ( (1999) ) Asymmetry of nucleotide composition of prokaryotic chromosomes. J. Appl. Genet., , 40, , 1–14.
Rocha,E.P., Danchin,A. and Viari,A. ( (1999) ) Universal replication biases in bacteria. Mol. Microbiol., , 32, , 11–16.
Mrazek,J. and Karlin,S. ( (1998) ) Strand compositional asymmetry in bacterial and large viral genomes. Proc. Natl Acad. Sci. USA, , 95, , 3720–3725.
Frank,A.C. and Lobry,J.R. ( (1999) ) Asymmetric substitution patterns: a review of possible underlying mutational or selective mechanisms. Gene, , 238, , 65–77.
Tillier,E.R. and Collins,R.A. ( (2000) ) The contributions of replication orientation, gene direction, and signal sequences to base-composition asymmetries in bacterial genomes. J. Mol. Evol., , 50, , 249–257.
Kowalczuk,M., Mackiewicz,P., Mackiewicz,D., Nowicka,A., Dudkiewicz,M., Dudek,M.R. and Cebrat,S. ( (2001) ) DNA asymmetry and the replicational mutational pressure. J. Appl. Genet., , 42, , 553–577.
Frank,A.C. and Lobry,J.R. ( (2000) ) Oriloc: prediction of replication boundaries in unannotated bacterial chromosomes. Bioinformatics, , 16, , 560–561.
Picardeau,M., Lobry,J.R. and Hinnebusch,B.J. ( (1999) ) Physical mapping of an origin of bidirectional replication at the centre of the Borrelia burgdorferi linear chromosome. Mol. Microbiol., , 32, , 437–445.
Picardeau,M., Lobry,J.R. and Hinnebusch,B.J. ( (2000) ) Analyzing DNA strand compositional asymmetry to identify candidate replication origins of Borrelia burgdorferi linear and circular plasmids. Genome Res., , 10, , 1594–1604.
Zawilak,A., Cebrat,S., Mackiewicz,P., Król-Hulewicz,A., Jakimowicz,D., Messer,W., Gociniak,G. and Zakrzewska-Czerwinska,J. ( (2001) ) Identification of a putative chromosomal replication origin from Helicobacter pylori and its interaction with the initiator protein DnaA. Nucleic Acids Res., , 29, , 2251–2259.
Gil,R., Silva,F.J., Zientz,E., Delmotte,F., Gonzalez-Candelas,F., Latorre,A., Rausell,C., Kamerbeek,J., Gadau,J., Holldobler,B. et al. ( (2003) ) The genome sequence of Blochmannia floridanus: comparative analysis of reduced genomes. Proc. Natl Acad. Sci. USA, , 100, , 9388–9393.
Richter,S., Hagemann,M. and Messer,W. ( (1998) ) Transcriptional analysis and mutation of a dnaA-like gene in Synechocystis sp. strain PCC 6803. J. Bacteriol., , 180, , 4946–4949.
Zawilak,A., Durrant,M.C., Jakimowicz,P., Backert,S. and Zakrzewska-Czerwinska,J. ( (2003) ) DNA binding specificity of the replication initiator protein, DnaA from Helicobacter pylori. J. Mol. Biol., , 334, , 933–947.
Lopez,P., Forterre,P., le Guyader,H. and Philippe,H. ( (2000) ) Origin of replication of Thermotoga maritima. Trends Genet., , 16, , 59–60.
Richter,D., Hess,W.R., Krause,M. and Messer,W. ( (1998) ) Unique organization of the dnaA region from Prochlorococcus marinus CCMP1375, a marine cyanobacterium. Mol. Gen. Genet., , 257, , 534–541.
Chen,C.W., Huang,Ch.-H., Lee,H.-H., Tsai,H.-H. and Kirby,R. ( (2002) ) Once the circle has been broken: dynamics and evolution of Streptomyces chromosomes. Trends in Genet., , 18, , 522–529.
Paulsen,I.T., Seshadri,R., Nelson,K.E., Eisen,J.A., Heidelberg,J.F., Read,T.D., Dodson,R.J., Umayam,L., Brinkac,L.M. and Beanan,M.J. ( (2002) ) The Brucella suis genome reveals fundamental similarities between animal and plant pathogens and symbionts. Proc. Natl Acad. Sci. USA, , 99, , 13148–13153.
Egan,E.S. and Waldor,M.K. ( (2003) ) Distinct replication requirements for the two Vibrio cholerae chromosomes. Cell, , 114, , 521–530.
Madiraju,M.V., Qin,M.H., Yamamoto,K., Atkinson,M.A. and Rajagopalan,M. ( (1999) ) The dnaA gene region of Mycobacterium avium and the autonomous replication activities of its 5' and 3' flanking regions. Microbiology, , 145, , 2913–2921.
Roth,A. and Messer,W. ( (1998) ) High-affinity binding sites for the initiator protein DnaA on the chromosome of Escherichia coli. Mol. Microbiol., , 28, , 395–401.
Boye,E., L?bner-Olesen,A. and Skarstad,K. ( (2000) ) Limiting DNA replication to once and only once. EMBO Rep., , 1, , 479–483.
Ogawa,T., Yamada,Y., Kuroda,T., Kishi,T. and Moriya,S. ( (2002) ) The datA locus predominantly contributes to the initiator titration mechanism in the control of replication initiation in Escherichia coli. Mol. Microbiol., , 44, , 1367–1375.
Muto,A. and Osawa,S. ( (1987) ) The guanine and cytosine content of genomic DNA and bacterial evolution. Proc. Natl Acad. Sci. USA, , 84, , 16616–16619.
Schaefer,Ch. and Messer,W. ( (1991) ) DnaA protein/DNA interaction. Modulation of the recognition sequence. Mol. Gen. Genet., , 226, , 34–40.
Moran,N.A. ( (2002) ) Microbial minimalism: genome reduction in bacterial pathogens. Cell, , 108, , 583–586.
Kumar,S., Tamura,K. and Nei,M. ( (1993) ) MEGA: Molecular Evolutionary Genetics Analysis. Pennsylvania State University, University Park, PA.
Tamura,K. and Nei,M. ( (1993) ) Estimation of the number of nucleotide substitutions in the control region of mitochondrial DNA in humans and chimpanzees. Mol. Biol. Evol., , 10, , 512–526.(Pawe Mackiewicz, Jolanta Zakrzewska-Czer)