YEAST VOL. 12: 673-682 (1996)

Svnchronization Affector of Autonomous Short-Period- Sbstained Oscillation of Saccharomyces cerevisiae MARC KEULERS, AIDAR D. SATROUTDINOVt, TAKA0 SUZUKI AND HIROSHI KURIYAMA*

‘Biochemi,cal Engineering Laboratory, National Institute of Bioscience and Human-Technology, Higushi 1-1, Tsukuba, Ibaraki 305, Japan

Received 25 August 1995; accepted 21 December 1995

When the yeast Succharomyces cerevisiae was grown under aerobic continuous culture, an autonomous short- period-sustained oscillation appeared. This oscillation was observed in concentrations of various extracellular and intracellular parameters, such as ethanol, acetate, glycogen, dissolved oxygen and intracellular pH. In this work the synchronization affecter of this oscillation was investigated. Ethanol was found not to be the synchronizer of the oscillation because a pulse of ethanol did not affect the phase or period of the oscillation. The oscillation was dependent on the aeration rate, i.e., the oscillation occurred only between 150 and 600 ml min- ’. However, the oxygen concentration did not influence synchronization as an upward shift in the oxygen concentration of the gas flow did not affect the sustainability of the oscillation. On the other hand, synchronization was stopped by an enhanced gas flow rate, keeping dissolved oxygen tension at the oscillatory condition, suggesting that synchronization was caused by a volatile compound in the culture. A stepwise increase in carbon dioxide concentration of the gas flow rate ceased synchronization, yet the oscillation seem to continue in each individual cell. Oscillatory behaviour of intracellular pH and carbon dioxide evolution rate showed a phase difference of 90 degrees. Based on these facts it is postulated that carbon dioxide, through the influence of its dissociation on intracellular pH, could be the synchronization affector of the autonomous short-period-sustained oscillation of S. cerevisiae.

KEY WORDS - Saccharomyces cerevisiae; metabolic oscillation; Synchronization affector; continuous culture

INTRODUCTION Analysis of synchronized cell behaviour, especially oscillatory behaviour, is very important for under- standing several biological mechanisms, ranging from proliferation to development (Lloyd et al., 1982). Cells are considered synchronized if the phase and period of oscillation are the same in individual cells. An oscillation is considered sus- tained if phase, period and amplitude are stable over a long time range, e.g. 100 times the period. An oscillation is called autonomous if no (continuous) external trigger is present to induce the oscillation. Cells remaining in phase with each other over many periods of a sustained autonomous oscillation must be synchronized by some mechanism.

Synchronization has been addressed for two different kinds of yeast-related oscillations: cell-

tPresent address: Institute of Biochemistry and Physiology of Microorganisms, Pushchino, Moscow Region, 142292, Russia. *Corresponding author.

cycle-dependent oscillation and glycolytic oscil- lation. One of the first papers on cell-cycle- dependent oscillation was from von Meyenburg (1969); afterwards an abundance of literature appeared (e.g., Parulekar et al., 1986; Porro et al., 1988; Chen and McDonald, 1990). Synchroniz- ation of cell-cycle-dependent oscillation was thought to be caused by ethanol (Martegani et al., 1990; Munch et al., 1992). The first papers on glycolytic oscillation were published by Chance et al. (1 964) and Hommes (1 964). During the early years it was postulated that the synchronization involved a diffusive glycolytic intermediate (Aldridge and Pye, 1976; Ghosh et al., 1971), but the metabolic basis of that interaction was not clear. Only recently have Aon and coworkers (Aon and Cortassa, 1991; Aon et al., 1991, 1992) postu- lated that extracellular ethanol affects the balance between oxidative and reductive fluxes of NADH, and ethanol was suggested to be the modulator of the oscillation. Later work by Richard et al. (1994)

CCC 0749-503X/96/070673-10 0 1996 by John Wiley & Sons Ltd

674 M. KEULERS ET AL.

MATERIALS AND METHODS

Strain and media The microorganism used in this study was S.

cevevisiae, diploid strain I F 0 0233, which was a wild-type strain used for distillery and bakers' yeast in Germany. The strain was kept on an agar slope at 4°C. Inocula were prepared by transferring a colony to a test tube containing 20cm3 YPG, (3 g dm-3 yeast extract, 5 g d m P 3 peptone and 20 g d m P 3 glucose). The culture was kept in a rotary shaker incubator at 30°C and 170rpm for 16 h. The experiments started with a batch culture for about 40 h. The medium composition of the batch culture was: glucose monohydrate, 22 g kg- I ; (NH,),SO,, 5 g kg- I ; KH,PO,, 2 g kg- I ;

CaC1,.2H,O, 0.1 g kg- MgS04.7H,0, 0.5 g kg - '; FeSO .7H,O, 0.02 g kg ~ I ; ZnS0,.7H20, 0.01 g kg- ; CuS04.5H20, 0.005 g kg-I; MnCl,.4H20, 0.001 g kg ~ '; 70% H2S04, 1 cm3 kg ~ I ; Antifoam agent (Adecanol LG-294; Asahidenka, Japan), 0.6 cm3 kg - I ; yeast extract (Difco), 1 g kg- I . At the end of the batch culture the culture was starved for about 4 h. Next the fermenter was operated in continuous mode using the same medium composition as with the batch culture. During the batch and the continuous culture the pH of the medium was kept at 4, using 2.5 N NaOH. If the pH was above set point no actions were taken. The fermenter (BioFlo, New Brunswick, NJ) was operated (oscillatory con- ditions) at a temperature set point of 30°C, a stirrer rate of 800 rpm, a working volume of 1.2 dm3, a gas flow rate of 180 cm3 min I , and a dilution rate of 0.08 h - ' (unless otherwise specified).

4

suggested that acetaldehyde might play a key role in the synchronization of glycolytic oscillation.

Autonomous short-period-sustained oscillation was first reported by Satroutdinov et al. (1992), who described synchronized oscillation during which many parameters changed cyclically, such as ethanol, acetate, glycogen, dissolved oxygen and intracellular pH. Besides our previous report, more detailed information about the autonomous short- period-sustained oscillation was obtained using a data acquisition system with a high sampling rate (Keulers et al., 1994). During continuous oper- ation of the fermenter no shift in period or phase was observable for more than 2 weeks, implying the sustainability of the oscillation. The oscillation was not caused by fluctuations in glucose feed; the concentration was constant at a level of 0.1 mM. It was not caused by any other external trigger and was thought to be autonomous. However, a cer- tain level of ethanol was necessary for the existence of the oscillation. Short-period oscillation was different in period, approximately 40 min, and sustainability from cell-cycle-dependent and glyco- lytic oscillation. The period of the oscillation found in cell-cycle synchronized culture has been reported to be between 2 and 16 h, depending on the dilution rate. Another difference between cell-cycle-dependent and short-period oscillation is in the role of trehalose. Trehalose is an im- portant parameter for regulation of cell-cycle- dependent oscillation (Munch et al., 1992). However, in short-period oscillation, the trehalose level does not change and has no influence on the regulation of the oscillation. The period of glyco- lytic oscillation varies from less than l min (Aon et al., 1992) to 30 min (Das et al., 1990). This oscillation was found under anaerobic conditions, in contrast to the aerobic conditions of the auton- omous short-period-sustained oscillation. Further- more, the sustainability of glycolytic oscillation is an order lower than that of short-period oscillation.

This work addresses the mechanism of synchro- nization of autonomous short-period-sustained oscillation of Saccharomyces cerevisiae. Ethanol is one of the synchronization candidates, conse- quently a series of pulse experiments was con- ducted to investigate the effect of ethanol. The effect of aeration on the synchronization of auton- omous short-period-sustained oscillation was also studied through a series of experiments and a hypothesis for a volatile synchronization affector, carbon dioxide, is proposed.

Instrumentat ion The same instrumentation was used as described

in Keulers et al. (1994).

Analytical procedures

previous paper (Satroutdinov et al., 1992). The same procedures were used as in our

Experimental procedures During low and high respiration rate, ethanol

concentration was instantaneously doubled by in- jection of 1.4 cm3 99.5% ethanol using an aseptic syringe. The oscillation was monitored for I day to investigate the effect of increased ethanol concentration on the phase, period and amplitude.

SYNCHRONIZATION OF S. CEREVISIAE OSCILLATION

To investigate the effect of aeration on the synchronization of oscillation, the gas flow was varied from 150 cm3 rnin - to 1000 cm3 min in several steps. To investigate whether dissolved oxygen tension played a role in the synchroniz- ation of oscillation, oxygen-enriched gas (30% 0,+70% N,) was supplied to the fermenter keep- ing the flow rate the same as under oscillatory conditions. The gas flow was changed at a specific moment from air to oxygen-enriched gas and after 72 h changed back to air again. The effect of the increase in inlet gas flow rate keeping dissolved oxygen tension at the oscillatory condition was investigated using gas mixtures of 0, and N,. Two different flow rates were chosen, one within the oscillatory range, 500 ml min I , and one just out- side this range, 700 ml rnin - '. The effect of carbon dioxide-enriched gas on oscillation was investigated using a mixture of air (180 ml min - ') and pure carbon dioxide gas (5.6 ml min ') resulting in a 2.5% increase in the inlet gas concentration of car- bon dioxide. The gas flow was changed at a specific moment from air to carbon dioxide-enriched gas and after 22 h changed back to air again.

675

Calculations For the calculation of intracellular concen-

trations of metabolites, the c tosolic volume was assumed to be 1.6 cm3 g (Pampulha and Loureiro-Dias, 1989). Acetate in the cell was cal- culated from the difference between the concen- trations in the culture filtrates and those in the perchlorate mixed cultures. Intracellular pH (pH,) was estimated from intracellular and cultural con- centrations of acetate by the following equation (Roos and Boron, 1981):

pHi = pK + log

(lOpH"pK+ 1) - 1) (1) I [intracellular acetate] [extracellular acetate]

Oxygen uptake rate (OUR) and carbon dioxide evolution rate (CER) were calculated from the oxygen and carbon dioxide content in the outlet gas respectively. Nitrogen was used as an inert gas to calculate the balance. The equations to calculate OUR and CER (Heinzle, 1987) are as follows, assuming a negligible amount of carbon dioxide in the inlet air:

(3)

Computations Parameter analyses of the sinusoidal oscillation

(period, phase) were performed using MatLab (Mathworks, USA). All calculations were done with an absolute numerical accuracy of 1 x 10 - '.

RESULTS

Influence of' ethanol The effect of ethanol on synchronization of

oscillation was investigated. In oscillatory con- ditions, respiratory-fermentative growth (produc- tion of ethanol) occurred during low respiration and respiratory growth on glucose and ethanol occurred during high respiration (Figure 1). Etha- nol concentration oscillated with a mean value of 32 mM and an amplitude of 1.9 mM. An ethanol pulse (25 mM) was injected at high respiration (low dissolved oxygen tension) during oscillation. Figure 2 shows the changes in ethanol concen- tration, dissolved oxygen tension and carbon di- oxide concentration in the outlet gas during the ethanol pulse. Ethanol concentration increased instantaneously; however, its oscillatory behaviour did not change. Carbon dioxide concentration at the top level increased immediately from 5.5 to 57%, and at the bottom level it decreased from 5.1 to 5% one peak after injection. Dissolved oxygen tension showed a smaller change compared to carbon dioxide. An increase in amplitude of a few per cent was observed several peaks after ethanol injection. However, it should be stressed that no changes in phase and period of the oscillation were observed after this ethanol injection. Another ethanol pulse (25 mM) was injected at low respir- ation (high dissolved oxygen tension) during oscil- lation. The changes that occurred were similar to ethanol injection at high respiration (data not shown).

lnfruence of aeration rate The influence of aeration rate on the oscillation

was investigated. During every aeration step the sustainability of the oscillation was tested and sev- eral parameters were determined (Table 1). The oscillation was sustained for aeration rates ranging from 150 to 600 ml min- but outside this range the oscillation ceased. The 150 to 600ml min- '

676 M. KEULERS ET AL.

70

60 C 0 .I z 3 50 E M x

a Ox 40 ? 2 3 30

-

20

34

33

v

32 5 g .f w

31

I I I I

0 40 80 120 160 Time (min)

Figure 1, Ethanol and dissolved oxygen tension as a function of time during oscillatory conditions. [ 01, measured ethanol (mM); solid line, dissolved oxygen tension as %I of air saturation (100%1=7.5 ppm).

44 E E v

42 0

Q

c .

.3 Y

&

30' s g 28 B w

- 0

26

I I I I I I I I 30 760 780 800 820 840 860 880

Time (min)

5.6 5 9 bD 0 u i Y

5.4 B C

c 0

.3

.3

5.2 8 8 2

5.0 0"

0 0

u

Figure 2. Response of ethanol concentration, dissolved oxygen tension and carbon dioxide concentration in the outlet gas as a function of time to a pulse of 25 mbi ethanol at high respiration during oscillatory conditions. The solid vertical line indicates the time of the injection of ethanol; [0] and dotted line. ethanol concentration (g dm-3); solid line, dissolved oxygen tension as YO of air saturation (100%=7.5 ppm); dashed line, carbon dioxide concentration in the outlet gas as 'YO of gas composition.

range had fairly constant parameters, except tration was around 8 g dm ~ and constant. Tre- dissolved oxygen tension in the culture and carbon halose level was also constant, around 6 mg g ~

dioxide content in the exhaust gas. Biomass concen- All the other parameters oscillated and average

SYNCHRONIZATION OF S. CEREVISIAE OSCILLATION 677

Table 1. Influence of aeration rate on several parameters during continuous culture of Succhuromyces cerevisiae IF0 0233, dilution rate 0.084 h - '. Oscillation occurred in the 150 to 600 ml rnin range.

Aeration rate (cm' rnin ') 150 200 400 600 1000

Biomass (g dm - ') 8.1 8.1 8.2 8.3 7.9 Ethanol (mM) 36* 25* 21* 27 * 0

Glycogen (mg g- I ) 41* 44* 35* 45* 50

DOT low (Yo) 9 22 50 70 77 DOT high (YO) 61 70 77 85 77

Trehalose (mg g ') 6.5 6 6 7 26

co, ("/I) 8.9* 6.8* 3.6* 2.2* 1.7

*Average value.

values were in the 150 to 600 ml min- range. Ethanol was around 27 mM and glycogen around 40 nig gpIcel,. No oscillation was observed at 1000 ml min - I , ethanol concentration was almost zero and trehalose concentration was high (26 mg g cell). These values differed remarkabl com- pared with those in the 150 to 600 ml min- 'range.

- 1

Influence of oxygen concentrution in the inlet gusflow The influence of a change in oxygen concen-

tration in the inlet gas flow was investigated. The

/ / / / I

- 120 '

E .a 100 * s c

c h x 0

% 80

60 e

2 5 4 0 . m

I I

12 15 18 78 81 84 87 Time (h)

Figure 3. Change in dissolved oxygen tension as a function of time. Areas 1 and 3 have a gas flow of 180 ml min- air, area 2 has a gas flow of 180 ml min - ' 30% 0, and 70Y0 N, mixture; solid line, dissolved oxygen tension as 9'0 of air saturation

inlet gas flow was changed from normal air to oxygen-enriched gas (02: 30% + N,: 70%), keeping the flow rate constant (180 ml min - '). Changes in dissolved oxygen tension from air to oxygen- enriched gas and back were very smooth (Figure 3) . No change was observable in the oscillation in terms of period or phase. This also held immedi- ately after the changes from normal air to oxygen- enriched air and back. The amplitude of the oscillation changed, but this was expected due to the change of oxygen content in the supplied gas. Dissolved oxygen tension oscillated between 80 and 130%, which was higher than dissolved oxygen tension at 1000 ml min aeration (Table l) , where oscillation did not occur. The oscillatory change of ethanol concentration in the culture was also measured and showed no difference to the oscillatory change found under oscillatory conditions (data not shown).

Influence ojpurging rate The effect of increase in the inlet gas flow rate

with dissolved oxygen tension in the range of standard oscillating conditions was investigated. At a certain time during oscillatory conditions normal inlet gas with a gas flow of 180 ml min- was changed to flow rate 1 (Table 2 and Figure 4). Oxygen content in the outlet gas decreased sharply (not shown), but the oscillation was sustained, although a period of change was noticeable. Dur- ing this higher through flow of inlet gas, no change in period or phase of the oscillation was measured, but after 1 day a change in amplitude was noticed. Yet the oscillation became stable after a day and remained stable. Changing the gas flow rate from flow rate 1 to flow rate 2 caused the oscillation to

(1 00% = 7.5 ppm). cease after less than 10 h. The oscillation ceased

678

Table 2. Settings of the different flow rates for the purging experiment.

M. KEULERS ET AL.

Flow mixture Flow rate sup p 1 i e d (cm3 min- '1 0 2 N Z CO, Ar

Normal flow 180 21% 78% 0.03'9'~ 0.93% Flow rate 1 500 15Yn 85% 0% 0% Flow rate 2 700 15% 8.5% 0% OY"

very gradually as phase and period did not change, only amplitude decreased. The mean value of dissolved oxygen tension remained approximately at the same level at the end of the oscillation (62 to 67 h) as during the end of the period with flow rate 1 (58 to 62 h).

Injluence of carbon dioxide concentration in the inlet gas flow

The influence of additional carbon dioxide in the inlet gas flow was investigated. The inlet gas flow was changed from normal air to carbon dioxide- enriched gas with a gas flow rate of approximately 180 ml min- '. Changing the gas flow caused the oscillation to cease after approximately ten periods (Figure 5). This was seen both in dissolved oxygen tension and carbon dioxide concentration in the outlet gas. The change in carbon dioxide con- centration in the inlet gas caused a phase shift in the dissolved oxygen tension of approximately 17 degrees, determined by numerical analysis. Note, however, that this phase shift is not visible in Figure 5. During 17 h (Figure 9, no stable oscil- latory pattern was recognized, except small changes in dissolved oxygen tension with an am- plitude of approximately 10% of the amplitude of the stable oscillation. After changing the gas flow back to normal air, the oscillation reappeared, becoming stable after 10-11 periods (Figure 5). The oscillatory change of ethanol concentration in the culture was also measured. Oscillatory behav- iour was as described above during normal con- ditions and a non-oscillating pattern was seen during carbon dioxide-enriched gas conditions (data not shown).

Production of carbon dioxide and internal p H during oscillatory conditions

Figure 6 shows the changes in internal pH, oxygen uptake rate and carbon dioxide evolution rate during oscillation. Changes in carbon dioxide evolution rate showed a shift of 90 degrees with

12 15 18 21 60 63 66 69

Time (h) Figure 4. Change in dissolved oxygen tension as a function of time. Area 1 has a gas flow of 180 ml min ~ ' air, area 2 has a gas flow of 500 ml min- ' 15% 0, and 85% N, mixture, and area 3 has a gas flow of 700 ml min - ' 15%) 0, and 85% N, mixture; solid line, dissolved oxygen tension as '4 of air saturation (100%=7.5 ppm).

oxygen uptake rate changes, i.e., carbon dioxide evolution rate had its maximum (highest value) 1/4 period (10 min) after oxygen uptake rate. Also, carbon dioxide evolution rate changes showed a phase difference approximately 90 degrees from changes in internal pH. Oxygen uptake rate and internal pH showed a phase difference of approxi- mately zero degrees, i.e., when internal pH was maximal, so was oxygen uptake rate.

DISCUSSION EfSect of ethanol on synchronization

Ethanol was considered to be the synchro- nization affector of both cell-cycle-dependent

SYNCHRONIZATION OF S. CEREVISIAE OSCILLATION 679

12 16 20 36 40 44

Time (h) Figure 5. Change of dissolved oxygen tension and carbon dioxide in the outlet gas as a function of time during enhanced carbon dioxide concentration in the inlet gas. Areas 1 and 3 have a gas flow of 180 ml min- ' air, area 2 has a gas flow of 180 ml min - air and 5.6 ml min ~ ' pure CO, gas; solid line, carbon dioxide concentration in the outlet gas; dashed line, dissolved oxygen tension as YO of air satura- tion (100%=7.5 ppm).

6.9

6.8 n I W

%. 3 6.7

% =:

B E:

a

U

6.6

6.5 0 40 80 120 160

Time (min) Figure 6 . Carbon dioxide evolution rate, oxygen uptake rate and intra- cellular pH (pH,) as a function of time during oscillatory conditions. A and dotted line, calculated pH, ( - ); solid line, carbon dioxide evolution rate; dashed line, oxygen uptake rate.

680 M. KEULERS ET AL.

oxygen-enriched gas mixture with the same gas flow rate, it is concluded that oxygen had no influence on synchronization (Figure 3). The oscil- lation showed no change in period or phase during a shift from air to oxygen-enriched gas or back. Ethanol concentration showed no change in its oscillatory behaviour either. Even under a high oxygen tension of 80-1 30%, which was much higher than the dissolved oxygen tension (77%) under 1OOOml min-' aeration in Table 1, no changes were noticeable. If there had been an oxygen influence, distinct changes in the phase of oscillation or ceasing of the oscillation should have occurred. It can be concluded that oxygen is not the synchronization affector of the oscillation.

Secondly, the effect of increase in inlet gas flow rate, keeping dissolved oxygen tension at osci- llatory conditions, was investigated (Figure 4). This experiment could indicate whether a volatile compound was the synchronization affector. In- creasing the flow rate should result in a ceasing of the oscillation if a volatile compound in the culture caused the synchronization. Changing the gas flow from flow rate 1 to flow rate 2 caused the oscilla- tion to cease smoothly and gradually (Figure 4). The mean value of dissolved oxygen tension re- mained at approximately the same level before and after the oscillation ceased. This change was con- sidered to be caused by ceasing of synchronization, but oscillation itself was considered to continue in each cell, suggesting that the ceasing of the synchronization was caused by stripping of a vola- tile compound. These observations suggest that a volatile compound could be the synchronization affector, but neither ethanol nor oxygen are suitable candidates.

(Martegani et al., 1990) and glycolytic oscillation (Aon et al., 1992) and was therefore the first synchronization candidate for short-period oscil- lation. However, an ethanol pulse during oscil- latory conditions, which increased ethanol concentration in the culture from approximately 2 7 m ~ to 4 3 m ~ , did not show any changes in period and phase of the oscillation (Figure 2). This result suggests that ethanol concentration could not be the synchronization affector considering the relatively small amplitude (1.9 mM) changes in ethanol concentration during oscillation. If etha- nol was the synchronization affector, a 1.9mM increase in ethanol during oscillation should stop ethanol production and force cells to enter the respiratory growth phase. Nevertheless, a 25 mM increase in ethanol through a pulse addition did not stop ethanol production and cells continued to oscillate under a rather high ethanol concentration of about 43 mM. However, the increase in ethanol concentration resulted in an immediate increase in carbon dioxide evolution rate at the top level (Figure 2), indicating increases in amplitude of the changes in carbon dioxide evolution rate. This fact suggests that ethanol concentration in this range has an influence on the ethanol assimilation rate and TCA rate. The decrease in value at the bottom level of carbon dioxide evoiution suggests a de- crease in TCA cycle turning rate, as the TCA cycle is considered to be the major carbon dioxide producer. The observation that the bottom level of carbon dioxide evolution rate decreased one peak after the ethanol injection suggests that the de- crease in TCA rate is the result of some changes in intra-mitochondria1 conditions caused by high carbon dioxide evolution rate or high ethanol assimilation rate.

Effect o j aeration, high oxygen tension and high f low rate on Oscillation

The effect of aeration on synchronization of oscillation was investigated. High aeration rate (1 000 ml min - ') stopped the oscillation (Table 1). Two conditions were changed by the increase in aeration rate: dissolved oxygen tension increased and volatile compounds present in the liquid were stripped by the higher gas flow rate. These two effects were tested separately using gas mixtures of oxygen and nitrogen.

First, the influence of a high oxygen concen- tration on synchronization of oscillation was investigated. From the result obtained with

Synchronization mechanism of oscillation Carbon dioxide in the culture is a possible

candidate for such a volatile compound. Here, we postulate a hypothesis for the synchronization of the short-period oscillation based on the effect of carbon dioxide. Regulation of relative concen- trations of carbon dioxide is via the plasma membrane. This membrane acts as a selective per- meability barrier that is impermeable to ionic and polar species (i.e., HCO-,) and allows only pas- sage of neutral or highly hydrophobic molecules (i.e,, CO,). In the oscillating culture (pH=4) car- bon dioxide exists as a non-dissociated form only (Jones and Greenfield, 1982), which can easily pass through the cell membrane. Carbon dioxide concentrations inside and outside the cell are

SYNCHRONIZATION OF S. CEREVISIAE OSCILLATION

considered to be almost in equilibrium. Conse- quently, during oscillation, intracellular and intra- mitochondrial carbon dioxide concentrations are considered to oscillate with the same size as the change of carbon dioxide in the culture. The three molecular species of carbon dioxide in the cell can be described according to:

C 0 2 + H 2 0 ~ H 2 C 0 , ~ H C 0 , - + H' (4)

An increase/decrease in carbon dioxide concen- tration would lead to an increase/decrease in HCO-, and H + (equation 4). This increase/ decrease is not instantaneous, as time constants related to the hydroxylation step in equation 4 are rather slow (Jones and Greenfield, 1982; Noorman et al., 1992). The phase difference of 90 degrees found between internal pH and carbon dioxide evolution rate (Figure 6 ) is considered to be caused by this slow time constant of the hydroxylation step. Due to higher pH levels in the cells (pH 6.5-7.0), the change in carbon dioxide could change intracellular and intra-mitochondria1 pH by carbon dioxide dissociating to HCO - and H+. Intracellular pH is known to be a regulator of metabolic rates; a lower intracellular pH leads to lower respiration (Madshus, 1988). Thus changes in carbon dioxide levels in the culture are consid- ered to regulate indirectly the respiratory activity and to force synchronization of the oscillation among cells in the culture. At a certain high gas flow rate, the magnitudes of the change in carbon dioxide level become small and it becomes difficult to synchronize the oscillation (Figure 2, Table 1). On the other hand, the results shown in Figure 5 suggest that when carbon dioxide concentration in the medium is high, synchronization is (gradually) lost due to a decrease in sensitivity of the synchro- nization affecter. Compare this with male fireflies congregated in trees that flash in synchrony only after the sun has gone down and the ambient light has decreased (Winfree, 1990). The 17 degrees phase shift occurring at the change from normal air to carbon dioxide-enriched air also suggests that carbon dioxide could be the synchronization affecter; however, the magnitude of the phase shift is rather small. Note in Figure 5 that the individual oscillation seems to continue, seen in hardly syn- chronized 10% dissolved oxygen tension amplitude changes, and synchronization is retained quite fast after switching back to the original settings (no addition of carbon dioxide in the inlet gas flow). Another indication for the probability of the con- tinuation of the individual oscillation is that the

68 1

mean of the dissolved oxygen tension during syn- chronization and during enhanced carbon dioxide flow remained the same (Figure 5). Moreover, ethanol was detected during the enhanced carbon dioxide flow (data not shown) and was not de- tected during 1000 ml min - ' aeration rate (Table l), at which setting no oscillation was detected. This suggests that individual oscillations could still be going on; however, no direct experimental evidence can be given as individual cell oscillation could not be measured directly.

Our main finding is that a volatile compound is the synchronization affector of the short-period- sustained oscillation of S. cerevisiae. Of all the synchronization candidates considered, carbon dioxide is suggested as the most likely volatile candidate for synchronization. This is based on the hypothesis that carbon dioxide will influence cell metabolism through dissociation to HCO - ,, which will affect the internal pH of the cell. The results shown in Figure 6 suggest that when intra- cellular pH is low, respiration rate is low. This observation coincides with the findings of Madshus (1988) and suggests that intracellular pH could be the modulator of the oscillation. How- ever, more detailed analysis of the oscillation mechanism is necessary. This is a subject for future research.

ACKNOWLEDGEMENTS M.K. and A.D.S. were recipients of an STA fellowship.

REFERENCES Aldridge, J. and Pye, E. K. (1976). Cell density depend-

ence of oscillatory metabolism. Nature, London 259, 670-671.

Aon, M. A. and Cortassa, S. (1991). Thermodynamic evaluation of energy metabolism in mixed substrate catabolism. Modelling studies of stationary and oscil- latory states. Biotechnol. Bioeng. 137, 197-204.

Aon, M. A,, Cortassa, S., Westerhoff, H. V., Berden, J. A., van Spronsen, E. and van Dam, K. (1991). Dy- namic regulation of yeast glycolytic oscillations by mitochondrial functions. J. Cell Sci. 99, 2325-2334.

Aon, M. A,, Cortassa, S., Westerhoff, H. V. and van Dam, K. (1992). Synchrony and mutual stimulation of yeast cells during fast glycolytic oscillations. J. Gen. Microbiol. 138, 2219-2227.

Chance, B., Hess, B. and Betz, A. (1964). DPNH oscillations in a cell-free extract of Saccharomyces carlsbergensis. Biochen?. Biophys. Res. Comm. 16, 182-187.

682 M. KEULERS ET AL.

Chen, C.-I. and McDonald, K. A. (1990). Oscillatory behavior of Saccharomyces cerevisiae in continuous culture: 11. Analysis of cell synchronisation and metabolism. Biotechnol. Bioeng. 36, 28-38.

Das, J., Timm, H., Busse, H.-G. and Degn, H. (1990). Oscillatory CO, evolution in glycolysing yeast ex- tracts. Yeast 6, 255-261.

Ghosh, A. K., Chance, B. and Pye, E. K. (1971). Metabolic coupling and synchronization of NADH oscillations in yeast cell populations. Arch. Biochem. Biophys. 145, 319-331.

Heinzle, E. ( I 987). Mass spectrometry for on-line moni- toring of biotechnical processes. Adv. Biochem. Eng. Biotechnol. 35, 1 4 5 .

Hommes, F. A. (1964). Oscillatory reductions of pyrid- ine nucleotides during anaerobic glycolysis in brewer's yeast. Arch. Biochem. Biophys. 108, 3646.

Jones, R. P. and Greenfield, P. F. (1982). Effect of carbon dioxide on yeast growth and fermentation. Enzyme Microb. Technol. 4, 210-223.

Keulers, M., Asaka, T. and Kuriyama, H. (1994). A versatile data acquisition system for physiological modelling of laboratory fermentation processes. Bio- technol. Tech. 8(12), 879-884.

Lloyd, D., Poole, R. K. and Edwards, S. W. (1982). The Cell Division Cycle, Temporal Organization and Con- trol of Cellulur Growth and Reproduction. Academic Press, London.

Madshus, I. H. (1988). Regulation of intracellular pH in eukaryotic cells. Biochem. J. 250, 1-8.

Martegani, E., Porro, D., Ranzi, B. M. and Albergina, L. (1 990). Involvement of a cell size control mechan- ism in the induction and maintenance of oscillations in continuous cultures of budding yeast. Biotechnol. Bioeng. 36, 453459.

Munch, T., Sonnleitner, B. and Fiechter, A. (1992). New insights into the synchronization mechanism with

forced synchronous cultures of Saccharomyces cerevisiae. J. Biotechnol. 24, 299-3 14.

Noorman, H. J., Luijkx, G. C. A,, Luyben, K. Ch. A. M. and Heijnen, J. J. (1992). Modeling and exper- imental validation of carbon dioxide evolution in alkalophic cultures. Biotechnol. Bioeng. 39, 1069- 1079.

Pampulha, M. E. and Loureiro-Dias, M. C. (1989). Combined effects of acetic acid, pH and ethanol on intracellular pH of fermenting yeast. Microbiol. Bio- technol. 31, 547-550.

Parulekar, S. J., Semones, G. B., Rolf, M. J., Lievense, J. C. and Lim, H. C. (1986). Induction and elimina- tion of oscillation in continuous cultures of Saccharo- myces cerevisiue. Biotechnol. Bioeng. 28, 700-710.

Porro, D., Martegani, E., Ranzi, B. M. and Alberghina. L. (1988). Oscillation in continuous cultures of bud- ding yeast: A segregated parameter analysis. Biotech- nol. Bioeng. 32, 41 1417.

Richard, P., Diderich, J. A., Bakker, B. M., Teusink, B., van Dam, K. and Westerhoff, H. V. (1994). Yeast cells with a specific cellular make-up and an environment that removes acetaldehyde are prone to sustained glycolytic oscillations. FEBS Letters 341, 223-226.

Roos, A. and Boron, W. F. (1981). Intracellular pH. Phys. Rev. 61(2), 296434.

Satroutdinov, A. D., Kuriyama, H. and Kobayashi, H. (1 992). Oscillatory metabolism of Saccharomyces cer- evisiae in continuous culture. FEMS Microbiol. Lett. 98, 261-268.

von Meyenburg, H. K. (1969). Energetics of the budding cycle of Saccharomyces cerevisiae during glucose lim- ited aerobic growth. Arch. Mikrobiol. 66, 289-303.

Winfree, A. T. (1990). The Geometry of Biological Time. Springer-Verlag, New York.