Pyridoxamine, a scavenger agent of carbohydrates

Article abstract of DOI:10.1002/kin.20223

Pyridoxamine has been found to inhibit protein glycation and to avoid the formation of advanced glycation end-products (AGEs). One of the mechanisms by which pyridoxamine can inhibit glycation involves the scavenger of carbonyl groups with glycation capacity. In this work, we conducted a kinetic study of the reactions of pyridoxamine with various carbohydrates under physiological pH and temperature. The reactions involving hexoses were found to give a tricyclic compound (5) in addition to pyridoxal and pyridoxine. Such a tricyclic compound inhibits the Amadori rearrangement and the formation of other carbonyl compounds with glycating properties. The reactions involving pentoses gave compound 7 and pyridoxal - by transamination of the Schiff base. The transamination reaction enhances the inhibitory action of pyridoxamine. The formation rate constants for the Schiff base, k₃, were found to be similar to those for the reactions of D-glucose with amino acids, which suggests competition between pyridoxamine and terminal amino residues in proteins for glycating sites in sugars. These constants are dependent on the electrophilic character of the carbonyl carbon in the carbohydrate.

Full text of DOI:10.1002/kin.20223

Pyridoxamine, A Scavenger
Agent of Carbohydrates
´
˜
MIQUEL ADROVER, BARTOLOME VILANOVA, FRANCISCO MUNOZ, JOSEFA DONOSO
Institut Universitari d’Investigacio´ en Cie`ncies de la Salut (IUNICS), Departament de Qu´ımica,
Universitat de les Illes Balears, E-07122 Palma de Mallorca, Spain
Received 5 June 2006; revised 10 October 2006; accepted 13 October 2006
DOI 10.1002/kin.20223
Published online in Wiley InterScience (www.interscience.wiley.com).
ABSTRACT: Pyridoxamine has been found to inhibit protein glycation and to avoid the formation
of advanced glycation end-products (AGEs). One of the mechanisms by which pyridoxamine
can inhibit glycation involves the scavenger of carbonyl groups with glycation capacity. In this
work, we conducted a kinetic study of the reactions of pyridoxamine with various carbohydrates
under physiological pH and temperature. The reactions involving hexoses were found to give
a tricyclic compound (5) in addition to pyridoxal and pyridoxine. Such a tricyclic compound
inhibits the Amadori rearrangement and the formation of other carbonyl compounds with
glycating properties. The reactions involving pentoses gave compound 7 and pyridoxal—by
transamination of the Schiff base. The transamination reaction enhances the inhibitory action
of pyridoxamine. The formation rate constants for the Schiff base, k3, were found to be similar
to those for the reactions of D-glucose with amino acids, which suggests competition between
pyridoxamine and terminal amino residues in proteins for glycating sites in sugars. These con-
stants are dependent on the electrophilic character of the carbonyl carbon in the carbohydrate.
ꢀ
C
2007 Wiley Periodicals, Inc. Int J Chem Kinet 39: 154–167, 2007
carbonyl group in the extended form of the sugar (or a
sugar derivative) that leads to the formation of a Schiff
base. The Schiff base is stable for a limited time after
which it undergoes internal rearrangement to a more
stable, ketoamine product: an Amadori compound [5].
The Amadori compound in turn can undergo various
complex transformations including oxidative decom-
position, rearrangement, condensation, and fragmen-
tation that ultimately produce advanced glycation end-
products (AGEs) in an irreversible manner [6].
By oxidation, the sugars, the Schiff base, and/or
the Amadori compound can form AGE precursors and
radical species accelerating the glycation process. The
formation of AGEs involves structural and functional
changes in proteins. Several studies have shown that
changes in protein structure by the effect of the for-
mation of AGEs are behind the development of vari-
ous pathologies associated with hyperglycemia (e.g.,
Alzheimer’s disease [7], arteriosclerosis [8], cataracts
[9], and kidney disorders [10]).
INTRODUCTION
The reaction between amino acids and sugars was first
reported by Maillard [1] and was supposed to be the
cause of the formation of colored compounds during
the cooking of food. However, it was not until 1949
that Sattler reported the end-products of the reaction
to be toxic [2]. Subsequently, the process was found to
also occur in biological processes, named as a nonen-
zymatic glycation of proteins [3,4], and was found to
be associated with various diseases.
As can be seen in Scheme 1, the first step in a nonen-
zymatic glycation process is a reversible reaction be-
tween the terminal amino residues in the protein and the
Correspondence to: Bartolome´ Vilanova; e-mail: bartomeu.
vilanova@uib.es.
Contract grant sponsor: DGICYT.
Contract grant number: CTQ 2005-00250.
2007 Wiley Periodicals, Inc.
c
ꢀ

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
155
cation processes). These reactions have been widely
examined in kinetic terms [19–21] and revealed that
most terminal amino residues have a similar affinity
for a given glycating species. Other kinetic studies
have examined the hemoglobin glycating properties
of various carbohydrates [22].
Most current research in this field is aimed at find-
ing inhibitors of the AGEs formation in order to restrict
their accumulation on proteins. The first such potential
inhibitors to be studied were aminoguanidine [23] and
carnosine [24], which were found to effectively scav-
enge glycating carbonyl groups. Subsequent studies
examined species, such as vitamin C and pyridoxine,
which can chelate metal ions acting as catalysts for the
electron transfer involved in the formation of AGEs
[25]. Work by Hudson and coworkers [26–29] showed
that pyridoxamine (one of the six natural vitamers of vi-
tamin B6) is a powerful in vivo and in vitro inhibitor for
AGEs formation. Onorato and coworkers [30] showed
that pyridoxamine can also inhibit ALEs formation.
The inhibitory action of pyridoxamine may be exerted
by (a) inhibiting the oxidation of the Amadori com-
pound via chelation of the metal ions that catalyze it
[29]; (b) scavenging carbonyl compounds with glycat-
ing properties [26] such as D-glucose or its autoxida-
tion products (glyoxal and glycoaldehyde) [27]; and (c)
scavenging oxygen radicals by donating the hydrogen
atom of the phenol group in pyridoxamine [26]. Re-
cently, a new generation of pyridoxamine derivatives
including dmaPM, diPM [25], and BST-4997 [31] has
been found to exhibit enhanced inhibitory properties
in various in vivo tests.
There have been virtually no kinetic studies on the
interactions between pyridoxamine and sugars. In pre-
vious work [32], our group studied the kinetics of the
interaction between pyridoxamine and D-glucose in a
physiological pH and temperature and detected a tri-
cyclic compound (5 in Scheme 2) as the main reac-
tion product; such a product precluded the formation
of other carbonyl sites with glycating properties from
the Schiff base. In addition, compound 5 decomposed
into pyridoxal and pyridoxine, which proved capable
of inhibiting glycation via other mechanisms [25,28].
In this work, we conducted a kinetic study under
physiological pH and temperature of the reaction of
pyridoxamine with various aldoses (hexoses and pen-
toses), some of them play a central role in the formation
of AGEs. Major reaction products were isolated, iden-
tified, and compared with those resulting from the in-
teraction between pyridoxamine and D-glucose. Based
on the results, a general kinetic mechanism for the in-
teraction of pyridoxamine with hexoses and pentoses
is proposed, and the kinetic constants of which are
determined. The experimental results were validated
Scheme 1
The chemical nature of AGEs depends on those
of the parent glycating carbohydrate and the protein
residue. Thus glucosepane [11], which is deemed the
most important AGE, and crossline [12] come from
hexoses, whereas vesperlysine [9] comes from pen-
toses. The dimers GOLD, GOLA, and MOLD are only
formed with glyoxal or methylglyoxal—both of which
result from the decomposition of hexoses [13]—as
the glycating agent. On the other hand, pentosidine
[14] results from the interaction of arginine and ly-
sine with pentoses (usually D-arabinose or D-ribose)
and hexoses. By contrast, the nature of some AGEs is
independent of those of the sugar and protein; thus,
Nε-(carboxymethyl)lysine (CML) results from the ox-
idative degradation of the Amadori compound [15]
formed between lysine and an aldose, and K2P, which
was recently detected in retinal proteins [16], forms
whatever the nature of the parent sugar.
In addition, glycation in hyperglycemic patients has
been found to involve nucleotides and alter DNA pat-
terns in a way dependent on the individual’s age and
also to induce congenital abnormalities in children of
diabetic mothers [17]. Likewise, proteins can be altered
through lipoxidation when the oxidant is a lipid, in
which case the resulting products are advanced lipoxi-
dation end-products (ALEs) [18].
The glycation process has been the subject of much
research aimed at unraveling its chemical mechanism
[19–22]. So far, most studies have focused on the for-
mation of the Schiff base (particularly on those formed
by various amino acids and proteins with glucose,
which is the sugar most frequently involved in gly-
International Journal of Chemical Kinetics DOI 10.1002/kin

156
ADROVER ET AL.
Scheme 2
by calculating the net natural bond orbital (NBO)
charge on the carbonyl carbon in the extended forms of
the aldoses using the B3LYP functional in conjunction
with the 6-31+G∗ basis set.
High-Performance Liquid Chromatography
High-performance liquid chromatography (HPLC)
analyses were conducted on a Shimadzu-LC 10AT
chromatograph equipped with a Rheodyne 7725i uni-
versal injector and a Shimadzu SPD-M6A UV/Vis
photodiode array detector. The column was a Tracer
Excel 120 ODSB model (25 × 0.46 cm, 5 µm). A
gradient of MeCN/water-50 mM potassium phos-
phate (pH 6.0) was used to separate the target
compounds.
EXPERIMENTAL
Material
D-Arabinose, D-lyxose, D-mannose, D-ribose, D-
glucose, D-xylose, pyridoxamine (2), pyridoxal (6),
pyridoxine (8), 4-pyridoxic acid (9), and D2O (99.9%
D) were purchased from Sigma-Aldrich Co. (Madrid,
Spain); and D-allose, D-altrose, D-galactose, D-gulose,
and D-talose were obtained from Acros Organics
(Geel, Belgium) (see Structural formulae 1). All
were used as received. The buffering material was
reagent grade, and freshly boiled distilled water was
used throughout. Reaction kinetics was studied by
monitoring changes in HPLC peak area for compounds
2, 5, 6, 7, 8, 9, and 10 at 321 nm. Pyridoxamine and
carbohydrates were used at a concentration of 0.005
and 0.2 M, respectively. The reaction mixture was
buffered at pH 7.4 with a 0.5 M phosphate solution and
kept at 37◦C. Sodium azide at a 0.02% concentration
was employed to prevent bacterial growth.
NMR Spectroscopy
NMR spectra were recorded on a Bruker AMX-300
spectrometer, using sample tubes 5 mm in diameter
and 3-(trimethylsilyl)-1-propanesulphonic acid (DSS)
1
as internal reference. All chemical shifts for H (δH)
and 13C (δC) are given in ppm. The solutions in D2O
were stabilized at pD 7.4 (pD = −log [D+]) by us-
ing a 0.5 M phosphate buffer. 13C multiplicities were
determined by using the distortionless enhancement
bipolarization transfer (DEPT) method [33], and 13C
1
and H correlations from two-dimensional heteronu-
clear measurements (H,C-COSY) [34].
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
157
Structural formulae 1
the different compounds present in the solutions were
isolated by liquid chromatography and freeze-dried.
The resulting residues were characterized spectroscop-
ically.
LC–MS
Mass analyses were performed on an Agilent 1110
Series LC–MS instrument. The mobile phase was
5 mM ammonium acetate at pH 6.0—NH4OAc
initiates the ionization for MSD detection—and cir-
culated at a flow-rate of 0.3 mL/min. The mass
spectra for the isolated compounds were obtained by
using a columnless flow injection analysis (FIA). No
Schiff base for the hexoses was isolated; therefore, their
mass spectra were obtained by injecting the reaction
mixture through a Tracer Excel 120 ODSB column
(25 × 0.46 cm, 5 µm). Mass spectral detection of the
compounds was done by using an electrospray ion-
ization interface and a quadrupole mass analyzer. The
mobile phase was nebulized into an electrospray mass
analyzer by using gaseous nitrogen at 350◦C at a flow-
rate of 13 mL/min. The detector was used to count
positive ions in the scan mode over the m/z range 100–
800. A nebulization pressure of 415.6 kPa, a fragmen-
tor voltage of 80 V, and a capillary voltage of 3000 V
were used.
3,3ꢁ,4,4ꢁ,5,6-Hexahydro-5ꢁ,6-bis(hydroxymethyl)-8ꢁ-
methylspiro[pyran-2,2ꢁ-pyrido[4,3-e][1,3]oxazine]-
3,4,5-triol (5). The chromatograms obtained for the
reaction mixtures of pyridoxamine with hexoses ex-
hibited a signal at tR 20.3 min (Fig. 1). This compound
was previously isolated, and its spectroscopic signals
were found to coincide with those of 5, which was
characterized in previous work [32] (see Table I).
2-[(3-Hydroxy-5-(hydroxymethyl)-2-methylpyridin-
4-yl)methylamino]acetic Acid (7). After a reaction
time of 30 days, an aliquot of the reaction mixture of
D-lyxose and pyridoxamine was purified by HPLC to
obtain 7. HPLC: tR 13.6 min (Fig. 2). For 1H and 13
C
NMR data, see Table I. EI–MS: 227.1 ([M + H]+).
The signals for 13C in the CH2-C(3ꢁ) group of the
isolated compound were shifted downfield relative to
compound 2 (δ(C) 38.9 ppm in 2 versus 46.4 ppm in
7). The 1H and 13C chemical shifts of the other signals
for the pyridine ring were similar to those obtained
for compound 2. Based on the DEPT and H,C-COSY
results, C(1) in compound 7 is a secondary carbon
the signal for which is shifted upfield relative to the
tertiary C(1) atom in 1 (δ(C) is 94.9 ppm for 1 and
51.0 ppm for 7). Also, C(2) in compound 7 belongs
to a carbonyl group since its 13C signal is shifted
Isolation and Identification of Products
A volume of 30 mL of each reaction mixture (0.2 M
in carbohydrate, 0.005 M in pyridoxamine, and 0.5 M
in phosphate buffer at pH 7.4) was heated at 37◦C.
Aliquots of 10 mL were collected at different times
for each carbohydrate—when the concentration of the
target analytes was maximum. In a subsequent step,
International Journal of Chemical Kinetics DOI 10.1002/kin

158
ADROVER ET AL.
Figure 1 Time-dependent HPLC chromatograms for the reaction between 0.005 M pyridoxamine (2) and 0.2 M D-allose
in phosphate-buffered aqueous solutions at pH 7.4 at 37◦C. UV/Vis detection at 321 nm. For compounds structures, see
Scheme 2.
downfield (δ(C) is 74.3 ppm for 1 and 174.2 ppm for
7). Based on these results, the structure of 7 must
coincide with that shown in Scheme 2, which was
previously proposed by Glomb and Pfahler for the
product of the reaction between pyridoxamine and
glyoxal [35].
is 38.9 ppm for 2 and 57.3 ppm for 10) by the effect
of the primary amino group in pyridoxamine becom-
1
ing a tertiary group in 10. The H and 13C shifts of
the other signals for the pyridine ring are essentially
similar to those for compound 2. Based on the DEPT
and H,C-COSY results, atoms C(1) and C(7) in com-
pound 10 are quaternary atoms with downfield-shifted
signals relative to the tertiary C(1) atom in compound
1 (δ(C) is 94.9 ppm for 1, and 100.2 and 100.1 ppm
for C(1) and C(7), respectively, in 10). These results
suggest that compound 10 results from condensation
of pyridoxamine with two aldose molecules. A solu-
tion of the isolated compound (10 in 0.5 M phosphate
buffer at pH 7.4) was heated at 37◦C. After 5 days of
Compound 10. After a reaction time of 3.5 days, an
aliquot of the reaction mixture of D-talose and pyri-
doxamine was purified by HPLC (tR 44.6 min; Fig. 3)
to obtain compound 10. Its 1H and 13C NMR data are
shown in Table I. EI–MS: 491.1 ([M + 2H]+). The
13C signals for the CH2-C(3ꢁ) group in this compound
were shifted downfield relative to compound 2 (δ(C)
Figure 2 Time-dependent HPLC chromatograms for the reaction between 0.005 M pyridoxamine (2) and 0.2 M
D-lyxose in phosphate-buffered aqueous solutions at pH 7.4 at 37◦C. UV/Vis detection at 321 nm. For compound structures, see
Scheme 2.
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
159
Table I 1H- and 13C-NMR Data of Compounds 1, 2, 5, 6, 7, 8, 9, and 10 (at 300 MHz in D2O; δ in ppm)
1d
2
5d
6a
7
8
9
10
13C 1Hb 13C
13C
134.9
1Hb 13C 1H
13C 1Hb
13C
1H
–
13C
1H
–
13C
1H
–
13C
1H
–
C(1ꢁ)
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
135.2
–
133.8
139.4
133.8
138.3
–
Me-C(1ꢁ)
C(2ꢁ)
2.48 17.8 2.52 17.5 2.53 17.3 2.51 17.4 2.57 18.3 2.58 136.3 2.56 17.1
–
–
165.5
138.5
–
–
165.4
138.8
–
–
–
160.1
147.6
–
–
–
165.5
138.9
–
–
163.2
141.5
–
–
–
–
–
–
19.6
–
–
161.5
138.8
C(3ꢁ)
175.8
157.1 4.40 57.3
CH2-C(3ꢁ)
HOC-C(3ꢁ)
HOOC-C(3ꢁ)
C(4ꢁ)
4.31 38.9 4.46 47.5
4.42 46.4 4.85 58.8
–
–
–
–
–
–
–
–
–
–
6.62 101.7
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
147.5
147.1
141.1
147.2
147.0
208.9
CH2-C(4ꢁ)
H-C(5ꢁ)
H-C(1)
H2-C(1)
H-C(2)
HOO-C(2)
H-C(3)
H-C(4)
H-C(5)
CH2-C(5)
C(7)
H-C(8)
H-C(9)
4.73 61.7 4.73 61.4 5-6c 72.5 4.72 61.4 4.79 61.6 4.81 149.9 4.75 61.1
7.74 126.1 7.72 125.7 7.74 121.1 7.70 125.7 7.92 127.8 7.81 63.3 7.90 128.6
5.21 94.9
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
97.9
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
130.8
–
–
100.2
–
–
–
3.63 51.0
–
–
–
–
–
–
–
–
–
–
–
–
–
3.51 74.3
3.75 71.6
–
–
–
–
–
–
–
–
–
–
–
–
–
–
4.15 72.3
–
–
–
174.2
–
3.6
3.7
4.05 72.3
3.82 65.2
–
68.7
72.7
3.69 75.6
3.39 72.5
3.81 74.3
3.81 63.5
3.92 72.2
3.33 71.3
4.05 71.6
4.05 66.5
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
100.1
3.93 74.4
3.98 72.3
3.86 72.5
3.9
3.85 65.6
H-C(10)
H-C(11)
CH2-C(11)
72.8
a The data for compound 6 correspond to its hydrated form, which was the major one at the working pH [49].
b The 1H coupling constants for compounds 1, 5 and 10 could not be determined.
c This signal could not be distinguished as it was overlapped with the solvent peak. Its presence was confirmed by DEPT and H,C-COSY
analysis.
d The chemical shifts for compounds 1 and 5 were taken from a previous paper [32].
Figure 3 Time-dependent HPLC chromatograms for the reaction between 0.005 M pyridoxamine (2) and 0.2 M D-talose
in phosphate-buffered aqueous solutions at pH 7.4 at 37◦C. UV/Vis detection at 321 nm. For compounds structures, see
Scheme 2.
International Journal of Chemical Kinetics DOI 10.1002/kin

160
ADROVER ET AL.
incubation, the solution was analyzed by column LC–
MS. The chromatogram obtained exhibited a signal at
tR 44.6 min corresponding to 10 in addition to one
at tR 20.3 min corresponding to 5. The mass spec-
trum included a signal at m/z 219.0 corresponding to
[M + K]+ for D-talose. The areas of the chromato-
graphic peaks leveled off after 3 days of incubation,
ratio):
−d[2b]
dt
= k3[1b] [2b] − k−3[3]
(1)
Under such conditions, the following steady-state ap-
proximation can be used for compound 3:
suggesting that the k12/k
equilibrium had been es-
−12
d[3]
tablished. One can therefore reasonably assume that
compound 10 is formed by condensation of 5 with
another sugar molecule.
= k3[1b][2b] − k−3[3] − k4[3]
dt
− k5[3] − k6[3] = 0
from which it follows that
(2)
(3)
3-Hydroxy-5-(hydroxymethyl)-2-methylpyridine-4-
carbaldehyde (6), 4,5-Bis-(hydroxymethyl)-2-methyl-
pyridin-3-ol (8), and 3-Hydroxy-5-(hydroxymethyl)2-
methylpyridine-4-carboxylic acid (9). After a reaction
time of 100 days, an aliquot of the reaction mixture of
pyridoxamine and D-mannose was purified by HPLC
to obtain compounds 6, 8, and 9, and the structures
of which were confirmed by comparing their NMR
and EI–MS data with those for commercial samples.
HPLC: tR 32.2, 40.6, and 47.4 min for 6, 8, and 9,
respectively (see Fig. 1). 1H and 13C NMR: see Table
I. EI–MS: 168.0 ([M + H]+) for 6, 170.0 ([M + H]+)
for 8, and 184.0 ([M + H]+) for 9.
k3[1b] [2b]
[3] =
k
−3 + k4 + k5 + k6
Substitution of Eq. (3) into (1) yields
−d[2b]
dt
k3[1b][2b]
= k3[1b][2b] − k
(4)
(5)
−3 k−3 + k4 + k5 + k6
which can be rearranged to
−d[2b]
dt
k3 (k4 + k5 + k6)
[1b] [2b]
=
k
−3 + k4 + k5 + k6
Based on the equilibria between forms 1a and 1b in
glucose and forms 2a and 2b in pyridoxamine,
5-(Hydroxymethyl)-2-methyl-4- {[(2,3,4,5,6-penta-
hydroxyhexylidene)amino]methyl}pyridin-3-ol (3).
A reaction mixture containing 0.4 M hexose (1) and
0.02 M pyridoxamine (2) in 0.5 M phosphate buffer
at pH 7.4 at 37◦C was used in order to increase the
concentration of Schiff base obtained. The solution
exhibited substantial accumulation of compound 3
after 3–8 days. HPLC: tR 44.2 min. EI–MS: 331.1
([M + H]+). The Schiff bases could not be character-
ized by NMR spectroscopy since, as found in previous
work [32], they decomposed rapidly.
[1b]
K1 =
K2 =
(6)
(7)
[1a]
[2b]
[2a]
substitution of which into Eq. (5) yields
−d[2b]
dt
k3(k4 + k5 + k6)
−3 + k4 + k5 + k6
=
K1K2[1a][2a] (8)
k
Since compound 1 is present in overstoichiometric
amounts in the reaction medium, its concentration can
be assumed to remain constant. By collecting all con-
stant terms in Eq. (8), one can define a new constant kr
such that
Kinetic Rate Constants
Rate constants were determined by using the kinetic
data processing software GIT [36,37], which performs
numerical integration of the differential equations cor-
responding to a given kinetic scheme.
k3 (k4 + k5 + k6)
kr =
K1K2[1a]
(9)
k
−3 + k4 + k5 + k6
substitution of which into Eq. (8) yields
Determination of Schiff
Base Formation (k3)
−d[2b]
= kr [2a]
dt
(10)
Based on Scheme 2, the following kinetic equation is
proposed under the operating conditions used to exam-
ine the reaction between compounds 1 and 2 (in a 40:1
Taking into account that the different tautomers of pyri-
doxamine are in mutual equilibrium, substituting [2b]
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
161
as a function of [2a] into Eq. (10) yields
RESULTS AND DISCUSSION
Interaction Between Pyridoxamine
and Pentoses
−d[2a]
= kr [2a]
dt
(11)
Figure 2 shows the temporal variation of the chro-
matographic signals during the reaction between pyri-
doxamine and D-lyxose. The assignation of chromato-
graphic signals to chemical structures is described in
the Experimental section, and the compounds corre-
sponding to each number are shown in Scheme 2. The
initial chromatogram consisted essentially of a signal
corresponding to pyridoxamine (2) (tR 11.4 min). After
9 days of reaction, the chromatogram included several
signals including the previous one for pyridoxamine,
one at tR 13.6 min corresponding to compound 7, an-
other at tR 32.3 min suggesting the presence of pyri-
doxal (6) and several other, minor signals correspond-
ing to compounds that could not be isolated. After
23 days, the signal for compound 7 was strongly in-
creased at the expense of that for compound 2.
which can be integrated to
[2a]
ln
= −kr t
(12)
[2a]0
A plot of ln [2] as a function of time was linear through-
out. This allowed us to determine the pseudo-first-order
rate constants kr from the experimental temporal vari-
ation of the concentration of compound 2, using Eq.
(12). Also, Eq. (9) allowed us to determine the forma-
tion constant for the Schiff base, k3, from the initial
concentration of 1 in the reaction mixture, the equilib-
rium constants K1 and K2, and the constants k−3, k4,
k5, and k6.
The chromatograms for the reactions of pyridox-
amine with other pentoses (D-arabinose, D-ribose, and
D-xylose) coincided with those for its reaction with
D-lyxose, which suggests that the interaction between
pyridoxamine and a pentose involves the same mecha-
nism and gives 6 and 7 as major compounds.
Theoretical Methodology
Initially, the conformational study was based on the
Monte Carlo simulation with the AMBER∗ force field
[38], and the continuum method for simulating the
aqueous phase as implemented in the software Macro-
Model v. 6.0. [39]. We selected those conformational
structures with a difference of less than 15 kJ/mol
in the overall energy minimum among those ob-
tained in the simulation. The geometry of each se-
lected structure was optimized semiempirically by us-
ing the AM1 Hamiltonian [40] included in the software
AMPAC v. 5.0. [41]. The solvent effect was included
in the calculations by using the continuum method
COSMO [42] as implemented in AMPAC. Subse-
quently, the three structures of minimum energy for
each sugar were selected and reoptimized by using
the DFT methodology in conjunction with the B3LYP
functional [43] and the bases in the 6-31+G∗ set in-
cluding the polarized and diffuse basis functions re-
quired for these specific computations on our system
[44]. Solvent effects were considered by using the On-
sager model [45] in order to optimize structures in
terms of geometry and energy. The resulting structures
were then energy relaxed by using Tomasi’s polarizable
continuum model (PCM) [46] without geometry opti-
mization. Charges were calculated by the NBO anal-
ysis [47]. DFT computations were done by using the
software Gaussian98 [48] on a Silicon Graphics Origin
200 R-10000 computer.
Scheme 2 with R = H illustrates the kinetics of the
reactions between pyridoxamine and pentoses. Like all
other aldoses, dissolved pentoses exhibit a tautomeric
equilibrium between their extended forms and various
cyclic forms that are defined by a constant K1. Also,
pyridoxamine exhibits an equilibrium between its tau-
tomers with a free amino group and those with a pro-
tonated group (K2). First, the primary amino group in
pyridoxamine (2b) performs a nucleophilic attack on
the carbonyl group of the pentose in its extended form
(1b). Condensation of 1 and 2 gives a Schiff base that
exhibits no signal in the chromatogram but undergoes
two different processes, namely (a) decomposition, by
transamination [50], into compound 6 and an alditol
derivative and (b) rearrangement into an Amadori com-
pound (4) [5] that can in turn decompose by cleavage
of the C(2)–C(3) bond to give compound 7 and a car-
boxylic acid by means of an oxidative mechanism.
This bond cleavage reaction has previously been put
forward by other authors to account for the formation
of CML through interaction between sugars and lysine
residues in proteins [15,51].
Interaction Between Pyridoxamine
and Hexoses
Figure 1 shows the temporal variation of the chro-
matogram for the reaction between pyridoxamine and
International Journal of Chemical Kinetics DOI 10.1002/kin

162
ADROVER ET AL.
D-allose. The assignation of chromatographic signals
to chemical structures is described in the Experimen-
tal section. The first chromatogram exhibited a sin-
gle signal that corresponded to pyridoxamine (2) (tR
11.4 min). That obtained after 25 days of reaction con-
tained several major signals, including that for pyridox-
amine, another for compound 5 (tR 20.6 min), and a
third for compound 6 (tR 32.3 min), in addition to three
minor signals for 7 (tR 13.6 min), 8 (tR 40.5 min), and
9 (tR 47.2 min). Finally, the chromatogram obtained at
a long-reaction time exhibited significantly decreased
signals for compounds 2 and 5, and markedly increased
signals for 6–9.
The chromatograms obtained for the reactions
of pyridoxamine with D-altrose, D-galactose, and D-
mannose were qualitatively consistent with those for
its reaction with D-allose. This suggests that the reac-
tions between pyridoxamine and the previous hexoses
take place via the same mechanism—the end-products
are identical in any case. In previous work, the reac-
tion between pyridoxamine and D-glucose was found
to give the same end-products except 7, which was not
detected [32].
We also studied the reactions of pyridoxamine with
D-talose and D-gulose; the chromatograms for which
differed from those obtained with the other hexoses.
Figure 3 shows the temporal variation of the chro-
matographic signals during the reaction between pyri-
doxamine and D-talose. The initial chromatogram ex-
hibited a single signal corresponding to pyridoxamine
(2) (tR 11.4 min). The chromatogram obtained after
1 day of reaction was consistent with partial reaction
of 2 and included three major signals corresponding to
compounds 5 (tR 20.6 min), 6 (tR 32.3 min), and 10
(tR 44.4 min). That obtained after 3 days of reaction
exhibited a further decreased signal for compound 2
and increased signals for 5, 6, and 10, in addition to a
new one corresponding to 7. The chromatograms for
the reaction between pyridoxamine and D-gulose were
similar to those for the previous reaction.
Scheme 2 with R = CH2OH illustrates the kinetics
of the reactions between pyridoxamine and hexoses.
Initially, the carbonyl group in the extended form of
the hexose (1b) undergoes a nucleophilic attack from
the primary amino group in pyridoxamine (2b). Then,
condensation of 1 and 2 gives compound 3, which is a
Schiff base. Figures 1 and 3 show no trace of this com-
pound because the reactant concentrations used pre-
cluded its accumulation. The Schiff base 3 evolved in
three different ways depending on the nature of the par-
ticular hexose, namely (a) decomposition, by transam-
ination, into compound 6 and an alditol derivative; (b)
a double internal cyclization and the result being com-
pound 5, that was previously isolated and character-
ized by our group [32]—the amount of 5 obtained was
greater in the sugars with a minor extended form (viz.,
D-glucose and D-mannose); and (c) rearrangement into
an Amadori compound (4).
Compound 5 decomposes into 6, 8, and 9 as ma-
jor products; compound 9, however, can also form by
oxidation of 6 [52]. Compound 7, which comes from
4, is obtained in a higher proportion from the hexoses
possessing a higher proportion of the aldehyde form
(viz., D-allose, D-altrose, and D-galactose). On the other
hand, compounds 8 and 9 are obtained in higher pro-
portions from hexoses with lower proportions of their
free aldehyde forms (viz., D-glucose and D-mannose).
The reactions of D-gulose and D-talose with pyri-
doxamine give compound 10, which is formed by the
condensation of the secondary amino group in 5 with
another sugar molecule, followed by internal cycliza-
tion between the hydroxyl group in C(5)-CH2OH and
C(1) in the new sugar added (C(7) in 10); this, in turn,
undergoes cyclization via the OH-C(11) group by a
similar mechanism to that proposed for the formation
of 5 [32]. Therefore, in reacting with D-gulose and D-
talose, pyridoxamine can scavenge two carbohydrate
molecules per vitamer molecule.
Determination of Rate Constants
Quantitative kinetic determinations were done by ex-
amining the temporal variation of the chromatographic
signals. The temporal variation of the concentrations
of the reaction products was established by fitting
changes in chromatographic peak area to the following
equation:
[A2]0 = c2A2 + c5A5 + c6A6 + c7A7 + c8A8
+ c9A9 + c10A10
(13)
where Ai is the peak area of the signal at 321 nm
for each species, [A2]0 is the initial pyridoxam-
ine concentration, and ci are constants of propor-
tionality between areas and concentrations. Com-
mercially available 2, 6, 8, and 9 were used to
determine the following ci values from a cali-
bration plot: c2 = 0.044 0.004, c6 = 0.091 0.003,
c8 = 0.111 0.005, and c9 = 0.132 0.002. Parame-
ters c5, c7, and c10 were determined by using the
software SigmaPlot 9.00 [53] to perform a nonlin-
ear multivariate regression of Eq. (13). The param-
eter values thus obtained were c5 = 0.059 0.006,
c7 = 0.177 0.007, and c10 = 0.266 0.006.
Kinetic constants were calculated by fitting the ex-
perimental temporal concentration changes for each
compound to Scheme 2, using numerical integration
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
163
Table II Kinetic Constants at 37◦C as Obtained by HPLC for a Reaction Mixture of 0.2M 1 and 0.005M 2 in 0.5M
Phosphate Buffer (pH 7.4)
k ×
3
K ×
10−4
k
(h−1
k ×
k5 ×
k
(h−1
k
×
k
×
k
×
kd ×
1
−3
4
6
10
12
−12
Aldose
105 a (M−1 h−1 c
)
)
102(h−1
)
102(h−1
)
)
10−3(h−1
)
10−2(M−1 h−1
)
102(h−1
)
104(h−1
)
Altrose
Arabinose
Galactose
Allose
Glucoseb
Lyxose
Manose
Xylose
36
33
17
11
2.2
25
5.3
16
19
47
33
5.9 (5.8)
3.4
25
3.0
1.5
16
17
13
13
20
2.1
51
0.96
1.9
–
57
4.2
49
7.2
–
9.1
5.2
110
–
83
–
133
–
–
0.34
–
0.036
–
0.43
–
0.32
–
0.87
1.4
1.0
1.7
–
1.4
2.1
2.5
1.8
1.7
1.4
–
–
–
–
–
–
–
–
5.7
–
–
–
–
–
–
–
–
–
4.5
–
98
–
96
123
27
–
37
–
9.6
–
5.5 (6.0)
5.7 (5.5)
9.6 (9.9)
9.2 (10)
9.7 (10)
9.0 (9.7)
9.9 (10)
13 (14)
14 (14)
13 (13)
Gulose
Ribose
Talose
3.5
30
1.9
8.9
8.3
0.44
–
110
6.1
6.9
17
a The K1 values for the sugars were those reported by Miller [54].
b The values for the reaction with D-glucose were taken from a previous paper [32].
c The values in brackets were calculated from the concentration decrease in 2 with time (see the Experimental section).
of the differential equations for the scheme with the
aid of the software GIT [36,37].
At pH 7.4, the proportion of aldose with a linearly
arrangedcarbonylgroup(1b)requiredtoreactwas very
low [54]. Table II lists the mutarotation constant of each
aldose used in the fitting (K1). The tautomeric con-
stant for pyridoxamine, K2, was taken to be 3.5 × 10−2
,
which was determined in previous work [55] and was
used in similar studies [32]. Finally, the pyridoxal ox-
idation constant, k11, was taken to be (2.6 0.2) ×
10−4 h−1, which was also determined in previous work
[32].
The final concentrations of pyridoxal, pyridoxine
and 4-pyridoxic acid in the reactions of pyridoxamine
with the hexoses were all very low; this precluded accu-
rate estimation of individual formation constants and
led us to calculate a general decomposition constant
for compound 5: kd = k7 + k8 + k9. Figure 4 shows the
fitting of the experimental data for the reactions of pyri-
doxamine with D-talose and D-xylose to Scheme 2, and
Table II presents the best constant values obtained in
the fitting.
Kinetic constant k3 was also estimated from the con-
centration change of compound 2 as a function of time
(see the Experimental section). The results (Table II)
were consistent with the k3 values obtained by fitting.
Thus, k3 ranged from 5.5 × 104 to 14 × 104 M−1 h−1
.
The values for D-altrose, D-arabinose, and D-galactose
were similar to one another and smaller than those
for D-allose, D-glucose, D-mannose, D-lyxose, and D-
xylose—the k3 values for which were in turn smaller
than those for D-gulose, D-ribose, and D-talose. These
results must be a consequence of the electrophilic char-
acter of the carbonyl carbon in the sugars.
Figure 4 Time course of the decomposition of pyridox-
amine (2) in the presence of D-xylose (a) or D-talose (b), in
phosphate buffer at pH 7.4 at 37◦C. (•) 2, ( ) 5, ( ) 7, (ꢁ)
◦
ꢀ
6, and (ꢂ) 10. Solid lines indicate the best theoretical fits.
International Journal of Chemical Kinetics DOI 10.1002/kin

164
ADROVER ET AL.
Interestingly, the previous k3 values are of the same
order of magnitude as the constants for the nucleophilic
attack of various amino acids on D-glucose previously
obtained by other authors (viz., 60 × 103 M−1 h−1 for
glycine [19], 12 × 103 M−1 h−1 for phenylalanine [20],
and 15 × 103 M−1 h−1 for hemoglobin [21]). On the
other hand, the reaction constants for hemoglobin with
various aldoses and ketoses determined by Bunn and
Higgins [22] were rather different from those obtained
in this work; however, their values cannot be directly
compared with ours as they failed to consider the pro-
portions of linear carbohydrate form, pKa for the at-
tacking amino group and the number of terminal amino
residues in hemoglobin.
The k4 values obtained suggest that the Schiff bases
formed between pentoses and pyridoxamine tend to
evolve to an Amadori compound more markedly than
do those formed with hexoses. The latter evolved to
compound 5 by virtue of the tendency of sugars to
cyclicizing (k5).
Decomposition of the Schiff base by transamination
(k6) was only substantial in the reactions between pyri-
doxamine and pentoses. As can be seen from Table II,
k6 was virtually identical for all pentoses and roughly
ten times smaller for D-allose.
Constant k10 corresponds to the oxidative degra-
dation of compound 4 to give compound 7. The k10
values obtained were virtually identical for all aldoses.
Also, k12 and k−12 were virtually identical for D-gulose
and D-talose.
must be a result of a disparate electrophilic character
of the carbonyl carbon, which in turn must be related
to its electron density. In order to obtain more accurate
information in this respect, we calculated the net NBO
charge on the carbonyl group of each sugar by using
the density functional theory (DFT) as described in the
Experimental section.
Figure 5 shows the variation of the formation con-
stant for the Schiff base (k3) for each sugar with the
net NBO charge on the carbonyl carbon in the sug-
ars. The results reveal a correlation between k3 and the
net NBO charge on the carbonyl carbon, C(1), in the
extended form of each sugar. The fact that D-gulose
and D-talose possess a much higher net charge than the
other hexoses may explain why these are the only two
sugars giving compound 10—the electrophilic char-
acter of the carbonyl carbon in the other sugars may
not be high enough to react with the secondary amino
group in 5. The absence of compound 10 among the
reaction products for D-ribose (a sugar with a similar
electron density) is quite consistent with the previously
noted fact that the Schiff bases of the pentoses evolved
to an Amadori compound (4) rather than to the tricyclic
compound 5.
Figure 6 shows the three structures of minimum en-
ergy identified by using the theoretical methodology
described in the Experimental section. The results sug-
gest that the variation of the electron density on the
carbonyl carbon in each aldose, C(1), must be related
to the hydrogen bond network formed by the linear
forms of the sugars.
In structure A, which corresponds to the energy
minimum for D-gulose, the oxygen atom in the car-
bonyl group forms a hydrogen bond with the hy-
droxyl group on C(2) that in turn is a member of a
Theoretical Calculations
As noted earlier, the disparate reactivity of the sug-
ars toward the nucleophile (pyridoxamine in our case)
Figure 5 Relationship between the rate constant of condensation of monosaccharides with pyridoxamine (k ) and NBO charges
3
on the carbonyl group of the reducing form of hexoses and pentoses.
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
165
CONCLUSIONS
The results obtained in this work for the reactions
of pyridoxamine with various sugars—hexoses and
pentoses—reveal marked differences in their mech-
anisms depending on the nature of the carbohydrate.
Initially, both types of sugars form a Schiff base by con-
densation of the amino group in the pyridine compound
with the carbonyl group in the linear form of the sugar
(1b in Scheme 2). The calculated net NBO charge on
the carbonyl carbon clearly indicates that the reactivity
of the sugars (k3) is related to such electronic density,
which is determined by the network of intramolecu-
lar hydrogen bonds formed with the carbonyl oxygen
but independent of the length of the sugar chain. Thus,
D-gulose, D-ribose, and D-talose, which possess a net
NBO charge on the carbonyl carbon in the region of
0.423 e, have k3 values one order of magnitude greater
than those for D-altrose, D-galactose, and D-arabinose,
the NBO charge on the carbonyl carbon of which is
0.407 e.
The Schiff bases can evolve in various ways depend-
ing on the length of the sugar chain. In hexoses, the
preferential reaction is the formation of a highly stable
tricyclic compound (5) by cyclization of the carbohy-
drate portion of the Schiff base. In pentoses, however,
the reaction takes place mainly by proton rearrange-
ment and gives a highly unstable Amadori compound
(k4) that is hydrolyzed to two new carboxylic acids.
The Schiff base formed in the reactions of pyridox-
amine with pentoses also decomposes, by transamina-
tion (k6), into pyridoxal (6). The presence of pyridoxal
in the reaction mixture suggests that the transamina-
tion reaction may contribute to the inhibitory action
of pyridoxamine on AGEs, particularly in those re-
actions involving pentoses. In hexoses, however, the
transamination reaction is insubstantial as their Schiff
bases evolve largely to compound 5. Even so, their re-
action products included pyridoxal (6) and pyridoxine
(8) resulting from the decomposition of 5, and these
compounds can further inhibit glycation via previously
reported mechanisms [25,28].
The reactions of pyridoxamine with D-talose and D-
gulose—two hexoses with a low-electron density on
the carbonyl carbon—give compound 10; therefore,
each molecule of pyridoxamine must be able to scav-
enge two hexose molecules with a low charge on their
carbonyl carbon.
In any case, the k3 values obtained for the reactions
of aldoses with pyridoxamine, and the fact that no new
glycating carbonyl groups are produced, confirm the
ability of pyridoxamine to inhibit protein glycation by
scavenging glycating carbonyl groups from aldoses.
Figure 6 Structures of D-gulose (A), D-mannose (B), and
D-galactose (C). O–H distances are given in angstroms.
[Color figure can be viewed in the online issue, which is
available at www.interscience.wiley.com.]
network of four hydrogen bonds with the other hy-
droxyl groups in the aldose which reduce the electron
density on the carbonyl carbon. In structure B, which
represents the minimum for D-mannose, the carbonyl
oxygen is also a part of a hydrogen bond with a hy-
droxyl group on C(2); however, such a group forms
a network of only two hydrogen bonds with the other
hydroxyls, so the electron density on the carbonyl car-
bon is less markedly reduced than in the previous case.
Finally, structure C corresponds to the energy mini-
mum for D-galactose. As can be seen, the carbonyl
oxygen forms no hydrogen bond with any hydroxyl
group on the hydrocarbon chain; as a result, this struc-
ture exhibits the lowest net NBO charge on the carbonyl
carbon.
International Journal of Chemical Kinetics DOI 10.1002/kin

166
ADROVER ET AL.
Altering its molecular structure by increasing the nu-
cleophilic character of the amino group and/or displac-
ing the equilibrium to its deprotonated form should
enhance its inhibitory power.
21. Bunn, H. F.; Higgins, P. J. J Biol Chem 1981, 256, 5204–
5208.
22. Bunn, H. F.; Higgins, P. J. Science 1981, 213, 222–224.
23. Brownlee, M.; Vlassara, H.; Kooney, A.; Ulrich, P.;
Cerami, A. Science 1986, 232, 1629–1632.
24. Hobart, L. J.; Seibel, I.; Yeargans, G. S.; Seidler, N. W.
Life Sci 2004, 75, 1379–1389.
25. Culbertson, S. M.; Vassilenko, E. I.; Morrison, L. D.;
Ingold, K. U. J Biol Chem 2003, 278, 38384–
38394.
BIBLIOGRAPHY
1. Maillard, L. C. C R Acad Sci 1912, 154, 66–68.
2. Sattler, L.; Zerban, F. W. Int Eng Chem 1949, 41, 1401–
1407.
26. Voziyan, P. A.; Hudson, B. G. Ann NY Acad Sci 2005,
1043, 807–816.
3. Ledl, F.; Schleicher, E. Angew Chem, Int Ed Engl 1990,
29, 565–594.
27. Voziyan, P. A.; Metz, T. O.; Baynes, J. W.; Hudson,
B. G. J Biol Chem 2002, 277, 3397–6403.
4. Monnier, V. M. Arch Biochem Biophys 2003, 419(1),
1–15.
5. Yaylayan, V. A.; Huyghues-Despointes, A. Crit Rev
Food Sci 1994, 34(4), 321–369.
28. Booth, A. A.; Khalifah, R. G.; Hudson, B. G. Biochem
Biophys Res Commun 1996, 220, 113–119.
29. Voziyan, P. A.; Hudson, B. G. Cell Mol Life Sci 2005,
62(15), 1671–1681.
6. Miyata, T.; Van Ypersele De Strihou, C; Ueda, Y.;
Ichimori, K.; Inagi, R.; Onogi, H.; Ishikawa, N.;
Nangaku, M.; Kurokawa, K. J Am Soc Nephrol 2002,
12, 2478–2487.
30. Onorato, J. M.; Jenkins, A. J.; Thorpe, S. R.; Baynes,
J. W. J Biol Chem 2000, 275, 21177–21184.
31. Khalifah, R. G.; Chen, Y.; Wassenberg, J. J. Ann NY
Acad Sci 2005, 1043, 793–806.
7. Prakash Reddy, V.; Obrenovich, M. E.; Atwood, C. S.;
Perry, G.; Smith, M. Neurotoxin Res 2002, 4(3), 191–
209.
8. Kume, S.; Takeya, M.; Mori, T.; Araki, N.; Suzuki, H.;
Horiuchi, S.; Kodama, T.; Miyauchi, Y.; Takahashi, K.
Am J Pathol 1995, 147, 654–667.
32. Adrover, M.; Vilanova, B.; Mun˜oz, F.; Donoso, J. Chem
Biodivers 2005, 2, 964–975.
33. Doddrell, D. M.; Pegg, D. T.; Bendall, M. R. J Magn
Reson 1982, 48, 323–327.
34. Derone, A. E. Modern NMR Techniques for Chemistry
Research; Pergamon Press: New York, 1987.
35. Glomb, M. A.; Pfahler, C. J Biol Chem 2001, 276,
41638–41647.
9. Tessier, T.; Obrenovich, M.; Monnier, V. M. J Biol Chem
1999, 274, 20796–20804.
10. Waanders, F.; Greven, W. L.; Baynes, J. W.; Thorpe,
S. R.; Kramer, A. B.; Nagai, R.; Sakata, N.; Van Goor,
H.; Navis, G. Nephrol Dial Transpl 2005, 20(10), 2060–
2070.
11. Reihl, O.; Biemel, K. M.; Lederer, M. O.; Schwack, W.
Carbohydr Res 2004, 339, 705–714.
36. Stabler, R. N.; Chesick, J. P. Int J Chem Kinet 1978, 10,
461–469.
37. Weigert, F. J. Comput Chem 1987, 11, 273–280.
38. Weiner, S. J.; Kollman, P. A.; Case, D. A.; Singh, U. C.;
Chio, C.; Alagona, G.; Profeta, S.; Weiner, P. J Am Chem
Soc 1984, 106, 765–784.
12. Obayashi, H.; Nakano, K.; Shigeta, H.; Yamaguchi,
M.; Yoshimori, K.; Fukui, M.; Fujii, M.; Kitagawa, Y.;
Nakamura, N.; Nakamura, K.; Nakazawa, Y.; Ienaga, K.;
Ohta, M.; Nishimura, M.; Fukui, I.; Kond M. Biochem
Biophys Res Commun 1996, 226, 37–41.
13. Wells-Knecht, K. J.; Zyzak, D. V.; Litchfield, J. E.;
Thorpe, S. R.; Baynes, J. W. Biochemistry 1995, 34,
3702–3709.
14. Dyer, D. G.; Blackledge, J. A.; Thorpe, S. R.; Baynes,
J. W. J Biol Chem 1991, 266, 11654–11660.
15. Ahmed, M. U.; Dunn, J. A.; Walla, M. D.; Thorpe,
S. R.; Baynes, J. W. J Biol Chem 1988, 263, 8816–8821.
16. Cheng, R.; Feng, Q.; Ognyan, K. A.; Ortweth, B. J. Ann
NY Acad Sci 2005, 1043, 184–194.
39. Mohamadi, F.; Richards, N. G. J.; Guida, W. C.;
Liskamp, R.; Caufield, C.; Chang, G.; Hendrickson, T.;
Still, W. C. J Comput Chem 1990, 11, 440–467.
40. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart,
J. J. P. J Am Chem Soc 1985, 107, 3902–3909.
41. AMPAC v. 6.0; Semichem, Shawnee, KS, 1997.
42. Klamt, A. J. J Phys Chem 1995, 99, 2224–2235.
43. Lee, C.; Yang, W.; Parr, R. G. Phys Rev B 1988, 37,
785–789.
44. Salva`, A.; Donoso, J.; Frau, J.; Mun˜oz, F. J Phys Chem
A 2004, 108, 11709–11714.
45. Onsager, L. J Am Chem Soc 1936, 58, 1486–1493.
46. Miertus, S.; Scrocco, E.; Tomasi, J. Chem Phys 1981,
55, 117–129.
17. Bucala, R.; Model, P.; Cerami, A. Proc Natl Acad Sci
1984, 48, 105–109.
47. Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem Rev
1988, 88, 899–926.
18. Baynes, J. W.; Thorpe, S. R. Free Radic Biol Med 2000,
28(12), 1708–1716.
19. Labuza, T. P.; Baisier, W. M. J Agric Food Chem 1992,
40, 707–713.
20. Ge, S. J.; Lee, T. C. J Agric Food Chem 1997, 45, 1619–
1623.
48. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Gill,
P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman,
J. R.; Keith, T.; Petersson, G. A.; Montgomery, J. A.;
Raghavachari, K.; Al-Laham, M. A.; Zakrzewski, V. G.;
Ortiz, J. V.; Foresman, J. B.; Cioslowski, J.; Stefanov,
B. B.; Nanayakkara, A.; Challacombe, M.; Peng, C. Y.;
International Journal of Chemical Kinetics DOI 10.1002/kin

PYRIDOXAMINE, SCAVENGER AGENT OF CARBOHYDRATES
167
Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.;
Replogle, E. S.; Gomperts, R.; Martin, R. L.; Fox,
D. J.; Binkley, J. S.; Defrees, D. J.; Baker, J.; Stewart,
J. P.; HeadGordon, M.; Gonzalez, C.; Pople, J. A.
Gaussian98; Gaussian, Inc.: Pittsburgh, PA,
1998.
51. Kasper, M.; Schieberle, P. Ann NY Acad Sci 2005, 1043,
59–62.
52. Gregory, J. F.; Kirk, J. R. J Food Sci 1978, 43, 1801–
1815.
53. SigmaPlot, v. 9.00; SPSS Inc., Erkrath, Germany, 2004.
54. Dworkin, J. P.; Miller, S. L. Carbohyd Res 2000, 329,
359–365.
55. Vilanova, B.; Adrover, M.; Mun˜oz, F.; Donoso, J. Chem
Biodivers 2004, 1, 1073–1090.
49. Snell, E. E. Transaminases; Wiley-Interscience: New
York, 1985.
50. Snell, E. E. J Am Chem Soc 1945, 67, 194–197.
International Journal of Chemical Kinetics DOI 10.1002/kin