Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effect of oil blending and use of natural antioxidants on the chemical composition and thermal stability… Du, Wenqun 2001

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2001-0185.pdf [ 8.35MB ]
Metadata
JSON: 831-1.0090037.json
JSON-LD: 831-1.0090037-ld.json
RDF/XML (Pretty): 831-1.0090037-rdf.xml
RDF/JSON: 831-1.0090037-rdf.json
Turtle: 831-1.0090037-turtle.txt
N-Triples: 831-1.0090037-rdf-ntriples.txt
Original Record: 831-1.0090037-source.json
Full Text
831-1.0090037-fulltext.txt
Citation
831-1.0090037.ris

Full Text

EFFECT OF OIL BLENDING AND USE OF NATURAL ANTIOXIDANTS ON THE CHEMICAL COMPOSITION AND THERMAL STABILITY OF OILS RICH IN POLYUNSATURATED FATTY ACIDS By WENQUN DU M. Sc., Wuxi University of Light Industry, 1993 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Food Science We accept this thesis as confirming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA April 2001 ©Wenqun Du, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada DE-6 (2/88) ABSTRACT The procedure of blending co-3 polyunsaturated fatty acids (PUFAs) oils with a medium chain, saturated butter oil was initiated for the purpose of minimizing the loss of co-3 fatty acids to oxidation reaction. The two sources of u)-3 PUFA used were flaxseed oil (rich in co-3 linolenic acid) and fish oil (rich in very long chain co-3, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA). A common consequence of the dilution of PUFAs by saturated fatty acids present in butter oil was a reduced peroxide value (PV), lower conjugated diene hydroperoxides (CDH) and thiobarbituric acid reactive substances (TBARS) following accelerated oxidation induced by thermal treatment at both 150°C and 180°C, for periods of 8, 16, and 24 h, respectively. The P/S (polyunsaturated to saturated fatty acid) and u)3/u)6 (u)3 to co6 fatty acid) ratios were both shown to represent predictive indicators of PUFA depletion when fitted to a quadratic model describing the influence of the blend proportions derived from fish oil and butter oil. No optimum could be reached for a blending composition to maximize these indicator ratios of PUFA retention. Moreover, it was determined that both EPA (C20:5co3) and DHA (C22:6co3) in fish oil were in fact, more susceptible to thermal oxidation after blending with butter oil. Blending of flaxseed oil with butter oil improved resistance to thermal oxidation as indexed by PV, CDH, and TBARS throughout 8, 16 and 24 h heating periods. A regression, quadratic model effectively optimized the blend of butter oil with flaxseed oil (i.e. 76% flaxseed oil; 24% butter oil) in respect to maximizing the co3/co6 ratio. • The procedure of cholesterol resulted in a greater stability of the flaxseed oil, as evidenced by significant reductions in TBARS and greater retention of co-3 and co-6 ii PUFAs (P<O.GT). Cholesterol oxidation products (COPs) were identified and shown to vary in production as a result of absolute temperature of heating, time of heating and source of oil used in the heating experiment. The combination of adding natural antioxidants, such as rosemary, sage extracts and citric acid was also investigated as a different blending strategy to prevent lipid oxidation in these co-3 rich oil sources. A quadratic model described the optimization of different proportions of the three antioxidants resulting in a combination of 0.067% - rosemary extract, 0.067% sage extract and 0.05% citric acid for a maximal stability of the oil (as evidenced by the lowest formation of CDH). The addition of an optimal mixture of 0.1% rosemary extract, 0.1% rosemary extract with 0.05% citric acid was found to also facilitate a greater retention of linoleic acid and a-linolenic acid in flaxseed oil. iii Table of Contents Abstract ii Table of Contents iv List of Tables x List of Figures xiii List of Abbreviations xix Acknowledgments xx General Introduction 1 Literature Review 4 I. Lipid Oxidation 4 II. Methodology of evaluating fatty acid composition and lipid oxidation 5 III. Exploration of natural antioxidants 9 IV. Heat induced oxidation in food processing and adverse effect of over-heated oil 15 V. Essential fatty acids, and related health benefits, or risk, of flaxseed oil, butter oil and fish oil 19 VI. Cholesterol and autoxidation of cholesterol 23 VII. Response surface methodology (RSM) 27 Chapter 1. Blending of Fish Oil with Butter Oil: Effect of blending on fatty acid profiles and thermal oxidative stability 28 Introduction 28 Materials and Methods 30 Oil sources 30 Block heater 30 iv Other chemicals 31 Experimental design 31 1. Constrained mixture design 31 2. Formulation of oil blends with fish and butter oils 32 3. Heating studies 32 Analysis of oxidation of oil blends 33 1. Thiobarbituric acid reactive substances (TBARS) 33 2. Conjugated diene hydroperoxides (CDH) 34 3. Peroxide value (PV) 35 4. Fatty acid quantification 36 a. Fatty acid calibration curve 36 b. Sample preparation 36 c. Gas chromatography 37 d. Fatty acid profile of native oils 38 Statistical analysis of data 38 Results 40 Fatty acid profiles of blended oils with fish oil and butter oil 40 Assessment of a quadratic model used to predict the optimal combinations of fish and butter oil blends, using P/S and co3/a)6 ratios 41 Depletion of PUFAs during thermal incubation 43 Evaluation of blending and heat processing on the absolute amount of PUFAs of butter-fish oil blends 46 Effect of blending strategy on the depletion of PUFA among butter-fish oil blends during thermal treatments 47 Information of lipid oxidation from chemical analyses 49 Discussion 52 co-3 to to-6 fatty acid ratio in native oils and oil blends 52 The effect of oil blending on the lipid oxidation 53 Effect of temperature and time on the thermal oxidation of butter-fish oil blends 54 Oxidation of co-3 fatty acids in butter-fish oil blends 55 Conclusion (1) 57 Chapter 2. Blending of Flaxseed Oil with Butter Oil: Effect of blending on fatty acid profiles and thermal oxidative stability, without normalization of cholesterol content 71 Introduction 71 Materials and Methods 73 Oil sources 73 Block heater 73 Other chemicals 73 Experimental design 73 1. Constrained mixture design 74 2. Formulation of oil blends with flaxseed and butter oils 75 3. Heating studies 75 Analysis of oxidation of oil blends 75 1. Thiobarbituric acid reactive substances (TBARS) 75 2. Conjugated diene hydroperoxides (CDH) 75 3. Peroxide value (PV) 75 4. Fatty acid quantification 75 a. Fatty acid calibration curve 75 b. Sample preparation 76 c. Gas chromatography 76 d. Fatty acid profile of native oils 76 Statistical analysis of data 76 Results 77 Fatty acid profiles of blended oils with flaxseed oil and butter oil 77 Assessment of a quadratic model used to predict the optimal combinations of flaxseed and butter oil blends, using P/S and co3/co6 ratios 78 Depletion of PUFAs during thermal incubation 79 Evaluation of blending and heat processing on the absolute amount of PUFAs of butter-flaxseed oil blends 81 Effect of blending strategy on the depletion of PUFA among butter-flaxseed oil blends during thermal treatments 82 Information of lipid oxidation from chemical analyses 84 Discussion 86 co-3 to co-6 fatty acid ratio in native oils and oil blends 86 Effectiveness of high temperature on the lipid oxidation 87 The effect of oil blending on the lipid oxidation 88 Prediction of the relative loss of co-3 PUFA 89 Relative oxidation rate of PUFAs in flaxseed and fish oils 90 Conclusion (2) 92 Chapter 3. Blending of Flaxseed and Fish oils with Butter oil, with normalization of cholesterol content 104 Introduction 104 Chapter 3a. Effect of blending on fatty acid profiles and thermal oxidative stability. 106 Materials and Methods 106 Oil sources 106 Block heater 106 vi Other chemicals 106 Experimental design 106 1. Constrained mixture design 107 2. Formulation of oil blends with flaxseed and butter oils 108 3. Heating studies 108 Analysis of oxidation of oil blends 108 1. Thiobarbituric acid reactive substances (TBARS) 108 2. Conjugated diene hydroperoxides (CDH) 108 3. Peroxide value (PV) 108 4. Fatty acid quantification 108 a. Fatty acid calibration curve 108 b. Sample preparation 108 c. Gas chromatography 108 d. Fatty acid profile of native oils 108 Statistical analysis of data 109 Results 110 Fatty acid profiles of blended oils with cholesterol modified flaxseed oil and butter oil 110 Evaluation of a quadratic model used to predict the optimal combinations of flaxseed and butter oil blends containing cholesterol, assessed by the P/S and 0)3/(o6 fatty acid ratio 110 Depletion of PUFAs during thermal incubation I l l Effect of blending strategy on the depletion of PUFAs among butter-cholesterol modified flaxseed oil blends during thermal treatments 112 Information of lipid oxidation from chemical analyses 113 Discussion 115 Conclusion (3a) 118 Chapter 3b. Characterization and quantitation of cholesterol oxidation products derived from thermally treated oil blends 119 Materials and Methods 119 Oil source 119 Other chemicals 119 Cholesterol and cholesterol oxidation products (COPs) determination 119 1. Cholesterol calibration curve 120 2. Sample preparation 120 3. Gas chromatography 121 Statistical analysis of data 122 Results 123 Cholesterol content of oils 123 Thermal degradation of cholesterol 123 vii 1. Cholesterol in butter-fish oil blends 123 2. Residual cholesterol concentration in butter-cholesterol modified flaxseed oil blends 124 3. Comparison of butter-fish oil blends and butter-cholesterol modified flaxseed oil blends in cholesterol content 125 Quantitation of butter-fish oil blends and butter-cholesterol modified flaxseed oil blends in cholesterol content 126 Discussion 128 Conclusion (3b) 135 Chapter 4. Blending of natural phytochemicals with citric acid on the fatty acid profiles and oxidative stability of flaxseed oil 145 Introduction 145 Materials and Methods 146 Oil sources 146 Block heater 146 Antioxidants 146 Other chemicals 140 Experimental design 1. Formulation antioxidant combinations 146 2. Heating studies 147 Analysis of lipid oxidation of oil with antioxidant combinations 148 1. Free radical scavenging capacity of rosemary and sage extracts 148 a. Free radical scavenging activity of rosemary, sage 148 b. Synergistic effect of rosemary and sage extracts on free radical scavenging capacity 148 2. Thiobarbituric acid reactive substances (TBARS) 149 3. Conjugated diene hydroperoxides (CDH) 149 4. Fatty acid quantification 149 Statistical analysis of data 149 Results 150 Free radical scavenging effects of rosemary and sage extracts and their synergistic effect 150 Changes in fatty acid profiles during heating 150 The formation of CDH and TBARS in antioxidant treated, heated flaxseed oil 152 Optimization of antioxidant combinations 152 Discussion 153 Conclusion (4) 157 viii General Conclusions and major Findings 173 Appendix 179 References 193 ix List of Tables Table 1. Combinations of butter-fish oil blends, containing fish oil blended with butter oil before thermal treatment 59 Table 2. Fatty acid composition of native fish oil, flaxseed oil and butter oil (mg fatty acid/gm oil, P/S ratio and 0)3/(o6 ratio) 60 Table 3. Fatty acid composition of butter-fish oil blends before heating (mg fatty acid/gm oil, P/S ratio and co3/oo6 ratio) 60 Table 4. Quadratic regression of butter-fish oil blends and the changes in P/S ratio at 150°C and 180°C 61 Table 5. Quadratic regression of butter-fish oil blends and the changes in co3/oo6 ratio at 150°Cand 180°C 61 Table 6. Polyunsaturated fatty acid composition (% of total fatty acids), P/S ratio and oo3/oo6 ratio of butter-fish oil blends before and after thermal treatment at 150°C 62 Table 7. Polyunsaturated fatty acid composition (% of total fatty acids), P/S ratio and co3/co6 ratio of butter-fish oil blends before and after thermal treatment at 180°C 63 Table 8. Effect of blending of fish oil and butter oil on peroxide value and the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) after 8, 16 and 24 h thermal treatment at 150°C and 180°C 64 Table 9. Quadratic regression of butter-fish oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8, 16 and 24 h thermal treatment at 150°C and 180°C 65 Table 10. Combinations of flax seed-butter oil blends containing flaxseed oil and butter oil before thermal treatment 94 Table 11. Fatty acid composition of flaxseed-butter oil blends before thermal treatment (mg fatty acid/gm oil, P/S ratio and co3/a>6 ratio) 94 Table 12. Quadratic regression of flaxseed-butter oil blends and the changes in P/S ratio at 150°C and 180°C 95 Table 13. Quadratic regression of flaxseed-butter oil blends and the changes in 033/0)6 ratio at 150°C and 180°C 95 Table 14. Polyunsaturated fatty acid composition (% of total fatty acids, w/w), P/S ratio and co3/co6 ratio of flaxseed-butter oil blends before and after thermal treatment at 150°C and 180°C 96 Table 15. Effect of blending of flaxseed oil and butter oil on the peroxide value (PV) and the formation of conjugated diene hydroperoxides (CDH) and thiobarbituric acid reactive substances (TBARS) after 8, 16 and 24 h thermal treatment at 150°C and 180°C 97 Table 16. Quadratic regression of flaxseed-butter oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8, 16 and 24 h thermal treatment at 150°C and 180°C 98 Table 17. Quadratic regression of cholesterol modified flaxseed-butter oil blends and the changes in P/S ratio at 150°C and 180°C 136 Table 18. Quadratic regression of cholesterol modified flaxseed-butter oil blends and the changes in 003/0)6 ratio at 150°C and 180°C 136 Table 19. Polyunsaturated fatty acid composition (% of total fatty acids, w/w), P/S ratio and 0)3/co6 ratio of cholesterol modified flaxseed-butter oil blends before and after thermal treatment at 150°C and 180°C 137 Table 20. Effect of blending of cholesterol modified flaxseed oil and butter oil on the peroxide value (PV) and the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) after 8, 16 and 24 h thermal treatment at 150°Candl80°C 138 Table 21. Quadratic regression of cholesterol modified flaxseed-butter oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8, 16 and 24 h thermal treatment at 150°Cand 180°C 139 Table 22. Residual cholesterol content of fish-butter oil blends and cholesterol modified flaxseed-butter oil blends after 8, 16, 24 h heating at 150°C and 180°C 140 Table 23. Formation of oxysterols from fish oil and cholesterol modified flaxseed oil during a 24 h thermal incubation at 150°C and 180°C 141 Table 24. Combinations of proportion of oleoresin rosemary, sage extracts and citric acid added into flaxseed oil before heating 159 Table 25. Synergistic effect of rosemary and sage extracts in scavenging free radical 160 xi Table 26. Quadratic regression coefficients derived from fatty acid composition, P/S ratio and n-3/n-6 ratio of flaxseed oil with antioxidant blends of rosemary extract, sage extract and citric acid following thermal treatment at 180°C for 2.5, 5, 10,15,20, and 25 h, respectively 161 Table 27. Quadratic regression coefficients derived from the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) of flaxseed oil with antioxidant blends of rosemary extract, sage extract and citric acid following thermal treatment at 180°C for 2.5, 5, 10,15,20, and 25 h, respectively 163 xii List of Figures Figure 1. Structures of antioxidant compounds identified from rosemary and sage 12 Figure 2. Biological transformation of linoleinic acid (18:3u)3) 20 Figure 3. Structure of cholesterol and its oxidative products 26 Figure 4. Comparison of experimental and hypothetical values on the changes in docosahexaenoic acid (DHA, C22:6co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of DHA, • = hypothetical value of DHA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butte oil, E = 100% butter oil 66 Figure 5. Comparison of experimental and hypothetical values on the changes in eicosapentaenoic acid (EPA, C20:5u)3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of EPA, • = hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butte oil, E = 100% butter oil 66 Figure 6. Comparison of experimental and hypothetical values on the changes in docosahexaenoic acid (DHA, C22:6co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of DHA, • = hypothetical value of DHA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butte oil, E = 100% butter oil 67 Figure 7. Comparison of experimental and hypothetical values on the changes in eicosapentaenoic acid (EPA, C20:5to3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of EPA, • = hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil 67 Figure 8. Comparison of experimental and hypothetical values on the changes in stearic acid of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of EPA, • = hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil 68 Figure 9. Comparison of experimental and hypothetical values on the changes in stearic acid of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of EPA, • = xiii hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil 68 Figure 10. The content of malonaldehyde (MDA) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% fish oil, • 66.7% fish oil and 33.3% butter oil, • 50% fish oil and 50% butter oil, X 33.3% fish oil and 66.7% butter oil, x 100% butter oil 69 Figure 11. The content of malonaldehyde (MDA) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% fish oil, • 66.7% fish oil and 33.3% butter oil, • 50% fish oil and 50% butter oil, X 33.3% fish oil and 66.7% butter oil, x l 0 0 % butter oil 69 Figure 12. The content of conjugated diene hydroperoxides (CDH) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% fish oil, • 66.7% fish oil and 33.3% butter oil, • 50% fish oil and 50% butter oil, X 33.3% fish oil and 66.7% butter oil, x 100% butter oil 70 Figure 13. The content of conjugated diene hydroperoxides (CDH) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% fish oil, A 66.7% fish oil and 33.3% butter oil, • 50% fish oil and 50% butter oil, X 33.3% fish oil and 66.7% butter oil, x l 0 0 % butter oil 70 Figure 14. Comparison of experimental and hypothetical values on the changes in linolenic acid (C18:3co3) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of linolenic acid, • = hypothetical value of linolenic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 99 Figure 15. Comparison of experimental and hypothetical values on the changes in linolenic acid (C18:3CD3) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of linolenic acid, • = hypothetical value of linolenic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 99 Figure 16. Comparison of experimental and hypothetical values on the changes in linoleic acid (C18:2co6) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of linoleic acid, • = hypothetical value of linoleic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil xiv + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 100 Figure 17. Comparison of experimental and hypothetical values on the changes in linoleic acid (C18:2u)6) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of linoleic acid, • = hypothetical value of linoleic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 100 Figure 18. Comparison of experimental and hypothetical values on the changes in palmitic acid (C 16:0) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of palmitic acid, • = hypothetical value of palmitic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 101 Figure 19. Comparison of experimental and hypothetical values on the changes in palmitic acid (C16:0) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of palmitic acid, • = hypothetical value of palmitic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil 101 Figure 20. The content of malonaldehyde (MDA) for butter-flaxseed oil blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% flaxseed oil, • 66.7% flaxseed oil and 33.3% butter oil, • 50% flaxseed oil and 50% butter oil, X 33.3% flaxseed oil and 66.7% butter oil, x 100% butter oil 102 Figure 21. The content of malonaldehyde (MDA) for butter-flaxseed oil blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% flaxseed oil, • 66.7% flaxseed oil and 33.3% butter oil, • 50% flaxseed oil and 50% butter oil, X 33.3% flaxseed oil and 66.7% butter oil, x 100% butter oil 102 Figure 22. The content of conjugated diene hydroperoxides (CDH) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% flaxseed oil, • 66.7% flaxseed oil and 33.3% butter oil, • 50% flaxseed oil and 50% butter oil, X 33.3% flaxseed oil and 66.7% butter oil, x 100% butter oil 103 Figure 23. The content of conjugated diene hydroperoxides (CDH) for butter-fish oil blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% flaxseed oil, • 66.7% flaxseed oil and 33.3% butter oil, • 50% flaxseed oil X V and 50% butter oil, X 33.3% flaxseed oil and 66.7% butter oil, x 100% butter oil 103 Figure 24. The comparison of residual linoleic acid (C18:2o)6) of flaxseed oil with and without cholesterol supplementation following thermal treatment at 150°C. • = without cholesterol, • = with cholesterol, * = P<0.05 142 Figure 25. The comparison of residual linolenic acid (C18:3ct>3) of flaxseed oil with and without cholesterol supplementation following thermal treatment at 180°C. • = without cholesterol, • = with cholesterol, * = P<0.05 142 Figure 26. The comparison of residual linoleic acid (C18:2oo6) of flaxseed oil with and without cholesterol supplementation following thermal treatment at 180°C. • = without cholesterol, • = with cholesterol, ** = P<0.01 143 Figure 27. The comparison of residual linolenic acid (C18:3oo3) of flaxseed oil with and without cholesterol supplementation following thermal treatment at 180°C. • = without cholesterol, • = with cholesterol, ** = P<0.01 143 Figure 28. Effect of heating on the loss of cholesterol from fish oil during a 24 h thermal incubation at 150°C and 180°C, respectively. • = 150°C, • = 180°C 144 Figure 29. Effect of heating on the loss of cholesterol from fish oil during a 24 h thermal incubation at 150°C and 180°C, respectively. • = 150°C, • = 180°C 144 Figure 30. Concentration-dependent effect of rosemary extract and sage extract on scavenging O.lmM DPPH radical in ethanol solution. • = rosemary extract, • = sage extract 164 Figure 31. The variation of residual linoleic acid (C18:2(o6) in flaxseed oil with and without addition of antioxidant following 5 h thermal treatment at 180°C. S = sage extract, R = rosemary extract and CA = citric acid 165 Figure 32. The variation of residual linoleic acid (C18:2oo6) in flaxseed oil with and without addition of antioxidant following 25 h thermal treatment at 180°C. S = sage extract, R = rosemary extract and CA = citric acid 166 Figure 33. The variation of residual linoleic acid (C18:3co3) in flaxseed oil with and without addition of antioxidant following 5 h thermal treatment at 180°C. S = sage extract, R = rosemary extract and CA = citric acid 167 Figure 34. The variation of residual linoleic acid (C18:3tt>3) in flaxseed oil with and without addition of antioxidant following 25 h thermal treatment at 180°C. S = sage extract, R = rosemary extract and CA = citric acid 168 xvi Figure 35. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the formation of conjugated diene hydroperoxides (pmol/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 169 Figure 36. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the formation of malonadehyde (MDA) (u,mol/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 170 Figure 37. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the amount of linoleic acid (mg/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 171 Figure 38. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the amount of linolenic acid (mg/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 172 xv i i Appendix figure A. Calibration curves for individual fatty acids 179 Appendix figure B. Mixture standards of fatty acid methyl esters 181 Appendix figure C. Quadratic model derived from co3/co6 ratio, or P/S ratio of butter-fish oil blends 182 Appendix figure D. Quadratic model derived from co3/co6 ratio, or P/S ratio of butter-flaxseed oil blends 183 Appendix figure E. The retention of PUFAs from fish oil or flaxseed oil following thermal treatment for 8, 16, and 24 h at 150°C, respectively. • eicosapentaenoic acid (C20:5co3) from fish oil, • docosahexsaenoic acid (C22:6co3) from fish oil, • linolenic acid (C18:3u)3) from flaxseed oil, X linoleic acid (C18:2oo6) from flaxseed oil 184 Appendix figure F. The retention of PUFAs from fish oil or flaxseed oil following thermal treatment for 8, 16, and 24 h at 180°C, respectively. • eicosapentaenoic acid (C20:5co3) from fish oil, • docosahexsaenoic acid (C22:6co3) from fish oil, • linolenic acid (C18:3co3) from flaxseed oil, X linoleic acid (C18:2co6) from flaxseed oil 185 Appendix figure G. The retention of PUFAs from fish oil or flaxseed oil following thermal treatment for 8 h at 150°C or 180°C, respectively. • = 150°C,Q= 180°C 186 Appendix figure F£. Thermal decomposition of hydroperoxy cyclic peroxidation from linoleate treated with l02 187 Appendix figure I. Calibration curve for cholesterol 188 Appendix figure J. • A representative chromatogram of derivatized cholesterol oxide standards 189 Appendix figure K. Major pathways of cholesterol oxidation 190 Appendix figure L. Hydrogen donation mechanism of rosmanol 191 Appendix figure M. Reaction between carnosic acid and lipid peroxyl radical 192 x v i i i List of Abbreviations CDH Conjugated diene hydroperoxides COPs Cholesterol oxidation products MDA Malonaldehyde DHA Docosahexaenoic acid (C22:6o)3) EPA Eicosapentaenoic acid (C20:5o)3) P/S Polyunsaturated to saturated fatty acid ratio PUFA Polyunsaturated fatty acid PV Peroxide value RSM Response surface methodology TBA Thiobarbituric acid TBARS Thiobarbituric acid reactive substances 0)3/0)6 0)-3 to 0)-6 fatty acid ratio xix Acknowledgments I would like to express my gratitude to many individuals that assisted me in finishing this thesis. I want to acknowledge Food Science for the use of facilities. I wish to express my sincere gratitude to Dr. D. Kitts, Professor, for his endless supervision and extreme patience with the preparation of my thesis. Also, his knowledge and friendship encouraged me during the course of this study. This thesis would not be completed without the support from the following individuals: Dr. E. Li-Chan, Dr. T. Durance, Dr. J. Richards, Dr. C. Seaman, Dr. B. Skura, Mr. Sherman Yee, Mrs. Val Skura, Mrs. Joyce Tom, Mrs. Jeannette Law and many other co-workers. I wish to acknowledge my parents for their support during the preparation of this thesis. In addition, the support of my husband Chun and the cooperation of my daughter Carolyn enabled me to finish this study. X X General Introduction Both marine oil and flaxseed oil are excellent sources of co-3 fatty acids. This particular group of fatty acids are generally regarded as essential fatty acids, because humans, like all mammals, cannot synthesize them endogenously and therefore must rely on the diet as the sole source. Linolenic acid (18:3co3) and very long chain co-3-eicosapentaenoic acid (EPA, C20:5co3) and docosahexanoic acid (DHA, C22:6co3), as well as co-6 PUFAs such as linoleic acid (C18:2co6) are particularly susceptible to oxidation, especially under thermal conditions. Currently, a higher demand for co-3 fatty acids in the North American diet has encouraged domestic animal producers to make available attractive dietary sources of co-3 fatty acids by using fish meal or flaxseed as a supplement for animal feeding (Aii et al., 1998; Lopez et al., 1999; Simopoulos, 1999; Enser et al., 2000; Sheard et al., 2000; Elmore et al., 2000). Potential alternation of fatty acid composition resulting in enhanced co-3 polyunsaturated fatty acids (PUFA) has raised the question about oxidative stability of foods that contain these PUFAs, since fish oil and flaxseed oil are generally not suitable for salad oil or cooking oil (Swern, 1979). Animal products are also a source of cholesterol; for example, beef contains 40mg cholesterol/lOOg muscle (Duckett and Wagner, 1997). It is necessary to assess the relative contribution of cholesterol and PUFAs in total lipid peroxidation reactions. Though PUFAs have been extensively studied and reviewed (Frankel, 1991; Coupland and McClements, 1996; Min, 1997), the study of the interaction between PUFA and cholesterol during food processing is still not clear. There is no doubt that cholesterol is oxidizable (Smith, 1987), and that cholesterol oxidative products have been found to be 1 associated with atherosclerosis (Steinberg, 1989) and food off-flavors (Kitts, 1996). Some of the oxidative derivatives of cholesterol such as cholestane triol, 25-hydroxyl-cholesterol are thought to be the most atherogenic oxysterol (Imai et al, 1976, Pend et al, 1979, Blankenship et al, 1991). In order to prevent, or retard, the oxidation of PUFAs, a number of strategies have been suggested and attempted. Besides the effect of blending PUFA with oils possessing lipid constituents that have a less tendency for oxidation, the addition of natural antioxidants may also be a more practical method to prevent lipid oxidation. Rosemary and sage extracts have been studied for more than twenty years (Chang et al, 1977). Rosemary has been studied for its antioxidant activity in stabilizing rapeseed oil and palm oil, during elevated temperature treatment (Reblova et al, 1999; Jaswir et al, 2000). Rosemary and sage have also been found to have synergistic effect, for example, rosemary and sage extracts are synergistic with BHA, BHT and ascorbyl palmitate, but not citric acid in lard (Banias et al, 1992). The objectives of this thesis were: a. To determine if butter oil, as a source of medium chain saturated fatty acids, can protect the oxidation of w-3 PUFAs derived from fish oil and flaxseed oil when blended with these oils. b. To determine if the presence of cholesterol in the blended oil mixture will influence the rate of lipid oxidation, and thus act to preserve the PUFAs in the oil blends. c. To determine if cholesterol oxidation reactions are influenced by the composition of the PUFA content in the lipid sources in which it is contained. 2 d. To determine if natural antioxidants, found in rosemary and sage extracts can protect against thermal oxidation of PUFA-rich oils. e. To demonstrate an antioxidant synergism between rosemary and sage extracts with citric acid and related effectiveness in stabilizing PUFA-rich oils. On the basis of these objectives, the hypotheses of this thesis are listed as follows: a. The blending of fish oil or flaxseed oil with butter oil will retain labile PUFAs and minimize lipid oxidation reactions associated with the unsaturated fatty acid profile of the blend. b. The presence of cholesterol in the mixed oil blend will not significantly influence the retention of labile PUFAs and generation of lipid oxidation products following a thermal process. c. The stability of cholesterol against oxidation reactions will depend on the fatty acid composition and the content of PUFAs in the mixed oil blend. d. Natural antioxidants derived from rosemary and sage extracts can be used to stabilize PUFA-rich oils from thermal-induced oxidation reactions. 3 Literature review I. Lipid oxidation Lipid oxidation occurs especially with lipids that contain high amounts of PUFAs and is a concern to food scientists in terms of food safety, nutritional value and organoleptic acceptance. Autoxidation of PUFAs in food involves a free radical chain reaction that is generally initiated by the exposure of lipids to light, heat, ionizing radiation, transition metal ions or metalloprotein catalysts (Shahidi and Nazck, 1995; Min, 1997). Autoxidation includes initiation (equation 1, production of lipid free radicals), propagation (equation 2-3, extension of lipid free radicals) and termination (equation 4-6, production of stable products) reactions, as illustrated in the following scheme: Initiation: Propagation: Termination: RH -» R* + H* R* +02 -> ROO* ROO* + RH -> R' + ROOH R* +R* -> R-R R' + ROO'^ROOR ROO* + ROO* -» ROOR + 0 2 equation 1 equation 2 equation 3 equation 4 equation 5 equation 6 of which, R\ ROO* stand for alkyl radical and lipid peroxy radical, respectively. The mechanism of phenolic antioxidant activity is believed to involve free radical termination; equation 7 was first proposed by Boland and ten-Have (1947, cited from Shahidi and Naczk, 1995). The phenolic antioxidant (AH) interferes with lipid oxidation by rapidly donating a hydrogen atom to the lipid radical that acts to neutralize the reactive activity of lipid radicals, as illustrated by: ROO* + AH -» ROOH + A* equation 7 RO* + AH —> ROH + A* equation 8 The lower reaction potential of antioxidant intermediate radical (A*) is important for the activity of primary antioxidant, therefore, antioxidant intermediate radical does not initiate further free radical chain reactions. The activity of phenolic antioxidants is also dependent on temperature and other oxidation conditions (Frankel, 1993). II. Methodology of evaluating fatty acid composition and lipid oxidation About 99% of fatty acid analysis in foods is executed by gas chromatography following methyl ester derivatization (Ackman 1992). Other ester derivatization procedures are used, for example, isopropyl esters (Peuchant et al., 1989). The reason for using fatty acid esters, rather than fatty acids themselves is that fatty acids may associate with the vapor phase and are absorbed to any convenient surface resulting in tailing and/or ghosting peaks (Ackman 1972). As well, much higher temperatures are required for GLC of free fatty acids than for methyl ester, thus granting the latter an advantage for GLC. During the period between 1980-1990, the flexible fused silica column replaced the packed column. The fused silica column is a modern version of the wall-coating open-tubular column (Ackman 1992). Various methods are available for the measurement of lipid oxidation in food. Changes in chemical, physical or organoleptic properties of fats and oils during oxidation 5 can be monitored to assess the extent of lipid oxidation; however, there are no standard and uniform methods to determine collectively all the oxidative changes in food systems. Methods that measure primary changes of lipids can be classified by those quantifying the loss of reactants (e. g. unsaturated fatty acids) (Shahidi and Wansundara, 1997); however, measuring changes in fatty acid composition is not widely used in assessing lipid oxidation because total lipid extraction from food, and subsequent conversion into derivatives suitable for gas chromatographic analysis may be required. This method is a useful technique for identifying the different potential classes of lipid and fatty acids that are involved in oxidation changes (Gray and Monahan, 1992). Regarding the oxidation of more saturated lipid, changes in fatty acid composition is debatable (Shahidi, et al., 1997). On the other hand, oxidative changes in marine oils and highly unsaturated vegetable oils can be monitored using this as the indicator of oxidative damage. Similarly, changes in iodine value due to loss of unsaturation during accelerated oxidation have also been used as an index of lipid oxidation. Peroxide value (PV), which is measured by thiosulfate titration, describes the formation of the allylic hydroperoxide primary product of lipid oxidation. PV is based on the determination of hydroperoxide content by a titration method, according to the reduction of the hydroperoxide group (ROOH) with iodine ion (T). The amount of iodine (I2) liberated is proportional to the concentration of peroxide present. Released I2 is assessed by titration against a standardized solution of sodium thiosulfate (Na2S203) using a starch indicator. The reactions involved are as follows: 2ROOH + 2H+ + 2KI - » I 2 +2ROH + H 2 0 +K20 equation 9 I2 +2Na2S203 -> Na 2S 40 6 + 2NaI. equation 10 6 Potential drawbacks of this method are the absorption of iodine at unsaturation sites of the fatty acids and the liberation of iodine from KI by the oxygen present in the solution to be titrated. This iodometric method, for the determination of PV is applicable to all normal fats and oils, and represents a highly empirical and variable procedure for monitoring lipid oxidation. This method also fails to adequately measure low PV, due to the difficulty in determining the titration end point (Shahidi and Wanasundara, 1997). Oils containing linoleate, or more highly unsaturated fatty acids, are oxidized to conjugated diene systems and absorption increases proportionately to the uptake of oxygen and to the formation of peroxides at the early stage of oxidation (Gray, 1978). Oxidation of PUFAs is accompanied by an increase in the ultraviolet absorption of product. Lipids containing methylene-interrupted diene or polyenes, show a shift in a double-bond position during oxidation that is due to isomerization and conjugate formation. The resulting conjugated diene exhibits an intense absorption at 234nm; similarly, conjugated trienes absorb at 268nm.The absorption increases due to the formation of conjugated diene and is proportional to the uptake of oxygen and the formation of peroxide during the early stage of oxidation. Shahidi et al. (1994) and Wanasundara et al. (1995) concluded that the conjugated diene method could be used as an index of stability of lipid in place of, or in addition to, PV. The advantage of the conjugated diene method is that it is faster than the PV determination. It is also much simpler, does not depend on the condition of the chemical reaction or color development, and requires a smaller sample size. However, the presence of compounds also absorbing in the UV region of conjugated diene formation will interfere with determinations. 7 The primary oxidation products (hydroperoxides) of lipids are transitional intermediates that decompose into various secondary products. These secondary products include aldehyde, ketones, hydrocarbons and alcohol. TBA (2-thiobarbituric acid) value or 2-thiobarbituric acid reaction substances (TBARS) is the most widely used method to evaluate oxidative stability of oil and the effectiveness of antioxidants to inhibit oxidation, by measuring the secondary products (mainly malonaldehyde) of oxidation (Gunstone 1996). The TBARS test is more sensitive when used with PUFAs that contain three or more double bonds (Frankel, 1993). Different procedures can be used for the TBA value. The TBARS method is sensitive and precise, but results can be misleading, since TBA reacts with a wide range of secondary oxidative products of lipid and other food ingredients, if the test is based on a complex food system (Gray, 1978; Halliwell et al., 1987; Frankel, 1993). For example, sucrose reacts with TBA to give a red color that interferes with the extent of oxidation (Baumgartner et al., 1975). Other factors regarding the precision of TBA value include many modifications of the methodology (Marcuse and Johansson, 1973; Ke and Woyewoda, 1979; Shahidi et al., 1985; Thomas and Fume, 1987; Schmedes and Holmer, 1989), and therefore it is hard to compare the results from different studies. However, the TBA value is still a widely used methodology in studying lipid oxidation, and is often regarded as an excellent means of evaluating the relative oxidative state of a food system, as affected by storage conditions or processing conditions (Gray and Pearson, 1987). The automatic Rancimat method also offers an automated method to determine the oxidative stability of oils at elevated temperature (100-140°C). This procedure monitors the change of conductivity derived from the formation of short-chain fatty acids 8 (Frankel, 1993). The disadvantage of the Rancimat method is that higher levels of lipid oxidation are required in order to reach the detectable change of conductivity (deMan et al, 1987). Sensory methods, based on odor and flavor evaluation, provide the most useful information related to consumer acceptance of food products; however, they are highly dependent of the quality of trained panelists (Frankel, 1993). Alternatively, volatile compounds derived from oxidative lipid can be analyzed by headspace-gas chromatography (Huang et al, 1996; Medina et al, 1999), and ethane and pentane are the usual end-products monitored for the following peroxidation of to-3 and to-6 PUFA (Kneepkens, 1997). Generally, the degree of oxidation should be determined at suitable time intervals by measuring different types of peroxidation products, including initial products and decomposition products of lipid oxidation. In order to save time and money, screening tests using accelerated model systems involving by heating, the addition of transitional metal ions and/or oxidative radical initiators, are required and are very practical (Chang et al, 1977; Kahl and Hildebrandt, 1986; Meyer, 1994; Plumb et al, 1996; Hu and Kitts, 2000). The most important consideration in evaluating the extent of oxidation is that single "standardized" methods cannot reflect the whole spectrum of lipid oxidative stability (Frankel, 1994; Meyer, 1994). Therefore, more than one method should be used to analyze both initial and decomposition products during lipid oxidation (Kitts, 1996). III. Exploration of natural antioxidants During the past 40 years, considerable interest in defining the benefits of plant secondary metabolites as food preservatives or antioxidants (Cuppett and Hall, 1998) has 9 occurred. Major breakthroughs in this area were achieved within the past 20 years, when more structured analytical techniques were introduced in areas such as the separation, isolation and testing of active components derived from plants. The reason behind using plant-derived compounds for application in the food industry is the increasing concern over the safety of the synthetic antioxidants, such as butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT) and tertiary butylated hydroquinone (TBHQ) (Iverson, 1995; Cuvelier et al, 1994; Banias et al, 1992, Wong et al, 1995). Besides the reasons stated above, BHT and BHA are also quite volatile and easily decompose at high temperature, and consequently do not provide satisfactory effectiveness for some common food products, such as French fries and potato chips that involve storage (Irwandi, 2000). Also, these antioxidants are not effective in vegetable oils and in preventing the development of initial off-flavors such as reversion flavor. Though addition of TBHQ results in a lower peroxide value in soybean oil, it also develops a strong objectionable flavor (Chang et al, 1977). Research on the antioxidant activities in both biological model and food model systems has been extensively conducted during the past decade. For example, over 13,800 papers have been cited in the Biological Abstracts from 1990 to the present date. Of the natural antioxidant extracts, studies have been reported on the Labiatae (Lamiaceae) family, which induces oregano (Origanum vulgare L) (Lolos et al, 1999; Marinova and Yanishilieva, 1997), rosemary (Rosmarinus officinalis L.) and sage (Salvia officinalis L) (Chipautl et al, 1952; Hall et al, 1962; Houlihan et al, 1984; Inatani et al, 1983; Schwarz et al, 1992) as well as thyme (Thymus vulgaris L.) (Farag et al, 1989; Deighton et al, 1993). More effort recently has been given to elucidate the antioxidant 10 mechanisms of rosemary and sage extracts; a possible reason for this being that both herbs are very good antioxidants as shown in the early pioneer work done by Chipault and coworkers in 1950s (Chipaults et al, 1952; Chipaults et al., 1956) and more recently by Jaswir et al. (2000). Studies of antioxidant activities of rosemary and sage cover a range of whole plant tissues to standardized isolates of single compounds. Dried leaves of rosemary were found to delay the development of warmed over flavor (WOF) of cooked minced meatball (Huisman et al., 1994). The development of commercialized rosemary and sage extract could be credited to the work of Chang et al. (1977), who tried to use different solvents with various polarity to extract rosemary and sage and evaluated the different antioxidant activities in a lard model system held at 60°C in the dark. A methanol extract was found to yield the greatest activity, although no identification of the active compounds was made at the time. The first identified antioxidant compound from rosemary leaves was carnosol (Houlihan et al., 1984). The active compounds in rosemary and sage with antioxidant activities include carnosol, carnosoic acid, rosmanol, rosmaridiphenol, rosemadial, rosemariquinone and rosemarinic acid (structures are shown in Figure 1). Carnosol and rosemanol are both more effective than oc-tocophenol, BHT and BHA when assessed with the active oxygen method (Nakatani and Inatani, 1981). Rosmaridiphenol (0.02%) was reported to possess equivalent antioxidant activity to BHT (Shahidi and Naczk, 1995). Similar results on the purpose of comparing antioxidant activity of different phytochemicals are also available elsewhere (Chen et al., 1992; Aruroma et al, 1996; Richheimer et al, 1996; Thomsen, 1999). 11 rosmarinic acid Figure L Structures of antioxidant compounds identified from rosemary and sage (Nakatani et al., 1981). 12 The use of sage and rosemary extract to replace synthetic antioxidant was initially proposed to overcome the disadvantage of BHT and BHA in thermal oxidation of edible oil (Chang et al, 1977). The loss of BHT, BHA and TBHQ at higher temperature (185°C), resulted from both evaporation and decomposition of the active constituents. The stability of these synthetic antioxidants, identified by thin-layer chromatography (TLC) and gas chromatography (GC), was found to be in the relative order of BHT>BHA>TBHQ (Hamama and Nawar, 1991). The proposed mechanisms of such decomposition-induced reactions of oxidation included rearrangement, etherification, dimerization and addition. In a paraffin-like medium, carnosol was the primary antioxidant derived from rosemary and sage (Wenkert et al, 1964). The rosemary and sage used in former experiments were basically oleoresins, made by solvent extraction of herbs and spices followed by solvent removal. Oleoresin was originally considered to be a concentrated mixture of natural colors, aromas and flavors in vegetable oil, which has been commercially available for more than 40 years (Guzinski 1996). Oleoresins are GRAS (generally recognized as safe) listed. The production procedures for manufacturing this product consist of grinding, extraction, desolvenization and solvent standardization. Oleoresin has been used as a form of natural antioxidant extracts, including oleoresins for rosemary and sage. Again, carnosol and carnosic acid are present at the highest concentration. In a liquid oleoresin, phenolics have a much larger effect than they would have when bound within the cell structure of the ground leaf (Guziniski, 1996). Generally, the commercial rosemary oleoresins are not flavorless, and thus the implication with use in a lipid system may be the potential taste of oleoresin, which should be considered before deciding to use oleoresins in foods as stabilizers. 13 The classification of antioxidant varies with the definition and function of such material in the underlying mechanism to delay oxidation of lipids. For example, Chipault (1962) classified antioxidants as primary antioxidant and synergist, whereas, they could be also classified as primary antioxidants, oxygen scavengers, secondary antioxidants, enzymatic antioxidants and synergists (Kochhar and Rossell, 1990). These compounds are regarded as antioxidants, if they are effective at low concentration, absent of undesirable effects regarding to color, taste and other objective properties of the food. They are also required to be effective and compatible with food, have stability under the conditions of processing and during storage, and ensure that both parent compounds and oxidation products are non-toxic (Pratt. 1996). Conflicting results of rosemary and sage extracts with other antioxidants, for example, a-tocopherol have been reported. Wong et al. (1995) found that the combination of rosemary extract with tocopherol actually exhibited similar antioxidant activity of the tocopherol when present alone in a homogenate beef model, thus indicating no synergism. On the other hand, Wada and coworkers (1992) reported that rosemary extract had a synergistic effect with a-tocopherol in a sardine oil model and frozen-crushed fish meat. This combination was comparable with BHA, from the aspect of delaying the onset of rancidity. Further research conducted by the same group proposed that such a synergism involved a-tocopherol as a free radical scavenger, whilst rosemary extract was a tocopherol regenerator (Fang and Wada, 1993). However, this result does not mean that rosemary and sage extracts do not function as free radical scavengers. In fact, Brand-Williams et al. (1995) demonstrated that one mole of rosmarnic acid could reduce 3.33 mole of DPPH radical. Hall and Cuppett (1997) 14 characterized the structure-activity relationship of rosemary and sage extracts and showed that carososic acid, carnosol, rosmanol, rosmarinic acid and rosmariquinone were all hydrogen donors, though an iron-chelating mechanism of carnosic acid and carosol (Auroma et al., 1992) should also be taken into account. Research from Rutgers University (Wang et al., 1998; Wang et al., 1999) reported that rosmarinic acid, luteolin-7-glucoside from sage were the most active compounds in scavenging a stable 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical, along with the other two less active compounds, i.e., 6-0-caffeoyl-beta-D-fructofuranosyl-(2->l)-alpha-D-glucopyranoside, l-0-caffeoyl-beta-D-apiofuranosyl-(l->6)-beta-D-glucopyranoside. However, no literature has been found on similar test results with rosemary extracts. Heat stability of rosemary and sage extracts were found to be dependent on both temperature and oxygen pressure (Cuppett and Hall, 1998). Schwarz et al. (1992), found that heat resulted in the depletion of rosmanol and carnosol in a lard-based model; however, the antioxidant activity was still detectable even though it was present at a reduced level. The thermal decomposition products of carnosol were recently identified as miltirone and dehydrorosmariquinone; both were identified as degradation products that still exhibited antioxidant activity (Wei et al., 1999). IV. Heat induced oxidation in food processing and adverse effects in over-heated oil Many food-processing protocols require thermal processing to produce desirable flavor and structure as well as safety of the final food products. One typical example is the oil frying of potato chips. One factor that contributes significantly to the quality of french fried potato chips is the frying medium. Vegetable oil, having a higher content of PUFAs, form a higher ratio of oxidized products during frying compared with oil having a higher oleic and saturated fatty acid content (Perez-Camino et al., 1991; Marquez-Ruiz and Dobarganes, 1996). Hydrogenated vegetable oil may also be used, which is less susceptible to oxidation, but results in a lower surface oil and lower mouthfeel quality product. A lower melting point oil generally impacts a better mouth-feel, since the melt occurs faster at body temperature (Stevenson et al., 1984). However, partially hydrogenated canola and soybean oils have been successfully used (Hawrysh et al., 1995; Melton et al., 1993). The oxidative stability of soybean oil stored at ambient temperature was found to be improved by blending with palmolein, which is the low melting liquid fraction of palm oil used as frying and cooking oil in tropical countries (Berry and Awang, 1983). For potato chips, frying temperatures (170°, 180° and 190°C) did not affect the oxidative rate during storage. Olive kernel oil and palmolein that were absorbed into the chips showed a better oxidative stability, whereas soybean oil presented a higher oxidation rate (Lolos et al., 1999). This result could be explained by the different fatty acid characteristics of oils. For example, 58% of the total fatty acid in soybean oil is linoleic and linolenic acids; the latter contributing to a greater rate of peroxide value increase. The primary fatty acids in both olive kernel oil and palmolein was the presence of oleic and palmitic acids, with both exhibiting a comparatively lower susceptibility to oxidation than linoleic and linolenic acids. To prolong the shelf life of fatty acid species during food processing and to conserve the nutritional value of food products, the inhibition of co-3 PUFA oxidation is critical (Hale and Brown, 1983; Caroll and Braden, 1986). In a baking test, flaxseed was baked along with other ingredients and headspace oxygen was noted to decrease 16 dramatically when the ground flaxseed was heated in sealed tubes. However, under typical baking conditions, there is a minimal loss of alpha-linolenic acid from flaxseed, although the manner of incorporation of flaxseed into food products should be considered to minimize oxidation of alpha-linolenic acids (Chen, et al., 1994). Oxidative changes in heated oils are accelerated, compared to similar reactions occurring at ambient temperature, from the standpoint that oxidation doubles for every 15°C increase in temperature (Kitts, 1996). Those changes, characteristic of the oil composition, result in a reduced number of usages for cooking oils. The type of fats (saturated vs. unsaturated) is a factor which modulates the composition of reactive products in over-heated fats; obviously the more unsaturated fatty acid content of the oil, the more easily it is oxidized. The geometry of double bonds on the fatty acid chain is another important factor to consider (Kitts, 1996). In addition to fatty acid oxidation, heat-treated fats can lead to the formation of cholesterol oxidation products (COP). For example, Park and Addis (1986a) found that the formation of 7-ketocholestrol in heated oil, and the formation of COPs was linearly related to the heating time (Park and Addis, 1986b). The first attempt to define a deteriorated frying fat was made by the German Society for fat research in 1973. These workers stated that a frying fat is deteriorated if its odor and taste are unacceptable; or if in case of doubtful sensory assessment, the concentration of petroleum ether-insoluble oxidized fatty acid content is 0.7% or higher, and the smoke point is lower than 170°C; or if the concentration of petroleum ether insoluble oxidized fatty acids is 1.0% or higher (Shahidi and Wanasundarea, 1997). During deep-fat frying, peroxide values go up to a maximum level and then decrease with 17 time, while total unsaturation decreases with increasing of frying time. Moreover, there will be more signs of free fatty acid oxidation including foaming, changes in color and viscosity of both polar and polymeric materials (Warner, 1997). Good quality edible fats and oils are bland, odorless, free of impurities and oxidatively stable. Oxidized fats and oils on the other hand develop objectionable tastes and odors during processing (e.g. frying and other heat processing in air). The result of this difference not only describes the problem of consumer acceptance of fried food, but also leads to the issue of potential serious biological effects that could originate from human consumption of the oxidized lipid (e.g. chemical food safety). Deteriorated frying oils are poorly digested and absorbed. About one-half of orally administrated linoleate-[14C] autoxidation products, recovered in the feces of rats has been attributed to the cyclic dimers and polymeric materials (Kanazawa et al., 1985). Although this result was based on a simulated model, there is evidence to show that badly abused frying oils can depress appetite and growth and will lead to depression, diarrhea, enlargement of tissues and interference with reproduction (Poling et al., 1970). Whether lipid hydroperoxides are absorbed and deposited in the adipose tissue or other organs of experiment subjects is uncertain (Chow, 1992), since peroxidized lipids are either poorly absorbed, or are metabolized in the intestinal mucosa (Glavinda and Sylven, 1970). In mammalian tissues, it is suspected that glutathione peroxidase, an antioxidant enzyme, which plays a role in reduction of lipid peroxides (Reddy and Tappel, 1974; Vilas et al., 1976), is important in reducing this risk. Such work in vivo; however, should be done with phospholipases (Chow, 1988), since glutathione peroxidase cannot reduce membrane lipid hydroperoxides (Grossman and Wendel, 1983). 18 V. Fatty acids and related health benefits or risks of flaxseed oil, butter and fish oil Essential dietary fatty acids have been known since the late 1920s (cited in Chapkin 1992), when Burr and Burr found that linoleic acid and possibly other fatty acids were essential to survival. Numerous researches have followed these initial observations. Two fatty acids, linoleic acid (18:2o>6) and linolenic acid (18:3co3), were recognized to be essential for numerous physiological functions in humans. In both plant and animal tissues palmitic acid and acyl modification systems are critical components for the regulation of cellular acyl composition and the desired chain length of different fatty acids. However, the initial substance required for biosynthesis of physiologically important fatty acids (for example, 20:4co6, 20:5co3 and 22:6to3) are 18:2co6 and 18:2co3, which are derived only from the diet. The processing of longer and more unsaturated fatty acids involves both desaturase and elongase enzymatic activity. The biological transformation of linolenic acid (18:3n3), for example, is illustrated in Figure 2. Therefore, a stable and sufficient quantity of dietary essential fatty acids is important to maintain a good health. For example, the 20:4OJ6, 20:5co3, 22:6co3 are similarly involved in the homoviscous control of membrane lipid bilayers of most cells (Bernsohn and Spitz, 1974; Mead, 1984). Various symptoms of 18:2to6 deficiency, include depressed growth, scaly dermatitis, increased permeability of skin, fatty liver, kidney damage, and impaired reproduction (Chapkin, 1992). The recommendation of dietary polyunsaturated fatty acid intake for a healthy adult is: 18:2w6-14g/day (4.8% of calories), 18:3co3-3g/day (1.0% of calories), 20:5w3 plus 22:6co3-0.8g/day (0.27% of calories). This is a total PUFA 19 C18:3eo3 slA-6-desaturase C18:4o)3 i-elongase C20:4co3 stA-5-desaturase 20:5co3 -lelongase 22:5co3 vlelongase 24:5oo3 slA-4-desaturase 24:60)3 IP-oxidation 22:60)3 Figure 2. Biological transformation of linoleinic acid (18:30)3) (Bernsohn et al., 1974). requirement of 18g/day (6-7% of calories), or an co3/co6 ratio of between 1:4 (Simopoulos 1989). Flaxseed oil, which is unique in its high linolenic acid (C18:3co-3) content, is an oil that is used in paint, varnishes and linoleum (Bhatty and Rowland, 1990). The only cultivated flaxseed (Linum usitatissimum) contains 40-63% linolenic acid or even higher (White, 1992). Generally, flaxseed is grown in northern, cooler climate in Canada, which produces a greater extent of unsaturated flaxseed oil compared to cultivate grown in southern warmer climates (McGregon and Carsen, 1961). Recent evidence on the nutritional value of flaxseed oil is derived from the interest in co-3 fatty acids. However, oxidative instability of oils with a high C18:3co3 content, results in a major limitation for the food industry (White, 1992). This is particularly true in cases where the oil kernel be exposed to heat. Milk fat, which is a readily available lipid is an extremely complex fat with respect to both lipid classes and component fatty acids. For example, 400 or more fatty acids have been identified from butter (Jensen et al., 1990, 1991). Typical North American butter contains about 10.4% (wt%) myristic acid (C14:0), 27.7% (wt%), palmitic acid (C16:0), 10.0% (wt%) stearic acid (C18:0) and 22.2% (wt%) linoleic acid (C18:2) (Smith, 1978). Ruminal biohydrogenation results in a low proportion of PUFAs, even though dairy cattle consume relatively large amounts of PUFAs from forage (Jensen, 1992). The concern about saturated fat, such as consumption in butter and tallow, also involves the issue of cholesterol content. High intake of saturated fatty acids (C14:0 and C16:0) has for example been confirmed to increase LDL cholesterol, by 21 reducing the activity of the LDL receptor (Grundy and Denke, 1990). Cholesterol will also produce a similar reduction in LDL receptor activity. Fish oils and other PUFAs have long been associated with a decreased risk of hypertriglyceridemia (Pillipson et al, 1985). The emphasis on co-3 fatty acids has led to the commercial availability of purified fish oil supplements in most health food and vitamin stores in North America. The primary benefit of fish oil comes from the unique, high concentration of long chain co3 fatty acids, such as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). Docosahexaenoic acid (DHA) is essential for the growth and functional development of the brain in infants, and it is also required for the maintenance of normal brain function in adults (Horrock and Yeo, 1999). DHA is also the principle active component in fish oil for cardiovascular protection (McLennan et al., 1996). Alpha-linolenic acid, which is high in flaxseed oil, is not effectively converted to DHA after elongation, desaturation, transportation to and from peroxisomes (where it is then shortened from 24:6 to 22:6) (Luthria et al, 1996). Recommendations to consume high levels of co3 fatty acid fish oil and flaxseed oil are based on several lines of epidemiological evidence, which show that co-3 fatty acid consumption (5500mg/month) was associated with a 50% reduction in the risk of primary cardiac arrest (Siscovick et al, 1995). Moreover, consumption of two or three portions of fish consumed per week could reduce mortality in men who have recovered from myocardial infarction (Burr et al, 1989). The World Health Organization and British Nutrition Foundation recommends ratio of 1:3 to 1:4 for co-3 to co-6 fatty acids (FAO/WHO, 1993; British Nutrition Foundation, 1992). 22 Most scientists distinguish the C8, CIO and C12 fatty acid from the common C16 and C18 fatty acid by calling the former " short chain" and the latter " long chain" fatty acids. The former are readily digested, and on absorption pass directly to the liver via the portal vein, whereas the latter tend to be transferred to the lymph in the form of chylomicrons (Nelson and Ackman, 1988; Welch and Borlakoglu, 1992). For most foods, C22 fatty acids represent the longest fatty acid chain length present in any quantity. Another group of fatty acids of very short chain lengths (C1-C6) are historically called, the volatile fatty acids, primarily because that they can be.steam-distilled at atmospheric pressure. These fatty acids are water-soluble, whereas fatty acids of C8 and longer chains become progressively less water-soluble. Myristic acid (14:0) and lauric acid (12:0) are potent to raise plasma total/LDL cholesterol saturated fatty acids (Zock et al., 1994; Temme et al., 1996). Stearic acid (C18:0) is neutral in cholesterol raising potential, whereas, linoleic acid (18:2co6) has a small LDL cholesterol lowering effect (Mattson and Grundy, 1985). A reduction in the intake of the cholesterol -raising saturated fatty acids and trans fatty acids is proposed to be more important for optimizing the plasma lipoprotein profile, than a reduction in total fat intake per se (Mensink et al., 1997). VI. Cholesterol and autoxidation of cholesterol Cholesterol in the North American diet has been identified as a contributing factor to the high death rate resulting from coronary heart disease (Dixon and Ernst, 2001). Cholesterol is predominant in animal lipids, such as lard (0.37-0.42%), beef fat (0.08-0.14%), mutton tallow (0.23-0.31%) and butter fat (0.2-0.4%). Hen eggs contain about 300mg of cholesterol per egg, which corresponds to approximately 5% of total lipids. A 23 typical UK diet contains 300-400mg cholesterol, which is found mainly from egg (~150mg), meat (-lOOmg), milk (~40mg) and butter spread (~30mg). Several reports have identified oxysterols, for example, 25-hydroxycholesterol from animal products, which has also been associated with atherosclerosis (Clare et al., 1995; Yuan et al., 1998). This particular steroid (25-hydroxycholesterol) has also been found linked to defects of the aortic surface when absorbed by mammals (Peng et al., 1982). The reason for oxidized cholesterol leading to atherosclerosis has also been explained by the oxidation of low density lipoprotein. Low density lipoprotein is the primary cholesterol carrier in the circulation and thus, oxidation of LDL can alter the pathway of macrophage regulation and generate foam cells, which in turn have receptors for the modified LDL. From this stage onward, fatty streaks, fibrous plaques and complication lessons, all of which are typical indices of atherosclerosis, have been reported to progress (Steinberg et al., 1989). The earliest evidence of cholesterol oxidation in vivo goes back to the 1940s (Bergstrom and Wintersteiner, 1941). Other adverse effects of cholesterol oxidation products are of cytotoxicity and inhibition of cholesterol synthesis enzyme (FfMG-CoA) activity (Smith, 1981b; Kitts, 1996; Paniangvait et al., 1996). The initiation factors of cholesterol oxidation are similar to the autoxidation of fatty acid (e.g. heating, oxidative radicals, transitional metal ions, radiation, etc.). This results in the formation of cholesterol oxidation products (COP) (Smith, 1981a). Commonly found cholesterol oxides include cholest-5-en-3p-ol, cholest-5-en-3p\ 7a-diol, cholest-5-en-3(3,7P-diol, 3(3-hydroxycholest-5-en-7-one, 5a,6a-epoxy-5-cholestan-3R-ol, 5p\6|3-epoxy-5-cholestan-3R-ol, cholest-5-en-3p\19-diol, choleset-5-24 en-3p\25-diol, 5a-cholestane-3p\5,6P-triol, etc. (Paniangvait et al., 1995). A series of COPs have been identified and are shown in Figure 3. Though dairy products are a potential source of COPs (Luby et al., 1986), fresh fluid milk and butter do not contain detectable levels of COPs (Kitts, 1996). Heating lard at 180°C for 150 h leads to about 70% loss of cholesterol and appearance of various COPs (Chen and Yen, 1994). The existence of a 5A-double bond makes the cholesterol prone to oxidation (Smith, 1987). The formation of different COPs has also been found during different stages of thermal oxidation of cholesterol. For example, cholestan-3p\5a-6-triol can be detected in lard after heat treatment for 20 minutes. This is in contrast to concentrations of 5,6oc-epoxycholesterol and 7P4iydroxycholesterol, which steadily increase during the first 100 minutes of heating. Though heating is an important factor influencing the oxidation of cholesterol, cholesterol is relatively stable if heated alone. However, the presence of other fats, particularly the unsaturated fats, provides a significant contribution to the promotion of cholesterol oxidation during heating (Osada et ah, 1993). This suggests that free radical generation from PUFA may contribute to the acceleration of radical centred cholesterol oxidation chain reactions (Korahani et al., 1982). A number of measures can be used to avoid and minimize the cholesterol oxidation in food processing. These include the feeding of antioxidants to live animals, or using antioxidants externally in food processing (Park and Addis, 1986b) In addition, minimum temperature exposure during storage (Luby et al., 1986) and exclusion of oxygen with packaging (Chan et ah, 1993) are effective ways to minimize oxidation reactions. Processed fish products also contain cholesterol oxidation products (COPs). 7a-hydroxycholestol O H cholestane-3P,5a,6P-triol 7-ketocholesterol H O ^ ^ ^ ' O H 7(3-hydroxycholesterol Figure 3. Structure of cholesterol and its oxidative products (Paniangvait et al., 1995). 26 Only one side-chain COP (e.g. 25-hydroxycholesterol) has been found in oxidized fish products (Oshima et al, 1993). VII. Response surface methodology (RSM) RSM was formally developed by G. Box and K. Wilson and other colleagues at Imperial Chemical Industries in England in 1952. The RSM technique uses simple polynomial models to develop a statistically precise predictive knowledge about a process, and it can find the answers efficiently and accurately (http://www.echip.com/response.htm). RSM uses a least-squares curve-fit regression analysis to calculate a system model, test its validity and analyze the model. RSM uses quantitative data from appropriate experiments to determine and simultaneously solve multivariate equations. Haryati et al. (1999) reported optimization of chemical transesterification of palm oil using RSM. Shieh et al. (1995) used RSM for optimizing enzymatic transesterification of triolein with capric acid, whereas Huang and Akoh (1996) used the program for the optimization of enzymatic transesterification of vegetable oils with ethyl caprylate. Application of this method has also been reported by Cho et al. (1993) for the formulation of partially interesterified canola/palm blends. Irwandi and co-workers (2000) also used the program to optimize a combination natural antioxidant combination for deep fat frying of palm oil. Thus, RSM has been used successfully for fat and oil research. 27 Chapter 1. Blending of Fish Oil with Butter Oil: Effect of blending on fatty acid profiles and thermal oxidative stability. Introduction Oil blending, the action of mixing two or more sources of oils together to form an oil mixture or blend, was considered in this thesis as one approach to produce an optimal fatty acid composition for an ultimate P/S or co3/co6 ratios, while minimizing the potential for lipid oxidation of a cooking oil. Previous studies by Berry and Awang (1983) on the physico-chemical characteristics of palm olein and soybean oil blends have reported this approach. Chemical analyses such as peroxide value and anisidine value were used in the former study to assess the apparent improvement in oxidative stability. In addition, Kim et al. (1988) also investigated the oxidative stabilities of oil mixtures made from soybean oil and coconut oil, or soybean oil and palm oil. Kupranycz et al. (1986) studied butter and butter oil blending to reveal the nutritional advantages and thermal stability contributed by both fats. These works suggested that saturated fatty acids from butter oil could possibly enhance the oxidative stability of oil blends. Butter oil has a relatively high content of saturated fatty acid, with short and medium chain fatty acids. Fish oil on the other hand is relatively rich in polyunsaturated fatty acids (PUFAs), especially the long chain co-3 fatty acids, eicosapentaenioc acid (EPA, C20:5co3) and docosahexaenoic acid (DHA, C22:6co3). The blending of butter oil to alter the fatty acid profile of fish oil may enhance the medium chain fatty acid content and improve the oxidative stability under typical thermal processing conditions. 28 The purpose of this experiment was to investigate the influence of blending butter oil with fish oil in minimizing the lipid oxidation and changes in PUFA profile. 29 Materials and Methods Oil sources Unsalted butter (4 blocks, 454 g/block) was obtained from Shady Lane, Lucerne Food Ltd. (Edmonton, Alberta) via Safeway (Vancouver, B. C). Butter oil was made from butter by melting butter in a warm water bath (40 °C) and filtering the melted mixture through cheesecloth to separate the butter oil from other non-lipid constituents. The filtrate was dried over anhydrous sodium sulfate. Butter oil was then collectively pooled, re-filtered through Whatman 41 filter paper under vacuum, flushed with nitrogen (Kupranycz, et al., 1986) and stored in darkness at -35 °C until the experiments started. Mixed salmon and herring oil (antioxidant free*) was a generous gift from EWOS (Surrey, B. C). Butter and fish oils were stored in the dark in a cold room (-35 °C) in suitable portions to avoid the freeze-thaw recycle during the course of the study. Block heater A DIGI-BLOCK digital block heater (Laboratory Devices Inc., USA) was used to provide a safe, compact and dry heat. The DIGI-BLOCK precision temperature probe was fitted directly into individual blocks, next to the samples, to provide the most accurate temperature readings and control possible. The temperature range of the block heater was ambient to 200 °C by 0.1 °C, while the maximum temperature ripple was ±0.1 °C. Blocks for 11 x 130 mm Pyrex culture tubes (Corning Incorporated, Corning, NY) were chosen for all experiments. * antioxidant free refers to no added a-tocopherol following refinery. 30 Other chemicals Fatty acid methylates, Tween-20 (polyoxyethylene sorbitan monolaurate), thiobarbituric acid, potassium iodide, sodium thiosulfate and starch were purchased from Sigma-Aldrich Chem. Co. (St. Louis, MO). Potassium hydroxide, sodium sulfate, sodium bicarbonate, H2SO4, methanol, acetic acid, chloroform, trichloroacetic acid and hexane were obtained from Fisher Scientific (Fair Lawn, NT). Experimental design The experimental design for selecting the different oil blends was determined using the Echip 6.04 (Echip, Inc., Hockessin, DE) program (Wheeler, 1993). This program also provided calculation equations and statistical analysis (Irwandi, et al., 1999). h Constrained mixture design Oil blends were studied using a constrained mixture design for two variables with different butter oil and fish oil combinations (Table 1). In this blending experiment, the sum of variables in the mixture design always totaled 100 % (or 1). The low limit of each component was set at 0, and the high limit was set at 1 or 100 %, thus: S l F 1.0 Furthermore, a quadratic response was regressed to predict the polyunsaturated fatty acid ratio, or co-3 to co-6 fatty acid ratio, as follows: Y = Ro +p \*XFish+p2 # XButter + Pl2 • X f i s h * Xfiut ter + CCi • X F i s h . 2 +V-2 ^ B u t t e r , in which, 31 Y = polyunsaturated to saturated fatty acid ratio, or, u)-3 to co-6 fatty acid ratio; Xfish + Xfiutter = 1 i X p i s h = composition of fish oil in percentage; X B m t e r = composition of butter oil in percentage; Po intercept; Pi = coefficient for fish oil at the first order; P2 = coefficient for butter oil at the first order; P12 = coefficient for the interaction between oils; a i = coefficient for fish oil at the second order; a 2 = coefficient for butter oil at the second order. 2. Formulation of oil blends with fish and butter oils Oil blends were formulated according to the computerized design derived from the ECHIP 6.04 program. A total of 13 oil blend samples, each representing different blending ratios, were tested (Table 1). Each oil blend sample contained 3 g of total oil. Therefore, 1.00±0.02 g fish oil and 2.00±0.02 g butter oil were sampled respectively, before they were vortexed for 1 minute to make a homogenous oil blend. Blended oils were weighed into 13x100mm Pyrex culture tubes (Corning Incorporated, Corning, NY). The oil blends consisting of 100% fish oil were called blend A, and blend E was made of 100% butter oil. The other oil blends B, C, and D were made of 67% fish oil with 33% butter oil, 50% fish oil with 50% butter, and 33% fish oil with 67% butter oil, respectively. 3. Heating studies Every butter-fish oil blend was used in a different thermal heating trials. Each sample was duplicated in the study. Oil blends made from butter oil and fish oil were 32 incubated at two temperatures, (150 °C and 180 °C) for three duration periods (8 h, 16 h and 24 h), respectively. Individual tubes, containing each oil samples, were inserted into the block heater. The level of oils was set to be lower than the depth of the hole in the block heater in order to get a uniform heat treatment for each sample. A marble was used to cover each tube for the purpose of condensing liquid evaporation and reducing oxygen exchange. After heating, all samples were taken out of the blocks and the duplicates were mixed together. They were left at room temperature for chemical analyses; including thiobarbituric acid reactive substances (TBARS), conjugated diene hydroperoxides (CDH) and peroxide value (PV), as well as fatty acid profile determination. The remainder of butter-fish oil blends were stored at -35 °C for further cholesterol content determination when applicable. Analysis of oxidation of oil blends All the analyses were conducted in duplicate. L Thiobarbituric acid reactive substances (TBARS) One hundred milligram of heated oil was transferred to a tissue culture tube, where a 50mg • ml"1 emulsion was made using lmL of lOmM phosphate buffer (pH 7.4) with 5 % Tween-20 and 1ml of absolute ethanol. Aliquots (100p,L) of this emulsion were then mixed with 900uL of lOmM phosphate buffer (pH7.4), followed by the addition of lmL of freshly prepared 0.8 % of thiobarbituric acid (in 25mM NaOH with 0.02 % of BHT) and lmL of 25 % (w/v) trichloroacetic acid. The reaction mixture was incubated in 33 a boiling water bath for 15 minutes, followed by cooling to room temperature with running tap water. A pre-wetted C18 Sep-Pak cartridge (Waters Corp., Milford, MA) was used to separate the lipophilic phase prior to taking the absorbance reading at 532nm (Shimadzu, Scientific Instruments Inc., Columbia, MD spectrophotometer 1600), against a blank tube where oil was replaced by phosphate buffer. The reading was converted to a malonaldehyde (MDA) content according to a standard curve obtained from the absorbance at 532nm vs. the concentration of 1,1,3,3-tetra propane. 2. Conjugated diene hydroperoxides (CDH) Heated oil (100 mg) was transferred to tissue culture tubes to make a 50 mg • ml"1 emulsion with 1:1 of 10 mM phosphate buffer (pH 7.4, containing 5 % Tween-20) and absolute ethanol. Each emulsion (100 uE) was then mixed with 900 pL of lOmM phosphate buffer (pH 7,4) and 1 mL of ethanol. A further dilution (20 times) with ethanol was necessary to obtain absorbance readings on scale at 234nm (Shimadzu, Scientific Instruments Inc., Columbia, MD, spectrophotometer 1600). Each reading was taken against a blank sample in which the oil sample was replaced by the phosphate buffer. The concentration of CDH was then obtained using an extinction coefficient for 234nm according to the following equation: CDH (Limol/g oil) = 106 x (A-A0) / 2.5 e. : where A = absorbance at 234 nm for oil sample, Ao = absorbance at 234 nm for the blank without oil, and e = 29500M''cm"' (extinction coefficient). 34 3. Peroxide value (PV) This method was adopted from the AOAC Official Method 965.33 (AOAC, 1980). Heated oil (1.5 g) was transferred into a 250 mL Erlenmeyer flask and dissolved in a 30mL of acetic acid-chloroform solution (3:2, v/v). After 0.5mL of saturated KI solution was added, with occasional shaking for 1 minute, 30 mL of distilled deionized water was added to the reaction. Samples were titrated with 0.01 N Na2S203 under vigorous shaking, until a light yellow color was obtained. A 1 % starch solution (0.5 mL) was then pipetted into each sample and titrated until the blue color disappeared. Peroxide values (milliequivalent peroxide per kilogram sample) were calculated according to the following equation: Peroxide value (milliequivalent peroxide/kg sample) = S x N x 1000 g"1 sample. : where 5 = mL Na2S203 (blank corrected), N - normality of Na2S203 solution, g = weight of oil (g). Sodium thiosulfate solution was made by dissolving Na2S203«5H20 in distilled deionized water and boiling the sample for about 5 minutes. The Na2S203 reagent was titrated with K 2 Cr 2 0 7 (freshly dried at 100 °C for 2 h). A stock solution of 0.1N Na 2S 20 3 was prepared and each concentration was calculated as follows: Normality = g X 1 0 0 0 x 49.032 . mL : where g represented the gram of K^C^O?, and mL represented volume of Na2S203 used in the titration. A 0.01 N fresh Na 2S 203 solution was diluted from the stock solution just prior to titration. 35 4. Fatty acid quantification The fatty acid profiles of all experimental oils and oil blends were quantitatively determined using methyl derivatized fatty acids and gas chromatography (GC). a. Fatty acid calibration curve Fatty acid standards and internal standard (IS) were ordered from Sigma, Inc. A fatty acid methyl ester standard mixture included methyl esters of lauric acid (C12:0), myristic acid (C14:0), palmitic acid (C16:0), palmitoleic acid (C16:lu)9), heptadecanoic acid (C17:0), stearic acid (C18:0), oleic acid (C18:lco9), linoleic acid (C18:2co6), linolenic acid (C18:3u)3), arachidonic acid (C20:4co6), eicosapentaenoic acid (C20:5co3), and docosahexaenoic acid (C22:6co3). Prior to the beginning of each fatty acid quantification, response linearity of gas chromatography (GC) determinations for each fatty acid methyl ester was confirmed by constructing a calibration curve for the response to mixtures of varying amounts of fatty acid methyl ester with a constant amount of internal standard (IS: heptadecanoic acid methyl ester). Mixtures of standards were prepared so that the concentrations of individual fatty acid methyl ester standards ranged from O.OlOpg pL"1 to 1.0 pg pL"1, with the ratio of fatty acid methyl ester standard / IS ranging from 0.1:1 to 10:1. Response linearity was analyzed by plotting the area response ratio of each fatty acid methyl ester to the IS, versus the weight ratio of the fatty acid methyl ester to the IS (Appendix A). b_. Sample preparation 36 Heated oil (100 mg) was weighed into a 20 mL screw-cap glass test tube, containing 5 mg of IS (heptadecanoic acid methyl ester). A 0.5 N potassium hydroxide aliquot (5 mL in methanol) was added to each test tube, and then each tube was vortexed and placed in the water bath at 40°C in order to dissolve oil into a KOH-CH3OH solution completely. Samples were incubated at 40°C for 4 h to accomplish saponification. Petroleum ether was used to extract the unsaponifized fractions. The petroleum ether layer was discarded and the saponifiables were kept in test tubes for methyl esterification. To methylate the fatty acids in the saponifiable fraction, 3 mL of 6 % H2SO4 in methanol (v/v) was pipetted into the test tube. The tube was sealed by a tight teflon-lined cap, vortexed and placed in a water bath set at 80 °C for 12 h (over night) (Ferrier et al., 1995). Upon completion, the tubes were removed from the water bath and cooled to room temperature. Hexane (2 mL) (Fisher Scientific, HPLC grade) was then added to extract the non-polar fatty acid methyl esters from the other constituents. Sample were separated into two layers, and the top hexane layer was collected and moved to a eppendorf microcentrifuge tube (1.5 mL), containing anhydrous sodium sulphate and sodium bicarbonate (ratio 4:1, w/w) to remove any residual water. The final solution was transferred into a 2.0 mL GC vial (12 x 32 mm, Alltech Associates, Inc., Deerfield, IL) and sealed with Teflon-lined septum. At this stage, the sample was ready for GC analysis. c. Gas chromatography Quantification of fatty acid methyl esters was determined using a GC-17A gas chromatography (Shimadzu, Scientific Instruments Inc., Columbia, MD), equipped with a flame ionization detector. A volume of 1 J IL sample was injected into a silicone fused 37 capillary column (Omegawax 320; 30 m x 0.32 mm i. d., 0.25 pm film thickness; Supelco Inc., Bellefomte, PA) using an autoinjector (AOC 1400, Shimadzu, Scientific Instruments Inc., Columbia, MD). Helium (He) was used as the carrier gas. The injection temperature was 200 °C, and the temperature at the detector was 220°C. The column temperature was programmed from 60 to 220°C at a rate of 10°C per minute, for 40 minutes to a final temperature 220°C, which was held for 10 minutes. The column flow rate was set at 1.9mL • min"1. The retention time of peaks was used to identify individual fatty acids in the sample, against those fatty acid methyl esters present in the standard. Chromatograms were stored and analyzed later using a Class-VP chromatography data system version 4.2 (Shimadzu, Scientific Instruments Inc., Columbia, MD). This program integrated the area under each peak. A GC chromatogram of the fatty acid methyl ester standards is shown in Appendix B. <± Fatty acid profile of native oils The fatty acid profiles of native fish oil and butter oil were obtained as shown in Table 2. This was performed prior to making the oil blends. Statistical analysis of data Experimental design software, Echip 6.04, was used to perform various statistical analyses of the data collected from each experiment. Also a quadratic regression was performed using the Echip software. SPSS (SPSS for Windows, version 7.5, SPSS Inc.) was used to determine the difference between paired treatments using student t-test with a significant level set at 38 0.05. ANOVA was used to analyze the effect of all various reaction conditions and interactions. Results Fatty acid profiles of blended oils with fish oil and butter oil Fish oil and butter oil, were chosen as the native oils due to their uniquely different fatty acid profiles. The fatty acid composition of fish oil and butter oil are listed in Table 2. EPA and DHA were the primary co-3 long-chain PUFAs in fish oil. There was more oleic acid present in fish oil, in addition to a relatively smaller amount of linoleic acid and linolenic acid than in butter oil. Myristic acid, palmitic acid and stearic acid were the major saturated fatty acids present in fish oil. Butter oil contained relatively higher amounts of saturated fatty acids, such as lauric acid (C12:0), myristic acid (C14:0), palmitic acid (C16:0) and stearic acid (C18:0) and monounsaturated fatty acid (e.g. oleic acid, C18:lco9). Trace amount of PUFA (<1%) was obtained in the butter oil. The ratio of polyunsaturated to saturated fatty acid (P/S ratio) of fish oil was calculated to be 0.508. Fish oil also had a high co3/co6 ratio of 7.403, which was attributed mainly by the presence of DHA and EPA. Thus fish oil is defined as a unique dietary source of long chain co-3 PUFAs. The dominance of myristic and palmitic acids in butter oil, and very low PUFA content resulted in a P/S ratio of 0.028 and a co3/co6 ratio of 0.202. Therefore, the fish oil used in this experiment was considered to be the primary source of co-3 fatty acids, as well as a source of co-6 PUFA and saturated fatty acids. The butter oil was chosen for its saturated fatty acid content. 40 With this initial information, a set of experiments with butter oil blended to fish oil was chosen to evaluate the effect of reducing susceptibility to heat-induced oxidation and depletion of PUFAs. Five blending ratios set between fish oil and butter oil were tested according to a computerized experimental design (ECFflP 6). The calculated fatty acid profiles for all butter-fish oil blends are shown in Table 3. Assessment of a quadratic model used to predict the optimal combinations of fish and butter oil blends, using P/S and to3/oo6 ratios The ratios of polyunsaturated to saturated fatty acids (P/S) and co-3 to to-6 fatty acids (co3/u)6) were two parameters used to assess the effect of blending on the change of fatty acid profile for each oil mixture. The coefficients of the quadratic regression model used to determine the optimal proportions of fish oil added to butter oil, as assessed by both the P/S and co3/oo6 ratios, following heat treatments at 150°C and 180°C, are presented in Tables 4 and 5, respectively. In recent years, a constrained mixture design has been used as an experimental technique for food formulations in the food industry (Wheeler, 1993). The quadratic model chosen in this study was successful (R2>0.979, p<0.001) at describing the changes in the P/S ratio (Table 4) for the different oil blend mixtures, following heat treatments for up to 24h, at both 150°C and 180°C, respectively. A similar result (R2>0.970, p<0.001) was obtained with the regression coefficients for the co3/u)6 ratio derived from different oil blends and time-temperature treatments (Table 5). Tables 4 and 5 confirm that the fish oil was an obvious contributor to a significant (P<0.001), positive, first order change in both P/S and co3/u)6 ratios in the butter-fish oil blend, while butter oil produced 41 a significant (P<0.001), negative effect on these same ratios. A significant (negative) interaction (P<0.05) was also observed for both P/S and co3/co6 ratios when butter oil was blended with fish oil, thus confirming that the loss in co-3 PUFAs of each blend, resulting in altered P/S and co3/co6 ratios, was attributed primarily to the addition of butter oil to the blend. However, differences in regression coefficient, Pi and p 2, for P/S and co3/co6 ratios of blended oils specified a greater dependence on duration of heating as well as the high temperature used for heating experiments. For example, a lower Pi coefficient for fish oil were obtained from the 150°C treatment, compared to the 180°C heat treatment, thus denoting that greater changes in the P/S and co3/co6 ratios were likely to occur at the higher temperature of 180°C. On the other hand, a greater coefficient p 2 for the butter oil was also derived from the 180°C heat treatment. Examining the different P/S and co3/co6 ratios listed in Tables 4 and 5, it can be seen that higher Pi coefficients were obtained from samples treated by the lower temperature and shorter heating time. The same result was obtained for butter oil, which produced higher p 2 coefficients for treatments that had lower temperature and shorter incubation time. This result signifies the fact that PUFAs, derived mainly from the fish oil, were relatively more labile at the higher temperature of processing. At this stage, it can not be determined if the presence of butter oil in the blend had a positive effect on P/S or co3/co6 ratio, because the generation of a higher p 2 may have occurred due to both the greater stability of butter oil during heating and the reduced proportion of fish oil in the blend, 42 The greater (31 value found from different samples tested was derived from a lower temperature and a shorter heating time. The opposite trend was found for butter oil, which produced the smaller (32 value in specific treatments that used lower temperatures and shorter incubation time. From Appendix C it can be seen that the quadratic regression model used in this experiment could not optimize the P/S and co3/u)6 ratios from different oil blends. This occurred because the contribution of fish oil in the blend produced a positive slope for both P/S and co3/co6 ratios regardless of the relative proportion of fish oil in the blend. Therefore instead of finding an optimum for the mixture of butter oil to fish oil, the data presented here only enables the user to predict the changes in co3/co6 or P/S ratio from the amount of fish oil added to the blend. Depletion of PUFAs during thermal incubation The heated oil blends were sampled every 8 h during the thermal incubation at both 150°C and 180°C, in separate experiments. Fatty acid data from gas chromatography analysis was quantitated and the results of the fatty acid profiles of butter-fish oil blends following thermal treatments at both temperatures, are shown in Tables 6 and 7, respectively. Following a thermal treatment at 150°C for up to 24 h, it was observed from the fatty acid composition (% of total fatty acids), that the depletion of DHA in the butter-fish oil blends occurred dramatically during the first 8 h of thermal incubation at 150°C (Table 6). By comparing the relative percentage of total fatty acids, it was also estimated 43 that DHA was destroyed to a greater extent than EPA during the same time period. In both cases, DHA (e.g. from 7.14 to 6.12% in blend A) and EPA (e.g. from 7.80 to 6.57% in blend A) were lost at a similar rate after the first 8 h in this experiment. The P/S ratio, also decreased.to a greater extent during the first 8 h, and thereafter showed only a small change. There was also a slight decrease in the observed co3/co6 ratio during the course of thermal treatment at 150°C. Relatively greater changes in EPA or DHA were obtained with the longer heating duration (e.g. 24 h). Table 7 presents the changes in PUFA composition of butter-fish oil blends during the thermal incubation at 180°C. A similar pattern to that observed with 150°C treatment was the fact that DHA depletion in all butter-fish oil blends also occurred mostly after the first 8 h of thermal incubation at 180°C. At this temperature, losses of DHA were relatively greater than EPA. As a result, the marked depletion of both EPA and DHA for the first 8 h at 180°C, resulted in a much lower P/S ratio and co3/co6 ratio, compared to the 150°C thermal treatment. From the P/S and co3/co6 ratios of the butter-fish oil blends, it could be predicted that greater loss in PUFA quality occurred at 180°C, thus indicating again, that the co-3 fatty acids were the primary sources of oxidation substrates lost during the heating of the butter-fish oil blend. Differences in P/S ratios between different blending combinations were also found, demonstrated that the more saturated the fatty acid composition, the more resistant the oil to the depletion PUFA (Tables 6 and 7). EPA and DHA were the dominant co-3 fatty acids in fish oil, although also present were the shorter chain PUFAs existing in fish oil, such as linolenic acid and linoleic acid. Comparatively, the amount of linolenic acid in fish oil was too low to be considered as an indicator of oxidation susceptibility. However, the percentage of linoleic acid to the total fatty acids (w/w), interestingly showed a slight increase after the first 8 h of thermal treatment at both 150°C and 180°C. Since this trend seems entirely impossible to have occurred, the interpretation of results using percentage of fatty acid is questionable and therefore could have a limited value for predicting changes in all fatty acids (Tables 6 and 7). It can be seen that, from examining the fatty acid profiles, blending fish oil with butter oil resulted in a greater exposure of oxidizable substrates (e.g. co-3 PUFAs) to heat. This was especially true if the specific fatty acids in the blends were more susceptible to heat induced damage. Since the absolute amount of the individual fatty acids in the butter-fish oil blends varied with different blending ratios, the relative proportion of individual fatty acids, when expressed as a percentage of total fatty acids in the butter-fish oil blends, could not be used as a sensitive indicator of fatty acid oxidation. It was therefore concluded that the percentage of a single fatty acid does not describe accurately the magnitude of change for that fatty acid following a thermal treatment. Accordingly, better results were obtained when the absolute amount of fatty acids were examined. As a result, the use of mass changes in fatty acids was felt to be the only true indicator of change, and was therefore used for the remainder of this study. 45 Evaluation of blending and heat processing on the absolute amount of PUFAs of butter-fish oil blends The results of ANOVA testing of experimental data showed that temperature, incubation time and blending ratio variables were all significant factors (P<0.001) in altering EPA and DHA content in butter-fish oil blends. Interactions between blending ratio x temperature (P<0.001), blending ratio x time (P<0.001) and blending ratio x temperature x time (P<0.05), significantly affected both EPA and DHA concentrations. No significant interaction was found for temperature x time. Similarly, temperature, incubation time and blending ratio significantly (P<0.0GT) influenced the retention of linolenic acid and linoleic acid. Significant interactions, such as blending ratio x temperature (P<0.01), temperature x time (P<0.05) and blending ratio x temperature x time (P<0.05) also characterized linolenic acid and linoleic acid residues. However, no significant interaction was obtained for blending ratio x time for these specific PUFAs. In addition, Pearson coefficients (r=-0.817 for EPA, r=-0.827 for DHA, P<0.01) derived from a correlation test for residual EPA and DHA, showed that the blending ratio produced a much stronger effect than the other two factors, namely, temperature (r=-0.230 for EPA, r=-0.212 for DHA, P<0.05) and time (r=-0.349 for EPA, r=-0.352 for DHA, P<0.01). Finally, it was found in this study that changes in EPA and DHA concentrations were temperature-dependent, as well as dependent on the blending ratio and time of heating. From the Pearson coefficient alone, the relative order of the treatment effects on residual EPA and DHA concentrations was: blending ratio > time > temperature. 46 Effect of blending strategy on the depletion of PUFA among butter-fish oil blends during thermal treatments Experimental values for the absolute amount of individual fatty acids (mg/gm oil), were presented with the averages from every butter-fish oil blend for the same blending ratio. On the other hand, the averages of blend A containing 100% fish oil, and the other extreme blend E, containing 100% butter oil, were used as sources to calculate hypothetical values. The hypothetical values, calculated subsequently from the averages of blending A (100% fish oil) and E (100% butter oil), were used to compare experimental values from different blends. For instance, hypothetical values of fatty acid retention in blend B are equal to the sums of 66.7% of the experimental value derived from blend A and 33.3% of experimental values derived from blend E. Similarly, hypothetical values of fatty acid retention in blend C were forced to be equal to the sum of 50% of experimental values from blend A and 50% of experimental values from blend E. Finally, hypothetical values of blend D were equal to the sums of 33.3% of experimental values from blend A and 66.7% of experimental values from blend E. Therefore, both the experimental values and the calculated hypothetical values for both extreme blends A and E are the same. Experimental values showed the subsequent changes in fatty acids in blended oil that occurred through the thermal treatment, whereas hypothetical values were derived to predict the changes occurring in fatty acids as a result of both blending and heating fish and butter oils. The comparison of the experimental and hypothetical values for the changes in DHA and EPA, following thermal incubations up to 24 h at 150°C, are given in Figures 4 47 and 5, respectively. The dominant PUFAs, namely, DHA and EPA, were dramatically lost during the first 8 h of heating. The hypothetical values for oil blends B, C, and D were actually higher than associated experimental values (Figures 4 and 5). This finding indicates that blending of oil did not proportionally reduce the amount of DHA or EPA under high temperature, but in fact, resulted in a greater loss of both DHA and EPA than was predicted. In another words, the blending of fish oil with butter oil predisposed both DHA and EPA to greater losses attributed to heating. Similarly, the results of the 180°C heating experiment indicated that higher temperature heating of oils resulted in more oxidation of both DHA and EPA, compared to results obtained at 150°C during the first 8h thermal treatment (Figures 6 and 7). It is important to note that both DHA and EPA were lost to a greater degree during the next 16 h at 180°C than during a similar time at 150°C. Therefore, higher temperature (180°C) heat treatment produced a greater decrease in residual EPA and DHA than a relatively lower temperature of processing. The hypothetical values for EPA and DHA obtained for 180°C heating were different from corresponding values obtained from 150°C heating, which suggests that the blending of oil (e.g. the dilution of some specific fatty acids), was again a factor in increasing the thermal oxidation of DHA and EPA. It is conceivable that blending of oil might retard shorter chain and more saturated fatty acids from oxidation. For example, there was a trend obtained for stearic acid (CT8:0), compared to DHA and EPA (Figures 8 and 9). The greatest loss of stearic acid occurred during the first 8 h and there was no distinguishable change thereafter. In addition to stearic acid, similar experimental and hypothetical values were seen for oleic, 48 palmitic, palmitoleic and myristic acids, which were also different from absolute DHA and EPA residual values. This finding suggests that blending of oil might benefit the shorter chain and more saturated fatty acids from degradation, relative to the changes seen with long chain PUFAs. Information of lipid oxidation from chemical analyses Table 8 shows the effect of blending of butter oil to fish oil on the formation of lipid hydroperoxides (PV), conjugated diene hydroperoxides (CDH) and thiobarbituric acid reactive substances (TBARS), following thermal treatments at 150°C and 180°C during 24 h heating. Coefficients of determinations (R2) and quadratic regression coefficients (p\, fi2, f3i2, cci and a2) for PV, the content of CDH and MDA are given in Table 9. Coefficients of determinations derived from PV of butter-fish oil blends resulted in smaller R 2 values, probably due to the thermal damage of peroxides formed during incubation at high temperature, which could adversely affect the accuracy of peroxide value measurements. Comparatively, the R 2 derived from TBARS (e.g. MDA) were greater than those from PV and CDH content. The greatest R 2 was noticed for TBARS (e.g. MDA) from the butter-fish oil blends treated at the higher temperature (180°C) for the longest time (24 h). Coefficients derived for PVs of oils treated for short heating periods (8 h), showed that fish oil and butter oil, as well as an interaction between the two, could be important factors affecting PV (Table 9). This trend was lost after longer (16 or 24 h) heating times, however. For both CDH or TBARS (e.g. MDA) parameters, the presence of fish oil and butter oil both exhibited a significant (P<0.05) effect according to first order reactions. 49 There was no interaction between the effects of fish oil and butter oil in different blends in the formation of both CDH and TBARS (e.g. MDA). The ANOVA test also showed that a PV response for the butter-fish oil blends was highly significant (P<0.001) and affected by blending ratio, temperature and time. In addition, significant two-way and three-way interactions (P<0.001) were obtained. A poor correlation for the PV response was obtained from the blending ratio with a lower Pearson coefficient (r=-0.218, P>0.05), which led to the conclusion that the influence of blending ratio on PV was less important than for the generation of CDH (Pearson coefficient, r=-0.624) or TBARS (e.g. MDA) (Pearson coefficient, r=-0.765). Furthermore, this result also meant that during thermal treatment unstable peroxides may be lost, which therefore restricts the usefulness of the PV measure for predicting quantitative changes in oxidation of the heat treated oil blends. Subsequent correlation tests indicated that PV was time-dependent (P<0.001) and temperature-related (P<0.05), but not dependent on the blending ratio (P>0.05) used in this study. However, the negative Pearson coefficient, suggests that the PV decreased at a smaller extent for butter-fish oil blends that contained a greater proportion of butter oil. Independent Sample T-test for PV indicated a significant difference between incubation times at both 8 h and 16 h, or 8 h and 24 h (P<0.001), but not between 16 h and 24 h (P=0.719). This finding confirmed the fact that oxidation of fatty acids occurred primarily up to 8 h or 16 h of heat treatment. A similar temperature-dependent (P<0.01) result for TBARS values was obtained when analyzed by ANOVA (Table 8 and Figures 10 and 11). Moreover, the correlation 50 obtained between PV and MDA content (Pearson coefficient, r=0.384), or CDH and MDA content (Pearson coefficient, r=0.369) was significant (P<0.001) in both cases. This result confirmed the presence of multiple oxidation products in these samples. The comparison of Figures 10 and 12, or Figures 11 and 13, is a good example of the above statistical result and lends a further support to the fact that the CDH content is a useful measure to predict primary lipid oxidation, and PV is applicable in comparison samples under same temperatures, whereas the MDA content is best used to estimate secondary lipid oxidation. 51 Discussion co-3 to co-6 fatty acid ratio in native oils and oil blends The metabolic pathways for both co-3 (e.g. from oc-linolenic acid) and co-6 fatty acids (e.g. from linoleic acid) have specific enzymes that are commonly involved in metabolizing both co-3 and co-6 polyunsaturated fatty acids (PUFAs) precursors. As a result, these PUFAs are in competition for these enzymes (Newton and Snyder 1997). The relative oversupply of co-6 PUFAs in the diet can impair the transformation of co-3 PUFAs into the longer-chain metabolites and lead to a relative imbalance in the different metabolic end-products (e.g. prostaglandins), which have many biological roles in metabolism and homeostasis (Shahidi and Wanasundara, 1998). Linoleic acid (C18:2co6), a predominant PUFA in many oil sources used in the Western diet, is consumed in higher quantities compared to co-3 fatty acids. Consequently, the co-3 to co-6 fatty acid ratio is usually in the range of 1:10 to 1:25. Presently there is some consensus that PUFAs should form at least 3%, and preferably 8-23%, of the total lipid intake, and that the co-3 to co-6 ratio should ideally lie between 1:4 and 1:10 (Marshall and Van Elswyk 1995). It is crucial to note that the relationship between chronic or acute disease states and the possible benefits of increasing dietary PUFA intake remains a highly contentious issue (Hummell, 1993). Generally, the ordinary sources of co-3 fatty acids are the marine oils and a few vegetable oils such as flaxseed oil. Fish oil has been considered a primary source of very long chain co-3 fatty acids, such as EPA and DHA. The supplementation of diets with 52 very long co-3 fatty acids from fish oil has been attributed to lower triglyceride levels (Rambjor, et al., 1996) and reduced hypertension (Smith et al., 1993). The fish oil chosen for the present experiment had a high co-3 to co-6 ratio, 7.4:1 (Table 2). With respect to lowering co-6 content of diets, fish oil is therefore a good source to achieve this purpose. However, fish oils are sensitive to oxidation reactions, which can result in loss of long chain PUFA. Butterfat may have a nutritional advantage over vegetable oils due to the prevalence of saturated fatty acids in butter, which are less prone to thermal oxidation (Kupranycz et. al., 1986). However, information on the thermal oxidative behavior of butter oil is lacking. In the present study, it was shown that butter oil could be blended to fish oil in such a manner that resulted in an acceptable co-3 to co-6 ratio, while also providing a greater degree of stability against thermal oxidation. The effect of oil blending on the lipid oxidation Berry and coworkers (1983) studied the physico-chemical characteristics of palm olein and soybean oil blends, and found that the oil blended samples containing 50% palm olein and 50% soybean oil (also tempered at 55°C for one hour and at 65°C for another one hour, prior to being stored at 9 ± 1°C) exhibited an improved oxidative stability for up to 6 weeks of storage. Peroxide and anisidine values were both used to access the overall oxidation value (Berry and Awang 1983). Theoretically, the blending of palm olein with soybean oil can produce an improved oxidative stability due to the lower degree of unsaturation of palm olein. This resulted in a lower peroxide value and a lower anisidine value as well. Similar results were achieved with the different oils used in 53 the present study. Blending butter oil with fish oil improved oxidative stability as assessed by lower peroxide value and reduced formation of CDH and TBARS. Effect of temperature and time on the thermal oxidation of butter-fish oil blends Regardless of how changes in EPA or DHA were evaluated, either using measurements for the relative amount in percentage (% of total fatty acids, w/w), or the absolute amount in mass (mg/gm oil), examination of these data indicated that the higher temperature (180°C) processing used to heat the oils resulted in the greatest oxidation of long chain co-3 fatty acids. The rate of loss in absolute amounts of PUFAs was greatest during the first 8 h before proceeding to relatively lower losses during the remainder of the experimental period. As the oxidation reaction proceeded, PUFAs were the first to be lost, followed by the more saturated fatty acids. Comparing the lag times for losses of individual fatty acids, as shown in Figures 4 to 9, it was clear that saturated fatty acids were the least labile to the high temperatures used in this experiment. As discussed, thermal treatment resulted in the rapid depletion of both EPA and DHA (Table 6 and 7). These specific PUFAs were oxidized first during heat-induced lipid oxidation and the experimental results in this study reflected the fact that temperature was the most significant factor in the thermal oxidation of fish oil for both the long chain PUFAs, (i.e., EPA, DHA, linolenic acid, and linoleic acid). These results which described the changes in individual PUFA content of the different oil blends correspond with results that described various chemical indices of oxidation, such as PV, the formation of CDH and TBARS. For example, the time of heating required to reduce residual fatty acids was comparable to the temporal pattern in which oxidation products, 54 such as peroxides, CDH and MDA, were generated during high temperature (150°C or 180°C) processing. Oxidation of co-3 fatty acids in butter-fish oil blends The present study also showed that the co-3 to co-6 fatty acid ratio followed a similar trend as far as heating temperature and time was concerned. The trend for a decrease in co-3 to co-6 fatty acid ratio with heating, was specific to the fact that co-3 fatty acids were lost relatively earlier than the co-6 fatty acids. This result reflects the greater susceptibility of co-3 fatty acids to oxidation, compared to co-6 fatty acids (Frankel, 1991). The majority of co-3 fatty acids in this study were EPA and DHA, which are very long chain fatty acids containing 5 or 6 double bonds. It is a well-known fact that the number, position and geometry of the double bonds can affect the oxidation rate of PUFA (Isbell, et al., 1999). Fennema (1985), furthermore, mentioned that relative rates of oxidation for arachidonic, linolenic, linoleic and oleic acid are approximately 40:20:10:1. In addition, cis fatty acids oxidize more readily than trans isomers, and conjugated double bonds are more reactive with oxygen than nonconjugated ones (Fennema 1985). Interesting research conducted by Kim and coworkers (1988) investigated the oxidation rates of mixed oils {e.g. 1:1 w/w mixtures of soybean oil and coconut oil, or soybean oil and palm oil) and the corresponding interesterified oils. These workers showed that the rate of oxidation of unsaturated fatty acids was consistently greater than that of saturated fatty acids (i.e., linolenic > linoleic > oleic acid) in these blends (Kim et al, 1988). It was concluded that the unsaturated fatty acids formed actively conjugated dienes during thermal oxidation. The hydroperoxides formed from the unsaturated acids 55 during thermal oxidation were decomposed faster than those formed from the saturated fatty acids. These results help explain why the blending of butter oil to fish oil resulted in a lower peroxide value and reduced formation of TBARS. Thus, thermal stability of oils can be characterized by the complete fatty acid composition of the lipid source. The findings herein also show that PUFAs are more susceptible to oxidation and are lost more rapidly than saturated fatty acids when present in a blend. Other workers have demonstrated that the more unsaturated fatty acids in oil actually oxidize faster in bulk oils, than the less unsaturated fatty acids (Rosao Romero et al, 1975; Cho et al., 1987). The opposite effect was found in an oil-in-water emulsion (Miyashita et al., 1993; Miyashita et al, 1994). 56 Conclusion (1) Marine oil composed of herring and salmon fish sources was analyzed for complete fatty acid composition and cholesterol content; the very long co-3 chain fatty acids, EPA and DHA were identified as being the primary sources of omega-3 fatty acids. Blending the fish oil with butter oil was performed using an optimization strategy with measured P/S and u)3/co6 PUFA ratios chosen as the endpoint measure estimates. The common employment of expressing fatty acid data on the basis of relative percentage of fatty acids was shown to be inferior to the quantitative measurement of fatty acids based on individual mass measures for the same fatty acids for measuring retention of fatty acids after thermal processing. Moreover, the null hypothesis established for enhancing retention of PUFAs in the butter-fish oil blends was rejected, as it was not possible to predict the best blend combination described by the designated quadratic model. The parallel measurement of a number of lipid oxidation product parameters generated from the thermal oxidation of the fish oil blends produced a different conclusion. Blending of fish oil by butter oil, was in fact successful at improving oxidative stability using initial (e.g. peroxide value (PV) and conjugated diene (CD)) and secondary lipid oxidation products (e.g. malonaldehyde (MDA)), as the endpoint measurement. However, blending of fish oil with butter oil was not a useful method to protect against the extent of thermal oxidation of PUFAs. It was concluded from this experiment that possibly the relatively small proportion of co-3, long chain PUFAs in the fish oil were not sufficient in concentration in the oil mixture and represented a limiting factor in successfully determining an optimal 57 proportion of butter oil required to protect PUFAs from thermal oxidation. In contrast, PUFAs were apparently more susceptible to oxidation and were lost more rapidly than those saturated fatty acids when present in the blend. 58 Table 1. Combinations of butter-fish oil blends containing fish oil and butter oil before thermal treatment.1 Butter-Fish O i l Blends Sample No. Fish O i l ( % ) Butter O i l ( % ) 1 100 2 0 2 100 0 3 50 50 4 50 50 5 0 100 6 0 100 7 33 67 8 33 67 9 33 67 10 33 67 11 67 33 12 33 67 13 33 67 Proportions were selected by the composition of fish o i l and butter o i l of an o i l blend in percentage (w/w). 2 Total proportion 100% (w/w) equals 3.00g of o i l . 59 Pi <N o C N O o T2 u o i f — ' u >i<=> fa 00 l-H O U © ve i-H © - 1 M >C CN —< O h Tf 0 0 0 0 f -(N O O O i n O 0 0 r-^  * O N * s * C S * C N 0 0 Tt O N C N 5 i n m -H i n m 3 r- C N — . 0 0 * . —1 rn C\ —1 rt m 0 0 O 0 0 0 0 i n C N ro O N r-~ vo rt *~* v o O rt I O m O N v d v d rn o \ Tt C N q # 0 O c o m —H m V O Tt i n Tt C N i n Tt Tt V O O O O N Tt i n V O *> pa fa fe u o. eo 6 c .0 rs o ca >-. rt T J ca b. 3 ca 3 .6 % b- J 3 ca R > > O I *S vo ^ ca —. 60 O e 0 5 v m b» w u ca CO co 60 4^ _ £ rn £ EC .3 J O o u u T J o c T3 C a o "•4—1 2 e 50 o cS (50 . s ' C <u x: s a CO J D . . 00 "O c x: 00 u rt 3 X ) o C r t O . & o H H 3 ca X) C 3 V S s m 3 cj HP. m C N i n C N O Tt Tt O O Tt V O Tt C N C N r ~ <n Tt rn o 0 0 O O N C N 0 0 o 0 0 O N m C N m C N —1 o 0 0 d o' o rt —< v o O O N v q ov rn —< Tt i n r ~ i n m C N —1 0 0 O N Tt O N v q r - ; 0 0 0 0 r~ —« rn i n Tt m C N C N i n vo t— O N O N Tt O N C N —; o Tt m 1-1 r ~ 0 0 C N v q p O N i n Tt cn rn Tt 0 0 vo m C N r» i n m 0 0 o C N m i n rn C N — q 0 0 rn cn 0 0 rn C N r~ vo un i n Tt O i n m — « vo rn p Tt 0 0 i n v d o-i V O O N V O Tt i n vo O N m r~ O N o q q i n -< 2 H O ft CA m [— i n m C N m vo m o m — I Tt V O 0 0 y-i rn i> Tt \6 i n 0 0 o C N m rt rt cN CN CN Tf rt Tt r~ Tt 0 0 C N rn v q Tt rn C N —< O N V O V O vo vo i n —1 r ~ m ^ O N 0 0 0 0 ~ n vi S ^ o E ca 5, b, « a. 60 S c >>.2 ,«* S rt I -cw •a 0 o c « 3 > > O tt E c « M -a 2 B 2. "5 CO X ca co ca CO CO >-> c CL, 3 2 >> co O U D. 3 . . l a Sr2 > EC .2 cn' + ' § 5 P>N. C/3 •2 ^ v o f . ca v d C II C rn -2 •-m ~ c a ' O 60 JS 4) CO Table 4. Quadratic regression of butter-fish oil blends and the changes in P/S ratio at 150°C and loTTC.1 Coefficients2 Time (h) 150°C 180°C 8 16 24 8 16 24 Po (intercept)3 0.127 0.0957 0.0916 0.0966 0.0747 0.0513 Pi 0.222"* 0.215*** 0.198*** 0.181*** 0.165*" 0.160*" Pi -0.172"* -0.167*" -0.154*** -0.141"* -0.128*** -0.124*** Pl2 -0.119"* -0.158*** -0.140"* -0.111*" -0.111*" -0.123*" CCl 0.153*** 0.203*** 0.180*" 0.143*** 0.143*** 0.159*** a 2 0.092*" 0.123*** 0.109*" 0.0861*** 0.0862"* 0.0958*" R 2 0.996*** 0.979"* 0.992*** 0.993*** 0.990*" 0.995*** ' Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***: P<0.001. 3 Po = intercept; $ } = coefficient for fish o i l at the first order; P 2 = coefficient for butter o i l at the first order; P 1 2 = coefficient for the interaction between oils; a! = coefficient for fish oil at the second order; a 2 = coefficient for butter oil at the second order. Table 5. Quadratic regression of butter-fish oil blends and the changes in co3/co6 ratio at 150°Candl80°C. a 150°C 180°C Coefficients'3 Time (h) 8 16 24 8 16 24 (3o (intercept) 2.401 1.592 1.948 1.836 1.381 0.867 p! 4.246*** 3.917*" 3.873*" 3.570*" 3.171*" 2.999*" p 2 -3.301*** -3.045"* -3.011"* -2.776*** -2.465*** -2.331*" p 1 2 -1.423"* -2.354*" -1.618*" -1.459*" -1.555*" -1.916*" ttl 1.830*** 3.028*" 2.081*" 1.877*** 2.000*** 2.464*** 02 1.106*** 1.830"* 1.258"* 1.134*" 1.209*" 1.489*" R 2 0.999*" 0.970*** 0.996*** 0.998*** 0.997*** 0.999*** 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***. p<o.001. 3 Po = intercept; P] = coefficient for fish o i l at the first order; p2 = coefficient for butter o i l at the first order; p l 2 = coefficient for the interaction between oils; = coefficient for fish oil at the second order; a 2 = coefficient for butter oil at the second order. 61 Table 6. Polyunsaturated fatty acid composition (% of total fatty acids), P/S ratio and co3/to6 ratio of butter-fish oil blends before and after thermal treatment at l s c c . 1 TIME C18:2 C18:3 C20:5 C22:6 P/S co3/a)6 Oh 8h 16 h 24 h A 2 1.88 0.88 7.98 8.69 0.508 7.403 B 1.81 0.70 5.32 5.79 0.278 5.516 C 1.78 0.61 3.99 4.34 0.197 4.413 D 1.75 0.52 2.66 2.90 0.130 3.177 E 1.68 0.34 0.00 0.00 0.028 0.201 A 2.10 0.73 7.94 7.51 0.391 7.256 ± 0 . 0 3 ± 0 . 0 2 ± 0 . 2 2 ± 0 . 2 8 ± 0 . 0 1 9 ± 0.034 B 2.33 0.66 5.03 4.67 0.204 4.737 ± 0 . 0 3 ± 0 . 0 1 ± 0 . 0 8 ± 0 . 0 7 ± 0 . 1 4 ± 0.021 C 2.41 0.59 3.47 3.28 0.179 2.997 ± 0 . 0 4 ± 0 . 0 1 ± 0 . 1 1 ± 0 . 1 0 ± 0.005 ± 0.035 D 2.44 0.51 2.13 1.96 0.092 1.957 ± 0 . 0 5 ± 0 . 0 3 ± 0 . 1 2 ± 0 . 1 2 ± 0.005 ± 0.067 E 2.61 0.49 0.00 0.00 0.035 0.213 ± 0 . 0 3 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 0 ± 0.001 ± 0.003 2.11 0.70 7.80 7.14 0.341 7.030 A ± 0 . 0 1 ± 0 . 0 3 ± 0 . 1 1 ± 0 . 0 7 ± 0 . 0 1 0 ± 0.094 B 2.15 0.60 4.10 3.96 0.153 4.257 ± 0 . 0 1 ± 0 . 0 2 ± 0 . 1 0 ± 0 . 0 9 ± 0.008 ± 0.077 C 2.31 0.56 2.88 2.75 0.122 2.669 ± 0 . 0 1 ± 0 . 0 7 ± 0 . 0 4 ± 0 . 1 2 ± 0 . 0 2 1 ± 0.024 D 2.32 0.46 1.98 1.84 0.089 1.895 ± 0 . 0 3 ± 0 . 0 7 ± 0 . 2 4 ± 0 . 0 8 ± 0 . 0 1 3 ± 0 . 0 1 3 E 2.36 0.38 0.00 0.00 0.031 0.174 ± 0 . 0 2 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.004 A 2.00 0.63 6.57 6.12 0.314 6.679 ± 0 . 0 2 ± 0 . 0 0 ± 0 . 0 2 ± 0 . 0 0 ± 0.006 ± 0.074 B 2.28 0.52 4.08 3.78 0.152 3.667 ± 0 . 0 3 ± 0 . 0 0 ± 0 . 0 1 ± 0 . 0 0 ± 0.002 ± 0.002 C 2.30 0.47 2.78 2.45 0.111 2.556 ± 0 . 0 5 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 8 ± 0.000 ± 0.096 D 2.33 0.41 1.64 1.37 0.072 1.508 ± 0 . 0 6 ± 0 . 0 2 ± 0 . 1 9 ± 0 . 0 9 ± 0.007 ± 0 . 0 6 1 E 2.39 0.37 0.00 0.00 0.031 0.168 ± 0 . 0 3 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 0 ± 0.001 ± 0.003 Values represent fatty acid composition in percentage of total fatty acids (w/w); P/S: polyunsaturated/saturated fatty acid ratio; co3/co6: omega 3/omega 6 fatty acid ratio. 2 A=100% fish oil; B=66.7% fish oil + 33.3% butter oil; C=50% fish oil + 50% butter oil; D=33.3% fish oil + 66.7% butter oil; E=100% butter oil. 62 Table 7. Polyunsaturated fatty acid composition (% of total fatty acids), P/S ratio and co3/co6 ratio of butter-fish oil blends before and after thermal treatment at iscrc. 1 TIME C18:2 C18:3 C20:5 C22:6 P/S co3/co6 Oh A 2 1.88 0.88 7.98 8.69 0.508 7.403 B 1.81 0.70 5.32 5.79 0.278 5.516 C 1.78 0.61 3.99 4.34 0.197 4.413 D 1.75 0.52 2.66 2.90 0.130 3.177 E 1.68 0.34 0.00 0.00 0.028 0.201 8h A 2.06 0.63 6.68 6.39 0.326 6.468 A ± 0 . 0 3 ± 0 . 0 1 ± 0 . 0 1 2 ± 0 . 1 6 ± 0 . 0 0 7 ± 0.040 B 2.25 0.54 3.81 3.67 0.158 3.840 ± 0 . 0 1 ± 0 . 0 1 ± 0 . 0 5 ± 0 . 0 9 ± 0.002 ± 0 . 0 1 1 C 2.24 0.42 2.45 2.35 0.104 2.345 ± 0 . 0 3 ± 0 . 0 0 ± 0 . 0 6 ± 0 . 0 8 ± 0.004 ± 0.030 D 2.34 0.48 1.66 1.54 0.077 1.631 ± 0 . 0 3 ± 0 . 0 2 ± 0 . 0 6 ± 0 . 0 4 ± 0.002 ± 0.039 E 2.09 0.30 0.00 0.00 0.026 0.144 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 16 h A 2.10 0.63 6.14 5.77 0.296 6.172 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 1 ± 0 . 0 4 ± 0.002 ± 0 . 0 1 3 B 2.11 0.50 2.86 2.69 0.121 3.074 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 2 ± 0.001 ± 0 . 0 1 2 C 2.67 0.50 2.22 2.01 0.094 2.136 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 5 ± 0 . 0 7 ± 0 . 0 0 1 ± 0.065 D 2.20 0.42 1.19 1.05 0.060 1.231 ± 0 . 0 6 ± 0 . 0 1 ± 0 . 0 1 ± 0 . 0 1 ± 0.005 ± 0.056 E 1.66 0.19 0.00 0.00 0.019 0.096 ± 0 . 0 0 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.004 24 h A 2.06 0.70 5.98 5.02 0.267 6.087 ± 0 . 0 2 ± 0 . 0 0 ± 0 . 0 4 ± 0 . 0 5 ± 0.004 ± 0 . 0 2 1 B 2.15 0.44 2.95 2.28 0.112 2.797 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 1 ± 0 . 0 0 ± 0.000 ± 0.007 C 2.17 0.41 1.82 1.38 0.076 1.673 ± 0 . 0 1 ± 0 . 0 0 ± 0 . 0 4 ± 0 . 0 2 ± 0.000 ± 0 . 0 1 5 D 2.02 0.31 0.82 0.42 0.041 0.733 ± 0 . 0 5 ± 0 . 0 2 ± 0 . 0 1 ± 0 . 0 2 ± 0.001 ± 0.026 E 1.37 0.00 0.00 0.00 0.013 0.000 ± 0 . 0 2 ± 0 . 0 0 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 Values represent fatty acid composition in percentage of total fatty acids (w/w); P /S : polyunsaturated/saturated fatty acid ratio; 0)3/co6: omega 3/omega 6 fatty acid ratio. A=100% fish o i l ; B=66.7% fish o i l + 33.3% butter o i l ; C=50% fish o i l + 50% butter o i l ; D=33.3% fish o i l + 66.7% butter o i l ; E=100% butter o i l . 63 Table 8. Effect of blending of fish oil and butter oil on peroxide value and the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) after 8, 16 and 24 h thermal treatment at 150°C and 180°C. 150°C 180"C 8hrs 16hrs 24hrs 8hrs 16hrs 24hrs PV1 A 4 1.82 1.89 1.74 1.80 4.34 4.94 ±0.075 ±0 .070 ±0 .130 ±0 .115 ±0 .140 ±0 .315 B 1.96 3.15 2.24 0.84 3.22 4.83 ±0 .097 ±0 .169 ±0 .152 ±0 .127 ±0 .164 ±0 .216 C 2.56 2.43 2.65 1.12 2.13 2.94 ±0.175 ±0 .190 ±0 .195 ±0 .140 ±0 .305 ±0 .000 D 2.12 1.95 1.61 1.15 3.31 2.29 ±0.146 ±0 .359 ±0 .351 ±0 .188 ±0 .272 ±0 .370 E 1.48 2.19 2.46 1.79 2.79 3.57 ±0.085 ±0 .025 ±0 .075 ±0 .130 ±0 .035 ±0 .210 C D H 2 A 4.502 3.776 5.031 3.919 4.190 5.153 ±0.285 ±0 .014 ±0 .339 ±0 .068 ±0 .054 ±0 .285 B 3.946 3.593 4.447 4.054 3.797 4.678 ±0 .258 ±0 .203 ±0 .244 ±0 .108 ±0 .081 ±0 .149 C 3.925 3.247 4.522 3.437 3.193 4.244 ±0 .081 ±0.407 ±0 .027 ±0 .136 ±0 .122 ±0 .203 D 3.871 3.010 4.185 3.295 2.899 4.047 ±0.312 ±0 .339 ±0 .244 ±0 .176 ±0 .176 ±0 .203 E 3.132 2.753 3.173 2.820 2.793 2.814 ±0.095 ±0 .027 ±0 .203 ±0 .014 ±0 .027 ±0 .027 M D A 3 A 2.153 2.156 1.769 2.435 2.774 2.630 ±0 .068 ±0 .002 ±0 .062 ±0 .002 ±0 .002 ±0 .056 B 1.946 2.084 1.646 1.976 2.120 2.048 ±0 .056 ±0 .122 ±0 .122 ±0 .038 ±0 .020 ±0 .026 C 1.781 1.748 1.598 2.033 2.033 1.937 ±0 .044 ±0 .0164 ±0 .164 ±0 .020 ±0 .044 ±0 .044 D 1.674 1.705 1.553 1.800 1.920 1.725 ±0.074 ±0 .068 ±0 .068 ±0 .110 ±0 .038 ±0 .026 E 1.295 1.289 1.262 1.607 1.499 1.430 ±0.002 ±0 .002 ±0 .002 ±0 .002 ±0 .062 ±0 .014 Values represent peroxide value (milliequiv. peroxide/kg oil) . 2 Values present conjugated diene hydroperoxide (umol/g oil) . 3 Values present thiobarbituric acid reactive substances ( T B A R S ) (umol/g oil) . 4 A=100% fish o i l ; B=66.7% fish o i l + 33.3% butter o i l ; C=50% fish o i l + 50% butter o i l ; D=33.3% fish o i l + 66.7% butter o i l ; E=100% butter o i l . 64 Table 9. Quadratic regression of butter-fish oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8, 16 and 24 h thermal treatment at 150°C and 180°C. 150°C 180°C Coefficients1 8hrs 16hrs 24hrs 8hrs 16hrs 24hrs PV 1 Po (intercept) 2.475 2.279 1.859 0.816 2.831 2.201 P i 0.436** 0.127 -0.231 -0.311* 0.449 0.549 P2 -0.339** -0.099 0.179 0.242* -0.349 -0.426 Pl2 0.867*** 0.260 -0.210 -0.975*** -0.719 -1 .918" -1.115*** -0.335 0.270 1.254*** 0.924 2.468** a 2 -0.674*** -0.202 0.163 0.758*" 0.559 1.419" R 2 0.695** 0.040 0.045 0.760*** 0.421 0 .698" CDH Po (intercept) 4.000 3.227 4.502 3.525 3.010 4.393 P i 0 .776" 0.727* 1.140"* 0.800** 0.793*** 1.424*** P2 -0 .602" -0.564* -0.887*" -0.622** -0.617*" -1.106*** Pl2 0.259 0.061 0.511 0.285 0.342 0.538 CCl -0.334 -0.079 -0.658 -0.366 0.439 -0.693 CC2 -0.202 -0.048 -0.397 -0.221 0.266 -0.419 R 2 0.657** 0.533* 0.701** 0.582* 0.889*** 0.759*** MDA Po (intercept) 1.800 1.926 1.644 1.884 1.992 1.812 P i 0.509*** 0.557** 0.311*** 0.437*** 0.654*** 0.624*** P2 -0.395*** -0.433** -0.242*" -0.340*** -0.508"* -0.484*** Pl2 0.114 0.240 0.143 -0.093 -0.087 -0.156* CCl -0.146 -0.308 -0.184 0.120 0.112 0.201* 0C2 -0.088 -0.187 -0.111 0.073 0.068 0.121* R 2 0.819*** 0.603** 0.703** 0.843*** 0.954*** 0.979*** 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 *: P<0.05; **: P<0.01; ***: P<0.001. 3 po = intercept; Pi = coefficient for fish oil at the first order; p 2 = coefficient for butter oil at the first order; Pi 2 = coefficient for the interaction between oils; ct] = coefficient for fish oil at the second order; a2 = coefficient for butter oil at the second order. 65 I l l l i Mk A B C D E A B C D E 16 h incubation 24 h incubation Figure 4. Comparison of experimental and hypothetical values on the changes in docosahexaenoic acid (DHA, C22:6co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of DHA, • = hypothetical value of DHA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil. 40.00 35.00 ^ 30.00 o 25.00 CL 20.00 tu "ro ^ 15.00 03 CD CC 10.00 5.00 0.00 8 h incubation Figure 5. Comparison of experimental and hypothetical values on the changes in eicosapentaenoic acid (EPA, C20:5co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of EPA, • = hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil. 66 45.00 40.00 35.00 B C D E A B C D E A B C D E 8 h incubation 16 h incubation 24 h incubation Figure 6. Comparison of experimental and hypothetical values on the changes in docosahexaenoic acid (DHA, C22:6co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of DHA, • = hypothetical value of DHA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil. 40.00 35.00 = 30.00 o co 1> 25.00 CL 20.00 LU § 15.00 ~a in rr 10.00 5.00 0.00 \ik L i * Ik A B C D E A B C D E A B C D E 8 h incubation 16 h incubation 24 h incubation Figure 7. Comparison of experimental and hypothetical values on the changes i  eicosapentaenoic acid (EPA, C20:5co3) of butter-fish oil blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of EPA, • = hypothetical value of EPA, A = 100% fish oil, B = 66.7% fish oil + 33.3% butter oil, C = 50% fish oil + 50% butter oil, D = 33.3% fish oil + 66.7% butter oil, E = 100% butter oil. 67 24 h incubation Figure 8. Comparison of experimental and hypothetical values on the changes in stearic acid of butter-fish oi l blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 150°C. • = experimental value of E P A , • = hypothetical value of E P A , A = 100% fish o i l , B = 66.7% fish oi l + 33.3% butter o i l , C = 50% fish oi l + 50% butter o i l , D = 33.3% fish oi l + 66.7% butter o i l , E = 100% butter o i l . 90.00 -i Figure 9. Comparison of experimental and hypothetical values on the changes in stearic acid of butter-fish oi l blends with different blending ratios after 8h, 16h, and 24h thermal treatment at 180°C. • = experimental value of E P A , • = hypothetical value of E P A , A = 100% fish o i l , B = 66.7% fish oi l + 33.3% butter o i l , C = 50% fish oi l + 50% butter o i l , D = 33.3% fish oi l + 66.7% butter o i l , E = 100% butter o i l . 68 0 ^ , . , , 0 8 16 24 Incubation time (h) Figure 10. The content of malonaldehyde ( M D A ) for butter-fish o i l blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% fish o i l , • 66.7% fish o i l and 33.3% butter o i l , • 50% fish oi l and 50% butter o i l , X 33.3% fish oi l and 66.7% butter o i l , x l 0 0 % butter o i l . 0.5 -, Incubation time (h) Figure 11. The content of malonaldehyde ( M D A ) for butter-fish o i l blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% fish o i l , • 66.7% fish o i l and 33.3% butter o i l , • 50% fish o i l and 50% butter o i l , X 33.3% fish oi l and 66.7% butter o i l , x l 0 0 % butter o i l . 69 6.000 0.000 8 16 Incubation time (h) 24 Figure 12. The content of conjugated diene hydroperoxides ( C D H ) for butter-fish o i l blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% fish o i l , • 66.7% fish oi l and 33.3% butter o i l , • 50% fish o i l and 50% butter o i l , X 33.3% fish o i l and 66.7% butter o i l , x l 0 0 % butter o i l . 6.000 n 0.000 8 16 Incubation time (h) 24 Figure 13. The content of conjugated diene hydroperoxides ( C D H ) for butter-fish o i l blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% fish o i l , • 66.7% fish o i l and 33.3% butter o i l , • 50% fish o i l and 50% butter o i l , X 33.3% fish oi l and 66.7% butter o i l , x l 0 0 % butter o i l . 70 Chapter 2. Blending of Flaxseed Oil with Butter Oil: Effect of blending on fatty acid profiles and thermal oxidative stability, without normalization of cholesterol content. Introduction Unlike fish oil, which is uniquely rich in co-3, long chain PUFAs, such as EPA (C20:5co-3) and DHA (C22:6co-3), flaxseed oil represents a vegetable oil that is especially abundant in alpha linolenic (C18:3 co-3) and to a lesser extent linoleic (C18:2 co-6) fatty acids. The difference between the two oils sources exist from the standpoint that there is marked difference in both the relative degree of unsaturation as well as the relative proportion of unsaturated fatty acids. In addition, the animal sterol, cholesterol is absent in vegetable oil. Recently, flaxseed oil has received considerable attention due its rich source of C 18:3 co-3 fatty acid (40-63% w/w) and possible non-lipid antioxidant materials. However, in the manufacturing of edible flaxseed oil, the fact that the oil is physically removed from the non-lipid fraction of the oil seed makes it particularly susceptible to oxidation reactions, which lead to the depletion of the essential co-6 linoleic acid and co-3 linolenic acid. Many oil extraction processes attempt to reduce the oxidation instability of the flaxseed oil by bottling the oil under nitrogen gas and recommending storage at refrigeration temperatures. The use of antioxidants has also been attempted to stabilize this oil, albeit to an extent that was not too successful (Chang et al., 1977). In experiment 1, blending of fish oil with butter oil was shown to provide a degree of protection for the long chain omega-3 fatty acids, however, it was concluded that the 71 absolute amounts of the omega-3 fatty acids in the oil were not sufficient to enable optimization of the blending with butter oil using the established quadratic model. The purpose of this experiment was to repeat the exercise used in experiment 1 with flaxseed oil, which contains a greater proportion of shorter chain omega 3 fatty acids, and a lower degree of unsaturation. The influence of blending butter oil with flaxseed oil will be tested to determine if lipid oxidation and changes in PUFA profile under selected time-temperature conditions can be minimized. 72 Materials and Methods Oil sources The source of butter oil was the same as reported in Chapter 1. Flaxseed oil (fresh, cold pressed, unrefined and antioxidant free*), was supplied by Flora Distributors Ltd. (Burnaby, B.C.). Both bulk oils were stored in the dark in a cold room (-35 °C) in suitable portions, until experiments started. This was done in order to avoid the freeze-thaw recycle during course of the study. Block Heater The block heater used in this Chapter was the same as the one reported in Chapter 1. Other chemicals All of the chemicals introduced into this chapter have been described in Chapter 1. Experimental design The Echip 6.04 (Echip, Inc., Hockessin, DE) program (Wheeler, 1993) was again simultaneously used to select the different oil blends and execute the experimental design. Calculation equations and statistical analysis were also provided by this program (Irwandi, et al, 1999). * antioxidant free refers to no added a-tocopherol following refinery. 73 L Constrained mixture design The oil blends were prepared in this chapter using a constrained mixture design for butter oil and flaxseed oil combinations (Table 10). The sum of variables in the mixture design totaled 100 % (or 1), as described in Chapter 1 previously. The low limit of each component was set at 0 % (or 0), and the high limit was set at 100 % (or 1), thus: 2ZXi= 100% (or 1.0) Consequently, a quadratic equation, which predicted the polyunsaturated to saturated fatty acid ratio, or the co-3 to co-6 fatty acid ratio was as follows: Y = Po +Pl#XFlaxseed+P2*XButter + Pl2 * X F l a x s e e c i • Xgutter + 0Ci * X Flaxseed + a2 •X2Butter> in which, Y = polyunsaturated to saturated fatty acid ratio, or, co-3 to co-6 fatty acid ratio; Xpiaxseed + X B utter = 1 (or 100%); Xnaxseed = composition of flaxseed oil in percentage; Xsutter = composition of butter oil in percentage; and Po intercept; Pi = coefficient for flaxseed oil at the first order; P 2= coefficient for butter oil at the first order; Pi2 = coefficient for the interaction between oils; 0Ci = coefficient for flaxseed oil at the second order; a2 = coefficient for butter oil at the second order. 74 2. Formulation of oil blends with flaxseed and butter oils Flaxseed oil and butter oil were used for the butter-flaxseed oil blends. As listed in Table 10, a total of 13 blended samples, each representing a different blending ratio, were tested. 3_, Heating studies Every butter-flaxseed oil blend formulated by the methods mentioned for butter-fish oil blends examined in Chapter 1 were used in different thermal heating trials. Each sample was duplicated in the study. Two temperatures (150 °C and 180 °C) and three heating duration periods (8 h, 16 h and 24 h) were used to test the different oil blends. Identical procedures as described in Chapter 1 were applied to the butter-flaxseed oil blends. Analysis of oxidation of oil blends The analytical protocols were as follows for each analysis. The methods have been described in Chapter 1. Detailed methods for each analysis can be found on the following pages: L Thiobarbituric acid reactive substances (TBARS) p.34 2. Conjugated diene hydroperoxide (CDH) p.34 3. Peroxide value (PV) p.35 4. Fatty acid quantification p.36 a. Fatty acid calibration curve p.36 75 b. Sample preparation p.37 c. Gas chromatography p.37 d. Fatty acid profile of native oils p.38 Statistical analysis of data The experimental design software, Echip 6.04, was used. Similar to Chapter 1, SPSS (SPSS for Windows, version 7.5, SPSS Inc.) was also used additionally to enhance the statistical analysis of this study. 76 Results Fatty acid profiles of blended oils with flaxseed oil and butter oil The fatty acid profile of the native flaxseed oil used in this study is given in Table 2. Flaxseed oil, typically contained a relatively high amount (58% of total fatty acids) of linolenic acid (C18:3co3), thus confirming that flaxseed oil is an excellent oil source of co-3 PUFA. The other noteworthy PUFA detected in flaxseed oil was linoleic acid (C18:2co6) (14% of total fatty acids). A relatively high (17%) level of oleic acid (C18:lco9), indicated that flaxseed oil is also a good source of monounsaturated fatty acid (MUFA). Overall, flaxseed oil contained a mixture of unsaturated fatty acids (i.e., 58% of linolenic acid, 14% of linoleic acid and 17% of oleic acid). From the results of this analysis, a relative high P/S ratio (e.g. 7.0:1) for flaxseed oil was obtained. The high content of linolenic acid, coupled with the relatively lower amount of linoleic acid also produced in a high to3/co6 ratio (e.g. 4.2:1) for flaxseed oil. Comparatively, butter oil displayed a low P/S ratio (e.g. 0.03:1) and a low co3/co6 ratio (e.g. 0.2:1), as mentioned earlier in Chapter 1. Flaxseed oil was considered in this study as another source of co-3 fatty acid, albeit with a dominance of shorter chain PUFA, as compared to fish oil. Similar to what was described in Chapter 1, butter-flaxseed oil blends were formulated to evaluate the effect of blending on reducing the susceptibility of PUFAs depletion resulting from heat-induced oxidation. According to the computerized experimental design (ECHTP 6), five blending ratios containing flaxseed oil and butter oil 77 in different proportions were tested. The calculated fatty acid profiles of all combinations are presented in Table 11. Assessment of a quadratic model used to predict the optimal combinations of flaxseed and butter oil blends, using P/S and co3/co6 ratios In RSM methodology, R 2 values greater than 0.75 are considered to adequately predict the effect of independent parameters on a measured response (Henika, 1982; Irwandi, et al., 2000). In this study, the ratios of the two oils used in oil blending represented various parameters and changes in both P/S ratio and co3/co6 ratio. The coefficients of the quadratic regression model used to determine the optimal proportions of flaxseed oil with butter oil in the oil blend are presented in Tables 12 and 13, respectively. A high level of significance (p<0.001) (Tables 12 and 13) was obtained for all of the individual treatments. The same quadratic model effectively predicted (R2>0.982, p<0.001) the P/S ratio (Table 12) that characterized the different butter-flaxseed oil blends, following different thermal treatments that included heating duration {e.g. up to 24 h) at both temperatures {e.g. 150°C and 180°C). Another quadratic model also significantly (R2>0.956, p<0.001) described the 0)3/co6 ratio (Table 13) obtained from the same blends and the same treatments mentioned above. The high correlation (R value) between these two data sets indicated that the optimization study supported the use of response surface methodology for predicting the blending of different butter oil and flaxseed oil combinations in this study. Tables 12 and 13 show that both flaxseed oil and butter oil contributed 78 significantly to P/S ratio and u)3/co6 ratio by both first and second order reactions, as well as exhibiting significant interaction (P<0.001). From Table 12, the values of pi and P2 for the P/S ratio, suggested that the greater effect of the flaxseed oil, or butter oil, occurred for the first 8 h treatment at 150°C. A similar result was obtained for the 16 h treatment at 180°C. It is of interest that the effect of these treatments on the P/S ratio was different in the butter-flaxseed oil blends compared to the butter-fish oil blends. Moreover, values of pi or P2 derived from the co3/co6 ratio showed less variance between treatment conditions at different incubation times (Table 13), compared to similar values obtained for the butter-fish oil blends. A blend optimized to contain 23% butter oil and 77% flaxseed oil was predicted to yield in the highest Cu3/u)6 ratio after heat processing. Obtaining a optimal blending ratio for the butter-flaxseed oil could not be done with the butter-fish oil experiment. This optimal blending ratio for butter-flaxseed oil was established regardless of both time and temperature of heating processing (Appendix D). Depletion of PUFAs during thermal incubation As mentioned previously, native flaxseed oil contained a high proportion of linolenic acid (58% of total fatty acids) and high P/S and co3/co6 ratios. The initial fatty acid profiles of butter-flaxseed oil blends are shown in Table 11. As to be expected, the portion of linolenic acid decreased with the addition of flaxseed oil in the formulation, as did the P/S ratio and co3/to6 ratio. The content of saturated fatty acids increased with the 79 decline in the relative proportion of flaxseed oil, reflecting fatty acids that were derived mostly from butter oil in these blends. A similar experimental approach was used to evaluate the effect of flaxseed oil blended with butter oil on the performance of lipid oxidation. The resultant fatty acid profiles following thermal treatment at 150°C and 180°C are shown in Tables 13 and 14, respectively. Thermal treatments at either 150°C or 180°C led to the depletion of (0-3 fatty acid (e.g. linolenic acid), the major PUFA present in the butter-flaxseed oil blends (Table 14). This change primarily happened during the first 8 h incubation, and it was more dramatic when the temperature was increased to 180°C. A similar decreasing trend could also be found with the 0)3/0)6 ratio, but not with the co-6 fatty acid alone (e.g. linoleic acid), or the P/S ratio. Therefore, the depletion of co-3 fatty acid was principally responsible for these changes due to the greater susceptibility to thermal degradation than co-6 fatty acids in the various butter-flaxseed oil blends. Decreasing P/S ratio with time of heating indicated the fact that PUFAs were lost at a much faster rate than saturated fatty acids in these blends. It is interesting to note that, on the contrary, the percentage of linoleic acid actually increased because of the considerable decrease in linolenic acid. This finding further highlights the somewhat restrictive limitation of using proportional changes of fatty acid composition to describe absolute changes of fatty acids that are lost due to heating. In the present study, the P/S and the 0)3/co6 ratios of the different blends were both higher after 150°C heating than 180°C, which reflected, in general the fact that unsaturated fatty acids, especially the PUFAs, are more susceptible to destruction at the higher temperature. 80 Evaluation of blending and heat processing on the absolute amount of PUFAs of butter-flaxseed oil blends As shown in Chapter 1, the absolute amount of fatty acids provided a more sensitive indicator of quantitative fatty acid change during thermal treatment than the relative percentage of total fatty acids. As a result, all of the statistical analyses were executed on the basis of the absolute amount of fatty acids (e.g. mass of fatty acids) for the purpose of revealing the relative susceptibility of individual fatty acid loss within the butter-flaxseed oil blends following a thermal treatment. The ANOVA test showed that temperature (P<0.001), incubation time (P<0.05) and blending ratio (P<0.001) were all significant factors affecting the linoleic acid and linolenic acid residues in the butter-flaxseed oil blends. An interaction between blending ratio x time was also significant (P<0.05) in regard to influencing the retention of the PUFAs. The interaction for blending ratio x temperature produced a significant effect on the rate of oxidation for linolenic acid (P<0.001), but not linoleic acid (N.S.). A correlation test for residual linoleic acid (C18:2co6) and linolenic acid (C18:3co3) was also accomplished. Negative Pearson coefficients were produced from residual linoleic acid (C18:2u)6) for temperature (r=-0.130, N.S.), time (r=-0.220, P<0.05), and blending ratio (r=-0.939, P<0.01). Similar coefficients derived from residual linolenic acid were as follows: r=-0.152 (N.S.) for temperature, r=-0.248 (P<0.05) for time and r=-0.919 (P<0.01) for blending ratio. Therefore, losses of either linoleic or linolenic PUFAs from flaxseed oil were time-dependent, as well as dependent on the blending ratio, but not dependent on the differences between 150°C and 180°C. 81 The finding that the 150°C temperature damaged the PUFAs in flaxseed oil to such a serious degree is one explanation for the fact that a higher temperature (180°C) did not produce a statistically stronger effect on the thermal oxidation of PUFAs. This is in contrast to the correlation analysis for butter-fish oil blends, which showed that the loss of EPA or DHA was higher at a higher temperature (180°C). Effect of blending strategy on the depletion of PUFA among butter-flaxseed oil blends during thermal treatments Following exactly the strategy for the butter-fish oil blends described in Chapter 1, hypothetical values were calculated subsequently from the averages of blending A (100% fish oil) and E (100% butter oil). These values were used to compare with experimental results. Both experimental and hypothetical values for fatty acids in blend A and blend E are identical. Experimental values were used to examine the degree of change in fatty acid concentration of blended oils after the thermal treatment, whereas hypothetical values were calculated to predict the changes in fatty acid concentration resulting from blending the heated flaxseed and butter oils together. Figures 14 and 16 illustrated the comparison of experimental and hypothetical values for the relative changes in linoleic acid and linolenic acid content of oils during a 24 h thermal incubation at 150°C. Similar to the butter-fish oil blends, the hypothetical values for different butter-flaxseed oil blends were markedly higher than the corresponding experimental values, thus indicating that linoleic acid and linolenic acid were both more heat liable following blending at a high temperature of 150°C. It is interesting to note that the butter oil blending of flaxseed oil did not proportionally reduce 82 the amount of PUFA when exposed to high temperature, but rather led to greater loss of both linoleic acid and linolenic acid. Figures 15 and 17 show the retention of linoleic and linolenic acids in the various blends after 180°C thermal incubation. At this higher temperature, greater losses of linoleic and linolenic acids occurred during the first 8h thermal treatment, compared to heating at 150°C. The hypothetical values obtained for PUFAs at 180°C followed the same trend as samples heated at 150°C. This finding further demonstrates that blending oil, resulting in the dilution of some specific fatty acids, actually enhances the susceptibility of PUFAs, such as linoleic and linolenic acids, to thermal oxidation when exposed to high temperatures. As far as the shorter chain and more saturated fatty acids were concerned, the hypothetical values were actually lower than experimental ones (Figures 18 and 19). Thus increasing temperatures to high levels had little effect on residual shorter chain and more saturated fatty acids in this study. Thus, as was the case in the butter-fish oil blends, the shorter chain and saturated fatty acid (e.g. palmitic acid (C16:0)), were more stable after blending the two principle native oils. For example, during the first 8 h of heating, the oxidation of palmitic acid dominated, and produced only minute changes thereafter for the rest of the incubation time. The experimental values for stearic acid (C18:0) were also comparable to hypothetical values. This result confirms that blending of oil retards the thermal oxidation of saturated fatty acids (e.g. palmitic acid, stearic acid). Information of lipid oxidation from chemical analyses The oxidation of butter-flaxseed oil blends was also evaluated measuring lipid peroxides (PV), conjugated diene hydroperoxide (CDH) and thiobarbituric acid reactive substances (TBARS) formation. Table 15 shows the effect of blending butter oil with flaxseed oil on the formation of peroxides, CDH and TBARS (e.g. MDA) following thermal treatments at 150°C and 180°C, respectively, for up 24 h. Coefficients of determinations (R2) and quadratic regression coefficients for PV, CDH content and TBARS (e.g. MDA) content are given in Table 16. The coefficients of determinations (R2) derived from the CDH content under higher temperature (180°C), were also comparatively greater than those derived from PV. This finding indicates that the formation of CDH was possibly a relatively more accurate estimate of lipid oxidation. Coefficients derived from PV, the formation of CDH and TBARS (e.g. MDA) indicated that oxidation reactions for flaxseed oil and butter oil were both significantly affected by first order reactions. And optimization could not be obtained. Analyzing data using ANOVA showed that the PV in the butter-flaxseed oil blends was significantly affected by temperature (p<0.001), blending ratio (p<0.001), but not incubation time (p>0.05). Similarly, temperature (p<0.01) and blending ratio (p<0.001) but not incubation time (P>0.05) were significant factors that affected the formation of CDH and TBARS of butter oil blended flaxseed oil. Significant interactions between the blending ratio x time (P<0.001), temperature x time (P<0.001) and blending ratio x temperature x time (P<0.001) were statistically significant to the formation of 84 hydroperoxides. No significant interactions between these main effects for the TBARS content (e.g. MDA) were observed. Performing a correlation test of the different variables showed that blending ratio was a significant factor affecting the content of hydroperoxide (r=0.281, Pearson coefficient, p<0.05) and TBARS (e.g. MDA) content (r=-0.475, Pearson coefficient, p<0.001). This finding confirmed the expectation that the total amount of unsaturated fatty acids in the oil overall affected the susceptibility of the blended oil to oxidation. Furthermore, temperature also affected PV with Pearson coefficient r=-0.422 (p<0.001) and TBARS (e.g. MDA) content with Pearson coefficient r=0.344 (p<0.001). Moreover, both PV and TBARS (e.g. MDA) content in this study were negatively correlated (r=-0.250, P<0.05), which probably illustrated that more decomposition of primary oxidation products may have occurred prior to the formation of secondary oxidation products (Figures 20 to 23). 85 Discussion co-3 to co-6 fatty acid ratio in native oils and oil blends The rich source of linolenic acid in flaxseed oil produced the characteristically high co3/co6 fatty acid ratio (1:0.241) and the uniquely high P/S ratio that is specific for many vegetable oils. The blending of flaxseed oil with butter oil resulted in distinctly different co3/co6 fatty acid ratios observed earlier with the butter-fish oil blends. These differences were attributed to the relatively higher proportions of linolenic acid in flaxseed oil, compared to the combined linolenic acid, EPA and DHA content in fish oil. The significance of these 0)3/co6 ratios has been reported in contemporary nutrition studies, which have recommended that the co3/co6 ideally fall within a 1:4 to 1:10 ratio (Marshall and Van Elswyk, 1995). In the present Western diet, however, the co3/co6 ratio is in the neighborhood of 1:10 to 1:25, with PUFAs contributions at least of 3%, and preferably 8-23% of the total lipid mixture. Lacking from these studies is information on the final co3/co6 ratio of food lipids that have undergone a thermal process and which have resulted in oxidative losses of individual essential co3 and 0)6 PUFAs. Flaxseed contains about 34-45% oil, of which oc-linolenic acid accounts for 50-55%. New flaxseed varieties are also available (for example, Linola™) which contain a fatty acid profile similar to that of sunflower oil (i.e., high linoleic acid (71.9-73.9%) and low linolenic acid (1.8-2.0%) (Haumann, 1993; Weiss, 1993). Generally, the high rate of autoxidation and thermal polymerization has made traditional flaxseed oil unsuitable for salad and cooking application (Swern 1979). 86 Effectiveness of high temperature on the lipid oxidation Monitoring of fatty acid changes in oils during high temperature (e.g. 180°C) processing is an effective method to assess thermal oxidative changes to oils. Thomson and coworkers (1983) reported that total linoleic and linolenic acids decreased by 50% after 100 h of intermittent frying. In this current work, linolenic acid and linoleic acid content of flaxseed oil depleted at characteristically higher rates, compared to less unsaturated fatty acids in the native flaxseed oil and showed a similar pattern to the depletion of EPA and DHA noticed in fish oil. Experimental results showed that temperature was a significant factor in the thermal oxidation of flaxseed oil as evidenced by reduced residual linolenic acid and linoleic acid retention and the corresponding PV, formation of TBARS (e.g. MDA) and CDH. The duration of heating was also another factor that was noted to affect the degree of oxidation of oils heated at a high temperature of 150°C or 180°C. Miller et al. (1988) reported in their study that a decrease in linoleic and linolenic acids was obtained in soybean oils after 40 h heating at 180°C. Jaswir et al. (2000) executed frying experiments on potato chips with a refined, bleached and deodorized palm olein. Thermal treatment at 180°C was chosen in this study to represent a temperature commonly used for frying foods. Repeating the experiment at a slightly lower temperature for similar time periods enabled a relative comparison of the stability of the different fatty acid components in the oil blends. One recent study has shown palm oil heated at 180°C showed a decrease in the relative composition of linoleic and linolenic acids after 5 h frying (Irwandi, 1999). Coupled with this finding is a marked increase in 87 the proportion of palmitic acid following frying at 180°C, presumably due to the breakdown of double bonds of higher carbon numbers fatty acids (Augustin et al., 1983). The effect of oil blending on the lipid oxidation The quadratic model used in this study enabled the prediction of thermal induced degradation of PUFA as measured by both P/S and 0)3/0)6 ratios. Whether the source of co-3 fatty acids was in the fish oil or flaxseed oil, blending with butter oil had no appreciable effect on the stability of PUFAs of varying chain length or degree of unsaturation. In the present study, blending the flaxseed oil with a saturated fat source, such as butter oil, resulted in a relatively more effective retention of co-3 PUFA than that observed with the butter-fish oil blends. The comparison of hypothetical and experimental values produced trends that were solely attributable to a dilution effect caused by blending of PUFA with butter oil. However, in the example with butter-flaxseed oil blends, optimization could be achieved in respect to maximum retention of co3/co6 ratio; a finding which was not possible with the butter-fish oil blends. The reason for successful optimal blending of the butter-flaxseed oil combination was the considerably high proportion of co-3 PUFA in flaxseed oil. Therefore, the starting concentration of co-3 PUFA in an oil blend is a more important factor than the specific degree of unsaturation of the co-3 PUFA in the oil blend used for optimization. For example, the total co-3 PUFAs in fish oil from linolenic acid, EPA and DHA represented only 20% of the total co-3 PUFA (e.g. linolenic acid) present in the flaxseed oil. 88 Prediction of the relative loss of co-3 PUFA Predicting the relative loss of co-3 PUFA in the butter-flaxseed oil blends could also be achieved from the MDA response generated from the quadratic model. This finding was similar to the results obtained in the butter-fish oil blends. The fact that, PV and CDH could not be efficiently used as indicators for predicting changes in co-3 fatty acids for these blends, may be explained by the volatile characteristic of both compounds which rendered susceptible to loss when oil blends were exposed to high temperature processing. Although the MDA parameter was more reliable than PV and CDH content, respectively, this chemical indicator was not as good as the co3/co6 ratio in predicting loss of co-3 PUFA. As a precursor of secondary oxidation products (Gunstone, 1996), MDA is claimed to be an important biological breakdown product expected from 5-membered cyclic peroxides of linoleate and linolenate because of its crosslinking ability with amino groups of proteins, enzymes and DNA (Frankel, 1984). Under high temperature, the thermal decomposition of cyclic peroxides from linoleate produce most of the same volatile cleavage products as the corresponding mono-hydroperoxides (Appendix H). The most important cleavage between the hydroperoxide group and the cyclic peroxide produces aldehydes and aldehyde esters. Unsaturated aldehydes and aldehyde esters are formed from the cleavage of the peroxide ring. Unsaturated methyl ketones are among some of the unique products of cyclic peroxides. Therefore, the opportunity to form MDA, a volatile oxidation product, is less than 19% of the total decomposition products (Appendix H, Frankel, 1984). 89 Relative oxidation rate of PUFAs in flaxseed and fish oils The butter oil blending of flaxseed oil resulted in reduced residual PUFAs, suggesting that the total oxidation of PUFAs was less due to the measure of a greater concentration of saturated fatty acids, which in turn was less susceptible to oxidation. However, the blending of butter oil with flaxseed oil actually resulted in more PUFAs becoming oxidized. The fact that similar high temperature (150°C) processing and shorter time (8-16 h) of incubation, resulted in a higher retention of DHA in fish oil than linolenic acid in flaxseed oil was a particularly interesting finding (Appendix E). DHA is a longer chain fatty acid compared to EPA and linolenic acid. It also contains more double bonds, and therefore should be more susceptible to thermal oxidation. But the findings indicated that DHA in fish oil had a comparatively greater oxidative stability to EPA in fish oil and to linolenic acid in flaxseed oil, at lower temperature and shorter time used in this study. Thus, it appears that the relative sensitivity of both classes of PUFA in the two different butter oil blends was due to more than just saturated fatty acid content and may involve some other factors endogenous to the oil composition. A similar finding, however, was not obtained at the higher temperature (e.g. 180°C) processing for both butter oil blends (Appendix F). Comparing the relative patterns of loss for both linolenic and linoleic acids in flaxseed oil, with that of EPA and DHA in fish oil following heat treatment at 180°C for 8,16, and 24h, it can be concluded that both oils lost most PUFAs during the first 8 hours. However, the depletion of EPA and DHA in fish oil was consistently greater than 90 the relatively shorter chain co-3 fatty acids in flaxseed oil, and therefore was more sensitive to thermal oxidative stress at 180°C (Appendix G). 91 Conclusion (2) The comparison of the relative changes in fatty acids in both butter-fish oil blends and butter-flaxseed oil blends following thermal oxidation indicated similar changes in decreased P/S and o)3/o)6 ratios with blending. This result was attributed to similar susceptibilities of the major 0)3 fatty acids of both oil blends, namely, linolenic acid for flaxseed oil and EPA and DHA for fish oil to thermal oxidation. A quadratic model was developed to describe the changes in these and other fatty acids. Residual PUFA content and lipid oxidation indices, such as the formation of peroxide, TBARS (e.g. MDA) and CDH, demonstrated that these indicators of lipid oxidation were not absolutely equivalent to both oil blends but did produce similar overall trends. Common to both oil blends was the finding that the greater the concentration of PUFAs in fish oil (Frankel, 1985) or flaxseed oil (Swern, 1979), the greater the relative oxidation of the fatty acid. However, an interesting finding came forth from these studies that rejected the null hypothesis that blending with butter oil could assist in preserving omega-3 PUFAs from oxidation. In fact, optimization results based on the u)3/o)6 ratio demonstrated that blending with butter oil actually produced a more profound effect on the omega-3 essential fatty acid retention in butter-flaxseed oil blends than the butter-fish oil blends. The difference between the results obtained from butter-fish oil and butter-flaxseed oil blends raises the question as to the role of other components in both oil sources, which may have affected the depletion of PUFA and subsequent oxidation. In addition to the different source and concentration of o)-3 fatty acids in the fatty acid compositions of both fish oil and flaxseed oil, it should also be pointed out that the higher cholesterol content of the fish oil 92 may have been an additional factor in the different results obtained between the two blends. This aspect will be investigated further in Chapter 3 of this thesis. Table 10. Combinations of butter-flaxseed oil blends containing flaxseed oil and butter oil before thermal treatment.1 Butter-flaxseed Oil Blends Sample No. Flaxseed Oil (%) Butter Oil(%) 1 100 2 0 2 100 0 3 50 50 4 50 50 5 0 100 6 0 100 7 33 67 8 33 67 9 33 67 10 33 67 11 67 33 12 33 67 13 33 67 Proportions were selected by the composition of flaxseed o i l and butter o i l of an o i l blend in percentage (w/w). 2 Total proportion 100% (w/w) equals 3g of o i l . Table 11. Fatty acid composition of butter-flaxseed oil blends before thermal treatment (mg fatty acid/gm oil, P/S ratio and co3/co6 ratio).1 Fatty Acids Oil C12:0 C14:0 C16:0 C16:l 08:0 C18:l C18:2 C18:3 P/S (o3/co6 A 2 * * 54.43 * 43.91 169.5 133.64 555.44 7.007 4.156 B 3.91 19.88 121.66 2.71 61.46 160.62 92.38 370.95 2.239 4.015 C 5.87 29.82 155.28 4.06 70.24 156.18 71.76 278.71 1.342 3.884 D 7.83 39.76 188.90 5.41 79.01 151.73 51.13 186.47 0.753 3.647 E 11.74 59.64 256.13 8.12 96.56 142.85 9.87 1.98 0.028 0.201 Values represent the absolute amount of fatty acid in mg per gram o i l (w/w); P/S: polyunsaturated/saturated fatty acid ratio; co3/co6: omega 3/omega 6 fatty acid ratio. 2 A=100% flaxseed o i l ; B=66.7% flaxseed o i l + 33.3% butter o i l ; C=50% flaxseed o i l + 50% butter o i l ; D=33.3% flaxseed o i l + 66.7% butter o i l ; E=100% butter o i l . * not detectable or < 0.01% of total fatty acids. 94 Table 12. Quadratic regression of butter-flaxseed oil blends and the changes in P/S ratio at 150°C and 180°C. 1 Coefficients2 Time (h) 150°C 180°C 8 16 24 8 16 24 Po (intercept)3 0.146 0.172 0.121 0.173 0.0213 0.0487 P i 2.806*** 2.527*** 2.514*** 2.337*** 2.451*** 2.278*** P2 -2.182*** -1.964*** -1.954*** -1.817*** -1.905*** -1.771*** Pl2 -2.898*** -2.538*** -2.596*** -2.338*** -2.675*** -2.423*** Oil 3.727*** 3.265*** 3.340*** 3.008*** 3.441*** 3.117*** <X2 2.252*** 1.973*** 2.018*** 1.818*** 2.080*** 1.884*** R 2 0.986*** 0.988*** 0.986*** 0.987*** 0.982*** 0.984*** Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***:P<0.001. 3 Po = intercept; Pi = coefficient for flaxseed oil at the first order; p 2 = coefficient for butter oil at the first order; P i 2 = coefficient for the interaction between oils; CC] = coefficient for flaxseed oil at the second order; a 2 = coefficient for butter oil at the second order. Table 13. Quadratic regression of butter-flaxseed oil blends and the changes in co3/u)6 ratio at 150°Cand 180°C. a Temperature (°C) 150 180 Coefficientsb Time (h) 8 16 24 8 16 24 Po (intercept) 4.017 3.812 3.650 3.636 3.510 3.476 P i 2.622*** 2.467*** 2.470*** 2.315*** 2.288*** 2.390*** p 2 -2.039*** -1.918*** -1.920*** -1.800*** -1.779*** -1.858*** p 1 2 2.128*** 2.057*** 1.866*** 1.976*** 1.949*** 1.858*** on -2.738*** -2.646*** -2.401*** -2.542*** -2.507*** -2.390*** 02 -1.655*** -1.599*** -1.451*** -1.536*** -1.515*'* -1.444*** R 2 0.968*** 0.978*** 0.974*** 0.954*** 0.966*** 0.975*** 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***: P<0.001. 3 Po = intercept; P, = coefficient for flaxseed oil at the first order; p 2 = coefficient for butter oil at the first order; p l 2 = coefficient for the interaction between oils; a! = coefficient for flaxseed oil at the second order; a 2 = coefficient for butter oil at the second order. 95 Table 14. Polyunsaturated fatty acid composition (% of total fatty acids, w/w), P/S ratio and co3/co6 ratio of butter-flaxseed oil blends before and after thermal treatment at 150°C and 180°C.1 TIME 150°C 180°C C18:2 C18:3 P/S co3/co6 C18:2 C18:3 P/S co3/co6 Oh A 2 13.97 58.04 7.007 4.155 13.97 58.04 7.007 4.155 B 11.08 44.50 1.574 4.016 11.08 44.50 1.574 4.016 C 9.30 36.11 0.897 3.883 9.30 36.11 0.897 3.883 D 7.20 26.25 0.491 3.646 7.20 26.25 0.491 3.646 E 1.68 0.34 0.028 0.202 1.68 0.34 0.028 0.202 8h 16.51 54.30 6.825 3.289 17.21 49.63 5.632 2.883 A ± 0 . 1 7 ± 0 . 0 5 ± 0.026 ± 0.037 ± 0 . 3 1 ± 0 . 0 3 ± 0 . 0 5 1 ± 0.050 B 12.49 37.99 1.841 3.043 12.32 34.24 1.601 2.779 ± 0 . 0 4 + 0.02 ± 0.000 ± 0.000 + 0.04 ± 0 . 0 1 ± 0.002 ± 0.001 C 9.78 28.25 1.021 2.889 10.03 24.64 0.872 2.457 ± 0 . 4 7 ± 0 . 0 9 ± 0.026 ± 0 . 1 2 9 + 0.20 ± 0 . 0 8 ± 0 . 0 1 1 ± 0.042 D 7.85 19.91 0.594 2.535 7.50 17.54 0.522 2.339 ± 0 . 1 8 ± 0 . 3 1 ± 0.008 ± 0 . 0 9 1 ± 0 . 3 1 + 1.10. ± 0.035 ± 0.057 E 2.62 0.47 0.045 0.178 2.19 0.33 0.037 0.150 ± 0 . 0 0 ± 0 . 0 4 ± 0.001 ± 0 . 0 1 6 ± 0 . 0 1 ± 0 . 0 0 ± 0.000 ± 0.001 16 h A 16.84 51.47 6.086 3.057 17.76 49.33 6.061 2.777 A ± 0 . 1 1 ± 0 . 0 4 ± 0.008 ± 0 . 0 1 7 ± 0 . 4 5 ± 0 . 0 1 ± 0.039 ± 0 . 0 7 1 B 12.30 35.54 1.662 2.891 12.20 32.28 1.427 2.646 ± 0 . 0 6 ± 0 . 0 2 ± 0.003 ± 0.002 ± 0 . 0 2 ± 0 . 0 1 ± 0.003 ± 0.005 C 9.94 27.76 1.018 2.792 9.78 24.20 0.838 2.475 ± 0 . 0 1 ± 0 . 4 4 ± 0 . 0 1 0 ± 0.049 ± 0 . 0 4 ± 0 . 8 3 ± 0.034 ± 0.075 D 7.50 17.83 0.528 2.377 7.23 15.95 0.468 2.206 ± 0 . 1 0 ± 0 . 4 4 ± 0 . 0 1 4 ± 0.079 ± 0 . 1 5 + 0.88 ± 0.028 + 0.083 E 2.48 0.45 0.044 0.181 1.69 0.19 0.027 0.115 ± 0 . 0 2 ± 0 . 0 1 ± 0.001 ± 0.003 ± 0 . 0 2 ± 0 . 0 0 ± 0.001 ± 0.001 24 h A B C D E 16.62 51.78 6.100 3.116 17.01 49.76 5.570 2.925 ± 0 . 4 2 ± 0 . 7 7 ± 0 . 1 4 4 ± 0 . 1 2 5 ± 0 . 0 8 ± 0 . 5 2 ± 0.061 ± 0 . 0 1 7 12.27 35.02 1.608 2.854 11.92 31.63 1.361 2.653 ± 0 . 0 0 ± 0 . 0 0 ± 0.002 ± 0 . 0 0 1 ± 0 . 0 1 ± 0 . 0 2 ± 0.003 ± 0.004 10.02 26.99 0.965 2.693 9.63 24.46 0.838 2.541 ± 0 . 0 3 ± 0 . 3 0 ± 0 . 0 1 4 ± 0.022 ± 0 . 3 4 ± 0 . 2 2 ± 0.009 ± 0 . 1 1 2 7.53 17.18 0.505 2.282 6.93 14.96 0.425 2.159 ± 0 . 1 2 ± 0 . 7 1 ± 0 . 0 1 8 ± 0 . 1 1 8 ± 0 . 1 8 ± 0 . 7 7 ± 0.024 ± 0.066 2.37 0.41 0.041 0.171 1.43 0.14 0.022 0.101 ± 0 . 0 2 ± 0 . 0 0 ± 0.000 ± 0.001 ± 0 . 0 1 ± 0 . 0 0 ± 0.000 ± 0.000 Values represent fatty acid composition in percentage of total fatty acids (w/w); P/S: polyunsaturated/saturated fatty acid ratio; u)3/co6: omega 3/omega 6 fatty acid ratio. A=100% flaxseed o i l ; B=66.7% flaxseed o i l + 33.3% butter o i l ; 0 5 0 % flaxseed o i l + 50% butter o i l ; D=33.3% flaxseed o i l + 66.7% butter o i l ; E=100% butter o i l . 96 Table 15. Effect of blending of flaxseed oil and butter oil on the peroxide value (PV) and the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) after 8, 16 and 24 h thermal treatment at 150°C and 180°C. 150°C 180°C 8 h 16 h 24 h 8 h 16 h 24 h PV 1 A 4 2.59 3.50 3.22 0.77 3.78 2.36 ±0 .070 ±0 .140 ±0 .210 ±0 .070 ±0 .070 ±0 .095 B 5.18 2.24 1.68 1.96 1.89 3.15 ±0 .293 ±0 .164 ±0 .277 ±0 .168 ±0 .173 ±0 .092 C 4.05 4.43 3.85 2.42 3.92 2.52 ±0.225 ±0 .085 ±0 .560 ±0 .315 ±0 .280 ±0 .070 D 3.67 3.68 2.46 2.16 2.37 2.79 ±0.605 ±0 .274 ±0 .488 ±0 .414 ±0 .237 ±0 .189 E 3.57 6.58 3.92 3.40 4.16 3.70 ±0 .280 ±0 .000 ±0 .000 ±0 .245 ±0 .180 ±0 .340 C D H 2 " A 3.647 3.620 6.061 4.061 5.105 6.034 ±0 .163 ±0 .136 ±0 .258 ±0 .136 ±0 .054 ±0 .339 B 3.390 3.512 5.098 3.729 4.786 5.695 ±0 .122 ±0 .190 ±0 .325 ±0 .190 ±0 .244 ±0 .231 C 3.937 3.078 4.102 3.437 4.190 5.437 ±0 .136 ±0 .054 ±0 .136 ±0 .271 ±0 .095 ±0 .312 D 3.512 3.471 4.529 2.834 4.081 4.502 ±0 .244 ±0 .244 ±0 .176 ±0 .136 ±0 .244 ±0 .312 E 4.014 2.780 3.363 2.807 2.766 2.997 ±0 .163 ±0 .041 ±0.231 ±0 .027 ±0 .014 ±0 .203 M D A 3 A 2.030 1.865 1.676 1.772 1.922 1.937 ±0 .094 ±0 .152 ±0 .020 ±0 .098 ±0 .062 ±0 .008 B 1.796 1.850 1.592 1.766 1.808 1.826 ±0 .104 ±0.098 ±0 .008 ±0 .032 ±0 .044 ±0 .014 C 1.565 1.895 1.454 1.772 1.898 1.841 ±0 .044 ±0 .158 ±0 .008 ±0 .002 ±0 .002 ±0 .002 D 1.579 1.778 1.522 1.740 1.778 1.704 ±0 .074 ±0 .062 ±0 .020 ±0 .086 ±0 .074 ±0 .038 E 1.304 1.295 1.247 1.757 1.481 1.475 ±0 .002 ±0 .002 ±0 .002 ±0 .002 ±0 .002 ±0 .092 Values represent peroxide value (milliequiv. peroxide/kg oil) . 2 Values present conjugated diene hydroperoxide (umol/g oil) . 3 Values present thiobarbituric acid reactive substances ( T B A R S ) (umol/g oil) . 4 A=100% flaxseed o i l ; B=66.7% flaxseed o i l + 33.3% butter o i l ; C=50% flaxseed o i l + 50% butter o i l ; D=33.3% flaxseed o i l + 66.7% butter o i l ; E=100% butter o i l . 97 Table 16. Quadratic regression of butter-flaxseed oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8, 16 and 24 h thermal treatment at 150°C and 180°C. 150°C 180°C Coefficients1 8h 16 h 24 h Sh 16 h 24 h PV 1 Po (intercept) 4.255 2.978 2.343 2.106 2.180 2.571 Pi 0.0606 -2.2620*** -0.637 -1.345*** -0.571 -0.805** P2 -0.0471 1.7580*** 0.495 1.045*** 0.444 0.626** Pl2 1.151 -2.172** -1.181 -0.0960 -1.714* -0.535 CCl -1.480 2.797** 1.519 0.123 1.205* 0.689 0C2 -0.894 1.690** 0.918 0.0746 • 1.332* 0.416 R 2 0.305 0.772*** 0.228 0.734** 0.407 0.629** CDH Po (intercept) 3.458 3.417 4.542 3.024 4.515 5.153 Pi -0.266 0.427 1.369** 0.793*** 1.559** 1.966*** P2 0.206 -0.331 -1.064** -0.617*** -1.214** -1.532*** Pl2 -0.366 0.277 -0.058 -0.252 0.720 0.868* CCl 0.471 -0.355 0.075 0.325 -0.926 -1.117* oc2 0.285 -0.214 0.045 0.197 -0.560 -0.675* R 2 0.241 0.166 0.711** 0.808*** 0.669** 0.881*** MDA Po (intercept) 1.740 1.926 1.572 1.830 1.926 1.818 Pi 0.338 0.374** 0.249*** 0.049 0.310*** 0.297*** P2 -0.263 -0.291** -0.194*** -0.038 -0.241*** -0.231*** Pl2 0.152 0.360* 0.114 0.062 0.237* 0.127 Ctl -0.195 -0.463* -0.146 -0.080 -0.305* -0.163 a 2 -0.118 -0.280* -0.089 -0.048 -0.184* -0.098 R 2 0.299 0.589* 0.793*** 0.048 0.733** 0.844*** Values represent coefficients of quadratic equation after 150°C and 180°C heating. *: P<0.05; **: P<0.01; ***: P<0.001. Po = intercept; Pi = coefficient for flaxseed o i l at the first order; p 2 = coefficient for butter o i l at the first order; p ) 2 = coefficient for the interaction between oils; a! = coefficient for flaxseed oil at the second order; oc2 = coefficient for butter oil at the second order. 98 450.00 _ 400.00 o 350.00 & 300.00 t j « 250.00 o - i 200.00 o c =^ 150.00 co s 100.00 H DC 50.00 0.00 8 h incubation Figure 14. Comparison of experimental and hypothetical values on the changes in linolenic acid (C18:3co3) of butter-flaxseed o i l blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of linolenic acid, • = hypothetical value of linolenic acid, A = 100% flaxseed oi l , B = 66.7% flaxseed oi l + 33.3% butter o i l , C = 50% flaxseed oi l + 50% butter o i l , D = 33.3% flaxseed o i l + 66.7% butter o i l , E = 100% butter oi l . 450.00 400.00 350.00 S 300.00 O ^ 250.00 o o 200.00 'c •§ 150.00 to 100.00 TJ 50.00 DC 0.00 8 h incubation Figure 15. Comparison of experimental and hypothetical values on the changes in linolenic acid (C18:3co3) of butter-flaxseed o i l blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of linolenic acid, • = hypothetical value of linolenic acid, A = 100% flaxseed oi l , B = 66.7% flaxseed o i l + 33.3% butter o i l , C = 50% flaxseed o i l + 50% butter o i l , D = 33.3% flaxseed o i l + 66.7% butter o i l , E = 100% butter o i l . 99 120.00 = 100.00 o -CO CO £ 80.00 •g o tc .2 60.00 <D O " i 40.00 •g w cr 20.00 0.00 I ^ B C D E A B C D E A B C D E 8 h incubation 16 h incubation 24 h incubation Figure 16. Comparison of experimental and hypothetical values on the changes in linoleic acid (C18:2co6) of butter-flaxseed oi l blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of linoleic acid, • = hypothetical value of linoleic acid, A = 100% flaxseed oi l , B = 66.7% flaxseed oi l + 33.3% butter o i l , C = 50% flaxseed oi l + 50% butter o i l , D = 33.3% flaxseed oi l + 66.7% butter o i l , E = 100% butter oi l . 120.00 -j o 100.00 -_o> "co E. 80.00 -T3 'o « o 60.00 -'<D o £ "to 40.00 -"O t/) CD DC 20.00 -0.00 -8 h incubation Figure 17. Comparison of experimental and hypothetical values on the changes in linoleic acid (C18:2co6) of butter-flaxseed oi l blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of linoleic acid, • = hypothetical value of linoleic acid, A = 100% flaxseed oi l , B = 66.7% flaxseed oi l + 33.3% butter o i l , C = 50% flaxseed oi l + 50% butter o i l , D = 33.3% flaxseed oi l + 66.7% butter o i l , E = 100% butter oi l . 100 250.00 i o •5" E 200.00 o 150.00 CD « 100.00 "(0 T J a> 50.00 rr 0.00 m -T - T - "T A B C D E A B C 8 h incubation 16hincuba I A | B | C I D | E 16 h incubation A | B | C I D | E 24 h incubation Figure 18. Comparison of experimental and hypothetical values on the changes in palmitic acid (C16:0) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 150°C. • = experimental value of palmitic acid, • = hypothetical value of palmitic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butter oil, E = 100% butter oil. 250.00 o 200.00 E T J O cd o 150.00 E « 100.00 Q. Hi 3 T J cu 50.00 rr 0.00 B C D E A B C D E A B C D E 8 h incubation 16 h incubation 24 h incubation ll)E A B C 24 h incul Figure 19. Comparison of experimental and hypothetical values on the changes in palmitic acid (C16:0) of butter-flaxseed oil blends with different blending ratios after 8 h, 16 h, and 24 h thermal treatment at 180°C. • = experimental value of palmitic acid, • = hypothetical value of palmitic acid, A = 100% flaxseed oil, B = 66.7% flaxseed oil + 33.3% butter oil, C = 50% flaxseed oil + 50% butter oil, D = 33.3% flaxseed oil + 66.7% butte oil, E = 100% butter oil. 101 2.500 Incubation time (h) Figure 20. The content of malonaldehyde ( M D A ) for butter-flaxseed oi l blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% flaxseed oi l , • 66.7% flaxseed o i l and 33.3% butter o i l , • 50% flaxseed oi l and 50% butter o i l , X 33.3% flaxseed o i l and 66.7% butter o i l , x 100% butter o i l . 2.500 n Incubation time (h) Figure 21. The content of malonaldehyde ( M D A ) for butter-flaxseed oi l blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% flaxseed o i l , • 66.7% flaxseed o i l and 33.3% butter o i l , • 50% flaxseed oi l and 50% butter oi l , X 33.3% flaxseed o i l and 66.7% butter o i l , x 100% butter o i l . 102 7.000 -i Incubation time (h) Figure 22. The content of conjugated diene hydroperoxides ( C D H ) for butter-flaxseed oi l blends following 8 h, 16 h, and 24 h thermal incubation at 150°C. • 100% flaxseed o i l , • 66.7% flaxseed oil and 33.3% butter o i l , • 50% flaxseed oi l and 50% butter o i l , X 33.3% flaxseed oi l and 66.7% butter oi l , x 100% butter o i l . Incubation time (h) Figure 23. The content of conjugated diene hydroperoxides ( C D H ) for butter-flaxseed o i l blends following 8 h, 16 h, and 24 h thermal incubation at 180°C. • 100% flaxseed oi l , • 66.7% flaxseed oi l and 33.3% butter o i l , • 50% flaxseed o i l and 50% butter o i l , X 33.3% flaxseed o i l and 66.7% butter o i l , x 100% butter o i l . 103 Chapter 3. Blending of Flaxseed and Fish oils with Butter oil, with normalization of cholesterol content. Introduction The information obtained from Experiments 1 and 2 brought forth the fact that the relative differences between flaxseed and fish oils may not be combined to only differences in individual fatty acid composition, but also the sterol content. The difference in cholesterol content between fish oil and flaxseed oil was not considered in the previous experimental study. As a matter of fact, fish oil used in Experiment 1 contained 721.8mg cholesterol/lOOg of fish oil, while, flaxseed oil used in Experiment 2 was of course cholesterol free. The result of the correlation test performed for both butter-fish oil blends in Chapter 1 and butter-flaxseed oil blends in Chapter 2, indicated that the lipid peroxidation indices were also related to the oil type. Therefore, the question of whether cholesterol in fish oil contributed to a reduced oxidation of unsaturated fatty acid oxidation needs to be answered, since fish oil represents a good source of cholesterol. Although it is controversial, it is reasonable to suspect that a potential antioxidant role of cholesterol may exist from the standpoint that sterols are also susceptible to oxidation reactions (Parasassi et al., 1995). If this is true, then it could be expected that cholesterol would compete with polyunsaturated fatty acids (PUFAs) for the oxygen required for oxidation reactions. Smith (1991) proposed that a protective effect of cholesterol existed in such a manner that gave rise to the various oxysterols. Rats fed diets with elevated cholesterol containing diets have been found to have an improved 104 resistance to erythrocyte prooxidant stress (Bereza et al., 1987; Yuan et al., 1998). A similar finding was also obtained from healthy human beings (Clemens et al., 1987). The potential effect of cholesterol on the oxidation of oil blends induced by thermal treatment therefore needs to be considered and evaluated. The purpose of the following experiment was to determine if cholesterol is a contributing factor to the susceptibility of labile PUFA to thermal oxidation To accomplish this objective, the effect of the presence of cholesterol on the co-3 fatty acid oxidation and associated elevated oxysterol level of oil blends was investigated. Cholesterol was added to flaxseed oil in an effort to match the level of cholesterol present in fish oil, and similar heating studies performed in Chapter 2 were conducted. In the following experiments reported in Chapter 3a and 3b the effect of cholesterol on the oxidative stability of different butter-flaxseed oil blends was examined along with quantitation of the potential oxysterols derived from cholesterol oxidation. 105 Chapter 3a. Effect of blending on fatty acid profiles and thermal oxidative stability. Materials and Methods Oil sources The source of butter oil and flaxseed oil used in this experiment were the same as that reported in Chapter 2, with the only exception that cholesterol was added into the flaxseed oil at a level of 750 mg per 100 g oil. This concentration of cholesterol was identical to the endogenous amount present in the fish oil (Experiment 1). The flaxseed oil with additional cholesterol was called thereafter cholesterol modified flaxseed oil. Both bulk oils were stored in a dark room held at -35 °C in suitable portions, until experiments started, in order to avoid the freeze-thaw recycle during the course of the study. Block Heater The block heater used in this Chapter was the same as described as Chapter 1. Other Chemicals All of the chemicals introduced into this chapter were those that had been mentioned in Chapter 2. U.S.P. pure cholesterol (ICN Biochemicals, U.S.A) was added to this model. Experimental Design Again, the Echip 6.04 (Echip, Inc., Hockessin, DE) program (Wheeler, 1993) was used to execute the experimental design for selecting the different oil blends. Calculation 106 equations and statistical analysis were also provided by this program (Irwandi, et al., 1999). L Constrained Mixture Design The oil blends were prepared using a constrained mixture design for two variables in this chapter with butter oil and cholesterol modified flaxseed oil combinations as described in Chapter 2 (Table 10). The sum of variables in the mixture design totaled 100 % (or 1), the low limit of each component was set at 0 % (or 0), and the high limit was set at 100 % (or 1), thus: ZXi= 100% (or 1.0) Consequently, a quadratic equation, which predicted the polyunsaturated to saturated fatty acid ratio, or co-3 to co-6 fatty acid ratio was also come up with as follows: Y = Po +Pl*XModified Flaxseed+P2*XB utter + P l2*X Modified Flaxseed • Xsutter + OCi »X" Modified Flaxseed 2 + 0C2 - X Butter, in which, Y = polyunsaturated to saturated fatty acid ratio, or, co-3 to co-6 fatty acid ratio; -^Modified Flaxseed + Xsutter = 1 (or 100%); XModified Flaxseed = composition of cholesterol modified flaxseed oil in percentage; Xsutter = composition of butter oil in percentage; and Po intercept; Pi = coefficient for cholesterol modified flaxseed oil at the first order; P 2= coefficient for butter oil at the first order; P12 = coefficient for the interaction between oils; OCi = coefficient for cholesterol modified flaxseed oil at the second order; a 2 = coefficient for butter oil at the second order. 107 2. Formulation of Oil Blends with cholesterol modified flaxseed oil and butter oil Similarly, cholesterol modified flaxseed oil and butter oil were used for the cholesterol modified butter-flaxseed oil blends. As listed in Table 10, a total of 13 blended samples, each representing a different blending ratio, were tested. 3. Heating studies Identical procedures to that used in the butter-flaxseed oil blends in Experiment 2 were applied to the cholesterol modified butter-flaxseed oil blends in all analyses. Analysis of oxidation of oil Blends The experimental protocol as follows for each analysis was conducted as described in Chapter 1: L Thiobarbituric acid reactive substances (TBARS) p.34 2. Conjugated Diene (CDH) p.34 3. Peroxide value (PV) p.35 4. Fatty acid quantification p.36 a. Fatty acid calibration curve p.36 b. Sample preparation p.37 c. Gas chromatography p.37 d. Fatty acid profile of native oils p.38 108 The fatty acid profile of the cholesterol modified flaxseed oil was technically regarded as the same for that of the native flaxseed oil (Table 2). No detectable quantities of plant sterol were found in the flaxseed oil used in this study. Statistical Analysis of Data Echip 6.04 and SPSS (SPSS for Windows, version 7.5, SPSS Inc.) were also both used for statistical analysis in this chapter. 109 Results Fatty acid profiles of blended oils with cholesterol modified flaxseed oil and butter oil The addition of cholesterol to the native flaxseed oil did not change the fatty acid profile of the native oil (Table 2). Similarly, the fatty acid profiles of the blended butter-flaxseed oil combination containing cholesterol were identical to the butter-flaxseed oil blend described in Chapter 2 (Table 11). Evaluation of a quadratic model used to predict the optimal combinations of flaxseed and butter oil blends containing cholesterol, assessed by the P/S and co3/co6 fatty acid ratio Both P/S and co3/co6 fatty acid ratios were chosen as end-point measures to assess the effect of blending on the fatty acid profile change. Flaxseed oil and butter oil both contributed to the specific changes in the fatty acid profiles of samples observed with the cholesterol modified butter-flaxseed oil blends. The coefficients of the quadratic regression model, used to determine the optimal proportions of cholesterol modified flaxseed oil with butter oil, are given in Tables 17 and 18. The significant (R2>0.980, p<0.001) result obtained for all coefficients derived from the P/S fatty acid ratio for predicting fatty acid modification after thermal treatment is given in Table 17. Similar significant (R2>0.963, p<0.001) coefficients were obtained for the co3/co6 fatty acid ratio parameter used to define changes in co-3 fatty acid concentration following heat processing (Table 18). The values of (31, (32, (312, al , and a2 produced from the P/S ratio and co3/co6 ratio indicators of co-3 fatty acid content of cholesterol modified butter-flaxseed oil blends were similar to the values obtained for the butter-flaxseed oil blends in 110 Experiment 2 (Tables 17 and 18). Consequently, the regression curves plotted from data listed in Tables 17 and 18 show similar trends for the change in the 0)3/0)6 ratio with blending. Optimization of co3/co6 ratio values produced from the various treatments revealed that the combination of 76% flaxseed oil with 24% butter oil produced the highest 0)3/0)6 ratio in the cholesterol modified butter-flaxseed oil blends. This result was slightly, but not significantly, different from butter-flaxseed oil blends that did not contain cholesterol. Depletion of PUFAs during thermal incubation The PUFA profiles for the different butter-cholesterol modified flaxseed oil blends following thermal treatments at 150°C and 180°C, respectively, are given in Table 19. During thermal treatment of oils at both temperatures for 24 hours, it was observed that significant depletions occurred in both the P/S and co3/u)6 ratios from all butter-flaxseed oil blends, supplemented with cholesterol (Table 19). Comparing these results with the butter-flaxseed oil blends which did not contain the added cholesterol, revealed that the addition of cholesterol had an effect on the oxidation of the PUFAs present in flaxseed oil was dependant on how the data was expressed. The indicator of fatty acid composition in percentage (of total fatty acids, w/w) produced questionable results, as to the decreasing trend observed for either linolenic acid or linoleic acid. On the other hand, expressing the data in terms of absolute amounts of fatty acids, demonstrated that the presence of sterol was indeed effective at protecting co-3 PUFAs from thermal oxidation at high temperatures. i l l The presence of cholesterol in the oil resulted in a significantly (P<0.05) greater • retention the linolenic acid during the first 8 h of heating (Figure 25), when expressed in absolute amounts. A similar trend (Figure 24) was seen for the co-6 fatty acid, linoleic acid, which was statistically significant (P<0.05) after 24 h thermal treatment. Higher temperature (180°C) heating revealed that both co-3 (Figure 27) and co-6 fatty acids (Figure 26) were protected from oxidation when in the presence of cholesterol. This was particularly evident with the shorter duration (8 h) of heat exposure of the cholesterol modified flaxseed oil. From the ANOVA results, it was confirmed that time of heating was a significant factor (P<0.001) affecting the retention of linolenic acid as well as linoleic acid. However, a different result was obtained with the addition of cholesterol to the highly unsaturated oil blend, whereby temperature was not a significant factor affecting PUFA depletion when the result form 150°C was compared that from 180°C. More specifically, the presence of cholesterol in the oil blend proved to be effective at protecting residual linolenic acid, but not linoleic acid from thermal oxidation or degradation. The interaction between added cholesterol and time or temperature was found for both PUFAs (P<0.05). Therefore, the presence of cholesterol was shown to be effective at reducing the effect of temperature on the depletion of different PUFAs in this study. Effect of blending strategy on the depletion of PUFAs among butter-cholesterol modified flaxseed oil blends during thermal treatments The comparison of experimental and hypothetical values for fatty acid retention was also made in this study. As noted, the presence of cholesterol provided a protective effect for PUFAs during thermal oxidation. This observation was established by the fact 112 that experimental values of residual linolenic acid or linoleic acid were not greater than hypothetical values. Information of lipid oxidation from chemical analyses The addition of cholesterol resulted in a significant improvement for the generation of peroxide (PV) or CDH (conjugated diene hydroperoxides) in cholesterol modified butter-flaxseed oil blends (Table 20), compared to butter-flaxseed oil blends (Table 15). MDA content was also considerably less in the oil blends containing cholesterol. Table 21 shows the effect of adding cholesterol to the modified flaxseed oil and butter oil blend on PV, the formation CDH and TBARS (e.g. MDA) following thermal treatment at 150°C and 180°C, up to 24 hours. Both the cholesterol modified flaxseed oil and the butter oil were susceptible to the formation of CDH and TBARS, but not PV, when heated at both temperatures (Table 21). Coefficients of determinations (R ) and derived quadratic regression coefficients generated to describe PV, CDH and MDA of heated oils are given in Table 21. The coefficients of determinations (R2) derived for MDA content from the cholesterol modified butter-flaxseed oil blends were superior to those derived from PV or CDH content. Moreover, these statistics were also improved for oil blends containing cholesterol, when compared to the cholesterol free butter-flaxseed oil blends (e.g. Experiment 2). This result suggests that the MDA content is a better indicator for evaluating the oxidative status of cholesterol supplemented butter-flaxseed oil blends in the quadratic model used in this study. Moreover, it is noteworthy that the greater comparative stability of secondary lipid oxidation products exceeded that of primary lipid oxidation products in thermal processed oils. 113 Analysis of data by ANOVA showed that the PV was significantly affected by temperature and time (P<0.05), while the formation of CDH was affected by time and blending ratio (P<0.05). The blending ratio (P<0.05) was the only factor that significantly influenced TBARS (e.g. MDA). In heated oil blends, a significant time-temperature treatment interaction was obtained for PV (P<0.001). In comparison, two interactions were noted between blending ratio x temperature, and between temperature x time (P<0.05) for TBARS. Similarly, interactions involving temperature x time and blending ratio x time (P<0.05) was derived for CDH. The addition of cholesterol to the butter-flaxseed oil blend was a significant factor affecting TBARS and POV (Pearson coefficient r=-0.361, P<0.001) and POV (Pearson coefficient r=-0.950, P<0.001). The negative coefficients for both measured parameters indicated that the addition of cholesterol lowered PV and reduced TBARS during thermal treatment of the oil blends. Moreover, a 2-way interaction between cholesterol and time of heating was also shown to affect PV in butter-cholesterol modified flaxseed oil blends (p<0.001). The duration of heating time was the only main effect that influenced the formation of CDH (p<0.001). As interaction between the time of heating x cholesterol content and time of heating x temperature x cholesterol were also factors that influenced the formation of CDH (p<0.001, p<0.05, respectively) in thermally treated oils. 114 Discussion The addition of cholesterol resulted in an improved preservation of PUFAs such as linolenic acid and linoleic acid in flaxseed oils. This finding is similar to the relatively greater retention observed with docohexaenoic acid (C22:6co3) in the butter-fish oil blends, following identical thermal treatments performed in Experiment 1. A similar conclusion for the protective effect of cholesterol against lipid oxidation was made from studies using a phospholipid bilayer model, exposed to irradiation treatment (Parasassi et ah, 1995). The hydroperoxide content produced by ionizing radiation was dramatically reduced by the presence of physiological concentrations of cholesterol. This phenomenon was explained by the steric hindrance that exists between a radical chain reaction and unsaturated fatty acids, due to the co-existence of cholesterol. A similar effect would be expected with the mechanism of thermal oxidation shown in the present study. Hydrogen abstraction by alkoxyl radicals of PUFA is, per se, relatively unimportant compared to peroxyl radical (Gardner 1989). There are several mechanisms that could lead to the formation of peroxyl radical as follows: ROOH + M n + 1 -» ROO* + H + + M n + Equation 11 ROOH + RZ -» RO* + RiOH Equation 12 where: ROOH= hydroperoxide; M n + 1 = metal ion in higher valent; ROO* = peroxyl radical; H + = proton; M n + = metal ion in lower valent; Ri* = alkyl radical; RO* = alkoxyl radical; R i O H = alkyl hydroxide. Fatty acid oxidation is controlled by the hydrogen abstraction from where R / radical is generated (Cosgrove et al., 1987). This reaction is more likely to occur 115 depending on the number of bis-al\y\ic methylenes present in the fatty acid skeletons, since the dissociation energy of the C-H bond (BDE) of to-allylic methylene is about 75 kcal/mol and lower than that of a mono-allylic methylene hydrogen (C-H BDE=88 kcal/mol) for a monounsaturated fatty acid, such as oleic acid. This different bond dissociation energy explains the reason why PUFAs, such as linoleic acid, linolenic acid, eicosapentaenoic acid and docosahexaenoic acid are relatively more susceptible to autoxidation reactions than monounsaturated fatty acids (German, 1999). Similar to the radical initiation mechanism reported with irradiation-induced lipid oxidation, thermal-induced lipid oxidation would also occur by involving hydrogen abstraction. However, the addition of cholesterol was found to delay such an oxidation reaction, even in the butter-flaxseed oil blending, which was abundant for both linolenic acid and linoleic acid. As discussed above, cholesterol itself can be oxidized under various conditions, which explains why higher temperatures result in greater cholesterol loss and more generation of oxysterols (Parasassi et al., 1992). It is apparent that cholesterol oxidation requires a greater oxidative stress than that required for the oxidation of PUFA double bonds (Parasassi et al, 1992; Maerker and Jones, 1991). Therefore, once the free radical chain reaction of PUFA is initiated by heating, cholesterol could also be susceptible to this reaction, by reacting with peroxyl radicals that are generated from oxidized PUFA. The lower oxidative reactivity of cholesterol, compared to PUFA would therefore explain the reduced rate of propagation noticed with lipid oxidation. Thus, cholesterol in this manner would exhibit an antioxidant potential under certain thermal oxidation conditions, such as in the case with PUFA present in flaxseed oil. 116 A detailed examination of the oxidation of cholesterol in oil blends will be further studied in Chapter 3 b . 117 Conclusion (3a) In experiment 3 a, a similar regression quadratic model was used to predict the changes in co-3 fatty acid (e.g. linolenic acid) in the butter-cholesterol modified flaxseed oil blends. It was determined that regression curves describing the changes in fatty acid composition of the butter- cholesterol modified flaxseed oil blends were similar to the curves generated from the butter-flaxseed oil blends reported in Chapter 2. It was concluded that the P/S and co3/co6 fatty acid ratios were both effective indicators of the depletion of PUFA in this model. For example, a similar optimized oil blends was obtained by the combination of 76% flaxseed oil with 24% butter oil, which was predicted to produce the highest co3/co6 ratio value. An additional conclusion from this study was that lipid oxidation parameters, such as PV, CDH and MDA were better indicators for oxidation stability of the oil blends in this experiment than Experiment 2. Much lower MDA content was achieved from the butter-cholesterol modified flaxseed oil blends than butter-flaxseed oil blends. The evidence of the improvement on the PUFA (e.g. linoleic acid and linolenic acid) depletion was observed by the normalization of cholesterol level in flaxseed oil used in Experiment 2, equivalent to that in fish oil used in Experiment 1. These results strongly indicate a possible role for cholesterol in stabilizing PUFA-rich vegetable oils that are exposed to thermal processing. 118 Chapter 3b. Characterization and quantitation of cholesterol oxidation products derived from thermally treated oil blends. Materials and Methods Oil sources The cholesterol modified flaxseed oil and butter oils used in this experiment were the same as that reported in Chapter 2a. The fish oil used here was the same as that mentioned in Chapter 1. Samples of the thermally treated butter-fish oil blends and cholesterol modified butter-flaxseed oil blends were stored in the dark room at -35 °C. Other chemicals U.S.P. pure cholesterol (ICN Biochemicals, U.S.A) was used in this experiment. Potassium hydroxide, methanol, diethyl ether, sodium sulfate and pyridine were obtained from Fisher Scientific (Fair Lawn, NJ). Sylon BTZ was used for this experiment (Supelco, Inc., Oakville, ON). Cholesterol and cholesterol oxidation products (COPs) determination Quantification of cholesterol and the identification of cholesterol oxidation products (COPs) were performed for both butter-fish oil blends and cholesterol modified butter-flaxseed oil blends by gas chromatography (GC). Residual cholesterol concentration in the natural oils prior to heating was determined according to a modified derivatized, trimethyl silyl ether cholesterol procedure. In the previous study, GC-MS was used to confirm the individual standard of different COPs (Yuan, 1995). 119 L Cholesterol calibration curve Sterol standards and internal standard were obtained from Stealoids, Inc. (Wilton, NH) with the exception of the cholesterol standard which was obtained from Matreya, Inc. (USA). 5oc-cholestane was used as the internal standard and cholesterol oxidation products (COPs) standards included 3,5-cholestadien-7-one, cholesterol, 5-cholesten-3p\7a-diol, cholestan-5R,6R-epoxy-3R-ol, cholestan-5a,6oc-epoxy-3R-ol, 5-cholesten-3R,7R-diol, 5-cholesten-3R,4R-diol, cholestan-3(3,5a,6R-triol, 5-cholesten-3R-ol-7-one, 5-cholesten-3R,25-diol. Quantification of cholesterol was performed using a response linearity curve derived for trimethyl silyl (TMS) ether cholesterol. A calibration curve of varying cholesterol concentrations and a constant amount of internal standard (IS: 5a-cholestane) was used to confirm the response linearity of trimethyl silyl (TMS) ether cholesterol. The concentrations of cholesterol were prepared to range from O.OlOpg pL"1 to 1.0 ug ' pL"1, with the ratio of cholesterol standardTS ranging from 0.1:1 to 10:1. Response linearity was demonstrated by plotting the area response ratio of cholesterol to the IS, versus the weight ratio of cholesterol to the IS (Appendix I). 2. Sample preparation The samples of butter-fish oil blends and cholesterol modified butter-flaxseed oil blends were removed from the cold room (-35 °C) and placed in a water bath at 25 °C to completely thaw. The melted oil samples were vortexed for 1 minute in order to make the oil system homogenous. 120 Oil (200mg) was sampled for analysis with the addition of 200 pg of the internal standard (IS; 5a-cholestane) and then subjected to a cold saponification with 1 N KOH (in methanol) over night at room temperature. The saponified sample was extracted using diethyl ether. The top layer of diethyl ether was washed with both distilled deionized water and 0.5 N KOH. Nonsaponifiables were dried using anhydrous Na2S0 4 and transferred into a 2 mL GC vial (12 x 32 mm, Alltech Associates, Inc., Deerfield, IL), before reducing the sample volume under a N 2 (Praxair) stream. Samples were dried under vacuum to remove trace amounts of moisture, before solubilization was made in 200 pL dry pyridine. A 100 pL aliquot sample or standard was derivatized with 50 pL Sylon BTZ (Supelco, Inc., Oakville, ON) and the reaction allowed to proceed to completion (30 min) at room temperature (Yuan, 1995). 3_, Gas chromatography After derivatization, all standards and samples were analyzed using a Heliflex AT-1 column (15 m x 0.5 mm i.d., 0.1 pm film thickness; Alltech Associates, Inc., Deerfield, IL), attached to a GC-17A gas chromatography (Shimadzu, Scientific Instruments Inc., Columbia, MD) that was equipped with a flame ionization detector (GC-FTD). A representative chromatogram of derivatized cholesterol oxide standards is shown in Appendix J. Carrier gas, helium (He) was used as described for the fatty acid determination. The injector and detector temperatures were 250 °C and 280 °C, respectively. Column temperature was programmed to increase from 180 °C to 250 °C at 3 °C per minute, reaching a final temperature for 15 minutes. The column flow was set at 1.1 mL • min"1. All chromatograms were integrated and analyzed by Class-VP 121 chromatography data software, version 4.2 (Shimadzu, Scientific Instruments Inc., Columbia, MD). Statistical analysis of data SPSS (SPSS for Windows, version 7.5, SPSS Inc.) was used for the statistical analysis. Results Cholesterol content of oils Butter oil, made from butter as described in Chapter 1, contained 48.8 mg cholesterol per 100 g oil. This amount of cholesterol was low compared to fish oil, which contained 721.8 mg per 100 g oil. In native flaxseed oil, cholesterol was not detectable by GC chromatography. In addition, 700.0 mg cholesterol per 100 g modified flaxseed oil was detected after the flaxseed oil had been supplemented with cholesterol. The cholesterol content in the self-made butter oil was very low, so that fish oil was considered as the primary source of cholesterol for butter-fish oil blends. Similarly, the cholesterol added to the modified flaxseed oil contributed to the primary amount of cholesterol present in the cholesterol modified butter-flaxseed oil blends and was equivalent to that of the butter-fish oil blends. Thermal degradation of cholesterol f, Cholesterol in butter-fish oil blends The reduction of cholesterol content from oil blends incubated at 150°C is listed in Table 22. Only 5-15% of the initial cholesterol concentration present in butter-fish oil blends was lost during the first 8 h incubation at 150°C. Cholesterol content gradually decreased afterwards, and was significantly lower than that of the original amount (p<0.05) after 16 h heating. The cholesterol content of butter-fish oil blends following 24 h incubation decreased to concentrations that were only 20-25% of the initial concentration (p<0.001). This observation confirmed the earlier suggestion that heat not 123 only induced fatty acid oxidation but also contributed to a decrease in the level of cholesterol in the oil blends. As shown in Table 22, butter-fish oil blends with varied blending ratios, lost more than 35% of cholesterol after the first 8 h of heating at 180°C. Similarly, after 16 h heating, the cholesterol present in the oil mixtures was further reduced in all butter-fish oil blends after 24 hours. In butter-fish oil blends, a greater amount of cholesterol was lost at 180°C compared to 150°C heating (Table 22), with less than one-third of cholesterol retained. A T-test analysis showed a significant (P<0.01) difference between residual concentration after 8-hour and 16-hour heating. Heated samples for 24-hour, proved to be the most harsh treatment resulting in cholesterol losses when compared to shorter time periods of heat treatment. The Pearson coefficients obtained from cholesterol content of butter-fish oil blends revealed that the cholesterol content in heated oils was negatively correlated with the heating time (r= -0.75, p<0.05) and temperature (r=-0.53, p<0.05), indicating that the retention of cholesterol in various butter-fish oil blends was both time-dependent and temperature-dependent (Figure 28). 2. Residual cholesterol concentration in butter-cholesterol modified flaxseed oil blends The amount of cholesterol in the butter-flaxseed oil blends was less than half that initially added following the first 8 h heating at 150°C. This degree of thermal induced degradation of cholesterol resulted in much lower concentrations, when compared to butter-fish oil blends that were treated at the same temperature and for the same duration. 124 The finding represented a unique result for the butter-flaxseed oil blends, not seen with the butter-fish oil blends. Higher temperature (180°C) processing resulted in even greater changes to the cholesterol retention (Table 22). The concentration of cholesterol in the cholesterol modified butter-flaxseed oil blends was reduced much more than that observed in the butter-fish oil blends (Figure 29). In addition, the time of heating was a predominant factor affecting the retention of cholesterol in the cholesterol modified butter-flaxseed oil blends, at both 150°C and 180°C. Further analysis of the cholesterol modified butter-flaxseed oil blends revealed a negative correlation between the residual cholesterol content and heating time (r=-0.58) and temperature (r=-0.44). It is noteworthy that smaller Pearson correlation coefficient values were obtained for the butter-fish oil blends and cholesterol modified flaxseed oil, suggesting that the two blends, despite having similar cholesterol concentrations to that of the fish oil, were not equivalent in regard to the apparent extent in which cholesterol remained stable during thermal incubation. 3. Comparison of butter-fish oil blends and butter-cholesterol modified flaxseed oil blends in cholesterol content Results from the correlation analysis of cholesterol content and PUFA content, showed that the cholesterol content of thermally treated oils was significantly (p<0.001) related to the specific oil blend examined in this study. This result indicated a potential relationship between the fatty acid composition of the oil blend and the susceptibility of cholesterol lost during thermal oxidation. The ANOVA results of this data showed that the main treatment variables of oil blend, incubation time and temperature of heating were significant (p<0.001) factors that 125 affected the retention of cholesterol after heating. Significant (p<0.01) interactions of oil type x time, oil type x temperature and time x temperature were also obtained. Quantitation on cholesterol oxidation products (COPs) generated from heated oil blends Quantitative analysis of generated COPs from the various oil blends is presented in Table 23. Typical cholesterol oxidation products such as 3,5-cholestadien-7-one, 5-cholesten-3P,7a-diol, cholestan-5P,6P-epoxy-3P-ol, cholestan-5a,6a-epoxy-3P-ol, 5-cholesten-3P,7P-diol, 5-cholesten-3P,4P-diol, 5-cholesten-3P-ol-7-one, 5-cholesten-3P,25-diol were detectable in both oil blends after 24 h heating, at both temperatures of 150°Cand 180°C. 5-choIesten-3P,7p-diol, 5-cholesten-3P,4P-diol, 5-cholesten-3P-ol-7-one were the three primary cholesterol oxidative products identified from the butter oil after heating at 150°C. Similar cholesterol oxides were detected from fish oil heated at both 150 and 180°C. A similar pattern of cholesterol oxides was observed for cholesterol modified butter-flaxseed oil blends and the butter-fish oil blends. In both oil blends, oxidation of cholesterol was evident after at least 8 h heating at 150°C. However, the higher temperature processing resulted in a greater degradation of COPs during the initial hours of heating. Exceptions were seen with 3,5-cholestadien-7-one, cholestan-5P,6P-epoxy-3P-ol and cholestan-5P,6P-epoxy-3p-ol, which were generated more so after the first 8 h of heating at 180°C. Detection of further, unidentified fragments of cholesterol degradation products was also evident at this time. Further heating led to a disappearance of all COPs, reflecting the chemical instability of the cholesterol oxides. 126 Oxidation of cholesterol in butter-fish oil blends resulted in the formation of different cholesterol oxides than that observed in the cholesterol modified flaxseed oil blends. For example, the formation of 3,5-cholestadien-7-one was higher at 180°C than at 150°C, and the generation of this oxysterin increased after 16 h at 180°C, before being destroyed. Figure 28 illustrates the effect of heating on the loss of cholesterol content of fish oil blending at both 150 and 180°C, showing that the loss of cholesterol was related to the time duration and temperature. A similar effect was found with the heating of cholesterol modified flaxseed oil (Figure 29). Table 22 shows the relative amount of cholesterol loss, however, compared to Table 23, the loss was not equivalent to the amount of cholesterol oxides generated during heating. 127 Discussion Cholesterol plays a key role in the development of atherosclerosis, which is a major health problem in industrialized countries (Stampler 1979). Indisputable evidence has implicated elevated blood cholesterol to be a major cause of atherosclerosis in both animal (Yuan et al., 1997) and human (Kuller, 1997) studies. In the past two decades, scientists have also reported a relationship between oxidative modified cholesterol and the onset of coronary heart disease. Human LDL is the major cholesterol carrier in blood circulation and the oxidation of human LDL is involved in the pathogenesis of atherosclerosis. Oxidized LDL is responsible for the formation of macrophagic "foam cells", which results in the early lesion of atheroma. This process acts according to three different mechanisms including modifying the scavenger receptor pathway of macrophages, recruitment of monocytes and inhibiting the mobility of resident macrophages (Steinberg et al., 1989), and eventually generate the fibrous plaques in the blood vessel. Cholesterol autoxidation occurs at the allylic C-7 and A5-olefin position of the B-ring in most cases, albeit, the minor A ring and side-chain are also prone to oxidation (Ansari and Smith, 1994). Generally, cholesterol autoxidation proceeds by a free radical processes similar to the antioxidation that occurs with PUFA (Smith, 1981c). Cholesterol autoxidation yields epimeric 7-hydroxycholesterols and related dehydrated 7-ketocholesterol, as well as epimeric 7-hydroperoxide as primary autoxidation products. In another possible pathway, cholesterol epoxidation yields epimeric 5,6-epoxides as the primary oxidation product and cholestane triol as a hydrated product. By using major cholesterol oxide products, such as 7-hydroxylcholesterol, epoxicholesterol, as indicators 128 of cholesterol oxidation, the major reaction pathways during heating of cholesterol has been shown to include free radical chain reaction, reduction, dehydration, dehydrogenation, and epoxidation (Chien et al., 1998). The amount of cholesterol lost in this retention is not equal to the amount of COPs formed, probably because cholesterol is also degraded to other compounds that are not easily measured. Animal studies have showed that feeding a high cholesterol diet will enhance the cholesterol level in Japanese quail and result in aortic plaque formation and detectable amount of cholesterol oxides (Yuan et al., 1998). Dietary consumption of cholesterol oxides in food, for example, deep frying fast food, also can contribute to an increase in the plasma cholesterol oxide level (Lake and Scholes 1997). The usual parameters that influence the autoxidation of cholesterol, including catalysts, initiation by heat (especially >50°C) and the presence of oxygen (Kim and Nawar, 1993) are common to many food processing methods. Food processing therefore may contribute to an increased presence of cholesterol oxides into food; for example, spray dried egg yolk powder (Li et al, 1996a), butter oil, frying and cooking animal oil, meat products, such as sausage etc., have been reported to contain oxidized cholesterol (Yan 1999). In addition to heating, benzoyl peroxide bleaching can also lead to COPs by the generation of 7oc-cholestenone and 7(3-cholestone, both of which are associated with peroxides generated from benzoyl peroxides. It is therefore reasonable to propose that fatty acid peroxides generated from lipid oxidation may also partially contribute to the formation of cholesterol oxidation products. Li et al. (1996b) reported that the total sterol oxide levels in fish oil were higher than sterol oxides (e.g. plant sterols) present in vegetable oil, and that the addition of 129 tocopherol was not effective in reducing oxidation. These results suggested that there was an interaction between fatty acid composition, especially PUFAs, and cholesterol oxidation existed. Osada et al. (1993), also found that cholesterol was quite stable when it was heated alone, but stability was lost when heated simultaneously with PUFAs (for example, fish oil and flaxseed oil). This finding further demonstrated that fatty acid oxidation precedes cholesterol oxidation. Moreover, the fatty acid composition of any given diet when exposed to cholesterol can also affect peroxidation of cholesterol to oxysterol which may lead to subsequent accumulation in aortic tissue (Bhadra 1993). Increased dietary antioxidant intake of domestic animals has been shown to prevent the formation of cholesterol oxides in muscle food (Monahan et al., 1992; Paniangvait et al., 1995). In this current study, the formation of COPs in oils was found to increase with higher temperature processing and longer heating periods, resulting in a decrease in cholesterol content. Statistical analysis revealed that the duration of heating oils was a significant factor affecting the cholesterol content of both oil blends (p<0.05). This result confirms that cholesterol degradation to oxides, or other unidentified fragments during heating, is the source for the marked loss of cholesterol in these oil products. This is a very important feature for the snack food processing industry, especially those that involve frying. Practically, any PUFA rich frying oil, if used for several cycles would lead to the accumulation of cholesterol oxides from animal products that were used in the frying process. It should be also pointed out that the loss of cholesterol was not equal to the detected amount of cholesterol oxides generated. Similar findings are also available from the work of Yan and White (1990), using cholesterol enriched lard at 180°C. The 130 fact that COPs formed was not equal to the amount of cholesterol lost during heating could be interpreted to indicate that some COPs formed during heating, are not detectable. For example, 7-hydroperoxide cholesterol (Yan and White, 1990, Chen et al, 1994) and other forms of cholesterol oxidative products are volatile (Chen et al., 1994). In this experiment, several cholesterol oxides were found to increase in oil blends with temperature and heating duration. For example, 7-keto-cholesterol consistently showed increases with temperature (p<0.05) in both oil blends. This result agreed with the work of Li et al. (1996a), who showed that cholesterol oxides dramatically increased during the cooking of egg yolk powder and that 7-keto-cholesterol was present during higher temperature processing. The formation of epimeric 7-hydroxycholesterol, as a decomposition product of 7-hydroperoxides also results because 7-hydroxycholesterols are easily dehydrated in the absence of water at elevated temperature (Ohshima et al., 1993). A relatively higher generation of 7-keto-cholesterol than 7-hydroxy-cholesterol during heating, likely reflected the heat stability of 7-keto-cholesterol. Though the formation of 7-keto-cholesterol and 7-hydroxyl cholesterol involves the initial formation of 7-hydroperoxide cholesterol, it is likely that 7-hydroperoxide cholesterol was not detected due to its thermal instability. Park and Addis (1985), found that the formation of 7-ketocholesterol was nearly linear with heating time, reaching about 10% of the initial cholesterol content after 376 hour heating at 155°C, before declining. In this work, the formation of 7-ketocholesterol was found to follow a similar pattern associated with a much shorter heating time. 25-hydroxycholesterol was the only side-chain oxidation product of cholesterol that was detected in both fish oil and the modified flaxseed oil supplied with cholesterol. 131 Together with cholestene triol, 25-hydroxyl-cholesterol is believed to be the most atherogenic oxysterol derivative (Imai et al., 1976; Pend et al, 1979; Blankenship et al., 1991). Though this side-chain oxidation product was detectable in this study, the relative amount was indeed,very low. It was also interesting to note that cholesterol-triol was detectable only in the butter-flaxseed oil blends, suggesting that more epoxidation occurred in this blending. Supporting evidence also comes from the higher level of epoxyl cholesterol derived from the flaxseed oil-cholesterol mixture compared to the fish oil. This difference indicates that fatty acid composition can affect the specific products of cholesterol oxidation during heating. Long chain fatty acids were found to accelerate the formation of cholesterol oxides in addition to storage and heating ( Li et al., 1996a). The accumulation of COPs in modified flaxseed oil apparently supports this finding, since flaxseed oil contains a high content of PUFAs, such as linolenic acid and linoleic acid. In my study, it can be seen (Figures 25 and 27) that flaxseed oil when supplied with cholesterol in oil blends, could indeed retard the losses of linolenic acid more so than in oil blends not containing cholesterol (e.g. the native plant oil). This finding clearly indicates that the presence of COPs was indeed related to fatty acid oxidation. Kim and Nawar (1991) reported that autoxidation of cholesterol is exaggerated when simultaneously present with triacylglycerols. This result suggests that cholesterol is oxidizable in the presence of lipids and oxygen, even under ambient temperatures. This finding was supported by the work of Park and Addis (1987), who previously noted that beef was less susceptible to fatty acid and cholesterol oxidation than turkey. It should be noted that turkey lipids are much more unsaturated than beef lipids and contain relatively 132 low vitamin E, which should favor the oxidation of cholesterol (Allen and Foegeding (1981). Interaction between fatty acids and cholesterol when in the presence of oxygen could possibly affect the stability of both fatty acid and cholesterol. In this study, the fatty acids present in the oil blends were also found to be more susceptible to the influence of oxidation (peroxide value, formation of CDH and TBARS) than similar fatty acids present in the cholesterol modified butter-flaxseed oil blends. This finding is strong evidence to indicate that cholesterol can protect highly unsaturated fatty acids from thermal induced oxidation, which otherwise can be propagated from the reaction of trace amounts of fatty acid hydroperoxide which exists in oil (Coupland and McClements, 1996). Smith (1981) also pointed out that hydroperoxides generated from PUFAs during oxidation may be necessary to initiate cholesterol oxidation as cholesterol contains one A5-double bond, which is prone to be oxidized. The generation of free radicals during thermal processing is illustrated by the following equation: LOOH-^LO*+ HO* Equation 13 After the generation of alkoxyl radical (LO*) and hydroxyl radical (HO*), a hydrogen atom is abstracted from a PUFA or cholesterol. Thus, in turn, cholesterol oxidation is induced by the initiation of free radicals. This is in agreement with the mechanism of cholesterol oxidation. Zarev et al. (1999) showed from evidence of hydroxyl radical-induced low-density lipoprotein oxidation, that this mechanism of action was effective but less so than copper ion induced Fenton reaction oxidation. In a heterogeneous model, water-soluble peroxyl radical (RO") was shown to initiate the oxidation of cholesterol and produced 7-keto-cholesterol as primary oxidative products, plus minor isomeric 7-hydroxycholesterol (Nielsen et al., 1996). In the present study, it 133 was confirmed that 7-keto cholesterol, a cholesterol oxidation product was formed in free radical induced cholesterol oxidation, following a thermal treatment of flaxseed oil containing cholesterol. Since cholesterol can be oxidized by a free radical mechanism, it can also be suggested that cholesterol possesses antioxidant capacity to a certain extent. As mentioned at an earlier stage of this thesis, an antioxidant can represent any molecule that can prolong lipid oxidation at a lower concentration. In this study, cholesterol loading was less than 1% of lipid. However, it is obvious that this is not a practical approach to prevent lipid oxidation, simply because there is extensive scientific evidence to show that the hypercholesterolemia and hyperoxysterolemia can lead to atherosclerosis (Westhuyzen, 1997; Ballantyne, 1998; Devaraj and Jialal, 1996). For the sake of preventing atherosclerosis related to LDL oxidation, it would be wise to use natural antioxidant supplements, for example, vitamin E and C, polyphenols etc. (Illingworth, 1993), that are not directly involved in the etiology of these diseases. Notwithstanding this however, cholesterol which co-exists with PUFA in formulated and native food products therefore could have a role in protecting essential fatty acid from oxidation reactions. 134 Conclusion (3b) In this experiment, the decrease in cholesterol content due to thermal heating of the cholesterol modified flaxseed oil was observed to correspond to a reduction in the susceptibility of omega-3 fatty acid depletion. From the results of this experiment, where higher temperature (180°C) caused greater cholesterol loss than a lower temperature (150°C), it can be concluded that both temperature of processing as well as duration of heating were significant independent variables that influenced the extent of cholesterol oxidation. In addition, it was also observed that the generation of specific COPs that were derived from the heating of the oil mixture also varied with the fatty acid composition of the oil. For example, a thermal treatment of oils for up to 24 hrs.at both 150°C and 180°C resulted in significant quantities of the cholesterol oxysterols. The fact that the relative concentration of oxysterols was also found to be significantly greater when cholesterol was heated in flaxseed oil, compared to fish oil, indicates that oils containing high concentrations of omega-3 PUFA (i.e. flaxseed oil) is important for the generation of cholesterol oxidation products than the degree of unsaturation of individual omega-3 fatty acids (i.e. EPA and DHA). This was particularly evident with the generation of cholesterol triol. 135 Table 17. Quadratic regression of butter-cholesterol modified flaxseed oil blends and the retention of P/S ratio at 150°C and 180°C. 1 Coefficients2 Time (h) 150°C 180°C 8 16 24 8 16 24 Po (intercept)3 0.156 0.261 0.115 0.122 0.0703 0.0931 P i 2.669*" 2.316*** 2.669*** 2.676*** 2.117*** 1.974*** P2 -2.075*** -1.800*** -2.175"* -2.080*** -1.646*** -1.535*** Pl2 -2.680*** -2.175*** -2.779*** -2.962*** -2.227*** -2.045*** O i 3.448*" 2.798*** 3.574*** 3.810*** 2.865*** 2.631*" a 2 2.084*** 1.691*" 2.160*** 2.302*** 1.731*** 1.590*** R 2 0.991*** 0.988*** 0.986*** 0.982*** 0.985*** 0.980*** 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***:P<0.001. 3 Po = intercept; Pi = coefficient for cholesterol modified flaxseed o i l at the first order; p 2 = coefficient for butter o i l at the first order; P i 2 = coefficient for the interaction between oils; = coefficient for cholesterol modified flaxseed oil at the second order; a2 = coefficient for butter oil at the second order. Table 18. Quadratic regression of butter-cholesterol modified flaxseed oil blends and the retention of co3/to6 ratio at 150°C and 180°C. a Temperature (°C) Coefficients" Time (h) 150 180 8 16 24 8 16 24 Po (intercept) 4.291 4.196 3.897 3.865 3.741 3.676 P i 2.911*** 2.806*** 2.747*" 2.852*** 2.716*** 2.694*** P2 -2.263*** -2.181*** -2.135*** -1.879*** -1.897*** -1.911"* Pl2 2.449*** 2.443*** 2.071*** 2.234*** 2.250*** 2.069*** CCl -3.150*** -3.143*** -2.664*** -2.864*** -2.819*** -2.690*** Ct-2 -1.904*** -1.899*** -1.610*** -1.725*" -1.693*** -1.687*** R 2 0.971*** 0.978*** 0.964*** 0.963*** 0.963*** 0.981*" 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 ***. p<n.001. 3 Po = intercept; PL = coefficient for cholesterol modified flaxseed o i l at the first order; p 2 = coefficient for butter o i l at the first order; P J 2 = coefficient for the interaction between oils; a] = coefficient for cholesterol modified flaxseed oil at the second order; oc2 = coefficient for butter oil at the second order. 136 Table 19. Polyunsaturated fatty acid composition (% of total fatty acids, w/w), P/S ratio and CO3/GO6 ratio of butter-cholesterol modified flaxseed oil blends before and after thermal treatment at 150°C and IS0°C.1 T I M E 150°C 180°C C18:2 C18:3 P/S o3/co6 C18:2 C18:3 P/S CO3/(D6 Oh A 2 13.97 58.04 7.007 4.155 13.97 58.04 7.007 4.155 B 11.08 44.50 1.574 4.016 11.08 44.50 1.574 4.016 C 9.30 36.11 0.897 3.883 9.30 36.11 0.897 3.883 D 7.20 26.25 0.491 3.646 7.20 26.25 0.491 3.646 E 1.68 0.34 0.028 0.202 1.68 0.34 0.028 0.202 8 h A 15.48 52.54 6.071 3.395 15.33 54.31 6.614 3.542 ± 0 . 0 9 ± 1.42 ± 0 . 4 1 0 ± 0 . 1 1 2 ± 0 . 0 3 ± 0 . 8 1 ± 0.288 ± 0.059 B 11.61 39.32 1.879 3.387 11.46 36.13 1.640 3.153 ± 0 . 0 0 ± 0 . 0 0 ±0 .000 ± 0.000 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 C 9.54 28.46 1.013 2.984 9.02 25.40 0.854 2.816 ± 0 . 0 4 ± 1.10 ± 0.046 ± 0 . 1 0 3 ± 0 . 0 3 ± 0 . 4 3 ± 0 . 0 1 6 ± 0.039 D 7.25 19.25 0.554 2.655 6.91 16.99 0.481 2.459 ± 0 . 2 1 ± 1.45 ± 0.048 ± 0 . 1 2 7 ± 0 . 2 1 ± 1.04 ± 0.034 ± 0 . 0 8 1 E 2.65 0.00 0.039 0.000 2.30 0.00 0.033 0.000 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 ± 0 . 0 4 ± 0 . 0 0 ± 0.001 ± 0.000 16 h A 15.58 50.39 5.473 3.234 15.47 49.21 5.180 3.182 + 0.03 ± 0 . 6 6 ± 0 . 1 5 6 ± 0.048 ± 0 . 1 5 ± 0 . 6 3 ± 0 . 1 3 1 ± 0.072 B 11.23 36.51 1.636 3.251 11.13 31.19 1.318 2.803 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 ± 0 . 0 0 + 0.00 ± 0.000 ± 0.000 C 9.50 27.98 0.988 2.945 8.89 23.69 0.784 2.664 ± 0 . 1 4 ± 0 . 8 6 ± 0.046 ± 0.049 ± 0 . 1 8 ± 1.10 ± 0.047 ± 0.069 D 7.16 18.47 0.530 2.579 6.59 14.86 0.416 2.256 ± 0 . 1 9 ± 1.08 ± 0.038 ± 0.084 ± 0 . 1 5 ± 0 . 5 7 ± 0 . 0 1 8 ± 0.058 E 2.58 0.00 0.037 0.000 1.91 0.00 0.027 0.000 ± 0 . 0 6 ± 0 . 0 0 ± 0.001 ± 0.000 ± 0 . 1 1 ± 0 . 0 0 ± 0.002 ± 0.000 24 h A 15.47 51.89 5.863 3.355 15.60 47.58 4.798 3.051 ± 0 . 0 4 ± 0 . 8 5 ± 0.246 ± 0.063 ± 0 . 0 5 ± 1.05 ± 0.265 ± 0.058 B 11.17 33.12 1.420 2.964 10.71 29.66 1.239 2.769 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 . ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 C 9.18 25.27 0.860 2.753 8.75 22.01 0.719 2.516 ± 0.08 ± 0 . 4 2 ± 0 . 0 2 1 ± 0.023 ± 0 . 0 7 ± 0 . 7 1 ± 0.025 ± 0.060 D 6.90 16.94 0.479 2.455 6.46 14.73 0.404 2.280 ± 0 . 1 7 ± 1.04 ± 0.033 ± 0.090 ± 0 . 3 3 ± 1.78 ± 0.056 ± 0 . 1 5 4 E 2.44 0.00 0.035 0.000 1.53 0.00 0.021 0.000 ± 0 . 0 0 ± 0 . 0 0 ± 0.000 ± 0.000 ± 0 . 3 3 ± 0 . 0 0 ± 0.005 ± 0.000 1 Values represent fatty acid composition in percentage of total fatty acids (w/w); P /S : polyunsaturated/saturated fatty acid ratio; to3/co6: omega 3/omega 6 fatty acid ratio. 2 A=100% cholesterol modified flaxseed o i l ; B=66.7% cholesterol modified flaxseed o i l + 33.3% butter o i l ; C=50% cholesterol modified flaxseed o i l + 50% butter o i l ; D=33.3% cholesterol modified flaxseed oi l + 66.7% butter o i l ; E=100% butter o i l . 137 Table 20. Effect of blending of cholesterol modified flaxseed oil and butter oil on the peroxide value (PV) and the formation of conjugated diene hydroperoxides (CDH) and malonaldehyde (MDA) value after 8-, 16- and 24-hour thermal treatment at 150°C and 180°C. 1 5 0 ° C 1 8 0 ° c 8h 16 h 24 h 8h 16 h 24 h P O V 1 A 4 2.63 3.57 3.47 0.98 1.51 1.05 ±0.18 ±0 .14 ±0 .04 ±0 .21 ±0 .11 ±0.07 B 1.40 2.31 4.41 1.40 1.89 1.26 ±0 .10 ±0.15 ±0 .20 ±0 .20 ±0 .15 ±0.07 C 1.47 1.86 3.96 1.09 1.72 1.16 ±0.07 ±0 .39 ±0 .11 ±0 .04 ±0 .74 ±0 .04 D 1.65 3.00 4.97 1.24 1.83 1.38 ±0.13 ±0 .35 ±0 .31 ±0 .27 ±0 .16 ±0 .20 E 2.59 3.12 2.94 1.61 2.87 2.49 ±0 .00 ±0 .04 ±0 .00 ±0 .21 ±1 .19 ±0 .46 C D H 2 A 3.932 4.915 5.702 3.932 4.915 5.702 ±0 .153 ±0 .585 ±0 .316 ±0 .336 ±0 .268 ±0 .489 B 3.376 3.905 4.827 3.376 3.905 4.827 ±0 .244 ±0 .203 ±0 .190 ±0 .122 ±0 .217 ±0 .203 C 3.363 3.892 4.664 3.363 3.892 4.664 ±0 .288 ±0 .479 ±0 .192 ±0 .278 ±0 .125 ±0 .518 D 3.245 3.882 4.400 3.245 3.882 4.400 ±0 .353 ±0 .451 ±0 .423 ±0 .498 ±0 .702 +0.602 E 3.064 3.275 3.620 3.064 3.275 3.620 ±0 .288 ±0 .547 ±0.038 ±0 .067 ±0 .249 ±0 .230 M D A 3 A 1.841 1.937 2.135 1.733 1.946 1.916 ±0.115 ±0 .039 ±0 .064 ±0 .056 ±0 .043 ±0 .069 B 1.592 1.652 1.856 1.736 1.790 1.886 ±0 .134 ±0 .080 ±0 .014 ±0 .080 ±0 .068 ±0 .074 C 1.535 1.589 2.060 1.712 1.811 1.781 ±0 .064 ±0 .030 ±0 .002 ±0 .001 ±0 .149 ±0 .056 D 1.484 1.560 1.727 1.646 1.697 1.675 ±0 .060 ±0 .082 ±0 .122 ±0 .073 ±0 .054 ±0 .072 E 1.418 1.412 1.478 1.427 1.391 1.460 ±0.001 ±0 .026 ±0 .043 ±0 .002 ±0 .073 ±0.001 1 Values represent peroxide value (milliequiv. peroxide/kg oil) . 2 Values present conjugated diene hydroperoxide (umol/g oil) . 3 Values present thiobarbituric acid reactive substances ( T B A R S ) (umol/g oi l ) . 4 A=100% cholesterol modified flaxseed o i l ; B=66.7% cholesterol modified flaxseed o i l + 33.3% butter o i l ; C=50% cholesterol modified flaxseed o i l + 50% butter o i l ; D=33.3% cholesterol modified flaxseed o i l + 66.7% butter o i l ; E=100% butter o i l . 138 Table 21. Quadratic regression of butter-cholesterol modified flaxseed oil blends for peroxide value (PV), conjugated diene hydroperoxides (CDH) content and malonaldehyde (MDA) content after 8-, 16- and 24-hour thermal treatment at 150°C and 180°C. 150°C 180°C Coefficients1 8 h 16 h 24 h 8 h 16 h 24 h P V 1 Po (intercept) 1.134 2.333 5.270 1.146 1.570 1.026 Pi -0.442"* -0.269 0.649 -0.363 -0.882 -0.987*** P2 0.343*** 0.209 -0.505 0.282 0.685 0.767*** Pl2 -1.476*** -1.051 2.022** -0.188 -0.690 -0.824*** OCl 1.899*** 1.352 -2.601** 0.241 0.888 1.060** oc2 1.147*** 0.817 -1 .572" 0.150 0.536 0.640** R 2 0.928*** 0.331 0.643** 0.324 0.324 0.772*** C D H Po (intercept) 3.254 3.892 4.583 3.403 4.610 4.597 Pi 0.412* 0.796* 1.139*" 1.082** 1.831*** 1.953" P2 -0.320* -0.618* -0.885*** -0 .841" -1.424*** -1 .519" Pl2 -0.193 -0.117 0.041 -0.055 0.163 -0.443 ai 0.247 0.151 -0.0052 0.071 -0.209 0.569 a2 0.149 0.091 -0.032 0.043 -0.126 0.344 R 2 0.517* 0.543* 0.807*** 0 .716" 0.791*** 0 .733" M D A Po (intercept) 0.299 0.383 0.696 0.547 0.612 0.586 Pi 0.199** 0.257*** 0.415*** 0.220*** 0.348*" 0.298*** P2 -1.542** -0.200*** -0.323*** -0.170*** -0.271*** -0.232*** Pl2 -0.120 -0.077 0.124 0.170* 0.165* 0.112 C t i 0.155 0.100 -0.160 -0.219* -0.212* -0.145 a 2 0.094 0.061 -0.097 -0.132 -0.128* -0.088 R 2 0.812*** 0.827*** 0.728** 0 .736" 0.825*** 0.824*** 1 Values represent coefficients of quadratic equation after 150°C and 180°C heating. 2 *: P<0.05; **: P<0.01; ***: P<0.001. 3 Po = intercept; p\ = coefficient for cholesterol modified flaxseed oil at the first order; p2 = coefficient for butter oil at the first order; P12 = coefficient for the interaction between oils; ai = coefficient for cholesterol modified flaxseed oil at the second order; a 2 = coefficient for butter oil at the second order. 139 Table 22. Residual cholesterol content of fish-butter oil blends and butter-cholesterol modified flaxseed oil blends after 8, 16, 24 h heating at 150°C/180°C. 1 Fish-butter oil blends 150°C 180°C Time (h) A 2 B c D E A B C D E 8 0.946 0.935 0.936 0.915 0.880 0.641 0.610 0.600 0.578 0.304 ± 0.006 ± 0 . 0 1 0 ± 0.002 ± 0 . 0 1 4 ± 0 . 0 1 4 ± 0.004 ± 0.003 ± 0.001 ± 0.004 ± 0 . 0 1 1 16 0.894 0.869 0.839 0.808 0.664 0.507 0.371 0.303 0.290 0.223 ± 0 . 0 1 6 ± 0.007 ± 0.009 ± 0.009 ± 0 . 0 1 2 ± 0 . 0 1 3 ± 0.006 ± 0.008 ± 0.003 ± 0.006 24 0.257 0.238 0.209 0.210 0.250 0.245 0.221 0.058 0.103 0.000 ± 0.009 ± 0.004 ± 0.002 ± 0.005 ± 0.008 ± 0.001 ± 0.004 ± 0.002 ± 0.003 ± 0.000 Butter-cholesterol modified flaxseed oil blends 150°C 180°C Time (h) A 3 B C D E A B C D E 8 0.342 0.343 0.370 0.411 0.867 0.312 0.303 0.306 0.295 0.294 ± 0.003 ± 0.006 ± 0.002 ± 0.007 ± 0.004 ±0 .002 ± 0.007 ± 0.003 ± 0.008 ± 0 . 0 1 1 16 0.294 0.310 0.313 0.336 0.678 0.199 0.276 0.285 0.295 0.205 ± 0.007 ± 0.003 ± 0.004 ± 0.006 ± 0.008 ±0 .001 ± 0.004 ± 0.005 ± 0.008 ± 0.004 24 0.189 0.198 0.262 0.227 0.248 0.153 0.146 0.141 0.155 0.000 ± 0.008 ± 0.003 ± 0.007 ± 0 . 0 1 3 ± 0.006 ±0 .004 ± 0.002 ± 0.002 ± 0.003 ± 0.000 Values represent the residual cholesterol content, regarding the original cholesterol content before thermal treatment as 1. 2 To fish-butter o i l blends, A=100% fish o i l ; B=66.7% fish o i l + 33.3% butter o i l ; C=50% fish o i l + 50% butter o i l ; D=33.3% fish o i l + 66.7% butter o i l ; E=100% butter o i l . 3 To butter-cholesterol modified flaxseed o i l blends, A=100% cholesterol modified flaxseed o i l ; B=66.7% cholesterol modified flaxseed o i l + 33.3% butter o i l ; C=50% cholesterol modified flaxseed o i l + 50% butter o i l ; D=33.3% cholesterol modified flaxseed o i l + 66.7% butter o i l ; E=100% butter o i l . 140 Table 23. Formation of oxysterols of fish oil and cholesterol modified flaxseed oil following up to 24 h thermal incubation at 150°C and 180°C.1 Fish o i l 3 7-keto 7-hydroxy Epoxy 25-hydroxy Cholesterol triol cholesterol cholesterol cholesterol cholesterol 150°C 8h 0.625 0.507 1.216 0.079 16 h 0.774 1.154 1.139 0.101 2 24 h 0.254 0.238 0.243 - -180°C 8h 2.528 0.479 2.563 0.980 -16 h 4.710 0.532 1.531 1.120 -24 h 14.242 17.193 - - -Cholesterol modified flaxseed oil 3 7-keto 7-hydroxy Epoxy 25-hydroxy Cholesterol triol cholesterol cholesterol cholesterol cholesterol 150°C 8h 0.944 6.207 3.604 6.100 30.036 16 h 0.964 6.108 6.114 1.957 27.493 24 h 1.190 1.028 2.759 2.757 12.186 180°C 8h 2.732 1.928 4.264 1.241 10.829 16 h 5.708 1.588 2.979 0.841 5.179 24 h 8.473 1.542 2.361 0.729 4.000 1 Values represent the content of cholesterol oxidation products (in mg/lOOg oil), the amount (in mg) of cholesterol oxidation products was not calibrated with internal standard (5a-cholestane). The cholesterol content of native fish oil is 721.8mg/gm oil and that of cholesterol modified flaxseed oil is 700.0mg/gm oil. Cholesterol oxides are not found in the native fish oil and flaxseed oil. 2 "-" not detectable. 3 7-keto cholesterol = 3,5-cholestadien-7-one + 5-cholesten-3(3-ol-7-one; 7-hydroxy cholesterol = 5-cholesten-3P,7a-diol + 5-cholesten-3P,7P-diol; Epoxy cholesterol = cholestan-5P,6P-epoxy-3P-ol + cholestan-5oc,6oc-epoxy-3P-ol; 25-hydroxy cholesterol = 5-cholesten-3P,25-diol; Cholesterol triol = cholestan-3P,5a,6P-triol. 120.00 l 100.00 E E o co o a> o co •g CO 05 cr 80.00 60.00 = 40.00 20.00 0.00 16 Incubation time (h) 24 Figure 24. The comparison of residual linoleic acid (C18:2co6) of 100% flaxseed oil with and without cholesterol supplementation following thermal treatment at 150°C. • = without cholesterol, • = with cholesterol, * = P<0.05. 600.00 o E E 500.00 400.00 300.00 o CO o 'c 03 O = 200 .00 C O 8 100.00 DC 0.00 16 Incubation time (h) 24 Figure 25. The comparison of residual linolenic acid (C18:3(u3) of 100% flaxseed oil with and without cholesterol supplementation following thermal treatment at 150°C. • without cholesterol, • = with cholesterol, * = P<0.05. 142 Incubation time (h) Figure 26. The comparison of residual linoleic acid (C18:2co6) of 100% flaxseed oil with and without cholesterol supplementation following thermal treatment at 180°C. • = without cholesterol, • = with cholesterol, ** = P<0.01. 8 16 24 Incubation time (h) Figure 27. The comparison of residual linolenic acid (C18:3to3) ofl00% flaxseed oil with and without cholesterol supplementation following thermal treatment at 180°C. • = without cholesterol, • = with cholesterol, ** = P<0.01. 143 100 n 90 -8 16 24 Incubation time (h) Figure 28. Effect of heating on the loss of cholesterol froml00% fish oil following up to 24 h thermal incubation at 150°C and 180°C, respectively. • = 150°C, • = 180°C. 40 -i 35 A 8 16 24 Incubation time (h) Figure 29. Effect of heating on the loss of cholesterol froml00% flaxseed oil following up to 24 h thermal incubation at 150°C and 180°C, respectively. • = 150°C, • = 180°C. 144 Chapter 4. Effect of blending of natural phytochemicals with citric acid on the fatty acid profiles and oxidative stability of flaxseed oil. Introduction Antioxidants are of interest to the food industry because they prevent rancidity of fats and offset losses of essential fatty acids caused by oxidation reactions. Dietary antioxidants play an important role in prolonging the shelf life and maintaining the nutritional quality of lipid-containing foods, in addition to potentially modulating the consequences of oxidative damage in the human body induced by reactive oxygen species (Ffalliwell et al., 1995). Although the mostly common used antioxidants at the present time are BHA (butylated hydroxyanisole), BHT (butylated hydroxytoluene), PG (propyl gallate) and TBHQ (tertbutylhydroquinone) (Sherwin, 1990), there has been a growing concern as to whether they are safe or whether they promote carcinogenesis. As a result, there is a general desire to replace synthetic food antioxidant additives with natural alternatives (Wanasundara et al., 1996; Irwandi et al., 2000). Many have studied that the efficacy of rosemary and sage extracts as natural antioxidants in heated oils (Che Man et al., 2000). Natural antioxidants, such as rosemary extract and sage extract were applied to saturated fat such as palm olein (Irwandi et al., 2000). There are no reports examining the usefulness of using these same plant derived antioxidants to stabilize an oil source rich in co-3 PUFA, such as flaxseed oil. The purpose of this study was to employ an optimization procedure to determine the best mixture of natural antioxidants required to inhibit thermal induced oxidation of flaxseed oil. This approach was considered as an alternative method to the oil blending experiments reported earlier in the thesis to preserve co-3 PUFA. 145 Materials and Methods Oil source Flaxseed oil (fresh, cold pressed, unrefined and antioxidant free*), supplied by Flora Distributors Ltd. (Burnaby, B.C.), was used as a heating medium in this study. Flaxseed oil was stored in the dark in a cold room (-35 °C) until utilization. Block heater The DIGI-BLOCK digital block heater (Laboratory Devices Inc., USA) was used as the controlled temperature incubator for oil samples as described in Chapter 1. Antioxidants The antioxidants, rosemary extract (Type O) and sage extract (Type S-O) were purchased from Herbalox Seasoning (Kalamazoo, MI). Citric acid was purchased from Fisher Scientific (Fair lawn, NJ). Other chemicals Other Chemicals such as fatty acid methylates and thiobarbituric acid used in this study, have been described in Chapter 1. l,l-diphenyl-2-picylhydrazyl (DPPH) was obtained from Sigma-Aldrich Chem. Inc. (St. Louis, MO). Experimental design L Formulation of antioxidant combinations The Echip 6.04 program (Echip, Inc., Hockessin, DE) and the central composite design (CCD) were again used to study the effect of different antioxidant combinations of rosemary extract, sage extract and citric acid in reducing lipid oxidation of flaxseed oil. The 146 optimal points of determining the effect of different antioxidant combinations on changes in lipid oxidation products were based on measurements of conjugated diene hydroperoxides (CDH) and thiobarbituric acid reactive substance (TBARS), as well as the fatty acid profile of flaxseed oil. In the rosemary-sage-citric acid combination study, concentrations of rosemary extract and sage extract ranged from 0 to 0.2%, and citric acid was set from 0 to 0.05% (Table 24). The central composition design (CCD) employed a quadratic model which best optimized a set of combinations of antioxidants that are required to stabilize the oil. There were a total of 20 different combinations of rosemary-sage extracts-citric acid combination. The quadratic model was expressed by the following equation: Y = Po +Pl,XRosemary+P2*Xsage + +P3 # Xcitric acid + Pl2 • X R o s e m a r y • X s a g e + P l 3 •XR o s e mary • Xcitric acid + P23 *Xs age • Xcitric acid + OCi • X R o s e m a r y 2 +CC2 # X s a g e 2 + CC3 •Xcitric acid , in which, Po = intercept; Pi = coefficient for rosemary extract at the first order; P2 = coefficient for sage extract at the first order; P3 = coefficient for citric acid at the first order; P12 = coefficient for the interaction between rosemary extract and sage extract; P13 = coefficient for the interaction between rosemary extract and citric acid; P 23 = coefficient for the interaction between sage extract and citric acid; cci = coefficient for rosemary extract at the second order; a2 - coefficient for sage extract at the second order; 0C3 = coefficient for citric acid at the second order. 2. Heating studies Different antioxidant combinations were added at specified concentrations to the flaxseed oil, according to the experimental design given in Table 24. Each tube contained 3 147 g of flaxseed oil and the level of the oil was set to be lower than the surface of the block to allow uniform heat for each sample. Test tubes were covered with a marble and inserted into the block heater. Heating studies in this model were performed using different times of 2.5, 5, 10, 15, 20, 25 hours. The temperature was set at 180 °C. Analysis of lipid oxidation of oil with antioxidant combinations All the sample analyses were conducted in duplicate according to the following methods: L Free radical scavenging capacity of rosemary and sage extracts a. Free radical scavenging activity of rosemary and sage extracts The method was adopted from Hu and Kitts (2000). Rosemary and sage extract was made to lOmg/ml in absolute ethanol. Rosemary, or sage extracts at different concentrations (0, 0.02, 0.04, 0.1, 0.2, 0.3 and 0.4mg/ml) were mixed with 0.1 mM of l,l-diphenyl-2-picrylhydrazyl (DPPH) in absolute ethanol solution. Absorbance at 519nm was taken after 20 minutes. The effect of antioxidant treatment on the scavenging power for DPPH radical was expressed as an inhibition percentage. This was calculated according to the following equation: I n h % = Abs—< ~ A Z ? w x 1 0 0 A^Sconlrol : where, Abs s a mpie and Abs c o n t r oi represent absorbance at 519nm of DPPH, with and without an antioxidant sample, respectively. b. Synergistic effect of rosemary and sage extracts on free radical scavenging capacity 148 Different absolute amounts of rosemary, sage and combinations of the two plant phytochemicals were mixed with DPPH. Free radical scavenging capacity was tested according to the method described above. Inhibition percentages of rosemary and sage combination were compared with the mathematical sum of individual antioxidant substances. A synergistic effect was confirmed when the combination of antioxidants produced a greater inhibition effect than the sum of individuals. 2. Thiobarbituric acid reactive substances (TBARS) p.34 3. Conjugate diene hydroperoxides (CDH) p.34 4. Fatty acid quantification p.36 The above methods were the same as described in Chapter 1. Statistical analysis of data The experimental design software, Echip 6.04, and SPSS (SPSS for Windows, version 7.5, SPSS Inc.) were used as described in Experimental 1 to execute the statistical analysis. . . 149 Results Free radical scavenging effects of rosemary and sage extracts and their synergistic effect Radical scavenging effects of rosemary and sage extracts in quenching stable free radical DPPH in an ethanol solution are shown in Figure 30. Both plant constituents exhibited a concentration-dependent free radical scavenging activity, which was not significantly different from each other. In relative terms, sage appeared to be always more effective at neutralizing DPPH radical; however, the inhibition percentages of free radical DPPH for both natural plant extracts similarly approached 100% scavenging activities at a concentration of 0.4mg/ml. The effect of citric acid on quenching of DPPH radical was also analyzed in the same model, but no inhibition was found up to a relatively high concentration of lOmg/ml. A synergistic effect between rosemary and sage extracts to quench DPPH radical was also observed (Table 25). The combination of the two plant extracts produced a higher free radical scavenging inhibition that was greater than the sum of individual antioxidants tested (P<0.05). For example, the combination of rosemary extract and sage extract at 0.05 mg/ml or 0.1 mg/mL, resulted in a stronger inhibition reaction than the sum of both antioxidants (P<0.05) added individually (Table 25). Changes in fatty acid profiles during heating The list of combinations of sage extract, rosemary extract and citric acid used in this study were derived from the Echip computerized program, which set the concentration of sage extract or rosemary extract at a maximum of 0.2% and that of citric acid at 0.05%. 150 The changes in fatty acid composition of flaxseed oil with different combinations of rosemary extract, sage extract and citric acid, during thermal oxidation at 180°C were characterized using the quadratic regression coefficients derived from the fatty acid composition, P/S ratio and u)3/u)6 ratio (Table 26). Samples heated up to 25 h showed a significant first order coefficient effect with both rosemary extract (P<0.01) and sage extract (P<0.001). There was no statistical effect of using citric acid on the retention, or the change, in fatty acid composition of different blends, thus demonstrating that citric acid had no independent antioxidant effect. Samples incubated for different durations presented distinct results, indicating potential interactions between the two antioxidants. The combination of 0.2% rosemary extract, 0.2% sage extract and 0.05% citric acid represented the best antioxidant mixture required to retain linoleic acid in flaxseed oil (Figure 31), following 180°C heating for 5 h. Extending the heat treatment to 25 h had very little effect in altering the efficacy in which the antioxidant mixture retained linoleic acid (Figure 32). The oil samples containing antioxidants and incubated for a shorter time period exhibited a considerable inter-sample variability. A trend towards greater protection against linolenic acid oxidation was obtained by the use of 0.2% rosemary extract and 0.1% sage extract (Figure 33). This trend is also shown in Figure 34. In all cases, however, the essential differences in PUFA retention within treatments was not statistically significant (Figures 31 and 34). This result may be due to the fact that the high temperature used caused a severe damage to the antioxidants employed in this study. The addition of rosemary, sage extracts with citric acid, however, lowered the rate of depletion for both linoleic acid (Figure 31) and linolenic acid (Figure 33) over a 5 h treatment duration. This reduction however, was not found to be significant. 151 The formation of CDH and TBARS in antioxidant treated, heated flaxseed oil Both the formation of CDH and TBARS from heated flaxseed oil increased with time, reflecting a time-dependent oxidation progression for the oil. The addition of a combination of natural antioxidant extracts and citric acid, however, resulted in a reduced (P<0.05) formation of both CDH and TBARS. In fact, a synergistic effect to reduce TBARS formation occurred with rosemary extract and citric acid. Specifically, TBARS from flaxseed oil were lower in samples treated with a combination of 0.2% rosemary extract and 0.05% citric acid, compared to the individual addition of each constituent. Table 27 also gives the regression coefficients derived from the quantitative generation of CDH and TBARS that describe the quadratic model. The sage extract was the only plant constituent that produced a significant effect on lowering both the generation of CDH and TBARS. Optimization of antioxidant combinations A contour map graphically representing the quality of heated flaxseed oil after addition of different combinations of rosemary extract-sage extract-citric acid showed that the presence of natural antioxidants, rosemary and sage, was more pronounced than the citric acid alone (Figures 35 and 36). The addition of 0.067% rosemary and 0.067% sage extracts, or 0.1% rosemary and 0.2% sage extracts, plus 0.05% citric acid was the optimal mixture that resulted in the lowest formation of CDH (Figure 35), or TBARS (Figure 36), in heated flaxseed oil, respectively. Varying the amount of citric acid had no synergistic effect in this model. The results of the addition of 0.1% rosemary, 0.1% sage extracts and 0.05% citric acid on lipid oxidation indices also correspond with the highest retention for linoleic acid (Figure 37) and linolenic acid (Figure 38). 152 Discussion The antioxidant activities of rosemary and sage extracts have been well summarized in several reviews (Hall and Cuppett, 1997; Cuppett and Hall, 1998). The interrelationships and mechanisms underlying antioxidant activity from rosemary and sage may be best understood by studying the biochemical mechanisms underlying the interactions. The biochemical conversion of carnosic acid to a variety of rosemary and sage antioxidants gives a direct indication as to how these compounds function as antioxidants (Hall and Cuppett, 1997). Polar compounds from rosemary extract (carnosic acid and romarinic acid) were found to be more effective in bulk oils, and less polar compounds (carnosol) were more effective in emulsified lipids (Frankel et al, 1996). As demonstrated herein, rosemary and sage extracts were both effective at directly quenching the 2,2-diphenyl-l-picryhydrazyl (DPPH) radical confirming the free radical scavenging activity of primary antioxidants properties for both spice extracts (Cuppett and Hall, 1998). The DPPH radical quenching capacity assay has been extensively used to assess antioxidant activity of many natural antioxidants (Brand-Williams et al, 1995; Kitts et al, 2000; Hu and Kitts, 2000). The principle of this assay is the extent of disappearance of color from DPPH, which is related to the potential to quench radicals (Brand-Williams et al, 1995; Kitts, et al, 2000; Espm et al, 2000; Pekkarinen et al, 1999). The mechanism of interaction between antioxidant (AH) and DPPH radical is shown below (Brand-Williams et al, 1995): DPPH* + AH - 4 DPPH2 + A* Equation 14 Pekkarinen et al. (1999) also demonstrated that this radical quenching activity, reflected the antioxidant activity in bulk oil, but not in an emulsion. Furthermore, the hydrogen donating capacity of the primary antioxidant activity of the rosemary and sage 153 extracts maybe due to either the multiple action of individual constituents or possible interaction of these constituents present in these herbs. For example, rosmanol can contribute a H-atom to DPPH radical (or lipid peroxyl radical LOO*), resulting in the formation of a semiquinone intermediate, which will further react with free radicals to generate a stable product, such as the formation of o-quinone rosemaquinone (Appendix L and M). The rosemary-sage extract with citric acid was effective at delaying the oxidation of flaxseed oil under higher incubation temperature. According to the indices of lipid oxidation used in this study (i.e., the generation of TBARS and CDH), a combination ratio of rosemary-sage extract-citric acid that produced the lowest formation of TBARS was 0.1%-0.2%-0.05%, and 0.067%-0.067%-0.05% for CDH. This confirms the finding that both rosemary and sage were free radical scavengers, as suggested by the DPPH model assay. Using this model, active components of rosemary and sage extract have been identified (Cuppett and Hall, 1998), but the impact of the contributing effect from citric acid was less significant. The original purpose of using citric acid in combination with primary natural antioxidants has been suggested for its potential synergistic effect, as previously shown with tocopherol in different circumstances (Cooney et al., 1958; Evans et al., 1959; Masuyama, 1965). Other workers reported a lower synergistic effect of citric acid when added with tocopherol in lard and palm oil (Kanematsu et al., 1983). These authors also concluded that the presence of a smaller amount of metal ions in refined oil may have contributed to a prooxidant effect. Citric acid is classified as a secondary antioxidant, which prevents oxidation by chelating transition metal ions (Gunstone, 1996). Although citric acid is known 154 to be a weaker metal chelator than EDTA, it has some affinity to prevent various transition metal ions from inducing lipid oxidation reactions (Gordon, 1990). A synergistic behavior of antioxidants requires that the effect of the combination of compounds is greater than the sum of the effects caused by either antioxidant (Niki 1996). A synergistic effect of rosemary extract with other antioxidants has been reported, with the example of rosemary extract in prime steam lard enhancing the effect of added ascorbic acid (Chang etal, 1977). Free radical scavenging capacity of citric acid against a water soluble free radical, such as 2,2'-azinobis (3-ethylbenzothiozoline-6-sulfonic acid) (ABTS*+) radical has not been demonstrated (Miller, 1998), therefore, the non-free radical scavenging capacity of citric acid in both water and organic solvent systems indicates a limited free radical scavenging capacity. Banias et al. (1992) found that citric acid exhibited a synergistic effect with a marjoram extract in heated lard (75°C) antioxidant reactions, but interfered with the antioxidant activities of rosemary and sage extracts. This has been attributed to the esterification of citric acid to phenolic antioxidant groups (Gordon 1990). Carnosic acid has long been recognized as a free radical scavenger, whereas citric acid is regarded primarily as a synergist (Schuler, 1990). The synergistic effect of citric acid should permit the reduction of the required concentration necessary for the antioxidant to function in a food formulation, which in turn lowers the cost of application and possible undesirable side effects attributed to the antioxidant agent (e.g. taste). Synergists are examples of metal chelating agents, peroxide decomposers or regenerators of oxidized antioxidant. In living tissue, for example, the affinity for ascorbic acid to regenerate oxidized tocopheroxyl, makes it a synergist for 155 tocopherol, in addition to providing direct antioxidant activity (Kitts, 1997). Unlike the findings of others that have reported a synergistic effect of citric acid in model lipid systems, this property was not observed in the present study. It is possible that the relatively greater proportion of PUFA present in the oils studied herein precluded an antioxidant effect of citric acid, which has been noted primarily with saturated fats. In addition, the oils used in the present study may not have contained the some level of transitional metal ions and this sequestering activity of citric acid was not an issue in respect to eliciting further antioxidant activity. 156 Conclusion (4) This experiment demonstrated that the combination of oleoresin rosemary and sage extracts produced an effective and enhanced free radical scavenging activity in a DPPH radical model in vitro test. It was concluded from these results that the measured antioxidant activity derived from the combination of antioxidant extracts exhibited a synergistic effect compared to the effect derived from the sum of individual compounds. The addition of rosemary extract and sage extract was shown to reduce loss of two PUFAs, namely, linoleic acid and linolenic acid to a level that corresponded to a reduction in the formation of TBARS (e.g. MDA) and CDH in this study. Citric acid on the other hand had no significant synergistic effect with the antioxidant combinations of rosemary and sage extracts. In addition, citric acid did not quench free radical DPPH. The optimizations made for the antioxidant combinations of the two plant extracts with citric acid were: 0.1% rosemary extract and 0.2% sage extract, or 0.067% rosemary extract and 0.067% sage extract, with 0.05% citric acid for producing the lowest formation of TBARS (e.g. MDA), or CDH, respectively. The optimum concentration of these two plant phenolics to protect against linoleic and linolenic acids degradation on thermally processed oil was 0.1% rosemary extract and 0.1% sage extract. The application of naturally occurring plant phytochemicals with antioxidant properties was shown to be effective at producing a trend in the retention of co-3 PUFA. This trend however was not statistically different and therefore it cannot be concluded that the mixture of plant antioxidants were effective in reducing co-3 PUFA oxidation and corresponding generation of lipid oxidation products. A possible explanation for this 157 conclusion could involve the stability of the antioxidant extract materials to thermal treatment. Table 24. Combinations of proportion of oleoresin rosemary, sage extracts and citric acid added into flaxseed oil before heating.1 Sample N o . Rosemary Extract (%) Sage Extract (%) Citr ic A c i d (%) I1 0 0.2 0.05 2 0 0.2 0.05 3 0.2 0.1 0 4 0.2 0.1 0 5 0 0.2 0 6 0 0.2 0 7 0 0 0.05 8 0 0 0.05 9 0.2 0.2 0.05 10 0.2 0.2 0.05 11 0.1 0.2 0.025 12 0 0.1 0.025 13 0.1 0.1 0.05 14 0 0 0 15 0.2 0 0.025 16 0.2 0.2 0.025 17 0.1 0.2 0 18 0.1 0 0 19 0.2 0.1 0.05 20 0 0.1 0 Proportions presented the composition of rosemary extract, sage extract and citric acid in flaxseed o i l in percentage (w/w). 2 Each sample contained equals 3g of o i l with/without any antioxidant. 159 Table 25. Synergistic effect of rosemary and sage extracts in scavenging free radical. Sample N o . Antioxidant combinations in ethanol Inhibit ion%' 1 Control (without any antioxidants) 0 2 Rosemary, 0.05mg/ml 5.7±2.2 3 Rosemary, 0.1 mg/ml 15.6+0.5 4 Rosemary, 0.2mg/ml 44 .3±2 .2 5 Sage, 0.05mg/ml 8.611.8 6 Sage, 0.1 mg/ml 22 .8±1 .8 7 Sage, 0.2mg/ml 55 .1±1 .6 8 0.025mg/ml rosemary + 0.025mg/ml sage 8.5±0.3 9 0.05mg/ml rosemary + 0.05mg/ml sage 2 2 . 3 ± 1 . 0 2 10 0.05mg/ml rosemary +0.15 mg/ml sage 54 .1±1.3 11 O. lmg/ml rosemary + O.lmg/ml sage 5 1 . 5 ± 2 . 6 3 12 0.15mg/ml rosemary + 0.05mg/ml sage 48.210.5 Values presented the scavenging effect of rosemary extract, sage extract and their combinations on the inhibition activity of stable free radical D P P H . 2 The sum of the values from sample 2 and 5 was 14.3. 3 The sum of the values from sample 3 and 6 was 38.4. 160 Table 26. Quadratic regression coefficients derived from fatty acid composition, P/S ratio and n-3/n-6 ratio of flaxseed oil with antioxidant blends of rosemary extract, sage extract and citric acid following thermal treatment at 180°C for 2.5, 5, 10,15,20, and 25hours, respectively. Time Coefficients C16:0 C18:0 C18:l C18:2 C18:3 P/S n-3/n-6 2.5 h (30 (intercept) 0.0575 0.0408 0.176 0.0963 0.372 4.702 3.855 Pi 3.46E-07 -3.77E-07 8.8E-06 1.46E-06 7.34E-06 0.0001 1.88E-05 P2 1.88E-06 -2E-07 -4.51E-06 5.05E-06 2.99E-05 0.000244 0.000073 P3 -5.7E-06 1.02E-06 -3.44E-05 -1.41E-05 -8.03E-05 -0.000669 -0.0002 Pl2 1.77E-09 -3.68E-10 1.62E-08 4.84E-09 3.02E-08 2.64E-07 8.42E-08 Pl3 -9.9E-09 4.92E-10 3.17E-08 -2.56E-08 -1.49E-07* -1.19E-06* -3.9E-07 P23 4.21E-09 -1.71E-09 -5.48E-08 1.53E-08 9.79E-08 1.01E-06 3.1E-07 CCl 1.66E-09 -1.33E-09 -4.95E-09 7.14E-09 4.77E-08 5.3E-07 1.6E-07 a-2 3.72E-09 -1.03E-09 2.15E-09 1.15E-08 6.94E-08* 6.77E-07* 2.1E-07 a 3 -3.73E-08 -1.74E-08 -2.03E-07 -2.59E-08 -1.05E-07 2.85E-06 4.76E-07 R 2 0.456 0.310 0.389 0.600 0.649 0.744* 0.723 5 h p 0 (intercept) 0.0699 0.0388 0.193 0.13 0.581 6.515 4.463 Pi -7.52E-07 1.41E-07 -1.68E-06 -1.59E-06 -1.51E-05 -0.000124 -6.86E-05 P2 1.48E-06 4.16E-07 3.41E-06 2.35E-06 1.57E-05 6.46E-05 4.39E-05 P3 1.64E-06 2.62E-06 5.02E-06 2.42E-06 2.67E-06 -0.000202 -6.82E-05 P12 2.35E-09 -2.17E-10 5.25E-09 5.73E-09 4.13E-08 3.47E-07 1.54E-07 Pl3 -8.91E-10 1.83E-09 1.16E-11 -4E-09 -2.86E-08 -3.84E-07 -1.03E-07 P23 3.41E-09 2.59E-09 1.07E-08 3.73E-09 1.52E-08 -1.62E-07 -8.49E-09 CCl -9.21E-10 -9.3E-10 -2.56E-09 -1.01E-09 -8.42E-10 1.16E-07 4.7E-08 a2 -8.35E-09* 8.44E-10 -1.85E-08* -1.89E-08* -1.19E-07* -8.97E-07* -3.19E-07 a3 -4.27E-08 -1.61E-09 -8.11E-08 -1.04E-07 -7.28E-07 -5.51E-06 -2.41E-06 R 2 0.502 0.658 0.502 0.527 0.526 0.546 0.523 10 h Po (intercept) 0.06 0.0397 0.169 0.103 0.394 4.99 3.842 Pi 3.54E-07 1.64E-07 1.19E-06 1.4E-06 6.97E-06 6.38E-05 1.57E-05 P2 1.25E-06 6.68E-07 3.2E-06 2.29E-06 1.02E-05 3.38E-05 1.34E-05 P3 -1.67E-06 -2.58E-06 -3.94E-06 1.1E-06 1.47E-05 0.000358 0.000101 P12 3.97E-10 2.38E-10 9.1E-10 1.52E-10 4.68E-10 -1.93E-08 2.39E-09 Pl3 -3.92E-09 -3.6E-11 -9.8E-09 -1.19E-08 -6.96E-08 -6.41E-07 -2.21E-07 P23 -3.42E-09 -4.42E-09 -1.16E-08 -2.74E-09 8.61E-10 3.5E-07 9.42E-08 «1 -3.63E-09 -3.71E-09 -9.13E-09 -1.91E-09 5.07E-09 3.91E-07 1.17E-07 a2 1.45E-09 1.08E-09 4.48E-09 2.96E-09 1.36E-08 4.34E-08 1.74E-08 a3 7.15E-09 2.53E-08 2.67E-08 -3.07E-08 -2.55E-07 -4.48E-06 -1.28E-06 R 2 0.307 0.504 0.295 0.236 0.276 0.573 0.515 15 h Po (intercept) 0.0555 0.039 0.158 0.0909 0.339 4.569 3.738 Pi -2.17E-06 -1.15E-07 -4.32E-06 -4.17E-06 -0.000022 -0.000116 -4.58E-05 P2 -2.61E-06 -2.06E-06 -7.43E-06 -3.52E-06 -1.25E-05 9.02E-05 1.98E-05 P3 -0.000011 -5.47E-06 -2.86E-05 -0.000021 -9.24E-05 -0.000222 -7.04E-05 Pl2 4.89E-09 2.24E-09 1.26E-08 9.35E-09 4.13E-08 1.41E-07 4.69E-08 Pl3 1.19E-08 9.58E-09 3.12E-08 1.5E-08 4.86E-08 -4.52E-07 -1.17E-07 P23 3.25E-08** 1.53E-08* 8.4E-08** 6.05E-08** 2.67E-07** 8.67E-07 2.97E-07 a i -1.05E-10 -1.03E-09 -1.99E-10 1.5E-09 1.09E-08 1.51E-07 3.85E-08 «2 5.88E-09 3.33E-09 1.55E-08 1.02E-08 4.19E-08 6.89E-08 2.55E-08 0:3 4.19E-08 2.59E-08 1.1E-07 7.29E-08 2.75E-07 4.29E-07 1.04E-08 R 2 0.724 0.630 0.713 0.745* 0.745* 0.408 0.417 161 (Continuation of Table 26) 20 h po (intercept) 0.0622 0.0396 0.173 0.104 0.398 4.908 3.804 P i -8.67E-07 1.43E-07 -2.18E-06 -2.17E-06 -1.11E-05 -9.69E-05 -2.94E-05 P2 -2.18E-07 -3.09E-07 6.2E-08 1.07E-06 2.83E-06 6.59E-05 -7.29E-06 P3 4.15E-06 -9.52E-07 6.57E-06 9.27E-06 6.41E-05 0.000559 0.000265 Pl2 5.28E-10 -2.3E-10 1.53E-09 2.14E-09 8.52E-09 9.29E-08 8.63E-09 Pl3 7.26E-10 -2.07E-09 -1.97E-09 1.32E-09 1.57E-08 2.2E-07 9.73E-08 P23 -1.46E-09 -9.14E-10 -4.55E-09 -4.86E-09 -2.52E-08 -1.82E-07 -6.5E-08 OCl -1.36E-09 2.57E-10 -5.01E-09 -5.98E-09 -3.41E-08 -3.32E-07 -9.36E-08 «2 -1.4E-09 8.39E-10 -3.48E-09 -5.91E-09 -3.35E-08 -3.53E-07 -9.74E-08 0:3 8.03E-09 1.57E-09 2.89E-08 2.88E-08 1.75E-07 1.55E-06 5.65E-07 R 2 0.358 0.358 0.501 0.375 0.360 0.344 0.304 25 h Po (intercept) 0.0581 0.0388 0.163 0.0892 0.322 4.411 3.597 P i -7.48E-06*** -2.4E-06 -1.48E-05** -7.68E-06** -2.64E-05* 0.00039 -1.66E-07 P2 1.09E-05*** 4.83E-06** 2.08E-05*** 1.44E-05*** 6.02E-05*** -0.000369 0.000113 P3 -9.05E-06 4.17E-07 -2.14E-05 -1.02E-05 -4.96E-05 0.000133 -0.00013 Pl2 1.14E-08*** 3.92E-09* 2.24E-08*** 1.17E-08*** 4.2E-08** -6E-07* 2.13E-08 Pl3 -1.25E-08 -5.61E-09 -2.53E-08 -2.28E-08* -1.14E-07* -1.16E-07 -3.35E-07 P23 7.46E-09 -3.55E-09 1.55E-08 8.21E-09 2.47E-08 -9.77E-08 -6.35E-08 « 1 -4.82E-09 -4.01E-09 -1.06E-08 -2.98E-09 -1.45E-08 4.24E-07 -3.55E-08 a 2 -1.04E-08* -4.15E-09 -2.07E-08** -1.47E-08** -5.93E-08** 2.94E-07 -9.42E-08 a 3 1.57E-07* 6.99E-08 2.75E-07* 1.67E-07* 7.07E-07 -8.38E-06 1.26E-06 R 2 0.863** 0.702 0.866** 0.872** 0.832** 0.531 0.410 1 Values represent coefficients of quadratic equation after 180°C heating. 2 *: P<0.05; **: P<0.01; ***: P<0.001. 3 Po = intercept; Pi = coefficient for rosemary extract at the first order; p 2 = coefficient for sage extract at the first order; p 3 = coefficient for citric acid at the first order; p ) 2 = coefficient for the interaction between rosemary extract and sage extract; P13 = coefficient for the interaction between rosemary extract and citric acid; p 2 3 = coefficient for the interaction between sage extract and citric acid; (Xi = coefficient for rosemary extract at the second order; oc2 = coefficient for sage extract at the second order; a 3 = coefficient for citric acid at the second order. 162 Table 27. Quadratic regression coefficients derived the formation of conjugated diene hydroperoxides (CDH) and malonadehyde (MDA) of flaxseed oil with antioxidant blends of rosemary extract, sage extract and citric acid following thermal treatment at 180°C for 2.5, 5, 10,15,20, and 25 h, respectively. Time Coefficients C D H T B A Time Coefficients C D H T B A 2.5 h Po (intercept) 3.932 0.660 15 h Po (intercept) 6.468 0.720 P i 1.102E-04 4.560E-05 P i -2.671E-05 7.080E-05 P2 -2.698E-04** -4.218E-05 P2 -2.332E-04 -1.002E-04* P 3 4.529E-04 -2.724E-04* P3 3.797E-05 1.248E-04 Pl2 -6.658E-08 2.970E-08 Pl2 -9.085E-08 -1.704E-08 Pl3 4.095E-07 3.570E-09 Pl3 -9.722E-08 5.490E-08 P23 3.037E-08 -3.696E-08 P * -8.041E-07 1.182E-07 CCl -2.847E-07 1.884E-08 OCl -3.444E-07 2.376E-08 a 2 1.098E-07* 1.236E-08 a 2 -5.722E-09 8.100E-08 a 3 9.749E-07 -7.440E-07 0C3 3.498E-06 2.556E-06 R 2 0.876 0.595 R 2 0.771 0.568 5 h Po (intercept) 4.922 0.888 20 h Po (intercept) 8.258 0.828 P i -7.336E-05 3.660E-06 P i -4.461E-04 -4.722E-05 P2 1.176E-04 -1.014E-04 P2 -3.702E-04 -4.980E-05 P3 7.647E-04* 2.868E-04 P3 1.369E-03 1.128E-04 Pl2 1.898E-07 -5.412E-08 Pl2 -1.614E-07 4.368E-08 Pl3 3.729E-07 1.860E-07 Pl3 -7.566E-08 -1.416E-07 P23 -2.197E-08 -5.730E-07* P23 3.146E-07 1.440E-07 a i 1.018E-07 -1.242E-07 a i -1.006E-06 1.068E-07 a 2 3.200E-07 4.020E-08 a 2 -3.919E-07 -5.082E-08 a 3 -2.942E-06 9.780E-07 a 3 -1.397E-06 1.020E-07 R 2 0.744 0.596 R 2 0.562 0.707 10 h Po (intercept) 5.058 0.780 25 h Po (intercept) 7.851 1.020 P i -2.644E-05 -1.116E-05 P i 5.858E-06 -5.142E-05* P2 -1.641E-04 -2.244E-05 P2 -2.454E-04 -3.060E-05 P3 3.742E-04 -1.398E-04 P3 1.017E-04 1.314E-04 Pl2 1.980E-08 2.784E-08 P12 5.153E-08 -3.768E-08 Pl3 3.512E-07 1.266E-07 Pl3 -2.373E-07 -4.908E-08 P23 -4.068E-07 9.000E-10 P23 1.101E-07 -8.280E-08 a i -4.854E-08 -5.160E-08 « l 4.881E-08 5.256E-08 a-l 4.610E-07 1.020E-07 a 2 -8.000E-09 -5.754E-08 a 3 -3.173E-06 -1.326E-06 0C3 -2.671E-06 -3.462E-07 R 2 0.478 0.594 R 2 0.466 0.697 1 Values represent coefficients of quadratic equation after 180°C heating. 2 *: P<0.05; **: P<0.01. 3 Po = intercept; PL = coefficient for rosemary extract at the first order; p2 = coefficient for sage extract at the first order; p3 = coefficient for citric acid at the first order; P i 2 = coefficient for the interaction between rosemary extract and sage extract; p13 = coefficient for the interaction between rosemary extract and citric acid; p23 - coefficient for the interaction between sage extract and citric acid; ax = coefficient for rosemary extract at the second order; a 2 = coefficient for sage extract at the second order; oc3 = coefficient for citric acid at the second order. 163 0 0.02 0.04 0.1 0.2 0.3 0.4 Concentration of extracts (mg/ml) Figure 30. Concentration-dependent effect of rosemary extract and sage extract on scavenging O.lmM DPPH radical in ethanol solution. • = rosemary extract, • = sage extract. 164 120.00 Trea tments Figure 31. The variation of residual linoleic acid (C18:2co6) in flaxseed oil with/without antioxidant following 5 h thermal treatment at 180°C. S = sage extract, R= rosemary extract, CA = citric acid. 165 120.00 Treatments Figure 32. The variation of residual linoleic acid (C18:2co6) in flaxseed oil with/without antioxidant following 25 h thermal treatment at 1 8 0 ° C . S = sage extract, R= rosemary extract, C A = citric acid. 166 -g o co o 'c 0) o c -g '</> CD oc 600. 500. 400. 300. 200. 100. 0. 00 00 00 00 00 00 00 O CO CO CO rr o o CM b °: CO 0s-CO § Treatments Figure 33. The variation of residual linoleic acid (C18:3co3) in flaxseed oil with/without antioxidant following 5 h thermal treatment at 180°C. S = sage extract, R= rosemary extract, CA = citric acid. 167 400.00 Treatments Figure 34. The variation of residual linoleic acid (C18:3o)3) in flaxseed oil with/without antioxidant following 25 h thermal treatment at 180°C. S = sage extract, R= rosemary extract, CA = citric acid. 168 CDH (Ltmol/g oil) Citric acid = 0.05% Rosemary extract Figure 35. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the formation of conjugated diene hydroperoxides (CDH) (pmol/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 169 MDA (umol/g oil) Rosemary extract Figure 36. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the formation of malonadehyde (MDA) (umol/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 170 linoleic acid (mg/g oil) Citric acid = 0.05% Rosemary extract Figure 37. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the amount of linoleic acid (mg/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 171 0.2$ linolenic acid (mg/g oil) 0.15%-! g 0.1%-] « 00 0.05%H X u u o-\ X N ">» \ / / / / / . / / s /////, < s f t / / / / / / A 0.05% 0.1%' 0.15% Rosemary extract p Citric acid = 0.05% 0.2% Figure 38. Contour maps describing the effect of rosemary extract, sage extract and citric acid on the amount of linolenic acid (mg/g oil) from the flaxseed oil samples following 20 h thermal treatment at 180°C 172 General Conclusions and Major Findings of the Thesis Polyunsaturated fatty acids (PUFA) found in vegetable oils and marine oils represent important sources of essential fatty acids for animal and human nutritional requirements. Plants can synthesize both co-6 and co-3 fatty acids; the latter often referred to as co-3 fatty acids are especially important for their role in composing different membrane phospholipids. Oil seeds such as flaxseed and canola represent excellent sources {e.g. 50% and 10%) of C18:3, co-3 PUFA in the form of cc-linolenic acid, which should not be confused with the very long chain co-3 PUFAs found in marine oils (e.g. eicosapentaenoic acid (EPA; C20:5, co3) and docosahexaenoic acid (DHA; C22:6, co3)). Paradoxically, these essential fatty acids are particularly susceptible to autoxidation reactions and oxidize readily in chemical systems (e.g. bulk oils or emulsions) and in foods. The result of these lipid peroxidation reactions contributes not only to the loss of essential fatty acids, but also the generation of lipid oxidation products which initiate and promote oxidation of other essential nutrients, such as vitamins and amino acids. Moreover, there is concern from the consumer point of view that chronic consumption of lipid oxidation products may lead to in vivo peroxidation reactions that underlie atherosclerosis, carcinogenesis and numerous other diseases and disorders. Food manufactures, processors and nutritionists have relied on chemical methods which include hydrogenation and interesterification, in addition to blending, to reduce the relative susceptibility of PUFAs to oxidation reactions. The purpose of this research thesis was to evaluate the usefulness of blending a vegetable oil source and a marine oil source, two distinctly different lipid sources (e.g. PUFA and sterol content), with a 173 medium chain saturated fatty acid source derived from butter oil for the purpose of reducing loss of essential fatty acids and generation of undesirable lipid oxidation products. Finally, a second strategy for stabilizing a PUFA-rich vegetable oil was to use a natural antioxidant mixture that consisted of extracts from known plant antioxidant sources. In both experiments, optimization protocols were used to identify the best combination of materials that could represent an alternative to the more traditional methods of stabilizing PUFA-rich oils. In Experiment 1, a marine (e.g. herring/salmon) oil was initially analyzed for fatty acid composition and cholesterol content. The primary PUFAs present in the fish oil were the very long co-3 chain fatty acids, EPA and DHA. This oil source was blended with butter oil, a source of medium to long chain saturated fatty acids, according to an optimization strategy to yield favorable P/S and co3/co6 PUFA ratios with relatively improved susceptibility to thermally-induced autoxidation. The results of this experiment first of all demonstrated that employing relative percentages of fatty acids, as a measure of chemical alteration, was not as accurate as using specific mass ratios for individual PUFAs for evaluating the effects of thermal oxidative induced losses in fatty acids. Notwithstanding this, was the finding that optimizing the butter-fish oil blends using maximum P/S or co3/co6 PUFA ratios was not possible for predicting the best blend combination described by a quadratic model. This result was substantiated further by measuring a number of lipid oxidation product parameters generated from the thermal oxidation of the fish oil blends. Quantitation of both initial (e.g. peroxide value (PV) and conjugated diene (CD)) and secondary lipid oxidation products (e.g. malonaldehyde (MDA)), confirmed the trend in stability of PUFA content present in butter-fish oil 174 blends. It was concluded from this experiment that possibly the relatively small proportion of co-3, long chain PUFAs present in the fish oil were not sufficient in concentration to enable an estimate of optimal blend of butter oil. This limitation resulted in the rejection of the null hypothesis set for this experiment. In fact, PUFAs seemed to be more susceptible to oxidation and were lost more rapidly than saturated fatty acids when present in the blend. Experiment 2 was initiated with flaxseed oil, a vegetable oil especially rich in co-3 fatty acid, C18:3co3. Unlike the fish oil used in Experiment 1, flaxseed oil was found to contain 58% a-linolenic acid along with 14% co-6, linoleic acid. Thus the P/S ratio of flax seed was approximately 15x greater than the fish oil and contained only half the co3/co6 PUFA ratio present in fish oil. Blending of the flax seed oil with butter oil at different combinations proved to be effective at deriving an optimal blend for the relative protection of PUFAs, as indexed by the co3/co6 fatty acid ratio following thermal-induced oxidation at both 150°C and 180°C. This result was not obtained with the fish oil experiments, and as a result addresses the need for a critical amount of co-3 PUFAs in the oil mix to predict an optimal stability. Moreover, it was especially interesting to note that a relatively higher percentage of co-3 PUFAs was retained following 150°C thermal oxidation from the fish oil compared to the flax seed oil within in 16 hours. This result led to the conclusion that other factors present in the crude lipid content of these oils may contribute to the stability of the co-3 PUFAs. One possible explanation for this observation was that the cholesterol content of fish oil represents a marked difference in the lipid composition of both fat sources. 175 Experiment 3 was therefore initiated to test the hypothesis that the presence of cholesterol was a factor in the relative susceptibility of PUFA to thermal-induced oxidative losses. To test this hypothesis, cholesterol was added to the flaxseed oil at the same concentration present in the fish oil, and the experiment was repeated with the blending of butter oil. The results of Experiment 3a showed that cholesterol indeed was a stabilizing factor for the retention of PUFA, especially the co-3 PUFA content of the flaxseed oil. Confirmation of this finding was made by the lipid oxidation data that identified both reduced primary and secondary lipid oxidation products from flaxseed oils that had been fortified with cholesterol. This observation is an important finding and identifies the importance of considering the complete lipid composition of a fat source when predicting the susceptibility of individual fatty acids to oxidative losses. Coupled with Experiment 3 was the identification of possible cholesterol oxidation products that were derived from the oxidation of cholesterol due to thermal oxidation. In Experiment 3b, nine different cholesterol oxidation products (COPS) were identified by gas chromatography and quantitated from individual response linearity curves, using 5a-cholestane as the internal standard. The findings from this experiment indicated that cholesterol oxidation occurred with thermal treatment; however, the extent of cholesterol degradation to oxidized sterols was dependent on the duration of heat treatment used on both the fish oil and flax seed oils. Of the different possible COPS formed with heating these oils, the oxysterol, 7-keto-cholesterol, consistently was found to be present in thermal treated oils. The oxysterol, 25-hydroxycholesterol was the only COPS found from the thermal treatment of both the fish oil and the flax seed oil blends. Since cholesterol is susceptible to oxidation reactions, the finding in this study that 176 showed the generation of COPS from PTJFA-rich oils on thermal processing indicate that cholesterol indeed can act as an antioxidant by utilizing oxygen for COPS formation that otherwise would be available for oxidation of PUFA. Although this finding supports previous suggestions that point to the same conclusion, it is obvious that the use of cholesterol as an exogenous antioxidant agent is not a practical solution to preserving PUFA rich oils. The collective findings of Experiments 1-3 indicated that blending of PUFA rich oils with saturated lipids or lipids containing cholesterol was only marginally effective at protecting against PUFA losses. An alternative approach was attempted in Experiment 4 with the blending of flaxseed oil with a combined mixture of rosemary, sage extracts and citric acid. Components of rosemary and sage extracts were found to be effective scavengers of the stable free radical DPPH. With this preliminary study, an optimization experiment followed, in which the best mixture of the three natural antioxidants was evaluated in flaxseed oil processed at 180°C. Results indicated that a mixture of 0.1% rosemary extract with 0.2% sage extract produced the optimal mixture for reduced malonaldehyde, whereas a mixture of 0.067% rosemary and 0.067% sage with 0.05% citric acid proved to be the best mixture for reduced conjugated diene formation. These mixtures of natural antioxidants also provided the maximum retention of PUFAs in the 180°C treated flaxseed oil. These studies have helped explain the importance of evaluating the total composition of crude lipid in a oil system whereby stability of PUFAs is a primary goal and the inhibition of generated lipid oxidized products associated with losses of PUFA is attempted. The use of blending PUFA rich oils with a more saturated fatty acid source 177 may not be as effective as the traditional use of hydrogenation. Notwithstanding this however, blending PUFA-rich oils with butter oil, or alternatively with natural antioxidants, does minimize the otherwise generation of trans-fatty acids that occur with hydrogenation and which also represent a primary health concern for the consumer. Future studies that involve similar experiments reported in this thesis but under conditions of ambient temperature and long term storage are needed to further evaluate this approach. 178 0 1 2 3 4 5 6 7 8 9 10 Area Ratio (C12U / C17:0) Appendix figure A . Calibration curves for individual fatty acids (to be continued). Area Ratio (C226/C17:0) Appendix figure A. Calibration curves for individual fatty acids. 0.025 0.020 0.015J1 o > 0.010 0.005 .3 12 10 11 0.000-t-6 1617 13 14 15 23 l"f> . •'>• 31 24 25 27 2cTi829 30 35 32 33 34 37 10' 15 20. 25* 30 " 35! 40 45 50 Retention time (min) Appendix figure B . Mixture standards of fatty acid methyl esters. 1= butyric acid (C4:0), 2= caproic acid (C6:0), 3= caprylic acid (C8:0), 4= capric acid (C10:0), 5= undecanoic acid (Cl 1:0), 6= lauric acid (C12:0), 7= tridecanoic acid (C13:0), 8= myristic acid (C14:0), 9= myristoleic acid (C14:l, cis-9), 10= pentadecanoic acid (C15:0), 11= cis-10-pentadecenoic acid (C15:l), 12= palmitic acid (C16:0), 13= palmitoleic acid (C16:l, cis-9), 14= heptadecanoic acid (17:0), 15= cis-10-heptadecenoic acid (C17:l), 16= stearic acid (C18:0), 17= oleic acid (C18:l, cis-9), 18= elaidic acid (C18:l, trans-9), 19= linoleic acid (C18:2, cis-9,12), 20= linolelaidic acid (C18:2, trans-9,12), 21= linolenic acid (C18:3, cis-9,12,15), 22= Y-linolenic acid (C18:3, cis-6,9,12), 23= arachidic acid (C20:0), 24= cis-11-eicosenoic acid (C20:l), 25= cis-11,14-eicosadienoic acid (C20:2), 26= cis-11,14,17-eicosatrienoic acid (20:3), 27= cis-8,ll,14-eicosatrienoic acid (20:3), 28= arachidonic acid (C20:4, cis-5,8,11,14), 29= cis-5,8,11,14.17-eicosapentaeaoic acid (C20:5), 30= heneicosanoic acid (C21:0), 31= behenic acid (C22:0), 32= erucic acid (C22:l, cis-13), 33= cis-13,16-docosadienoic acid (C22:2), 34= cis-4,7,10,13,16,19-docosahexaenoic acid (C22:6), 35= tricosanoic acid (C23:0), 36= lignoceric acid (C24:0), 37= nervonic acid (C24:l, cis-15) methyl esters. 181 v Content of fish oil (%) V, Appendix figure C-l. Experimental regression curve of P/S ratio and fish oil content in the butter-fish oil blends. 182 Content of flaxseed oil (%) Appendix figure D-l. Experimental regression curve of P/S ratio and flaxseed oil content in the butter-flaxseed oil blends. 0.2 0.3 0.4 0.5 0.6 0.7 Content of flaxseed oil (%) Appendix figure D-2. Experimental regression curve of co3/u)6 ratio and flaxseed oil content in the butter-flaxseed oil blends. 183 0.50 -I . . . 0 8 16 24 Time (h) Appendix figure E. The retention of PUFAs from fish oil or flaxseed oil following thermal treatment for 8, 16, and 24 h at 150°C, respectively. • eicosapentaenoic acid (C20:5co3) from fish oil, • docosahexsaenoic acid (C22:6co3) from fish oil, X linolenic acid (C18:3co3) from flaxseed oil, x linoleic acid (C18:2u)6) from flaxseed oil. 184 0.30 4-0 8 16 — | 24 Time (h) Appendix figure F. The retention of PUFAs from fish oil or flaxseed oil following thermal treatment for 8,16, and 24 h at 180°C, respectively. • eicosapentaenoic acid (C20:5co3) from fish oil, • docosahexaenoic acid (C22:6co3) from fish oil, X linolenic acid (C18:3co3) from flaxseed oil, x linoleic acid (C18:2co6) from flaxseed oil. 185 186 HO-«-C5 O-rV-O i (2.4%) GH3-(CH2)3-CH3^/ (45%) CH3-(CH2)4-CHO! y - H 3 (1.1 %) CH3-(CH2)3-CH=CH-CHO • O H ( C H ^ - c o o M e OHC-(CH2)7-COOMe (3.8%) OHC-C=C-(CH2)6-COOMe (29%) i O T T O I i \ O + O H ( c f * 2 > 6 — C O O M E C H A - ( C H A ) 3 . (2.5%) CH3-(CH2)4-CHO ^ \ CH3-(CH2)6-COOMe (5.0%) (27%) CH3-(CH2)3-CH=CH-CHO_%A OHC-(CH2)7-COOMe (39%) / - o 2 1 X ' OHC-C=C-(CH2)6-COOMe (2.6%) (4.9%) CH3-(CH2)3-CH=CH-C-CHO O Appendix figure H. Thermal decomposition of hydroperoxy cyclic peroxidation from linoleate treated with l02 (Frankel, 1984). 188 0.0012 Appendix figure J. A representative chromatogram of derivatized cholesterol oxide Standards. 1= 5a-cholestane, 2= 3,5-cholestadien-7-one, 3= cholesterol, 4= 5-cholesten-3p\7a-diol, 5= cholestan-5P,6p-epoxy-3p-ol, 6= cholestan-5a,6a-epoxy-3f3-ol. 7= 5-cholesten-3B,7p-diol, 8= 5-cholesten-3P,4p-diol, 9= cholestan-3P^a,6P-triol , 10= 5-cholesten-3P-ol-7-one, 11= 5-cholesten-3f3,25-diol. 189 190 191 192 References Ackman, R.G. 1972. Porous polymer bead packings and formic acid vapor in the GLC of volatile free fatty acids, J. Chromatogr. Sci, 10: 560-565. Ackman, R.G. 1992. Application of gas-liquid chromatography to lipid separation and analysis: qualitative and quantitative analysis. In: Fatty Acids in Foods and Their Health Implications, edited by Chow, C.K., Marcel Dekker Inc, New York, pp. 47-64. Allen, C.E. and Foegeding, E. A. 1981. Some lipid characteristics and interaction in muscle foods - review. Food Technol. 35: 253-257. Ansari, G.A.S. and Smith, L.L. 1994. Assay of cholesterol autoxidation. Methods in Enzymology. 233: 332-338. AOAC. 1980. Peroxide value - official final action. Official Methods for Analysis of the Association of Official Analytical Chemists. AOAC, Washington, D.C. pp. 440-441. Aruoma, O.I., Halliwell, B., Aeschbach, R. and Loliger, J. 1992. Antioxidant and prooxidant properties of active rosemary constituent: carnosol and carnosic acid. Xenbiotica. 22: 257-268. Aruoma, O.I., Spencer, J. P.E., Rossi, R., Aeschbach, R., Khan, A., Mahmood, N., Munoz, A., Murcia, A., Butler, J. and Halliwell, B. 1996. An evaluation of the antioxidant and antiviral action of extracts of rosemary and Provencal herbs. Food Chem Tox. 34: 449-456. Augustin, M.A. and Berry, S.K. 1983. Effectiveness of antioxidants in palm olein during heating and frying. J.AOCS. 60: 105-107. Ballantyne, C M . 1998. Low-density lipoproteins and risk for coronary artery disease. Am. J. of Cardiology. 82: 3-12. 193 Banias, C , Orepoulou, V. and Thompoulos, CD. 1992. The effect of primary antioxidants and synergists on the activity of plant extract in lard. J.AOCS. 69: 520-524. Baumgartner W.A., Baker, N., Hill, V.A. and Wright E.T. 1975. Novel interference in thiobarbituric acid assay for lipid peroxidation. Lipids. 10: 309-311. Bereza, U.L., Brewer, G.J. and Hill, G. 1987. Effect of dietary cholesterol on the erythrocyte peroxidant stress in vitro and in vivo. Biochim. Biophys. Acta. 835: 434-440. Bergstrom, S. and Wintersteiner, O. 1941. Autoxidation of sterols in colloidal aqueous solution. The nature of the products formed from cholesterol. J. Biol. Chem. 141: 597-610. Bernsohn, J. and Spitz, F.J. 1974. Linoleic and linolenic acid dependency of some brain membrane-bound enzymes after lipid deprivation in rats. Biochem. Biophys. Res Comm. 57: 293-298. Berry, S.K. and Awang, CR. 1983. Physico-chemical characteristics of palm olein and soybean oil blends. In: Proceeding of International Conference on Palm Oil Product Technology in the Eighties, pp. 483-498. Bhadra, S., Banavali, S.D., Agrawal, M. and Subbiah, M.T.R. 1993. Cholesterol peroxidation potential as influenced by dietary fat type. Int. J. Vit. Nutr. Res. 63: 223-228. Bhatty, R.S. and Rowland, G.G. 1990. Measurement of cc-linolenic acid in development of edible oil flax. J.AOCS. 67: 364-367. 194 Blankenship J.W., Van Gent C.M., Sandberg L.B., Roos P.J. and Scharffenberg J.A. 1991. Oxysterol incorporation into rat aorta resulting in elastin compositional change. Lipids. 26: 381-394. Blois, M.S. 1958. Antioxidant determinations by the use of a stable free radical. Nature. 181: 1199-1200. Brand-Williams, W., Cuvelier, M.E. and Berset, C. 1995. Use of a free radical method to evaluate antioxidant activity. Lebensm-Wiss. U.-Technol. 28: 25-30. The British Nutrition Foundation. 1992. Recommendations for intakes of unsaturated fatty acids. In: The Report of the British Nutrition Foundation's Task Force. Chapman and Hall, New York, 152-163. Burr, M.L., Fehily, A.M. and Gilbert, J.F. 1989. Effects of changes in fat, fish, and fibre intakes on death and myocardial reinfarction: diet and reinfarction trial (DARF). Lancet, 2: 757-761 Caroll, K.K. and Braden, L. 1986. Differing effects of dietary polyunsaturated vegetable and fish oil of mammary tumorigenesis. Prog. Lipid Res. 25: 583-585. Chan, S.H., Gray, J.I. and Gomaa, F.A. 1993. Cholesterol oxidation in whole milk powders as influenced by processing and packaging. Food Chem. 47: 321-328. Chang, S.S., Ostric-Matijasevic, B., Heish, O L . and Huang, C.L. 1977. Natural antioxidants from rosemary and sage. J. Food Sci. 42, 1102-1106. Chapkin R.S, 1992, Reappraisal of the essential fatty acids. In: Fatty Acids in Foods and Their Health Implications. Edited by Chow, C.K., Marcel Dekker Inc, New York. pp. 429-436. Che Man, Y.B. and Jaswir, I. 2000. Effect of rosemary and sage extracts on frying performance of refined, bleached and deodorized (RBD) palm olein during deep-fat frying. Food Chem. 69: 301-307. Chen, J.S. and Yen, G.C. 1994. Determination of cholesterol oxides in heated hard by liquid chromatography. Food Chem. 50:53-58. Chen, Q., Shi, H. and Ho, C T . 1992. Effects of rosemary extracts and major constituents on lipid oxidation and soybean lipoxygenase activity. J. AOCS. 69: 999-1002. Chen, Y . C , Chiu, CP. and Chen, B.H. 1994. Determination of cholesterol oxides in heated lard by liquid chromatography. J. Food Chem. 53-58. Chen, Z.Y., Ratnayake, W.M.N, and Cunnane, S.C 1994. Oxidative stability of flaxseed lipids during baking. J.AOCS. 71: 629-632. Chen, Z.Y., Chan, P.T., Ma, H.M., Fung, K.P., Wang, J. 1996. Antioxidative effect of ethanol tea extracts on oxidation of canola oil. J. AOCS. 73: 375-380. Chien, J.T., Wang, H.C. and Chen, B.H. 1998. Kinetic model of the cholesterol oxidation during heating. J. Agric. Food Chem. 46: 2572-2577. Chipault, J.P. 1962. In: Autoxidation and Antioxidants, edited by Lundberg, W.O. Vol. 2. Wiley -Interscience, New York, pp. 477-540. Chipault, J.R., Mizuno, G.R., Hawkins, J.M. and Lundberg, W.O. 1952. The antioxidant properties of natural spices. Food Res. 17: 46-55. Chipault, J.R., Mizuno, G.R. and Lundberg, O.W. 1956. The antioxidant properties of spices in foods. Food Tech. 10: 209-211. Cho, F., de Man, FM. and Allen, O.B. 1996. Application of Simplex-Centroid design for the formulation of partially interesterified canola/palm blends. J. Food Lipids. 1:25-52. Cho, S., Miyashita, K., Miyashita, T., Fujimoto, K. and Kaneda, T. 1987. Autoxidation of ethyl eicosapentaenoate and docosahexaenoate. J.AOCS. 64: 876-879. Chow, C K . 1988. Interrelationship of cellular antioxidant defense systems. In: Cellular Antioxidant Defense Mechanisms, Vol.2, edited by Chow, C.K., CRC, Boca Raton, pp. 217-253. Chow, C K . 1992. Biological effects of oxidized fatty acid. In: Fatty Acids in Foods and Their Health Implications, edited by Chow, C.K., Marcel Dekker Inc, New York, pp. 689-706. Clare, K., Hardwick, S.J., Carpenter, K.L., Weeratunge, N. and Mitchinson, M.J. 1995. Toxicity of oxysterols to human monocyte-macrophages. Atherosclerosis. 118: 67-75. Clemens, M.R., Ruess, M., Bursa, Z. and Waller, H.D. 1987. The relationship between lipid composition of red blood cells and their susceptibility to lipid peroxidation. Free Rad Res. Comm. 2: 265-271. Cosgrove, J.P., Church, D.F. and Pryor, W.A. 1987. The kinetics of the autoxidation of polyunsaturated fatty acids. Lipids. 22: 299-304. Coupland, J.N. and McClements, D.J. 1996. Lipid oxidation in food emulsions. Trends in Food Sci. and Tech. 7: 83-91. Cuppett, S.L. and Hall, C.A.III. 1998. Antioxidant activity of the Labiatate. Advances in Food and Nutrition Research. 42: 245-271. Cuvelier, H.E., Berset, C , Richard, H. 1994. Antioxidant constituents in sage (Salvia officinals). J. Agric. Food Chem. 42: 1665-1669. Deighton, N., Glidewell, S.M., Deans, S.G. and Goodman, B.A. 1993. Identification by EPR spectroscopy of carvacrol and thymol as the major sources of free radicals in the oxidation of plant essential oils. J. Sci. Food Agric. 63: 221-225. De Man, J.M., Tie, J. and de Man, I. 1987. Formation of short chain volatile organic acids in the automated AOM method. J.AOCS. 64: 993-996. Devaraj, S. and Jialal, I. 1996. Oxidized low-density lipoprotein and atherosclerosis. Inter. J. of Clinical & Laboratory Research. 26: 178-184. Dixon, L.B. and Ernst, N.D. 2001. Choose a diet that is low in saturated fat and cholesterol and moderate in total fat: Subtle changes to a familiar message. Journal of Nutrition. 131:510-526. Espfn, J.C, Soler-Rivas, C. and Wichers, H.J. 2000. Characterization of the total free radical scavenger capacity of vegetable oils and oil fractions using 2,2-diphenyl-l-picrylhydrazyl radical. J. Agric. Food Chem. 48: 648-656. Evans, CD. , Frankel, E.N. and Cooney, P.M. 1959. Tocopherol oxidation in fats. Hydrogenated soybean oil. J. AOCS. 36: 73-77. Fang, X. and Wada, S. 1993. Enhancing the antioxidant effect of a-tocopherol with rosemary in inhibiting catalyzed oxidation caused by Fe 2 + and hemoprotein. Food Res. Int. 26:405-411. FAO/WHO. 1993. Report, fats and oils in human nutrition. Rome, 49-55. 198 Farag, R.S., Badei, A.Z.M.A., Hewedi, F.M. and El-Baroty, G.S.A. 1989. Antioxidant activity of some spice essential oils on linoleic acid oxidation in aqueous media. J.AOCS. 66: 792-799. Fennema, O.R. 1985. In: Food Chemistry, Fennema, O.R. editor. Mecel Dekker Inc., New York, pp 196-221. Fritsch, CW. 1981. Measurement of frying fat deterioration: a brief review. J. AOCS. 58: 272-274. Frankel, E.N. 1984. Lipid oxidation: mechanisms, products and biological significance. J.AOCS. 61: 1908-1917. Frankel, E.N. and Gardner, H.W. 1989. Effect of alpha-tocopherol on the volatile thermal decomposition products of methyl linoleate hydroperoxides. Lipids 24: 603-608. Frankel, E.N. 1991. Recent advances in lipid oxidation. J. Sci. Food Agric. 54: 495-511. Frankel, E.N. 1993. In search of better method to evaluate natural antioxidant and oxidative stability in food lipids. Trends in Food Sci. and Tech. 4: 220-225. Frankel, E.N. 1994. Methods of evaluating food antioxidant. Reply. Trends in Food Sci. and Tech., 5: 57. Frankel, E.N., Huang, S.W., Aeschbach, R. and Prior, E. 1996. Antioxidant activity of a rosemary extract and its constituents, carnosic acid, carnosol and rosmarinic acid in bulk oil and oil-in-water emulsion. J. Agric. Food Chem. 44: 131-135. Gardner, H.W. 1989. Oxygen radical chemistry of polyunsaturated fatty acids. Free Rad. Biol. Med. 7: 65-86. 199 German, J.B. 1999. Food processing and lipid oxidation. Advances in Experimental Medicine & Biology. 459: 23-50. Glavind, J. and Sylven, C. 1970. Intestinal absorption and lymphatic transport methyl linoleate hydroperoxide and hydroxyoctadecadienoate in the rats. Acta Chem. Scand. 24: 3723-Glomset, J.A. 1985. Fish, fatty acids, and human health. New Eng. J. Med. 312: 1253-1254. Gordon, M.H. 1990. The mechanism of antioxidant involved in chelation of metal ions action in vitro. In: Food Antioxidant, edited by Hudson, B.J.F. Elsevier Applied Sci. New York, pp 1-18. Gray, J.I. 1978. Measurement of lipid oxidation: a review. J.AOCS. 55: 539-546. Gray, J.I. and Pearson, A.M. 1987. Rancidity and warmed-over flavor. In: Advances in meat research: restructured meat and poultry products, edited by Pearson, A. M. and Duston, T.R. Van Nostrand Reinhold, New York, pp 219-229. Gray, J.L and Monahan, J. 1992. Measurement of lipid oxidation in meat and meat products. Trends Food Sci Technol. 3: 320-324. Grundy, S.M. and Denke, M.A. 1990. Dietary influences on serum lipids and lipoproteins. J. Lipids Res. 31: 1149-1172. Grossman, A. and Wendel, A. 1983. Non-reactivity of the selenoenzyme glutathione peroxidase with enzymatically peroxidized phospholipids. Eur. J. Biochem. 135: 549-552. 200 Gunstone, F.D. 1996. Analytical procedures. In: Fatty Acid and Lipid Chemistry. Chapman & Hall, London, pp 83-84, 100-128. Guzinski, J. 1996. Oleoresins and essential oils. In: Bailey's Industrial Oil and Fat Products. Vol 1, edited by Y.H. Hui. Wiley-Interscience Publication, New York, pp. 145-158. Hale, M.B. and Brown, T. 1983. Fatty acid and lipid classes of three underutilized species and changes due to canning. Mar. Fish Rev. 45: 4-6. Hall, C A III. and Cuppett, S.L. 1997. Structure-activities of natural antioxidants. In: Antioxidant methodology: in vivo and in vitro concepts, edited by Auroma, O.I. and Cuppett, S.L., AOCS press, Champaign, IL. pp 141-172. Hall, J.L., Harrison, D.L., and Mackintosh, D.L. 1962. Counter effect of sodium chloride and sage on development of peroxide in frozen stored sausage. Food Tech. 16: 102-104. Halliwell, B., Gutteridge, J.M.C. and Aruoma, O.I. 1987. The deoxyribose method: a simple "test-tube" assay for determination of rate constants for reactions of hydroxyl radicals. Anal. Biochem. 165: 215-219. Halliwell, B., Aeschbach, R., Loliger, J. and Aruoma, O. I. 1995. The Characterization of antioxidants. Food and Chemical Toxicology. 33: 601-617. Hamama, A.A. and Nawar, W.F. 1990. Thermal decomposition of some phenolic antioxidants. J. Agric. Food Chem. 39: 1063-1069. Haumann BF. 1993. Linola approved for food use in Canada. Inform. 4: 907. 201 Hawrysh, Z.J., Erin, M.K., Kim, S.S. and Hardin, R.T. 1995. Sensory and chemical stability of tortilla chips fried in canola oil, corn oil and partially hydrogenated soybean oil. J.AOCS. 72: 1123-1130. Henika, R.G. 1982. Use of response surface methodology in sensory evaluations. Food Technol. 37: 96-101. Horrock, L.A. and Yeo, Y.K. 1999. Health benefits of docosahexaenoic acid. Pharmacological Res. 40: 211-225. Houlihan, C M . , Ho, C T . and Chang, S.S. 1984. Elucidation of the chemical structure of a novel antioxidant rosmaridiphenol, isolated from rosemary. J. AOCS. 61: 1036-1039. Hu, C. and Kitts, D.D. 2000. Antioxidant activity of Echinacea root extracts. J. Agric. Food Chem. 48: 1466-1472. Huang, S.W., Hopia, A., Schwarz, K., Frankel, E.N. and German, J.B. 1996. Antioxidant activity of oc-tocopherol and trolox in different lipid substrates: bulk oils vs oil-in-water emulsions. J. Agric. Food Chem. 44: 444-452. Huisman, M., Lindberg, M. H., Skibsted, L.H. and Bertelsn, G. 1994. The combined effect of rosemary (Rosmarinus officinalis L.) and modified atmosphere packaging as protection against warmed over flavor in cooked minced meat. Z. Lebensmittel. Untersuchung Forsch. 198: 57-59. Hummell, D. S. 1993. Dietary lipids and immune function. Prog. Food and Nutri. Sci. 17: 287-329. 202 Haryati, T., Che Man, Y.B., Jinap, S., Ghazali, H.M., and Buana, L. 1999. Optimization of chemical transesterification of palm oil using response surface methodology. J. Food Lipids. 6: 91-106. Huang, K.H., Akoh, C C . and Koehler, P.E. 1996. Optimization and scale-up of enzymatic synthesis of structured lipids using RSM. J. Food Sci. 61: 137-141. Illingworth, D.R. 1993. The potential role of antioxidants in the prevention of atherosclerosis. J. Nutr. Sci. Vit. 39: S43-S47. Imai, H., Werthessen, N.T., Taylor, C B . and Lee, K.T. 1976. Angiotoxicity and arteriosclerosis due to the contaminants of FJSP-grade cholesterol. Arch. Pathol. Lab. Med. 100: 565-572. Inatani, R., Nakatani, N. and Fuwa, H. 1983. Antioxidant effects of the constituents of rosemary (Rosmarinus officinalis L.) and their derivatives. Agric. Biol. Chem. 47: 521-528. Irwandi, J., Che Man, Y.B., Kitts, D.D., Bakar, J. and Jinap, S. 2000. Synergies between plant antioxidant blends in preventing peroxidation reactions in model and food oil systems. J.AOCS. 77: 945-950. Iverson, F. 1995. Phenolic antioxidants: Health Protection Branch studies on butylated hydroxyanisole. Cancer Letters. 93: 49-54. Jaswir, I., Che Man Y.B. and Kitts, D.D. 2000. Use of natural antioxidants in refined palm olein during repeated deep-fat frying. Food Res. International 2000: 1-8 Jensen, R.G. 1992. Fatty acids in milk and dairy products. In: Fatty Acids in Foods and Their Health Implication, edited by Chow, C K . Marcel Dekker Inc. New York, pp 95-135. Jensen, R.G., Ferris, A.M., Lammi-Keefe, CJ . and Henderson, R.A. 1990. Lipids of bovine and human milks: a comparison. J. Dairy Sci. 73: 223-228. Jensen, R.G., Ferris, A.M., and Lammi-Keefe, CJ . 1991. The composition of milk fat, J. Dairy Sci. 74: 3228-3243. Kahl, R. and Hildebrandt, A.G. 1986. Methodology for studying antioxidant activity and mechanisms of action of antioxidants. Food Chem. Toxic. 24: 1007-1014. Kanazawa, K., Kanazawa, E. and Natake, M. 1985. Uptake of secondary autoxidation products of linoleic acid by the rat. Lipids. 20: 412-419. Kanematsu, H., Aoyama, M., Maruyama,T., Niiya, I., Tsukamoto, M. Tokairin, S. and Matsumoto, T. Studies on the improvement of antioxidant effect of tocopherols. II. Synergistic effects of citric, malic and succinic acids. J. Japan Oil Chem. Soc. 1983: 32: 695-699. Ke, P.J. and Woyewoda, A.D. 1979. Microdetermination of thiobarbituric acid value in marine lipids by a direct spectrophotometric method with mono-phasic reaction system. Anal Chim. Acta. 106: 279-284. Kim, M., Matoba, T. and Hasegawa, K. 1988.Thermal oxidation stability of interesterified oils under continuous heating conditions. Agric. Biol. Chem. 52: 1239-1244. 204 Kim, S.K. and Nawar, W.W. 1991. Oxidative interaction of cholesterol with triacylglycerols. J.AOCS. 68: 931-934. Kim, S. K. and Nawar, W.W. 1993. Parameters influencing cholesterol oxidation. Lipids. 28:917-922. Kitts, D.D. 1996. Toxicology and safety of fats and oils, In: Bailey's Industrial Oil and Fat Products, V. edited by Hui, Y.H. Wiley-Interscience Publication., New York, pp 215-280. Kitts, D.D. 1997. An evaluation of the multiple effects of the antioxidant vitamins. Trends in Food Sci. and Tech. 8: 198-203. Kitts, D.D., Yuan, Y.V., Wijewickreme, A.N. and Thompson, L.U. 1999. Antioxidant activity of the flaxseed lignan secoisolariciresinol diglycoside and its mammalian lignan metabolites enterodiol and enterolactone. Mol. Cell. Biochem. 202: 91-100. Kneepkens, C.M.F. Assessment of oxidative stress and antioxidant status in humans: the hydrocarbon breath test. In: Antioxidant Methodology, In Vivo and In Vitro Concepts, edited by Aruoma, O.I. and Cuppett, S.L. AOCS Press, Champaign, IL., pp 23-38. Kochhar, S.P., and Rossell, J.R. 1990. Detection, estimation and evaluation of antioxidants in food system. In: Food Antioxidant, edited by Hudson, B.J.F., Elservier Applied Science Publishers, Ltd., Borking, UK. pp 19-64. Kohlmeier, L., Simonsen, N. and Mottus, K. 1995. Dietary modifiers of carcinogenesis. Environ. Health Persp. 103: 177-184. 205 Korahani, V., Bascoul, J. and Crastes de Paulet, A. 1982. Autoxidation of cholesterol fatty acid esters in solid states and aqueous dispersion. Lipids. 17: 703-708. Kuller LH. 1997. Dietary fat and chronic diseases: epidemiologic overview. J. American Dietetic Association. 97(7 Suppl):S9-15. Kupranycz, D.B. Amer, M.A. and Baker, B.E. 1986. Effects of thermal oxidation on the constitution of butterfat, butterfat fractions and certain vegetable oils. J. AOCS. 63: 332-337. Lake, R.J. and Scholes, P. 1997. Consumption of cholesterol oxides from fast foods fried in beef fat in New Zealand. J. AOCS. 74: 1069-1075. Li, S. X., Cherian, G. and Sim, J. S. 1996a. Cholesterol oxidation in egg yolk powder during storage and heating as affected by dietary oils and tocopherol. J. Food Sci. 61: 721-725. Li, S.X., Cherian, G., Ahn Du., Hardin, R.T. and Sim, J.S. 1996b. Storage, heating, and tocopherols affect cholesterol oxide formation in food oils. J. Agric. Food Chem. 44: 3830-3834. Lolos, M., Oreopoulous, V. and Tzia, C. 1999. Oxidative stability of potato chips: effects of frying oil type, temperature and antioxidants. J. Sci. Food Agric. 79: 1524-1528. Luby, J.M., Gray, J.I. and Harte, B.R.. 1986. Effects of packaging and light source on the oxidative stability of cholesterol in butter. J. Food Sc. 51: 908-911 Luthria, D.L., Mohammed, B.S. and Sprecher, H. 1996. Regulation of the biosynthesis of 4,7,10,13,16,19-docosahexaenoic acid. J. Biol. Chem. 271: 16020-16025. 206 Marcuse, R. and Johansson, L. 1973. Studies on the TBA test for rancidity grading: II TBA reactivity of different aldehyde classes. J.AOCS. 50: 387-391. Marinova, E.M. and Yanishilieva, N.V. 1997. Antioxidative activity of extracts from selected species of the family Lamiaceae in sunflower oil. Food Chem. 58: 245-248. Marquez-Ruiz, G. and Dobarganes, C. 1996. Short chain fatty acid formation during thermaoxidation and frying. J. Sci. Food Agric. 70: 120-126. Marshall, A. and Van Elswyk, M. E. 1995. Dietary polyunsaturated fats, is it beneficial? Nutr. Today. 30:207-213. Mattson, F.H. and Grundy, S.M. 1985. Comparison of effect of dietary saturated, monounsaturated and polyunsaturated fatty acids on plasma lipids and lipoprotein in man. J. Lipid Res. 26: 194-202. McGregon, W.G. and Carson, R. B. 1961. Fatty acid composition of flax varieties. Can. J. Plant Sci. 41: 814-818. McLennan, P.L., Howe, P. and Abeywardena, M. 1996. The cardiovascular protective role of docosahexaenoic acid. Eur. J. Pharmacol. 300: 83-89. Mead, J. F. 1984. The non-eicosanoid function of the essential fatty acids. J. Lipid Res. 25: 1517-1521. Medina, I., Satue-Gracia, S.T., German, J.B. and Frankel, E.N. 1999. Comparison of natural polyphenol antioxidants from extra virgin olive oil with synthetic antioxidants in tuna lipids during thermal oxidation. J. Agric. Food Chem. 47: 4873-4879. Melton, S.L., Trigiano, M.K., Penfield, M.P. and Yang, R. 1993. Potato chips fried in carnola and/or cottonseed oil maintain high quality. J. Food Sci. 58: 1079-1083. Mensink, R.P., Temme, E.H.M. and Plat, J. 1997. Dietary fats and coronary heart disease. In: Food Lipids-Chemistry, Nutrition and Biotechnology. Edited by Akoh, C C . and Min, D.B., Marcel Dekker Inc., New York, pp. 507-536. Meyer, A.S. 1994. Methods of evaluating food antioxidants. Trends in Food Sci. and Tech. 5: 56-57. Miller, L.A. and White, P.J. 1988. High temperature stability of low-linoleic, high stearate and common soybean oils. J. AOCS. 16: 23-28. Miller, N.J. 1998. Flavonoids and phenylpropanoids as contributors to the antioxidant activity of fruit juices. In: Flavonoids in Health and Diseases, edited by Rice-Evans C and Pakcer L. Marcel Dekker Inc. New York, pp 387-403. Min, D.B. 1997. Lipid oxidation of edible oil. In: Food Lipids- Chemistry, Nutrition and Biotechnology, edited by Akoh, C C and Min, D.B., Marcel Dekker Inc. New York, pp 283-297. Miyashita, K., Nara, E. and Ota, T. 1993. Oxidative stability of polyunsaturated fatty acids in aqueous solution. Biosci. Biotechnol. Biochem. 57: 1638-1640. Miyashita, K., Nara, E. and Ota, T. 1994. Comparative study on the oxidative stability of phosphatidylcholines from salmon egg and soybean in aqueous solution. Biosci. Biotechnol. Biochem. 57: 1638-1640. Monahan, F.J., Gray, J.L., Booren, A.M., Miller, E.R., Buckley, D.J., Morrissey, P.A. and Gomaa, E.A. 1992. Inflence of dietary treatment on lipid and cholesterol oxidation in pork. J. Agric Food Chem. 40: 1310-1315. 208 Nakatani, N, and Inatani, R. 1981. Structure of rosmanol, a new antioxidant from rosemary (Rosmarinus officinalis L). Agric. Biol. Chem. 47:2385-2386. Nelson, G.J. and Ackman,R.G. 1988. Absorption and transport of fat in mammals with emphasis on n-3 polyunsaturated fatty acids. Lipids. 23: 1005-1014. Newton, I. and Snyder, D. 1997. Nutritional aspects of long-chain omega-3 fatty acids and their use in bread enrichment. Cereal Food World. 42: 126-131. Nielsen, J.H., Olsen, C.E. and Skibsted, L.H. 1996. Cholesterol oxidation in a heterogeneous system initiated by water-soluble radicals. Food Chem. 56: 33-37. Niki, E. 1996. cc-Tocopherol. In: Handbook of antioxidants, edited by Cadenas E and Packer L. Marcel Dekker Inc. New York, p3-25. Odeh, R.M. and Cornish, L.A. 1995. Natural antioxidants for the prevention of atherosclerosis. Pharmacotherapy. 15: 648-659. Oomah, B.D. and Mazza, G. 1999. Health benefits of phytochemicals from selected Canadian crops. Trends in Food Sci. Tech. 10: 193-198. Osada, K., Kodama, T., Yamada, K. and Sugano, M. 1993. Levels and formation of oxidized cholesterols in processed marine foods. J. Agric. Food Chem. 41: 1893-1898. Oshima, T., Li, N. and Koizumi, C. 1993. Oxidative decomposition of cholesterol in fish products. J.AOCS. 70: 595-600 Paniangvait, P., King, A.J. and Jones, B.G. 1995. Cholesterol oxides in foods of animal original. J.Food Sci. 60:1159-1174. 209 Parassi, T., Ravagnan, G., Sapora, O. and Gratton, E. 1992. Membrane oxidative damage induced by ionizing radiation detected by di-phenylhexatriene fluorescence life time distributions. Int. J. Radiat. Biol. 69: 791-796. Parasassi, T., Giusti, A.M., Raimondi, M., Ravagnan, G., Sapora, O. and Gratton E. 1995. Cholesterol protects the phospholipid bilayer from oxidative damage. Free Rad. Biol Med. 19: 511-516. Park, S.W. and Addis, P.B. 1985. HPLC determination of C-7 cholesterol derivatives in foods. J. Food Sci. 50: 1437-1441, 1444. Park, S.W. and Addis, P.B. 1986a. Identification and quantitative estimation of oxidized cholesterol derivatives in heated tallow. J. Agric. Food Chem. 34: 653-659. Park, S.W. and Addis, P.B. 1986b. Further investigation of oxidized cholesterol derivatives in heated fats. J. Food Sci.51: 1380-1381 Park, S. W. and Addis, P. B. 1987. Cholesterol oxidation products in some muscle foods. J. Food Sci. 52: 1500-1503. Pekkarinen, S.S., Stockmann, H., Schwarz, K., Heinonen, I. M. and Hopia, A.I. 1999. Antioxidant activity and partitioning of phenolic acids in bulk and emulsified methyl linoleate. J. Agric. Food Chem. 46: 3036-3043. Peng, S. K., Tham, P.,Taylor, C. B. and Mikkelson, B. 1979. Cytotoxicity of oxidation derivatives of cholesterol on cultured aortic smooth muscle cells and their effect on cholesterol biosynthesis. Am. J. Clin. Nutr. 32: 1033-1042. 210 Peng, S.K., Taylor, C.B., Mosbach, E.H., Hwang, W.Y., Hill, J. and Mikkelson, B. 1982. Distribution of 25-hydroxycholesterol in phase lipoproteins and its role in atherogensis. Atherosclerosis. 41:395-402. Peuchant, E., Wolff, R., Salles, C. and Jensen, R. 1989. One-step extraction of human erythrocyte lipids allowing rapid determination of fatty acid composition. Anal. Biochem. 181: 341-344. Perez-Camino, M.C., Marques-Ruiz, G., Mendez-Ruiz, M.V. and Dobarganes, M.C. 1991. Lipid changes during frying of frozen prefried food. J. Food Sci. 56: 1644-1647. Phillipson, B.E., Rothrock, D.W., Connor, W.E., Harris, W.S. and Illingworth, D.R. 1985. Reduction of plasma lipids, lipoproteins, and apoproteins by dietary fish oils in patients with hypertriglyceridemia. N. Eng. J. Med., 312: 1210-1206 Plumb, G.W., Chambers, S.J., Lambert, N., Bartolome, B., Heaney, R.K., Wanigatunga, S., Auroma, O.I., Halliwell, B. and Williamson, G. 1996. Antioxidant actions of fruit, herb an spice extracts. J. Food Lipids. 3: 171-188. Poling, C. E., Eagle, E., Rice, E E . and Fisher, M. 1970. Long term responses of rats to heat-treated dietary fats. IV. Weight gains, food and energy efficiencies, longevity and histopathology. Lipids 5: 128-136. Pratt, D.E. 1996. Antioxidants: Technical and regulatory consideration. In: Bailey's Industrial Oil and Fat Products, edited by Hui,Y.H. Vol. 3. Wiley-Interscience Publication, New York, pp 523-546. 211 Privett, O.S. and Blank, M.L. 1962. The initial stages of autoxidation. J.AOCS. 39: 465-469. Reddy, K. and Tappel, A.L. 1974. Effect of dietary selenium and autoxidized lipids on glutathione peroxidase system of gastrointestinal tract and other tissues in the rat. J. Nutr. 104: 1069-1078. Richheimer, S.L., Bernart, M. W., King, G. A., Kent, C. and Bailey, D.T. 1996. Antioxidant activity of lipid soluble phenolic diterpenes from rosemary. J.AOCS. 73: 507-514. Rosao Romero, A. J. and Morton, I. D. 1975. A kinetic study of the competitive oxidaiotn of oleic-acid-linoleic acid mixtures. J. Sci. Food Agric. 26: 1353-1356. Schmedes, A. and Holmer, G. 1989. A new thiobarbituric acid method for determining free malonaldehyde and hydroperoxides selectively as a measure of lipid peroxidation. J.AOCS. 66: 813-817. Schuler P. 1990. Natural antioxidants exploited commercially. In: Food Antioxidants, edited by Hudson, N.J.F., Elsevier Applied Sci. New York, pp 99-170. Schwarz, K., Ternes, W. and Schmauderer, E. 1992. Antioxidant constituents of Rosmarinus officinalis and Salvia officinalis. III. Stability of phenolic diterpenes of rosemary extracts under thermal stress as required for technological processes. Z. Lebensmittel. Untersuchung. Forsch. 195: 104-107. Shahidi, F. Rubin, L.J., and Wood, D.F. 1985. Effect of sulfanimide on the TBA and oils. J Food Sci. 50: 274-274. Shahidi, F., Wanasundara, U.N. and Brunet, N. 1994.Oxidative stability of oil from blubber of harp seal (Phoca groenlandica) as assessed by NMR and standard procedures. Food Res Int. 27: 555-562. Shahidi, F. and Naczk, M. 1995. Antioxidant propertity of food phenolic. In: Food Phenolics: Sources, Chemistry Effect, Application. Technomic Publication Co., Lancaster, PA, pp 235-280. Shahidi, F. and Wanasundara, U.N. 1997. Oxidative rancidity in fats and oils. In: Food Lipids- Chemistry, Nutrition and Biotechnology. Edited by C C . Akoh and D.B. Min. Marcel Dekker Inc. New York, pp. 377-396. Shahidi, F., Wanasundara, U.N., He, Y. and Shukla, V.K.S. 1997. Marine lipids and their stabilization with green tea and catechins. In: Flavor and Lipid Chemistry of Seafood, edited by Shahidi, F. and Cadwallader, K. ACS, Washington DC. Shahidi, F. and Wanasundara,U.N., 1998. Omega-3 fatty acid concentrates: nutritional aspects and production technology. Trends in Food Sci. &Tech. 9: 230-240 Sheard, P.R., Enser, M., Wood, J.D., Nute, G.R., Gill, B.P. and Richardson, R.I. 2000. Shelf- life and quality of pork and pork products with raised n-3 PUFA. Meat Sci. 55: 213-221. Sherwin, E. R. 1990. Antioxidants. In: Food Additives, edited by Branen, R. Marcel Dekker. New York, pp 139-193. Shieh, C.J., Akoh, C C and Koehler, P.E. 1995. Four-factor response surface optimization of the enzymatic modification of triolein to structured lipids. J.AOCS. 72: 619-623. 213 Simopoulos, A.P. 1989. Executive summary. In: Dietary co3 and co6 Fatty Acids— Biological Effects and Nutritional Essentiality, edited by Galli, C. and Simopoulos, A.P., Plenum, New York, pp 391-404. Sisovick, D.S., Raghunatham, T.E., King, I. 1995. Dietary intake and cell membrane levels of long-chain n-3 polyunsaturated fatty acids and the risk of primary cardiac arrest. J. Am. Med. Assoc. 274: 1363-1367. Smith, L.L. 1981a. Initial events of autoxidaiton. In: Cholesterol Autoxidation, Plenum Press, New York, pp 125-200. Smith, L.L. 1981b. Biological effects. In: Cholesterol Autoxidation, Plenum Press; New York, pp 369-458. Smith, L.L. 1981c. Cholesterol autoxidation, Plenum Press, New York, p49-123. Smith, L.L. 1987. Cholesterol autoxidation. Chem. Phys. Lipids. 44: 87-125. Smith, L.L. 1991. Another cholesterol hypothesis: cholesterol as antioxidant. Free Rad. Biol. Med. 11:47-61. Smith, L. M., Dunkley, W. L., Franke, A. and Daikiri, T. 1978. Measurement of trans and other isomeric unsaturated fatty acids in butter and margarine. J.AOCS. 52: 257-261. Smith, J.M., Kopp, S.J., Paulson, D.J. and Daat, J.T. 1993. Effects of dietary fish oil on cardiovascular responsiveness to adrenergic agonists in spontaneously hypertensive rat. Can. J. Physiol. Pharmacol. 71: 432-438. Sreenivasan, B. 1978. Interesterification of fats. J.AOCS. 55: 796-805. 214 Stampler, J. 1979. Population studies. In: Nutrition, Lipids and Coronary Disease: a Global View, edited by Levy, R.I., Rifkind, B.M., Dennis, B.Ff. and Ernst, N. Raven Press, New York, pp 25-88. Steinberg, D., Parthasarathy, S., Carew, T.E., Koo, J.C. and Witzum, J.W. 1989. Beyond cholesterol modifications of low-density lipoprotein that increase arthrogenicity. New Eng. J. Med. 320: 915-924. Stevenson, S.G., Vaisey-Genser, M. and Eskin, N.A.M. 1984. Quality control in the use of deep frying oils. J. AOCS. 61: 1102-1108. Swern D. Bailey's Industrial Oil and Fat Products. 4th ed. Wiley and Sons, New York, vol. 1, 1979, pp 424-429. Temme, E.H.M., Mensink, R.P. and Hornstra, G. 1996. Comparison of the effect of diets enriched in lauric, palmitic or oleic acid on serum lipids and lipoproteins in healthy woman and men. Am. J. Clin. Nutr. 63:897-903. Thomas, M.C. and Fume, J. 1987. Application of 2-thiobarbituric acid reaction on exudates of frozen and refrigerated meats. J. Food Sci. 52: 575-579. Thomsen, M. K., Vedetesen, H. and Skibsted L.H. 1999. Quantification of radical formation in electron spin resonance spectroscopy. J. Food Lipids. 6: 149-155. Thomson, L.U. and Aust, R. 1983. Lipid changes in French fries and heated oils during commercial deep-fat frying and their nutritional and toxicological implications. Can. Inst. Food Sci. Technol. J. 16: 23-28. 215 Vilas, N.N., Bell, R.R. and Draper, H.H. 1976. Influence of dietary peroxides, selenium and vitamin E on glutathione peroxidase of the gastrointestinal tract. J. Nutri. 106: 589-596. Wada, S. and Koizumi, C. 1983. Influence of the position of unsaturated fatty acid esterified glycerol on the oxidation rate of triglyceride. J.AOCS. 60: 1105-1109. Wanasundara, U.N., Amarowicz, R. and Shahidi, F. 1996. Partial characterization of natural antioxidants in canola meal. Food Res. Intern. 28: 525-530. Wanasundara, U.N., Shahidi, F. and Jablonski, CR. 1995. Composition of standard and NMR methodologies for assessment of oxidative stability of canola and soybean oils. Food Chem. 52: 249-253. Wang, M.F., Li, J.G., Rangarajan, M., Shao, Y., LaVoie, E.J., Huang, T .C and Ho, C T . 1998. Antioxidative phenolic compounds from sage (Salvia officinalis). J. Agric. Food Chem. 46: 4869-4873. Wang, M.F., Shao, Y., Li, J.G., Zhu, N.Q., Rangarajan, M., LaVoie, E.J. and Ho, C T . 1999. Antioxidative phenolic glycosides from sage (Salvia officinalis). J. Natural Prod., 62: 454-456. Warner, K. 1997. Chemistry of frying oil. In: Food Lipids: Chemistry, Nutrition and Biotechnology, edited by Akoh, C C and Min, D.B. Marcel Dekker Inc., New York, pp 167-180. Wei, G.J., Wang, M.F., Rosen, R.T. and Ho, C T . 1999. Identification of thermal decomposition products of carnosol, an antioxidant in rosemary and sage. J. Food Lipids. 6: 173-179. 216 Weiss, E. 1993. Linola - a new edible oil. Oils Fats Intl. 9: 23-24. Welch, V.A. and Borlakoglu, J.T. 1992. Absorption and transport dietary lipid: effect on some lipid-related health problems. In: Fatty Acids in Foods and Their Health Implications, edited by Chow, C K . Marcel Dekker Inc, New York, pp 559-612. Wenkert, E., Fuchs, A. and McChesney, J.D. 1964. Chemical artifacts from the family of Labiatae. J. Org. Chem. 30: 2931-2934. Westhuyzen, J. 1997. The oxidation hypothesis of atherosclerosis: An update. Annals of Clinical & Laboratory Science. 27: 1-10. Wheeler, B. 1993. Echip-version 6.0 Windows. Reference manual. Hockessin, DE: Echip Inc. White, P.J. 1992. Fatty acids in oilseeds. In: Fatty Acids in Foods and Their Health Implications, edited by Chow, C K . Marcel Dekker Inc. New York, pp 237-262. Wong, J.W., Hashimoto, K. and Shibamoto, T. 1995. Antioxidant activities of rosemary and sage extracts and vitamin E in a model meat system. J. Agric. Food Chem. 43: 2707-2712. Yan, P.S. and White, P.J. 1990. Cholesterol oxidation in heated lard enriched with two level of cholesterol. J. AOCS. 67: 927-931. Yan, P.S. 1999. Cholesterol oxidation products. Their occurrence and detection in our foodstuffs. Advances in Experimental Medicine & Biology. 459:79-98. Yuan, Y.Y., Lipids and Antioxidant status. Ph.D thesis, University of BC, 1995. 217 Yuan, Y.Y., Kitts, D.D. and Godin, D.V. 1997. Influence of dietary cholesterol and fat source on atherosclerosis in the Japanese quail (Coturnix japonica). British J Nutr. 78: 993-1014. Yuan, Y.Y., Kitts, D.D. and Godin, D.V. 1998. Interactive effects of increased intake of saturated fat and cholesterol on atherosclerosis in the Japanese quail (Coturnix japonica). Brit. J. Nutr. 80: 89-100. Zarev, S., Therond, P., Bonnefont-Rousselot, D., Beaudeux, J.L., Gardes-Albert, M. and Legrand, A. 1999. Major differences in oxysterol formation in human low density lipoproteins (LDLs) oxidized by OH/O2'" free radicals or by copper. FEBS Letters. 451:103-108. Zock, P.L, de Vries, J.H.M. and Katan, M.B. 1994. Impact of myristic acid versus palmitic acid on serum lipid and lipoprotein levels in healthy women and men. Arterioscler. Thromb, 14: 567-575. 218 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0090037/manifest

Comment

Related Items