SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

To study the role of cell division in the process of nucleoid segregation, we measured the DNA content of individual nucleoids in isogenic Escherichia coli cell division mutants by image cytometry. In pbpB(Ts) and ftsZ strains growing as filaments at 42°C, nucleoids contained, on average, more than two chromosome equivalents compared with 1.6 in wild-type cells. Because similar results were obtained with a pbpB recA strain, the increased DNA content cannot be ascribed to the occurrence of chromosome dimers. From the determination of the amount of DNA per cell and per individual nucleoid after rifampicin inhibition, we estimated the C and D periods (duration of a round of replication and time between termination and cell division respectively), as well as the D′ period (time between termination and nucleoid separation). Compared with the parent strain and in contrast to ftsQ, ftsA and ftsZ mutants, pbpB(Ts) cells growing at the permissive temperature (28°C) showed a long D′ period (42 min versus 18 min in the parent) indicative of an extended segregation time. The results indicate that a defective cell division protein such as PbpB not only affects the division process but also plays a role in the last stage of DNA segregation. We propose that PbpB is involved in the assembly of the divisome and that this structure enhances nucleoid segregation.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

In bacteria, the chromosomal DNA is organized in a confined region, which gradually extends in size along with the growth of the cell (Mason and Powelson, 1958; Woldringh, 1976; Van Helvoort and Woldringh, 1994). Within this structure called the nucleoid, the majority of the DNA occurs in the form of supercoils, about half of which are restrained by relatively few bound proteins (Drlica and Riley, 1990). Interaction of the DNA supercoils with themselves and with the crowded environment of soluble proteins is thought to lead to DNA compaction and phase separation between nucleoid and cytoplasm (Valkenburg and Woldringh, 1984; Walter and Brooks, 1995; Odijk, 1998). To segregate the duplicating chromosomes, the cell has to overcome this physical compaction force. DNA or nucleoid segregation is defined as the process that moves the replicated daughter strands apart during a period (S) between initiation of DNA replication (i) and the event of structural separation of the nucleoids (ns).

As a possible mechanism of this segregation, it has been proposed that the nucleoids are expanded by numerous DNA loops attached to the plasma membrane through co-transcriptional and co-translational translocation of membrane proteins (Norris, 1995; Woldringh et al., 1995). The individually transient, but collectively continuous, attachments of the DNA could pull the nucleoids apart along with the elongating cell (Van Helvoort et al., 1998). The observation of fusion of separated nucleoids in Escherichia coli filaments treated with either rifampicin or chloramphenicol supports the existence of such a transcription/translation-mediated expansion force (Woldringh et al., 1994).

A more specific mechanism for segregation has been suggested on the basis of recent fluorescence microscope studies showing a bipolar movement of specific DNA regions, as visualized by antibodies in immunolabelling experiments (Sharpe and Errington, 1996; Niki and Hiraga, 1998). Abrupt movements of regions near the origin of replication have also been observed in living bacteria by labelling the Lac operator cassette with a fusion of the Lac repressor with green fluorescent protein (GFP; Gordon et al., 1997; Lin et al., 1997; Webb et al., 1997; 1998). However, the nature of such a dedicated mechanism that moves the DNA and helps to overcome the physical compaction force has remained elusive.

Apart from proteins involved in decatenation of the replicated chromosome, such as gyrase and topoisomerase IV, the proteins expressed by the E. coli muk operon have been suggested to form a motor for rapid DNA displacement (for a review, see Hiraga, 1992). Recently, the FtsK protein has also been implicated in chromosome segregation (Yu et al., 1998; see also Liu et al., 1998). In addition, Bouché and coworkers (Tétart et al., 1992) have implicated the cell division protein FtsZ in a role in the process of chromosome segregation. They observed a lengthening of the segregation period between termination of DNA replication and nucleoid separation in cells in which the level of FtsZ protein was reduced. However, their prediction that the cells would contain nucleoids with increased amounts of DNA was weakened because the use of chloramphenicol for microscopic visualization could have caused the fusion of nucleoids (Van Helvoort et al., 1996). Nevertheless, an involvement of the division process in the resolution of replicated chromosomes has been suggested recently by Steiner and Kuempel (1998). In addition, we have reported preliminary observations indicating that nucleoids, resegregating after release from chloramphenicol inhibition in E. coli pbpB cells growing as filaments, contained an increased amount of DNA, suggesting a delay in segregation (Van Helvoort et al., 1998).

For a better estimate of the efficiency of segregation in E. coli cells by microscopic methods, the amount of DNA per individual nucleoid has to be measured accurately. Therefore, we developed an image cytometric procedure for the determination of integrated fluorescence of nucleoids stained with DAPI (4′,6-diamidino-2-phenylindole dihydrochloride hydrate), with an accuracy similar to that of the flow cytometer (Vischer et al., 1999). With this method, we demonstrate in this paper that various cell division mutants (ftsZ, pbpB ), growing as filaments at 42°C, segregate their DNA in nucleoids containing on average more than two chromosome equivalents. In contrast to the other cell division mutants, E. coli pbpB cells already show an extended segregation period when growing under steady-state conditions at the permissive temperature (28°C). We propose that the formation of the FtsZ ring (Bi and Lutkenhaus, 1991) and the recruitment of other proteins for the assembly of the divisome (Nanninga, 1991), including the PbpB (FtsI) protein, enhance the final steps in nucleoid segregation, and we discuss a possible mechanism.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Increased amount of DNA per nucleoid in E. coli filaments growing at restrictive temperature

In previous work, we described the fusion and resegregation of nucleoids in E. coli pbpB filaments synchronized by centrifugal elutriation (Van Helvoort et al., 1998). We observed by image cytometry that, upon release from chloramphenicol inhibition, resegregating nucleoids presumably contained more than two chromosome equivalents of DNA. If nucleoids separated immediately upon termination of chromosome replication (Woldringh, 1976), the average DNA content per nucleoid is expected to be at most 1.5 chromosome equivalents. The increased amount of DNA per nucleoid could therefore reflect a delay in segregation. To verify whether this phenomenon also occurs in exponentially growing pbpB filaments and in filaments of other cell division mutants, we measured the amount of DNA per nucleoid with the image cytometric method described by Vischer et al. (1999). 1Figure 1A shows E. coli pbpB filaments grown at 42°C together with stationary phase pbpA cells used for calibration. In 1Fig. 1B–E, the distributions of DNA content per cell or filament and per nucleoid show that, compared with the parent strain growing at 42°C (Fig. 1B and C), the amount of DNA is higher in the pbpB strain (Fig. 1D and E). The average values, summarized in Table 1, confirm that nucleoids in pbpB and ftsZ filaments contain on average 2.2–2.6 chromosome equivalents versus 1.6 in the parent strain (also grown at 42°C), indicative of a delay in nucleoid segregation.

image

Figure 1. . Nucleoid segregation in cells growing at 42°C. A. Phase- contrast and fluorescence microscope images of pbpB filaments and pbpA calibration cells. Magnification bar represents 2 μm. B–E. Comparison of the distributions of DNA content of the parent strain E. coli LMC500 (B and C) and the pbpB strain (D and E), growing at 42°C for 130 min. Note the different scales. The bimodal, hatched distribution is from the pbpA cells in the same preparation. The average value of the left peak (CV = 6%) was assumed to represent one chromosome equivalent. See Table 1 for the average values of the distributions.

Download figure to PowerPoint

Table 1. . Image cytometric determination of the amount of DNA (in chromosome equivalents) per cell and per individual nucleoid in E. coli filaments growing at 42°C. a. Time at 42°C corresponded to 1.5–2 mass doublings.b. Stationary phase E. coli pbpA cells were used for calibration. Amounts of DNA were calculated by assuming that the value of the left peak of the DNA distribution of pbpA cells (see Figs 1D and 3B) equals one chromosome equivalent.c. Cells from a population synchronized by centrifugal elutriation as described previously (Van Helvoort et al., 1998).d. Cells were grown at 100 mosM.e. As a control, the isogenic parent strain E. coli LMC500 was grown at 42°C at 300 mosM. Similar results were obtained when grown at 42°C at 100 mosM.Thumbnail image of

Amount of DNA per nucleoid in E. coli cell division mutants growing at permissive temperature (28°C)

The image cytometric measurements of DNA per nucleoid are based on an evaluation of fluorescence images and on judgement as to whether nucleoids have separated or not (cf. Fig. 1A). In this regard, the fixation of the cells is important because the application of OsO4 causes some contraction of the nucleoids compared with nucleoid regions in living cells (Valkenburg et al., 1985). Estimates of the percentage of cells with separated nucleoids in living cells grown in the presence of 0.1 μg ml−1 DAPI showed a 15% increase after OsO4 fixation (E. Pas, unpublished). Consecutive measurements of the DNA content per nucleoid in samples taken from the same steady-state culture showed a coefficient of variation (CV) of 4% for the average values (see Table 4). Because there are no indications that various strains behave differently to OsO4 fixation, this method gives an accurate reflection of the segregation process.

Table 4. . Amount of DNA (in chromosome equivalents) and morphological parameters for the isogenic strains E. coli pbpB and pbpB recA, grown at 28°C (permissive temperature). a. Amounts of DNA calculated by assuming that the value of the left peak of the DNA distribution of stationary phase pbpA cells (see Fig. 1D) equals one chromosome equivalent.b. Same cell population as in previous row, measured after brief sonication of the cells. The data from this experiment have been used for the construction of 4Fig. 4B.c. Independent growth experiment; compare with the data of the two pbpB experiments in Table 2.d. Cells subjected to brief sonication. The longer doubling time is probably caused by filaments and DNA-less cells (see text and Fig. 2H).Thumbnail image of

From the above experiments with filaments, it appeared that nucleoids in the control cells of E. coli pbpB grown at 28°C also contained an increased amount of DNA compared with the parent strain. Therefore, we measured the nucleoid DNA content in different isogenic E. coli cell division mutants growing under steady-state conditions at permissive temperature (28°C).

The results summarized in Table 2 show that, in two independent growth experiments, pbpB cells contained an increased amount of DNA per cell (3.1 versus 2.0 chromosome equivalents in the parent) and per nucleoid (2.0 versus 1.6 chromosome equivalents in the parent). Table 2 also shows that pbpB cells were longer and had a higher percentage of constricted cells and of cells with separated nucleoids (54% versus 23–24% in the parent). The ftsZ and ftsA strains resembled the parent strain, whereas ftsQ also showed an increased amount of DNA per cell (2.6 chromosome equivalents) and an increased average cell length but a similar amount of DNA per nucleoid as the parent (1.7 chromosome equivalents).

Table 2. . Amount of DNA (in chromosome equivalents) and morphological parameters for E. coli LMC500 and its isogenic, temperature-sensitive derivatives grown at 28°C (permissive temperature). a. A doubling time of about 78 min for growth at 28°C is comparable with about 50 min for growth at 37°C.b. Amounts of DNA calculated by assuming that the value of the left peak of the DNA distribution of stationary phase pbpA cells (see Figs 1D and 3B) equals one chromosome equivalent.c. CV, coefficient of variation.d. The data from this experiment have been used for the construction of 4Fig. 4A. See Fig. 2 for distributions of DNA contents.e. Independent growth experiments.Thumbnail image of

Of all mutants examined at permissive temperature, E. coli pbpB cells showed the highest increase in the amount of DNA per nucleoid, suggesting a delay in nucleoid segregation even during growth at 28°C. In addition, the high percentage of constricted cells (45–51%; see below) indicates an extended constriction period. As will be argued below, both effects may have contributed to a postponement of cell division and to the observed increase in average cell length, which has already been reported in a previous electron microscope study by Taschner et al. (1988).

Parent strain, E. coli LMC500, treated with aztreonam

The different behaviour of E. coli pbpB cells compared with the other cell division mutants suggests that the temperature-sensitive mutation already affects the functioning of its gene product, penicillin binding protein 3 (PBP3 or FtsI, here referred to as PbpB) at 28°C. To test the possibility that the long cells and the concomitant increase in DNA content per nucleoid are caused by a reduced functioning of PbpB, we treated the parent strain LMC500 at 28°C with the PbpB-specific inhibitor aztreonam (2 μg ml−1; Adam et al., 1997). The results in Table 3 show that, after inhibition of cell division during one mass doubling, cell length increased from 2.55 μm to 4.16 μm and the average amount of DNA per cell from 1.9 to 3.7 chromosome equivalents. Remarkably, the DNA content of the segregating nucleoids increased from 1.5 to 2.1 chromosome equivalents after one mass doubling. This suggests that the inhibition of PbpB not only affects the division process but also causes a delay in nucleoid segregation.

Table 3. . Image cytometric determination of the amount of DNA (in chromosome equivalents) and morphological parameters for E. coli LMC500 treated with 2 μg ml−1 aztreonam at 28°C. a. Amounts of DNA calculated as indicated in Tables 1 and 2.b. Compare with the slightly higher values of Gc and Gn for the parent strain in Table 2.Thumbnail image of

Amount of DNA per nucleoid in E. coli pbpB recA growing at permissive temperature (28°C)

An increase in the average amount of DNA per nucleoid in E. coli pbpB cells compared with the parent strain growing at the same doubling time could be caused by (i) the formation of chromosome dimers by sister chromatid exchange in a fraction of the cells (Steiner and Kuempel, 1998); (ii) an increase in the DNA replication period, C, inducing multifork replication (Helmstetter et al., 1979); or (iii) a delay in the segregation of replicated chromosomes.

Steiner and Kuempel (1998) have demonstrated that cell division is required to resolve the dimers that form with a frequency of about 14% per chromosome per generation. It could therefore be argued that, in the case of an extended constriction period in the pbpB strain, impairment of the function of the PbpB protein slows down the resolution of dimers. To test this possibility, we compared the DNA content of pbpB cells with that of an isogenic pbpB recA strain in which dimers cannot be formed by homologous recombination (Steiner and Kuempel, 1998). Because in the previous experiments (Table 2), it was observed that the pbpB strain exhibited a high percentage of constricted cells (45–51%), many of which showed deep constrictions, we now subjected the fixed pbpB and pbpB recA cells to a brief sonication to obtain a better separation of divided cells. As shown in Table 4, in the pbpB cells, this caused a decrease of about 10% in the percentage of both constricted cells and cells with separated nucleoids, whereas it had no effect on the parent strain (result not shown). The results presented in Fig. 2 and Table 4 show that the DNA distributions and the average amounts of DNA per cell and per nucleoid in the pbpB and pbpB recA strains are similar. We therefore conclude that the increased DNA content of nucleoids compared with the parent strain cannot be ascribed to the persistence of dimers. This confirms our impression of a rather homogeneous population of nucleoids in E. coli pbpB (Fig. 2G), with no indication of special segregation problems in only a fraction of the cells.

image

Figure 2. . Nucleoid segregation in cells growing at 28°C. A–F. Comparison of the distributions of DNA of the parent strain E. coli LMC500 (A and B), the pbpB strain (C and D) and the pbpB recA strain (E and F) growing at 28°C. The last two cell populations were sonicated briefly (see text). Note the increased amount of DNA per cell (C and E) and per nucleoid (D and F) in the pbpB and pbpB recA strains (cf. Table 4). Note that the distributions of DNA per nucleoid are rather broad because they represent all the replicative states of the chromosomes. The vertical dashed lines indicate the peaks of the calibration cells. G and H. Phase-contrast and fluorescence microscope images of pbpB (G) and pbpB recA (H) cells. Arrows indicate nucleoids in a state of DNA degradation. Magnification bar represents 2 μm.

Download figure to PowerPoint

It should be noted that the recA mutation was observed to cause chromosome degradation exactly as reported by Skarstad and Boye (1993). From the distributions in 2Fig. 2E and F, we therefore deleted the cells (usually filaments; see arrows in Fig. 2H) with nucleoids containing less DNA than the minimal amount observed in the calibration cells (15% of the population). Probably, these aberrant cells are the cause of the longer doubling time of the culture (95 min; Table 4).

Determination of the C period by image cytometry

To investigate whether the increased DNA content of pbpB nucleoids can be ascribed to multifork replication, we determined the C period of the parent strain and of pbpB cells by applying a method similar to that described by Bipatnath et al. (1998). Exponentially growing cells were treated with 300 μg ml−1 rifampicin to stop further initiations, and the accumulation of DNA in individual nucleoids was determined by image cytometry.

The data in 3Fig. 3B (left) and Table 5 show that, in the LMC500 cells, the average amount of DNA per cell remained more or less constant (2.0 chromosome equivalents) because of a combination of run-out DNA synthesis and residual division. The average amount of DNA per individual nucleoid, however, increased gradually from 1.6 to 1.9 chromosome equivalents. Assuming that all existing replication forks terminate in the presence of rifampicin and that these fully replicated chromosomes do not segregate, all nucleoids with one replicating chromosome will accumulate DNA and will eventually consist of two chromosome equivalents. The time required for this accumulation to reach a plateau (Fig. 3A and Table 5[link]) and for the distribution of DNA content per nucleoid to reach an almost Gaussian shape (Fig. 3B) are measures of the C period (Bipatnath et al., 1998). According to this method, C was estimated to be 70 min for both the parent strain (see shape of distributions on the right of Fig. 3B) and the pbpB mutant.

image

Figure 3. . Determination of the C period from the time during which the amount of DNA per individual nucleoid accumulates after stopping new initiations with rifampicin (Rif) in E. coli LMC500 and pbpB. A. Average amount of DNA per nucleoid in chromosome equivalents before and after the addition of 300 μg ml−1 rifampicin. The accumulation reaches a plateau value after about 60 min. Open circles, LMC500; closed circles, pbpB strain. B. Distributions of DNA content per cell and per nucleoid in E. coli LMC500. Average values are given in Table 3. Note that there is no further change in the shape of the distribution of DNA per nucleoid after 70 min. The small fraction of cells with one chromosome equivalent can be attributed to cells that have just terminated DNA replication at the time of rifampicin addition and have subsequently divided (see Fig. 4A).

Download figure to PowerPoint

Table 5. . Image cytometric determination of the accumulation of DNA (in chromosome equivalents) and morphological parameters after treatment of E. coli LMC500 with rifampicin. a. Time (min) after the addition of 300 μg ml−1 rifampicin. See 3Fig. 3A and B.b. Mass doubling time was 78 min. Three consecutive samples were taken from the same steady-state culture before treatment with rifampicin. The CV of the average values of Gc and Gn obtained from the three independent measurements is 4%.Thumbnail image of

The average value of the distribution of DNA content per nucleoid in the parent strain (Fig. 3A) reached a value of 1.93 (instead of 2). This is the result of a small number of nucleoids containing one chromosome equivalent (see Fig. 3B, bottom). The DNA content per nucleoid in the pbpB (recA+) cells reached a value of 2.4 chromosome equivalents (Fig. 3A), being an average of nucleoids separated into entities with two and nucleoids with four chromosome equivalents (see also Fig. 4).

image

Figure 4. . Schematic representation of the change in nucleoid DNA content and of the cell cycles of E. coli LMC500 (A) and the pbpB mutant (B), both growing at 28°C with a doubling time of 78 min (see Tables 2 and 3 respectively). The cell cycles start with the initiation of DNA replication of a virtual, unreplicated chromosome (dashed lines). See Table 6 for the calculation of the various cycle periods. i1 and t1, initiation and termination of replication corresponding to division at 78 min; i2 and t2, initiation and termination corresponding to subsequent cycles; ns, nucleoid separation; ic, initiation of constriction; d, cell division; C, duration of a round of replication; D′, time between t and ns; T, duration of cell constriction; S, period of DNA segregation assumed to start at i1 and to end at ns.

Download figure to PowerPoint

The C value of 70 min for cells grown at 28°C compares well with the period of 55 min for a 37°C culture determined by Bipatnath et al. (1998). We conclude that the increased nucleoid DNA content in the pbpB strain does not result from multifork replication.

Calculation of the D and D′ periods

To establish the cause of the increased DNA content in pbpB nucleoids, we calculated the D, D′ and T periods for the various strains growing under steady-state conditions at 28°C. Similar to D, the time between termination of replication and cell division, the D′ period can be defined as the time between termination and nucleoid separation.

Knowing the amount of DNA per cell (Gc) and the duration of C, the D period can be derived using the formula given by Helmstetter et al. (1979). Using a similar formula (see note c in Table 6), the D′ period can be calculated from the amount of DNA per nucleoid (Gn) and C. The results summarized in Table 6 indicate that the pbpB cells have a long D period (82 min versus 42–47 min in the parent) largely exceeding the doubling time. Also, the D′ period is twice as long in pbpB (42 min) compared with the parent strain (18–20 min). This does not apply to the other cell division mutants, ftsQ, ftsA and ftsZ, which have D′ periods of 18–22 min, when assuming the same C period of 70 min.

Table 6. . Cell cycle periods calculated for E. coli LMC500 parent strain and different cell division mutants at 28°C. See Fig. 4 for a schematic representation of the cell cycles.a. C, the period required to replicate the chromosome. A value of 70 min was obtained for LMC500 and pbpB by image cytometric measurement of DNA accumulation in the nucleoids of cells treated with rifampicin (see Fig. 3; cf. Bipatnath et al., 1998).b. D, the period between the termination of DNA replication and cell division, was calculated from the DNA content per cell, Gc, using the relationship: Gc = {Td/Cln2}{2(C+D)/Td− 2D/Td} (Helmstetter et al., 1979), using C = 70 min.c. D′, the period between the termination of replication and nucleoid segregation, was calculated from the DNA content per nucleoid, Gn, using an equivalent formula to that in (b): Gn = {Td/Cln2}{2(C+D′)/Td− 2D′/Td}.d. T, the duration of the constriction period, was calculated from the fraction, F, of cells showing a visible constriction (Table 1), using the relationship: T = {Td ln(1 + F)} ln 2.e. Values of C assumed to be 70 min.f. Sonicated cells; see Table 4. Values used for the construction of 4Fig. 4B.Thumbnail image of

With the calculated values for D, D′ and T periods (Table 6), a schematic representation of the increase in DNA content per nucleoid and of the cell cycle has been constructed in Fig. 4 for the parent and the pbpB strains. It can be seen that pbpB cells (Fig. 4B) are born with a single nucleoid containing 2 × 1 chromosome equivalents, an amount that increases to three chromosome equivalents at the time of nucleoid separation, 40 min before division. After separation, the two nucleoids contain 1.5 chromosome equivalents increasing again to two per nucleoid at division. While the extended duration of the constriction or T period from 24 min in the parent to 37 min in the pbpB mutant (Table 6) is likely to be the primary defect caused by the mutation, it is insufficient to explain the present observations. We therefore propose that the mutation indirectly causes an additional delay in the process of segregation of replicated chromosomes, increasing the D′ period from 18 min in the parent to 42 min in pbpB.

The increased cell length of the pbpB cells (Tables 2 and 4) can be predicted by assuming that the mutation caused the (exponential) length increase of the parent strain to continue for an additional 40 min as a result of a delayed segregation and a slower constriction process (results not shown).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Determination of the amount of DNA per individual nucleoid in E. coli pbpB filaments growing at restrictive temperature revealed that the segregating nucleoids contained on average more than two chromosome equivalents. Surprisingly, this increased amount of DNA per nucleoid was also measured for pbpB cells in steady-state growth at permissive temperature. In contrast to other cell division mutants such as ftsA and ftsZ growing at 28°C, the pbpB cells not only showed an extended constriction period like ftsQ but also an extended D′ period, indicative of a delay in segregation (S in Fig. 4). Because similar results were obtained with the pbpB recA mutant (Fig. 2 and Table 4[link]), the increased DNA content cannot be ascribed to the occurrence of dimers that persist if cell division is inhibited (Steiner and Kuempel, 1998).

Assembly of the divisome

How can a defective cell division protein such as PbpB cause a delay in a process that precedes division? We propose that the assembly of the divisome occurs well before the initiation of constriction and that it enhances nucleoid separation. The FtsZ ring has been observed to assemble before nucleoid separation, around the time of termination of DNA replication in both E. coli K-12 and B/r strains (Den Blaauwen et al., 1999). Although cell division can take place in the absence of replication termination (Jafféet al., 1986; Cook and Rothfield, 1999), we speculate that, in a normal cell cycle, the replication forks collide at mid-cell (Lemon and Grossman, 1998) and that their dissociation represents a cytoplasmic signal (Woldringh and Nanninga, 1985) for the formation of a nucleation site initiating FtsZ assembly between the segregating nucleoids. After the formation of the Z-ring (Addinall et al., 1996), additional cytoplasmic (FtsA) and membrane-bound (ZipA, FtsQ, FtsL and PbpB) proteins are sequentially recruited, as indicated by recent immunofluorescence studies (Ma et al., 1996; Yu et al., 1998).

The observation that PbpB (FtsI) is one of the last to be located to the divisome (Weiss et al., 1999) does not necessarily imply that it joins the complex much later in time. Although PbpB is not required for assembly of the Z-ring (Pogliano et al., 1997), the divisome could already involve PbpB for its stability from the beginning. Therefore, we propose that, in a period of maturation (Addinall et al., 1996) in which the participation of PbpB in the divisome and its peptidoglycan synthesis activity gradually rises, the divisome already induces a local increase in envelope synthesis, leading eventually to the initiation of constriction (Nanninga, 1998). We envisage that this local increase in the rate of surface extension at mid-cell will generate more space along the length axis of the cell, allowing the nucleoids to move and separate. The observations of Nanninga and coworkers (unpublished) on synchronized E. coli cells have indicated that the increased rate of elongation induced by the addition of EDTA (cf. Nanninga et al., 1985) is indeed accompanied by an increase in the percentage of cells with separated nucleoids.

Delayed segregation

The present observations are consistent with an earlier suggestion by Tétart et al. (1992) that reduced levels of FtsZ caused a delay in nucleoid segregation. Although we criticized their observations because of their use of chloramphenicol to visualize the nucleoids (Van Helvoort et al., 1996), we now confirm their interpretation that nucleoids contain, on average, more DNA as a result of a delay between termination and the initiation of constriction. By growing cells at reduced levels of FtsZ, Palacios et al. (1996) only observed an increase in cell size and length, without a significant effect on the partition index (number of nucleoids μm−1 cell length). Because this parameter and the present measurement of DNA content per nucleoid both depend on the judgement as to whether nucleoids have separated or not, we do not understand this discrepancy.

Apart from recruiting Fts proteins, the Z-ring could also attract DNA-binding proteins such as MukB, which has been shown to bind with high affinity to FtsZ (Lockhart and Kendrick-Jones, 1998). If the function of MukB is similar to that of the Smc-like condensing protein in Bacillus subtilis (Graumann et al., 1998), it could cause a local compaction of DNA near the FtsZ ring. Such a local DNA compaction could exert a pulling force and a directionality for the strand-passing activity of topoisomerases involved in the decatenation of the replicated chromosomes. Such directionality is also required for the activity of the FtsK protein, which resolves chromosome dimers and prevents DNA from becoming trapped by the closing septum (Steiner et al., 1999). We thus propose that divisome assembly enhances the last stage of nucleoid segregation (t1 to ns in Fig. 4). The two mutations in pbpB2158(Ts) (Ayala et al., 1994) are localized in the so-called non-penicillin-binding module of the protein, and it has been suggested that this module provides specific interaction regions for other proteins of the divisome (Nguyen-Distèche et al., 1998); therefore, a defective PbpB protein could cause the divisome to be less stable, slowing down the surface extension necessary for undelayed nucleoid separation.

Segregation in filaments

What happens if the assembly of the divisome is inhibited altogether and cells continue to grow as filaments? It has generally been observed that even long filaments can contain evenly distributed nucleoids (Taschner et al., 1987; Mulder and Woldringh, 1989; Begg and Donachie, 1991; Tétart et al., 1992; Newman et al., 1998). However, such segregation also shows defects. Like others (Weiss et al., 1999), we have often observed irregularities in the segregation pattern of nucleoids in filaments, for instance an increase in the distance between the outer borders of the most distal nucleoids and the cell poles in pbpB filaments growing at 42°C (see Fig. 2C in Van Helvoort et al., 1998). Irregularities could also result from the persistence of dimers formed in a fraction of the cells by sister chromatid exchange (Steiner and Kuempel, 1998). Although we cannot exclude the presence of chromosome dimers, the present study demonstrates that segregation in the absence of division (at 42°C) affects all the nucleoids in the population and that segregation occurs in nucleoid entities that contain, on average, slightly more than two chromosome equivalents. Evidently, segregation of these nucleoids is eventually secured by the continuing length extension of the filaments.

Knowledge of the amount of DNA per nucleoid, and thus of the replicative state of the chromosome (cf. Fig. 4), is also essential to interpret fluorescence foci in bacteria labelled by fluorescence in situ hybridization (Sharpe and Errington, 1996; Niki and Hiraga, 1998) or with the repressor–GFP fusion system (Gordon et al., 1997; Lin et al., 1997; Webb et al., 1997; 1998). Integrated fluorescence measurements (Vischer et al., 1999) of the DNA content of individual nucleoids, as applied in this study, makes it possible to predict the number of fluorescent spots that can be expected in DNA localization studies.

Experimental procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

Bacterial strains and growth conditions

The following E. coli K-12 strains were used: LMC500, a lysA1 derivative of MC4100, and its isogenic derivatives pbpB2158(Ts), ftsQ, ftsA, and ftsZ, all temperature-sensitive mutants that have been previously described (Taschner et al., 1988). E. coli pbpA137 has been described by Woldringh et al. (1988). E. coli pbpB2158(Ts) recA56 (LMC520) was constructed from strain LMC510 by transduction using P1 grown on strain LMC133 (Taschner et al., 1988). All strains were grown in minimal medium containing (per l): 6.33 g of K2HPO4.3H2O, 2.95 g of KH2PO4, 1.05 g of (NH4)2SO4, 0.1 g of MgSO4.7H2O, 0.3 mg of FeSO4.7H2O, 7.1 mg of Ca(NO3)2.4H2O, 4 mg of thiamine, 4 g of glucose and 50 mg of lysine (pH 7.0). NaCl was added to adjust the osmolarity of the media to 300 mosM (Micro-Osmometer; Advanced Instruments). Only the ftsZ strain was grown at 100 mosM to prevent salt suppression (Taschner et al., 1988). Cell mass was monitored by optical density (OD) at 450 nm with a Gilford microsample spectrophotometer, and cell numbers were measured with an electronic particle counter (orifice diameter 30 μm). Cultures were considered to be in a steady state of growth if the average cell mass remained constant over time.

Preparation of cells

Cells were fixed by the addition of OsO4 to a final concentration of 0.1% and stored at 4°C for 2 weeks at most. For staining, DAPI was added to a final concentration of 2 μg ml−1. After at least 30 min at room temperature, cells were centrifuged in an Eppendorf centrifuge, resuspended in growth medium and concentrated about 50 × by a second centrifugation step. Of the concentrated cell suspension, 4 μl was placed on a ‘minislab’ of 1% agarose, prepared as described previously (Van Helvoort and Woldringh, 1994), to immobilize the cells on a flat agar surface. Where indicated, cells were sonicated for 2 s before preparation for light microscopy.

Integrated fluorescence measurements

For the quantification of the relative DNA content of individual nucleoids, we applied a newly developed method of image cytometry (Vischer et al., 1999). In short, we acquired digital images with a Princeton RTE-1317-k-1 cooled CCD camera (1317 by 1035 pixels; Princeton Instruments) mounted on an Olympus BH-2 conventional fluorescence microscope equipped with a 100× SPlan PL phase-contrast objective (NA = 1.25). Fluorescence was excited with a mercury high-pressure lamp using a dichroic filter cube containing a bandpass filter (transmission between 300 and 400 nm) and a dichroic mirror (cut-off wavelength at 400 nm). The fluorescence was separated from scattering light using a longpass filter with a cut-off wavelength at 420 nm.

An Apple Macintosh computer (Quadra 840AV) was used for image collection and exposure control. Images were recorded as 656 × 517 pixel arrays (2 × 2 binning) giving a final magnification of 53 nm per pixel.

Calibration cells

To estimate the relative DNA content in individual nucleoids, the cell suspensions to be measured were mixed with stationary phase E. coli pbpA cells. When grown in glucose minimal medium at 30°C, this strain has been shown to adopt a spherical shape, in contrast to its rod shape when grown in L broth at permissive temperature (Woldringh et al., 1988). Cells were used 48–96 h after entering the stationary phase. The DNA content of these cells reproducibly gave a bimodal distribution (Vischer et al., 1999). For unknown reasons, the number of cells in the two peaks varied between different experiments. However, the ratio of DNA content of the peaks was about two (Vischer et al., 1999). The left peak was assumed to consist of one, non-replicating chromosome and was used for calibration.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References

We thank Nanne Nanninga and LeRoy Bertsch for advice and critical reading of the manuscript, Tanneke den Blaauwen and Joop van Helvoort for suggestions and discussions, and E. Pas for experimental assistance. We thank two anonymous referees for suggesting the recA experiment.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  • 1
    Adam, M., Fraipont, C., Rhazi, N., Nguyen-Distèche, M., Lakaye, B., Frère, J.-M., et al (1997) The bimodular G57–V577 polypeptide chain of the class B penicillin-binding protein 3 of Escherichia coli catalyzes peptide bond formation from thiolesters and does not catalyze glycan chain polymerization from the lipid II intermediate. J Bacteriol 179: 60056009.
  • 2
    Addinall, S.G., Bi, E., Lutkenhaus, J. (1996) FtsZ ring formation in fts mutants. J Bacteriol 178: 38773884.
  • 3
    Ayala, J.A., Garrido, T., De Pedro, M.A., Vicente, M. (1994) Molecular biology of bacterial septation. In Bacterial Cell Wall. Ghuysen, J.M., and Hackenbeck, R. (eds). Amsterdam: Elsevier Science. pp. 73101.
  • 4
    Begg, K.J. & Donachie, W.D. (1991) Experiments on chromosome partitioning and positioning in Escherichia coli. New Biol 3: 475486.
  • 5
    Bi, E. & Lutkenhaus, J. (1991) FtsZ ring structure associated with division in Escherichia coli. Nature 354: 161164.
  • 6
    Bipatnath, M., Dennis, P.P., Bremer, H. (1998) Initiation and velocity of chromosome replication in Escherichia coli B/r and K-12. J Bacteriol 180: 265273.
  • 7
    Cook, W.R. & Rothfield, L.I. (1999) Nucleoid-independent identification of cell division sites in Escherichia coli. J Bacteriol 181: 19001905.
  • 8
    Den Blaauwen, T., Buddelmeijer, N., Aarsman, M.E.G., Hameete, C.M., Nanninga, N. (1999) Timing of the FtsZ assembly in Escherichia coli. J Bacteriol (in press).
  • 9
    Drlica, K. & Riley, M. (1990) The Bacterial Chromosome. Washington, DC: American Society for Microbiology Press.
  • 10
    Gordon, G.S., Sitnikov, D., Webb, Ch.D., Teleman, A., Straight, A., Losick, R., Murray, A.W., et al (1997) Chromosome and low copy plasmid segregation in E. coli : visual evidence for distinct mechanisms. Cell 90: 11131121.
  • 11
    Graumann, P.L., Losick, R., Strunnikov, A.V. (1998) Subcellular localization of Bacillus subtilis SMC, a protein involved in chromosome condensation and segregation. J Bacteriol 180: 57495755.
  • 12
    Helmstetter, C.E., Pierucci, O., Weinberger, M., Holmes, M., Tang, M.S. (1979) Control of cell division in Escherichia coli. In The Bacteria, Vol. VII. Ornston, L.N., and Sokatch, J.R. (eds). New York: Academic Press, pp. 517579.
  • 13
    Hiraga, S. (1992) Chromosome and plasmid partition in Escherichia coli. Annu Rev Biochem 61: 283306.
  • 14
    Jaffé, A., D'Ari, R., Norris, V. (1986) SOS-independent coupling between DNA replication and cell division in Escherichia coli. J Bacteriol 165: 6671.
  • 15
    Lemon, K.P. & Grossman, A.D. (1998) Localization of bacterial DNA polymerase: evidence for a factory model of replication. Science 282: 15161519.
  • 16
    Lin, D.C.-H., Levin, P.A., Grossman, A.D. (1997) Bipolar localization of a chromosome partition protein to Bacillus subtilis. Proc Natl Acad Sci USA 94: 47214725.
  • 17
    Liu, G., Draper, G.C., Donachie, W.D. (1998) FtsK is a bifunctional protein involved in cell division and chromosomal localization in Escherichia coli. Mol Microbiol 29: 893903.
  • 18
    Lockhart, A. & Kendrick-Jones, J. (1998) Interaction of the N-terminal domain of MukB with the bacterial tubulin homologue FtsZ. FEBS Lett 430: 278282.
  • 19
    Ma, X., Ehrhardt, D.W., Margolin, W. (1996) Colocalization of cell division proteins FtsZ and FtsA to cyto-skeletal structures in living Escherichia coli cells by using green fluorescent protein. Proc Natl Acad Sci USA 93: 1299813003.
  • 20
    Mason, D.J. & Powelson, D.M. (1958) Nuclear division as observed in live bacteria by a new technique. J Bacteriol 71: 474479.
  • 21
    Mulder, E. & Woldringh, C.L. (1989) Actively replicating nucleoids influence positioning of division sites in Escherichia coli filaments forming cells lacking DNA. J Bacteriol 171: 43034314.
  • 22
    Nanninga, N. (1991) Cell division and peptidoglycan assembly in Escherichia coli. Mol Microbiol 5: 791795.
  • 23
    Nanninga, N. (1998) Morphogenesis of Escherichia coli. Microbiol Mol Biol Rev 62: 120.
  • 24
    Nanninga, N., Den Blaauwen, T., Voskuil, J., Wientjes, F.B. (1985) Stimulation and inhibition of cell division in synchronized Escherichia coli. Ann Inst Pasteur Microbiol 136A: 139145.
  • 25
    Newman, E.B., Budman, L.I., Chan, E.C., Greene, R.C., Lin, R.T., Woldringh, C.L., et al (1998) Lack of S-adenosylmethionine results in a cell division defect in Escherichia coli. J Bacteriol 180: 36143619.
  • 26
    Nguyen-Distèche, M., Fraipont, C., Buddelmeijer, N., Nanninga, N. (1998) The structure and function of Escherichia coli penicillin-binding protein 3. Cell Mol Life Sci 54: 309316.
  • 27
    Niki, H. & Hiraga, S. (1998) Polar localization of the replication origin and terminus in Escherichia coli nucleoids during chromosome partitioning. Genes Dev 12: 10361045.
  • 28
    Norris, V. (1995) Hypothesis: chromosome separation in Escherichia coli involves autocatalytic gene expression, translation and membrane-domain formation. Mol Microbiol 16: 10511057.
  • 29
    Odijk, T. (1998) Osmotic compaction of supercoiled DNA into a bacterial nucleoid. Biophys Chem 73: 2330.
  • 30
    Palacios, P., Vicente, M., Sanchez, M. (1996) Dependency of Escherichia coli cell-division size, and independency of nucleoid segregation on the mode and level of ftsZ expression. Mol Microbiol 20: 10931098.
  • 31
    Pogliano, J., Pogliano, K., Weiss, D.S., Losick, R., Beckwith, J. (1997) Inactivation of FtsI inhibits constriction of the FtsZ cytokinetic ring and delays the assembly of FtsZ rings at potential division sites. Proc Natl Acad Sci USA 94: 559564.
  • 32
    Sharpe, M.E. & Errington, J. (1996) The Bacillus subtilis soj-spo0J locus is required for a centromere-like function involved in prespore chromosome partitioning. Mol Microbiol 21: 501509.
  • 33
    Skarstad, K. & Boye, E. (1993) Degradation of individual chromosomes in recA mutants of Escherichia coli. J Bacteriol 175: 55055509.
  • 34
    Steiner, W.W. & Kuempel, P.L. (1998) Cell division is required for resolution of dimer chromosomes at the dif locus of Escherichia coli. Mol Microbiol 27: 257268.
  • 35
    Steiner, W.W., Liu, G., Donachie, W.D., Kuempel, P.L. (1999) The cytoplasmic domain of FtsK protein is required for resolution of chromosome dimers. Mol Microbiol 31: 579583.
  • 36
    Taschner, P.E.M., Verest, J.G.J., Woldringh, C.L. (1987) Genetic and morphological characterization of ftsB and nrdB mutants of Escherichia coli. J Bacteriol 169: 1925.
  • 37
    Taschner, P.E.M., Huls, P., Pas, E., Woldringh, C.L. (1988) Division behavior and shape changes in isogenic ftsZ, ftsQ, ftsA, pbpB, and ftsE cell division mutants of Escherichia coli during temperature shift experiments. J Bacteriol 170: 15331540.
  • 38
    Tétart, F., Albigot, R., Conter, A., Mulder, E., Bouché, J.P. (1992) Involvement of FtsZ in coupling of nucleoid separation with septation. Mol Microbiol 6: 621627.
  • 39
    Valkenburg, J.A.C. & Woldringh, C.L. (1984) Phase separation between nucleoid and cytoplasm in Escherichia coli as defined by immersive refractometry. J Bacteriol 160: 11511157.
  • 40
    Valkenburg, J.A.C., Woldringh, C.L., Brakenhoff, G.J., Van Der Voort, H.T.M., Nanninga, N. (1985) Confocal scanning light microscopy of the Escherichia coli nucleoid: comparison with phase-contrast and electron microscope images. J Bacteriol 161: 478483.
  • 41
    Van Helvoort, J.M.L.M. & Woldringh, C.L. (1994) Nucleoid partitioning in Escherichia coli during steady-state growth and upon recovery from chloramphenicol treatment. Mol Microbiol 13: 577583.
  • 42
    Van Helvoort, J.M.L.M., Kool, J., Woldringh, C.L. (1996) Chloramphenicol causes fusion of separated nucleoids in Escherichia coli K-12 cells and filaments. J Bacteriol 178: 42894293.
  • 43
    Van Helvoort, J.M.L.M., Huls, P.G., Vischer, N.O.E., Woldringh, C.L. (1998) Fused nucleoids resegregate faster than cell elongation in Escherichia coli pbpB (Ts) filaments after release from chloramphenicol inhibition. Microbiology 144: 13391317.
  • 44
    Vischer, N.O.E., Huls, P.G., Ghauharali, R.I., Brakenhoff, G.J., Nanninga, N., Woldringh, C.L. (1999) Image cytometric method for quantifying the relative amount of DNA in bacterial nucleoids using Escherichia coli. J Microsc 196 (in press).
  • 45
    Walter, H. & Brooks, D.E. (1995) Phase separation in cytoplasm, due to macromolecular crowding, is the basis for microcompartmentation. FEBS Lett 361: 135139.
  • 46
    Webb, C.D., Teleman, A., Gordon, S., Straight, A., Belmont, A., Lin, D.C.-H., et al (1997) Bipolar localization of the replication origin regions of chromosomes in vegetative and sporulating cells of B. subtilis. Cell 88: 667674.
  • 47
    Webb, C.D., Graumann, P.L., Kabana, J.A., Teleman, A.A., Silver, P.A., Losick, R. (1998) Use of time-lapse microscopy to visualize rapid movement of the replication origin region of the chromosome during the cell cycle in Bacillus subtilis. Mol Microbiol 28: 883893.
  • 48
    Weiss, D.S., Chen, J.C., Ghigo, J.-M., Boyd, D., Beckwith, J. (1999) Localization of FtsI (PBP3) to the septal ring requires its membrane anchor, the Z ring, FtsA, FtsQ, and FtsL. J Bacteriol 181: 508520.
  • 49
    Woldringh, C.L. (1976) Morphological analysis of nuclear separation and cell division during the life cycle of Escherichia coli. J Bacteriol 125: 248257.
  • 50
    Woldringh, C.L. & Nanninga, N. (1985) Structure of nucleoid and cytoplasm in the intact cell. In Molecular Cytology of Escherichia coli. Nanninga, N. (ed.). London: Academic Press, pp. 161197.
  • 51
    Woldringh, C.L., Huls, P., Nanninga, N., Pas, E., Taschner, P.E.M., Wientjes, F.B. (1988) Autoradiographic analysis of peptidoglycan synthesis in shape and cell division mutants of Escherichia coli LMC500. In Antibiotic Inhibition of Bacterial Cell-Surface Assembly and Function. Actor, P., Daneo-Moore, L., Higgins, M.L., Salton, M.R.J., and Shockman, G.D. (eds). Washington, DC: American Society for Microbiology Press, pp. 6678.
  • 52
    Woldringh, C.L., Zaritsky, A., Grover, N.B. (1994) Nucleoid partitioning and the division plane in Escherichia coli. J Bacteriol 176: 60306038.
  • 53
    Woldringh, C.L., Jensen, P.R., Westerhoff, H.V. (1995) Structure and partitioning of bacterial DNA: determined by a balance of compaction and expansion forces? FEMS Microbiol Lett 131: 235242.
  • 54
    Yu, X.-C., Weihe, E.K., Margolin, W. (1998) Role of the C terminus of FtsK in Escherichia coli chromosome segregation. J Bacteriol 180: 64246428.