Polyploidy of Bacillus subtilis
We wished to determine the ploidy status of cells growing at a defined rate, and counted a) number of origins, b) of termini (stained by FROS), and c) DNA staining (DAPI intensity) and nucleoid morphology. We used all four parameters because counting of FROS signals leads to an underestimation of copy numbers, since duplicated regions that are separated less than 250 nm will appear as one signal. All strains are described in [6] and were characterized in this prior work.
The numbers of origins and termini (see Fig. 1 for examples of imaging) were counted under conditions of slow growth in S750 minimal medium (no amino acids) at 25 °C, with a doubling time of 90 minutes, or during fast growth at 37 °C in LB rich medium (doubling times of 26 min, similar to cells lacking a FROS system) at an OD600 ranging between 0.5 and 3, and the results are summarized in Table 1. The termini counts show that Bacillus subtilis is mostly diploid irrespective of the growth conditions, but that it is mero-polyploid in rich medium, where it can have multiple (up to 8) origins per cell (Table 1, first panel). The numbers of origins and termini were also counted every half an hour in in LB, starting from an OD600 of 0.1. Results are summarized in Table 1, second panel. The chromosome exists for the entire cell cycle as a distinct, compact structure termed the ‘nucleoid’. Near the end of DNA replication, the nucleoid adopts a bi-lobed, dumb-bell shape that ultimately splits into two independent structures prior to cytokinesis [7]. In the terminus-tag strain, chromosomes forming a dumb-bell shape were counted as two nucleoids when they corresponded with two foci for the terminus tag. However, when only one focus was detected in the GFP (terminus tag) channel, a dumb-bell was considered as two nuclei when each lobe measured > 0.55 µm, or when the lobes were separated by at least 3 dark pixels. The offset of 0.55 was chosen through measuring several small separated lobes and averaging them. For the origin tag strain, the number of origins was not used to assess polyploidy since multi-origination does not correspond to the completion of all replication cycles before cell division. In this strain, only the number of nucleoids was used for the classification of cells. Nucleoids with a dumb-bell shape were considered as one chromosome when the lobes were smaller than 0.7 µm and not separated by a minimum of 3 dark pixels. Cells with three or more distinct nucleoids were considered polyploid (Fig. 1), see also Fig. 1A in [6]. Elongated nucleoids occupying > 2.5 µm of the cell length were counted as 2 distinct chromosomes. The results are summarized in Table 1, second and third panels.
Table 1. Origins and termini count under slow and fast growth
|
Slow growth (25°C)
|
Fast growth (37°C)
|
|
Foci/cell
|
nt
|
Foci/cell
|
nt
|
PG26 (origin tag)
|
1.8±0.51
|
620
|
4.42±1.49
|
848
|
AT62 (terminus tag)
|
1.49±0.6
|
700
|
1.89±0.66
|
848
|
Percent polyploidy and average number of origins over time
25°C 37°C
PG26
|
Control
|
1h
|
1.5h
|
2h
|
2.5h
|
3h
|
OD
|
0.7
|
0.11
|
0.3
|
0.7
|
1.8
|
2.5
|
% monoploid1
|
37.1
|
11
|
9
|
11
|
10
|
13
|
% diploid2
|
59.2
|
70
|
71
|
70
|
68
|
68
|
% polyploid3
|
3.7
|
19
|
20
|
19
|
22
|
19
|
Origins per cell
|
1.8±0.51
|
4.8±1.6
|
4.9±2
|
4.9±1.1
|
4.7±0.55
|
4.1±0.89
|
nt
|
620
|
300
|
300
|
350
|
400
|
400
|
|
|
|
|
|
|
|
|
Percent ploididity and average number of termini over time
25°C 37°C
AT62
|
Control
|
1h
|
1.5h
|
2h
|
2.5h
|
3h
|
OD
|
0.7
|
0.13
|
0.4
|
1
|
1.9
|
3
|
% monoploid4
|
35
|
17
|
13
|
12
|
15
|
12
|
% diploid5
|
61
|
74
|
72
|
71
|
69
|
75
|
% polyploid6
|
4
|
9
|
15
|
17
|
16
|
13
|
Termini per cell
|
1.49±0.6
|
1.71±0.5
|
1.93±0.6
|
1.8±1.1
|
1.72±0.73
|
1.95±0.8
|
nt
|
700
|
200
|
300
|
300
|
400
|
400
|
|
|
|
|
|
|
|
|
Percent ploididity from automatically calculated average DAPI fluorescence
25°C 37°C
AT62
|
Control
|
1.5h
|
2h
|
2.5h
|
OD
|
0.7
|
0.4
|
1
|
1.9
|
% monoploid
|
27.8
|
11.7
|
15
|
11.5
|
% diploid
|
67.5
|
68.3
|
63.3
|
62.25
|
% polyploid
|
4.6
|
20
|
21.7
|
26.25
|
nt
|
540
|
300
|
300
|
400
|
% polyploidy calculated over all time points at 37°C = 23.65%.
*nt refers to the total number of cells counted in each experiment
1 - 1 or 2 origins
2 – 2 to 4 origins
3 - 4 to 8 origins
*nt refers to the total number of cells counted in each time interval.
4 - 1 terminus
5 - 2 termini
6 - > 2 termini
Interestingly, the number of termini did not increase strongly at fast growth compared to slow growth, while the number of origins increased with growth rate, consistent with the observation that B. subtilis is diploid most of the time. However, the number of termini could be underestimated because in many cases it was hard to tell if the focus between 2 nuclei corresponded to one or two termini. Based on the counted number of nuclei according to the above criteria, there was always a subset of polyploid cells under fast growth, but this subset never exceeded 22% for the origin tag strain (PG26), and 18% for the terminus tag (AT62) strain. The total polyploidy calculated over all time points under fast growth conditions for both strains was 17.2% and 15.06%, respectively. Fig. S1 (https://doi.org/10.6084/m9.figshare.12792407.v1) shows examples of monoploid, diploid, and polyploid cells. Most strikingly, under slow and even rapid growth conditions, fractions of monoploid, diploid and polyploidy cells were present, revealing a much higher heterogeneity of chromosome contents than previously anticipated.
Estimation of the number of nuclei based on DAPI fluorescence
Though the manual counting provided a good idea on the polyploidy state of B. subtilis, it was prone to subjectivity and hence needed to be corroborated with an automated count. For this purpose, the ChainTracer software (Norbert Vischer, University of Amsterdam) was used together with the ImageJ plugin ObjectJ [8] to automatically calculate the amount of DAPI fluorescence in the cell.
The software can detect cell chains in a hyperstack, then extract single cells from chains by detecting the septa and cell borders from the membrane stain images assigned to the blue channel. Finally, the fluorescence assigned to the green channel of the hyperstack is calculated, in our case the DAPI. The terminus tag strain was chosen to perform the count under slow and fast growth. A threshold was specified for monoploid cells in each condition by averaging through fluorescence intensity of at least 30 cells measuring between 1.1 µm and 2 µm (i.e. the smallest cells) in each condition.
The average intensity of a nucleoid in small cells was comparable in all counted samples. Monoploid cells were those having 1-1.69 genomes calculated from the threshold, diploid cells between 1.7 and 2.69, and polyploid cells contained ≥ 2.69 genomes. Results are shown in Table 1, fourth panel. The percent polyploidy was underestimated by the manual count when compared to the automated based count. However, from both cases, it can be clearly deduced that B. subtilis is preferably diploid under slow and fast growth conditions, and that the number of polyploid cells is very small under slow growth and increases with the growth rate under fast growth. Cell length in minimal medium ranged between 1.894 µm and 4.112 µm, with an average of 3.05 ± 0.47 µm (n = 252). At faster growth both strains showed a direct correlation between the chromosomal content and the cell length whereby the polyploid cells were considerably longer than the diploid and the monoploid cells.
Thus, a B. subtilis culture contains cells with strongly diverging numbers of chromosomes and origin regions. This has to be taken into account when calculating cell cycle parameters from doubling times and replication periods. Interestingly, it has recently been shown that small RNAs induce diversity in transcription repressor AbrB levels generating heterogeneity in growth rates during the exponential growth phase. This leads to subpopulations of fast- and slow-growing B. subtilis cells, which has been proposed to increase fitness of the population when it has to deal with changes between favourable and unfavourable conditions [9].
Effect of DNA double strand break on the polyploidy state of B. subtilis
In order to assess the effect of DNA double strand breaks on the chromosome count as well as the number of origins in B. subtilis, Mitomycin C (MMC) was added to a final concentration of 50 ng/ml to cells growing in minimal medium at 25 °C. In B. subtilis, In addition to the recA-dependent transcriptional response, replication arrest also induces a recA-independent response, which is mediated in part by DnaA [10]. MMC was shown to cause a relative increase of the dosage of the origin of replication proximal genes. It was concluded that this increase in gene dosage was most likely caused by a reduced rate of elongation of replication, rather than by over-initiation [10].
The addition of MMC resulted in a strong reduction of growth for three hours. Cells imaged 30 minutes after MMC addition showed nucleoids with a condensed morphology (Fig. 2), and the vast majority of cells (~ 70% of 250 cells counted) contained one nucleoid instead of two, and 2.5 origins on average. The rest of the cells had two nucleoids (Fig. 2). 90 minutes after MMC addition, around 15% of the cells were dead in each imaging field, and these were excluded from the count. The number of origins increased from around 2 per cell prior to the induction of damage to 3.5 ± 1.4 (n = 200 cells). The cells were elongated, average cell length increased from 2.87 ± 0.52 µm during exponential growth (n = 150 cells) to 4.4 ± 1.4 µm after MMC addition. The cell length fell in a range between 1.6 and 7.7 µm, with 20% of the cells being above 5 µm. 16% of the 200 cells counted had 5 origins or more, with a few having up to 8 origins. A subset of the cells was anucleate (around 3%), and cells longer than 5 µm contained an elongated unsegregated nucleoid. 0.5% of the cells had nucleoids bisected by a septum. The terminus tag strain showed a comparable phenotype, in that the cells were elongated and the nucleoids were decondensed 90 minutes after MMC addition. Excluding the dead cells, the average number of termini was 1.7 ± 0.6 (n = 224 cells), with 7.14% of the cells having 3 termini or even 4 in very rare cases.
These findings show that despite frequent replication arrests due to interstrand crosslinking by MMC and the atypical nucleoid morphology, many cells were either still able to complete a replication cycle and had a double set of chromosomes after induction of DNA damage, or that the cells simply retained the number of nucleoids and termini they had before damage induction. The latter is in accordance with the idea that MMC slows down the cell cycle considerably but does not stop it completely [6]. Goranov et al. showed that double strand breaks induced by MMC inhibit the replication elongation beyond the origin proximal region [10]. This is in agreement with our experiments show an increase in number of origins but not of termini.