Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Surimi-based product development and viscous properties of surimi paste Bouraoui, Moez Mohamed 1996

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

Item Metadata

Download

Media
831-ubc_1996-090442.pdf [ 4.28MB ]
Metadata
JSON: 831-1.0058575.json
JSON-LD: 831-1.0058575-ld.json
RDF/XML (Pretty): 831-1.0058575-rdf.xml
RDF/JSON: 831-1.0058575-rdf.json
Turtle: 831-1.0058575-turtle.txt
N-Triples: 831-1.0058575-rdf-ntriples.txt
Original Record: 831-1.0058575-source.json
Full Text
831-1.0058575-fulltext.txt
Citation
831-1.0058575.ris

Full Text

SURIMI-BASED PRODUCT DEVELOPMENT AND VISCOUS PROPERTIES OF SURIMI PASTE. by MOEZ M. BOURAOUI Engineering Degree, School of Agricultural Engineering, ESIER, Tunisia, 1988 M. A. Sc., University of British Columbia, Canada, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (Department of Chemical Engineering) We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA December 1995 © Moez M. Bouraoui, 1995 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of CM&MUU £NfyMg£#WQs The University of British Columbia Vancouver, Canada DE-6 (2/88) Abstract Raman spectroscopy was used to study the protein structure in raw and salted surimi from Pacific whiting, and in gels formed by setting (32°C), cooking (86°C) or setting followed by cooking. The intensity of the peaks assigned to disulfide bond stretching vibrations increased considerably in the cooked and set-cooked gels. A smaller increase was found in the gels that were subject to setting alone. Secondary structure estimation based on the amide I band indicated a change from predominantly a-helical structure in raw surimi to similar proportions of a-helical and anti-parallel p-sheet after setting. A further increase in anti-parallel p-sheet and decrease in a-helix content occurred during the kamaboko stage. The intensity of C-H stretching vibrations of the aliphatic residues decreased after salting, setting and cooking. The set and cooked gel had a better gel strength and fold score than the cooked gel which, in turn, had better properties than the set gel. Response surface methodology was used to determine the optimal setting and cooking time and temperature conditions resulting in a maximized gel strength, fold score and color, (whiteness index), and a minimized gel expressible liquid. Cooking temperature was the variable that had the strongest influence on the gel quality characteristics. Level set programming, ii a global optimization method, gave essentially the same results as a gradient based optimization method. The results obtained by these two methods were better than those generated by the Simplex technique. Using a formulation composed of pink salmon surimi, salt, whey protein concentrate and wheat starch, an optimal final product, kamaboko sausage, was developed. In addition, when this formulation was applied to herring surimi, the kamaboko obtained had similar gel strength and elasticity as a commercial surimi-based product. The effects of frozen storage conditions of roe herring on its gel making ability (GMA) were investigated. Frozen storage at around -45°C maintained the GMA for seventy days while storage at -83°C further increased the GMA period. The viscous properties of a salmon surimi paste were studied using a rotational viscometer. The paste behaved as a shear thinning fluid with a yield stress that increased with temperature up to 21 °C. Viscosity also increased with temperature up to 21 °C, possibly because of protein-protein interactions. The Theological data were reasonably well represented by a simple model which takes into consideration the effects of both shear rate and temperature. iii Table of Contents Abstract ii Table of contents iv List of Tables iv List of Figures xi Acknowledgments xiv Chapter 1: Introduction 1 1.1 Surimi manufacturing 3 1.2 Production of surimi-based products 6 1.3 Freeze denaturation of fish proteins 7 1.3.1 Moisture phase changes 9 1.3.2 Fish lipids 10 1.3.3 Enzymatic denaturation by TMAOase 10 1.4 Thesis outline 11 iv Chapter 2: Investigation of the Protein Structure of Pacific Whiting Surimi and Gels Using Raman Spectroscopy 13 2.1 Introduction 13 2.2 Materials and methods 18 2.2.1 Materials 18 2.2.2 Raman spectroscopic analysis 19 2.2.3 Measurement of surimi gel properties 22 2.3 Results and discussion 25 2.3.1 Disulfide (S-S) bond stretching vibrations in the 450-600 cm'1 region 25 2.3.2 Backbone and side chain vibrations in the 610-2000 cm"1 region 27 2.3.2.1 The tyrosine doublet near 830 and 850 cm"1 31 2.3.2.2 Secondary structure estimation from the amide I and III bands 35 2.3.3 C-H Stretching vibrations in the 2500-3400 cm"1 region 41 2.4 Conclusions 45 Chapter 3: Optimization of the Setting and Cooking Conditions for Kamaboko Making from Pacific Whiting Surimi 46 3.1 Introduction 46 3.1.1 The Nelder-Mead simplex optimization method 47 v 3.1.2 The Gradient based optimization methods 50 3.1.3 The Level Set Programming (LSP) method 50 3.2 Materials and methods 51 3.2.1 Gel strength (compression test) 53 3.2.2 Fold score 53 3.2.3 Gel color test 54 3.2.4 Expressible liquid measurement 54 3.3 Results and discussion 55 3.3.1 Curve fitting results 58 3.3.2 Optimization results 65 3.3.2.1 Results of Nelder-Mead Simplex optimization 65 3.3.2.2 Results of NLPQLO Gradient optimization 67 3.3.2.3 Results of LSP optimization 67 3.3.2.4 Multiobjective optimization 70 3.4 Conclusions and recommendations 73 Chapter 4: Kamaboko Production from Roe Herring and Pink Salmon 75 4.1 Introduction 75 4.2 Objectives 77 4.3 Materials and methods 78 4.3.1 Materials 78 vi 4.3.2 General procedures 79 4.3.2.1 Surimi making 79 4.3.2.2 Kamaboko making 80 4.3.2.3 Measurement of surimi gel properties 81 4.3.3 Roe herring samples 81 4.4 Results and discussion 82 4.4.1 Frozen storage experiments 82 4.4.2 Optimization experiments 85 4.5 Conclusions 92 Chapter 5: Viscous Properties of Salmon Surimi Paste 94 5.1 Introduction 94 5.2 Materials and methods 96 5.3 Results and discussion 98 5.4 Conclusions 126 Chapter 6: Summary of Conclusions and Proposed Future Work 128 6.1 Conclusions 128 6.2 Proposed future work 130 Nomenclature 132 vii Bibliography 136 viii List of Tables Tab le ! Surimi production in four countries, 1985-1991 2 Table 2. Rating system of the fold test 24 Table 3. Tentative assignment of some major bands in the Raman spectrum of the raw sample (610-2000 cm"1) 30 Table 4. Estimated proportions of the buried and the exposed tyrosine residues 34 Table 5. Secondary structure fractions estimated from the amide I band by RSAP for the five types of samples 38 Table 6. Experimental design for the optimization of surimi heating conditions 52 Table 7. Experimental results 56 Table 8. Regression results where "log (Gs)" is the response variable 60 Table 9. Experimental and estimated values of the 4 response variables 61 Table 10. Regression results where "F s" is the response variable 62 Table 11. Regression results where "C" is the response variable 63 Table 12. Regression results where "E" is the response variable 64 Table 13. Optimization of individual response variables using the Nelder-Mead Simplex method 66 ix Table 14. Optimization of individual response variables using the NLPQLO Gradient method 68 Table 15. Optimization of individual response variables using the LSP method 69 Table 16. Multiobjective optimization results 72 Table 17. Effects of different formulations on surimi gel texture 86 Table 18. Moisture content of different products 91 Table 19. Curve fitting results of the second ISR rheograms 110 Table 20. Curve fitting results of the DSR rheograms 112 Table 21. Results of the curve fitting of the yield stress 119 Table 22. Yield stress values from the first ISR rheograms 120 Table 23. Curve fitting results of Theological data by the ST model 122 x List of Figures Figure 1. Flow diagram of commercial surimi manufacturing process 5 Figure 2. Flow diagram of a crab analog production process 8 Figure 3. A schematic layout of a Raman spectrophotometer 20 Figure 4. Example of a force profile recorded by the Instron instrument 23 Figure 5. Raman spectra of surimi and gels in the 450-600 cm "1 region showing the S-S stretching band near 530 cm 1 26 Figure 6. Variation with treatment of the normalized intensity of the 530 cm"1 band 28 Figure 7. Raman spectra of raw Pacific whiting surimi in the 610-2000 cm' 1 region 29 Figure 8. Raman spectra of surimi and gels in the 820-912 cm"1 region, including the 830/850 tyrosine doublet, a) Original spectra, b) Deconvoluted spectra 32 Figure 9. Raman spectra of surimi and gels in the 1590-1720 cm'1 region which includes the amide I band, a) Original spectra, b) Deconvoluted spectra 37 Figure 10. Raman spectra of surimi and gels in the 1220-1275 cm"1 region which includes the amide III band, a) Original spectra, b) Deconvoluted spectra 40 xi Figure 11. Raman spectra of surimi and gels in the 2500-3400 cm"1 region showing the C-H stretching band 42 Figure 12. Variation with treatment of the intensity and of the area of the -2935 cm"1 peak 44 Figure 13. Variation of roe herring kamaboko gel strength with frozen storage 84 Figure 14. Combined effects of WPC and WS on salmon surimi gel strength, 87 Figure 15. Combined effects of WPC and WS on salmon surimi gel fold score 89 Figure 16. Rheograms of salmon surimi paste at 1 °C 99 Figure 17. Rheograms of salmon surimi paste at 6°C 100 Figure 18. Rheograms of salmon surimi paste at 11 °C 101 Figure 19. Rheograms of salmon surimi paste at 21 °C 102 Figure 20. Rheograms of salmon surimi paste at 26°C 103 Figure 21. DSR rheograms 106 Figure 22. Second ISR rheograms 107 Figure 23. DSR rheograms at 1°C 113 Figure 24. DSR rheograms at 6°C 114 Figure 25. DSR rheograms at 11 °C 115 Figure 26. DSR rheograms at 21 °C 116 Figure 27. DSR rheograms at 26°C 117 xii Figure 28. DSR rheograms and the predictions of the ST model 124 Figure 29. Second ISR rheograms and the predictions of the ST 125 model xiii Acknowledgments I would like to express my most sincere gratitude to my supervisors, Dr. Bruce Bowen and Dr. Shuryo Nakai, for their invaluable supervision, help and understanding. I greatly appreciate their contributions and their ideas. My special thanks go to Dr. Tim Durance and Dr. Ken Pinder, my committee members, for their contributions to this project and their guidance. I also thank Dr. Eunice Li-Chan for her great help with the Raman spectroscopy, Dr. Jaouad Fichtali for his contributions to this project and Mr. Dragan Macura for his help with the IRAP project. I extend my thanks to everyone affiliated with the Department of Chemical Engineering and the Department of Food Science, and to all my friends in Canada and Tunisia. I am unable to sufficiently thank my parents, Mansour and Radhia, my dear wife, Amna, my sister, Hella, and all my family members for their great love, their sacrifices and their encouragements. xiv Chapter 1. Introduction Chapter 1 Introduction Surimi is a Japanese term for the wet protein concentrate of mechanically deboned and water-washed fish flesh (Okada, 1992; Lee, 1984). It is often mixed with cryoprotectants to maintain the functional properties of its myofibrillar proteins. The high concentration of these proteins produces, upon cooking, an elastic and chewy texture that can be made to resemble that of shellfish (Lee, 1984). For centuries, the Japanese have been producing surimi mainly from Alaska pollock. Around the twelfth century, Japanese fishermen discovered that when surimi was salted, ground and steam or broil cooked, the end product, called kamaboko, could be kept longer (Lee, 1984). Surimi technology has evolved to produce such shellfish analogs as crab and shrimp. Recently, the USA has become interested in this technology especially since it is the major producer of Alaska Pollock. Table 1 shows the trends in surimi production in several countries (Sproul and Queirolo, 1994). The table demonstrates that surimi production in the USA jumped from none in 1985 to 160,000 metric tons in 1991. Other countries such as Canada, Norway and the UK have also started to produce surimi from under-utilized 1 Chapter 1. Introduction Table 1. Surimi production in four countries, 1985-1991 Year Production (1,0001) Japan USA Korea Thailand 1985 490 0 20 10 1987 480 20 30 20 1989 390 80 30 20 1991 190 160 10 20 2 Chapter 1. Introduction fish species (Hastings and Currall, 1989). In Japan, apart from pollock, other fish species are being used for surimi production. These species include sardine mackerel, atka mackerel and horse mackerel. Surimi of good quality can also be made from New Zealand hoki (Macruronus novaezelandiae). In 1988, New Zealand produced 22,800 metric tons of surimi from this species (Holmes et al., 1992). Pacific whiting is another under-utilized species that has recently been used for surimi production. It is very abundant on the West Coast of North America from British Columbia to California (Ohshima et al., 1993). The allowable catch of this species in the USA ranges between 140,000 and 250,000 metric tons per year (Yongsawatdigul et al., 1995). These trends should dramatically increase the amount of fish products available to humans; about half of the harvested fish in the world in 1986 was reduced to animal feed (Pigott, 1986). Because of the high nutritional and functional properties of surimi, extensive research is needed to further develop and improve the production of shellfish analogs as well as other non-seafood imitations from this important protein source. 1.1 Surimi manufacturing Because Alaska Pollock is highly susceptible to freeze denaturation (Matsumoto and Noguchi, 1992), surimi used to be produced and consumed on a same-day basis. In 1959, Nishiya's research team discovered that by 3 Chapter 1. Introduction washing out the water soluble components from the minced fish and adding cryoprotectants such as sugar compounds and polyphosphates, the functional properties of surimi could be maintained longer during frozen storage (Okada, 1992) . This was a threshold for the production of frozen surimi. The main cryoprotectants used are sucrose and sorbitol. Their presence, at low concentrations (i.e., 4% by weight) increases the amount of bound water which decreases the displacement of water from protein surfaces and hence reduces protein freeze denaturation (Arakawa and Timascheff, 1982; Matsumoto and Noguchi, 1992). Figure 1 summarizes the overall process of commercial surimi production. The process starts with deboning and gutting to remove the head, viscera and most of the backbone. The product is then minced and washed with water. Washing is very important because it removes most of the sarcoplasmic proteins, inorganic salts, and other undesirable water soluble compounds such as blood. It is usually carried out in three to four cycles (Ohshima et al., 1993) . Refining is then performed to remove any connective tissues, skin scales, small bones, etc. The soft meat is selectively forced through the perforations of a cylindrical screen [perforation size varies between 1 and 3 mm (Toyoda et al., 1992)]. The excess water due to the washing process is then eliminated by a screw press. This dewatering reduces the moisture content of the refined mince from 90% to around 80% (wet basis). The 4 Chapter 1. Introduction raw fish (100%) deboning and gutting \1/ mincing washing with water refining dewa ering blending of additives (i.e., cryoprotectants) filling and packaging freezing frozen surimi (20 to 30 %) Figure 1. Flow diagram of commercial surimi manufacturing process 5 Chapter 1. Introduction product is then mixed with cryoprotectants in a silent cutter. Then, the surimi is extruded as blocks (i.e., 10 kg) and packed in polyethylene bags. Surimi blocks are then frozen at -25°C or lower (Toyoda et al., 1992). The surimi yield is between 20 and 30% (raw material weight basis); it varies with the fish species, size and season as well as the manufacturing equipment used (Toyoda et al., 1992). Surimi is popular because of its high nutritive value, high protein content and low fat. For example, surimi produced by Alaska Pacific Seafoods (Kodiac, Alaska) from Alaska pollock contained 75.3% water, 15.4% protein, 4% sorbitol, 4% sucrose, 0.3% sodium tripolyphosphate, and 1.0% fat and ash (Nicklason and Pigott, 1989). 1.2 P r o d u c t i o n o f s u r i m i - b a s e d p r o d u c t s Most of surimi-based products are gelled surimi. The Japanese have been producing surimi-based products for centuries. The washed fish flesh, when combined with certain ingredients, mixed and kneaded, and steamed, forms a fish gel called kamaboko (Pigott, 1986). To name but a few, there is steamed kamaboko, broiled kamaboko and fried kamaboko (Pigott, 1986). More recently, shellfish analogs have been produced from surimi. These include imitations of crab, lobster, scallop and shrimp (Lee, 1984). Crab imitation has gained popularity not only in Japan and Korea but also in the 6 Chapter 1. Introduction USA. For instance, because of overfishing, the US production of King Crab plunged from 185 million pounds in 1980 to less than 40 million pounds in 1982 (Gwinn, 1992). It has continued to fall further in succeeding years. This shortage of crab meat, along with its high price, has created a significant market for the production of crab analogs. A crab analog manufacturing process is summarized in Figure 2 (Lee, 1984; Wu, 1992). Surimi is also used in such products as soups, meat pies and fish cakes. Lee et al. (1992b) stated that the texture of surimi-based products composed only of surimi and salt is inclined to be rubbery. To improve the texture quality and water binding capacity as well as the product stability for frozen distribution, other ingredients are added. These include starch (i.e., wheat or potato starch), egg white, etc., which become entrapped within the gel matrix thereby improving its strength and texture (Lee et al., 1992b). Modified starch was reported to improve the gel freeze-thaw stability (Lee et al., 1992b). Other ingredients may also be added depending on the final product desired. These include oil, flavorings and colorants, to name a few. 1.3 Freeze denaturation of fish proteins Frozen storage of fish products is widely used by the industry to prolong the storage life of seafood products. However, extended storage causes the 7 Chapter 1. Introduction tempering to -8 to -4°C comminution with other ingredients (i.e., salt, water, starch, flavor) pumping and sheet forming (extrusion) setting and cooking slitting bundling coloring wrapping 4 cut mg packaging I pasteurization si, coo mg freezing Figure 2. Flow diagram of a crab analog production process 8 Chapter 1. Introduction deterioration of the texture, the flavor and the color of the fish. Several mechanisms have been postulated to explain the causes of the loss of fish quality. Some of these mechanisms are discussed below. 1.3.1 Moisture phase changes The formation of ice crystals could cause damage to fish proteins. These ice crystals are larger when the product is frozen at slow freezing rates or when the storage temperature fluctuates (Wheaton and Lawson, 1985). Ice crystals can break cells, rupture membranes and cause drip loss upon thawing of the fish product (Shenouda, 1980). It is, therefore, recommended to freeze the product at fast rates allowing the passage of every part of the product through the critical freezing zone (-0.8 to -5°C) within five to ten hours (Wheaton and Lawson, 1985). Matsumoto (1980) and Matsumoto and Noguchi (1992) explained that the migration of water molecules to form ice crystals causes the dehydration of the protein molecules. This phenomenon triggers the denaturation of the fish myofibrillar proteins through aggregation and unfolding. A multitude of cryoprotectants such as sucrose, sorbitol, phosphates, can be used to reduce freeze denaturation during frozen storage (Matsumoto and Noguchi, 1992). 9 Chapter 1. Introduction As water in the fish freezes out, the salt concentration increases. This increase can cause protein aggregation. Connel (1964) reported that a salt concentration of 10% caused the most protein damage. This damage is reported to mainly affect the myosin proteins (Shenouda, 1980). 1.3.2 Fish lipids Dehydration (caused by freezing) as well as exposure to oxygen (i.e., due to poor packaging) can increase lipid oxidation during frozen storage of fish. Shenouda (1980) reported that the oxidized products of lipids interact with the proteins' functional groups decreasing the solubility of the proteins. It also causes rancidity and an undesirable fishy taste. Lipid oxidation increases with storage time and with the fat content of the fish. 1.3.3 Enzymatic denaturation by TMAOase During the frozen storage of some fish species, especially the gadoid family (cod, pollock, whiting, hake, etc.), an enzymatic reaction causes the degradation of trimethylamine oxide (TMAO) into dimethylamine (DMA) and formaldehyde (FrHO) (Shenouda, 1980). This reaction, caused by the TMAOase enzyme, is slowed at very low storage temperatures (Wheaton and Lawson, 1985). FrHO causes cross-linking of proteins and decreases their 10 Chapter 1. Introduction extractabilty. It mainly affects the tropomyosin and the heavy chains of myosin (Childs, 1973). There are still many uncertainties regarding the mechanisms responsible for fish protein denaturation during frozen storage, and especially concerning the interactions between the different mechanisms, their relative importance and the stages at which they occur. 1.4 Thesis outline This thesis was motivated by the need to provide alternative value-added fish products from under-utilized fish species. It consists of a comprehensive study of the various processing factors that play a role in the overall surimi gel production cycle. In Chapter 2, the mechanisms of Pacific whiting surimi gelation were studied using the Raman spectroscopy technique. This method makes it possible to study the proteins in intact pastes and gels without having to isolate or extract the myofibrillar proteins. The effects of setting and cooking conditions on the quality of the final product (Pacific whiting surimi gel) are studied in Chapter 3. The optimal 11 Chapter 1. Introduction setting and cooking temperatures and times were determined by finding the conditions that maximize three and minimize one gel texture quality variable. Surimi production from two under-utilized species, pink salmon and roe herring, is investigated in Chapter 4. The effects of frozen storage conditions on the gel making ability as well as the effects of gel improving ingredients on the gel texture are discussed. An optimal formulation for the production of kamaboko sausage was developed. The viscous properties of a surimi paste consisting of the optimal formulation obtained in Chapter 4 are studied in Chapter 5. 12 Chapter 2. Raman spectroscopy of surimi Chapter 2 Investigation of the Protein Structure of Pacific Whiting Surimi and Gels Using Raman Spectroscopy 2.1 Introduction Gelation is one of the most important functional properties of surimi. Gelation is accomplished by first grinding surimi with salt to increase the solubility or the extractability of the myofibrillar proteins. The resulting paste is then slowly set at temperatures below 50°C before being cooked at higher temperatures (usually above 70°C). The formed gel has an elastic and chewy texture that can be made to resemble the texture of shellfish flesh (Lee, 1984). A thorough understanding of the mechanisms that govern surimi gelation would be useful for the improvement of gel quality and texture. Niwa (1992) indicated that grinding surimi with salt causes the formation of actomyosin from myofibrils. Myosin has a molecular weight of roughly 480,00 to 500,000 and contains over 40 sulfhydryl residues but no disulfide bonds (Nakai and Li-Chan, 1988). Actin has a molecular weight of about 42,000 to 47,000 and contains between 8 and 13 sulfhydryl residues but no disulfide bonds (Nakai and Li-Chan, 1988). During the setting stage {suwari in Japanese), Gill and 13 Chapter 2. Raman spectroscopy of surimi Conway (1989) found that heavy meromyosin (HMM S-2) and light meromyosin (LMM) fragments of the cod myosin tail region were inaccessible to chymotryptic digestion. They assumed that tail-tail interactions took place. When isolated, only the LMM fragments formed a gel (Sano et al., 1990). Hydrophobic interactions have been proposed to be the most dominant mechanism for the formation of a three-dimensional gel network (Chan et al., 1992; Niwa, 1992; Stone and Stanley, 1992). Many studies have demonstrated the importance of hydrophobic interactions in isolated myosin or actin fractions, actomyosin, meat sols or salt extracts (Nakai and Li-Chan, 1988). In the presence of a hydrophobic probe, sodium anilino-naphthalene 8-sulfonate (ANS), the fluorometric intensity of actomyosin solutions from setting flatfish increased with heating (Niwa, 1992). During cooking (kamaboko making or gel strengthening), as more protein unfolding takes place, interactions between exposed hydrophobic sites become stronger. In addition, Itoh et al. (1979) postulated an increase in disulfide bonds, based on a decrease in the SH content of carp actomyosin solution with increasing temperatures up to 80°C. Niwa (1992) also proposed that some disulfide bonds may be formed due to oxidation of sulfhydryl residues. Other researchers, including Tsukamasa et al. (1993), have suggested that enzyme (transglutaminase) catalyzed cross-linking may occur during setting. 14 Chapter 2. Raman spectroscopy of surimi In most of the reported literature, the mechanisms of protein sol-to-gel transition have been postulated by studying the protein structure after dissolving the gel. A great deal of uncertainty, therefore, still exists regarding the exact chemistry of surimi gelation. Research is needed to elucidate the mechanisms involved at each stage of gelation. Protein molecules are polymers composed of twenty different amino acids linked together by peptide bonds. The function of each protein depends on its three-dimensional structure (Branden and Tooze, 1991). Due to the difficulties in determining a particular protein's tertiary structure, the knowledge of its secondary structure can be used to predict its overall three-dimensional structure and consequently the protein's function (Yada et al., 1988). The secondary structure is usually classified as a-helix, p-sheet or random coil. Branden and Tooze (1991) explained that, for the a-helices, the cp and \\t angles for the consecutive residues are around -60° and -50°, respectively. (q> is the angle of rotation around the N-Ca bond, whereas vy is the angle around the Ca-C bond from the same Ca-atom.) An a-helix has 3.6 residues per turn. On the other hand, p-sheets are composed of several regions, p-strands, of the polypeptide chain. Each strand contains from five to ten residues. The strands are aligned adjacent to each other. The p-sheet is called parallel when the amino acids in the aligned p-strands run in the same biochemical direction, amino terminal to carboxy terminal. When the amino 15 Chapter 2. Raman spectroscopy of surimi acids in successive strands run alternatively once from the amino terminal to the carboxy terminal and once from the carboxy terminal to the amino terminal, the sheet is called antiparallel (Branden and Tooze, 1991). Raman spectroscopy is a vibrational spectroscopic technique which can be used to monitor changes in chemical structure and environmental changes around atoms. The Raman scattering phenomenon which results in shifts in wavelength of an exciting incident beam is due to inelastic collisions of the incident photons with sample molecules. This technique is distinguished by its ability to analyze solids as well as solutions, and therefore, has great potential to study food systems (Li-Chan et al., 1994). Raman spectroscopy has been used to study the myosin substructure (Carew and Asher, 1975) and the effects of inorganic salts on myosin solutions (Barrett and Peticolas, 1978). Caille et al. (1983) reported on the effects of Mg2+, ATP and Ca 2 + on the Raman spectra of intact muscle fibers and internally perfused fibers. Raman spectroscopy was successfully employed to study the gelation of lysozyme (Li-Chan and Nakai, 1991) and the gelation of whey proteins (Nonakaetal., 1993). Since the Raman spectra of macromolecules such as proteins often consist of broad bands from overlapping components, many attempts have been made to apply curve^fitting or mathematical deconvolution techniques to aid in the 16 Chapter 2. Raman spectroscopy of surimi interpretation of the spectral data. Deconvolution is the unraveling of a band containing overlapping peaks into separate peaks. This method is utilized in many areas of science and engineering (i.e., telephone communication, seismology, etc.) where spectral peaks are overlapping and are therefore difficult to interpret. Deconvolution was successfully used by Luu et al. (1982) to study solvent-solute interactions in aqueous solutions, while Mathlouthi and Portmann (1990) used deconvolution to compare Raman spectra of water with aqueous solutions of sweeteners. While Fourier deconvolution has been the most commonly used method, maximum-likelihood spectral restoration or deconvolution has recently been reported to give superior resolution enhancement (deNoyer and Dodd, 1991). Unlike Fourier deconvolution which ignores noise, maximum-likelihood deconvolution takes noise into consideration. The maximum-likelihood deconvolution requires (Mendel, 1990): • the specification of a probability model for the measured output; • the determination of a formula for the likelihood function; and, • the maximization of the likelihood function. For noise with a normal distribution, the likelihood function to be maximized is (deNoyer and Dodd, 1991): 17 Chapter 2. Raman spectroscopy of surimi P = l 1 u e x p [ - — j ] 1 = i -42noi 2<jt where {y1( y2 l y„} is a measured data set, {CH, O2, on} is a more resolved spectrum, ® denotes convolution, a is the standard deviation and s is the peak shape function. To our knowledge, these techniques have not been applied before to the study offish myofibrillar proteins. The objective of this study was to investigate the mechanisms of Pacific whiting surimi gelation under different processing conditions using Raman spectroscopy, a method that offers the major advantage of being able to study changes in protein structure in intact pastes and gels, without having to isolate or extract the myofibrillar proteins. 2.2 Materials and methods 2.2.1 Materials Frozen surimi made from Pacific whiting was donated by Ucluelet Seafood Processors Ltd. (Ucluelet, B.C., Canada). Five types of samples were studied: raw surimi, salted surimi (ground with 3% salt in a vacuum cutter 18 Chapter 2. Raman spectroscopy of surimi [Stephan, Model VCM-5, Columbus, OH]), set gel (32°C for 19 min), cooked gel (86°C for 12 min) and set-cooked gel (setting at 32°C for 19 min followed by cooking at 86°C for 12 min). A more detailed description of the procedures used for preparing surimi gels is given in section 4.3.2.2. 2.2.2 Raman spectroscopic analysis The Raman spectrum of each sample was measured on a Jasco Model NR-1100 laser Raman spectrophotometer (Jasco Inc., Tokyo, Japan) with excitation from the 488-nm line of a Spectra-Physics Model 168B argon ion laser (Spectra-Physics, Mountain View, CA). A schematic layout of the Raman spectrophotometer used is shown in Figure 3. Laser light is focused on the sample. The resulting scattered light is accumulated by tollection optics" and directed to a double monochromator. The latter instrument separates the scattered light depending on its frequency. Then, the Raman spectrum is detected before being stored in a personal computer interfaced to the instrument (Li-Chan et al., 1994). The Raman scattering of samples placed in hematocrit capillary tubes in a transverse/transverse arrangement (capillary held horizontally and incident laser beam perpendicular to the capillary axis) was measured at ambient temperature under the following conditions: laser power, 100 mW; slit height, 4 mm; spectra resolution, 5 cm"1 19 Chapter 2. Raman spectroscopy of surimi Collection optics Laser 1—H Sample o Double Monochromator Detector Computer Figure 3. A schematic layout of a Raman spectrophotometer 20 Chapter 2. Raman spectroscopy of surimi at 19,000 cm"1; sampling speed, 120 cm"1 min"1 with data collected at every cm"1; 6 to 10 scans averaged per sample. These conditions were based on the recommendations of Nonaka et al. (1993), while taking into consideration the nature of our samples (surimi paste and gels). Duplicate samples were analyzed in all cases. Spectral smoothing (to improve the signal to noise ratio), baseline correction, normalization against the CH 2 bending vibration at -1450 cm"1, maximum-likelihood spectral restoration or deconvolution (to unravel overlapping peaks), integration (to determine peak areas) and other computations were performed using LabCalc (Galactic Industries Corp., Salem, NH) with Square Tools (Spectrum Square Associates, Ithaca, NY) software on an IBM compatible (486DX-33) personal computer. Estimation of the secondary structure composition of the samples based on the Raman spectra in the amide I region was carried out using the Raman spectral analysis package (RSAP) of Przybycien and Bailey (1989), which is based on the algorithms of Williams (1983) for least-squares analysis of the amide I band. Assignments of some peaks in the Raman spectra to specific vibrational modes of amino acid side chains or the polypeptide backbone were made according to published literature (Tu, 1986; Colthup et al., 1990; Li-Chan et al., 1994). 21 Chapter 2. Raman spectroscopy of surimi 2.2.3 Measurement of surimi gel properties The quality of the surimi gels obtained after setting alone, cooking alone, and setting followed by cooking needs to be studied in order to try to correlate it to the Raman spectroscopy results. The surimi gel quality tests performed were the compression test and the fold test. The compression test is an indicator of the gel cohesiveness. It was carried out using an Instron Universal Testing Instrument (Model 1122, Canton, Massachusetts). The test consisted of compressing a surimi gel cylinder, having a diameter of 30 mm and a height of 20 mm, to 90% deformation at a cross head speed of 20 mm/min. The variation of force with time was recorded. The gel strength was the product of the force and the distance at the point of rupture (Lee, 1984) (see Figure 4). The distance at the point of rupture was calculated by multiplying the time of rupture by the cross head speed. The fold test indicates gel elasticity. It was conducted by folding a 3 mm thick slice of surimi gel (diameter = 30 mm) slowly in half lengthwise and then in half again while examining it for structural failure. A rating score from 1 to 6 was used. The rating system is shown in Table 2 (Lanier, 1992). 22 Chapter 2. Raman spectroscopy of surimi Figure 4. Example of a force profile recorded by the Instron instrument 23 Chapter 2. R a m a n spectroscopy o f surimi Table 2. Rating system of the fold test Fold score Gel condition 6 No crack showing after folding twice 5 No crack showing after folding in half 4 Cracks gradually when folded in half 3 Cracks immediately when folded in half 2 Does not break by finger pressure 1 Breaks by finger pressure 24 Chapter 2. Raman spectroscopy of surimi 2.3 Results and discussion The set gel had a gel strength of 246 ± 35 N.mm (n=4) and a fold score of 3 while the cooked sample had a gel strength of 2972 + 87 N.mm (n=4) and a fold score of 4. The set and then cooked gel had the best properties consisting of a gel strength of 3897 + 104 N.mm (n=4) and a fold score of 6. Raman spectra were measured in three wavenumber regions for each sample. These regions were 450-600 cm'1 (to assess S-S stretching of disulfide bonds), 610-2000 cm"1 (to estimate secondary structure and various side chain vibrations) and 2500-3400 cm"1 (to evaluate C-H stretching vibrations of aliphatic residues). The Raman spectra were reasonably reproducible. 2.3.1 Disulfide (S-S) bond stretching vibrations in the 450-600 cm'1 region Figure 5 shows the Raman spectra of the five types of samples studied. A major peak was found at 529 cm"1 (raw, salted and set samples) or at 531 cm"1 (cooked and set-cooked samples), assigned to the S-S stretching vibrations of disulfide bonds in a gauche-gauche-trans conformation (Li-Chan et al., 25 Chapter 2. Raman spectroscopy of surimi Wavenumber ( c m " 1 ) Figure 5. Raman spectra of surimi and gels in the 450-600 cm "1 region showing the S-S stretching band near 530 cm"1 26 Chapter 2. Raman spectroscopy of surimi 1994). Although the normalized intensity of this band increased only slightly after salting and setting, a considerable increase was observed in the cooked gel and especially in the set-cooked gel (Figure 6). This may be attributed to either an increase in the number of disulfide bonds or a change in the environment around existing disulfide bonds during the gel strengthening stage at high temperatures. Increase in the intensity of the Raman band around 528 cm"1 was previously observed upon gelation of hen egg white lysozyme at high temperature (Li-Chan and Nakai, 1991), whereas thermally-induced gelation of whey proteins resulted in decreased intensity of the S-S stretching bands near 508 cm"1 (Nonaka et al., 1993). An increase in disulfide bonds due to gel strengthening of fish proteins was proposed by Itoh's group (1979) who found that the sulfhydryl content of carp actomyosin solution decreased with increasing temperature. Therefore, it is thought that the cooking treatment caused a considerable increase in the number of disulfide bonds of the surimi gels of this study. 2.3.2 Backbone and side chain vibrations in the 610-2000 cm'1 region The Raman spectrum of raw surimi in this region is shown in Figure 7. A tentative assignment of the major peaks of this spectrum to vibrational modes of the amino acid side chains or peptide backbone is depicted in Table 3. Further detailed analysis of selected bands in this region was carried out. 27 Chapter 2. Raman spectroscopy of surimi 2.5 Raw Salt Set Cook Set & cook Treatment Figure 6. Variation with treatment of the normalized intensity of the 530 cm band 28 Chapter 2. Raman spectroscopy of surimi CO to CD Wavenumber ( cm 1 ) Figure 7. Raman spectra of raw Pacific whiting surimi in the 610-2000 cm region 29 Chapter 2. Raman spectroscopy of surimi Table 3. Tentative assignment of some major bands in the Raman spectrum of the raw sample (610-2000 cm"1) wavenumber (cm"1) assignment 761 Trp 834, 850 Tyr 879 Trp 940 a-helix 1005 Phe 1062 p-sheet 1128 Backbone C-N stretch 1213 Tyr, Phe 1254 Amide III 1303 CH bending or amide III (a-helix) 1322, 1342 Trp 1403 Asp, Glu COO" 1450 CH 2 bending 1658 Amide I 30 Chapter 2. Raman spectroscopy of surimi 2.3.2.1 The tyrosine doublet near 830 and 850 cm"1 The original and the deconvoluted Raman spectra in the wavenumber region from 820 to 912 cm'1 for the five types of samples are shown in Figures 8a and 8b, respectively. Special attention was given to the bands located near 830 and 850 cm'1 because of their usefulness in monitoring the microenvironment around tyrosyl residues (Li-Chan et al., 1994). This tyrosine doublet band arises from vibrations of the para-substituted benzene ring which are affected by the environment and the involvement of the phenolic hydroxyl group in hydrogen bonding. In the case of tyrosine residues which are exposed to the aqueous or polar environment and act as simultaneous acceptor and donor of moderate to weak hydrogen bonds, the intensity ratio of the doublet bands (Uso/sa)) usually ranges from 0.90-1.45, but can be as high as 2.5. On the other hand, U50/830 values for tyrosine residues which are buried in a hydrophobic environment and which tend to act as hydrogen donors usually range between 0.7 and 1.0, and can be as low as 0.3 in the case of extremely strong hydrogen bonding to a negative acceptor (Tu, 1982; Li-Chan et al., 1994). The intensity ratio ISSO/KJO of raw and salted surimi were 1.18 and 1.14, respectively suggesting that the tyrosine residues of these samples were mainly exposed. The ISSO/KJO ratios decreased to 0.59 and 0.22 for the set and 31 Chapter 2. Raman spectroscopy of surimi gOO 860 820 890 870 850 830 Wavenumber (cm ) Wavenumber (cm 1) Figure 8. Raman spectra of surimi and gels in the 820-912 cm"1 region, including the 830/850 tyrosine doublet, a) Original spectra, b) Deconvoluted spectra 32 Chapter 2 . R a m a n spectroscopy o f surimi set-cooked samples, respectively. This suggests that some tyrosine residues became buried in a more hydrophobic environment and were involved as strong hydrogen bond donors after these samples were processed by setting as well as setting followed by cooking. On the other hand, the sample which received only the cook treatment had a high USO/KJO ratio of 3.1, indicating that some of the tyrosine residues became exposed to a more polar environment and were involved as strong hydrogen bond acceptors on the surface of the protein molecule. Tu (1986) developed an equation for estimating the proportions of the buried and the exposed tyrosine residues. The equation is: 0.5 N buried + 1.25N exposed ~ '850/830 where N is the mole fraction (N buhed + N exposed = 1). The numerical constants in this equation were obtained by fitting the general form to the X-ray diffraction data of Kannan et al. (1975). Table 4 lists the proportions of the buried and the exposed tyrosine residues (TR) estimated by Tu's equation. This table shows that the TR of the raw and the salted surimi were mainly exposed. Setting caused most of the TR to become buried. The negative numbers shown for the cooked and the set-cooked samples show the limitations of Tu's equations at too high or too low values of USO/KIO- However, these numbers might be an indication that all the TR of the cooked sample were exposed whereas all the TR of the set-cooked gel were buried. 33 Chapter 2. R a m a n spectroscopy o f surimi Table 4. Estimated proportions of the buried and the exposed tyrosine residues Sample '850/830 N exposed N buried Raw 1.18 0.91 0.09 Salted 1.14 0.85 0.15 Set 0.59 0.12 0.88 Cooked 3.1 3.47 -2.47 Set-cooked 0.22 -0.37 1.37 34 Chapter 2. R a m a n spectroscopy o f surimi 2.3.2.2 Secondary structure estimation from the amide I and ill bands The -CO-NH- amide or peptide bond has several distinct vibrational modes, with the amide I band near 1650 cm"1 and the amide III band near 1250 cm'1 being most easily identified in the Raman spectra of proteins. The exact location of these bands depends on the secondary structure of the polypeptide chain, and these features are therefore useful for estimating secondary structure fractions of proteins. Due to possible overlap of bands from the solvent water in the amide I region and from miscellaneous C-H bending and aromatic ring vibrations in the amide III region, it is recommended that both regions be analyzed in order to obtain a more reliable interpretation of the changes in the secondary structure of proteins. The Raman spectra of the surimi and gelled samples were, therefore, measured in both regions, and maximum-likelihood deconvolution was used to enhance the resolution of individual components contributing to the broad bands in each of these regions. The amide I band arises primarily from in-plane peptide C=0 stretching vibrations and also has contributions from in-plane N-H bending vibrations. Generally, proteins with high a-helical content show an amide I band centered around 1645-1657 cm"1 while those with predominantly p-sheet structures 35 Chapter 2. Raman spectroscopy of surimi have a strong band at 1665-1680 cm"1. Proteins with a high proportion of undefined or random coil structure have the amide I band centered near 1660 cm'1. The original and the deconvoluted Raman spectra in this region are shown in Figures 9a and 9b, respectively. For the raw surimi, a major band at 1656 cm"1 suggestive of a-helical structure and a minor band near 1679 cm"1 typical of 3-sheet structure are present. The salted surimi indicates a trend towards uncoiling of helices with the major band centered near 1661 cm'1, while the minor band near 1683 cm"1 is suggestive of p-sheet or turns. For each of the three remaining samples, increasing dominance of the band assigned to p-sheet structure was evident. The results of quantitative estimation of the secondary structure fractions using the RSAP program for least squares analysis of the amide I region are shown in Table 5. This algorithm has been reported to overestimate antiparallel p-sheet content and underestimate a-helical content, but is useful to assess relative changes in the secondary structure of proteins as a function of different treatments (Przybycien and Bailey, 1989). The results suggest that raw surimi contained predominantly a-helical structure with a smaller proportion of antiparallel p-sheet, in agreement with visual inspection of the deconvoluted spectra. Grinding with salt resulted in a slight increase in the random coil or unordered fraction, at the expense primarily of the a-helical fraction. After setting, similar proportions of a-helical and antiparallel p-sheet 36 Chapter 2. Raman spectroscopy of surimi Figure 9. Raman spectra of surimi and gels in the 1590-1720 cm"1 region which includes the amide I band, a) Original spectra, b) Deconvolved spectra' 37 Chapter 2. R a m a n spectroscopy o f surimi Table 5. Secondary structure fractions estimated from the amide I band by RSAP for the five types of samples Treatment total a-helix antiparallel parallel p- total random p-sheet sheet coil Raw surimi 0.44 0.24 0.00 0.32 Grinding with salt 0.37 0.26 0.00 0.37 Setting 0.36 0.41 0.00 0.23 Cooking 0.09 0.31 0.20 0.40 Setting & cooking 0.20 0.49 0.00 0.31 38 Chapter 2. R a m a n spectroscopy o f surimi were observed. Further increase in antiparallel p-sheet resulted upon cooking after setting, while the cooked gel without prior setting showed an increase in parallel p-sheet structure and a dramatic decrease in a-helix content. Similar trends were noted by examining the amide III band in the Raman spectra of the different samples. The amide III band results primarily from the C-N stretching and N-H in-plane bending vibrations of the peptide bond. It is less intense and located in a broad region from 1260-1300 cm"1 for proteins with high contents of a-helical structures, p-sheet structures usually lead to a more intense band near 1238-1245 cm"1 while random coil structures appear near 1250 cm"1. Figures 10a and 10b show the original and the deconvoluted Raman spectra in the region which includes the amide III bands from 1230-1300 cm"1. Detailed inspection of the bands suggests that raw and salted samples contain several bands in the amide III region typical of high a-helical contents. The heat-set sample has increased the intensity of the component bands near 1234-1242 cm'1 suggesting an increased proportion of p-sheet structures. The trend towards higher p-sheet structure at the expense of the a-helical content is even more evident in the cooked and set-cooked gels. The cooked gel shows two main component bands in this region, while the set-cooked gel shows only one major band at 1247 cm"1 with smaller bands near 1280-1290 cm"1. 39 Chapter 2. Raman spectroscopy of surimi Figure 10. Raman spectra of surimi and gels in the 1220-1275 cm"1 region which includes the amide III band, a) Original spectra, b) Deconvoluted spectra 40 Chapter 2. Raman spectroscopy of surimi The Raman spectra in the amjde I and III regions thus collectively indicate a trend toward decreasing a-helical content and increasing p-sheet content upon setting, which became more pronounced after the cooking treatments. These changes could be due to an unfolding of helical structures, followed by the formation of sheet structures possibly through intermolecular interactions between exposed hydrophobic residues. Chan et al. (1992) also reported an unraveling of the coiled helical structure of fish myosin after heating to 50°C, followed by exposure of hydrophobic residues. Moreover, Niwa (1992) reported an increase in the hydrophobicity of actomyosin solutions caused by setting. He also indicated that during cooking (kamaboko stage), the interactions between the exposed hydrophobic sites become stronger. 2.3.3 C-H Stretching vibrations in the 2500-3400 cm'1 region Figure 11 shows the Raman spectra of the five samples in this region, which corresponds to various C-H stretching vibrations of primarily aliphatic residues. For each sample, a major peak is shown between 2932 and 2939 cm"1. A slight increase in wavenumber of the Raman shift was observed in the gelled samples compared to the raw and salted surimi. However, information on interpretation of this shift is lacking in the literature. Arteaga (1994) reported that a slight shift to higher wavenumbers was observed upon addition of water or deuterium oxide to simple organic solvents such as 41 Chapter 2. Raman spectroscopy of surimi Raw -.-< W <D a i—i (D > -<-( -•-> cd «—i CD Salted Set Cooked I 1 I 1 1 3250 3000 2750 2500 Wavenumber ( c m - 1 ) Figure 11. Raman spectra of surimi and gels in the 2500-3400 cm"1 region showing the C-H stretching band 42 Chapter 2. Raman spectroscopy of surimi alcohols, dioxane and dimethyl sulfoxide. Several proteins also showed a shift to higher wavenumbers in the presence of urea compared to non-denaturing aqueous buffers. It was speculated that protein unfolding with resulting exposure of methyl and methylene groups should produce spectral changes to higher Raman shifts. The normalized intensity as well as the area of the peak assigned to C-H stretching vibrations decreased in the gelled samples compared to the raw and salted surimi (Figure 12). Again, the interpretation of these results is not clear, due to the scarcity of information in the literature on the study of amino acid C-H stretching vibrations. One may speculate that the higher Raman shifts and decrease in the intensity of the vibrations of the C-H stretching band demonstrate changes in the environment of aliphatic C-H groups caused by surimi gelation. Larsoon and Rand (1973) also found that the intensity of the 2930 cm'1 peak increased with increased degree of polar environment of the hydrocarbon chains. 43 Chapter 2. Raman spectroscopy of surimi Raw Salt Set Cook Set & cook Treatment Figure 12. Variation with treatment of the intensity and of the area of the ~2935 cm'1 peak 44 Chapter 2. Raman spectroscopy of surimi 2.4 Conclusions This study demonstrates the application of Raman spectroscopy to investigate changes in the protein structure of intact surimi and gels from Pacific whiting. Several important changes were observed during surimi processing. The intensity of the peak assigned to disulfide bond S-S stretching vibrations increased slightly after the setting treatment and considerably after the cooking treatment. Setting decreased a-helical content and increased p-sheet content. This trend was more pronounced after the cooking treatments. A slight increase in the Raman shift and a decrease in intensity of C-H stretching vibration of aliphatic residues was observed following the cooking operations. These changes could be caused by an unfolding of helical structures followed by interactions between the exposed hydrophobic residues through formation of intermolecular sheet structures. The gel properties, gel strength and fold score of the set and cooked gel were higher than those of the cooked gel. The properties of the latter gel were better than those of the gel that was subjected to the setting treatment alone. Therefore, setting followed by cooking is the recommended method for the manufacture of surimi gels. 45 Chapter 3. Optimization of processing conditions Chapter 3 Optimization of the Setting and Cooking Conditions for Kamaboko Making from Pacific Whiting Surimi 3.1 Introduction One of the most important functional properties of surimi is its gelation property. When surimi is chopped with salt (usually between 2 and 3% of surimi paste weight), its myofibrillar proteins become soluble in water (Lee, 1984). When salt is introduced, the salt ions bind to the oppositely charged groups exposed on the protein surface and the proteins dissolve because of their increased affinity for water (Niwa, 1992). When salted surimi is heated, it forms a fish gel structure called kamaboko. The formation of a three dimensional gel network requires the occurrence of at least three cross-links on every molecule (Niwa, 1992). Usually slow setting at temperatures below 50°C precedes cooking at higher temperatures. This results in a stronger gel than cooking without a slow set (Lee, 1984). 46 Chapter 3. Optimization of processing conditions Individual effects of heating parameters on surimi gelation have been studied. For instance, Douglas-Schwartz and Lee ( 1988) tested the effect of setting for 20 minutes at four temperatures; 40, 60, 80 and 90°C. They found that setting of Alaska pollock surimi and red hake surimi at 40°C resulted in gels with the best cohesiveness and water holding capacity. For testing the quality of surimi gels, Lanier (1992) suggested varying setting time and temperature but cooking temperature was constant (90°C for 15 minutes). Still, there is a lack of information about the combined effects of setting time and temperature as well as cooking time and temperature on gel quality especially for surimi made from Pacific whiting, an abundant species on the Pacific coast of North America. To obtain this information, an experiment was designed to study the individual and combined effects of setting and cooking conditions on the gel quality variables. " Then optimization was performed to determine the processing conditions which resulted in the best surimi gel. What follows is a brief description of some the optimization techniques that were used in this study. 47 Chapter 3. Opt imiza t ion o f processing condit ions 3.1.1 The Nelder-Mead simplex optimization method (Khuri and Cornell, 1987; Walters etal., 1991). This is a 'direct search method" For k independent variables, it uses a geometric figure; a simplex that is composed of k+1 points or vertices Pi (i=1,2,..,k+1). The objective function Yj = f(Pj) is evaluated at each vertex. For a minimization problem, the point having the maximum Y, P H , is replaced by a new point according to one of these operations: Reflection The reflection of P H is P* which is defined by: P* = (1+a )C -aP H where C is the centroid of the points other than P H , a > 0 is the reflection coefficient (i.e., a = 1/2, 2/3, or 1). If Y L< Y* = f(P*) < Y H , [YL = min (Yi), 1 < i < k+1], then P H is replaced by P* and the process is restarted. Expansion If Y* < Y L , then P* is expanded to P E by using the following equation: 48 Chapter 3. Optimization of processing conditions P E =YP* + ( 1 ^ ) C where y > 1 is the expansion coefficient (i.e., y = 1.5). If Y E = f(PE) < YL, then replace P H by P E and restart. Otherwise, if Y E = f(PE) > Y L, then replace P H by P* and restart the process. Contraction If the reflection of P H to P* results in Y* > Y| for all i * H, then a new P H is to be defined as the old P H or P* selecting the point that has the lower Y value. Let us define P c as: P C = P P H + (1 -P )C where p < 1. If Y c = f(Pc) > min (YH,Y*) then replace all the values of Pj with (Pi + PL)/2 and restart the process. Otherwise, replace P H with P C and restart. To check the optimization results obtained by the Nelder-Mead simplex method, another type of optimization technique was used. This method is a gradient based method that is described below. 49 Chapter 3. Optimization of processing conditions 3.1.2 The Gradient based optimization methods These optimization methods use the partial derivatives of the objective function in order to find the direction of search. For a minimization problem and when the steepest descent method (a common gradient based method) is employed, the gradient vector at a point gives the direction of the greatest decrease in the objective function and, therefore, guides the transition from one point to another until an optimal value is found (Edgar and Himmelblau, 1988). The NLPQLO gradient based method uses a quadratic approximation of the objective function (Vaessen, 1984). The simplex method and the gradient based method are local optimization methods, e.g., they are only able to find a single minimum even though other minima having lower objective function values may also be located in the constrained search field. A global optimization method was used to assess the results obtained by the two local techniques. This global method is the Level Set Programming technique. 3.1.3 The Level Set Programming (LSP) method The fact that LSP is a global optimization technique is very useful when dealing with problems having many local optima. The LSP theory was initially presented 50 Chapter 3. Opt imiza t ion o f processing condit ions by Chew and Zheng (1988). This nonlinear programming technique belongs to the direct search category. It does not need gradient evaluations and is mainly based on the determinations of the objective and constraint (if any) functions at solution points. An optimization software based on this method was developed by Yassien (1993). The method consists of randomly generating solutions having objective functions less than or equal to a pre-selected value called the level set value (LSV). The mean of these values defines a new LSV and hence a new level set interval. The search process continues at progressively decreasing LSV until the convergence criteria are satisfied. Then, the global solution or set of solutions is determined (Yassien, 1993). The objective of this study was to determine the optimal setting time and temperature and cooking time and temperature resulting in Pacific whiting surimi gel with the best combination of gel strength, fold score, color and expressible liquid. 3.2 Materials and methods The Response surface methodology (RSM), central composite design was used in order to correlate the response variables to the independent variables. Table 6 depicts the experimental design. For four independent variables, it consists of 51 Chapter 3. Optimization of processing conditions Table 6. Experimental design for the optimization of surimi heating conditions Factors CODES -2 -1 0 1 2 X, setting temperature (°C) 20 30 40 50 60 X 2 setting time (min.) 0 10 20 30 40 X3 cooking temperature (°C) 70 75 80 85 90 X4 cooking time (min.) 5 15 25 35 45 X1 X2 X3 X4 0 0 0 0 0 0 0 0 -2 0 0 0 2 0 0 0 0 -2 0 0 0 2 0 0 0 0 -2 0 0 0 2 0 0 0 0 -2 0 0 0 2 0 0 0 0 0 0 0 0 52 Chapter 3. Optimization of processing conditions 24 (16) factorial settings, 2x4 (8) axial settings and 6 center point replicates (Khuri and Cornell, 1987). The four independent variables were setting temperature (X^, setting time (X2), cooking temperature (X3) and cooking time (X4). Experiments were carried out on Pacific whiting surimi donated by Ucluelet Seafood Processors Ltd. (Ucluelet, B. C, Canada). For each set of processing conditions, surimi was mixed with 3% salt (total weight basis) before being set and cooked as in section 4.3.2.2. The setting and cooking conditions are depicted in Table 6. The quality of the obtained surimi gels was tested by performing the following quality tests. 3.2.1 Gel strength (compression test) This test is an indication of gel cohesiveness (Lee et al., 1987). The procedure for this test can be found in section 2.2.3. 3.2.2 Fold score This test indicates gel elasticity (Lee, 1984). It was performed according to the procedures outlined in section 2.2.3. 53 Chapter 3. Opt imiza t ion o f processing conditions 3.2.3 Gel color test The whiteness of a surimi gel sample (thickness = 50 mm, diameter = 30 mm) was measured using a color measurement instrument (Hunterlab, Model Labscan II, Reston, Virginia). The test consists of measuring L (lightness) on a 0-100 scale from black to white in shades of gray, a denoting redness (if positive) or greenness (if negative), and b indicating yellowness (if positive) or blueness (if negative). The "whiteness" index, C, for the overall color evaluation is (Lanier, 1992): C = 100-[(100-L)2+a2 + b 2] 0 5 3.2.4 Expressible liquid measurement Expressible liquid is a measure of the water binding ability of the gel. It was measured by compressing a surimi gel cylinder (20 mm height and 30 mm diameter), using the Instron Universal Instrument (Model 1122, Canton, Massachusetts) at a cross head speed of 10 mm / min, into a 1 mm thick disk in 1.9 minutes and absorbing the expressed moisture in two pieces of filter paper for five minutes. Free water is more easily expressed upon compression than bound water. Expressible liquid, Ei, is defined as (Niwa, 1992): 54 Chapter 3. Opt imiza t ion o f processing condit ions E, (%)=(Wi-Wf)x 100/ Wi where Wj is the weight of the surimi gel before compression while W f is the surimi gel weight after compression. These quality tests were conducted in triplicates. The optimization techniques used were the Nelder-Mead simplex method (Arteaga, 1992), the NLPQLO gradient based method (Vaessen, 1984) and the Level Set Programming method (Yassien, 1993). Xi had a lower limit of 22°C and an upper limit of 45°C. X 2 and X4 varied between 10 and 40 minutes. The lower and upper limits for X3 were 70 and 90°C, respectively. 3.3 Results and discussion Table 7 summarizes the results of the optimization experiments. For each set of independent variables, setting temperature (X^, setting time (X2), cooking temperature (X3) and cooking time (X4), it shows the corresponding response variables, namely gel strength (Gs), fold score (Fs), color (C) and expressible liquid (Ei). G s values were higher for cooking temperatures of 80°C and above. The fold score values were also better at these cooking temperatures. A lower 55 Chapter 3. Optimization of processing conditions Table 7. Experimental results Row X, x 2 Xa X 4 G s 1 F s c 4 j E, (°C) (min.) (°C) (min.) (N.mm) (%) 1 30 10 75 15 130 2.0 55.7 18.1 2 50 10 75 35 940 3.5 55.1 16.0 3 30 30 75 35 230 3.0 56.4 16.7 4 50 30 75 15 330 2.5 55.6 16.2 5 30 10 85 35 2860 5.0 55.8 19.1 6 50 10 85 15 3080 4.5 55.6 13.0 7 30 30 85 15 3950 6.0 57.6 19.3 8 50 30 85 35 720 3.0 55.2 14.5 9 40 20 80 25 2540 5.0 55.9 19.5 10 40 20 80 25 2160 5.0 56.3 19.5 11 30 10 75 35 520 4.0 56.5 16.5 12 50 10 75 15 150 2.5 54.1 17.4 13 30 30 75 15 110 1.5 56.9 18.9 14 50 30 75 35 720 3.0 56.8 14.4 15 30 10 85 15 3680 5.0 56.8 21.1 16 50 10 85 35 1470 3.5 55.6 14.9 17 30 30 85 35 3640 3.5 55.5 20.3 18 50 30 85 15 1270 3.0 55.4 14.3 19 40 20 80 25 2120 5.0 54.9 18.8 20 40 20 80 25 2940 6.0 59.1 12.4 21 20 20 80 25 3050 6.0 68.1 13.2 22 60 20 80 25 3050 6.0 65.4 12.7 23 40 0 80 25 2920 4.0 59.6 13.1 24 40 40 80 25 2800 6.0 59.9 12.1 25 40 20 70 25 340 2.0 67.2 16.7 26 40 20 90 25 1620 3.0 67.0 14.8 27 40 20 80 5 2810 6.0 58.3 10.4 28 40 20 80 45 2150 5.0 53.5 19.1 29 40 20 80 25 2960 6.0 60.7 12.5 30 40 20 80 25 2960 6.0 61.4 13.0 'Arithmetic mean (n=3) Chapter 3 .Optimization of processing conditions fold score, however, was observed for surimi cooked at 80°C without previous setting. This shows the effect of setting on the gel elasticity and supports the findings of Lanier et al. (1982) who reported that surimi gels prepared by setting prior to cooking were more elastic than those that were directly cooked. Setting at 50°C for 30 minutes followed by cooking at 85°C for 35 minutes resulted in weaker and less elastic gels than were obtained under the same cooking conditions but at a setting temperature of 30°C for 30 minutes. This could be due to the gel softening phenomenon, called modori in Japanese, which usually occurs when surimi is heated at temperatures between 50 and 70°C. Niwa explained that proteolytic degradation of myosin at the modori temperature range may be the cause of gel softening. Most of the C values were higher than 55 which corresponds to grade 4 according to the classification reported by Lanier (1992). (Grade 1 corresponds to C > 60.0.) Seven out of eight of the combinations that resulted in a fold score of 6 also resulted in low expressible moisture. This supports the deduction of Tagagi (1973) that the more elastic the surimi gel the lower the expressible moisture. The six center point replicates were reasonably reproducible for G s, F s and C. However, three points resulted in high E, whereas the other points gave low E, values. Since these six points resulted in high fold scores, it is thought that the high values of Et observed for three points could be due to experimental error. In most of the cases, less moisture was expressed from gels that were cooked at 80°C and above. 57 Chapter 3. Opt imiza t ion o f processing condit ions Each one of these response variables was fitted by an empirical equation in order to correlate them to the independent variables. The empirical equation has the general form: Y = Bo + £/3iXi + fjBnXi2 + 2 ^PaXiXj i=l i=l i=l ;'=i+l where X4 are the input variables which influence the response variable Y while Po, Pi(i=1,...,4) and Py (i=1,...,3; j=i+1,..., 4) are the model parameters. This is the second order model (SOM) used in the Response Surface Methodology (Khuri and Cornell, 1987). 3.3.1 Curve fitting results Curve fitting was carried out using the Systat 1990 (Systat, Inc., Evanston, IL) software. The parameter estimates obtained by the multiple, stepwise regression of the measured gel strengths are shown in Table 8. Only the terms that significantly affected the model (P < 0.05) are included in the table and were taken into account in the optimization procedures which are discussed in section 3.3.2. The logarithm of the gel strength was used in the regression analysis as the unaltered values, presumably because of their two-orders-of-magnitude range (from 110 to 3950 N.mm), yielded poor regression results (R2 < 0.50). Table 8 shows that cooking temperature had the strongest influence on gel 58 Chapter 3. Optimization of processing conditions Table 8. Regression results where "log (Gs)" is the response variable Coefficient Rz Standard Error of Estimate Po -114.024 0.858 0.180 01 0.284 p3 2.606 P4 0.285 Pl3 -0.004 P33 -0.014 P34 -0.004 59 Chapter 3. Opt imiza t ion o f processing conditions strength while setting temperature and cooking time had less effect on G s. However, setting time, at least in the range of 10 to 40 minutes, did not affect G s. Table 9, which compares the experimental and estimated values of G s, Fs, C and Ei, also demonstrates that the G s data were well fitted by the SOM. The experimental data of F s were also reasonably well fitted by the SOM (Tables 9 and 10). F s was mainly dependent on the cooking temperature. The three other independent variables affected F s only slightly. Table 11 shows that the curve fitting results for the color variable were worse than those of the other response variables. The higher the setting temperature and/or the cooking temperature, the lower the whiteness index of the surimi gel. Setting time and cooking time did not affect color. Measured E, data were well fitted by the SOM (Tables 9 and 12). The model parameters listed in Table 12 demonstrate that the higher the cooking temperature and/or time, the lower the expressible liquid. For all four response variables, the quadratic parameters had very little influence on the overall fit. The fact that cooking temperature was the variable that had the most influence on the gel cohesiveness, elasticity and water holding capacity shows the importance of the cooking stage on the gel texture. At the cooking stage, as found in Chapter 2, more disulfide interactions take place and the p-sheet content increases. This could result in better gel texture. 60 Chapter 3. Optimization of processing conditions Table 9. Experimental and estimated values of the 4 response variables Row# log G s (exp.) log G s (est.) F s (exp.) F s (est.) C (exp.) C (est.) E, (exp.) E, (est.) 1 2.12 2.11 2.00 1.96 55.7 56.8 18.1 19.2 2 2.97 2.88 3.50 2.76 55.1 54.9 16.0 15.2 3 2.37 2.55 3.00 1.61 56.4 55.5 16.7 17.3 4 2.52 2.44 2.50 1.68 55.6 56.2 16.2 17.1 5 3.46 3.52 5.00 3.53 55.8 55.4 19.1 19.7 6 3.49 3.39 4.50 4.17 55.6 68.7 13.0 14.4 7 3.60 3.78 6.00 4.64 57.6 56.7 19.3 19.6 8 2.86 3.14 3.00 1.61 55.2 67.4 14.5 14.6 9 3.40 3.33 5.00 4.61 55.9 58.3 19.5 13.0 10 3.33 3.33 5.00 4.61 56.3 58.3 19.5 12.0 11 2.72 2.55 4.00 2.33 56.5 55.5 16.5 17.3 12 2.17 2.44 2.50 2.40 54.1 56.2 17.4 17.1 13 2.02 2.11 1.50 1.25 56.9 56.8 18.9 19.2 14 2.86 2.88 3.00 2.05 56.8 54.9 14.4 15.2 15 3.57 3.78 5.00 5.36 56.8 56.7 21.1 19.6 16 3.17 3.14 3.50 2.32 55.6 67.4 14.9 14.6 17 3.56 3.52 3.50 2.80 55.5 55.4 20.3 19.7 18 3.10 3.39 3.00 3.45 55.4 68.7 14.3 14.4 19 3.33 3.33 5.00 4.61 54.9 58.3 18.8 13.0 20 3.47 3.33 6.00 4.61 59.1 58.3 12.4 13.0 21 3.48 3.36 6.00 4.99 68.1 52.6 13.2 21.9 22 3.48 3.30 6.00 4.23 65.4 64.0 12.7 14.7 23 3.47 3.33 4.00 2.54 59.6 58.3 13.1 13.0 24 3.45 3.33 6.00 1.10 59.9 58.3 12.1 13.0 25 2.53 0.95 2.00 -1.54 67.2 55.1 16.7 16.5 26 3.21 2.87 3.00 1.42 67.0 67.5 14.8 16.3 27 3.45 3.24 6.00 5.35 58.3 59.1 10.4 21.8 28 3.33 3.43 5.00 3.87 53.5 56.5 19.1 20.0 29 3.47 3.33 6.00 4.61 60.7 58.3 12.5 13.0 30 3.47 3.33 6.00 4.61 61.4 58.3 13.0 13.00 61 Chapter 3. Optimization of processing conditions Table 10. Regression results where "F s" is the response variable Coefficient R2 Standard Error of Estimate Po -354.198 0.80 0.627 Pi 0.631 p2 0.243 p3 8.211 P4 0.848 Pl3 -0.008 P22 -0.007 P33 -0.047 P34 -0.011 62 Chapter 3. Opt imiza t ion o f processing condit ions Table 11. Regression results where "C" is the response variable Coefficient R2 Standard Error of Estimate Po 388.221 0.667 1.911 Pi -4.776 p3 -6.602 Pl3 0.063 P33 0.029 P44 -0.001 63 Chapter 3. Opt imiza t ion o f processing condit ions Table 12. Regression results where "E" is the response variable Coefficient R2 Standard Error of Estimate Po 245 0.892 0.895 P3 -5.1 P4 -1.868 P11 0.013 Pl3 -0.016 P33 0.034 P34 0.010 P44 0.020 64 Chapter 3. Optimization of processing conditions The generated equations for the response variables were then used to determine the optimal processing conditions. 3.3.2 Optimization results Using the empirical equations found for each variable as objective functions, three optimization techniques were used to determine the best set of setting/cooking times and temperatures in each case. These were the Nelder-Mead Simplex method, the Level Set Programming method, and the NLPQLO Gradient method. 3.3.2.1 Results of Nelder-Mead Simplex optimization The results of optimizing one response variable at a time (maximizing G s, F s and C, and minimizing Ei) using the Simplex method are shown in Table 13. This table contains the optimal processing conditions (Xj, i=1,2 4) along with the corresponding values for the four response variables. The optimized values of log (Gs), F S | C and E, were 4.08, 6.41, 70.8 and 12.3, respectively. Apart from the Ei value, the other optimized values were better than the best experimental results. The processing conditions corresponding to a maximized C, 70.8, resulted in a negative F s value. This is a further indication that SOM fit of the color values was relatively poor. 65 Chapter 3. Optimization of processing conditions Table 13. Optimization of individual response variables using the Nelder-Mead Simplex method Variable Maximize G s Maximize F s Maximize C Minimize Ei * (°C) 22.0 24.1 45.0 45.0 X2(min.) 36.1 21.1 40.0 24.3 XaCC) 87.4 85.9 90.0 81.1 X 4 (min.) 10.0 10.0 40.0 25.8 log (Gs) 4.08 4.00 2.23 3.40 F s 3.98 6.41 -4.80 4.26 C 51.7 53.2 70.8 60.8 E, (%) 28.3 26.0 20.3 12.3 66 Chapter 3. Optimization of processing conditions 3.3.2.2 Results of NLPQLO Gradient optimization Table 14 depicts the results of optimizing the response variables using the NLPQLO gradient method. These results include the optimal processing conditions and the corresponding response variable values. The optimized values of log (Gs), Fs, C and Ei were 4.08, 6.64, 72.7 and 12.3, respectively. The maximized F s and C values were higher than those obtained by the Nelder-Mead Simplex method indicating that the Simplex optimizer likely found only local maxima for these two response variables. In addition, the processing conditions corresponding to a maximized C resulted in better values of log (Gs), F s and Ei than those obtained by the Simplex technique. 3.3.2.3 Results of LSP optimization The optimization results obtained using this method were almost the same as those obtained by the NLPQLO gradient method (see Table 15). The optimal setting time values for G s, C and Ei were considerably different from those obtained by the gradient method. However, these response variables, as indicated in Tables 8, 11 and 12, respectively, were independent of setting time over the range from 10 to 40 minutes. Hence, the results of the global LSP method confirmed the results of the local gradient method for the independent variables limits specified in section 3.2. 67 Chapter 3. Optimization of processing conditions Table 14. Optimization of individual response variables using the NLPQLO Gradient method Variable Maximize G s Maximize F s Maximize C Minimize Ei * (°C) 22.0 22.0 45.0 45.0 X2(min.) 20.0 20.0 20.0 20.0 XaCC) 87.5 84.9 90.0 81.2 X 4 (min.) 10.0 10.0 10.0 25.8 log (Gs) 4.08 3.99 3.14 3.40 F s 6.34 6.64 3.13 4.53 C 51.7 52.1 72.7 60.9 E, (%) 28.4 27.0 18.5 12.3 68 Chapter 3. Table 15. Optimization of processing conditions Optimization of individual response variables using the LSP method Variable Maximize G s Maximize F s Maximize C Minimize Et X, (°C) 22.0 22.0 45.0 45.0 X2(min.) 23.8 17.4 30.8 28.0 X3(°C) 87.5 85.0 90.0 81.2 X 4 (min.) 10.0 10.0 10.0 25.8 log (Gs) 4.08 3.99 3.14 3.40 F s 6.09 6.68 1.94 3.80 C 51.7 52.0 72.7 60.9 E. (%) 28.4 27.0 18.5 12.3 69 Chapter 3. Optimization of processing conditions 3.3.2.4 Multiobjective optimization As can be seen in Tables 13 to 15, the optimal conditions for the optimization of one response variable (e.g., Gs) did not give satisfactory results for the other response variables (i.e., Fs, C and Ei). Therefore, multiobjective optimization had to be performed in order to simultaneously optimize the four response variables in question. To fulfill this task, the multiobjective problem was transformed into a single objective problem. This was achieved by dividing each variable by its maximum experimental value and forming a single objective function, Z, as the weighted sum of these normalized variables (de Neufville, 1990). Each variable was normalized in this fashion in order to put all the response variables on the same scale. The expression for Z is: Z = w, [log(Gs )/3.60]+ w2 (Fs/6)+ w3 (C/68.1) - w4 (E,/21.1) where Wi,...,w4 are weights that are chosen arbitrarily taking into consideration the importance of each response variable so as to generate acceptable results for all the four response variables. The terms involving gel strength, fold score and color were added because they were to be maximized. However, the expressible liquid term was subtracted because it was to be minimized (Yassien (1993). 70 Chapter 3. Opt imiza t ion o f processing conditions Several combinations of wk (k=1 4) were tried. Some of these combinations along with the corresponding results are shown in Table 16. Those results were obtained using the LSP method. Emphasis was put on maximizing gel strength and fold score without excessively compromising color or the expressible liquid. In most cases, the color variable was assigned a low weight, 0.1, because its objective function was not well fitted to the measured color data. The Z values could not be compared because the weights in each row were different. The row in the bold font seems to give satisfactory results for all four response variables. These results consisted of a gel strength of 3980 N.mm, a fold score of 5.4, a gel whiteness index of 60.5 and an expressible liquid of 15.6%, corresponding to setting at 39.2°C for 17.4 minutes followed by cooking at 82.8°C for 14.7 minutes. 71 Chapter 3. Optimization of processing conditions Table 16. Multiobjective optimization results w2 W 3 w4 Xi x 2 Xa X4 log G s F s c E, z 1 1 1 1 44.8 17.0 83.2 16.9 3.5 5.0 62.4 13.9 2.0 1 6 0.1 3 36.6 17.4 83.5 11.3 3.6 5.8 59.7 18.2 4.4 1 5 0.1 3 39.2 17.4 82.8 14.7 3.6 5.4 60.5 15.6 3.4 1 6 0.1 3.5 38.5 17.8 83.0 14.1 3.6 5.5 60.3 16.1 4.0 1 6 0.1 4 40.9 17.4 82.6 16.8 3.5 5.2 61.0 14.5 3.6 1 8 0.1 5 40.3 17.4 82.6 15.9 3.5 5.3 60.7 14.9 4.7 1 9 0.1 5 38.9 17.5 83.1 14.1 3.6 5.5 60.6 16.0 5.6 72 Chapter 3. Optimization of processing conditions 3.4 Conclusions and recommendations The experimental results on the effects of processing conditions on surimi gel quality showed that high cooking temperatures (80°C and above) resulted in better gel cohesiveness, elasticity and water holding capacity. They also showed the importance of the setting treatment prior to sample cooking. However, setting should be performed at temperatures below 50°C. The data were well fitted by a second order model which allowed four objective functions to be generated for the four response variables; gel strength, fold score, color and expressible liquid. The optimal processing conditions corresponding to individually maximizing G s, F s and C, and minimizing Ei were determined using the Nelder-Mead Simplex optimization and the NLPQLO gradient optimization. The results obtained using the latter method appeared to be closer to the global optima than those of the Simplex method. The results obtained using a global optimization method, LSP, essentially confirmed the results of the gradient method. Multiobjective optimization was also conducted to simultaneously optimize the four response variables. It gave a set of setting and cooking times and temperatures which led to satisfactory values for all four response variables. However, these optimal processing conditions are only applicable to Pacific whiting kamaboko sausages having the size used in this project. The same 73 Chapter 3. Optimization of processing conditions methodology can be followed to determine the optimal processing conditions for other species and for other surimi-based products. The ingredients used for this work were Pacific whiting surimi and salt. For further studies, it is recommended that the gel texture improving ingredients (GTII) such as starch, egg white or whey protein concentrate be taken into account in the RSM experimental design. The concentrations of these ingredients would therefore be considered as independent variables to be optimized. This will allow a comprehensive study of the effects of setting and cooking temperatures and times along with the effects of the GTII on the surimi gel characteristics. 74 Chapter 4. Production of salmon and herring kamaboko Chapter 4 Kamaboko Production from Roe Herring and Pink Salmon 4 . 1 I n t r o d u c t i o n Alaska pollock has been the most popular species for the production of surimi. This is due to its accessibility, subtle flavor and relatively large size, among other characteristics (Holmes et al. 1992). However, with the increase in surimi production, the stocks of Alaska pollock have decreased dramatically. For instance, the pollock biomass in the eastern Bering Sea was estimated at 5.3 million metric tons in 1989, a 47% decline since 1986 (Holmes et al., 1992). This situation has encouraged the use of other species such as whiting, cod and New Zealand hoki. Roe herring (Clupea pallasi) is a species that is abundant on the coast of British Columbia, Canada. Its annual harvest is between 15,000 and 40,000 tons. The catch is almost exclusively aimed at collecting herring roe, a very 75 Chapter 4. Production of salmon and herring kamaboko popular commodity in Japan. The male fish and the remainder of the female fish, after roe collection, are either wasted or reduced to fish meal. Making surimi or a related commercial product from the leftover herring could be a good alternative. Kamaboko of average quality was obtained from fresh North Sea herring (Hastings et al., 1990). The quality of kamaboko was worse when North Sea herring was frozen at -30°C for six months. Nakai (1993) found that no kamaboko could be made from herring surimi that was frozen at -35°C for three months. In this study, an attempt was made to produce acceptable surimi-based products from herring that was frozen at different conditions in the form of raw fish. However, due to the limited availability of roe herring and because its small size makes surimi production difficult, another species was also tried. Pink salmon (Oncorynchus gorbuscha) is very abundant on the British Columbian coast and is mainly used for canning. It is characterized by its large size and relatively low cost. Therefore, Pink salmon was also utilized to produce surimi and kamaboko. When washed fish mince is the only component of a surimi-based product, the texture tends to be rubbery and freeze-thaw stability is often poor (Lee et al., 1992). To improve the texture of the final product, several ingredients can be added. In this study, the ingredients used were salt, whey protein 76 Chapter 4. P roduc t i on o f salmon and herring kamaboko concentrate and wheat starch. Salt was used to improve the solubility of surimi myofibrillar proteins. Kim and Lee (1987) explained that starch is mainly used as a filler ingredient. Upon heating, it swells and increases water uptake. This improves the firmness and the cohesiveness of the surimi gel matrix. Wu et al. (1985) reported that starch does not directly interact with the surimi protein matrix nor significantly influence its formation. Several studies have reported the use of whey protein concentrate in the manufacture of surimi-based products (Burgarella et al., 1985; Chung, 1990; Lee and Kim, 1986). It is also used as a filler. Its gel strengthening effect has been reported by Lee et al. (1992b). Thus, there was a need to investigate the combined effects of salt, wheat starch and whey protein concentrate on kamaboko made from pink salmon and roe herring. 4.2 Objectives a. To study the effect of frozen storage conditions on the kamaboko making ability of roe herring. b. To optimize a formulation consisting of surimi, salt, whey protein concentrate and wheat starch for kamaboko made from pink salmon surimi and from roe herring. 77 Chapter 4. Production of salmon and herring kamaboko 4.3 Materials and methods 4.3.1 Materials Frozen, whole, roe herring was purchased from a local fishing company (Sung Fish Company, Vancouver, B. C.) on May 17, 1994. The company kept this fish frozen at -40°C for approximately one month (since mid-April 1994). It was transported to our laboratory where 11 pounds (lb) were stored at -83°C, 70 lb at -45°C and 70 lb at -35°C. The fish comprising the last sample were soaked in glycerol (10% solution) for 24 hours before further frozen storage. The cryoprotective effect of glycerol has been reported by Matsumoto and Noguchi (1992). Due to the limited quantity of herring in our possession, pink salmon was selected for the tests relating to the optimization of the kamaboko formulation. Pink salmon is suitable for this type of study because the surimi preparation from it is less labor intensive and because there is also a commercial incentive to produce surimi from this species. It was therefore intended to optimize the kamaboko formulation with salmon surimi and to use the resulting optimum conditions to prepare roe herring kamaboko. Fresh pink salmon was purchased at the end of July 1994. It was kept frozen at -35°C for 78 Chapter 4. P roduc t i on o f salmon and herring kamaboko two months before salmon surimi was made from it. After four months of storage at -35°C, the salmon surimi was then used in the optimization experiments. To improve the texture of salmon kamaboko, several ingredients were used. These are salt, modified whey protein concentrate (WPC) (Alaco Surimi Plus), wheat starch (WS) and water. The ingredient combinations and their respective concentrations are listed in Table 17. Two Japanese surimi-based products (Maruu Kamaboko Shiro and Kibun Hanpen-L), which were made from Alaska pollock, were used as commercial target products. 4.3.2 General procedures 4.3.2.1 Surimi making The fish was filleted, minced, washed 4 times in a 0.3% salt solution (the volume of the solution was five times the volume of the fillets), and mixed with cryoprotectants [4% sucrose (meat weight basis), 4% sorbitol, 0.15% tripolyphosphate and 0.15% sodium pyrophosphate] in a silent cutter (Hobart Manufacturing Co., Model 84142, Don Mills, Ontario) for 5 minutes. These procedures were performed at 4°C. The surimi was stored frozen at -35°C. 79 Chapter 4. Production of salmon and herring kamaboko 4.3.2.2 Kamaboko making Surimi, partially thawed at room temperature for one hour, was chopped, in a vacuum cutter (Stephan, Model VCM-5, Columbus, Ohio), at low speed for two minutes. Then 3% salt (total weight basis) was added and the mixture was mixed at low speed for three minutes. Finally, mixing was performed under vacuum (to prevent oxidation) at high speed for one minute. Circulating ice chilled water in a jacket allowed the mixing to be carried out at temperatures below 10°C as recommended by several researchers (i.e., Lee et al., 1992a; Douglas-Schwartz and Lee, 1988; Lanier, 1992). When WS and WPC were used, the mixing procedure was different. Surimi chopped at low speed for two minutes, was mixed with salt at low speed for two additional minutes. The paste was then blended with WS (premixed with water) and WPC for two more minutes at low speed. The paste was finally mixed at high speed and under vacuum for one minute. Using a meat stuffer (Kitchenaid, Model K5-A, Troy, Ohio) the paste was stuffed into sausage casings having a diameter of 30 mm and a length of 140 mm. These sausages were then subjected to setting at 45°C for thirty minutes in a Blue M water bath (Model MW-1120A-1, Blue Island, IL) and then cooking at 90°C for half an hour (in a second Blue M water bath). The end product, kamaboko sausages, were immediately cooled in ice for fifteen minutes and kept at room temperature overnight before testing. The setting and cooking conditions were selected in 80 Chapter 4. Production of salmon and herring kamaboko light of previous experiments aimed at optimizing the processing conditions for kamaboko made from fresh herring (Nishimura et al., 1993). 4.3.2.3 Measurement of surimi gel properties The quality of the kamaboko sausages was tested by performing the punch and die test and the fold test. The punch test provides an indication of the gel firmness, while the fold test measures the elasticity of the product (Lanier, 1991). The punch tests were performed using an Instron Universal Testing Instrument (Model 1122, Canton, Massachusetts). A surimi gel cylinder, having a diameter of 30 mm and a height of 20 mm, was centered under a plunger (5 mm diameter). The plunger goes through the sample and then through an orifice (8 mm diameter). The cross head speed was 10 mm / min. The variation of the penetration force with time was recorded. The gel strength is the product of the force and the distance at the point of rupture (see Figure 4). The fold test was carried out according to the description given in section 2.2.3. 4.3.3 Roe herring samples On May 18,1994 surimi was made from about 30 pounds of whole herring (one day after its purchase). The surimi was stored frozen at -35°C. On the 81 Chapter 4. Production of salmon and herring kamaboko following day, kamaboko was made from this surimi and its quality was evaluated by performing the punch and the fold tests. Surimi and kamaboko were made from herring stored at -35°C and at -45°C for forty days. Fish stored at the latter temperatures for 130 days was used to produce surimi and kamaboko. Finally, surimi and kamaboko were made from frozen fish stored at -35°C, at -45°C and at -83°C for 190 days. 4.4 Results and discussion 4.4.1 Frozen storage experiments After 40 days of storage at -35°C the glycerol soaked roe herring gave a punch test score of 51 N.mm and a fold score of 3. The sample which was stored at -45°C for the same period of time resulted in a punch test score of 61 N.mm and a fold score of 3. By comparison, kamaboko made from frozen herring before the cold storage was initiated (i.e., at the time when the frozen herring was purchased from Sung Fish Co.) gave a punch test score of 61 N.mm and a fold test value of 3. From these results it appears that the fish stored at -45°C without the use of glycerol gave the same gel strength and fold test values as the sample before prolonged frozen storage. These results 82 Chapter 4. P roduc t ion o f salmon and herring kamaboko are similar to the findings of Fukuda (1991) in which he concluded that freeze damage of surimi made from cod, salmon, sardine and mackerel and frozen for six months could be reduced at cold storage temperatures lower than -AO°C. However, Figure 13 shows that after 130 days of storage, the sample stored at -45°C had a gel strength of 43 N.mm and a fold score of 2 while the sample soaked in glycerol and stored at -35°C had a gel strength of 37 N.mm and a fold score of 2. Thus, a notable decrease of gel strength occurred between 40 and 130 days of frozen storage. After 190 days of frozen storage, the sample stored at -83°C had a gel strength of 57 N.mm and a fold score of 3. This sample had maintained reasonable gel quality. By comparison, the gel strength values of the samples stored at -45°C and at -35°C for 190 days decreased further to 36 and 33 N.mm, respectively. These two samples had a fold score of 2. Hastings et al. (1990) also found that the frozen storage of whole herring for six months at -30°C resulted in a kamaboko having a gel strength about half of that of kamaboko made from fresh herring. The deterioration of the gel making ability of roe herring after prolonged frozen storage could be due to the denaturation of myofibrillar proteins, a phenomenon suggested by many researchers (i.e., Watabe and Hashimito, 83 Chapter 4. Production of salmon and herring kamaboko 70 - x — 3 5 C + glycerol —I 45 C -83 C 20 4-10 4-50 100 Time [days] 150 200 Figure 13. Variation of roe herring kamaboko gel strength with frozen storage 84 Chapter 4. Production of salmon and herring kamaboko 1987; Matsumuoto and Noguchi, 1992). This denaturation seems to be prevented by frozen storage at ~45°C up to forty days. Frozen storage at -83°C for 190 days maintained the gel firmness and elasticity. These findings are similar to those of Watabe and Hashimito (1987) who reported a decrease in the myofibrillar protein fraction caUsed by frozen storage of Alaska pollock surimi at -20°C and at -30°C. This decrease, however, did not occur for the storage temperatures of -40 and -80°C. These researchers concluded that frozen storage at -40°C or lower temperatures prevented protein denaturation and maintained the gel making ability of Alaska pollock surimi. Because of the complexity of the mechanisms responsible for the deterioration of the gel making ability of frozen fish muscle, it is recommended, for future studies, that Raman spectroscopy be used to investigate the changes that take place during the freeze denaturation of fish proteins. 4.4.2 Optimization experiments The results of the formulation optimization experiments are shown in Table 17. Figure 14 shows the variation of gel strength with whey protein concentrate and wheat starch. The combined effects of the latter ingredients 85 Chapter 4. Production of salmon and herring kamaboko Table 17. Effects of different formulations on surimi gel texture %WPC %WS % Water Gel Strength Fold % Moisture # Starting Material added (N.mm) Score 1 SSS 0 0 0.00 58.9*+ 5.5" 3 71.8 2 SS 2 3 2.24 65.4±14.7 4 69.6 3 SS° 2 5 7.08 79.6+12.9 6 69.2 4 SS 2 7 11.92 68.3±9.6 6 70:2 5 SS 4 3 7.80 86.2±7.7 4 70.4 6 SS 4 5 12.64 61.7±7.3 6 69.6 7 SS 4 7 17.49 45.6+4.5 6 69.3 8 SS 6 3 13.36 53.6±7.7 4 69.8 9 SS 6 5 18.20 50.5±6.1 5 70.1 10 SS 6 7 23.00 51.3+6.6 6 69.3 11 SS 4 0 0.00 72.2+6.2 3 69.4 12 SS 0 5 0.00 64.7±5.4 4 69.3 13 SS 2 5 0.00 104.7±7.8 4 66.7 14 HSh 2 5 7.00 74.4* ±3.3 6 NM" 15 MKSm 99.7*+8.6 6 74.2 16 KH-L k 84.2* ±2.3 6 71.5 8 Salmon surimi plus 3 % salt h Herring surimi plus 3 % salt 0 Optimal formulation mMaruu Kamaboko Shiro kKibun Hanpen-L (Boiled fish cake) " Not measured * Arithmetic mean (n=5-7) * Arithmetic mean (n=3) ** Standard deviation 86 Chapter 4. Production of salmon and herring kamaboko Figure 14. Combined effects of WPC and WS on salmon surimi gel strength 87 Chapter 4. Production of salmon and herring kamaboko on the fold score are depicted in Figure 15. Formulation 5, consisting of 82.2% salmon surimi (total weight basis), 3% salt, 4% WPC, 3% WS and 7.8% water, gave a gel strength of 86.2 N.mm and a fold score of 4. Formulation 3 composed of 83% salmon surimi (total weight basis), 3% salt, 2% WPC, 5% WS and 7% water resulted in a gel strength of 79.6 N.mm and a fold score of 6. Since the objective was to maximize both the gel strength and the fold score, formulation 3 was chosen to be the optimal formulation. The characteristics of the gel obtained from this optimal formulation are very close to those of commercial Kibun Hanpen-L (Boiled fish cake) (see Table 17). However, the gel strength of kamaboko produced from this optimal formulation was 20% lower than that of Maruu Kamaboko Shiro, the second target product. When compared to a control formulation (salmon surimi plus 3% salt), the optimal formulation caused a 35% increase in gel strength. It also increased the fold score from 3 to 6. Formulations having a 6% WPC consistently gave weak gels. Moreover, formulations consisting of 5% and 7% WS resulted in better fold score values than those containing 3% WS. When the cumulative proportions of WPC and WS reached 11 % or higher (i.e., formulations 7, 9 and 10), the corresponding fold score values were very good but the gel strength values were low. The individual gel texture improving effects of WPC and WS are demonstrated by formulations 11 and 12. These results confirm the findings of Kim and Lee (1987) who reported the gel strengthening effect of starch and Lee et al. (1992b) regarding the 88 Chapter 4. Production of salmon and herring kamaboko Figure 15. Combined effects of WPC and WS on salmon surimi gel fold score. 89 Chapter 4. P roduc t ion o f salmon and herring kamaboko gel strengthening effect of W P C . The amount of added water was calculated in a way that the final product moisture content had to be around 70% (wet basis). The moisture contents of the ingredients used are shown in Table 18. Although the addition of water decreased the gel strength, it improved the product elasticity. The optimal formulation (formulation 3) which was developed for surimi was applied to surimi made from fresh herring. The product had a gel strength of 74.4 N.mm and a fold score of 6. These gel quality values are close to those of Kibun Hanpen-L which is the lower quality target product. However, because the composition of roe herring (i.e., fat content, protein content, proportion of dark muscle) is different from that of pink salmon, the optimal formulation for salmon is not necessarily the optimal formulation for herring. Therefore, more research should be conducted to find the optimal formulation for herring kamaboko. 90 Chapter 4. P roduc t ion o f salmon and herring kamaboko Table 18. Moisture content of different products Sample % Moisture Salmon surimi 74.9 Herring surimi 73.6 WPC 5.1 WS 14.1 Salt 0.2 91 Chapter 4. P roduc t ion o f salmon and herring kamaboko 4.5 Conclusions Kamaboko made from an optimized formulation consisting of 83% pink salmon surimi (total weight basis), 3% salt, 2% WPC, 5% WS and 7% water gave gel strength and fold score values close to those of Kibun Hanpen-L (Boiled fish cake), which was used as a target product. When this formulation was applied to surimi made from fresh herring, similar gel characteristics were obtained. Roe herring frozen at -40°C for one month and then at -45°C for forty days maintained its gel making ability. Further storage for ninety more days resulted in surimi gels of lower quality. This could be due to protein denaturation due to frozen storage. However, when roe herring was frozen at -40°C for one month and then at -83°C for 190 days, the gel making ability was preserved. Pink salmon seems to have a great potential to be used for surimi-based products. Using pink salmon to make crab analogs, for instance, could be more profitable than using it for canning. Because of its darker color, roe herring could be successfully used in surimi-based products that do not require a white color. The problem of its small size can be overcome by using 9 2 Chapter 4. Production of salmon and herring kamaboko equipment that is designed for producing surimi from under-sized fish species such as sardines. Using male roe herring to make value-added surimi-based products is a better alternative than simply reducing it to fish meal. 93 Chapter 5. Sur imi viscous properties Chapter 5 Viscous Properties of Salmon Surimi Paste 5.1 Introduction The viscous properties of surimi are among the physical properties that need to be thoroughly studied before new and improved processes for surimi-based products can be developed. These flow properties of surimi pastes are of particular importance during the pumping and the extruding operations that occur during food processing. Several researchers have studied the viscous properties of actomyosin solutions. Nakayama et al. (1979) investigated the effects of incubation at 40°C and of the additions of Na4P207 and MgCI2 on the viscosity of carp actomyosin. The viscosity changes of croaker actomyosin solutions during gelation were studied by Wu et al. (1985). They found that viscosity decreased with temperature up to 31 °C. It then increased quickly, peaking at 36°C and then rapidly decreased. Owusu-Ansah et al. (1988) also observed 94 Chapter 5. Surimi viscous properties similar trends of viscosity variation with temperature for sucker myosin. They found that viscosity peaked at 43°C. Jimenez-Colmenero and Borderias (1983) and Borderias et al. (1985) stated that the apparent viscosity of proteins from frozen fish isolated in high ionic strength solutions can be used as a quality index for frozen fish proteins. Groninger et al. (1983) used viscosity measurement to assess changes in processed frozen fish muscle. A test for the viscosity of surimi diluted (599% by weight) in a 3.5% salt solution was used as a Japanese standard test for the quality evaluation of frozen surimi protein (Hamann and MacDonald, 1992). The apparatus to be used for this test is a Brookfield viscometer (Tokyo Keiki Type C). Scott et al. (1988) found that the viscosity of Alaska pollock surimi, diluted in accordance with the Japanese standard test, decreased with frozen storage time of the headed and gutted Alaska pollock up to 128 days. The effects of incubation time and temperature on the viscosity of a surimi paste (30% surimi, 2% salt, 1.2% beef plasma protein and 66.8% water) were studied by Park et al. (1994). The viscosity of the paste incubated at 5°C remained close to 40 Pas. When the paste was incubated at 25°C, the viscosity increased to 80 Pa.s after 20 hours of incubation. The viscosity of the paste incubated at 40°C for 30 minutes increased to 65 Pa.s. However, the viscosity of the paste incubated at 60°C for 30 minutes decreased to 10 Pa.s. The viscosity increases at 25 and 40°C were attributed to intramolecular binding of myofibrillar proteins whereas the 95 Chapter 5. Surimi viscous properties viscosity decrease at 60°C was thought to be due to proteolytic degradation of the myofibrillar proteins. In the manufacturing of surimi-based products, surimi is mixed with ingredients (i.e., salt, starch, etc.) in a silent cutter. The resulting viscous paste is then pumped before being cooked and extruded (Wu, 1992). Hence, it is important to study the viscous properties of surimi in the paste form. The objective of this work was to study the viscous properties of a salmon surimi paste having an optimum formulation for kamaboko making. The results should help to clarify the flow behavior of this minced fish muscle. 5.2 Materials and methods The surimi used was the same salmon surimi described in Chapter 4. The optimal formulation, which was selected in section 4.4.2, was used to make the surimi paste for the Theological studies. As in section 4.3.2.2, partially thawed surimi was cut in a vacuum cutter (Stephan, Model VCM-5, Columbus, Ohio) at low speed for two minutes. Then 3% salt (total weight basis) was added and the mixture was chopped at low speed for two minutes. Then 5% 96 Chapter 5. Sur imi viscous properties wheat starch, premixed with 7% water, and 2% whey protein concentrate (Alaco Surimi Plus) were added. This order was recommended by Wu (1992). The ingredients were blended at low speed for two minutes. Finally, mixing was performed under vacuum and at high speed for one minute. There are several methods for the measurement of viscosity. The glass capillary method is based on the measurement of the time for a known volume of fluid to pass through a length of capillary tubing. The orifice viscometers are simple instruments for quick measurements. The method consists of measuring the time needed for a volume of fluid to flow through an orifice. This method, however, is less accurate than the glass capillary method (Bourne, 1982). Other instruments include the capillary viscometer, the falling ball viscometer and the coaxial rotational viscometers. The latter type of viscometer is very popular because it allows the study of the time-dependent Theological behavior of the fluid, permits the continuous variation of the shear rate, can handle Newtonian and non-Newtonian fluids and can be equipped with a temperature control feature. For these reasons, the viscous properties of the salmon surimi paste (kept in ice) were studied using a rotational viscometer (Haake Rotovisko, Model VT500, Haake, Germany). A sensor with a profiled cup and spindle, SVP, was used to reduce slippage. The inner cylinder (rotor) has a radius of 10.65 mm and a height of 31.45 mm, while the outer cylinder has a radius of 11.55 mm. Experiments were carried out at five 97 Chapter 5. Sur imi v iscous properties temperatures; 1,6, 11, 21 and 26°C. For each temperature, the paste was put in the cup and kept there for ten minutes to come to temperature. Then, the sample was sheared at shear rates increasing linearly from 0 to 400 s"1 in ten minutes. After that, the shear rates decreased from 400 to 0 s"1 in ten minutes. Finally, the sample was sheared again for ten minutes at rates increasing from 0 to 400 s'\ Experiments were carried out in duplicate. A personal computer to which the viscometer was interfaced was used to vary shear rates and to record the corresponding shear stress values. The Systat 1990 (Systat, Inc., Evanston, IL) software was used for the regression analysis. 5.3 Results and discussion Figures 16 to 20 are surimi paste rheograms at temperatures of 1, 6, 11, 21 and 26°C, respectively. Each figure shows an increasing shear rate (ISR) rheogram, a decreasing shear rate (DSR) rheogram and a second ISR rheogram. For all the temperatures, the first ISR rheogram was not smooth or reproducible. This could be due to the fact that the paste was initially at rest or that the structure was being ruptured by the applied shear, possibly 98 Chapter 5. Surimi viscous properties 600 550 4-500 450 400 350 4-£. 300 (0 250 4-200 4-150 4-100 50 • Increase D, 1st • Decrease D A Increase D, 2nd T = 1 C •* • • • • • • • • • • A .OS A A A A A A A N A A A A n A * A * ° A A • A • A O A a n° A • A nD A • A C D H h 0 50 100 150 200 250 300 350 400 D [1/s] Figure 16. Rheograms of salmon surimi paste at 1°C 99 Chapter 5. Surimi viscous properties 650 600 550 500 450 400 „ 350 10 Q. (0 300 250 200 150 + 100 50 + Increase D, 1st a Decrease D A Increase D, 2nd T = 6C • • • • • ^ —6' • ^ r W 3 - 0 ° .^AA" A6 k • AAAA 2A A AD • A • •+-0 50 100 150 200 250 300 350 400 D [1/s] Figure 17. Rheograms of salmon surimi paste at 6°C 100 Chapter 5. Surimi viscous properties CO (0 650 600 550 500 450 400 350 300 250 4-200 150 100 4-50 • Increase D, 1st • Decrease D A Increase D, 2nd T = 11 C •o , f i o n A A A A A A A A A A A A A A A D A A A " A - A - A A -DDnO AA A A • _ D 2 a A A A A • • • A A A D • • A 0 a 0 ° O A 8A + 0 50 100 150 200 250 300 350 400 D [1/s] Figure 18. Rheograms of salmon surimi paste at 11°C 101 Chapter 5. Surimi viscous properties 650 600 550 500 450 + • Increase D, 1st • Decrease D A Increase D, 2nd • • • A* U •A A • ¥ * A a D = • A A A A A AA A A A D d O A A ^ A A AA HO A A A H D A A^ (0 0. (0 400 350 300 250 + 200 + 150 100 50 A " A ° + + 50 100 150 200 250 300 350 D [1/s] 400 Figure 19. Rheograms of salmon surimi paste at 21°C 102 Chapter 5. Sur imi viscous properties 500 450 400 350 300 co <L 250 W 200 150 4-100 50 04 • Increase D, 1st • Decrease D A Increase D, 2nd T = 2 6 C • • • • • • • nrf 3 A A A A A n 0 0 A A A A A A A A D D D A A A A A A A A D A A A A CP • " A A ° A A A ° A A & V A A A A • H h 0 50 100 150 200 250 300 350 400 D [1/s] Figure 20. Rheograms of salmon surimi paste at 26°C 103 Chapter 5. Surimi viscous properties causing oscillations. After ten minutes of shearing, however, the DSR rheogram and the second ISR rheogram became relatively smooth and monotonic. These rheograms show that shear stress did not vary linearly with shear rate. Therefore, surimi paste had a non-Newtonian behavior. Moreover, a yield stress [shear stress at zero shear rate, or the shear stress required for the material to flow (Tung, 1989)] was observed. This yield stress increased with increasing temperature (for T < 21 °C). For all the temperatures and for shear rates lower than 220 s"1, the DSR rheograms and the second ISR rheograms were very similar. However, at temperatures of 11°C and above and for shear rates higher than 220 s'1, the shear stress values of the second ISR rheograms became considerably lower than those of the DSR rheograms. This hysteresis could be due to the loss of structure of the paste that could occur at higher temperatures and especially after shearing at higher rates for more than 25 minutes. It can be deduced that this surimi paste behaved as a shear thinning fluid with a yield stress. The existence of hysteresis despite the low ramp speed (40 s"1 / min) that was used to increase and decrease the shear rate, demonstrates that the paste rheology was time dependent and had a slow recovery. Park et al. (1994) found that the viscosity of a Pacific whiting surimi paste started to increase after 20 hours of incubation at 25°C, and did not increase for an incubation temperature of 5°C. This is thought to be due to the 104 Chapter 5. Surimi viscous properties relatively low concentration of surimi (30% by weight) in the paste they studied. (The surimi concentration of the paste used in this project is 83% by weight.) The DSR rheograms for the five temperatures used are shown in Figure 21 while the second ISR rheograms are depicted in Figure 22. Both of these figures demonstrate that for the same shear rate, provided D < 200 s"1, shear stress values increased with temperature up to 21 °C. This is thought to be caused by the protein-protein interactions which increased with temperature. As suggested by Douglas-Schwartz and Lee (1988), actomyosin from salmon could be less thermally stable than actomyosin from species living in warmer waters. This increase of shear stress with temperature differs from the results of Wu et al. (1985) who found that the viscosity of croaker actomyosin solutions decreased with temperature up to 31 °C. This could be due to the high concentration of surimi used in this study. Shear stress values at 26°C were the lowest. This is possibly due to the fact that the paste started to gel and the shear stress readings could have been influenced by slippage or by the rupture of gel structure near the cup wall. It was attempted to run the experiment at 31 °C but the spindle could not rotate because the required torque was above the operating range for the rheometer. The gelling effect on the sample was apparent when it was taken from the cup. 105 Chapter 5. Surimi viscous properties 650 600 550 500 450 4-400 350 co CL CO 300 | A • 250 4-100 4-50 4-0 4 Decreasing D m — W MT~ . A A wX* • A A * * A A A A wXX* • X * * * * * A N w * A A A A no 0 X * A A • X * A • X A A • • A A •P •••• X A A A D A • * A x A n ° o • ••• • • • • T= 1 C 200 - • • T = 6 C •• • AT= 11 C 150 -• XT = 21 C - • • T = 26 C + + 50 100 150 200 D [1/s] 250 300 350 400 Figure 21. DSR rheograms 106 Chapter 5. Surimi viscous properties 600 550 500 450 4-400 Increasing D * X X * A A A A A A * A A £ : W a * * ° T * * * X X A A A x B»H 8 D » r V * \ X X A A A A A B p D • # j ; w X . . A • • ^ • % x Xx* * A A ° ¥ •••••• * \ 350 ra £L 300 co x x>A A A X X . A A • » X A A D D ™ / • A • • • * A D D • * 0*. • •••D**W X A A 250 + 200 150 + 100 4-50 + • • • • •••• •••• • T = 1 C • T = 6 C A T = 11 C X T = 21 C • T = 26 C + -+• + 50 100 150 200 250 D [1/s] 300 350 400 Figure 22. Second ISR rheograms 107 Chapter 5. Surimi viscous properties Since the first ISR rheograms were affected by the time dependency of the paste, only the yield stress values were calculated for these curves. The Theological data for the DSR rheograms and the second ISR rheograms were fitted by the Herschel-Bulkley model and by the Casson model. (Although only half of the results are shown, all the Theological data including the replicate results were used in the curve fitting.) These models were explained by Bourne (1982). The appropriate model should represent the flow behavior of the fluid over the shear rate range to which the fluid will be subjected during processing (i.e., pumping, mixing, etc.). It should also accurately fit the experimental data. The following is a brief description of the two models which were found to best represent the data. Herschel-Bulkley (HB) model: This model is also called the power-law plastic model. It is different from the standard power-law model in that it includes a yield stress term. It is therefore suitable for fluids having a yield stress. The equation of this model is: S = S0 + KD n 108 Chapter 5. Surimi viscous properties where S is the shear stress, K is a consistency coefficient, D is the shear rate, n is the flow behavior index and S0 is the yield stress. The parameters of this model can be determined by non-linear regression based on the least squares method. Casson model: This model was originally developed to simulate the flow of printing inks (Bourne, 1982). It is appropriate for some foods with a yield stress, such as tomato catsup and molten chocolate (Bourne, 1982). The equation for this model is: S 0 5 = So0 5 + (Kc D) 0 5 where Kc is a parameter. The parameters of this model can be determined by first plotting the square root of shear stress against the square root of shear rate and then finding the equation of the straight line which provides the best fit of the data points. Table 19 shows the curve fitting results of the second ISR rheograms. The data were well fitted by both models (R2 > 0.94). The values of n, the flow behavior index in the HB model, were close to 0.5. For both models, yield 109 Chapter 5. Sur imi v iscous properties Table 19. Curve fitting results of the second ISR rheograms HB model T[C] R2 So K n Note 1 0.963 219.235 11.822 0.534 6 0.988 253.365 8.302 0.597 11 0.948 309.752 12.65 0.505 21 0.957 319.549 17.117 0.477 D<220 26 0.970 134.562 8.137 0.513 Casson model T[C] R2 So Kc 1 0.956 215.591 0.164 6 0.989 245.365 0.153 11 0.944 311.571 0.102 21 0.954 333.741 0.119 D<220 26 0.955 137.223 0.092 110 Chapter 5. Sur imi v iscous properties stress values increased with temperature up to 21 °C. The yield stress values at 26°C were the lowest. These trends were also observed for the curve fitting results of the DSR rheograms (Table 20). These rheograms were also very well fitted by the HB model and by the Casson model (R2 > 0.96). The curve fitting results of the DSR rheograms were slightly better than those of the second ISR rheograms. Figures 23 to 27 show the DSR rheograms along with the predictions of the fitted models for the five temperatures used. The curve fitting results obtained with the HB model demonstrate the existence of a yield stress that increased with temperature up to 21 °C. The yield stress values of the second ISR rheograms were higher than those of DSR rheograms. The yield stress values, S0 [Pa], predicted by the HB model for various temperatures, T [°C], was fitted by the following equation: So = C y T m where C y and m are the equation parameters, and T < 21 °C. A logarithmic transformation of this equation yields the following linearized form: 111 Chapter 5. Surimi viscous properties Table 20. Curve fitting results of the DSR rheograms HB model T[C] R2 So K n 1 0.997 183.145 10.021 0.576 6 0.995 217.791 14.381 0.548 11 0.994 264.433 13.383 0.562 21 0.990 285.473 19.03 0.478 26 0.965 129.023 3.336 0.751 Casson model T[C] R* So Kc 1 0.996 184.311 0.196 6 0.994 227.624 0.228 11 0.994 267.594 0.215 21 0.988 310.236 0.137 26 0.972 103.338 0.274 112 Chapter 5. Sur imi v iscous properties 150 4-100 4-50 4-0 4 1 1 1 1 1 1 1 1 0 50 100 150 200 250 300 350 400 D [1/s] Figure 23. DSR rheograms at 1°C 113 Chapter 5. Surimi viscous properties 114 Chapter 5. Surimi viscous properties 700 650 600 550 500 450 + 400 n SL 350 (0 300 | D 250 200 4/ 150 100 50 • Deacrease D H.B. model Casson model T = 11 C — i — 200 D [1/s] 50 100 150 250 300 350 400 Figure 25. DSR rheograms at 11°C 115 Chapter 5. Surimi viscous properties 450 + 400 + 200 + 150 4-• Decrease D H.B. model Casson model T = 21 C •+-50 100 150 200 D [1/s] 250 300 350 400 Figure 26. DSR rheograms at 21 °C 116 Chapter 5. Surimi viscous properties 50 4-0 -I 1 1 1 1 1 1 1 1 0 50 100 150 200 250 300 350 400 D [1/s] Figure 27. DSR rheograms at 26°C 117 Chapter 5. Sur imi viscous properties log(So) = log(Cy) + m log(T) The results obtained by fitting the above logarithmic equation to the S0 values predicted by the HB model for both the second ISR and the DSR rheograms are listed in Table 21. The fit was good for both the second ISR and the DSR rheograms. The latter rheograms resulted in a better fit (R2> 0.90). However, for the pumping of surimi paste, the yield stress value that must be overcome at the beginning of the process is the yield stress determined at the first increasing shear rate rheogram. Table 22 lists these yield stress values for different temperatures. These were the recorded shear stress values corresponding to the lowest shear rate, 8.97 s~1. Generally, yield stress increased with increasing temperature up to 21 °C. The lowest yield stress value was observed at 26°C. Moreover, for each temperature the initial yield stress was much higher than that of the DSR rheograms and the second ISR rheograms. The Herschel-Bulkley and Casson models depend only on the shear rate. They do not account for the influence of temperature. The viscosity of Newtonian fluids is found to vary with temperature according to the following equation (Morgan et al., 1989): 118 Chapter 5. Surimi viscous properties Table 21. Results of the curve fitting of the yield stress Rheograms Cy m R2 Second ISR 215.406 0.130 0.874 DSR 179.205 0.148 0.915 119 Chapter 5. Sur imi viscous properties Table 22. Yield stress values from the first ISR rheograms Temperature, T (°C) Initial yield stress, S 0 (Pa) 1 371 6 462 11 457 21 522 26 329 120 Chapter 5. Sur imi viscous properties ri = C T exp (AEV / RT) where r| is the viscosity, T is the absolute temperature, AE V is the molar free energy of activation in a stationary fluid, R is the universal gas constant and C T is a constant. A new model, taking into consideration both the shear rate effect and the temperature effect, is proposed. The equation of this model is assumed to be: S = e D f + g T h where S is the shear stress in Pa, T is the temperature in °C, D is the shear rate in s"1 and e, f, g and h are the model parameters. The form of this equation derives from the fact that only the yield stress appears, in most experiments, to be a strong function of temperature. Usually, the viscosity of most fluids decreases with temperature (Streeter and Wylie, 1985). The system used in this study, however, was likely affected by temperature because of changes in protein-protein interactions. For temperatures ranging from 1 to 21°C, the DSR rheograms and the second ISR rheograms were fitted by the above 'shear and temperature (ST) model" Table 23 summarizes the curve fitting results. The model parameters show 121 Chapter 5. Surimi viscous properties Table 23. Curve fitting results of Theological data by the ST model Rheograms e f g h R DSR 9.945 0.590 178.382 0.185 0.968 Second ISR 96.915 0.240 75.893 0.251 0.903 122 Chapter 5. Surimi viscous properties that shear stress increases with increasing temperature and shear rate. Figure 28 shows the DSR rheograms as well as the predictions of the fitted ST model. The second ISR rheograms along with the predictions of the ST model are depicted in Figure 29. Curve fitting by the ST model was better for the DSR rheograms (R2 > 0.96). The curve fitting was also better for shear rates from 0 to 200 s"1. The time dependency of the paste was mainly present after shearing for 25 minutes. Therefore, the ST model should be applicable for typical residence times of the paste as it is pumped between the mixing tank and the extruder. 123 Chapter 5. Surimi viscous properties 150 4-100 4-50 4-0 -1 1 1 1 1 1 1 1 1 0 50 100 150 200 250 300 350 400 D [1/s] Figure 28. DSR rheograms and the predictions of the ST model 124 Chapter 5. Surimi viscous properties 600 50 100 150 200 D [1/s] 250 300 350 400 Figure 29. Second ISR rheograms and the predictions of the ST model 125 Chapter 5. Sur imi viscous properties 5.4 Conclusions The viscous properties of a salmon surimi paste composed of surimi, salt, whey protein concentrate and wheat starch were studied using a rotational viscometer. The paste behaved as a time dependent non-Newtonian shear thinning fluid with a yield stress. For the same shear rate, shear stress increased with temperature up to 21 °C. This could be caused by the protein-protein interactions that were facilitated by grinding the surimi with salt. The yield stress values also increased with temperature up to 21 °C. The DSR rheograms and the second ISR rheograms, after surimi had undergone shear for sufficient time, were well fitted by the Herschel-Bulkley model and by the Casson model. A new model taking into account the shear rate effect and the temperature effect has been developed. For temperatures up to 21 °C, the DSR rheograms and the second ISR rheograms were well fitted by this new model. Predicted shear stress increased with increasing temperature and shear rate. Finally, it is important to mention that the paste used in this study was similar to what is used in the commercial manufacture of surimi-based products (i.e., the paste was not first diluted with water as in most previous Theological 126 Chapter 5. Surimi viscous properties studies). The Theological information obtained should be useful for the engineering design of the equipment needed for the pumping and extruding operations that immediately follow the mixing process. However, the results obtained here apply only for a salmon surimi paste having a specific formulation and for temperature ranging between 1 an 26°C. A similar approach can be followed to determine the rheological properties of other surimi pastes. 127 Chapter 6. Conclusions and future work Chapter 6 Summary of Conclusions and Proposed Future Work 6.1 Conclusions Raman spectroscopy was successfully used to investigate the effects of grinding with salt as well as setting and cooking on the protein structure of raw surimi, surimi paste and surimi gels. The disulfide bond stretching intensity increased slightly after setting and greatly after cooking at higher temperature. These results explain the importance of the disulfide bonds and the stages at which they occur. Secondary structure analysis showed that setting caused a decrease in the a-helical content and an increase in the p-sheet content. Cooking alone or preceded by setting resulted in a greater increase of the p-sheet fraction and a further decrease of the a-helical content. These findings suggest an unfolding of proteins allowing hydrophobic interactions between exposed hydrophobic residues to take place. Studies of gel properties also showed that cooking, whether preceded by setting or not, resulted in a better surimi gel texture than the setting 128 Chapter 6. Conclusions and future work treatment alone. In comparison with cooking alone, setting followed by cooking resulted in a higher disulfide peak intensity and in stronger and more elastic surimi gels. Experiments on optimizing the setting and cooking conditions for kamaboko making showed that setting at temperatures below 50°C followed by cooking at 80°C and above resulted in surimi gels with the best gel texture characteristics. The optimal processing conditions corresponding to individually maximized gel strength, fold score and color, and minimized expressible liquid were determined. Multiobjective optimization also gave the optimal processing conditions resulting in a satisfactory combination of all four response variables. An optimized formulation composed of pink salmon surimi, salt, whey protein concentrate and wheat starch has been developed. The final product, kamaboko sausage, had gel strength and fold score values close to those of a commercial surimi-based product. This formulation was successfully applied to roe herring surimi. Freezing roe herring at around -45°C maintained the gel making ability for up to seventy days. Freezing at -83°C preserved the gel making ability for a longer period of time. However, this temperature is too low to be used by the industry. Using male herring for surimi-based products has a great potential in British Columbia. 129 Chapter 6. Conclusions and future work A rotational viscometer was used to study the viscous properties of a paste made from the formulation optimized in Chapter 4 and containing pink salmon surimi, salt, whey protein concentrate and wheat starch. This paste had a non-Newtonian shear thinning behavior with a yield stress. Apparent viscosity and yield stress increased with temperature up to 21 °C. This could be due to protein-protein interactions that increased the viscosity with temperature. A new model taking into consideration the effects of shear rate and temperature provided a reasonable representation of the rheological data and demonstrated that up to 21 °C, shear stress increased with increasing shear rate and temperature. 6.2 Proposed future work For future studies, it is recommended that the mechanisms of surimi gelation of other species be investigated using Raman spectroscopy. This method could also be used to assess the changes of protein structure caused by the frozen storage offish or surimi. More work needs to be done in order to commercialize the use of herring, pink salmon and Pacific whiting in the manufacture of surimi-based products. For instance, new storage conditions for surimi should be determined in order to 130 Chapter 6. Conclusions and future work prolong the frozen storage at reasonable temperatures (i.e., -20°C). Raman spectroscopy could be used to investigate the changes in fish protein structure due to frozen storage. The viscous properties of raw surimi and of surimi paste should be studied further using a measuring instrument, such as a capillary rheometer, that could handle very viscous materials and hence could allow higher incubation temperatures. 131 Nomenclature Nomenclature a redness index b yellowness index C whiteness index CT model parameter of the Morgan model D shear rate, s"1 e model parameter of the ST model Ei expressible liquid, % AE V molar free activation energy, J mol"1 f model parameter of the ST model F s fold score g model parameter of the ST model G s gel strength, N.mm h model parameter of the ST model •850/830 intensity ratio of the tyrosine doublet K consistency coefficient of the HB model Kc parameter of the Casson model m parameter of the yield stress model n flow behavior index 132 Nomenclature N buried mole fraction of the buried tyrosine residues N exposed mole fraction of the exposed tyrosine residues Oi more resolved spectrum P Likelihood function Pc contracted vertex PE expanded vertex PH vertex with the highest objective function value P* reflected vertex R universal gas constant, J mol"1 K"1 R2 regression coefficient s peak shape function S shear stress, Pa So yield stress, Pa T temperature, °C setting temperature, °C x2 setting time, min. X3 cooking temperature, °C X4 cooking time, min. yi measured spectral data Yj objective function of the vertex Pj Y objective function of the SOM Wj weighting coefficient 133 Nomenclature WPC whey protein concentrate, % WS wheat starch, % Z multiobjective function 134 Nomenclature Greek letters a reflection coefficient 0 contraction coefficient 00 parameter of the SOM 0. parameter of the SOM 0H parameter of the SOM 0« parameter of the SOM Y expansion coefficient a standard deviation viscosity, Pa.s ® convolution 135 Bibliography Bibliography Arakawa, T.; Tirriasheff, S. N. Stabilization of protein structure by sugars. Biochemistry 1982, 21, 6536-6544. Arteaga, G. E. Assessment of protein surface hydrophobicity by spectroscopic methods and its relation to emulsifying properties of proteins. Ph. D. Thesis, Department of Food Science, The University of British Columbia, Vancouver, B. C, Canada, 1994. Arteaga, G. E. Personal communication. 1992 Barrett, T. W.; Peticolas, W. L. Laser Raman light-scattering observations of conformational changes in myosin induced by inorganic salts. Biophys. J. 1978, 23, 349-358. Borderias, A. J.; Jimenez-Colmenero; F.; Tejada, M. Parameters affecting viscosity as a quality control for frozen fish. Marine Fisheries Review 1985, 47,43-45. Bourne, M. C. Viscosity and Consistency. In Food Texture and Viscosity, G. F. Stewart, B. S. Schweigert and J. Hawthorn, Ed.; Academic Press, Inc., New York, N. Y., 1982; pp 215-219. Branden; C ; Tooze, J. Introduction to Protein Structure. Garlan Publishing, Inc., New York, N. Y., 1991. Burgarella, J. C; Lanier, T. C; Hamann, D. D.; Wu, M. C. Gel strength development during heating of surimi in combination with egg white or whey protein concentrate. J. Food Sci. 1985, 50, 1595-1597. Caille, J. P.; Pigeon-Gosselin, M.; Pezolet, M. Laser Raman study of internally perfused muscle fibers. Effect of Mg2+, ATP and Ca 2 +. Biochim. Biophys. Acta 1983, 758, 121-127. Carew, E. B.; Asher, I. M. Laser Raman spectroscopy - New probe of myosin substructure. Science 1975, 188, 933-936. Chan, J. K.; Gill, T. A.; Paulson, A. T. The dynamics of thermal denaturation offish myosins. Food Res. Int. 1992, 25, 117-123. 136 Bibliography Chew, S. H.; Zheng, Q. Integral Global Optimization: Theory, Implementation and Applications. Lecture Notes in Economics and Mathematical Systems. Vol 298, Springer-Verlag, Berlin, Germany, 1988. Childs, E. A. Interaction of formaldehyde with fish muscle in vitro. J. Food Sci. 1973, 38, 1009-1011. Chung, K. H. Texture-modifying effect ofnonfish proteins in surimi gel. Ph. D. Thesis, Department of Food Science and Nutrition, University of Rhode Island, Kingston, Rhode Island, USA, 1990 Colthup, N. B.; Lawrence, H. D.; Wiberley, S. E. Introduction to Infrared and Raman Spectroscopy. Academic Press, Inc., New York, N. Y., 1990. Connel, J. J. Fish muscle proteins and some effects on them of processing. In Symposium on Foods: Proteins and their Reaction, H. W. Schultz and A. F. Anglemier, Ed.; AVI Publishing Co., Westport, Connecticut, 1964. Quoted by Shenouda (1980). de Neufville, R. Multiobjective optimization. In Applied Systems Analysis. McGraw-Hill Inc., New York, N. Y, 1990; pp 173-188. deNoyer, L. K.; Dodd, J. G. Maximum likelihood deconvolution for spectroscopy and chromatography. Am. Lab. 1991, 23 (12), 24D-24H. Douglas-Schwartz, M; Lee, C. M. Comparison of thermostability of red hake and Alaska pollock surimi during processing. J. Food Sci. 1988, 53, 1347-1351. Edgar, T. F.; Himmelblau, D. M. Optimization of Chemical Processes. McGraw-Hill Inc., New York, N. Y, 1988. Fukuda, K. Fish Processing and Prevention of Protein Denaturation, K. Arai, Ed.; Koseisha-Koseikaku, Tokyo, Japan, 1991. Gill, T. A.; Conway, J. T. Thermal aggregation of cod (Gadus morhua) muscle proteins using 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide as a zero length cross linker. Agric. Biol. Chem. 1989, 53, 2553-2562. Groninger, H.; Hawkes, J. W.; Babbitt, J. K. Functional and morphological changes in processed frozen fish muscle. J. Food Sci. 1983, 48, 1388-1390. 137 Bibliography Gwinn, S. Development of surimi technology in the United States. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y., 1992; pp 23-39. Hamann, D. D.; MacDonald, G. A. Rheology and texture properties of surimi and surimi based foods. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc. New York, N. Y., 1992; pp 429-500. Hastings, R. J.; Keay, J. N.; Young, K. W. The properties of surimi and kamaboko gels from nine British species offish. Int. J. Food Sci. Technol. 1990, 25, 281-294. Hastings, R. J.; Currall, J. E. P. Effects of water, oil, egg white and starch on the texture of cod surimi gels by response surface methodology. J. Text. Stud. 1989, 19, 431-451. Holmes, K. L; Noguchi, S. F; MacDonald, G. A. The Alaska pollock resource and other species used for surimi. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y., 1992; pp41-76. Itoh, Y.; Yoshinaka, R.; Ikeda, S. Behaviour of the sulfhydryl groups of carp actomyosin by heating. Bull. Jap. Soc. Sci. Fish. 1979, 45, 1019-1022. Jimenez-Colmenero, F.; Borderias, A. J. A study of the effects of frozen storage on certain functional properties of meat and fish protein. J. Food Technol. 1983, 18, 731-737. Kannan, K. K.; Nostrand, B.; Fridborg, K.; Lovgren, S.; Oilsson, A.; Petef, M. Crystal structure of human erythrocyte carbonic anhydrase B. Three dimensional structure at a nominal 2.2-A resolution. Proc. Nat. Acad. Sci. 1975, 51. Quoted by Tu (1982). Khuri, A. I.; Cornell, J. A. Response Surfaces, Designs and Analysis. Marcel Dekker, New York, N. Y, 1987. Kim, J. M.; Lee, C. M. Effect of starch on textural properties of surimi gel. J. Food Sci. 1987, 52, 722-725. Lanier, T. C. Measurement of surimi composition and functional properties. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y, 1992; pp 123-163. 138 Bibliography Lanier, T. C; Hart, K.; Martin, R. E. A Manual of Standard Methods for Measuring and Specifying the Properties of Surimi. Published by: University of North Carolina Sea Grant College Program, Raleigh, N. C, 1991. Lanier, T. C; Lin, T. S.; Liu, Y. M.; Hamann, D. D. Heat gelation properties of actomyosin and surimi prepared from Atlantic croaker. J. Food Sci. 1982, 47, 1921-1925. Larsoon, K.; Rand, R. P. Detection of changes in the environment of hydrocarbon chains by Raman spectroscopy and its application to lipid-protein systems. Biochim. Biophys. Acta 1973, 326, 245-255. Lee, H. G.; Lee, C. M.; Chung, K. H.; Lavery, S. A. Sodium ascorbate affects surimi gel-forming properties. J. Food Sci. 1992a, 57, 1343-1347. Lee, C. M.; Wu, M. C; Okada, M. Ingredient and formulation technology for surimi-based products. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y, 1992b; pp 273-302. Lee, C. M.; Whiting, R.; Jenkins, R. Texture and sensory evaluation of frankfurters made with different formulations and processes. J. Food Sci. 1987, 52, 896-900. Lee, C. M.; Kim, J. M. Texture and freeze-thaw stability of surimi gels in relation to ingredient and formulation. In International Symposium on Engineered Seafoods Including Surimi, R. Martin and R. Collette, Ed.; National Fisheries Institute, Washington, D. C, 1986; pp 168-187. Lee, C. M. Surimi process technology. Food Technol. 1984, 38 (11), 69-80. Li-Chan, E.; Nakai, S.; Hirotsuka, M. Raman spectroscopy as a probe of protein structure in food systems. In Protein Structure-Function Relationships in Foods, R. Y. Yada, R. L. Jackman, and J. L. Smith, Ed.; Chapman & Hall Inc., New York, N. Y, 1994; pp 163-197. Li-Chan, E.; Nakai, S. Raman spectroscopic study of thermally and/or dithiothreitol induced gelation of lysozyme. J. Agric. Food Chem. 1991, 39, 1238-1245. Luu, C; Luu, D. V.; Rull, F.; Sopron, F. Etude par effet Raman de la perturbation structurale de I'eau liquide par une substance etrangere. J. Mol. Struct. 1982, 81, 1-10. 139 Bibliography Mathlouthi, M.; Portmann, M. O. Hydrogen-bonding and the sweet taste mechanism. J. Mol. Struct. 1990, 237, 327-338. Matsumuoto, J. J.; Noguchi, S. F. Cryostabilization of protein in surimi. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y, 1992; pp 357-388. Matsumuoto, J. J. Chemical deterioration of muscle proteins during frozen storage. In Chemical Deterioration of Proteins, J. Whitaker and M. Fujimaki, eds. ACS Symp. No. 123, American Chemical Society, Washington, D. C, 1980; pp 97. Quoted by Shenouda (1980). Mendel, J. M. Maximum-Likelihood Deconvolution. Springer-Verlag, New York, N. Y., 1990. Morgan, R. G.; Steffe, J. F.; Ofoli, R. Y. A generalized viscosity model for extrusion of protein doughs. J. Food Process Eng. 1989, 11, 55-78. Morrissey, M. T.; Wu, J. W.; Lin, D.; An, H. Protease inhibitor effects on torsion measurements and autolysis of Pacific whiting surimi. J. Food Sci. 1993, 58, 1050-1054. Nakai, S. Personal communication. 1993 Nakai, S.; Li-Chan, E. Hydrophobic interactions in muscle proteins and their role in comminuted meat systems. In Hydrophobic Interactions in Food Systems, CRC Press Inc., Boca Raton, FL, 1988; pp 63-127. Nakayama, T.; Niwa, E.; Hamada, I.; Shin, C. Viscosity changes in carp actomyosin solutions. J. Food Sci. 1979, 44, 1106-1109. Nicklason, P.M.; Pigott, G.M. Simultaneous heating, forming and extruding of surimi based products. J. Food Eng. 1989, 9, 219-229. Nishimura, K.; Bouraoui, M.; Nakai, S. Personal communication. 1993. Niwa, E. Chemistry of surimi gelation. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y., 1992; pp 389-427. Nonaka, M.; Li-Chan, E.; Nakai, S. Raman spectroscopic study of thermally induced gelation of whey proteins. J. Agric. Food Chem. 1993, 41, 1176-1181. 140 Bibliography Ohshima, T.; Suzuki, T.; Koizumi, C. New developments in surimi technology. Trends Food Sci. Technol. 1993. 4, 157-163. Okada, M. History of surimi technology in Japan. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y., 1992; pp 3-21. Owusu-Ansah, Y. J.; McCurdy, A. R.; Kopp, T. G. Viscosity changes of actomyosin and properties of surimi from sucker (Catostomus commersoni). Can. Inst. Food Sci. Technol. J. 1988, 21, 531-533. Park, J. W.; Yongsawatdigul, J.; Lin, T. M. Rheological behavior and potential cross-linking of Pacific whiting (Merluccius productus) surimi gel. J. Food Sci. 1994, 59, 773-776. Pigott, G. M. The high tech"raw materials from minced fish flesh. Food Rev. Int. 1986, 2, 213-246. Przybycien, T. M.; Bailey, J. E. Structure-function relationships in the inorganic salt-induced precipitation of a-chymotrypsin. Biochim. Biophys. Acta 1989, 995, 231-245. Sano, T.; Noguchi, S. F.; Matsumoto, J. J.; Tsuchiya, T. Thermal gelation characteristics of myosin subfragments. J. Food Sci. 1990, 55, 55-58. Scott, D. N.; Porter, R. W.; Kudo, G.; Miller, R.; Koury, B. Effect of freezing and frozen storage of Alaska pollock on the chemical and gel-forming properties of surimi. J. Food Sci. 1988, 53, 353-358. Shenouda, S. Y. K. Theories of protein denaturation during frozen storage of fish flesh. Advances Food Res. 1980, 26, 275-311. Sproul, J. T.; Queirolo, L. E. Trade and management: exclusive economic zones and the changing Japanese surimi market. Marine Fisheries Review. 1994, 56, 31-39. Stone, A. P.; Stanley, D. W. Mechanisms offish muscle gelation. Food Res. Int. 1992, 25, 381-388. Streeter, V. L.; Wylie, E. B. Fluid properties. In Fluid Mechanics, V. L. Streeterand E. B. Wylie, Ed.; McGraw-Hill, Inc., New York, N. Y, 1985; pp 3-24. 141 Bibliography Tagagi, I. State of water in kamaboko. In Water in Food. Japan. Soc. Sci. Fish., Ed.; Koseisha, Koseikaku, Tokyo, Japan, 1973, pp 95-100. Quoted by Niwa (1992). Toyoda, K.; Kimura, I.; Fujita, T.; Noguchi, S. F.; Lee, C. M. 1992. The surimi manufacturing process. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc., New York, N. Y., 1992; pp 79-112. Tsukamasa, Y; Sato, K.; Shimizu, Y; Imai, C ; Sugiyama, M.; Minegishi, Y; Kawabata, M. e-(y-Glutamyl) lysine crosslink formation in sardine myofibril sol during setting at 25°C. J. Food Sci. 1993, 58, 785-787. Tu, A. T. Peptide backbone conformation and microenvironment of protein side chains. In Spectroscopy of Biological Systems, R. J. H. Clark and R. E. Hester, Ed.; John Wiley & Sons Ltd., New York, N. Y., 1986; pp 47-112. Tu, A. T. Proteins. In Raman Spectroscopy in Biology: Principles and Applications, A. T. Tu, Ed.; John Wiley & Sons Ltd., New York, N. Y, 1982; pp 65-116. Tung, A. T.; Speers, R. A.; Britt, I. J.; Owen, S. R.; Wilson, L. L. Yield stress characterization of structured foods. In Engineering and Food, v.1, W. E. L. Spiess and H. Schubert, Ed.; Elsevier Applied Science, New York, N. Y, 1989; pp 79-88. Vaessen, W. Nonlinear Function Optimization. Published by: The University of British Columbia Computing Centre, Vancouver, B. C, Canada, 1984. Walters, F. H.; Parker, L. R.; Morgan, S. L; Deming, S. N. Sequential Simplex Optimization. CRC Press, Inc., Boca Raton, FL, 1991. Watabe, S.; Hashimito, K. Temperature conditions for the frozen storage of representative marine products in Japan. Food Rev. Int. 1987, 2, 353-393. Wheaton, F. W.; Lawson, T. B. Processing Aquatic Food Products. John Wiley & Sons Ltd., New York, N. Y, 1985. Williams, R. W. Estimation of protein secondary structure from the laser Raman Amide I spectrum. J. Mol. Biol. 1983, 166, 581-603. 142 Bibliography Wu, M. C. Manufacture of surimi-based products. In Surimi Technology, T. C. Lanier and C. M. Lee, Ed.; Marcel Dekker, Inc. New York, N. Y., 1992; pp 245-272. Wu, M. C; Hamann, D. D.; Lanier, T. C. Rheological and calorimetric investigations of starch-fish protein systems during thermal processing. J. Text. Stud. 1985, 16, 53-74. Yada, R. Y.; Jackman, R. L.; Nakai, S. Secondary structure prediction and determination of proteins, a review. Int. J. Peptide Protein Res. 1988, 31,98-108. Yassien, H. A. A level set global optimization method for nonlinear engineering problems. Ph. D. Thesis, Department of Civil Engineering, The University of British Columbia, Vancouver, B. C, Canada, 1993. Yongsawatdigul, J.; Park, J. W.; Kolbe, E.; Abu Dagga, Y.; Morrissey, M. T. Ohmic heating maximizes gel functionality of Pacific whiting surimi. J Food Sci. 1995, 60, 10-14. 143 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0058575/manifest

Comment

Related Items