Identification and Quantitation of Four New 2-Alkylthiazolidine-4-carboxylic Acids Formed in Orange Juice by a Reaction of Saturated Aldehydes with Cysteine

Article abstract of DOI:10.1021/acs.jafc.8b04333

Despite several technological efforts to maximize the quality and shelf life of chilled stored not-from-concentrate orange juice, changes in the overall aroma profile might occur during storage. Besides the degradation of terpenoids, a loss of the aroma-active aldehydes, hexanal, octanal, nonanal, and decanal as well-as of 1-penten-3-one were recently confirmed as a major cause for the changes in the aroma profile of orange juice even during storage under aseptic conditions at 0 °C. To unravel the fate of the aroma-active aldehydes, model experiments were carried out considering the oxidation into the corresponding acids as well as a reaction with free amino acids present in orange juice. The oxidation into the acids could be confirmed by isotope labeling experiments; additionally, the reaction of the four aldehydes mentioned above with l-cysteine yielded four new compounds identified as 2-alkylsubstituted thiazolidine-4-carboxylic acids. Their quantitation in orange juice samples by newly developed stable isotope dilution assays revealed that these acids were already present in the fresh samples but were considerably increased after storage. Labeling experiments in orange juice administered with either labeled octanal or labeled cysteine confirmed that the reaction quickly occurs in the juice. The data contribute another puzzle piece to the loss of aroma-active aldehydes during orange juice storage, which may also be relevant in other foods.

Full text of DOI:10.1021/acs.jafc.8b04333

Subscriber access provided by University of Sunderland
Chemistry and Biology of Aroma and Taste
Identification and Quantitation of Four New 2-
Alkylthiazolidine-4-carboxylic Acids Formed in Or-ange
Juice by a Reaction of Saturated Aldehydes with Cysteine
Veronika Mall, Ines Sellami, and Peter Schieberle
J. Agric. Food Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jafc.8b04333 • Publication Date (Web): 21 Sep 2018
Downloaded from http://pubs.acs.org on October 4, 2018
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a service to the research community to expedite the dissemination
of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in
full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully
peer reviewed, but should not be considered the official version of record. They are citable by the
Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore,
the “Just Accepted” Web site may not include all articles that will be published in the journal. After
a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web
site and published as an ASAP article. Note that technical editing may introduce minor changes
to the manuscript text and/or graphics which could affect content, and all legal disclaimers and
ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or
consequences arising from the use of information contained in these “Just Accepted” manuscripts.
is published by the American Chemical Society. 1155 Sixteenth Street N.W.,
Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 43
Journal of Agricultural and Food Chemistry
Identification and Quantitation of Four New 2-
Alkylthiazolidine-4-carboxylic Acids Formed in Or-
ange Juice by a Reaction of Saturated Aldehydes
with Cysteine
#
,
§
§*
Veronika Mall Ines Sellami and Peter Schieberle
#
LeibnizꢀInstitute for Food Systems Biology at the Technical University of Munich
(formerly Deutsche Forschungsanstalt für Lebensmittelchemie), LiseꢀMeitnerꢀStraße
34, Dꢀ85354 Freising, Germany
§
Lehrstuhl für Lebensmittelchemie, Technische Universität München, Department
für Chemie, LiseꢀMeitnerꢀStraße 34, Dꢀ85354 Freising, Germany
*
Corresponding Author
Phone: +49 871 97698160
Eꢀmail: peter.schieberle@ch.tum.de
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 2 of 43
2
1
ABSTRACT
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
Despite several technological efforts to maximize the quality and shelf life of chilled
stored notꢀfromꢀconcentrate (NFC) orange juice, changes in the overall aroma profile
might occur during storage. Besides the degradation of terpenoids, a loss of the aroꢀ
maꢀactive aldehydes, hexanal, octanal, nonanal and decanal as wellꢀas of 1ꢀpentenꢀ
3ꢀone were recently confirmed as a major cause for the changes in the aroma profile
of orange juice even during storage under aseptic conditions at 0 °C. To unravel the
fate of the aromaꢀactive aldehydes, model experiments were carried out considering
the oxidation into the corresponding acids as well as a reaction with free amino acids
present in orange juice. The oxidation into the acids could be confirmed by isotope
labeling experiments, additionally the reaction of the four aldehydes mentioned above
1
1
1
1
1
1
1
1
1
1
with Lꢀcysteine yielded four new compounds identified as 2ꢀalkylsubstituted thiazoliꢀ
dineꢀ4ꢀcarboxylic acids. Their quantitation in orange juice samples by newly develꢀ
oped stable isotope dilution assays revealed that these acids were already present in
the fresh samples, but were considerably increased after storage. Labeling experiꢀ
ments in orange juice administered with either labeled octanal or labeled cysteine
confirmed that the reaction quickly occurs in the juice. The data contribute another
puzzle piece to the loss of aromaꢀactive aldehydes during orange juice storage,
which may also be relevant in other foods.
20
2
1
2
KEY WORDS: Aldehydes, chilled storage, 2ꢀalkylsubstituted thiazolidineꢀ4ꢀ
2
carboxylic acids, stable isotope dilution assays, Lꢀcysteine
ACS Paragon Plus Environment

Page 3 of 43
Journal of Agricultural and Food Chemistry
3
23
INTRODUCTION
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
Many consumers enjoy a glass of orange juice as a healthy and tasty part of their
daily diet. To meet their demand of a premium orange juice throughout the year, sevꢀ
eral technological processes have been developed to maximize the quality and shelf
life in particular for premium notꢀfromꢀconcentrate (NFC) orange juice, such as chilled
storage under aseptic conditions.
But, despite the high technological standards, in particular extended storage time
1
may lead to changes in the overall aroma. In the past, depending on time,
temperature, oxygen content and light exposure, the loss of ascorbic acid and juice
darkening were the first observed reactions during orange juice storage. Additionally,
systematic model studies using forced storage showed the formation of offꢀflavor
compounds, such as carvone,
αꢀterpineol, 2ꢀmethoxyꢀ4ꢀvinylphenol or 4ꢀhydroxyꢀ2,5ꢀ
1ꢀ3
4,5
dimethylꢀ3(2H)ꢀfuranone. Averbeck and Schieberle for the first time applied a
comprehensive approach, the molecular sensory concept, to characterize the effect
of storage on the overall aroma of orange juice. Next to the formation of offꢀflavor
compounds, also a rapid degradation of terpenoids and saturated aldehydes were
6ꢀ8
confirmed as a function of storage temperature and time. While other studies also
reported a decline in concentration of saturated aldehydes during storage, Peterson
9
et al. were the first to propose an oxidation to their corresponding acids. However, in
5
model trials, Averbeck and Schieberle could not confirm such oxidation reactions for
juice from concentrate.
Although it is commonly known that the degradation of saturated aldehydes mostly
occurs at elevated temperatures, losses in the aldehydes also occurred in aseptically
10
storage juice at 0 °C as recently shown by us using the Sensomics approach. Esꢀ
pecially, saturated aldehydes, such as hexanal, octanal, nonanal and decanal as well
as 1ꢀpentenꢀ3ꢀone and acetaldehyde showed a significant decrease already after a
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 4 of 43
4
1
0
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
short time of chilled storage. Sensory experiments, i.e. aroma recombinates, conꢀ
11
firmed their importance in the aroma of the unstored juice.
Besides the assumed oxidation of the aldehydes to their corresponding acids, no
other degradation pathways of these impact orange juice aroma compounds have yet
been proposed.
Thus, to unravel the fate of the saturated aldehydes during the storage of orange
juice, several model experiments should be carried out as isotope labeling experiꢀ
ments considering a possible oxidation.
Additionally, as already reported in the literature, Lꢀcysteine serves as a nucleoꢀ
phile in many reactions, for example in a reaction with 2,3ꢀbutanedione to 5ꢀacetylꢀ
12
13
2,3ꢀdihydroꢀ1,4ꢀthiazine or in the reaction of 2ꢀoxopropanal. Hence a reaction of
the aldehydes with free amino acids or peptides present in orange juice, such as
lysine, ꢀcysteine and ꢀglutathione should to be studied as well. Using UPLCꢀ
Lꢀ
L
L
TOF/MS, LCꢀMS/MS and GC×GCꢀTOF/MS newly formed compounds should be
identified and quantitated in NFC orange juice.
ACS Paragon Plus Environment

Page 5 of 43
Journal of Agricultural and Food Chemistry
5
64
MATERIALS AND METHODS
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
Materials. Orange juice not from concentrate (NFC) was produced by Tropicana
Products Inc. from Hamlin oranges, and the pasteurized juice was aseptically stored.
in tanks at 0 °C. The fresh juice (on the day of storage) as well as samples after one,
two, four and six months of chilled storage were frozen and send to Germany by air
freight. The samples were further stored at ꢀ80 °C prior to analysis.
Reference Compounds and Chemicals. Reference compounds and chemicals
were obtained from the sources given in parentheses: hexanal, decanal, octanal, acꢀ
etaldehyde, decanoic acid, nonanoic acid, octanoic acid, hexanoic acid, ethanol, citric
acid, hydroxycinnamic acid, ethyl 3,4,5ꢀtrihydroxybenzoate, potassium acetate, disoꢀ
dium hydrogen phosphate, Lꢀcysteine, Lꢀlysine, Lꢀglutathione, dansyl chloride, and
iodoacetic acid (SigmaꢀAldrich Chemie, Taufkirchen, Germany); nonanal (Roth,
Karlsruhe, Germany); acetonitrile LiChrosolv ®, methanol LiChrosolv ®, formic acid,
hydrochloric acid, 2ꢀpropanol, sodium sulfate, ethyl acetate and sulfuric acid (Merck,
Darmstadt). Diethyl ether and dichloromethane (Merck) were freshly distilled before
use.
13
2
8
Isotopically Labeled Standards. [ C ]ꢀOctanoic acid, [ H12]ꢀhexanal were purꢀ
13
15
chased from SigmaꢀAldrich Chemie (Taufkirchen, Germany). [ C N]ꢀ
Lꢀcysteine and
3
2
[ H
2
]ꢀhexanoic acid were from Cambridge Isotope Laboratories Inc. (Andover, Masꢀ
2
2
sachusetts). [ H
6 6 3 3
]ꢀDimethyl sulfoxide (DMSOꢀd ) and [ H ]ꢀacetonitrile (acetonitrileꢀd )
were from Euroisotop (Saarbrücken, Germany).
1
3
10
2
14
2
15
2
4
2
[ C
8
]ꢀoctanal, [ H
4
]ꢀoctanal, [ H
4
]ꢀnonanal, [ H
2
]ꢀdecanal
,
[ H2ꢀ4]ꢀdecanoic
16
2
17
acid and [ H ]ꢀnonanoic acid were synthesized according to the references given
2
in superscript.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 6 of 43
6
88
89
90
91
92
93
94
95
96
97
98
99
Syntheses of 2-Alkylsubstituted Thiazolidine-4-Carboxylic Acids. The 2ꢀ
alkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids (1 – 4, Figure 1) were prepared by
18
condensation of
Lꢀcysteine with the respective aldehyde as previously described.
A solution of the respective unlabeled and labeled aldehyde (hexanal, octanal,
13
2
2
2
8 4 4
nonanal, decanal, [ C ]ꢀoctanal, [ H ]ꢀoctanal, [ H12]ꢀhexanal, [ H ]ꢀnonanal and
2
[ H
2
]ꢀdecanal (each 0.67 mmol) in 2.5 mL of 95% ethanol was added to either a soluꢀ
13
15
tion of Lꢀcysteine or [ C N]ꢀLꢀcysteine (0.57 mmol each) and potassium acetate
3
(0.7 mmol) in 2.5 mL of water. Precipitation of the reaction product occurred after a
few minutes of agitation. The reaction mixture was stirred for 1 h at room temperature
and frozen overnight. The precipitate was filtered with gentle suction and washed
with water, cold ethanol and finally diethyl ether before lyophilization. The dry comꢀ
pound was recrystallized from isopropyl alcohol.
1
1
1
1
1
1
1
1
1
1
1
1
1
1
00
For the preparation of labeled standards, lower amounts of the reactants were
01 used and the purification step was changed. In such cases, the reaction mixture was
02 extracted with ethyl acetate (total volume 25 mL), and the ethyl acetate layer was
03 vaporized. The residue was dissolved in acetonitrile containing a few drops of formic
04 acid.
05
The concentrations of the unlabeled reference compounds were determined by
06 qNMR, while the concentrations of the labeled standards were determined by means
07 of LCꢀMS/MS.
08
Characterization of 2-Alkylsubstituted Thiazolidine-4-Carboxylic Acids.
09 Characterization of the synthesized compounds, was carried out by UPLCꢀTOF/MS,
10 LCꢀMS and one and twoꢀdimensional NMR experiments. The ring closure during the
11 formation of the 2ꢀalkylsubstituted thiazolineꢀ4ꢀcarboxylic acids created a new chiral
12 center at position 2 (Figure 1) yielding two diastereomers, here assigned as a and b,
13 found in a ratio of 1:1 (Table 1 and 2, Supporting Information Table S1 – S6)
ACS Paragon Plus Environment

Page 7 of 43
Journal of Agricultural and Food Chemistry
7
114
2-Pentylthiazolidine-4-carboxylic acid (2ꢀPTꢀ4ꢀCA), (1, Figure 1); UPLCꢀTOF
+
+
1
1
1
1
15 (ESI ), m/z 204.1061 (measured), m/z 204.1058 (calculated for [C H O NS+H] );
9 17 2
+
+
16 MS (ESI ), m/z 204 (100, [M+H] ; Supporting Information Figure S1 a).
1
17
Diastereomer 1a: HꢀNMR (400 MHz, DMSOꢀd
6
, COSY), δ 0.86 [t, 3H, J= 6.76, Hꢀ
18 C(5’)], δ 1.18 – 1.47 [m, 8H, HꢀC(2’, 3’, 4’)], δ 1.61 – 1.83 [m, 1H, HαꢀC(1’)], δ 1.83 –
19 1.96 [m, 1H, HβꢀC(1’)], δ 2.75 [dd, 1H, J= 9.24; 9.84, HαꢀC(5)], δ 3.18 [dd, 1H,
20 J=6.93; 9.84, HβꢀC(5)], δ 3.70 [dd, 1 H, J=6.96; 9.24, HꢀC(4)], δ 4.40 [dd, 1H, J=5.80;
1
1
1
1
1
1
1
13
6
21 7.40, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.3 [C(5’)], δ 22.4/
22 22.5 [C(4’)], δ 27.5/ 27.7 [C(2’)], δ 31.5 [C(3’)], δ 35.3 [C(1’)], δ 37.5 [C(5)], δ 65.7
23 [C(4)], δ 71.6 [C(2)], δ 172.8 [C(6)].
1
24
Diastereomer 1b: HꢀNMR (400 MHz, DMSOꢀd
6
, COSY), δ 0.83 [t, 3H, J= 6.76, Hꢀ
25 C(5’)], δ 1.18 – 1.47 [m, 8H, HꢀC(2’, 3’, 4’)], δ 1.47 – 1.57 [m, 1H, HαꢀC(1’)], δ 1.61 –
26 1.83 [m, 1H, HβꢀC(1’)], δ 2.93 [dd, 1H, J= 5.12; 10.24, HαꢀC(5)], δ 3.08 [dd, 1H,
27 J=7.04; 10.24, HβꢀC(5)], δ 4.06 [dd, 1 H, J=5.16; 6.96, HꢀC(4)], δ 4.50 [dd, 1H,
1
1
1
1
1
1
1
1
1
1
13
6
28 J=6.28; 7.44, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.3 [C(5’)],
29 δ 22.4/ 22.5 [C(4’)], δ 27.5/ 27.7 [C(2’)], δ 31.5 [C(3’)], δ 37.1 [C(1’)], δ 37.1 [C(5)], δ
30 64.6 [C(4)], δ 70.9 [C(2)], δ 173.4 [C(6)].
+
31
2-Heptylthiazolidine-4-carboxylic acid (2ꢀHTꢀ4ꢀCA) (2, Figure 1) UPLCꢀTOF (ESI ),
+
32 m/z 232.1377 (measured), m/z 232.1371 (calculated for [C H O NS+H] ); MS
11
21
2
+
+
33 (ESI ), m/z 232 (100, [M+H] ; Supporting Information Figure S1 b).
1
34
Diastereomer 2a: HꢀNMR (400 MHz, DMSOꢀd
6
, COSY), δ 0.86 [t, 3H, J= 6.76, Hꢀ
35 C(7’)], δ 1.15 – 1.47 [m, 10H, HꢀC(2’, 3’, 4’, 5’, 6’)], δ 1.60 – 1.83 [m, 1H, HαꢀC(1’)], δ
36 1.83 – 1.96 [m, 1H, HβꢀC(1’)], δ 2.75 [dd, 1H, J= 9.20; 9.84, HαꢀC(5)], δ 3.18 [dd, 1H,
37 J=6.97; 9.84, HβꢀC(5)], δ 3.69 [dd, 1 H, J=6.96; 9.20, HꢀC(4)], δ 4.40 [dd, 1H, J=5.80;
1
1
1
13
6
38 7.40, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4 [C(7’)], δ 22.5
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 8 of 43
8
1
1
1
1
1
39 [C(6’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3 [C(2’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3
40 [C(3’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3 [C(4’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3
41 [C(5’)], δ 35.3 [C(1’)], δ 37.5 [C(5)], δ 65.7 [C(4)], δ 71.6 [C(2)], δ 172.8 [C(6)].
1
42
Diastereomer 2b: HꢀNMR (400 MHz, DMSOꢀd
6
, COSY), δ 0.83 [t, 3H, J= 6.76, Hꢀ
43 C(5’)], δ 1.15 – 1.47 [m, 10H, HꢀC(2’, 3’, 4’, 5’, 6’)], δ 1.47 – 1.60 [m, 1H, HαꢀC(1’)], δ
44 1.60 – 1.83 [m, 1H, HβꢀC(1’)], δ 2.93 [dd, 1H, J= 5.12; 10.24, HαꢀC(5)], δ 3.08 [dd,
45 1H, J=7.00; 10.24, HβꢀC(5)], δ 4.06 [dd, 1 H, J=5.12; 6.97, HꢀC(4)], δ 4.50 [dd, 1H,
1
1
1
1
1
1
1
1
1
1
1
1
13
6
46 J=6.24; 7.44, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4 [C(7’)],
47 δ 22.5 [C(6’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3 [C(2’)], δ 27.9/ 28.1/ 29.0/ 29.1/
48 29.2/ 29.3 [C(3’)], δ 27.9/ 28.1/ 29.0/ 29.1/ 29.2/ 29.3 [C(4’)], δ 27.9/ 28.1/ 29.0/ 29.1/
49 29.2/ 29.3 [C(5’)], δ 37.1 [C(1’)], δ 37.1 [C(5)], δ 64.6 [C(4)], δ 70.9 [C(2)], δ 173.4
50 [C(6)].
+
51
2-Octylthiazolidine-4-carboxylic acid (2ꢀOTꢀ4ꢀCA), (3, Figure 1) UPLCꢀTOF (ESI ),
+
52 m/z 246.1528 (measured), m/z 246.1528 (calculated for [C12
H
23
O
2
NS+H] ); MS
+
+
53 (ESI ), m/z 246 (100, [M+H] ; Supporting Information Figure S1 c).
1
54
Diastereomer 3a: HꢀNMR (400 MHz, DMSOꢀd
6
, COSY), δ 0.86 [t, 3H, J= 7.07, Hꢀ
55 C(8’)], δ 1.15 – 1.45 [m, 12H, HꢀC(2’, 3’, 4’, 5’, 6’, 7’)], δ 1.63 – 1.78 [m, 1H, Hαꢀ
56 C(1’)], δ 1.83 – 1.94 [m, 1H, HβꢀC(1’)], δ 2.70 [dd, 1H, J= 9.20; 9.77, HαꢀC(5)], δ 3.16
57 [dd, 1H, J=6.97; 9.77, HβꢀC(5)], δ 3.60 [dd, 1 H, J=6.97; 9.20, HꢀC(4)], δ 4.38 [dd, 1H,
1
1
1
1
1
1
1
1
13
6
58 J=5.75; 7.36, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4 [C(8’)],
59 δ 22.5 [C(7’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(2’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(3’)], δ 27.8/
60 28.1/ 29.3/ 29.4 [C(4’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(5’)], δ 31.7 [C(6’)], δ 35.4 [C(1’)], δ
61 37.8 [C(5)], δ 66.5 [C(4)], δ 71.7 [C(2)], δ 173.0 [C(6)].
1
62
Diastereomer 3b: HꢀNMR (400 MHz, DMSOꢀd , COSY), δ 0.86 [t, 3H, J= 7.07, Hꢀ
6
63 C(8’)], δ 1.15 – 1.45 [m, 12H, HꢀC(2’, 3’, 4’, 5’, 6’, 7’)], δ 1.45 – 1.57 [m, 1H, Hαꢀ
ACS Paragon Plus Environment

Page 9 of 43
Journal of Agricultural and Food Chemistry
9
164 C(1’)], δ 1.63 – 1.78 [m, 1H, HβꢀC(1’)], δ 2.88 [dd, 1H, J= 5.52; 10.18, HαꢀC(5)], δ
1
1
1
1
1
1
1
1
1
1
65 3.08 [dd, 1H, J=6.97; 10.18, HβꢀC(5)], δ 3.99 [dd, 1 H, J=5.52; 6.97, HꢀC(4)], δ 4.55
13
6
66 [dd, 1H, J=6.50; 7.22, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4
67 [C(8’)], δ 22.5 [C(7’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(2’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(3’)],
68 δ 27.8/ 28.1/ 29.3/ 29.4 [C(4’)], δ 27.8/ 28.1/ 29.3/ 29.4 [C(5’)], δ 31.7 [C(6’)], δ 37.5
69 [C(1’)], δ 37.5 [C(5)], δ 64.9 [C(4)], δ 70.9 [C(2)], δ 173.4 [C(6)].
+
70
2-Nonylthiazolidine-4-carboxylic acid (2ꢀNTꢀ4ꢀCA) (4, Figure 1; UPLCꢀTOF (ESI ),
+
71 m/z 260.1684 (measured), m/z 260.1684 (calculated for [C13
H
25
O
2
NS+H] ); MS
+
+
72 (ESI ), m/z 260 (100, [M+H] ; Supporting Information Figure S1 d).
1
73
Diastereomer 4a: HꢀNMR (400 MHz, DMSOꢀd , COSY), δ 0.86 [t, 3H, J= 6.98, Hꢀ
6
74 C(9’)], δ 1.15 – 1.46 [m, 12H, HꢀC(2’, 3’, 4’, 5’, 6’, 7’, 8’)], δ 1.62 – 1.81 [m, 1H, Hαꢀ
75 C(1’)], δ 1.84 – 1.95 [m, 1H, HβꢀC(1’)], δ 2.74 [dd, 1H, J= 9.20; 9.83, HαꢀC(5)], δ 3.18
76 [dd, 1H, J=6.93; 9.83, HβꢀC(5)], δ 3.69 [dd, 1 H, J=6.93; 9.20, HꢀC(4)], δ 4.40 [dd, 1H,
1
1
1
1
1
1
1
1
1
13
77 J=5.86; 7.46, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4 [C(9’)],
6
78 δ 22.6 [C(8’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(2’)], δ 27.9/ 28.1/ 29.1/ 29.2/
79 29.3/ 29.4 [C(3’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(4’)], δ 27.9/ 28.1/ 29.1/ 29.2/
80 29.3/ 29.4 [C(5’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(6’)], δ 31.8 [C(7’)], δ 35.3
81 [C(1’)], δ 37.5 [C(5)], δ 66.7 [C(4)], δ 71.6 [C(2)], δ 172.8 [C(6)].
1
82
Diastereomer 4b: HꢀNMR (400 MHz, DMSOꢀd , COSY), δ 0.86 [t, 3H, J= 6.98, Hꢀ
6
83 C(9’)], δ 1.15 – 1.46 [m, 12H, HꢀC(2’, 3’, 4’, 5’, 6’, 7’, 8’)], δ 1.48 – 1.58 [m, 1H, Hαꢀ
84 C(1’)], δ 1.62 – 1.81 [m, 1H, HβꢀC(1’)], δ 2.93 [dd, 1H, J= 5.09; 10.22, HαꢀC(5)], δ
85 3.07 [dd, 1H, J=7.02; 10.22, HβꢀC(5)], δ 4.06 [dd, 1 H, J=5.09; 7.02, HꢀC(4)], δ 4.53
1
1
1
1
1
13
86 [dd, 1H, J=6.30; 7.41, HꢀC(2)]; CꢀNMR (100 MHz, DMSOꢀd , HSQC, HMBC), δ 14.4
6
87 [C(9’)], δ 22.6 [C(8’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(2’)], δ 27.9/ 28.1/ 29.1/
88 29.2/ 29.3/ 29.4 [C(3’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(4’)], δ 27.9/ 28.1/ 29.1/
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 10 of 43
10
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
89 29.2/ 29.3/ 29.4 [C(5’)], δ 27.9/ 28.1/ 29.1/ 29.2/ 29.3/ 29.4 [C(6’)], δ 31.8 [C(7’)], δ
90 35.3 [C(1’)], δ 37.1 [C(5)], δ 64.6 [C(4)], δ 70.7 [C(2)], δ 173.4 [C(6)].
91
92
Model Reactions.
Stability of Aldehydes under Nitrogen or Oxygen. An aqueous solution of hexanal,
93 octanal, nonanal and decanal (~5 ꢁmol/L), respectively, adjusted to the natural pH of
94
95
orange juice (pH 3.6) with citric acid, was stored in gas tight brown glass vessels at
37 °C (in an warming cabinet) for 4 weeks either under nitrogen or an air atmosꢀ
96 phere. In further trials, hydroxycinnamic acid or ethyl 3,4,5ꢀtrihydroxybenzoate (~0.05
97 ꢁmol/L) were added as antioxidants. The stored solutions were extracted with diethyl
98 ether, subjected to distillation by Solvent Assistant Flavor Evaporation (SAFE) and
99 the volatiles formed were analyzed by GC/MS.
00
01 solution of octanal and the free amino acid,
02 respectively, was adjusted to pH 3.6 with citric acid and stored in gas tight brown
Reaction of Aldehydes with Free Amino Acids and Glutathione. First, an aqueous
L
ꢀlysine, ꢀcysteine, and ꢀglutathione,
L
L
03 glass bottles at 37 °C (in an warming cabinet) for 14 days. After dilution with doubly
04 distilled water (1:1000) and subsequent filtration (0.45 ꢁm, Schleicher & Schuell,
05 Dassel, Germany), the solutions were subjected to a screening by UPLCꢀTOF/MS.
06 For hexanal, nonanal and decanal only the model trial with
L
ꢀcysteine was carried
ꢀcysteineꢀtrials was
07 out. For NMR characterization the precipitate of the aldehyde/
L
08 washed with water, cold ethanol and finally diethyl ether, followed by lyophilization
09 and recrystallization. The purified crystals were dissolved in deuterated dimethyl sulꢀ
10 foxide (DMSOꢀd ) and subjected to NMR analysis.
6
11
12 [ C
13 tight brown glass bottle was stored either for 14 d at 37 °C (in a warming cabinet) or
Formation of 2-Alkylsubstituted Thiazolidine-4-Carboxylic Acids in Orange Juice.
13
8
]ꢀOctanal was spiked to a sample of fresh orange juice. The sample in a gas
15
13
14 for 14 d at 7 °C (in a refrigerator), respectively. In a third trial, [ N C ]ꢀ
Lꢀcysteine
3
ACS Paragon Plus Environment

Page 11 of 43
Journal of Agricultural and Food Chemistry
11
2
2
2
2
15 was spiked to the juice and stored for 14 d at 7 °C. Several samples were taken durꢀ
16 ing the storage period and the formation of the formed, differently labeled 2ꢀ
2
13
17 heptylthiazolidineꢀ4ꢀcarboxylic acids was quantitated using either [ H
4
]ꢀ or [ C
8
]ꢀ
18 labeled 2ꢀheptylthiazolidineꢀ4ꢀcarboxylic acid as internal standard.
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
19
Isolation of the Volatiles from Orange Juice. The workꢀup procedure was carꢀ
10
20 ried out as described previously. After the separation of neutral/ basic (NBF) and
21 acidic fraction (AF), the aldehydes were recovered in the NBF, while the acids were
22 found in AF.
23
Quantitation of Aldehydes and Their Corresponding Acids in Orange Juice
24 and Model Solutions.
25
Stable isotope dilution assays of the four aldehydes and four acids under considꢀ
10
26 eration were carried out as described previously.
27
28
29
Quantitation of
was carried out as described previously.
Quantitation of 2-Alkylsubstituted Thiazolidine-4-Carboxylic Acids. An aliquot
L-Cysteine. For the quantitation, a stable isotope dilution assay
18
30 of the isotopically labeled standards was added to the orange juice (10 g). After 30
31 minutes of equilibration, the sample was centrifuged (4600 g, 20 min, 6 °C), and the
32 supernatant was filtered (0.45 ꢁm, Schleicher & Schuell) and finally analyzed by
33 means of HPLCꢀMS/MS. A response factor was determined by monitoring the most
34 abundant transitions (Table 3) of analyte/ standard mixtures in different ratios (1:20,
35 1:10, 1:5, 1:3, 1:1, 3:1, 5:1, 10:1, 20:1) in water/acetonitrile (80:20).
36
High Resolution Gas Chromatography–Mass Spectrometry (HRGC/MS).
37 HRGC/MS was performed by using a HewlettꢀPackard gas chromatograph 5890 seꢀ
38 ries II (Waldbronn, Germany) connected to a Finnigan sector field mass spectrometer
39 type MAT 95 S (Bremen, Germany). 0.5 ꢁL were injected onꢀcolumn on an Agilent
40 DBꢀFFAP (30 m, 0.25 mm i.d., 0.25 ꢁm film thickness) (Waldbronn, Germany) with
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 12 of 43
12
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
41 the following temperature program: 40 °C held for 2 min, 20 °C/min rate up to 60 °C
42 (held for 2 min), 6 °C/min rate up to 180 °C, followed by 20 °C/min rate up to 230 °C
43 (held for 5 min). Mass spectra were generated in the electron impact mode (MSꢀEI) at
44 70 eV with a scan range from 35 – 250 m/z (scan rate 0.8 s/d) and a resolution of
45 5000 with an accuracy error < 5 ppm.
46
Comprehensive Two-Dimensional Gas Chromatography–Time of Flight/Mass
47 Spectrometry (GC×GC-TOF/MS). A Lego Pegasus 3D GC×GCꢀTOFꢀMS instrument
48 (St. Joseph, MI, USA) was used consisting of an Agilent GC model 6890N (Waldꢀ
49 bronn, Germany) equipped with a dual stage quadꢀjet thermal modulator and a secꢀ
50 ondary oven coupled to the mass spectrometer providing unit mass resolution. A
51 PTVꢀinlet was used for coldꢀonꢀcolumn injection, operated by a CTCꢀAnalytics GCꢀ
52 PAL auto sampler (Zwingen, Switzerland). In the first dimension an Agilent DBꢀFFAP
53 (30 m, 0.25 i.d., 0.25 ꢁm film thickness) equipped with a deactivated precolumn (5 m,
54 0.53 mm i.d.) was connected to an Agilent DBꢀ1701 column (2 m, 0.1 i.d., 0.1 ꢁm film
55 thickness) in the second dimension. The primary oven temperature was held for 2
56 min at 40 °C, then raised with 9 °C/min to 60 °C, held for 2 min, then raised with 9
57 °C/min to 180 °C and then with 6 °C/min to the final temperature 230 °C, held for 5
58 min. The secondary oven started at 47 °C, held for 2 min, was then raised with 9
59 °C/min to 60 °C, held isothermal for 2 min, then raised with 6 °C/min to 187 °C, folꢀ
60 lowed by a rise of 9 °C/min to the final temperature of 250 °C, held for 1 min. Modulaꢀ
61 tion period (4 s) and temperature programming rate were adjusted to obtain a miniꢀ
62 mum of three modulations per peak at the concentration level equal to the limit of
63 quantitation (LoQ). The detector voltage was optimized and set to 1750 V. Mass
64 spectra were acquired within m/z 30 − 350 at a rate of 100 spectra/s. Data were
65 elaborated using GC Image (GCꢀImage, Lincoln, NE, USA).
ACS Paragon Plus Environment

Page 13 of 43
Journal of Agricultural and Food Chemistry
13
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
66
Ultra-Performance Liquid Chromatography/ Time of Flight–Mass Spectrome-
67 try (UPLC/TOF-MS). High resolution mass spectra were measured on a Waters
68 Synapt G2 HDMS mass spectrometer (Manchester, UK) coupled to a Waters Acquity
69 UPLC core system (Milford, MA, USA) consisting of a binary solvent manager, a
70 sample manager and column oven. All compounds were dissolved in 1 mL waꢀ
71 ter/methanol (9:1, v/v) and aliquots (1ꢀ5 ꢁL) were injected onto a 2.1 x 150 mm i.d.,
72 1.7 ꢁm, Waters ACQUITY CSHT C18 column operated with a flow rate of 0.4 mL/min
73 at a temperature of 40 °C. For chromatography the following gradient was used:
74 starting with 100% water, increasing the acetonitrile concentration until 100% within 4
E
75 min and then kept constant for 30 s. Scan time for the MS method (centroid) was set
76 to 0.1 s. Analyses were performed in the positive ESI and in resolution mode using
77 the following ion source parameters: capillary voltage, +2.5 kV; sampling cone, 50 V;
78 extraction cone, 4.0 V; source temperature, 120 °C; desolvation temperature, 450 °C;
79 gone gas, 10 L/h; and desolvation gas, 850 L/h. Data processing was performed by
80 using Waters MassLynx 4.1 SCN 779 and the elemental composition tool for deterꢀ
81 mining the accurate mass. All data were lock mass corrected to the pentapeptide
+
82 leucine enkephaline (TyrꢀGlyꢀGlyꢀPheꢀLeu, m/z 556.2771, [M+H] ) in a solution (2
83 ng/ꢁL) of acetonitrile/ 0.1% formic acid (1:1, v/v). Scan time for the lock mass was set
84 to 0.3 s, an interval of 15 and 3 scans to average with a mass window of ±0.3 Da.
85 Calibration of the Synapt G2 in the range form m/z 50 to 1300 was performed using a
86 solution of sodium formate (0.5 mmol/L) in 2ꢀpropanol/water (90:10, v/v). The UPLC
87 and Synapt G2 systems were operated with the Waters MassLnyx software.
88
High Performance Liquid Chromatography–Tandem Mass Spectrometry
89 (HPLC-MS/MS). A Finnigan TSQ Quantum Discovery quadrupole mass spectrometer
90 coupled with a Finnigan Surveyor Plus HPLC system (Thermo Electron Corporation,
91 Waltham, MA) was used for LCꢀMS/MS analysis. The stationary phase was a Pheꢀ
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 14 of 43
14
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
92 nomenex Synergi HydroRP C18 column (2.0 mm × 150 mm, 4 ꢂm particle size, 8 nm
93 pores) (Aschaffenburg, Germany), which was equipped with a Phenomenex C18
94 guard column. For the separation of the 2ꢀsubstituted thiazolidineꢀ4ꢀcarboxylic acids,
95 gradient elution (flow rate: 0.2 ml/min) was employed with aqueous 0.1% formic acid
96 (solvent A) and 0.1% formic acid in acetonitrile (solvent B). The initial condition of
97 80% A and 20% B were held for 2 min, then the proportion of solvent B was raised to
98 90% within the next 16 min and held for 5 min. Before each injection, the column was
99 equilibrated for 15 min. Injection was carried out at full loop mode, and injection volꢀ
+
00 ume was set at 10 ꢁL. The mass spectrometer was operated in ESI mode with the
01 following conditions: spray needle voltage, 3.2 kV; capillary temperature, 300 °C; caꢀ
02 pillary offset, 35 V; sheath gas, 30 arbitrary units (arb); auxiliary gas, 15 arb; collision
03 gas (argon) pressure, 1.2 mTorr; scan time, 0.2 s, peak width of both quadrupole filꢀ
04 ters 1 and 3, 0.7 amu. During method development, the unlabeled compounds were
05 subjected to MS/MS using a full scan in the product mode to find the two most inꢀ
06 tense and specific product ions for selected reaction monitoring (SRM). To optimize
07 the intensity of these product ions, different collision energies in quadrupole 2 were
08 tested using SRM mode (Table 3).
09
Nuclear Magnetic Resonance Spectroscopy (NMR). All NMR spectra were proꢀ
10 duced with a Bruker Avance III 400 spectrometer (400MHz) equipped with a
11 BBFOplus probe (298 K; Rheinstetten, Germany). The samples were dissolved in
6
12 deuterated dimethyl sulfoxide (DMSOꢀd ) and filled in NMRꢀtubes (178 x 5 mm;
13 Schott E, Mainz, Germany). The chemical shift was referenced to TMS. The Bruker
14 software TopSpin (Version 3.0, Rheinstetten, Germany) was used for data proꢀ
1
13
15 cessing. The appropriate standard pulse sequences for Hꢀ and CꢀNMR measureꢀ
2,3
16 ments, as well as for the standard 2D NMR experiments gsꢀCOSY ( JH,H correlaꢀ
1
2,3
17 tion), gsꢀHSQC ( J correlation) and gsꢀHMBC ( J correlation) were taken from
H,C
H,C
ACS Paragon Plus Environment

Page 15 of 43
Journal of Agricultural and Food Chemistry
15
3
18 the library of the TopSpin software. Data interpretation was then carried out with
19 MestReNova V.5.1.0ꢀ2940 (Mestrelab Research, La Coruña, Spain).
3
3
3
3
20
Quantitative NMR (qNMR). Determination of the concentrations of newly syntheꢀ
ꢀtyrosine as reference as
21 sized compounds was carried out by means of qNMR with
L
20
22 described previously.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 16 of 43
16
323
RESULTS AND DISCUSSION
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
24
Oxidation of Aldehydes into Corresponding Acids. Quantitative data obtained
25 in our recent study on the storage of NFC orange juice, indicated considerable losses
10
26 of saturated and unsaturated aldehydes. With a series of sensory trials, it was conꢀ
27 firmed that these aldehydes have a positive impact on the overall aroma of the unꢀ
10,11
28 stored orange juice.
But, although the degradation of aldehydes during storage of
5
ꢀ9
29 orange juice has already been observed in several former studies, the reason for
30 the losses has not yet been clarified. In order to get to the bottom of this phenomeꢀ
13
31 non, [ C
8
]ꢀoctanal was spiked to an authentic fresh orange juice and the sample was
32 stored for four weeks at 37 °C. The volatiles were then extracted and isolated accordꢀ
10
33 ing to the previously described protocol and subjected to GC×GCꢀTOF/MS analysis.
34 For comparison, a nonꢀspiked juice was stored under the same conditions. Monitorꢀ
35 ing the volatiles present in the spiked juice for isotopical shifts, indicated traces of
13
13
8 8
36 [ C ]ꢀoctanol next to greater amounts of [ C ]ꢀoctanoic acid (data not shown). These
37 findings confirmed an oxidation of octanal to its corresponding acid as it was previꢀ
9
38 ously suggested by Peterson et al. To get more information on the responsible
39 mechanism of this oxidation, aqueous solutions of octanal (pH 3.6) were stored either
40 under nitrogen or an air atmosphere for 14 days at 37 °C. In a third trial, waterꢀ
41 soluble antioxidants were added to the octanal solutions. The concentrations of the
42 remaining aldehyde and its corresponding acids were monitored by means of stable
43 isotope dilution assays using HRGCꢀMS. Only 47% or 27% of the initial octanal conꢀ
44 centration were recovered in the samples stored under nitrogen or air, respectively.
45 However, adding the antioxidants resulted in a stable solution of octanal. In this trail,
46 octanoic acid was not detectable after storage (data not shown). This can only be
47 explained by a radical autoxidation of octanal to its corresponding acid via a peroxyꢀ
21–23
48 acid as intermediate (Figure 2).
ACS Paragon Plus Environment

Page 17 of 43
Journal of Agricultural and Food Chemistry
17
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
49
However, the quantitative data showed an imbalance in their relationship. While
50 the increase of hexanoic acid as well as decanoic acid were lower than the losses in
51 the aldehydes, the concentrations of octanoic acid and nonanoic acid were even
52 higher than the decrease in both aldehydes (Table 4). These observations led to the
53 conclusion that also other mechanisms may lead to the formation of the organic acids
54 during storage, and that there must be another degradation pathway besides the radꢀ
55 ical induced oxidation of the aldehydes.
56
Because the experiment with labeled octanal (see above) did not hint at any other
57 possible degradation product in sufficient concentrations, possible reactions with the
58 nonꢀvolatile fraction of the juice were investigated.
59
60 ange Juice. First, model solutions of octanal with different single free amino acids,
61 such as ꢀcysteine and ꢀlysine, and glutathione showing a reactive side chain, reꢀ
Characterization of 2-Alkylsubstituted Thiazolidine-4-Carboxylic Acids in Or-
L
L
62 spectively, were prepared in an acidic aqueous solution (pH 3.6) and stored for 14
63 days at 37 °C. The samples were diluted 1:1000 with water and subjected to a
64 screening by means of UPLC/TOFꢀMS. Monitoring of the solutions with
65 ꢀglutathione resulted only in the identification of the respective amino acid and glutaꢀ
66 thione disulfide, but no reaction product with octanal was found. However, in the susꢀ
Lꢀlysine and
L
67 pension of the octanal/
Lꢀcysteine model solution an unknown reaction product with a
+
68 m/z ratio of 232.136576 was found with the sum formula [C11
H
21
O
2
NS+H] . This forꢀ
69 mula was predicted to be caused by 2ꢀheptylthiazolidineꢀ4ꢀcarboxylic acid, a condenꢀ
70 sation product of ꢀcysteine and octanal. The experiment was extended to the other
L
71 saturated aldehydes hexanal, nonanal and decanal and the corresponding 2ꢀ
72 alkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids could be predicted on the basis of the
73 UPLC/TOFꢀMS monitoring (Table 5). For a full characterization of the newly found
74 compounds, NMR experiments were carried out and compared to the synthesized 2ꢀ
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 18 of 43
18
3
3
3
3
3
3
3
3
3
3
3
3
3
3
75 alkylthiazolidineꢀ4ꢀcarboxylic acids. While a synthesis of these acids was already deꢀ
18
76 scribed in the late 1950ies, no full NMR characterization was available in the literaꢀ
77 ture.
78
Exemplarily the characterization of 2ꢀpentylthiazolidineꢀ4ꢀcarboxylic acid (1; Figure
79 1) (Table 1 and 2) will thus be explained in more detail. To characterize the linkage
80 between the
Lꢀcysteine moiety and the saturated aldehydes, oneꢀ and two dimenꢀ
1
81 sional NMR experiments were carried out. In the H spectrum, the two doublets of
82 doublets observed at 3.70 ppm (HꢀC(4a)) and 4.06 ppm (HꢀC(4b)), both integrating
83 for one proton, were assigned to the Lꢀcysteineꢀmoiety. Application of the HMBC exꢀ
84 periment showed a linkage of these protons to two carbon signals at 64.6 ppm
85 (C(4b)) and 65.7 ppm (C(4a)). In the COSY spectrum (Figure 3) these two protons
86 showed correlations with the two diastereotopic methylene groups, HꢀC(4a) featured
1
1
87
H, H couplings with HαꢀC(5a) (J=9.24) and HβꢀC(5a) (J=6.96), while HꢀC(4b) was
88 correlated with HαꢀC(5b) and HβꢀC(5b) with respective coupling constants of J=6.96
1
1
3
3
3
3
3
3
3
3
3
3
3
4
89 and J=5.12. The H, H connectivity between HαꢀC(5) and HβꢀC(5) was found to be
90 J=9.84 (a) or J=10.24 (b), respectively. The identification of two individual proton sigꢀ
91 nals originating from C(4) and the chemical shift of the diastereotopic methylene
92 group into four signals with integrals of each one proton suggested that a new chiral
93 center was introduced to the molecule and that two diastereomers were formed durꢀ
94 ing the reaction. The chemical shift of the carbon signal of C(6) (172.8 ppm (a) and
13
95 173.4 ppm (b)) in the C spectrum and the DEPT experiment led to the conclusion
96 that the carboxylic acid sits on this position. Furthermore, the doublets of doublet at
97 4.40 ppm (a) and 4.50 ppm (b) were assigned to the new chiral center at HꢀC(2). The
98 HMBC experiment showed the correlation to C(2a) (71.6 ppm) and C(2b) (70.9) and
1
1
99 the COSY experiment revealed H, H connectivities to the diastereotopic methylene
00 groups at C(1’). Hence, HꢀC(2a) showed couplings to HαꢀC(1’a) and HβꢀC(1’a) with
ACS Paragon Plus Environment

Page 19 of 43
Journal of Agricultural and Food Chemistry
19
4
4
4
4
4
4
4
4
01 J=5.80 and J=7.40, while HꢀC(2b) featured coupling constants of J=6.28 and J=7.44
02 to HαꢀC(1’b) and HβꢀC(1’b). Additionally, the long range couplings of C(2) to C(5) and
03 C(4) and C(1’) proved the new linkage between the carbon chain and the thiazolidine
04 ring (Figure 1). Due to the similar chemical shift of HꢀC(2’), HꢀC(3’) and HꢀC(4’), a
05 multiplet at 1.18 to 1.47 ppm with an joint integral of 12 protons for a and b was found
1
06 in the H spectrum. Thus, a correct assignment of the individual carbon atoms to the
13
07 signals in the C spectrum was not possible for C(2’a) and C(2’b) or C(4’a) and
08 C(4’b).
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
09
The other three 2ꢀalkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids formed from hexꢀ
10 anal, nonanal and decanal featured similar NMR patterns. Only the integral of the
11 multiplet for the elongated carbon chain was found for the larger number of protons
12 (Supporting Information Table S1 – S6).
13
To screen the authentic reference orange juices for these newly found 2ꢀ
14 alkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids, a fresh and a 6 month stored juice
15 were subjected to HPLCꢀMS/MS. After optimizing the LCꢀMS/MS conditions as well
16 as the collision energy to monitor the four 2ꢀalkylsubstituted thiazolidineꢀ4ꢀcarboxylic
17 acids, all four compounds were found to be present in the stored orange juice.
18
Quantitation of 2-Substituted Thiazolidine-4-Carboxylic Acids. To quantitate
19 the four thiazolidine carboxylic acids, stable isotope dilution assays were developed
13
20 using the newly synthesized deuterated or [ C]ꢀlabeled analogues, respectively. All
21 four 2ꢀalkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids could be quantitated in one run
22 within less than 25 min. Monitoring the specific mass transitions of analyte and laꢀ
23 beled standards (Table 3), the concentration of the acids could be determined in a
24 fast and precise way (Figure 4).
25
Following this approach, concentrations ranging between 14.2 ꢁg/kg 2ꢀ
26 octylthiazolidineꢀ4ꢀcarboxylic acid (2ꢀOTꢀ4ꢀAC) and 160 ꢁg/kg 2ꢀheptylthiazolidineꢀ4ꢀ
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 20 of 43
20
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
27 carboxylic acid (2ꢀHTꢀ4ꢀAC) were already found in the fresh orange juice. However,
28 after six months of chilled storage higher concentrations ranging from 20.1 ꢁg/kg (2ꢀ
29 OTꢀ4ꢀAC) to 202 ꢁg/kg (2ꢀHTꢀ4ꢀAC) were found, equivalent to an increase of 26% to
30 55% (Table 6).
31
Additionally, the free cysteine was determined in the fresh and six month stored
32 juice to see if the storage had an impact on the amount of the free amino acid. For
19
33 this purpose, the SIDA approach according to Reinbold et al. was followed using a
13
15
34 derivatization with dansyl chloride and [ C
3
N]ꢀLꢀcysteine as internal standard. After
35 the chilled storage of the juice for six months the concentration of free Lꢀcysteine was
36 decreased from 2.89 mg/kg to 2.03 mg/kg.
37
Studies on the Formation of 2-Alkylsubstituted Thiazolidine-4-Carboxylic
38 Acids (2-AT-4-CA). While the concentrations of saturated aldehydes and free Lꢀ
39 cysteine declined after chilled storage, the concentrations of 2ꢀATꢀ4ꢀCAs went up.
40 Thus, the formation can be proposed as a condensation reaction between the aldeꢀ
41 hyde and Lꢀcysteine leading to a ring formation between the carbonyl function of the
42 aldehyde and the aminoꢀ and sulfanyl group of
43 ure 5).
Lꢀcysteine by water elimination (Figꢀ
44
45 ments with orange juice using stable isotope labeled analogues of
46 octanal, respectively, were carried out. For this purpose, the juice was either spiked
To confirm that the same reaction also takes place in orange juice, spiking experiꢀ
Lꢀcysteine and
13
15
13
47 with [ C
3 8
N]ꢀLꢀcysteine or with [ C ]ꢀoctanal, and stored for several days at 7 °C or
48 37 °C, respectively. The concentrations of the unlabeled and labeled 2ꢀ
2
49 heptylthiazolidineꢀ4ꢀcarboxylic acid were measured using [ H
4
]ꢀ2ꢀHTꢀ4ꢀCA as internal
13
13
50 standard to monitor the formation of [ C
8 8
]ꢀ2ꢀHTꢀ4ꢀCA formed from the added [ C ]ꢀ
13
51 labeled octanal (Table 7). At 7 °C, the concentration of the labeled [ C
8
]ꢀ2ꢀHTꢀ4ꢀCA
52 quickly raised to 262 ꢁg/kg within the first hour, but then dropped and remained at
ACS Paragon Plus Environment

Page 21 of 43
Journal of Agricultural and Food Chemistry
21
1
3
15
4
4
4
4
4
4
4
4
4
4
4
4
4
3
53 about 30 to 35 ꢁg/kg during further storage. But, spiking the juice with [ C N]ꢀLꢀ
13
15
54 cysteine only resulted in lower concentrations of the corresponding [ C N]ꢀ2ꢀHTꢀ4ꢀ
3
55 AC after 6 days of storage (Table 8). A possible reason for the lower conversion rate
56 maybe the lower amount of the available reaction partner octanal compared to the
57 higher amount of free available
L
ꢀcysteine (2.89 mg/kg), which may quickly react with
13
58 the added [ C
8
]ꢀoctanal (Table 7). In the storage trial at 37 °C, a drastic decrease of
59 the labeled as well as the unlabeled 2ꢀHTꢀ4ꢀAC following the fast formation of the
60 condensation product in the very beginning. Nevertheless, screening a 2ꢀHTꢀ4ꢀAC
61 solution at higher temperatures did neither result in octanal nor Lꢀcysteine suggesting
62 a formation of yet unknown degradation products. The results confirmed that the forꢀ
63 mation of the 2ꢀHTꢀ4ꢀAC also occurs in orange juice following the mechanism shown
64 in Figure 5. So, this reaction clearly contributes to the loss of the aromaꢀactive aldeꢀ
65 hydes during chilled storage of NFC juice.
4
4
4
4
4
4
4
4
4
4
4
4
66
Although the reaction of Lꢀcysteine with saturated as well as unsaturated aldeꢀ
18,24–26
67 hydes, ketones or steroids is longꢀknown,
in previous studies the focus was
68 mainly on structural characterization and biochemical functionality. For example, the
69 formation of steroid thiazolidines showed reactivity towards proteins and enzymes
25
70 proposing steroid hormone functionality as prosthetic groups or coꢀenzymes. Peniꢀ
71 cillin, also containing a thiazolidine moiety, aroused the interest in these heterocyclic
27
72 acids and their bioactivity and pharmaceutical properties.
73 thiazolidineꢀ4ꢀcarboxylic acid, the reaction product of formaldehyde and
74 belonging to the group of reactive carbonyl metabolites in the human body, has been
For example, 2ꢀ
Lꢀcysteine,
28
75 found to have antiꢀinflammatory, antiꢀdiabetes or antiꢀcancer properties. However,
76 to our knowledge, this study is the first report on a clear characterization of these
77 heterocyclic acids in a food.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 22 of 43
22
2
9,30
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
4
5
5
5
5
78
Baert et al.
recently monitored the amounts of some aldehydes in beer during
ꢀcysteine led to the
79 storage. The decrease in aldehyde concentrations after adding
L
80 assumption on the formation of 2ꢀsubstituted thiazolidineꢀ4ꢀcarboxylic acids. Howevꢀ
81 er, the individual condensation products in beer were not characterized on the basis
82 of synthesized reference compounds. Only 2ꢀ(furanꢀ2ꢀyl)ꢀ1,3ꢀthiazolidineꢀ4ꢀcarboxylic
83 acid was characterized by means of UPLCꢀUV in a pale lager beer in comparison to
84 a reference solution, but no quantitative data were reported.
31
85
86
Kim and Shin developed an LCꢀMS/MS method with an online derivatization of
D
ꢀcysteine for the determination of aldehydes in water and alcoholic beverages. For
87 thiazolidineꢀ4ꢀcarboxylic acid also a stable isotope dilution assay was developed usꢀ
2
31
2
88 ing [ H ]ꢀformaldehyde for the synthesis of the labeled standard. In our study, new
89 stable isotope dilution assays were developed for the simultaneous quantitation of
90 four different 2ꢀalkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids using individually laꢀ
+
91 beled internal standards. Monitoring not only the [M+H] mass traces, but also charꢀ
92 acteristic mass transitions at an optimized ionization energy, the analysis was made
93 specific and yielded very precise results, even in a complex matrix like orange juice.
94
95 introducing free
96 of beer during storage.
In beer, aldehydes contribute to a staling or aging offꢀflavor during storage. Hence,
Lꢀcysteine into beer was suggested to positively influence the aroma
29,30
32,33
Naim et al.
had already proposed a similar approach
97 to inhibit the generation of offꢀflavor compounds during storage of orange juice. They
98 found a significantly lower formation of pꢀvinylguaiacol and 2,5ꢀdimethylꢀ4ꢀhydroxyꢀ
99 3(2H)ꢀfuranone, next to a decreased degradation of ascorbic acid after adding Lꢀ
00 cysteine to orange juice. However, the effect of a Lꢀcysteine administration on other
01 key aroma compounds in orange juice has not been investigated until now. As aldeꢀ
02 hydes were confirmed to have a positive impact on the desirable aroma of fresh orꢀ
11
03 ange juice, this wellꢀmeant approach of adding Lꢀcysteine to stabilize the aroma of
ACS Paragon Plus Environment

Page 23 of 43
Journal of Agricultural and Food Chemistry
23
5
04 orange juice could backfire and result in losses of key aroma compounds, and conꢀ
05 sequently the loss of freshness in the juice aroma profile.
5
5
5
5
5
5
5
5
5
5
5
5
06
The results show that an oxidation of aromaꢀactive aldehydes to the corresponding
07 acids also occurs during chilled storage of NFC orange juice. But, in addition the forꢀ
08 mation of 2ꢀalkylsubstituted thiazolidineꢀ4ꢀcarboxylic acids was characterized as a
09 new pathway leading to losses of four saturated aldehydes during juice storage. The
3
2,33
10 previously suggested addition of Lꢀcysteine to stabilize the juice
was not conꢀ
11 firmed as a useful idea for juice stabilization, because 2ꢀATꢀ4ꢀCAs are immediately
12 formed, four of which were identified for the first time in orange juice. The results clarꢀ
13 ified a new puzzle piece in the degradation of aromaꢀactive saturated aldehydes durꢀ
14 ing storage of orange juice. Results of studies on the stability and degradation pathꢀ
15 ways of 1ꢀpentenꢀ3ꢀone, recently identified also as important indicator of juice freshꢀ
16 ness, will be published separately.
517 AUTHOR INFORMATIONS
5
5
5
5
5
5
5
5
5
5
5
18 Corresponding Author
19 Email: peter.schieberle@ch.tum.de
20 Author Contributions
#
21
I.S. and V.M. contributed equally to this work.
22 Funding
23 We thank Tropicana Products Inc./ PepsiCo Global R & D for funding this project.
24 Notes
25 The authors declare no competing financial interest.
26 Please note that Deutsche Forschungsanstalt für Lebensmittelchemie was recently
27 renamed as LeibnizꢀInstitute for Food Systems Biology at the Technical University
28 Munich.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 24 of 43
24
529 ACKNOWLEDGMENTS.
530 We are grateful to S. Kaviani and I. Otte for excellent technical assistance with the
531 LCꢀMS measurements and thanks to C. Dawid for her help with the UPLCꢀTOF/MS
532 measurements. We also thankfully acknowledge J. Kreissl for the skillful acquisition
533 of the NMR spectra.
ACS Paragon Plus Environment

Page 25 of 43
Journal of Agricultural and Food Chemistry
25
534 REFERENCES
535 (1) PerezꢀCacho, P.R.; Rousseff, R. Processing and storage effects on orange juice
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
36
aroma: a review. J. Agric. Food Chem. 2008, 56, 9785–9796.
37 (2) Kirchner, J. G.; Miller, J. M. Volatile waterꢀsoluble and oil constituents of Valencia
38
39
40
orange juice. J. Agric. Food Chem. 1957, 5, 283–291.
(3) Tatum, J. H.; Nagy, S.; Berry, R. E. Degradation products formed in canned sinꢀ
gleꢀstrength orange juice during storage. J. Food Sci. 1975, 40, 707–709.
41 (4) Averbeck, M.; Schieberle, P. Characterisation of the key aroma compounds in a
42
43
freshly reconstituted orange juice from concentrate. Eur. Food Res. Technol.
2009, 229, 611–622.
44 (5) Averbeck, M.; Schieberle, P. Influence of different storage conditions on changes
45
46
in the key aroma compounds of orange juice reconstituted from concentrate.
Eur. Food Res. Technol. 2011, 232, 129–142.
47 (6) Moshonas, M. G.; Shaw, P. E. Flavor evaluation of fresh and aseptic packaged
48
49
orange juices. In Frontiers of Flavor, Charalambous, G., Ed.; Elsevier Science
Publisher B. V.: Amsterdam, the Netherlands; 1988, pp 133–145.
50 (7) Moshonas, M. G.; Shaw, P. E. Changes in composition of volatile components in
51
52
aseptically packaged orange juice during storage. J. Agric. Food Chem. 1989,
37, 157–161.
53 (8) Farnworth, E. R.; Lagacé, M.; Couture, R.; Yaylayan, V.; Stewart, B. Thermal proꢀ
54
55
cessing, storage conditions and the composition and physical properties of orꢀ
ange juice. Food Res. Int. 2001, 34, 25–30.
56 (9) Petersen, M. A.; Tønder, D.; Poll, L. Comparison of normal and accelerated storꢀ
57
58
age of commercial orange juice – changes in flavour and content of volatile
compounds. Food Qual. Prefer. 1988, 9, 43–51.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 26 of 43
26
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
5
59 (10) Sellami, I.; Mall, V.; Schieberle, P. Changes in the key odorants and aroma proꢀ
60
61
files of Hamlin and Valencia orange juices not from concentrate (NFC) during
chilled storage. J. Agric. Food Chem. 2018, 66, 7428–7440.
62 (11) Mall, V.; Sellami, I.; Schieberle, P. The Importance of Aldehydes to the Fresh
63
64
65
Aroma of Orange Juice. In Flavour Science: Proceedings of the XIV Weurman
Flavour Research Symposium, Taylor, A. J., Mottram, D. S., Eds.; Context
Products Ltd: Packington, UK; 2015, pp 507 – 510.
66 (12) Hofmann, T.; Münch, P.; Schieberle, P. Quantitative model studies on the forꢀ
67
68
mation of aromaꢀactive aldehydes and acids by Streckerꢀtype reactions. J.
Agric. Food Chem. 2000, 48, 434–440.
69 (13) Hofmann, T.; Schieberle, P. Evaluation of the key odorants in a thermally treated
70
71
solution of ribose and cysteine by aroma extract dilution techniques. J. Agric.
Food Chem. 1995, 43, 2187–2194.
72 (14) Blekas, G.; Guth, H. Evaluation and quantification of potent odorants of Greek
73 virgin olive oils. Dev. Food Sci. 1995, 37, 419–427.
74 (15) Kerscher, R. Objectivation of species specific aroma differences in heatꢀ
75
76
processed meat (in German). Ph.D. Thesis, Technical University of Munich,
Germany, 2000.
77 (16) Gröhnke, G. Decoding the aroma of balsamic vinegar and Parmigiano Reggiano
78
79
cheese by concepts of molecular sensory science. Ph.D. Thesis, Technical Uniꢀ
versity of Munich, Germany, Germany, 2010.
80 (17) Rota, V. Characterization of key aroma compounds in raw and cooked sheep
81
82
meat by means of a structureꢀresponse relationship concept (in German). Ph.D.
Thesis, Technical University of Munich, Germany, Germany, 2003.
83 (18) Schmolka, I.; Spoerri, P. Thiazolidine chemistry. II. The preparation of 2ꢀ
84
substituted thiazolidineꢀ4ꢀcarboxylic acids. J. Org. Chem. 1957, 8, 943–946.
ACS Paragon Plus Environment

Page 27 of 43
Journal of Agricultural and Food Chemistry
27
585 (19) Reinbold, J.; Rychlik, M.; Asam, S.; Wieser, H.; Köhler, P. Concentrations of total
5
5
5
5
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
86
glutathione and cysteine in wheat flour as affected by sulfur deficiency and corꢀ
87
relation to quality parameters. J. Agric. Food Chem. 2008, 56, 6844–6850.
88 (20) Frank, O.; Kreissl, J. K.; Daschner, A.; Hofmann, T. Accurate determination of
1
89
90
reference materials and natural isolates by means of quantitative H NMR specꢀ
troscopy. J. Agric. Food Chem. 2014, 62, 2506–2515.
91 (21) Loury, M. Possible mechanisms of autoxidative rancidity. Lipids, 1972, 7, 671ꢀ
92 675.
93 (22) Palamand, S. R.; Dieckmann, R. H. Autoxidation of nꢀhexenal. Identification and
94
95
flavor properties of some products of autoxidation. J. Agric. Food Chem. 1974,
22, 503–506.
96 (23) Schieberle, P.; Grosch, W. Model experiments about the formation of volatile
97
98
99
carbonyl compounds. J. Am. Oil Chem. Soc. 1981, 58, 602–607.
(24) Schubert, M.P. Compounds of thiol acids with aldehydes. J. Biol. Chem. 1936,
114, 341–350.
00 (25) Lieberman, S. Steroid thiazolidines, their possible biochemical significance. Ex-
01
02
03
perientia 1946, 2, 411–412.
(26) Esterbauer, H.; Ertl, A.; Scholz, N. The reaction of cysteine with
aldehydes. Tetrahedron. 1976, 32, 285–298.
α,βꢀunsaturated
04 (27) Soloway, H.; Kipnis, F.; Ornfelt, J.; Spoerri, P. E. 2ꢀSubstitutedꢀthiazolidineꢀ4ꢀ
05 carboxylic acids. J. Am. Chem. Soc., 1948, 70, 1667–1668.
06 (28) Liu, J.; Chan, W. Quantification of thiazolidineꢀ4ꢀcarboxylic acid in toxicantꢀ
07
08
09
exposed cells by isotopeꢀdilution liquid chromatographyꢀmass spectrometry reꢀ
veals an intrinsic antagonistic response to oxidative stressꢀinduced toxicity.
Chem. Res. Toxicol. 2015, 28, 394–400.
ACS Paragon Plus Environment

Journal of Agricultural and Food Chemistry
Page 28 of 43
28
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
10 (29) Baert, J. J.; De Clippeleer, J; De Cooman, L.; Aerts, G. Exploring the binding
11
12
behavior of beer staling aldehydes in model systems. J. Am. Soc. Of Brew.
Chem. 2015, 73, 100–108.
13 (30) Baert, J. J.; De Clippeleer, J; JaskulaꢀGoiris, B.; Van Opstaele, F.; De Rouck, G.;
14
15
16
Aerts, G.; De Cooman, L. Further elucidation of beer flavor instability: the potenꢀ
tial role of cysteineꢀbound aldehydes. J. Am. Soc. Brew. Chem. 2015, 73, 243–
252.
17 (31) Kim, H.; Shin, H. Simple derivatization of aldehydes with
D
ꢀcysteine and their
determination in beverages by liquid chromatographyꢀtandem mass spectromeꢀ
try. Anal. Chim. Acta 2011, 702, 225–232.
18
19
20 (32) Naim, M.; Wainish, S.; Zehavi, U.; Peleg, H.: Rouseff, R. L.; Nagy, S. Inhibition
21
22
23
by thiol compounds of offꢀflavor formation in stored orange juice. 1. Effect of
cysteine and Nꢀacetylꢀ ꢀcysteine on 2,5ꢀdimethylꢀ4ꢀhydroxyꢀ3(2H)ꢀfuranone
formation. J. Agric. Food Chem. 1993, 41, 1355–1358.
Lꢀ
L
24 (33) Naim, M.; Zuker, I.; Zehavi, U.; Rouseff, R. L. Inhibition by thiol compounds of
25
26
27
offꢀflavor formation in stored orange juice. 2. Effect of Lꢀcysteine and NꢀacetylꢀLꢀ
cysteine on pꢀvinylguaiacol formation. J. Agric. Food Chem. 1993, 41, 1359–
1361.
ACS Paragon Plus Environment

Page 29 of 43
Journal of Agricultural and Food Chemistry
29
628 FIGURE CAPTIONS
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
6
29
30 Figure 1. Chemical structures of 2ꢀpentylthiazolidineꢀ4ꢀcarboxylic acid (1),
31 2ꢀheptylthiazolidineꢀ4ꢀcarboxylic acid (2), 2ꢀoctylthiazolidineꢀ4ꢀcarboxylic acid (3) and
32 2ꢀnonylthiazolidineꢀ4ꢀcarboxylic acid (4).
33
34 Figure 2. Proposed radical oxidation of octanal to the corresponding octanoic acid
35 via an intermediate peroxyacid.
36
2,3
37 Figure 3. gsꢀCOSY ( JH,H correlation) spectrum of 2ꢀheptylthiazolidineꢀ4ꢀcarboxylic
38 acid.
39
40 Figure 4. LCꢀMS/MS run showing the specific mass transitions of 2ꢀ
13
8
41 heptylthiazolidineꢀ4ꢀcarboxylic acid and the respective internal standard [ C ]ꢀ2ꢀ
42 heptylthiazolidineꢀ4ꢀcarboxylic acid in an authentic orange juice sample.
43
44 Figure 5. Proposed mechanism of the formation of 2ꢀsubstituted thiazolidineꢀ4ꢀ
45 caboxylic acids.
ACS Paragon Plus Environment