Food Control 29 (2013) 131e137

Contents lists available at SciVerse ScienceDirect

Food Control
journal homepage: www.elsevier.com/locate/foodcont

A versatile real-time PCR method to quantify bovine contamination in buffalo

products
M.G. Drummond a, b, *,1, B.S.A.F. Brasil a, b,1, L.S. Dalsecco a, R.S.A.F. Brasil c, L.V. Teixeira d, D.A.A. Oliveira a
a

Laboratrio de Gentica Animal, Departamento de Zootecnia, Escola de Veterinria, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
Valid Biotechnology research team, INOVA, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
c
Escola de Engenharia, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
d
Departamento de Tecnologia e Inspeo de Produtos de Origem Animal, Escola de Veterinria, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais, Brazil
b

a r t i c l e i n f o

a b s t r a c t

Article history:
Received 8 March 2012
Received in revised form
15 May 2012
Accepted 22 May 2012

Fraudulent species substitution in food products is a reality in many markets around the world. The
consequences are serious and impact governments, consumers and producers. Molecular methods,
including PCR-based techniques, are available to prevent species substitutions and to authenticate food
content. Here, we describe a method for calculating the bovine and buffalo content in milk- and meatderived food products. The method applies real-time PCR using primers designed to specically amplify
bovine or buffalo DNA. The amplication efciencies of both primer sets were assessed using the TaqMan
and SYBR Green systems. Both sets of primers showed satisfactory results. A quantication procedure is
proposed and involves amplifying a sample with both primer sets and then normalizing the total DNA using
the total non-normalized bovine and buffalo DNA. To correct for the potential deviations between the real
and measured DNA quantity caused by biological differences between species, we propose the use of calibration curves generated from each analyzed matrix. These curves include a set of controlled admixtures of
bovine and buffalo material. Using this method for dairy samples containing known bovine and buffalo
content, the use of the calibration curve always approximated the measured to the expected quantities. This
technique was then tested on commercial samples. Our method was efcacious and reliable and can be
applied to the routine analysis of these products. Although we only tested the methodology on dairy
products, it can also be applied to other food matrices, such as meat-derived products. A patent document for
the technique has been submitted to the Brazilian Patent Ofce (INPI e Instituto Nacional da Propriedade
Industrial e www.inpi.gov.br) under the number 014110002844.
2012 Elsevier Ltd. Open access under the Elsevier OA license.

Keywords:
Food authentication
Bovine
Buffalo
Real-time PCR
Quantication

1. Introduction
The adulteration of foods with material from other species with
greater availability and/or lower cost is observed worldwide. An
example of this process is the addition of bovine milk to sheep, goat
and water buffalo dairy products (Herman, 2001). Other animal
products, such as seafood and meat-derived products, are also
subjected to content substitution (Carvalho, Pimenta-Neto, Brasil, &
Oliveira, 2011; Gonzlez-Crdova, Caldern de la Barca, Cota,
& Vallejo-Crdoba, 1998; Lahiff et al., 2001; Soares, Amaral,

* Corresponding author. Laboratrio de Gentica Animal, Escola de Veterinria,

Universidade Federal de Minas Gerais (UFMG), Avenida Antnio Carlos 6627, 31270010 Belo Horizonte, Minas Gerais, Brazil. Tel.: 55 31 34092206; fax: 55 31
34913963.
E-mail addresses: marcela.drummond@validbiotech.com, marceladrummond@
gmail.com (M.G. Drummond).
1
These authors contributed equally to this work.
0956-7135 2012 Elsevier Ltd. Open access under the Elsevier OA license.
doi:10.1016/j.foodcont.2012.05.051

Mafra, & Oliveira, 2010). These adulterations may pose health

risks to individuals with metabolic disorders or particular allergies,
such as to bovine-derived milk (Woolfe & Primrose, 2004). The food
industry may also suffer from nancial losses as a result of the
ethical, religious and cultural objections of consumers (Woolfe &
Primrose, 2004). The negative impacts of food adulteration
compel governing bodies to create more efcient public policies
that combat food mislabeling and fraud (Herman, 2001). Analytical
tools are therefore necessary, particularly if the food changes
drastically after processing, thus hampering or precluding the
visual identication of the product (Woolfe & Primrose, 2004).
In recent years, a variety of methods have been developed to
detect the fraudulent admixture of milk or meat from different
species. These techniques include lipid (Saeed, Ali, Rahman, &
Sawaya, 1989), protein (Addeo, Pizzano, Nicolai, Caira, & Chianese,
2009; Cozzolino, Passalacqua, Salemi, & Garozzo, 2002; Cuollo
et al., 2010; Gonzlez-Crdova et al., 1998; Guarino et al., 2010)
and DNA assays (Lanzilao, Burgalassi, Fancelli, Settimelli, & Fani,

132

M.G. Drummond et al. / Food Control 29 (2013) 131e137

2005; Maccabiani et al., 2005; Reale, Campanella, Merigioli, & Pilla,

2008; Zhang, Fowler, Scott, Lawson, & Slater, 2007). The lipid and
protein analyses, however, are laborious and most of the time
cannot be applied to heated or chemically treated products, such as
cooked or fermented foods (Lpez-Calleja et al., 2007; Mayer,
2005). In addition, they cannot distinguish between phylogenetically related species, such as cattle and buffalo (Bellis, 2003;
Branciari, Nijman, Plas, Di Antonio, & Lenstra, 2000; Zhang et al.,
2007). In contrast, the DNA-based methods are more practical,
sensitive and robust (Cuollo et al., 2010; Guarino et al., 2010).
Among the DNA-based techniques, PCR is often the method of
choice for identifying the species present in food products. Minimal
DNA quantities can be detected using species-specic primers
(Mafra, Ferreira, & Oliveira, 2007). The use of uorescent probes or
chemicals during a PCR reaction (real-time PCR) permits DNA
detection during the amplication cycles. This technique can be
used to quantify the initial DNA in a given sample (Kubista et al.,
2006). Real-time PCR is highly sensitive and is useful for situations in which occasional contamination is acceptable (LpezCalleja et al., 2007; Sawyer, Wood, Shanahan, Gout, & McDowell,
2003).
In the present study, we propose a real-time PCR system for
quantifying the bovine and buffalo material present in animalderived products. The TaqMan system was used, even though we
demonstrate that the technique described can also be used with the
SYBR Green system. Although previous studies have shown the
utility of real-time PCR for quantifying the contaminants in products
derived from diverse species (Lpez-Andreo, Lugo, Garrido-Pertierra,
Prieto, & Puyet, 2005; Lpez-Calleja et al., 2007; Sawyer et al., 2003;
Zhang et al., 2007), only one study specically addressed the adulteration of buffalo products (Lopparelli, Cardazzo, Balzan, Giaccone,
& Novelli, 2007). Nonetheless, the reported techniques exhibit high
measure variability mainly due to species-specic DNA content
variations in the diverse food matrices. In contrast, our method is
calibrated using a wide range curve of controlled frauds to normalize
the DNA quantity according to the food matrix. This process,
together with the proposed calculation method, makes the present
methodology more accurate and versatile than the previously
described techniques.

DNA was extracted from dairy products using methods described

by Boom et al. (1990), with some modications. The cheese samples
were sliced, and 6 g (total weight) was incubated overnight with
5 mL of extraction buffer (5 M guanidinium thiocyanate, 50 mM
TriseHCl pH 8.0, 20 mM EDTA, 1.3% Triton X-100). The samples were
then centrifuged (5 min, 2500 g), and the supernatant was diluted
with 5 mL of extraction buffer. One hundred and fty microliters of
activated silica (prepared according to Boom et al., 1990) was added
to each sample. The mixture was vortexed and incubated under
constant agitation at room temperature for 10 min. The silica, now
bound to DNA molecules, was then rapidly centrifuged. It was
washed three times with 500 mL of washing buffer (5 M guanidinium
thiocyanate, 50 mM TriseHCl pH 8.0), twice with 500 mL of 70%
ethanol and once with 500 mL of acetone. The silica particles were
dried by incubation at 95  C for 5 min. Finally, the DNA was recovered by adding 150 mL Milli-Q water, vortexing and incubating for
10 min at 95  C. The silica was spun down and discarded, and the
DNA-containing supernatant was recovered.
The DNA was extracted from meat products (5 mg) using
a phenolechloroform protocol, as described by Sambrook, Fritsch,
and Maniatis (1989).
The DNA quantity and quality were assessed using a NanoVue
Spectrophotometer (GE Healthcare, Waukesha, WI, USA). The 260/
280 nm measurement was considered a marker of DNA quality.
Negative controls (containing no dairy or meat sample) were
included in each extraction procedure.
2.2. Primer and probe design

2. Methodology

Bovine (Bos)-specic primers (Table 1) were designed to amplify

the mitochondrial gene CytB. They were modied from that
described by Zhang et al. (2007) for use with MGB TaqMan probes
(Applied Biosystems, Carlsbad, CA, USA). Buffalo (Bub)-specic
primers (Table 1) were designed to amplify the mitochondrial gene
16S. The MGB TaqMan probes were designed based on the
conserved CytB and 16S gene regions of Bos taurus, Bos indicus,
Bubalus bubalis, Capra hircus, Ovis aries and Sus scrofa. Sequences of
the CytB and 16S genes were obtained from the NCBI GenBank
database (Benson, Karsch-Mizrachi, Lipman, Ostell, & Sayers, 2010)
and aligned using the ClustalW algorithm (Chenna et al., 2003) in
MEGA version 5 (Tamura et al., 2011). The Primer Express Software
(Applied Biosystems) was used to design primers and probes.

2.1. Sample collection and DNA extraction

2.3. Real-time PCR reactions

Pure dairy and meat samples derived from bovine (Bos taurus
taurus or Bos taurus indicus) and buffalo (Bubalus bubalis) material
were used to test the efciency of the primer/probe sets (described
in section 2.4). These samples were also applied to create a calibration curve (section 2.6) and to test the technique presented here
(section 2.7). Commercial samples were acquired from local
markets to test the usefulness of the method.

All real-time PCR reactions were conducted in triplicates using

the 7500 Real-Time PCR System (Applied Biosystems). Each reaction contained 12.5 mL of TaqMan Universal PCR Master Mix
(Applied Biosystems), 400 nM of each specic primer, 200 nM of
specic probe and 2 mL of target DNA. The DNA concentrations used
in each reaction are described in the Results section. Five micrograms of BSA (Bovine Serum Albumin) were added to the reactions

Table 1
Primer/probe sets used in this study.
Primer/probe

Typea

Sequence (50 / 30 )

Target

BosF
BosR
MamCytB
Bub2F
Bub2R
Bub16S

Forward primer
Reverse primer
TaqMan e MGB probe
Forward primer
Reverse primer
TaqMan e MGB probe

CGGAGTAATCCTTCTGCTCACAGT
GGATTGCTGATAAGAGGTTGGTG
FAM e CATGAGGACAAATATC e MGBb
TCAGCCCAAAGAAAAATAAACCA
GTCACCCCAACCGAAACTGT
FAM e TAAGGARTAACAACAMTCT e MGBc

CytB (bovine)
CytB (bovine)
CytB (mammals)
16S (buffalo)
16S (buffalo)
16S (mammals)

a
b
c

Only the primer sets were used with the SYBR Green system.
In the test, the FAM uorophore was used to target the probes. Other uorophores can be used in its place.
The Bub16S probe was designed to undergo degeneration. According to IUPAC, R A or G, M A or C.

M.G. Drummond et al. / Food Control 29 (2013) 131e137

containing meat-derived DNA. The reaction conditions were

applied as follows: 50  C for 2 min, 95  C for 10 min and 40 cycles of
95  C for 15 s, 58  C for 15 s and 60  C for 1 min. In the SYBR Green
reactions, 12.5 mL of SYBR Green PCR Master Mix (Applied Biosystems) were used instead of the TaqMan mix, and no probe was
added. A dissociation curve was added at the end of PCR cycling, as
follows: 95  C for 15 s, 60  C for 1 min, 95  C for 30 s and 60  C for
15 s. The dissociation curves were analyzed after each experiment
to avoid non-specic amplication.
The reactions were analyzed using the 7500 Software (Applied
Biosystems). The program automatically corrected the basal uorescence measurement. The optimal thresholds were xed for each
primer/probe set (0.127651 for Bos and 0.025158 for Bub).
2.4. Amplication efciency
Dilution curves using pure bovine or buffalo DNA were used to
verify the amplication efciency of each primer/probe set (TaqMan) or primer (SYBR Green). The template DNA was diluted in a 1/
2 series and amplied as specied above. Using the 7500 Software,
Cycle threshold (Ct)  DNA quantity curves were constructed. The
curve slope, amplication efciency (E [101/slope]  1) and R2
were also calculated using the program. The R2 values were used to
verify the t of the trend line to the points on the curve. The Ct
range was dened as the interval between the Ctmin and Ctmax of
the curve. Similarly, the initial DNA quantity range was dened as
the lower and upper limits of total DNA on the curve. The specicity
(given in Ct) of each primer/probe set was measured using nontargeted DNA at the highest quantity used in the target DNA
curve. Similarly, the primer/probe sets were also tested against
non-target DNA derived from goats (C. hircus) and sheep (O. aries).
2.5. Data analysis
The quantication protocol used in this study included the realtime amplication of each sample using the bovine and buffalo
primer/probe sets. Thus, it was possible to determine the quantity
of cattle or buffalo DNA relative to the total material derived from
both species. The quantity of material derived from each species
was calculated using the following equations:

Relative quantity of Bos material 2

CtBos

=2

CtBos

CtBub

(1)
or

Relative quantity of Bub material 2CtBub =2CtBos 2CtBub

(2)
where CtBos and CtBub indicate the reaction cycle in which the
sample reached the pre-dened cycle threshold for the Bos or Bub
primer/probe set, respectively.
Equations (1) and (2) are based on the non-normalized quantities of Bos and Bub material in a given sample. According to
Schmittgen and Livak (2008), the non-normalized quantity N of
a specic DNA target in a sample is given by 2Ct for an approximately 100% efcient amplication reaction. Thus,

NBos 2CtBos

2.6. Calibration curves and automated quantication

Calibration curves were constructed for the dairy samples using
points of controlled ratios. We used this procedure to correct the
measured to the real quantity of material from bovine and buffalo
species. Pure cheese samples were weighed (nal weight: 6 g) and
mixed in the following proportions of Bos/Bos Bub content: 100%,
99%, 95%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, 20%, 10%, 5%, 1%, and 0%.
Each ratio was tested through the DNA extraction and amplication
procedures (TaqMan system) specied in sections 2.1 and 2.3.
Two independent curves were built using the same points of
controlled ratios. The mean relative bovine quantity measured for
each ratio was plotted against its real bovine quantity. Using
Microsoft Excel 2010, a trend line was added to the curve, and the
rst degree equation describing the curve was recorded. The
equation obtained from the calibration curve was then used to
adjust the measurements from the assays of other buffalo/bovine
dairy samples. A software program (named Calibration Software)
was designed for Microsoft Excel 2010 using Visual Basic for
Applications. This program automates the quantication calculations. The Results le exported from the 7500 Software containing Cts data must be used as the input le.
2.7. Method application
To test the developed method, controlled ratios derived from
pure cow and buffalo cheeses were assembled as described in
section 2.6. They consisted of Bos/Bos Bub content in the
following ratios: 7, 16, 30, 66, 78, 87, 97%. Then, 10 commercial
samples were analyzed. The 0, 100 and 50% Bos/Bos Bub samples
were processed as controls with each experiment. The DNA was
extracted, and the PCR reactions were performed (TaqMan system)
as previously described. The Result les from the 7500 Software
were used with our Calibration Software to calculate the real
bovine content of the tested samples.
3. Results
3.1. Validation of amplication efciency
The amplication efciencies of the Bos and Bub primer/probe
set (TaqMan) and primers (SYBR Green) were analyzed by testing
the DNA extracted from pure meat and cheese samples from both
species. The amplication efciencies, slopes, R2, specicity limits
and ranges of Ct and template DNA quantities obtained for each set
are summarized in Table 2. All combinations of primers, PCR
systems and food matrices yielded efciency values between 90
and 100% (Table 2). R2 values were 0.97 at minimum (Table 2). The
melting curves of the Bos and Bub primers yielded only one peak
with the expected TM values (Bos 77  C; Bub 75  C). Both
primer/probe sets showed a Ct  31 when tested against non-target
DNA from the other species (Table 2). DNA derived from pure sheep
and goat dairy samples were also tested. The same level of specicity was found (Ct  31 e data not shown). Thus, Cts above the
threshold of cross-species DNA amplication (31) were considered
to be negative.
3.2. Specic DNA quantication

(3)

and

NBub 2CtBub :

133

(4)

By normalizing using the sum of the non-normalized quantities

of Bos and Bub material, we can deduce the Equations (1) and (2).

To test the accuracy of the quantication method, DNA samples

obtained from two independent sets of controlled bovine/buffalo
cheese mixtures (calibration curves) were amplied using the
TaqMan system. Dairy samples were chosen only to demonstrate
the utility of our method. Any other food matrix, such as
hamburgers and sausages, could also have been used. Equation (1),

134

M.G. Drummond et al. / Food Control 29 (2013) 131e137

Table 2
Data from efciency experiments using Bos and Bub amplication sets.a
PCR system

Primer/probe
set

Sample
material

Efciency (%)

Slope

R2

Ct range

Initial DNA
amount range

Number of dilution
points in the curve

Specicity
limit (Ct)

TaqMan

Bos
Bos
Bub
Bub
Bos
Bub

Meat
Cheese
Meat
Cheese
Cheese
Cheese

96.0
91.6
90.1
92.7
93.1
92.9

3.42
3.54
3.58
3.51
3.49
3.50

0.99
0.99
0.99
0.97
0.99
0.98

25.2e30.5
16.8e24.0
28.7e32.9
17.8e25.2
16.7e22.0
15.7e20.0

0.098e0.003 ng
10e0.08 mg
0.08e0.005 ng
10e0.08 mg
1.25e0.04 mg
1.25e0.08 mg

6
8
5
8
6
5

37
31
35
35
40
40

SYBR Greenb
a
b

Each variable was measured as described in materials and methods section 2.3.
The SYBR Green system was used to verify the amplication efciency of both primer sets on dairy samples to illustrate the versatility of the technique.

described in the materials and methods section, was used to

calculate the bovine DNA in the mixtures. To maintain the Cts
within the efciency range of both primer sets, 220 ng/sample (2 mL
of a 110 ng/mL sample DNA, as measured by spectrophotometric
analysis) were used in all experiments. Good correlations between
the measured and expected values were observed for both sets of
samples (Fig. 1). Only minor deviations were observed. Interestingly, with the exception of the 10%, 20% and 40% Bos/Bos Bub
ratios, the results of the two independent test sets showed similar
patterns. The plotted points for the respective ratios in the two sets
were either both above or both below the expected values.
3.3. Specic DNA calibration curves
The deviations from the expected values described above (Fig. 1)
seemed to follow reproducible patterns. Thus, they may have been
a result of intrinsic systemic characteristics or variations in the
analyzed species and/or food matrices. In contrast, variations due to
experimental manipulation or technical errors would have been
expected to vary randomly. The observed phenomenon was
predictable because the number of somatic cells in a sample and
the number of mitochondria per cell may vary among the milk of
different species (Lopparelli et al., 2007; Zhang et al., 2007). To
correct for these intrinsic variations, we averaged the measured
relative quantity of Bos material for each ratio in the two test sets to
build a Bos/Bub reference curve for the dairy samples. These values
were plotted, and a trend line was added using Microsoft Excel

2010 (Fig. 2). The measured R2 value (0.998) showed that the
points were well represented by the straight trend line. The derived
trend line equation was then applied to the following experiments
using an automated spreadsheet (Calibration Software). This step
allowed us to correct the measured to the real quantity of the
bovine and buffalo material present in a sample.
3.4. Method validation
Admixtures of pure bovine and buffalo dairy samples were again
tested in controlled ratios. DNA was extracted and amplied from
these samples using the methodology described above for TaqMan
system. The Calibration Software was then used to calculate the real
bovine content in each sample. As shown in Table 3, the quantication method was effective in determining the real bovine content
of a given sample. The accuracy of the method was very high. The
measured value was at most 12.5% greater or less than the expected
value (e.g., expected 0.16 / measured 0.14). Importantly, the
calibration curve successfully corrected the calculated quantication to the expected value in all cases (Table 3).
Commercial samples purchased from local retailers were also
analyzed using our method (Table 4) to show its applicability to the
routine analysis of commercial buffalo products. Controls with 50,
100 and 0% Bos/Bos Bub content were tested with each experimental sample. A maximum of 6% deviation from the expected
value was observed for these controls (e.g., expected
0.5 / measured 0.47).

Fig. 1. Expected and measured bovine content in the diverse ratios of controlled mixtures of two calibration curves. Known amounts of bovine and buffalo cheese pure samples
were weighed and mixed to create ratios of controlled contamination for two calibration curves. DNA from the resulting admixtures was extracted and amplied using the Bos and
Bub primer/probe sets and the TaqMan system. Equation (1) was used to calculate the proportion of the bovine content in each sample.

M.G. Drummond et al. / Food Control 29 (2013) 131e137

135

Table 4
Proportion of Bos content in commercial samples.

Fig. 2. Calibration curve used to correct the error in the measured bovine quantity. The
mean bovine quantity measured for the different ratios in the two controlled
contamination experiments was plotted against the expected Bos content. A trend line
was added using Microsoft Excel 2010. The trend line equation was applied in the
following experiments to adjust for the real bovine content of buffalo-derived dairy
samples. R2 was calculated to assess the t of the trend line to the measured points.

4. Discussion
Several methods are currently being applied to authenticate
food products, as reviewed by Reid, ODonnell, and Downey (2006).
However, the complexity of food matrices and small quantities of
the contaminant present challenges to the sensitivity of the available techniques (Vallejo-Cordoba & Gonzlez-Crdova, 2010).
Among the DNA-based technologies, PCR was chosen for the
present study because it can quickly and specically amplify small
DNA quantities (Mafra et al., 2007) e even after heavy food processing (Rodrguez-Ramrez, Gonzlez-Crdova, & Vallejo-Cordoba,
2011). Mitochondrial DNA was chosen as the amplication target
because of its relative cellular abundance. This attribute increases
assay sensitivity (Lopparelli et al., 2007). Furthermore, the genetic
regions that were targeted are relatively small (approximately
100 bp). Thus, even DNA molecules degraded by the thermal,
chemical and/or physical processes through which food is prepared
can be detected.
Real-time PCR is commonly used to measure the quantity of
a DNA template in a given sample. Several methods are used to
calculate DNA quantity based on real-time PCR Ct results. One such
method is the relative quantication using comparative Ct (Livak &
Schmittgen, 2001). For this method, an internal control must be
applied to normalize the DNA quantities in different samples (Livak
& Schmittgen, 2001). In the majority of the studies that use realtime PCR to quantify contaminant species, normalization is

Table 3
Proportion of expected and measured Bos content in controlled contamination
experiments.
Ratios

Expecteda

Measuredb

After correctionc

0%
7%
16%
30%
50%
66%
78%
87%
97%
100%

0.00
0.07
0.16
0.30
0.50
0.66
0.78
0.87
0.97
1.00

0.00
0.07
0.13
0.30
0.46
0.65
0.75
0.85
0.96
1.00

0.00
0.07
0.14
0.30
0.47
0.66
0.76
0.86
0.97
1.00

Expected Bos content based on the bovine material added to the mixture.
Proportion of Bos content calculated using the method described in this
manuscript and Equation (1).
c
Corrected proportion of Bos content adjusted using the calibration curve.
b

Sample

Measureda

After correctionb

1
2
3
4
5
6
7
8
9
10

0.00
0.00
0.05
0.25
0.39
0.04
1.00
0.00
0.00
0.00

0.00
0.01
0.06
0.25
0.40
0.05
1.00
0.00
0.00
0.00

a
Proportion of Bos content as calculated using the method described in this
manuscript and Equation (1).
b
Proportion of Bos content corrected using the calibration curve.

performed using a eukaryotic sequence common to all involved

species (Lpez-Andreo et al., 2005; Lpez-Calleja et al., 2007;
Lopparelli et al., 2007; Sawyer et al., 2003). However, the amplication efciency of universal primers can be affected by the presence of more than one type of DNA in the sample and can vary
depending on the species involved (Lpez-Andreo et al., 2005). This
deviation in amplication efciency may be a result of minor
sequence variations between or within species that are out of the
researchers control. Taris, Lang, and Camara (2008) showed that
a large number of polymorphisms can interfere with real-time PCR.
In fact, when we used universal primers to normalize the DNA
quantities of diverse samples, we found a signicant difference in
the amplication efciencies obtained with pure or mixed samples
(data not shown). Thus, we used the sum of the non-normalized
Bos- and Bub-specic measures to normalize DNA quantity. Our
method is corroborated by the propositions that the nonnormalized quantity of a given DNA molecule can be calculated
using 2Ct (Schmittgen & Livak, 2008).
To validate our technique, we analyzed the amplication efciencies of the primer/probe sets. This standard validation step is
crucial for real-time PCR experiments in which more than one
target is amplied. In such cases, it is essential that both primer/
probe sets exhibit amplication efciencies of 100% (10%). This
condition was true for the two sets used in the present study. The
test developed here can be conducted using the Bos and Bub sets
with the TaqMan or SYBR Green systems for dairy products (as
shown in Table 3). The amplication efciencies were also tested
using the TaqMan system on meat-derived products, and the
results were satisfactory.
The quantication method proposed here was efcient in
calculating the relative bovine and buffalo quantities. We showed
high correlations between the measured and expected values in
the calibration curves (Fig. 1). The minor deviations may have
been a result of differences in the number of mitochondria per cell
in the different species. They may also have been attributable to
the variations of somatic cell numbers (and therefore DNA quantity) in the milk of the different species (Lopparelli et al., 2007;
Zhang et al., 2007). To minimize these deviations, we applied
calibration curves to adjust the measured to the real relative
content of bovine or buffalo material in a sample. Interestingly,
when testing random ratios of controlled contamination, we
observed that the calibration curve always corrected the calculated quantity to the expected quantity (Table 3). Although the
measurement deviations of the bovine/buffalo mixtures appear to
be minor, the variability intrinsic to other mixtures may be
signicant. One example of this latter case may be mixtures of
sheep and bovine dairy products (Drummond, M. G. and Brasil, B.
S. A. F. et al., e unpublished results).

136

M.G. Drummond et al. / Food Control 29 (2013) 131e137

Other authors have already applied similar approaches for

quantifying the contamination present in complex mixtures (LpezAndreo et al., 2005; Lpez-Calleja et al., 2007; Lopparelli et al., 2007;
Sawyer et al., 2003). The method described by Sawyer et al. (2003)
and Lopparelli et al. (2007) requires a curve to be constructed each
time an experiment is performed. Furthermore, the authors report
that their method is inefcient for dealing with between-species
variations that can lead to inaccurate quantication. Lpez-Andreo
et al. (2005) used a calibration curve similar to the described here
to test their quantication method in products derived from pork,
chicken, turkey, cattle and lamb. The results showed greater error
than those in our study, especially when the contamination level was
below 5%. Lpez-Calleja and colleagues (2007) used a calibration
curve to quantify the amount of cows milk in ewes milk. However,
their method only covers 0.5e10 % of contamination levels.
Dairy samples were chosen to show an application of our
method because the difference between the expected and observed
quantities may be large in these samples. This assumption is based
on the knowledge of the somatic cell number variation in the milk
of different species (Boutinaud & Ammes, 2002). Furthermore,
cheeses are complex matrices and are processed by pasteurization
processes that may damage DNA to a greater extent than processes
applied to produce sausages and hamburgers, for example. Interestingly, the DNA damage in cheese samples may explain the
relatively large template DNA quantities required to achieve highquality efciency curves in comparison with the meat-derived
curves (Table 2).
Overall, the technique described here improves upon the
currently available methods for quantifying the bovine and buffalo
material present in processed food products. Our method is applicable to forensic investigations of food mislabeling and can aid
regulatory agencies in controlling food fraud.
Disclosure statement
The method described here is the subject of the patent document 014110002844 at Brazilian Patent Ofce (INPI e www.inpi.
gov.br) and is protected under Brazilian regulations. MGD, BSAFB,
RSAFB and DAAO are the inventors on this patent document.
Funding source
This work was supported by the Brazilian agencies CNPq (INCT
573899/2008-8) and FAPEMIG (INCT APQ-0084/08). The funders
had no role in study design, data collection, analysis and interpretation, writing of the report or the decision to publish.
Acknowledgments
The authors thank Dr. Eduardo Bastianetto for his valuable
contributions to this work.
References
Addeo, F., Pizzano, R., Nicolai, M. A., Caira, S., & Chianese, L. (2009). Fast isoelectric
focusing and antipeptide antibodies for detecting bovine casein in adulterated
water buffalo milk and derived mozzarella cheese. Journal of Agricultural and
Food Chemistry, 57(21), 10063e10066.
Bellis, C. (2003). A molecular genetic approach for forensic animal species identication. Forensic Science International, 134(2e3), 99e108.
Benson, D. A., Karsch-Mizrachi, I., Lipman, D. J., Ostell, J., & Sayers, E. W. (2010).
GenBank. Nucleic Acids Research, 39(Database issue), D32eD37.
Boom, R., Sol, C. J., Salimans, M. M., Jansen, C. L., Wertheim-van Dillen, P. M., & van
der Noordaa, J. (1990). Rapid and simple method for purication of nucleic
acids. Journal of Clinical Microbiology, 28(3), 495e503.
Boutinaud, M. B., & Ammes, H. J. (2002). Potential uses of milk epithelial cells:
a review. Reproduction, Nutrition, Development, 42, 133e147.

Branciari, R., Nijman, I. J., Plas, M. E., Di Antonio, E., & Lenstra, J. A. (2000). Species
origin of milk in Italian mozzarella and Greek Feta cheese. Journal of Food
Protection, 63(3), 408e411, International Association for Food Protection.
Carvalho, D. C., Pimenta-Neto, D. A., Brasil, B. S. A. F., & Oliveira, D. A. A. (2011). DNA
barcoding unveils a high rate of mislabeling in a commercial freshwater catsh
from Brazil. Mitochondrial DNA, 22(S1), 97e105.
Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T. J., Higgins, D. G., et al. (2003).
Multiple sequence alignment with the Clustal series of programs. Nucleic Acids
Research, 31(13), 3497e3500.
Cozzolino, R., Passalacqua, S., Salemi, S., & Garozzo, D. (2002). Identication of
adulteration in water buffalo mozzarella and in ewe cheese by using whey
proteins as biomarkers and matrix-assisted laser desorption/ionization mass
spectrometry. Journal of Mass Spectrometry, 37(9), 985e991.
Cuollo, M., Caira, S., Fierro, O., Pinto, G., Picariello, G., & Addeo, F. (2010). Toward
milk speciation through the monitoring of casein proteotypic peptides. Rapid
Communications in Mass Spectrometry, 24(11), 1687e1696.
Gonzlez-Crdova, A. F., Caldern de la Barca, A. M., Cota, M., & Vallejo-Crdoba, B.
(1998). Immunochemical detection of fraudulent adulteration of pork chorizo
(sausage) with soy protein. Food Science and Technology International, 4(4),
257e262.
Guarino, C., Fuselli, F., La Mantia, A., Longo, L., Faberi, A., & Marianella, R. M.
(2010). Peptidomic approach, based on liquid chromatography/electrospray
ionization tandem mass spectrometry, for detecting sheeps milk in goats
and cows cheeses. Rapid Communications in Mass Spectrometry, 24(6),
705e713.
Herman, L. (2001). Determination of the animal origin of raw food by speciesspecic PCR. Journal of Dairy Research, 68(03), 429e436.
Kubista, M., Andrade, J. M., Bengtsson, M., Forootan, A., Jonk, J., Lind, K., et al.
(2006). The real-time polymerase chain reaction. Molecular Aspects of Medicine,
27(2e3), 95e125.
Lahiff, S., Glennon, M., OBrien, L., Lyng, J., Smith, T., Maher, M., et al. (2001). Speciesspecic PCR for the identication of ovine, porcine and chicken species in meat
and bone meal (MBM). Molecular and Cellular Probes, 15(1), 27e35.
Lanzilao, I., Burgalassi, F., Fancelli, S., Settimelli, M., & Fani, R. (2005). Polymerase
chain reaction-restriction fragment length polymorphism analysis of mitochondrial cytb gene from species of dairy interest. Journal of AOAC International,
88(1), 128e135.
Livak, K. J., & Schmittgen, T. D. (2001). Analysis of relative gene expression data
using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods,
25(4), 402e408.
Lpez-Andreo, M., Lugo, L., Garrido-Pertierra, A., Prieto, M. I., & Puyet, A. (2005).
Identication and quantitation of species in complex DNA mixtures by realtime polymerase chain reaction. Analytical Biochemistry, 339(1), 73e82.
Lpez-Calleja, I., Gonzalez, I., Fajardo, V., Martin, I., Hernandez, P., Garcia, T., et al.
(2007). Real-time TaqMan PCR for quantitative detection of cows milk in ewes
milk mixtures. International Dairy Journal, 17(7), 729e736.
Lopparelli, R. M., Cardazzo, B., Balzan, S., Giaccone, V., & Novelli, E. (2007). Real-time
TaqMan polymerase chain reaction detection and quantication of cow DNA in
pure water buffalo mozzarella cheese: method validation and its application on
commercial samples. Journal of Agricultural and Food Chemistry, 55(9),
3429e3434.
Maccabiani, G., Pavoni, E., Tilola, M., Agnelli, E., Simoni, M., DAbrosca, F., et al.
(2005). Setting-up a PCR-based method for species identication in milk
products. Veterinary Research Communications, 29(Suppl. 2), 327e329.
Mafra, I., Ferreira, I., & Oliveira, B. (2007). Food authentication by PCR-based
methods. European Food Research and Technology, 227(3), 649e665.
Mayer, H. (2005). Milk species identication in cheese varieties using electrophoretic, chromatographic and PCR techniques. International Dairy Journal, 15(6e9),
595e604.
Reale, S., Campanella, A., Merigioli, A., & Pilla, F. (2008). A novel method for species
identication in milk and milk-based products. Journal of Dairy Research, 75(1),
107e112.
Reid, L. M., ODonnell, C. P., & Downey, G. (2006). Recent technological advances for
the determination of food authenticity. Trends in Food Science & Technology,
17(7), 344e353.
Rodrguez-Ramrez, R., Gonzlez-Crdova, A. F., & Vallejo-Cordoba, B. (2011).
Review: authentication and traceability of foods from animal origin by polymerase chain reaction-based capillary electrophoresis. Analytica Chimica Acta,
685(2), 120e126.
Saeed, T., Ali, S. G., Rahman, H. A., & Sawaya, W. N. (1989). Detection of pork and lard
as adulterants in processed meat: liquid chromatographic analysis of derivatized triglycerides. Association of Ofcial Analytical Chemists, 72(6), 921e925.
Sambrook, J., Fritsch, E. F., & Maniatis, T. (1989). Molecular cloning: A laboratory
manual (2nd ed.). Cold Spring Harbor, NY: Cold Spring Harbor Laboratories.
Sawyer, J., Wood, C., Shanahan, D., Gout, S., & McDowell, D. (2003). Real-time PCR
for quantitative meat species testing. Food Control, 14(8), 579e583.
Schmittgen, T. D., & Livak, K. J. (2008). Analyzing real-time PCR data by the
comparative CT method. Nature Protocols, 3(6), 1101e1108.
Soares, S., Amaral, J. S., Mafra, I., & Oliveira, M. B. P. P. (2010). Quantitative detection
of poultry meat adulteration with pork by a duplex PCR assay. Meat Science,
85(3), 531e536.
Tamura, K., Peterson, D., Peterson, N., Stecher, G., Nei, M., & Kumar, S. (2011).
MEGA5: molecular evolutionary genetics analysis using maximum likelihood,
evolutionary distance, and maximum parsimony methods. Molecular Biology
and Evolution, 28(10), 2731e2739.

M.G. Drummond et al. / Food Control 29 (2013) 131e137

Taris, N., Lang, R. P., & Camara, M. D. (2008). Sequence polymorphism can produce
serious artefacts in real-time PCR assays: hard lessons from Pacic oysters. BMC
Genomics, 9, 234.
Vallejo-Cordoba, B., & Gonzlez-Crdova, A. F. (2010). Capillary electrophoresis for
the analysis of contaminants in emerging food safety issues and food traceability. Electrophoresis, 31(13), 2154e2164.

137

Woolfe, M., & Primrose, S. (2004). Food forensics: using DNA technology to
combat misdescription and fraud. Trends in Biotechnology, 22(5),
222e226.
Zhang, C., Fowler, M., Scott, N., Lawson, G., & Slater, A. (2007). A TaqMan real-time
PCR system for the identication and quantication of bovine DNA in meats,
milks and cheeses. Food Control, 18(9), 1149e1158.