International Journal of Food Microbiology 111 (2006) 280 – 287

www.elsevier.com/locate/ijfoodmicro

Short communication
Rapid discrimination of lactobacilli isolated from kefir grains
by FT-IR spectroscopy
Alejandra Bosch a , Marina A. Golowczyc b , Analía G. Abraham b , Graciela L. Garrote b ,
Graciela L. De Antoni b , Osvaldo Yantorno a,⁎
a
Centro de Investigación y Desarrollo de Fermentaciones Industriales, CINDEFI, CONICET y Facultad de Ciencias Exactas,
UNLP, calle 50 y 115, La Plata (1900), Argentina
b
Centro de Investigación y Desarrollo en Criotecnología de Alimentos, CIDCA (CONICET y Facultad de Ciencias Exactas,
UNLP), 47 y 116, La Plata (1900), Argentina
Received 2 December 2005; received in revised form 3 May 2006; accepted 27 May 2006

Abstract

Fourier transform infrared (FT-IR) spectroscopy was used in combination with multivariate statistical analysis for differentiation of lactic bacteria
isolated from kefir grains. Twelve reference strains and 42 lactobacilli isolates from four local kefir grains, previously identified by biochemical
traditional techniques at species level were included in this study. The spectra were analysed by hierarchical clustering analysis (HCA) using
Pearson's product–moment correlation coefficient and Ward's algorithm. The differentiation between homo- and heterofermentative lactobacilli,
proposed as a first level in the classification scheme, was performed with vector normalized first derivatives spectra in the windows 1789–1700,
1059–935, 3000–2927 and 896–833 cm− 1. For heterofermentative lactobacilli the windows 1780–1750, 1500–1200, 2950–2930 and 900–
700 cm− 1 were found to contribute to the maximal separation among L. kefir, L. parakefir and Lactobacillus brevis. It was also demonstrated that
although this model was robust against small variations in growth temperature (± 5 °C) and growth time (± 5 h), the make of culture medium used
(Biokar or Difco) affected the separation of heterofermentative lactobacilli at species level. For homofermentative lactobacilli the spectral regions
1230–900, 1777–1690, 1357–1240 and 2960–2870 cm− 1, were selected for discrimination among 5 different species that are normally present in
kefir grains: L. plantarum, L. acidophilus, L. kefirgranum, L. kefiranofaciens and L. cassei. The classification and discrimination schemes proposed
in this work completely matched with the identification obtained by classical biochemical techniques at species level.
© 2006 Elsevier B.V. All rights reserved.

Keywords: Kefir grains; Lactobacillus; Discrimination; FT-IR spectroscopy

1. Introduction Marshall et al., 1984), L. plantarum (Garrote et al., 2001,

Kefir is a type of sour fermented milk in which kefir grains are
employed as starters. These grains are white or lightly yellowish
irregular masses in which bacteria and yeast are contained in a
matrix of proteins and polysaccharides. The most common
lactobacilli isolated from kefir grains comprise: Lactobacillus
brevis (Ottogalli et al., 1973; Rosi and Rossi, 1978; Marshall et al.,
1984; Angulo et al., 1993), L. casei (Molska et al., 1983; Angulo
et al., 1993), L. kefir (Kandler and Kunath, 1983; Marshall et al.,
1984; Angulo et al., 1993; Pintado et al., 1996, Garrote et al.,
2001), L. acidophilus (Ottogalli et al., 1973; Angulo et al., 1993;
⁎ Corresponding author. Tel./fax: +54 221 483 3794.
E-mail address: yantorno@quimica.unlp.edu.ar (O. Yantorno).
0168-1605/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijfoodmicro.2006.05.010
Witthuhn et al., 2005a,b) L. kefiranofaciens (Fujisawa et al., 1988;
Toba et al., 1991; Mukai et al., 1992), L. kefirgranum and L.
parakefir (Takizawa et al., 1998, Garrote et al., 2001).
We have recently reported the chemical and microbiological
composition of local kefir grains obtained from different
sources. Biochemical tests and whole-cell protein pattern pro-
files were employed to characterize and identify the microflora
present in these grains (Garrote et al., 2001). However, as these
methodologies are tedious, expensive and time-consuming the
application of other techniques should be considered.
Fourier transformed infrared (FT-IR) spectra of intact
bacteria are highly specific patterns which may be unique for
individual strains (Helm et al., 1991; Naumann et al., 1991). FT-
IR spectroscopy is easy to implement, allows analysis of small
 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287 281

quantities of biomass and requires no specific consumables or
reagents (Helm et al., 1991; Naumann et al., 1991; Naumann,
2000; Maquelin et al., 2002). FT-IR spectroscopy has been
applied in the identification of lactic acid bacteria present in
breweries, soft cheese and meat (Curk et al., 1994; Amiel et al.,
2000; Oust et al., 2004a). The aim of this work was to develop
an approach based on FT-IR spectroscopy in combination with
multivariate statistical analysis for rapid differentiation of
lactobacilli isolated from kefir grains.
2. Materials and methods
2.1. Microorganisms
A total of forty-two lactobacilli isolated from four local kefir
grains belonging to CIDCA (Centro de Investigación y
Desarrollo en Criotecnología de Alimentos, Argentina) collec-
tion, were used in this study (Table 1). According to their ability
to produce gas from glucose, these lactobacilli were classified as
heterofermentative (21 isolates) or homofermentative (21
isolates). These bacteria were identified at the species level by
biochemical tests and whole-cell protein patterns by SDS-
PAGE (Garrote et al., 2001). In addition twelve reference strains
(Table 1) chosen according to the Lactobacillus species
previously reported in different kefir grains (Marshall et al.,
1984; Angulo et al., 1993; Simova et al., 2002) were included.
2.2. Growth conditions and sample preparation
(Bruker Optics, Germany) (Helm et al., 1991) were used in the
analysis. The first derivative of every spectrum was calculated
using Savitzky–Golay algorithm with 9 smoothing points to
increase the number of discriminative features present in the
spectra. In order to avoid variation due to differences in biomass
among the different samples, the first derivative of each respec-
tive spectrum was vector normalized in the full range (Naumann
et al., 1991; Mariey et al., 2001).
All data were subjected to multivariate statistical analysis
using the cluster analysis module of OPUS software version 4.0
(Bruker Optics, Germany). Hierarchical clustering analyses
(HCA) using the first derivatives of the original spectra as input,
were carried out looking for the best combination of wave
number regions that produced the desirable discrimination.
Ward's algorithm was applied to construct dendrograms, using
Pearson's product–moment correlation coefficient as a distance
measure (Helm et al., 1991; Naumann, 2000; Kirschner et al.,
2001).
3. Results and discussion
3.1. Discrimination between homo- and heterofermentative
lactobacilli spectra
When almost 400 absorption spectra acquired from twelve
reference strains and 42 different lactobacilli isolated from kefir
grains were overlaid, the spectra showed two main different
features, which corresponded to homofermentative and

All bacteria were grown on MRS agar (Biokar, France) Table 1

plates at 30 °C for 48 ± 2 h for homofermentative and 72 ± 2 h Strains and isolates of Lactobacillus used in this study
for heterofermentative lactobacilli. Approximately 2 full loops Species Strains and isolates
of bacteria were directly harvested from isolated colonies using Heterofermentative
a calibrated 1 mm-diameter platinum loop and suspended in lactobacillus
150 μl of distilled water. An aliquot (120 μl) was transferred to a L. kefir JCM 5818a
ZnSe optical plate and dried under moderate vacuum (0.1 bar) 8310, 8314, 8315, 8317, 8319, 8321, 8325, 8326,
8332, 8335, 8343, 8344, 8345, 8347, 8348, 83110,
to obtain a transparent bacterial film (Helm et al., 1991). Three
83111, 83113, 3115, 83116b
replicate spectra obtained from at least 4 independent experi- L. parakefir JCM 8573a
ments were averaged for each reference strain or isolate (a 8328b
minimum of 4 averaged spectra were analysed for each strain). L. brevis JCM 1059a
Effects of i) small variations in the growth temperature (±5 °C), Homofermentative
lactobacillus
ii) incubation time (±5 h) and, iii) the make of MRS medium
L. plantarum 335, 337c
used (Biokar, France, or Difco, Detroit, MI, USA) on the DSMZ 20174d
spectral features of the reference strains of heterofermentative 8311, 8313, 8316, 8318, 8323, 8324, 8327, 8329,
lactobacilli (L. kefir JCM 5818, L. parakefir JCM 8573 and L. 8331, 8333, 8334, 8336, 8337, 8338, 8341, 8342,
brevis JCM 1059) and three homofermentative lactobacilli (L. 8346, 8349, 83112, 83114, 83210b
L. acidophilus DSMZ 20071d
acidophilus ATCC 314, L. plantarum DSMZ 20174 and L.
ATCC 314, ATCC515e
kefiranofaciens JCM 6985), were studied through four L. kefirgranum JCM 8572a
independent experiments. L. cassei DSMZ 20011d
L. kefiranofaciens JCM 6985a
2.3. Acquisition of IR spectra and data processing a
JCM Japan Collection of Microorganisms, Institute of Physical and
Chemical Research, Japan.
FT-IR absorption spectra between 4000 and 650 cm− 1 were
b
Local isolates of kefir grains, identified by biochemical tests and whole-cell
acquired in a Spectrum One FT-IR spectrometer (Perkin Elmer protein pattern by SDS-PAGE (Garrote et al., 2001).
c
Daeschel et al., 1991.
Inst. USA) under the following conditions: 6 cm− 1 spectral d
DSMZ Deustche Sammlung von Mikroorganismen und Zellkulturen Gmbh,
resolution and 64 scan co-additions. Only the spectra that Germany.
e
surpassed the quality test assessed by the OPUS software ATCC American Type Culture Collection.
282 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287

heterofermentative lactobacilli. These differences became more
significant when the first derivative was applied. The maximum
differences among first derivatives of homo- and heterofermenta-
tive lactobacilli spectra were found in four main windows of the
whole spectral range (Fig. 1A): the 3000–2800 cm− 1 region (W1),
dominated by –C–H stretching vibrations of the functional
groups usually present in the fatty acid components; the 1800–
1600 cm− 1 window (W2), which includes the ester and Amide I
regions; the 1200–900 cm− 1 window (W3), known as polysac-
charide region and, the 900–700 cm− 1 window (W4), which
contains the specific spectral pattern known as “true fingerprint
region” (Helm et al., 1991; Naumann et al., 1991). Sub-ranges
included in these four windows were selected as being the ones
that contributed maximally to grouping the spectra into 2 distinct
clusters and used as input data for spectral distances calculation
and cluster analysis. The dendrogram obtained exhibited two
main clusters clearly separating homo- and heterofermentative
lactobacilli (Fig. 1B). Therefore, phenotypic differences between
the two groups of bacteria could be detected by FT-IR
spectroscopy discriminating the organisms belonging to each
group. This first discrimination step was essential for the
subsequent differentiation among the bacteria belonging to each

Fig. 1. (A) Vector normalized first-derivatives spectra of lactobacilli reference strains and isolates from local kefir grains. The most discriminatory spectral sub-region
used for homofermentative and heterofermentative differentiation is indicated in each spectral window (W1–4). (B) Dendrogram obtained from hierarchical clustering
analysis of vector normalized first derivative spectra applying Pearson correlation coefficient and Ward's algorithm in the sub-region: 3000–2927, 1789–1700, 1059–
935 and 896–833 cm− 1.
 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287 283

group. It is important to note that the same procedure had to be
applied when biochemical assays were employed for kefir
lactobacilli classification (Garrote et al., 2001).
3.2. Differentiation among heterofermentative lactobacilli
spectra
Three heterofermentative reference strains (Table 1) repre-
sentative of the most frequently species found in kefir grains
(Angulo et al., 1993; Takizawa et al., 1994, 1998; Garrote et al.,
2001) were included in the present analysis. In a first set of
experiments, all the measurements were performed under strict
standardised conditions: 30 ± 2 °C, 72 ± 2 h, and same make of
MRS medium. The first derivatives features in several spectral
windows were compared and analysed. After performing
different combinations of spectral ranges, four equally weighted
frequencies 1780–1720, 1500–1200, 2950–2930 and 900–
700 cm− 1, were selected to best contribute to the discrimination
among the spectra of the three species studied (Fig. 2A).
Then, 4 average spectra of each of the heterofermentative
lactobacilli isolated from kefir grains were included in the
cluster analysis described above. The dendrogram obtained
using the first derivatives in the regions previously selected is
shown in Fig. 2B. Most of these lactobacilli clustered with L.
kefir JCM 5818 strain (showed with arrows in Fig. 2B). Only
one isolate (CIDCA 8328) was grouped with L. parakefir JCM
8573 strain. The discrimination obtained by this model
completely matched with the identification achieved by
conventional biochemical procedures (Garrote et al., 2001).
3.3. Effects of variations in growth conditions
Since the biochemical composition of cells depends on
cultivation conditions, the importance of using highly

Fig. 2. Dendrograms obtained with vector normalized first derivative spectra of heterofermentative lactobacilli, arising from hierarchical cluster analysis using Pearson
coefficient (1780–1720, 1500–1200, 2950–2930 and 900–700 cm− 1 windows) and Ward's clustering algorithm: (A) developed with the 3 reference strains indicated
in Table 1. (B) Performed with the 21 local isolates from kefir grains and reference strains. Arrows indicate the eight L. kefir JCM 5818 average spectra.
284 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287

standardised procedures when applying vibrational spectroscopy
techniques in the characterization and discrimination of micro-
organisms has long been recognised (Helm et al., 1991; van der
Mei et al., 1993; Oust et al., 2004a,b). The robustness of our
model in the discrimination of heterofermentative lactobacilli
was validated under small variations in culture conditions.
Although in an identification procedure all parameters have to be
kept constant, the heterogeneity due to culture conditions cannot
be completely eliminated and some spectral variability may
occur. FT-IR spectra of the heterofermentative reference strains
(Table 1) were evaluated against changes in cultivation
temperature (25, 30 and 35 °C) and incubation time (67, 72
and 77 h). The maximum changes in the first derivatives spectra
of reference strains were observed in the spectral window 1500–
850 cm− 1 (data not shown). This region was used in previous
reports to study reproducibility and differentiation of homo-
fermentative lactobacilli (Oust et al., 2004a,b). Nevertheless,
when HCA was applied in the windows used to differentiate
heterofermentative lactobacilli (1780–1720, 1500–1200, 2950–
2930 and 900–700 cm− 1), the species analysed were accurately
differentiated although their spectra were clustered according to
growth temperature (Fig. 3). These results indicated that, in
accordance to previous results obtained for homofermentative
lactobacilli (Oust et al., 2004b), small changes in time of culture
and temperature of growth produced a reduction in reproduc-
ibility of the FT-IR spectra, although they did not affect the
differentiation among the heterofermentative lactobacilli studied.
In conclusion, the developed model may be applied as a potential
method for the rapid routine differentiation of heterofermentative
lactobacilli isolated from kefir grains at species level and showed
to be robust against small variations in growth conditions.
However, when the reference strains were cultivated in different
make of MRS medium (Difco and Biokar) the respective spectra
were significantly affected. In the case of L. kefir and L. parakefir
the HCA clustered the spectra in two groups associated with the
two media. The changes in the phenotypic expression of both
species were so different that they grouped themselves primarily
according to the kind of growth medium where they were grown
(data not shown). Then, the results obtained with heterofermen-
tative lactobacilli in different make of culture medium affected
the differentiation at species level, in particular the ones that were
not separated by a high distance in a FT-IR analysis (L. kefir and
L. parakefir). The susceptibility of FT-IR spectroscopy to
discriminate biochemical changes in whole cells associated to
modifications in culture media was previously analysed in
homofermentative lactobacilli (L. sakei, L. plantarum and L.
curvatus) (Oust et al., 2004b), in different Listeria monocyto-
genes strains, and in oral Streptococci (van der Mei et al., 1993).
In these cases the authors reported very little or no variations
when culturing on different types of growth media. The use of
different make of culture media did not affect the differentiation
at species of the homofermentative lactobacilli assayed in this
work (data not shown).
3.4. Differentiation of kefir homofermentative species
The first derivative spectra of the homofermentative
lactobacilli reference strains (Table 1) were analysed. These
species were selected taking into account the ones previously
found in other kefir grains (Angulo et al., 1993; Takizawa et al.,
1998). The spectral features of the first derivatives were
compared in the whole spectral range. A differentiation model
for these homofermentative Lactobacillus species could be
developed on the basis of a combination of four spectral
windows in which the maximum differences among the species
were found: 1230–900 cm− 1 (included the window mainly

Fig. 3. Robustness of the classification model for heterofermentative lactobacilli. Hierarchical cluster analysis produced by the model developed for the discrimination
of heterofermentative lactobacilli. Pearson's coefficient and Ward's algorithm in the spectral regions 1780–1720, 1500–1200, 2950–2930 and 900–700 cm−1 obtained
at different temperatures (25, 30 and 35 °C) and after different incubation times (67, 72 and 77 h).
 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287 285

assigned to carbohydrates vibration), 1777–1690 cm− 1 (in the
lipid ester region containing the absorption of fatty acid esters
C_O bond), 1357–1240 cm− 1 (included in the so-called
“mixed” region) and 2960–2870 cm− 1 (inside the window
assigned to lipids vibrations). The hierarchical classification
obtained using the combination of these four windows resulted
in the dendrogram displayed in Fig. 4A. Five distinct clusters
were produced and a clear discrimination among the species
could be observed. It is important to note that L. kefirgranum
JCM 8572 and L. kefiranofaciens JCM 6985 were closely
clustered, which is consistent with the fact that L. kefirgranum
has recently been reclassified as a sub-species of L. kefirano-
faciens (Vancanneyt et al., 2004).
When the spectra of the 21 kefir homofermentative isolates
were tested in the developed model, all the isolates clustered
with L. plantarum reference strains (Fig. 4B). These results are

Fig. 4. (A) Dendrogram of a hierarchical cluster analysis performed on first derivative spectra of 5 species (9 strains) of homofermentative lactobacilli reference strains.
Cluster analysis was performed using Ward’s algorithm and Pearson’s correlation considering the spectral ranges: 1230–900, 1777–1690, 1357–1240 and 2960–
2870 cm− 1 equally weighted. (B) Dendrogram obtained with first derivatives vector normalized spectra of kefir grains isolates and homofermentative reference strains
applying the model mentioned above.
286 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287
consistent with the identification obtained in the routine
phenotypic characterization (Garrote et al., 2001).
FT-IR spectroscopic identification techniques for Lactoba-
cillus spp. applying different multivariante algorithms were
previously reported by Curk et al. (1994). Amiel et al. (2000)
studied different homofermentative lactobacilli species like L.
paracasei, L. casei, L. zeae and L. rhammosus. While in these
previous works the spectral region 1400–720 cm− 1 was found
to be relevant for the discrimination among those species, in the
present report four different windows spread along the whole
spectral range were found to be important for the differentiation
of the 5 homofermentative species studied (Fig. 4A). Oust et al.
(2004a,b) have recently developed a robust identification
technique based on the combination of FT-IR and partial least
squares regression (PLSR), a supervised multivariate regression
technique, for the identification of three closely related
homofermentative lactobacilli typically found in meat and
cheese products (L. sakei, L. curvatus and L. plantarum), at
species level. They also reported the differentiation among these
three species by FT-IR identification techniques based on HCA
using three windows in the spectral region 1400–720 cm− 1
(Oust et al., 2004a). However, this model was reported to be
suited for the discrimination of these three species when less
than 29 strains were considered. On the other hand, the results
of our study reflect the high discriminatory power of the FT-IR
technique in combination with HCA for the accurate differen-
tiation of five homofermentative lactobacilli species commonly
present in kefir grains (L. plantarum, L. acidophilus, L.
kefirgranum, L. cassei, and L. kefiranofaciens) when 4 spectral
windows distributed along all the spectral range (3000–
700 cm− 1) were used.
In conclusion, our results proved that FT-IR is a powerful
tool for an easy and rapid characterization of both homo- and
heterofermentative lactobacilli present in kefir grains. It is
interesting to notice that for taxonomic discrimination, FT-IR
spectroscopy and biochemical methods (both phenotypic
methodologies) required a prior differentiation between homo-
fermentative and heterofermentative lactobacilli. Simple iden-
tification techniques of lactic acid bacteria are important at the
time of studying microbial diversity of kefir grains.
Acknowledgements
AG Abraham and GL Garrote are members of the Scientific
career of CONICET and GL De Antoni of CIC-PBA. MA
Golowczyc is fellowship of CONICET. A. Bosch is member of
CIC-PBA. O. Yantorno is Professor of the Facultad Ciencias
Exactas, UNLP. The authors are grateful to Julio Figari for his
technical assistance. This work was supported by the Agencia
Nacional de Promoción Científica y Tecnológica (ANPCyT),
CONICET, CIC-PBA and Universidad Nacional de La Plata
(UNLP).
References
Amiel, C., Mariey, L., Curk-Daubie, M.-C., Pichon, P., Travert, J., 2000.
Potentiality of Fourier transform infrared spectroscopy (FT-IR) for
discrimination and identification of dairy lactic acid bacteria. Lait 80,
445–459.
Angulo, L., Lopez, E., Lema, C., 1993. Microflora present in kefir grains of the
Galician Region (north-west of Spain). Journal of Dairy Research 60,
263–267.
Curk, M.C., Peledan, F., Hubert, J.C., 1994. Fourier transform infrared (FTIR)
spectroscopy for identifying Lactobacillus species. FEMS Microbiology
Letters 123 (3), 241–248.
Daeschel, M.A., Torres, J.A., Martino, M., Youssef-Hakimi, N., 1991. Producción
y aspectos fisicoquímicos de un hidrocoloide obtenido por fermentación en
sistema modelo y suero ácido de leche con Lactobacillus plantarum. Revista de
Agroquímica y Tecnología de Alimentos 31, 379–392.
Fujisawa, T., Adachi, S., Toba, T., Arihara, K., Mitsuoka, T., 1988. Lactoba-
cillus kefiranofaciens sp. nov. isolated from kefir grains. International
Journal of Systematic Bacteriology 38, 12–14.
Garrote, G., Abraham, A., De Antoni, G., 2001. Chemical and microbiological
characterization of kefir grains. Journal of Dairy Research 68, 639–652.
Helm, D., Labischinski, H., Schallehm, P., Naumann, D., 1991. Classification
and identification of bacteria by Fourier-transform infrared spectroscopy.
Journal of General Microbiology 137, 69–79.
Kandler, O., Kunath, P., 1983. 1 Lactobacillus kefir sp. nov., a component of the
microflora of kefir. Systematic and Applied Microbiology 4, 286.
Kirschner, C., Maquelin, K., Pina, P., Ngo Thi, N.A., Choo-Smith, L.P.,
Sockalingum, G.D., Sandt, C., Ami, D., Orsini, F., Doglia, S.M., Allouch, P.,
Manfait, M., Puppels, G.J., Naumann, D., 2001. Classification and
identification of enterococci: a comparative phenotypic, genotypic, and
vibrational spectroscopy study. Journal of Clinical Microbiology 39 (5),
1763–1770.
Maquelin, K., Choo-Smith, L.P., Kirschner, C., Ngo Thi, N.A., Naumann, D.,
Puppels, G., 2002. Vibrational spectroscopic studies of microorganisms. In:
Chalmers, J.M., Griffiths, P.R. (Eds.), Handbook of Vibrational Spectros-
copy. John Wiley & Sons, Ltd, Chichester, pp. 1–27.
Mariey, L., Signolle, J.P., Amiel, C., Travert, J., 2001. Discrimination,
classification, identification of microorganisms using FT-IR spectroscopy
and chemometrics. Vibrational Spectroscopy 26, 151–159.
Marshall, V.M., Cole, W., Brooker, B.E., 1984. Observations on the structure of
kefir grains and the distribution of the microflora. Journal of Applied
Bacteriology 57, 491–497.
Molska, I., Moniuszco, I., Komorowas, K., Merilanen, V., 1983. Characteristics
of bacilli of Lactobacillus casei species appearing in kefir grains. Acta
Alimentaria Polónica IX, 80–88.
Mukai, T., Onose, Y., Toba, T., Itoh, T., 1992. Presence of glycerol teichoic acid
in the cell wall of Lactobacillus kefiranofaciens. Letters in Applied
Microbiology 15, 29–31.
Naumann, D., 2000. Infrared Spectroscopy in Microbiology. In: Meyers (Ed.),
Encyclopedia of Analytical Chemistry. John Wiley & Sons, Ltd, Chichester,
pp. 1–19.
Naumann, D., Helm, D., Labischinski, H., 1991. Microbiological characteriza-
tions by FT-IR spectroscopy. Nature 351, 81–85.
Ottogalli, G., Galli, A., Resmini, P., Volonerio, G., 1973. Composizione
Microbiologique, chimica ed ultraestructtura dei granuli di Kefir. Annali di
Microbiologia ed Enzimologia 23, 109–121.
Oust, A., Moretro, T., Kirschener, K., Narvhus, J.A., Kohler, A., 2004a. FT-IR
spectroscopy for identification of closely related lactobacilli. Journal of
Microbiological Methods 59, 149–162.
Oust, A., Moretro, T., Kirschner, C., Narvhus, J.A., Kohler, A., 2004b.
Evaluation of the robustness of FT-IR spectra of lactobacilli towards changes
in the bacterial growth conditions. FEMS Microbiology Letters 239,
111–116.
Pintado, M.E., Lopes da Silva, J.A., Fernandes, P.B., Malcata, F.X.y.J.T.A.,
1996. Microbiological and rheological studies on Portuguese kefir grains.
International Journal of Food Science & Technology 31, 15–26.
Rosi, J., Rossi, J., 1978. I microorganismi del kefir: I fermenti lattici. Sienza e
Tecnica Lattiero-Casearia 29, 291–305.
Simova, E., Beshkova, D., Angelov, A., Hristozova, T.S., Frengolva, G.,
Spasov, Z., 2002. Lactic acid bacteria and yeasts in kefir grains and kefir
made from them. Journal of Industrial Microbiology and Biotechnology 28,
1–6.
 A. Bosch et al. / International Journal of Food Microbiology 111 (2006) 280–287
Takizawa, S., Kojima, S., Tamura, S., Fujinagasa, S., Benno, Y., Nakase, T.,
1994. Lactobacillus kefirgranum sp. nov. and Lactobacillus parakefir sp.
nov., two new species from kefir grains. International Journal of Systematic
Bacteriology 44, 435–439.
Takizawa, S., Kojima, S., Tamura, S., Fujinagasa, S., Benno, Y., Nakase, T.,
1998. The composition of the Lactobacillus flora in kefir grains. Systematic
and Applied Microbiology 21, 121–127.
Toba, T., Uemura, H., Mukai, T., Fujii, T., Itoh, T., Adachi, S., 1991. A new
fermented milk using capsular polysaccharide-producing Lactobacillus
kefiranofaciens isolated from kefir grains. Journal of Dairy Research 58,
497–502.
Vancanneyt, M., Mengaud, J., Cleenwerck, I., Vanhonacker, K., Hoste, B.,
Dawyndt, P., Degivry, M.C., Ringuet, D., Janssens, D., Swings, J., 2004.
Reclassification of Lactobacillus kefirgranum Takizawa et al 1994 as Lac-
287
tobacillus kefiranofaciens subsp. kefirgranum subsp. nov. and emended
description of L. kefiranofaciens Fujisawa et al 1988. International Journal
of Systematic and Evolutionary Microbiology 54, 551–556.
van der Mei, H.C., Naumann, D., Busscher, H., 1993. Grouping of oral
streptococcal species using Fourier-transform infrared spectroscopy in
comparison with classical microbiological identification. Archives of Oral
Biology 38, 1013–1019.
Witthuhn, R.C., Schoeman, T., Cilliers, A., Britz, T.J., 2005a. Impact of
preservation and different packaging conditions on the microbial community
and activity of Kefir grain. Food Microbiology 22, 337–344.
Witthuhn, R.C., Schoeman, T., Britz, T.J., 2005b. Characterization of the
microbial population at different stages of Kefir production and Kefir grain
mass cultivation. International Dairy Journal 15, 383–389.