Vol.1, No.3, 146-151 (2009)
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Openly accessible at
2DGE-coomassie brilliant blue staining used to
differentiate pasteurized milk from reconstituted milk
Yajun Wu1, Ying Chen1*, Bin Wang1, Haiyan Wang2, Fei Yuan1, Guiming Zhao1
1Chinese Academy of Inspection and Quarantine, No. 3, Gaobeidian North Street, Chaoyang District, Beijing, 100123, China
2Internal Mongolia CIQ, No. 12, Erduosi Street, Huhehaote, 010020, China; yqychen@yahoo.com.cn
Received 27 October 2009; revised 4 September 2009; accepted 7 September 2009.
Differentiating pasteurized milk and reconsti-
tuted milk by scientific approach was necessary
to defend consumer from economic fraud of
wrong labeling. In this paper 2DGE (2 Dimen-
sional Gel Electrophoresis)-coomassie brilliant
blue staining method was employed and sig-
nificant color intensity changing was observed
among raw milk, pasteurized milk, UHT milk and
reconstituted milk. For example, the intensity of
10 protein spots including casein and lac-
toglobulin reduced more than two folds from
pasteurized milk to reconstituted milk. However,
DIGE (Differential Gel Electrophoresis) assay
showed that the majority protein remained simi-
lar level from pasteurized milk to reconstituted
milk. Therefore the color fading of coomassie
brilliant blue stained 2D gels may be due to other
biochemical reaction, such as Maillard reaction,
instead of protein degradation. Stability of 2DGE
pattern was confirmed by running six gels of the
same sample in parallel and software analysis
showed that all proteins were at similar level. Two
commercialized pasteurized milk samples and
one reconstituted milk sample were tested by
2DGE-coomassie blue staining method and re-
constituted milk could be easily identified.
Keywords: 2DGE; Coomassie Brilliant Blue;
Pasteurized Milk; Reconstituted Milk
China has recently become one of the ten biggest milk
producers in the world [1]. Fluid milk products include
pasteurized milk and reconstituted milk. Chinese product
standard of pasteurized milk (GB5408.1) commanded that
only fresh raw milk could be used as raw material for
pasteurized milk, while reconstituted milk made by re-
solving milk powder in water and sterilization was cate-
gorized as sterilized milk (GB5408.2). Compared to pas-
teurized milk, reconstituted milk underwent more com-
plicated thermal process including spray drying, pasteuri-
zation, UHT-treatment or in-bottle sterilization [2]. Be-
cause of higher cost at factory location, seasonal variation
and transportation, pasteurized milk claims higher price
than reconstituted milk. It is reported that the price for 8
tons of raw milk in China is about 20,000 YUAN, while the
price for 1 ton of imported milk powder is 14,000-15,000
YUAN which could be made into 8 tons of reconstituted
milk [3]. However, as a lot of literature pointed out, inten-
sive thermal treatment would compromise milk nutrition
and flavor [4], thus consumers prefer pasteurized milk to
reconstituted milk and were concerned at possible eco-
nomic fraud by labeling reconstituted milk as pasteurized
A few analysis techniques such as CE (Capillary Elec-
trophoresis), HPLC (High performance liquid chromatog-
raphy), ELSD (Evaporative Light-scattering Detector)
have been applied in differentiating pasteurized milk and
reconstituted milk [2,5]. In these methods, individual pro-
tein or sugar ingredient, for example furosine, lactoglobu-
line, HMF (hydroxymethylfurfural) is quantified, which
demands complicated pre-procession of milk sample. Re-
sults of above-mentioned studies revealed significant
change of protein component during the procession of milk.
In this paper, we reported the application of 2DGE (2 Di-
mension Gel Electrophoresis) technique in an overall
analysis of protein profile change related to milk thermal
procession, revealing a significant alteration of protein
component between pasteurized milk and reconstituted
milk. Compared to other methods, 2DGE is characterized
by simplicity in sample preparation, ability of parallel
treatment of several samples and being information-rich. In
recent years, 2DGE have been widely applied for food
analysis [6-8]. A number of research work have been done
in milk proteome such as Equidae milk [9], marsupial
Trichosurus vulpecula milk [10], early lacatation milk of
the tammar wallaby [11], κ-casein micro-heterogeneity in
bovine milk [12] and whey protein [13]. In our proteomic
*Corresponding author. Tel: 0086-10-85783587; Fax: 0086-10-85774634
Y. J. Wu et al. / HEALTH 1 (2009) 146-151
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/Openly accessible at
study of milk product, it showed that 2D patterns after
coomassie brilliant blue staining could differentiate pas-
teurized milk and reconstituted milk according to the
change of color intensity of some protein spots.
2.1. Material
Pooled raw milk sample was collected from Sanyuan
Dairy Company and immediately sent to milk processing
laboratory in Food Institute of China Agricultural Uni-
versity for heat processing. Dry milk powder was also
collected from the company and reconstituted in accor-
dance with the original milk/water ratio, then pasteurized.
After preparation, total protein concentration of each
sample was determined. Raw milk was centrifuged at
1100g, 20mins and fat cream was removed. Three com-
mercialized milk samples were bought from local su-
permarket including two pasteurized milk samples from
different supplier and one reconstituted milk sample. All
samples were stored at 4C for immediate use or at -80C.
2.2. Total Protein Concentration Determination
Total protein concentration was determined using Protein
Assay Kit (NoVagen, Merk, Darmstadt, Germany) fol-
lowing the instruction. The optical absorbance value was
recorded on ELISA reader (Thermo Fisher Scientific, MA,
USA). Protein concentration was calculated on the basis
of Absorbance-Concentration curve of reference BSA
2.3. 2DGE
In preparation for IEF running, 10μL milk was mixed
with 440μL of solubilization buffer consisting of 8 M urea,
400mg/L CHAPS, 40 mM Tris, 50mg/L pH 4.7–5.9 car-
rier ampholytes (Bio-rad, Hercules, California, USA) and
100 mM DTT. The sample was used to hydrate a 17cm
pH4.7–5.9 IPG strip for 12 h at room temperature. Hy-
drated IPG strips were focused in a PROTEAN IEF Sys-
tem (Bio-rad, Hercules, California, USA) at 100 V for 1 h
followed by 500 V for 1 h and 1 kV for 1 h before the
voltage was increased to 8 kV for a total of 100 kVh. In
the second dimensional SDS-PAGE assay, focused strips
were first balanced in equilibrium buffer I and buffer II,
then embedded with 0.5% agarose on top of 14% poly-
acrylamide gels (18×18 cm). Electrophoresis was per-
formed in PROTEAN II XL Cell (Bio-rad, Hercules,
California, USA) at 5 mA/gel for 2 h followed by 20
mA/gel for 16 h. Gels were stained with Coomassie Bril-
liant Blue G-250 and destained in 1% acetic acid. Images
were captured on Versadoc Imager (Bio-rad, Hercules,
California, USA) in transmission mode.
2.4. DIGE
Milk samples were labeled with Cy dye (CyDye DIGE
Fluors, GE Healthcare, Buckinghamshire, UK) according
to the instruction. Sample pooling strategy was modified
as 5μL milk labeled with 1μL cy working solution
(400pmol/μL). All of the 10μL labeled sample comprised
with 5μL pasteurized milk and 5μL reconstituted milk
was pooled together. 2DGE was run following above-
mentioned procedure.
2.5. Data Analysis
2DGE profiles caught by Versadoc imager were analyzed
with PDQuest software 7.4.0 (Bio-rad, Hercules, Cali-
fornia, USA). After automatic spot detection, spot view
was performed to display quantity of protein spots. For
DIGE imaging, specific cy channel was selected.
2.6. Spot Digestion
In-gel digest was conducted following procedure of lit-
erature (Holland etal. 2004). Digestion product was pu-
rified using Ziptip C18 pipette tip (Millipore, Danvers,
MA, USA) following instruction. In the final elution step,
peptide was dissolved in 10mg/mL a-cyano-4-hydroxy-
cinnamic acid in 0.1%TFA/50%ACN and directly applied
to MALDI-TOF analysis.
2.7. MALDI-TOF and Database Search
One microliter purified peptide solution was spotted onto
a stainless steel MALDI target. Spectra were acquired
using a 4700 MALDI-TOF mass spectrometer (Applied
Biosystems, Foster City, CA, USA) in delayed extraction
mode. Tryptic digests were analyzed in positive ion re-
flectron mode with an accelerating voltage of 20 kV, grid
voltage at 64% and a delay time of 165 ns. One hundred
laser shots were accumulated for each spectrum. Peptide
mass fingerprint (PMF) of cut protein spot was analyzed
by MS-fit program of the ProteinProspector software
(University of California, USA). SwissProt.20071010
database was searched and searching parameter was set as
follow: Bos Taurus species, Tol 1 Da, Min matched pep-
tide set as 6.
Raw milk sample was taken from pooled milk container
to minimize heterogeneity of protein composition. Milk
was processed in lab to ensure the authenticity of proc-
essing condition. Different PI ranges were tried to de-
termine the best 2DGE condition (2DGE of methods
section). As shown in Figure 1, pH4.9-5.7 IPG strip pro-
duced the most satisfied 2D pattern in terms of protein
spot quantity and separating size among these spots.
Total protein concentration of raw milk, pasteurized
milk, UHT milk and reconstituted milk was calculated by
Biuret method as shown in Table 1. After comparison of
the 2D-coomassie brilliant blue staining patterns of the
four milk samples shown in Figure 2, we found that for
the majority of protein spots, color intensity decreased si
gnificantly from raw milk, pasteurized milk, UHT milk to
M. C. Hsiung et al. / HEALTH 1 (2009) 146-151
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Openly accessible at
SDS-PAGE was run in 7cm×7cm gels.
Figure 1. 2DGE profiles at different IPG range.
Ten microliter milk of each sample the protein concentration of
which has been modulated at the same level was loaded.
Figure 2. 2DGE profiles of raw milk, pasteurized milk, UHT
milk and reconstituted milk.
Those spots were begot by PDQuest software analysis. The analysis
set manager was defined as pasteurized milk two times higher in
quantity than reconstituted milk.
Figure 3. Quantitative comparison of ten protein spots among
raw milk, pasteurized milk, UHT milk and reconstituted milk.
reconstituted milk. Quantitative analysis by PDQuest
confirmed the trend. As shown in Figure 3, after the
analysis set template was defined as pasteurized milk
being two fold above reconstituted milk, ten spots were
Table 1. Protein concentration of raw milk, pasteurized milk, UHT
milk and reconstituted milk.
Assay Raw milkPasteurized milk UHT milk Reconstituted milk
A1 0.227 0.237 0.233 0.228
A2 0.224 0.235 0.227 0.231
A3 0.23 0.224 0.24 0.225
0.227 0.232 0.233 0.228
C(mg/ml) 23.5 23.8 23.8 23.6
detected. The quantity of the protein spots as represented
by the Y axis decreased when milk was heat treated and
the difference between pasteurized milk and reconstituted
milk was remarkable.
In order to prove that under standard operation proce-
dure 2DGE pattern was characterized by good repeat-
ability, six 2D gels of pasteurized milk were run in par-
allel. Raw maps of coomassie blue staining were shown in
Figure 4. All gels presented similar pattern and PDQuest
analysis showed similar quantity level for all protein
Protein spots were extracted, digested and identified by
MALDI-TOF. As shown in Figure 5, 16 spots were suc-
cessfully identified as casein and its isomers, lactoglobu-
lin and lactate dehydronese-like protein.
The above experiments demonstrated that after
coomassie blue staining, color intensity of most proteins
decreased when milk sample was processed under thermo
condition and the difference between pasteurized milk
and reconstituted milk was significant enough to be used
in product identification. As for the reason behind the
changing trend, Maillard reaction should be considered
other than proteins degradation since Maillard was the
most significant biochemical process during heat treat-
ment of milk [4,14]. In Maillard reaction, the reducing
sugar covalently binds to the epsilon amide residue of
lysine. We inferred that the intensity decrease of protein
spots was related to coomassie brilliant blue staining.
DIGE (2D Difference Gel Electrophoresis) was then
conducted to understand the true situation of protein
M. C. Hsiung et al. / HEALTH 1 (2009) 146-151
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Openly accessible at
2DGE profile was shown as raw map. Spot quantity was shown in the small box below the map.
Figure 4. Six 2DGE profiles of pasteurized milk run in parallel to confirm the stability of the method.
The left figure was peptide mass fingerprint of one protein after in-gel digestion. The right table showed identification re-
sults of several proteins after database search.
Figure 5. Identification of milk proteins by MALDI-TOF.
quantity changing between pasteurized milk and reconsti-
tuted milk. As shown in Figure 6, after pre-stained with
cy3 and cy5 respectively, pasteurized milk and reconsti-
tuted milk were mixed together and run in the same 2D
gel. Images of cy3 and cy5 showed that the majority of
protein spots had similar level of fluorescence intensity.
SSP IDProtein function
3701 Beta-casein
5901 Alpha-S2-casein
4901 Alpha-S2-casein
7901 Alpha-S2-casein
8901 Alpha-S2-casein
5701 L-lactate dehydrogenase A-like 6B
9401 Kappa-casein
3302 Bate-lactoglobulin
4101 Bate-lactoglobulin
1701 L-lactate dehydrogenase A-like 6B
4701 L-lactate dehydrogenase A-like 6B
6901 Alpha-S2-casein
6801 Alpha-S2-casein
7601 L-lactate dehydrogenase A-like 6B
3001 Alpha-S1-casein
3903 Alpha-S2-casein
M. C. Hsiung et al. / HEALTH 1 (2009) 146-151
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Profiles were shown as Gaussian map. Spot quantity was shown in the small box below the map. In each box, left column corresponded to recon-
stituted milk and right column to pasteurized milk.
Figure 6. DIGE profiles of pasteurized milk and reconstituted milk.
These three samples were bought from local supermarket.
Figure 7. 2DGE-coomassie brilliant blue staining profiles of three commercialized milk samples.
Openly accessible at
M. C. Hsiung et al. / HEALTH 1 (2009) 146-151
SciRes Copyright © 2009 http://www.scirp.org/journal/HEALTH/
Openly accessible at
PDQuest analysis indicated that the quantity difference
between pasteurized milk and reconstituted milk was less
than two folds. Thus we suppose the remarkable diverge
between cydye labeling and coomassie blue staining was
due to different staining mechanism. Cydye DIGE fluors
were designed to covalently attach to the epsilon amino
group of lysine of proteins in a “minimal labeling” way
which means the dyes labeled only on a single lysine per
protein molecule. Therefore Maillard reaction would not
affect staining efficiency. As for coomassie blue, the
mechanism of staining is still not well understood ever
since it was used in protein staining 45 years ago. How-
ever some literatures proved that coomassie blue varied
widely in its ability to bind proteins due to its affinity for
protein rich in basic amino acids such as lysine, arginine
and histidine and its poor ability to bind with glycoprotein
[15]. Moreover, since coomassie blue staining was not a
“minimal labeling” technique, its labeling efficiency
would be affected significantly by Maillard reaction. In
fact the content of available lysine has often been used as
marker of heat damage affecting dairy protein [16].
In order to verify the practicality of the method, three
commercialized milk samples plus the reference pas-
teurized milk sample were tested in parallel. As shown in
Figure 7, two pasteurized milk samples resulted in
similar 2D pattern as reference sample. The reconstituted
milk sample showed remarkably different 2D pattern.
The seriously reduced level of protein staining indicated
that the reconstituted milk sample might undergo stronger
thermo treatment than pasteurization.
Therefore, 2DGE-coomassie brilliant blue staining
technique was proved to be very useful in differentiation
between pasteurized milk and reconstituted milk in that
color intensity of most proteins decreased significantly
after heat treatment. DIGE experiment showed that true
quantity of these proteins did not change much. Change
of color intensity of different samples may be due to the
decrease of labeling efficiency of coomassie blue caused
by Mailard reaction after heat treatment. Therefore we
could apply 2DGE-coomassie brilliant blue staining
method for identification of pasteurized milk and recon-
stituted milk. Moreover 2DGE could be used to detect
adulterated milk product according to different 2D pattern
of protein from different organism or tissue.
This work was supported by the CAIQ (China Academy of Inspection and
Quarantine) Research Fund (2007JK005) and the Eleventh Five-year Plan
Research Fund of the Ministry of Science and Technology, People’s
Republic of China (2006BAD27B02). We greatly appreciate the kindness
of professor Ren Zhenfa of Chinese Agriculture University in helping with
sample preparation.
[1] A. Fagerberg, (2008) Field study on milk quality in China,
Internatinal Dairy Topics, 7(1), 11-12.
[2] J. D. Block, M. Merchiers, L. Mortier, A. Braekman, W.
Ooghe, R. V. Renterghem, (2003) Monitoring nutritional
quality of milk powders: capillary electrophoresis of the
whey protein fraction compared with other methods, In-
ternational Dairy Journal, 13, 87-94.
[3] G. X. Liu, (2003) Reconstituted milkPure fresh milk,
City Quality Supervision, 4, 46.
[4] M. Lacroix, J. Leonil, C. Bos, G. Henry, G. Airinei, J. Fau-
quant, D. Tome, C. Gaudichon, (2006) Heat markers and
quality indexes of industrially heat-treated [15N] milk protein
measured in rats, J. Agric. Food Chem., 54, 1508-1517.
[5] F. Guyomarch, F. Warin, D. Muir, J. Leave, (2000) Lac-
tosylation of milk proteins during the manufacture and
storage of skim milk powders, International Dairy Journal,
10, 863-872.
[6] M. Carbonaro, (2004) Proteomics: present and future in
food quality evaluation, Trends in Food Science &
Technology, 15, 209-216.
[7] M. A. Manso, J. Leonil, G. Jan, V. Gagnaire, (2005) Ap-
plication of proteomics to the characterization of milk and
dairy products, International Dairy Journal, 15, 845-855.
[8] R. O. Donnella, J. W. Hollanda, H. C. Deethb, P. Alewo-
oda, (2004) Milk proteomics, International Dairy Journal,
14, 1013-1023.
[9] G. Miranda, F. M. Mahé, C. Leroux, P. Martin, (2004)
Proteomic tools to characterize the protein fraction of
Equidae milk, Proteomics, 4, 2496-2509.
[10] S. Kuy, V. C. Kelly, A. M. Smit, D. J. Palmer, G. J. Coo-
per, (2007) Proteomic analysis of whey and casein pro-
teins in early milk from the marsupial Trichosurus
vulpecula, the common brushtail possum, Comparative
Biochemistry and Physiology, 2, 112-120.
[11] J. Joss, M. Molloy, L. Hinds, E. Deane, (2007) Proteomic
analysis of early lactation milk of the tammar wallaby
(Macropus eugenii), Comparative Biochemistry and
Physiology, 2, 150-164.
[12] J. W. Holland, H. C. Deeth, P. F. Alewood, (2004) Pro-
teomic analysis of k-casein micro-heterogeneity, Pro-
teomics, 4, 743-752.
[13] C. J. Hogarth, J. L. Fitzpatrick, A. M. Nolan, F. J. Young,
A. Pitt, P. D. Eckersall, (2004) Differential protein com-
position of bovine whey: A comparison of whey from
healthy animals and from those with clinical mastitis,
Proteomics, 4, 2094-2100.
[14] P. Resmini, L. Pellegrino, F. Masott, A. Tirelli, F. Prati,
(1992) Detection of reconstituted milk powder in raw and
in pasteurized milk by HPLC of furosine, Scienzae Tec-
nica Lattiero Casearia, 3(3), 169-186.
[15] K. McDonald, (2009) Overcoming the Coomassie blues.
Drug Discovery & Development.
http://www.dddmag.com/Article-Overcoming the
Coomas-Sie Blues-060409.aspx
[16] S. M. Rutherfurd, P. J. Moughan, (1997) Application of a
new method for determining digestible reactive lysine to
variably heated protein sources, J. Agric. Food Chem.,
45(5), 1582-1586.