Investigation of reactive α-dicarbonyl compounds generated from the maillard reactions of L -Methionine with reducing sugars via their stable quinoxaline derivatives

Article abstract of DOI:10.1021/jf1008988

The formation of α-dicarbonyl compounds was investigated in methionine-catalyzed (Maillard reaction) thermal degradation of d-glucose, maltose, and dextrin 10 at three different pH values (3, 5, and 8). The α-dicarbonyl compounds were trapped as quinoxalines and could be quantified by HPLC and GC-MS. Formation of 1,4-dideoxypentodiulose from hexoses and disaccharides was evidenced for the first time. The use of L-methionine as the amino compound for the formation of 1,4-dideoxypentodiulose in model systems is explained. Furthermore, it could be shown that methionine has great effect on the formation of specific α-dicarbonyl compounds. At various pH values and also by application of mono-, di-, and oligosaccharides in all model reactions, 3-deoxyhexosulose and 1,4-dideoxypentodiulose were generated preferentially.

Full text of DOI:10.1021/jf1008988

J. Agric. Food Chem. 2010, 58, 8293–8299 8293
DOI:10.1021/jf1008988
Investigation of Reactive r-Dicarbonyl Compounds Generated
from the Maillard Reactions of -Methionine with Reducing
Sugars via Their Stable Quinoxaline Derivatives
L
Y
VONNE V. PFEIFER AND LOTHAR W. KROH*
Institute of Food Technology and Food Chemistry, Technical University of Berlin,
Gustav-Meyer-Allee 25, 13355 Berlin, Germany
The formation of R-dicarbonyl compounds was investigated in methionine-catalyzed (Maillard reaction)
thermal degradation of D-glucose, maltose, and dextrin 10 at three different pH values (3, 5, and 8).
The R-dicarbonyl compounds were trapped as quinoxalines and could be quantified by HPLC and GC-
MS. Formation of 1,4-dideoxypentodiulose from hexoses and disaccharides was evidenced for the
first time. The use of L-methionine as the amino compound for the formation of 1,4-dideoxypento-
diulose in model systems is explained. Furthermore, it could be shown that methionine has great effect
on the formation of specific R-dicarbonyl compounds. At various pH values and also by application of
mono-, di-, and oligosaccharides in all model reactions, 3-deoxyhexosulose and 1,4-dideoxypento-
diulose were generated preferentially.
KEYWORDS: L-Methioninemethional; aroma; Maillard reaction; R-dicarbonyl compounds; 1,4-dideoxy-
pentodiulose
INTRODUCTION
-Methionine is an essential, proteinogenic amino acid for
humans (1). -Methionine is contained in many different foods;
mainly it can be found in fish, meat, vegetables, egg, whole-grain
bread, and rice. Furthermore, -methionine is used in the flavor
without
research about carbohydrate and amino acid degradation path-
ways in -methionine/ -glucose model systems. They showed
that -glucose and -methionine are precursors of thiofurans,
thiopyrroles, and thiopyrazines, important compounds of meat
aroma (11).
D-glucose (10). Yaylayan and Keyhani accomplished
L
L
D
L
D
L
L
industry to produce aroma compounds, which smell like cooked
potatoes, coffee, or roasted meat (2, 3). The contribution of
Although many investigations about the reaction of L-methio-
nine with carbohydrates in thermally treated food and also in
model systems were carried out, the role of specific R-dicarbonyl
compounds is still unknown. Furthermore, it is ambiguous which
of the multitude of reactive compounds arising from L-methio-
nine in combination with carbohydrates generate the R-dicarbo-
nyl compounds. The aim of this work is to gain deeper insights
L
-methionine to flavoring proceeds predominantly through
thermal degradation or as a consequence of thermal inter-
actions with other food ingredients such as reducing sugars.
Investigations about reactions between amino acids and
R-dicarbonyl compounds showed that oxidative degradation of
the amino acid occurs and the so-called Strecker aldehydes are
generated (4). The Strecker aldehydes of specific amino acids
into the reaction between L-methionine and carbohydrates with
respect to the formation of specific R-dicarbonyl compounds.
especially from
L-leucine,
L-phenylalanine, and
L-methionine, but
also from -valine,
L
L
-isoleucine, and L
-alanine, are well-known for
MATERIALS AND METHODS
their significant odors. These aldehydes are formed in food also
from reactions between R-dicarbonyl compounds and amino
acids (5). In recent years investigations have corroborated that
Chemicals. The following chemicals were obtained commercially:
D-glucose, D-fructose, maltose, sucrose, dextrin 10, L-methionine, 3-pen-
tyn-1-ol, and ruthenium(IV)oxide hydrate were purchased from Fluka
(Neu-Ulm, Germany); maltotriose, sodium periodate, 1,2-diaminoben-
the Strecker aldehydes of these three amino acids,
-phenylalanine, and -methionine, are key aromas of many ther-
mally treated foods (6-8). Reactions between -methionine
and monosaccharides were studied by Tressl et al. (3) and Rijke
et al. (9). Both groups identified 3-(methylthio)propylamine,
methional, methanethiol, and acrolein as reactive intermediates
that are responsible for most of the end-products occurring in the
model systems. Yu and Ho analyzed the formation of various
heterocycles, for example, pyrazines generated from the thermal
L-leucine,
L
L
zene,
steine,
L
-phenylalanine,
L
-alanine, γ-aminobutyric acid,
L-leucine, L-cy-
L
L
-asparagine, and [13C1]-
D-glucose were purchased from Aldrich
(Steinheim, Germany). 1,2-Diaminobenzene was recrystallized twice from
methanol.
Synthesis. (2S,3R)-1-(2-Quinoxalinyl)-2,3,4-trihydroxybutane.
3-Deoxy-D-erythro-2-hexosulose (3-DG) was synthesized largely according
to the method of Madson and Feather (12). Interception of 3-deoxy-
D-
erythro-2-hexosulose as stable (2S,3R)-1-(2-quinoxalinyl)-2,3,4-trihydroxy-
butane was carried out according to the method of Glomb and Tschirnich (13).
GC-MS (3-DG-quinoxaline after acetylation): tR, 27.99 min; m/z 360
(0.2%, Mþ), 301 (7), 300 (9), 258 (12), 241 (49), 199 (33), 181 (24), 157 (34),
144 (100), 102 (11), 43 (89).
reactions of L-methionine and L-methionine sulfoxide with or
*Corresponding author (e-mail lothar.kroh@tu-berlin.de; phone
þ49 30 314 72 583; fax þ49 30 314 72 585).
© 2010 American Chemical Society
Published on Web 06/24/2010
pubs.acs.org/JAFC

8294 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Pfeifer and Kroh
(1S,2R)-1-(3-Methyl-2-quinoxalinyl)-1,2,3-propanetriol (1-DG-quinox-
aline) was prepared largely according to the method of Beck and Ledl (14)
via the Amadori compounds.
GC-MS (1-DG-quinoxaline after acetylation): tR, 24.38 min; m/z 360
(0.2%, Mþ), 301 (4), 300 (4), 241 (18), 199 (51), 198 (12), 174 (100), 157 (5),
143 (23), 107 (22), 103 (9), 43 (96).
was trapped with 500 μL of 0.05 M 1,2-diaminobenzene solution
to intercept the R-dicarbonyls as quinoxalines. After 3 h at 25 °C,
quinoxalines were analyzed by HPLC-DAD and GC-MS after acetyla-
tion. The molecular ions (MS-EI) of labeled and unlabeled acetylated
(1S,2R)-1-(3-methyl-2-quinoxalinyl)-2-ethyl alcohol [230 (10, [Mþ])/
231 (10, [Mþ])] were used for the quantification.
(1R,2S,3R)-1-(2-Quinoxalinyl)-1,2,3,4-tetrahydroxybutane (glucosone-
quinoxaline) was carried out according to the method of Morita et al. (15).
GC-MS (glucosone-quinoxaline after acetylation): tR, 30.66 min; m/z
418 (0.2%, Mþ), 359 (2), 358 (2), 299 (63), 257 (11), 215 (14), 213 (9), 202
(50), 196 (8), 160 (100), 143 (4), 129 (10), 115 (12), 102 (7), 43 (97).
(1S,2R)-1-(3-Methyl-2-quinoxalinyl)-2,3-propanediol (1,4-DDH-quinox-
aline) was prepared largely according to the method of Morita et al. (15).
GC-MS (1,4-DDH-quinoxaline after acetylation): tR, 22.31 min; m/z
302 (0.2%, Mþ), 243 (16), 242 (7), 183 (100), 171 (40), 158 (77), 143 (7), 117
(9), 102 (9).
Model Reaction Food Items. In a typical experiment 5.00 g of food
sample was milled and roasted at 150/170 °C. After a defined time, the
samples were derivatized for 30 min with 10 mL of 0.05 M 1,2-diamino-
benzene solution, filtered, centrifuged at 6000 U, and analyzed by HPLC-
DAD and GC-MS after acetylation.
Chromatography. Thin-layer chromatography (TLC) was performed
on silica gel 60 F254 plates (Merck, Darmstadt, Germany). Preparative
column chromatography was performed on silica gel 60, 40-63 μm (Merck).
All solvents were of chromatographic grade.
Nuclear Magnetic Resonance Spectroscopy (NMR). NMR spectra
were recorded on a Bruker AC 400 instrument (Rheinstetten, Germany).
Chemical shifts are given in parts per million relative to residual non-
deuterated solvent as an internal reference.
High-Performance Liquid Chromatography (HPLC-DAD) In-
strumentation: degasser, Gegasys DG-13000 (Knauer); pump, Shimadzu
LC-10 AT; thermostat, 30 °C, Shimadzu CT0-6A; guard column, Nucleosil
120-5 C18 Macherey-Nagel; column, Nucleosil 5 C18 (250 mm ꢀ 4.6 mm);
injection volume, 40 μL; eluent, methanol/water gradient, 1 mL/min.
Gas Chromatography (GC-MS). The samples were extracted with
n-butanol. The solvent was dried off, the residue was dissolved in toluene,
and acetic anhydride was added (18). Gas chromatography instrumenta-
tion: Finnigan GCQ; capillary column, BPX 5 (SGE, 30 m, 0.25 mm i.d.,
0.5 μm film thickness); carrier gas, helium 4.6; detector, Finnigan Ion Trap
Mass Analyzer GCQ; injection temperature, 270 °C; split 1:10; tempera-
ture program, initial temperature, 95 °C, held for 1 min, raised from 95 to
200 °C at 15 °C/min, 200 °C held for 1 min, raised from 200 to 280 °C at
3 °C/min, 280 °C held for 5 min, raised from 280 to 300 °C at 5 °C/min,
300 °C held for 5 min. Column effluents were analyzed by selected ion
monitoring (SIM). Quantification was carried out by comparison of peak
areas obtained in the TIC with those of standard solutions containing
amounts of pure authentic reference compounds.
(2S,3R)-1-(2-Quinoxalinyl)-2,3dihydroxypropane (3-DP-quinoxaline).
This synthesis was performed as described by Hollnagel and Kroh (16).
GC-MS (3-DP-quinoxaline after acetylation): tR, 21.28 min; m/z 288
(0.3%, Mþ), 229 (5), 228 (4), 169 (42), 157 (24), 144 (37), 129 (3), 117 (4),
43 (100).
1,4-Dideoxypentodiulose (17, 19, 20). To a solution of 3-pentyn-1-ol
(2 mL, 26 mmol) in CHCl3 (60 mL) and MeCN (60 mL) was added NaIO4
(12.84 g, 60 mmol) in H2O (120 mL). The mixture was vigorously stirred,
and RuO2 H2O (120 mg, 0.79 mmol) was added. Stirring was continued
3
for 15 min in the presence of air at ambient temperature. Then the reaction
mixture was filtered through a pad of silica and rinsed with CH2Cl2 as the
eluent. The solution was dried with Na2SO4 and concentrated in vacuo to
give a yellow oil. Purification via flash chromatography on silica (EtOAc/
pentane, 1:1) afforded pure 1,4-dideoxypentodiulose (2.02 g) in 67% yield.
1H NMR (400 MHz, CDCl3): δ 1.45 (s, 2.5, -CH3) þ 2.34 (s, 0.5,
-CH3), 2.60 (m, 1.6, -CH2) þ 2.99 (m, 0.4, -CH2), 3.92 (m, 0.4, -CH2) þ
4.24 (m, 1.6, -CH2).
13C NMR (75 MHz, CDCl3): δ 21.3, 23.5, 33.6, 38.4, 57.3, 61.9, 96.5,
197.0, 198.4, 209.9.
(1S,2R)-1-(3-Methyl-2-quinoxalinyl)-2-ethyl Alcohol. 1,4-Dideoxy-
pentodiulose (200 mg, 1.72 mmol) and 1,2-diaminobenzene (167 mg, 1.55
mmol) were stirred for 30 min at 100 °C in an evacuated 5 mL flask. Puri-
fication via flash chromatography on silica (EtOAc/pentane, 1:1, EtOAc)
afforded pure (1S,2R)-1-(3-methyl-2-quinoxalinyl)-2-ethyl alcohol (313 mg)
in 98% yield as a bright orange solid.
RESULTS AND DISCUSSION
Various studies concerning the relevance of
thermally treated food have been performed to date. It is well-
known that by thermal treatment of -methionine both in the
L-methionine in
1H NMR (400 MHz, D2O): δ 2.62 (s, 3, -CH3), 3.12 (t, 2, J = 6.8 Hz,
-CH2), 3.98 (t, 2, J = 6.8 Hz, -CH2), 7.70 (m, 2, Ar-H), 7.77 (m, 2,
Ar-H).
L
absence or in the presence of carbohydrate/dicarbonyl com-
pound one of the main products is methional (10). Many
investigations have been accomplished, although only limited
information about the reaction pathways of the carbohydrate
compounds in these model systems is available. Does the
formation of specific R-dicarbonyl compounds depend on the
13C NMR (75 MHz, D2O): δ 21.5, 37.2, 60.0, 126.7, 127.0, 129.7, 129.9,
139.3, 139.6, 154.3, 154.5.
GC-MS (1,4-DDP-quinoxaline after acetylation): tR, 16.06 min; m/z
230 (0.2%, Mþ), 188 (13), 187 (100), 169 (100), 159 (44), 158 (36), 144 (8),
117 (9), 77 (7).
Model Reactions. In a typical experiment an aqueous solution of
presence of
the amino acid? Furthermore, the interest is directed toward
the use of different carbohydrates. -Glucose was used as a
L-methionine, or is there no specific influence by
carbohydrate (1 mmol) and amino acid (1 mmol) (L-methionine,
L-phenyl-
alanine, -alanine, γ-aminobutyric acid, -leucine,
L
L
L-cysteine, L-aspar-
D
agine) in 10 mL of H2O was adjusted to a pH of 3 or 5 with 3 N HCl or
to a pH of 8 with 3 N NaOH.
typical monosaccharide in food, maltose as an agent for
disaccharides, and dextrin 10 as an agent for oligosaccharides,
which showed in previous works that reaction behavior is
different compared to monosaccharides (16).
Within this work various model reactions with different
carbohydrates (mono-, di-, and oligosaccharides) and the amino
The model solutions were heated in sealed ampules at 100 ( 1 or 130 (
1 °C for up to 300 min in a thermo block (Behr Labor Technik, behrotest
ET2). After a defined reaction time, 500 μL of the samples was trapped
with 500 μL of 0.05 M 1,2-diaminobenzene solution to intercept the
R-dicarbonyls as quinoxalines. After 3 h at 25 °C, quinoxalines were
analyzed by HPLC-DAD and GC-MS after acetylation. Values were
expressed as the means of at least three independent determinations.
Model Reaction Diacetyl/Formaldehyde. In a typical experiment
an aqueous solution of diacetyl (1 mmol) and formaldehyde (37%) (1 mmol)
in 1.00 mL of H2O was adjusted to a pH of 8. The model solution was stirred
for 10 min at ambient temperature and then trapped with 1.00 mL of 0.05 M
1,2-diaminobenzene solution to intercept the R-dicarbonyls as quinoxalines.
acid L-methionine were performed in aqueous solutions (starting
with pH 3, 5, and 8). The samples were heated in sealed ampules
at 130 °C for a defined time and then treated with 0.5 M 1,2-
diaminobenzene solution to trap the arising R-dicarbonyl com-
pounds as stable, measurable quinoxalines (16).
In line with our investigations with L-methionine as amino
component, an unidentified peak could be found in the HPLC
chromatogram, which showed the typical UV spectrum of a
quinoxaline (Figure 1).
Remarkably, this unknown component is only generated in the
model system with L-methionine as the amino compound and
Labeling Experiments. An aqueous solution of [13C1]-
(1 mmol) or [13C]-
-glucose (1 mmol) and -methionine (1 mmol) in
D-glucose
D
L
10 mL of H2O was adjusted to a pH of 5. The model solution was heated
in sealed ampules at 130(1 °C for up to 120 min in a thermo block
(Behr Labor Technik, behrotest ET2). Afterward, 500 μL of the sample

Article
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8295
Figure 1. HPLC chromatogramm of a glucose/L-methionine model system.
Figure 3. Synthesis of 1,4-dideoxypentodiulose and its quinoxaline.
Figure 2. GC-MS spectrum (SIM-MS) of the acetylated quinoxaline of
1,4-dideoxypentodiulose.
with no other amino acid, that is,
-asparagine, -cysteine, or glycine.
It is a crucial point why this unknown R-dicarbonyl compound
is formed only in the presence of -methionine and what the
L-phenylalanine, L-leucine,
L
L
L
structure is about. When the chromatographical retention time of
this quinoxaline is compared with the established quinoxalines, it
becomes clear that this component must be more nonpolar than
the quinoxaline of 3-deoxyhexosulose and more polar than the
quinoxaline of methylglyoxal. To obtain more information about
the structure, the samples were acetylated and then measured as
acetylated quinoxalines by GC-MS.
All quinoxalines show the same typical mass fragments of m/z
144, 117, and 77, which are caused by the quinoxaline skeleton
(Figure 2). Thus, the task consisted of searching for more
substances in the GC spectrum that show these typical fragments
and that have retention indices in accord with the results from the
HPLC. Such a substance, eluting on the GC column at 16.06 min,
could be found with a m/z ratio from 230. With regard to the
appropriated fragmentation and its retention indices, the un-
known dicarbonyl was posited to be 1,4-dideoxypentodiulose.
For confirmation of the proposed structure 1,4-dideoxypento-
diulose was synthesized in one step. Various syntheses have been
described by de Kimpe et al. (19, 20), but they are all more
complicated.
Figure 4. HPLC chromatogram of the synthesized new quinoxaline (A)
and of themodel system D-glucose/L-methionine, at 130 °C and with60min
of thermal treatment (B).
Table 1. Influence of Carbohydrate and pH Value on the Formation of
R-Dicarbonyl Compounds (Milligrams per Liter) in Model Systems from
L-Methionine/Carbohydrate at 130 °C after 180 min
D-glucose
maltose
dextrin10
pH 3 pH 5 pH 8 pH 3 pH 5 pH 8 pH 3 pH 5 pH 8
glucoson
1-DH
3-DH
0
0
0
0
0
1.9
0
0
0
6.9 36.7
0
0
0
0
0
0
7.5
40.1 32.8 27.3 8.2 15.6 19.6 10.3 15.3 15.9
3-DP
1,4-DDH
glyoxal
0
0
0
0
0
0
0
4.4
0
0
0
0
0
0
0
0.5
0
0.5
0.1
0.7
0
0.8
0
0
0.2
6.8
0
0.7
7.4
0
0.3
4.3
0
0.4
4.7
0.4
0.9
0
1,4-DDP
methylglyoxal
diacetyl
2.4 0.1
7.4
0
3.9
1.9
1.1
2.0
1.1
0
0
0
0
0
0
0
The oxidation of 3-pentyn-1-ol with NaIO4 and RuO2 H2O as
3
catalyst afforded 1,4-dideoxypentodiulose in pure form in 67%
yield. In a next step the R-dicarbonyl was trapped with 1,2-
diaminobenzene to yield the stable, measurable quinoxaline in
nearly quantitative yield of 98% (Figure 3).
is predominant (Table 1). That is not the case when using other
amino compounds (16, 21).
The higher amount of 3-deoxyhexosulose at acidic pH values is
in accordance with the literature, whereby the formation of
3-deoxyhexosulose is preferred (Figure 5) (5).
At a pH of 8 1-deoxy-2,3-hexodiulose, glyoxal, methylglyoxal,
and diacetyl were formed in addition. 1,4-Dideoxypentodiulose
arose just like 3-deoxyhexosulose at an acidic pH value in higher
amounts than in alkaline values (Figure 6).
Comparison of the synthesized dicarbonyl compound with the
model systems revealed that the new compound is actually 1,4-
dideoxypentodiulose (Figure 4).
To investigate the differences in reaction between mono- and
oligosaccharides, the reaction behavior of D-glucose with L-methionine
was investigated thoroughly. The reactions were performed at various
pH values (3, 5, and 8). Noticeably, independent from the pH value
the formation of 3-deoxyhexosulose and 1,4-dideoxypentodiulose
From the given standard reference quinoxalines, 3-deoxyhexo-
sulose and 1,4-dideoxypentodiulose were also dominant R-dicar-
bonyl compounds inmodel reactions withmaltose. Using maltose

8296 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Pfeifer and Kroh
Figure 6. Formation of 1,4-dideoxypentodiulose (1,4-DDP) from
maltose, and dextrin 10 with
300 min time period.
D
-glucose,
Figure 5. Formation of 3-deoxyhexosulose (3-DH) from
tose, and dextrin10 with
300 min time period.
D
-glucose, mal-
L
-methionine at pH 3, 5, and 8 and 130 °Covera
L
-methionine at pH 3, 5, and 8 and 130 °C over a
as carbohydrate compound the percentage of these two R-dicar-
bonyl compounds differs from -glucose model systems under
By comparison of the formation of the R-dicarbonyl com-
pounds in dependence on the carbohydrates, it becomes clear that
D
various pH values (see Figures 5 and 6). Whereas at a pH of 3 after
300 min 17 mg/L 3-deoxyhexosulose was produced, at a pH of
8 25 mg/L was generated. Thus, in this case formation of 3-deoxy-
hexosulose is preferred at alkaline pH values. Furthermore,
1-deoxy-2,3-hexodiulose (the dominant R-dicarbonyl compound
in this reaction), 3-deoxypentodiulose, 1,4-dideoxyhexodiulose,
glyoxal, and diacetyl were formed at a pH of 8. 1,4-Dideoxypen-
todiulose arose from maltose at a pH of 3 in low concentrations
by using D-glucose the formation of 3-deoxyhexosulose and 1,4-
dideoxypentodiulose is preferred. Using the disaccharide maltose
or oligosaccharides such as dextrin 10 no pH dependence of
the formation of 3-deoxyhexosulose can be detected (22). The
generation of 1,4-dideoxypentodiulose is favored at pH values of
5 and 8.
The experiments also show that there is a difference between
the formation of 3-deoxyhexosulose and the formation of 1,4-di-
deoxypentodiulose. Whereas the concentration of 3-deoxyhexo-
sulose arises directly, 1,4-dideoxypentodiulose is formed in higher
amounts not before 90 min.
compared to
forms 6 mg/L 1,4-dideoxypentodiulose,
The behavior of dextrin 10 in model systems with
D
-glucose. At a pH of 8 the ratio is reversed; maltose
-glucose only 4 mg/L.
-methionine is
D
L
similar to that of the disaccharide maltose (see Figures 5 and 6).
3-Deoxyhexosulose is generated at a pH of 3, 5, and also 8 in nearly
equal concentrations. After 300 min of thermal treatment, about
12 mg/L of 3-deoxyhexosulose could be detected. At pH values of
3 and 5 1,4-dideoxy-2,3-hexodiulose is formed and additionally
3-deoxyhexosulose and 1,4-dideoxypentodiulose are formed in small
amounts. At a pH of 8 1-deoxy-2,3-hexodiulose, 3-deoxypentodiu-
lose, glyoxal, methylglyoxal, and diacetyl were produced as well.
1,4-Dideoxypentodiulose shows the same behavior in its for-
Now the question arises: which role does 1,4-dideoxypento-
diulose play in methionine-containing food? Does it influence the
aroma properties itself, or does it generate resulting products,
which form intensive, characteristic aroma compounds? In real
food samples 1,4-dideoxypentodiulose was generated: sunflower
seeds, roasted for 15 min at 170 °C, yielded 1.76 mg/kg, and in
cashews, 1.4 mg/kg was formed (Figure 7).
The well-known aroma compound 2,3-pentanedione could be
formed from 1,4-dideoxypentodiulose by water elimination as
reported by Weenen et al. (23). 2,3-Pentanedione has aroma
properties similar to those of diacetyl and provides a buttery
aroma. It is formed in addition to diacetyl, for example, during
mation as in the model system maltose/L-methionine. At a pH of
3, 2 mg/L was produced after 300 min; at a pH of 5 or 8, 5 mg/L
was formed.

Article
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8297
Figure 7. Formation 1,4-dideoxypentodiulose (1,4-DDP) from cashew nuts and sunflower seeds in roasted model systems at 150/170 °C over a 20 min
time period.
Figure 8. Generation of formaldehyde from methionine via Strecker degradation and retro-Michael reaction.
beer aging (24). Furthermore, it is likewise that 1,4-dideoxypen-
todiulose yields 2-methyl-3-furanone by β-elimination. 2-Methyl-
3-furanone has a sweet aroma and could be found in baked
potatoes and chips as an aroma compound (25).
Formation of 3-Deoxyhexosulose. On the basis of the investiga-
tions of various model systems with different carbohydrates and
carbohydrate compound up to now (26). Presumably it is
generated via a formation pathway similar to that of the produc-
tion of 1,4-dideoxy-2,3-hexodiulose from hexoses. First, 1-deoxy-
2,3-hexodiulose is formed via the 2,3-endiol of the Amadori
product by β-elimination. Keto-enol tautomerization and a
second β-elimination yield 1,4-dideoxy-2,3-hexodiulose.
diverse pH values it can be deduced that the amino acid
L-methio-
The key question is whether 1,4-dideoxypentodiulose can be
nine produces 3-deoxyhexosulose and 1,4-dideoxypentodiulose
in all model systems independently from the carbohydrate pre-
sent. The formation of 3-deoxyhexosulose in the model system
generated from hexoses and what role L-methionine plays in the
formation of this R-dicarbonyl compound. The importance of
this amino acid becomes apparent when the results are compared
with those obtained from investigations with other amino acids.
The first hypothesis was that the formation of 1,4-dideoxypento-
D-glucose/L-methionine at acidic pH values is well-known. At
alkaline pH ranges other R-dicarbonyl compounds can be de-
tected as well. An explanation is the catalysis of enolization,
rearrangement, and fragmentation reactions in the course of
the Maillard reaction at alkaline conditions. Therefore, the low
concentration of 3-deoxyhexosulose at this pH value can be
elucidated. Considering the total R-dicarbonyl concentration at
different pH ranges, the amount is higher at alkaline values,
in accordance with the literature (5).
diulose was conditional on the sulfur of
reactions with -cysteine as amino compound could not approve
this assumption. Also, the use of -phenylalanine, -asparagine,
γ-aminobutyric acid, -leucine, glycine, and glutamine did not
L-methionine, but model
L
L
L
L
yield 1,4-dideoxypentodiulose.
So what differentiates the properties of L-methionine from the
other amino acids? Considering the amino acids in thermally
treated model reactions with carbohydrates, it is remarkable that
By the use of the disaccharide maltose or the oligosaccharide
dextrin 10 initially the glycosidic bond must be cleaved to obtain
reducing sugars. Therewith, the lower amount of 3-deoxyhexo-
L-methionine indeed like all other amino acids forms the corres-
ponding Strecker aldehyde (Figure 8). Methional is an unstable
aldehyde, which decomposes very quickly into methanethiol,
dimethyl disulfide, and acrolein via a retro-Michael reaction (28).
From the produced acrolein formaldehyde is generated in the
next step via hydration to 3-hydroxypropanal and subsequent
retro-Aldol reaction.
sulose in comparison to D-glucose at acidic pH values can be
explained. At a pH of 8 there is only a slight difference in the
3-deoxyhexosulose concentration of various carbohydrates, be-
cause the actual concentration of glucose is lower. In this pH
regimen color and melanoidin formation from D-glucose is faster.
The approach of postderivatization allows the determination of
the actual concentration of the R-dicarbonyl compounds, so that
the formed 3-deoxyhexosulose reacts to melanoidins or decom-
position products and cannot be detected as quinoxaline any
more.
It is conceivable that 1,4-dideoxypentodiulose is generated in
an Aldol reaction from diacetyl with formaldehyde (Figure 9)
Diacetyl can be produced in the Maillard reaction via different
pathways (see Figure 9). It can be formed via a retro-Aldol
reaction from
glycolaldehyde and
elimination steps diacetyl is produced (27). Oligosaccharides
D-glucose. Thereby, the C6-frame is cleaved into
Formation of 1,4-Dideoxypentodiulose. 1,4-Dideoxypentodiu-
lose could be detected only in model systems with pentoses as
D-erythrose. In following oxidation and

8298 J. Agric. Food Chem., Vol. 58, No. 14, 2010
Pfeifer and Kroh
values and also by the application of mono-, di-, and oligo-
saccharides in all model reactions 3-deoxyhexosulose and 1,4-
dideoxypentodiulose were generated preferentially. The identity
was unequivocally confirmed by chemical synthesis of 1,4-di-
deoxypentodiulose and its quinoxaline. Formation of 1,4-dideoxy-
pentodiulose from hexoses was evidenced the first time and could
be quantified. Furthermore, we could show a new reaction
pathway for R-dicarbonyl compounds in the Maillard reaction
via Aldol condensation.
ACKNOWLEDGMENT
We are indebted to Dr. Reinhard Zeisberg, Technical Univer-
sity Berlin, for NMR measurement.
LITERATURE CITED
(1) Meister, A. The Role of Amino Acids in Nutrition; Academic Press:
New York, 1965; pp 201-230.
(2) Hodge, J. E. Origin of Flavour in Food. Nonenzymatic Browning
Reactions; AVI Publishing: Wesport, CT, 1967.
(3) Tressl, R.; Helak, B.; Martin, N.; Kersten, E. Formation of Amino
Acid Specific Maillard Products and Their Contribution to Thermally
Generated Aromas; ACS Symposium Series; American Chemical
Society: Washington, DC, 1989; pp 156-169.
€
(4) Hofmann, T.; Munch, P.; Schieberle, P. Quantitative model studies
on the formation of aroma-active aldehydes and acids by Strecker-
type reactions. J. Agric. Food Chem. 2000, 48, 434–440.
(5) Ledl, F.; Schleicher, E. The Maillard reaction in food and human
body - new results in chemistry, biochemistry and medicine. Angew.
Chem. 1992, 102, 597–626.
(6) Kerler, J.; Grosch, W. Character impact odorants of boiled chicken:
changes during refrigerated storage and reheating. Z. Lebensm.
Unters. Forsch. 1997, 205, 232–238.
Figure 9. Generation of diacetyl from oligosaccharides via two different
formation pathways: (A) pathway via glycosidic cleavage and retro-Aldol
reaction of D-glucose; (B) pathway via the peeling-off mechanism and
retro-Aldol reaction of 1,4-dideoxy-2,3-hexodiulose. Subsequent Aldol
reaction of diacetyl with formaldehyde yields 1,4-dideoxypentodiulose (B).
(7) Blank, I.; Sen, A.; Grosch, W. Potent odorants of the roasted powder
and brew of Arabica coffee. Z. Lebensm. Unters. Forsch. 1992, 195,
239–245.
(8) Wagner, R.; Grosch, W. Evaluation of potent odorants of french
fries. Lebensm. Wiss. Technol. 1997, 30, 164–169.
(9) Rijke, D.; Dort, J. M.; Boelnes, H. Shigematsu Variation of the
Maillard Reaction; de Gruyter: Berlin, Germany, 1981; pp 417-431.
(10) Yu, T.; Ho, C. T. Volatile compounds generated from thermal
reaction of methionine and methionine sulfoxide with or without
glucose. J. Agric. Food Chem. 1995, 43, 1641–1646.
Table 2. Amounts of 1,4-Dideoxypentodiulose Generated upon Heating of a
Glucose/L-Methionine, [13C1]Glucose/L-Methionine, or [13C]Glucose/L-Methio-
nine Model System
R-dicarbonyl isotopomer
yield (mg/L)
1,4-dideoxypentodiulose
[13C1]-1,4-dideoxypentodiulose
[13C1-4]-1,4-dideoxypentodiulose
5.6
not detectable
5.4
(11) Yaylayan, V. A.; Keyhani, A. Carbohydrate and amino acid
can react in an alternative route to Amadori products, and via the
peeling-off mechanism 1,4-dideoxy-2,3-hexodiulose is formed
and hence, by retro-Aldol cleavage, diacetyl (16). For reassurance
of the postulated mechanism diacetyl(1 mmol) and formaldehyde
(1 mmol) were dissolved in water (1 mL) and stirred for 10 min at
a pH of 8 at ambient temperature. After 10 min, the reaction
mixture was trapped with 1 mL of 0.05 M 1,2-diaminobenzene
solution to yield the quinoxaline of 1,4-dideoxypentodiulose.
To confirm these results, the following experiment was per-
degradation pathways in L-methionine/D-[13C]glucose model sys-
tems. J. Agric. Food Chem. 2001, 49, 800–803.
(12) Madson, M. A.; Feather, M. S. An improved preparation of
3-deoxy-D-erythro-2-hexosulose via the bis(benzoylhydrazone) and
some related constitutional studies. Carbohydr. Res. 1981, 94,
183–197.
(13) Glomb, M. A.; Tschirnich, R. Detection of R-dicarbonyl compounds
in Maillard reaction systems. J. Agric. Food Chem. 2001, 49, 5543–
5550.
(14) Beck, J.; Ledl, F.; Severin, T. Formation of 1-deoxy-D-erythro-
formed: Isotopically labeled [13C1]-
was added to an aqueous solution of
D
-glucose or [13C]-
-methionine. The reaction
D-glucose
2,3-hexodiulose from Amadori compounds. Carbohydr. Res. 1988,
177, 240–243.
L
was started and, after thermal treatment for 120 min, the amounts
of unlabeled 1,4-dideoxypentodiulose, which was expected to be
formed, as well as [13C1]-1,4-dideoxypentodiulose, formed only
via another reaction pathway, and [13C1-4]-1,4-dideoxypentodiu-
lose should quantified. As shown in Table 2, 1,4-dideoxypento-
diulose and [13C1-4]-1,4-dideoxypentodiulose wereformed in high
amounts, whereas the 13C1-labeled compound was not detectable.
This result supports the hypothesis that 1,4-dideoxypentodiulose
is formed via an Aldol reaction from diacetyl and formaldehyde
(15) Morita, N.; Inoue, K.; Takagi, M. Quinoxalines derived from
D
-glucose and o-phenylendiamine in a weakly acidic medium. Agric.
Biol. Chem. 1981, 45, 2665–2888.
(16) Hollnagel, A.; Kroh, L. W. Degradation of oligosaccahrides in
nonenzymatic browning by formation of R-dicarbonyl compounds
via a “peeling off” mechanism. J. Agric. Food Chem. 2000, 48, 6219–
6226.
(17) Semmelhack, M. F.; R. C., S.; Federle, M. J.; Bassler, B. L. An
expeditious synthesis of DPD and boron binding studies. Org. Lett.
2005, 7, 569–572.
(18) Nedvidek, W.; Fischer, F. L. P. Detection of 5-hydroxymethyl-2-
methyl-3(2H)-furanone and of R-dicarbonyl compounds in reaction
mixtures of hexoses and pentoses with different amines. Z. Lebensm.
Unters. Forsch. 1992, 194, 228–228.
during the thermal treatment of
systems.
D-glucose/L-methionine model
This work points out that
L-methionine has a great effect on the
formation of specific R-dicarbonyl compounds. At various pH

Article
J. Agric. Food Chem., Vol. 58, No. 14, 2010 8299
(19) de Kimpe, N.; et al. Synthesis of laurencione, a labile dihydrofur-
anone derivative from red alga Laurencia spectabilis. J. Org. Chem.
1995, 47, 3280–3284.
during aging of top-fermented beer. J. Agric. Food Chem. 2003, 51,
6782–6790.
(25) Oruna-Concha, M. J.; Duckham, S. C.; Ames, J. M. Comparison of
volatile compounds isolated from skin and flesh of four potato
cultivars after baking. J. Agric. Food Chem. 2001, 49, 2414–2421.
(26) Cerny, C.; Davidek, T. Formation of aroma compounds from ribose
and cysteine during the Maillard reaction. J. Agric. Food Chem.
2003, 51, 2714–2721.
(20) de Kimpe, N.; et al. New formal synthesis of laurencione, a labile
dihydrofuranone derivative from red alga Laurencia spectabilis.
Synthesis 1995, 1995, 1131–1134.
(21) Fiedler, T.; Moritz, T.; Kroh, L. W. Influence of R-dicarbonyl
compounds to the molecular weight distribution of melanoidins in
sucrose solutions: part 1. Eur. Food Res. Technol. 2006, 223, 837–842.
(22) Kroh, L. W.; Wagner, T. F. J. The formation of coloured com-
pounds in technical sucorse solutions - new knowledge about their
molecular basics and molecular size distribution. Sugar Ind. 2007,
132, 25–32.
(27) Yaylayan, V. A.; Keyhani, A. Origin of 2,3-pentanedione and
2,3-butanedione in
D-glucose/L-alanine Maillard model systems.
J. Agric. Food Chem. 1999, 47, 3280–3284.
(28) Pripis-Nicolau, L. Formation of flavor components by the reaction
of amino acid and carbonyl compound in mild conditions. J. Agric.
Food Chem. 2000, 48, 3761–3766.
(23) Weenen, H. Reactive intermediates and carbohydrate fragmentation
in Maillard chemistry. Food Chem. 1998, 62, 393–401.
Received for review March 10, 2010. Revised manuscript received
May 30, 2010. Accepted June 03, 2010.
(24) Vanderhaegen, B.; Neven, H.; Coghe, S.; Verstrepen, K. J.;
Verachtert, H.; et al. Evolution of chemical and sensory properties