Oxidation kinetics of sardine oil in the presence of commercial immobilized lipases commonly used as biocatalyst

Accepted Manuscript Oxidation kinetics of sardine oil in the presence of commercial immobilized lipases commonly used as biocatalyst Ángela García Solaesa, María Teresa Sanz, Rodrigo Melgosa, Sagrario Beltrán PII: S0023-6438(18)30446-8 DOI: 10.1016/j.lwt.2018.05.032 Reference: YFSTL 7142 To appear in: LWT - Food Science and Technology Received Date: 22 January 2018 Revised Date: 10 May 2018 Accepted Date: 11 May 2018 Please cite this article as: Solaesa, Á.Garcí., Sanz, Marí.Teresa., Melgosa, R., Beltrán, S., Oxidation kinetics of sardine oil in the presence of commercial immobilized lipases commonly used as biocatalyst, LWT - Food Science and Technology (2018), doi: 10.1016/j.lwt.2018.05.032. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 1 Oxidation kinetics of sardine oil in the presence of 1 commercial immobilized lipases commonly used as biocatalyst 2 Ángela García Solaesa, María Teresa Sanz ∗ , Rodrigo Melgosa, Sagrario Beltrán 3 Department of Biotechnology and Food Science (Chemical Engineering Section), 4 University of Burgos, 09001 Burgos. Spain 5 Abstract 6 Oxidation kinetics of sardine oil have been determined at 40, 65 and 90ºC by measuring 7 concentration of primary and secondary oxidation products in the presence of 8 commercial immobilized lipases (Lipozyme 435, Lipozyme RM and Lipozyme TL) 9 commonly used as biocatalyst in lipid modification reactions. Oxidation products 10 concentration was found to be lower when the immobilized lipases were added at the 11 highest temperatures studied. The lowest oxidation indices were observed in the 12 presence of Lipozyme RM. 13 Although the mechanism to explain this decrease in the oxidation products is not still 14 clear, these results might indicate that the use of these immobilized lipases in lipase15 catalyzed reactions of fish oils at high temperature (90ºC) will yield higher reaction 16 rates but also a reduction of the oxidation products formed due to oxidation of 17 polyunsatured fatty acids. 18 Keywords: fish oil, omega-3, oxidation products, commercial immobilized lipases. 19 ∗Corresponding author. Tel.: +34 947 258810. Fax: ++34947258831. E-mail address [email protected] MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 2 1. Introduction 20 Fish oil has high amounts of omega-3 long-chain polyunsaturated fatty acids (n-3 21 PUFA), mainly eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22 22:6n-3) which have been reported to protect against the development of many diseases 23 (Solaesa, Sanz, Falkeborg, Beltrán, & Guo, 2016). However, despite the high nutritional 24 value of these products, the high degree of unsaturation makes fish oils very prone to 25 autoxidation. The resulting breakdown products cause off-flavors and rancidity, loss of 26 nutritional value and finally consumer rejection (Gómez-Alonso, Mancebo-Campos, & 27 Salvador, 2004). The degree and rate of lipid oxidation is influenced by the unsaturation 28 of fatty acids, oxygen concentration, temperature, surface area, water activity and 29 presence of anti– and prooxidants. Temperature has also an important impact on lipid 30 oxidation since an increase in temperature accelerates oxidation rates. 31 The process of lipid oxidation can be described in three general steps: initiation, 32 propagation and termination. The concentrations of primary and secondary oxidation 33 products can be measured quantitatively and therefore give an indication of the 34 oxidative status of the oil. The hydroperoxide content, as primary oxidation products, is 35 usually determined by the peroxide value (PV) assay. Secondary oxidation products can 36 be determined by anisidine value (AV) and thiobarbituric acid reactive substances 37 (TBARS) assays. AV measures mainly 2-alkenals and 2,4-dienals. On the other hand, 38 the cyclic peroxides formed by autoxidation from polyunsaturated fatty acids with three 39 or more double bonds are the most important precursors of malonaldehyde and hence 40 source of TBARS (Hoyland & Taylor, 1991). Polyene index (PI) is also used as a good 41 indicator of PUFA deterioration in fish oils, defined as (EPA + DHA): palmitic acid 42 ratio (Pazhouhanmehr, Farhoosh, Sharif, & Esmaeilzadeh, 2016). 43 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 3 Omega-3 concentrates can be produced by lipase-catalyzed reactions since they can be 44 carried out under mild conditions in short reaction times compared with chemical 45 processes. Some widely used commercial immobilized lipases used in enzymatic 46 reactions are from Candida antarctica (Novozym 435 and Lipozyme 435), Rhizomucor 47 miehei (Lipozyme RM IM) and Thermomyces lanuginose (Lipozyme TL IM). In a 48 previous work (Solaesa, Sanz, Melgosa, & Beltrán, 2017), it was found that the 49 oxidation status of the final reaction products of glycerolysis of sardine oil at 60 and 50 90ºC catalyzed by the commercial immobilized lipase, Lipozyme 435, was even lower 51 than the oxidation status of the initial sardine oil, determined as PV and AV. Especially, 52 hydroperoxide content decreased in a greater extend at the highest reaction temperature 53 assayed. Therefore, an increase in reaction temperature, not higher than the maximum 54 denaturation temperature for the lipase, was positive since reaction rate increased and 55 lower oxidation status of the final products was determined. 56 Based on those previous results, in this work, a systematic study of the oxidation 57 products formation of sardine oil in the presence of three commercial immobilized 58 lipases commonly used in lipid modification reactions has been carried out in the 59 temperature range where lipase catalyzed reactions usually take place. The commercial 60 lipases used, immobilized onto different supports, were Lipozyme 435, Lipozyme RM 61 and Lipozyme TL. Autoxidation kinetics of sardine oil, as a control, were first 62 determined at the temperatures selected in this work (40, 65 and 90ºC) to determine the 63 net formation rate of the primary and secondary oxidation products. Afterwards, sardine 64 oil was incubated at the same temperatures in contact with the different immobilized 65 lipases to evaluate the reduction of the oxidation products concentration in the presence 66 of these immobilized lipases. 67 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 4 2. Materials and methods 68 2.1 Materials 69 Refined sardine oil was provided by Industrias Afines S.L. (Spain). Three commercial 70 immobilized lipases were used in this work. Lipozyme 435 from Candida antarctica B 71 and Lipozyme TL IM from Thermomyces lanuginosus were kindly donated by 72 Novozymes A/S (Bagsvaerd, Denmark). Lipozyme RM IM, from Rhizomucor miehei, 73 was purchased from Sigma Aldrich. The support characteristics of these immobilized 74 lipases are summarized in Table 1. All other chemicals used in the different analyses 75 performed in this work, were of analytical or HPLC grade. 76 2.2 Oxidation kinetics of sardine oil 77 First, autooxidation kinetics of sardine oil were evaluated at three different temperatures 78 40, 65 and 90 ºC as control samples. 79 For each experiment, stoppered erlenmeyer glass flasks were filled with 70 g of refined 80 sardine oil and immediately applied a nitrogen stream in the flask. Afterwards they were 81 perfectly closed and covered with foil paper to avoid the light exposure. Sardine oil 82 samples were incubated in a water bath with orbital agitation at the corresponding 83 temperature. At different time intervals, during10 h, aliquots were withdrawn to follow 84 the oxidation kinetics. 85 When the oxidation kinetics were determined in the presence of the immobilized 86 lipases, the sardine oil was put in contact with the different immobilized commercial 87 lipases, Lipozyme 435, Lipozyme RM and Lipozyme TL, at 3.5% w/w concentration. 88 This concentration was found to be suitable for lipid modification reactions (Solaesa et 89 al., 2016). As for the autooxidation kinetics, samples were withdrawal at different time 90 intervals during 10 h. 91 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 5 All samples were stored at -18ºC up to oxidation products analysis. Duplicate 92 experiments were carried out at each temperature. 93 2.3 Determination of primary oxidation products: peroxide value (PV) 94 PV was determined by iodometric titration following the AOAC Official Method 95 (AOAC Official Method 965.33, 2000) by an automatic titrator Methrom 905 Titrando. 96 PV was expressed as milliequivalents of O 2 per kilogram of sample. All samples were 97 analyzed in triplicate. 98 2.4 Determination of secondary oxidation products 99 2.4.1 Anisidine value (AV) 100 The AV was measured according to AOCS Official Method (AOCS Official Method Cd 101 18-90, 2017), using a UV-Visible spectrophotometer at 350nm. All samples were 102 analyzed in triplicate. 103 2.4.2 Thiobarbituric acid reactive substances (TBARS) assay 104 TBARS were determined according to the spectrophotometric method described by 105 Norveel Semb in her Master's Thesis (Norveel Semb, 2012). The method is based on the 106 formation of a pink complex with strong absorbance at 532-535nm due to the presence 107 of thiobarbituric acid reactive substances (TBARS). TBARS is expressed as mg of 108 malondialdehyde (MDA) per kilogram of oil. All samples were analyzed in triplicate. 109 2.5 Fatty acid analysis 110 The initial sardine oil and the oil samples after 10 h of incubation at 40, 65 and 90ºC 111 were analyzed by the AOAC method (AOAC Official Method 991.39, 2000) to evaluate 112 the fatty acid profile and the polyene index (EPA+DHA/16:0). An Agilent gas 113 chromatograph (6890N Network GC System) equipped with a flame ionization detector 114 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 6 (FID) and a fused silica capillary column (OmegawaxTM-320, 30 m x 0.32 mm i.d.) 115 was used. The method and the calibration procedure was developed previously (Solaesa 116 et al., 2016). The samples were analyzed in triplicate. 117 2.6 Statistical analysis 118 Statistical analyses were conducted using software Statgraphics X64. Control 119 experiments and oxidation status of sardine oil in contact with the immobilized lipases 120 were carried out in duplicate. For each oxidation product, analysis was performed in 121 triplicate. The results are presented as a mean ± standard deviation. The significance of 122 the differences was determined based on an analysis of the variance with the Tukey's 123 honestly significant difference (HSD) method at p-value ≤ 0.05. For each oxidation 124 product, significance difference has been determined among the three lipases at the 125 three operating temperatures at a specific contact time. ANOVA was also performed to 126 analyze the statistical significance of time by comparing oxidation product 127 concentration for each enzyme and temperature along time. 128 The estimation of the parameters for the models tested in this work was performed by 129 using the Marquardt algorithm (Statgraphics X64). 130 3. Results and discussion 131 3.1 Chemical quality of initial sardine oil 132 The fatty acid profile and some quality parameters of the supplied refined sardine oil 133 used in this work are presented in Table 2. The fatty acid profile was similar to those 134 previously reported for sardine oil (Homayooni, Sahari, & Barzegar, 2014; Noriega135 Rodríguez et al., 2009; Okada & Morrisset, 2007; Solaesa, Bucio, Sanz, Beltrán, & 136 Rebolleda, 2014). The polyene index (PI) of the supplied refined sardine oil used in this 137 work was 1.39 (Table 2) being similar to the PI value reported for sardine oil 138 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 7 (Homayooni et al., 2014). Other important quality parameters such as acid value, PV 139 and AV for the supplied refined sardine oil were 0.45 ± 0.04 mg KOH/g, 4.8 ± 0.6 meq 140 O 2 /kg and 19.7 ± 0.4, respectively. These values were very close to the limits allowed 141 by GOED Voluntary Monograph (GOED, 2015), which are 0.5 mg KOH/g for acid 142 value, 5 meq O 2 /kg oil for PV and 20 for AV. Therefore, the supplied refined sardine oil 143 used in this work was partially oxidized and oxidation reaction rates could be faster due 144 to the presence in the medium of oxidation products. According to Toro-Vázquez et al. 145 (1993) the induction period was reduced as initial peroxide value increased, based on 146 their study of corn oil oxidation. The refined sardine oil used in this work presented a 147 value of 46 mg of MDA/kg. Regarding TBARS content, the maximum is not clearly 148 specified. A document by FAO establishes for fresh fish oil a TBARS value of 50 mg of 149 MDA/kg fish oil, although it is referred to fishes feed (Masson S, 1994). The induction 150 period (h) by Rancimat, determined at 70ºC and 20 L/h of air flow rate was 7.2 hours, 151 similar to that obtained by Noriega et al. (Noriega-Rodríguez et al., 2009) for crude and 152 deodorized sardine oil (10.4 and 17.7 h, respectively) using Rancimat at 60ºC and 7 L/h 153 of air flow rate. 154 3.2 Oxidation kinetics of sardine oil at different temperatures 155 3.2.1 Kinetic aspects of oxidation products 156 Peroxide values were determined over time during 10 h at 40, 65 and 90ºC (Fig. 1a). At 157 any incubation time, hydroperoxide content increased with incubation temperature. Fig. 158 1a also shows that peroxide value continuously increased with time during 10 h. This 159 means that hydroperoxide rate formation was still higher than hydroperoxide rate 160 decomposition in the period of time covered in this work for all the temperatures. 161 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 8 The kinetics of lipid oxidation is not an easy task. Reactions rates are usually described 162 by a pseudo-zero, pseudo-first and pseudo second order reactions. In any case, the order 163 or the reaction rate does not comply with the stoichiometry of the reaction (Kamal-Eldin 164 & Yanishlieva, 2005). Labuza and Bergquist (Labuza & Bergquist, 1983) found that 165 lipid oxidation were half-order with respect to pure lipids in model systems. However, 166 in complex food systems the data sometimes fit zero order as well (Labuza & Bergquist, 167 1983). In this work, formation of primary oxidation products followed a pseudo-zero 168 order kinetic model: 169 PV =PV o + k·t [1] 170 where PV o denotes the initial peroxide value and k is the reaction rate constant for a 171 zero order reaction (meq O 2 ·kg oil -1 ·h -1 ). Reaction rate constants, k, were determined 172 from the slope of plotting the PV as a function of incubation time. Table 3 lists the 173 reaction rate constant for the three temperatures assayed in this work, together with the 174 quality of the fitting. Pseudo-first and pseudo-second order kinetics were also tried, but 175 the fitting was worse than for pseudo-zero order reaction. Gomez-Alonso et al. (Gómez176 Alonso et al., 2004) also found that formation of primary oxidation products of olive oil 177 in the temperature range from 25 to 75ºC followed a pseudo-zero order kinetics. 178 Fig. 1b and 1c show the variation of the AV and the TBARS content, respectively, over 179 time at the three temperatures studied in this work. It can be observed that, both AV and 180 TBARS content, continuously increased with time and temperature in the first 10 h of 181 incubation time, mainly at the highest temperatures, 90ºC and 65ºC. At 40ºC the 182 oxidation rates were slower and the secondary oxidation products concentration did not 183 increase in a great extent. Comparing Fig. 1a, 1b and 1c it can be observed that 184 generation of secondary oxidation products was taking place simultaneously with 185 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 15 regard, Wang et al. (1991) in the study of transesterification reaction of triolein with 328 lauric acid catalyzed by Lipozyme IM 20 observed that, when testing the operation 329 stability of the immobilized lipase at 65 ºC over 10 number of reaction batches, 10 h 330 each batch, hydroperoxide content decreased through the repeated batches, but so did 331 the lipase activity. These authors proposed that hydroperoxide decomposition products 332 could cause inactivation of the enzyme. 333 Although, mechanism is not still clear, among the different phenomenon that could be 334 taking place, adsorption of the oxidation products on the lipase support could play an 335 important role. Lipozyme RM is immobilized onto a weak-base anion exchange resin, 336 having good capability for adsorption process. Its hydrophilic nature could result in a 337 higher affinity for hydrophilic oxidation products. On the contrary, Lipozyme 435 was 338 immobilized onto a hydrophobic support (Chen et al., 2008). The secondary oxidation 339 products, as well as hydroperoxides, are considered polar molecules and, it is well 340 known that the higher the degree of oxidation of an oil, the more polar the oxidation 341 products usually are (Kamal-Eldin & Yanishlieva, 2005). This could partially explain 342 the difference observed among both immobilized lipases. In case of Lipozyme TL, 343 although it was immobilized on silica gel, a hydrophilic material, its particle size range 344 was bigger (250-1000 µm) as well as the true density of the support (1830 kg/m 3 , 345 Zhang, 2007). Therefore, its surface area was considerably lower. In Table 1 it can be 346 observed that the average value of Lipozyme RM particle size was smaller (200~600 347 µm) than the particle size of the others two immobilized lipases (300-1000 µm). 348 Therefore, the adsorption could be increase due to a higher frequency of collisions 349 between adsorbate and adsorbent. In any case, it must be highlighted that the average 350 particle of these immobilized lipases was much higher than the particle size for 351 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 16 commercial absorbents such as Tonsil ® bleaching earths which range from 25 to 150 352 µm (Strieder et al., 2017). 353 The oxidation status for sardine oil in contact with the immobilized lipases (C immob-lipase ) 354 has been related to the corresponding data for the control sardine oil previously 355 determined in section 3.2 at the same temperature and incubation times (C control ), C immob356 lipase /C control , for each type of oxidation product. Figures 3a, 3b, 4a, 4b, 5a and 5b show 357 the ratio C immob-lipase /C control for the PV, AV and TBARS level change at 90 and 65ºC, 358 respectively for the three immobilized lipases. The oxidation status of sardine oil in 359 contact with the immobilized lipases at 40ºC showed no difference with that of the 360 sardine oil in the absence of immobilized lipases, or even slightly higher oxidation 361 status. Therefore, the results have not been plotted. For the three immobilized lipases 362 tested in this work at 65 and 90ºC, the ratio C immob-lipase /C control decreased over time. 363 Based on the shape of the Fig. 3-5 the following equation has been used to correlate the 364 data over time: 365 C  !"#$%&',) C *+),#,) -=exp.−kt + 0 [5] 366 where C immob-lipase,t is the concentration of oxidation compounds in sardine oil in contact 367 with immobilized lipases at a certain contact time, t, C control,t is the concentration of 368 oxidation compounds in control sardine oil (in the absence of immobilized lipases) at 369 the same incubation time, t, n is an adjustment parameter and k is an empirical constant. 370 This equation was based on the modified model of Brimberg proposed by Monte et al. 371 (Monte et al., 2015) in the study of reduction of color and oxidation products of carp oil 372 with blends of bleaching earth and activated carbon. The dependence of k parameter on 373 temperature was assumed to follow an Arrhenius type relationship (Eq. 2), although 374 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 17 only the highest temperatures, 90 and 65º C were considered in the fitting procedure. 375 Table 6 lists the parameters obtained for equation 5 to relate the oxidation products 376 concentration in contact with the three immobilized lipases and the control oil for the 377 three measured parameters, PV, AV and TBARS content. For all the oxidation products, 378 the highest E a was found for Lipozyme TL proving that this immobilized lipase is more 379 temperature-dependent on the reduction of oxidation products in the bulk oil. The 380 continuous lines in Fig. 3-5 represent the equation 5, showing good fitting. 381 4. Conclusions 382 The oxidation kinetics of sardine oil revealed that hydroperoxide and TBARS formation 383 was higher than that of secondary oxidation products determined as AV, due to the high 384 PUFA content. Oxidation status was lower in the presence of three commercial 385 immobilized lipases, in the same temperature range, especially at high temperatures, 65 386 and 90ºC. However, at 40ºC, oxidation products formation seemed to be still promoted. 387 Lipozyme RM yielded the lower oxidation indices, quite below to those of the initial 388 fish oil. 389 Although exact mechanism of reduction of oxidation products in the presence of 390 commercial immobilized lipases is not yet clear. It is an important finding since high 391 temperatures are usually avoided when dealing with lipid modification of fish oil due to 392 its high content of PUFA, very prone to oxidation. By using these type of lipases, 393 temperatures, at least, up to 90ºC can be used having a double benefit of temperature; 394 on one hand higher reaction rates and, on the other hand lower oxidation status. 395 However, reusability of the enzyme should be further considered. 396 Acknowledgements 397 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 18 The authors thank the European Regional Development Fund (ERDF) and Junta de 398 Castilla y León [grant number BU055U16] for financial support. AGS acknowledges 399 University of Burgos and RM MINECO [grant number BES-2013-063937] for their 400 pre-doctoral contracts. 401 402 403 404 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 19 References 405 AOAC Official Method 965.33. (2000). Peroxide Value of Oils and Fats. 406 AOAC Official Method 991.39. (2000). Fatty Acids in Encapsulated Fish Oils and Fish 407 Oil Methyl and Ethyl Esters. 408 AOCS Official Method Cd 18-90. (2017). p-Anisidine Value. 409 Chen, B., Hu, J., Miller, E. M., Xie, W., Cai, M., & Gross, R. A. (2008). Candida 410 antarctica Lipase B Chemically Immobilized on Epoxy-Activated Microand 411 Nanobeads : Catalysts for Polyester Synthesis. Biomacromolecules, 9, 463–471. 412 GOED. (2015). Voluntary Monograph Version 5 (Issue Date November 19). 413 Gómez-Alonso, S., Mancebo-Campos, V., & Salvador, M. D. (2004). Oxidation 414 kinetics in olive oil triacylglycerols under accelerated shelf-life testing (25 – 75 o C). 415 European Journal of Lipid Science and Technology, 106, 369–375. 416 Guillen, M.D., Cabo N. (2002). Fourier transform infrared spectra data versus peroxide 417 and anisidine values to determined oxidative stability of edible oils. Food 418 Chemistry, 77, 503-510. 419 Homayooni, B., Sahari, M. A., & Barzegar, M. (2014). Concentrations of omega-3 fatty 420 acids from rainbow sardine fish oil by various methods. International Food 421 Research Journal, 21(2), 743–748. 422 Hoyland, D., & Taylor, A. (1991). A Review of the Methodology of the 2423 Thiobarbituric Acid Test. Food Chemistry, 40, 271–291. 424 Kamal-Eldin, A., & Yanishlieva, N. (2005). Chapter 10. Kinetic Analysis of Lipid 425 Oxidation Data. In AOCS Lipid Library. 426 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 20 Labuza, T., & Bergquist, S. (1983). Kinetics of Oxidation of Potato Chips under 427 Constant Temperature and Sine Wave Temperature Conditions. Journal of Food 428 Processing and Preservation, 48, 712–715. 429 Lee, J., Lee, Y., & Choe, E. (2007). Temperature dependence of the autoxidation and 430 antioxidants of soybean, sunflower, and olive oil. European Food Research and 431 Technology, 226, 239-246. 432 Masson S, L. (1994). Control de Calidad de Insumos y Dietas Acuicolas 433 http://www.fao.org/docrep/field/003/ab482s/ab482s10.html. 434 Monte, M. L., Monte, M. L., Pohndorf, R. S., Crexi, V. T., & Pinto, L. A. A. (2015). 435 Bleaching with blends of bleaching earth and activated carbon reduces color and 436 oxidation products of carp oil. European Journal of Lipid Science and Technology, 437 117, 829–836. 438 Noriega-Rodríguez, J. A., Ortega-García, J., Angulo-Guerrero, O., García, H. S., 439 Medina-Juárez, L. A., & Gámez-Meza, N. (2009). Oil production from sardine ( 440 Sardinops sagax caerulea ) Producción de aceite a partir de sardina ( Sardinops 441 sagax caerulea). CyTA -Journal of Food, 7(3), 173–179. 442 Norveel Semb, T. (2012). Analytical Methods for Determination of the Oxidative Status 443 in Oils (Master’s Thesis; NTNU). 444 Okada, T., & Morrisset, M. (2007). Recovery and Characterization of Sardine Oil 445 Extracted by pH Adjustment. Journal of Agricultural and Food Chemistry, 55, 446 1808–1813. 447 Pazhouhanmehr, S., Farhoosh, R., Sharif, A., & Esmaeilzadeh, R. (2016). Oxidation 448 kinetics of common Kilka ( Clupeonella cultiventris caspia ) oil in presence of 449 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 21 bene oils ’ unsaponifiable matter. Food Chemistry, 190, 748–754. 450 Rodrigues, R. C., & Fernandez-Lafuente, R. (2010). Lipase from Rhizomucor miehei as 451 a biocatalyst in fats and oils modification. Journal of Molecular Catalysis B: 452 Enzymatic, 66, 15–32. 453 Sathivel, S., Huang, J., & Prinyawiwatkul, W. (2008). Thermal properties and 454 applications of the Arrhenius equation for evaluating viscosity and oxidation rates 455 of unrefined pollock oil. Journal of Food Engineering, 84, 187–193. 456 Silva Rodrigues, J., Peixoto do Valle, C., Pinheiro Guerra, P. D. A. G., De Sousa Rios, 457 M. A., Queiroz Malveira, J. De, & Ricardo, N. M. P. S. (2017). Study of kinetics 458 and thermodynamic parameters of the degradation process of biodiesel produced 459 from fi sh viscera oil. Fuel Processing Technology, 161, 95–100. 460 Solaesa, Á. G., Bucio, S. L., Sanz, M. T., Beltrán, S., & Rebolleda, S. (2014). 461 Characterization of Triacylglycerol Composition of Fish Oils by Using 462 Chromatographic Techniques. Journal of Oleo Science, 63(5), 449–460. 463 Solaesa, Á. G., Sanz, M. T., Falkeborg, M., Beltrán, S., & Guo, Z. (2016). Production 464 and concentration of monoacylglycerols rich in omega-3 polyunsaturated fatty 465 acids by enzymatic glycerolysis and molecular distillation. Food Chemistry, 190, 466 960–967. 467 Solaesa, Á. G., Sanz, M. T., Melgosa, R., & Beltrán, S. (2017). Substrates 468 emulsification process to improve lipase-catalyzed sardine oil glycerolysis in 469 different systems. Evaluation of lipid oxidation of the reaction products. Food 470 Research International, 100(May), 572–578. 471 Strieder, M. M., Pinheiro, C. P., Borba, V. S., Pohndorf, R. S., Cadaval, T. R. S., & 472 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 22 Pinto, L. A. A. (2017). Bleaching optimization and winterization step evaluation in 473 the refinement of rice bran oil. Separation and Purification Technology, 175, 72– 474 78. 475 Sullivan, J. C., Suzanne, R., Jovica, F., & Jean, A. (2015). Oxidation Rates of 476 Triacylglycerol and Ethyl Ester Fish Oils. Journal of the American Oil Chemists’ 477 Society, 92, 561–569. 478 Toro-Vazquez, J.F., Castillo-M., A.A., & Hernández-C., R. (1993). A multiple-variable 479 approach to study corn oil oxidation. Journal of the American Oil Chemist's 480 Society, 70(3), 261-267. 481 Wang, Y., & Gordon, M.H. (1991). Effect of lipid oxidation products on the 482 transesterification activity of an immobilized lipase. Journal of Agriculturad and 483 Food Chemistry, 39, 1693-1695. 484 Xu, X., Timm-H., M., Skall Nielsen, N., Porsgaard, T., & Jacobsen, Ch. (2005) Effects 485 of antioxidants on the lipase-catalyzed acidolysis during production of structured 486 lipids. European Journal of Lipid Science and Technology, 107, 464-468. 487 Yin, H., & Sathivel, S. (2010). Physical Properties and Oxidation Rates of Unrefined 488 Menhaden Oil ( Brevoortia patronus ). Journal of Food Science, 75(3), 163–168. 489 Zhang, H. (2007). Evaluation of Practical Process Aspects for Lipozyme TL IM 490 Catalyzed Bulk Fat Modification in a Batch Reactor. The Open Biotechnology 491 Journal, 1, 72–80. 492 493 494 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 23 495 496 497 Fig. 1. (a) PV, (b) AV and (c) TBARS of sardine oil during incubation time at 40ºC (∆), 498 65ºC (□) and 90ºC (◊). Continuous lines represent the pseudo-zero order kinetic model (Table 499 3). 500 0 10 20 30 40 50 60 0 2 4 6 8 10 12 PV, meqO2/kg oil time, h 0 10 20 30 40 50 60 70 0 2 4 6 8 10 12 AV time, h 0 25 50 75 100 125 150 175 200 0 2 4 6 8 10 12 TBARS (mg MDA/kg oil) time, h (a) (b) (c) MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 24 501 502 503 504 Fig. 2. (a) Arrhenius plot and (b) Eyring plot of the oxidation of sardine oil, PV (○), AV (□), 505 TBARS (∆). Units of k are listed in Table 3. Time units in Eyring plot for k values have been 506 expressed as s. 507 508 509 -2 -1 0 1 2 3 0.0027 0.0028 0.0029 0.0030 0.0031 0.0032 0.0033 ln k 1/T (K-1) (a) -16.0 -15.2 -14.4 -13.6 -12.8 -12.0 -11.2 0.0027 0.0028 0.0029 0.0030 0.0031 0.0032 0.0033 ln (k/T) 1/T (K-1) (b) MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 31 Table 4. Activation energy from the Arrhenius equation for the formation of hydroperoxides in different kind of oils found in literature. Type of oil T range (ºC) E a , kJ/mol Reference Refined sardine oil 40 - 90 37.1 This work Unrefined menhaden oil 45 - 85 30.9 (Yin & Sathivel, 2010) Unrefined pollock oil 24 - 90 33.2 (Sathivel et al., 2008) Refined soybean oil 25 - 80 73.6 (Lee et al., 2007) Refined sunflower oil 79.5 Virgin olive oil 52.3 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 32 Table 5. Oxidation products concentration for sardine oil samples in contact with Lipozyme 435, Lipozyme RM and Lipozyme TL at 40, 65 and 90ºC at different incubation times. Oxidation Index Immobilized lipase T (ºC) Contact time (h) 1 2.5 5 7.5 10 PV Lipozyme 435 90 B 4.3 ± 0.4 a AB 3.8 ± 0.1 ab A 3.1 ± 0.1 ab A 3.1 ± 0.1 ab A 3.3 ± 0.3 b 65 A 4.6 ± 0.1 a A 4.6 ± 0.1 bc A 4.3 ± 0.5 cd A 3.9 ± 0.1 b A 3.9 ± 0.1 c 40 A 4.6 ± 0.2 a B 5.2 ± 0.0 cd C 6.5 ± 0.0 e B 5.3 ± 0.5 c AB 5.0 ± 0.1 d Lipozyme RM 90 B 4.8 ± 0.3 a AB 3.6 ± 0.3 a A 2.7 ± 0.2 a A 2.2 ± 0.1 a A 1.9 ± 0.1 a 65 C 4.4 ± 0.3 a B 3.6 ±0.2 a A 2.9 ± 0.1 ab A 2.5 ± 0.2 a A 2.4 ± 0.1 a 40 A 4.6 ± 0.1 a A 5.4 ± 0.2 cd B 8.3 ± 0.3 f B 7.4 ± 1.0 d A 5.1 ± 0.4 d Lipozyme TL 90 C 5.0 ± 0.0 a C 5.3 ± 0.2 cd B 3.6 ± 0.0 bc A 2.9 ± 0.1 ab B 4.0 ± 0.3 c 65 C 5.4 ± 0.2 a C 5.6 ± 0.2 d B 4.6 ± 0.2 d A 3.9 ± 0.1 bc A 3.5 ± 0.2 bc 40 A 5.2 ± 0.0 a C 11.9 ± 0.7 e D 15.0 ± 0.1 g C 11.9 ± 0.5 e B 9.1 ± 0.0 e AV Lipozyme 435 90 A 22.0 ± 0.0 de BC 26.0 ± 0.0 e B 25.1 ± 0.4 d D 28.2 ± 0.2 e C 26.5 ± 0.5 c 65 A 22.1 ± 0.5 de A 22.3 ± 1.2 cd A 22.2 ± 0.2 c AB 23.4 ± 0.6 bc B 25.9 ± 0.8 c 40 A 21.0 ± 0.4 cd AB 21.5 ± 0.3 bc AB 21.6 ± 0.2 c AB 21.5 ± 0.7 b B 22.1 ± 0.9 b Lipozyme RM 90 B 19.5 ± 0.4 ab AB 19.2 ± 0.7 a AB 19.0 ± 0.3 ab AB 17.9 ± 0.7 a A 17.4 ± 0.2 a 65 B 20.0 ± 0.3 bc B 19.8 ± 0.3 ab A 18.2 ± 0.1 a A 17.7 ± 0.7 a AB 18.9 ± 0.5 a 40 A 18.8 ± 0.1 a A 19.4 ± 0.2 a A 19.7 ± 0.1 b A 18.2 ± 1.5 a A 19.2 ± 0.0 a Lipozyme TL 90 A 22.8 ± 0.1 e A 24.1 ± 0.1 de AB 24.8 ± 0.1 d B 26.8 ± 0.7 de C 33.1 ± 1.2 d 65 A 21.4 ± 0.3 d A 21.7 ± 0.2 bc B 24.8 ± 0.5 d B 24.5 ± 0.6 cd B 25.1 ± 0.5 c 40 A 21.4 ± 0.5 d A 22.0 ± 0.3 c A 23.1 ± 0.8 c A 22.3 ± 0.5 bc A 22.4 ± 0.2 b TBARS Lipozyme 435 90 B 60 ± 4 cd B 58 ± 4 cd A 37 ± 3 ab A 34 ± 6 b A 42 ± 2 b 65 A 55 ± 4 abcd A 48 ± 4 abc A 43 ± 2 bc A 49 ± 5 c A 54 ± 4 c 40 AB 57 ± 3 bcd AB 56 ± 3 bcd A 54 ± 0 c AB 64 ± 0 d B 70 ± 2 d Lipozyme RM 90 D 42 ± 4 a CD 39 ± 4 a BC 28 ± 1 a AB 20 ± 0 a A 15 ± 1 a 65 C 51 ± 4 abc C 45 ± 4 ab B 33 ± 2 ab A 22 ± 3 ab A 14 ± 0 a 40 A 43 ± 3 ab B 59 ± 3 cd CD 73 ± 4 d D 85 ± 2 e BC 69 ± 3 d Lipozyme TL 90 B 66 ± 3 d B 61 ± 4 d A 30 ± 1 a A 24 ± 1 ab A 22 ± 2 a 65 C 64 ± 3 cd BC 61 ± 1 d BC 54 ± 2 c B 50 ± 4 cd A 37 ± 5 b 40 A 52 ± 9 abcd BC 110 ± 11 e D 143 ± 7 e C 120 ± 5 f B 105 ± 2 e Values with different small letters in the same column for each oxidation product are significantly different when applying the Tukey's honestly significant difference (HSD) method at p-value ≤ 0.05. Values with different capital letters in the same row for each lipase at a certain incubation temperature are significantly different when applying the Tukey's honestly significant difference (HSD) method at p-value ≤ 0.05. MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT 33 Table 6. Parameters of equation 5 with an Arrehnius type dependence on temperature for k parameter for oxidation products quantified by PV, AV and TBARS for the samples in contact with Lipozyme 435, Lipozyme RM and Lipozyme TL. Immobilized lipase PV AV TBARS k o E a , kJ/mol n R 2 k o E a , kJ/mol n R 2 k o E a , kJ/mol n R 2 Lipozyme 435 5.9·10 3 26.5 0.59 0.998 2.3·10 5 45.2 1.03 0.968 2.2·10 2 19.9 0.78 0.973 Lipozyme RM 9.6·10 2 21.2 0.68 0.999 8.9·10 4 32.03 0.76 0.989 3.2·10 2 19.6 0.78 0.993 Lipozyme TL 9.0·10 3 28.7 0.76 0.992 5.3·10 5 46.7 0.84 0.919 1.0·10 4 33.5 1.28 0.995 MANUS CRIP T ACCEP TED ACCEPTED MANUSCRIPT Highlights • PV and TBARS increase faster than AV in the autoxidation of sardine oil. • Oxidation products concentration decreased in the presence of immobilized lipases. • The decrease was higher for PV and TBARS than anisidine reacted compounds. • Lipozyme RM presented the best results to obtain lower oxidation indices.