J. Cell Sci. 42, 329-356 (1980)
Printed in Great Britain © Company of Biologists Limited 1980
MAGNESIUM IONS AND THE CONTROL OF
THE CELL CYCLE IN YEAST
GRAEME M. WALKER* AND JOHN H. DUFFUS
Department of Brewing and Biological Sciences, Heriot-Watt University,
Edinburgh, EHi iHX, Scotland
SUMMARY
A study has been made of the role of magnesium ions in cell division cycle control in the
fission yeast, Schizosaccharomyces pombe, and the budding yeast, Kluyveromyces fragilis.
Synchronization of cell division in these organisms can be induced by restoring magnesium
to magnesium-exhausted cultures. In S. pombe, a correlation exists between the time taken for
cells to enter the first synchronous division and the period of magnesium exhaustion. During
short-term incubation in magnesium-deficient media, S. pombe cells are observed to continue
growth in length, but they fail to make a cell plate and divide; long-term magnesium deficiency
results in the production of aberrant cell forms, and a reduction in viability. Analysis of total
cell magnesium in cultures of both S. pombe and K. fragilis, synchronized by various induction
and selection procedures, revealed that there is a fairly steady fall in magnesium concentration
as cells grow, terminating in a rapid influx of magnesium just before cell division. This leads to
the hypothesis that falling magnesium concentration may act as a transducer of cell size, eventually triggering spindle formation and a membrane change which permits rapid uptake of
magnesium to a concentration which brings about spindle breakdown. The hypothesis was
tested directly using the divalent cation ionophore, A23187, in the absence of calcium ions;
the results obtained showed that a short pulse of A23187, very late in the cell cycle, accelerated
cells into division and shortened the subsequent cycle. The hypothesis is discussed in relation
to current models of cell cycle regulation.
INTRODUCTION
Over the past years, there has been great interest in the role of calcium ions in the
control of growth and cell division (Heilbrunn & Daugherty, 1931; Mazia, 1937;
Heilbrunn, 1952; Durham, 1974; Berridge, 1975; Mazia, 1977). Very few studies of
dividing cells, however, have been concerned with the role of magnesium ions, despite
the fact that magnesium is known to be of widespread importance in metabolic and
physiological control systems (see e.g. Vernon & Wacker, 1978). Magnesium and
calcium act antagonistically towards each other in many biological processes. This
is due to differences in the structural, thermodynamic, and kinetic parameters of
these ions (Williams, 1976). Calcium and magnesium exhibit differential binding
affinities for naturally occurring chelatable ligands. In calcium complexes, bond distances are flexible, bond angles adjustable, and co-ordination numbers variable from
6 to 10; magnesium, on the other hand, has a much more restricted range of structures, with a fixed co-ordination number of 6. Thus, calcium can displace magnesium
• Present address: Biological Institute of the Carlsberg Foundation, 16, Tagensvej, Copenhagen, N., Denmark, DK 2200.
22
CEL42
330
G. M. Walker and J. H. Duffus
from most binding sites composed of complex sets of multidentate ligands (Williams,
1974). Because of this, cross-linking of biopolymers is effectively achieved by calcium,
but magnesium may antagonize this action by binding to certain groups and making
them unavailable. Rubin (1975) has postulated that the effects observed in numerous
cell types in response to varying calcium concentrations are produced by calcium
competing with magnesium for membrane sites. Certainly, many enzymes show
variable activity depending on the magnesium/calcium ratio in their environment
(Dixon & Webb, 1964; Bygrave, 1976). More significantly for the work to be described
in this paper, calcium inhibits polymerization of tubulin in a way which is regulated
by the magnesium concentration (Rosenfeld, Zackroff & Weisenberg, 1976). Thus,
many workers have suggested that control of the calcium/magnesium ratio helps to
govern microtubule assembly in vivo (Fuller et al. 1975; Harris, 1975; Hepler, 1976;
Schliwa, 1976). If this is so, the calcium/magnesium ratio must also control spindle
formation, which depends on microtubule assembly, and hence control the onset
of cell division.
There are very few data regarding metal ion fluxes during the cell cycle. Magnesium
and calcium levels in synchronized yeast cultures were studied by Duffus & Reid
(1973) but the present paper represents the first detailed investigation of magnesium
flux. It was initiated following the discovery that agents which were capable of
altering magnesium levels in yeast cells, such as the ionophore A23187, and various
magnesium-chelating agents, were also capable of controlling progress through the
cell cycle (Duffus & Paterson, 1974a, b; Penman & Duffus, 1975; Ahluwalia, Duffus,
Paterson & Walker, 1978; Walker & Duffus, 1979). The object was to determine
whether cell magnesium concentration varied in any way which could be related to
cell division. This paper describes the results obtained, the hypothesis that has been
formed to explain these results, and an experiment with the ionophore, A23187,
which appears to support the hypothesis.
MATERIALS AND METHODS
Organisms, media and growth conditions
Schizosaccharomyces pombe NCYC132 (ATCC24751) and Kluyveromyces fragilis NCYC100
(formerly Saccharomyces fragilis (Van der Walt, 1970)) were originally obtained from the
National Collection of Yeast Cultures at Nutfield, Surrey, England. S. pombe haploid strain
972h~ was originally obtained from Dr Jorgen Friis, University of Odense, Denmark. The
growth medium used throughout is Edinburgh Minimal Medium No. 2 (EMM2) (Mitchison,
1970). Medium referred to as 'magnesium-free' or 'calcium-free' is prepared by omitting the
MgCl 2 or CaCl2 components of EMM2 and ensuring that only AnalaR grade reagents (BDH,
Poole, Dorset, England) and thrice glass-distilled water constitute the remainder of the
medium, which is autoclaved in scrupulously-clean Pyrex culture vessels. Viable stock cultures
of both yeast species were stored on EMM2/agar plates at 4 °C. Monthly subculturing was
carried out by growing cells for 3-4 days at 30 °C prior to cold storage. Weekly subculturing of
the cells was carried out by inoculating stationary phase cells into 10 ml EMM2 in McCartney
bottles. Small volumes (generally 1 0 ml) of actively dividing precultures were used to initiate
experimental cultures which were aerobically propagated in cotton-plugged Erlenmeyer flasks
in an orbital incubator operating at 160 rev/min. S. pombe was grown at 32 °C and K. fragilis
at 30 °C.
Magnesium and control of the cell cycle
331
Cell number determination
Cell numbers were determined using an improved Neubauer haemocytometer (Weber,
England). Three to four counts of 150-200 cells were made on each sample and the mean
calculated. For S. pombe, the criterion employed by Mitchison (1970) was adopted, in that
cells were not taken as having divided until the constriction or 'notch' appearing between
2 daughter cells was observed. The cell plate index is the percentage of cells of the total
population which show a cell plate, and is equivalent to the mitotic index of higher eukaryotes.
For cell counting in K. fragilis cultures, the criterion employed by Penman (1974) was
adopted in that cytokinesis is not regarded as complete until the daughter buds reach twothirds or more the size of the mother cells. To counteract cell clumping for accurate number
determination in S. pombe cultures, i-o ml samples were routinely sonicated at an amplitude of
16/tm (peak to peak) for 30 s using the exponential microprobe of an MSE 150-W ultrasonic
disintegrator.
Nuclear staining
Yeast nuclei were stained using the following Giemsa method, modified from Nurse, Thuriaux
& Nasmyth (1976). Samples (50 ml) from experimental cultures were centrifuged and the
pellet immediately frozen in an ethanol cold bath ( — 45 °C). After thawing at room temperature,
the cells were suspended in distilled water, briefly (5 s) sonicated and heat fixed onto glass
slides. The smears were washed with tapwater and postfixed in Carnoy's fixative (60 ml
ethanol, 30 ml chloroform, 10 ml acetic acid) for 30 min. After further washing, the RNA
content of the cells was digested by incubating at 30 °C with bovine pancreatic ribonuclease B
(1 mg/ml in sodium phosphate buffer, pH 68) for 3 h. The cells were washed again and
extracted for 30 min with 1 N HC1 at 60 °C. A final rinsing was followed by staining with
0-05 M solium phosphate buffer (pH 6-8) containing 4% freshly added Giemsa stain.
Establishment of synchronous cultures: induction methods
Magnesium exhaustion-synchronization. Exponentially-dividing cultures of yeast were harvested by Millipore membrane (i-2/tm pore size) filtration and washed thrice with glassdistilled water at room temperature. Washed cells were then reinoculated into prewarmed
magnesium-free EMM2 and propagated as described above. After set periods of time, sufficient
AnalaR grade MgCl2 was added back to the culture to restore the magnesium level to that in
normal EMM2 (0-3 mM).
EDTA pulse synchronization. Cell division synchrony was induced in asynchronous yeast
cultures by EDTA using a modification of the method described by Ahluwalia, Duffus,
Paterson & Walker (1978). AnalaR grade EDTA (disodium salt) was added to exponentially
growing cultures to give a final concentration of 50 mM. After i-h exposure to the chelator,
cells were harvested by Millipore membrane filtration, washed with glass-distilled water at
room temperature andfinallyreinoculated into fresh, prewarmed EMM2 (which lacked EDTA).
The time taken for this procedure varied with culture volume and cell density but generally
took about 5 min to complete. Control cultures, without EDTA treatment, were treated
identically.
Deoxyadenosine pulse synchronization. Synchrony was established in S. pombe and K. fragilis
cultures using the DNA synthesis inhibitor, 2'-deoxyadenosine (Adr) in modifications of the
methods employed by Mitchison & Creanor (1971) (S. pombe) and Penman & Duffus (1975)
(K. fragilis). Adr (Sigma Chemical Co. Ltd., London) was added to exponentially-dividing
cultures to a final concentration of 20 mM. Cells were exposed to the inhibitor until cell
division ceased, after which they were harvested by Millipore membrane filtration, washed with
glass-distilled water at room temperature andfinallyreinoculated into fresh, prewarmed EMM2.
Control cultures, without Adr treatment, were treated identically.
Studies with ^23187 (Efrapeptin). Stock solutions of the ionophore (which was a generous
gift from Dr R. L. Hammill, Eli Lilly and Co., Indianapolis, U.S.A.) were prepared by
dissolving 05 mg in 05 ml AnalaR acetone and then adding 10 ml absolute ethanol (Prince,
Rasmussen & Berridge, 1973). Small volumes of the stock solution were added directly to
exponentially dividing S. pombe cultures in Ca-free EMM2; equal volumes of acetone/ethanol
332
G. M. Walker andj. H. Duffus
mixtures were added to control cultures without effect on cell division. In pulse experiments,
cultures were exposed to the ionophore for 5 min, after which cells were rapidly harvested
by Millipore membrane filtration, washed twice with prewarmed Ca-free EMM2, and finally
reinoculated into fresh prewarmed medium. These operations took about 2 min to complete.
Establishment of synchronous cultures: selection methods
Density gradient selection synchronization. S. pombe cells were selection-synchronized using
a modification of the density gradient sedimentation procedure first described by Mitchison &
Vincent (1965). Cells from an exponentially dividing culture were filtered and concentrated to
about 5 x io8 cells/ml by resuspending in 50 ml EMM2. This suspension was carefully layered
on top of a linear lactose gradient (15-30 % w/v), prepared in EMM2 and centrifuged at
1000 rev/min in an MSE 6L centrifuge for 6 min. Using a long, thin syringe needle, about
2'0 ml of culture (containing small cells at an early cell cycle stage) were removed from the
uppermost layer of the gradient cell suspension and reinoculated into fresh EMM2 to initiate
synchrony. As far as possible, all steps were performed at 32 °C to avoid major temperature
shifts.
Continuousflowsize-selection synchronization. Synchronous cultures of both S. pombe and
K. fragilis were established using a modification of the continuous flow size-selection centrifugation method first described by Lloyd, John, Edwards & Chagala (1975). Exponentiallydividing cultures were siphoned at a set flow rate (monitored by a rotameter) into a continuousaction rotor (MSE) operating at constant revolution in the bowl of an MSE 18 centrifuge.
Overflow effluent containing small cells was retained and allowed to grow as a synchronous
culture. All operations prior to this were carried out 30 °C. Operating conditions for S. pombe
and K. fragilis are described in the legends to Figs. 11 and 12, respectively.
Analysis of total magnesium
Instrumentation, icorking conditions and sample preparation:
The total magnesium content of yeast cells has been determined using flameless (electrothermal) atomic absorption spectrophotometry (FAAS). The advantages of electrothermal
atomization over the older flame technique include: a greater intrinsic sensitivity (less than 1 ng
of element can be measured during each analytical sequence), a simplified sample preparation
(see below) and a greatly reduced sample volume (typical determinations require 5 /tl of sample,
compared to the minimum requirement of the flame system of about 500 /tl for one integrated
reading). The electrothermal carbon-rod atomization technique as employed in this study is
especially suited for the rapid analysis of the magnesium content of yeast cells since the cell
suspension is dried, ashed and atomized automatically in a single preset programme. The
instrumentation consisted of a Varian Techtron Model 1100 atomic absorption spectrophotometer, a Model 63 carbon-rod atomizer and a BC-6 Module background corrector. Absorption
readings were recorded on an Oxford 3000 series recorder. The optimal working conditions
and range of linear response for magnesium were determined and are shown in Table 1. The
data represent modifications of the guidelines suggested by Helin & Slaughter (1977) for the
determination of magnesium in brewing materials by FAAS.
Preparation of cells for magnesium analysis was undertaken by pipetting a 1 -o-ml sample of
culture into microcentrifuge tubes and immediately freezing the tubes in an ethanol cold bath
at —45 °C. Samples were thawed at room temperature and quickly centrifuged at 14 000 g in a
Quickfit microcentrifuge for 2 min, again at room temperature. The supernatant medium was
decanted and the pellet resuspended in an equal volume of glass-distilled water, using a vortex
mixer. The cells were washed a total of 4 times with glass-distilled water using the same
technique. Five-microlitre samples of final-washed cells were injected into the carbon-rod
atomizer using a platinum-tipped microsyringe. As a check on magnesium contamination from
the syringe and other laboratory hardware, several injections of glass-distilled water were
routinely carried out in order to establish a baseline.
Magnesium ion content of yeast cells is expressed as fg/cell. To obtain an accurate calculation
of this, cell numbers in the final washed cell suspension were directly determined immediately
preceding FAAS analysis. This counteracts error which may result from cell loss during the
washing procedure.
Magnesium and control of the cell cycle
333
Table 1. FA AS working conditions for Mg analysis of yeast cells
Wavelength HCL 1 H2L2
202-5 nm
5-0
5-0
Low3 High4
0-5
30
Ash5
Atomize6
A
,
*
* ,
v
volt time volt time
62
15
6-6
27
Standard'
range
002-025
1
HCL - Hollow cathode lamp current in mA.
H2L - Hydrogen continuum lamp intensity (in arbitrary units), used to check for nonatomic absorption.
3
Low - Scale expansion setting.
4
High - Curve linearization setting.
5
Ash - Ashing conditions in volts (arbitrary units) and time (seconds).
6
Atomize - Atomization conditions in volts (arbitrary units) and time (seconds)
7
The concentration (in /tg/ml) of lowest standard represents the effective sensitivity of the
method. Standard MgCl2 solutions (BDH chemicals Ltd.) were prepared by appropriate
dilution using thrice glass-distilled water from a Scorah all-glass still and the instrument was
calibrated with fresh standards each time an analysis was performed.
2
RESULTS
Studies on yeast cell division in magnesium-deficient media; magnesium exhaustionsynchronization
When washed cells of S. pombe are inoculated into a minimal medium devoid of
magnesium ions (or rather containing an irreducible minimum level of magnesium),
after a lag phase, cell division commences normally then ceases. This cessation of
cell division occurs after about 3 h and cells are permanently prevented from dividing
in this environment, even after 24-h periods of incubation (Fig. 1). One of the most
striking microscopic observations to be made during prolonged incubation of S.
pombe in magnesium-deficient media is that cells continue lengthwise growth with a
constant cell diameter. Giemsa nuclear staining reveals that these elongated cells are
mononucleate implying that in the absence of magnesium ions S. pombe cells fail to
carry out nuclear division, synthesize 1 new cell plate, and constrict their cytoplasm.
After 24 h in magnesium-free EMM2, the majority of cells appear 3 or 4 times longer
than those in control cultures grown with magnesium for similar periods (Figs. 2, 3).
The latter appear small and rounded, characteristic of stationary phase cells of S.
pombe (Mitchison, 1970). Fig. 3 shows that the nuclei in elongated cells are asymmetrically elongated, suggesting that some nuclear changes have occurred preparatory
to division, but that complete chromosome segregation has failed to occur. The
elongation process observed in magnesium-starved cultures of fission yeast does not
continue indefinitely since it has been observed that the majority of cells attain a maximum constant length after 24 h, and that they fail to elongate further after extended
incubation periods. In fact, extended periods of magnesium-deficiency result in the
production of aberrant cell forms, with some cells showing signs of hyphal-type
filamentous morphology. A 'granularization' of the cytoplasm and enlarged vacuoles
are also observed when examined with phase contrast. The long periods of magnesium
deficiency are associated with a reduction in viability, as judged by the granular
334
G. M. Walker and J. H. Dujfus
appearance of the cytoplasm, and also by a failure of cells to recover on the restoration of normal magnesium levels.
Concerning K. fragilis, incubation for a few days in Mg-free EMM2 produces
cells which characteristically form buds which fail to separate from the mother cells.
There is also a high incidence of bilateral budding (and, occasionally, multilateral
budding), but again with a failure of cell separation. Giemsa staining of magnesiumstarved K. fragilis revealed a high frequency of mononucleate cells showing incomplete nuclear migration, as with S. pombe.
7-5 r
70
6-9
6-8
6-7
|
6-6
S 6-5
3
6-4
6-3
6-2
6-1
60
5-9
5-8
1
4
1
5
24
Time, h
Fig. 1. The Mg ion dependence of cell division in S. pombe. At zero time, washed
cells of 5. pombe 132 were inoculated into EMM2 which contained: ( • ) the
normal complement of Mg (0-3 IMI), and ( • ) no added Mg (trace contaminant
level of O-I6/*M).
If magnesium ions are restored to starved cultures, cell division recommences
synchronously. This phenomenon is referred to as magnesium exhaustion synchronization, and the effect is shown for S. pombe in Fig. 4 B-H, and for K. fragilis in Fig. 5.
The addition of an excess supply of magnesium ions to a normally dividing fission
yeast culture (one containing the normal magnesium complement) has no effect of
phasing cell division, as seen in Fig. 4 A. Further, increases in cell number occurring
after transfer to Mg-deficient EMM2 (Fig. 4B, c) are still asynchronous, as can be
seen from the cell plate index and the slope of increase in cell number, neither of
which is significantly different from those in Fig. 4A. Figs. 4 B-H show that in S.
pombe the period of magnesium exhaustion prior to magnesium replenishment has the
effect of varying both the timing and degree of the resulting synchrony patterns. For
example, a proportionality exists between the time of magnesium exhaustion and the
Magnesium and control of the cell cycle
335
Fig. 2. Giemsa-stained cells of S. pombe grown in complete EMM2 for 24 h. x 1000
approx.
Fig. 3. Giemsa-stained cells of S. pombe grown in Mg-free EMM2 for 24 h. x 1000
approx.
time taken for cells to commence their first synchronous division. Generally, the
longer a cell is deprived of magnesium, the longer it takes to divide. In addition,
after the first division in cultures synchronized by relatively short periods of magnesium exhaustion, cells rapidly enter a second synchronous division. This occurs in
only a fraction of the normal cell cycle time, approximately 1-5 h compared to 2-5 h.
G. M. Walker and J. H. Duffus
336
16
7-2
12
70
6-8
6-6
0
I 16
6-8
12
6-6
°, 66-4
o-d
4
o
6-2
0
6-8
16
6-6
12
6-4
6-2
-O--O--O-
6-0
0
1
2
3
4
5
6
7
8
9
10 11
Time, h
Fig. 4. Mg ion exhaustion synchronization of cell division in 5. pombe. Washed cells
of S. pombe 132 were inoculated (at zero time) into Mg-deficient EMM2 and incubated
for various periods of time (B-H, 5, 65, 12, 14, 16, 24 and 42 h respectively), after
which the Mg level in the cultures was restored to normal by the addition of MgCl2
(indicated by the arrows), A shows the effect of adding the same concentration of
MgCl2 to a normally-dividing culture (i.e. one already containing magnesium ions).
# , log cell no. (left-hand ordinates); O, cell plate index (right-hand ordinates).
Fig. 6 illustrates this by summarizing the timing of cell plate index peaks in cultures
of S. pombe synchronized by various periods of magnesium exhaustion. The pattern
of magnesium exhaustion synchrony in K. fragilis (Fig. 5) differs from that in S.
pombe. K. fragilis is synchronized at 2 different points in the cell cycle, so that two
synchronous divisions are required before doubling of cell number is attained.
Magnesium and control of the cell cycle
Time, h
G. M. Walker and J. H. Duffus
338
Changes in cell magnesium levels during synchronous yeast growth
Regarding the preparation of yeast cells for magnesium ion analysis, the validity of
the water wash technique (see Materials and methods) was checked by measuring cell
magnesium levels in cells subjected to a series of multiple washes. Fig. 7 shows that,
after the first resuspension in glass-distilled water, cell magnesium falls, and reaches
7-5
7-4
7-3
7-2
7-1
70
6-9
6-8
6-7
6-6
o
c
6-5
g, 7-4
7-3
7-2
7-1
70
6-9
6-8
6
I
I
10
11
12
13
Time, h
14
15
16
17
Fig. 5. Mg ion exhaustion synchronization of cell division in K. fragilis. • , washed
cells of K. fragilis 100 were inoculated at zero time into Mg-deficient EMM2 and
incubated for 12-5 h, after which the Mg level in the culture was restored to normal
(indicated by arrow). • , cells inoculated into EMM2 containing the normal complement of Mg ions.
a fairly constant value, which remains at this level, irrespective of prolonged washing.
The routine technique of sample preparation for magnesium analysis involves 4
water washes. Although some intracellular magnesium may be lost during the first
wash, there can be no doubt that a substantial fraction is retained, and that this fraction
is unaffected by subsequent washing. It may be that the magnesium lost in the first
wash is the intracellular pool component, but we believe that this fraction represents
magnesium loosely retained within the cell wall.
Magnesium and control of the cell cycle
CP 1
1
339
CP,
I
CP,
1
CP,
1
CP,
6-5
CP,
1
CP,
1
CP,
1
12
CP2
1
CP,
14
CP,
16
CP,
24
CP,
I
42
1
1
Period of
Mg-exhaustion, h
1
1
1
2
3
4
Time after Mg restoration, h
1
1
Fig. 6. Schematic representation of the timing of cell plate index peaks in magnesium
exhaustion synchronized cultures of 5. pombe. CP1( CP2, and CP3 with arrows indicate
the appearance of the first, second and third cell plate index peaks, respectively.
64
60
_ 56
QJ
O
I 1 52
£71
"" 48
44
40 L l
First 1
resuspension
3
4
5
No. of washes
Fig. 7. The effect of prolonged water-washing on the Mg ion content of stationary
phase cells of S. pombe. Cells were repeatedly washed by microcentrifugation and
resuspended by vortex mixing in glass-distilled water.
Changes in cell magnesium throughout the growth cycle of S. pombe were monitored
as described, and, from Fig. 8, it is evident that during true exponential and asynchronous cell division, magnesium levels remain constant at about 10 fg per cell. Fluctuations in cell magnesium were observed during the lag phase of growth; one distinct
peak occurred at the transition between lag phase and log phase, at the point corresponding to the onset of cell division. Concerning cell cycle-dependent changes in
340
G. M. Walker and J. H. Duffus
magnesium, it is apparent that in cultures of both S. pombe and K. fragilis, synchronized by various induction and selection methods, qualitative similarities in the
patterns of change in total cell magnesium exist (Figs. 9-12 and Fig. 14). Because of
the discontinuous sampling regime, precise timing of the magnesium and cell plate
peaks and quantitation of the maxima are impossible. The apparent timing of the
peaks could vary considerably, say +10 min, from the true value. Hence, seeming
140
120
100
1
I
80
60
40
20
0
7-1
7-0
6-9
.
o
6-8
I 6-7
S
o.
-
1
6-6
6.5
6-4
6-3
6-2
6-1 0 1
2
3
4
5
6
7
8
9
10
Time, h
Fig. 8. Cell Mg levels in an asynchronously grown batch culture of S. pombe. At zero
time, washed stationary-phase cells were inoculated into fresh EMM2 and the culture
aerobically propagated at 32 °C.
discrepancies in their relationship to other cell cycle events, in Figs. 9 A, B, 10 A, and
1 IB, are readily explained. The general trend is one in which total cell magnesium
increases sharply at a time just prior to cell division, and then falls equally dramatically.
This has been schematically summarized in cell cycle maps (Figs. 15, 16). With 5.
pombe, the correlation between magnesium and cell division also extends to the timing of
the appearance of cell plate index peaks, total magnesium levels exhibiting maxima just
prior to this cell cycle stage (Fig. 15). Apart from this short-lived spike of total magnesium, the amount remains fairly constant through the cell cycle, resulting in a gradual
fall in concentration as the cell volume increases. Quantitatively, the largest values of
magnesium per cell are found in cultures to which magnesium has been restored
after relatively short periods of magnesium exhaustion (Fig. 9 A, B). Fig. 9 A shows
Magnesium and control of the cell cycle
341
that, during incubation in magnesium-deficient medium, cell magnesium remains at
a low and fairly constant level. S. pombe continues lengthwise growth in such conditions, and so again the intracellular concentration of magnesium will drop as cell
water volume, by way of an increase in cell length, increases. By the time the magnesium complement is restored to initiate synchrony, therefore, the effective magnesium
concentration in cells will be very low, or exhausted.
Some points of detail in the results shown in Figs. 9C and 11 A deserve special
mention. In Fig. 9 c there is a peak in magnesium between 25 and 26 h, which precedes
an increase in cell plate index, without any immediate increase in cell numbers. This
suggests that there are 2 magnesium-dependent processes in division of S. pombe,
the formation of the cell plate and the separation of the daughter cells, which are
distinguishable here because of the long period of magnesium exhaustion (24 h). In
the same figure, the absence of a magnesium peak before the second round of division
can be explained only on the premise that it has been missed in sampling. In Fig. H A ,
the magnesium peaks are much smaller than in Fig. 9 A, where the degree of synchrony
is similar. This must reflect a difference in cells subjected to magnesium exhaustion
(Fig. 9 A) and 2'-deoxyadenosine treatment (Fig. 11 A).
With regard to cultures synchronized by continuous flow size selection, observable
magnesium levels must be viewed in the context of proper controls. For S. pombe,
Fig. 13 A represents a control culture for this method. No significant differences in cell
magnesium patterns between the asynchronous control and the partially synchronized
culture (Fig. 13B) could be detected. However, the degree of synchrony attained was
low, and any cell cycle-related peaks would, therefore, be difficult to detect.
Control of cell division with ionophore A23187 and magnesium ions
Studies on S. pombe cell division with the ionophore A23187 were carried out in
calcium-free EMM2. This medium contained a trace contaminant level of 0-4 /HM
calcium, as detected by FAAS, and the magnesium to calcium concentration ratio was
calculated as about 750:1. A23187 can act as an ionophore for both these ions, but in
this environment it is assumed that its major effects lie in an alteration of the magnesium ion homeostasis. When cells of S. pombe were washed and inoculated into
calcium-free EMM2 the rate of cell division did not differ from that of cells grown in
complete EMM2 (Fig. 17 A). When A23187 was applied to an exponentially dividing
S. pombe culture at 0-5 /tg/ml, and in the absence of calcium, cell division was immediately inhibited. Cells were apparently arrested late in the cell cycle, since the cell
plate index increased about 4-fold after 2 h exposure to the ionophore; continued
exposure at this concentration resulted in some residual cell division (Fig. 17B).
Calcium ions added to 5 rail during ionophore arrest produced no effect (Fig. 17 c),
but magnesium ions at the same concentration reversed the ionophore-mediated
arrest, since cells recovered quickly upon magnesium addition, displaying a reduction
in cell plate index and a corresponding increase in cell numbers (Fig. 17D).
Concerning the effects of A23187 in selection-synchronized S. pombe cultures,
Fig. I8B shows that, when applied in mid cell cycle, cell division is inhibited and
about one third of the cells contain cell plates. A 5-min pulse of A23187 applied just
G. M. Walker andj. H. Duffus
342
560
360
60
50
40
30
20
10
0
18
6-8
6-7
16
6-6
14
6-5
12
6-4
10
6-3
8
6-2
6
6-1
4
O
2
60
5-9
I
I
5
6
8
9
10 11
Time, h
Fig. 9. Fluctuations of cell Mg levels in cultures of S. pombe 132 synchronized by:
6"5 (A), 14 (B), and 24 (c) h periods of Mg-exhaustion. # , log cell no.; O. cell plate
index; • , fg Mg/cell. Arrows indicate addition of MgCl2.
after the first cell plate index peak in a synchronous culture has the effect of accelerating
cells through their second synchronous division (Fig. 19B) and approximately halving
the cell cycle time.
DISCUSSION
The magnesium ion dependence of cell division in yeast
Magnesium ions are essential for cell division in yeast. This absolute requirement
has been demonstrated in the present study using magnesium-deficient media, and
in previous reports by exposing cells to magnesium-chelating agents (Ahluwalia et al.
1978; Walker & Duffus, 1979). The observation that cells of S. pombe complete about
one doubling in numbers in magnesium-free EMM2 before cell division ceases (Fig. 1)
may reflect a utilization of endogenous stores of magnesium known to occur in yeast
(Lichko & Okorobov, 1976). Cell division is inhibited by limiting the magnesium
Magnesium and control of the cell cycle
Time, h
Time, h
Fig. g, cont'd
G. M. Walker andj. H. Duffus
344
60
50
40
30
20
10
0
7-4
7-3
7-2
7-1
70
69
6-8
0
120
180 240
Time, min
300
40
30
20
10
0
7-1
70
6-9
6-3
6-7
i
6-6
60
120
180
240
j
300
Time, min
Fig. io. A. Fluctuations of cell Mg levels in a culture of S. pombe 132 synchronized
by a i-h pulse of 50 mM EDTA. Zero time marks the end of the chelating agent
pulse by re-inoculation into fresh EMM2. B. AS in A, but the culture is K. fragilis
100. Control cultures (not shown), which lacked EDTA treatment, grew asynchronously. # , log cell no; O, cell plate index; • , fg Mg/cell.
supply, but growth by way of continued protein synthesis is not inhibited, either by
chelating agents like EDTA (Ahluwalia et al. 1978) or by specific magnesiumdeficiency (Walker, 1978). DNA synthesis, however, is inhibited. Moreover, while
cells of S. pombe continue to elongate in the absence of magnesium, nuclear division
and subsequent cell plate formation is arrested (Fig. 3). This, together with other
work using A23187 reported here (Figs. 17, 18) and elsewhere (Duffus & Paterson,
19746), showing that the ionophore prevents progress through the cell cycle at a
point just before cell division, makes it likely that specific magnesium limitation in
yeast arrests cells somewhere late in the G2 phase of the cell cycle.
Magnesium and control of the cell cycle
120
180
Time, min
240
345
300
Fig. I I . A. Fluctuations of cell Mg levels in a culture of S. pombe 132 synchronized
by a S-h pulse of 2 mM 2'-deoxyadenosine. Zero time marks the end of the inhibitor
pulse by reinoculation into fresh EMM2. B. As in A, but the culture is K.fragilis 100
which was synchronized by a 2-h pulse of 2'-deoxyadenosine. Control cultures
(not shown), which lacked inhibitor treatment, grew asynchronously. # , log cell no;
O, cell plate index; • , fg Mg/cell.
23
CEL 12
346
G. M. Walker and J. H. Dujfus
The observation that cell division can be synchronized by periods of magnesium
exhaustion has not previously been reported. Synchrony is only established by
replenishment of magnesium to previously-exhausted cells, since addition of this
ion to normally-dividing cultures (with sufficient magnesium) does not result in
synchrony (Fig. 4A). The response to magnesium is also a specific one because it was
found (Walker, 1978) that cations of similar ionic configuration (Ca, Sr and Be) and
ones with similar effects on enzyme activation (Mn and Zn) all fail to substitute for
magnesium in the induction of synchrony.
120
180
Time, min
240
300
Fig. 12. Fluctuations of cell Mg levels in a culture of S. pombe 972h~ synchronized
by size-selection. A concentrated suspension of asynchronously-dividing cells was
centrifuged at low speed through a linear (15-30% w/v) lactose gradient and small
cells from the top layer were selected off and used to initiate the synchronous culture
(zero time). # , log cell no.; O, cell plate index; • , fg Mg/cell.
Magnesium ions as transducers of size control of cell division
In 5. pombe, the cell cycle maps of Fig. 6 reveal that periods of magnesium exhaustion affect subsequent synchrony patterns, with a correlation existing between
exhaustion time and the timing of the first synchronous division. This correlation also
extends to cell size; S. pombe elongates in magnesium-free EMM2 and so, generally,
the longer a cell is the longer it takes to divide after restoration of normal magnesium
levels. This observation, together with the other that intervals between first and second
synchronous divisions in magnesium-exhausted cells are shorter than normal cell
cycle times (Fig. 6), is relevant to current 'timer' and 'sizer' models of cell cycle
Magnesium and control of the cell cycle
347
control in S. pombe (reviewed by Mitchison, 1977a). It has been suggested, for example, that in S. pombe there exists a homeostatic size control over cell division
(Nurse, 1975; Fantes, 1977) where large daughters have short cycles and small
daughters have long cycles. This inverse proportionality does not hold if one considers
the present results, where the converse appears to be true; that is, the larger the cells,
the longer they take to divide. However, magnesium exhaustion must be regarded as
O
60
120
180
240
Time, min
60
120
180
240
300
Time, min
Fig. 13. A. Levels of cell Mg in an asynchronous culture of S. pombe 132. Cells were
siphoned at 600 ml/min into a continuous-action centrifuge rotor operating at 3000
rev/min. The effluent culture (45 % of the original cell density) was collected and
propagated normally. Zero time marks the completion of centrifugation. The culture
is taken as being representative of a ' control' for other cultures which had been synchronized by modifications of the same technique. B. Levels of cell Mg in a culture
of iS. pombe 132 synchronized by continuous-flow size selection. Cells from an asynchronously dividing culture were siphoned at 400 ml/min into a continuous-action
centrifuge rotor operating at 3000 rev/min. The effluent (24 % of the original cell
density) was collected and propagated as a synchronous culture after completion of
centrifugation (zero time). • , log cell no.; O, cell plate index; • , fg Mg/cell.
23-2
348
G. M. Walker andj. H. Duffus
a special case in this instance, since distortions in cell size are not brought about by
genetic lesions (as in the paper by Nurse), and it is possible that long periods of
magnesium-deficiency produce an abnormal effect through a general depression of
cell metabolism (Rubin, 1975). More significant in relation to normal division control
are probably the abbreviated intermitotic times between first and second synchronous
divisions. They represent only a fraction of the normal cell cycle time (about 60%).
The cells produced by the first division are still exceptionally large, and thus the
shortened intermitotic times conform with the observations of Fantes and Nurse cited
40
_ 30
I 20
•S1 10
120
Time, min
180
240
Fig. 14. Fluctuations of cell Mg in a culture of K. fragilis 100 synchronized by continuous-flow size selection. Cells from an asynchronously-dividing culture were
siphoned at 200 ml/min into a continuous-action centrifuge rotor operating at 6000
rev/min. The effluent (8 % of original cell density) was propagated as a synchronous
culture after completion of centrifugation (zero time). # , log cell no.; • , fg Mg/cell.
above. This has interesting corollaries in experiments with S. pombe involving metabolic inhibitors (Mitchison & Creanor, 1971), nutritional shifts (Fantes & Nurse, 1977)
and temperature shocks (Kramheft, Nissen & Zeuthen, 1976) where cell size is
also altered at the time of division. For example, Mitchison & Creanor (1971) found
that 2'-deoxyadenosine inhibition of S. pombe produced oversized cells, and postulated
that shortened intermitotic times observed following induction of synchrony with this
inhibitor were due to accumulation of 'division proteins' during the induction period,
in a manner similar to that postulated for heat-shocked Tetrahymena cells (Zeuthen,
1964). Perhaps a similar explanation applies here.
A size control model may closely correlate results from the division of S. pombe
cells altered in size either by genetic lesions (Nurse, 1975) or by nutrient depletion
Magnesium and control of the cell cycle
CP
349
CP
6-5 h Mg-exhaustion
D,
D;
CP
CP
J_^
14 h Mg-exhaustion
LJJ.
CP
-I-4-
24 li Mg-exhaustion
CP
CP
EDTA-pulse induction
D,
D;
CP
CP
Adr-pulse induction
CP
Density gradient
sizu-selection
CP
'l
60
120
180
Time, min
240
300
Fig. 15. Cell cycle maps of S. pombe from cultures synchronized by induction and
selection. Dx and D 2 indicate midpoints of synchronous divisions. Closed triangles
above the lines show the time of the major peaks in cell magnesium and ' CP' with an
arrow represents the time of appearance of cell plate index peaks. Zero time in each case
indicates the end of the treatment employed to synchronize the cells.
D,
EDTA-pulse induction
Adr-pulse induction
Continuous flow
size-selection
60
I
120
180
240
300
Time, min
Fig. 16. Cell cycle maps of K. fragilis from cultures synchronized by induction and
selection. See legend of Fig. 15 for an explanation.
G. M. Walker and J. H. Duffus
35°
70
6-8
6-6
6-4
6-2
6 0 L6-8
A23187
25
6-6
20
6-4
15
6-2
10
cf
g 60 >o-a o0 -o-o
5
u
co
ro 6-8
o
25
_i
20
6-6
15
6-4
10
6-2
^o-Coo-o-otf
5
6-0
6-8
A23187
25
20
6-6
15
6-4
10
6-2
6-0
•
3
4
Time, h
5
Fig. 17. Effect of ionophore A23187 with Ca and Mg ions during asynchronous growth
of S. pombe in Ca-free EMM2. A. Ca-free growth of S. pombe. Where indicated by the
arrow, normally-dividing cells were filtered, washed and reinoculated into Ca-free
EMM2 ( • ) . An untreated control culture ( • ) was grown throughout in complete
EMM2. B. Effect of ionophore A23187 on Ca-free growth of S. pombe. Where indicated, A23187 was added (final concentration of 0-5 /tg/ml) to an asynchronously-dividing culture in Ca-free EMM2. C. Effect of Ca on reversal of A23187 arrest. As in B,
but where indicated, Ca ions were added to the culture (final concentration 50 mni).
D. Effect of Mg on reversal of A23187 arrest. As in B, but where indicated, Mg ions
were added to the culture (final concentration 5-0 HIM). # , log cell no.; O, cell plate
index.
Magnesium and control of the cell cycle
351
(Fantes & Nurse, 1977), but it fails to indicate the underlying control mechanism.
The observed influence of magnesium on cell elongation in S. potnbe and also in
bacterial cells (Webb, 1949; Brock, 1962; Kennel & Kotoulas, 1967) may mean that
under normal steady-state conditions, changes in intracellular magnesium concentrations directly exert a size control over cell division. Further work is necessary in
order to substantiate this.
60
120
180
240
300
360
Time, min
Fig. 18. Effect of continual exposure to ionophore A23187 during synchronous growth
of S. pombe in Ca-free EMM2. A. Control culture synchronized by lactose gradient
size-selection. B. As in A, but where indicated, A23187 was added (final concentration
05 /tg/ml). • , log cell no.; O, cell plate index.
Magnesium ions are ' timers' of cell cycle progression
The qualitative similarities in the timing of magnesium influx in cultures of both
fission and budding yeasts synchronized in various ways encourages us to believe that
these are related to the fundamental cell division cycle. Quantitative differences in
cell magnesium may be explained by the fact that different methods were used to
synchronize the cells, or by the discontinuous sampling regime which would frequently
miss short-lived peak values. The necessity of establishing suitable control cultures
during cell cycle studies has recently been emphasized by Mitchison (19776), and it
is, therefore, important to note that the dramatic peaks in magnesium levels observed
in the synchronous cultures were absent from corresponding controls (for example,
Fig. 8). That they are also absent from the culture of S. pombe synchronized by
continuousflowsize-selection (Fig. 13 B), probably reflects the low degree of synchrony
attained. In successfully synchronized cultures, the relevant increases in magnesium
352
G. M. Walker andj. H. Duffus
per cell occur transiently prior to division, and in most cases are too large (i.e. more
than a doubling) to be simply due to the parallel increase in the volume of cell water
which doubles during the cell cycle as the cells grow from one division to the next.
In other words, the observable fluctuations represent genuine physiological responses
by the cells, since, even if magnesium were to be expressed on a per ml of culture
basis, or on a cell concentration basis, peaks would still occur at "the time point
indicated.
120
180
Time, min
240
300
Fig. 19. Effect of a 5-min pulse of ionophore A23187 during synchronous growth of
5. pombe in Ca-free EMM2. A. Control culture synchronized by lactose gradient
size-selection, B. AS in A, but where indicated, A23187 was added (final concentration
4'° /*g/ m l) t 0 the culture and pulsed for 5 min. Long arrow, beginning of pulse. Short
arrow represents the time when cells were reinoculated into fresh Ca-free EMM2
after filtration and washing. # , log cell no.; O, cell plate index.
From the results obtained, it may be seen that the predominant pattern of change
in magnesium concentration through the cell cycle is a fall (total cell magnesium
constant, with cell volume increasing) followed by a sudden increase which precedes
cell division, the concentration thereafter declining to a level characteristic for
dividing cells under a given set of conditions. Although these results show the
possibility of a cell cycle 'timer' based on fall in magnesium concentration, and a
close temporal relationship between magnesium uptake and cell division, they do not,
of themselves, allow us to distinguish between change in magnesium concentration as
a cause of cell cycle events and as a response to them. Therefore, in an attempt to test
the hypothesis that the fall in magnesium concentration with cell growth is the prime
basis of both size and time control, we undertook some experiments with the ionophore, A23187. The effects of A23187 are assumed to be due mainly to an equilibration of magnesium ions between the cytosol and surrounding medium; calcium effects
Magnesium and control of the cell cycle
353
are regarded as negligible in media containing very high magnesium to calcium
concentration ratios. Support for this assumption comes from observations (Fig.
17 c, D) where magnesium, but not calcium under the same conditions, rapidly reversed
A23187 arrest of cell division. In the continuing presence of A23187 in a synchronized
S. pombe culture, where cells lose magnesium in equilibrating with the medium
(Duffus & Paterson, 19746), cells accumulated at the end of the cell cycle with a fair
percentage managing to form cell plates (Fig. 18). This is what one would expect if
the end of progression through the cell cycle to cell plate formation were characterized
by a minimal magnesium concentration. On the other hand, a 5-min pulse treatment
of A23187 applied very late in the cell cycle of S. pombe (just before fission) results in
an acceleration of cell division (Fig. 19B). Although it is not unknown for S. pombe
cells to be accelerated into division (Smith &Mitchison, 1976; Kramheft, unpublished),
it is unusual that so short a pulse treatment should produce such a dramatic response.
In our opinion, what happens in this experiment is that the ionophore, applied at a
time when the intracellular magnesium concentration is high, accelerates the normal
subsequent fall in magnesium concentration and this prematurely generates the
conditions necessary for the next division. In other words, the normal relationship
between magnesium concentration and cell growth has been upset in such a way as
to shorten the cell cycle. This effect is reminiscent of the retiming of the cell cycle
observed in the synchronization of cell division using chelating agents (Ahluwalia
et al. 1978; Walker & Duffus, 1979).
CONCLUSION
The results reported in this paper and discussed above all support the previously
suggested hypothesis that magnesium concentration is the transducer for size, and
consequently time, related control of the cell cycle. The fundamental proposition is
that a fall in magnesium concentration, accompanying an increase in cell size, eventually reaches a level which permits tubulin polymerization and spindle formation.
Once chromosome separation is complete, a rapid influx of magnesium causes breakdown of the spindle, and nuclear and cell division ensue. The results quoted here show
a much greater magnesium influx than had previously been envisaged, and elucidation
of the uptake mechanism involved must be a high priority for future research. Further
work is also needed on the tubulins known to occur in yeast (Water & Kleinsmith,
1976; Baum, Thorner & Honig, 1978), since the hypothesis depends at present on the
assumption that yeast tubulin has similar properties to tubulin from sea urchins and
mammals. Finally, there is the postulated correlation of tubulin with the ' division
proteins' suggested by Zeuthen and his coworkers (Zeuthen & Ramussen, 1972). If
such a correlation can be demonstrated, it should be possible to integrate our hypothesis into the currently accepted cell cycle control models without too much difficulty.
This work was supported by grants from the Science Research Council. Part of the work
was carried out at the Biological Institute of the Carlsberg Foundation in Copenhagen, and
for this G. M. W. thanks NATO and the Royal Society for a Postdoctral Fellowship. G.M.W.
also thanks his colleagues in Copenhagen for helpful discussions, notably Professor Erik
354
G. M. Walker andj. H. Duffus
Zeuthen and Dr Birte Kramhoft. We wish to thank the Suntory Co., Osaka, Japan, for provision
of the Varian Atomic Absorption Spectrophotometer and the SRC for the Carbon Rod Atomizer. (Grant No. B/RG 5746.7).
REFERENCES
B., DUFFUS, J. H., PATERSON, L. J. & WALKER, G. M. (1978). Synchronization of
cell division in the fission yeast, Schizosaccharomyces pombe, by ethylenediaminetetra-acetic
acid. JF. gen. Microbiol. 106, 261-264.
BAUM, P., THORNER, J. & HONIG, L. (1978). Identification of tubulin from the yeast Saccharomyces cerevisiae. Proc. natn. Acad. Sci. (U.S.A.) 75, 4962-4966.
BERRIDGE, M. J. (1975). Control of cell divisions: a unifying hypothesis. J. Cyclic Nucleotide
Res. 1, 305-321.
BROCK, T. D. (1962). Effects of magnesium ion deficiency on Escherichia coli and possible
relation to the mode of action of novobiocin. J. Bad. 84, 679-682.
BYGRAVE, F. L. (1976). Mitochondrial calcium transport. In Current Topics in Bioenergetics
(ed. D. R. Sanadi), pp. 259-318. New York, London and San Francisco: Academic Press.
DIXON, M. & WEBB, E. (1964). Enzymes, 2nd edn. New York and London: Academic Press.
DUFFUS, J. H. & PATERSON, L. J. (1974a). The cell cycle in the fission yeast Schizosaccharomyces pombe: changes in activity of magnesium dependent ATP'ase and in total internal
magnesium in relation to cell division. Z. allg. Mikrobiol. 14, 727-729
DUFFUS, J. H. & PATERSON, L. J. (19746). Control of cell division in yeast using the ionophore,
A23187, with calcium and magnesium. Nature, Lond. 251, 626-627
DUFFUS, J. H. & REID, L. (1973). Changes in the cell content of free magnesium and calcium
in fission yeast during the cell cycle. Proc. 2nd Eur. Symp. Cell Cycle (Innsbruck), p. 21.
DURHAM, A. C. H. (1974). A unified theory of the control of actin and myosin in nonmuscle
movements. Cell 2, 123-136.
FANTES, P. A. (1977). Control of cell size and cycle time in Schizosaccharomyces pombe. jf. Cell.
Sci. 24, 51-67.
FANTES, P. A. & NURSE, P. (1977). Control of cell size at division in fission yeast by a growth
modulated size control over nuclear division. Expl Cell Res. 107, 377-386.
FULLER, G. M., ELLISON, J. J., MCGILL, M., SORDAHL, L. A. & BRINKLEY, B. R. (1975).
Studies on the inhibitory role of calcium in the regulation of microtubule assembly in vitro
and in vivo. In Microtubules and Microtubule Inhibitors (ed. M. Borgers & M. De Brabander),
pp. 379-390. Amsterdam and Oxford: North-Holland.
HARRIS, P. (1975). The role of membranes in the organization of the mitotic apparatus. Expl
Cell Res. 94, 409-425.
HEILBRUNN, L. V. (1952). An Outline of General Physiology, 3rd edn. Philadelphia and London:
Saunders.
HEILBRUNN, L. V. & DAUGHERTY, K. (193 I). The action of chlorides of sodium, potassium,
calcium and magnesium on the protoplasm of Amoeba dubia. Physiol. Zool. 4, 635-651.
HELIN, T. R. M. & SLAUGHTER, J. C. (1977). Determination of metals in brewing materials by
ftameless atomic absorption spectrophotometry. J. Inst. Brewing 83, 15-16.
HEPLER, P. K. (1976). Membranes in the spindle apparatus: their possible role in the control
of microtubule assembly. In Mechanisms and Control of Cell Division (ed. T. L. Rost &
E. M. Gifford), pp. 212-232. Stoudsburg (Pennsylvania): Dowden, Hutchinson and Ross.
KENNEL, D. & KOTOULAS, A. (1967). Magnesium starvation of Aerobacter aerogenes IV. Cytochemical changes. J. Bad. 93, 367-378.
KRAMHOFT, B., NISSEN, S. B. & ZEUTHEN, E. (1976). The cell cycle in heat and selection
synchronized Schizosaccharomyces pombe. Carlsberg Res. Commun. 41, 15-25.
LICHKO, L. P. & OKOROKOV, L. A. (1976). Compartmentalization of magnesium and orthophosphate ions in cells of Saccharomyces carlsbergensis. Dokl. Akad. JSfauk. SSSR 227,
756-758.
LLOYD, D., JOHN, L., EDWARDS, C. & CHAGALA, A. (1975). Synchronous cultures of microorganisms: large scale preparation by continuous flow size selection. J. gen. Microbiol. 88,
153-158.
AHLUWALIA,
Magnesium and control of the cell cycle
355
MAZIA, D. (1937). The release of calcium in Arbacia eggs on fertilization. J. cell. con:p. Physiol.
10, 291-304.
MAZIA, D. (1977). Future research on mitosis. In Mitosis Facts and Questions (ed. M. Little,
N. Paweletz, C. Petzelt, H. Ponstingl, D. Shroeter & H.P.Zimmerman), pp. 198-213.
Berlin, Heidelberg and New York: Springer-Verlag.
MITCHISON, J. M. (1970). Physiological and cytological methods for Schizosaccharomyces
pombe. Meth. Cell Physiol. 4, 131-165.
MITCHISON, J. M. (1977a). The timing of cell cycle events. In Mitosis Facts and Questions
(ed. M. Little, N. Paweletz, C. Petzelt, H. Ponstingl, D. Shroeter & H. P. Zimmerman),
pp. 1-19. Berlin, Heidelberg and New York: Springer-Verlag.
MITCHISON, J. M. (19776). Enzyme synthesis during the cell cycle. In Cell Differentiation in
Microorganisms, Plants and Animals (ed. L. Nover & K. Mothes), pp. 377-401. Jena: VEB
Gustav Fisher Verlag.
MITCHISON, J. M. & CREANOR, J. (1971). Induction synchrony in the fission yeast Schizosaccharomyces pombe. Expl Cell Res. 67, 368-374.
MITCHISON, J. M. & VINCENT, W. S. (1965). Preparation of synchronous cell cultures by
sedimentation. Nature, Lond. 205, 987-989.
NURSE, P. (1975). Genetic control of cell size at cell division in yeast. Nature, Lond. 256, 54755iNURSE, P., THURIAUX, P. & NASMYTH, K. (1976). Genetic control of the cell division cycle in
the fission yeast Schizosaccharomyces pombe. Molec. gen. Genet. 146, 167-178.
PENMAN, C. S. (1974). Some Aspects of the Cell Cycle in the Budding Yeast, Klnyveromyces
fragilis. Ph.D. Thesis, Heriot-Watt University, Edinburgh.
PENMAN, C. S. & DUFFUS, J. H. (1975). 2'-Deoxyadenosine and A23187 as agents for inducing
synchrony in the budding yeast, Kluyveromyces fragilis. J. gen. Microbiol. 90, 76-80.
PRINCE, W. T., RASMUSSEN, H. & BERRIDGE, M. J. (1973). The role of calcium in fly salivary
gland secretion analysed with the ionophore A23187. Biochim. biophys. Ada 329, 98-107.
ROSENFELD, A., ZACKROFF, R. V. & WEISENBERG, R. C. (1976). Magnesium stimulation of
calcium binding to tubulin and calcium induced depolymerization of microtubules. FEBS
Letters, Amsterdam 65, 144-147.
RUBIN, H. (1975). Central role for magnesium in coordinate control of metabolism and growth
in animal cells. Proc. natn. Acad. Sci. U.S.A. 72, 3551-3555.
SCHLIWA, M. (1976). The role of divalent cations in the regulation of microtubule assembly.
In vivo studies on microtubules of the Heliozoan Axopodium using the ionophore A23187.
J. Cell Biol. 70, 527-54OSMITH, H. T . B. & MITCHISON, J. M. (1976). Anaesthetics delay and accelerate division in the
fission yeast, Schizosaccharomyces pombe. Expl Cell Res. 99, 432-434.
VAN DER WALT, J. P. (1970). Kluyveromyces van der Walt emend, van der Walt. In The Yeasts,
A Taxonomic Study (ed. J. Lodder), pp. 316-378. Amsterdam and London: North-Holland.
VERNON, W. B. & WACKER, W. E. C. (1978). Magnesium metabolism. In Recent Advances in
Clinical Biochemistry (ed. K. G. M. M. Alberti), pp. 39-71. Edinburgh, London and New
York: Churchill Livingstone.
WALKER, G. M. (1978). Metal Ions and the Control of the Cell Cycle in Fission and Budding
Yeasts. Ph.D. Thesis, Heriot-Watt University, Edinburgh.
WALKER, G. M. & DUFFUS, J. H. (1979). An investigation into the potential use of chelating
agents and antibiotics as synchrony inducers in the fission yeast, Schizosaccharomyces pombe.
J. gen. Microbiol. 114, 391-400.
WATER, R. D. & KLEINSMITH, L. J. (1976). Identification of a and/? tubulin in yeast. Biochem.
biophys. Res. Commun. 70, 704-708.
WEBB, M. (1949). The influence of magnesium on cell division. I. The effect of magnesium in
the growth and cell division of various bacterial species in complex media. J.gen. Microbiol. 3,
410-417.
WILLIAMS, R. J. P. (1974). Calcium ions: their ligands and their functions. Biochem. Soc.
Symp. 39, 133-138.
WILLIAMS, R. J. P. (1976). Calcium chemistry and its relation to biological function. In
Calcium in Biological Systems (ed. C. J. Duncan), pp. 1-17. Cambridge University Press.
356
G. M. Walker andj. H. Duffus
E. (1964). The temperature induced division synchrony in Tetrahymena. In Synchrony in Cell Division and Growth (ed. E. Zeuthen), pp. 99-158. New York, London and
Sydney: Interscience.
ZEUTHEN, E. & RASMUSSEN, L. (1972). Synchronized cell division in protozoa. In Research in
Protozoology, vol. 4 (ed. R. T. Chen), pp. 11-145. Oxford: Pergamon.
ZEUTHEN,
{Received 3 August 1979 -Revised
15 October 1979)