Journal of Applied Microbiology 2003, 95, 301–309

Optimization of growth conditions of the postharvest

biocontrol agent Candida sake CPA-1 in a lab-scale fermenter

M. Abadias, N. Teixidó, J. Usall and I. Viñas

Postharvest Area, CeRTA, Centre UdL-IRTA, 177 Rovira Roure, 25198 Lleida, Catalonia, Spain

2002/368: received 9 October 2002, revised 6 March 2003 and accepted 19 March 2003

ABSTRACT
M . A B A D I A S , N . T E I X I D Ó , J . U S A L L A N D I . V I Ñ A S . 2003.
Aim: To maximize the growth (expressed as number of viable cells per millilitre) of the postharvest biocontrol
agent Candida sake CPA-1 at laboratory scale conditions.
Methods and Results: Growth conditons (aeration, agitation speed and inoculum size) were studied in batch
conditions in a 5 l fermenter using molasses and urea as growth medium. Consumption of sugars and urea were
analysed. Fed-batch studies were also carried out. Glucose and fructose were consumed during the exponential
growth phase and were depleted after 18 h of growth. On the contrary, C. sake cells assimilated sucrose during the
stationary phase. There was not growth improvement when fed-batch technology was used. Addition of an extra
amount of glucose or molasses after 18 h of growth did not contribute to increase final population.
Conclusions: Maximum growth (about 8 · 108 CFU ml)1) was obtained at batch fermentation after 30 h growth
at 400 rev min)1, 150 l h)1 of air and initial concentration of 106 CFU ml)1.
Significance and Impact of the Study: The results obtained in this study are an approach for further upscaling
of C. sake production.

Keywords: aeration, batch, fed-batch, Gompertz kinetics parameters, inoculum size, molasses, stirring speed.

commercial use in the USA and Israel. YieldPlus contains

INTRODUCTION
Biological control of postharvest diseases has been exten-
sively studied in recent years (Pusey and Wilson 1984;
Janisiewicz 1987; Roberts 1990; Janisiewicz and Marchi
1992; Filonow et al. 1996; Viñas et al. 1998; Nunes et al.
2001). Concerns regarding human health and environmental
risks associated with chemical residues in foods and in the
environment have been the main driving force of the search
of new and safer control methods (Droby et al. 1998). Many
microbial antagonists of postharvest pathogens have been
studied and are effective biocontrol agents. However, few
have been commercialized. At present, only two yeasts and
one bacteria have been developed commercially to control
postharvest diseases. Decco I-182 (formerly named Aspire)
contains the yeast Candida oleophila and is registered for
Correspondence to: M. Abadias, Postharvest Area, CeRTA, Centre UdL-IRTA, 177
Rovira Roure, 25198 Lleida, Catalonia, Spain (e-mail: isabel.abadias@irta.es).
ª 2003 The Society for Applied Microbiology
the yeast Cryptococcus albidus and is registered in South
Africa. BioSave 100 and 110 are frozen formulations of the
bacterium Pseudomonas syringae ESC10 and 11 and are
registered in USA.
In Europe, detailed studies have shown that the strain
CPA-1 of Candida sake is an effective antagonist to the major
postharvest pathogens on pome fruits at laboratory, pilot
(Teixidó et al. 1998; Viñas et al. 1998; Usall et al. 2000;
Nunes et al. 2002) and commercial scale applied at
107 CFU ml)1 (Usall et al. 2001). Moreover, the strain
CPA-2 of Pantoea agglomerans has been demonstrated to be
an effective antagonist to the major postharvest pathogens
on citrus and pome fruits (Viñas et al. 1999; Costa et al.
2001, 2002; Nunes et al. 2001; Teixidó et al. 2001).
Mass production of yeast cells is an essential step in the
commercialization of a biocontrol agent. A key factor to
consider is the development of a suitable low-cost medium
for cell production. A low-cost medium, which could be
 302 M . A B A D I A S ET AL.
used at industrial scale, has been developed for C. sake in
our laboratory. This medium is based on cane molasses, a
by-product from the sugar industry (Abadias et al. 2000,
2001). However, efforts must be made to produce C. sake
in a laboratory scale fermenter to provide relevant infor-
mation for further scale-up production. Operating condi-
tions (aeration, agitation, pH and temperature) as well as
medium constituents may affect the quality and quantity of
the tested microorganisms (Churchill 1982). The purpose
of aeration is to provide micro-organisms in submerged
culture with sufficient oxygen for metabolic requirements,
while agitation should ensure that a uniform suspension of
microbial cells is achieved in a homogeneous nutrient
medium (Stanbury et al. 1995). The inoculum is an
important factor for an optimal production of a microor-
ganism. It must be in a healthy, active state thus
minimizing the length of the lag phase in the subsequent
fermentation, available in sufficiently large volumes to
provide an inoculum of optimum size, free of contamin-
ation and must retain its product-forming capabilities.
Some studies have indicated that high cell density
fermentations can be achieved in Saccharomyces cerevisiae
using fed-batch technology and molasses as the growth
substrate (Beudeker et al. 1989; Reed and Nagodawithana
1991; Walker 1998).
Thus, it is the aim of this paper to find the optimal
conditions (aeration, agitation and initial inoculum) for the
maximum production of C. sake in a lab-scale fermenter.
Moreover the effect of using fed-batch technology will also
be studied.
M A T E R I A LS A N D M E T H O D S
Microorganism and medium
The biocontrol yeast C. sake CPA-1 obtained from UdL-
IRTA (Lleida, Catalonia, Spain) was used in this study. It
was isolated from the apple surface and it is deposited in
Colección Española de Cultivos Tipo, CECT-10817 (Uni-
versidad de Valencia, Campus de Burjasot, Burjasot,
Valencia, Spain). Stock cultures were stored at 4C and
were subcultured on nutrient yeast dextrose agar, NYDA
(nutrient broth, 8 g l)1; yeast extract, 5 g l)1; dextrose,
10 g l)1 and agar, 15 g l)1) at 25C.
The growth medium of all batch fermenter experiments
was composed of 40 g l)1 cane molasses and 1Æ2 g l)1 urea
and 0Æ1 ml antifoam (DF7960, Chemitrade, S.A.B. Braun
1 Biotech, Melsungen, Germany) per litre of medium.
Inoculum preparation
A 48 h preculture of C. sake was transferred from NYDA-
plates to 5 ml potassium phosphate buffer (PB; pH 6Æ5).
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
Fifty millilitres of molasses medium in 250 ml conical flasks
were inoculated with 0Æ5 ml of this freshly prepared C. sake
solution. After incubation at 25C for 24 h on a rotatory
shaker at 150 rev min)1, cells were used as starter in the
fermentation experiments.
Fermentation conditions
All batch and fed-batch fermentations were performed in a
BIOSTAT-A modular fermenter (Braun Biotech Interna-
tional, Melsungen, Germany) with 5 l working volume.
Each fermenter (except those in which the effect of
inoculum was studied) was inoculated with 10 ml of a
starter prepared as described above. Temperature of all
experiments was set at 25C, and controlled within 0Æ1C
of the set point. Initial pH of molasses medium was in the
range of 6Æ5–7Æ0 and was not controlled during the process.
In all experiments and during cultivation, pH, temperature
and dissolved oxygen, pO2 were constantly registered with
the micro DCU-300 unit (Braun Biotech International).
When necessary, antifoam was added to the culture by the
controller unit.
Effect of aeration and agitation rates
In order to examine the influence of agitation and aeration
rates on growth, C. sake was grown in the fermenter as
described previously, setting the aeration at 50 l h)1
[0Æ17 vvm (air volume/liquid volume per minute)] or
150 l h)1 (0Æ50 vvm) and the agitation at 200, 400 or
600 rev min)1. A subsample of the culture (10 ml) was
taken regularly in order to determine the number of viable
cells (CFU) per millilitre.
Effect of initial inoculum
To study the effect of initial inoculum concentration on
C. sake growth, the Biostat A fermenter set at 25C,
400 rev min)1 and 150 l h)1 was inoculated with 1, 10 or
50 ml of preculture (0Æ02, 0Æ2 or 1% of inoculum, respect-
ively). The number of viable cells was determined from a
10 ml subsample of the culture at the beginning of the
experiment and every 2 h from 18 to 30 h of growth.
Fed-batch experiments
In the first fed-batch experiment, 4Æ7 l of growth medium
were inoculated with 10 ml of 24 h starter culture. After 8,
16 and 24 h of the inoculation, fermenter was fed with
batches of 100 ml of a cane molasses : water solution
(50 : 50), which gave a final concentration of 40 g l)1 of
cane molasses. Fermenter aeration was set at 150 l h)1 and
400 rev min)1 and was inoculated with 10 ml of the

preculture as described above. The number of viable cells
was determined from a 10 ml subsample of the culture at the
beginning of the experiment and every 2 h until 48 h of
growth.
Two more experiments were carried out using residual
glucose/fructose content as a signal to feed substrate to
the culture. In one of them, after depletion of glucose and
fructose, concentrated synthetic glucose medium (250 g l)1
of glucose) was added to the fermentation broth by a
peristaltic pump at a flow rate of 4Æ5 ml min)1, to bring
the concentration in the bioreactor to 5 g l)1. On the other
hand, when glucose/fructose present in the molasses
medium had been consumed, an additional 200 g of cane
molasses were supplied in order to bring the concentration
of glucose and fructose to the same as the initial one. A
10 ml subsample was taken from the culture every 2 h
from 18 h of growth to 28 h and then the number of
viable cells was determined. Glucose, fructose, sucrose and
urea content was also analysed at the same time of viable
cells.
Analytical procedures
Candida sake growth was followed from a 10 ml subsample.
Serial dilutions were made with 0Æ05 M PB (pH, 6Æ5) and
plated on NYDA, obtaining the number of CFU per
millilitre of medium. When sucrose, glucose, fructose and
urea were analysed, an extra 10 ml subsample from the
culture was taken. Sucrose, D-fructose, D-glucose and urea
were analysed enzymatically using Boehringer Mannheim
Test combination kit (catalogue number 716260 for sucrose,
glucose and fructose and 542946 for urea). Before the
enzymatic analysis, samples were placed in a waterbath at
80C for 15 min to stop enzymatic reactions. Then samples
were centrifuged (4000 · g, 10 min) and the supernatant
was used for the analysis.
Modelling of microbial growth
One of the most used and recommended model for microbial
growth is the modified Gompertz equation whose expression
is: log N ¼ a + c exp()exp()b(t ) m))), where log N: deci-
mal logarithm of microbial counts [log(CFU ml)1)] at time
t; a: log No asymptotic log count as time decreases
indefinitely [log(CFU ml)1)]; c: number of log cycles of
growth [log(CFU ml)1)], m: time required to reach the
maximum growth rate [h]; b: relative growth rate at time
m [h)1].
From these parameters, the exponential growth rate
(l ¼ bc/e) [log(CFU ml)1) h1], with e ¼ 2Æ7182), lag
phase duration (LPD ¼ m ) (1/b) + (log No ) a)/(bc/
e)[h]) and maximum population density (MPD ¼ a + c
[log(CFU ml)1)]) were derived. The equation was fitted to
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
OPTIMIZATION OF CANDIDA SAKE PRODUCTION 303
growth data using non-linear regression of the Statgraphics
2 Plus 4Æ1 software (Statistical Graphics Corp.).
RESULTS
Effect of aeration and agitation speed
on C. sake growth
When C. sake cells were inoculated into molasses medium
and incubated under different aeration and agitation condi-
tions in a fermenter, a typical batch growth curve resulted
when the viable population of cells was plotted against time
(Fig. 1). Gompertz parameters are shown in Table 1. The
LPD was between 4Æ3 and 8Æ9 h and l was between 0Æ17 and
0Æ26 log(CFU ml)1)h)1, being the maximum growth rate
when agitation and aeration were set at 400 rev min)1 and
150 l h)1, respectively. At these conditions the maximum
population growth was achieved (8Æ94 log CFU ml)1)
after 30 h of growth. Similar countings were found when
the fermenter was set at 400 rev min)1 (50 l h)1) or
600 rev min)1 (150 l h)1) (8Æ82 and 8Æ89 log CFU ml)1,
respectively). However, time to reach this population was
longer (36–40 h).
During the exponential growth phase, pO2 value
decreased up to its minimum value depending on aeration
and agitation conditions. As aeration and agitation speed
decreased, the minimum pO2 value and time to reach this
minimum decreased. For example, at 50 l h)1, when
agitation speed was 600 rev min)1, minimum pO2 value
was 80%, it was 20 and 0% when agitation was 400 and
200 rev min)1, respectively.
Regardless of growth conditions, the pattern of pH was
similar: it decreased during exponential phase, reached its
minimum and then increased again. Minimum pH values
were between 5Æ40 and 5Æ93.
Effect of initial inoculum
Initial C. sake population in the fermenter in different
experiments was 1Æ5 · 105, 9Æ0 · 105 and 5Æ8 ·
106 CFU ml)1 when the initial inoculum was 1, 10 or
50 ml, respectively (Fig. 2). An initial inoculum of 1 ml
gave the lowest C. sake growth, obtaining the maximum
after 30 h of culturing (3Æ7 · 108 CFU ml)1), meanwhile
similar counts were obtained when the inoculum used was
10 or 50 ml, with the maximum growth obtained after 26–
30 h of culturing (6Æ0–7Æ4 · 108 CFU ml)1). Thus, we
decided to continue the experiments with an initial inocu-
lum of, approx. 106 CFU ml)1, which corresponded to
10 ml of the 24 h preculture.
Depending on the inoculum used, marked differences on
pO2 and pH patterns were observed. Although maximum
and minimum pH and pO2 values were similar, the required
304 M . A B A D I A S ET AL.

(a) 200 rpm 50 l h–1 (b) 200 rpm 150 l h–1

9·0 100·0
r 2 = 0·9913 r 2 = 0·9942

8·0 75·0
log10 CFU ml–1

pO2 (%)
pH

7·0 50·0

6·0 25·0

5·0 0· 0

(c) 400 rpm 50 l h–1

(d) 400 rpm 150 l h–1
9·0 100·0
r2= 0·9851 r 2 = 0·9958

8·0 75·0
log10 CFU ml–1

pO2 (%)
pH

7·0 50·0

6·0 25·0

5·0 0.0

(e) 600 rpm 50 l h–1 (f) 600 rpm 150 l h–1

9· 0 100·0

r 2 = 0·9939 r 2 = 0·9849
8· 0 75·0
log10 CFU ml–1

pO2 (%)
pH

7· 0 50·0

6· 0 25·0

5· 0 0· 0
0 4 8 12 16 20 24 28 32 36 40 0 4 8 12 16 20 24 28 32 36 40
Time (h) Time (h)

Fig. 1 Effect of aeration (50 or 150 l h)1) and stirring speed (200, 400 or 600 rev min)1) on Candida sake growth (n), pH (h) and pO2 (s).
A 5 l fermenter with molasses-based medium was inoculated with 10 ml of starter culture and was set at 25C. Growth curves were fitted using
the Gompertz modified equation. R-square coefficient is shown

time to reach these values were markedly different. Mini- when the initial inoculum was 1, 10 or 50 ml, respectively. It
mum pH value was obtained at 24, 18 or 12 h of culture and seemed that when lower initial inoculum was used, the
minimum pO2 value was obtained after 28, 20 and 14 h process was delayed.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
 OPTIMIZATION OF CANDIDA SAKE PRODUCTION 305

Table 1 Gompertz parameters (a, b, c, m), derived parameters (exponential growth rate, l; lag phase duration, LPD; maximum population density,
MPD, of Candida sake grown in a 5 l fermenter at different agitation and aeration conditions, corresponding to the fitted Gompertz equation
log N ¼ a + c exp()exp()b(t ) m))). Fitted curves and R-square values are represented in Fig. 1

a c b m l LPD MPD
Growth conditions (log CFU ml)1) (log CFU ml)1) (h)1) (h) (log(CFU ml)1)h)1) (h) (log CFU ml)1)

200 rev min)1 (50 l h)1) 5Æ66 2Æ97 0Æ22 8Æ15 0Æ24 4Æ54 8Æ64
200 rev min)1 (150 l h)1) 5Æ44 3Æ22 0Æ16 10Æ55 0Æ19 4Æ39 8Æ66
400 rev min)1 (50 l h)1) 5Æ87 2Æ95 0Æ21 8Æ74 0Æ23 4Æ29 8Æ82
400 rev min)1 (150 l h)1) 5Æ99 2Æ96 0Æ23 11Æ08 0Æ26 6Æ71 8Æ94
600 rev min)1 (50 l h)1) 6Æ01 2Æ68 0Æ20 9Æ15 0Æ19 4Æ69 8Æ69
600 rev min)1 (150 l h)1) 5Æ90 2Æ99 0Æ15 14Æ88 0Æ17 8Æ90 8Æ89

9·0 100·0

8·0 75·0
log10 CFU ml–1

pO2 (%)
pH

7·0 50·0

Fig. 2 Effect of initial inoculum on pH value

(circles), pO2 value (squares) and growth
(triangles) of Candida sake in a 5-l fermenter 6·0 25·0
set at 400 rev min)1, 150 l h)1 and 25C.
Fermenter was inoculated with 1 ml of starter
(white symbols), 10 ml of starter
(grey symbols) or 50 ml of inoculum 5·0 0·0
(black symbols) grown at 150 rev min)1 for 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30
24 h at 25C Time (h)

sake cells did not assimilate urea until 12 h cultivation. At

Consumption of glucose, fructose, sucrose
and urea by C. sake in batch conditions
Concentration of glucose, fructose, sucrose and urea during
cultivation at 400 rev min)1 and 150 l h)1 at 25C were
determined. The fermenter was inoculated with 10 ml of a
24 h preculture. Initial concentration of glucose, fructose
and sucrose was 1Æ4, 2Æ6 and 12Æ0 g l)1, respectively (Fig. 3).
Glucose and fructose were assimilated before sucrose.
Maximal amount of glucose and fructose was consumed
during the exponential growth phase of C. sake cells. After
18–20 h there was neither residual glucose nor fructose. At
this moment C. sake cells begun to assimilate sucrose
gradually, which was assimilated basically during the
stationary growth phase of C. sake cells. After 28 h
cultivation, 3Æ8 g l)1 residual sucrose was present in the
medium.
Some urea may degrade during the sterilization process,
as after sterilization, 0Æ85 g l)1 of urea were detected in the
medium instead of 1Æ2 g l)1 that had been added. Candida
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
the end of the analysed period, there was 0Æ03 g l)1 of urea
in the medium. Thus, urea was not a limiting nutrient
during fermentation at these conditions.
The pH and pO2 patterns were similar to those shown in
Figs 1d and 2, as aeration, agitation conditions and inocu-
lum size were the same.
Fed-batch experiments
Although temperature, initial inoculum, aeration and
agitation conditions were the same, growth of C. sake in
the first fed-batch experiment was lower than that obtained
in batch conditions up to 32 h of fermentation (Fig. 4).
After 38 h of growth, C. sake population reached the same
level in fed-batch conditions as in batch conditions.
Moreover, growth of C. sake stopped after 14 and 20 h
and did not started again until more medium was added to
the fermenter. During the period in which growth was
stopped, it could be observed that pO2 decrease was also
306 M . A B A D I A S ET AL.

9·00 12·5 1·25

8·00 10·0 1·00

Suc, fru, gluc (g l–1)

log10 CFU ml–1

7·00 7·5 0·75

Urea (g l–1)
6·00 5·0 0·50
Fig. 3 Consumption of sucrose (r), glucose
(j), fructose (·) and urea (d) during Candida
5·00 2·5 0·25 sake growth in a fermenter with 5 l of
molasses-based medium and set at
400 rev min)1, 150 l h)1 and 25C.
4·00 0·0 0·00 Fermenter was inoculated with 10 ml of
0 4 8 12 16 20 24 28
starter culture (150 rev min)1 for 24 h at
Time (h) 25C). Results of growth (n) are also reported

9·0 120
11 0
100
8·0 90
80
log10 CFU ml–1

70
pO2 (%)
pH

7·0 60
50
40 Fig. 4 Growth of Candida sake in a 5 l
fermenter in fed-batch conditions (m) com-
6·0 30 pared with batch conditions (n). Fermenter
20 was fed with 100 ml of a molasses : water
(50 : 50) at 8, 16 and 24 h (indicated with an
10
arrow). Fermenter was inoculated with 10 ml
5·0 0 of a starter culture and set at 150 l h)1 of air,
0 4 8 12 16 20 24 28 32 36 40 44 48 400 rev min)1 and 25C. pO2 (s) and pH
Time (h) values (+) for all the process are represented

stopped and reached its minimum (50%) after 28 h of
growth.
In the second series of experiments, after 18 h of growth
the 5 l fermenter (400 rev min)1, 150 l h)1, 25C) was
supplemented with a concentrated solution of glucose or
molasses in order to have a glucose concentration in the
medium of 5 g l)1, or the same initial molasses concentra-
tion, respectively. Regarding fed-batch with glucose, results
shown that although glucose was added in the moment that
it was depleted (18 h) C. sake growth was not significantly
increased (Fig. 5) reaching final population similar to that
obtained with no glucose added (6Æ5–7Æ3 · 108 CFU ml)1 in
26–28 h). After 26 h of growth, glucose was again depleted.
Candida sake cells first assimilated glucose and then sucrose.
Final concentration of sucrose was higher in this experiment
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
(7Æ5 g l)1) than in that which no glucose added (residual
sucrose 3Æ8 g l)1). The pH and pO2 pattern was similar to
that obtained with no glucose added. Similar pattern of
growth, pO2 and pH were obtained when the fermenter was
fed with molasses (data not shown). However, residual
sucrose after 28 h growth was 7Æ0 g l)1, approximately.
DISCUSSION
As the commercial production of C. sake for its use as a
biocontrol agent in postharvest requires low-cost and high
cell densities, optimization of growth conditions (aeration,
agitation speed and inoculum size) in a 5 l fermenter have
been studied. Moreover, sugar and urea consumption
studies have also been carried out.

9·00
8·00
log10 CFU ml–1
7·00
pH
Fig. 5 Effect of fed-batch of the fermenter
6·00
with glucose after 18 h of fermentation on
Candida sake growth (m), glucose (·), sucrose
(r), fructose (j) and urea (d) consumption. 5·00
pO2 (s) and pH (+) values for fed-batch
experiement are reported. Candida sake
growth at the same conditions in the batch 4·00
0 4
experiment is shown (n). Vertical bars
indicate S.E. of mean values
When yeast cells were inoculated into the fermenter, a
typical batch growth curve was observed when the viable
population was plotted against time. Maximum exponential
growth rate and maximum population density, which is the
most important parameter for our purpose, were obtained
when agitation was set at 400 rev min)1 and aeration was
150 l h)1, achieving a population of 8Æ0 · 108 CFU ml)1 in
30 h. These growth conditions were chosen for subsequent
experiments, as they seem to be the best agitation–aeration
combination. It is possible that at high aeration rates, when
low agitation speed is used, the impeller is unable to disperse
the incoming air and low oxygen transfer rates may be
achieved as a result of the impeller being flooded. Flooding
is the phenomenon where the air-flow dominates the flow
pattern and is due to an inappropriate combination of air
flow rate and speed of agitation (Stanbury et al. 1995). If air-
flow dominates the flow profile, the air could escape without
being distributed by agitation and it could not be used by
cells.
The dissolved O2 rapidly decreased during exponential
growth phase because of respiration of cells. During the
stationary phase, pO2 level started increasing probably
because of a decrease in the respiration rate of cells.
Meesters et al. (1996) attributed this increase of dissolved
O2 to the lack of substrate and used the increase of pO2 as
signal to feed the fermenter. However, our studies of
consumption of sugars have demonstrated that even pO2
increased at the stationary phase, there is still sucrose in
the medium, and this increase could not be attributed to
lack of substrate but probably to the decrease of growth
rate. The pH value of molasses medium decreased during
the exponential growth phase of C. sake, reached a
minimum almost simultaneously to the pO2 minimum
value, and then increased again during the stationary
growth phase (generally at the same time of pO2 increase).
Actively growing yeast acidifies their growth medium
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
OPTIMIZATION OF CANDIDA SAKE PRODUCTION 307
125·0 12·5 1·25
100·0 10·0 1·00
Suc, fru, gluc (g l )
–1
75·0 7·5 0·75
Urea (g l–1)
pO2
0·50
50·0 5·0
25·0 2·5 0·25
0·0 0·0 0·00
8 12 16 20 24 28
Time (h)
through a combination of differential ion uptake, proton
secretion during nutrient transport, direct secretion of
organic acids and CO2 evolution (Walker 1998). It is
possible that active respiration of C. sake yeast during
exponential phase, with a decrease of dissolved O2 and
consequently an increase of CO2 production, caused the
acidification of medium. On the contrary, when respiration
rate decreased during the stationary phase of growth, CO2
production decreased and consequently, pH increased
again.
An inoculum of 10 ml of the starter culture (initial C. sake
population in the fermenter of 106 CFU ml)1) was chosen
for the optimization in the 5 l fermenter. One millilitre
inoculum gave lower final counts and, although 50-ml
inoculum gave similar counts, a higher amount of starter was
required. Minimum pH and pO2 values were reached later
when the inoculum was lower. This could be probably
because of the fact that exponential growth phase is delayed
when there is lower initial population.
In the consumption studies, it could be seen that glucose
and fructose were used by C. sake cells before sucrose.
Glucose is considered the main carbon source by all
microorganisms because of its size, rapid uptake, utilization
and cellular energy conversion. Once glucose and fructose
were consumed, C. sake cells started to assimilate sucrose.
Growth of C. sake in fed-batch conditions was not
improved compared with that obtained in batch conditions.
Moreover, growth was stopped after 14 and 20 h of growth
and did not started again until the fermenter was fed with
more molasses. This could be probably because of the fact
that carbon source is depleted and the microorganism is
not able to grow. The use of fed-batch culture takes the
advantage of the fact that the concentration of the limiting
substrate may be maintained at a very low level, thus
avoiding the repressive effects of high substrate concen-
tration. It is also used in controlled glucose-feeding
308 M . A B A D I A S ET AL.
regimes to avoid the Crabtree effect and to maximize
respiratory growth (Walker 1998). In our case, the high
amount of sucrose in the molasses medium does not seem
to have a repressive effect on C. sake growth and the
population achieved was not improved by using fet-batch
technology. In contrast, some studies have indicated that
high cell density fermentations can be achieved in
S. cerevisiae using fed-batch technology and molasses as
the growth substrate (Beudeker et al. 1989; Reed and
Nagodawithana 1991; Walker 1998).
As glucose and fructose were depleted after 18 h of
growth, addition of glucose by means of a concentrated
glucose solution or by adding more molasses was investi-
gated. This extra glucose or sucrose supply did not increase
significantly the final population of C. sake cells. Urea was
not completely used by C. sake cells during the fermentation
process and consequently, feeding the fermenter with an
extra amount of urea was not considered necessary for
further investigation.
These studies have demonstrated that C. sake cells can be
produced rapidly in a liquid culture. High cell concentra-
tions are produced in simple medium using inexpensive
carbohydrates and nitrogen source. Moreover, in this study,
using a molasses-based medium and batch fermentation in a
5 l fermenter set at 25C, 400 rev min)1 of stirring speed,
aeration of 150 l h)1 and an initial population of about
106 CFU ml)1, approx. 8 · 108 CFU ml)1 were obtained
after 28 h of growth. At this time, cells are in the stationary
phase of growth. Some studies (Mackenzie et al. 1988;
Jackson et al. 1998; Abadias et al. 2001) suggested that
harvesting stationary phase cells is desirable to enhance cell
survival under stress conditions such as low water potential
or drying. This could be an advantage, as after production
C. sake cells have to be formulated and consequently, in the
most cases, dried.
Taking into account the commercial applicability of this
biocontrol agent in the packinghouses and the results
obtained in this study, it should be noted that with the
cells produced with 12Æ5 l of low-cost molasses medium in a
relative short-time, it would be possible to treat approx.
30 000 kg of fruit, in a 1000-l drencher, which is the system
used in Spain for treating pome fruits. Moreover, these
results could also be used to provide a reliable basis for a
scaling-up of this fermentation process to a pilot scale level.
Future research should be focused on studing potential of
scaling up.
ACKNOWLEDGEMENTS
The authors are grateful to the European Community
(QLK5-CT-1999-01065) for their financial support and to
the Spanish Government (CICYT, Comisión Interministe-
rial de Ciencia y Tecnologı́a grant ALI99-0652.C02-01).
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
REFERENCES
Abadias, M., Teixidó, N., Usall, J., Viñas, I. and Magan, N. (2000)
Solute stresses affect growth patterns, endogenous water potentials
and accumulation of sugars and sugar alcohols in cells of the
biocontrol yeast Candida sake. Journal of Applied Microbiology 89,
1009–1017.
Abadias, M., Teixidó, N., Usall, J., Viñas, I. and Magan, N. (2001)
Improving water stress tolerance of the biocontrol yeast Candida sake
grown in molasses-based media by physiological manipulation.
Canadian Journal of Microbiology 4, 123–129.
Beudeker, R.F., Van Dam, H.W., Van Der Plaat, J.B. and Vellenga, K.
(1989) Developments in baker’s yeast production. In Yeast.
Biotechnology and Biocatalysis ed. Verachert, H. and De Mot, R.
pp. 103–146. New York and Basel: Marcel Dekker, Inc.
Churchill, B.N. (1982) Mass production of microorganisms for
biological control. In Biological Control of Weeds with Plant Pathogens
ed. Charudattan, R. and Walker, H.L. pp. 157–173. New York: John
Wiley and Sons.
Costa, E., Teixidó, N., Usall, J., Atarés, E. and Viñas, I. (2001)
Production of the biocontrol agent Pantoea agglomerans strain CPA-2
using comercial products and by-products. Applied Microbiology and
Biotechnology 56, 367–371.
Costa, E., Usall, J., Teixidó, N., Torres, R. and Viñas, I. (2002) Effect
of package and storage conditions on viability and efficacy of the
freeze-dried biocontrol agent Pantoea agglomerans strain CPA-2.
Journal of Applied Microbiology 92, 873–878.
Droby, S., Cohen, L., Daus, A., Weiss, B., Horev, B., Chalutz, E.,
Katz, H., Keren-Tzur, M. and Shachnai, A. (1998) Commercial
testing of Aspire: a yeast preparation for the biological control of
postharvest decay of citrus. Biological Control 12, 97–101.
Filonow, A.B., Vishniac, H.S., Anderson, J.A. and Janisiewicz, W.J.
(1996) Biological control of Botrytis cinerea in apple by yeasts from
various habitats and their putative mechanisms of antagonism.
Biological Control 7, 212–220.
Jackson, M.A., Frymier, J.S., Wilkinson, B.J., Zorner, P. and Evans, S.
(1998) Growth requirements for production of stable cell of the
bioherbicidal bacterium Xanthomonas campestris. Journal of Industrial
Microbiology and Biotechnology 21, 237–241.
Janisiewicz, W.J. (1987) Postharvest biological control of blue mold on
apples. Phytopathology 77, 481–485.
Janisiewicz, W.J. and Marchi, A. (1992) Control of storage rots on
various pear cultivars with a saprophytic strain of Pseudomonas
syringae. Plant Disease 76, 555–560.
Mackenzie, K.F., Singh, K.K. and Brown, A.D. (1988) Water stress
plating hipersensitivity of yeasts: protective role of trehalose in
Saccharomyces cerevisiae. Journal of General Microbiology 134, 1661–
1666.
Meesters, P.A.E.P., Huijberts, G.N.M. and Eggink, G. (1996) High-
cell-density cultivation of the lipid accumulating yeast Cryptococcus
curvatus using glycerol as carbon source. Applied Microbiology and
Biotechnology 45, 575–579.
Nunes, C., Usall, J., Teixidó, N. and Viñas, I. (2001) Biological control
of postharvest diseases using a bacterium Pantoea agglomerans CPA-
2. International Journal of Food Microbiology 70, 53–60.
Nunes, N., Usall, J., Teixidó, N., Abadı́as, M. and Viñas, I. (2002)
Improved control of postharvest decay of pears by the combination

of Candida sake (CPA-1) and ammonium molybdate. Biological
Control 92, 281–287.
Pusey, P.L. and Wilson, C.L. (1984) Postharvest biological control of
stone fruit brown rot by Bacillus subtilis. Plant Disease 68, 753–756.
Reed, G. and Nagodawithana, T.W. (1991) Yeast Technology. New
York: Van Nostrand Reinhold.
Roberts, R.G. (1990) Postharvest biological control of gray mold of
apple by Cryptococcus laurentii. Phytopathology 80, 526–530.
Stanbury, P.F., Whitaker, A. and Hall, S.J. (1995) Principles of
Fermentation technology, 2nd edn. Oxford: Butterworth-Heinemann.
Teixidó, N., Viñas, I., Usall, J. and Magan, N. (1998) Control of
blue mold of apples by preharvest application of Candida sake
grown in media with different water activity. Phytopathology 88,
960–964.
Teixidó, N., Usall, J., Palou, L., Asensio, A., Nunes, C. and Viñas, I.
(2001) Improving control of green and blue molds of oranges by
combining Pantoea agglomerans (CPA-2) and sodium bicarbonate.
European Journal of Plant Pathology 107, 685–694.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 95, 301–309
OPTIMIZATION OF CANDIDA SAKE PRODUCTION 309
Usall, J., Teixidó, N., Fons, E. and Viñas, I. (2000) Biological control
of blue mold on apple by a strain of Candida sake under several
controlled atmosphere conditions. International Journal of Food
Microbiology 58, 83–92.
Usall, J., Teixidó, N., Torres, R., Ochoa de Eribe, X. and Viñas, I.
(2001) Pilot tests of Candida sake (CPA-1) applications to control
postharvest blue mold on apple fruit. Postharvest Biology and
Technology 21, 147–156.
Viñas, I., Usall, J., Teixidó, N. and Sanchis, V. (1998) Biological
control of major postharvest pathogens on apple with Candida sake.
International Journal of Food Microbiology 40, 9–16.
Viñas, I., Usall, J., Nunes, C. and Teixidó, N. (1999) Nueva cepa
bacteriana Pantoea agglomerans, Beijerinck (1888) Gavini, Mergaert,
Beji, Mielcareck, Izard, Kerstersy, De Ley (1989) y su utilización
como agente de control biológico de las enfermedades fúngicas de
fruta. Spanish Patent P9900612.
Walker, G.M. (1998) Yeast Physiology and Biotechnology. Chichester:
John Wiley and Sons.