v VLFXLDVIS-I'é—iv-IHEC’: v'w-K...vl . A . é .. ._. . . , an 1 4%.: "3:! z .7; fimflflfifii . ., _ . ‘ E. .. .i . . ‘ ...a .i .n... :34 .6. ‘ , .1; r 3.332.: . V .V : .. i. .. 1 ‘ -1... I . , i. .“ (43. .E Li: Jim . - . a 3... .... 4:95. 2.». (‘2‘ .1v 1 I121... :29. I .. flirt-$3. .23....1 !. 53:91.. :3. 5;... .1303 x141! .i.§.. ) :- x. .w, z-.. 1mg. r v... 53.53. [-1. $.25 .hoflrlvbz .11. z t. f. 13“ .,. -4.,.. .3 llllllllllllllllllllllll‘llllllllllllllllllllllllll 31293 01417 2294 This is to certify that the dissertation entitled EFFECT OF LIPIDS ON FLAVOR DEVELOPMENT IN A MAILL ARD REACTION MODEL SYSTEM presented by SHAUN CHENGHSIUNG CHEN has been accepted towards fulfillment of the requirements for Ph. D. degree in FOOD SCIENCE Major professor Date 7- /‘2- ?5’ MSU is an Affirmative Action /Equal Opportunity Institution 0-12771 LIERARY Michigan State Universlty PLACE II RETURN BOX to ro'movothb checkout from your rooord. TO AVOID FINES Mum on or bdoro onto duo. DATE DUE DATE DUE DATE DUE MSU lo An Afflmntivo ActlorVEquol Opportunlty lnotltwon Wanna-m EFFECT OF LIPIDS ON FLAVOR DEVELOPMENT IN A MAILLARD REACTION MODEL SYSTEM by SHAUN CHENGHSIUNG CHEN A DISSERTATION SUBMITTED TO MICHIGAN STATE UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF FOOD SCIENCE AND HUN/LAN NUTRITION 1995 ABSTRACT EFFECT OF LIPIDS ON FLAVOR DEVELOPMENT IN A MAILLARD REACTION MODEL SYSTEM by SHAUN CHENGHSIUNG CHEN Maillard reactions between amino acids and reducing sugars occur in foods during cooking, and volatile compounds developed in these reactions contribute to the flavor. Sulfur-containing compounds generated via the Maillard reaction are important to meat flavor. However, the formation of these compounds is influenced by the presence of lipids during heating. This study investigated the effect of phospholipids of different degrees of unsaturation on flavor development in a Maillard model system containing cysteine and ribose. The influence of antioxidants on this system was also studied. A headspace analysis method was developed using Tenax traps to collect volatiles which were subsequently desorbed into a solvent by a combination of heat and solvent extraction. More than 90 compounds were isolated from a Maillard model system containing cysteine and ribose heated at 140°C for 1 h. Thirty-one compounds were identified by their Kovats indices and mass spectrometry. Compounds identified included thiophenes, thiophenone, thiopyran, bicyclic thiophenes, thiazole, disulfides, heterocyclic thiols, thiolanone, mercapto carbonyls, methylthio acetates, and non-sulfur-containing compounds such as fiJran derivatives and sugar-derived carbonyls. Fifieen of them were selected as markers to evaluate the effect of lipids on the Maillard reaction. Cysteine, ribose and phospholipids were heated at 140°C for 1 h to produce Maillard reaction volatiles. Sensory evaluation of the headspace of the heated reaction mixtures revealed that the aromas developed in the systems containing only cysteine and ribose were more meaty and sulfury than those detected in systems containing phospholipids. More meaty and sulfury aromas were generated in the reaction mixtures containing phosphatidylcholine-distearoyl than in the systems containing phosphatidylcholine-dioleoyl. However, no discernible differences in aroma were observed between systems containing phosphatidylethanolanune and phosphatidylcholine. Phospholipids suppressed the formation of selected Maillard reaction products. In addition, the degree of unsaturation of the phospholipids influenced the extent to which Maillard reaction volatiles were generated. However, the effects of phospholipids on the formation of volatile compounds was not related to the amino moieties of the phospholipids. Suppression of the formation of Maillard volatiles was also observed in model systems containing aldehydes. Phospholipids isolated from nitrite-cured and uncured hams were heated in a model system with cysteine and ribose. The formation of Maillard reaction volatiles was affected less in the presence of phospholipids from cured hams. When antioxidants were added to the model systems of cysteine, ribose and phosphatidylcholine, the impact of the phospholipids on the Maillard reaction was reduced. A sensory evaluation of volatiles produced from the cysteine-ribose-phosphatidylcholine system indicated flavor differences as a result of adding antioxidants to the systems. These observations were confirmed by gas chromatographic - mass spectrometric analyses. Carbonyl compounds were predominant in the flavor profiles of both cured and uncured hams. Of the carbonyl compounds identified, pentanal and hexanal were more abundant in the uncured hams. The addition of nitrite to hams reduced the formation of these carbonyl compounds. Two nitriles and two thiazoles were also detected in the hams cured with nitrite. © Copyright by SHAUN CHENGHSIUNG CHEN 1995 Dedicated to My Parents, My Sister and My Lovely Wife. iv ACKNOWLEDGMENTS I wish to express my sincere thanks and deepest gratitude to my major professor, Dr. J. Ian Gray for his continued guidance, assistance and encouragement in the course of my graduate study, and for his genuine concern and valuable friendship during these years. His enthusiasm for scientific research and motivation for independent thinking will always be an inspiration to me. I also appreciate very much his patience in helping me develop my dissertation. Appreciation is also extended to Drs. A. M. Booren, D. A. Gage, B. R. Harte and M. J. Zabik for serving on my guidance committee and their critical review of this manuscript. Special thanks go to Dr. Gage for his technical advice on the gas chromatographic and mass spectrometric analyses, and to Dr. Booren for his assistance in the processing of the hams. A special acknowledgment is due to Dr. D. S. Mottram of University of Reading, United Kingdom for his generosity in inviting me to work in his laboratory and for providing a mass spectral library software. Sincere thanks and appreciation are given to Rhonda Crackel, Stephanie Smith and Ross Santell for their participation in my sensory evaluations. I wish to thank my laboratory colleagues, Shu-Mei Lai, Zsuzsana Balogh, Vicki Lin, Coco Saba, Virginia Vega, Pervaiz Akhtar, Eric Cole, and Enayat Gomma, and Drs. Bill Helfen'ch and Gale Strasberg for their help and friendship during this research. Last but not least, I wish to express my deepest appreciation to my parents and sister for their long term support, and to my lovely wife, Angela. TABLE OF CONTENT Page LIST OF TABLES ............................................................................. xii LIST OF FIGURES .......................................................................... xiv INTRODUCTION .............................................................................. 1 REVIEW OF LITERATURE ............................................................ 5 Chemistry of Meat Flavor .................................................................. 5 General aspects of meat flavor ....................................................... 5 Precursors of meat flavor ............................................................... 6 Reactions leading to meat flavor production ................................... 8 A. Sugar degradation .................................................................... 9 B. Maillard reaction ..................................................................... 11 C. Strecker degradation ................................................................ 11 D. Lipid oxidation ........................................................................ 12 Flavor compounds in meats ............................................................ 12 Importance of the Maillard reaction to meat flavor ............................. 16 Chemistry of the Maillard reaction ................................................. 17 1. Initial stage of Maillard reaction ............................................... 17 H. Intermediate stage of the Maillard reaction ............................... 20 Strecker degradation ................................................................. 21 111. Final stage of the Maillard reaction .......................................... 22 Factors influencing the Maillard reaction in foods ............................ 24 A. Duration and temperature of heating .......................................... 24 B. Water content ............................................................................ 24 C. pH ............................................................................................. 25 D. Sugar ......................................................................................... 26 Maillard reaction products and meat aroma ...................................... 27 Classification of flavor compounds produced from the Maillard reaction ......................................................................................... 28 Oxygen-containing flavor compound from the Maillard reaction 29 Nitrogen-containing compounds .................................................... 29 Sulfur-containing compounds ........................................................ 31 Role of Lipids in Meat Flavor ............................................................. 33 Significance of lipid oxidation in flavor formation .......................... 33 Mechanism of lipid oxidation .......................................................... 36 Lipid oxidation during cooking ....................................................... 39 Interaction of Lipids in the Maillard Reaction ..................................... 41 Studies of the involvement of lipid oxidation in the Maillard reaction ......................................................................................... 42 A. Interaction of amino acid degradation products with carbonyl compounds ............................................................................... 42 B. Interaction between amino acids and carbonyl compounds ....... 43 C. Interaction of Maillard reaction products and niacylglycerols or phospholipids in model systems ........................................... 44 Possible mechanisms for the interaction of lipid oxidation products with Maillard reaction products ..................................................... 48 Cured Meat Flavor .............................................................................. 50 Role of nitrite in cured meats .......................................................... 50 Antioxidant role of nitrite in cured meats ........................................ 51 Formation of nitrite-derived antioxidants ..................................... 52 Stabilization of the heme pigments ............................................... 53 Chelation of metal ions ................................................................ 53 Stabilization of lipids ................................................................... 54 Chemistry of cured meat flavor ....................................................... 54 REFERENCES .................................................................................... 57 CHAPTER 1 ANALYSIS OF VOLATILE COMPOUNDS PRODUCED FROM THE MAILLARD REACTION BETWEEN CYSTEINE AND RIBOSE ABSTRACT ........................................................................................ 74 INTRODUCTION ............................................................................... 75 EXPERIMENTAL .............................................................................. 77 Materials ............................................................................................. 77 vii Preparation of Maillard reaction mixtures and collection of volatile compounds ........................................................................................ 77 Desorption of trapped volatile compounds from the Tenax trap ........... 78 Gas chromatographic - mass spectrometric analysis ............................ 82 RESULTS AND DISCUSSION .......................................................... 83 Development of analytical procedures for the identification of volatile compounds in a heated Maillard reaction system ............................... 83 Isolation and identification of Maillard reaction products .................... 84 Selection of marker compounds for use in subsequent studies ............. 93 CONCLUSION ................................................................................... 93 REFERENCES .................................................................................... 95 CHAPTER 2 EFFECTS OF PHOSPHOLIPIDS AND ALDEHYDES ON THE FORMATION OF MAILLARD REACTION VOLATILES IN MODEL SYSTEMS CONTAINING CYSTEINE AND RIBOSE ABSTRACT ........................................................................................ 98 INTRODUCTION ............................................................................... 99 EXPERIMENTAL .............................................................................. 101 Materials ............................................................................................. 101 Preparation of Maillard reaction volatiles ............................................ 102 Isolation and identification of Maillard reaction volatiles .................... 102 Desorption of the trapped volatile compounds from the Tenax trap ..... 103 Gas chromatography - mass spectrometry ........................................... 104 Evaluation of odor characteristics ....................................................... 105 Determination of the effects of phospholipids and aldehydes on the Maillard reaction ............................................................................... 105 viii RESULTS AND DISCUSSION .......................................................... 106 Sensory evaluation of compounds developed from cysteine and ribose in the presence of phospholipids ....................................................... 106 Effect of phospholipids on the formation of Maillard reaction volatiles fi'om heated cysteine and ribose ........................................................ 108 The effects of various phospholipids on the Maillard reaction involving cysteine and ribose ............................................................ 115 Effects of aldehydes on the production of Maillard reaction volatiles in a heated cysteine and ribose model ............................................... 118 CONCLUSION ................................................................................... 124 REFERENCES .................................................................................... 127 CHAPTER 3 THE EFFECTS OF STABILIZED LIPIDS ON THE FORMATION OF FLAVOR VOLATILES IN A HEATED MODEL SYSTEM CONTAINING CYSTEINE AND RIBOSE ABSTRACT ........................................................................................ 130 INTRODUCTION ............................................................................... 13 1 EXPERIMENTAL .............................................................................. 133 Materials ............................................................................................. 133 Preparation of cured hams ............................................................ 134 Quantitation of phospholipids in hams ................................................ 136 Nitrosation of morpholine by phospholipids extracted fiom ham ......... 137 Preparation and isolation of Maillard reaction volatiles ....................... 137 Statistical analysis ............................................................................... 140 RESULTS AND DISCUSSION .......................................................... 141 Quantitation of phospholipids in ham samples .................................... 141 Nitrosation of morpholine ................................................................... 141 ix Volatiles generated form the Maillard reaction in the presence of phospholipids .................................................................................. Effect of antioxidants on the interaction between the Maillard reaction and lipids in a heated model system .................................................. CONCLUSION ................................................................................... REFERENCES .................................................................................... CHAPTER 4 FLAVOR PROFILES OF CURED AND UN CURED HAMS ABSTRACT ........................................................................................ INTRODUCTION ............................................................................... EXPERIMENTAL .............................................................................. Materials ............................................................................................. Preparation of cooked ham samples .................................................... Collection of flavor compounds .......................................................... Gas chromatography - mass spectrometric analysis ............................. RESULTS AND DISCUSSION .......................................................... Volatile compounds isolated from nitrite-free hams ............................ Flavor profile of cooked cured hams ................................................... CONCLUSION ................................................................................... REFERENCES .................................................................................... SUMMARY AND CONCLUSIONS .................................................. FUTURE RESEARCH ....................................................................... 145 152 158 158 163 164 166 166 166 166 167 I70 170 173 175 176 179 182 APPENDIX A. List of authentic flavor compounds used as references ................. 185 LIST OF TABLES TABLES REVIEW OF LITERATURE 1. Volatile compounds with meaty aroma ........................................ 2. Volatile compounds of different chemical classes reported in cooked meats ................................................................................ CHAPTER ONE 1. Maillard reaction products isolated from a model system of cysteine and ribose heated at 140°C for 1 h ................................. 2. Sensorial properties of Maillard reaction volatiles selected as marker compounds to study the effect of phospholipids on volatile production ....................................................................... CHAPTER TWO 1. Summary of odor descriptions attributed to the headspace of cysteine and ribose heated in the presence of phospholipids ........ 2. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phosphatidylethanolamines .......................................................... 3. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phosphatidylcholines ................................................................... xii Page 13 15 85 94 107 109 112 4. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of hexanal ........... 120 5. Relative amounts of Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of dodecanal ..................... 122 CHAPTER THREE 1. Composition of the curing brines used to process hams ................ 135 2. Lipid contents of hams cured with varying levels of nitrite ..... 142 3. Formation of N-nitrosomorpholine on heating morpholine with phospholipids extracted from hams cured with different levels of nitrite ........................................................................................... 144 4. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phospholipids extracted from hams .................................................................... 146 5. Relative amounts of lipid oxidation products and heterocyclic compounds derived from the reaction of cysteine and ribose in the presence of phospholipids extracted from hams cured with different levels of nitrites ............................................................. 149 6. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence phosphatidylcholine and antioxidants .......................................... 153 7. Summary of odor descriptions attributed to the headspace of heated cysteine and ribose in the presence of phosphatidylcholine and antioxidants .................................... 154 CHAPTER FOUR 1. Volatile compounds isolated from uncured and nitrite-cured ham by purge-and-trap method ............................................................ 172 xiii LIST OF FIGURES FIGURES REVIEW OF LITERATURE 1. Formation of furan and thiophene derivatives from the reaction of 5-methyl-4-hydroxy-3(2H)-firranone with hydrogen sulfide ....... 2. Simplified scheme of the Maillard reaction ................................... 3. Initial steps of the Maillard reaction ............................................. 4. Strecker degradation of cysteine ................................................... 5. Pathway for the formation of pyrazine .......................................... 6. Formation of thiazole and thiazoline form the intermediates of the Maillard reaction ........................................................................ 7. Mechanism of lipid oxidation and formation of carbonyl compounds ................................................................................. 00 . Hydroperoxides formed by the oxidation of oleic acid .................. 9. General reaction pathway for the homoytic cleavage of hydroperoxides of unsaturated lipids .......................................... 10. Mechanism for the formation of nitrite with polyunsaturated lipids .......................................................................................... CHAPTER ONE 1. Apparatus used for the collection of the volatile compounds generated by heating cysteine and ribose at 140°C for l h .......... xiv Page 10 18 19 23 3O 32 35 38 4O 55 79 2. Desorption of the volatile Maillard reaction compounds from the Tenax trap .................................................................................. 81 3. Formation of thiazoles and thiazolines from intermediates of the Maillard reaction ........................................................................ 89 4. Structures and sensorial properties of disulfide compounds isolated from a model system containing cysteine and ribose ..... 91 CHAPTER TWO 1. Relative amounts of selected Maillard reaction volatiles isolated fiom the reaction mixture of cysteine and ribose in the presence of phospholipids compared to cysteine and ribose alone ............. 116 2. Relative amounts of selected Maillard reaction volatiles isolated from the reaction mixture of cysteine and ribose in the presence of 0.1 mmole aldehydes compared to cysteine and ribose alone .. 125 CHAPTER THREE 1. Suggested routes for the formation of 2-pentylthiapyran, 2- hexylthiophene and 2-pentylpyridine from 2,4-decadienal .......... 151 2. Relative amounts of selected Maillard reaction volatiles isolated from model systems containing cysteine and ribose in the presence of phospholipids with or without antioxidants .............. 156 CHAPTER FOUR l. The scheme of the Short Path Thermal Desorption Unit and the procedures for the desorption of trapped volatile compounds into a GC/MS ............................................................................. 168 XV 2. A schematic diagram of the procedure used to desorb the trapped volatile compounds from the Tenax trap. The trapped flavor compounds were thermally desorbed from the Tenax trap and swept into the injection port of the GC via a needle .................... 169 xvi INTRODUCTION Flavor is one of the important sensory attributes that contribute to the overall acceptability of meat products. Raw meat itself has only a blood-like odor. However, it is a rich reservoir of non-volatile compounds with taste tactile properties, flavor enhancers, and aroma precursors. Flavor is perceived as the simultaneous stimulation of taste and odor senses caused by the volatile chemicals present in cooked meats (Shahidi, 1989). Cooked meat flavor is principally derived from non-volatile water-soluble precursors (Homstein and Crowe, 1960; Macey gt Q, 1964a, b; Hornstein and Wasserman, 1987; Shahidi, 1989, Mottrarn, 1994c), and fat-soluble constituents (Sanders _e_t_ g, 1966; Sink, 1973; Shahidi, 1989). The heating process develops complex components which evoke characteristic sensory responses (W asserrnan, 1979). On cooking, free amino acids and reducing sugars (e.g., glucose or fiuctose produced by glycogen breakdown) together with other low molecular weight water-soluble materials, react with one another to produce flavor compounds. The reaction between amino acids and reducing sugars is defined as the Maillard reaction. Products from one reaction may often become the precursors for others. Thus, the Maillard reaction and its subsequent reactions lead to the formation of a large number of volatile components, which are contributory to cooked meat flavor development (Shahidi e_t al_., 1986). The oxygen-, nitrogen- and sulfur- containing heterocyclic volatiles contribute significantly to the overall desirable flavor of meat (Bailey, 1983; Shahidi, 1989; Werkhoff e_t g, 1993 ). In particular, sulfur compounds play a major role in meat flavor due to their distinctive olfactory 1 properties and generally low taste thresholds (Gi‘mtert 91 g, 1990). Although many volatiles are derived via lipid oxidation, it is believed that compounds produced fiom the water-soluble precursors are the major contributors to meat flavor (Nursten, 1986; Shahidi, 1989; Mottram, 1991). Lipid is a major component in muscle foods and provides characteristic flavors after cooking (Homstein and Crowe, 1960). Lipids are important in flavor development because of the oxidation products generated during cooking (Mottram and Edwards, 1983; Shahidi, 1989). Lipids also serve as a depot for fat- soluble compounds which are volatilized during heating and strongly affect flavor (Shahidi, 1989). Additionally, lipid oxidation products, such as carbonyl compounds, contribute to species-specific meat flavors (Mottram e_t g, 1982; Mottram and Edwards, 1983; Shahidi, 1989). For example, the "goaty" or "chickeny" flavors are strongly related to carbonyls, both qualitatively and quantitatively (Shahidi, 1989). Lipids also are involved in the Maillard reaction. Suppression of Maillard reaction products was observed when polyunsaturated lipids, e.g., phospholipids were heated in a model system containing an amino acid and a reducing sugar. These results indicated that lipid oxidation derivatives may participate in the Maillard reaction (Mottram and Whitfield, 1987b; Whitfield e_t g, 1988; Farmer e_t g, 1989; Farmer and Mottram, 1990a, 1994). It was concluded that Maillard reaction intermediates (e.g., hydrogen sulfide or ammonia) react with carbonyl compounds derived from lipid oxidation, thus reducing the amounts of Maillard reaction products formed (Salter e_t fl, 1988; Farmer and Mottram, 1990a). It is generally believed that the curing process alters the flavor of flesh muscle by changing either the nature of the flavor volatiles, or the quantities of the individual volatiles formed. This modification is achieved through either the addition of extraneous compounds such as salt or sugar, or by the reduction of flavor-producing components (Swain, 1972). The so-called cured meat flavor has actually been defined as the basic meat flavor which describes the flavor derived from precursors other than triacylglycerols (Cross and Ziegler, 1965). The differences in the volatile compounds contributing to cured and uncured meat flavor are quantitative rather than qualitative. For example, the concentrations of aldehydes are smaller in cured meats (Cross and Ziegler, 1965; Ramarathnam gt g, 1991a, b), an observation which has been ascribed to lipid-nitrite interactions (Swain, 1972; Mottram, 1985a; Freybler, 1989). Oxidation of the unsaturated lipids results in the formation of carbonyl compounds which have been implicated as significant contributors to the flavor of uncured meat products, but not to cured meat flavor (Ramarathnam e_t e_rl_., 1991a). The role of nitrite in cured meat flavor development has not been clearly defined. Nitrite functions as an antioxidant, thus preventing the formation of carbonyl compounds through lipid oxidation. Based on the studies of Farmer and Mottram (1990a), it is possible that nitrite stabilization of lipids in cured meats (Freybler _e_t g, 1993) may play a significant role in modifying the flavor compounds generated by the Maillard reaction in heat-processed meat products. This study is based on the hypothesis that the difference in the flavors of cured and uncured meat is due, in part, to the differences in the concentrations of carbonyl compounds and Maillard reaction heterocyclic compounds. Carbonyl compounds, generated through thermal processing (cooking), may modify the Maillard reaction and subsequently suppress the formation of heterocyclic compounds. Stabilization of meat lipids through the addition of nitrite should theoretically produce fewer carbonyl compounds during the cooking of meat. Therefore, the Maillard reaction between amino acids and reducing sugars should proceed without competition from the carbonyl compounds and generate the optimum quantities of Maillard reaction products. Specific objectives of this study are: 1. To develop a dynamic headspace analytical technique for the collection and identification of Maillard reaction products in a model system containing ribose and cysteine. 2. To identify some Maillard reaction volatile compounds as marker compounds in the studies of the impact of lipids on the formation of Maillard reaction products. 3. To determine the effect of the presence of lipids of varying degrees of unsaturation on Maillard reaction volatiles produced from heating a mixture of cysteine and ribose in a model system. 4. To ascertain the role of nitrite in cured meat flavor by characterizing the flavor profiles obtained on heating cysteine and ribose with phospholipids isolated from cured and uncured hams. 5. To investigate the effects of antioxidants on the formation of Maillard reaction products in a model system containing cysteine, ribose and phospholipids. 6. To compare the total flavor profiles of cured and uncured meat products. LITERATURE REVIEW Chemistry of Meat Flavor General aspects of meat flavor Flavor is defined as the sensation caused by those properties of any substance taken into the mouth which stimulate the senses of taste and smell, and also the general pain, tactile and temperature receptors in the mouth (Farmer, 1992). Odor and taste are the main contributors to flavor (Shahidi _e_t Q, 1986). Flavor is an essential sensory aspect of the overall acceptability of meat products. It is perceived as the simultaneous stimulation of taste and odor senses by the high molecular weight components and volatile compounds present in cooked meats. The flavor volatiles have a significant effect on the sensory acceptability of food even before it is consumed (Shahidi, 1989). Meat has always constituted an important part of our diet. It is an excellent source of protein and provides certain essential vitamins and minerals. Its unique sensory properties make it the focus of culinary art in many cultures. Meat may account for nearly one-third of the total food expenditure in western societies, and therefore flavor becomes a very important component of its eating quality. Many studies have focused on understanding the chemistry of meat flavor. The very desirable characteristics of meat flavor have also been sought in the production of simulated meat flavorings, and these are of significant commercial importance in convenience and processed foods as well as in meat substitutes (Mottram, 1991). Although a number of factors are known to influence the flavor of meat, no single group has been assigned a major role. Genetic as well as environmental factors are probably the most important (Shahidi e_t fl, 1986). Of the former, the "species flavor" has the most pronounced effect. Diet is the most important 5 environmental factor, the type of diet and nutrient source exhibiting significant influence on the flavor of muscle foods (Shahidi e_t al., 1986; Monahan, 1992; Monahanet a, 1993). The amount and type of protein in the diet have little effect, but lipids do make a significant contribution (Sink, 1979). The method of cooking also afi‘ects the formation of meat flavor. Boiling, roasting, frying, and pressure- cooldng, contribute significantly to the volatile compounds that are formed in the heating process and hence relate to differences in overall meat flavor (Shahidi e_t g, 1986). Precursors of meat flavor Meat flavor is principally derived from non-volatile precursors which are both water-soluble (Homstein and Crowe, 1960; Macey e_t _a_1., 1964a, b; Homstein and Wasserman, 1987) and fat-soluble (Sanders et _a_1., 1966; Sink, 1973; Lawrie, 1974; Shahidi, 1989). The development of meat flavor begins with raw meat which exhibits a simple blood-like odor (Shahidi et a_l., 1986). During cooking, a series of reactions occur between non-volatile components of the lean and adipose tissues that evoke characteristic sensory responses (W asserrnan, 1979; Mottram, 1991). The flavor of cooked meats is due to a combination of thermal degradation products of sugars, amino acids and nucleotides, and Maillard reaction and lipid oxidation products (Shahidi et aL, 1986; Shahidi, 1989; Mottram, 1991). Meat tissue consists primarily of water, proteins, fats and carbohydrates, plus lesser quantities of non-protein nitrogen-containing compounds, minerals, trace amounts of vitamins and organic substances. On heating, these compounds react with one another to produce complex volatile mixtures that contribute to meat flavor (Homstein and Wasserman, 1987). Though free amino acids constitute only about 0.1% of the fresh weight of beef muscle, these components appear to be responsible for the flavor substances formed by the Maillard reaction on heating (Shahidi e; g, 1986). Therefore, meat flavor precursors are low molecular weight water-soluble compounds. Higher molecular weight muscle fibrils and sarcoplasmic proteins are thought to be unimportant (Mottram, 1991). The amino acid and carbohydrate compositions of lean meat from different animals (beef, pork or lamb) are similar; and thus, similar aromas are produced when lean tissue extracts of the various species are heated (Homstein and Crowe, 1960, 1963; Homstein and Wasserman 1987; Shahidi, 1989; Mottram, 1991). Homstein and Wasserman (1987) reported that no meaty aroma was produced on heating individually amino acid and reducing sugar fractions obtained from ion-exchange chromatographic separation of meat dialysis diffusates. When the two fractions were heated together, a meaty aroma was generated, thus demonstrating the importance of the Maillard reaction. The main flavor precursors are regarded as free sugars, nucleotide-bound sugars, free amino acids, peptides, nucleotides, glycopeptides, creatine and creatinine (Mottram, 1991). The role of lipids in meat, both adipose tissue and fat contained within the lean portion, is to provide the species-characteristic aroma. Homstein and Crowe (1963) and Homstein and Wasserman (1987) found that, while aqueous extracts of beef, pork and lamb had similar aromas after heating; however, heating the fats yielded species-specific aromas. These investigators concluded that lipids provide volatile compounds with the characteristic flavors of different species, and that the lean gave a basic meaty flavor common to all species. The interaction between compounds derived from both the adipose and lean tissues of meat may also contribute to flavor. The direct reaction of free amino groups of proteins and carbonyl compounds derived from lipid oxidation could produce odorous components (Wasserman and Spinelli, 1972). Pippen and Mecchi (1969) suggested that hydrogen sulfide derived from sulfur amino acids reacts with carbonyls produced from the thermal degradation of lipids to form volatile compounds which contribute to chicken aroma. The involvement of phospholipids, the essential structural components of all cells, has been well documented in the development of warmed-over-flavor (Pearson et al, 1977). However, phospholipids also play an important role in the formation of desirable flavor. Mottram and Edwards (1983) demonstrated that lipids were essential for the development of cooked meat aroma, and that the phospholipids alone could provide sufficient lipid for meat aroma production. Removal of triacylglycerols from meat had only a minor effect on the formation of oxidative volatile compounds during cooking. However, the additional removal of phospholipids changed the aroma from a meaty to a roast or biscuit-like note. This change caused a significant increase in certain heterocyclic volatile compounds such as alkylpyrazines, alkylthiazoles and thiophenes (Whitfield e_t fl, 1988). As the primary source of the heterocyclic compounds in food is the Maillard reaction, it appears that the involvement of phospholipids or their decomposition products in the Maillard reaction may play an important role in the development of the characteristic aroma of cooked meat (Mottram, 1991). Reactions leading to meat flavor production Many components in meat undergo a series of physical and chemical changes when heated, that are governed by temperature, duration of heating, and water content (W asserman, 1979). Heat serves many functions including releasing flavor precursors from fat, allowing intimate mixing of fat and water-soluble components, and accelerating non-enzymatic browning reactions (Herz and Chang, 1970). The primary reactions during cooking that lead to the generation of cooked meat flavor include pyrolysis of amino acids and peptides, carbohydrate degradation, interaction of reducing sugars with amino acids and peptides, degradation of thiamine, and thermal degradation of lipids (MacLeod and Seyyedian-Ardebili, 1981). More complicated reactions arise by secondary reactions which occur between the products of the initial reactions, and which give rise to a vast number of volatile compounds contributing to meat flavor (Mottram, 1991). Some of the reactions that produce flavor compounds are discussed briefly below: A. Sugar degradation. When sugars are heated, caramelized flavors are formed; but, in general, relatively high temperatures are required. At temperature above 150°C, a sugar molecule may lose a unit of water and form an anhydride. The anhydride from pentose can be further dehydrated to a furfural (Feather and Harris, 1973). Subsequent heating results in the formation of many odorous compounds, including furan derivatives, carbonyl compounds, and both aliphatic and aromatic hydrocarbons (Feather and Harris, 1973). Though the concentrations of sugars in meat are low, cararnelization reactions can possibly occur on the surface of roasted and grilled meat when dehydration is accomplished (Mottram, 1991). A major group of sugar degradation products, the furanones, are produced by aqueous degradation. For example, 2,5-dimethyl-4-hydroxyl-3 (2H)-furanone results from the base-catalyzed degradation of fi'uctose in a boiling aqueous solution (van den Ouweland e_t fl, 1975). Hydroxymethylfuranones further react with hydrogen sulfide to produce volatiles with meaty aromas (Figure l). lO H,opo,H2 H2O HZS H <———> H20 0 H S 4-hydroxy-5-methyl- 4-hydroxy—5-methyl- 3(2H)-furanone 3(2H)—thiophenone HZS H25 H H H x j x H H SH SH 016 (1+0? X x X X X=OorS Figure 1 Formation of furan and thiOphene derivatives from the reaction of 5- methyl-4-hydroxy-3 (2H)-furanone with hydrogen sulfide (van den Ouweland et al., 1975). 11 B. Maillard reaction The Maillard reaction between reducing sugars and amino acids is one of the most important routes to flavor compounds in cooked foods. The Maillard reaction does not require the high temperatures normally associated with sugar caramelization and protein pyrolysis (Hurrel, 1982). However, the reaction rate increases markedly with temperature, and the formation of flavor compounds generally occurs at elevated temperatures. The Maillard reaction occurs in aqueous systems, but it proceeds more readily in dried and concentrated foods with lower moisture contents (Hurrel, 1982). Therefore, flavor volatiles produced in meat by the Maillard reaction tend to be associated with those parts of the meat that have been dehydrated by the heat source (van den Ouweland gt g, 1978). The Maillard reaction and its involvement in meat flavor formation will be discussed in more detail in a later section. C. Strecker degradation. The Strecker degradation is one of the most important elements associated with the Maillard reaction. It involves the oxidative dearnination and decarboxylation of an a-amino acid in the presence of a dicarbonyl compound. The degradation results in the formation of an aldehyde containing one carbon atom less than the original amino acid and an arninoketone. Aminoketones can rearrange to produce several heterocyclic compounds such as pyrazines, oxazoles and thiazoles (MacLeod and Seyyedian-Ardebili, 1981; Vemin and Parkanyi, 1982; Baines and Mlotkiewicz, 1984). Further details will be provided in a later section. 12 D. Lipid oxidation. An important route of meat flavor generation during cooking is the thermally-induced oxidation/degradation of the unsaturated acyl chains in lipids. However, oxidation of these unsaturated fatty acids is also responsible for the development of some undesirable off flavors usually associated with rancidity (Mottram, 1991). Lipid oxidation during cooking plays an important role in the development of cooked meat aroma, and its oxidative products have been found predominantly in lightly grilled or boiled meat (Mottram et al., 1982; Mottram and Edwards, 1983). The role of lipid oxidation in meat flavor formation will be discussed in more detail later. flavor compounds in meats In general, the flavor of cooked meat is due to a mixture of compounds including non-volatile or water-soluble compounds with taste-tactile characteristics, potentiators or synergists that enhance the flavor contributions of other agents, and volatiles which give rise to the odor properties (Shahidi et al., 1986). Nitrogen, sulfur and oxygen heterocyclic compounds, along with noncyclic sulfur compounds and hydrocarbons, are the predominant "meaty" flavor volatiles (Table 1). Heterocyclic aroma compounds in meat primarily arise fi'om interactions between mono- and dicarbonyl compounds with hydrogen sulfide and ammonia. These carbonyl compounds are derived from the Maillard reaction, lipid oxidation, and aldolization reactions. The low-molecular weight intermediates, such as hydrogen sulfide and ammonia, are produced by thermal degradation of sulfur-containing amino acids, and pyrolysis of amino acids, respectively (Bailey and Einig, 1989). 13 Table 1. Volatile compounds with "meaty" aromaa Compound Number Example Flavor note Sulfides 7 Mercaptan meaty (1-5 ppb) Thiols (acyclic) Methylthioethane onion (hydro) Furan with sulfur 17 Furfuryl thiol; meaty (0.5 - 1 ppb) containing side chain 5-methyl sulfurous (> 1 ppb) Thiophenes l 1 Thiophene—Z-methyl roast meat -3-thiol di-, tri-Thiolanes 5 1,2,4-Trithiolane roast meat Trithianes 3 l ,3,5-Trithiane meaty 2,4,6-Trimethyl trithiane Thiazol(in)es l4 Thiazole meaty, nutty, pyridine-like Thialdines 4 Thialdines meaty, roast beef Pyrazine-firran sulfide 3 Furfurythio-Z- cooked meat (3-methyl) pyrazine (< 1 ppb) Furans 2 F uran meaty, pleasant, 2-Methylfuran slightly sulfurous sickly Oxazol(in)es 2 Oxazole boiled beef, nutty, 2,4,5-Trimethyl oxazole sweet, green Ketones 4 Cyclopentanone roast beef 3-Methyl Hydrocarbons l n-Octane meaty Miscellaneous 5 2-Ethyl thiophenol a Shahidi, 1989. l4 Qualitative information on the presence of volatile compounds in different species is available in the literature (Table 2). For example, the proportion of sulfur compounds in beef (20% of the total number of compounds reported) is much higher than that in other meats. Few furans have been found in lamb, yet lamb contains more carboxylic acids than other species. The volatiles in chicken flavor contain a higher proportion of lipid oxidation derivatives (hydrocarbons, alcohols, carbonyl compounds), with aldehydes and ketones particularly abundant. Additionally, a large numbers of alcohols and phenols in cured pork are mainly due to phenols derived from the smoking of the meat (Shahidi e_t al., 1986; Mottram, 1991). Over 100 hydrocarbons, mainly branched alkanes, alkenes and alicyclics have been reported in beef volatiles (MacLeod and Ames, 1986, 1987; Umano and Shibamoto, 1987; Vercellotti e_t al., 1987). Two classes of compounds that are believed to make important contributions to characteristic beef aroma have been recently reported by MacLeod and Amos (1987). One of these is a series of methyl-substituted cyclopentanones, cyclopentenones and cyclohexenones. The other group is the sulfur-substituted furans such as 2-methyl-3 -(methylthio)furan, 2-methyl-3-furanthiol and the corresponding disulfides. Volatiles from pork, including fried bacon and other cured and uncured pork products, are mainly aliphatic compounds, with aldehydes, acids and esters being the most abundant (Mottram e_t _a_1., 1982, 1984; Mottram, 1984, 1985b; Ho e1 g, 1983). Some heterocyclic compounds have been reported in grilled pork and fried bacon, including pyrazines, pyridines, thiophenes, thiazoles, and oxazoles. Many of them have also been reported in beef flavor (Mottram, 1991). Phenols 15 Table 2. Volatile compounds of different chemical classes reported in cooked meatsa Compounds beef pork lamb chicken cured uncured mutton Hydrocarbons 193 39 37 43 84 Alcohols and phenols 82 64 25 20 53 Aldehydes 65 38 41 39 83 Ketones 76 32 3 1 20 53 Carboxylic acids 24 29 30 51 22 Esters 59 21 33 11 16 Lactones 38 8 12 14 24 Furans and pyrans 47 16 28 5 16 Pyrroles and pyridines 39 12 16 19 24 Pyrazines 51 22 44 16 22 Other nitrogen compounds 28 22 9 2 7 Oxazoles and oxazolines 13 3 1 4 5 Non-heterocyclic sulfur 72 20 17 7 17 compounds Thiophenes 35 4 15 2 7 Thiazoles and thiazolines 29 6 17 13 18 Other heterocyclic sulfur 13 4 1 4 6 compounds Miscellaneous compounds 16 7 4 1 11 Total 880 347 361 271 468 3 Mottram, 1991. l6 and guaiacols have been found in bacon, and are believed to originate from the smoke treatment. A series of organic nitrates and nitriles have been reported in boiled bacon, and arosed from the interaction of sodium nitrite with lipids (Mottram e_t g, 1984; Mottram, 1984). Many of the compounds reported in chicken flavor have also been found in other meats, including hydrocarbons, acids, lactones, saturated aliphatic alcohols and ketones, pyrazines and oxazoles (Tang et _a_1., 1983; Hartrnen e_t a_1_., 1984). The most noticeable feature of chicken flavor is the large number of unsaturated aliphatic aldehydes, alcohols and ketones found in both roast chicken meat and fat (Mottram, 1991). These meats contain significant numbers of thiazoles. Thiazoles have low odor threshold values, green and nutty aromas. These compounds may be important in both meaty aroma and the differences in aroma between species (Mottram, 1991). The Importance of the Maillard Reaction to Meat Flavor The Maillard reaction is one of the primary routes for the development of flavor associated with cooked meat and other heat-processed foods such as bread, cereal products, and roasted coffee (Apriyantono and Ames, 1990; Knoch and Baltes, 1992; Tressl e_t fl» 1993; Mottram and Whitfield, 1994). The reaction is named after the French chemist Louis Maillard, who first described the formation of browning pigments or melanoidins upon heating a solution of glucose and glycine (Maillard, 1912). The Maillard reaction in food products is important for a number of reasons: it is involved in the production of color and flavor compounds, it reduces the nutritive value of foods by the involvement of essential amino acids such as lysine, it is also involved in the formation of heterocyclic aromatic amines, some of which are mutagenic and carcinogenic, and it may impart antioxidant 17 properties to heat processed foods (Nursten, 1986; O’Brien and Morrissey, 1989; Skog, 1993, Vemin et al., 1993). The carbonyl-amino reaction occurs widely during the heating or prolonged storage of foodstuffs, and is a major source of browning and flavor production (Hurrell, 1982). In food systems, the reactions normally occur between reducing sugars and amino acids or proteins, although aldehydes formed from lipid oxidation may react in a similar way with amino acids or proteins (Kwon gt A, 1965; Montgomery and Day, 1965). Chemism of the Maillard reaction The Maillard reaction can be divided into three stages. The initial stage involves a sugar-amine condensation reaction to form a glycosylamine which then undergoes rearrangement. The intermediate stage involves the dehydration of sugar, either by loss of three molecules of water to form furfural or by loss of two water to produce reductones, fission, mainly by dealdolization, and the Strecker degradation. The final stage consists of the conversion of carbonyl compounds into high molecular weight products such as melanoidins (Hodge, 1953). A simplified scheme of the Maillard reaction is shown in Figure 2 (Hodge, 1967). 1. Initial stage of the Maillard reaction The first step involves the addition of the amine to the carbonyl group of a reducing sugar, followed by elimination of water to form a Schiff base. The Schiff base cyclizes to an N-substituted glycosylamine which undergoes an acid- catalyzed Amadori rearrangement to form the N-substituted-l-amino-l-deoxy-Z- ketose (Figure 3). Amadori products are predominantly degraded to deoxyosones which are reactive a-dicarbonyl compounds and are involved in the formation of 18 REDUCING SUGAR + AMINO ACID DEOXYKETOSYL COMPOUND i CH3 HC=O d=o methyl l— deoxyosone dicarbonyl -0 intermediates STRECKER i=0 intermediates in DEGRADATION 2 HHOH lHOH amino acid ¢ ‘ + dicarbonyl FISSION DEHYDRATION Strecker . , aldehyde yields yields short chain 5-hydroxy carbonyls, methyl-2- dicarbonyls furaldehyde amino compound +CO2 v i v i i Melanoidin formation by the polymerization of intermediate compounds, production of heterocyclic compounds Figure 2 Simplified scheme of the Maillard reaction (Hodge, 1967). 19 REE RN HOH fl 1‘00 +RNH2 -H20 H (CHOH)n ( )n l HIOH H2 OH CHZOH aldose Schiff base / \+ RNH RNH RNH .1 .1 l. 0 —>(HCOH) 0—> HOH —_l H ———i ‘— Hi—J ‘— (H 0H)n HZOH HZOH HZOH N-suhstituted N-substituted \ cation of / glycosylamine aldosylamine Schiffbase RNH RNH _H. d1! 1H2 -> iiiOH ‘ > (i=0 N-substituted (H£OH)D I l-amino-l-deoxy (HCOH)n -2-ketosez keto form HZOH J: (Amadorr product) HZOH enol form Figure 3. Initial steps of the Maillard reaction (Hodge, 1953). 20 many heterocyclic volatile compounds (Ledl, 1990). The sugar-amine condensation is the first step in the so-called catalysis of amino compounds. After condensation and rearrangement, the sugar moiety is dehydrated and easily polymerizable, and unsaturated compounds are formed in which the amine moiety is labile (Hodge, 1953). In the condensation step, the amines act as a nucleophiles, their activities depending on the pH of the system (Ledl, 1990). Amines catalyze the enolization of sugar at pH values greater than 8, with the resultant production of low molecular weight fi'agrnents (Ledl, 1990). These early Maillard reactions do not cause browning or produce flavor compounds in food systems, although they may severely reduce the nutritive value (Hurrell and Carpenter, 1974). H. Intermediate stage of the Maillard reaction There are three main pathways in the advanced Maillard reaction, and all lead to the production of brown pigments or melanoidins. Two of these pathways begin with the Amadori compounds, while the third pathway is the Strecker degradation (Figure 2). In the first pathway, the Amadori product undergoes 2,3-enolization which eliminates the amine fi'om the C-1 position to form a methyl dicarbonyl intermediate (Hodge, 1953). This compound undergoes fission to produce C- methyl aldehydes, keto aldehydes, dicarbonyls and reductones (Hodge, 1967). The reaction products also include some flavor compounds such as acetaldehyde, pyruvaldehyde, diacetyl and acetic acid (Hurrel, 1982). In the second pathway, 3- deoxyosone, a compound formed from the enol form of the Amadori product by the elimination of the hydroxyl group at 03, may lose three molecules of water to 21 yield furfural (Hodge, 1953). The reactions which follow the formation of these intermediates may result in the production of dark brown nitrogen-containing pigments. The reactions involved are the aldol condensation and aldehyde-amino polymerization, and result in the formation of nitrogen-containing heterocyclic compounds such as pyrazines, pyrroles and pyridines (Hodge, 1953). These compounds are largely responsible for the roasted, bready and nutty flavor of heated foods. Strecker degradation The Strecker degradation is a reaction where the amino acids undergo oxidative decarboxylation and dearnination to produce carbon dioxide and Strecker aldehydes (Baltes _e_t_ Q, 1989). The reaction involves the oxidative degradation of free amino acids by the or-dicarbonyls and other conjugated dicarbonyl compounds produced by the breakdown of the Amadori product (Hurrel, 1982). Generally, the Strecker degradation takes place at temperature above 100°C, due to the requirement of a relative high activation energy associated with the amino acid decarboxylation step (Rizzi, 1987). The aldehydes formed from the Strecker degradation are a source of browning precursors. They can condense with themselves, with sugar fragments such as furfural and other dehydration products, and with aldimines and ketimines to form brown pigments (Hodge, 1953). With the liberation of carbon dioxide, aldehydes produced via the Strecker degradation contain one carbon less than the amino acid (Hodge, 1953). Maillard (1912) concluded that the carbon dioxide liberated in the Strecker degradation was from the carboxyl group of the or-amino acid, but not from the sugar radical. 22 With cysteine, the Strecker degradation yields hydrogen sulfide (Figure 4). Other than elemental sulfur, hydrogen sulfide is the simplest sulfur-containing chemical which can be obtained. In general, this chemical plays an important role in the formation of sulfur-containing heterocyclic compounds such as mercaptoketones, thiazoles and thiazolines, all of which contribute to meat and poultry flavor. It can also react with ammonia, aldehydes, dicarbonyl compounds, or carbonyls derived from lipid oxidation (Katz, 1981). Though the quantities of these heterocyclic sulfur compounds are not significant, they play an important role in the aroma of cooked meats due to their low threshold values (Guntert e_t Q, 1990). HI. Final stage of the Maillard reaction The major reactions involved in the final stage of the Maillard reaction are aldol condensation, aldehyde-amine polymerization, and the formation of heterocyclic compounds such as pyrroles, irnidazoles, pyridines and pyrazines (Hodge, 1953). With amine catalysts present, aldehdyes undergo an aldol condensation to produce melanoidins (Hodge, 1953). The condensation proceeds initially through the formation of the ordinary aldol between two molecules, then intramolecular condensation takes place to yield 2,5-dimethyl-p-quinone. The intermediate compound then decomposes to form color pigments at room temperature (Hodge, 1953). In the subsequent steps, reactions involve the interaction of furfural, firrfuranones, and dicarbonyl compounds with other reactive products such as ammonia, hydrogen sulfide, and thiols. These reactions lead to the formation of 23 HS - CHZCH - NH, Loon HS ' CH2\ H/ N\ ’ Cysteine O \ / \R” + -H 0 2 ’ OH O = ' R, Schiffbase o = c - R” Dicarbonyl 'Coz CH2=CH-\/R’ HS-CHZ-CH=N\ /R’ C C (fa R” H“ \R” l . H R’ 2N1}? + CH,CHo / O/ \R” aminoketone i O = C - R’ NH3 + O = _ R” Figure 4. Strecker degradation of cysteine (Kobayashi and Fujirnaki, 1965). 24 many important classes of heterocyclic flavor compounds such as pyrazines, thiazoles, thiazolines, oxazoles, thiophenes, and other sulfur-containing compounds (Mottram, 1991). F_actors influencing the Maillard reaction in foods A. Duration and temperature of heating Temperature has the greatest influence on the Maillard reaction, an increase in temperature resulting in an increased reaction rate (Leahy and Reineccius, 1989; Shu and Ho, 1989; Shaw and Ho, 1989; Ames and Apriyantono, 1994a; Spanier and Drurn-Boylston, 1994). However, the Maillard reaction can also occur in foods stored at refiigerated temperatures (Whitfield, 1992). The Maillard reaction is also affected by the duration of heating. For example, Apriyantono and Ames. (1990) found that the amounts of furans, pyrroles, and furfural from the Lickens and Nickerson extraction increased in a model system containing lysine and xylose, and then decreased after 2 h. It is suggested that the time and temperature of heating influence the advanced stage of the Maillard reaction (e.g., the Strecker degradation), because high energy is required in such reactions (Hurrell, 1982). B. Water content The Maillard reaction and the Strecker degradation have maximum reaction rates at the intermediate moisture level in most food products (Eichner and Karel, 1972). The rate of the Maillard reaction increases from the dry state, starting at critical water activities of 0.2 - 0.3 for most foods to a maximum at water activities of 0.5 - 0.8, then decreases at higher water activities (Baisier and Labuza, 1992; Labuza and Baisier, 1992). The reaction rate follows zero-order kinetics, with the 25 rate constants being drastically reduced with the addition of small amounts of water (Peterson e_t Q, 1994). The decreased reaction rates at higher water activities have generally been attributed to the dilution of the reactants. The decreased reaction rate at low water activities (< 0.1) has been ascribed to an increasing diffusion resistance which lowers the mobility of the reactants (Labuza _e_t Q, 1970). Therefore, the optimum Maillard reaction conditions are determined by the amount of water and state of water binding in a distinct system, and by the mobility of reactants (Mottram and Whitfield, 1995). Since moisture loss is a typical feature of thermally processed foods, the water content strongly affects volatile production through the thermal reaction. Zhang e_t Q (1994) determined that the greatest concentrations of volatiles were produced in model food systems with water contents ranging from 20% to 50%. When the water content was greater than 50%, smaller quantities of volatile compounds were produced, which might be due to the substrate dilution effect or the inhibition of condensation steps (Zhang _e_t Q, 1994). C. pH Model systems studies have established that the Maillard reaction is afi’ected by pH as both carbonyl and amino groups have the potential to be charged or uncharged depending on the hydrogen concentration of the system. For example, greatest quantities of pyrazine were observed at pH 9 - 10 (Bemis-Young et Q, 1993). The pH of the system influences the extent to which the Amadori rearrangement products degrade by 1,2-enolization (Baltes _e_t Q, 1989; Ames and Apriyantono, 19943). As the pH increases, the quantities of colored and polymeric compounds increase (Mauron, 1981; Ames and Apriyantono, 1994b). Alkaline conditions favor the formation of pyrazines, pyridines and some aliphatic sulfur 26 compounds including 2—acetylthiazole and dimethyl disulfide (Leahy and Reineccius, 1989; Mottram, 1990; Ames and Apriyantono, 1994a). However, the production of 2-furfural, furanthiols, methional and thermal degradation products of cysteine such as 2-thiophenethiol and trithiacycloheptene compounds, is reduced with increasing pH (Shu _e;t_ Q, 1985; Mottram, 1990). Some heterocyclic sulfur compounds (e. g., 1,2-dithia-4-one and furanmethanol) are largely unafiected by variation in pH (Mottram, 1990). In meat systems, meat is characterized by a pH value in the range of 5.5 - 6.0 and a high buffering capacity, therefore, very little change in pH occurs during cooking (Mottram, 1990). Thus, sulfur-containing heterocyclic volatile compounds may be predominant in meat products in which pH values are low (Ames and Apriyantono, 1994a). D. Sugars Only reducing sugars can take part in the Maillard reaction as they provide the necessary carbonyl groups (Hurrell, 1982). At 37°C and 15 % moisture, the order of reactivity of reducing sugars is ribose > xylose > glucose > lactose. Moreover, aldopentoses are more reactive than aldohexoses (Lewis and Lea, 1950) The Maillard reaction is also influenced by the relative concentrations of sugars and amino acids (W armbier g; Q, 1976; Shaw and Ho, 1989). The reaction rate increases linearly as the ratio of reducing sugar to free amino acid increased fiom 0.5 to 3.0. The rate does not change above this range. As the ratio increases, more sugar molecules may be in close proximity to the amino groups and overcome the diffusion barrier caused by the viscosity, thus increasing the reaction rate (Warmbier e_t Q, 1976). 27 Maillard reaction products and meat arom_a The main types of heat-induced reactions leading to the formation of meat flavor volatiles from non-volatile precursors have been summarized as follows (van den Ouwland e_t Q, 1978; Baltes e_t Q, 1989): 1. Pyrolysis of amino acids and peptides. 2. Sugar fragmentation. 3. Interactions involving sugars, peptides, amino acids, or their degradation products (Maillard reaction). 4. Degradation and reaction of ribonucleotides. 5. Reaction of hydrogen sulfide, ammonia, and thiols with non-volatile and volatile components. 6. Oxidation, hydrolysis, dehydration, and decarboxylation of lipids. 7. Thiamin degradation. Although many volatiles in meat flavor are derived by lipid oxidation, many researchers believe that the volatiles produced from water-soluble precursors may be the major contributors to meat flavor, especially heterocyclic compounds arising fi'om the Maillard reaction (Bailey, 1983; Fors, 1983). Many researchers have concluded that oxygen-, nitrogen- and sulfur- containing heterocyclic compounds contribute significantly to the overall desirable flavor of meat (Bailey, 1983; Whitfield @1411, 1988; Salter e_t Q, 1988; Farmer e_t Q, 1989). Lactones, acyclic sulfur-containing compounds (e. g. mercaptans and sulfides), non-aromatic heterocyclic compounds containing sulfur, nitrogen and oxygen (e. g. hydrofuranoids) and aromatic heterocyclic compounds containing sulfur, nitrogen and oxygen (e. g. pyrazines, thiazoles and furans) are probably the dominant contributors to meat flavor. In particular, sulfur compounds play a major role in 28 meat flavor due to their distinctive olfactory properties and generally low flavor thresholds (Guntert e_t Q, 1990). Classification of flavor compounds produced from the Maillard reaction The aroma compounds produced in the Maillard reaction can be classified into three groups as follows (Nursten , 1986): A. Simple sugar dehydration or fiagmentation products: furans pyrones cyclopentenes carbonyl compounds acids B. Simple amino acid degradation products: aldehydes sulfur-containing compounds (including hydrogen sulfide, methanethiol) nitrogen-containing compounds (including ammonia, amies) C. Volatiles produced by further interaction: pyrroles, pyridines, pyrazines imidazoles, oxazloles thiazole, thiophene di-, trithiolanes di-, trithianes furanthiols 29 Oxygen-conflingflvor compound from the Maillard reaction. F urans with oxygenated substituents such as furfural and furanones, contribute to flavor of all heated foods and are among the most abundant products of the Maillard reaction. These oxygenated furans generally impart caramel-like, sweet or fruity characteristics to foods (Mottram, 1994a). However, they are important intermediates in the formation of other flavor compounds including thiophenes, furanthiols and other sulfur containing compounds (Ruther and Baltes, 1994) Aliphatic carbonyl compounds, such as diacetyl which has a butter-like aroma, may contribute to the aromas derived from Maillard reaction. Many of the Strecker aldehydes also have characteristic sensory notes. For example, 3- methylbutanal which is derived from leucine, has a malt flavor (Mottram, 1994a). Nitrggpn—co taining compounds. Pyrazines are important aroma compounds and are believed to contribute to the pleasant and desirable flavor of many foods (Leahy and Reineccius, 1989; Maga, 1992; Weenen e_t Q, 1994). Several mechanisms have been proposed for the formation of pyrazines in cooked foods including their genesis from or- aminoketones. Aminoketones undergo self-condensation or condensation with other aminoketones to form a dihydropyrazine which is further oxidized to the pyrazine (Figure 5). The alkylpyrazines generally have nutty, roast aromas with some eliciting earthy or potato-like notes. However, the odor threshold values of the mono-, di-, tri- and tetrarnethylpyrazines are all relative high (> 1 ppm), and these pyrazines probably only play minor roles in food aromas (Fors, 1983). The other major nitrogen-containing compounds in food aromas are the pyrroles. Pyrroles have been reported to contribute to the caramel-like, sweet and 30 RR NH2 OVR NH3 fi'om + Strecker degradation . HH dicarbonyls Rt/ \0 HzN/ \R’ R’ R” R’ R” r ” My 1 IX MINI —> J: r. 0.... HM. dihydropyrazine -2H base R’ R” IIfla—w II: I R’ CH3 CHO R R” R’ alkylpyrazine -HZO RIN H,CH, R” R’ ethyl substituted pyrazine Figure 5. Pathway for the formation of pyrazine (Maga, 1992)- 31 com-like flavor (Fors, 1983), and may be important in defining roast beef aroma (MacLeod and Coppock, 1977). They may be formed from the reaction of 3- deoxyketose with ammonia or an amino compound followed by dehydration and ring closure (Mottram, 1991). Sulfur-containing compounds. Sulfur-containing compounds contribute both pleasant and unpleasant aroma to many foods. In meat, both aliphatic and heterocyclic sulfur compounds impart the characteristic flavor. MacLeod (1986) reported that 78 sulfur-containing compounds possess meat-like aroma, including 7 aliphatic and 65 heterocyclic sulfur compounds . Thiazoles and thiazolines are more profilic in food flavors than their oxygenated analogues such as oxazoles and oxazolines, and have lower threshold values (Mottram, 1994a, b). Both groups of compounds are important constituents of food aroma, especially in roast, grilled or fiied meat products (Maga, 1975; Mottram 1991). Some dialkylthiazoles have been reported in roast pork that are not present in other meats, while beef and chicken contain more trialkylthiazoles (Tang _e_t Q, 1983). The thermal degradation of thiamine is one of the sources of thiazole in heated foods. The formation of thiazole involves the reaction of hydrogen sulfide and ammonia with aliphatic aldehydes and dicarbonyl compounds (Figure 6), which is closely related to the pathway of oxazole (Mottram, 1994a). A number of polysulfur heterocyclic compounds containing two or three sulfur atoms in five or six membered ring have been identified as meat flavor volatiles (Mottram, 1994a; Mottram and Madruga, 1994; Mottram _e_t Q, 1995). 32 O NH3 NH R —-> R—< H H aldehyde imine R O I NH2 R S/ \ R R o st R o X ——> I R 0 R SH dicarbonyl mercapto ketone -H,o V [or R “x ‘ Z? S R R S R thim'e thiazoline Figure 6. Formation of thiazole and thiazoline form the intermediates of the Maillard reaction (Vemin and Parkanyi, 1982). 33 The most frequently found is 3, 5-dimethy1-1, 2, 4-trithiolane which was first isolated from boiled beef. Other sulfur-heterocyclics found in meat include trithioacetaldehyde, trithioacetone and thialdine (Mottram, 1991). Thiophenes have odor threshold values in the low part per billion (ng/g) ranges. Thus, they are considered to be major contributors to the aroma of foods (Mottram, 1994a). Thiophenes with a thiol group in the 3-position possess meaty aroma characteristics. Several acylthiophenes have been reported in pork liver, but 2-acetylthiophene has been found only in chicken (Mottram, 1991). The thiophenes with long alkyl chains may arise from lipid sources, probably through the interaction of hydrogen sulfide with unsaturated fatty acids or their oxidation products. These compounds have been formed upon heating a model system containing cysteine, ribose and phospholipid (Whitfield e_t al., 1988; Farmer _t Q, 1989; Farmer and Mottram, 1990a). A pathway has been proposed that involves the addition of hydrogen sulfur across the 4-ene group in 2,4-dienal, followed by ring closure and dehydration (Mottram and Salter, 1989). Role of Lipids in Meat Flavor Sigfl cance of lipid oxidation in flaLvor form_ation Lipids, proteins and carbohydrates are the major structural components of living cells. Lipids not only play a vital role in the metabolism of cells by providing a source of energy via oxidation, but also are an important source of food flavors (Ho and Chen, 1994). Lipids serve as a depot for fat-soluble compounds which are volatilized when foods are heated. This particular function of lipids strongly affects the flavor of cooked meats (Shahidi, 1989). In addition, 34 lipids may act as a solvent for the accumulation of flavor compounds during meat processing and cooking (Mottram and Edwards, 1983). Lipid oxidation products can also contribute to species-specific meat flavors, in that a "chickeny" or a "beefy" flavor is strongly associated with special carbonyl compounds (Shahidi, 1989) There are many catalytic systems in foods that oxidize lipids such as light, temperature, enzymes, metal and metalproteins, and microorganisms (Vercellotti _e_t Q, 1992). As lipids oxidize, they form hydroperoxides which are susceptible to further oxidation or decomposition to secondary reaction products (Figure 7). Lipid oxidation affects the flavor, aroma, taste, nutritional value and overall quality of foods (Vercellotti e_t Q, 1992; Ho and Chen, 1994); however, only the volatile lipid oxidation compounds impart the desirable meat aroma. Mottram et Q (1982) reported that alcohols and aldehydes were the predominant components in the volatile isolated from cooked beef and pork. These investigators further (1983) investigated the roles of triacylglycerols and phospholipids in the formation of meat aroma, and observed that the removal of the triacylglycerols had little effect on the aroma of the cooked meat or on the pattern of volatile compounds observed using gas chromatography - mass spectrometry. The additional extraction of phospholipids, however, eliminated the "meaty" character of the odor and resulted in a marked alteration of the volatile components. It was concluded that the presence of lipids is necessary for the development of the full meaty aroma. Phospholipids, being more susceptible to oxidation, appear to provide sufficient flavor formation, while triacylglycerols apparently are not essential. This is due to the unsaturated carbon chains of the lipids reacting with oxygen to form a number 35 polyunsaturated fatty acid (PUFA) -H free radicals l activated peroxide O 2 ) free radicals hydroperoxides hydrocarbon aldehydes ketones alcohols other products Figure 7. Mechanism of lipid oxidation and formation of carbonyl compounds (Shahidi, 1989). 36 of volatile compounds which can contribute either a desirable or an undesirable flavor (Mottram, 1987). Mechanism of lipid oxidation Lipid oxidation is generally considered to be a free radical process. The reaction of unsaturated lipids with oxygen to form hydroperoxides involves three basic steps, initiation, propagation and termination, as shown below (Frankel, 19soy Initiation: initiator (heat, light) RH 4, R. Propagation: R0 + 02 4, ROO- (fast) ROOO + RH 4, ROOH + R0 (slow) ROOH 4, R0. + OH. Termination: ROOo + R000 47 ROOR + 02 R000 + Ro 4, ROOR R0 + R0 A, R - R R0 (R000 + R00 + 0H.) + AH ----—+ RH (ROOH + ROH + H20) + A- Initiation occurs when a labile hydrogen atom is abstracted from a fatty acid molecule (RH) in the presence of an initiator (such as heat or light) to form a carbon-centeredalkyl radical (Ru). This alkyl radical reacts rapidly with oxygen to form a peroxy radical (R000). The peroxy radical in turn is capable of abstracting 37 a hydrogen atom from another unsaturated fatty acid, and hence propagates the chain reaction. The propagation results in the formation of hydroperoxides (ROOH) which then decompose to a variety of secondary products. The termination step involves the interaction of free radicals, e. g. R0 and R000, to form non-initiating and non-propagating products (Gray and Crackel, 1992; Ho and Chen, 1994). In the presence of antioxidants (AH), the chain-breaking antioxidants may donate hydrogen atoms to free radical species (R0, ROOO, R00), and form less reactive products (e. g. RH, ROH, ROOH, A0), thus, terminate the propagating reaction (Monahan, 1992). The primary site for oxidative attack is the methylene group (CH2) adjacent to a double bond, with the initial step involving abstraction of a labile hydrogen from the allyl methylene group (Mottram, 1991). The resulting radical undergoes rearrangement prior to reaction with oxygen, giving rise to a number of different hydroperoxides (Figure 8). Hydroperoxides are the primary initial products of lipid oxidation, and are essentially odorless (Paquette p; Q, 1985). Degradation of hydroperoxides involves homolytic cleavage of the -OOH group, giving rise to an alkoxy radical and a hydroxyl radical (shown as below). R- (IZH-R' 4. R-CIZH-R‘ + OH- OOH Oo alkoxy fiee + hydroxy fi'ee radical radical The alkoxy radical then undergoes B-scission on the C-C bond, with the formation of an aldehyde and an alkyl or vinyl radical (Mottrarrr, 1991; Ho and Chen, 1994). A general reaction scheme for the formation of volatile aldehydes, 38 CH,- --(CH,)6 (2H,- CH= CH CH, --(CH,), COOH / \ -CH-CH=CH-CH- -CH-CH=CH-CH- / 02 02 \ -CH=CH-CH-CH- -CH—CH-CH=CH- -CH-CH=CH-CH- -CH-CH=CH-CH- 00H 00H ()2 ()2 -CH=CH-CH-CH— -CH-CH-CH= - OH OOH Figure 8. Hydroperoxides formed by the oxidation of oleic acid (Mottram, 1987) 39 alkenes, alkanes and alcohols is illustrated in Figure 9. These secondary lipid oxidation products are the main contributors to off-flavor in meats (Gray and Crackel, 1992). Lipid oxidation during cooki_ng Lipids in meat undergo oxidation during cooking to produce a range of volatile compounds. Compounds found from the reaction are dominated by aldehydes, alcohols, furans and hydrocarbons, and are similar to those arising from the autoxidation of stored fats (Mottram, 1987). In general, volatiles produced from the thermal oxidation of lipids in cooked meats contribute to the desirable flavor and do not impart off-flavors. However, if the off-flavors have already developed before cooking, they can’t be removed by the cooking process (Mottram, 1987). Thus, subfle differences exist in the mechanisms of oxidation occurring during thermal processing and storage. These may be explained by the formation of hydroperoxides. The amounts of hydroperoxides in heated fat are always small because they are extremely heat labile. At low temperatures, hydroperoxides are more stable and may be found in significant high concentrations before degrading to other volatile compounds. This will lead to different proportions of the various radical intermediates, and will result in differences in the relative amounts of volatiles and varieties of compounds formed. A breakdown mechanism unique to the heated system is the oxidation of saturated lipids which does not occur at storage temperatures and will give compounds not produced by low temperature oxidation. These two systems will lead to the generation of a number of volatiles; however, both qualitative and 4O R,-CH=CH-CH-R, _OH, OOH (A) (B) R,- CH= CH CH-R, Scission0 . Scission (B) R,-CH= CH* + R, CHo Rl CH= CH CHO + R; 03 N M02 R1'CH=CH'O'O* R1“CH=CH2 R211 Rz‘O'O‘ alkene alkane 1m .. R,-CH=CH-o-0H Rz'O'OH ‘-0H* -OH‘¢ R- CH= CH-O" 112-0: Rim RH -=CH CH- OH Rz-OH alcohol Rl-CHZ-CHO aldehyde Figure 9. General reaction pathway for the hemolytic cleavage of hydroperoxides of unsaturated lipids (Ho and Chen, 1994). 41 quantitative are different (Mottram, 1987). Interaction of Lipids in the Maillard Reaction The main sources of flavor formation in heated foods are the Maillard reaction and the thermal degradation of lipids. Lipids, particularly the structural phospholipids, are essential for the development of the characteristic flavor in cooked meat, while the interaction of the products of lipid oxidation with Maillard reaction products is also an important route to flavor development (Mottram and Salter, 1989). The precise lipid composition of a food will affect the balance of the many flavor-forming reactions occurring during cooking and hence will influence the overall flavor and aroma of the food (Farmer and Mottram, 1994). Though the interaction of lipid oxidation and Maillard reaction products plays an important role in meat flavor, lipids are capable of suppressing the formation of certain heterocyclic compounds by competing with sugar-derived dicarbonyl for the Maillard reaction intermediates (Farmer and Mottram, 1990a, 1994). Mottram and Whitfield ( 1987b) found that the removal of phospholipids, as well as triacylglycerols from meat products before cooking, resulted in marked differences in aroma and an increase in the concentrations of volatile heterocyclic compounds, particularly alkylpyrazines. The enhanced formation of these compounds implied that lipids or their degradation products may suppress their formation by participating in the Maillard reaction. This effect could be due to lipid oxidation products reacting with Strecker degradation products such as ammonia and hydrogen sulfide, thus reducing the extent of the Maillard reaction (Mottram and Whitfield, 1987b). There is also evidence of a distinct flavor change 42 in cooked food caused by the interaction of lipids with Maillard reaction products. For instance, Saittagaroon e_t Q (1984) found that defatted ground coconut lost its characteristic sweet aroma after roasting, and developed a burnt, nut-like aroma. The uncooked coconut contained significant amounts of lactones which provided the sweet aroma note. On roasting, pyrazines, pyrroles and furans were formed and added a strong nut-like characteristic to the sweet aroma of the unroasted coconut. Studies of the involvement of lipid oxidation in the Maillard reaction The formation of volatiles from the interaction between lipid and its oxidation derivatives with Maillard reaction intermediates have been studied in detail in many model systems. Examples are as follows: A. Interaction of amino acid degradation products with carbonyl compounds. The involvement of lipid oxidation carbonyls in the Maillard reaction was first reported by Barch (1952), who claimed that the reaction of gaseous hydrogen sulfide (a specific Strecker degradation product from cysteine) with an aliphatic aldehyde developed from lipid oxidation led to the formation of flavoring compounds. Though compounds responsible for flavor were not identified, onion- and bacon-like flavors were produced by this reaction. Other researchers have investigated the reactions of acetaldehyde with hydrogen sulfide and ammonia; acetaldehyde or propanal with hydrogen sulfide, methanethiol, ethanethiol or propanethiol; and 2-hexenal or 2,4-decadienal with hydrogen sulfide and methanethiol (Boelens e_t Q, 1974). Volatile compounds generated from the reaction of saturated aldehydes with hydrogen sulfide at atmospheric pressure and temperature included dioxathianes and trithianes. Many of the compounds f0! 0X ICE Da cat rea llpj 0011 low all}: 43 synthesized in these studies have been found in flavor extracts, particularly those from allimns and boiled meat and beef broth (Boelens gt Q, 1974). An extension of this study was the reaction of pentanal or hexanal, common lipid oxidation products, with ammonium sulfide (Hwang e_t Q, 1986). F orty— seven compounds, including forty-two containing at least one hetero atom, were identified, and the most abundant compounds were 2,3,5-trialkylpyridines. Many of the compounds formed are present in the volatile extracts of fiied chicken (Hwang e_t Q, 1986). The identification of heterocyclic compounds from reactions between fatty aldehydes and Maillard reaction products indicates that they could be formed in foods during cooking. Studies have shown that aldehydes arising via lipid oxidation, viz. pentanal, hexanal, 2-hexenal, 2-decenal and 2,4-decadienal, can readily enter into these reactions (Whitfield, 1992). I. Interaction between amino acids and carbonyl compounds. When heptanal was used to react with glycine at temperatures between 10 and 550C, 2-alkenals were the only carbonyl compounds found (Montgomery and Day, 1965). It was suggested that the amino group of glycine acted as a basic catalyst in an aldol condensation reaction. These investigators concluded that the reaction could lead to the polymerization of carbonyl compounds generated during lipid oxidation, which in turn would lead to the depletion of carbonyl flavor components in some foods. Zhang and Ho (1989) studied the volatiles produced from deep-fat fiying foods by heating a mixture of 2,4-decadienal with cysteine in an aqueous system. A total of 45 compounds was obtained, including a large number of long-chain alkyl-substituted heterocyclic compounds such as thiophenes, pyridines, thiazoles, 44 and other sulfur-containing compounds. It was concluded that short-chain aldehydes such as acetaldehyde, butanal and hexanal, were probably derived fi'om the decomposition of 2,4-decadienal. These aldehydes then participated in the formation of the heterocyclic compounds, although the mechanism is unclear. The variety of volatile compounds formed during the reaction of aliphatic aldehydes with amino acids suggests that these reactions could be an important part of the interactions of Maillard reaction products and lipids. Many studies have shown the ease with which carbonyl compounds react with amino acids to form heterocyclic compounds (Ho e_t Q, 1989; Zhang and Ho, 1989). In addition, those reactions involving 2,4—decadienal showed that the carbonyl compounds can be converted into a number of volatile compounds through degradation and condensation reactions (Whitfield, 1992). Future studies in this field should study the reaction of other major lipid oxidation products such as 2-alkenals and l-alken- 3-ones with amino acids. Such reactions could lead to the identification of many new potent flavor compounds, as yet not detected in cooked or processed foods (Whitfield, 1992). C. Interaction of Maillard reaction products and triacylglycerols or phospholipids in model systems. Ledl and Severin (1973) heated cysteine, xylose and tributyrin at 200°C to study the volatiles formed from the interaction of Maillard reaction products with lipids. Sixty-nine compounds were identified including 19 with sulfur, 13 with nitrogen, and 22 with both sulfur and nitrogen. Compounds identified included furans or furanones, pyrroles, pyridines, thiophenes, thiazoles, and a number of other heterocyclic compounds. The detection of five propylthiazoles and one propenylthiazole also indicated that the triacylglycerol components of the reaction ICE C011 as pie: \W \e: i 45 ‘ mixture could be involved in the formation of these compounds. Model system studies of the effect of lipids on the formation of flavor of roasted cocoa were investigated using different amino acids, fi'uctose in water and deodorized cocoa- butter heated at 120°C for 3 h (Amolidi et al., 1987, 1990). A total of 78 compounds including furans, pyrroles and pyrazines were identified; however, they did not report any information on the possible interaction between Maillard reaction products and lipids. The effect of phospholipids on the formation of volatile heterocyclic compounds produced during the cooking of meat was investigated using model systems containing glycine, lysine or cysteine, and ribose that were heated in the presence and absence of egg yolk lecithin (Mottram and Whitfield, 1987, Whitfield e1 Q, 1988). In the simple Maillard reaction mixtures (i.e., those without lecithin), alkylpyrazines were the predominant volatile products when either glycine or lysine was used as the nitrogen source. In the system with cysteine, sulfur-containing heterocyclic compounds such as alkylthiophenes and alkylthiazoles, were predominant. When lecithin was added to these Maillard reaction mixtures, the quantity of the heterocyclic compounds formed was decreased (Whitfield et Q, 1988). In general, the addition of lecithin to simple Maillard reaction mixtures exerted some quenching efl‘ect on the production of heterocyclic volatiles. For example, smaller quantities of the alkylpyrazines were produced in the presence of lipids; the trend was more obvious with the alkylthiophenes and alkylthiazoles (Whitfield _e_t _a_1., 1988). The decrease was due to preferential reaction of fiee ammonia or the amino acid nitrogen with compounds derived from lipid oxidation rather than with sugar-derived carbonyls. The aliphatic aldehydes were considered to be the most likely compounds to react with ammonia and amino acids to form 46 polymeric materials. In a continuation of these studies, efforts were made to investigate the effect of lipid composition on the production of specific compounds from the Maillard / lipid interaction (Mottram and Salter, 1989). It was found that the fatty acid composition of phospholipids, as well as the phosphatidyl group, would produce different outcome fiom that of triacylglycerols in defatted meat. Farmer and Mottram (1990a) heated cysteine and ribose in the presence of beef triacylglycerols and three phospholipids (phosphatidylcholine (PC), phosphatidylethanolamine (PE) and beef phospholipid) to study the interaction between lipids and Maillard reaction products. They classified four groups of compounds which were produced as follows: [A] Compounds formed only in the presence of lipid: The greatest differences between the effects of the four lipids were observed for those compounds which required lipid precursors for their formation. Such compounds, e.g., 2-pentylpyridine, 2-alkylthiophenes, 2-akenylthiophens, were present in greater quantities in the systems containing the phospholipids. Of the three phospholipids investigated, systems containing PC produced the greatest amounts of the compounds. [B] Compounds produced from the reaction of cysteine and ribose that were suppressed by the presence of lipids: (a). Phospholipids showed ggeater suppression thap beef triacylglycerols. Heterocyclic compounds, e. g. mercaptocarbonyls, tended to be reduced considerably by phospholipids. These compounds were probably formed by the reaction of hydrogen sulfide with dicarbonyl or or,B-unsaturated 47 carbonyl compounds (Takken et al., 1976). Reduction in their formation may be due to the removal of hydrogen sulfide fi'om solution by its reaction with the polyunsaturated fatty acids of the phospholipids. (b). PE and beef phospholipids showed ggeater suppression then beef triacylglycerols. The quantities of some thiol-substituted compounds were markedly reduced in the presence of PC and beef triacylglycerols, and suppressed even more in the presence of PE and beef phospholipids. Compounds classified in this group include 3-furanthiols and 2-methyl-3-furanthiols. (c). Beef triacylgIvcerols showed greQer suppression than phospholipids. Compounds suppressed include 2,3-dihydro-6-methylthieno[2,3]firran and isomers of thienothiophenes, as well as a series of methyl-substituted thienothiophenes and dihydrothienothiophenes. The amounts of some of these substances were reduced by phospholipids, but they were all strongly suppressed by the presence of triacylglycerols. (d). Triacylglycerols caused similar reductions to Qrospholipids. In contrast to the above compounds, only small variations between the different lipids were observed for other compounds. For example, with 3(2H)-thiophenones the effects of lipid were relative small, although there was a consistent trend towards reduced levels in the presence of all the lipids. 48 [C] Compounds largely unaflected by lipid: Three dithianones were included in this group and their mechanism of formation has been proposed in relation to the thiophenones (Hartman gt Q, 1984). Other compounds such as certain thiazoles were reduced slightly by the addition of lipid, especially PC (Whitfield 91 Q, 1988); however, the differences were small and within the range of standard deviation. Thus, no clear tendencies can be inferred. [D] Compounds not included in previous categories: 2-furfural was markedly influenced by the presence of the various lipids. Concentrations were approximately halved by the addition of PC or PE, while beef triacylglycerols and beef phospholipids gave lesser extents of suppression (Whitfield e_t Q, 1988). The addition of lipids would not be expected to affect the formation of furfural from ribose; however, in the presence of lipid degradation products further reactions may occur to form polymeric materials or volatile fragmentation products (Shibarnoto and Bernhard, 1977). Two ethylmethyloxazoles also displayed patterns similar to that of 2-furfural. Possible mechanisms for the interaction of lipid oxidation products with Maillard reafiction products The most likely mechanism by which lipids can interact in the Maillard reaction may be summarized as follows: 1. The reaction of lipid-derived carbonyl products with the amino group(s) of an anrino acid and the ammonia produced by its Strecker degradation. 49 2. The reaction of the amino group of ethanolarnine (in phosphatidylethanolamine) with lipid-derived carbonyl compounds. 3. The interaction of free radicals from peroxidized lipids in the Maillard reaction. 4. The reaction of hydroxyl and carbonyl lipid oxidation products with free hydrogen sulfide which is derived from the Strecker degradation of cysteine (Mottram and Salter, 1989; Farmer and Mottram, 1990a; Whitfield, 1992). The different effects of the triacylglycerols and phospholipids may be due to the dissirnilarities in fatty acid composition and the polar moieties of the phospholipids (Farmer and Mottram, 1990a). Less than 2% of the fatty acids in beef triacylglycerols contain two or more double bonds, while in beef phospholipids, at least 20% of the fatty acids are polyunsaturated (Farmer and Mottram, 1990a). Thus, triacylglycerols lack the most reactive precursors for both the formation of long chain heterocyclic compounds and for the reaction with Maillard reaction intermediates such as free hydrogen sulfide and ammonia. The participation of triacylglycerols in the Maillard reaction may also differ from that of phospholipids due to the fact that triacylglycerols are less miscible with the aqueous Maillard reactants than the phospholipids. It is possible that these physical effects limit the participation of triacylglycerols in the Maillard reaction in both model systems and meat (Farmer and Mottram, 1990a). 50 Cured Meat Flavor Cured meats are distinctively attractive to consumers because of their pink color, characteristic texture and unique flavor. The curing process is achieved by adding a number of curing agents to the meat, each ingredient imparting its own unique characteristics (Gray and Pearson, 1984). The origin of the use of nitrite in the curing process is lost in history, but it is believed that the preservation of meat with salt has been used for many centuries. Kramlich e_t Q (1973) stated that salting of fish as a means of preservation can be traced back to before 3500 BC, and the use of nitrite was discovered probably through the presence of impurities in the salt. Currently, meat curing involves the addition of sodium nitrite and salt along with sugar, some reducing agents such as sodium erythorbate, phosphates, and seasonings to impart the characteristic properties to cured meat products. Role of nitrite in cured meats Nitrite is a unique curing ingredient that imparts many characteristics to cured meats, such as color, texture, and flavor. In addition, it has a potent antibacterial effect. Nitrite produces nitric oxide in the absence of light and air, which then combines with myoglobin to form nitric oxide metrnyoglobin. This compound is then reduced to nitric oxide myoglobin and is connected to nitrosohemochromogen under the influence of smoke or heat. Nitrosohemochromogen is the typical cured meat pigment which imparts the desirable pink color to cured meat (Kramlich at Q, 1973; MacDougall, e_t Q, 1975). 51 Nitrite also improves the texture of cured meat (Eakes e_t Q, 1975) and has been reported to increase the firmness of country-style hams. Nitrite serves as an antimicrobial agent and prevents the outgrowth of Clostridium botulinum spores and the subsequent formation of a lethal toxin, particularly under conditions of mishandling (Hotchkiss and Gassens, 1987). In the aspect of cured flavor, Cho and Bratzler (1970) and MacDougall at Q (1975) suggested that nitrite reacts with tissue components such as thiol- and amino-containing compounds, to produce hitherto unidentified products which might contribute to the unique flavor. Moreover, nitrite provides oxidative stability to meat by preventing lipid oxidation. Antioxidant role of nitrite in cured meats A number of studies have shown that nitrite retards lipid oxidation or the development of warmed-over flavor (WOF) in cooked meat and meat products (Sato and Hegarty, 1971; Swain, 1972; Fooladi e_t Q, 1979; Gray and Pearson, 1984; Ramarathnam e_t Q, 1991a, b). Sato and Hegarty (1971) found that WOF was completely eliminated by adding nitrite at a level of 2000 mg/kg meat. F ooladi _e_t Q (1979) confirmed the effect by demonstrating that nitrite (156 mg/kg) inhibited WOF in cooked meat with a 2-fold reduction in TBA (2-thiobarbituric acid) values for beef and chicken, and a 5-fold reduction in pork. Recently, Ramarathnam e_t Q (1991a) reported that the concentration of hexanal which is produced from lipid oxidation was approximately 400 times greater in uncured pork compared to cured pork. These results confirm the data of many previous studies that formation of hexanal was suppressed in the presence of nitrite (Cross and Ziegler, 1965; Swain, 1972). The mechanism by which nitrite prevents or retards lipid oxidation in cured meat is not fully understood. Results suggest that more one mechanism is involved 52 (Gray and Pearson, 1994). Four possible mechanisms have been proposed as follows: 1. Stabilization of the unsaturated fatty acids by forming nitrite-derived antioxidants such as S-nitrosocysteine or nitrogen oxide myoglobin (Kanner e_t Q, 1980). 2. Formation of a strong complex with the heme pigments, thereby, preventing the release of ferrous ion during processing / cooking with its attendant catalysis of the propagation stage of lipid oxidation (Igene e_t Q, 1985). 3. "Chelation" of metal ions, e.g., ferrous ions, thus rendering them unavailable for catalysis of oxidation reaction (Monissey and Tichivangana, 1985). 4. Stabilization of polyunsaturated lipids within the membranes (Freybler 91 Q, 1993). Formafin of nitrite-derived antioxida_nt_§ The proposed mechanism involves the formation of certain nitrosyl compounds. For example, nitrogen oxide myoglobin can quench free radicals, thus lowering the amounts of prooxidants in the meat (Kanner e_t Q, 1980). Other nitrite reaction products such as S-nitrosocysteine also possess antioxidant properties. Kanner (1979) demonstrated the antioxidant potential of this compound in an aqueous linoleate model system. Kanner e_t Q (1984) found that both hemin nitroxide and cysteine-iron-nitroxide have antioxidant properties. These compounds are believed to act in a similar manner as nitric oxide myoglobin by quenching free radicals. 53 Stabilization of the heme pigments Zipser _e_t Q (1964) first proposed that nitrite can form a stable complex with the iron porphyrins in heated meat systems, thus, preventing heme-catalyzed lipid oxidation. A study of the effect of nitrite on the stability of heme pigments in beef extract was acomplished by Monissey and Tichivagana (1985). They reported that extracts without nitrite contained greater amounts of free irons after heating; however, beef extracts with nitrite were not found any changes in the amounts of free or heme irons after heating. They also suggested that nitrite reacted with myoglobin to form nitric oxide myoglobin, and a corresponding stable complex of nitric oxide hemochromogen was formed after heating, thus preventing the release of free iron. Chelafion of metal ion_s It has also been proposed that nitrite can inhibit the lipid oxidation by chelating trace metals (MacDonald e_t Q, 1980). When a small amount of nitrite (< 25 mg/kg) was added to a system containing linoleic acid, phosphate buffer, Tween 20 and ferrous iron, the amount of lipid oxidation was greatly reduced. Morrissey and Tichivangana (1985) also postulated that nitrite probably formed inactive 'chelates' or complexes with nonhaem iron, copper and cobalt. These studies used water-extracted muscle fibers to which was then added the various prooxidants (ferrous, copper and cobalt irons) and nitrite. The results indicated that nitrite at concentrations as low as 20 mg/kg, caused a significant reduction in TBA values. Maximum inhibition of lipid oxidation was observed when 50 mg/kg was used. Thus, the investigators concluded that nitrite must react with these ions, thus preventing lipid oxidation. 54 StabilizQion of h'pids Pearson e_t Q (1977) suggested that nitrite stabilizes the membrane lipids in cured meat, thereby inhibiting lipid oxidation. Goutefongea 91 Q (1977) demonstrated that nitrite reacted with unsaturated lipids, and reported the amount of binding of nitrite was related to the degree of unsaturation of the lipids. They hypothesized that nitrite or a derivative reacted with one or more of the carbon- carbon double bonds in the adipose tissues. Liu e_t Q (1988) further investigated the stabilization of lipids in cured meats and concluded that nitrite reacted with unsaturated fatty acids to form nitro-nitrosite derivatives (Figure 10). F reybler e_t Q (1993) reacted unsaturated lipids with dinitrogen trioxide and found an increase of stability of these lipids to peroxidative changes. They also reported that phospholipids, and rnicrosomes and mitochondria fiom cured pork were less susceptible to metrnyoglobin / hydrogen peroxide-catalyzed peroxidation than their counterparts from nitrite-free pork samples. These results indicate that nitrite reacts with unsaturated lipids to form nitro-nitroso derivatives, thus stabilizing the lipids toward peroxidation changes (Freybler e_t Q, 1993). Chemispy of cured meat flavor Although nitrite is clearly associated with cured meat flavor, the chemical changes for the unique flavor are not entirely understood (Gray and Pearson, 1984). Consequently, many researchers have attempted to identify the volatile compounds isolated from cured and uncured meat (Swain, 1972; Bailey and Swain, 1973; Mottram and Rhodes, 1973; MacDougall e_t Q, 1975; Gray e_t Q, 1981; Shahidi, 1989; Neol e_t Q, 1990; Ramarathnam e_t Q, 1991a, b, 1993a, b, 1994). 55 NaNO2 H N203 + Na Figure 10. Mechanism for the formation of nitrite with polyunsaturated lipids (Liu e_t Q, 1988) 56 Cross and Ziegler (1965) analyzed the volatile compounds from both uncured and cured ham and found quantitative differences. Pentanal and hexanal were present in appreciable quantities in the uncured product, whereas the cured hams contained smaller amounts. It was suggested that the smaller concentrations of these aldehydes in cured meats were responsible, in part, for the flavor differences between uncured and cured products. The investigators also found that the volatiles, after passing through a solution of 2,4-dinitropheny1hydrazine, still had a characteristic cured ham aroma regardless of whether they originated from cured or uncured ham. They concluded that "cured-meat" flavor was composed of essentially the same constituents that were generated by a combination of several reactions, in addition to suppression of lipid oxidation. Similar results were observed by Swain (1972), who also found that volatiles from hams treated with and without nitrite were qualitatively similar but quantitatively different. Several alkyl nitrates, and aryl- and alkane-nitriles have been found in pork cured with nitrite (Mottram _e_t Q, 1973, 1978, 1984a, b, 1985a). The most likely origin of these compounds is from the reaction between fatty acids and sodium nitrite. The thermal oxidation of unsaturated fatty acids generally proceeds via a free radical mechanisms, thus alkyl nitrates could result from the participation of nitrous acid, or a free radical derived from nitrite, in these decomposition reactions (Mottram, 1984b). A mechanism for the formation of nitriles from the interaction between nitrite and lipids involves the C-nitrosation of a methylene group in the aliphatic chain of a fatty acid. The reaction is activated by an adjacent carbonyl, carboxyl or similar group. Thermal degradation of the resulting oxirne could give the CN group as shown bellows. 57 nitrosation R-CHz-CO- > R-CHz-CO- +R- C-CO I ll NO NOH R - C - CO heat II > R - C E N NOH BIBLIOGRAPHY Ames, J. M. and Apriyantono, A. 1994a. Effect of pH on the volatile compounds formed in a xylose-lysine model system. In "Thermal Generated Flavor: Maillard, Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 228-239. Ames, J. M. and Apriyantono, A. 1994b. Comparison of the non-volatile ethyl acetate-extractable reaction products formed in a xylose-lysine model system heated with and without pH control. Food Chem. 50(3): 289-292. Ames, J. M. and Apriyantono, A. 1993. Volatile reaction products from a heated xylose-lysine model system. Food Chem. 48(3): 271-277. Apriyantono, A. and Ames, J. M. 1990. Volatile compounds produced in heating lysine with xylose. In "Flavor Science and Technology", Y. Bessiére and A. F. Thomas (Ed), Wiley Chichester, pp 117-120. Amolidi, A., Amolidi, C., Baldi, O. and Ghizzoni, C. 1990. Effect of lipid in the Maillard reaction. In the "Maillard Reaction in Food Processing, Human Nutrition and Physiology", P. A. Finot, H. U. Aeschbacher, R. F. Hurrell and R. Liardon (Ed), Birkhauser, Switzerland, pp 133-138. Amolidi, A., Amolidi, C., Baldi, O. and Griffrni, A. 1987. Strecker degradation of leucine and valine in a lipidic model system. J. Agric. Food Chem. 35(6): 1035-1038. 58 Bailey, M. E. 1994. Maillard reactions and meat flavor development. In "Flavor of Meat and Meat Products", F. Shahidi, Ed, Blackie Academic and Professional, Glasgow, UK, pp 153-173. Bailey, M. E. and Um, K. W. 1992. Maillard reaction products and lipid oxidation. In "Lipid Oxidation in Food", A. J. St. Angelo (Ed), ACS Symposium Series 500, Washington DC, pp 122-139. Bailey, M. E. and Einig, R. C. 1989. Reaction flavor of meat. In “Thermal Generation of Aromas” , T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 421-432. Bailey, M. E. 1983. The Maillard reaction and meat flavor. In "The Maillard Reaction in Food and Nutrition", G. R. Waller and M. S. Feather (Ed), ACS Symposium Series 215, Washington, DC, pp 169-184. Bailey, M. E. and Swain, J. W. 1973. Influence of nitrite on meat flavor. In "Proceedings of the Meat Industry Research Conference", American Meat Science Association, Chicago, pp 29-45. Baines, D. A. and Mlotkiewicz, J. A. 1984. The chemistry of meat flavor. In "Recent Advances in the Chemistry of Meat", A. J. Bailey (Ed), Royal Society of Chemistry, London, pp 119-164. Baisier, W. M. and Labuza T. P. 1992. Maillard browning kinetics in a liquid model system. J. Agric. Food Chem. 40(5): 707-713. Baltes, W. 1990. Roast aroma formation: the role of amino acids during the Maillard reaction. In "the Maillard Reaction in Food Processing, Human Nutrition and Physiology, P. A. F inot, H. U. Aeschbacher, R. F. Hurrell and R. Liardon (Ed), Birkhauser, Switzerland, pp 43-62. Baltes, W., Kunert-Kirchhoff, J. and Reese, G. 1989. Model reactions on generation of thermal aroma compounds. In "Thermal Generation of Aromas" , T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 143-155. Batch, W. E. 1952. U. 8. Patent 2,594,379. 59 Bemis-Young, G. L., Huang, J. and Bernhard, R. A. 1993. Effect of pH on pyrazine formation in glucose-glycine model systems. Food Chem. 46(4): 383-3 87. Boelens, M., van der Linde, L. M., de Valois, P. J ., van Dort, H. M. and Takken, H. J. 1974. Organic sulfur compounds from fatty aldehydes, hydrogen sulfide, thiols and ammonia as flavor constituents. J. Agric. Food Chem. 22(6): 1071-1076. Cho, 1. C. and Bratzler, L. J. 1970. Effect of sodium nitrite on flavor of cured pork. J. Food Sci. 35: 668. Cross, C. K. and Ziegler, P. 1965. A comparison of the volatile fractions from cured and uncured meat. J. Food Sci. 30: 610-614. Eakes, B. D., Blumer, T. N. and Monroe, R. J. 1975. Effect of nitrate and nitrite on color and flavor of country-style hams. J. Food Sci. 40: 973-976. Eichner, K and Karel, M. 1972. The influence of water content and water activity on the sugar-amino browning reaaction in model systems under various conditions. J. Agric. Food Chem. 20: 218-223. Elizalde, B. E., Dalla Rosa, M. and Lerici, C. R. 1991. Effect of Maillard reaction volatiles products on lipid oxidation. JAOCS 68(10): 758-762. Farmer, L. J. and Mottram, D. S. 1994. Lipid-Maillard interactions in the formation of volatile aroma compounds. In “Developments in Food Science 35: Trends in Flavor Research”, H. Maarse and D. G. van der Heij (Ed), pp 3 13-3 26. Farmer, L. J. 1992. Meat flavour. In "the Chemistry of Muscle-Based Foods", D. E. Johnson, M. K. Knight and D. A. Ledward (Ed), Royal Society of Chemistry, Science Park, Cambridge, UK, pp 169-182. Farmer, L. J. and Mottram, D. S. 1992. Effect of cysteine and ribose on the volatiles thermal degradation products of a triglyceride and three phospholipids. J. Sci. Food Agric. 60(4): 489-497. Farmer, L. J. and Patterson, R. L. S. 1991. Compounds contributing to meat flavor. Food Chem. 40: 201-205. 60 Farmer, L. J. and Mottram, D. S. 1990a. Interaction of lipid in the Maillard reaction between cysteine and ribose: the effect of a triglyceride and three phospholipids on the volatile products. J. Sci. Food Agric. 53(11): 505-525. Farmer, L. J. and Mottranr, D. S. 1990b. Recent studies on the formation of meat- like aroma compounds. In "Flavor Science and Technology", Y. Biseeiere and A. F. Thomas (Ed), Wiley, New York, NY, pp 113-116. Farmer, L. J ., Mottram, D. S. and Whitfield F. B. 1989. Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid. J. Sci. Food Agric. 49: 347-368. Feather, M. S. and Harris, J. F. 1973. Dehydration reaction of carbohydrates, Adv. Carbohyd. Chem. Biochem. 28: 161-224. Fooladi, M. H., Pearson, A. M., Coleman, T. H. and Merkel, R. A. 1979. The role of nitrite in preventing development of warrned-over flavor. Food Chem. 4: 293-292. Fors, S. 1983. Sensory properties of volatile Maillard reaction products and related compounds. In "The Maillard Reaction in Food and Nutrition", G. R. Waller and M. S. Feather (Ed), ACS Symposium Series 215, Washington, DC, pp 185-286. Frankel, E. N. 1991. Recent advances in lipid oxidation. J. Sci. Food Agric. 54(4): 495-51 1. Frankel, E. N. 1980. Lipid oxidation. Prog. Lipid Res. 19: 1-22. Freybler, L. A., Gray, J. 1., Ashgar, A., Booren, A. M., Pearson, A. M. and Buckley, D. J. 1993. Nitrite stabilization of lipids in cured pork. Meat Sci. 33(1): 85-96. Freybler, L. 1989. Antioxidant Role of Nitrite in Meat Systems. MS. Thesis, Michigan State University, East Lansing, Michigan. Goutefongea, R. Cassens, R. G. and Woolford, G. 1977. Distribution of sodium nitrite in adipose tissue during curing. J. Food Sci. 42: 1637. 61 Gray, J. I. and Pearson, A. M. 1994. Lipid-derived off-flavors in meat - formation and inhibition. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 116-143. Gray, J. I. and Crackel, R. L. 1992. Oxidative flavor changes in meats: their origin and prevention. In "The Chemistry of Muscle-based Food", D. E. Johnson, M. E. Knight and D. A. Ledward (Ed), Royal Society of Chemistry, Cambrige, UK, PP 145-168. Gray, J. I. and Pearson, A. M. 1984. Cured meat flavor. Adv. Food Res. pp 1-86. Gray, J. I., MacDonald, B., Pearson, A. M. and Morton, I. D. 1981. Role of nitrite in cured meat flavor. J. Food Prot. 44(4): 302-312. Gfintert, M., Bri'mug, R. E., Kopsel, M., Kuhn, W., Thielrnann, T. and Werkhoff, P. 1990. Identification and formation of some selected sulfur-containing flavor compounds in various meat model systems. J. Agric. Food Sci. 38(11): 2027-2041. Hartrnen, G. J., Charlin, J. T., Hwang, S. S., Bao, Y., Tang, J. and Ho, C.-T. 1984. Identification of 3,5-diisobutyl-l, 2, 4-trithiolane and 2-isobutyl-3, 5- diisopropylpyridine in fiied chicken flavor. J. Food Sci. 49: 1398-1400. Herz, K. O. and Chang, S. S. 1970. Meat flavor. Adv. Food Res. 18: 1-83. Ho, C.-T. and Chen, Q. 1994. Lipids in food flavors. In "Lipids in Food Flavors", C.-T. Ho and T. G. Hartman (Ed), ACS Symposium Series 558, ACS, Washington, D. C., pp 2-14. Ho, C.-T., Bruechert, L. J ., Zhang, Y. and Chiu, E.-M. 1989. Contribution of lipids to the formation of heterocyclic compounds in model system. In "Thermal Generation of Aromas" , T. H. Parliment, R. J. McGonin and C- T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 104. Ho, C.-T., Lee, K. N. and Jin, Q. Z. 1983. Isolation and identification of flavor compounds in fried bacon. J. Agric. Food Chem. 31: 336-342. Hodge, J. E., Mills, J. D. and Fischer, B. E. 1972. Compounds of browned flavor derived from sugar-amine reactions. Cereal Sci. Today 17: 34-38. 62 Hodge, J. E. 1967. Origin of flavor in foods. Nonenzymatic browning reactions. In "Chemistry and Physiology of Flavors", H. W. Schultz, E. A. Day and L. M. Libbey (Ed), AVI Publishing Co., Westport, CT, pp 465-491. Hodge, J. E. 1953. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1: 928-943. Homstein, I. and Wasserman, A. 1987. Chemistry of meat flavor. In "The Science of Meat and Meat Products", J. F. Price and B. S. Schweigert (Ed). Third Edition, Food and Nutrition Press, Inc., Westport, CT, pp 329-347. Homstein, I and Crowe, P. F. 1963. Meat flavor: lamb. J. Agric. Food Chem. 11: 147-149. Homstein, I and Crowe, P. F. 1960. Flavor study on beef and pork. J. Agric. Food Chem. 8: 494-498. Hotchkiss, J. H., and Gassens, R. G. 1987. Nitrate, nitrite and nitroso compounds in foods. Food Technol. 41(4): 127-131. Hurrell, R. F. 1982. Maillard reaction flavor. In "Food Flavour", I. D. Morton and A. J. MacLeod (Ed), Elsevier Scientific Publishing Co., London, UK, pp 399-437. Hurrell, R. F. and Carpenter, J. K 1974. Mechanism of heat damage in protein. 4. The reactive lysine content of heat-damaged material as measured in different ways. Br. J. Nutrition 32: 589-604. Hwang, S. S., Carlin, J. T., Bao, Y., Hartman, G. J. and Ho, C. T. 1986. Characterization of volatile compounds generated from the reactions of aldehydes with ammonia sulfide. J. Agric. Food Chem., 34: 538-542. Igene, J. 0., Yarnauchi, K., Pearson, A. M., Gray, J. I. and Aust, S. D. 1985. Mechanisms by which nitrite inhibits the development of warmed-over flavor (W OF) in cured meat. Food Chem. 18: 1-18. Kanner, J ., Harel, S., Shagalovich, J. and Berman, S. 1984. Antioxidant effect of nitrite in cured meat products: Nitric oxide-iron complexes of low molecular weight. J. Agric. Food Chem. 32: 512-515. 63 Kanner, J ., Ben-Gera, I. and Berman, S. 1980. Nitric-oxide myoglobin as an inhibitor of lipid oxidation. Lipids 15(11): 944-948. Kanner, J. 1979. S-Nitrosocysteine, an effective antioxidant in cured meats. JAOCS 56: 74-76. Katz, 1. 1981. Recent progress in some aspects of meat flavor chemistry. In "Flavor Research: Recent Advances", R. Teranishi, R. A. Flath and H. Sugisawa (Ed), Marcel Dekker, Inc., New York, pp 217-230. Knoch, E. and Baltes, W. 1992. Model reactions of roast aroma formation: X. Amino acid - specific products after roasting of tryptophan with reducing sugars and sugar degradation products. Food Chem. 44(4): 243-250. Kobayashi, N. and Fujimaki, M. 1965. Formation of mercaptoacetaldehyde, hydrogen sulfide and acetaldehyde by boiling cysteine and carbonyl compounds. Agric. Biol. Chem. 29: 698-699. Kramlich, W. E., Pearson, A. M. and Tauber, F. W. 1973. “Processed Meats”, AVI Publishing Co., Westport, Connecticut. Kwon, T.-W., Menzel, D. B. and Olcott, H. S. 1965. Reactivity of malonaldehyde with food constituents. J. Food Sci. 30: 808-813. Labuza, T. P. and Baisier, W. M. 1992. The kinetics of nonenzymatic browning. In "Physical Chemistry of Foods", H. G. Schwartzenburg and R. W. Hartel (Ed), Marcel Dekker, Inc., New York, pp 595-649. Labuza, T. P., Tannenbaunr, S. R. and Karel, M. 1970. Water content and stability of low-moisture and intennediate-moisture foods. Food Technol. 24: 35-42. Lawrie, R. A. 1974. Meat Science. Pergamon Press, Oxford, UK. Leahy, M. M. and Reineccius, G. A. 1989. Kinetics of the formation of alkylpyrazine, effect of pH and water activity. In "Thermal Generation of Aromas" T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 196-208. 64 Ledl, F. 1990. Chemical pathways of the Maillard reaction. In "the Maillard Reaction in Food Processing, Human Nutrition and Physiology, P. A. Finot, H. U. Aeschbacher, R. F. Hurrell and R. Liardon (Ed), Birkhauser, Switzerland, pp 19-42. Ledl, F. and Severin, T. 1973. Thermische zersetzung von cystein und xylose in tributyrin. Chem. Mikrobiol. Technol. Lebensm. 2: 155. Lee, M. L. 1990. Meat Flavor Generation in Extruded Soy Protein Products. Ph.D. Dissertation, Department of Food Science and Human Nutrition, Michigan State University, East Lansing, Michigan. Lewis, V. M. and Lea, C. H. 1950. A note of the relative rates of reaction of several reducing sugars and sugar derivatives with casein. Biochem. Biophys. Acta. 4: 532-534. Liu, R. H., Conboy, J. J. and Hotchkiss, J. H. 1988. Nitrosation by nitro-nitrosite derivatives of olefins: a potential mechanism for N-nitrosamine formation in fiied bacon. J. Agric. Food Chem. 36: 984-987. MacDonald, B., Gray, J. I., Kukuda, Y. and Lee, M. L. 1980. Role of nitrite in cured meat flavor: 11. Chemical analyses. J. Food Sci. 45(4): 889-892. MacDougall, D. B, Mottram, D. S. and Rhodes, D. N. 1975, Contribution of nitrite and nitrate to the colour and flavour of cured meat. J. Sci. Food Agric. 26: 1743-1754. Macey, R. L. Jr., Naumann, N. D. and Bailey, M. E. 1964a. Water-soluble flavor and odor precursors of meat. 1. Quantitative study of certain amino acids, carbohydrates, non-amino nitrogen compounds and phosphoric acid esters of beef, pork, and lamb. J. Food Sci. 29: 136-141. Macey, R. L. Jr., Naumann, N. D. and Bailey, M. E. 1964b. Water-soluble flavor and odor precursors of meat. I. Effect of heating on amino acids, nitrogen constituents and carbohydrate in lypophilized diffusates fiom aqueous extracts of beef, pork, and lamb. J. Food Sci. 29: 142-148. MacLeod, G. and Ames, J. M. 1987. Effect of water on the production of cooked beef aroma compounds. J. Food Sci. 52(1):42-45, 56. 65 MacLeod, G. 1986. The scientific and technological basis of meat flavours. In "Developments in Food Flavours", G. G. Birch and M. G. Lindley (Ed), Elservier Applied Science, London, UK, pp 191-223. MacLeod, G. and Ames, J. M. 1986. Capillary gas chromatography - mass spectrometric analysis of cooked ground beef aroma. J. Food Sci. 51(6): 1427-1434. MacLeod, G. and Seyyedian-Ardebili, M. 1981. Natural and simulated meat flavor. CRC Crit. Rev. Food Sci. Nutr. 14: 309-437. MacLeod, G. and Coppock, B. M. 1977 . A comparison of the chemical composition of boiled and roast aromas of heat beef. J. Agric. Food Chem. 25(1): 113-117. Maga, J. A. 1992. Pyrazine update. Food Reviews International 8(4): 479-558. Maga, J. A. 1975. The role of sulfur compounds in food flavor, Part I. Thiazole. CRC Crit. Rev. Food Sci. Nutr. 6: 153-176. Maillard, L. C. 1912. Action des acides amines sur les sucres: formation des melanoidines par voie methodique. C. R. Nat. Acad, Paris, 154: 66-68. Mauron, J. 1981. The Maillard reaction in food, a critical review from the nutrition standard. Prog. Food Nutr. Sci. 5: 5-35. Meynier, A. and Mottram, D. S. 1992. The effect of lipid on the Maillard reaction between methionine and ribose: a model system to investigate the formation of meat flavour, 38th International Conference of Meat Science and Technology (ICoMST), Clermont-F errand, France, pp 535-538. Montgomery, M. W. and Day, E. A. 1965. Aldehyde-amine condensation reaction: a possible fate of carbonyls in food. J. Food Sci. 30: 828-832. Monahan, F. J ., Gray, J. I., Asghar, A., Haug, A., Shi, B. and Buckley, D. J. 1993. Effect of dietary lipid and vitamin E supplementation on fiee radical production and lipid oxidation in porcine muscle rrricrosomal fractions. Food Chem. 46 (1): 1-6. 66 Monahan, F. J. 1992. Dietary Factors Affecting Lipid Oxidation in Pork. Ph.D. Dissertation, Department of Food Technology, University College Cork, Ireland. Montgomery, M. W. and Day, E. A. 1965. Aldehyde-amine condensation reaction: a possible fate of carbonyls in foods. J. Food Sci. 30: 828. Monissey, P. A. and Tichivangana, J. Z. 1985. The antioxidant activities of nitrite and nitrosyhnyoglobin in cooked meats. Meat Sci. 14(3): 175-190. Mottrarrr, D. S., Madruga, M. S. and Whitfield, F. B. 1995. Some novel meatlike aroma compounds from the reactions of alkanediones with hydrogen sulfide and firranthiols. J. Agric. Food Chem. 43(1): 189-193. Mottram, D. S. and Whitfield, F. B. 1995. Volatile compounds form the reaction of cysteine, ribose and phospholipid in low-moisture systems. J. Agric. Food Chem, 43(4): 984-988. Mottram, D. S. 1994a. Flavor compounds formed during the Maillard reaction. In "Thermal Generated Flavor: Maillard, Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington DC, pp 104-126. Mottranr, D. S. 1994b. Some aspects of the chemistry of meat flavor. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 210-230. Mottram, D. S. and Madruga, M. S. 1994. Important sulfur-containing aroma volatiles in meat. In "Sulfur Compounds in F oods", C. J. Mussinan and M. E. Keelan (Ed), ACS Symposium Series 564, Washington DC, pp 180- 187. Mottram, D. S. and Whitfield, F. B. 1994. Aroma volatiles from Maillard systems, in "Thermal Generated Flavor: Maillard, Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington DC, pp 180-191. Mottranr, D. S. 1991. Meat. In "Volatile Compounds in Foods and Beverages", H. Maarse (Ed), Marcel Dekker, Inc., New York, pp 107-177. 67 Mottram, D. S. 1990. The effect of pH on the formation of aroma volatiles in meat-like Maillard systems. In "Flavor Science and Technology", Y. Bessiére and A. F. Thomas (Ed), Wiley Chichester, London, UK, pp 121- 124. Mottram, D. S. and Salter, L. J. 1989. Flavor formation in meat-related Maillard systems containing phospholipids. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGonin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 442-451. Mottram, D. S. 1987. Lipid oxidation and flavor in meat and meat products. Food Sci. Technol. Today 1(3): 159-162. Mottrarrr, D. S. and Whitfield, F. B. 1987. The effect of lipid on the formation of volatile heterocyclic compounds in the Maillard reaction. In "Flavour Science and Technology", M. Martens, G. A. Daien and H. Russwurm Jr. (Ed), John Wiley & Sons Ltd, New York, pp 29-34. Mottram, D. S. 1985a. Some new observations of the flavour chemistry of cured meats. In "Progress in Flavor Research", J. Adda (Ed), Elsevier, Amsterdam, the Netherlands, pp 323-328. Mottram, D. S. 1985b. The effect of cooking conditions on the formation of volatile heterocyclic compounds in pork. J. Sci. Food Agric. 36: 377-382. Mottram, D. S. 1984. Organic nitrates and nitriles in the volatiles of cooked cured pork. J. Agric. Food Chem. 32(2): 343-345. Mottram, D. S., Croft, S. E. and Patterson, R. L. S. 1984. Volatile components of cured and uncured pork: the role of nitrite and the formation of nitrogen compounds. J. Agric. Food Chem. 35: 233-239. Mottram, D. S. and Edwards, R. A. 1983. The role of triglycerides and phospholipids in the aroma of cooked beef. J. Sci. Food Agric. 34: 517-522. Mottrarrr, D. S., Edwards, R. A. and MacFie, H. J. H. 1982. A comparison of the flavour volatiles for cooked beef and pork meat system. J. Sci. Food Agric. 33: 934-944. 68 Mottram, D. S. and Puckey, D. J. 1978. Artifact formation during the extraction of bacon volatiles in a Liken-Nickerson apparatus, Chem. Ind. (London), 3 June, pp 385-386. Mottranr, D. S. and Rhodes, D. N. 1973. Nitrite and the flavour of cured meat 1, Proc. Int. Symp. Nitrite Meat Prod. Zeist, pp 161-171. Neol, P., Briand, E. and Dumont, J. P. 1990. Role of nitrite in flavor development in uncooked cured meat products: sensory assessment. Meat Sci. 28: 1-8. Nursten, H. E. 1986. Aroma compounds from the Maillard reaction. In "Developments in Food Flavours", G. G. Birch and M. G. Lindley (Ed), Elservier Applied Science, London, pp 173-190. Nursten, H. E. 1981. Workshop on volatile products. Prog. Food Nutr. Sci. 5: 491- 496. O'Brien, J. and Morrissey, P. A. 1989. Nutritional and toxicological aspects of the Maillard browning reaction in foods. CRC Crit. Rev. Food Sci. Nutr. 28(3): 21 1-243. Paquette, G., Kupranycz, D. B., and van de Voort, F. R. 1985. The mechanism of lipid autoxidation. 1. Primary oxidation products. Can. Inst. Food Sci. Technol. J. 18: 112-118. Pearson, A. M., Love, J. D. and Shorland, F. B. 1977 . Warmer-over flavor in meat, poultry and fish. Adv. Food Res. 23: 1-74. Peterson, B. 1., Tong, C.-H., Ho, C.-T. and Welt, B. A. 1994. Effect of moisture content on Maillard browning kinetics of a model system during microwave heating. J. Agric. Food Chem. 42(9): 1884-1887. Pippen, E. L. and Mecchi, E. P. 1969. Hydrogen sulfide, a direct and potentially indirect contributor to cooked chicken aroma. J. Agric. Food Chem. 34: 443-446. Price, J. F. and Schweigert, B. S., 1987. “The Science of Meat and Meat Products”. Third Edition, Food & Nutrition Press, Inc., Westport, CT. 69 Ramarathnam, N. and Rubin, L. J. 1994. The flavor of cured meat. In "Flavor of Meat and Meat Products", F. Shahidi, (Ed), Blackie Academic and Professional, Glasgow, UK, pp 174-198. Ramarathnam, N., Rubin, L. J. and Diosady, L. L. 1993a. Studies on meat flavor. IH. A novel method for t rapping volatiles components from uncured and cured pork. J. Agric. Food Chem. 41(6): 933-938. Ramarathnam, N., Rubin, L. J. and Diosady, L. L. 1993b. Studies on meat flavor. IV. Fractionation, characterization, and quantitation of volatiles from uncured and cured beef and chicken. J. Agric. Food Chem. 41(6): 939-945. Ramarathnam, N., Rubin, L. J. and Diosady, L. 1991a. Studies on meat flavor: I. Qualitative and quantitative differences in uncured and cured pork. J. Agric. Food Chem. 39: 344-350. Ramarathnam, N., Rubin, L. J. and Diosady, L. 1991b. Studies on meat flavor: 11. Quantitative approach to the estimation and differentiation of volatiles in uncured and cured beef and chicken. J. Agric. Food Chem. 39: 1839-1847. Rizzi, G. P. 1987. New aspects on the mechanism of pyrazine formation in the Strecker degradation of amino acids. In "Flavour Science and Technology", M. Martens, G. A. Dalen and H. Russwurm Jr. (Ed), John Wiley and Sons, Ltd, New York, pp 23-28. Rubin, L. J. and Shahidi, F. 1988. Lipid oxidation and the flavor of meat products. Proceedings, 34th International Congress of Meat Science Technology, Brisbane, Australia, pp 295-301. Ruther, J. and Baltes, W. 1994. Sulfur-containing furans in commercial meat flavorings. J. Agric. Food Chem. 42(10): 2254-2259. Saittagaroon, S., Kawakishi, S. and Nanriki, M. 1984. Aroma constituents of roasted coconut. Agric. Biol. Chem. 48(9): 2301-2307. Salter, L. J ., Mottram, D. S. and Whitfield, F. B. 1988. Volatiles compounds produced in Maillard reactions involving glycine, ribose and phospholipid. J. Sci. Food Agric. 46: 227-242. 7O Sanders, A, Pearson, A. M. and Schweigert, B. S. 1966. Effect of cooking procedures on flavor components of beef - carbonyl compounds. J. Agric. Food Chem. 14:256. Sato, K. and Hegarty. G. R. 1971. Warmer-over flavor in cooked meats. J. Food Sci. 36: 1098-1102. Shahidi, F. 1989. Flavor of cooked meats. In "Thermal Generation of Aromas", T. H. Parliament, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 188-201. Shahidi, F ., Rubin, L. J. and D'Souza, L. A. 1986. Meat flavor volatiles: A review of the composition, techniques of analysis and sensory evaluation. CRC Crit. Rev. Food Sci. Nutr. 24(2): 141-243. Shaw, J. J. and Ho, C.-T. 1989. Effect of temperature, pH and relative concentration on the reaction of rharnnose and proline. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 217-228. Shibamoto, T. and Bernhard, R. A. 1977. Investigation of pyrazines in sugar- ammonia model systems. J. Agric. Food Chem. 25(3): 609-614. Shu, C.-K. and Ho, C.-T. 1989. Parameter effects on the thermal reaction of cystine and 2,5-dimethyl-4-hydroxy-3(2H)—furanone. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 229-241. Shu, C.-K., Hagedorn, M. L., Mookherjee, B. D. and Ho, C.-T. 1985. pH Efl‘ect on the volatile components in the thermal degradation of cysteine. J. Agric. Food Chem. 33(3): 442-446. Sink, J. D. 1979. Factors influencing the flavor of muscle foods. J. Food Sci. 44(1): 1-5. Sink, J. D. 1973. Lipid-soluble components of meat flavors/odors and their biochemical origin. JAOCS. 50: 470. Skog, K. 1993. Cooking procedures and food mutagens, a literature review. Food Chem. Toxicol. 31(9): 655-675. 71 Spanier, A. M. and Drurn-Boylston, T. 1994. Effect of temperature on the analysis of beef volatiles: focus on carbonyl and sulfur-containing compounds. Food Chem. 50(3); 251-259. Swain, J. W. 1972. Volatile Flavor Constituent of Pork Cured with and without Nitrite. Ph.D. dissertation, Department of Food Science, University of Missouri, Columbia, Missouri. Takken, H. J ., van der Linde, L. M., de Volois, P. J ., van Dort, J. M., Boelens, M. 1976. Reaction product of a-dicarbonyl compounds, aldehydes, hydrogen sulfide and ammonia. In "Phenolic, Sulfur and Nitrogen Compounds in Food F lavor", G. Charalambous and I. Katz (Ed), American Chemistry Society, Washington, DC, pp 114-121. Tang, J ., Jin, Q. Z., Sherr, G. H., Ho, C.-T. and Chang, S. S. 1983. Isolation and identification of volatiles compounds form fried chicken. J. Agric. Food Chem. 31: 1287-1292. Tressl, R., Helak, B., Kersten, E. and Nittka, C. 1993. Formation of flavor compounds by Maillard reaction. In "Recent Developments in Flavor and Fragrance Chemistry", R. Hopp and K. Mori (Ed), VCH, New York, pp 161-181. Umano, K. and Shibamoto, T. 1987. Analysis of headspace volatiles from overheated beef fat. J. Agric. Food Chem. 35(1): 14-18. van der Ouweland, G. A. M., Peer, H. G. and Tjan, S. B. 1978. Occurrence of Amadori and Heynes rearrangement products in processed foods and their role in flavor formation. In “Flavor of Foods and Beverages”, Charalambous, C. and Inglett, G. E. (Ed), Academic Press, New York, pp 131. van der Ouweland, G. A. M. and Peer, H. G. 1975. Components contributing to beef flavor: volatile compounds produced by the reaction of 4-hydroxy-5- methyl-3(2H)-furanone and its thiol analog with hydrogen sulfide. J. Agric. Food Chem. 23: 501-505. Vercellotti, J. R., St. Angelo, A. J. and Spanier, A. M. 1992. Lipid oxidation in foods. In "Lipid Oxidation in Food", A. J. St. Angelo (Ed), ACS Symposium Series 500, Washington DC, pp 1-11. 72 Vercellotti, J. R., Kuan, J. W., Liu, R. H., Legendre, M. G., St. Angelo, A. J. and Dupuy, H. P. 1987. Analysis of volatile heteroatomic meat flavor principles by purge-and-trap gas chromatography-mass spectrometry. J. Agric. Food Chem. 35: 1030-1035. Vemin, G., Metzger, J ., Sultan, A. A. M., El-Shafei, A. K. and Parkanyi, C. 1993. Heterocyclic compounds in model systems for the Maillard reaction. In "Food Flavor and Safety", A. M. Spanier, H. Okai and M. Tamura, Eds, ACS Symposium Series 528, Washington, DC, pp 36-55. Vemin, G. and Parkanyi, C. 1982. Mechanisms of formation of heterocyclic compounds in Maillard and pyrolysis reaction. In "Cherrristry of Heterocyclic Compounds in Flavor and Aromas", G. Vemin (Ed), Ellis Horwood, Chichester, UK, pp 151-207. Warmbier, H. C., Schnickels, R. A. and Labuza, T. P. 1976. Nonenzymatic browning kinetics in an intermediate moisture model system: effect of glucose to lysine ratio. J. Food Sci. 41: 981-983. Wasserman, A. E. 1979. Symposium of meat flavor chemical basis for meat flavor: a review. J. Food Sci. 44(1): 6-11. Wasserman, A. E. and Spinelli, A. M. 1972. Effect of some water-soluble components on aroma of heated adipose tissue. J. Agric. Food Chem. 20: 171-174. Weenen, H., Tjan, S. B., de Valois, P. J ., Pos, A. and Vonk, H. 1994. Mechanism of pyrazine formation. In "Thermal Generated Flavor: Maillard, Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. MacGorrin (Ed), ACS Symposium Series 543, pp 142-157. Werkhofi, P., Brfining, J ., Emberger, R, Giintert, M. and Hopp, R. 1993. Flavor chemistry of meat volatiles: new results on flavor components from beef, pork and chicken. In "Recent Developments in Flavor and Fragrance Chemistry", R. Hopp and K. Mori (Ed), VCH, New York, pp 183-213. Whitfield, F. B. 1992. Volatiles for interactions of Maillard reactions and lipids. CRC Crit. Rev. Food Sci. Nutr. 31(1/2): 1-58. Whitfield F. B., Mottrarrr, D. 8., Brock, S., Puckey, D. J. and Salter, L. J. 1988. Effect of phospholipid on the formation of volatile heterocyclic compounds 73 in heated aqueous solutions of amino acids and ribose. J. Sci. Food Agric. 42: 261-272. Zhang, Y., Ritter, W. J ., Barker, C. C., Traci, P. A. and Ho, C.-T. 1994. Volatile formation by lipid-mediated Maillard reaction in model system. In "Lipids in Food Flavors", C.-T. Ho and T. G. Hartman (Ed), ACS Symposium Series 558, ACS, Washington, D. C., pp 2-14. Zhang, Y. and Ho. C.-T. 1989. Volatile compounds from thermal interaction of 2,4-decadienal with cysteine and glutathione. J. Agric. Food Chem. 37: 1016-1020. Zipser, M. W., Kwon, T. W. and Watts, B. M. 1964. Oxidative changes in cured and uncured frozen cooked pork. J. Agric. Food Chem. 12: 105-109. (SHAPED I ANALYSIS OF VOLATILES PRODUCED FROM THE MAILLARD REACTION BETWEEN CYSTEINE AND RIBOSE ABSTRACT A headspace collection technique was developed in this study. Volatile compounds produced in a Maillard model system containing cysteine and ribose were collected in a Tenax trap and then desorbed into ethyl acetate by a combination of heat and solvent elution. Percent recoveries of the internal standards, butyl acetate, methyl docanoate, thiazole, furfuryl mercaptan and l-(2- thienyl)-propanone were 90%, 82%, 89%, 85% and 72%, respectively. More than 90 compounds were isolated, thirty-one of which were identified via Kovats indices and mass spectrometry. The compounds identified included thiophenes (8), thiophenone (1), thiopyran (bi-heterocyclic compound) (1), bicyclic thiophenes (3), thiazole (1), disulfides (4), heterocyclic thiols (3), thiolanone (1), mercapto carbonyls (2), methylthio acetates (2), and non-sulfur-containing compounds such as furan derivatives (3) and sugar-derived carbonyls (2). Fifteen of these compounds were selected as markers in subsequent studies to evaluate the effect of lipids on the Maillard reaction. This selection was based on previous literature findings fiom model systems involving lipids and their contribution to meat flavor. 74 INTRODUCTION The Maillard reaction is one of the most important routes for the generation of flavor compounds in cooked foods. It is responsible for the production of many heterocyclic compounds with distinctive aromas and low thresholds (Mottrarrr, 1994a). The Maillard reaction can be conveniently divided into three stages. The initial reaction is the condensation of an amino acid and a reducing sugar to form a glycosylamine which then undergoes the Amadori rearrangement. This is followed by the dehydration of intermediate endiol structures to form furan derivatives and dicarbonyl compounds. The final stage involves the conversion of the furan and dicarbonyl intermediates into aroma compounds by reaction with other intermediates such as amino compounds or Strecker degradation products (Baltes e_t al., 1989; Mottram 1994a). More than 1000 volatile compounds developed in the Maillard reaction have been isolated. These aroma compounds have been classified into three groups based on the origin of the complex mixture of volatiles (Nursten 1981; 1986). 1. Compounds derivedfrom "simple " sugar dehydration /frag2nentation. The predominant products consist of furans and furanones, pyranones, cyclopentenones and cyclopentanones with or without hydroxyl groups, and some aliphatic carbonyl compounds. Many of these compounds possess aroma and may contribute to food flavor. However, their roles are more important in serving as precursors for other flavor compounds. 75 76 H. Compounds floducedfrom "simple" amino acid degradation. This group comprises simple aldehydes, hydrogen sulfide or amino compounds which are formed via the Strecker degradation. 1H. Volatiles compounds generated byfurther interactions. Pyrroles, pyridines, pyrazines, imidazoles, oxazoles, thiazoles, thiophenes, di- and trithiolanes, di- and trithianes, furanthiols and compounds from aldol condensations are primary products in this group. These compounds are important to the flavor of meat products, and are particular, the sulfur- containing heterocyclic compounds (Mottram 1991; 1994a, b). Model systems containing amino aicds and reducing sugars have been investigated extensively to determine the major compounds contributing to meat flavor (Whitfield 1992). In an early attempt to synthesize meat flavor, cysteine, and amino acids or hydrolyzed proteins were heated with a pentose or hexose (Morton _e_t fl, 1960). Farmer and Mottram (1990a) established a model system containing cysteine and ribose to study the effect of three phospholipids and a triacylglycerol on the formation of Maillard reaction volatiles. The Maillard reaction products were collected by a purge-and-trap method, and the trapped compounds were then thermally desorbed into a gas chromatograph / mass spectrometer (GC / MS) system via a "Unijector" inject port. Solid carbon dioxide was also used to condense the volatile compounds in the front end of the GC column. Sixty volatile compounds were identified and used to demonstrate the effect of lipids on aroma production. The purpose of this study was to develop a dynamic headspace technique to collect the Maillard reaction volatiles, and to isolate and identify these compounds using gas chromatography / mass spectrometry. Another objective was to identify 77 some Maillard reaction products to be used as "marker" compounds for the subsequent investigation of the role of lipids in modifing Maillard reaction flavors. EXPERIMENTAL Materials Cysteine and ribose, straight chain alkanes (Cg - C24), butyl acetate and methyl decanoate were purchased from Sigma Chemical Co. (St. Louis, MO). Ethyl acetate, a solvent for the extraction of volatile compounds, and disodium hydrogen phosphate and sodium dihydrogen phosphate for phosphate buffer preparation, were obtained from J. T. Baker Chemical Co. (Phillipsburg, NJ). Tenax GC (60/80 mesh) was obtained from Alltech Associates Inc. (Deerfield, IL). A “flavors and fragrances kit” containing 24 heterocyclic compounds (Appendix A) and three sulfur-compounds (1-(2-thienyl)-propanone, 2-methyl thiophene, and furfuryl mercaptan) were used as standard compounds, and purchased from Aldrich Chemical Co. (Milwaukee, WI). The reagents and solvents used in this study were analytical and/or HPLC grade. Preparation of Maillard reaction mixtures and collection of volatile compgyn_ds Cysteine (0.61 g) and ribose (0.75 g) were dissolved in 2 ml phosphate buffer (0.5M, pH 5.7), and placed in 10 ml clear glass vials (25 mm dia x 54 mm length, mouth 20 mm o.d. x 13 mm id) (Pierce Chemical Co., Rockford, IL). The vials were flushed with nitrogen for 2 min before sealing with Tuf-bond Teflon- silicon discs and aluminum caps (Pierce Chemical Co., Rockford, IL). The reaction mixtures were heated at 140°C in an autoclave under a pressure of 0.28 78 MPa for l h. After heating, the vials containing the Maillard reaction products were stored in a freezer (-10°C) prior to volatile collection. Tenax traps were prepared by packing 200 mg of the Tenax GC porous polymer into a glass tube (5 mm id. x 70 mm length x 1 mm thick, custom made in the Glass Shop, Department of Chemistry, Michigan State University) and plugging both ends with glass wool. The Tenax traps were conditioned overnight in an oven at 180°C, with a nitrogen flow (30 ml / min) prior to volatile collection. For the collection of the volatile compounds, the contents of the reaction vials (ca. 2 ml) were transferred to a 500-ml two-neck flask (Kontes, Vineland, NJ) containing 18 ml phosphate buffer (pH 5.7) (Figure 1). The flask was fitted with a L-shape glass joint (Kontes, Vineland, NJ), and the Tenax trap was attached by a stainless steel Swagelok reducing union (Cajon Ultra-Torr® fitting, Cajon, Macedonia, OH) to the outlet of the joint. During collection, the flask was heated at 60°C using a heating mantle, and the Tenax trap was maintained at room temperature (approximately, 20°C 3: 2°C). The volatile compounds were swept into the Tenax trap using a stream of nitrogen (30 ml /min), and the collection was continued for 1 h. The flask was then removed and the Tenax trap directly connected to the nitrogen supply for 5 min to remove water vapor (Whitfield e_t a1 ., 1988; Farmer _e_t _a_1., 1990a). Desogption of trapped volatile compo_ur£ls from the Tenax trap A new desorption technique was developed in this study which involved the simultaneous use of heat and solvent extraction. This method utilized the properties of the Tenax polymer whose affinity for volatile compounds decreases at high temperatures. Solvent extraction was also used to elute the trapped compounds as it has been reported that there is no decomposition of Tenax in the 79 Tenax trap 123E“: Swagelok reducing union 3/8 ' - 1/4' flow gauge / ’ Teflon tube N2 two—neck flask (500 ml) 2 ml reaction mixture + 18 ml phosphate buffer heating mantle (60°C) Figure 1. Apparatus used for the collection of the volatile compounds generated by heating cysteine and ribose at 140°C for l h. 80 absence of heat (Toyoda e_t Q, 1993). In this study, the desorption process involved heating at 150°C (which is approximately 100°C lower than the temperature usually employed in commercial thermal desorption equipment) with simultaneous flushing of the Tenax trap with a stream of nitrogen saturated with solvent (ethyl acetate). The released compounds were then collected in a cold trap. During desorption, the Tenax trap was connected between a solvent (ethyl acetate) source and a cold trap (Figure 2). The Tenax trap was wrapped with a heating strip and heated at 150°C. The nitrogen was passed through the ethyl acetate reservoir (30 - 40 ml/min), and then the ethyl acetate - nitrogen gas was introduced into the heated Tenax trap. The nitrogen stream was dried by passing through a desiccant (anhydrous calcium sulfate) (W. A. Hammond Drierite Co., Xenia, OH ) before entering the ethyl acetate reservoir. The desorbed flavor compounds were then condensed in a cold trap using dry ice - acetone. The desorption process was continued for 30 min, after which the ethyl acetate extract was concentrated to 0.1 ml under nitrogen. The concentrate was transferred to a screw-cap vial and stored in a refiigerator prior to gas chromatographic (GC)/ mass spectrometric (MS) analysis. To measure the recovery of the trapped compounds, solution of ester compounds (butyl acetate and methyl decanoate, 1 pg each in 1 ml ethanol) and sulfur-containing heterocyclics (thiazole, furfuryl mercaptan and l-(2-thienyl)- propanone, 5 ug each in 1 ml ethanol) were prepared. An aliquot (1 ul) of the internal standard solution was deposited into the trap using a 10 pl syringe (Hamilton, Reno, NV), and the solvent was removed by purging with nitrogen (30 ml / min) for 5 min. The recovery of each internal standard was calculated by comparing the relative GC peak areas of the compounds in the original standard solvent reservoir Figure 2. 81 heating strip Tenax trap dry ice + acetone Desorption of the volatile Maillard reaction compounds fi'om the Tenax trap. The Tenax trap was simultaneously heated at 150°C and extracted by an ethyl acetate-saturated nitrogen flow (30 ml/min) for 30 min. The solvent and desorbed volatile compounds were collected in dry ice - acetone cold trap. 82 solution to those of the compounds recovered from the Tenax traps. To measure the Kovats index (KI) of a compound, straight-chain alkanes (C3 - C24, 15 ng each) were added to the internal standard solution before applying to the Tenax traps. The KI of each compound was calculated using the retention times of the compound and the two adjacent n-alkanes. Gas chromatographic - mass spectrometric analysis (GC / MS) The volatile compounds were separated and identified using a HP 5890 gas chromatograph (Hewlett Packard, Avondale, PA) interfaced with a HP 5970 MSD (mass selective detector) mass spectrometer. The GC / MS system was equipped with a HP 59970 Chemstation Data System. A Carbowax 20M fused silica capillary column (60m x 0.32 mm id, Supelco Inc., Bellefonte, PA) was used for the separation of the volatile compounds. The GC oven temperature was programmed as follows: initial temperature, 40°C; initial time, 5 min; rate 2°C / min; final temperature, 200°C; final time, 10 min. Helium was used as a carrier gas with a flow rate of 10 ml/min. The temperatures of the injection port and transfer lines were both set at 220°C. The MS was operated in the electron impact mode with an electron energy of 70 eV and an ion source temperature of 250°C. Compounds were introduced to the ion source directly from the capillary column in the GC using an open-split interface. A continuous scan mode with a scan time of 1 see over a mass range of 40-3 00 was employed. The GC / MS data were monitored, stored and analyzed using a HP Chemstation data system. The Maillard reaction products were identified by comparing their mass spectra and Kl’s to those of authentic standard compounds. Two mass spectral libraries were used in the study. One was the INRA Mass Spectra Library - 83 "INRAMASS", - a computer program created by R. Ahnanza (Laboratoire de Recherche sue les Aromes, Dijon, France) and donated by Professor D. S. Mottram Wniversity of Reading, Reading , UK). The other was the NIST / EPA / MSDC Mass Spectral Database purchased from the ACS Publication Co., (Washington, DC). Where reference compounds were not available, identifications were made on the basis of mass spectral data. RESULTS AND DISCUSSION Development of ar_ralytical procedures for the identification of volatile compounds in a heated Maillard reaction system A new desorption technique was developed by combining thermal desorption and solvent extraction. The desorption efficiency of the method was determined by measuring the recovery of internal standards. For ester compounds, the recoveries were 90 i 8% and 82 :l: 6% for butyl acetate and methyl decanoate, respectively. Recoveries of sulfur-containing compounds were 89 i 5% , 85 :l: 4% and 72 i 7% for thaizole, furfuryl mercaptan and 1-(2-thienyl)-propanone, respectively. The difference in the recoveries obtained may be due to the relative affinities of the compounds for the Tenax absorbent. The volatile Maillard reaction compounds were collected by a purge-and- trap method using Tenax trap as the absorbent (Whitfield at g, 1988). As the flavor/aroma compounds in food products are often present at concentrations too small to permit direct analysis by GC, a preconcentration step is normally required. Tenax (p-2,6-diphenyl-p-phenylene oxide), a porous polymer absorbent, has been widely used to concentrate headspace volatiles because of its sensitivity 84 and mobility (Toyoda e_t a_l., 1993; MacCaffrey e_t al_., 1994). This absorbent has been reported to have low affinities for organic compounds at high temperatures (Buchholz at al_., 1980; Barnes _e_t_ fl, 1981). Therefore, thermal desorption is routinely used to release the absorbed volatile compounds. A direct-inject type of thermal desorption system requires substantial heating (at temperatures in excess of 250°C) to completely desorb the trapped volatile compounds. This process fiequently causes isomerization and decomposition of heat labile compounds (Toyoda e_t fl, 1993). The solvent desorption method described here utilizes a lower temperature in order to minimize such changes. Solvent extraction, however, produces a dilution effect and results in the non-detection of compounds present in trace amounts. The desorption technique developed in this study combined thermal desorption and solvent extraction, and provided excellent recoveries of the internal standards. Additionally, no Tenax breakdown products (such as phenyl compounds) were observed using the method. Therefore, this new desorption procedure was applied in all subsequent studies. Mtion and identification of Maillard reaction products Cysteine and ribose were used as reactants to produce Maillard reaction products. The choice of cysteine as the Maillard reactant was based on the observation that sulfur-containing compounds are dominant in meat flavor, and that hydrogen sulfide, formed from the Strecker degradation of this amino acid, is essential for the formation of these compounds (Mottram, 1991). More than 90 compounds were isolated from the reaction mixture of cysteine and ribose heated at 140°C for 1h. Thirty-one compounds were identified (Table 1), confirmation of their identities being achieved by Kovats indices (KI) 85 Table 1. Maillard reaction products isolated from a model system of cysteine and ribose heated at 140°C for l h. Compound MW K1 K1 (ref) Relative Confirmationb amount (%)a cyclopentanone 84 1237 2.69 MS thiazole 85 1247 1248 0.78 KI, MSc methylthio] acetate 90 1332 1.67 MS firrfural 96 1453 1455 9.68 KI, MS 2,5-firrandione 98 13 71 0.55 MS 4,5-dihydro-2-methylthiophene 100 l 145 1 142 0.82 KI, MS 2-ethyl butanal 100 1286 4.73 MS dihydro-3(2I-I)-thiophenone 102 1558 1556 0.19 KI, MS 3-thiolanone 102 1592 1.32 MS 2-formyl thiophene 112 1700 1688 0.23 KI, MS firrfuryl mercaptan 114 1423 1426 18.82 KI, MSc 2-methyl-3-firranthiol 114 1296 1295 13.82 KI, MS tetrahydro-4H-thiopyran-4-one 116 1705 0.82 MS 3-thiophenethiol 1 16 1576 1581 10.32 KI, MS 2-mercapto-3-pentanone 1 18 1481 1.20 MS 3-mercapto-2-pentanone 118 1353 1350 16.47 K1, MS 2-acetyl thiophene 126 1775 1772 0.06 KI, MSc 3-methyl-2-formylthiophene 126 1813 18 13 0. l6 KI, MS 2-ethyl tetrahydrothiophene 130 1654 0.07 MS Table 1. continued. ethyl 2-(methylthio)-acetate 1 -(2-furyl)-3 -butanone 1 -(2-thineyl)-2-propanone dimethyl forrnylthiophene 3-ethyl-2-formylthiophene dihydrothienothiophene methyl dihydrothienothiophene dimethyl dihydrothienothiophene Isulfide compounds 3-[(2-methyl-3-firryl)dithio]-2- butanone bis-(2-methyl-3-furyl) disulfide 2-methyl-3-[(2-firrylmethyl) dithio] firran 3-[(2-methyl-3-firryl) dithio]-2- pentanone 134 138 140 140 140 142 156 170 216 226 226 230 86 1328 1653 1846 1924 1869 2029 2109 2082 2147 2120 2233 2174 1657 1842 1871 2022 2103 2077 2115 3.54 0.29 0.58 0.06 0.20 0.03 0.22 0.01 0.15 0.53 0.05 0.08 MS KI, MS KI, MSc MS KI, MS KI, MS KI, MS KI, MS MS K1, MS MS MS 3 replications. b KI : Kovats index. MS : confirmed by MS using the standard mass spectrum library. c Confirmation was made by comparing the reference compounds. The relative amount was calculated by expressing the peak area of each compound as a percent of the total peak area. Numbers represent the average of 87 and mass spectral data. The relative amount of each compound was calculated by expressing its peak area as a percentage of the total peak area of the isolated compounds. An identical model system was used by Farmer and Mottram (1990a) to study the volatile compounds in meat flavor. Fewer compounds were identified in this study than what was reported by Farmer and Mottram (1990a). These investigators isolated 60 compounds and subsequently used them to compare the effects of beef triacylglycerols, beef phospholipids, egg phosphatidylethanolamine and egg phsphatidylcholine on the production of Maillard reaction volatiles from heated cysteine and ribose. The qualitative differences in the compounds isolated in the two studies is due to the different experimental procedures involved. Farmer and Mottram (1990a) used an "Unijector" injector port linked with a cryogenic trap in the front end of the GC column to concentrate the volatile compounds. When solvent is used in the final extraction (concentration phase), many flavor compounds present in very low concentrations will be diluted. The present study, however, was not intended to repeat the definitive investigations of Farmer and Mottram (1990a), but rather to select "marker" compounds so that the effect on stabilized lipids in the Maillard reaction could be determined. Compounds identified included thiazole, furan derivatives, thiophenes and thiophenones, heterocyclic thiols, mercapto carbonyls, disulfides and nonsulfur- containing compounds (Table l). Sulfur-containing compounds (26) were predominant in the isolate. Twelve thiophenes or their derivatives were confirmed. These compounds are widely present in beef products and many have been reported to have odor threshold values in the pig / kg range (Mottram, 1991). Of the thiophene compounds isolated, an alkylthiophene (4,5-dihydro-2-methyl thiophene), six acylthiophenes (2-formy1 thiophene, 2-acetyl thiophene, 3-methyl- 2-formyl thiophene, 1-(2-thienyl)-2-propanone, a dimethyl formyl thiophene and 88 3-ethyl-2-formyl thiophene), one thiophenone (dihydro-3(2H)-thiophenone) and three bicyclic compounds (dihydrothienothiophene, methyl dihydrothienothiophene and dimethyl dihydrothienothiophene) were reported by Farmer and Mottram (1990a). These compounds were produced in a heated (140°C, 1 h) model system containing cysteine and ribose; however, their formation were noticeably suppressed in the presence of lipids. These investigators assumed that carbonyl compounds produced by the thermal oxidation of lipids may compete with sugar-derived dicarbonyls for free hydrogen sulfide (a Maillard reaction intermediate), thus reducing the production of Maillard reaction volatile compounds. One thiophene compound, 2-ethyl tetradihydrothiophene, was not reported by Farmer and Mottram (1990a). However, this compound is unlikely to play an important role in subsequent studies because it was only present in a relatively small amount (0.07%). Other sulfur-containing compounds identified in this study were one thiazole, one thiolanone, three heterocyclic thiols (fiufuryl mercaptan, 2-methyl-3- furanthiol, 3-thiophenethiol), and four disulfide compounds (Table 1). Though thiazole was present in only very small concentrations (0.78%), it is important to meat flavor and has been reported in many meat products (Mottram, 1991). Several mechanisms have been proposed for the formation of thiazoles. A major route involves the reaction of hydrogen sulfide and ammonia with aliphatic aldehydes and 1,2-dicarbonyl compounds (Figure 3) by a mechanism similar to that for oxazole formation (Mottram, 1991). Smaller amounts of thiazole were produced in a Maillard reaction mixture containing lipids; however, the difference in quantities formed in the presence and absence of lipids was small and within the 89 l 0 NH, NH R ———> R'_< H H aldehyde imine R2 f“ NH, R, R3 s/ \ rt . I R 0 R3 SH di-carbonyl mercapto ketone -HZO V 2 R2 PM ‘L A bit 3 s>\rt R3 s R thiazole thiazoline Formation of thiazoles and thiazolines from intermediates of the Figure 3. Maillard reaction. 90 range of standard deviation (Farmer and Mottram, 1990a). Heterocyclic thiol compounds were the most abundant compounds formed in the system containing cysteine and ribose (Table 1). Furfuryl mercaptan was the most dominant (18.82 %) compound in the model system and is believed to be a product of the reaction between furfural and hydrogen sulfide (Farmer and Mottram, 1990a). Two other heterocyclic thiols, 2-methyl-3-furanthiol and 3- thiophenethiol, are also very important because of the quantities produced (13.82 % and 10.32 %, respectively). These compounds are reported to be important contributors to meat aroma. Furans and thiophenes with a thiol substitution in the 3-position possess meat-like aromas (Mottram and Madruga, 1994). However, the quantities of these heterocyclic thiol compounds are also reduced when lipids are introduced into the model system of cysteine and ribose (Farmer and Mottram, 1990a). Four disulfide compounds were identified in the present study (Figure 4), albeit in low concentrations (total relative amounts less than 1%, Table 1). However, they are very important contributors to meaty aroma (Farmer and Mottram, 1990b; Farmer and Patterson, 1991; Mottram and Whitfield 1994; Mottram e_t fl, 1995). The disulfide compounds are reportedly derived from 2- methyl-3 -thiophenethiol, 2-methyl-3 -furanthiol and 2-furylmethanethiol (W erkhoff e_t al., 1990; Farmer and Mottranr, 1990b; Mottram e_t a, 1995). Farmer a1 al_. (1990b, 1991) reported that phospholipids were involved in the formation of these disulfide compounds, and reduced the quantities formed in the model system. Several non-heterocyclic sulfur-containing compounds were also identified. Among these were mercapto carbonyls (2-mercapo-3 -pentanone and 3-mercapto- 91 O S -—S 0 H3 0 02H5 3-[(2methyI-3-furyl) dithio]-Z-pentanone (meaty, boiled meat, roast beef)‘ ('81:) bis (2-methyl-3-furyl) disulfide (beefy, meaty, meat soup)‘ [XS—NC? 2-methyl-3-[(2-furylmethyl) dithio] furan (meat, burnt meat, roast meat)‘ (s—s fikws 3-[(2methyl-3-furyl) dithio]-Z-butanone (meaty, burnt meat)‘ Figure 4. Structures and sensorial properties of disulfide compounds isolated from a model system containing cysteine and ribose. “ (Mottram et al., 1995). 92 2-pentanone) and methylthio] acetates (methylthiol acetate and ethyl 2- (methylthiol) acetate). The second most abundant compound produced in the model system was 3-mercapto-2-pentanone (16.47%) and possessed a sulfury note (Mottram e_t al_., 1995). Farmer and Mottram (19903) concluded that mercapto carbonyls were derived from the reaction of hydrogen sulfide and dicarbonyls, and that their formation tended to be reduced considerably by the presence of phospholipids in a model system. Two methylthio acetates were also identified and were present at approximately 5.0 % of the total amount. The majority of methyl sulfides are generated from the reaction of hydrogen sulfide or methanethiol with aliphatic aldehydes or alcohols, and their further reactions give rise to a range of acyclic and heterocyclic compounds (MacLeod, 1986). However, these methylthio acetates are believed to contribute to the aroma of meat (Mottram, 1991). Three furan-related compounds were identified: furfural, 2,5-furandione and 1-(2-furyl)-3 -butanone. These compounds represented more than 10.5% of the total amount of compounds isolated in the model system (Table l). Furfural is a degradation product of ribose and its formation is also affected negatively by the presence of phospholipids (Farmer and Mottram, 1990a). The other two compounds, 2,5-furandione and 1-(2-furyl)-3-butanone, were not reported by Farmer and Mottram (1990a); however, they could be derived from furfural. Two other non-heterocyclic compounds, cyclopentanone and 2-ethyl butanal, were identified and could be derived from sugar degradation (Nursten, 1986). No pyrazines were isolated in this study. This could be explained by the fact that the formation of pyrazine is favored by alkaline conditions (Maga, 1992). The model system study was performed in an aqueous phase with a pH value of 93 5.6 immediately after heating. The slightly acidic reaction medium could lead to minimal formation of pyrazines. Selection of Ledger cormpounds for use in subsequent studies In subsequent studies, the effect of phospholipids on the generation of Maillard reaction volatiles in a model system containing cysteine and ribose will be investigated. Therefore, some compounds isolated in this model system and shown to be influenced by the presence of lipids by Farmer and Mottram (1990a) were selected as "marker" compounds. Fifteen compounds were selected 14 of which contain sulfur (Table 2). The role of sulfur-containing heterocyclic compounds in meat flavor is well documented (Mottram, 1991). Their contribution is due to their meaty notes and low threshold levels. The formation of the majority of these compounds is affected by the presence of lipids. Lipid-derived carbonyls compete for free hydrogen sulfide with sugar-derived dicarbonyls and reduces / suppresses the formation of Maillard reaction volatiles (Farmer and Mottram, 1990a; Whitfield, 1992). Therefore, these compounds will be used to assess the effect of unsaturated fatty acids on the suppression of Maillard reaction products in model systems containing phospholipids. CONCLUSION A new technique for the analysis of volatiles was developed using a combination of heating and solvent extraction to desorb volatiles fiom Tenax 94 Table 2. Sensorial properties of Maillard reaction volatiles selected as marker compounds to study the effect of phospholipids on volatile production. Compound Odor description Source Thiophenes 2-acetyl thiophene l-(2-thienyl)-2-propanone dimethyl formylthiophene 3-ethyl-2-fonnylthiophene Icyclic compounds dihydrothienothiophene Thiazoles thiazole Disulfides 3-[(2-methyl-3-furyl) dithio]-2- butanone bis-(2-methyl-3-fmyl) disulfide 2-methyl-3-[(2-fi1rylmethyl) dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- pentanone Heterocyclic thiols furfuryl mercaptan 2-methyl-3-furanthiol 3-thiophenethiol Mercapto carbonyls 3-mercapto-2-pentanone Non-sulfur compounds furfural chicken, beef, porkza3 beef2 roastedlr2 creamy, caramel-like1 nutty, meaty1 nutty, meaty1 meaty, sulfuryl,3 meaty, nuttyla2 chicken, beef, pork2,3 meaty, burnt meat4 beefy, meaty, broth4 meat, burnt meat, roast beef4 meaty, broiled meat, roast beef4 sweet, caramel-like1 beef3 beef broth, roast meatl meaty1 sulfurous, rotten meat3 pungent but sweet,3 mutton, beef, pork3 1 Fors, 1983. 2 Shahidi at. a, 1986. 3 Mottram, 1991. 4 Mottram e_t a, 1995. 95 traps. Thirty-one compounds were identified in a mdoel system containing cysteine and ribose. Of these, fifteen compounds were selected as "marker" compounds. The selection was based on literature reports of compounds whose formation in similar model systems was suppressed by the presence of lipids. REFERENCES Baltes, W., Kunert-Kirchhoff, J. and Reese, G. 1989. Model reactions on generation of thermal aroma compounds. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 143-155. Barnes, R. D., Law, L. M. and McLeod A. J. 1981. Comparison of some porous polymers as adsorbants for collection of odour samples and the application of the technique to an enviromental maldour, Analyst 106: 412-418. Buckholz, L. L., Withycombe, D. A. and Daun, H. 1980. Application and characteristics of polymer adsorption method used to analyze flavor volatile from peanuts. J. Agric. Food Chem. 28(4): 760-765. Farmer, L. J. and Patterson, R. L. S. 1991. Compounds contributing to meat flavor. Food Chem. 40: 201-205. Farmer, L. J. and Mottram, D. S. 1990a. Interaction of lipid in the Maillard reaction between cysteine and ribose: the effect of a triglyceride and three phospholipids on the volatile products. J. Sci. Food Agric. 53(11): 505-525. Farmer, L. J. and Mottram, D. S. 1990b. Recent studies on the formation of meat- like aroma compounds. In "Flavor Science and Technology", Y. Bseeiére and A. F. Thomas (Ed), Wiley, New York, pp 113-116. Farmer, L. J ., Mottram, D. S. and Whitfield F. B. 1989. Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid. J. Sci. Food Agric. 49: 347-368. 96 Hodge, J. E. 1953. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1: 928-943. Maga, J. A. 1992. Pyrazine update. Food Reviews International 8(4): 479-558. McCarffrey, C. A., MacLachlan, J. and Brooks, B. I. 1994. Absorbent tube evaluation for the preconcentration of volatile organic compounds in air for analysis by gas chromatography - mass spectrometry. Analyst 119 : 897- 902. McLeod G. 1986. The scientific and technological basis of meat flavours. In "Developments in Food Flavours", G. G. Birch and M. G. Lindley (Ed), Elservier Applied Science, London, UK, pp 191-223. Morton, I. D., Akroyd P. and May, C. G. Great Britian Patent 836,694, 1960. Mottram, D. S., Madruga, M. S. and Whitfield F. B. 1995. Some novel meatlike aroma compounds from the reactions of alkanediones with hydrogen sulfide and furanthiols. J. Agric. Food Chem. 43(1): 189-193. Mottram, D. S. 1994a. Flavor compounds formed during the Maillard reaction. In "Thermal Generated Flavor: Maillard Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 104-126. Mottram, D. S. 1994b. Some aspects of the chemistry of meat flavor. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 210-230. Mottram, D. S. and Madruga, M. S. 1994. Important sulfur-containing aroma volatiles in meat. In "Sulfur Compounds in Foods", C. J. Mussinan and M. E. Keelan (Ed), ACS Syrnposiourn Series 564, Washington, DC, pp 180- 187. Mottram, D. S. and Whitfield F. B. 1994. Aroma volatiles from Maillard systems. In "Thermal Generated Flavor: Maillard Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 180-191. 97 Mottram, D. S. 1991. Meat. In "Volatile Compounds in Foods and Beverages", H. Maarse (Ed), Marcel Dekker, Inc., New York, pp 107-177. Nursten, H. E. 1986. Aroma compounds fiom the Maillard reaction. In "Developments in Food Flavours", G. G. Birch and M. G. Lindley (Ed), Elservier Applied Science, London, UK, pp 173-190. Nursten, H. E. 1981. Workshop on volatile products. Progr. Food Nutr. Sci. 5: 491-496. Toyoda, T., Nohara, I. and Sato, T. 1993. Headspace analysis of volatile compounds emitted from various citrus blossoms. 1n "Bioactive Volatile Compounds from Plant", R. Teranishi, R. G. Buttery and H. Sugisawa (Ed), ACS Symposium Series 525, Washington, DC, pp 205-219. Werkhoff, P., Brfining, J ., Emberger, R., Gitntert, M., Kopsel. M., Kuhn, W. and Surburg, H. 1990. Isolation and characterization of volatile sulfur containing meat flavor components in model systems. J. Agric. Food Chem. 38: 777-791. CHAPTER 2 EFFECTS OF PHOSPHOLIPIDS AND ALDEHYDES ON THE FORMATION OF MAILLARD REACTION VOLATILES IN MODEL SYSTEMS CONTAINING CYSTEINE AND RIBOSE ABSTRACT Cysteine and ribose were heated to produce Maillard reaction volatiles, and fifteen compounds were selected as "markers" to study the influence of lipids on their formation. Sensory assessment revealed that the main difference between the aromas developed in model systems containing cysteine and ribose was that more meaty and sulfury notes were detected in those systems which did not contain phospholipids. No discernible differences in aroma were observed between samples containing phosphatidylethanolamine (PE) and phosphatidylcholine (PC). More meaty and sulfury aromas were generated in the reaction mixture containing PC-distearoyl than in the system containing PC-dioleoyl. Phospholipids suppressed the formation of selected Maillard reaction products. In addition, the degree of unsaturation of the phospholipids influenced the extent to which Maillard reaction volatiles were generated. However, the eflects of phospholipids on the formation of volatile compounds was not related to the amino moieties of the phospholipids. Suppression of the formation of Maillard volatiles was also observed in model systems containing aldehydes. 98 INTRODUCTION Maillard-derived heterocyclic compounds are key components in the desirable flavors of processed and cooked foods, particularly meat products. The primary source of these compounds is the Maillard reaction between amino acids and reducing sugars (Whitfield e_t a, 1988). The Maillard reaction does not require high temperatures and readily produces aroma compounds at temperatures associated with the cooking of food. It also occurs much more readily at low moisture levels; thus, flavor compounds generated in meat by the Maillard reaction tend to be associated with the dehydrated areas during cooking (Mottrarrr, 1994c). Recently, model system studies have implicated lipids in the Maillard reaction, thus affecting the production of certain heterocyclic compounds (Salter at g, 1988; Whitfield at Q, 1988; Farmer e_t Q, 1989; Farmer and Mottram, 1990a, b). When lipids were added to model systems containing cysteine and ribose, a noticeable quenching effect on the formation of Maillard reaction products was observed. Many of those compounds whose mechanism of formation was susceptible to lipid intervention, were affected more by the presence of a phospholipid than by a triacylglycerol (Farmer and Mottram, 1990a). A mechanism by which phospholipids suppress the formation of Maillard reaction products in model system was proposed by Farmer and Mottram (1990a). It was hypothesized that carbonyl compounds formed by the thermal oxidation of lipids competed with the reducing sugars for Maillard reaction intermediates such as hydrogen sulfide and ammonia, thus reducing the formation of heterocyclic compounds. Because carbonyl compounds are derived from lipid oxidation, the 99 100 degree of unsaturation of lipids may play an important role in this lipid - Maillard interaction. Compounds isolated from a heated Maillard model system containing cysteine and ribose in the presence of beef triacylglycerols (BTG), beef phospholipids (BPL), phosphatidylcholine (PC) and phosphatidylethanolamine (PE) were classified into four categories by Farmer and Mottram (1990a). These categories are: [A] Compounds formed only in the presence of lipids. The greatest differences between the effects of lipids were observed among those compounds which required lipid precursors for their formation. [B] Compounds suppressed by the presence of lipids. These compounds were further subdivided based on the effects of different lipids on their formation. (a) Phospholipids showed greater suppression than triacylglycerols. (b) PE and BPL showed greater suppression than BTG and PC. (c) Triacylglycerols showed greater suppression than phospholipids. (d) Triacylglycerols caused similar reductions to phospholipids. [C] Compounds largely unaffected by the nature of lipids. Compounds such as certain thiazoles showed some reduction on the addition of lipid especially PC (Whitfield at a, 1988); however, the differences were small. Thus, no clear trends were inferred [D] Compounds not included in the previous categories. The four lipids influenced the formation of 2-furfural by diflerent extents. The final concentration was approximately halved by the addition of PC or PE to the model systems, while BPL and BTG suppresse to a lesser extent (Whitfield e_t al., 1988). 101 As previous results have shown, the formation of Maillard reaction products is influenced by the presence of lipids in a Maillard model system. The degree of lipid unsaturation may play an important role in this interaction. Thus, this study was designed to address the effects of adding lipids of different degrees of unsaturation to the model system on the production of Maillard reaction volatiles. The investigation involved heating phospholipids with different fatty acids with cysteine and ribose. The influence of lipid derived carbonyls on the Maillard reaction was also determined by including selected aldehydes in the model systems. EXPERIMENTAL Materials Cysteine and ribose, two phosphatidylethanolamines (PE) (L-or- phosphatidylethanolamine: dioleoyl and Don-phosphatidylethanolamine, distearoyl) and three phosphatidylcholines (PC) (Lo-phosphatidylcholine, dilinoleoyl, Ira-phosphatidylcholine, dioleoyl and L-a-phosphatidylcholine, distearoyl) were purchased from Sigma Chemical Co. (St. Louis, MO). Hexanal and dodecanal were obtained fi'om Aldrich Chemical Co., Inc. (Milwaukee, WI). Ethyl acetate and sodium phosphates (monobasic and dibasic) were obtained from J. T. Baker Chemical Co. (Phillipsburg, NJ). Tenax GC was obtained fiom Alltech Associates Inc. (Deerfield IL). All the reagents and solvents used in this study were analytical and/or HPLC grade. 102 Preparation of Maillard reaction volatiles Cysteine and ribose were heated together in the presence of phospholipids / aldehydes in a phosphate buffer. Cysteine and ribose served as the control system. Each reaction mixture was prepared in triplicate. The reaction mixtures were prepared by dissolving cysteine (0.61 g) and ribose (0.75 g) in 2 ml phosphate buffer (0.5M, pH 5.7) in a 10 ml clear glass vial (Pierce Chemical Co., Rockford, IL). For the phospholipid systems, 100 mg of each phospholipid were added to the reaction mixture. The vials were immersed in an ultrasonic bath (Branson 2000, Danbury, CT), and held at 50°C to disperse the lipids prior to heating. These reaction vials were then flushed with a stream of nitrogen (30 ml/min) for 2 min, and sealed using Tuf-bond Teflon-silicon discs and aluminum caps (Pierce Chemical Co., Rockford IL). The reactants were heated at 140°C in an autoclave under a pressure of 0.28 MPa for l h. After heating, the vials containing the Maillard reaction products were stored in a fieezer (-10°C) prior to the volatile collection. For the aldehyde systems, hexanal (0.1 and 1 mole) and dodecanal (0.1, 0.5, 1 and 2 mole) were added to the glass vials containing cysteine and ribose. The reaction vials were sonicated for 2 min and flushed with nitrogen prior to scaling. The Maillard reaction was carried out at 140°C in an autoclave under a pressure for 1 h. The vials were then cooled and stored in a freezer prior to analysis. Isolation and identification of Maillard reaction volatiles Volatile compounds developed from the Maillard reaction were collected using a purge-and-trap procedure, and then desorbed into ethyl acetate using a technique combining thermal desorption-solvent extraction. Tenax traps were used 103 to collect the volatile compounds, and were conditioned overnight at 180°C with nitrogen flowing at 30 ml / min prior to volatile collection. The contents of the reaction vials (ca. 2 ml) were transferred to a two-neck flask (500 ml, Kontes, Vineland NJ) containing 18 ml phosphate buffer (0.5M, pH 5.7). The flask was fitted with a L-shape glass joint (Kontes, Vineland NJ), and a Tenax trap was attached by a stainless steel Swagelok reducing union (Cajon Ultra-Torr® fitting, Cajon, Macedonia, OH) to the outlet of the joint. During collection, the flask containing the Maillard reaction mixture was heated at 60°C in a heating mantle, and the Tenax trap was maintained at room temperature (approximately 20°C i 2°C). Volatile compounds were swept into the Tenax trap using a stream of nitrogen (30 ml /min), and the collection was continued for 1 h. At the end of the collection period the Tenax trap was directly connected to the nitrogen supply for 5 min to remove water vapor. A more detailed description of this procedure is provided in Chapter 1. Desorption of the trapped volatile compoands from the Tenaxlap During desorption, the Tenax trap was connected between a solvent (ethyl acetate) source and a cold trap. The Tenax trap was wrapped with a heating strip and heated at 150°C. The nitrogen was passed through the ethyl acetate reservoir (30 - 40 ml/min), and then the ethyl acetate - nitrogen gas was introduced into the heated Tenax trap. The nitrogen stream was maintained water-free by passing through a desiccant (anhydrous calcium sulfate) (W. A. Hammond Drierite Co., Xenia, OH ) before passing through the ethyl acetate reservoir. The desorbed volatile compounds were then condensed in a cold trap using dry ice - acetone. The desorption process was continued for 30 min, after which the ethyl acetate extract was concentrated to 0.1 ml under nitrogen. The concentrate was transferred 104 to a screw-cap vial and stored in a refiigerator prior to gas chromatographic / mass spectrometric analysis. Gas chromatogpaphy - mass spectrometry (GC / MS) The volatile compounds were identified using a HP 5890 gas chromatograph (Hewlett Packard Avondale, PA) interfaced with a HP 5970 MSD (Mass Selective Detector) mass spectrometer. The GC / MS system was equipped with a HP 59970 Chemstation Data System. A Carbowax 20M fused silica capillary column (60m x 0.32 mm id Supelco Inc., Bellefonte, PA) was used to separate the compounds. The GC oven temperature was programmed as follows: the initial temperature was set at 40°C for 5 min, the temperature was increased at a rate 2°C / min to a final temperature of 200°C, and maintained for 10 min. Helium was the carrier gas with a flow rate of 10 ml/min. The temperatures of the injection port and transfer lines were both set at 220°C. The mass spectrometer was operated in the electron impact mode with an electron energy of 70 eV and an ion source temperature of 250°C. Compounds were introduced to the ion source directly from the capillary column in the GC using an open-split interface. A continuous scan mode with a scan time of 1 sec over a mass range of 40-300 was employed. The GC / MS data were monitored stored and analyzed using an HP Chemstation data system. Two mass spectral libraries, the INRA Mass Spectra Library - "INRAMASS" (Laboratoire de Recherche sue les Aromes, Dijon, France) and the NIST / EPA / MSDC Mass Spectral Database (ACS Publication Co., Washington, DC), were used to identify the compounds isolated. 105 Evaluation of odor characteristics The odor characteristics of each sample were determined by three flavor chemists in the Department of Food Science and Human Nutrition, Michigan State University. When each sample was opened for volatile collection, the headspace in the reaction vials was sniffed by the panelists to assess the odor. They were asked to describe freely the odor characteristics and their comments were noted (Farmer and Mottram, 1990a). Determination of the effects of phosphohrids and aldehde on the MM mam The effects of phospholipids and aldehydes on the formation of Maillard reaction volatiles were determined by comparing the relative amounts of each compound produced in these systems to those in the controls. The calculation of the relative amounts was accomplished by using the peak areas on the gas chromatograms. The peak area of each compound fiom the control system (cysteine and ribose heated alone) was standardized as 100%, and the relative amounts of the same compound generated from the reaction mixture containing either phospholipid or aldehyde was determined as: peak area (with phospholipids or aldehydes) relative amount (%) = X 100 % peak area (control) RESULTS AND DISCUSSION SensorLevalufation of compounds developed from cysteine and ribose in the presence of phospholipids A summary of the odor descriptions attributed to the reaction mixtures is presented in Table l. The control reaction mixtures, i.e., cysteine and ribose, had pleasant meaty aromas; however, sulfirry notes were also detected. Upon adding phospholipids to the reaction mixtures, the intensities of the meaty and sulfury notes were decreased. No discernible differences between samples containing phosphatidylethanolamine (PE) and phosphatidylcholine (PC) were noticed. More meaty and sulfury aromas were generated from the reaction mixtures containing PC-distearoyl than the systems containing PC-dioleoyl. These results imply that the susceptibility of the fatty acid to oxidation may play a major role in determining the nature of the aroma compounds developed in the Maillard reaction. These observations were similar to those previously reported by Farmer and Mottram (1990a), who concluded that the major difference between the odors of reaction mixtures with and without lipids was in the relative intensities of the various meaty and sulfury notes. Thus, lipids, in particular phospholipids, may influence the Maillard reaction and the subsequent development of key flavor compounds, thus affecting the intensity of the flavor generated. 106 107 Table 1. Summary of odor descriptions attributed to the headspace of cysteine and ribose heated in the presence of phospholipids. REACTION MIXTURE DESCRIPTORS cysteine + ribose pleasant, brothy-beefy, slightly sulfury cysteine + ribose pleasant, slightly meaty, + phosphatidylethanolamine - dioleoyl slightly sweet, slightly "cooked green pepper" cysteine + ribose pleasant, slightly meaty, + phosphatidylcholine - dioleoyl slightly sweet, slightly caramelization cysteine + ribose pleasant, slightly meaty, + phosphatidylcholine - distearoyl sweet, slightly sulfury 108 Effect of phospholipids on the forrgation of Mflard reaction vomles from heate_d cysteine and ribose More than 30 compounds developed from the reaction of cysteine and ribose were identified. This number is smaller than those identified by Farmer and Mottram (1990a) using a sirrrilar model system. This may be due to the different equipment employed in these studies. Farmer and Mottram (1990a) used an "Unijector" (Scientific Glass Engineering Ltd Milton Keynes) in place of a conventional GC injector, and a cold trap (solid carbon dioxide) to increase the efficiency of desorption. However, the present study was not intended to repeat the definitive studies of Farmer and Mottram (1990a), but rather to determine the efl‘ects of phospholipids on the formation of selected Maillard reaction products. To do this, fifteen compounds whose formation was modified by the presence of lipids were selected as "marker" compounds (Whitfield e_t a, 1988; Farmer at a, 1989; Farmer and Mottranr, 1990a, b; Mottram e_t fl, 1995). These marker compounds were considered adequate to study the effects of lipids and their oxidation products on the Maillard reaction. The formation of the selected compounds was influenced by the presence of PE (Table 2). In general, their formation of the selected compounds was reduced when PE was present in the reaction mixtures. The relative amounts of the majority of the marker comopunds were suppressed by up to 25 % in the presence of PE-distearoyl (Table 2). Heterocyclic thiols, 3-mercapto-2-pentanone, and furfiual were noticeably more reduced than were thiazole, dihydrothienothiophene, 1-(2-thienyl)-2-propanone and three disulfides. The quantities of the latter compounds were reduced by less than 10 %. 109 Table 2. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phosphatidylethanolamines. Compound PE-distearoyl PE-dioleoyl Thiophenes 2-acetyl thiophene --_-a ---- 1-(2-thienyl)-2-propanone 90b 60 dimethyl formylthiophene ---- ---- 3-ethyl-2-formylthiophene ---- ---- Iicyclic compounds dihydrothienothiophene 95 98 Thiazoles thiazole 87 79 Disulfides 3-[(2-methyl-3-furyl) dithio]-2- 99 90 butanone bis-(2-methyl-3-furyl) disulfide 74 82 2-methyl-3-[(2-fi1rylmethyl) 96 88 dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- 99 79 pentanone Heterocyclic thiols furfuryl mercaptan 76 77 2-methyl-3-furanthiol 77 50 3-thiophenethiol 8 1 55 Mercapto carbonyls 3-mercapto-2-pentanone 75 33 Non-sulfur compounds furfural 75 75 110 Table 2. (continued) a "----" indicates not detected. b The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction mixture without phospholipid (i.e., control). The calculation was accomplished by standardizing the peak area of the selected compound in the control system as 100%. The peak area of the Maillard reaction product was then divided by the peak area of the same compound in the control. 111 Similar suppressive effects were found in the Maillard reaction system containing PE-dioleoyl (Table 3). However, more discernible reductions in the formation of some compounds were observed with the more unsaturated phospholipid. Only dihydrothienothiophene and 3-[(2-methyl-3 -furyl) dithio]-2- butanone were reduced by less than 10 % by this phospholipid while the amount of 3-mercapto-2-pentanone formed was approximately one third of that produced in the control model system. Two heterocyclic thiols, l-(2-thienyl)-2-propanone and 3-[(2-methyl-3-furyl) dithio]-2-pentanone, were also considerately reduced while the remaining compounds listed in the Table 2 suffered smaller reductions. There was no apparent difference in the formation of dihydrothienothiophene, furfuryl mercaptan and furfural between the two PEs. These results indicated that lipids, particularly those that are unsaturated suppress the formation of selected Maillard reaction volatiles developed from the reaction of cysteine and ribose. The data generated also confirmed the observations of Farmer and Mottram (1990a) that phospholipids from eg and beef sources have a greater suppressive effect on the Maillard reaction than do triacylglycerols. Phosphatidylcholine (PC) also suppressed the formation of selected Maillard reaction products when included in the model system containing cysteine and ribose (Table 3). Three PCs with different degrees of unsaturation were used and each influenced the formation of Maillard reaction volatiles to a different extent. Compounds developed in the system containing PC-distearoyl were reduced by 11 - 93% relative to the control. Several compounds such as 2-acetyl thiophene, dimethyl formylthiophene and 3-ethyl-2-formylthiophene, were not detected in 112 Table 3. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phosphatidylcholines. Compound PC-distearoyl PC-dioleoyl PC-dilinoleoyl Thiophenes 2-acetyl thiophene _---a ---- ---- 1-(2-thienyl)-2-propanone 80b 75 12 dimethyl formylthiophene ---- ---- ---- 3-ethyl-2-formylthiophene ---- ---- ---- Iicyclic compounds dihydrothienothiophene 85 77 ---- Thiazoles thiazole 83 89 84 Disulfides 3-[(2-methyl-3-furyl) dithio]-2- 79 59 20 butanone bis-(2-methyl-3-furyl) disulfide 63 37 15 2-methyl-3-[(2-furylmethyl) 93 73 66 dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- 88 78 81 pentanone Heterocyclic thiols furfuryl mercaptan 76 75 76 2-methyl-3 -furanthiol 1 1 ---- ---- 3-thiophenethiol 76 78 78 Mercapto carbonyls 3-mercapto-2-pentanone 76 57 ---- Non-sulfur compounds furfural 75 82 79 113 Table 3. (continued) a "----" indicates not detected. b The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction mixture without phospholipid (i.e., control). The calculation was accomplished by standardizing the peak area of the selected compound in the control as 100%. The peak area of the Maillard reaction product was then divided by the peak area of the same compound in the control. 114 this system. The compound suppressed to the greatest extent was 2-methyl-3- furanthiol (11 % of control), while the smallest suppression was observed for 2- methyl-3-[(2-fiuylmethyl) dithio] furan (93% of control). The relative amounts of the remaining selected compounds generally ranged between 75 - 90%. The data in Table 3 also show that the extent of suppression of the Maillard reaction volatiles in the presence of PCs increased with increasing unsaturation of the fatty acid moieties. Greater suppression was observed in the systems containing PC-dioleoyl or PC-dilinoleoyl compared to the systems with PC- distearoyl (Table 3). Three compounds, dihydrothienothiophene, 2-methyl-3- furanthiol and 3-mercapto-3 -pentanone, were not detected in the presence of PC- dilinoleoyl, while 2-methyl-3-furanthiol was not detected in the reaction mixture containing PC-dioleoyl. Noticeable decreases in the formation of l-(2-thienyl)-2- propanone and three disulfides were observed with PC-dioleoyl. These studies with PEs and PCs indicate that the generation of Maillard reaction products from cysteine and ribose is influenced by the presence of lipids. Greater suppressive effects were observed with increasing unsaturation of the phospholipids. Similar results were reported by Farmer and Mottram (1990a) who demonstrated that lipids were capable of reducing the formation of Maillard reaction products. They also reported that beef phospholipids exhibited a greater suppressive effect than did beef triacylglycerols, and suggested that this could be due to the difference in the fatty acid composition of the lipids. These investigators also suggested the most likely pathways by which lipids could interact in the Maillard reaction. Lipid-derived carbonyls could react with the amino groups of amino acids and free ammonia produced by the Strecker degradation, while hydroxyl and carbonyl products derived via lipid oxidation could react with free hydrogen sulfide. 115 Results in this study indicate that the formation of Maillard reaction compounds is impacted by the presence of phospholipids. The greater influence of phospholipids with more unsaturation implies that thermal oxidation during the heating of the model systems may be the reason for this effect. To verify this hypothesis further, the effect of aldehydes on the formation of Maillard reaction products was investigated. The effects of vgious phospholipids on the MMd reaction involvi_ng cysteine and ribose As described previously, both PE and PC exerted a quchening effect on the formation of Maillard reaction products (Tables 2 and 3). However, dissirrrilarities in the behavior of PE and PC with the same degree of unsaturation were observed. Greater suppression of the formation of three disulfides, 2-methyl-3-furanthiol and dihydrothienothiophene was observed in the presence of PC-dioleoyl compared to PE-dioleoyl (Figure 1). However, smaller amounts of 1-(2-thienyl)-2-propanone, thiazole, 3-thiophenethiol, 3-mercapto-2-pentanone and furfural were formed in the system containing PE-dioleoyl. One long-chain alkylthiophene, 2- hexylthiophene, was detected in the reaction systems containing PC-dioleoyl and PE-dioleoyl. The route of formation for this chemical was suggested to involve the reaction of 2,4-decadienal and hydrogen sulfide. Therefore, it was classified in the category of compounds formed only in the presence of lipids (Farmer and Mottram, 1990a). The amount of 2-hexylthiophene isolated from PE-dioleoyl was approximately 40% of that from PC-dioleoyl. Farmer and Mottram (1990a) also reported that a greater amount of this compound was formed in the Maillard 116 Figure 1. Relative amounts of selected Maillard reaction volatiles isolated from the reaction mixture of cysteine and ribose in the presence of phospholipids compared to cysteine and ribose alone. 117 ococuseflafifiiaatsoeaxa \ «Esau AEaniscEaIn MAA/ :83 32% AsseraéIisoea ococscoxdozgu AEaatfiEéI 58:85.: 58.2 _e_scwéatsaea _a_sucoiowaa o:o:£:oa.~.2nwucoE.m _wsta ozofio_§>E§.~-_>£o.m ococaogocflfiongcfi \ gaggfl/ X 9.2325 Soufw M a. runoulu 311981911 118 reaction system in the presence of egg PC compared to egg PE. These was no apparent trend in the suppression of the formation of Maillard reaction products in the two phospholipid systems. Similar results were reported in a study of the browning reaction of various unsaturated fatty acids heated with either ethanolamine or choline (Husain at g, 1986). It was shown that the resultant color intensity increased with the degree of fatty acid unsaturation in the presence of both ethanolamine and choline. However, for the same fatty acid choline gave greater browning intensity than ethanolarrrine. Ethanolamine showed much greater reactivity than choline when heated with the conjugated carbonyl derivatives of methyl linoleate. This is due to the reaction of the amino group of ethanolarrrine and carbonyls to yield a Schiffs base and subsequently the colored products (Husain e_t al., 1986). Since the amino group of ethanolamine in PE may react with sugar-derived dicarbonyls (Farmer and Mottram, 1990a), greater suppression of the Maillard reaction compounds formed in the presence of PE may be predicted. In a previous study, egg PE and egg PC were used in a Maillard system containing cysteine and ribose (Farmer and Mottram 1990a). Greater quantities of l-heptanol, l-octanol and 2,4-decadienal were produced from the thermal degradation of PC than PE, and thus might cause greater suppression of the Maillard reaction. Therefore, the suppressive effects of phospholipids on the formation of Maillard reaction products may be due to a number of factors, the most important being the formation of lipid-derived carbonyl compounds. Effecps of aldehydes on the production of Maillard reaction volatiles in a heated gasteineg and ribose model The results of the phospholipid study indicate that phospholipids have a quenching effect on the formation of Maillard reaction products in a heated model 119 system containing cysteine and ribose. It is assumed that the unsaturated lipid moieties undergo thermal oxidation druing heating of the model system, and that the carbonyl compounds produced may be involved in suppressing the Maillard reaction. Two aldehydes, hexanal and dodecanal, were used to investigate the effects of lipid oxidation products on the Maillard reaction containing cysteine and ribose. The formation of selected Maillard reaction products was influenced by the addition of hexanal in the model system (Table 4). Four disulfides, three thiophenes, dihydrothienothiophene and one heterocyclic compounds were not detected in the presence of 0.1 mole of hexanal. One of the thiophene compounds, 1-(2-thienyl)-2-propanone, was remarkably influenced by the difference concentrations of hexanal used. However, increasing the amount of hexanal in the model system containing cysteine and ribose, resulted in a similar suppression of the formation of selected Maillard reaction volatiles (Table 4). Generally, these results confirm that lipid oxidation, or more accurately carbonyl compounds derived by thermal oxidation, influences the formation of Maillard volatiles. In a similar investigation, a longer chain aldehyde, dodecanal, was also heated with cysteine and ribose in a model system. Thiophenes and thienothiophene was not detected in any of the systems with dodecanal (Table 5). It was also observed that the greater the concentration of dodecanal in the model system, the greater the reduction of the quantities produced. The most obvious difference was observed between the model systems containing 0.1 and 0.5 mole of dodecanal (Table 5). Disulfide compounds were sensitive to the presence of 120 Table 4. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of hexanal. Compound 0.1 (mmole) 1 (mmole) Thiophenes 2-acetyl thiophene ----a ---- l-(2-thienyl)-2-propanone 105b 29 dimethyl formylthiophene ---- ---- 3-ethyl-2-formylthiophene ---- -_-- Iicyclic compounds dihydrothienothiophene ---- ---- Thiazoles thiazole 84 92 Disulfides 3-[(2-methyl-3-furyl) dithio]-2- --..- ---- butanone bis-(2-methyl-3-furyl) disulfide ---- ---- 2-methyl-3-[(2-firrylmethyl) ---- ---- dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- ---- ---- pentanone Heterocyclic thiols furfuryl mercaptan 57 36 2-methyl-3 -furanthiol 7 1 ---- 3-thiophenethiol ---- 45 Mercapto carbonyls 3-mercapto-2-pentanone 38 58 Non-sulfur compounds furfural 61 63 121 Table 4. (continued) a "----" indicates not detected. b The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction mixture without hexanal (i.e., control). The calculation was accomplished by standardizing the peak area of the selected compound in the control as 100%. The peak area of the Maillard reaction product was then divided by the peak area of the same compound in the control. 122 Table 5. Relative amounts of Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of dodecanal. Compound 0. l 0.5 1 2 (mole) Thiophenes 2-acetyl thiophene _-__a ---- ---- ---- l-(2-thienyl)-2-propanone ---- ---- ---- ---- dimethyl formylthiophene ---- ---- ---— ---- 3-ethyl-2-fonnylthiophene ---- ---- ---- ---- Iicyclic compounds dihydrothienothiophene ---- m- an ---- Thiazoles thiazole 10419 39 34 27 Disulfides 3-[(2-methyl-3-firry1) dithio]-2- 23 14 ---- ---- butanone bis-(2-methyl-3-furyl) disulfide 25 19 ---- ---- 2-methyl-3-[(2-furylmethyl) ---- ---- ---- ---- dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- ---- ---- ---- ---- pentanone Heterocyclic thiols furfuryl mercaptan 71 35 43 16 2-methyl-3-furanthiol 98 28 29 5 3-thiophenethiol 58 43 31 21 Mercapto carbonyls 3-mercapto-2-pentanone 46 27 1 1 4 Non-sulfur compounds furfural 39 36 23 16 123 Table 5. (continued) a "----" indicates not detected. b The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction mixture without dodecanal (i.e., control). The calculation was accomplished by standardizing the peak area of the selected compound in the control as 100%. The peak area of the Maillard reaction product was then divided by the peak area of the same compound in the control. 124 dodecanal, none being detected when 1 mole was added to the model system. Thiazole was found in a relatively high amount in the system containing 0.1 mole of dodecanal. The suppressive effects of different aldehydes on the formation of selected Maillard reaction volatiles in a model system of cysteine and ribose are presented in Figure 2. Both hexanal and dodecanal suppressed of the formation of selected Maillard reaction products. Two disulfides and four thiophenes were not detected in both systems, and, to some degrees, hexanal produced more suppression of the selected Maillard reaction volatiles than dodecanal at same concentration (0.1 mmole) (Figure 2). CONCLUSION These studies show that the susceptibility of phospholipids to thermal oxidation (i.e., degree of unsaturation) influences the extent that lipids suppress the formation of Maillard reaction volatiles in model systems containing cysteine and ribose. The inclusion of aldehydes in the model system also resulted in suppression of the production of Maillard reaction volatiles, giving indirect evidence that thermal oxidation of lipids is involved in this suppressive effect. It is our hypothesis that the stabilization of lipids in meat products such as through the use of nitrite as a curing adjunct (Freybler at al_., 1993), could reduce the effect of lipids and/or their oxidation carbonyl products on the development of flavor 125 Figure 2. Relative amounts of selected Maillard reaction volatiles isolated fiom the reaction mixture of cysteine and ribose in the presence of 0.1 mmole aldehydes compared to cyseine and ribose alone. 126 are 32% cgoébeazadaeosa 80583- 0:05.33- N _oeefiéirflsoaa 6368533: ~ ~-oa8noE-m K 0:282“ tavern oaoauaoa-m-:>=oiruv-_ Suicide—head 35056 - r”. ozonqofla_>§om-m-_xfio-m oaoamofluonofiuoaxfiu e_oufifi oqouSB-~-moE€ :ba-m-_E~oE-mz-m 823% eba-m-_§oeé-ms 32% ebaaéeoeaza 533808 b . \ _ aha lAmeix V. _ \K 4». Eeé ll! [I 'V on ,8 runounr angels“ 127 compounds via the Maillard reaction. The stabilization of lipids could contribute, in part, to the difference in the flavors of cured and uncured meats. A study of the influence of the nitrite-stabilized phospholipids and antioxidants on the Maillard reaction is reported in the next section (Chapter 3). REFERENCES Farmer, L. J. and Patterson, R. L. S. 1991. Compounds contributing to meat flavor. Food Chem. 40: 201-205. Farmer, L. J. and Mottram, D. S. 1990a. Interaction of lipid in the Maillard reaction between cysteine and ribose: the effect of a triglyceride and three phospholipids on the volatile products. J. Sci. Food Agric. 53(11): 505-525. Farmer, L. J. and Mottram, D. S. 1990b. Recent studies on the formation of meat- like aroma compounds. In "Flavor Science and Technology", Y. Bseeiére and A. F. Thomas (Ed), Wiley, New York, pp 113-116. Farmer, L. J ., Mottram, D. S. and Whitfield F. B. 1989. Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid. J. Sci. Food Agric. 49: 347-368. Fors, S. 1983. Sensory properties of volatile Maillard reaction products and related compounds. In "The Maillard Reaction in Food and Nutrition", G. R. Waller and M. S. Feather (Ed), ACS Symposium Series 215, Washington, DC, pp 185-286. Hartrnen, G. J., Charlin, J. T., Hwang, S. S., Bao, Y., Tang, J. and Ho, C.-T. 1984. Identification of 3,5-diisobutyl-1, 2, 4-trithiolane and 2-isobutyl-3, 5- diisopropylpyridine in fiied chicken flavor. J. Food Sci. 49: 1398-1400. Hodge, J. E. 1953. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1: 928-943. 128 Husain, S. R., Terao, J. and Matushita, S. 1986. Comparison of phosphatidylcholine and phosphatidylethanolamine affecting the browning reaction of heated oil. In "Amino-Carbonyl Reactions in Food and Model Systems", M. Fujimaki, M. Namiki and H. Kato (Ed), Elsevier, Amsterdam, Netherlands, pp 301-309. Mottram, D. S., Madruga, M. S. and Whitfield F. B. 1995. Some novel meatlike aroma compounds from the reactions of alkanediones with hydrogen sulfide and furanthiols. J. Agric. Food Chem. 43(1): 189-193. Mottram, D. S. 1994a. Flavor compounds formed during the Maillard reaction. In "Thermal Generated Flavor: Maillard Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 104-126. Mottram, D. S. and Whitfield F. B. 1994b. Aroma volatiles from Maillard systems. In "Thermal Generated Flavor: Maillard Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 180-191. Mottranr, D. S. 1994c. Some aspects of the chemistry of meat flavor. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 210-230. Mottram, D. S. and Madruga, M. S. 1994d. Important sulfur-containing aroma volatiles in meat. In "Sulfur Compounds in Foods", C. J. Mussinan and M. E. Keelan (Ed), ACS Symposium Series 564, Washington DC, pp 180- 187. Salter, L. J ., Mottram, D. S. and Whitfield F. B. 1988. Volatiles compounds produced in Maillard reactions involving glycine, ribose and phospholipid. J. Sci. Food Agric. 46: 227-242. Shahidi, F ., Rubin, L. J. and D'Souza, L. A. 1986. Meat flavor volatiles: A review of the composition, techniques of analysis and sensory evaluation. CRC Crit. Rev. Food Sci. Nutr. 24(2): 141-243. Takken, H. J ., van der Linde, L. M., de Volois P. J ., van Dort, J. M., Boelens, M. 1976. Reaction product of or-dicarbonyl compounds, aldehydes, hydrogen sulfide and ammonia. In "Phenolic, Sulfur and Nitrogen Compounds in 129 Food F lavor", G. Charalambous and I. Katz (Ed), American Chemistry Society, Washington, DC, pp 114-121. Werkhoff, P., Brfining, J ., Emberger, R., Gi'mtert, M., Kopsel. M., Kuhn, W. and Surburg, H. 1990. Isolation and characterization of volatile sulfur containing meat flavor components in model systems. J. Agric. Food Chem, 38: 777-791 Whitfield F. B. 1992. Volatiles for interactions of Maillard reactions and lipids. CRC Crit. Rev. Food Sci. Nutr. 31(1/2): 1-58. Whitfield F. B., Mottram, D. 8., Brock, S., Puckey, D. J. and Salter, L. J. 1988. Effect of phospholipid on the formation of volatile heterocyclic compounds in heated aqueous solutions of amino acids and ribose. J. Sci. Food Agric. 42: 261-272. CHADTIED 3 THE EFFECTS OF NITRITE-STABILIZED LIPIDS ON THE FORMATION OF FLAVOR VOLATILES FROM A HEATED MODEL SYSTEM INVOLVING CYSTEINE AND RIBOSE ABSTRACT Phospholipids (PLs) were isolated from cured and uncured hams and heated in a model system with cysteine and ribose. Nitrite stabilization of the phospholipids was confirmed by heating them with morpholine and measuring the formation of the corresponding N-nitrosanrine. The concentrations of N- nitrosomorpholine produced by PLs from cured and uncured hams were significantly different (p < 0.005). Volatiles produced by heating cysteine and ribose were less suppressed in the presence of PLs from cured hams. When antioxidants were added to the model systems of cysteine, ribose and phosphatidylcholine before heating, less suppression of the Maillard reaction was also observed. It was concluded that nitrite not only prevents lipid oxidation, but also modifies the formation of Maillard flavor compounds. This modification of the Maillard reaction may be partially responsible for the difference in flavors of cured and uncured meats. A sensory evaluation of volatiles produced from the cysteine-ribose-phosphatidylcholine system indicated flavor differences as a result of adding antioxidants to the system. These observations were confirmed by gas chromatographic analysis 130 INTRODUCTION The Maillard reaction is a complex reaction which yields both high molecular weight browning products and volatile aroma compounds (Hodge, 1953). Among the latter are heterocyclic compounds which play a major role in meat flavor (Mottram, 1994). Recently, the interaction between lipids and Maillard reaction mixtures (cysteine and ribose) in a heated model system has been studied extensively (Whitfield at a, 1988; Farmer e_t al_., 1989; Farmer and Mottranr, 19903). These studies established that lipids or their degradation products can alter the composition of the volatile products formed, both qualitatively and quantitatively. In general, the addition of phospholipids to the simple Maillard reaction mixtures exerted a quenching efiect on the production of the heterocyclic compounds derived solely from the reaction of the amino acids with ribose (Whitfield 1992). For example, alkylpyrazines were produced in slightly smaller concentrations in the presence of the lipids. The trend was more obvious with many of alkylthiophenes and alkylthiazoles (Whitfield _e_t a, 1988; Farmer and Mottram, 1990a). The observed reduction in the concentrations of the alkylpyrazines, alkylthiazoles and alkylthiophenes in the reactions containing phospholipids was thought to be due to the preferential reaction of free hydrogen sulfide or ammonia with compounds derived from the lipids (Whitfield _e_t al_., 1988; Farmer and Mottram, 1990a). Phospholipids in meat systems contain unsaturated fatty acids, particularly those with three or more double bonds. The unsaturated moieties undergo oxidative breakdown during heating to give carbonyl compounds which could compete with sugar-derived carbonyls for Maillard reaction intermediates 131 132 such as free hydrogen sulfide. This will affect the relative proportions of the heterocyclic compounds produced from the Maillard reaction (Farmer and Mottram, 1990a; Mottram, 1994). The use of antioxidants to stabilize lipids in a heated model system containing cysteine and ribose may minimize the effect of lipid oxidation on the formation of heterocyclic compounds. It is well established that nitrite functions as an antioxidant in meats (Gray and Crackel, 1992), and several mechanisms have been proposed to explain its antioxidant effects: 1. Nitrite forms stable complexes with the heme pigment in meat systems, thereby inhibiting heme-catalyzed lipid oxidation (Igene e_t al_., 1985). 2. Nitrite chelates metal ions such as ferrous ions, thus rendering these ions unavailable for catalysis of oxidation (Morrissey and Tichivangana, 1985). 3. Nitrite forms certain compounds, such as nitrite oxide - myoglobin, which can quench free radicals (Kanner e_t al., 1980). 4. Nitrite stabilizes unsaturated lipids in meat systems (Freybler at al_., 1993). Stabilization of the porphyrin ring during cooking appears to be the most plausible mechanism (Gray and Pearson, 1987). The mechanism involving nitrite addition to the double bonds of unsaturated fatty acids also merits consideration (F reybler at a, 1993). The antioxidant activities of nitrite may modify the flavor compounds generated by the Maillard reaction in heat-processed meat products, and could explain, in part, the difference in the flavors of cured and uncured meats. There is considerable evidence that smaller quantities of carbonyl compounds are formed in nitrite-cured meats during cooking compared to those that are nitrite-free (Bailey and Swain, 1973; Shahidi, 1989; Gray and Pearson, 1 994; Ramarathnam e_t al_., 1991a, b; Ramarathnam and Rubin, 1994). 133 Consequently, less suppression or quenching of the Maillard reaction may be obtained resulting in the generation of a more intense meaty aroma. The purpose of this study was to determine the effect of nitrite-stabilized phospholipids on the formation of the Maillard reaction products in model systems containing cysteine and ribose. Another objective was to investigate the efi‘ect of adding antioxidants to the model system of cysteine, ribose and phospholipids on the formation of flavor compounds. EXPERIMENTAL Materials Pork ham samples were obtained from a local commercial processor within 48 hr of slaughter. Chemicals used in the curing brine were salt (International Salt Co., Clark Summit, PA), sodium ascorbate (Sigma Chemical Co., St. Louis, MO), sodium nitrite (J. T. Baker Inc., Phillipsburg, NJ) and sodium tripolyphosphate (Stauffer Chemical Co., Westport, CT). Anhydrous sodium sulfate, calcium phosphate and dichloromethane were purchased from Mallinckrodt Inc. (Paris, KY), and the solvent was redistilled before using. Celite 545 and methanol were purchased from Fisher Scientific Co. (Fair Lawn, NJ). Phosphatidylcholine (egg yolk, type IH-E) and a-tocopherol were purchased from Sigma Chemical Co. (St. Louis, MO). tert-Butylhydroquinone (Tenox® TBHQ, food-grade antioxidant) was obtained from Eastman Chemical Products Co. (Kingsport, TN). All the chemicals and solvents used were reagent grade. 134 Preparation of cured hams Three pork leg-quarter samples (ca. 50 lb each) were processed in the Meat Laboratory, Michigan State University. Each sample was divided into three equal portions for the different curing processes. The composition of the curing brines is listed in the Table 1. The pork samples were stitch pumped with the curing brine to 120% (wt/wt) of their green weights. The pumped hams were then placed in large plastic containers. The hams were transferred to a cooler set at 2-4°C and allowed to equilibrate for overnight. The cured samples were then removed from the cooler and freshened using cold nmning water. The hams were then placed in stockinettes and hung in the smokehouse. Samples were placed in an Elec-Trol laboratory smokehouse (Drying System Inc., Chicago, IL), and processed under the following conditions: 1. 140°F (dry bulb temperature) / 110°F (wet bulb temperature) (60°C/43°C) for 2 hr. 2. 160°F / l30°F (71°C/54°C) for 2-4 hr. 3. 175°F / 155°F (79°C/68°C) until a temperature of 152°F (67°C) was reached. After cooking, the hams were held overnight in a cooler (approximately 4°C) prior to slicing and packing. The hams were sliced into 1 in (2.54 cm) slices, and each slice was packed in a polyethylene - laminated nylon pouch (Koch, Kansas City, MO). These pouches were 90 pm thick and had a water vapor transmission 135 Table 1. Composition of the curing brines used to process hams. Ingredients weight (g/kg) brine 1 brine 2 brine 3 water 882.9 882.9 882.9 salt 100.0 100.0 100.0 sugar 34.0 34.0 34.0 sodium 15.0 15.0 15.0 tripolyphosphate sodium ascorbate 1.3 1.3 1.3 sodium nitrite 0 156 1560 (mg fkg) 136 rate of 0.041 ml / mzodayommHg and an oxygen transmission rate of 0.124 ml / mzodayommHg at 227°C and 50% RH. The packaged ham samples were sealed under vacuum and stored in a fieezer (-15°C) until required for analysis. Qaantitation of phospholipids in hams Total phospholipids in the hams were quantitatively isolated using the dry- column procedure of Marmer e_t a. (1981). The hams were trimmed of excess adipose tissue prior to lipid extraction. The lean portions were ground using a food processor (Model K5-A, Kitchen Aid Troy, Ohio). A 5g sample was ground with 20g anhydrous sodium sulfate and 15g Celite 545 in a porcelain mortar until a fiee-flow powder was produced. The mixture was then transferred to a glass chromatography column (16 mm id. x 25 cm with 8 mm id. x 5 cm drip tip, Kontes, Vineyard NJ) containing 10g of a Celite / calcium phosphate mixture (9:1 w/w), and lightly tamped to obtain a uniform bed using a glass rod. The neutral lipids were eluted first with 150 ml dichloromethane, followed by the phospholipids (polar fraction) using 150 ml of the solvent mixture of dichloromethane / methanol (9:1 v/v). The solvent extracts were concentrated to approximately 5 ml using a rotary vacuum evaporator (Btichi Rotavapor, Biichi Inc., Postfach, Switzerland). The concentrates were quantitatively transferred into 4-ml glass vials (Kimble Opticlear glass vials, 15 mm o.d. x 45 mm, Supelco Inc., Bellefonte, PA), and flushed with nitrogen to remove the solvent. After weighing, the vials were immediately sealed with screw-caps and stored in a fi'eezer (-15°C) until required for further analyses. 137 Nitroatfion of morpholine by phospholprids extracted from hfl To confirm that nitrite reacted with the phospholipids to form nitro-nitrosite derivatives, a secondary amine (morpholine) was heated with the extracted phospholipids to form the corresponding N-nitrosarnine (F reybler at Q, 1993). Aliquots (180 pg) of the phospholipids were transferred to a glass ampule to which were added morpholine (500 pg) and 0.6 pl heptane. Six replications of each reaction mixture were prepared. The ampules were flushed with nitrogen for 5 min, flarne-sealed and heated in a Reactitlrerrn Heating Module (Pierce Chemical Company, Rockford IL) at 130 °C for 30 min. The reaction mixtures were then allowed to cool, and the contents in each ampule dissolved in 1 ml dichloromethane (Ross e_t Q, 1987). Quantitation of N-nitrosomorpholine was achieved using a GC (Varian Model 3700) / thermal energy analyzer system (TEA) (Thermal Electron Corporation Model 502, Walthanr, Mass.) linked to a Hewlett Packard 3390a integrator (Freybler e_t Q, 1993). A glass chromatographic column (2 mm id. x 3 m) packed with 10% Carbowax 20m TPA on 80/ 100 Chromosorb WHP was used for analysis. The GC operating conditions were programmed as follows: an initial temperature of 140°C for l min, followed by temperature increase up to 180 °C at a rate of 15 °C/min, holding at this temperature for 7 min. A 5 pl aliquot of the concentrate was injected. A standard solution of N-nitrosomorpholine (NMOR) (0.1 pg / ml) was used for quantitation of the N-nitrosamine produced. Preparation and isoirtion of Maillard reaction volafil§§ Cysteine and ribose were heated together in the presence and absence of phospholipids in a phosphate buffer. Each reaction mixture was prepared in triplicate. 138 The reaction mixtures were prepared by dissolving cysteine (0.61 g) and ribose (0.75 g) in 2 ml phosphate buffer (0.5M, pH 5.7) in a 10 ml clear glass vial (Pierce Chemical Co, Rockford IL). Phospholipids (100 mg) were weighed into the reaction vials which were then were immersed in an ultrasonic bath (Branson 2000, Danbury, CT). The reaction mixtures were held at 45°C until the reactants were uniformly dispersed throughout the aqueous phase. The vials were flushed with a stream of nitrogen (30 ml/min) for 2 min prior to sealing with a Tuf-bond Teflon-silicon disc and aluminum cap (Pierce Chemical Co, Rockford IL). The reaction vials were heated at 140°C in an autoclave under a pressure of 0.28 MP3 for l h. After heating, they were stored in a freezer (-10°C) prior to volatile collection. To study the effect of antioxidants on flavor production, a model system containing phosphatidylcholine (PC) (100 mg), cysteine (0.61 g) and ribose (0.75g) in 2 ml phosphate buffer (0.5M, pH 5.7) was heated at 140°C. Three antioxidants: a-tocopherol, TBHQ, and sodium nitrite (150 pg/g lipid) were investigated and added individually to the reaction mixture. Three replications were carried out for each antioxidant. The aroma generated in these model systems was evaluated by three flavor chemists in the Department of Food Science and Human Nutrition, Michigan State University. Each sample was presented to the panelists in the reaction vial at room temperature (22 :l: 2°C). The panelists were asked to open the vials and sniff the headspace. The panels were allowed to describe the odor characteristics fieely. Volatile compounds were collected using the purge-and-trap procedure which was described in Chapter 1. The contents of the reaction vials (ca. 2 ml) were transferred to a 500-ml two-neck flask (Kontes, Vineland NJ) containing 18 ml phosphate buffer (0.5M, pH 5.7). The flask was fitted with a L—shape glass 139 joint (Kontes, Vineland NJ), and the Tenax trap was attached by a stainless steel Swagelok reducing union (Cajon Ultra-Torr® fitting, Cajon, Macedonia, OH) to the outlet of the joint. During collection, the flask was heated at 60°C using a heating mantle, and the Tenax trap was maintained at room temperature (approximately, 20°C i 2°C). The volatile compounds were swept into the Tenax trap using a stream of nitrogen (30 ml /min), and the collection was continued for 1 h. The flask was then removed and the Tenax trap was connected directly to the nitrogen supply for 5 min to remove water vapor (Whitfield at Q, 1988; Farmer e_t Q, 1990a). For the desorption of volatile compounds, the Tenax trap was connected between a solvent (ethyl acetate) source and a cold trap. The nitrogen was passed through the ethyl acetate reservoir (30 - 40 ml/min), and the ethyl acetate - nitrogen gas was introduced into the heated (150°C) Tenax trap. The desorbed volatile compounds were then condensed in a dry ice / acetone cold trap. The desorption process was continued for 30 min, after which the ethyl acetate extract was concentrated to 0.1 ml under nitrogen. The concentrate was transferred to a screw-cap vial and stored in a refiigerator prior to gas chromatographic / mass spectrometeric (GC/MS) analysis. The volatile compounds were separated and identified using a HP 5890 GC (Hewlett Packard Avondale, PA) interfaced with a HP 5970 MSD (Mass Select Detector) MS. The GC / MS system was equipped with a HP 59970 Chemstation Data System. A Carbowax 20M fused silica capillary column (60m x 0.32 mm id Supelco Inc., Bellefonte, PA) was used for the separation of the volatile compounds. The GC oven temperature was programmed initially at 40°C for 5 min, and increased to 200°C at a rate of 2°C / min, then held at this temperature for 10 min. Helium was used as a carrier gas with a flow rate of 10 ml/min. The 140 temperatures of the injection port and transfer lines were both set at 220°C. The MS was operated in the electron impact mode with an electron energy of 70 eV and an ion source temperature of 250°C. Compounds were introduced to the ion source directly from the capillary column in the GC using an open-split interface. A continuous scan mode with a scan time of 1 see over a mass range of 40-300 was employed. The GC / MS data were monitored and stored in the HP Chemstation data system. The Maillard reaction products were identified by comparing their mass spectra and Kovats indices (KI) to those of authentic standard compounds. Two mass spectral libraries, the INRA Mass Spectra Library - "INRAMASS" (Laboratoire de Recherche sue les Aromes, Dijon, France) and the NIST / EPA / MSDC Mass Spectral Database (ACS Publication Co, Washington, DC), were used to confirm the identity of selected compounds. Statistical analysis A one way AN OVA (Analysis of Variance) was used to determine the significance of the differences of the phospholipid contents in the hams and the amounts of NMOR produced by the various phospholipids. The significance between treatment means was determined by the Bonferroni-t test for non- orthogonal comparisons (Gill, 1978). Six replications were performed in each treatment, and the data were calculated using the MSTATC microcomputer statistical program (Michigan State University, 1991). RESULTS AND DISCUSSION mutation of phospholipids in ham samples The quantities of phospholipids extracted from cured and uncured hams are listed in Table 2. The phospholipid contents of the lean portion of the hams were in the range 0.8 - 1.2 % (w/w). The phospholipid contents in hams cured with 150 and 1500 mg/kg nitrite averaged of 1.2% i 0.2% (w/w), while the uncured hams appeared to contain smaller quantities ( 0.8% i 0.2%, w/w). The phospholipid contents expressed as a percentage of the total lipids were also determined (Table 2). An average of 10.6 % i 3.4% was obtained for the nitrite-free hams, while the percentages of phospholipids in the nitrite-cured hams averaged 13.0 i 3.3 and 13.2 i 2.7% respectively for the hams cured with 150 and 1500 (mg nitrite / kg meat), respectively. The concentrations of phospholipids in the lean tissue of meats are relatively constant (Bodwell e_t Q, 1986). Their proportions vary as a fimction of the total lipids in muscle (Anderson, 1988). As the total lipid content of meat decreases fiom 12% to 1%, the percentage of phospholipid in the total fat increases from less than 10% to over 80% (W eihrauch e_t Q, 1983). For example, Anderson (1988) reported that the phospholipid contents in lean tissues of beef and pork represented approximately 12% and 15% of the total lipid respectively. These values are comparable to those reported in Table 2. Nitrosation of moppholine To confirm that the stabilization of the phospholipids was due to the 141 142 Table 2. Lipid contents1 of hams cured with varying levels of nitrite. Nitrite Phospholipids Total lipids phospholipid %2 target level (mg/kg ham) wt3 % wt % 0 0.04 i 0.01 0.8 i 0.2 0.39 i 0.10 7.75 i 0.02 10.6 i' 3.4 (control) 150 0.06 i 0.01 1.2 i 0.2 0.46 i 0.04 9.20 i 0.01 13.0 i 3.3 1500 0.06 i 0.01 1.2 i 0.2 0.44 i 0.04 8.97 i 0.01 13.2 i 2.7 1 Lipid contents are expressed as an average wt 1 standard deviation (SD) (g) in a 5 g ham sample. Two slices were taken from each ham sample. There were three hams for each level of nitrite. Each determination was performed in duplicate. 2 Phospholipid contents are expressed as a percentage of the total lipid content. 3 No significant difference in the average phospholipid weights between hams cured with different levels of nitrite was obtained (Gill, 1978). 143 interaction between nitrite / oxides of nitrogen and the double bonds of the unsaturated fatty acids, phospholipids extracted from nitrite-flee and nitrite-cured hams were heated with morpholine to form the corresponding N-nitroso compound. Nitrite, per se, is not a nitrosating agent. However, a number of nitrite derivatives can nitrosate amines under certain conditions. Oxides of nitrogen (N Ox) are believed to be responsible for N-nitrosamine formation in cured meat products. Liu et Q. (1988) proposed that dinitrogen trioxide is formed when nitrite is added to cured meat and can react with the carbon-carbon double bonds of unsaturated lipids to form nitro-nitroso derivatives. This derivative can decompose during cooking to release the NOx compounds which are capable of nitrosating secondary amines. Significant differences (p < 0.005) in NMOR formation was observed using phospholipids from the various hams (Table 3). The greatest quantity of NMOR was formed (1858.60 at 451.23 pg / g lipid) by the phospholipids extracted fi'om the hams cured with 1500 mg / kg nitrite. Phospholipids from the hams cured with 150 mg / kg nitrite produced 695.99 i 319.00 pg NMOR / g lipid. However, NMOR was not detected when phospholipids from the uncured hams were heated with morpholine. Results of this study clearly indicate that phospholipids extracted from the nitrite-cured hams were capable of nitrosating morpholine, whereas those from the uncured samples could not. Similar results were obtained by F reybler e_t Q. (1993). When phospholipids from nitrite-cured pork (156 mg / kg) were heated with morpholine, a significantly (p < 0.05) greater amount of NMOR (1433.5 :l: 82.5 pg / g lipid) was produced compared to that formed by phospholipids from 144 Table 3. Formation of N-nitrosomorpholine on heating morpholine with phospholipids extracted from hams cured with different levels of nitritel. Sample pg N-nitrosamine I g lipid phospholipids from not detecteda uncured ham phospholipids from 695.99 :1: 319.00b 150 mg / kg nitrite ham (337.96 - 950.00) phospholipids from 1858.60 1- 451.23C 1500 mg / kg nitrite ham (1353.58 - 2222.22) Values (i.e., the means of 6 replications) having different superscripts within column are significantly different (p < 0.005) in a one-way ANOVA analysis (Gill, 1978). 145 uncured pork (226.8 i 79.4 pg / g lipid). These investigators concluded that nitrite stabilizes the double bonds of unsaturated fatty acids against peroxidative changes through the formation of nitro-nitroso derivatives. Volatiles generated from the Maillard reaction in the presence of phospholipids The formation of selected Maillard reaction volatiles was reduced when phospholipids from uncured were present in the model system containing cysteine and ribose (Table 4). Fifteen volatile compounds generated from the reaction of cysteine and ribose upon heating were previously identified in order to study the efiect of ham phospholipids on the Maillard reaction (Chapter 2). Formation of these compounds in an identical model system were reported to be influenced by the presence of phospholipids (Farmer and Mottram, 1990a, b), an observation that was confirmed in Chapter 2. Most of the selected Maillard reaction volatiles were reduced in the presence of phospholipids from uncured hams, the relative amounts being in the range of 12 - 84 % compared to those in the phospholipid-free control. Furthermore, thiazole and two thiophenes were not detected in this system (Table 4). The amounts of a bicyclic compound heterocyclic thiols, two disulfides, 2- mercapto-3-pentanone and furfural were reduced to less than 30%, while two disulfides were reduced to approximately 55% (Table 4). A smaller effect on the formation of Maillard reaction volatiles was observed when phospholipids from the cured hams were heated with cysteine and ribose (Table 4). An obvious difference was noticed in the formation of thiazole. Concentrations of this compound in the systems containing phospholipids fi'om 146 Table 4. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence of phospholipids extracted from hams. Compound phospholipids nitrite-free nitrite-cured nitrite-cured ham (150 mg/kg) (1500 mg/kg) Thiophenes 2-acetyl thiophene --_-a ---- ---- l-(2-thienyl)-2-propanone 80b 88 82 dimethyl formylthiophene ---- ---- ---- 3-ethyl-2-formylthiophene 84 84 87 Iicyclic compounds dihydrothienothiophene 24 25 3 1 Thiazoles thiazole ---- 39 48 Disulfides 3-[(2-methyl-3-furyl) dithio]-2- 20 27 33 butanone bis-(2-methyl-3-furyl) disulfide 26 37 35 2-methyl-3-[(2-fluylmethyl) 58 63 66 dithio] furan 3-[(2-methyl-3-furyl) dithio]-2- 53 56 89 pentanone Heterocyclic thiols furfuryl mercaptan 25 31 36 2-methyl-3-furanthiol 12 27 28 3-thiophenethiol 18 40 38 Mercapto carbonyls 3-mercapto—2-pentanone 25 30 33 Non-sulfur compounds furfural 24 24 22 147 Table 4. (continued) a "----" indicates not detected. b The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction mixture without phospholipid (i.e. control). The calculation was accomplished by standardizing the peak area of the selected compound in the control as 100%. The peak area of the Maillard reaction product was then divided by the peak area of the same compound in the control system. 148 hams cured with 150 and 1500 mg /kg nitrite were 61% and 52% of the concentration in the control system (i.e. cysteine and ribose). Thiazole was not detected in the system containing phospholipids from the uncured ham (Table 4).Two thiophenes (1-(2-thienyl)-2-propane and 3-ethyl-2-formyl thiophene) were not detected in both the cured ham phospholipid systems. Similar results were obtained for phospholipids from uncured hams. Greater quantities of disulfides, heterocyclic thiols and mercapto carbonyl compounds were observed in the reactions with phospholipids extracted from the nitrite-cured hams compared to those with phospholipids from nitrite-free samples (Table 4). No apparent difference in furfural formation was observed between systems. Previous results have indicated that the formation of selected Maillard reaction volatiles were suppressed in the presence of phospholipids (Farmer and Mottram, 1990a, b). It was proposed that this quenching / suppression was due to lipid-derived carbonyls reacting with the amino group of cysteine and/or fiee ammonia or hydrogen sulfide from the Strecker degradation. As nitrite stabilizes the unsaturated fatty acids in phospholipids (Freybler e_t Q, 1993), it was anticipated that the effect of phospholipids from nitrite-cured hams on the Maillard reaction would be less than phospholipids from the nitrite-free hams. This was confirmed by the results in Table 4. This study also revealed that smaller quantities of lipid oxidation products, notably hexanal (Shahidi at Q, 1987), were produced in those systems containing the nitrite-stabilized phospholipids (Table 5). Hexanal and three long-chain alkyl heterocyclic compounds were isolated from the reaction between cysteine, ribose and phospholipids. Hexanal is a major lipid oxidation carbonyl product and is used routinely to monitor the extent of lipid 149 Table 5. Relative amounts of lipid oxidation products and heterocyclic compounds derived from the reaction of cysteine and ribose in the presence of phospholipids extracted from hams cured with different levels of nitrites. Compound Phospholipids nitrite-flee nitrite-cured nitrite-cured ham (150 mg/kg) (1500 mg/kg) hexanal 100 67a 28 2-pentyl furan 100 40 48 2-pentyl thiophene 100 ----b 12 2-hexyl thiophene 100 ---- 3 a The numbers represent the average percent (3 replications) of the relative amounts of the compounds formed compared to the same compound isolated from the system containing phospholipids from uncured hams. The calculation was accomplished by standardizing the peak area of selected compound fi'om the system containing phospholipids from the uncured hams as 100%. The peak area of the volatile compound was then divided by the peak area of same compound in the uncured ham system. b "----": not detected. 150 oxidation in foods (Frankel, 1991; Shahidi and Pegg, 1994). It has been proposed that the long-chain alkyl thiophenes are derived from the reaction between hydrogen sulfide and the corresponding furans at high temperatures (Vemin at Q, 1982). A more probable route is the reaction of hydrogen sulfide with diean (Farmer and Mottranr, 1990a), and the alkyl substituent orginated fi'om the omega end of fatty acid chain (Farmer and Whitfield 1993). Suggested routes for the formation of long-chain alkyl heterocyclic compounds from 2,4-decadienal, a thermal oxidation product of unsaturated fatty acids, are shown in Figure 1. As 2,4-decadienal, like hexanal, is known to be produced from linoleic acid (Frankel, 1991), it is expected that smaller quantities of this aldehyde would be produced in the system containing nitrite-stabilized lipids. Thus, it is expected that smaller concentrations of 2-pentyl furan, 2-pentyl thiophene and 2-hexyl thiophene were produced in the model systems containing phospholipids fiom nitrite-cured hams relative to systems containing phospholipids from the uncured hams (Table 5). The results indicated that the production of selected Maillard reaction volatiles is quenched to a lesser extent by phospholipids that have been stabilized by the addition of nitrite during curing of hams. It has also been observed that smaller quantities of carbonyl compounds, i.e., pentanal and hexanal, are produced in the model system containing the nitrite-stabilized phospholipids. Therefore, it is evident that when lipid oxidation is minimized through the stabilization of lipids by nitrite, the extents of quenching/suppression of the Maillard reaction is also minimized. Consequently, the antioxidant role of nitrite in cured meats may play an important role in the formation of cured meat flavor, i.e., by balancing the degree of interaction between lipid oxidation and products of the Maillard reaction. 151 oxidation Lipid ’ Csflrr'SH=fH-CH=EHO Hf‘ E H941 .. CH” S ett ( CSHll S \j 6 U CSHII N H l -H.0 J-H,o 101 CSHll N 2-pentylthiapyran 2-hexylthiophene 2-pentylpyrazine Figure 1. Suggested routes for the formation of 2-pentylthiapyran, 2-hexylthiophene and 2-pentylpyridine from 2,4-decadienal (Farmer and Mottram, 1 990a). 152 Effect of antioxidants on the interaction between the Maillard reaction and lipids [gagged model system The previous results indicated that nitrite stabilization of phospholipids from cured hams minimizes the suppression of selected Maillard reaction volatiles by thermal oxidation of phospholipids. Thus, antioxidants (e. g, a-tocopherol and TBHQ) when added to a model system containing cysteine, ribose and phospholipids, should reduce the extent of lipid oxidation and ultimately the effect of the carbonyls on the Maillard reaction. The antioxidants provided different degrees of modification of the lipid- Maillard interaction (Table 6). Compounds including two disulfides, 2-methyl-3- furanthiol, mercapto carbonyl and furfural were produced in similar quantities in systems with different antioxidants. However, nitrite had a greater effect in minimizing the suppression of thiazole, two disulfides and two heterocylic thiols than did a-tocopherol and TBHQ (Table 6). Hexanal concentrations were smallest in the systems containing a-tocopherol, but the lipid-derived long-chain thiophene (2-hexylthiophene) was also present in the greatest amounts in these systems. A sensory evaluation of the odor of systems containing cysteine, ribose and phosphatidylcholine with different antioxidants revealed no obvious difference, through the use of the antioxidants (Table 7). All samples had slightly meaty aroma; however, one panelist noted that the most meaty note was in the system containing TBHQ. The antioxidants provided different degrees of influence on lipid oxidation and subsequently in the formation of the Maillard reaction products (Table 6). There was no obvious trend on the formation of volatile compounds from heating 153 Table 6. Relative amounts of selected Maillard reaction volatiles from the reaction of cysteine and ribose in the presence phosphatidylcholine and antioxidants. Compound antioxidants (0.015%, w/w PC) nitrite or-tocopherol TBHQ Thiazoles thiazole 42a ----b ---- Bisulfides 3-[(2-methyl-3-furyl) dithio]-2- 35 7 7 butanone bis-(2-methyl-3-furyl) disulfide 20 6 12 2-methyl-3-[(2-furylmethyl) dithio] 25 27 24 furan 3-[(2-methyl-3-furyl) dithio]-2- 46 49 26 pentanone Heterocyclic thiols furfuryl mercaptan 33 15 20 2-methyl-3-furanthiol 38 27 40 3-thiophenethiol 50 33 37 Mercapto carbonyls 3-mercapto-2-pentanone 4 l 43 44 Non-sulfur compounds furfural 31 34 32 Compounds related to lipid oxidation hexanal 100 78° 97 2-hexylthiophene 100 214 192 a The number represents the average percent (3 replications) of the compound compared to the same compound isolated from the reaction of cysteine and ribose alone (i.e. control). The calculation was accomplished by standardizing the peak area of the selected compound in the control as 100%. The peak area of Maillard reaction product was then divided by the peak area of same compound from the control. b "----" means did not detected. c Numbers indicated the relative amounts of compounds isolated fi'om heated cysteine and ribose in the presence of antioxidants compared with nitrite where the peak areas were standardized as 100%. 154 Table 7. Summary of odor descriptions attributed to the headspace of heated cysteine and ribose in the presence of phosphatidylcholine and antioxidants. Reaction mixture Descriptors cysteine + ribose slightly meaty, + 100 mg phosphatidylcholine slightly vinegar + 0.015% (w/w PC) nitrite cysteine + ribose slightly meaty, + 100 mg phosphatidylcholine slightly sweet + 0.015% (w/w PC) a-tocopherol cysteine + ribose slightly meaty + 100 mg phosphatidylcholine most meaty/ beefy in all + 0.015% (w/w PC) TBHQ 155 cysteine, ribose and PC with different antioxidants. However, the relative amounts of the selected Maillard reaction products in the various model systems were in the order: control (cysteine and ribose) > control + phospholipids from nitrite- cured (150 mg/kg) ham, control+ PC + 0.015 % nitrite (w/w PC), control + PC + 0.015% or-tocopherol (w/w PC) > control + phospholipids from nitrite-free hams; however, no differences were apparent in the quantities formed in the following systems: control + phospholipids from nitrite-cured (150 mg/kg) ham, control + PC + 0.015 % nitrite (w/w PC), and control + PC + 0.015% a-tocopherol (w/w PC) (Figure 2). The greatest quantities of hexanal and 2-hexylthiophene were found in the system containing phospholipids from the uncured hams. The results of this study indicate that antioxidants inhibit lipid oxidation, and thus, minimize the quenching of the formation of Maillard reaction products. These results indicate that phospholipids from uncured hams suppress the formation of Maillard reaction volatiles in a model system containing cysteine and ribose. This suppression is caused by oxidation of the fatty acids (Farmer and Mottram, 1990a). If the involvement of lipids is due to their thermal oxidation during the reaction process, then a "stabilized" lipid such as that shown by Freybler at Q. (1993) for phospholipids in cured pork, would have a smaller impact on the production of Maillard reaction volatiles than phospholipids from uncured meats. This hypothesis was established by the results using phospholipids from hams cured with different levels of nitrite. The use of nitrite or other antioxidants (such as a-tocopherol and TBHQ) minimizes the suppression of the formation of Maillard reaction products in the model systems. The results confirm the assumption that prevention of lipid oxidation leads to less quenching of the Maillard reaction volatiles. Figure 2. 156 Relative amounts of selected Maillard reaction volatiles isolated from model systems containing cysteine and ribose in the presence of phospholipids with or without antioxidants. (Cotnrol: cysteine and ribose heated alone; PL (cured-ham): control + phospholipids from 150 (mg/kg) nitrite-cured hams; PC + nitrite: control + phosphatidylcholine (PC) + 0.015% nitrite (w/w PC); PC + Vit E: control + PC + 0.015% a-tocopherol (w/w PC); PL (pork): control + phospholipids from nitrite-free hams). i 3-[(2-rnethyl-3-fuyl) dlthlol-Z-butanone bls (2-methyl-3-ftnyl) diarlfide 2-methy1-3—{(2-furylmethyl) dithio] furan 3-{(2-meflryl-3—furyl) dithio]-Z-pentanone furfural mercaptan: / 2-methyl-3—furanthiol 157 / 3-thloph neth ol 3-mercapto-2—pentanone runomu oarrupu (410d) M 311A + DJ 9191111 + DJ (well-POM) ”Id [orruoo CONCLUSION Results in this study indicate that prevention of lipid oxidation may minimize the suppression of the Maillard reaction products in the model systems. Thus, it may be concluded that the importance of nitrite in cured meat flavor is due, in part, to its modefication of the relative concentrations of Maillard reaction products and lipid carbonyl compounds. REFERENCES Anderson, B. 1988. Composition and nutritional value of edible meat by-products. In "Advances in Meat Research, Vol. 5: Edible Meat By-Product", A. M. Pearson and T. R. Dutson (Ed), Elsevier, London, UK, pp 15-45. Bailey, M. E. and Swain, J. W. 1973. Influence of nitrite on meat flavor. In "Proceedings of the Meat Industry Resarch Conference", American Meat Science Association, Chicago, IL, pp 29-45. Bodwell, C. E. and Anderson, B. A. 1986. Nutritional composition and value of meat and meat products. In "Muscle as a Food", P. J. Bechtal (Ed), Academic Press, London, UK, pp 321-369. Farmer, L. J. and Whitfield F. B. 1993. Some aroma compounds formed from the interaction of lipid in the Maillard reaction. In “Proceeding Progress in Flavour Precusor Studies”, Wurzburg, Germany, P. Schreier (Ed), Allured Carol Stream, IL, pp 387-390. Farmer, L. J. and Mottram, D. S. 1990a. Interaction of lipid in the Maillard reaction between cysteine and ribose: the effect of a triglyceride and three phospholipids on the volatile products. J. Sci. Food Agric. 53(11): 505-525. 158 159 Farmer, L. J. and Mottram, D. S. 1990b. Recent studies on the formation of meat- like aroma compounds. In "Flavor Science and Technology", Y. Bseeiere and A. F. Thomas (Ed), Wiley, New York, pp 113-116. Farmer, L. J ., Mottram, D. S. and Whitfield F. B. 1989. Volatile compounds produced in Maillard reactions involving cysteine, ribose and phospholipid. J. Sci. Food Agric. 49: 347-368. Frankel, E. N. 1991. Recent advances in lipid oxidation. J. Sci. Food Agric. 54(4): 495-5 1 1. Freybler, L. A, Gray, J. I, Ashgar, A, Booren, A. M., Pearson, A. M. and Buckley, D. J. 1993. Nitrite stabilization of lipids in cured pork. Meat Sci. 33(1): 85-96. Freybler, L. 1989. Antioxidant Role of Nitrite in Meat System. MS. Thesis, Michigan State University, East Lansing, Michigan. Fors, S. 1983. Sensory properties of volatile Maillard reaction products and related compounds. In "The Maillard Reaction in Food and Nutrition", G. R. Waller and M. S. Feather (Ed), ACS Symposium Series 215, Washington, DC, pp 185-286. Gill, J. L. 1978. Design and Analysis of Experimental in the Animal and Medical Science, Vol. 1 and 11. Iowa State University Press, Ames, IA. Gray, J. I. and Pearson, A. M. 1994. Lipid-derived off-flavors in meat - formation and inhibition. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Acadenric and Professional, Glasgow, UK, pp 116-143. Gray, J. I. and Crackel, R. L. 1992. Oxidative flavor changes in meats: their origin and prevention. In "Thr Chemistry of Muscle-based Food", D. E. Johnson, M. E. Knight and D. A. Ledward (Ed), Royal Society of Chemistry, Cambrige, UK, PP 145-168. Gray, J. I. and Pearson, A. M. 1987. Rancidity and warrned-over flaovr. In "Advanced in Meat Resarch. VOL. 3. Restructed Meat and Poultry Products", A. M. Pearson and TR. Dutson (Ed), Van Nostrand Reinhold Co, New York, PP 221-269. 160 Hodge, J. E. 1953. Chemistry of browning reactions in model systems. J. Agric. Food Chem. 1: 928-943. Hotchkiss, J. H, Vecchio, A. J. and Ross, H. D. 1985. N-nitrosamine formation in flied out bacon fat: Evidence for nitrosation by lipid-bound nitrite. J. Agric. Food Chem. 33:5-8. Igene, J. O, Yarnauchi, K, Pearson, A. M., Gray, J. I. and Aust, S. D. 1985. Mechanisms by which nitrite inhibits the development of warmed-over flavor (W OF ) in cured meat. Food Chem. 18: 1-18. Ikins, W. G, Gray, J. I, Mandagere, A. K, Booren, A. M., Pearson, A. M. and Stachiw, MA. 1986. N-nitrosamine formation in fiied bacon processed with liquid smoke preparation. J. Agric. Food Chem. 34(6): 980-985. Kanner, J ., Ben-Gera, I. and Berman, S. 1980. Nitric-oxide myoglobin as an inhibitor of lipid oxidation. Lipids 15(11): 944-948. Liu, R. H, Conboy, J. J. and Hotchkiss, J. H. 1988. Nitrosation by nitro-nitrosite derivatives of olefms: a potential mechanism for N-nitrosamine formation in fiied bacon. J. Agric. Food Chem. 36: 984-987. Manner, W. N. and Maxwell, R. J. 1981. Dry column method for the quantitative extraction and simultaneous class separation of lipids from muscle tissue. Lipids 16(5): 365-371. Mottram, D. S, Madruga, M. S. and Whitfield F. B. 1995. Some novel meaflike aroma compounds from the reactions of alkanediones with hydrogen sulfide and furanthiols. J. Agric. Food Chem. 43(1): 189-193. Mottram, D. S. 1994. Some aspects of the chemistry of meat flavor. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 210-230. Morrissey, P. A. and Tichivangana, J. Z. 1985. The antioxidant activities of nitrite and nitrosylmyoglobin in cooked meats. Meat Sci. 14(3): 175-190. Ramarathnam, N. and Rubin, L. J. 1994. The flavor of cured meat. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 174-198. 161 Ramarathnam, N, Rubin, L. J. and Diosady, L. 1991a. Studies on meat flavor: I. Qualitative and quantitative differences in uncured and cured pork. J. Agric. Food Chem. 39: 344-350. Ramarathnam, N, Rubin, L. J. and Diosady, L. 1991b. Studies on meat flavor: 11. Quantitative approach to the estimation and differentiation of volatiles in uncured and cured beef and chicken. J. Agric. Food Chem. 39: 1839-1847. Ross, H. D, Henion, J ., Babish, J. G. and Hotchkiss, J. H. 1987. Nitrosating agents from the reaction between methyl oleate and dinitrogen trioxide: identification and mutagenicity. Food Chem. 23: 207-222. Shahidi, F. and Pegg, R. B. 1994. Hexanal as an indicator of the flavor deterioration of meat and meat products. In "Lipids in Food Flavors", C.-T. Ho and T. G. Hartman (Ed), ACS Symposium Series 558, ACS, Washington, D. C, pp 256-279. Shahidi, F. 1989. Flavor of cooked meats. In "Thermal Generation of Aromas", T. H. Parliament, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 188-201. Shahidi, F, Yun, J., Rubin, L. J. and Wood D. F. 1987. The hexanal content as an indicator of oxidative stability and flavour acceptability in cooked ground pork. J. Can. Inst. Food Sci. Technol. 20(2): 104-106. Shahidi, F, Rubin, L. J. and D'Souza, L. A. 1986. Meat flavor volatiles: A review of the composition, techniques of analysis and sensory evaluation. CRC Crit. Rev. Food Sci. Nutr. 24(2): 141-243. Vemin, G. and Parkanyi, C. 1982. Mechanisms of formation of heterocyclic compounds in Maillard and pyrolysis recation. In "Chemistry of Heterocyclic Compounds in Flavor and Aromas", G. Vemin (Ed), Ellis Horwood Chichester, UK, pp 151-207. Weihrauch, J. L. and Son, Y.-S. 1983. The phospholipid content of foods. JAOCS 60(12): 1971-1978. Whitfield F. B. 1992. Volatiles for interactions of Maillard reactions and lipids. CRC Crit. Rev. Food Sci. Nutr. 31(1/2): 1-58. 162 Whitfield F. B, Mottram, D. S, Brock, S, Puckey, D. J. and Salter, L. J. 1988. Effect of phospholipid on the formation of volatile heterocyclic compounds in heated aqueous solutions of amino acids and ribose. J. Sci. Food Agric. 42: 261-272. CHADIED 4 FLAVOR PROFILES OF CURED AND UNCURED HAMS ABSTRACT Flavor compounds from uncured and cured hams were collected by Tenax traps and then directly desorbed into a gas chromatograph - mass spectrometer system via a thermal desorption system. Carbonyl compounds were predominant in the flavor profiles of both cured and uncured hams. Of the carbonyl compounds identified pentanal and hexanal were the more abundant in the uncured hams. The addition of nitrite to hams reduced the formation of these carbonyl compounds. Two nitriles and two thiazoles were also detected in the hams cured with nitrite. 163 INTRODUCTION The curing of meat is an ancient art which originated in salting, one of the first methods used for preserving meat (Gray and Pearson, 1984). Nitrite is the most important ingredient in curing brines and imparts the characteristic flavor and pink color to cured meats. It has an antimicrobial effect and prevents the outgrowth of Clostridium botulinum spores. It also serves as an antioxidant by minimizing lipid oxidative changes during storage of meat. Although nitrite is closely associated with cured meat flavor, the chemical changes that are responsible for the unique aroma are not clearly understood (Gray and Pearson, 1984). Noel e_t Q. (1990) studied the role of nitrite in flavor development in uncooked cured meat products via a sensory evaluation, and found that the flavor of the samples containing nitrite was significantly different from that of nitrite-free samples. They concluded that nitrite plays a preeminent role in the development of the unique flavor notes in cured meats. The general chemical composition of volatile compounds in cured and uncured pork was summarized by Shahidi e_t Q. (1986). They indicated that the total number (118) of carbonyl compounds and hydrocarbons in uncured pork was greater than that of cured pork (45). The number of sulfur-containing compounds in uncured and cured pork was essentially the same. However, the concentration of sulfirr-containing compounds in cured pork was approximately two and half times greater than the concentrations of similar compounds in uncured pork. The types of sulfur-containing compounds were also different. Tritlriahexanes and thiophenes were predominant in uncured pork; however, more mercaptans, alkyl- sulfides, -disulfides and -trisulfides were isolated from cured pork. Ramarathnam 164 165 and Rubin (1994) reported the presence of 30 aldehydes, 20 carboxylic acids, 9 alcohols, 9 esters, 6 furans, 4 hydrocarbons, 12 ketones, a dimethylphenol, a 2- methylpyridine, a 2-methylpyrazine, 3 nitrogen-containing compounds, 31 sulfur- containing compounds, 6 nitriles and 4 nitrates in nitrite-cured hams. However, they concluded that no single group of volatile compounds was responsible for the unique cured meat flavor. Many studies have indicated that carbonyl compounds make a significant contribution to uncured meat flavor (Homstein and Crowe, 1960; Cross and Ziegler, 1965; Bailey and Swain, 1973; Shahidi at Q, 1986; Shahidi, 1989). However, the nature of cured meat flavor, which is made apparent by suppression of lipid oxidation by nitrite and which is the basic meat flavor of cooked meat, remains a mystery (Rubin and Shahidi, 1988). The presence of phospholipids in model systems containing cysteine and ribose suppressed the formation of selected Maillard reaction volatiles (Chapter 2). The stabilization of the phospholipids by the addition of antioxidants such as nitrite, a-tocopherol and tert—butylhydroquinone (TBHQ) reduced the quenching effect of unsaturated lipids (Chapter 3). Nitrite prevents lipid oxidation in cured meat products and thus may minimize the negative impact of lipids on the development of the Maillard reaction products. Therefore, the uniqueness of cured meat flavor may be due, in part, to the presence of greater amounts of Maillard reaction products, along with the formation of fewer carbonyl compounds. The objective of this study was to investigate the differences in the flavor profiles of cured and uncured hams, particularly with the respect to the relative concentrations of carbonyl compounds and Maillard reaction products. The effect of different concentrations of nitrite in the hams on the formation of selected Maillard reaction volatiles was also studied. EXPERIMENTAL Materials Hams were obtained from a local processor within 48 hr of slaughter and processed in the Meat Laboratory (Michigan State University, East Lansing, MI. Detailed were provided in Chapter 3). Sodium phosphate (monobasic and dibasic) were obtained fi'om J. T. Baker Chemical Co. (Phillipsburg, NJ). All the reagents and solvents used in this study were analytical and/or HPLC grade. mum of cooked hapr samples Hams, processed to contain 0, 150, 1500 mg nitrite /kg meat, were sliced into a 1/ " (0.6 cm) thick slices and cooked in an electric frying pan set at 350°F ( 177°C) for 5 nrin on each side. The cooked slices were allowed to cool to room temperature, and homogenized in a Waring blender (New Hartford CT). The samples were transferred to low density polyethylene pouches (8" x 10", 75 pm thickness, Whirl-Pak, Fisher Scientific Co, Fair Lawn, NJ) and stored in a freezer (- 10°C) prior to use. Collection of mvor compoands Volatile compounds were collected using a purge-and-trap procedure. Cooked hams (50 g, homogenized) were placed in a two-neck flask (500 ml, Kontes, Vineland NJ) containing 150 ml phosphate buffer (0.5M, pH 5.7). The flask was fitted with a L-shape glass joint (Kontes, Vineland NJ), and a Tenax trap (glass lined stainless steel tube containing 200 mg Tenax, 4 mm id. x 100 mm) threaded on both ends , Scientific Instrument Services, Ringoes, NJ) was 166 167 attached by a stainless steel Swagelok reducing union (Cajun Ultra-Torr® fitting, Cajun, Macedonia, OH) to the outlet of the joint. Tenax traps were conditioned overnight at 180°C with nitrogen flowing at 30 ml / min prior to volatile collection. During collection, the flask containing the ham sample was heated at 60°C in a heating mantle, while the Tenax trap was maintained at room temperature (approximately 20°C i 2°C). Volatile compounds were swept into the Tenax trap using a stream of nitrogen (30 ml /min), and the collection was continued for 1 h. At the end of the collection period the Tenax trap was directly connected to the nitrogen supply for 5 min to remove water vapor. More details of this procedure were provided in Chapter 1. Gas chromatoggaphy - mass spectrometric analysis The trapped flavor compounds were directly desorbed into a gas chromatograph - mass spectrometer (GC / MS) system using a thermal desorption unit (Short Path Thermal Desorption, Scientific Instrument Services, Ringoes, NJ). For desorption, the Tenax trap was fitted with a syringe needle and attached to the desorption unit (Figure 1). This permitted the trapped volatile compounds to be heated by the desorption heater block and desorbed from the Tenax absorbent into the injection port of the gas chamber (GC) via the shortest path possible, i.e. direct injection into the GC via a syringe (Figure 2). The trapped flavor compounds in the Tenax trap were heated at 250°C with nitrogen flushing (10 ml/min) for 5 min. The GC / MS analysis was simultaneously started with the desorption process. The flavor compounds were separated and identified using a HP 5890 gas chromatograph (Hewlett Packard Avondale, PA) interfaced with a HP 5970 MSD 168 mgUU a 3.: 352—88 2232, coma—«.5 do 5:982. 05 .5.— moaatoooa 2: 33 3:3 =2E38n 13:2; :25 team 2: do 952.3. 2a. A 2sz 929.50 . :oEmod 828:: m 958:: Uncl— Figure 2. 169 / Carrier Gas Flow ,. /- Desorption Tube r" i V‘ ‘— Desorption Tube 5 b l: l: Heater Blocks / / / r / 6 x V /JM Au“ Needle -* GC lnjection Port To GC Detector or Mass Spec A schematic diagram of the procedure used to desorb the trapped volatile compounds from the Tenax trap. The trapped flavor compounds were thermally desorbed from the Tenax trap and swept into the injection port of the GC via a needle. 170 (Mass Selective Detector) mass spectrometer. The GC / MS system was equipped with a HP 59970 Chemstation Data System. A Carbowax 20M fused silica capillary column (60m x 0.32 mm id Supelco Inc., Bellefonte, PA) was used to separate the compounds. The GC oven temperature was programmed as follows: the initial temperature was set at 40°C for 5 rrrin, the temperature was increased at a rate 2°C / min to a final temperature of 200°C, and maintained for 10 rrrin. Helium was the carrier gas with a flow rate of 10 ml/min. The temperatures of the injection port and transfer lines were. both set at 220°C. The mass spectrometer was operated in the electron impact mode with an electron energy of 70 eV and an ion source temperature of 250°C. Compounds were introduced to the ion source directly from the capillary column in the GC using an open-split interface. A continuous scan mode with a scan time of 1 sec over a mass range of 40-300 was employed. The GC / MS data were monitored, stored and analyzed using an HP Chemstation data system. Two mass spectral libraries, the IN RA Mass Spectra Library - "IN RAMASS" (Laboratoire de Recherche sue les Aromes, Dijon, France) and the NIST / EPA / MSDC Mass Spectral Database (ACS Publication Co, Washington, DC), were used to identify the compounds isolated. RESULTS AND DISCUSSION Volatile compounds isolatml from nitrite-free hams Volatile compounds were collected by the Tenax traps and directly desorbed into the GC via a thermal desorption system. The major advantage of using direct thermal desorption was the reduction of loss of trace compounds. The 171 concentrations of free amino acids and reducing sugars in meats are much smaller than those used in the model systems, thus much smaller concentrations of Maillard reaction volatiles were expected to be generated in the cooked meat. The use of the thermal desorption system was an attempt to provide a more efficient system for the analysis of the volatile compounds than the system used in the model system studies. The volatiles collected from the cooked uncured ham contained more than 75 compounds, twelve of which were identified and used to study the differences in uncured and cured meat flavor profiles. The volatile compounds identified were dominated by carbonyl compounds (Table l). Pentanal and hexanal were the primary compounds, and are derived from the oxidation of unsaturated fatty acids in meats (Cross and Ziegler, 1965). One sulfur-containing compound ethylthioacetate, was also identified. This compound reported in cured hams by Golovnya e_t Q. (1982), represented less than 1% of the total volatiles. l-Octene, the only hydrocarbon isolated from the uncured hams, was also identified in an uncured pork sample by Mottram M. (1982). Very few Maillard reaction products were identified in the ham samples. As discussed by Farmer and Mottram (19903) and confirmed in Chapters 2 and 3, lipid oxidation suppresses the formation of Maillard reaction volatiles in model systems containing cysteine and ribose. However, the failure to detect the Maillard reaction products in the uncured ham samples may be due to the lack of sensitivity of the analytical procedure. In addition, the amount of sample (50 g) could also play an important role. Ramarathnam at Q. (1993) used a similar system containing 250-400 g of ground meat in distilled water (meat-to-water ratio was 172 Table 1. Volatile compounds isolated from uncured and nitrite-cured ham by purge-and-trap method. Compound MW KI Ham (nitrite mg/kg) 0 1 50 l 500 pentanal 86 982 100 33a 33 3-methyl-1-butanol 88 1 120 100 60 37 5-hexen-2-one 98 14 l 3 100 ---- ---- hexanal 100 1090 100 36 30 benzonitrile 103 1605 ----b 100 130 ethylthioacetate 104 1792 100 105 131 benzaldehyde 106 1524 100 165 145 2,3-dimethyl-2-cyclopenten-l-one 110 1642 100 83 ---- 1-octene 1 12 1359 100 97 98 2-methyl-tetrahydrothiophene-l -one 1 l 6 l 521 l 00 ---- ---- methyl bezonitrile 1 17 191 7 ---- 100 3 36 acetyl thiazole 127 1714 ---- 100 120 benzothiazole 1 3 5 1972 ---- 100 1 8O 1-nonen-3-ol 142 1510 100 44 43 2,4-decadienal 1 52 1 8 l 6 100 3 8 l 5 decyl acetate 200 1675 100 6 5 a The numbers represent the average percent (3 replications) of the compound in the cured ham compared to the same compound isolated from the uncured ham. The calculation was accomplished by standardizing the peak area of the flavor compound from the uncured ham as 100%, and expressing the peak area of the same compound in cured hams as a percentage. When the compound was not detected in the uncured ham, the percentage was expressed by comparing to 150 (mg/kg) nitrite-cured ham (i.e. peak area standardized as 100%). b "----" : indicates not detected. 173 4:1 w/w) to collect the volatile compounds from uncured and cured pork. They, too, did not isolate any sulfur-containing compounds. They reported some nitrogen-containing Maillard reaction compounds such as pyridines, pyrazines and pyrroles, but these compounds were present at concentrations less than 0.4 part per million in the cured products. Thus, the smaller sample size used in this study did not permit the detection of compounds that are formed in only small amounts. 13mm profile of cooked cured harfi The volatile compounds isolated from the cured hams (150 and 1500 mg/kg) were also dominated by carbonyl compounds. However, their concentrations were considerably smaller than those isolated fiom the uncured hams (Table l). The concentrations of pentanal and hexanal were approximately 30-36 % of those in the uncured hams. Increasing the nitrite level in cured hams reduced the formation of the carbonyls (Table 1). Similar results were reported by Cross and Ziegler (1965) regarding the concentrations of pentanal and hexanal in cured and uncured meats. Pentanal and hexanal were present in appreciable quantities in the uncured products, but were barely detectable in the volatiles of the cured meats. Other carbonyls derived from lipid oxidation were also reduced by the presence of nitrite in cured meat (Ramarathnam Q Q, 1991a, b). Nitrite serves as an antioxidant and prevents lipid oxidation in meats during cooking and subsequent storage (Gray and Pearson, 1994). However, carbonyl compounds are still predominant in the cured meat flavor profiles, Bailey and Swain (1973) reported that the volatile compounds from ham treated with and without nitrite were qualitatively similar, but quantitatively different. The presence and absence of certain carbonyls, or the differences in their concentrations, can be a major contributing factor to the differences in the aromas 174 of cured and uncured hams. Carbonyl compounds have been implicated as significant contributors to the flavor of uncured meat, but not to cured meat flavor (Ramarathnam and Rubin, 1994). Greater amounts of benzaldehyde were isolated from cured hams compared to the uncured samples (Table 1). The quantities were 65% and 45% greater in the 150 and 1500 (mg/kg) nitrite-cured hams relative to the uncured hams. However, Ramarathnam at Q. (1993, 1994) reported smaller concentrations of benzaldehyde in cured pork than in uncured meat. The differences in the two studies may be due to the relatively small amounts present in the cured hams, thus a possible instrumental error may be involved. Two nitriles (benzonitrile and methyl benzonitrile) and two thiazoles (2- acetyl thiazole and benzothiazole) were isolated fi'om the cured hams (Table l). The amounts of these compounds increased with increasing nitrite content. Nitrile compounds were also reported in the volatiles of cooked cured pork by Mottram (1984b; 1985). The origin of these nitrogen compounds is from the reaction of fatty acids and sodium nitrite (Mottram, 1985). Benzonitrile appear to have a low odor threshold with alrnond-like aromas similar to that of benzaldehyde. Thus, nitriles may contribute to the characteristic cured meat flavor (Mottranr, 1984b). Both thiazole compounds identified in this study were previously reported in bacon aroma (Mottram, 1984a), and possess strong nutty-roasted and smoky to meaty notes (F ors, 1983). These compounds may also contribute to the unique cured meat flavor. The curing process simply modifies the composition of the volatile constituents. Cured meat flavor is believed to be the basic meat flavor (Noel _t Q, 1990), and possibly is derived fi'om Maillard reaction products. The failure to detect the selected Maillard reaction volatiles which were isolated from a model 175 system (Chapter 1) could be due to several factors. First, the sample size was inadequate to produce appreciable concentrations of volatile compounds for the flavor collection studies. The free amino acid and reducing sugar contents in meats are too small to provide sufficient Maillard reaction products for adequate analysis. Much greater concentrations of these reactants were used in the model system. Second the temperatures and time of cooking also play an important role in the formation of the Maillard reaction products in meats. Though the Maillard reaction may occur at refrigerated temperatures (Hurrel, 1982), temperature has the greatest influence on the Maillard reaction (Leahy and Reineccius, 1989; Shu e_t Q, 1989; Shaw and Ho, 1989; Ames and Apriyantono, 1994). The Maillard reaction may be initiated at temperatures greater than 90°C and the reaction rate is markedly increased as the temperature rises (Hurrel, 1982). Hams were fried at 350°F for 5 min per side; however, the internal temperature of the ham samples would be much lower than this. Thus, smaller concentrations of Maillard reaction products would be produced. Future studies of cooked meat flavor profiles should employ better and more efficient methods for collecting and identifying trace volatiles compounds in meat products. The use of the cryogenic traps to concentrate the volatile compounds is a necessity. CONCLUSION In our hypothesis, carbonyl compounds derived from lipid oxidation may suppress the formation of Maillard reaction products which contribute to the basic meat aroma. Nitrite functions as an antioxidant in cured meat, thus preventing lipid oxidation and benefiting the formation of Maillard reaction volatiles. Though, the 176 quantities of the Maillard reaction volatiles isolated from the ham samples were too small for positive identification, it is still our speculation that the inhibition of lipid oxidation will reduce quenching of the formation of Maillard reaction products. Model system studies confinned this hypothesis. However, it remains to be proven in cured meat systems. REFERENCES Ames, J. M. and Apriyantono, A. 1994. Effect of pH on the volatile compounds formed in a xylose-lysine model system. In "Thermal Generated Flavor: Maillard Microwave and Extrusion Process", T. H. Parliment, M. J. Morello and R. J. McGorrin (Ed), ACS Symposium Series 543, Washington, DC, pp 228-239. Bailey. M. E. and Swain, J. W. 1973. Influence of nitrite on meat flavor. Proceedings of Meat Industry Research Conference, American Meat Institute Foundation, Chicago, IL, pp 29-45. Cross, C. K. and Ziegler, P. 1965. A comparison of the volatile fractions from cured and uncured meat. J. Food Sci. 30: 610-614. F ors, S. 1983. Sensory properties of volatile Maillard reaction products and related compounds. In "The Maillard Reaction in Food and Nutrition", G. R. Waller and M. S. Feather (Ed), ACS Symposium Series 215, Washington, DC, pp 185-286. Golovnya, R. V, Garbuzov, V. G, Grigor'eva, 1. Ya, Zharich, S. L. and Bol'shakov, A. S. 1982. Gas chromatographic analysis of volatile sulfur- containing compounds of salted and salted-boiled pork, Nahrung 26: 89-96. Gray, J. I. and Pearson, A. M. 1994. Lipid-derived off-flavors in meat - formation and inhibition. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 116-143. 177 Gray, J. I. and Pearson, A. M. 1984. Cured meat flavor. Adv. Food Res. pp 1-86. Homstein, I and Crowe, P. F. 1960. Flavor study on beef and pork. J. Agric. Food Chem. 8: 494-498. Hurrel, R. F. 1982. Maillard reaction flavor. In "Food Flavour", I. D. Morton and A. J. MacLeod (Ed), Elsevier Scientific Publishing Co, London, UK, pp 399-437. Leahy, M. M. and Reineccius, G. A. 1989. Kinetics of the formation of alkylpyrazine, effect of pH and water activity. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 196-208. Mottrarrr, D. S. 1985. Some new observations of the flavour chemistry of cured meats. In "Progress in Flavor Research", J. Adda (Ed), Elsevier, Amsterdam, the Netherlands, pp 323-328. Mottram, D. S, Croft, S. E. and Patterson, R. L. S. 1984a. Volatile components of cured and uncured pork: the role of nitrite and the formation of nitrogen compounds. J. Agric. Food Chem. 35: 233-239. Mottram, D. S. 1984b. Organic nitrates and nitriles in the volatiles of cooked cured pork. J. Agric. Food Chem. 32(2): 343-345. Mottranr, D. S, Edwards, R. A. and MacFie, H. J. H. 1982. A comparison of the flavour volatiles for cooked beef and pork meat system. J. Sci. Food Agric. 33: 934-944. Neol, P, Briand E. and Dumont, J. P. 1990. Role of nitrite in flavor development in uncooked cured meat products: sensory assessment. Meat Sci. 28: 1-8. Ramarathnam, N. and Rubin, L. J. 1994. The flavor of cured meat. In "Flavor of Meat and Meat Products", F. Shahidi (Ed), Blackie Academic and Professional, Glasgow, UK, pp 174-198. Ramarathnam, N, Rubin, L. J. and Diosady, L. L. 1993. Studies on meat flavor. III. A novel method for t rapping volatiles components from uncured and cured pork. J. Agric. Food Chem. 41(6): 933-938. 178 Ramarathnam, N, Rubin, L. J. and Diosady, L. 1991a. Studies on meat flavor: I. Qualitative and quantitative differences in uncured and cured pork. J. Agric. Food Chem. 39: 344-350. Ramarathnam, N, Rubin, L. J. and Diosady, L. 1991b. Studies on meat flavor: 11. Quantitative approach to the estimation and differentiation of volatiles in uncured and cured beef and chicken. J. Agric. Food Chem. 39: 1839-1847. Rubin, L. J. and Shahidi, F. 1988. Lipid oxidation and the flavor of meat products. Proceedings, 34th International Congress of Meat Science Technology, Brisbane, Australia, pp 295-301. Shahidi, F. 1989. Flavor of cooked meats. In "Thermal Generation of Aromas", T. H. Parliament, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 188-201. Shahidi, F, Rubin, L. J. and D’Souza, L. A. 1986. Meat flavor volatiles: A review of the composition, techniques of analysis and sensory evaluation. CRC Crit. Rev. Food Sci. Nutr. 24(2): 141-243. Shaw, J. J. and Ho, C.-T. 1989. Effect of temperature, pH and relative concentration on the reaction of rharnnose and proline. In "Thermal Generation of Aromas" , T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 217-228. Shu, C.-K. and Ho, C .-T. 1989. Parameter effects on the thermal reaction of cysteine and 2,5-dimethyl-4-hydroxy-3 (2H)-furanone. In "Thermal Generation of Aromas", T. H. Parliment, R. J. McGorrin and C-T. Ho (Ed), ACS Symposium Series 409, Washington, DC, pp 229-241. SUMMARY AND CONCLUSIONS A series of studies were conducted to investigate the effect of lipids on the formation of Maillard reaction products in a model system containing cysteine and ribose. Maillard reaction products are major contributors to many food flavors; however, the formation of these compounds in the model systems are influenced by the presence of lipids because lipid oxidation-derived carbonyls compete with sugar-derived dicarbonyl compounds for hydrogen sulfide and ammonia. A new technique for the analysis of volatiles was developed using Tenax traps to collect volatile compounds which were subsequently desorbed into a solvent by a combination of heat and solvent extraction. Percent recoveries of five internal standards were greater than 70%. More than 90 compounds were isolated from the heated model system using this method. Thirty-one compounds were identified via the mass spectral library softwares. Of these, fifteen compounds were selected as "marker" compounds. This selection was based on literature reports of compounds whose formation in similar model systems was suppressed by the presence of lipids. The effect of the degree of unsaturation of phospholipids on the formation of Maillard reaction volatile compounds in a model system containing cysteine and ribose were studied. The extent of suppression or quenching of the formation of Maillard reaction products was related to the increase of unsaturation in the phospholipids. However, the suppressive effect was not related to the amino moieties of the phospholipids. The inclusion of aldehydes in the model system suppressed of the production of Maillard reaction volatiles, thus providing indirect evidence that thermal oxidation of lipids is involved in this suppressive efl‘ect. It is 179 180 our hypothesis that the stabilization of lipids in meat products, such as through the use of nitrite as a curing adjunct, could reduce the effect of lipids and/or their oxidation carbonyl products on the development of flavor compounds via the Maillard reaction. The stabilization of lipids could contribute, in part, to the difference in the flavors of cured and uncured meats. The influence of nitrite-stabilized lipids on the formation of Maillard reaction volatiles in a model system containing cysteine and ribose was investigated. Nitrite stabilization of the phospholipids was confirmed by the formation of N-nitrosomorpholine on heating the phospholipids isolated from hams with morpholine. A significant difierence (p < 0.005) in the quantities of N- nitrosomorpholine formed between cured and uncured hams was obtained. The production of Maillard reaction volatiles in the model systems with phospholipids from the cured hams was less suppressed than by phospholipids from the uncured hams. However, there was no apparent difference in the formation of Maillard reaction products in the model systems with phospholipids from two nitrite-cured (150 and 1500 mg/kg) hams. Results indicated that nitrite prevents lipid oxidation and subsequently minimizes the extent of quenching of the Maillard reaction in the model systems. The addition of antioxidants to a model system containing cysteine, ribose and phosphatidylcholine also resulted in less suppression of the Maillard reaction. Thus, it may be concluded that the importance of nitrite in the formation of cured-meat flavor is due, in part, to the control of lipid oxidation and its possible suppression of the development of Maillard reaction volatile compounds. The total flavor profiles of cured and uncured hams were studied. A thermal desorption system linked to GC / MS was employed to prevent the loss of trace compounds. Carbonyl compounds were predominant in the flavor profiles of both 181 cured and uncured hams. However, less lipid oxidation was observed in the cured hams. Two nitriles and two thiazole compounds were identified in the cured ham flavor. It is our hypothesis that carbonyl compounds derived from lipid oxidation may suppress the formation of Maillard reaction products which contribute to basic meat aroma. Nitrite functions as an antioxidant in cured meat, thus preventing lipid oxidation and benefiting the formation of Maillard reaction volatiles. Though the quantities of the Maillard reaction volatiles isolated fiom the ham samples were too small for positive identification, it is still our hypothesis that the inhibition of lipid oxidation will reduce the quenching of the formation of Maillard reaction products. Model system studies confirmed this hypothesis. It remains to be proven in cured meat systems using a more sensitive analytical system. FUTURE RESEARCH The effect of lipids on flavor development in a Maillard reaction model system containing cysteine and ribose was studied. Results indicate that lipids influence the formation of Maillard reaction volatile compounds, and the extent of suppression is associated with the degree of unsaturation of the lipids. The amounts of Maillard reaction products formed were reduced in model system containing aldehydes. These results demonstrated that lipids suppress the formation of Maillard reaction volatile compounds. However, this study revealed other issues that require some further investigation. 1. Future studies should employ a more sensitive system for volatile compounds to ascertain whether heterocyclic compounds identified in the model systems (Chapter 2 and 3) are present in cured and uncured meats. Since relatively low concentrations of amino acids and reducing sugars are present in foods compared to those used in the model systems, a more efficient headspace collection method is necessary. A cryogenic trap linked to the injection port of GC may prevent the loss of volatile compounds from Tenax traps during desorption. Using a supercritical fluid to extract the volatile compounds fi'om food systems may be another alternative for aroma research in the future. 2. This study focused on cured hams. Other investigations could be carried out with meat products of higher fat content such as cured and uncured pork belly. The higher fat contents would permit the attainment of higher temperatures 182 183 during cooking and might lead to the generation of greater concentrations of Maillard reaction products. . It is well established that the reaction between sugar, amino acids and creatinine in meats lead to the formation of carcinogenic / mutagenic heterocyclic aromatic amines. These compounds generally are formed from pyrazines and pyridines that are produced by the Maillard reaction. This study and those of Farmer and Mottram (1990a) revealed the presence of sulfur-containing heterocyclic compounds in model systems containing cysteine and ribose. It remains to be determined whether heterocyclic aromatic amines containing sulfur heteroatoms are produced in cooked meat products such as fiied ground beef. APPENDIX 184 List of authentic flavor compounds used as standards. 2-Acetyl heterocycles: 2-acetylpyrrole 2-acetylfuran 2-acetythiophene 2-acetylthiazole 2-acetylpyrridine 2-acetylpyrazine Thizoles: thiazole 4-methylthiazole 4,5-dimethylthiazole 2,4,5-trimethylthiazole 2-isobutylthiazole 2-ethoxythiazole Methylated pyrazines: 2-methylpyrazine 2,3-dimethylpyrazine 2,5-dimethylpyrazine 2,6-dimethylpyrazine 2,3,5-trimethylpyrazine 2,3,5,6-tetramethylpyrazine Miscellaneous pyrazines: 2-ethy1pyrazine 2,3-diethylpyrazine 3-ethy1-2-methylpyrazine 2-methoxypyrazine 2-methoxy-3methylpyrazine 2-isobutyl-3 -methoxypyrazine HICHIGAN STATE UNIV. LIBRARI l”11111111111111”WWI 111111ll11111111111111 31293014172294