Understanding non-enzymatic aminophospholipid glycation and its inhibition. Polar head features affect the kinetics of Schiff base formation

Article abstract of DOI:10.1016/j.bmc.2011.06.018

Non-enzymatic aminophospholipid glycation is an especially important process because it alters the stability of lipid bilayers and interferes with cell function and integrity as a result. However, the kinetic mechanism behind this process has scarcely been studied. As in protein glycation, the process has been suggested to involve the formation of a Schiff base as the initial, rate-determining step. In this work, we conducted a comparative kinetic study of Schiff base formation under physiological conditions in three low-molecular weight analogues of polar heads in the naturally occurring aminophospholipids O-phosphorylethanolamine (PEA), O-phospho-dl-serine (PSer) and 2-aminoethylphenethylphosphate (APP) with various glycating carbonyl compounds (glucose, arabinose and acetol) and the lipid glycation inhibitor pyridoxal 5′-phosphate (PLP). Based on the results, the presence of a phosphate group and a carboxyl group in α position respect to the amino group decrease the formation constant for the Schiff base relative to amino acids. On the other hand, esterifying the phosphate group with a non-polar substituent in APP increases the stability of its Schiff base. The observed kinetic formation constants of aminophosphates with carbonyl groups were smaller than those for PLP. Our results constitute an important contribution to understanding the competitive inhibition effect of PLP on aminophospholipid glycation.

Full text of DOI:10.1016/j.bmc.2011.06.018

Bioorganic & Medicinal Chemistry 19 (2011) 4536–4543
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Understanding non-enzymatic aminophospholipid glycation
and its inhibition. Polar head features affect the kinetics
of Schiff base formation
⇑
Catalina Caldés, Bartolomé Vilanova , Miquel Adrover, Francisco Muñoz, Josefa Donoso
Institut Universitari d’Investigació en Ciències de la Salut (IUNICS), Departament de Química, Universitat de les Illes Balears, Ctra Valldemosa km 7.5,
E-07122 Palma de Mallorca, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Non-enzymatic aminophospholipid glycation is an especially important process because it alters the sta-
bility of lipid bilayers and interferes with cell function and integrity as a result. However, the kinetic
mechanism behind this process has scarcely been studied. As in protein glycation, the process has been
suggested to involve the formation of a Schiff base as the initial, rate-determining step. In this work, we
conducted a comparative kinetic study of Schiff base formation under physiological conditions in three
low-molecular weight analogues of polar heads in the naturally occurring aminophospholipids O-phos-
phorylethanolamine (PEA), O-phospho-DL-serine (PSer) and 2-aminoethylphenethylphosphate (APP) with
various glycating carbonyl compounds (glucose, arabinose and acetol) and the lipid glycation inhibitor
pyridoxal 50-phosphate (PLP). Based on the results, the presence of a phosphate group and a carboxyl
Received 5 April 2011
Revised 1 June 2011
Accepted 8 June 2011
Available online 16 June 2011
Keywords:
Aminophospholipid glycation
Schiff base
Pyridoxal 50-phosphate
group in
a position respect to the amino group decrease the formation constant for the Schiff base rela-
tive to amino acids. On the other hand, esterifying the phosphate group with a non-polar substituent in
APP increases the stability of its Schiff base. The observed kinetic formation constants of aminophos-
phates with carbonyl groups were smaller than those for PLP. Our results constitute an important contri-
bution to understanding the competitive inhibition effect of PLP on aminophospholipid glycation.
Ó 2011 Elsevier Ltd. All rights reserved.
1. Introduction
toxic aggregates.8 It also induces nucleic acids mutations during
DNA replication,9 and structural changes in the lipid bilayer.10
The reactions between a carbohydrate and a amino group in a
protein, lipid or nucleic acid are known as ‘non-enzymatic
glycation of biomolecules’ and their products have been deemed
responsible for a number of medical pathologies associated to
aging,1 Alzheimer’s disease,2 artherosclerosis3 and diabetes.4
Non-enzymatic glycation starts with the formation of a Schiff
base which then rearranges into a more stable ketoamine: an
Amadori product. The Amadori product subsequently evolves via
a complex series of reactions leading to the formation of a hetero-
geneous family of compounds known as ‘advanced glycation
end-products’ (AGEs). By oxidation, the sugars, the Schiff base,
and or the Amadori compound can form AGE precursors and radi-
Non-enzymatic glycation in aminophospholipids has been the
subject of study since the 1990s. The naturally occurring amino-
phospholipids acting as targets in glycation reactions are phosphati-
dylethanolamine (PE) and phosphatidylserine (PS), both of which
are present in mammal cell membranes. Bucala et al.11 were the first
to show that PE reacts with glucose to form phospholipid-linked
AGEs. Subsequently, Ravandi et al.12 examined PE glycation in red
cells and plasma from diabetic individuals and found 10–16% of all
PE in their samples to be in glycated form. Other authors have char-
acterized various AGEs formed in PE–glucose reaction models,13–16
and Fountain et al.17 quantified glucose-linked PE and PS in samples
from diabetic patients. In 2001, Breitling et al.18 identified and quan-
tified Schiff-PE and Amadori-PE adducts in erythrocytes from
healthy and diabetic individuals. The last years, Miyazawa et al. have
analysed Amadori-glycated phosphatidylethanolamine in human
erythrocytes and blood plasma.19,20 Recently, our group by using
theoretical calculations has studied Schiff base formation between
various carbonyl and amino compounds,21–24 and on the surface of
an aminophospholipid layer.25
cal species (OÅ2ꢀ and OH) which accelerate the glycation process
Å
and have pro-oxidant effects on other molecules.5 In fact, some
studies showed that glycated aminophospholipids increase oxida-
tion reactions and degradation either of the phospholipid or other
biomolecules.6,7 It has been reported that the glycation of proteins
causes fragmentation of the polypeptide chain and formation of
One of the therapeutic strategies used to prevent the diseases
derived from lipid glycation is based on the use of molecules that
can inhibit the glycation process. In 1993, Bucala et al.11 found
⇑
Corresponding author. Tel.: +34 971 173005; fax: +34 971 173426.
E-mail address: bartomeu.vilanova@uib.es (B. Vilanova).
0968-0896/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2011.06.018

C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
4537
aminoguanidine to inhibit protein and lipid glycation. Recently,
Higuchi et al.26 detected in vivo formation of PE–PLP adducts in hu-
man red cells and found the use of a PLP supplement in the diet for
diabetic rats to substantially reduce their physiological levels of
Amadori-PE adducts; based on these results, they suggested that
pyridoxal 50-phosphate (PLP, a vitamin B6 derivative) might be an
effective inhibitor of lipid glycation. Subsequently, other authors
provided in vitro confirmation that urea protects from lipid-de-
rived AGEs,27 albeit less efficiently than PLP.28
In any case, lipid glycation has been the subject of little kinetic
research.12,15,18,29 In this work, we undertook a study of the glyca-
tion kinetics of aminophospholipids analogues with a view to
improving current understanding of the mechanism behind lipid
glycation; to this end, we focused on lipid polar heads in order to
facilitate their study in aqueous media. Figure 1 shows the
structure of the three model substances used [O-phosphorylethan-
olamime (PEA), O-phospho-DL-serine (PSer) and 2-aminoethyl-
phenethylphosphate (APP)], which were reacted with three
strong glycating carbonyl compounds (glucose, arabinose and
acetol) in addition to PLP.
We also determined pKa for the amino group in APP, which was
9.00 0.05.
We examined the reactivity of the three model compounds
against three carbonyl compounds of biomedical interest, namely:
glucose, which promotes the glycation of biomolecules (particularly
in hyperglycemic individuals33); arabinose, which is a sugar formed
by glucose degradation in the body and a suggested precursor of
some AGEs;34 and acetol, which is an inhibitor of glyceraldehyde
3-phosphatedehydrogenase35 and present at elevated concentra-
tionsin diabeticketosis.36 We alsostudied thereactionsof themodel
compounds with PLP, which is another potent inhibitor of lipid
glycation by virtue of its protecting aminophospholipid polar heads
from carbonyl glycants via the formation of a Schiff base.26
2.1. Reactions between glycating carbonyl compounds and
aminophosphates
Scheme 1 shows the proposed kinetic mechanism for the reac-
tions of the model compounds with the carbonyl reagents. The
nucleophilic attack of the amino group on the carbonyl group pro-
duced a Schiff base (SB) in hydrolytic equilibrium with the reac-
tants. These reactions were studied in the presence of NaCNBH3,
which is a selective reducing agent for imino groups at neutral
pH,37,38 in order to facilitate scavenging of the resulting Schiff base
and the determination of its kinetic formation constant.30
Figure 2A shows the temporal variation of the chromatograms
for the reaction between glucose and APP. The initial chromato-
gram contained a major signal for APP (tR 14.3 min) that decreased
with time as a new signal at tR 12.3 min corresponding to the re-
duced Schiff base (redSB_1) appeared. After a long enough incuba-
tion time, a new signal at tR 10.8 min corresponding to the reaction
product between redSB_1 and a second glucose molecule, redSB_2,
was observed. All reaction products were isolated and character-
ized by 1D and 2D-NMR spectroscopy, and mass spectrometry
(see Section 5). Condensation of a second molecule of carbonyl
compound with a secondary amine (carbinolamine) was previ-
ously demonstrated by other authors.39
A comparison of our kinetic results allowed us to elucidate the
influence of the chemical nature of aminophospholipid polar heads
on their reactivity, and to assess the reactivity of PLP with amino-
phospholipid polar heads in relation to carbonyl glycants. These
results, together with others obtained in previous work,30 provide
useful insight into the inhibitory effect of vitamin B6 derivatives on
biomolecular glycation.
2. Results
Aminophospholipid glycation was studied in three model com-
pounds of low molecular weight: PEA, PSer and APP. PEA and PSer
are analogues of polar heads in PE and PS, respectively. Also, APP
has the same polar head as PEA, but possesses a hydrophobic tail
by effect of its terminal phosphate group being esterified with a
phenethyl group that confers the compound the typical amphiph-
ilicity of naturally occurring aminophospholipids. The three model
compounds are water-soluble and possess an amino group acting
as glycation target.
Accurately fitting the experimental kinetic results required the
potentiometric determination of the ionization constants of the
three model compounds at 37 °C. In this way, we found pKa for
the carboxyl group in PSer to be 2.04 0.05; the second ionization
constant for the phosphate group in PEA and PSer 5.81 0.02 and
6.02 0.05, respectively; and the ionization constant for the amino
group (pKa4 in Scheme 1) 10.24 0.02 for PEA and 10.10 0.04 for
PSer. These values are similar to previously reported values.31,32
Figure 2B shows the chromatograms for the reaction between
arabinose and APP. As can be seen, the signal for APP (tR
14.1 min) decreased with time as that for the reduced Schiff base
(tR 12.6 min) appeared. RedSB_2 formation was not observed dur-
ing the studied time frame. The results for the reaction between
APP and acetol were similar to those for the reaction with arabi-
nose (data not shown).
Figure 3 shows the temporal variation of the 1H NMR spectra for
the reaction of PEA with arabinose in the region from 3.0 to
3.45 ppm. As can be seen, the triplet at 3.21 ppm, corresponding
to H2–C(1) in PEA, decreased simultaneously with the appearance
of two signals downfield. By analogy with the 1H NMR results for
the reduced Schiff bases isolated in the reaction of APP with glu-
cose (see Section 5), which exhibited a downfield shift in the signal
for H2–C(1), these signals were assigned to the reduced Schiff bases
redSB_1 and redSB_2 from PEA and arabinose. The reaction kinetics
was monitored via temporal changes in the H2–C(1) signal for the
reactant. The reaction between PEA and acetol was studied by
monitoring changes in the H2–C(1) signal for PEA by 1H NMR.
The reaction of PEA with glucose, and those of PSer with glucose
and arabinose, was monitored by 1H, 13C-HMQC since the 1D-1H
spectra exhibited overlapped signals for the reactants. Figure 4
shows the 2D-1H,13C-HMQC spectra for the PSer–glucose reaction
mixture in the presence of NaCNBH3; the initial spectrum is shown
in blue and that recorded after 16 days of reaction in red. HC(1)–
C(1) and H2C(2)–C(2) cross-peaks in PSer decreased with time,
simultaneously with the appearance of new signals at 68 ppm for
the species contributing to the formation of redSB_1. The signals
for glucose exhibited no change.
O
4
6
9
1
O
O
5
7
8
P
NH3
2
3
O
10
APP
O
O
3
O
P
O
P
1
O
O
O
O
1
NH3
NH3
2
2
O
O
PEA
PSer
Figure 1. Structural formulae of APP (2-aminoethylphenethylphosphate), PEA (O-
phosphorylethanolamine) and PSer (O-phospho-DL-serine).

4538
C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
R4
R4
O
P
O
O
R1
R2
k1
R3
O
- H2O
+ H2O
R3
O
+
N
P
NH2
k-1
R1
R2
O
O
R1-CO-R2
SB
R-NH2
k2
NaCNBH3
R4
K4
K3
R4
O
P
O
R1
HO
R1
OR'
R3
O
R3
O
NH3
P
N
H
H
R2
R2
O
O
+
R1-C(OH)(OR')-R2
R-NH3
redSB_1
O
R1
R2
R1:
k5
(glucose),
(CH2OH)4 CH2OH
(CH2OH)3 CH2OH
(arabinose),
NaCNBH3
CH2OH
(acetol)
R2:
CH3
(glucose and arabinose),
(acetol)
H
O
R4
O
P
R1
R3:
R4:
R':
(PEA and PSer),
(PEA and APP),
O
CH2 CH2
Ph (APP)
R3
O
N
H
H
H
COO-
(PSer)
R2
O
R2
R1
(acetol), hidrocarbon chain (glucose, arabinose)
H
redSB_2
Scheme 1.
The temporal variation of the HPLC, 1D and 2D-NMR signals
was used to determine aminophosphate concentration in each
reaction. Fitting such a variation to Eq. 8 allowed the correspond-
ing observed rate constants, kobs, to be calculated. The rate con-
stants obtained from NMR values are affected by the isotopic
effect of the solvent (D2O). In order to compare the results in
H2O and D2O, we determined such an effect by using the Schiff
base formation between APP and arabinose as a model reaction.
The kobs value thus obtained was 0.35 Mꢀ1 hꢀ1 in D2O and 0.47 M
ꢀ1 hꢀ1 in H2O. The resulting isotopic effect, kH/kD = 1.3, is consistent
with previously reported values for ketimines.40 Table 1 shows the
kobs values for PEA and PSer after correction for the isotopic effect.
The microscopic kinetic formation constants for the Schiff bases
(kl) shown in Table 1 were calculated by introducing the corre-
sponding kobs values in Eq. 7.
bases (k1) of the aldehyde groups in glucose and arabinose were 3
orders of magnitude greater than that for the ketone group in ace-
tol, which is suggestive of a difference in electrophilicity between
the reactive carbonyl groups. A similar trend was previously ob-
served in their reactions with pyridoxamine (PM).42 The increased
reactivity of glucose in these systems is a result of their electrophi-
licity being raised by effect of intramolecular hydrogen bonding
net.42
APP was the least reactive of the three model aminophosphates.
In fact, its kl value was 2–5 times smaller than that for PEA and
PSer. Also, the three aminophosphates were less reactive by up
to one order of magnitude than the dipeptide Ac-Phe-Lys against
the same glycants.30 Therefore, the presence of phosphate groups
in the polar heads of aminophospholipids reduces the reactivity
of their amino groups. Also, similarly to amino acids in relation
to primary amines,43,44 the presence of a carboxyl group in PSer re-
sulted in slightly reduced formation constants for its Schiff bases in
relation to PEA.
2.2. Reactions between PLP and aminophosphates
Scheme 2 shows the kinetic mechanism for the Schiff base for-
mation between PLP and the aminophosphates studied. The forma-
tion process was monitored via the absorbance at 420 nm, which
has been associated to the formation of PLP Schiff bases.41 Fitting
the experimental values to Eq. 10 allowed the kinetic formation
(kf) and hydrolysis (kh) constant, and the equilibrium constant
(Keq), at pH 7.4 for the Schiff base of each aminophosphate to be
calculated (see Table 1).
Based on the kobs values obtained (Table 1), acetol was the
strongest glycant. The efficiency of the glycants was dependent
on the proportion of the reactive carbonyl form with respect to
the cyclic or hydrated form. Accordingly, glucose was the least effi-
cient sugar owing to the very high proportion of its non-reactive
cyclic form (99.99%).45 kobs was roughly five times greater for
APP than for PEA and PSer by effect of its increased proportion of
free amino groups at physiological pH (2.45% vs 0.14% in PEA and
0.20% in PSer). The fact that kobs was slightly smaller for PSer than
for PEA is consistent with the results of in vitro tests conducted by
Fountain et al.17 which revealed the amount of glucose-glycated
product to be greater with PE than with PS. However, these authors
found the amount of glycated PS in cell red membranes, where PE
and PS are present at a 2:1 ratio, to be three times greater than that
of glycated PE. The microscopic constants (kl) obtained in this work
suggest that the polar heads in PS are less reactive than those in PE;
3. Discussion
3.1. Influence of the chemical nature of the aminophosphate on
its chemical reactivity
The quantitative results obtained are shown in Table 1. As can
be seen, the microscopic kinetic formation constants for the Schiff

C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
4539
A
100x103
80x103
redSB_2
redSB_1
60x103
40x103
20x103
0
APP
80
60
6
8
40
20
10
12
14
16
18
0
20
22
B
redSB_1
80x103
Figure 4. 2D-1H, 13C-HMQC spectra for a reaction mixture of 20 mM PSer and
0.42 M glucose in the presence of 0.1 M NaCNBH3 at pD 7.4 at 37 °C. The blue line
represents the spectrum at t = 0 and the red line that after 16 h of reaction.
60x103
40x103
20x103
0
APP
their pKa and suggested that the amino group in PE has a slightly
higher pKa in PS-containing vesicles.46,47
8
3.2. PLP competes with carbonyl groups for aminophospholipid
polar head groups
6
6
4
8
10
12
2
14
PLP is one of the strongest inhibitor of aminophospholipid gly-
cation known to date.28 Careful study of the reactions between
aminophosphates and PLP has allowed its inhibitory effect under
physiological pH and temperature conditions to be assessed. Ta-
ble 1 shows the observed formation (kf) and hydrolysis constants
(kh) for its Schiff base and the equilibrium constant (Keq) for the
process. As can be seen, kf for PLP was 4–5 orders of magnitude
greater than the kobs values for the other carbonyl compounds un-
der physiological conditions. The high stability of the Schiff base of
PLP is a result of intramolecular hydrogen bonding between the
phenolate ion and the protonate imino nitrogen,48 which occurs
in none of the Schiff bases formed with the other carbonyl groups.
Similarly to the reactions with the previous glycants, the near-
ness of the phosphate group to the amino group in aminophos-
phates also influences their reactions with PLP. A comparison of
the results for the reactions of PLP and PSer (Table 1) with reported
data for serine49 (kf = 3.0 ꢁ 103 Mꢀ1 hꢀ1 and kh = 19 hꢀ1 at 25 °C,
which, based on the activation energies determined in similar
studies,50,51 should be roughly twice greater at 37 °C), reveals that
the presence of the phosphate group reduces kf and increases kh. In
fact, the equilibrium constants, Keq, for PEA and PSer are smaller
16
0
18
20
22
Figure 2. Time-dependent HPLC chromatograms for the reactions of 10 mM APP
with 400 mM glucose (A) and 400 mM arabinose (B) in the presence of 100 mM
NaCNBH3 in phosphate buffer at pH 7.4 at 37 °C as obtained with UV–vis detection
at 254 nm.
than those for glycine (100 Mꢀ1 50 and serine (160 Mꢀ1 49
) at
)
25 °C. These results are consistent with those of a recent study
where excess negative charge in the amino acid chain was found
to decrease the equilibrium constant for the reaction with PLP at
pH 7.35,52 as well as with the fact that Keq was smaller for PSer
than for PEA by effect of the former additionally bearing a carbox-
ylate group. This effect was previously observed in the reaction of
PLP with various amino acids and primary amines,43,44 where the
presence of the carboxyl group hindered formation of an imino
bond but facilitated hydrolysis of the resulting Schiff base.
Figure 3. 1H NMR spectra for a reaction mixture of 10 mM PEA and 200 mM
arabinose in the presence of 100 mM NaCNBH3 at pD 7.4 at 37 °C.
therefore, the differences observed in in vivo tests can be ascribed
to a change in pKa for the amino group of these compounds in the
lipid membrane. Studies on vesicles from various aminophospho-
lipid mixtures have shown that membrane composition influences
The increased kf value of APP relative to PSer and PEA can be
ascribed to its smaller pKa and its containing a non-polar group.
The kf value for APP was similar to that for tryptophan—an amino
acid also possessing an aromatic ring—under identical pH and

4540
C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
Table 1
Kinetic and equilibrium constants for the reactions of aminophosphates with glycating carbonyl compounds and PLP at pH 7.4 at 37 °C
Compound
APP
PEA
PSer
cAc-Phe-Lys
cPM
Acetol (aK3 = 4.9 ꢁ 101)
kobs (Mꢀ1 hꢀ1
k1 (Mꢀ1 hꢀ1
)
)
)
9.3 ꢁ 10ꢀ1
1.8 ꢁ 10ꢀ1d
—
—
8.6 ꢁ 10ꢀ1
2.9
)
3.9 ꢁ 101
1.3 ꢁ 102
4.3 ꢁ 102
8.5 ꢁ 101
Arabinosa (bK3 = 3.0 ꢁ 10ꢀ4
)
kobs (Mꢀ1 hꢀ1
k1 (Mꢀ1 hꢀ1
4.7 ꢁ 10ꢀ1
6.4 ꢁ 104
8.7 ꢁ 10ꢀ2d
2.0 ꢁ 105
7.8 ꢁ 10ꢀ2d
5.4 ꢁ 10ꢀ1
8.7 ꢁ 105
3.4
)
1.3 ꢁ 105
2.9 ꢁ 105
Glucosa (bK3 = 2.0 ꢁ 10ꢀ5
)
kobs (Mꢀ1 hꢀ1
k1 (Mꢀ1 hꢀ1
5.3 ꢁ 10ꢀ2
1.7 ꢁ 10ꢀ2d
9.2 ꢁ 10ꢀ3d
2.3 ꢁ 105
4.6 ꢁ 10ꢀ2
2.9 ꢁ 10ꢀ1
)
1.1 ꢁ 105
5.9 ꢁ 105
1.0 ꢁ 106
3.7 ꢁ 105
PLP
kf (Mꢀ1 hꢀ1
kh (hꢀ1
Keq (Mꢀ1
)
1.4 ꢁ 104
5.9 ꢁ 101
2.3 ꢁ 102
4.5 ꢁ 103
8.2 ꢁ 101
5.5 ꢁ 101
3.4 ꢁ 103
1.1 ꢁ 102
3.0 ꢁ 101
—
—
—
—
—
—
)
)
a
b
c
From Ref. 59.
From Ref. 45.
From Ref. 30.
d
These values are corrected for the solvent isotope effect, calculated as kH/kD = 1.3.
O
P
R4
O
R3
H
O
O
R4
H
N
O
P
R3
kf
H
O
O
+
2-O3PO
+
H2O
NH2
O
kh
2-O3PO
O
CH3
N
H
R-NH2
CH3
N
H
PLP
SB
R3:
R4:
O
CH2 CH2
(APP)
Ph
O
(PEA and PSer),
(PEA and APP),
H
COO-
(PSer)
Scheme 2.
temperature conditions (kf = 1.2 ꢁ 104 Mꢀ1 hꢀ1).50 The last result is
consistent with those of previous studies by our group, which re-
vealed the formation rate constant for the Schiff base of PLP with
the non-polar compound n-dodecylamine at neutral pH to exceed
that for polar amino acids.43,53 On the other hand, the presence of
non-polar groups hinders the hydrolysis of the imino bond (kh). As
a result of both effects, the equilibrium constant for APP was great-
er than those for PEA and PSer. Similar Keq values were previously
carboxyl group in a with the amino group further hinders the reac-
tion. Also, esterifying the phosphate group with a substituent bear-
ing a hydrophobic chain (e.g., APP, which is amphiphilic) favours
formation of the Schiff base.
The kobs values for the reactions of PLP with aminophosphates
at physiological pH and temperature were 4 orders of magnitude
greater than those for the reactions of aminophosphates with the
other carbonyl compounds. The difference can be ascribed to intra-
molecular hydrogen bonding in the Schiff base of PLP. kobs for the
reactions of PM with the glycants exceeded the values for those
involving the aminophosphates. Based on the results, amino-
phospholipid glycation can be inhibited by the action of PM and
PLP, which act synergistically via two different mechanisms. Thus,
PLP reacts reversibly with aminophospholipids to form a Schiff
base which does not evolve to an Amadori compound—and hence
prevents the formation of AGEs. On the other hand, PM reacts more
rapidly with glycating carbonyl compounds than do aminophosp-
holipids, thereby inhibiting structural alterations in the latter.
These results constitute a substantial contribution to a better
understanding of the inhibitory effect of B6 vitamer on amino-
phospholipid glycation.
reported for n-hexylamine at 30 °C (245 Mꢀ1 51
)
and e-aminoca-
proic acid at 25 °C (220 Mꢀ1).53
Pyridoxamine (PM), a B6 vitamer with a methylamino group in-
stead of the carbonyl group in PLP, is one other effective inhibitor
of glycation and lipoxidation processes in biomolecules. In previ-
ous work, we determined the kinetic constants for the reactions
of PM with various carbonyl compounds and provided quantitative
evidence for the origin of competitive inhibition of protein glyca-
tion by PM.30,54,55 As can be seen from Table 1, kl was 2–5 times
greater for PM than for APP, but only two times greater than for
PSer and 1.5 than for PEA. Also, kobs for PM (Table 1) exceeded
the values for the aminophosphates, which can be ascribed to its
increased proportion of free amino form at pH 7.4 (3.5%).42 There-
fore, PM can competitively inhibit aminophospholipid glycation
under physiological conditions.
5. Experimental
5.1. Materials
4. Conclusions
Based on the kinetic formation constants (kobs) determined in
this work, the presence of a phosphate group in the model com-
pounds studied hinders the formation of a Schiff base in relation
to amino acids and peptides, and the additional presence of a
D-Glucose, D
-arabinose, D2O (99.9% D), pyridoxal 50-phosphate
hydrate (PLP), O-phosphorylethanolamine (PEA) and O-phospho-
DL-Serine (PSer) were purchased from Sigma–Aldrich; acetol was
obtained from Fluka; sodium cyanoborohydride (NaCNBH3),

C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
4541
potassium dihydrogen phosphate and sodium hydroxide 0.1 M
standard solution were obtained from Acros Organics. All reagents
were used as received.
cose, D-arabinose and acetol) in the presence of NaCNBH3 for
selective reduction of imino groups at neutral pH and scavenging
of Schiff bases.37,38
The reactions between APP and carbonyl compounds were
studied by HPLC. Samples were mixtures of the reactants in
0.5 M phosphate buffer at pH 7.4 that were thermostated at
37 °C. Reactant concentrations were 10 mM APP, 100 mM
NaCNBH3 and 400 mM carbonyl compound—by exception, acetol
was used at 200 mM.
5.2. High performance liquid chromatography (HPLC)
HPLC analyses were conducted on a Shimadzu-LC 10AT chro-
matograph equipped with a Rheodyne 7725i universal injector
and a Shimadzu SPD-10AV UV/Vis detector. The column was a Tra-
cer Excel 120 ODSB model (25 ꢁ 0.46 cm, 5
l
m). The target com-
Reactions of PEA and PSer with carbonyl compounds were stud-
ied from 1H NMR, and 1H, 13C-HMQC spectra, all recorded at 37 °C.
Reaction mixtures were prepared in 0.5 M phosphate buffer, using
D2O at pD 7.4 as solvent throughout. The reactant concentrations
used in the reactions studied by 1D-1H NMR were 10 mM PEA
and 0.80 M acetol (or 0.20 M arabinose); and those used for 2D
pounds in the reaction mixtures were separated by using MeOH/
water-20 mM potassium phosphate (pH 6.5) in various gradient
modes (flow rate 1 ml/min) and detection at 254 nm.
5.3. UV–vis spectra
monitoring 20 mM PEA (or PSer) and 0.40 M
-arabinose). A NaCNBH3 concentration of 0.10 M was added in
all cases.
Formation of the reduced form of the Schiff base (redSB_1) was
demonstrated by using NMR in 0.5 M phosphate buffer at pD 7.4 to
D-glucose (or 0.40 M
Absorption spectra were recorded on a Shimadzu UV-2401 PC
double-beam spectrophotometer. The buffer solution background
spectrum was used as spectral reference. Quartz cells of 1 cm path
length were used to obtain electronic spectra. The cell’s tempera-
ture were adjusted at 37 0.1 °C using a Shimadzu thermostat
TCC-240A model.
D
characterize the major product of the reaction between APP and D-
glucose following isolation by HPLC after 4 days of reaction and
freeze-drying. 1H NMR: d 7.37 (m, 5H, H-C(6)–H-C(10)), 4.11 (H-
C(3)), 4.11 (H-C(20)), 3.82 (H-C(2)), 3.80 (H-C(30)), 3.80 (Ha-C(60)),
3.67 (H-C(50)), 3.64 (H-C(40)), 3.63 (Hb-C(60)), 3.20 (Ha-C(10)), 3.16
(H-C(1)), 3.13 (Hb-C(10)), 2.96 (H-C(4)); 13C NMR, DEPT-135, HSQC:
d 141.43 (s, C(5)), 131.90 (s, C(6) and C(10)), 131.37 (s, C(7) and
C(9)), 129.35 (s, C(8)), 73.56 (s, C(40)), 73.38 (s, C(50)), 73.18 (s,
5.4. NMR spectroscopy
NMR spectra were recorded on a Bruker AMX-300 spectrome-
ter, using sample tubes of 5 mm in diameter and 3-(trimethyl-
silyl)-1-propanesulphonic acid (DSS) as an internal reference. All
the chemical shifts for 1H (dH) and 13C (dC) are given in parts per
million and coupling constants (J) in Hertz. The solutions used
for products characterization were prepared in D2O, and were ad-
justed at pD 7.4 (pD = ꢀlog [D+]) with 0.5 M phosphate buffer.
Tests were of the one-dimensional 1H NMR, 13C NMR and polariza-
tion transfer (DEPT-135) and two-dimensional type (1H,13C-
HMQC).
2
C(30)), 70.72 (s, C(20)), 69.55 (d, J3,P = 5.8 Hz, C(3)), 65.36 (s,
2
C(60)), 63.08 (d, J2,P = 5.2 Hz, C(2)), 52.07 (s, C(10)), 50.22 (d,
3
3J1,P = 7.6 Hz, C(1)), 38.80 (d, J4,P = 7.4 Hz, C(4)). ESI+: m/z 486.28
[MꢀH+2K]+ (100%), 470.28 [MꢀH+K+Na]+ (24%). Assignation of
the spectral signals confirmed that the isolated product was the
Schiff base of APP with glucose (Fig. 5A).
The reaction of APP with glucose gave a second product that
was isolated by HPLC after 17 days of reaction and freeze-drying.
The relative intensities of the NMR signals for that compound were
consistent with the addition of a second glucose molecule to the
initial reduced form of the Schiff base to form redSB_2 (Fig. 5B).
13C NMR, DEPT-135, HSQC: d 141.54 (s, C(5)), 131.94 (s, C(6) and
C(10)), 131.39 (s, C(7) and C(9)), 129.36 (s, C(8)), 73.60 (s, C(40)),
73.55 (s, C(50)), 73.27 (s, C(30)), 69.89 (s, C(20)), 69.58 (d,
5.5. Synthesis of 2-aminoethylphenethylphosphate (APP)
APP was prepared by following the procedure of Argirov56 with
slight modifications. The product was isolated by decantation with
benzene to facilitate extraction of all water-soluble compounds.
The aqueous fraction was freeze-dried and redissolved in the min-
imum volume of water. This was followed by HPLC purification on
2
2J3,P = 5.8 Hz, C(3)), 65.40 (s, C(60)), 62.58 (d, J2,P = 5.4 Hz, C(2)),
3
a Kromasil 100 semi-preparative column (25 ꢁ 1.00 cm, 5
l
m) that
59.07 (s, C(10)), 57.03 (d, J1,P = 7.7 Hz, C(1)), 38.81 (d,
was eluted with a MeOH/Milli-Q water binary gradient at a flow-
rate of 2.5 ml/min for detection at 254 nm. Finally, the collected
fraction was freeze-dried to obtain a product of mp 220–223 °C
in 5% yield. 1H NMR: d 7.37 (m, 5H, H-C(6)–H-C(10)), 4.11 (q, 2H,
3JH3-H4 = 6.4 Hz, 3JH3-P = 6.4 Hz, H-C(3)), 3.79 (q, 2H, 3JH2-
3J4,P = 7.4 Hz, C(4)). ESI+: m/z 596.25 [M+Na]+ (100%), 574.26
[M+H]+ (84%).
5.7. Reaction mixtures for the kinetic study of Schiff base
formation between PLP and aminophosphates
3
3
H1 = 5.5 Hz, JH2-P = 5.5 Hz, H-C(2)), 3.07 (t, 2H, JH1-H2 = 5.1 Hz, H-
3
C(1)), 2.96 (t, 2H, JH4-H3 = 6.5 Hz, H-C(4)); 13C NMR, DEPT-135,
PLP solutions were prepared daily in 0.5 M phosphate buffer at
pH 7.4 and stored in the dark. Their exact concentrations, which
spanned the range (1.2–1.4) ꢁ 10ꢀ4 M, were determined by dilu-
tion with 0.1 M NaOH and measurement of their absorbance at
HSQC: d 141.43 (s, C(5)), 131.91 (s, C(6) and C(10)), 131.38 (s,
2
C(7) and C(9)), 129.36 (s, C(8)), 69.51 (d, J3,P = 5.8 Hz, C(3)),
2
3
64.33 (d, J2,P = 5.2 Hz, C(2)), 42.54 (d, J1,P = 8.0 Hz, C(1)), 38.83
(d, J4,P = 7.2 Hz, C(4)); ESI+: m/z 246.11 [M+H]+ (100%), 268.09
388 nm (e
= 6600 molꢀ1 cmꢀ1).57
3
[M+Na]+ (50%), 491.21 [2M+H]+ (32%), 513.22 [2M+Na]+ (44%).
Theoretical elemental analysis for C10H16NO4P (MW 245.21): C,
48.98; N, 5.71; H, 6.58; O, 26.10. Experimental Elemental Anal.:
C, 48.73; N, 5.51; H, 6.43; O, 25.86.
Solutions containing PEA, PSer or APP at a 0.1 M concentration
were made in 0.5 M phosphate buffer adjusted to pH 7.4 with
NaOH. Reaction mixtures were prepared by diluting an appropriate
volume of concentrated aminophospholipid model solution with
pre-thermostated buffer and adding PLP last. The aminophosphate
concentrations in the measuring cell were 50–260 times greater
than the PLP concentrations. Reactions were monitored by
UV–vis spectroscopy, by measuring the absorbance at 420 nm. As
confirmed in each run, the differences between the initial and final
pH of the reaction mixture never exceeded 0.04 pH units. pH
measurements were made with a Crison GLP21+ pH-meter, using
5.6. Reaction mixtures for the kinetic study of Schiff base
formation between aminophosphates and glycating carbonyl
compounds
The kinetic study of Schiff base formation with aminophos-
phates involved three different carbonyl compounds (viz. D-glu-

4542
C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
OH
A
4
6
1
O
O
5
7
8
P
NH
2
3
1'
CH2
O
10
2'
9
H
OH
H
3'
4'
5'
HO
H
OH
OH
H
6'
CH2OH
OH
H
OH
OH
OH
P
B
4
6
9
1
O
O
H2
5
7
C
CH2OH
6'
N
3'
OH
4'
5'
2'
1'
2
3
1'
CH2
H
H
H
O
8
10
2'
H
OH
H
3'
4'
5'
HO
H
OH
OH
H
6'
CH2OH
Figure 5. Structure of the compounds redSB_1 and redSB_2, obtained as products of the reaction between APP and glucose in the presence of NaCNBH3 at pH 7.4 at 37 °C.
½R1—CO—R2ꢂ
a model 52 09 glass electrode 6 mm in diameter that was previ-
ously calibrated with aqueous buffers at 37 °C.
K3
¼
ð3Þ
½R1—CðOHÞðOR0Þ—R2ꢂ
The protonation equilibrium constant for the aminophosphates can
be expressed as
5.8. Determination of ionization equilibrium constants
The ionization constants for PEA, PSer and APP were determined
by potentiometric titration with a Metrohm Titrino 718 pH-Stat of
10 ml of an 0.01 M aqueous solution of each compound adjusted to
a total ionic strength of 0.1 M at 37 °C. The titrant was 0.1 M NaOH.
Titrations were performed in duplicate and ionization equilibrium
constants calculated by using the software SigmaPlot v. 10.0.58
½R—NH2ꢂ½H3Oþꢂ
K4
¼
ð4Þ
½R—NHþ3 ꢂ
Application of a mass balance and substitution of Eq. 2 yields
ꢀd½R—NH2ꢂ
k1K3K4½R1—CO—R2ꢂT ½R—NH2ꢂT
½H3Oþꢂ þ K4 þ ½H3OþꢂK3 þ K3K4
¼
ð5Þ
dt
5.9. Determination of Schiff base formation rate constants
which can be rearranged to
Scheme 1 depicts the kinetic mechanism of the formation of the
Schiff bases of the aminophosphates with the carbonyl compounds
and their subsequent reduction by NaCNBH3 added to the solu-
tions. The rate of aminophosphate disappearance fitted the follow-
ing equation:
ꢀd½R—NH2ꢂ
¼ kobs½R1—CO—R2ꢂT ½R—NH2ꢂT
ð6Þ
ð7Þ
dt
where kobs is defined as:
k1K3K4
ꢀd½R—NH2ꢂ
kobs
¼
½H3Oþꢂ þ K4 þ ½H3OþꢂK3 þ K3K4
¼ k1½R—NH2ꢂ½R1—CO—R2ꢂ ꢀ k ½SBꢂ
ð1Þ
ꢀ1
dt
Under pseudo first-order conditions (i.e., [R1–CO–R2 ꢃ [R–NH2]),
which can be simplified to the following expression by applying the
steady-state approximation to the Schiff based under the assump-
Eq. 6 can be integrated to obtain
tion k2 >>> k
:
ꢀ1
½R—NH2ꢂt
ln
¼ kobs½R1—CO—R2ꢂTt
ð8Þ
½R—NH2ꢂ0
ꢀd½R—NH2ꢂ
¼ k1½R—NH2ꢂ½R1—CO—R2ꢂ
ð2Þ
dt
Fitting the experimental data for the reactions with the carbonyl
compounds to Eq. 8, and using the initial concentration of each, al-
lowed the observed kinetic constant (kobs) to be calculated.
Scheme 2 depicts the mechanism for the reaction between PLP
and the aminophosphates. The mathematical treatment used to
determine the kinetic constants is described in detail elsewhere.49
Briefly, integration of the kinetic equations for this mechanism and
However, each reagent was involved in another equilibrium. Thus,
carbonyl compounds in solution can be in a reactive form (R1–
CO–R2) or non-reactive form [R1–C(OH)(OR0)–R2]. Also, the latter
can be a hydrated species (R0 = H) as in acetol or a cyclic species
(R0 = hydrocarbon chain) as in glucose and arabinose. Therefore,
the equilibrium constant (K3 in Table 1) is given by

C. Caldés et al. / Bioorg. Med. Chem. 19 (2011) 4536–4543
4543
18. Breitling-Utzmann, C. M.; Unger, A.; Friedl, D. A.; Lederer, M. O. Arch. Biochem.
Biophys. 2001, 391, 245.
19. Nakagawa, K.; Oak, J.-H.; Higuchi, O.; Tsuzuki, T.; Oikawa, S.; Otani, H.; Mune,
M.; Cai, H.; Miyazawa, T. J. Lipid Res. 2005, 46, 2514.
20. Shoji, N.; Nakagawa, K.; Asai, A.; Fujita, I.; Hashiura, A.; Nakajima, Y.; Oikawa,
S.; Miyazawa, T. J. Lipid Res. 2010, 51, 2445.
21. Ortega-Castro, J.; Adrover, M.; Frau, J.; Salvà, A.; Donoso, J.; Muñoz, F. J. Phys.
Chem. A 2010, 114, 4634.
application of the Beer–Lambert law at a wavelength of zero
absorption for each aminophosphate yielded
A
A
1 ꢀ A0
ab ꢀ xxe
ln
¼ ꢀ ln
þ kap
t
ð9Þ
1 ꢀ A
xe2
where A0, A and A are the absorbances at time 0, t and equilibrium;
1
22. Salvà, A.; Donoso, J.; Frau, J.; Muñoz, F. J. Phys. Chem. A 2004, 108, 11709.
23. Salvà, A.; Donoso, J.; Frau, J.; Muñoz, F. J. Phys. Chem. A 2003, 107, 9409.
24. Salvà, A.; Donoso, J.; Frau, J.; Muñoz, F. J. Mol. Struct. (Theochem) 2002, 577, 229.
25. Solís-Calero, C.; Ortega-Castro, J.; Muñoz, F. J. Phys. Chem. B 2010, 114, 15879.
26. Higuchi, O.; Nakagawa, K.; Tsuzuki, T.; Suzuki, T.; Oikawa, S.; Miyazawa, T. J.
Lipid Res. 2006, 47, 964.
27. Aguilar-Hernández, M.; Méndez, J. D. Biomed. Pharmacother. 2007, 61, 693.
28. Nakagawa, K.; Ibusuki, D.; Yamashita, S.; Miyazawa, T. Ann. N.Y. Acad. Sci. 2008,
1126, 288.
29. Ravandi, A.; Kuksis, A.; Marai, L.; Myher, J. Lipids 1995, 30, 885.
30. Adrover, M.; Vilanova, B.; Frau, J.; Muñoz, F.; Donoso, J. Amino Acids 2009, 36,
437.
31. Dawson, R. M. C.; Elliott, D. C.; Elliott, W. H.; Jones, K. M. Data for Biochemical
Research; Clarendon Press: Oxford, 1986.
parameters a and b are the initial concentrations of PLP and amino
compound, respectively; and x and xe are the Schiff base concentra-
tions at time t and equilibrium, respectively. Taking into account
that in our experimental conditions ab ꢃ xxe, kap was calculated
from the slope of a plot of ln (A1 ꢀ At) versus time. Changing the
aminophosphate concentration over a wide range while keeping
the PLP concentration constant provided kap values that were fitted
to
2
2
2
kap ¼ ½kh þ kf ða þ bÞꢂ ꢀ 4abkf
ð10Þ
were kf and kh are the kinetic formation and hydrolysis constant,
respectively, for the Schiff base. Finally, the equilibrium constant,
Keq, was calculated as the ratio kf/kh.
32. CRC Handbook of Chemistry and Physics, 89th Edition (Internet Versión 2009);
Lide, D. R., Ed.; CRC Press/Taylor and Francis: Boca Raton, FL, 2009.
33. Suji, G.; Sivakami, S. Biogerontology 2004, 5, 365.
34. Dyer, D. G.; Blackledge, J. A.; Thorpe, S. R.; Baynes, J. W. J. Biol. Chem. 1991, 266,
11654.
35. Morgan, P. E.; Dean, R. T.; Davies, M. J. Arch. Biochem. Biophys. 2002, 403, 259.
36. Kuksis, A.; Ravandi, A.; Scheneider, M. Ann. N.Y. Acad. Sci. 2005, 1043, 417.
37. Borch, R. F.; Bernstein, M. D.; Durst, H. D. J. Am. Chem. Soc. 1971, 93, 2897.
38. Bunn, H. F.; Higgins, P. J. Science 1981, 213, 222.
39. Verardo, G.; Gorassini, F.; Giumanini, A. G.; Scubla, T.; Tolazzi, M.; Strazzolini,
P. Tetrahedron 1995, 51, 8311.
40. Pollack, R.; Ritterstein, S. J. Am. Chem. Soc. 1972, 94, 5064.
41. Metzler, C. M.; Cahill, A.; Metzler, D. E. J. Am. Chem. Soc. 1980, 102, 6075.
42. Adrover, M.; Vilanova, B.; Donoso, J.; Muñoz, F. Int. J. Chem. Kinet. 2007, 39, 154.
43. Vázquez, M. A.; Muñoz, F.; Donoso, J. Int. J. Chem. Kinet. 1990, 22, 905.
44. Vázquez, M. A.; Muñoz, F.; Donoso, J.; García Blanco, F. J. Chem. Soc., Perkin
Trans. II 1991, 2, 275.
Acknowledgments
This work was made possible by a grant from the Spanish Gov-
ernment (Project CTQ-2008-02207/BQU). One of us (C.C.) wishes to
acknowledge the award of a fellowship by the Regional Govern-
ment of the Balearic Islands.
References and notes
1. Bailey, A. J.; Paul, R. G.; Knott, L. Mech. Ageing Dev. 1998, 106, 1.
2. Prakash, V.; Obrenovich, M. E.; Atwood, C. S.; Perry, G.; Smith, M. Neurotoxin.
Res. 2002, 4, 191.
3. Kume, S.; Takeda, M.; Mori, T.; Araki, N.; Suzuki, H.; Horiuchi, S.; Kodama, T.;
Miyauchi, Y.; Tkahashi, K. Am. J. Pathol. 1995, 147, 645.
4. Baynes, J. W.; Thorpe, S. R. Diabetes 1999, 48, 1.
5. Ulrich, P.; Cerami, A. Recent Prog. Horm. Res. 2001, 56, 1.
6. Ravandi, A.; Kuksis, A.; Shaikh, N. A. Arterioscler. Thromb. Vasc. Biol. 2000, 20,
467.
45. Dworkin, J. P.; Miller, S. L. Carbohydr. Res. 2000, 329, 359.
46. Tsui, F.; Ojcius, D.; Hubbell, W. Biophys. J. 1986, 49, 459.
47. Moncelli, M. R.; Becucci, L.; Guidelli, R. Biophys. J. 1994, 66, 1969.
48. Crugeiras, J.; Rios, A.; Riveiros, E.; Richard, J. P. J. Am. Chem. Soc. 2009, 131,
15815.
49. Vázquez, M. A.; Muñoz, F.; Donoso, J.; García Blanco, F. J. Mol. Catal. 1991, 68,
105.
50. Echevarría, G. R.; Santos, J. G.; Basagoitia, A.; García Blanco, F. J. Phys. Org. Chem.
2005, 18, 546.
7. Oak, J.; Nakagawa, K.; Miyazawa, T. FEBS Lett. 2000, 481, 26.
8. Stadtman, E. R.; Levine, R. L. Amino Acids 2003, 25, 207.
9. Nakabeppu, Y.; Sakumi, K.; Sakamoto, K.; Tsuchimoto, D.; Tsuzuki, T.; Nakatsu,
Y. Biol. Chem. 2006, 387, 373.
51. García del Vado, M. A.; Donoso, J.; Muñoz, F.; Echevarría, G. R.; García Blanco, F.
J. Chem. Soc., Perkin Trans. II 1987, 2, 445.
52. Barannikov, V. P.; Badelin, V. G.; Venediktov, E. A.; Mezhevoi, I. N.; Guseinov, S.
S. Russ. J. Phys. Chem. A 2011, 1, 16.
53. Vázquez, M. A.; Muñoz, F.; Donoso, J.; García Blanco, F. Int. J. Chem. Kinet. 1992,
24, 67.
54. Voziyan, P. A.; Mettz, T. O.; Baynes, J. W.; Hudson, B. G. J. Biol. Chem. 2002, 277,
3397.
55. Onorato, J. M.; Jenkins, A. J.; Thorpe, S. R.; Baynes, J. W. J. Biol. Chem. 2000, 275,
21177.
56. Arginov, O. K.; Kerina, I. I.; Uzunova, J. I.; Argirova, M. D. From Special
Publication – Royal Society of Chemistry (1998), 223 (The Maillard Reaction in
Foods and Medicine), 245.
ˇ
10. Obšil, T.; Amler, E.; Obšilová, V.; Pavlícek, Z. Biophys. Chem. 1999, 80, 165.
11. Bucala, R.; Makita, A.; Koschinky, T.; Cerami, A.; Vlassara, H. Proc. Natl. Acad. Sci.
U.S.A. 1993, 90, 6434.
12. Ravandi, A.; Kuksis, A.; Marai, L.; Myher, J. J.; Steiner, G.; Lewisa, G.; Kamido, H.
FEBS Lett. 1996, 381, 77.
13. Pamplona, R.; Bellmunt, M. J.; Portero, M.; Riba, D.; Prat, J. Life Sci. 1995, 57,
873.
14. Lederer, M. O.; Dreisbusch, C. M.; Bundschuh, R. M. Carbohydr. Res. 1997, 301,
111.
15. Lederer, M. O.; Baumann, M. Bioorg. Med. Chem. 2000, 8, 115.
16. Lertsiri, S.; Shiraishi, M.; Miyazawa, T. Biosci. Biotechnol. Biochem. 1998, 62, 893.
17. Fountain, W. C.; Requena, J. R.; Jenkins, A. J.; Lyons, T. J.; Smyth, B.; Baynes, J.
W.; Thorpe, S. R. Anal. Biochem. 1999, 272, 48.
57. Peterson, E. A.; Sober, H. A. J. Am. Chem. Soc. 1954, 76, 169.
58. SigmaPlot Version 10.0. SPSS, Chicago, Illinois, USA, 2005.
59. Glushonok, G. K.; Glushonok, T. G.; Maslovskaya, L. A.; Shadyro, O. I. Russ. J. Gen.
Chem. 2003, 73, 1027.