Precise determinations of C and D periods by flow cytometry in Escherichia coli K-12 and B/r

Ole Michelsen1, M. Joost Teixeira de Mattos2, Peter Ruhdal Jensen1 and Flemming G. Hansen1

1 Section of Molecular Microbiology, BioCentrum – DTU, Technical University of Denmark – DTU, Building 301, DK-2800 Lyngby, Denmark
2 Department of Microbiology, E. C. Slater Institute, BioCentrum Amsterdam, University of Amsterdam, 1018 WS Amsterdam, The Netherlands

Correspondence
Ole Michelsen
Ole.Michelsen{at}BioCentrum.dtu.dk


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The C and D cell cycle periods of seven Escherichia coli K-12 strains and three E. coli B/r strains were determined by computer simulation of DNA histograms obtained by flow cytometry of batch cultures grown at several different generation times. To obtain longer generation times two of the K-12 strains were cultivated at several different dilution rates in glucose-limited chemostats. The replication period (C period) was found to be similar in K-12 and B/r strains grown at similar generation times. At generation times below 60 min the C period was constant; above 60 min it increased linearly with increasing generation time. The period from termination of replication to cell division (D period) was more variable. It was much shorter in B/r than in K-12 strains. Like the C period it was relatively constant at generation times below 60 min and it increased with increasing generation times at longer generation times. In glucose-limited chemostats good correlation was found between D periods and generation times, whereas batch cultures exhibited carbon-source-dependent variations. Chemostat cultures showed cell cycle variations very similar to those obtained in batch cultures. These flow cytometric determinations of cell cycle periods confirm earlier determinations of the C period and establish that the D period also varies with generation time in slowly growing cultures. In addition they extend the range of growth rates at which cell cycle periods have been determined in E. coli K-12.


Abbreviations: {tau}, generation time; CV, coefficient of variation


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The cell cycle of slowly growing bacteria consists of three time periods. The B period is the time from cell division to initiation of chromosome replication. The C period is the time it takes to replicate the chromosome, and the D period is the time from termination of replication to cell division. In a balanced culture of slowly growing bacteria (Fig. 1) all the cells in the B period should contain 1 genome equivalent of DNA, the cells in the C period should contain between 1 and 2 genome equivalents, and the cells in the D period should contain 2 genome equivalents. In faster-growing bacteria, where the generation time becomes shorter than the C+D period, initiation of replication is moved into the preceding cell cycle, and the B period disappears (Helmstetter, 1996).



View larger version (11K):
[in this window]
[in a new window]
 
Fig. 1. Age distribution of an ideal slowly growing bacterial culture (Powell, 1958). The times of initiation (i), termination of replication (t) and cell division (d) are shown above the age distribution. The cell cycle intervals B, C and D are shown; the corresponding areas represent the relative number of cells present in the respective cell cycle interval.

 
Several methods have been used to estimate the B, C and D periods in cultures of Escherichia coli. The membrane elution technique was used to obtain values for C and D periods in slowly growing strains of E. coli B/r (Cooper & Helmstetter, 1968; Helmstetter et al., 1968; Helmstetter & Cooper, 1968). The B period, which was not observed in these early studies, was introduced in a later study where the membrane elution technique was used to analyse different substrains of E. coli B/r (Helmstetter & Pierucci, 1976). Other frequently used methods only allow an incomplete determination of one or more of these cell cycle periods. The C period has been determined from the residual DNA synthesis obtained after inhibition of initiation of chromosome replication (Bipatnath et al., 1998; Churchward & Bremer, 1977); from marker frequency analysis determining the origin to terminus ratio (Atlung & Hansen, 1993; Yoshikawa & Sueoka, 1963); or from determinations of the fraction of replicating cells in slowly growing bacterial cultures (Kubitschek & Newman, 1978). The D period has been reported to be constant in E. coli B/r at different generation times, estimated either from residual division in the presence of chloramphenicol (Kubitschek, 1974), or from residual division during thymine starvation (Bremer & Chuang, 1981). The C+D period can be estimated from flow cytometric analysis of rifampicin-treated cells (Løbner-Olesen et al., 1989). However, flow cytometry has also been used for a direct determination of the B, C and D periods by analysis of DNA distributions of exponentially growing cultures either directly (Skarstad et al., 1983) or by computer simulations (Skarstad et al., 1985).

The majority of the research into the cell cycle periods of bacteria has focused on relatively fast-growing cultures, and only a few studies have been concerned with slowly growing cultures (Helmstetter, 1996). But bacteria such as E. coli do experience conditions of substrate limitation resulting in slow growth in parts of their natural environment, and it is therefore of interest to analyse to what extent the cell cycle periods change in response to slow growth on poor carbon sources or to substrate-limited growth in the chemostat.

The study by Skarstad et al. (1985) strongly suggests that computer simulation of DNA distributions obtained by flow cytometry might be a precise way to determine the B, C and D cell cycle periods in slowly growing cultures. However, except for a few determinations of the C and D periods in relatively fast-growing cultures of E. coli strains (Allman et al., 1991; Skarstad et al., 1985) computer simulation of DNA distributions obtained by flow cytometry has not been used for a systematic study of cell cycle periods in E. coli. We have developed new computer software for the accurate cell cycle analysis of DNA distributions obtained by flow cytometry, and here we describe the use of our software for a comprehensive analysis of cell cycle periods in slowly growing cultures of E. coli B/r and K-12.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strains and growth of bacterial cultures.
The Escherichia coli strains used in this study are listed in Table 1. Batch cultures were grown with the necessary requirements in A+B medium (Clark & Maaløe, 1967) enriched with 0·2–0·5 % of one of the following carbon sources: alanine, pyruvate, acetate, succinate, glycerol or glucose. To obtain faster growth Casamino acids or amino acid mixtures were added, or the cultures were grown in buffered LB medium (Miller, 1972) with 0·2 % glucose. Chemostat cultures were grown exactly as described by Calhoun et al. (1993) except that the glucose input was 2 g l-1 instead of 5 g l-1. The glucose concentrations were measured immediately after sampling and were found to be not detectable, which is compatible with the very low expected concentration (1–20 µM), as no special effort was made to prevent glucose uptake by the cells during sampling. The carbon balance was calculated from the CO2 production and the cell dry weight content of the chemostat and was found to account for the carbon input within the experimental error (mean 97 ± 6 %). The amount of cell dry weight per g glucose in the faster-growing chemostat cultures (0·45 g dry weight per g glucose) approached that found in glucose-grown batch cultures (Jensen & Michelsen, 1992).


View this table:
[in this window]
[in a new window]
 
Table 1. E. coli strains

 
Flow cytometric procedures.
Samples were prepared and flow cytometry was performed essentially as described by Løbner-Olesen et al. (1989) based on the procedures described by Skarstad et al. (1983, 1985) except that we used 36 µg cephalexin ml-1 to stop cell division. We used the Argus (Skatron) or the Bryte (Bio-Rad) flow cytometers.

Computer simulation of DNA distributions.
Our computer software to simulate experimental DNA distributions of slowly growing bacterial cultures is described in Results and Discussion. For faster-growing bacterial cultures where initiation of replication is moved into the preceding cell cycles the variation in doubling times for the individual bacteria in a culture should be taken into account (Skarstad et al., 1985). This was done as follows. The culture was divided into up to 41 subpopulations of cells which initiated chromosome replication at slightly different times in the cell cycle, and it was assumed that the numbers of cells in these subpopulations were normally distributed around the normal initiation time with a standard deviation which can be chosen as a percentage of the generation time. Additionally, and in contrast to Skarstad et al. (1985), who kept the C period constant and varied the D period according to two different models, it was assumed that the C and D periods were affected proportionally in cells which initiated replication too early or too late in the cell cycle and therefore had a longer or a shorter period (C+D) to cell division. Input parameters for the software are standard deviations for the lower and upper parts of the DNA distribution, or standard deviations obtained from rifampicin-treated samples, and the length of the C and D periods. These values are varied by manual iteration until a best fit of the calculated distribution to the DNA histogram was obtained (see Results and Discussion for further details). To evaluate our simulations we calculated the deviation s using the formula presented by Skarstad et al. (1985):

where yi is the observed number of cells, Ni the expected number of cells and m the number of channels.

The program can be used to determine accurate cell cycle periods as long as C+D<2{tau}; for faster-growing cultures the program can be used – and it was used in this study – but other methods are recommended. The computer program can be obtained from F. G. Hansen (FGH@BioCentrum.dtu.dk).


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Resolving a DNA histogram obtained by flow cytometry into cell cycle periods
The DNA distribution in an ideal culture can be predicted using the model by Cooper & Helmstetter (1968) and this approach was used by Skarstad et al. (1985) to simulate DNA distributions obtained by flow cytometry. In that work it was stated that: ‘If there is no variability from cell to cell in C, D, and {tau} and if initiation is synchronous at all origins within a cell, then the rate of DNA synthesis during the cell cycle can be described by a step function with two discontinuities, one at the time of initiation and the other one at termination.’ The simplest DNA histograms are obtained by flow cytometry of slowly growing cultures of cells, which are born with one chromosome. Fig. 2(a) shows that the DNA histogram of such a culture exhibits two peaks, representing cells with 1 and 2 genome equivalents of DNA. These cells represent cells in the B and D periods, respectively, as they are not engaged in DNA replication. A ridge representing the replicating cells (cells in the C period) connects the two peaks (cf. Fig. 1).



View larger version (21K):
[in this window]
[in a new window]
 
Fig. 2. The cell cycle in a slowly growing bacterial culture analysed by flow cytometry. (a) The DNA distribution (black circles) of a culture of strain RH147 grown in succinate medium with a doubling time of 88 min was computer simulated using the standard deviations obtained from the rifampicin-treated cells (panel b). The simulations of cells in the B, C and D periods, respectively, are shown with thin curves. The final simulation, which is the sum of the cells in the different cell cycle periods, is shown with a thick grey curve. The deviation s (see Methods) is indicated in this panel and in panels (c–e), which are described below. (b) A part of the same culture was incubated with rifampicin (to stop initiation of replication) and cephalexin (to stop cell division). The number of cells with one chromosome represents cells in the B period and the number of cells with two chromosomes represents cells in the C and D periods. Using the age distribution these cell numbers can be used to calculate the length of the B period and the C+D period, respectively. Alternatively, the C+D period can be determined from the origins per cell ratio, which equals 2(C+D)/{tau} (Helmstetter, 1996). (c–d) Simulations for which the coefficient of variation (CV) was kept constant. The CV values were based on the standard deviations at the arrow positions. (c) CV based on the standard deviation of cells in the 1 chromosome peak in (b). The simulation was carried out to obtain perfect fit to the left part of the distribution. (d) CV based on the standard deviation of cells in the 2 chromosome peak in (b). The simulation was carried out to obtain perfect fit to the right part of the distribution. (e) CV based on the mean of the standard deviations used above, i.e. the standard deviation of the mid distribution. This simulation was carried out to obtain the lowest deviation s between the simulated and the experimental curve.

 
We have developed new computer software similar to that described by Skarstad et al. (1985) to analyse DNA distributions obtained by flow cytometry. In experimental DNA distributions obtained using a flow cytometer, staining as well as instrumental variabilities should be taken into account in the simulation of theoretical DNA distributions. These variabilities can be determined by a flow cytometric analysis of the experimental DNA distribution of cultures incubated with rifampicin for long enough to terminate ongoing rounds of replication. Such distributions show peaks representing integral numbers of genome equivalents (Fig. 3a), which can be approximated by normal distributions, with characteristic standard deviations. These standard deviations were found to increase in a close to linear fashion with increasing amount of DNA (Fig. 3b). However, in terms of the coefficient of variation (CV) the results obtained with the rifampicin-treated cultures indicate that the CV decreases in cell populations with increasing amounts of DNA. Thus, in contrast to Skarstad et al. (1985), we have developed our computer software to simulate flow cytometric DNA distributions by increasing the standard deviation non-proportionally, but linearly with increasing amount of DNA in accordance with the result from the rifampicin-treated cultures. As pointed out by Skarstad et al. (1985) there might be a variance in individual generation times, as well as initiation and termination times, within a culture. However in slowly growing cultures where C+D<{tau}, and where the cells have either 1, between 1 and 2, or 2 genome equivalents, such variations would not affect the DNA distribution. For faster-growing cultures where C+D>={tau} these variations could affect the DNA distribution and the software can handle these situations.



View larger version (21K):
[in this window]
[in a new window]
 
Fig. 3. The standard deviation as a function of the amount of DNA (fluorescence) in a flow cytometric DNA histogram. (a) A culture grown in glucose+Casamino acids at 30 °C of strain FH1674 carrying the dnaA606 mutation was incubated with rifampicin to stop initiation of transcription and thereby initiation of chromosome replication. This strain was chosen because it exhibits the asynchrony initiation phenotype (Skarstad et al., 1988). (b) The standard deviations obtained from the cell populations with different numbers of genome equivalents in (a) are plotted as a function of the channel number (black circles). The open circles represent standard deviations obtained in a similar experiment, and the open squares the experiment shown in Fig. 2. Solid line: coefficient of variation decrease with increasing DNA. Dotted line: constant coefficient of variation.

 
We determined the standard deviation of cells containing 1 or 2 genome equivalents of DNA in the culture shown in Fig. 2(a) by analysing the individual distributions obtained after incubating a portion of the culture with rifampicin and cephalexin to stop initiation of chromosome replication and cell division, respectively (Fig. 2b). We used the computer software with the standard deviations determined above to resolve the DNA distribution (Fig. 2a). Fig. 2(a) also shows the simulation curves and the values obtained for the B, C and D periods as well as the deviation s (Skarstad et al., 1985). The C+D and B periods, respectively, were obtained from rifampicin+cephalexin-treated culture (Fig. 2b) (Løbner-Olesen et al., 1989), and the C+D period was found to be identical to the sum of the C and D periods estimated from the exponential culture. Several samples were taken from the culture shown in Fig. 2(a). The simulation results of all of these samples gave B, C and D values which were close to the values reported in Fig. 2(a) (less than 2 % variation). It should be noted that the DNA histogram in Fig. 2(a) (and Fig. 2b) contains fluorescence signals of particles with more than 2 genome equivalents of DNA. The size distribution of the sample showed that these particles were bigger than the cells containing between 1 and 2 genome equivalents of DNA (data not shown). Thus, they are either filaments or cells sticking together, consistent with the presence of a small fraction of particles containing 3 genome equivalents in the rifampicin-treated culture (Fig. 2b). Microscopic inspection of samples prepared for flow cytometry also showed the presence of filaments or more often cells sticking together (data not shown). These cells (and signals from electronic noise) were removed by a background subtraction feature (see below) and were not taken into consideration in this simulation or in other simulations discussed in this study.

Preferably, and as is done above, standard deviations should be determined for cells with different amounts of DNA obtained from rifampicin-treated cultures. These samples should be measured immediately before or after that of the exponential culture from which the DNA histogram should be simulated. However, this poses a problem for the reanalysis of older data where only the DNA distribution of the exponential culture is available, and for the data of all of the batch and chemostat cultures of strain PJ4004, which will be analysed later in this article, as it turned out that this strain continued initiation of chromosome replication in the presence of rifampicin.

The background subtraction feature of the software allows the user to approximate normal distributions to the left (1 chromosome) and right (2 chromosome) peaks, to subtract cells (signals) which are outside of the main DNA distribution, and to calculate the corresponding standard deviations. These ‘standard deviations' will be slightly higher (5–15 % for the 1 chromosome and 15–25 % for the 2 chromosome peaks) than the standard deviations obtained by analysing corresponding rifampicin- and cephalexin-treated samples containing cells with either 1 or 2 chromosomes simply because the left and right curvatures of an exponential DNA distribution are formed by the presence of non-replicating as well as replicating cells (cf. Fig. 2a). Thus, in cases where rifampicin samples are not available (and often in our daily use of the software also when rifampicin samples are available) the ‘standard deviations' obtained from the DNA distribution of the exponentially growing cells are used as a guide for making ‘qualified guesses' of the ‘true’ standard deviations, which are then used in the simulation. In the present example the standard deviation for the 1 chromosome peak in the rifampicin-treated sample (Fig. 2b) was 9 channels and the standard deviation obtained from the left curvature of the cells containing preferentially 1 chromosome was 9·5 channels. To demonstrate the necessity to use a non-proportional increase in standard deviations we also present simulations where a proportional increase in standard deviations (constant CV) was used (Fig. 2c–e). The CV values were based on the standard deviations determined from Fig. 2(b) of either the 1 (Fig. 2c) or the 2 (Fig. 2d) chromosome peaks or the mean of these standard deviations (Fig. 2e). It is clear that we obtain the best curve fit (the lowest s) for the simulation presented in Fig. 2(a) (variable CV); however, the calculated cell cycle periods are similar for the calculations based on a constant CV.

It is our experience that only when ‘correct’ standard deviations are used will it be possible to get a perfect fit of the calculated to the experimental DNA distribution.

Recalculation of cell cycle data
The B, C and D periods of slowly growing E. coli B/r A and K strains were determined by flow cytometry by Skarstad et al. (1983). These estimations were based on the number of cells present in the peaks representing cells with one and two chromosomes, respectively. However, replicating cells will also contribute to the number of cells in these two peaks, making it very difficult to put precise borders between cells which have not started replication, and cells which have just started. The same will be true for cells which are just about to finish and cells that have finished replication. Later these authors (Skarstad et al., 1985) developed a computer program to simulate DNA distribution data obtained by flow cytometry. This program can handle flow cytometric DNA distributions for both slowly growing and fast-growing bacteria and the B, C and D periods were determined for an E. coli B/r A strain growing with a 330 min doubling time. The B, C and D periods determined in these two articles (Skarstad et al., 1983, 1985), are presented in Table 2, together with the values obtained using our newly developed simulation routine to estimate the cell cycle periods from the same DNA distributions. The new simulations (Fig. 4a, b) of the data presented in the early article (Skarstad et al., 1983) gave cell cycle periods (Table 2) which were significantly different from the results presented earlier. It was clear that the B and D periods were overestimated and thus the C period became too short. However, we also found significant differences from the results presented in the article where Skarstad et al. (1985) took advantage of a computer program to simulate the DNA distribution. Our simulation suggests that the D period might be overestimated. Fig. 5(a) shows two simulations of the data obtained from an E. coli B/r culture growing with 330 min doubling time (Skarstad et al., 1985) using our program. The enlarged area (Fig. 5b) emphasizes the differences in the two simulations. For the light grey thick curve we varied the C and D periods as well as the standard deviations until we obtained the best fit to the experimental data (deviation s=0·35). For the simulation presented by the black curve we used the C and D periods estimated by Skarstad et al. (1985) and we varied the standard deviations with the restriction that the CV should be constant. This latter simulation obtained with our program was identical to the simulation presented by Skarstad et al. (1985), but the fit to the experimental data was not as good (s=1·01) as for the one where we used a variable CV. Thus, we suggest that the approach we have taken to simulate DNA histograms obtained by flow cytometry might give more precise data for the different cell cycle periods.


View this table:
[in this window]
[in a new window]
 
Table 2. The periods of the bacterial cell cycle determined by flow cytometry: a comparison between old and new determinations based on the same data

 


View larger version (19K):
[in this window]
[in a new window]
 
Fig. 4. Resolving the DNA histograms of slowly growing E. coli B/r strains. The experimental distributions, of glucose-limited chemostat cultures of (a) E. coli B/r K (17 h doubling time) and (b) E. coli B/r A (16 h doubling time), were digitized from the data presented by Skarstad et al. (1983), and resolved into B, C and D periods (thin curves). The thick grey curve represents the sum of the cells in the different cell cycle periods.

 


View larger version (17K):
[in this window]
[in a new window]
 
Fig. 5. Resolving the DNA histograms of a slowly growing E. coli B/r strain. (a) The experimental distributions of a glucose-limited chemostat culture of E. coli B/r A (5·5 h doubling time), were digitized from the data presented by Skarstad et al. (1985), and resolved into B, C and D periods. The thick light grey curve represents the sum of the cells in the different cell cycle periods simulating the DNA histogram by our program varying C and D periods until the best fit was obtained. The thin black curve presents a simulation using C and D values from Skarstad et al. (1985) but varying the standard deviation to give a constant CV. The boxed area in (a) is magnified in (b) to emphasize the difference between the two simulations.

 
The C and D periods increase with increasing generation time
We used flow cytometry to determine the cell cycle periods C and D in steady-state batch cultures of seven strains of E. coli K-12 and three of E. coli B/r (Table 3). The different generation times were obtained by using different carbon sources. In addition two of the K-12 strains were grown at different dilution rates in glucose-limited chemostats (Table 4). One sample was taken directly from the respective cultures for flow cytometric analysis of the DNA distribution of cells in the population. Another sample was incubated with rifampicin and cephalexin for long enough to terminate ongoing rounds of replication and to stop cell division before the samples were prepared for flow cytometry. The DNA histograms of the respective samples were obtained by flow cytometry and resolved into cell cycle periods as described above. Most DNA histograms were resolved more than once. Our results from Tables 3 and 4 are presented with filled symbols in Fig. 6. Most of the previous determinations of the C and D periods have been carried out using different E. coli B/r strains, but there are also a number of determinations from a variety of different K-12 strains. These data were compiled by Helmstetter (1996) and are presented in Fig. 6 for comparison (open symbols). We find a very good correlation between our flow cytometric determinations of C and the data compiled by Helmstetter (1996). In agreement with earlier findings the C period is virtually constant at generation times shorter than 60–70 min. At longer generation times we find that the C period increases linearly with increasing generation time and with similar slopes for the E. coli K-12 and B/r strains.


View this table:
[in this window]
[in a new window]
 
Table 3. Determination of cell cycle periods in batch cultures

DNA distributions obtained by flow cytometry were resolved with our computer software into the cell cycle periods C, D and B.

 

View this table:
[in this window]
[in a new window]
 
Table 4. Determination of cell cycle periods in chemostat cultures

DNA distributions obtained by flow cytometry were resolved with our computer software into the cell cycle periods C, D and B.

 


View larger version (37K):
[in this window]
[in a new window]
 
Fig. 6. The C and D cell cycle periods in E. coli. The cell cycle periods, C and D respectively, estimated by computer simulation of DNA distributions obtained by flow cytometry of different E. coli strains grown at different generation times (data from Tables 2, 3 and 4) are presented as a function of the generation time (filled symbols). Filled circles and triangles show the cell cycle periods of batch and chemostat cultures, respectively, of K-12. Filled squares and diamonds show the cell cycle periods of batch and chemostat cultures, respectively, of B/r strains. Open symbols represent the cell cycle periods compiled by Helmstetter (1996) from K-12 strains (open circles) and B/r strains (open squares and open diamonds). From the compiled data only those cell cycle periods are presented in which both the C period and the D period were determined. Furthermore we have omitted the data in which the generation time was given as an interval. The data of Skarstad et al. (1983, 1985) and our recalculated values (Table 2) are presented by open and filled diamonds, respectively.

 
The D period, on the other hand, shows much variation. Flow cytometry shows clearly that the D period increases with increasing generation time for both K-12 and B/r strains and it is also clear that the D period is significantly shorter in B/r than in K-12 at the various growth rates, but we also find carbon-source-dependent variations in the D periods of the different strains. For generation times above 100 min we find virtually no correlation between our data and the data compiled by Helmstetter (1996). Most of these data originated from a study in which the C period was determined directly but the D period was calculated based on the C period, the generation time and the DNA content of the culture (Kubitschek & Newman, 1978). We feel confident that our direct approach to measure the D period will give more precise estimates than those calculated by Kubitschek & Newman (1978).

Variation in the cell cycle periods
We observed minor and major adaptations or mutations of the cultures during prolonged growth in the chemostat. As can be seen from Table 4 the duration of the individual cell cycle periods changed during cultivation in the chemostat, and from Fig. 6 it can be seen that the values for the different periods scatter compared to the simple linear relationship. Part of this scatter is probably real and might be a reflection of adaptive mutations in the chemostats (Notley-McRobb & Ferenci, 2000). As the most extreme example, Fig. 7 shows two histograms of samples from chemostat 2 (generation times 105 and 103 min, Table 4). There are 140 generations between the two samples and during this time only a small change in dilution rate (growth rate) had occurred. However, the last sample showed a dramatic change in the shape of the DNA histograms. The calculations indicate that this is due to a significant decrease in the C period and a concomitant increase in the D period. We have not investigated this phenomenon in more detail. However, it is known that insertional inactivation of the ihfA or ihfB alleles (von Freiesleben et al., 2000) and of the hns gene (Atlung & Hansen, 2002) can decrease the C period. Although the DNA distributions of the two cultures look very different, the durations of the cell cycle periods do fall within the general scatter in Fig. 6.



View larger version (19K):
[in this window]
[in a new window]
 
Fig. 7. Variations in DNA distributions during prolonged chemostat growth. The experimentally obtained DNA distributions from strain PJ4004 (chemostat 2) growing with a 105 min (a) doubling time and after 140 generations with a 103 min (b) doubling time (black dots) are shown together with the computer simulations (light grey curve).

 
Concluding remarks
We have used flow cytometry and newly developed computer software to determine the C and D periods in E. coli K-12 and B/r strains. We found that they are constant below and increase linearly above 60–70 min generation time. With respect to the C period, several E. coli K-12 and B/r strains of different origin behaved identically in this analysis and our results are similar to those compiled by Helmstetter (1996). This indicates that the average rate at which the DNA polymerase adds nucleotides to the replicating chromosome is highly variable. This variation may reflect the fact that the abundance of deoxyribonucleotides and/or the energy status of the cells are similar under the conditions of slow growth on poor carbon sources or on limiting glucose concentrations in the chemostat cultures (Jensen & Pedersen, 1990).

We find that the flow cytometric approach is very accurate for determining the C period for slowly growing bacteria. Marker frequency analysis of the origin to terminus ratio by hybridization is another method, which gives very precise measurements of the C period (Atlung & Hansen, 1993). However, this method is especially good for determining the C period in fast-growing bacteria where the origin to terminus ratio is high. In slow-growing bacteria with a long B period the origin to terminus ratio is very low and the method becomes less accurate. Although we have used flow cytometry in this study to estimate cell cycle periods also from faster-growing cultures, we recommend the use of different methods for bacteria growing at different generation times. Flow cytometry can be used successfully if the C+D period is less than two doubling times, and marker frequency analysis if the C period is longer than the doubling time.

As mentioned before, many of the previous determinations of the D period indicate that D is constant (Helmstetter, 1996). Some of these determinations are based on the assumption that protein synthesis is not required for cell division after termination of chromosome replication (Kubitschek, 1974); others were calculated based on the fraction of replicating cells (Kubitschek & Newman, 1978). However, early studies using the membrane elution technique (Helmstetter et al., 1968) suggested that the D period increases with increasing generation time at generation times above 60 min. Also, a few flow cytometric determinations of D suggested that the D period increases with increasing generation time in E. coli B/r (Skarstad et al., 1983, 1985). In the light of these data and our measurements of the D periods in many different strains growing with different generation times, we conclude that the D period increases with increasing generation time.

However, there are clear differences between the lengths of the D period in the different strains at the long generation times, and the scatter of lengths of D periods is much greater than that for C periods, possibly caused by carbon-source-dependent variations of the D period. In general it is clear that the D period is much shorter in B/r strains than in K-12 strains. Compare for example the D period of 42 min in B/r A at 17 h doubling time with the D period of 90 min in B/r K at 16 h doubling time and with the 210 min D period we observed in the K-12 strain (PJ4004) at 13 h doubling time. We suggest that flow cytometric analysis is the best (and the easiest) way to get precise measurements of the D period.


   ACKNOWLEDGEMENTS
 
We thank Vibeke Tjell and Søs Koefoed for technical assistance and Tove Atlung for editorial advice. This work was supported by the Danish Natural Science Research Council.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Allman, R., Schjerven, T. & Boye, E. (1991). Cell cycle parameters of Escherichia coli K-12. J Bacteriol 173, 7970–7974.[Medline]

Atlung, T. & Hansen, F. G. (1993). Three distinct chromosome replication states are induced by increasing concentrations of DnaA protein in Escherichia coli. J Bacteriol 175, 6537–6545.[Abstract]

Atlung, T. & Hansen, F. G. (2002). Effect of different concentrations of H-NS protein on chromosome replication and the cell cycle in Escherichia coli. J Bacteriol 184, 1843–1850.[Abstract/Free Full Text]

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, 265–273.[Abstract/Free Full Text]

Braun, R. E., O'Day, K. & Wright, A. (1985). Autoregulation of the DNA replication gene dnaA in E. coli. Cell 40, 159–169.[Medline]

Bremer, H. & Chuang, L. (1981). Cell division after inhibition of chromosome replication in Escherichia coli. J Theor Biol 93, 909–926.[Medline]

Calhoun, M. W., Oden, K. L., Gennis, R. B., de Mattos, M. J. & Neijssel, O. M. (1993). Energetic efficiency of Escherichia coli: effects of mutations in components of the aerobic respiratory chain. J Bacteriol 175, 3020–3025.[Abstract]

Casadaban, M. J. & Cohen, S. (1980). Analysis of gene control signals by DNA fusion and cloning in Escherichia coli. J Mol Biol 138, 179–207.[Medline]

Churchward, G. & Bremer, H. (1977). Determination of deoxyribonucleic acid replication time in exponentially growing Escherichia coli B/r. J Bacteriol 130, 1206–1213.[Medline]

Clark, D. J. & Maaløe, O. (1967). DNA replication and the division cycle in Escherichia coli. J Mol Biol 23, 99–112.

Cooper, S. & Helmstetter, C. E. (1968). Chromosome replication and the division cycle of Escherichia coli B/r. J Mol Biol 31, 519–540.[Medline]

Guyer, M. S., Reed, R. R., Steitz, J. A. & Low, K. B. (1981). Identification of a sex-factor-affinity site in E. coli as gamma delta. Cold Spring Harb Symp Quant Biol 45, 135–140.[Medline]

Hansen, F. G., Koefoed, S. & Atlung, T. (1992). Cloning and nucleotide sequence determination of twelve mutant dnaA genes of Escherichia coli. Mol Gen Genet 234, 14–21.[Medline]

Helmstetter, C. E. (1996). Timing of synthetic activities in the cell cycle. In Escherichia coli and Salmonella Cellular and Molecular Biology, pp. 1627–1639. Edited by F. C. Neidhardt and others. Washington, DC: ASM Press.

Helmstetter, C. E. & Cooper, S. (1968). DNA synthesis during the division cycle of rapidly growing Escherichia coli B/r. J Mol Biol 31, 507–518.[Medline]

Helmstetter, C. E. & Pierucci, O. (1976). DNA synthesis during the division cycle of three substrains of Escherichia coli B/r. J Mol Biol 102, 477–486.[Medline]

Helmstetter, C. E., Cooper, S., Pierucci, D. & Revelas, E. (1968). The bacterial life cycle. Cold Spring Harbor Symp Quant Biol 33, 809–822.[Medline]

Jensen, K. F. & Pedersen, S. (1990). Metabolic growth rate control in Escherichia coli may be a consequence of subsaturation of the macromolecular biosynthetic apparatus with substrates and catalytic components. Microbiol Rev 54, 89–100.[Medline]

Jensen, P. R. & Michelsen, O. (1992). Carbon and energy metabolism of atp mutants of Escherichia coli. J Bacteriol 174, 7635–7641.[Abstract]

Jensen, P. R., Westerhoff, H. V. & Michelsen, O. (1993). Excess capacity of H(+)-ATPase and inverse respiratory control in Escherichia coli. EMBO J 12, 1277–1282.[Abstract]

Kubitschek, H. E. (1974). Estimation of the D period from residual division after exposure of exponential phase bacteria to chloramphenicol. Mol Gen Genet 135, 123–130.[Medline]

Kubitschek, H. E. & Freedman, M. L. (1971). Chromosome replication and the division cycle of Escherichia coli B-r. J Bacteriol 107, 95–99.[Medline]

Kubitschek, H. E. & Newman, C. N. (1978). Chromosome replication during the division cycle in slowly growing, steady-state cultures of three Escherichia coli B/r strains. J Bacteriol 136, 179–190.[Medline]

Løbner-Olesen, A., Skarstad, K., Hansen, F. G., von Meyenburg, K. & Boye, E. (1989). The DnaA protein determines the initiation mass of Escherichia coli K-12. Cell 57, 881–889.[Medline]

Miller, J. H. (1972). Experiments in Molecular Genetics. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Notley-McRobb, L. & Ferenci, T. (2000). Experimental analysis of molecular events during mutational periodic selections in bacterial evolution. Genetics 156, 1493–1501.[Abstract/Free Full Text]

Powell, E. O. (1958). An outline of the pattern of bacterial generation times. J Gen Microbiol 18, 382–417.[Medline]

Rasmussen, L. J., Møller, P. L. & Atlung, T. (1991). Carbon metabolism regulates expression of the pfl (pyruvate formate-lyase) gene in Escherichia coli. J Bacteriol 173, 6390–6397.[Medline]

Skarstad, K., Steen, H. B. & Boye, E. (1983). Cell cycle parameters of slowly growing E. coli B/r studied by flow cytometry. J Bacteriol 154, 656–662.[Medline]

Skarstad, K., Steen, H. B. & Boye, E. (1985). Escherichia coli DNA distributions measured by flow cytometry and compared with theoretical computer simulations. J Bacteriol 163, 661–668.[Medline]

Skarstad, K., von Meyenburg, K., Hansen, F. G. & Boye, E. (1988). Coordination of chromosome replication initiation in Escherichia coli: effects of different dnaA alleles. J Bacteriol 170, 852–858.[Medline]

von Freiesleben, U., Rasmussen, K. V., Atlung, T. & Hansen, F. G. (2000). Rifampicin resistant initiation of chromosome replication from oriC in ihf mutants. Mol Microbiol 37, 1087–1093.[CrossRef][Medline]

Yoshikawa, H. & Sueoka, N. (1963). Sequential replication of the Bacillus subtilis chromosome. I. Comparison of marker frequencies in exponential and stationary growth phases. Proc Natl Acad Sci U S A 49, 559–566.[Medline]

Zeuthen, J. & Pato, M. L. (1971). Replication of the F'lac sex factor in the cell cycle of Escherichia coli. Mol Gen Genet 111, 242–255.[Medline]

Received 17 October 2002; revised 9 December 2002; accepted 24 December 2002.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Michelsen, O.
Articles by Hansen, F. G.
Articles citing this Article
PubMed
PubMed Citation
Articles by Michelsen, O.
Articles by Hansen, F. G.
Agricola
Articles by Michelsen, O.
Articles by Hansen, F. G.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2003 Society for General Microbiology.