pcr–dgge fingerprints of microbial succession during a manufacture of traditional water buffalo...
TRANSCRIPT
PCR–DGGE fingerprints of microbial succession duringa manufacture of traditional water buffalo mozzarella cheese
D. Ercolini, G. Mauriello, G. Blaiotta, G. Moschetti and S. CoppolaDipartimento di Scienza degli Alimenti, Sezione di Microbiologia Agraria, Alimentare e Ambientale e di Igiene, Universita degli Studi di
Napoli �Federico II�, Portici, Italy
2003/0417: received 20 May 2003, revised 4 September 2003 and accepted 3 October 2003
ABSTRACT
D. ERCOLINI , G . MAURIELLO, G. BLA IOTTA, G. MOSCHETTI AND S. COPPOLA. 2003.
Aims: To monitor the process and the starter effectiveness recording a series of fingerprints of the microbial
diversity occurring at different steps of mozzarella cheese manufacture and to investigate the involvement of the
natural starter to the achievement of the final product.
Methods and Results: Samples of raw milk, natural whey culture (NWC) used as starter, curd after ripening
and final product were collected during a mozzarella cheese manufacture. Total microbial DNA was directly
extracted from the dairy samples as well as bulk colonies collected from the plates of appropriate culture media
generally used for viable counts of mesophilic and thermophilic lactic acid bacteria (LAB) and used in polymerase
chain reaction–denaturing gradient gel electrophoresis (PCR–DGGE) experiments. The analysis of the DGGE
profiles showed a strong influence of the microflora of the NWC on the whole process because after the starter
addition, the profile of all the dairy samples was identical to the one shown by the NWC. Simple indexes were
calculated for the DGGE profiles to have an objective estimation of biodiversity and of technological importance of
specific groups of organisms. LAB grown on Man Rogosa Sharp (MRS) and Rogosa agar at 30�C showed high
viable counts and the highest diversity in species indicating their importance in the cheese making, which had not
been considered so far. Moreover, the NWC profiles were shown to be the most similar to the curd profile
suggesting to be effective in manufacture.
Conclusions: The PCR–DGGE analysis showed that in premium quality manufacture the NWC used as starter
had a strong influence on the microflora responsible for process development.
Significance and Impact of the Study: The molecular approach appeared to be valid as a tool to control process
development, starter effectiveness and product identity as well as to rank cheese quality.
Keywords: microbial diversity, natural whey culture, PCR–DGGE analysis, product identity, quality control,
starter effectiveness, tracing system, water buffalo mozzarella cheese.
INTRODUCTION
Mozzarella cheese is perhaps the most popular non-ripened
cheese in the world. Traditional mozzarella is mainly
produced in Italy although it is widely exported and also
industrially produced in other countries. Water buffalo
mozzarella cheese is a high moisture (55–62%) and high fat
in dry matter (>45%) cheese; it is characterized by a soft
body and a juicy appearance and by a pleasant, fresh, sour
and slightly nutty flavour. Water buffalo mozzarella cheese
from Campania (�Mozzarella di Bufala Campana�) received
the European certification Product of Designated Origin
Correspondence to: Dr Danilo Ercolini, Dipartimento di Scienza degli Alimenti,
Sezione di Microbiologia, Universita di Napoli �Federico II�, 80055 Portici, Italy
(e-mail: [email protected]).
ª 2003 The Society for Applied Microbiology
Journal of Applied Microbiology 2004, 96, 263–270 doi:10.1046/j.1365-2672.2003.02146.x
(DOP, EEC Regulation no. 1107 12th June 1996) and it is
regarded as a typical product of southern Italy.
The manufacture has been described in detail in previous
works (Coppola et al. 1988, 1990). Briefly, the cheese is
made from whole raw water buffalo milk by adding a natural
whey culture (NWC, from the previous day manufacture) as
starter. The raw milk is heated at 37�C, then rennet and
NWC are added. After a curd-ripening phase (4Æ0–4Æ5 h at
35–37�C), which occurs under whey, the optimal pH (4Æ9–
5Æ1) is reached and the drained curd is stretched in hot water
(90–95�C). The elastic product formed is then hand-
moulded in order to get the final typical round shape with
a hand-cut on one side, which gives it the name mozzarella
(from the Italian �mozzare� for hand-cutting).
The specific characteristics of the final product mainly
arise from the raw materials employed, the agro-ecosystem
of the area of production and the traditional technology of
manufacture. The traditional mozzarella is made from raw
water buffalo milk and the microflora occurring in such
complex environment is certainly one of the parameters
affecting the dairy manufacture. In the traditional proce-
dure the NWC is a natural microbial culture occurring in
the whey drained after curd ripening. Part of this whey is
stored and employed as starter in the manufacture of the
next day. In previous studies this starter has been
characterized using both traditional and molecular proce-
dures (Coppola et al. 1988, 1990; Ercolini et al. 2001b) and
defined as complex consortium of micro-organisms of great
importance for the quality of the traditional product. The
traditional technologies of cheese making are the most
difficult to control, especially when the cheese is produced
from raw milk or by adding natural starters. Therefore, it
would be interesting to develop new methods of quality
control capable of supporting a standardization of the
process for good quality products, while preserving their
typical traits. Tracing processes and traditional products
identity would be furthermore helpful for the protection of
territory claims and for consumer protection against
frauds.
A molecular evaluation of the microbiota of several
mozzarella cheeses has been reported showing the potential
of a polymerase chain reaction–denaturing gradient gel
electrophoresis (PCR–DGGE) approach in discriminating
different qualities of cheese (Coppola et al. 2001).
In this study the succession of microbial populations
during the whole process of a premium quality traditional
water buffalo mozzarella cheese manufacture has been
monitored by cultivation coupled with molecular methods.
The aim was to monitor the process and the starter
effectiveness by collecting a series of fingerprints of the
microbial diversity occurring at different steps to control the
contribution of lactic acid bacteria (LAB) to the achievement
of the final product.
MATERIALS AND METHODS
Dairy samples
The samples were collected from a dairy producing top
quality traditional water buffalo mozzarella cheese PDO
(protected designation origin), located in Campania region,
southern Italy. Samples of raw milk before and after the
starter addition, starter (NWC), curd at the end of the
ripening, whey after draining, stretched curd and final
mozzarella cheese were aseptically collected, cooled at 4�C,
and analysed within 6 h.
Microbial enumeration and collectionof cells in bulk
Serial dilutions of each sample in quarter strength Ringer’s
solution (Oxoid) were used to inoculate plates of: MRS
agar; Rogosa agar and M17 agar (Oxoid) as culture media
widely employed to cultivate LAB. Two series of agar
plates were inoculated and incubated at 30 and 44�C for
48 h. Rogosa agar plates were incubated anaerobically
using an Anaerogen kit (Oxoid). Portions (0Æ1 ml) of
appropriate dilutions were spread plated in triplicate.
Colonies were counted and the results were calculated as
the means of three determinations. After the counts, the
plates were used for bulk formation as previously des-
cribed (Ercolini et al. 2001b). For each dilution, all the
colonies present on the surface of the plate were suspen-
ded in a suitable volume of quarter strength Ringer’s
solution to reach 1 unit of optical density (600 nm),
harvested with a sterile pipette and stored by freezing at
)20�C (Ercolini et al. 2001b). When necessary, 1 ml of the
bulk was used for DNA extraction as described below.
The bulk analysis was performed only on the samples of
initial raw water buffalo milk (M), NWC, curd at the end
of ripening (C) and final product (FP).
DNA extraction
Total DNA extraction from the dairy samples was conducted
as previously described (Ercolini et al. 2001b). The dairy
samples were fivefold diluted in TE (Tris–EDTA) buffer and
the protocol was applied to 1 ml of suspension. Moreover, the
protocol was also applied to aliquots of 1 ml bulk suspension
of colonies from the plates (Ercolini et al. 2001b). The
protocol described by the manufacturer of the Wizard DNA
purification kit (Promega, Madison, WI, USA) was applied as
follows: 1 ml of sample was centrifuged at 17 000 g for 5 min
at 4�C and the resulting pellet was resuspended in 100 ll of
TE buffer (100 mmol l)1 Tris, 10 mmol l)1 EDTA); then
160 ll of 0Æ5 mol l)1 EDTA/nuclei lysis solution (Wizard
DNA purification kit; Promega) in 1/4Æ16 ratio, 5 ll of
RNAse (10 mg ml)1; Sigma) and 20 ll of pronase E
264 D. ERCOLINI ET AL.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
(20 mg ml)1; Sigma) were added, and the mixture was
incubated for 60 min at 37�C. After incubation, 1 volume of
ammonium acetate 5 mol l)1 was added to the sample that was
then centrifuged at 17 000 g for 5 min at 4�C. The superna-
tant was precipitated with 0Æ7 volume of isopropanol and
centrifuged at 29 000 g for 5 min. Finally, the pellet was
dried and resuspended in 50 ll of DNA rehydration
solution by incubation at 55�C for 45 min.
PCR–DGGE analysis
Primers spanning the 200-bp V3 region of the 16S ribosomal
DNA of Escherichia coli were used in PCR amplification as
previously described (Ercolini et al. 2001a). A GC-clamp was
added to the forward primer, according to Muyzer et al.
(1993). Amplification was performed in a programmable
heating incubator (MJ Research Inc., Waltham, MA, USA).
Each mixture (final volume, 25 ll) contained 20 ng of tem-
plate DNA, each primer at a concentration of 0Æ2 lmol l)1,
each deoxynucleoside triphosphate at a concentration of
0Æ25 mmol l)1, 2Æ5 mmol l)1 MgCl2, 2Æ5 ll of 10x PCR
buffer and 2Æ5 U of Taq polymerase (Invitrogen). Template
DNA was denatured for 5 min at 94�C. A �touchdown� PCR
was performed (Muyzer et al. 1993) to increase the specificity
of amplification and to avoid the formation of spurious
by-products. PCR products were analysed by DGGE using a
Bio-Rad Dcode apparatus (Bio-Rad, Hercules, CA, USA).
Parallel electrophoresis experiments were performed at 60�Cby using gels containing a 25–50% urea-formamide denatur-
ing gradient [100% corresponded to 7 mol l)1 urea and 40%
(w/v) formamide] increasing in the direction of electrophor-
esis. After the electrophoresis the gels were stained in
ethidium bromide solution for 5 min, washed in distilled
water for 15 min and observed. Bands were automatically
detected by using the software Phoretic 1 advanced version
3Æ01 (Phoretix International Limited, Newcastle upon Tyne,
UK). The PCR–DGGE protocol was applied to the DNA
directly extracted from the dairy samples as well as from
DNA extracted from colonies collected in bulk.
Indexes
Simple mathematical indexes were calculated for each
fingerprint arising from PCR–DGGE analysis of the bulk
suspension from countable plates of appropriate media. The
bands considered in the analysis were only the ones
automatically detected by the software Phoretic 1.
Indexes of biodiversity:
IB1 ¼ n=nM
IB2 ¼ n=nB
IB3 ¼ IB1=nB
The biodiversity indexes (IB1, IB2 and IB3) were meant to
express the degree of microbial complexity for each bulk
analysed. The discriminative value was the number of bands in
the profile compared with the top of biodiversity detected (IB1)
or corrected for the microbial diversity relative to the specific
medium to which the profile belonged to (IB2 and IB3).
Index of similarity with the curd profile (Nei and Li
1979):IC1 ¼ 2nCs=n þ nC
where n is the number of DGGE bands in the profile; nM is
the number of bands counted in the bulk profile with the
maximum number of bands; nB is the total number of bands
counted on the specific medium the profiles refer to; nCs is
the number of bands of the profile also detected in the
profile of the curd from the same medium; nC is the number
of bands in the profile of the curd from the same medium. In
calculating the nB value, the bands migrating the same
distance in the gel were counted only once.
RESULTS
The results of viable counts are summarized in Table 1.
The raw water buffalo milk was found to be rich of
LAB. The curd and the NWC used as starter were found
to have a strong concentration of both thermophilic and
mesophilic micro-organisms; the whey collected after curd
Table 1 Viable counts of bacterial groups for different samples collected during water buffalo mozzarella cheese manufacture
Sample
Log CFU g)1 or ml)1 (S.DS.D.)
M17 at 30�C M17 at 44�C MRS at 30�C MRS at 44�C Rogosa at 30�C
Raw water buffalo milk 6Æ30 (0Æ22) 4Æ45 (0Æ20) 6Æ23 (0Æ20) 4Æ48 (0Æ16) 4Æ88 (0Æ21)
Milk after starter addition 6Æ39 (0Æ12) 4Æ64 (0Æ22) 6Æ28 (0Æ22) 4Æ63 (0Æ12) 4Æ69 (0Æ13)
Natural whey culture 7Æ78 (0Æ15) 6Æ34 (0Æ17) 7Æ49 (0Æ19) 6Æ25 (0Æ11) 6Æ50 (0Æ23)
Curd 8Æ08 (0Æ28) 6Æ32 (0Æ15) 8Æ17 (0Æ25) 6Æ20 (0Æ14) 7Æ04 (0Æ28)
Whey from curd draining 7Æ15 (0Æ24) 5Æ18 (0Æ18) 6Æ71 (0Æ17) 5Æ87 (0Æ24) 4Æ07 (0Æ27)
Stretched curd 5Æ36 (0Æ13) 2Æ39 (0Æ25) 5Æ76 (0Æ23) 3Æ39 (0Æ24) 3Æ82 (0Æ25)
Final product 4Æ04 (0Æ13) 2Æ58 (0Æ21) 3Æ78 (0Æ20) 2Æ43 (0Æ24) 2Æ84 (0Æ19)
The data are the mean values based on three replicates. The values in parentheses are standard deviations.
MICROBIAL SUCCESSION IN MOZZARELLA CHEESE 265
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
ripening had a concentration of about 1 log lower than the
curd, probably as result of cell retention in the curd
matrix. The microflora counted in the process significantly
decreased after the hot stretching step and the brine
soaking. On the basis of the counts on M17 agar, both
mesophilic and thermophilic lactic streptococci were
affected by the stretching, which displayed a significant
decrease in counts. Rogosa agar after incubation at 44�Cdid not give countable plates and thus was not considered
for the bulk analysis (data not shown).
Figure 1 shows the fingerprints obtained after the PCR–
DGGE analysis of the DNA directly extracted from the
dairy samples. The raw water buffalo milk profile consisted
of at least 15 detectable bands (Fig. 1; lane 1); all the other
samples displayed a fingerprint identical to the NWC profile
(lanes 3–6).
The analysis of the microbial diversity was also extended
to the cultivable community and the results are displayed
in Fig. 2a–e. Each panel shows profiles of the countable
dilutions from specific culture media. Only the countable
plates were considered for the analysis because only those
dilutions would have been taken into account in a
traditional isolation procedure. Simple indexes were calcu-
lated for each profile in order to achieve an objective
interpretation of the results in terms of degree of
biodiversity and technological importance of the sample.
Two sets of indexes, described in the �Materials and
methods� section, were taken into account. The biodiversity
indexes (IB1, IB2 and IB3) were meant to express the
degree of microbial complexity for each bulk analysed. Of
course, the discriminative value was the number of bands
in the profile which was considered as compared with the
top of biodiversity detected (IB1) or corrected for the
microbial diversity relative to the specific medium which
the profile belonged to (IB2 and IB3). A Pearson correlation
index (CI) was calculated for IB1 vs IB2, IB1 vs IB3 and
IB2 vs IB3; it was found that the CI was always very close
to 1 (data not shown). As the IB indexes were related, the
use of the three of them could not be useful because they
would express the same degree of diversity. Therefore, the
simplest IB1 was chosen for the interpretation of data. The
IB1 values for all the samples and media are depicted in
Fig. 3. The index is comprised between 0 and 1 and
according to the IB1 values the degree of diversity of each
sample can be arbitrarily divided in three groups: (i) high,
(ii) medium and (iii) low microbial diversity (IB1 > 0Æ7,
0Æ4 < IB1 < 0Æ7 and IB1 < 0Æ4, respectively). As shown in
Fig. 3, dairy samples plated on the same medium showed
about the same degree of complexity except for the samples
from MRS agar at 44�C which gave different IB1 values.
Moreover, the highest diversity was shown by the
mesophilic LAB, giving IB1 values higher than 0Æ7 for all
the samples on MRS and Rogosa at 30�C. The samples of
milk plated onto MRS and M17 and incubated at 44�Calso gave medium values of IB1 indicating that a consid-
erable diversity of thermophilic species was present in the
raw water buffalo milk. In addition, mesophilic streptococci
showed low microbial diversity.
IB1 values were plotted vs IC1 values and the resulting
graph is depicted in Fig. 4. This graph allows the
interpretation of two sets of data at the same time. The
plot is divided in order to highlight the area characterized
by high microbial complexity (IB1 > 0Æ7) and high
similarity with the curd after ripening (IC1 > 0Æ7). As
already observed in Fig. 3, the mesophilic LAB collected
from MRS and Rogosa incubated at 30�C displayed the
highest degree of microbial diversity grouping in the top
part of the plot. However, in the right part of the plot all
the samples with high IC1 values are gathered. Of course,
all the curd samples have IC1 values equal to 1. All the
NWC patterns showed high similarity with the curd
laying in the right part of the plot. Particularly, NWC
bulk from M17 at 44�C showed profiles identical to the
corresponding curd. Among all the other samples, only
the milk sample from M17 at 30�C displayed high
similarity with the curd.
1 2 3 4 5 6
Fig. 1 PCR-DGGE profiles of dairy samples during water buffalo
mozzarella cheese manufacture. Lanes: 1, raw milk; 2, milk after NWC
addition; 3, NWC; 4, curd after ripening; 5, stretched curd; 6, final
product
266 D. ERCOLINI ET AL.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
DISCUSSION
In this study the microbial succession during the manufac-
ture of traditional water buffalo mozzarella cheese was
investigated by PCR–DGGE analysis of the DNA directly
extracted from the dairy samples. Moreover, LAB popula-
tions were also monitored by analysing the PCR–DGGE
profiles of bulk colonies recovered from the countable plates
of three different media incubated at 37 and 44�C.
The PCR–DGGE analysis of mixed bacterial popula-
tions is widely used in environmental microbiology (Muy-
zer 1999). The amplification of variable regions of the 16S
rDNA followed by DGGE analysis leads to fingerprints of
the microbial community characterized by a correspon-
dence between bands and microbial species. This approach
has been recently applied to food and food-related
ecosystems where the microbial community was identified
from PCR–DGGE profiles after a direct DNA extraction
from food samples (Ampe et al. 1999; Cocolin et al. 2001;
Ercolini et al. 2001b, 2003; Randazzo et al. 2002). How-
ever, in some of these studies it was demonstrated that the
analysis of the cultivable community can give support to
the analysis of the sole microbial DNA directly extracted
from food (Ercolini et al. 2001b, 2003; Randazzo et al.
2002). In this study, PCR–DGGE fingerprints were
obtained from countable bulk colonies from M17, MRS
and Rogosa agar plates incubated at different temperatures.
This approach was previously explored (Ercolini et al.
2001b, 2003) and it was considered a rapid and useful
support for the PCR–DGGE analysis performed in situ.
Moreover, in previous studies we demonstrated that
statistical analysis of PCR–DGGE profiling results has
good potential in differentiating dairy products (Coppola
et al. 2001; Ercolini et al. 2002) and also in ascertaining the
geographical origin of natural starters (NWC) for mozza-
rella cheese PDO (Mauriello et al. 2003). This was applied
to monitor a mozzarella cheese making process in this
study.
Fig. 2 PCR-DGGE profiles of bulk from
media used for viable counts of LAB. (a) Bulk
on Rogosa agar after incubation at 30�C;
(b) bulk on MRS agar after incubation at
30�C; (c) bulk on MRS agar after incuba-
tion at 44�C; (d) bulk on M17 agar after
incubation at 44�C; (e) bulk on M17 agar after
incubation at 30�C. Lanes: 1, raw water
buffalo milk; 2, NWC; 3, curd after ripening;
4, final product
MICROBIAL SUCCESSION IN MOZZARELLA CHEESE 267
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
Analysing the Fig. 1 it is immediately clear that, as
expected, the NWC used as starter had a great influence on
the whole process. In fact, the complex pattern of the raw
water buffalo milk is soon simplified after the starter
addition. Throughout the process, all the samples displayed
a fingerprint identical to the one shown by the NWC.
Noteworthy, the Fig. 1 shows that all the microbial entities
composing the NWC are detected from the starter addition
until the end of the process.
The analysis of the cultivable community provided with
the following data. According to IB1 values, it was shown
that the mesophilic LAB represented the group with the
highest diversity in this process. It has been generally
recognized that the microflora involved in the manufactur-
ing of water buffalo mozzarella cheese mainly consisted of
thermophilic species (Limsowtin et al. 1995; Ottogalli 1998),
while the mesophilic lactobacilli were usually not considered
(Coppola et al. 1988, 1990; Parente et al. 1997). However, in
this case, mesophilic LAB able to grow on Rogosa and MRS
at 30�C appeared to be a significant group. They gave a high
number of CFU and also displayed the highest number of
bands in DGGE analysis of bulk (Fig. 2) resulting in high
values of IB1 indexes (Fig. 3). This is consistent with the
high levels of the mesophilic Lactobacillus fermentum and
L. crispatus in NWC found in a previous study (Ercolini
et al. 2001b).
0 0·2 0·4 0·6 0·8 1
Rogosa at 30°C
MRS at 30°C
MRS at 44°C
M17 at 44°C
M17 at 30°C
IB1
Fig. 3 Bars diagram showing the distribution
of IB1 values of DGGE profiles of bulk from
all the dairy samples collected during the
cheese making. Lines indicate the arbitrary
subgroups according to IB1 values. (j), Milk;
(j), NWC; (() curd; (j), final product
0·1
0·2
0·3
0·4
0·5
0·6
0·7
0·8
0·9
1
0 0·1 0·2 0·3 0·4 0·5 0·6 0·7 0·8 0·9 1IC1
IB1
Curd
Curd
Curd
Curd
Curd
NWC
NWC
NWC
NWC
NWC
Milk
Milk
Milk
Milk
Milk
FP
FP
FP FP
FP
Fig. 4 Scatter plot of the similarity indexes IC1
and IB1 of DGGE profiles of bulk from different
media and growth condition. (d), Rogosa at
30�C; (j), MRS at 30�C; (m), MRS at 44�C;
((), M17 at 44�C; (s), M17 at 30�C
268 D. ERCOLINI ET AL.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
Mesophilic lactic streptococci monitored on M17 incuba-
ted at 30�C showed high viable counts but low microbial
diversity (IB1 values lower than 0Æ4; Fig. 3), suggesting that
only a narrow group of species is involved in the process. By
contrast, Morea et al. (1999) found a significant diversity of
species of both mesophilic and thermophilic streptococci in
mozzarella cheese made from cow milk.
The index IC1 (Nei and Li 1979) was calculated in order
to estimate the similarity of each dairy sample with the curd
after ripening. This index had a technological meaning. For
milk and NWC samples the IC1 values represented an
estimation of the technological importance of that type of
organisms for the process. In fact, the micro-organisms still
present in the curd are considered to be responsible for the
acidification of milk and curd ripening with the consequent
release of compounds important for texture and flavour of
the final product (Moio et al. 1993; Mauriello et al. 2001,
2003). However, for the final product samples, IC1 values
represented the influence of the hot-stretching phase and its
effect on the microflora of the curd. Showing the highest
IC1 values, the NWC profiles contained a significant
number of bands that occurred also in the curd profile.
Hence, the microbial species of the NWC strongly
contribute to the process and, once added to the milk, out
compete the other species arising from milk and environ-
ment of production leading to curd ripening and influencing
the quality of the final product. The competition is mainly
because of the technological conditions such as temperature,
short time of ripening (4 h) and use of raw milk as
substrate. Traditional mozzarella cheese is made from
nonpasteurized milk and this is one of the conditions
recommended by the official rules of production followed
by the PDO water buffalo mozzarella producers. However,
milk profiles showed high or medium level of microbial
diversity but often had only few bands in common with
curd suggesting that the species occurring in the milk did
not grow during the ripening remaining at concentrations
lower than the species occurring in the NWC. The
technological influence of the NWC as resulted from the
IC1 estimation is absolutely consistent with the results of
the PCR–DGGE analysis of the DNA directly extracted
from the dairy samples (Fig. 1). The NWC has been
already defined as a complex ecosystem (Coppola et al.1988, 1990) and it is believed to represent the strength of
the traditional mozzarella cheese manufacture. The micro-
flora of the NWC mainly arises from milk as well as the
environment of the farm and processing areas. However,
only few species are naturally selected as capable of utilizing
milk nutrients, providing acidification and metabolic activ-
ities leading to typical flavours and texture. The use of
the NWC also identifies the product identity and it has
been shown to be linked to the geographical origin
(Mauriello et al. 2003).
In this study, the fate of the starter and its effectiveness
could be checked in 24 h and it has been ascertained that the
microflora of the NWC dominates in all the samples
collected during the process. The water buffalo mozzarella
cheese manufacture monitored in this study led to a top
quality product. Therefore the microbial succession could be
registered as fingerprints of microbial groups involved in a
premium quality production. This procedure might find a
useful application for the general monitoring of nonpremi-
um quality products where the poor quality arises from the
lack of development of the NWC or one or more of
the microbial groups targeted in this study. Consequently,
the procedure might allow ranking product quality when
nonpremium products are found.
This method can be easily applied to other plants allowing
process development and starter effectiveness to be checked
by analysing dairy samples by PCR–DGGE. The molecular
approaches can be considered a step forward for the
innovation of tracing systems in food technology and may
play an important role in the quality control of traditional
products allowing the preservation of their typical identity
and the consumer protection when territory claims are
involved.
ACKNOWLEDGEMENTS
This work was financed by a grant of National Research
Council (CNR), Rome, Italy (grant Agenzia 2000 no.
G00B58E) and by a grant of MURST (Rome, Italy). The
authors would like to thank Immacolata Tagliamonte for
technical collaboration.
REFERENCES
Ampe, F., Ben Omar, N., Moizan, C., Wacher, C. and Guyot, J.-P.
(1999) Polyphasic study of the spatial distribution of microorganisms
in Mexican pozol, a fermented maize dough, demonstrates the need
for cultivation-independent methods to investigate traditional
fermentations. Applied and Environmental Microbiology 65, 5464–
5473.
Cocolin, L., Manzano, M., Cantoni, C. and Comi, G. (2001)
Denaturing gradient gel electrophoresis analysis of the 16S rRNA
gene V1 region to monitor dynamic changes in the bacterial
population during fermentation of Italian sausages. Applied and
Environmental Microbiology 67, 5113–5121.
Coppola, S., Parente, E., Dumontet, S. and La Peccerella, A. (1988)
The microflora of natural whey cultures utilized as starter in the
manufacture of mozzarella cheese from water-buffalo milk. Le Lait
68, 295–310.
Coppola, S., Villani, F., Coppola, R. and Parente, E. (1990)
Comparison of different starter systems for water-buffalo mozzarella
cheese manufacture. Le Lait 70, 411–423.
Coppola, S., Blaiotta, G., Ercolini, D. and Moschetti, G. (2001)
Molecular evaluation of microbial diversity occurring in different
MICROBIAL SUCCESSION IN MOZZARELLA CHEESE 269
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x
types of mozzarella cheese. Journal of Applied Microbiology 90, 414–
420.
Ercolini, D., Moschetti, G., Blaiotta, G. and Coppola, S. (2001a)
Behavior of variable V3 region from 16S rDNA of lactic acid bacteria
in denaturing gradient gel electrophoresis. Current Microbiology 42,
199–202.
Ercolini, D., Moschetti, G., Blaiotta, G. and Coppola, S. (2001b) The
potential of a polyphasic PCR-DGGE approach in evaluating
microbial diversity of natural whey cultures for water-buffalo
mozzarella cheese production: bias of �culture dependent� and
�culture independent� approaches. Systematic and Applied Microbio-
logy 24, 610–617.
Ercolini, D., Blaiotta, G., Moschetti, G. and Coppola, S. (2002)
Molecular typing of cheeses on the basis of their microflora as
detected by PCR-DGGE analysis. Annals of Microbiology 52, 81–87.
Ercolini, D., Hill, P.J. and Dodd, C.E.R. (2003) Bacterial community
structure and location in Stilton cheese. Applied and Environmental
Microbiology 69, 3540–3548.
Limsowtin, G.K.Y., Powell, I.B. and Parente, E. (1995) Types of
starters. In Dairy Starter Cultures ed. Cogan, T.M. and Accolas, J-P.
pp. 101–129. New York, NY: VCH.
Mauriello, G., Moio, L., Moschetti, G., Piombino, P., Addeo, F. and
Coppola, S. (2001) Characterization of lactic acid bacteria strains on
the basis of neutral volatile compounds produced in whey. Journal of
Applied Microbiology 90, 928–942.
Mauriello, G., Moio, L., Genovese, A. and Ercolini, D. (2003)
Relationships between flavouring capabilities, bacterial composition
and geographical origin of natural whey cultures (NWCs) used for
traditional water-buffalo mozzarella cheese manufacture. Journal of
Dairy Science 86, 486–497.
Moio, L., Dekimpe, J., Etievant, P. and Addeo, F. (1993) Volatile
flavour compounds of water buffalo mozzarella cheese. Italian
Journal of Food Science 5, 57–68.
Morea, M., Baruzzi F. and Cocconcelli, P.S. (1999) Molecular and
physiological characterisation of dominant bacteral populations in
traditional mozzarella cheese processing. Journal of Applied Micro-
biology 87, 574–582.
Muyzer, G. (1999) DGGE/TGGE a method for identifying
genes from natural ecosystems. Current Opinion in Microbiology 2,
317–322.
Muyzer, G., De Waal, E.C. and Uitterlinden, A.G. (1993) Profiling of
complex microbial populations by denaturing gradient gel electro-
phoresis analysis of polymerase chain reaction-amplified genes
coding for 16S rRNA. Applied and Environmental Microbiology 59,
695–700.
Nei, M. and Li, W. (1979) Mathematical model for studying genetic
variation in terms of restriction endonucleases. Proceedings of the
National Academy of Science 76, 5269–5273.
Ottogalli, G. (1998) A global comparative method for the classification
of world cheeses (with special reference to microbiological criteria).
Annals of Microbiology 48, 31–57.
Parente, E., Rota, M.A., Ricciardi, A. and Clementi, F. (1997)
Characterization of natural starter cultures used in the manufacture
of Pasta Filata cheese in Basilicata (Southern Italy). International
Dairy Journal 7, 775–783.
Randazzo, C.L., Torriani, S., Akkermans, A.D.L., de Vos, W.M. and
Vaughan, E.E. (2002) Diversity, dynamics and activity of bacterial
communities during production of an artisanal Sicilian cheese as
evaluated by 16S rRNA analysis. Applied and Environmental
Microbiology 68, 1882–1892.
270 D. ERCOLINI ET AL.
ª 2003 The Society for Applied Microbiology, Journal of Applied Microbiology, 96, 263–270, doi:10.1046/j.1365-2672.2003.02146.x