AN ABSTRACT OF THE DISSERTATION OF

Tomoko Okada for the degree of Doctor of Philosophy in

Food Science and Technology presented on May 5, 2006.

Title: Extraction and Production of n-3 Polvunsaturated Fatty Acid Concentrate from Pacific Sardines (Sardinops sasax)

Abstract approved:

MicEael T. Morfissfcy--

Intrinsic characteristics of Pacific sardines were determined. The lipid content in sardines was initially low (6.79%) in the beginning of the season, increased to 22.95% in mid-August, and decreased slightly by the end of September. An inverse correlation (R2 = -0.90) was found between lipid and moisture content. Analysis showed that 20:5n3 was the most abundant fatty acid in sardine oil followed by 16:0 and22:6n3. Microbial Upases from Candida Rugosa (CR), Candida cylindracea (CC), Mucorjavanicus (MJ), and Aspergillus niger (AN) were used for enzymatic hydrolyses of sardine oil to concentrate n-3 polyunsaturated fatty acids (n-3 PUFAs). Hydrolysis with CR and CC Upases resulted in a two-fold increase of docosahexaenoic acid while MJ and AN were less effective in concentrating n-3 PUFAs. Thin layer chromatography analysis showed that triglycerides were the predominant fraction in the final products. A new oil extraction method involving isoelectric point (pi) adjustment was compared with the heat processing method in terms of lipid recovery and quality. Oil extracted by adjusting pH to pi using tartaric acid and citric acid showed the highest oil recovery, followed by calcium tartarate (Ca-TA) and calcium citrate (Ca-CA). The oil extracted by heat contained the lowest total fatty acids and resulted in the lowest total n-3 PUFAs. Oil extracted by the heat process showed the highest TBA (value 12.24) on 0 day while the lowest thiobarbituric acid reactive substance values were found in the oil extracted with Ca-TA and Ca-CA. A study was conducted to develop an immobilized-enzyme system to entrap CR lipase in chitosan-alginate-CaCl2 beads for the purpose of concentrating n-3 PUFA from sardine oil. Lipase was immobilized by an ionotropic gelatin method. Scanning electron microscopy analysis revealed that lipase had a strong influence on bead structure. Optimum pH of immobilized lipase shifted to pH 6.0 from that of free lipase (pH 7.0), and immobilized lipase showed higher stability against pH and temperature changes. Both free and immobilized lipases increased total n-3 PUFAs from 38.13% to 65.00% after 90 min of repeated hydrolysis. Among different lipid fractions, the diglyceride fraction contained the highest n-3 PUFAs. © Copyright by Tomoko Okada May 5, 2006 All Rights Reserved Extraction and Production of n-3 Polyunsaturated Fatty Acid Concentrate from Pacific Sardines (Sardinops sagax)

by Tomoko Okada

A DISSERTATION

submitted to

Oregon State University

in partial fulfillment of the requirements for the degree of

Doctor of Philosophy

Presented May 5, 2006 Commencement June 2006 Doctor of Philosophy dissertation of Tomoko Okada presented on May 5, 2006.

APPROVED:

-1^-

Major Professor, representing Food Science and Technology

KJ Head of the Department of Food Science and Technology

V^ Dean of the Graduate School

I understand that my dissertation will become part of the permanent collection of Oregon State University libraries. My signature below authorizes release of my dissertation to any reader upon request.

Tomoko Okada, Author ACKNOWLEDGEMENTS

First and foremost, I would like to express my sincere appreciation to my major professor, Dr. Michael T. Morrissey for his generous and enthusiastic guidance throughout the Ph.D. program. It has been very fun to work with him, and I have learned a lot, both in regards to carrying out scientific research and maintaining a healthy mindset. I will always have fond memories of my experiences with him, his family and everyone at the OSU Seafood Laboratory. I am also deeply grateful to the members of my graduate committee (Dr. Charles D. Boyer, Dr. Robert J. McGorrin, Dr. Yi-Cheng Su, and Dr. Andrew S. Ross) for their generous advice and suggestions on my research. I very much appreciate Dr. Jae W. Park and Dr. Yi-Cheng Su for their kindness and encouragements during my stay in Astoria. My thanks to Sue Hansell, Toni Miethe, Dr. Craig Holt, Angee Hunt, Michael Thompson, Nancy Chamberlain, and the students at the OSU Seafood Laboratory for their help, support, and warm friendship. It always felt like home while working at the office surrounded by such nice people. I will carry many nice memories from Astoria when I return to my country. I would like to express my deep gratitude to Aki Hill for her tremendous amount of help and support on everything during my stay in the US from the very beginning to the end. Without her support, it would be hard to imagine how I would have managed my whole new life in the US. Thanks to the National Oceanic and Atmospheric Administration and the Oregon State University Sea Grant College Program. This work was supported by grant No.NA16RG1039 from the National Oceanic and Atmospheric Administration to the Oregon State University Sea Grant College Program and by appropriations made by the Oregon State Legislature. Finally, sincere appreciation goes to my parents in Japan for their love and encouragement, without whom I would never have enjoyed so many opportunities. TABLE OF CONTENTS Page

1. INTRODUCTION 1

2. LITERATURE REVIEW 4

2.1 METHODS FOR THE PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID CONCENTRATE 4

2.2 MICROBIAL LIPASES FOR N-3 POLYUNSATURATED FATTY ACID CONCENTRATE PRODUCTION 5

2.3 INTERESTERDFICATION 7

2.3.1 Acidolysis 8

2.3.2 Glycerolysis 9

2.3.3 Alcoholysis 10

2.3.4 Transesterification 11

2.4 HYDROLYSIS 11

3. SEASONAL CHANGES IN INTRINSIC CHARACTERISTICS OF PACIFIC SARDINE (Sardinops sagax) 17

3.1 ABSTRACT 18 TABLE OF CONTENTS (Continued) Page

3.2 INTRODUCTION 19

3.3 MATERIALS AND METHODS 21

3.4 RESULTS AND DISCUSSION 23

3.5 CONCLUSIONS 38

4. PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID CONCENTRATE FROM SARDINE OIL BY LIPASE-CATALYZED HYDRLYSIS 39

4.1 ABSTRACT 40

4.2 INTRODUCTION 41

4.3 MATERIALS AND METHODS 42

4.4 RESULTS AND DISCUSSION 47

4.5 CONCLUSIONS 59

5. RECOVERY AND CHARACTERIZATION OF SARDINE OIL EXTRACTED BY pH ADJUSTMENT 60

5.1 ABSTRACT 61 TABLE OF CONTENTS (Continued) Page

5.2 INTRODUCTION 62

5.3 MATERIALS AND METHODS 63

5.4 RESULTS AND DICUSSION 68

5.5 CONCLUSIONS 81

6. PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID CONCENTRATE FROM SARDINE OIL BY IMMOBILIZED CANDIDA RUGOSA LIPASE IN ALGINATE-CHITOSAN-CaC^ BEADS 82

6.1 ABSTRACT 83

6.2 INTRODUCTION 84

6.3 MATERIALS AND METHODS 87

6.4 RESULTS AND DISCUSSION 92

6.5 CONCLUSIONS 109

7. GENERAL CONCLUSIONS 110

BIBLIOGRAPHY 112 APPENDIX 124 MARINE ENZYMES 124 LIST OF FIGURES Figure Page

2.1 Reaction schemes for the lipase-catalyzed hydrolysis offish oil and concentration of n-3 PUFA 13

2.2 Production of n-3 PUFA concentrate oil with free lipase system 14

3.1 Changes in sardine landings and vessels targeting sardines from 1999 to 2005 in Oregon-Washington coastal area 19

3.2 Weights of sardines collected in the Oregon-Washington coastal area from June to September2004 25

3.3 Lengths of sardines collected in the Oregon-Washington coastal area from June to September 2004 25

3.4 Lipid and moisture content of Pacific sardines over the 2004 fishery season 28

3.5 Correlation between lipid and moisture content of Pacific sardines 30

3.6 Effect of sardine size on lipid and moisture content 37

4.1 Degree of hydrolysis (%) of hydrolyzed sardine oil by Upases at370C 50 LIST OF FIGURES (Continued) Figure Page

4.2 Changes in C16:0, C16:ln7, EPA, and DHA concentration in final n-3 PUFA concentrate with lipases from (a) CR, (b) CC, (c) MJ and (d) AN during hydrolysis at 3 TC 51

4.3 Changes in TG, DG, MG fractions of final n-3 PUFA concentrate, and acid values in hydrolyzed sardine oil with lipase from (a) CR and (b) CC during hydrolysis reaction at 370C 57

4.4 Changes in TG, DG, MG fractions of final n-3 PUFA concentrate, and acid values in hydrolyzed sardine oil with lipase from (a) MJ and (b) AN during hydrolysis reaction at 370C 58

5.1 Changes in solubility of sardine muscle protein at different pH values 69

5.2 Percent oil recovery from sardine mince and lipid distribution in centrifuge residues and emulsions 70

5.3 Lipid compositions of sardine oils extracted by heat and pH adjustment methods 74

5.4 Protein content in sardine oils extracted by heat and pH adjustment methods 78 LIST OF FIGURES (Continued) Figure Page

5.5 Changes in TBA values of sardine oil extracted by heat and pH adjustment methods over the storage period 81

6.1 Lipase immobilization by an ionotropic gelatin method utilizing chitosan and alginate 86

6.2 Mechanical strength of the CR lipase-immobilized beads prepared with 0.5% alginate-0.5% chitosan (Al-Ch), 0.5% alginate-1.0% CaCfc (Al-CaCl2 1%), 0.5% alginate-2.0% CaCk (Al-CaCfe 2%), 0.5% alginate-0.5% chitosan-1.0% C&Ch (Al-Ch-CaCU 1%) or 0.5% alginate-0.5% chitosan-2.0% CaCh (Al-Ch-CaCh 2%) 93

6.3 SEM micrographs of alginate beads prepared with (a) 2.0% alginate + 0.5% chitosan, (b) 2.0% alginate + 1.0% CaCfe (c) 2.0% alginate + 0.5% chitosan + 1.0 % CaCb, (d) 2.0% alginate + 0.5% chitosan +1.0% CaCU + CR lipase 94

6.4 Percent lipase retained in alginate-chitosan-CaCk beads and percent released into chitosan solution 96

6.5 Effect of reaction pH values on hydrolytic activities of free lipase and immobilized lipase 98

6.6 Effect of reaction temperatures on hydrolytic activities of free lipase and immobilized lipase 100 LIST OF FIGURES (Continued) Figure Page

6.7 Changes in 16:0 and 16:ln7 in final n-3 PUFA concentrate hydrolyzed (1st and 2nd hydrolysis) with (a) free lipase or (b) immobilized lipase, and 20:5n3 and 22:6n3 in oils hydrolyzed (1st and 2nd hydrolysis) by (c) free lipase or (d) immobilized lipase 102

6.8 Changes in lipid composition of final n-3 PUFA concentrate with either (a) free or (b) immobilized lipase system after 90 min with 1st or 2nd hydrolysis 106

6.9 Fatty acid composition of total lipid, TG, DG and MG fractions of original oil and final n-3 PUFA concentrate with immobilized lipase after a 90 min reaction time (1st hydrolysis) 108 LIST OF TABLES Table Page

2.1 Commercially available microbial lipases for lipid modification 7

3.1 Changes in proximate composition of Pacific sardines over summer 2004 27

3.2 Changes in lipid composition of Pacific sardines 32

3.3 Changes in fatty acid composition (wt/wt%) of total sardine lipid 34

4.1 Microbial lipases and their characteristics 44

4.2 Fatty acid profiles of crude and refined sardine oil (wt/wt%) 48

5.1 Color analysis of sardine oils extracted by heat and pH adjustment methods 72

5.2 Fatty acid profiles of sardine oils extracted by heat and pH adjustment methods 76

5.3 Absorbance of conjugated dienes and trienes of oil extracted by heat or pH adjustment methods 79

6.1 Fatty acid profile of free fatty acid fraction of sardine oils hydrolyzed with immobilized CR lipase 104 DEDICATION This dissertation is dedicated to my parents, Hiroshi and Noriko Okada EXTRACTION AND PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID CONCENTRATE FROM PACIFIC SARDINES (SADINOPS SAGAX)

CHAPTER 1 INTRODUCTION The Pacific sardine (Sardinops sagax) is a coastal pelagic fish found from the Gulf of California to Southeastern Alaska (Hammann, 1998; Hammann et al., 1988; Schweigert, 2002). Pacific sardines have recently made a strong biological comeback in the Oregon and Washington (OR-WA) coastal area, with an increase in catch volume from 776 metric tons (mt) with 3 vessels in 1999 to over 45,000 mt in 2005 with 20 vessels. The Pacific sardine fishery is concentrated at the mouth of the Columbia River and most of the fish are 3-5 years old (McCrae and Smith, 2006). Currently, the majority of the sardine catch is frozen whole and sold at very low ex- vessel prices to Asian markets for use as long-line bait in the tuna fishery. Although Pacific sardines have valuable nutritional qualities, they have yet to be exploited for human consumption, especially here in the U.S. Sardines contain high amounts of lipids as well as high levels of n-3 polyunsaturated fatty acids (n-3 PUFA), particularly eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). Consumption of n-3 PUFAs has been associated with a variety of health benefits, including reduced incidence of coronary heart disease and mental health disorders as well as partial remediation of diabetes and various types of cancers (Alasalvar et al., 2002; Nettleton, 1995; Nettleton and Katz, 2005). Recent work has also demonstrated positive effects of these fatty acids on the development of brain and nervous tissue in infants (Innis, 2003; Innis, 2004). In order to maximize utilization of Pacific sardines for human consumption, sardine oil recovery and subsequent production of n-3 PUFA concentrate was investigated over the course of four research projects presented here. In Project 1, the intrinsic characteristics of Pacific sardines were studied. Determination of intrinsic factors including proximate composition and fatty acid profiles is important because these factors include inherent chemical/physiological characteristics in fish species, 2 and capturing sardines at their peak physiological condition can result in maximum utilization and increased economic benefits. The objective of this project was to determine intrinsic characteristics in sardines harvested off the OR-WA coast during the 2004 season and track changes in these characteristics as the season progressed. In Project 2, oil extraction from sardines by a conventional method using heat (wet reduction method) and production of n-3 PUFA concentrate from extracted sardine oil were investigated. Among a variety of possible methods to concentrate n-3 PUFA levels in sardine oils, lipase-catalyzed hydrolysis was selected because the reaction is relatively simple and faster than other available methods, such as interesterification (Mohapatra and Hsu, 1999). Commercially available microbial lipases (1,3-specific and non-specific) were evaluated for efficiency in concentrating n-3 PUFAs. The objectives of this project were: to recover oil from sardines by the wet reduction method; develop n-3 PUFA concentrate using lipase-catalyzed hydrolysis; and determine the optimal processing parameters for n-3 PUFA production. Traditionally, industrial fish oil extraction uses a wet reduction method, which includes high temperature treatments (Aidos et al., 2003; Chantachum et al., 2000). The high levels of PUFAs in sardine oil are susceptible to oxidation, and it is known that high temperatures facilitate lipid oxidation. Oil extraction with high temperatures results in decomposition of PUFAs, including EPA and DHA, and production of trans- isomers of these fatty acids, resulting in decreased nutritional value (Toyoshima et al., 2004). In Project 3, a new method for oil extraction without the use of heat was investigated. The objectives were to develop a new oil extraction method using pH adjustment and compare the new oil extraction method with the conventional wet reduction method in terms of lipid recovery and quality. Typical n-3 PUFA concentrate production requires addition of organic solvent (typically ethanol) to inactivate water-solubilized free lipases in the reaction medium. However, recovery and reuse of these lipases after each application can be difficult. If the enzyme can be immobilized, a simple n-3 PUFA concentrate production method without the use of organic solvent can be developed, where the lipase can be easily removed by filtration from the reaction medium and reused in another hydrolysis 3 reaction. A study was conducted to develop an inunobilized-enzyme system to entrap lipase from Candida rugosa in chitosan-alginate-CaCl2 beads for the purpose of concentrating n-3 PUFAs from sardine oil (Project 4). The objectives of this project were to develop a new n-3 PUFA concentrate production method using immobilized lipase; determine the characteristics of the immobilized-lipase system, such as hydrolytic activity; determine the efficiency of immobilized lipase in n-3 PUFA concentration; and compare these characteristics with a free lipase system. The over all objective was to increase sardine utilization for human consumption through oil extraction and modification for nutritional purposes. A better understanding of seasonal variation of sardines and developing extraction and production of n-3 PUFA concentrate will provide useful information for the growing sardine industry and assist its expansion into new markets. CHAPTER 2

LITERATURE REVIEW

2.1 METHODS FOR THE PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID CONCENTRATE With growing public awareness of the clinical benefits of n-3 polyunsaturated fatty acid (n-3 PUFA), researchers have been developing a variety of techniques to produce n-3 PUFA concentrate from fish oil (Shahidi and Wanasundara, 1998). These methods include supercritical extraction with CO2 (Hsieh et al., 2005; Riha and Brunner, 2000), fractional solvent crystallization (Chen and Ju, 2000; Chen and Ju, 2002; Ganga et al, 1998), urea complexation (Adachi et al., 1993; Guil-Guerrero and Belarbi, 2001; Hwang and Liang, 2001; Wanasundara and Shahidi, 1999), esterification (Akimoto et al., 2003; Haraldsson et al., 1997; Robles et al., 1999; Schmitt-Rozieres et al., 2000), hydrolysis (Linder et al., 2005; Rakshit et al., 2000; Sun et al., 2002; Ustun et al., 1997; Zheng et al., 2005), supercritical fluid and silver- resin chromatography (Alkio et al., 2000; Nieto et al., 1997), acidolysis (Adachi et al., 1993; Garcia et al., 2000; Gonzalez Moreno et al., 2004; Ramirez Fajardo et al., 2003), and any combination of such methods (Adachi et al., 1993; Gamez-Meza et al., 2003; Ganga et al., 1998). A goal in concentrating fish n-3 PUP As from fish oil is to maximize the triglyceride (TG) moiety, which is considered to be more nutritionally favorable than methyl or ethyl esters of these fatty acids. Experimental results have shown impaired intestinal absorption of methyl or ethyl esters of n-3 PUFAs (as compared to TGs) in laboratory animals (Lawson and Hughes, 1998; Lin et al., 2006). Among the available methods for production of n-3 PUFAs concentrate in glyceride form, enzymatic processes using lipases, including interesterification and hydrolysis, have been extensively studied. Interesterification of glycerides involves rearrangement of fatty acids in TG with the addition of PUFA concentrate in the free fatty acid (FFA) form. This reaction is carried out in organic solvents where lipase stimulates acyl-group migration 5 between the individual glyceride molecules (Bomscheuer, 2000; Uhlig, 1998). Hydrolysis is a simple hydrolase-catalyzed reaction involving the enzymatic cleavage of specific fatty acids (Bomscheuer, 2000). This reaction results in a mixture of mono-, di-, triglycerides and FFAs depending on degree of hydrolysis. Both the interesterification and hydrolysis reactions are reversible and, under low water activity conditions, the enzyme functions 'in reverse', that is the synthesis of an ester bond rather than its hydrolysis (Miller et al., 1988; Shahidi and Wanasundara, 1998). The use of lipases as biocatalysts has considerable merits: superior efficacy under mild reaction conditions reduced environmental pollution, few side products, high availability from a wide range of sources, and high adaptability for modifications (Rakshit et al., 2000; Xuebing, 2000). Production of n-3 PUFA by utilizing lipases has advantages over traditional methods such as chemical hydrolysis and chemical interesterification since such methods involve high temperatures which may partially destroy the natural a\\-cis n-3 PUFA by oxidation and by cis-trans isomerization (Senanayake and Shahidi, 1999). Enzymatic reactions can be carried out under mild conditions such as neutral pH and low temperature, which could save energy and increase product selectivity. Also, use of lipase for n-3 PUFA concentration has a potential benefit in that the n-3 PUFA concentrates are in glyceride form. Therefore, the procedures have become popular and their industrial production is gaining momentum (Shahidi and Wanasundara, 1998).

2.2 MICROBIAL LIPASES FOR N-3 POLYUNSATURATED FATTY ACID CONCENTRATE PRODUCTION Lipases (triacylglycerol acylhydrolases) are hydrolases that act on carboxylic ester bonds resulting in a separation of the fatty acid from the glyceride backbone (Houde et al., 2004). They posses the unique feature of acting at the aqueous and non- aqueous interface which distinguishes them from esterases (Vakhlu and Kour, 2006) Lipases can be obtained from either animals, plants or microbes; however, microbial lipases have gained much attention due to their multifold properties, easy extraction 6 procedures, virtually unlimited supply, thermostable nature, and ability to carry out reactions without use of a co-lipase (Xuebing, 2000). Microbial Upases can be divided into two categories: regiospecific (non-specific and 1,3-specific) and fatty acid specific (Gupta et al., 2003). Non-specific Upases show no specificity regarding the position of the ester bond in the glycerol molecule. Non-specific Upases cleave TG into FFA producing intermediate 1,2-, or 2,3-diglycerides (DG). Examples of enzymes in this category are Upases from the microorganisms Candida cylindracae and Candida rugosa. The 1,3-specific Upases possess stereospecificity with respect to positions 1 and 3 of the glycerol residue, resulting in 2-monoglyceride (MG) and a mixture of fatty acids (Uhlig, 1998). This type of specificity is common among microbial Upases and is the result of the sterically-hindered ester bond at the 2- position, which is unable to enter the active site of the enzyme. These Upases can be obtained from Aspergillus niger, Mucor javanicus, and Rhizopus arrhizus (deMan, 1999). Table 1 shows commercially available microbial Upases and their specificities (modified from Xuebing, 2000). Fatty acid specific Upases show specificity for particular fatty acids. This category includes lipase from C. rugosa, which has fatty acid chainlength selectivity, showing higher activity with relatively short chain fatty acids such as C18 or below (McNeill et al., 1996). In contrast, lipase from Geotrichum candidum possesses specificity for long chain fatty acids that contain a cis double bond in the 2- position of TG (deMan, 1999). Table 2.1 Commercially available microbial lipases for lipid modification. Lipase source Specificity Aspergillus niger 1,3 » 2 Candida cylindcea non-specific Candida lipolytica 1,3 > 2 Candida rugosa non-specific Mucorjavanicus 1,3 » 2 Rhizomucor miehei 1>3 » 2 Penicillium roquefortii 1,3-specific Rhizopus delemar 1,3 » 2 Rh izopus javan icus 1,3 > 2 Rhizopus japonicus 1,3 > 2 Rhizopus niveus 1,3 > 2 Pseudomonas fluorescens 1,3 > 2 Rhizopus arrhizus 1,3 > 2

2.3 INTERESTERIFICATION Interesterification can be defined as modifying the properties of a TG molecule by changing its fatty acid composition (Uhlig, 1998). Enzymatic interesterification systems consist of the lipase catalyst and a very small amount of water dispersed in a continuous organic phase comprising the reactants and an optional solvent such as hexane. Lipases bind at the interface between aqueous and organic phases and catalyze hydrolysis. In the presence of organic solvents, lipases catalyze the synthesis of new esters and modify lipid composition of the glycerides (Benjamin and Pandey, 1998). Interesterification can be broadly categorized into four groups involving exchange of acyl moieties between (1) an ester and an acid (acidolysis), (2) an ester 8 and a glycerol (glycerolysis), (3) an ester and an alcohol (alcoholysis), or (4) an ester and an ester (transesterification) (Nawar, 1996).

2.3.1 Acidolysis Among the available methods to concentrate n-3 PUFAs, acidolysis is one of the most intensively studied applications. Acidolysis reactions generally consist of oil, lipase and an excess level of PUFA/n-3 PUFA concentrates or eicosapentaenoic acid (EPA)/docosahexaenoic acid (DHA) as FFAs. The use of an organic solvent, usually ethanol, propanol, octanol or glycerol, is common for this application. Lipase from Pseudomonas spp. is often used for acidolysis reactions, and it has been reported to be the most effective lipase for incorporation of n-3 PUFA into borage and evening primrose oil as compared to lipases from Mucor miehei and Candida Antarctica (Senanayake and Shahidi, 1999). The optimum conditions reported for lipase from Pseudomonas spp. were 278-299 U lipase per g oil at 42-430C with 24-26 h reaction time. According to the results, 35.5% and 33.6% EPA + DHA incorporation into borage and evening primrose oils were achieved, respectively. Rakshit (2000) studied acidolysis of hydrolyzed tuna oil using Pseudomonas fluorescens lipase and demonstrated that fatty acid specificity of the lipase is highly influenced by the alcohol used in the reaction, where long chain alcohols such as octanol showed the best result. This method resulted in a yield of 80%) for DHA and 70% for EPA with an enrichment of 45% EPA and 11% DHA in the unesterified FFA. Adachi et al. (1993) reported that concentration of EPA and DHA from sardine oil was achieved by acidolysis with Pseudomonas spp. lipase. Acetone was more effective than n-hexane as a solvent for dissolving the reactants and concentrating EPA and DHA. The amount of lipase used was 10,000 U per g oil, and the total percent of EPA and DHA in the glycerides reached 65%. Enrichment of DHA in sardine oil from sardine cannery effluents using Lipozyme™ was studied and shown to be highly efficient for DHA enrichment (80%), whereas EPA was not concentrated significantly (Schmitt-Rozieres et al., 2000). In general, factors that affect acidolysis efficiency include use of different kinds of alcohols, the presence of water, temperature and 9 lipase concentration. Also, Tanaka et al. (1994) suggested two important factors for obtaining TGs rich in DHA: 1) the original TG has to be of high DHA, and 2) the DHA content of FFA used in the acydolysis reaction has to be high.

2.3.2 Glycerolysis A typical glycerolysis reaction can be carried out with glycerol, PUFA concentrate, hexane, water and lipase. Optimum reaction conditions reported by Robles Medina et al. (1999) were 100 mg of lipase, 9 ml hexane, and a molar ratio of 1.2 to 3 for glycerol and PUFA concentrate. PUFA concentrate obtained from cod liver oil was esterified with glycerol by Novozyme 435 with hexane. This reaction resulted in high levels of EPA and DHA (25.7% and 44.7% of total fatty acid, respectively) in TG form. Robles et al. (1998) also investigated lipase-catalyzed glycerolysis with n-3 PUFA concentrate from winterized menhaden oil. The reaction was run with different organic solvents (hexane, isooctane, ethyl acetate, t-butanol, 2- propanol or chloroform), glycerol, winterized menhaden oil, water and lipase from Pseudomonas spp. Of all the organic solvents tested, hexane was selected as the most suitable. Optimum reaction conditions were reported to be T = 390C for 18 h in a reaction mixture containing 2 ml hexane, 1.5 ml glycerol (47.3%), 0.1 ml water, 0.1 ml lipase solution (1%» w/v) and 1 ml n-3 enriched oil {2% in hexane). The total concentrations of EPA and DHA in the final product were 20 and 15%), respectively. He and Shahidi (1997) also studied enzymatic synthesis of glycerides directly from glycerol and n-3 PUFA concentrate prepared from sea blubber oil with organic solvent. Glycerolysis catalyzed by lipase from Chromobacterium viscosum resulted in the highest esterification rate. The highest glycerides synthesis reached 94.3%, where the DG fraction was the highest (43.1%) of all fractions. This fraction contained fairly high levels of EPA and DHA (25.25% and 49.71%, respectively). Esterification of hydrolyzed oil (FFA form) from sardine, tuna and herring oil with glycerol was studied using Rhizomucor miehei lipase (Halldorsson et al., 2003). While most of the study focused on increasing both EPA and DHA in the final product, 10 this study was rather unique in that EPA concentrate in glyceride form and DHA concentrate in FFA form were generated separately from the same hydrolyzed fish oil. The purpose was to meet a demand both for EPA concentrates with low DHA content for specific EPA-related health effects and DHA concentrates with low EPA content for infant formulas. The mechanism utilized the different kinetic resolutions of EPA and DHA, which occur during the esterification reaction when the rate constant for EPA is not equal to the rate constant for DHA. In the typical procedure, the mixture for glycerolysis consists of 1.0 g lipase, a 10.0 g mixture offish oil FFA and 1.59 g glycerol. It was reported that after 48 h of reaction with hydrolyzed tuna oil containing 5% EPA and 25% DHA (area %), the residual FFA contained 78% DHA and only 3% EPA with 19% DHA recovery, and the glycerides contained 6.7% EPA and 5.0% DHA with 90% EPA recovery. It was also suggested that the type offish oil and extent of conversion were highly important parameters in controlling the degree of concentration. In regards to production of structured lipids, several studies have reported incorporation of n-3 PUFA into plant oil such as palm oil and soybean oil. Ramirez Fajardo et al. (2003) reported that lipase-catalyzed acidolysis using a commercially available FFA mix along with organic solvent could result in the successful incorporation of n-3 PUFA into palm oil. Akimoto et al. (2003) was able to incorporate n-3 PUFAs into soybean oil with acidolysis using a commercially available immobilized R. miehei lipase and PUFA concentrate prepared from sardine oil.

2.3.3 Alcoholysis Typically, an alcoholysis reaction consists of oil, lipase and alcohol, mainly in the form of ethanol (ethanolysis). The saturated fatty acid (SFA) and mono unsaturated fatty acid (MUFA) fractions can be converted to ethyl esters, while n-3 PUFAs including EPA and DHA remain attached to the residual acylglycerides. In addition to acidolysis reactions, lipases from Pseudomonas spp. have been reported to be efficient catalysts in alcoholysis reactions. Production of EPA and DHA 11 concentrate from sardine oil has been investigated by ethanolysis with 17 different commercially available Upases (Haraldsson et al., 1997). In accordance with a typical ethanolysis procedure, 8500 U of lipase was added to a mixture of the sardine oil TG (5.0 g) and absolute ethanol (0.8 g), and the reaction was carried out at room temperature for a prolonged time. The study demonstrated that Pseudomonas Upases have the highest activity toward SFA and MUFA, and the lowest activity toward PUFAs. After 24 h at 20oC, there was a 50% conversion of SFA and MUFA into ethyl esters, while the n-3 PUFAs remained in the residual mixture in glyceride form, with 50% of total EPA and DHA content. In a study examining ethanolysis with lipase from P. fluorescenes, DHA and EPA were enriched to total of 74% in the glyceride mixture (Rakshit et al., 2000).

2.3.4 Transesterification Transesterification for the production of n-3 PUFA concentrate from fish oil has not been as thoroughly studied as other methods discussed here. Lin et al. (2006) investigated enrichment of n-3 PUFAs of menhaden oil by transesterification using the 1,3-specific lipase M. miehei and supercritical carbon dioxide as a solvent. Free n-3 PUFA was prepared from menhaden oil by the urea inclusion method involving saponification, formation of the urea inclusion complex, and extraction of free n-3 PUFAs. The resulting n-3 PUFA concentrate, which contained 71.2% EPA + DHA (wt/wt%), was subjected to transesterification withM miehei lipase and TG in menhaden oil. The optimum temperature for this process was determined to be 50oC. A positive effect was reported with a 10%) ethanol addition, and the n-3 PUFA content in TG was reported to increase from 35%) to 56%) (wt/wt%>).

2.4 HYDROLYSIS In hydrolysis reactions, Upases catalyze the hydrolysis of fatty acid ester bond in the TG and release FFA (Sheldon, 1993). Factors that have been suggested to affect the reaction rates of Upases include: the positional specificity of the enzyme, the rate of the reverse reaction, the reactivity of lipase toward specific fatty acids, lipase 12 concentration, substrate concentration, temperature, pH and presence or absence of metallic ions (Sun et al., 2002; Tanaka et al, 1992; Uhlig, 1998; Uhlig, 1998). Concentration of n-3 PUFA utilizing lipases is based on the fatty acid specific mechanism of lipase toward the SFA and MUFA, leaving n-3 PUFAs intact on the glyceride moiety as shown in Fig. 2.1 (Wanasundara and Shahidi, 1998). The molecular conformations of cis carbon-carbon double bonds in PUFAs, particularly EPA and DHA, result in bending of the chains. This brings the terminal methyl group of these fatty acids close enough to the ester bond to cause a steric hindrance effect on lipase. Therefore, the lipase active site cannot interact with the ester linkage between these fatty acids and glycerol, protecting EPA and DHA from lipase-catalyzed hydrolysis. However, hydrolysis with lipase is not hindered in the case of the relatively straight chains of SFAs and MUFAs (Carvalho et al., 2002; Shahidi and Wanasundara, 1998). Lipase-catalyzed enzymatic hydrolysis holds significant advantages over competing methods in that hydrolysis reactions are generally faster and simpler than other reactions such as interesterification (Mohapatra and Hsu, 1999) which generally involve organic solvents (e.g. ethanol) as substrates in the reaction mixture, requiring increased caution with handling and discarding procedures. In general, enzymatic lipid hydrolysis systems consist of the oil, a lipase catalyst and a selected buffer at the optimum pH for the lipase (Fig. 2.2). After hydrolysis, lipase is inactivated by addition of alcohol or high heat treatment. FFA released by hydrolysis in the obtained oil are removed by neutralization, typically with potassium hydroxide, followed by liquid-liquid extractions using water and a solvent such as hexane. The hexane layer, which contains higher levels of n-3 PUFAs compared to the original unhydrolyzed oil, can be separated by separatory funnel, and an n-3 PUFA concentrated oil is obtained after the hexane is evaporated. In the case of immobilized lipase, a simple lipase removal by filtration can be carried out after the hydrolysis reaction, with no need for an organic solvent. Fish oil Lipase KOH n-3 PUFA concentrate

v"

Buffer having the Water optimum pH of lipase 16:0 (COO)+ K+

Aqueous phase

Figure 2.1 Reaction schemes for the lipase-catalyzed hydrolysis offish oil and concentration of n-3 PUFA. Buffer Sardine oil -► I Lipase Hydrolysis H 95% EtOH Inactivation of lipase KOH Hexane Neutralizing free FA Water I Liquid-liquid extraction x 3 I L Hexane layer Aqueous layer Anhydrous Na2S04 Filtration ► Hydrated Na2S04 Evaporation Hexane n-3 PUFA concentrate

Figure 2.2 Production of n-3 PUFA concentrate oil with free lipase system. 15 Linder et al. (2005) extracted oil from salmon heads enzymatically using commercially available proteases, and recovered oil was hydrolyzed with lipase from Aspergillus oryzae (1,3-specific lipase) to concentrate n-3 PUFAs. A reaction mixture containing 1000 g oil, 200 Kilo Lipase Units, and 1000 ml of distilled water was incubated at 370C, pH 7.0 for 12 h. The resulting hydrolyzed oil was then subjected to membrane filtration in order to separate FFA from glycerides. The results show that the total PUFA levels increased by hydrolysis reaction followed by FFA removal from 41.6% in the crude oil to 46.5% in refined oil, with increases in both DHA (from 9.9 to 11.6%) and EPA (from 3.6 to 5.6%). Carvalho et al. (2002) investigated EPA and DHA enrichment from sardine oil by hydrolysis with four different microbial lipases. Enzyme (150 U of each) was reacted with 500 mg sardine oil in 3.5 ml potassium phosphate buffer (pH 7.0, 0.08 M). The results showed that C. cylindracea (non-specific) lipase increased total levels of EPA and DHA from 25.5% in the original oil to 40.4% in the refined oil. Most of the enrichment took place with DHA, which experienced a 2-fold increase in concentration; EPA content, on the other hand, did not change significantly following hydrolysis. Non-specific lipases were reported to be more effective for the enrichment of DHA compared to 1,3-specific lipases, such as Aspergillus oryzae lipase, in agreement with other studies (Sun et al., 2002). Sun et al. investigated n-3 PUFA concentrate production from salmon viscera oil using lipase-catalyzed hydrolysis with 6 different commercially available lipases. Hydrolysis reaction mixtures contained 4 g viscera oil, 3200 U lipase, and phosphate buffer, and the reaction was carried out for a prolonged period of time at 350C with constant agitation. The results demonstrated that non-specific lipases, such as those from Pseudomonas cepacia and C. rugosa, were the most effective of all lipases tested. After 12 h of hydrolysis, EPA content increased from 64 to 91 mg/g, and DHA content increased from 53 to 91 mg/g. Wanasundara and Shahidi (1998) also reported that hydrolysis with non-specific lipase was more effective in concentrating n-3 PUFA levels in marine oil (seal blubber oil and menhaden oil) among a number of commercial microbial lipases {A. niger, C. cylindracea, C. viscosum, G. candidum, M. miehei, Pseudomonas spp., Rhizopus 16 oryzae, and Rhizopus niveus). C. cylindracea was found to be the most effective at n- 3 PUFA concentration, increasing total n-3 PUFAs from 20.18% to 43.50% in seal blubber oil and from 30.95% to 44.10% in menhaden oil after 40 h hydrolysis. Zhen et al. (2005) concentrated PUFAs in commercially available fish oil (species undefined) by hydrolysis with five commercially available lipases (Penicillium expansum, P. cyclopium, Candida lipolytica, A. oryzae and Pancreatic lipases). Lipase from A. oryzae showed the highest efficiency in regards to PUFA concentration, with increases in both EPA (from 3.0 to 9.0%) and DHA (from 4.3% to 16.5%)) after a 24 h hydrolysis reaction. 17

CHAPTER 3

SEASONAL CHANGES IN INTRINSIC CHARACTERISTICS

OF PACIFIC SARDINE (SARDINOPS SAGAX)

Tomoko Okada and Michael T. Morrissey

Submitted to Journal of Aquatic Food Product Technology Haworth Press, Binghamton, NY, USA 18 3.1 ABSTRACT Intrinsic characteristics of sardines harvested off the Oregon-Washington coast were determined. Analyses included physical measurements, proximate composition, lipid composition and fatty acid profiles during the 2004 season. The average weight of sardines was 189.18 ± 33.26 g and length was 27.41 ± 1.76 cm (n= 230) throughout the season. The lipid content was initially low (6.79%) in the beginning of the season in June, increased to 22.95% in mid-August, and decreased slightly by the end of September. An inverse correlation (R = -0.90) was found between lipid and moisture content. Lipid composition analysis showed triacylglycerides as the predominant lipid fraction ranging from 78.62% to 80.93% while the phospholipids fraction was the lowest (1.83% to 2.13%)). Free fatty acid levels decreased slightly as the season progressed from 6.81%) to 5.05%. Analysis showed that 20:5n3 was the most abundant fatty acid in sardine oil followed by 16:0 and 22:6n3 fatty acids. The level of 20:5n3 fatty acid was slightly higher in August - September (ranging from 23.91% to 24.51%)) than earlier in the season (21.77% to 22.45%).

Keywords: Sardine, lipid, proximate analysis, gas chromatography, thin layer chromatography 19 3.2 INTRODUCTION

Sardines are subtropical species existing in areas of subarctic and tropical ocean mixing, where they constitute some of the most voluminous fisheries species in the world (Leonardis and Macciola, 2004; Lluch-Belda et al., 1991). They are gregarious plankton-eating fish that live in large schools in the open sea and are usually found in the upper trophic levels of the ocean (Leonardis and Macciola, 2004). The Pacific sardine {Sardinops sagax) is a coastal pelagic fish found from the Gulf of California to southeastern Alaska (Hammann, 1998; Hammann et al., 1988; Schweigert, 2002). Pacific sardines were first landed commercially in the Pacific Northwest during the 1935-36 season and continued to be a very active fishery until collapsing in the early 1940s (Emmet et al., 2005). The abrupt decline in abundance of Pacific sardines is largely related to ocean temperature changes linked to climate or ocean regime changes that vary over periods of about fifty years (Chavez et al., 2003). A new Pacific sardines fishery began in the Oregon-Washington (OR-WA) coastal area in 1999 and the catch volume increased from 776 metric tons (mt) in 1999 to over 45,000 mt in 2005 (Fig. 3.1) (McCrae and Smith, 2006). Sardine landings generally begin in June and peak in August-September depending on catch rates and fishery quotas (Emmet et al., 2005).

O ■4-» •co 6 o on o en (U > c/5 G

1999 2000 2001 2002 2003 2004 2005 Figure 3.1 Changes in sardine landing&ajid vessels targeting sardines from 1999 to 2005 in Oregon-Washington coastal area. 20 Sardines and other pelagic fish are often termed fatty fish with lipid content that can reach levels in excess of 20% but can also drop below 4%, depending on time of year (Caponio et al., 2004). This variability depends on several factors including seasonality, metabolic function, food availability and type, sea water temperature, maturity, spawning and geographic location (Aidos et al., 2002; Leonardis and Macciola, 2004). Sardines contain high levels of lipids and high amounts of long- chain n-3 polyunsaturated fatty acids (LC-PUFA) including 20:5n3 and 22:6n3, also known as eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA), respectively. These fatty acids play important physiological roles in fish as components of membrane phospholipids and as precursors of biologically active eicosanoids (Rodriguez et al., 2004). The nutritional importance of fish oil for human nutrition is well known and is mainly associated with EPA and DHA. These fatty acids have been shown to exhibit several beneficial effects on human health including reduction of coronary heart disease, partial remediation of diabetes, delay of the onset of Alzheimer's disease, and inhibition of various types of cancers (Alasalvar et al., 2002; Holub and Holub, 2004; Nettleton and Katz, 2005). Recent work also has demonstrated positive effects of these fatty acids in the development of brain and nervous tissue for infants (Innis, 2003; Innis, 2004). It has been reported that n-3 PUFAs are effective in reducing the risk of behavioral and learning disorders including attention deficit hyperactivity disorder, dyslexia, dyspraxia and autism (Ruxton et al., 2004) as well as mental illness such as depression, suicidal tendencies and other violent behaviors (Hibbeln and Salem, 2001). Studies suggest that intake of n-3 PUFAs has beneficial effects on symptoms of Alzheimer's disease such as short-term memory loss, depressed mood, and inability to sleep, as well as end-stage Huntington's disease (Haag, 2003; Tiemeier et al., 2003). Although the sardine biomass for the U.S. fishery stretches along the West Coast from Mexico to Alaska, the fishing activity is concentrated at the mouth of the Columbia River, which borders WA and OR. Sardine schools are located by fixed- wing airplane and harvested by large purse seining vessels (25-40 m) which can 21 capture 20-40 mt each set. The fish are pumped directly from the nets into refrigerated seawater holds on board the vessels and rapidly cooled to < 4°C (Dion, 2003). The fish are normally harvested in the afternoon and off-loaded at processing plants within 8 h of capture, graded by size, packed into 10 kg boxes and frozen. This allows for rapid turnover from harvests to final product and helps maintain a very high quality. However, most landings at Astoria, OR (located near the mouth of the Columbia River) have been processed as bait for the overseas long-line tuna fishery with low ex-vessel prices of approximately $100-120/metric ton. Sardines landed in Astoria tend to be relatively large in size, with average weights of about 150-180 g. While this has caused some concern in the bait-fish industry, which prefers smaller sardines (130-150 g), it does provide opportunities for developing new products and directing these sardines toward human consumption. The objective of this study was to determine intrinsic characteristics in sardines harvested off the OR-WA coast during the 2004 season and track changes in these characteristics as the season progressed. A better understanding of seasonal changes that occur in the proximate composition and other parameters of sardines will provide useful information for the growing sardine industry and assist its expansion into new markets.

3.3 MATERIALS AND METHODS 3.3.1 Fish and sample preparation Sardines (Sardinops sagax) were harvested off the OR-WA coast by commercial purse seiners and samples were collected and analyzed between June 8 and September 23, 2004. The sardines were pumped from the seine nets directly into refrigerated sea water holds aboard the vessels and rapidly cooled to < 40C. The sardines were off-loaded within 8 h of capture at local processing plants in Astoria, and separated into three different size groups. Seven of the smallest-, middle- and largest-sized sardines were randomly chosen, packed in ice and delivered to the laboratory on the day of catch. Analysis of the sardines began the following day and physical measurements (weight and length) were taken. The head, viscera and 22 backbone were removed from each fish, and skin-on fillets were obtained. The fillets from all three sizes were mixed together and minced by a vertical-cutter/mixer (Stephan Food Processing Technology, West Germany) for 2 min and kept in vacuum- sealed bags at -30oC for further analysis. In a separate set of experiments, sardines were separated by weight (150-199.9 g and 200-249.9 g) during the period of July to August to determine if sardine size had an effect on total lipid content.

3.3.2 Proximate analysis Moisture analysis was carried out according to Association of Official Analytical Chemists (AOAC) method 934.01 (1998). Total nitrogen content was determined by Kjeldal method according to AOAC method 940.25 (1998) and multiplied by 6.25 to estimate the crude protein content. Ash content was gravimetrically determined using a muffle furnace by heating at 5150C to constant weight according to AOAC method 938.08 (1998). Total lipid content was determined by hydrolyzing the sample with hydrochloric acid in 100oC water bath and extracting the lipid fraction with diethyl ether and petroleum ether. The lipid was obtained after heating on 100oC hot plate and in an oven to remove the ether layers according to AOAC method 948.15 (1998).

3.3.3 Thin-layer chromatography separation Extracted sardine lipids were subjected to thin layer chromatography (TLC) separation for lipid composition determination. Lipids were separated into triacylglyceride (TG), monoacylglyceride (MG), diacylglyceride (DG), free fatty acid (FFA), and phospholipids (PL). Samples were dissolved in hexane (10% solution, v/v), and an aliquot of 5 (j.1 sample mixture was loaded onto the silica plate (K6 silica gel 60, 200x200x0.25 mm, Whatman Inc., Brentford, UK). The developing solvent was hexane:diethylether:acetic acid (60:40:2, volume basis). Spots were visualized by spraying 5% sulfuric acid solution followed by drying at 100oC on the hot plate for 30 min. Quantification was done by densitometory ChemDoc with Quantity One software (Bio-Rad, Hercules, CA). Commercially available TLC standard (Cat. No. 18-1A and 23 18-5 A, Nu Check Prep, Inc., Elysian, MN) was used for qualification, and fraction levels were expressed as wt/wt%.

3.3.4 Determination of fatty acid composition by gas chromatography Fatty acids in sardine oil were converted into fatty acid methyl esters (FAME) according to AOAC method 991.39 (1998), and their composition was determined by gas chromatography (GC). A Hewlett-Packard 5890 Series II gas chromatograph (Palo Alto, CA) equipped with a flame-ionization detector and capillary column (EC- wax, 30 m X 0.25 mm i.d; split ratio, 100:1; Alltech, Deerfield, IL) was used for analyzing FAME. Parameters of the GC system were set as follows: injector and detector temperatures 250oC and 270oC, respectively; the column temperature was set to 50oC and gradually heated to 180oC at a rate of 50C/ min, then slowed to a rate of 0.8oC/ min until it reached 220oC; helium was used as a carrier gas. The fatty acid concentrations were calculated by comparison of their retention times with those of the reference standards (Supelco, Bellefonte, PA).

3.3.5 Statistical analysis The results were presented as average and standard deviations of each experiment with triplicates. Statistical comparisons were made between treatments by ANOVA and Tukey's test; linear regression analyses were applied using S-PLUS software package. The results were presented in terms of p values, with significant mean differences at thepO.OS level.

3.4 RESULTS AND DISCUSSION 3.4.1 Physical measurement The weight and length of sardines collected for this study in Astoria during 2004 are shown in Fig. 3.2 and 3.3, respectively. The average (ave) weight was 189.18 ± 33.26 g, and the ave length was 27.41 ± 1.76 cm (n= 230) throughout the fishing season. This is within the range reported by Oregon Dept. of Fisheries and Wildlife for the same season (31.3 g - 293.6 g, ave 154.4 g) (McCrae and Smith, 2005). 24 Fish size depends on the year-class of the fish being harvested and the large sardines are associated with five-year old fish. Sardines from the Astoria area were larger in size compared to reports in regions such as Japan (Sardinops melanostictus), with a weight range of 30.6-90.7 g and length range of 12.2-19.7 cm (Shirai et al., 2002); Turkey, with a range of 60-80 g in weight and 14-16 cm in length (Serdaroglu and Felekogglu, 2005); and Brazil {Sarinella sp), with weight range of 48.6 g to 98.4 g and length of 14.6 cm to 19.1 cm (Luzia et al., 2003). Pacific sardines from the Gulf of California, a similar specie as those in the present study, were also smaller, ranging from less than 12 cm to 22 cm (Cisner Os-Mata and Nevarez-Martinez, 1995). The sardines in this study showed consistent weight and length measurements early in the season and gradually showed a broader weight range. Emmett (2005) also reported that large- (>18 cm) and medium- (14-16 cm) sized sardines were found in June offshore near Astoria, while small sardines (< 14 cm) appeared with high density in the Astoria area in September 2004. It is thought that mature sardines migrate north early in the season and appear in the Astoria area after spawning, as oppose to the juveniles, which migrate later in the season. Mature sardines are thought to migrate south to California during the winter and generally arrive to Pacific Northwest coastal feeding grounds when sea surface temperatures exceed 120C (June-July). Small sardines in contrast, do not appear to migrate but over-winter in near shore coastal waters, including the Columbia River estuary and Willapa Bay (Emmet et al., 2005). 25

Z.JU T lJ Tm r T T 200 ' HflM-t- > ■ 3 150 35 JL ' L | 100

50

0 1 1 1 * 1 1 1 1 1 1 6/12 6/22 7/2 7/12 7/22 8/1 8/11 8/21 8/31 9/10 9/20 9/30

Date

Figure 3.2 Weights of sardines collected in the Oregon-Washington coastal area from June to September 2004.

E

c

6/12 6/22 7/2 7/12 7/22 8/1 8/11 8/21 8/31 9/10 9/20 9/30

Date

Figure 3.3 Lengths of sardines collected in the Oregon-Washington coastal area from June to September 2004.

3.4.2 Changes in proximate composition The composition of several fish species varies from season to season due to natural cycles, maturity, spawning, geographic location, and food availability. Pelagic fish, such as mackerel, bluefish (Imre and Saglik, 1998), salmon (Katikou et al., 2001), 26 herring (Aidos et al, 2002) and sardines (Gokodlu et al., 1998), typically contain relatively high levels of lipid, and go through a natural cycle showing considerable variations in lipid content and composition. Ground fish such as sole (Gokce et al., 2004), whiting (Morrissey and Sylvia, 2004), deep-sea cod, black dogfish, sharks (Okland et al., 2005), and halibut (Olsson et al., 2003) contain low levels of lipid but will still show some seasonal variation. Sardine proximate composition was measured every 14 days and is shown in Table 3.1. The lipid and moisture content of sardine samples analyzed more frequently throughout the season are shown in Fig. 3.4. In fatty fish, such as sardines, most of the lipid reserves are found in the flesh while non-fatty species carry most of their lipid reserves in their livers (Love, 1997). In general, lipids are mobilized during the spawning cycle and are at their lowest post-spawning. The lipid content in sardines was lowest in June (6.97%), and gradually increased through the season, reaching its highest level (22.95%) in mid-August. By the end of the season in September, the lipid content decreased slightly to 18.68%. The maximum level of lipid content in August may be explained by an enhanced feeding activity due to higher sea water temperatures and increased day length, both of which have a positive influence on appetite (Gokce et al., 2004). The two main spawning grounds for Pacific sardines are areas off southern California and Baja California, and spawning occurs both in the spring and fall. The area between Point Conception and Ensenada is the major spawning area (Hernandez- Vazquez, 1994); however, spawning has also occurred on the southern Oregon coast in recent years with the expansion of the biomass geographical area (Schweigert, 2002). The low lipid content of sardines at the beginning of the season (June) is most likely due to the spawning season in the previous months resulting in fat mobilization associated with gametogenesis (Hardy and Keay, 1972). Studies have also shown that lower content of lipid in fish can be observed in the reproductive season of the species since reduced feeding level and sexual maturation contribute to reduced lipid levels during winter (Gokce et al, 2004). Table 3.1 Changes in proximate composition of Pacific sardine over summer 2004. 6/24 7/7 7/23 8/3 8/18 9/7 9/23

Moisture 68.58±0.25a 66.21±0.07b 61.53±0.19d 60.01±0.19f 60.52±0.22ef 60.94±0.35de 62.6(^0.43°

Protein 17.70±0.26abc 17.81±0.13ab 18.33±0.48a 17.04±0.26cde 16.64±0.07de 16.32±0.27e n.\S±0.25cd

Lipid 13.73±0.21a 13.30±0.34a 18.60±0.34b 19.87±0.18d 21.02±0.38e I938±0.l9ai 18.68±0.08bc

Ash 1.68±0.03ab 1.60±0.11ab 1.63±0.06ab 1.76±0.11b 1.67±0.01ab 1.50±0.09a 1.63±0.01ab Values are % mean ± S.D. of twenty-one homogeneous samples consisting of seven sardines from each size category (100-150g, 150-200g and 200-250g). The values in the same row labeled with a different superscript letter are significantly different (p<0.05). 80

A Lipid % 70 ^ Moisture % * • •{ *.' i 1-H 60 d * * KTl •^H O 6 50 T3

T3 40 ■ PH OH •—-H 30

a 6 ^ ^6 a A 10 A

—i 1 1 \ 1 1 1 1 1 1 1— 6/2 6/12 6/22 7/2 7/12 7/22 8/1 8/11 8/21 8/31 9/10 9/20 9/30 Date

Figure 3.4 Lipid and moisture content of Pacific sardines over the 2004 fishery season. oo 29 Lower lipid content of sardines at the end of season could be explained by a sexual maturation period towards the winter season and lipid migration from the flesh to the gonads. Changes in moisture content were proportional to lipid content, and the highest moisture level was found in June (72.50%) while the lowest (59.80%) was in August (Fig. 3.4). Protein analysis showed the sardines contained 16.32% to 18.33% crude protein, indicating sardines are a good source of protein. Ash content showed only a slight variation from 1.50% to 1.16%. Results of proximate analysis were generally similar to other studies using Sardina pilchardus (Castrillon et al., 1997; Gokodlu et al., 1998; Nunes et al., 1992; Serdaroglu and Felekogglu, 2005). Protein and ash content remained constant though out the season, accounting for approximately 18- 20% of the total composition. The protein levels tend to be constant since they are mobilized from the flesh late in the depletion process and the first to be restored after the completion of spawning due to their structural importance (Love, 1997). An inverse correlation (R2= -0.90) was found between lipid and moisture content (Fig. 3.5), which is in agreement with other sardine studies such as Sardina pilchardus (Nunes et al., 1992) and Sardinops melanosticta (Sato et al., 1978). This correlation has also been found in other fish species including albacore tuna {Thunnus alalunga) (Wheeler and Morissey, 2003) and hoki (MacDonald et al., 2002). It has been suggested that lipid content is inversely proportional to moisture content in most pelagic species (Castrillon et al., 1997; Love, 1997). A simple linear regression equation was developed that shows the relation between lipid and moisture content as: H {lipid I moisture} =35.23+ Pi (- 0.32) moisture %. 55 60 65 70 75 Moisture content (%)

Figure 3.5 Correlation between lipid and moisture content of Pacific sardines. o 31 3.4.3 Changes in lipid composition Changes in the lipid composition of sardines are shown in Table 3.2. Neutral lipids ranged from 97.43% to 98.56% with TG being the predominant fraction varying from 78.62% to 80.93%. The lipid composition levels remained constant throughout the season except FFA which increased from 5.05% to 7.15% with the highest occurring towards the end of July (p<0.05). DG and MG content ranged from 4.30% to 5.54% and 2.02%) to 2.76%), with relatively constant levels during the season. PL fractions existed in very low levels varying slightly from 1.83%) to 2.13%. It has been suggested that fatty and lean sardines have different lipid composition, and PLs can be considered structural lipids whereas neutral lipids act mainly as reserve lipids (Bandarra et al., 1997; Leonardis and Macciola, 2004). These researchers showed that fatty fish had higher amounts of TG lipids compared to lean fish and lower amounts of free fatty acids, cholesterol and polar lipids. Leonardis and Macciola (2004) also reported different lipid composition in lean and fat sardines. PLs were quantitatively equal in both lean and fatty sardines (0.7 and 0.6 g/100 g sardines, respectively) where TG levels were significantly higher in fatty sardines (12.1 g/100 g sardines) compared to lean sardines (1.3 g/100 g sardine). They suggested that during hotter summer months, sardines accumulate reserve lipids, especially TG, which is metabolized for energy during the cold season. Fatty fish in the Bandarra et al. study (1997) contained 96.35% neutral lipids, which agrees with our study except that their study showed higher TG levels (91.63%). The difference could be due to environmental and physiological differences including diet availability and the sexual maturity of sardines. Also, lipid content of lean fish in their study was 1.20%) which was considerably lower than the early season fish in this study. Table 3.2 Changes in lipid composition of Pacific sardines.

6/24 7/7 7/23 8/3 8/18 9/7 9/23

Neutral lipids 98.05 ± 0.29a 97.87 ±0.37a 98.14±0.40a 98.56±0.46a 97.43±0.78a 98.05±0.31a 97.93±0.27a TG 78.62 ±0.52a 78.62 ±0.24a 78.81 ±0.28a 78.25±2.21" 79.26±1.76a 79.80±1.40a 80.93±1.76a DG 4.84±0.39a 4.56 ±0.52a 4.87±0.61a 5.54±0.51a 4.52±0.60a 4.70±0.88a 4.30±0.78a MG 2.02±0.91a 2.19±0.19a 2.41±0.68a 2.76±0.90a 2.66±0.58a 2.63±0.43a 2.56±0.25a FFA 6.81±0.32ab 6.68 ±0.40abc 7.15±0.68a 6.41±0.56abc 5.83±0.99abc 5.34±0.09bc 5.05±0.77c Phospholipids 1.95±0.29a 2.13±0.37a 1.83±0.40a 1.83±0.52a 1.90±0.39a 1.95±0.31a 2.07±0.27a Values are % mean ± S.D. of twenty-one homogeneous samples consisting of seven sardines from each size category (100-150g, 150- 200g and 200-250g). Values in the same row labeled with a different superscript letter are significantly different (p<0.05).

to 33 3.4.4 Changes in fatty acid composition The fatty acid composition of sardine lipids between June and September is shown in Table 3.3. The major fatty acids in sardine lipids were 16:0, 16:ln7, 20:5n3, and 22:6n3 with 20:5n3 being the highest which is in agreement with several other studies (Bandarra et al., 1997; Beltran and Moral, 1991; Leonardis and Macciola, 2004; Shirai et al., 2002). The saturated fatty acid (SFA) fraction increased from 33.30% to 38.63%) as the season progressed with 16:0 as the predominant fatty acid. Later in the season there were significantly higher levels of 16:0 (22.72%) compared to early season (20.41%), while other fatty acids such as 14:0 and 15:0 did not show significant differences as the season progressed. The monounsaturated fatty acid (MUFA) fraction ranged from 18.27%) to 21.12%), with the predominant fatty acid being 16:ln7 followed by 18:ln9. Minor increases of these fatty acids were found as the season progressed. On the other hand, the levels of 20:ln9 and 22:ln9 decreased from June to September (3.60%) to 2.79%, and 1.33%) to 0.61%, respectively). As the highest fatty acid fraction, PUFA,fraction ranged from 41.69% to 46.23% with the lowest concentration in mid-August and the highest in early July. The PUFA fraction consisted mainly of n-3 PUT As (18:3n3, 20:3n3, 20:5n3, 22:5n3, and 22:6n3) with high levels of 20:5n3 (range 21.77% to 24.51%)) throughout the season. Sardines harvested off the OR-WA coast showed relatively higher amounts of n-3 PUFA levels compared to other species such as Sardina pilchardus W caught in Spain (23.21%) n-3 PUFAs) and Sardinops melanostictus caught in Japan (28.8%) to 42.0% n-3 PUFAs) (Beltran and Moral, 1991; Shirai et al., 2002). A similar species (Sardinops sagas caeruleus) harvested in Mexican waters in the Gulf of California showed similar concentrations to our study of both 20:5n3 and 22:6n3 ranging from 18.26% to 23.91% and 9.61% to 13.70%, respectively (Gamez-Meza et al., 1999). Our study also showed higher 20:5n3 levels than 22:6n3 throughout the season, and the same tendency was found by others (Bandarra et al., 1997; Kotb and Abu Hadeed, 1991). Table 3.3 Changes in fatty acid composition (wt/wt%) of total sardine lipid.

FFA 6/24 111 7/23 8/3 8/18 9/7 9/23 C14:0 7.42±0.54a 7.09±0.04a 7.19±0.01a 7.33±0.09a 7.59±0.34a 7.57±0.42a 8.01±0.83a C15:0 0.41±0.03a 0.48±0.00a 0.39±0.00a 0.30±0.06a 0.33±0.09a 0.35±0.10a 0.35±0.11a C16:0 20.41±0.98a 20.90±0.10ab 21.40±0.03abc 2.50±0.11bc 21.84±0.35abc 22.03±0.46bc 22.72±1.01c C17:0 0.91±0.13a l.SgiO.Ol* 1.89±0.00bc 1.83±0.01b 1.89±0.04bc 1.86±0.04bc 2.02±0.08c C18:0 3.97±0.12a 4.36±0.01ab 5.34±0.02bc 5.81±0.07c S-lliO^S1"1 4.43±0.42ab 5.39±0.80bc C22:0 0.07±0.01a 0.05±0.00ab 0.05±0.00b 0.00±0.00c 0.04±0.00b 0.04±0.00b 0.05±0.01b C23:0 0.11±0.01a 0.11±0.00a 0.10±0.00ab O.OSiO.OO1 0.07±0.00c 0.07±0.00c 0.08±0.01bc SFA 33.30±1.42a 34.88±0.14ab se.seiO.oe^ 7.85±0.19cd 36.90±0.34bcd 36.34±0.39bc 38.63±1.16d C14:ln5 o.o4±o.or 0.04±0.00ab 0.03±0.00b 0.03±0.00b 0.03±0.00b 0.04±0.00ab 0.04±0.01ab C16:ln7 9.83±0.51ab 9.25±0.05a 10.06±0.01ab 10.43±0.07b 10.68±0.29b 10.29±0.39ab 10.49±0.84b C17:ln9 0.95±0.17a l.ll±0.01ab 1.34±0.00c l^SiO-Ol"0 1.33±0.03bc l^SdzCOS1"1 1.33±0.03c C18:ln9 3.90±0.07b 3.48±0.01a 4.58±0.00': 5.24±0.03d 4.53±0.08c 4.51±0.07c 4.45±0.23c C20:ln9 3.60±0.08a 2.84±0.01b 2.75±0.00b 2.31±0.02c 1.66±0.16d 1.66±0.03d 1.33±0.14e C22:ln9 2.79±0.22a 2.17±0.01b 1.87±0.00c 1.18±0.01d 0.90±0.02e 0.72±0.01ef 0.61±0.05f 024:1 n9 0.03±0.00a 0.03±0.00a 0.02±0.00bc 0.02±0.00cd 0.02±0.00b<:d 0.03±0.01b 0.02*0.00°" MUFA 21.12±0.40a 18.90±0.05b 20.65±0.01a 20.45±0.09a 19.18±0.48b 18.47±0.45b 18.27±0.92b C18:2n6 1.84±0.03a 1.72±0.01b 1.50±0.00d 1.49±0.00d 1.49±0.01d 1.61±0.00c 1.69±0.02b C18:3n6 0.33±0.02ab 0.37±0.00c 0.38±0.00c 0.36±0.00c 0.36±0.00bc 0.32±0.01ab 0.31±0.03a C18:3n3 0.79±0.01a 0.74±0.02b 0.55±0.00c 0.50±0.01d 0.47±0.00e 0.52±0.02cd 0.50±0.01de C20:2n6 0.14±0.00a 0.12±0.00b 0.10±0.00d O.lOiO.OO"1 0.08±0.00e 0.10±0.00c 0.00±0.00f C20:3n6 0.20±0.01a 0.21±0.01a 0.17±0.01b 0.20±0.00ab 0.17±0.02ab 0.19±0.00ab 0.20±0.02ab C20:3n3 0.72±0.02a 0.71±0.07a 0.65±0.00ab 0.61±0.00b 0.65±0.02ab 0.72±0.04a o.esio.os3" C20:4n6 0.10±0.02a 0.03±0.01b 0.03±0.01b 0.06±0.01b 0.05±0.01b 0.05±0.01b 0.03±0.01b C20:5n3 22.02±0.54a 21.77±0.03a 22.45±0.00a 22.14±0.08a 24.51±0.27b 23.91±0.25b 24.18±0.87b C22:2n6 0.89t0.05a 0.93±0.01a 0.89±0.00a 0.87±0.01a 0.91±0.01a 0.88±0.02a 1.03±0.05b C22:5n3 4.96±0.34a 4.67±0.03a 5.06±0.01a 5.08±0.06a 5.14±0.12a 5.22±0.10a 4.90±0.42a C22:6n3 13.61±0.92b 14.96±0.10a 11.22±0.03cd 10.27*0.11^ 10.15±0.23dc 11.79±0.22c 9.58*0.77* PUFA 45.58±1.81ab 46.23±0.09a 42.9ftt0.06bc 41.69±0.27c 43.98±0.62abc 45.30±0.65ab 43.09±2.10bc n-3 42.09±1.82a 42.85±0.10a 39.93±0.05ab 38.61±0.26b 40.92±0.63ab 42.15±0.63a 39.83±2.09ab Values are % mean ± S.D. of twenty-one homogeneous samples consisting of seven sardines from each size category (100-150g, 150-200g and 200-250g). Values in the same row labeled with a different superscript letter are significantly different (p<0.05). uo -^ 35 However, some researchers reported higher 22:6n3 levels than 20:5n3 in sardine lipid, such as: the Beltran and Moral study (1991) which showed 22:6n3 levels of 11.30% and 20:5n3 of 9.44%; Castrillon et al. (1997), reporting 22:6n3 levels of 16.92% and 20:5n3 of 12.44%, and Leonardis and Macciola(2004), reporting 22:6n3 levels of 11.3% and 20:5n3 of 6.5%. Researchers have suggested that the proportional changes of 20:5n3 and 22:6n3 can be influenced by the plankton species in their diet. In the post-spawning period when sardines start to feed eagerly, 20:5n3 levels increase with a proportional decrease of 22:6n3, as the predominant fatty acid in plankton is 20:5n3 (Bandarra et al, 1997). Minor increases of 20:5n3 levels were found as the season progressed, from approximately 22% in June-August to 24% in August-September. A similar tendency has been observed in other studies and it was suggested that seasonal blooms of phytoplankton and spawning cycles could be the cause for these changes (Gamez- Meza et al., 1999). The relationship among 20:5n3 and 22:6n3, season and total lipid content was evaluated by multivariable linear regression using the model 20:5n3= P + Pi (days) + p2 (22:6n3) + Ps (lipid). Results (R2 = 0.67,/?<0.001) showed a positive correlation between season and 20:5n3 content (Pi= 0.03,/?<0.001) at constant 22:6n3 and lipid content, showing an increase in 20:5n3 was associated with a statistically significant progression of season. In contrast, there were no significant relationships between 20:5n3 and 22:6n3 (p2= 0.24,p< 0.17), and 20:5n3 and lipid content (p3= 0.18, p< 0.11). The impact of season, 20:5n3, and lipid content on 22:6n3 level was also evaluated with the model 22:6n3= p + Pi (days) + p2 (20:5n3) + p3 (lipid). The result (R2= 0.81,p<0.001) indicated an inverse correlation between 22:6n3 and lipid content (Pi= 0.03, jt?<0.001) where other factors were found to have no significant relationship with 22:6n3 level (pi= -0.024,p<0.06, $2= 0A3,p

3.4.5 The effect of sardine sizes on lipid and moisture content The relationship between % lipid and sardine size over the period of July 1st to August 9th is shown in Fig. 3.6. Sardines were separated into two size categories by weight (150.0-199.9 g and 200.0-249.9 g) with an ave weight of 220.8 g and 179.8 g, while lipid and moisture content were determined separately. There was a significant difference found in lipid content between these two size groups. The highest difference in lipid content was found in July, with the 200.0-249.9 g size group having higher lipid levels than 150.0-199.9 g size group. 50 60

40

K— -- 55

30 - 0s- (D

300 -- 50 o 20 --- rt

T 45 10 f\

'§6 ■:-' 1 1 ' 1 1 40 7/1 7/6 7/8 7/12 7/15 7/19 7/26 8/2 8/9 Date

D Hpid (150-199.9 g sample) 1=11 lipid (200-249.9 g sample) - moisture (150-199.9 g sample) —•— moisture (200-249.9 g sample)

Figure 3.6 Effect of sardine size on lipid and moisture content. 38 3.5 CONCLUSIONS

This study showed seasonal differences in sardine composition during the 2004 harvest season off the OR-WA coasts. With proportional decrease of moisture content, sardine lipid content increased from June to August and slightly decreased in September. The high lipid content resulting in high LC-PUFA levels in sardines suggests that sardines captured in August - September have higher potential for oil recovery and should be preferred as raw material for human consumption or development of nutraceuticals. The consumption of 100 g of sardine provides higher levels of n-3 PUFAs, in particular 20:5n3 and 22:6n3 fatty acids, than human daily requirements. 39

CHAPTER 4

PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID

CONCENTRATE FROM SARDINE OIL BY

LIPASE-CATALYZED HYDROLYSIS

Tomoko Okada and Michael T. Morrissey

Submitted to Journal of Food Chemistry Elsevier Science Ltd., Exeter, Devon, UK 40 4.1 ABSTRACT The production of n-3 polyunsaturated fatty acids (n-3 PUFAs) concentrate from oil extracted from Pacific sardines {Sardinops sagax) was studied using lipase- catalyzed hydrolysis. Sardines were obtained from a local fish plant in Astoria, OR, and the oil was extracted by wet reduction method (heat extraction) and refined by degumming, bleaching and deodorization. Commercially available microbial Upases, from Candida Rugosa (CR), Candida cylindracea (CC), Mucorjavanicus (MJ), and Aspergillus niger (AN) were used for enzymatic hydrolyses, run at 370C with constant stirring for 1.5, 3, 6, and 9 h. The highest degree of hydrolysis was shown by CR lipase (500 U) after 9 h (78.40%). Fatty acid composition analysis by gas chromatography showed that the refined unhydrolyzed oil contained 26.86% of eicosapentaenoic acid (EPA) and 13.62% of docosahexaenoic acid (DHA) (wt/wt%). CR lipase was the most effective in concentrating n-3 PUFA. Hydrolysis with 250 U CR lipase followed by free fatty acid removal increased EPA concentration to a relatively constant level of 33.74% after 1.5 h, with no further significant increases with time. DHA levels were also significantly increased after 1.5 h to 23.12% with 250 U and 23.78% with 500 U. Hydrolysis with CC resulted in maximum levels of EPA (31.91%) and DHA (22.65%) after 1.5 h with 250 U. Compared to CR and CC Upases, MJ and AN Upases resulted in low n-3 PUFA concentration in the final products. Thin layer chromatography analysis showed that the original unhydrolyzed sardine oil contained 86.20% triglyceride (TG), 13.40% diglyceride, and 0.51% monoglyceride. TG levels in final n-3 PUFA concentrate decreased significantly as reaction time progressed. This study demonstrated that lipase-catalyzed hydrolysis is a viable method to concentrate n-3 PUFAs from sardine oil and has potential for use in the sardine industry.

Keywords: n-3 polyunsaturated fatty acid, EPA, DHA, lipase, hydrolysis 41

4.2 INTRODUCTION Interest in fish consumption has been increasing recently due to the wide range of health benefits being discovered in long chain marine fatty acids, specifically eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). There is general consensus about the beneficial effects of fish consumption and the reduction of coronary heart disease (CHD). These include studies from over two decades ago when epidemiologists observed the low rate of coronary heart disease among Alaskan and Greenland Eskimos who consumed large amounts offish (Nettleton, 1995). In both the U.S. and Europe, several studies have also confirmed the positive effects of fish consumption in reducing CHD among diverse populations (He et al., 2004; Mozaffarian et al., 2005). More recently a Japanese study involving more than 40,000 individuals showed a 40% reduction in CHD among those who ate fish more than four times per week (Iso et al., 2006). The American Heart Association has recommended that all individuals should eat fish, especially fatty fish, twice a week and those with documented CHD should eat a fish diet or take fish oil supplements that supply 1 g/day of EPA and DHA (Kris-Etherton et al., 2002). There has also been research in the last decade showing the positive effects of fish oils on cognitive development and vision enhancement in newboms as well as young children (Colombo et al., 2004; Daniels et al., 2005; Innis, 2003). The long chain fatty acids EPA and DHA have also been shown to contribute to the reduction of certain types of cancers, diabetes, mental health disorders and asthma (Alasalvar et al., 2002; Nettleton and Katz, 2005). Kalmijn et al. (2004) showed that consumption of fatty fish high in EPA and DHA reduced the risk of impaired cognitive function in a study of middle-aged Europeans, while Morris et al. (2003) showed that elderly study participants had 60% less risk of developing Alzheimer's disease if they consumed fish one or more times per week. With growing public awareness of the clinical benefits of EPA and DHA, various types of products and supplements have been developed. Sales offish oil supplements have grown almost 10-fold over the last decade, reaching sales of $310 million in 2004 (Fiorillo, 2005). Most of these represent fish oil capsules made from 42 various seafood sources including fish byproducts. Consumption of EPA and DHA in a concentrated form may be more effective than general fish oil itself because these concentrates contain less saturated fatty acid (SFA). Lipase-catalyzed enzymatic production of EPA and DHA concentrate from fish oil has shown potential in producing a high quality product due to the mild conditions (e.g., neutral pH and low temperatures) of the process (Breivik et al., 1997; Sun et al., 2002). A consistently available source of high quality raw material is critical for production of n-3 polyunsaturated fatty acid (n-3 PUFA) concentrate. A new sardine (Sagax sardinops) fishery off the Oregon-Washington coast could provide a sustainable supply of the necessary raw material. In 2005 alone, 45,000 metric tons of sardines were landed in Oregon, the majority of which was sold for non-human food uses (McCrae and Smith, 2006). Previous study has shown that during peak harvest months these sardines have high lipid content (15-23%) with concomitant high levels of EPA and DHA (Chapter 3). The objectives of this study were to: recover the lipids from sardines, develop EPA and DHA concentrates using lipase-catalyzed hydrolysis, and determine the optimal processing parameters for production of a fish oil concentrate.

4.3 MATERIALS AND METHODS 4.3.1 Sample preparation Sardines were harvested by purse-seine fishing boats and pumped directly from the nets into refrigerated seawater holds on board the vessels followed by rapid cooling to 0 to 40C. The area of catch was approximately 40 kilometers offshore from the mouth of Columbia River. Sardines were off-loaded at a processing plant in Astoria and transported to Oregon State University Seafood Laboratory in ice on the same day of capture. The head, viscera and backbone were removed from each fish, and skin-on fillets were obtained. The fillets were mixed together, minced by a vertical-cutter/mixer for 2 min and kept in vacuum-sealed bags at -30oC for further analysis. 43 4.3.2 Extraction and purification of sardine oils Sardine oil was extracted from minced sardines followed by bleaching and deodorization according to Sathivel et al. (2003) with some modifications. Deionized water was added to the minced sardine in a 1:1 ratio, and the mixture was heated at 850C for 30 min. The lipid phase was removed from other particles by centrifuging at 5000 rpm for 30 min and was then stored as a crude oil at -80oC until tested. The crude oil was refined by filtrating it under vacuum followed by neutralization with KOH. The extracted oils were bleached by adding 4% activated carbon, and the oil was heated in a water bath at 70oC for 10 min. The activated carbon and impurities were removed by centrifugation at 5000 rpm for 30 min. After deodorization at 100oC for 30 min under vacuum (5 mm Hg), tertiary butyl hydroquinone (200 ppm) was added to oils as an antioxidant with constant agitation with N2. The oil was collected in 5 ml vials, the air in the vials was replaced by N2 and the vials were stored at -80oC until further use. The initial amount of sardine flesh was 16.38 kg and the yields of crude and refined oils were 1.54 kg and 1.18 kg, respectively.

4.3.3 Determination of enzyme activities Four commercially available microbial lipases were used for testing: Candida Rugosa (CR) lipase (Sigma-Aldrich Co., St. Louis, MO.), Candida cylindracea (CC) lipase (Fluka Chemie AG, Buchs, Switzerland), Mucorjavanicus (MJ) lipase (Aldrich Chemical Company, Inc., Milwaukee,WI), and Aspergillus niger (AN) lipase (Aldrich Chemical Company, Inc., Milwaukee, WI). The characteristics of these lipases are shown in Table 4.1. CR and CC lipases were non-specific lipases and MJ and AN lipases were 1,3-specific lipases. Enzyme activities were determined by the Japanese industrial standard method substituting sardine oil for olive oil (JIS, 1995). Free fatty acid (FFA) released by the hydrolysis reaction (20 min) were titrated against 0.5 N sodium hydroxide and the pH changes were monitored by adding 1% methanolic phenolphthalein solution. One unit of enzyme activity (U) was defined as the amount of enzyme that liberated 1 (amol of fatty acid per min at 370C. Table 4.1 Microbial Upases and their characteristics.

Enzyme origin Abbreviations Optimum temp. Optimum pH Specificity Activity (U/g lipase)

Candida rugosa CR 30-50 0C 7.0 None 10355

Candida cylindracea CC 30-50 0C 6.5 None 34436

Mucor javanicus MJ 30-45 0C 7.0 1-, 3-»2- 2352

Aspergillus niger AN 30-50 0C 6.5 1-, 3-»2- 4741

4^ 45 4.3.4 Hydrolysis reaction The hydrolysis of sardine oil by microbial lipases and separation of the n-3 PUFAs enriched fraction were carried out according to a modified procedure of Sun et al. (2002). Enzyme powder representing 250 U and 500 U per g oil of lipase with 8 ml of 0.1 N phosphate buffer at the optimum pH of each enzyme were mixed with refined sardine oil (2 g) and placed in a 25 ml flat-bottomed flask. The air in the flask was replaced by N2, and the flask was capped with a glass lid covered by parafilm to minimize lipid oxidation. The mixture was agitated with a magnetic stir bar at 500 rpm. The hydrolysis reaction was maintained at 370C for 1.5, 3, 6, and 9 h, and lipase was inactivated by addition of 10 ml of 95%. A previously determined amount of KOH necessary to neutralize FFA released by hydrolysis was added to the mixture. Aliquots of 100 ml hexane and 50 ml deionized water were added to remove neutralized FFA with two additional water extractions. The hexane layer, including mono-, di- and triglycerides (MG, DG, and TG, respectively) with n-3 PUFAs, was collected and filtered through a bed of anhydrous sodium sulfate. Hexane was removed by a rotary evaporator at 450C for 30 min, and final n-3 PUFA concentrate was flushed with N2 and stored at -80oC until further use.

4.3.5 Determination of degree of hydrolysis Degree of hydrolysis (DH) was determined by measuring the acid value of both unhydrolyzed and hydrolyzed oil at different treatment times (1.5, 3, 6, and 9 h) and concentrations as well as saponification value of unhydrolyzed oil according to American Oil Chemists' Society (AOCS) methods (AOCS, 1998). Blanks (no enzyme) were determined at each treatment. DH was calculated according to the following equation. _ .„.. acid value (hydrolyzed oil - blank at each condition^ ..„ DH(%)TT := ^ x 100 saponification value (original oil) - acid value (original oil) 46 4.3.6 Determination of fatty acid composition by gas chromatography Fatty acids in original sardine oil and final n-3 PUFA concentrate were converted into fatty acid methyl esters (FAME) according to AOCS method 991.39 (AOCS, 1998), and their composition was determined by gas chromatography (GC). A Hewlett-Packard 5890 Series II gas chromatograph (Hewlett-Packard, Palo Alto, CA), equipped with a flame-ionization detector, capillary column (EC-wax, 30 m x 0.25 mm i.d.; split ratio, 100:1; Alltech, Deerfield, IL) was used for analyzing FAME. GC parameters were set as follows: injector and detector temperatures set at 250oC and 270oC, respectively; column temperature set at 50oC, with gradual heating to 180oC at a rate of 50C / min, followed by slow heating to 220oC at a rate of 0.8oC / min; and helium was used as a carrier gas. The fatty acid concentrations were calculated by comparison of their retention times with those of the reference standards (Supelco, Bellefonte, PA).

4.3.7 Thin-layer chromatography separation The original sardine oil and final n-3 PUFA concentrate were separated into MG, DG, and TG fractions by thin layer chromatography (TLC). Samples were dissolved in hexane (10% solution, v/v) and an aliquot of 5 |a.l hexane-sample mixture was loaded onto a silica plate (K6 silica gel 60, Whatman Inc., Florham Park, NJ). Solvent development was applied with hexane/diethylether/acetic acid (60:40:2, volume basis), and the spots were visualized by spraying 5% sulfuric acid solution followed by drying at 100oC on the hot plate for 30 min. Quantification was carried out according to a modified method from Helmerich and Koehler (2003). TLC plates with dark brown spots were scanned (CanoScanLide 30, Canon U.S.A., Inc., Lake Success, NY), converted to a bitmap file (8 bit gray scale, 100 dpi) and analyzed by a picture evaluation program (Image J: http://rsb. info.nih. gov/ii/) to obtain the chromatograph chart from each fraction. The standard curve was made by using commercially available TLC standard (Cat. No. 18-3A, Nu Check Prep, Inc., Elysian, MN), and each measurement was made by comparison with the standard. 47 4.3.8 Statistical analysis Each hydrolysis reaction was run in duplicate and all analyses were run in triplicate. The results were presented as average and standard deviations. Statistical comparisons were made between treatments by analysis of variance (ANOVA) and Tukey's test, and linear regression was applied using S-PLUS software. The results were presented in terms of p values (p < 0.05).

4.4 RESULTS AND DISCUSSION 4.4.1 Fatty acid composition in extracted sardine oil Fatty acid compositions of crude and refined unhydrolyzed oil (wt/wt%) are shown in Table 4.2. The SFA fractions of crude and refined oils were 32.95 and 36.87%, respectively. The predominant fatty acid in both oils was 16:0, accounting for approximately 67% of each SFA fraction. The monounsaturated fatty acid (MUFA) fraction was 18.34% and 19.27% for crude and refined oil, respectively, with the predominant fatty acid being 16:ln7 followed by 18:ln9. The PUFA fractions of crude and refined oils were 48.29 and 43.17%o, respectively. Total n-3 PUFAs (18:3n3, 20:3n3, 20:5n3, and 22:6n3) accounted for 96% of the PUFA fractions of both crude and refined oil, with high levels of 20:5n3 (EPA) and 22:6n3 (DHA). A study by Shirai et al. (2002) using Sardinops melanostictus caught in Japan showed SFA fractions of 34.3-41.4%, MUFA fractions of 15.1-22.4%, and PUFA fractions of 37.3- 49.4%). These levels were in agreement with the present study, except that Shirai et al. (2002) reported a higher DHA concentration compared to EPA. Sardines caught in California (Sardinops sagax caeruleus) showed a similar ratio of EPA to DHA as in the present study; however, the California sardines had lower overall levels of EPA and DHA (Gamez-Meza et al., 1999). Sardines caught near Astoria showed higher amounts of n-3 PUFAs compared to other species such as Sardina pilchardus W caught in Spain (Beltran and Moral, 1991) and Sardinops melanostictus caught in Japan (Shirai et al., 2002). A certain amount of n-3 PUFAs may be lost due to oil refining processes that involve high temperature treatments, such as bleaching and deodorization. A study by 48

Table 4.2 Fatty acid profiles of crude and refined sardine oil (wt/wt%). Fatty acid Crude Refined C14:0 6.34±0.41a 7.38±0.15b C16:0 22.22 ±0.51a 24.97 ± 0.28b C17:0 0.33±0.01a 0.36 ± 0.00b C18:0 4.06 ± 0.15a 4.16±0.03a Total SFA 32.95a 36.87b C16:ln7 9.58±0.52a 10.68 ±0.12a C18:ln9 5.23 ± 0.09a 5.27 ± 0.06a C20:ln9 2.87 ± 0.04a 2.77±0.03a C24:ln9 0.66±0.05a 0.55 ± 0.02a Total MUFA 18.34a 19.27a C18:2n6 1.03±0.03a 0.29±0.01b C18:3n6 0.33 ± 0.02a 0.34 ± 0.02a C18:3n3 0.57±0.01a 0.62±0.01a C20:2n6 0.22±0.01a 0.20 ± 0.00a C20:3n6 0.20±0.01a 0.20 ± 0.00a C20:3n3 0.78±0.01a 0.77±0.01a C20:4n6 0.28 ± 0.04a 0.27 ± 0.04a C20:5n3 28.20 ±0.17a 26.86 ±0.17b C22:6n3 16.68 ±0.68a 13.62 ±0.16b Total PUFA 48.29a 43.17b Total n-3 PUFA 46.23a 41.87b Values in the same row labeled with a different superscript letter are significantly different (p<0.05).

Sathivel et al. (2003) reported an approximate 19% reduction in the n-3 PUFAs of catfish viscera after oil refining processes including degumming, neutralizing, bleaching and deodorization. Most of the reduction occurred at the deodorization step. In the present study, the fraction of total n-3 PUFAs decreased 49 significantly by 9.43% following the refining steps. The greatest reductions in the PUFA fraction were observed with DHA (-3.06%) and EPA (-1.34%). There were slight increases in 14:0, 16:0 and 17:0 in the refined SFA fraction compared to the crude fraction and no significant difference in fatty acid composition of the MUFA fraction.

4.4.2 Enzyme activity and degree of hydrolysis (DH) CC lipase showed the highest enzyme activity (34436 U/g lipase) among all lipases followed by CR lipase (10355 U/g lipase), while AN (2352 U/g lipase) and MJ (4741 U/g lipase) lipases showed much lower levels (Table 4.1). These results demonstrated that the non-specific lipases used in this study had higher activity than the 1,3-specific lipases in the presence of sardine oil. DHs of sardine oils from different lipases, concentrations and reaction times are shown in Fig. 4.1. Sardine oil was hydrolyzed rapidly during the initial 1.5 h for all treatments except AN 250 U. Afterwards, hydrolysis was more gradual, and only minor increases were shown throughout the 3, 6, and 9 h time periods. The highest DH (78.40%) was shown by CR 500 U lipase treatment at 9 h followed by CR 250 U (69.70%) and CC 250 U (69.33%) at 9 h. The lowest DH (12.68%) was shown by 250 U AN lipase after 9 h of hydrolysis. In general, the DH was higher in oil hydrolyzed by non-specific lipases (CR and CC) compared with 1,3-specific lipases (AN and MJ), which is in agreement with other studies on cod liver oil, salmon viscera oil and menhaden oil (Hoshino et al., 1990; Sun et al., 2002; Wanasundara and Shahidi, 1998). Wanasundara and Shahidi (1998) showed that the DH of seal blubber oil treated with 200 U CC lipase after 9 h was greater than 70%, while menhaden oil showed a lower DH. They reported that the difference was due to the presence of higher amounts of PUFA in menhaden oil than seal blubber oil. However, refined unhydrolyzed sardine oil in the present study had a higher DH and had higher amounts of EPA and DHA (26.86% and 13.62%, respectively) than seal blubber oil (4.66% EPA and 7.58% DHA) and menhaden oil (2.40% EPA and 10.06% DHA). 50

100

80 -

60 ffl 40

20

0 0

Reaction time (h)

Figure 4.1 Degree of hydrolysis (%) of hydrolyzed sardine oil by Upases at 370C (individual incubations for each lipase at each reaction time). O : CR 250 U, # : CR 500 U, O : CC 250 U, ♦ : CC 500 U, A: MJ 250 U, ▲ : MJ 500 U, □ : AN 250 U, ■ : AN 500 U.

4.4.3 Fatty acids profiles after hydrolysis All Upases caused changes in fatty acid composition, but in varying amounts. Ideally, enzymatic activity followed by FFA removal will increase EPA and DHA concentrations while reducing SFA and MUFA. Fig. 4.2 shows the changes in 16:0, 16: ln7, EPA, and DHA content in the final product obtained after lipase-catalyzed hydrolysis followed by FFA removal with neutralization. Linear regression analysis showed that EPA content was affected by enzyme and enzyme concentration, while 3 6 3 6 Reaction time (h) Reaction time (h)

3 6 3 6 Reaction time (h) Reaction time (h)

Figure 4.2 Changes in C16:0, C16:ln7, EPA, and DHA concentration in final n-3 PUFA concentrate with Upases from (a) CR, (b) CC, (c) MJ and (d) AN during hydrolysis at 370C. O : C16:0 with 250 U, ♦ :C16:0 with CR 500 U, O : C16:l with 250 U, • : C16:l with 500 U, □ : EPA with 250 U, ■ : EPA with 500 U, A : DHA with 250 U, A : DHA with 500 U. 52 DHA content was affected by enzyme and reaction time. With CR 250 U, EPA content significantly increased from 26.87 to 33.74% after 1.5 h and remained at relatively constant levels after 3, 6 and 9 h (33.45, 34.45 and 33.17%, respectively). A similar tendency was observed with CR 500 U except that EPA was present at slightly lower levels, ranging from 30.69 to 33.06%. DHA levels also increased significantly after 1.5 h, from an original of 13.63% to 23.12% with 250 U and to 23.78% with 500 U. Minor increases in DHA concentration from 24.27% to 28.02% with 250 U and 26.87% to 29.94% with 500 U occurred as hydrolysis continued to 9 h, although there was no significant difference between 1.5 h and longer reaction times. The highest total n-3 PUFA fraction (63.86%) was found in the oil hydrolyzed with 250 U CR for 6 h (data not shown). The 16:0 content decreased significantly from 24.99%) to 17.07% after 1.5 h and gradually decreased to 13.24% after 6 h with CR 250 U; the same pattern was observed with CR 500 U. Also, levels of 16:ln7 decreased significantly after 1.5 h from 10.68% to 5.11% and 5.78% with 250 U and 500 U, respectively; however the levels remained relatively constant for the rest of the hydrolysis reaction time (up to 9 h). Hydrolysis with CC lipase also resulted in high concentrations of n-3 PUFAs in the final products. EPA content increased significantly after 1.5 h to 31.91% and 31.19% with 250 U and 500 U, respectively. DHA content also increased significantly after 1.5 h to 22.65% and again after 6 and 9 h (to 26.16% and 26.54%, respectively) with 250 U O<0.05). With 500 U, DHA increased significantly to 21.14% after 1.5 h and then increased significantly at 9 h (26.32%). These EPA and DHA changes were concomitant with significant decreases of both 16:0 and 16:ln7 after 1.5 h (from 24.99% to 17.19% and 10.68% to 6.03% with 250 U, and 24.99% to 19.08% and 10.68% to 6.15% with 500 U, respectively). AN lipase was found to be less effective at increasing n-3 PUFA concentration compared with CR and CC lipases. Oil hydrolyzed by AN lipase exhibited only a minor increase in DHA content and no significant changes in EPA content. This result was in agreement with Hoshino et al. (1990). Although the changes in DHA concentration were much smaller than changes resulting from either 53 CR or CC lipases, DHA levels did increase slightly from 13.63% to 18.46% with 250 U after 1.5 h and to 18.72% with 500 U after 3 h. However, at 6 and 9 h, DHA levels remained relatively constant with no significant differences. Minor decreases in 16:0 content were observed after 1.5 h with 250 U (20.41%) and after 3 h with 500 U (20.88%). There were also reductions in 16:ln7 levels after 9 h with 250 U (8.44%) and after 3 h with 500 U (9.11%)). Despite the low activity of AN lipase in concentrating n-3 PUFAs in sardine oil, it did cause slight changes in levels of other fatty acids, such as 18:2n6, which increased from 0.19% to 1.06% and 1.11% (250 U and 500 U, respectively) after 1.5 h. The 20:3n3 content also increased slightly after 9 h, going from 0.78%) to 0.92% (data not shown). Among all lipases, MJ lipase showed the lowest effect on n-3 PUFAs concentration. No significant increases were found in EPA content by treatment with either 250 U or 500 U of this lipase. Only minor increases in DHA were observed after 1.5 h (21.34%) with 250 U and there was no significant change in DHA content with 500 U. Levels of 16:0 and 16:ln7 decreased slightly with 250 U after 1.5 h, from 24.99% to 19.70% and from 10.68% to 8.45%, respectively; however, there were no significant differences with 500 U. This study showed that lipase-catalyzed hydrolysis, especially with CR and CC lipases, resulted in successful concentration of both EPA and DHA with different rates and efficiencies depending on conditions. In general, there were significant increases in levels of both EPA and DHA after 1.5 h of hydrolysis with just slight increases afterwards, suggesting 1.5 h of reaction time may be sufficient enough to achieve fairly high levels of n-3 PUFA concentration in the final product. A further benefit is in the reduction of SFA as well as MUFA fractions, particularly 16:0 and 16:ln7. This indicates an ability of the CR and CC lipases to discriminate SFAs and MUFAs from EPA and DHA in fish oils, most likely due to the reduced steric hindrance observed with SFAs and MUFAs when linked to a glycerol backbone (Gamez-Meza et al., 2003). 54 The molecular conformation of cis carbon-carbon double bonds in PUFAs, particularly EPA and DHA, causes steric hindrance and subsequent bending of the fatty acid chains, bringing the terminal methyl groups very close to the ester bonds. Because of this steric hindrance effect, enzymatic active sites cannot reach the ester- linkages of these fatty acids with their glycerol backbones, thereby protecting EPA and DHA from lipase-catalyzed hydrolysis. However, this does not occur with the relatively straight chains of SFAs and MUFAs, and therefore hydrolysis is not hindered (Carvalho et al., 2002; Shahidi and Wanasundara, 1998). Also, it has been suggested that TGs without EPA and DHA are hydrolyzed in the first phase, and TGs with EPA and DHA are hydrolyzed later, indicating that the lipase recognizes the whole molecular structure, not only its ester bonds (Hoshino et al., 1990). In this study, non-specific Upases showed a higher ability to concentrate EPA and DHA than 1,3-specific Upases, which is in agreement with other studies (Carvalho et al., 2002; Sun et al., 2002). One reason for the high efficiency of the non-specific CR lipase in the production of an n-3 PUFA concentrate may be found in its chainlength selectivity. CR lipase is known to possess fatty acid chainlength selectivity, showing higher activity with relatively short-chain fatty acids such as C18 or below (McNeill et al., 1996). Additional possible reasons for higher efficiency of non-specific Upases might be lipase regiospecificity and/or original fatty acid profiles offish oil (Hoshino et al., 1990; McNeill et al., 1996). Preferential fatty acid hydrolysis could also depend on fish species, which may have varied positional arrangements of PUFAs. For example, the non-specific CR and CC Upases were not able to increase EPA content in salmon oil (Linder et al., 2005), salmon viscera oil (Sun et al., 2002) or Brazilian sardine oil (Carvalho et al., 2002), but did cause an increase in EPA content in seal blubber oil (Wanasundara and Shahidi, 1998) and in Pacific sardine oil (present study). Also, 1,3-specific lipase, which was found to be ineffective at n-3 PUFA concentration from sardine oil in this study, increased EPA and DHA levels in anchovy oil (Ustun et al., 1997). In the case offish oil, the second position of the glycerol moiety is usually more enriched with n-3 PUFAs (Bomscheuer, 2000), although this may vary 55 depending on species (Gamez-Meza et al., 1999). For example, stereospecific analysis has revealed that EPA is preferentially located in position 3 of the glycerol backbone in sardine oil (Ando et al., 1996). The 1,3-specific Upases possess stereospecificity with respect to the 1- and 3-positions of the acylglycerides, hydrolyzing the fatty acids at these positions easier than at the 2- position, resulting in 2-MGs and a mixture of DGs and fatty acids (Uhlig, 1998). If the 2- position is high in EPA and DHA, 1,3- specific lipase should be more effective at concentrating n-3 PUFAs; however, if SFA or MUTA are located at the 2-position and PUFAs are located at the 1- and 3-positions of triglycerides, hydrolysis with 1,3-specific lipase would most likely result in low n-3 PUFA levels in the final product (Sun et al., 2002). This could partially explain why 1,3-specific lipase was less effective at n-3 PUFA concentration compared to non- specific lipases in this study.

4.4.4 Changes in mono-, di- and triglyceride distribution and acid values after hydrolysis The changes in levels of MG, DG, and TG as well as acid values are shown in Figs. 4.3 and 4.4. The unhydrolyzed, refined sardine oil contained 86.20% TG, 13.40% DG, 0.51% MG, and had an acid value of 2.06. TG levels were significantly less in all final n-3 PUFA concentrate compared to the original sardine oil. TG levels were significantly reduced from 86.20% to 70.46% with 250 U after 1.5 h and gradually decreased to 65.94% after 9 h, although there was no significant difference between 1.5 h hydrolysis and longer reaction times on TG levels. CR lipase at 500 U showed a similar trend as at 250 U, and TG levels decreased to 70.31% at 1.5 h. DG levels increased from 13.40% to 26.89% after 1.5 h and again (to 32.33%) after 9 h hydrolysis (p<0.05). MG levels also increased after 1.5 h treatment from 0.51% to 2.08%) at 250 U, with constant levels at longer reaction times. Acid values of oil hydrolyzed with CR lipase increased from 2.06 to 83.61 and 92.11 after 1.5 h and gradually increased to 108.22 and 122.15 after 9 h by CR lipase at 250 U and at 500 U, respectively. With CC lipase, TG levels were significantly reduced from 86.20% to 66.09% after 1.5 h by 250 U and the levels gradually decreased to 63.38% at 9 h. DG 56 levels increased from 13.40% to 30.67% after 1.5 h and again increased to 3335% after 9 h with 250 U, and the same tendency was observed with 500 U. MG levels increased from 0.51% to 3.75% after 1.5 h and remained constant for the rest of the hydrolysis time with 250 U. Acid values also increased from 2.38 to 48.92 after 1.5 h and gradually increased up to 104.83 after 9 h. Hydrolysis with 500 U showed similar acid values as with 250 U (61.58 after 1.5 h, with a gradual increase up to 100.71 after 9h). In the case of the MJ lipase, TG levels decreased significantly to 70.90% after 1.5 h and again to 60.57% after 6 h, with an increase in DG content after 1.5 h (24.53%) and after 9 h (33.75%), as well as an increase in MG levels after 1.5 h (4.22%) and after 9 h (5.7%). The acid values increased from 2.06 to 70.66 and 61.99 after 9 h (250 U and 500 U, respectively), and were much lower than in oils hydrolyzed with non-specific lipases. With AN lipase at 250 U, changes in fatty acid fractions showed somewhat different patterns and had the lowest acid values among all treatments. However, acid values did increase somewhat, going from 2.06 to 26.11 and to 31.26 (at 250 U and 500 U, respectively) after 1.5 h. TG fractions decreased slightly, with an increase in DG and MG fractions, but fewer changes were observed compared to treatments with all other enzymes. 1.5 3 6 1.5 3 6 Reaction time (h) Reaction time (h)

a >■

i r 0 1.5 3 6 9 1.5 3 6 Reaction time (h) Reaction time (h) -O-TG250U -A-DG250U -O-MG250U C-250U -^-500U -TG500U -DG500U - MG 500 U

Figure 4.3 Changes in TG, DG, MG fractions of final n-3 PUFA concentrate, and acid values in hydrolyzed sardine oil with lipase from (a) CR and (b) CC during hydrolysis reaction at 370C. □: TG, 250 U, ■: TG, 500 U, A: DG, 250 U, A: DG, 500 U, O: MG, 250 U, •: MG, 500 U, O: acid value, 250 U, ♦: acid value, 500 U. 100

0 1.5 3 6 1.5 3 6 Reaction time (h) Reaction time (h)

0 1.5 3 6 1.5 3 6 Reaction time (h) Reaction time (h) -TG250U -6-DG250U -MG250U -250 U -•-SOOU -TG500U -DG500U -MG500U

Figure 4.4 Changes in TG, DG, MG fractions of final n-3 PUFA concentrate, and acid values in hydrolyzed sardine oil with lipase from (a) MJ and (b) AN during hydrolysis reaction at 370C. □: TG, 250 U, ■: TG, 500 U, A: DG, 250 U, A: DG, 00 500 U, O: MG, 250 U, •: MG, 500 U, O: acid value, 250 U, ♦: acid value, 500 U. 59 4.5 CONCLUSIONS Non-specific lipases used in this study were significantly better than 1,3- specific lipases at concentrating n-3 PUFAs from sardine oil. Hydrolysis with CR lipase and CC lipase followed by FFA removal resulted in improved total n-3 PUFAs while AN and MJ lipases had little effect on concentrating n-3 PUFAs in sardine oil. It should be noted that n-3 PUFA levels increased significantly after 1.5 h, after which point only slight increases were observed. Thus, for commercial production runs, which depend on efficiency and through-put, 1.5 h might be the optimum reaction time for producing n-3 PUFA concentrates. The majority of the concentrated product had a relatively high amount of TG fraction even after 9 h hydrolysis. Overall, analysis showed that 250 U of lipase performed equal or better than 500 U of lipase, maintaining high TG levels and increasing n-3 PUFAs content in the final concentrated products. Sardines caught in the Oregon-Washington coastal area were shown to be a good source offish oil, containing high levels of n-3 PUFAs, and lipase-catalyzed hydrolysis was demonstrated to be a feasible method for concentration of n-3 PUFAs from sardine oil for use in nutraceuticals and other products. 60

CHAPTER 5

RECOVERY AND CHARACTERIZATION OF SARDINE

OIL EXTRACTED BY pH ADJUSTMENT

Tomoko Okada and Michael T. Momssey 61 5.1 ABSTRACT Sardine oil contains high amounts of n-3 polyunsaturated fatty acids (n-3 PUP As), which have great potential applications in nutraceuticals and value-added products. A new oil extraction method involving pH adjustment was developed and compared with the conventional heat processing method in terms of lipid recovery and quality. Sardine oil was obtained either by traditional heat extraction methods or by adjusting pH to the pi of sardine muscle (pH 5.5) using HC1 or food-grade organic acids. The effect of calcium addition on oil recovery and quality was also evaluated. Use of the pH adjustment method with addition of tartaric acid (TA) and citric acid (CA) resulted in the highest oil recoveries (73.73% and 72.24%), respectively), while heat processing resulted in the lowest percent oil recovery (52.95%). pH adjustment in conjunction with HC1 treatment resulted in 59.65% oil recovery. Addition of calcium tartarate prior to pH adjustment with TA (Ca-TA) and addition of calcium citrate prior to pH adjustment with CA (Ca-CA) showed high oil recoveries (68.92% and 68.84%), respectively). Color analysis showed that oil extracted by Ca-CA had the lowest haze value (4.58), suggesting high purity in the final product. Oil extractions using TA and CA resulted in the highest levels of triglyceride (TG) followed by Ca- TA and Ca-CA treatments, with the lowest TG levels in oil extracted by heat. Fatty acid analysis showed that heat-extracted oil contained the lowest total fatty acids (513.49 mg/g oil) and resulted in the lowest total n-3 PUFAs (207.40 mg/g oil). The oil extracted by pH adjustment contained higher levels of n-3 PUFAs, ranging from 215.01 to 251.04 mg/g oil. The heat-extracted oil had the highest protein contamination, at 448 jig protein/g oil. Heat-extracted oil had high thiobarbituric acid (TBA) values as well as the highest levels of conjugated dienes and trienes, illustrating the effects of heat on lipid oxidation. This study demonstrated that a new oil extraction method by direct acidification to the pi can result in increased oil recovery and improved quality compared to the conventional heat extraction method. Keywords: Sardines, oil extraction, isoelectric point, n-3 PUFA, TBARS 62 5.2 INTRODUCTION The need for high quality fish oils is increasing due to growing nutraceutical markets as well as increased demand in the aquaculture industry. A number of methods, such as physical fractionation, low temperature solvent fractionation and supercritical fluid extraction, have been used to produce fish oils. However, wet reduction, which includes cooking, pressing and/or centrifiigation, is the traditional method used to produce fish oil (Aidos et al., 2003; Chantachum et al., 2000). The processing of pelagic fish is normally carried out by cooking the fish in large continuous cookers to temperatures of 95-100oC for 15-30 min (FAO, 1986; Sobstad, 1992). This causes coagulation of the fish muscle proteins, thereby facilitating mechanical separation of liquids and solids (FAO, 1986), and it disrupts fat cells, resulting in the release of oil into the liquid phase (Bimbo, 1990). The chemical and physical properties of edible oils depend primarily on the composition of original materials and processing temperatures (Rudan-Tasic and Klofutar, 1999). The crude oil obtained by traditional heat processing tends to be dark in color and contain a large amount of impurities. Therefore, refining steps including degumming, neutralization, bleaching, and deodorization are necessary (Bimbo, 1990). However, the high temperatures and chemical treatments utilized during the refining steps may cause isomerization of PUFAs such as eicosapentaenoic acid (EPA: 20:5n3) and docosahexaenoic acid (DHA: 22:6n3), producing trans-isomers and resulting in decreased nutritional value (Toyoshima et al., 2004). It is generally recognized that temperature influences fish oil deterioration (Aidos et al., 2002; Luten et al., 2003), as high temperatures can accelerate lipid oxidation offish oil (Al-Saghir et al., 2004). When fish oils oxidize, they produce unstable intermediary compounds, such as free radicals and hydroperoxides, which are susceptible to further decomposition into products such as aldehydes and ketones. These decomposition products adversely affect flavor, taste, nutritional value, and the overall quality offish oils. Volatile compounds such as pentanal can also develop, resulting in strong fishy odors in the final product (Aidos et al., 2002; Eymard et al., 2005). 63 The aims of oil processing are to maximize yield while maintaining high quality and to produce highly stable oils by eliminating undesirable compounds (Hidalgo and Zamora, 2006). The high levels of PUFAs in sardine oil are subject to rapid and extensive oxidation by exposure to air, light or heat during processing (Alamed et al, 2006). Oil extraction by non-heat processes would be expected to result in higher quality compared to heat-based extraction by minimizing the decomposition of PUFAs and the formation of various oxidation products. The objective of this study was to develop an oil extraction method using pH adjustment to isoelectric point (pi) and to compare this method with the heat processing method in terms of oil recovery and quality. The use of acids in pH adjustment, as well as addition of calcium, was also investigated.

5.3 MATERIALS AND METHODS 5.3.1 Fish and sample preparation Sardines {Sardinops sagax) were harvested off the Northwest Oregon coast in September 2004 by commercial purse seiners. The sardines were pumped from the seine nets directly into refrigerated seawater holds aboard the vessels and rapidly cooled to < 40C. The sardines were off-loaded within eight hours of capture at a local processing plant in Astoria, OR. Sardines were randomly chosen, packed in ice and delivered to the Oregon State University Seafood Laboratory on the day of catch. The head, viscera and backbone were removed from each fish, and skin-on fillets were obtained. The fillets were mixed together and minced with a vertical-cutter/mixer (Stephan Food Processing Technology, West Germany) for 2 min and kept in vacuum- sealed bags at -30oC for further analysis. Lipid content of sardine mince sample was determined according to Association of Official Analytical Chemists (AOAC) method 934.01 (1998). 64 5.3.2 Determination of isoelectric point (pi) of sardine muscle Citrate-phosphate buffers of varying pH (5.0-5.9) were prepared using 0.1 M citric acid (CeHgOy • H2O, J.T. Baker Chemical Co., Phillipsburg, NJ) and 0.2 M dibasic sodium phosphate (NaaHPCXj • 7H2O, Matheson Coleman & Bell, East Rutherford, NJ). Sardine mince (5 g) was mixed with citrate-phosphate buffers at varying pHs and the final volume of homogenates was brought to 25 ml. The mixture was centrifuged at 20,000 x g at 40C for 30 min, and three layers were obtained: the supernatant lipid fraction in the upper layer, an aqueous fraction in the middle and a residue fraction in the bottom layer. Soluble protein concentration in the aqueous layer was determined by the Bradford dye binding method using Bio-Rad Protein Assay kit (Bio-Rad Laboratories, Hercules, CA). The absorbance was measured at 595 nm (UV-VIS Spectrophotometer; UV 2401 PC, Shimadzu Corp., Kyoto, Japan) using bovine albumin as the standard.

5.3.3. Oil extraction Sardine mince was mixed with deionized water (1:3, v/v), and the pH was adjusted to the pi of sardine muscle protein with 2.0 M hydrochloric acid (HCl), 2.0 M tartaric acid (TA) or 2.0 M citric acid (CA). It has been reported that Ca++ ions facilitate the separation of the oil and water phases in fish mince. To test this, a separate set of mince samples were treated with either calcium tartarate (1.5% of vol.) followed by TA (Ca-TA) or calcium citrate (1.5% of vol.) followed by CA (Ca-CA) for pH adjustment. The mixtures were maintained at 50C for 2 h and oils were separated by centrifiigation at 10,000 x g for 20 min. After centrifugation, the upper and middle layers containing oil and aqueous fractions, respectively, were transferred to a separatory funnel and the oil was removed and weighed. In the control group, sardine mince was heated with deionized water (1:3, v/v) at 950C for 30 min and then centrifuged at 10,000 x g for 20 min. The resulting supernatant was transferred to a separatory funnel and the oil was removed and weighed. All extracted oils were filtered under vacuum with hardened ashless filter paper (type 542, Whatman Inc., 65 Brentford, UK) to remove contaminants and particles from the final products. Oil recovery (%) was determined by dividing the removed oil in the different phases by the total lipid content in sardine muscle. Six replicates were run, and the extracted oils were stored under nitrogen at -80oC until further analysis. Oil content in the precipitated residue after oil extraction was determined by AOAC method 934.01 (1998) to measure the oil retained in the centrifuged residues. The oil content (%) in the emulsion layer was calculated by subtracting the sum of the recovered oil and precipitated residue oil from 100%.

5.3.4 Color analysis Color analysis of extracted oils was carried out using a spectrophotometer (Hunter lab Color Quest, Hunter Associates Laboratory, Inc., Reston, VA). The color of oils was determined based on 10° standard observation and D 65 illuminants. Values were expressed as L, a, b and haze. L value indicates lightness (100: White, 0: Black), a value indicates either redness (+) or greenness (-), and b value indicates either blueness (+) or yellowness (-). Haze represents cloudiness of the products possibly as a result of particles such as pigments or contaminants. Hue angle, which represents spectral distribution (actual color), and chroma, which evaluates intensity of the color, were also calculated.

5.3.5 Thin-layer chromatography separation Extracted sardine oils were subjected to thin layer chromatography (TLC) separation for lipid composition determination. Lipids were separated into triacylglyceride (TG), monoacylglyceride (MG), diacylglyceride (DG), free fatty acids (FFA), and phospholipids (PL). Samples were dissolved in hexane (10% solution, v/v), and an aliquot of 5 p.1 sample mixture was loaded onto the silica plate (K6 silica gel 60, 200 x 200 x 0.25 mm, Whatman Inc., Brentford, UK). The developing solvent was hexane: diethylether: acetic acid (60:40:2, volume basis). Spots were visualized by spraying 5% sulfuric acid solution followed by drying at 100oC on a hot plate for 66 30 min. Quantification was carried out with densitometory ChemDoc with Quantity One software (Bio-Rad, Hercules, CA) using a commercially available TLC standard (Cat. No. 18-1A and 18-5A, Nu Check Prep, Inc., Elysian, MN), and each fraction level was expressed as wt/wt%.

5.3.6 Determination of fatty acid composition by gas chromatograpby Fatty acids in sardine oils were converted into fatty acid methyl esters (FAME) according to AOAC method 991.39 (1998), and their composition was determined by gas chromatography (GC). A Shimadzu GC-2010 (Shimadzu Corp., Kyoto, Japan) equipped with a flame-ionization detector, capillary column (Omegawax™ 250 capillary column, 30 m x 0.25 x 0.25 jim film thickness, Supelco, Bellefonte, PA) was used for analyzing FAME. Injector and detector temperatures for the GC system were set at 250oC and 270oC, respectively. The column temperature was set at 170oC with 8 min of hold time and gradually heated to 2450C at a rate of 10C/ min with 2 min of hold time. Helium was used as a carrier gas. The fatty acid concentrations were calculated by comparison of their retention times with those of the reference standards (Supelco, Bellefonte, PA).

5.3.7 Protein content determination in extracted oils Extraction of protein from oils was carried out according to a modified method of Zitouni et al. (2000). A volume of 2 ml of 50 mM phosphate-buffered saline solution was added to 5 g of extracted sardine oils and the mixture was maintained at 60oC for 24 h. After centrifugation at 1500 x g for 15 min, 2 ml aqueous layers were collected and dialyzed (dialysis membrane with 12,000 to 14,000 Dalton MW; 45 mm flat; 20 (im wall thickness; Fisher Scientific, Pittsburgh, PA) against deionized water for 24 h. The protein concentration was determined with the Lowry method (Lowry et al., 1951) using bovine albumin as the standard. 67 5.3.8 Conjugated diene and triene determination Conjugated dienes and trienes were measured as indicators of free radical production. The progress of the peroxidation reaction was estimated by measuring absorbance (Abs) at 234 nm for conjugated dienes and 268 nm for conjugated trienes using a spectrophotometer (UV 2401PC, Shimadzu Corp., Kyoto, Japan). Abs was measured using 10 mm pathlength quartz UV cuvettes with reference readings made on spectroscopic-grade n-hexane with a 1% oil dilution.

5.3.9 Thiobarbituric acid measurement Degree of oxidation based on a secondary oxidation product (malonaldehyde) was determined by the thiobarbituric acid (TBA) method (Sinnhuber and Yu, 1977). Extracted oils (approximately 0.05 g) in antioxidant solution were mixed with 3 ml of TBA reagent, and the mixture was heated in boiling water for 30 min. After cooling, 5 ml of chloroform was added and the mixture was shaked vigorously. Next, the mixture was centrifuged at 3500 rpm for 10 min and the absorbance of the upper layer was measured at 535 nm by spectrophotometer (UV 2401PC, Shimadzu Corp., Kyoto, Japan). TBA values were determined at day 0 and then every two weeks over an eight week time period. The results were expressed as mg of malonaldehyde per kg of sardine tissue.

5.3.10 Statistical analysis All analyses were run in triplicate, and the results were presented as averages and standard deviations. Statistical comparisons were made between treatments by analysis of variance (ANOVA) and Tukey's test using SPSS software. The results were presented in terms ofp values (p < 0.05). 68 5.4 RESULTS AND DISCUSSION 5.4.1 pi of sardine Solubility of sardine protein at pH 5.0 to 5.9 is shown in Fig 5.1. Solubility ranged from a high of 5.22 mg/g sardine tissue (pH 5.0) to a low of 3.77 mg/g sardine tissue (pH 5.5), indicating a pi of 5.5 for sardine muscle protein. In general, the pi of fish muscle ranges between pH 4.5 to 5.5 depending on species (Huss, 1995). Several other studies have shown similar pis for fish muscle protein. Pacific whiting and rockfish muscle protein had a pi of 5.0 (Choi and Park, 2002; Yongsawatdigul and Park, 2004), and mackerel and cod muscle protein had a pi of 5.5 (Park et al., 2003; Stefansson and Hultin, 1994). The pH shift method has been shown to be a feasible process for recovering protein from various fish species including sardines (Hultin and Kelleher, 1999). By this method, protein is first solubilized at extreme pH and then precipitated at the pi and removed by centrifugation. It was expected that this method might also be practical for oil recovery from sardines. However, preliminary experiments showed that pH shifting to either extreme acid or alkaline endpoints (pH 2-3 or 10-11, respectively) resulted in lower oil recovery due to a presence of white membrane-like emulsions formed at the oil-water interface. This was perhaps due to high emulsifier effects of partially denatured proteins exposing both hydrophilic and hydrophobic amino acid sequences. By going directly to the pi, electrostatic charges can be minimized and the proteins quickly aggregate and begin to precipitate out of solution, thereby limiting emulsion-type interactions (Damodaran, 1996; Pace et al., 2004). Because the emulsion capacity of sardine protein is minimal at the pi, separation of the oil and water phases was greatly facilitated in the present study by adjusting a sardine mince: water slurry to pH 5.5 without first solubilizing the proteins. This method resulted in high recovery and quality of sardine oil. 69

6.00

5.00 -

4.00

3.00 -

2.00

1.00

0.00

Figure 5.1 Changes in solubility of sardine muscle protein at different pH values.

5.4.2 Oil recovery and lipid distribution The oil recoveries (%) of different extraction methods are shown in Fig. 5.2. The oil extracted by adjusting pH to pi of sardine muscle protein with TA and CA showed the highest recovery (73.73% and 72.24%, respectively) followed by Ca-TA and Ca-CA. Furthermore, oil separation after adjusting pH to pi with organic acid, with or without calcium, was notably easier than in the heating process due to fewer particles and contaminants at the oil-water interface. This may be due to effective removal of emulsifiers including PL and protein residues in the pH adjustment method, demonstrating advantageous effects on emulsion breaking and clear separation of the oil and water phases. Emulsified oil can be localized, with particle-like aggregations of compatibly charged molecules existing in an incompatible aqueous phase. Substances in the slurry such as amphiphilic proteins and phospholipids can act as aids in forming emulsions. Coagulants including metal salts have been used for removal of emulsified 70 oil from large amounts of oily wastewater generated by various industries (Suzuki and T., 2005). Likewise, it was reported that addition of calcium ions in the presence of organic acid greatly facilitated removal of PLs from cod muscle homogenates (Liang and Hultin, 2005). In order to achieve high oil recovery, emulsifiers including PLs should be removed. If emulsions are already formed, addition of cations, including organic acid and calcium, as emulsion-breaking agents can disrupt the charge attraction of the hydrophobic molecules and coagulate oil droplets via charge neutralization, resulting in fewer emulsions at the oil-water interface (EPA, 2000). This was most likely the reason that pH adjustment methods demonstrated effective removal of emulsifiers including PL and protein residues, facilitating separation of the oil and water phases and allowing for higher oil recovery compared to heat extraction. In contrast, oil extraction by heat resulted in more emulsion layers at the oil-water interface, and the lowest oil recovery (52.95%) was observed. Heating the slurry causes coagulation of the sardine muscle proteins, and the liquids and solids must be mechanically separated (FAO, 1986).

100%

80%

60% □ % lipid in emulsion □ % lipid in residue 40% O % recovery

20%

0% Heat HC1 TA Ca-TA CA Ca-CA

Figure 5.2 Percent oil recovery from sardine mince and lipid distribution in centrifuge residues and emulsions. 71 The results of lipid analysis of centrifuged residues also suggest increased efficiency of oil removal with pH adjustment methods. Centrifuged residues from heat extraction contained the highest amount of lipid (31.83%)) followed by residues from HC1 treatment (27.36%) whereas pH adjustment methods with organic acids resulted in lower lipid in the residues, ranging from 21.87%) to 25.90%. Lipid content in the emulsion layer also showed significantly higher levels in heat-processed oil (15.22%) as well as in oil extracted with HC1 (12.99%) compared to oils extracted by pH adjustment with organic acids (range 4.04-5.89%)).

5.4.3 Color analysis The color of oil is a factor in determining quality because dark-colored oils require high-cost processing to achieve an acceptable light-colored product (Low et al., 1998). Color analysis (Table 5.1) indicated that all sardine oils had high L values, suggesting light color characteristics of the products. All oils had a negative a value, indicating a slight greenish color, and a positive b color, indicating yellowish colors. Chroma represents intensity of the color and also showed similar values among treatments. Hue angle values were in the range of 98.33 to 99.02°, which implies greenish-yellow color of the products. There were significant differences found in haze values among treatments. Haze represents cloudiness of the oils, perhaps as a result of particles such as pigments or contaminants. The lowest haze value (4.58) was recorded for the oil treated with Ca-CA, suggesting that this treatment results in a relatively high quality product. 72

n O r- ON m m m oo (N o o

XI XI o ro NO •<+ r- oo o O O o o o u o d d o o O ^ -H -H -H +1 -H -H CN r«-) ro ON a rt r<] in NO Os 00 00 oo od oo ON ON ON ON ON ON

M x) o O O O o o -fl4^ 2 o (U B d d d d d d -H.. -H -H .. -H -H -H s NO NO Ul -4—» 1 'sl- r^ NO u CN (N CS CS CS CS 1 M^ '•&a X o o O o o OH d d d d d o T3 Xi -H -H -H .. -H -H.. —i en (N ON 00 O ^ ro u-i rt •*-*u a cs fs cs cs cs rs u ^3 >. £> -o CO -*-»(D o o O o o o O C X r- Tt r-~ --i D (U ON 00 00 00 ON u> H o (U 1=1 '•3 S3 O o o o a, m M o <4H d d d d ■fi o h-1 -H -H -H -H CM o o\ oo <-<-> in O V .23 oo ^H o ON 00 00 'w 3. in NO NO in >n Q ^ ON ON ON ON ON 13 00 S -H IB t-i fci ,0 u C/3 . £0 (1) -I-' i—i U s U *< H < i v^ ffi X H ^ U a S3 S3 13 CO 0) U VHP s H « 13 I H > CO 73 5.4.4 Lipid composition Lipid compositions (TG, DG, MG, FFA and PL) of extracted oils are shown in Fig. 5.3. Oil extracted by heat showed significantly higher levels of FFA, DG, MG, and PL (7.29%, 3.41%, 0.99% and 1.33%, respectively) with a lower level of TG (86.98%) compared to oils extracted by pH adjustment. This is in agreement with a study on oil separation from precooked and raw skipjack tuna heads (Chantachum et al., 2000), which showed that oils prepared at high temperatures had high resulting FFA levels. The high levels of FFA as well as DG and MG in oil extracted by heat suggest that hydrolysis of the ester bonds of TGs might have occurred during the heat extraction process, as oil can undergo hydrolysis in the presence of moisture and heat. There were no significant differences found in TG levels among the pH adjustment methods, and the levels were significantly higher than those in heat extraction, ranging from 90.92% to 92.24%. PL was non-detectable in oil extracted with Ca-CA, showing efficient removal of PL from the final products. Cationic salts, such as calcium chloride, aluminum sulfate and ferric chloride, are used extensively by wastewater treatment plants for the removal of phosphorus (Galameau and Ronald, 1997; Georgantas and Grigoropoulou, 2005). Calcium ions have also been used for whey protein purification by promoting the aggregation and precipitation of phospholipoproteins that originate from fat globule membranes in cheese whey (Liang and Hultin, 2005). Cations are known to form precipitates with phosphorus, creating an insoluble material (EPA, 2000; Zvomuya et al., 2006). A study on membrane sedimentation from acid-solubilized cod protein also showed that significant amounts of membrane PL were removed from muscle homogenates by adding Ca in the presence of CA prior to solubilization at pH 3 (Liang and Hultin, 2005). This observation was similar to results of the present study, and a possible mechanism suggested was that the polycarboxylic groups of CA induce membrane removal, while Ca2+ ions allow for close contact between the vesicles through the formation of anhydrous calcium- phospholipid complexes (Liang and Hultin, 2005). 74 100.0

95.0

90.0

85.0 Heat HCl TA Ca-TA CA Ca-CA

fflTG 0FFA EDO IMG DPL

Figure 5.3 Lipid compositions of sardine oils extracted by heat and pH adjustment methods.

5.4.5 Fatty acid profiles GC analysis showed that oil extracted from sardines harvested off the OR-WA coast had high levels of n-3 PUFAs. PUFAs represented the highest oil fraction, ranging from 219.19 mg/g oil to 264.40 mg/g oil, followed by the saturated fatty acid (SFA) fraction, which ranged from 185.70 mg/g oil to 219.95 mg/g oil (Table 5.2). The major fatty acids in sardine oils were 16:0, 16:ln7, 20:5n3, and 22:6n3, with 20:5n3 being the highest, in agreement with several other studies (Bandarra et al., 1997; Leonardis and Macciola, 2004; Shirai et al., 2002). These results show relatively similar amounts of fatty acids in all extracted oils in terms of wt/wt%; however, there were differences found in terms of fatty acid mg per g oil. Oil extracted by heat and HCl showed the lowest levels of total fatty acids (513.49 mg/g oil and 515.02 mg/g oil, respectively) and n-3 PUFAs (207.40 mg/g oil and 209.54 mg/g oil, respectively). The pH adjustment procedure with CA, TA, Ca-CA and Ca- TA showed higher amounts of total fatty acids, ranging from 530.36 to 597.54 mg/g 75 oil, with n-3 PUFAs ranging from 221.18 to 361.62 mg/g oil. Oil extracted with Ca- CA contained the highest levels of 20:5n3 and 22:6n3 (146.75 mg/g oil and 63.12 mg/g oil, respectively), whereas oil extracted by heat showed significantly lower levels of 20:5n3 and 22:6n3 (123.09 mg/g oil and 50.58 mg/g oil, respectively), demonstrating reduced purity of heat-extracted oil. Edible oils are complexes of glycerides and other minor components, including hydrocarbons, waxes, sterols, alcohols, vitamins (tocopherols), pigments, terpenic acids and other contaminants such as proteins and phenols. Each of these components can alter the texture and/or flavor of the oil (Ruiz-Gutierrez and Perez-Camino, 2000; Sobstad, 1992). The higher total fatty acids content, and therefore lower contaminant levels, found in oil extracted by pH adjustment with organic acids, especially with Ca- CA, suggests higher oil quality. 76

O ON 0O ON ON < ON NO O O CM —i o o o ^ o O o o o o (N o H « -H -H -H -H -H +1 -H -H in■ IJ m NO ro oo r- ON NO m m ON NO r-- oo oo O 00 O ON r*i NO CO ^H ■>*

o O t~- m ON NO —< O P- ON ON NO NO o ON ^ m -H — NO < O o O O o O o O o o o -H +1 +1 -H -H •"* eg (N en ON 00 NO fN NO oo H fN ^ O ON o O r- 00 o •^t CO (N NO CM O

T3

NO o 00 00 o NO 00 ON NO CN NO 00 (N NO o ^ NO o ID < o o O O o o o o o o tN -H -H -H -H -H -H -H -H >. m oo m m r-- (N NO 00 cs 00 rM ON CN NO ^^ o o m oo ON — ■^t U NO CN '— is x

o in OO ON ON NO O NO ON O P^ O -H NO O U O O CN o CN —< O O p- -H 41 -H +1 4) c X 5 * ON NO in ON CN CN —i NO OO (N —; O m oo rn —< en la O -H m 00 O l> p^ NO Os -^ o in CO in CN o O C/3

r- —^^ NO (N 04 CN CN o NO CN CN p- o X n p > O p > O o > O > O p & H- ft? o to ►—* ►—» o to o OO 00 p o I*) ft? H- If i—» o i—» VO o b b OS o b U) P H- oo o o o ►— 4i. O ,_t ?f? IP f §■ to o b ^ fh o b o\ to o b »—* o £ H- H- ■■ 4^ o IP O b (—» o o o w ^ so os to ir\ ? O b ►—* P 4^ i-i oo o o o o — 4^ 0\ p o b p C* 00 so H- H- H- to H- p u> to vo ►—» H-* o b OS 4^ b '<-* b to to O b U) P O b to B? o b — o to 4^ iP H- to H- o to to -^j w o b u> o b OJ » o b i^i u> u> Oi ff o oo s- H- to b IP o O 00 00 OS --J o 4^. to ^J b 4^ VO 4i. 4^ to ^-J £ ~-J to 4^ to o o O <~n o\ OJ >-i p OJ to so -J 00 ff u H-ffH-H-H-lfH-lf H- to <1 fr ? p --J O H—t to bo o l/i o B? o 4^ o to bs oo p Lfi u , H-H-H-K-H-ffffH-lf totototototo^i-'^ totoooopoooooo ooooonnoo ^L^LJioJwtoijjLiJto 3033313333 U)OJU>OJOSO\OJOSO\ p o 4i.^btoto4i.b\\ob OS <^> 4^. Os vo to H^ u> bs OJ bs as ?OlVpOSH--~0tOU)V£> 10^4^.000000 O >-' 0\ ~ ~o OJ os OJ vo b I__I -J <-* to to 00H-000S4i>.ppptO 0\OJ^-000000 OS OJ \t O i— U\ OJ 4^ <^< to so Ln OJ UJ Ul o bo JP t» » b) §• o- o* if H- '-d to to ►—' VO ^ to ^ o to 1—* IF to u> to OJ 4* U) 1 o bo U) OJ 4s> to 4^. ^ 00 u> to to C\ ff OJ b ^o to VO to u> 1 s 4^ ? --J u> -J p §• 03 a* P O" H- 3 to o -~4 ^ fi? o Ln fr s> 4^ > ? O 4^ u^ *>. Lfi 00 b> OS K) b i—» NJ VO S) os ^o )0 ^-J -1^ K) 4^ O U) i KJ )P OJ VO K) L^i i—* o fr o t-fi OJ NJ u< 01 S- a" a> a* p l— if If U\ ►—- .fc. ■^ ' ^ U) ff l_> t^) ►—* Ln b fi? *>. (O 4^ OJ sr u> ^o LK> ji. ~-J t—• ui g VO "O b\ <% U\ U) p ON t—* l_>l -li. 00 '*> (-K VO ff VO ^J Lh 00 4 M 3 &• |. r-t- tJ* o p < $ n> 7, o 3. r-f i ■ n ft « 4 P a 3 i P cr p. S P s*. s CD H cr CT £L c* 0) M B' CO c § CD Valu es are an± i 1 13 •a A o D- a> o *+> signi ©a V}Xi ficant «< 3 iffere Z.Z. 78 5.4.6 Protein content in extracted oils Protein content of sardine oil extracted by both heat and pH adjustment methods are shown in Fig. 5.4. Oil extracted by heat showed the highest protein content (448 |^g/g oil) followed by oil extracted with HC1 (151 jag/g oil). The lowest protein content was found in oil extracted with CA (43 (xg/g oil) and Ca-CA (47 jxg/g oil) followed by oil extracted with TA and Ca-TA (80 |ig/g oil and 73 jig/g oil, respectively). Traditionally, minor oil components, such as proteins, have been considered impurities (Hidalgo and Zamora, 2006). Other nitrogen-containing intermediates including trimethylamine oxide, a hydrolysis product of phospholipids, have been identified in the oils. These intermediates react with hydroperoxides, giving formaldehyde and dimethylamine, which are responsible for fishy odors (Debruyne, 2004). In order to achieve high quality oil recovery from sardines, protein contaminants should be removed during the extraction process; however, during conventional oil extraction processes, partial proteins derived from original material (sardine mince) can be solubilized and still remain in the final product. Protein content in refined/crude oil ranges from 0.1 ug to 148 ug/g oil depending on the nature of the source and the processing method (Hidalgo and Zamora, 2006). It was demonstrated that extraction with CA and Ca-CA could result in high quality oils with less impurities, as supported by low levels of protein contamination in the final products.

-£- 400 ■

300

200 -

100 • r-n r—-i

0 • Heat HCI TA Ca-TA CA Ca-CA

Treatments

Figure 5.4 Protein content in sardine oils extracted by heat and pH adjustment methods. 79 5.4.7 Conjugated dienes and trienes determination Results of conjugated diene and triene measurements are shown in Table 5.3. Conjugated dienes and trienes are typically generated during initial oxidation processes, when cis-methyl interrupted double bonds are converted into conjugated double bonds (Vaya and Aviram, 2001). These conjugated double bonds have strong UV absorption at 234 nm and 268 nm, and therefore relative levels of conjugated dienes and trienes can be determined by measuring the Abs of the samples at these wavelengths. Conjugate dienes were found to be the highest in oil extracted by heat (Abs 1.52). Among the oils extracted by pH adjustment, the lowest values were found in oil extracted with TA (Abs 1.06) followed by oil extracted with CA (Abs 1.24). In the case of conjugated trienes, the highest value was also found in oil extracted by heat (Abs 0.25) followed by HC1 (Abs 0.23), and the lowest was in oil extracted with TA, corresponding to the results for conjugated diene levels.

Table 5.3 Absorbance of conjugated dienes and trienes in oil extracted by heat or pH adjustment methods. Conjugated dienes Conjugated trienes Heat 1.522 0.253 HC1 1.304 0.231 TA 1.065 0.128 Ca-TA 1.447 0.209 CA 1.245 0.206 Ca-CA 1.415 0.163 Abs at 234 nm Abs at 268 nm

5.4.8 TBA values At day 0, the oil extracted by heat showed the highest TBA value (12.24), while the lowest TBA value was found in the oil extracted by Ca-TA (8.50) followed by Ca-CA (9.40) (Fig. 5.5). Moreover, TBA values of oil extracted by heat steadily 80 increased with storage time and reached the highest level at day 42 (18.87). On the other hand, TBA values of oil extracted by pH adjustment remained low throughout the storage period. TBA values were significantly reduced with the addition of calcium in both the CA and TA treatments. Organic acids including CA and TA interact with metals, such as copper, iron, magnesium and calcium, to form a wide range of metallic salts that exhibit antioxidant properties in oils (Gil et al., 2000) by reducing metal-catalyzed oxidation through chelation of trace metals such as iron (Debruyne, 2004; Mattey and Kristiansen, 1999). Approaches to inhibit oxidation offish oil include maintaining minimal temperatures during processing and retaining natural antioxidants by avoiding conditions that destroy them (Hultin, 1994). Fish oil contains natural antioxidants, such as a- tocopherol, which is known to play an essential role in protecting PUFAs against oxidation by trapping free radials (McGuire et al., 1997). Radicals trapped by a- tocopherol are extremely stable, making it a highly efficient antioxidant (Berges, 1996). a-tocopherol can be destroyed by processing conditions including heat, freezing temperatures and the presence of oxygen, iron, chlorine, and/or mineral oil. Cooking of minced channel catfish resulted in a 40% loss of tocopherol content after 5 min of heating at 1770C (Erickson, 1992). A time-course study of the TBA value of meat and the effect of tocopherol in preventing oxidation demonstrated that tocopherol levels of 4.4 g/kg meat reduced TBA values by approximately 80% following a 3 month storage period (Faustman et al., 1989). The significantly high TBA values observed in oils extracted by heat in the present study might be due to partial decomposition of endogenous antioxidants such as tocopherol in sardine oil. 81

o c

-^60 n. u T3>^ (L>

60

10 20 30 40 50 Days -♦-Heat ^^HCL -TA

^-Ca-TA -•-CA -Ca-CA

Figure 5.5 Changes in TBA values of sardine oil extracted by heat and pH adjustment methods over the storage period.

5.5 CONCLUSIONS This study showed that oil extraction by pH adjustment followed by centrifugation can result in higher recovery and increased oil quality as compared to heat extraction. Oil extracted with Ca-CA showed the best quality, along with good recovery, the lowest haze value, and the highest n-3 PUFA content. All oils extracted by pH adjustment exhibited high stability against oxidation with fewer impurities compared to the heat process. There was a significant beneficial effect of calcium addition prior to pH adjustment in terms of lipid oxidation stability and color of the final products. The pH adjustment method is a novel process for oil extraction and a promising method that can be utilized for high-oil pelagic species such as sardines. 82

CHAPTER 6

PRODUCTION OF N-3 POLYUNSATURATED FATTY ACID

CONCENTRATE FROM SARDINE OIL BY IMMOBILIZED CANDIDA

RUGOSA LIPASE IN ALGINATE-CHITOSAN-CaCb BEADS

Tomoko Okada and Michael T. Morrissey 83 6.1 ABSTRACT A study was conducted to develop an immobilized-enzyme system to entrap lipase in chitosan-alginate-CaCh beads for the purpose of concentrating n-3 polyunsaturated fatty acids (n-3 PUFAs) from sardine oil. The objectives were to determine the characteristics of the system, hydrolytic activity and efficiency for n-3 PUFA concentration compared with a free lipase system. Lipase was immobilized by an ionotropic gelatin method, and beads composed of alginate-chitosan, alginate-CaC^, and alginate-chitosan-CaCh were analyzed for characteristics including mechanical strength and morphology. Immobilized lipase beads composed of alginate-chitosan- CaCb were analyzed for lipase loading efficiency, optimum pH and temperature and fatty acid profiles following hydrolysis. Lipase-immobilized beads prepared with alginate-chitosan-CaCl2 showed the highest mechanical strength. Scanning electron microscopy analysis revealed that lipase had a strong influence on bead structure. Optimum pH of immobilized lipase shifted from 7.0 to pH 6.0 and immobilized lipase showed higher stability against pH and temperature changes. Original sardine oil contained 38.13% n-3 polysunsaturated fatty acids (PUFAs) (25.20% 20:5n3 and 7.2% 22:6n3) and the concentration was significantly increased to 65.32% (40.24% 20:5n3 and 15.46% 22:6n3) with free lipase and to 64.77% (39.64% 20:5n3 and 15.33% 22:6n3) with immobilized lipase after 90 min of repeated hydrolysis. Fatty acid of the free fatty acid (FFA) fraction of hydrolyzed oil showed that lipase preferably hydrolyzed 16:0, 16:ln7 and 18:0 accounting for 76.56% and 69.51% of total FFAs (after first and second hydrolysis, respectively). Among different lipid fractions, the diglyceride fraction contained the highest n-3 PUFAs. This study shows that use of immobilized lipase systems for increasing n-3 PUFA concentration in sardine oil provides new process opportunities for the industry.

Keywords: sardine oil, chitosan, sodium alginate, enzyme immobilization, lipase 84 6.2 INTRODUCTION With growing public awareness of the clinical benefits of n-3 polysunsaturated fatty acids (PUFAs), researchers have been developing techniques for production of n- 3 PUFA concentrate via a variety of methods. These methods include supercritical extraction with CO2 (Hsieh et al., 2005; Riha and Brunner, 2000), fractional solvent crystallization (Chen and Ju, 2000; Chen and Ju, 2002; Ganga et al., 1998), urea-complexation (Adachi et al., 1993; Guil-Guerrero and Belarbi, 2001; Hwang and Liang, 2001; Wanasundara and Shahidi, 1999), esterification (Akimoto et al., 2003; Haraldsson et al., 1997; Robles et al., 1999; Schmitt-Rozieres et al., 2000), hydrolysis (Linder et al., 2005; Rakshit et al., 2000; Sun et al., 2002; Ustun et al, 1997; Zheng et al., 2005), chromatography techniques such as supercritical fluid chromatography and silver-resin chromatography (Alkio et al., 2000; Nieto et al., 1997), acidolysis (Adachi et al., 1993; Garcia et al., 2000; Gonzalez Moreno et al., 2004; Ramirez Fajardo et al., 2003), and any combination of such methods (Adachi et al., 1993; Gamez-Meza et al., 2003; Ganga et al., 1998). The use of lipases as biocatalysts has considerable merits, including superior efficacy under mild reaction conditions, reduced environmental pollution, high availability from a wide range of sources, and high adaptability for modifications (Xuebing, 2000). Due to its promising advantages, lipase-catalyzed enzymatic hydrolysis has received increasing attention as a currently available method for n-3 PUFA concentrate production. Hydrolysis reactions are generally faster compared to other concentration methods such as esterification and interesterification (Mohapatra and Hsu, 1999). Also, the process outlined in this study is relatively simple, can be carried out under mild conditions of pH and temperature, and no organic solvent is required. The majority of studies on production of n-3 PUFA concentrate from fish oil by lipase-catalyzed hydrolysis utilize commercially available free lipases derived from microorganisms such as Candida, aspergillus, mucor species and others (Hoshino et al., 1990; Linder et al., 2005; Rakshit et al., 2000; Sun et al., 2002; Wanasundara and Shahidi, 1998; Zheng et al., 2005). However, use of enzymes in chemical and 85 biochemical reactions, as well as their disposal following each application can be costly. Also, the recovery and reuse of enzymes from the reaction medium can be difficult. For these reasons, enzyme immobilization has attracted a wide range of interest from fundamental academic research to industrial applications (Betigeri and Neau, 2002). Immobilized enzymes have several known advantages over free enzymes: 1) the enzyme can be reused, 2) the process can be readily controlled, 3) enzyme properties (activity and thermostability) can be favorably altered, 4) the process is more cost effective, and 5) immobilized enzyme has more adaptability to a variety of engineering designs (Benjamin and Pandey, 1998; Hayashi and Ikada, 1991). Entrapment of enzymes can be defined as the coupling of enzymes to the lattices of a polymer matrix or enclosing them in semi-permeable membranes tight enough to prevent the leakage of enzymes while allowing the diffusion (mass transfer) of substrates and products (Benjamin and Pandey, 1998). Enzymes can be immobilized by a variety of methods broadly categorized as either physical (weak interactions between support and the enzyme) or chemical (covalent bonds between support and the enzyme) (Berger et al., 2004). The ionotropic gelatin method (or coacervation) is a physical enzyme immobilization technique (Krajewska, 2004). It is a simple procedure carried out under mild conditions and without the often-toxic auxiliary molecules (e.g., glutaraldehyde) used in the covalent crosslinking method (chemical method) (Berger et al., 2004). This technique is carried out by preparing an enzyme-containing anionic polyelectrolyte solution prior to gelation, and then adding the solution dropwise to a cationic polyelectrolyte solution. Consequently, the enzyme is immobilized by inclusion in the interior of the beads/capsules (Krajewska, 2004). This method has been successfully used for the immobilization of enzymes and cells, encapsulation of nutraceuticals and for controlled drug release (Dumitriu et al., 1994; Il'ina and Varlamov, 2005; Shahidi etal., 1999). It has been reported that the amino groups of chitosan, when present in solutions in a cationic polyelectrolyte form, are capable of interacting with an anionic 86 polymer that has carboxylic groups, such as sodium alginate, by ionic binding (Coppi et al., 2002; Il'ina and Varlamov, 2005). The mechanism of lipase immobilization by an ionotropic gelatin method utilizing alginate and chitosan is depicted in Fig. 6.1. Chitosan is a natural polymer that has gained tremendous interest for use in immobilization technologies (Krajewska, 2004). It is nontoxic (food grade), available at low costs and in different forms (powder, gel, fibers and membranes), easy to handle, easily derived, and has high protein affinity (Gomes et al., 2004; Pereira et al., 2003). Chitosan, (l-^4)-2-amino-2-deoxy-P-D-glucan, is commercially produced from chitin, the second most abundant polysaccharide in nature, by alkali deacetylation. The presence of amino groups in chitosan causes the molecule to be a cationic polyelectrolyte (pKa ~ 6.5) and ultimately gives rise to unique properties amenable to chemical modifications (Baxter et al., 2005; Krajewska, 2004; Lopez- Caballero et al., 2005).

Anionic polyelectrolyte solution (sodium-alginate) containing lipase

Crosslinking between carboxyl group of a-L- gluluronic acid and

Cationic polyelectrolyte © Negative charge of alginate solution © Positive charge of chitosan — Chitosan polymer (pH< 6.0) Figure 6.1 Lipase immobilization by an ionotropic gelatin method utilizing chitosan and alginate.

In addition to chitosan, alginate is a widely utilized polymer for enzyme immobilization (Hari et al., 1996). It is non-toxic, common food additive and 87 commercially available in a variety of sodium salt forms. Alginate is a seaweed extract composed of chains of alternating a-L-gluluronic acid and P-D-mannuronic acid residues. Alginate encapsulation has been reported to be suitable for many biological compounds due to advantages including mild (room temperature) encapsulation conditions and high gel porosity allowing for rapid diffusion rates of macromolecules (Srivastava et al., 2005). Because alginate has a pKa of 3.8, it is negatively charged at neutral pH and has high affinity for polycations (Breguet et al., 2005). Therefore, a strong ionic network can be developed in the presence of Ca2+ or other multivalent cations at neutral pH (Hari et al., 1996). Previously reported immobilized enzymes utilizing both alginate and chitosan include lipase (Won et al., 2005), p-galactosidase (Taqieddin and Amiji, 2004), urease (DeGroot and Neufeld, 2001), tannase (Boadi and Neufeld, 2001; Yu et al., 2004), rhodanese (De Riso et al., 1996), aminoacylase (Lee et al., 1992), invertase (Gomez et al., 2006), and amylase (Shukla and Jajpura, 2005). However, no studies have been published investigating production of n-3 PUFA utilizing immobilized lipase in chitosan-alginate-CaC^ beads. This paper presents the development of an immobilized-enzyme system to entrap lipase in chitosan-alginate-CaC^ beads for the purpose of concentrating n-3 PUFAs from sardine oil. Lipase from Candida rugosa (CR lipase) (Amano Enzyme, Aichi, Japan) was used because a prior study showed high efficiency in concentrating n-3 PUFA using the free form of the lipase (Chapter 4). The objectives were to determine the characteristics of the system, hydrolytic activity and efficiency for n-3 PUFA concentration compared with a free CR lipase system. The effect of repeated hydrolysis with first (1st) and second (2nd) hydrolysis was also studied.

6.3 MATERIALS AND METHODS 6.3.1 Sample preparation and oil extraction Sardines (Sardinops sagax) were harvested off the Northwest Oregon coast in September 2005 by commercial purse seiners. The sardines were pumped from the seine nets directly into refrigerated seawater holds aboard the vessels and rapidly 88 cooled to < 40C. The sardines were off-loaded within 8 hr of capture at a local processing plant in Astoria. Sardines were randomly chosen, packed in ice and delivered to the laboratory on the day of catch. The head, viscera and backbone were removed from each fish, and skin-on fillets were obtained. The fillets were mixed together and minced using a vertical-cutter/mixer (Stephan Food Processing Technology, West Germany) for 2 min. The minced sardines were mixed with deionized water (3:1, v/v), and pH was adjusted to the isoelectric point of sardine muscle protein (5.5) by 2.0 M citric acid with addition of calcium citrate (1.5% of vol.) prior to pH adjustment. The mixture was kept at 50C for 2 h, transferred to a separatory funnel after centrifugation at 10,000 * g for 20 min, and then oil was recovered and filtered under vaccum with hardened ashless filter paper (type 542, Whatman Inc., Brentford, UK). Tertiary butyl hydroquinone (200 ppm) was added to the extracted oil as an antioxidant with constant agitation with N2. The oil was collected in 5 ml vials and the air in the vials was replaced by N2. The vials were stored at -80oC until further use.

6.3.2 Preparation of immobilized Upases Immobilized lipases were prepared by an ionotropic gelatin method. Alginate solution was prepared with different concentrations of sodium alginate (0.5, 1.0, 1.5, and 2.0%) in deionized water (w/v). The amount of 350 U CR lipase (0.07 g) was selected for use according to the results of a preliminary experiment. CR lipase was dissolved in 7 g of 2.0% alginate solution and solubilized by constant agitation for approximately 15 min. Chitosan solution (0.5%) was prepared by dissolving high molecular weight (MW) chitosan (degree of deacetylation (DD): >85%, MW: 1.0 x 106, Kyowa, Tokyo, Japan) in 3.0% acetic acid solution. The alginate solution containing CR lipase was added dropwise into chitosan solution with/without 1.0 or 2.0% CaCh. For alginate-CaCh beads, CR lipase-containing alginate solution was added dropwise into either 1.0 or 2.0% CaCh solution. After rapid curing, the lipase- immobilized beads were obtained by filtration and immediately used for further analysis. 89 6.3.3 Measurement of mechanical strength of immobilized Upases The mechanical strength of lipase-immobilized beads made with alginate- chitosan, alginate-CaCb, or alginate-chitosan-CaCh was determined by a texture analyzer (TA.XT.Plus, Scarsdale, NY) equipped with a 50 kg load cell (Khan et al., 2000). Solutions containing 0.5% alginate, 0.5% chitosan and either 1.0% or 2.0% CaCh were used for this assay. Each bead was held between two clamps separated by a distance of 2 cm. The beads were pulled by the top clamp at a rate of 1 mm/s, and the force (g) was measured when the beads were broken.

6.3.4 Scanning electron microscopy analysis Scanning electron microscopy (SEM) analysis of the surface and cross-sections of the dried beads prepared with alginate-chitosan, alginate-CaCb, or alginate- chitosan-CaC^ with/without lipase was performed with an AMR - 3300 SEM (Amray Instruments, Bedford, MA). All samples were chemically dried according to Nation (1983) by transferring alginate beads for 5 min to 1.0% glutaraldehyde in 0.1 M cacodylate buffer (pH 7) followed by washing in deionized water for 5 min. Beads were dehydrated with ethanol solutions of 70%, 85%, 95% and 100% for 5 min each, then immersed in hexamethyldisilazane for 5 min, followed by air drying at room temperature.

6.3.5 Determination of lipase loading efficiency Effect of alginate concentration (0.5, 1.0, 1.5, and 2.0%) on lipase loading efficiency (percent of total lipase entrapped) was investigated. Immobilization of lipase in alginate-chitosan-1.0% CaC^ beads was selected for further analysis based on its high mechanical strength as well as high efficiency in concentrating n-3 PUFAs (data not shown). CR lipase (350 U) was immobilized in alginate-chitosan-1.0% CaCl2 beads at different alginate concentrations, and loading efficiency was determined by measuring protein content in beads and chitosan solution by the Lowry method (Lowry et al., 1951) using bovine albumin as a standard. 90 6.3.6 Determination of lipase activities at different pH values and temperatures Free and immobilized lipase hydrolytic activities were measured according to the Japanese Industrial standard method (JIS, 1995) at different pH values in the range of 2.0 to 10.0 at 370C. The effect of temperature on both free and immobilized lipase activity at optimum pH values (pH 7.0 for free lipase, and pH 6.0 for immobilized lipase) was also determined within a temperature range of 20-60oC at 50C intervals. Results were expressed as relative activities where the highest original activities were defined as 100%.

6.3.7 Hydrolysis reaction A 20 ml aliquot of phosphate buffer (0.1 M) and 2 g extracted sardine oils were added to a flask along with either free or immobilized lipase. The hydrolysis reaction was run at the pre-determined optimum temperatures and pHs of free or immobilized lipase (350 U/g sardine oil). Following hydrolysis, deionized water (10- fold), potassium hydroxide and 1.0% methanolic phenolphthalein solution were added to the flask. The potassium hydroxide was added for neutralization of free fatty acids (FFA) released by hydrolysis, and the 1.0% methanolic phenolphthalein solution was added to monitor pH. The final n-3 PUFA concentrate was obtained after centrifugation at 14,000 rpm for 5 min. After the 1st hydrolysis for 30, 60 or 90 min, immobilized lipase was removed via filtration, and was reused for the 2nd hydrolysis (30, 60 or 90 min). Unrecoverable free lipase was discarded with the water phase; therefore, fresh free lipase was added to the new oil -buffer solution after removing FFA fractions for a 2nd hydrolysis reaction.

6.3.8 Determination of fatty acid composition by gas chromatography After the 2nd hydrolysis, the fatty acids in original sardine oil and final n-3 PUFA concentrate were converted into fatty acid methyl esters (FAME) according to American Oil Chemists' Society method 991.39 (AOCS, 1998), and their composition was determined by gas chromatography (GC). A Shimadzu GC-2010 (Shimadzu Corp., Kyoto, Japan), equipped with a flame-ionization detector and capillary column 91 (Omegawax™ 250 capillary column, 30 m x 0.25 x 0.25 jxm film thickness, Supelco, Bellefonte, PA) was used for analyzing FAME. GC system parameters were set as follows: injector and detector temperatures of 250oC and 270oC, respectively; column temperature of 170oC with 8 min hold time; gradual heating to 2450C at a rate of 10C/ min with 2 min hold time; and helium as a carrier gas. The fatty acid concentrations were calculated by comparison of their retention times with those of the reference standards (Supelco, Bellefonte, PA).

6.3.9 Thin-layer chromatography separation Original sardine oil and final n-3 PUFA concentrate were subjected to thin layer chromatography (TLC) separation for lipid composition determination. Lipids were separated into monoglyceride (MG), diglyceride (DG), and triglyceride (TG) fractions. Samples were dissolved in hexane (10% solution, v/v), and an aliquot of 5 |xl sample mixture was loaded onto a silica plate (K6 silica gel 60, Whatman Inc, Park, NJ). The developing solvent was hexane: diethylether: acetic acid (60:40:2, volume basis). Spots were visualized by spraying 5% sulfuric acid solution followed by drying at 100oC on a hot plate for 30 min. Quantification was carried out with a densitometory ChemDoc with Quantity One software (Bio-Rad, Hercules, CA) using a commercially available TLC standard (Cat. No. 18-1A and 18-5A, Nu Check Prep, Inc., Elysian, MN), and each fraction level was expressed as wt/wt%. Fatty acid profiles in FFA fraction recovered from oil hydrolyzed by CR lipase (after both the 1st and 2nd 90 min hydrolyses) were also determined to investigate which fatty acids were favorably hydrolyzed by immobilized CR lipase. To determine which lipid fraction contained the highest levels of n-3 PUFAs, different lipid fractions (TG, DG, and MG) of final n-3 PUFA concentrate treated with immobilized CR lipase (after 2nd hydrolysis of 90 min) were separated by TLC and visualized with iodine vapor. Each fraction was removed from the TLC plates and converted into FAME, and fatty acid profiles in each fraction were determined by GC. 92 6.3.10 Statistical analysis All analyses were run in triplicate, and the results were presented as average and standard deviations. Statistical comparisons were made between treatments by ANOVA and Tukey's test using SPSS software package. The results were presented in terms ofp values (p < 0.05).

6.4 RESULTS AND DISCUSSION 6.4.1 Mechanical strength of immobilized Upases As shown in Fig. 6.2, lipase-immobilized beads prepared with alginate- chitosan had the lowest breaking force (2.90 g). The low mechanical strength of these beads was most likely due to their liquid core. Because the liquid core was too weak to support the integrity of the capsules against mechanical stress during the hydrolysis reaction, alginate-chitosan beads were deemed unsuitable for use in batch production of n-3 PUFA concentrate. Alginate-CaCh beads showed higher breaking force value than alginate-chitosan beads; however, their mechanical strength was not strong enough and they were easily broken down during the hydrolysis reaction when constant agitation resulted in leakage of the lipase. Among the different preparations, lipase-immobilized beads prepared with alginate-chitosan-CaCb had the highest breaking force (3.35g). The addition of CaC^ to alginate-chitosan beads might have induced gelling of the liquid core, thereby increasing the breaking force of the beads (Gasered et al., 1998). The results showed that gel strength was not increased by a change of CaCh concentration from 1.0% to 2.0%, suggesting that addition of Ca2+ ions above a certain level does not strengthen the formation of gel-like networks. These findings are in agreement with previous reports (Won et al., 2005). 93

■39

o

Al-Ch Al-CaC12 Al-CaC12 Al-Ch-CaC12 Al-Ch-CaC12 (1%) (2%) (1%) (2%)

Figure 6.2 Mechanical strength of the CR lipase-immobilized beads prepared with 0.5% alginate-0.5%chitosan (Al-Ch), 0.5% alginate-1.0% CaCfc (Al-CaCfe 1%), 0.5% alginate-2.0% CaCb (Al-CaCh 2%), 0.5% alginate-0.5% chitosan-1.0% CaCh (Al- Ch-CaCh 1%) or 0.5% alginate-0.5% chitosan-2.0% CaCh (Al-Ch-CaCh 2%).

6.4.2 Morphology of the beads SEM images of beads prepared with alginate-chitosan, alginate-CaC^ or alginate-chitosan-CaCU with/without lipase are shown in Fig. 6.3. In general, the beads were about 1-2 mm in size and relatively ovoid or egg-shaped. However, the alginate-CaCl2 beads were very spherical in shape. Alginate-CaCh beads also had a distinct surface image that was clear and smooth compared to the mesh-like surface of chitosan-coated beads. In general, the shape of the beads changed to a spherical disc with a collapsed center after chemical drying. According to previous reports, calcium alginate beads usually have a heterogeneous structure with a dense surface layer and are loosely cored due to the heterogeneous gelation mechanism, which results in the collapse of beads during the drying process (Sankalia et al., 2005). Alginate-chitosan beads posses a highly porous core, resulting from drying of liquid alginate (Taqieddin and Amiji, 2004), and a uniform chitosan layer unlike other beads. 94

Figure 6.3 SEM micrographs of alginate beads prepared with (a) 2.0% alginate + 0.5 % chitosan, (b) 2.0% alginate + 1.0% CaCh (c) 2.0% alginate + 0.5% chitosan + 1.0% CaCh, (d) 2.0% alginate + 0.5% chitosan +1.0% CaCh + CR lipase. 95 On the other hand, all other preparations formed a semi-solid gel core, shown as a compact dense structure surrounded by a chitosan layer (Fig. 6.3b, c and d). Lipase seemed to have a strong influence on bead structure. The structure of the alginate-chitosan-CaC^ bead without lipase (Fig. 6.3c) was more compact than the structure of alginate-chitosan-CaCl2 bead with lipase (Fig. 6.3d). This implies that lipase increases the porosity of the complex, which is in agreement with Magnin et al.

6.4.3 Lipase loading efficiency Because immobilized lipase prepared with alginate-0.5% chitosan-1.0% C&Ch exhibited the highest mechanical strength, this mixture was selected for lipase loading efficiency tests. Different alginate concentrations (0.5 to 2.0%) were used in order to find an optimum alginate concentration. Lipase loading efficiency was determined by measuring the percent lipase retained in the beads and the percent lost into the chitosan solution (Fig. 6.4). The results showed that percent lipase loading increased as alginate concentration increased. However, alginate concentrations higher than 2.0% are not practical for use in lipase immobilization because they cause strong gelation resulting in solutions that are too thick to stir. The highest percent lipase loading (87.41%) was found in beads prepared with 2.0% sodium alginate. Alginate anions are cross-linked to Ca cations and chitosan, and it was expected that an increase in alginate concentration would make it difficult for CR lipase (MW= 5.7 x 104) to leak out of the gel network, as suggested by a similar study (Won et al., 2005). Furthermore, a preliminary experiment showed that resulting levels of n-3 PUFA concentrate were higher when 2.0% alginate was used compared to 1.0% or 1.5% alginate concentration; therefore, immobilized lipase prepared with 2.0% alginate was used in all further analyses including determination of optimum pH, temperature, and hydrolysis reaction. 96

□ lipase lost into the chitosan solution M lipase retained in the beads

0.5 1.0 1.5 2.0 Alginate (%)

Figure 6.4 Percent lipase retained in alginate-chitosan-CaC^ beads and percent released into chitosan solution.

In general, the main disadvantage to enzyme immobilization is the low retention yield (about 60%) of enzymes successfully loaded into beads (Dumitriu and Chomet, 1997). Alsarra et al. (2002) reported that lipase loading efficiency ranged from 23.9% to 44.3% depending on the DD and MW of the chitosan used, and that higher DD and MW provided a higher lipase loading efficiency. It was suggested that this was because chitosan samples with higher DD would have more positively- charged amine groups upon being dissolved in the acetic acid solution; therefore, more positive charges would be available to interact with negatively-charged ions, leading to the formation of more ionic bonds and a more complex intricate network than that of lower DD chitosan. This might be the reason for the high lipase loading efficiency shown in this study as the chitosan used possesses not only high DD (>85%) but also high MW (1.0 x 106). Lipase loss of only 12.59% was found in the bead prepared with 2.0%) alginate concentration. Enzyme loss during the immobilization process has been reported by other studies, and it generally varies from 33%) to 10% depending on immobilization conditions (Betigeri and Neau, 2002). 97 6.4.4 Optimum pH of free and immobilized Upases The effect of pH on the hydrolytic activity of free and immobilized lipase is depicted in Fig 6.5. The optimum pH of free lipase was 7.0, in agreement with another study (Pereira et al., 2001) as well as a supplier's report (Amano Enzyme, Aichi, Japan). Although initial enzyme activity of free lipase was higher than that of immobilized lipase, the results showed that relative activities of immobilized lipase possessed higher stability against pH changes. The decreased activity of immobilized lipase might be due to interactions between the lipase and Ca2+ ions (Lee et al., 1993). Optimum pH of immobilized lipase (pH 6) shifted slightly from that of free lipase (pH 7.0). This was in agreement with other studies (Gomes et al., 2004; Pereira et al., 2003), and it has been suggested that this is most likely due to the positively-charged amino groups of chitosan (pKa=6.5). Immobilized lipase maintained high levels of relative activity even at extreme pH values (approximately 60% relative activity at both pH 2 and 10), whereas activity of free lipase was significantly decreased at pH lower than 4.0 and higher than 8.0 (6.55% and 1.33%, respectively). According to the results, lower pH values (less than 4.0) should be avoided in the free lipase system: the pi of CR lipase is at about 4.5, and the pH of the aqueous phase should be maintained above 5.5 to prevent serious desorption of the enzyme in the free lipase system (Minovska et al., 2005). By maintaining activity over a wide pH range, the immobilized lipase provides a significant advantage over the free lipase system for the large-scale concentration of n-3 PUFAs. 100

80

> £o 60 A -•— Free lipase a -a— Immobilized lipase ■S 40

* 20

0 6 7 8 10 pH

Figure 6.5 Effect of reaction pH values on hydrolytic activities of free lipase and immobilized lipase. The highest original activity of each system was defined as 100%.

00 99 6.4.5 Optimum temperature of free and immobilized Upases The temperature dependence of the hydrolytic reaction with free and immobilized lipase is shown in Fig. 6.6. Within the temperature range measured (20- 60oC), immobilized lipase showed higher stability than free lipase, especially as temperatures increased: at 60oC, the relative activity of immobilized lipase was 84%, compared to 46% relative activity of free lipase. The optimum temperature range of immobilized lipase was broader (35-40oC), and temperature changes had less influence on the immobilized lipase compared with that of free lipase throughout the temperature range measured. According to these results, the immobilization procedure results in higher temperature stability. Ability to operate under a wide temperature range is a highly desirable property within the food industry: higher optimum temperatures reduce the risk of microbial contamination, while lower temperatures require less energy input (Minovska et al., 2005)

6.4.6 Fatty acid composition of oils after 1st and 2nd hydrolysis The original sardine oil contained 23.06% 16:0, 12.01% 16:ln7, 25.21% 20:5n3 and 7.18% 22:6n3, with 38.13% total n-3 PUFAs. The levels of 16:0 and 16: ln7 in sardine oil after hydrolysis followed by FFA removal were significantly decreased as reaction time progressed with both free and immobilized lipase (Fig. 6.7a, b). Levels of 20:5n3 in the final n-3 PUFA concentrate were significantly increased (Fig. 6.7c,) after the 1st free lipase hydrolysis at 30 min (30.2%) and 90 min (37.23%). After two 30 min free lipase hydrolyses, the same EPA levels were achieved as with one 90 min free lipase hydrolysis. The levels were significantly increased after 60 and 90 min (38.71% and 40.24%, respectively) of the 1st hydrolysis. After the 1st immobilized lipase hydrolysis follwed by FFA removal, 20:5n3 content was significantly increased to 32.02% and 35.76%) after 60 min and 90 min, respectively (Fig. 6.7d). In the 2nd hydrolysis 20:5n3 content increased to 31.01%, 37.00% and 39.64%) after 30, 60, and 90 min reaction times, respectively. Free Ipase O cd Immobilized lq)ase >

Pi

20 25 30 35 40 45 50 55 60 Temperature (0C)

Figure 6.6 Effect of reaction temperatures on hydrolytic activities of free lipase and immobilized lipase. The highest original activity of each system was defined as 100%.

o o 101 Levels of 22:6n3 also increased significantly (to 9.25%) after the 1st 30 min free lipase hydrolysis and increased again to reach a maximum of (15.02%) after 90 min. In the immobilized lipase hydrolysis, 22:6n3 content increased to 9.50% after 30 min and reached 14.09%) after 90 min. There was only a slight difference in 22:6n3 content between the 1st and 2nd hydrolyses with both free and immobilized lipases; however, a repeated hydrolysis (2nd hydrolysis) resulted in significantly higher levels of n-3 PUFAs, which increased from 61.97%) to 65.32%) with free lipase and from 59.18%) to 64.77%) with immobilized lipase. This was mainly due to an increase in 22:6n3 and other n-3 PUFAs. Although the final fatty acid concentrations were similar for both free and immobilized lipase after a 90 min hydrolysis, the rate of changes in fatty acid composition, including levels of 20:5n3 and 22:6n3, was relatively slower with the immobilized lipase system than with the free lipase system. The effect of repeated hydrolysis on enrichment of PUFA in tuna oil using lipase from Geotrichum candidum was reported by Shimada et al. (1994). The study indicated that a 2nd hydrolysis did not show a significant effect when the reaction was continued with the same amount of lipase without removing FFA from the final product; however after removing FFA, degree of hydrolysis was increased by lipase addition, in agreement with this study. From the results, it should be noted that removal of FFA released by hydrolysis after reaching the maximum level of n-3 PUFA level is essential for achieving a higher n-3 PUFA concentration. This could be due to an inhibitory effect of FFA on lipase activity. 102

30 60 30 60 Reaction time (min) Reaction time (min) B Free 1st C16:0 B Free 2 ndC16:0 ■ Immobilized 1st C16:0 B Immobilized 2nd CI6:0 D Free 1st C16:ln7 Q Free 2 ndCI6:ln7 D Immobilized lstCI6:ln7 Q Immobilized 2nd CI6:ln7

r

30 60 30 60 Reaction time (min) Reaction time (min) ■ Free 1st C20:5n3 19 Free 2nd C20:5n3 G Immobilized 1st C20:Sn3 S Immobilized 2nd C20:5n3 Q Free I st C22:6n3 £3 Free 2nd C22:6n3 D Immobilized 1st C22:6n3 EJ Immobilized 2nd C22:6n3

Figure 6.7 Changes in 16:0 and 16:ln7 in final n-3 PUFA concentrate hydrolyzed (1st and 2n hydrolysis) with (a) free lipase or (b) immobilized lipase, and 20:5n3 and 22:6n3 in oils hydrolyzed (1st and 2nd hydrolysis) by (c) free lipase or (d) immobilized lipase

6.4.7 Fatty acid composition in free fatty acid fraction of hydrolyzed oil Fatty acid composition in the FFA fraction of hydrolyzed oil with immobilized lipase (90 min reaction time for both 1 st and 2nd hydrolyses) was investigated in order to determine which fatty acids were hydrolyzed preferably by CR lipase (Table 6.1). The FFA fraction was obtained by TLC after hydrolysis without neutralization, and the fatty acids were converted to FAME. With the 1st hydrolysis, the result suggested that the saturated fatty acid (SFA) fraction was the highest (61.29%) containing 103 34.19% 16:0, 16.55% 18:0 and 8.79% 14:0, in the FFA fraction of hydrolyzed oil. The total monounsaturated fatty acid (MUFA) fraction was 36.58%, containing 25.82%) 16:ln7 and 10.02% 18:ln9. Compared with these fractions, the PUFA fraction showed significantly low levels (2.12%), with 18:2n6 being the highest (0.88%)) followed by 18:3n6 (0.58%). A similar tendency was found after the 2nd hydrolysis, with the SFA fraction becoming significantly higher (71.44%) and 16:0 being the most prevalent (37.04%), while the MUFA fraction levels were found to be lower than that in oil obtained from the 1st hydrolysis. Levels of 22:6n3 and 22:5n3 increased slightly from 0.35%) to 0.55% and 0.06% to 1.29%), respectively, with an overall increase in total PUFAs (2.12% to 4.26%) in the FFA fraction of oil after a 2nd hydrolysis. These results show that n-3 PUFAs, such as 22:5n3, that are present at low levels could possibly be hydrolyzed with prolonged hydrolysis reactions. Taken together, the results from Fig. 6.7 and Table 6.1 show that enrichment of n-3 PUFAs was accompanied by a reduction in fatty acid levels in the SFA and MUFA fractions, particularly in the case of 16:0, 16:ln7 and 18:0. A possible reason for the initial reduction of SFA and MUFA may be that TG without n-3 PUFAs is hydrolyzed in the first phase, and TG with n-3 PUFAs is hydrolyzed later. This indicates that the lipase recognizes the whole molecular structure, not only the ester bonds as suggested by Hoshino et al. (1990). 104

Table 6.1 Fatty acid profile of free fatty acid fraction of sardine oils hydrolyzed with immobilized CR lipase. Hydrolyzed oil3 Fatty acid wt/wt% 1st hydrolysis 2nd hydrolysis 14:0 8.79 ±0.03 15.51 ±0.08 15:0 0.19 ±0.01 0.12±0.11 16:0 34.19 ±0.42 37.04 ±0.12 17:0 0.45 ±0.19 0.25 ± 0.04 18:0 16.55 ±1.39 16.57 ±0.59 SFA 61.29 ±2.27 71.44±1.31 16:ln7 25.82 ±0.54 15.90 ±0.28 17:ln9 0.38 ±0.01 0.17 ±0.01 18:ln9 10.02 ±0.35 7.36 ±0.27 20:ln9 1.27 ±0.78 2.91 ±0.09 22:ln9 0.00 ± 0.00 0.28 ± 0.01 MUFA 36.58 ±1.96 24.30 ±1.89 18:2n6 0.88 ±0.00 0.34 ±0.00 18:3n6 0.58 ±0.02 0.57 ±0.02 18:3n3 0.06 ± 0.00 0.37 ±0.14 20:2n6 0.07 ± 0.00 0.10 ±0.00 20:3n6 0.09 ± 0.00 0.04 ± 0.00 20:3n3 0.05 ± 0.00 0.46 ± 0.01 20:5n3 0.20 ± 0.00 0.15 ±0.01 22:5n3 0.06 ± 0.02 1.29 ±0.04 22:6n3 0.35 ±0.00 0.55 ±0.03 PUFA 2.12 ±0.37 4.26 ±0.58 FFA content 147 mg/ g 151mg/g Glycerides obtained after 90 min hydrolysis (1st and 2n ) by immobilized CR lipase 105 6.4.8 Lipid composition of oils obtained after 1st and 2nd hydrolysis Changes in the lipid composition (MG, DG, and TG levels) after the 1st and 2nd hydrolyses with (a) free lipase or (b) immobilized lipase followed by FFA removal are shown in Fig. 6.8. Prior to the 1st hydrolysis, the oil contained 89.47% TG, 5.89% DG, and 0.75% MG. With the free lipase system, TG content decreased significantly after 30 min to 67.55% with increases of DG and MG (30.13% and 2.32%, respectively) after the 1st hydrolysis followed by FFA removal. A prolonged reaction time (90 min) resulted in slight changes in lipid composition (TG 64.35%, DG 32.40% and MG 3.24%). In the 2nd hydrolysis, the TG content in final n-3 PUFA concentrate was reduced (61.77%), while the DG and MG levels increased and remained constant throughout the rest of the reaction time. On the other hand, after a 30 min reaction time in the 1st immobilized lipase hydrolysis, TG content decreased to 77.74% and DG and MG levels increased to 19.20% and 3.05%, respectively. Changes in lipid composition during the immobilized lipase hydrolysis were rather slow compared to those observed in the free lipase hydrolysis. However, TG content decreased continuously to 67.15% after the 1st hydrolysis and 65.23% after the 2nd hydrolysis, resulting in relatively similar levels with that of free lipase system after 90 min. The slower changes in TG, DG and MG by immobilized lipase compared with free lipase system were accordance with the rate of fatty acid profile changes (Fig. 6.8). Lipid compositions of hydrolyzed oil as well as n-3 PUFA concentration resulted in similar values after 90 min but not shorter time (30 and 60 min), suggesting that immobilized lipase systems require slightly longer reaction times to achieve the same result as obtained from free lipase system (90 min) perhaps because they are covered in chitosan-alginate materials. 100 (a) . (b) 80 80 = 1 n 1 m 60 60 ^i r I 40 is 40

20 20 :|: m - 0 - m - 30 60 90 ll ILIIL 30 60 90 Reaction time (min) Reaction time (min)

DlstTG Q 2nd TO DlstTG Q2ndTG ■ 1 st DG @ 2nd DG B 1 st DG 0 2nd DG ■ IstMG B2ndMG ■ IstMG B2ndMG

Figure 6.8 Changes in lipid composition of final n-3 PUFA concentrate with either (a) free or (b) immobilized lipase system after 90 min with 1 or 2 nd hydrolysis.

o ON 107 6.4.9 Fatty acid profiles of mono-, di-, and triglyceride fractions of hydrolyzed sardine oil Fatty acid profiles in different fractions of final n-3 PUFA concentrate hydrolyzed with immobilized lipase after a 90 min reaction time (1st hydrolysis) were determined in order to find the fraction with the highest n-3 PUFA content (Fig. 6.9). Compared with levels in the total lipid fraction, the TG fraction contained higher levels of 16:0 (19.67%) and 16:ln7 (6.46%) and lower levels of 20:5n3, 22:6n3, and total n-3 PUFAs (30.52%, 10.75%, and 49.18%, respectively). The lowest n-3 PUFA levels were found in the MG fraction (31.98%), with 20:5n3 and 22:6n3 present at 17.29% and 6.50%, respectively. The DG fraction contained the highest levels of n-3 PUFAs (62.91%), having 37.53% 20:5n3 and 15.02% 22:6n3. These values were similar to the n-3 PUFA concentrate oil prepared with two 90 min hydrolyses, suggesting that fractionation and recovery of the DG fraction could be an alternative application to produce highly-concentrated n-3 PUFA oil. Similar results were shown in a study by McNeill et al. (1996), where the highest n-3 PUFA levels were found in the DG fraction compared to the TG and FFA fractions in hydrolyzed fish oil (species were undefined). Also in agreement with the present study, Halldorsson et al. (2004) reported that the MG fraction of hydrolyzed herring oil contained the highest SFA, especially in the case of 16:0, with lower levels of 20:5n3 and 22:6n3. 70 60 50 ■ C16:0 ^ 40 ■ C16:ln7 ED C20:5n3 I 30 5 B C22:6n3 20 D n-3 PUFA 10 0 Total Total TG DG MG

Original oil Hydrolyzed oil by immobilized lipase

Figure 6.9 Fatty acid composition of total lipid, TG, DG and MG fractions of original oil and final n-3 PUFA concentrate with immobilized lipase after a 90 min reaction time (1st hydrolysis).

o oc 109 6.5 CONCLUSIONS This study outlines a simple, non-toxic, versatile method to immobilize lipase in an alginate-chitosan-CaCU matrix driven by polyionic interactions for the purpose of producing n-3 PUFA concentrate. The complexation of chitosan, alginate, CaC^ and lipase provides high thermal protection and stability against changes in pH and temperature. The immobilized lipase system allows for easy recovery of both lipase and product, reuse of the enzyme, rapid termination of reactions and increased concentration of n-3 PUFAs. Use of the immobilized lipase system to increase n-3 PUFA concentration from sardine oil provides new processing opportunities for the fats and oils industry and a hopeful, more efficient alternative to existing enzymatic methods. 110 7. GENERAL CONCLUSIONS Seasonal differences were found in the proximate composition of Pacific sardines during the 2004 harvest season off the Oregon-Washington (OR-WA) coasts. Sardines had the lowest lipid content in June, and reached maximum levels in the middle of August. Sardines caught in the OR-WA coastal area also contained high levels of n-3 PUFAs with EPA and DHA having peak levels of 3.36 g and 1.39 g per 100 g sardine tissue, respectively. Because increases in lipid content correspond to increases in n-3 PUFA levels, sardines captured in August-September have higher potential for oil recovery for nutritional purposes. Lipase-catalyzed hydrolysis was found to be a viable method for concentrating n-3 PUFAs from sardine oils, resulting in a 2-fold increase in the total levels of these fatty acids. This study showed that nonspecific Upases were more efficient at concentrating n-3 PUFAs in sardine oils than 1,3-specific lipases. The optimum enzyme concentration was 250 U/g sardine oil, and optimum reaction time was found to be 1.5 h. The new oil extraction method by pH adjustment developed in study 3 resulted in increased oil recovery and quality as compared to conventional heat extraction. A pH adjustment method with citric acid + calcium citrate resulted in high oil recovery and high quality of final products as compared to other methods. The phospholipids (PLs) fraction was non-detectable in oil extracted by this method, thereby eliminating the need for the degumming process, which normally serves to remove the PLs fraction from the final products. There was a significant beneficial effect of calcium addition prior to pH adjustment in terms of stability against lipid oxidation. The results of this study show that the pH adjustment method has advantages over traditional oil recovery processes, especially in the quality of the oil recovered from sardines. A simple, non-toxic, versatile method to immobilize lipase in an alginate- chitosan-CaCh matrix was developed using polyionic interactions. The triple complexation among chitosan, alginate and lipase provides high thermal protection and stability across a wide range of pH values. The immobilized lipase significantly Ill increased n-3 PUFAs level in sardine oils while allowing easy recovery of both lipase and product, reuse of enzyme and rapid termination of reactions. The use of developed immobilized lipase systems for increasing n-3 PUFA concentration in sardine oils will provide new opportunities for the growing sardine industry and promote its expansion into new markets. 112 BIBLIOGRAPHY Adachi S, Okumura K, Ota Y and Mankura M. 1993. Acidolysis of sardine oil by lipase to concentrate eicosapentaenoic and docosahexaenoic acids in glycerides. J Perm Bioeng 75: 259-264. Aidos I, Louren90 S, Luten JB, Van der Padt A and Boom RM. 2002. Stability of crude herring oil produced from fresh byproducts: Influence of temperature during storage. J Food Sci 67: 3314-3320. Aidos I, Luten JB, Boom RM, Jacobsen C, Jensen B and Van Der Padt A. 2002. Volatile oxidation products formed in crude herring oil under accelerated oxidative conditions. Eur J Lipid Sci Technol 104: 808-818. Aidos I, Luten JB, Kreb N, Boonman M, Boom RM and Van Der Padt A. 2003. Influence of production process parameters on fish oil quality in a pilot plant. J Food Sci 68: 581-587. Aidos I, Padt AVD, Luten JB and Boom RM. 2002. Seasonal changes in crude and lipid composition of herring fillets, byproducts, and respective produced oils. J Agric Food Chem 50: 4589-4599. Aidos I, Schelvis Smit R, Veldman M, Luten JB, Boom RM and Der Padt AV. 2003. Chemical and sensory evaluation of crude oil extracted from herring byproducts from different processing operations. J Agric Food Chem 51: 1897- 1903. Akimoto M, Izawa M, Hoshino K, Abe K-I and Takahashi H. 2003. Lipase-catalyzed interesterification of soybean oil with an omega-3 polyunsaturated fatty acid concentrate prepared from sardine oil. Appl Biochem Biotechnol 104: 105-118. Alamed J, McClements DJ and Decker EA. 2006. Influence of heat processing and calcium ions on the ability of EDTA to inhibit lipid oxidation in oil-in-water emulsions containing omega-3 fatty acids. Food Chem 95: 585-590. Alasalvar C, Shahidi F and Quantick P. 2002. Food and health applications of marine nutraceuticals: a review, p. 175-195. In: Alasalvar C, Taylor T (eds.). Seafoods-Quality, Technology and Nutraceutical Applications. Springer, New York. Alkio M, Gonzalez C, Jantti M and Aaltonen O. 2000. Purification of polyunsaturated fatty acid esters from tuna oil with supercritical fluid chromatography. J Amer Oil Chem Soc 77: 315-321. Al-Saghir S, Thumer K, Wagner KH, Frisch G, Luf W, Razzazi-Fazeli E and Elmadfa I. 2004. Effects of different cooking procedures on lipid quality and cholesterol oxidation of farmed salmon fish (Salmo salar). J Agric Food Chem 52: 5290- 5296. Alsarra IA, Betigeri SS, Zhang H, Evans BA and Neau SH. 2002. Molecular weight and degree of deacetylation effects on lipase-loaded chitosan bead characteristics. Biomaterials 23: 3637-3644. Ando Y, Ota T, Matsuhira Y and Yazawa K. 1996. Stereospecific analysis of triacyl- sn-glycerols in docosahexaenoic acid-rich fish oils. J. Amer. Oil Chem 73: 483-487. 113 AOAC. 1998. Official Methods of Analysis of AOAC International. AOAC International, VA. AOCS. 1998. Official methods and recommended practices of the American Oil Chemists's Society. Champaign, IL. Bandarra NM, Batista I, Nunes ML, Empis JM and Christie WW. 1997. Seasonal changes in lipid composition of sardine (Sardina pilchardus). J Food Sci 62: 40-42. Baxter S, Zivanovic S and Weiss J. 2005. Molecular weight and degree of acetylation of high-intensity ultrasonicated chitosan. Food Hydrocoll 19: 821-830. Beltran A and Moral A. 1991. Changes in fatty acid composition of fresh and frozen sardine {Sardinapichardus W.) during smoking. Food Chem 42: 99-109. Benjamin S and Pandey A. 1998. Candida rugosa lipases: Molecular biology and versatility in biotechnology. Yeast 14: 1069-1087. Berger J, Reist M, Mayer JM, Felt O, Gumy R and Peppas NA. 2004. Structure and interactions in covalently and ionically crosslinked chitosan hydrogels for biomedical applications. Europ J Pharm Biopharm 57: 19-34. Berges E. 1996. Importance of vitamin E in the oxidative stability of meat: organoleptic qualities and consequences. CIHEAM- Options Mediterraneennes: 347-363. Betigeri SS and Neau SH. 2002. Immobilization of lipase using hydrophilic polymers in the form of hydrogel beads. Biomaterials 23: 3627-3636. Bimbo AP. 1990. Processing offish oils, p. 181-225. In: Standsby ME (ed.). Fish Oils In Nutrition. Van Nostrand Reinhold, New York. Bimbo AP. 1990. Production offish oil, p. 141-225. In: Standsby ME (ed.). Fish Oils In Nutrition. Van Nostrand Reinhold, New York. Boadi DK and Neufeld RJ. 2001. Encapsulation of tarmase for the hydrolysis of tea tannins. Enz Microb Technol 28: 590-595. Bomscheuer UT. 2000. Enrichment of lipids with EPA and DHA by lipase, p. 170. In: Bomscheuer UT (ed.). Enzymes in Lipid Modification. WILLEY-VCH Verlag GmbH, Weinheim. Breguet V, Gugerli R, Von Stockar U, Marison IW and Pemetti M. 2005. Formation of microcapsules from polyelectrolyte and covalent interactions. Langmuir 21: 9764-9772. Breivik H, Haraldsson GG and Kristinsson B. 1997. Preparation of highly purified concentrates of eicosapentaenoic acid and docosahexaenoic acid. J Amer Oil Chem Soc 74: 1425-1429. Caponio F, Summo C, Bilancia MT, Lestingi A and Laudadio V. 2004. Chemical characteristics and lipid fraction quality of sardines (Sardina pilchardus W.): Influence of sex and length. J Appl Ichthyol 20: 530-535. Carvalho PO, Campos PRB, Noffs MD, Bastos DHM and de Oliveira JG. 2002. Enzymatic enhancement of omega-3 polyunsaturated fatty acids content in Brazilian sardine oil. ActaFarm Bonaerense 21: 85-88. Castrillon AM, Navarro P and Alvarez-Pontes E. 1997. Changes in chemical composition and nutritional quality of fried sardine (Clupea pilchardus) 114 produced by frozen storage and microwave reheating. J Sci Food Agric 75: 125-132. Chantachum S, Benjakul S and Sriwirat N. 2000. Separation and quality offish oil from precooked and non-precooked tuna heads. Food Chem 69: 289-294. Chen T-C and Ju Y-H. 2000. Enrichment of eicosapentaenoic acid and docosahexaenoic acid in saponified menhaden oil. J Amer Oil Chem Soc 77: 425-428. Chen T-C and Ju Y-H. 2002. Concentration of n-3 PUFA in saponified Menhaden oil by fractional solvent crystallization. J Chin Inst Chem Eng 33: 143-147. Choi YJ and Park JW. 2002. Acid-aided protein recovery from enzyme-rich Pacific whiting. J Food Sci 67: 2962-2967. Cisner Os-Mata MA and Nevarez-Martinez MO. 1995. The rise and fall of the Pacific sardine, Sadinops sagax caeruleus girard, in the gulf of California, Mexico. CalCOFI Rep 36: 136-143. Colombo J, Shaddy DJ, Anderson CJ, Blaga OM, Karmass KN, Kundurthi S, Maikranz JM and Carlson SE. 2004. Maternal DHA and the development of attention in infancy and toddlerhood. Child Development 75: 1254-1267. Coppi G, lannuccelli V, Leo E, Bemabei MT and Cameroni R. 2002. Protein immobilization in crosslinked alginate microparticles. J Microencap 19: 37-44. Damodaran S. 1996. Amino acid, peptides, and proteins., p. 370-371. In: Fennema OR (ed.). Food Chemistry. Marcel Dekker, New York. Daniels JL, Longnecker MP, Rowland AS and Golding J. 2005. Fish intake during pregnancy and early cognitive development of offspring. Obstetrical & gynecological survey 60: 80-81. De Riso L, De Alteriis E, Parascandola P, La Cara F and Sada A. 1996. Immobilization of Bacillus acidocaldarius whole-cell rhodanese in polysaccharide and insolubilized gelatin gels. Biotechnol Appl Biochem 23: 127-131. Debruyne I. 2004. Soybean oil processing: quality criteria and flavor reversion. Oil Mill Gazetteer 110: 10-11. DeGroot AR and Neufeld RJ. 2001. Encapsulation of urease in alginate beads and protection from chymotrypsin with chitosan membranes. Enz Microb Technol 29: 321-327. deMan JM. 1999. Enzymes, p. 389-423. In: deMan JM (ed.). Principles of Food Chemistry. Springer Science and Business Media, Inc., New York. Dion P. 2003. On-board handling of sardines in the Oregon sardine fishery. Astoria, OR. Dumitriu S and Chomet E. 1997. Immobilization of xylanase in chitosan-xanthan hydrogels. Biotechnol Prog 13: 539-545. Dumitriu S, Magny P, Montane D, Vidal PF and Chomet E. 1994. Polyionic hydrogels obtained by complexation between xanthan and chitosan: Their properties as supports for enzyme immobilization. J Bioactive Comp Polym 9: 184-209. Emmet RTL, Broudeur RD, Miller TW, Pool SS, Krutzikowsky GK, Bentley PJ and McCrae J. 2005. Pacific sardine {Sardinops sagax) abundance, distribution, and ecological relatioships in the Pacific northwest. CalCOFI Rep 46: 122-143. 115 EPA, 2000. Wastewater technology fact sheet chemical precipitation, EPA 832-F-00- 018 Vol. 2006. Erickson MC. 1992. Changes in lipid oxidation during cooking of refrigerated minced channel catfish muscle, p. 344-350. In: St Angelo AJ (ed.). Lipid oxidation. American chemical society, Washington D.C. Eymard S, Carcouet E, Dumay J, Chopin C, Rochet M-J and Genot C. 2005. Development of lipid oxidation during manufacturing of horse mackerel surimi. J Sci Food Agric 85: 1750-1756. FAO. 1986. The production offish meal and oil. Food and Agriculture Ortanization of the United Nations, Rome. Faustman C, Cassens RG, Schaefer DM, Bueger DR, Williams SN and Scheller KK. 1989. Improvement of pigment and lipid stability in Holstein steer beef by dietary supplementation with vitamin E. J Food Sci 54: 858-863. Fiorillo J. 2005. Are fish oil supplements a threat? Intrafish 3: 18-20. Galameau E and Ronald G. 1997. Phosphorus removal from wastewaters: Experimental and theoretical support for alternative mechanisms. Water Res 31:328-338. Gamez-Meza N, Noriega-Rodriguez J, Higuera-Ciapara I, Calderon De La Barca AM, Vazquez-Moreno L and Angulo-Guerrero O. 1999. Seasonal variation in the fatty acid composition and quality of sardine oil from Sardinops sagax caeruleus of the Gulf of California. Lipids 34: 639-642. Gamez-Meza N, Noriega-Rodriguez JA, Medina-Juarez LA, Ortega-Garcia J, Monroy-Rivera J, Toro-Vazquez FJ, Garcia HS and Angulo-Guerrero O. 2003. Concentration of eicosapentaenoic acid and docosahexaenoic acid from fish oil by hydrolysis and urea complexation. Food Res Intern 36: 721-727. Ganga A, Nieto S, Sanhuez J, Romo C, Speisky H and Valenzuela A. 1998. Concentration and stabilization of n-3 polyunsaturated fatty acids from sardine oil. J Amer Oil Chem Soc 75: 733-736. Garcia HS, Arcos JA, Ward DJ and Hill C.G. J. 2000. Synthesis of glycerides containing n-3 fatty acids and conjugated linoleic acid by solvent-free acidolysis offish oil. Biotechnol Bioeng 70: 587-591. Gaserad O, Smidsrad O and Skjak-Braek G. 1998. Microcapsules of alginate-chitosan -1. A quantitative study of the interaction between alginate and chitosan. Biomaterials 19: 1815-1825. Georgantas DA and Grigoropoulou HP. 2005. Phosphorus removal from synthetic and municipal wastewater using spent alum sludge. Water Sci Technol 52: 525-532. Gil MI, Tomas-Barberan FA, Hess-Pierce B, Holcroft DM and Kader AA. 2000. Antioxidant activity of pomegranate juice and its relationship with phenolic composition and processing. J Agric Food Chem 48: 4581-4589. Gokce AM, Tasbozan O, Celik M and Tabakoglu S. 2004. Seasonal variation in proximate and fatty acid compositions of female common sole {Solea soled). Food Chem 88: 419-423. Gokodlu N, Ozden O and Erkan N. 1998. Physical, chemical and sensory analyses of freshly harvested sardines {Sardina pichardus) stored at 4 0C. J Aquatic Food Prod 7: 5-15. 116 Gomes FM, Pereira EB and de Castro HF. 2004. Immobilization of lipase on chitin and its use in nonconventional biocatalysis. Biomacromolecules 5: 17-23. Gomez L, Ramirez HL, Villalonga ML, Hernandez J and Villalonga R. 2006. Immobilization of chitosan-modified invertase on alginate-coated chitin support via polyelectrolyte complex formation. Enz Microb Technol 38: 22-27. Gonzalez Moreno PA, Robles Medina A, Camacho Rubio F, Camacho Paez B and Molina Grima E. 2004. Production of structured lipids by acidolysis of an EPA-enriched fish oil and caprylic acid in a packed bed reactor: Analysis of three different operation modes. Biotechnol Prog 20: 1044-1052. Guil-Guerrero JL and Belarbi E-H. 2001. Purification process for cod liver oil polyunsaturated fatty acids. J Amer Oil Chem Soc 78: 477-484. Gupta R, Rathi P and Bradoo S. 2003. Lipase mediated upgradation of dietary fats and oils. Crit Rev Food Sci Nutr 43: 635-644. Haag M. 2003. Essential fatty acids and the brain. Canad J Psychol 48: 195-203. Halldorsson A, Kristinsson B, Glyrm C and Haraldsson GG. 2003. Separation of EPA and DMA in Fish Oil by Lipase-Catalyzed Esterification with Glycerol. J Amer Oil Chem Soc 80: 915-921. Halldorsson A, Kristinsson B and Haraldsson GG. 2004. Lipase selectivity toward fatty acids commonly found in fish oil. Europ J Pharm Biopharm 106: 79-87. Hammann MG. 1998. Spawning habitat of the Pacific sardine {Sardinops Sagax) in the gulf of California: egg and larval distribution 1956-1957 and 1971-1991. CalCOFI Rep 39: 169-179. Hammann MG, Baumgartner TR and Bandan-Dangon A. 1988. Coupling of the Pacific sardine (Sardinops Sagax Caeruleus) life cycle with the gulf of California pelagic environment. CalCOFI Rep XXIX: 102-109. Haraldsson GG, Kristinsson B, Sigurdardottir R, Gudmundsson GG and Breivik H. 1997. The preparation of concentrates of eicosapentaenoic acid and docosahexaenoic acid by lipase-catalyzed transesterification of fish oil with ethanol. J Amer Oil Chem Soc 74: 1419-1424. Hardy R and Keay JN. 1972. Seasonal variations in the chemical composition of Cornish mackerel, Scomber scombrus (L) with detailed reference to the lipids. J Food Technol 7: 125-137. Hari PR, Chandy T and Sharma CP. 1996. Chitosan/calcium alginate microcapsules for intestinal delivery of nitrofurantoin. J Microencap 13: 319-329. Hayashi T and Ikada Y. 1991. Protease immobilization onto porous chitosan beads. J Appl Polym Sci 42: 85-92. He K, Daviglus ML, Liu K, Van Horn L, Dyer AR, Greenland P and Song Y. 2004. Accumulated evidence on fish consumption and coronary heart disease mortality: A meta-analysis of cohort studies. Circulation 109: 2705-2711. He Y and Shahidi F. 1997. Enzymatic esterification of omega-3 fatty acid concentrates from seal blubber oil with glycerol. J Amer Oil Chem Soc 74: 1133-1136. Hernandez-Vazquez S. 1994. Distribution of eggs and larvae from sardine and anchovy off California and Baja California, 1951-1989. CalCOFI Rep 35: 94- 107. 117 Hibbeln JR and Salem N. 2001. Omega-3 fatty acids and psychiatric disorders, current statutes of the field, p. 311-330. In: Mostofsky DJ, Yehuda S, N S (eds.). Fatty acids/Physiological and Behavioral Functions. Humana Press, New Jersey. Hidalgo FJ and Zamora R. 2006. Peptides and proteins in edible oils: Stability, allergenicity, and new processing trends. Trends Food Sci Technol 17: 56-63. Holub DJ and Holub BJ. 2004. Omega-3 fatty acids from fish oils and cardiovascular disease. Mol Cell Biochem 263: 217-225. Hoshino T, Yamane T and Shimazu S. 1990. Selective hydrolysis offish oil by lipase to concentrate n-3 polyunsaturated fatty acids. Agric Biolog Chem 54: 1459- 1467. Houde A, Kademi A and Leblanc D. 2004. Lipases and their industrial applications. Appl Biochem Biotech 118: 155-170. Hsieh C-W, Chang CJ and Ko W-C. 2005. Supercritical CO2 extraction and concentration of n-3 polyunsaturated fatty acid ethyl esters from tuna cooking juice. Fish Sci 71: 441-447. Hultin HO. 1994. Oxidation of lipids in seafoods, p. 49-74. In: Shahidi F, Botta JR (eds.). Seafoods: Chemistry, Processing Technology and Quality. Blackie Academic and Professional, New York. Hultin HO and Kelleher SD, 1999. Process for isolating a protein composition from a muscle source and protein composition, Adv Protein Technol, US. Huss HH. 1995. Quality and quality changes in fresh fish. Food and Agriculture Organization of United Nations, Rome. Hwang LS and Liang J-H. 2001. Fractionation of urea-pretreated squid visceral oil ethyl esters. J Amer Oil Chem Soc 78: 473-476. H'ina AV and Varlamov VP. 2005. Chitosan-based polyelectrolyte complexes: A review. Appl Biochem Microb 41: 5-11. Imre S and Saglik S. 1998. Fatty acid composition and cholesterol content of some Turkish fish species. Turk J Chem 22: 321-324. Innis SM. 2003. Perinatal biochemistry and physiology of long-chain polyunsaturated fatty acids. J Pediatric 143: Sl-8. Innis SM. 2004. Polyunsaturated fatty acids in human milk: an essential role in infant development. Adv Exp Med Biolog 554: 27-43. Iso H, Kobayashi M, Ishihara J, Sasaki S, Okada K, Kita Y, Kokubo Y and Tsugane S. 2006. Intake offish and n-3 fatty acids and risk of coronary heart disease among Japanese: the Japan public health enter-based (JPHC) study cohort I. Circulation 113: 195-202. JIS. 1995. Determination of lipolytic activity of lipase for industrial use. Japanese Industrial Standards Committee, Tokyo, Japan. Katikou P, Hughes SI and Robb DHF. 2001. Lipid distribution within Atlantic salmon (Salmo salar) fillets. Aquacul 202: 89-99. Khan TA, Peh Kk and Chng HS. 2000. Mechanical, bioadhesive strength and biological evaluations of chitosan films for wound dressing. J Pharm PhamaceuSci3: 303-311. Kotb AR and Abu Hadeed AF. 1991. Omega-3 polyunsaturated fatty acid content of some popular species of Arabian gulf fish. Food Chem 40: 185-190. 118 Krajewska B. 2004. Application of chitin- and chitosan-based materials for enzyme immobilizations: A review. Enz Microb Technol 35: 126-139. Kris-Etherton PM, Harris WS and Appel LJ. 2002. Fish consumption, fish oil, omega- 3 fatty acids, and cardiovascular disease. Circulation 106: 2747-2757. Kris-Etherton PM, Taylor DS, Yu-Poth S, Huth P, Moriarty K, Fishell V, Hargrove RL, Zhao G and Etherton TD. 2000. Polyunsaturated fatty acids in the food chain in the United States. Amer J Clin Nutr 71: 179S-188S. Lawson LD and Hughes BG. 1998. Absorption of eicosapentaenoic acid and docosahexaenoic acid from fish oil triacylglycerols or fish oil ethyl esters co- ingested with a high-fat meal. Biochem Biophys Res Comm 156: 960-963. Lee KH, Lee PM and Saiw YS. 1992. Studies of L-phenylalanine production by immobilised aminoacylase in stabilized calcium alginate beads. J Chem Technol Biotechnol 54: 375-382. Lee KH, Lee PM and Siawa YS. 1993. Immobilization of aminoacylase by encapsulation in poly-L-lysine-stabilized calcium alginate beads. J Chem Technol Biotechnol 57: 27-32. Leonardis AD and Macciola V. 2004. A study on the lipid fraction of Adriatic sardine fillets {Sardinapilchardus). Nahrung/Food 48: 209-212. Liang Y and Hultin HO. 2005. Separation of membranes from acid-solubilized fish muscle proteins with the aid of calcium ions and organic acids. J Agric Food Chem 53: 3008-3016. Liang Y and Hultin HO. 2005. Separation of muscle membrane from alkali- solubilized fish muscle proteins. J Agric Food Chem 53: 10012-10017. Lin T-J, Chen S-W and Chang A-C. 2006. Enrichment of n-3 PUFA contents on triglycerides of fish oil by lipase-catalyzed trans-esterification under supercritical conditions. Biocheml Eng J 29: 27-34. Linder M, Fanni J and Parmentier M. 2005. Proteolytic extraction of salmon oil and PUFA concentration by Upases. Mar Biotechnol 7: 70-76. Lluch-Belda D, Lluch-Cota DB and Schwartzlose RA. 1991. Sardine and anchovy spawning as related to temperature and upwelling in the California current system. CalCOFIRep 32: 105-111. Lopez-Caballero ME, Gomez-Guillen MC, Perez-Mateos M and Montero P. 2005. A chitosan-gelatin blend as a coating for fish patties. Food Hydrocoll 19: 303- 311. Love RM. 1997. Biochemical dynamics and the quality of fresh and frozen fish, p. 1- 26. In: Hall GM (ed.). Fish Processing Technology. Blackie Academic and Professional pub, London. Low K-S, Lee C-K and Kong L-Y. 1998. Decolorisation of crude palm oil by acid- activated spent bleaching earth. J Chem Technol Biotechnol 72: 67-73. Lowry OH, Rosebrough NJ, Fair AL and Randall RJ. 1951. Protein measurement with the Folin phenol reagent. J Biol Chem 193: 265-275. Luten JB, Aidos I, Boom RM and Van Der Padt A. 2003. Quality of crude fish oil extracted from herring byproducts of varying states of freshness. J Food Sci 68: 458-465. 119 Luzia LA, Sampaio GR, Castellucci CMN and Torres EAFS. 2003. The influence of season on the lipid profiles of five commercially important species of Brazilian fish. Food Chem 83: 93-97. MacDonald G, Hall BI and Vlieg P. 2002. Seasonal changes in hoki (Macruronus novaezelandaie)- implications for quality and yield. J Aquatic Food Prod Technol11:35-51. Mattey M and Kristiansen B. 1999. A Brief Introduction to Citric Acid Biotechnology, p. 1-9. In: Kristiansen B, Mattey M, Linden J (eds.). Citric Acid Biotechnology. Taylor and Francis Inc., Philadelphia. McCrae J and Smith J. 2005. Oregon's sardine fishery 2004 summary Newport, OR. McCrae J and Smith J. 2006. Oregon's sardine fishery 2005 summary Newport, OR. McGuire SO, Alexander DW and Fritsche KL. 1997. Fish oil source differentially affects rat immune cell a-tocophenol concentration. Nutr Reqs Interact: 1388- 1394. McNeill GP, Ackman RG and Moore SR. 1996. Lipase-catalyzed enrichment of long- chain polyunsaturated fatty acids. J Amer Oil Chem Soc 73: 1403-1407. Miller C, Austin H, Posorske L and Gonzlez J. 1988. Characteristics of an immobilized lipase for the commercial synthesis of esters. J Amer Oil Chem Soc 65: 927-931. Minovska V, Winkelhausen E and Kuzmanova S. 2005. Lipase immobilized by different techniques on various support materials applied in oil hydrolysis. J Serbian Chem Soc 70: 609-624. Mohapatra SC and Hsu JT. 1999. Optimizing lipase activity, enantioselectivity, and stability with medium engineering and immobilization for ?-blocker synthesis. Biotechnol Bioeng 64: 213-220. Morris MC, Evans DA, Bienias JL, Tangney CC, Bennet DA, Wilson RS, Aggarwal N and Schneider J. 2003. Consumption offish and n-3 fatty acids and risk of incident Alzheimer disease. Archives of Neurology 60: 940-946. Morrissey MT and Sylvia G. 2004. Intrinsic and extrinsic factors affecting efficient utilization of marine resources, p. 37-43. In: Sakaguchi M (ed.). More Efficient Utilization of Fish and Fisheries Products. Elsvier, Amsterdam. Mozaffarian D, Bryson CL, Lemaitre RN, Siscovick DS and Burke GL. 2005. Fish intake and risk of incident heart failure. J Amer College Cardiol 45: 2015-2021. Nation JL. 1983. A new method using hexamethyldisilazane for preparation of soft insect tissues for scanning electron microscopy. Stain Technol 58: 347-351. Nawar WW. 1996. Lipids, p. 225-319. In: Fennema OR (ed.). Food Chem. Marcel Dekker, Inc., New York. Nettleton JA. 1995. Omega-3 fatty acids and health. Chapman and Hall, New York. Nettleton JA and Katz R. 2005. n-3 long-chain polyunsaturated fatty acids in type 2 diabetes: A review. J Amer Diet Associ 105: 428-440. Nieto S, Cordoba AM, Sanhueza J and Valenzuela A. 1997. Obtention of highly purified fractions of eicosapentaenoic acid and docosahexaenoic acid from sardine oil by silver-resin chromatography: A semi-preparative procedure. Grasas y Aceites 48: 197-199. 120 Nunes ML, Batista I and Morao de Campos R. 1992. Physical, chemical and sensory analysis of sardine (Sardina pilchardus) stored in ice. J Sci Food Agric 59: 37- 43. Okland AMW, Stoknes IS, Remme JF, Kjerstad M and Synnes M. 2005. Proximate composition, fatty acid and lipid class composition of the muscle from deep- sea teleosts and elasmobranchs. Comp Biochem Physiol 140: 437-443. Olsson BG, Olsen LR, Carlehog M and Ofstad R. 2003. Differentiation of cultured and wild sea b&ss(Dicentrachus labrax): total lipid content, fatty acid and trace mineral composition. Food Chem 79: 145-150. Pace CN, Trevino S, Prabhakaran E, Scholtz JM, Franks F, Wilson K, Daniel RM, Hailing PJ, Clark DS and Purkiss A. 2004. Protein structure, stability and solubility in water and other solvents. Philosophy Transac Royal Soc London Ser B Biological Sciences 359: 1225-1235. Park JD, Jung CH, Kim JS, Cho DM, Cho MS and Choi YJ. 2003. Surimi processing using acid and alkali solubilization offish muscle protein. J Korean Soc Food Sci Nutr 32: 400-405. Pereira EB, De Castro HF, De Moraes FF and Zanin GM. 2001. Kinetic studies of lipase from Candida rugosa: A comparative study between free and immobilized enzyme onto porous chitosan beads. App Biochem Biotechnol . 91-93: 739-752. Pereira EB, Zanin GM and Castro HF. 2003. Immobilization and catalytic properties of lipase on chitosan for hydrolysis and esterification reactions. Braz J Chem Eng 20: 343-355. Rakshit SK, Vasuhi R and Kosugi Y. 2000. Enrichment of polyunsaturated fatty acids from tuna oil using immobilized Pseudomonas fluorescens lipase. Bioprocess Eng 23: 251-255. Ramirez Fajardo A, Akoh CC and Lai OM. 2003. Lipase-Catalyzed Incorporation of n-3 PUFA into Palm Oil. J Amer Oil Chem Soc 80: 1197-1200. Riha V and Brunner G. 2000. Separation offish oil ethyl esters with supercritical carbon dioxide. J Supercri Flu 17: 55-64. Robles MA, Esteban Cerdan L, Gimenez Gimenez A, Camacho Paez B, Ibanez Gonzalez MJ and Molina Grima E. 1999. Lipase-catalyzed esterification of glycerol and polyunsaturated fatty acids from fish and microalgae oils. J Biotechnol 70: 379-391. Robles VJ, Garcia HS, Monroy JA and Angulo O. 1998. Lipase-catalyzed esterification of glycerol with n-3 polyunsaturated fatty acids from winterized fish oil. Food Sci Technol Intern 4: 401-405. Rodriguez C, Acosta C, Badia P, Cejas JR, Santamaria FJ and Lorenzo A. 2004. Assessment of lipid and essential fatty acids requirements of black seabream (Spondylipsoma cantharus) by comparison of lipid composition in muscle and liver of wild and captive adult fish. Comp Biochem Physiol 139: 619-629. Rudan-Tasic D and Klofutar C. 1999. Characteristics of vegetable oils of some Slovene manufacturers. Acta Chim Slov 46: 511-521. Ruiz-Gutierrez V and Perez-Camino MC. 2000. Update on solid-phase extraction for the analysis of lipid classes and related comounds. J Chromato A 885: 321-341. 121 Ruxton CH, Reed SC, Simpson MJ and Millington KJ. 2004. The health benefits of omega-3 polyunsaturated fatty acids: a review of the evidence. J Hum Nutr Diet 17: 449-459. Sankalia MG, Mashru RC, Sankalia JM and Sutariya VB. 2005. Papain entrapment in alginate beads for stability improvement and site-specific delivery: Physicochemical characterization and factorial optimization using neural network modeling. AAPS Pharm Sci Tech 6: E209-222. Sato B, Sasaki Y and Abe S. 1978. Developing technology of utilization of small pelagic fish. Fish Agency: 105-116. Schmitt-Rozieres M, Deyris V and Comeau L-C. 2000. Enrichment of polyunsaturated fatty acids from sardine cannery effluents by enzymatic selective esterification. J Amer Oil Chem Soc 77: 329-332. Schweigert JF. 2002. Update COSEWIC status report on the Pacific Sardine {Sardinops sagax) in Canada Nanaimo. Senanayake SPIN and Shahidi F. 1999. Enzyme-assisted acidolysis of borage {Borago officinalis L.) and evening primrose (Oenothera biennis L.) oils:Incorporation of omega-3 polyunsaturated fatty acids. J Agric Food Chem 47: 3105-3112. Serdaroglu M and Felekogglu E. 2005. Effects of using rosemary extract and onion juice on oxidative stability of sardine (Sardina pilchardus) mince. J Food Qual 28: 109-120. Shahidi F, Arachchi JKV and Jeon Y-J. 1999. Food applications of chitin and chitosans. Trends Food Sci Technol 10: 37-51. Shahidi F and Wanasundara UN. 1998. Omega-3 fatty acid concentrates: nutritional aspects and production technologies. Trends in Food Science and Technology 9: 230-240. Sheldon RA. 1993. Chirotechnology-industrial synthesis of optically-active compounds. Marcel Dekker Ltd., New York. Shimada Y, Maruyama K, Okazaki S, Nakamura M, Sugihara A and Tominaga Y. 1994. Enrichment of polyunsaturated fatty acids with Geotrichum candidum lipase. J Amer Oil Chem Soc 71: 951-954. Shirai N, Terayama M and Takeda H. 2002. Effect of season on the fatty acid composition and free amino acid content of the sardine Sardinops melanostictus. Comp Biochem Physiol 131: 387-393. Shukla SR and Jajpura L. 2005. Immobilisation of amylase by various techniques. Indian J Fibre Textile Res 30: 75-81. Siddiqui RA, Shaikh SR, Sech LA, Yount HR, Stillwell W and Zaloga GP. 2004. Omega 3-fatty acids: health benefits and cellular mechanisms of action. Mini Rev Med Chem 4: 859-871. Sinnhuber RO and Yu TC. 1977. The 2-Tiobarbituric acid reaction, an objective measure of the oxidative deterioration occuring in fats and oils. J Japan Oil Chem Soc 26: 259-267. Smith KM and Sahyoun NR. 2005. Fish consumption: Recommendations versus advisories, can they be reconciled? Nutri Rev 63: 39-46. Sobstad GE. 1992. Marine oil separation, purificatin technology. INFORM 3: 827-830. 122 Srivastava R, Brown JQ, Zhu H and McShane MJ. 2005. Stable encapsulation of active enzyme by application of multilayer nanofilm coatings to alginate microspheres. Macromol Biosci 5: 717-727. Stefansson G and Hultin HO. 1994. On the solubility of cod muscle proteins in water. J Agric Food Chem 42: 2656-2664. Sun T, Pigott GM and Herwig RP. 2002. Lipase-assisted concentration of n-3 polyunsaturated fatty acids from viscera of farmed Atlantic salmon (Salmo salar L.). J Food Sci 67: 130-136. Suzuki Y and T. M. 2005. Removal of emulsified oil from water by coagulation and foam separation. Separation Sci Technol 40: 3407-3418. Tanaka Y, Hirano J and Funada T. 1994. Synthesis of docosahexaenoic acid-rich triglyceride with immobilized chromobacterium viscosum lipase. J Amer Oil Chem Soc 71: 331-334. Tanaka Y, Hirano J and Tunada T. 1992. Concentration of docosahexaenoic acid in glyceride by hydrolysis offish oil with Candida cylindracea lipase. J Amer Oil Chem Soc 69: 1210-1214. Taqieddin E and Amiji M. 2004. Enzyme immobilization in novel alginate-chitosan core-shell microcapsules. Biomaterials 25: 1937-1945. Tiemeier H, Tujil HR, Hofinan A, Kiliaan AJ and Breteler MM. 2003. Plasma fatty acid composition and depression are associated in the elderly: the Rotterdam study. Amer J Clin Nutr 78: 40-46. Toyoshima K, Noguchi R, Hosokawa M, Miyashita K, Fukunaga K, Nishiyama T and Takahashi R. 2004. Separation of Sardine Oil without Heating from Surimi Waste and Its Effect on Lipid Metabolism in Rats. J Agric Food Chem 52: 2372-2375. Uhlig H. 1998. Description of enzymes: Ester cleavage, p. 179-193. In: Linsmaier- Bednar EM (ed.). Industrial Enzymes and Their Applications. Jhon Wiley and Sons, Inc., Tronto. Uhlig H. 1998. Lipase-catalyzed hydrolysis and modification of fats and oils, p. 381- 390. In: Linsmaier-Bednar EM (ed.). Industrial Enzymes and Their Applications. Jhon Wiley and Sons, Inc., Tronto. Ustun G, Guner S, Arer G, Turkay S and Erciyes AT. 1997. Enzymatic hydrolysis of anchovy oil: Production of glycerides enriched in polyunsaturated fatty acids. ApplBiochemBiotechnol68: 171-186. Vakhlu J and Kour A. 2006. Yeast lipases: enzyme purification, biochemical properties and gene cloning. J Biotechnol 9: 69-85. Vaya J and Aviram M. 2001. Nutritional antioxidants: Mechanisms of action, analyses of activities and medical application. Curr Med Chem - Imm Endoc Metab Agents 1:99-117. Wanasundara UN and Shahidi F. 1998. Lipase-assisted concentration of n-3 polyunsaturated fatty acids in acylglycerols from marine oils. J Amer Oil Chem Soc 75: 945-951. Wanasundara UN and Shahidi F. 1999. Concentration of omega 3-polyunsaturated fatty acids of seal blubber oil by urea complexation: Optimization of reaction conditions. Food Chem 65: 41-49. 123 Wheeler SC and Morissey MT. 2003. Quantification and distribution of lipid, moisture, and fatty acids of West coast albacore tuna (Thunnus alalunga). J Aquatic Food Prod Technol 12:3-16. Won K, Kim S, Kim K-J, Moon S-J and Park HW. 2005. Optimization of lipase entrapment in Ca-alginate gel beads. Pro Biochem 40: 2149-2154. Xuebing X. 2000. Production of specific-structured triacylglycerols by lipase- catalyzed reactions: a review. Europ J Pharma Biopharma 102: 287-303. Yongsawatdigul J and Park JW. 2004. Effects of alkali and acid solubilization on gelation characteristics of Rockfish muscle proteins. Food Chem Toxicol 69: C499-539. Yu X, Li Y and Wu D. 2004. Microencapsulation of tannase by chitosan-alginate complex coacervate membrane: Synthesis of antioxidant propyl gallate in biphasic media. J Chem Technol Biotechnol 79: 475-479. Zheng Y, Zheng N, Zhuo J and Ma L. 2005. Concentration of PUFAs in glyceride by hydrolysis offish oil with lipase. Chin J Appl Environ Biology 11: 571-574. Zitouni N, Errahali Y, Metche M, Kanny G, Moneret-Vautrin DA, Nicolas JP and Fremont S. 2000. Influence of refining steps on trace allergenic protein content in sunflower oil. J Aller Clin Immun 106: 962-967. Zvomuya F, Rosen CJ and Gupta SC. 2006. Phosphorus sequestration by chemical amendments to reduce leaching from wastewater applications. J Environ Qual 35: 207-215. 124 APPENDIX

MARINE ENZYMES

Tomoko Okada and Michael T. Morrissey

In: Maximizing the Value of Marine Byproducts

Woodhead Publishing Limited, Cambridge, UK

In press 125 ABSTRACT

Marine-based enzymes often have unique physical, chemical and catalytic properties compared to homologous enzymes from terrestrial animals and plant sources. Extractable enzymes from seafood and seafood byproducts include digestive and cellular proteinases, extracellular gastric proteinases, chitinases, lipases, phospholipases, transglutaminases, polyphenoloxidase and others. Enzyme extraction and utilization from seafood byproducts include enzymes from both solid and liquid waste streams that are often discarded or traditionally processed into fish feed or fertilizer. Seafood byproducts include viscera, skins, bones, heads, and frames from fish as well as shells and exoskeletons from shellfish in both traditional and modem processing operations. Whether it is economically feasible to commercially extract enzymes from these byproducts depends on the volume and quality of raw material, extraction technology and potential markets. This paper describes several different categories of marine enzymes and some of their unique properties. Extraction and purification methods are discussed followed by some examples of commercial operations as well as a discussion of limitations in the use of byproducts for marine enzymes. The advances of biotechnology and the production of specific compounds through gene transfer could provide unique opportunities for marine enzymes in the food and medical fields.

Keywords: seafood, enzyme characterization, enzyme purification, byproducts, biothechnology 126 INTRODUCTION

World fish and shellfish production has increased from 40 million metric tonnes (MMT) in 1962 to 133 MMT in 2002, with a concomitant increase in processed fish and production offish byproducts (Vannuccini, 2004). Frozen and fresh fish were the main fishery products accounting for over 60% of total production followed by canning/curing (7% and 4%, respectively) and non-food uses (26%). While the wild- caught fisheries production appears to have leveled off by the mid 1990s, aquaculture has shown considerable growth and now represents about 30% of the total production (Johnson, 2004). Currently, only about 50 to 60% of total catch is used for direct human consumption, and the annual discards from the world fisheries have been estimated to be approximately 25-30 MMT offish and shellfish caught each year (Sovik et al., 2005). While much of these discards can be used to address increasing demands for fish meal there is also potential for the use of certain byproducts for the extraction of unique compounds such as marine enzymes. Enzyme technology has been used extensively in the food processing industry including fish processing. Enzymes are used as processing aid and can be produced from byproduct material from both new and traditional fish processing operations (Vecchi et al., 1996). There are several excellent reviews on the utilization of enzymes from aquatic organisms (Gildberg et al., 2000; Haard, 1998; Shahidi et al., 2001). Marine enzymes have drawn the attention of numerous researchers as they can possess unique specificity and characteristics. Enzyme recovery from seafood byproducts could also help the environmental and ethical concerns surrounding discards and improve the bottom line for seafood companies wishing to exploit new technologies and markets. In this chapter, the current literature on the extraction and purification of various marine enzymes is reviewed and their potential use and limitations in their utilization is discussed. Extractable enzymes from seafood and seafood byproducts include digestive and cellular proteinases, extracellular gastric proteinases, chitinases, lipases, phospholipases, transglutaminases, polyphenoloxidase and others. An excellent reference source on this topic is the text Seafood Enzymes by Haard and Simpson 127 (2000). A partial list of enzymes which have been purified from fish and processing byproducts include lipase from tilapia stomach (Taniguchi et al, 2001), amino peptidase, carboxypeptidases from tilapia intestine (Taniguchi et al., 2001; Taniguchi et al., 2002), glycogenolytic and alpha glucosidase from Pacific mackerel, Japanese eel, black sea bream and frog flounder (Nakagawa et al., 1996), carbohydrases from butterfish, silver drummer, and marble fish, (Skea et al., 2005), transglutaminase from walleye pollock liver (Kumazawa et al., 1996) and oyster gill (Kumazawa et al., 1997), protease from arrowtooth flounder (Wilson et al., 2004), myosin heavy chain- degrading proteinase from squid muscle (Ehara et al., 1994), cathepsin S from carp hepatopancreas (Pangkey et al., 2000), a-glucosidase, P-glucosidase and amylase from digestive caecum of scallop (Nakai et al., 2005), amylases from hard clam viscera (Tsao et al., 2004), acid and alkaline phosphatase from intestine, liver, kidney of mackerel, sea robin, tongue sole, and others (Kuda et al., 2004), and cellulase and xylanase from crayfish and marine prawns (Crawford et al., 2005). Marine-based enzymes often have unique physical, chemical and catalytic properties compared to corresponding enzymes from terrestrial animals and plant sources (Shahidi et al., 2001). Fish muscle, for example, contains approximately 10 times more catheptic enzyme activities than mammalian muscles (Haard et al., 1994), and marine enzymes are generally more susceptible to hydrostatic pressure inactivation than their mammalian counterparts (Ashie et al., 1996). Since the habitat of many marine animals tends to be in cold temperatures, their enzymes often have cold-adapted properties. They are known to be more catalytically active at low temperatures and possess lower thermal stability making them suitable for enzymatic applications in food processing as they can often be inactivated at lower temperatures (Gerday et al, 2000). Several enzymes have been shown to be salt tolerant as well, which can be an advantage in certain food applications (Cavicchioli et al., 2002; Haard, 1998). Compared to their mesophilic analogues, cold-adapted enzymes tend to have a lower number of hydrogen bonds, less densely packed structures, increased surface hydrophilicity, a higher number of methionine residues, a different fold of the autolysis loop as well as the carboxyterminal region (Gudmundsdottir et al., 2005). 128 Enzyme extraction and utilization from seafood byproducts include enzymes from both solid and liquid waste streams that are often discarded or traditionally processed into fish feed or fertilizer. Seafood byproducts include viscera, skins, bones, heads, and frames from fish as well as shells and exoskeletons from shellfish in both traditional and modem processing operations. Fish and shellfish processing discards can consist of two thirds of incoming raw materials (Lee, 2000) depending on the process and species. Whether it is economically feasible to commercially extract enzymes from these byproducts depends on the volume and quality of raw material, extraction technology and potential markets. The following sections will describe several different categories of marine enzymes and some of their unique properties. Extraction and purification methods will then be discussed followed by some examples of commercial operations as well as a discussion of limitations in the use of byproducts for marine enzymes.

MARINE ENZYMES Proteolytic enzymes Among the extractable enzymes from seafood and seafood byproducts, digestive proteolytic enzymes have received a considerable amount of interest from researchers over the last two decades due to the availability of raw materials such as viscera and their high rate of enzymatic activity. Proteases hydrolyze the peptide bonds in proteins and polypeptides and are characterized as endo- (proteinases) and exo- (peptidases) proteases (Shahidi et al., 2001). The increased production of value- added products such as fresh and frozen fillets and others has created large volumes of discards such as viscera that could be used as a raw material source. Proteolytic enzymes can also be found in muscle cells, the extracellular matrix or other organs and the hepatopancreas in shellfish. They can be extracted from seafood digestive tract and frame muscle as well as waste water from fish, shellfish and surimi processing. Digestive proteolytic enzymes include pepsin, trypsin, chymotryosin, elastase, gastricin and others. Pepsin can be categorized as aspartic proteinase and formed by an autocatalytic reaction from pepsinogen, which is normally located within the fish 129 stomach and has peak activity at acidic conditions (pH 2.0 for most of substrates). Trypsin can be categorized as serine proteinase and it has a dual role in that it cleaves ingested proteins and activates the precursor forms of several other digestive proteinases including chymotrypsin. Trypsin is normally located in pyloric cecum and shows its activity at neutral and alkaline conditions (Naz, 2002). Muscle proteinases include catepsin A, B, C, D, H, and L, calpains and muscle collagenases, and they are mainly located in lysosomes, the sacroplasm, and extracellular matrix of the connective tissue (An et al., 2000). Proteolytic enzymes have been purified and characterized from many species and seafood byproducts. These include Atlantic cod (Amiza et al., 2002; Jonsdottir et al., 2004), skipjack tuna, yellowfin tuna, tongol tuna (Suppasith et al., 2004), anchovy (Heu et al., 1995), white croaker skeletal muscle (Makinodan et al., 1987), salmon (Yamashita et al., 1995), decapoda muscle (Ehara et al., 1992), Pacific whiting (An et al., 1994), orange roughy stomach (Xu et al., 1996), Atlantic menhaden muscle (Choi et al., 1999), Asian bony tongue (Natalia et al., 2003), mackerel white muscle (Aoki et al., 1995), gastric fluid of crab (Sabrowski et al., 2004), Antarctic krill (Sjodahk et al., 2002; Yoshitomi, 2005) jumbo squid (Cardenas-Lopez et al., 2005), and surimi waste water (Benjakul et al., 1996; Mireles-DeWitt et al., 2002). Proteolytic activity of Atlantic cod byproduct from different fishing location was studied by Sovik and Rustad (2005). They found that the proteolytic activity in viscera and frame muscle was the highest at pH 3 and 35 0C, whereas the proteolytic enzyme from liver showed its maximum activity at pH 3 and 50 0C. The proteolytic activity in viscera was 20 times higher than liver and 250 times higher than that of frame muscle. A significant difference in enzymatic activity was also reported for cod viscera harvested from different locations. There were seasonally differences in cod from the Icelandic and Barents Sea in April-June as well as differences from cod in these locations and cod from the South coast of Ireland during October-December. These researchers also studied seasonal changes in trypsin and chymotrypsin activity in viscera from cod, saithe, haddock, tusk and ling viscera from different locations (Sovik et al., 2004) and found that enzyme activities differed among species. Krill 130 proteases have been researched thoroughly and are well defined. Sjodhk et al (2002) have identified three trypsin-like proteinases, and four carboxypeptidases. The trypsin-like proteinases were as much as 60 times as active as bovine trypsin. Yoshitomi (2005) found that crude krill digestive proteases were much more active in the Antarctic summer season (December to February) than the winter season and was correlated with phytoplankton abundance. Benjakul et al. (1996) recovered a major proteinase from Pacific whiting surimi wash water which was cathepsin L with molecular weight of 39.5 kDa. These researchers also characterized biochemical properties and stability of the proteinase recovered from Pacific whiting surimi wash water by ohmic heating, ultrafiltration, and freeze-drying (Benjakul et al., 1997) and found that they were stable at pH range of 3.0-9.0, having the highest activity at pH 4.0.

Collagenolytic Enzymes Collagenolytic enzymes are capable of degrading the polypeptide backbone of native collagen under conditions that do not denature the protein (Kim et al., 2002). They can be divided into serine collagenases and metallocollagenases due to different physiological functions that they posses. Serine collagenases show their greatest activity in protein digestion, blood-clotting, fibrinolysis, complement activation and fertilization whereas metallocollagenases are involved in remodeling the extracellular matrix (Park et al., 2002). Collagenolytic enzymes have been extracted from shrimp (Brauer et al., 2003; Oh et al., 2000; Van Wormhoudt et al., 1992), cod (Kristjansson et al., 1995), filefish (Kim et al., 1991), yellow-tail pyloric caeca (Yoshinaka et al., 1972), crab (Grant et al., 1981; Grant et al., 1983; Klimova et al., 1990; Roy et al., 1996), winter flounder (Teruel et al., 1995), and mackerel (Park et al., 2002). Related studies regarding its effect on autolysis of fish muscle tissues during storage period include cod (Hemandez-Herrero et al., 2003), and salmon (Hultmann et al., 2004). These enzymes were most active in the pH range of 6.5-8.0 and are inactivated at pH < 6.0 (Haard et al., 1994). Collagenase from internal organs of mackerel was purified and 131 characterized, and its molecular weight was estimated to be 14.8 kDa, and the optimum pH and temperature were approximately pH 7.5 and 55 0C, respectively (Park et al, 2002).

Lipases Lipase is an enzyme produced primarily in the pancreas and hydrolyzes fatty acids from the glycerol backbone at the hydrophobic/hydrophilic interface of the lipid substrate. It is necessary for the absorption and digestion of lipid including mono-, di-, and triacylglycerols. Lipase has been purified and characterized from various aquatic organisms including cod (Gijellesvik et al., 1992), salmon (Gjellesvik et al., 1994), stomach, pyloric caeca, liver, and intestine of rohu, oil sardine, mullet, Indian mackerel (Nayak et al., 2003), red sea bream hepatopancreas (lijima et al., 1998) and others. The optimum pH of lipase ranges from 6.0 to 9.0, and lipase is normally stable between 30-45 0C. These lipases could be suitable for various kinds of food applications because of their high activity at mild pH and temperature conditions. Phospholipases have also been extracted from pollock muscle (Audley et al., 1978), Atlantic cod muscle (Chawla et al., 1987), gill membranes of red sea bream (Uchiyama et al., 2005), and hepatopancreas of red sea bream (lijima et al., 1997; lijima et al., 1990). Recently, phospholipase A2 isozyme from starfish {Asterina pectinifera) was characterized (Kishimura et al., 2004) demonstrating that phospholipase A2 isozyme II had an optimum temperature of 50 0C and pH of 9.0. It did not show fatty acid specificity for hydrolysis of phosphatidylcholine; however, it had 10 times the activity than that of commercially available porcine pancreatic phospholipase A2 and has been suggested as a potential source for phospholipase A2 for industrial production. Lipase from tilapia stomach and intestine were extracted and characterized by Taniguchi et al. (2001). The study showed that this lipase had a molecular weight of 54 kDa with optimum pH of 6.5, and optimum temperature of 40 0C. The lipases were stable at the pH range of 5.0 to 7.0 at 40 0C for 30 min while lipase from tilapia intestine showed its molecular weight of 46 kDa, optimum temperature of 35 0C, and 132 was stable in the pH range of 6.5 to 8.5. The Upases were identified as non-specific in terms of position on the glycerol backbone but preferentially did hydrolyzed triacylglycerol rather than di- and monoacylglycerol. The highest lipase activity was found when soybean oil was used for substrate showing 100% hydrolysis. A bile salt-activated lipase from hepatopancreas of red sea bream was extracted and characterized (lijima et al., 1998). Molecular weight of this lipase was determined as approximately 64 kDa and had a pH optimum of 7.0 to 9.0. This lipase was homologous to mammalian bile salt-activated lipase which preferentially hydrolyzed ethyl esters of polyunsaturated fatty acids.

Chitinolytic Enzymes Chtinolytic enzymes are widely distributed in crustaceans and play an important role in the degradation of chitin. Two chitinolytic enzymes are involved in crustaceans, which are endo-type (chitinase) and exo-type (p-N-acetylhexosaminidase), which are related to molting in several crustaceans (Kono et al., 1995). These enzymes have been purified mainly from processing byproducts such as squid liver (Matsumiya, 2004), shrimp waste silage (Matsumoto et al., 2004), and shrimp byproducts (Olsen et al., 1990). Chitinase was purified from the stomach of red sea bream (Pagrus major) and kinetic analysis (Karasuda et al., 2004) showed high enzyme activity at pH 2.5 and 9.0 toward glycolchitin, and at pH 2.5 and 5.0 toward N-acethylchitopentasaccharide. These multiple pH-peak activities are unique for chitinases. Prawn shell waste was used as a raw material for chitinase production by the marine fungus Beauveria bassiana BTMF S10 by solid state fermentation (Suresh et al., 1998). Likewise, shrimp waste silage was used for production of P-N- acetylhexosaminidase by Verticillium lecanii in submerged and solid state fermentations taking advantage of the abundance and composition of crustacean wastes (Matsumoto et al, 2004). They concluded that shrimp silage was an efficient inducer of the extracellular enzyme, compared with media supplemented with sucrose where enzymic activity was not detected. 133 Transglutaminases Transglutaminase is an aminoacyltransferase that catalyzes an acyl transfer reaction between y-carboxyamide groups of glutamine residues in polypeptides and proteins. When the e-amino group of lysine acts as an acyl acceptor, £-(y-glutamyl) lysine crosslinks are formed in proteins which change protein texture and structure (Kumazawa, 2002). Protein properties, gelation capability, thermal stability, and water-holding capacity are uniquely affected by this crosslinkage (Kuraishi et al.,

2001). A study showed that seafood products contained the highest level of S-(Y- glutamyl) lysine crosslinks (43 (a.moles/100 g protein) followed by meat products (38 |a,moles/100 g protein) and soy product (15 [imoles/100 g protein) (Kuraishi et al., 2001). Transglutaminases have been purified and characterized from walleye pollock liver (Kumazawa et al., 1996), Japanese oyster (Kumazawa et al., 1997), red sea bream muscle (Noguchi et al., 2001; Yasueda et al., 1994), botan shrimp squid, carp, rainbow trout, atka mackerel (Nozawa et al., 1997), squid gill (Nozawa et al., 2001), scallop striated adductor muscle (Nozawa et al., 2001), and tilapia (Worratao et al., 2003; Worratao et al., 2005). Two different kinds of transglutaminases were extracted from the Japanese oyster (Kumazawa et al., 1997) and had molecular weights of 84 and 90 kDa with optimum pH of 8.0 for both enzymes. The optimum temperature differed and showed 40 and 25 0C, respectively. Squid gill was used for transglutaminase extraction due to its high activity compared with other tissues (Nozawa et al., 2001). Its molecular weight was estimated to be 94 kDa and its activity showed an optimum pH and temperature at 8.0 and 20 0C, respectively with 10 mM CaC^. Tissue transglutaminase from scallop striated adductor muscle was purified (Nozawa et al., 2001), and its molecular weight was estimated to be 95 kDa. Optimum pH and temperature were pH 8.0 and 35 0C, respectively. This enzyme was Ca dependent and inactivated by p-chloromercuribenzoic acid, N-ethylmaleimide, Cu , and Zn showing that it belongs to the thiol group of enzymes. 134 Polyphenoloxidases Polyphenoloxidase which is also known as tyrosinase, polyphenolase, phenolase, catechol oxidase, cresolase, and catecholase is widely distributed in nature (Chen et al., 1991). Polyphenoloxidase oxidizes diphenols to quinones, which undergo autoxidation and polymerization to form dark pigments in fruits, vegetables, and crustacean species (Bartolo et al., 1998). Of these, crustacean species can be a source for polyphenoloxidase from seafood byproducts of the shellfish processing industry. Polyphenoloxidase can be found mainly in shellfish and byproducts including shrimp (Nakagawa et al., 1992), prawns (Montero et al., 2001), lobster (Opoku-Gyamfua et al, 1992) and cuttlefish (Zhou et al., 2004). A polyphenoloxidase fraction was isolated and characterized from lobster (Opoku-Gyamfua et al, 1992) using the skin layer between the muscle and the exoskeleton (Homarus americanus) and compared with those of commercial tyrosinase. The lobster polyphenoloxidase fraction was activated by trypsin, and it showed a more heat labile nature as compared with commercial tyrosinase. The enzyme was most stable at pH 7.5, while tyrosinase exhibited a much broader pH stability ranging between 6.5 and 10.0. The thermotolerance of polypenoloxidase activity could depend on its source and the environmental factors under which the species are grown. One of the significant environmental factors is water temperature although polyphenoloxidase from shrimp are usually stable between 30 "and 50 0C. Polyphenoloxidase activity from Florida spiny lobster (Panulirus argus) and Western Australian lobster (Panuliruscygnus) was studied (Chen et al., 1991). Both enzymes showed similar characteristics that catalyzed oxidation of catechol and dl-p-3, 4- dihydroxyphenylalanine and showed optimum pH stability at pH 7.0. However, they differed with respect to activation energy and thermal stability. Western Australian lobster polyphenoloxidase showed decreased activity when preincubated at temperatures greater than 30 0C whereas that of Florida spiny lobster showed greater stability at 35 0C. Chen et al. (1991) suggested that this is because Florida spiny lobsters live in warm water areas while Western Australian lobsters live in cold water 135 areas. Authors stated that these differences in environmental conditions of their natural habitats may account for the difference in optimal thremostability between the enzymes.

PRODUCING ENZYMES FROM SEAFOOD PROCESSING BYPRODUCTS Enzyme extraction/solubilization, concentration, fractionation, and purification are the main steps of enzyme production. To produce enzymes at the industrial level from seafood byproducts, there are several requirements including: 1) a large quantity of raw material is needed because the target enzymes are usually in very low concentration; 2) facilities are required to efficiently remove significant amounts of water to concentrate/purify enzyme; and 3) the removal of particulate matter (including cell and/or cell debris) from the final product is required. The following applications are examples achieving these objectives using a combination of several techniques.

Enzyme extraction/ solubilization Extraction of enzymes largely depends on the localization of the target enzymes. For instance, if the starting material contains intracellular enzymes, cell membrane should be disrupted and homogenized in a buffer. The membrane disruption can be done by various methods; alkalization, addition of Ethylenediaminetetraacetic acid (EDTA), detergents or osmotic shock, or by physical methods, including sonication, alternating freezing and thawing phases, solid or liquid shear, or grinding or agitation with abrasives. Then, insoluble impurities and cellular waste should be removed by filtration or centrifiigation (Barthomeuf, 1989). On the other hand, if the enzyme is located in the extracellular fluid, cells should be removed by centrifiigation to eliminate contaminants.

Enzyme concentration As mentioned earlier, the target enzyme is normally of very low concentration in a large volume of raw material; therefore, the concentration of enzymes after 136 extraction is necessary. To maintain enzyme activity, harsh conditions such as high temperature and strong mechanical stress should be avoided. The enzyme concentration process includes removal of water and other molecules or use of precipitation methods. The following methods are examples that can be applied with relatively mild condition throughout the process.

Membrane Technology The food industry started incorporating membrane technology with reverse osmosis technology for water purification as well as ultrafiltration technology for concentration of products (Cuperus et al., 1993). Since then, membrane processes have become major tools in food processing and for extraction of bioactive ingredients. A membrane is a thin sheet of artificial or natural polymeric material with pores distributed throughout the material. Under certain pressure, the membrane rejects large particles, and allows species smaller than the pores to pass through (Morrissey et al, 2005). Membrane technology includes microfiltration, ultrafiltration, nanofiltration, reverse osmosis and electrodialysis. Ultrafiltration, where particles smaller than 0.2 jam pass through the membrane, is the most widely used process for protein and enzyme recovery. Ultrafiltration has been used extensively in many industries including the pharmaceutical as well as the food and beverage industry. A good example of membrane technology and use is the dairy industry to recover and separate out different whey proteins (Dsouza et al., 2005). Ultrafiltration is a pressure-driven process and works by passing low-molecular-weight products through the membrane while retaining high-molecular-weight compounds such as protein. Cellulose acetate is one of the traditional materials for ultrafiltration membrane. Compared to others such as polyacrylnitrile and polyethersulphone, it is extremely hydrophilic. Having less fouling problems, it tends to have weak resistance against heat and chemicals. With the development of chemical and heat stable membrane such as polysulphone membrane, the feasibility of this technology has greatly increased. 137 Protease recovery from surimi wash water containing various cathepsins was studied using ultrafiltration technology with various pretreatments (DeWitt et al., 2002). Preliminary tests showed that fish proteins will rapidly clog membrane pores actually changing the dynamics of the filtration (Huang et al., 1998). Acidification of wash water (pH 4.0) and mild heat treatment (60 0C) followed by centrifugation were necessary as pretreatments before ultrafiltration was applied. This helped to remove high-molecular-weight proteins that might interfere with ultrafilation without reducing protease activity. The supernatant was then subjected to both high-molecular-weight ultramicrofiltration and low-molecular-weight ultrafiltration. The result showed that concentration of protease using 50 kDa ultrafiltration polysulphone membranes was successful in recovering approximately 80 % of original protease activity (DeWitt et al., 2002).

Enzyme concentration by precipitation An alternative and relatively traditional way to concentrate enzymes is to use salts for enzyme precipitation. This is due to its ability to change electrostatic forces which affect enzyme solubility. Enzymes are usually soluble in water/buffer solutions because hydrophobic residues tend to locate interior of the globular proteins (enzymes). At low salt concentration (0.5 -1 M), protein solubility increases (salting in) as ions from salt decrease the electrostatic attraction between opposite charges of neighboring molecules. At high salt concentration (>1M), protein solubility decreases resulting in protein precipitation. The protein can also be precipitated by adjusting pH to pi of proteins which is usually between 5.2 and 5.5. However, use of salt, especially ammonium sulphate (NH^SC^, is the most common way to precipitate and concentrate enzymes. Ammonium sulfate possesses relatively strong "salting out" effect without causing significant levels of protein denaturation. It has been intensively used in research area including seafood and seafood byproducts as a first step to accomplish separation of crude enzymes. The advantages of using these tecniques include: 1) high solubility, 2) high commercial availability with inexpensive costs, 3) lack of toxicity, and 4) its stabilizing effect on precipitates. A disadvantage 138 of using mineral salts is that the method requires another step such as dialysis or gel filtration (Barthomeuf, 1989). In industries where enzymes are purified on a large scale, the corrosion of stainless steel by ammonium sulphate is a disadvantage and possibly causes additional environmental concerns. Sodium sulphate may be more suitable for this point of view, but it is not as effective (Naz, 2002).

Freeze-drying or lyophilization Freeze-drying, also known as lyophilization, has been widely used in the food industry for various protein products including enzymes. Freeze-drying removes water from a frozen sample by sublimation and desorption in a three-step process, which includes freezing, primary drying and secondary drying. Freeze-drying technology is used in combination with other concentration/purification methods including ultrafiltration in seafood enzyme recovery research. Dry powder forms of enzymes are more stable than enzymes in aqueous solution, since water can facilitate enzyme denaturation. However, it is also known that conformational changes of enzymes by freeze-drying can result in lower enzyme activity. Alternative forms of enzyme preparations have been developed to increase enzyme stabilization, including immobilization with sol-gel methods, cross-linked enzyme crystals (Altus Biologies, Inc. Cambridge, MA), soluble enzymes with polymers such as ethylene glycol, and surfactant-modified enzymes such as sorbitan monostearate-modified lipase (Roy et al., 2004).

Fractionation/purification Concentration techniques will remove considerable extraneous material but only partially purify the enzyme. The degree of purity requires further downstream processing that often include chromatography and electrophoresis. Since electrophoresis is mainly an analytical procedure, only chromatography will be discussed here. 139 Chromatography Although there are limitations for industrial scale use of chromatography, it is frequently used in the laboratory to fractionate components including enzymes with high degrees of purity. Chromatography technology can be categorized into various methods based on protein size, charge, hydrophobicity and molecular recognition. These include size exclusion chromatography, ion exchange chromatography, and affinity chromatography. Enzymes can be fractionated depending on their hydrodynamic volume by size exclusion chromatography whereas affinity chromatography is based on specific recognition between two relevant biomolecules (An et al., 2000). Affinity chromatography is commonly used for the separation of one enzyme from others. Fractionation of target enzyme by ion exchange chromatography is based on the electrostatistic interactions between the enzyme and charged groups on the exchangers. Cellulosic ion exchange chromatography has been one of the most common methods for protein separation (Naz, 2002). Separation of enzymes can be controlled by changing pH of the elution buffer. Gastricsin-like proteinase was purified from Atlantic cod viscera after lipophilization with a single step purification scheme on ion-exchange of Amberlite CG-50 with efficient recovery (Amiza et al., 2002). Crude cod pepsinogen was dissolved in 10 ml of 0.2 M sodium citrate buffer (pH 2.1) and introduced to the column. The pH of elution buffer was changed to 3.8, 4.2, and 4.6 to fractionate proteinases followed by pooling the target peaks, dialyzing, and freeze-drying to produce purified proteinaes. Pepsin A was eluted at a lower pH (pH 4.0) while gastricsin appeared at higher pH (pH 4.4). The study demonstrated that final proteinase recovery was high which could be related to the salt activation by citrate buffer or removal of inhibitors. Chen et al. (1997) investigated the purification methods for shrimp polyphenoloxidase from frozen powder of white shrimp {Penaeus setiferus) and pink shrimp {P duorarum) and found that the use of butanol treatment followed by Phenyl Sepharose CL-4B chromatography was better than ammonium sulphate fractionation and then phenyl sepharose chromatography. They also found that different species exhibited different activity through the purification process 140 which demonstrated that activity of white shrimp polyphenoloxidase was more susceptible than that of pink shrimp during the process. Taniguchi and coworkers (2001) extracted lipase from the stomach and intestines of tilapia. They extracted crude lipase by chromatofocusing and applying the extarct on a polyexchanger column previously equilibrated with 25 mM imidazole- acetic acid buffer. The lipase was eluted with polybuffer 96-acetic acid. lijima et al. (1998) characterized a bile salt-activated lipase from hepatopancreas of red sea bream. A dehpidated powder processed from red sea bream hepatopancreas was used as a raw material. Lipase was extracted by fractional precipitation with ammonium sulphate and sequential chromatography.

MARINE BYPRODUCT ENZYME UTILIZATION Despite extensive research in marine enzyme technology, there are only a few applications in the food processing sector. Limitations of marine byproduct enzyme utilization are often due to the cost of enzyme recovery and competition with more mainstream enzyme sources. However, there are some commercial operations that use marine byproduct enzymes such as the enzyme recovery process from cod viscera decribed in Fig.l (Gildberg, 2004). This figure shows a Norwegian muti-purpose processing plant manufacturing several products including a protein hydrolysate and crude pepsin extract as a fish processing aid for fish caviar and descaling operations. The crude extract will vary in pepsin concentration from 2 to 10%, depending on the quality and source of raw material and operating conditions. Cod viscera (or stomach) + 2% formic acid (pH 3) Storage: 3 days at 25 0C (Autolysis)

Protein Oil/emulsion Decanter hydrolysate

Steam Peptone Separator Pepsin U.F. Extract

Protein Oil Microbial Caviar sludge Feed gr. media production

Figure 1 Flowchart of byproduct processing and crude extract production using cod viscera (Gildberg 2004). 142 Crude extracts can undergo further separation by chromatography or other methods to produce a purified enzyme. However, these are often expensive procedures and most commercial uses require only the crude extract form except in the medical/biotechnology field. Several of these uses from traditional seafood products to advanced biochemical/medical applications are described below.

Fermented fish products Fish fermentation products, such as fish sauces, are expanding in the marketplace and there have been increased efforts to better define their manufacturing parameters and control the fermentation process (Lopetcharat et al., 2001). More success has been found using enzymes to hasten the process and quality of fish sauce production. Researchers (Chaveesuk et al., 1993; Tungkawachara et al., 2003) have shown the efficiency of using specific enzymes for the production offish sauce and producing an acceptable product. Maatjes is a fermented product using the natural viscera enzymes in herring and several efforts have been made to describe the enzymatic reaction and mimic the end result (Nielsen et al., 1997; Nunes et al., 1997; Olsenetal., 1997).

Protein hydrolysates There is an increasing amount of interest in protein hydrolysate production as this product has shown unique biological properties such as stimulating the immune system as well as having peptide fractions that stimulate growth. A good general review on hydrolysates can be found by Kristinsson and Rasco (2000). Recent research in the area of production hydrolysates from byproducts include threadfin bream (Normah et al., 2005), cod byproducts (Slizyte et al., 2005; Slizyte et al., 2004), gold carp (Sumaya-Martinez et al., 2005), salmon heads (Gbogouri et al., 2004), mackerel (Wu et al., 2003), and herring (Sathivel et al., 2003). Although enzymes from plant and microbial sources are commonly used for the production of fish protein hydrolysates, research with marine enzymes has shown comparable production rates. Shahidi and his coworkers have produced fish protein 143 hydrolysates using mixtures of extracted and endogenous enzymes in both capelin and seal meat (Shahidi et al., 1995; Shahidi et al., 1997).

Deskinning The enzymatic removal of fish skin from certain species has proven advantageous as an alternative method for mechanical deskinning (Kim et al., 1993; Tschersich et al., 1998) since mechanical methods tend to be harsh processes resulting in lower fillet recovery. The use of proteases for removal of squid skin is a standard process, although non-marine proteases are often used. Several researchers have shown that marine enzymes can be used and often leaves a product that has several advantages over papain or ficin proteases (Wray, 1988).

Fish roe (Caviar) production Caviar is the salt-cured and preserved eggs of aquatic animals that have been separated from the supporting connective tissue. The most widely recognized and valued caviar is made from sturgeon harvested from the Caspian Sea (Bledsoe et al, 2003). There are enzyme-based processes for removing the connective tissue that surrounds the eggs, which reduces human handling and increases caviar recovery. Commercially available fish enzymes such as Rozym™ (Biotec-Mackzymal AS, Tromso, Norway) and Digestase™ (Alaska Russia Salmon Caviar Co., Anchorage, AK, U.S.A.) have been used for salmon caviar. Enzymes from fish viscera such as pepsin can be also used for skein removal. Pepsins split the linkages that adhere the egg cells to the roe sack without affecting the eggs. Such application with marine derived pepsin has advantages over the enzymes from mammalian origin including higher optimum pH and higher activity at lower temperatures (Raa, 1996). Research has shown the efficacy of proteolytic enzymes for higher yields in fish caviar for salmon and lump fish (Raa, 1997). 144 Gel formation of food Transglutaminase can be used in certain food products such as surimi, seafood and meat products that require improved gel formation and gel strength resulting in better texture. The advantages of transglutaminase addition in the surimi products include: 1) to increase gel strength of surimi resulting in higher quality, and 2) to lower production costs by increasing water content in surimi (Kumazawa, 2002). Currently, transglutaminase has also been used for meat products, noodles, soy protein products such as tofu, dairy and baked products. Commercially available trasglutaminase, such as Activa TG-K has a high potential for food application but to date are only from microbial sources (Kuraishi et al, 2001).

Meat tenderization Currently, proteases from plant sources, such as papain and bromelain, are the principle enzymes used as meat tenderizers (Haard et al., 1994). However, these enzymes attack both connective and myofibrillar proteins which often lead to over- tenderization whereas marine collagenase could only hydrolyze connective tissue protein possibly resulting in optimum tenderization of meat products. Aoki et al. (2004) extracted collagenase from Northern shrimp by products and suggested potential use for the meat industry in reducing toughness of the meat products cased by connective tissues claiming that it works better than currently available enzymes from plant sources.

Biotechnology/medical Although many of the marine enzyme technologies are in the initial phases, there have been notable successes in biochemistry/biotechnology field. The isolation and purification of shrimp alkaline phosphatase (SAP) from cold water shrimp {Pandalus borealis) has led to the use of this compound in gene splicing with plasmid and bacteriophage vectors in most biotechnology laboratories (Olsen et al., 1991; Sambrook et al., 2001). SAP completely dephosphorylates DNA has the advantage over other alkaline phosphatases as it can be inactivated at lower temperatures (65 0C 145 for 15 min) and thus not denature DNA materials. The enzyme is recovered and purified from the shrimp processing waste water and is a highly sought after chemical. This success has created other potential opportunities with byproduct enzymes and Biotec, ASA in Norway also lists cod uracil-DNA glycosylase and other enzymes for use in biotechnology and food processing fields. Atlantic cod viscera is an abundant fishery byproduct in Iceland and the purification and characterization of trypsin followed by chymotrypsin from the viscera has led to its utilization (Asgeirsson et al., 1991; Asgeirsson et al., 1989). After ten years of research and collaboration between University of Virginia and University of Iceland, researchers developed a new product using both trypsin and chymotrypsin from cod viscera. The product, called Penzim, is used as a gel and lotion in the treatment of skin ailments including psoriasis and other skin conditions. Recently, researchers have reported the successful cloning and expression of trypsin I from cod in E. coli (Jonsdottir et al., 2004). There also has been active research in investigating krill proteolytic enzymes as an active ingredient in wound debridement or the removal of necrotic tissue in wounds (Mekkes et al., 1997). Research suggests that it is more active than several plant proteases and can help accelerate wound healing.

FUTURE TRENDS Despite tremendous scientific strides in identification and purification of marine enzymes and technological advances in the recovery of specific compounds from fish byproducts, the question remains concerning the economic feasibility of enzyme recovery from fish and shellfish byproducts. Many of the specific enzymatic activities, e.g. proteases for fish hydrolysates, have been preempted by other enzymes from plant and microbial sources that often prove to be less expensive. In addition to scale-up limitations, there is also strong market competition for the byproduct raw material. Many fish processing operations already use solid byproducts for established-market products such as fish meal, fertilizers, silage and more recently hydrolysates. As aquaculture expands over the next decade, demand for the raw 146 material for producing fish meal and other products for fish feed formulations will also increase (Kilpatrick, 2002). These products have relatively stable global markets in which there are known risks, capital investment needs and available technology which facilitates entry into the marketplace. Other markets for fish byproducts have also developed over the last decade. Viscera byproducts, e.g. stomachs from the Alaska pollock industry and other organs, are now being marketed into Asian niche markets (Morrissey et al., 2005). Extraction of specific enzymes from fish byproducts requires considerable investment in technologies that often have high costs, varying efficiencies and require skilled technicians which can be problematic in remote areas. Some of the best examples of utilization of marine enzymes from seafood byproducts come from the biotechnology/medical field. Companies such as Biotec Pharmacon ASA in Tromso, Norway, which produces shrimp alkaline phosphatase for biotechnology laboratories and Zymetec in Iceland currently marketing a marine trypsin as a skin healing lotion, are examples of companies that have successfully taken extractive byproduct research into the marketplace. Perhaps the future of marine enzymes utilization rests more in biotechnology/medical uses than in food processing; however, even this is a two-edged sword. Rapid advances in biotechnology have also revolutionized the field of enzyme prodcution providing researchers and companies with the potential to produce specific enzymes more economically. Haard (1998) addressed the uniqueness of aquatic enzymes in their diverse environments in his review of "specialty enzymes". It is unlikely that many of these marine enzymes could be produced in large enough quantities due to limitations in obtaining sufficient raw material. However, their unique properties may have industrial or medical applications that warrant production through biotechnological techniques. The advent of biotechnology and the production of specific compounds through gene transfer and use of microbial organisms for production are making inroads in the commercial production of enzymes. As this technology continues to develop over the next decade, marine enzymes with unique characteristics could be cloned and produced more economically through biotechnology than from byproduct recovery operations. Even in the case of shrimp alkaline phosphatase and penzim, there is active research to 147 transfer these genes to bacteria which would be able to produce commercial quantities of this valuable enzyme. In many cases, the scientific information about biochemical properties of the marine enzymes themselves might prove to be the most valuable component of the byproduct itself. Although the science of seafood enzyme research remains an exciting one due, in part, to the uniqueness of its resources and the opportunities in may provide in the biotechnology/medicine fields, the question remains whether it will reach its true potential beyond the laboratory setting and become a viable force in the global marketplace. 148

REFERENCES Amiza MA, Owusu R. 2002. A single step purification of gastrinsin-like proteinease from Atlantic Cod (Gadus morhua). J Bio Sci 2: 591-595. An H, Seymour TA, Wu J, Morrissey MT. 1994. Assay systems and characterization of Pacific Whiting {Merluccus prodctus) protease. J Food Sci 59: 277-281. An H, Visessanguan W. 2000. Recovery of enzymes from seafood-processing wastes, p. 641-664. In: Haard NF, Simpson BK (eds.). Seafood Enzymes. Marcel Dekker Inc., New York. Aoki H, Ahsan N, Matsuo K, Hagiwara T, Watabe S. 2004. Partial purification of proteases that are generated by processing of the Northern shrimp Pandalus borealis and which can tenderize beef. Inter J Food Sci Technol 39: 471-480. Aoki T, Yamashita T, Ueno R. 1995. Purification and characterization of a cathepsin B-like enzyme from mackerel white muscle. Fish Sci 61: 121-126. Asgeirsson B, Bjamason JB. 1991. Structural and kinetic properties of chymotrypsin from Atlantic cod (Godus morhua). Comp Biochem Physiol 99: 327-335. Asgeirsson B, Fox JW, Bjamason JB. 1989. Purification and characterization of trypsin from the poilikotherm Gadus morhua. Eur J Biochem 189: 85-94. Ashie INA, Simpson BK. 1996. Application of high hydrostatic pressure to control enzyme related fresh seafood texture deterioration. Food Res Inter 29: 569-575. Audley MA, Shetty KJ, Kinsella JE. 1978. Isolation and properties of phospholipase A from pollock muscle. J Food Sci 43: 1771-1775. Barthomeuf C. 1989. Review application of enzyme purification process to proteolytic enzymes. J Chromato 473: 1-26. Bartolo I, Birk E. 1998. Some factors affecting Norway lobster (Nephrops norvegicus) cuticle polyphenol oxidase activity and blackspot development. J Food Sci Technol 33: 329-336. Benjakul S, Seymour TA, Morrissey MT, An H. 1996. Proteinase in Pacific whiting surimi wash water: identification and characterization. J Food Sci 61: 1165- 1170. Benjakul S, Seymour TA, Morrissey MT, An H. 1997. Characterization of proteinase recovered from pacific whiting surimi wash water. J Food Biochem 22: 1-16. Bledsoe GE, Bledsoe CD, Rasco B. 2003. Caviars and Fish Roe Products. Crit Rev in Food Sci Nutr 43: 317-356. Brauer JME, Leyva JAS, Alvarado LB, Sandez OR. 2003. Effect of dietary protein on muscle collagen, collagenase and shear force of farmed white shrimp (Litopenaeus vannamei). Eur Food Res Technol 217: 277-280. Cardenas-Lopez JC, „ Haard NF. 2005. Cysteine proteinase activity in jumbo squid (Dosidicus gigas) hepatopancrease extracts. J Food Biochem 29: 1745-1753. Cavicchioli R, Siddiqui KS, Andrews D, Sowers KR. 2002. Low-temperature extremophiles and their applications. Curr Opini Biotechnol 13: 253-261. Chaveesuk R, Smith JP, Simpson BK. 1993. Production offish sauce and acceleration of sauce fermentation using proteolytic enzymes. Aquat Food Prod Tech 2: 59- 77. 149 Chawla P, Ablett RP. 1987. Detection of microsomal phospholipase activity in myotomal tissue of Atlantic cod (Gradus morhua). J Food Sci 52: 1194-1197. Chen JS, Charest DJ, Marshall MR, Wei CI. 1997. Comparison of two treatment methods on the purification of shrimp polyphenol oxidase. J Sci Food Agric 75: 12-18. Chen JS, Rolle RS, Marshall MR, Wei CI. 1991. Comparison of phenoloxidase activity from Florida spiny lobster and Western Australian lobster. J Food Sci 56: 154-157. Choi YJ, Cho Y-J, Lanier TC. 1999. Purification and characterization of proteinase from Atlantic menhaden muscle. J Food Sci 64: 772-775. Crawford AC, Richardson NR, Mather PB. 2005. A comparative study of cellulase and xylanase activity in freshwater crayfish and marine prawns. Aquacul Res 36: 586-592. Cuperus FP, Nijhuis HH. 1993. Applications of membrane technology to food processing. Trends Food Sci Technol 4: 277-282. DeWitt CAM, Morrissey MT. 2002. Parameters for the recovery of proteases from surimi wash water. Bio Technol 81: 241-247. DsouzaNM, Mawson AJ. 2005. Membrane cleaning in the dairy industry: review. Crit Rev Food Sci Nutr 45: 125-134. Ehara T, Tamiya T, Tsuchiya T. 1992. Investigation of myosin heavy chain-degrading proteinase in decapoda muscle. Nippon Suisan Gakkaishi 58: 2379-2382. Ehara T, Tamiya T, Tsuchiya T. 1994. Investigation of myosin heavy chain-degrading proteinase in squid muscle. Nippon Suisan Gakkaishi 60: 527-528. Gbogouri GA, Linder M, Fanni J, Parmentier M. 2004. Influence of hydrolysis degree on the functional properties of salmon byproducts hydrolysates. J Food Sci 69: 615-622. Gerday C, Aittaleb M, Bentahir M, Chessa JP, Claverie P, Collins T, D'Amico S, Dumont J, Garsoux G, Georlette D, Hoyoux A, Lonhienne T, Meuwis MA, Feller G. 2000. Cold-adapted enzymes: from fundamentals to biotechnology. Tibtech 18: 103-107. Gijellesvik DR, Lombardo D, Walther BT. 1992. Pancreatic bile salt dependent lipase from cod (Gadus morhua): purification and properties. Biochem Biophys Acta 1124: 123-134. Gildberg A. 2004. Enzymes and bioactive peptides from fish waste related to fish silage, fish feed and fish sauce production. J Aquat Food Prod Technol 13:3- 11. Gildberg A, Simpson BK, Haard NF. 2000. Use of enzymes from marine organisms, p. 619-639. In: Haard NF, Simpson BK (eds.). Seafood Enzymes. Marcel Dekker, New York. Gjellesvik DR, Lorens JB, Male R. 1994. Pancreatic carboxylester lipase from Atlantic salmon (Salmo salar) cDNA sequence and computer-assisted modeling of tertiary structure. Eur J Biochem 226: 603-612. Grant GA, Eisen AZ, Bradshaw RA. 1981. Collagenolytic protease from fiddler crab (Ucapugilator). Methods Enz 80: 722-734. 150 Grant GA, Sacchettini JC, Welgus HG. 1983. A collagenolytic serine protease with trypsin-like specificity from fiddler crab Ucapugilator. Biochem 22: 354-356. Gudmundsdottir A, Palsdottir HM. 2005. Atlantic cod trypsins: From basic research to practical applications. Mar Biotechnol 7: 77-88. Haard NF. 1998. Specialty enzymes from marine organisms. Food Technol 52: 64-67. Haard NF, Simpson BK. 1994. Proteases from aquatic organisms and their uses in seafood industry. Chapman and Hall, New York. Haard NF, Simpson BK. 2000. Seafood Enzymes. Marcel Dekker, New York. Haard NF, Simpson BK, Sikorski ZE. 1994. Biotechnological applizations of seafood proteins and other nitrogenous compounds, p. 194-222. In: Sikorski ZE, Sun PB, Shahidi F (eds.). Seafood Proteins. Chapman and Hall, New York. Hemandez-Herrero MM, Duflos G, Malle P, Bouquelet S. 2003. Collagenase activity and protein hydrolysis as related to spoilage of iced cod (Gadus morhua). Food Reslnt36: 141-147. Heu MS, Kim HR, Pyeun JH. 1995. Comparison of trypsin and chymotrypsin from the viscera of anchovy, Engraulis japonica. Comp Biochem Physiol 112: 557-567. Huang L, Morrissey MT. 1998. Fouling of membranes during microfiltration of surimi wash water: Roles of pore blocking and surface cake formation. J Membrane Scil44: 113-123. Hultmann L, Rustad T. 2004. Iced storage of Atlantic salmon {Salmo salar)- effects on endogenous enzymes and their impact on muscle proteins and texture. Food Chem 87: 31-41. lijima N, Nakamura M, Uematsu K, Kayama M. 1990. Partial purification and characteriazation of phospholipase A2 from the hepatopancreas of red sea bream. Nippon Suisan Gakkaishi 56: 1331-1339. lijima N, Chosa S, Uematsu K, Goto T, Hoshita T, Kayama M. 1997. Purification and characterization of phospholipase A2 from the pyloric caeca of red sea bream, Pagrus major. Fish Physiol Biochem 16: 487-498. lijima N, Tanaka S, Ota Y. 1998. Purification and characterization of bile salt- activated lipase from the hepatopancreas of red sea bream. Fish physiol and Biochem 18: 59-69. Johnson HM. 2004. Annual report on the United States Seafood Industry (11th edition) Jacksonville, OR. Jonsdottir G, Bjamason JB, Gudmundsdottir A. 2004. Recombinant cold-addapted trypsin I from Atlantic cod - expression, purification and identification. Protein ExprPurif33: 11-22. Karasuda S, Yamamoto K, Kono M, Sakuda S, Koga D. 2004. Kinetic analysis of a chitinase from red sea bream, Pagrus major. Biosci Biotechnol Biochem 68: 1338-1344. Kilpatrick JS. 2002. Fish processing waste: opportunity or liability. In: Advances in Seafood Byproducts. Conference Proceedings, Fairbanks, AK. Kim SK, Byun HG, Choi KD, Roh HS, Lee WH, Lee EH. 1993. Removal of skin from filefish using enzymes. Bull Korean Fish Soci 26: 159-172. 151 Kim SK, Park PJ, Kim JB, Shahidi F. 2002. Purification and characterization of a collagenolytic protease from the filefish, Novoden modestrus. J Biochem Mol Biol 35: 165-171. Kim YT, Kim SK. 1991. Purification and characterization of the collagenase from the tissue of filefish, Novoden modestrus. Korean Biochem J 24: 401-409. Kishimura H, Hayashi K. 2004. Purification and properties of phospholipase A2 isozymes from pyloric ceca of the starfish {Asterina pectinifera). J Food Biochem 28: 181-194. Klimova OA, Borukhov SI, Solovyeva NI, Balaevskaya TO, Strongin AY. 1990. The isolation and properties of collagenolytic proteases from crab hepatopancreas. Biochem Biophys Res Commun 166: 1411-1420. Kono M, Wilder MN, Matsui T, Furukawa K, Koga D, Aida K. 1995. Chitinolytic enzyme activities in the hepatopancreas, tail fan and hemolymph of kuruma prawnpenaeus japonicus during the molt cycle. Fish Sci 61: 727-728. Kristinsson HG, Rasco BA. 2000. Fish protein hydrolysates: production, biochemical, and functional properties. Crit Rev Food Sci Nutr 40: 43-81. Kristjansson MM, Gudmundsdottir S, Bjamason JB, Fox JW. 1995. Characterization of a collagenolytic serine proteinase from the Atlantic cod (Gadus morhua). Comp Biochem Physiol B 110: 707-717. Kuda T, Tsuda N, Yano T. 2004. Termal inactivation characteristics of acid and alkaline phosphotase in fish and shellfish. Food Chem 88: 543-548. Kumazawa Y. 2002. Development and application of transglutaminase for seafood products. Nippon Suisan Gakkaishi 68: 633-636. Kumazawa Y, Nakanishi K, Yasueda H, Motoki M. 1996. Purification and Characterization of Transglutaminase from Walleye Pollack Liver. Fish Sci 62: 959-964. Kumazawa Y, Sano K-I, Seguro K, Yasueda H, Nio N, Motoki M. 1997. Purification and Characterization of Transglutaminase from Japanese Oyster {Crassostrea gigas). J Agric Food Chem 45: 604-610. Kuraishi C, Yamazaki K, Susa Y. 2001. Transglutaminase: Its utilization in the food industry. Food Rev Intern 17: 221-246. Lee CM. 2000. Technology development for flavor production from seafood processing wastes. Lopetcharat K, Choi YJ, Park JW, Daeschel MA. 2001. Fish sauce products and manufacturing: A review. Food Rev Intern 17: 65-88. Makinodan Y, Yokoyama Y, Kinoshita M, Toyohara H. 1987. Characterization of an alkaline proteinase offish muscle. Comp Biochem Physiol 87: 1041-1046. Matsumiya M. 2004. Enzymatic production of N-Acetyl-D-Glucosamine using crude enzyme from the liver of squids. Food Sci Technol Res 10: 296-299. Matsumoto Y, Saucedo-Castaneda G, Revah S, Shirai K. 2004. Production of B-N- acehylhexosaminidase of Verticillium lecanii by solid state and submerged fermentations utilizing shrimp waste silage as substrate and inducer. Biochem 39:665-671. 152 Mekkes JR, Le Poole IC, Das PK, Kammeyer A, Westerhof W. 1997. In vitro tissue- digesting properties of krill enzymes compared with Fibrinolysin/DNAse, papain and placebo. Intern J Biochem Cell Biol 29: 703-706. Mireles-DeWitt CA, Morrissey MT. 2002. Parameters for the recovery of proteases from surimi wash water. Bioresource Technol 81: 241-247. Montero P, Avalos A, Perez-Mateos M. 2001. Characterization of polyphenol oxidase from prawns (Penaeus japonicus). Alternatives to inhibition: additives and high-pressure treatment. Food Chem 75: 317-324. Morrissey MT, Lin J, Ismond A. 2005. Waste management and by-product utilization, p. 279-323. In: Park JW (ed.). Surimi and Surimi Seafood. Taylor and Francis, New York. Nakagawa T, Ando M, Mainodan Y. 1996. Properties of glycogenolytic enzyme from fish muscle. Nippon Suisan Gakkaishi 62: 434-438. Nakagawa T, Makinodan Y, M. H. 1992. Distribution and properties of polyphenol oxidase from shrimp Penaeus spp. Kinki University Agriculture Department 25: 27-32. Nakai H, lizuka T, Fukukawa T, Nishioka K, Tanizawa S, Okuyama M, Mori H, Chiba S, Kimura A. 2005. Screening of enzymes having excellent functions from the digestive caecum of scallop. 21st COE program. Marine Bio- manipulation frontier for food production, Hokkaido, Japan. Natalia Y, Hashim R, Ali A, Chong A. 2003. Characterization of digestive enzymes in a carnivorous ornamental fish, the Asian bony tongue Scleropages formosus (Osteoglossidae). Aquacul 233: 305-320. Nayak J, Viswanathan NP, Ammu K, Mathew S. 2003. Lipase activity in different tissues of four fish. J Sci Food Agric 83: 1139-1142. Naz S. 2002. Enzymes and Food. Oxford, New York. Nielsen HH, Borrensen T. 1997. The influence of intestinal proteinases on ripening of salted herring, p. 293-303. In: Luten JB, Brresen T, Oehlenschlager J (eds.). Seafood from producer to consmer, integrated approch to quality. Elsevier Science, Amsterdam. Noguchi K, Ishikawa K, Yokoyama KI, Ohtsuka T, Nio N, Suzuki EL 2001. Crystal structure of red sea bream transglutaminase. J Biolog Chem 276. Normah I, Jamilah B, Saari N, Yaakob BCM. 2005. Optimization of hydrolysis conditions for the production of threadfm bream (Nemipterus japonicus) hydrolysate by alcalase. J Muscle Foods 16: 87-102. Nozawa H, Cho S-Y, Seki N. 2001. Purification and characterization of transglutaminase from squid gill. Fish Sci 67: 912-919. Nozawa H, Mamegoshi S, Seki N. 1997. Partial purification and characterization of six transglutaminases from ordinary muscles of various fishes and marine invertebrates. Comp Biochem Physiol 118: 313-317. Nozawa H, Seki N. 2001. Purification of transglutaminase from scallop striated adductor muscle and NaCl-induced inactivation. Fish Sci 67: 493-499. Nunes ML, Campos RM, Batista I. 1997. Sardine ripening: evolution of enzymatic, seasonal and biochemical aspects, p. 319-329. In: Luten JB, Borrensen T, 153 Oehlenschlager J (eds.). Seafood from producer to consumer, integrated approach to quality. Elsvier Science, Amsterdam. Oh ES, Kim DS, Kim JH, Kim HR. 2000. Enzymatic properties of a protease from the hepatopancreas of shrimp, Penaeus orientalis. J Food Biochem 24: 251-264. Olsen R, Overbo Mymes B. 1991. Alkaline phosphatase from the hepatopancrease of shrimp (Pandalus borealis), a dimmer enzyme with catalytically active subunits. Comp Biochem Physiol 99B: 755-761. Olsen RL, Johansen A, Mymes B. 1990. Recovery of enzymes from shrimp waste. Process Biochem 25: 67-68. Olsen SO, Skara T. 1997. Chemical changes during ripening of North Sea herring, p. 305-317. In: Luten JB, Borresen T, Oehlenschlager J (eds.). Seafood from producer to consumer, integrated approach to quality. Elsvier Science, Amsterdam. Opoku-Gyamfua A, Simpson BK, Squires EJ. 1992. Comparative studies on the polyphenoloxidase fraction from lobster and tyrosinase. J Agric Food Chem 40: 772-775. Pangkey H, Hara K, Tachibana K, Cao M-J, Osatomi K, Ishihara T. 2000. Purification and characterization of cathepsin S from hepatopancreas of carp Cyprinus carpio. Fish Sci 66: 1130-1137. Park PJ, Lee SH, Byun HG, Kim SH, Kim SK. 2002. Purification and characterization of a collagenase from the mackerel, scomber japonicus. J Biochem Mol Biol 35: 576-582. Raa J. 1996. Biotechnology and aquaculture and the fish processing industry: A success story in Norway Tromso, Norway, 15. Raa J. 1997. New commercial products from waste of the fish processing industry. Making the most of the catch. Proceedings of the Seafood Symposium, Brisbane, Australia, p. 33-36. Roy I, Gupta MN. 2004. Freeze-drying of proteins: Some emerging concerns. Biotechnol Appl Biochem 39: 165-177. Roy P, Colas B, Durand P. 1996. Purification, kinetical and molecular characterizations of a serine collagenolytic protease from greenshore crab (Carcinus maenas) digestive gland. Comp Biochem Physiol 115B: 87-95. Sabrowski R, Shaling G, Navarette dTM, Walter I, Garcia-Carreno FL. 2004. Stability and effects of organic solvents on endopeptidases from the gastric fluid of the marine crab Cancer pagurus. J Mol Catal 30. Sambrook J, Russell DW. 2001. Molecular cloning: A Laboratory Manual. Laboratory Press, Cold Springs Harbor, New York. Sathivel S, Babbitt J, Smiley S, Crapo C, Reppond KD, Prinyawiwatkul W. 2003. Biochemical and functional properties of herring (Clupea harengus) byproduct hydrolysates. J Food Sci 68: 2196-2200. Shahidi F, Han XQ, Synowiecki J. 1995. Production and characteristics of protein hydrolysates from capelin (Mallotus villosus). Food Chem 53: 285-293. Shahidi F, Kamil YVA. 2001. Enzymes from fish and aquatic invertebrates and their application in the food industry. Trends Food Sci Technol 12: 435-464. 154 Shahidi F, Synowiecki J. 1997. Protein hydrolysate from seal meat as phosphate alternative in food processing application. Food Chem 60: 29-32. Sjodahk J, Emmer A, Vincent J, Roeraade J. 2002. Charcterization of proteinases from Antarctic krill (Euphausia superba). Protein Expr Purif 26: 153-161. Skea LG, Mountfort OD, Clements DK. 2005. Gut carbohydrases from the New Zealand marine herbivorous fishes Kyphosus sydneyanus (Kyphosidae), Aplodactylus arctidens (Aplodactylidae) and Odaxpullus (Labridae). Comp Biochem Physiol Part B 140: 259-269. Slizyte R, Dauksas E, Falch E, Storro I. 2005. Characteristics of protein fractions generated from hydrolysed cod (Gadus morhua) by-products. Process Biochem 40: 2021-2033. Slizyte R, Van NJ, Rustad T. 2004. Hydrolysis of cod (Gadus morhua) by-products: influence of initial heat inactivation, concentration and separation conditions. J Aquat Food Prod Techno 113: 31-48. Sovik SL, Rustad T. 2004. Seasonal changes in trypsin and chymotrypsin activity in viscera from cod species. J Aquatic Food Product Technol 13: 13-30. Sovik SL, Rustad T. 2005. Proteolytic activity in byproducts from cod species caught at three different fishing grounds. J Agric Food Chem 53: 452-458. Sumaya-Martinez T, Castillo-Morales A, Favela-Torres E, Huerta-Ochoa S, Prado- Barragan LA. 2005. Fish protein hydrolysates from gold carp (Carassius auratus). J Sci Food Agric 85: 98-104. Suppasith k, Sootawat B, Wonnop V. 2004. Comparative studies on proteolytic activity of splenic extract from three tuna species commonly used in Thailand. J Food Biochem 28: 355-372. Suresh PV, Chandrasekaran M. 1998. Utilization of prawn waste for chitinase production by the marine fungus Beauveria bassiana by solid state fermentation. World J Micro Biochem 14. Taniguchi A, Takano K. 2001. Purification and properties of carboxypeptidase from tilapia intestine-digestive enzyme of tilapia VIII. Nippon Suisan Gakkaishi 67: 1096-1102. Taniguchi A, Takano K. 2002. Purification of properties of aminopeptidase from Tilapia intestine digestive enzyme of tilapia-IX. Nippon Suisan Gakkaishi 68: 382-388. Taniguchi A, Takano K, Kamoi I. 2001. Purification and properties of lipase from tilapia stomach -digestive enzyme of tilapia-VII-. Nippon Suisan Gakkaishi 67: 96-101. Teruel SRL, Simpson BK. 1995. Characterization of the collagenolytic enzyme fraction from winter flounder (Pseudopleuronectes americanus). Comp Biochem Physiol 112: 131-136. Tsao CY, Hsu YH, Chao LM, Jiang ST. 2004. Purification and characterization of three amylases from viscera of hard clam Meretrix lusoria. Fish Sci 70: 174- 182. Tschersich P, Choudhury GS. 1998. Arrowtooth flounder (Atheresthes stomias) protease as a processing aid. J Aquat Food Prod Technol 7: 77-89. 155 Tungkawachara S, Park JW, Choi YJ. 2003. Biochemical properties and consumer acceptance of Pacific whiting fish sauce. J Food Chem 68: 855-860. Uchiyama S, Nozaki I. 2005. Partial purification and characterization of pro- phospholipase A2 activating proteases from gill membranes of the red sea bream Chrysophrys major. Comp Biochem Physiol 141B: 121-127. Van Wormhoudt A, Le Chevalier P, Sellos D. 1992. Purification, biochemical characterization and N-terminal sequence of a serine-protease with chymotrypsic and collagenolytic activities in a tropical shrimp, Penaeus vannamei (Crustacea, Decapoda). Comp Biochem Physiol 103B: 675-680. Vannuccini S. 2004. Overview offish production, utilization, consumption and trade based on 2002 data Fishery information, data and statistics unit Food and Agriculture Organization of United Nations. Vecchi SD, Coppes Z. 1996. Marine Fish Digestive Proteases-Relevance to Food Industry and the South-West Atlantic Region- A Review. J Food Biochem 20: 193-214. Wilson DD, Choudhury GS. 2004. Modulation and inhibition of arrowtooth flounder (Atheresthes stomias) protease. J Aquat Food Prod Technol 13: 24-45. Worratao A, Yongsawatdigul J. 2003. Cross-linking of actomyosin by crude tilapia (Oreochromis niloticus) transglutaminase. J Food Biochem 27: 35-51. Worratao A, Yongsawatdigul J. 2005. Purification and characterization of transglutaminase from tropical tilapia {Oreochromis niloticus). Food Chem 93: 651-658. Wray T. 1988. Fish processing: new uses for enzymes. Food Manufac 63: 64-65. Wu HC, Shiau CY, Chen HM. 2003. Free amino acids and peptides as related to antioxidant properties in protein hydrolysates of mackerel {Scomber austriasicus). Food Res Int 36: 949-957. Xu RA, Wong RJ, Rogers ML, Fletcher GC. 1996. Purification and characterization of acidic proteases from the stomach of the deepwater fmfish orange roughy {Hoplostethus Atlanticus). J Food Biochem 20: 31-48. Yamashita M, Konagaya S. 1995. Purification and characterization of cathepsin L from the white muscle of chum salmon, Oncorhynchus keta. Comp Biochem Physiol B 111: 587-596. Yasueda H, Kumazawa Y, Motoki M. 1994. Purification and characterization of a tissue-type transglutaminase from red sea bream {Pagrus major). Biochimica Biophysica 58: 2041-2045. Yoshinaka R, Sato M, Ikeda S. 1972. Studies on collagenase of fish-I. Existence of collagenolytic enzyme in pyloric caeca of seriola quinqueradiata. Bull Jap Soc Sci Fish 39: 275-281. Yoshitomi B. 2005. Seasonal variation of crude digestive protease activity in Antarctic krill Euphausia Superba. Fish Sci 71. Zhou P, Qi X, Zheng X. 2004. Purification and some properties of cuttlefish ink polyphenol oxidase, p. 223-241. In: Sakaguchi M (ed.). More efficient utilization of fish and fisheries products. Elesvier, New York.