H
OH
OH
metabolites
Article
Complex Mixture Analysis of Organic Compounds in
Yogurt by NMR Spectroscopy
Yi Lu, Fangyu Hu, Takuya Miyakawa and Masaru Tanokura *
Department of Applied Biological Chemistry, Graduate School of Agricultural and Life Sciences,
The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan; [email protected] (Y.L.);
[email protected] (F.H.); [email protected] (T.M.)
* Correspondence: [email protected]; Tel.: +81-3-5841-5165; Fax: +81-3-5841-8023
Academic Editor: Jun Kikuchi
Received: 26 April 2016; Accepted: 13 June 2016; Published: 16 June 2016
Abstract: NMR measurements do not require separation and chemical modification of samples and
therefore rapidly and directly provide non-targeted information on chemical components in complex
mixtures. In this study, one-dimensional (1 H, 13 C, and 31 P) and two-dimensional (1 H-13 C and 1 H-31 P)
NMR spectroscopy were conducted to analyze yogurt without any pretreatment. 1 H, 13 C, and 31 P
NMR signals were assigned to 10 types of compounds. The signals of α/β-lactose and α/β-galactose
were separately observed in the 1 H NMR spectra. In addition, the signals from the acyl chains of milk
fats were also successfully identified but overlapped with many other signals. Quantitative difference
spectra were obtained by subtracting the diffusion ordered spectroscopy (DOSY) spectra from the
quantitative 1 H NMR spectra. This method allowed us to eliminate interference on the overlaps;
therefore, the correct intensities of signals overlapped with those from the acyl chains of milk fat
could be determined directly without separation. Moreover, the 1 H-31 P HMBC spectra revealed for
the first time that N-acetyl-D-glucosamine-1-phosphate is contained in yogurt.
Keywords: NMR; assignment; yogurt; complex mixture analysis; existing state; DOSY; quantification
1. Introduction
NMR spectroscopy is highly quantitative and reproducible, and its sensitivity does not depend on
the types of metabolites [1]. As a non-targeted method, NMR measurements do not require separation
and chemical modification; therefore, comprehensive information regarding the chemical components
of mixtures can be rapidly and directly provided [2,3]. For the past decade, NMR has been recognized
as a powerful technique for discerning the chemical properties of complex mixtures and has been
widely applied to identify organic compounds in foods, such as milk [4], soy sauce [5], coffee [6],
wine [7], mango juice [8], and green tea [9]. Among NMR-based comprehensive analyses, 1 H NMR
is the most useful tool because of its informative spectral patterns and high-throughput acquisition.
Compared with the NMR signals of isolated compounds, chemical shift changes are often caused by
interactions with other compounds in food mixtures [10,11]. In addition, food mixtures can also lead
to extreme signal overlaps. Therefore, 13 C NMR spectra and two-dimensional NMR spectroscopy are
necessary to help assign each NMR signal in complex food mixtures.
Yogurt is traditionally made from milk fermentation by thermophilic cultures of Lactobacillus
delbrueckii subsp. bulgaricus and Streptococcus thermophilus. L. delbrueckii subsp. bulgaricus metabolizes
sugar to lactic acid with small amounts of by-products [12]. The benefits of yogurt include lactose
digestion, intestinal microflora modulation, cholesterol reduction, immune system stimulation,
and cancer prevention [13]. Texture is also a factor that influences the quality of yogurt and is
related to sensory perception of food products. Therefore, the chemical properties of yogurt can help
us analyze and improve the functional and gustatory qualities of yogurt [14].
Metabolites 2016, 6, 19; doi:10.3390/metabo6020019
www.mdpi.com/journal/metabolites
Metabolites 2016, 6, 19
2 of 11
Metabolites
2016, 6, 19 of food products. Therefore, the chemical properties of yogurt can help us analyze
2 of 11
sensory perception
and improve the functional and gustatory qualities of yogurt [14].
Many yogurt studies have been dedicated to identifying the chemical compositions of yogurt
Many yogurt studies have been dedicated to identifying the chemical compositions of
[15–17]. In these studies, organic acids, acetaldehyde, and other compounds in yogurt have been
yogurt [15–17]. In these studies, organic acids, acetaldehyde, and other compounds in yogurt have
analyzed with conventional techniques, such as GC-MS [16] and LC-MS [17]. These methods require
been analyzed with conventional techniques, such as GC-MS [16] and LC-MS [17]. These methods
appropriate extraction, separation and chemical derivatization of individual components. However,
require appropriate extraction, separation and chemical derivatization of individual components.
qualitative and quantitative modifications of the original mixture can even be performed by a simple
However, qualitative and quantitative modifications of the original mixture can even be performed
treatment without any separation. A method for the simultaneous and nondestructive identification
by a simple treatment without any separation. A method for the simultaneous and nondestructive
of compounds in yogurt has not been established.
identification of compounds in yogurt has not been established.
In a previous study, we reported the signals assignments of NMR spectra of milk, and the
In a previous study, we reported the signals assignments of NMR spectra of milk, and the analysis
analysis of the existing states of components [10]. Here, we assigned the signals of yogurt by
of the existing states of components [10]. Here, we assigned the signals of yogurt by combining 1D
combining 1D and 2D NMR spectroscopy without changing the chemical compositions. In addition,
and 2D NMR spectroscopy without changing the chemical
compositions. In addition, quantitative
quantitative difference spectra between
quantitative 1H NMR spectra and diffusion ordered
1
difference spectra between quantitative H NMR spectra and diffusion ordered spectroscopy (DOSY)
spectroscopy (DOSY) spectra were applied to quantitatively analyze the fermented metabolites
spectra were applied to quantitatively analyze the fermented metabolites whose signals overlapped
whose signals overlapped with broad signals from milk fat.
with broad signals from milk fat.
2. Results and Discussion
2. Results and Discussion
2.1. 1H NMR Spectra of Yogurt
2.1. 1 H NMR Spectra of Yogurt
A typical 1H NMR spectrum of yogurt is shown in Figure 1A. Proton resonances of aliphatic
A typical 1 H NMR spectrum of yogurt is shown in Figure 1A. Proton resonances of aliphatic
groups were observed with considerable overlap in the region from 0.6 to 2.2 ppm. These signals
groups were observed with considerable overlap in the region from 0.6 to 2.2 ppm. These signals
were assigned to the acyl chains of fatty acids from milk fats. In addition, the signals at 1.23 and 1.30
were assigned to the acyl chains of fatty acids from milk fats. In addition, the signals at 1.23 and
ppm were assigned to lactic acid and alanine, respectively. Compared with the chemical shifts on the
1.30 ppm were assigned to lactic acid and alanine, respectively. Compared with the chemical shifts
database, the weak signals from 2.2 to 3.1 ppm were suggested to be methyl or methylene groups of
on the database, the weak signals from 2.2 to 3.1 ppm were suggested to be methyl or methylene
citrate, creatine, and lecithin, which are minor components of yogurt. These signals were finally
groups of citrate, creatine, and lecithin, which are minor components of yogurt. These signals were
assigned by interpreting the cross peaks on the 1H-13C HSQC, 1H-1H DQF-COSY, and 1H-13C CTfinally assigned by interpreting the cross peaks on the 1 H-13 C HSQC, 1 H-1 H DQF-COSY, and 1 H-13 C
HMBC spectra. The signals from 3.1 to 5.1 ppm were assigned to D-lactose and D-galactose. These
CT-HMBC spectra. The signals from 3.1 to 5.1 ppm were assigned to D -lactose and D-galactose.
signals were sharp and highly sensitive but heavily overlapped with one another. Therefore, their
These signals were sharp and highly sensitive but heavily overlapped with one another. Therefore,
precise assignments were achieved by the 13C NMR and 1H-13C HSQC spectra. However, the region
their precise assignments were achieved by the 13 C NMR and 1 H-13 C HSQC spectra. However,
from 5.5 to 9.0 ppm contained several weak signals that were considered to be the amide protons of
the region from 5.5 to 9.0 ppm contained several weak signals that were considered to be the
proteins. Caseins and lactoglobulins are the major proteins of cow's milk [18,19]. These proteins and
amide protons of proteins. Caseins and lactoglobulins are the major proteins of cow's milk [18,19].
their degradation products may be observed in the 1H NMR spectra of yogurt. The details of the
These proteins and their degradation products may be observed in the 1 H NMR spectra of yogurt.
signal assignments are summarized in Table 1.
The details of the signal assignments are summarized in Table 1.
Figure 1. Cont.
Metabolites 2016, 6, 19
3 of 11
Metabolites 2016, 6, 19
3 of 11
Figure 1. (A) The 11H NMR spectrum of yogurt. (B) The 13
C NMR spectrum of yogurt. Gal and lac are
Figure 1. (A) The H NMR spectrum of yogurt. (B) The 13 C NMR spectrum of yogurt. Gal and lac are
abbreviations for galactose and lactose, respectively.
abbreviations for galactose and lactose, respectively.
Table 1. Assignment of 1H and 13C signals of compounds in yogurt.
Table 1. Assignment of 1 H and 13 C signals of compounds in yogurt.
H
Chemical Shift (ppm)
1
Compound
Compound
D-lactose
-lactose
D
1H
Multiplicity
Multiplicity
J (Hz)
(Hz)
d(7.70)
d(7.70)
d(3.05)
t(8.85)
d(3.05)
t(8.85)
d(7.90)
t(7.83)
d(7.90)
t(7.83)
D -galactose
d(3.35)
D-galactose
d(3.35)
d(7.80)
d(7.80)
J
Chemical Shift (ppm)
H1
1
H
4.33
4.33
3.43
3.43
3.54
3.54
3.80
3.80
3.61
3.61
3.64
5.11
3.64
3.47
5.11
3.71
3.47
3.54
3.71
3.82
3.54
3.76
4.55
3.82
3.16
3.76
3.53
4.55
3.54
3.16
3.52
3.82
3.53
5.14
3.54
3.69
3.52
3.72
3.82
3.85
5.14
3.96
3.66
3.69
4.46
3.72
3.36
3.85
3.53
3.96
3.82
3.60
3.66
3.63
4.46
3.36
3.53
3.82
C
13
13
C
103.57
103.57
71.79
71.79
73.31
73.31
69.38
69.38
76.15
76.15
61.94
92.62
61.94
71.98
92.62
72.24
71.98
78.89
72.24
70.86
78.89
60.66
96.55
70.86
74.66
60.66
75.17
96.55
78.75
74.66
75.56
60.80
75.17
93.06
78.75
69.18
75.56
69.99
60.80
70.11
93.06
71.25
62.02
69.18
97.23
69.99
72.70
70.11
75.17
71.25
69.56
75.94
62.02
61.84
97.23
72.70
75.17
69.56
P
31
P
31
a
Assignment
Assignment a
CH-1′
CH-11
CH-21
CH-2′
CH-31
CH-3′
CH-41
CH-4′
CH-51
CH-5′
CH2 OH-61
CHCH-α1
2OH-6′
CH-α2
CH-α1
CH-α3
CH-α2
CH-α4
CH-α3
CH-α5
CH-α4
CH
2 OH-α6
CH-β1
CH-α5
CHCH-β2
2OH-α6
CH-β3
CH-β1
CH-β4
CH-β2
CH-β5
CH
CH-β3
2 OH-β6
CH-α1
CH-β4
CH-α2
CH-β5
CH-α3
CHCH-α4
2OH-β6
CH-α1
CH-α5
CH
2 OH-α6
CH-α2
CH-β1
CH-α3
CH-β2
CH-α4
CH-β3
CH-α5
CH-β4
CHCH-β5
2OH-α6
CH
2 OH-β6
CH-β1
CH-β2
CH-β3
CH-β4
Metabolites 2016, 6, 19
4 of 11
Table 1. Cont.
Compound
1H
Multiplicity J
(Hz)
acyl chains of fatty acids
Chemical Shift (ppm)
1H
13 C
0.76
1.17
1.45
14.74
23.54
25.60
1.89
27.93
29.78–30.90
32.79
34.37
130.18
62.54
69.59
22.72
175.40
44.93
75.80
177.70
181.40
20.84
68.95
182.50
17.09
51.49
176.4
37.40
54.78
158.30
175.40
54.28
acetic acid
s
1.17
1.14
2.11
5.19
3.98, 4.16
5.10
1.93
citric acid
AB
2.53,2.69
lactic acid
d(6.60)
dd
1.23
4.08
alanine
d(7.15)
1.36
3.66
creatine
s
s
2.90
3.81
lecithin
s
3.10
3.74, 3.80
4.18
3.54
3.74, 3.80
3.97
glycerol backbone of fats
cephalin
N-acetyl-D-glucosamine-1-phosphate
5.30
3.83
60.22
67.09
31 P
Assignment a
´0.20
´0.20
´0.20
0.33
0.33
CH3- ω1
CH2 -ω2
CH2 -∆3
oleate
CH2- ω8,11
CH2 (ω4-n)
CH2 -ω3
CH2 -∆2
oleate HC=CH
glycerol-1,3
glycerol-2
CH3
COOH
CH2
C
CH2 -COOH
COOH
CH3
CH
COOH
CH3
CH
COOH
CH3
CH2
C
COOH
trimethylamine
CH2 -3
CH2 -4
CH2 -5
CH2 -3
CH2 -4
´1.44
´1.44
O-CH
N-CH
a
The symbol “ω” indicates the position from the methyl group; the symbol “∆” indicates the position from the
ester group.
2.2.
13 C
NMR Spectra of Yogurt
A typical 13 C NMR spectrum of yogurt is shown in Figure 1B. The signals were well separated and
were completely assigned by utilizing the 1 H-13 C HSQC, 1 H-1 H DQF-COSY, and 1 H-13 C CT-HMBC
spectra. The signals from 10 to 55 ppm were assigned to the primary carbon atoms from the acyl
chains of fatty acids, lactic acid, and acetic acid. This region also contained some signals from minor
components, which were assigned to be alanine, creatine, and citrate. There were signals of tertiary
carbon atoms from the aliphatic rings of lactose and galactose in the region from 55 to 110 ppm.
The signals from 160 to 190 ppm were assigned to the quaternary carbon atoms of organic acids
and alanine based on the 1 H-13 C CT-HMBC spectrum. The details of the signal assignments are
summarized in Table 1.
2.3. Identification of Several Components with 2D NMR Spectra
Lactic acid. In the 1 H-1 H DQF-COSY spectra of yogurt (Figure 2A), the cross-peaks at 1.23 and
4.08 ppm were the correlations between –CH3 and –CH (Figure 2D), while the 1 H signals at 1.23 ppm
Metabolites 2016, 6, 19
2.3. Identification
Metabolites
2016, 6, 19 of Several Components with 2D NMR Spectra
5 of 11
5 of 11
Lactic acid. In the 1H-1H DQF-COSY spectra of yogurt (Figure 2A), the cross-peaks at 1.23 and
4.08 ppm were the correlations13between –CH3 and –CH (Figure 2D), 1while the 1H signals at 1.23 ppm
showed correlations with the 13C signals at 20.84 ppm, and the 1 H signals at 4.08 ppm showed
showed correlations with the C signals at 20.84 ppm, and the H signals at 4.08 ppm showed
correlations with the 13 C13signals at 68.95 ppm in the HSQC spectrum (Figure 2B). The 1 H1 signals at 1.23
correlations with the C signals at 68.9513ppm in the HSQC spectrum (Figure
2B). The H signals at
and 4.08 ppm were also connected to the C signals
at 182.41 ppm in the 1 H-13 C CT-HMBC spectrum
1.23 and 4.08 ppm were also connected to the 13C signals at 182.41 ppm in the 1H-13C CT-HMBC
(Figure
2C,D),
which2C,D),
indicated
correlations
of –CH3 and
spectrum
(Figures
which
indicated correlations
of–CH
–CH3with
and –COOH.
–CH with –COOH.
Figure 2. (A) The 1H–1H COSY NMR spectrum of yogurt. (B) The 1H–13C HSQC NMR spectrum of
Figure 2. (A) The 1 H–1 H COSY NMR spectrum of yogurt. (B) The 1 H–13 C HSQC NMR spectrum of
yogurt. (C) The 1H–13C HMBC NMR spectrum of yogurt. (D‒H) Correlations of lactic acid (D), acetic
yogurt. (C) The 1 H–13 C HMBC NMR spectrum of yogurt. (D-H) Correlations of lactic1 acid
(D), acetic
13C HMBC
acid (E), alanine (F), citrate (G), and creatine (H). Arrows show the cross-peaks in the H–
1 H–13
acid (E), alanine (F), citrate (G), and creatine (H). Arrows show1 the
cross-peaks
in
the
C HMBC
1H COSY NMR spectrum. Gal and
NMR spectrum, and bold bonds show the cross-peaks in the
H–
1
1
NMR spectrum, and bold bonds show the cross-peaks in the H– H COSY NMR spectrum. Gal and lac
lac are abbreviations for galactose and lactose, respectively.
are abbreviations for galactose and lactose, respectively.
Acetic acid. The 1H signals at 1.93 ppm showed correlations with the 13C signals at 22.72 ppm in
13 C signals at 22.72 ppm in
Acetic
acid.
The 1 H(Figure
signals2B)
at 1.93
correlations
with the
the HSQC
spectrum
and ppm
were showed
connected
to the 13C signals
at 175.46
ppm in the 1H-13C
13 C signals at 175.46 ppm in the 1 H-13 C
theCT-HMBC
HSQC spectrum
(Figure
and
were connected
to thebetween
spectrum
(Figure2B)
2C).
Therefore,
the correlation
–CH3, and –COOH of acetic acid
CT-HMBC
spectrum
(Figure
was confirmed
(Figure
2E). 2C). Therefore, the correlation between –CH3 , and –COOH of acetic acid
1
1
was confirmed
(Figure
2E).
Alanine. In the H- H DQF-COSY spectra of yogurt (Figure 2A), the cross peaks at 1.35 and 3.65
Alanine.
In the 1 HH DQF-COSY
spectra of
yogurt
the13Ccross
peaks at
1.35 and
ppm
were assigned
to 1the
correlations between
–CH
3 and (Figure
–CH. In 2A),
the 1HCT-HMBC
spectrum
1 H13 C13C
3.65(Figure
ppm were
to theatcorrelations
between
–CH3 and
–CH. In the
CT-HMBC
2C), assigned
the 1H signals
1.35 and 3.65
ppm showed
correlations
with
the
signals atspectrum
176.40
1
13
ppm,
which
confirmed
the
connectivity
among
–CH
3
,
–CH,
and
–COOH
of
alanine
(Figure
2F).
(Figure 2C), the H signals at 1.35 and 3.65 ppm showed correlations with the C signals at 176.40 ppm,
1H signals at 2.53 and 2.69 ppm showed correlations with the 13C signals at 44.93
Citrate. Thethe
which confirmed
connectivity among –CH3 , –CH, and –COOH of alanine (Figure 2F).
1
13 C signals
ppm
in the
HSQC
spectrum
andand
were
also
connected
the 13C signals
75.90,
177.18, and
180.81
Citrate.
The
H signals
at 2.53
2.69
ppm
showedtocorrelations
withatthe
at 44.93
ppm
1
13
13
ppm
(Figures
2C,G)
in
the
HC
CT-HMBC
spectrum,
which
suggested
that
the
signals
should
be
in the HSQC spectrum and were also connected to the C signals at 75.90, 177.18, and 180.81 ppm
1H signals at 2.40 and 2.54
1
13
assigned
to
citrate.
In
whole
milk,
the
citrate
signals
were
assigned
as
the
(Figure 2C,G) in the H- C CT-HMBC spectrum, which suggested that the signals should be assigned
13C signals at 44.93 ppm [4]. Citrate is the main organic acid in milk and binds to calcium
ppm, and
to citrate.
Inthe
whole
milk, the citrate signals were assigned as the 1 H signals at 2.40 and 2.54 ppm,
ions to
act as a component of casein micelles [20]. The variety of chemical shifts may reflect differences
13
and the C signals at 44.93 ppm [4]. Citrate is the main organic acid in milk and binds to calcium ions
in the existing states of casein micelles, pH, and temperature.
to act as a component of casein micelles [20]. The variety of chemical shifts may reflect differences in
the existing states of casein micelles, pH, and temperature.
Creatine. In the 1 H-13 C CT-HMBC spectrum (Figure 2C), the 1 H signals at 3.81 ppm were connected
to the 13 C signals at 37.40, 158.30, and 175.48 ppm, while the 1 H signals at 2.90 ppm showed correlations
Metabolites 2016, 6, 19
6 of 11
Metabolites 2016, 6, 19
6 of 11
Creatine. In the 1H-13C CT-HMBC spectrum (Figure 2C), the 1H signals at 3.81 ppm were
connected to the 13C signals at 37.40, 158.30, and 175.48 ppm, while the 1H signals at 2.90 ppm showed
with the 13 C signals at1354.78 and 158.30 ppm (Figure 2C,H), which indicated correlations of –CH3 with
correlations with the C signals at 54.78 and 158.30 ppm (Figures 2C,H), which indicated correlations
–CH
2 , –C[(NH)NH
2 ], and –COOH.
of –CH
3 with –CH2, –C[(NH)NH2], and –COOH.
1 H-13 C HSQC spectrum (Figure 2B), the 1 H signal at 3.10 ppm
Lecithin
and Cephalin.
Cephalin. InInthe
the1H13C HSQC spectrum (Figure 2B), the 1H signal at 3.10 ppm was
Lecithin and
13
13 C signals
was
correlated
with
the
C
signal
at
54.28and
ppm
and
also showed
a correlation
with
13C the
correlated with the 13C signal at 154.28
ppm
also
showed
a
correlation
with
the
signals
at 54.28
13 C HMBC spectrum (Figure 2C). In addition, the 1 H signal at
at
54.28
and
67.09
ppm
in
the
H1
13
1
and 67.09 ppm in the H- C HMBC spectrum
(Figure 2C). In addition, the H signal at 3.54 ppm
3.54
ppmashowed
a cross-peak
the 13 C
67.09
ppm.correlations
These correlations
indicated
that
13C signal
showed
cross-peak
with the with
atsignal
67.09 at
ppm.
These
indicated
that yogurt
1 H-31 P HMBC spectrum (Figure 3B),
yogurt
contained
compounds
with
trimethylamine
groups.
In
the
contained
compounds with trimethylamine groups. In the 1H-31P HMBC spectrum (Figure 3B), the
1 H signals at 3.54 and 4.18 ppm connected with the 31 P signal at ´0.20 ppm supported the
the
1H signals
at 3.54 and 4.18 ppm connected with the 31P signal at −0.20
ppm supported the existence of
existence
of
trimethylamine
groups(Figures
in yogurt
(Figure
The 1 H
at 3.74
ppm
1H signals
trimethylamine
groups in yogurt
3D,E).
The3D,E).
at signals
3.74 and
3.80 and
ppm3.80
showed
31 P signals at both ´0.20 and 0.33 ppm in the 1 H-31 P HMBC spectrum,
showed
correlations
with
the
correlations with the 31P signals at both −0.20 and 0.33 ppm in the 1H-31P HMBC spectrum, which
which
suggested
that
the signals
derived
compounds
similar
structures.
Lecithin
suggested
that the
signals
were were
derived
from from
compounds
with with
similar
structures.
Lecithin
and
23 and have similar structures. Enhancements of the
and
cephalin
are
the
main
phospholipids
in
milk
cephalin are the main phospholipids in milk23 and have similar structures. Enhancements of the
1 H-31 P HMBC spectrum were observed when standard reagents were added. Therefore,
signals
31P HMBC spectrum were observed when standard reagents were added. Therefore,
signals in
in the
the 1H31
the
P
signal
at
´0.2
ppmthat
thatconnected
connected the
the 11H
Hsignals
signalsat
at3.55,
3.55, 3.74,
3.74, 3.80,
3.80, and
and 4.19
4.19 ppm
ppm was
was assigned
assigned
31
the P signal at −0.2 ppm
31
1 H signals at 3.74, 3.80, and 3.97 ppm
to
lecithin,
and
the
P
signal
at
´0.2
ppm
that
connected
the
to lecithin, and the 31P signal at −0.2 ppm that connected the 1H signals at 3.74, 3.80, and 3.97 ppm
was
was assigned
assigned to
to cephalin.
cephalin.
Figure 3. (A) The 31P NMR spectrum of yogurt. (B) The 1H-31P HMBC NMR spectrum of yogurt. (C‒
Figure 3. (A) The 31 P NMR spectrum of yogurt. (B) The 1 H-31 P HMBC NMR spectrum of yogurt.
E) Correlations of N-acetyl-D-glucosamine-1-phosphate (C), lecithin (D), and cephalin (E). Arrows
(C-E) Correlations of N-acetyl-D -glucosamine-1-phosphate (C), lecithin (D), and cephalin (E). Arrows
31
show the
the cross-peaks
cross-peaks in
in the
the 11H–
show
H–31PPHMBC
HMBC NMR
NMR spectrum.
spectrum.
N-acetyl-D-glucosamine-1-phosphate. There were some signals that were not assigned in the 1H-31P
N-acetyl-D -glucosamine-1-phosphate. There were some signals that were not assigned in the
spectrum of yogurt. In a previous study, the signals of N-acetyl carbohydrates were detected
1HMBC
H-31 P HMBC spectrum of yogurt. In a previous study, the signals of N-acetyl carbohydrates
in whole milk [4]. We hypothesized that the compounds were phosphorylated during fermentation.
were detected in whole milk [4]. We hypothesized that the compounds were phosphorylated
To assign the remaining signals, we performed spiking experiments with the candidate compounds.
during fermentation. To assign the remaining signals, we performed spiking experiments with
The 1H-31P HMBC spectrum of N-acetyl-D-glucosamine-1-phosphate was observed using standard
the candidate compounds. The 1 H-31 P HMBC spectrum of N-acetyl-D -glucosamine-1-phosphate
reagents. Based on the 1H-31P HMBC spectrum (Figure 3B), the 31P signal at −1.44 ppm that connected
was observed using standard reagents. Based on the 1 H-31 P HMBC spectrum (Figure 3B),
the 1H signals at 3.83 and 5.30 ppm was assigned to N-acetyl-D-glucosamine-1-phosphate. N-acetylthe 31 P signal at ´1.44 ppm that connected the 1 H signals at 3.83 and 5.30 ppm was assigned
D-glucosamine-1-phosphate is synthesized from D-glucosamine-1-phosphate, which is one of the
to N-acetyl-D -glucosamine-1-phosphate. N-acetyl-D -glucosamine-1-phosphate is synthesized from
products made from lactic acid bacteria and is needed for the synthesis of UDP-α-N-acetyl-DD -glucosamine-1-phosphate, which is one of the products made from lactic acid bacteria and is
glucosamine-1-phosphate, which is required for UDP-N-acetyl-D-glucosamine biosynthesis [21].
needed for the synthesis of UDP-α-N-acetyl-D -glucosamine-1-phosphate, which is required for
UDP-α-N-acetyl-D-glucosamine-1-phosphate is an essential precursor of cell wall peptidoglycans,
UDP-N-acetyl-D -glucosamine biosynthesis [21]. UDP-α-N-acetyl-D -glucosamine-1-phosphate is an
lipopolysaccharides, and enterobacterial common antigens [22].
Metabolites 2016, 6, 19
7 of 11
essential precursor of cell wall peptidoglycans, lipopolysaccharides, and enterobacterial common
antigens [22].
Metabolites 2016, 6, 19
7 of 11
2.4. Concentrations of Yogurt Components
2.4. Concentrations of Yogurt Components
To control the quality and content of components in yogurt, we used a Bulgarian yogurt culture
To control the quality and content of components in yogurt, we used a Bulgarian yogurt culture˝
thatthat
was
prepared
byby
adding
a Bulgarian
milk. After
Aftercultivating
cultivatingatat4040
was
prepared
adding
a Bulgarianyogurt
yogurtinoculum
inoculum into
into whole
whole milk.
°C C
for for
24 h,
sample
was
applied
Thecomponents
componentswhose
whose
24 the
h, the
sample
was
appliedtotoNMR
NMRmeasurements
measurements for
for quantification.
quantification. The
1 H NMR spectra, while the other compounds were
signals
did
not
overlap
were
quantified
by
the
1
signals did not overlap were quantified by the H NMR spectra, while the other compounds were
quantified
using
thethe
difference
subtractingthe
theDOSY
DOSYspectra
spectrafrom
from
the
quantified
using
differencespectra
spectrathat
thatwere
wereobtained
obtained by
by subtracting
the
1 H 1NMR spectra.
H NMR spectra.
1 H1 NMR working curves of the citrate protons are shown in Figure 4A. We used the capillary
TheThe
H NMR working curves of the citrate protons
shown in Figure 4A. We used the capillary
containing
1,1,2,2-tetrachloroethane
the working
workingcurves
curves
and
quantified
the concentrations
containing
1,1,2,2-tetrachloroethane for the
and
quantified
the concentrations
of theof
TheThe
working
curves
showed
good linearity
and therefore
were usedwere
to calculate
theyogurt
yogurtcomponents.
components.
working
curves
showed
good linearity
and therefore
used to
the concentrations
of yogurt
components.
In addition,
the 1H NMR
spectra
the self-made
yogurt
calculate
the concentrations
of yogurt
components.
In addition,
the 1 H
NMRofspectra
of the self-made
˝
were
measured
at
40
°C
(Figure
4B),
and
the
signals
in
the
difference
spectra
were
used
forfor
yogurt were measured at 40 C (Figure 4B), and the signals in the difference spectra were used
quantification
because
signals
recovered
to original
their original
4B,D).
The
quantification
because
the the
signals
werewere
recovered
to their
shapesshapes
(Figure(Figures
4B,D). The
inoculum
inoculum
samples
were
prepared
in
triplicate
with
the
same
fermentation
process.
samples were prepared in triplicate with the same fermentation process.
1H NMR spectrum. The integral value of the signal
Figure
4. (A)
Working
curvesofofcitrate
citratefor
forthe
the1 H
Figure
4. (A)
Working
curves
NMR spectrum. The integral value of the signal
1H NMR
1 H NMR
arising
from
the
protons
of
1,1,2,2-tetrachloroethane
was
set to
(B) Quantitative
arising from the protons of 1,1,2,2-tetrachloroethane was set
to 100.
(B)100.
Quantitative
spectrum
spectrum
yogurt.
(C) DOSY
NMR
of yogurt.
(D) The spectrum
differencefor
spectrum
for quantifying
of yogurt.
(C)ofDOSY
NMR
spectrum
ofspectrum
yogurt. (D)
The difference
quantifying
the integral
the signal
integralatof4.08
the ppm
signalbetween
at 4.08 ppm
between
(C). (E) Concentrations
of yogurt components
of the
(B) and
(C). (B)
(E) and
Concentrations
of yogurt components
quantified
1H NMR signals. The standard deviation is shown in parentheses (n = 3). Gal and
quantified
using
the
1
using the H NMR signals. The standard deviation is shown in parentheses (n = 3). Gal and lac are
lac are abbreviations
for galactose
and respectively.
lactose, respectively.
abbreviations
for galactose
and lactose,
The concentrations of several yogurt components are shown in Figure 4E. The concentrations of
The concentrations of several yogurt components are shown in Figure 4E. The concentrations
α-D-galactose, β-D-galactose, α-D-lactose, β-D-lactose, citric acid, and lactic acid were measured using
of α-D -galactose, βD -galactose, α- D -lactose, β- D -lactose, citric acid, and lactic acid were measured
the quantitative 1H NMR
spectrum. During milk fermentation, lactic acid bacteria convert part of αusing
the quantitative 1 H NMR spectrum. During milk fermentation, lactic acid bacteria convert
D-lactose and β-D-lactose into α-D-galactose and β-D-galactose, which are finally broken down to
become lactic acids, causing a pH decrease that is responsible for casein coagulation [23]. These
results showed that, even if the signals in the 1H NMR spectrum overlapped with the signals of the
Metabolites 2016, 6, 19
8 of 11
part of α-D -lactose and β-D-lactose into α-D -galactose and β-D -galactose, which are finally broken
down to become lactic acids, causing a pH decrease that is responsible for casein coagulation [23].
Metabolites 2016, 6, 19
8 of 11
These results showed that, even if the signals in the 1 H NMR spectrum overlapped with the signals of
theacyl
acylchains
chainsofoffatty
fattyacids
acidsfrom
frommilk
milkfats,
fats,this
thisquantitative
quantitativemethod
methodusing
usingquantitative
quantitative
difference
difference
spectra
can
eliminate
interference
due
to
overlapping
spectra
to
obtain
the
correct
concentrations
spectra can eliminate interference due to overlapping spectra to obtain the correct concentrations of of
yogurt
components.
yogurt
components.
1 H1 NMR spectra of the ready yogurt measured at 40 ˝ C in different lots were shown in
The
The H NMR spectra of the ready yogurt measured at 40 °C in different lots were shown in
Figure
5. 5.The
types of
ofdetected
detectedcomponents
components
were
quite
similar
among
the yogurt
Figure
Thespectra
spectraand
and the
the types
were
quite
similar
among
the yogurt
in
and
thethe
self-made
oneone
from
inoculum
contained
two types
of
in different
different lots.
lots.Both
Boththe
theready
readyyogurt
yogurt
and
self-made
from
inoculum
contained
two types
1 H NMR
subsp.
bulgaricus
andand
S. thermophiles.
The 1The
H NMR
spectra
and the
of bacterial
bacterialstrains:
strains:L.L.delbrueckii
delbrueckii
subsp.
bulgaricus
S. thermophiles.
spectra
and
NMR-detected
components
were
also
similar
between
the
ready
yogurt
and
the
self-made
one
the NMR-detected components were also similar between the ready yogurt and the self-made one
(Figures
and
(Figures
4B4B
and
5).5).
Figure 5. The 1H NMR spectra of the ready yogurt measured at 40 °C in different lots.
Figure 5. The 1 H NMR spectra of the ready yogurt measured at 40 ˝ C in different lots.
3. Experimental Section
3. Experimental Section
3.1. Materials and Sample Preparation
3.1. Materials and Sample Preparation
Bulgarian yogurt and its inoculum were purchased at a local market. Bulgarian yogurt contained
Bulgarian yogurt and its inoculum were purchased at a local market. Bulgarian yogurt contained
two types of bacterial strains: L. delbrueckii subsp. bulgaricus and S. thermophilus. For signal
two types of bacterial strains: L. delbrueckii subsp. bulgaricus and S. thermophilus. For signal assignments,
assignments, D2O was added to yogurt at a final concentration of 10% (v/v). Bulgarian yogurt samples
D2 O was added to yogurt at a final concentration of 10% (v/v). Bulgarian yogurt samples (0.6 mL)
(0.6 mL) were then placed in a 5-mm NMR tube (Kanto Chemical Co., Inc., Tokyo, Japan). For
were
then placed
in a 5-mm NMR
tube
(Kanto
Chemical
Tokyo,
Japan). from
For quantitative
quantitative
measurements,
triplicate
yogurt
samples
were Co.,
madeInc.,
for this
experiment
inoculum.
measurements,
triplicate
yogurt
were
made milk
for this
experiment
Bulgarian yogurt
inoculum
(1 g)samples
was added
to whole
(1 L).
The samplefrom
was inoculum.
immediatelyBulgarian
mixed
yogurt
(1 concentration
g) was added of
to 10%
whole
milk
(1 was
L). The
mixed
with
D2 O at
with inoculum
D2O at final
(v/v)
and
thensample
placedwas
in a immediately
5-mm NMR tube.
The
volume
final
10% (v/v)
and
was
then placedwas
in aperformed
5-mm NMR
tube.
The24volume
the sample
of concentration
the sample wasofadjusted
to 0.6
mL.
Fermentation
at 40
°C for
h after of
putting
the
was
adjusted
mL.equipment.
Fermentation was performed at 40 ˝ C for 24 h after putting the sample into
sample
intoto
the0.6
NMR
the NMR equipment.
3.2. NMR Spectroscopy
3.2. NMR Spectroscopy
NMR experiments were performed at 4 °C on a Unity INOVA-500 spectrometer (Agilent
31P 1D
NMR experiments
were
performed
at 4 1˝H,
C 13
on
Unity
INOVA-500
spectrometer
(Agilent
Technologies,
Santa Clara,
CA,
USA) to obtain
C, aand
NMR spectra
and 1H-13C HSQC,
13 C, and 31 P 1D NMR spectra and 1 H-13 C HSQC,
1H-1H DQF–COSY,
1H-31PCA,
1H-131C
Technologies,
Santa Clara,
USA) toand
obtain
H,CT-HMBC
FG-HMBC,
spectra.
1 H-1 H DQF–COSY,
31 P FG-HMBC,
13 C CT-HMBC
The 1H NMR1 Hspectra
of yogurt and
were1 Hmeasured
at 499.87
MHz, and the water signal was
spectra.
suppressed by the pre-saturation method. 4,4-dimethyl-4-silapentane-1-sulfonic acid (Wako Pure
Metabolites 2016, 6, 19
9 of 11
The 1 H NMR spectra of yogurt were measured at 499.87 MHz, and the water signal was
suppressed by the pre-saturation method. 4,4-dimethyl-4-silapentane-1-sulfonic acid (Wako Pure
Chemical Industries, Ltd., Osaka, Japan) was used as an internal reference, and its chemical shift was
set to 0 ppm. For assignments, the acquisition parameters were as follows: number of data points,
30,272; spectral width, 8000 Hz; acquisition time, 1.892 s; delay time, 2.000 s; and number of scans, 8.
For quantification, NMR experiments were performed at 40 ˝ C, and the acquisition parameters were
as follows: number of data points, 32,768; spectral width, 8000 Hz; acquisition time, 2.000 s; delay
time, 15.000 s; and number of scans, 32. The free induction decays (FIDs) were zero-filled to 65,536
data points and multiplied by an exponential window function with a 0.25-Hz line-broadening for all
1 H NMR spectra by NMR software.
The 13 C NMR spectra were measured at 125.71 MHz. Dioxane was used as an external reference,
and its chemical shift was set to 67.5 ppm. The parameters of the 13 C NMR spectrum were as follows:
number of data points, 65,536; spectral width, 31,422 Hz; acquisition time, 1.043 s; delay time, 2.000 s;
and number of scans, 83,392.
The 31 P NMR spectra were measured at 202.35 MHz. Potassium phosphate was used as an
external reference, and its 31 P chemical shift was set to 0 ppm. The 31 P NMR spectra were measured
with the following parameters: number of data points, 30,272; spectral width, 11,999 Hz; acquisition
time, 1.043 s; delay time, 2.000 s; and number of scans, 1024. The FIDs were zero-filled to be 65,536
data points and multiplied by an exponential window function with a 0.25-Hz line-broadening for 31 P
NMR spectra by NMR software.
The 1 H-1 H DQF-COSY spectra were obtained by suppressing the water signal with the
pre-saturation method, and the acquisition parameters were as follows: number of data points,
2048 (F2) and 512 (F1); spectral width, 5911 Hz (F1 and F2); acquisition time, 0.202 s; delay time, 2.000 s;
and number of scans, 48.
The 1 H-13 C HSQC spectra of yogurt were generated in the phase-sensitive mode with the
following acquisition parameters: number of data points, 512 for 1 H and 256 for 13 C; spectral widths,
5498 Hz for 1 H and 20,110 Hz for 13 C; acquisition time, 0.186 s; delay time, 2.000 s; and number of
scans, 80.
The 1 H-13 C CT-HMBC spectra were measured in the absolute mode with the following parameters:
number of data points, 4096 for 1 H and 512 for 13 C; spectral widths, 5498 Hz for 1 H and 27,643 Hz for
13 C; acquisition time, 0.402 s; delay time, 3.000 s; and number of scans, 80.
The 1 H-31 P CT-HMBC spectra of yogurt were obtained in the absolute mode. Potassium phosphate
was used as an external reference, and its 31 P chemical shift was set to 0 ppm. The acquisition parameters
were as follows: number of data points, 2048 for 1 H and 128 for 31 P; spectral widths, 5004 Hz for 1 H and
4047 Hz for 31 P; acquisition time, 0.402 s; delay time, 1.000 s; and number of scans, 200.
3.3. NMR Signal Assignments
The signals in the NMR spectra of yogurt were assigned in reference to the databases, Spectral
Database for Organic Compounds (SDBS, http://sdbs.db.aist.go.jp/sdbs) and Biological Magnetic
Resonance Data Bank (BMRB Metabolomics, http://www.bmrb.wisc.edu/metabolomics). Because the
chemical shifts of several signals were significantly different than those in the published data, their
correlations were confirmed in the 2D NMR spectra. Finally, authentic standard compounds were
added to yogurt for further confirmation of assignments.
3.4. Quantification of Yogurt Components
Because several signals overlapped with signals of the acyl chains of fatty acids from milk
fats, quantitative analyses of yogurt were performed using the difference spectrum between the
quantitative 1 H NMR spectrum and the DOSY spectrum. The spin-lattice relaxation times (T1 ) for
the quantitative 1 H NMR spectrum were measured by the partial relaxation Fourier-transform (FT)
method [4]. The delay time (d1 ) was determined with aq being the acquisition time.
Metabolites 2016, 6, 19
10 of 11
d1 ě 5 ˆ T1 ´ aq .
A capillary containing 20% (v/v) 1,1,2,2-tetrachloroethane and 80% (v/v) chloroform-d (CDCl3 )
was inserted into the NMR sample tubes as the concentration standard. It has been reported that
T1 of 1,1,2,2-tetrachloroethane was decreased from 2.0 to 0.4 s by adding the relaxation reagent
Cr(AcAc)3 at a concentration of 1 mg/mL [4]. Therefore, 1 mg/mL Cr(AcAc)3 was added to a 20%
(v/v) 1,1,2,2-tetrachloroethane solution. The integral value of 1,1,2,2-tetrachloroethane was measured
and compared with those of the compounds in yogurt to determine their concentrations.
3.5. Diffusion Ordered Spectroscopy (DOSY)
DOSY separates individual components in mixtures based on their diffusion. Molecules in
solution are in constant motion and are accompanied by both rotational and translational motions.
Diffusion experiments were performed in the z-direction using the DgcsteSL_dpfgse sequence [24].
The acquisition parameters were as follows: number of data points, 32,768; spectral width, 8000 Hz;
acquisition time, 2.048 s; delay time, 15.000 s; number of scans, 16; diffusion delay, 0.400 s; total
diffusion-encoding gradient pulse duration, 0.002 s; gradient stabilization delay, 0.0003 s. The signals
that overlapped with those of milk fats were quantified using difference spectra that were obtained by
subtracting the DOSY spectra from the quantitative 1 H NMR spectra.
4. Conclusions
In conclusion, 1 H, 13 C, and 31 P NMR spectra, as well as 1 H-13 C HSQC, 1 H-1 H DQF-COSY,
1 H-31 P CT-HMBC, and 1 H-13 C CT-HMBC spectra, of yogurt were successfully obtained without any
separation or pretreatment. In addition, the quantification of yogurt components was conducted using
the 1 H NMR spectra and the difference spectra between the quantitative 1 H NMR spectra and the
DOSY spectra. Therefore, this study of yogurt using NMR spectroscopy provides a promising method
to monitor the various components produced during yogurt fermentation and to classify yogurt types.
Moreover, the assignment data and quantitative method can be utilized for quality control and other
applications. The proposed method could be useful to analyze the differences between various kinds
of yogurt.
Acknowledgments: We thank K. Furihata and F. Wei for assisting with the NMR measurements. This work was
supported by a Grant-in-Aid for Scientific Research (S) from the Japan Society for the Promotion of Science (JSPS)
(Grant No. 23228003).
Conflicts of Interest: The authors declare no conflict of interest.
References
1.
2.
3.
4.
5.
Gowda, G.N.; Raftery, D. Can NMR solve some significant challenges in metabolomics? J. Magn. Reson. 2015,
260, 144–160. [CrossRef] [PubMed]
Monakhova, Y.B.; Ruge, W.; Kuballa, T.; Ilse, M.; Winkelmann, O.; Diehl, B.; Lachenmeier, D.W. Rapid
approach to identify the presence of Arabica and Robusta species in coffee using 1 H NMR spectroscopy.
Food Chem. 2015, 182, 178–184. [CrossRef] [PubMed]
Hohmann, M.; Koospal, V.; Bauer-Christoph, C.; Christoph, N.; Wachter, H.; Diehl, B.; Holzgrabe, U.
Quantitative 1 H NMR analysis of egg yolk, alcohol, and total sugar content in egg liqueurs. J. Agric.
Food Chem. 2015, 63, 4112–4119. [CrossRef] [PubMed]
Hu, F.; Furihata, K.; Kato, Y.; Tanokura, M. Nondestructive quantification of organic compounds in whole
milk without pretreatment by two-dimensional NMR spectroscopy. J. Agric. Food Chem. 2007, 55, 4307.
[CrossRef] [PubMed]
Li, Y.; Teng, Z.; Parkin, K.L.; Wang, Q.; Zhang, Q.; Luo, W.; Zhao, M. Identification of bioactive metabolites
dihydrocanadensolide, kojic acid, and vanillic acid in soy sauce using GC-MS, NMR spectroscopy,
and single-crystal X-ray diffraction. J. Agric. Food Chem. 2014, 62, 8392–8401. [CrossRef] [PubMed]
Metabolites 2016, 6, 19
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
11 of 11
Wei, F.; Furihata, K.; Miyakawa, T.; Tanokura, M. A pilot study of NMR-based sensory prediction of roasted
coffee bean extracts. Food Chem. 2014, 152, 363–369. [CrossRef] [PubMed]
Vázquez-Fresno, R.; Llorach, R.; Urpi-Sarda, M.; Khymenets, O.; Bulló, M.; Corella, D.; Andres-Lacueva, C.
An NMR metabolomics approach reveals a combined-biomarkers model in a wine interventional trial with
validation in free-living individuals of the PREDIMED study. Metabolomics 2015, 11, 797–806. [CrossRef]
Koda, M.; Furihata, K.; Wei, F.; Miyakawa, T.; Tanokura, M. Metabolic discrimination of mango juice from
various cultivars by band-selective NMR spectroscopy. J. Agric. Food Chem. 2012, 60, 1158–1166. [CrossRef]
[PubMed]
Yuan, Y.; Song, Y.; Jing, W.; Wang, Y.; Yang, X.; Liu, D. Simultaneous determination of caffeine, gallic acid,
theanine, (–)-epigallocatechin and (–)-epigallocatechin-3-gallate in green tea using quantitative 1 H-NMR
spectroscopy. Anal. Methods 2014, 6, 907–914. [CrossRef]
Hu, F.; Furihata, K.; Ito-Ishida, M.; Kaminogawa, S.; Tanokura, M. Nondestructive observation of bovine
milk by NMR spectroscopy: Analysis of existing states of compounds and detection of new compounds.
J. Agric. Food Chem. 2004, 52, 4969–4974. [CrossRef] [PubMed]
Wei, F.; Furihata, K.; Hu, F.; Miyakawa, T.; Tanokura, M. Complex mixture analysis of organic compounds in
green coffee bean extract by two-dimensional NMR spectroscopy. Magn. Reson. Chem. 2010, 48, 857–865.
[CrossRef] [PubMed]
Da Costa, M.P.; da Silva Frasao, B.; da Costa Lima, B.R.C.; Rodrigues, B.L.; Junior, C.A.C. Simultaneous
analysis of carbohydrates and organic acids by HPLC-DAD-RI for monitoring goat’s milk yogurts
fermentation. Talanta 2016, 152, 162–170. [CrossRef] [PubMed]
Weerathilake, W.A.; Rasika, D.M.; Ruwanmali, J.K.; Munasinghe, M.A. The evolution, processing, varieties
and health benefits of yogurt. Int. J. Sci. Res. Publ. 2014, 4, 1–10.
Siefarth, C.; Tran, T.B.T.; Mittermaier, P.; Pfeiffer, T.; Buettner, A. Effect of radio frequency heating on yoghurt,
I: Technological applicability, shelf-life and sensorial quality. Foods 2014, 3, 318–335. [CrossRef]
Lobato-Calleros, C.; Ramírez-Santiago, C.; Vernon-Carter, E.J.; Álvarez-Ramírez, J. Impact of native and
chemically modified starches addition as fat replacers in the viscoelasticity of reduceD-fat stirred yogurt.
J. Food Eng. 2014, 131, 110–115. [CrossRef]
Pan, X.D.; Wu, P.G.; Yang, D.J.; Wang, L.Y.; Shen, X.H.; Zhu, C.Y. Simultaneous determination of melamine
and cyanuric acid in dairy products by mixeD-mode solid phase extraction and GC–MS. Food Control. 2013,
30, 545–548. [CrossRef]
Kunda, P.B.; Benavente, F.; Catalá-Clariana, S.; Giménez, E.; Barbosa, J.; Sanz-Nebot, V. Identification of
bioactive peptides in a functional yogurt by micro liquid chromatography time-of-flight mass spectrometry
assisted by retention time prediction. J. Chromatogr. A 2012, 1229, 121–128. [CrossRef] [PubMed]
Coscia, A.; Orrù, S.; Di Nicola, P.; Giuliani, F.; Rovelli, I.; Peila, C.; Bertino, E. Cow’s milk proteins in human
milk. J. Biol. Regul. Homeost. Agents 2012, 26, 39–42. [PubMed]
Ahrens, B.; Lopes de Oliveira, L.C.; Grabenhenrich, L.; Schulz, G.; Niggemann, B.; Wahn, U.; Beyer, K.
Individual cow’s milk allergens as prognostic markers for tolerance development? Clin. Exp. Allergy 2012,
42, 1630–1637. [CrossRef] [PubMed]
Mekmene, O.; Gaucheron, F. Determination of calcium-binding constants of caseins, phosphoserine, citrate
and pyrophosphate: A modelling approach using free calcium measurement. Food Chem. 2011, 127, 676–682.
[CrossRef] [PubMed]
Milewski, S.; Gabriel, I.; Olchowy, J. Enzymes of UDP-GlcNAc biosynthesis in yeast. Yeast 2006, 23, 1–14.
[CrossRef] [PubMed]
Ardawatia, H.; Liberles, D.A. A systematic analysis of lineage-specific evolution in metabolic pathways.
Gene 2007, 387, 67–74. [CrossRef] [PubMed]
Shiby, V.K.; Mishra, H.N. Fermented milks and milk products as functional foods—A review. Crit. Rev. Food
Sci. Nutr. 2013, 53, 482–496. [CrossRef] [PubMed]
Johnson, C.S. Diffusion ordered nuclear magnetic resonance spectroscopy: Principles and applications.
Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 203–256. [CrossRef]
© 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC-BY) license (http://creativecommons.org/licenses/by/4.0/).