Intermediate role of α-keto acids in the formation of Strecker aldehydes

Article abstract of DOI:10.1016/j.foodchem.2013.03.068

The ability of α-keto acids to covert amino acids into Strecker aldehydes was investigated in an attempt to both identify new pathways for Strecker degradation, and analyse the role of α-keto acids as intermediate compounds in the formation of Strecker aldehydes by oxidised lipids. The results obtained indicated that phenylalanine was converted into phenylacetaldehyde to a significant extent by all α-keto acids assayed; glyoxylic acid being the most reactive α-keto acid for this reaction. It has been proposed that the reaction occurs by formation of an imine between the keto group of the α-keto acid, and the amino group of the amino acid. This then undergoes an electronic rearrangement with the loss of carbon dioxide to produce a new imine. This final imine is the origin of both the Strecker aldehyde and the amino acid from which the α-keto acid is derived. When glycine was incubated in the presence of 4,5-epoxy-2-decenal, the amino acid was converted into glyoxylic acid, and this α-keto acid was then able to convert phenylalanine into phenylacetaldehyde. All these results suggest that Strecker aldehydes can be produced by amino acid degradation initiated by different reactive carbonyl compounds, included those coming from amino acids and proteins. In addition, α-keto acids may act as intermediates for the Strecker degradation of amino acids by oxidised lipids.

Full text of DOI:10.1016/j.foodchem.2013.03.068

Food Chemistry 141 (2013) 1140–1146
Contents lists available at SciVerse ScienceDirect
Food Chemistry
journal homepage: www.elsevier.com/locate/foodchem
Intermediate role of
a
-keto acids in the formation of Strecker aldehydes
⇑
Francisco J. Hidalgo, Rosa M. Delgado, Rosario Zamora
Instituto de la Grasa, Consejo Superior de Investigaciones Científicas, Avenida Padre García Tejero 4, 41012-Seville, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
The ability of
to both identify new pathways for Strecker degradation, and analyse the role of
diate compounds in the formation of Strecker aldehydes by oxidised lipids. The results obtained indicated
that phenylalanine was converted into phenylacetaldehyde to a significant extent by all -keto acids
assayed; glyoxylic acid being the most reactive -keto acid for this reaction. It has been proposed that
the reaction occurs by formation of an imine between the keto group of the -keto acid, and the amino
a
-keto acids to covert amino acids into Strecker aldehydes was investigated in an attempt
Received 21 December 2012
Received in revised form 11 February 2013
Accepted 18 March 2013
a-keto acids as interme-
a
Available online 26 March 2013
a
a
Keywords:
group of the amino acid. This then undergoes an electronic rearrangement with the loss of carbon dioxide
to produce a new imine. This final imine is the origin of both the Strecker aldehyde and the amino acid
Amino acid degradation
Amino acid formation
Carbonyl–amine reactions
Flavours
from which the
a-keto acid is derived. When glycine was incubated in the presence of 4,5-epoxy-2-dec-
enal, the amino acid was converted into glyoxylic acid, and this
a
-keto acid was then able to convert
phenylalanine into phenylacetaldehyde. All these results suggest that Strecker aldehydes can be pro-
a-Keto acids
Maillard reaction
Strecker aldehydes
duced by amino acid degradation initiated by different reactive carbonyl compounds, included those
coming from amino acids and proteins. In addition,
degradation of amino acids by oxidised lipids.
a
-keto acids may act as intermediates for the Strecker
Ó 2013 Elsevier Ltd. All rights reserved.
1. Introduction
acids, which is considered one of the main pathways by which fla-
vour compounds are produced in the Maillard reaction (Kerler &
Wimkel, 2002). In fact, Strecker degradation has been suggested
to play a switching role in directing the Maillard reaction towards
aromagenic rather than chromogenic pathways (Yaylayan, 2003).
However, Strecker degradation of amino acids is not only produced
Flavour is one of the main sensory properties by which consum-
ers select, accept and ingest foods. Many compounds responsible
for the flavour of foods occur naturally, but many others are pro-
duced during food processing. At present, many different pathways
for flavour formation have been described. Among them, Maillard
reaction and lipid oxidation are widely accepted as major routes
for flavour formation during food processing and storage (Heredia,
Peinado, Rosa, Andres, & Escriche, 2012; Wang, Yang, & Song, 2012;
Yarnpakdee, Benjakul, Nalinanon, & Kristinsson, 2012).
The Maillard reaction has been traditionally related to the for-
mation of pleasant flavours during food processing, and lipid oxi-
dation is a well-known pathway responsible for the appearance
of rancid flavours during storage. However, this separation is not
so clear because many flavours are produced simultaneously in
both routes, by analogous reaction mechanisms (Zamora & Hidal-
go, 2005). One example of this is the Strecker degradation of amino
by a-dicarbonyl or hydroxycarbonyl compounds formed by carbo-
hydrate degradation (Granvogl, Beksan, & Schieberle, 2012). Other
reactive carbonyls, such as lipid-derived reactive carbonyls (Hidal-
go & Zamora, 2004) or the quinones produced by polyphenol oxi-
dation (Rizzi, 2008), have also been described as significant
producers of Strecker aldehydes. All these results suggest that
the production of flavour compounds may sometimes be the result
of different interrelated reactions, that can even work in a co-ordi-
nated way.
In an attempt to identify new pathways by which Strecker alde-
hydes are produced, and how different reactions can be co-ordi-
nated to produce flavour compounds, this study investigated the
role of lipid oxidation as an inducer of the formation of
a-keto
acids, and how these amino acid derived carbonyl compounds
were able to produce the Strecker degradation of other amino
acids.
Abbreviations: AlaKA, pyruvic acid; CysKA, mercaptopyruvic acid; GluKA,
ketoglutaric acid; GlyKA, glyoxylic acid; IleKA, 3-methyl-2-oxovaleric acid; IS,
internal standard; LeuKA, 4-methyl-2-oxovaleric acid; MetKA, -keto- -meth-
ylthiobutyric acid; OHP, 6-oxoheptanoic acid; OVA, 2-oxovaleric acid; PheKA,
phenylpyruvic acid; TyrKA, 4-hydroxyphenylpyruvic acid; ValKA, 3-methyl-2-
oxobutyric acid.
a-
a
c
a-Keto acids have been shown to be formed in foods either
enzymatically by the action of transaminases (Gao et al., 1997),
or chemically by reaction of amino acids with lipid-derived reac-
tive carbonyls (Zamora, Navarro, Gallardo, & Hidalgo, 2006). In
⇑
Corresponding author. Tel.: +34 954 611 550; fax: +34 954 616 790.
E-mail address: rzamora@ig.csic.es (R. Zamora).
0308-8146/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.foodchem.2013.03.068

F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
1141
addition, these compounds have been shown to contribute signifi-
2.4. Phenylacetaldehyde determination
cantly to the formation of food flavours by either enzymatic or
chemical decomposition (Chu & Yaylayan, 2008; Smit et al.,
2004). Furthermore, these compounds have a carbonyl group
and, thus, they produce a Strecker degradation product when re-
acted with amino acids (Guerra & Yaylayan, 2010). In addition,
they have been shown to convert asparagine into acrylamide, a
reaction in which the presence of reactive carbonyls is required
(Zamora, Delgado, & Hidalgo, 2011). Therefore, an alternative route
for production of flavour compounds, as a consequence of lipid oxi-
For phenylacetaldehyde determination, the cooled sample was
treated with sodium borohydride (ꢁ1 mg) for 10 min at room tem-
perature to convert phenylacetaldehyde into phenylethanol. Ex-
cess of sodium borohydride was eliminated by adding acetone
(300
ll). Thirty microlitres of an internal standard (a solution of
15 l of cis-3-nonenol in 25 ml of ethanol) was then added, and
l
the resulting mixture was analysed by GC–MS. The ions monitored
for the identification of the analyte, 2-phenylethanol, were
[C8H10O]+ = 122 and [C7H7]+ = 91, using m/z 122 for quantification.
The ions monitored for the identification of the internal standard
were [C9H16]+ = 124, [C7H11]+ = 95 and [C6H9]+ = 81, using m/z 81
for quantification.
dation, would be the formation of a-keto acids. These would later
produce a Strecker degradation when reacted with amino acids.
This study will first describe the Strecker degradation of amino
acids produced by
Strecker aldehydes can be produced as a consequence of lipid oxi-
dation, in which -keto acids are active intermediates.
a-keto acids, and then will investigate how
GC–MS analyses were conducted with a Hewlett–Packard 6890
GC Plus coupled with an Agilent 5973 MSD (Mass Selective Detec-
tor-Quadrupole type). In most experiments, a 30 m ꢂ 0.25 mm
a
i.d. ꢂ 0.25
lm HP5-MS capillary column was used. Working condi-
tions were as follows: carrier gas helium (1 ml/min at constant
flow); injector, 250 °C; oven temperature: from 40 (1 min) to 240
at 5 °C/min and then to 300 at 10 °C/min; transfer line to MSD,
280 °C; and ionisation EI, 70 eV.
2. Materials and methods
2.1. Materials
Quantification of phenylacetaldehyde was carried out by pre-
paring standard curves of the aldehyde in the 320 ll of the buffer,
which was submitted to the whole process described above. For
each curve, eight different concentration levels of the aldehyde
were used. Phenylacetaldehyde content was directly proportional
to the phenylethanol/internal standard area ratio (r > 0.98,
p < 0.0005). The coefficients of the variation were <10%.
4,5-Epoxy-2-decenal was prepared by epoxidation of (E,E)-2,4-
decadienal with 3-chloroperoxybenzoic acid as described previ-
ously (Zamora et al., 2006).
The commercial
acid (GlyKA, -Keto Acid derived from Glycine), pyruvic acid (Ala-
KA, -Keto Acid derived from Alanine), 3-methyl-2-oxobutyric acid
(ValKA, -Keto Acid derived from Valine), 2-oxovaleric acid (OVA),
4-methyl-2-oxovaleric acid (LeuKA, -Keto Acid derived from Leu-
cine), 3-methyl-2-oxovaleric acid (IleKA, -Keto Acid derived from
Isoleucine), -keto- -methylthiobutyric acid (MetKA, -Keto Acid
a-keto acids used in this study were: glyoxylic
a
a
a
a
2.5. Amino acid and 4,5-epoxy-2-decenal determination
a
a
c
a
For amino acid determination, the method described by Villas-
Boas, Delicato, Akesson, and Nielsen (2003) was employed, which
was slightly modified. This method also allowed the determination
of 4,5-epoxy-2-decenal. Briefly, the cooled sample was treated
derived from Methionine), mercaptopyruvic acid (CysKA,
Acid derived from Cysteine), phenylpyruvic acid (PheKA,
a
a
-Keto
-Keto
Acid derived from Phenylalanine), 4-hydroxyphenylpyruvic acid
(TyrKA,
(GluKA,
a
a
-Keto Acid derived from Tyrosine) and
-Keto Acid derived from Glutamic acid). Although 6-oxo-
-keto acid, it was also studied for
a-ketoglutaric acid
with 20
anol], 100
pyridine and 20
red for 10 s. After 30 s, other 20
l
l of the IS [a solution of 10% (v/v) octanoic acid in meth-
l of methanol, 100 l of 1 M sodium hydroxide, 35 l of
l of methyl chloroformate. The mixture was stir-
l of methyl chloroformate was
l
l
l
heptanoic acid (OHP) is not an
a
l
comparison purposes. All these chemicals, as well as other chemi-
cals employed in this study, were of the highest available analytical
grade, and were purchased from Aldrich (Milwakee, WI, USA), Sig-
ma (St. Louis, MO, USA), Fluka (Buchs, Switzerland) or Merck
(Darmstadt, Germany).
l
added and the reaction mixture was stirred for a further 10 s. Then,
the reaction mixture was maintained for 1 min at room tempera-
ture. After that, 400
ll of chloroform was added and the mixture
was shaken for 10 s. Next, 400
l
l of 50 mM potassium bicarbonate
solution was added and the mixture was shaken for 10 s. The reac-
tion mixture was then centrifuged at 2000g for 5 min to separate
the layers, and the organic layer was analysed by GC–MS.
2.2. Phenylalanine/a-keto acid reaction mixtures
By using this methodology, 4,5-epoxy-2-decenal could be
determined in its intact form, but glycine was transformed into
methyl-2-(methoxycarbonylamino) acetate, phenylalanine into
A solution of phenylalanine (20
lmol) and the a-keto acid (0–
20
l
mol) in 320 l of 0.3 M buffer (sodium citrate for pH 2.15–6,
l
sodium phosphate for pH 6–8 and pH 11–12, and sodium borate
for pH 8–10) was heated at 37–140 °C in closed test tubes for 1 h
(in most experiments). After cooling (5 min at room temperature
and 10 min at ꢀ20 °C), samples were used for either phenylacetal-
dehyde or amino acid determination.
methyl-2-(methoxycarbonylamino)-3-phenylpropanoate
and
octanoic acid into methyl octanoate. The ions monitored for the
identification of methyl 2-(methoxycarbonylamino) acetate were
[C5H9NO4]+ = 147, [C3H6NO2]+ = 88 and [C2H3O2]+ = 59, using m/z
88 for quantification. The ions monitored for the identification of
methyl-2-(methoxycarbonylamino)-3-phenylpropanoate
were
[C10H12NO2]+ = 178, [C10H10O2]+ = 162, and [C7H7]+ = 91, using m/z
162 for quantification. The ions monitored for 4,5-epoxy-2-decenal
were [C10H16O2]+ = 168, [C8H11O2]+ = 139 and [C4H4O]+ = 68, using
m/z 68 for quantification. The ions monitored for the identification
of the internal standard were [C8H15O]+ = 127, [C4H7O2]+ = 87 and
[C3H6O2]+ = 74, using m/z 74 for quantification.
GC–MS analyses were conducted with a Hewlett–Packard 6890
GC Plus coupled with an Agilent 5973 MSD (Mass Selective Detec-
tor-Quadrupole type) using the column and conditions described
above. Quantification of glycine, phenylalanine and 4,5-epoxy-2-
2.3. In situ formation and reaction of a-keto acids
A
solution of glycine (20
mol) in 320 l of 0.3 M sodium citrate buffer (pH 3) was
heated at 37–60 °C in closed test tubes for 2 days. Phenylalanine
(20 mol) was then added and the reaction was maintained at
lmol) and 4,5-epoxy-2-decenal
(20
l
l
l
37–60 °C for other 3 days. At the end of the heating time, samples
were cooled and used for either glyoxylic acid (GlyKA), phenylacet-
aldehyde, 4,5-epoxy-2-decenal, or amino acid determination.

1142
F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
decenal was carried out by preparing standard curves of the amino
acids and the lipid oxidation product in the 320 l of the buffer,
3. Results
l
which were submitted to the whole process described above. For
each curve, six different concentration levels of the amino acids
and the lipid were used. Amino acid and lipid contents were di-
rectly proportional to the amino acid/internal standard (or lipid/
internal standard) area ratios (r > 0.99, p < 0.0001). The coefficients
of variation were <10%.
3.1. Formation of phenylacetaldehyde in phenylalanine/
reaction mixtures
a-keto acid
When phenylalanine was heated alone, the formation of a small
amount of phenylacetaldehyde was observed (Fig. 1). However, the
amount of the produced Strecker aldehyde was much higher when
the amino acid was heated in the presence of
a-keto acids. Among
the different -keto acids assayed, GlyKA was the most reactive
a
2.6. GlyKA determination
compound, followed by MetKA and ValKA. On the other hand, Leu-
KA and IleKA did not produce a significant (p < 0.05) increase in the
amount of the Strecker aldehyde produced. Surprisingly, OHP
For GlyKA determination, 100
ted sequentially with 50 l of a 0.25 nmol/ml solution of 2-oxova-
leric acid (IS, internal standard) in methanol, 30 of
trifluoromethanesulfonic acid solution (3% in methanol) and
200 l of dansylhydrazine solution (2 mg/ml in methanol) (Appel-
ll of the cooled sample was trea-
l
(which is not a a-keto acid) also produced phenylacetaldehyde to
ll
a significant extent, therefore suggesting that carbonyl and carbox-
ylic groups do not need to be conjugated to react, at least for this
carbonyl compound. Because of the high reactivity of GlyKA in this
l
blad, Pontén, Jaegfeldt, Bäckström, & Irgum, 1997; Hyytiaïnen
et al., 1996). The resulting solution was kept at 25 °C for 1 h, then
diluted with 1 ml of eluent A (a 30:70 mixture of 0.2% formic acid
in acetonitrile and 4 mM ammonium acetate), and analysed by LC–
MS/MS.
Samples were analysed using an Agilent liquid chromatography
system (1200 Series) consisting of binary pump (G1312A), degas-
ser (G1379B) and autosampler (G1329A), connected to a triple-
quadrupole API 2000 mass spectrometer (Applied Biosystems)
using an electrospray ionisation interface in positive ionisation
mode (ESI+). Compounds were separated on a Zorbax Eclipse
reaction, this was the
In addition to the formation of the Strecker aldehyde of phenyl-
alanine, the reaction also converted the -keto acid into the corre-
sponding amino acid from which it derived. Thus, in the reaction
between GlyKA and phenylalanine, the formation of glycine was
observed (data not shown).
a-keto acid selected for these studies.
a
3.2. Effect of reaction conditions (pH, a-keto acid concentration,
atmosphere, time and temperature) on phenylacetaldehyde formation
in phenylalanine/GlyKA reaction mixtures.
XDB-C18 (150 mm ꢂ 4.6 mm, 5
lm) column from Agilent. As elu-
Phenylacetaldehyde was mainly produced at acidic pH values
(pH < 4) and decreased at higher pH values. Although produced
to a lower extent than at pH 2.15–3, it was similarly observed at
pH 4–9, and the amount of phenylacetaldehyde decreased linearly
ent A, a 30:70 mixture of 0.2% formic acid in acetonitrile and
4 mM ammonium acetate was used. As eluent B, a 0.2% formic acid
solution in acetonitrile was employed. The mobile phase was deliv-
ered at 0.5 ml/min in linear gradient mode: 0 min 0% B, 40 min 30%
B, 45 min 100% B, 50 min 100% B, 51 min 0% B, 60 min 0% B. Mass
spectrometric acquisition was performed by using multiple reac-
tion monitoring (MRM). The nebulizer gas (synthetic air), the cur-
tain gas (nitrogen) and the heater gas (synthetic air) were set at 45,
30, and 50 (arbitrary units), respectively. The collision gas (nitro-
gen) was set at 3 (arbitrary units). The heater gas temperature
was set at 500 °C and the electrospray capillary voltage to 5.5 kV.
The focusing potential was 370 V and the declustering potential
was 21 V. The fragment ions in MRM mode were produced by col-
lision-activated dissociation of selected precursor ions in the colli-
sion cell of the triple quadrupole, and the selected products
analysed with the second analyzer of the instrument. Three transi-
tions were acquired for the identification of the dansylhydrazone
derivative of GlyKA. These transitions were 322?236.1,
322?171.1 and 322?156.2, using the first transition for quatifica-
tion. The transition 364?236.1 was employed for the IS.
Quantification of GlyKA was carried out by preparing standard
curves of this
a-keto acid in the 320 ll of the buffer, which was
submitted to the whole process described above. For each curve,
five different concentration levels of GlyKA were used. GlyKA con-
tent was directly proportional to the GlyKA/internal standard area
ratios (r > 0.99, p < 0.0001). The coefficients of variation were <15%.
2.7. Statistical analysis
Fig. 1. Phenylacetaldehyde (PAC) formation in the Strecker degradation of phen-
ylalanine (Phe) by
-keto acids (20
acids assayed were: CO, control reaction of phenylalanine heated in the absence of
-keto acids; GlyKA, glyoxylic acid; AlaKA, pyruvic acid; ValKA, 3-methyl-2-
oxobutyric acid; OVA, 2-oxovaleric acid; LeuKA, 4-methyl-2-oxovaleric acid; IleKA,
3-methyl-2-oxovaleric acid; MetKA, -keto- -methylthiobutyric acid; CysKA,
mercaptopyruvic acid; PheKA, phenylpyruvic acid; TyrKA, 4-hydroxyphenylpyruvic
acid; GluKA, -ketoglutaric acid; and OHP, 6-oxoheptanoic acid. Bars with different
letters are significantly different (p < 0.05).
a
-keto acids. Equimolecular mixtures of phenylalanine and the
All data given are mean SD values of, at least, four indepen-
dent experiments. Statistical comparisons among different groups
were made using analysis of variance. When significant F values
were obtained, group differences were evaluated by the Tukey’s
test (Snedecor & Cochran, 1980). Statistical comparisons were car-
ried out using Origin v.7.0 (OriginLab Corporation, Northampton,
MA). The significance level is p < 0.05 unless otherwise indicated.
a
l
mol of each) were heated at 100 °C for 1 h at pH 3. The -keto
a
a
a
c
a

F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
1143
Fig. 2. Effect of: A, reaction pH; B, concentration of the a-keto acid; C, concentration of oxygen; and D, time and temperature; on the phenylacetaldehyde (PAC) produced in
phenylalanine/glyoxylic acid reaction mixtures. In panels A and C, samples were heated at 100 °C for 1 h. In panels B, C, and D, the pH was 3. The buffers assayed in panel A
were: sodium citrate (s), sodium phosphate (4), and sodium borate (5). The heating times and temperatures assayed in panel B were: 100 °C for 1 h (s) or 37 °C for 6 days
(4). The temperatures assayed in panel D were: 140 (s), 120 (4), 100 (5), and 80 °C (}).
(r = ꢀ0.98, p = 0.016) between pH 9 and pH 12 (Fig. 2A). Because
previous studies of Strecker degradation of amino acids by lipid
oxidation products were carried out at pH 3, the rest of this study
was also carried out at this pH for comparison.
The amount of phenylacetaldehyde increased linearly
(r = 0.998, p < 0.0001) as a function of the amount of GlyKA present
in the mixture (Fig. 2B), therefore confirming that the formation of
in Fig. 2B, phenylacetaldehyde was also produced at 37 °C and the
amount of phenylacetaldehyde increased linearly as a function of
the amount of GlyKA present in the reaction mixture (r = 0.996,
p = 0.0003). In fact, the amount of phenylacetaldehyde produced
at 37 °C for 6 days was only 30% lower than the amount of phenyl-
acetaldehyde produced at 100 °C for 1 h.
the Strecker aldehyde was a consequence of the ability of the
keto acid to degrade the amino acid.
a-
3.3. Formation of phenylacetaldehyde as a consequence of lipid
oxidation via an intermediate
a-keto acid
Differently to the reaction pH and the amount of
a-keto acid
present in the mixture, the composition of the reaction atmosphere
seemed to have very little effect on the amount of phenylacetalde-
hyde produced (Fig. 2C). Thus, the amount of phenylacetaldehyde
only seemed to increase slightly (19%) as a function of oxygen con-
centration. This increase was linear (r = 0.94, p = 0.006) within the
0–100% O2 range.
Phenylacetaldehyde concentration also increased linearly as a
function of time and temperature between 80 and 140 °C, and
reaction rates increased with the temperature (Fig. 2D). Reaction
rates at different temperatures were calculated by using this
equation:
To confirm that -keto acids were not required to be added, but
a
could be formed in situ and degrade surrounding amino acids, an
experiment in which two amino acids (glycine and phenylalanine)
were added sequentially was assayed (Fig. 3). At the beginning of
the experiment, only glycine and a lipid oxidation product (4,5-
epoxy-2-decenal) were added. The reaction was kept at 37 or
60 °C for 2 days (Fig. 3A/B and C/D, respectively). This resulted in
the disappearance of the lipid at 60 °C, its decrease at 37 °C
(0.23 0.04 lmol of the lipid was still present after 2 days at
37 °C in the presence of glycine) and the decrease of glycine con-
tent. The decrease of the amino acid was higher in the samples
containing the lipid oxidation product than in the samples in
which only the amino acid was heated.
½phenylacetaldehydeꢃ ¼ ½phenylacetaldehydeꢃ0 þ kt
where [phenylacetaldehyde]0 represents the intercept, k is the rate
constant and t is time. These rate constants were used in an Arrhe-
nius plot to calculate the activation energy (Ea) of phenylacetalde-
hyde formation in the phenylalanine/glyoxylic acid system. From
the slope of the obtained line (r = ꢀ0.9999, p < 0.0001), the Ea was
determined. Ea was 38.2 kJ/mol.
This decrease of glycine content was a consequence of the reac-
tion between the amino acid and epoxyalkenal, which have been
partly reviewed by Hidalgo and Zamora (2008). This reaction af-
fords different amino acid degradation products. Among them,
the formation of GlyKA was observed (Fig. 3B and D at 37 and
60 °C, respectively). As observed in Fig. 3B and D, GlyKA was
formed to a higher extent at 60 °C than at 37 °C. This is in agree-
ment with the higher loss of glycine (in comparison to the control
This relatively low Ea suggested that the reaction product
should be produced even at relatively low temperatures. As shown

1144
F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
Fig. 3. Changes observed in glycine (s,d), phenylalanine (4,N), glyoxylic acid (},ꢀ) and phenylacetaldehyde (5,.) contents as a function of heating time in mixtures of
glycine, phenylalanine, and 4,5-epoxy-2-decenal. A mixture of glycine was heated either alone (control, open symbols) or in the presence of 4,5-epoxy-2-decenal (closed
symbols), for 2 days at 37 (panels A and B) or 60 °C (panels C and D) and pH 3. After that time, phenylalanine was added to all reaction mixtures and these mixtures were
heated for another 3 days at the same temperature.
reaction in which glycine was heated in the absence of the lipid)
observed at 60 °C than the observed at 37 °C.
with previous studies (see, for example, Guerra & Yaylayan, 2010).
Among the different -keto acids assayed in this study, GlyKA was
a
After 2 days, phenylalanine was added and the reaction was
maintained at 37 or 60 °C for other 3 days. The addition of this sec-
ond amino acid produced significant changes in the reaction mix-
tures. A continued decrease in glycine was observed. This decrease
was always linear for the different controls and also for the sam-
ples incubated in the presence of the epoxyalkenal at 37 °C. How-
ever, the loss of glycine, incubated in the presence of the
epoxyalkenal at 60 °C, was reduced during the second heating.
Analogous losses of glycine were observed in the presence and in
the absence of the oxidised lipid after 5 days at 60 °C. In addition
to glycine, phenylalanine also decreased with heating. However,
compared with the decrease in glycine after the first 2 days, the
loss of phenylalanine at the end of the experiment appeared to
be slightly higher at 37 than at 60 °C.
These changes in amino acid content at the end of the experi-
ment, also produced changes in the concentration of GlyKA and
in the formation of phenylacetaldehyde, the Strecker aldehyde of
phenylalanine (Fig. 3B and D at 37 and 60 °C, respectively). At
37 °C, the concentration of GlyKA only decreased slightly. On the
other hand, at 60 °C this second incubation, in the presence of
phenylalanine, consumed the GlyKA formed during the first 2 days.
In addition to the changes in the concentration of GlyKA, this sec-
ond incubation also produced phenylacetaldehyde. As observed in
Fig. 3B and D, phenylacetaldehyde was produced to a much higher
extent at 60 than at 37 °C.
the reactive carbonyl compound that produced the highest concen-
tration of phenylacetaldehyde (Fig. 1).
The degradation of amino acids by a-keto acids may follow two
alternative reaction pathways as shown in Fig. 4. Both of these pro-
duce an imine in the first step by reaction of the keto group of the
a-keto acid, with the amino group of the amino acid (Hidalgo &
Zamora, 2004). The imine can then undergo two alternative reac-
tions. On one hand, an electronic rearrangement with the loss of
carbon dioxide would produce a new imine, from which the Strec-
ker aldehyde is liberated by hydrolysis. Alternatively, the loss of
carbon dioxide may be facilitated by the formation of a relatively
stable azomethine ylide, through a 5-oxazolidinone intermediate
(Chu & Yaylayan, 2009; Tsuge, Kanemasa, Ohe, & Takenaka,
1987). This ylide would then produce the conjugated imines which
are the origins for both phenylacetaldehyde and phenylethyl-
amine. The first mechanism would be favoured if the compound
possess a carboxylic group at the
a-position of the carbonyl. This
would facilitate the electronic rearrangement. Therefore, this path-
way is not expected to be favoured for OHP. On the other hand, the
reaction producing the intermediate azomethine ylide would pro-
duce the corresponding phenylethylamine, in addition to phenyl-
acetaldehyde. This amine was not observed in the reactions of
phenylalanine with
in this study. Independently of the reaction pathway followed,
the -keto acid is converted into its corresponding -amino acid.
a-keto acids under the conditions employed
a
a
For this reason, when phenylalanine was degraded in the presence
4. Discussion
of GlyKA, the formation of glycine was observed.
Previous studies have shown that lipid oxidation products are
able to degrade amino acids in multiple ways, including their con-
a
-Keto acids are reactive carbonyl compounds, which have
been shown to convert asparagine into acrylamide (Zamora et al.,
2011). In addition, the results obtained in this study show that they
can also convert amino acids into Strecker aldehydes in agreement
version into Strecker aldehydes (Hidalgo & Zamora, 2004), a-keto
acids (Zamora et al., 2006), amines (Zamora, Delgado, & Hidalgo,
2012) and olefins (Hidalgo, Delgado, & Zamora, 2009). In addition

F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
1145
Fig. 4. Proposed reaction pathways for the Strecker degradation of phenylalanine initiated by glyoxilic acid.
to these degradations produced directly by oxidised lipids, the re-
sults obtained in this study suggest that oxidised lipids (or other
reactive carbonyl compounds) can also degrade amino acids in a
temperature. At 60 °C, all the epoxyalkenal was consumed, and a
significant decrease in glycine content after 2 days was observed.
The reaction between the epoxyalkenal and glycine produced Gly-
KA to a significant extent. When phenylalanine was added, the Gly-
KA present should produce the Strecker degradation product of this
amino acid, at the same time that GlyKA is converted into glycine.
For that reason, phenylacetaldehyde was observed at the same time
that GlyKA disappeared. In addition, similar results were obtained
for glycine content when the reaction was heated at 60 °C for
5 days, with and without epoxyalkenal. The only surprising result
obtained at 60 °C was the amount of phenylacetaldehyde produced,
which was higher than the amount of GlyKA. This suggests that
other glycine-epoxyalkenal reaction products also contribute to
the formation of phenylacetaldehyde at this temperature.
The reaction at 37 °C was different. After the first 2 days of heat-
ing, 4,5-epoxy-2-decenal was still observed, and the decrease in
glycine for both reaction with and without lipid, was similar. For
this reason, the formation of a smaller amount of GlyKA at 37 °C
was observed compared with reactions at 60 °C. When phenylala-
nine was added, the reaction mixture then contained two reactive
carbonyls: the remaining epoxyalkenal and GlyKA. Both could have
reacted with the new amino acid but the epoxyalkenal appeared to
delayed way via
duced by oxidised lipids, and the resultant
a
-keto acids. Thus,
a
-keto acids would be pro-
-keto acid would then
a
degrade amino acids when the oxidised lipids have been depleted.
Fig. 5 shows a general scheme by which lipid oxidation prod-
ucts (an epoxyalkenal in the figure) are able to produce Strecker
aldehydes from amino acids. The reaction is initiated by formation
of an imine between the carbonyl group of the lipid oxidation
product, and the amino group of the amino acid 1. This imine can
undergo two different electronic rearrangements, depending on if
decarboxylation occurs. If carbon dioxide is lost, the reaction prod-
ucts are the Strecker aldehyde 1 and the compound derived from
the lipid oxidation product (2-alkypyridines for epoxyalkenals). If
carbon dioxide is not lost, the reaction products are the
and the same compound derived from the lipid oxidation product.
This -keto acid can either decompose thermally to produce the
a-keto acid,
a
Strecker aldehyde 1, or react with a molecule of amino acid (amino
acid 1 or other amino acid 2, as indicated in the figure) to produce
its Strecker degradation product. This reaction produces a new
Strecker aldehyde, which may be the Strecker aldehyde 1 if pro-
duced from amino acid 1, but can also occur with amino acid 2.
The Strecker aldehyde 2 would be different to the Strecker aldehyde
1. In any case, this reaction allows the recovery of the initial amino
acid 1.
be more reactive than the a-keto acid as it was exhausted but only
a small decrease in the concentration of GlyKa was observed.
All these results show both the complexity of the chemistry be-
hind the formation of some flavours, and the difficulty to control
these reactions, as these flavours can be formed by alternative
pathways. Thus, the results obtained in this study provide a new
The results obtained are in agreement with this reaction path-
way, and suggest different alternatives depending on the reaction

1146
F.J. Hidalgo et al. / Food Chemistry 141 (2013) 1140–1146
Fig. 5. Alternative pathways for the Strecker degradation of amino acids produced in mixtures of amino acids and oxidised lipids.
Heredia, A., Peinado, I., Rosa, E., Andres, A., & Escriche, I. (2012). Volatile profile of
proof that Strecker aldehydes can be produced as a consequence of
amino acid degradation initiated by a high number of reactive car-
bonyl compounds. These include not only carbohydrates, lipids, or
phenols, but also carbonyl derivatives of amino acids and proteins.
In addition, and because amino acid-derived carbonyl compounds
can be produced by carbonyl–amine reactions of oxidised lipids (or
carbohydrates) with amino acids and proteins, Strecker degrada-
tion of new amino acids by these amino acid-derived carbonyl
compounds should be considered in many cases as a delayed con-
sequence of reactions produced by lipids and carbohydrates. Final-
dehydrated cherry tomato: Influences of osmotic pre-treatment and microwave
power. Food Chemistry, 130, 889–895.
Hidalgo, F. J., Delgado, R. M., & Zamora, R. (2009). Degradation of asparagine to
acrylamide by carbonyl-amine reactions initiated by alkadienals. Food
Chemistry, 116, 779–784.
Hidalgo, F. J., & Zamora, R. (2004). Strecker-type degradation produced by the lipid
oxidation products 4,5-epoxy-2-alkenals. Journal of Agricultural and Food
Chemistry, 52, 7126–7131.
Hidalgo, F. J., & Zamora, R. (2008). Food anoxia and the formation of either flavour or
toxic compounds by amino acid degradation initiated by oxidized lipids. Annals
of the New York Academy of Sciences, 1126, 25–29.
Hyytiaïnen, M., Appelblad, P., Pontén, E., Stigbrand, M., Irgum, K., & Jaegfeldt, H.
(1996). Trifluoromethanesulfonic acid as
dansylhydrazone derivatives. Journal of Chromatography A, 740, 279–283.
Kerler, J., Wimkel, C. (2002). The basic chemistry and process conditions
a catalyst for the formation of
ly, the
a-keto acids produced by carbonyl–amine reactions, are
transformed into their corresponding amino acids by reactions
with other amino acids to produce Strecker aldehydes. This sug-
gests that amino acids may be not uniformly lost during processing
or storage, and that the content of some amino acids might de-
crease very slightly, or even increase, as a consequence of these
reactions.
&
underpinning reaction flavour production. In A. J. Taylor (Ed.), Food Flavour
Technology (pp. 27–59). Sheffield, UK: Sheffield Publishers.
Rizzi, G. (2008). The Strecker degradation of amino acids: Newer avenues for flavour
formation. Food Reviews International, 24, 416–435.
Smit, B. A., Engels, W. J. M., Alewijn, M., Lommerse, G. T. C. A., Kippersluijs, E. A. H.,
Wouters, J. T. M., et al. (2004). Chemical conversion of a-keto acids in relation to
flavour formation in fermented foods. Journal of Agricultural and Food Chemistry,
52, 1263–1268.
Acknowledgements
Snedecor, G. W., & Cochran, W. G. (1980). Statistical methods (7th ed.). Ames, IA:
Iowa State University Press.
Tsuge, O., Kanemasa, S., Ohe, M., & Takenaka, S. (1987). Simple generation of
We are indebted to José L. Navarro for technical assistance. This
study was supported in part by the European Union (FEDER funds)
and the Plan Nacional de I + D of the Ministerio de Economía y
Competitividad of Spain (projects AGL2009-07638 and AGL2012-
35627).
nonstabilized azomethine ylides through decarboxylative condensation of
a-
amino acids with carbonyl compounds via 5-oxazolidinone intermediate.
Bulletin of the Chemical Society of Japan, 60, 4079–4089.
Villas-Boas, S. G., Delicato, D. G., Akesson, M., & Nielsen, J. (2003). Simultaneous
analysis of amino and nonamino acids as methyl chloroformate derivatives using
gas chromatography–mass spectrometry. Analytical Biochemistry, 322, 134–138.
Wang, R., Yang, C., & Song, H. L. (2012). Key meat flavour compounds formation
mechanism ina glutathione-xylose Maillardreaction.Food Chemistry, 131, 280–285.
References
Yarnpakdee, S., Benjakul, S., Nalinanon, S.,
& Kristinsson, H. G. (2012). Lipid
oxidation and fishy odour development in protein hydrolysate from Nile tilapia
(Oreochromis niloticus) muscle as affected by freshness and antioxidants. Food
Chemistry, 132, 1781–1788.
Appelblad, P., Pontén, E., Jaegfeldt, H., Bäckström, T.,
&
Irgum, H. (1997).
using
Derivatization of steroids with dansylhydrazine
trifluoromethanesulfonic acid as catalyst. Analytical Chemistry, 69, 4905–4911.
Chu, F. L., & Yaylayan, V. A. (2008). Model studies on the oxygen-induced formation
of benzaldehyde from phenylacetaldehyde using pyrolysis GC–MS and FTIR.
Journal of Agricultural and Food Chemistry, 56, 10697–10704.
Chu, F. L., & Yaylayan, V. A. (2009). FTIR monitoring of oxazolidin-5-one formation
and decomposition in a glycolaldehyde-phenylalanine model system by isotope
labeling techniques. Carbohydrate Research, 334, 229–236.
Gao, S., Oh, D. H., Broadbent, J. R., Johnson, M. E., Weimer, B. C., & Steele, J. L. (1997).
Aromatic amino acid catabolism by lactococci. Lait, 77, 371–381.
Granvogl, M., Beksan, E., & Schieberle, P. (2012). New insights into the formation of
aroma-active Strecker aldehydes from 3-oxazolines as transient intermediates.
Journal of Agricultural and Food Chemistry, 60, 6312–6322.
Yaylayan, V. A. (2003). Recent advances in the chemistry of Strecker degradation
and Amadori rearrangement: Implications to aroma and color formation. Food
Science and Technology Research, 9, 1–6.
Zamora, R., Delgado, R. M., & Hidalgo, F. J. (2011). Strecker aldehydes and
a-keto
acids, produced by carbonyl–amine reactions, contribute to the formation of
acrylamide. Food Chemistry, 128, 465–470.
Zamora, R., Delgado, R. M., & Hidalgo, F. J. (2012). Formation of b-phenylethylamine
as a consequence of lipid oxidation. Food Research International, 46, 321–325.
Zamora, R., & Hidalgo, F. J. (2005). Co-ordinate contribution of lipid oxidation and
Maillard reaction to the nonenzymatic food browning. Critical Reviews in Food
Science and Nutrition, 45, 49–59.
Zamora, R., Navarro, J. L., Gallardo, E., & Hidalgo, F. J. (2006). Chemical conversion of
Guerra, P. V., & Yaylayan, V. A. (2010). Dimerization of azomethine ylides: An
alternate route to pyrazine formation in the Maillard reaction. Journal of
Agricultural and Food Chemistry, 58, 12523–12529.
a
-amino acids into
a-keto acids by 4,5-epoxy-2-decenal. Journal of Agricultural
and Food Chemistry, 54, 6101–6105.