Institute of Food Research
Orange g juice j authentication by 1H NMR G n Le Gwen L Gall, G ll Marion M i n Cuny C n and nd Ian Colquhoun
Need for authentication • Orange Juice consumption is increasing world wide and freshly squeezed juices (Not from Concentrate,, NFC)) is a growing g g market in US, Canada and Europe Million SSE M E Gallons
1800 1600
U S consumption of orange juice by category U.S.
1400
From Conc OJ (FCOJ)
1200
Reconst Conc OJ ((RCOJ)) (chilled)
1000
Not From Conc (NFC) (chilled) TOTAL
800 600 400 200 0
89
90
91
92
93
94
95
96
97
98
99
00
YEAR
http://edis.ifas.ufl.edu/FE195#TABLE_4
Need for authenticating orange juice • Some techniques exist to detect fraud (HPLC SNIF (HPLC, SNIF-NMR) NMR) but still need for more Aim: would it possible to easily implement a technique such as 1H Nuclear Magnetic Resonance (NMR) spectroscopy to address authentication issues?
Techniques for measuring wide ranges off metabolites t b lit 1H H-NMR NMR
major compounds (fingerprinting) (f g p g)
GC/MS GC/ S
Ma n y primary Mainly pr mary metabolites, polar & nonpolar: amino-acids amino acids, organic acids, sugars, fatty acids acids…
METABOLITES
LC/MS
Secondary S d metabolites, aromatic cpds, phenolics, glycoalkaloids
Preparation • pH adjust 10 mL of juice to 4 00±0 03 add D2O to supernatant 4.00±0.03, (50-100 samples a day) • Proton NMR: 10 min recording per sample Le Gall G, Puaud M, Colquhoun IJ, Discrimination between orange juice and pulp wash by H-1 nuclear magnetic resonance spectroscopy: Identification of marker compounds JOURNAL OF AGRICULTURAL AND FOOD CHEMISTRY 49 (2): 580-588 FEB 2001
1H NMR of fruit juice
citric dmp asn ? asp
suc + gluc
OVERALL
malic
SPECTRA
10
8
suc gluc gluc luc
6
4
2
0
ppm 3
ppm
2.5
ethanol
malic gaba dmp pro dmp pro dmp gaba
arg
ala
?
fatty val acids sterol Mean Orange Juice
Mean Grapefruit 2.6
2.4
2.2
2
1.8
1.6 ppm
1.4
1.2
1
0.8
Mean spectra Low field region formic acid niacin niacin
? ?
? ?
tyr * tyr * ?* hesphesp h ? ?
naringin g
* polyphenols
? *
h hesp hl phlorin
Mean OJ
naringin
M n GF Mean 9
8.5
8
7.5 ppm
7
6.5
6
Study on orange and grapefruit juice • 93 juices bought in local supermarkets, Norwich ,UK •
58 – – –
orange g juices j 24 freshly squeezed 12 fresh concentrates 22 p pasteurised concentrates
•
10 mixtures orange/grapefruit juice – 3 freshly squeezed – 7 pasteurised concentrates
•
23 – – – •
grapefruit juices 9 freshly squeezed 4 fresh concentrates 10 p pasteurised concentrates
2 freshly squeezed Clementine juices
• Can we differentiate orange g juice j from g grapefruit p by 1H NMR? Markers? • Orange juice from mixtures? • Freshly squeezed from concentrate? Markers?
Approaches for data analysis
sample l 15_4 30_4 13_4
t tyr 84089 153982 140869
multivariate, integrated sample l 15_4 30_4 13_4
t tyr
phe h
c. putr t
84089
20591
4937
153982
46018
11968
140869
21517
12189
multivariate, whole trace measurements s sample
univariate, integrated
Data reduction by PCA, whole traces traces measurements
PCs
sam mples
samp ples
z11 z12 z13
z21 z22 z31 scores
zn1 zn2
Principal Component Analysis on 1H NMR spectra of fruit juices -3
Normalised covariance
x 10
3
3
2
2
1
1 PC C2
PC C2 10% var
x 10
0
-1
-2
-2
-3
-3
grapefruit orange
0 0.005 0.01 PC1 79% var
mixtures clementine
Normalised covariance
0
-1
-0.01 -0.005
-3
-0.01 -0.005
0 0.005 0.01 PC1
fresh squeezed (NFC) fresh concentrate (RCOJ) pasteurised concentrate (FCOJ)
Origin of the discrimination -
ethanol P PC2
grapefruit orange
0
polyphenols ?
0 PC1
naringin
2.5
9 8.5 8 7.5 7 6.5 6 ppm
1.5
suc
unk 4.4 4.3 ppm
2 ppm
malic li
4.5
?
malic pro
loading 1
citric
glucose 5.5
5
4.5
4 ppm
3.5
3
1
Approaches for data analysis
sample l 15_4 30_4 13_4
t tyr 84089 153982 140869
multivariate, integrated sample l 15_4 30_4 13_4
t tyr
phe h
c. putr t
84089
20591
4937
153982
46018
11968
140869
21517
12189
multivariate, whole trace measurements s sample
univariate, integrated
Any markers? • Anova was performed on 35 signals representing most of the metabolites detected by 1H NMR – F value was significant for 32 signals – 25 showed h d hi high h F values l = potential t ti l markers discriminating OJ/GF
ANOVA AND INTENSITY PLOTS ANOVA on niacin 8.81ppm x 10
-7
Intensity plot on niacin -6 6
1.4
x 10
mixtures
12 10 8
90% 50%OJ
1.2
NMR In ntensitty
NMR In N ntensitty
F 52.29
1
0.8
6
0.6
4
0.4 0.2
2
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
0
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
• Even if niacin is 4 fold-diff fold diff to the mean, mean mixtures do not separate from pure OJ
Principal Component Analysis on 1H NMR spectra of fruit juices -3
Normalised covariance
x 10
3
3
2
2
1
1 PC2
PC2 2 10% var
x 10
0 -1
0 -1
-2 2
-2
-3
-3 -0.01 -0.005
grapefruit orange
0 0.005 0.01 PC1 79% var
-3
zoom
%OJ
70% 60% 50% 75% 90% 85% 72% ?% 51% 70% -4 4
-2 2
0 2 PC1
4
6 x 10
-3
mixtures % of grapefruit in mixture clementine needs to be high (>40%) for mixture to be discriminated
PCA on parts of spectra PCA on low f field region g x 10
0.01
-3
zoom
6
0.005 4
PC2 P
PC2 P
0 -0.005
85%OJ 70%OJ ?%
0
-0.01 -0 0.015 015 -0.02 -0.03
2
-2 -0.02
grapefruit rapefruit orange
-0.01 PC1
0
0.01
-10
-5
0 PC1
mixtures clementine
5 x 10
-3
PCA on high g field region g not as good (not shown)
Authenticity by liquid chromatography HESPERIDIN NARINGIN
700
Pea ak Intensityy
600 500 400 300 200 100 0
Pure Orange Juice
Pure Grapefruit Mixture (60% 60% OJ)
Detection of hesperidin (OJ marker) and naringin (GF marker) L field Low fi ld region i h hesp hesp naringin
h hesp
Mean OJ
naringin
7.08
Mean GF 8
7.5 ppm
7
6.5
7.06 7.04 ppm
7.02
6
HPLC markers seen by y NMR but signals affected by pH
6.19 6.18 6.17 6.16 6.15 ppm
Peak alignment done by Matlab macro written in house
Hesperidin (OJ) before
7.08
7.06 7.04 ppm
after
Matlab macro
7.02
N i i (GF) Naringin
7.08
7.06
7.04
Matlab ‘collation’
7.02
7.08
ppm
7.06
7.04
7.02
ppm
after
before Matlab macro
6.19 6.18 6.17 6.16 6.15
6.2
6.18
6.16 ppm
6.14
ANOVA HESPERIDIN 7.04
ppm pp
x 10
7.02 -6 6
ANOVA
5
6
F 61.27
NMR In ntensitty
7.06
NMR I N Intensiity
7.08
4
3
2
1
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
x 10
-6 6
Intensity plot
mixtures 5 4 3 2 1 0
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
ANOVA NARINGIN
x 10
-7
15 10 5 0
6.18 6.16 ppm x 10
6.14 -6 6
ANOVA
7
NMR In N ntensitty
6.2
6
F642.07
7
x 10
-6
I t Intensity it plot l t
6 5
5
4
4
3
3
2 2 1 1 0 0
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
-1
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
Discrimination according to process x 10
-3
Normalised covariance
ethanol
3 2
PC C2
1 0 -1
loading g 1
-2 -3
2.5 -0.01 -0.005
grapefruit orange
0 0.005 0.01 PC1
mixtures clementine
2
1.5
1
ppm
fresh squeezed q fresh concentrate Pasteurised concentrate
ANOVA according to process • At least two signals seem to be potential
markers for (not from concentrate) freshly squeezed juices x 10
8
F29.63
6
4
2
-4
2.5
ethanol
NM MR Intensity y
NM MR Inttensity y
x 10 -4
? 3.33ppm pp F22.94 prob-14
2
15 1.5
1
0.5
0
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
ANOVA according to process • Half a dozen signals g seem to show a trend in difference between NFC, RCOJ and FCOJ especially p y for grapefruit g p x 10
1.8
x 10
b-glucose 4.614ppm F81.71
1.6 1.4
-3
sucrose 5.25 ppm
2.5
NM MR Inte ensity
NM MR Inttensity y
2
-3
F65.3 2
1.5
1.2 1
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
1
NFC RCOJ FCOJ NFC RCOJ FCOJ
OJ
GF
•
1H
CONCLUSIONS
NMR spectra easy to obtain (same results for 400 MHz, not shown)
• Orange juice discriminated from grapefruit juice • Some new markers but mixtures with
