Antioxidant activity of a combinatorial library of emulsifier-antioxidant bioconjugates

Article abstract of DOI:10.1021/jf8024826

A combinatorial chemistry approach was employed for the design and systematic synthesis of antioxidant-emulsifier bioconjugates to improve antioxidant activity in the interface between oil and water. A combinatorial library of 12 bioconjugates was synthesized from the phenolic antioxidants Trolox (a water-soluble α-tocopherol analogue), dihydroferulic acid, dihydrocaffeic acid, and gallic acid in combination with serine ethyl ester, serine lauryl ester, and lauroyl serine by esterification of the serine side chain or amidation, respectively. The bioconjugates were characterized by NMR and mass spectrometry, and each was tested for antioxidant activity by measuring the radical scavenging rate of 2,2-diphenyl-1-picrylhydrazyl (DPPH ^?) in methanol, the radical scavenging rate of DPPH ^? in a heterogeneous solvent system, the rate of oxygen consumption in an oil-in-water emulsion with metmyoglobin initiated oxidation, and the lag phase for diene formation in unilamellar liposomes with free radical initiation in the aqueous phase; each case was compared to the antioxidant activity of the parent antioxidant. In general, the conjugates with longer chain derivatives exhibited improved antioxidative activity in heterogeneous systems, whereas no improvement was observed in homogeneous solution. The rate of oxygen consumption in oil-in-water emulsion was found to decrease with increasing octanol/water partition coefficient, which is suggested to correspond to a saturation of the water/oil interface with antioxidant bioconjugate to approach maximal protection.

Full text of DOI:10.1021/jf8024826

9258 J. Agric. Food Chem. 2008, 56, 9258–9268
Antioxidant Activity of a Combinatorial Library of
Emulsifier-Antioxidant Bioconjugates
CHARLOTTE S. HUNNECHE,†,# MARIANNE N. LUND,§ LEIF H. SKIBSTED,§ AND
JOHN NIELSEN*,†
Bioorganic Chemistry, Department of Basic Sciences and Environment, Faculty of Life Sciences,
University of Copenhagen, Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark, and Food
Chemistry, Department of Food Science, Faculty of Life Sciences, University of Copenhagen,
Rolighedsvej 30, DK-1958 Frederiksberg C, Denmark
A combinatorial chemistry approach was employed for the design and systematic synthesis of
antioxidant-emulsifier bioconjugates to improve antioxidant activity in the interface between oil and
water. A combinatorial library of 12 bioconjugates was synthesized from the phenolic antioxidants
Trolox (a water-soluble R-tocopherol analogue), dihydroferulic acid, dihydrocaffeic acid, and gallic
acid in combination with serine ethyl ester, serine lauryl ester, and lauroyl serine by esterification of
the serine side chain or amidation, respectively. The bioconjugates were characterized by NMR and
mass spectrometry, and each was tested for antioxidant activity by measuring the radical scavenging
rate of 2,2-diphenyl-1-picrylhydrazyl (DPPH•) in methanol, the radical scavenging rate of DPPH• in a
heterogeneous solvent system, the rate of oxygen consumption in an oil-in-water emulsion with
metmyoglobin initiated oxidation, and the lag phase for diene formation in unilamellar liposomes with
free radical initiation in the aqueous phase; each case was compared to the antioxidant activity of
the parent antioxidant. In general, the conjugates with longer chain derivatives exhibited improved
antioxidative activity in heterogeneous systems, whereas no improvement was observed in
homogeneous solution. The rate of oxygen consumption in oil-in-water emulsion was found to decrease
with increasing octanol/water partition coefficient, which is suggested to correspond to a saturation
of the water/oil interface with antioxidant bioconjugate to approach maximal protection.
KEYWORDS: Antioxidant; emulsifier; combinatorial chemistry; bioconjugates
INTRODUCTION
for emulsions. Nonpolar antioxidants are predominantly more
effective in emulsions because they are retained in the oil
droplets and/or accumulate at the oil-water interface, the
location where interactions between lipid hydroperoxides and
pro-oxidants such as transition metal ions originating in the
aqueous phase occur. Likewise, the effectiveness of chain-
breaking antioxidants in retarding lipid oxidation in oil-in-water
emulsions increases as their polarity decreases or their surface
activity increases, because they become more likely to be
localized at the oil-water interface where oxidation occurs (4-6).
Incorporation of antioxidants into membranes or liposomes as
models for cellular membranes is improved when the water-soluble
antioxidant is conjugated with a longer chain ester derivative (7).
Improvement of antioxidant activity by altering the polarity of the
antioxidant (e.g., adding a lipophilic molecule to a hydrophilic
antioxidant) has previously been observed only in a few studies
(8). The antioxidant efficiency of retinyl ascorbate, which is a
conjugate of the lipophilic retinoic acid and the water-soluble
antioxidant ascorbic acid, is improved compared to that of
the parent antioxidant (9). Conversely, conjugating a lipid-soluble
antioxidant (R-tocopherol) with phosphatidylcholine increases
the antioxidative activity in lard compared to the addition of
Deterioration of food due to lipid oxidation is an inevitable
but natural process, which results in chemical and physical
changes with consequences to both human health and production
costs. Radical processes in biological systems such as food are
mediated or may be controlled by vitamin antioxidants and
various polyphenols such as anthocyanins and flavonoids
occurring widespread in the plant kingdom (1-3). Lipids in
most foods are dispersed in water as emulsions, and a major
cause of quality deterioration in emulsion is the susceptibility
of lipids to oxidation in the lipid-water interface. Emulsions
are extremely sensitive toward oxidation as the total surface
area of lipid droplets is large compared to the surface of bulk
oils (4). The effectiveness of an antioxidant in bulk oils is mainly
dependent on the air-oil interface affinity of the antioxidant,
whereas the affinity for the oil-water interface is more important
* Corresponding author (telephone +45 3533 2436; fax +45 3533
2398; e-mail jn@life.ku.dk).
† Department of Basic Sciences and Environment.
# Present address: BioProcess Laboratories, Novo Nordisk A/S,
Brudelysvej 20, DK-2880 Bagsværd, Denmark.
§ Department of Food Science.
10.1021/jf8024826 CCC: $40.75  2008 American Chemical Society
Published on Web 09/11/2008

Emulsifier-Antioxidant Bioconjugates
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9259
alteration of the lipid/water partition coefficient of the water-soluble
antioxidants by conjugation of fatty acids of different chain lengths
to form esters has been shown gradually to alter the antioxidant
activity of the compound (11, 12). However, this effect is highly
dependent on which antioxidant assay is used and in which system
the antioxidant efficiency is investigated. Thus, in addition to the
radical scavenging activity of an antioxidant, both its polarity and
spatial interaction with lipid bilayers become important for the
antioxidant activity during lipid oxidation in heterogeneous systems
(11).
Hence, a systematic approach is required to clarify which
structures for specific phenolic groups most efficiently postpone
or terminate oxidation in heterogeneous systems. In the present
study we have used combinatorial chemistry to establish a library
of antioxidant-emulsifier bioconjugates from Trolox (a water-
soluble vitamin E analogue), dihydroferulic acid, dihydrocaffeic
acid, and gallic acid. Systematic variation in each of the four
antioxidants was obtained by conjugation with serine ester of
ethanol, serine ester of lauryl alcohol, and serine amide of lauric
Figure 1. Synthesis of serine building blocks.
R-tocopherol and phosphatidylcholine alone or in combination. The
increased antioxidative effect of R-tocopherol-phosphatidylcholine
has been explained by its ability to form reverse micelles in oil
with a trace amount of water, resulting in better accessibility of
the radical scavenging chroman-6-ol group to the polar site, where
iron-dependent initiation reactions take place (10). In addition,
Figure 2. Synthesis of conjugates: method A, 1, 2, or 3, DCC, DMAP in CH2Cl2, 20 °C, 16 h; method B, (a) TFA, CH2Cl2, 20 °C, 1.5 h, (b) 10% Pd/C,
1 atm H2, EtOAc, MeOH, 16 h; method C, 10% Pd/C, 1 atm H2, EtOAc, MeOH, 16 h.

9260 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Hunneche et al.
Figure 3. Alternative method resulting in a deprotected conjugate with the double bond intact.
(Shimadzu Corp., Kyoto, Japan).
acid to obtain both various alkyl chain lengths (ethyl vs lauryl)
and different charge distributions for similar structures, that is,
free amine versus free carboxylic acid for the lauryl derivatives.
Serine was chosen as a versatile natural building block easily
forming both ester and amide derivatives (Figure 1). Each of
the 12 library members has been synthesized, as outlined in
Figure 2, and the antioxidative activity evaluated by their ability
to scavenging rate of the DPPH radical in methanol (MeOH)
as a homogeneous solvent; increase the scavenging rate of the
DPPH radical in a heterogeneous solvent system; decrease the
rate of oxygen consumption in an oil-in-water emulsion; and
reduce formation of conjugated dienes in liposomes and
accordingly prolong the lag time for onset of oxidation. By
synthesis of such emulsifier-antioxidant conjugates, which
depending on structure and charge will concentrate in the
lipid-water interface to various degrees, oxidation of lipids in
heterogeneous systems such as emulsions, membranes, and
liposomes can be investigated with respect to reaction mecha-
nism. Throughout this study, our new bioconjugates were
compared to the parent antioxidants from which they were
derived. Notably, all of the bioconjugates synthesized and
investigated are assembled from natural building blocks and
would during digestion hydrolyze to naturally occurring com-
pounds. This investigation represents to the best of our
knowledge the first such example of the application of combi-
natorial chemistry and design to solve problems in food
chemistry. Such combinatorial libraries should be useful for
further systematic evaluation of the properties and effectiveness
of antioxidants in a relevant series of assays rather than the more
classical approaches based on more random selection of potential
antioxidants.
Synthesis. N-tert-Butoxycarbonyl-L-serine Lauryl Ester, 2. N-tert-
Butoxycarbonyl-L-serine ethyl ester, 1 (3.32 g, 14.2 mmol), lauryl
alcohol (31 g, 166 mmol), and dibutyltin oxide (Bu2SnO) (0.35 g, 1.42
mmol) were mixed and stirred under N2 at 100 °C overnight. Lauryl
alcohol was distilled off at reduced pressure (1 Torr). The residue was
dissolved in CH2Cl2, washed with saturated NaHCO3, dried (Na2SO4),
and concentrated and purified by column chromatography [gradient of
ethylacetate (EtOAc) in heptane] to give 2 as an oil: yield, 3.84 g,
72%; TOF ESMS (m/z) [M + H]+ calcd for C15H32NO3 274.2382;
found, 274.2387.
N-Lauroyl-L-serine Benzyl Ester, 3. L-Serine benzyl ester hydro-
chloride (2.5 g, 10.8 mmol), 4-dimethylaminopyridine (DMAP) (1.32
g, 10.8 mmol), triethylamine (Et3N) (1.5 mL, 10.8 mmol), and CHCl3
(85 mL) were mixed under Ar, and lauroyl chloride (2.5 mL, 10.8
mmol) was added dropwise at room temperature. The reaction
mixture was stirred overnight. Water (200 mL) was added, and the
two phases were separated. The organic phase was dried (Na2SO4)
and concentrated to a yellow solid. The solid was triturated with
diethyl ether, resulting in a light yellow powder: yield, 3.86 g, 94%;
mp, 74-75 °C; TOF ESMS (m/z) [M + H]+ calcd for C22H36NO4
378.2644; found, 378.2679.
Method A. Serine building block (1, 2, or 3, 1.5 mmol), benzylated
antioxidant (1.5 mmol), DMAP (0.18 g, 1.5 mmol), and dicyclohexy-
lcarbodiimide (DCC) (0.31 g, 1.5 mmol) were dissolved in dry CH2Cl2
(20 mL).
The reaction mixture was stirred overnight and filtered. The filtrate
was washed with 10% KHSO4, dried, and concentrated. The residue
was crystallized from ethanol (EtOH) to give white crystals.
Method B. Fully protected conjugate (0.76 mmol) and trifluoroacetic
acid (TFA) (1 mL) were dissolved in dry CH2Cl2 (4 mL), and the
reaction mixture was stirred for 1.5 h. The solvents were evaporated,
and the residue was dissolved in EtOAc (40 mL) and MeOH (10 mL).
Ten percent Pd/C was added, and the mixture was stirred under H2 (1
atm) overnight. The mixture was filtered through silica gel, which was
washed with MeOH. The filtrate was concentrated and the residue
purified by HPLC.
MATERIALS AND METHODS
General Experimental Information. Commercially available re-
agents (Sigma-Aldrich Chemie GmbH, Steinheim, Germany) were used
without further purification unless otherwise noted. Trolox was used
as the racemic mixture. The reference compound dihydrocaffeic acid
was obtained from a hydrogenation of caffeic acid. Solvents used for
the synthesis were of analytical grade, dried over activated 4 Å
molecular sieves when necessary (all solvents used under dry conditions
had a water content of <25 ppm measured by coulometric Karl Fischer
titration). Analytical TLC was performed using precoated silica gel 60
F254 plates (Merck, Darmstadt, Germany) and visualized using either
UV light or potassium permanganate stain. Flash chromatography was
performed on silica gel 60 (0.040-0.063 mm) (Merck). HPLC was
performed on a Waters 2525 system equipped with a Waters 2996
photodiode array detector and a Waters 2767 Sample Manager using a
100 mm × 19 mm i.d. XTerra prep MS C18 column (Waters) with a
gradient of acetonitrile in Milli-Q water with a flow of 15 mL/min
(Waters Corp., Milford, MA). Melting points were measured on a
Reichert melting point microscope, model N254-1R (Vienna, Austria).
1H and 13C NMR spectroscopic data (see the Supporting Information
for detailed data) were recorded on a Bruker Avance 300 (Bruker
BioSpin MRI GmbH, Ettlingen, Germany) using deuterated solvents
as a lock. Chemical shifts are reported in parts per million relative to
the residual solvent peak (1H NMR) or the solvent peak (13C NMR)
as the internal standard. Accurate mass determinations were performed
on a Micromass LCT apparatus (Manchester, U.K.) equipped with an
AP-ESI probe calibrated with Leu-Enkephalin (556.2771 g/mol). All
spectrophotometric measurements were performed on a Shimadzu UV-
2101PC UV-vis scanning spectrophotometer with automatic cell
changer and a temperature-controlled water-jacket-regulated cell holder
Method C. Fully benzylated conjugate (0.24 mmol) was dissolved
in EtOAc (20 mL) and MeOH (5 mL). Ten percent Pd/C was added,
and the mixture was stirred under H2 (1 atm) overnight. The mixture
was filtered through silica gel, which was washed with MeOH. The
filtrate was concentrated and the residue purified by HPLC.
O-(6-Benzyloxy-2,5,7,8-tetramethylchroman-2-carbonyl)-N-tert-bu-
toxycarbonyl-L-serine Ethyl Ester, 4. Method A. The residue was purified
by flash chromatography (EtOAc/heptane) followed by HPLC to give
4 as a viscous oil: yield, 17%; TOF ESMS (m/z) [M + H]+ calcd for
C26H34NO6 456.2386; found, 456.2393.
O-(6-Benzyloxy-2,5,7,8-tetramethylchroman-2-carbonyl)-N-tert-bu-
toxycarbonyl-L-serine Lauryl Ester, 5. Method A. The residue was
purified by flash chromatography (EtOAc/heptane) to give 5 as an oil:
yield, 65%; TOF ESMS (m/z) [M + H]+ calcd for C36H54NO6 596.3951;
found, 596.3979.
O-(6-Benzyloxy-2,5,7,8-tetramethylchroman-2-carbonyl)-N-lauroyl-
L-serine Benzyl Ester, 6. Method A. The residue was purified by flash
chromatography to give an oil: yield, 60%; TOF ESMS (m/z) [M +
H]+ calcd for C43H58NO7 700.4213; found, 700.4272.
O-(6-Hydroxy-2,5,7,8-tetramethylchroman-2-carbonyl)-L-serine Ethyl
Ester, 7. Method B: orange oil: yield, 25%; TOF ESMS (m/z) [M +
H]+ calcd for C19H28NO6 366.1917; found, 366.1906.
O-(6-Hydroxy-2,5,7,8-tetramethylchroman-2-carbonyl)-L-serine Lau-
ryl Ester, 8. Method B: yield, 100%; TOF ESMS (m/z) [M + H]+
calcd for C29H48NO6 506.3476; found, 506.3473.
O-(6-Hydroxy-2,5,7,8-tetramethylchroman-2-carbonyl)-N-lauroyl-L-
serine, 9. Method C: yield, 70%; TOF ESMS (m/z) [M + H]+ calcd
for C29H46NO7 520.3274; found, 520.3252.

Emulsifier-Antioxidant Bioconjugates
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9261
O-(4-Benzyl-3-methylcaffeoyl)-N-tert-butoxycarbonyl-L-serine Ethyl
Ester, 1. Method A: yield, 82%; mp, 99-100 °C; TOF ESMS (m/z)
[M + H]+ calcd for C22H26NO6 400.1760; found, 400.1784.
O-(4-Benzyl-3-methylcaffeoyl)-N-tert-butoxycarbonyl-L-serine Lauryl
Ester, 11. Method A: yield, 75%; mp, 81-82 °C; TOF ESMS (m/z)
[M + H]+ calcd for C37H54NO8Na 663.3747; found, 663.3705.
O-(4-Benzyl-3-methylcaffeoyl)-N-lauroyl-L-serine Benzyl Ester, 12.
Method A: yield, 100%; mp, 87-88 °C; TOF ESMS (m/z) [M + H]+
calcd for C39H50NO7 644.3587; found, 644.3607.
O-(3,4-Dibenzylcaffeoyl)-N-tert-butoxycarbonyl-L-serine Ethyl Ester,
13. Method A: yield, 75%; mp, 99-100 °C; TOF ESMS (m/z) [M +
H]+ calcd for C33H40NO8 578.2754; found, 578.2745.
O-(3,4-Dibenzylcaffeoyl)-N-tert-butoxycarbonyl-L-serine Lauryl Es-
ter, 14. Method A. The residue was purified by flash chromatography
(EtOAc/heptane) to give 14 as an oil: yield, 27%; mp, 99-100 °C;
TOF ESMS (m/z) [M + H]+ calcd for C38H50NO6 616.3638; found,
616.3619.
O-(3,4-Dibenzyloxycinnamoyl)-N-lauroyl-L-serine Benzyl Ester, 15.
Method A: yield, 89%; mp, 56-60 °C; TOF ESMS (m/z) [M + H]+
calcd for C45H54NO7 720.3900; found, 720.3970.
O-(4-Hydroxy-4-methoxyphenyl-3-propanoyl)-L-serine Ethyl Ester,
16. Method B. The intermediate was isolated and characterized by
NMR: yield, 83%; TOF ESMS (m/z) [M + H]+ calcd for C15H22NO6
312.1447; found, 312.1477.
O-(4-Hydroxy-4-methoxyphenyl-3-propanoyl)-L-serine Lauryl Ester,
17. Method B: overall yield, 21%; TOF ESMS (m/z) [M + H]+ calcd
for C25H42NO6 452.3012; found, 452.2987.
O-(4-Hydroxy-4-methoxyphenyl-3-propanoyl)-N-lauroyl-L-serine, 18.
Method C: yield, 98%; TOF ESMS (m/z) [M + H]+ calcd for
C25H40NO7 466.2805; found, 466.2780.
O-(3,4-Dihydroxyphenyl-3-propanoyl)-L-serine Ethyl Ester, 19. Method
B: overall yield, 100%; TOF ESMS (m/z) [M + H]+ calcd for
C14H20NO6 298.1291; found, 298.1283.
O-(3,4-Dihydroxyphenyl-3-propanoyl)-L-serine Lauryl Ester, 20.
Method B: overall yield, 34%; TOF ESMS (m/z) [M + H]+ calcd for
C24H40NO6 438.2856; found, 438.2852.
O-(3,4-Dihydroxyphenyl-3-propanoyl)-N-lauroyl-L-serine, 21. Method
C: yield, 31%; TOF ESMS (m/z) [M + H]+ calcd for C24H38NO7
452.2648; found, 452.2618.
O-Tribenzylgalloyl-N-tert-butoxycarbonyl-L-serine Ethyl Ester, 22.
Method A: yield, 62%; mp, 98-100 °C; TOF ESMS (m/z) [M + H]+
calcd for C33H34NO7 556.2335; found, 556.2284; calibrated with Fmoc-
Cys(Acm)-OH (414.1249 g/mol).
O-Tribenzylgalloyl-N-tert-butoxycarbonyl-L-serine Lauryl Ester, 23.
Method A: yield, 52%; mp, 81-82 °C; TOF ESMS (m/z) [M + H]+
calcd for C43H54NO7 696.3900; found, 696.3879.
O-Tribenzylgalloyl-N-lauroyl-L-serine Benzyl Ester, 24. Method A:
yield, 56%; mp, 106-108 °C; TOF ESMS (m/z) [M + H]+ calcd for
C50H58NO8 800.4162; found, 800.4159.
O-Galloyl-L-serine Ethyl Ester, 25. Method B: overall yield, 75%;
TOF ESMS (m/z) [M + H]+ calcd for C12H16NO7 286.0927; found,
286.0955.
O-Galloyl-L-serine Lauryl Ester, 26. Method B: overall yield, 60%;
TOF ESMS (m/z) [M + H]+ calcd for C22H36NO7 426.2492; found,
426.2456.
The reaction mixture was stirred for 1 h. The solvents were evaporated,
and the residue was freeze-dried overnight. The residue was purified
by flash chromatography (EtOAc/heptane) and then by HPLC to give
44 mg of amorphous powder: overall yield, 20%; TOF ESMS (m/z)
[M + H]+ calcd for C14H18NO6 296.1134; found, 296.1115.
Antioxidant and Radical Scavenging Assays. Chemicals. Horse
heart myoglobin (MMb, type III), methyl linoleate, and Tween 20 were
obtained from Sigma (St. Louis, MO). 2,2-Diphenyl-1-picrylhydrazyl
hydrate (DPPH•) and L-R-phosphatidylcholine were from Sigma-Aldrich
Chemie GmbH (Steinheim, Germany). 2,2′-Azobis(2-amidinopropane)
(AAPH) was from Wako Chemicals, Richmond, VA. Trolox (A0),
ferulic acid (B0′), dihydroferulic acid (B0), caffeic acid (C0′), dihy-
drocaffeic acid (C0), and gallic acid (D0) were also tested as reference
compounds. The emulsifier-antioxidant bioconjugates synthesized and
examined are all shown in Figure 4.
Radical ScaVenging of DPPH• in MeOH. Scavenging of DPPH• by
the antioxidants was followed at 515 nm (25 °C) in MeOH (13). DPPH•
(0.1 mM) and antioxidant (0.98 mM) were mixed directly in disposable
cuvettes, and absorbance at 515 nm was measured every 0.1 s for 60 s.
The measurements were performed in triplicate. With large excess of
antioxidant, the radical scavenging was expected to follow pseudo-
first-order kinetics. However, as some deviation from first-order kinetics
was seen for certain bioconjugates, probably due to secondary reactions
following the initial scavenging reactions, the initial rate of reaction
(Vi in mol/L/s) was determined from the slope of the initial linear part
of the curve in the plot of [DPPH] versus time by division with the
molar absorptivity of DPPH of ε ) 1.25 × 104 L/mol/cm.
Radical ScaVenging of DPPH• in a Heterogeneous SolVent System.
All reactions were performed in sodium phosphate buffer (pH 7.0, 5
mM) with 4% (w/v) Tween 20 (14). DPPH• was dissolved in EtOH
and then diluted 40 times in 5 mM sodium phosphate buffer (pH 7.0)
with 4% (w/v) Tween 20 to give a concentration of 0.125 mM. The
antioxidants were dissolved directly in buffer with 4% (w/v) Tween
20 to give a concentration of 5 mM. DPPH (0.1 mM) and antioxidant
(1.0 mM) were mixed carefully in disposable cuvettes, and absorbance
at 515 nm was measured every 0.1 s for 60 s.
In the case of Trolox and its derivatives (group A compounds) it
was necessary to dissolve the antioxidants in EtOH and then dilute the
solution in buffer with 4% (w/v) Tween 20. The scavenging of DPPH•
by group A compounds was much faster than that shown by the rest
of the investigated compounds, and the experiments were performed
using a DX-17MV stopped-flow spectrofluorometer (Applied Photo-
physics, London, U.K.). Each syringe was filled with 0.2 mM DPPH•
in 5 mM sodium phosphate buffer (pH 7.0) with 4% (w/v) Tween 20
prepared from a 3 mM DPPH in EtOH solution and 2.0 mM antioxidant
in 5 mM sodium phosphate buffer (pH 7.0) with 4% (w/v) Tween 20
prepared from a solution of 5 mM antioxidant in EtOH giving the same
reaction concentrations as above (0.10 mM DPPH• and 1.0 mM
antioxidant). Absorbance at 515 nm was measured every 0.0025 s for
1.000 s. The measurements were performed in triplicate. The initial
rate of reaction was calculated as for DPPH• scavenging in MeOH.
Oxygen Consumption. Oxygen consumption was measured as
described by Hu and Skibsted (15). Methyl linoleate was mixed with
Tween 20 and air-saturated thermostated (25 °C) phosphate buffer (pH
6.8), and 20 µL of antioxidant solution (1.0 mM) in MeOH was added
to give a final concentration of antioxidant of 7.9 µM. The oxidation
was initiated by the addition of 25 µL of 0.20 mM MMb aqueous
solution, and immediately thereafter measurements of the oxygen
consumption were started. As a positive blank 20 µL of MeOH was
used instead of antioxidant solution, and as a negative blank no
antioxidant or no MMb was added. The relative oxygen consumption
was measured with oxygen microsensors (Unisense, Aarhus N,
Denmark) and recorded at time intervals of 10 s for 10 min. Profix
Software v. 3.05 (Unisense, Denmark) was used for data handling. The
initial rate of consumption [V(O2)] was calculated from the slope of
the oxygen consumption versus time curve in the linear region. All
measurements were performed in duplicate. The influence of each of
the antioxidants on the initial rate of oxygen consumption was expressed
as an antioxidative index relative to the rate in the absence of antioxidant
according to eq 1.
O-Galloyl-N-lauroyl-L-serine, 27. Method C: yield, 86%; TOF ESMS
(m/z) [M + H]+ calcd for C22H34NO8 440.2284; found, 440.2256.
O-(3,4-Diethoxycarbonylcaffeoyl)-N-tert-butoxycarbonyl-L-serine Eth-
yl Ester, 28. N-tert-Butoxycarbonyl-L-serine ethyl ester (0.48 g, 2.04
mmol) was dissolved in pyridine (5 mL) under N2 and cooled on ice.
3,4-Diethoxycarbonylcaffeoyl chloride (28)(0.70 g, 2.04 mmol) was
dissolved in toluene (5 mL) and added. After 10 min at 0 °C, the
reaction mixture was stirred at room temperature overnight. The solvents
were evaporated and coevaporated with toluene twice. The residue was
purified by flash chromatography (gradient of EtOAc in heptane) to
give 28 as an oil: 0.41 g; yield, 37%; TOF ESMS (m/z) [M + H]+
calcd for C20H26NO10 440.1557; found, 440.1533.
O-Caffeoyl-L-serine Ethyl Ester, 29(Figure 3). 28 (0.41 g, 0.76
mmol) was dissolved in CH2Cl2 (4 mL) and TFA (1 mL), and the
reaction mixture was stirred for 3 h. The solvents were evaporated,
and the residue was dissolved in MeOH (14.5 mL) and NH3 (aq, 25%).

9262 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Hunneche et al.
Figure 4. Emulsifier-antioxidant bioconjugate library: overview of all the compounds tested in the antioxidant assays (the serine building blocks were
not tested). The numbers in parentheses refer to the compound numbering under Materials and Methods.
differences, and least-squares differences (LSD) were applied to
compare the mean values of the tested antioxidants.
V(O2) with antioxidant
Ioxygen ) 1 -
(1)
V(O2) without antioxidant
Preparation of Liposomes. Liposomes were prepared as described
by Roberts and Gordon (16). Briefly, 1.35 mg of L-R-phosphatidyl-
choline (PC) from soybean in 2 mL of CHCl3 and 15 µL of antioxidant
(1 mM) in MeOH (or 15 µL of MeOH for blank) were transferred to
a round-bottom flask, which was covered with aluminum foil, and the
solvents were evaporated at reduced pressure (water bath at 30 °C).
Nitrogen was introduced when atmospheric pressure was established,
and 10 mL of 10 mM sodium phosphate buffer (pH 7.4) was added.
The content of the flask was vortexed for 10 min followed by 30 s in
an ultrasonic bath. This produced a homogeneous white suspension.
Large unilamellar liposomes were obtained by transferring the liposome
suspension to a small volume extrusion device (Liposofast Basic,
Aventin, Mannheim, Germany). The suspension was passed 20 times
through a double layer of polycarbonate membranes (100 nm pore size).
Peroxidation of Liposomes. The liposome suspension (2.5 mL) was
pipetted into a quartz cuvette with a stopper and incubated for 10 min
at 37 °C in the spectrophotometer (Shimadzu UV-2101PC UV-vis
scanning spectrophotometer). Lipid peroxidation was initiated by the
addition of 25 µL of 75 mM AAPH in 10 mM sodium phosphate buffer
(pH 7.4). The cuvettes were sealed with the stopper and inverted three
times. The absorbance was recorded at 234 nm every 10 min against
a blank of sodium phosphate buffer (pH 7.4). Each antioxidant was
tested in duplicate. The contribution of AAPH in buffer to the
absorbance at 234 nm was subtracted from the absorbance obtained
for the antioxidants. The lag phase was measured as the time in minutes
to the point where a tangent to the propagation phase intercepted the
x-axis (17).
RESULTS AND DISCUSSION
Synthesis. Three different amino acid building blocks were
selected representing a short-chain and a long-chain serine ester
as well as a long-chain serine amide. The first L-serine building
block, N-tert-butoxycarbonyl-L-serine ethyl ester, 1 (18), was
synthesized from L-serine ethyl ester and di-tert-butyl pyrocar-
bonate according to standard procedures (19). Transesterification
of 1 to yield the corresponding lauryl ester 2 was done using
dibutyltin oxide as catalyst using a method hitherto described
for methanol, ethanol, and n-butanol as well as allyl, isopropyl,
and benzyl alcohol (18). The third serine building block, 3, was
obtained by acylation of L-serine benzyl ester hydrochloride in
the presence of triethylamine and DMAP using standard
procedures (20).
Four different well-established antioxidants were converted
into the appropriately protected building blocks required for the
subsequent formation of the bioconjugates. Racemic 6-benzy-
loxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (21), ben-
zylferulic acid (22), dibenzylcaffeic acid (23), and tribenzylgallic
acid (24) were all synthesized using methods described by
Moulin et al. (21). Then the benzyl ethers of the four antioxi-
dants were coupled with the three serine building blocks (1-3)
using DCC and DMAP in DCM, resulting in 12 different
protected antioxidant-emulsifier conjugates. The yields varied
from 17 to 100%.
Statistical Analysis. Results were analyzed statistically by analysis
of variance using the general linear models (GLM) procedure in the
SAS 9.1 package, SAS Institute, Inc. Antioxidant was included as a
fixed effect. In some cases it was necessary to transform data
logarithmically to obtain homogeneous variance, and data were
subsequently analyzed as described above. Means were used to compare
Because a racemic mixture of 6-hydroxy-2,5,7,8-tetrameth-
ylchroman-2-carboxylic acid (Trolox) was used and serine was
used, as the L-isomer, the three conjugates containing the Trolox
building block were obtained as a mixture of the two diaste-

Emulsifier-Antioxidant Bioconjugates
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9263
Table 1. Overview of All Results in the Four Antioxidant or Radical Scavenging Assaysa
vi in MeOHa (10-6 L/mol/s)
antioxidant
vi in micelles (10-6 L/mol/s)
Ioxygen
lag phase (min)
ClogP
A0
A1
A2
A3
4.6 ( 0.5 b
4.2 ( 0.2 bc
18 ( 3 a
490 ( 10 a
220 ( 20 d
378 ( 8 b
340 ( 10 c
0.75 ( 0.08 b
0.97 ( 0.03 a
0.96 ( 0.04 a
0.95 ( 0.01 a
67 ( 6 ns
67 ( 3 ns
88 ( 24 ns
81 ( 4 ns
3.0888
3.1349
8.4249
7.8394
3.5 ( 0.1 c
B0′
B0
B1
B2
B3
1.1 ( 0.2 a
0.29 ( 0.04 c
0.80 ( 0.07 a
0.115 ( 0.009 d
0.55 ( 0.02 b
2.1 ( 0.1 a
1.86 ( 0.04 b
1.51 ( 0.04 d
1.8 ( 0.1 bc
1.63 ( 0.07 cd
0.45 ( 0.15 b
0.50 ( 0.03 b
0.46 ( 0.05 b
0.96 ( 0.03 a
0.96 ( 0.02 a
172 ( 33 a
79 ( 4 c
28 ( 2 d
80 ( 3 c
130 ( 7 b
1.4212
1.0852
1.1523
6.4423
5.8568
C0′
C1′
C0
C1
C2
C3
4.1 ( 0.4 a
3.1 ( 0.5 b
4.9 ( 0.4 a
1.4 ( 0.7 c
1.2 ( 0.2 c
1.5 ( 0.1 c
3.8 ( 0.2 c
3.83 ( 0.09 c
4.9 ( 0.6 b
3.5 ( 0.2 c
10.0 ( 0.7 a
9.2 ( 0.5 a
0.49 ( 0.09 c
0.71 ( 0.01 b
0.49 ( 0.02 c
0.58 ( 0.04 c
0.995 ( 0.002 a
0.96 ( 0.01 a
152 ( 9 b
107 ( 4 e
177 ( 3 a
66 ( 3 e
131 ( 10 c
144 ( 3 bc
0.975
0.7061
0.639
0.6475
5.9961
5.4106
D0
D1
D2
D3
3.9698 ( 0.0004 a
3.3 ( 0.2 b
4.01 ( 0.06 a
1.055 ( 0.007 c
4.8 ( 0.4 c
4.0 ( 0.2 d
10.8 ( 0.6 a
6.7 ( 0.2 b
0.34 ( 0.07 c
0.526 ( 0.005 b
0.95 ( 0.01 a
0.87 ( 0.04 a
47 ( 1 ns
75 ( 26 ns
66 ( 9 ns
74 ( 4 ns
0.42549
0.33253
5.62253
5.03703
a The initial rates (vi) in MeOH and micelles were determined for DPPH• scavenging, Ioxygen was determined from electrochemical oxygen consumption measurements,
and lag phase was obtained from diene measurements in the liposome assay. The calculated 1-octanol/water partition coefficient (ClogP) is also given for all compounds.
Values are given as mean ( standard deviation (n ) minimum 2). Within the same antioxidant group, values bearing different letters are significantly different (P < 0.05)
within the same column (ns, nonsignificant).
reoisomers. No attempts to separate the diastereoisomers were
performed. This inherently complicates the NMR spectra,
because the two isomers may give different chemical shift
values.
included the use of ethyl carbonate as the phenol protecting
group (28). The antioxidant and the serine building block (1)
were coupled via the acid chloride (28) in toluene and pyridine
to give the conjugate 28 in moderate yield (37%). Deprotection
was done in two steps: first, a removal of the Boc group as
described above followed by hydrolysis of the carbonate in
aqueous ammonia and MeOH, resulting in 29 in moderate
overall yield (20%).
Antioxidant and Radical Scavenging Activity of Biocon-
jugates. All results of the antioxidant or radical scavenging
assays are summarized in Table 1. The octanol/water partition
coefficient (ClogP) of each bioconjugate and parent antioxi-
dant was calculated using ChemDraw 10.0 (CambridgeSoft
Corp., Cambridge, MA), and these values are also included
in Table 1.
Initially, we attempted to deprotect the compounds in one
step using thioanisole and TFA (25), but these conditions
resulted in several byproducts, and in the case of the caffeic
acid and ferulic acid derivatives, the conditions led to a
modification/reaction at the double bond (the structures of the
byproduct were not determined). Instead, a number of different
standard methods for the selective removal of benzyl groups in
the presence of double bonds were attempted, but, unfortunately,
none of these proved to be successful. Therefore, it was finally
decided to remove the benzyl groups by hydrogenolysis, which
concomitantly resulted in reduction of the double bonds in
caffeic acid and ferulic acid. It has been shown that the
antioxidative effects of caffeic acid and ferulic acid and their
reduced analogues, respectively, are comparable (26, 27), and
accordingly we expected that it would not have a significant
impact on the results in the antioxidant assays.
ScaVenging of DPPH• in MeOH. The concentration of DPPH
was chosen to obtain an intermediate rate of scavenging of
DPPH• for most of the antioxidants, with the ferulic acid
derivatives being slowest (Figures 5 and 6) and Trolox
derivatives being fastest (Figure 6). Ferulic acid scavenged
DPPH• with the fastest rate compared to the other group B
compounds including dihydroferulic acid. However, conjugating
dihydroferulic acid with serine ethyl ester (1) and lauroyl serine
(3) improved the scavenging rate from 2.9 × 10-7/s for B0 to
8.0 × 10-7 and 5.5 × 10-7/s for B1 and B3, respectively.
The correlation between the initial rate (Vi) for scavenging
of DPPH• in MeOH and the type of conjugate is shown in
Figure 6. For group A compounds, only A2 exhibited a faster
scavenging rate than Trolox. Negative or no improvement of
conjugating the antioxidants on the rate of scavenging DPPH•
was found for groups C and D. Improvement of the radical
scavenging activity of the conjugates was expected in hetero-
geneous systems as emulsions containing interfaces. Hence, the
lack of improvement of the radical scavenging activity in a
homogeneous solution such as MeOH could be expected. The
presence of the double bond in caffeic acid did not seem to
have any significant positive effect on the scavenging rate of
DPPH• as the scavenging rate by dihydrocaffeic acid was slightly
higher than the scavenging rate by caffeic acid.
In the case of compounds A3-D3, which contained only
benzyl groups as protecting groups, the deprotection was done
in a single step (method C) involving standard hydrogenation
conditions at atmospheric pressure using 10% Pd/C in EtOAc
and MeOH. The yields ranged from 31 to 98%. In the case of
compounds A1-D1 and A2-D2, containing both a Boc group
and a benzyl group, the protecting groups were removed in two
steps (method B): first, the Boc group was removed using TFA
in CH2Cl2 (20% v/v). Normally, the Boc-deprotected compounds
were used in the next deprotection step without further purifica-
tion, but in one case (the deprotection of 10) the intermediate
1
was isolated and characterized by H NMR and 13C NMR to
validate the method. Finally, the benzyl groups were removed
using the same method as described under Method C. The
overall yields ranged from 21 to 100% for the two deprotection
steps.
The importance of the double bond in caffeic acid derivatives
was assessed by synthesizing one caffeic acid conjugate using
a method that would leave the double bond untouched. This

9264 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Hunneche et al.
Figure 7. Scavenging of DPPH• (0.1 mM) in emulsion with 0.1 M sodium
phosphate buffer (pH 7.0) and 4% (w/v) Tween 20 at 25 °C by 1.0 mM
of gallic acid (9), D1 (2), D2 (1), and D3 ([) and in absence of
antioxidant (b).
Figure 5. Scavenging of DPPH• (0.1 mM) in MeOH at 25 °C by 0.98
mM of ferulic acid (b), dihydroferulic acid (9), B1 (2), B2 (1), or B3
([).
Figure 6. Correlation between the initial rate (vi) for scavenging of DPPH•
in MeOH and type of conjugate for each antioxidant and its derivatives.
A group (9): A0 (Trolox), A1, A2, and A3. B group (b): B0 (dihydroferulic
acid), B1, B2, and B3. B0′ (O) (ferulic acid). C group (2): C0
(dihydrocaffeic acid), C1, C2, and C3. C′ group (4): C0′ (caffeic acid)
and C1′. D group (1): D0 (gallic acid), D1, D2, and D3.
Figure 8. Correlation between the initial rate (vi) for scavenging of DPPH•
in aqueous micelles (pH 7.0) and type of conjugate for each antioxidant
and its derivatives. A group (9): A0 (Trolox), A1, A2, and A3. B group
(b): B0 (dihydroferulic acid), B1, B2, and B3. B0′ (O) (ferulic acid). C
group (2): C0 (dihydrocaffeic acid), C1, C2, and C3. C′ group (4): C0′
(caffeic acid) and C1′. D group (1): D0 (gallic acid), D1, D2, and D3.
ScaVenging of DPPH• in a Heterogeneous SolVent. The
compounds were designed to improve the antioxidative effect
in systems with interfaces, and therefore the DPPH assay
was repeated in an aqueous solution, where the radicals are
solublized in micelles. DPPH was dissolved in a small volume
of EtOH and subsequently diluted in phosphate buffer (pH
7.0) with Tween 20 to avoid precipitation of DPPH. Similarly,
Trolox and its derivatives were dissolved in EtOH and
subsequently diluted in buffer as these compounds are more
lipophilic than the rest of the conjugates. Furthermore,
scavenging of DPPH• by Trolox and its derivatives was fast,
and it was necessary to follow the reaction using a stopped-
flow spectrofluorometer.
Comparison of the rate of DPPH• scavenging in this
heterogeneous solvent in the presence of gallic acid and its
derivatives showed that conjugating gallic acid with serine lauryl
ester (2) or lauroyl serine (3) increased the scavenging rate
(Figure 7). However, conjugation of gallic acid with serine ethyl
ester (1) decreased the DPPH• scavenging rate, indicating an
expected reduction of the antioxidative activity.
The correlation between Vi in emulsion and the type of
conjugate is illustrated in Figure 8. A similar effect of
conjugating dihydrocaffeic acid (C0) as for gallic acid was
observed. In contrast, no effect of conjugating the B group
compounds on the scavenging rate of DPPH• was found, and
the conjugates of Trolox (group A compounds) exhibited slower
DPPH• scavenging than Trolox. The presence of the double bond
in ferulic acid seems to have a slightly positive influence on
the scavenging rate of DPPH• as a small, but significant,
difference was observed in Vi between B0 and B0′. In contrast,
the DPPH• scavenging rate observed in the presence of
dihydrocaffeic acid (C0) was higher than that in the presence
of caffeic acid (C0′), indicating that removal of the double bond
in caffeic acid improved the radical scavenging activity sig-
nificantly. These observations are similar to the results obtained
for DPPH• scavenging in homogeneous MeOH.
Oxygen Consumption. Initially, the oxygen consumption at
different concentrations of antioxidant was measured. The final
concentration of 7.9 µM antioxidant was chosen because it gave
an intermediate inhibition of the oxidation of methyl linoleate.

Emulsifier-Antioxidant Bioconjugates
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9265
Figure 9. Oxygen consumption in an emulsion of methyl lineolate, Tween
20, and phosphate buffer (25 °C, pH 6.8) with initiation of oxidation by
MMb in the presence of 7.9 µM of ferulic acid (b), dihydroferulic acid
(9), B1 (2), B2 (1), and B3 ([) in the absence of antioxidant (0) or
both antioxidant and MMb (O).
Figure 11. Oxygen consumption index plotted as a function of the
calculated 1-octanol/water partition coefficient (ClogP). Curve is calculated
by nonlinear regression analysis according to the empiric equation Ioxygen
) k(ClogP)/[1 + (ClogP)] to yield the parameter k ) 1.10 ( 0.04.
Figure 12. Formation of conjugated dienes in liposomes (37 °C, pH 7.4)
in the presence of 1.5 µM gallic acid (9), D1 (2), D2 (1), or D3 ([) and
in the absence of antioxidant (b) measured as absorbance at 234 nm
after the addition of 0.75 mM AAPH as water-soluble radical initiator.
The lag phase in minutes obtained by determining the intercept between
the linear part of the curve with the x-axis is found in Table 1.
between the antioxidative activity of the parent antioxidants with
double bonds (B0′ and C0′) and the corresponding reduced
compounds (B0 and C0) was found.
The oxygen consumption index (Ioxygen) was considered as a
function of the calculated octanol/water partition coefficient
(ClogP) to study the antioxidant efficiency in an emulsion
depending on the hydrophobicity of the bioconjugates (Figure
11). It is suggested that Ioxygen depends on ClogP according to
the empiric eq 2.
Figure 10. Correlation between Ioxygen and type of conjugate for each
antioxidant and its derivatives. A group (9): A0 (Trolox), A1, A2, and
A3. B group (b): B0 (dihydroferulic acid), B1, B2, and B3. B0′ (O) (ferulic
acid). C group (2): C0 (dihydrocaffeic acid), C1, C2, and C3. C′ group
(4): C0′ (caffeic acid) and C1′. D group (1): D0 (gallic acid), D1, D2,
and D3.
In the presence of the compounds B0′ (ferulic acid), B0
(dihydroferulic acid), and B1, an inhibition of the oxygen
consumption was found compared to the measurement without
antioxidant (Figure 9). The oxidative indices (Ioxygen, Table 1)
for these compounds were similar (0.45, 0.50, and 0.46,
respectively), which indicates that conjugation with serine ethyl
ester (1) does not improve the antioxidative activity of dihy-
droferulic acid. In contrast, there was a remarkable reduction
in oxygen consumption in the presence of the conjugates B2
and B3 (Ioxygen ) 0.96 and 0.96, respectively), indicating a
significant effect on the antioxidative activity of conjugation
with longer chain building blocks.
A similar tendency was found for group A, C, and D
conjugates as illustrated in Figure 10 by correlation between
Ioxygen and type of conjugate. In most cases (groups A, C′, C,
and D) an improvement of antioxidative activity was found for
all conjugates as compared to the parent antioxidants, and the
most significant improvements were generally seen for the
longer chain derivatives (2 and 3). No notable difference
k
Ioxygen
)
(ClogP)
(2)
1 + (ClogP)
The parameter k was determined by nonlinear regression. As
may be seen from Figure 11, Ioxygen is approximating unity
(k ) 1.10 ( 0.04) for the most hydrophobic bioconjugates,
corresponding to a very efficient antioxidant. Other effects of
the antioxidants on lipid oxidation could be possible through a
direct interaction between antioxidants and myoglobin leading
to protein denaturation or heme dissociation. Such effects would,
however, enhance lipid oxidation.

9266 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Hunneche et al.
Trolox and Its DeriVatiVes (Group A). Improvement of the
antioxidative activities of conjugates of Trolox was observed
for oxygen consumption in a methyl linoleate emulsion and to
a smaller extent on the formation of conjugated dienes in
liposomes with the longer chain conjugates of Trolox (A2 and
A3). The scavenging rate of DPPH• was highest for Trolox itself
compared to the derivatives of Trolox.
Trolox is more lipophilic than ferulic acid, caffeic acid, and
gallic acid, and Trolox was therefore expected to exhibit better
antioxidative activities in emulsions compared to the other
conjugates, as was also observed in the oxygen consumption
assay and scavenging of DPPH• in micelles. The effect of adding
a lipophilic part to Trolox to align the antioxidant in the interface
between oil and water is accordingly not as significant as
observed for the other conjugates.
Ferulic Acid and Its DeriVatiVes (Group B). Improvement
of the antioxidative activity was found when dihydroferulic acid
was conjugated with longer chain derivatives [serine lauryl ester
(2) and lauroyl serine (3)] on the formation of conjugated dienes
in liposomes and for oxygen consumption in oil-in-water
emulsion. The results indicated that conjugation of a lipophilic
molecule to the hydrophilic antioxidant reduces oxidation in
the heterogeneous systems. No improvement on antioxidative
activity was, however, observed when dihydroferulic acid was
conjugated with serine ethyl ester (1), confirming that improve-
ment depends on conjugation with longer hydrophobic chains.
In general, the scavenging rate of DPPH• was slow, and no
marked difference was observed between the individual group
B compounds, indicating that ferulic acid and derivatives are
not good radical scavengers under the actual assay conditions.
In the study by Kikuzaki et al. (11), the percentage of DPPH
radical scavenging was determined for ferulic acid and related
compounds in EtOH. Ferulic acid scavenged more DPPH
radicals than various ferulate esters such as dodecyl ferulate,
which is consistent with the results in the present study (Figure
5).
Figure 13. Correlation between the lag phase obtained from peroxidation
of liposomes and the type of conjugate for each antioxidant and its
derivatives. A group (9): A0 (Trolox), A1, A2, and A3. B group (b): B0
(dihydroferulic acid), B1, B2, and B3. B0′ (O) (ferulic acid). C group (2):
C0 (dihydrocaffeic acid), C1, C2, and C3. C′ group (4): C0′ (caffeic acid)
and C1′. D group (1): D0 (gallic acid), D1, D2, and D3.
Peroxidation of Liposomes. The antioxidative effect of the
bioconjugates on the formation of conjugated dienes in lipo-
somes was investigated by measuring the absorbance at 234
nm. Results for gallic acid and its derivatives (group D
compounds) are shown in Figure 12. In the absence of
antioxidant the lag phase was 14 min, whereas the presence of
gallic acid increased the lag phase (47 min), showing a small
protection of gallic acid against oxidation of liposomes. The
presence of the derivatives of gallic acid (D1, D2, and D3)
further increased the lag phase (75, 66, and 74 min, respec-
tively), indicating a better protection against oxidation than gallic
acid and increased antioxidative effect of the bioconjugates
compared to the parent antioxidant, although these observations
were not statistically significant.
A marked difference in antioxidative activity between ferulic
acid and dihydroferulic acid was found only for development
of conjugated dienes in liposomes, where ferulic acid more
efficiently protected the liposomes from oxidation than dihy-
droferulic acid. No obvious difference was found in the other
antioxidant assays apart from DPPH• scavenging in MeOH, but
the rate of DPPH• scavenging in presence of group B compounds
was generally very low.
Correlation between results from the liposome assay and type
of conjugation is summarized in Figure 13. Overall, the
conjugates with serine ethyl ester as building block (1) exhibit
lower antioxidative activity than the rest of the compounds. A
slight improvement of the antioxidative activity was found for
Trolox (group A) with the long-chain building blocks, although
these results are not statistically significant. Improvement of
antioxidative activity is found when dihydroferulic acid is
conjugated with lauroyl serine (B3), compared to the compound
without conjugation. However, the compounds with the best
antioxidative effect are nonconjugated compounds: B0′ (ferulic
acid), C0 (dihydrocaffeic acid), and C0′ (caffeic acid). The effect
of the double bond of the side chain on the antioxidative
properties in the liposome assay is not clear. Removing the
double bond from ferulic acid decreases the antioxidant capacity
significantly (the lag phase is reduced from 172 to 79 min),
whereas removing the double bond from caffeic acid increases
the lag phase from 152 to 177 min.
Caffeic Acid and Its DeriVatiVes (Group C). Conjugation of
dihydrocaffeic acid with the longer chain derivatives (2 and 3)
significantly improved the antioxidative and radical scavenging
activity as measured by oxygen consumption in an oil-in-water
emulsion and as measured as scavenging rate of DPPH• in the
heterogeneous system. On the contrary, the scavenging rate of
DPPH• in MeOH, in micellular solution, and inhibition of
conjugated diene formation in liposomes were reduced when
dihydrocaffeic acid was conjugated with the tested emulsifier
building blocks. In agreement with the present study, dihydro-
caffeic acid has been found to scavenge DPPH• in EtOH more
effectively than its alkyl esters (methyl, ethyl, and propyl
dihydrocaffeate) (29). In the same study, the caffeic acid alkyl
esters were found to scavenge DPPH• more effectively than
caffeic acid itself, which is in contrast to the present study that
found the scavenging rate of DPPH• by caffeic acid (C0′) to be
faster than that by C1′. Caffeic acid and dihydrocaffeic acid
have, to our knowledge, not previously been synthesized with
longer chain conjugates than the n-C3 derivatives. The longer
conjugation seems to improve the effect for the reduced caffeic
In the present study, conjugation of water-soluble antioxidants
with longer chain derivatives was found overall to improve the
antioxidative effect in heterogeneous systems compared to the
parent antioxidants, similar to what has been observed in other
studies with retinyl ascorbate, ferulic acid esters, and gallic acid
esters (9, 11, 12). However, the actual improvement of anti-
oxidative activity is highly dependent on which assay is used
for evaluation of the antioxidants.

Emulsifier-Antioxidant Bioconjugates
J. Agric. Food Chem., Vol. 56, No. 19, 2008 9267
acid apart from the inability of the conjugates to inhibit the
formation of conjugated dienes in liposomes and to scavenge
DPPH• in methanol.
reasons does not lead to differences in the overall pattern in
effects of bioconjugation.
Combinatorial chemistry principles have been used to design
and synthesize an array of emulsifier-antioxidant bioconjugates.
Such combinatorial libraries are useful for systematic evaluation
of bioactive compounds such as antioxidants compared to a more
classical trial and error approach. In homogeneous solution only
minor improvement could be obtained for conjugation of the
antioxidants. In contrast, prevention of oxidation in heteroge-
neous systems was clearly improved when a lipophilic append-
age was added to a water-soluble polyphenolic antioxidant,
indicating that oxidation takes place in the interface between
oil and water in heterogeneous systems, and conjugation of
polyphenols with emulsifier compounds provides better protec-
tion against oxidation as the antioxidant-emulsifier conjugates
are aligned in the interface between oil and water. The
improvement of protection was shown to depend on the octanol/
water partition coefficient, and we suggest that optimal protec-
tion corresponds to saturation of the water-oil interface.
Interestingly, the reduction of the double bond in caffeic acid
improves antioxidative activity to a small extent in all applied
assays, which is consistent with the results of Silva et al. (27),
who found that propyl dihydrocaffeate showed improved
antioxidant activity compared to propyl caffeate. However, in
another study it was concluded that both caffeic acid and
dihydrocaffeic acid are efficient antioxidants for lard and human
plasma low-density lipoprotein, but the presence of the double
bond in the side chain of the catechol group affects the efficiency
of antioxidant activity depending on the environment in which
the oxidation takes place (26).
Gallic Acid and Its DeriVatiVes (Group D). Conjugation of
gallic acid showed the most significant improvement of the four
groups of antioxidants as oxygen consumption in oil-in-water
emulsions and formation of conjugated dienes clearly decreased.
The rate of scavenging of DPPH• in the heterogeneous system
was likewise increased as seen by a comparison of the results
obtained for the derivatives of serine lauryl ester (2) and lauroyl
serine (3) with the parent gallic acid. In the study of Kikuzaki
et al. (11), gallic acid and derivatives inhibited PC hydroperoxide
formation in liposomes in the order lauryl gallate > methyl
gallate > gallic acid, also indicating improved antioxidant
activity for longer chain conjugates. Furthermore, methyl gallate
was found to scavenge a larger percentage of DPPH radicals
compared to gallic acid and lauryl gallate in EtOH, showing
that no effect of the longer chain conjugates was found in a
homogeneous system (11). This observation is in agreement with
the results obtained in the present study, where no effect of
conjugating gallic acid was found on scavenging rate of DPPH•
in MeOH. However, in a study where the DPPH radical
scavenging efficiency of gallic acid and longer chain derivatives
was compared by electron spin resonance spectroscopy in
liposomes, gallic acid was found to exhibit the best antioxidative
activity compared to the longer chain derivatives (12).
ACKNOWLEDGMENT
The technical assistance of Henriette Erichsen, Andrew
Hunwald, Stig Jacobsen, and Thure F. Skovsgaard is gratefully
acknowledged.
Supporting Information Available: Detailed NMR data from
compounds 2-29. This material is available free of charge via
the Internet at http://pubs.acs.org.
LITERATURE CITED
(1) Halliwell, B.; Murcia, M. A.; Chirico, S.; Aruoma, O. I. Free
radicals and antioxidants in food and in vivo: what they do and
how they work. Crit. ReV. Food Sci. Nutr. 1995, 35, 7–20.
(2) Decker, E. A.; Livisay, S.; Zhou, S. Mechanisms of endogenous skeletal
muscle antioxidants: chemical and physical aspects. In Antioxidants in
Muscle Foods; Decker, E. A., Faustman, C., Lopez-Bote, C. J.,
Eds.; Wiley: Chichester, U.K., 2000; pp 25-59.
Effect of Charge of the Serine Building Blocks. At neutral
pH, which was selected as the reaction condition in the
antioxidant assays used in the present study, serine ethyl and
lauryl esters (1 and 2, respectively) are positively charged,
whereas lauroyl serine amides (3) are negatively charged. Group
2 compounds generally exhibited better DPPH• scavenging
activity compared to group 3 compounds, whereas group 3
compounds generally more effectively inhibited the formation
of conjugated dienes in liposomes than group 2 compounds.
No notable difference was observed for the effect on oxygen
consumption in the oil-in-water emulsions between group 2 and
3 compounds. The different assays evaluate effects of antioxi-
dants at different stages in the overall oxidation process. The
DPPH• scavenging assay relates to early stages in oxidative
damage, and it is not surprising that positively charged anti-
oxidant derivatives may scavenge radicals with a partial negative
charge better than negatively charged antioxidants. In contrast,
a negative charge seems to improve antioxidant activity toward
the formation of conjugated dienes in liposomes. Notably, in
the liposome assay, oxidation was initiated by AAPH, an azo
initiator, which forms cationic radicals. In foods and other
biological systems, Fe(III)/Fe(II) is a common catalyst for
radical generation, and both in the present assay and in foods,
negatively charged antioxidants seem to be more effective
against lipid oxidation initiated by cations (30). It should further
be noted that the use in the two assays, where lipids are
becoming oxidized, of different initiators for experimental
(3) Schwarz, K.; Bertelsen, G.; Nissen, L. R.; Gardner, P. T.;
Heinonen, M.; Hopia, A.; Huynh-Ba, T.; Lambelet, P.; McPhail,
D.; Skibsted, L. H.; Tijburg, L. Investigation of plant extracts for
the protection of processed foods against lipid oxidation. Com-
parison of antioxidant assays based on radical scavenging, lipid
oxidation and analysis of the principal antioxidant compounds.
Eur. Food Res. Technol. 2001, 212, 319–328.
(4) McClements, D. J.; Decker, E. A. Lipid oxidation in oil-in-water
emulsions: impact of molecular environment on chemical reactions
in heterogeneous food systems. J. Food Sci. 2000, 65, 1270–1282.
(5) Huang, S.-W.; Hopia, A.; Schwarz, K.; Frankel, E.; German, J. B.
Antioxidant activity of R-tocopherol and Trolox in different lipid
substrates: bulk oils vs oil-in-water emulsions. J. Agric. Food
Chem. 1996, 44, 444–452.
(6) Yuji, H.; Weiss, J.; Villeneuve, P.; Giraldo, L. J. L.; Figueroa-
Espinoza, M.-C.; Decker, E. A. Ability of surface-active antioxi-
dants to inhibit lipid oxidation in oil-in-water emulsion. J. Agric.
Food Chem. 2007, 55, 11052–11056.
(7) Nakayama, T.; Ono, K.; Hashimoto, K. Affinity of antioxidative
polyphenols for lipid bilayers evaluated with a liposome system.
Biosci., Biotechnol., Biochem. 1998, 62, 1005–1007.
(8) Figueroa-Espinoza, M.-C.; Villeneuve, P. Phenolic acids enzymatic
lipophilization. J. Agric. Food Chem. 2005, 53, 2779–2787.
(9) Abdulmajed, K.; McGuigan, C.; Heard, C. M. Topical delivery
of retinyl ascorbate: 4. Comparative anti-oxidant activity towards
DPPH. Free Radical Res. 2005, 39, 491–498.
(10) Koga, T.; Terao, J. Antioxidant activity of a novel phosphatidyl
derivative of vitamin E in lard and its model system. J. Agric.
Food Chem. 1994, 42, 1291–1294.

9268 J. Agric. Food Chem., Vol. 56, No. 19, 2008
Hunneche et al.
(11) Kikuzaki, H.; Hisamoto, M.; Hirose, K.; Akiyama, K.; Taniguchi,
H. Antioxidant properties of ferulic acid and its related com-
pounds. J. Agric. Food Chem. 2002, 50, 2161–2168.
(12) Lu, Z.; Nie, G.; Belton, P. S.; Tang, H.; Zhao, B. Structure-
activity relationship analysis of antioxidant ability and neuropro-
tective effect of gallic acid derivatives. Neurochem. Int. 2006,
48, 263–274.
(13) Brand-Williams, W.; Cuvelier, M. E.; Berset, C. Use of a free
radical method to evaluate antioxidant activity. Lebensm.
Wiss.Technol. 1995, 28, 25–30.
(14) Cardoso, D. R.; Frederiksen, A. M.; da Silva, A. A.; Franco, D. W.;
Skibsted, L. H. Sugarcane spirit extracts of oak and Brazilian
woods: antioxidant capacity and activity. Eur. Food Res. Technol.
2008, 227, 1109–1116.
(15) Hu, M.; Skibsted, L. H. Antioxidative capacity of rhizome extract
and rhizome knot extract of edible lotus (Nelumbo nuficera). Food
Chem. 2002, 76, 327–333.
(16) Roberts, W. G.; Gordon, M. H. Determination of the total
antioxidant activity of fruits and vegetables by a liposome assay.
J. Agric. Food Chem. 2003, 51, 1486–1493.
(17) Peyrat-Maillard, M. N.; Cuvelier, M. E.; Berset, C. Antioxidant
activity of phenolic compounds in 2,2′-azobis(2-amidinopropane)
dihydrochloride (AAPH)-induced oxidation: synergistic and an-
tagonistic effects. J. Am. Oil Chem. Soc. 2003, 80, 1007–1012.
(18) Baumhof, P.; Mazitschek, R.; Giannis, A. A mild and effective
method for the transesterification of carboxylic acid esters. Angew.
Chem. Int. Ed. 2001, 40, 3672–3674.
(19) Jackson, R. F. W.; Wishart, N.; Wood, A.; James, K.; Wythes,
M. J. Preparation of enantiomerically pure protected 4-oxo-R-
amino acids from serine. J. Org. Chem. 1992, 57, 3397–3404.
(20) Yu, L.; Deems, R. A.; Hajdu, J.; Dennis, E. A. The interaction of
phospholipase A2 with phospholipid analogues and inhibitors.
J. Biol. Chem. 1990, 265, 2657–2664.
(21) Moulin, C.; Duflos, M.; Le Baut, G.; Grimaud, N.; Renard, P.;
Caignard, D.-H. Synthesis and anti-inflammatory activity of
N-(aza)arylcarboxamides derived from Trolox. Eur. J. Med. Chem.
1998, 33, 321–329.
(23) Lamidey, A.-M.; Fernon, L.; Pouyse´gu, L.; Delattre, C.; Quideau,
S.; Pardon, P. A convenient synthesis of the echinacea-derived
immunostimulator and HIV-1 integrase inhibitor (-)-(2R,3R)-
chicoric acid. HelV. Chim. Acta 2002, 85, 2328–2334.
(24) Goetz, G.; Fkyerat, A.; Metais, N.; Kunz, M.; Tabacchi, R.; Pezet,
R.; Pont, V. Resistance factors to grey mould in grape berries:
identification of some phenolic inhibitors of Botrytis cinerea
stilbene oxidase. Phytochemistry 1999, 52, 759–767.
(25) Kiso, Y.; Ukawa, K.; Nakamura, S.; Ito, K.; Akita, T. Efficient
removal of protecting groups by a push-pull mechanism. 2.
Deprotection of O-benzyltyrosine with a thioanisole-trifluoroacetic
acid system without O-to-C rearrangements. Chem. Pharm. Bull.
1980, 28, 673–676.
(26) Moon, J.-H.; Terao, J. Antioxidant activity of caffeic acid and
dihydrocaffeic acid in lard and human low-density lipoprotein. J.
Agric. Food Chem. 1998, 46, 5062–5065.
(27) Silva, F. A. M.; Borges, F.; Ferreira, M. A. Effects of phenolic
propyl esters on the oxidative stability of refined sunflower oil.
J. Agric. Food Chem. 2001, 49, 3936–3941.
(28) Calatayud, P.-A.; Auger, J.; Thibout, E.; Rousset, S.; Caicedo,
A. M.; Calatayud, S.; Buschmann, H.; Guillaud, J.; Mandon, N.;
Belloti, A. C. Identification and synthesis of a kairomone
mediating host location by two parasitoid species of the cassava
mealybug Phenacoccus herreni. J. Chem. Ecol. 2001, 27, 2203–
2217.
(29) Silva, F. A. M.; Borges, F.; Guimara˜es, C.; Lima, J. L. F. C.;
Matos, C.; Reis, S. Phenolic acids and derivatives: studies on the
relationship among structure, radical scavenging activity, and
physicochemical parameters. J. Agric. Food Chem. 2000, 48,
2122–2126.
(30) Sørensen, A.-D. M.; Haahr, A.-M.; Becker, E. M.; Skibsted, L. H.;
Bergenståhl, B.; Nilsson, L.; Jacobsen, C. Interactions between
iron, phenolic compounds, emulsifiers, and pH in omega-3-
enriched oil-in-water emulsions. J. Agric. Food Chem. 2008, 56,
1740–1750.
Received for review August 11, 2008. Revised manuscript received
August 12, 2008. Accepted August 12, 2008. This work was supported
by The Danish Agency for Science, Technology, and Innovation.
(22) Leschot, A.; Tapia, R. A.; Eyzaguirre, J. Efficient synthesis of
4-methylumbelliferyl dihydroferulate. Synth. Commun. 2002, 32,
3219–3223.
JF8024826