International Journal of Food Microbiology 94 (2004) 269 – 278

www.elsevier.com/locate/ijfoodmicro

Isolation, identification and characterisation of the dominant

microorganisms of kule naoto: the Maasai traditional fermented
milk in Kenya
Julius Maina Mathara a,*, Ulrich Schillinger a, Phillip Museve Kutima b,
Samuel K. Mbugua c, Wilhelm H. Holzapfel a
a
Federal Research Centre for Nutrition, Institute of Hygiene and Toxicology, Haid-und-Neu-Str. 9, D-76131 Karlsruhe, Germany
b
Department of Food Science and Technology, Jomo Kenyatta University of Agriculture and Technology, P.O. Box 62000 Nairobi, Kenya
c
Department of Food Technology and Nutrition, University of Nairobi, P.O. Box 29053 Nairobi, Kenya
Received 28 April 2003; received in revised form 2 November 2003; accepted 20 January 2004

Abstract

From 22 samples of kule naoto, the traditional fermented milk products of the Maasai in Kenya, 300 lactic acid bacterial
strains were isolated and phenotypically characterised by their ability to ferment different carbohydrates and by additional
biochemical tests. Lactic acid bacteria (LAB), especially the genus Lactobacillus, followed by Enterococcus, Lactococcus and
Leuconostoc, dominated the microflora of these samples. The major Lactobacillus species was Lactobacillus plantarum (60%),
with a lower frequency of isolation for Lactobacillus fermentum, Lactobacillus paracasei and Lactobacillus acidophilus. Most
strains produced enzymes such as h-galactosidase and peptidases, which are of relevance to cultured dairy product processing,
and exhibited similar patterns of enzymatic activity between species. Enterobacteriaceae could not be detected in 15 out of 22
samples (detection level 102/ml). Conversely, yeasts (detection level 101/ml) were detected in those samples in which
Enterobacteriaceae were not found. The pH values of all these samples were < 4.5.
D 2004 Elsevier B.V. All rights reserved.

Keywords: Maasai; Lactic acid bacteria; Traditional fermented milk; Kule naoto

1. Introduction of 8 – 15 huts per kraal, known as Manyatta. Kule

The Maasai community is a Nilo-hamitic tribe
living a nomadic life in the East Africa Rift-Valley
of southern Kenya and northern Tanzania. The com-
munity has a conserved rich traditional heritage that
revolves around cattle. They live in small settlements
* Corresponding author. Tel.: +49-721-6625-465; fax: +49-721-
6625-453.
E-mail address: juliusmaina_2000@yahoo.com (J.M. Mathara).
naoto, the traditional lactic fermented milk product, is
the major daily diet of the Maasai community, which
rarely consume fruits or grains. On average, 2 –3 l of
the fermented product is consumed per person per
day. The product is produced from unpasteurised
whole milk from the zebu breed of cows using
centuries old practices. The product whose rich taste
and consistency is similar to unsweetened commercial
yoghurt or fresh cheese, is spontaneously fermented
for at least 5 days, though a longer period is usually

0168-1605/$ - see front matter D 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijfoodmicro.2004.01.008
270 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
preferred (Mathara et al., 1996; Mathara, 1999).
Whole fresh raw milk is filled into a custom-made
specially treated gourd made from the hollowed out
dried fruit of the plant Lagenaria siceraria. The dried
calabash, used as fermentation gourd, is gently rubbed
with a burning end of a chopped stick from the tree
Olea africana locally known as Enkidogoe, or also
from other trees allowing charcoal to break inside
(Mathara, 1999). This procedure is repeated at least
three times. The gourd is filled with milk and then
closed by a special cap obtained from the same gourd
during its preparation. After fermentation, the product
is gently shaken before consumption.
These traditional lactic fermented milk products
are popular among the Maasai community and people
living in eastern Africa. Kenyans, in general, prefer it
due to its excellent natural taste and aroma, among
other functional benefits. Also, because of their cul-
tural associations, the people believe in a therapeutic
value towards curing or protection from ailments such
as diarrhoea and constipation (Mathara, 1999). So far,
there is limited scientific information to justify these
claims. According to Holzapfel (1997, 2002), there is
an increasing need to select microbial strains with
functional properties for commercial production and
for improvement of quality and safety of existing
traditional fermented food products. This study was
conducted against the background of the key role kule
naoto plays in the nutrition of the Maasai people and
the fact that large numbers of viable lactic acid
bacteria (LAB) are consumed daily by the Maasai.
The objectives of this study were to determine and
study the predominant microbial groups in kule naoto
and to identify and characterise the dominant (LAB)
from representative product samples obtained from
the Maasai community in Kenya.
2. Material and methods
2.1. Collection of samples
Twenty-two traditional fermented milk samples
were collected from the Maasai manyattas in the rural
plains of Maasailand in Kenya. Samples were collect-
ed, both in their original fermentation vessels and in
sterile sample bottles. Two sampling regimes were
used, 1 during the rainy season (April and May) 2001,
with 13 samples and the second during the dry period
(August and September) 2002, with 9 samples The pH
of the samples was determined at the sampling site
using a calibrated portable pH-meter (Messkoffer Qph
70, WWR-International, Germany). The samples were
kept at 4 –8 jC and transported in special cool boxes
within hours to the Federal Research Centre for
Nutrition, Karlsruhe, Germany for analyses.
2.2. Microbial enumeration and isolation
Ten grams of milk sample was transferred asepti-
cally into 90-ml Ringer’s solution and mixed thor-
oughly. Serial dilutions (10 1 – 10 8) were made for
each sample and 0.1 ml of the appropriate dilution
spread plated on universal and selective media. Plate
count agar (Merck, Darmstadt, Germany) was used
for enumeration of aerobic mesophilic counts as
described in the International Dairy Federation refer-
ence method (IDF 100B: 1991). MRS agar pH 6.4
(Merck) was used for enumeration of total LAB,
Rogosa agar (Merck) for enumeration of lactobacilli
and M17 (Merck) for enumeration of lactococci.
Kanamycin aesculine azide agar (Merck) (KAA)
was used for enumeration of enterococci and Violet
Red Bile Dextrose Agar (VRBD) (Merck) for enu-
meration of Enterobacteriaceae. Acidified Potato
Dextrose Agar (PDA) (Merck), pH 3.7, was used for
enumeration of yeasts. MRS and M17 plates were
incubated anaerobically at 30 jC using anaerobic jars
together with the Merck Anaerocult A GasPak anaer-
obic system (Merck 1.13829). Representative colonies
were isolated from MRS, M17, PDA and Kanamycin,
aesculine azide agar and purified by streak plating
using the same medium. Gram-positive, catalase-neg-
ative bacteria were purified by re-streaking on MRS
agar pH 6.4 (Merck) and M17 (Merck), and were re-
suspended and preserved in the same medium con-
taining 15% glycerol at 18 jC. Representative
yeast colonies on PDA were examined by phase
contrast microscopy, purified by successive streaking
on PDA and stored on slants at 2 jC.
2.3. Phenotypic characterisation
Initial characterisation of Gram-positive, catalase-
negative bacterial isolates was by microscopy (cell
morphology and arrangements) (Gerhardt et al.,
 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278 271

1981). Growth at 10, 15 and 45 jC in MRS broth was
evaluated visually after 24, 48 and 72 h of incubation.
Tests for presence of catalase, production of ammonia
from arginine and type of fermentation were carried
out as described by Harrigan and McCance (1976).
Salt tolerance was determined using MRS broth con-
taining 6.5% (w/v) NaCl with incubation for 72 h at
37 jC. Growth at pH 3.9 and 9.6 was determined in
MRS with the pH adjusted by using HCl and NaOH,
respectively. Dextran production was determined by
roppy colony formation on MRS agar pH 6.5 supple-
mented with 10% sucrose (Merck, 1.05323). Ap-
proaches followed in the phenotypic differentiation
were according to the information supplied by Wood
and Holzapfel (1995) and Stiles and Holzapfel (1997).
2.4. Lactic acid configuration
The type and amount of D( ) and L(+) isomers of
lactic acid produced from glucose was assayed in
modified MRS broth without beef extract and acetate
and using D- and L-lactate dehydrogenase from a
commercial kit (Hoffman La Roche Diagnostic Man-
nheim, Germany containing D-lactate dehydrogenase
code no. 1585436, L-lactate dehydrogenase code no.
127221, glutamate pyruvate transaminase (GPT) code
no. 737127 and nicotinamide adenine dinucleotide
(NAD) code no. 127990 following instructions of the
supplier). The modified MRS medium had the follow-
ing composition in g/l: casein peptone 1.0, yeast
extract 4.0, glucose 20, di-potassium hydrogen phos-
phate 2.0, Tween 80 1.0, di-ammonium hydrogen
citrate 2.0, magnesium sulphate 0.2 and manganese
sulphate 0.04.
2.5. Identification of meso-diaminopimelic acid
(meso-DAP)
Presence of meso-DAP in the bacterial cell wall was
determined according to a modified method described
by Schillinger and Lücke (1987). Briefly, isolates were
grown in 1.0 ml MRS broth (Merck), for 48 h,
harvested and washed with 1 ml of distilled water.
The sediment was re-suspended in 1.0 ml 6 N HCl and
transferred to a capped test tube. The cells were
hydrolysed overnight at 100 jC. The content of the
tube was dried in an air stream. The sediment was re-
suspended in 1.0 ml of distilled water and dried in the
same manner. Finally, the sediment was re-suspended
in 0.1 ml distilled water and samples were spotted on

Table 1
Enzymatic profiles of lactic acid bacterial strains isolated from kule naoto, the Maasai traditional fermented milk product of Kenya
Enzyme/genera Leuconostoc Lb. casei group Lb. rhamnosus Lb. acidophilus group Lb. plantarum
(n = 3) (n = 5) (n = 6) (n = 10) (n = 10)
Alkaline phosphatase 5 2 20 0 0
Esterase (C4) 23 18 30 20 10
Esterase lipase (C8) 3.3 8 30 8.5 10
Lipase (C14) 3.3 10 5 6.5 0
Leucine arylamidase 5 z 40 z 40 z 40 z 40
Valine arylamidase 5 z 40 z 40 7.5 z 40
Cystine arylamidase 3.3 10 30 12 20
Trypsin 0 5 5 0 0
a-Chymotrypsin 5 5 10 0 0
Acid phosphatase 10 30 30 19 5
Naphthol-AS-BIphosphohydrolase 5 z 40 20 19 z 40
a-Galactosidase 8.3 14 20 4.5 0
h-Galactosidase 21.7 0 z 40 16.5 z 40
h-Glucuronidase 3.3 0 5 0 0
a-Glucosidase z 40 5 20 7.5 10
h-Glucosidase 3.3 z 40 z 40 z 40 z 40
N-Acetyl-h-glucosaminidase 0 5 0 12.5 30
a-Mannosidase 0 5 0 0 0
a-Fucosidase 0 0 z 40 0 0
Results (nmol of chromophore released) are averaged and n is the number of strains tested. Enzyme activity is shown as nM of chromophore
released after 6 h of incubation at 37 jC.
272 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
thin layer plates [DC Plastic Folien: Cellulose (20 
20) Merck 5577]. Ascending one-dimensional chro-
matography was done in a modified solvent solution
containing methanol/pyridine/10 N HCl/water in
(32:4:1:7) v/v/v/v. The solvent was prepared one hour
before use. After drying, the chromatogram was de-
veloped with ninhydrin solution and placed for 5 min
in a 100 jC oven. Spots representing meso-diamino-
pimelic acid appeared dark green to grey and turned
yellow within 24 h in the dark.
2.6. Carbohydrate fermentation assays
The first determination of the carbohydrate fermen-
tation profile of all the strains was done in MRS
fermentation broth (Merck) without glucose and con-
taining chlorophenol red (0.04 g/l) as pH indicator.
The individual sugars were prepared as 2.5% (m/v)
solutions and filter sterilised using a 0.5-Am filter. A
micro-titre plate was used whereby 25 Al of sugar
solution was added to 100 Al of basal medium con-
taining the washed cell suspension. Strains were
grown overnight at 30 jC in MRS and the cell pellet
obtained by centrifugation at 10,000  g for 5 min.
The pellet was washed twice in sterilised double
distilled water before re-suspending it in basal me-
dium. Subsequently, API 50CH galleries and API
CHL medium (bioMerieux Marcy-I’Etoile, France)
were used according to the manufacturer’s instructions
for determining the sugar fermentation spectrum of
some selected strains. APILAB PLUS V3.2.2 software
database (bioMerieux) was used for interpretation of
results.
2.7. Enzymatic profiles of predominant organisms
The enzymatic profiles of the predominant LAB
were assayed using API Zym galleries (bioMerieux)
according to the manufacturer’s instructions. The
activities of 19 enzymes were tested (see Table 1).
3. Results
3.1. pH of the fermented milk samples
The average pH of the fermented milk samples was
4.4 with pH values ranging from 4.17 to 5.19 with
only four samples in which the pH exceeding 4.5
(Table 2).
3.2. Enumeration of microorganisms
Table 2 summarises the microbial counts of the
fermented milk samples obtained from 22 sampling
regions (manyattas), and Fig. 1 shows the distribution
of LAB and yeasts according to their levels of viable
counts. LAB were the dominating organisms in all the
samples with average values of 8.0 log10 cfu ml 1 on
MRS and 7.9 log10 cfu ml 1 on M17 agar, with
corresponding average mesophilic bacterial counts of
8.1 log10 cfu ml 1. In general, lactobacilli and lacto-
cocci were the predominant LAB and were mostly
detected in a range from 107 to 109 ml 1, whereas
enterococci were detected at a lower range with
numbers >104/ml in 15 samples (Table 2). The mean
Enteroccocus count was 5.5 log10 cfu ml 1 with a
range of 3.3– 9.0 log10 ml 1, an indication that this
group of LAB may contribute to some extent to the
fermentation of these products. Only in seven samples
Enterobacteriaceae were detected with counts ranging
from 4.9 to 9.4 log10 cfu ml 1 (Table 2). In none of
these seven samples, yeasts were detected. One sam-
ple had Enterobacteriaceae counts of 5.84 log10 cfu
ml 1 and a mould count of 4.26 log10 cfu ml 1, but
with no yeasts detected. The pH of these samples with
detectable numbers of Enterobacteriaceae was z 4.5
with the exception of one sample. The remainder of
the samples (15) had yeast counts with values ranging
from 4.24 to 7.44 log10 cfu ml 1, with average of 6
log10 cfu ml 1. The pH of these samples was V 4.5.
The observation that no Enterobacteriaceae were
detected in samples with detectable numbers of yeasts
is remarkable. This observation suggests a possible
interaction between bacteria and yeasts involved in
the Maasai milk fermentation.
3.3. Identification of lactic acid bacteria to genus
level
In total, 520 strains were isolated from the Maasai
traditional fermented milk products and identified to
genus level on the basis of cell morphology, gas
production from glucose, growth behaviour at 10
jC, 45 jC and in presence of 6.5% NaCl and at pH
9.6 according to Wood and Holzapfel (1995) and
 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278 273

Table 2
Microbial numbers distribution (log10 cfu/ml) determined in Maasai traditional fermented milk (kule naoto) samples from different villages
(Manyatta) in Kenya
Sampling pH Mesophilic LAB Lactobacillus Lactococcus Enterobacteriaceaea Yeastsb Enterococcus Lb.
region bacterial count plantarumc
(Manyatta) count (detected at
levels z 106)
1 4.4 8.4 8.2 8.3 6.2 < 2.0 6.6 5.6 +
2 4.4 7.4 7.5 5.1 7.3 < 2.0 4.2 3.6
3 5.0 9.2 8.5 8.5 9.4 8.4 < 1.0 9.0
4 4.8 9.2 9.2 9.0 9.0 8.2 < 1.0 9.0
5 4.7 8.7 8.8 8.5 8.5 6.8 < 1.0 4.5
6 4.5 8.4 8.0 6.6 8.7 6.4 < 1.0 6.2
7 4.3 8.5 8.2 7.3 7.2 < 2.0 6.5 3.9 +
8 4.2 6.5 6.6 6.5 7.4 < 2.0 6.1 4.6 +
9 4.3 7.5 7.4 5.5 6.2 < 2.0 5.2 3.6 +
10 4.3 6.1 6.6 5.5 7.3 < 2.0 5.4 4.2 +
11 4.4 8.6 8.3 7.6 8.2 < 2.0 6.3 3.7 +
12 4.4 8.0 7.3 5.4 7.2 < 2.0 5.0 3.3 +
13 4.4 8.3 8.0 7.0 8.3 < 2.0 6.2 3.4 +
14 4.3 8.3 7.9 7.5 8.2 < 2.0 5.3 5.7 +
15 4.3 7.8 8.0 7.7 7.7 < 2.0 7.4 5.8 +
16 5.2 9.0 8.9 8.7 9.4 9.4 < 1.0 7.4
17 4.3 7.9 8.1 8.9 7.7 < 2.0 6.9 7.2 +
18 4.3 7.9 8.2 7.8 7.7 4.9 < 1.0 6.4 +
19 4.2 7.8 7.9 7.8 7.7 < 2.0 6.3 3.7 +
20 4.2 7.8 7.6 7.7 7.7 < 2.0 6.5 6.0 +
21 4.3 8.1 8.0 8.0 7.4 < 2.0 6.2 7.4 +
22 4.5 8.6 8.5 7.8 8.5 5.8 < 1.0 7.2 +
Mean 4.4 8.1 8.0 7.4 7.9 7.1 5.9 5.5
Standard 0.3 0.8 0.6 1.2 0.9 1.6 0.9 1.8
deviation
Enterobacteriaceae were only detected in samples with pH z 4.5 with exception of sample from region 18.
a
Detection level of 100 cfu/ml and 10 cfu/ml were used for detection of Enterobacteriaceae and yeasts, respectively.
b
Only moulds were detected in samples from the Manyattas 18 and 22, compromising a log10 6.5 and 4.3, respectively.
c
Positive sign indicates positive detection of Lb. plantarum strains at detection levels of z 106.

Stiles and Holzapfel (1997). Sampling was conducted
in two seasons, during the rainy season April – May
2001, 13 samples, and during the dry season August –
September 2002, 9 samples. All the LAB isolates
were Gram-positive and catalase – negative and 55%
were found to belong to the genus Lactobacillus, 14%
to Lactococcus, 25% to Enterococcus and 6% to
Leuconostoc. Strains of Lactobacillus plantarum con-
stituted 60% of the lactobacilli.
3.4. Identification of the dominant lactic acid bacteria
to species level
In 20 samples, the lactobacilli were found to pre-
dominate the LAB. One hundred and thirty homofer-
mentative Lactobacillus strains were identified as
plantarum, as shown in Fig. 2 and Table 3. They all
produced DL-lactic acid, contained meso-DAP in the
cell wall, and all could grow in MRS at pH 3.9 and in
the presence of 6.5% NaCl. Lb. plantarum strains were
the dominant lactobacilli in all the samples except
those in which yeasts were not detected (Fig. 2 and
Table 2). Lb. plantarum strains were obtained from the
samples where Enterobacteriaceae were not detected
with the exception of sample number 18. All Lb.
plantarum strains were able to ferment ribose, galac-
tose, fructose, mannose, mannitol, N-acetyl-glucos-
amine, amygdaline, esculin, cellobiose, maltose,
melibiose, sucrose, trehalose, melezitose, raffinose
and gentiobiose. Other sugars utilised by some Lb.
plantarum strains were: glycerol (25%), L-arabinose
(70%), rhamnose (5%), sorbitol (65%), methyl-D-man-
274 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278

Fig. 1. Microbial distribution in 22 Maasai traditional fermented milk (kule naoto) samples from different villages (Manyatta) in Kenya.

noside (50%), methyl-D-glucoside (5%), turanose
(70%), D-arabitol (90%) and gluconate (80%). Eryth-
ritol, D-arabinose, D-xylose, L-xylose, adonitol, h-
methyl-xyloside, dulcitol, inositol, inulin, amidon,
glycogen, xylitol, D-lyxose, D-tagatose, D-fucose, D-
arabitol, 2-ceto-gluconate and 5-ceto-gluconate were
not fermented. Six Lactobacillus trains were identified
as Lactobacillus rhamnosus. They all produced L(+)
lactic acid, were DAP negative and were able to
ferment rhamnose. Ten Lactobacillus strains were
identified as Lactobacillus acidophilus. They all pro-
duced DL-lactic acid, they were DAP negative, grew
well at 45 jC and were arginine negative. Tests using
API 50CHL sugar profiles indicated with >90% iden-
tification probability of these strains to be as Lb.
acidophilus. Fourteen Lactobacillus strains were as
Lactobacillus paracasei collectively grouped here as
Lb. casei group. They produced exclusively L(+)-lactic
acid and had no DAP in the cell wall. Seventy-eight
heterofermentative Lactobacillus strains were identi-
fied as Lactobacillus fermentum on the basis of the
sugar fermentation pattern and biochemical properties.
Among the coccoid LAB, over 100 isolates were
classified as Enterococcus. Sixty of which were iden-
tified to species level and were found to be Entero-
coccus faecium. They fermented arabinose, grew well
at 45 jC and in MRS adjusted to pH 9.6 and in MRS
broth containing 6.5% NaCl. Sixty coccoid LAB
strains belonged to Lactococcus and 35 of them were
identified as Lactococcus lactis.
Six Leuconostoc strains were identified as Leuco-
nostoc mesenteroides subsp. dextranicum. They pro-

Fig. 2. Distribution of 238 Lactobacillus strains isolated from the Maasai traditional fermented milk products in Kenya. For sampling region
(Manyatta), see text.
 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
Table 3
Identity and distribution of 339 lactic acid bacteria strains isolated
from Maasai traditional fermented milk in Kenya
Species Number of
strains
identified
Lactobacillus plantarum 130
Lactobacillus fermentum 78
Lactobacillus casei 14
Lactobacillus rhamnosus 6 107 – 108
Lactobacillus acidophilus 10
Leuconostoc mesenteroides 6 106 – 107
Lactococcus lactis 35
Enterococcus faecium 60
duced D( )-lactic acid and produced dextran from
glucose. One strain of Leuconostoc did not produce
dextran from sucrose and could be identified as either
Weissella paramesenteroides or Leuconostoc mesen-
teroides subsp. mesenteroides.
3.5. Enzymatic profiles of the lactic acid bacteria
isolates
The activity of the 19 enzymes of technological
importance were tested. Results indicate a high h-
galactosidase activity among the Lb. plantarum and
Lb. rhamnosus strains. But none for the Lb. paracasei
strains Relatively low h-galactosidase activity was
observed among the Leuconostoc strains. Lb. rham-
nosus strains had high esterase and a-fucosidase
activity compared to other groups of strains tested.
Leuconostoc strains had high a-glucosidase activity
compared to other groups. The Lb. casei group, Lb.
rhamnosus and Lb. plantarum strains had high leucin
and valine arylamidase activities.
4. Discussion
With pH values ranging from 4.17 to 5.16, the
acidity of Maasai traditional fermented milk products
is comparable with that of most fermented dairy
products. However, those with pH values >4.5
contained moderate to high numbers of Enterobacter-
iaceae and therefore do not appear to be sufficiently
safe. An example is the sample from region 16 with a
pH value of 5.2 and an Enterobacteriaceae mean
count of 9.4 log10 cfu/ml. The lactic acid, together
275
with other metabolites, and the strong competitive
effects of the LAB population, may be mainly re-
sponsible for the extended shelf life of up to 3 months
Level of
of fermented products with sufficiently low pH. The
isolation
(per ml) LAB dominated the microbial population and viable
counts of 6.6– 9.15 log10 cfu ml 1, with mean vales
107 – 108
107 – 108 of 8.0 log10 cfu/ml, were recorded. The counts com-
107 – 108 pared with findings of similar studies on fermented
milks within the East African region. According to
106 – 107 Isono et al. (1994), counts of LAB and lactobacilli
reach 108 –109/g and 107 – 109/g, respectively. Miya-
107 – 108
106 – 107 moto et al. (1986) and Mathara (1999) reported
similar range of bacterial counts. In this study, the
dominant Lactobacillus strains apparently contributed
mainly to the technological quality attributes of the
fermented product. Also, Lb. plantarum strains clearly
seem to dominate lactobacilli in Maasai fermented
milk. This observation may be related to the fact that
the fermentation of the milk is exclusively carried out
in a plant gourd previously prepared by treatment
inside with splints of pre-heated smoking wood
(Mathara, 1999) and to the adaptation of the particular
Lb. plantarum strains to milk. According to Stiles and
Holzapfel (1997), and Wood and Holzapfel (1995),
Lb. plantarum strains are known to be commonly
associated with plant based food fermentations. From
cultured milk in Cameroon, Jivoua and Milliere
(1990) identified 47 out of 426 isolates as Lb. plan-
tarum. From a total of 26 Lactobacillus isolates from
fermented milk in northern Tanzania, Isono et al.
(1994) identified only four strains as Lb. plantarum.
Out of 21 isolates from naturally fermented milk in
Zimbabwe (Matukumira, 1996), only 3 strains were
identified as Lb. plantarum, whilst only 3 out of 336
bacteria isolates from South African traditional fer-
mented milks were identified as Lb. plantarum
(Beukes et al., 2001). Medina et al. (2001) identified
92% of all lactobacilli in ewe’s milk and cheese from
Northern Argentina as Lb. plantarum. Olasupo et al.
(2001) reported that, all the Lactobacillus isolates
from four different fermented products in Nigeria,
including the fermented milk products wara and
nono, were found to belong to Lb. fermentum.
This is the first report on the occurrence of high
levels of Lb. plantarum strains associated with Maasai
traditional fermented milk products in Kenya. All the
Lb. plantarum strains tested were able to ferment
raffinose. This property indicates good bacterial tech-
276 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
nological potential associated with consumption of the
fermented milk products together with legume based
products like beans high in flatulence causing sugars.
It also shows potential for use as a starter culture in
development of soy based fermented milk products
(Xanthopoulos et al., 2000).
Yeasts appear to be commonly associated with
traditional fermented dairy products and have been
reported in several studies (Isono et al., 1994; Gadaga
et al., 2001; Beukes et al., 2001). Isono et al. (1994)
reported occurrence of yeasts in seven of 10 samples
of traditional fermented milk in northern Tanzania
with the mean counts ranging from 6.0 to 8.0 log10 cfu
ml 1. In the current study, yeast counts ranged from
4.3 to 7.4 log10 cfu ml 1. However, a definite cor-
relation was observed in all the samples with regard to
the occurrence of yeasts and Enterobacteriaceae. No
Enterobacterieaceae were detected in any of samples
where yeasts were detected. The pH values of these
samples were less or equal to 4.5. However, one
sample (pH 4.5) had Enterobacteriaceae alongside
with moulds but no yeasts were detected. This obser-
vation indicates a possible interaction of yeasts and
the bacterial flora in the fermentation of fermented
Maasai milk and may suggest a relation to the pH.
Enterobacteriaceae, especially coliforms, are associ-
ated with poor hygiene and their occurrence in the
product may indicate a potential health risk (Beukes et
al., 2001). However, there are no reported cases of
diarrhoea associated with the consumption of Maasai
traditional fermented milk products (Mathara et al.,
1996; Mathara, 1999; Nakamura et al., 1999). This is
the first report indicating a possible interaction of
Enterobacteriaceae and yeasts in the Maasai tradi-
tional fermented milk products. In this report, practi-
cally no Lb. plantarum strains were detected in those
samples where Enterobacteriaceae were found. Fur-
ther investigation is necessary to establish the extent
and possible mechanisms of these interactions which
most probably involve the LAB, and their particular
domination relative to metabolic activity and acid
production.
In addition to Lb. plantarum, Lb. acidophilus and
Lb. rhamnosus were also found among the lactobacilli
in Maasai traditional fermented milk products. Strains
of Lb. acidophilus have been reported only for a few
other traditional fermented milk products. They com-
prised 8.5% of the bacterial flora isolated from the
traditional fermented milk product rob from Sudan
(Abdelgadir et al., 2001). Matukumira (1996) isolated
Lb. acidophilus from amasi, a Zimbabwean naturally
fermented milk product. Eight percent of the Lacto-
bacillus isolates in ewe’s milk and cheese from
Northern Argentina (Medina et al., 2001) were iden-
tified as Lb. acidophilus. This is the first report on the
occurrence of Lb. rhamnosus and Lb. acidophilus
strains (with isolation range of 107 – 108/ml) in Maasai
traditional fermented milk. The isolated Lb. rhamno-
sus strains are well adapted to the milk as indicated by
the high h-galactosidase activity.
According to Isono et al. (1994), Weissella confusa
(Lactobacillus confusus) was the dominant LAB spe-
cies isolated from the Maasai fermented milk products
in northern Tanzania. Out of 32 Lactobacillus isolates,
26 were identified as W. confusa and, of the remaining
6, 1 as W. viridescens, 1 strain as Lb. brevis and 4
strains as Lb. plantarum. In our study, 78 strains out
of 130 heterofermentative Lactobacillus strains were
identified as Lb. fermentum. Most of the other hetero-
fermentative Lactobacillus strains may be Weissella
species according to preliminary data from molecular
confirmatory tests. Lb. fermentum and species of
Weissella are occasionally found in raw milk, but
their technological role in fermentation of milk has
not been widely reported.
Only 6% of the isolated strains were identified as
Leuconostoc species. Both dextran forming and non-
dextran forming strains were identified. Leuconostoc
strains may contribute to the development of flavour
quality attributes of fermented Maasai milk. The low
percentage of Leuconostoc strains isolated from the
fermented milk samples could partly be explained by
their complex nutritional requirements (Medina et al.,
2001), but probably also by their lower adaptation to
milk. Leuconostoc species generally show a weak
competitive ability during fermentation of milk (Wood
and Holzapfel, 1995).
The absence of proteases (trypsin and chymotryp-
sin), the high activities of peptidases (leucine, valine
and cystine-aminopeptidase) and the low esterase-
lipase (C4 and C8) activities among the strains tested
appear desirable traits for flavour and texture devel-
opment in milk fermentation. Starters with low pro-
teinase and strong peptidase activities are useful in
reducing bitterness and improving body and texture
defects (Arora et al., 1990) The high h-galactosidase
 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
and h-glucosidase activities among the Lb. plantarum
and Lb. rhamnosus strains suggest their general ability
to ferment lactose. These strains are probably mainly
responsible for the characteristic flavour, body and
taste of Maasai traditional fermented milk products.
All strains showed variable activities of lipase and
esterase enzymes. Such strains could be involved in
removal of free fatty acids due to their high esterase
and lipase activities. No strain tested for its enzymatic
profiles from wara and nono (Olasupo et al., 2001)
showed any h-glucosidase or lipase activities. Similar
enzymatic profiles among LAB isolated from soft
variety Chhurpi, a traditional cheese typical of Sikkim
Himalayas, was reported by Tamang et al. (2000).
5. Conclusions
In this study on traditional Maasai fermented milk
products, the domination of Lb. plantarum and Lb.
fermentum among the associated LAB was estab-
lished. Other strains occurring in relatively high
numbers were identified as representatives of Lacto-
coccus lactis, E. faecium, Leuc. mesenteroides, Lb.
paracasei, Lb. rhamnosus and Lb. acidophilus. Data
from this study suggest a possible interaction be-
tween Enterobacteriaceae; yeasts and the pH during
the fermentation of Maasai milk. Enterobacteriaceae
were not detected simultaneously with yeasts. This
may reflect the inhibitory effect of reduced pH
values (to which yeast are more resistance) and/or
a probable antagonistic effect of yeasts and LAB
against Enterobacteriaceae, thereby contributing to
the quality and safety of the fermented milk prod-
ucts. The occurrence of a wide diversity of LAB in
Maasai traditional fermented milk could in part be
responsible for the desirable quality attributes asso-
ciated with the products. Most representatives are
well adapted to the milk environment as is evidenced
by their enzymatic profiles. According to Rolle and
Satin (2002), traditional small scale fermentation
technologies offer considerable potential for stimu-
lating development in the food industry of develop-
ing countries in light of their low cost, scalability,
minimal energy and the infrastructure requirements
and the wide consumer acceptance of such traditional
fermented products. LAB strains associated with
Maasai traditional fermented milk products, and with
277
desirable technological properties, show potential for
use in small scale milk fermentation.
Acknowledgements
The authors wish to thank the German Academic
Exchange Services (DAAD) for financial support of
this study. Thanks also to all the Maasai families in
Kenya for providing the research samples. The tech-
nical staff at the Institute of Hygiene and Toxicology,
Federal Research Centre for Nutrition, Karlsruhe,
Germany is also gratefully acknowledged for their
concerted effort and support during the experimental
work.
References
Abdelgadir, W.S., Hamad, S.H., Möller, P.L., Jakobsen, M., 2001.
Characteristics of the dominant microbiota of Sudanese fer-
mented milk Rob. International Dairy Journal 11, 63 – 70.
Arora, G., Lee, B.H., Lamoureux, M., 1990. Characterisation of
enzyme profiles of Lactobacillus casei species by rapid API
ZYM system. Journal of Dairy Science 73, 264 – 273.
Beukes, E.M., Bester, B.H., Mostert, J.F., 2001. The microbiology
of South African traditional fermented milks. International Jour-
nal of Food Microbiology 63, 189 – 197.
Gadaga, T.H., Mutukumira, A.N., Narvhus, J.A., 2001. The growth
and interaction of yeasts and lactic acid bacteria isolated from
Zimbabwean naturally fermented milk in UHT milk. Interna-
tional Journal of Food Microbiology 68, 21 – 32.
Gerhardt, P., Murray, R.G.E., Costilow, R.N., Nester, E.W., Wood,
W.A., Krig, N.R., Phillips, G.B., 1981. Manual of Methods
for General Bacteriology. American Society for Microbiology,
Washington, DC.
Harrigan, W.F., McCance, M.E., 1976. Laboratory methods in Food
and Dairy Microbiology. Academic Press, London.
Holzapfel, W., 1997. Use of starter cultures in fermentation on a
household scale. Food Control 8 (5/6), 158 – 241.
Holzapfel, W.H., 2002. Appropriate starter culture technologies for
small-scale fermentation in developing countries. International
Journal of Food Microbiology 75 (3), 197 – 212.
International Dairy Federation, 1991. Milk and Milk Products-Enu-
meration of Microorganisms. Colony Count at 30 jC (IDF Stan-
dard 100b: 1991). IDF, Brussels.
Isono, Y., Shingu, I., Shimizu, S., 1994. Identification and charac-
teristics of lactic acid bacteria isolated from Maasai fermented
milk in northern Tanzania. Bioscience, Biotechnology, and Bio-
chemistry 58 (4), 660 – 664.
Jivoua, C., Milliere, J.B., 1990. Lactic flora and enterococci in
cultured milk (Pindidam) manufacture in Adamaoua, Cameroon.
Lait 70, 475 – 486 (abstract).
278 J.M. Mathara et al. / International Journal of Food Microbiology 94 (2004) 269–278
Mathara, J.M., 1999. Studies on Lactic acid producing microflora in
Mursik and Kule naoto, traditional fermented milks from Nandi
and Maasai communities in Kenya. MSc. Thesis. University of
Nairobi, Kenya.
Mathara, J.M., Miyamoto, T., Koaze, H., Kiiyukia, C., Yoneya, T.,
1996. Production of traditional fermented milk in Kenya. Review.
Shizuoka Prefectural Univ. Publication, Japan, pp. 257 – 265.
Matukumira, A.N., 1996. Investigation of some prospects for the
development of starter cultures for industrial production of tradi-
tional fermented milk in Zimbabwe. Doctor Scientiarum Thesis.
Department of Food Science, Agricultural University of Norway.
Medina, R., Katz, M., Gonzalez, S., Oliver, G., 2001. Character-
isation of the lactic acid bacteria in ewe’s milk and cheese
from Northwest Argentina. Journal of Food Protection 64 (4),
559 – 563.
Miyamoto, T., Gichuru, S.G.G., Akimoto, T., Nakae, T., 1986.
Identification and properties of lactic acid bacteria isolated from
fermented beverages in East Africa. Japanese Journal of Zoo-
technical Sciences 57 (3), 265 – 276.
Nakamura, T., Sugai, M., Ozawa, A.N., Ariga, H., Koaze, H.,
Kiiyukia, C., Arai, I., Urashima, T., 1999. Microbiological prop-
erties of maziwa lala, a Kenyan traditional fermented milk of
Maasai community in Kenya. Milk Science 48 (1), 9 – 14.
Olasupo, N.A., Schillinger, U., Holzapfel, W.H., 2001. Studies on
some technological properties of predominant lactic acid bacte-
ria isolated from Nigeria fermented foods. Food Biotechnology
15 (3), 157 – 167.
Rolle, R., Satin, M., 2002. Basic requirements for the transfer of
fermentation technologies to the developing countries. Interna-
tional Journal of Food Microbiology 75 (3), 181 – 187.
Schillinger, U., Lücke, F.K., 1987. Identification of lactobacilli
from meat and meat products. Food Microbiology 4, 199 – 208.
Stiles, M.E., Holzapfel, W., 1997. Lactic acid bacteria of food and
their current taxonomy. Review. International Journal of Food
Microbiology 36, 1 – 27.
Tamang, J.P., Dewan, S., Thapa, S., Olasupo, N.A., Schillinger, U.,
Wijaya, A., Holzapfel, W.H., 2000. Identification and enzymatic
profiles of the predominant lactic acid bacteria isolated from
soft-variety Chhurpi, a traditional cheese typical of the Sikkim
Himalayas. Food Biotechnology 14 (1&2), 99 – 112.
Wood, B.J.B., Holzapfel, W.H., 1995. The Genera of Lactic Acid
Bacteria, vol. 2. Blackie Academic & Professional, Glasgow.
Xanthopoulos, V., Hatzikamari, M., Adamidis, T., Tsakalidou, E.,
Tzanetakis, N., Tzanetaki, E.L., 2000. Heterogeneity of Lac-
tobacillus plantarum isolated from Feta cheese throughout
ripening. Journal of Applied Microbiology 88, 1056 – 1064.