Orderly Replication and Segregation of the Four Replicons of Burkholderia cenocepacia J2315

Abstract

Bacterial genomes typically consist of a single chromosome and, optionally, one or more plasmids. But whole-genome sequencing reveals about ten per-cent of them to be multipartite, with additional replicons which by size and indispensability are considered secondary chromosomes. This raises the questions of how their replication and partition is managed without compromising genome stability and of how such genomes arose. Vibrio cholerae, with a 1 Mb replicon in addition to its 3 Mb chromosome, is the only species for which maintenance of a multipartite genome has been investigated. In this study we have explored the more complex genome of Burkholderia cenocepacia (strain J2315). It comprises an extra replicon (c2) of 3.21 Mb, comparable in size to the3.87Mb main chromosome (c1), another extra replicon(c3) of 0.87 Mb and a plasmid of 0.09 Mb. The replication origin of c1 is typically chromosomal and those of c2 and c3 are plasmid-like; all are replicated bidirectionally. Fluorescence microscopy of tagged origins indicates that all initiate replication at mid-cell and segregate towards the cell quarter positions sequentially, c1-c2-p1/c3. c2 segregation is as well-phased with the cell cycle as c1, implying that this plasmid-like origin has become subject to regulation not typical of plasmids; in contrast, c3 segregates more randomly through the cycle. Disruption of individual Par systems by deletion of parAB or by addition of parS sites showed each Par system to govern the positioning of its own replicon only. Inactivation of c1, c2 and c3 Par systems not only reduced growth rate, generated anucleate cells and compromised viability but influenced processes beyond replicon partition, notably regulation of replication, chromosome condensation and cell size determination. In particular, the absence of the c1 ParA protein altered replication of all three chromosomes, suggesting that the partition system of the main chromosome is a major participant in the choreography of the cell cycle.

Fig 1. Location of replication regions on the Bcen replicons.

A. Base disparity along c1, c2 and c3, shown as Z-curves (Ori-finder; [22]). Each chromosome sequence is arrayed to place nucleotide 1 at the centre of the x-axis. Z-curves display base distribution asymmetry (left ordinate): with x_n = purine vs. pyrimidine (blue trace) and y_n = amino vs. keto (yellow trace), GC- disparity (green trace) and AT-disparity (red trace) are defined as (x_n—y_n)/2; see [22] for a full explanation. The density of DnaA boxes (right ordinate) is calculated as follows: the distance from each DnaA box (≤ 1 deviation from TTATCCACA) to its adjacent bases are summed and the reciprocals (b values) plotted [24]; each diamond shows a cluster, defined as ≥ 3 DnaA boxes within 100 bp. B. Distribution of putative KOP sites on the Bcen chromosomes, opened at nucleotide 1 and read clockwise. The sites are assumed to correspond to the E.coli KOP sequence GGG(C/A/T)AGGG on the basis of nearly complete sharing of conserved and contact residues with the KOPS interaction domain of E.coli FtsK (S3 Fig; [25]). Arrowheads show dif sites corresponding to the Burkholderiales family consensus [26]. C. Ori-par regions of each chromosome, showing the principal elements (detailed maps are given in S1 and S2 Figs) denoted as follows: “ori”—sequence deduced to contain the replication origin by Ori-finder; nt1—nucleotide 1 in the genome database; par–parAB partition genes; rep–homologue of genes of the replication control protein (RepA) family; iterons—cluster (green rectangles) of 19-21bp sequence repeats characteristic of plasmid replication control elements (detailed in S1 Table); GC min—minimum in GC disparity curve; A—DnaA boxes (those with >1 deviation from the E.coli TTATCCACA consensus are omitted); S–parS sites of ParB binding; A/S–parS sites denoted as DnaA boxes by Ori-finder.

Fig 2. Replication characteristics determined by high-throughput sequencing.

A. Base-pair frequency gradients. Data points are averages of binned read numbers representing successive blocks of 10kbp for c1 and c2 and 1kbp for c3. The top panels show the raw data for DNA from exponentially-growing cells, the third row shows the same data after division by the correspondingly binned data obtained from stationary-phase cells (second row). Nucleotide positions on the abscissa are reversed to conform to the intuitive sense of right and left chromosome arms. Because the data are plotted as raw read frequencies, relative copy numbers of the replicons can be read from the ordinates. B. Calculation of chromosome replication period, C, and speed. Origin/terminus ratios were used to calculate the time taken to replicate each chromosome arm from ori/ter = 2^C/τ. In the case of c1 and c2, the concavity of the stationary-phase base-pair frequency curves would falsify calculation of origin/terminus ratios from normalized data, necessitating use of the raw data plot. For this, raw data ori- and ter-proximal points that corresponded to points intersected by the linear regression plot of the normalized data were connected by lines whose upper extremities and intersection were taken as the ori and ter values respectively. The c3 stationary-phase bp frequency curve is essentially flat, validating the normalized data. The * values for c3 replication speed are calculated on the basis of terminus displacement creating arms estimated to be 340 and 530 kb long.

Fig 3. Position of origins relative to cell poles.

A. Graphical summary of distances from the pole nearest a focus to the foci in one- and two-focus cells, and below, a plot of inter-focus distances. B. The positions of foci formed on ori-proximal sites by (c1) ParBc1::Gfp, (c2) ParBc2::Chfp, (c3) ParBP1::Gfp and (p1) ParB::Chfp were measured relative to the cell pole nearest a focus and plotted against cell length. Distances in cells with a single focus are shown as black-bordered circles in all cases; distances in cells with two foci are shown as coloured symbols. Arrows indicate the beginning of the two-focus clusters (left) and the end of the main one-focus clusters (right). Insets show numbers of cells scored, in corresponding colours. C. Examples of cells showing ParB::fp-marked origin regions. Scale bar is 1 μm.

Fig 4. Positions of two origins visualized simultaneously.

A. The origins of two replicons are plotted relative to the distance from the pole nearest any focus. Insets show numbers of cells with two foci of the replicon correspondingly coloured in the plot, comprising those with either one or two foci of the other replicon visualized; cells lacking foci of either replicon were excluded from the analysis. Red and green discs indicate respectively Chfp and Gfp fusions used to mark the origins shown. B. Examples of cells with origins of two replicons co-visualized. The separate components of the overlays are shown as examples of the images on which length measurements were made. Scale bar is 1 μm. C. Frequencies of focus combinations, shown as percentage of total cells scored.

Fig 5. Effect of ParABS disruption on distribution of replication origins.

The chromosome origins of Δpar and corresponding wild type strains were marked with ParB::Fp proteins, and the distances of fluorescent foci from the most focus-proximal pole were measured and binned in intervals of one-tenth cell length. The ParABS disrupting factors are shown at the right: deletion mutations in oric1- and oric2-marked cells (grown in MglyC), plasmid-borne parS sites in oric3-marked cells (grown in SOB). Origin distributions in cells with undisturbed Par function are shown as shaded areas. Numbers of one- and two-focus cells analyzed are shown at the right.

Fig 6. Replication characteristics of chromosomes in ΔparAc1 mutant cells.

A. DNA purified from exponentially-growing FBP47 (Nel13 ΔparAc1) cells was purified, processed and analyzed in parallel to wild type (Nel13) DNA, as outlined in the Fig 2 legend. The FBP47 data are shown in orange superimposed on the Nel13 data in blue (transposed from Fig 2). Raw read numbers are shown for c1 and c2: these reflect relative quantities of loci within a given strain but not of a given replicon between wt and mutant strains. B. Positions of c1 and c2 origins in wild type and ΔAc1 mutant cells growing in MglyC medium at 30°C. Pole-to-focus distances are shown black for one-focus cells and coloured for two-focus cells, as in Fig 3. In these conditions, cells grow more slowly (τ ~140 mins) than in the previous analyses of focus position shown in Figs 3 and 4 (τ ~110 mins), and the cells are correspondingly smaller.

Fig 7. Role of Par systems in cell growth and morphology.

A. Size of colonies formed at 3 days from cells transformed by plasmids carrying the parS sites indicated, either as single sites (e.g. Sc2) or as the natural clusters (e.g. Sc2+). Bcen makes pigment as cells enter stationary phase: faster-growing colonies are thus brown, slower-growing colonies still white. B: left panels Time of appearance of colonies after spreading on the agar media shown. Colonies were scored every 8 hours until the count reached its maximum; middle panels Cells grown exponentially in SOB were fixed, stained with DAPI and scored by fluorescence microscopy; right panels Viability on solid media. Cells grown in SOB to OD₆₀₀ ~1.0 were diluted and plated on LB agar either as a 0.1ml sample spread with glass beads or as a 10μl drop. The numbers of colonies are expressed as the ratio to the wild type colony count corrected for differences in OD. C. Cell dimensions. Cells growing exponentially in SOB at 37°C were fixed for viewing by phase contrast microscopy and their length (top panels) and width (bottom panels) were measured. The reference strains for the ΔA(B) mutants and parS-transformants were Nel13 without and with pMMBΔ respectively. D. Illustrative examples of abnormal cell phenotypes characteristic of mutants ΔparAc1 (top, and see S8 Fig) and ΔparABc2 (bottom). The nucleoids of the ΔparABc2 cells are revealed by DAPI staining. Scale bars are 2 μm.

Fig 8. Segregation and positioning of replication origin regions during the Bcen cell cycle.

Sequential partition in the order c1 –c2 –c3 –p1 is depicted. All origins are shown as gravitating to the centre of cell halves before division, as deduced from their central position in single-focus cells.