Está en la página 1de 16

BIOTECHNOLOGY AND BIOENGINEERING, VOL. XVIII, PAGES 1001-1016

(1976)

Fed Batch Culture of Saccharomyces cerevisiae:

A Perspective of Computer Control to Enhance the Productivity in Baker’s Yeast Cultivation

SHUICHI AIBA, SHIRO NAGAI,* and YOSHINORI NISHI-

ZAWA, Institute of Applied Microbiology, University of Tokyo, Tokyo, Japan

Summary

A means to avoid the glucose effect in the production of baker’s yeast from glucose and/or molasses in a fed batch culture by controlling the feed rate of fresh medium with an ad hoc measurement of the respiratory quotient, RQ, is presented. The feed rate is changed stepwise here such that the value of RQ ranges from 1.0 to 1.2 throughout the cultivation. Thus far, the specific growth rate based on the total cell mass and the growth yield obtained throughout are 0.24 hr-l and 0.55 g cell/g glucose. Prior to the experimental run mentioned a,bove, equations to predetermine the feed rate and concentration of glucose in the feed are derived from the mass balance of limiting substrates (glucose). Since values of either RQ or lo, (Qo22, oxygen consumption rate with respect to the total cell mass in the fer- menter) can be measured quite easily and reliably, computer control of the fer- mentation in light of this information is discussed.

INTRODUCTION

The well-known phenomenon termed “glucose effect” cannot be prevented in the aerobic cultivation of Xaccharom yces cerevisiae in a glucose medium. This phenomenon is the so-called “aerobic fermen- tation.”’ When glucose concentration in an aerobic culture medium reaches 70 mg/liter,2 glucose tends to be partly metabolized during the fermentation to ethanol and COz; the fermentation is claimed to cease if glucose concentration in the medium is less than the specific level of 70 mg/liter. Whereas a larger than expected yield of cells in the absence of aero- bic fermentation deteriorates the specificgrowth rate, the fact that the

* Present

address:

Department

of

Fermentation

University, Hiroshima, Japan.

1001

@ 1976 by John Wiley & Sons, Inc.

Technology,

Hiroshima

1002 AIBA, NAGAI, AND NISHIZAWA

increase of the specific growth rate is accompanied by necessity by a lower yield of cells is a characteristic of the aerobic cultivation of baker's yeast. From the viewpoint of securing the most favorable productivity of an aerobic cultivation of the yeast, there should be a mediation in uncompromising phenomena between fermentation and respiration.

Maxon and Johnson3 reported earlier a diauxic

growth of the X.

cerevisiae on glucose in an aerobic and batch culture; the maximum value of the specific growth rate and growth yield reported therein was 0.41 hr-1 and 0.14 g cell/g glucose when fermentation prevailed.3 Von Meyenburg' published his work on a chemostat culture with S. cerevisiae using glucose as the limiting substrate, in which the value of the growth yield deteriorated from 0.50 to 0.145 'g cell/g glucose, depending on the dilution rates ranging from 0.24 to 0.45 hr-1. However, the data on the residual concentrations of glucose in the culture medium are not described in the original paper;' it could easily be envisaged that the residual concentration of glucose in- creased when the dilution rate was enhanced. It is considered significant at this point to minimize the glucose effectin order to produce a higher yield of cells and, in fact, empiricism has contributed to the development of the fed batch culture in the production of baker's yeast. In addition, Pirt presented some dis- cussion on the theoretical aspect of the fed batch culture,4 although a formula to define the feed rate required for negating the glucose effect has not been established. In this connection, either the ethanol concentration in exit gas from the aerobic cultivation or the respiratory quotient (RQ)value with respect to the exhaust gas may be employed as parameters in the fed batch culture. If the concentration of ethanol in the exit line be- comes detectable, the feed rate of fresh medium into the culture might be squeezed down and vice versa; indeed, this idea has already been materialized.5 However, the installation of gas chromatograph on the exit line is not always allowable from the viewpoint of economy. In this work, the RQ value which can be determined by measur- ing the partial pressures of oxygen and carbon dioxide in the exhaust gas will be used as a parameter to control and minimize the aerobic fermentation. The purpose of this work is, first, to demonstrate the controllability of this fermentation via RQ values after formulat- ing the feed rate and concentration of glucose in the feed and second, to discuss a perspective of computer control of baker's yeast produc- tion which guarantees the maximal yield and productivity.

FED BATCH CULTURE OF S. CEREVISIAE

1003

THEORETICAL CONSIDERATION

In line with the experimental setup appearing later on, suppose that a fresh feed is charged into a culture vessel at a constant rate of Fi during Ati (= ti+* - ti). The subdivision of the cultivation period of time from 1 = to to t = tp into At; (i = 0 to i = n), culminating in the increase of broth volume from V = Vo to V = Vp as shown in Figure 1, is only for the convenience of experimentation as well as the derivation of some equations which follow. Assuming a complete mixing of the medium in a well-agitated and aerated fermenter, the rate of change in concentrations of the cell mass and growth-limiting subst,rate (glucose) are :

PI,

AX;

~ At;

= pixi

1 dx;

- --

a-

X;

=pi

dt

+

=

1

~~

XiV;

Fi

Vo;+ Fi Ati

)Xi

d(X;V;)

dt

-

-

1 dX;

Xi

dt

+--

Vi

1 dV;

dt

1

Voi + F; At;

Fi

to

tl

kAto+-Atq

72

4

------

t

tl

tl+l

bat14 ---

tn

tnti

CAtn 4

1004 AIBA, NAGAI, AND NISHIZAWA

where X is the concentration of cell mass in culture medium (g/liter), J: is the total cell mass in fermenter (g), V is the broth volume (liter),

P is the feed rate of fresh medium (liter/hr), p is the specific growth

rate based on cell mass concentration (hr-l), p' is the specific growth

rate based on total cell mass (hr-l), t is the time (hr). The subscript iis the ith interval of time (see Fig. 1) and the sub- script 0 is the initial of the ith interval. The mass balance for the limiting substrate during Ato (cf. Fig. 1) yields,

Si { Vo + Fo(ti -

to))

-

Sovo = S~Po(ti- to)

(eccumuhtion)

(input)

-

1

- [Xi { vo + Fo(t1 - to) 1 - XoVoI

Yo

(consumption due to cell growth)

(3)

Rearranging eq. (3),

(Si -

So) { Vo + Fo (ti

-

to))

= Po(ti -

to)(S~- So)

provided

1

dX

Po= --=-

Xo

dt

1

Xo

(X, - XO)

(tl -

to)

Where S is the concentration of glucose in the culture medium

(g/liter), SR is the concentration of glucose in fresh feed (g/liter),

Y is the growth yield (g cell/g glucose). Modifying eq. (4) so that SI - SO= AS0 and tl - to = Ato,

AS0

--

Ate

-

Fo

Vo + Po Ato

(SR - SO)

1

--

Yo POX0 -

1

Yo Vo + Fo Po Ato

-

xo

A general expression for eq. (5) is:

(5)

Now the concentrationof glucose, SR, in a fresh feed and the feed

In quasi-steady

rate, Fi of the fresh medium are predetermined.

FED BATCH CULTURE OF S. CEREVISIAE

1005

state, where ASi/At; ‘v 0 and assuming that SR >> S;, the following equation is derived from eq. (6).

Fix;

F;SR = - Yi

1

+ - Yi (Vo;+ Fi At;)

AX.

At;

(7)

If Y;is constant throughout the fed batch culture,

therefore

FOSR(ti -

to) =

1’

-

Y

(xi -

XO)

i =

1

1, F~SR(tz - ti) = - (xZ- xl)

Y

i = n, FnSR (&+I

-

tn)

=

1

7 (Xn+1

2Fi AtiSR = - Y (Xn+l - 20)

i =O

1

-

Xn)

SR

=

Xn+1

- ZO -

XF

-

XO

Y 2 Fi At;

- y (VF - VO)

i =O

(9)

(10a)

(lob)

(11)

where XF is the final value of the total cell mass, xn+lin the fermenter (g), VF is the final value of the broth volume (liter). Equation (11) suggests that the value of SR in the fresh medium can be estimated once the target of production is established, i.e.:

XF starting from zo in the total cell mass and the broth volume in- creases, VF - Voreach expected values in the fed batch culture, and the value of Y is given, respectively. The feed rate F; for the ith time interval can also be assessed as shown below from eqs. (2) and (6) when the quasi-steady state and the specific case in which SR >> are dealt with.

MATERIALS AND METHODS

Organism and Medium Composition

The strain used in this work was baker’s yeast (8.cereerisiae) from Oriental Yeast Co., Ltd., Tokyo. The two kinds of media used,

1006 AIBA, NAGAI, AND NISHIZAWA

semisynthetic and molasses, were composed of the following. Semi- synthetic medium: glucose, 20 g; (NH2)&O, 2.15 g; NaHP04.2Hz0,

1.0 g; MgS04-7Hz0,0.38 g; KC1, 0.22 g; sodium citrate, 2.5 g; yeast extract, 0.5 g; 1 ml of vitamin solution (biotin, 0.04 g; vitamin B1, 0.08 g; vitamin B6, 2.0 g; calcium pantothenate, 1.0 g and inositol,

20 g/liter), 1 ml of mineral solution (CuS04.5H@, 0.05 g, ZnS04.7Hz0,

0.8 g; and Fe(S04)z(NH4)2-6Hz0,0.3 g/liter), tap water, 1,000 ml; and pH = 5.0 adjusted with an aqueous solution of HzS04 (2N). Molasses medium: molasses was treated with steam for 1 hr at about 80°C and centrifuged for 1 min at 4,000 x g to be free from solid ingredients and then diluted to about a 30% sugar content (as glucose). Urea was supplemented to the medium with a ratio of 0.5 g (urea) to 30 g (glucose).

Fed Batch Culture

A cell suspension obtained from the Oriental Yeast Co. was inocu- lated into a bench-scale fermenter (nominal volume = 10 liters,

initial working volume = 3.5 liters, L. E. Marubishi Co., Ltd., Tokyo, Model MD-500) in order to have 4 g/liter in the cell concen-

tration just prior to the start of the fed batch culture.

in the batch culture (initial glucose concentration = 1 g/liter), the

feeding of the fresh medium into the fermenter was begun at an adequate rate using two peristaltic pumps (A and B, Taiyo Kagaku Co., Tokyo), where the feed rate of Pump B was about 10 to 20% of that of Pump A. Pump A, as a staple control, was operated con- tinuously to feed the fresh medium at a feed rate which was changed

stepwise periodically to achieve a cell growth such that the RQ value did not fluctuate too much from 1.0. Pump B, as a fine control,

was operated intermittently ; the time of

value of RQ, i.e., this latter pump was switched on when the RQ value was below the datum of 1.0 and/or turned off when the RQ value exceeded another datum of 1.10. Usually, about 60 sec was needed to change the feed rate after confirming the varying RQ values. Obviously, using two pumps made it convenient to substitute for a single one which could vary the feed rate continuously in light of information originating from the experimentation (see Figs. 2 and 3). As was mentioned earlier, the improvised use of this pump (or pumps) necessitated the change in the feed rate stepwise rather than con- tinuously. The rotation speed of an impeller and the temperature of the culti- vation were kept at 600 rpm and 30°C, respectively. The air flow rate was controlled with a specific miniflow valve set at 2.45 and/or

After 0.5 hr

operation depended on the

FED BATCH CULTURE OF S. CEREVISIAE

1007

3.27 liters/min with respect to the cell growth; the value of pH was also controlled at 4.5 with an aqueous solution of NaOH (2N).

Analytical Methods

The optical densities measured at 610 nm in wavelength with a spectrophotometer (Hitachi Works, Model 101) were converted to dry cell mass concentration after establishing a calibration chart. The calibration was made by filtering the cells through a Millipore filter (pore size = 1.2 pm) and drying the cake at 105°C for 1 hr in a semisynthetic culture medium of glucose. In the molasses medium, the cell mass concentration was determined directly by filtering a sample broth through the Millipore filter (pore size = 1.2 pm), followed by drying at 105°C for 1 hr. The concentration of glucose in the glucose medium was deter- mined with the Glucostat reagent (Fujisawa Medical Supply Co., Ltd., Osaka), whereas the molasses was first hydrolyzed with a concentration of HC1 in a boiling water bath for 40 min and then analyzed by the Somogyi method to determine the concentration of glucose. The concentration of ethanol in the culture medium was deter- mined by the microdiffusion method. The dissolved oxygen concen- tration was measured occasionally with a membrane electrode (L. E.

Marubishi Co., Ltd.) ; actually, the oxygen concentration

was well

above the level which might have limited the cell growth. The respiration rate, Qo2,the total oxygen consumption rate, Io2, the specific rate of carbon dioxide evolution, &con,and the respiratory quotient, RQ, were estimated by calculating the difference of the partial pressures of oxygen and carbon dioxide in air between input and output through the fermenter, with a Beckman oxygen analyzer (Type 777) and an infrared gas analyzer (Shimadzu Works, Tokyo, Model URA-2), respectively. The response times of these analyzers were quite short, approximately 30 sec in preliminary experiments.

RESULTS AND DISCUSSION

Fed Batch Culture of Glucose Medium

An example of fed batch culture of glucose medium is shown in Figure 2. The total cell mass, 5, the total oxygen consumption rate, Io2, and the respiratory quotient, RQ, are on the left-hand side of the ordinate, while concentrations of glucose, S, and ethanol, P in the culture medium, the specific rate of respiration, Qo2, the specific rate

1008 AIBA, NAGAI, AND NISHIZAWA

0

U

0‘

0

I

I

I

1

2

3

t

Ihr)

‘0

4

Fig. 2. Growth patterns of baker’s yeast in the fed batch culture of the glucose medium. The differences in partial pressures of oxygen, Ape, and carbon dioxide, Apco, between input and output air were recorded continuously. RQ values as the ratio of A~co,/A~o,and lo, (Ape, times air flow rate) are then recorded continuously. Since the total cell mass, x,was determined inter- mittently, Qo, (Zo,/z) and QCO,(Qo, RQ) were both observed in discrete rather than continuous terms. The flow rate of fresh feed, F, was observed directly by using a measuring cylinder, but the data in the figure are described sche- matically. M’ (l/z.Az/At), Y’ (FiS~/zi),and Y (r’/180 Y’) values were assessed

from the curve drawn through the data points of x,respectively. (0)P, (0)Qo,, ((3)Qco,.

(0)z, (A)S,

of carbon dioxide evolution, Qco,, the specific rate of glucose consump- tion, v‘, the specific rate of increase in the total cell mass, p’, the feed rate of the fresh medium, F, and the growth yield, Y,are on the right- hand side of the ordinate. The data observed are shown in the lower diagram of the figure, whereas in the upper diagram the characteristic of the fed batch culture is demonstrated by respective calculations from the data given in the lower diagram.

FED BATCH CULTURE OF S. CEREVISIAE

1 om

Prior to the experimental run, the concentration

of glucose, SR in

the feed was calculated from eq. (11) since SR = 24 g/liter, assuming x0 = 14 g, Vo = 3.5 liters, XF = 26 g, VF = 4.5 liters, and Y = 0.5 g cell/g glucose. The actual data of SR prepared thus far was SR = 23.2 g/liter. Yeast cells inoculated into the fermenter were cultivated for 0.5 hr in batch as stated earlier, followed by the fed batch culture at t = 0

(see Fig. 2). The initial feed rate, Fo = 0.22 liters/hr at t = 0 was derived from eq. (la), assuming SR = 23.2 g/liter, 50 = 12.8 g, pa' = 0.20 hr-l, and Y = 0.5 g cell/g glucose. The feed rate which should have been periodically changed was arbitrarily manipulated stepwise such that the RQ values in situ would be within the preset

boundaries (1.0 - 1.10) (see Materials and Methods).

from the figure that RQ values, though oscillated, could be controlled within the preset range by changing the feed rate stepwise; the response of RQ to the feed-rate change was fairly rapid. Thus far, as concerns this experimental set-up (analyzer per se plus dead-space above the culture medium in the fermenter, etc.), the time required for the analyzer to respond to the change in C02 emergence from the yeast was estimated to be of the order of 60 sec. Since the time required for manipulating the pump(s) was also of the same order of magnitude as referred to earlier, the change in RQ values observed could be commensurable with that of the feed rate. In fact, when RQ > 1, the change in F entails fairly rapid emer- gence of QCO,(see Figs. 2 and 3). When RQ < 1, on the other hand, the change in F was accompanied by a short delay (see Figs. 2 and 3), implying that the yeast utilized, most probably ethanol, formed due to the lack of glucose in the culture medium. The delay was hardly relevant considering the instrument that was used.

It is clear

The fact that concentrations of glucose, S, and ethanol, P, were rather high during the early period of the fed batch culture might have been attributed to the glucose initially added (about 1,000 mg/liter) in the batch culture and the ethanol that then emerged.

However, the value of S was kept low at about 10 mg/liter throughout the rest of the fed batch culture. Therefore, from this experimental fact, the quasi-steady state assumed earlier in deriving relevant equations appears to have been warranted.

Values for the specific rate of respiration, Qo, (I~,/x)in Figure 2

remained nearly unchanged at about 7.8 mmol 02/g cell hr through- out. Accordingly, the pattern of the t.otal oxygen consumption rate, lo,most likely represented the rate of increase in the total cell mass and the oscillation of the RQ values was caused principally be Qco,.

1010 AIBA. NAGAI, AND NISHIZAWA

For the oscillation of QCO~it had also been confirmed from preliminary experiments that the construction of the infrared analyzer used here was not responislbe at all; the change in feed rate followed by the physiological activities of the yeast was considered to reflect the oscillation. Values for the specific growth rate, p', the specific glucose con- sumption rate, v', and the growth yield, Y, shown in the upper diagram of the figure were assessed here as follows: pr (1/x . Az/At) is the curve through the data points of the total cell mass, Y' is F;SR/x; (therefore, SR >> S) (cf. eq. (la)), and Y is p'/v' . 180 (on a gram basis) derived from p' and v' assessed earlier. Values of p' increased from 0.20 to 0.24 hr-' in Figure 2, while V' values were between 2.0 and 2.4 mmol glucose/g cell hr, depending on the feed rate. Accordingly, values of Y increased most likely from 0.48 to 0.55 g cell/g glucose, although they were a bit modulated. These values of pr and Y observed in this fed batch culture are nearly the same as the maximum values reported for aerobic growth (RQ = 1.0) of the S. cerevisiae in glucose-limited chemostat cultures.'

Fed Batch Culture of Molasses Medium

The experimental conditions employed in this run were as follows:

xo = 15 g, VO= 3.5 liters, XF = 35 g, and VF = 5.0 liters. If Y is

taken as 0.5 g cell/g glucose, SR is assumed to be 26.7 g/liter from

eq. (11). However, the actual data of SR prepared and g/liter and 14.1 g, respectively.

By and large, the results of this run in Figure 3 resembled that of glucose in Figure 2 except for S, which will be elaborated below. The higher concentration of residual sugar might have originated from the nonfermentative sugar which accumulated as the fed batch

culture progressed. Values of p', v',

and Y were assessed as shown in

the upper portion of Figure 3 exactly by the same procedure men- tioned earlier in Figure 2. Incidentally, p' ranged from 0.20 to 0.22 hr-l, while the growth yield, Y fluctuated around 0.48 to 0.52 g cell/g glucose.

xo were 28.4

Y,I, and RQ

It was confirmed from Figures 2 and 3 that the feed rate of the fresh medium could be controlled solely by the observed value of RQ. However, it is deemed more desirable when considering computer control to formulate the feed rate in close connection with the meta- bolic activity of baker's yeast represented by the value of RQ.

Fig. 3.

FED BATCH CULTURE OF S. CEREVISIAE

0 0

1

2

(hr)

3

1011

Growth patterns of baker’s yeast in the fed batch culture of the molasses

medium.

(0)2, (A)3, (O)P,(0)Qo,, ((3)Qco,.

With respect to the aerobic fermentation, another yield, Yp,ais defined as follows:

YPIS

=

Qp

-

V’

1012 AIBA, NAGAI, AN11 NISHIZAWA

provided that Q, is the specific rate of ethanol production (mol ethanol/g cell hr), v’ is the specific rate of glucose consumption (mol glucose/g cell hr), klis the stoichiometric constant which correlates C02evolution with ethanol production in the fermentation of glucose (1 mol ethanol/mol CO,), k2 is the stoichiometric constant which correlates 02 consumption with C02evolution in complete oxidation of glucose (1 mol C02/mol 02), lo, is the total oxygen consumption rate (Qo,a)(mol 02/hr). Equation (13) can be translated to the ith interval of time as follows :

( yp/s)

=

k~ (I0,)i (RQi -

1)

(diXi/

Y ;) (1/ 180)

Assuming again the quasi-steady state and SR >> Si, the following equation which assesses the value of Fi is derived from eqs. (12)and

(14).

(15)

Equation (15) clearly denotes that the information in Io2 and RQ stemming from the analysis of the exhaust gas from the fermenter at each time-interval defines the feed rate, if and only if the values of Yplaand SRgiven earlier from eq. (11) are made available. The relationship between Yplaand RQ calculated from the experi- mental data in Icigurcs 2 and 3 is shown in Figure 4. The linear

correlation could be assumed to be in the range of

RQ (RQ =

1.0 -

1.2) which was employed here in each fed batch culture.

Then

(16)

where K is the proportionality constant (empirical) (mol ethanol mol 02/mol glucose mol C02). The values of K given from the slope of the lines in Figure 4 were 3.SO for glucosc and 2.65 for molasses media. Substituting eq. (16) into cq. (15).

(Yp/a) i

=

K(RQi -

1)

Equation (17) means that the feed rate in order to maintain the quasi-steady state can be defined exclusively with I,,, values at that time.

Procedures to Control the Feed Rate

Two ways to control the feed rate in the f(.d batch culture were already proposed here in this work. On(. is to change the rate

FED BATCH CULTURE OF S. CEREVISIAE

"O

0.8

0.6

K= 3.80

(semi - synthetic 1

1013

Fig. 4. The relationship between Y,,, and RQ in the fed batch culture of

Y,,* was assessed

baker's yeast in semisynthetic (glucose) and molasses media

from the data of QCO,(Qo, RQ), Qo, (ZoJz), and Y' (FiS~/zi)(cf. eq. (13)); the data points do not correspond exactly to those of z in the original figures (Fig. 2 or 3), because the interpolated data (see curves drawn through the data of z) pertaining to RQ > 1 were ta.ken. (0)Semisynthetic medium; (A)mo- lasses medium.

referring to the values of RQ, the other is by affecting lo2 using eq.

(17).

The feed rate, Fc,l, estimated from lo, (eq. (17)) and another rate, F,,,, employed in the experimental runs which refer to instan- taneous values of RQ are sporadically compared in Figure 5a (semi- systematic medium) and b (molasses medium). The value of Fcal should define the feed rate once the value of RQ, approximately be- tween 1.0 and 1.15 or 1.20 (see Fig. 4) in these examples which warrant the linearity between Ypl8and (RQ - 1)) is given. In other words, control based solely on lozis essentially provided with freedom to choose an appropriate value of RQ within the specific range. Conversely, the control of the feed rate based on lo2 should be coupled with a sophisticated mechanism for controlling the feed pump, keeping the value of RQ as closely as possible around a par-

1014 AIBA, NAGAI, AND NISHIZAWA

L

z

4 -

LL

0.6-

0.4 -

I

(0 ) semi - synthetic

I

I

0

rn

Fig. 5. The feed rate, Fexpreproduced from Figs. 2 and 3 is compared with the rate, Fa!,as assessed from eq. (17): (a) semisynthetic medium, (b) molasses medium. In the assessment of Fa,, K values shown in Fig. 4 were used, respec- tively. For ease of discussion, ZZQ values are also reproduced from previous figures.

ticular value employed; otherwise, the RQ values oscillate greatly around RQ = 1.0 as illustrated in the figure. If the above situation where a precise and instantaneous control of the pump had been understood, the variations of RQ in the figure would have been minimized. In carrying out the experiments in this work, thc feed rate, which was far above the value of F,,,,was acceptable so far as the require- ment to minimize the glucose effect in terms of RQ from 1.0 - 1.15 was considered. Briefly, it goes without saying that the value of Fcz,l serves as datum to define the flow rate of the medium, leaving the exact value of RQ undefined. It is shown from Figure .5 that the upper and lower

FED BATCH CULTURE OF S. CEREVISIAE

1015

diagrams dealt with F,,, values rather than with the Fcalvalues, respectively, as a whole, even in the absence of any automatic con- trol of the feed pump where the deviation between Fcsl and F,, in both diagrams in Figure 5 seems to support the appropriate range of RQ from the viewpoint of minimizing the glucose effect. The possibility and advisability of controlling the fed batch culture using either lorand/or RQ coupled with a computer is now self-evident.

Nomenclature

feed rate of fresh medium (liters/hr)

total oxygen consumption

proportionality constant (empirical) (mol ethanol mol 02/mol glucose mol

COr) stoichiometric constant to correlate COZ evolution with ethanol produc- tion in fermentation of glucose (1 mol ethanol/mol COZ) stoichiometric constant to correlate 02 consumption with COZevolution in complete oxidation of glucose (1 mol COr/mol 02) specific rate of COZevolution (mol CO2/g cell hr) specific rate of respiration (mol 02/g cell hr) specific rate of ethanol production (mol ethanol/g cell hr) respiratory quotient (mol COn/mol 02) concentration of glucose in culture medium (g/liter) concentration of glucose in fresh feed (&liter) time (hr) broth volume (liter) final value of broth volume (liter) concentra.tion of cell mass in culture medium (g/liter) total cell mass in fermenter (9) final value of total cell mass in fermenter (g) growth yield (g cell/g glucose) yield coefficient of ethanol produced to glucose consumed (mol ethanol/ mol glucose)

rate (QOZz) (mol Ot/hr)

Greek Letters

p specific growth rate based on cell mass concentration (hr-I)

specific growth rate based on total cell mass (hr-1) specific rate of glucose consumption (mol glucose/g cell hr)

Y’

p‘

Subscript

i

ith interval of time

0

initial of ith interval

The authors are indebted to Dr. M. Ohashi, Oriental Yeast Co., Ltd. for the baker’s yeast and molasses used and to Mr. T. Yamagata, L. E. Marubishi Co., Ltd. for some instruments used throughout this work. They are also grateful to Mr. T. Karasawa, Oriental Yeast Co., Ltd. and Mr. H. Sakuma, L. E. Maru- bishi Co., Ltd. for their technical assistance.

1016 AIBA, NAGAI, AND NISHIZAWA

References

1. H. K. von Meyenburg, Arch. Mikrobiol., 66, 289 (1969).

2. F. J. Moss, P. A. D. Rickard, F. E. Bush, and P. Caiger, Biotechnol.Bioeng.,

13, 63 (1971).

3. W. I>. Maxon and M. J. Johnson, Znd. Eng. Chem., 45,2554 (1953).

4. S. J. Pirt, J. Appl. Chem. Biotechnol., 24, 415 (1974).

5. Austria Patent No. A3194-70.

6. T. Ozawa, S. Nagaoka, and K. Sumino, Hyg. Chem. (Japan), 10, 17 (1964).

Accepted for Publication February 18, 1976