Structure-Antioxidant Activity Relationships in a Series of NO-Donor Phenols

Food Chemistry 194 (2016) 128–134

Contents lists available at ScienceDirect

Food Chemistry

journal homepage: www.elsevier .com/locate / foodchem

Structure–antioxidant activity relationships of o-hydroxyl, o-methoxy,and alkyl ester derivatives of p-hydroxybenzoic acid

http://dx.doi.org/10.1016/j.foodchem.2015.08.0030308-8146/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author.E-mail address: [email protected] (R. Farhoosh).

Reza Farhoosh ⇑, Saeed Johnny, Maryam Asnaashari, Najme Molaahmadibahraseman, Ali SharifFerdowsi University of Mashhad, Faculty of Agriculture, Department of Food Science and Technology, P.O. Box 91775-1163, Mashhad, Iran

a r t i c l e i n f o

Article history:Received 7 November 2014Received in revised form 30 July 2015Accepted 1 August 2015Available online 4 August 2015

Keywords:Antioxidant activityEmulsionKilka fish oilp-Hydroxybenzoic acidPolar paradox theory

a b s t r a c t

Anti-DPPH radical effect as well as anti-peroxide activity of o-hydroxyl, o-methoxy, and alkyl esterderivatives of p-hydroxybenzoic acid in a bulk fish oil system and its O/W emulsion were investigated.Electronic phenomena, intra- and/or intermolecular hydrogen bonds, interfacial properties, and chemicalreaction of the solvent molecules with phenolic compounds were considered to be mainly involved in theantiradical activities observed. Antioxidant activity of the phenolic acids derivatives as a function of thesefactors was variously affected by the environmental conditions which may occur in practice.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Phenolic acids as a major class of bioactive phenolic compoundsare widely distributed in plant kingdom and are known to haveantioxidant activity (Frankel, 1998). Apart from other mechanismsthat may be involved, the antioxidant activity of these compoundshas been considered to be a very important cause of many biolog-ical capabilities, including anti-inflammatory, antiviral, antiathero-genic, antibacterial, and anticancer effects (Cos, Calomme, Pieters,Vlietinck, & Vanden Berghe, 2000; Middleton, Kandaswami, &Theoharides, 2000). p-Hydroxybenzoic acid (p-HBA) can be takeninto account as the common monophenolic structure of a largenumber of derivatives obtained from phenolic acids.Monophenols have been shown to possess less efficient radicalscavenging activity than polyphenols (Hsieh, Yen, & Chen, 2005).However, the structural variations due to the introduction of dif-ferent electron-donating/withdrawing groups to the various posi-tions of the phenolic ring can promote the antioxidant potency ofthe resulted compounds (Shahidi &Wanasundara, 1992). The func-tional groups substituted to the ortho or para positions of phenolicrings have been shown to be more effective than those attached tometa position in changing the performance of phenolic antioxi-dants (Chen & Ho, 1997). The majority of studies on the struc-ture–antioxidant activity relationship have individuallyconsidered the influence of different functional groups or yielded

the data which are less comparable because of the different meth-ods used to evaluate the antioxidant capacity.

o-Hydroxyl, o-methoxy, and alkyl ester groups are among thefunctional groups that are frequently seen in a wide range ofphenolic compounds. Protocatechuic acid (3,4-dihydroxybenzoicacid, PCA), gallic acid (3,4,5-trihydroxybenzoic acid, GA), vanillicacid (3-methoxy,4-hydroxybenzoic acid, VA), syringic acid(3,5-dimethoxy,4-hydroxybenzoic acid, SA), methyl p-hydroxy-benzoate (p-HBM), ethyl protocatechuate (EPC), and methyl gallate(MG) are recognized as the most well-known derivatives of p-HBAcontaining these functional groups (Fig. 1). In the present study,the antioxidant activity of these phenolic derivatives wasevaluated by different methods, in order to get an insight intothe structure–activity relationships responsible for the observedperformances.

At the first step of the evaluation, the reactivity of the p-HBAderivatives was measured in terms of the strength to scavengethe stable DPPH (1,1-diphenyl-2-picrylhydrazyl) free radical(Lima, Fernandes-Ferreira, & Pereira-Wilson, 2006), which is con-sidered as the main mechanism of action of phenolic antioxidants.The second step was to investigate the inhibitory effect of the com-pounds on the formation of hydroperoxides, which is considered tobe one of the most commonly used methods to evaluate the inhi-bitory effect of antioxidants, during the oxidation of fish oil as anedible oil with extremely beneficial effects on human health(Arkhipeko & Sazontova, 1995; Cleland, James, & Proudman,2003; Hye-Kyeong, Della-Fera, Lin, & Baile, 2006) but high suscep-tibility to oxidative deteriorations. Finally, a dispersed system

R1 R2 R3

p-Hydroxybenzoic acid Protocatechuic acid Gallic acid Vanillic acid Syringic acid Methyl p-hydroxybenzoateMethyl gallate Ethyl protocatechuate

HOHOH

OCH3OCH3

HOHOH

HH

OHH

OCH3H

OHH

OHOHOHOHOH

OCH3OCH3

OCH2CH3

Fig. 1. Molecular structure of p-hydroxybenzoic acid and its derivatives.

R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134 129

consisting of two immiscible phases (the fish oil-in-water, O/W,emulsion) was used to gain a better understand of thepartitioning-based performance of the phenolic antioxidants infood matrices which are usually multicomponent systems.

2. Methods and materials

2.1. Materials

Crude Kilka fish oil (Table 1) was supplied by Khazar company(Babolsar, Iran). p-HBA and its derivatives (p-HBM, PCA, EPC, GA,MG, VA, and SA) as well as a-tocopherol (a-Toc) were purchasedfrom Sigma–Aldrich (St. Louis, MO). All chemicals and solventsused in this study were of analytical reagent grade and purchasedfrom Merck (Darmstadt, Germany) and Sigma–Aldrich (St. Louis,MO).

2.2. Fatty acid composition

The fatty acids methyl ester (FAME) preparation of the oil sam-ples was carried out to determine the fatty acids composition inlipid fraction. The oil (0.1 ml) was pipetted into the clean 10 mlof screw-top glass bottles, dissolved in 1 ml of hexane and con-verted to the methyl esters by reaction with 0.5 ml of sodiummethoxide. The mixture was homogenized using the vortex for10–15 s. The clear upper phase layer was pipetted out and injectedto the gas chromatograph (Sharina & Jumat, 2006). Menhaden oilwas used as the standard of PUFA and the identity of individualFAME was compared after conversion to equivalent chain length.

Table 1Fatty acid composition (%w/w) of the Kilka fish oil.

Parameter

C14:0 6.22 ± 0.05C16:0 17.31 ± 0.01C16:1 13.23 ± 0.07C17:0 1.89 ± 0.06C18:0 3.23 ± 0.04C18:1 27.51 ± 0.06C18:2 8.16 ± 0.15C18:3 1.17 ± 0.09C20:0 1.16 ± 0.06C20:4 0.21 ± 0.03C20:5 (EPA) 6.35 ± 0.05C22:6 (DHA) 5.89 ± 0.03

Mean ± SD (standard deviation) of triplicatedeterminations.

Routine GC analyses were performed on a Shimadzu GC-17A GasChromatography equipped with FID detector. The column usedwas BPX-70 (60 m length � 0.32 mm i.d � 0.25 lm thickness), splitratio 100:1. The analyses were performed using programmed tem-perature at the initial temperature of 120 �C, with the temperatureincrement rate at 3 �C min�1 and final temperature at 245 �C. Theinjection port temperature was set at 260 �C and detector temper-ature, was at 280 �C. Nitrogen gas was used as a carrier gas.

2.3. Purification of the fish oil sample

To remove indigenous antioxidants, 120 g of the oil samplewere applied to a glass chromatographic column (50 � 5 cm i.d.)packed sequentially with three adsorbents. The bottom layer wasaluminum oxide 60 (50 g, active, neutral) activated at 200 �C for3 h immediately before use. The middle layer was 80 g of the acti-vated silica gel (60–200 mesh) activated at 160 �C for 3 h immedi-ately before use. The upper layer was 2 g of the activated carbon.The column and collection vessels were wrapped in aluminum foil,and the oil was drawn through the column by suction without sol-vent (Belhaj, Arab-Tehrany, & Linder, 2010).

2.4. Preparation of the fish oil-in-water emulsion

The O/W emulsion was prepared by gently adding 10% of thepurified fish oil containing 200 ppm of each antioxidant into asolution containing 5% of soy protein isolate. To obtain a stableemulsion, the mixture was vortexed by ultra-Turrax (5 min,�3000g), afterwards was sonicated for 4 min in an ice bath. Tomonitor lipid oxidation, the emulsion samples were kept in anoven at 55 �C. Oil extraction from the emulsions for analysis wascarried out by mixing a chloroform/methanol solvent system(1:1, v/v) and the emulsion in a shaker (1 min) and then centrifug-ing for 1 min at �700g. The lower lipid layer was collected and itssolvent evaporated using a stream of nitrogen.

2.5. Partition coefficient (log P)

Solutions (0.3 mM) of each compound in 1-octanol were kept at60 �C for 1 h. The maximum absorbance was read by UV spectrum(A0). Equal volumes of this solution and acetate buffer (0.1 M, pHs3.5 and 5.5) were vortexed (69.875 g) for 1 min. The UV spectrumof the 1-octanol layer was determined after 30 min (Ax). The parti-tion coefficient (log P) of antioxidant was calculated according tothe following equation (Gorden, Paivia-Martins, & Almeida, 2001):

P ¼ Ax=ðA0 � AxÞ ð1Þ

2.6. Radical scavenging activity

The antioxidants in different concentration ranges dependenton the antioxidative power were reacted with the stable DPPH freeradical in a methanol solution (60 lM). After 30 min at room tem-perature in the dark, the absorbance of the samples was readagainst a blank at 517 nm. Inhibition of the DPPH free radical inpercent (I%) was calculated as follows:

I% ¼ 100� ðABlank � ASampleÞ=ABlank ð2Þwhere ABlank is the absorbance of the control reaction (containing allreagents except the test compound), and ASample is the absorbanceof the test compound. The concentration of antioxidant requiredfor 50% inhibition of the DPPH free radical (IC50 value) was calcu-lated by linear regression analysis of dose–response curve plottingbetween the % inhibition and concentrations (Lima et al., 2006).Radical scavenging activity (RSA) was calculated from the IC50 valueas follows:

Table 2Radical scavenging activity (RSA, 100/IC50) and partition coefficient (log P) of thephenolic compounds.a

Compound RSA Log P at pH

3.5 5.5

p-Hydroxybenzoic acid – 0.75 ± 0.02 d 0.07 ± 0.03 eMethyl p-hydroxybenzoate – 0.85 ± 0.04 bc 0.87 ± 0.04 cProtocatechuic acid 1.50 ± 0.01 d 0.55 ± 0.02 e –0.39 ± 0.02 gEthyl protocatechuate 2.01 ± 0.02 c 0.95 ± 0.07 b 1.03 ± 0.03 bGallic acid 3.40 ± 0.00 a 0.31 ± 0.02 f –0.74 ± 0.03 hMethyl gallate 2.63 ± 0.01 b 0.77 ± 0.04 cd 0.74 ± 0.03 dVanillic acid 0.24 ± 0.01 f 0.58 ± 0.02 e –0.32 ± 0.02 fSyringic acid 2.07 ± 0.02 c 0.37 ± 0.03 f –0.65 ± 0.04 ha-Tocopherol 0.95 ± 0.00 e 2.50 ± 0.06 a 2.30 ± 0.04 a

a Means ± SD (standard deviation) within a column with the same lowercaseletters are not significantly different at p < 0.05.

130 R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134

RSA ¼ 100� ð1=IC50Þ ð3Þwhere the larger the RSA value, the more efficient the antioxidant.

2.7. Oxidative stability test

The purified fish oil samples (5 g) containing 200 ppm of thephenolic antioxidants were stored in a 1-mm layer in a Petri dishwith a diameter of 9 cm at 35 �C. Progress of lipid oxidation wasmonitored by the spectrophotometrical determination of peroxidevalue (PV) at 500 nm by UV–VIS instrument (Model 160AShimadzu, Kyoto, Japan) (Shantha & Decker, 1994). Changes in PVversus storage time of the fish oil sample and its O/W emulsion(see the section 2.3) were separately plotted (Fig. 2). Thex-coordinate of intersection point of two straight lines fitted onthe initiation and propagation stages of the oil oxidation was calcu-lated as induction period, IP (Farhoosh & Hoseini-Yazdi, 2013).Antioxidant activity was calculated as follows:

A ¼ IPinh �W0ð Þ= IP0 �W inhð Þ ð4Þwhere IPinh and IP0 are the IPs in presence and absence of an antiox-idant, respectively, and Winh and W0 are the oxidation rates (theslope of the linear initiation stage of lipid oxidation) in presenceand absence of an antioxidant, respectively.

2.8. Statistical analysis

All determinations were carried out in triplicate, and data weresubjected to analysis of variance (ANOVA). ANOVA analyses wereperformed according to SAS software. Significant differencesbetween means were determined by Duncan’s multiple rangetests; p values less than 0.05 were considered statisticallysignificant.

3. Results and discussion

3.1. Partition coefficient (Log P)

Measuring the log P value provides helpful information aboutthe polarity-based antioxidative behavior of phenolic antioxidants.

Fig. 2. A schematic kinetic curve of peroxide accumulation during o

The log P value was measured in pH values of 3.5 and 5.5; the ion-ization of carboxyl group with pKa values of around 4.5 in phenolicacids is suppressed at pKa � 1 and is complete at pKa + 1. The val-ues of log P ranged from 0.31 to 2.50 at pH 3.5 and from �0.74 to2.30 at pH 5.5 (Table 2). More hydrophobic phenolic compoundspossessed the higher values of log P. The highest value at bothpH belonged to a-Toc, followed by EPC, p-HBM, MG, p-HBA, VA,PCA, SA, and GA. The phenolic acids containing protonated car-boxyl group at the lower pH value exhibited higher hydrophobici-ties, whereas the esterified ones were undergone no significantvariation in the log P value as the pH changed.

The hydrophilic nature of p-HBA changed in a different mannerby esterifying carboxyl group and also introducing hydroxyl andmethoxy substituents to the ortho position of phenolic hydroxylgroup on the aromatic ring. o-Hydroxyl groups led to progressivedecreases in the log P value from 0.07 in p-HBA to �0.39 in PCAand �0.74 in GA (pH 5.5). o-Methoxy groups also caused the logP value to decrease but in a lower rate, so that the log P value ofp-HBA reached �0.32 in VA and �0.65 in SA. This indicates thato-hydroxyl groups and, to a lesser amount, o-methoxy groupsmake p-HBA more hydrophilic. Less effective role of o-methoxygroups in increasing the polarity of phenolic acids can be

xidation of lipid systems, and the corresponding kinetic values.

R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134 131

attributed to the fact that o-hydroxyl groups are easily capable ofcreating linear intermolecular hydrogen bonds in polar proticand/or aprotic solvents but o-methoxy groups are less likely to cre-ate such hydrogen bonds, which will be established only in polarprotic solvents (de Heer, Korth, & Mulder, 1999). In contrast, thealkyl ester groups increased the hydrophobicity of the p-HBAderivatives, so that the log P value at pH 5.5 increased from 0.07in p-HBA to 0.87 in p-HBM, from �0.74 in GA to 0.74 in MG, andfrom �0.39 in PCA to 1.03 in EPC. As can be seen, the rate ofchanges in the hydrophobicity as affected by the alkyl ester groupswas higher than that in the hydrophilicity due to the o-hydroxyland o-methoxy groups. It is expected that the bigger alkyl estergroup is, the more hydrophobic is the molecule.

3.2. Radical scavenging activity

The RSA of the p-HBA derivatives in the study is shown inTable 2. The results were compared to a-tocopherol as control withthe RSA = 0.95. There was found no concentration of p-HBA and itsmethyl ester to scavenge 50% of the DPPH free radical, and there-fore, no RSA value was calculated for them. GA was found to bethe most active antiradical agent, followed by MG, SA � EPC, PCA,a-Toc, and VA.

GA with an extra hydroxyl group in ortho position to the pheno-lic hydroxyl group had a RSA value more than two times that ofPCA (3.40 vs. 1.50). This was in agreement with previous findingsindicating that a pyrogallol moiety in the molecular structure ofphenolic compounds leads to a better antioxidant activity than acatechol moiety (Kawabata, Okamoto, Kodama, Makimoto, &Kasai, 2002; Siquet, Paiva-Martin, Lima, Reis, & Borges, 2006).More potent antioxidants have lower O–H bond dissociationenthalpies (BDE), facilitating the process of direct hydrogen trans-fer to a radical (Zhang, Sun, & Wang, 2003). Electron-donatingo-hydroxyl substituents on the aromatic ring lower the O–H BDEvalue and thus enhance the antiradical activity (Wright, Johnson,& DiLabio, 2001). Following the hydrogen abstraction, polyphenolswith ortho or para position heteroatoms may produce quinone orsemiquinone free radicals, and then form o- or p-benzoquinonestructures through resonance, which can stabilize the phenolicradical to a certain extent (Friedman & Jurgens, 2000)(Scheme 1). Kawabata et al. (2002) showed that the protocate-chuquinone resulted from PCA is able to provide an extra inhibi-tion by slowly reacting with DPPH free radicals, whereas GA iscapable of producing highly reactive hydroxyquinone intermedi-ates which can scavenge the radicals more efficiently. Another pos-sible explanation for the antiradical activity difference is that thepotential intramolecular hydrogen bonds (IHB) in the catechol(Scheme 2) or pyrogallol (Scheme 3) moieties and also the

-Hydroxyphenol Semiquinone reso

-Hydroxyphenol Semiquinone reso

o

p

Scheme 1. Formation of semiquinone radical interm

intermolecular hydrogen bonds between these functional groupsand polar solvents can play an important role in the DPPHradical-scavenging activity. IHBs have been shown to possess alarge contribution to lowering the O–H BDE value of polyphenols.The O–H BDE value for catechol has been calculated to be about10 (82.8–72.8) kcal/mol lower than that of phenol, being the sum-mation of the IHB (5.5 kcal/mol) and o-hydroxyl electronic(4.5 kcal/mol) effects (Zhang et al., 2003). Hence, it is expected thatGA capable of forming a more IHB than PCA lowers the O–H BDEvalue more significantly. Meanwhile, the semiquinone free radicalresulted from the hydrogen abstraction in PCA and GA can be sta-bilized by forming one and two IHBs (Schemes 2 and 3), respec-tively, with OH at ortho position (Cheng, Ren, Li, Chang, & Chen,2003; Siquet et al., 2006). It has been reported that the introduc-tion of more than two hydroxyl groups to a monophenol will notimprove the antioxidant efficiency (Pokorny, 1987).

o-Methoxy groups could promote the antiradical activity ofp-HBA as well but they were not able to create the same improve-ments in the radical-scavenging activities as the o-hydroxylgroups, so that the RSA value decreased from 1.50 and 3.40 inPCA and GA to 0.24 and 2.07 in VA and SA, respectively (Table 2).As can be seen, VA presented a much lower RSA value than PCAand also the lowest quantity among the phenolic derivatives stud-ied. The O–H BDE value for o-methoxyphenol has been estimatedto be 6.6 kcal/mol lower than that for phenol (de Heer et al.,1999), which is markedly lower than that for catechol(10 kcal/mol). This can be attributed to a weaker electron-donatingstrength of o-methoxy group than o-hydroxyl group and also theopposite effect of the IHB formed in o-methoxyphenol (Scheme 4).In their study on the antioxidant activity of ferulic acid derivatives,Nenadis, Zhang, and Tsimidou (2003) showed that the O–H BDEvalue increased (from 81 to 84 kcal/mol) as the IHB strengthbetween the phenolic O–H and o-methoxy groups increased (from3.7 to 4.6 kcal/mol). Besides, the semiquinone free radical of VA pro-duced after the abstraction of hydrogen from the only phenolichydroxyl group cannot be stabilized by forming IHB (Scheme 4).Furthermore, the catechol moiety of PCA has a further hydroxylgroup with a O–H BDE value approximately similar to that of thefirst one (73.3 kcal/mol) which allows its hydrogen to be readilyabstracted and employed by a second reactive radical (Zhang et al.,2003).

It is noteworthy to mention that the second o-methoxy group ofSA could increase the RSA value much better than the first one. Thiscan be due to the more effective role of o-dimethoxy moiety inreducing the O–H BDE value (10.1 kcal/mol vs. 6.6 kcal/mol foro-methoxy moiety) (Zhang et al., 2003). However, despite roughlythe same reduction in the O–H BDE value (�10 kcal/mol), SAshowed a significantly higher activity than PCA to scavenge the

nance hybrids -Benzoquinone

nance hybrids

o

p-Benzoquinone

ediates and quinones from dihydroxyphenols.

Scheme 2. Intramolecular hydrogen bonds in the catechol moiety of polypheols.

Scheme 3. Intramolecular hydrogen bonds in the pyrogallol moiety of polypheols.

Scheme 4. Intramolecular hydrogen bond in an o-methoxyphenol.

Methyl protocatechuate

Protocatechuquinone methyl ester

Methyl 2-methoxyprotocatechate

Methyl 2-methoxyprotocatechuquinone

3-hemiacetal

Scheme 5. Plausible radical scavenging mechanism of methyl protocatechuate inalcoholic solvents.

Table 3Kinetic parameters of hydroperoxide formation during the oxidation of the fish oilsample containing the phenolic compounds at 35 �C.a

IPinhb Winhc Ad

Control 10.17 ± 1.89 ef 0.92 ± 0.07 d 1.00 ± 0.00 fp-Hydroxybenzoic acid 10.07 ± 0.65 f 1.37 ± 0.15 c 0.66 ± 0.04 gMethyl p-

hydroxybenzoate11.21 ± 1.29 ef 1.53 ± 0.30 bc 0.66 ± 0.08 g

Protocatechuic acid 28.59 ± 1.46 d 0.93 ± 0.21 cd 2.75 ± 0.52 eEthyl protocatechuate 35.80 ± 1.19 c 0.43 ± 0.02 f 7.65 ± 0.26 dGallic acid 61.24 ± 1.17 a 0.17 ± 0.01 g 31.68 ± 0.63 bMethyl gallate 63.01 ± 2.72 a 0.15 ± 0.01 g 38.75 ± 1.69 aVanillic acid 12.32 ± 0.36 e 3.65 ± 0.14 a 0.30 ± 0.01 iSyringic acid 10.91 ± 0.21 f 1.93 ± 0.06 b 0.51 ± 0.01 ha-Tocopherol 55.67 ± 0.78 b 0.54 ± 0.01 e 9.27 ± 0.14 c

a Means ± SD (standard deviation) within a column with the same lowercaseletters for a system are not significantly different at p < 0.05.

b Induction period (hours) of inhibited oxidation; IP0 = 10.17 h.c Oxidation rate (meq/kg h) of inhibited oxidation; W0 = 0.92 meq/kg h.d Antioxidant activity, dimensionless.

132 R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134

DPPH free radical (RSA = 2.07 vs. 1.50). This may be owing to thefact that PCA can establish linear intermolecular hydrogenbonds with many polar solvents, leading to the prevention ofhydrogen abstraction to a certain extent, and therefore reducedantioxidant efficiency. The negative effect of the IHBs formed ino-methoxyphenols on the O–H BDE value can be removed to someextent by the intermolecular hydrogen bonds with polar solvents,leading to a raised antiradical activity.

Interestingly, the comparison between the RSA value of PCA andGA with their esters indicated different influence of the esterifica-tion reaction on the antiradical behavior of catechol-type PCA(increase from 1.50 to 2.01) and pyrogallol-type GA (decrease from3.40 to 2.63). Saito, Okamoto, and Kawabata (2004) reported thatmethyl protocatechuate dramatically inhibits DPPH free radicalswhen changing solvent from aprotic (e.g. acetone and acetonitrile)to protic (e.g. methanol, ethanol, and 1-propanol) ones. Theyshowed that the intermediate protocatechuquinone methyl esteris highly susceptible to nucleophilic attack of the alcohol moleculeto C-2 position, resulting in a regeneration of the catechol struc-ture capable of scavenging two additional radicals (Scheme 5).However, it has not been provided yet a clear reason for the oppo-site effect of esterification on the DPPH radical-scavenging activityof GA. It is possible that the additional hydroxyl group present inthe o-quinone produced from GA could enhance the reactivity ofthe quinone product to undergo further complex reactions.

3.3. Antioxidant activity in the bulk fish oil system

The rate of hydroperoxides formation in the initiation stage oflipid oxidation (W0 and Winh values) and also the duration of thisstage (IP0 and IPinh values) are affected by inhibitors and therefore,a combined kinetic parameter defined as antioxidant activity (Avalue) was employed to more comprehensively evaluate the per-formance of the phenolic compounds. The kinetic parameters ofhydroperoxide formation during the oxidation of fish oil samplecontaining the phenolic compounds at 35 �C are shown inTable 3. The results found in the bulk oil system for p-HBA andits methyl ester was very similar to those observed in the DPPHfree radical assay, so that they showed no stabilizing ability in

the fish oil and even exerted pro-oxidant activity to some extent.In their study on a series of synthesized phenolic-based antioxi-dants in olive oil using Rancimat test, Torres de Pinedo, Penalver,and Morales (2007) observed no comparable stabilizing potencyfor the monophenolic antioxidants experimented.

GA and its methyl ester with activity values of 31.7 and 38.8,respectively, could stabilize the fish oil more tangibly than theother phenolics studied. These values were much greater than thatof a-tocopherol (A = 9.3) which is often found in the mixed-tocopherol formula conventionally used to oxidatively stabilizefish oils. As can be seen in Tables 2 and 3, the antioxidant perfor-mance of GA in preventing the formation of hydroperoxides wasby far higher than in scavenging the DPPH free radical, so thatthe reactivity ratio between GA and PCA increased considerablyfrom �2.3 in the methanolic medium of the DPPH free radicalassay to �11.5 in the lipid matrix of the bulk fish oil system.Hydrogen bonds are primarily electrostatic interactions that theirstability depends on the dielectric constant of the environment.Lipid systems with dielectric constants extremely lower than thoseof aqueous systems provide anhydrous conditions in which IHBscapable of lowering the O–H BDE value of polyphenols (two bondsfor GA vs. one bond for PCA) are created more extensively(Fennema, 1996). Opposite results were obtained regarding theo-methoxy derivatives of p-HBA, so that the enhancement of theIHBs in VA and SA under the anhydrous conditions of the bulk oilsystem led to appear pro-oxidant activities to some extent (A val-ues <1, Table 3). Similar results have been reported by Torres dePinedo et al. (2007), indicating that the methoxylated antioxidantshad no tendency for stabilizing olive oil, in contrast to data fromthe DPPH free radical assay.

Unlike showing the different behaviors in the DPPH free radicalassay, the alkyl esters of both PCA and GA presented higher antiox-idant activities than their parent molecules in the bulk fish oil

R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134 133

system. However, EPC possessed a much higher stabilizing effectthan its parent molecule (a ratio of 2.8) when compared with theeffectiveness of MG relative to that of GA (a ratio of 1.2). It canbe ascribed to the fact that esterification reaction causes theelectron-withdrawing activity of carboxyl group attached to thephenolic ring to decrease with increasing the length of the alkylchain, resulting in a well-stabilized phenoxy radical (Torres dePinedo et al., 2007). In general, the performance of the phenolicacid antioxidants and their alkyl esters in the bulk oil systemimplies that their stabilizing strength is essentially related to theelectronic properties of the characteristic functional groups andIHBs. Meanwhile, the polar paradox theory, denoting that lesspolar antioxidants show higher strength in more polar media(Porter, 1993), cannot provide a suitable interpretation for thebehavior observed in many cases.

3.4. Antioxidant activity in the O/W emulsion system

The relative order of decreasing activity in the emulsion system,which was physically stable during the experiment, was found tobe MG � a-Toc > SA > EPC > GA � PCA � VA, with no antioxidantactivity observed for p-HBA and p-HBM (Table 4). The antioxidantefficiency of the phenolic compounds in the emulsion was almostcompletely different from that in the alcoholic environment ofthe DPPH free radical assay and the anhydrous conditions of thebulk fish oil system as well, implying that the performance of agiven antioxidant depends considerably on the method and/or sys-tem used for evaluation. In contrast to the high activities observedin the previous systems, GA was among the weakest antioxidantsin the emulsion system with no significant difference with PCA.Studying on a polarity-based range of GA derivatives in an O/Wemulsion system, Schwarz et al. (2000) showed that GA as themost polar member of the group exhibited pro-oxidant activity.In fact, the activity of antioxidants in dispersed systems relatesto their radical scavenging activities as well as their affinitytowards the water–oil interface, the site where oxidation occurs(Kikuzaki, Hisamoto, Hirose, Akiyama, & Taniguchi, 2002). Hence,it is expected that GA with the highest polarity among the pheno-lics studied (Table 2) incorporates a larger portion of its moleculesto the aqueous phase of the emulsion system than to the water–oilinterface. This is in agreement with the polar paradox theorydenoting that more polar antioxidants are less effective in morepolar media (Porter, 1993).

VA, which had the lowest antioxidant efficiency in the alcoholicand lipid media experimented, exerted the same performance asPCA statistically and an activity close to that of GA in the emulsion

Table 4Kinetic parameters of hydroperoxide formation during the oxidation of the fish oil-in-water emulsion containing the phenolic compounds at 55 �C.a

IPinhb Winhc Ad

Control 26.69 ± 0.26 g 0.28 ± 0.01 a 1.00 ± 0.00 fp-Hydroxybenzoic acid 18.31 ± 0.88 i 0.27 ± 0.03 ac 0.72 ± 0.04 hMethyl p-

hydroxybenzoate20.82 ± 0.57 h 0.25 ± 0.02 ac 0.89 ± 0.07 g

Protocatechuic acid 33.16 ± 2.05 e 0.25 ± 0.04 abce 1.39 ± 0.09 deEthyl protocatechuate 33.88 ± 1.19 e 0.18 ± 0.02 ef 2.02 ± 0.06 cGallic acid 28.42 ± 0.69 f 0.19 ± 0.01 bc 1.57 ± 0.08 dMethyl gallate 51.59 ± 0.40 b 0.12 ± 0.00 g 4.48 ± 0.17 aVanillic acid 42.87 ± 0.67 c 0.34 ± 0.04 a 1.33 ± 0.04 eSyringic acid 39.78 ± 1.05 d 0.18 ± 0.03 cef 2.29 ± 0.07 ba-Tocopherol 60.54 ± 1.51 a 0.15 ± 0.01 f 4.34 ± 0.21 a

a Means ± SD (standard deviation) within a column with the same lowercaseletters for a system are not significantly different at p < 0.05.

b Induction period (hours) of inhibited oxidation; IP0 = 26.69 h.c Oxidation rate (meq/kg h) of inhibited oxidation; W0 = 0.28 meq/kg h.d Antioxidant activity, dimensionless.

system. This can be due to the hydrophobic nature of its methylgroup, which must have shifted a larger contribution of the VAmolecules to the interface, agreeing with the polar paradox theory.Such an effect should have created a further interfacial activity inSA with a performance even more than those of polyphenolic acidsstudied. Therefore, when the affinity to interface predominates, theultimate efficiency of antioxidants possessing different polaritydepends to a lesser amount on the electronic phenomena of thefunctional groups attached.

The comparison between the activities due to o-methoxy (from1.39 in PCA to 1.33 in VA and from 1.57 in GA to 2.29 in SA) andalkyl ester (from 1.39 in PCA to 2.02 in EPC and from 1.57 to4.48 in MG) groups indicates the more effective role of the latterin increasing the antioxidant activity of phenolic acids in the emul-sion system. As can be seen in Table 2, the overall hydrophobicityof the phenolic ring has increased more by introducing alkyl esterthan methoxy groups. In addition, the nature of emulsifier used tostabilize the emulsion can greatly affect the contribution of antiox-idant incorporated to the water–oil interface (Velasco, Dobarganes,& Marquez-Ruiz, 2004). Soy protein isolate is considered to be anegatively charged emulsifier at neutral pH. This may account fora lower tendency to absorb phenolic acids existing as negativelycharged molecules, or a higher tendency to incorporate theuncharged alkyl esters of the phenolic acids.

4. Conclusions

The influence of o-hydroxyl, o-methoxy, and alkyl ester groupson the antioxidant activity of p-HBA as a function of the type ofreaction medium was investigated in the present study. An amal-gam of electronic phenomena, intra- and/or intermolecular rela-tionships, interfacial properties, and chemical reaction of thesolvent molecules with phenolic compounds were considered tobeing mainly involved in the antiradical activities observed.o-Hydroxyl groups were considered to be highly powerful electron-donors when compared with o-methoxy groups. Due to the weak-ening effect on the electron-withdrawing activity of the carboxylsubstituent, alkyl ester groups were able to increase the antioxi-dant activity, although their effectiveness was extensively affectedby the environmental conditions of the test used to evaluate theantioxidant performance. As for the polyphenolic derivatives,IHBs were considered as improving causes of the antioxidant per-formance, while the intermolecular hydrogen bonds were referredas the possible reason for the diminished efficiency. The obtainedresults indicated the opposite roles for the hydrogen bonds in theo-methoxy phenolic derivatives. The anhydrous conditions of thebulk oil system caused the IHBs to form more extensively, whereasthe polar medium of the solvent used enhanced the intermolecularhydrogen bonds. In the O/W emulsion system, interfacial proper-ties of the p-HBA derivatives as affected by the type of substitutinggroup, in a decreasing order of alkyl esters, o-methoxy, ando-hydroxyl groups, was taken into account to be the predominantfactor to exert antioxidant activity.

References

Arkhipeko, Y. V., & Sazontova, T. G. (1995). Mechanisms of the cardioprotectiveeffect of a diet enriched with x-3 polyunsaturated fatty acids. Pathophysiology,2, 131–140.

Belhaj, N., Arab-Tehrany, E., & Linder, M. (2010). Oxidative kinetics of salmon oil inbulk and in nanoemulsion stabilized by marine lecithin. Process Biochemistry,45, 187–195.

Chen, J. H., & Ho, C. T. (1997). Antioxidant activities of caffeic acid and its relatedhydroxycinnamic acid compounds. Journal of Agricultural and Food Chemistry,45, 2374–2378.

Cheng, Z., Ren, J., Li, Y., Chang, W., & Chen, Z. (2003). Establishment of a quantitativestructure–activity relationship model for evaluating and predicting the

134 R. Farhoosh et al. / Food Chemistry 194 (2016) 128–134

protective potentials of phenolic antioxidants on lipid peroxidation. Journal ofPharmaceutical Sciences, 92, 475–484.

Cleland, L., James, M., & Proudman, S. (2003). The role of fish oils in the treatment ofrheumatoid arthritis. Drugs, 63, 845–853.

Cos, P., Calomme, M., Pieters, M., Vlietinck, A. J., & Vanden Berghe, D. (2000). Studiesin Natural Products Chemistry. In Atta-Ur-Rahman (Ed.) (22, pp. 307–341).Amsterdam: Elsevier Science.

de Heer, M. I., Korth, H.-G., & Mulder, P. (1999). Poly methoxy phenols in solution:O–H bond dissociation enthalpies, structures, and hydrogen bonding. Journal ofOrganic Chemistry, 64, 6969–6975.

Farhoosh, R., & Hoseini-Yazdi, S. Z. (2013). Shelf-life prediction of olive oils usingempirical models developed at low and high temperatures. Food Chemistry, 141,557–565.

Fennema, O. R. (1996). Food Chemistry. New York: Marcel Dekker, Inc.Frankel, E. N. (1998). Lipid Oxidation. West Ferry, Dundee (Scotland): The Oily Press

Ltd.Friedman, M., & Jurgens, H. S. (2000). Effect of pH on the stability of plant phenolic

compounds. Journal of Agricultural and Food Chemistry, 48, 2101–2110.Gorden, M. H., Paivia-Martins, F., & Almeida, M. (2001). Antioxidant activity of

hydroxytyrosol acetate compared with that of other olive oil polyphenols.Journal of Agricultural and Food Chemistry, 49, 2480–2485.

Hsieh, C.-L., Yen, G.-C., & Chen, H.-Y. (2005). Antioxidant activities of phenolic acidson ultraviolet radiation-induced erythrocyte and low density lipoproteinoxidation. Journal of Agricultural and Food Chemistry, 53, 6151–6155.

Hye-Kyeong, H., Della-Fera, M. A., Lin, J., & Baile, C. A. (2006). Docosahexaenoic acidinhibits adipocyte differentiation and induces apoptosis in 3T3-L1preadipocytes. Journal of Nutrition, 21, 2965–2969.

Kawabata, J., Okamoto, Y., Kodama, A., Makimoto, T., & Kasai, T. (2002). Oxidativedimers produced from protocatechuic and gallic esters in the DPPH radicalscavenging reaction. Journal of Agricultural and Food Chemistry, 50, 5468–5471.

Kikuzaki, H., Hisamoto, M., Hirose, K., Akiyama, K., & Taniguchi, H. (2002).Antioxidant properties of ferulic acd and its related compounds. Journal ofAgricultural and Food Chemistry, 50, 2161–2168.

Lima, C. F., Fernandes-Ferreira, M., & Pereira-Wilson, C. (2006). Phenolic compoundsprotect HepG2 cells from oxidative damage: relevance of glutathione levels. LifeScience, 79, 2056–2068.

Middleton, E. J., Kandaswami, C., & Theoharides, T. C. (2000). The effects of plantflavonoids on mammalian cells: Implications for inflammation, heart disease,and cancer. Pharmacological Reviews, 52, 673–751.

Nenadis, N., Zhang, H.-Y., & Tsimidou, M. Z. (2003). Structure–antioxidant activityrelationship of ferulic acid derivatives: effect of carbon side chain characteristicgroups. Journal of Agricultural and Food Chemistry, 51, 1874–1879.

Pokorny, J. (1987). In H. Chan (Ed.), Autoxidation of Unsaturated Lipids (pp. 141–206).London: Academic Press.

Porter, W. L. (1993). Paradoxical behavior of antioxidants in food and biologicalsystems. In G. M. William (Ed.), Antioxidants: Chemical, physiological, nutritionaland toxicological aspects (pp. 93–122). Princeton Scientific: Princeton.

Saito, S., Okamoto, Y., & Kawabata, J. (2004). Effects of alcoholic solvents onantiradical abilities of protocatechuic acid and its alkyl esters. Bioscience,Biotechnology, and Biochemistry, 68, 1221–1227.

Schwarz, K., Huang, S.-W., German, J. B., Tiersch, B., Hartmann, J., & Frankel, E. N.(2000). Activities of antioxidants are affected by colloidal properties of oil-in-water and water-in-oil emulsions and bulk oils. Journal of Agricultural and FoodChemistry, 48, 4874–4882.

Shahidi, F., & Wanasundara, P. K. J. P. D. (1992). Phenolic antioxidants. CriticalReviews in Food Science and Nutrition., 32, 67–103.

Shantha, N. C., & Decker, E. A. (1994). Rapid, sensitive, iron-basedspectrophotometric methods for determination of peroxide values of foodlipids. Journal of AOAC International, 77, 421–424.

Sharina, S., & Jumat, S. (2006). Physicochemical characteristics of aji-fish seriolanigrofasciata lipid. Malaysian Journal of Analytical Sciences, 10, 55–58.

Siquet, C., Paiva-Martin, F., Lima, J. F. C., Reis, S., & Borges, F. (2006). Antioxidantprofile of dihydroxy- and trihydroxyphenolic acids: A structure–activityrelationship study. Free Radical Research, 40, 433–442.

Torres de Pinedo, A., Penalver, P., & Morales, J. C. (2007). Synthesis and evaluation ofnew phenolic-based antioxidants: Structure–activity relationship. FoodChemistry, 103, 55–61.

Velasco, J., Dobarganes, M. C., & Marquez-Ruiz, G. (2004). Antioxidant activity ofphenolic compounds in sunflower oil-in-water emulsion containing sodiumcaseinate and lactose. European Journal of Science and Technology, 106, 325–333.

Wright, J. S., Johnson, E. R., & DiLabio, G. A. (2001). Predicting the activity of phenolicantioxidants: Theoretical method, analysis of substituent effects, andapplication to major families of antioxidants. Journal of the American OilChemists’ Society, 123, 1173–1183.

Zhang, H.-Y., Sun, Y.-M., & Wang, X.-L. (2003). Substituent effects on O–H bonddissociation enthalpies and ionization potentials of catechols: A DFT study andits implications in the rational design of phenolic antioxidants and elucidationof structure–activity relationships for flavonoid antioxidants. Chemistry – AEuropean Journal, 9, 502–508.