Molecular Microbiology (1992) 6(15), 2073-2083
Cell division in Escherichia coli minB mutants Thomas Akerlund, Rolf Bernander and Kurt Nordstrom* Department of Mierobiology, Biomedical Centre, Uppsala University, Box 581. S-751 23. Uppsala, Sweden. Summary In Escherichia coli minB mutants, cell division can take place at the cell poles as well as non-polarly in the cell. We have examined growth, division patterns, and nucleoid distribution in individual cells of a minC point mutant and a minB deletion mutant, and compared them to the corresponding wild-type strain and an intRI strain in which the chromosome is overreplicated. The main findings were as follows. In the minB mutants, polar and non-polar divisions appeared to occur independently of each other. Furthermore, the timing of cell division in the cell cycle was found to be severely affected. In addition, nucleoid conformation and distribution were considerably disturbed. The results obtained call for a reevaluation of the role of the MinB system in the E coli cell cycle, and of the concept that limiting quanta of cell division factors are regularly produced during the cell cycle.
Introduction Cell division in exponentially growing Escherieiiia coli strains normally takes place in the centre of the cells, betv^'een two newly replicated and segregated nucleoids. A mutant E, co//strain has been isolated which, in addition to non-polar divisions, also carries out polar divisions that result in the formation of DNA-less minicells (Adier et ai. 1967). Furthermore, in addition to normal-si2ed cells. elongated and filamentous cells are frequently produced by this mutant. The locus was denoted minB, and is located at 26 min on the genetic map of the E. coli chromosome (Bachmann, 1990). A model was proposed for how the minB operon may act during the E. co/;cell cycle {Teather etai, 1974). The key concept in this model is that the MinB system functions to direct cell division to the centre of the cell, by inactivating old' division sites at the cell poles. For each unitcell doubling, enough division factors {a 'quantum') are Received 29 January. 1992; revised and accepted 30 March, 1992. *For correspondence. Tel. (18) 174526: Fax (18)530396.
produced to carry out a single division. The division factors can normally only be used in the centre of the cell. whereas in minB mutants the old division sites at the cell poles are equally available for division and thus a 'choice' between which site to use is possible. The factors are entirely consumed in the division process, and a polar division is, therefore, carried out at the expense ot a cen* tral division. Hence, a cell that is two units in size should normally divide into two cells, each of which would be one unit in size, if, instead, a polar division occurred, the cell would have to grow in size from two to four units, i.e. double its cell mass, before new divisions could occur (cf,. Donachie. 1991). Since the cell size in this case has increased by two units, two quanta of division factors are available, and two new divisions may be carried out, again either polarly or non-poiarly. Consequently, both minicells and elongated cells are expected to be present in cultures of minB mutants. In the mode!, chromosome replication as well as nucleoid structure, segregation and partition are not expected to be affected. Importantly, the timing of chromosome replication and cell division in the cell cycle are also assumed to function normally: only the localization of the divisions is affected. Although this model has gained widespread acceptance, biochemical data in support of the various proposed functions are lacking; e,g. specific association between any of the Min proteins (below) and the cell poles has not been demonstrated (cf. de Boer et ai, 1991). The minB operon was cloned and sequenced and shown to code for three proteins: MinC, MinD, and MinE (de Boer etai. 1989). The MinC and MinD proteins were suggested to perform inhibitory functions, i.e. to inactivate the old division sites discussed above. The MinE protein was proposed to be the specificity factor that prevents MinCD from acting at non-polar division sites. During studies of the E. eoli celt cycle using the socalled intRI strains (Nordstrom etai. 1991) developed in our laboratory, we noticed that during over-replication of the chromosome also the intm strains readily produced minicells (Bernander ef ai, 1989). We decided to further examine minicell production in the intHI strains, using the well-characterized minB mutants as reference strains. During these studies we noted several discrepancies between our experimental findings and the proposed model for the function of the minB operon (above). Here, we report our investigations of the division pattern and nucleoid distribution in individual cells carrying minB mutations, as well as in intRI strains. We discuss the
2074
T. Akerlund, R. Bernander and K. Nordstrom
proposed role of the MinB system as a positional specificity system in E. coli cell division, and the concept of a quantal behaviour of ceil division factors during the cell cycle. Results Experimental outline Earlier reports have indicated that the cell size distribution and the division pattern vary in different minB mutants, and also depend upon the genetic background and on the growth conditions used (Frazer and Curtiss, 1975; Jaffe et al.. 1988; Woidringh et al., 1991). Consequently, the experiments were performed under similar growth conditions and in the same genetic background. To be able to compare our results with earlier reports, we used the same strain, P678-54. that was used in the experiments that form the basis for the model of the role of the MinB system (Teather etal.. 1974)- Therefore this strain, which carries the originally isolated mInC point mutation, was chosen, although additional mutations may be present because of the mutagenization that gave rise to the strain. To investigate possible differences between the point mutant and a mutant in which the entire minB locus has
Table 1. Growth rate and division capacity of the wild-type strain and the minB mutants. Division Capacity"^
Growth Rate (Doublings Per Hour) LB + 0.27., Glucose
Strain
batch culture^
2.5 Wild type mmC point 2.1 mutant minB deletion 2.3
M9CA
on top of agar''
within batch softagar^ culture^
22
1.9
2.0 2.1
1.6 1.9
Cell Sizes short cells
long cells
1.3
0.72
_
0.9 1.1
0 73 073
0.55 0.88
a. The growth rates in liquid media were determined by measuring cell density in a Klett-Summerson colorimeter. Each value is a mean of at least three individually performed experiments. M9CA is M9 medium (Maniatis er a/., 1982) supplemented with 1% casamino acids and 0.2% glycerol. b. The growth rates of individual clones were determined by adding together the lengths of all the cells in a microcolony at each time point, and plotting the sum of the cell lengths versus time. By measuring the slopes of the curves obtained, the doubling times in the microcolonies could be estimated. The values are mean values of all clones followed for each method. c. The division capacity was defined as the number of divisions per length unit and cell-length doubling. One length unit was defined as the average size of a newborn wtld-type cell (3.4 nm). Randomly chosen cells from pedigrees exemplified in Fig. 2 were foHowed for exactly one doubling in length, and the sum of the lengths of the mother cells was divided with the total number of divisions that took place during the doubling period. These values were in turn divided with one length unit. 3.4 |am. For the minB mutants, short and long cells were defined as cells shorter or ionger than 6.4 |am, respectively. The average cell size fof the short and long cells was about 5 and 9 |im, respectively, for both mutants.
Fig. 1. Time-lEyase microphotography of E. coliceWs growing on top ot agaf media containing Luria-Bertani medium and 0,2% glucose. The columns show one microcolony o( the wild type (A), two microcolonies of the m;nSdeletion mutant (B and C). and one microcolony of the minC point mutant (D).
been deleted (de Boer et al.. 1989), the deletion was introduced into strain P678, the parent of P678-54, by PI transduction. For direct microscopic observation of the growth and division patterns of the various strains, the cells were immobilized on solid growth media. Two microculture techniques were used; the cells were either grown on flat agar surfaces, or within 0.6% soft agar. The growth rates in the microcolonies were determined by measuring the lengths of all the cells in a microoolony at each time point, and plotting the sum of the cell lengths versus time. The doubling times in each of the microcolonies was estimated by measuring the slopes of the curves obtained. The average doubling times were generally longer on solid media than in liquid media, and the minB mutants usually displayed decreased growth rates (Table 1) as well as a larger variation of growth rates in individual microcolonies (not shown) relative to the wild type. The cells growing on solid media were microscopically inspected every five minutes, and micrographs were taken at every second observation event. A total of 38 microcolonies of the wild type, 74 of the minC point mutant, and 58 of the minB deletion mutant were studied for 50-70 min. fvlicrographs that illustrate different representative growth patterns are shown in Fig. 1, and stylized pedigrees constructed from such micrographs are presented in Fig. 2.
Cell division in Escherichia coli minB mutants
0
AminB
wt
'ml'l
A
"
F
10 10 30
JL J . . . . .
so
0 10 10
-
-
B
•
*
-
p
ZL. w
•
G -
7
- — X -)-*.
• •
-
Q
I
—
-*-
X X . . . .
40 SO
no
"
K
-
JL __
w M 70
minC point mutation
Jm _ _
— . JL JL
70 0 10 10
C
40 .
50 M 70
0
H -
T "
_
i
•
I 1 .
.
_
_
M
JUL. . ^ ^ .
a _
X - _
*
• ^^ ^
D 7
I
iO 40 60 70
0
. . J. 1 . . . . » ^ ^ ^ X X X _ E
10 10 40 JO M) 70
R
-
.
•
—
*
-
z z *_
^— t
t
.
N
.
—
•'•
S ^
^
t
t*. 1 ..
.
JLJ
Jm. J L .
o 2
;
•
_
J. .
.
•
Fig. 2. Pedigrees ol E co/i microcolonies derived from direct observaiion of growing cells as exemplided in Fig 1 Microcolonies A-O were grown on top of 1 2% agar surfaces whereas colonies P-T were grown within 0.6 % soft agar. Pedigrees A-E show five wiW-type microcolonies. F-J five colonies ol the m;nBdeletion mulanl. and K-T 10 colonies of the mmC point mutant. Minicells are not drawn to scale
^
. J ^ -*- ^
10 10
-
t.
•
-
3.
^ ^ .
•
2075
T t
• ^ t
X
^ ^ ^
^
_
^ ^_ .
JL JL
. J;
_t JJ^_ , ,
Celt sizes In minB mutants, cell division can take place at the cell poles as well as non-polarly. It has been proposed that cell division in minB mutants results in daughter cells (minicells excluded) whose lengths are multiples of the average length of a newborn wild-type cell (Teather etal.. 1974). We wanted to determine whether this was the case, and whether both polar and non-polar divisions resulted in daughter cells of distinct ceil lengths. First, we measured the cell lengths of more than 120 newborn cells of each strain after non-polar divisions. The newborn wildtype cells formed a uniform size class with a mean cell length of 3.4 ±0.2 nm (Fig. 3A). For the minB mutants, the cell sizes fell into two major classes, with a mean length of 4.7±0.3 |jm tor the class containing the smallest cells (Figs 3B and 3D). Thus the average cell size was larger than that of the wild type, despite the loss of cell material due to minicell production, and we could not detect any cells in the minB mutants that were smaller than the smallest cells of the wild type. Furthermore, non-polar divisions could take place centrally in the cell (Fig. 1C at 20 min) as well as at asymmetric positions (Fig. 1C at 50
and 70 min) in the minB mutants. The divisions did not appear to be randomly positioned in the cells, since both types of divisions gave rise to daughter cells of distinct size classes (Figs. 3B and 3D). In contrast to the daughter cells produced after nonpolar divisions, daughter cells (minicells excluded) resulting from polar divisions could not be grouped into distinct size classes (Figs. 3C and 3E). Furthermore, the minicells produced by the polar divisions were found to vary greatly in size, from a diameter close to that of the rod-shaped mother cells down to one-fifth or less of this diameter (cf. Fig. 1C at 70 min). The heterogeneity was confirmed by flow cytometry, in which the minicells displayed a considerable variation in size (not shown). In conclusion, an average newborn cell of the smallest size class from the minB mutants was longer than that of the corresponding wild-type strain. Non-polar divisions were found at both symmetric (central) and asymmetric positions along the length axis of the cells. The non-polar divisions did not take place at positions that were multiples of the average length of a newborn wild-type cell, although they appeared to be non-randomly positioned
2076
T. Akerlund. R. Bemander and K. Nordstrom
wt
R g . 3. Sizes of newborr) cells. The cells were grown on top of agar medium, and the sizes of newborn cells were measured directly under (he microscope. The panels show newborn wild-type cells (A), newborn ceils of the minCpoint mutant and the mrnSdeletion mutant after non-polar divisjons (B and D, respectively), and after polar divisions only (C and E, respectively). The open areas in panels B and D represent newborn cells that resulted from asymmetric divisions in the mother cells. The numbers of cells analysed were 120, 181, 150, 160, and 130 in panels A, B, C, D. and E. respectively.
non-polar divisions
minC point mutation
minC point mutation
non-polar divisions
polar divisions
Q lEl non-poUidivWoni Q uymmitric dlvidoiu
10
15
20
AmmB
AntinB
poUr divisions
non-polar divisions 0,3 n
•
ill non-poUi diviiioiu
E 0,2
0,1
0,0 5
10
15
20
0
size (|xm)
5
1•
along the length axis. Daughter cells produced after polar divisions could not be grouped into distinct size classes. The minicells displayed a broad cell-size distribution, indicating that the polar divisions that resulted in minicell formation were not initiated from a fixed position at the cell pole. Minicell divisions and filamentation According to the model described in the Introduction, a polar division is formed at the expense of a non-polar division. Therefore, the next division(s} cannot take place until the cell has doubled its mass, leading to filamentation. We wanted to investigate whether we could confirm such a relationship between the production of minicells
10
15
20
and the tendency to form elongated cells and filaments. Several observations relevant to this question are evident from Figs 1 and 2. First, it was clear that a polar division did not necessarily result in a delayed non-polar division, A common division pattern is shown in Fig. 1D, in which a cell had completed a polar division at 20 min and another at 30 min, after which a non-polar division had already occurred at 40 min. Clone N (Fig. 2) shows a polar division that took place at 30 min, which was followed shortly after by a non-polar division. Similar examples can be found in Figs 1C,2Jand2M, Second, cell elongation without division could occur in the absence of minicell production, as exemplified by clone J in Fig, 2, At 10 min, the mother cell divided and
Cell division in Escherichia coli minB mutants
i cells
minC point mutation JL ,r\. 11
\i,^-
Cell division in Escherichia coli minB mutants Thomas Akerlund, Rolf Bernander and Kurt Nordstrom* Department of Mierobiology, Biomedical Centre, Uppsala University, Box 581. S-751 23. Uppsala, Sweden. Summary In Escherichia coli minB mutants, cell division can take place at the cell poles as well as non-polarly in the cell. We have examined growth, division patterns, and nucleoid distribution in individual cells of a minC point mutant and a minB deletion mutant, and compared them to the corresponding wild-type strain and an intRI strain in which the chromosome is overreplicated. The main findings were as follows. In the minB mutants, polar and non-polar divisions appeared to occur independently of each other. Furthermore, the timing of cell division in the cell cycle was found to be severely affected. In addition, nucleoid conformation and distribution were considerably disturbed. The results obtained call for a reevaluation of the role of the MinB system in the E coli cell cycle, and of the concept that limiting quanta of cell division factors are regularly produced during the cell cycle.
Introduction Cell division in exponentially growing Escherieiiia coli strains normally takes place in the centre of the cells, betv^'een two newly replicated and segregated nucleoids. A mutant E, co//strain has been isolated which, in addition to non-polar divisions, also carries out polar divisions that result in the formation of DNA-less minicells (Adier et ai. 1967). Furthermore, in addition to normal-si2ed cells. elongated and filamentous cells are frequently produced by this mutant. The locus was denoted minB, and is located at 26 min on the genetic map of the E. coli chromosome (Bachmann, 1990). A model was proposed for how the minB operon may act during the E. co/;cell cycle {Teather etai, 1974). The key concept in this model is that the MinB system functions to direct cell division to the centre of the cell, by inactivating old' division sites at the cell poles. For each unitcell doubling, enough division factors {a 'quantum') are Received 29 January. 1992; revised and accepted 30 March, 1992. *For correspondence. Tel. (18) 174526: Fax (18)530396.
produced to carry out a single division. The division factors can normally only be used in the centre of the cell. whereas in minB mutants the old division sites at the cell poles are equally available for division and thus a 'choice' between which site to use is possible. The factors are entirely consumed in the division process, and a polar division is, therefore, carried out at the expense ot a cen* tral division. Hence, a cell that is two units in size should normally divide into two cells, each of which would be one unit in size, if, instead, a polar division occurred, the cell would have to grow in size from two to four units, i.e. double its cell mass, before new divisions could occur (cf,. Donachie. 1991). Since the cell size in this case has increased by two units, two quanta of division factors are available, and two new divisions may be carried out, again either polarly or non-poiarly. Consequently, both minicells and elongated cells are expected to be present in cultures of minB mutants. In the mode!, chromosome replication as well as nucleoid structure, segregation and partition are not expected to be affected. Importantly, the timing of chromosome replication and cell division in the cell cycle are also assumed to function normally: only the localization of the divisions is affected. Although this model has gained widespread acceptance, biochemical data in support of the various proposed functions are lacking; e,g. specific association between any of the Min proteins (below) and the cell poles has not been demonstrated (cf. de Boer et ai, 1991). The minB operon was cloned and sequenced and shown to code for three proteins: MinC, MinD, and MinE (de Boer etai. 1989). The MinC and MinD proteins were suggested to perform inhibitory functions, i.e. to inactivate the old division sites discussed above. The MinE protein was proposed to be the specificity factor that prevents MinCD from acting at non-polar division sites. During studies of the E. eoli celt cycle using the socalled intRI strains (Nordstrom etai. 1991) developed in our laboratory, we noticed that during over-replication of the chromosome also the intm strains readily produced minicells (Bernander ef ai, 1989). We decided to further examine minicell production in the intHI strains, using the well-characterized minB mutants as reference strains. During these studies we noted several discrepancies between our experimental findings and the proposed model for the function of the minB operon (above). Here, we report our investigations of the division pattern and nucleoid distribution in individual cells carrying minB mutations, as well as in intRI strains. We discuss the
2074
T. Akerlund, R. Bernander and K. Nordstrom
proposed role of the MinB system as a positional specificity system in E. coli cell division, and the concept of a quantal behaviour of ceil division factors during the cell cycle. Results Experimental outline Earlier reports have indicated that the cell size distribution and the division pattern vary in different minB mutants, and also depend upon the genetic background and on the growth conditions used (Frazer and Curtiss, 1975; Jaffe et al.. 1988; Woidringh et al., 1991). Consequently, the experiments were performed under similar growth conditions and in the same genetic background. To be able to compare our results with earlier reports, we used the same strain, P678-54. that was used in the experiments that form the basis for the model of the role of the MinB system (Teather etal.. 1974)- Therefore this strain, which carries the originally isolated mInC point mutation, was chosen, although additional mutations may be present because of the mutagenization that gave rise to the strain. To investigate possible differences between the point mutant and a mutant in which the entire minB locus has
Table 1. Growth rate and division capacity of the wild-type strain and the minB mutants. Division Capacity"^
Growth Rate (Doublings Per Hour) LB + 0.27., Glucose
Strain
batch culture^
2.5 Wild type mmC point 2.1 mutant minB deletion 2.3
M9CA
on top of agar''
within batch softagar^ culture^
22
1.9
2.0 2.1
1.6 1.9
Cell Sizes short cells
long cells
1.3
0.72
_
0.9 1.1
0 73 073
0.55 0.88
a. The growth rates in liquid media were determined by measuring cell density in a Klett-Summerson colorimeter. Each value is a mean of at least three individually performed experiments. M9CA is M9 medium (Maniatis er a/., 1982) supplemented with 1% casamino acids and 0.2% glycerol. b. The growth rates of individual clones were determined by adding together the lengths of all the cells in a microcolony at each time point, and plotting the sum of the cell lengths versus time. By measuring the slopes of the curves obtained, the doubling times in the microcolonies could be estimated. The values are mean values of all clones followed for each method. c. The division capacity was defined as the number of divisions per length unit and cell-length doubling. One length unit was defined as the average size of a newborn wtld-type cell (3.4 nm). Randomly chosen cells from pedigrees exemplified in Fig. 2 were foHowed for exactly one doubling in length, and the sum of the lengths of the mother cells was divided with the total number of divisions that took place during the doubling period. These values were in turn divided with one length unit. 3.4 |am. For the minB mutants, short and long cells were defined as cells shorter or ionger than 6.4 |am, respectively. The average cell size fof the short and long cells was about 5 and 9 |im, respectively, for both mutants.
Fig. 1. Time-lEyase microphotography of E. coliceWs growing on top ot agaf media containing Luria-Bertani medium and 0,2% glucose. The columns show one microcolony o( the wild type (A), two microcolonies of the m;nSdeletion mutant (B and C). and one microcolony of the minC point mutant (D).
been deleted (de Boer et al.. 1989), the deletion was introduced into strain P678, the parent of P678-54, by PI transduction. For direct microscopic observation of the growth and division patterns of the various strains, the cells were immobilized on solid growth media. Two microculture techniques were used; the cells were either grown on flat agar surfaces, or within 0.6% soft agar. The growth rates in the microcolonies were determined by measuring the lengths of all the cells in a microoolony at each time point, and plotting the sum of the cell lengths versus time. The doubling times in each of the microcolonies was estimated by measuring the slopes of the curves obtained. The average doubling times were generally longer on solid media than in liquid media, and the minB mutants usually displayed decreased growth rates (Table 1) as well as a larger variation of growth rates in individual microcolonies (not shown) relative to the wild type. The cells growing on solid media were microscopically inspected every five minutes, and micrographs were taken at every second observation event. A total of 38 microcolonies of the wild type, 74 of the minC point mutant, and 58 of the minB deletion mutant were studied for 50-70 min. fvlicrographs that illustrate different representative growth patterns are shown in Fig. 1, and stylized pedigrees constructed from such micrographs are presented in Fig. 2.
Cell division in Escherichia coli minB mutants
0
AminB
wt
'ml'l
A
"
F
10 10 30
JL J . . . . .
so
0 10 10
-
-
B
•
*
-
p
ZL. w
•
G -
7
- — X -)-*.
• •
-
Q
I
—
-*-
X X . . . .
40 SO
no
"
K
-
JL __
w M 70
minC point mutation
Jm _ _
— . JL JL
70 0 10 10
C
40 .
50 M 70
0
H -
T "
_
i
•
I 1 .
.
_
_
M
JUL. . ^ ^ .
a _
X - _
*
• ^^ ^
D 7
I
iO 40 60 70
0
. . J. 1 . . . . » ^ ^ ^ X X X _ E
10 10 40 JO M) 70
R
-
.
•
—
*
-
z z *_
^— t
t
.
N
.
—
•'•
S ^
^
t
t*. 1 ..
.
JLJ
Jm. J L .
o 2
;
•
_
J. .
.
•
Fig. 2. Pedigrees ol E co/i microcolonies derived from direct observaiion of growing cells as exemplided in Fig 1 Microcolonies A-O were grown on top of 1 2% agar surfaces whereas colonies P-T were grown within 0.6 % soft agar. Pedigrees A-E show five wiW-type microcolonies. F-J five colonies ol the m;nBdeletion mulanl. and K-T 10 colonies of the mmC point mutant. Minicells are not drawn to scale
^
. J ^ -*- ^
10 10
-
t.
•
-
3.
^ ^ .
•
2075
T t
• ^ t
X
^ ^ ^
^
_
^ ^_ .
JL JL
. J;
_t JJ^_ , ,
Celt sizes In minB mutants, cell division can take place at the cell poles as well as non-polarly. It has been proposed that cell division in minB mutants results in daughter cells (minicells excluded) whose lengths are multiples of the average length of a newborn wild-type cell (Teather etal.. 1974). We wanted to determine whether this was the case, and whether both polar and non-polar divisions resulted in daughter cells of distinct ceil lengths. First, we measured the cell lengths of more than 120 newborn cells of each strain after non-polar divisions. The newborn wildtype cells formed a uniform size class with a mean cell length of 3.4 ±0.2 nm (Fig. 3A). For the minB mutants, the cell sizes fell into two major classes, with a mean length of 4.7±0.3 |jm tor the class containing the smallest cells (Figs 3B and 3D). Thus the average cell size was larger than that of the wild type, despite the loss of cell material due to minicell production, and we could not detect any cells in the minB mutants that were smaller than the smallest cells of the wild type. Furthermore, non-polar divisions could take place centrally in the cell (Fig. 1C at 20 min) as well as at asymmetric positions (Fig. 1C at 50
and 70 min) in the minB mutants. The divisions did not appear to be randomly positioned in the cells, since both types of divisions gave rise to daughter cells of distinct size classes (Figs. 3B and 3D). In contrast to the daughter cells produced after nonpolar divisions, daughter cells (minicells excluded) resulting from polar divisions could not be grouped into distinct size classes (Figs. 3C and 3E). Furthermore, the minicells produced by the polar divisions were found to vary greatly in size, from a diameter close to that of the rod-shaped mother cells down to one-fifth or less of this diameter (cf. Fig. 1C at 70 min). The heterogeneity was confirmed by flow cytometry, in which the minicells displayed a considerable variation in size (not shown). In conclusion, an average newborn cell of the smallest size class from the minB mutants was longer than that of the corresponding wild-type strain. Non-polar divisions were found at both symmetric (central) and asymmetric positions along the length axis of the cells. The non-polar divisions did not take place at positions that were multiples of the average length of a newborn wild-type cell, although they appeared to be non-randomly positioned
2076
T. Akerlund. R. Bemander and K. Nordstrom
wt
R g . 3. Sizes of newborr) cells. The cells were grown on top of agar medium, and the sizes of newborn cells were measured directly under (he microscope. The panels show newborn wild-type cells (A), newborn ceils of the minCpoint mutant and the mrnSdeletion mutant after non-polar divisjons (B and D, respectively), and after polar divisions only (C and E, respectively). The open areas in panels B and D represent newborn cells that resulted from asymmetric divisions in the mother cells. The numbers of cells analysed were 120, 181, 150, 160, and 130 in panels A, B, C, D. and E. respectively.
non-polar divisions
minC point mutation
minC point mutation
non-polar divisions
polar divisions
Q lEl non-poUidivWoni Q uymmitric dlvidoiu
10
15
20
AmmB
AntinB
poUr divisions
non-polar divisions 0,3 n
•
ill non-poUi diviiioiu
E 0,2
0,1
0,0 5
10
15
20
0
size (|xm)
5
1•
along the length axis. Daughter cells produced after polar divisions could not be grouped into distinct size classes. The minicells displayed a broad cell-size distribution, indicating that the polar divisions that resulted in minicell formation were not initiated from a fixed position at the cell pole. Minicell divisions and filamentation According to the model described in the Introduction, a polar division is formed at the expense of a non-polar division. Therefore, the next division(s} cannot take place until the cell has doubled its mass, leading to filamentation. We wanted to investigate whether we could confirm such a relationship between the production of minicells
10
15
20
and the tendency to form elongated cells and filaments. Several observations relevant to this question are evident from Figs 1 and 2. First, it was clear that a polar division did not necessarily result in a delayed non-polar division, A common division pattern is shown in Fig. 1D, in which a cell had completed a polar division at 20 min and another at 30 min, after which a non-polar division had already occurred at 40 min. Clone N (Fig. 2) shows a polar division that took place at 30 min, which was followed shortly after by a non-polar division. Similar examples can be found in Figs 1C,2Jand2M, Second, cell elongation without division could occur in the absence of minicell production, as exemplified by clone J in Fig, 2, At 10 min, the mother cell divided and
Cell division in Escherichia coli minB mutants
i cells
minC point mutation JL ,r\. 11
\i,^-
