JOURNAL OF BACTERIOLOGY, Nov. 1973, p.

736-740 Copyright 1973 American Society for Microbiology

Vol. 116, No. 2 Printed in U.S.A.,

Synchronous Cultures of Bacillus subtilis Obtained by Filtration with Glass Fiber Filters
MICHAEL G. SARGENT

National Institute for Medical Research, Mill Hill, London, NW7 IAA, England
Received for publication 2 March 1973

A simple method of potentially wide applicability for obtaining synchronous cultures of Bacillus subtilis based on size selection is described. Using glass fiber filters, a population (about 1 to 2% of the parent population) can be obtained substantially enriched for small cells which grow synchronously. A method for rapidly concentrating dilute suspensions of cells is described.

Cultures of synchronously dividing bacteria can be obtained either by growing cells of a particular age class or by imposing synchrony (e.g., amino acid starvation). The latter methods must distort temporal relationships of cell processes, although damaging stresses may be incurred with selection methods (1, 6, 8). Phased cell division of bacilli has been obtained using starvation and stationary phase methods (4, 9, 14), and Imanaka et al. (7) have used a filtration method to obtain synchronous cultures of Bacillus cereus. A density gradient method of isolating synchronous cells of Bacillus subtilis has been described (2). Germinating spores may have synchronous chromosome replication and cell division, but these divisions are not typical of the normal process in exponential cultures (10). Enteric bacteria have been used more widely than bacilli for synchrony studies (6). This investigation describes a simple reproducible method for obtaining synchronous cultures based on size selection that causes minimum stress and gives sufficient yield for biochemical experiments. Growth conditions are indicated that eliminate chain formation and restrict septum formation to a small part of the division cycle. MATERIALS AND METHODS
Bacteria, growth, and media. B. subtilis 168 thytrp- was kept as a spore suspension in distilled water. Cells were grown on minimal medium containing phosphate buffer (pH 7.3) (7.1 g of Na2HPO4 and 1.35 g of KH2PO4 per liter of Analar grade chemicals), which was sterilized alone. Ammonium sulfate (2 mg/ml), magnesium sulfate (0.25 mg/ml), glucose (5 mg/ml), ferrous sulfate (7H20) (1.0 Ag/ml), manganese sulfate (4H,O) (0.1 ;g/ml), thymine (10 ug/ ml), and tryptophan (10 ig/ml) were subsequently added aseptically to give the final concentrations
736

shown. Spores were germinated overnight in shaken culture in minimal medium supplemented with 0.01% casein hydrolysate. Cultures to be used for synchrony experiments (250 ml in 1-liter flasks) were inoculated from an exponential culture so as to obtain growth equivalent to an absorbancy (540 nm) of about 0.8 on the following morning. Exponential growth was maintained for at least 12 generations and was not allowed to exceed an absorbancy of 2.0 at any stage. Bacteria were counted by use of a model FN Coulter electronic particle counter, fitted with a 30-gm probe. Dilutions were made in membrane-filtered (Millipore Corp.) 2% NaCl containing sodium azide (0.001 M) as a preservative. Samples were fixed with an equal volume of 10% Formalin and were counted usually within 3 h of sampling. Storage overnight had no effect on the count or on absorbancy. Bacterial numbers determined with the Coulter counter (referred to as Coulter units) represent single or septate units. Numbers of individual cells (c) can be calculated from the Coulter count (n) by the formula c = n (1 + s), where s is the average number of septa per Coulter unit. Septa were counted with a miscroscope (using either phase or normal optics) on Formalin-fixed cells, dried onto slides at room temperature, and mounted under cover slips with crystal violet (0.02%). About 200 cells in several random fields were counted. Selection of synchronous cells. The smallest cells in a population were selected by filtering a culture under vacuum through a pad of glass fiber filters. Three grass fiber filters (Whatman GF/C, 15 cm), sandwiched between two Whatman no. 1 paper filters, were mounted in the filtration unit. Glass fiber filters are more fragile than paper filters and must be protected from sharp edges by use of filter papers. The design of the filtration unit was similar in principle to that described by Helmstetter (6) (internal diameter 12 cm). To allow greater speed in changing filters, the top part of the unit was made of stainless steel of sufficient weight of insure a tight seal without clamps. Batches of cells (500 ml) at an optical density (OD)

VOL. 116, 1973

SYNCHRONOUS CULTURES OF B. SUBTILIS

737

of 1.5 to 2.0 (540 nm) were placed in the filtration unit, and the water pump was started. For most purposes the negative pressure was increased steadily from 0 to about 15 cm of mercury. By changing the filters once, 1 liter of selected cells (yield about 4%) could be obtained in 2 min. Further selection and concentration was achieved by passing the filtrate through another pad of glass fiber filters, a collection filter (GF/C, 5.5 cm) (mounted in a smaller apparatus), at a negative pressure of 30 to 40 cm of mercury. At this higher negative pressure, the entire population adheres to the filter, but can be eluted with fresh medium at low pressures (2 to 4 cm of mercury). Greater selectivity was obtained if the bacteria were eluted from this collection filter through a second pad of three filters. A final yield of 1 to 2% was obtained within 5 min. Radioactive methods. Synchronous cells were grown in the presence of '4C-tryptophan (0.05 qCi/ml) and 3H-thymine (0.gCi/ml). Incorporation was stopped by transferring samples (1 ml) to 1 ml of an ice-cold solution of thymine (200 gg/ml) and tryptophan (400 sg/ml), chloramphenicol (200 AsgAiter), lysozyme (100 ig/ml), sodium azide (0.002 M), and 0.01 M trisodium citrate in sodium phosphate buffer (pH 7.3) (0.05 M). Cells were lysed by incubating the sample at 35 C for 15 min, after which samples were collected and 10% trichloroacetic acid (2 ml) was added. After 30 min, precipitates were collected on GF/A filters and counted in toluene scintillation fluid.

FIG. 1. Size distribution of exponential phase cells and cells selected with glass fiber filters. Heavy line: selected cells, fine line: exponential phase cells.

TABLE 1. Factors affecting selection and yield of cells RESULTS when using glass fiber filters Selection of small cells by use of glass fiber Final Treatment filters. When suspensions of B. subtilis were (no. and kind Culture negative Selection Yield (%) applied passed through glass fiber filters under suction, vola of filters) pressureb index' only a small proportion of the bacteria were not 100 20 2.6 3.0 retained, and these could be substantially 1.3 GF/C 20 100 2.0 5.0 smaller than those in the parent population. 2. 2 GF/C 1.5 10.4 100 20 The -size distribution of bacteria from a filtrate 3. 1 GF/C 4. 3 50 20 2.9 2.1 and the parent population (with no correction 5.3 GF/C 40 100 < 0.3 GF/C _d for septation) is illustrated in Fig. 1. Selected 6. 2 GF/C 100 40 2.3 2.2 bacteria usually contained less than 2 septa per 7. 2 GF/C 10 50 1.8e 8.1 100 bacteria, whereas unselected bacteria con- 8.8 GF/A 100 20 2.3 2.5 tained 25 per 100. To quantitate selection, a The volume of culture shown was applied to a frequencies of small Coulter units of a specific size have been compared. The size class used 5.5-cm diameter filter. b Cells were drawn onto the filters starting at zero includes the smallest 25% of the parent populanegative pressure rising steadily to the value shown tion (referred to as first quartile size class). The of mercury). selection index (SI) used equals the number of (centimeters of Coulter units in selected population in c Number Coulter units in a selected population in the first quartile size class divided by number of Coulter first quartile size class divided by the number of units in parent population in first quartile size class. d Yield too small to give meaningful size distribuCoulter units in the parent population in the first quartile size class. The SI for the data in tion. eSelection index would be even lower if 100-ml Fig. 1 is 2.75. Table 1 illustrates a number of factors and elution volumes were used. conditions affecting the quality and yield of selected cells. Yields evidently decreased with volumes. GF/C filters gave more selection than increasing selection. Good selection was favored GF/A. Cell density did not affect the quality of by larger numbers of filters, greater negative selection, but there was a considerable reducpressures during filtration, and low elution tion in speed! of filtration when the number of

738

SARGENT

J. BACTERIOL.

cells applied to the filter exceed 2 x 101 cells per cm2. Selection was greater and yields lower at high negative pressures, but in either case better results were obtained if selection was performed with a progressive increase in vacuum from zero to the appropriate level. Negative pressures greater than 45 cm of mercury were not used, so that damage to the bacteria would be avoided, although there was no decrease in growth after exposure to such a vacuum for 5 min. The relationship between negative pressure and retentiveness of glass fiber filters has been exploited in collecting and concentrating cells from large volumes of medium. Thus, cells were collected on three GF/C filters at high negative pressures (40 cm mercury) and were then eluted with a small volume of medium at 1 to 2 cm of mercury to give a concentrated suspension. Further selection was obtained at this stage by eluting bacteria from the collection filters through more filters (see Materials and Methods). Cultural conditions for synchronous growth. In many strains of bacilli the time taken from septum formation to cell separation must be long in relation to the generation time as they are frequently multiseptate (12). Once the septum has formed, the daughter cells can be regarded as at cell age zero. Data based on numbers of growth units (Coulter or viable units, which correspond closely when using B. subtilis) must be corrected for the degree of septation to give cell number. Conditions have been sought that minimize cell separation time. Initial experiments with Spizizen minimal medium containing glucose (16) indicated that levels of magnesium, citrate, trace metals, and thymine, and pH value all affected separation time. Medium 9 (Table 2) is essentially the same as Spizizen medium. Citrate had a beneficial effect on growth without ion supplements (cf. media 8 and 9). Iron and manganese supplements stimulated growth to the same extent as citrate (cf. medium 1), although, even in the presence of these ions, citrate stimulated growth to a small extent (medium 7). However, the presence of citrate increased the cell separation time. Magnesium was required for most rapid cell separation (cf. media 1 and 2), so the effect of citrate can probably be ascribed to magnesium chelation. Without iron and magnesium supplements, either in the presence of absence of citrate, growth rate and separation time were strongly pH dependent, with a sharp optimum for growth at about pH 7.0 and with faster cell separation at high pH values (media

11 and 13). The effect of pH was markedly reduced in the presence of ion supplements (media 10 and 12). Reduced thymine levels increased the cell separation time and reduced the growth rate (media 1, 4, 5, and 6). However, at 4 gg of thymine per ml, the cell separation time is almost doubled without a significant effect on the growth rate. At 10 gg of thymine per ml, the mass growth rate of B. subtilis 168 trp- thy- is not significantly different to the thy+ parent strain. Figure 2 illustrates growth of bacteria selected, as described above in medium 1 (Fig. 1). Cell numbers and OD are shown on a,logarithmic scale. The importance of correcting Coulter counts for the number of unseparated cells is shown. A higher degree of synchrony is seen in the corrected curve than the uncorrected one. The number of bacteria with septa is also shown, again illustrating that cell division occurs periodically and therefore synchronously. Synthesis of macromolecules. '4C-tryptophan and 3H-thymine were incorporated in proportion to cell mass (OD) (Fig. 3), which suggests that deoxyribonucleic acid (DNA) and protein synthesis was exponential in synchronous cultures. With a generation time of 81 min, there was a 63% increase in DNA (in about 3 h) in the presence of chloramphenicol (100 ,ug/ml). This gives a DNA replication time of 125 min (13). As this is substantially greater than a cell generation time, chromosomes under these conditions must contain three forks (10).

DISCUSSION In principle, filtration methods of selecting uniform populations of small cells that will grow synchronously should be quicker than centrifugation and should not introduce the osmotically active solutes necessary for gradient centrifugation methods. With good selection, yields of cells must be small, so methods of handling and rapidly concentrating large numbers of cells must be available. A technique has been devised using glass fiber filters which meets these requirements. The degree of size selection and the total load that glass fiber filters will accommodate at a reasonable flow rate is much greater than with paper filters. For example, from 1 liter of culture containing about 4 x 108 cells per ml, a good synchronous population containing about 1 to 2% of the parent population can be obtained within 5 min. The increase in retentiveness at high negative pressures provides an especially useful method of combining a selection and concentration step. Cells ob-

VOL. 116, 1973

SYNCHRONOUS CULTURES OF B. SUB TILIS

TABLE 2. Factors affecting septation in minimal medium
Citrate

739

MediuMa

Conc of Mg (mM)

Fe-Mnb
+ + + + + + + _ _ + _ +
-

(0.1%)
-

pH

Thymine
(g)
10 10 10 4 2 1 10 10 10 10 10 10 10

Generation Coulter units time (min) with septa (%)

Separation
time (min)

1 2 3 4 5 6 7 8 9 10 11 12 13

1.0 0.1 0.01 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0

+ _ + _ + +

7.3 7.3 7.3 7.3 7.3 7.3 7.3 7.3 7.3 6.0 6.0 7.6 7.6

63 69 80 66 84 132 63 126 70 78 84 72 77

25 42 29 36 42 48 53 20 75 36 99 31 29

20 35 30 30 37 63 38 42 60 36 >84 37 27

aBasal medium contains phosphate buffer, ammonium sulfate, glucose, and tryptophan as shown in Materials and Methods. b Concentrations as in Materials and Methods.

FIG. 3. Synthesis of protein and DNA in synchronous cells. Radioactivity from 3H-thymine (0) and l4C-tryptophan (0) incorporated into trichloroacetic acid-preciptable material from 1 ml of lysed cells. Bars indicate interdivisional periods. FIG. 2. Growth of cells selected by filtration. Symbols: *, % of Coulter units with septa; , optical density (540 nm) (log scale); 0, cell number determined by Coulter counter (log scale); 0, cell number corrected for septation (log scale).

doubling time and time between divisions are equal to the preselection growth rate. (ii) The degree of synchrony is related to the degree of selection. (ii) DNA synthesis occurs in a manner compatible with the chromosome replication
time. The method described above meets most of these criteria, but growth is not completely balanced. Postselection mass growth rates are typically 20% lower than before selection, whether tested in previously used culture supernatant fluid or in fresh medium. Each step in the growth curve corresponds to a doubling in cell number, and the times between periods of cell division are equal to the preselection dou-

tained by filtration in this way grow synchronously (Fig. 2). Studies of filtration synchrony have indicated that unusual phenomena can be induced by stresses such as temperature changes, shocks on entering fresh media, and delays in preparation (1, 6, 8). Synchronous cultures can be regarded as "unstressed" if they meet the following criteria. (i) Growth is balanced, i.e., mass

740

SARGENT

J. BACTERIOL.

ACKNOWLEDGMENTS bling time. The decrease in mass growth rate is reflected in a decrease in cell volume. Advice and encouragement from H. J. Rogers and the In filtration, stress could arise from (i) physi- technical assistance of M. F. Bennett are gratefully acknowlcal effects of passing through a filter, (ii) expo- edged. sure to low pressure, and (iii) damage during colLITERATURE CITED lection. Criteria (i) and (iii) are difficult to test rigorously, although exponentially growing 1. Abbo, F. E., and A. B. Pardee. 1960. Synthesis of cells, collected on filters and eluted into fresh macromolecules in synchronously dividing bacteria. Biochem. Biophys. Acta 39:478-485. medium, show a small reduction in growth rate. Taber, and F. Exposure to reduced oxygen tension at the 2. Chattergee, A. N., H.synchronisation E. Young. 1971. A of Staphylococcus rapid method for growth temperature for 5 min has no adverse aureus and Bacillus subtilis. Biochem. Biophys. Res. effect on the subsequent growth rate. Commun. 44:1125-1130. The degree of synchrony observed (Fig. 2) is 3. Cummings, D. J. 1970. Synchronisation of Escherichia coli K 12 by membrane selection. Biochem. Biophys. less than might have been expected from the Res. degree of selection, although the size selection 4. Daniels,Commun. 7:448-452. M. J. 1969. Lipid synthesis in relation to the cell obtained is slightly greater than with the Helmcycle of Bacillus megaterium KM and Escherichia coli. Biochem. J. 115:697-701. stetter method (3, 8, 15). There are indications E., and S. Borenstein, 1971. that the distribution of interdivision times in 5. Ephrati-Elizur, replication in Bacillus subtilis.Velocity of J. Bactechromosome in Eschbacilli is substantially greater than riol. 106:58-64. erichia coli (11). The cell number curve (Coulter 6. Helmstetter, C. E. 1963. Methods for studying the microbiol. division cycle, p. 327. In J. R. Norris and D. count corrected for septation) appears more W. Ribbons (ed.), Methods in microbiology, vol. 1. synchronous than the uncorrected Coulter Academic Press Inc., New York. count, indicating that variation in septation 7. Imanaka, H., J. R. Gillis and R. A. Slepecky. 1967. time is less than variation in separation time. Synchronous growth and sporulation of Bacillus megaterium. J. Bacteriol. 93:1624-1630. By measuring the amount of DNA synthesized 1956. aspects of cell in the presence of chloramphenicol, a chromo- 8. Maruyama, Y. coli asBiochemical the method ofgrowth of studied by synchroEscherichia some replication time of 125 min has been nous culture. J. Bacteriol. 72:821-826. estimated, in agreement with the estimates of 9. Masters, M., P. L. Kuempel, and A. Pardee. 1964. Enzyme synthesis in synchronous cultures of bacteria. Ephrati-Elizur and Borenstein (5). This is subCommun. 15:38-42. Biochem. stantially longer than is found in thy+ strains 10. Oishi, M., H.Biophys. Res.and N. Sueoka. 1964. SynchroYoshikawa, (5) and is dependent on the thymine concentranous and dichotomous replications of the Bacillus tion in the medium (13). Therefore, under the subtilis chromosome during spore germination. Nature (London) 204:1069-1973. growth conditions used, the chromosome must Mathematics of and A. G. have three forks and DNA synthesis should 11. Painter, P. R.,populations. Marr. 1968.Rev. Microbiol. Annu. microbiol occur throughout the cell cycle. It must be 22:519-548. assumed that DNA synthesis proceeds linearly 12. Paulton, R. J. L. 1970. Analysis of the multiseptate potential of Bacillus subtilis. J. Bacteriol. 104:762-767. in any chromosome, and it would be expected and A. Zaritsky. that, in cells in perfect synchrony, the DNA 13. Pritchard, R. H., on the replication1970. Effect of thymine velocity of DNA in a concentration synthesis would observe sharp rate changes in thymineless mutant of Escherichia coli. Nature (Loneach cycle in the ratio 1:3:2 (6). During steadydon) 226:127-131. state labeling of synchronous cells, DNA syn- 14. Reaveley, D. A., and H. J. Rogers. 1969. Some enzymic activities and chemical properties of the mesosomes thesis is apparently exponential (proportional and cytoplasmic membranes of Bacillus licheniformis to OD). However, this method is probably not 6346. Biochem. J. 113:67-79. to detect such rate 15. Shehata, T. E., and A. G. Marr. 1970. Synchronous sufficiently sensitive growth of enteric bacteria. J. Bacteriol. 103:789-792. changes. Transformation of Continuous protein and DNA synthesis can 16. Spizizen, J. 1958. Bacillus subtilis by biochemically deficient strains of deoxyribonucleate. be regarded as evidence that no extreme stresses Proc. Nat. Acad. Sci. U.S.A. 44:1072-1078. have been incurred in selection.