Monitoring carbonyl-amine reaction and enolization of 1-hydroxy-2- propanone (acetol) by FTIR spectroscopy

Article abstract of DOI:10.1021/jf9812836

Infrared absorption bands characteristic of the aldehydo, keto, and enediol forms of 1-hydroxy-2-propanone (acetol) were identified and used to study the effect of solvent on the absorption frequencies and the effect of temperature and acid/base catalysis on the enolization reactions. The data indicated that, in addition to water, acids and bases can catalyze the enolization of 1-hydroxy-2-propanone and that the temperature inversely effects the rate of enolization under basic conditions. However, under acidic conditions, increasing the temperature favors the enolization process. In addition, the reaction of 1-hydroxy-2-propanone with a primary and a secondary amine was also monitored by Fourier transform infrared spectroscopy. The data indicated that at room temperature the rate of amine reaction was faster than the rate of its catalysis of enolization; however, below room temperature, the rate of base-catalyzed enolization became comparable with the rate of carbonyl-amine reaction forming both Heyns and Amadori adducts.

Full text of DOI:10.1021/jf9812836

J. Agric. Food Chem. 1999, 47, 2335−2340
2335
Mon itor in g Ca r bon yl-Am in e Rea ction a n d En oliza tion of
1-Hyd r oxy-2-p r op a n on e (Acetol) by F TIR Sp ectr oscop y
Varoujan A. Yaylayan,* Susan Harty-Majors, and Ashraf A. Ismail
Department of Food Science and Agricultural Chemistry, McGill University, 21,111 Lakeshore,
Ste. Anne de Bellevue, Quebec, Canada H9X 3V9
Infrared absorption bands characteristic of the aldehydo, keto, and enediol forms of 1-hydroxy-2-
propanone (acetol) were identified and used to study the effect of solvent on the absorption frequencies
and the effect of temperature and acid/base catalysis on the enolization reactions. The data indicated
that, in addition to water, acids and bases can catalyze the enolization of 1-hydroxy-2-propanone
and that the temperature inversely effects the rate of enolization under basic conditions. However,
under acidic conditions, increasing the temperature favors the enolization process. In addition, the
reaction of 1-hydroxy-2-propanone with a primary and a secondary amine was also monitored by
Fourier transform infrared spectroscopy. The data indicated that at room temperature the rate of
amine reaction was faster than the rate of its catalysis of enolization; however, below room
temperature, the rate of base-catalyzed enolization became comparable with the rate of carbonyl-
amine reaction forming both Heyns and Amadori adducts.
Keyw or d s: FTIR; R-hydroxycarbonyl moiety; enediol; enolization; 1-hydroxy-2-propanone (acetol);
carbonyl-amine reactions; Heyns and Amadori product formation
INTRODUCTION
infrared spectroscopy (FTIR) has been used to study the
effect of temperature (Yaylayan and Ismail, 1992) on
acyclic forms of D-fructose, mutarotation of D-glucose
and D-fructose (Back et al., 1984), and enolization and
carbonyl group migration in selected sugars (Yaylayan
and Ismail, 1995). Kobayashi et al. (1976) studied
dimeric structures of D,L-glyceraldehyde and dihydroxy-
acetone by infrared and Raman spectroscopy. FTIR
spectroscopy is ideally suited to study the different
forms of R-hydroxycarbonyl moiety. Recently, the sim-
plest R-hydroxyaldehyde (glycolaldehyde) was investi-
gated by FTIR (Yaylayan et al., 1998). The data
indicated that the glycolaldehyde cyclic dimer (2,5-
dihydroxy-1,4-dioxane) undergoes a ring opening to form
the acyclic dimer, which can recyclize into the 2-hy-
droxymethyl-4-hydroxy-1,3-dioxolane structure. The acy-
clic dimer can dissociate into monomeric glycoladehyde
in equilibrium with the enediol form. In this paper, the
enolization and its effect on the carbonyl-amine reac-
tions of the simplest R-hydroxyketone (acetol) is inves-
tigated.
Many chemical transformations of reducing sugars,
such as cyclizations, enolizations, and isomerizations,
are initiated by the presence of the R-hydroxycarbonyl
moiety. The stabilization of the initial imine adduct
formed during Maillard reaction, through rearrange-
ment reactions (Amadori and Heyns), is also due to the
presence of the R-hydroxyl groups (Yaylayan and Huy-
ghues-Despointes, 1994). The physical and chemical
properties of reducing sugars in solution depend on the
relative concentrations of different forms originating
from the R-hydroxycarbonyl moiety. Their biological
properties can also have similar dependence. It appears
that the enediol forms of certain sugar derivatives in
biological systems are the active forms with which
enzymes react. The conversion of cytotoxic methylgly-
oxal into lactate, for example, by the enzyme glyoxylase
I (thioester hydrolase glyoxylase II, EC 3.1.2.6), involves
an enediolate intermediate of the methylglyoxal deriva-
tive (Hamilton and Creighton, 1992). Rubisco (ribulose-
1,5-bisphosphate carboxylase, EC 4.1.1.39) adds carbon
dioxide to the 2,3-enediolate form of ribulose-1,5-bis-
phosphate (Lorimer et al., 1993). Enediol forms are
known to play an important role in metal-catalyzed
oxidative degradation of reducing sugars specially in
biological systems (Thornalley et al., 1984). The com-
plexity in the population of reducing sugars in solution
makes their study a difficult task, especially in hexoses
and pentoses, where the presence of more stable fura-
nose and pyranose forms renders the R-hydroxycarbonyl
moiety undetectable at room temperature due to their
lower concentrations (<1%). Solvent interference, es-
pecially water, makes the detection of aldehydo forms
even more difficult due to hydration. Fourier transform
MATERIALS AND METHODS
All reagents and chemicals were purchased from Aldrich
Chemical Co. (Milwaukee, WI) and used without further
purification. All solvents used were of HPLC grade; D2O was
purchased from MSD Isotopes (Montreal, Quebec, Canada).
F TIR An a lysis. Infrared spectra were recorded on a Nicolet
8210 Fourier transform infrared spectrometer (Madison, WI)
purged with dry air and equipped with a deuterated triglycine
sulfate (DTGS) detector. The spectra were acquired on a CaF2
IR cell with a 25-µm Teflon spacer at room temperature unless
otherwise specified. A total of 128 scans at 4 cm-1 resolution
were co-added. Processing of the FTIR data was performed
using GRAMS/386 version 3.01. Second-order derivatization
was performed using the Savitsky-Golay function (9 points)
to enhance closely absorbing peaks.
* Corresponding author. Tel: 514-398-7918. Fax 514-398-
7977. E-mail: yaylayan@agradm.lan.mcgill.ca.
Tem p er a tu r e Stu d ies. Sample solutions were placed in a
CaF2 IR cell with a 25-µm Teflon spacer. The temperature of
10.1021/jf9812836 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/14/1999

2336 J. Agric. Food Chem., Vol. 47, No. 6, 1999
Yaylayan et al.
Sch em e 1. Molecu la r Tr a n sfor m a tion s of
1-Hyd r oxy-2-p r op a n on e (Acetol)
Sch em e 2. Rea ction of P yr r olid in e w ith
1-Hyd r oxy-2-p r op a n on e (Acetol) a n d F or m a tion of
Am a d or i (5) a n d Heyn s Rea r r a n gem en t (3) P r od u cts,
in th e Absen ce of a Solven t
the sample was regulated by placing the IR cell in a temper-
ature-controlled cell holder. Infrared spectra were recorded as
described above. The initial temperature of the cell was raised
by 5 °C/min, and every 10 min the temperature was kept
constant for 2 min to record the spectra. A total of 128 scans
at 4 cm-1 resolution were co-added.
GC/MS An a lysis. A Hewlett-Packard GC/mass selective
detector (5890 series II GC/5971B MSD, Palo Alto, CA) was
used for the GC/MS analysis. The samples were introduced
through the cool on-column injection port at constant pressure
of 34 psi. The pressure was regulated by an electronic pressure
controller (Hewlett-Packard, Palo Alto, CA). Capillary direct
MS interface temperature was 180 °C; ion source temperature
was 280 °C. The ionization voltage was 70 eV, and the electron
multiplier was 2047 V. The mass range analyzed was 20-200
amu. The column used was Q-PLOT (30 m × 0.2 mm × 0.32
µm; Hewlett-Packard). The column initial temperature was
set at -10 °C for 2 min and then increased to 50 °C at a
heating rate of 30 °C/min; immediately the temperature was
further increased to 250 °C at a rate of 8 °C/min and kept at
250 °C for 5 min.
are as important as aldehydo or keto forms of reducing
sugars in aqueous solutions. In addition, knowledge of
characteristic absorption frequencies can also allow
monitoring of the carbonyl-amine interaction of 1-hy-
droxy-2-propanone. By observing changes in the inten-
sities of the carbonyl bands of 1-hydroxy-2-propanone
and the emergence of new bands, conclusions can be
drawn regarding the formation of the so-called Heyns
and Amadori rearrangement products (3 and 5 in
Scheme 2). 1-Hydroxy-2-propanone, similar to reducing
sugars, can undergo Maillard reaction with amines to
produce 2-aminopropionaldehyde (Heyns product) in
addition to 3-amino-2-propanone (Amadori product).
Effect of Solven t on th e Ca r bon yl Absor p tion
Ba n d of 1-Hyd r oxy-2-p r op a n on e. In general, ketones
absorb at a lower frequency than aldehydes. However,
the introduction of electron-withdrawing groups such
as halogens or a hydroxyl group at the R-carbons is
known to lead to a shift of the carbonyl stretching
frequency to higher values, provided that the halogen
can rotate to eclipse the carbonyl group. The magnitude
of this shift depends on the torsional angle. The effect
of the R-hydroxyl groups on the absorption frequency
of sugar carbonyls was dependent on their ability to
attain eclipsed conformation with respect to the carbo-
nyl group. Sugars having more rotational freedom to
attain an eclipse or near-eclipse conformation exhibited
a higher shift. 1-Hydroxy-2-propanone in the gas phase
has been reported to absorb at 1750 cm-1 (Coleman and
Gordon, 1989); 10 cm-1 higher than acetone in the gas
phase. Neat 1-hydroxy-2-propanone exhibits a strong
carbonyl absorption band at 1726 cm-1 shifted by 9 cm-1
as compared to neat acetone that absorbs at 1717 cm-1
(see Table 1). In D2O, this band was shifted to 1720 cm-1
due to solvent effect; however, the extent of this shift
due to water indicates that 1-hydroxy-2-propanone is
mainly intramolecularly hydrogen bonded (1), as has
been confirmed by other investigators (Yasuda et al.,
1995). On the other hand, the carbonyl band of acetone
in D2O is shifted to 1698 cm-1, a 19 cm-1 shift as
compared to only 6 cm-1 in the case of 1-hydroxy-2-
propanone. In dioxane, the carbonyl group of 1-hydroxy-
RESULTS AND DISCUSSION
1-Hydroxy-2-propanone (1 in Scheme 1), similar to
other R-hydroxycarbonyl compounds, can undergo 1,2-
and 2,3-enolizations to produce the enol 1a and the
enediol 1b. The enediol 1b can exist in equilibrium with
2-hydroxypropionaldehyde (1c) due to the reversal of
the enediol-carbonyl equilibrium through both hydroxyl
groups. However, unlike other short-chain R-hydroxy-
carbonyl compounds, 1-hydroxy-2-propanone is not
known to exist in dimeric forms (Davis, 1973). Under-
standing the factors that influence this equilibrium can
further our ability to rationalize the different products
obtained during the interaction of 1-hydroxy-2-pro-
panone with nucleophiles such as amines or amino
acids. Identification of characteristic absorption fre-
quencies of different forms of 1-hydroxy-2-propanone
can aid in studying the effect of solvent, pH, and
temperature on the extent of enolization. The absorption
frequencies of the carbonyl and enediol moieties of
hexoses and ketoses have been confirmed by studies
with selected 13C-substituted sugars (Yaylayan and
Ismail, 1995). The presence of R-hydroxyl groups in
reducing sugars was found to shift the stretching
frequencies of sugar carbonyl bands to higher values
relative to R-deoxy-derivatives. Using D-[2-13C]ribose
and D-[1-13C]ribose, further evidence for the occurrence
of enolization was provided by observing the migration
of carbonyl group from C-1 to C-2 atom in D-ribose and
subsequent formation of D-ribulose. These studies have
indicated that the relative concentrations of the enediols

FTIR Analysis of Acetol Reaction with Amines
J. Agric. Food Chem., Vol. 47, No. 6, 1999 2337
Ta ble 1. Ca r bon yl Absor p tion F r equ en cies of Releva n t
2-P r op a n on e (Aceton e) Der iva tives
carbonyl absorption
acetone derivativea
frequency (cm-1
)
acetone (g)
acetone (D2O)
acetone (l)
1-chloroacetone (g)
1-chloroacetone (l)
1,3-dichloroacetone (l)
1,3-diaminoacetone (l)
1-hydroxyacetone (D2O)
1-hydroxyacetone (g)
1-hydroxyacetone (l)
1740
1698
1717
1743
1736
1743
1743
1720
1750
1726
a
g, gas phase; l, liquid.
F igu r e 2. Second-derivative spectra (1600-1800 cm-1) of a
2% solution of 1-hydroxy-2-propanone in 5% DCl in D2O at 80
°C (dotted line) and in 5% NaOD in D2O at 30 °C (solid line).
in D2O due to the presence of low concentrations of
enolized species. Enolization in D2O can be verified
however by observing deuterium exchange at the R-hy-
drogens by GC/MS analysis. A solution of 1-hydroxy-2-
propanone in D2O was equilibrated overnight and
analyzed by GC/MS. The data indicated the occurrence
of deuterium exchange of all the six hydrogens of
1-hydroxy-2-propanone in the following percentages: M,
2%; M + 1, 10%; M + 2, 22%; M + 3, 28%; M + 4, 22%;
M + 5, 12%; M + 6, 4%. When 1-hydroxy-2-propanone
was equilibrated in 5% NaOD solution at 30 °C, the
carbonyl peak was shifted to 1712 cm-1, and two new
bands appeared in the alkene region, a major band at
1694 cm-1 and a weaker band at 1667 cm-1 (Figure 2).
Under basic conditions, the negatively charged alkoxide
group should destabilize the developing partial negative
charge on C-1 and make it less acidic relative to C-3.
Accordingly, the stronger band at 1694 cm-1 was
assigned to the enol 1a , and the weaker band at 1667
cm-1was assigned to the enediol 1b. Under acidic
conditions, the reverse trend should be observed where
C-1 protons will be more acidic relative to C-3 protons
due to the formation of alkyl oxonium ions. 1-Hydroxy-
2-propanone was equilibrated in a 5% DCl solution at
30 °C, no absorption was observed in the alkene region,
only the carbonyl band was observed at 1719 cm-1. At
temperatures higher than 65 °C, a new band appeared
at 1667 cm-1 and increased with temperature with
corresponding decrease in the intensity of carbonyl
band. This observation is consistent with the assign-
ment of the enol and enediol bands under basic condi-
tions. The process was not reversible during cooling
cycle due to the formation of stable intramolecular
H-bonding in 1b. To prevent hydration and conse-
quently to observe the carbonyl band due to the forma-
tion of 2-hydroxypropionaldehyde (1c), 1-hydroxy-2-
propanone was analyzed in 100% triethylamine solution.
The appearance of two bands in the carbonyl region (see
Figure 3) indicated the occurrence of enolization. The
addition of a small amount of D2O to this solution
caused the aldehydo band (1735 cm-1) to disappear and
F igu r e 1. Effect of temperature on the intensity of carbonyl
absorption bands (1680-1740 cm-1) of 1-hydroxy-2-propanone
(2% D2O).
2-propanone absorbs at 1726 cm-1. In 5% NaOD, due
to the formation of a bulky sodium alkoxide at the
R-carbon, prevented 1-hydroxy-2-propanone to attain
eclipsed conformation, and consequently, the absorption
frequency was lowered to 1712 cm-1
.
A solution of 1-hydroxy-2-propanone in D2O, also
exhibited a weak absorption band in the carbonyl region
centered at 1702 cm-1 (see Figure 1) very close to the
absorption frequency of acetone in D2O (1698 cm-1).
Tentatively, this peak was assigned to the 1-hydroxy-
2-propanone rotomer in which the hydroxyl group is
near-staggered conformation (1′). By increasing the
temperature of the 1-hydroxy-2-propanone solution, the
intensity of non-H-bonded rotomer (1′) should increase
on the expense of H-bonded rotomer (1). The initial
temperature (30 °C) of the cell was raised by 1 °C/min,
and every 5 min the temperature was kept constant for
15 min to record the spectra. During the heating cycle,
the intensity of the band at 1702 cm-1 increased during
the heating cycle and decreased during the cooling cycle,
indicating reversibility of the process.
Effect of Acid /Ba se Ca ta lysis a n d Tem p er a tu r e
on th e En oliza tion of 1-Hyd r oxy-2-p r op a n on e.
Enolization of 1-hydroxy-2-propanone (1, acetol) can be
initiated by both R-carbons to produce the enediol 1b
and the enol 1a . This process can be catalyzed by acids
and bases. Bases in general are more efficient catalysts
than acids. Under basic conditions, enolization initiated
from the more acidic R-hydrogens will predominate
(Rappe and Sachs, 1967). In acidic and basic conditions,
1-hydroxy-2-propanone exhibited peaks in the alkene
region; however, unlike reducing sugars (Yaylayan and
Ismail, 1995), no evidence of enolization was observed
the keto (1726 cm-1) band to shift to 1720 cm-1
.
However, neither enol nor enediol bands were observed
in the alkene region, indicating the presence of very low
concentrations of enol or enediols forms in the equilib-
rium mixture under basic conditions. This might be due
to the enhanced acidity (∼ 4-5 pKa units) of enediol
protons relative to R-hydrogens of their corresponding

2338 J. Agric. Food Chem., Vol. 47, No. 6, 1999
Yaylayan et al.
F igu r e 4. Time-dependent second-derivative spectra (1660-
1760 cm-1) of a equimolar mixture of amylamine and 1-hy-
droxy-2-propanone at room temperature.
F igu r e 3. Absorption of the carbonyl and enediol region
(1660-1760 cm-1) of a 2% solution of 1-hydroxy-2-propanone
in (a) D2O (solid line), (b) triethylamine (dotted line), and (c)
NaOD (dashed line).
carbonyl forms (Bender and Williams, 1966). Although
enol and enediol forms were not detected, they can be
trapped as stable trimethylsilyl enol ethers using tri-
methylchlorosilane (TMS), a reagent commonly used for
this purpose under alkaline conditions. A solution of
1-hydroxy-2-propanone in excess trimethylchlorosilane
dissolved in acetonitrile was monitored at 30 °C for 6
h, after addition of a catalytic amount of triethylamine
to initiate enolization. As expected, no aldehydic peak
was detected due to trapping by TMS of the enediol
intermediate (1b) leading to it. By the time of the first
scan, a new band had appeared at 1698 cm-1, in
addition to the keto band at 1724 cm-1, indicating the
formation of trimethylsilyl ether of 1b instead of alde-
hyde 1c. However, over time, the keto band at 1724 cm-1
decreased in intensity with an increase of both 1698
cm-1 and a new band at 1714 cm-1. The latter was
assigned to the trimethylsilyl ether of 1a .
To study the effect of temperature on enolization,
1-hydroxy-2-propanone solutions in TEA were equili-
brated separately for 20 min at 30 and 60 °C, and the
relative percentages of keto and aldehydo forms were
estimated based on the peak heights of the second-
derivative spectra. At 30 °C, the 1-hydroxy-2-propanone
solution exhibited ∼40% keto form and ∼60% aldehydo
form. However, at 60 °C, the percent of keto form
increased to ∼55%, indicating a slower rate of enoliza-
tion at higher temperatures. At higher temperatures,
the rate of enolization was also slowed in reducing
sugars (Yaylayan and Ismail, 1995). It is interesting to
note that under acid catalysis the rate of enolization
increased with increasing temperature due to different
mechanisms of catalysis (Lienhard and Wang, 1969).
Mon itor in g Ca r bon yl-Am in e Rea ction s of 1-Hy-
d r oxy-2-p r op a n on e. To provide further evidence for
the formation of 2-hydroxypropionaldehyde (1c) in the
presence of basic amines, 1-hydroxy-2-propanone was
reacted with amylamine (a reactive 1° amine) and
pyrrolidine (a reactive 2° amine) in separate time-run
experiments. These reactive amines, in addition to their
ability to catalyze the enolization of 1-hydroxy-2-pro-
panone (Bender and Williams, 1966), can also react with
carbonyl compounds and form stable Amadori and
Heyns rearrangement products (Scheme 2). Conse-
quently, the carbonyl peaks due to 1 and 1c should
decrease over time with concomitant formation and an
increase of new peaks in the carbonyl region during
F igu r e 5. Time-dependent second-derivative spectra (1660-
1760 cm-1) of a equimolar mixture of pyrrolidine and 1-hy-
droxy-2-propanone at room temperature
time-run experiments. When equimolar amount of amyl-
amine was added to neat 1-hydroxy-2-propanone, di-
rectly on the IR cell, the keto band of 1-hydroxy-2-
propanone centered at 1726 cm-1 decreased rapidly over
time with the formation and increase of a new band at
1711 cm-1 as shown in Figure 4. The reaction was
completed within 35 min. The same experiment was
repeated with pyrrolidine. Again, the keto band of
1-hydroxy-2-propanone decreased with the formation of
a new band at 1712 cm-1. However, at the end of 35
min, a new band at 1724 cm-1 started to form as shown
in Figure 5. In both systems, the aldehydo band
expected at 1734 cm-1 was not detected. This could
mean either enolization did not occur fast enough or the
aldehyde formed reacted very fast. To study this inter-
action in more detail, the reaction of pyrrolidine with
1-hydroxy-2-propanone was further investigated at a
lower initial temperature, and the reactants were mixed
prior to FTIR analysis to avoid diffusion control of the
reaction rates. Both reactants were cooled separately
to -2 °C and vortexed outside of the IR cell for 20 s,
and a sample was removed for time-run analysis. The
first scan, obtained 5 min after mixing, indicated the
formation of a new peak at 1712 cm-1 and the presence
of an unreacted keto peak of the 1-hydroxy-2-propanone
at 1726 cm-1 (see Figure 6). The second scan, obtained
10 min after mixing, indicated the complete disappear-
ance of the keto peak and the formation of a new band
centered at 1724 cm-1. For the next 50 min both new
peaks increased continuously as shown in Figure 6.
Monitoring the reaction for another hour indicated that

FTIR Analysis of Acetol Reaction with Amines
J. Agric. Food Chem., Vol. 47, No. 6, 1999 2339
F igu r e 6. Time-dependent second-derivative spectra (1660-
1760 cm-1) of a equimolar mixture of pyrrolidine and 1-hy-
droxy-2-propanone mixed at -2 °C and vortexed before
monitoring at room temperature.
the new peaks reached their optimum concentrations
after around 80 min. A small amount of D2O was added
to the IR cell after the final spectrum had been collected
to identify the presence of aldehydo bands. The band
at 1712 cm-1 disappeared due to hydration, and the
band at 1724 cm-1 was shifted to 1719 cm-1 due to
hydrogen bonding with D2O. A similar shift of the keto
band and disappearance of the aldehyde peak was also
observed when D2O was added to a solution of 1-hy-
droxy-2-propanone in TEA. Consequently, the band
centered at 1712 cm-1 was assigned to the Heyns
product (3).
When 1-hydroxy-2-propanone was equilibrated in
TEA, a nonreactive tertiary amine, FTIR analysis
indicated the presence of an aldehyde (1c) and a ketone
(1) in the equilibrium mixture (Figure 3). Decreasing
the temperature favored the enolization process and the
formation of the aldehyde (1c). When a reactive primary
amine (amylamine) was added to the neat 1-hydroxy-
2-propanone, on the IR cell, and the reaction mixture
was immediately scanned, only one carbonyl product
was formed on the expense of 1-hydroxy-2-propanone
(see Figure 4). Disappearance of this band after the
addition of D2O to the IR cell at the end of the last scan
indicated the formation of an aldehyde (Heyns product
3) rather than a ketone. Evidently, the rate of nucleo-
philic addition of the primary amine reaction with
1-hydroxy-2-propanone was much faster than the rate
of its catalysis of enolization. As a result, only a Heyns
product was formed. When the primary amine was
replaced by a less nucleophilic but a more basic second-
ary amine (pyrrolidine), enolization was catalyzed to a
small extent to generate toward the end of the reaction
a new peak centered at 1724 cm-1 in addition to the
major peak at 1712 cm-1. The rate of formation of the
new peak (1724 cm-1) increased, as expected, by lower-
ing the initial temperature of the reaction due to the
decreased rate of the nucleophilic addition and the
increased rate of enolization. Lowering the initial tem-
perature allowed the enolization to proceed to a greater
extent relative to the rate of nucleophilic addition and
establish higher equilibrium concentrations of 2-hy-
droxypropionaldehyde (1c). This enabled the amine to
react and produce the Amadori product 5. The data
indicated that within 5 min all the ketone 1 and the
aldehyde 1c have already been converted into their
respective tetrahedral intermediates 2 and 4 and that
the rate of conversion of tetrahedral intermediate into
F igu r e 7. Mass spectra of two products extracted from the
reaction mixture of pyrrolidine and 1-hydroxy-2-propanone. A,
Amadori product (5); B, Heyns product (3).
iminium ion and subsequently into 3 and 5 is much
slower. Consequently, the rate of increase of the bands
centered at 1712 and 1724 cm-1 over time, shown in
Figure 6, corresponds to the rate of conversion of the
tetrahedral intermediates into their corresponding
Heyns (3) and Amadori (5) products. This observation
is consistent with the literature data (Williams and
Bender, 1966), which indicate that the conversion of the
tetrahedral intermediate into imine is the rate-limiting
step during the reaction of acetone with reactive amines.
To provide direct evidence for the formation of Heyns
(3) and Amadori products (5), the reaction solution was
extracted with chloroform and analyzed by GC/MS. The
analysis indicted the presence of two products with
molecular weights of 127 amu and with mass spectral
fragmentation patterns consistent with the assigned
structures (see Figure 7).
ACKNOWLEDGMENT
V.Y. acknowledges funding for this research by the
Natural Sciences and Engineering Research Council of
Canada (NSERC).
LITERATURE CITED
Back, D. M.; Michalska, D. F.; Polavarpu, P. L. Fourier
Transform Infrared spectroscopy as a potential tool for the
study of carbohydrates in aqueous solutions. Appl. Spectrosc.
1984, 38, 173-178.
Bender, M. L.; Williams, A. Ketimine intermediates in amine-
catalyzed enolization of acetone. J . Am. Chem. Soc. 1966,
88, 2502-2508.

2340 J. Agric. Food Chem., Vol. 47, No. 6, 1999
Yaylayan et al.
Coleman, W. M., III; Gordon B. M. Examination of the matrix
isolation infrared spectra of organic compounds: part XIII.
Appl. Spectrosc. 1989, 43, 305-310.
Williams, A.; Bender, M. L. Studies on the mechanism of oxime
and ketimine formation. J . Am. Chem. Soc. 1966, 88, 2508-
2513.
Yasuda, T.; Okuyama, M.; Tanimoto, N.; Sekiguchi, S.; Ikawa,
S. Intramolecular hydrogen bonding and molecular confor-
mation. J . Chem. Soc., Faraday Trans. 1995, 91, 3379-
3383.
Yaylayan, V.; Huygues-Despointes, A. Chemistry of Amadori
Rearrangement products. C. R. C. Crit. Rev. Food Sci. Nutr.
1994, 34, 321-369.
Yaylayan, V. A.; Ismail, A. A. Determination of the effect of
temperature on the concentration of keto form of D-fructose
by FT-IR spectroscopy. J . Carbohydr. Chem. 1992, 11 (2),
149-158.
Yaylayan, V. A.; Ismail, A. A. Investigation of the enolization
and carbonyl group migration in reducing sugars by FTIR
spectroscopy. Carbohydr. Res. 1995, 276, 253-265.
Yaylayan, V. A.; Harty-Majors, S.; Ismail, A. A. Investigation
of the mechanism of dissociation of glycolaldehyde dimer
(2,5-dihydroxy-1,4-dioxane) by FTIR spectroscopy. Carbo-
hydr. Res. 1998, 309, 31-38.
Davis, L. The structure of dihydoxyacetone in solution. Bioorg.
Chem. 1973, 2, 197-201.
Hamilton, D. S.; Creighton, D. J . Inhibition of glyoxalase I by
the enediol mimic S-(N-hydroxy-N-methylcarbamoyl)glutha-
thione. The possible basis of a tumor-selective anticancer
strategy. J . Biol. Chem. 1992, 267, 24933-36.
Kobayashi, Y.; Igarashi, T.; Takahashi, H.; Higasi, K. Infrared
and Raman Studies of the Dimeric Structure of 1,3-Dihy-
droxyacetone, D(+)- and DL-Glycerladehyde. J . Mol. Struct.
1976, 35, 85-99.
Leinhard, G. E.; Wang, T.-C. On the mechanism of acid-
catalyzed enolization of ketones. J . Am. Chem. Soc. 1969,
91, 1146-1153.
Lorimer, G. H.; Chen, Y.-R.; Hartman, F. C. A role for the
-amino group of lysine-334 of ribulose-1,5-bisphosphate
carboxylase in the addition of carbon dioxide to the 2,3-
enediol(ate) of ribulose-1,5-bisphosphate. Biochemistry 1993,
32, 9018-24.
Rappe, C.; Sachs, W. H. Enolization of ketones III. Rate and
orientation of base-catalyzed deuteration of some methyl-
ketones. J . Org. Chem. 1967, 32, 4127-4128.
Thornalley, P.; Wolf, S.; Crabe, J .; Stern, A. The autoxidation
of glyceraldehyde and other simple monosaccharides under
physiological conditions catalyzed by buffer ions. Biochim.
Biophys. Acta. 1984, 797, 276-287.
Received for review November 23, 1998. Revised manuscript
received March 24, 1999. Accepted April 2, 1999.
J F9812836