International Journal of Food Microbiology 87 (2003) 17 27

www.elsevier.com/locate/ijfoodmicro

Growth and metabolism of selected strains of

probiotic bacteria in milk
Hilde M. stlie *, Merete H. Helland, Judith A. Narvhus
Department of Food Science, Agricultural University of Norway, P.O. Box 5036, N-1432 As, Norway
Received 28 March 2002; received in revised form 23 October 2002; accepted 14 December 2002

Abstract
Growth and metabolism of five probiotic strains with well-documented health effects were studied in ultra-high temperature
(UHT) treated milk, supplemented with 0.5% (w/v) tryptone or 0.75% (w/v) fructose. The probiotic strains were Lactobacillus
acidophilus La5, Lb. acidophilus 1748, Lactobacillus rhamnosus GG, Lactobacillus reuteri SD 2112 and Bifidobacterium
animalis BB12. Fermentation was followed for 72 h at 37 jC and the samples were analysed for pH, log cfu ml 1, volatile
compounds, organic acids and carbon dioxide. The strains reduced pH from 6.7 to between 3.9 and 4.4 after 24 h of incubation.
All strains attained viable cell counts above 8.7 9.18 log cfu ml 1 after 6 16 h of incubation. The two Lb. acidophilus strains
showed a stable level of viable cells during 12 72 h of incubation but the three other strains showed a reduction of 0.4 1.1 log
cfu ml 1 from 24 to 72 h of incubation. However, all strains showed cell levels between 7.8 and 8.7 log cfu ml 1 after 72 h of
incubation. After 48 h of incubation, the amount of lactic acid produced varied according to strain from 6949 to 14,000 mg
kg 1 and acetic acid produced varied from 0 to 6901 mg kg 1. Three of the strains metabolised citrate but only low amounts of
diacetyl and acetoin were detected within strains, 0.2 0.8 and 6.5 10 mg kg 1, respectively. Carbon dioxide produced varied
from 221 to 3942 mg kg 1 and was connected to the citrate-fermenting ability of the strain used and their carbohydrate
fermentation pathway. Three of the strains produced detectable levels of acetaldehyde and the concentration varied from 9.4 to
12.6 mg kg 1 after 24 h of incubation. All five probiotic strains showed very different profiles of metabolites during
fermentation; however, the two Lb. acidophilus strains were the most alike. Our findings show the importance of controlling the
fermentation time since the probiotic strains produced different amounts of metabolic products according to fermentation time.
D 2003 Elsevier Science B.V. All rights reserved.
Keywords: Fermented milk; Probiotic bacteria; Organic acids; Volatile compounds; CO2

1. Introduction
Probiotic bacteria and their health effects are a
focus of intensive international research. Evidence is
* Corresponding author. Tel.: +47-64-94-85-78; fax: +47-6494-37-89.
E-mail address: hilde.ostlie@inf.nlh.no. (H.M. stlie).

currently accumulating that documents probiotic,

health promoting effects for a few well-characterized
lactic acid bacteria strains (Lee and Salminen, 1995;
Salminen et al., 1996, 1998; Mattila-Sandholm et al.,
1999; Saarela et al., 2000). In order to exert positive
health effects, the microorganisms need to be viable,
active and abundant in the concentration of at least
106 cfu g 1 in the product throughout the specified

0168-1605/03/$ - see front matter D 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0168-1605(03)00044-8

18

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

shelf life according to Samona and Robinson (1991)

and Vinderola et al. (2000).
Over the past two decades, there has been a
worldwide increase in the consumption of dairy
products containing probiotic bacteria isolated from
humans (OSullivan et al., 1992; Saarela et al.,
2000). Focus has generally been on incorporation
of selected strains of Bifidobacterium spp. and
Lactobacillus spp. into milk and fermented milk
products. Although a certain amount of attention
has been directed towards the organoleptic characteristics of the product, most publications concerning
probiotic bacteria have focused on the human health
aspect. To succeed in promoting the consumption of
functional probiotic products, the food industry has
to satisfy the demands of the consumer. All probiotic foods should be safe and have good sensory
properties (Saarela et al., 2000). Safety aspects
reviewed lately by Salminen et al. (1998) and
Saarela et al. (2000) include specifications such as
human origin, nonpathogenicity and antibiotic resistance characteristics. Little information is available
concerning the growth and metabolism of these
strains in milk. Production of organic acids and
volatile compounds directly influences the product
flavour and consistency, and hence, the organoleptic
quality of the product.
The ability to metabolise milk proteins and lactose
varies considerably among probiotic bacteria. In this
work, the growth and metabolism of five probiotic
strains with well-documented health effects were
studied in ultra-high temperature (UHT) semiskimmed milk supplemented with 0.5% tryptone
(w/v) or 0.75% fructose (w/v). Their ability to
produce organic acids, volatile compounds and carbon dioxide will be focused upon.

2. Materials and methods

2.1. Probiotic strains
Lactobacillus rhamnosus GG (ATCC 53103) was
kindly supplied by Valio, Helsinki, Finland; Lactobacillus reuteri SD 2112 from Biogaia Biologics,
Stockholm, Sweden; and Lactobacillus acidophilus
LA-5 and Bifidobacterium animalis BB12 from
Christian Hansen, Oslo, Norway. Lb. acidophilus

NCFB 1748 was obtained from the National Collection of Food Bacteria, Reading, England.
2.2. Preparation of inocula
The strains were subcultured three times in Man
Rogosa Sharpe (MRS) broth (Difco Labs., Detroit,
MI, USA) at 37 jC overnight, before a final
inoculation for making concentrated stock cultures.
Cysteine hydrochloride (0.05%, w/v, Sigma, St.
Louis, MO, USA) was added to MRS broth for
culturing B. animalis BB12. Early stationary phase
cells (17 22 h of incubation) were harvested by
centrifugation at 14,000  g for 10 min at 4 jC. The
pellet was washed once in 0.05 M potassium phosphate buffer (pH 7.0) before resuspension in UHT
milk (1.5% fat, TINE Norwegian Dairies, Oslo,
Norway) to 1:10 of the original volume. UHT milk
with sucrose (5% and 10%, w/v) and without
sucrose were tested as cryoprotectants. The concentrated cultures (10 ) were stored in 3-ml aliquots at
80 jC. The cell counts of the frozen cultures were
analysed after 1 day, 1 month and 6 months of
storage.
2.3. Production of fermented milk
UHT milk (1.5% fat, TINE) was inoculated with
1% (v/v) of the frozen culture. The milk to be
inoculated with Lb. rhamnosus GG was supplemented in advance with 0.75% (w/v) filter-sterilized
fructose (Merck, Darmstadt, Germany) and the milk
for the other strains was supplemented with 0.5%
(w/v) filter-sterilized tryptone (Oxoid, Hampshire,
England). Forty-milliliter aliquots of the inoculated
milk were aseptically distributed in 50-ml sterile
bottles. One bottle was prepared for each sampling
time and was used for all analyses except the CO2
measurement. The inoculated milk was incubated at
37 jC for 0 72 h. For the CO2 measurement, 10-ml
aliquots of the freshly inoculated milk were aseptically distributed into sterile headspace vials (20 CV,
Chromacol, Trumbell, USA) and sealed with sterile
septa (20-CB3, Chromacol) and aluminium crimp
caps (20-ACB3, Chromacol). One vial was prepared
for each sampling time and the vials were incubated
at 37j for 0 72 h. The viable microbial counts, pH,
volatile compounds, organic acids and carbon diox-

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

ide were determined in the incubated milk after 0

12 h (every 2 h), 12 24 h (every 4 h), 48 and 72 h
of incubation.
2.3.1. Viable microorganisms
Samples were diluted in peptone saline water
(0.9%, w/v saline; 0.1%, w/v peptone) and viable
counts of the probiotic strains were determined on
MRS agar (Difco) after anaerobic incubation at 37 jC
for 3 days (BBL GasPakPlus System, Becton Dickinson Microbiology Systems, Sparks, MD, USA).
2.3.2. Chemical analysis
pH was measured during fermentation with a
Radiometer (pHM 92) pH meter with a combined
glass electrode and temperature probe (Radiometer,
Copenhagen, Denmark). The pH meter was calibrated
using standard buffer solutions (Merck) at pH 4.0 and
7.0.
Volatile compounds were analysed by headspace
gas chromatography according to the method of
Narvhus et al. (1998) with the following modifications: Volatile compounds were separated on a GC
column, Supelco SPB-1k: 30 m  0.53 mm I.D., film
thickness 5 Am (Supelco, Bellefonte, PA, USA).
Detector response was monitored by Turbochrom
Chromatography workstation v4.1 (Perkin Elmer,
Norwalk, USA). The analysis was externally calibrated using standard solutions prepared in sterile
(10%, w/v) reconstituted dried milk (TINE). The
following compounds were used as reference: acetaldehyde (Fluka, Buchs, Switzerland), ethanol (Vinmonopolet, Oslo, Norway), acetone (Tokyo Kasei,
Japan), diacetyl (Sigma) and acetoin (Merck).
Organic acids were analysed by high pressure
liquid chromatography (HPLC) using a modification
of the method of Marsili et al. (1981) as described by
Narvhus et al. (1998). One gram of sample was added
0.2 ml 0.5 M H2SO4 and 8 ml acetonitrile (Rathburn
Chemicals, Scotland) and mixed for 30 min. Following centrifugation (1500  g, 10 min) and filtration
(0.2 Am MFS-13 PTFE filter, Advantec MFS, California, USA), 25 Al of the sample was injected onto
an HPLC (Perkin Elmer), with a UV/VIS detector at
210 nm. Separation was achieved using a cation-Hmicroguard precolumn and Aminex HPX-87H cation
exchange column (Bio-Rad Laboratories, California,
USA). The mobile phase used was 5 mM H2SO4 at a

19

flow of 0.4 ml min 1 and the column temperature

was 30 jC. Detector response was recorded by
Turbochrom Chromatography workstation v4.1. (Perkin Elmer). Organic acids were identified according to
their retention times compared with standard solutions
of the following acids: citric, orotic, pyruvic, succinic,
lactic, formic, acetic, uric, propionic and butyric
(Sigma). The analysis was externally calibrated using
mixed standard solutions in deionised water, prepared
as for the samples.
Quantitative analysis of carbon dioxide production was determined by an infrared CO2 gas analyser (ADC 225 Mk3, Analytical development,
Hoddesdon, Hertfordshire, UK) by the method of
Narvhus et al. (1991). A steady stream of nitrogen
(0.3 l min 1) was passed through the detector cell
and the stability of the flow was visually checked by
a Gilmont flowmeter (Cole-Parmer Instrument, Chicago, USA). The control cell was continuously
flushed with CO2-free air. Detector response was
monitored by a Shimatzu C-R3A Chromatopac integrator (Shimatzu, Kyoto, Japan). Calibration of the
detector response to CO2 concentration in the analysis was based upon liberation of CO2 (0 4000 mg
kg 1) from solutions of known concentration of
sodium bicarbonate. The vial to be analysed was
shaken manually for 2 min. Headspace gas samples
for analysis (0.1 0.5 ml, depending on the amount
of CO2 expected) were withdrawn using a 1-ml
disposable plastic syringe and immediately injected
into the rubber tubing supplying nitrogen to the CO2
analyser. Duplicate injections were made from each
vial and the mean value was calculated. The stability of the analysis system was checked hourly
between analyses by injection of a 0.5-ml standard
gas containing 10% CO2 in nitrogen (Hydrogas,
Porsgrunn, Norway).

3. Results
Results presented are based on UHT milk inoculated with concentrated cell culture made of UHT
milk without supplement since the same cell counts
were determined in UHT milk with and without
sucrose after 1 day, 1 month and 6 months of storage
at 80 jC. Survival and storage stability are shown
in Table 1. The growth and metabolism experiment

20

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

Table 1
Survival and storage stability of probiotic strains in UHT milk without supplement and with supplement (5%, 10% sucrose) during freezing at
80 jC for 1 day, 1 month and 6 months of storage
Strains

Viable counts in UHT milk (log cfu ml

1 a

Time of freezing
0 hb

1 day

1 month

6 months

Sucrose (%)

Lb. acidophilus La5

Lb. acidophilus 1748
Lb. rhamnosus GG
Lb. reuteri SD 2112
B. animalis BB12

10

10

10

8.78
10.11
9.08
9.17
8.86

8.63

8.70

8.76

8.84

8.81

8.70

8.79

8.80

9.04
9.05
8.84

9.06
9.08
8.83

9.09
9.02
8.80

8.79
10.18
9.15
9.15
8.93

9.05
9.05
8.93

9.08
9.08
8.87

9.17
9.17
8.75

9.16
9.16
8.71

9.25
9.25
8.76

Experiments in milk are repeated three times while experiments in milk supplemented with sucrose are repeated one time; representative
results are shown.
b
Before freezing.

was replicated three times, and the mean values are

reported.
3.1. Viable microorganisms
Preliminary growth studies of five probiotic strains
in milk supplemented with 0 1% tryptone or fructose showed that the growth varied considerably
according to the concentration of the added supplements (results not shown). Lb. rhamnosus GG was
unable to ferment lactose, and B. animalis BB12 and
Lb. reuteri SD 2112 did not grow well in milk not
added tryptone. A supplementation of 0.5% tryptone
was selected for studying the growth and metabolism
of Lb. acidophilus La5, Lb. acidophilus 1748, Lb.
reuteri SD 2112 and B. animalis BB12 in UHT milk.
For growth studies of Lb. rhamnosus GG in UHT
milk, 0.75% fructose was added. Growth of the
strains is shown in Fig. 1a. The initial viable cell
counts were between 7.7 and 8.51 log cfu ml 1. All
strains attained viable cell numbers above 8.7 9.18
log cfu ml 1 after 6 16 h of incubation. Lb. reuteri
SD 2112 showed the fastest growth of all strains,
showing maximum viable cells after only 6 h of
incubation. The two strains of Lb. acidophilus
showed a stable level of viable cells during 12 72
h of incubation. The other strains showed a reduction
of 0.4 1.1 log cfu ml 1 from 24 to 72 h of
incubation, most markedly for Lb. reuteri SD 2112
which showed a decrease of 1.1 log cfu ml 1
between 48 and 72 h of incubation.

3.2. pH
Depending on the probiotic strain used, the pH
of the UHT milk decreased from 6.7 initially to
3.9 4.4 after 24 h of incubation (Fig. 1b). Milk
inoculated with Lb. reuteri SD 2112, B. animalis
BB12 and Lb. rhamnosus GG had a pH of 3.9 4.1
after 24 h of incubation, and this remained stable
during further incubation. Lb. acidophilus 1748 and
Lb. acidophilus La5 decreased the pH on prolonged incubation from 4.2 and 4.4 to 3.7 and 3.9,
respectively.
3.3. Organic acids, volatile organic compounds and
carbon dioxide
The amount of lactic acid produced varied according to strain (Fig. 2a), from 6949 to 14,000 mg kg 1
after 48 h of incubation. The homofermentative Lb.
acidophilus La5 and Lb. acidophilus 1748, and the
heterofermentative Lb. reuteri SD 2112 produced the
highest amounts ranging from 9993 to 14,000 mg
kg 1. B. animalis BB12 and Lb. rhamnosus GG
produced the lowest amounts, ranging from 6949 to
7820 mg kg 1. Lb. reuteri SD 2112 and Lb. rhamnosus GG ceased lactic acid production completely after
24 h of incubation whereas the other three strains
continued producing lactic acid during further incubation. In fact, the two Lb. acidophilus strains nearly
doubled the lactic acid production during 24 72 h of
incubation.

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

21

Fig. 1. Extent of (a) growth and (b) changes in pH during fermentation at 37 jC with Lb. acidophilus La5 (x), Lb. acidophilus 1748 (n),
B. animalis BB12 (E) and Lb. reuteri SD2112 (.) in UHT milk supplemented with 0.5% (w/v) tryptone, and with Lb. rhamnosus GG (5) in
UHT milk supplemented with 0.75% (w/v) fructose. Vertical lines represent standard deviations.

Acetic acid levels were highest in milk inoculated

with B. animalis BB12, which produced 6901 mg
kg 1 after 48 h of incubation (results not shown). Lb.
acidophilus La5, Lb. acidophilus 1748 and Lb. reuteri
SD2112 produced 590 1010 mg kg 1 acetic acid
after 48 h of incubation, and Lb. rhamnosus GG did
not produce any acetate during fermentation.
The concentrations of pyruvate, orotic acid, succinic
acid and uric acid were also monitored during fermentation (results not shown). Pyruvate increased during
early fermentation (0 20 h) for all strains, but
decreased after prolonged incubation. Lb. acidophilus
1748 produced most pyruvate, 142 mg kg 1 after 20 h.
The other strains investigated produced from 5 to 31
mg kg 1 and these maximum values were achieved

between 5 and 12 h. A reduction in the level of orotic

acid from approximately 80 to 40 50 mg kg 1 was
observed during fermentation for Lb. acidophilus La5,
Lb. acidophilus 1748 and B. animalis BB12. Lb.
rhamnosus GG showed an even greater reduction,
down to 24 mg kg 1, but Lb. reuteri SD 2112 showed
a smaller reduction from 80 to 67 mg kg 1. Succinic
acid increased during fermentation for all strains from
initially 750 to 1017 1960 mg kg 1 and these maximum values were achieved between 16 and 72 h. The
concentration of uric acid was stable at approximately
16 mg kg 1 during the 72-h fermentation for all strains.
The initial level of citric acid in the milk was
1967 2223 mg kg 1 (Fig. 2b). Three of the strains
were able to metabolise citrate, but the speed of

22

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

Fig. 2. Production of (a) lactic acid and development of (b) citric acid during fermentation at 37 jC with Lb. acidophilus La5 (x),
Lb. acidophilus 1748 (n), B. animalis BB12 (E) and Lb. reuteri SD2112 (.) in UHT milk supplemented with 0.5% (w/v) tryptone, and with
Lb. rhamnosus GG (5) in UHT milk supplemented with 0.75% (w/v) fructose. Vertical lines represent standard deviations.

reduction varied considerably. Lb. reuteri SD2112

metabolised citrate completely within 4 6 h but Lb.
acidophilus 1748 and Lb. acidophilus La5 had metabolised nearly all citrate after 20 and 48 h, respectively.
Three of the strains produced acetaldehyde (Fig.
3a). Concentrations of acetaldehyde produced by the
homofermentative Lb. acidophilus La5 and Lb. acidophilus 1748 were 9.7 and 12.6 mg kg 1, respectively, after 24 h of incubation. On further incubation
the acetaldehyde production of these strains increased
considerably, to 71 and 116 mg kg 1, respectively,
after 72 h. In addition, B. animalis BB12 produced 9.4
mg kg 1 acetaldehyde after 10 h of incubation, and
this level remained relatively stable during further
incubation.

Heterofermentative Lb. reuteri SD2112 produced

considerably higher amounts of ethanol than the
other strains (Fig. 3b). After 72 h of incubation,
Lb. reuteri SD2112 produced 7050 mg kg 1 of
ethanol, while the other strains produced 5 40 mg
kg 1 of ethanol.
Diacetyl and acetoin are normally produced at rates
dependent upon the rate of citrate metabolism.
Although Lb. acidophilus La5, Lb. acidophilus 1748
and Lb. reuteri SD 2112 metabolised citrate, only low
amounts of diacetyl were detected with these strains
(0.2 0.8 mg kg 1). In addition, Lb. rhamnosus GG
(citrate negative) produced 0.6 mg kg 1 diacetyl after
8 h of incubation. Diacetyl was reduced completely for
all strains after 72 h of incubation. B. animalis BB12

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

23

Fig. 3. Production of (a) acetaldehyde, (b) ethanol and (c) carbon dioxide during fermentation at 37 jC with Lb. acidophilus La5 (x),
Lb. acidophilus 1748 (n), B. animalis BB12 (E) and Lb. reuteri SD2112 (.) in UHT milk supplemented with 0.5% (w/v) tryptone, and with
Lb. rhamnosus GG (5) in UHT milk supplemented with 0.75% (w/v) fructose. Vertical lines represent standard deviations.

did not produce diacetyl. No acetoin production was

detected for Lb. acidophilus La5 and Lb. acidophilus
1748 during the 72-h incubation period. The other
strains tested did not produce more than 10 mg kg 1
acetoin at any time during the incubation period (results
not shown) regardless of their ability to degrade citrate.
Carbon dioxide produced by the strains tested
varied from 221 to 3942 mg kg 1 (Fig. 3c) and was

clearly connected with whether the individual strain

could degrade citrate and also which carbohydrate
fermentation pathway was used. The heterofermentative and citrate-degrading Lb. reuteri SD 2112 produced the highest amount of carbon dioxide (3942 mg
kg 1). The homofermentative and citrate-degrading
Lb. acidophilus La5 and Lb. acidophilus 1748 produced 497 586 mg kg 1 CO2. The homofermenta-

24

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

tive, citrate-negative Lb. rhamnosus GG and B. animalis BB12 produced the lowest amounts of CO2, 221
and 240 mg kg 1, respectively.

4. Discussion
Production of the frozen concentrated probiotic
cultures was successful since the storage stability
was excellent for cultures frozen at
80 jC for 6
months, both with and without added cryoprotectants.
Saxelin et al. (1999) studied survival of frozen probiotic cultures with cryoprotectants added and
observed that the storage temperature needs to be
below
35 jC in order for the frozen cultures to
maintain a high cell count for 12 months.
There is limited published information concerning
the technological production of fermented probiotic
milk products and of the metabolic pathways followed
by specific probiotic organisms during the fermentation of milk. The production of organoleptically acceptable and stable probiotic fermented milks can be
improved by knowledge of which important metabolic
products are produced and in what concentrations.
Production of acceptable and stable fermented milk
products using probiotic lactic acid bacteria is a major
challenge to dairies as milk is not, on the whole, a
good growth medium for these organisms. To produce
a satisfactory fermented probiotic milk product, the
viable cell count should be above 6 log cfu g 1 in
order to supply a sufficient daily dose of viable
bacteria (Samona and Robinson, 1991; Vinderola et
al., 2000). In addition, the lactate concentration should
be about 8000 mg kg 1 and the pH between 4.2 and
4.4 in order for the organoleptic qualities of sourness
and firm coagulum to be satisfactory (Narvhus, 1996).
Lb. acidophilus 1748, Lb. acidophilus La5, Lb. reuteri
SD 2112 and B. animalis BB12 needed tryptone as a
supplement and Lb. rhamnosus GG needed fructose as
supplement to the milk. The optimum concentration
of these supplements was found in preliminary studies
(unpublished information) and under these conditions
all the strains showed growth satisfactory for the
production of a probiotic fermented milk. Differences
in incubation time required to reach pH 4.2 4.4 was
observed between organisms. Lb. reuteri SD 2112
needed 8 10 h of incubation, Lb. rhamnosus GG and
B. animalis BB12 needed 16 h, and the Lb. acid-

ophilus strains 1748 and La5 needed 20 and 24 h,

respectively.
Lactose is the primary substrate for acid production
in milk. However, its fermentation is not limited by
the amount of lactose available, but by the production
of lactic acid and the concomitant lowering of pH
which increasingly inhibits the starter organisms long
before the lactose is exhausted (Narvhus et al., 1998).
Homofermentative lactic acid bacteria (LAB) strains
metabolise lactose to lactic acid and heterofermentative LAB strains metabolise lactose to lactate, ethanol,
acetate and carbon dioxide. Bifidobacteria metabolise
lactose to lactate and acetate (bifidus pathway). Lb.
acidophilus and Lb. rhamnosus ferment lactose homofermentatively, while Lb. reuteri ferments lactose
heterofermentatively and B. animalis ferments lactose
by the bifidus pathway (Kandler and Weiss, 1986;
Scardovi, 1986). Different amounts of lactic acid were
produced by the strains studied; however, after 24 h of
incubation, all strains except B. animalis BB12 had
produced above 8000 mg kg 1. After 48 and 72 h of
incubation, B. animalis BB12 produced 6949 and
7825 mg kg 1 lactic acid, respectively. Such an
extended period would be inconvenient from the point
of view of product production. Acetic acid can be
produced by Lb. acidophilus from degradation of
citrate and by Lb. reuteri from degradation of citrate
and by heterofermentation. Bifidobacterium spp. produce acetate using the bifidus pathway. Approximately 3:2 mol of acetic/lactic acid were produced
after 48 h in our study, observations that concur with
the theoretical amounts reported by Scardovi (1986).
The two homofermentative, citrate-degrading organisms, Lb. acidophilus La5 and Lb. acidophilus 1748,
produced 590 642 mg kg 1 of acetic acid after 48 h
which is near the theoretical amount that would be
produced from 2000 to 2200 mg kg 1 citrate in milk.
Citrate-degrading, heterofermentative Lb. reuteri SD
2112 produced more acetic acid (1010 mg kg 1)
because of heterofermentation of lactose in addition
to citrate degradation.
According to Okonkwo and Kinsella (1969), orotic
acid is reduced by ca. 50% during yoghurt fermentation, from 83 to 34 42 mg kg 1. A decrease (30%)
has also been reported in yoghurt by FernandezGarcia and McGregor (1994). The same tendency
was observed in this study. Lb. rhamnosus GG
showed the highest reduction of orotic acid, from

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

approximately 80 down to 24 mg kg 1, and Lb.

reuteri SD 2112 showed the least reduction, from 76
to 67 mg kg 1. Orotic acid is an intermediate in the
synthesis of nucleotides and also a growth factor for
Lactobacillus delbrueckii subsp. bulgaricus and
Streptococcus salvarius subsp. thermophilus in
yoghurt starter cultures (Fernandez-Garcia and
McGregor, 1994). According to Tamime and Robinson (1999), the level of succinic acid in yoghurt
increases during fermentation, and the same was
noticed for all of the probiotic strains in our experiments. Succinic acid production is fairly common
among heterofermentative lactobacilli isolated from
the intestine of several animals (Axelsson, 1998).
Axelsson (1998) proposed a pathway for succinic
acid production in heterofermentative lactobacilli
growing on glucose and citrate, and the end products
formed are lactate, acetate, carbon dioxide and succinate. Succinate was also produced in MRS broth from
citrate by a number of Lactobacillus strains (Kaneuchi
et al., 1988).
Lactic acid bacteria produce varying amounts of
acetaldehyde and ethanol during growth (Gonzalez et
al., 1994). The accumulation of acetaldehyde in the
growth medium depends on whether the organism has
active enzymes that convert this compound to other
metabolites, principally ethanol (Gonzalez et al.,
1994). The amount of acetaldehyde required for the
development of a characteristic flavour varies widely
among dairy products. In cheese, cultured buttermilk
and butter, acetaldehyde is considered to be an undesirable compound giving an atypical taste whereas in
yoghurt, relatively large amounts are required to give
the characteristic flavour of that product (Tamime and
Robinson, 1999). In the present study, Lb. acidophilus
La5, Lb. acidophilus 1748 and B. animalis BB12
produced acetaldehyde in amounts that would have
an influence on the sensory profile of the product, and
Lb. acidophilus is of special interest for its ability to
produce both acetaldehyde and diacetyl. After 24 h of
incubation, Lb. acidophilus La5 and Lb. acidophilus
1748 had produced 9.7 and 12.6 mg kg 1 acetaldehyde, respectively, values which are similar to those
found in commercial yoghurt (Tamime and Robinson,
1999). However, on further incubation, the levels of
acetaldehyde increased greatly to 70.7 and 116 mg
kg 1. Such levels would probably not be organoleptically acceptable and these results therefore indi-

25

cate that the acetaldehyde production by these strains

must be limited by controlling the fermentation time.
The accumulation of acetaldehyde in the growth
medium can occur when the specific activities of
enzymes which form acetaldehyde are higher than
those able to convert it to ethanol (Gonzalez et al.,
1994). Gonzalez et al. (1994) studied acetaldehyde
production by probiotic bacteria in fermented milk
and found that Lb. acidophilus can form acetaldehyde
from different sources such as carbohydrates, amino
acids (such as threonine) and nucleic acids. Marshall
and Cole (1983) examined threonine metabolism as a
pathway for producing acetaldehyde and found that
strains of Lb. acidophilus contained modest activities
of threonine aldolase and alcohol dehydrogenase.
They found that addition of threonine to the milk
fortified with whey proteins caused Lb. acidophilus to
produce a well-flavoured product with a taste and
aroma similar to that of yoghurt. They reported an
acetaldehyde production by Lb. acidophilus of 3 41
mg kg 1, which was not the case in milk without
threonine addition. The tryptone added to the milk in
our experiments contains 1.87% (w/w) threonine
(Oxoid Manual, 1995). Three of the strains incubated
in milk supplemented with tryptone showed acetaldehyde production and the amounts reported from two
of the strains after 24 48 h of incubation are close to
the theoretical amount (34.5 mg kg 1) produced from
tryptone. Lb. rhamnosus GG incubated in milk supplemented with fructose did not produce any acetaldehyde during fermentation.
Another very important compound determining the
specific characteristics of fermented milks is diacetyl
(Oberman and Libudzisz, 1998). At very low concentration (up to 5 mg kg 1), diacetyl is responsible for
the characteristic buttery aroma in milk products
(Oberman and Libudzisz, 1998). Milk contains an
average of 0.2% citrate which is co-metabolised
together with sugars by many lactic acid bacteria
(Cocaign-Bousquet et al., 1996). The ability to metabolise citrate to CO2, acetate, diacetyl, acetoin and
sometimes 2,3-butandiol is a common trait of many
lactic acid bacteria, especially Lactococcus and Leuconostoc spp. (Hugenholtz, 1993). All the lactic acid
bacteria that can produce diacetyl are also able to
reduce it to butandiol (Hugenholtz, 1993). Diacetyl is
irreversibly reduced to acetoin by acetoin reductase,
and in a further step, acetoin is reversibly reduced to

26

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

butandiol by the same enzyme (Hugenholtz and

Starrenburg, 1992). The production rates and amounts
of diacetyl and acetoin in mesophilic starter cultures
are dependent upon the rate of citrate metabolism.
Three of our tested strains metabolised citrate but only
low amounts of diacetyl and acetoin were produced
by these strains, 0.2 0.8 and 6.5 10 mg kg 1,
respectively. It is possible that diacetyl and acetoin
were further metabolised to 2,3-butandiol, which was
not measured in this study. The citrate may also have
been metabolised to pyruvate and then further to other
products. The loss of flavour in fermented milk
products due to the reduction of diacetyl is a problem
and is affected by the redox potential of the milk and
the requirement for the regeneration of NAD (Oberman and Libudzisz, 1998). Several approaches have
been used in order to achieve high levels of diacetyl in
certain fermented milk products. The addition of
citrate to the milk, either pre- or post-fermentation,
immediate cooling of the fermented milk before
complete citrate utilization, temperature variation,
aeration and strain selection have been examined as
mechanisms for increasing the level of diacetyl in the
fermented milk product (Rattray et al., 2000). However, aeration of milk as a way to increase the amount
of diacetyl formed by mesophilic starter cultures
would be unsuitable in the case of the less oxygentolerant probiotic strains (Hugenholtz, 1993). In complex or synthetic media, Lb. acidophilus produced
more diacetyl or acetoin in the presence of pyruvate
and also by elevation of temperature, reaching an
optimum at 45 jC (Benito de Cardenas et al., 1989,
1990, 1991).
In the present work, growth and metabolism
(investigated by quantification of key metabolites)
were followed for 72 h to investigate the stability of
the probiotic cells and their metabolites during fermentation. Normally, growth and metabolism is
mostly stopped after 24 h of incubation; however,
for the lactic acid bacteria, it is known that growth and
metabolism can be uncoupled. Lb. acidophilus La5
and Lb. acidophilus 1748 clearly show this phenomenon. After 24 h of incubation, the log cfu ml 1 was
stable for both strains; however, the production of
lactic acid continued and the production of acetaldehyde also increased during further incubation.
The survival and viable cell count of probiotic
bacteria vary depending on the strains and manufac-

turers (Saxelin et al., 1999; Shillinger, 1999). In this

investigation, the survival of the probiotic strains in
the fermented milk was excellent during the 72-h
fermentation. All strains showed cell levels between
7.8 and 8.7 log cfu ml 1. The various probiotic
strains tested in this study had very different metabolic profiles in fortified milk with respect to known
organoleptically important compounds; however, the
two Lb. acidophilus strains were most alike. These
differences in profiles would undoubtedly affect the
sensory quality of products made using these different
organisms. Our findings show the importance of
controlling the fermentation time since the probiotic
strains produced different amount of metabolic products after a defined fermentation time. Further studies
will address the shift in metabolism of the probiotic
bacteria in response to environmental changes and
also the effect of different milk treatment regimes on
the metabolism of probiotic bacteria in milk.
Acknowledgements
The authors wish to acknowledge the expert
technical help of Kari Olsen for the GC and HPLC
analyses. This work was financed by the Norwegian
Research Council.
References
Axelsson, L., 1998. Lactic acid bacteria: classification and physiology. In: Salminen, S., von Wright, A. (Eds.), Lactic Acid
Bacteria. Marcel Dekker, New York, pp. 1 72.
Benito de Cardenas, I.L., Ledesma, O., Oliver, G., 1989. Effect of
pyruvate on diacetyl and acetoin production by Lactobacillus
casei and Lactobacillus plantarum. Milchwissenschaft 44,
347 350.
Benito de Cardenas, I.L., Cerutti de Guglielmone, G., Ledesma,
O., Oliver, G., 1990. Diacetyl and acetoin production from
pyruvate by cell suspensions of lactobacilli. Milchwissenschaft
45, 775 777.
Benito de Cardenas, I.L., Peral de Portillo, M.C., Albizzatti de
Rivadeneira, M.C., Oliver, G., 1991. Diacetyl and acetoin production by thermophilic starter cultures. M.A.N. Microbiol. Aliment. Nutr. 9, 369 374.
Cocaign-Bousquet, M., Garrigues, C., Loubiere, P., Lindley, N.D.,
1996. Physiology of pyruvate metabolism in Lactococcus lactis.
Antonie van Leeuwenhoek 70, 253 267.
Fernandez-Garcia, E., McGregor, J.U., 1994. Determination of organic acids during the fermentation and cold storage of yogurt.
J. Dairy Sci. 77, 2934 2939.

H.M. stlie et al. / International Journal of Food Microbiology 87 (2003) 1727

Gonzalez, S., Morata de Ambrosini, V., Manca de Nadra, M., Pesce
de Ruiz Holgado, A., Oliver, G., 1994. Acetaldehyde production
by strains used as probiotics in fermented milk. J. Food Prot. 57,
436 440.
Hugenholtz, J., 1993. Citrate metabolism in lactic acid bacteria.
FEMS Microbiol. Rev. 12, 165 178.
Hugenholtz, J., Starrenburg, M.J.C., 1992. Diacetyl production by
different strains of Lactococcus lactis subsp. lactis var. diacetylactis and Leuconostoc spp. Appl. Microbiol. Biotechnol. 38,
17 22.
Kandler, O., Weiss, N., 1986. Genus Lactobacillus Beijerinck 1901,
212AL. In: Sneath, P.H.A., Mair, N.S., Sharpe, M.E., Holt, J.G.
(Eds.), Bergeys Manual of Systematic Bacteriology, vol. 2.
Williams and Wilkins, Baltimore, MD, pp. 1208 1234.
Kaneuchi, C., Seki, M., Komagata, K., 1988. Production of succinic
acid and related acids by Lactobacillus strains. Appl. Environ.
Microbiol. 54, 3053 3056.
Lee, Y.-K., Salminen, S., 1995. The coming of age of probiotics.
Trends Food Sci. Technol. 6, 241 245.
Marshall, V.M., Cole, W.M., 1983. Threonine aldolase and alcohol
dehydrogenase activities in Lactobacillus bulgaricus and Lactobacillus acidophilus and their contribution to flavour production
in fermented milks. J. Dairy Res. 50, 375 379.
Marsili, R.T., Ostapenko, H., Simmons, R.E., Green, D.E., 1981.
High performance liquid chromatography of organic acids in
dairy products. J. Food Sci. 46, 52 57.
Mattila-Sandholm, T., Blum, S., Collins, J.K., Crittenden, R., de
Vos, W., Dunne, C., Fonden, R., Grenov, G., Isolauri, E., Kiely,
B., Marteau, P., Morelli, L., Ouwehand, A., Reniero, R., Saarela,
M., Salminen, S., Saxelin, M., Schiffrin, E., Shanahan, F.,
Vaughan, E., von Wright, A., 1999. Probiotics: towards demonstrating efficacy. Trends Food Sci. Technol. 10, 393 399.
Narvhus, J., 1996. Probiotiske bakterier-metabolisme i melk. Meieriposten 12, 341 343.
Narvhus, J.A., Hulbkdal, A., Baugerd, H., Abrahamsen, R.K.,
1991. Measurement of CO2 production and O2 metabolism by
pure and mixed cultures of lactic acid bacteria growing in milk.
Actes du colloque Lactic, vol. 91. Centre de Publications de
lUniversite de Caen, Caen, France, p. 371.
Narvhus, J.A., steraas, K., Mutukumira, T., Abrahamsen, R.K.,
1998. Production of fermented milk using a malty compoundproducing strain of Lactococcus lactis subsp. lactis biovar. diacetylactis isolated from Zimbabwean naturally fermented milk.
Int. J. Food Microbiol. 41, 73 80.

27

Oberman, H., Libudzisz, Z., 1998. Fermented milks. In: Wood,

B.J.B. (Ed.), Microbiology of Fermented Foods. Blackie Academic and Professional, London, pp. 309 350.
Okonkwo, P., Kinsella, J.E., 1969. Orotic acid in yoghurt. J. Dairy
Sci. 52, 1861 1862.
OSullivan, M.G., Thornton, G., OSullivan, G.C., Collins, J.K.,
1992. Probiotic bacteria: myth or reality. Trends Food Sci. Technol. 3, 309 314.
Oxoid Manual, 1995. Bridson, E.Y. (compiler), 7th ed. Unipath,
Hampshire, England.
Rattray, F.P., Walfridsson, M., Nilsson, D., 2000. Purification and
characterization of a diacetyl reductase from Leuconostoc pseudomesenteroides. Int. Dairy J. 10, 781 789.
Saarela, M., Mogensen, G., Fonden, R., Matto, J., Mattila-Sandholm, T., 2000. Probiotic bacteria: safety, functional and technological properties. J. Biotechnol. 84, 197 215.
Salminen, S., Isolauri, E., Salminen, E., 1996. Clinical uses of probiotics for stabilizing the gut mucosal barrier: successful strains
and future challenges. Antonie van Leeuwenhoek 70, 347 358.
Salminen, S., von Wright, A., Morelli, L., Marteau, P., Brassart, D.,
de Vos, W.M., Fonden, R., Saxelin, M., Collins, K., Mogensen,
G., Birkeland, S.-E., Mattila-Sandholm, T., 1998. Demonstration of safety of probioticsa review. Int. J. Food Microbiol.
44, 93 106.
Samona, A., Robinson, R.K., 1991. Enumeration of bifidobacteria
in dairy products. J. Soc. Dairy Technol. 44, 64 66.
Saxelin, M., Grenov, B., Svensson, U., Fonden, R., Reniero, R.,
Mattila-Sandholm, T., 1999. The technology of probiotics.
Trends Food Sci. Technol. 10, 387 392.
Scardovi, V., 1986. Genus Bifidobacterium Orla Jensen 1924,
472AL. In: Sneath, P.H.A., Mair, N.S., Sharpe, M.E., Holt, J.G.
(Eds.), Bergeys Manual of Systematic Bacteriology, vol. 2. Williams and Wilkins, Baltimore, MD, pp. 1418 1434.
Shillinger, U., 1999. Isolation and identification of lactobacilli from
novel-type probiotic and mild yoghurts and their stability during
refrigerated storage. Int. J. Food Microbiol. 47, 79 87.
Tamime, A.Y., Robinson, R.K., 1999. Biochemistry of fermentation. In: Tamime, A.Y., Robinson, R.K. (Eds.), Yoghurt: Science
and Technology. Pergamon, Oxford, England, pp. 433 485.
Vinderola, C.G., Bailo, N., Reinheimer, J.A., 2000. Survival of probiotic microflora in Argentinean yoghurts during refrigerated
storage. Food Res. Int. 33, 97 102.