Phenolic antioxidants

5 downloads 0 Views 718KB Size Report
Received: January 10, 2009; accepted: March 27, 2009 ... 2009 WILEY-VCH Verlag GmbH & Co. .... and/or diphenyl methane (DPM) at 50 7C under air atmos-.
1072

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Research Paper Phenolic antioxidants – radical-scavenging and chain-breaking activity: A comparative study* Vessela D. Kancheva Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Sofia, Bulgaria

Fifty phenolic antioxidants (AH) (42 individual compounds and 8 binary mixtures of two antioxidants) were chosen for a comparative analysis of their radical-scavenging (H-donating) and chain-breaking (antioxidant) activity. Correlations between experimental (antiradical and antioxidant) and predictable (theoretical) activities of 15 flavonoids, 15 hydroxy cinnamic acid derivatives, 5 hydroxy chalcones, 4 dihydroxy coumarins and 3 standard antioxidants (butylated hydroxytoluene, hydroquinone, DL-atocopherol) were summarized and discussed. The following models were applied to explain the structureactivity relationships of phenolic antioxidants of natural origin: (a) model 1, a DPPH assay used for the determination of the radical-scavenging capacity (AH 1 DPPH. ? A. 1 DPPH-H); (b) model 2, chemiluminescence of a model substrate RH (cumene or diphenylmethane) used for the determination of the rate constant of a reaction with model peroxyl radicals (AH 1 RO2. ? ROOH 1 A.); (c) model 3, lipid autoxidation used for the determination of the chain-breaking antioxidant efficiency and reactivity (AH 1 LO2. ? LOOH 1 A.; A. 1 LH (1O2) ? AH 1 LO2.); and (d) model 4, theoretical methods used for predicting the activity (predictable activity). The highest lipid oxidation stability was found for antioxidants with a catecholic structure and for their binary mixtures with DL-a-tocopherol, as a result of synergism between them. Keywords: Antioxidant activity / Phenolic antioxidants / Radical-scavenging activity

Received: January 10, 2009; accepted: March 27, 2009 DOI 10.1002/ejlt.200900005

1 Introduction The human body is constantly subjected to significant oxidative stress as a result of a misbalance between antioxidative protective systems and the formation of strong oxidizing substances, including free radicals. The free-radical formation process results in damage and death of cells, accelerates aging and initiates many diseases such as cardiovascular diseases, cancer, Parkinson’s disease, etc. In this respect, the medical treatment of most of these illnesses includes formulations based on a combination of traditional drugs with targeted functionality and different antioxidants [1–5]. Plant polyphenolics are multifunctional antioxidants and act as reducing

Correspondence: Assoc. Prof. Vessela D. Kancheva (formerly Kortenska), Lipid Chemistry Department, Institute of Organic Chemistry with Centre of Phytochemistry, Bulgarian Academy of Sciences, Acad. “Georgi Bonchev” Street, Block 9, Sofia 1113, Bulgaria. E-mail: [email protected] Fax: 1359 2 8700225

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

agents (free-radical terminators), metal chelators, and singlet oxygen quenchers. The activity of antioxidants depends on complex factors including the nature of the antioxidants, the conditions of oxidation, the properties of the oxidizing substrate and the stage of oxidation [6–10]. The problem of the antioxidant capacity of food has at least two sides: (a) the antioxidant potential, which is determined by the antioxidant composition and the antioxidative properties of the constituents and is the subject of food chemistry, and (b) biological effects which depend, among other things, on the bioavailability of antioxidants, and pose a medico-biological problem [3, 6, 10]. The inherent compositional and structural complexity of real foods means that systematic studies of lipid oxidation must first be carried out in model systems. These

* Parts of this paper has been reported at the 5th Euro Fed Lipid Congress 2007 in Gothenburg, Sweden: “Structure-activity relationships of phenolic antioxidants with a biological activity from natural origin” and at the 6th Euro Fed Lipid Congress 2008 in Athens, Greece: “Chain-breaking antioxidant activity of two new chalcones of propolis of El Salvador in homogeneous and micellar media”. www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

systems should have the basic features of real systems but welldetermined compositions and structural properties [11–15]. Structure-activity relationship studies using theoretical methods are gaining interest among scientists for the prediction or elucidation of differences in the activity of series of molecules. Phenolic antioxidants are the most intensively examined category because of their broadly accepted biological activity [16– 22]. The aim of this article is to present a comparative analysis of the activity of antioxidants as radical scavengers and chainbreaking antioxidants as well as their structure-activity relationships, using various experimental and theoretical methods. For these reasons, 42 individual phenolic antioxidants with different numbers of functional phenolic groups (mono-, bi- and polyphenols), different hindrance, i.e. reactivity, and with a wide spectrum of their hydrophilic-lipophilic balance, as well as 8 binary mixtures of two antioxidants were chosen (Fig. 1). They were separated into the following groups: group A: hydroxy cinnamic acids (1–15); group B: standard antioxidants (16–18); group C: flavonoids and related compounds (19–33); group D: hydroxy chalcones (34–38); group E: dihydroxy coumarins (39–42); and group F: binary mixtures of two antioxidants (43–50). Correlations between experimental (model 1, DPPH assay; model 2, chemiluminescence method; and model 3, lipid autoxidation) and theoretical (model 4, QSAR) methods applied to explain the structure-activity relationships of antioxidants and the obtained data are summarized and discussed.

2 Materials and methods 2.1 Phenolic antioxidants Fifteen flavonoids and related compounds were isolated from Bulgarian medicinal plants [21, 22]. Prenyl derivatives of 4hydroxy cinnamic acid were obtained from Brazilian propolis ethanol extract [23]. Two new chalcones (34–36) were isolated from El Salvador propolis [24]. DL-a-Tocopherol (aTOH), butylated hydroxytoluene (BHT), hydroquinone (HQ) and caffeic acid (CA) were obtained from E. Merck (Darmstadt, Germany); p-coumaric acid (p-CumA), ferulic acid (FA) and sinapic acid (SA) were obtained from Fluka (Buchs, Switzerland). 6-Hydroxy-2,2,5,7,8-pentamethylchromate (CrC1), a synthetic analogue of a-TOH, was synthesized in the Institute of Chemical Physics, Russian Academy of Sciences. Some of the phenolic compounds under study were synthesized [25–27] with the purpose of yielding a sufficient amount needed for the determination of their antioxidant activity. These compounds are hydroxy chalcones (34–36) [26], dihydroxy coumarins (39–42) [27] and Nhydroxy-cinnamoyl amino acid amides (5–14) [25]: (six feruloyl (6–10) and three sinapoyl (12–14) derivatives) the accumulation of which in plant material is assumed to be a part of the defense system of the plants displayed as a protective reaction against stress. Binary mixtures of two anti© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Phenolic antioxidants

1073

oxidants with equimolecular concentration of 0.1 and 0.5 mM were prepared in a 1 : 1 ratio. Al2O3 (neutral; Merck, Germany) and n-hexane (Riedel-de-Häen) were used for column chromatography. 2,2-Diphenyl-1-picrylhydrazyl (DPPH) radical (approximately 90%) was obtained from Sigma Chemical Co. All solvents used were of the highest quality (HPLC grade) available from Merck (Darmstadt, Germany). The substrates of oxidation, diphenylmethane and cumene (Aldrich), chlorobenzene (Aldrich), the initiator azo-bis-isobutyronitrile (AIBN; Merck) and the chemiluminescence activator 9,10-dibromoanthracene (Merck) were purified by standard methods.

2.2 Models 1–3 Model 1: Estimation of radical-scavenging activity (% RSA) by rapid DPPH radical test, which gives information about the H-donating capacity of the phenolic compounds towards free (DPPH) radicals. The radical scavenging activity (%RSA) of the phenolic compounds (AH) under study was based on the rapid DPPH test of Nenadis and Tsimidou [28]. The decrease in the absorbtion at 516 nm of the DPPH radical solution in ethanol after addition of the AH was measured in a glass cuvette (1 cm long), automatically. Different concentrations (expressed as the number of antioxidant moles per mole of DPPH ([AH]/ [DPPH] = 0.15 and 0.25) were used and the % RSA of antioxidants under study was calculated from the absorbance at the start and after 10 and 20 min of reaction time and is expressed as % RSAexp = [Abs516nm(t=0) – Abs516nm(t=20)6100/ Abs516nm(t=0)] [28, 29]. Absorbance values were corrected for radical decay using blank solutions. All tests were performed in triplicate at 25 7C. Model 2: Estimation of the rate constant of antioxidant with peroxide radicals (kA) using the kinetic chemiluminescence (CL) method [26, 30, 31], which provides information about the antioxidant efficiency (kA) and the stoichiometry (n). The absolute value of the rate constant of the inhibitor reactions with peroxyl radicals (kA) can be determined only by using this highly sensitive CL method. The reaction between a phenolic antioxidant and peroxyl radicals (AH 1 RO2. ? ROOH 1 A.) is known as the key reaction of the inhibited oxidation process. In CL experiments, the duration of the induction period (IP) is equal to the time from the moment of the inhibitors’ injection to 50% restoration of the CL intensity (for experimental details see [30, 31]). The kinetics of CL was studied by monitoring the autoxidation of cumene (60 7C) and/or diphenyl methane (DPM) at 50 7C under air atmosphere. The rate of free-radical generation (rate of initiation, RIN) was determined from the induction period, IP, caused by adding the standard inhibitor CrC1 at known concentrations, [CrC1]0: RIN = 2[CrC1]0/IP. Model 3: Estimation of the chain-breaking antioxidant activity of the phenolic compounds during lipid autoxidation, www.ejlst.com

1074

V. D. Kancheva

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Phenolic antioxidants

1075

www.ejlst.com

1076

V. D. Kancheva

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Phenolic antioxidants

1077

Figure 1. Structures of the phenolic antioxidants under study

which provides information about the efficiency (RAE) and reactivity (ID) of the phenolic antioxidants to inhibit the lipid oxidation. Triacylglycerols of commercially available sunflower oil (TGSO) were cleaned from pro- and antioxidants by adsorption chromatography and stored under nitrogen at –20 7C. Lipid samples (kinetically pure TGSO) containing various inhibitors were prepared directly before use. Aliquots of the antioxidant solutions in purified acetone or methanol and 1– 2 droplets of DMSO, depending on the solubility, were added to the lipid sample. Solvents were removed in a nitrogen flow (99.99%). The fatty acid composition of the lipid sample was determined by GC analysis of the methyl esters of the total fatty acids obtained according to Christie [32] with GC-FID Hewlett-Packard 5890 equipment (Hewlett-Packard GmbH, Austria) and a capillary column HP INNOWAX (polyethylene glycol mobile phase; Agilent Technologies, USA) 30 m60.25 mm60.25 mm. Lipid autoxidation was carried out at 80 7C (60.2) by blowing air through the samples (2.0 mL) in the dark at a rate of 100 mL/min (kinetic regime of oxidation). The oxidation process was monitored by withdrawing samples at various time intervals and determining the primary oxidation products (hydroperoxides) as peroxide value (PV) [33]. All kinetic data were calculated as the mean result of two or three independent experiments and were processed using the computer programs Origin 6.1 and Microsoft Excel-97. The standard deviation (SD) for different mean values of the induction period (IP) © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

was (in h): IP = 2.0, SD = 0.2; IP = 5.0, SD = 0.3; IP = 15.0, SD = 1.0; IP = 25, SD = 1.5; IP = 50.0, SD = 3.0. The SD of the PV determination (in meq/kg), according to the modified iodometric method for different mean values of PV, was: PV = 12.0, SD = 1.0; PV = 30.0, SD = 2.0; PV = 70.0, SD = 5.0; PV = 150.0, SD = 10; PV = 250.0, SD = 20. The RA and RC were quite constant, varying by less than 2% [9, 34–37].

2.3 Determination of the main antioxidative properties of the tested compounds The relative antioxidant efficiency (RAE) expresses the relative increase in the oxidation stability of the sample in the presence of antioxidant, taking into account the oxidizability of the control sample, i.e. RAE = (DIP/IPC) = (IPA – IPC)/IPC. IPC and IPA represent the induction periods in the absence and in the presence of an antioxidant, respectively. The inhibition degree (ID) is a measure of the antioxidant reactivity or strength, e.g. how many times the antioxidant shortens the oxidation chain length, i.e. ID = Rc/RA. RC and RA are the initial oxidation rate in the absence and in the presence of an antioxidant, respectively. Synergism, additivism, and antagonism of binary mixtures [38]: Synergism is observed when the inhibiting effect of the binary mixtures (IP112) is higher than the sum of the induction periods of the individual phenolic antioxidants (IP1 1 IP2), i.e. IP112 . IP1 1 IP2. The percent of the synergism is www.ejlst.com

1078

V. D. Kancheva

presented by the following formula [39]: % Synergism = {[IP112 – (IP1 1 IP2)]/(IP1 1 IP2)}6100, %. Additivism is observed when the inhibiting effect of the binary mixtures (IP112) is equal to the sum of the induction periods of the phenolic antioxidants alone (IP1 1 IP2), i.e. IP112 = IP1 1 IP2. Antagonism is observed when the inhibiting effect of the binary mixtures (IP112) is lower than the sum of the induction periods of the individual phenolic antioxidants (IP1 1 IP2), i.e. IP112 , IP1 1 IP2.

2.3 Model 4: Theoretical methods (predictable activity) Statistical analysis: The statistical analysis gives information about the correlation between the experimental and the theoretical (calculated/predictable) radical-scavenging activity of flavonoids and related compounds [21, 40]. A correlation between the antiradical activities with various physicochemical parameters of flavonoids, presented as theoretical values, i.e. QSAR, is based on the statistical analysis of Amic et al. [40] considering the well-known fact that the position of the phenolic groups in flavonoids could be more important for the radical-scavenging activity than the number of the phenolic groups. Thus, the indicator variable I is defined as the sum of the following molecular features: (i) the presence of a 2,3-double bond (I = 1) or (ii) of two 3,5,7-OH groups (I = 1) or (iii) of two 3’,4’,5’-OH groups (I = 1), or (iv) in the case of the absence of the above situations (I = 0): % RSAcalc = 3.954 1 75.950 I3’,4’-di-OH or 3-OH 1 8.499 I5-OH (I = 1 for 3’,4’-di-OH and/or 3-OH) and I = 1 for 5-OH). TOPS-MODE QSAR approach: TOPS-MODE is the acronym of topological substructural molecular design. The philosophy of this approach consists of deriving linear quantitative structure-activity relationship models using moments of the bond matrix. These models contain encrypted structural information. The key advantage of the TOPS-MODE approach is that the developed encrypted models may be transformed into quantitative contributions of the different bonds in the molecule. These bond contributions can then be used in formulating rules regarding the influence of key features of the molecule in a mechanism of action. The TOPS-

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

MODE descriptors account for hydrophobicity/polarity, electronic and steric features of molecules on the basis of bond weights. The bond weights used in this work describe the noctanol/water partition coefficient (H), polar surface, polarizability, Gasteiger-Marsilli charges, van der Waals atomic radii and molar refraction. Selected phenolic compounds differing significantly in their antioxidant activity and hydrophilic-lipophilic balance were tested using this TOPS-MODE QSAR approach according to Estrada [41–43] (27 of these phenolic compounds are presented in this paper).

2.4 Quantum-chemical calculations – determination of the O-H bond dissociation energy The most important factor was considered for the characterization of the free-radical scavenging activity of antioxidants, for instance, the difference in the heat of formation (DHOF) of the antioxidant molecule and its corresponding radical, i.e. the O-H bond dissociation energy. Due to the large size of the antioxidant molecules, semi-empirical quantum-chemical methods were employed in most cases; DFT calculations (B3LYP, and basic sets 6-31* and 6-31**) were also applied [5, 6, 26, 44, 45].

3 Results and discussion Figure 1 presents the structure of all phenolic antioxidants in this study. They were separated into four main groups, based on their structural characteristics. The following fatty acid composition (in wt-%) was determined: 16:0, 6.7%; 18:0, 3.6%; 18:1, 25.1%; 18:2, 63.7%; 20:0, 0.2%; 22:0, 0.7%.

3.1 Model 1 A comparative analysis for the radical-scavenging activity (%RSAexp) of 20 phenolic antioxidants (11 flavonoids, 4 cinnamic acids, 3 chalcones and 2 standard antioxidants) was made based on the experimental results obtained with the DPPH rapid test at the ratios of [AH]/[DPPH] = 0.15 and 0.25 (Fig. 2). It can be seen that (a) antioxidants with the

Figure 2. Bar graph of the radical-scavenging activity (%RSA) obtained by the DPPH radical test of various antioxidants. Ratios between the antioxidant and DPPH concentrations: [AH]/[DPPH] = 0.15 and 0.25.

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

highest activity (% RSA .70) for ratios of 0.25 and 0.15 (% RSA .50) are compounds with a catecholic structure (Qu, Qu-3Gu, Ru, CA, Ch3); (b) antioxidants with a moderate activity (% RSA = 30–60) for ratios of 0.25 and 0.15 (% RSA = 10–40) are monophenolic antioxidants with an electrondonating substituent in the ortho position (a-TOH, SA, FA) and with a free 3-OH phenolic group of flavonoids (Kf, Isrh); and (c) antioxidants with the lowest activity (% RSA ,20) for ratios of 0.25 and 0.15 (% RSA ,10) are monophenols without an ortho substituent (p-CumA, Ch1, Ch2) and 3-O-glycosides of Kf and Isorh (Kf-3Glu, Isrh-3-Glu, Kf-3-Rut, Isrh3-Rut, Kf-3-Cum-Glu, Isrh-3-Cum-Glu). An unexpected result is obtained for BHT with a low % RSAexp. From this comparative analysis, it could be concluded that all phenolic compounds with catecholic moieties demonstrate much higher activity as radical scavengers than those of monophenols. The highest RSA of quercetin and its derivatives is due to the reaction of bimolecular recombination with disproportionation of the semiquinone radicals formed from them (Qu, Qu-3Glu, Ru), which results in regeneration of the initial antioxidant molecule and also in formation of ortho- and para-quinone [21]. The stoichiometry (n) of the reaction between AH and DPPH., i.e. how many radicals (DPPH) were scavenged by one molecule of the antioxidant (AH), depends on the antioxidant’s structure and the mechanism. The following stoichiometries are possible: Stoichiometry n = 1 DPPH. 1 AH ? DPPH2 1 A. A. 1 A. ? A-A ———————————————— 2 DPPH. 1 2 AH ? 2 DPPH2 1 A-A Stoichiometry n = 2 DPPH. 1 AH ? DPPH2 1 A. DPPH. 1 A. ? DPPH-A ——————————————— 2 DPPH. 1 AH ? DPPH2 1 DPPH-A Stoichiometry n = 3 DPPH. 1 AH ? DPPH2 1 A. DPPH. 1 A. ? DPPH-A DPPH. 1 DPPH-A ? DPPH2-Q ————————————————– 3 DPPH. 1 AH ? DPPH2 1 DPPH-Q Stoichiometry n = 4 DPPH. 1 AH ? DPPH2 1 A. DPPH. 1 A. ? DPPH-A (P1) DPPH. 1 P1 ? DPPH2 1 P2 DPPH. 1 P2 ? DPPH2 1 P3 ——————————————— 4 DPPH. 1 AH ? 3 DPPH2 1 P These mechanisms of action of AH with DPPH. explain the higher stoichiometry (n = 3.6) for Qu compared to the calculated theoretical value of 2, and for Kf and Isrh as .1 (n = 1.5 and 1.6, respectively) [6, 45]. These higher values of stoichiometry confirm the results obtained earlier [46, 47]. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Phenolic antioxidants

1079

3.2 Model 2 The oxidation mechanism of model hydrocarbons (RH) (cumene or DPM), initiated by the free-radical initiator AIBN in the absence (non-inhibited) and in the presence (inhibited) of antioxidants (AH), may be described with the following kinetic scheme (see Table 1). The comparison of the chain-breaking activity of 15 phenolic antioxidants (5 chalcones, 4 flavonoids, 2 standard antioxidants and 4 cinnamic acids) under the same experimental conditions and antioxidant concentrations is presented in Table 2. The results obtained by both kinetic models (cumene and DPM) demonstrated that the antioxidant activity (determined as the rate constant of the reaction with peroxyl radicals, kA) decreases in the following sequence: Cumene: a-TOH (1.36105) . CA (8.36104) . SA (2.56104) . Ch4 (1.56104)  BHT (1.46104) = Ru (1.46104) = Ch5 (1.36104) . FA (0.86104) . p-CumA (0.76104); and DPM: Lu (2.26107) = Qu (2.16107) . Ch3 (1.36107) . a-TOH (8.56106) . CA (4.66106) . Kf (1.06106) . Ch1 (1.16105) . FA (3.66104) = BHT (3.66104) .. pCumA (2.26104) . Ch2 (9.66103). The obtained kA for the same phenolic antioxidant is much lower in case of cumene, due to the lower rate constant of the termination of the cumene peroxyl radicals (kt) [30, 38]. In case of DPM as oxidizing substrate, the most powerful antioxidants are those with a catecholic structure of the phenolic groups: Qu, Lu, CA, Ch3 (kA < 107), and also a-TOH and Kf (kA < 106) and Ch1 (kA < 105). In case of cumene as oxidizing substrate, the most powerful antioxidant is a-TOH (kA < 105), but its value is much lower than that determined with DPM. CA, Ru, BHT, SA, Ch4 and Ch5 demonstrated a moderate activity (kA < 104) [26, 31]. It could be concluded that only the antioxidants with weak and moderate rate constants of reactions with peroxyl radicals should be studied using a cumene as oxidizing substrate. In case of DPM as oxidizing substrate, higher values of kA were obtained for the antioxidants with moderate and weak activity (see data for cumene of the same antioxidants). The antioxidants with highest activity, kA, demonstrated almost twofold lower values with cumene than with DPM, which makes DPM a much better substrate for studying the powerful antioxidants.

3.3 Model 3 Lipid autoxidation is a radical process involved in a chain reaction including induction, propagation and termination steps. During the induction period, various lipid peroxyl radicals are formed. These highly reactive chemical species produce hydroperoxides (LOOH) during the propagation stage. Termination consists of the association of two radicals together to form more stable products. The basic kinetic scheme of lipid (LH) autoxidation in the absence (non-inhibited oxidation) and in the presence (inhibited oxidation) of phenolic antioxidants (AH) is presented in Table 1. Figwww.ejlst.com

1080

V. D. Kancheva

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Table 1. Kinetic models and their characteristics. Kinetic models and characteristics

Model 2

Model 3

Name Initiator Oxidizing substrate Experimental conditions Kinetic scheme of oxidation In the absence of the antioxidant Chain generation Chain propagation Chain termination Chain branching 1 Chain branching 2 Chain branching 3 In the presence of the antioxidant Inhibition1 Inhibition 2 Inhibition 3 (Recombination) Inhibitipn 4 (Disproportionation) Side reactions of phenolic antioxidant of phenoxyl radical

Kinetic CL method AIBN Cumene/DPM T = 60 7C/T = 50 7C; homogeneous solution Initiated oxidation Non-inhibited process AIBN 1 RH(O2) ? RO2. (RIN) RO2.1RH (1O2) ? ROOH 1 RO2. (kp) RO2. 1 RO2. ? P* ? P 1 hn (kt)

Inhibited process RO2. 1 AH ? ROOH 1 A. (kA) RO2.1 A. ? ROO-A (kA’) A. 1 A. ? A-A (kR) A. 1 A. ? AH 1 Q (kD)

Lipid autoxidation Without an initiator TGSO kinetically pure T = 80 7C; homogeneous solution Autoxidation Non-inhibited process LH 1 O2 (Y) ? LO2. (RIN) LO2. 1 LH (1O2) ? LOOH 1 LO2. (kp) LO2. 1 LO2. ? P (kt) LOOH (1O2) ? d1 LO2. 1 P1 (kb1) LOOH 1 LH (1O2) ? d2 LO2. 1 P2 (kb2) LOOH 1 LOOH (1O2) ? d3 LO2. 1 P3 (kb3) Inhibited process LO2. 1 AH ? LOOH 1 A. (kA) LO2. 1 A. ? LOO-A (kA’) A. 1 A. ? A-A (kR) A. 1 A. ? AH 1 Q (kD)

AH 1 ROOH ? d1 RO2. 1 PA A. 1 ROOH ? RO2. 1 AH (k-A) A. 1 O2 ? AO2. A. 1 RH (1O2) ? AH 1 RO2. A. 1 ROOH (1O2) ? AH 1 RO2. Constant and well controlled Constant and well controlled R0 = kp[RH](RIN/kt)0.5 RA = kp[RH]RIN/nkA[AH]0 n0 = R0/RIN nA = RA/RIN

AH 1 LOOH ? d1 LO2. 1 PA A. 1 LOOH ? LO2.1 AH (k-A) A. 1 O2 ? AO2. A. 1 LH (1O2) ? AH 1 LO2. A. 1 LOOH (1O2) ? AH 1 LO2. Accelerated during the process Accelerated during the process R0 = kp[LH](RIN/kt)0.5 RA = kp[LH]RIN/nkA[AH]0 n0 = R0/RIN nA = RA/RIN

ID = n0/nA IP = n[AH]0/RIN kA

ID = R0/RA IP = n[AH]0/RIN RAE = (IPA – IP0)/IP0

Rate of initiation (RIN) Rate of oxidation (R0) and (RA) Rate of non-inhibited oxidation (R0) Rate of inhibited oxidation (RA) Oxidation chain length (n0) and (nA) Chain length of non-inhibited oxidation (n0) Chain length of inhibited oxidation (nA) Inhibition degree (ID) Induction period (IP) Antioxidant efficiency

ures 3 and 4 present the chain-breaking antioxidant activity of the 31 studied phenolic compounds (6 flavonoids, 15 cinnamic acids, 3 standard antioxidants, 3 chalcones and 4 coumarins), using the main kinetic parameters, RAE and ID, during lipid autoxidation at the same antioxidant concentration (0.1 mM). It has been established (Fig. 3) that the antioxidant efficiency (RAE) and the inhibition degree (ID) of the tested compounds (groups A and B) at a concentration of 0.1 mM decrease in the following sequences: Groups A and B (RAE): CA (8.8) . a-TOH (4.5) . HQ (3.3) . SA (3.2) . PHC (2.6) . N9 (2.0) . BHT(1.5) . N8 (1.4) . N4 (1.3) . DPHC (1.2) = N7 (1.2) . N3 (1.1) = N1 (1.1) . N5 (0.8) = N2 (0.8) . N6 (0.7) = FA (0.7) . p-CumA (0); Groups A and B (ID): CA (11.0) . N3 (6.8) . N4 (6.4) = N6 (6.4) . aTOH (6.0) = N9 (6.0) . PHC (4.9) . N1 (4.6) .N7 (4.4) . N2 (3.5) . DPHC (3.4) . HQ (3.3) . N8 (3.2) . N5 (3.1) . SA (2.9) . FA (1.9) . BHT (1.7) . p-CumA (1.0). The antioxidant efficiency (RAE) and the inhibition degree (ID) of the tested compounds (groups C–E) at a con© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

centration of 0.1 mM decrease in the following sequence (Fig. 4): Groups C, D and E (RAE): Ch3 (10.7) . Qu (6.6) . Cum1 (4.7) . Cum0 (3.0) . Ch1 (1.8) . Lu (1.7) . (Qu-7) (1.2) . Ru (1.1) . Cum2 (0.3) . (Lu-7) (0.5) . (Qu-3) (0.1) . Ch2 = Cum3 (0); Groups C, D and E (ID): Ch3 (25) . Qu (22.5) . Cum0 (11.2) . Ru (8.0) . Cum1 (6.1) . Lu (5.5) . (Qu-7) (2.2) . Ch1 (1.6) . Cum2 (1.3) = (Lu-7) (1.3) . Cum3 (1.0) . Ch2 (0.8) . (Qu-3) (0.5). It can be seen from the presented data that the antioxidants Qu, Ch3, CA, Cum0, Ru, Lu and Cum1 (all of them have a catecholic moiety) demonstrate the strongest antioxidant efficiency (RAE) and the highest values of their inhibition degree (ID). The latter can be explained with the possibility of these biphenolic groups to produce semiquinone radicals (QH.) during the reaction between bi- and polyphenolic antioxidants with a catecholic structure (QH2) and lipid peroxyl radicals (LO2.): QH2 1 LO2. ? QH. 1 LOOH. Semiquinone radicals recombine with the highest rate constant (almost 109 M– 1 –1 s , [38]) by two possible reaction mechanisms: (1) bimolewww.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Phenolic antioxidants

1081

Figure 3. The main kinetic parameters, RAE and ID, for the phenolic antioxidants of group A (cinnamic acid derivatives) and of group B (standard antioxidants) (a) and of group C (flavonoids), of group D (chalcones) and of group D (coumarins) (b), determined during TGSO autoxidation at 80 7C and 0.1 mM antioxidant concentration. For abbreviations see Fig. 1.

Figure 4. Effect of the antioxidant concentration ([AH] = 0.1 mM and 1.0 mM) on the main kinetic parameter RAE (a) and ID (b) for selected phenolic antioxidant, determined during TGSO antioxidation at 80 7C. For abbreviations please see Fig. 1.

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.ejlst.com

1082

V. D. Kancheva

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Table 2. Kinetic parameters characterizing the chain-breaking activity of the antioxidants studied during the oxidation of cumene and DPM calculated from the kinetics of CL. No.

Antioxidants

kA (Ms)0.5

Substrate

Ref.

1

p-CumA

4

FA

11 15

SA CA

17

BHT

18

TOH

19 21 25 28 34 35 36 37 38

Lu Qu Kf Ru Ch1 Ch2 Ch3 Ch4 Ch5

0.76104 2.26104 0.86104 3.66104 2.56104 8.36104 4.66106/1.56107 1.46104 3.66104 1.36105 8.56106 2.26107 2.16107 1.06106 1.46104 1.16105 9.66103 1.36107 1.56104 1.36104

Cumene DPM Cumene DPM Cumene Cumene DPM Cumene DPM Cumene DPM DPM DPM DPM Cumene DPM DPM DPM Cumene Cumene

[31] [5, 26] [31] [5, 26] [31] [31 [5, 26, 30] [21, 31] [30] [31] [30] [30] [30] [30] [21] [5, 26] [5, 26] [5, 26] [5, 24] [5, 24]

cular recombination with dimerization of semiquinone radicals (path A), 2 QH. ? QH-QH; and (2) bimolecular recombination with homo-disproportionation of semiquinone

radicals (path B), 2 QH. ? QH2 1 Q. Path B is very important, due to the possibility to regenerate the initial molecule of the phenolic compounds [21, 48]. In case of flavonoid compounds (QH2), the ortho- and para-flavonoid quinones are formed and the initial molecule of QH2 is regenerated during this reaction of homo-disproportionation of two QH. as shown in Fig. 5. The effect of the antioxidant concentration ([AH] = 0.1 and 1.0 mM) is presented in Fig. 4 for 17 selected compounds: 7 cinnamic acids, 6 flavonoids and 4 coumarins. It is seen that the most important kinetic parameters of lipid autoxidation, RAE and ID, at [AH] = 1.0 mM increase in all cases. However, RAE decreases in case of Lu-7and Qu-7 and ID decreases in case of Ru and Qu-7. The following new sequences for the RAE and ID decreases at 1.0 m M of AH were found: RAE: Qu (38.5) . CA (33.0) . Cum1 (16.9) . N9 (14.5) . N7 (14.1) . N8 (14.0) . SA (13.6) . Cum0 (12.0) . Cum2 (4.9) . Lu (3.8) .Qu-3 (2.2) . FA (2.0) = Cum3 (2.0) . Ru (1.9) . p-CumA (1.0) . Qu-7 (0.2) . Lu-7 (0.1). ID: Qu (80) . Cum1 (56.0) . CA (42) . Cum0 (28) . N7 (9.6) . Lu (8.6) . SA (8.2) . N8 (8.1) = N9 (8.1) . Cum2 (7.0) . Lu-7 (5.5) . Cum3 (4.3) . Ru (4.2) . Qu-3 (3.6) . p-CumA (2.0) . FA (1.9) . Qu-7 (1.2). In order to study the possible synergism between two phenolic antioxidants, various binary mixtures of two antioxidants were tested and compared: (a) two flavonoid agly-

Figure 5. Path B. Bimolecular recombination with disproportionation of semiquinone radicals from quercetin (QH.) and formation of ortho-quinones (o-Q) or para-quinones (p-Q).

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Phenolic antioxidants

cones (Qu 1 Lu); (b) flavonoid aglycon and glycoside (Qu 1 Ru); (c) two flavonoid glycosides (Qu-Lu); (d) antioxidants with a catecholic structure and a-TOH (Qu 1 aTOH, Ru 1 a-TOH, Cum1 1 a-TOH, Cum2 1 a-TOH). Table 3 presents the main results obtained, as well as the synergism and/or antagonism obtained for the binary mixtures. It is found that only binary mixtures with a-TOH demonstrate synergism. Binary mixtures without tocopherol do not show any synergism; even an antagonism between them may be observed. The following new sequences of % synergism and of the inhibiting activity (IP112) are: Group F: Binary mixtures (%Synergism): Ru 1 a-TOH . Qu 1 a-TOH . Cum11a-TOH. Group F: Binary mixtures (IP112): Qu 1 a-TOH . Ru 1 aTOH . Cum2 1 a-TOH . Cum1 1 a-TOH . Qu 1 Ru . Qu 1 Lu . Qu-Lu. This synergism of binary mixtures with tocopherol may be explained taking into account that the catecholic moiety of the flavonoid and coumarine molecules allows the formation of semiquinone radicals. These semiquinone radicals may regenerate the initial antioxidant molecule during the reaction of bimolecular recombination with homo- and cross-disproportionation of semiquinone radicals [38, 48]: (a) Homo-disproportionation reaction: 2 QH. ? QH2 1 Q regeneration of QH2 (flavonoid or coumarine) and 2 a-TO. ? a-TOH 1 a-T=O regeneration of a-TOH. (b) Cross-disproportionation reaction: QH. 1 a-TO. ? QH2 1 a-T=O regeneration of QH2 and a-TO. 1 QH. ? aTOH 1 Q regeneration of a-TOH.

1083

These reactions demonstrate that, during the oxidation process, the initial molecules of flavonoids and a-TOH are regenerated by different mechanisms and this makes these binary mixtures the most powerful antioxidant compositions. It must be noted that a-TOH regeneration is evidently responsible for the highest % synergism obtained. For binary mixtures of Qu 1 Lu, Qu 1 Ru, and Qu-Lu at the same equimolar (1 : 1) concentrations (0.1 and 0.5 mM), neither synergism nor an additive effect (IP112 = IP1 1 IP2) was observed [22]. In this case, only reactions of homo-disproportionation are possible and only the initial molecule of the corresponding flavonoid antioxidants is regenerated. Evidently, the lack of a tocopherol molecule during the lipid autoxidation is the reason that no synergisms were found in these binary systems.

3.4 Model 4a – QSAR statistical analysis The relationship between the structure and radical-scavenging activity of 16 flavonoids (5 aglycones, 6 glucosides, 3 rutinosides and 2 p-coumaroyl-glucosides) was studied. The statistical analysis of the theoretically calculated (% RSAcalc) and the experimentally obtained (% RSAexp) values for the radical-scavenging activities of all studied flavonoids is presented in Table 4. The experimental data are presented as % RSA obtained by using the rapid DPPH radical test. Data for an additional 7 flavonoids were taken from the paper of Amic et al. [40] who calculated the theoretical values of % RSAcalc using the statistical model: % RSAcalc = 3.954 1 75.950

Table 3. Inhibiting the efficiency of various binary mixtures of two antioxidants in equimolar concentrations (0.1 and 0.5 mM) at a 1 : 1 ratio and 80 7C during TGSO autoxidation (for SA 1 a-TOH, TGL autoxidation, 100 7C). % Synergism = [IP112 – (IP1 1 IP2)/(IP1 1 IP2)]6100, %. Binary mixtures (1 : 1)

[AH] [mM]

IP112 [h]

IP1 [h]

IP2 [h]

Synergism Antagonism

% Synergism

Refs.

Quercetin 1 Luteolin (Qu 1 Lu)1 (Qu 1 Lu)5 Quercetin 1 Rutin (Qu 1 Ru)1 (Qu 1 Ru)5 Qu-7Glu 1 Lu-7Glu (Qu-7 1 Lu-7)5 (Qu-7 1 Lu-7)10 Quercetin 1 a-Tocopherol (Qu 1 a-TOH)1 Rutin 1 a-Tocopherol (Ru 1 a-TOH)1 7,8-diOH-Cumarin 1 a-TOH (Cum2 1 a-TOH)1 6,7-diOH-Cumarin 1 a-TOH (Cum1 1 a-TOH)1 Sinapic acid 1 a-TOH (SA 1 a-TOH)1*

0.1 0.5

7.5 6 0.4 12.3 6 0.5

9.9 6 0.9 24.5 6 0.6

2.2 6 0.2 6.3 6 0.4

Antagonism Antagonism

– –

[22] [22]

0.1 0.5

8.3 6 0.8 21.5 6 0.6

9.9 6 0.9 24.5 6 0.6

2.7 6 0.2 2.8 6 0.2

Antagonism Antagonism



[22] [22]

0.5 1.0

2.0 6 0.2 2.1 6 0.2

1.5 6 0.2 1.0 6 0.2

1.8 6 0.2 1.4 6 0.2

Antagonism Antagonism



[22] [22]

0.1

29.7 6 1.5

9.9 6 0.9

10.5 6 0.9

Synergism

46%

[22]

0.1

24.9 6 1.5

2.7 6 0.2

10.5 6 0.9

Synergism

87%

[22]

0.1

14.2 6 0.9

2.0 6 0.2

10.5 6 0.9

Synergism

14%

[5, 27]

0.1

12.7 6 0.9

7.1 6 0.5

10.5 6 0.9

Antagonism



[5, 27]

0.1

45.0 6 1.0

8.5 6 0.5

21.0 6 1.5

Synergism

52%

[36]

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.ejlst.com

1084

V. D. Kancheva

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Table 4. Substitution pattern of the series of flavonoids examined for their radical-scavenging activity: % RSAexp and % RSAcalc.

Flavonoids

% RSAexp

% RSAcalc

R3

R5

R7

R3’

R4’

Ref.

Kaempferol (Kf) Kf-3Glu Kf-3Rut Kf-3Glu-Cum Kf Kf-3,7diGly Apigenin Isorhamnetin (Isrh) Isrh-3Glu Isrh-3Rut Isrh-3Glu-Cum Quercetin (Q) Q-3Glu Q-3Rut Quercetin (Q) Q-3,7diGly

54.4 6 0.8 1.7 6 0.1 0.7 6 0.2 0.8 6 0.4 93.5 70.6 0.7 19.2 6 1.4

88.4 12.75 12.75 12.75 88.40 12.75 12.75 88.40

OH OGlu ORut OGlu-Cum OH OGly H OH

OH OH OH OH OH OH OH OH

OH OH OH OH OH OGly OH OH

H H H H H H H OCH3

OH OH OH OH OH OH OH OH

[21] [21] [21] [21] [40] [40] [40] [21]

4.4 6 0.9 2.1 6 0.4 2.8 6 0.5 62.2 6 4.4

12.75 12.75 12.75 88.40

OGlu ORut OGlu-Cum OH

OH OH OH OH

OH OH OH OH

OCH3 OCH3 OCH3 OH

OH OH OH OH

[21] [21] [21] [21]

63.9 6 2.8 59.0 6 1.4 89.8 86.8

88.40 88.40 88.40 88.40

OGlu ORut OH OGly

OH OH OH OH

OH OH OH OGly

OH OH OH OH

OH OH OH OH

[21] [21] [40] [40]

Glu, b-D-Glucoside; Rut, rutinoside; Glu-Com, p-coumaroyl-glucosides; Gly, glycosides [40].

I3’,4’-diOH or 3-OH 1 8.499 I5-OH. This model divides the compounds into two main groups [21, 40]: (i) active (% RSAcalc = 88.4%) for all flavonoids with catecholic structure in ring B and free 3-OH and 5-OH groups, i.e. for these compounds I = 1 (for 5 flavonoids: Qu, Qu-3-Glu, Ru, Kf and Isrh); (ii) inactive (% RSAcalc = 12.45%). This value is found only for flavonoid compounds without catecholic structure and without free 3-OH (for 6 compounds: Kf-3-Glu, Kf-3-Rut, Kf-3-Cum-Glu, Isrh-3-Glu, Isrh-3-Rut, Isrh-3-Cum-Glu). A very good correlation between experimental and theoretical RSA was found. The new data obtained are in agreement with those published earlier about structure-radical-scavenging activity relationships [49, 50].

3.5 Model 4b – TOPS-MODE QSAR Phenolic compounds with a wide range of antioxidant activities (presented as inhibition degree, ID) were chosen for the TOPS-MODE QSAR study. The ID parameter is of great importance for the practice because it demonstrates the pos© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

sibility of the antioxidant molecule to significantly reduce the length of the lipid autoxidation chain and to terminate the oxidation process. In this comparative analysis, 6 flavonoids, 15 cinnamic acid derivatives, 3 standard antioxidants and 3 chalcones were included with a wide range of inhibiting degree values and then separated into three main groups: Group 1, compounds with the highest values of their inhibition degree, strong activity, i.e. ID .8.0;. in this group, 3 compounds: CA, Qu, and Ch3 were included. Group 2, compounds with moderate values of their inhibition degree, moderate activity, i.e. 8.0. ID .3.0; this group included 14 compounds: PHC, N1, N2, N3, N4, N5, N6, N7, N8, N9, HQ, a-TOH, Lu, Ru. Group 3, compounds with the lowest values of their inhibition degree, weak activity or inactive, i.e. ID ,3.0; this group consisted of 10 compounds: DPHC, SA, Qu-7, FA, BHT, Lu-7, Ch1, Ch2, Qu-3, p-CumA. The obtained theoretical model demonstrates an excellent agreement between the experimental and the predicted inhibition degree of the studied compounds, using the following theoretical parameters of the model: a catecholic moiety, a dipole www.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Phenolic antioxidants

1085

moment component and a steric effect component. The classification percent correct for the very active compounds is 100%; percent correct for the moderate active group is 92.7%; percent correct for the weakly active/inactive group is 83.3%, and percent correct total is 89.65% [6, 43].

method, DPPH test and PM3 calculations) Chalcone Ch2 (3hydroxy chalcone) without a catecholic structure and ortho substituent showed the lowest activity by all methods applied. However Ch1 (4-hydroxy chalcone) has a moderate activity against peroxyl radicals (CL method).

3.6 Model 4c – Quantum chemical calculations

3.7 Comparison of experimental and theoretical parameters

The highest radical-scavenging activity of quercetin derivatives correlates with quantum-chemical calculations [44, 45], indicating that, in case of a catecholic structure of ring B, the rest of the molecule does not affect its ability to scavenge free radicals. The latter can be explained with the possible formation of ortho-quinone as a result of a homo-disproportionation reaction of semiquinone radicals, formed in ring B. The important role of the catecholic structure in ring B of a flavonol was demonstrated in case of kaempferol and isorhamnetin, the antiradical capacities of which are significantly lower (no formation of o-quinone is possible) than those for quercetin and its 3-O-glycosides. It has been demonstrated that the substitution of the 3-OH hydrogen with a sugar moiety (glucose, rutinose and/or p-coumaroylglucose) in case of kaempferol and isorhamnetin leads to a strong decrease of their antiradical activity. The poor radicalscavenging activities of these compounds could be explained by the substitution of the 3- and 3’-hydroxyl groups by sugar and OCH3 or H groups. Thus, these compounds cannot be oxidized to o- and p-quinones. The 3-OH group fixes the B ring by a hydrogen bond formation with the 4-keto group of ring C, and when the flavonoid radical is formed, electron delocalization over the whole molecule is possible. As a result, a higher stability of the formed radical was observed and also higher antiradical activity of Qu than those of Qu-3, Ru and Lu. In case of Qu-3, Ru, and Lu, the formation of semiquinone radicals in the B ring is possible. However, the lack of a 3-OH group (Lu) or glycoside moiety (Qu-3 and Ru) makes electron delocalization impossible because ring B is twisted [45]. As a result, their chain-breaking antioxidant activity decreases. In case of Kf and Isrh without a catecholic structure (FlOH) in the B ring, only the formation of the para-flavonoid quinone (p-Q) is possible in the reaction of homo-disproportionation of the flavonoid radicals (2 FlO. ? FlOH 1 Fl=O). For this reason, the substitution of the 3-OH group with a glycoside moiety strongly decreases their antioxidant activity [5, 6, 45, 48]. We have used a semi-empirical PM3 method as the most reliable for molecules with filled non-bonding p-orbitals on adjacent atoms [5, 6, 26]. The following sequence of DH(OH) (kcal/mol) was found for the chalcones: Ch3 (81.42) < CA (81.46) , FA (83.91) , Ch1 (86.93) < Ch2 (86.95) , pCumA (87.29). The main conclusion of this comparative analysis is that both compounds with the catecholic structure in the B ring (CA and Ch3) demonstrate the strongest activity assessed by all methods applied (experimental: lipid autoxidation, CL © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Table 5 presents structure-activity relationships. The main structural fragments of phenolic antioxidants and their experimental (antiradical and chain-breaking antioxidant) and theoretical (predictable) activities were compared using different methods: DPPH test (% RSAexp), CL experiments (kA), lipid autoxidation (RAE, ID), statistical analysis (% RSAcalc), TOPS-MODE QSAR (ID predictable) and quantum-chemical calculations (bond dissociation energy). All compounds were separated into three main groups: (a) with strong activity; (b) with moderate activity; and (c) with low activity. A good correlation between % RSAexp and H-donating ability could be seen, calculated as % RSAcalc, and with H bond dissociation energy as DO-H [26, 40]. The best correlation was found between experimental inhibition degrees with the predicted values found by using the TOPSMODE QSAR approach. There is a good correlation also between the absolute rate constants of the antioxidants with peroxyl radical, determined using the CL approach, with the H-bond-donating ability and with antioxidant efficiency during lipid autoxidation.

4 Conclusions The radical-scavenging activity (% RSA) differs significantly from the chain-breaking antioxidant activity. RSA towards the DPPH radical gives information only about the H-donating capacity of the studied compounds and some preliminary information for their possibility to be used as antioxidants. This comparative study showed a good correlation between experimental and theoretical/predictable RSA. Antioxidant activity is the capacity of a compound to shorten the oxidation chain length as a result of its reaction with peroxyl radicals. For this reason, we mean by antioxidant activity the chainbreaking activity of the compounds. This comparative study showed an excellent correlation between the experimental inhibition degree of phenolic compounds and their predictable activity by using the TOPS-MODE QSAR approach. The obtained theoretical model demonstrated an excellent agreement between the experimental and the predictable inhibition degree of the studied compounds, using the following theoretical parameters of the model: a catecholic moiety, a dipole moment component and a steric effect component. It has been demonstrated that phenolic compounds with catecholic moiety are the most powerful scavengers of free www.ejlst.com

Main structural fragment

Moderate

Weak

Weak

SA, N7, N8, N9

BHT

PHC

II.2 Di-methoxyphenols

II.3 tert-Butyl phenols

II.4 Mono-prenylated phenols

Moderate Strong/weak

Strong Strong

DPPH test (% RSAexp)

FA, N1, N2, N3, N4, N5, N6 Isrh, Isrh-3-Glu, Isrh-3-Rut, Isrh-3-Cum-Glu

CA, Lu, Qu, Ch3, Cum0, Cum1 Lu-7, Qu-7, Qu-3-Glu, Qu-3-Rhm, Ru

Compounds

II.1 Mono-methoxyphenols

II. Monophenolic antioxidants

With a catecholic moiety

I. Bi- and polyphenolic antioxidants

Phenolic antioxidants

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Moderate

Moderate

Moderate

Moderate nd

Strong Weak/moderate

Lipid autoxidation (RAE, ID)

Moderate

Moderate

Moderate

Moderate Strong/weak

Strong Strong

Theoretical QSAR

V. D. Kancheva

nd

Moderate

Moderate

Weak nd

Strong Moderate

CL (kA)

Activities/methods

Table 5. Correlations between the main structural fragments of phenolic antioxidants and their experimental (antiradical and antioxidant) and predictable (theoretical) activities. 1086 Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

www.ejlst.com

Weak Strong Weak Weak Strong nd Weak Strong Weak

Weak Strong Weak

Moderate Moderate nd Moderate

Lipid autoxidation (RAE, ID) CL (kA) DPPH test (% RSAexp)

Activities/methods

Theoretical QSAR

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

Phenolic antioxidants

1087

radicals and that they may be used as effective chain-breaking antioxidants. The highest antiradical and antioxidant activities of phenolic antioxidants with catecholic moiety are explained by the possible mechanism of homo-disproportionation of their semiquinone radicals formed. Thus, the regeneration of the antioxidant molecule during the oxidation process is possible. It has been found that the rest of the antioxidant molecule is of no significance for the radical-scavenging activity of phenolic antioxidants with a catecholic moiety. This comparative analysis demonstrates also that the activity of the antioxidant significantly depends not only on its structural characteristics but also on the properties of the substrate being oxidized and the experimental conditions applied. Structural characteristics of the complex system, oxidizing substrate-antioxidant, must be considered. The synergism has been found only for the binary mixtures of antioxidants with catecholic structure and tocopherol in equimolar concentrations. On the basis of this comparative analysis, the most effective individual antioxidants and binary mixtures are proposed for highest and optimal lipid oxidation stability.

p-CumA, Ch1, Ch4 Kf Kf-3-Glu, Kf-3-Rut, Kf-3-Cum-Glu

DPHC

The author has declared no conflict of interest.

II.6,7 Simple monophenols

References

II.5 Di-prenylated phenols

Phenolic antioxidants

Table 5. Continued.

This study was supported in part by Grant BIn4/04 awarded by the National Fund of Scientific Research in the Bulgarian Ministry of Science and Education. The author gratefully acknowledges the possibility to run the TOPS-MODE theoretical study and the statistical analysis at the University of Santiago de Compostela, Spain (Bilateral Joint Research Project between Spain and Bulgaria: “New healthy antioxidants for improvement human nutrition and health”), and quantum-chemical calculations at the Aristotle University of Thessaloniki (Grant of Foundation IKYGreece), and the anonymous reviewers for their helpful comments on the manuscript.

Conflict of interest statement

Main structural fragment

Compounds

Acknowledgments

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

[1] O. Potterat: Antioxidants and free radical scavengers of natural origin.Curr Org Chem. 1997, 1, 415–440. [2] E. B. Burlakova: Bioantioxidants. Molecular cell biophysics. Russ Chem J. 2007, 51, 3–12 (in Russian). [3] F. Shahidi: Natural antioxidants, a review. In: Natural Antioxidants Chemistry, Health Effects and Applications. Ed. F. Shahidi, AOCS Press, Champaign, IL (USA) 1999, pp. 1–11. [4] V. G. Kondratovich, V. D. Kortenska, Z. S. Kartasheva, N. V. Yanishlieva, I. R. Totzeva, M. I. Boneva, O. T. Kasaikina: Kinetics of lipid oxidation and lipid hydroperoxide decomposition in the presence of amphiphilic compounds. In: Perwww.ejlst.com

1088

V. D. Kancheva

oxides at the Beginning of the Third Millenium: Synthesis, Properties, Application. Eds. V. A. Antonovsky, O. T. Kasaikina, G. E. Zaikov, Nova Science Publ., New York, NY (USA) 2004, pp. 261–267. [5] V. D. Kancheva: Effective antioxidants from natural origin to increase the lipid oxidation stability. Bulgarian-Indian Workshop on Medical Sciences Conference, 16th October 2008, Sofia (Bulgaria), Book of Abstracts, p 3. [6] V. D. Kancheva: Structure-activity relationship of phenolic antioxidants of natural origin. 5th Euro Fed Lipid Congress and the Symposium of the Nordic Lipid forum: “Oils, Fats and Lipids – from Science to Applications, Innovations for a better World”, 16–19 September 2007, Gothenburg (Sweden), Book of Abstracts, p 307, OXI-022. [7] W. Bors: Antioxidants in dietary plants. DGQ 35 Vortragstagung, Karlsruhe (Germany) 2000, pp. 3–16. [8] V. D. Kortenska-Kancheva, V. S. Bankova: A review on the antioxidant activity of propolis and propolis constituents from different geographic origin. Recent Prog Med Plants. 2006, 14, 81–98. [9] D. Huang, B. Ou, R. L. Prior: The chemistry behind antioxidant capacity assays. J Agric Food Chem. 2005, 53, 1841– 1856. [10] N. Yanishlieva-Maslarova: Inhibiting oxidation. In: Antioxidants in Food. Practical Applications Eds. J. Pokorny, N. Yanishlieva, M. Gordon, CRC Press, Boca Raton, FL (USA), Woodhead Publishing, Cambridge (UK) 2001, pp. 22–70. [11] V. A. Roginsky, E. A. Lissi: Review of methods to determine chain-breaking antioxidant activity in food. Food Chem. 2005, 92, 235–254. [12] V. A. Beljakov, V. D. Kortenska, V. S. Rafikova, N. V. Yanishlieva: Kinetics and mechanism of initiated oxidation of hydrocarbons in the presence of fatty alcohols. Oxid Commun. 1991, 14, 132–138. [13] V. D. Kortenska, N. V. Yanishlieva, V. A. Roginskii: Kinetics of inhibited oxidation of lipids in the presence of 1-octadecanol and 1-palmitoylglycerol. J Am Oil Chem Soc. 1991, 68, 888– 890. [14] V. A. Belyakov, V. D. Kortenska, V. S. Rafikova, N. V. Yanishlieva: Kinetics of the initiated oxidation of model lipid systems. Role of fatty alcohols. Kinet Catal. 1992, 33, 611– 616. [15] P. C. H. Hollmann, J. H. M. de Vries, S. D. Leeuwen, M. J. B. Mengelers, M. B. Katan: Absorption of dietary quercetin glycosides and quercetin in healthy ileostomy volunteers. Am J Clin Nutr. 1995, 62, 1276–1282. [16] W. Bors, S. Kazazic, C. Michel, V. D. Kortenska, K. Stettmaier, L. Klasinc: Methoxyphenols – antioxidant principles in food plants and spices: Pulse radiolysis, EPR spectroscopy and DFT calculations. Int J Quantum Chem. 2002, 90, 969– 979. [17] N. V. Yanishlieva, E. M. Marinova: Activity and mechanism of action of natural antioxidants in lipids. Res Dev Oil Chem. 1998, 2, 1–14. [18] D. Amic, D. Davidovic-Amic, D. Beslo, V. Rastija, B. Lucic, N. Trinajstic: SAR and QSAR of the antioxidant activity of flavonoids. Curr Top Med Chem. 2007, 14, 827–845. [19] H.-F. Ji, G.-Y. Tang, H.-Yu Zhang: Theoretical elucidation of DPPH radical-scavenging activity difference of antioxidant xanthones. QSAR Comb Sci. 2005, 24, 826–830. © 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

[20] J. Nushida, J. Kawasata: DPPH radical scavenging reaction of hydroxy- and methoxy chalcones. Biosci Biotechnol Biochem. 2006, 70, 193–202. [21] V. D. Kancheva, D. Dinchev, M. Tsimidou, I. Kostova, N. Nenadis: Antioxidant properties of Tribulus terrestris from Bulgaria and radical scavenging activity of its flavonoid components. Riv Ital Sost Grasse. 2007, 54, 10–19. [22] V. D. Kancheva, R. Taskova, I. Totseva, N. Handjieva: Antioxidant activity of extracts, fractions and flavonoid constituents from Carthamus lanatus L. Riv Ital Sost Grasse. 2007, 54, 77–86. [23] V. D. Kortenska, M. P. Velikova, N. V. Yanishlieva, I. R. Totseva, V. S. Bankova, M. C. Marcucci: Kinetics of lipid oxidation in the presence of cinnamic acid derivatives. Eur J Lipid Sci Technol. 2000, 104, 19–28. [24] V. D. Kancheva, V. S. Bankova: Chain-breaking antioxidant activity of two new chalcones from propolis of El Salvador in homogeneous and micellar media. Bulg Chem Commun. 2008, 40, 546–555. [25] M. Spasova, V. D. Kortenska-Kancheva, I. Totseva, G. Ivanova, L. Georgiev, Ts. Milkova: Synthesis of cinnamoyl- and hydroxy-cinnamoyl-amino acid conjugates and evaluation of their antioxidant activity. J Peptide Sci. 2006, 12, 369–375. [26] V. D. Kancheva, R. F. Vasilev, I. R. Totseva, D. I. Batovska, G. F. Fedorova: Antioxidant and antiradical activity of some chalcones. Correlations between experimental and theoretical data. 6th Euro Fed Lipid Congress: “Oils, Fats and Lipids in the 3rd Millennium: Challenges, Achievements and Perspectives, 7–10 September 2008, Athens (Greece), Book of Abstracts, p. 435. [27] P. Boranova, L. Saso, V. Kancheva, I. Totseva: Structure-activity relationship of nine dihydroxy coumarins from India. 6th International Conference of the Chemical Societies of the SouthEastern European Countries, 10–14 September 2008, Sofia (Bulgaria), Book of Abstracts, p. 130. [28] N. Nenadis, M. Tsimidou: Observations on the estimation of scavenging activity of phenolic compounds using rapid 1,1diphenyl-2-picrylhydrazyl (DPPH) test. J Am Oil Chem Soc. 2002, 79, 1191–1195. [29] A. T. Lithoxoidou, E. G. Bakalbasis: Liquid-phase theoretical antioxidant activity trend of some cinnamic acid antioxidants. J Am Oil Chem Soc. 2004, 81, 799–802. [30] V. A. Belyakov, V. A. Roginsky, W. Bors: Rate constants for the reaction of peroxyl free radical with flavonoids and related compounds as determined by the kinetic chemiluminescence method. J Chem Soc Perkin Trans II 1995, 2, 2319–2326. [31] I. F. Rusina, M. I. Boneva, O. T. Kasaikina, V. D. Kortenska, N. V. Yanishlieva:Evaluation of the antiradical efficiency of cinnamic acid derivatives using a chemiluminescence method. Oxid Commun. 2004, 27, 562–570. [32] W. W. Christie: Lipid Analysis. 3rd Edn. The Oily Press, Bridgwater (UK) 2003, p. 208. [33] N. Yanishlieva, A. Popov, E. Marinova: Eine modifizierte jodometrische Methode zur Bestimmung der Peroxidzahl in kleinen Lipidproben. Compt Rend Acad Bulg Sci. 1978, 31, 869–871. [34] V. D. Kortenska, N. V. Yanishlieva: Effect of the phenol antioxidant type on the kinetics and mechanism of inhibited lipid oxidation in the presence of fatty alcohols. J Sci Food Agric. 1995, 68, 117–126. [35] V. D. Kortenska, N. V. Yanishlieva, O. T. Kasaikina, I. R. Totzeva, M. I. Boneva, I. F. Rusina: Phenol antioxidant effiwww.ejlst.com

Eur. J. Lipid Sci. Technol. 2009, 111, 1072–1089

[36]

[37]

[38]

[39] [40]

[41]

[42]

ciency in presence of lipid hydroxy compounds in various lipid systems. Eur J Lipid Sci Technol. 2002, 104, 513–519. V. D. Kortenska-Kancheva, N. V. Yanishlieva, K. S. Kyoseva, M. I. Boneva, I. R. Totzeva: Antioxidant activity of cinnamic acid derivatives in presence of a fatty alcohol during the lard autoxidation. Riv Ital Sost Grasse. 2005, 52, 87–92. V. D. Kortenska-Kancheva, V. A. Belyakov: Simulation of lipid oxidation kinetics in various mechanisms of hydroperoxides decomposition. Riv Ital Sost Grasse. 2005, 52, 177– 185. E. T. Denisov, T. G. Denisova: Handbook of Antioxidants. Bond Dissociation Energies, Rate constants, Activation Energies and Enthalpies of Reactions. 2nd Edn. CRS Press, New York, Washington, D.C. (USA) 2001. E. N. Frankel: Lipid Oxidation. The Oily Press, Dundee (UK) 1998. D. Amic, D. Amic-Davidovich, D. Beslo, N. Trinjastic: Structure-radical scavenging activity relationships of flavonoids. Croat Chem Acta CCACAA 2003, 76, 55–61. E. Estrada, E. Molina: QSPR/QSAR by Graph Theoretical Descriptors beyond the Frontiers. In: QSAR/QSPR Studies by Molecular Descriptors. Ed. M. Diudea, NovaScience, New York, NY (USA) 2000, pp. 83–103. E. Estrada, J. A. Quincoces, G. Patlewicz: Creating molecular diversity from antioxidants in Brazilian propolis. Combination of TOPS-MODE QSAR and virtual structure generation. Mol Divers. 2004, 8, 21–33.

© 2009 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Phenolic antioxidants

1089

[43] E. Estrada, V. Kancheva: Structure-antioxidant activity relationship of some phenolic antioxidants of natural origin by using TOPS-MODE QSAR approach, in preparation. [44] H.-Yu Zhang: Selection of theoretical parameter characterizing scavenging activity of antioxidants on free radicals. J Am Oil Chem Soc. 1998, 75, 1705–1709. [45] V. D. Kancheva, M. Tsimidou, D. Dinchev, N. Nenadis, I. Kostova: Structure-antiradical activity relationship of some flavonoids and related compounds, in preparation. [46] W. Brand-Williams, M. E. Cuvelier, C. Berset: Use of free radical method to evaluate antioxidant activity. Lebensm Wiss Technol. 1995, 28, 25–30. [47] D. Rosch, M. Bergmann, D. Knorr, L. W. Kroh: Structureantioxidant efficiency relationships of phenolic compounds and their contribution to the antioxidant activity of sea Buckthorn juice. J Agric Food Chem. 2003, 51, 4233–4239. [48] O. T. Kasaikina, V. D. Kortenska, N. V. Yanishlieva: Effects of chain transfer and recombination/disproportionation of inhibitor radicals on inhibited oxidation of lipids. Russ Chem Bull. 1999, 48, 1891–1896. [49] C. A. Rice-Evans, N. J. Miller, G. Papanga: Structure-antioxidant activity relationships of flavonoids and phenolic acids. Free Radic Biol Med. 1996, 20, 933–956. [50] P. Goupy, C. Dufour, M. Loonis, O. Dangles: Quantitative kinetic analysis of hydrogen transfer reactions from dietary polyphenols to the DPPH radical. J Agric Food Chem. 2003, 51, 615–622.

www.ejlst.com