3-Deoxypentosulose: An α-dicarbonyl compound predominating in nonenzymatic browning of oligosaccharides in aqueous solution

Article abstract of DOI:10.1021/jf011107o

The thermal degradation of D-glucose, maltose, and maltotriose in aqueous solution was investigated under caramelization (no glycine) and Maillard (with glycine) conditions. Degradation of the sugar and α-dicarbonyls product was monitored. Under both caramelization and Maillard reaction conditions, 3-deoxypentosulose was the predominating α-dicarbonyl compound formed from maltose and maltotriose. In the absence of an amino compound, however, 3-deoxypentosulose is formed in much lower concentration. It was concluded that 3-deoxypentosulose is formed by a pathway specific for oligo- and polysaccharides since this α-dicarbonyl is formed from the α-1→4 glucans such as maltose and maltotriose but not from glucose. For its formation, a retro Claisen reaction of an enolization product of 1-amino-1,4-dideoxyhexosulose is proposed as the route to its formation. 1-Amino-1,4-dideoxyhexosulose could be formed by vinylogous α-elimination from the 2,3-enediol structure after Amadori rearrangement, favored by planar alignment of the bonds between C1 and C4. Subsequent rearrangement by keto - enoltautomerization leads to a 1-imino-3-keto structure. In this structure, attack of a hydroxyl anion, provided by water at neutral pH, could cause a splitting off of the C1. This reaction gives rise to formic acid or formamide and a pentose derivative, which reacts further to give 3-deoxypentosulose.

Full text of DOI:10.1021/jf011107o

J. Agric. Food Chem. 2002, 50, 1659−1664 1659
3-Deoxypentosulose: An r-Dicarbonyl Compound
Predominating in Nonenzymatic Browning of Oligosaccharides
in Aqueous Solution
ANKE HOLLNAGEL AND LOTHAR W. KROH*
Institut f u¨ r Lebensmittelchemie, Technische Universit a¨ t Berlin, Gustav-Meyer-Allee 25,
D-13355 Berlin, Germany
The thermal degradation of D-glucose, maltose, and maltotriose in aqueous solution was investigated
under caramelization (no glycine) and Maillard (with glycine) conditions. Degradation of the sugar
and R-dicarbonyls product was monitored. Under both caramelization and Maillard reaction conditions,
3-deoxypentosulose was the predominating R-dicarbonyl compound formed from maltose and
maltotriose. In the absence of an amino compound, however, 3-deoxypentosulose is formed in much
lower concentration. It was concluded that 3-deoxypentosulose is formed by a pathway specific for
oligo- and polysaccharides since this R-dicarbonyl is formed from the R-1f4 glucans such as maltose
and maltotriose but not from glucose. For its formation, a retro Claisen reaction of an enolization
product of 1-amino-1,4-dideoxyhexosulose is proposed as the route to its formation. 1-Amino-1,4-
dideoxyhexosulose could be formed by vinylogous R-elimination from the 2,3-enediol structure after
Amadori rearrangement, favored by planar alignment of the bonds between C1 and C4. Subsequent
rearrangement by keto-enoltautomerization leads to a 1-imino-3-keto structure. In this structure,
attack of a hydroxyl anion, provided by water at neutral pH, could cause a splitting off of the C1. This
reaction gives rise to formic acid or formamide and a pentose derivative, which reacts further to give
3-deoxypentosulose.
KEYWORDS: Nonenzymatic browning; Maillard reaction; r-dicarbonyl compounds; thermal degradation;
oligosaccharides; 3-deoxypentosulose
INTRODUCTION
The reaction of pentoses to form R-dicarbonyls was also
investigated. 3-Deoxy-pentosulose was synthesized first (4) and
was found in Maillard reaction mixtures, caramelization, and
pyrolysis of pentoses (5, 6). When Beck et al. (7) published
the formation of 1-amino-1,4-dideoxyhexosuloses, they specu-
lated that it might be formed preferentially from disaccharides
and secondary amines. However, 1,4-dideoxyhexosulose, as a
reaction product of this R-dicarbonyl, was also detected in
aqueous mixtures of glucose/â-alanine (8).
Recently, the formation of aminoreductones with a 4-deoxy-
moiety from lactose was found by Pischetsrieder et al. (9). The
reductone was detected in an early stage of the Maillard reaction
and was proposed to be the first stable intermediate of the
The Maillard reaction between monosaccharides and amino
acids has been investigated thoroughly in aqueous solution. Ledl
(1) reviewed the investigation that had been done and stressed
the role of compounds with an R-dicarbonyl structure as key
intermediates in the initial and advanced stage. The formation
of reaction products with an R-dicarbonyl moiety was reported
mainly for the Maillard reaction of monosaccharides. Besides
the formation of osuloses by the elimination of water, Weenen
(2) suggested pathways in which monosaccharides are degraded
into sugar fragments with an R-dicarbonyl moiety by retro
aldolization or R-cleavage. Hofmann (3) focused on the forma-
tion of glyoxal in the reaction of D-glucose and D-xylose with
L-alanine. Because it was formed to a considerable amount in
a very early stage of the reaction, they proposed its formation
via cleavage of glucosone and xylosone, respectively, which
they postulated to be oxidation products of the Schiff’s base.
Also, 3-deoxyhexosulose and 1-deoxyhexosulose were quanti-
fied, reaching a maximum concentration at a later stage of the
reaction.
1
-amino-1,4-dideoxyhexosulose.
The degradation of disaccharides in aqueous solutions during
the Maillard reaction has so far been described via two major
pathways. Kroh (10) reported the breakdown of oligo- and
polysaccharides in aqueous solution to 1,4-linked malto-
oligosaccharides and glucose by hydrothermolysis. Moreover,
fission of the acetal bond can be achieved by elimination
reactions after the glucose unit at the reducing end has reacted
to give heterocyclic rings (11).
In earlier investigations, it was demonstrated that o-phen-
ylenediamine is an appropriate means of monitoring R-dicar-
*
To whom correspondence should be addressed. Tel: +49 30 31472584.
Fax: +49-30-31472585. E-mail: lothar.kroh@tu-berlin.de.
1
0.1021/jf011107o CCC: $22.00 © 2002 American Chemical Society
Published on Web 02/07/2002

1660 J. Agric. Food Chem., Vol. 50, No. 6, 2002
Hollnagel and Kroh
bonyl production in the Maillard reaction (12). For the reaction
of oligosaccharides during Maillard reaction in quasi water free
systems, a specific degradation pathway was previously identi-
fied (13). The formation of specific R-dicarbonyl compounds
in the degradation of oligosaccharides in aqueous solution has
now been investigated. o-Phenylenediamine was used as a
trapping reagent in our investigations on the Maillard reaction
in a quasi water free system.
MATERIALS AND METHODS
Thermal Treatment. Caramelization: Mono-, Di-, or Trisaccharide
with o-Phenylenediamine. Samples of 0.5 mL of 0.25 M solution of
each sugar D-glucose (Merck, water free), maltose monohydrate
Figure 1. Formation of R-dicarbonyls detected as quinoxalines in the
Maillard reaction (D-glucose/glycine/o-phenylenediamine, solid lines) and
in caramelization (D-glucose/o-phenylenediamine, dotted line). 1-DH,
(Merck), or maltotriose (Fluka, >93% high-performance liquid chro-
matography (HPLC)) and 0.25 M trapping reagent (o-phenylenediamine,
recrystallized from water) were heated for timed intervals up to 240
min at 100 ( 1 °C in sealed tubes by means of a thermoblock (behrotest
ET 1, behr Labor Technik). Experiments were carried out in duplicate.
After the samples were heated, the samples were cooled, and after
appropriate dilution (methanol; Merck, suprasolv), they were subjected
to HPLC and high-perfomance thin-layer chromatography (HPTLC).
Maillard Reaction: Mono-, Di-, or Trisaccharide with Glycine and
o-Phenylenediamine. Samples (0.5 mL) of 0.25 M solutions of
D-glucose, maltose monohydrate, or maltotriose, respectively, and 0.5
mL each of 0.25 M glycine (Serva) and 0.25 M o-phenylenediamine
were heated for timed intervals up to 240 min at 100 ( 1 °C in sealed
tubes. Samples were treated and analyzed as above.
1-deoxyhexosulose; 3-DH, 3-deoxyhexosulose; 1,4-DH, 1,4-dideoxyhexo-
sulose; all compounds as quinoxalines.
Di- and Trisaccharides. HPTLC plates were developed twice in an
AMD chamber (CAMAG) using an eluent mixture of chloroform,
methanol, and water (50/40/8; v/v/v) in which 2.5-2.8 mg of boric
acid was dissolved. Detection was performed in both systems with
diphenylamine/aniline/phosphoric acid as a spray reagent and heating
(120 °C for 5 min).
HPAEC/PAD. Instrumentation: pump (Dionex GP 40); column (2
×
PA-100, 250 mm × 4 mm i.d., Dionex); detector (PAD, Dionex);
oven (HPLC column oven 2155, Pharmacia); temperature, 25 °C;
injection volume, 20 µL; flow, 0.5 mL/min; eluent: solvent A, 0.15
M NaOH; solvent B, 1 M sodium acetate in 0.15 M NaOH; gradient,
Synthesis of Reference Material. The synthesis of 2-methyl-3-
(
1′,2′,3′-trihydroxypropyl)quinoxaline, 2-(2′,3′,4′-trihydroxybutyl)quin-
oxaline, 2-(arabino-1′,2′,3′,4′-tetrahydroxybutyl)quinoxaline, and 2-methyl-
-(2′,3′-dihydroxypropyl)quinoxaline was performed as described by
0
-10 min 100% A and 10-60 min 0-60% B.
3
Hollnagel and Kroh (13). The synthesis of 2-(2′,3′-dihydroxypropyl)-
RESULTS AND DISCUSSION
quinoxaline was performed with slight variations according to Glomb
1
(
14). H nuclear magnetic resonance (NMR) (Brucker AM 300, 300
To investigate the identity and the concentration of R-dicar-
bonyls, a trapping agent (o-phenylenediamine) was directly
added to the reaction mixture. With this approach, all of the
R-dicarbonyls formed could be analyzed in their accumulated
concentration. The trapping agent was added prior to the reaction
in order to trap all of the R-dicarbonyl compounds formed. In
preliminary experiments with post derivatization, a lower
concentration of R-dicarbonyls was detected. This method
further allows the difference due to the varying reactivity of
R-dicarbonyls, as was reported by Glomb and Pfahler (4), to
be minimized. Using reference material for quinoxalines of
hexosuloses as well as of sugar fragments with an R-dicarbonyl
moiety, a wide range of R-dicarbonyl compounds could be
determined quantitatively. To investigate the effect of the amino
acid on the formation of R-dicarbonyls only, experiments were
carried out with and without glycine whereas OPD was present
in both types of reaction mixtures.
Characterization of the complex Maillard reaction mixture
was performed by observation of (i) the formation of R-dicar-
bonyls, (ii) degradation of the starting carbohydrate, and (iii)
subsequent formation of carbohydrates with a lower degree of
polymerization (dp) than the starting compound by hydrother-
molysis.
Formation of r-Dicarbonyls from Mono- and Oligomeric
Carbohydrates. For the quantification of R-dicarbonyl com-
pounds formed, the reaction of D-glucose under caramelization
and Maillard reaction conditions was investigated as a starting
point for further interpretation of the results from oligosaccha-
rides. In the caramelization model, only glyoxal was detected
MHz) in CD OD: δ 3.09 (dd, 1H), 3.24 (dd, 1H), 3.61 (dd, 2H), 4.15
3
1
3
(m, 1H), 7.76 (m, 2H), 8.02 (m, 2H), 8.83 (s, 1H). C NMR (300
MHz) in CD OD: δ 41.05, 67.07, 72.9, 129.5, 129.62, 130.65, 131.4,
3
1
42.0, 143.18, 147.9, 156.97. Mass spectrum of the acetylated
+
compound (Hewlett-Packard 5989B): m/z 288 (0.03%, M ), 229 (5),
2
1
28 (4), 215 (0.2), 187 (15), 186 (10), 169 (42), 157 (24), 144 (37),
29 (3), 117 (4), 43 (100).
HPLC/DAD. Instrumentation: degasser (Degasys DG-13000, Knau-
er); pump (Shimadzu LC 10 AT); thermostat (Haake F3, Fisons); guard
column (Nucleosil 120-5 C18, Macherey-Nagel); column (Nucleosil 5
C
18, Macherey-Nagel, 250 mm × 4.6 mm i.d., 5 µm); detector (Kontron
4
40); flow, 1.0 mL/min; temperature, 30 °C; injection volume, 20 µL;
eluent: solvent A, methanol; solvent B, water (both HPLC grade);
detection wavelength, 320 nm, full scan 190-440 nm; gradient, 0-5
min 5% A, 5-25 min 5-50% A, 25-30 min 50-100% A, and 30-
4
0 min 100% A.
Gas Chromatography/Mass Spectroscopy (GC/MS) and Gas
Chromatography/Flame Ionization Detector (GC/FID). Extraction
and derivatization of the reference material was carried out according
to ref 12. GC/MS: Analytical GC was performed on a Hewlett-Packard
5
890 Series II gas-liquid chromatograph equipped with a Hewlett-
Packard 5989B mass spectrometer (in EI mode) and a fused-silica
capillary column DB-5HT (J&W; 30 m × 0.32 mm i.d., 0.1 µm film).
Carrier gas, He; detector/injector temperature, 280 °C; temperature
program, initial temperature 120 °C, held for 5 min, 120-200 °C at
1
0 °C/min, held at 200 °C for 5 min, 200-280 °C at 10 °C/min, held
at 280 °C for 9 min. Column effluents were analyzed by electron
ionization mass spectrometry (range: m/z 40-800 amu). The identities
of the separated compounds were confirmed by comparison with those
of independently synthesized standards.
HPTLC. Monosaccharides. HPTLC plates (Merck, Kieselgel 60,
(
(
as its quinoxaline) with a yield of 0.6 mol % after 240 min
Figure 1).
2
0 cm × 10 cm) were used; the eluent consisted of chloroform (Merck,
HPLC grade), acetic acid (Merck, 96%), methanol (Merck suprasolv),
and water (60/18/12.5/5; v/v/v/v); and the plates were developed twice
according to ref 10.
In the presence of glycine, the formation of R-dicarbonyls
was enhanced as expected. Glyoxal formation from D-glucose

R-Dicarbonyl Compounds in Nonenzymatic Browning
J. Agric. Food Chem., Vol. 50, No. 6, 2002 1661
Table 1. Formation of R-Dicarbonyls Detected as Quinoxalines (mol
Table 2. Formation of R-Dicarbonyls Detected as Quinoxalines (mol
%) in Caramelization (Maltotriose/o-Phenylenediamine; cara) and in the
Maillard Reaction (Maltotriose/glycine/o-Phenylenediamine; MR)a
%
) in Caramelization of Maltose (Maltose/o-Phenylenediamine)
time (min)
glyoxal
methylglyoxal
3-deoxypentosulose
glyoxal
3-deoxypentosulose
1,4-dideoxyhexosulose
3
0
0.03
0.07
0.20
0.29
0.31
0.02
0.07
0.03
0.02
0.02
0.04
0.06
0.23
0.45
0.63
time
(min)
6
0
cara
MR
cara
MR
cara
MR
1
1
2
20
80
40
3
0
0.02
0.06
0.16
0.27
0.36
0.09
0.18
0.45
0.52
0.59
0.01
0.05
0.19
0.35
0.52
0.07
0.28
0.92
1.73
2.42
nd
nd
nd
nd
nd
0.02
0.05
0.09
0.11
0.11
6
0
1
1
2
20
80
40
a
nd is not detected.
compounds were formed at considerably lower concentrations
than 3-deoxypentosulose. The starting pH measured prior to
reaction (D-glucose/glycine/o-phenylenediamine) was neutral
and dropped slowly over the reaction time to 5. When glycine
was added, glyoxal reached a concentration of 0.55 mol % after
2
1
40 min, and also, methylglyoxal formation was increased.
-Deoxyhexosulose and 3-deoxyhexosulose were formed under
these conditions as well but in relatively low concentrations of
.24 mol % each after 240 min (Figure 2). The production of
0
all detected R-dicarbonyl compounds except 3-deoxypentosulose
decreased after 120 min. In contrast to the reaction of maltose/
glycine/o-phenylenediamine in a quasi water free system, where
Figure 2. Formation of R-dicarbonyl compounds detected as quinoxalines
in the Maillard reaction of maltose/glycine/o-phenylenediamine. 1-DH,
1-deoxyhexosulose; 3-DH, 3-deoxyhexosulose; 3-DP, 3-deoxypentosulose;
1
,4-dideoxyhexosulose was formed to 18 mol % (13), this
1,4-DH, 1,4-dideoxyhexosulose; all compounds as quinoxalines.
R-dicarbonyl does not play an important role in aqueous
solution. Under the chosen reaction conditions, 3-deoxypento-
sulose is the predominating R-dicarbonyl, and it can be assumed
that 3-deoxypentosulose is formed from the reducing end of
disaccharides.
In caramelization models with maltotriose, similar observa-
tions were made as for maltose. In addition to glyoxal (0.36
mol % after 240 min), only 3-deoxypentosulose (0.52 mol %
after 240 min) could be identified and determined (Table 2).
The addition of glycine to the reaction mixture strongly
accelerated the formation 3-deoxypentosulose to give 2.4 mol
% after 240 min. Together with 3-deoxyhexosulose, only glyoxal
(0.6 mol % after 240 min) and a small amount of 1,4-
dideoxyhexosulose (0.1 mol % after 240 min) were formed
(Table 2). This observation underlines the importance of the
formation of 3-deoxypentosulose in the Maillard reaction of
oligosaccharides in aqueous solution.
increased to 1.6 mol %, and also, methylglyoxal was formed in
considerable quantities (1.2 mol % after 240 min; Figure 1).
The formation of glyoxal was fastest during the first 60 min;
subsequently, the rate of its formation decreased slowly.
However, the concentration of glyoxal and methylglyoxal still
increased throughout the last interval of the investigated time
period.
In contrast to caramelization, in the Maillard reaction, hexose
derivatives with an R-dicarbonyl moiety could also be deter-
mined. 1-Deoxyhexosulose was formed to 1.1 mol % after 240
min, whereas 3-deoxyhexosulose was formed in a slightly lower
amount at the same reaction time (0.9 mol %, Figure 1). The
formation of 1,4-dideoxyhexosulose was almost negligible with
0
.2 mol % after the observed time range. In general, the
formation of hexosuloses was sluggish in the early stage of the
reaction and increased after 120 min.
With disaccharides, solutions containing maltose and o-
phenylenediamine showed the formation of glyoxal (0.3 mol
Degradation of the Starting Compound by Caramelization
and Maillard Reaction. Degradation of D-glucose in models
resembling caramelization proceeded rather slowly as compared
to quasi water free reaction mixtures. This finding is consistent
with the relatively low concentrations of quinoxalines detected.
When glycine was present, D-glucose was degraded faster than
in absence of glycine. After 60 min, almost half of the D-glucose
was converted, whereas in caramelization at that reaction time
more than two-thirds of the carbohydrate remained unreacted.
After 240 min in Maillard reaction mixtures, (D-glucose/glycine/
o-phenylenediamine) there was still 35.7 mol % glucose present
in contrast to caramelization reaction mixtures, which contained
44.9 mol % D-glucose (Table 3). The slower degradation of
the carbohydrate in aqueous solution can be assigned to the high
dilution (0.25 M) of the starting compound, resulting in a low
reaction rate. This finding is consistent with the relatively low
concentrations of quinoxalines that were formed in aqueous
solution.
%
after 240 min; Table 1) and traces of methylglyoxal.
However, the HPLC chromatogram indicated that a so far
unidentified quinoxaline was predominating. The retention time
suggested that this quinoxaline should be less polar than that
of 3-deoxyhexosulose and more polar than that of 1,4-dideoxy-
hexosulose. GC/MS revealed that it was the quinoxaline of a
pentose derivative. Comparison with reference material obtained
by independent synthesis confirmed by several chromatographic
and spectroscopic measures that the observed peak has to be
assigned to the quinoxaline of 3-deoxypentosulose. It was
quantified with 0.63 mol % formed after 240 min in caramel-
ization of maltose (maltose/o-phenylenediamine; Table 1). After
an induction phase of 60 min, the formation of 3-deoxypento-
sulose proceeded at a more or less constant rate.
In the presence of glycine, the concentration of 3-deoxypen-
tosulose was significantly higher; after 240 min, 1.86 mol %
(
Figure 2) was formed. The shape of the graph suggests that
The degradation of maltose was not significantly effected by
glycine. Although conversion of the disaccharide proceeded
somewhat faster in the Maillard reaction, the values determined
its formation did not stop at that stage of the reaction, and the
reaction was still proceeding. Other detected R-dicarbonyl

1662 J. Agric. Food Chem., Vol. 50, No. 6, 2002
Hollnagel and Kroh
Table 3. Degradation (mol %) of D-Glucose, Maltose, and Maltotriose
in the Maillard Reaction (MR) with Glycine (D-Glucose/Glycine/
o-Phenylenediamine) and in Caramelization (cara) without Glycine
a
(
D-Glucose/o-Phenylenediamine)
D-glucose
maltose
maltotriose
time (min)
0
MR
cara
MR
cara
MR
cara
100
46
54
45
nr
100
71
68
59
55
45
100
52
53
51
43
33
100
67
57
53
43
34
100
41
38
36
29
28
100
71
63
55
44
49
3
0
6
0
1
1
2
20
80
40
34
a
nr is not resolved.
Table 4. Formation of D-Glucose in Caramelization of Maltose
Maltose/o-Phenylenediamine) and in the Maillard Reaction of Maltose
Maltose/Glycine/o-Phenylenediamine) in mol %
(
(
D-glucose from
maltose/
o-phenylenediamine
D-glucose from
maltose/glycine/
o-phenylenediamine
time
min)
(
3
60
0
7.58
7.83
5.41
7.39
1
1
2
20
80
40
9.34
11.22
11.41
9.60
10.01
9.68
Figure 3. Proposed reaction pathway for the formation of 3-deoxypen-
tosulose and formic acid.
for maltose after 240 min of reaction time did not differ
significantly (Table 3). However, with maltotriose, the enhanc-
ing effect of glycine on carbohydrate degradation was observed
as for D-glucose. Especially after 30 min, the difference was
very distinct, whereas in caramelization at that time 70.7 mol
%
was still present, in the Maillard reaction only 45.8 mol %
remained (Table 3). After the first hour in both cases, the rate
of degradation decreased markedly, which might be due to the
diminished concentration of starting compound. In the absence
of glycine, somewhat more than half of the maltotriose was
converted after 240 min (48.7 mol %). In Maillard reaction
mixtures, 28.5 mol % maltotriose was found in the reaction
mixture after 240 min. The degradation of the trisaccharide
proceeded at a similar rate as D-glucose, although monosac-
charides are expected to be more reactive. It can be assumed
that degradation of oligosaccharides in aqueous solution occurs
not only by reaction of the OH group at the reducing end with
the amino compound but also by hydrolysis of the glycosidic
bonds (15).
Hydrothermolysis of Oligomeric Carbohydrates. To elu-
cidate the relevance of hydrolytic cleavage, carbohydrates with
a lower dp than the starting oligosaccharides were determined.
In caramelization mixtures containing maltose, D-glucose was
formed in considerable quantities, up to 11.4 mol % after 240
min (Table 4). That means that at least 5.7% of the maltose
was converted by hydrolytic cleavage at the end of the observed
time range.
In the Maillard reaction, about 10 mol % D-glucose was
detected as well (Table 4), but it has to be considered that now
other degradation mechanisms besides hydrolysis take place,
which include the reaction of the glucose unit at the reducing
end to 3-deoxypentosulose and formic acid as well as to
aminoreductones (9). Therefore, it can assumed that under
Maillard reaction conditions more disaccharide is degraded by
fission of the glycosidic bond than in caramelization. Moreover,
in the presence of glycine, D-glucose is able to react in terms
of the Maillard reaction given the higher reactivity as compared
Figure 4. Formation of maltose in the Maillard reaction from maltotriose/
glycine/o-phenylenediamine (solid line) and in caramelization maltotriose/
o-phenylenediamine (dotted line).
with disaccharides, reducing the detected value of glucose
stronger than in the absence of an amino acid.
The cleavage of the glycosidic bond in maltotriose gives rise
to maltose. In caramelization, maltose was formed by hydro-
thermolysis to maximal 6 mol % maximum after 240 min
(Figure 4). With participation of glycine, maltose concentration
rose to 13.7 mol % at 240 min reaction time (Figure 4). This
increase of maltose formation in the Maillard reaction cannot
be assigned only to the formation of formic acid and 3-deoxy-
pentosulose at the reducing end of the sugar molecule. Ad-
ditionally, it has to be considered that oligosaccharides can also
be degraded by the formation of aminoreductones (9); the
R-dicarbonyl structure is not necessarily involved.
For better understanding of nonenzymatic browning, it is
important to study the occurring reaction pathways in detail.
Information about the mechanisms that lead to the formation
of the various R-dicarbonyls could be of particular value.
The formation of sugar fragments with an R-dicarbonyl
moiety was higher under Maillard reaction conditions than
during caramelization from glucose. Whereas they are formed
by retro aldolization of the sugar molecule itself under cara-
melization conditions, in the Maillard reaction, the Schiff’s base
as well as the ene-aminol can be starting compounds for
cleavage reactions. Moreover, the amino acid can act as a base
and catalyze retro aldolization reactions (2).

R-Dicarbonyl Compounds in Nonenzymatic Browning
J. Agric. Food Chem., Vol. 50, No. 6, 2002 1663
In both systems, an increase of the glyoxal concentration was
observed throughout the reaction time when the trapping agent
was present. In contrast, Hofmann et al. (3) observed a
maximum concentration of glyoxal at a very early stage of the
reaction with the observed yields of quinoxalines much lower
as compared to the data presented here. Hofmann and co-
workers suggested that glyoxal is formed via glucosone;
however, under the chosen reaction conditions, no glucosone,
as its corresponding quinoxaline, could be detected by GC/MS.
The glyoxal we detected should therefore be formed by
pathways suggested by Weenen (2).
neutral pH (and in weak acid pH), or the participation of the
carboxylate functionality of glycine could cause a splitting off
of the C1. This reaction can be characterized as a retro Claisen
ester condensation, giving rise to formic acid or formamide and
a pentose derivative. The first reaction product that is suggested
to be produced by this reaction pathway, formic acid, is
frequently found in the Maillard reaction and causes, among
other formed acids, the drop in pH that is typical for the Maillard
reaction. The essential role of hydroxyl ion or carboxylate ion
in retro Claisen ester condensation could explain why this
reaction only predominates in aqueous systems (Figure 3).
The subsequent reaction of this pentose derivative can lead
to 3-deoxypentosulose as it was detected in reaction mixtures
of maltose. Other possible reaction products of this derivative
such as 3-deoxypentulose, as ketose derivatives, or furfuryl
alcohol have been found in model systems and food, e.g., in
heat-treated milk. For example, Troyano et al. (18) published
the formation of 3-deoxypentulose in milk and in alkaline model
mixtures containing maltose or cellobiose. Formic acid was also
proposed as a secondary product. Furfuryl alcohol, another
possible product of the further conversion of the pentose
derivative, was detected by Patton (19), who found furfuryl
alcohol in milk as well as in neutral and alkaline model systems
containing lactose or lactose and lysine. The formation was
postulated to occur via the formation of 3-deoxyhexosulose at
the reducing end of lactose and R-dicarbonyl cleavage yielding
a galactosylated pentose derivative that eliminates the galactosyl
residue and gives rise, after ring closure and dehydration, to
furfuryl alcohol. The pH necessary for these reactions ranged
between 6 and 8 and 4-6, respectively, which might be
accomplished by the increase of hydrogen ion concentration
throughout the Maillard reaction. In caramelization, one could
imagine that a similar reaction pathway occurs also in the
absence of an amino acid. Enolization, which is essential to
form the 2,3-ene-diol as a precursor to the â-elimination, is also
known to occur in the absence of the amino acid under the given
reaction conditions. However, this reaction occurs much more
slowly than in the presence of the amino compound, acting as
a catalyst in enolization and â-elimination.
In contrast to caramelization, in the Maillard reaction, hexose
derivatives with an R-dicarbonyl moiety could also be deter-
mined. 1-Deoxyhexosulose was formed at 1.1 mol % after 240
min, whereas 3-deoxyhexosulose was formed in a slightly lower
amount (0.9 mol %; Figure 1). The formation of 1,4-dideoxy-
hexosulose was almost negligible, with 0.2 mol % after the
observed time period.
Hofmann (3) obtained different results for the reaction of
D-glucose with alanine. They found that four times more
3
-deoxyhexosulose was formed than 1-deoxyhexosulose, when
the trapping reagent was added after thermal treatment. How-
ever, in this case, differences in the reactivity of the various
R-dicarbonyls might not be fully accounted for. Nevertheless,
when the trapping reagent was present throughout the entire
reaction time, Feather et al. and Nedvidek (16, 17) obtained
results that confirm the present data. More 1-deoxyhexosulose
than 3-deoxyhexosulose was found in mixtures of fructosyl-
glycine and D-glucose/propylamine and fructosylpropylamine
reaction mixtures as well (17). The fact that 1,4-dideoxyhexo-
sulose was formed in such small amounts suggests that Strecker
degradation or other reduction steps, which are necessary for
the formation of 1,4-dideoxyhexosulose (13), are not favored
in aqueous solution or that these reactions proceed more slowly
than the reaction with o-phenylenediamine. In general, the
formation of hexosuloses was sluggish in the early stage of the
reaction and increased after 120 min.
In comparison with the experiments under quasi water free
conditions (13), yields of the quinoxalines of the D-glucose
model system were relatively low. This effect is most probably
due to the higher dilution of starting material reducing the
reaction rate. Furthermore, the product spectrum was shifted;
while in quasi water free mixtures, 1,4-dideoxyhexosulose,
methylglyoxal, and 3-deoxyhexosulose were formed, and in
aqueous solution, glyoxal was the predominating R-dicarbonyl,
followed by methylglyoxal and 1-deoxyhexosulose.
Reactions of oligosaccharides in aqueous solution proceed
via various pathways. In the absence of an amino compound,
the formation of R-dicarbonyls plays a minor role; therefore,
subsequent reaction products of such reactive intermediates are
formed to a minor extent throughout caramelization. Accord-
ingly, the formation of brown pigments and flavor is rather slow
when oligosaccharides are heated in an aqueous solution.
The addition of glycine enhanced the formation of R-dicar-
bonyl compounds markedly. For the reaction of oligosaccha-
rides, the formation of 3-deoxypentosulose was described as
the most abundant R-dicarbonyl, and a reaction pathway was
proposed yielding the R-dicarbonyl, a carbohydrate with a lower
dp, and formic acid as degradation products. Besides the
formation of formic acid, the generation of some pentose
derivatives such as 3-deoxypentulose or furfuryl alcohol can
be explained by the proposed pathway. Under the chosen
reaction conditions, glyoxal is most probably not formed from
glucosone but is rather a reaction product of retro aldolization.
Under caramelization and Maillard reaction conditions,
-deoxypentosulose is the predominating R-dicarbonyl from
3
maltose. The presence of glycine strongly enhanced its forma-
tion. The fact that it is not formed from D-glucose indicates
that the glycosidic bond causes a change in the type of
degradation reactions taking place. Taking into account that
fission of the hexose between C1 and C2 must have occurred
for the formation of the pentose derivative, the deoxy moiety
at C3 in the reaction product was originally the site of the
glycosidic linkage.
A feasible reaction pathway towards 3-deoxypentosulose
could be the formation of 1-amino-1,4-dideoxyhexosulose by
vinylogous â-elimination from the 2,3-ene-diol structure after
Amadori rearrangement, favored by planar alignment of the
bonds between C1 and C4. Subsequent rearrangement by keto-
enoltautomerization leads to a 1-imino-3-keto structure. In this
structure, attack of a hydroxyl anion, provided by water at
Because of the reaction with the amino compound, the
formation of carbohydrates with a lower dp increased. Appar-
ently, in the presence of an amino compound, conversion of
the oligosaccharide and formation of color are not only
accelerated as compared with caramelization but also there are
some qualitative differences in the spectrum of intermediates
formed in caramelization and the Maillard reaction. By the

1664 J. Agric. Food Chem., Vol. 50, No. 6, 2002
Hollnagel and Kroh
reaction with the amino acid, certain degradation pathways are
favoured in the degradation of oligosaccharides, which are
significant neither in the Maillard reaction of monosaccharides
nor in the caramelization of mono- or oligosaccharides. These
findings are relevant for the understanding of changes of food
components occurring throughout food processing, such as in
milk and other dairy products.
(9) Pischetsrieder, M.; Schoetter, C.; Severin, T. Formation of an
R
aminoreductone during the Maillard reaction of lactose with N -
acetyllysine or proteins. J. Agric. Food Chem. 1998, 46, 928-
9
31.
(
10) Kroh, L. W.; Jalyschko, W.; H a¨ seler, J. Nonvolatile reaction
products by heat-induced degradation of R-glucans. Part I:
analysis of oligomeric maltodextrins and anhydrosugars. Starch/
St a¨ rke 1996, 48, 426-433.
11) Kramh o¨ ller, B.; Pischetsrieder, M.; Severin, T. Maillard reactions
of dextrins and starch. Z. Lebensm.-Unters. Forsch. 1993, 197,
(
ACKNOWLEDGMENT
2
27-229.
The authors express their sincere gratitude to Mrs. Anette
Bergh a¨ user for excellent technical assistance. The authors are
also indebted to Prof. Dr. Christian Vogel, University Rostock,
for valuable discussions on the reaction mechanisms.
(
12) Hollnagel, A.; Kroh, L. W. Formation of R-dicarbonyl fragments
from mono- and disaccharides under caramelization and Maillard
reaction conditions. Z. Lebensm.-Unters. Forsch. A 1998, 207,
50-54.
13) Hollnagel, A.; Kroh, L. W. Degradation of oligosaccharides in
nonenzymatic browning by formation of R-dicarbonyl com-
pounds via a “peeling off” mechanism. J. Agric. Food Chem.
2000, 48, 6219-6226.
(14) Glomb, M. A. Synthese, Strukturaufkl a¨ rung und Nachweis
reaktiver Pentoseabbauprodukte. Thesis, Stuttgart, 1992.
(15) H a¨ seler, J. Nichtfl u¨ chtige polare Verbindungen des thermisch
induzierten Abbaus von R-Glucanen. Thesis, Berlin, 2000.
(16) Feather, M. S.; Mossine, V.; Hirsch, J. The Use of Aminoguani-
dine To Trap and Measure Dicarbonyl Intermediates Produced
During the Maillard Reaction. In Chemical Markers for Pro-
cessed and Stored Foods; Lee, T.-C., Kim, H.-J., Eds.; ACS
Symposium Series 631; American Chemical Society: Washing-
ton, DC, 1996.
(
LITERATURE CITED
(
1) Ledl, F.; Schleicher, E. Die Maillard-Reaktion in Lebensmitteln
und im menschlichen K o¨ rper- neue Ergebnisse zu Chemie,
Biochemie und Medizin. Angew. Chem. 1990, 102, 597-734.
2) Weenen, H. Reactive intermediates and carbohydrate fragmenta-
tion in Maillard chemistry. Food Chem. 1998, 62, 393-401.
3) Hofmann, T.; Bors, W.; Stettmaier, K. Studies on radical
intermediates in the early stage of the nonenzymatic browning
of carbohydrates and amino acids. J. Agric. Food Chem. 1999,
(
(
4
7, 379-390.
(4) Glomb, M. A.; Pfahler, C. Synthese eines bedeutenden Maillard-
Intermediats beim Abbau von Hexosen. Lebensmittelchemie
1
999, 53, 16-17.
(17) Nedvidek, W. Zuckerabbauprodukte mit R-Dicarbonylstruktur.
Thesis, Stuttgart, 1992a.
(
5) Lederer, M.; Glomb, M.; Fischer, P.; Ledl, F. Reactive Maillard
intermediates leading to colored products- 5-hydroxymethyl- and
(18) Troyano, E.; Olano, A.; Jimeno, M. L.; Sanz, J.; Martinez-Castro,
I. Isolation and characterization of 3-deoxypentulose and its
determination in heated milk. Dairy Res. 1992, 59, 507-515.
(19) Patton, S. Studies of heated milk III. Mode of formation of certain
furan compounds. Dairy Sci. 1950, 33, 904-910.
5-(1,2-dihydroxyethyl)-3(2H)-furanone. Z. Lebensm.-Unters. For-
sch. 1993, 197, 413-418.
(
6) Kato, K.; Komorito, H. Pyrolysis of cellulose. V. Isolation and
identification of 3-deoxyglycosones produced from D-glucose,
D-xylose, and R-cellulose by heating. Agric. Biol. Chem. 1968,
3
2, 715-720.
Received for review August 14, 2001. Revised manuscript received
November 14, 2001. Accepted December 3, 2001. The study was carried
out with financial support from the Commission of the European
Communities, Agriculture and Fisheries (FAIR) specific program,
FAIR-CT96-1080, “Optimization of the Maillard Reaction. A Way to
Improve Quality and Safety of Thermally Processed Food.”
(
7) Beck, J.; Ledl, F.; Severin, T. Formation of 1-deoxy-D-erythro-
2
1
,3-hexodiulose from Amadori compounds. Carbohydr. Res.
988, 177, 240-243.
(
8) Nedvidek, W.; Ledl, F.; Fischer, P. Detection of 5-hydroxym-
ethyl-2-methyl-3(2H)-furanone and of R-dicarbonyl compounds
in reaction mixtures of hexoses and pentoses with different
amines. Z. Lebensm.-Unters. Forsch. 1992b, 194, 222-228.
JF011107O