. ‘ h-.. . Ad ' _ ' A532,”. ”if. Vv‘. .. .A, v ( .q 3521:“ ('5: k' “3;: m 1.243.114. ‘1 m...“ , 4 . ‘- M 5‘3“? ~;;»- An—th ' \:.'.;v u - . “.5 u. '.-L {a m. .».r 2 «run I ‘Jx‘ ( I r a K A? .r . "at" x I - .‘ mu . --.. u r. v.. -‘ ” ' V ' r :.-. N I, . » * > , , » AnnA , V - vux . , . 3,. r:- u «m‘ "L uur'l‘ , . . . r. r 'I ‘ -, .V,,_ . .. :I I .1 . :3 ~ ‘ ,, x x J _ . - . . , . ' . r * fir'lri‘ : ' , " ‘ . J I ._.-..~ 1:. «at: ‘ v - .r ' '-" .,.-.. T. - , :4 ‘IF ‘ 1.! sh ‘ :1 l P. . . .. ., .‘ 'vrvr , . Law-My ,. , , ‘ - -»--rr— TL?» ...-. '~ " ‘7.- >uv>r(-r' _ '.’..«’L..’... r» -...._. y ' ’0 ‘ ' ‘ I. I V - , _ .- ' r . u " V " " ' . n- t- _ _ l ' " , , . ’ . -. -- , 4,1,, ".4 ‘ - n ‘ r '1' ”I. x . -. .._-..',.,..‘. m " " " ' ’ ,. "0.... .,-.;. o r— ~r'- lr’ . .. , . . .. r 1 , Wm... . .. .._ . . ...., a, .., .., . w .. V... -9.-.” ‘ ”Jana. -v , . ,. ..... . , u . .r. .. 4"...” -.,.~u... -. .9 ’ ' .-.".';'.’.-".~ . . . u .. ,.. "___ '~'...'.'. . .1. .1... ”up," > ' M..-” M. , .. _ 4 ........... , . P ...4' ~u." .- . . 4 ¢-:-‘:-y- .110 .......a.. . ... .a. '...- V . ,-. ......,.,,..-,.-,r, 2‘ _ .*.. up... ..».u.- 4.“.‘.’..'.."..,- ,- »-- '9': ‘t' . ,.,_;.~ ~ \. \ ll/I/lll/llI/l/llllI/ll/Il/Illl/lllll/llll/ll ' ' 3 129 l E mirfile 30 This is to certify that the thesis entitled CHOLESTEROL OXIDATION IN WHOLE MILK POWDER AS INFLUENCED BY PROCESSING AND PACKAGING presented by Shu-Hui Chan has been accepted towards fulfillment of the requirements for Master of Science degeejn Food Science 7 / / / fl Major professor U L/ 0-7639 MSU is an Affirmative Action/Equal Opportunity Institution LIBRARY Michigan State University PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution chddodmnomts-ni CHOLESTEROL OXIDATION IN WHOLE MILK POWDER AS INFLUENCED BY PROCESSING AND PACKAGING BY SHU-HUI CHAN A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 1992 5 3 67/— ‘ ABSTRACT CHOLESTEROL OXIDATION IN WHOLE MILK POWDER AS INFLUENCED BY PROCESSING AND PACKAGING BY SHU-HUI CHAN This study was designed to examine the effects of various spray—drying processes (direct firing, high levels of oxides of nitrogen (Nox); direct firing, low NOx; indirect electric heating) and packaging systems (polyethylene pouches and crimp-sealed glass vials, with and without oxygen absorbers) on the oxidative stability of lipids, including cholesterol, during the storage of whole milk powders. Lipid oxidation, including the generation of cholesterol oxidation products, was greatest in samples processed by high NOx direct~fired dryers. Oxygen absorbers effectively controlled cholesterol oxidation during the entire storage period, even in those samples from the high NOx drying system. There was a positive correlation (r= +0.89) between the extent of lipid oxidation and cholesterol oxidation in all the samples. It was concluded that the stability of whole milk powder during storage can be increased by using low NOx drying processes and by packaging in oxygen-impermeable packages containing oxygen absorbers. ACKNOWLEDGEMENTS I wish to express my sincere gratitude to my academic advisor, Dr. J. Ian Gray, for his unlimited input, inspiring guidance and consistent encouragement throughout my graduate studies. .Appreciation is extended to my committee, Dr. B. R. Harte, Dr. J. A. Partridge, and Dr. M. R. Bennink, for their technical guidance and critical review of this manuscript. Further acknowledgements are also due to Ms. Shu-Mei Lai and Dr. E. Gomma for their technical assistance. Special thanks are extended to Dr. Phil Kelly of the National Dairy Products Research Centre, Moorepark Research Center in Fermoy, Ireland, for his preparation of the whole milk powders and to Dr. Joe Buckley, University of Cork, Ireland, for his assistance in sending the samples here. My warmest thanks are also extended to Wei—Chuan Food Co., Taiwan, for providing financial support. Finally, I would like to express my heartfelt appreciation to my parents and family for all of the love, without their support and encouragement this degree would not have been so smoothly completed. TABLE OF CONTENTS Page LIST OF TABLES ....................................... V1 LIST OF FIGURES ........................ . ...... . ...... viii INTRODUCTION .......................... ............... 1 REVIEW OF LITERATURE ........................ ......... 3 Methods of Drying ....................... ........ 3 Mechanism of Lipid Oxidation ..... .............. .. 8 Stability of Whole Milk Powder during Storage .... 10 Composition of milkfat . ........ ... ....... . ..... 10 Oxidation of milkfat ........................... 13 Prevention of oxidation by packaging ... ........ 15 Oxidation of Cholesterol ............... ........ .. 18 Description of cholesterol ........... ..... ..... 18 Mechanism of cholesterol oxidation . ....... ..... 20 Biological effects of cholesterol oxides ....... 23 Determination of cholesterol oxides in foods ... 26 Cholesterol Oxidation Products in Foods .... ..... . 29 Cholesterol Oxidation Products in Whole Milk Powder .................................... 31 MATERIAL AND METHODS ... .......................... ..... 35 Cholesterol oxides standards ......... . ....... .... 35 Fatty acids standards ............................ 35 Packaging materials ........................ . ..... 36 Reagents ..................... . ........ . ...... .... 36 Preparation of whole milk powders ............. ... 36 Measurement of lipid oxidation .. ....... . ..... .... 37 Lipid extraction .... ................... . ......... 38 Total cholesterol determination .................. 39 Cholesterol oxide determination .......... ..... .. 40 Fatty acid composition of whole milk lipids ...... 41 Nitrite and nitrate analyses ... ............. ..... 42 Color measurement ... ........ .... ............. .... 43 Headspace oxygen measurement . .................... 43 Statistical analysis ................ . ............ 44 iv Page RESULTS AND DISCUSSION ................................ 45 Nitrite and Nitrate Concentrations in Whole Milk Powders .. ................ . ........ ... ...... ..... 45 Headspace Oxygen Content as Influenced by Packaging Systems ....... ...... .................. 47 Lipid Oxidation in Whole Milk Powder during Storage .......................... ........ 48 Effect of storage conditions on lipid stability in whole milk powders ............. 50 Influence of drying method on lipid oxidation in whole milk powders ....... ...... 56 Influence of packaging on lipid stability ...... 59 Change in fatty acid profiles of whole milk powders during storage ........ ...... ... 60 Total Cholesterol Content of Whole Milk Powders .. 68 Cholesterol Oxidation in Whole Milk Powders ...... 68 Effects of storage conditions on cholesterol stability ............ .. ......... 70 Effects of drying method on cholesterol oxidation .................................. . 80 Effects of packaging on cholesterol oxidation .................................... 85 Correlation of TBARS Values and Cholesterol Oxide Concentrations ...... ......... ............. 89 Color Changes in Whole Milk Powder ..... ...... .... 89 SUMMARY AND CONCLUSIONS ............................... 96 APPENDICES .... ........................................ 99 1. Diagram of a direct low NOx CXA gas burner .... 99 2. TBARS development in whole milk powders processed by different drying methods, when packaged in various packagings and stored for 6 months at 20°C and 40°C ...... 100 REFERENCES .... ....... ..... .............. .. ............ 101 LIST OF TABLES Table Page 1. The major fractions of bovine milk lipids ........ 12 2. The major fatty acids in bovine milkfat ... ....... 12 3. Cholesterol oxidation products in foods .......... 30 4. Headspace oxygen contents of whole milk powders packaged in various packaging systems and held at 20°C and 40°C for 6 months ....... ........ ..... 48 5. The relationship between storage time and TBARS values of whole milk powders processed by different drying methods, packaged in various packaging systems and stored for 6 months at 20°C and 40°C ..................... .. .......... 52 6. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in PE pouches and stored for 6 months at 20°C .. .............................. . 62 7. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in PE pouches and stored for 6 months at 40°C ............ . ........... . ........ 63 8. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in glass vials without oxygen absorbers and stored for 6 months at 20°C ................ .. 64 9. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in glass vials without oxygen absorber and stored for 6 months at 40°C ........... ..... .. 65 10. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in glass vials with oxygen absorbers and stored for 6 months at 20°C .......... . ....... 66 11. Fatty acid profiles of whole milk powders produced by different drying methods, packaged in glass vials with oxygen absorbers and stored for 6 months at 40°C ................ .. 67 vi 12. 13. 14 15. 16. 17. 18. Cholesterol oxide concentrations in whole milk powders packaged in PE pouches and held at two temperatures (20°C and 40°C) for 6 months ........ 71 Cholesterol oxide concentrations in whole milk powders packaged in glass vials without oxygen absorbers and held at two temperatures (20°C and 40°C) for 6 months ..................... 72 Cholesterol oxide concentrations in whole milk powders packaged in glass vials with oxygen absorbers and held at two temperatures (20°C and 40°C) for 6 months .... ................. 73 The relationship of TBARS values and cholesterol oxide concentrations in whole milk powders manufactured by different drying method, when packaged in various packaging systems and held at 20°C and 40°C for 6 months ...... .......... .... 90 Changes in color powders prepared when packaged in 6 months at 20°C Changes in color powders prepared when packaged in and stored for 6 Changes in color powders prepared when packaged in and stored for 6 characteristics of whole milk by different drying methods, PE pouches and stored for and 40°C ............ ............. 91 characteristics of whole milk by different drying methods, glass vials without oxygen absorbers, months at 20°C and 40°C ........... characteristics of whole milk by different drying methods, glass vials with oxygen absorbers, months at 20°C and 40°C . .......... 93 Figure 1. 2. 10. LIST OF FIGURES Structure of cholesterol ..... . .................. 19 The major pathways of cholesterol oxidation ..... 21 Development of TBARS in whole milk powder in PE pouches during storage as influenced by method of drying and storage temperature ........ 53 Development of TBARS in whole milk powder in glass vials without oxygen absorbers during storage as influenced by method of drying and storage temperature ................. ........ 54 Development of TBARS in whole milk powder in glass vials with oxygen absorbers during storage as influenced by method of drying and storage temperature ................ ......... 55 Total cholesterol oxidation products (COPs) (expressed as a percentage of total cholesterol) in whole milk powder packaged in PE pouches as influenced by storage time and temperature ...... 74 Total cholesterol oxidation products (COPS) (expressed as a percentage of total cholesterol) in whole milk powder packaged in glass vials without oxygen absorbers as influenced by storage time and temperature ............ .. ...... 75 Total cholesterol oxidation products (COPs) (expressed as a percentage of total cholesterol) in whole milk powder packaged in glass vials with oxygen absorbers as influenced by storage time and temperature ........ ....... ..... 76 Concentration of cholesterol oxidation products (COPs) in 6 month-old whole milk powder processed by various drying methods, packaged in PE pouches, and stored at 40°C ........... .... 82 Concentration of cholesterol oxidation products (COPS) in 6 month-old whole milk powder processed by various drying methods, packaged in glass vials without oxygen absorbers, and stored at 40°C ........ ..... . ........ 83 11. 12. 13. Concentration of cholesterol oxidation products (COPS) in 6 month-old whole milk powder processed by various drying methods, packaged in glass vials with oxygen absorbers, and stored at 40°C .............................. Total cholesterol oxides in whole milk powders processed by different drying methods when packaged in various packaging systems and stored at 40°C for 6 months ............ . ........ Total cholesterol oxides in whole milk powders processed by different drying methods when packaged in various packaging systems and stored at 20°C for 6 months ..................... ix INTRODUCTION Cholesterol is susceptible to oxidation in the presence of oxygen, heat, light and radiation. Some of the cholesterol oxidation products (COPS) have been implicated in adverse biological effects, e.g., atherogenesis, cytotoxicity, carcinogenesis and inhibition of cholesterol synthesis (Kumar and Singhal, 1991). Foods of animal origin may contain varying concentrations of COPS, depending upon the severity of processing or storage conditions. Whole milk powders contain 0.2 to 0.4 % cholesterol and when subjected to various thermal treatments and storage at ambient temperature for prolonged periods, may undergo oxidation. Until recently, however, cholesterol oxidation in whole milk powders has not received much attention in comparison to comparable studies with egg products. The author hypothesizes that the formation of COPs in whole milk powder may be eliminated or reduced by developing better methods of processing, packaging, and storage of the products. Recently, the influence of spray drying method on the formation of Cholesterol oxides in egg powder has been examined (Tsai and Hudson, 1985; Missler et a1., 1985). It was demonstrated that the increase in the levels of nitrogen oxides (NOX) generated during the spray drying process, 1 2 enhances the oxidation of cholesterol in egg powder. The major objective of this study was to examine the influence of three methods of spray drying (indirect, electric; direct, low Nox; direct, high NOx) on the formation of cholesterol oxides in whole milk powder. In addition, the oxidative changes of milkfat and cholesterol in whole milk powders during storage under different conditions were also evaluated. Another objective of the study was to evaluate various packaging systems in suppressing lipid oxidation in whole milk powders during storage. The effect of three different packaging systems (polyethylene pouches, crimp—on glass vials with or without oxygen absorbers) on the storage stability of whole milk powders was determined. REVI EW OF LITERATURE Methods of Drying Powdered milk production is an important segment of the dairy industry. Milk powders provide not only a means of handling excess milk produced for other dairy products, especially market milk, but also provide better keeping quality, less storage space, and lower shipping costs. AS a result of dry milk production, it is possible to make milk available in regions that are not suitable for dairying, or where the milk production is insufficient. Milk or milk products can be dried in several ways: 1) roller drying, 2) spray drying, 3) freeze drying, 4) vacuum drying, 5) foam mat drying, and 6) by a vortex method (Hall and Hedrick, 1966). Roller drying and spray drying are the major methods of producing milk solids for milk and milk products. However, roller drying does not meet today's requirements with regard to powder quality, high capacity, high preconcentration and low operating and maintenance costs. Therefore, spray drying has become the most important method in dry milk production (Knipschildt, 1986). The spray dryer system is somewhat complicated. According to the method of furnishing heat and the method of heating the air, spray driers can be classified as: 1) direct- fired spray driers -- burning gas or fuel oil and the products 4 of combustion heating the air directly; 2) indirect-fired spray driers —- using steam or electricity as heat sources and the heat being transferred across heat exchanger plates or coils to the air (Hall and Hedrick, 1966). Most spray driers tend to be the indirect type as a result of the better quality of the products produced, and the relatively nonhazardous nature of the heat source. However, the direct-fired units are more efficient due to less heat loss during heat transfer. Further, the capital investment and maintenance costs in a direct system are substantially less than for an indirect system. For these reasons, there is a growing interest in the use of direct—fired heating systems to produce dried milk products (Kelly and Slattery, 1985; Jansen and Elgersma, 1985; Kelly et a1., 1989), although they are not yet widely used in the food industry. with a direct—fired unit, the selection of fuel is based primarily on the cost and the effect of fuel combustion products on the drying product. In recent years, the availability of natural gas and high fuel prices have created considerable interest in the direct gas-fired burner. However, the danger of contaminating milk powder by products arising from direct gas firing has also been recognized (Knipschildt, 1986). The contamination is produced by oxides of nitrogen, e.g., nitrogen monoxide (NO) and nitrogen dioxide (N02) (commonly known as NOx gases), in the combustion gases. During dehydration, these gases dissolve in the moist 4_____¥ 5 atmosphere of the drier to form nitric and nitrous acids, and contribute to the nitrate and nitrite contents in milk powders. Research has shown that the nitrite and nitrate contents in milk powders prepared in a direct gas-fired dryer, can be as much as double those present in powders produced by indirect heating (Harding and Gregson, 1978). N-Nitrosamines may also be present in the products and are formed from the reaction of nitrite with amines associated with the proteins of the products (Libbey et al., 1980; Challis et al., 1982; Havery et al., 1982). It has been shown that when direct gas—fired spray dryers are used, nitrite, nitrate and N-nitrosamines are formed in the powder at a rate directly proportional to the amount of nitrogen oxides in the drying air (Kinpschildt, 1986). In order to prevent the formation of these undesirable compounds in food products, many researches have attempted to minimize contamination of the drying air by the combustion products in the flue gases from the burner. Rothery (1968) found that the nitrite contents of skim milk powder and sodium caseinate could be significantly reduced by injection of steam into the flame when direct gas—firing is used. However, a burner using steam injection to obtain lower NOx levels could raise the moisture content in the dried products, and may offset some of the energy savings associated with direct firing. Therefore, special burners which produce minimum amounts of combustion— generated NOX are required. ______—__4 6 Wheeler (1980) summarized the mechanism of NO formation from air as a result of combustion. According to his theory, the rate of thermal formation of NO from air is strongly temperature—dependent and does not take place by a simple combination of nitrogen and oxygen, but through a: set of inter-related reactions in which the predominantly active reactants are atomic oxygen, atomic nitrogen and hydroxyl radicals. N2 + o- —---> NO + N' 02 + N' -—--> NO + 0- OH + N' —---> NO + H' At flame temperatures in the range 1500 to 1600°C, the formation of NO starts to be significant. The transformation of nitrogen monoxide to nitrogen dioxide is a low temperature reaction which takes place at approximately 600°C. Therefore, it was suggested that the combustion requirements for minimizing the formation of NOx are low burning temperatures, but high enough to ensure complete combustion, and a Short duration of the top temperature (Knipschildt, 1986) . Conventional gas burners will produce an NOx level of between 50 parts per million (ppm) and several hundred ppm (typically about 150 ppm) (Knipschildt, 1986). In 1982, Altermark and Hess developed a low-NOx burner which could reduce the NOx level to 2-6 ppm. The low—NOx burner incorporated several important features: (1) flame cooling by means of inert or recycled flue gases; (2) homogeneous 7 premixing of fuel gas, combustion air and cooling gas to prevent temperature peaks during combustion; and (3) an almost adiabatic environment of the flame to prevent temperature gradients and to ensure complete combustion at low temperature. The studies of Altermark and Hess (1982) clearly demonstrated the benefits resulting from reduced levels of combustion-generated nitrogen oxides, and this led to the development of the CXA low-NOx gas burner (Urquhart Engineering Co. Ltd, England). This burner uses high excess air levels to burn the gas at the lower limit of its inflammability. Thus, the combustion temperature is lower and the duration of the top temperature in the flame is minimized. To ensure the absence of pockets of gas mixture, the combustion of which would result in localized high temperatures and high NOx levels, gas and air are premixed using a multi-venturi gas/air mixer. Because of the Slower flame speeds associated with combustion at high excess air levels, specially designed flame stabilizers are used to enhance flame stability. The combustion chamber is refractory-lined in order to increase flame stability by reradiation and also to ensure combustion completeness before the quenching of the combustion products by the air to be heated. A diagram of the CXA low-NOx gas burner is shown in Appendix 1. A previous study (Kelly et al., 1989) on the use of the low-NOx CXA burner for milk drying has shown that the NOx 1 _ ‘ 8 levels at the burner throat were in the range of 1.0 to 1.65 ul/l which are superior to those reported by Jansen and Elgersma (1985). Moreover, there was no significant difference between the nitrate contents of skim milk powders produced by direct firing, electrical heating, and commercial producers (indirect heating), and the levels of N- nitrosodimethylamine (NDMA) were unaffected. Only small increases (0.2-1.7 ppm) in the concentrations of nitrite were detected. Mechanism of Lipid Oxidation Lipid oxidation is a major deteriorative reaction that can occur in foods during storage. It is responsible for a wide variety of undesirable reactions such as flavor and color changes, loss of nutritive value as a result of the reaction of oxidation products with proteins, and possible adverse biological effects (Addis, 1986). To prevent the onset of oxidation, an understanding of the initiation of oxidative changes in foods is necessary. As a result of the many investigations over the past fifty years, the fundamental mechanisms of lipid oxidation are well established (Korycka—Dahl and Richardson, 1980; Nawar, 1985; Chan, 1987). However, many details of the reaction and its consequences remain unknown. Lipid oxidation proceeds via a free-radical chain reaction which can be divided into three separate steps as indicated below: Initiation —- formation of free radicals RH ------------- >R'+H‘ RH + 02 --------- > Roo- + H' Propagation —— the free radical chain reaction R- + 02 ————————— > R00' R00' + RH ------- > ROOH + R' Termination -- formation of non—reactive products R' + R' ---------- > RR R' + ROO‘ -------- > ROOR ROO' + ROO' ------ > ROOR + 02 where RH refers to any unsaturated fatty acid in which the H is labile due to the activating influence of the adjacent double bond. R' and R00' refer to the lipid alkyl and peroxy radicals, respectively. The rate of oxidation of fatty acids is dependent on the degree of unsaturation. The reaction can be accelerated by pro-oxidant factors such as metals, free radicals, light, elevated temperature and moisture. Activated oxygen species, including singlet oxygen, hydroxyl radical, ozone, superoxide anion and hydrogen peroxide, may be important in initiating oxidative changes in foods. Rawls and Van Santen (1970) indicated that unless a catalyst is involved, singlet oxygen (102) is believed to be responsible for initiation. Once the reaction has been initiated, hydroperoxides (ROOH) are formed 10 and subsequently undergo homolytic cleavage, forming two radicals which induce the autocatalytic propagation of the oxidative mechanism and yield a wide variety of secondary products. Termination reactions occur when the concentration of radicals are sufficiently great to permit kinetically preferential radical—radical interactions. The final products of oxidation, although dependent on the type of oxidized fat or oil, generally include short chain aldehydes,ketones, acids, alcohols and other carbonyl compounds (Nawar, 1985). Stability of Whole Milk Powder during Storage Comnosition of milkfat Milkfat consists of approximately 95-96% triacylglycerols, 1.3-1.6% diacylglycerols, and 0.8-1.0% phospholipids. It also contains varying quantities of other compounds such as monoacylglycerols, carotenoids, free fatty acids, sterols and vitamins (Kurtz, 1974). The major fractions of bovine milk lipids are listed in Table 1. Fatty acids, the major components of triacylglycerols in milkfat, account for over 85% of the total weight. The relative proportions of the various fatty acids in milkfat are affected by many factors such as feeding conditions and Seasonal variations. Milkfat with a linoleic acid content as high as 30 % has been reported by Hill et al. (1977). Approximately 500 fatty acids have been detected in milkfats 11 (Sonntag, 1979). It is probable that additional fatty acids remain to be identified. The major fatty acids in bovine milkfat are listed in Table 2. Milkfat is distinguished from other fats by the low average molecular weight of its fatty acids and by its high content of steam volatile acids, including butyric acid, which is unique to milkfat from ruminants. Milkfat contains predominantly saturated fatty acids ranging from C4 to C18. The unsaturated fatty acids of milkfat are mainly monounsaturated. Oleic acid (Clan) is the principal unsaturated fatty acid, with small amounts of di-unsaturated fatty acids, principally C18:2 (linoleic acid) and trace amounts of other' polyunsaturated fatty acids. For lipid oxidation, the important lipids in foods are the unsaturated fatty acids, particularly oleate, linoleate, and linolenate (Labuza, 1971). The susceptibility and rate of oxidation of these fatty acids increase with their degree of unsaturation. Therefore, measuring the extent of oxidation of these fatty acids can help determine the point at which rancidity occurs. Most fatty acids of milkfat are present in triacylglycerols, but small proportions of free fatty acids are always present in fresh milkfat. Larger percentages of free fatty acids in milk are found in samples stored over time. Ritchie (1967) found little or no qualitative difference between the fatty acid profiles of free and total fat in whole milk powder. 12 Table 1. The major fractions of bovine milk lipids. Component % Total milk lipids Triacylglycerols 95—96 Diacylglyerols 1.26—1.59 Monoacylglycerols 0.016-0.038 Phospholipids (total) 0.8—1.0 sterols 0.22—0.41 Free fatty acids 0.10-0.44 Adapted from Kurtz (1974). Table 2. The major fatty acids in bovine milkfata Fatty acid Wt % Fatty acid Wt % C 4:0 2'79 C 15.0 0.79 C 6:0 2'34 C 16:0 23.8 C 8:0 1'06 C 16:1 1.78 C 10:0 3'04 C 17:0 0~7 C 10:1 0.27 C 18:0 13.2 C 12:0 2.87 C 18:1 29.63 C 14:0 8.94 C 1822 2.85 C 14:1 0'75 C 18:3 0-38 aAdapted from Kurtz (1974) 13 Oxidation of milkfat Dry whole milk has a shelf life of less than 6 months, mainly as a result of its susceptibility to oxidation (Boon et al., 1976). The oxidation of fat in whole milk powder has recently been reviewed by Tuohy (1987). Oxidation of milkfat occurs according to the classical mechanism. However, the complex composition of dairy products as well as processing , manufacturing, and storage conditions, tend to influence both the rate of oxidation and the composition and percentage of oxidation products formed. In whole milk powder, the triacylglycerols are relatively susceptible to oxidation, whereas the phospholipids are more stable (Patton, 1962). Conversely, in fluid milk, phospholipids will undergo oxidation more readily than the triacylglycerols (Lea, 1953). The fatty acids in milkfat are mainly saturated. From a practical viewpoint, the oxidation of saturated fatty acids at ambient temperatures can be ignored, as they remain unchanged relative to the unsaturated fatty acids (Nawar, 1985). Milkfat contains unsaturated fatty acids, e.g., oleic acid, linoleic acid and linolenic acid, and these are susceptible to oxidation. Pathways for the formation of hydroperoxides from these acids have been outlined by Frankel (1962). Oleic acid, a major unsaturated fatty acid of milkfat, has two a-methylene groups and these are the points of attack in the free radical chain reaction. Hydrogen abstraction at C8 and C11 results in the formation of two allylic radicals. 14 Oxygen attack at the end carbons of each radical produces an isomeric mixture of C8-, C9-, C10-, and Cll-allylic hydroperoxides. It has been shown that these hydroperoxides are formed in about equal amounts (Frankel, 1962, 1979). For linoleic acid and linolenic acid, polyene non— conjugated systems, the preferential points of attack are the a-methylene groups located between the double bonds. These a- methylene groups are doubly activated by the two adjacent double bonds, which makes them much more susceptible to oxidation than the a-methylene groups in fatty acids with one double bond. In linoleic acid, hydrogen abstraction at the C11 position along with oxygen attack produces an equal mixture of conjugated C9- and Cl3-diene hydroperoxides (Chan and Levett, 1977). The preferential formation of the free radicals at the C9 and C13 positions is explained by the fact that these products could be resonance stabilized by a conjugated diene system. In linolenic acid, hydrogen abstraction at C11 and C14 results in the formation of a mixture of isomeric C9-, C12-, C11" and (as—hydroperoxides (Chan and Levett, 1977). In theory, six isomeric hydroperoxides are possible from linolenic acid oxidation, however, C11- and C14-hydroperoxides have not been found. This phenomenon is also explained by the theory of resonance stabilization. Hydroperoxides are the primary products resulting from the autoxidation of unsaturated fatty acids. These 15 hydroperoxides are unstable and readily decompose. In particular, the primary hydroperoxides of linolenate easily decompose to secondary dihydroperoxides because of the presence of active methylene groups in their structures. The main products of hydroperoxide decomposition are saturated and unsaturated aldehydes. Other products, such as unsaturated ketones, saturated and unsaturated alcohols, saturated and unsaturated hydrocarbons, and semi-aldehydes, have been observed in the decomposition of hydroperoxides of oxidized lipid systems. In addition to the major fatty acids, milk also contains many minor polyunsaturated acids, hence the oxidation of dairy products can lead to a multitude of saturated and unsaturated aldehydes (Parks, 1974). Prevention of oxidation by packaging Since reactive oxygen species are instrumental in initiating many of the oxidatively damaging reactions in food systems, it would be desirable to inhibit these reactions. For the long-term storage of foods, it often requires almost complete removal of headspace and dissolved oxygen. Some antioxidants such as butylated hydroxyanisole (BHA), butylated hydroxytoluene (BHT), and tocopherols probably can react with a variety of activated oxygen species including free radicals (ROO', RO') and singlet oxygen to minimize the oxidation of milkfat (Abbot and Waite, 1962; Hammond, 1970). However, addition of antioxidants in standard dairy products is not 16 permitted, except for a-tocopherol which is less effective in retarding lipid oxidation compared to the synthetic phenolic antioxidants. Alternatively, the shelf life of many food products has been greatly extended through improved packaging technology. One approach in food packaging is to store the food in an atmosphere as low in oxygen as possible to prevent or retard lipid deterioration. Taylor et al. (1979) recommended low temperatures coupled with oxygen barrier packaging or nitrogen flushing for the storage of dried foods. Warmbier and Wolf (1976) found that the oxygen levels can be reduced to 2% by flushing with inert gases. The use of nitrogen flushing or vacuum packaging has been found to be effective in minimizing the formation of undesirable flavor compounds in dry whole milk during storage (Tamsma et al., 1967,1973). Min et al.(1989) also demonstrated the beneficial effects of gas flushing with a mixture of 92% nitrogen and 8% hydrogen in retarding lipid oxidation in dry whole milk by measuring formation of volatile compounds in the headspace of packaging system. Samples packed in air were much more oxidized. Although the oxidation of lipids may be controlled by gas flushing, residual oxygen may still cause deterioration in enclosed systems, especially for dry milk products. Mucha et al. (1961) noted that extremely low oxygen concentrations are necessary to prevent oxidative flavors in reconstituted foam spray-dried whole milk. Berlin and Pallansch (1963) reported 17 that it is possible to entrap oxygen inside dry milk particles during processing. In order to reduce the residual oxygen level in packaging systems to satisfy product requirements, it is necessary to obtain an efficient inert gas flush system and a scavenging system to tie up any oxygen which may be entrapped in the product or which may permeate through the packaging material (Berlin and Pallansch, 1963). Many researchers have developed oxygen absorbers, also referred to as oxygen removers and oxygen scavengers, and evaluated the effect on food product quality and stability. A free oxygen absorber which consisted of iron powder, ferrous sulfate and a hygroscopic substance was developed by Maude et al. (1925). The first application of oxygen absorbers in preserving dry food quality was reported by Isherwood (1943). Later, a new oxygen scavenging system was devised by King et a1. (1955). The system involved flushing the pack containing palladium pellets and product with a combination of hydrogen and nitrogen gas. Palladium catalyzed the reaction: 2H2 + 02 -—> 2H20 to remove residual oxygen from the headspace. Kuhn et al. (1970) demonstrated that a scavenger system containing polyvinyl alcohol as the principal gas barrier medium was very effective in achieving and maintaining very low residual oxygen concentrations in packages containing whole milk powder. In order to improve the reliability of the scavenging activity, Zimmerman et al. (1974) changed the film structure of the oxygen scavenger pouch to polyester/foil/ionomer/ 18 catalyst/ionomer. This oxygen scavenging system significantly improved the flavor stability of whole milk powder during storage. Another type of oxygen absorber consisting of dithionite (sodium hyposulfite and hydrosulfite), iron powder and organic reductions was developed by Fujishima (1977). It was claimed that sulfur dioxide (802) was formed during the oxygen scavenging reaction and was trapped in the food product. This type of oxygen absorber has been applied successfully to the storage of vegetable oil, fried beans and dry instant noddles (Saito, 1979). In 1981, Rooney developed a new in-pack deoxygenation system, by using organic compounds immobilized in polymer media, for oxygen scavenging from headspaces in sealed packages. The results demonstrated that the oxygen content of plastic pouches containing air was reduced from 21 % to 1. % in 6 ndnutes. Recently, Sakamaki et a1. (1988) conducted a study with oat cereal packaged in materials of different oxygen barrier properties with or without pouched— type oxygen absorbers, which consisted of a mixture of iron powder, catalysts and water. The absorbers were effective in delaying lipid oxidation in the oat cereal during storage. Oxidation of Cholesterol Description of cholesterol Cholesterol (C27 H45 OH), cholest—S—en—3B-ol, is the major sterol in mammalian tissues. It is present in all 19 cellular membranes, especially the plasma membrane, and found in association with phospholipids, may control the passage of substances across such membranes (Dugan, 1987). By regulating fluidity of the membrane, cholesterol regulates membrane permeability, thereby exercising some control over what may pass into and out of the cell (Hunt and Groff, 1990). The total cholesterol content in all mammalian species is between 0.1% to 0.2% of the body weight of adults (Gibbons et al., 1982). Nervous tissue and the brain together contribute the largest proportion of the total body cholesterol. Cholesterol is a non—polar simple lipid and contains a cyclopentanophenanthrene ring structure. The ring structure has an eight carbon side chain at the C17 position of the D ring, two methylated tertiary carbons at the C10 and C13 positions, a double bond at the C5-C6 position and a hydroxyl group at C3. The main features of its stereochemistry are shown in Figure 1. Figure 1. Structure of cholesterol 20 Mechanism of cholesterol oxidation Being an unsaturated lipid, cholesterol readily undergoes oxidation in the presence of oxygen and light by a free radical process. Exposure of cholesterol to air, heat, light and irradiation may result in a variety of oxidation products. Free cholesterol is more stable towards oxidation than its esters (Korahani et al., 1982). The major pathways of cholesterol oxidation have been elucidated and reviewed by Smith (1981) (Figure 2). He proposed two free radical mechanisms for cholesterol oxidation: (1) hydroperoxide formation in the B-ring and side- chain; and (2) formal dehydrogenation of the 3B-alcohol group. RH + 02 ------ > ROOH (1) RCH(OH) + 02 ----- > RC=O + H202 _______ (2) In equation 1, the major mode of cholesterol oxidation, R mostly refers to C7, C20 and C25; moreover, it also refers to C24, C26 or other carbons in the side chain. The C7 position, the carbon adjacent to the double bond, is most sensitive to molecular attack by oxygen. The initial reaction involves the abstraction of the allylic C7 hydrogen and reaction with ground state dioxygen (302) to form C7 peroxy radicals. These peroxy radicals are stabilized by hydrogen abstraction to form the epimeric 7a- and 70- hydroperoxides (Smith and Hill, 1972), which are the first detectable stable products. Both 21 Ammma .>Q5AV :ofluccflxo Houmummaono mo m>m3nuom HOmoE one .m ousmflm uoiozuaozo>:.us-oo-an-zupmuoozo-16.m uzo-~-wzyo<»mwoozo.nn moiozuao¢o>:.ah.oo.nn-zuhmuoozo-rn.< o0 ,/ 5;: .3 ....ndn 2:35:65 / v . :o .10 o: 65-..; $2583.55 soaasanéoaooozo-“ . o :1 20:. oz :0 o: / mzo-s-oo-nn-zo»muoo:o-n O” O: . :oox.:z zoox.zz O I 0.1 0:. Joan 58$ twin 253905 John 58% no.9. zfimuooxo V4 .w o 0 ® \ n O: o: :00... O: :00 0: ‘ll 0 :oo:.:x VY/(\/mwmwwmwyflwq\ o VY/\)MWW%WWWMHHH\ zoom .1: .o 493-233.1983 /o o: o: 4 Alllll uhxomua . 22 hydroperoxides are subsequently decomposed to form therespective diols, or yield 7-ketocholesterol via direct dehydration (Van Lier and Smith, 1970; Smith et al., 1973). Epoxides, secondary oxidation products of cholesterol, are formed by the reaction of an unoxidized cholesterol molecule with 7-hydroperoxides. In terms of thermodynamic stability, the levels of 7B-diol and B-epoxide are predominant over those of the 7a—diol and a-epoxide, respectively, during the cholesterol oxidation process. Further hydration of these epoxides results in the formation of cholestan-3B,5,6B-triol (Smith and Kulig, 1975). The presence of the tertiary C20 and C25 atoms in the side chain adds to the centers that are sensitive to oxidation. Oxidation reactions occurring at these positions tend to form relatively stable radicals via hydrogen abstraction by other free radicals ( Van Lier and Smith, 1970, 1971). These radicals can then react with molecular oxygen to form hydroperoxides. These hydroperoxides then continue the chain propagation reactions, involving other cholesterol radicals to yield the corresponding hydroxycholesterols upon reduction or ketones via dehydration. In the minor mode of cholesterol oxidation (equation 2), the 3fi-alcohol group may undergo oxygen-dependent dehydrogenation involving initial C3 radical formation and reaction with 3O2 to form 3-hydroxy—3—peroxyl radicals. The 3—peroxyl radicals are then stabilized by elimination of the 23 elements of H202, resulting in the formation of cholest-S-en- 3-one (Ansari and Smith, 1978). Biological effects of cholesterol oxides Cholesterol oxidatitwiproducts have received considerable attention in recent years because of the biological activities associated with the etiology of certain human diseases. Based on studies of the biological effects and metabolic pathways, cholesterol oxidation products have been linked with carcinogenesis, cytotoxicity, atherogenesis, and inhibitioniof cholesterol synthesis (Smith, 1981; Addis and Park, 1989; Kumar and Singhal, 1991). The atherosclerotic effects of diets containing cholesterol have well documented for many decades. However, the possible correlation of atherosclerosis and cholesterol oxides was not established until the reports of Imai et al. (1976). By feeding concentrated impurities from USP-grade cholesterol to rabbits, they observed an increase in frequency of dead or dying aortic smooth muscle cells (an index of angiotoxicity) and induced focal edema 24 hours after administration of the contaminant at a gavage of 250 mg/kg body weight. Administration of new and 5—year—old cholesterol produced similar effects, with the older sample exhibiting the greater atherosclerotic response. However, purified cholesterol showed no increase in degenerative cells. Imai and co-workers (1980) conducted a similar experiment to 24 further confirm the angiotoxicity of cholesterol oxides by injecting synthetic oxygenated sterols into rabbits. They concluded that oxygenated sterols, not cholesterol, may play a primary role in arterial wall injury and lesion development. More recent studies have also provided evidence that certain cholesterol oxides Show atherogenic effects (Baranowski et al., 1982; Peng et al., 1982; Addis, 1986). A large number of cholesterol oxides have been evaluated for their atherogenicity. Addis (1986) concluded that the cholesterol oxides regarded as most atherogenic are cholestane-triol and 25-hydroxycholesterol. The inhibitory effect of cholesterol oxides on cholesterol biosynthesis has also been thoroughly investigated during the past two decades. Kandutsch and Chen (1973, 1974) demonstrated that certain oxygenated sterols influenced cholesterol biosynthesis by inhibiting the activity of HMG-CoA reductase (3-hydroxy-3-methylglutaryl-CoA), a role-limiting enzyme in cholesterol biosynthesis. According to these reports, 25—hydroxycholesterol was the most potent inhibitor of cholesterol biosynthesis. However, Brown and Goldstein (1974) reported that the most potent inhibitor of cholesterol biosynthesis was probably 7-ketocholesterol, which was 100 times more potent than cholesterol. Parish et al. (1986) stated that the structural features of cholesterol oxides are related to their ability to repress the activity of HMG-CoA reductase. As a general trend, the inhibitory effect 25 increases as the distance between the C3 and the second oxygen function becomes greater. Because of the important biological activities of cholesterol in cells and cell membranes, the cytotoxic effects of cholesterol oxidation products have also been the major focus of many studies. Higley and Taylor (1984) demonstrated that some cholesterol oxides can cause inhibition of cell growth and cell death on cultured medial smooth muscle cells. Addition of cholestan—triol and 25-hydroxycholesterol to the cultured smooth muscle cells results in cell death, whereas purified cholesterol showed no cytotoxicity (Peng et al., 1979, 1982). The inhibitory effect of cholesterol oxides on the activities of the enzyme Na+, K+ ATPase and 5'- nucleotidase has also been observed (Peng et a1, 1985; Peng and Morin, 1987). Another risk associated with the intake of cholesterol oxides is the suspected carcinogenicity of a-epoxide. Black and Douglas (1972, 1973) observed the formation of a-epoxide in the skin of human and hairless mice after exposure to UV radiation. Chan and Black (1974) also showed that increases of the a-epoxide level in the skin of hairless mice subjected to UV light coincided with a rapid increase in the number of tumor incidences. Although there is no direct evidence to link a—epoxide to development of cancer, the presence of epoxycholesterol may still play an indirect toxicological role. 26 Determination of cholesterol oxidation products in foods During the past decade, various methods have been developed for the identification and quantification of cholesterol oxidation products in foods. These methods include thin—layer chromatography (TLC), packed column gas chromatography (GC), capillary column GC, and high-performance liquid chromatography (HPLC). Although TLC remains one of the best methods for rapid analyses of cholesterol oxides (Smith, 1987), a few compounds, e.g., 7-ketocholesterol and 5,6-epoxides, are poorly resolved by TLC (Maerker, 1987). HPLC has been successfully used for identification and quantitation of some cholesterol oxides. The procedure is highly effective for the separation of diols and hydroperoxides, but triol is often not determined by adsorption HPLC (Maerker, 1987). Many workers have demonstrated that satisfactory resolution of the major oxidation products of cholesterol can be achieved by capillary GC [(Gumulka et al., 1982; Missler et al., 1985; Park and Addis, 1985b, 1986; Maerker and Unruh, 1986; Nourooz-Zadeh and Appelqvist, 1988 a,b; Sander et al., 1988, 1989; Morgan and Armstrong, 1989; Engeseth, 1990; Pie et al., 1991)]. Smith (1987) reviewed the analytical methods for oxysterols and concluded that GC analysis of the oxysterol trimethylsilyl ethers on capillary columns with suitable bonded or liquid phases is the most powerful technique at the present time. The quantitative determination of cholesterol oxides in 27 foods is prone to several potential errors. One is the generation of artifacts during the analytical procedure. Another is the breakdown of oxides during analysis (Smith, 1981). It is especially important to observe the formation of 5a-cholestane-3B,5,6B-triol which is a spontaneous breakdown product of the two epoxy compounds. Another error arises from the breakdown of cholesterol oxides during saponification (Park and Addis, 1985). The presence of large quantities of cholesterol in foods interferes with the quantification of some cholesterol oxidation products. Therefore, cholesterol must be removed prior to analysis. Previous workers have frequently employed hot alkaline saponification to hydrolyze lipids and to concentrate non—saponifiable sterols (Flanagan et al., 1975; Ryan et al., 1981; Finocchiaro et al., 1984). Several reports, however, have demonstrated that some cholesterol oxides suffered from structural alterations under such harsh treatment (Tsai and Hudson, 1981; Maerker and Unruh, 1986; Park and Addis, 1986). For example, the isomeric epoxides and 7-ketone can decompose to various degrees during this procedure, thereby generating the triol and the diene, respectively. There are many literature reports detailing losses of Single cholesterol oxides. Chicoye et al. (1968) reported that when 7-ketocholesterol was heated with a potassiunihydroxide solution, cholesta-3,5—dien-7-one was the major degradation product produced. Tsai et al. (1980) ..‘A‘ , 7.“. 28 reported a loss of approximately 75% of the a-epoxide after saponification, due to the hydrolysis of the epoxide ring. Maerker and Unruh (1986) reported no losses of the isomeric epoxides, but a substantial loss of 7—ketocholesterol during hot saponification. Moreover, they also demonstrated that 6- ketocholestanol, a compound commonly used as an internal standard, is affected by base and is partially destroyed by hot alkali as well. Therefore, it is necessary to develop an alternative technique for more accurate quantitation of cholesterol oxides. Park and Addis (1985a) developed a mild method, avoiding saponification, for the analysis of C7 cholesterol oxides. Total lipid extracts were fractionated on silica gel columns to concentrate trace sterol oxides from triacylglycerols, cholesterol, and phospholipids. These results showed that the combination of silica gel column chromatography and HPLC with UV detection was quite reliable and reproducible. Later, they prepared cholesterol and its oxidized derivatives through cold saponification and found that no detectable amounts of artifactual oxidation products were produced during the sample preparation. Nourooz-Zadeh and Appelqvist (1987) also demonstrated that the breakdown of cholesterol oxides was minimized by the elimination of the saponification step. Therefore, Smith (1987) has emphasized that some of the quantitative data of cholesterol oxides must be viewed with skepticism because of the insidious nature of cholesterol 29 oxidation. Cholesterol Oxidation Products in Foods Food products of animal origin are often subjected to various processing treatments which may induce the oxidation of cholesterol. Until recently, however, the deterioration of foods due to the oxidation of cholesterol was seldom examined. Because cholesterol oxides are present in foods in rather low concentration, their isolation and determination presented challenging analytical problems. In the past decade, as a result of the development of appropriate analytical methods, cholesterol oxides have been found in a variety of foodstuffs such as egg products (Tsai and Hudson, 1984; Missler et al., 1985; Sugino et al., 1986; Nourooz—Zadeh and Appelqvist, 1987; Morgan and Armstrong, 1989; Sander et al., 1989), dairy products (Finocchiaro and Richardson, 1983; Luby et al., 1986; Nourooz—Zadeh and Appelqvist, 1988 a,b; Sander et al., 1988; Sander et al., 1989), heated fats (Ryan et al., 1981; Bascoul et al., 1986; Park and Addis, 1986 a,b), meat products (Park and Addis, 1985 a; Higley et al., 1986; Park and Addis, 1987; Engeseth and Gray, 1989; Zubillaga and Maerker, 1991; Pie et al., 1991), and other deep-fried foods (Lee et al, 1985; Zhang et al., 1991). Some of the cholesterol oxidation products isolated from foods are listed in Table 3. Formation of cholesterol oxidation products in foods may occur from oxidation before, during, or after processing. Table 3. Cholesterol oxidation products in foods 30 a Food Oxysterols found (ug/g) Egg Products: Dried egg yolk Spray-dried egg Milk Products: Whole milk powder0 Cheeses Butteroil Meat Products: Heated beef tallow Minced pork (raw, cooked)d 38,7a-diol (50) 3B,7B—diol (25) 3B,25-diol (25) cholest-S-ene-3B,4B-diol 7-ketone (22)b 5a,6a-epoxide (79)b 58,6B-epoxide (28)b 35,7a-diol (2.2-20.2) 3B,7B-diol (1.6-24.6) 5,6-epoxides (1.6-24.9) 7-ketone 50,6a—epoxide (17.4) 58,6fi-epoxide (31.8) triol (0-1.2) 3B,7a-diol (0.3—1.0) 33,7p-diol (0.7-1.5) 5a,6a-epoxide (0.5-2.5) 58,6B-epoxide (1.1—3.2) 7-ketone (5.1-9.2) 3B,7a-diol (0.1-0.8, 3 3B,7B-diol (0.2—0.9, 3 5,6-epoxides (0.1-0.9, triol 3B,7a—diol (0.1—0.4, 20-60) 38,7B-diol (0.3-1.2, 30-90) 7-ketone cholesta-3,5-dien-7-one 5,6-epoxides (0.1-0.2, 20-30) triol (0.1) 3B,25-diol (0.17) _6) -6) 6-32) 3B,7-diols 7-ketone cholesta-3,5-dien—7—one 5,6-epoxides triol 38,7a-diol (0.19, 0.64) 3B,7B-diol (0.28, 0.85) 50,6a—epoxide (0.22, 0.39) 58,6B-epoxide (0.35, 1.01) 7-ketone (0.92, 2.25) 3B,20-diol (0, 0.14) 39,25-diol (0.13, 0.38) triol (0.04, 0.06) 31 Raw chicken musclee 50,6a-epoxide (0.066) 58,6B-epoxide (0.058) 7—ketone (0.129) Other Products: French fried potatoesf 3B,7-diols 5,6-epoxides 7-ketone 3B,25-diol triol aAdapted from Smith (1987) bFrom Sander et al. (1989) cFrom Nourooz—Zadeh and Appelqvist (1988) dFrom Pie et al. (1991) eFrom Zubillaga and Maerker (1991) fFrom Zhang et al. (1991) Smith (1981) reported that the major oxidation products of cholesterol in foodstuffs include 25—hydroxycholesterol, cholestane-triol, 7a-hydroxycholesterol, 7B- hydroxycholesterol, 7-ketocholesterol, a—epoxide, B-epoxide, and cholesta—3,5-dien-7-one. Heating foodstuffs in air or placing foodstuffs under illumination can increase the levels of cholesterol oxides. Missler et al. (1985) reported that the spray-drying of eggs using direct heat (gas-fired) leads to increased cholesterol oxide concentrations in comparison to the use of indirect heat (steam). Cholesterol Oxidation Products in Whole Milk Powder Cholesterol is a minor but essential component of all mammalian milks (Hilditch and Williams, 1964; Sabine, 1977). The amount of cholesterol in milk varies among species and tends to vary proportionally to the triacylglycerol 32 concentration in different species (Moore and Richardson, 1965; Sabin, 1977). The cholesterol content of cow's milk is small, 0.20 %-O.4 % of the total lipids (Nataf et al., 1944; Moore and Richardson, 1965; Brink, 1968; Sweeney and Weihrauch, 1977) and varies with the seasons (Nieman and Groot, 1950). Cholesterol in milkfat is esterified or unesterified (Nataf et al., 1948). However, most cholesterol (85 to 90%) in milk exists as free cholesterol (Patton and McCarthy, 1963). The bulk of this cholesterol is associated with the phospholipids in the formation of the surface membrane of fat droplets. It is derived from the plasma membrane of the secretory cells of the mammary gland. A minor portion of the cholesterol exists in an esterified form, usually combined with long-chain fatty acids. There are many investigations of the properties of the fatty acid in cow's milk cholesteryl esters (Patton and McCarthy, 1963; Keenan and Patton, 1970; Parks, 1980; Wood and Bitman, 1986). Recently, Wood and Bitman (1986) found that in the cholesteryl esters in the milk of mature cows, 62 % of the fatty acids were saturated. Linoleic acid represented 27.1% of the fatty acids present. As dried milk powders contain cholesterol, it has been suggested that they may have an adverse impact on human health. Wilson (1976) considered that the consumption of old powdered whole milk may be a significant risk factor in human atherosclerosis. Taylor et al. (1979) reported that dried 33 powdered foods of animal origin stored in air at room temperature may be major sources of angiotoxic sterol oxidation products in the diet. Leduc (1980), as cited by (Luby, 1982), also demonstrated that ingestion of dehydrated milk powder caused a high incidence of tumors in mice. Dried milk products are processed under various thermal conditions and are stored at ambient temperature for prolonged periods. These treatments provide ample opportunities for the oxidation of cholesterol to occur. However, much of the research regarding cholesterol oxidation in dairy products has focused on butter and cheese. Although there are reports on the occurrence of cholesterol oxides in milk powder products (Flanagan et al., 1975; Finocchiaro and Richardson, 1983), no quantitative data are available on the levels of these oxides in milk powder. Nourooz—Zadeh and Appelqvist (1988a) reported the concentrations of cholesterol oxides in milk powder products (cream, whole milk and skim milk powder) as related to processing technology and storage. Their results revealed that all freshly-processed milk powder products contained less than 0.1 mg cholesterol oxides/ kg total lipids for low and medium heat powders. On the other hand, freshly spray-dried whole milk and skim milk powders from "high heat" conditions had quantifiable amounts of some important cholesterol oxides, e.g., cholest-5-ene-3B,7a-diol, cholest-S-ene-3B,7B—diol, 5,6a-epoxy-5a-cholestan-3B—ol,5,6B-epoxy-SB-cholestan-3B-ol, 34 and 3B—hydroxycholest-5-en-7—one. The epimeric 7- hydroxycholesterols and 7—ketocholesterol were the major oxidation products, followed by the isomeric 5,6— epoxycholesterols. Furthermore, one—year-old Spray-dried whole milk powder samples exhibited increased levels of cholesterol oxides. The quantitative pattern for the concentration of the oxidation products of cholesterol in milk powder products was almost similar to that observed in the analysis of dehydrated egg yolk and egg yolk mix products (Nourooz-Zadeh and Appelqvist, 1987). However, 7— ketocholesterol was found in relatively higher concentrations in milk powder products than in egg yolk powders. MATERIALS AND METHODS Materials Cholesterol oxide standards Cholesterol and Sa-cholestan-3B—ol-6-one (6- ketocholesterol) were purchased from Sigma Chemical Co. (St. Louis, MO). The following cholesterol oxide standards were purchased from Steraloids Inc., Wilton, NH: 5a-cholestane, cholestan-S,6a-epoxy-3B—ol(a-epoxide),cholestan-5,6B-epoxy- 3B-ol (B-epoxide), 5-cholesten-3B-ol-7—one (7- ketocholesterol), cholestan-3B,5a,6B-triol, 5-cholesten- 3B,ZOa-diol (20a—hydroxycholesterol), 5-cholesten—BB,25-diol (ZS—hydroxycholesterol), 5-cholesten-3B,7a-diol (7a- hydroxycholesterol) and 5-cholesten-3B,7B—diol (75- hydroxycholesterol). Fatty acid standards The methyl esters of butyric acid, stearic acid, oleic acid, linoleic acid, and linolenic acid were purchased from Sigma Chemical Co. (St. Louis, MO). Standard reference mixtures of fatty acid methyl esters (C6, C8—C16) were also obtained from Supelco, Inc. (Bellefonte, PA). 35 Packaging materials The pouched oxygen absorbers (consisting of iron powder, water and catalysts) were obtained from Multiform Desiccants, Inc., Buffalo, NY. Polyethylene pouches (Ziploc freezer bags, 5"x7", 2 mil thickness) were purchased from a local supermarket. Crimp seal glass vials (100 ml), teflon septa, and aluminum crimp-on caps were obtained from Supelco, Inc. (Bellefonte, PA). Reagents Pyridine and bis-(trimethylsilyl)—trifluoroacetamide (BSTFA) were obtained from Pierce Chemical Co. (Rockford, IL). All other reagents used in the experiments were of analytical grade. Preparation of Whole Milk Powders Liquid whole milk was supplied by a local producer in County Cork, Ireland. A pilot-scale Anhydro Lab3 spray drier with pneumatic nozzle atomization was used to spray dry the milk at the National Dairy Products Research Centre (Moorepark, Fermoy, Co. Cork, Ireland). The drier was equipped with electrical heating elements for indirect air heating. It also had the capability of being converted to direct gas-fired heating by attaching a gas burner to the extended air inlet duct. Whole milk powder samples were manufactured by direct low-NOx gas—fired heating and direct 37 high-NOx gas-fired heating of the drier, respectively, with the fitting of a CXA gas burner (digram shown in appendix 1) and a standard gas burner (Radiant Superjet model GX 25) to the drier. Samples of whole milk powder produced by the indirect (electric) air heating system were used as experimental controls. Whole milk powder samples were vacuum packaged on the same day as drying and air freighted to Michigan. On receiving at MSU, the whole milk powder samples were refrigerated at 4°C for one week until required for analysis. After the initial analyses (TBA test, color measurement, total lipid and total cholesterol analysis), duplicate samples (60 g) produced by all drying methods were placed into polyethylene pouches (5" x 7", 2 mil thickness), crimp seal glass vials (100 ml) with oxygen absorbers, and crimp seal glass vials without oxygen absorbers, respectively. After filling, the sample pouches were heat sealed and the sample vials were sealed by aluminum crimp—on caps with teflon septum. The packed samples were stored in the dark for six months at 20°C and 40°C. The whole experiment was repeated twice. Methods Measurement of lipid oxidation Lipid oxidation in the stored powder samples was measured using the thiobarbituric acid (TBA) procedure. Ten grams of 38 the whole milk powder were mixed with 97.5 ml of deionized water and added to a distillation flask containing silicone emulsion antifoam (Thomas Co., Swedesboro, NJ), glass beads, and 2.5 ml of hydrochloric acid (HZO/HC1= 2:1, v/v). By using the distillation method of Tarladgis et al. (1964), 50 ml of distillate were collected. A 5 ml aliquot of the distillate was added to a test tube containing 5 ml of aqueous TBA solution (Crackel et al., 1988). The capped tubes were lightly vortexed, and then heated in a water bath at 100°C for 35 minutes. The tubes were cooled to room temperature and the color intensity quantitated by measuring the absorbance at 532 nm using a double beam Bausch and Lomb Spectronic 2000 spectrophotometer (Rochester, N.Y.). Lipid extraction Total lipids were extracted from the milk powders using a slightly modified version of the method of Folch et al. (1957). Whole milk powder (1 g) was homogenized using a Ultra—Turrax type of homogenizer (Tekmar Co., Cinn., OH) in a 150 ml beaker for l min with 15 ml methanol, then chloroform (30 ml) was added and the process continued for an additional 2 min. The homogenate was filtered through Whatman filter paper (No.1) to remove whole milk powder particles, which were then washed with 30 ml chloroform/methanol (2:1, v/v). After washing, the combined filtrates were added to 10 ml deionized water and shaken by hand for thirty seconds to ensure mixing. 39 The mixture was then transferred to a 100 ml centrifuge tube and centrifuged (1000 x g) for 10 min. The upper aqueous layer was removed by aspiration. The inside wall of the centrifuge tube was further rinsed with 3 x 1.5 ml aliquots of a solvent mixture of chloroform/methanol/water (3:48:47, by volume). The lower cholroform phase was collected, dried over anhydrous Nazso4, and evaporated to dryness using a rotorary evaporator. The lipid extract was redissolved in 5 ml hexane and stored at —20°C until required for further analysis. Total cholesterol determination The total cholesterol content of the whole milk powders was determined by the method of Adams et al. (1986). Milk powder (3-5 g) was accurately weighed into a 250 ml flat bottom refluxing flask to which was added 8 ml 50 % KOH and 40 ml alcohol. This mixture was saponified by refluxing and stirring for one hour. After the addition of 60 ml alcohol, the mixture was allowed to cool down and toluene (100 ml) was added and vigorously stirred for one min. The mixture was extracted in a 500 ml separatory funnel with 110 ml 1N KOH, shaking vigorously for 30 sec and then the lower aqueous layer was discarded upon separation. To the toluene layer, 40 ml 0.5N KOH were added and the separatory funnel was rotated gently. Again the lower layer was discarded and the toluene extract was washed five times with water (100 ml). The toluene extract was dried over anhydrous NaZSO4. A 50 ml 40 aliquot was removed and rotary evaporated to dryness, washed with acetone, dried, and redissolved in 3 ml dimethylformamide. Gas chromatographic (GC) analysis was carried out using a packed (1% SE-30 on 100/120 Gas Chrom Q) column and a Hewlett Packard 5840A gas chromatograph (Avondale, PA) under isothermal conditions (230°C for 20 min). The injector and detector temperatures were held at 275°C and 300°C, respectively. Helium with a flow rate of 25 ml/min was used as the carrier gas. Cholesterol oxide determination The procedures used to concentrate and derivatize the cholesterol oxides were slightly modified from the method of Morgan and Armstrong (1989). The lipid extract (250 mg) with 15 ug of internal standard (6-ketocholesterol) was transferred to EH1 Supelclean LC-Si SPE tube (Supelco, Bellefonte, PA) which had been prewashed with 5 ml hexane. The Supelclean LC- Si SPE tube was then washed with 10 ml hexane and 15 ml hexane/diethyl ether (95:5, v/v) to elute most of the triacylglycerols from the column. The remaining triacylglycerols and cholesterol were eluted by washing the column with 25 ml hexane/diethyl ether (90:10, v/v) and 15 ml of hexane/diethyl ether (80:20, v/v). The cholesterol oxides were eluted off the column with 10 ml acetone. The acetone fraction was evaporated to dryness under nitrogen and 41 redissolved in 50 pl pyridine plus 50 ul BSTFA in a 1 ml vial. This mixture was placed in the dark at room temperature for 1 hr to form the trimethylsilyl ether derivatives of the cholesterol oxides. GC analysis of the cholesterol oxide derivatives was performed using a Hewlett Packard 5890A gas chromatograph (Avondale, PA) equipped with a flame ionization detector. The cholesterol oxide derivatives were separated on a DB-l capillary column (15 m x 0.25 mm i.d., 0.1 um film thickness; J&W Scientific Inc., Ann Arbor, MI). Helium was used as carrier gas at a head pressure of 50 psi and with a flow rate of 27 ml/min. The oven temperature was programmed from 170°C to 220°C at a rate of 10°C per min, then increased to 234°C at a rate of 0.4°C per minute, held for 5 minutes, and then increased to 256°C at a rate of 2°C. The temperatures of the injection and detector ports were 270°C and 300°C, respectively. Fatty acid composition of whole milk lipids The determination of the fatty acid profiles of milkfat was performed by the method of Badings and Jong (1983). Whole milk powder lipids (100 mg) were dissolved in 6 ml pentane in a screw capped test tube, to which was added 0.06 ml 2M sodium methoxide solution. The contents of the capped tube were stirred vigorously for l min at room temperature on a vortex mixer. The sediment of sodium glycetolate was separated by 42 centrifugation at 1000 x g for 3 min. One ul sample of the clear supernatant was then taken for GC analysis. Fatty acid analyses were conducted on a HP 5890A system with a DB—225 capillary column (30 m x 0.25 mm i.d., 0.25 um film thickness; J&W Scientific Inc., Ann Arbor, MI) and a flame ionization detector. An initial oven temperature of 40°C was held for 5 min, then programmed to 200°C at a rate of 10°C per minute and held for 20 min. Injector and detector temperatures were 275 °C and 300 °C, respectively. The GC was run in the split mode with a split ratio of 16. Peaks were identified by comparison of retention times with standard reference mixtures of fatty acid methyl esters. Nitrite and nitrate analyses Extraction of nitrite and nitrate from whole milk powder was achieved using the method of the International Dairy Federation Standard 95 (1980). A sample of whole milk powder (10 g) was dissolved in 156 ml warm deionized water (SO-55°C). Then 12 ml 53.5 % (w/v) zinc sulfate solution, 12 ml 17.2 % (w/v) potassium hexacyanoferrate solution and 20 ml ammonium chloride buffer solution (pH 9.6-9.7) were added in the order described and mixed thoroughly using a vortex mixer after each addition. After 15 min, the mixture was filtered through Whatman filter paper (No.1) and the filtrate was collected. Nitrate is quantitatively reduced to nitrite by passing the filtrate through a copperized cadmium column. Nitrite 43 analyses were performed using a Lachat Quikchem Automated Flow Injection Ion Analyzer (Mequon, WI) in which nitrite was diazotized with sulfanilamide, and then coupled with N-(l— naphthyl)ethylenediamine dihydrochloride to produce a water- soluble magenta dye. Absorbance was measured at 520 nm. The nitrate content was calculated from the difference between the total nitrite (reduced nitrate plus original nitrite) content and the original nitrite content. Color measurement Color differences in the whole milk powder samples, as affected by storage and packaging, were determined using a Hunterlab ColorQUEST 45°/0° spectrophotometer (Hunter Assoc. Lab. Inc., Reston, VA). A standardized white tile (X=81.73, Y=86.51, Z=92.74) was used as a color reference. Headspace oxygen measurement The headspace oxygen contents of the packed samples were determined using a Carle 2153-B gas chromatography (Carle Inc., Anaheim, CA) with a molecular sieve 5A column (60/80 mesh, 8 m) and a column (50/80 mesh, 8 m) consisting of 20 % PPQ (Porapak Q) and 80 % PPN (Porapak N). A 500 ml sample of gas was withdraw from the package by a gas-tight syringe and injected into the sample loop of the GC. Helium was used as the carrier gas at a head pressure of 40 psi. The oven temperature was maintained at 50°C. 44 Statistical analysis The experiment was conducted as a four factor (drying method x packaging x temperature x time) split-plot design with two replications. Statistical analysis of the data for cholesterol oxides, TBARS values, and color differences was performed using a Bonferroni t-test to analyze specific contrasts among temperature treatments and Duncan's multiple comparisons test for drying methods and packaging systems (Ott, 1988). The analysis of variances were performed using MSTAT-C microcomputer statistical program (Michigan State University, East Lansing, MI, 1989). RESULTS AND DISCUSSION Nitrite and Nitrate Concentrations in Whole Milk Powders Oxides of nitrogen (NOx), generated by the spray drying process, may serve as initiators of lipid oxidation in whole milk powders. To evaluate the efficiency of different spray drying techniques in reducing the NOx level in the drying air, the concentrations of NOx were measured using Draegar tubes during the spray drying process (Kelly, personal communication). In addition, nitrite and nitrate contents, a reflection of NOx levels present in the spray drying chamber, in whole milk powders were also determined. By using the low NOx gas burner, the concentrations of NOx in the flue gases at the burner throat were between 1.0 and 1.65 ppm. Following dilution of the flue gases by the drying air, the levels of NOX were in the range 0.13 to 0.23 ppm. For the high NOx burner, 8 ppm nitrogen oxides were detected in the air at the inlet of the dryer. The average concentration of nitrite (NOZ') in direct high NOx gas~fired whole milk powders was 0.11 ug/g. There was no detectable nitrite in samples prepared by the direct low NOx and electrically heated processes. These data are not in agreement with the report of Kelly et al. (1989), who observed nitrite values of 0.81 and 0.59 ug/g for direct low NO gas-fired and electric heated samples of skim milk X 45 46 powders, respectively. The mean nitrate (NO3’) values were 5.33, 5.50 and 7.41 ug/g for electric heating, direct low NOx and direct high NOx gas-fired heating whole milk powders, respectively. The nitrate content in powders processed by the direct high NOx gas-fired dryer was significantly (P < 0.05) higher than those of samples processed by direct low NOx gas-fired and indirect electric heating. There was no significant difference in the nitrate contents of the direct lOW'NOx.and electrically heated samples. These results were generally consistent with the report of Kelly et al. (1989), however, their data were approximately two-fold higher than those presented here. Knipschildt (1986) noted that NOx in the drying air are not only created from the air by thermal processes, but can also originate from nitrogen in the fuel during the spray drying process. Kelly et al. (1989) also demonstrated that backgroundNOx levels in the ambient air affected the flue gas NOx levels and that the lowest values of NOx were obtained during the winter period. Therefore, the variations in NOx contents in spray-dried milk powders could be due to the different NOx levels in the ambient air, and to the fluctuation of nitrogen content in the natural gas used during the combustion process. In this study, it was apparent that the NOx in whole milk powders could be effectively reduced using a direct low NOx drying process. Based on these results, it was anticipated 47 that the oxidative stability of lipids in whole milk powders would also be improved by using low NOx drying processes, based on the fact that NOx are known initiators of lipid oxidation. Headspace Oxygen Content as Influepged by Packaging System Toiexamine.the scavenging ability'of oxygen.absorbers and the influence of oxygen on the oxidative stability of whole milk powders, headspace oxygen contents of packaged samples were measured during the storage period. The oxygen concentrations in the headspaces of the PE pouches, glass vials without oxygen absorbers, and glass vials with oxygen absorbers were 20.1 95, 20.2 % and 0.03 95, respectively, 3 tuyurs after' packaging (Table ‘4). The headspace oxygen contents of whole milk powders packaged in PE pouches were approximately equal to the percent oxygen in air and remained constant during the storage period. This was due to the oxygen permeability of the polyethylene film which allowed oxygen to continuously penetrate through the packaging material from the atmosphere. However, the oxygen contents in the headspace of whole milk powders packaged in glass vials ‘without.oxygen.absorbers.decreased.with.storage time, with the samples stored at 40°C showing the greater decreases. After 6 months storage at 20°C and 40°C, the oxygen contents in the headspaces were 18.8 % and 6.8 %, respectively. It is 48 Table 4 - Headspace oxygen contents of whole milk powders packed in various packaging systems and held at 20°C and 40°C for 6 months. Storage Packaging Headspace oxvqen ( % ) condition system 0 mo 6 mo 20°C PE 20.1:0.3 20.1:0.8 w/out OA 20.2:0.4 l8.8:1.0 w/ OA 0.03:0.01 0-24i0-07 40°C PE 20.1:0.3 20.2:0.5 w/out 0A 20.2io.4 6.8i2.5 w/ 0A 0.03:0.01 0.24:0.07 *PE — polyethylene pouch; w/out OA — glass vial without oxygen absorber; w/ OA - glass vial with oxygen absorber. postulated that the oxygen in the headspace reacted with the whole milk powders and the reactive rate was accelerated by the higher temperature. On the other hand, oxygen absorbers reduced headspace oxygen in the glass vials to almost zero percent within 3 days. The headspace oxygen content increased to 0.24 % after 6 months of storage. This could be attributed to a small leakage in the seals of the vials. Lipid Oxidation in Whole Milk Powders during Storage Lipid oxidation is a major problem associated with the long-term storage of whole milk powders (Coulter et al., 1951; Lea et al., 1953; Shipstead and Tarassuk, 1953; Boon et al., 49 1976; Tuohy, 1987). Many methods have been used for the measurement of lipid oxidation in such powders, including the 2-thiobarbituric acid test, peroxide value test, carbonyl measurement, oxygen absorption and sensory assessment. According to a comprehensive study by Tuohy (1987), each of the methods has its limitations IHowever, he reported that the TBA-reactive compounds (TBARS) value was a better index of milkfat oxidation than the peroxide value, oxygen absorption, or sensory assessment procedures. Tuohy (1987) also reported a reasonable correlation between the TBARS value and flavor acceptability of whole milk powder. Furthermore, the development of TBARS in whole milk powders during storage tended to precede an increase in the content of total carbonyl compounds, which are believed to be responsible for the off- flavors associated with oxidized milkfat (Boon, 1976). Therefore, it was concluded that the TBARS value is a more sensitive.index of oxidative changes in whole milk powder than the total carbonyl content. Based on the results reported by Tuohy (1987), only the TBARS value was used as a measure of lipid oxidation in this study. TBA-reactive compounds were isolated from whole milk powder by the steam distillation method of Tarladgis et al. (1964). 50 Effect of storaqepconditionslpn lipid stability in whole milk powders The initial TBARS values of whole milk powder samples produced by indirect electric heating, direct low NOx gas- fired and direct high NOx gas-fired heating were 0.11, 0.12 and 0.13 (obtained by multiplying absorbance at 532 nm by 6.2, a factor determined by Crackel et al., 1988), respectively. TBARS development at the initial stages of lipid oxidation in the powders was not significantly influenced (P < 0.05) by the drying procedure. Steen (1977) reported TBARS values in the range of 0.020 to 0.042 (expressed as absorbance at 530 nm) in commercially produced whole milk powders immediately after manufacture. These results generally agree with the TBARS values reported here, even though the procedure used by Steen (1977) was different from that used in this study. Using a different TBA procedure, Mettler (1973) also indicated that a TBARS value of 0.008 (absorbance at 530 nm) was representative of freshly produced whole milk powder. Ward (1985) concluded that the TBA assay was both operator- and method-dependent. Therefore, comparison of TBARS results from different laboratories must be undertaken with caution. The change in TBARS values of the milk powders with storage time was very distinct. During storage, the TBARS values increased in a linear manner for the samples packaged in PE pouches and glass vials without oxygen absorbers. A highly significant correlation was found between TBARS value 51 and storage time. When TBARS values were plotted against storage time, correlation coefficients of 0.87 to 0.96 were obtained indicating a strong linear relationship. The best- fit regression lines of TBARS value versus storage time at 20°C and 40°C for the samples manufactured by different drying methods are shown in Table 5. However, a poor correlation (r= 0.55) between TBARS value and storage time was found in the samples packaged in glass vials with oxygen absorbers. The damaging effect of temperature on the oxidative stability of whole milk powders packaged in PE pouches and in glass vials without oxygen absorbers was reflected in the development of TBARS, as illustrated in Figures 3 and 4. As expected, TBARS values of the powders stored at 40°C increased more rapidly than those of samples stored at 20°C. The means were significantly different (P < 0.01). However, the higher temperature did not influence TBARS development in powders packaged in glass vials with oxygen absorbers (Figure 5). After 6 months storage, the TBARS values of powders manufactured by direct high NOx gas-fired heating, packaged in PE pouches, and stored at 20°C and 40°C were 0.45 and 1.41, respectively. When the same powders were packaged in glass vials without oxygen absorbers, similar results were observed. On the other hand, there was only a slight decrease in TBARS values when the samples were packaged in glass vials with oxygen absorbers and stored at 40°C, compared to samples stored at 20°C. 52 Table 5 - The relationship between storage time and TBARS values of whole milk powders processed by different drying methods, packaged in various packaging systems and stored for 6 months at 20°C and 40°C. Storage Packaging Drying Best fit ** r condition system* method regression line 20°C PE electric Y=0.10+0.03X 0.867 low-NOx Y=O.ll+0.04X 0.927 high-NOx Y=0.13+0.05X 0.909 w/out 0A electric Y=0.12+0.04X 0.961 low-NOx Y=O.12+0.05X 0.941 high-NOx Y=O.l4+0.06X 0.882 w/ OA electric Y=0.11+0.004X 0.276 low-NOx Y=0.10+0.02X 0.701 high-NOx Y=0.12+0.02X 0.762 40°C PE electric Y=0.07+0.12X 0.958 low-NOx Y=0.24+0.20X 0.938 high-NOx Y=0.09+0.21X 0.885 w/out 0A electric Y=0.11+0.08X 0.942 low-NOx Y=O.l4+0.11X 0.916 high-NOX Y=0.16+0.14X 0.926 w/ 0A electric Y=O.10+0.01X 0.493 low-NOx Y=O.ll+0.008X 0.560 high-NOx Y=O.13+0.007X 0.500 TPE - polyethylene pouch; w/out 0A - glass vial without oxygen absorber; w/ OA - glass vial with oxygen absorber. HrX= storage time (months), Y= TBARS value. TBARS value. mg IDA/kg 1.8 1.8 1.4 1.2 1.0 0.8 0.0 0.4 0.2 0.0 63 zo’c 4o’c . o—o Low—N0: o—o "' A — A High-N0! A — A ’ D—D Electric l — l 'r I .- I .l. L A l . I/. i- - t!- ’- ,w/‘ 0 1 z 3 4 5 8 7 Storage time (months) Figure 3. Development of TBARS in whole milk powders in PE pouches during storage as influenced by method of drying and storage temperature TBARS value. mg IDA/kg 1.8 O O 1.5 _ 200 40c )- O—O Low—N0! .—. 1.4 - A—A High—N01 A—A ' ill—D ectric —— 1.2 _ E1 I I 1.0 . 1 t i 0.8 - g 0.6 P- ¥/2 0.4 _ // 8 - - . 0.2 - l 0.0 J 1 1 1 1 1 O 1 2 3 4 5 5 Storage time (months) Figure 4. Development of TBARS in whole milk powders in glass vials without oxygen absorbers during storage as influenced by method of drying and storage temperature TBARS value. mg MDA/ kg 1.8 1.8 - 1.4 1.2 1.0 0.8 0.8 0.4 0.2 0.0 65 20 °c 40°C 0 -— O Low—NO: O — O A — A High-NO! A — A D — El Electric I — l I fl.£8 T l l l J J 8 0 1 2 3 4 5 Storage time (months) Figure 5. Development of TBARS in whole milk powders in glass vials with oxygen absorbers during storage as influenced by method of drying and storage temperature 56 An average temperature coefficient (Q10) of 1.55 was reported by Tuohy (1987) for TBARS development in whole milk powders packaged in PE pouches over the temperature range 4- 45°C. The magnitude of the Q10 value for increase in TBARS values illustrates the profound effect of storage temperature on lipid oxidation in whole milk powders. Although the determination of Q10 for TBARS development could not be conducted in this study, general rates of increase of TBARS were obtained by comparing the TBARS values of samples stored at 40°C to those at 20°C. An average factor of 1.64 per 10°C was estimated for TBARS development in the samples packaged in PE pouches over a 6 month storage period. However, a mean value of 1.01 per 10°C was found for the samples packaged in glass vials without oxygen absorbers, which could indicate an insufficient oxygen content in the headspace. With samples packaged in glass vials in the presence of oxygen absorbers, an average value of 0.48 per 10°C was observed. Because the oxygen absorbers reduced the oxygen content of the headspace to less than 2 %, this limited amount of oxygen, therefore, slowed down the rate of lipid oxidation in.whole milk powders. Influence of drying method on lipid oxidation in whole milk powders The initial TBARS values of whole milk powder samples produced by the various drying methods were not significantly different from each other (Appendix 2). However, during the 57 storage of these powders, the drying method had a significant effect (P < 0.05) on TBARS development. In addition, the interactions between drying method and temperature of storage, drying method and packaging, drying method and storage time, were also pronounced. As shown in Figures 3 and 4, whole milk powder produced by direct high NOx gas-fired heating was found to deteriorate at a faster rate than powders from direct low NOx gas-fired and indirect electric heating systems. Powders manufactured by direct high NOx gas-fired heating and stored at 40°C had the highest TBARS values. Alternatively, samples produced by indirect electric heating and stored at 20°C showed the least increase in TBARS value over the storage period. After 6 months at 40°C, samples produced by direct high NOx gas-fired heating and packaged in PE pouches had a TBARS value of 1.41 compared to values of 1.33 and 0.86 for the direct low NOx gas-fired and indirect electric heating powders, respectively. These data were significantly different (P < 0.05). When the samples were packaged in glass vials without oxygen absorbers, the development of TBARS showed a Similar pattern, although the values were somewhat lower. However, there was only a relatively small change in TBARS values for powders packaged in the glass vials with oxygen absorbers. In these cases, the influence of drying methods was not significant (P < 0.05), even in those samples stored at 40°C. Using the TBARS value as an index of lipid oxidation, it 58 is apparent that whole milk powder manufactured by electric heating had the best oxidative stability, followed by powders produced by direct low NOx and high NOx gas—fired heating. It is assumed that the levels of nitrogen oxides generated during the drying processes are mainly responsible for the difference in lipid oxidation in the powders (Lightsey, 1982; Missler et al., 1985). As discussed earlier, by using the low NOx drying process, the levels of NOx in the drying air were substantially lower than those generated by the high NOx drying procedure. Moreover, powders prepared by direct high NOX gas-fired heating contained the highest levels of nitrate and nitrite, followed by the samples produced by direct low N0x gas-fired and indirect electric heating. These results coincide with the rate of lipid oxidation in these samples, the extent of oxidation increasing with increasing nitrogen oxide content (as nitrite and nitrate) in the powders. The initiation of oxidation in model systems of lipids and cholesterol by oxides of nitrogen has been demonstrated by Kamel et al. (1971) and Rhoem et al. (1971). Smith (1981) also pointed out that a variety of free radicals, including oxides of nitrogen are involved in the initiation of lipid oxidation. Thus, the difference in the levels of nitrogen oxides formed. during processing seems to be an important factor influencing the oxidative stability of the whole milk powders in this study. 59 Influence of packaging on lipid stability The effect of packaging on the development of TBARS in whole milk powders produced by the various drying methods was significant (P < 0.05). TBARS values were greatest in samples packaged in PE pouches, followed by the samples packaged in glass vials without and with oxygen absorbers. Temperature of storage and drying method also showed significant interaction effects (P < 0.05) the TBARS development, as described previously. When whole milk powder was packaged in PE pouches and stored at 40°C, a large increase in TBARS value was noted, irrespective of drying method used (Figure 3). A similar increase in the TBARS value of whole milk powder packaged in glass vials without oxygen absorbers was also observed, although not to the same extent as for powders packaged in PE pouches (Figure 4). This difference was attributed to a lower oxygen content in the headspace of the glass vials. As discussed earlier, the concentration of oxygen in the headspace of PE pouches was almost equal to that in the atmosphere. Therefore, powders packaged in PE pouches were continuously exposed to oxidative attack by atmospheric oxygen. In glass vials, however, the oxidative reaction could be attributed to the oxygen incorporated in the powder particles and the residual oxygen in the headspace. Furthermore, the larger surface area of the PE pouches could provide more access for the interaction of headspace oxygen 60 and the powdered milk samples (Nawar, 1985). On the other hand, TBARS development in samples in the packaging system containing the oxygen absorbers did not show the same trend as in those samples packaged in the other two systems (Figure 5). Powders manufactured by indirect electric heating showed a slight decrease in TBARS value after 3 months of storage, following by an insignificant (P < 0.05) increase after 6 months storage. A similar trend was also found in samples processed by the direct low NOx drying method. Although the TBARS values of samples prepared by direct high NOx drying consistently increased during the storage period, this increase was not significant (P < 0.05). Moreover, with the presence of oxygen absorber, the TBARS values of the powders were Significantly lower than those of powders in other packaging systems (P < 0.05). These results demonstrated that oxygen absorbers effectively retard or delay lipid oxidation in whole milk powders during storage, regardless of the processing method used. Change in fatty acid profiles of whole milk powders during storage Fatty acid methyl esters of total lipid extracted from whole milk powders were analyzed by gas liquid chromatography. Fatty acid profiles of whole milk powders over the storage period are presented in Tables 6—11. The initial fatty acid profiles of whole milk powder samples produced by the 61 9 .Uoom um mEDCOE 6 MOM pououm one monocom mm :H pomexomm .mponuos pooscoum mumozom EAHE macs; Mo moHHwouQ pace hunch 6 wanes 63 66.66 06.66 06.66 66.66 66.66 66.66 .60.66 66.06 60.66 .0660: 60.06 00.66 66.66 60.06 00.06 60.66 66.66 60.66 66.06 .066 06.0 06.0 66.0 06.0 06.0 66.0 06.0 06.0 00.0 6.660 60.6 06.6 66.6 60.6 66.6 66.6 60.6 66.6 66.6 6;:0 66.66 06.06 06.06 66.06 00.66 06.06 00.66 66.06 60.66 6.660 60.66 66.66 66.06 66.66 60.06 06.06 00.66 66.66 00.66 0.660 60.66 66.06 66.66 06.66 00.66 00.66 60.66 66.06 06.06 0.660 66.06 60.06 66.6 00.06 66.06 60.06 00.66 66.06 66.6 0 060 06.6 60.6 66.6 06.6 06.6 06.6 66.6 66.6 00.6 0 660 66.6 66.6 66.6 66.6 06.6 66.6 66.6 66.6 66.6 0.660 66.6 00.6 66.6 66.6 06.6 06.6 60.6 66.6 00.6 0H60 06.6 06.6 66.6 66.6 66.6 66.6 06.6 66.6 60.6 0H60 00.0 00.0 06.6 00.0 00.0 66.6 00.0 00.0 66.6 0.00 0E 6 OE m OE o OE 6 OE 6 GE 0 OE 6 OE m OE 0 73606.06 00066 xoz1066a xOZiso6 0600066o .0000 06 0:66:00: 6 you 60.0000 one monosom mm :6 609606066» .moonuoE 9.1.656 #:0065qu >9 COOSCOHQ 6.606309 0:..6E 06023 no 0066.60.69 06.06 >006; I N. OHQMB 64 66.66 66.06 06.66 60.66 60.06 66.66 06.66 00.06 60.66 .0660: 60.66 66.66 66.66 66.66 66.66 60.66 60.66 66.66 66.06 .066 66.0 66.0 66.0 06.0 06.0 66.0 06.0 66.0 00.0 6w660 06.6 60.6 66.6 66.6 66.6 66.6 66.6 66.6 66.6 6 660 60.06 66.66 06.06 66.06 66.66 06.06 00.66 60.66 60.66 6.660 00.06 66.66 66.06 00.66 60.66 06.06 06.66 00.66 00.66 0 660 66.06 60.06 66.66 06.06 06.06 00.66 06.06 66.06 06.06 0 660 66.66 06.06 66.6 66.06 66.06 60.06 66.06 66.06 66.6 0 060 66.6 66.6 66.6 66.6 66.6 06.6 66.6 66.6 00.6 0.660 66.6 06.6 66.6 60.6 66.6 66.6 06.6 60.6 66.6 0.660 60.6 66.6 66.6 66.6 66.6 06.6 66.6 66.6 00.6 0H60 66.6 66.6 66.6 66.6 06.6 66.6 66.6 66.6 60.6 0H60 00.0 00.0 06.6 00.0 00.0 66.6 00.0 00.0 66.6 0.00 OE 6 0E m OE o OE 6 OE 6 OE o OE 6 OE 6 OE o vamp606 x x 00066 0210660 oz:3o6 OMRUOQHO .0006 um mEUEOE 6 606 000000 6:6 mnmnuomnc comwxo 050:063 6666> 00660 :6 commxocm .mponuoE oc6>up #:0006660 Eonu 0600300 x66E 06023 HO 60666009 6606 xuusm i 6 06069 66.66 66.66 66.66 06.66 66.06 66.66 66.66 60.66 60.66 .0660: 66.66 66.66 66.66 06.06 06.66 60.66 06.06 06.66 66.06 .066 06.0 66.0 66.0 06.0 66.0 66.0 66.0 06.0 00.0 6 660 06.6 66.6 66.6 06.6 06.6 66.6 60.6 06.6 66.6 6 660 06.06 66.06 06.06 66.06 66.66 06.06 06.66 00.06 60.66 6H660 66.66 06.66 66.06 06.66 00.66 06.06 60.66 00.66 00.66 0 660 06.66 00.06 66.66 00.66 66.06 00.66 66.66 60.06 06.06 “#660 60.6 00.06 66.6 60.06 66.6 60.06 66.06 06.06 66.6 0 060 5 06.6 66.6 66.6 66.6 60.6 06.6 66.6 60.6 00.6 0 660 6 66.6 66.6 66.6 66.6 66.6 66.6 06.6 06.6 66.6 0.660 00.6 06.6 66.6 60.6 66.6 06.6 00.6 60.6 00.6 0.60 66.6 60.6 66.6 06.6 66.6 66.6 66.6 66.6 60.6 0W60 00.0 00.0 06.6 00.0 00.0 66.6 00.0 00.0 66.6 0.00 1 05 0 CE m CE 0 . CE 0 CE M CE 0 CE 0 CE m 05 O A VWUHUM _ 00066 xozi0660 x02.306 06000660 .0000 um wnucoE 6 How 60.6006 can 6606660686 cmmxxo 066020.63 6666> 666.3 :6 pomexoem .mconuoE 9.1.6.066 00.60.609.066 >0 couscoum 6.606309 026E 060663 60 606660.60 0606 00.060 1 6 0606.6. 66 00.66 00.66 06.66 «0.66 06.06 00.66 00.66 00.06 N0.m6 .ummc: 00.00 00.00 00.00 00.00 00.00 00.00 06.00 00.00 00.00 .000 00.0 00.0 00.0 00.0 00.0 60.0 00.0 00.0 00.0 6 060 06.6 06.0 06.6 06.6 06.m 06.6 06.6 66.6 06.6 6.060 00.06 00.06 06.06 60.06 66.66 00.06 60.06 60.66 60.06 .2060 66.06 00.06 60.06 00.06 00.06 00.06 06.06 06.06 00.06 0w06u 00.06 00.06 00.06 00.06 00.06 00.06 60.06 00.06 00.06 0.060 00.06 60.06 60.0 00.06 06.06 00.06 00.06 00.06 66.0 ou060 06.6 06.6 06.6 66.6 66.6 06.6 06.6 00.6 00.6 o 660 60.6 60.6 00.6 60.6 60.6 N0.m 00.6 00.6 60.6 0H06o 06.6 60.6 60.6 06.6 00.6 06.6 06.6 00.6 00.6 oumu 06.6 06.6 06.6 06.0 00.0 66.6 66.6 66.6 60.6 ou0o 00.0 00.0 00.6 00.0 00.0 06.6 00.0 00.0 66.6 ou0o OE 0 OS 6 OE o 08 0 CE m OE 0 GE 0 CE 0 CE 0 vamv6ow xuumm xozuc06n x02.306 06000060 .Uoom um maucofi 0 How Umuoum Ucm wumnuomnm cmmxxo £u63 w606> mwMHm :6 memxomm ~moonumfi @C6wuu ucwuwuu6c >9 Umosvoum mumvzom x665 060:3 mo mm66woum U600 xuumm I 06 @6208 67 60.66 00.06 06.66 06.66 00.06 00.66 00.66 60.06 60.66 .0000: 00.00 66.00 00.00 00.00 60.00 00.00 60.00 00.00 00.00 .000 60.0 00.0 00.0 00.0 00.0 60.0 60.0 00.0 00.0 6 060 60.6 60.6 06.6 06.6 60.6 06.6 66.6 00.0 06.6 6H060 00.06 00.66 06.06 00.06 00.06 00.06 66.66 00.66 60.06 6H060 00.06 00.06 60.06 00.06 00.06 00.06 06.06 60.06 00.06 0H060 66.06 66.06 00.00 00.06 00.06 00.06 00.06 00.06 00.06 6x060 60.06 66.06 60.0 60.06 00.0 00.06 60.06 06.06 66.0 o 060 06.6 60.6 06.6 06.6 00.6 06.6 06.6 00.6 00.6 0.660 60.6 00.6 00.6 60.6 00.6 60.6 60.6 00.6 60.6 o 060 60.6 06.6 60.6 06.6 00.6 06.6 06.6 00.6 00.6 0H00 06.6 06.6 06.6 06.6 00.6 66.6 06.6 00.6 60.6 0H00 00.0 00.0 06.6 00.0 00.0 06.6 00.0 00.0 66.6 0.00 CE 0 CE M 05 0 CE 0 03 M CE 0 CE 0 CE m OE O AwmvwmuHUM 00000 xozunm6n x02.066 06000060 .0000 um M63655 0 .60“ Umuouw 6cm mumnuomnw £093.60 £6063 m606> www.mm 666 00936065 .moocumfi 90606.60 068.000.6660 an 000660.69 muwv3om 066600 0606.3 00 000660009 60600 >uumm I 66 0690.6. 68 reason, the TBA procedure is a more sensitive indicator of the extent of lipid oxidation in milk powders than fatty acid profile determination. Total Cholesterol Content of Whole Milk Powders Based on a previous study by Engeseth (1990), a very efficient and reproducible method for total cholesterol determination is the modified method of Adams et al. (1986) . This method involves direct saponification of the samples, without derivatization, followed by gas liquid chromatographic analysis using packed columns. The total amounts of cholesterol found in whole milk powders ranged from 0.85 to 0.90 mg/g powder (3.39-3.52 mg/g fat). These results are quite similar to that (4 mg/g fat) presented in the Swedish Food Composition Tables (Anonymous, 1986) . However, they are higher than the values (2.0 mg/g fat) reported by Nourooz- Zadeh and Appelquvist (1988a) . Cholesterol Oxidation in Whole Milk Powders The formation of cholesterol oxides during the storage of dried milk products is not addressed in any detail in the scientific literature. Nourooz-Zadeh and Appelqvist (1988a) examined the occurrence of Cholesterol oxides in milk powder products during storage as related to processing technology, however, their research focused on the influence of thermal treatment prior to spray drying. In the present study, 69 samples of whole milk powders manufactured by the three different spray drying methods were analyzed for the presence of cholesterol oxidation products over the 6 month storage period. The influence of the three spray drying procedures as well as various packaging systems on cholesterol oxidation was investigated . The identification of cholesterol oxides was based on the the trimethylsilyl ether of relative retention times of derivatives of sample components compared to those synthetic cholesterol oxides. Quantification of the cholesterol oxides was based on peak area measurement of the sample peaks relative to that of the internal standard, 6— ketocholesterol. Since cholesterol is a labile compound and improper treatment during the analytical procedure may generate artifacts (Smith, 1981, 1987) , all the analyses were done under reduced light and without interruption. For the analysis of cholesterol oxides in whole milk powder, milkfat was extracted from the samples. The fat contents ranged from 24.9 to 26.5 %, and are similar to those reported previously for commercially prepared whole milk powders (Kurtz, 1974). The cholesterol oxide concentrations occurring in whole milk powders during storage as function of packaging and processing are shown in Tables 12, 13 and 14. Treatment (processing, packaging, temperature and storage time) effects, as well as significant interactions among treatments were 70 found (p < 0.01). The effects of each individual treatment on the formation of cholesterol oxides are discussed as follows. Effects of storage conditions on cholesterol stability The development of cholesterol oxidation products in whole milk powders during storage is illustrated in Figures 6, 7 and 8. Cholesterol oxides were not detected in freshly- processed whole milk powder samples, regardless of drying procedure. These results are consistent with those obtained by Nourooz—Zadeh and Appelqvist (1988a). The increase in the concentrations of cholesterol oxides in milk powders during storage was significant (Figures 6, 7, 8). Total cholesterol oxide contents in whole milk powders stored at 40°C for 3 months and 6 months were 0.61 and 4.45 %, respectively, when expressed as a percentage of the total cholesterol. This difference is due to the fact that time had a significant influence (P < 0.05) on the generation of cholesterol oxides in whole milk powders during storage. Cholesterol oxides were not generally detected in samples until they were stored for at least 3 months. Some cholesterol oxides were detected in samples stored for 3 months at 40°C, 7-ketocholesterol and the epimeric 7- hydroxycholesterols being the major oxidation products, followed by the isomeric 5,6-epoxycholestanols (Tables 12 and 13). The concentrations of these cholesterol oxides increased with further storage. When the samples were stored for six 71 Table 12 - Cholesterol oxide concentrations in whole milk powders packaged in PE pouches and held at two temperatures (20° and 40°C) for 6 months. Cholesterol oxides (pg/g lipid) Total Storage* condition 7-keto 7a-OH 7B-OH a-epo B-epo 20°C 3 month EH ND ND ND ND ND 0.0a L 1.1 ND ND ND ND 1.1a H ND ND ND ND ND 0.0a 6 month E 3.6 0.7 4.9 ND ND 9.1a L 3.7 ND 4.8 ND 3.1 11.7a H 17.5 3.3 10.2 0.8 6.4 38.1ab 40°C 3 month E 8.0 14.5 1.2 ND 3.1 26.8ab L 9.6 3.6 1.4 ND 4.8 19.4a H 21.4 8.9 14 3 1.4 13.5 59.5bc 6 month E 49.2 5.3 13.4 2.7 27.5 73.6C L 132.0 25.3 30.7 22.9 112.1 322.9d H 317.0 56.2 45.7 7.1 113.7 539.8e ND = Not detected, detection limit 0.1 pg. *No COPs were detected in whole milk powders at 0 month (i.e., immediately after processing) **E = electric, L = low-NOX, H = high-NOx abodeMean values in the column with different superscripts are significantly different (P <0.05). 72 Table 13 — Cholesterol oxide concentrations in whole milk powders packaged in glass vials without oxygen absorbers and held at two temperatures (200 and 40°C) for 6 months. Cholesterol oxides (pg/g lipid) Total Storage* condition 7-keto 7a-OH 7B-OH a-epo B-epo 20°C 3 month En ND ND ND ND ND 0.0a L ND ND ND ND ND 0.0a H ND ND ND ND ND 0.0a §_mgnth E 2.2 0.6 5.5 ND 2.9 11.23 L 7.6 1.0 6.2 ND 3.4 18.1a H 12.0 4.8 11.7 ND 4.6 33.0ab 40°C 3 month E 5.4 2.6 3.6 ND 6.8 18.5a L 15.1 3.2 5.8 2.6 3.8 30.4ab H 13.6 6.7 9.4 ND ND 29.7ab 6 month E 35.0 8.2 16.6 1.4 21.8 82.9c L 54.9 14.8 25.5 4.2 39.5 138.9d H 82.1 22.7 38.7 4.6 55.6 203.5e ND = Not detected, detection limit 0.1 pg. *No COPs were detected in whole milk powders at 0 month (i.e., immediately after processing) **E = electric, L = low-NOX, H = high-NOx adeeMean values in the column with different superscripts are significantly different (P <0.05). 73 Table 14 - Cholesterol oxide concentrations in whole milk powders packaged in glass vials with oxygen absorbers and held at two temperatures (20° and 40°C) for 6 months. Cholesterol oxides (pg/g lipid) Total Storage* condition 7-keto 7a—OH 7B-OH a—epo B-epo 20°C 3 month EH ND ND ND ND ND 0.0a L ND ND ND ND ND o.oa H ND ND ND ND ND 0.0a §_mgnth E ND ND 2.9 ND ND 2.9a L ND ND 4.2 ND ND 4.2a H 9.4 ND 5 4 ND ND 14.8a 40°C 3 month E ND ND ND ND ND 0.0a L ND ND ND ND ND 0.0a H ND ND ND ND ND 0.0a 6 month E 1.2 ND ND ND ND 1.2a L ND ND ND ND ND 0.0a H 2.5 ND ND ND ND 2.5a ND = Not detected, detection limit 0.1 pg. *No COPs were detected in whole milk powders at 0 month (i.e., immediately after processing) ** E = electric, L = low-NOX, H = high—NOx aMean values in the column with different superscripts are significantly different (P <0.05). Total COPs ( x ) 74 20.0 zo'c 40°C O—O Low—NO: 0—0 'I’ 16.0 — A —A High—N01 A—A A El—El Electric l—l J. 12.0 — I 8.0 ~ 1 4.0 - 0 E ' I L ///:s _______/A 0.0 . L J— o 1 2 3 5 a 4 Storage tlme (months) Figure 6. Total cholesterol oxidation products (COPs) (expressed as a percentage of total cholesterol) in whole milk powders packaged in PE pouches as influenced by storage time and temperature Total COPs ( Z ) 1. 75 20.0 18.0 - H 9’ O n— 5” O l 4.0 - ...l 20 ”c 4o'c O —- O Low—NOx O — O A —— A High—NOx A —— A El — El Electric I —- I 0 ... to to .p 0| 0 «1 Storage time (months) Figure 7. Total cholesterol oxidation products (COPs) (expressed as a percentage of total cholesterol) in whole milk powders packaged in glass vials without oxygen absorbers as influenced by storage time and temperature Total COPs ( x ) '78 20.0 20'c 40°C O—O Low—NOx e—e 15'0 ‘ A—A High—NOx x—x D—D Electric l—l 12.0 - 0.0 - 4.0 - col 1 ' I i 0 1 2 s 4 5 6 Storage time (months) Figure 8. Total cholesterol oxidation products (COPs) (expressed as a percentage of total cholesterol) in whole milk powders packaged in glass vials with oxygen absorbers as influenced by storage time and temperature 77 months at 20°C and 40°C, all showed an increase in the concentration of individual cholesterol oxides. The predominant cholesterol oxide in all samples was 7— ketocholesterol, which agreed with the observation of Nourooz- Zadeh and Appelqvist (1988a). These researchers reported higher concentrations of 7—ketocholesterol compared to other cholesterol oxides in 12-month—old whole milk powder stored at 20°C. The concentrations of 7B-hydroxycholesterol and B- epoxide were higher than those of 7a-hydroxycholesterol and a- epoxide, respectively. Side-chain hydroxylated cholesterol oxides and Sa—cholestane,35,5,6B-triol were not detected in the samples after 6 months storage. A similar finding was reported by Nourooz—Zadeh and Appelqvist (1988a). The data indicate that the ratio of each cholesterol oxide to cholesterol increased during storage. However, on average, the secondary cholesterol oxides did not increase to the same extent as the primary COPs, i.e., those generated through oxidation of cholesterol at C7. The total concentration of C7 oxidation products represented 0.63 % and 2.57 % of the total cholesterol in whole milk powder samples stored for 3 and 6 months, respectively. On the other hand, the concentration of secondary cholesterol oxides accounted for only 0.15 and 1.05 % of the original cholesterol in samples stored for 3 months and 6 months, respectively. These results suggest that cholesterol oxidation products at the allylic position, i.e., C7, will be formed more abundantly 78 than those at the double bond itself, i.e., positions 5 and 6, in cholesterol of whole milk powder. Smith (1981) noted that this is probably because epoxidation is a secondary process which is dependent on the presence of 7-hydroxycholesterols. It was observed that 7-ketocholesterol was formed more extensively than any other cholesterol oxide throughout the storage period, especially when the samples were stored at 40°C. The average percentages of 7-ketocholesterol in the total cholesterol oxides were 39.6 and 48.2 % for samples stored at 20°C and 40°C, respectively. According to Smith (1973), thermal decomposition of C7 hydroperoxides results in the formation of 7a- and 7B-hydroxycholesterol and 7- ketocholesterol. Therefore, it is assumed that at the higher temperature or in the absence of water, the decomposition of the 7—hydroperoxides via dehydration to 7-ketocholesterol is more likely than the formation of epimeric 7- hydroxycholesterols. Moreover, the concentration of 7B— hydroxycholesterol was predominant over the 70- hydroxycholesterol, as was B-epoxide over the a-epoxide. The ratio of B-epoxide to a—epoxide was approximately 2.9/1, and the ratio of 7B-hydroxy to 7a-hydroxycholesterol was 1.5/1. These data were similar to the findings of Nourooz-Zadeh and Appelqvist (1988a), who reported ratios of 1.6/1 and 1.9/1 for B-epoxide/a—epoxide and 7B-hydroxy/7a-hydroxycholesterol, respectively, in whole milk powders stored for 12 months. These results could be explained by the greater thermodynamic 79 stability of the equatorial over the axial conformation (Smith, 1981). Formation of 7-ketocholesterol was nearly linear (r= 0.88) with respect to storage time at 40°C. The formation of 7B-hydroxycholesterol and B—epoxide followed similar trends as well, with correlations of 0.84 and 0.77, respectively. Nevertheless, the linear relationships of storage time and the generation of 7—ketocholesterol, 7B-hydroxycholesterol and B— epoxide did not exist in samples stored at 20°C. The data presented in Tables 12 and 13 indicate that exposure to a higher temperature during storage caused noticeable concentrations of cholesterol oxides in whole milk powders. The difference between means of temperature treatments was significant (P < 0.05). When samples were packaged in PE pouches or in glass vials without oxygen absorbers, the average concentrations of cholesterol oxides in samples stored at 40°C for 6 months were approximately three times greater than those in samples stored at 20°C. Furthermore, the quantities of secondary cholesterol oxides were increased at 40°C. The rate of formation of 7B- hydroxycholesterol was greater than that of B-epoxide at 20°C, however, B-epoxide was formed in higher quantities than 7B- hyniroxycholesterol when the samples were stored at 40°C. Because B-epoxide is a secondary oxidation product of cholesterol, it is apparent that the elevated temperature accelerated the oxidative process. Smith (1981) indicated 80 that the exposure of cholesterol to high temperatures for a relatively long period in the presence of oxygen may initiate the allylic free radical reaction at C7. Following the formation of C7 peroxy radicals, a series of free radical reactions take place which results in the formation of stable oxidative products. Thus, storage of samples at 40°C can lead to a more intense oxidative degradation of cholesterol than when stored at 20°C. However, the higher temperature did not have the same impact on the oxidative stability of cholesterol when the samples were packaged in glass vials with oxygen absorbers, which was mainly due to insufficient oxygen in the headspace. In conclusion, the levels of cholesterol oxides in whole milk powders are greatly affected by temperature and storage time. The rate of oxidation for all cholesterol oxides was greatest between month 3 and month 6. Effect of drying method on cholesterol oxidation As mentioned previously, there was no detectable cholesterol oxides in freshly processed whole milk powder samples. These results indicated that the drying method had initially no influence on the presence of cholesterol oxides in fresh products. Similar results were found for samples stored for 3 months at 20°C. Although cholesterol oxides in the samples stored at 40°C were detected after 3 months of storage, the drying methods did not show any significant 81 effect (P < 0.05) on the formation of cholesterol oxides. After 6 months of storage at 20°C, powders prepared by direct high N0x gas-fired heating had the highest cholesterol oxide content, followed by the samples produced by direct low NO gas-fired and indirect electric heating systems. X Nevertheless, there was no significant difference (P < 0.05) among the means. The methods of drying, however, did significantly (P < 0.05) affect the concentrations of cholesterol oxides in whole milk powders stored at 40°C for 6 months. As shown in Figures 9, 10 and 11, the difference in the total cholesterol oxide contents and the individual cholesterol oxides is related. to drying' methods Total cholesterol oxides in powders manufactured by the direct high NOx gas-fired heating process increased rapidly from 59.5 to 539.8 pg/g lipid for samples packaged in PE pouches, and from 29.7 to 203.5 pg/g lipid for samples packaged in glass vials without oxygen absorbers when stored for 3 months at 40°C. The powders prepared by the direct low NOx gas-fired heating process followed a similar pattern but contained lower amounts of the cholesterol oxides. There was only a relatively small change in the total cholesterol oxide contents of powders produced by indirect electric heating compared to those from direct gas-fired heating. The higher concentrations of cholesterol oxides in direct gas-fired whole milk powders can be related to the processing ‘A—La“ Concentration of COPS (ug/ g lipid) 82 90- 80‘ 70- 60- 50- 40‘ 30- 20— 10~ §nl1fl [:1 Low—NOx - High—NOx [SS Electric 7-Keto 7a—OH 70—011 a—epo p—epo Cholesterol oxides Figure 9. Concentration of cholesterol oxidation products (COPs) in 6 month-old whole milk powders processed by various drying methods. packaged in PE pouches. and stored at 400C Concentration of COPS (ug/ g lipid) 35 30 _ D Low—NOx 25 .. ES! Electric 20- 5-1 83 - High—NOx 7—Keto 7a—OH 7p—0H a—epo p—epo Cholesterol oxides Figure 10. Concentration of cholesterol oxidation products (COPs) in 6 month-old whole milk powders processed by various drying methods. packaged in glass vials without oxygen absorbers. and stored at 40 C Concentration of COPS (ug/g lipid) 1.6 1.4- 1.2- 1.0- 0.8- 0.6- 0.4- O.2~ [:1 Low—NOx - High—NOx IE Electric 0.0 7-Keto "la-0H 7p—0H a—epo fi—epo Cholesterol oxides Figure 11. Concentration of cholesterol oxidation products (COPs) in 6 month-old whole milk powders processed by various drying methods, packaged in glass vials with oxygen absorbers. and stored at 40’c 85 conditions which generated greater quantites of NOx, known initiators of lipid oxidation. The levels of NOx in the drying air and the resultant concentrations of nitrite and nitrate in powders prepared by the different drying processes tended to parallel lipid oxidation (as measured by TBARS) and cholesterol oxidation. Effect of packaging on cholesterol oxidation The extent of cholesterol oxidation in whole milk powders was related to the packaging conditions, as shown in Figures 12 and 13. The interactions between packaging and processing, temperature and storage time were also significant (P < 0.01). It is apparent that samples packaged in PE pouches had the highest concentrations of cholesterol oxides, while those in the glass vials with oxygen absorbers had the least oxides. Total cholesterol oxide concentrations in milk powders processed by direct high NOx gas-fired heating and packaged in PE pouches increased from 1.7 % of the original cholesterol content after 3 months at 40°C, to 15.6 % after 6 months (Figure 6). In the glass vials without oxygen absorbers, the milk powders processed by the same drying procedure had cholesterol oxide concentrations of 0.86 % to 5.9 % of the original cholesterol after 3 months and 6 months of storage, respectively (Figure 7). In the packaging system with oxygen absorbers, the cholesterol oxidation products accounted for only 0.07 % of the total cholesterol content after 6 months of 86 storage at 40°C (Figure 8). Considering the interaction of packaging and processing, powders produced by direct gas—fired heating and packaged in PE pouches exhibited the highest concentrations of cholesterol oxides, followed by the samples packaged in glass vials without. oxygen absorbers (Figures 12 and 13). Moreover, samples from direct high NOX gas-fired heating had greater concentrations of cholesterol oxides than those from direct low NOx gas—fired heating. Samples prepared by indirect electric heating and packaged in PE pouches or in the system without oxygen absorbers achieved moderate levels of oxidation of cholesterol, but there was no significant difference between the mean values at the 95 % confidence level. In the packaging systems with the oxygen absorbers, samples showed the least extent of cholesterol oxidation, irrespective of the drying method used in powder preparation. Moreover, as discussed above, cholesterol oxides started to appear in detectable concentrations in samples packaged in PE pouches and glass vials without oxygen absorbers after 3 months, and the amounts increased substantially over the next 3 months. On the other hand, the oxygen absorbers effectively prevented oxidative changes in cholesterol during the 6 month storage period. The results demonstrated that the oxygen content in the headspace plays an important role in determining the susceptibility of cholesterol to oxidation, and the formation of cholesterol oxides could be significantly Total COPS content (ug/g lipid) 50- 60 8'7 40- 30- 20- 10- [:1 PE pouch - glass vial w/out 0A [3 glass vial w/ 0A l his 0 electric low-NOx high—NOx Drying methods Figure 12. Total cholesterol oxides in whole milk powders processed by different drying methods when packaged in various packaging systems and stored at 20°C for 6 months Total COPS content (ug/g lipid) 700 88 600 - 500 - 400 — 300 - 200 - 100- [:1 PE pouch — glass vial w/out 0A l 531 glass vial w/ 0A fill electric low—NOx high—NOx Drying methods Figure 13. Total cholesterol oxides in whole milk powders processed by different drying methods when packaged in various packaging systems and stored at 40°C for 6 months 89 minimized by packing whole milk powders with oxygen absorbers. Correlation of TBARS Values and Cholesterol Oxide Concentrations The correlation of TBARS values in relation to cholesterol oxide concentrations in whole milk powders is illustrated in Table 15. It is apparent that when the concentrations of cholesterol oxides increased, the TBARS values in the whole milk powders increased linearly for those powders packaged in PE pouches and in glass vials without oxygen absorbers. Generally, the changes of cholesterol oxide content over storage times agreed with those for the TBARS values for all treatments, except for the samples packaged in glass vials with oxygen absorbers. Correlations of 0.86, 0.82 and 0.79 were found between the TBARS values and the total cholesterol oxide contents in whole milk powders manufactured by direct high NOX, low NOx gas-fired heating and indirect electric heating, respectively. These results indicate that the oxidative rate of cholesterol in whole milk powder is proportional to milkfat oxidation under storage and packaging conditions with sufficient headspace oxygen. Color Change in Whole Milk Powders during Storage Changes in color (expressed as L, a, b Hunter values) of whole milk powders during storage are listed in Tables 16, 17 and 18. These data indicate that color changes in the samples 90 Table 15 - The relationship of TBARS values and cholesterol oxide concentrations in whole milk powders manufactured by different drying methods, when packaged in various packaging systems and held at 20°C and 40°C for 6 months. Storage Packaging Drying Best fit * r condition system* method regression line 20°C PE electric Y=-0.33+19.31X 0.861 low-NOx Y=-0.32+18.52X 0.755 high-NOx Y=-O.63+34.34X 0.876 w/out OA electric Y=-O.20+13.11X 0.668 low-NOx Y=-0.43+23.76X 0.760 high-NOx Y=-0.70+33.61X 0.835 w/ 0A electric Y=-0.l4+14.95X 0.700 low-NOx Y=-0.14+12.30X 0.655 high-NOx Y=-O.ll+l4.15X 0.295 40°C PE electric Y=-0.57+43.86X 0.900 low-NOX Y=-l.67+80.47X 0.920 high-NOx Y=-2.48+112.8X 0.923 w/out 0A electric Y=-0.51+44.81X 0.739 low-NOx Y=-0.87+52.20X 0.832 high-NOx Y=-l.23+60.74X 0.789 w/ 0A electric Y= 0.02-0.70X —0.106 low-NOX ——- high-NOx Y=0.035-O.79 -0.051 ?PE — polyethylene pouch, w/out 0A - glass bottle without oxygen absorber, w/ 0A - glass bottle with oxygen absorber. **X= TBARS value, Y= cholesterol oxide concentration (expressed as a percentage of original cholesterol content). 91 Table 16 — Changes in color characteristics of whole milk powders prepared by different drying methods, when packaged in PE pouches and stored for 6 months at 20°C and 40°C. Storage Hunter Drying methods condition parameter electric low-NOx high-NOx 20°C (months) 0 L-value 92.25 91.38 92.46 3 91.07 90.71 91.02 6 92.36 91.79 92.82 0 a—value -2.94 -3.18 —3.10 3 -3.00 -3.23 -3.10 6 —3.11 -3.34 -3.30 0 b-value 13.29 15.36 13.63 3 12.91 14.94 13.13 6 13.05 15.10 13.40 40°C (months) 0 L-value 92.25 91.38 92.46 3 90.88 91.11 91.04 6 91.68 91.17 91.90 0 a-value -2.94 —3.18 -3.10 3 -3.06 -3.40 -3.48 6 -3.22 -3.49 -3.56 0 b-value 13.29 15.36 13.63 3 12.79 15.15 13.25 6 13.00 14.77 13.44 L: measures lightness and varies from 100 for perfect white to zero for black. a: measures redness when positive, grey when zero and greeness when negative. measures yellowness when positive, grey when zero, and blueness when negative. U 92 Table 17 - Changes in color characteristics of whole milk powders prepared by different drying methods, when packaged in glass vials without oxygen absorbers and stored for 6 months at 20°C and 40°C. Storage Hunter Drying methods condition parameter electric low-NOx high-NOX 20°C (months) 0 L—value 92.25 91.38 92.46 3 90.86 90.59 91.27 6 91.53 90.81 91.65 0 a—value —2.94 -3.18 -3.10 3 -2.93 -3.27 -3.10 6 -3.11 -3.35 -3.28 0 b-value 13.29 15.36 13.63 3 13.10 15.29 13.33 6 13.24 15.23 13.65 40°C (months) 0 L-value 92.25 91.38 92.46 3 91.12 90.69 91.01 6 91.61 90.54 91.29 0 a-value -2.94 —3.18 -3.10 3 -3.08 -3.35 -3.19 6 -3.17 -3.40 -3.28 0 b-value 13.29 15.36 13.63 3 12.97 15.02 13.33 6 13.12 15.05 13.68 L: measures lightness and varies from 100 for perfect white to zero for black. a: measures redness when positive, grey when zero and greeness when negative. b: measures yellowness when positive, grey when zero, and blueness when negative. 93 Table 18 - Changes in color characteristics of whole milk powders prepared by different drying methods, when packaged in glass vials with oxygen absorbers and stored for 6 months at 20°C and 40°C. Storage Hunter Drying methods condition parameter electric low-NOx high-NOx 20°C (months) 0 L-Value 92.25 91.38 92.46 3 91.36 90.90 91.13 6 91.72 90.91 91.78 0 a-value —2.94 -3.18 -3.10 3 -2.93 -3.14 —2.93 6 —3.04 -3.26 -3.10 0 b-Value 13.29 15.36 13.63 3 13.17 15.17 13.38 6 13.35 15.35 13.59 40°C (months) 0 L-value 92.25 91.38 92.46 3 91.09 89.87 90.88 6 91.08 89.73 90.72 0 a—value -2.94 -3.18 —3.10 3 -2.85 -2.92 -2.82 6 —2.97 -2.91 -2.86 0 b—value 13.29 15.36 13.63 3 13.98 16.08 14.53 6 14.33 17.10 15.21 L: measures lightness and varies from 100 for perfect white to zero for black. a: measures redness when positive, grey when zero and greeness when negative. b: measures yellowness when positive, grey when zero, and blueness when negative. 94 were not related to the drying procedures, packaging systems employed, or the temperatures of storage. The change of color in milk powders is mainly attributed to the nonenzymatic browning reaction (Maillard reaction) between lactose and milk protein (Palombo et al., 1984; Min et al., 1989). High processing temperatures, moisture content and prolonged storage are the major factors involved in the susceptibility of dry milk powders to the Maillard reaction (Bender, 1972; Labuza, 1972). Based on these observations, whole milk powders stored at 40°C were expected to have more brown color than those stored at 20°C. However, the changes in L value (lightness) and b value (yellowness) of the whole milk powders stored at the two temperatures were not significantly different (P < 0.05) during the entire storage period. On the other hand, powders produced by the low NOx drying procedure had significantly ( P < 0.05) higher b values and lower L values than those produced by the other drying methods, which was unexpected. These results indicated that the Hunter parameters may be not sensitive enough to detect the color changes in the whole milk powders of this experiment. Furthermore, a poor correlation (r= 0.13) between b value and TBARS value was observed in all the samples, which suggested that color changes in whole milk powders as measured by Hunter parameters were not related to the extent of lipid oxidation. Min et al. (1989) reported that the brown color in whole milk powder increased with storage time over a 72 hour 95 storage period at 65°C. Their results demonstrated that measuring the brown color could be a more suitable method for detecting color change in whole milk powders during storage. SUMMARY AND CONCLUSIONS The effects of various drying' methods and packaging systems on cholesterol oxidation, as well as the relationship between milkfat oxidation and the formation of cholesterol oxides in whole milk powders were the focal points of this study. Powders manufactured by direct high NOx gas—fired heating had the highest TBARS values, followed by the powders produced by the direct low NOx gas-fired and indirect electric heating processes. Lipid oxidation proceeded at a significantly (P < 0.05) higher rate in all samples at 40°C than at 20°C, with the powders produced by direct high NOx drying process undergoing the greatest amount of oxidation. The method of drying also significantly (P < 0.05) affected the concentrations of COPs in whole milk powders. Total COPs in powders produced by the direct high NOx system increased most rapidly during the storage period, followed by the samples prepared by the low NOx drying procedure. There was only a relatively small change in total COPs during storage of the powders produced by indirect electric heating. The differences in the TBARS values and the concentrations of cholesterol oxides in powders manufactured by the direct high NO direct low NOX, and indirect electrically heating x! processes indicated that nitrogen oxides play an important 96 97 role in the initiation of lipid and cholesterol oxidation reactions. The nature of the packaging systems greatly influenced the oxidative stability of whole milk powders during the storage period. Oxygen absorbers reduced headspace oxygen in the glass vials to approximately 3 % within 3 hours and to less than 2 % after 6 months of storage. Packaging whole milk powders in glass vials with oxygen absorbers was observed to have a pmotective effect on the stability of milkfat and cholesterol. TBARS values of powders packaged in PE pouches and in glass vials without oxygen absorbers significantly increased (P < 0.05) during the storage period. Samples packaged in glass vials with oxygen absorbers exhibited a much slower increase in TBARS values compared to those packed in the other two packaging systems. The oxygen absorbers effectively prevented oxidative changes in cholesterol over the 6 months storage period, even in those samples stored at 40°C. However, cholesterol oxides started. to appear’ in detectable concentrations 111 powders packaged in PE pouches after 3 months of storage, and the amounts increased substantially over the next 3 months. Similar oxidation products were present in powders packaged in glass vials without oxygen absorbers, but at lower concentrations. Exposure to the higher temperature facilitated the 98 oxidative degradation of cholesterol in whole milk powders. Storage of powders at 40°C led to a more rapid oxidation of cholesterol than in powders stored at 20°C. Formation of 7- ketocholesterol, 7B—hydroxycholesterol.and.B-epoxide increased linearly with respect to storage time at 40°C. In general, the rate of oxidation of cholesterol in whole milk powders was parallel to that of lipid oxidation under conditions with sufficient headspace oxygen. A strong linear relationship (r = 0.82) was observed between the TBARS values and the total cholesterol oxide contents in whole milk powders packaged in PE pouches and in glass vials without oxygen absorbers over the storage time . The conclusions drawn from this study are summarized as follows: 1. The large differences in the TBARS values of the powders confirmed that the TBARS value was a satisfactory indicator of lipid oxidation in whole milk powder. 2. The oxidation of lipid and cholesterol in whole milk powders can be minimized by packaging the powders in oxygen-impermeable packages containing oxygen absorbers. 3. Using a direct low NOx drying process can improve the oxidative stability of whole milk powders. 4. The extent of oxidation of lipid and cholesterol in whole milk powders can be reduced by lowering the storage temperature. APPENDICES 99 6066.566 mom 58 x02 306 000.660 0 ..60 50.66060 .6 6660:6396 €3.09 0.5:. 3 ammzu 100 .muwnuomnc cco>xo £063 m606> mmch :6 00m0xocm I <0 \3 “muwnuomnm commxo 050:063 m6c6> 00060 :6 oomcxomm I do uno\3 “£0509 0co6>n00>60m :6 pommxowm I ma.. .>60>6uocmm06 .66.o one N6.o .66.0 060 moonuoE mc6>uo X OZInm6£ 0:0 xOZI306 .06600060 >9 pmmmwooum mumpBOQ x66E 060:3 :6 demB 60c6m6uo 068a mo.on6.o no.0HmH.o 60.0H5H.O N0.0HNH.O m0.0HhH.O no.0HOH.O <0 \3 mo.OHhm.o NN.OHN®.O NH.OHN®.O hH.onm.O HH.Ome.O N0.0H®M.O do U50\3 mmooflHv.H om.onw.O HN.OHMM.H v0.0HV¢.O OH.OH®m.O vo.onm.o mm 0°00 ¢0.0HmN.O mo.onH.o no.0HHN.O M0.0HHH.O v0.0HMH.O N0.0HOH.O <0 \3 NH.OHF¢.O no.0HNm.o 00.0Hmm.o N0.0H@N.O v0.0Hmm.o 60.0HmN.O do v50\3 OH.OHm¢.O OmoonN.O 00.0me.o H0.0HMN.O wo.OHNm.O H0.0HON.O mm 0o0~ OE 0 OE 6 OE 0 OE m OE 0 OE m ..co606oc00 cocuoum xOZIco6n xOZI306 06600060 .0000 0:0 Doom 00 mauCOE 0 606 006000 0:0 woc6mcxocm m50660> c6 owmcxomm C053 .woonqu m m x6ocwmmd c6mup accumuu6o >0 owwmoooum mMOUBOQ x66E 060:3 :6 uc0E9060>mo mmdma I REFERENCES Abbot, J. and R. Waite. 1962. The effects of antioxidants on keeping quality of whole milk powder. I. Flavones, butylhydroxyanisole and nordihydroguaiaretic acid. J. Dairy Res. 29:55. Adams, M. L., D. M. Sullivan, R. L. Smith and E. F. Richter. 1986. Evaluation of direct saponification method for determination of cholesterol in meats. J. Assoc. Off. Anal. Chem. 69:844. Addis, P. B. 1986. Occurrence of lipid oxidation products in foods. Food Chem. Toxic. 24:1021. Addis, P. B. and S. W. Park. 1989. Role of lipid oxidation products in atherosclerosis. In Food Toxicology: A Perspective On The Relative Risks. Ed. R. A. Scanlan and S. L. Taylor. Marcel Dekker, Inc., New York. Altermark, D. and R. Hess. 1982. A new burner generation for clean air and energy conservation. 15th World Gas Conference, p.1. LausannezlGU/FI-82, C 018.4. Anonymous. 1986. "Swedish Food Composition Tables." "The National Food Administration, Uppsala, Sweden. Ansari, G. A. S. and L. L. Smith. 1978. Sterol metabolism. XII Cholesterol A-ring autoxidation. Chem. Phy. Lipids 22:55. Badings, H. T. and C. DE Jong. 1983. Glass capillary gas chromatography of fatty acid methyl esters. A study of conditions for the quantitative analysis of short— and long- chain fatty acids in lipids. J. Chromatogr. 279:493. Baranowski, A., C. W. M. Adams, 0. B. Bayliss High, and D. B. Bowyer. 1982. Connective tissue response to oxysterols. Atherosclerosis. 41:255. Bascoul, J., N. Domergue, M. Olle, and A. C. de Paulet. 1986. Autoxidation of cholesterol in tallows heated under deep frying conditions: Evaluation of oxysterols by GLC and TLC-FID. Lipids 21:383. Bender, V. H. 1972. Processing damage to protein food. J. Food Technol. 7:239. 101 .1 102 Berlin, E. and M. J. Pallansch. 1963. Influence of drying methods on density and porosity of milk powder granules. J. Dairy Sci. 46:8. Black, H. S.and D. R. Douglas. 1972. A model system for the evaluation of the role of cholesterol a—oxide in ultraviolet carcinogenesis. Cancer Res. 32:2630. Black, H. S.and D. R. Douglas. 1973. Formation of a carcinogen of natural origin in the etiology of ultraviolet light-induced carcinogenesis. Cancer Res. 33:2094. Blackburn, G. M., A. Rashid, and M. H. Thompson. 1979. Interaction of 50,60-cholesterol oxide with DNA and other nucleophiles. J. Chem. Soc. Chem. Commun. p.420. Boley, N. P. and R. K. Colwell. 1987. Comparison of methods for the determination of butyric acid in foodstuffs by gas-liquid chromatography and high-performance liquid chromatography. J. Chromatogr. 410:109. Boon, P. M., A. R. Keen and N. J. Walker. 1976. Off-flavor in stored whole milk powder. II. Separation and identification of individual monocarbonyl components. N. Z. J. Dairy Sci. Technol. 11:189. Brink, M. F. 1968. Symposium: Dairy Lipids and Lipid Metabolism. AVI Publishing Co., Westport, CT. Brown, S., and J. L. Goldstein. 1974. Suppression 3- hydroxy-3-methylglutaryl coenzyme A reductase activity and inhibition of growth of human fibroblasts by 7- ketocholesterol. J. Biol. Chem. 249:7306. Challis, B.C. 1982. Amine nitration and nitrosation by gaseous nitrogen dioxide. IARC Scientific Publications 41, p.11. Chan, J. T. and H. S. Black. 1974. Skin carcinogenesis: cholesterol—50,6a—epoxide hydrase activity in mouse skin irradiated with ultraviolet light. Science 186:1216. Chan, H. W. S. and G. Levett. 1977. Autoxidation of methyl linolenate: Analysis of methyl hydroxylinolenate isomers by high performance liquid chromatography. Lipids 12:837. Chan, H. W. S. 1987. Autoxidation of Unsaturated Lipids. Academic Press Inc., Orlando, FL. 103 Chen, H. W., A. A. Kandutsch, and C. Waymouth. 1974. Inhibition of cell growth by oxygenated derivatives of cholesterol. Nature 251:419. Chicoye, E., W. D. Powrie, and O. Fennema. 1968. Photooxidation of cholesterol in spray-dried egg yolk upon irradiation. J. Food Sci. 33:581. Christie, W. W., K. Connor and R. C. Noble. 1984. Preparative separation of milk fatty acid derivatives by high- performance liquid chromatography. J. Chromatogr. 298:513. Coulter, S. T., R. Jenness and W. F. Geddes. 1951. Physical and chemical aspects of the production, storage and utility of dry milk products. Adv. Food Res. 3:45. Crackel, R. L., J. I. Gray, A. M. Pearson, A. M. Booren, and D. J. Buckley. 1988. Some further observations on the TBA test as an index of lipid oxidation in meats. Food Chem. 28:187. Cullinane, N., S. Aherne, J. F. Connolly, and J. A. Phelan. 1984. Seasonal variation in the triglyceride and fatty acid composition of Irish butter. Ir. J. Fd. Sci. Technol. 8:1. Dugan, L. R. 1987. Fats. In: The Science of Meat and Meat Products. Ed. J. F. Price and B. S. Schweigert. Food and Nutrition Press, Inc., Westport, CT. Engeseth, N. J. 1990. Membranal lipid oxidation in muscle tissue: mechanism and prevention. Ph. D. Dissertation, Michigan State Univ., E. Lansing, MI. Ferretti, A., V. P. Flanagan, and J. M. Ruth. 1970. Nonenzymatic browning in lactose-casein model system. J. Agr. Food Chem. 18:13. FIL-IDF. 1980. Dried milks-determination of nitrate and nitrite contents method by cadmium reduction and photometry. Provisional International IDF standard 95:1980. Finocchiaro, E. T., and T. Richardson. 1983. Sterol oxides in foodstuffs: a review. J. Food Protect. 46:917. Finocchiaro, E. T., K. Lee and T. Richardson. 1984. Identification and quantification of cholesterol oxides in grated cheese and bleached butter oil. J. Am. Oil Chem. Soc. 61:877. 104 Flanagan, V. P., A. Ferretti, D. P. Schwartz, and J. M. Ruth. 1975. Characterization of two steroidal ketones and two isoprenoid alcohols in dairy products. J. Lipid Res. 16:97. Folch, J., M. Lees, and G. H. S. Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226:497. Frankel, E. N. 1979. Autoxidation. In Fatty Acids, Ed. E. H. Pryde, Am. 011 Chem. Soc. Champaign, I11., p.353. Frankel, E. N. 1962. Hydroperoxides. In Symposium on Foods: Lipids and Their Oxidation. Ed. H. W. Schultz, E. A. Day and R. O. Sinnhuber, AVI Publishing Co., Westport, CT. Chapter 3. p.51. Fujishima, D. 1977. Japan Patent No. 19:729. Gibbons, G. F., K. A. Mitropoulos and N. B. Myant. 1982. Biochemistry of Cholesterol. Elsevier Biomedical Press. New York, NY. Gumulka, J., J. St. Pyrek and L. L. Smith. 1982. Interception of discrete oxygen species in aqueous media by cholesterol: formation of cholesterol epoxides and secosterols. Lipids 17:197. Hall, C. W. and T. I. Hedrick. 1966. Drying Milk and Milk Products. AVI Publishing Co., Westport CT, U.S.A. Hammond, E. G. 1970. Stabilizing milk fat with antioxidants. Amer. Dairy Review. June, p.40. Harding, F. and R. Gregson. 1978. Nitrates and nitrites in skim milk and whey powders in the United Kingdom. Proceedings of the XXth International Dairy Congress. p.356. Havery, D. C., J. H. Hotchkiss and T. Fazio. 1982. Rapid determination of volatile N-nitrosamines in nonfat dry milk. J. Dairy Sci. 65:182. Higley, N. A., S. L. Taylor, A. M. Herian, and K. Lee. 1986. Cholesterol oxides in processed meats. Meat Sci. 16:175. Hilditch, T. P., and P. N. Williams. 1964. The Chemical Constitution of Natural Fats. 4th ed. John Wiley and Sons, Inc., New York, NY. 105 Hill, R. D., V. Van Leeuwen, and R. A. Wilkinson. 1977. Some factors influencing the autoxidation of milks rich in linoleic acid. New Zealand J. Dairy Sci. Technol. 12:69. Hunt, S. M. and J. L. Groff. 1990. The cell: a microcosm of life. In Advanced Nutrition and Human Metabolism. West Publishing Co., St. Paul, MN, Chapter 1. p.5. Imai, H., N. T. Werthessen, C. B. Taylor, and K. T. Lee. 1976. Angiotoxicity and atherosclerosis due to contaminants of USP—grade cholesterol. Arch. Pathol. Lab Med. 100:565. Imai, H., N. T. Werthessen, P. W. Subramanyam, P. W. Lequesne, and A. H. Soloway. 1980. Angiotoxicity of oxygenated sterol and possible precursors. Science 207:651. Isherwood, F. A. 1943. United Kingdom Patent No. 553991. Jansen, L. A. and R. H. C. Elgersma. 1985. Direct heating of drying air with natural gas in the preparation of milk powder. J. Soc. Dairy Tech. 38:134. Kamel, A. M., N. D. Weiner and A. Felmeister. 1971. Indentification of cholesterol nitrate as a product of the reaction between N02 and cholesterol monomolecular films. Chem. Phys. Lipids 6:225. Kandutsch, A. A. and H. W. Chen. 1973. Inhibition Of sterol synthesis in cultured mouse cells by 7a-hydroxy cholesterol, 7B-hydroxy cholesterol and 7-keto cholesterol. J. Biol. Chem. 248:8404. Kandutsch, A. A. and H. W. Chen. 1974. Inhibition of sterol synthesis in cultured mouse cells by cholesterol derivatives oxygenated in the side chains. J. Biol. Chem. 249:6057. Kandutsch, A. A. and H. W. Chen. 1978. Inhibition of cholesterol synthesis by oxygenated sterols. Lipids 13:704. Keenan, T. W., and S. Patton. 1970. Cholesterol esters of milk and mammary tissue. Lipids 5:42. Kelly, P. M. and J. Slattery. 1985. Direct gas-fired heating of a pilot-scale spray drier using natural gas for milk powder manufacture. J. Soc. Dairy Tech. 38:55. Kelly, P. M., J. I. Gray and J. Slattery. 1989. Direct '1ow- NOx'gas combustion heating of a spray drier during milk 106 powder manufacture. J. Soc. Dairy Tech. 42:14. King, J. 1955. Catalytic removal of oxygen from food containers. Food Mfg. 30:441. Knipschildt, M. E. 1986. Drying milk and milk products. In Modern Dairy Technology. Vol. 1. Ed. R. K. Robinson, Elsevier Applied Science Publishers LTD., England, Chapter 4, p.131. Korahani, V., J. Bascoul and A. Crastes De Paulet. 1982. Autoxidation of fatty acid esters in solid state and aqueous dispersions. Lipid 17:703. Korycka-Dahl, M. and T. Richardson. 1980. Symposium: oxidative changes in milk. Initiation of oxidative changes in foods. J. Dairy Sci. 63:1181. Kuhn, P. E., K. F. Weinke and P. L. Zimmerman. 1970. Oxygen scavenging system for flexible packaging of whole dry milk. Presented in 9th Milk Concentrates Conference, Pennsylvania State University, September, 15. Kumar, N. and O. P. Singhal. 1991. Cholesterol oxides and atherosclerosis: A review. J. Sci. Food Agric. 55:497. Kurtz, F. E. 1974. The lipids of milk: composition and properties. In Fundamentals of Dairy Chemistry (2nd Edition). Ed. B. H. Webb, A. H. Johnson and J. A. Alford. Pub. by AVI Pub. Co., Westport. CT, Chapter 4. p.125. Labuza, T.P. 1971. Kinetics of lipid oxidation in foods. CRC Crit. Rev. Food Technol. 2:355. Lea, C. H. 1953. Oxidation defects in milk and dairy products. Proc. XIII Int. Dairy Congress. 5:315. Leduc, E. H. 1980. Carcinogenic potential of cholesterol oxidation products. Final Report, U.S. Army Research Office. Grant NO. DAAG-29-77-G-OO73. Lee, K., A. M. Herian, and N. A. Higley. 1985. Sterol oxidation products in french fries and in stored potato chips. J. Food Protect. 48:158. Libbey, L. M., R. A. Scanlan, and J. F. Barbour. 1980. N- Nitrosodimethylamine in dried dairy products. Food Cosmet. Toxicol. 18:559. Lightsey, J. W. 1982. A mechanistic study of free-radical initiation of alkene oxidation by nitrogen dioxide. .4 107 Ph. D. Dissertation, Louisiana State University, Baton Rouge, LA. Luby, J. M., J. I. Gray, B. R. Harte, and T. C. Ryan. 1986. Photooxidation of cholesterol in butter. J. Food Sci. 51:904. Luby, J. M. 1982. Photooxidation of cholesterol in butter. Master Thesis, Michigan State Univ., E. Lansing, MI. Maerker, G. and J. Unruh. 1986. Cholesterol oxides. 1. Isolation and determination of some cholesterol oxidation products. J. Am. Oil Chem. Soc. 63:767. Maerker, G. 1987. Cholesterol autoxidation - current status. J. Am. Oil Chem. Soc. 64:388. Maude, A. H., C. J. Rodman, C. A. Styer, and W. C. Wilharm. 1925. United Kindom Patent No. 226512. Mettler, A. E. 1973. Observation on the use of 2- thiobarbituric acid for estimating deterioration in powder milk products. J. Soc. Dairy Technol. 26:84. Min, D. B., S. H. Lee, J. B. Lindamood, K. S. Chang, and G. A. Reineccius. 1989. Effects of packaging conditions on the flavor stability of dry whole milk. J. Food Sci. 54:1222. Missler, S. R., B. A. Wasilchuk, and C. Merritt. 1985. Separation and identification of cholesterol oxidation products in dried egg preparations. J. Food Sci. 50:595. Moore, J. L., T. Richardson, and C. H. Amundson. 1965. The chemical and physical properties of interesterified milk fat fractions. J. Am. Oil Chem. Soc. 42:796. Morgan, J. N. and D. J. Armstrong. 1989. Wide-bore capillary gas chromatography method for quantification of cholesterol oxidation products in egg york powder. J. Food Sci. 54:427. Mucha, T. J., N. J. Pallansch, W. I. Patterson, and A. Tamsma. 1961. Factors related to the flavor stability of foam—dried whole milk. J. Dairy Sci. 44:91. Nataf, B., O. Micklsen, A. Keys, and W. E. Peterson. 1948. The cholesterol content of cow's milk. J. Nutr. 36:495. 108 Nawar, W. W. 1985. Lipids. In Food Chemistry, Ed. 0. R. Fennema. Marcel Dekker, Inc., Nieman, C., E. H. Groot and W. J. Rooselaar. 1950. Acta Physiol. Pharmacol. Neerl., 1:488. Nourooz-Zadeh, J. and L-A Appelqvist. 1987. Cholesterol oxides in Swedish foods and food ingredients: Fresh eggs and dehydrated egg products. J. Food Sci. 52:57. Nourooz-Zadeh, J. and L-A Appelqvist. 1988a. Cholesterol oxides in Swedish foods and food ingredients: milk powder products. J. Food Sci. 53:74. Nourooz-Zadeh, J. and L-A Appelqvist. 1988b. Cholesterol oxides in Swedish foods and food ingredients: butter and cheese. J. Am. Oil Chem. Soc. 65:1635. Ott, L. 1988. An Introduction to Statistical Methods and Data Analysis. 3rd Ed., PWS—Kent Pub. Co., Boston, MA. Palombo, R., A. Gertler, and I. Saguy. 1984. A simplified method for determination of browning in dairy powders. J. Food Sci. 49:1609. Parish, E. J., V. B. B. Nanduri, H. H. Kohl and F. R. Taylor. 1986. Oxysterolszchemical synthesis, biosynthesis and biological activities. Lipids 21:27. Park, S. W. and P. B. Addis. 1987. Cholesterol oxidation products in some muscle foods. J. Food Sci. 52:1500. Park, S. W. and P. B. Addis. 1986a. Identification and quantitative estimation of oxidized cholesterol derivative in heated tallow. J. Agri. Food Chem. 34(4):53. Park, S. W. and P. B. Addis. 1986b. Further investigation of oxidized cholesterol derivatives in heated fat. J. Food Sci. 51:1380. Park, S. W. and P. B. Addis. 1985a. HPLC determination of C-7 oxidized cholesterol derivatives in foods. J. Food Sci. 50:1437. Park, S. W. and P. B. Addis. 1985b. Capillary column gas— liquid chromatographic resolution of oxidized cholesterol derivative. Anal. Biochem. 149:275. Parks, 0. W. 1980. Cholesterol esters of skim milk. J. Dairy Sci. 63:295. 109 Parks, 0. W., 1974. The lipids of milk: deterioration. In Fundamentals of Dairy Chemistry. (2nd Edition). Ed. B. H. Webb, A. H. Johnson and J. A. Alford, AVI Publishing. Co., Westport. CT, Chapter 5. p.240. Patton, S., and R. D. McCarthy. 1963. Structure and synthesis of milk fat. IV. Role of the mammary gland with special reference to the cholesterol esters. J. Dairy Sci. 46:396. Patton, S. 1962. Dairy products. In Symposium on Foods: Lipids and Their Oxidation. Ed. H. W. Schultz, E. A. Day and R. O. Sinnhuber, AVI Publishing Co., Westport, CT, Chapter 10, p.190. Peng, S. K., C. B. Taylor, and B. Mikkelson. 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. Peng, S. K., C. B. Taylor, E. H. Mosbach, W. Y. Huang, J. Hill, and B. Mikkelson. 1982. Distribution of 25- hydroxycholesterol in plasma lipoproteins and its role in atherogenesis. Atherosclerosis. 41:395. Peng, S. K., J. C. Hill, R. J. Morin and C. B. Taylor. 1985. Influence of cholesterol oxidative derivatives on membrane bound enzymes in cultured aortic smooth muscles. Proc Soc. Exp. Biol. Med. 180(1):126. Peng, S. K. and R. J. Morin. 1987. Effect of membrane function of cholesterol oxidative derivatives in cultured aortic smooth muscle cells. Artery 14(2):85. Pie, J. E., K. Spahis, and C. Seillan. 1991. Cholesterol oxidation in meat products during cooking and frozen storage. J. Agr. Food Chem. 39:250. Rawls, H. R. and P. J. Van Santen. 1970. A possible role for singlet oxygen in the initiation of fatty acid autoxidation. J. Amer. Oil Chem. Soc. 47:121. Rhoem, J. N., J. G. Hadley, and D. B. Menzel. 1971. Oxidation of unsaturated fatty acids by ozone and nitrogen dioxide: A common mechanism of action. Arch. Environ. Health. 23:142. Ritchie, J. J. 1967. The study of the factors influencing the oxidation of fat in dry whole milk. Ph. D. Dissertation, Univ. of Minnestoa, U.S.A. 110 Rooney, M. L. 1981. Oxygen scavenging from air in packaging headspaces by singlet oxygen reactions in polymer media. J. Food Sci. 47:291. Rothery, G. B. 1968. Reduction of nitrite content of products dried in direct- fired spray dryers. Austr. J. Dairy Technol. 23: 76. Ryan, T. C., J. 1. Gray and I. D. Morton. 1981. Oxidation of cholesterol in heated fat. J. Sci. Food Agric. 32:305. Sabine, J. R., 1977. Cholesterol. Marcel Dekker Inc., New York, NY. Saito, M. 1979. Food quality preservation by means of free- oxygen absorber. Shokuhin to Kogyo. (Food and Industry) May 2nd, Vol. 65. Sakamaki, C., J. I. Gray, and B. R. Harte. 1988. The influence of selected barriers and oxygen absorbers on the stability of oat cereal during storage. J. Packaging Technol. 2(3):98. Salih, A. M. A., M. Anderson, and B. Tuckley. 1977. The determination of short- and long-chain free fatty acids in milk. J. Dairy Res. 44:601. Sander, B. D., D. E. Smith, and P. B. Addis. 1988. Effects of processing stage and storage conditions on cholesterol oxidation products in butter and cheddar cheese. J. Dairy Sci. 71:3173. Sander, B. D., D. E. Smith, P. B. Addis and S. W. Park. 1989. Effects of prolonged and adverse storage conditions on levels of cholesterol oxidation products in dairy products. J. Food Sci. 54:874. Sander, B. D., P. B. Addis, S. W. Park, and D. E. Smith. 1989. Quantification of cholesterol oxidation products in a variety of foods. J. Food Prot. 52:109. Shipstead, H. and N. P. Tarassuk. 1953. Dehydrated foods: Chemical changes in dehydrated milk during storage. J. Agr. Food Chem. 1:613. Shoptaugh, N. H. 1982. Cholesterol oxides: toxicological and analytical aspects. Food Research Institute, Univ. of WI-Madison; 1982 Fall Meeting. Smith, L. L. and F. L. Hill. 1972. Detection of sterol hydroperoxides on thin-layer chromatoplates by means of the Wurster dyes. J. Chromatogr. 66:101. 111 Smith, L. L., M. J. Kulig, and J. I. Teng. 1973. Sterol metabolism. XXIV. Pyrolysis of some sterol allylic alcohols and hydroperoxides. Steroids, 22:627. Smith, L. L. and M. J. Kulig. 1975. Sterol metabolism. XXXIV. on the derivation Of carcinogenic sterols from cholesterol. Cancer Biochem. Biophys. 1:79. Smith, L. L. 1981. Cholesterol Autoxidation. Plenum Press. New York and London. Smith, L. L. 1987. Cholesterol autoxidation 1981-1986. Chem. Phys. Lipids. 44:87. Sonntag, N. O. V. 1979. Structure and composition of fats and oils. In Bailey's Industrial Oil and Fat Products. 4th Ed. Ed. D. Swern, John Wiley and Son, New York. Chapter 1. p.1. Steen, K. 1977. Seasonal variations in the composition of dry whole milk and its keeping qualities under storing. Maelkeritidende, 90:152. Sugino, K., J. Terao, H. Murakami, and S. Matsushita. 1986. High—performance liquid chromatographic method for the quantification of cholesterol epoxides in spray-dried egg. J. Agr. Food Chem. 34:36. Sweeney, J. P., and J. L. Weihrauch. 1977. Summary of available data for cholesterol in foods and methods for its determination. Crit. Rev. Food Sci. Nutr. 8:131. Tamsma, A., F. E. Kurtz, and M. J. Pallansch. 1967. Effect of oxygen removal technique on flavor stability of low heat foam spray-dried whole milk. J. Dairy Sci. 50:1562. Tamsma, A., F. E. Kurtz, and M. J. Pallansch. 1973. Comparison of flavor and physical properties of foam spray—dried whole milks prepared from concentrates foamed with air and with nitrogen. J. Dairy Sci. 56:161. Tarladgis, B. G., A. M. Pearson, and L. R. Dugan, J. 1964. Chemistry of the 2—thiobarbituric acid test for determination of oxidative rancidity in foods. II. Formation of the TBA—malonaldehyde complex without acid-heat treatment. J. Sci. Food Agric. 15:602. Taylor, C. B., S. -K. Peng, W. T. Werthessen, P. Tham, and K. T. Lee. 1979. Spontaneously occurring angiotoxic derivatives of cholesterol. Am. J. Clin. Nutr. 32:40. 112 Tsai, L. S., K. Ijichi, C. A. Hudson, and J. J. Meehan. 1980. A method for the quantitative estimation of cholesterol oxides in eggs. Lipids 15:124. Tsai, L. S. and C. A. Hudson. 1981. High-performance liquid chromatography of oxygenated cholesterols and related compounds. J. Am. Oil Chem. Soc. 58:931. Tsai, L. S. and C. A. Hudson. 1984. Cholesterol oxides in commercial dry egg products: Isolation and identification. J. Food Sci. 49:1245. Tuohy, J. J. 1987. Milkfat autoxidation in whole milk powder during storage. Ph. D. Dissertation, Univ. of Ireland, Ireland. Van Lier, J. E. and L. L. Smith. 1970. Autoxidation of cholesterol via hydroperoxide intermediates. J. Org. Chem. 35:2627. Van Lier, J. E. and L. L. Smith. 1971. Sterol metabolism. XI. Thermal decomposition of some cholesterol hydroperoxides. Steroids. 15:485. Ward, D. D. 1985. The TBA assay and lipid oxidation: an overview of the relevent literature. Milchwissenschaft, 40:583. Warmbier, H. C. and M. J. Wolf. 1976. A pouch for oxygen- sensitive products. Mod. Package. 49:38. Wheeler, W. H. 1980. Chemical and engineering aspects of low NOx concentration. Chem. Eng. 362:693. Wilson,R. B. 1976. Lipid peroxidation and atherosclerosis. CRC Crit. Rev. Food Sci. Nutr. 7(4):325. Wood, D. L. and J. Bitman. 1986. Cholesteryl esters of cow's milk. J. Dairy Sci. 6922203. Zhang, W. B., P. B. Addis, and T. P. Krick. 1991. Quantification of Sa-cholestane-3B,5,6B-triol and other cholesterol oxidation products in fast food french fried potatoes. J. Food Sci. 56:716. Zimmerman, P. L., L. J. Ernst and W. F. Ossian. 1974. Scavenger pouch protects oxygen-sensitive foods. Food Technol. 28:63. Zubillaga, M. P. and G. Maerker. 1991. Quantification of three cholesterol oxidation products in raw meat and chicken. J. Food Sci. 56:1194. W03" .- g I... 706, J 7L ;:‘_.I’ .1111. ‘I: 6. I. . I‘ 2 ‘ .6. .. “...... ‘(II3 .Ia . ‘ . 3:)“ ~ - -.II'. ... .|“:WII\'.~W ‘ jWI.’ I ...m... . ‘ n-uy‘tII-IiIun. \ r _ m. II I I,. ‘-~a(I|:J‘:.’. y,“ __ V I “lama”... L3. q .0 0.. \ ‘04. I III-VINIIIII. . II» ’- ~' ' . _ mun I “. . .I . r; “I .I v ...-4.0.. . t“"l'nr"\w~ Il|l'.l:‘ I VI In III...«._-.-.I,I..:II,II.~I-.III”:- ': ”pun-u; .