~1 [NSTITUTPASrEUR/ELSEVIER Paris 1991
ReS. MicrobioL i991, 142, I.~tl-115
Bacterial growth control studied by flow cytometry E. Boy¢ "~ a n d A. Lobner-Olescn t2)
m Department of Biophysics, Institute of Cancer Research, Montebello, 0310 Oslo J and ¢2)Department of Microbiology, Technical University o f Denmark, 2800 Lyngby (Denmark)
SUMMARY By employing flow cytometry, the DNA content and cell size of individual hactarial calls may be determined rapidly and with high precision. Also, the number of DNA replication origins in Escherinbia cob cells can be measured after treating the cells with rifampicin together with the cell division inhibitor cepbale~in. As opposed to wild type cells, certain mutants contain, with high frequency, a number of origins different from 2", indicating that the mutants do not initiate DNA replication at all origins simultaneously. Here we give evidence that this asynehrony phenotype cannot occur as e consequence of aberrant chromosomal segregation or call dh/Islon, but can only be caused by defeotive coordination of multiple initiation events within one and the same cell. Flow cytometry has been used to perform exact and detailed analyses of the growth and cell cycle of E. coil While the DNA distribution of a bacterial culture was unchanged as long as eteedy-etate growth was maintained, the cellular ONA content was reduced when the culture approached and entered stationary phase. Only after prolonged incubation in stationary phase did the cells contain fully replicated chromosomes, and rapidly growing cells ended up with either 2 or 4 chromosomes in stationary phase. Key-words : Flow oytometry, Escherichia co/i, DNA, Replication, Cell cycle control;
Rifampicin.
IN/RODUCT;Olq Flow cylometry is an elegant method to measure the contents of specific macromolecules in individual cells. In constrast to classical genetic and biochemical techniques, flow cytometry reveals the distribution of parameter values, thereby adding important information to measurements of heterogeneous populations. During the last decade, flow cytometry has been successfully applied tO m i c r o b i o l o g y (Boye a n d Lobner-Olesen, 1990a). If bacteria are properly stained with a DNA-specific fluorescent dye,
the DNA content of each bacterium may bc determined (Steen and Boye, 1980). The amount of scattered light emanating from a bacterium as it passes through the excitation focus of a flow eytometer is a good measure of the cell mass (Boye et aL, 1983). Thus , the D N A content and mass of individual bacterial cells may be measured simultaneously (Steen and Boye, 1981). By using a flow cytometer with high sensitivity and low background noise, the two p a r a m e t e r s may be determined with a coefficient of variation (CV) of a few percent. The precision and versatility of
f~ow cytol~etry m a k e s it a powerful method for studies of bacterial growth comrol. Here we describe some of the microbiological applications of flow cytometry and discuss the implications and limitations of the data emerging from such studies. MATERIALS AND METHODS Bacteria and growth media The Eschericbia coil KI2 strains AB1157 ( B a c h m a n n , 1987) or CM742 dnciA46 ( H a w
t~, BOYE AND A. LOBNER-OLE3EN
132 sen et al.. 1984) were grown in shaker flasks at 37 °C (ABIi57) or 30 ~C (CM742) itl either AB nlinio|;d mcdiunt (Clark and Maaloe, 196"/) supplemented ~ith I% casamino acids and 0.20/0 glucose or in LB with 0.2070 glucose. Baclerial growth was tnot:ilored by measuring the O D at 450 nm for ntinimal medium and al 600 nm for LB.
Drag treelmenls Rifampicin was added at a final concentration of 150 # g / m i ; c~phalexin (Eli Lilly, B a s i n g s l o k e , E n g l a n d ) was added at 10 # g /ml. The drugs were added when the cell culture had reached an O D of 0.1-0.2, whereafler ioeubalion was continued f o r 3-4 h before samples were taken fur flow cytometry.
Flow cytomelr), Cell preparation and flow cytometry with an "Argus 100" flow cytometer (Skatrou, Tranby. Norway) was performed as described previously (Skarslad et aL, 1985) Particle counting was based on the light scatter signal and was performed on unstained ethanol-fixed bacteria in proper dilutions.
RESULTS When E. coil cells are treated with the transcription inhibitor rifampiein (RIF), initiation of D N A replication is inhibited. while ongoing rounds of replicalioo are allowed to proceed to the termini. Therefore, treatment with RIF results in calls with an integral number of fully replicated chromosomes. After RiF treatment, a wild type cell conCV : coelt~ciem or ~ariation. LB = Lmia broth.
talus 2" chromosomes ( n = 0 . I, 2, 3), showing that each initiation event results in a doubling of the number of replication forks per cell. The conclusion is that all re pl i c a t i on o r i g i n s present in each individual cell are initiated simultaneously (Skarstud et at., 1986). S o m e dnaA(Ts) mutants perform asynchronous initiations even at the fully permissive t e m p e r a t u r e (Skgrstad et aL, 1986, 1988). The lack of synchrony in a danA46 mutant is demonstrated by the high frequency of cells with 3, 5, 6 and 7 fully replicated chromosomes after R i F treatment (fig. 1). Cells with different integral numbers of funy replicated chromosomes are clearly resolved by flow cytometry. The CV of the individual peaks in the D N A histogram shown in between 2.3 and 2.5%, demonstrating how precisely the D N A content of individual cells ego be measured. T h e C V of the D N A measurements may be reduced by about 25070 by using D20 instead of H20 Ln the flow (Stokke and Boye, unpublished.).
t f O rain
It may be argued that the above histogram (fig. i) is a result of aberrant chromosomal s e g r e g a t i o n a n d not async h r o n o u s initiation o f D N A replication. Consider the ceil cycle of a cell dividing every 30 min (fig. 2). For simplicity, let us assume that the D N A replication period (C) is 50 rain and the postreplication period (D) is 20 min. With synchronous initiation, the
ONA
fluor|lce~¢#
Fig. I. DNA histogram of strain CM742 danA46 grown in LB plus glucose at 3O°C and treated with RIF. The abscissa unit refers to number of fully replicated chromosomes.
t : 10min
t = 20 rain
U) ~
RIF
i RiF
0000
0000
0000 0000
Fig. 2. Dependence of chromosomal configuration on cell age. The cells were assumed to grow with a doubling time of 30 min, with a C period of 50 rain and a D period of 20 rain. The configurations at cell ages 0 min, tO min and 20 rain are shown in the upper row, and after treatment with RIF in the lower row. Replication forks are marked with filled circles. I
OD = optical density,
I
RIF=rifampicin,
B A C T E R I A L GROWTH CONTROL STUDIED BY FLOW CYTOMETRY
cell will contain 2n origins o f replication at any point in the cell cycle: a newborn cell will contain one replicating structure with 4 origins (upper left); in a 10-min old cell this structure is about to be parted n two (middle), while a 20-rain old cell (right) has two replicating structures, each with 2 origins just being initiated. At no point in the cell cycle can these chromosomes be segregated to give anything but 2 n origins per cell, and the origins will give rise to 2" fully replicated c h r o m o s o m e s after replication run-out in the presence of RIF (fig. 2, lower row). The only way aberrant cell segregation, in theory, can result in a histogram like the one shown in figure 1 is if the D period is longer than the doubling time, so that termination o f replication occurs in the generation preceding its corresponding cell division. The fact that some residual cell division does occur even after the addition of RIF may cause problems in interpreting D N A histograms of RIF-treated cells. If division were blocked immediately, the number o f fully replicated chromosomes per cell would reflect the number of replication origins present at the time o f drug addition. A low concentration of the antibiotic cephalexin was used to inhibit cell division, and the effect was monitored by using the flow cytometer to measure the concentration of bacteria with and without drug treatment. Addition of RIF caused the increase in cell number to slow down and to stop after about 10 rain (fig. 3). When cephalexin was added together with RIF, the cell concentration stopped its increase within less than a minute, in evidence of an almost instantaneous inhibition o f cell division. No signs o f cell lysis could be detected even aftel continued incubation for several hours. T h e effect o f e m p l o y i n g cephalexin in addition to RIF is
demonstrated by an increased number of chromosomes per ceil (fig. 4). When strain ABII57 was grown in LB supplemented with 0.2o70 glucose (doubling t i m e ~ 2 6 rain) and treated with RIF, most of the cells ended up with 4 fully replicated chromosomes, while a few ended up with 2 or 8 c h r o m o s o m e s (panel A). Addition of cephalexin together with RIF blocked all cell division, and the frequency of g-chromosome cells increased markedly, while virtually no cells contained only 2 chromosomes (panel B), Thus, many of the 2and 4-chromosome cells in panel A originated from cells containing 4 and 8 origins, respectively, at the time of drug addition, but which divided afterwards. The cellular D N A distribution in a culture depends on the growth medium, on the temperature and on the stage of growth. Cells of strain ABl157 growing with a doubling time of 34 min contained an amount of D N A equivalent to about 3 fully replicated chromosomes when newborn and divided with twice that amount of D N A (fig. 5A). This is consistent with the durations of the C and D periods in this s t r a i n ( A l l m a n a n d Boye, manuscript in preparation). The optical density of the culture increased exponentially with time up tO an OD value of 0.6, whereafter the growth rate decreased continuously and the OD ended up at a value of 1.8 in stmionary phase (data not shown). The D N A histogram is predicted to be independent of cell density as long as steady-state growth is maintained. Nonetheless, the D N A histograms for cells m OD 0.11 (fig. 5A) and 0.36 (fig. 5B) are different, even though the OD increase was strictly exponential in this range. The average DNA content of the ceils shown in panel B is 5o7o higher than of those in panel A, which is mainly due to an increased frequency of cells with more than 8
,
,~,
133
.~ .....
,~
l~'~ [n'Hn]
FI~, 3. Cell number as a function c.f time after addition of RIF or RIF plus cephalexin. Strain ABII57 was grown in LB plus glucose and relative cell concentration measured by flow cytometry after no additions (-), addition of RIF (o), and addition of RIP and cephalexin (n).
A
o B
Z
ii 5O DNA
1C9
lso 2oo ~s (channel no.)
fluorescence
Fig. 4. DNA histograms of drug treated cells. Strain ABl157 was grown in LB plus gluc:~se and treated with either RIF (pcmel AI or RIF plus cephalexin (panel 13) For 3 h.
E. BOYE AND A. L~)BNER-OLESEN
134 c b r o m o s o m e equivalents in panel B. Such differences have been observed frequently, but only when the cells approach stationary phase, and cannot be uttribued In poor reproducibility. A similar effect is pronounced if LB medium without glucose is used, ia whicil case steady-state growtb is very hard Io obtain, even t h o u g h the O D m a y increase exponentially for several generations (data not shown). At an O D of 0.56, the growth rate of the cells just started In decrease and at this point, the DNA histogram (fig. 5C) was markedly changed from the one representing steady-state growlh. Tile average DNA content was decreased by 20o'0 compared to that at an O D of OAt. At an O D of 1.3, the histogram displayed a promine~; peak corresponding to a DNA camera of 2 fully replicated chromosomes (fig. 5D), indicative of an inhibition of initiation of DNA replication. After I h in stationary phase, most of the cells contained 2 fully replicated ch r o mo so mes Ifig. 5E). Continued incubation for fifteen hours in stationary phase resulted in cells containing either 2 or 4 fully replicated chromosomes (fig. 5F). Further incubation allowed some residual cell division to occur, whereafter some l-chromosome cells could be detected (data not shown).
the cells within seconds to stop transcription and thereby initiation of D N A replication. Conceivably, intermediate RIF concentrations in these seconds may perturb initiation and produce artifacts. H o w e v e r , evidence favours the interpretation that RIF action is rapid and complete. First) other drugs inhibiting transcription or protein synthesis give similar results to what RIF does. Second, the resulting D N A histogram after RIF treatment is independent of RIF concentration as long as it is high enough (above 100 #g/ml).
quick[y and precisely measure the concentration of bacteria (fig. 3). Certain mutations have been found to confer an asynchrony phenotype, i.e, all replication origins present are not initiated simultaneously (Skarstad et al.. 1988, Skarstad and Boye, i988, Boye et al., [088). It is argued here that the asynchrony pheaotype c a n n o t o r i g i n a t e f r o m defective chromosome segregation or cell division. The molecular bases for the asynehrony p h c n o t y p e of the d i f f e r e n t mutants are not known, hut a model explaining the phenomenon for dam mutant ceils has been p r o p o s e d (Boye and Lobner-Olesen, i99Ob). Addition of RIF and cephalexin to a bacterial culture stops initiation of D N A replication and cell division rapidly (fig. 3). RIF presumably enters
|
.
.
.
The present data demonstrate that flow cytometry is a very sensitive method to monitor the growth state of a bacterial culture (fig. 5). Even small deviations from steady-state growth can he detected. It was revealed that exponentially growing E. coil cells deviate from steady-
.
DISCUSSION The data presented demonstrate the precision and sensitivity with which flow cytometry cgr, d e t e r m i n e bacterial cell parameters. Not only can be DNA content and cell size be measured, but also the number of replication origins present in each individual cell. This feature has proven invaluable in characterizing the control o f D N A replication in E. c~li (reviewed in Boye and Lobner-Olesen, 1990b). It may also be noted that flow cytometry can be used to
°l:i °
t
B
c
E
F
: 1o0
~e
DNA fluorescence(channelno, Fig. 5. DNA histograms of cells at different stages of growth. Strain ABI157 was grown in minimal medium with glucos~ and samples were removed for flow cytometry at different cell densities. The optical densities of the cultures were 0.ll (A), 0.36 (B), 0.56 (C), 1.3 (D), 1.8 (E) and 1.8 (F). In pand E. the culture had just reached its maximal OD; the hislogram of cells 15 h later is show in panel F. The position of cells containing 8 fully replicated chromosomes is indicated in each panel.
B A C T E R I A L G R O W T H C O N T R O L S T U D I E D B Y [:LOW C Y T O N I E T R Y
state growth at least one generation before any deviations can be seen on the increase in OD. Thus, it is imperative that experiments requiring steady-state growth are performed at least one generation before the growth rate starts to decrease. Cell division and initiation of DNA replication are inhibited as a rapidly growing culture of E . c e l l is allowed to grow into stationary phase. Such cells eventually contain an integral number of folly replicated chromosomes (fig. 5F). Contrary to common conceptions, they do not contain 9he but rather lw¢~ J r tour chromosomes. This effect is growthrate-dependent; slowly growing cells will contain one or two chromosomes in stationary phase (data not shown). A similar growth rate dependence has been observed for B a c i l l u s spores, which may contain either one, two or three chromosomes, depending on the growth condilions (AIIman and Boye, manuscript in preparation).
Acknowledgements This work was supported by grant no. GM39582 from the National Institutes of Health(to E.B.) and by the Danish Center for Microbiology (to A. L.
O.), and by a NATO Collaboralive ge~earch Grant to both. We are grateful to Harold B. Sreen and Trend Stokke rot assistance'~ith the flow cylOmetry.
References Bachmann, B.J. (1987), DerivaTions and genotypes of some mutam derivatives of Escherichia coil K-t2, in "Escherichia coil and Salmonella typhimurium: Cellular and Molecular Biulogy" (F.C. Neidhardt) (pp. 1190-1219). American Society for Microbiology, Washington, D.C. Boye, E., Lobner-Ole~cn, A. & Skarstad, K. 0988), Timing of chromosomal replication in Escherichia coIL Biochim. Biophys. Acta. (Amst.), 951, 359-364. Buye, E, & Lobner-Olesen, A. (1990a), Flow cytometry: illuminating microbiologY. New Biologist, 2, 119-125. Boye, E. & Lobner-Olesen, A. (1990bl, The role of d a m methyltransferase in the control of DNA replication in E. coil Cell, 62, 981-989. Boye, E., Steen, H.B. & Skarstad, K. 11983l, Flow cytometry of bacteria: a promising tool in experimental and clinical microbiology. J. ; en. MicrobioL, 129, 973-980. Clark, D,J. & Maaloe, O. 0967), DNA replication and the divi-
~ion cycle in Escherichia coli. J. moL Biol., 23, 99-112 Hansen, E.B., Atlung, T., Hansen, F.G., Skovgaard, O, & yon Meyenburg, K, (1984), Fine structure genetic map and complementation analysis of mlllatiuns in the dnttA gene of Escherichio c o i l Mol. gen. Genetics, 196, 387-396. Skarslad, S. & Boye, E. (1988), Periurbed chromosomal replication in recA mutants of Escherichia coil J. Bact., 170. 2549-2554. Skarstad, K., Boye, E. & Steen, H.B. (1986), Timing of initiation of chromosome replication in individual Eseherichia cob. cells. E M R O J., 5, 1711-1717. Skarstad, K., Steen, H.B. & goye, E, (1985), Escherichia coil DNA distributions measured by flow cytometry and compared with theoretical computer simulations. J. Bact., 163, 661-668. Skarstad, K., von Meyenburg, K., Hansen, F.G. & Buye, E. (1988), Coordination of chromosome replication initiation in Escherichia cull: effects of different dnaA alleles. J. Bact., 170, 2549-2554. Steen, H.B. & Boye, E. (1980), Bacterial growlh studied by flow cytometry. Cytometry, I, 32-36. Steen, H.B. & Boye, E. (1981l, Growth of Escherichia coli studied by dual parameter flow cytometry. J. Boot., 145, 1091-1094.
ReS. MicrobioL i991, 142, I.~tl-115
Bacterial growth control studied by flow cytometry E. Boy¢ "~ a n d A. Lobner-Olescn t2)
m Department of Biophysics, Institute of Cancer Research, Montebello, 0310 Oslo J and ¢2)Department of Microbiology, Technical University o f Denmark, 2800 Lyngby (Denmark)
SUMMARY By employing flow cytometry, the DNA content and cell size of individual hactarial calls may be determined rapidly and with high precision. Also, the number of DNA replication origins in Escherinbia cob cells can be measured after treating the cells with rifampicin together with the cell division inhibitor cepbale~in. As opposed to wild type cells, certain mutants contain, with high frequency, a number of origins different from 2", indicating that the mutants do not initiate DNA replication at all origins simultaneously. Here we give evidence that this asynehrony phenotype cannot occur as e consequence of aberrant chromosomal segregation or call dh/Islon, but can only be caused by defeotive coordination of multiple initiation events within one and the same cell. Flow cytometry has been used to perform exact and detailed analyses of the growth and cell cycle of E. coil While the DNA distribution of a bacterial culture was unchanged as long as eteedy-etate growth was maintained, the cellular ONA content was reduced when the culture approached and entered stationary phase. Only after prolonged incubation in stationary phase did the cells contain fully replicated chromosomes, and rapidly growing cells ended up with either 2 or 4 chromosomes in stationary phase. Key-words : Flow oytometry, Escherichia co/i, DNA, Replication, Cell cycle control;
Rifampicin.
IN/RODUCT;Olq Flow cylometry is an elegant method to measure the contents of specific macromolecules in individual cells. In constrast to classical genetic and biochemical techniques, flow cytometry reveals the distribution of parameter values, thereby adding important information to measurements of heterogeneous populations. During the last decade, flow cytometry has been successfully applied tO m i c r o b i o l o g y (Boye a n d Lobner-Olesen, 1990a). If bacteria are properly stained with a DNA-specific fluorescent dye,
the DNA content of each bacterium may bc determined (Steen and Boye, 1980). The amount of scattered light emanating from a bacterium as it passes through the excitation focus of a flow eytometer is a good measure of the cell mass (Boye et aL, 1983). Thus , the D N A content and mass of individual bacterial cells may be measured simultaneously (Steen and Boye, 1981). By using a flow cytometer with high sensitivity and low background noise, the two p a r a m e t e r s may be determined with a coefficient of variation (CV) of a few percent. The precision and versatility of
f~ow cytol~etry m a k e s it a powerful method for studies of bacterial growth comrol. Here we describe some of the microbiological applications of flow cytometry and discuss the implications and limitations of the data emerging from such studies. MATERIALS AND METHODS Bacteria and growth media The Eschericbia coil KI2 strains AB1157 ( B a c h m a n n , 1987) or CM742 dnciA46 ( H a w
t~, BOYE AND A. LOBNER-OLE3EN
132 sen et al.. 1984) were grown in shaker flasks at 37 °C (ABIi57) or 30 ~C (CM742) itl either AB nlinio|;d mcdiunt (Clark and Maaloe, 196"/) supplemented ~ith I% casamino acids and 0.20/0 glucose or in LB with 0.2070 glucose. Baclerial growth was tnot:ilored by measuring the O D at 450 nm for ntinimal medium and al 600 nm for LB.
Drag treelmenls Rifampicin was added at a final concentration of 150 # g / m i ; c~phalexin (Eli Lilly, B a s i n g s l o k e , E n g l a n d ) was added at 10 # g /ml. The drugs were added when the cell culture had reached an O D of 0.1-0.2, whereafler ioeubalion was continued f o r 3-4 h before samples were taken fur flow cytometry.
Flow cytomelr), Cell preparation and flow cytometry with an "Argus 100" flow cytometer (Skatrou, Tranby. Norway) was performed as described previously (Skarslad et aL, 1985) Particle counting was based on the light scatter signal and was performed on unstained ethanol-fixed bacteria in proper dilutions.
RESULTS When E. coil cells are treated with the transcription inhibitor rifampiein (RIF), initiation of D N A replication is inhibited. while ongoing rounds of replicalioo are allowed to proceed to the termini. Therefore, treatment with RIF results in calls with an integral number of fully replicated chromosomes. After RiF treatment, a wild type cell conCV : coelt~ciem or ~ariation. LB = Lmia broth.
talus 2" chromosomes ( n = 0 . I, 2, 3), showing that each initiation event results in a doubling of the number of replication forks per cell. The conclusion is that all re pl i c a t i on o r i g i n s present in each individual cell are initiated simultaneously (Skarstud et at., 1986). S o m e dnaA(Ts) mutants perform asynchronous initiations even at the fully permissive t e m p e r a t u r e (Skgrstad et aL, 1986, 1988). The lack of synchrony in a danA46 mutant is demonstrated by the high frequency of cells with 3, 5, 6 and 7 fully replicated chromosomes after R i F treatment (fig. 1). Cells with different integral numbers of funy replicated chromosomes are clearly resolved by flow cytometry. The CV of the individual peaks in the D N A histogram shown in between 2.3 and 2.5%, demonstrating how precisely the D N A content of individual cells ego be measured. T h e C V of the D N A measurements may be reduced by about 25070 by using D20 instead of H20 Ln the flow (Stokke and Boye, unpublished.).
t f O rain
It may be argued that the above histogram (fig. i) is a result of aberrant chromosomal s e g r e g a t i o n a n d not async h r o n o u s initiation o f D N A replication. Consider the ceil cycle of a cell dividing every 30 min (fig. 2). For simplicity, let us assume that the D N A replication period (C) is 50 rain and the postreplication period (D) is 20 min. With synchronous initiation, the
ONA
fluor|lce~¢#
Fig. I. DNA histogram of strain CM742 danA46 grown in LB plus glucose at 3O°C and treated with RIF. The abscissa unit refers to number of fully replicated chromosomes.
t : 10min
t = 20 rain
U) ~
RIF
i RiF
0000
0000
0000 0000
Fig. 2. Dependence of chromosomal configuration on cell age. The cells were assumed to grow with a doubling time of 30 min, with a C period of 50 rain and a D period of 20 rain. The configurations at cell ages 0 min, tO min and 20 rain are shown in the upper row, and after treatment with RIF in the lower row. Replication forks are marked with filled circles. I
OD = optical density,
I
RIF=rifampicin,
B A C T E R I A L GROWTH CONTROL STUDIED BY FLOW CYTOMETRY
cell will contain 2n origins o f replication at any point in the cell cycle: a newborn cell will contain one replicating structure with 4 origins (upper left); in a 10-min old cell this structure is about to be parted n two (middle), while a 20-rain old cell (right) has two replicating structures, each with 2 origins just being initiated. At no point in the cell cycle can these chromosomes be segregated to give anything but 2 n origins per cell, and the origins will give rise to 2" fully replicated c h r o m o s o m e s after replication run-out in the presence of RIF (fig. 2, lower row). The only way aberrant cell segregation, in theory, can result in a histogram like the one shown in figure 1 is if the D period is longer than the doubling time, so that termination o f replication occurs in the generation preceding its corresponding cell division. The fact that some residual cell division does occur even after the addition of RIF may cause problems in interpreting D N A histograms of RIF-treated cells. If division were blocked immediately, the number o f fully replicated chromosomes per cell would reflect the number of replication origins present at the time o f drug addition. A low concentration of the antibiotic cephalexin was used to inhibit cell division, and the effect was monitored by using the flow cytometer to measure the concentration of bacteria with and without drug treatment. Addition of RIF caused the increase in cell number to slow down and to stop after about 10 rain (fig. 3). When cephalexin was added together with RIF, the cell concentration stopped its increase within less than a minute, in evidence of an almost instantaneous inhibition o f cell division. No signs o f cell lysis could be detected even aftel continued incubation for several hours. T h e effect o f e m p l o y i n g cephalexin in addition to RIF is
demonstrated by an increased number of chromosomes per ceil (fig. 4). When strain ABII57 was grown in LB supplemented with 0.2o70 glucose (doubling t i m e ~ 2 6 rain) and treated with RIF, most of the cells ended up with 4 fully replicated chromosomes, while a few ended up with 2 or 8 c h r o m o s o m e s (panel A). Addition of cephalexin together with RIF blocked all cell division, and the frequency of g-chromosome cells increased markedly, while virtually no cells contained only 2 chromosomes (panel B), Thus, many of the 2and 4-chromosome cells in panel A originated from cells containing 4 and 8 origins, respectively, at the time of drug addition, but which divided afterwards. The cellular D N A distribution in a culture depends on the growth medium, on the temperature and on the stage of growth. Cells of strain ABl157 growing with a doubling time of 34 min contained an amount of D N A equivalent to about 3 fully replicated chromosomes when newborn and divided with twice that amount of D N A (fig. 5A). This is consistent with the durations of the C and D periods in this s t r a i n ( A l l m a n a n d Boye, manuscript in preparation). The optical density of the culture increased exponentially with time up tO an OD value of 0.6, whereafter the growth rate decreased continuously and the OD ended up at a value of 1.8 in stmionary phase (data not shown). The D N A histogram is predicted to be independent of cell density as long as steady-state growth is maintained. Nonetheless, the D N A histograms for cells m OD 0.11 (fig. 5A) and 0.36 (fig. 5B) are different, even though the OD increase was strictly exponential in this range. The average DNA content of the ceils shown in panel B is 5o7o higher than of those in panel A, which is mainly due to an increased frequency of cells with more than 8
,
,~,
133
.~ .....
,~
l~'~ [n'Hn]
FI~, 3. Cell number as a function c.f time after addition of RIF or RIF plus cephalexin. Strain ABII57 was grown in LB plus glucose and relative cell concentration measured by flow cytometry after no additions (-), addition of RIF (o), and addition of RIP and cephalexin (n).
A
o B
Z
ii 5O DNA
1C9
lso 2oo ~s (channel no.)
fluorescence
Fig. 4. DNA histograms of drug treated cells. Strain ABl157 was grown in LB plus gluc:~se and treated with either RIF (pcmel AI or RIF plus cephalexin (panel 13) For 3 h.
E. BOYE AND A. L~)BNER-OLESEN
134 c b r o m o s o m e equivalents in panel B. Such differences have been observed frequently, but only when the cells approach stationary phase, and cannot be uttribued In poor reproducibility. A similar effect is pronounced if LB medium without glucose is used, ia whicil case steady-state growtb is very hard Io obtain, even t h o u g h the O D m a y increase exponentially for several generations (data not shown). At an O D of 0.56, the growth rate of the cells just started In decrease and at this point, the DNA histogram (fig. 5C) was markedly changed from the one representing steady-state growlh. Tile average DNA content was decreased by 20o'0 compared to that at an O D of OAt. At an O D of 1.3, the histogram displayed a promine~; peak corresponding to a DNA camera of 2 fully replicated chromosomes (fig. 5D), indicative of an inhibition of initiation of DNA replication. After I h in stationary phase, most of the cells contained 2 fully replicated ch r o mo so mes Ifig. 5E). Continued incubation for fifteen hours in stationary phase resulted in cells containing either 2 or 4 fully replicated chromosomes (fig. 5F). Further incubation allowed some residual cell division to occur, whereafter some l-chromosome cells could be detected (data not shown).
the cells within seconds to stop transcription and thereby initiation of D N A replication. Conceivably, intermediate RIF concentrations in these seconds may perturb initiation and produce artifacts. H o w e v e r , evidence favours the interpretation that RIF action is rapid and complete. First) other drugs inhibiting transcription or protein synthesis give similar results to what RIF does. Second, the resulting D N A histogram after RIF treatment is independent of RIF concentration as long as it is high enough (above 100 #g/ml).
quick[y and precisely measure the concentration of bacteria (fig. 3). Certain mutations have been found to confer an asynchrony phenotype, i.e, all replication origins present are not initiated simultaneously (Skarstad et al.. 1988, Skarstad and Boye, i988, Boye et al., [088). It is argued here that the asynchrony pheaotype c a n n o t o r i g i n a t e f r o m defective chromosome segregation or cell division. The molecular bases for the asynehrony p h c n o t y p e of the d i f f e r e n t mutants are not known, hut a model explaining the phenomenon for dam mutant ceils has been p r o p o s e d (Boye and Lobner-Olesen, i99Ob). Addition of RIF and cephalexin to a bacterial culture stops initiation of D N A replication and cell division rapidly (fig. 3). RIF presumably enters
|
.
.
.
The present data demonstrate that flow cytometry is a very sensitive method to monitor the growth state of a bacterial culture (fig. 5). Even small deviations from steady-state growth can he detected. It was revealed that exponentially growing E. coil cells deviate from steady-
.
DISCUSSION The data presented demonstrate the precision and sensitivity with which flow cytometry cgr, d e t e r m i n e bacterial cell parameters. Not only can be DNA content and cell size be measured, but also the number of replication origins present in each individual cell. This feature has proven invaluable in characterizing the control o f D N A replication in E. c~li (reviewed in Boye and Lobner-Olesen, 1990b). It may also be noted that flow cytometry can be used to
°l:i °
t
B
c
E
F
: 1o0
~e
DNA fluorescence(channelno, Fig. 5. DNA histograms of cells at different stages of growth. Strain ABI157 was grown in minimal medium with glucos~ and samples were removed for flow cytometry at different cell densities. The optical densities of the cultures were 0.ll (A), 0.36 (B), 0.56 (C), 1.3 (D), 1.8 (E) and 1.8 (F). In pand E. the culture had just reached its maximal OD; the hislogram of cells 15 h later is show in panel F. The position of cells containing 8 fully replicated chromosomes is indicated in each panel.
B A C T E R I A L G R O W T H C O N T R O L S T U D I E D B Y [:LOW C Y T O N I E T R Y
state growth at least one generation before any deviations can be seen on the increase in OD. Thus, it is imperative that experiments requiring steady-state growth are performed at least one generation before the growth rate starts to decrease. Cell division and initiation of DNA replication are inhibited as a rapidly growing culture of E . c e l l is allowed to grow into stationary phase. Such cells eventually contain an integral number of folly replicated chromosomes (fig. 5F). Contrary to common conceptions, they do not contain 9he but rather lw¢~ J r tour chromosomes. This effect is growthrate-dependent; slowly growing cells will contain one or two chromosomes in stationary phase (data not shown). A similar growth rate dependence has been observed for B a c i l l u s spores, which may contain either one, two or three chromosomes, depending on the growth condilions (AIIman and Boye, manuscript in preparation).
Acknowledgements This work was supported by grant no. GM39582 from the National Institutes of Health(to E.B.) and by the Danish Center for Microbiology (to A. L.
O.), and by a NATO Collaboralive ge~earch Grant to both. We are grateful to Harold B. Sreen and Trend Stokke rot assistance'~ith the flow cylOmetry.
References Bachmann, B.J. (1987), DerivaTions and genotypes of some mutam derivatives of Escherichia coil K-t2, in "Escherichia coil and Salmonella typhimurium: Cellular and Molecular Biulogy" (F.C. Neidhardt) (pp. 1190-1219). American Society for Microbiology, Washington, D.C. Boye, E., Lobner-Ole~cn, A. & Skarstad, K. 0988), Timing of chromosomal replication in Escherichia coIL Biochim. Biophys. Acta. (Amst.), 951, 359-364. Buye, E, & Lobner-Olesen, A. (1990a), Flow cytometry: illuminating microbiologY. New Biologist, 2, 119-125. Boye, E. & Lobner-Olesen, A. (1990bl, The role of d a m methyltransferase in the control of DNA replication in E. coil Cell, 62, 981-989. Boye, E., Steen, H.B. & Skarstad, K. 11983l, Flow cytometry of bacteria: a promising tool in experimental and clinical microbiology. J. ; en. MicrobioL, 129, 973-980. Clark, D,J. & Maaloe, O. 0967), DNA replication and the divi-
~ion cycle in Escherichia coli. J. moL Biol., 23, 99-112 Hansen, E.B., Atlung, T., Hansen, F.G., Skovgaard, O, & yon Meyenburg, K, (1984), Fine structure genetic map and complementation analysis of mlllatiuns in the dnttA gene of Escherichio c o i l Mol. gen. Genetics, 196, 387-396. Skarslad, S. & Boye, E. (1988), Periurbed chromosomal replication in recA mutants of Escherichia coil J. Bact., 170. 2549-2554. Skarstad, K., Boye, E. & Steen, H.B. (1986), Timing of initiation of chromosome replication in individual Eseherichia cob. cells. E M R O J., 5, 1711-1717. Skarstad, K., Steen, H.B. & goye, E, (1985), Escherichia coil DNA distributions measured by flow cytometry and compared with theoretical computer simulations. J. Bact., 163, 661-668. Skarstad, K., von Meyenburg, K., Hansen, F.G. & Buye, E. (1988), Coordination of chromosome replication initiation in Escherichia cull: effects of different dnaA alleles. J. Bact., 170, 2549-2554. Steen, H.B. & Boye, E. (1980), Bacterial growlh studied by flow cytometry. Cytometry, I, 32-36. Steen, H.B. & Boye, E. (1981l, Growth of Escherichia coli studied by dual parameter flow cytometry. J. Boot., 145, 1091-1094.
