Journal of Applied Microbiology 2002, 92, 851–859

Growth studies of potentially probiotic lactic acid bacteria

in cereal-based substrates

D. Charalampopoulos, S.S. Pandiella and C. Webb

Satake Centre for Grain Process Engineering, UMIST, Department of Chemical Engineering, Manchester, UK

2001/21: received 19 March 2001, revised 31 July 2001 and accepted 16 November 2001

D . C H A R A L A M P O P O U L O S , S . S . P A N D I E L L A A N D C . W E B B . 2002.
Aims: The overall growth kinetics of four potentially probiotic strains (Lactobacillus fermentum,
Lact. reuteri, Lact. acidophilus and Lact. plantarum) cultured in malt, barley and wheat media
were investigated. The objectives were to identify the main factors influencing the growth and
metabolic activity of each strain in association with the cereal substrate.
Methods and Results: All fermentations were performed without pH control. A logistic-type
equation, which included a growth inhibition term, was used to describe the experimental data.
In the malt medium, all strains attained high maximum cell populations (8Æ10–10Æ11
log10 cfu ml)1, depending on the strain), probably due to the availability of maltose, sucrose,
glucose, fructose (approx. 15 g l)1 total fermentable sugars) and free amino nitrogen (approx.
80 mg l)1). The consumption of sugars during the exponential phase (10–12 h) resulted in the
accumulation of lactic acid (1Æ06–1Æ99 g l)1) and acetic acid (0Æ29–0Æ59 g l)1), which
progressively decreased the pH of the medium. Each strain demonstrated a specific preference
for one or more sugars. Since small amounts of sugars were consumed by the end of the
exponential phase (17–43%), the decisive growth-limiting factor was probably the pH, which at
that time ranged between 3Æ40 and 3Æ77 for all of the strains. Analysis of the metabolic products
confirmed the heterofermentative or homofermentative nature of the strains used, except in the
case of Lact. acidophilus which demonstrated a shift towards the heterofermentative pathway.
All strains produced acetic acid during the exponential phase, which could be attributed to the
presence of oxygen. Lactobacillus plantarum, Lact. reuteri and Lact. fermentum continued to
consume the remaining sugars and accumulate metabolic products in the medium, probably
due to energy requirements for cell viability, while Lact. acidophilus entered directly into the
decline phase. In the barley and wheat media all strains, especially Lact. acidophilus and Lact.
reuteri, attained lower maximum cell populations (7Æ20–9Æ43 log10 cfu ml)1) than in the malt
medium. This could be attributed to the low sugar content (3–4 g l)1 total fermentable sugar
for each medium) and the low free amino nitrogen concentration (15Æ3–26Æ6 mg l)1). In all
fermentations, the microbial growth ceased at pH values (3Æ73–4Æ88, depending on the strain)
lower than those observed for malt fermentations, which suggests that substrate deficiency in
sugars and free amino nitrogen contributed to growth limitation.
Conclusions: The malt medium supported the growth of all strains more than barley and
wheat media due to its chemical composition, while Lact. plantarum and Lact. fermentum
appeared to be less fastidious and more resistant to acidic conditions than Lact. acidophilus and
Lact. reuteri.
Significance and Impact of the Study: Cereals are suitable substrates for the growth of
potentially probiotic lactic acid bacteria.

Correspondence to: S.S. Pandiella, Satake Centre for Grain Process Engineering, UMIST, Department of Chemical Engineering, PO Box 88, Manchester M60 1QD,
UK (e-mail: s.pandiella@umist.ac.uk).

ª 2002 The Society for Applied Microbiology

852 D . C H A R A L A M P O P O U L O S ET AL.
INTRODUCTION
Many lactic acid bacteria strains, characterized as probiotics,
have been proven to beneficially affect humans or animals by
improving the properties of their indigenous gut flora
(Fuller 1989; Havenaar and Huis in’t Veld 1992). The
incorporation of potentially probiotic lactic acid bacteria of
human origin in traditional fermented foods has been
established in the dairy industry, leading to the production
of different types of fermented milks and yoghurts (Gomes
and Malcata 1999). Strains of the Lactobacillus genus, such
as Lactobacillus acidophilus, Lact. casei, Lact. reuteri, Lact.
gasseri and Lact. paracasei, constitute a significant proportion
of the lactic acid bacteria used in commercial probiotic milk-
based products (Shortt 1999). New applications of probiotic
micro-organisms in foods have been introduced into the
market or are still in the development phase, such as frozen
yoghurt, soy yoghurt, dairy desserts, cheese, ice-cream,
bread and chocolate (De Vuyst 2000).
This work is part of a project studying the development of
a novel food fermentation process, based on the use of
cereals as substrates for potentially probiotic lactic acid
bacteria. Cereals, such as wheat and rye, are used for
sourdough production, a fermentation process that now-
adays is performed by using defined mixed starter cultures,
consisting of lactic acid bacteria strains isolated from natural
environments and yeasts (Lönner and Åkesson 1988; Gänzle
et al. 1998). Other cereal-based fermented foods are pro-
duced indigenously in Asia and Africa by natural lactic acid
fermentation under uncontrolled conditions (Oyewole
1997). However, as natural fermentations rely on microbial
populations present in the raw material, these products
exhibit substantial variations in flavour and quality (Meraz
et al. 1992; Giraud et al. 1998). The good adaptation of
lactic acid bacteria in cereals suggests that the utilization of a
potentially probiotic strain as starter culture in a cereal
substrate would produce a fermented food with defined and
consistent characteristics and possibly health-promoting
properties. However, several technological aspects have to
be considered in the design of such a novel food fermen-
tation process, such as the composition and processing of the
raw material, the growth capacity and productivity of the
starter culture and the stability of the final product during
storage (De Vuyst 2000).
Probiotic products are usually standardized, based on the
presumption that culture viability is a reasonable measure of
probiotic activity, thus the ability of the strain to attain a
high cell population is of primary importance. A concen-
tration of approx. 107 cells ml)1 at the time of consumption
is considered functional (Gomes and Malcata 1999; Shortt
1999). High cell growth and acidification rates would also
result in the reduction of fermentation times and enhance
the viability of the specific strain by preventing growth of
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
undesirable micro-organisms present in the raw material
(Marklinder and Lönner 1992). Lactobacilli are very
fastidious micro-organisms that require fermentable carbo-
hydrates, amino acids, vitamins of the B-complex, nucleic
acids and minerals to grow, regardless of the specific
nutrient requirements of the strain (Gomes and Malcata
1999). Thus, the substrate composition and nutritional
requirements of the strain considerably affect the overall
performance of the fermentation. Microbial growth also
depends on environmental factors, such as pH, temperature
and accumulation of metabolic end-products. Several
investigators have demonstrated the effects of these para-
meters on the growth of lactic acid bacteria based on
experiments and have used empirical mathematical models
to describe them (Mercier et al. 1992; Gänzle et al. 1998;
Lejeune et al. 1998).
The distribution of the metabolic end-products is another
important parameter when evaluating a microbial food
process. Lactic acid-fermented foods are usually character-
ized by a sour taste, which is mainly attributed to lactic and
acetic acids produced via the homo- or heterofermentative
metabolic pathways. Consequently, an appropriate selection
of the strain is necessary to efficiently control the distribu-
tion of the metabolic end-products (Lönner and Åkesson
1988; De Vuyst 2000).
In this work, the overall growth kinetics of the potentially
probiotic Lact. fermentum, Lact. reuteri, Lact. acidophilus and
Lact. plantarum strains in malt, barley and wheat media have
been studied. The main objectives were to identify the key
factors influencing the growth and metabolic activity of each
strain in association with the cereal substrate in order to
evaluate the production process of a novel cereal-based
probiotic food. Subsequent research will aim to develop a
mathematical model that could predict the growth and
acidifying properties of specific probiotic lactic acid bacteria
in different cereal substrates, taking into account their
different sugar compositions.
MATERIALS AND METHODS
Micro-organisms and culture conditions
The micro-organisms used in this study were as follows:
Lact. reuteri NCIMB 11951 (National Collection of Indus-
trial and Marine Bacteria, Aberdeen, Scotland, UK),
isolated from human intestine; Lact. acidophilus NCIMB
8821, isolated from human saliva; Lact. fermentum NCIMB
12116, isolated from ogi (fermented corn gruel indigenous to
South Nigeria) and Lact. plantarum NCIMB 8826, isolated
from human saliva. The strains were maintained at 4C and
subcultured monthly on slants prepared from MRS agar
(Oxoid, Basingstoke, Hampshire, UK). Colonies isolated
from MRS agar plates were precultured twice in MRS broth
 GROWTH OF LACTIC ACID BACTERIA IN CEREAL-BASED SUBSTRATES
(Oxoid) for approx. 12 h at 37C. The 12-h precultured cells
were then centrifuged (5000 g, 10 min, 4C), washed twice
with sterile quarter-strength Ringer’s solution and resus-
pended in Ringer’s solution. The bacterial suspensions were
then used to inoculate the fermentation media at 2Æ5% (v/v)
for Lact. fermentum, 3% (v/v) for Lact. reuteri, 7Æ5% (v/v)
for Lact. acidophilus and 2Æ5% (v/v) for Lact. plantarum. In
all cases, the initial microbial concentration was approxi-
mately 107 cfu ml)1.
Cereal-based fermentation media
Malt, wheat and barley grains were used to prepare the
fermentation media following the same procedure. The
grains were ground in a laboratory Falling Number hammer
mill (Perten Instruments, Huddinge, Sweden) with a sieve
of size 0Æ5 mm. A sample (50 g) of the flour obtained was
mixed with 450 ml tapwater and the resulting slurry
centrifuged (6000 g) for 30 min at room temperature. The
starch-free supernatant fluid was collected and immediately
sterilized at 121C for 45 min. Sedimentation of solids
(possibly protein mixtures) was observed after sterilization
and approx. 4–5% (w/w) of solids were present in the final
fermentation media. The extraction and sterilization pro-
cedures were repeated four times. The sugar content, pH,
free amino nitrogen (FAN) and buffering capacity (mmol
HCl per pH unit per unit volume) of the three media were
analysed. The small differences observed in the initial
composition of the media were probably due to reactions
between sugars and amino acids during sterilization. The
malt medium had a pH value between 5Æ12 and 5Æ21 and
contained maltose (4Æ82–5Æ43 g l)1), sucrose (6Æ36–
6Æ98 g l)1), glucose (2Æ00–2Æ05 g l)1) and fructose (0Æ95–
1Æ12 g l)1); the FAN concentration was between 75Æ2
and 82Æ1 mg l)1 and the buffering capacity was
8Æ45 mmol pH)1 l)1. The barley medium had a pH value
between 5Æ69 and 5Æ83 and contained maltose (0Æ91–
1Æ19 g l)1), sucrose (1Æ56–1Æ78 g l)1), glucose (0Æ17–
0Æ20 g l)1) and fructose (0Æ16–0Æ18 g l)1); the FAN
concentration was between 15Æ3 and 15Æ9 mg l)1 and the
buffering capacity was 6Æ54 mmol pH)1 l)1. The wheat -
medium had a pH value between 6Æ23 and 6Æ38 and contained
maltose (0Æ82–1Æ09 g l)1), sucrose (1Æ88–2Æ12 g l)1), glucose
(0Æ15–0Æ18 g l)1) and fructose (0Æ14–0Æ16 g l)1); the FAN
concentration was between 20Æ1 and 26Æ6 mg l)1 and the
buffering capacity was 3Æ30 mmol pH)1 l)1. All fermenta-
tions were performed under no pH control in 500-ml screw-
capped bottles, without agitation. The temperature was set
at 37C, which is usually the optimum for lactic acid
bacteria of human origin. Samples were collected approx.
every 2 h for the first 12 h and then at intervals of 8–12 h
during the next 36 h. All fermentations were performed in
duplicate.
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
853
Analytical methods and statistics
Bacterial enumeration. Fermentation samples were deci-
mally diluted in sterile quarter-strength Ringer’s solution
and four 0Æ15-ml aliquot dilutions were plated on MRS agar
using a precalibrated pipette and incubated at 37C for 48 h.
Colony-forming units were counted (cfu ml)1) and the
results expressed as their log10 values. Throughout this
work, cell values are given as mean values of eight replicate
measurements and the S.E. of the mean was calculated with
95% confidence.
Buffering capacity. The buffering capacity of each cereal
medium was determined by titrating 100 ml of the medium
with HCl (1 mol l)1). The values were expressed as the
amount of HCl (mmoles) required to drop 1 pH unit per
unit volume (l) (Pai et al. 2001).
Determination of ethanol and acetic acid. Ethanol and
acetic acid were determined in sample supernatant fluids by
gas chromatography using a gas chromatograph (5890;
Hewlett-Packard, Palo Alto, CA, USA) equipped with a
flame ionization detector, set at 250C, a ChemStation
Integrator (Hewlett-Packard) and a capillary column
(30 m · 0Æ25 mm; HP-5; Hewlett-Packard). The tempera-
ture was initially held at 90C for 2Æ5 min and then
increased at 8C min)1 to 138C; the total time of analysis
was 10 min. The retention times were approximately 1Æ1
and 3Æ8 min for ethanol and acetic acid, respectively.
The data presented are mean values of three replicate
measurements.
Determination of free amino nitrogen, sugars and lactic
acid. The FAN content was determined by the ninhy-
drin method (Magne and Larher 1992). Glucose, fruc-
tose, sucrose, maltose, D- and L-lactic acid were
determined using enzymatic methods (Boehringer
Mannheim). The total lactic acid concentration was
calculated by adding the values for D- and L-lactic acids.
The inaccuracy and imprecision of each of the enzymatic
methods was determined by conducting a block design
experiment, consisting of three sets of six replicate
measurements. The one-way AnoVa test applied to the
data showed that variations between block means were, in
all cases, not significant (P > 0Æ05), which suggests that
they were probably due to imprecision and not inaccur-
acy. Thus, the S.D. of each method was calculated after
treating all of the measurements from the three blocks as
a single set. The S.D. values obtained were 0Æ10 for lactic
acid, 0Æ12 for maltose, 0Æ14 for sucrose, 0Æ08 for glucose
and 0Æ09 for fructose. Each value was regarded as
expressing the S.E. of all individual measurements with
95% confidence.
854 D . C H A R A L A M P O P O U L O S ET AL.

RESULTS exponential growth phase of 6–8 h was observed for Lact.

Growth kinetics of lactic acid bacteria in malt,
wheat and barley media
A common model used for the simulation of microbial
growth is the logistic equation (Mercier et al. 1992; Özen
and Özilgen 1992; Lejeune et al. 1998):
dX
¼ lX ¼ lmax Xð1  X=Xmax Þ
dt
where l is the specific growth rate, lmax is the maximum
specific growth rate predicted by the model, t is time, X is
the cell concentration and Xmax the maximum attainable cell
concentration. This equation represents both the exponen-
tial (X < < Xmax) and the stationary phase (X ¼ Xmax) of
growth. The term (1 ) X/Xmax) implies that micro-organ-
isms inhibit their own growth either by competing for the
same carbon and nitrogen sources present in the fermenta-
tion medium or by accumulating inhibitory compounds.
Integration of eqn (1) yields the following expression for cell
concentration:
X0 Xmax expðlmax tÞ
X¼
Xmax  X0 þ X0 expðlmax tÞ
The kinetic parameters of eqn (2) (see Table 1) were
estimated from the experimental data by selecting the set of
parameters that minimize the sum of the square deviation
between experimental and computed values. To measure the
precision of the model an AnoVa was performed using as
null hypothesis that the residuals from the non-linear
regression are non-zero only because of uniform imprecision
in the measurement of the response variable (cell number)
and not due to the lack of fit. The ratio F was in all cases less
than Ftable (Draper and Smith 1981).
The experimental data and simulations of Lact. fermentum,
Lact. reuteri, Lact. plantarum and Lact. acidophilus in malt,
wheat and barley media are presented in Fig. 1. An
fermentum and Lact. acidophilus, while Lact. reuteri and Lact.
plantarum grew exponentially until 10–12 h of fermentation.
The viable cell densities of Lact. fermentum, Lact. plantarum
and Lact. reuteri declined slightly during the stationary
phase (12–48 h); however, the cell population of Lact.
acidophilus significantly declined after the end of the
exponential phase. Since all fermentations were performed
ð1Þ under no pH control, the organic acids formed via the
metabolic pathways decreased the pH of the media.
In malt medium Lact. fermentum exhibited a higher
maximum specific growth rate and population density than
Lact. reuteri (see Table 1). Lactobacillus plantarum reached
the highest population densities (10Æ11 ± 0Æ18 log10
cfu ml)1). At the end of the exponential phase pH values
dropped from 5Æ20 ± 0Æ09 to 3Æ77 ± 0Æ09, 3Æ72 ± 0Æ09,
3Æ73 ± 0Æ09 and 3Æ40 ± 0Æ09 for Lact. fermentum, Lact.
reuteri, Lact. acidophilus and Lact. plantarum, respectively.
Lactobacillus fermentum and Lact. plantarum attained high
cell populations growing in barley medium (see Table 1).
However, Lact. reuteri and Lact. acidophilus did not grow
well. Based on the model parameters, the increase in their
ð2Þ cell population was 1Æ16 ± 0Æ08 and 0Æ73 ± 0Æ09 log10
cfu ml)1, respectively. At the end of the exponential phase
pH values decreased from 6Æ20 ± 0Æ09 to 4Æ61 ± 0Æ09,
4Æ88 ± 0Æ09, 3Æ93 ± 0Æ09 and 3Æ92 ± 0Æ09 for Lact. fermen-
tum, Lact. reuteri, Lact. acidophilus and Lact. plantarum,
respectively.
In the wheat medium the four lactic acid bacteria strains
displayed similar growth patterns to those observed in barley
medium (see Table 1). The growth of Lact. reuteri and Lact.
acidophilus was again inhibited, leading to an increase in cell
population of only 1Æ03 ± 0Æ07 and 0Æ70 ± 0Æ08 log10
cfu ml)1. At the end of the exponential phase pH values
dropped from 5Æ85 ± 0Æ09 to 4Æ50 ± 0Æ09, 4Æ40 ± 0Æ09,
3Æ73 ± 0Æ09 and 3Æ83 ± 0Æ09 for Lact. fermentum, Lact.
reuteri, Lact. acidophilus and Lact. plantarum, respectively.

Table 1 Numerical values of estimated

Medium Micro-organism lmax (h)1) X0 (log10 cfu ml)1) Xmax (log10 cfu ml)1)
microbial growth parameters in malt, barley
Malt Lact. fermentum 0Æ62 ± 0Æ04 6Æ85 ± 0Æ08 9Æ68 ± 0Æ03 and wheat media based on eqn (2)
Lact. plantarum 0Æ41 ± 0Æ03 6Æ90 ± 0Æ10 10Æ11 ± 0Æ18
Lact. reuteri 0Æ38 ± 0Æ02 6Æ20 ± 0Æ07 8Æ86 ± 0Æ06
Lact. acidophilus 0Æ19 ± 0Æ02 6Æ89 ± 0Æ06 8Æ10 ± 0Æ06
Barley Lact. fermentum 0Æ43 ± 0Æ05 6Æ90 ± 0Æ11 9Æ12 ± 0Æ05
Lact. plantarum 0Æ20 ± 0Æ02 6Æ71 ± 0Æ13 9Æ43 ± 0Æ10
Lact. reuteri 0Æ13 ± 0Æ01 6Æ14 ± 0Æ04 7Æ28 ± 0Æ05
Lact. acidophilus 0Æ18 ± 0Æ03 7Æ02 ± 0Æ04 7Æ73 ± 0Æ03
Wheat Lact. fermentum 0Æ53 ± 0Æ05 6Æ93 ± 0Æ09 9Æ28 ± 0Æ04
Lact. plantarum 0Æ23 ± 0Æ02 7Æ21 ± 0Æ05 9Æ29 ± 0Æ06
Lact. reuteri 0Æ13 ± 0Æ01 6Æ22 ± 0Æ03 7Æ20 ± 0Æ04
Lact. acidophilus 0Æ15 ± 0Æ01 7Æ02 ± 0Æ02 7Æ71 ± 0Æ03

ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
 GROWTH OF LACTIC ACID BACTERIA IN CEREAL-BASED SUBSTRATES
log10 cfu ml–1
log10 cfu ml–1
log10 cfu ml–1
Fig. 1 Growth of lactic acid bacteria in (a) malt, (b) barley and (c)
wheat media. The solid lines represent the growth curves predicted
with eqn (2). Experimental data symbols: j, Lactobacillus plantarum;
d, Lact. fermentum; m, Lact. reuteri and e, Lact. acidophilus
Substrates and products
In order to investigate the ability of each strain to utilize the
available carbohydrates and to study the distribution of the
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
855
primary products of lactic acid metabolism, samples from all
fermentations were taken after the end of the exponential
phase (12 h) and at the end of the fermentation (48 h). They
were analysed for their content of sugars, organic acids and
ethanol. The results from malt fermentations are shown in
Tables 2 and 3.
Lactobacillus fermentum and Lact. reuteri produced mainly
lactic and acetic acids and ethanol (see Table 2). The molar
ratios of lactic acid produced to ethanol plus acetic acid at
the end of the exponential phase were 0Æ72 ± 0Æ10 and
0Æ62 ± 0Æ11, respectively. During the stationary phase
(12–48 h) the lactic acid and ethanol concentrations
increased, while the acetic acid concentration remained
constant for both microbes, giving an overall molar ratio
throughout the fermentation of 0Æ90 ± 0Æ11 and 0Æ95 ± 0Æ10,
respectively. Analysis of the samples taken at the end of the
exponential phase of Lact. fermentum indicated a steady
decrease in all sugars, but mainly maltose, while Lact. reuteri
consumed primarily glucose and fructose (see Table 3).
Lactobacillus plantarum demonstrated a homofermentative
metabolic pattern producing mainly lactic acid and small
amounts of acetic acid. Sugar analysis indicated a significant
depletion of glucose and a similar consumption of fructose,
maltose and sucrose (see Table 3). The growth of Lact.
acidophilus was associated with the production of lactic acid
and comparably significant amounts of acetic acid. The low
concentrations of metabolites coincided with a relatively low
consumption of the available carbohydrates, mainly glucose
(see Table 3). In contrast to the other microbes, sugar
consumption was not observed between 12 and 48 h.
DISCUSSION
The aim of this study was to monitor and compare the
growth characteristics of potentially probiotic strains of
Lact. fermentum, Lact. reuteri, Lact. plantarum and Lact.
acidophilus species in natural cereal substrates formulated
without the addition of supplements. Important technolo-
gical points in the design and evaluation of this type of food
fermentation process are the composition of the raw
material, specific growth rate of the starter culture applied,
final cell population, acidification rate and distribution of the
primary metabolic products.
The empirical logistic-type model (eqn (2)), that was used
to simulate the growth of each of the four microbes in malt,
barley and wheat media, described well the clear exponential
phase (l ¼ lmax, X < < Xmax), the deceleration stage of the
exponential phase (l < lmax, X < Xmax) and the constant
stationary phase (l ¼ 0 h)1, X ¼ Xmax). This could be due
to the fact that the model can describe both the effect of a
product inhibition and the depletion of an essential nutrient
on microbial growth (Lejeune et al. 1998). This model
cannot, however, describe the declining stationary phase.
856 D . C H A R A L A M P O P O U L O S ET AL.

Table 2 Primary metabolites produced during growth of Lactobacillus fermentum, Lact. reuteri, Lact. plantarum and Lact. acidophilus in the malt
medium. The prefermented malt medium contained 0Æ05 g l)1 lactic acid (La); acetic acid (Ac) and ethanol (Et) were not detected

Lact. fermentum Lact. reuteri Lact. plantarum Lact. acidophilus

12 h 48 h 12 h 48 h 12 h 48 h 12 h 48 h
)1
Lactic acid (g l ) 1Æ39 ± 0Æ1 2Æ41 ± 0Æ1 1Æ06 ± 0Æ1 2Æ96 ± 0Æ1 1Æ99 ± 0Æ1 5Æ74 ± 0Æ1 0Æ81 ± 0Æ1 1Æ07 ± 0Æ1
Acetic acid (g l)1) 0Æ59 ± 0Æ02 0Æ57 ± 0Æ01 0Æ50 ± 0Æ02 0Æ62 ± 0Æ02 0Æ29 ± 0Æ06 0Æ32 ± 0Æ05 0Æ38 ± 0Æ04 0Æ37 ± 0Æ03
Ethanol (g l)1) 0Æ53 ± 0Æ02 0Æ93 ± 0Æ06 0Æ48 ± 0Æ02 1Æ11 ± 0Æ04 ND ND ND ND
Molar ratio 0Æ72 ± 0Æ10 0Æ90 ± 0Æ11 0Æ62 ± 0Æ11 0Æ95 ± 0Æ10 4Æ56 ± 1Æ17 11Æ95 ± 2Æ07 1Æ42 ± 0Æ32 1Æ92 ± 0Æ33
La : Ac + Et
YP/Sb 0Æ84 ± 0Æ29 0Æ42 ± 0Æ06 0Æ88 ± 0Æ39 0Æ42 ± 0Æ04 0Æ66 ± 0Æ21 0Æ76 ± 0Æ14 0Æ68 ± 0Æ42 0Æ62 ± 0Æ28

ND, Not detected.

YP/Sb, Product yield.

Table 3 Sugar content of the malt medium during fermentation with Lactobacillus fermentum, Lact. reuteri, Lact. plantarum and Lact. acidophilus

Concentration (g l)1) at different times (h)

Lact. fermentum Lact. reuteri Lact. plantarum Lact. acidophilus

Substrates 0 12 48 0 12 48 0 12 48 0 12 48

Maltose 5Æ27 3Æ49 2Æ10 5Æ53 5Æ17 1Æ86 4Æ78 4Æ06 2Æ25 4Æ82 4Æ67 4Æ80
Sucrose 7Æ33 6Æ88 3Æ85 6Æ98 6Æ40 2Æ08 6Æ36 3Æ67 2Æ30 6Æ88 6Æ38 6Æ42
Glucose 2Æ10 1Æ64 0Æ03 2Æ14 1Æ14 0Æ08 2Æ11 0Æ28 0Æ17 2Æ16 1Æ03 1Æ00
Fructose 1Æ14 0Æ73 0Æ08 1Æ08 0Æ32 0Æ09 0Æ98 0Æ16 0Æ05 0Æ95 0Æ28 0Æ09

The malt medium supported well the growth of Lact.
plantarum, Lact. fermentum and Lact. reuteri, which showed
increases in their cell populations of 3Æ21 ± 0Æ28, 2Æ83 ± 0Æ16
and 2Æ66 ± 0Æ13 log10 cfu ml)1 at the end of the exponential
phase (12 h), respectively. This could be attributed to the
simultaneous presence of considerable amounts of monosac-
charides (glucose and fructose) and disaccharides (maltose and
sucrose) in the malt medium (approximately 3 and 12 g l)1,
respectively). Our results showed that each microbe demon-
strated a specific preference for one or more sugars during the
exponential phase (see Table 3). The preference of Lact.
plantarum towards glucose has also been suggested by Samuel
et al. (1980) and Gobbetti et al. (1994). The preferential
utilization of maltose by the sourdough-originated Lact.
fermentum strain and glucose and fructose by the intestinal
Lact. reuteri strain could be due to the origins of these strains,
since sourdough strains of both species have been shown to
effectively metabolize maltose (Stolz et al. 1995). Regarding
Lact. acidophilus, the small increase in cell populations
(1Æ21 ± 0Æ12 log10 cfu ml)1) could be explained by the
possible absence of specific nutrients in the malt medium,
such as free amino acids, B-vitamins or minerals. This species
has complex growth requirements (Gomes and Malcata 1999),
usually exhibiting poor growth in synthetic media without the
addition of large amounts of supplements, such as yeast
extract and peptone (Taillandier et al. 1996; Elli et al. 1999).
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
The small amounts of the available sugars consumed
during the exponential phase by all strains (19% for Lact.
fermentum, 17% for Lact. reuteri, 43% for Lact. plantarum
and 16% for Lact. acidophilus) indicated that the sugar
content of the malt medium is not the decisive growth-
limiting factor. In agreement with the above results, Passos
et al. (1994) demonstrated a 45% reduction in sugar during
the Lact. plantarum exponential phase (10 h) in cucumber
juice; the pH was 3Æ53 when growth ceased. In addition,
Venkatesh et al. 1993) reported an incomplete consumption
of the available sugars (17%) and a fast cessation of Lact.
bulgaricus growth (approx. 10 h) in fermentations performed
in synthetic media without pH control, while at constant pH
(5Æ6) a 90% reduction in sugar and a longer exponential phase
(approx. 18 h) were observed. Since all experiments were
performed under uncontrolled conditions, the accumulation
of lactic and acetic acids produced via the metabolic pathways
progressively decreased the pH of the medium. These
organic acids can inhibit microbial growth in their undisso-
ciated form, dissociated form or indirectly by the protons
(H+) that are released in the medium (Passos et al. 1993).
The relatively low concentrations of total lactic acid at the
end of the exponential phase (from 0Æ81 to 1Æ82 g l)1) are not
considered inhibitory to cell growth (Giraud et al. 1991;
Passos et al. 1993; Gänzle et al. 1998). Therefore, the most
significant factor in growth limitation was probably pH
 GROWTH OF LACTIC ACID BACTERIA IN CEREAL-BASED SUBSTRATES
which, at the end of the exponential phase, ranged between
3Æ40 and 3Æ77 for all strains. Thus, the higher cell populations
of Lact. plantarum compared with the other microbes could
be attributed to its ability to continue to grow in malt
medium until the pH value dropped to 3Æ40, similar to the
values reported by Giraud et al. (1991) and Passos et al.
(1993). Lactobacillus plantarum maintains a proton (pH) and
charge gradient between the inside and outside of the cells
even in the presence of large amounts of lactate and protons
(Giraud et al. 1998). No information was found in the
literature regarding the growth-limiting pH values of Lact.
fermentum and Lact. reuteri, which in this study were
3Æ77 ± 0Æ09 and 3Æ72 ± 0Æ09, respectively. Gänzle et al.
(1998) reported a value of 3Æ8 for Lact. sanfranciscensis, a
heterofermentative Lactobacillus, in sourdough fermentation
performed under similar non-nutrient-limiting conditions.
In this study, the limiting pH value for the growth of Lact.
acidophilus was 3Æ83 ± 0Æ09, while Lönner and Åkesson
(1988) reported pH values between 3Æ4 and 3Æ6, depending on
the type of cereal substrate used. However, in that study
these values did not correspond to the end of the exponential
phase. Our results showed that, during the stationary phase,
all strains except Lact. acidophilus continued to consume
sugars and accumulate organic acids (see Tables 2 and 3),
decreasing the pH of the medium to approx. 3Æ10, probably
due to energy requirements for the preservation of cell
viability (Passos et al. 1994). Lactobacillus acidophilus entered
directly into the decline phase, presumably due to its
inability to withstand the low pH of the medium
(3Æ83 ± 0Æ09) at the end of the exponential phase.
The low pH values, compared with the moderate concen-
trations of lactic and acetic acids detected at the same time
(12 h), could be attributed to the low buffering capacity of
the malt medium used in this study (8Æ45 mmol pH)1 l)1),
confirming the significance of substrate composition on the
overall fermentation process. Therefore, the supplementa-
tion of malt medium with additives that enhance its buffering
capacity would result in reduced acidification rates and
increased fermentation times. In milk and vegetable lactic
acid fermentations, citric and acetic acids are added in order
to enhance the buffering capacity of the substrates (Boquien
et al. 1988; Passos et al. 1994). In the prefermented malt
medium acetic acid was not detected, while the amount of
citric acid present in cereals is usually small (Tamine et al.
1997). Other factors contributing to the buffering capacity
could be the protein and ash content of the malt medium
(Marklinder and Johansson 1995).
The similar fermentation patterns of the four strains in
barley and wheat media could be associated with the FAN
content (15Æ3–26Æ6 mg l)1) and sugar composition of the
media, consisting mainly of maltose (0Æ9–1Æ3 g l)1) and
sucrose (1Æ4–2Æ0 g l)1). In accordance with existing reports
(Salovaara and Valjakka 1987; Gänzle et al. 1998; Bvochora
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
857
et al. 1999), the concentrations of these constituents were
significantly lower than those in malt medium (75Æ2–
82Æ1 mg l)1 FAN and 14Æ5–16Æ8 g l)1 total available carbo-
hydrates). Interestingly, the microbial growth ceased at
higher pH values than those observed for malt fermenta-
tions, which suggests that the growth-limiting factor was not
only pH but that deficiency in nutrients also contributed to
growth limitation. A deficiency in specific vitamins or
minerals could contribute to growth limitation but barley
and wheat contain significant amounts of these nutrients
(Palmer 1989). Therefore, the poor growth of Lact. reuteri
(1Æ16 ± 0Æ08 and 1Æ03 ± 0Æ07 log10 cfu ml)1 increase in
barley and wheat, respectively) and Lact. acidophilus
(0Æ70 ± 0Æ08 and 0Æ73 ± 0Æ09 log10 cfu ml)1 increase in
barley and wheat, respectively) could be attributed to the
low concentrations of sugars (mainly glucose and fructose)
and FAN. These observations are in good agreement with
the literature (Marklinder and Lönner 1992). In the case of
Lact. plantarum and Lact. fermentum, although maltose and
sucrose were not completely depleted during the exponential
phase, their initial low concentrations were probably limit-
ing, resulting in lower final cell counts when compared with
malt fermentations. However, these cell populations, ran-
ging between 9Æ1 and 9Æ5 log10 cfu ml)1, still enhance the
potentially probiotic activities of these strains. The increase
in the cell population of Lact. plantarum (2Æ72 ±
0Æ23 cfu ml)1 in barley and 2Æ08 ± 0Æ11 cfu ml)1 in wheat)
and Lact. fermentum (2Æ22 ± 0Æ16 cfu ml)1 in barley and
2Æ35 ± 0Æ09 cfu ml)1 in wheat) suggests that these strains
are less fastidious than Lact. reuteri and Lact. acidophilus
strains, being able to grow better under nutrient-limiting
conditions.
When designing and evaluating a novel fermentation
process, besides considering the factors influencing micro-
bial growth, the distribution of the end metabolic products is
also very important. The organoleptic properties of cereal-
based and dairy fermented foods produced by lactic acid
fermentation are very much dependent on the amounts of
organic acids produced. For example, in wheat sourdough a
molar ratio of lactic to acetic acid from 2 to 2Æ7 is considered
an optimal value for its sensory quality (Gobbetti et al. 1995)
while, in yoghurt, a 0Æ85–0Æ90% lactic acid content is
considered an optimum value (Oberman and Libudzisz
1998). In general, lactic acid is described as ‘acid-sour’ and
acetic acid as ‘acid-sharp’ (De Vuyst 2000).
Theoretically, under strict anaerobic conditions, hetero-
fermentative lactic acid bacteria produce 1 mole of lactic acid
and 1 mole of ethanol per mole of glucose utilized. However,
in the presence of oxygen or compounds that can serve as
electron acceptors, such as fructose, citric and malic acids,
there is a shift of the metabolic pathway towards acetic acid
production instead of ethanol (Stolz et al. 1995). Lactobacillus
fermentum and Lact. reuteri showed a heterofermentative
858 D . C H A R A L A M P O P O U L O S ET AL.
pattern by producing approx. equimolar amounts of lactic
acid and ethanol plus acetic acid (see Table 2). Both strains
indicated an overproduction of acetic acid during the
exponential phase, which could be attributed to the presence
of small amounts of oxygen and fructose. Since acetic acid
production persisted in wheat and barley fermentations (data
not shown), where very small amounts of fructose were
present, it is more likely that acetic acid is formed via oxygen
utilization. In addition, Stolz et al. (1995) demonstrated the
inability of a Lact. reuteri strain isolated from sourdough to
utilize fructose as an electron acceptor. After cell growth
ceased, the formation of lactic acid and ethanol continued
following a heterofermentative pattern, and no acetic acid was
produced, which suggests that there is an association between
acetic acid and growth (Gobbetti et al. 1995). However, the
total product yields based on sugar consumption were very
low, less than 50% (see Table 2), probably due to the fact that
substrate consumption was mainly associated with cell
maintenance and survival (Gonçalves et al. 1997; Fu and
Mathews 1999). The small amount of acetic acid observed
with the homofermentative Lact. plantarum could be attrib-
uted to the presence of oxygen (Murphy and Condon 1984;
Bobillo and Marshall 1991), while the resulting product yield
of 0Æ76 ± 0Æ14 was similar to those found in other studies
(Saucedo et al. 1990; Fu and Mathews 1999). The high
concentration of acetic acid detected in the case of Lact.
acidophilus, shifting the metabolism to an almost heterofer-
mentative pathway, cannot be sufficiently explained.
In this study, the growth characteristics of four potentially
probiotic lactic acid bacteria in cereal substrates were
investigated. A general conclusion is that in a non-pH-
controlled lactic acid fermentation of cereals the main
inhibitors of microbial growth are pH and nutrient limita-
tions, probably sugars or FAN. Further research will aim to
develop a mathematical model that will enable prediction of
the growth of potentially probiotic lactic acid bacteria in
cereal substrates, taking into account the effect of pH and
the substrate composition.
ACKNOWLEDGEMENTS
The authors acknowledge the financial support provided to
D.C. by the Hellenic State Scholarships Foundation. They
are also grateful to Elizabeth Davenport and Janet Wilson
for their help in performing the GC analysis and Konstan-
tinos Stathopoulos for his comments during the preparation
of the manuscript.
REFERENCES
Bobillo, M. and Marshall, V.M. (1991) Effect of salt and culture
aeration on lactate and acetate production by Lactobacillus plantarum.
Food Microbiology 8, 153–160.
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
Boquien, C.Y., Corrieu, G. and Desmazeaud, M.J. (1988) Effect of
fermentation conditions on growth of Streptococcus cremonis AM2
and Leuconostoc lactis CNRZ 1091 in pure and mixed cultures.
Applied and Environmental Microbiology 54, 2527–2531.
Bvochora, J.M., Reed, J.D., Read, J.S. and Zvauya, R. (1999) Effect of
fermentation processes on proanthocyanidins in sorghum during
preparation of Mahewu, a non-alcoholic beverage. Process Biochem-
istry 35, 21–25.
De Vuyst, L. (2000) Technology aspects related to the application of
functional starter cultures. Food Technology and Biotechnology 38,
105–112.
Draper, N.R. and Smith, H. (1981) Checking the straight line fit. In
Applied Regression Analysis, pp. 33–40. New York, USA: John
Wiley.
Elli, M., Zink, R., Reniero, R. and Morelli, L. (1999) Growth
requirements of Lactobacillus johnsonii in skim and UHT milk.
International Dairy Journal 9, 507–513.
Fu, W. and Mathews, A.P. (1999) Lactic acid production from lactose
by Lactobacillus plantarum: kinetic model and effects of pH, substrate
and oxygen. Biochemical Engineering Journal 3, 163–170.
Fuller, R. (1989) Probiotics in man and animals. Journal of Applied
Bacteriology 66, 365–378.
Gänzle, M.G., Ehmann, M. and Hammes, W.P. (1998) Modelling of
growth of Lactobacillus sanfranciscensis and Candida milleri in
response to process parameters of sourdough fermentation. Applied
and Environmental Microbiology 64, 2616–2623.
Giraud, E., Champailler, A., Moulard, S. and Raimbault, M. (1998)
Development of a miniaturized selective counting strategy of lactic
acid bacteria for evaluation of mixed starter in a model cassava
fermentation. Journal of Applied Microbiology 84, 444–450.
Giraud, E., Lelong, B. and Raimbault, M. (1991) Influence of pH and
initial lactate concentration on the growth of Lactobacillus plantarum.
Applied Microbiology and Biotechnology 36, 96–99.
Gobbetti, M., Corsetti, A. and Rossi, J. (1994) The sourdough
microflora. Interactions between lactic acid bacteria and yeasts:
metabolism of carbohydrates. Applied Microbiology and Biotechnology
41, 456–460.
Gobbetti, M., Corsetti, A. and Rossi, J. (1995) Maltose-fructose
co-fermentation by Lactobacillus brevis subsp. linderi CB1 fructose-
negative strain. Applied Microbiology and Biotechnology 42, 939–944.
Gomes, A.M.P. and Malcata, F.X. (1999) Bifidobacterium spp. &
Lactobacillus acidophilus: biological, biochemical, technological and
therapeutical properties relevant for use as probiotics. Trends in Food
Science and Technology 10, 139–157.
Gonçalves, L.M.D., Ramos, A., Almeida, J.S., Xavier, A.M.R.B. and
Carrondo, M.J.T. (1997) Elucidation of the mechanism of lactic acid
growth inhibition and production in batch cultures of Lactobacillus
rhamnosus. Applied Microbiology and Biotechnology 48, 346–350.
Havenaar, R. and Huis in’t Veld, J.H.J. (1992) Probiotics: a general
view. In The Lactic Acid Bacteria ed. Wood, B.J.B. pp. 209–224.
New York: Chapman & Hall.
Lejeune, R., Callewaert, R., Crabbé, K. and de Vuyst, L. (1998)
Modelling the growth and bacteriocin production by Lactobacillus
amylovorus DCE 471 in batch cultivation. Journal of Applied
Microbiology 84, 159–168.
Lönner, C. and Åkesson, P.K. (1988) Acidification properties of lactic
acid bacteria in rye sour doughs. Food Microbiology 5, 43–58.
 GROWTH OF LACTIC ACID BACTERIA IN CEREAL-BASED SUBSTRATES
Magne, C. and Larher, F. (1992) Higher sugar content of extracts
interferes with the colorimetric determination of amino acids and
proline. Analytical Biochemistry 200, 115–118.
Marklinder, I. and Johansson, L. (1995) Sour dough fermentation of
barley flours with varied content of mixed-linked (1fi3), (1fi4)b-D-
glucans. Food Microbiology 12, 363–371.
Marklinder, I. and Lönner, C. (1992) Fermentation properties of
intestinal strains of Lactobacillus, of a sour dough and of a yoghurt
starter culture in an oat-based nutritive solution. Food Microbiology
9, 197–205.
Meraz, M., Shirai, K., Larralde, P. and Revah, S. (1992) Studies on the
bacterial acidification process of cassava (Manihot esculenta). Journal
of the Science of Food and Agriculture 60, 457–463.
Mercier, P., Yerushalmi, L., Rouleau, D. and Dochain, D. (1992)
Kinetics of lactic acid fermentation on glucose and corn by
Lactobacillus amylophilus. Journal of Chemical Technology and
Biotechnology 55, 111–121.
Murphy, M.G. and Condon, S. (1984) Comparison of aerobic and
anaerobic growth of Lactobacillus plantarum in a glucose medium.
Archives in Microbiology 138, 49–53.
Oberman, H. and Libudzisz, Z. (1998) Fermented milks. In Micro-
biology of Fermented Foods ed. Wood, B.J.B. pp. 308–349. London:
Blackie Academic and Professional.
Oyewole, O.B. (1997) Lactic fermented foods in Africa and their
benefits. Food Control 8, 289–297.
Özen, S. and Özilgen, M. (1992) Effects of substrate concentration on
growth and lactic acid production by mixed cultures of Lactobacillus
bulgaricus and Streptococcus thermophilus. Journal of Chemical Tech-
nology and Biotechnology 54, 57–61.
Pai, S.C., Tsa, Y.J. and Yang, T.I. (2001) pH and buffering capacity
problems involved in the determination of ammonia in saline water
using the indophenol blue spectrophotometric method. Analytica
Chimica Acta 434, 209–216.
Palmer, G.H. (1989) Cereals in malting and brewing. In Cereal Science
and Technology ed. Palmer, G.H. pp. 168–174. Aberdeen: University
Press.
ª 2002 The Society for Applied Microbiology, Journal of Applied Microbiology, 92, 851–859
859
Passos, F.V., Fleming, H.P., Ollis, D.F., Felder, R.M. and McFeeters,
R.F. (1994) Kinetics and modeling of lactic acid production by
Lactobacillus plantarum. Applied and Environmental Microbiology 60,
2627–2636.
Passos, F.V., Fleming, H.P., Ollis, D.F., Hassan, H.M. and Felder,
R.M. (1993) Modeling the specific growth rate of Lactobacillus
plantarum in cucumber extract. Applied Microbiology and Biotechno-
logy 40, 143–150.
Salovaara, H. and Valjakka, T. (1987) The effect of fermentation
temperature, flour type, and starter on the properties of sour wheat
bread. International Journal of Food Science 22, 591–597.
Samuel, W.A., Lee, Y.Y. and Anthony, W.B. (1980) Lactic acid
fermentation of crude sorghum extract. Biotechnology and Bioengin-
eering 22, 757–777.
Saucedo, G.C., González, P.B., Revah, S.M., Viniegra, G.G. and
Raimbault, M. (1990) Effect of Lactobacilli inoculation on cassava
(Manihot Esculenta) silage: Fermentation pattern and kinetic
analysis. Journal of the Science of Food and Agriculture 50, 467–
477.
Shortt, C. (1999) The probiotic century: historical and current
perspectives. Trends in Food Science and Technology 10, 411–417.
Stolz, P., Vogel, R.F. and Hammes, W.P. (1995) Utilization of electron
acceptors by lactobacilli isolated from sourdough. II. Lactobacillus
pontis, Lact. reuteri, Lact. amylovorus and Lact. fermentum. Zeitschrift
fur Lebensmittel Untersuchung und Forschung 201, 402–410.
Taillandier, P., Gilis, F., Portugal, F.R., Laforce, P. and Strehaiano, P.
(1996) Influence of medium composition, pH and temperature on
the growth and viability of Lactobacillus acidophilus. Biotechnology
Letters 18, 775–780.
Tamine, A.Y., Muir, D.D., Barclay, M.N.I., Khaskheli, M. and
McNulty, D. (1997) Laboratory-made Kishk from wheat, oat and
barley: 2. Compositional quality and sensory properties. Food
Research International 30, 319–326.
Venkatesh, K.V., Okos, M.R. and Wankat, P.C. (1993) Kinetic model
of growth and lactic acid production from lactose by Lactobacillus
bulgaricus. Process Biochemistry 28, 231–241.