Portal de Periódicos da CAPES

The two chromosomes of Vibrio cholerae are initiated at different time points in the cell cycle

2007; Springer Nature; Volume: 26; Issue: 13 Linguagem: Inglês

10.1038/sj.emboj.7601747

1460-2075

Tue Rasmussen, Rasmus Bugge Jensen, Ole Skovgaard,

Salmonella and Campylobacter epidemiology

Article7 June 2007free access The two chromosomes of Vibrio cholerae are initiated at different time points in the cell cycle Tue Rasmussen Tue Rasmussen Department of Science, Systems and Models, Roskilde University, Roskilde, DenmarkPresent address: Department of Pharmaceutics and Analytical Chemistry, University of Copenhagen, Universitetsparken 2, 2100 København Ø, Denmark Search for more papers by this author Rasmus Bugge Jensen Rasmus Bugge Jensen Department of Science, Systems and Models, Roskilde University, Roskilde, DenmarkPresent address: LEO Pharma, Industriparken 55, 2750 Ballerup, Denmark Search for more papers by this author Ole Skovgaard Corresponding Author Ole Skovgaard Department of Science, Systems and Models, Roskilde University, Roskilde, Denmark Search for more papers by this author Tue Rasmussen Tue Rasmussen Department of Science, Systems and Models, Roskilde University, Roskilde, DenmarkPresent address: Department of Pharmaceutics and Analytical Chemistry, University of Copenhagen, Universitetsparken 2, 2100 København Ø, Denmark Search for more papers by this author Rasmus Bugge Jensen Rasmus Bugge Jensen Department of Science, Systems and Models, Roskilde University, Roskilde, DenmarkPresent address: LEO Pharma, Industriparken 55, 2750 Ballerup, Denmark Search for more papers by this author Ole Skovgaard Corresponding Author Ole Skovgaard Department of Science, Systems and Models, Roskilde University, Roskilde, Denmark Search for more papers by this author Author Information Tue Rasmussen1, Rasmus Bugge Jensen1 and Ole Skovgaard 1 1Department of Science, Systems and Models, Roskilde University, Roskilde, Denmark *Corresponding author. Department of Science, Systems and Models, 18-1, Roskilde University, Universitetsvej 1, Roskilde 4000, Denmark. Tel.: +45 4674 2405; Fax: +45 4674 3011; E-mail: [email protected] The EMBO Journal (2007)26:3124-3131https://doi.org/10.1038/sj.emboj.7601747 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info The bacterium Vibrio cholerae, the cause of the diarrhoeal disease cholera, has its genome divided between two chromosomes, a feature uncommon for bacteria. The two chromosomes are of different sizes and different initiator molecules control their replication independently. Using novel methods for analysing flow cytometry data and marker frequency analysis, we show that the small chromosome II is replicated late in the C period of the cell cycle, where most of chromosome I has been replicated. Owing to the delay in initiation of chromosome II, the two chromosomes terminate replication at approximately the same time and the average number of replication origins per cell is higher for chromosome I than for chromosome II. Analysis of cell-cycle parameters shows that chromosome replication and segregation is exceptionally fast in V. cholerae. The divided genome and delayed replication of chromosome II may reduce the metabolic burden and complexity of chromosome replication by postponing DNA synthesis to the last part of the cell cycle and reducing the need for overlapping replication cycles during rapid proliferation. Introduction Most bacteria contain only one circular chromosome and bidirectional replication is initiated at a single origin. However, a few bacteria contain a genome distributed on multiple chromosomes and some have unusual genome structures like linear chromosomes. What are the evolutionary advantages of dividing the genome into several chromosomes, when most bacteria have not evolved so? Multiple chromosomes are found in diverse prokaryotic phyla including the Deinococcus–Thermus, Spirochaetes and the α-, γ- and β-subgroups of Proteobacteria. Interestingly, most of the bacteria with multiple chromosomes have several different life cycles and interact with host organisms, like Vibrio cholerae that lives in brackish water and within human hosts, making it likely that dividing the genome into several different replicons may be favourable for organisms living under variable environmental conditions (Egan et al, 2005). The genome of V. cholerae is distributed onto two circular chromosomes (Trucksis et al, 1998; Heidelberg et al, 2000), with a large 2.96 Mb chromosome I and a small 1.07 Mb chromosome II. This division is seen in other Vibrio bacteria as well, but the relative sizes of the secondary chromosomes vary significantly between different Vibrio species (Okada et al, 2005). The splitting of the V. cholerae genome onto two chromosomes could facilitate an up- or downregulation of genes present on one of the chromosomes by an amplification or elimination of one of the chromosomes under certain environmental conditions (Heidelberg et al, 2000). A study of gene expression in V. cholerae shows a preferential increase in the transcription of genes from the small chromosome II during colon colonisation compared to cells growing in vitro under aerobic conditions (Xu et al, 2003). Thus, possessing multiple chromosomes may give an additional mechanism for switching between different portfolios of genes needed in various environments. Bacteria tightly regulate genome replication and this process is primarily controlled at the level of initiation of chromosomal DNA replication. The presence of several chromosomes makes the coordination of genome replication more difficult compared to when only a single chromosome is present. Some bacteria grow very fast under optimal conditions, with a doubling time much shorter than the time required for DNA replication and segregation, and therefore have overlapping DNA replication cycles. V. cholerae cells are able to divide every 18 min under optimal conditions. How do bacteria with multiple chromosomes and fast cell cycles ensure the faithful replication and distribution of several chromosomes to daughter cells? Most of the experimental data on chromosome replication and segregation in bacteria with divided genomes have been obtained using V. cholerae as model organism. A Meselson–Stahl density shift experiment showed that both V. cholerae chromosomes are replicated only once per cell cycle and with an inter-replication time equal to the doubling time of the cells (Egan et al, 2004). Thus, the replication of both chromosomes in V. cholerae are linked to the cell cycle (Egan et al, 2005). In contrast, plasmid replication is normally not tied to the cell cycle; plasmids tend to replicate at random time points during the cell cycle. Flow cytometry analysis of samples of slow growing cells treated with rifampicin and cephalexin indicated that the two chromosomes are initiated at the same time in the cell cycle (Egan et al, 2004). Examination of the replication control systems of the two V. cholerae chromosomes showed that each chromosome is controlled by its own initiator molecule: replication of chromosome I is controlled by the DnaA protein, similarly to the control of chromosome replication in Escherichia coli (Egan and Waldor, 2003). Increased expression of the DnaA protein stimulates initiation of chromosome I, but not of chromosome II (Duigou et al, 2006). Chromosome II has a plasmid-like origin, its replication is controlled by the RctB protein and requires the DnaA protein (Egan and Waldor, 2003). P1 and F plasmids shows a similar requirement for DnaA protein (Hansen and Yarmolinsky, 1986) despite that replication is controlled by the respective replication proteins RepA and RepE. The role of DnaA protein in these and other plasmids is strand opening, helicase loading or an architectural role in forming the replication complex (reviewed by Messer, 2002) and a similar role for the DnaA protein in chromosome II replication is probable. The RctB protein is autoregulated and the activity of RctB is further controlled by the rctA titration sites. The titration activity of rctA is controlled by the RctB protein (Venkova-Canova et al, 2006). Induction of the RctB protein stimulates initiation of chromosome II, but not chromosome I (Duigou et al, 2006). The separate control of replication initiation of the two V. cholerae chromosomes makes it difficult to understand how initiation synchrony can be obtained in this organism. This led us to re-evaluate the finding of initiation synchrony by Egan et al (2004). Using novel flow cytometry methods and marker frequency analysis, we show that initiation of the smaller chromosome II is delayed compared to initiation of chromosome I, resulting in completion of replication of the two chromosomes at approximately the same time in the cell cycle. Results and discussion Models for coordinating replication of the two chromosomes in V. cholerae The cell cycle of slowly growing bacteria is divided into three time periods with respect to chromosomal DNA replication: B is the period from cell birth to initiation of chromosome replication, C is the period from initiation until termination of chromosome replication and D is the period between termination and cell division (Figure 1). Thus, cells in the B period are small and contain one genome equivalent of DNA, cells in the C period are intermediate in size and contain between one and two genome equivalents of DNA and cells in the D period are large and contain two genome equivalents of DNA. Figure 1.Models of the DNA replication patterns during the V. cholerae cell cycle assuming either initiation or termination synchrony. Shown are chromosome I (large circles) and chromosome II (small circles). Origins are marked with filled circles and termini are marked with triangles. The replication speed is assumed to be equal for both chromosomes and to be constant during the entire C period. Download figure Download PowerPoint Both of the chromosomes in V. cholerae are replicated once and only once during the cell cycle and replication run-off experiments support an 'initiation synchrony' model (see Figure 1), where the two origins are initiated at the same time and DNA replication proceeds bidirectionally on both chromosomes until the terminus of each chromosome is reached (Egan et al, 2004). Owing to the smaller size, replication of chromosome II will terminate earlier than replication of chromosome I. According to the initiation synchrony model, the number of oriCI will equal the number of oriCII, whereas the number of terI will be lower than the number of terII. The total rate of DNA accumulation will be highest at the onset of the C period when both chromosomes are replicating and decrease when replication of chromosome II has terminated. Alternatively, initiation of the smaller chromosome II could take place at a later point in the cell cycle. In the 'termination synchrony' model (see Figure 1), chromosome II is initiated when the size of the non-replicated part of chromosome I is roughly equivalent to the size of chromosome II. Replication of both chromosomes then proceeds until the replication forks reach the termini at approximately the same time in both chromosomes. According to the termination synchrony model, the number of terI will equal the number of terII, whereas the number of oriCI will be higher than the number of oriCII. The rate of DNA accumulation will be highest late in the C period when both chromosomes are replicating. The DNA replication rate increases abruptly during the C period To determine which of the two models in Figure 1 that best describes the timing of replication of the two chromosomes in V. cholerae, we measured the DNA accumulation rate as a function of cell age. The DNA content and cell size of exponentially growing cells were measured using flow cytometry and the data are visualised by the 3-dimensional (3D) presentations in Figure 2A. The ridges of the 3D distributions were computationally determined as described in Materials and methods. The graph for exponentially growing Caulobacter crescentus cells, a one-chromosome bacteria with distinct B, C and D periods, is shown in Figure 2A. The first horizontal part of the ridge graph represents the B period, where only cell size increases. This is followed by the C period where both cell size and DNA content gradually increase, and the final horizontal line is the D period. Similar results were obtained with an E. coli strain (Supplementary Figure S1). When the same analysis is performed on an exponentially growing V. cholerae culture, the graph still displays B, C and D periods, but a bend is evident in the C period part of the graph (Figure 2A and Supplementary Figure S1). Since the cell size most likely increases gradually during the cell cycle, this bend must correspond to a change in the rate of DNA synthesis (Figure 2B). The slope of the ridge plot increases in the last part of the C period, indicating that V. cholerae cells abruptly increase their rate of DNA synthesis late in the C period. We repeated this experiment with varying carbon sources to change the growth rate (Table I) and in all cases we observed a very similar bend of the ridge representing the cells in the C period (Supplementary Figure S1). Figure 2.Flow cytometry analysis of chromosome replication. (A) 3D flow cytometry analysis of DNA replication patterns in C. crescentus and V. cholerae. The left panels show the raw 3D flow cytometry data with cell size on the x-axis, the DNA contents on the y-axis and the number of cells colour-coded as shown in the bar. The right panels show the ridges of the 3D graphs. The ridge graph for C. crescentus, a bacterium with one chromosome, shows three distinct parts, corresponding to the birth (B), chromosome replication (C) and division (D) periods of the cell cycle. The graph for V. cholerae shows a bend in the C period part of the ridge graph. (B) Models of the expected ridge graphs for V. cholerae with either initiation or termination synchrony. A higher slope of the ridge graph is expected in the part of the C period where both chromosomes replicate, since more replication forks are present. (C) Fitting of an experimental V. cholerae DNA histogram of exponentially growing cells to computer simulations of DNA contents in cells of an ideal culture assuming an equal speed of the replication forks throughout the entire C phase and different initiation times of chromosome II. Purple dots are actual DNA contents data, green lines are simulated B period cells, blue lines are simulated C period cells, red lines are simulated D period cells and black lines are the combined simulated cells. The left graph shows a simulation based on both chromosomes initiating at the same time. In the right graph, the simulation is based on a delay of initiation of chromosome II until two-thirds of chromosome I is replicated, resulting in simultaneous termination of the two chromosomes. Download figure Download PowerPoint Table 1. Estimates of the B, C and D periods of V. cholerae cells growing at moderate growth rates Growth medium Doubling time (min) B period (min) C period (min) D period (min) AB+sucrose 42 2 27 13 AB+maltose 55 10 31 14 AB+fructose 55 11 32 12 AB+glycerol+serine 80 22 39 19 AB+glycerol 114 40 49 25 The cell-cycle parameters were estimated by fitting simulations of DNA contents distributions (Figure 2C and Supplementary Figure S2) assuming termination synchrony to experimental flow cytometry DNA histograms. Simultaneous initiation of the two chromosomes would have led to a high rate of DNA replication at the onset of the C period and a decreased rate of DNA replication after termination of chromosome II (Figures 1 and 2B, Initiation synchrony) in contradiction with our results. In contrast, the increased rate of DNA replication late in the C period corresponds well with later initiation of chromosome II than chromosome I, leading to a doubling of the number of active DNA replication forks late in the C period (Figures 1 and 2B, Termination synchrony). Computer simulations of DNA contents distributions To support our conclusion that chromosome II replicate during the late part of the C phase and to determine basic cell-cycle parameters of V. cholerae, we compared the experimental DNA histograms obtained by flow cytometry of exponentially growing cells to computer simulations of DNA contents in ideal cultures using the approach described by Michelsen et al (2003) that was modified to accommodate two-chromosome bacteria. In these simulations, the DNA histograms are resolved into the contributions from cells in the B, C and D periods by varying the basic cell-cycle parameters and the time of initiation of chromosome II. We simulated the DNA contents assuming that replication of the two chromosomes either initiate or terminate simultaneously. We only obtained good fits between experimental and simulated DNA histograms when the two chromosomes terminate at approximately the same time in the computer simulations (Figure 2C). Similar results were obtained when V. cholerae cells were grown in other media, resulting in different doubling times (Supplementary Figure S2). The fitting of simulations of the DNA histograms to experimental data gave estimates of the basic cell-cycle parameters of V. cholerae cells growing at medium and slow growth rates (Table I). In cells growing at a moderate rate (42–55 min doubling time), the replication time for chromosome I was nearly 30 min (27–32 min) and the replication time for chromosome II was nearly 10 min (9–11 min), giving a replication speed of approximately 800 bp per second per replication fork, very similar to the replication speed in E. coli when grown at the same temperature (37°C). The D period was 12–14 min, which is similar to that observed in E. coli B/r cells and shorter than that observed in E. coli K-12 cells at similar growth rates (Michelsen et al, 2003). The oriCI to oriCII ratio is increased in fast-growing V. cholerae cells and is growth rate dependent The different sizes of the two chromosomes will lead to different copy numbers of either the origins or the termini of the two chromosomes (Figure 1). The initiation synchrony model predicts an oriCI to oriCII ratio of one and a terI to terII ratio less than one (Figure 1). The termination synchrony model predicts an oriCI to oriCII ratio higher than one and a terI to terII ratio of one (Figure 1). The differences in ter and oriC ratios, respectively, are most pronounced during rapid growth, where a larger fraction of the cells will be in the part of the cell cycle where chromosome I is under replication and replication of chromosome II is either finished or not initiated yet. This difference is modelled as described in Materials and methods and shown in Figure 3B and C. Figure 3.Relative marker frequencies (oriCI/oriCII and terI/terII ratios) at different growth rates. (A) Wild-type V. cholerae cells were grown in media containing different carbon sources, resulting in a range of doubling times. Cells from exponentially growing cultures were fixed and the relative marker frequencies were determined using quantitative PCR. Shown are the relative oriCI/oriCII (▪) and terI/terII(▴) marker frequencies. The line shows the predicted oriCI/oriCII ratio and the broken line shows the predicted terI/terII ratio assuming termination synchrony. (B) Predicted oriCI/oriCII ratio (line) and predicted terI/terII ratio (broken line) assuming initiation synchrony. (C) Predicted oriCI/oriCII ratio (line) and predicted terI/terII ratio (broken line) assuming termination synchrony. Predicted relative marker frequencies were calculated as described in Materials and methods assuming a constant C period at all growth rates. Download figure Download PowerPoint We measured the ratios of oriCI to oriCII and the ratios of terI to terII in V. cholerae cells where the doubling time was modulated between 18 and 164 min by varying the carbon source and the supplements of amino acids in the culture medium (Figure 3A). The oriCI to oriCII ratio dropped from ∼2 at fast growth rates to just above one at doubling times longer than 100 min, whereas the terI to terII ratio was almost invariant and near 1 for all doubling times. The results from the marker frequency analysis fit a termination synchrony model significantly better than it fit an initiation synchrony model (Figure 3). Initiation of replication at oriCI and oriCII responds differently to the transcription inhibitor rifampicin Our finding of later initiation at oriCII than at oriCI contradicts with the initiation synchrony observed by Egan et al (2004). Their observation was based on flow cytometry analysis of slowly growing V. cholerae cells followed by treatment with rifampicin and cephalexin. Rifampicin inhibits transcription, which then inhibits initiation of DNA replication, and cephalexin inhibits cell division. If chromosome II is initiated later than chromosome I, it was expected to observe cells that had started replication of chromosome I but not replication of chromosome II in the rifampicin run-off experiments. These cells should contain approximately 7.2 Mb DNA after rifampicin incubation, corresponding to two copies of chromosome I and one copy of chromosome II. However, Egan et al (2004) observed that all cells contained either 4.1 or 8.3 Mb DNA after rifampicin treatment, corresponding to either one or two copies of each chromosome and they therefore concluded that the two chromosomes are initiated synchronously. To resolve this contradiction, we measured the ratio of oriCI to oriCII in very fast-growing V. cholerae cells (doubling time of 18 min) at different time points after addition of rifampicin (Figure 4). We found that the oriCI to oriCII ratio was ∼2.6 before addition of rifampicin as expected from the experiment in Figure 3. After addition of rifampicin, this ratio dropped to near 1 after 10 min of incubation with rifampicin. This indicates that initiation at oriCII continued for a longer period after addition of rifampicin than initiation of chromosome I, until the two chromosomes had the same copy number. A double-sided t-test shows that the means of the 1 and 30 min points are different with 99% confidence. Similarly, the ratio of oriCI to oriCII in V. cholerae cells growing with doubling time of 55 min dropped from 1.4 to 1.0 after addition of rifampicin (data not shown). These results can only be explained if rifampicin has a delayed effect on initiation of chromosome II or if the chromosome I DNA is degraded during the rifampicin incubation. This second option is highly unlikely as the raw data show a constant amount of oriCI and an increasing amount oriCII. We therefore conclude that the initiations at the two origins are not equally sensitive to rifampicin, which explains the discrepancy between our results and the finding of initiation synchrony by Egan et al (2004). Figure 4.Differential effect of rifampicin addition on the two V. cholerae chromosomes. The oriCI/oriCII ratios at various time points after addition of rifampicin to a culture growing in LB glucose media (τ=18 min). The measurements were repeated at least four times and the s.e.m. are represented on the graphs. Download figure Download PowerPoint Bremer and Churchward (1977b) measured a delay of 6–11 min before the inhibition of initiation of E. coli chromosomal DNA replication by either rifampicin (inhibiting transcription) or chloramphenicol (inhibiting translation) was complete, indicating that cells near initiation can complete this process in the presence of the drugs. Further, there are several demonstrations of rifampicin-resistant initiation in E. coli strains where DnaA is in surplus (Pierucci et al, 1987; Atlung and Hansen, 1993; von Freiesleben et al, 2000; Morigen et al, 2005), indicating that the primary action of rifampicin in chromosome replication is the inhibition of de novo DnaA protein synthesis. As initiation at oriCI is controlled by the activity of the DnaA protein and initiation at oriCII is controlled by the activity of the RctB protein (Duigou et al, 2006), initiation at both origins is expected to require ongoing protein synthesis and therefore they should both be sensitive to rifampicin. Our demonstration that rifampicin inhibition is delayed for oriCII may indicate that the RctB protein bound to the rctA titration sites can be released during transcription arrest and initiate replication. Mechanisms ensuring approximately simultaneous termination of the two chromosomes A mechanistically simple mechanism for ensuring approximately simultaneous termination of the two chromosomes would be to link initiation of both chromosomes to cell mass, but with a different initiation mass for the two chromosomes. The two chromosomes are independently controlled by distinct replication initiators, DnaA for chromosome I and RctB for chromosome II (Duigou et al, 2006). The activity of RctB is controlled in much of the same fashion as the activity of DnaA is controlled in E. coli with both initiator autorepression and titration (Pal et al, 2005; Egan et al, 2006; Venkova-Canova et al, 2006). Thus, it is likely that chromosome II initiation is linked to cell mass in a similar fashion as the chromosome of E. coli (Donachie, 1968) or to a lesser degree plasmid P1 (Keasling et al, 1992; Bogan et al, 2001). Another possible molecular mechanism ensuring termination of the two chromosomes at approximately the same time during the cell cycle could be the presence of a timer on chromosome I that when replicated sends a signal to chromosome II. In other bacteria, a change in methylation status is known to effect transcription of a number of genes. In C. crescentus, the passage of the replication fork leads to the presence of hemimethylated DNA, which activates the promoter of the cell-cycle control gene ctrA (Reisenauer and Shapiro, 2002; reviewed by Wion and Casadesus, 2006). Dam methylation could have a similar function at an unknown activator gene in V. cholerae. Concluding remarks In the current work, we have determined the replication timing for both of the chromosomes in V. cholerae by flow cytometry analysis of samples harvested from exponentially growing cultures. Two different data analysis methods led to the conclusion that the rate of DNA replication increases late in the C period in congruence with a model where approximately two-thirds of the large chromosome I is replicated before the smaller chromosome II is initiated, leading to approximately simultaneous termination of replication of the two chromosomes. We have further tested this conclusion independently by measuring the ratio of the copy numbers of the two chromosomal origins, oriCI to oriCII, and the ratio of the two termini, terI to terII, as a function of the doubling time. These data show that the origin ratio varies with growth rate, whereas the termini ratio is invariant. This finding further supports that replication of the two chromosomes is initiated at different points of the cell cycle, whereas termination of the two chromosomes happens at approximately the same time in the cell cycle. Our finding of approximate termination synchrony conflicts with the finding of initiation synchrony by Egan et al (2004) based on flow cytometry analysis of a rifampicin-treated sample. We show here that it takes longer time before rifampicin inhibits initiation at oriCII than at oriCI and that the copy number of the two chromosomes will be the same after a long rifampicin treatment, explaining the discrepancy between the results. Fogel and Waldor (2005) studied the intracellular locations of the two origins in fast-growing cells and found that 78% of the cells had two oriCI foci and only 30% of the cells had two oriCII foci. Consistent with this, Fiebig et al (2006) found that mainly mid-sized cells have two oriCI foci and one oriCII focus, long cells had two foci of each origin and short cells had one focus of each origin. The finding of fewer oriCII foci than expected from the initiation synchrony model was explained by much earlier segregation of oriCI than of oriCII during the cell cycle. Our finding of later initiation at oriCII than at oriCI and the subsequently higher copy number of oriCI compared to oriCII is consistent with the origin localisation data. Thus, both oriCI and oriCII foci are likely to segregate away from each other shortly after initiation. Srivastava et al (2006) studied the intracellular locations of the two termini in cells growing in either L broth or minimal glucose medium. They observed that most cells had only one focus of each terminus and that these foci colocalised in these cell. Only 2–3% of the cells had two terI foci and 17–18% of the cells had two terII foci, indicating that terII foci separates earlier than terI foci. With the age distribution (Powell, 1958), the cell age at terI and terII foci separation can be calculated to be later in the cell cycle than the termination of chromosome replication (calculations not shown). The earlier segregation of terII is therefore not in conflict with an approximately termination synchrony of the two chromosomes. Evolutionary advantages of having a divided genome Over-initiation in E. coli leads to a decreased replication speed (Atlung et al, 1987) and under-initiation results in an increased DNA replication rate (Atlung and Hansen, 2002; Morigen et al, 2003; Skovgaard and Løbner-Olesen, 2005), demonstrating that the normal replication rate is limited by the supply of energy, nucleotides, replication factors or other components. By delaying initiation at oriCII until as late as possible in the cell cycle, the cell postpones DNA synthesis unti