I'llllll lllllfllllllll 1:22;; 1 beg/i '\ :- fi—w This is to certify that the dissertation entitled THE EFFECT OF PROCESSING VARIABLES ON THE SAFETY AND ACCEPTABILITY 0F SMOKED GREAT LAKES NHITEFISH presented by Susan L. Cuppett has been accepted towards fulfillment of the requirements for Ph . D . degree in Food Science ///ém/ Major wry: sor Date 4/11/85 MS U i: an Afiimativc Action/Sq ual Opportunity Institution 0-12771 MSU RETURNING MATERIALS: Place in book drop to LIBRARIES remove this checkout from 4--3--L your record. FINES will be charged if book is returned after the date stamped below. fi?k,f ‘ 1??§ CF? 2‘ “ Due 00"? 99.99% ' THE EFFECT OF PROCESSING VARIABLES ON THE SAFETY AND ACCEPTABILITY OF SMOKED GREAT LAKES WHITEFISH By Susan L. Cuppett 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 Human Nutrition 1985 ©1986 SUSAN LEA CUPPETT All Rights Reserved ABSTRACT THE EFFECT OF PROCESSING VARIABLES ON THE SAFETY AND ACCEPTABILITY OF SMOKED GREAT LAKES NHITEFISH By Susan L. Cuppett The effect of various levels of salt concentrations (water-phase) on the lipid stability and the botulinal safety of smoked whitefish was investigated. A 4 percent (water-phase) salt concentration prevented the outgrowth of Clostridium botulinum type E spores in smoked whitefish held for 42 days at 27°C. The addition of nitrite to smoked salted (4%) whitefish prevented spore outgrowth for 63 days at 27°C. Unsalted smoked whitefish prepared with or without nitrite was significantly (P<0.0l) less rancid after 22 days of refrigerated storage than smoked whitefish containing salt concentrations (water-phase) of 2, 4 or 6 percent. However, sensory evaluations of these treatment samples throughout the refrigerated storage period indicated a significant (P<0.05) panelist preference for smoked whitefish containing 4 percent salt (water-phase). The effect of smoke type (woodsmoke or liquid) and the level of liquid smoke on lipid stability and organoleptic acceptability of smoked whitefish was investigated. Hoodsmoke-treated whitefish prepared with or without nitrite Susan L. Cuppett had significantly (P<0.0l) lower levels of rancidity than fish treated with increasing levels of liquid smoke during 22 days of refrigerated storage. Sensory evaluation of the whitefish samples treated with woodsmoke or liquid smoke indicated a significant (P<0.05) panelist preference for the woodsmoke-treated samples on day 0. However, by the l4th day of refrigerated storage, panelists were unable to distinguish differences between sample treatments. Nitrite was shown to be an effective antioxidant and antibotulinal agent in smoked whitefish. In addition, nitrited whitefish did not contain any detectable levels of volatile N-nitrosamines or N-nitrosothiazolidine carboxylic acid. Therefore, the addition of nitrite in combination with four percent salt (water-phase) can be recommended for use in the production of smoked whitefish. To my mother for her continued support of this endeavor and to Dr. J.F. Price for his friendship ii ACKNOWLEDGMENTS The author wishes to express her appreciation to her major professor, Dr. J.I. Gray, for his guidance, sugges- tions, encouragement and support throughout the research program and for his assistance in the preparation of the dissertation. Appreciation is expressed to Dr. A.M. Booren, Dr. A.M. Pearson, Dr. B.R. Harte and Dr. M.R. Bennink for serving on the guidance committee. Special thanks to Dr. J.J. Pestka for his guidance and cooperation during the microbiological evaluation of the smoked whitefish. Thanks are also extended to Ms. M. Dynnik for her patience and technological support. The author would also like to thank Mr. w.G. Ikins, Ms. M.L. Lee and Ms. C. Kutil for their assistance during the preparation of the smoked whitefish. The author also wishes to express her gratitude to the Michigan Fish Producers Association, Manistique, M1 for their generous support of this research project. TABLE OF CONTENTS LIST OF TABLES. LIST OF FIGURES INTRODUCTION. LITERATURE REVIEW . Michigan Fishing Industry . . Regulation of Smoked Fish Production. The Salting/Brining of Fish . . . Incidence of Clostridium botulinum Type E in Smoked Whitefish . . Factors Affecting the Outgrowth of C. botulinum Type E Spores Salt. . . Cooking/Liquid Smoke. Temperature . Nitrite . Sorbate . . Hater Activity and pH Mechanism of Lipid Oxidation. . Measurement of Lipid Oxidation. Lipid Oxidation in Muscle Foods . . . . Lipid Oxidation in Red Meat and Poultry Lipid Oxidation in Fish . Mechanism of Narmed- over Flavor Development in Muscle Foods. . . Narmed- -over Flavor in Red Meats and Poultry . Harmed- over Flavor in Fish. . . Antioxidant Role of Nitrite in Meat Systems Mechanism of N- nitrosamine Formation. . N- Nitrosamine in Fish Products. . N- Nitrosothiazolidine in Smoked Meats Smoking of Meat Systems Woodsmoking Process . . Smoke Absorption of Fish. Composition of Noodsmoke. . Sensory Attributes of Smoked Foods. Antimicrobial Activity of Noodsmoke Antioxidant Activity of Noodsmoke Liquid Smoke in Foods . . . Functionality of Liquid Smokes. iv Page vii Flavor Characteristics of Whitefish. METHODS AND MATERIALS. Materials. Experimental . . . Salt Uptake by Whitefish . . The Role of Nitrite, Salt and Sorbate. on the Incidence of C. botulinum in Smoked Whitefish. . . . The Role of Salt in the Stability of Smoked Whitefish . . The Role of Smoke Type and Level on the Stability of Smoked Whitefish. . . The Evaluation of Lipid Oxidation in Baked Whitefish by TBA Number and Volatile Analysis . . Methods of Analysis. . . Assay for Clostridium botulinum Type E Toxin Proximate Analysis . . Sorbate Analysis Water Activity . . Thiobarbituric Acid Test (TBA) Lipid Extraction and Fractionation Fatty Acid Analysis. N- Nitrosamines Phenol Analysis. . Gas Chromatography of. the Volatiles from Baked Whitefish. . GC- MS Analyses of the Baked Whitefish Volatile Extract. . . . Statistical Analysis RESULTS AND DISCUSSION The Effect of Time, Filleting and Brine Salt Concentration on Salt Uptake in Whitefish. . . The Effect of Salt, Nitrite and Sorbate on Toxin Production by Clostridium botulinum Type E Spores in Smoked Whitefish . . The Role of Salt Concentration (wp) on the Rate of Toxin Production by C. botulinum Type E Spores in Smoked Whitefish. . The Effect of Residual Nitrite on Toxin Production by C. botulinum Type E Spores The Effect of Residual Sorbate on Toxin Production by C. botulinum Type E Spores The Effect of pH on Toxin Production by C. botulinum Type E Spores . . The Effect of Water Activity (Aw) on Toxin Production by g. botulinum Type E Spores V Page 60 63 63 64 9O 94 97 101 103 105 Effect of Salt Level and Nitrite on the Lipid Stability and Organoleptic Acceptability of Smoked Whitefish. . . . Chemical Analyses . . Residual Nitrite Levels of Smoked Whitefish During Storage. . Stability of Smoked Whitefish Lipids During Storage Sensory Analysis of Smoked Whitefish Samples: Effect of Smoke Type and Level on the Lipid Stability and Organoleptic Acceptability of Smoked Whitefish. . Chemical Analyses Stability of Whitefish Lipids During Storage: Sensory Analysis of Smoked Whitefish. Evaluation of Lipid Oxidation in Baked Whitefish: SUMMARY AND CONCLUSIONS . PROPOSAL FOR FUTURE RESEARCH. APPENDIX. BIBLIOGRAPHY. vi Page 110 110 117 120 125 132 132 136 152 158 161 162 171 Table 10 11 LIST OF TABLES Summary of the growth characteristics of Clostridium botulinum types A and E. Unsaturated fatty acid content of lipid in some muscle foods Compounds identified in curing smoke classified by chemical group. Cook cycle used in the smoking of whitefish. Brine composition for determining the effect of salt concentration on the rate of salt uptake by whitefish fillets during a 14 hour brining period at 4°C.. . . . . Brine composition for determining the effect of salt, nitrite and sorbate on the outgrowth of Clostridium botulinum type E spores in smoked whitefish. . . . . . . . Brine composition for determining the effect of salt and nitrite with ascorbate on the develop- ment of oxidative rancidity in smoked whitefish. Brine compositions for determining the effect of smoke type and level and nitrite with ascorbate on the development of oxidative rancidity in smoked whitefish. Salt and moisture levels in whole, eviscerated and filleted whitefish brined in 30° salometer brine for different times. . . . . . . . Salt, moisture and salt (water-phase) concen- trations of whitefish fillets brined for 14 hours in different degree salometer brines Presence of toxin in temperature abused (27°C) smoked whitefish inoculated with Clostridium botulinum type E spores. . . . vii Page 14 34 52 66 67 69 72 74 83 87 91 Table Page 12 Percent salt (water-phase) of smoked whitefish brined in different degree salometer brines prepared with or without nitrite or nitrite and sorbate. . . . . . . . . . . . . . . . . . . . . 95 13 Residual nitrite levels in smoked whitefish brined in different degree salometer brines containing nitrite and stored at 27°C. . . . . . 98 14 Residual sorbate levels in smoked whitefish brined in different degree salometer brines with nitrite and sorbate and stored at 27°C. . . 102 15 Effect of storage at 27°C on the pH values of smoked whitefish brined in different degree salometer brines . . . . . . . . . . . . . . . . 104 l6 Water activity levels in smoked whitefish brined in different degree salometer brines prepared with or without nitrite or nitrite and sorbate and stored at 27°C . . . . . . . . . . . . . . . ‘07 17 Percent salt (water-phase) of smoked whitefish brined in different degree salometer brines prepared with and without nitrite. . . . . . . . 112 l8 Percent fat in smoked whitefish brined in different degree salometer brines prepared with or without nitrite . . . . . . . . . . . . . . . ‘13 l9 Fatty acid composition of the total, triglycer- ide and phospholipid fractions of whitefish lipids . . . . . . . . . . . . . . . . . . . 114 20 Residual nitrite levels of smoked whitefish brined in different degree salometer brines containing nitrite . . . . . . . . . . . . . . . 118 21 TBA values of smoked whitefish brined in different degree salometer brines prepared with or without nitrite . . . . . . . . . . . . . . . 12] 22 Average sensory scores at day 0 for smoked whitefish samples brined in different degree salometer brines . . . . . . . . . . . . . . . . 127 23 Average sensory scores at day 7 for smoked whitefish brined in different degree salometer brines . . . . . . . . . . . . . . . . . . . . . 128 viii Table Page 24 Average sensory scores at day 14 for smoked whitefish brined in different degree salometer brines. . . . . . . . . . . . . . . . . . . . . 129 25 Percent salt concentration (water-phase) of brined whitefish treated with various levels of liquid smoke or smoked with woodsmoke . . . . . 133 26 Percent fat in brined whitefish treated with various levels of liquid smoke or smoked with woodsmoke . . . . . . . . . . . . . . . . . . . 134 27 Residual nitrite levels of brined whitefish with or without nitrite and treated with varying levels of liquid smoke or smoked with woodsmoke . . . . . . . . . . . . . . . . . . . 135 28 Effect of smoke type and level on the TBA values of whitefish brined with or without nitrite....................137 29 Total phenol content of whitefish smoked with different levels of liquid smoke or smoked with woodsmoke. . . . . . . . . . . . . . . . . 14] 30 Relative percent of phenols tentatively iden- tified in whitefish prepared with 1.4% liquid smoke or smoked with woodsmoke. . . . . . ‘45 3l Average sensory scores at day 0 for brined whitefish samples treated with varying levels of liquid smoke or smoked with woodsmoke. . . . 148 32 Average sensory scores at day 5 for brined whitefish samples treated with varying levels of liquid smoke or smoked with woodsmoke. . . . 149 33 Average sensory scores at day 14 for brined whitefish samples treated with varying levels of liquid smoke or smoked with woodsmoke. . . . 15° 34 The TBA number, hexanal changes and the sensory scores of baked whitefish refrigerated for O, 1, 2 and 3 days . . . . . . . . . . . . . . . . 153 ix LIST OF FIGURES Figure Page 1 The effect of brining time on salt uptake in whole, eviscerated and whitefish fillets brined in 30° salometer brine. . . . . . . . . . . . . 84 2 The effect of brine salometer concentration on salt uptake of whitefish fillets brined for 14 hours at 4°C. . . . . . . . . . . . . . . . . . 88 3 The effect of brine salometer and nitrite on salt uptake in whitefish fillets brined for l4 hours at 4°C. . . . . . . . . . . . . . . . . . 89 4 Gas chromatographic analysis of whitefish smoked with woodsmoke . . . . . . . . . . . . . 143 5 Gas chromatographic analysis of smoked white- fish prepared with l.4% liquid smoke in the brine . . . . . . . . . . . . . . . . . . . 144 6 Gas chromatographic analysis of the volatiles from baked whitefish. . . . . . . . . . . . . . 155 INTRODUCTION Considerable amounts of processed smoked foods including meats, cheese and fish are consumed in the United States. These products are smoked by traditional wood smoking methods or by the use of liquid smokes (Daun, 1979). Smoking of foods was originally a means of food preservation. However, smoking of foods today is done primarily to add smoke flavor and color to the product (Sink, 1979). The production of smoked fish in the state of Michigan requires that the finished product contain 5% salt in the water-phase portion of the finished product (Stachiw et al., 1984). This stipulation was established to prevent botulism. Botulism is a food borne disease caused by the ingestion of the potent neurotoxin produced by the bacterium Clostridium botulinum. Because of the widespread distribu- tion of C. botulinum in marine and freshwater environments, contamination of fish products by this organism is difficult to avoid (National Academy of Sciences, 1981). The most prevalent type of g. botulinum in the marine environment is type E (Schmidt et al., 1962; Pelroy et al., 1982). Type E is non-proteolytic and can grow at refrigerator temperatures. Therefore, it can be present in refrigerated foods without any obvious signs of putrefaction and its associated off-odors (Eklund, 1982). The safety of smoked fish relies on the ability of the salt concentration in the water-phase of the finished product to inhibit the outgrowth of C. botulinum type E spores. Since the consumer is becoming more aware of the relationship between dietary salt intake, hypertension and its associated health problems, long term acceptability of this salty product could be limited in today's society. Warmed-over flavor (WOF) is a term used to describe the rapid development of oxidative rancidity occurring in meats that are cooked, held at refrigerator or freezer temperatures, and then reheated before serving. WOF is important in red meats and poultry. Research involving red meats and poultry indicates that the phospholipid fraction of the animal fat is primarily responsible for the develop- ment of WOF (Igene et al., 1979). In these products, the phospholipids contain a greater percentage of the unsaturated fatty acids than the triglyceride fraction. It has been shown that the rate of lipid oxidation is affected by the degree of unsaturation occurring in the lipid system (Dawson and Gartner, 1983). WOF is anticipated to occur in smoked fish products since fish lipids contain a significant amount of long chain polyunsaturated fatty acids (Khayat and Schwall, 1983). In addition, smoked fish can be held at refrigerator temperatures for as long as 14 days prior to consumption. A primary catalyst of WOF has been shown to be the heme compounds of the muscle tissue (Pearson et al., 1977) or free iron released from the meat pigments during cooking (Igene et al., 1979). It has been suggested that during cooking, the heme-containing proteins denature and unfold, exposing the iron molecule. The free iron molecule is a strong catalyst of lipid oxidation. Nitrite has been shown to be an effective inhibitor of WOF development in both red meats and poultry, but it has not been tested in fish products. Nitrite is believed to bind the iron in the heme-containing pigments of meats forming a non-reactive species, thus preventing the accelerated rate of lipid oxidation normally seen in the develOpment of WOF. The addition of nitrite to smoked fish has been clearly demonstrated to allow a decrease in the salt concentration while maintaining safety against botulinal toxin production in temperature abused products (Pelroy et al., 1982). However, the addition of nitrite to fish products could result in the formation of N-nitrosamines from the reaction of the added nitrite and the free amines of fish (Shewan, 1951; Castell et al., 1971; Golovyna, 1976). Research has shown that N-nitrosamines can be formed in fish products, although the data are extremely variable (Malins et al., 1970; Sen et al., 1970; Fazio et al., 1971; Gadbois et al., 1975). The formation of N-nitrosamines in fish products occurs primarily in saltwater species. This is understand- able, since the saltwater species contain high levels of trimethylamine and trimethylamine oxide, which readily break down to produce formaldehyde and dimethylamine. Dimethylamine is readily nitrosated to produce N-nitroso- dimethylamine (NDMA) (Dyer and Mounsey, 1949; Spinelli and Koury, 1979; Sikorski and Kostuch, 1982). The purpose of this project was to investigate the effect of brining and brine salometer levels on the salt uptake in whitefish and to study the feasibility of lowering the level of salt in the finished smoked fish product, while maintaining organoleptic acceptability and safety against botulinal toxin production. This required investigating the antibotulinal and the antioxidant activity of nitrite in the smoked fish product. Another objective of this study was to investigate the effect of smoke (wood and liquid), with or without the addition of nitrite, on the organoleptic acceptability of smoked fish. Before advocating the use of nitrite in smoked fish, it was necessary to assess whether the addition of nitrite would result in the formation of N-nitrosamines in the fish products. Research was conducted to correlate the TBA values of baked whitefish, stored in the refrigerator, with the sensory evaluation of rancidity of the product. Finally, this study involved profiling and identifying the major flavor volatiles in baked whitefish, particularly as they pertain to lipid oxidation. LITERATURE REVIEW Michigan Fishing Industry The state of Michigan is surrounded by four of the five Great Lakes; Superior, Michigan, Huron and Erie (Stachiw et al., 1984). Therefore, it is understandable that Michigan has a strong fishing industry. Michigan's first commercial fishing industry was established on Lake Erie about 1820. By 1840, fisheries were established on Lakes Michigan and Huron (Bulletin E-lOOO, 1977). Whitefish was the mainstay of these early fisheries (Bulletin E-lOOO, 1977), and whitefish has remained the major utilized species in the Great Lakes, with an average annual catch of 3.6 million pounds (Stachiw et al., 1984). Approximately two- thirds of the whitefish caught in Michigan are from Lake Michigan. Lake Superior and Lake Huron each produce about one half of the remaining catch. These figures do not account for the catch taken by the Indian tribes along the Lakes by Canadian commercial catches. Their catch is estimated to account for approximately 1.9 million pounds (Stachiw et al., 1984). Approximately half of all whitefish caught commercially are eventually processed and sold as a smoked product (Booren, 1984). Regulation of Smoked Fish Production in Michigan Fish processors are being confronted with an increasing demand for higher quality product by the consumer. Because the fish processing industry is a food industry, it is subject to food regulations at both federal and state levels (Stachiw et al., 1984). In Michigan, the production of smoked fish is controlled by Michigan Regulation 541 (Michigan Department of Agriculture, 1965). This regulation requires that the fish processor must (I) cook all product to an internal temperature of 180°F in the coldest part of the fish, and maintain that temperature for not less than 30 minutes; (2) insure that all smoked fish product contains in its water phase portion, a salt content of not less than 5%; (3) maintain the product under refrigeration (36°F) at all times, excluding the time necessary for smoking operations; and (4) must provide a label on the finished product which carries a warning state- ment indicating smoked fish should not be sold or consumed after 14 days of refrigerated storage (Michigan Department of Agriculture, 1965). The Salting/Brining of Fish Salting is one of the oldest methods used for the preservation of meats and, in particular, fish. Tradition- ally, the high levels of salt used in smoked fish acted to preserve the product by lowering its water activity (Deng, 1977). Today, because of the availability of refrigeration, there is less need for heavy salting and the levels of salt and smoke used are primarily for flavor acceptability (Deng, 1977). There is little information on the rate of salt pene- tration into muscle. Kormendy and Gartner (1958) reported that the diffusion of salt into meat depended on the ratio of the amount of brine to the amount of meat within certain limits, and on the duration of the curing time. They also noted that during prolonged curing, the amount of salt penetration into the muscle would cease, although equilibrium had not been established. They theorized that the swelling that occurred during salting causes the external layers of the muscle to become closed to further salt penetration. Crean (1961) studied the problem of salt penetration in fish muscle. He reported that during dry-salting, the average water and salt contents of fish muscle were inversely and linearly related, indicating that the rate of salt uptake was in constant ratio with the rate of water loss. This led to the postulation that the salt and water exchange was primarily confined to a region of "denatured" muscle which acted to createla"front" at which denaturation occurred. This front could move into the muscle and it was behind this front that the bulk of the salt and water exchange occurred. A final suggestion by Crean (1961) was that the driving force for the salt penetration into fish muscle might be the difference in the concentration between the average internal fish brine and the ambient brine. Hamm (1960) studied the factors which affected the amount of swelling occurring in fish muscle during brining. He reported that the amount of tissue swelling depended upon the manner in which the brine immersion was carried out. If the strength of the brine was increased gradually, the muscle's weight gain was much larger than if the brining occurred in full strength brine. During the brining equilibration, one of two events could occur. These were: a) at low or intermediate salt concentrations, the water is transferred from the brine into the muscle and the muscle swells, or b) at salt concentrations beyond a certain point, water is transferred from the muscle to the brine. If the salt content of the brine is high enough, the muscle protein will coagulate or salt-out. Final equilibrium between the brine of a given concentration and the muscle which is immersed in the brine has been taken as an equality of the salt concentration of the brine with the total water inside the muscle (Del Valle and Nickerson, 1967a). The factors which affect the rate of salt penetration in fish muscle have been studied. However, most of the research has been directed towards producing a dried salted fish such as found in countries where fish is a staple in the diet, i.e., developing countries (Del Valle and Nickerson, 1967; Deng, 1977; Doe et al., 1982; Poulter et al., 1982). Del Valle and Nickerson (1967) in a series of studies investigated the equilibrium considerations that apply to the salting and drying of fish. They found that the coefficient for the penetration of salt into fish muscle (swordfish) was not constant but depended upon the salt concentration of the brine and the temperature. In a second set of studies, Del Valle and Nickerson (1967) determined that the primary variables affecting salt uptake by the fish muscle were: 1) salt concentration of the brine, 2) volume of the muscle, 3) distribution coefficient of the salt between the muscle volume and the brine and 4) distribution coefficient of the salt between the muscle tissue water and the brine. Secondary variables affecting the salting equilibria were functions of the salt concentration in the brine. Deng (1977) found that salt penetration into whitefish muscle can be affected by the storage conditions prior to the actual salting. Freezing acted to increase the rate of salt penetration. Fish that had been processed fresh had a salt uptake rate of 0.006 g salt/g sample/min and a rate constant of 0.018 min". Fish frozen for 1 week had a salt uptake rate of 0.014 and a rate constant of 0.029. However, fish frozen for 3 and 5-9 weeks had salt uptake rates of 0.011 and 0.009 and rate constants of 0.025 and 0.018, respectively. These changes in salt penetration closely 10 followed changes in the amount of extractable actinomyosin in the muscle, indicating that the rate of salt uptake depends on the degree of denaturation of the fish proteins that occurred during freezing. Deng (1977) reported that when the brine concentration used with fresh fish exceeded 20% salt, the water migrated from the muscle into the brine, and the reverse occurred in brine less than 15% salt. In fish frozen for 2 months, a brine of 25% or more salt was required before the water was transferred from the muscle to the brine. Fish could be brined in salt concen- trations up to 20% and still have salt penetration. Incidence of Clostridium botulinum Type E in Smoked Whitefish Botulism is a foodborne disease caused by ingesting the potent neurotoxin produced by the bacterial genus Clostridium botulinum. C. botulinum is a rod shaped, anaerobic, spore-forming bacterium of which there are 7 types (Kautter and Lynt, 1978). The toxins of all seven types of C. botulinum are probably toxic to man, but in the United States only types A, B and E are epidemiologically important. This is a result of the natural distribution of the g. botulinum types in the environment, the food habits of various ethnic groups and the method(s) of food processing used, especially prior to consumption (Pivnick and Bird, 1965). 11 Type E is the most prevalent type of C. botulinum in the marine environment, except in Southern California where type A predominates (Schmidt et al., 1962; Pelroy et al., 1982). C. botulinum type E was first isolated and identified from Russian sturgeon in 1936 (Kautter, 1964). C. botulinum type E has never been isolated in the southern hemisphere (Insalata et al., 1967). Epidemiological evidence indicates that C. botulinum type E occurs in significant numbers only in latitudes higher than 40° North. Type E C. botulinum occurs most frequently in Northern Japan, British Columbia, Alaska, the Soviet Union and Western Europe. It occurs less frequently in other areas such as the Pacific Northwest, along the sub-Arctic and Arctic Perimeters, and it may occur in such areas as the Mediterranean, the Gulf of Mexico and the Great Lakes (Johannsen, 1965). In 1957, it was established that C. botulinum type E was distributed in the marine sediments of the Canadian Pacific Coast (Johannsen, 1965; Emodi and Lechowich, 1969). The organism was not identified in the Great Lakes until 1960, when a botulinal outbreak was traced to smoked ciscoes (Christiansen et al., 1968; Pace et al., 1972; Pace and Krumbiegel, 1973). Further evidence of the presence of C. botulinum type E in the Great Lakes was provided in 1963 when two additional outbreaks of type E botulism were traced to smoked fish that had been processed in the Great Lakes area (Bott et al., 1966; Christiansen, 1968; Pace 12 et al., 1972). Bott et a1. (1966) reported that fish caught in Lakes Erie, Superior, Huron and Michigan had C. botulinum spores and vegetative cells in their intestinal tract, at levels of contamination of 1%, 1%, 4% and 9%, respectively. They also noted that 56% of the 835 fish samples taken from the Green Bay area of Lake Michigan harbored C. botulinum type E. Pace et a1. (1968) examined 1071 whitefish chub samples throughout the eight stages of processing in order to demon- strate the frequency or source of C. botulinum contamination in smoked fish products. They tested samples from different stages aboard ship, the routine in the smoking plant, and on display in retail cases. They reported that 13-14% of freshly caught, eviscerated chubs were contaminated. The highest level of contamination (20%) was found among chubs sampled at the brining stage of processing. Levels of contamination in the chubs at processing stages before smoking ranged from 6-14%. After smoking at 180°F for 30 minutes, 10 out of the 858 samples tested were found to be contaminated. Nine of these samples contained C. botulinum type E, while one had type B spores. Pace et a1. (1972) attributed post smoking contamination to the processor's practice of taking the cooked product back into the raw processing area before packaging. Regardless of process treatment, this problem will not change and must be corrected by improved handling practices. 13 The differences in the growth requirements of Clostri- dium botulinum types A and E are summarized in Table l. The types of C. botulinum can be characterized as proteo- lytic or nonproteolytic. The proteolytic types (A, B and F) are capable of producing putrefactive growth and its associated off-odors. They have a minimum required growth temperature of 10°C (50°F) and growth is inhibited by pH — levels of less than 4.6. These types of C. botulinum are the most heat resistant. High water-phase salt concentra— tions (8-9%) and water activities of 0.93 are necessary for ~ their inhibition (Eklund, 1982; Sperber, 1982). The non-proteolytic types of C. botulinum (B, E and G) are more sensitive to heat. They are inhibited by pH levels of less than 4.6, water-phase salt concentrations of 5-6% - and water activities of 0.96. However, they can grow at - temperatures as low as 3.3°C (38°F), and because they are non-proteolytic, their growth is not accompanied by detectable putrefaction and off-odors (Eklund, 1982). Fresh or frozen fish products have never been implicated in any recorded human botulism outbreaks in the United States. There are several reasons for this: (1) most fish products are cooked prior to consumption; (2) the endogenous microflora causes rapid quality deterioration, preventing consumption of the fish product; and (3) the spoilage organisms usually outnumber C. botulinum and, in some cases, their proteolytic enzymes inhibit and/or inactivate the 14 Table 1. Summary of the growth characteristics of Clostri- dium botulinum types A and E.a Characteristic Type A Type E Proteolytic activity + - Heat-resentant spores + - Minimum rowth temperature °F§ 50 38 (°C) 10 3.3 Minimum water activity 0.94 0.97 Salt, water-phase (8-9%) (5-6%) Minimum pH 4.6 4.6 Maximum Eh (mV) -250 —250 aFrom Sperber (1982). 15 botulinal toxin (Eklund, 1982). Preservation of fish by salting and/or smoking dates back into antiquity. The development of refrigeration has 1 resulted in modifications of the traditional salt-smoking procedures. There is less need for the heavy salt and , smoke levels, and the product does not need to be so severely dehydrated (Eklund, 1982). The hot-process smoking procedures used today inactivate the endogenous spoilage microflora and/or inhibit their growth. However, the hot-process smoking procedures do not destroy even the most heat sensitive of the C. botulinum spores. This can possibly lead to botulinal outbreaks (Eklund, 1982). Most of the reported botulinal outbreaks have been traced to a lack of under- standing by both the consumer and the retailer of the proper handling procedures for this product. Salted smoked fish must be handled in the same manner as any perishable food product (Morbidity and Mortality, 1963). Factors Affecting the Outgrowth of Clostridium botulinum Type E Spores in Smoked Whitefish The safety of salted-smoked fish products against the outgrowth of C. botulimum spores is dependent on the inhibitory effects of the salt and the cooking parameters - used during processing. Other factors which affect the rate of spore outgrowth are storage temperature, water ~ activity, pH of the product and whether any antibotulinal 16 agents such as nitrite or sorbate are used (Sperber, 1982). Salt Abrahamsson et a1. (1966) reported that the lowest level of salt necessary to inhibit toxin production by C. botulinum type E spores varied with incubation temperature and length of storage. The inhibitory effect of salt on toxin production was greatest at lower temperatures. Using Robertson's meat medium inoculated with 105 spores/g, they reported that a salt concentration of 4.5% was completely inhibitory to C. botulinum type E toxin formation even at optimal temperature (30°C. Cann and Taylor (1979) reported that in naturally contaminated hot-smoked trout and mackerel, a minimum of 2.5% water-phase salt concentration was needed to prevent toxin production for 30 days by g. botulimum type E spores. However, when processed fish were inoculated with 103 9 spores/g, vacuum packaged and stored at 10°C and 20°C, a minimum of 3.0% water-phase salt concentration was necessary to prevent toxin production for 30 days and 1 day, respec- tively. Roberts and Ingram (1973) studied the effects of various levels of pH, salt and sodium nitrite on the inhibi- tion of C. botulinum types A, B, E and F vegetative cells. Type E vegetative cells were the most sensitive to each of the variables tested (pH, salt and nitrite) and to combina- tions of these variables. Using a medium containing trypticase, bacto-peptone, yeast extract and cysteine HC1, 17 C. botulimum type E spores were added at a level of 105 vegetative cells/20 m1 of medium and incubated at 35°C for 3 months. The results indicate that a 4% salt concentration, a nitrite level of 156 mg/kg or a pH of 5.4 inhibited toxin - production by the type E vegetative cells. CookingjLiquid Smoke The survival of g. botulimum type E spores during the cooking and smoking of chub loin muscles was studied by Christiansen et a1. (1968). Raw chub loin muscles, which had been brined in a 300 salometer brine (7.8% salt), were inoculated with 106 spores of C. botulinum type E per gram of muscle tissue. The fish were cooked to an internal temperature of 180°F and held for 30 minutes. After proces- sing, the chubs were sealed under vacuum in triple laminated plastic bags and incubated at room temperature (20-25°C) for ‘ 7 days. Due to the variability of brine uptake, the tested samples had salt contents that ranged from 1.17-5.06%. Toxin was found in all samples containing less than 2.75% ~ salt, but spores survived in all the samples tested. Christiansen et a1. (1968) also tested the effect of a moist atmosphere during cooking on the survival of C. botulinum type E spores. Using steam, hot air or a combina- tion of the two, unbrined chubs that had been inoculated with 10° spores/g were heated to 180°F and held for 30 minutes. The samples were sealed in plastic and stored at room temperature. They reported that moisture in the heated 18 atmosphere did not reduce the incidence of C. botulinum type E spore survival. Pace et a1. (1972) reported that the destruction of C. botulimum spores on whitefish chubs depended on the relative humidity (RH) in the smoke chamber during the cook. 1 Using spore types 8 and E and conditions designed to simulate those used in commercial fish smoking plants, low numbers of the type E spores were destroyed within 30 minutes in fish held at an internal temperature of 77°C (170.60F) in an atmosphere of at least 70% RH. However, when several hundred thousand type E spores were present, an internal temperature of 82°C (179.6°F) and a minimum RH of 70% were required for spore destruction. Pace et a1. (1972) made quantitative estimates of spore destruction as a function of cooking conditions. Type E spore populations heated in an atmosphere of 70% RH were reduced by 2 to 4 logarithms when heated at 77°C (170.6°F), by 5 to 6 logarithms when ‘ heated at 82°C (179.6°F) and by more than 6 logarithms when . heated at 880C (190.4°F). Kosak and Toledo (1981) develOped a heating/smoking process for fish that was equivalent to a 12 logarithmic reduction of C. botulinum type E spores. The process, although severe, resulted in a product that did not appear excessively dehydrated. However, no attempt was made to evaluate the effect of their process on the overall organo- leptic acceptability of the fish, initially or over time. 19 Eklund et a1. (1982) studied the ability of liquid smoke, in combination with salt, to inhibit the outgrowth of C. botulinum type E spores in hot-processed whitefish. A mixture of C. botulinum type E spores was added to white- fish at a level of 103 spores/100 9 just before processing. They reported that the combination of liquid smoke (2.0%) 3 and a 3.7% water-phase salt concentration effectively inhibited spore outgrowth in hot-processed whitefish during 14 days of temperature abuse (25°C). However, liquid smoke (2.0%) in combination with 2.0% water-phase salt concentra- tions did not inhibit the outgrowth of the C. botulinum type E spores in temperature abused (25°C) hot-processed smoked whitefish. Temperature Schmidt et a1. (1961) evaluated growth and toxin pro- duction of four strains of C. botulinum type E (VH, Beluga, Iwanai and BE) at temperatures less than 40°F (i.e., 34, 36 and 38°F). They inoculated a sterilized beef stew with mildly heat shocked spores at a level of 4-12 million spores/30 9. They reported that the strains of C. botulinum differed in their rate of toxin production. However, the results of this study clearly showed that at 38°F (3.30C), toxin production was detected by day 31-45 in all strains tested, while no toxin was detected at day 104 at either 36°F (1.1°C) or 34°F (2.2°C). These results indicate that type E C. botulinum is a risk in products stored at normal 1 20 refrigerator temperatures (4°C). Abrahamsson et al. (1966) studied the effect of tempera- ture on the formation of toxin by C. botulinum type E spores 5 (lO spores/g) in both Robertson's chopped meat medium and a fish dialysate medium. They found that the C. botulinum type E spores/vegetative cells were able to grow and produce toxin at temperatures between 3° and 30°C. No toxin was found in the samples incubated at 1°C after 1 year. Nitrite Considerable research has been conducted to elucidate the role of nitrite in controlling the growth of C. botulinum in cured meat systems (Christiansen, 1980). Most cured meat 3 products do not have the levels of acid and/or salt required to completely inhibit the growth of C. botulinum spores (Riemann et al., 1972). The use of nitrite in smoked fish products has been limited. This is unfortunate because the control of Q. botulinum type E in this product requires a salt level which could be considered excessive. The addition of nitrite to - smoked fish could allow a reduction of the salt level in the product, while still maintaining safety against botulinal outgrowth. Pelroy et a1. (1982) evaluated sodium nitrite and sodium chloride as inhibitors of C. botulinum types A and E spore outgrowth and toxin production in temperature-abused (25°C), hot-processed salmon steaks. The fish samples were 21 inoculated intramuscularly before processing. After processing they were vacuum packaged in oxygen-impermeable plastic packages and incubated at 25°C. Pelroy et al. (1982) reported that the addition of nitrite permitted a decrease in the salt concentration required to maintain safety against botulinum toxin production. Salmon steaks processed without nitrite required a 3.8% water-phase salt concentration in order to inhibit C. botulinum type E spore outgrowth. A 6.1% water-phase salt concentration was required to inhibit C. botulinum type A toxin production. 9 If nitrite was added at a level of 100 mg/kg or greater, a 2.5% water-phase salt concentration was required to inhibit the type E spore outgrowth. In order to inhibit the out- growth of the type A spores, a minimum of 150 mg/kg nitrite and a 3.5% water-phase salt concentration was required. These values were valid for samples inoculated with 102 spores/gram of tissue and held for 7 days at 25°C. As the level of inoculation and/or the storage time increased, more nitrite and salt were required to inhibit toxin pro- duction. Holley (1981) stated that initial concentrations of 100-150 mg/kg of sodium nitrite are inhibitory to C. CQCC; ljflgm_in processed meats and that the maximum inhibition of C. botulinum occurred under anaerobic conditions at pH values of 4.5-5.5. L i 22 Nitrite inhibition of C. botulinum occurs by inhibiting or delaying the emergence of the vegetative cell from the spore and during cellular division (Sofos et al., l979c). Pivnick et a1. (1970) showed that nitrite acts to inhibit C. botulinum between the stages of germination and out- growth. Research by Duncan and Foster (1968) revealed that nitrite allows germination and swelling but prevents emergence from the spore coat or elongation. The inhibitory effect of nitrite can be increased by the addition of reducing agents such as ascorbic acid, cysteine and thioglycolate (Johnston and Loynes, 1971). The exact mechanism(s) of nitrite inhibition of C. botulinum has not been fully elucidated, since the nitrite ion is capable of a variety of reactions in a given system and because meat products are also complex and lack uni- formity (National Academy of Sciences, 1981). Some of the mechanisms proposed for the antibotulinal effects of nitrite are that nitrite (1) could react with other components during heating forming a substance capable of inhibiting spore outgrowth; (2) acts as either an oxidant or reductant on cellular components such as enzymes, enzyme cofactors, nucleic acids and cellular membranes; (3) reacts with thiols, which can react with components of the spore membrane and interfere with spore metabolism or (4) reacts with cellular iron, interfering with energy metabolism and repair mechanisms (Benedict, 1980). 23 In meat systems, the inhibitory effect of nitrite on the outgrowth of C. botulinum has been related to the ability of nitrite to bind iron. Tompkin et al. (1978b) investigated the role of iron in controlling the inhibition of C. botulinum in perishable canned meats cured with 156 mg/kg of sodium nitrite. They inoculated cured pork ham, beef round, and beef and pork organ meats (liver and heart) with a mixture of C. botulinum spores (types A and B) at a level of 102 spores/g sample. The inoculated samples were sealed in retort cans and stored at 27°C for up to 110 days. When supplemental iron was added to these systems, there was an increase in the rate of toxin production by the C. botulinum spores, indicating that nitrite acts to inhibit C. botulinum by binding the iron necessary for sporulation and/or cell outgrowth. Freeze et a1. (1973) studied the role of nitrite in C. botulinum inhibition and reported that nitrite, as undissociated nitrous acid, appeared to inhibit the energy dependent transport systems within the cell. Nitric oxide is theorized to react with essential iron containing compound(s) within the botulinal cell and therefore prevents cell outgrowth (Tompkin et al., 1978b). Tompkin et al. (1978b) reported that there appeared to be an inverse relationship between the amount of muscle pigment and the degree of inhibition by nitrite on the occurrence of C. botulinum outgrowth. 24 Added iron has been found to reduce the action of the nitrite by replacing the cation that had been bound (scavenged) by the nitrous oxide. The effect of added iron can be countered by the addition of metal chelators, such as polyphOSphate or ethylenediaminetetraacetic acid (Tompkin et al., 1978, 1979). Reddy et al. (1984) showed that nitric oxide forms an iron-nitrite complex which results in the destruction of iron—sulfur enzymes. They asserted that nitrite, therefore, inhibits the outgrowth of C. botulinum cells by interferring with their iron-sulfur enzymes. The duration of nitrite inhibition is temperature and/or spore load dependent (Geingeorgis and Riemann, 1979). The level of residual nitrite has been considered to be important in the ability of sodium nitrite to maintain its inhibitory effect, and the level of residual nitrite is time and temperature dependent (Christiansen et al., 1982). The inhibition of C. botulinum by nitrite essentially is a balance between nitrite depletion and the death of the germinated botulinal spores present in the system. This means that the safety of a meat product depends on having a high enough level of ingoing nitrite so that there is sufficient residual nitrite present until the number of viable botulinal cells has decreased to a point at which growth can no longer occur (Christiansen et al., 1978; Christiansen, 1980; Cook and Pierson, 1983; Reddy et al., 25 1984). Sorbate During the last two decades, the role of nitrite as a precursor in the formation of N-nitrosamines in foods has received much attention, and research has been directed towards finding a substitute for nitrite in meat systems. Sorbic acid and its potassium salt have been investigated as possible replacements for nitrite. The use of sorbates in conjunction with small amounts of nitrite (40 mg/kg) for color and flavor purposes has also been suggested (National Academy of Sciences, 1981). Research has shown that sorbate when used at a level of a 200 mg/kg is as effective as 100-120 mg/kg nitrite in delaying toxin production by C. botulinum in a variety of temperature-abused cured meats (Sofos and Busta, 1980). Tompkin et a1. (1974) added 1000 mg/kg of potassium sorbate to skinless, precooked, uncured sausage links inoculated with C. botulinum. They reported that potassium sorbate delayed the growth of the normal spoilage organisms. Sorbate also reduced the growth of C. botulinum, delaying toxin production for 6 days. The results of this study were in contrast to the findings of previous investigators and acted to stimulate further studies (Sofos and Busta, 1980). The significant (P<0.01) effect of potassium sorbate on the occurrence of gas production, package swelling (bloating) and toxin production has been shown in bacon 26 (Ivey et al., 1978; Sofos et al., l979a), in chicken frank- furters (Robach et al., 1978a; Sofos et al., 1979c; Huhtanen and Feinberg, 1980), and in turkey frankfurters (Huhtanen and Feinberg, 1980). Tanaka et a1. (1977) demonstrated that when potassium sorbate (2600 mg/kg) was added to frankfurters, it had an antibotulinal effect as effective as 1000 mg/kg of nitrite. Robach (1980) reported that the addition of l to 5% salt increased the ability of potassium sorbate to inhibit clostridia. It has been shown by many workers that the combination of nitrite (40 mg/kg) and sorbate (2600 mg/kg) is more effective in retarding C. botulinum than either nitrite or sorbate used individually (Sofos and Busta, 1980; Widdus and Busta, 1982). The inhibition of C. botulinum by sorbate is pH dependent, with an upper limit of pH 6.5 necessary for inhibitory action (Smoot and Pierson, 1981; Cook and Pierson, 1983). Smoot and Pierson (1981) reported that potassium sorbate is a strong inhibitor of C. botulinum at a pH of 5.7. Because the undissociated acid form is responsible for the inhibition of C. botulinum, the hydrogen ion concentration (pH) is a major factor in the antibotulinal efficacy of sorbate thus the inhibitory effect of sorbate increases as the pH decreases (Samson et al., 1955). This change in inhibitory effect occurs because the cell is only permeable to the acid in the undissociated 27 form. Raevori (1976) stated that sorbic acid as the free acid enters the bacterial cell and is then capable of inhibiting several of the cells enzyme systems. Researchers have linked sorbic acid to inhibiting fumarase activity (York and Vaughn, 1955b), sulfhydryl-containing enzymes, ficin and alcohol dehydrogenases (Whitaker, 1959), aspartase and succinic dehydrogenase (York and Vaughn, 1964), and malate and a-ketoglutarate dehydrogenases (Rhem, 1967). Harada et a1. (1968) suggested that sorbate competitively combines with Coenzyme A and acetate, inhibiting all reactions involving these compounds. Water Activity and pH Two main factors controlling the growth of spore forming organisms are water activity (Aw) and pH. The Aw of foods influences the growth of microorganisms by affecting their metabolic activity, reproduction, as well as their resistance to environmental conditions (Leistner et al., 1981). The Aw affects the lag and stationary growth phases as well as the death rate of the organism (Troller and Christian, 1978). Since most of the organisms associated with foods require a relatively high Aw level to grow, lowering the Aw in foods would reduce the organisms ability to grow or multiply (Leistner et al., 1981). Pace et al. (1972) noted fish processed in low moisture environments would undergo desiccation causing a gradient in water activity to occur at the surface of the 28 fish. This gradient would move into the fish until an equilibrium is established with the relative humidity of the environment. The loss of water from the fish could result in a reduction of the Aw to levels as low as 0.2 to 0.4. Murrell and Scott (1966) have shown that C. botulinum type E spores, while relatively intolerant to heat, are the most resistant to heat at Aw of 0.2 to 0.4. Baird-Parker and Freame (1967) investigated the inter- relationships of Aw, pH and temperature in controlling the outgrowth of C. botulinum type E spores. Type E spores were incubated in Reinforced Clostridial Medium (RCM) that had pH values between 5 and 7, and Aw's between 0.997 and 0.890. The spores were incubated at either 20° or 30°C. They reported that C. botulinum type E spores grew at Aw's between 0.99 and 0.997 at pH 5.3, 0.98 and 0.997 at pH 6.0 and 5.5 and at 0.97 and 0.997 at pH 7. These data indicate that as the pH increased, a lower water activity was required to inhibit the outgrowth of the C. botulinum type E spores. Ohye and Christian (1967) reported that the minimal Aw levels for C. botulinum types A, B and E were 0.95, 0.94 and 0.97, respectively. Riemann (1967) reported that toxin production by C. botulinum type E in brain heart infusion, ceased at a higher Aw than did growth. Emodi and Lechowich (1969) studied the interaction of water activity and temperature on the outgrowth of C. botu- linum type E spores in a TPSY (trypticase-peptone-sucrose-yeast 9 29 extract) medium. They reported that at 30°C, C. botulinum could grow at a water activity of 0.97, but growth only occurred at 3.3°C if all other factors were optimal. They also reported that at the same Aw, the time for spore outgrowth increased as the temperature decreased. The use of salt to lower the Aw of a system is an effective means of inhibiting Clostridia. It has been reported that salt concentrations which have little inhibi- tory effect at Optimal temperatures (30°C) have a strong inhibitory effect at lower temperatures (Roberts and Smart, 1976). This is important in the control of C. botulinum type E since type E is more sensitive to salt than types A and B. This means that at lower temperatures, the limiting Aw for C. botulinum type E would increase (Roberts and Smart, 1976). Mechanism of Lipid Oxidation Lipid oxidation is one of the most deteriorative processes occurring in food systems. It can, if allowed, result in reduction of quality, nutritional value and safety of foods (Dugan, 1968). Lipid oxidation is respon- sible for the formation of short chain aldehydes, ketones and fatty acids in fat-containing foods. These products of lipid oxidation are believed to be responsible for the development of the oxidized flavors in red meats, poultry and fish (Watts, 1962; Pearson et al., 1983). 30 The classical mechanism of lipid oxidation involves three steps: 1) initiation, 2) propagation and 3) termina- tion (Farmer and Sutton, 1943; Uri, 1961; Lundberg, 1962; Labuza, 1971). This mechanism is a free radical process that utilizes unsaturated fatty acids as the initial substrate. It has been accepted as the primary event in autoxidation. The three steps of the lipid oxidation mechanism can be illustrated as follows: Initiation: RH ................ R. + H. RH + 02 -------- R00 + H Pr0pagation: R- + 02 ------ R00- R00 + RH --------- ROOH + R- Termination: ROO- + R- -------- ROOR R +R- ---------- RR ROO- + ROO- ------ ROOR + 02 where R00- is a lipid peroxy radical, R- is an alkyl radical and RH is an unsaturated fatty acid. Once initiated, the reaction is propagated by the level of hydroperoxides produced due to their ability to decompose to free radicals. Termination begins at the point where the level of free radicals is such that they begin interacting and forming non-reactive species. The rate of lipid oxidation is affected by such factors as lipid composition, temperature, presence or absence of light, metal catalysts, inhibitory compounds and oxygen (Lea, 1962; Labuza, 1971). Under 31 extreme conditions, the hydrOperoxides formed during the autoxidation of unsaturated fatty acids can undergo further free radical chain reactions that can result in the forma- tion of polymers, other oxygenated compounds, cleavage products and/or reaction products with proteins (Gardner, 1979; Pearson et al., 1983). Measurement of Lipid Oxidation The major effect of lipid oxidation on food accepta- bility is the development of off-odors and off-flavors. These time-related organoleptic changes occur in food products as a result of oxidative rancidity. They have been associated with the accumulation of malonaldehyde and other oxidative reaction products some of which may be potentially harmful to human health (Mukai and Goldstein, 1976; Sham- berger et al., 1974, 1977; Caldironi and Bazan, 1982; Pearson et al., 1983). Malonaldehyde has been shown to be a secondary oxidation product of polyunsaturated fatty acids containing two or more double bonds (Dahle et al., 1962). To date the most widely used test for measuring the malonaldehyde content of muscle foods has been the 2-thiobarbituric acid (TBA) test (Gray, 1978; Rhee, 1978a; Melton, 1983). This test expresses the malonaldehyde content of foods in mg/kg values and this value is commonly referred to as the TBA number (Melton, 1983). The TBA test has been used by many researchers to follow lipid oxidation in cooked beef, pork and poultry 32 (Huang and Greene, 1978; Igene et al., 1979), in refrigerated and frozen stored beef, pork and poultry (Igene et al., 1979; Judge and Aberle, 1980; Younathan et al., 1980; Drerup et al., 1981) in freeze-dried beef and pork (Chipault and Hawkins, 1971) and in fish (Lee and Toledo, 1977). Despite abundant literature on its use as a measure of rancidity, the validity of the test has been questioned, particularly in food systems containing nitrite. Nitrite has been reported to decrease the TBA values in oxidized meat systems. Nitrite decreases the TBA value by nitrosating the malonaldehyde during the distillation step in the analysis (Hougham and Watts, 1958; Swain, 1972). The reduction of TBA . values by nitrite can be overcome by the addition of sulfanilamide before the distillation step (Zipser and Watts, 1962). It has been shown that the TBA test is subject to interference by the reaction of the TBA reagent with other substances. Aldehydes have been reported to react with the TBA reagent to form red complexes that absorb at the same wavelength (532) as the malonaldehyde-TBA complex (Melton, 1983). It has also been reported that the TBA reagent can react with alkanals, 2-alkenes and 2,4-alkadienals to form a yellow complex that absorbs at 452 nm (Marcuse and Johansson, 1973; Patton, 1974). The significance of these other TBA reactive substances (TBARS) to the quantitation of malonaldehyde has been 33 studied. Igene et al. (1985a) reported that in cooked chicken meat (white and dark), the TBA values had a correla- tion of -O.87 with the sensory evaluation of rancidity in the cooked product (warmed-over flavor). They noted that the malonaldehyde-TBA complex accounted for 93.3% and 83.0% of the total TBARS in the distillates from the cooked white meat after 0 and 3 days, respectively. Corresponding values for the dark meat were 95.5 and 94%, respectively, indicating that in cooked chicken, malonaldehyde is the primary TBARS. Yamauchi (1972a) reported that malonaldehyde is responsible for 99.2% of the TBARS in cooked rancid pork. Lipid Oxidation in Muscle Foods Table 2 lists the unsaturated fatty acid composition of the major muscle foods. From the data shown, it is understood why fish and poultry lipids are so vulnerable to lipid oxidation. The fish lipid system contains significant quantities of the long chain polyunsaturated fatty acids (Khayat and Schwall, 1983). Lipid Oxidation in Red Meat and Poultry Igene et a1. (1980) using a beef and a poultry model system, reported that during frozen storage both the tri- glycerides and the phospholipids contribute to the develOp- ment of rancidity in meat products. The influence of the triglycerides on the development of rancidity during frozen storage depended on the degree of unsaturation of the 34 Table 2. Unsaturated fatty acid content of lipids in some muscle foods (Melton, 1983). Content (%) Fatty AC‘d Lamb Beef Pork Chicken Fish C18:1 9.51 33.44 12.78 20.25 19.59 C18:2 18.49 10.52 35.08 14.20 5.88 C18:3 0.43 1.66 0.33 0.90 8.07 020:2 0.34 0.69 -- -- 0.20 020:3 0.62 2.77 1.31 1.30 0.36 020:4 13 20 8.51 9.51 11.60 3.75 020:5 -- 0.76 1.31 1.55 7.16 022:4 -- 0.88 0.98 2.10 0.65 022:5 -- 0.92 2.30 5.75 2.39 022:6 -- -- 2.30 5.75 2.39 35 component fatty acids and the length of storage of the meat and poultry products. They also found that during frozen storage of intact raw meat, the major changes in total lipids occurred in the triglyceride content, while the phospholipid content remained relatively unchanged over time. Lipid Oxidation in Fish The development of rancidity in fish muscle from fatty species such as salmon and whitefish can be rapid. It is governed by complex factors common to many biological systems such as inherent metal ions, presence of natural antioxidants, kind and amount of fatty acids, age, season of harvest and storage conditions before and after processing (Braddock and Dugan, 1972). The effects of frozen storage on the overall quality of fish have been investigated and several reviews are available (Mills, 1975; Kelly, 1969; Olley et al., 1969; Shewfelt, 1981; Khayat and Schwall, 1983). Most of the studies concerning the oxidation of fish lipids have been conducted with lean fish whose primary lipid constituents are phos- pholipids. Therefore little research has been concerned with the triglyceride fraction of fish lipids. Fish phos- pholipids can oxidize during frozen storage. This has been noted by many researchers and in a variety of species including cod (Lovern et al., 1969), lemon sole, halibut (Olley et al., 1962), trout (Jonas and Bilinski, 1976b), \ 36 herring (Bosund and Ganrot, 1969b), freshwater whitefish (Awad et al., 1969), salmon (Botta et al., 1973), silver hake (Hiltz et al., 1976), carp, red sea beam (Toyomizu et al., 1977) and capelin (Botta and Shaw, 1978). Mechanism of Warmed-over Flavor Development in Muscle Foods A primary catalyst to warmed-over flavor in meat systems has been shown to be the heme compounds of the muscle tissue. This is the topic of an in-depth review by Love (1983). The ability of heme to accelerate the oxidation of lipids was shown by Tappel (1962). Sato and Hegarty (1971) and Igene et a1. (1979) have shown that beef muscle thoroughly extracted with water does not develop warmed-over flavor. This indicates that the fact0r(s) responsible for initiating warmed-over flavor are water soluble, as would be the heme and nonheme irons. Younathan and Watts (1959) reported that cooked meats develop rancid off-flavors and TBARS more rapidly than raw meat. They proposed that the Fe3+ hemes were active catalysts of lipid oxidation in cooked meat systems. Brown et al. (1963) reported that ferric and ferrous hemes were both catalysts, but that the ferric heme was the most active form of heme in lipid oxidation. Fox (1966) stated that during cooking, heme pigments are converted to the ferric hemochromogen. 37 It has been speculated that during cooking, the protein portion of the hemoprotein is denatured and unfolds. This unfolding of the protein would expose/release the iron mole- cule and allow contact with the lipids of the system, increasing the rate of lipid oxidation (Labuza, 1971; Pearson et al., 1977). Igene et a1. (1979) demonstrated that the myoglobin is not responsible for the increased rate of lipid oxidation in cooked meats. Rather, it is the Fe2+ released from the pigment that causes the accelerated rate of lipid oxidation. The role of nonheme iron as a prooxidant in meats has been investigated (Sato and Hegarty, 1971; Love and Pearson, 1974). Torrance and Bothwell (1968) and Schricker (1982) have reported that the heme iron represents 62% of the total iron in beef muscle, while nonheme iron accounts for only 5.4- 5.5%. Igene et a1. (1979) reported that heating of meat results in the release of iron from the heme molecules. They noted that when extracts of beef pigments were heated to 70°C, the amount of nonheme iron increased from 8.7 to 27% of the total iron in the system. Schricker et a1. (1982) reported that the nonheme iron content in beef heated to 100°C increased from 9.9 to 20.9 mg/kg. Chen et a1. (1984) showed that slow cooking of meat pigment extracts increased the amount of nonheme iron released more rapidly than did fast cooking. They theorized 38 that since meats generally are cooked slowly, the increased rates of lipid oxidation could be the result of the increased nonheme levels that would be produced. Warmed-over Flavor in Red Meats and Poultry Warmed-over flavor in cooked meats has been well documented (Timms and Watts, 1958; Younathan and Watts, 1960; Ruenger et al., 1978) and has been attributed to lipid oxidation. Oxidation of the tissue lipids appears to occur in two stages, i.e., the phospholipids are oxidized first followed by the triglycerides (Igene, l979). Younathan and Watts (1960) demonstrated that flavor deterioration in cooked meats involved the unsaturated fatty acids of the lean tissue or cellular lipids, which exist primarily in the form of phospholipids. Wilson et a1. (1976), however, reported that while the phospholipids are major contributors to rancidity in beef and lamb, the total lipids are more important in pork. Igene and Pearson (1979) have also shown that the total phospholipids are primarily responsible for the develop- ment of warmed-over flavor in cooked beef and poultry. Being less susceptible to oxidation, the triglycerides appear to exert only a minor influence on the development of warmed-over flavor. The level of phospholipids in muscle foods remains relatively constant when they are expressed as a function of the total lipid (Dugan, 1971). However, between species 39 there is considerable variation in the actual phospholipid content (Pearson et al., 1977). Pearson et al. (1977) found that poultry meat and fish are higher in phospholipids than are red meats. Within the same species, it has been found that phospholipids generally have a higher level of unsatu- rated fatty acids than the triglycerides. Also within the same species, the red muscles are more prone to the develop- ment of WOF than are the white muscles. Between species there is a great difference in the susceptibility to the deve10pment of WOF, with turkey being the most susceptible followed by chicken, pork, beef and mutton (Pearson et al., 1977). Warmed-over Flavor in Fish Based on the high correlation between the TBA number and the oxidation of polyunsaturated fatty acids, fish would suffer extensively from WOF (Pearson et al., 1977). Theo- retically, fish would behave more like turkey than red meats in this aspect. Wide differences would be expected among the fish species in their susceptibility to the development of WOF. These differences would be due to the variability in the lipid content and composition among the species of fish (Pearson et al., 1977). The role of the phospholipids and the triglycerides in the deve10pment of WOF in fish is unclear. Since there is little difference in the fatty acid composition between the 40 phOSpholipid and the triglyceride fractions in fish (Braddock and Dugan, 1972), both lipid fractions contribute to the rapid deve10pment of oxidative rancidity in fish (Zipser and Watts, 1951; Lovern, 1959; Braddock and Dugan, 1972; Pearson et al., 1977). In addition, the occurrence of warmed-over flavor in fish has not received much attention since few fish products are cooked and then rewarmed prior to serving. However, a study by Sen and Bhandary (1978) showed that when sardine fish were cooked and then stored at refrigerated temperatures there was a significant decrease in the rate of lipid oxida- tion when compared to raw fish stored under the same conditions. The raw sardines became rancid in 2-3 days while the cooked product became rancid in 6 days as evaluated by TBA numbers and peroxide values. They concluded that this decrease in lipid oxidation was due to: l) the destruction of a "lipoxygenase" inherent to the fish, 2) the formation of water-soluble antioxidants as a result of the cooking process and/or 3) the destruction of heme compounds. The role of metal ions, Fe2+, Fe°+, Cu°+, and hemin in the oxidation of lipids in frozen fish and fish oils have been investigated. Ke and Ackman (1976) found that when copper, iron and zinc were added to mackerel skin and meat there was acceleration of the lipid oxidation process in this nonaqueous system. Mizushima et al. (1977) added Fe°+, Fe3+ 2+ and Cu to fish homogenates and found that the relative effectiveness of these metals to oxygen uptake, i.e., 41 oxidation, was in decreasing order: Fe3+ > Fe2+ > Cu°+. They went on to investigate the effect of iron, copper and hemin on lipid oxidation in fish homogenates and found all three were capable of accelerating the process. The relative 3+. The activity of these ions were: Fe2+ > hemin > Cuz+ > Fe addition of up to 50 mg/kg copper to fish muscle prior to freezing acts to reduce the induction period for lipid oxidation to a few days (MacLean and Castell, 1964). Cations have also been shown to be capable of accelerating the formation of secondary reaction products (Seblacek, 1974). Various biochemical compounds such as heme proteins, organic acids, amino acids, pigments and various fish tissues have been shown to catalyze lipid oxidation reactions, alone or in combination with trace metals (Castell and Bish0p, 1969; Jurewicz and Salmonowicz, 1973; El-Zeany et al., 1974; Yu et al., 1974). Antioxidant Role of Nitrite in Meat Systems Warmed-over flavor can be inhibited in meat systems by the addition of 50 mg/kg nitrite and eliminated by the addition of 220 mg/kg of nitrite (Sato and Hegarty, 1971). Fooladi et a1. (1979) reported that nitrite-free cooked pork had TBA values 5 times higher than cooked pork containing nitrite. The addition of nitrite to cooked beef and chicken reduced their TBA values by half. 42 The mechanism by which nitrite inhibits the development of WOF has never been fully explained. Igene et al. (19850) suggested that nitrite acts as an antioxidant by (1) strongly binding the heme pigments, thus preventing the release of the nonheme iron during the cooking process; (2) serving as a chelator or metal sequestor and/or (3) to a lesser degree, nitrite could act to stabilize the lipids in muscle membranes. The primary process occurring in meat systems appears to be the stabilization of the heme porphyrin ring. Chen et a1. (1984) heated meat pigments extracts, with or without added nitrite. They reported that the nonheme content of the extract heated to 62°C without added nitrite, increased from 1.31 to 1.78 mg/kg. When the same system was heated to 88°C, the nonheme content increased to 2.34 mg/kg, an overall increase of 78.6%. However, when a similar extract of pigments were heated to 88°C in the presence of nitrite, the nonheme content went from 1.29 to 1.14 mg/kg, an 11.6% decrease. They attributed this finding to the ability of nitrite to stabilize the heme proteins, thus preventing the release of nonheme iron. Chen et al. (1984) asserted that the increase in nonheme iron in the heated pigment extract, without added nitrite, indicated that the iron molecule was being released from the porphyrin ring. Schricker and Miller (1983) have also suggested that the nonheme iron originates from the porphyrin ring of the heme molecule. 43 The role of nitrite as a membrane stabilizer has been discussed in the literature. Liu and Watts (1970) noted that in meat muscle, myoglobin is in solution in the cyt0plasm and is separated from the phospholipids in the membranes. During cooking the membranes are destroyed and contact between the heme components and the phospholipids is established. Liu and Watts (1970) asserted that the addition of nitrite to this system would have the same effects as stabilization of the membrane. Love and Pearson (1976) suggested that it seemed more probable that nitrite complexes and stabilizes the membranal lipids, thus preventing the development of WOF in cooked meats. Evidence indicates that nitrite can react with heme proteins, nonheme proteins, low molecular weight peptides, amino acids and trace metals in meats (MacDonald et al., 1980). Nitrite has also been found in adipose tissue and may react with unsaturated fatty acids (Frouin et al., 1975; Walters et al., 1979). Woolford and Cassens (1977) reported finding levels as high as 20-25% of the nitrite added to bacon in the adipose tissue. Goutefongea et al. (1977) recovered 35% of the added nitrite in whole adipose tissue. They also reported that 80-90% of the added nitrite found was in the free state. 44 Mechanism of N-nitrosamine Formation Nitrite contributes to the color, flavor, lipid stabil- ity and botulinal safety of cured meats (Gray, 1981). However, its continued use has been debated as it has been shown that nitrite reacts with secondary amines found in foods to form N-nitrosamines (Gray, 1976). It has been shown that the majority of the N-nitrosamines tested in animal experiments are carcinogenic (Crosby and Sawyer, 1976; Preussman et al., 1976; Gray and Randall, 1979; Sen, 1980). N-Nitrosamines are formed primarily from the reaction between secondary amines and nitrous acid. In this reaction, R1 is an alkyl group, while R may be an alkyl, aryl or a wide variety of functional groups. R R\ ;::::::>NH + NOE l///////NN=O + H20 R R The extent of N-nitrosamine formation is governed by a variety of factors such as the basicity of the amine, concentration of the reactants, pH, temperature and the presence or absence of catalysts and inhibitors. N-Nitroso compounds have been identified in a variety of food systems, including cured meat products, non-fat dried milk, dried malt and beer (Gray and Randall, l979). 45 N-Nitrosamines in Fish Products In the state of Michigan, nitrite has clearance for use in smoked chub, and it may be added to levels up to 200 mg/kg (Michigan Department of Agriculture, 1965). This often results in initial nitrite levels of 75 to 150 mg/kg and a residual nitrite level of 10-20 mg/kg (Holley, 1981; Cook and Pierson, 1983). Because of these levels of nitrite it is possible that the formation of carcinogenic N-nitrosamines could occur from the reaction of the added nitrite, and the free amines of the fish. This would be especially true in some marine species of fish which have been shown to have high levels of trimethylamine which readily breaks down to dimethylamine and formaldehyde (Dyer and Mounsey, 1949; Shewan, 1951; Castell et al., 1971; Spinelli and Koury, 1979; Sikorski and Kostuch, 1982). Malins et al. (1970) investigated the possibility of nitrosamine formation in smoked chub and found that in all probability N-nitrosodimethylamine (NDMA) does not form in concentrations greater than 10 ug/kg during the processing of the fish. Malins et a1. (1970) were unable to find any evidence of amine nitrosation in smoked chub. Gadbois et a1. (1975) treated sablefish with 0-1300 mg/kg - nitrite before cold-smoking. The product was analyzed for N-nitrosamine levels at day 0 and after 2 weeks storage at 40°F. They reported that in fish containing 0-500 mg/kg nitrite only trace amounts (<10 pg/kg) of NDMA were present in the samples. Increasing levels of nitrite did not increase 46 the levels of NDMA detected, and there was a slight decrease in the amount of NDMA during storage at 40°C. Fazio et a1. (1971) investigated the effect of processing, with or without nitrite and/or nitrate, on N-nitrosamine levels in commercially processed shad, sable and salmon. Raw and processed samples of these fish were collected from 2 different commercial processing plants. Analysis of these species of fish showed that the raw sable contained 4 ug/g NDMA while no NDMA was found in the other species of fish in the raw state. The processing of these three species of fish with nitrite 0r nitrate, increased the NDMA level in sable to 9-26 ug/g. Salmon and shad, processed with nitrite, were found to contain O-l7 ug/kg and 0-12 ug/g NDMA, respectively. Sen et al. (1970) analyzed 23 samples of fish products, i.e., 18 smoked and 5 canned. All of the 23 samples had been cooked with or without added nitrite (200 mg/kg). The 18 smoked fish samples included cod, haddock, hake, mackerel and salmon, while the canned samples included salmon and mackerel. They reported that when the fish were cooked without added nitrite no N-nitrosamines were detected. When the fish were cooked in the presence of the added nitrite it was found that the cod contained 0.5 ug/kg N-nitrosodipropyla- mine (NDPA), the haddock contained 1 ug/kg N-diethylnitrosa- mine (NDEA), the mackerel contained 1 ug/kg NDMA, the hake contained 4 ug/kg NDMA and the salmon contained wumnvxo mo ucmanFm>mv mg» :0 munncoomm so“: muvcuwc ecu Fm>mp new max» mxosm Co uumwwm on» mcpzvscmpmn gee mcovuvmoasoo mcvcm .w mpnmp 75 prior to analyzing, samples were baked on days 3, 2, 1 and 0. On day 0, the samples were analyzed for their TBA value and their flavor volatiles. At the same time, taste panelists were asked to evaluate the overall desirability of samples from days 0, l, 2 and 3 using a ranking procedure (Appendix 3). Using the same panelists, triangle tests on samples baked on days 0, l and 2 were used to determine the threshold value for rancidity in this product based on the TBA value. In the triangle test, the panelists were asked to find the "odd" sample and then indicate if it was more or less desirable than the other two samples on the plate (Appendix 3). Methods of Analyses Assay for Clostridium botulinum Type E Toxin The inoculated, temperature abused (27°C) whitefish samples were assayed for the presence of C. botulinum type E toxin using the FDA standard procedure (FDA Bacteriological Analytical Manual, 1978). Proximate Analysis Moisture, fat, salt, pH and nitrite determinations were performed according to standard AOAC procedures (1975). In the nitrite analysis, N-l-naphthylethylene diamine was used to replace the carcinogenic a-naphthylamine (Usher and Telling, 1975). 76 Sorbate Analysis Residual potassium sorbate was determined using the method described by Robach (1980). Water Activity Water activity was determined using a thermocouple psychro- meter (Decagon Devices, Inc., Seattle, WA). A confirmation of the data was made using the freezing point depression method of Lerici et al. (1983). Thiobarbituric Acid Test (TBA) The TBA distillation method of Tarladgis et a1. (1960) was used to measure the deve10pment of rancidity in smoked fish. The modification of Zipser and Watts (1962) to prevent interference by nitrite was utilized for all samples contain- ing nitrite. TBA numbers were expressed as mg malonaldehyde/ kg fish. Lipid Extraction and Fractionation The total lipid was extracted from the whitefish using the procedure of Bligh and Dyer (1959). Separation of the total lipid into triglyceride and phospholipid fractions was accomplished using the method of Choudhury et al. (1960). Fatty Acid Analysis The methylation of the total, triglyceride and phospho- lipid fractions of the raw whitefish was done according to the boron-trifluoride-methanol procedure outlined by Morrison and 77 Smith (1964). The methylated samples were analyzed for their fatty acid composition using a Hewlett Packard gas chromato- graph (Model 5840A) equipped with a flame ionization detector (F10) and a Hewlett Packard 18850A GC integrator. The glass column (2m x 2 mm i.d.) was packed with 10% SP 2330 on Chromosorb W. (Supelco, Bellefonte, PA). Operating conditions were as follows: Nitrogen was used as the GC carrier gas and the flow rate was set as 30 ml/min; the injection port tempera- ture was 250°C; the detector temperature was set at 350°C; a temperature program that started at 140°C for one minute then went to 190°C at a rate of 10°/min was used. Identification of the fatty acid peaks was made using the retention times of standard fatty acids assayed under identical conditions. N-Nitrosamines Smoked whitefish were analyzed for N-nitrosamines using the extraction procedure of White et al. (1974). The only modification was the addition of 1 gram of ammonium sulfamate prior to the distillation step to minimize possible arti- factual formation of N-nitrosamines during sample work-up. The quantitative determination of volatile N-nitrosamines was carried out using a gas chromatograph-thermal energy analyzer (GC-TEA) system comprised of a Varian 3700 GC coupled with a TEA model 502 LC (Thermo Electron Corp., Waltham, MA) via a 1/8” glass-lined stainless steel transfer line. The GC column was glass (3m x 2mm i.d.) packed with 10% Carbowax 20 MM TPA on Chromosorb WHP. The operating conditions were 78 as follows: nitrogen flow was 30 ml/min; initial temperature was 140°C held for 1 minute; the rate of temperature change was 15°/min to a final temperature of 180°C for 7 minutes. The injector temperature was 150°C and the TEA pyrolyzer furnace was set at 425°C. The TEA reaction chamber pressure was 1.5 Torr and the GC-TEA transfer line was heated to 200°C. Identification and quantitation of the N-nitrosamines were made by injecting known amounts of standards and comparing retention times and peak areas. N-Nitrosothiazolidine and N-nitrosothiazolidine carboxylic acid were determined using the procedure of Sen et al. (1985). Phenol Analysis The total phenol content was determined spectrophotometri- cally using the method of Bratzler et a1. (1969). The analysis of the samples for the individual phenols present involved the extraction method outlined by Lustre and Issenberg (1970). In this procedure, 100 g of smoked whitefish were blended with 200 m1 of 5% sodium hydroxide until homogenous. Trichlora- cetic acid (40%, 200 ml) was slowly added during the blending to precipitate the proteins. The sample was transferred to 250 m1 glass centrifuge bottles and centrifuged at 1800 rpm for 15 minutes. After centrifugation, the top liquid layer was decanted off and then filtered under suction through Whatman #1 filter paper. The samples were placed into a l 1 beaker and placed overnight at -20°0 to solidify the lipid layer so that it could be physically removed from the sample. 79 The extract was taken to pH 12 with 40% sodium hydroxide (approximately 30 ml). The alkaline solution was then extracted with two 300 ml and one 150 m1 volumes of ethyl ether and the ether was discarded. The remaining solution was taken to pH 6.8 by saturation with carbon dioxide. This step regenerated the phenols from their sodium salts. After regeneration, the sample was extracted with two 300 ml and one 150 ml volumes of ether. The ether was dried over anhydrous sodium sulfate and then concentrated to a volume of 20 ml using a rotary evaporator at room temperature. The concentrate was transferred to a centrifuge tube and dried to 2 ml under a stream of nitrogen. The extract of the samples were analyzed on a Hewlett Packard gas chromatograph (Model 5840A) equipped with a flame ionization detector (F10) and a Hewlett Packard 18850A GC integrator. A 50 meter 10% Carbowax 20 M column (Alltech Assoc. Inc., Deerfield, IL) was used under the following operating conditions. Helium flow was set at 28 psi; the injector temperature was 250°C; detector temperature was 350°C; initial oven temperature was 90°C for 1 minute; rate was 5°/minute to a final temperature of 190°C for 60 minutes. Specific phenols present were tentatively identified by injecting standards and comparing retention times. Gas Chromatography of the Volatiles from Baked Whitefish Two hundred grams of the baked whitefish were placed in a 200000 boiling flask, 500 ml of water were added and the 80 mixture was placed in a Likens-Nickerson apparatus. Ethyl ether (25 ml) was used as the extracting solvent. The system was allowed to reflux for 6 hours followed by concentration of the ether to a final volume of 0.5 ml. Sodium sulfate was used to dry the ether extract of the fish volatiles. The extract was stored in a screw-top vial and held at -20°C until analysis. The extracts were analyzed using a 3m x 2mm (i.d.) glass column packed with a 10% Carbowax 20M TPA on Chromosorb WHP (80/100 mesh) (Supelco, Bellefonte, PA). A Hewlett Packard 5840A Gas chromatograph was used as follows: the initial temperature was set at 50°C and held for 2 minutes. The temperature changed at a rate of 5°/minute until a final temperature of 190°C was reached followed by a 20 minute hold at this temperature. Helium carrier gas was regulated at 30 ml/min, the injection port temperature was set at 275°C and the flame ionization temperature was at 300°C. Injection volume was 3 ul. GC-MS Analyses of the Baked Whitefish Volatile Extract GC-MS analyses of the extracts of the volatiles from baked whitefish cooked on day 0, l and 2 were performed. Three ul of the volatile extract were injected into a Hewlett Packard 5840A gas chromatograph equipped with a 3Mx 2mm (i.d.) glass column packed with 10% Carbowax 20 M TPA on Chromosorb WHP (80/100 mesh). The GC was operated using the conditions described previously. The injection effluent passed into a Hewlett Packard 5985A mass spectrophotometer having the 81 following parameters: electron impact voltage, 70 e V; electron multiplier voltage, 2410 e V; threshold 0.6; source temperature, 200°C; analog/digital measurements, 3/sec; and ion detection in the positive mode. Statistical Analysis Statistical analyses of the data in this study were made using the ranking procedure of Kramer (1963), ANOVA analysis (Gill, 1978) and Bonferonni's t-statistic comparison (Gill, 1978). RESULTS AND DISCUSSION The Effect of Time, Filleting and Brine Salt Concentration on Salt Uptake in Whitefish Whitefish fillets and whole, eviscerated whitefish were brined in 30° salometer brine at 4°C and sampled every two hours for 14 hours in order to establish the rate of salt uptake. The percent salt, moisture and salt (wp) concentra- tions of the fillets and whole, eviscerated whitefish over the brining period are listed in Table 9. The relationship between salt uptake and time in both the fillets and the whole eviscerated whitefish is illustrated in Figure 1. The fillets had a significantly (P<0.01) faster rate of salt uptake than the whole, eviscerated fish. This is expected, since filleting of the fish increases the surface area, thus allowing the salt solution greater access to the fish tissue. The maximum level of salt uptake appeared to occur after 10 hours, indicating that an equilibrium between the salt in the brine and that in the fish tissue might exist. A similar phenomenon was observed with the whole, eviscerated fish, but at a much lower level of salt. The salt concentra- tions in the whole, eviscerated fish appeared to reach equilibrium with the salt in the brine after 10-12 hours. 82 83 Table 9. Salt and moisture levels in whole, eviscerated and filleted whitefish brined in 30° salometer brine for different times. Fish Form Time of Salta Moistureb Salt brinin (%) (%) (water-phase) (hours? (%) Fillets 2 3.0 57.2 5.0 4 3.8 60.0 5.9 5 3.8 59.5 6.0 6 4.5 58.7 7.1 10 4.9 58.5 7.7 12 4.6 59.8 7.2 14 4.7 58.7 7.4 Whole, 2 0.4 69.6 0.6 eviscerated 4 0.8 70.6 1.1 5 0.9 68.9 1.3 6 1.1 70.3 1.5 10 1.1 67.9 1.6 12 1.2 70.0 1.6 14 1.0 69.3 1.4 aEach value represents the mean of six analyses. bEach value represents the mean of triplicate analyses. 84 IBJD r 7.0 '- ./.@\)/© Fillets ’8 g 6.0 '- O O .i: a. 4 5.0 - . o H to E 4.0 - = a 00 ._, 3.0 - c o 2 o 2-0 * Whole, eviscerated fish a. O o o O O 1.0 - O O o 2 4 e e 10 12 14 Hours of brining Figure l. The effect of brining time on salt uptake in whole, eviscerated and whitefish fillets brined in 30° salometer brine. 85 In both brinings, there appeared to be a slight loss of salt from the fish tissue once equilibrium was established. However, a longer period of brining would be necessary to establish this loss as being real. If the salting of fish follows the laws of simple dif- fusion as it has been suggested (Torry Research Station, 1960), then any loss of salt from the fish tissue could be the result of a higher salt concentration in the tissue relative to that in the brine. Crean (1961) reported that when muscle is immersed in a brine solution and allowed to equilibrate, two events can occur. At low or intermediate brine salt concentrations, water/salt is transferred from the brine to the muscle and the muscle swells. This swelling is the result of the adsorption of the chloride ions on the surface of the protein chains, which increases the net negative charge on the chains. The change in charge causes increased repulsion between and within the proteins so that protein structure expands. There is an increase in the water of hydration needed to hydrate the new negative charges and this results in muscle swelling. The second event that can occur during the brining of muscle is that at salt concen- trations beyond a certain point, water/salt leaves the muscle and goes into the brine. The percent salt, moisture and salt (wp) concentration of whitefish fillets brined for 14 hours at 4°C in 10, 20, 30, 40 and 50° salometer brines prepared with or without nitrite, 86 are summarized in Table 10. The relationship between the degree salometer of the brine and percent salt (wp) content of the finished product are shown in Figures 2 and 3. There was no significant difference (P<0.01) in the regression curve coefficients of the rate of salt uptake between the fillets brined with or without nitrite in the brine. There was a high correlation (r=0.981) between the salt uptake by the fish and the salt content of the immersion brine. When nitrite was added to the brines there was a slight decrease in the correlation (r=0.971) between the salt uptake by the fish and the salt content of the brine. This difference was not found to be significant (P>0.0l) as would be expected. These data agree with the findings of Del Valle and Nickerson (1967b) who showed that the salt content in swordfish muscle slices increased with increasing concentra- tions of salt in the brine. Their work showed a correlation of 0.99 between the rate of salt uptake in the fish tissue and the salt content of the brines. Statistical analysis of the taste panel evaluations of the smoked whitefish prepared in the 10, 20, 30, 40 and 500 salometer brines, with and without added nitrite, was carried out using an ANOVA analysis (Gill, 1978). There was no significant difference in the flavor, mouthfeel or overall acceptability of fish brined with and without nitrite. However, there was a significant (P<0.01) difference in the ability of the panelists to discern different salt levels in 87 Table 10. Salt, moisture and salt (water-phase) concentra- tions of whitefish fillets brined for 14 hours in different degree salometer brines. Brine Composition Salta Moistureb Salt (%) (%) (water-phase) (%) 10O Brine 1.3 62.1 2.0 200 Brine 2.2 62.6 3.4 30° Brine 3.5 61.3 5.4 40° Brine 4.0 61.9 6.1 50° Brine 5.9 59.0 9.1 100 Brine + nitrite 1.0 62.3 1.5 20° Brine + nitrite 2.6 61.6 4.0 30° Brine + nitrite 3.7 61.0 5.7 400 Brine + nitrite 3.8 62.4 5.8 50° Brine + nitrite 6.3 63.1 9.0 °Each value represents the mean of six analyses. bEach value represents triplicate analyses. 88 1(3‘3 r a e/ 8 4: 1343 - Cl L .2 6.0 - 9 g :2 = 4.0 _ Y=O.168X+ 0.14 a o 00 5 2.0 - © 2 13 //////l a, . . . . . <3 10 20 30 40 so Brine salometer Figure 2. The effect of brine salometer concentration on salt uptake of whitefish fillets brined for 14 hours at 4°C. 89 10.0 - 9' o 6.0 F P o ?=o.168x+o.18 2.0 Percent salt (water-phase) l l l I O ‘10 2O 30 4O 50 Brine salometer Figure 3. The effect of brine salometer and nitrite on salt uptake in whitefish fillets brined for 14 hours at 4°C. 90 the smoked whitefish. There was a correlation of 0.91 between the salt (wp) content of the sample and the panelists' response to the level of saltiness in the fish brined in the non-nitrited brines. When nitrite was added to the brines, the correlation between panelists response to the saltiness of the fish and the salt (wp) content of the fish was 0.95. These data reflected the ability of panelists to distinguish different salt levels, but it did not establish which level of salt was considered to be the most desirable. Nitrite appeared to heighten the perception of the panelists to salt in the smoked whitefish in this study. The Effect of Salt, Nitrite and Sorbate on Toxin Production by Clostridium botulinum Type E Spores in Smoked Whitefish The ability of salt, nitrite and sorbate to inhibit C. botulinum type E spore outgrowth and toxin production in smoked whitefish was evaluated. The presence of toxin in smoked whitefish samples inoculated with a 10° spores/m1 mixture of C. botulinum type E spores and incubated at 27°C is shown in Table 11. Due to a limited supply of fish, only 20 inoculated and 18 uninoculated samples were prepared for the treatments prepared in the 10° salometer brines. How- ever, for the treatments prepared in the 20 and 30° salometer brines, a total of 48 inoculated and 36 uninoculated samples were prepared for each treatment. The uninoculated samples 91 .uoceaoca no—asou on coats: puuoaxmu—aaom u.x0pu .u.xou on o» vo>osa was» mo—atou cause—o .moconm u max» s=:.—:uoa I:.v—cumopu so ecu-n puxnusoou no. a so .I m~.o no.) a noun—:uoc_o ov\o ov\o oe\o oe\o moxo ov\o ovxo ovxo ovxo coxo ovxo ovxo ouoacom + ou.su.: + 0:.Lo con oo\o o¢\o oc\o oexo oc\o oe\o ocxo mv\o ovxo ovxo ocxo ovxo uu—ga—c + 0:.59 com oe\o ocxo ooxo ov\o oc\o oc\o ovxo cvxo ocxo ovxo ovxo ocxo 0:.Lo con oexo ov\o ov\n oq\o ov\o ov\o mvxo ccxo ocxo cvxo ovxo . ov\o nuance» + no.5“.c + 0:.so cow oe\o ov\o QQ\n ov\o ov\o oc\o ovxo ov\o ov\o oc\c ovxo oc\o ou.Lu—: + ozone cow oc\e_ oc\_— ov\o oe\m ov\~ oo\~ ov\o ovxo ovxo ov\o ocxo ov\o u:.so cow -- -- it -- i- i- .. no~\c— ao~\o— ao~\o nouxn o~\o oueacom + 00.50.: + oc—Lo oo— -- s- i: i- i- i- i: .: ao~\o_ nouxo— ao~xo ouxo no.5».c + oc—Lo eon -- -- i- -- -- i- i- i- i- ao~\o— ao~\~— uo~xc oc—Lo cop no as no om o. N: on oN pm o- s o «cosy-och monocum uo «zoo i--Ii:lii I i--ll:.ii a.nocoam m was» E:c..=ooo E:_o_cum .u so.) coon—39oz. mo_nsum zw.oou.gs voxosm “oeswv canon. osaaoconsou e. :.xou we mucomosa .—p o—aoh 92 were used for chemical analysis. Twelve of the 20 inoculated samples prepared in the 10° salometer brine (without added nitrite or sorbate) were bloated and toxic by day 7. By day 14, the remaining samples (6) had bloated and were also found to be toxic. Only the 2 samples taken, just prior to placing the samples in the 27°C incubator, on day 0 were non-toxic. The addition of nitrite to the 10° salometer brine decreased the rate of toxin production, but by day 21, sixteen 0f the prepared 20 samples had bloated and were toxic. Only four samples, two taken on day 0 and two on day 7, had not bloated and were non-toxic. The addition of sorbate in combination with nitrite, to the samples prepared in the 10° salometer brine, decreased the rate of toxin production to a greater extent than did nitrite alone. Bloated toxic samples did occur by day 7, in the samples treated with nitrite and sorbate, but to a lesser extent than seen in the other samples prepared in the 100 salometer brines. These results indicate that at this low brine salt concentration, the addition of nitrite or the combination of nitrite and sorbate decreased the rate of toxin production by C. botulinum type E spores in temperature abused smoked whitefish, but none of these treatments produce a safe product. In the set of samples brined in the 20° salometer brine, without nitrite or sorbate, a set of two and four bloated samples were found on days 14 and 21, respectively. However, 93 these bloated samples did not contain toxin and therefore were not recorded in the data presented in Table 11. In the triplicate samples taken for toxin analysis, toxicity was found in the samples taken on day 42. Toxicity did not occur again in this sample treatment until day 56 at which time all three samples taken were toxic. Thereafter, each set of three samples taken were toxic. The addition of nitrite, or nitrite with sorbate, to the 20° salometer brine delayed toxin production until day 63. At this time, the triplicate samples taken for both treat- ments were toxic. Toxicity was evident in all subsequent sets of samples taken until the end of the study on day 83. The data indicate that in the samples brined in the 20° brine, the addition of nitrite, and the combination of nitrite and sorbate, resulted in a reduced rate of toxin production by the C. botulinum type E spores (Table 11). These data also suggest that the use of a 20° salometer brine, especially with the addition of nitrite, does produce a safer product than does the use of a 100 salometer brine. Sorbate, at the level present in the fish, does not appear to greatly affect the rate of toxin production in smoked white- fish prepared in the 20° salometer brine. In the sets of samples prepared using the 30° salometer brines, with or without added nitrite or sorbate, no toxic samples were found even by day 83. The addition of nitrite or nitrite in combination with sorbate in these higher salt 94 concentration treatments did not appear to influence the rate of toxin production by the spores of C. botulinum type E during the temperature abuse storage period of 83 days. As a means of explaining the effect of salt, nitrite and sorbate on the rate of toxin production of C. botulinum type E spores in smoked whitefish, a set of uninoculated samples were prepared at the same time. These samples were analyzed for moisture, salt, pH and Aw values. The samples treated with nitrite and nitrite with sorbate were analyzed for residual nitrite and sorbate. The Role of Salt Concentration (wp) on the Rate of Toxin Production by C. botulinum type E Spores in Smoked Whitefish The percent salt in the water-phase (wp) of the smoked whitefish, prepared in 10, 20 or 30° salometer brines with or without added nitrite or nitrite and sorbate is shown in Table 12. This set of data was generated using the actual salt (Appendix 4) and moisture (Appendix 5) values in the following equation (Bratzler and Robinson, 1967). % salt x 100 % salt + % moisture % salt, water-phase = The samples brined in the 100 salometer brines had an average salt concentration (wp) of 2.1 - 2.3%. Samples prepared in the 200 salometer brine contained an average of 3.4 - 4.5% salt, (wp). while the samples prepared in the 30° 95 .oueu—pn—cu no uao v0.55-u use: mumapac3.5%, there was a decrease in the rate of toxin production by C. botulinum type E spores. In the present study, it was found that toxin production occurred in 7 days in the whitefish samples incubated at 27°C when the salt concentrations (wp) were between 2.1 and 2.3% (Tables 11 and 12). These results are in agreement with the findings of Christiansen et al. (1968) who reported toxin production by C. botulinum type E spores at salt concentrations (wp) less than 2.75%. The delayed toxin production by the C. botulinum type E spores in the samples prepared in the 200 salometer brine occurred in a system containing an average 4.5% salt concen- tration (wp). These results agree with those of Abrahamsson et al. (1966) who showed that in Robertson's chopped meat medium, a 4.5% salt (wp) inhibited toxin production by C. botulinum type E spores for 90 days at 30°C. The addition of nitrite and sorbate to the 20° salometer brines decreased the rate of toxin production in the finished product, but the nitrite and/or sorbate at this salt level did not totally prevent toxin production by the C. botulinum type E spores. 97 In this study only the samples brined in the 30° salo- meter brines would be legal in the state of Michigan which requires a minimum 5% salt in the water-phase portion of the fish. The present study confirms the safety of a fish product containing >5% salt (wp), but it also indicates that a product containing 4% (wp) with added nitrite would be a safe product for up to 56 days at 27°C. The state of Michigan requires that smoked fish products must be processed and stored at refrigerator temperatures and must be sold or consumed within 14 days after processing. A lower salt containing smoked fish product, especially one containing nitrite, would be a safe product under these stipulations. The Effect of Residual Nitrite on Toxin Production by C. botulinum type E Spores The residual nitrite levels in the smoked whitefish samples are shown in Table 13. On day 0, i.e., immediately after the fish was packaged, there was an average nitrite concentration of 63 mg/kg fish. The nitrite depleted gradually until on day 42, nitrite was no longer detectable by the method used. These results are in agreement with those of Christiansen (1980) who reported that at two ingoing nitrite levels (156 and 50 mg/kg), the residual nitrite declines to approximately 5 mg/kg in meat systems after 21 days of storage at 0°C. He also found that the inhibition of C. botulinum was much greater with the higher concentration 98 .me\me m.NF me: eoeeooeoe to “weep .opeeeooeoe see u oz .mumowporcp a? use omvgcoo mmmxpmcmp mupguwc Pmoommmz .mp «Pooh 99 of residual nitrite. The decrease of detectable nitrite levels in cured meats during storage has been attributed to the reactivity of the nitrite with the meat components (Cassen et al., 1979). Cassens et a1. (1974) reported that because nitrite reacts with various components of the muscle, less than 50% of the added nitrite can be detected immediately after processing. However, the measurement of residual nitrite does not always accurately measure all of the nitrite present. The sensi- tivity of the official AOAC method for residual nitrite analysis can be affected by the ascorbate in the meat system. The ascorbate can destroy some of the nitrite before the color reaction step. It has also been reported that breakdown products of other chemicals can be detected by this method (National Academy of Sciences, 1981). The concentration of residual nitrite required to insure safety against the outgrowth of C. botulinum spores in cured meats has not been fully established. Bowen and Deibel (1974) and Hustad et a1. (1973) reported that residual nitrite concentrations as low as 50 mg/kg were inhibitory to toxin production in frankfurters for 56 days at 27°C. These results generally agree with the results of the present study on fish. It should also be recognized that as nitrite concentration decreases over time, the level of viable C. botulinum cells also decrease (Christiansen, 1980). This means that the 100 botulinal safety of the meat system depends on the presence of a residual nitrite concentration high enough to be main- tained in the system until the viable cell level has decreased to a point where growth can no longer occur (Christiansen et al., 1978; Cook and Pierson, 1983). The inhibition of C. botulinum type E spores by nitrite in combination with 4.3 percent salt (wp) as observed in this study is in agreement with the data reported by Pelroy et al. (1982). These authors reported that with the addition of nitrite (100 mg/kg, residual), the salt level could be lowered to 2.5% (wp) and still inhibit C. botulinum type E spores (102 spores/g) in smoked salmon steaks held for 7 days at 27°C. When the salmon was inoculated with 104 spores/g, a 3.1% salt (wp) was required with 100 mg/kg residual nitrite, to inhibit the type E botulinal outgrowth for 25 days at 27°C. Therefore, it is not unrealistic that with an inoculation 6 level of 10 spores/g, the inhibition of outgrowth of C. botulinum type E spores for 56 days at 27°C would require a percent salt (wp) of 4.3 and a residual nitrite level of 63 ' mg/kg. The addition of sorbate with the nitrite provided the same inhibitory effect as the nitrite alone, but at a salt concentration (wp) of 3.4%. 101 The Effect of Residual Sorbate on Toxin Production by C. botulinum type E Spores The residual sorbate concentrations of the smoked white- fish treated with nitrite and sorbate are shown in Table 14. Initially, the smoked whitefish had an average residual sorbate level of 195 mg/kg. Sorbate levels decreased and were no longer detectable by day 42. Since the sorbate was added to the brines at a level of 2600 mg/kg, the uptake of sorbate by the fish tissue appeared to be approximately 10% of the amount added to the brine. Although the level of residual sorbate in the smoked whitefish was low, the addition of sorbate in combination with nitrite appeared to enhance the inhibition of toxin production by the C. botulinum type E spores. As discussed earlier, the nitrite/sorbate combina- tion prevented toxin production for 63 days at 27°C at a salt concentration (wp) of 3.4%, thus indicating that this treat- ment was equally as effective as the nitrite treatment. The ability of this low level of sorbate in combination with nitrite to inhibit toxin production at a lower salt level than the salt and nitrite combination indicates the possi- bility that a stronger, synergistic effect would be possible if the level of sorbate in the fish had been greater, i.e., closer to the 2600 mg/kg level recommended for inhibition of toxin production in cured meats (Robach, 1980). The synergistic interaction between nitrite and sorbate noted in the literature is thought to be the result of the 102 .o¥\os op we: oozums we xuv>vuvmcmm .mpomoomumo no: u oz .mumovpaov c? poo vmpccau mew: mmmxpmcm~ wumncom —m:uommm .ep «Pomp 103 combination of the individual effects of these compounds. Sorbate/sorbic acid affects spore germination and retards cell development. Nitrite inhibits the outgrowth of the germinated cell, therefore the combination of these two compounds results in an increased inhibition of botulinal toxin production (Sofos et al., 1979a). However, this effect has only been shown with sorbate/sorbic acid levels of 2000-2600 mg/kg (Sofos et al., l979a, 19790). The Effect of pH on toxin production by C. botulinum type E Spores The inhibitory effect of both nitrite and sorbate on C. botulinum type E spore outgrowth is pH dependent. With both compounds, the undissociated acid form is responsible for the inhibitory effect on C. botulinum outgrowth and toxin production (Samson et al., 1955; Freese et al., 1973). Initially, the smoked whitefish had an average pH value of 6.4 (Table 15). However, during storage there was a gradual decrease to an average of 5.4 for the samples prepared in the 10° salometer brines, and to 5.5 for the remaining samples. The pH was never acid enough to totally inhibit botulinal growth. It has been indicated that as the pH decreased from 7.0 to 5.5, the inhibitory capacity of nitrite on botulinal outgrowth increases (Pelroy et al., 1982; Roberts and Ingram, 1966). 104 .mumum—aau cw use nmvggmu mum: z mom pmc In em me we mm «P o ucmspmmc» moogoum we mxmo .mmcvco comeo_mm mmcmmu ucmcmmewc cw vmcpca gmwwmuwgz umxosm do mmnpm> :g mg» co cum an mmmcoum mo uummmm .mp mpnmh w 105 The maximum inhibitory effect of nitrite has been shown to occur at pH values between 4.5-5.5 (Holley, T981). Potas- sium sorbate is a strong inhibitor of C. botulinum at a pH value of 5.7 (Smoot and Pierson, l981). In the present study, the whitefish had an overall initial pH value of 6.4. The samples became more acidic during their storage at 27°C, reaching pH values of 5.4-5.5. These pH values are ideal for the maximization of the inhibitory effect of both the nitrite and the sorbate on the botulinal spores and cells present in the samples. The decrease in the pH during the storage of the samples is presumably due to the presence of lactic-acid producing bacteria. The Effect of Water Activity (Aw) on Toxin Production by C. botulinum type E Spores There are very few literature reports describing the effect of Aw on the outgrowth and toxin production by C. botulinum type E in smoked fish. However, it has been reported that lowering the Aw through the use of salt is an effective means of inhibiting clostridia (Roberts and Smart, 1976). In addition, C. botulinum type E has been shown to be more sensitive to salt than are types A and B (Roberts and Smart, 1976). Therefore, in the present study, the samples were analyzed for their Aw levels during the first 56 day of incubation at 27°C. There was an apparent difference in the initial water activities of samples containing the three different levels 106 of salt (Table 16). The samples brined in the 10, 20 and 30° salometer brines had average Aw's of 0.988, 0.976 and 0.962, respectively. The decrease in the initial Aw's with increas- ing levels of salt observed in this study is not unexpected since salt has been shown to act as a preservative by lowering the Aw of food (Deng, 1977). During storage (56 days) at 27°C, there was a gradual decrease in the Aw levels in the samples. As noted earlier, there was also a decrease in the pH of these samples over time in these samples. Baird- Parker and Freame (1967) reported that at a pH value of 7.0, an Aw of 0.97 or greater was required for the outgrowth of C. botulinum type E spores. They also noted that at pH values between 5.5-6.0, an Aw of 0.98 or greater was necessary for botulinal type E outgrowth. In the initial (day 0) sample which had an overall pH value of 6.4, C. botulinum spores should be able to grow and produce toxin in the samples prepared in the 10 and 200 salometer brines since the Aw's averaged 0.988 and 0.978, respectively. In this study, toxicity did occur in the samples prepared in the 10° salometer brines within the first seven days at 27°C (Table ll). However, no toxicity was observed in the samples prepared in the 200 brines until day 42 (Table ll). The absence of toxicity in the sample prepared in the 200 salometer brine alone requires an explanation. There are two events that could be occurring based on the results presented by Baird-Parker and Freame 107 .mumuwpaau :. use umpgsmu mew: mmmapmcmp app>vuuo noun: .0? manh 108 (1967) that at pH values of 7 and 6, the level of Aw required for botulinal growth would be 0.97 and 0.98, respectively. Since the sample containing salt alone (200 salometer brine) had an Aw value of 0.974, this level of Aw could be low enough to inhibit the growth of the botulinal spores present since the pH was closer to pH 6 than pH 7. By day 7, this sample had a pH of 6.0 (Table 15) and an Aw of 0.968 (Table 16). At this point, the Aw was well below the 0.98 required for botulinal growth at pH 6. The samples prepared in the 200 brine alone became toxic on day 42. This indi- cates that the C. botulinum spores/cells were able to adapt to the lower Aw level and were able to grow and produce toxin. This finding is in agreement with those of Troller (l983) who reported that bacteria in Aw values lower than minimum growth level suffer a prolonged lag or "resting" phase in their growth curve. This phase is therefore more an "adaptation" phase because the organism undergoes an adapta- tion in order to survive the lower Aw level. Once the adaptation occurs, growth can resume, although usually at a slower rate initially (Sperber, 1982). Sperber (1982) also noted that spores can usually germinate at Aw's substantially lower than would permit growth. Baird-Perker and Freame (1976) reported that the minimum germination Aw for C. botulinum type E spores was 0.93, while growth required a minimum of 0.97 Aw, when salt was used to establish the Aw level. 109 The production of toxin in the samples prepared in the 200 salometer brines with added nitrite and nitrite/sorbate on day 63 was not unexpected. Both residual nitrite and sorbate were not detectable on day 42. Once these compounds were depleted, the surviving viable cells could begin growing and toxin production would occur (Christiansen et al., 1978). The lack of toxin in the samples prepared in the 300 brines indicates that the salt level initially reduced the Aw to a level where germination could possibly occur but growth could not. The combined effects of the decreasing pH and Aw's further inhibited the outgrowth and toxin production by the botulinal cells. The ability of salt at this level (6% wp) to inhibit botulinal outgrowth has been discussed previously. The addition of nitrite and nitrite/sorbate did not affect the rate of toxin production at this high level of salt. This indicates that the salt level/Aw effect in these treatments was the primary inhibitor of the C. botulinum spores. These results agree with the findings of Segner et al. (1966), who reported that in a trypticase- peptone-glucose medium, a salt concentration (wp) of 5% inhibited toxin production by C. botulinum type E spores for 1 year in samples incubated at 37°C. 110 Effect of Salt Level and Nitrite on the Lipid Stability and Organoleptic Acceptability of Smoked Whitefish Chemical Analyses The state of Michigan (Regulation 541) requires that smoked fish must contain at least 5% salt (wp) in the finished product. However, the acceptance of this high salt product in today's society could be limited as consumers are becoming increasingly aware of the relationship between dietary salt and the incidence of hypertension and strokes or heart attacks. Therefore, the following study was conducted to determine the effect of salt and nitrite on the lipid stability and organoleptic acceptability of smoked whitefish. Chemical analyses, including salt, moisture, fat, TBA number residual nitrite and N-nitrosamine determinations were carried out on samples taken on days 0, 7, l4 and 22. Day 0 represents the initial day of the storage study, approximately 20 hours post smoking (cooking). Taste panel evaluations were made on all sets of samples except day 22, because of the 14 day limitation dictated by Regulation 541. The percent salt in the water—phase (wp) of the white- fish was calculated from the salt and moisture data (Appendices 6 and 7, respectively). These were obtained by chemical analysis. The percent salt levels (wp) of whitefish brined in different degree salometer brines prepared with 111 and without nitrite are presented in Table 17. The control samples, which were soaked in water during the brining period contained minimal salt levels (0.34 to 0.40%) in the water- phase portion of the fish. The average percent salt levels (wp) in samples brined in the 10, 20 and 300 salometer brines were 1.8-2.1, 4.0-4.3 and 6.3—6.9, respectively. These values agree with the data presented earlier. There was a wide variation in the levels of fat in the smoked whitefish samples that had been prepared with different levels of salt in the brine (Table 18). Even with sample randomization, the fat contents ranged from 7.95-12.30 percent, with overall average of 9.43 percent. These fat levels are higher than the 3.84% and 5.2% values reported by Awad et al. (1969) and Exler and Heichrauch (1976), respectively. This difference in fat levels could be due to seasonal variation in the fish since there is a seasonal cycle that occurs in the catching of the fish (Stachiw et al., 1984). However, the fat levels reported in this study agree with other fat values for whitefish that have been determined in this laboratory in other studies (Maruf, 1983, unpublished data). Fish lipids have been reported to contain significant quantities of long chain polyunsaturated fatty acids (Khayat and Schwall, 1983). The complexity of the fatty acids in the total, triglyceride and phospholipid fraction of whitefish lipid are shown in Table 19. The data presented do not 112 Table 17. Percent salt (water-phase) of smoked whitefish brined in different degree salometer brines prepared with and without nitrite. Days of Storage Treatment _ o 7 14 22 x Percent salt (water-phase) Controla 0.3 0.5 0.3 0.5 0.4 Contro1a + nitrite 0.5 0.3 0.2 0.4 0.3 100 Brine 1.9 1.7 1.6 2.0 1.8 10° Brine + nitrite 2.1 2.1 2.1 2.1 2.1 20° Brine 3.3 3.9 4.6 4.2 4.0 20° Brine + nitrite 4.2 4.3 4.7 4.0 4.3 30° Brine 6.3 7.3 7.1 6.9 6.9 30° Brine + nitrite 6.0 6.5 6.4 6.4 6.3 aNo salt used in the brine. 113 Table 18. Percent fata in smoked whitefish brined in different degree salometer brines prepared with or without nitrite. Days of Storageb Treatment 0 7 14 22 7 Percent fat (%) Controlc 9.22 14.66 8.26 10.01 10.54 Controlc + nitrite 12.85 14.80 8.13 13.43 12.30 100 Brine 8.57 5.53 7.60 10.10 7.95 10° Brine + nitrite 7.77 7.87 8.86 9.70 8.55 200 Brine 8.23 8.87 8.51 8.56 8.54 20° Brine + nitrite 7.79 8.57 8.52 9.22 8.53 300 Brine 9.34 7.85 10.78 9.97 9.49 30° Brine + nitrite 9.36 11.20 7.44 10.25 9.56 aAnalyses were carried out in duplicate. bStored at refrigerated temperature (4°C). CNo salt used in the brine. 114 Table 19. Fatty acid composition of the total, triglyceride and phospholipid fractions of whitefish lipid. Fatty acid Total Triglyceride Phospholipid Area percent 12:0 0.19 0.10 1.05 12:1 0.03 -- 0.71 13:0 0.04 0.21 0.56 14:0 4.31 1.94 3.38 14:1 0.43 0.19 0.71 15:0 0.34 0.23 0.66 16:0 13.59 19.55 8.51 16:1 10.57 7.93 13.60 18:0 2.45 4.78 2.13 18:1 24.30 11.75 16.29 18:2 2.23 0.82 2.32 18:3 2.36 1.56 2.93 20:0 1.39 0.50 1.60 20:1 1.05 0.24 2.00 20:2 0.19 0.31 0.85 22:0 0.61 0.35 1.06 22:1 8.63 10.68 6.33 22:5 0.52 0.52 1.47 24:0 2.40 14.49 2.43 24:1 5.79 1.38 4.19 115 represent all of the fatty acids present in the lipid frac- tions and only includes those that could be identified by fatty acid standards. While similar fatty acids are found in the total, tri- glyceride and phospholipid fractions of the whitefish lipid, there were some differences in the quantities of specific fatty acids in each fraction. The predominant unsaturated fatty acids in the triglyceride fractions were, in order of prominence, Cl8:l (11.75%),C22zl (10.68%), and C16:1 (7.93%), while in the phospholipid fraction C18:1 (16.29%) was most predominant followed in sequence by C16:l (13.6%), C22:1 (6.33%), C24:1 (4.19%) and C18:3 (2.93%). Based on the fatty acid data generated, the phospholipid fraction contained 21.4% of saturated fatty acids and 54.0% of unsaturated fatty acids. This gives an unsaturated to saturated fatty acid ratio of 2.5. The triglyceride fraction contained 42.2% of saturated fatty acids and 35.4% of unsaturated fatty acids. Therefore the triglyceride fraction had an 0.8 ratio of unsaturated fatty acids to saturated fatty acids. The higher level of unsaturation in the phospho- lipid fraction is not unexpected. It has been reported that within the same species, the phospholipids generally contain a higher level of unsaturated fatty acids (Pearson et al., 1977). Exler and Weichrauch (1976) analyzed whitefish for its fatty acid composition and reported values as follows for 116 total lipids: C14:0, 2.1%; C16:0, 12%; C16:1, 10.6%; 018:0, 1.05%; C18zl, 27.4%, C18:2, 5.48%; 018:3, 3.59%; C20:0, 2.11% and C22:l, 4.2%. In comparison, the total lipid fraction of the whitefish analyzed in the present study had a much higher level of C22:1 (8.63%); however, there were lower levels of C18:2 (2.23%) and C18:3 (2.36%). Braddock and Dugan (1969) studied the fatty acid compo- sition of Coho salmon and reported that fatty acids ranged from Cl4:0 to C22:6 in chain length and in degree of unsatu- ration. In the Coho salmon, the levels of C16:0 and Cl6:l were 10.8 and 10.9%, respectively. These values were compar- able to those reported for the total lipid fraction in the present study. Literature values for the levels of C16:0 and Cl6:l, in a variety of fish species, range from 10.8 to 24.8% and 4.0 to 20.8%, respectively (Braddock and Dugan, 1969; Ackman and Eaton, 1971; Ackman et al., 1975; Exler and Weichrauch, 1976; Mai and Kinsella, 1979; Leu et al., 1981; Gall et al., 1983). Therefore, the levels of 016:0 and C16:l reported for the whitefish lipids in the present study are reasonable. The whitefish used in this study had an average fat content of 9.43%. When this lipid was separated into its triglyceride and triglyceride fractions, it was found to contain 78% triglyceride and 22% phospholipid. These values agree with those of Awad et a1. (1969), who reported that the whitefish lipid contained 80% triglyceride and 20% phospholipid. 117 In contrast, Igene et al. (1979) reported that beef lipids contain 8% phospholipids, while triglycerides account for the other 92%. In chicken, the phospholipid is 14% of the total lipid and the triglyceride is 86%. Pearson et al. (1977) indicated that poultry and fish contain more phospho- lipids than red meats. The results of this study support this statement. Residual Nitrite Levels of Smoked Whitefish during Storage Residual nitrite levels in these samples at the start of the refrigerated storage period ranged from 37.5 to 53.1 mg/kg with an average of 46 mg/kg (Table 20). There was a gradual depletion of nitrite during refrigerated storage as was expected. Residual nitrite is the amount of nitrite that is analytically detectable or recoverable by chemical analysis (Ito et al., 1983). The depletion of nitrite with time in cured meat is well established in the literature. The fate of the nitrite in the cured meat system has not been clearly demonstrated. Mirna and Hofmann (1969) reported that nitrite and thiol groups disappeared equimolarly in a minced meat system containing nitrite. They suggested that nitrosothiol formation was occurring and was responsible for nitrite depletion. Kubberod et al. (1974) showed that nitrosothiol formation did not play a major role in the nitrite loss from cured meats. Woolford et a1. (1976) reported that one of the major pathways for nitrite loss in cured meats could be the 118 Table 20. Residual nitrite levelsa of smoked whitefish brined in different degree salometer brines containing nitrite. Days of Storageb Treatment 0 7 14 22 Nitrite (mg/kg) Contro1C + nitrite 46.9 36.3 28.8 10.0 10° Brine + nitrite 37.5 36.3 28.8 7.5 20° Brine + nitrite 53.1 37.5 31.3 16.3 30° Brine + nitrite 46.3 37.5 31.3 6.0 aAnalyses were carried out in triplicate. bStored at refrigerator temperature (4°C). CNo salt used in the brine. 119 reaction between nitrite and heme proteins. Ito et a1. (1983) noted that an assumption has been made that the protein-bound nitrite is unavailable for further reactions. However, they showed that the protein-bound nitrite c0mplex is not stable and is potentially reversible. They suggested that the “bound" nitrite could function in transnitrosation reactions. The use of nitrite in meat systems has been under scrutiny since it has been shown that nitrite reacts with secondary amines and amino acids to produce N-nitrosamines (Gray, 1976). N-Nitrosopyrollidine (NPYR) and NDMA, both commonly identified in cured meats, have been shown to be carcinogenic (Gray, 1976). Therefore, before advocating the use of nitrite in smoked whitefish, a product that could have high levels of amines (Castell et al., 1971), it was considered necessary to establish whether N-nitrosamines could be formed in nitrite-treated smoked whitefish. Smoking has been shown to result in the formation of N-nitrosothiazolidine (NTHZ) in smoked cured meats (Pensabene and Fiddler, 1983a). Mandagere et a1. (1984) has related the formation of NTHZ to the reaction of formaldehyde in smoke with cysteamine and/or cysteine in the meat. This reaction can subsequently result in the formation of NTHZ and/or N-nitrosothiazolidine carboxylic acid (NTCA). Since the production of smoked fish uses a heavy smoking process, it was necessary to evaluate the possible risk of NTHZ 120 formation in the finished product. The nitrite-treated samples in this study were analyzed for volatile N-nitrosamines (NDMA and NTHZ) and NTCA. No detectable level of these compounds were found in any of the samples tested. The absence of NDMA in these samples is not unexpected since most of the N-nitrosamines reported in fish have been isolated in marine species fish. Marine species fish have been shown to contain high levels of trimethylamine which is degraded to formaldehyde and dimethylamine. The dimethylamine is readily nitrosated (Dyer and Mounsey, 1945; Shewan, 1951; Castell et al., 1971; Spinelli and Koury, 1979; Sikorski and Kostrich, 1982). Stability of Smoked Whitefish Lipids during Storage The stability of lipids in whitefish prepared in brines containing different salt concentrations was evaluated by the TBA procedure. The results of these analyses are presented in Table 21. The control samples prepared without nitrite had an average TBA number of 2.04 after 22 days of storage at 4°C. The addition of salt to the smoked whitefish resulted in increased levels of rancidity as measured by TBA number. The samples brined in the 10, 20 and 300 salometer brines without nitrite had TBA numbers of 2.85, 2.54 and 2.68, respectively, after storage for 22 days. These data indicate that each level of salt had a significant (P<0.01) prooxidant effect in this system; however, no correlation was 121 Table 21. TBA valuesa of smoked whitefish brined in different degree salometer brines prepared with or without nitrite. Days of Storageb Treatment 0 7 14 22 TBA value Controlc 1.06 2.07 1.94 2.04d Controlc + nitrite 0.32 0.42 0.34 0.41e 10° Brine 1.06 2.30 2.16 2.85f 10O Brine + nitrite 0.45 0.67 0.52 0.879 20° Brine 0.97 1.76 1.60 2.54h 20° Brine + nitrite 0.54 0.59 0.41 0.839 30° Brine 0.85 1.77 2.27 2.68h 30° Brine + nitrite 0.22 0.48 0.44 1.861 aAnalyses were carried out in duplicate. bSamples were stored at 4°C. cNo salt used in the brine. Values with different letters are significantly different (P<0.01). 122 found between the increasing levels of salt and the increased rate of lipid oxidation (TBA number). A prooxidant effect of salt has not been clearly shown in the literature. Originally, it was postulated that salt promoted the activity of lipoxidases in meats (Banks, 1937; Lea, 1939). However, Banks (1944) and Tappel (1952, 1953) reported that meat did not contain lipoxidases. They felt the prooxidant effect of salt was due to a catalytic effect of the heme pigments. These researchers failed to demon- strate any evidence of heme pigment's ability to catalyze the prooxidant effect of salt. Gibbons et al. (1951) reported that when salted bacon sides were held at frozen temperatures, rancidity increased as the temperature decreased. Chang and Watts (1950) reported that the addition of salt, at levels of 15% or greater, to lard had a direct effect on the rate of rancidity development in the lard. Ellis et al. (1968) reported that in freezer stored salt cured pork, the rate of oxidation occurred at a rapid rate. Ellis et a1. (1968) also reported that as the level of salt in the pork increased so did the rate of oxidation. They found that high pr0portions of lean increased the rate of oxidation and that the direct effect of the salt on oxidation did not appear to involve a reactive chloride ion. The prooxidant effect of salt has been suggested to be the result of heavy metal contamination of the salt. Denisov and Emanuel (1960) noted that heavy metals 123 (iron, copper and chromium) can catalyze oxidative rancidity. The flake salt used in curing has been found to contain iron and copper (Olson and Rust, 1973). More research is needed to clarify the effect of salt on the rate of lipid oxidation in food systems (Love, 1983). The addition of nitrite to the unsalted whitefish samples resulted in a significant (P<0.01) decrease in the rate of lipid oxidation as measured by the TBA number (Table 21). This indicates the strong antioxidant effect of nitrite in this system. The efficacy of nitrite as an antioxidant was diminished as the level of salt in the sample increased. The TBA numbers of the samples brined in the 10, 20 and 300 salometer brines with added nitrite were 0.87, 0.83 and 1.86, respectively. These data again suggest a prooxidant effect of salt in this system, although further studies are necessary to confirm this observed trend since only one experiment was carried out. The antioxidant effect of nitrite on the development of WOF in cooked meats has been shown in red meats and poultry. Fooladi et al. (1978) reported that the addition of nitrite (156 mg/kg) to cooked beef, pork and chicken resulted in a two-fold reduction in their TBA numbers. MacDonald et al. (1980) reported that the addition of nitrite to cured hams significantly reduced TBA numbers. Chen et al. (1984) showed that the addition of nitrite to meat pigment extracts pre- vented the release of heme iron from the porphyrin ring. It 124 has been suggested that nitrite inhibits the development of WOF by reacting with iron porphyrins to form a stable non- reactive complex (Zipser et al., 1964; Igene et al., 1979). Kanner et al. (1981) proposed that the antioxidant effect of nitrite in cured meat could be attributed to the formation of nitric oxide. Nitric oxide is capable of interacting with metals, especially heme and nonheme compounds. Shahidi et a1. (1985) reported that the addition of nitric oxide gas to a buffered solution containing hemin resulted in the formation of cooked cured-meat pigment, dinitrosyl ferro- hemochrome. Igene et a1. (1985) suggested that in addition to its ability to bind heme and nonheme pigments and to chelate metals, nitrite can inhibit the development of WOF by stabilizing the unsaturated lipids in the membranes of the meat tissue. Overall the TBA numbers observed in this set of samples agree with literature values pertaining to the oxidation of fish, including whitefish. Biggar et al. (1975) reported that in canned whitefish that were opened and exposed to refrigerator temperatures for four days, the TBA values reached a maximum value of 1.97. These fish were assumed to contain 2% lipid. Awad et al. (1969) studied lipid oxi- dation in whitefish held frozen at -10°c for up to 16 weeks. They found a maximum TBA value of 2.26 at the end of the 16 weeks of storage. These fish contained 3.8% lipid. 125 Mechanically deboned dogfish containing 8.7% lipid was stored at -20°C for 6 months had TBA values as high as 6-7 (Nakayama and Yamamoto, 1977). In contrast, it has been shown that cod containing 0.6% lipid and stored at -20°C for 6 months had TBA values between 0.5 and 1.0 (Nakayama and Yamamoto, 1977). Sensory Analysis of Smoked Whitefish Samples On each taste panel day, the treatment samples were divided into two sets of four in which all salt levels were represented as were nitrited samples. The panelists were asked to evaluate which treatment(s) produced the most desirable product relative to the perceived desirability of specific traits, which included odor, saltiness, texture, flavor and overall acceptability. The procedure of Kramer (1963) was used to statistically analyze the sensory evaluation data. An ANOVA (Gill, 1978) was run using all eight samples on each day. This was done in order to determine if there were any significant differ- ences between the eight samples on each test day. Bonferroni's t—statistic comparisons (Gill, 1978) were made on specific sets of treatment samples to determine if the panelists could detect significant differences between specific treat- ments. Bonferroni's comparisons were used because of the greater sensitivity to differences when limited comparisons are made between treatments (Gill, 1978). 126 On day 0, the first day of the study, no differences were recorded in the odor of any of the sample treatments (Table 22). The samples found to be significantly (P<0.05) more desirable in their saltiness, flavor and overall acceptability were those prepared in the 200 salometer brine, with or without nitrite. There was no significant difference between textures of the samples. On day 7, no significant (P<0.05) differences in odor of the sample treatments were perceived (Table 23). The most desirable (P<0.05) level of saltiness, texture, flavor and overall acceptability was found in the samples prepared in the 200 salometer brine containing nitrite. The panelists found the unsalted sample containing no nitrite, signifi- cantly (P<0.05) less desirable in all of the traits except odor. On day 14, no differences in odor were detected among any of the eight treatment samples (Table 24). The samples found to be significantly (P<0.05) most desirable in all traits except odor were those prepared in the 200 salometer brines, with or without nitrite. The sensory data from all eight treatments were analyzed by ANOVA and Bonferroni's comparisons. On day 0, significant differences were found in the desirability of saltiness (P<0.01), texture (P<0.01), flavor (P<0.01) and overall acceptability (P<0.001) in all samples evaluated. Bonfer- roni's contrasts were made between the samples brined in the 20 and 300 salometer brines only. The Bonferroni's contrasts 127 Amvmapmcm mcwxceg gmsmsxv mpnogrmmu ammo; Amvmxpmce mcpxcag gmsmng mpnegwmmu umozr beneaecee Apo.ovav speeeewceem.m nee meeeeep eeneecc.e get: mePeEem».x maven ecu cw 6mm: upmm oz re a mPQMmemt umwmp u e mmpnmgwmmv umoE n F ”wpmumm xm.m N.N o.~ xo.m o.~ muvsuwc + maven com x.~.~ em.F m._ x.~.~ N.~ ee_eeee + eewem com m.~ m.N ~.N m.m ~.N muvguw: + mcwgm cop ano.m «n~.m «nm.m new.m m.~ apogucou xw.~ m.~ o.N xm.m w.~ mcwgm com x.m.~ .m.~ m.~ xem.F N.m eeeem com m.~ N.N n.~ n.~ m.~ mcwgm cop m.m o.v m.m N.m ~.N Fogucou «e «B #e n seePBDeeeeoeq Fmem>o Lo>mru mLzume mmmcwupmm LOUD ucmEumeh nonexcmm .mmcpgn Lmumsopmm mmemmu ucwgmmewu c, umcwcn mmFgEem zmvwmawzz uwonm to» o xmu no museum xcomcmm mmmgm>< .NN mpomh 128 Amwmxpece mcpxcms cmsmcxv wpnmcwmmc anew; «r Ampmxpmcm unexceg LmEmng m-nmgwmmc amaze .eeeeeceee Amo.ovev speeeeeceem_m ate meeeeep eeeeeceee ewe: nepee~mx.x m:_gn us» cw new: uFem ozn mpnmgwmmu umemp n e "mpnaspmmu pmos n F "mpmumm trN.N F.~ m.p ~.N m.~ muwguwc + mcvgm oom nv.p xrm.p «e p km P ~.N mapsuw: + mcwgm com m.F o.~ e.m F.N o.~ mavsupc + mcwgm ooP erm.m «rm.m ~.m p.m m.~ muwguwc + apogucou P.N ~.N m.~ ¢.~ m.~ mcwgm oom m.m xm.~ ¢.~ ¢.~ m.~ mcvgm cow ~.m P.m F.m m.~ _.~ mcwgm cop ato.m t.2..m.m are.m ne~.m o.~ nFOLucoo xuwppnmuamuu< Fpocm>o Lo>mpd mczuxm» mmmcvupmm coco pcmspcmgp mmcwxcmm .mmcwcn Lmumso—mm mmgmmv acmcmmwvn cw uwcwen zmvmmurz; nmonm co» m man an mweoum xgomcmm wmmgm>< .mm «Fame 129 Amwmxpmco mcwxceg Lmsmng mpnegvmmu ummmgna Amwmxpmcm mcmxcog emancxv mpnmgwmmv umoze ucwgmmmmu Amo.oqu xpucoUVmpcmwm mew mgmaump acmgmmmwu saw: mmpnsmmx.x maven mg» cw cam: upom oz 5 mpnmgwmmu ummmp n e "mpnmgwmmv umos F "mpouma mm.~ F.N o.~ »N.N F.m avenue: + maven com x«P.P «m.p m.F xnm.P m.~ ouvcuvc + mcwgm com .4. o.~ m.~ m.P F.N ¢.~ muFLHVc + mcrgm oop nam.m ««N.m new.m ero.m m.~ muwsuv: + apogucoo m.~ F.~ o.~ o.m m.~ mcwcm com no.~ «o.N «o.~ «o.~ F.N mcwcm com m.m o.~ o.m p.~ m.~ mc+gm oo— ««N.m rem.m eno.m «rp.m m.~ apocucou »e_P,eeeeeeeq FPMLO>O Lo>wpm szuxmh mmGCquwm LOUD HcmEummLP mmcwxcma .mwcPcn cmumeopmm mmgmmu ucmcmwewu cw cmcwgn gmrmmuwgz umeEm Lo; op xmu um museum xgomcmm mmagm>< .em mpneh 130 showed that the samples brined in the 200 salometer brine with or without nitrite had a significantly more desirable (P<0.01) level of saltiness than the sample prepared in the 300 salometer brines prepared with or without added nitrite (Table 22). No differences in texture were found between the samples prepared in the 20 and 300 salometer brines. However, the 200 salometer brined samples were significantly (P<0.01) more desirable in flavor and overall acceptability than the 300 salometer brined sample. On each sampling day, the eight treatments were tested for sensory differences that could result from the presence of nitrite in combination with the four levels of salt. However, no nitrite effect was found in any of the sensory data, indicating that the panelists were unable to distinguish between the nitrite and non-nitrited smoked whitefish. On day 7, significant differences were found between the eight samples in texture (P<0.05), flavor (P<0.05) and overall acceptability (P<0.001) by ANOVA. However, the Bonferroni's comparison showed that the 200 salometer brined sample without nitrite was significantly (P<0.05) less desirable than the sample prepared in the 200 salometer brine with nitrite (Table 23). 4 On day 14, significant differences were found in salti- ness (P<0.01), texture (P<0.01), flavor (P<0.05) and overall acceptability (P<0.05). Again it appeared that most of the significant differences occurred in the samples prepared in 131 the control and the 100 salometer brines. The fish sample prepared in the 200 salometer brine with nitrite was signifi- cantly (P<0.05) more desirable in saltiness and overall acceptability than the sample prepared in the 300 salometer brine with nitrite (Table 24). Over time, the majority of the differences seen in the ANOVA analysis was due to the unacceptability of those samples prepared in the control and the 100 salometer brines. These samples were not tested using the Bonferroni's comparison because they would not be a commercially viable product. These data indicate the preference by the panelists for a smoked whitefish containing 2.5 to 2.85% salt or 4.0 to 4.3% salt (wp). The 4% salt (wp) concentration is obviously lower than the 5% salt (wp) that is required by Michigan Regulation 541. The botulinal safety of the whitefish brined in a 200 salometer brine and which contain 4 percent salt (wp) in the finished product has been demonstrated in the initial phase of this study. It was also demonstrated that the inclusion of nitrite increases the safety of this product by delaying the outgrowth of C. botulinum type E spores. In addition, the panel data indicate that the addition of nitrite to smoked whitefish does not affect the desirability of the product. Nitrite in the smoked whitefish did not result in the forma— tion of detectable levels of N-nitrosamines. Therefore, the addition of nitrite in combination with a 4% salt (wp) concentration could be recommended for use in the production 132 of smoked whitefish. Effect of Smoke Type and Level on the Lipid Stability and Organoleptic Acceptability of Smoked Whitefish Chemical Analyses The salt concentration in the water-phase of smoked white- fish prepared with different levels of liquid smoke or smoked with woodsmoke was calculated from the salt and moisture data (Appendices 8 and 9, respectively). The results are presented in Table 25. The fish samples had salt (wp) concentrations that ranged from 4.8 - 6.5% salt (WP) with an overall average of 5.9% salt (wp). These results are consistent with the values reported in previous phases of this study. Only one treatment had an average salt concentration (wp) that was less than the 5% required by Michigan law. There was much variation in the level of fat found in these samples, even with randomization of samples (Table 26). The average fat content ranged from 9.11 to 12.72% with an overall average of 10.47%. The fat levels in these samples are comparable to the levels reported previously in this study. The initial average residual nitrite levels in the samples, i.e. at day 0, brined with different levels of liquid smoke or smoked with woodsmoke was 53.8 mg/kg (Table 27). There was a gradual depletion of nitrite during the refrigera- ted storage period to a final average level of 27.6 mg/kg 133 Table 25. Percent salt concentration (water-phase) of brined whitefisha treated with various levels of liquid smoke or smoked with woodsmoke. Days of Storageb Treatment _ O 7 14 22 X Percent saltC Woodsmoke 6.6 7.7 6.3 5.8 6.2 Woodsmoke + nitrite 6.2 6.8 6 .4 6.5 0.7% Liquid smoke 5.7 6.0 5.8 5.9 5.9 0.7% Liquid smoke + nitrite 6.1 6.0 6.4 6.8 6.3 1.4% Liquid smoke 5.3 5.9 6.1 6.6 6.0 1.4% Liquid smoke + nitrite 5.6 6.4 6.6 6.7 6.3 2.1% Liquid smoke 4.7 6.5 5.2 5.1 5.4 2.1% Liquid smoke + nitrite 5.3 4.9 4.5 4.5 4.8 aBrine contained 7.89% salt (300 salometer) bStored at refrigerated temperature (4°C) cAnalyses were carried out in triplicate 134 Table 26. Percent fat in brined whitefisha treated with various levels of liquid smoke or smoked with woodsmoke. Days of Storageb Treatment _ 0 7 14 22 X Percent fatC Woodsmoke 9.89 10.65 11.99 10.51 10.76 Woodsmoke + nitrite 9.77 9.58 10.54 8.51 9.60 0.7% Liquid smoke 10.26 12.14 11.50 11.06 11.24 0.7% Liquid smoke 9.52 9.18 9.42 9.33 9.11 + nitrite 1.4% Liquid smoke 7.35 9.25 10.62 10.11 9.33 1.4% Liquid smoke 11.26 10.95 10.87 11.29 11.09 + nitrite 2.1% Liquid smoke 8.52 13.12 8.05 9.93 9.93 2.1% Liquid smoke 11.77 13.54 12.08 13.48 12.72 + nitrite aBrine contained 7.89% salt (300 salometer) bStored at refrigerated temperature (4°C) CAnalyses were carried out in duplicate 135 Table 27. Residual nitrite levels of brined whitefisha with or without nitrite and treated with varying levels of liquid smoke or smoked with woodsmoke. Days of Storageb Treatment 0 7 14 22 Nitrite (mg/kg)c Woodsmoke + nitrite 63.8 41.3 36.3 28.7 0.7% Liquid smoke + nitrite 51.5 42.5 32.5 26.3 1.4% Liquid smoke + nitrite 46.3 45.0 37.5 28.8 2.1% Liquid smoke + nitrite 53.8 47.5 39.8 26.5 aBrine contained 7.89% salt (300 salometer) bStored at refrigerator temperature (4°C) cAnalyses were carried out in triplicate 136 after storage for 22 days. The samples treated with nitrite were analyzed for their N-nitrosamine (NDMA and NTHZ) and NTCA content. No detect- able levels of these compounds were found in any of the. samples tested. These results indicate that the addition of nitrite to smoked whitefish would not result in the forma- tion of any detectable levels of N-nitrosamines. Stability of Whitefish Lipids During Storage The TBA values for whitefish brined with different levels of liquid smoke or smoked with woodsmoke during refrigerated storage are shown in Table 28. The greater antioxidant activity of the woodsmoke was evident by the 5th day of the study. On day 14, it was found that as the level of liquid smoke increased, the rate of lipid oxidation significantly (P<0.01) decreased. However, the same level of effective- ness shown by woodsmoke as an antioxidant was never achieved by the liquid smoke. After 22 days of refrigerated storage, there was a decreased level of oxidation (as measured by TBA number) in the samples brined without nitrite. These lower levels of oxidation indicate a reduced level of malonalde- hyde in the samples. Malonaldehyde in the presence of water exists mainly as the enolate ion. The enolate ion can react with amino acids, proteins, glycogen and other food constituents to form products in which the malonaldehyde exists in a bound 137 Table 28. Effect of smoke type and level on the TBA values of whitefish brineda with or without nitrite. Days of Storageb Treatment 0 7 14 22 TBA value (mg/kg)C Woodsmoke 0.75 0.99 1.17d 1.00d Woodsmoke + nitrite 0.44 0.52 0.54e 0.47e 0.7% Liquid smoke 1.80 3.07 5.17f 3.25f 0.7% Liquid smoke + nitrite 0.80 0.72 1.45d 1.549 1.4% Liquid smoke 0.99 1.77 3.099 3.07f 1.4% Liquid smoke + nitrite 0.65 0.61 1.39d 1.459 2.1% Liquid smoke 1.29 1.37 2.52h 3.22f 2.1% Liquid smoke + nitrite 0.66 0.51 0 59e 0.99d aBrine contained 7.89% salt (300 salometer) bStored at refrigerated temperature (4°C) CAnalyses were carried out in duplicate Values with different letters on a single day cantly different (P<0.01) are signifi- 138 form (Kwon et al., 1965). Buttkus (1967) showed that malon- aldehyde reacted with the e-amino groups in frozen trout myosin. Braddock and Dugan (1973) reported that in freeze- dried salmon steaks and salmon steaks held at -20°C for 1 year, there was an initial rapid increase in the level of oxidation as measured by the TBA number. However, over time the rate of oxidation as measured by increase in TBA number decreased. These researchers reported the presence of C=N functional groups, i.e. Schiff's base type compounds. When the freeze dried and frozen salmon samples were analyzed using UV, visible and fluorescence spectra and IR spectra before and after borohydride reduction, Kikugawa and Ido (1984) reported the formation of fluorescent products as a result of the reaction between malonaldehyde and primary amines. The reactivity of malonaldehyde with amino compounds has been related to the occurrence of browning (Maillard) in food systems (Porter et al., 1983). This reactivity of malon- aldehyde with food components may account for the lower levels of detectable malonaldehyde seen in this study. Reduced levels of rancidity (TBA numbers) over storage time has been shown in fish. Botta et al. (1973) studied the rate of oxidation (TBA number) in Pacific halibut and Chinook salmon during frozen (-30°C) storage for 81 weeks. The maximum level of oxidation in the Pacific halibut occurred on week 62, thereafter the TBA number decreased to a level almost equal to the initial level. In the Chinook salmon, Botta et al. 139 (1973) found that the highest TBA number occurred during the 26th week of frozen storage and that this level of oxidation (TBA number) decreased until on week 77, it was almost equal to the initial level. The addition of nitrite produced a significant (P<0.01) reduction in the rate of lipid oxidation in all systems tested (Table 28). The greatest effects were seen in the combinations of nitrite with woodsmoke and with 2.1% liquid smoke. These data confirm the antioxidant effect of nitrite in smoked whitefish during 22 days of refrigerated storage and confirm the results reported in the previous study. Day 14 data indicate that the 2.1% liquid smoke in combination with nitrite was equal to the woodsmoke with nitrite treatment in inhibiting oxidative rancidity in the smoked whitefish. Bailey and Swain (1973) reported that woodsmoke in combination with nitrite inhibited oxidation in hams to a greater extent than woodsmoke alone. These results are in agreement with the findings of this study. The greater ability of woodsmoke to inhibit oxidation in the smoked whitefish may be related to the method of appli- cation. The antioxidant effect of woodsmoke has been related to a surface effect. Draudt (1963) in a review of the smoking process in meats, reported that oxidation occurred most rapidly in the first half inch of the surface of unsmoked bacon. In smoked bacon, the surface oxidation rate was greatly reduced. Chen and Issenberg (1972) noted that 140 the preservative (antioxidant) effect of woodsmoke in foods was the result of partial surface dehydration and deposition of antioxidant compounds from the smoke onto the surface of the product. The antioxidant effect of smoke has been related to the presence of phenols in the smoke (Fretheim et al., 1980; Toth and Potthast, 1984). Bratzler et al. (1969) showed that the greatest quantity of phenols in smoked bacon occurred on the surface of the product. In the present study, the outer layer of the whitefish treated with woodsmoke received a heavy smoking. This coat- ing of the outer layer with smoke acted to prevent surface oxidation in these samples. In the samples treated with the liquid smoke, there was no surface protection, because the phenols were dispersed uniformly throughout the product. Therefore the samples treated with liquid smoke had no surface-barrier and oxidation could occur on the surface to a greater extent than in the woodsmoke-treated whitefish. The whitefish samples prepared by the various smoking procedures were analyzed for their total phenol content by the method of Bratzler et a1. (1969). The results of these analyses are presented in Table 29. The samples smoked with woodsmoke had a total phenol content of 79.2 mg/100 g of fish. The samples prepared in brines containing 0.7, 1.4 and 2.1% liquid smoke had total phenol contents of 46.1, 82.1 and 112.3 mg/100 9 sample, respectively. As shown in Table 29, the levels of total phenols in the woodsmoke and 141 Table 29. Total phenol content of whitefish smoked with different levels of liquid smoke or smoked with woodsmoke. Treatment Total Phenolsa’b (mg/100 9) Woodsmoke 79.2 0.7% Liquid smoke 46.1 1.4% Liquid smoke 82.1 2.1% Liquid smoke 112.3 aDetermined by the method of Bratzler et a1. (1969). bAnalyses were carried out in duplicate on the sample taken on day 0. 142 the sample prepared with brine containing 1.4% liquid smoke were essentially the same. However, their effect on the rate of lipid oxidation were not equal (Table 28). In an attempt to explain the significant difference- (P<0.01) between the antioxidant effect of the woodsmoke and the liquid smoke (Table 28), the respective samples were analyzed for their phenol profiles (Figures 4 and 5). Some of the major differences in the relative percent concentra- tions of the individual phenols that could be tentatively identified through the use of standards and their retention times are shown in Table 30. The woodsmoke samples contained greater amounts of phenols which were tentatively identified as m-phenylphenol (1.1%), eugenol (0.5), isoeugenol (52.4%), 2,2-bisquinoline (9.2%) and o—hydroxyphenol (3.7%) than the samples prepared with the liquid smoke. The liquid smoke samples appeared to contain more guaiacol, 2-methoxy-4 methylphenol (10.4%), phenol (26.2%), and 4-methylsyringol (1.7%). Knowles et a1. (1975) studied the phenol uptake by bacon prepared by traditional kiln smoking and bacon smoked using electrostatic application of liquid smoke. They reported that bacon smoked using the traditional kiln smoking contained greater levels of eugenol, isoeugenol and 4-methylsyringol. The samples prepared with the liquid smoke had greater levels of guaiacol, phenol, o-cresol, 2-methoxy-4-methylphenol and m/p-cresols. These results agree with the findings of the 143 Figure 4. Gas chromatographic analysis of whitefish smoked with woodsmoke. 144 b- 'G---I: u-ifi °o ___ an 4h-¢t =======; _ .i =5 7r Figure 5. Gas chromatographic anal sis of smoked whitefish prepared with .4% liquid smoke in the brine. 145 Table 30. Relative percent of phenols tentatively identified in whitefish prepared with 1.4% liquid smoke or smoked with woodsmoke. Phenol Woodsmoke 1-4% Liqud Smoke (74) (%) Guaiacol 10.0 44.6 2-Methoxy-4-methylphenol 4.8 10.4 m-Phenylphenol 1.1 ND Phenol 12.6 26.2 p-Cresol 0.7 0.7 m-Cresol . 1.5 1.4 2,4-Dimethylphenol 2.6 2.5 Eugenol 0.5 ND Isoeugenol (cis and trans) 52.4 12.2 4-methylsyringol 0.4 1.7 2,2-Biquiniline 9.2 ND O-Hydroxyphenol 3.7 ND N0 = not detected, limit of detection was 1000 mg/kg 146 present study. Daun (1979) noted that the antioxidant activity of smoke is due to the higher, rather than the lower, boiling point phenols. The lower boiling point phenols include phenol, cresols and guaiacol, while the higher boiling point phenols include syringol and its derivatives (Toth and Potthast, 1984). In the present study, fish samples prepared with the liquid smoke had greater levels of guaiacol, 2-methoxy-4- methylphenol (4-methylguaiacol), phenol and the cresols. All of these compounds are considered to be part of the lower boiling point phenols which have less antioxidant activity. Toth and Potthast (1984), in a review on the chemical aspects of meat smoking, reported that pyrocatechol, guaiacol, eugenol and isoeugenol have been found to have antioxidant activity; however, no information of their relative effectiveness as antioxidants was reported. In the present study, the whitefish smoked with woodsmoke had higher levels of eugenol and isoeugenol which could help explain the greater antioxidant activity of the woodsmoke. Sensory Analysis of Smoked Whitefish The results of the statistical analysis of the taste panel evaluations of the samples prepared with different levels of liquid smoke in their brines or smoked with wood— smoke was carried out as described previously. On day 0, 147 the samples prepared with woodsmoke, with and without nitrite, were found to be significantly (P<0.05) more desirable in odor and in flavor (Table 31). The non-nitrited woodsmoked sample was also found to be significantly (P<0.05) more desirable in texture and overall acceptability relative to the other samples. In contrast, both samples prepared with 2.1% liquid smoke in their brine were found to be signifi- cantly (P<0.05) less desirable in their odor, flavor and overall acceptability (Table 31). These data indicate that there is an upper limit on the amount of liquid smoke that can be used in this product with respect to consumer accept- ance. By day 5, both woodsmoke treated samples were still significantly (P<0.05) more desirable in odor, but only the non-nitrited woodsmoke treated sample was significantly (P<0.05) more desirable in flavor and overall acceptability (Table 32). The samples prepared with the 2.1% liquid smoke in the brine containing nitrite were found to be significantly (P<0.05) less desirable in flavor and overall acceptability. By day 14, the panelists were unable to distinguish any specific sample treatment as being consistently most or least desirable (Table 33). Significant differences in the samples tested appeared to be occurring randomly indicating that the panelists were unable to identify any sample as being consistently more or less desirable. 148 Amwmxpmcm memxcmc gmsmsxv mpnmgpmmv ummmbkk Amwmxpmcm unexcuc gmsmggv mpnmcwmmu “moss ucmgmwmmu Amo.ovav zpuceuwwwcmmm one memuump ucwgmmwwv saw: mmpasmmx.x mpnmgpmmu umemp u e ”mpnmgwmmv umos u p "upmumn upmm Agmpwsopmm oomv nmm.m umcwmucou mmcwsa Ppo co>mpu wgzuxw» mmmcvapmm Louo pcmsummcp nonexcmm .monmvooz cue: uwonm Lo mxosm cwacwp mo mpm>mp chXLm> new; nmummgp mm—asmm zmwmmuwcz mumc.gn to» o xmu um mmcoum acomcmm mmmcw>< .pm mpnm» 149 Amwmapmcm mcmxcmg cosmcxv mpnmgwmmv ummmbkk Amwmxpmcm mcmxcmg gmsmng mpnmevmmu umozk ucmgmmmwu Amo.ovav xFucmquwcmmm mum msmuump ucmgmmewu gum: meQEmma.x wpnmswmmu ummmp u e “mpnmcpmmu umoe n F ”upmumn upem Agmumsome oomv wow.~ umcwmucou mmcmen ~Po Lo>w—u 9:3.me mmmaupmm LOUD acmEummLH nonexcmm .mxosmvooz new: vmxosm to monm cvsovp mo mpm>mp mcpxcm> saw: umuemgu mwpqum smwwmuvgz mumcwga to» m xmu um mmgoum agomcmm mmmgm>< .Nm «ppm» 150 mppmcwmmv amemp e umpnagmmmu amos open AeoooEopom oomv Row.~ ooeveoeoo Amvmapmce mcwxcmg ngmva mpnmsvmmv ammo; «a. Amwmxpmcm mcvxcmc gmseng mpaecvmmv umozk u no no p P ma moePse P~o co>mpu mgauxm» mmmcwupmm Loco ucmEummL» omcwxcmm .mxosmvooz nae: nmonm to exosm upchp mo m~o>mp ocvxgm> .I sue: emummcu mm—aEmm zmvmmuwsz mvmcvga to» «F xmu um mmcoum xgomcmm mmmcm>< .mm epoch 151 ANOVA analysis, in combination with Bonferroni's contrasts, of the taste panel data revealed essentially the same trends as seen in the ranking analysis. On day 0, the woodsmoke treated samples were significantly (P<0.05) more desirable in odor, while the non-nitrited woodsmoke treated sample had the most desirable flavor (P<0.05) (Table 31). Also on day 0, the samples treated with the 2.1% liquid smoke were significantly (P<0.05) less desirable in flavor. On day 5, the woodsmoke treated samples were still more desirable (P<0.05) in odor and the non-nitrited woodsmoke treated sample was more desirable in flavor (P<0.05) (Table 32). By day 15, there was no significant distinction made between the samples (Table 33). The results of this study indicate that woodsmoke treated whitefish undergo less oxidative rancidity than samples prepared with liquid smoke. The addition of nitrite significantly (P<0.01) reduced the level of oxidation occur- ring in the smoked whitefish. The addition of nitrite did not result in the formation of N-nitrosamines in the finished smoked whitefish product. Taste panel evaluation of the samples prepared with increasing levels of liquid smoke in the brine or smoked with woodsmoke indicate that there was an upper limit to the addition of liquid smoke in this product. The addition of 2.1% liquid smoke was found to be undesirable on the first day of the study. However, over time the panelists were unable to distinguish the samples 152 containing the 2.1% liquid smoke from the other treatments. Evaluation of Lipid Oxidation in Baked Whitefish The development of WOF in fish has not received much attention because fish, generally, is not cooked, refrigera- ted or frozen and then reheated prior to serving. However, smoked fish can be held at refrigerator temperatures for up to 14 days before consumption. That smoked fish undergoes oxidative rancidity has been shown in the present study. However, the relationship between the sensory evaluation of rancidity in the cooked, refrigerated whitefish and a chemical parameter of rancidity, such as TBA number and/or hexanal levels has not been established. Therefore this study was designed to establish the level of TBA number in baked, refrigerated whitefish that is detectable by taste panelists and to investigate the changes in hexanal levels in the refrigerated baked whitefish. Whitefish that was baked then held at refrigerated temperatures for 0, 1, 2 and 3 days had TBA values of 1.51, 4.77, 5.41 and 7.53, respectively (Table 34). The samples from day 0, l and 2 were analyzed for relative changes in their hexanal content. These values are shown in Table 34. Statistical analysis of the taste panel evaluation of the baked whitefish held at refrigerated temperatures for 0, 1, 2 and 3 days, using the ranking procedure of Kramer (1963) showed that the day 0 sample was the most desirable in flavor 153 Table 34. The TBA number, hexanal changes and the sensory scores of baked whitefish refrigerated for 0, l, 2 and 3 days. Days Refrigerated TBA Numbera Hexanalb Sensory ScoreC (mg/kg) (area) 0 1.51 1854x105 0.54* 1 4.77 2075x105 2.14** 2 5.41 1423x105 1.00 *‘k 3 7.53 -- 2,14 aAnalyses were carried out in triplicate bHexanal area corresponds to computer printout of GLC analysis CScale: l = most desirable; 4 = least desirable *Most desirable (Kramer ranking analysis) **Least desirable (Kramer ranking analysis) 154 (P<0.05) (Table 34). The sample that was 1 day old and the sample that was 3 days old were found to be the less desirable (P<0.05). The sample that was 2 days old was desirable but to a lesser degree than the day 0 sample. When the samples that were 0, l and 2 days old were used in a triangle test, as a means of establishing a threshold level for the TBA value, it was found that the panelists were able to determine a significant (P<0.05) difference between the day 0 sample and the day 1 sample. However, they were unable to distinguish between the day 0 sample and the day 2 sample. The data indicates that the panelists were able to detect TBA values of 4.77 and 7.53. However, the panelists were unable to distintuish a TBA value of 5.41. The sensory evaluation of off-flavor (rancidity) thresh- olds have been reported in the literature. Yu et a1. (1969) reported that fish having a TBA value of 3.1 or greater were considered very rancid and unacceptable by panelists. How- ever, TBA values of 2.4 in the same fish were judged as being acceptable. The correlation between the TBA value and taste panel evaluation of rancidity in cooked meats has been shown. Igene (1979) reported correlation coefficients between TBA numbers and taste panel scores for beef, the dark meat of chicken and the white meat of chicken were -O.74, -0.91 and -0.87, respectively. Bussey (1981) reported that in turkey hams, the TBA number and the taste panel score for off-flavor 155 and acceptability had correlation coefficients of +0.81 and -0.87, respectively. Bussey (1981) also reported that in turkey bologna the correlation coefficient between TBA number and off-flavor and acceptability were +0.91 and -0.87, respectively. Zipser et al. (1964) reported a correlation coefficient of 0.92 between TBA number and rancid odor in cured and uncured pork. The volatile extract from the baked whitefish (day 0) was subjected to gas chromatographic-mass spectrometric (GC-MS) analysis in order to identify the volatiles occur- ring in the baked whitefish (Figure 6). The volatiles that ' were identified in the baked whitefish volatiles were: decane, 1-penten-3-ol, hexanal, undecane, heptanal, 1- pentanol, 1,3,5-trimethylbenzene and benzaldehyde. These compounds were found by Josephson et a1. (1984) in cooked whitefish; however, they identified 63 different volatiles in their sample(s). They reported that hexanal accounted for 40% of the volatiles in cooked whitefish. In the present study it was found that hexanal accounted for greater than 50% of the volatiles in baked whitefish; however, the sample analyzed by GC-MS only contained 14 compounds. Therefore the hexanal would be more predominant. The concentration of hexanal iri the samples of baked whitefish held for 0, l and 2 days (Table 34) increased from the initial level on day 0, then decreased on day 2. It appears that changes in the hexanal levels in the 156 ‘--3: [ ‘ 0 :{J— m <—< Figure 6. Gas chromatographic analysis of the volatiles from baked whitefish. 157 refrigerated baked whitefish should be investigated for their role in the evaluation of rancidity in this system. The presence of hexanal as an indicator of oxidation in food systems has been investigated by a number of workers. MacLeod and Coppock (1976) reported that hexanal was produced during the cooking of beef, and that hexanal has been described to have a strong, unpleasant rancid odor. Gaddis and Ellis (1957) reported that hexanal has been identified as the major volatile in rancid turkey and pork fat. Cross and Ziegler (1965) noted that in uncured ham there were higher levels of hexanal and valeraldehyde than in cured ham. These hams had been cooked to an internal temperature of 70°C, then held at 3°C for several days before analysis. Based on the work by Fooladi et al. (1979), uncured cooked pork was shown to have a TBA number of 7.85 within 48 hours after cooking. Therefore it can be assumed that the uncured ham analyzed by Cross and Ziegler (1965) was rancid and the hexanal detected in the meat's volatiles was the result of this rancidity. More research is needed to clarify the relationship between taste panel evaluation of rancidity with TBA and hexanal values in baked whitefish during refrigerated storage. SUMMARY AND CONCLUSIONS The salt uptake of whitefish during brining is affected by the salt concentration of the brine, the length of the brining period and whether the fish is brined whole or as a fillet. Fillets, due to their increased surface area, take up greater amounts of salt from the brine than the whole, eviscerated fish. Higher brine salt concentrations produce fish containing higher levels of salt. The level of salt in smoked whitefish affected the rate of Clostridium botulinum type E spore outgrowth and toxin production. Fish containing 2 percent salt (wp) became toxic within 21 days at 27°C. Increasing the salt concentration to 4 percent (wp) inhibited toxin production for 42 days at 27°C. The addition of nitrite to the fish containing 4 per- cent salt (wp) prevented toxin production until day 63. Smoked whitefish containing 6 percent salt (wp) were not toxic even after 83 days at 27°C. In another study concerned with the lipid stability and organoleptic acceptability of smoked whitefish containing various levels of salt (wp), it was found that unsalted smoked whitefish had a significantly (P<0.01) lower rate of oxidation (TBA number) during refrigerated storage. However, the unsalted product was found to be organoleptically 158 159 unacceptable initially (day 0) and over time (day 14). Taste panel evaluation showed that the smoked whitefish containing 4 percent salt (wp), with and without nitrite, was signifi- cantly (P<0.05) more desirable initially and over time. The addition of nitrite significantly (P<0.01) reduced the rate of oxidation in all salt treatment levels. The role of smoke treatments on the lipid stability and organoleptic acceptability of smoked whitefish was investi- gated. The results showed that woodsmoke had a greater antioxidant effect than did the liquid smoke. Nitrite strongly inhibited the deve10pment of warmed-over flavor in this product. Taste panel evaluation indicated that initially (day 0) there was a significant (P<0.05) preference for the samples treated with woodsmoke but over time (day 14) the panelists were unable to distinguish between the woodsmoke and the liquid smoke treatments. A Smoked whitefish containing nitrite was shown to contain no detectable levels of N-nitrosamine or NTCA. The evaluation of oxidative rancidity using taste panel evaluation in conjunction with TBA and hexanal analyses in whitefish baked then held at refrigerator temperatures for 3 days was studied. Sensory evaluation data indicated that the panelists could identify rancidity in fish with TBA numbers of 4.77 and 7.53. However, the panelists were unable to detect rancidity in fish having a TBA number of 5.41. Hexanal analysis indicated that in the samples with the 4.77 160 and 7.53 TBA number contained the highest level of hexanal. The sample with the TBA number of 5.41 had the lower level of hexanal. These data indicate that both chemical parameters, i.e, TBA and hexanal, appear to affect the panelists' ability to detect rancidity in whitefish that had been baked then held at refrigerated temperatures. It can be concluded from these studies that it is feasible to produce smoked whitefish that contain a lower level of salt (4% water-phase) and still maintain safety against botulism. The addition of nitrite to smoked whitefish containing less salt increased the safety of the product against botulism. Taste panel data indicate that the addition of nitrite did not appear to affect the desirability of the smoked whitefish. Nitrite in the smoked whitefish did not result in the formation of detectable levels of N-nitrosamines. Therefore, the addition of nitrite in combination with a 4 percent salt (water-phase) concentration can be recommended for use in the production of smoked whitefish. PROPOSALS FOR FUTURE RESEARCH These studies on the safety and acceptability of smoked whitefish and the brining of whitefish have raised some questions which merit further study. These include: (1) Determining the role of each lipid fraction (trigly- ceride and phospholipid) in the development of warmed-over flavor in smoked whitefish; (2) Further clarification of the factors affecting salt uptake in whitefish tissue so that a more uniform product can be produced. This work should include an investigation of the cost effectiveness of using fillets as opposed to the whole, eviscerated fish in the production of smoked white- fish; (3) To evaluate the effect of nitrite on the flavor of smoked whitefish; (4) To evaluate the influence of higher levels of sorbate in combination with nitrite on the outgrowth of Clostridium botulinum type E spores. 161 APPENDIX 162 mpnmgwmmcc: xgm> mpnmgwmmuca xpmueemuo: mpnmcwmmvca xeeemepm Fmgpamz opnmgwmmo spoemepm mpnmcwmmo xpmuecmuoz mpnmcvmmo xgm> xuwpwnwuqmuu< Fpmcm>o xupmm cove: xcm> mupem Lows: xpmumgmvoz xupmm Love: »_oem.Fm pmguamz xepem spdemvpm xupmm apmuecmvoz Appmm xgm> mmmcwupom mama zmvemuvzz umonm mo cowumapm>m xcomcmm F xpucmaq< Ago xgm> xgo xpmumgmuoz xco spoemwpm Fmguzmz xovzn speeoepm hows» zpmueemwoz movaq xgm> _ooceo=oz nus: agm> mxwpmwo spoemVFm oxepneo apoea__m oxpmeQ oxe_m_o to expo toee.oz »_oeme_m o¥_o apmumgmuoz ox“; eoaz xeo> ox_o .xoveeeeodooo< Lo>mpm * mpasom msmz 163 Appendix 2 Taste panel form for the evaluation of the desirability of smoked whitefish using ranking analysis SMOKED FISH EVALUATION In this evaluation of smoked whitefish, you are asked to taste all of the samples, evaluate them and then decide how they relate to each of the traits listed. Using their assigned numbers, rank the samples from the most desirable to the least desirable in the spaces provided. There is no right or wrong answer, only your opinion. Trait Most Desirability Least l Odor i l Saltiness Texture Flavor Overall Acceptability 164 Appendix 3 Cooked fish evaluation COOKED FISH EVALUATION The following taste panels areodesigned to evaluate the desirability of cooked fish. In order to complete this evaluation, you are asked to taste three (3) sets of samples. The first set is to be ranked as being most to least desirable. The next two sets will be used in triangle tests. Ranking: Please rank the samples as to their desirability. Most desirable Least desirable Triangle Tests: In these sets of samples, two of the samples are identical and one is different. Please taste and then decide which is the different sample; also indicate if the different sample is more or less desirable. Test #1: Please list the number of the odd sample Is this sample more or less desirable Test #2: Please list the number of the odd sample Is this sample more or less desirable Name Date 165 .uoNN e. notoome .ounu._n.cu :. uao go’ssou ago: momapuc