Degradation of the Amadori compound N-(1-deoxy-d-fructos-1-yl)glycine in aqueous model systems

Article abstract of DOI:10.1021/jf025561j

The fate of the Amadori compound N-(1-deoxy-D-fructos-1-yl)glycine (DFG) was studied in aqueous model systems as a function of time and pH. The samples were reacted at 90 °C for up to 7 h while maintaining the pH constant at 5, 6, 7, or 8. Special attenti

Full text of DOI:10.1021/jf025561j

5472 J. Agric. Food Chem. 2002, 50, 5472−5479
Degradation of the Amadori Compound
N-(1-Deoxy-D-fructos-1-yl)glycine in Aqueous Model Systems
TOMAS DAVIDEK,† NATHALIE CLETY,† SANDRA AUBIN,† AND IMRE BLANK*,‡
Centres R&D Nestle´ SAS, Amiens, P.O. Box 47, 80800 Aubigny, France, and Nestle´ Research
Center, P.O. Box 44, 1000 Lausanne 26, Switzerland
The fate of the Amadori compound N-(1-deoxy-D-fructos-1-yl)glycine (DFG) was studied in aqueous
model systems as a function of time and pH. The samples were reacted at 90 °C for up to 7 h while
maintaining the pH constant at 5, 6, 7, or 8. Special attention was paid to the effect of phosphate on
the formation of glycine and the parent sugars glucose and mannose, as well as formic and acetic
acid. These compounds and DFG were quantified by high-performance anion-exchange chroma-
tography. The rate of DFG degradation increased with pH. Addition of phosphate accelerated this
reaction, particularly at pH 5-7. The rate of glycine formation increased with pH in both the absence
and presence of phosphate. High glycine concentrations (60-70 mol %) were obtained, preferably
at pH 6-8 with phosphate. However, the yield of glycine formed from DFG decreased at the advanced
reaction stage for all pH values studied, both in water and in phosphate buffer. The rate of parent
sugar formation increased from pH 5 to pH 7 in the absence of phosphate, leading to glucose and
mannose in a constant ratio of 7:3. Addition of phosphate accelerated this reaction, yielding up to
18% parent sugars, most likely formed by reverse Amadori rearrangement. The formation rate of
acetic and formic acid increased with increasing pH. The sum of both acids attained 76 mol %.
However, the acetic acid concentrations were much higher than those of formic acid.
KEYWORDS: Maillard reaction; reverse Amadori rearrangement; Amadori compound; N-(1-deoxy-D-
fructos-1-yl)glycine; anion-exchange chromatography
INTRODUCTION
(5). In parallel to these pathways, other R-dicarbonyls can be
formed by enolization. For example, transition-metal-catalyzed
oxidation of 1,2-enaminol IV can lead via pathway C to osones
such as glucosone (IX) (6, 7). Pathway E gives rise to the
1-amino-1,4-dideoxy-2,3-diulose (XIII) by elimination of the
C4-OH group of the 2,3-enediol VI (8). If the iminoketone VIII
formed by oxidation of 1,2-enaminol IV is not hydrolyzed, then
Strecker aldehydes can be formed in the course of pathway C
by direct oxidative degradation of the Amadori compound and
decarboxylation of X, as recently proposed by Hofmann and
Schieberle (9).
Amadori compounds are N-substituted 1-amino-1-deoxyke-
toses, V (Figure 1), representing an important class of Maillard
intermediates (1). They are formed in the initial phase of the
Maillard reaction by Amadori rearrangement of the correspond-
ing N-glycosylamines II (2), the latter obtained by condensation
of amino acids and aldoses such as glucose (I) as shown in
pathway A. The importance of Amadori compounds stems from
the fact that their formation as well as decomposition can be
initiated under physiological conditions and during food pro-
cessing and storage. Thus, the formation of Amadori compounds
represents a pathway requiring less energy for sugar degradation
as compared to caramelization. The chemistry of Amadori
compounds has recently been reviewed, including analysis,
synthesis, and physical and spectrometric properties (3, 4).
One of the important characteristics of Amadori compounds
is their tendency to easily undergo enolization (Figure 1).
Degradation of the Amadori compound V by 1,2-enolization
(pathway B) and 2,3-enolization (pathway D) leads to the
formation of 3-deoxy-2-hexosulose (VII) and 1-deoxy-2,3-
hexodiulose (XII), respectively, as already suggested by Hodge
Other mechanisms of degradation of Amadori compounds
involve retro-aldol cleavage of further aminated Amadori
compounds yielding diimine derivatives of methylglyoxal (10)
and dehydration of Amadori compounds in the cyclic form (11).
Besides the above-mentioned reactions, Amadori compounds
can also act as nucleophiles and react with another sugar
molecule to form diketosyl derivatives (1).
Despite significant effort in the past decades, many questions
still persist concerning the degradation of Amadori compounds.
Among others, the reversibility of Amadori rearrangement and
formation of carboxylic acids at early stages of the Maillard
reaction need further clarification. The formation of parent
sugars from the corresponding Amadori compound has been
reported under physiological conditions (12, 13) and more
* To whom correspondence should be addressed (telephone, +21/785-
8607; fax, +21/785-8554; e-mail, imre.blank@rdls.nestle.com).
† Centres R&D Nestle´ SAS.
‡ Nestle´ Research Center.
10.1021/jf025561j CCC: $22.00 © 2002 American Chemical Society
Published on Web 08/01/2002

Degradation of Fructosylglycine
J. Agric. Food Chem., Vol. 50, No. 19, 2002 5473
Figure 1. Reaction of glucose (I) and glycine leading to the Amadori compound N-(1-deoxy-D-fructos-1-yl)glycine (V, open-chain form) and related
degradation reactions.
recently under pyrolysis conditions (14). However, no informa-
tion is available about the yields of parent sugars formed under
food processing conditions or about the effect of reaction
conditions on the yields.
to the effect of pH and phosphate on the formation of the parent
sugars glucose and mannose as well as acetic and formic acid.
MATERIALS AND METHODS
Generation of high levels of acetic and formic acid at early
stages of the degradation of protein-bound Amadori compounds
has recently been reported by Brands and van Boekel (15). C1-
C2 cleavage of 3-deoxy-2-hexosulose (VII) and C2-C3 cleavage
of 1-deoxy-2,3-hexodiulose (XII) were proposed as mechanisms
leading to the formation of formic acid and acetic acid,
respectively. However, as no labeling experiments were per-
formed and neither 2-deoxypentose (remaining C5 fragment of
VII) nor erythrose (remaining C4 fragment of XII) were
detected, further experiments are necessary to confirm the
proposed mechanisms.
Chemicals. The following compounds were obtained commer-
cially: glycine, D-mannose, sodium acetate (Merck, Darmstadt, Ger-
many); D-glucose, acetic acid, and formic acid (Sigma-Aldrich,
Steinheim, Germany); maltitol (Fluka, Buchs, Switzerland); sodium
hydroxide (NaOH), 46-48% solution (Fisher Scientific, Pittsburgh,
PA); sodium hydrogen carbonate (Prolabo, Paris, France). DFG was
prepared from D-glucose and glycine as reported by Staempfli et al.
(16). The solutions and eluents were prepared using ultrapure deionized
water (specific resistance g18.2 MΩ‚cm) from a Milli-Q system
(Millipore, Bedford, MA). NaOH solutions used as eluents were
prepared by diluting a carbonate-free 46-48% (w/w) NaOH solution
in water previously degassed with helium gas.
The aim of our study was to gain more insight into the fate
of the Amadori compound N-(1-deoxy-D-fructos-1-yl)glycine
(DFG) in aqueous model systems. Special attention was paid
Model Reactions. Solutions of DFG (15 mmol) in water (150 mL)
or in a phosphate buffer (150 mL, 100 mmol/L) were heated at 90 °C
for 7 h. Experiments were performed at constant pH (5, 6, 7, or 8),

5474 J. Agric. Food Chem., Vol. 50, No. 19, 2002
Davidek et al.
Table 1. Integrated Amperometry Waveform
electrode
time (s)
potential (V)
comment
0
−0.1
−0.1
0.35
0.35
−0.1
−0.1
0.9
0.9
−0.9
−0.9
0.2
0.3
0.4
0.5
0.7
0.71
0.9
0.91
1.0
start of signal integration
stop of signal integration
measured at 90 °C, using a pH-stat device (Metrohm, Herisau,
Switzerland) filled with a NaOH solution (3 mol/L) and equipped with
an electrode permitting measurements at 90 °C. Aliquots of the reaction
mixture were taken at regular time intervals and analyzed by high-
performance anion-exchange chromatography (HPAEC) and UV/vis
spectroscopy.
Quantification of DFG, D-Glucose, D-Mannose, and Glycine. An
aliquot of the reaction mixture (1 mL) was diluted 1000 times with
deionized water, filtered through a PVDF filter (polyvinylidene fluoride,
0.22 µm/25 mm), and analyzed by HPAEC. The analyses were
performed using the DX-500 system from Dionex (Sunnyvale, CA)
composed of an autosampler (model AS-50 with a 50 µL sample loop),
a gradient pump (model EG-40) with on-line degas, and an electro-
chemical detector (model ED-40). The separation was accomplished
on a 250 × 4 mm i.d. CarboPac PA1 anion-exchange column (Dionex)
and a 50 × 4 mm i.d. CarboPac PA1 guard column (Dionex) using the
following gradient: starting with a mixture (0/96/4, v/v/v) of sodium
acetate (300 mmol/L), water, and NaOH (300 mmol/L), the sodium
acetate was first increased to 3% (3/93/4, v/v/v) in 1 min and kept
isocratic for 28 min and then increased to 80% (80/16/4, v/v/v) in 5
min and kept isocratic for 5 min. Each analytical cycle was followed
by cleaning and regeneration of the column with NaOH (300 mmol/L)
for 15 min and equilibration of the column with the initial gradient
condition for 10 min. The flow rate was kept constant at 1 mL/min
throughout the program. D-Glucose (RT 18.07 min), D-mannose (RT
19.47 min), glycine (RT 27.35 min), DFG (RT 36.53 min), and maltitol
(internal standard, RT 11.48 min) were quantified with an electrochemi-
cal detector working in the integrated amperometry mode and equipped
with a gold working electrode. The waveform used is shown in Table
1. All compounds were quantified on the basis of calibration curves
by comparing the peak areas with those of standard solutions containing
known amounts of pure compounds. All samples were analyzed in
duplicate.
Figure 2. Thermal degradation of N-(1-deoxy-D-fructos-1-yl)glycine (DFG)
in water (s) and in phosphate buffer (‚‚‚) at pH 5 (b), pH 6 (9), pH 7
(2), and pH 8 ([). The coefficient of variation was lower than 10% for
concentrations above 1.5 mmol/L.
the same initial concentration of DFG (0.1 mol/L). All experi-
ments were carried out at constant pH (5, 6, 7, or 8) to obtain
comparable data between experiments performed in water and
in phosphate buffer. Quantification of DFG was achieved by
HPEAC coupled with an electrochemical detector working in
integrated amperometry mode. This method, developed in our
laboratories, will be published elsewhere in detail. It permits
quantification of not only DFG but also its parent compounds
(glucose, mannose, and glycine) in the same analytical run.
Degradation of N-(1-Deoxy-D-fructos-1-yl)glycine (DFG).
The rate of DFG degradation in water was relatively slow at
pH 5 (Figure 2). After 7 h of heating at 90 °C, about 70% of
DFG remained unreacted. Increasing the pH considerably
favored the degradation rate of DFG. While 6 h was required
to degrade 25% of DFG at pH 5, only 135, 40, and 20 min
were needed to reach the same degree of degradation at pH 6,
7, and 8, respectively.
Quantification of Acetic and Formic Acid. An aliquot of the
reaction mixture (1 mL) was diluted 50 times with deionized water
and analyzed by HPAEC. The analyses were performed using a DX-
120 Dionex system (Dionex Corp., Sunnyvale, CA) equipped with an
autosampler (model AS-40), a 25 µL sample loop, and a conductivity
detector. The anion self-regenerating suppressor (ASRS-Ultra, 4 mm)
was working in the autosuppression recycle mode. The isocratic
separation was accomplished on a 250 × 4 mm i.d. IonPac AS-14
anion-exchange column (Dionex) and a 50 × 4 mm i.d. IonPac AG-
14 guard column (Dionex) using sodium hydrogen carbonate (2.5 mmol/
L) as the mobile phase at a flow rate of 1.2 mL/min. Acetic acid (RT
11.07 min) and formic acid (RT 11.81 min) were quantified on the
basis of calibration curves by comparing the peak areas with those of
standard solutions containing known amounts of pure compounds. All
samples were analyzed in duplicate.
A similar trend was observed for the effect of pH on the
degradation of DFG in the presence of phosphate buffer.
However, the rate of DFG degradation increased with pH,
particularly in the pH range 5-7, and there was only a small
increase between pH 7 and pH 8. Interestingly, the addition of
an equimolar amount of phosphate showed a similar effect on
the decomposition of DFG as increasing the pH of the
unbuffered solution by one unit. These findings are in agreement
with data reported in the literature (17). The effect of phosphate
on the reaction of glucose with glycine was described as a
general base catalysis with an optimum in the pH range between
5 and 7. Dihydrogen phosphate ions, which act as a base
abstracting a proton during the Amadori rearrangement, were
found to be the principal catalytic species leading to enhanced
degradation of Amadori compounds.
Browning of Reaction Mixtures. UV/vis spectrometry measure-
ments were performed on a Uvikon UV-942 (Kontron Instruments,
Milan, Italy). The absorbance of the 50 times diluted reaction mixtures
was measured at λ ) 420 nm in a 10 mm cell.
Formation of Glycine. The rate of glycine formation
increased with pH in both the absence and presence of
phosphate. However, there was almost no difference between
the rate of formation of glycine at pH 7 and 8 in the presence
of phosphate (Figure 3). This is in line with the data shown in
Figure 2 as glycine is liberated during the decomposition of
RESULTS AND DISCUSSION
The effect of pH and phosphate on the decomposition of N-(1-
deoxy-D-fructos-1-yl)glycine (DFG) was studied at 90 °C using

Degradation of Fructosylglycine
J. Agric. Food Chem., Vol. 50, No. 19, 2002 5475
Table 2. Amount of Glycine Generated from
N-(1-Deoxy-D-fructos-1-yl)glycine (DFG) at Different Reaction Stages
amount of glycine
amount of glycine
(mol/mol of DFG)
generated in water
(mol/mol of DFG)
generated in
phosphate buffer
amount
of DFG
(%) de-
composed pH 5 pH 6 pH 7 pH 8 pH 5 pH 6 pH 7 pH 8
5
25
50
75
95
0.74 0.74 0.80 0.84 0.88 0.88 0.80 0.78
0.54 0.62 0.77 0.84 0.86 0.85 0.81 0.79
0.46 0.69 0.77 0.79 0.78 0.79 0.77
0.64 0.68
0.55 0.59
0.74 0.74 0.69
0.62 0.68 0.60
not surprising as new compounds, such as R-dicarbonyls and
R-hydroxycarbonyls, are formed with progressing reaction time
and glycine is more importantly involved in reactions such as
Strecker degradation.
Similarly, the yield of glycine decreased as the pH value of
the phosphate buffer increased. These data are coherent with
the fact that increasing pH favors retro-aldol reactions, which
leads to the formation of short-chain R-dicarbonyls or R-hy-
droxycarbonyls such as glyoxal, methylglyoxal, and glycol
aldehyde (7, 19). These compounds are known to be highly
reactive in Strecker degradation of amino acids (23, 24) and
also in other reactions such as the formation of 1,4-dialkyl-
pyrazinium radicals (7). Similarly to the observation that the
formation of carboxymethyllysine increased with phosphate
concentration and pH (13), the formation of carboxymethyl-
glycine (CMG) is another reaction that could cause decrease in
yield of glycine with increasing pH. However, this reaction does
not seem to be significant under our reaction conditions as
enhanced formation of CMG should lead to lower yields of
glycine in phosphate buffer as compared to water. However,
opposite results were found as shown in Table 2.
Contrary to the data obtained in the presence of phosphate
buffer, the yield of glycine in water increased with increasing
pH (Table 2). For example, when 25% of DFG was reacted,
the yield of glycine reached 84% at pH 8 but only 54% at pH
5. These results are surprising, as the opposite trend would be
expected as explained above. Comparison of the yields obtained
in the presence and in the absence of phosphate buffer indicates
that phosphate slowed the degradation of glycine at lower pH
values (mainly pH 5 and 6). These results could indicate that
reactions such as decarboxylation of the Schiff base III and
subsequent hydrolysis gain in importance in the absence of
phosphate (Figure 1, pathway G). The relative importance of
this reaction should decrease at higher pH values as the Schiff
base formation is less favored. If phosphate decreases the relative
importance of pathway G, then the increase of the yield of
glycine in the presence of phosphate should be more visible at
lower pH values, which was observed in this study. However,
more experiments are necessary to validate these hypotheses.
Formation of Glucose and Mannose. The formation of
glucose and mannose was studied in parallel to the degradation
of DFG. The ratio of glucose to mannose was 7:3 and was stable
under different reaction conditions (Table 3). Therefore, only
the sums of glucose and mannose concentrations are presented.
In the absence of phosphate, the formation rate of these sugars
increased with pH in the pH range 5-7 (Figure 4). Further
increase to pH 8 led to lower concentrations of glucose and
mannose as compared to pH 7. This could be explained by the
lower stability of the sugars formed. However, a decrease in
the formation of glucose and mannose from DFG at pH 8 cannot
be excluded. The data also illustrate that addition of phosphate
Figure 3. Formation of glycine during thermal degradation of N-(1-deoxy-
D-fructos-1-yl)glycine in water (s) and in phosphate buffer (‚‚‚) at pH 5
(b), pH 6 (9), pH 7 (2), and pH 8 ([). The coefficient of variation was
lower than 10% for concentrations above 2.5 mmol/L.
DFG. Formation of glycine from DFG mainly occurs by 1,2-
enolization (pathway B) and 2,3-enolization (pathway C),
yielding 3-deoxy-2-hexosulose (VII) and 1-deoxy-2,3-hexodiu-
lose (XII), respectively (Figure 1). In theory, 1 mol of DFG
should yield 1 mol of glycine. However, the yield of glycine
was generally lower, especially for the advanced reaction stages.
This indicates that part of the liberated glycine reacted further
with other compounds present in the reaction mixture (e.g.,
R-dicarbonyls and R-hydroxycarbonyls). These reactions may
include Strecker degradation and formation of imidazolium
compounds (18) and 1,4-dialkylpyrazinium radicals (7) as well
as incorporation of glycine into melanoidins.
However, glycine can also be degraded by mechanisms that
do not include glycine liberation from DFG. For example, the
mechanism reported by Namiki and co-workers (19-21)
involves cleavage of the sugar moiety of Schiff base III (Figure
1, pathway F), leading to a C2 fragment (XIV, glycolaldehyde
imine in the enol form), which upon dimerization and oxidation
forms a 1,4-dialkylpyrazinium radical (XV). Alternatively, the
Schiff base III can also be decarboxylated, resulting in XVI,
which hydrolyzes to form a Strecker aldehyde as shown in
pathway G (22). Another mechanism (9) involves direct
degradation of the Amadori compound in the presence of oxygen
via oxidation of enaminol IV to iminoketone VIII, which is
stabilized by cyclization (X). Decarboxylation and dehydration
lead to XI, resulting in the Strecker aldehyde upon hydrolysis
(Figure 1, pathway C).
To obtain more insight into the importance of glycine
degradation under different reaction conditions, the amount of
glycine generated per 1 mol of DFG decomposed was calculated
for five reaction stages (Table 2), i.e., at the beginning of the
reaction when only 5% of DFG was decomposed and at the
time when 25%, 50%, 75%, and 95% of DFG was decomposed.
The data set is incomplete for some reaction stages, as the
reaction at pH 5 or 6 did not proceed far enough under the
conditions chosen (90 °C, 7 h). The yield of glycine formed
from DFG decreased at the advanced reaction stage for all pH
values studied both in water and in phosphate buffer. This is

5476 J. Agric. Food Chem., Vol. 50, No. 19, 2002
Davidek et al.
Table 4. Amount of Glucose and Mannose Generated from
Table 3. Ratio of Glucose to Mannose Generated from
N-(1-Deoxy-D-fructos-1-yl)glycine (DFG) under Different Reaction
Conditions
N-(1-Deoxy-D-fructos-1-yl)glycine (DFG) at Different Reaction Stages
glucose and mannose
glucose:mannose ratio
glucose and mannose
(mol/mol of DFG)
generated in water
(mol/mol of DFG)
generated in
phosphate buffer
amount
of DFG
(%) de-
pH value
in water
in phosphate buffer
5
6
7
8
67:33
67:33
69:31
69:31
69:31
68:32
69:31
71:29
composed pH 5 pH 6 pH 7 pH 8 pH 5 pH 6 pH 7 pH 8
5
25
50
75
95
0.03 0.06 0.07 0.03 0.18 0.15 0.09 0.05
0.03 0.06 0.06 0.03 0.13 0.12 0.11 0.06
0.07 0.06 0.03 0.11 0.13 0.11 0.06
0.05 0.03
0.04 0.02
0.12 0.1
0.1 0.08 0.03
0.05
acetic acid
acetic acid
(mol/mol of DFG)
generated in water
(mol/mol of DFG)
generated in
phosphate buffer
amount
of DFG
(%) de-
composed pH 5 pH 6 pH 7 pH 8 pH 5 pH 6 pH 7 pH 8
5
25
50
75
95
0
0.13 0.36 0.60
0
0.24 0.40 0.55
0.05 0.24 0.39 0.60 0.06 0.27 0.43 0.54
0.42 0.42 0.60 0.10 0.32 0.46 0.54
0.48 0.58
0.51 0.64
0.37 0.53 0.59
0.46 0.59 0.63
formic acid
(mol/mol of DFG)
generated in
formic acid
(mol/mol of DFG)
generated in water
amount
of DFG
(%) de-
phosphate buffer
composed pH 5 pH 6 pH 7 pH 8 pH 5 pH 6 pH 7 pH 8
5
25
50
75
95
0
0.05 0.03 0.06
0
0
0
0.02 0.08
0.03 0.07 0.04 0.06
0.01 0.03 0.09
0.09 0.05 0.07 0.01 0.04 0.05 0.09
0.07 0.08
0.11
0.06 0.10 0.10
0.1 0.14 0.13
Figure 4. Formation of glucose and mannose during the thermal
degradation of N-(1-deoxy-D-fructos-1-yl)glycine in water (s) and in
phosphate buffer (‚‚‚) at pH 5 (b), pH 6 (9), pH 7 (2), and pH 8 ([).
The curves represent the sum of glucose and mannose concentrations.
The coefficient of variation was lower than 10% for concentrations above
1.5 mmol/L.
which has also been explained by reverse Amadori rearrange-
ment (14). The formation of parent sugars by aldol-type
condensations was excluded by labeling experiments. However,
reduction of glucosone has not been considered as a mechanism
leading to the parent sugars, and the design of labeling
experiments does not exclude this possibility.
increases the formation rate of glucose and mannose at all pH
values studied.
The results shown in this study represent for the first time
quantitative data concerning the formation of parent sugars from
an Amadori compound under conditions relevant to food
processing (aqueous system, cooking temperature). Similarly,
the effect of pH and phosphate on the yield of glucose and
mannose has not been reported previously. Although the high
yields of parent sugars found already at the early stage of DGF
degradation support the hypothesis of reverse Amadori rear-
rangement under our experimental conditions, additional experi-
ments are currently being performed to substantiate this
hypothesis. Irrespective of the formation mechanism, our data
clearly demonstrate that rather significant amounts of the parent
sugars can be regenerated at the early stages of the Maillard
reaction under conditions relevant to food processing.
Formation of Acids. The concentrations of acetic and formic
acid were monitored by HPAEC to verify their formation from
DFG. The rate of formation of both acids increased with
increasing pH over the whole pH range studied (Figure 5).
However, in the presence of phosphate there was only a small
difference between the formation rate at pH 7 and 8. Thus, the
effect of pH was similar on both the formation of these acids
and on the decomposition of DFG. Phosphate accelerated the
formation of acetic acid in the pH range studied, as it promoted
the degradation of DFG. However, this effect was most
pronounced at pH 5-7. Phosphate also accelerated the formation
To gain more insight into the importance of glucose and
mannose formation under different reaction conditions, the yields
of these sugars per mole of DFG were calculated as shown in
Table 4. In water, the highest yields of glucose and mannose
were observed at pH 6 and 7, reaching up to 7%. However,
much higher yields were obtained in phosphate buffer, indicating
that phosphate favors the formation of glucose and mannose
from DFG, particularly at pH 5. At this pH, the yields were
13-18%, about four to six times higher than in water. The data
also indicate that high yields were obtained already at very early
stages of the reaction.
Theoretically, the formation of glucose and mannose from
DFG can be explained by several pathways: (i) reverse Amadori
rearrangement (Figure 1, pathway A), (ii) aldol-type condensa-
tions of sugar fragments, and (iii) reduction of glucosone (IX)
by reducing substances such as reductones and aminoreductones.
The formation of parent carbonyls from Amadori compounds
under physiological conditions has been attributed to reverse
Amadori rearrangement (12, 13). However, the other mecha-
nisms cannot be excluded as no labeling experiments were
performed. Thermal decomposition of the Amadori compound
N-(deoxy-D-fructos-1-yl)proline in N,N-dimethylformamide (130
°C, 20 min) resulted in the parent sugars glucose and mannose,

Degradation of Fructosylglycine
J. Agric. Food Chem., Vol. 50, No. 19, 2002 5477
Table 5. Comparison of Amounts of Acetic and Formic Acid
Generated with Amounts of NaOH Added To Keep the pH Value
Constant
sum of acetic and
formic acid (mmol)
reaction in water
NaOH added (mmol)
reaction in water
reaction
time (h)
pH 5 pH 6 pH 7 pH 8 pH 5 pH 6 pH 7 pH 8
1
2
3
4
5
6
7
0.0
0.0
0.1
0.2
0.3
0.3
0.5
0.3
0.9
1.7
2.4
3.0
3.7
4.1
2.7
5.0
6.5
7.4
8.1
8.7
8.9
6.6
9.7
0.1
0.2
0.4
0.5
0.7
0.9
1.1
1.2
2.5
3.2
4.1
5.0
5.8
6.5
4.3
6.8
8.5
5.0
7.8
10.9
11.5
11.7
11.7
11.8
10.1
11.1
11.9
12.5
13.1
9.8
10.8
11.5
12.0
sum of acetic and
formic acid (mmol)
NaOH added (mmol)
reaction in
reaction in phosphate buffer
phosphate buffer
reaction
time (h)
pH 5
pH 6
pH 7
pH 8 pH 5 pH 6 pH 7a pH 8
1
2
3
4
5
6
7
0.0
0.2
0.3
0.5
0.9
1.3
1.7
1.7
4.0
5.7
6.9
7.8
8.3
8.7
6.1
9.2
7.8
10.5
11.5
11.9
11.8
12.4
11.9
0.6
1.2
2.0
2.6
3.5
4.2
4.2
4.1
6.5
8.4
5.6
8.6
9.9
10.6
10.6
10.6
10.6
10.6
10.9
11.2
11.5
11.7
9.7
11.2
11.6
12.1
a Data not available.
the yield of formic acid in water increased only with increasing
DFG degradation, but pH had almost no effect.
The yield of acetic acid was much higher than that of formic
acid. The sum of the yields of both acids reached up to 0.76
mol/mol of DFG, which indicates that fragmentation of DFG
was considerable. These high yields are in line with literature
data (15, 25). The degradation of protein-bound Amadori
compound (0.43 mmol/L, 120 °C, pH 6.7) yielded more than
0.22 mmol/L acetic acid and 0.15 mmol/L formic acid. The
authors explained the formation of formic acid by a C1-C2
cleavage of 3-deoxy-2-hexosulose. A C2-C3 cleavage of
1-deoxy-2,3-hexodiulose and a C1-C2 cleavage of methylgly-
oxal were proposed to explain the formation of acetic acid. Our
data on acetic acid fit well the proposed mechanism as the yield
of acetic acid increased with increasing pH. This is coherent
with the fact that sugar fragmentation as well as 2,3-enolization
is favored at higher pH values. Moreover, the formation of acetic
acid was very rapid at the beginning of the reaction. However,
it almost stopped when no more DFG was present in the reaction
mixture. This implies direct involvement of DFG or its early
degradation product in the formation of acetic acid.
Figure 5. Formation of (A) acetic acid and (B) formic acid during the
thermal degradation N-(1-deoxy-D-fructos-1-yl)glycine in water (s) and
in phosphate buffer (‚‚‚) at pH5 (b), pH 6 (9), pH 7 (2), and pH 8 ([).
The coefficient of variation was lower than 10% for concentrations above
3 mmol/L.
In contrast, the formation of formic acid by the proposed
mechanism is more difficult to explain. The yield of formic
acid either was not affected by pH (reaction in water) or was
even enhanced at higher pH values (reaction in phosphate
buffer). However, the formation of 3-deoxy-2-hexosulose, the
proposed precursor of formic acid (15), via 1,2-enolization
should be favored at lower pH values. This discrepancy could
be explained by the favored fragmentation at higher pH values.
Thus, although less 3-deoxy-2-hexosulose is formed at higher
pH values, it is more efficiently fragmented, resulting finally
in the same or even higher amounts of formic acid as compared
to lower pH values.
of formic acid, but only in the pH range 6-8. With the exception
of pH 8, the total amounts of acetic and formic acid were
generally lower than the amounts of NaOH added to keep pH
constant (Table 5). Hence other acids had to be formed. On
the other hand, the amounts of acetic and formic acid at pH 8
were higher than the amounts of NaOH added. This means that
part of the acids was neutralized by compounds created during
the reaction.
The yields of both acids were calculated at different reaction
stages to gain more insight into the formation of acetic and
formic acid under different reaction conditions. The yields of
acetic acid increased with increasing pH and with increasing
degree of DFG degradation in both water and phosphate buffer
(Table 4). The same trend was observed for the yield of formic
acid in the presence of phosphate buffer (Table 4). Interestingly,
However, the proposed mechanism does not explain why the
ratio of acetic acid to formic acid decreased as the reaction
progressed (Table 4). This indicates that, apart from the possible
formation by a C1-C2 cleavage of 3-deoxy-2-hexosulose, formic

5478 J. Agric. Food Chem., Vol. 50, No. 19, 2002
Davidek et al.
(6) Kawakishi, S.; Ushida, K. Autoxidation of Amadori compounds
in the presence of copper ion and its effect on the oxidative
damage of protein. In Maillard reaction in food processing,
human nutrition and physiology; Finot, P. A., Aeschbacher, H.
U., Hurrel, H. U., Liardon, R., Eds.; Birkha¨user: Basel, 1990;
pp 475-480.
(7) Hofmann, T.; Bors, W.; Stettmaier, K. Studies on radical
intermediates in the early stage of the nonenzymatic browning
reaction of carbohydrates and amino acids. J. Agric. Food Chem.
1999, 47, 379-390.
(8) Hubert, B.; Ledl, F. Formation of 1-amino-1,4-dideoxy-2,3-
hexodiuloses and 2-aminoacetylfurans in the Maillard reaction.
Carbohydr. Res. 1990, 204, 215-220.
(9) Hofmann, T.; Schieberle, P. Formation of aroma active Strecker-
aldehydes by a direct oxidative degradation of Amadori com-
pounds. J. Agric. Food Chem. 2000, 48, 4301-4305.
(10) Hayashi, T.; Namiki, M. Role of sugar fragmentation in the
Maillard reaction. In Amino carbonyl reactions in foods and
biological systems; Fujimaki, M., Namiki, M., Kato, H., Eds.;
Elsevier: Amsterdam, 1986; pp 29-38.
(11) Yaylayan, V. A.; Sporn, P. Novel mechanism for the decomposi-
tion of 1-(amino acid)-1-deoxy-D-fructoses (Amadori com-
pounds): A mass spectrometric approach. Food Chem. 1987,
26, 283-305.
(12) Acharya, A. S.; Sussman, L. G. The reversibility of ketosamine
linkages of aldose with proteins. J. Biol. Chem. 1984, 259, 4372-
4378.
Figure 6. Color development during the thermal degradation of N-(1-
deoxy-D-fructos-1-yl)glycine in water (s) and in phosphate buffer (‚‚‚) at
pH 5 (b), pH 6 (9), pH 7 (2), and pH 8 ([).
(13) Ahmed, M. U.; Thrope, S. R.; Bayens, J. W. Identification of
Nꢀ-carboxymethyllysine as degradation product of fructoselysine
in glycated protein. J. Biol. Chem. 1986, 261, 4889-4894.
(14) Yaylayan, V. A.; Huyghues-Despointes, A. Identification of per-
O-(trimethylsilyl) derivatives of aldoses generated from thermal
decomposition of N-(1-deoxy-D-fructopyranos-1-yl)proline: Re-
versibility of the Amadori rearrangement. Carbohydr. Res. 1996,
286, 179-187.
(15) Brands, C. M. J.; van Boekel, M. A. J. S. Reaction of
monosaccharides during heating of sugar-casein system: Build-
ing of reaction network model. J. Agric. Food Chem. 2001, 49,
4667-4675.
(16) Staempfli, A. A.; Blank, I.; Fumeaux, R.; Fay, L. B. Study on
the decomposition of the Amadori compound N-(1-deoxy-D-
fructos-1-yl)-glycine in model systems: Quantification by fast
atom bombardment tandem mass spectrometry. Biol. Mass
Spectrom. 1994, 23, 642-646.
(17) Potman, R. P.; van Wijk T. A. Mechanistic studies of the Maillard
reaction with emphasis on phosphate-mediated catalysis. In
Thermal Generation of Aromas; Parliment, T. H., McGorrin, R.
J., Ho, C.-T., Eds.; ACS Symposium Series 409; American
Chemical Society: Washington, DC, 1989; pp 182-195.
(18) Velisek, J.; Davidek, T.; Davidek, J.; Trska, P.; Kvasnicka, F.;
Velcova, K. New imidazoles in nonenzymatic browning reac-
tions. J. Food Sci. 1989, 54, 1544-1646.
(19) Hayashi, T.; Namiki, M. Formation of two-carbon sugar fragment
at an early stage of the browning of sugar with amine. Agric.
Biol. Chem. 1980, 44, 2575-2580.
(20) Namiki, M.; Hayashi, T. Formation of novel free radical products
in early stage of Maillard reaction. Prog. Food Nutr. Sci. 1981,
5, 81-91.
(21) Namiki, M.; Hayashi, T. A new mechanism of the Maillard
reaction involving sugar fragmentation and free radical formation.
In The Maillard reaction in foods and nutrition; Waller, G. R.,
Feather, M. S., Eds.; ACS Symposium Series 215; American
Chemical Society: Washington, DC, 1983; pp 21-46.
(22) Cremer, D. R.; Vollenbroeker, M.; Eichner, K. Investigation of
the formation of Strecker aldehydes from the reaction of Amadori
rearrangement products with R-amino acids in low moisture
model system. Eur. Food Res. Technol. 2000, 211, 400-403.
(23) Cremer, D. R.; Eichner, E. The influence of pH value on the
formation of Strecker aldehydes in low moisture model systems
and plant powders. Eur. Food Res. Technol. 2000, 211, 247-
251.
acid is formed by additional pathways. These pathways, active
in advanced reaction stages, could include further fragmentation
of intermediates arising from the sugar molecule but also
Cannizzaro reaction of formaldehyde, the Strecker aldehyde of
glycine, or formation of formic acid by Strecker-type reaction
(24). To clarify the formation mechanism of formic acid, further
experiments using labeled compounds are necessary.
Development of Browning. The rate of browning of unbuf-
fered DFG solutions increased with pH over the whole pH range
studied (Figure 6). In the presence of phosphate, the browning
rate also increased with pH, but only between pH 5 and pH 7.
These data are well in line with the data obtained by Westphal
et al. (26). These authors reported acceleration of the browning
of fructosylalanine (100 °C) in phosphate buffer at pH 7 as
compared to pH 5, but no further acceleration was observed at
pH 8. The presence of phosphate accelerated the browning of
DFG over the whole pH range studied. However, at pH 8 the
effect of phosphate was lower than at pH 5-7.
ACKNOWLEDGMENT
We thank Dr. Elizabeth Prior for linguistic proofreading of the
manuscript and Dr. Fabien Robert for helpful discussions.
LITERATURE CITED
(1) Ledl, F.; Schleicher, E. New aspects of the Maillard reaction in
foods and in the human body. Angew. Chem. 1990, 29, 565-
594.
(2) Hodge, J. E. The Amadori rearrangement. AdV. Carbohydr. Res.
1955, 10, 169-205.
(3) Yaylayan, V. A.; Huyghues-Despointes A. Chemistry of Amadori
rearrangement products: analysis, synthesis, kinetics, reactions
and spectroscopic properties. Crit. ReV. Food Sci. Nutr. 1994,
34, 321-369.
(4) Wrodnigg, T. M.; Eder, B. The Amadori and Heyns rearrange-
ments: Landmarks in the history of carbohydrate chemistry or
unrecognized synthetic opportunities? Top. Curr. Chem. 2001,
215, 115-152.
(5) Hodge, J. E. Dehydrated foods: Chemistry of browning reactions
in model systems. J. Agric. Food Chem. 1953, 1, 928-943.

Degradation of Fructosylglycine
J. Agric. Food Chem., Vol. 50, No. 19, 2002 5479
(24) Hofmann, T.; Mu¨nch, 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.
(26) Westphal, G.; Kroh, L.; Fo¨llmer, U. Studies of the Maillard
reaction. Part 16. Reactivity of Amadori compounds dependent
on the reaction medium. Food (Nahrung) 1988, 32, 117-120
(in German; Chem. Abstr. 1988, 109, 127413m).
(25) van Boekel, M. A. J. S.; Brands, C. M. J. Heating of sugar casein
solutions: Isomerisation and Maillard reaction. In The Maillard
reaction in food and medicine; O’Brien, J., Nursten, H. E.,
Crabbe, M. J. C., Ames, J. M., Eds.; The Royal Society of
Chemistry: Cambridge, U.K., 1998; pp 154-159.
Received for review March 27, 2002. Revised manuscript received June
24, 2002. Accepted June 25, 2002.
JF025561J