Synthesis of polyhydroxylated aromatic mandelic acid amides and their antioxidative potential

Article abstract of DOI:10.1016/S0040-4020(00)01136-4

Six hydroxymandelic acid amides of phenolic amines were synthesized by condensation of 3,4-dihydroxymandelic or 4-hydroxy-3-methoxymandelic acid N-succinimidyl esters and several phenolic benzylamino and phenethylamino hydrochlorides in moderate to good y

Full text of DOI:10.1016/S0040-4020(00)01136-4

TETRAHEDRON
Pergamon
Tetrahedron 57 (2001) 1277±1282
Synthesis of polyhydroxylated aromatic mandelic acid amides
and their antioxidative potential
Jakob P. Leyp and Heinz-JuÈrgen Bertram
Haarmann & Reimer GmbH, Flavor Division, Research and Development Flavors, Synthesis New Flavor Ingredients,
P.O. Box 1253, D-37601 Holzminden, Germany
Received 15 November 2000; accepted 4 December 2000
AbstractÐSix hydroxymandelic acid amides of phenolic amines were synthesized by condensation of 3,4-dihydroxymandelic or
4-hydroxy-3-methoxymandelic acid N-succinimidyl esters and several phenolic benzylamino and phenethylamino hydrochlorides in
moderate to good yield. The radical scavenging activities determined by 2,2-diphenyl-1-picrylhydrazyl assay and superoxide trapping
assay were superior compared to standards l-ascorbic acid, a-tocopherol, and butylated hydroxytoluene (BHT). The antioxidative activities
tested by accelerated autoxidation of bulk lipids were 2±3.5 times more potent compared to standards a-tocopherol and BHT. 3,4-Dihy-
droxymandelic acid dopamide (4b) and 4-hydroxy-3-methoxymandelic acid dopamide (4a) showed the best overall performance as anti-
oxidants. q 2001 Elsevier Science Ltd. All rights reserved.
1. Introduction
(SO) radical anions from triplet oxygen.7 Additionally, SO
radical anions are formed as intermediates in the respiratory
chain. The SO can be accidentally liberated from mito-
chondria.
Antioxidants play a crucial role in organisms or tissues like
human skin as well as in non-living systems like foodstuffs.
In man, photoinduced oxidative stress is suspected as the
main cause of skin ageing.1±3 In addition, it is assumed that
many severe diseases like arteriosclerosis, cataract for-
mation and cancer are caused or accelerated by oxidative
stress, which can be generated by natural and arti®cial
sources.4±6
Under physiological conditions, the SO is converted mainly
by SO dismutase into hydrogen peroxide, which is degraded
by catalase or by glutathione peroxidase. Alternatively,
hydrogen peroxide can be converted to the highly aggres-
sive hydroxyl radical by heavy metal ions, e.g. Fe21 (Fenton
reaction). Hydroxyl radicals are especially responsible for
oxidative damage of the lipids in cell membranes or inter-
cellular lipids. The hydroperoxides of the lipids are able to
propagate the radical chain mechanism of lipid autoxi-
dation.
A complex antioxidant defense system has evolved in
organisms, which protects against reactive oxygen species
(ROS). However, a high level of ROS can tilt the balance
toward a prooxidant state. The resulting oxidative stress is
responsible for the damage of cellular molecules, especially
DNA, proteins and unsaturated lipids, e.g. human sebum
lipid squalene is oxidized to its corresponding hydro-
peroxide. This situation leads to the degeneration of cellular
and interstitial structures and, therefore, to apoptosis or
necrosis of cells or tissue. Finally, the organs develop patho-
logical conditions such as the sagging or wrinkling of skin,
lesions of blood vessels, clouding of eye lens and even
cancer.
The most important molecular antioxidants in mammals are
a-tocopherol, l-ascorbic acid, ubiquinone, lipoic acid and
glutathione, which are able to scavenge the ROS directly.
The highly reactive hydroxyl radical is also trapped by many
other organic cell compounds, e.g. glucose and proteins.
If cells are exposed to strong oxidative stress, the high
amount of ROS depletes the endogenous molecular radicals
that accelerate the development of cell damage.8 It is
assumed that the administration of antioxidants or radical
scavengers to tissues like human skin can shift the anti-
oxidative balance back towards protection against oxidative
stress. In fact, recent clinical studies showed that topically
applied antioxidants were able to prevent skin lipid peroxi-
dation and photoageing.9,10 For arteriosclerosis, some facts
suggest that dietary antioxidants like tocopherols and ¯avo-
noids prevent or retard the development of typical sclerotic
lesions of blood vessels.6
In cells, ROS are formed by different pathways. For
example, mediated by photosensitizers (e.g. porphyrins),
UVA radiation generates singlet oxygen or superoxide
Keywords: antioxidants; radical scavengers; hydroxymandelic acid amides;
polyphenols.
Corresponding author. Fax: 149-5531-903883;
e-mail: jakob.ley.jl@hr-gmbh.de
p
0040±4020/01/$ - see front matter q 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4020(00)01136-4

1278
J. P. Ley, H.-J. Bertram / Tetrahedron 57 (2001) 1277±1282
Scheme 1. Synthesis of hydroxymandelic acid amides 4a±4f. (i) 1,4-Dioxan, NHOSu (1 equiv.), DCC (1 equiv.), rt, 16 h; (ii) H2O, NaHCO3 (1 equiv.), 1 h,
508C, then HCl (5% w/w).
Antioxidants are added to food mainly to inhibit autoxi-
dation of lipids or proteins. In particular, the autoxidation
of lipids generates typical off-¯avors in many foodstuffs.
The long chain polyunsaturated fatty acids are degraded to
aliphatic C-9 or C-6 aldehydes like (E)-2-nonenal, (2E,4E)-
2,4-nonadienal and different hexenals like n-hexanal, which
cause the characteristic rancid ¯avor of oxidized fats and
oils.11
®ltrates were treated immediately with the appropriate
phenolic amino hydrochlorides 3a±3d in aqueous solution.
3,4,5-Trihydroxybenzylamino hydrochloride (3d), which is
not available commercially, was obtained from 3,4,5-tri-
hydroxybenzaldehyde via the benzaldoxime and subsequent
Pd±C-catalysed reduction by hydrogen in hydrochloric acid
containing ethanol.12
In our efforts to develop new antioxidants for cosmetic or
food use we have synthesized and investigated new phenolic
acid amides of different phenolic amines as potent anti-
oxidants.12,13 As most of these amides showed a limited
water solubility (,0.1%), we decided to prepare the
hydroxymandelic acid amides of phenolic amines as more
hydrophilic derivatives.
The isolated yield of hydroxymandelic acid amides 4a±4f
varied widely between 25 and 77%. In particular, the yield
of the amide of 3,4,5-trihydroxybenzylamine (4d) was only
moderate. In order to remove all the remaining DCU, the
crude products have to be chromatographed in two steps:
the ®rst eluent used was ethyl acetate; in a second step, the
solvent system chloroform/methanol was used for further
puri®cation. Especially in solution, the products are sensi-
tive toward oxidation. Therefore, during reaction and
isolation, precautions against oxygen have to be taken,
such as degassing of solvents and inert gas.
2. Results and discussion
2.1. Synthesis
For the chromatographed products, the purity detected by
HPLC was higher than 95% (UV detection by a diode array
detector). The amount of remaining DCU was determined
In the literature, 3,4-dihydroxymandelic or 4-hydroxy-3-
methoxymandelic acid amides of phenolic amines have
not been described so far. In our earlier studies regarding
phenolic acid amides, we used protected active esters for
synthesis.12 The classical coupling of acid chlorides with
amines in basic mixtures was not applicable because self-
condensation of the phenolic acids occurred. The phenolic
acids, e.g. caffeic acid, were protected at the phenolic
hydroxyl moieties by methoxycarbonyl groups and
esteri®ed with N-hydroxysuccinimide (NHOSu) by N,N0-
dicyclohexylcarbodiimide (DCC) methodology. The active
esters were isolated, puri®ed, and subsequently condensed
with phenolic amines at pH 8±9.
1
by H NMR spectroscopy because the urea showed no
characteristic absorption in the UV detection system. The
chromatographic puri®cation steps yielded products with a
DCU amount lower than the detection limit.
The hydroxymandelic acid amides 4a±4f were identi®ed by
NMR, LC±MS and HRMS. For recording the H NMR
1
spectra in CD3OD, a suppressing sequence in the range
4.9±5.0 ppm for the CD3OH signal was used. Because the
a-H signal of the mandelic acid moiety occurs in the same
region (4.95 ppm), it was at least partially suppressed by the
sequence and the integral was much too low. All other
signals and their integrals in 1H NMR spectra corresponded
to the proposed structures. The characteristic 13C NMR
signal of mandelic acid a-C could be detected for each
product from 74 to 75 ppm. The identi®cation of HPLC
peaks was performed by MS coupling measuring in atmos-
pheric pressure chemical ionization (APCI) negative mode.
The [M2H]2 ions were found for all compounds. With the
For the synthesis of the new hydroxymandelic acid amides
4a±4f, we have used a variation of this method (cf. Scheme
1). The active succinimidyl esters 2a and 2b were prepared
directly from the non-protected hydroxymandelic acids 1a
and 1b by the NHOSu/DCC method in 1,4-dioxan as
solvent. The reaction mixtures were only ®ltered to remove
the precipitated N,N0-dicyclohexylurea (DCU) and the

J. P. Ley, H.-J. Bertram / Tetrahedron 57 (2001) 1277±1282
1279
Table 1. Results of evaluation of radical scavenging and antioxidative potential for hydroxymandelic acid amides 4a±4f. Conditions for Rancimat assay: A,
soybean oil, 1008C, 0.05%; B, evening primrose oil, 1008C, 0.05%; C, squalene, 808C, 0.005%
DPPH assay
EC50/mmol/mmol
SO assay, IC50/nM
Rancimat assay AOI
B
Solubility (%),
5% DMSO in H2O
s.s.a/h
A
C
4a
4b
4c
4d
4e
4f
0.13^0.01
0.10^0.01
0.20^0.01
0.20^0.01
0.20^0.01
0.07^0.01
0.27^0.02
0.25^0.02
0.24^0.01
24
5
7
7
7
150
80
110
800
140
70
700
±
±
13^1
17^1
9.8^1.3
11^1
12^1
17^2
±
7.7^0.5
6.4^0.8
8.0^0.4
9.3^1.4
4.4^0.6
7.4^1.0
±
28^1
7.1^0.8
15^3
7.5^1.4
13^1
6.4^0.6
±
5±10
.10
.10
0.5±1
0.05±0.1
0.1±0.5
.10
7
ascorbic acid
a-tocopherol
BHT
0.2
0.3
7
5.1^0.1
4.6^0.7
4.6^1.3
±
39^4
19^1
,0.05
,0.05
a
s.s.: steady state.
exception of 4c and 4e, the [2M2H]2 ions could be
detected, too. For identi®cation, satisfactory HRMS spectra
could be recorded. The difference between the calculated
and observed molecular weights was below 1.3£1023 mass
units.
50% of SO radicals had been trapped. l-Ascorbic acid
was used as standard for both assays.
For the determination of the antioxidative potential for lipid
systems, the accelerated autoxidation of bulk lipids by air
was used.12,16 The oxidation was carried out with or without
test substance in the Rancimat apparatus with stripped
soybean oil and stripped evening primrose oil at 1008C
and squalene at 808C. a-Tocopherol and butylated hydroxy-
toluene (BHT) were used as standards. Squalene was chosen
as an example of a typical skin lipid component, soybean oil
as an important nutritional fat, and evening primrose oil as a
valuable cosmetic lipid, which is very sensitive towards
oxidation. Commercial squalene was used after puri®cation
over alumina type N and stabilization with 1 ppm a-toco-
pherol. Oxidation of the unsaturated lipids proceeds only
very slowly for the induction period (IP) and then suddenly
increases when autoxidation accelerates. The IP of the
stripped soybean oil at 1008C was between 2 and 3 h.
Squalene stabilized with 1 ppm a-tocopherol showed IPs
between 0.5 and 1 h at 808C. The antioxidative indices
2.2. Antioxidative activity
The free radical scavenging action of the amides 4a±4f was
determined by the 2,2-diphenyl-1-picrylhydrazyl (DPPH)
assay14 and an SO radical trapping assay.15 For performing
the DPPH assay, a solution of the purple coloured DPPH
radical was mixed with the test compound and the decrease
of the absorption was determined photometrically until a
steady state was reached. The results were calculated as
the effective concentration EC50 of the test compound in
relation to starting DPPH concentration, at which 50% of
the DPPH radicals remained. The SO radical was generated
in a hydrogen peroxide/horseradish peroxidase system. The
SO radical was detected by recording chemiluminescence
ampli®ed by luminol. The SO scavenging activity of lipo-
philic test compounds such as a-tocopherol could not be
measured correctly in the SO assay because of their poor
solubility in the required aqueous solvent. Results were
expressed as the inhibitor concentration IC50, at which
(AOIs) were calculated by dividing the IPtest
IPcontrol.
by
In soybean and evening primrose oil, we have
compound
12
used test concentrations of 0.05% (500 mg g21) and
0.005% squalene (50 mg g21).
Figure 1. Correlation between radical scavenging activity determined by DPPH assay and antioxidative potential determined in soybean oil for amides 4a±4f.
ECp50 is the mass related EC50 calculated by ECpˆEC£MW (mg mmol21).

1280
J. P. Ley, H.-J. Bertram / Tetrahedron 57 (2001) 1277±1282
Solubilities of the test compounds were determined in water
containing 5% DMSO. This aqueous system was chosen
because it would be tolerated by cell cultures, which will
be performed in the future. The results of all tests are listed
in Table 1.
investigated amides are equal or more potent antioxidants
for soybean and evening primrose oil with respect to toco-
pherol. They are able to protect squalene against oxidation
fairly well, but are inferior to classic antioxidants like
a-tocopherol or BHT.
The amides 4a±4f are superior radical scavengers in respect
to l-ascorbic acid, a-tocopherol and BHT. One equivalent
of 4c±4e is able to scavenge 5 equiv. of DPPH radicals, and
for 4a, 4b, and 4f the ratio is still 1:10. In the SO assay, only
the 4-hydroxy-3-methoxymandelic acid amide 4d shows a
lower radical trapping activity. All other amides are 4±10
times more potent than l-ascorbic acid. The AOIs of
hydroxymandelic acid amides in soybean oil are in the
range of 10±17 and are 2±3.5 times higher compared to
a-tocopherol and BHT, respectively. In evening primrose
oil, the AOI of 4e is comparable to tocopherol, but the other
amides are superior to the standards. In squalene, only 4a
has an antioxidative potential of the same order compared to
a-tocopherol. 4c and 4e show good action similar to BHT,
but the remaining amides 4b, 4d, and 4f are only weak to
moderate antioxidants in squalene.
4. Experimental
4.1. General
Squalene was obtained from E. Merck (Darmstadt,
Germany) and evening primrose oil and soybean oil from
Henry Lamotte (Bremen, Germany). All other chemicals
were obtained from Sigma±Aldrich (Deisenhofen,
Germany) or Lancaster Synthesis (MuÈlheim, Germany).
Flash chromatography was performed using Biotage Flash
40 equipment (LCTech, Dorfen, Germany) with disposable
pre-packed columns. NMR spectra were recorded using
Varian VXR400S (1H: 400 MHz) or Gemini 2000 (1H:
200 MHz) spectrometers (Varian, Darmstadt, Germany) at
258C using tetramethylsilane as internal standard. LC±MS
spectra were recorded using the LCQ HPLC system
Finnigan MAT HP1100 (Finnigan MAT, Egelsbach,
Germany; APCI). High resolution mass spectra were
recorded using a Finnigan MAT 8200 by ¯ash evaporation
of samples after dissolving in methanol (resolution .8000).
The oxidative stability of bulk lipids was measured
with a Rancimat apparatus (Deutsche Metrohm GmbH,
Filderstadt, Germany). Chemiluminescence was determined
In Fig. 1, a plot of mass related ECp (calculated by
50
ECpˆEC£MW) against AOI of soybean oil is shown.
There is a fairly good linear correlation with a Pearson
coef®cient of 0.86. In contrast, we could not detect any
correlations between ECp50 and AOI for squalene or evening
primrose oil, and between EC50 of DPPH assay and IC50 of
SO assay. The used test systems are not redundant
and therefore they show different aspects of antioxidative
activity.
È
using a Berthold LB96P (Perkin±Elmer GmbH, Uberlingen,
Germany).
In addition, we found some structure±activity relations for
hydroxymandelic acid amides 4a±4f. In soybean oil and
evening primrose oil, the methoxy groups of 4a, 4c, 4d,
and 4e show a strong suppressing effect on the AOI.
Additionally, the number of catechol moieties has a nega-
tive in¯uence on AOI in squalene and evening primrose oil.
In contrast, the catechol number correlates positively with
AOI in soybean oil and the effectiveness as DPPH radical
scavengers. A trihydroxy group in the molecule decreases
the SO trapping ability as well as the potential to protect
squalene.
4.1.1. N-(3,4-Dihydroxyphenethyl)-2-hydroxy-2-(4-hy-
droxy-3-methoxyphenyl)acetamide (4a). 2-(4-Hydroxy-
3-methoxyphenyl)-2-hydroxyacetic acid (1b) (400 mg,
2.02 mmol) and NHOSu (232 mg, 2.02 mmol) were
dissolved in 1,4-dioxan (20 ml) under nitrogen. N,N0-Di-
cyclohexylcarbodiimide (416 mg, 2.02 mmol) was added
to the mixture at room temperature, and the mixture was
stirred at that temperature for 48 h. The by-product that
precipitated out was ®ltered off, and the ®ltrate was added
to a solution of 2-(3,4-dihydroxyphenyl)-ethylamino hydro-
chloride (3a) (383 mg, 2.02 mmol) in a mixture of water and
1,4-dioxan (2:1, 20 ml) under nitrogen, followed by sodium
hydrogen carbonate (187 mg, 2.22 mmol). The mixture was
stirred at 508C for 4 h, then rendered acidic using hydro-
chloric acid (cˆ10%), and the reaction solution was
extracted 3 times with ethyl acetate (a total of 90 ml). The
organic phase was washed with saturated NaCl solution,
dried over Na2SO4, ®ltered, and the ®ltrate was evaporated
under reduced pressure. Puri®cation was carried out by ¯ash
chromatography on silica gel using the eluent ethyl acetate.
The yield was 170 mg of an amorphous, colourless solid
The solubilities of compounds 4a±4f in water/DMSO were
much better compared to the phenolic acid amides synthe-
sized earlier. Most of the hydroxymandelic acid amides
show solubilities greater than 1% in the mixture. As
expected, the a-hydroxy group of the mandelic acid moiety
has a positive effect upon the solubility. When the results of
the assays were scored, 3,4-dihydroxymandelic acid dopa-
mide (4b) shows the best performance as antioxidant and
radical scavenger, followed by the 4-hydroxy-3-methoxy-
mandelic acid dopamide (4a).
1
(25% of theory based on the acid used). HPLC .99%; H
NMR (400 MHz, CD3OD with CD3OH suppression):
dˆ6.94 (1H, d, 1.9 Hz), 6.79 (1H, ddd, 8, 2, 0.5 Hz), 6.74
(1H, dd, 8.1 Hz), 6.654 (1H, d, 7.9 Hz), 6.647 (1H, d, 2 Hz),
6.48 (1H, dd, 8.1, 1.9 Hz), 4.87 (partially suppressed, s),
3.82 (3H, s), 3.46±3.36 (2H, m), 2.68±2.63 (2H, m) ppm;
13C NMR (100 MHz, CD3OD): dˆ175.7 (C), 149.0 (C),
133.1 (C), 131.8 (C), 121.0 (2£CH), 116.8 (CH), 116.4
(CH), 116.1 (CH), 111.6 (CH), 75.5 (CH), 56.4 (CH3),
3. Conclusions
The new hydroxymandelic acid amides of phenolic amines
4a±4f are ef®cient radical scavengers and antioxidants and
more effective as radical scavengers when compared with
the standards a-tocopherol, BHT and l-ascorbic acid. The

J. P. Ley, H.-J. Bertram / Tetrahedron 57 (2001) 1277±1282
1281
41.8 (CH2), 35.9 (CH2) ppm; MS (APCI-): m/eˆ332.1
(100%, [M2H]2), 664.6 (17%, [2M2H]2); HRMS:
MWfoundˆ333.1216; MWcalcd for C17H19NO6ˆ333.1213.
(290 mg, 1.52 mmol) as a colourless amorphous solid. The
product was puri®ed by ¯ash chromatography on silica gel
two times using subsequently the eluent ethyl acetate and a
gradient chloroform/methanol 10:1 to 2:1. Yield: 200 mg
1
4.1.2.
N-(3,4-Dihydroxyphenethyl)-2-(3,4-dihydroxy-
(39% of theory based on the acid used). HPLC .98%; H
phenyl)-2-hydroxyacetamide (4b). 2-(3,4-Dihydroxy-
phenyl)-2-hydroxyacetic acid (1b) (300 mg, 1.63 mmol)
and NHOSu (188 mg, 1.63 mmol) were dissolved in 1,4-
dioxan (20 ml) under nitrogen, and N,N0-dicyclohexyl-
carbodiimide (336 mg, 1.63 mmol) was added to the
mixture at room temperature, which was stirred at that
temperature for 16 h. The by-product which precipitated
out was ®ltered off, and the ®ltrate was added to a solution
of 2-(3,4-dihydroxyphenyl)ethylamino hydrochloride (3a)
(309 mg, 1.63 mmol) in water (20 ml). Sodium hydrogen
carbonate (151 mg, 1.8 mmol) was added, and the reaction
mixture was stirred under nitrogen at 508C for 1.5 h. The
mixture was rendered acidic using hydrochloric acid
(cˆ5%) and extracted 3 times with ethyl acetate (a total
of 90 ml). The organic phase was washed with saturated
NaCl solution, dried over Na2SO4, ®ltered, and the ®ltrate
was evaporated under reduced pressure. The product was
puri®ed by ¯ash chromatography on silica gel using the
eluent ethyl acetate. The yield was 266 mg of a colourless
solid (51% of theory based on the acid used). HPLC .96%;
1H NMR (400 MHz, CD3OD with CD3OH suppression):
dˆ6.83 (1H, d, 2 Hz), 6.72 (1H, d, 8 Hz), 6.674 (1H, ddd,
8, 2, 0.5 Hz), 6.672 (1H, d, 8 Hz), 6.65 (1H, d, 2 Hz), 6.50
(1H, dd, 8, 2.1 Hz), 3.43±3.35 (2H, m), 2.65 (t, 7.6 Hz)
ppm; 13C NMR (100 MHz, CD3OD): dˆ175.8 (C), 146.5
(C), 146.3 (C), 146.26 (C), 144.9 (C), 133.3 (C), 131.9 (C),
121.2 (CH), 119.9 (CH), 116.9 (CH), 116.5 (CH), 116.2
(CH), 115.4 (CH), 75.4 (CH), 41.9 (CH2), 36.0 (CH2)
ppm; MS (APCI-): m/eˆ318.8 (100%, [M2H]2), 636.4
NMR (400 MHz, CD3OD with CD3OH suppression):
dˆ6.93 (1H, d, 2.1 Hz), 6.86 (1H, ddd, 8.1 Hz, 2.1 Hz,
0.5 Hz), 6.75 (1H, d, 8.1 Hz), 6.31 (2H, t, 0.6 Hz), 4.93
(partially suppressed, s), 4.27 (1H, d, 14.3 Hz), 4.14 (1H,
d, 14.3 Hz), 3.79 (3H, s) ppm; 13C NMR (100 MHz,
CD3OD): dˆ175.5 (C), 149.0 (C), 147.7 (C), 147.1
(2£C), 133.3 (C), 130.9 (C), 121.3 (CH), 116.0 (CH),
111.4 (CH), 107.9 (2£CH), 75.5(CH), 56.4 (CH3), 43.7
(CH2) ppm; MS (APCI-): m/eˆ334.0 (100%, [M2H]2),
668.7 (5%, [2M2H]2); HRMS: MWfoundˆ335.1010;
MWcalcdˆ335.1005 for C16H17NO7.
4.1.5. 2-(3,4-Dihydroxyphenyl)-2-hydroxy-N-(4-hydroxy-
3-methoxybenzyl)acetamide (4e). This compound was
obtained analogous to 4b from 2-(3,4-dihydroxyphenyl)-2-
hydroxyacetic acid (1b) (300 mg, 1.63 mmol) and 4-hydroxy-
3-methoxybenzylamino hydrochloride (3b) (309 mg,
1.62 mmol) as a colourless amorphous solid. The product
was puri®ed by ¯ash chromatography on silica gel using the
eluent ethyl acetate. Yield: 296 mg (57% of theory based on
the acid used). HPLC .98%; 1H NMR (400 MHz, CD3OD
with CD3OH suppression): dˆ6.90 (1H, dt, 2.1, 0.4 Hz),
6.78 (1H, ddd, 8.2, 2.1, 0.6 Hz), 6.74 (1H, d, 1.6 Hz), 6.72
(1H, d, 8.2 Hz), 6.71 (1H, dd, 8.0, 0.4 Hz), 6.68 (1H, dd, 8.0,
1.8 Hz), 4.90 (partially suppressed, s), 4.36 (1H, d, 14.7 Hz),
4.28 (1H, d, 14.7 Hz), 3.73 (3H, d, 0.4 Hz) ppm; 13C NMR
(100 MHz, CD3OD): dˆ175.8 (C), 149.1 (C), 146.7 (C),
146.5 (C), 146.4 (C), 133.4 (C), 131.5 (C), 121.1 (CH),
119.7 (CH), 116.1 (CH), 116.0 (CH), 115.1 (CH), 112.0
(CH), 75.4 (CH), 56.3 (CH3), 43.4 (CH2) ppm; MS (CI-):
m/eˆ318.0 (100%, [M2H]2), 301.7 (22%), 300.5
(34%); HRMS: MWfoundˆ319.1054; MWcalcdˆ319.1056
for C16H17NO6.
(23%, [2M2H]2); HRMS: MWfoundˆ319.1043; MWcalcd
ˆ
319.1056 for C16H17NO6.
4.1.3. N-(3,4-Dihydroxybenzyl)-2-hydroxy-2-(4-hydroxy-
3-methoxyphenyl)acetamide (4c). This compound was
obtained analogous to 4a from 2-(4-hydroxy-3-methoxy-
phenyl)-2-hydroxyacetic acid (1a) (500 mg, 2.53 mmol)
and 3,4-dihydroxybenzylamino hydrobromide (3c)
(667 mg, 3.03 mmol) as a colourless amorphous solid. The
product was puri®ed by ¯ash chromatography on silica gel
using the eluent ethyl acetate. Yield: 618 mg (77% of theory
based on the acid used). HPLC .98%; 1H NMR (400 MHz,
CD3OD with CD3OH suppression): dˆ6.94 (1H, d, 2.1 Hz),
6.87 (1H, ddd, 8.1, 2.0, 0.5 Hz), 6.75 (1H, d, 8.1 Hz), 6.73
(1H, d, 2.0 Hz), 6.68 (1H, d, 8.0 Hz), 6.60 (1H, ddd, 8.0, 2.1,
0.6 Hz), 4.94 (partially suppressed, s), 4.32 (1H, d, 14.5 Hz),
4.21 (1H, d, 14.5 Hz), 3.78 (3H, s) ppm; 13C NMR
(100 MHz, CD3OD): (dˆ175.5 (C), 149.0 (C), 147.7 (C),
146.4 (C), 145.7 (C), 133.3 (C), 131.6 (C), 121.3 (CH),
120.3 (CH), 116.3 (CH), 116.1 (CH), 115.9 (CH), 111.4
(CH), 75.5 (CH), 56.4 (CH3), 43.5 (CH2) ppm; MS
4.1.6. N-(3,4-Dihydroxybenzyl)-2-(3,4-dihydroxyphenyl)-
2-hydroxyacetamide (4f). This compound was obtained
analogous to 4b from 2-(3,4-dihydroxyphenyl)-2-hydroxy-
acetic acid (1b) (1.00 g, 5.4 mmol) and 3,4-
dihydroxybenzylamino hydrochloride (3c) (1.15 g,
6.6 mmol) as a colourless amorphous solid. The product
was puri®ed by ¯ash chromatography on silica gel using
the eluent ethyl acetate. Yield: 640 mg (39% of theory
based on the acid used). HPLC .99%; 1H NMR
(400 MHz, CD3OD with CD3OH suppression): dˆ6.87
(1H, dt, 1.9, 0.5 Hz), 6.75 (1H, ddd, 8.1, 1.9, 0.5 Hz), 6.72
(1H, dd, 8.1, 0.5 Hz), 6.72 (1H, dm, 2.1, 0.3 Hz), 6.69 (1H,
d, 8.0 Hz), 6.59 (1H, ddt, 8.0, 2.1, 0.6 Hz), 4.88 (partially
suppressed, s), 4.25 (2H, s) ppm; 13C NMR (100 MHz,
CD3OD): dˆ175.6 (C), 146.6 (C), 146.5 (C), 146.4 (C),
145.8 (C), 133.2 (C), 131.3 (C), 120.2 (CH), 119.8 (CH),
116.3 (CH), 116.1 (CH), 116.0 (CH), 115.3 (CH), 75.4
(CH), 43.6 (CH2) ppm; MS (APCI-): m/eˆ304.1 (10%,
(APCI-): m/eˆ318.1 (100%, [M2H]2); HRMS: MWfound
ˆ
319.1063; MWcalcdˆ319.1056 for C16H17NO6.
[M2H]2),
608.7
(100%,
[2M2H]2);
HRMS:
MWfoundˆ305.0894; MWcalcdˆ 305.0899 for C15H15NO6.
4.1.4. 2-Hydroxy-2-(4-hydroxy-3-methoxyphenyl)-N-(3,4,5-
trihydroxybenzyl)acetamide (4d). This compound was
obtained analogous to 4a from 2-(4-hydroxy-3-methoxy-
phenyl)-2-hydroxyacetic acid (1a) (300 mg, 1.52 mmol)
and 3,4,5-trihydroxybenzylamino hydrochloride12 (3d)
4.2. DPPH assay
In disposable polystyrol cuvettes, 2.5 ml of a methanolic

1282
J. P. Ley, H.-J. Bertram / Tetrahedron 57 (2001) 1277±1282
DPPH solution (100 mmol l21) was mixed with 500 ml of
test solutions of different concentrations (methanol as
control) and the decrease of the purple coloured radical
was determined photometrically at 514 nm until the
decrease was less than 2% per hour against negative control
(steady state). The concentration dependent activity of the
test compounds as free radical scavengers was calculated
using the following equation:
measured continuously and recorded. The IPs were calcu-
lated automatically by the Rancimat. All tests were run in
duplicate and averaged before calculation. The following
equation was used to calculate the AOI:
AOI ˆ IPtest=IPcontrol
ꢀ3†
a-Tocopherol and BHT were used as standards.
activityconcꢀ%† ˆ 100 2 ꢀremaining DPPH†
ꢀ1†
4.5. Solubility
The remaining DPPH was calculated from absorptions by
Of the test compound, 5 mg was weighed into a glass vessel,
mixed with a de®ned amount of 5% DMSO in water (®nal
masses: 50 mg, 100 mg, 500 mg, 1 g, 5 g, 10 g), and treated
with ultrasound at room temperature for 5 min. This pro-
cedure was repeated until the compound had completely
dissolved. At a solubility of ,0.05% by weight, the experi-
ment was terminated.
remaining DPPH ꢀ%†
ˆ absorptionꢀtest at steady state†=absorptionꢀcontrol at t †
ꢀ2†
0
The activity (%) in a series of dilutions of one test
compound was used to calculate the effective relative
concentration EC50 (based on the starting concentration of
DPPH, ECˆctest/c
), at which 50% of DPPH has
ꢀDPPH at t0†
Acknowledgements
been removed. The assay was performed in triplicate and
the absorptions were averaged before calculation. a-Toco-
pherol, l-ascorbic acid and BHT were used as standards.
I am very grateful to Mr Strempel and Mrs Schlichter for
their excellent technical assistance and to Dr Werkhoff, Dr
Krammer, Dr Schmidt and Dr Sommer for the analysis of
the compounds. Special thanks are dedicated to Professor Dr
MuÈller-Peddinghaus and Dr GruÈtzmann (both Bayer AG)
for conducting the SO assay.
4.3. SO assay
The test compounds were diluted using a dimethyl sulfoxide
stock solution (10 mM) with phosphate buffered saline
(PBS). Each well of a 96-well plate of white polystyrene
was subsequently charged with 50 ml luminol (200 mM in
PBS), 50 ml horseradish peroxidase (HRPO, 1 U ml21 in
PBS), and 50 ml of the test solutions of different concen-
trations (e.g. 40 mM). The mixtures were incubated at 378C
for 5 min. Finally, 50 ml aliquots of H2O2 (40 mM)
were added and the chemiluminescence was recorded
immediately for 10 s. The activity was calculated as
inhibitor concentration IC50, at which 50% of SO generated
was depleted. l-Ascorbic acid was used as standard.
References
1. Halliwell, B.; Murcia, M. A.; Chirico, S.; Aruoma, O. I. Crit.
Rev. Food Sci. Nutr. 1995, 35 (1/2), 7±20.
2. Tyrrell, R. M. Biochem. Soc. Symp. 1995, 61, 47±53.
3. Scharffetter-Kochanek, K. Photoaging of the connective tissue
of skin: its prevention and therapy. In Antoxidants in Disease
Mechanisms and Therapy; Sies, H., Ed.; Academic: San
Diego, 1997; Vol. 38, pp 639±655.
4. Ames, B. N.; Shigenaga, M. K.; Hagan, T. M. Proc. Natl.
Acad. Sci. USA 1993, 90, 7915±7922.
4.4. Rancimat assay
5. Cheeseman, K. H. Toxicol. Ind. Health 1993, 9 (1/2), 39±51.
6. Gordon, M. H. Nat. Prod. Rep. 1996, 265±273.
7. Carbonare, M. D.; Pathak, M. A. J. Photochem. Photobiol. B
1992, 14, 105±124.
Prior to the assays, the lipids were puri®ed as follows:
soybean oil, evening primrose oil and squalene were
dissolved in n-heptane. The solutions were passed through
a glass column charged with activated alumina type N by
pressurized nitrogen. The eluent was collected under nitro-
gen. The solvents of the puri®ed solutions were distilled off
under reduced pressure. In the case of squalene, a solution of
a-tocopherol in n-heptane was added to give a ®nal concen-
tration of 1 ppm tocopherol prior to evaporation.
8. Podda, M.; Traber, M. G.; Weber, C.; Yan, L.-J.; Packer, L.
Free Radical Biol. Med. 1998, 24 (1), 55±65 (publ. 1997).
9. Pelle, E.; Muizzuddin, N.; Mammone, T.; Marenus, K.; Maes,
D. Photodermatol. Photoimmunol. Photomed. 1999, 15 (3/4),
115±119.
È
È
È
È
10. Stab, F.; Lanzendorfer, G.; Schonrock, U.; Wenck, H. SOFW J.
1998, 124 (10), 604 (see also pages 606, 608±610, 612, 613).
11. Fielder, S.; Rowan, D. D. J. Labelled Compd. Radiopharm.
1999, 42 (1), 83±92.
The test compounds were dissolved in methanol or acetone
(15 mg ml21 for soybean and evening primrose oil and
1.5 mg ml21 for squalene, respectively) and 100 ml of
these solutions were added to a pre-prepared 3 g lipid
sample in the reaction vessels of the Rancimat apparatus.
Pure solvent was added to control samples. A constant dry
air stream (20 l h21) was blown through the sample while
heating to 1008C for soybean oil and evening primrose oil
and 808C for squalene, respectively. The volatile oxidation
products were collected in the measuring vessels containing
water and the conductivity of the aqueous solution was
12. Ley, J. P. Int. J. Cosm. Sci. 2001, 23 (1) (in press).
13. Ley, J. P. Haarmann and Reimer GmbH EP 0,900,781, 1998;
Chem. Abstr. 1999, 130 (16), 209510.
14. Brand-Williams, W.; Cuvelier, M.-E.; Berset, C. Lebensm.-
Wiss.u.-Technol. 1995, 28, 25±30.
15. MuÈller-Peddinghaus, R.; GruÈtzmann, R. Personal communi-
cation.
16. Hadorn, H.; ZuÈrcher, K. Deutsche Lebensmittel-Rundschau
1974, 70 (2), 57±65.