Initiation and propagation kinetics of inhibited lipid peroxidation

Effect of hydroxytyrosol (HT) and tert-butylhydroquinone (TBHQ) on the kinetics of lipid hydroperoxides (LOOH) accumulation during the initiation and propagation peroxidations of canola and fish oils at 60 °C was studied. The initiation kinetics of the inhibited peroxidation indicated considerable relative activities, A, for HT and TBHQ in the canola (> 3200 and > 27,000, respectively) and fish (> 120 and > 5000, respectively) oils. The critical concentrations of LOOH reverse micelles (CMCL = 33 mM and 57 mM in the canola and fish, respectively, oils) significantly decreased, on average, to about one-third and 8% of the initial values for HT and TBHQ, respectively. Interestingly, the propagation kinetics of the inhibited peroxidation demonstrated that the antioxidants were still able to inhibit peroxidation, so that the relative propagation oxidizability parameter Rn′ was significantly improved to < 0.5 for HT and to < 0.2 for TBHQ in the canola and fish, respectively, oils.

chromatography. Glass columns (40 × 3.5 cm i.d.) for the canola oil were packed by aluminum oxide 60 (active, neutral, 120 g) activated at 240 °C for 4 h right before use. As for the fish oil sample, the activated aluminum oxide 60 (55 g) was the bottom layer, and the middle and top layers were silica gel (60-200 mesh, 80 g) activated at 160 °C for 3 h right before use and activated carbon (2 g), respectively. The chromatographic columns and collection vessels were wrapped in aluminum foil, and the oils were drawn through the column by suction without solvent 9 . To make sure complete purification of the oils, the contents of total phenolics, total tocopherols, and hydroperoxides (see below) were determined in the purified oils. The oils containing HT (0.6 and 1.2 mM) and TBHQ (0.6 mM) were prepared by adding aliquots of their solutions in acetone. The acetone was removed under a steam of nitrogen.
Total phenolics. The oil (2.5 g) was dissolved in 2.5 ml of n-hexane and extracted three times by 5-min centrifugations (2700×g) with CH 3 OH:H 2 O (80:20 v/v). After adding 2.5 ml of Folin-Ciocalteau reagent and 5 ml of 7.5% Na 2 CO 3 to the extract, the solution was made up to 50 ml with distilled water. The solutions were stored overnight and the spectrophotometric analysis was performed at 765 nm. To plot calibration curve, 1 ml of methanolic solutions of gallic acid (0.04-0.40 mg/ml), 6 ml of methanol, 2.5 ml of the Folin-Ciocalteau reagent, and 5 ml of 7.5% Na 2 CO 3 were made up to 50 ml with distilled water. TP was reported as milligrams of gallic acid per kilogram of oil 10 . Total tocopherols. A solution of 100 ± 10 mg of the oil in 5 ml of toluene was prepared. After successive addition of 3.5 of 2,2′-bipyridine (0.07% w/v in 95% aqueous ethanol) and 0.5 ml of FeCl 3 .6H 2 O (0.2% w/v in 95% aqueous ethanol), the solution was made up to 10 ml with 95% aqueous ethanol. After 1 min, the absorbance was read at 520 nm against a blank containing all the reagents except the oil. All the operations were carried   www.nature.com/scientificreports/ out under subdued light. A calibration curve of α-tocopherol in toluene (0-240 µg/ml) was prepared. TT was reported as milligrams of α-tocopherol per kilogram of oil 11 .
Partition coefficient (log P). Solutions  Fatty acid composition. Fatty acids were transesterified into methyl esters by vigorous shaking of a solution of oil in n-hexane (0.3 g in 7 ml) with 2 ml of 7 N methanolic potassium hydroxide at 50 °C for 10 min. The methyl esters were identified using an HP-5890 gas chromatograph (Hewlett-Packard, CA, USA) equipped with a CP-FIL 88 (Supel Co., Inc., Bellefonte, PA) capillary column of fused silica, 60 m in length × 0.22 mm I.D., 0.2 µm film thickness, and a flame ionization detector. Nitrogen was used as carrier gas with a flow rate of 0.75 ml min −1 .
The temperature of oven, injector and detector was maintained at 198, 250 and 250 °C, respectively. The fatty acid compositions were reported in relative area percentages with the average of duplicate samples 13 .
Peroxidation. The concentration (mM) of LOOH (known as peroxide value, PV; see below) was determined over time in a kinetic regime 14 in which the reaction medium is saturated with oxygen through performing the process in layers of a thickness less than 1 mm. In such a condition, more reproducible kinetic parameters are achieved and the rate of LOOH accumulation is independent of oxygen concentration 15 . The 1-mm layers of the oils (4 g) in Petri dishes of 9 cm in diameter were stored in a dry oven set at 60 °C. ) were added, respectively, and after adding each of them, the sample was mixed on a vortex mixer for 2-4 s. Then, the absorbance of the sample was read, after 5 min incubation at room temperature 16 . Results in milliequivalents of oxygen per kilogram of oil were reported as LOOH molarity (1 meq kg -1 = 0.504 mM) 15 .

PV measurement.
Kinetic parameters derived from the LOOH accumulation curves. The maximum rate of LOOH production (R max , mM h −1 ) and the propagation oxidizability parameter (R n , h −1 ) were calculated from Eqs. (5) and (6), respectively. Equations (7) and (8)   www.nature.com/scientificreports/ The initiation oxidizability parameter O i (mM −1 h 2 ), representing lipid oxidizability only with respect to the initiation phase, and t p (h) were calculated as follow: Initiation kinetics of inhibited peroxidation. According to the terminology introduced by Yanishlieva and Marinova 6 , the effectiveness of an antioxidant, which means its capability to scavenge peroxyl radicals (LOO • ), is calculated by the stabilization factor F: where IP AH and IP C are the IPs in the presence and absence (control) of the antioxidants (AH), respectively. Oxidation rate ratio (ORR) as an inverse measure of antioxidant strength was generated by Eq. (12): where k IP,AH and k IP,C are the values of k IP in the presence and absence (control) of AH, respectively. Antioxidant activity was calculated by the unifying parameter A: Statistical analysis. All determinations were carried out in triplicate and data were subjected to analysis of variance (ANOVA). ANOVA and regression analyses were carried out by the MStatC and Slide Write 7.0. Significant differences were determined by Duncan's multiple range tests. P values < 0.05 were considered statistically significant.

Results and discussion
Chemical composition of the oil samples. The chromatographic technique yielded the stripped oils containing no detectable LOOH, tocopherols, and phenolic compounds. The canola and Kilka fish oils possessed the same fatty acid compositions as the corresponding ones usually reported in literature (Table 1). Due to the content of mainly palmitic acid (C16:0), the fish oil had significantly higher saturation degree (SFA) than the canola oil. The canola oil was constituted of more remarkable content of monounsaturated fatty acids (MUFA, mainly oleic acid, C18:1) compared to the fish oil (mainly palmitoleic, C16:1, and oleic acids). While linoleic (C18:2) and linolenic (18:3) acids were the main types of polyunsaturated fatty acids (PUFA) measured in the canola oil, the majority of PUFA found in the fish oil were linoleic acid as well as the highly oxidizable eicosapentaenoic (EPA, C20:5) and docosahexaenoic (DHA, C22:6) acids. Naturally, the degree of unsaturation plays an important role in susceptibility of lipid matrices to oxidation. The relative rate of oxidation for stearic (C18:0), oleic, linoleic, and linolenic acids has been reported to be 1:100:1200:2500 17 . Arachidonic acid (C20:4) has been shown to be oxidized 2.9 times faster than linoleic acid 18 . Oxygen uptake of EPA and DHA esters after the IPs has been reported to be 5.2 and 8.5 times, respectively, faster than that of ethyl linolenate 19 . Figure 3 illustrates the LOOH kinetic curves over the whole range of the lipid peroxidation of the stripped canola and fish oils at 60 °C. Table 2 provides a wide range of kinetic data calculated from Eqs. Initiation phase of lipid peroxidation. As expected, the canola oil was of significantly better initiation oxidizability than the fish oil (O i = 2.23 vs. 0.11 mM −1 h 2 , Table 2). This was obviously due to its less highly polyunsaturated fatty acid composition (Table 1; C20:4, C20:5, and C22:6) which naturally prolongs IP and decreases k IP . Lipid systems of higher capability to generate less reactive LOO • show greater IPs 6 . Also, those less prone to produce free radicals of more reactivity and diversity, including LOO • (E 0 = 1000 mV), alkoxyl (LO • , E 0 = 1600 mV), and/or hydroxyl ( • OH, E 0 = 2320 mV) 20,21 , being able to initiate and propagate peroxidation 22 , provide smaller k IP values.

Kinetic data analysis.
The parameter O i was extraordinarily improved in the presence of the antioxidants added (Table 2), indicating their high potency to scavenge the reactive free radicals LOO • , LO • , and/or • OH. Considering the unifying parameter A given in Table 3, the antioxidants acted more dramatically against the initiation peroxidation in the canola oil which possessed a reaction environment of relatively lower oxidative instability than that of the fish oil. This clearly implies the more intensive consumption of the antioxidant molecules to reduce the radicals of (8) [LOOH] IP = k IP (IP) + [LOOH] 0 . www.nature.com/scientificreports/ higher reactivity and diversity in the fish oil. Similar results have been reported to show the superior antioxidant activity of HT 9 and TBHQ 3,8,23 in the linoleic/linolenic acid group of edible oils. In general, TBHQ with two hydroxyl (-OH) groups in para position and a tertiary butyl group [-C(CH 3 )] around the phenolic -OH group (Fig. 2) exhibited quite higher antioxidant activities than HT with two -OH groups in ortho position and a hydroxyethyl group (-CH 2 CH 2 OH) far from the phenolic -OH group (Fig. 2) in preventing the initiation peroxidation (Table 3). It has been postulated that the alcoholic -OH group in HT is able to orient towards the aromatic ring and to establish an intramolecular hydrogen bond with the catecholic (1,2-dihydroxybenzene) hydroxyl groups. This accordingly leads to form minimum-energy conformers which have lower tendency to reduce the oxidizing radicals 24 .
As shown in Table 2, the parameter [LOOH] IP of the control samples were significantly different. According to Ghnimi et al. 25 , [LOOH] IP actually represents the critical concentration of reverse micelles (CMC L ) composed basically of the LOOH accumulated during IP. It may indicate the level of LOOH amphiphilicity and their spatially alignment in water-oil interfaces 24 . The much bigger [LOOH] IP in the fish oil control can be attributed to its fatty acid composition of higher diversity and the lower amphiphilic character resulting from the lesser extent of unsaturation degree (Table 1). Diversity in fatty acid compositions arises from the length of acyl chains as well as the number of double bonds. More unsaturated fatty acids generate LOOH molecules of higher polarity and, therefore, of higher levels of surface activity, facilitating their incorporation into the water-oil interfaces 25 . However, more unsaturated LOOH would be bulkier and occupy more space in the interface, decreasing their compact aggregations 26 . The value of [LOOH] IP decreased considerably when adding the antioxidants ( Table 2). Phenolic antioxidants are considered as amphiphilic molecules likely to have surfactant or co-surfactant properties 25 , enabling them to decrease interfacial tension and establish more stable and organized reverse micelles. TBHQ with a hydrophobic tertiary butyl substituent showed a partition coefficient of three times higher than that of HT with a hydrophilic ethyl alcohol substituent (log P = 1.38 vs. 0.46), denoting its lesser sharing in the water-oil interfaces. Similar results were observed when evaluating the antioxidant potency of gallic acid and methyl gallate compared to TBHQ in sunflower oil triacyl glycerols 3 .
Propagation phase of lipid peroxidation. The duration of the propagation phase (t p ) is normally quite smaller than IP under mild oxidative conditions. However, they might approach together as the oxidative conditions become harsher. As can be seen in Table 2, the antioxidants could significantly change the values of t p in both the oil samples. By analogy with the initiation kinetics of the inhibited peroxidation (Table 3), the relative quantities of t p (t p ′, Table 4) indicated more remarkable effect of HT and TBHQ on the propagation time of the fish oil possessing the fatty acid composition of naturally higher susceptibility to peroxidation (Table 1). This was in contrast to the antioxidants performance in terms of the stabilization factor F in the two oils (Table 3). TBHQ, www.nature.com/scientificreports/ in general, exerted significantly better inhibitory effects than HT on the oxidizing radicals in the propagation phase of lipid peroxidation. The maximum rate of LOOH formation during t p (R max ) was significantly affected by the antioxidants added ( Table 2). This demonstrates that the antioxidants were still able to scavenge the reactive free radicals that propagate the oxidation reaction chains. Besides, the value of [LOOH] max as a measure of the potency of lipid systems to create LOOH compositions of different stability 4 significantly improved under the inhibited peroxidations. As given by Eq. (4), [LOOH] max is affected by the balance between the overall formation rate of LOOH molecules (k c ) and the rate of LOOH decomposition (k d ) 5 , which both in turn could show well the better antioxidant performance of TBHQ than HT in the two oils ( Table 2). The composite rate constant k c has been shown to be fully correlated with the parameter R n , unifying the values of R max and [LOOH] max 5 . R n can be taken into account as a comprehensive kinetic parameter encompassing the values of every single kinetic parameter and rate constant noted above. It could significantly differentiate the antioxidant potencies in inhibiting the propagation peroxidations as affected by the type of antioxidant as well as oxidative system. With respect to the relative quantities of R n (R n ′, Table 4), HT and especially TBHQ were able to protect better the fish oil, which was more oxidizable than the canola oil, from propagation peroxidation. www.nature.com/scientificreports/   www.nature.com/scientificreports/

Conclusions
The present study indicated how a wide range of kinetic parameters and rate constants characterizing the initiation and propagation phases of lipid peroxidation may be changed by adding an antioxidant to the lipid systems of different degrees of unsaturation. Unlike the conventional methodology focusing on the antioxidants performance exclusively during the initiation phase of lipid peroxidation, this study demonstrated that the activity of antioxidants in the propagation phase must be taken into account as well. Interestingly, the performance of an antioxidant in the initiation and propagation phases might be quite different from each other as a function of the degree of unsaturation and the diversity in the fatty acid composition. This is of extremely high importance because the secondary oxidation products, which lead to many dramatic negative effects on sensory attributes and toxicity of lipid matrices, are likely to significantly be produced in the propagation phase of lipid peroxidation. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.