Model studies on the pattern of volatiles generated in mixtures of amino acids, lipid-oxidation-derived aldehydes, and glucose

Article abstract of DOI:10.1021/jf104091p

The development of flavor and browning in thermally treated foods results mainly from the Maillard reaction and lipid degradation but also from the interactions between both reaction pathways. To study these interactions, we analyzed the volatile compounds resulting from model reactions of lysine or glycine with aldehydes originating from lipid oxidation [hexanal, (E)-2-hexenal, or (2E,4E)-decadienal] in the presence and absence of glucose. The main reaction products identified in these model mixtures were carbonyl compounds, resulting essentially from amino-acid-catalyzed aldol condensation reactions. Several 2-alkylfurans were detected as well. Only a few azaheterocyclic compounds were identified, in particular 5-butyl-2-propylpyridine from (E)-2-hexenal model systems and 2-pentylpyridine from (2E,4E)-decadienal model reactions. Although few reaction products were found resulting from the condensation of an amino acid with a lipid-derived aldehyde, the amino acid plays an important role in catalyzing the degradation and further reaction of these carbonyl compounds. These results suggest that amino-acid-induced degradations and further reactions of lipid oxidation products may be of considerable importance in thermally processed foods.

Full text of DOI:10.1021/jf104091p

ARTICLE
pubs.acs.org/JAFC
Model Studies on the Pattern of Volatiles Generated in Mixtures
of Amino Acids, Lipid-Oxidation-Derived Aldehydes, and Glucose
An Adams,† Vaida Kitryte_,†,‡ Rimantas Venskutonis,‡ and Norbert De Kimpe*,†
†Department of Sustainable Organic Chemistry and Technology, Faculty of Bioscience Engineering, Ghent University, Coupure links
653, B-9000 Ghent, Belgium
‡Department of Food Technology, Kaunas University of Technology, Radvilenu pl. 19, Kaunas LT-3028, Lithuania
ABSTRACT: The development of flavor and browning in thermally treated foods results mainly from the Maillard reaction and
lipid degradation but also from the interactions between both reaction pathways. To study these interactions, we analyzed the
volatile compounds resulting from model reactions of lysine or glycine with aldehydes originating from lipid oxidation [hexanal, (E)-
2-hexenal, or (2E,4E)-decadienal] in the presence and absence of glucose. The main reaction products identified in these model
mixtures were carbonyl compounds, resulting essentially from amino-acid-catalyzed aldol condensation reactions. Several
2-alkylfurans were detected as well. Only a few azaheterocyclic compounds were identified, in particular 5-butyl-2-propylpyridine
from (E)-2-hexenal model systems and 2-pentylpyridine from (2E,4E)-decadienal model reactions. Although few reaction products
were found resulting from the condensation of an amino acid with a lipid-derived aldehyde, the amino acid plays an important role
in catalyzing the degradation and further reaction of these carbonyl compounds. These results suggest that amino-acid-
induced degradations and further reactions of lipid oxidation products may be of considerable importance in thermally processed
foods.
KEYWORDS: Aroma volatiles, Maillard reaction, lipid oxidation, lysine, glycine, hexanal, (E)-2-hexenal, (2E,4E)-decadienal
’ INTRODUCTION
unsaturated aldehydes decreased (almost 100 times), the authors
concluded that secondary reactions occurred between glycine
and those aldehydes. Similar compounds were detected in the
headspace of soybean oil heated with glycine. In this case, also the
addition of cane sugar was investigated, which affected flavor
formation.9
To identify the origin of several of these volatiles, model
reactions were performed using amino acids (lysine and glycine)
and long-chain aldehydes with a varying degree of unsaturation,
in particular, hexanal, (E)-2-hexenal, and (2E,4E)-decadienal,
known as common lipid oxidation products.10 In addition, the
influence of the addition of glucose to the model systems on the
volatiles produced was studied, as well as the interaction between
the different constituents. The high-molecular-weight polycon-
densation products of the same model reactions have been
characterized before by a combination of different analytical
methods. These results have been discussed elsewhere.10
Two of the most important reactions responsible for the
development of flavor and color in processed foods are lipid
degradation, on the one hand, and the Maillard reaction,
occurring between a carbohydrate and an amino compound,
on the other hand. The interaction between these two reactions
has been the subject of several studies, because the course of both
reactions can be modified by the reactants, intermediates, and
products of the other.1,2
In this respect, it was recently shown that secondary lipid
oxidation products having two oxygenated functions are able to
degrade amino acids by a Strecker-type mechanism.3,4 This
reaction results in the formation of the Strecker aldehyde of
the corresponding amino acid and an unsaturated imine, which
cyclizes to 2-alkylpyridine (in the case of 4,5-epoxy-2-alkenal) or
2-pentyl-1H-pyrrole (in the case of 4-hydroxy-2-nonenal).
Most studies on the interaction between the Maillard reaction
and lipid degradation focused on meat flavor and, thus, the
reactions between carbonyl compounds and cysteine or its
degradation products, hydrogen sulfide and ammonia.1 Studies
on the thermal interaction of amino acids with triglycerides
reported N-substituted amides and nitriles as the major
products,5,6 while 2-pentylpyridine was the major product from
the reactions of valine with linoleic acid and its esters.7 Macku
and Shibamoto8 investigated the volatile components from
heated corn oil with and without glycine. Several nitrogen-
containing compounds were formed from these interactions,
such as nitriles, pyridines, and pyrroles. As a result of the addition
of glycine (0.5-10%) to the oil, the amount of released volatiles
decreased to almost 50%. Especially because the amount of
’ MATERIALS AND METHODS
Materials. Glycine (>99%), L(þ)-lysine monohydrochloride (>99%),
and hexanal (>98%) were obtained from Sigma-Aldrich (Bornem,
Belgium). (2E,4E)-Decadienal (95%) was obtained from Acros Organ-
ics (Geel, Belgium). (E)-2-Hexenal (>99%) was obtained from Janssen
Chimica (Beerse, Belgium). D(þ)-Glucose (>99%, anhydrous) was
obtained from Fluka (Bornem, Belgium). Solvents were of analytical
Received: October 19, 2010
Accepted: December 23, 2010
Revised:
December 20, 2010
Published: January 25, 2011
r
2011 American Chemical Society
1449
dx.doi.org/10.1021/jf104091p J. Agric. Food Chem. 2011, 59, 1449–1456
|

Journal of Agricultural and Food Chemistry
ARTICLE
Table 1. Volatile Compounds (Percentage of Total GC Peak Area as Measured by SPME-GC-MS) Identified after Thermal
Interaction of Glycine (Gly) or Lysine (Lys) with Hexanal (Hexa) or (E)-2-Hexenal (Hexe) in the Absence and Presence of
Glucose
125 °C, 20 min
Lys Gly
125 °C, 120 min
Lys Gly
125 °C, 20 min, þglucose
LRIa
LRI
(lit)
Lys
Gly
Lys
Gly
Lys
Lys
Gly
Gly
compound
acetaldehydeb
ethanolb
1-pentened
pentaneb
acetic acidb
butanalb
2-butanoneb
(exp)
Hexa Hexe Hexa Hexe Hexa Hexe Hexa Hexe
Hexa
Hexe
Hexa
Hexa
c
<600
<600
<600
<600
<600
<600
600
699
703
708
709
710
728
744
761
767
771
786
794
801
804
817
828
851
857
860
861
866
886
894
899
900
903
905
909
912
925
929
930
948
953
983
999
-
-
-
-
-
-
-
-
-
0.64
-
-
-
0.30
-
-
-
-
-
-
-
0.42
-
0.62
-
1.36
-
3.33
-
1.17
-
-
-
0.53
-
0.60
-
-
3.03
-
-
-
-
-
2.73
-
1.53
-
600e
596f
592g
694h
703h
704i
0.14
-
-
-
-
-
-
-
-
-
-
-
0.02
0.71
-
0.26
0.03
0.09
-
0.02
-
1.46
-
1.84
-
-
-
0.04
0.09
0.34
-
0.24
-
-
-
-
1-penten-3-ol
2-pentanoneb
pentanalb
1-ethoxy-1-butened
2-ethylfuranb
2-methyl-3-pentanoned
(E)-2-methyl-2-butenal
2-ethylbutanalb
3-methylpentanald
1-pentanolb
3-hexanoneb
2-hexanoneb
-
0.18
-
0.25
-
0.25
-
0.10
-
0.08
-
0.17
-
0.31
-
0.61
-
0.88
-
0.96
-
0.11
-
0.25
0.44
-
0.15
-
0.67
-
-
-
0.15
-
0.19
-
-
-
-
-
-
-
-
-
-
0.85
-
1.44
702i
-
3.65
-
1.41
-
6.96
-
0.71
-
1.44
-
2.76
-
0.30
-
-
-
-
-
-
-
0.41
-
0.73
745h
-
0.36
-
0.21
-
-
-
-
-
-
-
0.14
0.52
0.08
1.31
-
-
1.81
-
1.32
-
0.15
1.49
-
1.07
-
1.60
-
-
-
-
-
-
-
-
-
-
-
772g
788g
792g
-
2.35
-
4.71
0.65
0.67
-
-
2.65
-
2.99
-
3.30
-
-
-
0.68
-
1.50
-
1.19
-
0.45
-
0.64
-
0.17
800i 78.28
0.80
0.24
1.07
-
2.62
0.41
-
0.60
-
13.43
0.36
4.37
-
hexanalb
4.32
0.26
2.94
-
4.98
0.22 22.72
0.61
27.24
14.92
31.59
ethyl butanoateb
2-ethyl-2(E)-butenalb
pentyl formated
(E)-2-hexenalb
ethylbenzene
2-methylhexanald
2-methylenehexanald
(Z)-2-hexenolb
3-heptanone
804i
-
-
-
-
0.56
0.40
-
-
0.37
0.27
-
-
0.22
-
-
-
6.27
-
-
-
2.27
-
-
-
-
0.92
-
0.98
-
0.20
846i
864j
-
0.20
0.14
-
-
-
-
-
-
-
-
35.34
-
1.87
0.23
-
-
-
-
-
0.42
-
-
0.39
-
-
0.15
-
-
0.11
-
-
0.17
-
-
0.13
-
-
0.36
-
-
-
-
-
-
-
1.12
0.19
1.79
0.78
-
859i
888h
904k
900
896i
898i
-
0.96
0.57
-
-
-
-
2.10
1.90
-
-
2.93
1.81
-
-
0.45
-
-
-
0.41
-
-
-
0.36
-
pentanoic acid
nonaneb
-
-
-
0.42
-
-
-
0.36
0.25
-
0.16
-
-
-
0.35
-
0.13
-
0.41
-
propyl butanoate
ethyl pentanoateb
1,1-diethoxybutaneb
2-butylfuranb
2-methyl-2-cyclopenten-1-oneb
methyl hexanoateb
2-ethyl-2-pentenald
3-methyl-4-heptanoned
3-ethyl-2-methyl-1,3-hexadiened
2-ethylhexanal
hexanoic acidb
2-ethyl-2(Z)-hexenald
decaneb
2-ethyl-2(E)-hexenalb
2-hexen-1-yl acetate
5-ethyl-2(5H)-furanoned
butylbenzene
-
-
-
-
-
-
-
-
-
-
0.15
0.10
16.82
-
-
-
-
0.51
0.22
38.34
-
-
0.38
0.18
45.69
-
-
0.12
-
0.25
0.18
20.87
-
-
-
-
-
-
0.08
-
-
-
-
-
-
11.17
-
893h
912h
916f
-
-
-
-
-
-
0.10
-
-
-
-
0.55
-
-
0.05
-
-
0.18
-
0.16
-
-
0.16
-
-
3.47
-
-
-
0.74
-
-
1.17
-
0.37
0.19
2.28
0.46
-
-
-
-
-
-
-
-
-
-
-
-
-
0.16
0.23
2.21
-
-
-
-
0.28
-
-
-
-
-
-
-
-
-
0.33
-
-
-
-
0.58
-
956j
0.13
-
-
0.29
-
-
-
-
967i 12.01
0.17
-
15.89
-
14.91
-
8.41
-
-
3.55
-
-
-
-
-
-
-
-
26.38
-
4.63
-
2.75
-
-
0.57
-
-
-
-
-
-
-
1.29
-
1001 1000
1006
1017 1010i
-
-
-
-
-
-
-
1.53
1.32
-
4.12
1.37
0.14
-
5.39
-
-
-
-
-
0.70
2.25
-
0.11
2.05
-
-
-
-
-
-
0.34
0.86
1.44
-
-
-
-
-
-
-
-
-
1037
1056 1067g
1057 1057f
0.59
-
0.48
-
-
(E)-2-octenalb
0.10
-
1450
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
Table 1. Continued
125 °C, 20 min
Lys Gly
Hexa Hexe Hexa Hexe Hexa Hexe Hexa Hexe
125 °C, 120 min
125 °C, 20 min, þglucose
LRIa
LRI
(lit)
Lys
Gly
Lys
Lys Gly
Gly
Lys
Lys
Gly
Gly
compound
(exp)
Hexa
Hexe
Hexa
Hexa
4-propyl-1,3-benzenediold
2-butyl-2-pentenald
4-undecened
1080
1085
1090
1091 1094l
1101 1100
1165 1171m
1176
-
1.59
-
0.35
-
2.02
-
-
-
0.15
-
0.18
-
0.87
-
2.12
-
-
-
-
-
-
-
-
0.12
-
0.52
-
4.34
-
1.39
-
0.16
-
2.34
-
5-undecene
undecaneb
0.16
-
0.69
-
5.71
-
1.83
-
0.21
-
3.15
-
-
-
-
-
-
-
-
-
-
-
-
-
1.51
-
-
-
-
-
ethylbenzaldehyde
5-decanoned
2-butyl-2-hexenald
5-undecanoned
4-ethyl-(2E,4E,6E)-decatrienald
(Z)-2-butyl-2-octenalb
(E)-2-butyl-2-octenalb
4-ethyl-(2E,4Z,6E)-decatrienald
5-butyl-2-propylpyridined
total GC peak area (ꢀ108)
unidentified (%)
-
-
-
-
-
-
-
-
-
0.14
-
-
-
4.03
-
-
0.37
-
-
-
-
-
-
1178
-
8.83
-
-
2.69
-
-
0.43
-
-
0.18
-
-
0.97
-
1276
-
3.83
-
1.78
-
-
-
-
-
1361
-
11.69
6.31
4.18
1.17
0.21
-
5.89
0.41
-
10.59
1.31
-
1374 1378n
1388 1385o
1398
3.35
0.26
-
3.78 78.08
3.23 24.78
2.38 56.38
38.53
2.73
-
37.10
6.08
-
0.15
4.07
0.04
0.48
-
1.68
-
-
-
-
-
-
-
1.73
-
3.32
-
1400
-
-
-
-
-
2.28
2.06
-
2.83
1.77
-
-
5.58
4.32
3.55
2.80
2.32
2.13
2.69
2.86
2.43
2.71
2.89 34.11
1.88 32.31 10.87 19.07 22.57 32.08
5.99
19.53
1.15
26.01
a Comparison of calculated (calc) linear retention index (LRI) on the AT-5MS stationary phase to literature data (lit). b Identification confirmed by a
comparison to the authentic reference substance. c Not detected. d Tentatively identified. e From ref 27. f From ref 28. g From ref 8. h From ref 29. i From
ref 30. j From ref 32. k From ref 33. l From ref 10. m From ref 34. n From ref 31. o From ref 35.
grade. All standards used for compound identification were of high
purity (>95%), with 2,5-dimethylfuran, 2-acetylfuran, (E)-2-hexenol,
(E)-2-hexenoic acid, (E)-2-octenal, heptanal, and methyl hexanoate
from Sigma-Aldrich, crotonaldehyde, valeraldehyde, benzaldehyde,
(E)-2-heptenal, (E)-2-decenal, 1-butanol, 1-pentanol, 1,1-diethoxybu-
tane, 2-butanone, 2-pentanone, 2-hexanone, 3-hexanone, 2-methyl-
2-cyclopenten-1-one, 2-methylfuran, and 2-ethylfuran from Acros
Organics, butyraldehyde and hexanoic acid from Janssen Chimica,
acetaldehyde, 2-heptanone, 2-furaldehyde, ethyl butyrate, and ethyl
valerate from Fluka, 2-ethylcrotonaldehyde, 2-ethylbutyraldehyde,
and 2-ethyl-2-hexenal (mixture of E/Z isomers) from TCI Europe
(Zwijndrecht, Belgium), and 2-propylfuran, 2-butylfuran, 2-pentylfuran,
and 2-hexylfuran (97%) from Alfa Aesar (Karlsruhe, Germany).
Model Reactions. For this purpose, 5 mmol of amino acid and 2.5
mmol of aldehyde [hexanal or (E)-2-hexenal] with or without 5 mmol of
D(þ)-glucose were well-mixed in a 20 mL solid-phase microextraction
(SPME) vial, covered with a magnetic crimp cap with a septum (natural
rubber red orange/Teflon; 55° Shore A; 1.3 mm, Supelco, Bornem,
Belgium) and placed in an oven (equipped with a fan, Memmert,
Schwabach, Germany), which had been preheated to and stabilized at
125 °C. Because of the high reactivity of (2E,4E)-decadienal-containing
model mixtures, twice lower amounts of initial reactants were used. The
amino acid/carbonyl compound mixtures were heated for exactly 20 or
120 min. After heating, the vials with brown-colored reaction products
were removed from the oven and placed in a desiccator to cool to room
temperature. Most model reactions were performed in duplicate
(Table 1), while (2E,4E)-decadienal-containing model reactions were
performed in triplicate (Table 2).
the solvent, the E/Z isomer mixture of 2-butyl-2-octenal was obtained in
67% yield. The gas chromatography (GC) (E/Z, 4:96; linear retention
index (LRI, AT5-MS), 1384:1370), mass spectrometry (MS), and
nuclear magnetic resonance (NMR) data correspond well with the
literature, as reported by Robert et al.12
SPME extraction (5 min) and desorption (2 min) were performed
automatically by means of a MPS-2 autosampler (Gerstel, M€ulheim an
der Ruhr, Germany). A divinylbenzene/Carboxen/polydimethylsilox-
ane (DVB/Carboxen/PDMS) fiber (Supelco) was used.
Gas Chromatography-Mass Spectrometry (GC-MS). For
the analysis of the SPME extracts, a Hewlett-Packard 6890 GC Plus
coupled with a HP 5973 mass selective detector (MSD-quadrupole
type), equipped with a CIS-4 programmed temperature vaporization
(PTV) injector (Gerstel) and an AT5-MS capillary column (30 ꢀ 0.25
mm inner diameter; coating thickness of 0.25 μm) was used. Working
conditions were as follows: injector, 250 °C; transfer line to MSD,
260 °C; oven temperature, start at 40 °C, hold for 2 min, programmed
from 40 to 120 °C at 4 °C min-1, programmed from 120 to 240 °C at
30 °C min-1, and hold for 2 min; split ratio, 20:1; carrier gas (He),
1.2 mL min-1; ionization electron impact (EI), 70 eV; and scanned
acquisition parameters, m/z 40-200 (0-10 min), m/z 40-300 (10-20
min), and m/z 40-400 (>20 min). Volatile substances were identified by
a comparison of their mass spectra and retention times to those of
reference substances and a comparison to the Wiley (6th edition) and
National Institute of Standards and Technology (NIST) Mass Spectral
Library (version 1.6d) and particular literature data, if available. When
only MS data were available, identifications were indicated as tentative.
Synthesis of 2-Butyl-2-octenal. 2-Butyl-2-octenal was pre-
pared as follows. A three-neck flask (50 mL) was fitted with a sealed
stirrer, an efficient reflux condenser, and a dropping funnel containing
freshly redistilled hexanal (30 mmol). A NaOH solution (1 M, 10 mL)
was placed in the flask and heated to 80 °C. The aldehyde was added
rapidly under vigorous stirring, after which the reaction mixture was
refluxed for 1 h. After cooling to room temperature, the organic layer was
extracted with diethylether and dried with MgSO4. After evaporation of
’ RESULTS AND DISCUSSION
The volatiles detected in the headspace of heated model
mixtures of lysine or glycine with hexanal, (E)-2-hexenal, or
(2E,4E)-decadienal (2:1) were investigated in the presence and
absence of glucose. The model mixtures were heated in closed
vials at 125 °C for 20 min. In the absence of glucose, a high
amount of starting material remained, and therefore, the amino
1451
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
Table 2. Volatile Compounds (Percentage of Total GC Peak Area as Measured by SPME-GC-MS) after Heating (125 °C) of
Glycine (Gly) or Lysine (Lys) with (2E,4E)-Decadienal (Dcdnl) in the Absence and Presence of Glucose
125 °C, 20 min
125 °C, 120 min
125 °C, 20 min, þglucose
125 °C, 120 min, þglucose
LRIa LRI
compound
acetaldehydeb
ethanolb
(calc) (lit) Lys/Dcdnl Gly/Dcdnl Lys/Dcdnl Gly/Dcdnl Lys/Dcdnl
Gly/Dcdnl
Lys/Dcdnl
Gly/Dcdnl
<600
<600
<600
<600
<600
<600
602 610g
603 612g
611 606h
650 651g
660
0.34
0.52
0.23
tad
-
1.79
0.72
3.32
1.18
0.23
-
1.38
3.92
1.73
0.73
1.01
1.04
4.75
3.18
2.38
-
0.58
1.53
0.83
0.61
-
2.28
1.20
1.15
pentaneb
1.07
0.79
0.94
ethyl formatec
methyl acetatec
butanalb
0.35
-
0.60
0.48
e
-
tad
4.33
-
-
596f
tad
0.17
-
0.54
-
-
-
2.22
1.21
-
acetic acidb
-
-
-
-
0.45
-
1.04
1.32
1.47
-
2-methylfuranb
ethyl acetateb
2-butenalb (E/Z)
1-butanolb
-
-
-
-
-
-
-
-
-
-
-
-
0.21
-
tad
0.49
-
0.19
0.14
0.14
3.26
1.21
-
-
-
tad
0.19
0.91
-
-
tad
-
-
0.10
0.70
1.00
-
-
0.20
4.50
1.15
tad
-
1-heptene
pentanalb
2,5-dimethylfuranb
1-methyl-1H-pyrrole
tolueneb
1-pentanolb
2-hexanoneb
hexanalb
2-ethyl-2(E)-butenalb
methyl pentanoate
1,3-octadienec
furfuralb
687
694i
0.25
0.68
2.51
0.55
0.20
-
699 704h
710 711g
2.66
-
-
-
732
760
765
788
750i
-
tad
-
tad
36.55
1.82
1.53
4.52
0.11
0.90
-
-
-
4.46
0.12
tad
7.33
-
33.15
29.41
1.32
1.26
11.81
-
15.26
0.67
tad
34.53
-
9.84
27.30
1.30
0.56
6.39
-
6.89
0.48
0.54
28.20
-
772i
792i
1.08
0.30
0.19
-
800 800h
17.16
12.36
817
-
-
822 821h
823
-
-
-
-
-
-
-
-
-
-
-
-
-
12.45
-
833 830h
-
-
-
-
-
0.29
-
ethylbenzene
1,3-dimethylbenzenec
2-heptanoneb
2-butylfuranb
3(E)-noneneb
heptanalb
2-acetylfuranb
1,2-nonadienec
2(E)-heptenalb
benzaldehydeb
2-pentylfuranb
2-ethyl-2(E)-hexenalb
octanal
856
866
889
889
897
902 908k
912 912g
924
956 947h
959 952h
989 993k
1004
1004 998h
1053 1067i
1058 1049h
1083
860j
893i
900j
tad
tad
-
0.46
0.36
0.33
-
0.47
0.18
2.33
-
0.28
0.18
-
0.13
0.45
0.31
-
0.37
0.30
-
0.31
0.11
0.82
-
0.32
tad
0.15
-
-
3.20
0.54
0.39
-
0.65
0.14
0.81
-
0.24
tad
2.32
0.27
0.59
0.36
4.82
-
1.90
tad
1.23
0.21
1.10
-
0.36
0.50
0.45
-
0.76
0.50
-
-
0.21
tad
6.12
0.68
0.22
-
1.99
2.16
-
6.40
-
0.36
0.10
4.91
0.65
0.69
-
0.34
-
-
5.96
5.97
3.23
1.25
-
0.51
0.20
0.61
-
7.26
0.25
2.41
0.56
-
8.23
0.61
-
0.18
0.28
0.64
-
tad
-
0.23
0.77
-
butylbenzene
2(E)-octenalb
2,4-decadienolc
2-hexylfuranb
1,1-diethoxyethanec
4-undecenec
0.22
0.58
-
0.33
2.61
-
0.94
-
0.44
2.33
-
0.13
1.93
-
-
0.33
-
0.22
0.47
-
-
-
0.10
0.45
-
-
1089
0.18
-
-
-
-
-
0.18
-
1090
0.37
0.57
-
0.27
-
0.10
1091
1093 1094j
-
-
0.21
0.21
0.10
-
-
0.21
0.21
0.22
-
0.14
0.14
-
5-undecene
-
tad
-
-
-
tad
-
2-hex-1-enylfuranc (E/Z)
pentylbenzene
2-hex-1-enylfuranc (E/Z)
1,3,5-undecatriene
4(Z)-decenal
2-pentylpyridine
decanal
1135
-
tad
-
-
1154 1152j
1157
1184 1183j
1195 1193h
1196 1205i
1205 1201h
-
0.10
-
-
tad
0.19
-
-
0.61
-
-
0.12
-
-
0.66
-
-
0.17
tad
-
-
-
9.53
-
-
-
5.23
-
7.62
2.05
-
2.66
-
1.04
-
0.53
-
0.62
-
-
-
0.12
-
0.20
-
1452
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
Table 2. Continued
125 °C, 20 min
125 °C, 120 min
125 °C, 20 min, þglucose
125 °C, 120 min, þglucose
LRIa LRI
compound
2(E)-decenalb
2-pentyl-2-cyclopentene-1-one 1263 1288h
(calc) (lit) Lys/Dcdnl Gly/Dcdnl Lys/Dcdnl Gly/Dcdnl Lys/Dcdnl
Gly/Dcdnl
Lys/Dcdnl
Gly/Dcdnl
1261 1260h
0.33
-
-
-
-
-
0.23
-
0.33
-
-
-
-
0.26
-
2.28
-
-
-
(2E,4Z)-decadienal
(2E,4E)-decadienalb
1293 1292h
1320 1319h
1325
6.93
54.85
-
-
2.71
18.68
0.33
0.13
0.11
0.65
-
5.71
40.37
0.50
0.64
0.48
3.92
-
-
0.63
0.72
1.68
1.44
11.05
tad
0.13
0.28
1.83
2.36
10.64
tad
0.19
0.33
1.23
1.79
9.24
tad
-
-
2,4-decadienal (Z,E/Z)c
2,6-dodecadienalc (E/Z)
2,6-dodecadienalc (E/Z)
2,6-dodecadienalc (E/Z)
2-butyl-2(Z)-octenalg
total GC peak area (ꢀ108)
unidentified (%)
-
-
1334
0.42
0.32
2.75
-
2.21
2.87
14.89
tad
3.10
7.67
2.28
2.65
11.21
-
1335
1354
1370
7.29
2.65
2.73
2.41
5.51
5.09
2.56
2.18
10.66
8.2
11.93
4.79
3.96
4.4
a Comparison of calculated (calc) linear retention index (LRI) on the AT-5MS stationary phase to literature data (lit). b Identification confirmed by
a comparison to the authentic reference substance. c Tentatively identified. d Trace amounts (<0.1 %). e Not detected. f From ref 28. g From ref 36.
h From ref 30. i From ref 8. j From ref 10. k From ref 29.
acid/lipid oxidation product model mixtures were also heated for
120 min. After cooling to room temperature, the headspace of the
heated vials was directly analyzed by means of SPME-GC-MS.
The volatiles identified in the different model reactions are
presented in Tables 1 and 2. The results are mean values of
two (Table 1) or three (Table 2) replicated experiments. Con-
cerning the variability on the results obtained, the relative
standard deviation ranges from 0 to 110%. The high standard
deviations were noted for compounds present in low amounts.
For the most important compounds (abundance >5%), the
relative standard deviation usually remained lower than 10%,
apart from a few exceptions (such as hexanoic acid, which gives
an irregular chromatographic peak).
The amounts of volatiles detected in the headspace of amino
acid/lipid oxidation product mixtures heated at 125 °C for 120
min (total GC peak area of around 1.8-2.3 ꢀ 108) were lower
than for the mixtures heated at the same temperature for a
shorter time of 20 min (total GC peak area of 2.8-5.6 ꢀ 108).
This was mainly due to the decrease of the volatile starting
material hexanal with longer heating times. In the lysine/hexanal
model system, for example, 78.3% of the total GC peak area was
due to hexanal remaining after 20 min of reaction. After 120 min
of reaction time, this percentage had decreased to 22.7%. Much
lower amounts of (E)-2-hexenal remained after heating
(Table 1). This can be explained by the R,β unsaturation of
the carbonyl group, which makes (E)-2-hexenal more reactive.
Only in the presence of glucose, a relatively high amount of (E)-
2-hexenal remained unreacted after heating, especially in the
lysine/(E)-2-hexenal/glucose model mixture (35.3%), indicating
that lysine reacts preferably with glucose or its degradation
products, as compared to (E)-2-hexenal. In the case of
(2E,4E)-decadienal, considerable amounts of unreacted starting
material were found in lysine model systems heated for 20 min
(54.9%). Upon prolonged heating, most of the starting material
disappeared (Table 1).
Results on the formation and characterization of these polycon-
densation products of high molecular weight have been discussed
elsewhere.10
The most important part of the volatiles identified were
carbonyl compounds, resulting from the starting aldehydes and
aldol condensation reactions thereof. The hexanal aldol con-
densation products (Z)-2-butyl-2-octenal and (E)-2-butyl-2-oc-
tenal dominated in the headspace of most hexanal-containing
heated samples (Table 1). The identification and elution order of
both isomers was established based on the synthesized reference
compounds.12 For (Z)-2-butyl-2-octenal, a low odor threshold
(20 μg/L water) and an attractive ham-like, meaty, savory odor
have been reported.12
These aldol condensation products were not detected when
hexanal was heated alone. Hence, the aldol condensation reac-
tion of this aldehyde is strongly catalyzed by the presence of an
amino acid. It has been reported in the literature that amino acids
as well as small peptides can catalyze aldol reactions.13-15 To
explain the stereoselectivity obtained in specific cases, a mecha-
nism was proposed in which the ketone donor reacts with the
amino acid with the formation of an enamine. Afterward, the
acceptor aldehyde reacts with the chiral enamine and, upon
hydrolysis, an enantiomerically enriched aldol product can be
generated.15 In the case of the basic amino acids lysine and
arginine (when unprotonated), however, reported yields were
low and no diastereoselectivity was obtained, indicating that only
a general base aldolization mechanism took place in these cases.16
The conversion of hexanal to 2-butyl-2-octenal in our study
occurred much more readily in the presence of glycine than in the
presence of lysine, which can be explained by the lower nucleo-
philicity of the primary amino group of lysine monohydrochlor-
ide (pKa = 8.95), as compared to glycine (pKa = 9.78). Thus, after
20 min of reaction, lysine/hexanal model mixtures contained
78.3% hexanal and only 3.6% 2-butyl-2-octenal, while glycine/
hexanal model mixtures contained 5.0% hexanal and 82.2% of the
aldol product, 2-butyl-2-octenal (sum of E and Z isomers as
measured by the GC peak area). In the presence of glucose,
however, reaction mixtures of both amino acids showed equal
reactivity toward aldol condensation, because the headspace
extracts of both glycine and lysine reaction mixtures contained
As the prepared reaction mixtures were heated in hermetically
closed vials, a decreasing amount of low-molecular volatiles
results from reactions incorporating low-molecular-weight vola-
tiles into nonvolatile reaction products and, eventually, in the
skeleton of the final colored polycondensation reaction products.
1453
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
about 30% hexanal and a little more than 40% 2-butyl-2-octenal
after 20 min of reaction.
Different isomers of the aldol condensation product of (E)-2-
hexenal, 4-ethyl-2,4,6-decatrienal, were identified in the (E)-2-
hexenal model mixtures. Both 4-ethyl-(2E,4E,6E)-decatrienal
and 4-ethyl-(2E,4Z,6E)-decatrienal were tentatively identified,
but correct integration of the 4-ethyl-(2E,4Z,6E)-decatrienal
peak was not always possible because of the overlap with
5-butyl-2-propylpyridine. 4-Ethyl-2,4,6-decatrienal has only been
reported once in the literature, as a result of the self-condensation
of (E)-2-hexenal by means of diethylamine.17
Figure 1. Formation of 2-alkyl-2-cyclopenten-1-one from the corre-
sponding alkadienal.
Acetaldehyde and butanal are produced by the retro-aldol
reaction of hydrated 2-hexenal (that is, 3-hydroxyhexanal).
Because acetaldehyde is difficult to detect because of its high
reactivity and low molecular weight, only trace amounts were
found in the lysine/(E)-2-hexenal/glucose model system. Buta-
nal, on the other hand, was detected in all (E)-2-hexenal model
reactions, albeit in low amounts ranging from 0.02 to 1.8%.
Surprisingly, large amounts of butanal were present as its diethyl
acetal, 1,1-diethoxybutane, which dominated in the headspace
of the amino acid/(E)-2-hexenal mixtures heated for 120 min
(38.3 and 45.7% for the lysine and glycine model systems,
respectively). This compound can result from the nucleophilic
addition of ethanol across butanal. Ethanol itself was detected,
particularly in the glucose/(E)-2-hexenal model reactions. These
model systems also contained considerable amounts of 1,1-
diethoxybutane, in addition to 1-ethoxy-1-butene. The latter
results from the elimination of ethanol from 1,1-diethoxybutane,
because it is less plausible that it is formed from the addition of
ethanol to butanal followed by dehydration. Additionally, butanal
is subjected to aldol condensation reactions, resulting in the
formation of, on one hand, 2-ethyl-2-butenal (0.3-6.3%) from
butanal and acetaldehyde and, on the other hand, 2-ethyl-2-
hexenal (0.9-31.8%, E and Z isomers together) from self-
condensation of butanal.
Upon hydration and retro-aldol reaction, (2E,4E)-decadienal
yields hexanal and 2-butenal via 5-hydroxy-2-decenal and
2-octenal and acetaldehyde via 3-hydroxy-4-decenal (Table 2).
Correspondingly, 2-octenal itself yields hexanal and acetaldehyde
upon hydration and retro-aldolization. All of these aldehydes
were present in the headspace of heated (2E,4E)-decadienal
model systems. It must be noted that the aldehydes mentioned
can also be formed from (2E,4E)-decadienal by means of
autoxidation, which will in part account for their presence as
well.18 Substantial amounts of dodecadienal were discovered in
all of the model systems analyzed. The formation of dodecadienal
in these model systems can be explained by aldol reactions of
acetaldehyde with 2-decenal, resulting in 2,4-dodecadienal, or
with 4-decenal, resulting in 2,6-dodecadienal. On the basis of the
mass spectrum and the higher amounts of 4-decenal detected as
compared to 2-decenal, this compound was tentatively identified
as 2,6-dodecadienal.
Considerable amounts of hexanal were also found in (E)-2-
hexenal model systems, especially in the (E)-2-hexenal/glucose
models. According to Rizzi,19 the reaction of an R,β-unsaturated
carbonyl compound with an amino acid can result in the
formation of the reduced carbonyl compound, following a
Strecker-type degradation. In this reaction, imine formation is
followed by decarboxylation and subsequent formation of the
Strecker aldehyde and 1-hexenylamine. By means of isomeriza-
tion to the corresponding imine and hydrolysis, 1-hexenylamine
is converted into hexanal. Detection of considerable amounts of
hexanal in (E)-2-hexenal model systems clearly demonstrates the
occurrence of this reaction, which has been hypothesized but
rarely observed and only for compounds with an aromatic
substituent.19
A considerable part of hexanal was oxidized to hexanoic acid, a
compound that is not very sensitive to headspace SPME extrac-
tion; detection thus indicates the presence of considerable
amounts of acid. It was an important constituent of the headspace
of lysine/hexanal model mixtures (12.0-15.9%) and, to a lesser
extent, of the glycine/hexanal/glucose reaction mixture (8.4%)
as well.
In addition, some methyl ketones, which are important
volatiles in the flavor deterioration of dairy products, were
detected (2-butanone, 2-pentanone, and 2-hexanone). Macku
and Shibamoto8 previously showed in a model mixture of corn oil
and glycine that secondary reactions occurred between glycine
and the volatile aldehydes formed from the heated corn oil. A
positive linear correlation was found between methyl ketone
formation and the amount of glycine added. The formation of
methyl ketones can be partly explained by the principle of chain
elongation, first described by Yaylayan and Keyhani20 in a
glucose/alanine Maillard model system. Following this mecha-
nism, an aldehyde is converted into the corresponding methyl
ketone by reaction with glycine. Thus, for example, 2-pentanone
can be formed from the reaction of butanal with glycine,
2-hexanone from pentanal and glycine, and 2-heptanone from
hexanal and glycine. Because 2-heptanone was, however, only
detected in the glycine/(2E,4E)-decadienal model systems, the
mechanism of chain elongation is probably of minor importance
in these model systems.
Intramolecular cyclization of (2E,4E)-decadienal yields 2-pen-
tyl-2-cyclopenten-1-one (up to 2.3%), according to the mechan-
ism shown in Figure 1. In this hypothesized mechanism, the
alkadienal is protonated, inducing ring closure after E to Z
isomerization. Subsequently, rearrangement of the double bonds
leads to the corresponding 2-alkyl-2-cyclopenten-1-one. The
intermediacy of 2-alkyl-3-cyclopenten-1-one and its transforma-
tion to the thermodynamically more stable 2-alkyl-2-cyclopen-
ten-1-one has been described before.21 Correspondingly, 2-methyl-
2-cyclopenten-1-one was detected in some of the (E)-2-hexenal
model systems (up to 0.5%). In this case, dehydrogenation of
Apparently, a large part of the volatiles identified can be
explained by means of aldol-type reactions occurring among the
carbonyl compounds involved. The same aldol-type reactions
have been identified as the predominant mechanisms respon-
sible for the formation of brown-colored high-molecular-weight
polycondensation products.10 The larger part of the com-
pounds that remain unidentified are compounds of intermedi-
ate molecular weight and, on the basis of the mass spectrum, are
assumed to be aldol products of the many carbonyl compounds
involved.
1454
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
these volatiles was favored at the final stages of the reaction. They
probably result from enhanced degradation and recombination
of fragments of the lipid-derived aldehydes. Among these, 4- and
5-undecene were repeatedly found in hexanal model systems.
Also benzaldehyde needs to be mentioned as a known autoxida-
tion product of (2E,4E)-decadienal.26
When these lipid oxidation aldehydes [hexanal, (2E)-hexenal,
and (2E,4E)-decadienal] were heated alone, no browning oc-
curred and no significant degradation took place. The results
presented thus demonstrate that, although few reaction products
were found resulting from the actual condensation of an amino
acid with a lipid oxidation aldehyde, the amino acid plays a very
significantrolein catalyzingthe degradation and further reaction of
these carbonyl compounds. It seems that aldol condensation
reactions of lipid oxidation aldehydes catalyzed by an amino acid
are very important in the overall formation of aroma volatiles in
such reaction mixtures. Thus, in addition to the degradation of
amino acids induced by lipid oxidation products, which has been
reported beforeby Hidalgoand co-workers,3,4 amino-acid-induced
degradations and further reactions of lipid oxidation products may
be of considerable importance in thermally processed foods as
well. In addition, most of the volatiles identified in this study have
been reported as volatile constituents of thermally processed lipid-
rich food products,21,29,32 indicating that the mechanisms ex-
plained play a role in the formation of food volatiles.
The amino-acid-induced degradation of lipid oxidation pro-
ducts is comparable to the amino-acid-induced carbohydrate
degradation as commonly described in Maillard-type reactions.
The heating conditions applied (125 °C and 20 min) are also not
severe enough to induce caramelization or (detectable) forma-
tion of furans upon heating of glucose alone. At these heating
conditions, also the extent of the Maillard reaction in the
corresponding glycine/glucose and lysine/glucose model reac-
tions was rather limited. Typical Maillard flavor compounds, in
particular, 2-methylfuran, 2,5-dimethylfuran, furfural, and 2-acet-
ylfuran, were only detected after prolonged heating for 120 min
(data not shown). No nitrogen-containing compounds were
detected under these conditions. In comparison to this, a higher
degree of browning and volatile formation was observed in the
model systems containing lipid oxidation products, thus illus-
trating their potential importance in real food systems, which, of
course, consist of a complex mixture of several compounds.
Comparison of the amino acid/lipid oxidation product model
systems in the absence and presence of glucose reveals differ-
ences mainly in the relative composition of the headspace extracts.
Specific compounds indicative of an interaction between the
three components in these amino acid/aldehyde/glucose model
reactions were not found. Again, especially aldol products of the
carbonyl compounds involved were predominant in the head-
space extract, indicating the importance of these reactions in
thermally processed foods.
Figure 2. Formation of 5-butyl-2-propylpyridine from (E)-2-hexenal
and ammonia.
2-hexenal to 2,4-hexadienal is required in a first step. 2-Pentyl-2-
cyclopenten-1-one is one of the volatile constituents detected in
used frying oils,22 as were most of the compounds identified in
Tables 1 and 2. In addition, intramolecular cyclization of
(2E,4E)-decadienal toward a six-membered ring results in the
formation of butylbenzene, which was detected in the headspace
fraction of all (2E,4E)-decadienal model reaction mixtures.
Oxidation of (E)-2-hexenal in the allylic position followed by
cyclization results in the formation of 2-ethylfuran, the main
oxygen-containing heterocyclic compound among the (E)-2-
hexenal headspace constituents. According to the same reaction
mechanism, 2-butylfuran is formed from 2-octenal, 2-pentylfuran
from 2-nonenal, 2-hexylfuran from 2-decenal, 5-ethyl-2(5H)-
furanone from hexanoic acid, and 2-hex-1-enylfuran from 2,4-
decadienal. Especially in the (2E,4E)-decadienal model systems,
these 2-alkylfurans were quite important headspace constituents
and have also been identified before in 2,4-decadienal model
mixtures.23 A detailed study of the formation of 2-alkylfurans in
model reactions of lipid-derived R,β-unsaturated aldehydes with
amino acids will be reported elsewhere.
Only one nitrogen-containing aromatic volatile compound
was identified in the headspace of the (E)-2-hexenal model
reactions, in particular 5-butyl-2-propylpyridine. A mechanism
of formation of this compound, resulting from the condensation
of two molecules of (E)-2-hexenal with ammonia, is proposed in
Figure 2. In this hypothesized reaction mechanism, the imine
formed from (E)-2-hexenal with ammonia performs a Michael-
type addition reaction toward a second molecule of (E)-2-
hexenal. After rearrangement of the double bonds, cyclization
occurs, which yields 5-butyl-2-propylpyridine after elimination of
water. The release of ammonia from amino acids upon heating
has been described. Model studies have shown that especially
glutamine released considerable amounts of ammonia upon
moderate heating (110 °C), while glycine and lysine yielded
low amounts of ammonia upon more elevated heating at 180 °C
for 2 h in aqueous solution (pH 8).24 In our model experiments,
5-butyl-2-propylpyridine was only detected after prolonged
heating (120 min, 125 °C, and dry). Its formation might still
increase with stronger heating conditions.
2-Pentylpyridine was detected in the headspace of most
(2E,4E)-decadienal model systems in the absence of glucose.
This compound was also detected after thermal degradation of
(2E,4E)-decadienal/amino acid polycondensation products, in a
relatively large quantity compared to the other volatile azahe-
terocycles detected.10 This very potent odorant is responsible for
grassy aroma defects in soybean products.11 Its formation results
from the reaction of (2E,4E)-decadienal with either free ammo-
nia (released from the amino acids) or the R-amino group bound
in amino acids.25
’ AUTHOR INFORMATION
Corresponding Author
*Telephone: þ32-9-264-59-51. Fax: þ32-9-264-62-43. E-mail:
norbert.dekimpe@ugent.be.
Funding Sources
Besides the compounds described above, several volatile
alkanes, alkenes, and alcohols were detected in the headspace
of the amino acid/carbonyl compound samples. The release of
The authors are indebted to the Research Foundation-Flanders
(Belgium) (FWO-Vlaanderen) for a postdoctoral fellowship to
An Adams.
1455
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456

Journal of Agricultural and Food Chemistry
ARTICLE
’ REFERENCES
(1) Whitfield, F. B. Volatiles from interactions of Maillard reactions
and lipids. Crit. Rev. Food Sci. Nutr. 1992, 31, 1–58.
(2) Zamora, R.; Hidalgo, F. J. Coordinate contribution of lipid
oxidation and Maillard reaction to the nonenzymatic food browning.
Crit. Rev. Food Sci. Nutr. 2005, 45, 49–59.
(3) Hidalgo, F. J.; Zamora, R. Strecker-type degradation produced
by the lipid oxidation products 4,5-epoxy-2-alkenals. J. Agric. Food Chem.
2004, 52, 7126–7131.
(4) Hidalgo, F. J.; Gallardo, E.; Zamora, R. Strecker-type degradation
of phenylalanine by 4-hydroxy-2-nonenal in model systems. J. Agric.
Food Chem. 2005, 53, 10254–10259.
(5) Lien, Y. C.; Nawar, W. W. Thermal interaction of amino acids
and triglycerides. Valine and tricaproin. J. Food Sci. 1974, 39, 917–919.
(6) Breitbart, D.; Nawar, W. W. Thermal interaction of lysine and
triglycerides. J. Agric. Food Chem. 1981, 29, 1194–1196.
(7) Henderson, S. K.; Nawar, W. W. Thermal interaction of linoleic
acid and its esters with valine. J. Am. Oil Chem. Soc. 1981, 58, 632–635.
(8) Macku, C.; Shibamoto, T. Headspace volatile compounds
formed from heated corn oil and corn oil with glycine. J. Agric. Food
Chem. 1991, 39, 1265–1269.
(9) Wang, Z.; Wu, Q. H.; Wu, C.-M. Effect of sucrose on the volatile
compounds produced from heated glycine and soybean oil. Food Sci.
Technol. Res. 1999, 5, 393–397.
(10) Adams, A.; Kitryte, V.; Venskutonis, R.; De Kimpe, N. Forma-
tion and characterization of melanoidin-like polycondensation products
from amino acids and lipid oxidation products. Food Chem. 2009, 115,
904–911.
(25) Kim, Y.-S.; Hartman, T. G.; Ho, C.-T. Formation of 2-pentyl-
pyridine from the thermal interaction of amino acids and 2,4-decadienal.
J. Agric. Food Chem. 1996, 44, 3906–3908.
(26) Bruechert, L. J.; Zhang, Y.; Huang, T. C.; Hartman, T. G.;
Rosen, R. T.; Ho, C.-T. Contribution of lipids to volatile generation in
extruded corn-based model systems. J. Food Sci. 1988, 53, 1444–1447.
(27) Schnermann, P.; Schieberle, P. Evaluation of key odorants in
milk chocolate and cocoa mass by aroma extract dilution analysis.
J. Agric. Food Chem. 1997, 45, 867–872.
(28) Fuhrmann, E.; Grosch, W. Character impact odorants of the
apple cultivars Elstar and Cox orange. Nahrung 2002, 46, 187–193.
(29) Elmore, J. S.; Mottram, D. S.; Enser, M.; Wood, J. D. The effects
of diet and breed on the volatile compounds of cooked lamb. Meat Sci.
2000, 55, 149–159.
(30) Adams, R. P. Identification of Essential Oil Components by Gas
Chromatography/Mass Spectrometry, 4th ed.; Allured Publishing Cor-
poration: Carol Stream, IL, 2006.
(31) Pino, J. A.; Mesa, J.; Munoz, Y.; Marti, M. P.; Marbot, R. Volatile
components from mango (Mangifera indica L.) cultivars. J. Agric. Food
Chem. 2005, 53, 2213–2223.
(32) Hierro, E.; de la Hoz, L.; Ordꢀo~nez, J. A. Headspace volatile
compounds from salted and occasionally smoked dried meats (cecinas)
as affected by animal species. Food Chem. 2004, 85, 649–657.
(33) Machiels, D.; Van Ruth, S. M.; Posthumus, M. A.; Istasse, L. Gas
chromatography-olfactometry analysis of the volatile compounds of
two commercial Irish beef meats. Talanta 2003, 60, 755–764.
(34) Triqui, R.; Reineccius, G. A. Changes in flavor profiles with
ripening of anchovy (Engraulis encrasicholus). J. Agric. Food Chem. 1995,
43, 1883–1889.
(11) Belitz, H.-D.; Grosch, W.; Schieberle, P. Food Chemistry, 3rd
revised ed.; Springer-Verlag: Heidelberg, Germany, 2004.
(12) Robert, F.; Hꢀeritier, J.; Quiquerez, J.; Simian, H.; Blank, I.
Synthesis and sensorial properties of 2-alkylalk-2-enals and
3-(acetylthio)-2-alkyl alkanals. J. Agric. Food Chem. 2004, 52, 3525–3529.
(13) Pokorny, J.; Janitz, W.; Viden, I.; Velisek, J.; Valentova, H.;
Davidek, J. Reaction of oxidized lipids with protein: Part 14. Aldolization
reactions of lower alkanes in the presence of nonlipidic substances.
Nahrung 1987, 1, 63–70.
(35) Bredie, W. L. P.; Mottram, D. S.; Guy, R. C. E. Effect of
temperature and pH on the generation of flavour volatiles in extrusion
cooking of wheat flour. J. Agric. Food Chem. 2002, 50, 1118–1125.
(36) Adams, A.; Abbaspour Tehrani, K.; Kersiene, M.; Venskutonis,
R.; De Kimpe, N. Characterization of model melanoidins by the thermal
degradation profile. J. Agric. Food Chem. 2003, 51, 4338–4343.
(14) List, B.; Lerner, R. A.; Barbas, C. F. Proline-catalyzed direct
asymmetric aldol reactions. J. Am. Oil Chem. Soc. 2000, 122, 2395–2396.
(15) Cordova, A.; Zou, W.; Dziedzic, P.; Ibrahem, I.; Reyes, E.; Xu,
Y. Direct asymmetric intermolecular aldol reactions catalyzed by amino
acids and small peptides. Chem.—Eur. J. 2006, 12, 5385–5397.
(16) Lombardo, M.; Easwar, S.; Pasi, F.; Trombini, C.; Dhavale,
D. D. Protonated arginine and lysine as catalysts for the direct asym-
metric aldol reaction in ionic liquids. Tetrahedron 2008, 64, 9203–
9207.
(17) Tadahiko, K.; Akikazu, H.; Yuzo, I.; Minoru, O. Leaf alcohol.
XVIII. Condensation of leaf aldehyde by means of diethylamine. Agric.
Biol. Chem. 1969, 33, 409–413.
(18) Josephson, D. B.; Lindsay, R. C. Retro-aldol related degrada-
tions of 2,4-decadienal in the development of staling flavors in fried
foods. J. Food Sci. 1989, 52, 1186–1190.
(19) Rizzi, G. The Strecker degradation of amino acids: Newer
avenues for flavor formation. Food Rev. Int. 2008, 24, 416–435.
(20) Yaylayan, V.; Keyhani, A. Origin of 2,3-pentanedione and 2,3-
butanedione in glucose/L-alanine Maillard model systems. J. Agric. Food
Chem. 1999, 47, 3280–3284.
(21) Elia, G. R.; Childs, R. F.; Shaw, G. S. Thermal cyclizations of
protonated polyunsaturated aldehydes. Can. J. Chem. 1992, 70, 2065–
2069.
(22) Takeoka, G.; Perrino, C., Jr.; Buttery, R. Volatile constituents of
used frying oils. J. Agric. Food Chem. 1996, 44, 654–660.
(23) Zhang, Y.; Ho, C.-T. Volatile compounds formed from thermal
interaction of 2,4-decadienal with cysteine and glutathione. J. Agric. Food
Chem. 1989, 37, 1016–1020.
(24) Sohn, M.; Ho, C.-T. Ammonia generation during thermal
degradation of amino acids. J. Agric. Food Chem. 1995, 43, 3001–3003.
1456
dx.doi.org/10.1021/jf104091p |J. Agric. Food Chem. 2011, 59, 1449–1456