Study on decomposition products of norbixin during bleaching with hydrogen peroxide and a peroxidase by means of UPLC-UV and mass spectrometry

Article abstract of DOI:10.1016/j.foodchem.2011.11.118

The decomposition products of norbixin, a component of the natural colouring agent annatto, have been studied under bleaching conditions in water and in a whey matrix. In water, several unsaturated aldehydes and ketones of carboxylic acids were identified with UPLC-UV/MS and high resolution mass spectrometry techniques. Based on these products a reaction scheme for the decomposition of norbixin is proposed. In whey, the norbixin is also degraded during bleaching, but no decomposition products are detected. Most likely these products react with endogenous compounds from the whey matrix. For one of these compounds, i.e. cysteine, the formation of a reaction product with 3-acetylacrylic acid (decomposition product of norbixin) was shown.

Full text of DOI:10.1016/j.foodchem.2011.11.118

Food Chemistry 132 (2012) 1354–1359
Contents lists available at SciVerse ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Study on decomposition products of norbixin during bleaching with hydrogen
peroxide and a peroxidase by means of UPLC-UV and mass spectrometry
⇑
Wibo B. van Scheppingen , Ilco A.L.A. Boogers, Alexander L.L. Duchateau
DSM Biotechnology Center, Analysis Dept., P.O. Box 1, 2600 MA, Delft, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 13 May 2011
Received in revised form 13 October 2011
Accepted 28 November 2011
Available online 6 December 2011
The decomposition products of norbixin, a component of the natural colouring agent annatto, have been
studied under bleaching conditions in water and in a whey matrix. In water, several unsaturated alde-
hydes and ketones of carboxylic acids were identified with UPLC-UV/MS and high resolution mass spec-
trometry techniques. Based on these products a reaction scheme for the decomposition of norbixin is
proposed. In whey, the norbixin is also degraded during bleaching, but no decomposition products are
detected. Most likely these products react with endogenous compounds from the whey matrix. For
one of these compounds, i.e. cysteine, the formation of a reaction product with 3-acetylacrylic acid
(decomposition product of norbixin) was shown.
Keywords:
Norbixin
Annatto
Bleaching
Whey
Ó 2011 Elsevier Ltd. All rights reserved.
Carotenoids
Peroxidase
Hydrogen peroxide
1. Introduction
taining (several isomers of) bixin or norbixin in different concen-
trations and as potassium and sodium salts (Joint FAO/WHO,
Annatto (E160b) is an orange/red natural carotenoid colouring
agent obtained from the seeds of the tropical plant Bixa orellana
L. The major pigment in the seed coat is 90-cis-bixin (>80%) with
smaller amounts of other carotenoids (Preston & Rickard, 1980).
It has been suggested that biosynthesis of bixin, with a C25 skele-
ton, takes place by the oxidation of a normal C40 carotenoid like
lycopene (Bouvier, Dogbo, & Bilal, 2003). The presence of small
amounts of b-carotene, cryptoxanthin, lutein, zeaxanthin, and sev-
eral lycopenoate analogues has been reported (Mercadante, Steck,
& Pfander, 1997; Tirimanna, 1981). Several processes are used to
extract the pigment from dried annatto seeds: extraction in oil,
in aqueous alkali or with solvents. The composition of an annatto
solution depends on the extraction method and parameters such
as: temperature, heating period, origin of the seed, and shelf-life
of the products. During the extraction processes isomerisation of
90-cis-bixin can take place. By heating a suspension of the seeds
in oil, the more stable and better soluble all-trans-isomer is
formed. In aqueous alkali the methyl ester of bixin is hydrolysed
to give (cis- and trans-) norbixin isomers (Preston & Rickard,
1980). Different commercial annatto colourants are available con-
2007). Bixin is a fat soluble compound while norbixin dissolves
better in an aqueous environment (for structures see Fig. 1).
Annatto is used as a colouring agent in a wide range of food-
stuffs such as butter, margarine, cheese, fats, cereals, baked goods,
snacks, ice-cream, salad dressings, sugar confectionary, yoghurts,
drinks, meat, and fish (Mercadante et al., 1997; Scotter, 2009). In
most foods the annatto extract used contains norbixin, with the
exception of margarine, which mainly contains bixin (Tennant &
O’Callaghan, 2005). In the European Union the use of colourants
in foodstuffs is regulated by the European Commission and for an-
natto (E160b) the permitted maximum level in several food prod-
ucts ranges from 10 to 50 mg/kg (for red Leicester cheese)
(European Commission, 1994). The JECFA committee established
an ADI (Acceptable Daily Intake) for norbixin and its sodium and
potassium salts of 0.6 mg/kg of body weight (expressed as nor-
bixin) (Joint FAO/WHO, 2007). In several publications the esti-
mated intake of annatto by food consumption is lower than the
ADI (Joint FAO/WHO, 2007; Scotter, 2009; Tennant & O’Callaghan,
2005).
The thermal degradation of annatto is studied in detail by Scot-
ter and co-workers (Scotter, 1995, 2009; Scotter, Castle, & Apple-
ton, 2001). Depending on (thermal) processes during production
the annatto will be contaminated with aromatic hydrocarbons as
a result of the decomposition of both bixin and norbixin (Scotter,
2009).
⇑
Corresponding author. Tel.: +31 15 2792718.
E-mail addresses: wibo.scheppingen-van@dsm.com (W.B. van Scheppingen),
Lucien.Duchateau@dsm.com (A.L.L. Duchateau).
0308-8146/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2011.11.118

W.B. van Scheppingen et al. / Food Chemistry 132 (2012) 1354–1359
1355
trans-bixin
H3C
H3C
O
HO
O
O
CH3
CH3
CH3
trans-norbixin
H3C
H3C
O
HO
OH
O
CH3
CH3
Fig. 1. Chemical structures of all-trans-isomers of bixin and norbixin.
About 15% of world cheese production is coloured and often an-
natto is used (El-Samragy, 2004). In cheese, annatto combines with
casein and gives it a yellow colour. In some cases annatto gives a
pink colour in the cheese. A part of the annatto added to the cheese
milk will be present in the resulting whey. Approximately 15–20%
of the bixin is recovered in the whey (El-Samragy, 2004; Smith,
2008). Not much is known about the partition of norbixin between
cheese and whey (Smith, 2008). Besides the added annatto, natural
colour compounds can be present in the whey, such as b-carotene
from the natural feed of cows.
Before whey can be used in other products the colour has to be
removed. Light is reported as the most destructive agent, followed
by benzoyl peroxide, while air is much less effective (Najar, Bobbio,
& Bobbio, 1988). For bleaching whey, two compounds are legally
permitted: benzoyl peroxide and hydrogen peroxide. Typical con-
ditions for the use of benzoyl peroxide are a concentration of less
than 20 ppm (<0.002%) and heating at 60 °C for 15 min at pH 6–
7. During this process benzoyl peroxide reacts almost quantita-
tively to benzoic acid (El-Samragy, 2004). Saiz et al. postulated that
after a free radical oxidative attack a part of the peroxyl radical is
trapped in the polyene system and forms a resonance stabilized
peroxyl radical or stable adduct (Saiz, Manrique, & Fritz, 2001).
The use of benzoyl peroxide is a subject of international concern
because a small percentage of the population (asthmatics) are sen-
sitive to benzoic acid related compounds (Smith, 2008).
Hydrogen peroxide may be used in concentrations up to 0.05%
and residual hydrogen peroxide must be removed by addition of
a maximum concentration of 20 ppm catalase. The use of hydrogen
peroxide can alter the functionality of whey proteins and is known
to inhibit browning in milk systems. In a recent study on the effect
of bleaching agents on the flavour of whey, higher concentrations
of oxidation reaction products were found in products bleached
with hydrogen peroxide compared with unbleached or benzoyl
peroxide bleached whey. In the bleached whey off-flavours were
identified by sensory and instrumental analysis and compounds
like heptanal, hexanal, octanal, and nonanal were detected (Crois-
sant, Kang, Campbell, Bastian, & Drake, 2009).
2008; Zelena, Hardebusch, Hülsdau, Berger, & Zorn, 2009; Zorn,
Langhoff, Scheibner, & Berger, 2003; Zorn, Langhoff, Scheibner,
Nimtz, & Berger, 2003). This enzyme is now commercially available
for the bleaching of annatto in whey under the brand name Maxi-
Bright™. This enzymatic bleaching only needs small amounts of
hydrogen peroxide and works efficiently at room temperature.
Moreover, no catalase is needed to remove residues of hydrogen
peroxide.
In a recent review it is concluded that there is a need for scien-
tific evaluation of bleaching and its effect on whey (Kang et al.,
2010). Despite the worldwide bleaching of norbixin with hydrogen
peroxide there is a lack of knowledge about the mechanisms, the
reaction products and the effects of these products on the whey
matrix. In our study, we investigated the fate of norbixin in aque-
ous solution and in whey during chemical and enzymatic treat-
ment. For this, LC-UV/MS and high resolution mass spectrometry
were used as instrumental analytical techniques.
2. Materials and methods
2.1. Chemicals
Hydrogen peroxide 30% Merck, (Darmstadt, Germany), diluted
with Milli-Q water to 1% just before usage. Annatto (E160b) con-
taining 1.45% norbixin, CeskaÒ batch 02/06/2008 (CSK Food Enrich-
ment, Leeuwarden, The Netherlands). MaxiBrightTM (MsP1), batch
ODL GRZ 0927 (DSM Food Specialties, Delft, The Netherlands). 3-
acetylacrylic acid (4-oxopent-2-enoic acid), Alfa Aeser (Karlsruhe,
Germany).
L
-Cysteine, Sigma (Steinheim, Germany).
L-Lysine
monohydrochloride, Sigma (Steinheim, Germany). Acetonitrile,
LiChrosolvÒ, Merck (Darmstadt, Germany). Phosphate buffer
50 mM pH 6.6, 34.6 mM potassium dihydrogen phos-
phate + 16.1 mM di-potassium hydrogen phosphate, both from
Merck (Darmstadt, Germany). Whey from Cheddar cheese pre-
pared with a DelvoTEC MT53A starter culture and MaxirenÒ as
the rennet (DSM Food Specialties, Delft, The Netherlands).
A patent has been issued on the biochemical process of bleach-
ing dairy products such as whey with the enzyme lipoxygenase
(Roos, Dijk, & Folkertsma, 2005). Lactoperoxidase, a natural com-
ponent of milk, is also capable of bleaching carotenoids but only
at low concentrations of hydrogen peroxide (Kang, Campbell, Bas-
tian, & Drake, 2010). Recently, Zorn and co-workers reported about
bleaching by several fungal peroxidases. The peroxidase isolated
from a mushroom (Marasmius scorodonius) is capable of degrading
several carotenes and xanthophylls like b-carotene, lutein, zeaxan-
thin, violaxanthin, neoxanthin, and lycopene (Scheibner et al.,
2.2. Methods
The UPLC-UV/MS system consisted of a Waters Acquity UPLC
with an Acquity Photo Diode Array (PDA) detector and Micromass
Quattro Micro triple quadrupole mass spectrometer, controlled
with MassLynx software (Waters, Milford, MA, USA). Two different
columns were used: AccQ.TagTM
,
100 ꢀ 2.1 mm, particle size
m (C18 phase) and HSS-T3, 100 ꢀ 2.1 mm, particle size
m (both Waters, Dublin, Ireland). The mobile phases used
1.7
1.8
l
l

1356
W.B. van Scheppingen et al. / Food Chemistry 132 (2012) 1354–1359
4.07
1.0e+1
8.0
7.90
8.23
8.67
6.20
0.96
6.0
5.53
5.78
6.85
7.11
9.37
11.47
11.46
8.92
9.00
4.38
7.34
6.46
2.33
4.0
4.89
2.86
1.46
1.95
2.0
0.68
0.0
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
10.00
11.00
12.00
13.00
13.00
13.00
14.00
1.0e+2
8.0e+1
6.0e+1
4.0e+1
2.0e+1
0.0
10.40
10.64
0.98
1.00
12.12
12.00
4.07
4.00
0.79
0.00
2.00
3.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
14.00
0.79
1.5e+1
1.25e+1
1.0e+1
7.5
0.98
4.07
4.89
5.78
9.37
5.53
2.86
2.61
5.0
4.38
6.46
3.56
3.87
1.96
2.00
2.5
1.47
0.68
Time
14.00
0.00
1.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
12.00
Fig. 2. UPLC-UV chromatograms on a C18-columna of samples with MsP1 (top), with 17 ppm H2O2 (centre) and with 500 ppm H2O2 (bottom). a(For conditions see Section 2.)
were water with 0.1% formic acid (A) and acetonitrile with 0.1%
formic acid (B), both purchased from Biosolve (Valkenswaard,
The Netherlands). For the C18 column the gradient was pro-
grammed from 90% A/10% B increasing in 15 min to 10% A/90% B
at a flow rate of 0.30 ml/min and a column temperature of 40 °C.
To get more retention for the compounds in the first few minutes
of the chromatogram the samples were also analysed using an-
other column with a different stationary phase having more inter-
action for polar compounds. For this HSS-T3 column the linear
gradient starts with 100% A and changes in 15 min to 10% A/90%
B with a flow rate of 0.25 ml/min and a column temperature of
40 °C. Data were recorded by the PDA in the scan range of 200–
500 nm (resolution 1.2 nm) with a sampling rate of 10 points/s.
The mass spectrometer was equipped with an electrospray
source operated in the negative ionisation mode (ESIꢁ) using a
desolvation temperature of 350 °C and a capillary voltage of
3.00 kV. The source temperature was 120 °C, and the cone voltage
was kept constant at 40 V. The mass spectra were recorded over a
scan range of m/z 50–400 Da, using a scan time of 0.25 s and an
interscan delay of 50 ms. The electron multiplier detector was set
at 650 V.
The accurate mass spectrometric analysis was performed on a
LTQ-orbitrap equipped with a Thermo Accela pump and autosam-
pler (Thermo Scientific GmbH, Bremen, Germany). The HSS-T3 col-
umn was used with the same chromatographic conditions as
described above for the UPLC-UV/MS. The LTQ-Orbitrap was oper-
ated with an electrospray source in the negative ionisation mode
(ESIꢁ), using a capillary temperature of 300 °C and a voltage of
5.00 kV. A scan range of m/z 100–800 Da was used at a resolution
of 60,000 with automatic gain control at 200,000 for full MS in the
orbitrap and 10,000 for MS/MS in the ion trap. For identification of
norbixin degradation products accurate mass analysis, using the
LTQ-orbitrap, was performed after a daily mass calibration of the
orbitrap with sodium trifluoroacetate cluster ions as described by
Moini, Jones, Rogers, and Jiang (1998). For further characterisation
MS/MS was performed, using an isolation width of 3.00 m/z with a
normalised collision energy of 35%. In MS/MS the most abundant
ions were used from the full scan and as activation type Collision
Induced Dissociation (CID) was used.
2.3. Experimental design
In order to investigate the chemical and enzymatic decomposi-
tion of norbixin, three mixtures were prepared:
(1) 100 ml of phosphate solution + 300
MsP1 + 170 l of hydrogen peroxide 1% (17 ppm, v/v)
(2) 100 ml of phosphate solution + 300 l of annatto + 170
hydrogen peroxide 1% (17 ppm, v/v), control sample
(3) 100 ml of phosphate solution + 300 l of annatto + 5 ml of
hydrogen peroxide 1% (500 ppm, v/v)
ll of annatto + 20
l
l of
l
l
ll of
l
The mixtures contained 0.3% annatto, corresponding to about
40 ppm (v/v) of norbixin, and were kept at room temperature in
the dark under continuous stirring. At regular intervals, samples
of these mixtures were analysed with UPLC-MS without any fur-
ther sample pre-treatment.
For whey, the experiments were repeated by substituting the
phosphate solution in the mixtures by the same volume of whey.
These whey samples were mixed with an equal volume of acetoni-
trile in order to precipitate proteins. After centrifugation, the
supernatant was injected on the UPLC system.
One of the identified decomposition products, 4-oxopent-2-
enoic acid (3-acetylacrylic acid), is commercially available. This

W.B. van Scheppingen et al. / Food Chemistry 132 (2012) 1354–1359
1357
acid was added to whey and stored under dark conditions at room
temperature under continuous stirring. Samples from the spiked
whey were analysed as described above. The fate of 3-acetylacrylic
acid was also studied in a phosphate buffer at pH 6.6 (as in whey)
in the presence of two amino acids: cysteine and lysine. The mix-
tures were placed in the autosampler of the UPLC-MS at 30 °C and
the reactions were followed by continuous injections and analysis
using a short method.
(mixture 2) remained yellow. In the UPLC-UV chromatograms the
norbixin isomers (retention time > 10 min) were not present in
the samples with MsP1 and with high concentrations of hydrogen
peroxide, but are still detected in the control sample (Fig. 2). The
sample treated with MsP1 contained more decomposition prod-
ucts with retention times between 5 and 10 min than the hydrogen
peroxide samples. The norbixin treated with a high concentration
of hydrogen peroxide shows some large unidentified peaks eluting
within the first minute of the chromatogram and likely belonging
to polar compounds. The different patterns of decomposition prod-
ucts after MsP1 and hydrogen peroxide treatment might indicate
that the degradation mechanisms for chemical and enzymatic
breakdown are different. Proposed structures of the decomposition
products are given in Table 1. The UV–visible spectrum of norbixin
is characteristic and in close agreement with the literature, show-
ing high absorption values at wavelengths in the visible range
(>400 nm) and maximum absorptions at 460 and 488 nm (Scotter,
2009). The two carboxyl groups at either end of the norbixin mol-
ecule are conjugated with the double bonds. This results in a spec-
tral shift to a longer wavelength. In the decomposition products
the amount of chromophores is reduced and the absorption max-
ima (kmax) change from visible to the UV region (<400 nm). For
3. Results and discussion
The concentration of 0.3% annatto (alkalinity 0.3 M) results in a
high pH of the experimental mixtures, but in a buffer with a high
salt concentration the norbixin is no longer soluble. The pH values
of the mixtures (about 9–10) did not represent optimal conditions
for bleaching with MsP1, but they provided the opportunity to de-
tect (intermediate) decomposition products. In an experiment with
a lower concentration (0.03%) of annatto at pH 6.5, bleaching was
much faster but gave the same chromatograms with decomposi-
tion products.
The decolourisation in mixture 1 (with MsP1) was faster than in
mixture 3 (500 ppm hydrogen peroxide), and the control sample
Table 1
UPLC-UV/MS identification of norbixin and decomposition products after hydrogen peroxide and peroxidase treatment.a
Retention times (min)
C18 column
Structurea
Mass
kmax
nm
HSS T3 column
3.44
m/z (Da)
113.1
0.70
223
2.30
4.75
139.1
155.1
179.1
205.2
379.2
281
268
320
355
460
1.93
4.33
4.06 + 4.39
5.55 + 5.79
10.66 + 11.48
6.56 + 6.81
7.90 + 8.04
12.56 + 13.30
a
The chemical names of the identified decomposition compounds are (top–bottom): 4-oxopent-2-enoic acid (3-acetylacrylic acid), 4-methyl-6-oxohexa-2,4-dienoic acid,
3-methylhexa-2,4 dienedioic acid, 4-methyl-8-oxonona-2,4,6-trienoic acid, 4,8-dimethyl-10-oxodeca-2,4,6,8-tetraenoic acid, 4,8,13,17-tetramethylicosa-2,4,6,8,10,12,14,
16,18-nonaenedioic acid (norbixin).

1358
W.B. van Scheppingen et al. / Food Chemistry 132 (2012) 1354–1359
Table 2
methyl-4-oxopent-2-enoate, methyl 4-methyl-6-oxohexa-2,4-
dienoate, 6-methoxy-3-methyl-6-oxohexa-2,4-dienoic acid
(methyl trans-b-muconate) and 2-oxopropanal (methylglyoxal). A
complete oxidation of norbixin with permanganate will give acetic
acid (Singh, 2007).
Based on the detected products, a decomposition scheme for
norbixin by MsP1 is proposed (see Fig. 3). The reaction products
indicate the same reaction mechanisms as for the decomposition
of other carotenoids by peroxidases reported by Zorn and co-work-
ers. They proposed an abstraction of hydrogen from an allylic
methyl group, resulting in a resonance-stabilized carbon radical,
followed by a reaction with oxygen. The resulting hydroperoxide
splits into two carbonyl compounds via the Hock cleavage mecha-
nism (Zorn, Langhoff, Scheibner, & Berger, 2003; Zelena et al.,
2009).
High resolution MS identification of norbixin and decomposition products after
hydrogen peroxide and peroxidase treatment.
Detected mass Molecular formula Theoretical mass Differences
RDBa
(m/z)
[M-H]ꢁ
(Da)
mDa
0.20
0.07
ꢁ0.01 ꢁ0.06
0.04 0.22
ꢁ0.01 ꢁ0.05
d ppm
113.02462
139.04014
155.03497
179.07141
205.08701
379.19127
C5H5O3
C7H7O3
C7H7O4
113.02442
139.04007
155.03498
179.07137
205.08702
379.19148
1.77
0.50
3.5
4.5
4.5
5.5
6.5
C
C
C
10H11O3
12H13O3
24H27O4
ꢁ0.21 ꢁ0.55 11.5
a
RDB = Ring plus Double Bond equivalent, including 0.5 for the negative ion.
the different decomposition products, absorption maxima between
223 and 355 nm were obtained.
Mass spectrometric detection was carried out with an electro-
spray ionisation interface in the negative ionisation mode (ESIꢁ).
For carboxylic acids the deprotonated molecular ion [M-H]ꢁ was
detected. The loss of CO2 (M-44) observed during fragmentation
is typical of these compounds. With dicarboxylic acid compounds
the loss of two times 44 Da occurs during fragmentation. The loss
of a fragment of 92 Da is characteristic of the polyene chain, due
to elimination of toluene. By means of accurate mass measure-
ments using high resolution MS the molecular formula of the com-
pounds and the RDB equivalent (Ring plus Double Bonds) can be
calculated. For all the identified molecules the differences between
the detected mass and theoretical mass (d ppm) are within the
specifications of 2 ppm (see Table 2). Based on all the analytical
data (retention times on two different columns, UV-spectra, MS
& MS/MS spectra, and accurate mass data) and the literature, sev-
eral decomposition products were identified. Isomeric decomposi-
tion products (with similar UV- and MS spectra) were detected at
different retention times (see Table 1). Around the double bonds
of polyenes several cis- and trans-isomers can be formed and these
different configurations may result in separate peaks in the
chromatograms. In an ozonolysis experiment with bixin, similar
oxidation products (as methylesters) were detected, such as
The bleaching of whey (with added norbixin) with the peroxi-
dase MsP1 is very efficient. At room temperature and in the pres-
ence of a small amount (17 ppm) of hydrogen peroxide the
colour disappears within 30 min. In the bleached samples none
of the proposed decomposition products of norbixin were detected.
A possible explanation is that all decomposition products of nor-
bixin contain aldehyde or ketone groups which are known to be
very reactive. Most likely, these compounds will react with whey
components. This hypothesis was tested by the addition of one of
the decomposition compounds, i.e. 3-acetylacrylic acid, to the
whey. After one hour, this compound was no longer detected by
UV/MS. To prove that a reaction takes place, a less complex matrix
of a phosphate buffer at pH 6.5 was chosen. In this medium, the
reactive amino acids cysteine and lysine were incubated at 30 °C
with 3-acetylacrylic acid. It was observed that the lysine concen-
tration in the mixture remains constant during the experiment
and is not reactive under these conditions. Otherwise, full conver-
sion occurred with cysteine. The corresponding reaction product,
with a molecular weight of 235 Da, showed up in the chromato-
gram as a separate peak. Besides the reaction with 3-acetylacrylic
acid, the oxidation of two cysteine molecules to cystine takes place
(see Fig. 4). In the presence of hydrogen peroxide both reaction
Molecular Formula: C24H28O4
O
H3C
H3C
Norbixin
OH
OH
CH3
CH3
O
H3C
H3C
H3C
HO
HO
O
O
Molecular Formula: C10H12O3
4-methyl-8-oxonona-
2,4,6-trienoic acid
O
O
O
H3C
H3C
H3C
Molecular Formula: C5H6O3
4-oxopent-2-enoic acid
HO
HO
O
Molecular Formula: C12H14O3
4,8,-dimethyl-10-oxodeca-
2,4,6,8-tetraenoic acid
O
O
Molecular Formula:C7H8O3
4-methyl-6-oxohexa-2,4-dienoic acid
HO
H3C
HO
O
O
Molecular Formula: C7H8O4
3-methylhexa-2,4-dienedioic acid
Fig. 3. Proposed scheme for the decomposition of norbixin by the peroxidase MsP1.

W.B. van Scheppingen et al. / Food Chemistry 132 (2012) 1354–1359
1359
O
C
H2N
O
C
OH
O
H2N
HS
CH2
OH
OH
O
S
+
OH
O
O
2-(2-amino-2-carboxyethylthio)-
4-oxopentanoic acid
3-acetylacrylic acid
Cysteine
O
H2N
C
OH
O
O
C
NH2
SH
H2N
HS
C
S
S
HO
OH
+
HO
C
O
NH2
Cysteine
Cysteine
Cystine
Fig. 4. Proposed reaction for cysteine with 3-acetylacrylic acid and oxidation to cystine.
ionization mass spectrometry in the mass range of 100–4000 Da. Journal of
the American Society for Mass Spectrometry, 9, 977–980.
Najar, S. V., Bobbio, F. O., & Bobbio, P. A. (1988). Effects of light, air, anti-oxidants
and pro-oxidants on annatto extracts (Bixa orellana). Food Chemistry, 29,
283–289.
Preston, H. D., & Rickard, M. D. (1980). Extraction and chemistry of annatto. Food
Chemistry, 5, 47–56.
Roos, A. L. de, Dijk, A. A. van, & Folkertsma B. (2005). Bleaching of dairy products.
International application published under the patent cooperation treaty,
International publication number WO 2005/004616 A1.
products with cysteine are detected but relatively more cystine is
detected. The fast reaction of cysteine with carbonylacrylic deriva-
tives containing an unsubstituted double bond is also described in
the literature (Dal Pozzo, Acquasaliente, Donzelli, DeMaria, &
Nicoli, 1987). Besides the reaction with cysteine the 3-acetylacrylic
acid possibly also reacts with other carbonyl reactive molecules
present in the whey. It is likely that also other decomposition
products of norbixin can react with the whey, although so far no
experiments have been performed to prove this.
Saiz, A. I., Manrique, G. D., & Fritz, R. (2001). Determination of benzoyl peroxide and
benzoic acid levels by HPLC during wheat flour bleaching process. Journal of
Agricultural and Food Chemistry, 49, 98–102.
Scheibner, M., Hülsdau, B., Zelena, K., Nimtz, M., De Boer, L., Berger, R. G., et al.
(2008). Novel peroxidases of Marasmius scorodonius degrade b-carotene. Applied
Microbiology and Biotechnology, 77, 1241–1250.
References
Scotter, M. J. (1995). Characterisation of the coloured thermal degradation products
of bixin from annatto and a revised mechanism for their formation. Food
Chemistry, 53, 177–185.
Scotter, M. J., Castle, L., & Appleton, G. P. (2001). Kinetics and yields for the
formation of coloured and aromatic thermal degradation products of annatto in
foods. Food Chemistry, 74, 365–375.
Scotter, M. (2009). The chemistry and analysis of annatto food colouring: A review.
Food Additives & Contaminants, 26, 1123–1145.
Singh, G. (2007). Chemistry of terpenoids and carotenoids. India: Discovery publishing
house (pp. 272–274).
Smith, K. (2008). Bleaching. Wisconsin: Center for Diary Research, pp. 1–13.
Tennant, D. R., & O’Callaghan, M. (2005). Survey of usage and estimated intakes of
annatto extracts. Food Research International, 38, 911–917.
Tirimanna, A. S. L. (1981). Study of the carotenoid pigments of Bixa orellana L. seeds
by thin layer chromatography. Mikrochimica Acta, 2, 11–16.
Zelena, K., Hardebusch, B., Hülsdau, B., Berger, R., & Zorn, H. (2009). Generation of
norisoprenoid flavours from carotenoids by fungal peroxidases. Journal of
Agricultural and Food Chemistry, 57, 9951–9955.
Zorn, H., Langhoff, S., Scheibner, M., & Berger, R. G. (2003). Cleavage of b,b-carotene
to flavor compounds by fungi. Applied Microbiology and Biotechnology, 62,
331–336.
Zorn, H., Langhoff, S., Scheibner, M., Nimtz, M., & Berger, R. G. (2003). A peroxidase
from Lepista irina cleaves b,b-carotene to flavor compounds. Biological
Chemistry, 384, 1049–1056.
Bouvier, F., Dogbo, O., & Bilal, C. (2003). Biosynthesis of the food and cosmetic plant
pigment bixin (Annatto). Science, 300, 2089–2091.
Croissant, A. E., Kang, E. J., Campbell, R. E., Bastian, E., & Drake, M. A. (2009). The
effect of bleaching agent on the flavor of liquid whey and whey protein
concentrate. Journal of Diary Science, 92, 5917–5927.
Dal Pozzo, A., Acquasaliente, M., Donzelli, G., DeMaria, P., & Nicoli, M. C. (1987). 3-
Carbonylacrylic derivatives as potential antimicrobial agents. Correlations
between activity and reactivity towards cysteine. Journal of Medicinal
Chemistry, 30, 1674–1677.
El-Samragy, Y. (2004). Benzoyl peroxide, chemical and technical assessment (CTA).
61st Joint FAO/WHO Expert Committee on Food Additives, First draft, 1–6.
European Commission (1994). European parliament and council directive 94/36/EC,
of 30 June 1994 on colours for use in foodstuffs. Official Journal European
Commission, L237/13–29.
Joint FAO/WHO (2007). Expert Committee on Food Additives (Meeting 67th: 2006,
Rome, Italy), Evaluation of certain food additives and contaminants, WHO
Technical report series no. 940, 11–15.
Kang, E. J., Campbell, R. E., Bastian, E., & Drake, M. A. (2010). Invited review: Annatto
usage and bleaching in dairy foods. Journal of Diary Science, 93, 3891–3901.
Mercadante, A. Z., Steck, A., & Pfander, H. (1997). Isolation and identification of new
apocarotenoids from annatto (Bixa orellana) seeds. Journal of Agricultural and
Food Chemistry, 45, 1050–1054.
Moini, M., Jones, B. L., Rogers, R. M., & Jiang, L. (1998). Sodium trifluoroacetate as a
tune/calibration compound for positive- and negative-ion electrospray