Formation of odorants in Maillard model systems based on L-proline as affected by pH

Article abstract of DOI:10.1021/jf034077t

Formation of the odorants acetic acid, 4-hydroxy-2,5-dimethyl-3-(2H)-furanone (HDMF), 6-acetyl-1,2,3,4-tetrahydropyridine (ATHP), and 2-acetyl-1-pyrroline (AP) was monitored by isotope dilution assays at pH 6, 7, and 8 in Maillard model reactions containing glucose and proline (Glc/Pro) or the corresponding Amadori compound fructosyl-proline (Fru-Pro). In general, higher yields were obtained at pH 7 and 8. Acetic acid was the major odorant with up to 40 mg/mmol precursor followed by HDMF (up to 0.25 mg/mmol), the formation of which was favored in the Fru-Pro reaction systems. On the contrary, ATHP (up to 50 μg/mmol) and AP (up to 5 μg/mmol) were more abundant in Glc/Pro. However, the sensory relevance of the two N-heterocycles was more pronounced on the basis of odor activity values, confirming their contribution to the overall roasty note of the reaction samples. It was also found that formation and decomposition of Fru-Pro were faster at pH 7 as compared to pH 6, explaining in part the preferred formation of the four odorants studied under neutral and slightly alkaline conditions. After 4 h of reaction at pH 7 in the presence of proline, about one-fourth of the glucose was consumed leading to acetic acid with a transformation yield of almost 40 mol %.

Full text of DOI:10.1021/jf034077t

J. Agric. Food Chem. 2003, 51, 3643−3650 3643
Formation of Odorants in Maillard Model Systems Based on
L-Proline as Affected by pH
IMRE BLANK,* STEÄ PHANIE DEVAUD, WALTER MATTHEY-DORET, AND
FABIEN ROBERT
Nestec Ltd., Nestle´ Research Center, P.O. Box 44, 1000 Lausanne 26, Switzerland
Formation of the odorants acetic acid, 4-hydroxy-2,5-dimethyl-3-(2H)-furanone (HDMF), 6-acetyl-
1,2,3,4-tetrahydropyridine (ATHP), and 2-acetyl-1-pyrroline (AP) was monitored by isotope dilution
assays at pH 6, 7, and 8 in Maillard model reactions containing glucose and proline (Glc/Pro) or the
corresponding Amadori compound fructosyl-proline (Fru-Pro). In general, higher yields were obtained
at pH 7 and 8. Acetic acid was the major odorant with up to 40 mg/mmol precursor followed by
HDMF (up to 0.25 mg/mmol), the formation of which was favored in the Fru-Pro reaction systems.
On the contrary, ATHP (up to 50 µg/mmol) and AP (up to 5 µg/mmol) were more abundant in Glc/Pro.
However, the sensory relevance of the two N-heterocycles was more pronounced on the basis of
odor activity values, confirming their contribution to the overall roasty note of the reaction samples.
It was also found that formation and decomposition of Fru-Pro were faster at pH 7 as compared to
pH 6, explaining in part the preferred formation of the four odorants studied under neutral and slightly
alkaline conditions. After 4 h of reaction at pH 7 in the presence of proline, about one-fourth of the
glucose was consumed leading to acetic acid with a transformation yield of almost 40 mol %.
KEYWORDS: Maillard reaction; L-proline; Amadori compound; N-(1-deoxy-D-fructos-1-yl)-L-proline; isotope
dilution assay; GC-MS; odorants; anion exchange chromatography
INTRODUCTION
Maillard reaction systems based on L-proline (Pro) are well-
known to generate roasty notes (1-4) that contribute to the
aroma of many thermally treated foods such as cereal products
(5, 6). 2-Acetyl-1-pyrroline (AP) and 6-acetyl-1,2,3,4-tetrahy-
dropyridine (ATHP), also referred to as 2-acetyltetrahydropy-
ridine in the literature, are two roasty smelling impact odorants
(Figure 1) identified in proline-containing foods such as Basmati
rice (7), bread crust (8), toast (9), popcorn (10), and sweet corn-
based products (11). These N-containing volatile compounds,
along with 2,3-butanedione (diacetyl) and 4-hydroxy-2,5-
dimethyl-3-(2H)-furanone (HDMF), have been reported as
important odorants generated in dry-heated model systems
containing D-glucose (Glc) and proline (12). In addition,
Hofmann and Schieberle (13) found further N-containing
compounds such as 2-propionyl-1-pyrroline and 2-propionyltet-
rahydropyridine in Glc/Pro systems under boiling conditions.
The formation mechanisms of these N-heterocyclic odorants
have been studied in great detail; their common precursor is
1-pyrroline, the Strecker degradation product of proline and
ornithine, which reacts with 2-oxopropanal (2-oxobutanal) and
1-hydroxy-2-propanone (1-hydroxy-2-butanone) to yield AP,
ATHP, and their higher homologues (5, 13-16). An excellent
Figure 1. Chemical structures of AP, HDMF, and ATHP.
review on R-acetyl-N-heterocycles has recently been published
by Kerler and co-workers (17).
On the other hand, Amadori compounds have been claimed
as nonvolatile flavor precursors in processed foods (18). These
N-substituted 1-amino-1-deoxy-ketoses represent an important
class of Maillard intermediates (19, 20). They are formed in
the initial phase of the Maillard reaction by Amadori rearrange-
ment of the corresponding N-glycosylamines, the latter obtained
by condensation of amino acids and aldoses, such as proline
and glucose, for example (21). The Amadori compound N-(1-
deoxy-D-fructos-1-yl)-L-proline (Fru-Pro) is a well-known con-
stituent of dried apricots and peaches (22), cured tobacco leaves
(23), white wine (24), malts, and beers (25). Fru-Pro has been
the subject of several investigations dealing with its thermally
* To whom correspondence should be addressed. Tel: +21/785-8607.
Fax: +21/785-8554. E-mail: imre.blank@rdls.nestle.com.
10.1021/jf034077t CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/18/2003

3644 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Blank et al.
Table 1. Waveforms Used for the Analysis of Glc, Pro, and the
Amadori Compound Fru-Pro by HPAEC Coupled with Electrochemical
Detection
induced degradation. Many volatiles have been identified after
pyrolysis (26-29) or autoclave treatment (30) of Fru-Pro.
However, there is not much known about their sensory relevance
and the key odorants generated from Fru-Pro. Only recently,
the roasty smelling odorants AP and ATHP were studied in Glc/
Pro and Fru-Pro reaction samples (31).
potential (V)
potential (V)
time (s)
Fru-Pro
Glc, Pro
time (s)
Fru-Pro
Glc, Pro
0.00
0.20
0.30
0.40
0.41
0.50
0.60
+ 0.05
−0.10
−0.10a
+ 0.35
+ 0.35
0.61
0.70
0.71
0.90
0.91
1.00
−0.15
−0.15
In our previous work (32), we reported on the identification
of odor active volatile compounds in Maillard systems based
on Glc/Pro and Fru-Pro. Odorants of high intensity were
screened by gas chromatography-olfactometry (GC-O) and
identified by GC-mass spectrometry (GC-MS). The roasty
smelling odorants AP and ATHP were perceived as more intense
in Glc/Pro, whereas HDMF (caramel-like) and sotolone (season-
ing-like) were more pronounced in Fru-Pro. No differences were
observed for acetic acid (AcOH) and diacetyl. In this work, we
focus on the quantification of four key odorants generated in
Glc/Pro and Fru-Pro Maillard systems as affected by the pH of
the reaction sample. In addition, an attempt is made to link the
formation of key odorants to nonvolatile Maillard intermediates
such as the Amadori compound Fru-Pro.
+ 0.05a
−0.10b
+ 0.90
+ 0.90
−0.90
−0.90
+ 0.05b
+ 0.75
−0.10
+ 0.75
a Starting with integration. b Integration stopped.
chromatographic conditions were as follows: temperature program, 35
°C (2 min), 40 °C/min to 50 °C, 8 °C/min to 180 °C, 10 °C/min to 240
°C (10 min).
GC-MS. This was carried out on a HP-6890A GC coupled to a
HP-5973N mass selective detector (Hewlett-Packard). Samples were
injected splitless (1 µL). The electron impact (EI) MS spectra were
generated at 70 eV. Qualitative measurements were performed on a
DB-Wax (60 m × 0.25 mm, 0.25 µm film thickness) or HP-PONA
(60 m × 0.25 mm, 0.25 µm film thickness) using the temperature
program 20 °C (1 min), 70 °C/min to 60 °C, 4 °C/min to 240 °C.
Quantification by IDA (14) was performed in the EI-MS mode using
the chromatographic conditions described above. AcOH and HDMF
were quantified on the DB-Wax capillary column by measuring the
molecular ions of analyte and labeled internal standard at m/z 60 (61)
and m/z 128 (130), respectively. The amounts of AP and ATHP were
determined on the HP-PONA by monitoring the molecular ions at m/z
111 (113-115) and m/z 125 (127-130), respectively. Quantification
of the N-heterocycles was based on the first eluting tautomer. Each
sample was injected twice.
HPAEC. The analyses were performed using a Dionex ion chro-
matography system (DX500, Dionex, Sunnyvale, CA) composed of
an autosampler (model AS-50 with a 25 µL sample loop), a gradient
pump (model GP-50) with on-line degas, and an electrochemical
detector (model ED-40). The separation was accomplished on a 250
mm × 4 mm i.d. CarboPac PA10 anion exchange column (Dionex)
and a 50 mm × 4 mm i.d. CarboPac PA10 guard column (Dionex).
The solutions and eluents were prepared using ultrapure deionized water
(specific resistivity g 18.2 MΩ cm) from a Milli-Q-system (Millipore,
Bedford, MA). NaOH solutions used as eluents were prepared by
diluting a carbonate free 50-52% (w/w) NaOH solution in water
previously degassed with helium gas. The following gradients were
used, composed of water, NaOH (500 mM), and NaOAc (1 M) in %:
(i) for Fru-Pro: 0.0 (87.5/10/2.5), 12.0 (52/40/8), 14.0 (30/50/20), 23.0
(30/50/20), 25.0 min (87.5/10/2.5); (ii) for glucose and proline: 0.0
(82/10/8), 7.0 (82/10/8), 9.0 (30/50/20), 18.0 (30/50/20), 20.0 min
(82/10/8). The flow rate was 1 mL/min separating Glc (RT ) 2.5 min),
Pro (RT ) 3.4 min), and Fru-Pro (RT ) 9.0 min), which were quantified
with an electrochemical detector equipped with a gold working
electrode. Each sample was injected twice. Quantification was based
on calibration curves by comparing the peak areas with those of standard
solutions containing known amounts of pure compounds. The wave-
forms used for the detection of Fru-Pro, glucose, and proline are shown
in Table 1.
MATERIALS AND METHODS
Chemicals. Glc, disodium hydrogenphosphate, sodium dihydrogen-
phosphate, sodium sulfate, sodium acetate (NaOAc), diethyl ether
(Et2O), methanol (MeOH), and n-hexane were from Merck (Darmstadt,
Germany). Pro, HDMF, 2-acetylpyridine, rhodium (5% on active
alumina, Rh/Al2O3), d1-methanol (MeO2H), heavy water (2H2O),
deuterochloroform (C2HCl3), acetone, Celite 560, Florisil (100-200
mesh), silver carbonate (50% on Celite, Ag2CO3), sodium hydroxide
(50-52%, NaOH), and AcOH were from Aldrich/Fluka (Buchs,
Switzerland). [13C1]Acetic acid (c-AcOH) was from S. I. C. (Inneberg,
Switzerland). Argon, hydrogen, and deuterium (grade 28) were obtained
from Carbagaz (Lausanne, Switzerland). The chemicals were of
analytical grade. The solvents were distilled prior to use on a Vigreux
column (100 cm × 1 cm). AP and 4-hydroxy-2(or 5)[13C]methyl-5(or
2)-methyl-3-(2H)-[2(or 5)-13C]furanone (c-HDMF) were prepared as
previously described (33, 34).
Model Reactions. Quantification of Odorants. Equimolar amounts
(2 mmol) of glucose and proline or Fru-Pro (2 mmol) were dissolved
in a phosphate buffer (20 mL, 0.2 M), and the pH was adjusted to 6,
7, or 8. Then, the solutions were refluxed for 1, 2, or 4 h. Quantification
was achieved using isotope dilution assay (IDA) (14). Defined amounts
of labeled internal standards were added to the cooled reaction
mixture: 0.4 mL of c-AcOH (1100 µg/mL Et2O), 0.2 mL of c-HDMF
(182.5 µg/mL Et2O), 0.05 mL of d-AP (47.7 µg/mL PG), and 0.05 mL
of d-ATHP (289.0 µg/mL PG). The samples were separated in acidic
(pH 3) and basic (pH 10) fractions using HCl (2 N) and NaOH (2 N),
and the odorants were extracted with Et2O (2×, 5 mL). Finally, the
combined organic extracts were dried over anhydrous Na2SO4 and
concentrated to about 1 mL with a stream of nitrogen prior to analysis
by GC-MS. Each sample was prepared at least in duplicate.
Quantification of NonVolatiles. Samples for monitoring the formation
of Fru-Pro were prepared by refluxing glucose (0.9 g, 5 mmol) and
proline (0.575 g, 5 mmol) in a phosphate buffer (50 mL, 0.2 mol/L)
for 7 h. Aliquots (0.6 mL) were taken at defined time intervals and
passed through a filter (0.45 µm). Samples were diluted with water, to
be in the linear range of the calibration curves, and then analyzed by
high-performance anion exchange chromatography (HPAEC). Similarly,
the decomposition of Fru-Pro was studied by refluxing Fru-Pro (1.39
g, 5 mmol) in a buffered solution as described above and analyzing
Fru-Pro by HPAEC. All samples were prepared at least in duplicate.
NMR Spectroscopy. The samples for NMR spectroscopy were
prepared in WILMAD 528-PP 5 mm Pyrex NMR tubes, using
deuterated water or chloroform (0.7 mL) as solvent. The NMR spectra
were acquired on a Bruker AM-360 spectrometer, equipped with a
1
quadrinuclear 5 mm probe head, at 360.13 MHz for H and at 75.56
MHz for 13C under standard conditions (35). Chemical shifts were
measured relative to the solvent signal.
GC. This was performed on a Carlo Erba Mega 2 gas chromatograph
equipped with a cold on-column injector and a flame ionization detector
(FID). A DB-1701 or DB-Wax fused silica capillary column was used,
30 m × 0.32 mm, with a film thickness of 0.25 µm (J&W Scientific,
Folsom, CA). A splitter was attached to the end of the capillary column
to split the effluent 1:1 into the FID and the sniffing port. The
Syntheses. Fru-Pro. The Amadori compound Fru-Pro was synthe-
sized as described in the literature (20, 26, 29) using some modifications.
Glucose (39.3 g, 0.22 mol) and proline (28.2 g, 0.25 mol) were placed
in a round flask. Anhydrous MeOH (300 mL) was added under a stream
of argon and then refluxed for 7 h in an oil bath (72 °C). The reaction

Odorants from Proline Maillard Systems
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3645
Figure 2. Chemical structure of the Amadori compound Fru-Pro.
mixture was cooled in a refrigerator, and the solvent was evaporated
under reduced pressure to obtain about 200 mL of solution. The target
compound was isolated by precipitation with dry acetone (350 mL) in
a desiccator under argon. This procedure was repeated twice to purify
Fru-Pro, finally obtaining a white powder after drying under reduced
pressure (14 g, 20% yield) with a purity of 95% (HPAEC) containing
residual glucose (4%) and proline (1%). The product is hygroscopic
2
and must be stored under argon. 13C NMR (90 MHz, H2O, δ/ppm):
23.8 (4′-CH2), 28.7 (3′-CH2), 57.8 (5′-CH2), 61.1 (1-CH2), 64.2
(6-CH2), 69.3-70.4 (3-CH, 4-CH, 5-CH), 71.9 (2′-CH), 96.1 (2-CH2),
174.3 (1′-COOH). The assignment of the signals was verified by a
HETCOR experiment, in particular 2′-CH, 5′-CH2, and 1-CH2 (data
not shown). In general, these data are in good agreement with those
reported in the literature (29, 36) and indicate that Fru-Pro occurred
mainly in the â-pyranoid form (Figure 2), which usually represents
the major conformation of Amadori compounds (37).
[2H2-5]ATHP. d-ATHP was prepared as a tautomeric mixture of
isotopomers according to the Bu¨chi and Wu¨st method (38) using some
modifications. 2-Acetylpyridine (3 g, 24.8 mmol) was deuterated in
MeO2H (70 mL) in the presence of Rh/Al2O3 (3.54 g) for 12 h under
pressure (5 bar) and gentle stirring using an autoclave with a glass
vessel (100 mL). The reaction product was filtered through Celite (30
g) and concentrated at room temperature in vacuo (20 mbar, Rotavap,
Bu¨chi) yielding [2H5-11]-2-(1-hydroxyethyl)piperidine (3.7 g) as a
crude orange oil with a purity of 95%. (GC): RI(DB-1701) ) 1231,
RI(HP-PONA) ) 1078. MS-EI (m/z, %): ion clusters 134-140 (1-2,
M+), 115-123 (2-3, [M - H2O]+), 86-94 (10-100), 90 (100),
56-62 (5-30), 40-50 (2-10). 1H NMR (360 MHz, C2HCl3, δ/ppm):
0.88 (s, 3H, CH3), 0.90-1.00 (m, 1H, CH), 1.16-1.31 (m, 1H, CH),
1.40-1.52 (m, 1H, CH), 2.80-2.97 (m, 1H, N-CH). 13C NMR (90
MHz, C2HCl3, δ/ppm): 18.8 (CH3), 24.0, 25.6, 28.6, 46.8 (6-C), 61.9
(2-C), 70.0 (CH-O).
The intermediate (3.4 g) was dissolved in n-hexane (200 mL) and
oxidized under a stream of argon with Ag2CO3 (13.5 g) by refluxing
the solution at 72 °C overnight. The reaction product was filtered
through Florisil (90 g) and eluted with hexane/Et2O (1:1, by volume,
200 mL). The target compound was purified by distilling off the solvent
under high vacuum (10-5 mbar) using the SAFE apparatus (39), leading
to a solution of d-ATHP (215.5 µg/mL) with an overall yield of 1.4%
and a purity of 85% (GC) represented by two tautomers, each of them
composed of several isotopomers: [2H2-5]-6-acetyl-2,3,4,5-tetra-
hydropyridine (60%, RI(HP-PONA) ) 1019) and [2H2-5]ATHP (40%,
RI(HP-PONA) ) 1115). The mass spectra of both tautomers (Figure
3) show characteristic ion clusters of deuterated compounds. On the
DB-Wax column, d-ATHP was obtained as a broad peak (RI ≈ 1610),
which is not recommended for quantification purposes due to poor
chromatographic properties. The concentration of the d-ATHP solution
was determined using N-acetylpiperidine as internal standard.
ATHP. ATHP was synthesized as a tautomeric mixture as described
for d-ATHP by hydrogenation of 2-acetylpyridine in MeOH yielding
2-(1-hydroxyethyl)piperidine with 95% purity: RI(DB-1701) ) 1231,
RI(HP-PONA) ) 1078, followed by oxidation of the intermediate
leading to the target compound with a conversion yield of 98% (GC)
and a purity of 97% (GC). The mass spectra of the two tautomers were
in good agreement with literature data (13).
Figure 3. Mass spectra of the deuterated acetyltetrahydropyridine
tautomers: (A) [2H2-5]-6-acetyl-2,3,4,5-tetrahydropyridine (60%); (B)
[2H2-5]ATHP (40%).
ion clusters 118-123 (1, M+), 101-106 (2-3, [M - H2O]+), 71-76
(10-100), 75 (100), 40-48 (2-10). 1H NMR (360 MHz, C2HCl3)
δ/ppm: 1.10 (s, 3H, CH3), 1.56 (m, 1H, CH), 1.72 (m, 2H).
Oxidation of the intermediate was carried out in n-hexane (75 mL)
with Ag2CO3 (13.5 g) by refluxing the solution at 72 °C overnight.
Purification over Florisil and elution with hexane/Et2O (1:1, by volume,
150 mL) gave a solution of d-AP (47.7 µg/mL) with an overall yield
of 0.4% and a purity of 95% (GC) represented by two tautomeric
forms: [2H3-5]AP (93%, RI(HP-PONA) ) 891) and [2H3-5]-2-acetyl-
2-pyrroline (7%, RI(HP-PONA) ) 921). The mass spectra of both
tautomers (Figure 4) show characteristic ion clusters of deuterated
compounds. Only one peak was observed on the DB-Wax column at
RI ) 1355. The concentration of the d-AP solution was determined
using N-methyl-2-acetylpyrrol as internal standard.
RESULTS
The effect of pH on the formation of odorants from Glc/Pro
and Fru-Pro was studied in buffered aqueous solutions under
boiling conditions using the same initial concentration of the
precursors (0.1 mol/L). The experiments were carried out at pH
6, 7, and 8. As GC-O data only provide a rough estimation of
the concentration, four odorants with high sensory intensities,
AcOH, HDMF, AP, and ATHP (32), were quantified by IDA
using the labeled analogues as internal standards (14) to
substantiate differences between the Maillard systems Glc/Pro
and Fru-Pro as affected by the pH. An HPEAC method coupled
with electrochemical detection was developed to quantify the
Amadori compound Fru-Pro.
[2H2-4]AP. d-AP was synthesized as reported by Buttery et al. (7),
which is based on the Bu¨chi and Wu¨st method (38), following the
procedure described above for d-ATHP using some modifications:
Deuteration of 2-acetylpyrrole (2 g, 18.3 mmol) was performed with
Rh/Al2O3 (2.35 g) in MeO2H (50 mL) resulting in [2H2-9]-2-(1-
hydroxyethyl)pyrrolidine (1.7 g) as an orange oil with a purity of 80%
(GC): RI(DB-Wax) ) 1591, RI(HP-PONA) ) 993. MS-EI (m/z, %):
Quantification of Selected Odorants. AcOH was the major
odorant formed in both Maillard systems at all pH values studied
(Figure 5A). The concentrations after 2 h of reaction varied
from about 400 to 38 000 µg/mmol, which corresponds to up

3646 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Blank et al.
from Glc/Pro as compared to the Amadori compound Fru-Pro,
in particular at pH 6 and 7, whereas similar amounts were found
in both Maillard systems at pH 8. These data also suggest that
ATHP may be formed by other reaction pathways than AcOH
and HDMF. Another difference to AcOH and HDMF was the
observation that the formation of ATHP was favored at pH 7.
It seems that at pH 8 more ATHP is decomposed than formed.
As shown in Figure 5D, AP was generated in very low
amounts, i.e., about 0.5-4 µg/mmol, with overall yields of up
to 0.004 mol % after a reaction period of 2 h, which is about
10 times less as compared to ATHP. Similar to ATHP, the
formation of AP was favored in Glc/Pro as compared to Fru-
Pro at all pH values studied. However, under slightly alkaline
conditions (pH 8), the amounts of AP generated were similar
in both Maillard systems. This was also observed for ATHP,
but there, the concentrations were lower as compared to pH 7.
The Amadori system gave the highest AP amounts at pH 8.
Quantification of Nonvolatile Flavor Precursors. The
formation of the Amadori compound Fru-Pro in Maillard
systems containing glucose and proline was monitored by
HPAEC (Figure 6A). The concentrations were very low, up to
1.3 mg of Fru-Pro per mmol glucose, representing about 0.5
mol % yield. A continuous formation was observed at pH 6
with a rapid increase within 2 h that then flattened off to reach
a plateau after about 4 h. Therefore, Fru-Pro can be accumulated
at pH 6. On the contrary, at pH 7, the increase in Fru-Pro was
faster, reaching a maximum after already 1 h. However,
thereafter, the Amadori compound that formed was more rapidly
decomposed: only half of the maximum amount could be
detected after 7 h of reaction.
The amount of glucose was monitored during the Maillard
reaction as it is consumed by the formation of Fru-Pro, which
in turn can further react and give rise to many sugar degradation
products. As shown in Table 3, more glucose was consumed
at pH 7 and more rapidly than at pH 6. After a reaction period
of 7 h, about 40% of glucose was lost at pH 7 as compared to
10% at pH 6.
Degradation of Fru-Pro was more pronounced at pH 7
(Figure 6B). Less than 10% of the Amadori compound was
left at pH 7 after a reaction period of 1 h, whereas at pH 6
about 40% was still available. Therefore, it is not surprising
that rapid formation of Fru-Pro at pH 7 cannot compensate for
the losses.
Figure 4. Mass spectra of the deuterated AP tautomers: (A) [2H3-5]AP
(93%); (B) [2H3-5]-2-acetyl-2-pyrroline (7%).
to 60 mol % yield. However, in all cases, more AcOH was found
in the samples containing the Amadori compound (Fru-Pro) as
compared to those containing Glc/Pro. As a consequence, the
pH in these samples dropped after 4 h of reaction by 1 unit
from the original pH value (Table 2). In contrast, the pH
remained unchanged in the Glc/Pro reaction systems over the
whole reaction period, most likely due to the lower amounts of
AcOH formed. The increase in AcOH concentration from Glc/
Pro was particularly drastic at pH 8 (Figure 5A) where almost
20 mg/mmol was formed already after 1 h of reaction. Such
high amounts were produced from Fru-Pro under milder
conditions, namely, at lower pH, thus suggesting that the
Amadori compound is readily decomposed leading to fragmen-
tation products such as AcOH.
After 2 h of reaction, 0 to about 250 µg/mmol of the caramel-
like smelling odorant HDMF was produced (Figure 5B), which
is 100 times lower than AcOH, representing yields of up to 0.2
mol %. In general, the formation of HDMF was favored from
the Amadori compound as compared to Glc/Pro. However, at
pH 8, both Maillard systems resulted in comparable amounts
of HDMF with similar kinetic curves. The Amadori system
showed a drastic increase at pH 7 whereas Glc/Pro required
pH 8 to significantly enhance the HDMF concentration. This
behavior was similar to that observed for AcOH, indicating that
their formation may follow similar pathways. However, HDMF
can be generated by several mechanisms: with or without sugar
fragmentation.
DISCUSSION
Reaction of glucose and proline leads via pathway A to the
corresponding N-glycosyl derivative I, which then enolizes to
various intermediates such as the enaminol derivative II, the
Amadori compound III (Fru-Pro), and the endiol derivative IV
as shown in Figure 7 (19, 40). Pathway B is favored under
acidic conditions and results by 1,2-enolization in 3-deoxy-2-
hexosulose (V), with formic acid and 5-hydroxy-2-furaldehyde
as characteristic degradation products. Pathway C preferably
proceeds under alkaline conditions leading by 2,3-enolization
to 1-deoxy-2,3-hexodiulose (VI) with AcOH and 2,3-dihydro-
3,5-dihydroxy-6-methyl-4(H)-pyran-4-one as typical reaction
products.
The highest concentrations of AcOH were obtained at pH 8
as compared to pH 7 and 6, the latter being in general the least
efficient reaction system in generating odor active compounds.
The formation of AcOH was favored from Fru-Pro, in particular
at pH 6 and 7, which means under milder conditions, suggesting
that the Amadori compound is readily decomposed leading to
AcOH by fragmentation. This reaction is one of the main
The formation of the roasty smelling odorant ATHP is shown
in Figure 5C. Concentrations after 2 h of reaction varied from
2 to 45 µg/mmol, which was approximately 5 times less as
compared to HDMF, representing yields of up to 0.04 mol %.
In contrast to AcOH and HDMF, higher amounts were obtained

Odorants from Proline Maillard Systems
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3647
Figure 5. Generation of AcOH (A), HDMF (B), ATHP (C), and AP (D) from Glc/Pro (‚ ‚ ‚) and the Amadori compound Fru-Pro (s) under boiling
conditions at pH 6 (4 2), pH 7 (0 9), and pH 8 (O b). The coefficient of variation was lower than 20% for concentrations (µg/mmol) above 300 (A),
5.0 (B), 1.0 (C), and 0.3 (D).
Table 2. Evolution of the pH in the Maillard Reaction Samples Based
on Glc/Pro and the Amadori Compound Fru-Proa
pH 6
pH 7
pH 8
reaction
time (h)
Glc/Pro
Fru-Pro
Glc/Pro
Fru-Pro
Glc/Pro
Fru-Pro
0
1
2
4
6.00
6.15
6.16
6.13
6.00
5.80
5.42
5.05
7.00
7.04
7.00
7.00
7.00
6.70
6.30
6.20
8.00
8.04
8.02
8.02
8.00
7.30
7.02
6.97
a The samples were buffered with phosphate (0.2 mol/L). The pH values were
measured in an aliquot of the reaction sample after rapidly cooling to room
temperature.
degradation pathways of 1-deoxy-2,3-diuloses resulting in up
to 60 mol % of AcOH, which is mainly responsible for the pH
drop observed particularly in the Amadori system (Table 2).
These high yields are in line with recent literature data.
Degradation of the casein-bound Amadori compound (120 °C,
40 min, pH 6.8) yielded 50 mol % of AcOH (41) and N-(1-
Deoxy-D-fructos-1-yl)glycine gave rise to about 60 mol % of
AcOH (90°C, 5 h, pH 7.0) (40).
The relatively high amounts of AcOH in the Glc/Pro system
at pH 8 can be explained by the preferred formation and
decomposition of the Amadori compound under alkaline condi-
tions (Figure 7). In the Maillard system Glc/Pro at pH 7, AcOH
was continuously generated reaching 3.3 and 5.8 mg/mmol sugar
after 2 and 4 h (Figure 5A), respectively, which corresponds
to about 5 and 10 mol % yield (Table 4). In the same period of
time, about 10 and 25 mol % of glucose was consumed after 2
and 4 h of reaction (Table 3), respectively, thus indicating that
almost half of the sugar consumed was transformed into AcOH
at pH 7. Fru-Pro is rapidly decomposed under the same reaction
conditions (Figure 6B), of which 50-60 mol % was trans-
formed to AcOH (Table 4). The absolute yields at pH 6 in the
Amadori system increased from 17 (1 h) to 25 mol % (4 h)
with transformation yields of about 30 mol % (Table 4). These
data suggest that at pH 6 less AcOH is formed in total and that
the transformation rate from sugar to AcOH is half as much as
compared to pH 7.
Figure 6. (A) Formation of the Amadori compound Fru-Pro from Glc and
Pro and (B) degradation of Fru-Pro under boiling conditions at pH 6 (2)
and pH 7 (9).
at pH 6 and 7. The Amadori compound N-(1-deoxy-D-fructos-
1-yl)glycine has also been demonstrated to generate more
HDMF at pH 6 and 7 as compared to its precursors, glucose
and glycine (42). The reaction goes through 1-deoxy-2,3-
hexodiuloses (Figure 7, pathway C) with acetylformoine as the
key intermediate (not shown). This pathway requires a reduction
step, which may take place in the presence of reductones, readily
HDMF showed a behavior similar to that of AcOH. The
formation of HDMF was also favored from Fru-Pro, in particular

3648 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Table 3. Consumption of Glc in the Presence of Proa
Blank et al.
Table 4. Absolute and Relative Yields (mol %) of AcOH Generated
from Glc/Pro and the Amadori Compound Fru-Pro in the Maillard
Reactiona
concentration
(mmol/50 mL)
loss of glucose
(mol %)
reaction
Glc/Pro
(pH 6)
Glc/Pro
(pH 7)
Fru-Pro
(pH 6)
Fru-Pro
(pH 7)
time (min)
pH 6
5.00
pH 7
pH 6
pH 7
0
30
60
5.00
0
1
1
8
5
7
7
13
11
11
0
11
5
10
11
20
26
31
33
39
time (t)
abs
rel
abs
rel
abs
rel
abs
rel
4.95 ± 0.10
4.93 ± 0.06
4.59 ± 0.10
4.74 ± 0.09
4.66 ± 0.05
4.65 ± 0.05
4.33 ± 0.11
4.44 ± 0.08
4.44 ± 0.13
4.47 ± 0.10
4.75 ± 0.33
4.49 ± 0.10
4.44 ± 0.24
4.02 ± 0.15
3.68 ± 0.40
3.45 ± 0.05
3.37 ± 0.06
3.05 ± 0.07
1
2
4
nd
nd
1.0
2.5
5.4
9.7
50
50
37
16.6
17.8
24.5
28
25
30
51.6
54.6
58.7
54
55
59
90
120
180
240
300
360
420
13
a The absolute (abs) yield was calculated on the basis of AcOH determined in
the sample as related to the amount of Glc or Fru-Pro used in the reaction. The
relative (rel) yield was calculated on the basis of AcOH determined in the sample
as related to the amount of Glc or Fru-Pro consumed in the reaction. Not detected,
nd.
a Glucose and proline were refluxed in a phosphate buffer (0.2 mol/L) for up to
7 h. Aliquots were analyzed by HPAEC as described in the Experimental Section.
Figure 8. Formation of HDMF by recombination of the C sugar
3
fragmentation products VII and VIII.
HDMF found in the Glc/Pro system at pH 8. However, this
may also be due to the preferred formation and decomposi-
tion of the Amadori compound under alkaline conditions
(Figure 7).
The concentrations of the roasty smelling odorants ATHP
and AP were significantly lower as compared to AcOH and
HDMF. They were preferably generated from Glc/Pro, less from
Fru-Pro, thus suggesting that there is no benefit in using the
Amadori compound for the formation of ATHP and AP, which
is in agreement with the recent observation of Weenen and van
der Ven (31). All samples contained more ATHP than AP,
confirming data reported by Schieberle (14) who obtained 38
µg of ATHP per mmol proline (pH 7, 2 h, boiling), which is
close to the 45 µg/mmol found in this work.
In contrast to AcOH and HDMF, N-heterocyclic odorants
were more abundant in the Glc/Pro samples, indicating that they
are formed by other reaction pathways. As shown in Figure 9,
the common precursor of ATHP and AP is 1-pyrroline (IX),
the Strecker degradation product of proline, which reacts with
the sugar degradation products VII (pathway D) and VIII,
shown in the hydrated form (pathway E), to generate ATHP
and AP, respectively (16). It seems that Glc/Pro represents a
better system than the Amadori compound for generating these
intermediates in high concentrations at a given reaction time.
However, the reaction yields reported in the literature are
moderate, about 1 mol % ATHP and 5 mol % AP when reacting
equimolar amounts of IX with VII and VIII (pH 7, 30 min,
boiling), respectively (16). In view of the high instability of IX
and the probably low amounts generated in the Maillard reaction,
IX seems to be the limiting factor in forming ATHP and AP.
However, reliable quantitative data of IX under Maillard reaction
conditions are missing in the literature.
Figure 7. Simplified schematic presentation of parts of the Maillard reaction
focusing on the formation of early Maillard intermediates (A) and their
degradation via the 1,2- (B) and 2,3-enolization (C) pathways. The
structures are shown in the open chain form, which are in equilibrium
with cyclic species.
available in Maillard reactions (42, 43). On the other side,
HDMF can also be formed by recombination of sugar frag-
mentation products such as hydroxy-2-propanone (VII) and
2-oxopropanal (VIII) (Figure 8), which are readily available
in Maillard reaction samples (44). Aldol reaction of VII and
VIII leads to an intermediate with two methyl groups that can
easily generate HDMF by enolization, dehydration, and cy-
clization. Indeed, HDMF could easily be identified by GC-MS
after reacting VII and VIII at 90 °C in a phosphate-buffered
solution at pH 7 for 1 h (data not shown). Thus, the favored
sugar fragmentation under alkaline conditions leading to various
reactive C2 and C3 units may explain the high amounts of
Despite the low yields, both N-heterocyclic odorants play an
important role for the overall aroma due to their low odor
thresholds. The sensory relevance of odorants can be estimated
by comparing their odor activity values (OAV), i.e., the ratio

Odorants from Proline Maillard Systems
J. Agric. Food Chem., Vol. 51, No. 12, 2003 3649
various intermediates to be available at a given time. Conse-
quently, aroma generation of such odorants may be improved
on the basis of a more precise insight into dynamic changes of
key intermediates as affected by pH, temperature, and other
reaction parameters.
ACKNOWLEDGMENT
We are grateful to Dr. Tomas Davidek for critical discussions
and Dr. Elizabeth Prior for linguistic proofreading of the
manuscript.
LITERATURE CITED
(1) Kiely, P. J.; Nowlin, A. C.; Moriarty, J. H. Bread aromatics from
browning systems. Cereal Sci. Today 1960, 5, 273-274.
(2) Shigematsu, H.; Shibata, S.; Kurata, T.; Kato, H.; Fujimaki, M.
5-Acetyl-2,3-dihydro-1H-pyrrolizines and 5,6,7,8-tetrahydroin-
dilizin-8-ones, odor constituents formed on heating L-proline with
D-glucose. J. Agric. Food Chem. 1975, 23, 233-236.
(3) Tressl, R.; Helak, B.; Martin, N. Formation of flavor compounds
from L-proline. In Topics in FlaVour Research; Berger, R. G.,
Nitz, S., Schreier, P., Eds.; Eichhorn: Marzling-Hagenham,
Germany, 1985; pp 139-159.
(4) Schieberle, P.; Hofmann, T.; Mu¨nch, P. Studies on potent aroma
compounds generated in Maillard-type reactions using the odor-
activity-value concept. In FlaVor Chemistry; Risch, S. J., Ho,
C.-T., Eds.; ACS Symposium Series 756; American Chemical
Society: Washington, DC, 2000; pp 133-150.
(5) Schieberle, P. The role of free amino acids in yeast as precursors
of the odorants 2-acetyl-1-pyrroline and 2-acetyltetrahydropy-
ridine in wheat bread crust. Z. Lebensm. Unters. Forsch. 1990,
191, 206-209.
Figure 9. Simplified scheme showing the formation of AP and ATHP by
reacting IX with VII and VIII, the latter shown in the hydrated form (adapted
from (4)).
(6) Grosch, W.; Schieberle, P. Flavor of cereal products. Cereal
Chem. 1997, 74, 91-97.
Table 5. OAV of AcOH, HDMF, ATHP, and AP in Maillard Reaction
Samples Based on Glc/Pro and the Amadori Compound Fru-Proa
(7) Buttery, R. G.; Ling, L. C.; Juliano, B. O.; Turnbaugh, J. G.
Cooked rice aroma and 2-acetyl-1-pyrroline. J. Agric. Food
Chem. 1983, 31, 823-826.
(8) Schieberle, P.; Grosch, W. Identification of the volatile flavour
compounds of wheat bread crust - Comparison with rye bread
crust. Z. Lebensm. Unters. Forsch. 1985, 180, 474-478.
(9) Rychlik, M.; Grosch, W. Identification and quantification of
potent odorants formed by toasting of wheat bread. Lebensm.-
Wiss. Technol. 1996, 29, 515-525.
AcOH
HDMF
ATHP
AP
Maillard
systems concn OAV concn OAV concn OAV concn OAV
pH 6 Glc/Pro
Fru-Pro
41 <1
648 13
<0.05
4
<1
67
1.3
0.2
810 0.25 2500
125 0.03 300
pH 7 Glc/Pro
325
6.5
1.6
20
26
27
4.5 2810 0.36 3600
Fru-Pro 3274 65
pH 8 Glc/Pro 2310 46
Fru-Pro 3810 76
330
430
380
0.5
1.1
1.3
310 0.04
690 0.33 3300
810 0.22 2200
400
23
(10) Schieberle, P. Primary odorants of popcorn. J. Agric. Food Chem.
1991, 39, 1141-1144.
a The concentrations were taken from samples reacted for 2 h. The OAVs were
calculated by dividing the concentration (concn, in mg/L) of each odorant by its
nasal odor threshold (mg/L) determined in water: AcOH, 50 (45); HDMF, 0.06
(46); ATHP, 0.0016 (47); and AP, 0.0001 (7). Note that the odor thresholds in
water were determined without adjusting the pH to 6, 7, and 8.
(11) Buttery, R. G.; Stern, D. J.; Ling, L. C. Studies on the flavor
volatiles of some sweet corn products. J. Agric. Food Chem.
1994, 42, 791-795.
(12) Roberts, D. D.; Acree, T. E. Gas chromatography-olfactometry
of glucose-proline Maillard reaction products. In Thermally
Generated FlaVors; Parliament, T. H., Morello, M. J., McGorrin,
R., Eds.; ACS Symposium Series 543; American Chemical
Society: Washington, DC, 1993; pp 71-79.
(13) Hofmann, T.; Schieberle, P. Flavor contribution and formation
of the intense roast-smelling odorants 2-propionyl-1-pyrroline
and 2-propionyltetrahydropyridine in Maillard-type reactions. J.
Agric. Food Chem. 1998, 46, 2721-2726.
(14) Schieberle, P. Quantification of important roast-smelling odorants
in popcorn by stable isotope dilution assays and model studies
on flavor formation during popping. J. Agric. Food Chem. 1995,
43, 2442-2448.
of concentration to odor threshold determined in water. As
shown in Table 5, both N-heterocyclic odorants showed high
sensory relevance based on high OAVs and dominated espe-
cially in the Glc/Pro samples. These results also show that AP,
occurring in very low concentrations, does significantly con-
tribute to the roasty note. On the contrary, AcOH plays a minor
role as aroma compound. Its major contribution is guiding
Maillard reaction pathways by lowering the pH. Finally, the
importance of HDMF is evidenced by its high OAV, particularly
in the Fru-Pro samples.
In conclusion, AcOH was the most abundant odorant in both
Glc/Pro and Fru-Pro Maillard systems formed by major reaction
pathways in the Maillard reaction. On the contrary, HDMF and
in particular the two N-heterocyclic odorants were minor
reaction products formed by side reactions, which require
(15) Schieberle, P.; Hofmann, T. Mechanistic studies on the formation
of the cracker-like aroma compounds 2-acetyltetrahydropyridine
and 2-acetyl-1-pyrroline by Maillard-type reactions. In The
Maillard Reaction in Foods and Medicine; O’Brian, J., Nursten,
H. E., Crabbe, M. J. C., Ames, J. M., Eds.; The Royal Society
of Chemistry: Cambridge, U.K., 1998; pp 209-215.

3650 J. Agric. Food Chem., Vol. 51, No. 12, 2003
Blank et al.
(16) Hofmann, T.; Schieberle, P. 2-Oxopropanal, hydroxy-2-pro-
panone, and 1-pyrroline - Important intermediates in the
generation of the roast-smelling food flavor compounds 2-acetyl-
1-pyrroline and 2-acetyltetrahydropyridine. J. Agric. Food Chem.
1998, 46, 2270-2277.
(17) Kerler, J.; van der Ven, J. G. M.; Weenen, H. R-Acetyl-N-
heterocycles in the Maillard reaction. Food ReV. Int. 1997, 13,
553-575.
(18) Mills, F. D.; Baker, B. G.; Hodge, J. E. Amadori compounds as
nonvolatile flavor precursors in processed foods. J. Agric. Food
Chem. 1969, 17, 723-727.
(19) Ledl, F.; Schleicher, E. New aspects of the Maillard reaction in
foods and in the human body. Angew. Chem. 1990, 29, 565-
594.
(20) Yaylayan, V. A.; Huyghues-Despointes, A. Chemistry of Ama-
dori rearrangement products: Analysis, synthesis, kinetics,
reactions, and spectroscopic properties. CRC ReV. Food Sci. Nutr.
1994, 34, 321-369.
(21) Hodge, J. E. The Amadori rearrangement. AdV. Carbohydr.
Chem. 1955, 10, 169-205.
(22) Anet, E. F. L. J.; Reynolds, T. M. Chemistry of nonenzymic
browning. I. Reactions between amino acids, organic acids, and
sugars in freeze-dried apricots and peaches. Aust. J. Chem. 1957,
10, 182-192.
(33) Duby, P.; Huynh-Ba, T. Preparation and encapsulation of the
flavour enhancer 2-acetyl-1-pyrroline. European Patent Applica-
tion, EP 545085 A1, 1992.
(34) Blank, I.; Fay, L. B.; Lakner, F. J.; Schlosser, M. Determination
of 4-hydroxy-2,5-dimethyl-3-(2H)-furanone and 2(or 5)-ethyl-
4-hydroxy-5(or 2)-methyl-3-(2H)-furanone in pentose sugar-
based Maillard model systems by isotope dilution assays. J.
Agric. Food Chem. 1997, 45, 2642-2648.
(35) Lin, J.; Welti, H. D.; Arce Vera, F.; Fay, L. B.; Blank, I.
Synthesis of deuterated volatile lipid degradation products to be
used as internal standards in isotope dilution assays. 1. Alde-
hydes. J. Agric. Food Chem. 1999, 47, 2813-2821.
(36) Vernin, G.; Debrauwer, L.; Vernin, G. M. F.; Zamkotsian, R.-
M.; Metzger, J.; Larice J. L.; Pa´rka´nyi, C. Heterocycles by
thermal degradation of Amadori intermediates. In Off-FlaVors
in Foods and BeVerages; Charalambous, G., Ed.; Elsevier
Publishers: Amsterdam, The Netherlands, 1992; pp 567-623.
(37) Ro¨per, H.; Ro¨per, S.; Heyns, K.; Meyer, B. N. M. R. spectros-
copy of N-(1-deoxy-D-fructos-1-yl)-L-amino acids (“Fructose-
amino acids”). Carbohydr. Res. 1983, 116, 183-195.
(38) Bu¨chi, G.; Wu¨st, H. Synthesis of 2-acetyl-1,4,5,6-tetrahydropy-
ridine, a constituent of bread aroma. J. Org. Chem. 1971, 36,
609-610.
(39) Engel, W.; Bahr, W.; Schieberle, P. Solvent assisted flavour
evaporation. A new and versatile technique for the careful and
direct isolation of aroma compounds from complex food
matrixes. Z. Lebensm. Unters. Forsch. 1999, 209, 237-241.
(40) Davidek, T.; Clety, N.; Aubin, S.; Blank, I. Degradation of the
Amadori compound N-(1-deoxy-D-fructos-1-yl)glycine in aque-
ous model systems. J. Agric. Food Chem. 2002, 50, 5472-5479.
(41) Brands, C. M. J.; van Boekel, M. A. J. S. Reactions of
monosaccharides during heating of sugar-casein systems: Build-
ing of a reaction network model. J. Agric. Food Chem. 2001,
49, 4667-4675.
(23) Tomita, H.; Noguchi, M.; Tamaki, E. Chemical studies on
ninhydrin-positive compounds in cured tobacco leaves. Part II.
Isolation of 1-deoxy-1-L-prolino-D-fructose. Agric. Biol. Chem.
1965, 29, 515-519.
(24) Hashiba, H. Isolation and identification of Amadori compounds
from miso, white wine and sake. Agric. Biol. Chem. 1978, 42,
1727-1731.
(25) Wittmann, R.; Eichner, K. Detection of Maillard products in
malts, beers, and brewing colorants. Z. Lebensm. Unters. Forsch.
1989, 188, 212-220 (in German; Chem. Abstr. 1989, 111,
55721b).
(42) Blank, I.; Devaud, S.; Fay, L. B. Study on the formation and
decomposition of the Amadori compound N-(1-deoxy-D-fructos-
1-yl)glycine in Maillard model systems. In The Maillard Reaction
in Foods and Medicine; O’Brian, J., Nursten, H. E., Crabbe, M.
J. C., Ames, J. M., Eds; The Royal Society of Chemistry:
Cambridge, U.K., 1998; pp 43-50.
(43) Schieberle, P. Studies on the formation of furaneol in heat-
processed foods. In FlaVor PrecursorssThermal and Enzymatic
ConVersions; Teranishi, R., Takeoka, G. R., Gu¨ntert, M., Eds.;
ACS Symposium Series 490; American Chemical Society:
Washington, DC, 1992; pp 164-175.
(44) Hofmann, T. Quantitative studies on the role of browning
precursos in the Maillard reaction of pentoses and hexoses with
L-alanine. Eur. Food Res. Technol. 1999, 209, 113-121.
(45) Larsen, M.; Poll, L. Odour thresholds of some important aroma
compounds in strawberries. Z. Lebensm. Unters. Forsch. 1992,
195, 120-123.
(46) Buttery, R. G.; Takeoka, G. R.; Krammer, G. E.; Ling, L. C.
Identification of 2,5-dimethyl-4-hydroxy-3(2H)-furanone (Fura-
neol) and 5-methyl-4-hydroxy-3(2H)-furanone in fresh and
processed tomato. Lebensm.-Wiss. Technol. 1994, 27, 592-594.
(47) Belitz, H.-D.; Grosch, W. Food Chemistry, 2nd English ed.;
Springer: Berlin, Germany, 1999; p 340.
(26) Mills, F. D.; Hodge, J. E. Amadori compounds: Vacuum
thermolysis of 1-deoxy-1-L-prolino-D-fructose. Carbohydr. Res.
1976, 51, 9-21.
(27) Shigematsu, H.; Shibata, S.; Kurata, T.; Kato, H.; Fujimaki, M.
Thermal degradation products of several Amadori compounds.
Agric. Biol. Chem. 1977, 41, 2377-2385.
(28) Yaylayan, V. A.; Sporns, P. Novel mechanisms for the decom-
position of 1-(amino acid)-1-deoxy-D-fructoses (Amadori com-
pounds): A mass spectrometric approach. Food Chem. 1987,
26, 283-305.
(29) Debrauwer, L.; Vernin, G.; Metzger, J.; Siouffi, A. M.; Larice,
J. L. Kinetics of formation and thermal degradation of the
fructose-proline Amadori intermediate (glucose + proline). Bull.
Soc. Chim. Fr. 1991, 128, 244-254 (in French; Chem. Abstr.
1991, 115, 114893s).
(30) Huyghues-Despointes, A.; Yaylayan, V. A.; Keyhani, A. Py-
rolysis GC/MS analysis of 1-[(2′-carboxyl)pyrrolidinyl]-1-deoxy-
D-fructose (Amadori-Proline). J. Agric. Food Chem. 1994, 42,
2519-2524.
(31) Weenen, H.; van der Ven, J. G. M. The formation of Strecker
aldehydes. In Aroma ActiVe Compounds in Foods; Takeoka, G.
R., Gu¨ntert, M., Engel, K.-H., Eds.; ACS Symposium Series 794;
American Chemical Society: Washington, DC, 2001; pp 183-
195.
(32) Blank, I.; Devaud, S.; Matthey-Doret, W.; Pollien, P.; Robert,
F.; Yeretzian, C. Formation of odour-active compounds in
Maillard model systems based on proline. In 10th Weurman
FlaVour Research Symposium; Le Que´re´, J.-L., Etie´vant, P. X.,
Eds; Lavoisier: Paris, France, 2003; in press.
Received for review January 27, 2003. Revised manuscript received
March 10, 2003. Accepted March 23, 2003.
JF034077T