Differential Partitioning of Bioantioxidants in Edible Oil-Water and Octanol-Water Systems: Linear Free Energy Relationships

Article abstract of DOI:10.1021/acs.jced.8b00258

Partition coefficients, P_W^O, of antioxidants (AOs) between edible oils and water are scarce in the literature, despite that AOs are widely used to control lipid oxidation and the oxidative stress in cells. P_W^O values have a great importance to predict the efficiency and distribution of bioactives at different levels of biological organization from binary oil-water systems to living cells. Here, we determined the partition coefficients, P_W^O, of four series of potent, natural AOs of increasing lipophilicity between vegetable oils (olive, soybean, and corn) and water and, for the sake of comparison, between octanol-water, P_W^OCT. Results indicate that the contribution of the -CH₂ groups to the overall lipophilicity of the AO is the same irrespective of the oil employed. The P_W^O values were compared with P_W^OCT values, and linear relationships were obtained for each series of AOs. Results indicate that, in general, P_W^O values cannot be predicted from the P_W^OCT values, making necessary to determine the P_W^O values for each oil and antioxidant.

Full text of DOI:10.1021/acs.jced.8b00258

Article
pubs.acs.org/jced
Cite This: J. Chem. Eng. Data XXXX, XXX, XXX−XXX
Differential Partitioning of Bioantioxidants in Edible Oil−Water and
Octanol−Water Systems: Linear Free Energy Relationships
†
,†,‡
J. Freiría-Gandara, S. Losada-Barreiro,* F. Paiva-Martins,‡ and C. Bravo-Díaz†
́
†Fac. Química, Departamento Química Física, Universidad de Vigo, 36310 Vigo, Spain
‡
̂
REQUIMTE-LAQV, Departamento de Química e Bioquímica, Faculdade de Ciencias, Universidade do Porto, 4169-007 Porto,
Portugal
ABSTRACT: Partition coefficients, PWO, of antioxidants (AOs) between edible oils and
water are scarce in the literature, despite that AOs are widely used to control lipid
O
oxidation and the oxidative stress in cells. PW values have a great importance to predict
the efficiency and distribution of bioactives at different levels of biological organization
from binary oil−water systems to living cells. Here, we determined the partition
coefficients, PWO, of four series of potent, natural AOs of increasing lipophilicity between
vegetable oils (olive, soybean, and corn) and water and, for the sake of comparison,
between octanol−water, PWOCT. Results indicate that the contribution of the −CH2 groups
O
to the overall lipophilicity of the AO is the same irrespective of the oil employed. The PW
values were compared with PWOCT values, and linear relationships were obtained for each
O
series of AOs. Results indicate that, in general, PW values cannot be predicted from the
PWOCT values, making necessary to determine the PWO values for each oil and antioxidant.
O
coefficients between edible oils (e.g., olive) and water, Pw
,
INTRODUCTION
■
correlate pretty well with the partitioning of drugs into rats and
human tissues and, in some instances, they are even superior
for the prediction of lipid tissue plasma partition coefficients
over the widely used octanol−water partition coefficient,
Antioxidants (AOs) present in the human body can be either
biosynthesized or obtained from the diet.1,2 They play an
important role in controlling the oxidative stress of cells,
helping to minimize the production of harmful radicals, and
the intake of antioxidants is recommended because they also
have well-recognized beneficial health effects, including
anticancer, antitumor, and anti-inflammatory properties.3−5
AOs are also added to lipid-based foods to control the
oxidative stability of the lipids.6 Most AOs are polyphenolic in
nature and inhibit the oxidation of lipids because they are
excellent hydrogen donors that are accepted by reactive
radicals to yield less reactive radical and nonradical species.1,2
Knowledge on how AOs partition in different systems and
the effects of lipophilicity of AOs on their distribution in binary
oil−water systems is important in relevant economic areas such
as medicine, pharmacy, and food technology.7,8 For example,
the preparation of olive oil requires the use of large amounts of
water which are subsequently discarded, leading to a significant
loss of important endogenous antioxidants present in the olive
fruit.9,10 Olive oil formulations are prime examples of lipid-
based delivery systems such as emulsions and nanoemul-
sions,11,12 and bio-oils are receiving attention as green solvents
for liquid−liquid extractions.13
15−17
OCT
PW
.
Note that partition coefficient values are
determined by all intermolecular interactions, including
electrostatic, hydrogen bonding, and dispersion forces between
the solute and the two phases in which it is dissolved. Its value
thus depends on the balance of all intermolecular forces
involving a solute and the two phases between which it
partitions.7,18,19
The differential partitioning of AOs makes them to be
present at a wide range of concentrations in body fluids,
tissues, and in the different regions of lipid-based emulsions,
affecting largely their efficiency in the system.20,21 In general,
water-soluble AOs react with oxidants in the cell cytoplasm,
the blood plasma, or in the aqueous and interfacial regions of
lipid-based emulsions; meanwhile, lipid-soluble antioxidants
are oil-soluble and protect the lipids in the cell membranes or
in the food emulsions because they are mostly located in the
interfacial or oil regions of the system.20−22
In this work, we determined the distribution in olive−,
soybean−, and corn oil−water binary systems of series of
homologous AOs of different hydrophobicity derived from
potent, natural AOs. For the purpose, we grafted gallic (GA),
caffeic (CA), and protochateuic (PT) acids and hydroxytyrosol
(HT) with alkyl chains of increasing number of carbon atoms
Partitioning of AOs in multiphasic systems such as living
cells is important for understanding their efficiency in
inhibiting lipid oxidation and their bioavailability in the
human body.1,14 The partition coefficient values between
octanol and water, PWOCT, are widely use in predictive
environmental and pharmaceutical studies8 and are important
for a rational description of the partitioning behavior in
multiphasic biological systems. However, the partition
Received: March 29, 2018
Accepted: July 3, 2018
© XXXX American Chemical Society
A
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
Table 1. Composition (Given As Percentage) of the Main Fatty Acids Found in the Oils Employed in This Work
palmitic (C16:0)
palmitoleic (C16:1)
octadecanoic (C18:0)
oleic (C18:1)
linoleic (18:2)
linolenic (18:3)
arachidonic (C20:4)
corn
olive
soybean
12.6
13.4
9.0
0.0
0.9
0.0
2.3
2.4
3.7
30.2
72.1
27.6
53.4
2.9
53.7
0.8
3.7
5.1
0.6
0.5
0.0
Table 2. Partition Coefficients Values for the Distributions of AOs of Different Hydrophobicities in the Oil−Water and
Octanol−Water Mixtures at T =25°C
(1−8) so that four series of homologous AOs bearing the same
reactive group but of different hydrophobicities were prepared.
Data on partitioning of bioantioxidants in vegetable oils, and
particularly in olive oil, are scarce and much more sparse than
partitioning data in octanol−water systems.23 Determining the
partition coefficients and the distribution of AOs in oil−water
relationship can be established between both parameters.
OCT
PW
values can be determined by a variety of exper-
imental24,25 and theoretical methods, including ACD/log D
software, ADMET predictors etc., which provide PWOCT values
very close to the experimental ones.7,26 We employed the free
(online) computer software Molinspiration27,30 to determine
the PWOCT values. The software estimates PWOCT values on the
basis of the group contribution concept. This method assumes
that each molecule is considered a mixture of simple groups
(−CH2, −COOH, −OH, etc.) whose thermodynamic
property parameters are calculated by the summation of the
contributions of the various groups.24
O
systems may be of general interest because analyses of the Pw
values contribute to the general understanding of the
interactions of the molecules in the different oil phases. All
the oils employed here are premium functional foods of
enormous economic and social importance. Particularly, olive
oil is important in both the Mediterranean region and in the
healthy Mediterranean “lifestyle” because of its well-balanced
fatty acid composition/profile with oleic acid as the main
component, Table 1, and because of the presence of minor
compounds such as polar phenols, polyphenols, tocopherols,
and sterols. The composition of the oils employed in this work
is given in Table 1, and the chemical structures of the AOs are
shown in Table 2.
Moreover, some of the parent AOs employed here are
phenolic acids that may be partially ionized at the typical pH
(4−6) range of foods,6 affecting their partition coefficients.
O
Thus, we analyzed the effects of acidity on the Pw values of
two representative phenolic acids, GA and CA, and of ascorbic
acid, representative of nonphenolic antioxidants.
EXPERIMENTAL SECTION
It may be particularly important for predictive purposes to
■
OCT
investigate if there is any relationship between PwO and PW
values. Thus, we also determined the PW
Materials. All chemicals were of the highest purity available
and used as received. Aqueous buffered (citric acid/citrate
sodium; 0.04M; pH 2−6) solutions were prepared by using
OCT
values for the
same series of AOs and explored if any extrathermodynamic
B
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
Figure 1. (A) Spectra of aqueous solutions of GA in aqueous solutions at different pH values. (B) Calibration curves of CA in BuOH/EtOH
mixtures obtained from independent experiments carried out on different days. The slopes are essentially the same (differences less than 7%). (C)
Standard addition curves obtained from aliquots (40 μL) of the oil phase of oil/acidic water mixtures obtained after phase separation by
centrifugation.
O
●
Figure 2. Variations of the natural log(PW ) with the number of −CH2 groups in the alkyl chain ( ) for AO ester derivatives of gallic (A), caffeic
OCT
(B), and protocatechuic (C) acids and hydroxytyrosol (D). For the sake of comparisons, the partition coefficients PW
between octanol and
■
water ( ) are also displayed.
Mili-Q grade (conductivity <0.1 mS·cm−1). Commercial
antioxidants were purchased from Sigma-Aldrich or Acros.
Gallic acid (GA), methyl gallate (C1), ethyl gallate (C2),
propyl gallate (C3), butyl gallate (C4), octyl gallate (C8), and
lauryl gallate (C12) were purchased from Sigma-Aldrich and
used as received. Caffeates were synthesized from manomal-
onates through Verley−Doebner modification of Knoevenagel
condensation with the 3,4-dihydroxybenzaldehyde and 4-
dihydroxybenzaldehyde.28 Protocatechuates (C1−C16) were
synthesized by chemical acylation of the carboxylic group
C
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
following literature procedures.14 The hydroxytyrosol esters
were synthesized by direct esterification of HT with the
corresponding fatty acid (C1−C6 derivatives) or by trans-
esterification with the corresponding fatty ester (C8−C16) in
the presence of Novozym 435 according to published
procedures.29 Final products were purified by chromatography
with silica gel 60 (0.040−0.063 mm, Merck). Nuclear magnetic
resonance (NMR) spectra were recorded on 100 or 400 MHz
adding increasing amounts of a AO stock solution ([AO] = 6.8
× 10−3 M) to EtOH/BuOH mixtures containing 40 μL of oil
(added to minimize a potential solvent effect). Figure 2B
shows two of the typical calibration curves obtained. The
average values of molar extinction coefficient for CA is εCA
=
13 700 88 M−1 cm−1.
B. Standard Addition Method. After the centrifugation of
the binary oil/water mixtures (different pH and prepared with
AO), an aliquot (40 μL) of the oil phase was removed and
diluted in a 50:50 (v/v) EtOH/BuOH mixture. Addition
standard curves were prepared by adding increasing volumes of
a CA stock solution ([CA] = 6.8 × 10−3 M) as illustrated in
Figure 2C, from where %CAO was calculated.
1
NMR spectrometers, and the H and 13C NMR spectra of the
synthesized AOs (in CDCl3) were in accordance with the
literature reports.
O
Methods. The partition coefficients PW (at constant
acidity) in binary olive oil−water systems were determined by
employing the shake-flask method.30 Hydrophilic AOs were
dissolved in water, while the hydrophobic ones were dissolved
directly in the oil. 1:9 or 4:6 (oil:water, v/v) mixtures (VT = 10
mL) were prepared, stirred with a high-speed rotor for 1 min
and allowed stand for at least 4 h to permit phase separation
and to reach thermal equilibrium (T = 25 °C). The phases
were separated by centrifugation, and the AO concentrations
in the aqueous and oil phases were determined by UV
spectrometry with the aid of a previously prepared calibration
curves. The calibration curves were obtained by measuring the
absorbance of increasing amounts of the antioxidants ([AO] =
10−4 − 10−2 M) both in the water and in the oil phase of the
binary mixtures.
Figure 1A is representative and show the spectrum of
aqueous solutions of GA at different pH values . The observed
spectral changes are due to the ionization of the carboxylic
group because the first pKa of the −COOH groups of phenolic
acids range 4−5, meanwhile that of the aromatic −OH groups
range 8.3−8.7.31,32 We note that the −OH groups of the alkyl
ester derivatives do not significantly ionize at the typical
physiological food acidities (2−7).
Calibration curves were obtained at selected pH values by
measuring the absorbance of increasing amounts of the
antioxidants. Vegetable oils absorb strongly in the same UV
region as that of AOs, making the determination of the
percentage of some AOs in this region difficult if not
impossible. In some cases (for example, caffeic acid), the
maximum in the absorbance is at higher wavelengths than that
of the components of the oils, and in these cases, the fraction
of CA in the oil, %CAO, could be determined directly in the oil
phase after dilution in a BuOH/EtOH mixture. However, after
dilution, the measured absorbance values are very low, and this
creates some uncertainty in %CAO values. To minimize
inasmuch as possible this problem, %AOO was determined
by employing two complementary methods, (a) interpolation
in a calibration curve and (b) a standard addition method.
Both gave similar results, yielding values of the percentage of
AO in the oil phase with differences less than 5%. Figures 1B
and 1C are illustrative and show the calibration curves
obtained at selected pH. Values of the intercepts and the
slopes (molar extinction coefficients, ε), not shown here, were
obtained by least-squares fitting of the (A, [AO]) pairs of data
obtained under the different experimental conditions. In all
cases, square correlation coefficients, r2, were higher than r2 =
0.999.
For those AOs with low oil solubility, the fraction of AO in
the oil region (%AOo) could not be determined by
interpolation in a calibration curve (as the percentage in the
aqueous phase, %AOW) because of the high absorption of the
oils. In these cases, %AOO values were estimated by difference
between the total added amount of AO to the determined
amount in water, i.e., %AOO = 100 − %AOW. On the contrary,
for the most hydrophobic AOs (with alkyl chains of eight or
more carbon atoms), the fraction of AO in the aqueous region
could not be determined accurately because of their extremely
low solubility in water. For example, the solubility of OG and
LG have been reported to be less than 0.001 g/L.33,34 For
these AOs, the reported partition coefficients need to be taken
with extreme caution because values of (AOW) in the oil/water
mixture will also be very small, and small errors in the
denominator of eq 1 may have a large effect on the value of
O
PW of the large errors in their determinations.
The partition coefficients PWO (at constant acidity) in binary
oil−water systems were determined by employing a shake-flask
method30 and by employing eq 1, where (hereafter),
parentheses, (), indicate molar concentrations in moles per
liter of a particular region’s volume, and VW and VO are the
volumes of the aqueous (W) and oil (O) regions, respectively.
(AOO)
(AOW )
%AO
%AOW
VW
VO
o
PWO
=
=
×
(1)
When investigating the effects of acidity, the apparent
partition coefficients PWO(app) for gallic, caffeic, and ascorbic
acids, chosen as representatives of phenolic and nonphenolic
acids, were determined at different acidities by employing the
same shake-flask method. The percentages of GA in the
aqueous phase, %AOW were determined by interpolation of the
absorbance of aliquots (30 μL) of the aqueous phase of the
binary olive oil−water mixtures by using previously prepared
calibration curves at each pH as described elsewhere,35,36 and
the PWO(app) values were calculated by employing eq 2.
O
Reported PW and PWO(app) values in Table 2 were obtained
in duplicate or triplicate with deviations less than 5%.
(AOO)
(AOW ) + (AOW− )
PWO(app) =
(2)
Statistical Analysis. Reported partition coefficients in
binary oil−water systems were the average of 6−8 replicates.
The Dixon’s Q-test was employed in deciding whether to
accept or reject the datum before calculating the average of the
set of replicates. Data are presented as means confidential
interval. Square correlation coefficients, r2, of linear relation-
ships were in all cases r2 > 0.999.
A. Interpolation in a Calibration Curve. The absorbance of
an aliquot (40 μL) .of the oil fraction of the binary oil/acidic
water mixtures prepared without antioxidant was measured
after dilution in a 50:50 (v/v) EtOH/BuOH mixture.
Calibration curves for the AOs were prepared in triplicate by
D
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
Theory. On thermodynamic grounds, the transfer of a
solute from an aqueous (W) to an organic (oil, O) phase (both
in equilibrium) can be described by the change in the Gibbs
free energy, ΔGW→O, that accompanies the transfer, providing
a quantitative measurement of the relative hydrophobicity of
the phases.37 At a constant temperature T and pressure P,
ΔG0,W→O is given by the difference in the chemical potentials
of the compound i being transferred, ΔG0,W→O = μiW − μiO,
where μi is the chemical potential of the component i. At
equilibrium, ΔG0,W→O = 0, and the chemical potentials equal
each other. The chemical potentials depend on the activities of
the solute, and at the low concentration limit, the activity
coefficients can be assumed to be close to the unit. Therefore,
it is safe to use concentrations instead of activities, and eq 3
can be derived. Eq 3 relates the Gibbs free energy of transfer of
the component i from one phase to another with its partition
coefficient between both phases. A deeper explanation on the
assumptions and equations involved can be found else-
where.19,37,38
Scheme 1. Oil−Water Biphasic System (Employed in This
Work) Showing the Equilibrium Contribution of an
a
Ionizable Antioxidant AO, Its Conjugated Base, AO−
a
O
Pw is the partition coefficient, and KaW is the aqueous ionization
constant of the AO. Ionization is assumed to take place exclusively in
the aqueous region, and the formation of potential ion pairs is
neglected.
Scheme 2. Distribution of an Antioxidant AO between the
Oil, Interfacial, and Aqueous is Described by the Partition
I
Coefficients between the Aqueous−Interfacial, PW , and
a
I
Oil−Interfacial, PO , Regions of the Emulsion
ΔG0,W→0 = −RTln PWO
(3)
Contributions to partition coefficient values can be
factorized into various parameters, including the molar volume,
which comprises hydrophobic and dispersion forces, solvatoc-
romic parameters, defined as the solute’s H-bond donor
acidity, H-bond acceptor basicity, and polarizability, which
embraces the orientation and induction forces.7,24 Thus,
studies of the partitioning of solutes between water and
nonpolar solvents provide estimates for the energy cost of
exposing nonpolar molecules to water creating hydrophobic
water contacts.19,23
When low solute concentrations are employed, the
distribution of the AOs is described by the partition coefficient
PWO defined by eq 1. Hereafter, parentheses, (), indicate molar
concentrations in moles per liter of a particular region’s
volume, and VW and VO are the volumes of the aqueous (W)
and oil (O) regions, respectively, meanwhile stoichiometric
concentrations in moles per liter of total volume of the oil−
water mixture will be indicated by the commonly employed
brackets [ ]. If the AO is partially ionized under the
experimental conditions (e.g., a phenolic acid), then an
apparent partition coefficient, PWO(app), eq 2, may be defined
in terms of all ionized and neutral forms present at a particular
pH.39 The ionization of weak acids in the oil phase is negligible
because literature reports indicate that the ionization constants
of weak acids in oils are typically 5−6 orders of magnitude
smaller than those in the aqueous phase,32 (i.e., pKa(O) ≫
pKa(W)). Thus, the distribution of the AO can be illustrated as
in Scheme 1, and the apparent (measured) partition coefficient
a
For the sake of clarity, the potential ionization of the AO in the
aqueous and interfacial regions is not shown.
to determine the partition coefficients POI and PwI needs to be
done in the intact emulsions.40,41 We note that the ratio
between POI and PwI is equal to the partition coefficient in the
absence of surfactant, PWO, eq 6, highlighting the importance of
determining partition coefficients in binary systems. Therefore,
determining partition coefficients in binary systems provides
useful information on the behavior of chemicals of interest in
both binary oil−water systems and in food emulsions.
(AOI)
(AOW )
PWI
=
(4)
(5)
(AOI)
(AOO)
POI =
is defined by eq 2. Note that at the low pH limit, the
PWI
POI
(AOI)/(AOW )
(AOI)/(AOO)
(AOO)
(AOW )
O
antioxidant is completely neutral, and PWO(app) = PW
.
=
=
= PWO
Importance of Determining PWO Values in Oil−Water
Systems to Estimate the Distribution of AOs in
Multiphasic Systems. Partitioning in multiphasic systems is
more complex and requires definition of new distribution
constants.40,41 For example, addition of surfactants to oil−
water binary systems to create kinetically stable emulsions
makes the antioxidant to partition between the oil, water, and
interfacial regions, and their distribution is now described by
two partition coefficients, Scheme 2, that between the oil-
(6)
RESULTS AND DISCUSSION
■
Table 2 shows the values of the partition coefficients for the
distributions of AOs in the oil−water binary mixtures and, for
the sake of comparisons, in octanol−water. To a first
approximation, each of the homologous series of AOs
employed have two distinct regions within their molecules
that influence their partitioning: an aromatic residue, which is
common to all individuals of each series, and an alkyl ester tail
containing a variable number of methylene groups and the
I
I
interfacial, PO , and aqueous-interfacial, Pw , regions, eqs 4 and
5. Because of the physical impossibility of separating the
interfacial region from the oil and aqueous ones, any attempt
E
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
Table 3. Slopes and Intercepts for the Variations of log(PWO) with the Number of Methylene Groups According to Eq 7
oil
gallic
caffeic
protocatechuic
hydroxytyrosol
octanol (OCT)
aOCT
bOCT
aOL
bOL
aSO
bSO
aCO
bCO
0.77 0.03
0.51 0.01
−1.15 0.04
0.50 0.01
−1.18 0.06
0.58 0.03
−1.03 0.10
0.54 0.06
1.53 0.05
0.49 0.01
0.40 0.03
0.51 0.01
0.44 0.03
0.52 0.03
0.52 0.01
0.53 0.01
1.06 0.03
0.50 0.01
−0.18 0.07
0.55 0.04
1.15 0.04
0.50 0.01
−0.16 0.02
0.44 0.02
olive (OL)
soybean (SO)
corn (CO)
O
■
Figure 3. Variations of log(PW ) with the number of −CH2 groups in the alkyl chain derivatives of caffeic (A) and gallic (B) acids in olive ( ),
43
●
▲
̈
soybean ( ), and corn ( ) oils. Data for the partition coefficients of gallic derivatives in corn oil were obtained from Stockman et al.
terminal methyl group. The aromatic moiety and the methyl
group make a constant contribution.
value for the slope is quite similar to the average value obtained
in octanol, boct = 0.50 0.01, suggesting that the contribution
of the methylene group to the hydrophobicity of the AOs is
very similar in both solvents.
However, the intercept values are different for each series as
a consequence of the different structure of the aromatic
Using the basic group-contribution concept, the logarithm of
O
the partition coefficient log PW can be computed as the
summation of two contributions, eq 7, that of the alkyl part of
the molecule and that of the aromatic and methyl residues. In
eq 7, usually called the Colander equation,42 nCH is the
moieties of the AOs employed, following the order aGA < aPT
≈
2
aHT < aCA. These aO values cannot be compared directly to
each other because the nonalkyl moieties contain different
groups and refer to different oils with different compositions.
However, for the same oil (olive), their relative values reflect
the big effect of the addition of an −OH group to their
chemical structure on the overall hydrophobicity of the
nonalkyl portion of the AOs (compare, for instance, GA and
PT) and also the big effect of the −OH group on the
hydrophobicity of the molecule compared to the addition of a
single (HT) or double bond (CA) to the alkyl chain.
number of methylene groups in the AO, and the parameter a
stands for the contribution of the nonalkyl part of the molecule
to PWO. Eq 7 predicts that an increase in the number of C
atoms in the alkyl chain increases the hydrophobicity of the
AOs and its solubility in oil.
log(PWO) = aO + bOn
CH2
(7)
Figure 2 shows that plots of log(PWO) (olive and octanol) vs
the number of −CH2 groups for the esters derived of gallic
(2A), caffeic (2B), and protochateuic (2C) acids and of
hydroxytyrosol (2D) are linear except for the most hydro-
phobic AOs (n > 7), where deviations from the linearity are
evident. Deviations from linearity are probably caused by the
uncertainty in their values, affected by experimental errors in
the determination of their concentration or percentages in the
water phase. These values were not considered in determining
the intercepts and slopes collected in Table 3. For the sake of
clarity, the variations of log(PWO) with the number of −CH2
groups for the caffeic and gallic esters in olive, corn, and
soybean oils are shown in Figure 3.
Because PWOCT is a key physicochemical parameter for drug
OCT
discovery, design, and development, databases with PW
values are available for several thousands of compounds in the
literature (for example, logKoW, PHYSPROP, etc.). If not
available for a given species, the values can be easily calculated
with the aid of various computer programs that employ
different descriptors. Thus, for predictive purposes, it may be
useful to convert the partition coefficients from the octanol−
water system to the olive oil-water system. This can be easily
done bearing in mind that partition coefficients are equilibrium
constants, and therefore, extrathermodynamic relationships
between values in different solvents can be established.44,45
Table 3 shows that the slopes b obtained are very similar to
each other (with differences less than 10%, except for HT in
olive, which is 20%) regardless of the oil or the AO series, with
an average value of bO = 0.53 0.02. Remarkably, this average
O
OCT
The linearity in the variations of both PW and PW
with
the number of CH2 atoms in the alkyl chain of each series of
F
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
AOs, Figures 2A−D, with very similar slopes suggests that
general, the values of the partition coefficients of a given
antioxidant in vegetable oil−water systems cannot be predicted
from the values of the partition coefficient in octanol−water.
pH Effects on the Partition Coefficients between
Olive Oil and Water. Phenolic and nonphenolic AOs may be
partially ionized under the typical pH of foods, and their
degree of ionization depends, therefore, on the particular
experimental conditions (acidity) employed. Because the
neutral and ionic species exhibit different polarities, their
partition coefficients are pH dependent, and an apparent
partition coefficient, PWO(app), may be defined in terms of all
ionized and neutral forms present at a particular pH. Ionization
constants of phenolic and nonphenolic AOs range 4−9, but the
typical pH of foods is 4−7. The pKa of aromatic −OH groups
is around 8−9, and that of phenolic and nonphenolic acids is
around 4−5. Thus, as a first approach and for the sake of
simplicity, we consider only those AOs that can be ionized in
the pH 4−6 range, e.g., phenolic and nonphenolic acids. The
mathematical treatment can be easily extended to other
situations, for example, if the aqueous pH is higher than 7, then
the second ionization equilibria should be eventually
considered.
Literature reports indicate that the ionization constants of
weak acids in oils are typically 5−6 orders of magnitude
smaller than those in the aqueous phase,32 (i.e., pKa(O) ≫
pKa(W)). Thus, one can safely assume that the ionization of
weak acids in the oil phase is negligible, and the distribution of
the AO can be illustrated as in Scheme 2. The apparent
(measured) partition coefficient PWO(app) is defined by eq 1,
which assumes that activity effects are negligible, i.e., the
activity coefficients are very close to the unit. If the ionic
species may be formed in the oil phase, a more complex
approach is required.39 Note that eq 9 predicts that at the low
pH limit, the antioxidant is completely neutral, and PWO(app)
= PWO, eq 1.
there must be a linear relationship such as that given by eq 8
O
between PW and PWOCT. Figure 4 shows such linear
O
Figure 4. Extrathermodynamic linear relationships between log(PW
and log(PWOCT) for homologous series of AOs.
)
relationship. The values for the intercept and slope were
obtained by fitting the corresponding (PWO, PWOCT) pairs of
data to eq 8 and are given in Table 4.
log(PWO) = A + Blog(PWOCT
)
(8)
Table 4. Values of the Slopes and Intercepts of the Linear
Plots in Figure 4
antioxidant series
A
B
GA
PT
CA
HT
−2.56 0.16
−1.45 0.04
−1.16 0.18
−2.60 0.09
1.22 0.10
1.18 0.02
1.04 0.08
0.84 0.03
Bearing in mind the ionization of the AO in the aqueous
O
phase, eq 9, the relationship between PwO(app) and Pw is
Both A and B values are different for each series of AOs but
are in line with those reported by Leo et al.45 when analyzing
the distribution of a number of compounds between vegetable
oils and octanol. Because the A and B values are different, the
use of linear relationships such as that indicated by eq 8 seems
to be limited to series of the same AOs, suggesting that, in
given by eq 10:
(AO‐W )(H+)
Ka =
(AOW )
(9)
O
■
●
▲
Figure 5. (A) Variation of PW (app) with pH for GA( ), AA ( ), and CA ( ) for olive oil in water mixtures. The solid line is the theoretical
curve obtained by fitting the experimental data to a sigmoidal Henderson−Hasselbach type equation. (B) Linear plots of 1/PWO(app) with pH
according to eq 11. (C) Theoretical variations of PW (app) in octanol−water mixtures with pH for GA( ), AA ( ), and CA ( ).
O
■
●
▲
G
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
PWO
AUTHOR INFORMATION
Corresponding Author
*E-mail: sonia@uvigo.es.
ORCID
C. Bravo-Díaz: 0000-0002-9468-0881
Funding
PWO(app) =
■
Ka
1 +
[H+]
(10)
(11)
Ka
PWO
1
1
1
=
+
[H+]
PWO(app)
PWO
Financial support of the following institutions is also
Eq 10 predicts a sigmoidal variation of the apparent
(measured) partition coefficient value with acidity, Figure
5A. When pH is, at least, two pH units below the pKa of the
ionizable AO (i.e., [H+] ≫ Ka,) PWO(app) = PWO, and
therefore, there is an upper limit in the experimentally
measured partition coefficients which corresponds to that of
the neutral molecule. The reverse of eq 10 is eq 11, which
predicts a linear variation of 1/PWO(app) with 1/[H+], Figure
acknowledged: FEDER (COMPETE) and FCTFundaca
̧ o
̃
̂
para a Ciencia e a Tecnologia (UID/QUI/50006/2013 POCI-
01-0145-FEDER-007265), Red de Uso Sostenible de los
Recursos Naturales y Agroalimentarios (REDUSO, Xunta de
Galicia, Grant number ED431D 2017/18), and University of
Vigo and Porto. S.L.-B. thanks Xunta de Galicia for a
postdoctoral grant (POS-B/2016/012).
Notes
O
O
5B, from where value of PW = 0.096 0.009 (GA), PW
=
O
The authors declare no competing financial interest.
0.090 0.004 (CA), and PW = 0.110 0.002 (AA) can be
O
determined for the partition coefficient PW of the totally
REFERENCES
neutral AOs. The value for CA is very similar to a literature
■
value reported by Rodis et al.,9 PW = 0.089. For the sake of
O
(1) Losada-Barreiro, S.; Bravo-Díaz, C. Free radicals and
polyphenols: the redox chemistry of neurodegenerative diseases.
Eur. J. Med. Chem. 2017, 133, 379−402.
(2) Shahidi, F. Handbook of Antioxidants for Food Preservation, 1st
ed.; Woodhead Pub.: 2015.
OCT
comparisons, sigmoidal variation of the PW
values with
acidity is illustrated in Figure 5C.
The variation of PWO(app) with pH reflects the difference in
lipophilicity between the neutral and ionic species. We note
(3) Rajendran, P.; Nandakumar, N.; Rengarajan, T.; Palaniswami, R.;
Gnanadhas, E. N.; Lakshminarasaiah, U.; Gopas, J.; Nishigaki, I.
Antioxidants and human diseases. Clin. Chim. Acta 2014, 436, 332−
347.
O
that the actual values of both PW and PWO(app) are
determined by all intermolecular interactions which may
include electrostatic, hydrogen bonding, and dispersion forces
between the solute and the two phases in which it is dissolved.
Thus, the difference between PWO(app) (partially or totally
́
(4) Menaa, F.; Menaa, A.; Treton, J., Chapter 63: Polyphenols
against Skin Aging. In Polyphenols in Human Health and Disease;
Academic Press: San Diego, 2014; pp 819−830.
O
ionized substrate) and PW (totally neutral substrate) is a
́
(5) Morales-Gonzalez, J. A. Oxidative Stress and Chronic Degenerative
parameter containing important information on the intermo-
lecular forces and intramolecular interactions of a given solute.
Diseases: A Role for Antioxidants; InTech: 2013.
(6) Frankel, E. N. Lipid Oxidation; The Oily Press, PJ Barnes &
Associates: Bridgwater, England, 2005.
CONCLUSIONS
(7) Leo, A. J., Octanol/Water Partition Coefficients. In Encyclopedia
of Computational Chemistry; John Wiley & Sons, Ltd: 2002.
(8) Korinth, G.; Wellner, T.; Schaller, K. H.; Drexler, H. Potential of
the octanol−water partition coefficient (logP) to predict the dermal
penetration behaviour of amphiphilic compounds in aqueous
solutions. Toxicol. Lett. 2012, 215 (1), 49−53.
(9) Rodis, P. S.; Karathanos, V. T.; Mantzavinou, A. Partitioning of
olive oil antioxidants between oil and water phases. J. Agric. Food
Chem. 2002, 50, 596−601.
(10) De Leonardis, A. Virgin Olive Oil: Production, Composition, Uses
and Benefits for Man; Nova Science Publishers, Incorporated: 2014.
(11) McClements, D. J. Enhanced delivery of lipophilic bioactives
using emulsions: a review of major factors affecting vitamin,
nutraceutical, and lipid bioaccessibility. Food Funct. 2018, 9 (1),
22−41.
(12) Kumar, K. S.; Dhachinamoorthi, D.; Saravanan, R.; Gopal, U.;
Shanmugam, V. Microemulsions as carriers for novel drug delivery: A
review. Int. J. Pharm. Sci. Rev. Res. 2011, 10, 37−45.
(13) Spear, S. K.; Griffin, S. T.; Granger, K. S.; Huddleston, J. G.;
Rogers, R. D. Renewable plant-based soybean oil methyl esters as
alternatives to organic solvents. Green Chem. 2007, 9 (9), 1008−1015.
(14) Silva, R.; Losada-Barreiro, S.; Paiva-Martins, F.; Bravo-Díaz, C.
Partitioning and antioxidative effect of protocatechuates in soybean
oil emulsions: Relevance of emulsifier concentration. Eur. J. Lipid Sci.
Technol. 2017, 119 (6), 1600274.
(15) Chamberlin, A. C.; Levitt, D. G.; Cramer, C. J.; Truhlar, D. G.
Modeling Free Energies of Solvation in Olive Oil. Mol. Pharmaceutics
2008, 5 (6), 1064−1079.
(16) Coe, E. L.; Coe, M. H. A hypothesis relating oil: Water
partition coefficients and vapor pressures of nonelectrolytes to their
penetration rates through biological membranes. J. Theor. Biol. 1965,
8 (2), 327−343.
■
We determined the partition coefficients for series of potent,
natural AOs of different hydrophobicity in oil−water mixtures
and, for the sake of comparison, in octanol−water systems.
The variations of the partition coefficients with the number of
C atoms in the alkyl chain were linear and parallel to those
obtained in octanol−water, indicating a similar contribution of
the −CH2 groups to the total hydrophobicity of the molecule
in both solvents. As expected, the intercept values, which
reflect the contribution of the nonalkyl portion of the
molecule, are different as a consequence of the different
groups in the chemical structures of the AOs. Extrathermody-
namic linear relationships between the partition coefficients in
olive oil−water and octanol−water systems could be
established. They are only valid for a given series of
O
compounds, suggesting therefore that, in general, PW values
cannot be predicted from PW
determined for each series of AOs employed.
OCT
values and should be
When AOs contain ionizable groups (phenolic and ascorbic
acids), the values of the partition coefficients depend on the
actual value of the acidity, decreasing upon increasing pH
following a sigmoidal-like curve. Literature reports indicate
that PWO(app) values can differ by more than one order of
magnitude, reflecting the changes in the hydrophobicity of the
AO because of its ionization. Thus, determination of their
variation is highly recommended because once the PWO values
for the neutral and fully ionized species are known, theoretical
equations can be used to calculate the change in hydro-
phobicity with pH.
H
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX

Journal of Chemical & Engineering Data
Article
(17) Chiou, C. T. Partition coefficients of organic compounds in
lipid-water systems and correlations with fish bioconcentration
factors. Environ. Sci. Technol. 1985, 19 (1), 57−62.
(18) Liu, X.; Testa, B.; Fahr, A. Lipophilicity and Its Relationship
with Passive Drug Permeation. Pharm. Res. 2011, 28 (5), 962−977.
(19) Sangster, J. Octanol-Water Partition Coefficients, Fundamentals
and Physcial Chemistry; J. Wiley & Sons: Chichester, 1997.
(37) Leo, A.; Hansch, C.; Elkins, D. Partition Coefficients And Their
Uses. Chem. Rev. 1971, 71 (6), 525−616.
́
(38) Silverio, S. C.; Rodríguez, O.; Teixeira, J. A.; Macedo, E. A.
Gibbs free energy of transfer of a methylene group on {UCON
+(sodium or potassium) phosphate salts} aqueous two-phase systems:
Hydrophobicity effects. J. Chem. Thermodyn. 2010, 42 (8), 1063−
1069.
(39) Kah, M.; Brown, C. D. LogD: Lipophilicity for ionisable
compounds. Chemosphere 2008, 72 (10), 1401−1408.
(40) Bravo-Díaz, C.; Romsted, L. S.; Liu, C.; Losada-Barreiro, S.;
́
(20) Losada-Barreiro, S.; Sanchez-Paz, V.; Bravo-Díaz, C. Transfer of
antioxidants at the interfaces of model food emulsions: distributions
and thermodynamic parameters. Org. Biomol. Chem. 2015, 13, 876−
885.
́
Pastoriza-Gallego, M. J.; Gao, X.; Gu, Q.; Krishnan, G.; Sanchez-Paz,
V.; Zhang, Y.; Ahmad-Dar, A. To Model Chemical Reactivity in
Heterogeneous Emulsions, Think Homogeneous Microemulsions.
Langmuir 2015, 31, 8961−8979.
(21) Costa, M.; Losada-Barreiro, S.; Paiva-Martins, F.; Bravo-Diaz,
C. Optimizing the efficiency of antioxidants in emulsions by
lipophilization: tuning interfacial concentrations. RSC Adv. 2016, 6
(94), 91483−91493.
(22) Costa, M.; Losada-Barreiro, S.; Paiva-Martins, F.; Bravo-Díaz,
C. Physical evidence that the variations in the efficiency of
homologous series of antioxidants in emulsions are a result of
differences in their distribution. J. Sci. Food Agric. 2017, 97 (2), 564−
571.
(23) Schwarzenbach, R. P.; Gschwend, P. M.; Imboden, D. M.
Organic Liquid-Water Partitioning. In Environmental Organic
Chemistry, 2nd Ed.; J. Wiley & Sons, Inc.: 2003.
(24) Molyneux, P. Octanol/water partition coefficients Kow: A
critical examination of the value of the methylene group contribution
to logKow for homologous series of organic compounds. Fluid Phase
Equilib. 2014, 368, 120−141.
̈
(25) Cumming, H.; Rucker, C. Octanol−Water Partition Coefficient
Measurement by a Simple 1H NMR Method. ACS Omega 2017, 2
(9), 6244−6249.
(41) Romsted, L. S.; Bravo-Díaz, C. Modelling chemical reactivity in
emulsions. Curr. Opin. Colloid Interface Sci. 2013, 18, 3−14.
(42) Beezer, A. E.; Gooch, C. A.; Hunter, W. H.; Volpe, P. L. O. A
thermodynamic analysis of the Collander equation and establishment
of a reference solvent for use in drug partitioning studies. J. Pharm.
Pharmacol. 1987, 39 (10), 774−779.
̈
(43) Stockmann, H.; Schwarz, K.; Huynh-Ba, T. The Influence of
Various Emulsifiers on the Partitioning and Antioxidant Activity of
Hydrobenzoic acids and Their Derivatives in Oil-in-Water Emulsions.
J. Am. Oil Chem. Soc. 2000, 77, 535−542.
(44) Goss, K.-U.; Schwarzenbach, R. P. Linear Free Energy
Relationships Used To Evaluate Equilibrium Partitioning of Organic
Compounds. Environ. Sci. Technol. 2001, 35 (1), 1−9.
(45) Leo, A. J.; Hansch, C. Linear free energy relations between
partitioning solvent systems. J. Org. Chem. 1971, 36 (11), 1539−1544.
(26) Klopman, G.; Zhu, H. Recent Methodologies for the
Estimation of N-Octanol/Water Partition Coefficients and their Use
in the Prediction of Membrane Transport Properties of Drugs. Mini-
Rev. Med. Chem. 2005, 5, 127−133.
(27) Molinspiration. http://www.molinspiration.com/.
(28) Costa, M.; Losada-Barreiro, S.; Paiva-Martins, F.; Bravo-Díaz,
C.; Romsted, L. S. A direct correlation between the antioxidant
efficiencies of caffeic acid and its alkyl esters and their concentrations
in the interfacial region of olive oil emulsions. The pseudophase
model interpretation of the ‘‘cut-off’’ effect. Food Chem. 2015, 175,
233−242.
(29) Almeida, J.; Losada-Barreiro, S.; Costa, M.; Paiva-Martins, F.;
Bravo-Díaz, C.; Romsted, L. S. Interfacial Concentrations of
Hydroxytyrosol and Its Lipophilic Esters in Intact Olive Oil-in-
Water Emulsions: Effects of Antioxidant Hydrophobicity, Surfactant
Concentration, and the Oil-to-Water Ratio on the Oxidative Stability
of the Emulsions. J. Agric. Food Chem. 2016, 64, 5274−5283.
(30) Berthod, A.; Carda-Broch, S. Determination of liquid-liquid
partition coefficients by separation methods. J. Chromatog. A 2004,
1037, 3−14.
(31) CRC. Handbook of Chemistry and Physics, 78th ed.; CRC Press:
Boca Raton, 1997.
́
(32) Dean, J. A. Langes Handbook of Chemistry; McGraw-Hill, Inc.:
New York, 1992.
(33) Lu, L.-L.; Lu, X.-Y. Solubilities of gallic acid and its esters in
water. J. Chem. Eng. Data 2007, 52, 37−39.
(34) Takai, E.; Hirano, A.; Shiraki, K. Effects of alkyl chain length of
gallate on self-associationand membrane binding. J. Biochem. 2011,
150 (2), 165−171.
(35) Galan, A.; Losada-Barreiro, S.; Bravo-Díaz, C. A Physicochem-
ical Study of the Effects of Acidity on the Distribution and
Antioxidant Efficiency of Trolox in Olive Oil-in-Water Emulsions.
ChemPhysChem 2016, 17, 296−304.
(36) Losada-Barreiro, S.; Bravo-Díaz, C.; Romsted, L. S.
Distributions of phenolic acid antioxidants between the interfacial
and aqueous regions of corn oil emulsions: effects of pH and
emulsifier concentration. Eur. J. Lipid Sci. Technol. 2015, 117, 1801−
1813.
I
DOI: 10.1021/acs.jced.8b00258
J. Chem. Eng. Data XXXX, XXX, XXX−XXX